Author Response

Reviewer #2 (Public Review):

Zylbertal and Bianco propose a new model of trial-to-trial neuronal variability that incorporates the spatial distance between neurons. The 7-parameter model is attractive because of its simplicity: A neuron's activity is a function of stimulus drive, neighboring neurons, and global inhibition. A neuroscientist studying almost any brain area in any model organism could make use of this model, provided that they have access to 1) simultaneously-recorded neurons and 2) the spatial locations of those neurons. I could foresee this model being the de-facto model to compare to all future models, as it is easy to code up and interpret. The paper explores the effectiveness of this distance model by modeling neural activity in the zebrafish optic tectum. They find that this distance-based model can capture 1) bursting found in spontaneous activity, 2) ongoing co-fluctuations during stimulus-evoked activity, and 3) adaptation effects during prey-catching behavior.

Strengths:

The main strength of the paper is the interpretability of the distance-based model. This model is agnostic to the brain area from which the population of neurons is recorded, making the model broadly applicable to many neuroscientists. I would certainly use this model for any baseline comparisons of trial-to-trial variability.

The model is assessed in three different contexts, including spontaneous activity and behavior. That the model provides some prediction in all three contexts is a strong indicator that this model will be useful in other contexts, including other model organisms. The model could reasonably be extended to other cognitive states (e.g., spatial attention) or accounting for other neuron properties (such as feature tuning, as mentioned in the manuscript).

The analyses and intuition to show how the distance-based model explains adaptation were insightful and concise.

We thank the reviewer for these supportive comments.

Weaknesses:

Model evaluation and comparison: The paper does not fully evaluate the model or its assumptions; here, I note details in which evaluation is needed. A key assumption of the model - that correlations fall off in a gaussian manner (Fig. 1C-E - is not supported by Fig. 1C, which appears to have an exponential fall-off. Functions other than gaussian may provide better fits.

A key feature of our model is that connection strengths smoothly decrease with distance. However, we did not intend to make strong claims about the exact function parametrizing this distance relationship. In light of the reviewer’s comment, we have additionally tested an exponential function and find that it too can describe activity correlations in OT with a negligible decrease in r2 (Figure 1 – figure supplement 1A-C). The main purpose of the analysis was to show that the correlation is maximal around the seed and decays uniformly with distance from it (i.e. no sub-networks or cliques are detected). We have emphasized this in a revised conclusion paragraph and note that while multiple functions can be used to parameterize the relationship, they are nonetheless certainly simplifications. Secondly, we also ran a version of the network simulation where the connections decay in space according to an exponential rather than Gaussian function and show that, as expected, tectal bursting is robust to this change.

Furthermore, it is not clear whether the r^2s in Fig. 1E are computed in a held-out manner (more details about what goes into computing r^2 are needed).

These values are computed by fitting the 2-d Gaussian (or exponential function) to all neurons excluding the seed itself (added a short clarification in the Methods).

Assessing the model based on peak location alone (Fig. 1E) is not sufficient, as other smooth monotonically-decreasing functions may perform similarly.

As discussed above, an exponential function indeed performs similarly to a Gaussian. However, goodness of fit is secondary to the main aim of Fig 1E, which is to show that the correlation peak tends to fall near the seed cell.

Simulating from the model greatly improves the reader's understanding (Fig. 2D), but no explanation is given for why the simulations (Fig. 2D) have almost no background spikes and much fewer, non-co-occurring bursts than those of real data (Fig. 2E).

In part this is because the simulation results depicted in Fig 2D were derived from the ‘baseline model’, prior to optimizing to match biological bursting statistics. It is thus expected that activity will differ from experimental observation and was our main motive to tune the model parameters (now emphasized in the text). However, the model will certainly not account for all aspects of tectal activity; rather, it was designed to reproduce bursting as a prominent feature of ongoing activity and in the second part of the paper we explore the extent to which it can account for other phenomena. As noted above, in the revised abstract, introduction and discussion we have tried to clarify the motivation for developing the model and how it was used to gain insight into activity-dependent changes in network excitability.

A key assumption of the distance model (Fig. 2A) is that each neuron has the same gaussian fall-off (i.e., sigma_excitation and sigma_inhibition), but it is unclear if the data support this assumption.

We intentionally opted for a simple model (i.e. described by few parameters), in part due to the lack of connectivity data and additionally to set a lower bound on the extent to which multiple features of tectal activity could be accounted for. More complex models with additional degrees of freedom (such as cell-specific connectivity) may well describe the data better, but likely at the cost of interpretability. We consider such extensions are beyond the scope of the present study but might be fruitful avenues for future research.

Although an excitatory and inhibitory gain is assumed (Fig. 2A), it is not clear from the data (Fig. 1C) that an inhibitory gain is needed (no negative correlations are observed in Fig. 1C-D).

This is now explored in the revised Figure 3A which includes the condition of zero inhibition gain. See also response to reviewer 1.

After optimization (Fig. 3), the model is evaluated on predicting burst properties but not evaluated on predicting held-out responses (R^2s or likelihoods), and no other model (e.g., fitting a GLM or a model with only an excitatory gain) is considered. In particular, one may consider a model in which "assemblies" do exist - does such an assembly model lead to better held-out prediction performance?

The model we developed is a mechanistic, generative model. In contrast to Pillow et al 2008, we did not fit the model to data but rather we used it to simulate network activity and tuned the seven parameters (using EMOO) to best match biological observations. Thus, rather than assessing goodness-of-fit using cross-validation, our approach involved comparison of summary statistics related to the target emergent phenomenon (tectal bursting). This was necessary as bursting appears highly stochastic. Further to the comments above, we have expanded the parameter space to include instances with only an excitatory gain (where bursting failed) and no distance-dependence (again, busting failed). Introducing assemblies into the model will inevitably support bursting (and introduce many more free parameters), but one of our key observations is that such assemblies are not required for this aspect of spontaneous activity. Again, our aim was not to produce a detailed picture of tectal connectivity, but rather to develop a minimal model and estimate the extent to which it can account for observed features of activity. Note that the second half of the paper (Figure 4 onwards) shows the model can explain phenomena that were not considered during parameter tuning.

It is unclear why a genetic algorithm (Fig. 1A-C) is necessary versus a grid search; it appears that solutions in Generation 2 (Fig. 3C, leftmost plot, points close to the origin) are as good as solutions in Generation 30 and that the spreads of points across generations do not shrink (as one would expect from better mutations). Given the small number of parameters (7), a grid search is reasonable, computationally tractable, and easier to understand for all readers (Fig. 3A).

Perhaps in hindsight a grid search would have worked, but at increased computational cost (each instantiation of the model is computationally expansive). At the time we chose EMOO, and since it produced satisfactory results, we kept it. As often happens with multi-objective optimization, an improvement in one objective usually happens at the expense of other objectives, so the spread of the points does not shrink much but they move closer to the axes (i.e. reduced error). The final parameter combination is closer to the origin than any point in generation 2, though admittedly not by much. Importantly, however, optimizing the model using the training features generalized to other burst-related statistics.

It is unclear why the excitatory and inhibitory gains of the temporal profiles (Fig. 3I) appear to be gaussian but are formulated as exponential (formula for I_ij^X in Methods).

The interactions indeed have exponential decay in time. These might appear Gaussian because the axis scale is logarithmic.

Overall, comparing this model to other possible (similar) models and reporting held-out prediction performance will support the claim that the distance model is a good explanation for trial-to-trial variability.

See comments above. A key point we want to stress is that we intentionally explored a minimal network model and found that, despite obvious simplifications of the biology, it was nonetheless able to explain multiple aspects of tectal physiology and behaviour. We hope that it inspires future studies and can be extended, in parallel to experimental findings, to more accurately represent the cell-type diversity and cell-specific connectivity of the tectal network.

Data results: Data results were clear and straightforward. However, the explanation was not given for certain results. For example, the relationship between pre-stimulus linear drive and delta R was weak; the examples in Fig. 4C do not appear to be representative of the other sessions. The example sessions in Fig. 4C have R^2=0.17 and 0.19, the two outliers in the R^2 histogram (Fig. 4D).

The revised figure 4 is based on new data and new analysis (see below), and the presented examples no longer represent the extreme tail of the distribution (they still, however, represent strong examples, as is now explicitly indicated in the figure legend).

The black trace in Fig. 4D has large variations (e.g., a linear drive of 25 and 30 have a change in delta R of ~0.1 - greater than the overall change of the dashed line at both ends, ~0.08) but the SEMs are very tight. This suggests that either this last fluctuation is real and a major effect of the data (although not present in Fig. 4C) or the SEM is not conservative enough. No null distribution or statistics were computed on the R^2 distribution (Fig. 4C, blue distribution) to confirm the R^2s are statistically significant and not due to random fluctuations.

We agree that this was not sufficiently robust and in response to this comment we undertook a significant revision to figure 4 and the associated text:

i) The revised figure is based on an entirely new dataset, allowing us to verify the results on independent data. We used 5 min ISI for all stimulus presentations, regardless of stimulus type (high or low elevation), thus ensuring that we are only examining differences in state brought about by previous ongoing activity, without risk of ‘contamination’ by evoked activity.

ii) As per the reviewer’s suggestion, we compared model-estimated pre-stimulus state to a null estimate using randomly sampled time-points. We additionally compared the optimised model with the baseline model. Whereas the null (random times) estimates had no predictive power, both models using pre-stimulus activity were able to explain a fraction of the response residuals with the optimised model performing better.

iii) We refined the binning process by first computing, for each response, the mean of response residuals across neurons for each bin of estimated linear drive, and then averaging across responses. This prevents the relationship being skewed by rare instances involving unusually large numbers of neurons for a particular linear drive bin, and thereby eliminates the fluctuations the reviewer was referring to.

The absence of any background activity in Fig. 6B (e.g., during the rest blocks) is confusing, given that in spontaneous activity many bursts and background activity are present (Fig. 2E).

The raster only presents evoked responses and no background activity is shown. This has been clarified in the revised figure and legend.

Finally, it appears that the anterior optic tectum contributes to convergent saccades (CS) (Fig. 7E) but no post-saccadic activity is shown to assess how activity changes after the saccade (e.g., plotting activity from 0 to 60).

Activity before and after the saccade is shown in Fig 7A. Fig 7E shows the ‘linear drive’ (or ‘excitability’), and how it changes leading up to the saccade. Since we were interested in the association between pre-saccade state and saccade-associated activity, we did not plot post-saccadic linear drive. However, as can be seen in the below figure for the reviewer, linear drive is strongly suppressed by the saccade, as expected due to CS-associated activity.

No explanation is given why activity drops ~30 seconds before a convergent saccade (Fig. 7E).

This is no longer shown after we trimmed the history data in Fig 7E in accordance with a comment from reviewer 1. We speculate, however, that the mean linear drive of a compact population of neurons would be somewhat periodical, since a high linear drive leads to a burst which results in a prolonged inhibition (low linear drive) with a slow recovery and so on.

No statistical test is performed on the R^2 distribution (Fig. 7H) to confirm the R^2s (with a mean close to R^2=0.01) are meaningful and not due to random fluctuations.

We revised the analysis in Fig 7 along the same lines as the revision of Fig 4. Model-estimated linear drive predicts CS-associated activity whereas a null estimate (random times) shows no such relationship.

Presentation: A disjointed part of the paper is that for the first part (Figs. 1-3), the focus is on capturing burst activity, but for the second part (Figs. 4-7), the focus is on trial-to-trial variability with no mention of bursts. It is unclear how the reader should relate the two and if bursts serve a purpose for stimulus-evoked activity.

In the first part of the paper (Figs. 1-3), we use ongoing activity to develop an understanding (formulated as a network model) of how activity modulates the network state. In the second part, we test this understanding in the context of evoked responses and show that model-estimated network state explains a fraction of visual response variability and experience-dependent changes in activity and behaviour. In the revised MS we further emphasize this idea and have edited the results text to strengthen the connections between these parts of the study. See also comments above.

Citations: The manuscript may cite other relevant studies in electrophysiology that have investigated noise correlations, such as:

• Luczak et al., Neuron 2009 (comparing spontaneous and evoked activity).

• Cohen and Kohn, Nat Neuro 2011 (review on noise correlations).

• Smith and Kohn, JNeurosci 2008 (looking at correlations over distance).

• Lin et al., Neuron 2015 (modeling shared variability).

• Goris et al., Nat Neuro 2014 (check out Fig. 4).

• Umakantha et al., Neuron 2021 (links noise correlation and dim reduction; includes other recent references to noise correlations).

We agree that the manuscript could benefit from citing some of these suggested studies and have added citations accordingly.

Author Response

Reviewer #1 (Public Review):

It is well established that valuation and value-based decision-making is context-dependent. This manuscript presents the results of six behavioral experiments specifically designed to disentangle two prominent functional forms of value normalization during reward learning: divisive normalization and range normalization. The behavioral and modeling results are clear and convincing, showing that key features of choice behavior in the current setting are incompatible with divisive normalization but are well predicted by a non-linear transformation of range-normalized values.

Overall, this is an excellent study with important implications for reinforcement learning and decision-making research. The manuscript could be strengthened by examining individual variability in value normalization, as outlined below.

We thank the Reviewer for the positive appreciation of our work and for the very relevant suggestions. Please find our point-by-point answer below.

There is a lot of individual variation in the choice data that may potentially be explained by individual differences in normalization strategies. It would be important to examine whether there are any subgroups of subjects whose behavior is better explained by a divisive vs. range normalization process. Alternatively, it may be possible to compute an index that captures how much a given subject displays behavior compatible with divisive vs. range normalization. Seeing the distribution of such an index could provide insights into individual differences in normalization strategies.

Thank you for pointing this out, it is indeed true that there is some variability. To address this, and in line with the Reviewer’s suggestion, we extracted model attributions per participant on the individual out-of-sample log-likelihood, using the VBA_toolbox in Matlab (Daunizeau et al., 2014). In experiment 1 (presented in the main text), we found that the RANGE model accounted for 79% of the participants, while the DIVISIVE model accounted for 12%. The relative difference was even higher when including the RANGEω model in the model space: the RANGE and RANGEω models account for a total of 85% of the participants, while the DIVISIVE model accounted only for 5%.

In experiment 2 (presented in the supplementary materials), the results were comparable (see Figure 3-figure supplement 3: 73% vs 10%, 83% vs 2%).

To provide further insights into the behavioral signatures behind inter-individual differences, we plotted the transfer choice rates for each group of participants (best explained by the RANGE, DIVISIVE, or UNBIASED models), and the results are similar to our model predictions from Figure 1C:

Author Response Image 1. Behavioral data in the transfer phase, split over participants best explained by the RANGE (left), DIVISIVE (middle) or UNBIASED (right) model in experiment 1 (A) and experiment 2 (B) (versions a, b and c were pooled together).

To keep things concise, we did not include this last figure in the revised manuscript, but it will be available for the interested readers in the Rebuttal letter.

One possibility currently not considered by the authors is that both forms of value normalization are at work at the same time. It would be interesting to see the results from a hybrid model. R1.2 Thank you for the suggestion, we fitted and simulated a hybrid model as a weighted sum between both forms of normalization:

First, the HYBRID model quantitatively wins over the DIVISIVE model (oosLLHYB vs oosLLDIV : t(149)=10.19, p<.0001, d=0.41) but not over the RANGE model, which produced a marginally higher log-likelihood (oosLLHYB vs oosLLRAN : t(149)=-1.82, p=.07, d=-0.008). Second, model simulations also suggest that the model would predict a very similar (if not worse) behavior compared to the RANGE model (see figure below). This is supported by the distribution of the weight parameter over our participants: it appears that, consistently with the model attributions presented above, most participants are best explained by a range-normalization rule (weight > 0.5, 87% of the participants, see figure below). Together, these results favor the RANGE model over the DIVISIVE model in our task.

Out of curiosity, we also implemented a hybrid model as a weighted sum between absolute (UNBIASED model) and relative (RANGE model) valuations:

Model fitting, simulations and comparisons slightly favored this hybrid model over the UNBIASED model (oosLLHYB vs oosLLUNB: t(149)=2.63, p=.0094, d=0.15), but also drastically favored the range normalization account (oosLLHYB vs oosLLRAN : t(149)=-3.80, p=.00021, d=-0.40, see Author Response Image 2).

Author Response Image 2. Model simulations in the transfer phase for the RANGE model (left) and the HYBRID model (middle) defined as a weighted sum between divisive and range forms of normalization (top) and between unbiased (no normalization) and range normalization (bottom). The HYBRID model features an additional weight parameter, whose distribution favors the range normalization rule (right).

To keep things concise, we did not include this last figure in the revised manuscript, but it will be available for the interested readers in the Rebuttal letter.

Reviewer #2 (Public Review):

This paper studies how relative values are encoded in a learning task, and how they are subsequently used to make a decision. This is a topic that integrates multiple disciplines (psych, neuro, economics) and has generated significant interest. The experimental setting is based on previous work from this research team that has advanced the field's understanding of value coding in learning tasks. These experiments are well-designed to distinguish some predictions of different accounts for value encoding. However there is an additional treatment that would provide an additional (strong) test of these theories: RN would make an equivalent set of predictions if the range were equivalently adjusted downward instead (for example by adding a "68" option to "50" and "86", and then comparing to WB and WT). The predictions of DN would differ however because adding a low-value alternative to the normalization would not change it much. Would the behaviour of subjects be symmetric for equivalent ranges, as RN predicts? If so this would be a compelling result, because symmetry is a very strong theoretical assumption in this setting.

We thank the Reviewer for the overall positive appraisal concerning our work, but also for the stimulating and constructive remarks that we have addressed below. At this stage, we just wanted to mention that we also agree with the Reviewer concerning the fact that a design where we add "68" option to "50" and "86" would represent also an important test of our hypotheses. This is why we had, in fact, run this experiment. Unfortunately, their results were somehow buried in the Supplementary Materials of our original submission and not correctly highlighted in the main text. We modified the manuscript in order to make them more visible:

Behavioral results in three experiments (N=50 each) featuring a slightly different design, where we added a mid value option (NT68) between NT50 and NT87 converge to the same broad conclusion: the behavioral pattern in the transfer phase is largely incompatible with that predicted by outcome divisive normalization during the learning phase (Figure 2-figure supplement 2).

Reviewer #3 (Public Review):

Bavard & Palminteri extend their research program by devising a task that enables them to disassociate two types of normalisation: range normalisation (by which outcomes are normalised by the min and max of the options) and divisive normalisation (in which outcomes are normalised by the average of the options in ones context). By providing 4 different training contexts in which the range of outcomes and number of options vary, they successfully show using 'ex ante' simulations that different learning approaches during training (unbiased, divisive, range) should lead to different patterns of choice in a subsequent probe phase during which all options from the training are paired with one another generating novel choice pairings. These patterns are somewhat subtle but are elegantly unpacked. They then fit participants' training choices to different learning models and test how well these models predict probe phase choices. They find evidence - both in terms of quantitive (i.e. comparing out-of-sample log-likelihood scores) and qualitative (comparing the pattern of choices observed to the pattern that would be observed under each mode) fit - for the range model. This fit is further improved by adding a power parameter which suggests that alongside being relativised via range normalisation, outcomes were also transformed non-linearly.

I thought this approach to address their research question was really successful and the methods and results were strong, credible, and robust (owing to the number of experiments conducted, the design used and combination of approaches used). I do not think the paper has any major weaknesses. The paper is very clear and well-written which aids interpretability.

This is an important topic for understanding, predicting, and improving behaviour in a range of domains potentially. The findings will be of interest to researchers in interdisciplinary fields such as neuroeconomics and behavioural economics as well as reinforcement learning and cognitive psychology.

We thank Prof. Garrett for his positive evaluation and supportive attitude.

Author Response

Reviewer #1 (Public Review):

While the mechanism about arm-races between plant and specialist herbivores has been studied, such as detoxification of specific secondary metabolites, the mechanism of the wider diet breadth, so-called generalist herbivores have been less studied. Since the heterogeneity of host plant species, the experimental validation of phylogenetic generalism of herbivores seemed as hard to be conducted. The authors declared the two major hypotheses about the large diet breadth ("metabolic generalism" and "multi-host metabolic specialism"), and carefully designed the experiment using Drosophila suzukii as a model herbivore species.

By an untargeted metabolomics approach using UHPLC-MS, authors attempted to falsify the hypotheses both in qualitative- and quantitative metabolomic profiles. Intersections of four fruit (puree) samples and each diet-based fly individual samples from the qualitative data revealed that there were few ions that occur as the specific metabolite in each diet-based fly group, which could reject the "multi-host metabolic specialism" hypothesis. Quantitative data also showed results that could support the "metabolic generalism" hypothesis. Therefore, the wide diet breadth of D. suzukii seemed to be derived from the general metabolism rather than the adaptive traits of the diverse host plant species. On the other hand, the reduction of the metabolites (ions) set using GLM seemed logical and 2-D clustering from the reduced ions set showed that quantitative aspects of diet-associated ions could classify "what the flies ate". These interesting results could enhance the understanding of the diet breadth (niche) of herbivorous insects.

The authors' approach seemed clear to falsify the hypotheses based on the appropriate data processing. The intersection of shared ions from the qualitative dataset could distinguish the diet-specific metabolites in flies and commonly occurring metabolites among flies and/or fruits. Also, filtering on the diet-specific ions seemed to be a logical and appropriate way. Meanwhile, the discussion about the results seemed to be focused on different points regarding the research hypotheses which were raised in the introduction part. Discussion about the results mainly focused on the metabolism of D. suzukii itself, rather than the research hypotheses and questions that were raised from the evolution of the wide diet breadth of generalist herbivores. In particular, the conclusion seems to be far from the main context of the authors' research; e.g. frugivory. It makes the implication of the study weaker.

We wish to thank Reviewer #1 for their appreciation of our study. As recommended, we now focus our discussion more on the general aspect of our findings (relevant to insects, herbivores, or frugivores), and less on the peculiarities of the metabolism of D. suzukii itself. Specifically, we now only mention D. suzukii in one section (two sentences) of our Discussion, to serve as an example (l.387-396). Thanks to this comment, the Discussion may interest a broader readership, on the evolution of diet breadth in generalist herbivorous species and offers a better understanding of the general implications of our findings.

Reviewer #2 (Public Review):

The manuscript: "Metabolic consequences of various fruit-based diets in a generalist insect species" by Olazcuaga et al., addresses an interesting question. Using an untargeted metabolomics approach, the authors study how diet generalism may have evolved versus diet specialization which is generally more commonly observed, at least in drosophila species. Using the phytophagous species Drosophila suzukii, and by directly comparing the metabolomes of fruit purees and the flies that fed on them, the authors found evidence for "metabolic generalism". Metabolic generalism means that individuals of a generalist species process all types of diet in a similar way, which is in contrast to "multi-host metabolic specialism" which entails the use of specific pathways to metabolize unique compounds of different diets. The authors find strong evidence for the first hypothesis, as they could easily detect the signature of each fruit diet in the flies. The authors then go on to speculate on the evolutionary ramifications of this for how potentially diet specializations may have evolved from diet generalism. Overall, the paper is well written, the experiments well documented, and the conclusions convincing.

We thank Reviewer #2 for their comments and appreciation of our work.

Reviewer #3 (Public Review):

Laure Olazcuaga et al. investigated the metabolomes of four fruit-based diets and corresponding individuals of Drosophila suzukii that reared on them using comparative metabolomics analysis. They observed that the four fruit-based diets are metabolically dissimilar. On the contrary, flies that fed on them are mostly similar in their metabolic response. From a quantitative point of view, they find that part of the fly metabolomes correlates well with that of the corresponding diet metabolomes, which is indicative of insect ingestive history. By further focusing on 71 metabolites derived from diet-specific fly ions and highly abundant fruit ions, the authors show that D. suzukii differentially accumulates diet metabolism in a compound-specific manner. The authors claim that the data support the metabolic generalism hypothesis while rejecting the multi-host metabolic specialism hypothesis. This study provides a valuable global chemical comparison of how diverse diet metabolites are processed by a generalist insect species.

Strengths:

The rapid advances in high-resolution mass spectrometry have recently accelerated the discovery of many novel post-ingestive compounds through comparative metabolomics analysis of insect/frass and plant samples. Untargeted metabolomics is thus a very powerful approach for the systematic comparison of global chemical shifts when diverse plant-derived specialized metabolites are further modified or quantitatively metabolized after ingestion by insects. The technique can be readily extended to a larger micro- or macro-evolutionary context for both generalist and specialist insects to systematically investigate how plant chemical diversity contributes to dietary generalism and specialism.

We would like to thank Reviewer #3 for their insightful comments on the power of untargeted metabolomics to evaluate the fate of plant metabolites and their use by herbivores. We also agree that these techniques can be used to tackle eco-evolutionary issues, such as the origin and maintenance of dietary generalism and specialism here. We hope that our study will inspire other researchers to explore such techniques and experiments to gain a global overview of biochemistry fluxes and their evolution. We now mention it in the conclusion (L454-459).

Weaknesses:

The authors claim that their data support the hypothesis of metabolic generalism, however, a total analysis of insect metabolism may not generate a clean dataset for direct comparison of fruit-derived metabolites with those metabolized by D. suzukii, given that much of these metabolites would be "diluted" proportionally by insect-derived metabolites. If the insect-derived metabolites predominate, then, as the authors observed, a tight clustering of D. suzukii metabolomes in the PCA plot would be expected. It is therefore very difficult to interpret these patterns.

We agree with Reviewer #3 that a careful examination of the different possible origins of metabolites should take place to distinguish between our two competing hypotheses.

The only source of metabolites for insects in our experimental setup is a mixture of (i) a large proportion of fruit purees and (ii) a minor proportion of artificial medium consisting mainly of yeast. Our goal is thus to understand the fate of (i) “fruit-derived” metabolites (transformed and untransformed), while controlling for (ii) “artificial media-derived” metabolites, that constitute a nuisance signal but are necessary for a complete development in our system.

By “fruit-derived” and “insect-derived” metabolites, it is our understanding that Reviewer #3 means “fruit” metabolites (when in insects, untransformed “fruit-derived” metabolites) and “artificial medium-derived” metabolites. It is true that we do wish to avoid a predominance of “artificial medium-derived” metabolites and focus on “fruit-derived” metabolites in insects. We also want to note that it is of primary importance in our study to distinguish between “fruit” metabolites that are carried as is (“fruit” metabolites present in insects, ie untransformed “fruit-derived” metabolites), and “fruit” metabolites that are used after transformation by the insect (i.e., transformed “fruit-derived” metabolites).

We agree with Reviewer #3 that the presence of “artificial medium-derived” metabolites could be problematic in direct comparisons of fruits and insects (and not among fruits or among insects’ comparisons).

However, we took some steps to avoid such problems:

1. We included control fly samples in our experiment: at each experimental generation, flies developed only on artificial medium (without fruit puree) were collected and processed simultaneously with flies that developed on fruit media. Results using these artificial medium-reared flies as controls (by subtracting their ions levels and removing ions that were similar, respective of their generation) were similar to results using raw data and conclusions were identical (see below).

2. We lowered the proportion of artificial medium in our fruit media so that it was kept to a minimum, compatible with larval development and adult survival.

Consistent with the low impact of this “artificial medium” component on our conclusions, we also wish to point out the presence pattern of metabolites found only in flies and never in fruits when using raw data (Figure 3, yellow stack). Even in the most conservative hypothesis of 100% of these metabolites originating from our artificial medium (which is probably not the case), we observe that it constitutes only a minor proportion of metabolites common to all flies (15.7%).

For your consideration, we include below the main Figures, using both raw data and artificial medium-controlled:

Figure 2, left = raw data; right = artificial-media controlled:

Figure 3, left = raw data; right = artificial-media controlled:

Figure 3S1, left = raw data; right = artificial-media controlled:

Figure 4, above = raw data; below = artificial-media controlled:

We hope that we convinced the Editor/Reviewers that raw data and artificial-medium controlled data provide a single and same answer to all our analyses. We chose to present only raw data, to simplify the Materials & Methods section.

We however modified the current version of the manuscript to inform the reader that proper controls were done and that their inclusion do not modify any of our conclusions (l.110-113 and l.583-589).

We also wish to point out two additional comments:

• As Reviewer #1 also recommended, we modified the expectations drawn in Fig1G to better consider the general comment of “insect derived” metabolites being fundamentally different from plant metabolites (even if we do show in our study that only approx. 9% of metabolites are private to flies).

• The main part of our care in the use of this global PCA analysis is that it follows two other analyses (global intersection and comparison of intersections among fruits and among flies) and precedes another one (fly-focused PCA). We hope that all these analyses help the readers get a comprehensive overview of the dataset and associated results, avoiding reliance on a single analysis.

• We also help readers to explore and visualize all analyses presented in our manuscript by setting up a shiny application (in addition to our available dataset and R code), at https://fruitfliesmetabo.shinyapps.io/shiny/. This is now mentioned in the main text (l.588-589).

We thank the Reviewer for their comment that greatly improved the manuscript.

The authors generated a qualitative dataset using the peak list produced by XCMS which contains quantitative peak areas, it is unclear how the threshold was selected to determine if a peak is present or absent in a given sample. The qualitative dataset would influence the output of their data analysis.

The referee is right in pointing out that the threshold used to determine if a peak is present or absent in a given sample was not clearly specified. This has now been corrected in the “Host use” section of the Materials & Methods (l.513-516). Briefly, a given replicate of a compound was considered present if the corresponding peak area following XCMS quantification was > 1000. This threshold was selected to be close to the practical quantification threshold of the Thermo Exactive mass spectrometer used in this study. This threshold was selected in order to allow the quantification of low-abundance compounds, as many plant-derived diet compounds were expected to be present in trace amounts in flies. We additionally applied a stringent rule for presence of any given compound (presence in at least 3 biological replicates).

The authors reply on in-source fragmentation for peak annotation when authentic standards are not available. The accuracy of the annotation thus requires further validation.

The Supplementary Table 1 was unfortunately omitted in the first submission of the manuscript. This oversight has been now corrected and the Supplementary Table 1 details all information used for metabolite annotation. In particular, MS/MS data comparison with mass spectral databases as well as with published literature have been added to substantiate metabolite identifications. This MS/MS data was produced thanks to the comment of the Reviewer. We also provide four more annotations from standards to attain 30 / 71 identifications validated through chemical standards.

Author Response

Reviewer #1 (Public Review):

Overall, the science is sound and interesting, and the results are clearly presented. However, the paper falls in-between describing a novel method and studying biology. As a consequence, it is a bit difficult to grasp the general flow, central story and focus point. The study does uncover several interesting phenomena, but none are really studied in much detail and the novel biological insight is therefore a bit limited and lost in the abundance of observations. Several interesting novel interactions are uncovered, in particular for the SPS sensor and GAPDH paralogs, but these are not followed up on in much detail. The same can be said for the more general observations, eg the fact that different types of mutations (missense vs nonsense) in different types of genes (essential vs non-essential, housekeeping vs. stress-regulated...) cause different effects.

This is not to say that the paper has no merit - far from it even. But, in its current form, it is a bit chaotic. Maybe there is simply too much in the paper? To me, it would already help if the authors would explicitly state that the paper is a "methods" paper that describes a novel technique for studying the effects of mutations on protein abundance, and then goes on to demonstrate the possibilities of the technology by giving a few examples of the phenomena that can be studied. The discussion section ends in this way, but it may be helpful if this was moved to the end of the introduction.

We modified the manuscript as suggested.

Reviewer #2 (Public Review):

Schubert et al. describe a new pooled screening strategy that combines protein abundance measurements of 11 proteins determined via FACS with genome-wide mutagenesis of stop codons and missense mutations (achieved via a base editor) in yeast. The method allows to identify genetic perturbations that affect steady state protein levels (vs transcript abundance), and in this way define regulators of protein abundance. The authors find that perturbation of essential genes more often alters protein abundance than of nonessential genes and proteins with core cellular functions more often decrease in abundance in response to genetic perturbations than stress proteins. Genes whose knockouts affected the level of several of the 11 proteins were enriched in protein biosynthetic processes while genes whose knockouts affected specific proteins were enriched for functions in transcriptional regulation. The authors also leverage the dataset to confirm known and identify new regulatory relationships, such as a link between the SDS amino acid sensor and the stress response gene Yhb1 or between Ras/PKA signalling and GAPDH isoenzymes Tdh1, 2, and 3. In addition, the paper contains a section on benchmarking of the base editor in yeast, where it has not been used before.

Strengths and weaknesses of the paper

The authors establish the BE3 base editor as a screening tool in S. cerevisiae and very thoroughly benchmark its functionality for single edits and in different screening formats (fitness and FACS screening). This will be very beneficial for the yeast community.

The strategy established here allows measuring the effect of genetic perturbations on protein abundances in highly complex libraries. This complements capabilities for measuring effects of genetic perturbations on transcript levels, which is important as for some proteins mRNA and protein levels do not correlate well. The ability to measure proteins directly therefore promises to close an important gap in determining all their regulatory inputs. The strategy is furthermore broadly applicable beyond the current study. All experimental procedures are very well described and plasmids and scripts are openly shared, maximizing utility for the community.

There is a good balance between global analyses aimed at characterizing properties of the regulatory network and more detailed analyses of interesting new regulatory relationships. Some of the key conclusions are further supported by additional experimental evidence, which includes re-making specific mutations and confirming their effects on protein levels by mass spectrometry.

The conclusions of the paper are mostly well supported, but I am missing some analyses on reproducibility and potential confounders and some of the data analysis steps should be clarified.

The paper starts on the premise that measuring protein levels will identify regulators and regulatory principles that would not be found by measuring transcripts, but since the findings are not discussed in light of studies looking at mRNA levels it is unclear how the current study extends knowledge regarding the regulatory inputs of each protein.

See response to Comment #10.

Specific comments regarding data analysis, reproducibility, confounders

1) The authors use the number of unique barcodes per guide RNA rather than barcode counts to determine fold-changes. For reliable fold changes the number of unique barcodes per gRNA should then ideally be in the 100s for each guide, is that the case? It would also be important to show the distribution of the number of barcodes per gRNA and their abundances determined from read counts. I could imagine that if the distribution of barcodes per gRNA or the abundance of these barcodes is highly skewed (particularly if there are many barcodes with only few reads) that could lead to spurious differences in unique barcode number between the high and low fluorescence pool. I imagine some skew is present as is normal in pooled library experiments. The fold-changes in the control pools could show whether spurious differences are a problem, but it is not clear to me if and how these controls are used in the protein screen.

Because of the large number of screens performed in this study (11 proteins, with 8 replicates for each) we had to trade off sequencing depth and power against cell sorting time and sequencing cost, resulting in lower read and barcode numbers than what might be ideally aimed for. As described further in the response to Comment #5, we added a new figure to the manuscript that shows that the correlation of fold-changes between replicates is high (Figure 3–S1A). The second figure below shows that the correlation between the number of unique barcodes and the number of reads per gRNA is highly significant (p < 2.2e-16).

2) I like the idea of using an additional barcode (plasmid barcode) to distinguish between different cells with the same gRNA - this would directly allow to assess variability and serve as a sort of replicate within replicate. However, this information is not leveraged in the analysis. It would be nice to see an analysis of how well the different plasmid barcodes tagging the same gRNA agree (for fitness and protein abundance), to show how reproducible and reliable the findings are.

We agree with the reviewer that this would be nice to do in principle, but our sequencing depth for the sorted cell populations was not high enough to compare the same barcode across the low/unsorted/high samples. See also our response to Comment #5 for the replicate analyses.

3) From Fig 1 and previous research on base editors it is clear that mutation outcomes are often heterogeneous for the same gRNA and comprise a substantial fraction of wild-type alleles, alleles where only part of the Cs in the target window or where Cs outside the target window are edited, and non C-to-T edits. How does this reflect on the variability of phenotypic measurements, given that any barcode represents a genetically heterogeneous population of cells rather than a specific genotype? This would be important information for anyone planning to use the base editor in future.

We agree with the reviewer that the heterogeneity of editing outcomes is an important point to keep in mind when working with base editors. In genetic screens, like the ones described here, often the individual edit is less important, and the overall effects of the base editor are specific/localized enough to obtain insights into the effects of mutations in the area where the gRNA targets the genome. For example, in our test screens for Canavanine resistance and fitness effects, in which we used gRNAs predicted to introduce stop codons into the CAN1 gene and into essential genes, respectively, we see the expected loss-of-function effect for a majority of the gRNAs (canavanine screen: expected effect for 67% of all gRNAs introducing stop codons into CAN1; fitness screen: expected effect for 59% of all gRNAs introducing stop codons into essential genes) (Figure 2). In the canavanine screen, we also see that gRNAs predicted to introduce missense mutations at highly conserved residues are more likely to lead to a loss-of-function effect than gRNAs predicted to introduce missense mutations at less conserved residues, further highlighting the differentiated results that can be obtained with the base editor despite the heterogeneity in editing outcomes overall. We would certainly advise anyone to confirm by sequencing the base edits in individual mutants whenever a precise mutation is desired, as we did in this study when following up on selected findings with individual mutants.

4) How common are additional mutations in the genome of these cells and could they confound the measured effects? I can think of several sources of additional mutations, such as off-target editing, edits outside the target window, or when 2 gRNA plasmids are present in the same cell (both target windows obtain edits). Could some of these events explain the discrepancy in phenotype for two gRNAs that should make the same mutation (Fig S4)? Even though BE3 has been described in mammalian cells, an off-target analysis would be desirable as there can be substantial differences in off-target behavior between cell types and organisms.

Generally, we are not very concerned about random off-target activity of the base editor because we would not expect this to cause a consistent signal that would be picked up in our screen as a significant effect of a particular gRNA. Reproducible off-target editing with a specific gRNA at a site other than the intended target site would be problematic, though. We limited the chance of this happening by not using gRNAs that may target similar sequences to the intended target site in the genome. Specifically, we excluded gRNAs that have more than one target in the genome when the 12 nucleotides in the seed region (directly upstream of the PAM site) are considered (DiCarlo et al., Nucleic Acids Research, 2013).

We do observe some off-target editing right outside the target window, but generally at much lower frequency than the on-target editing in the target window (Figure 1B and Figure 1–S2). Since for most of our analyses we grouped perturbations per gene, such off-target edits should not affect our findings. In addition, we validated key findings with independent experiments. For our study, we used the Base Editor v3 (Komor et al., Nature, 2016); more recently, additional base editors have been developed that show improved accuracy and efficiency, and we would recommend these base editors when starting a new study (see, e.g., Anzalone et al., Nature Biotechnology, 2020).

We are not concerned about cases in which one cell gets two gRNAs, since the chance that the same two gRNAs end up in one cell repeatedly is low, and such events would therefore not result in a significant signal in our screens.

We don’t think that off-target mutations can explain the discrepancy between pairs of gRNAs that should introduce the same mutation (Figure 3–S1. The effect of the two gRNAs is actually well-correlated, but, often, one of the two gRNAs doesn’t pass our significance cut-off or simply doesn’t edit efficiently (i.e., most discrepancies arise from false negatives rather than false positives). We may therefore miss the effects of some mutations, but we are unlikely to draw erroneous conclusions from significant signals.

5) In the protein screen normalization uses the total unique barcode counts. Does this efficiently correct for differences from sequencing (rather than total read counts or other methods)? It would be nice to see some replicate plots for the analysis of the fitness as well as the protein screen to be able to judge that.

We made a new figure that shows a replicate comparison for the protein screen (see below; in the manuscript it is Figure 3–S1A) and commented on it in the manuscript. For this analysis, the eight replicates for each protein were split into two groups of four replicates each and analyzed the same way as the eight replicates. The correlation between the two groups of replicates is highly significant (p < 2.2e-16). The second figure shows that the total number of reads and the total number of unique barcodes are well correlated.

For the fitness screen, we used read counts rather than barcode counts for the analysis since read counts better reflect the dropout of cells due to reduced fitness. The figure below shows a replicate comparison for the fitness screen. For this analysis, the four replicates were split into two groups of two replicates each and analyzed the same way as the four replicates. The correlation between the two groups of replicates is highly significant (p < 2.2e-16).

6) In the main text the authors mention very high agreement between gRNAs introducing the same mutation but this is only based on 20 or so gRNA pairs; for many more pairs that introduce the same mutation only one reaches significance, and the correlation in their effects is lower (Fig S4). It would be better to reflect this in the text directly rather than exclusively in the supplementary information.

We clarified this in the manuscript main text: “For 78 of these gRNA pairs, at least one gRNA had a significant effect (FDR < 0.05) on at least one of the eleven proteins; their effects were highly correlated (Pearson’s R2 = 0.43, p < 2.2E-16) (Figure 3–S1B). For the 20 gRNA pairs for which both gRNAs had a significant effect, the correlation was even higher (Pearson’s R2 = 0.819, p = 8.8e-13) (Figure 3–S1C). These findings show that the significant gRNA effects that we identify have a low false positive rate, but they also suggest that many real gRNA effects are not detected in the screen due to limitations in statistical power.”

7) When the different gRNAs for a targeted gene are combined, instead of using an averaged measure of their effects the authors use the largest fold-change. This seems not ideal to me as it is sensitive to outliers (experimental error or background mutations present in that strain).

We agree that the method we used is more sensitive to outliers than averaging per gene. However, because many gRNAs have no effect either because they are not editing efficiently or because the edit doesn’t have a phenotypic consequence, an averaging method across all gRNAs targeting the same gene would be too conservative and not properly capture the effect of a perturbation of that gene.

8) Phenotyping is performed directly after editing, when the base editor is still present in the cells and could still interact with target sites. I could imagine this could lead to reduced levels of the proteins targeted for mutagenesis as it could act like a CRISPRi transcriptional roadblock. Could this enhance some of the effects or alter them in case of some missense mutations?

To reduce potential “CRISPRi-like” effects of the base editor on gene expression, we placed the base editor under a galactose-inducible promoter. For both the fitness and protein screens we grew the cultures in media without galactose for another 24 hours (fitness screen) or 8-9 hours (protein screens) before sampling. In the latter case, this recovery time corresponded to more than three cell divisions, after which we assume base editor levels to have strongly decreased, and therefore to no longer interfere with transcription. This is also supported by our ability to detect discordant effects of gRNAs targeting the same gene (e.g., the two mutations leading to loss-of-function and gain-of-function of RAS2), which would otherwise be overshadowed by a CRISPRi effect.

9) I feel that the main text does not reflect the actual editing efficiency very well (the main numbers I noticed were 95% C to T conversion and 89% of these occurring in a specific window). More informative for interpreting the results would be to know what fraction of the alleles show an edit (vs wild-type) and how many show the 'complete' edit (as the authors assume 100% of the genotypes generated by a gRNA to be conversion of all Cs to Ts in the target window). It would be important to state in the main text how variable this is for different gRNAs and what the typical purity of editing outcomes is.

We now show the editing efficiency and purity in a new figure (Figure 1B), and discuss it in the main text as follows: “We found that the target window and mutagenesis pattern are very similar to those described in human cells: 95% of edits are C-to-T transitions, and 89% of these occurred in a five-nucleotide window 13 to 17 base pairs upstream of the PAM sequence (Figure 1A; Figure 1–S2) (Komor et al., 2016). Editing efficiency was variable across the eight gRNAs and ranged from 4% to 64% if considering only cases where all Cs in the window are edited; percentages are higher if incomplete edits are considered, too (Figure 1B).”

Comments regarding findings

10) It would be nice to see a comparison of the results to the effects of ~1500 yeast gene knockouts on cellular transcriptomes (https://doi.org/10.1016/j.cell.2014.02.054). This would show where the current study extends established knowledge regarding the regulatory inputs of each protein and highlight the importance of directly measuring protein levels. This would be particularly interesting for proteins whose abundance cannot be predicted well from mRNA abundance.

We agree with the reviewer that it would be very interesting to compare the effect of perturbations on mRNA vs protein levels. We have compared our protein-level data to mRNA-level data from Kemmeren and colleagues (Kemmeren et al., Cell 2014), and we find very good agreement between the effects of gene perturbations on mRNA and protein levels when considering only genes with q < 0.05 and Log2FC > 0.5 in both studies (Pearson’s R = 0.79, p < 5.3e-15).

Gene perturbations with effects detected only on mRNA but not protein levels are enriched in genes with a role in “chromatin organization” (FDR = 0.01; as a background for the analysis, only the 1098 genes covered in both studies were considered). This suggests that perturbations of genes involved in chromatin organization tend to affect mRNA levels but are then buffered and do not lead to altered protein levels. There was no enrichment of functional annotations among gene perturbations with effects on protein levels but not mRNA levels.

We did not include these results in the manuscript because there are some limitations to the conclusions that can be drawn from these comparisons, including that our study has a relatively high number of false negatives, and that the genes perturbed in the Kemmeren et al. study were selected to play a role in gene regulation, meaning that differences in mRNA-vs-protein effects of perturbations are limited to this function, and other gene functions cannot be assessed.

11) The finding that genes that affect only one or two proteins are enriched for roles in transcriptional regulation could be a consequence of 'only' looking at 10 proteins rather than a globally valid conclusion. Particularly as the 10 proteins were selected for diverse functions that are subject to distinct regulatory cascades. ('only' because I appreciate this was a lot of work.)

We agree with this, and we think it is clear in the abstract and the main text of the manuscript that here we studied 11 proteins. We made this point also more explicit in the discussion, so that it is clear for readers that the findings are based on the 11 proteins and may not extrapolate to the entire yeast proteome.

Reviewer #3 (Public Review):

This manuscript presents two main contributions. First, the authors modified a CRISPR base editing system for use in an important model organism: budding yeast. Second, they demonstrate the utility of this system by using it to conduct an extremely high throughput study the effects of mutation on protein abundance. This study confirms known protein regulatory relationships and detects several important new ones. It also reveals trends in the type of mutations that influence protein abundances. Overall, the findings are of high significance and the method appears to be extremely useful. I found the conclusions to be justified by the data.

One potential weakness is that some of the methods are not described in main body of the paper, so the reader has to really dive into the methods section to understand particular aspects of the study, for example, how the fitness competition was conducted.

We expanded the first section for better readability.

Another potential weakness is the comparison of this study (of protein abundances) to previous studies (of transcript abundances) was a little cursory, and left some open questions. For example, is it remarkable that the mutations affecting protein abundance are predominantly in genes involved in translation rather than transcription, or is this an expected result of a study focusing on protein levels?

We thank the reviewer for pointing out that this paragraph requires more explanation. We expanded it as follows: “Of these 29 genes, 21 (72%) have roles in protein translation—more specifically, in ribosome biogenesis and tRNA metabolism (FDR < 8.0e-4, Figure 5C). In contrast, perturbations that affect the abundance of only one or two of the eleven proteins mostly occur in genes with roles in transcription (e.g., GO:0006351, FDR < 1.3e-5). Protein biosynthesis entails both transcription and translation, and these results suggest that perturbations of translational machinery alter protein abundance broadly, while perturbations of transcriptional machinery can tune the abundance of individual proteins. Thus, genes with post-transcriptional functions are more likely to appear as hubs in protein regulatory networks, whereas genes with transcriptional functions are likely to show fewer connections.”

Overall, the strengths of this study far outweigh these weaknesses. This manuscript represents a very large amount of work and demonstrates important new insights into protein regulatory networks.

Author Response

Reviewer #2 (Public Review):

The authors seek to determine how various species combine their effects on the growth of a species of interest when part of the same community.

To this end, the authors carry out an impressive experiment containing what I believe must be one of the largest pairwise + third-order co-culture experiments done to date, using a high-throughput co-culture system they had co-developed in previous work. The unprecedented nature of this data is a major strength of the paper. The authors also discover that species combine their effect through "dominance", i.e. the strongest effect masks the others. This is important as it calls into question the common assumption of additivity that is implicit in the choice of using Lotka-Volterra models.

A stronger claim (i.e. in the abstract) is that joint effect of multiple species on the growth of another can be derived from the effect of individual species. Unless I am misunderstanding something, this statement may have to be qualified a little, as the authors show that a model based on pairwise dominance (i.e. the strongest pairwise) does a somewhat better job (lower RMSD, though granted, not by much, 0.57 vs 0.63) than a model based on single species dominance. This is, the effect of the strongest pair predicts better the effect of a trio than the effect of the larger species.

This issue makes one wonder whether, had the authors included higher-order combinations of species (i.e. five-member consortia or higher), the strongest-effect trio would have predicted better than the strongest-effect pair, which in turn is better predictor than the strongest-effect species. This is important, as it would help one determine to what extent the strongest-effect model would work in more diverse communities, such as those one typically finds in nature. Indeed, the authors find that the predictive ability of the strongest effect species is much stronger for pairs than it is for trios (RMSD of 0.28 vs 0.63). Does the predictive ability of the single species model decline faster and faster as diversity grows beyond 4-member consortia?

Thank you for raising this important point. It is true that in our study we see that single species predict pairs better than trios, and that pairs predict trios better than single species. As we did not perform experiments on more diverse communities (n>4), we are not sure if or how these rules will scale up. We explicitly address these caveats in our revised discussion.

Reviewer #3 (Public Review):

A problem in synthetic ecology is that one can't brute-force complex community design because combinatorics make it basically impossible to screen all possible communities from a bank of possible species. Therefore, we need a way to predict phenomena in complex communities from phenomena in simple communities. This paper aims to improve this predictive ability by comparing a few different simple models applied to a large dataset obtained with the use of the author's "kchip" microfluidics device. The main question they ask is whether the effect of two species on a focal species is predicted from the mean, the sum, or the max of the effect of each single "affecting" species on the focal species. They find that the max effect is often the best predictor, in the sense of minimizing the difference between predicted effect and measured effect. They also measure single-species trait data for their library of strains, including resource niche and antibiotic resistance, and then find that Pearson correlations between distance calculations generated from these metrics and the effect of added species are weak and unpredictive. This work is largely well-done, timely and likely to be of high interest to the field, as predicting ecosystem traits from species traits is a major research aim.

My main criticism is that the main take-home from the paper (fig 3B)-that the strongest effect is the best predictor-is oversold. While it is true that, averaged over their six focal species, the "strongest effect" was the best overall predictor, when one looks at the species-specific data (S9), we see that it is not the best predictor for 1/3 of their focal species, and this fraction grows to 1/2 if one considers a difference in nRMSE of 0.01 to be negligible.

As suggested, we have softened our language regarding the take-home message. This matter is addressed in detail above in response to 'Essential Revisions'. Briefly, we see that the strongest model works best when both single species have qualitatively similar effects, but is slightly less accurate when effects are mixed. We also see overall less accurate predictions for positive effects. In light of these findings, we propose that focal species for which the strongest model is not the most accurate is due to the interaction types, and not specific to the focal species.

We made substantial changes to the manuscript, including the first paragraph of the discussion which more accurately describes these findings and emphasizes the relevant caveats:

"By measuring thousands of simplified microbial communities, we quantified the effects of single species, pairs, and trios on multiple focal species. The most accurate model, overall and specifically when both single species effects were negative, was the strongest effect model. This is in stark contrast to models often used in antibiotic compound combinations, despite most effects being negative, where additivity is often the default model (Bollenbach 2015). The additive model performed well for mixed effects (i.e. one negative and one positive), but only slightly better than the strongest model, and poorly when both species had effects of the same sign. When both single species’ effects were positive, the strongest model was also the best, though the difference was less pronounced and all models performed worse for these interactions. This may be due to the small effect size seen with positive effects, as when we limited negative and mixed effects to a similar range of effects strength, their accuracy dropped to similar values (Figure 3–Figure supplement 5). We posit that the difference in accuracy across species is affected mainly by the effect type dominating different focal species' interactions, rather than by inherent species traits (Figure 3–Figure supplement 6)." (Lines 288-304)

The same criticism applies to the result from figure 2-that pairs of affecting species have more negative effects than single species. Considered across all focal species this is true (though minor in effect size, Fig 2A). But there is only a significant effect within two individual species. Again, this points to the effects being focal-species-specific, and perhaps not as generalizable as is currently being claimed.

Upon more rigorous analysis, and with regard to changes in the dataset after filtering, we see that the more accurate statement is that effects become stronger, not necessarily more negative (in line with the accuracy of the strongest model). The overall trend is towards more negative interactions, due to the majority of interactions being negative, but as stated this is not true for each individual focal. As such the following sentence in the manuscript has been changed:

"The median effect on each focal was more negative by 0.28 on average, though the difference was not significant in all cases; additionally, focals with mostly positive single species interactions showed a small increase in median effect (Fig. 2D)" (Lines 151-154)

As well as the title of this section: "Joint effects of species pairs tend to be stronger than those of individual affecting species" (Lines 127-128)

Another thing that points to a focal-species-specific response is Fig 2D, which shows the distributions of responses of each focal species to pairs. Two of these distributions are unimodal, one appears bimodal, and three appear tri-modal. This suggests to me that the focal species respond in categorically different ways to species addition.

We believe this distribution of pair effects is related to the distribution of single species effects, and not to the way in which different focal species respond to the addition of second species. Though this may be difficult to see from the swarm plots shown in the paper, below is a split violin plot that emphasizes this point.

Fig R1: Distribution of single species and pair effects. Distribution of the effect of single and pairs of affecting species for each focal species individually. Dashed lines represent the median, while dotted lines the interquartile range.

These differences occur even though the focal bacteria are all from the same family. This suggests to me that the generalizability may be even less when a more phylogenetically dispersed set of focal species are used.

We have added the following sentence to the discussion explicitly emphasizing the phylogenetic limitations of our study:

"Lastly, it is important to note that our focal species are all from the same order (Enterobacterales), which may also limit the purview of our findings." (Lines 364-366)

Considering these points together, I argue that the conclusion should be shifted from "strongest effect is the best" to "in 3 of our focal species, strongest effect was the best, but this was not universal, and with only 6 focal species, we can't know if it will always be the best across a set of focal species".

As mentioned above, we have softened our language regarding the take-home message in response to these evaluations.

My second main criticism is that it is hard to understand exactly how the trait data were used to predict effects. It seems like it was just pearson correlation coefficients between interspecies niche distances (or antibiotic distances) and the effect. I'm not very surprised these correlations were unpredictive, because the underlying measurements don't seem to be relevant to the environment tested. What if, rather than using niche data across 20 nutrients, only the growth data on glucose (the carbon source in the experiments) was used? I understand that in a field experiment, for example, one might not know what resources are available, and so measuring niche across 20 resources may be the best thing to do. Here though it seems imperative to test using the most relevant data.

It is true that much of the profiling data is not directly related to the experimental conditions (different carbon sources and antibiotics), but in addition to these we do use measurements from experiments carried out in the same environment as the interactions assays (i.e. growth rate and carrying capacity when growing on glucose), which also showed poor correlation with the effects on focals. Additionally, we believe that these profiles contain relevant information regarding metabolic similarity between species (similar to metabolic models often constructed computationally). To improve clarity, we added the following sentence to the figure legend of Figure 3–Figure supplement 1:

"The growth rate, and maximum OD shown in panel A were measured only in M9 glucose, similar to conditions used in the interaction assays." (Lines 591-592)

Additionally and relatedly, it would be valuable to show the scatterplots leading to the conclusion that trait data were uninformative. Pearson's r only works on an assumption of linearity. But there could be strong relationships between the trait data and effect that are monotonic but not linear, or even that are non-monotonic yet still strong (e.g. U-shaped). For the first case, I recommend switching to Spearman's rho over Pearson's r, because it only assumes monotonicity, not linearity. If there are observable relationships that are not monotonic, a different test should be used.

Per your suggestion, we have changed the measurement of correlation in this analysis from Pearson's r, to Spearman's rho. As we observed similar, and still mostly weak correlations, we did not investigate these relationships further. See Figure 3–Figure supplement 1.

Additionally, we generated heat maps including scatterplots mapping the data leading to these correlations. We found no notable dependency in these plots, and visually they were quite crowded and difficult to interpret. As this is not the central point of our study, we ultimately decided against adding this information to the plots.

In general, I think the analyses using the trait data were too simplistic to conclude that the trait data are not predictive.

We agree that more sophisticated analyses may help connect between species traits and their effects on focal species. In fact, other members of our research group have recently used machine learning to accomplish similar predictions (https://doi.org/10.1101/2022.08.02.502471). As such we have changed the wording in to reflect that this correlation is difficult to find using simple analyses:

"These results indicate that it may be challenging to connect the effects of single and pairs of species on a focal strain to a specific trait of the involved strains, using simple analysis." (Lines 157-159)

Author Response

Reviewer #1 (Public Review):

Slusarczyk et al present a very well written manuscript focused on understanding the mechanisms underlying aging of erythrophagocytic macrophages in the spleen (RPM) and its relationship to iron loading with age. The manuscript is diffuse with a broad swath of data elements. Importantly, the manuscript demonstrates that RPM erythrophagocytic capacity is diminished with age, restored in iron restricted diet fed aged mice. In addition, the mechanism for declining RPM erythrophagocytic capacity appears to be ferroptosis-mediated, insensitive to heme as it is to iron, and occur independently of ROS generation. These are compelling findings. However, some of the data relies on conjecture for conclusion and a clear causal association is not clear. The main conclusion of the manuscript points to the accumulation of unavailable insoluble forms of iron as both causing and resulting from decreased RPM erythrophagocytic capacity.

We are proposing that intracellular iron accumulation progresses first and leads to global proteotoxic damage and increased lipid peroxidation. This eventually triggers the death of a fraction of aging RPMs, thus promoting the formation of extracellular iron-rich protein aggregates. More explanation can be found below. Besides, iron loading suppresses the erythrophagocytic activity of RPMs, hence further contributing to their functional impairment during aging.

In addition, the finding that IR diet leads to increased TF saturation in aged mice is surprising.

We believe that this observation implies better mobilization of splenic iron stores, and corroborates our conclusion that mice that age on an iron-reduced diet benefit from higher iron bioavailability, although these differences are relatively mild. More explanation can be found in our replies to Reviewer #2.

Furthermore, whether the finding in RPMs is intrinsic or related to RBC-related changes with aging is not addressed.

We now addressed this issue and we characterized in more detail both iron and ROS levels in RBCs.

Finally, these findings in a single strain and only female mice is intriguing but warrants tempered conclusions.

We tempered the conclusions and provided a basic characterization of the RPM aging phenotype in Balb/c female mice.

Major points:

1) The main concern is that there is no clear explanation of why iron increases during aging although the authors appear to be saying that iron accumulation is both the cause of and a consequence of decreased RPM erythrophagocytic capacity. This requires more clarification of the main hypothesis on Page 4, line 17-18.

We thank the reviewer for this comment. It was previously reported that iron accumulates substantially in the spleen during aging, especially in female mice (Altamura et al., 2014). Since RPMs are those cells that process most of the iron in the spleen, we aimed to explore what is the relationship between iron accumulation and RPM functions during aging. This investigation led us to uncover that indeed iron accumulation is both the cause and the consequence of RPM dysfunction. Specifically, we propose that intracellular iron loading of RPMs precedes extracellular deposition of iron in a form of protein-rich aggregates, driven by RPMs damage. To support this, we now show that the proteome of RPMs overlaps with those proteins that are present in the age-triggered aggregates (Fig. 3F). Furthermore, corroborating our model, we now demonstrate that transient iron loading of RPMs via iron-dextran injection (new Fig. 3G) leads to the formation of protein-rich aggregates, closely resembling those present in aged spleens (new Fig. 3H). This implies that high iron content in RPMs is indeed a major driving factor that leads to aggregation of their proteome and cell damage. Importantly, we now supported this model with studies using iRPMs. We demonstrated that iron loading and blockage of ferroportin by synthetic mini-hepcidin (PR73)(Stefanova et al., 2018) cause protein aggregation in iRPMs and lead to their decreased viability only in cells that were exposed to heat shock, a well-established trigger of proteotoxicity (new Fig. 5K and L). We propose that these two factors, namely age-triggered decrease in protein homeostasis and exposure to excessive iron levels, act in concert and render RPMs particularly sensitive to damage during aging (see also Discussion, p. 16).

In parallel, our data imply that the increased iron content in aged RPMs drives their decreased erythrophagocytic activity, as we now better documented by more extensive in vitro experiments in iRPMs (new Fig 6E-H). We cannot exclude that some of the senescent splenic RBCs that are retained in the red pulp and evade erythrophagocytosis due to RPM defects in aging, may also contribute to the formation of the aggregates. This is supported by the fact that mice that lack RPMs as well exhibit iron loading in the spleen (Kohyama et al., 2009; Okreglicka et al., 2021), and that the proteome of aggregates overlaps to some extent with the proteome of erythrocytes (new Fig. 3F).

We believe that during aging intracellular iron accumulation is chiefly driven by ferroportin downregulation, as also suggested by Reviewer#3. We now show that ferroportin drops significantly already in mice aged 4 and 5 months (new Fig. 4H), preceding most of the other impairments. This drop coincides with the increase in hepcidin expression, but if this is the sole reason for ferroportin suppression during early aging would require further investigation outside the scope of the present manuscript.

In sum, to address this comment, we now modified the fragment of the introduction that refers to our hypothesis and major findings to be more clear (p. 4), we improved our manuscript by providing new data mentioned above and we added more explanation in the corresponding sections of the Results and Discussion.

2) It is unclear if RPMs are in limited supply. Based on the introduction (page 4, line 13-15), they have limited self-renewal capacity and blood monocytes only partially replenished. Fig 4D suggests that there is a decrease in RPMs from aged mice. The %RPM from CD45+ compartment suggests that there may just be relatively more neutrophils or fewer monocytes recruited. There is not enough clarity on the meaning of this data point.

Thank you for this comment. We fully agree that %RPMs of CD45+ splenocytes, although well-accepted in literature (Kohyama et al., 2009; Okreglicka et al., 2021), is only a relative number. Hence, we now included additional data and explanations regarding the loss of RPMs during aging.

It was reported that the proportion of RPMs derived from bone marrow monocytes increases mildly but progressively during aging (Liu et al., 2019). This implies that due to the loss of the total RPM population, as illustrated by our data, the cells of embryonic origin are likely even more affected. We could confirm this assumption by re-analysis of the data from Liu et al. that we now included in the manuscript as Fig. 5E. These data clearly show that the representation of embryonically-derived RPMs drops more drastically than the percent of total RPMs, whereas the replenishment rate from monocytes is not affected significantly during aging. Consistent with this, we have not observed any robust change in the population of monocytes (F4/80-low, CD11b-high) or pre-RPMs (F4/80-high, CD11b-high) in the spleen at the age of 10 months (Figure 5-figure supplement 2A and B). We also have detected a mild decrease, not an increase, in the number of granulocytes (new Figure 5-figure supplement 2C). Furthermore, we measured in situ apoptosis marker and found a clear sign of apoptosis in the aged spleen (especially in the red pulp area), a phenotype that is less pronounced in mice on an IR diet (new Fig. 5O). This is consistent with the observation that apoptosis markers can be elevated in tissues upon ferroptosis induction (Friedmann Angeli et al., 2014) and that the proteotoxic stress in aged RPMs, which we now emphasized better in our manuscript, may also lead to apoptosis (Brancolini & Iuliano, 2020). Taken together, we strongly believe that the functional defect of embryonically-derived RPMs chiefly contributes to their shortage during aging.

3) Anemia of aging is a complex and poorly understood mechanistically. In general, it is considered similar to anemia of chronic inflammation with increased Epo, mild drop in Hb, and erythroid expansion, similar to ineffective erythropoiesis / low Epo responsiveness. It is not surprising that IR diet did not impact this mild anemia. However, was the MCV or MCH altered in aged and IR aged mice?

We now included the data for hematocrit, RBC counts, MCV, and MCH in Figure 1-figure supplement 5. Hematocrit shows a similar tendency as hemoglobin levels, but the values for RBC counts, MCV, and MCH seem not to be altered. We also show now that the erythropoietic activity in the bone marrow is not affected in aged versus young mice. Taken together, the anemic phenotype in female C57BL/6J mice at this age is very mild, which we emphasized in the main text, and is likely affected by other factors than serum iron levels (p. 6).

4) Page 6, line 23 onward: the conclusion is that KC compensate for the decreased function of RPM in the spleen, based on the expansion of KC fraction in the liver. Is there evidence that KCs are engaged in more erythrophagocytosis in aged mice? Furthermore, iron accumulation in the liver with age does not demonstrate specifically enhanced erythrophagocytosis of KC. Please clarify why liver iron accumulation would not be simply a consequence of increased parenchymal iron similar to increased splenic iron with age, independent of erythrophagocytic activity in resident macrophages in either organ.

Thanks for these questions. For the quantification of the erythrophagocytosis rate in KC, we show, as for the RPMs (Fig. 1K), the % of PKH67-positive macrophages, following transfusion of PKH67-stained stressed RBCs (Fig. 1M). The data implies a mild (not statistically significant) drop (of approx. 30%) in EP activity. We believe that it is overridden by a more pronounced (on average, 2-fold) increase in the representation of KCs (Fig. 1N). The mechanisms of iron accumulation between the spleen and the liver are very different. In the liver, we observed iron deposition in the parenchymal cells (not non-parenchymal, new Fig. 1P) that we currently characterizing in more detail in a parallel manuscript. Our data demonstrate a drop in transferrin saturation in aged mice. Hence, it is highly unlikely that aging would be hallmarked by the presence of circulating non-transferrin-bound iron that would be sequestered by hepatocytes, as shown previously (Jenkitkasemwong et al., 2015). Thus, the iron released locally by KCs is the most likely contributor to progressive hepatocytic iron loading during aging. The mechanism of iron delivery to hepatocytes from erythrophagocytosing KCs was demonstrated by Theurl et al.(Theurl et al., 2016), and we propose that it may be operational, although in a much more prolonged time scale, during aging. We now discussed this part better in our Results sections (p. 7).

5) Unclear whether the effect on RPMs is intrinsic or extrinsic. Would be helpful to evaluate aged iRPMs using young RBC vs. young iRPMs using old RBCs.

We are skeptical if the generation of iRPMs cells from aged mice would be helpful – these cells are a specific type of primary macrophage culture, derived from bone marrow monocytes with MCSF1, and exposed additionally to heme and IL-33 for 4 days. We do not expect that bone marrow monocytes are heavily affected by aging, and would thus recapitulate some aspects of aged RPMs from the spleen, especially after 8-day in vitro culture. However, to address the concerns of the reviewer, we now provide additional data regarding RBC fitness. Consistent with the time life-span experiment (Fig, 2A), we show that oxidative stress in RBCs is only increased in splenic, but not circulating RBCs (new Fig. 2C, replacing the old Fig. 2B and C). In addition, we show no signs of age-triggered iron loading in RBCs, either in the spleen (new Fig. 2F) or in the circulation (new Fig. 2B). Hence, we do not envision a possibility that RPMs become iron-loaded during aging as a result of erythrophagocytosis of iron-loaded RBCs. In support of this, we also have observed that during aging first RPMs’ FPN levels drop, afterward erythrophagocytosis rate decreases, and lastly, RBCs start to exhibit significantly increased oxidative stress (presented now in new Fig. 4H, J and K).

6) Discussion of aggregates in the spleen of aged mice (Fig 2G-2K and Fig 3) is very descriptive and non-specific. For example, if the iron-rich aggregates are hemosiderin, a hemosiderin-specific stain would be helpful. This data specifically is correlatory and difficult to extract value from.

Thanks for these comments. To the best of our knowledge Prussian blue Perls’ staining (Fig. 2J) is considered a hemosiderin staining. Our investigations aimed to better understand the nature and the origin of splenic iron deposits that to some extent are referred to as hemosiderin. Most importantly, as mentioned in our reply R1 Ad. 1. to assign causality to our data, we now demonstrated that iron accumulation in RPMs in response to iron-dextran (Fig. 3G) increases lipid peroxidation (Fig. 5F), tends to provoke RPMs depletion (Fig. 5G) and triggers the formation of protein-rich aggregates (new Fig. 3H). Of note, we assume that the loss of embryonically-derived RPMs in this model may be masked by simultaneous replenishment of the niche from monocytes, a phenomenon that may be addressed by future studies using Ms4a3-driven reporter mice (as shown for aged mice in our new Fig. 5E).

7) The aging phenotype in RPMs appears to be initiated sometime after 2 months of age. However, there is some reversal of the phenotype with increasing age, e.g. Fig 4B with decreased lipid peroxidation in 9 month old relative to 6 month old RPMs. What does this mean? Why is there a partial spontaneous normalization?

Thanks for this comment and questions. Indeed, the degree of lipid peroxidation exhibits some kinetics, suggestive of partial normalization. Of note, such a tendency is not evident for other aging phenotypes of RPMs, hence, we did not emphasize this in the original manuscript. However, in a revised version of the manuscript, we now present the re-analysis of the published data which implies that the number of embryonically-derived RPMs drops substantially between mice at 20 weeks and 36 weeks (new Fig. 5E). We think that the higher proportion of monocyte-derived RPMs in total RPM population later in aging (9 months) might be responsible for the partial alleviation of lipid peroxidation. We now discussed this possibility in the Results sections (p. 12).

8) Does the aging phenotype in RPMs respond to ferristatin? It appears that NAC, which is a glutathione generator and can reverse ferroptosis, does not reverse the decreased RPM erythrophagocytic capacity observed with age yet the authors still propose that ferroptosis is involved. A response to ferristatin is a standard and acceptable approach to evaluating ferroptosis.

We fully agree with the Reviewer that using ferristatin or Liproxstatin-1 would be very helpful to fully characterize a mechanism of RPMs depletion in mice. However, previous in vivo studies involving Liproxstatin-1 administration required daily injections of this ferroptosis inhibitor (Friedmann Angeli et al., 2014). This would be hardly feasible during aging. Regarding the experiments involving iron-dextran injection, using Liproxstatin-1 would require additional permission from the ethical committee which takes time to be processed and received. However, to address this question we now provide data from iRPMs cell cultures (new Fig.5 K-L). In essence, our results imply that both proteotoxic stress and iron overload act in concert to trigger cytotoxicity in RPM in vitro model. Interestingly, this phenomenon does not depend solely on the increased lipid peroxidation, but when we neutralize the latter with Liproxstatin-1, the cytotoxic effect is diminished (please, see also Results on p. 13 and Discussion p. 15/16).

9) The possible central role for HO-1 in the pathophysiology of decreased RPM erythrophagocytic capacity with age is interesting. However, it is not clear how the authors arrived at this hypothesis and would be useful to evaluate in the least whether RBCs in young vs. aged mice have more hemoglobin as these changes may be primary drivers of how much HO-1 is needed during erythrophagocytosis.

Thanks for this comment. We got interested in HO-1 levels based on the RNA sequencing data, which detected lower Hmox-1 expression in aged RPMs (Figure 3-figure supplement 1). We now show that the content of hemoglobin is not significantly altered in aged RBCs (MCH parameter, Figure 1-figure supplement 5E), hence we do not think that this is the major driver for Hmox-1 downregulation. Likewise, the levels of the Bach1 message, a gene encoding Hmox-1 transcriptional repressor, are not significantly altered according to RNAseq data. Hence, the reason for the transcriptional downregulation of Hmox-1 is not clear. Of note, HO-1 protein levels in the total spleen are higher in aged versus young mice, and we also detected a clear appearance of its nuclear truncated and enzymatically-inactive form (see a figure below, we opt not to include this in the manuscript for better clarity). The appearance of truncated HO-1 seems to be partially rescued by the IR diet. It is well established that the nuclear form of HO-1 emerges via proteolytic cleavage and migrates to the nucleus under conditions of oxidative stress (Mascaro et al., 2021). This additionally confirms that the aging spleen is hallmarked by an increased burden of ROS. Moreover, we also detected HO-1 as one of the components of the protein iron-rich aggregates. Thus, we propose that the low levels of the cytoplasmic enzymatically active form of HO-1 in RPMs (that we preferentially detect with our intracellular staining and flow cytometry) may be underlain by its nuclear translocation and sequestration in protein aggregates that evade antibody binding [this is also supported by our observation that the protein aggregates, despite the high content of ferritin (as indicated by MS analysis) are negative for L-ferritin staining. Of note, we also cannot exclude that other cell types in the aging spleen (eg. lymphocytes) express higher levels of HO-1 in response to splenic oxidative stress.

Fig. Total splenic levels of HO-1 in young, aged IR and aged mice.

Reviewer #2 (Public Review):

Slusarczyk et al. investigate the functional impairment of red pulp macrophages (RPMs) during aging. When red blood cells (RBCs) become senescent, they are recycled by RPMs via erythrophagocytosis (EP). This leads to an increase in intracellular heme and iron both of which are cytotoxic. The authors hypothesize that the continuous processing of iron by RPMs could alter their functions in an age-dependent manner. The authors used a wide variety of models: in vivo model using female mice with standard (200ppm) and restricted (25ppm) iron diet, ex vivo model using EP with splenocytes, and in vitro model with EP using iRPMs. The authors found iron accumulation in organs but markers for serum iron deficiency. They show that during aging, RPMs have a higher labile iron pool (LIP), decreased lysosomal activity with a concomitant reduction in EP. Furthermore, aging RPMs undergo ferroptosis resulting in a non-bioavailable iron deposition as intra and extracellular aggregates. Aged mice fed with an iron restricted diet restore most of the iron-recycling capacity of RPMs even though the mild-anemia remains unchanged.

Overall, I find the manuscript to be of significant potential interest. But there are important discrepancies that need to be first resolved. The proposed model is that during aging both EP and HO-1 expression decreases in RPMs but iron and ferroportin levels are elevated. In their model, the authors show intracellular iron-rich proteinaceous aggregates. But if HO-1 levels decrease, intracellular heme levels should increase. If Fpn levels increase, intracellular iron levels should decrease. How does LIP stay high in RPMs under these conditions? I find these to be major conflicting questions in the model.

We thank the Reviewer for her/his valuable feedback. As we mentioned in our replies we can only assume that a small misunderstanding in the interpretation of the presented data underlies this comment. We show that ferroportin levels in RPMs (Fig. 1F) are modulated in a manner that fully reflects the iron status of these cells (both labile and total iron levels, Figs. 1H and I). FPN levels drop in aged RPMs and are rescued when mice are maintained on a reduced iron diet. As pointed out by Reviewer#3, and explained in our replies we believe that ferroportin levels are critical for the observed phenotypes in aging. We now described our data in a more clear way to avoid any potential misinterpretation (p.6).

Reviewer #3 (Public Review):

This is a comprehensive study of the effects of aging of the function of red pulp macrophages (RPM) involved in iron recycling from erythrocytes. The authors document that insoluble iron accumulates in the spleen, that RPM become functionally impaired, and that these effects can be ameliorated by an iron-restricted diet. The study is well written, carefully done, extensively documented, and its conclusions are well supported. It is a useful and important addition for at least three distinct fields: aging, iron and macrophage biology.

The authors do not explain why an iron-restricted diet has such a strong beneficial effect on RPM aging. This is not at all obvious. I assume that the number of erythrocytes that are recycled in the spleen, and are by far the largest source of splenic iron, is not changed much by iron restriction. Is the iron retention time in macrophages changed by the diet, i.e. the recycled iron is retained for a short time when diet is iron-restricted (making hepcidin low and ferroportin high), and long time when iron is sufficient (making hepcidin high and ferroportin low)? Longer iron retention could increase damage and account for the effect. Possibly, macrophages may not empty completely of iron before having to ingest another senescent erythrocyte, and so gradually accumulate iron.

We are very grateful to this Reviewer for emphasizing the importance of the iron export capacity of RPMs as a possible driver of the observed phenotypes. Indeed, as mentioned above, we now show in the revised version of the manuscript that ferroportin drops early during aging (revised Fig. 4). Importantly, we now also observed that iron loading and limitation of iron export from iRPMs via ferroportin aggravate the impact of heat shock (a well-accepted trigger of proteotoxicity) on both protein aggregation and cell viability (new Fig. 5K and L). Physiologically, recent findings show that aging promotes a global decrease in protein solubility [BioRxiv manuscript (Sui X. et al., 2022)], and it is very likely that the constant exposure of RPMs to high iron fluxes renders these specialized cells particularly sensitive to proteome instability. This could be further aggravated by a build-up of iron due to the drop of ferroportin early during aging, ultimately leading to the appearance of the protein aggregates as early as at 5 months of age in C57BL/6J females. Based on the new data, we emphasized this model in the revised version of the manuscript (please, see Discussion on p. 16)

Author Response

Reviewer #1 (Public Review):

1) It would be helpful to include some sort of comparison in Fig. 4, e.g. the regressions shown in Fig 3, to indicate to what extent the ICCl data corresponds to the "control range" of frequency tuning.

Figure 4 was modified to show the frequency range typically found in the ICCls. This range is based on results from Wagner et al., 2007, which extensively surveyed ICCls responses. This modification shows that our ICCls recordings in the ruff-removed owls cover the normal frequency hearing range of the owl.

2) A central hypothesis of the study is that the frequency preference of the high-frequency neurons is lower in ruff-removed owls because of the lowered reliability caused by a lack of the ruff. Yet, while lower, the frequency range of many neurons in juvenile and ruff-removed owls seems sufficiently high to be still responsive at 7-8 kHz. I think it would be important to know to what extent neurons are still ITD sensitive at the "unreliable high frequencies" even if the CFs are lower since the "optimization" according to reliability depends not on the best frequency of each neuron per se, but whether neurons are less ITD sensitive at the higher, less reliable frequencies.

The concern regarding the frequency range that elicits responsivity was largely addressed above. Specifically, Figure L1 showing frequency tuning of frontally tuned ICx neurons in ruff-removed owls indicates that while there is some variability of tuning across neurons, there is little responsivity above 6 kHz. In contrast, equivalent analysis in juvenile owls (Figure L3), shows there is much more responsiveness and variability across neurons to high and low frequencies. This evidence supports our hypothesis that the juvenile owl brain is still highly plastic, which facilitates learning during development. Although the underlying data was already reported in Figure 7 of our previously submitted manuscript, we can include Figures L1 and L2, potentially as supplemental figures, if considered useful by editors and reviewers. Nevertheless, this argumentation was further expanded in the revised text (Line 229).

Figure L1. Frequency tuning of frontally-tuned ICx neurons in ruff-removed owls. Tuning curves are normalized by the max response. Thick black line indicates the average tuning curve. Dashed black line indicates basal response.

Figure L2. ITD sensitivity across frequencies in ruff-removed owl. Two example neurons shown in a and b. ITD tuning for tones (colored) and broadband (black) plotted by firing rate (non-normalized). Solid colored lines indicate responses to frequencies that are within the neuron’s preferred frequency range (i.e. above the half-height, see Methods), dashed lines indicate frequencies outside of the neuron’s frequency range.

Figure L3. Frequency tuning of frontally-tuned ICx neurons in juvenile owls. Tuning curves are normalized by the max response. Thick black line indicates the average tuning curve. Dashed black line indicates basal response.

3) It would be interesting to have an estimate of the time scale of experience dependency that induces tuning changes. Do the authors have any data on this question? I appreciate the authors' notion that the quantifications in Fig 7 might indicate that juvenile owls are already "beginning to be shaped by ITD reliability" (line 323 in Discussion). How many days after hearing onset would this correspond to? Does this mean that a few days will already induce changes?

While tracking changes induced by ruff-removal over development were outside of the scope of this study, many other studies have assessed experience-dependent plasticity in the barn owl. The recordings in this study were performed approximately 20 days after hearing onset, suggesting that the juveniles had ample time to begin learning. These points were expanded upon in the discussion (Lines 254, 280-283).

Reviewer #2 (Public Review):

1) Why is IPD variability plotted instead of ITD variability (or indeed spatial reliability)? The relationship between these measures is likely to vary across frequency, which makes it difficult to compare ITD variability across frequency when IPDs are plotted. Normalizing data across frequencies also makes it difficult to compare different locations and acoustical conditions. For example, in Fig.1a and Fig.1b, the data shown for 3 kHz at ~160 degrees seems quantitatively and visually quite different, but the difference (in Fig.1c) appears to be negligible.

Justification of why IPD variability is used as an estimate of ITD variability was added to introduction (Lines 55-60), results (Line 100) and methods (Lines 371-374) sections of the manuscript, explaining the fact that because ITD detection is based on phase locking by auditory nerve and ITD detector neurons tuned to narrow frequency bands, responses of ITD detector neurons forwarded to downstream midbrain regions are therefore determined by IPD variability. Additionally, ITD is calculated by dividing IPD by frequency, which makes comparisons of ITD reliability across frequency mathematically uninformative.

2) How well do the measures of ITD reliability used reflect real-world listening? For example, the model used to calculate ITD reliability appears to assume the same (flat) spectral profile for targets and distractors, which are presented simultaneously with the same temporal envelope, and a uniform spatial distribution of sounds across space. It is therefore unclear how robust the study's results are to violations of these assumptions.

While we agree that our analysis cannot completely capture real-world listening for the barn owl, a general analysis using similar flat spectral profiles for targets and concurrent sounds provides a broad assessment of reliability of ITD cues. While a full recapitulation of real-world listening is beyond the scope of this study (i.e. recording natural scenes from the ear canals of wild barn owls), we included additional analyses of ITD reliability in Figure 1-figure supplement 1, described above.

3) Does facial ruff removal produce an isolated effect on ITD variability or does it also produce changes in directional gain, and the relationship between spatial cues and sound location? Although the study considers this issue in some places (e.g. Fig.2, Fig.5), a clearer presentation of the acoustical effects of facial ruff removal and their implications (for all locations, not just those to the front), as well as an attempt to understand how these acoustical changes lead to the observed changes in ITD reliability, would greatly strengthen the study. In addition, Fig.1 shows average ITD reliability across owls, but it would be helpful to know how consistent these measures are across owls, given individual variability in Head-Related Transfer Functions (HRTFs). This potentially has implications for the electrophysiological experiments, if the HRTFs of those animals were not measured. One specific question that is potentially very relevant is whether the facial ruff attenuates sounds presented behind the animal and whether it does so in a frequency-dependent way. In addition, if facial ruff removal enables ILDs to be used for azimuth, then ITDs may also become less necessary at higher frequencies, even if their reliability remains unchanged.

Additional analysis was conducted to generate representation of changes in directional gain induced by ruff removal, added to new figure (Fig 5). This analysis shows that changes in gain following ruff-removal are largely frequency-independent: there is a de-attenuation of peripherally and rearwardly located sounds, but the highest gain remains for high frequencies in frontal space. There is an additional increase in gain for high frequencies from rearward space, these changes would not explain the changes in frequency tuning we report. As mentioned in new additions to the manuscript, the changes at the most rearward-located auditory spatial locations are unlikely to have an effect on the auditory midbrain. No studies in the barn owl have found neurons in the ICx or optic tectum tuned to >120° (Knudsen, 1982; Knudsen, 1984; Cazettes et al., 2014). In addition, variability of IPD reliability across owls was analyzed and reported in the amended Figure 1, which notes very little changes across owls. In this analysis, we did realize that the file of one of the HRTFs obtained from von Campenhausen et al. 2006 was mislabeled, which explains slight differences in revised Fig 1b. Nevertheless, added analysis of IPD reliability across owls indicates that the pattern in ITD reliability is stable across owls (Fig. 1d,e), which supports our decision to not record HRTFs from owls used in this study. Finally, we added to the discussion that clarifies that the use of ILD for azimuth would not provide the same resolution as ITD would (Lines 295-303). We also do not believe that the use of ILD for azimuth would make “ITDs… less necessary at higher frequencies”, given that the ICCls is still computing ITD at these high frequencies (Fig 4), and that ILDs also have higher resolution at higher frequencies, with and without the facial ruff (Olsen et al, 1989; Keller et al., 1998; von Campenhausen et al., 2006).

1) It is unclear why some analyses (Fig.5, Fig.7) are focused on frontal locations and frontally-tuned neurons. It is also unclear why neurons with a best ITDs of 0 are described as frontally tuned since locations behind the animal produce an ITD of 0 also. Related to this, in Fig.1, facial ruff removal appears to reduce IPD variability at low frequencies for locations to the rear (~160 degrees), where the ITD is likely to be close to 0. Neurons with a best ITD of 0 might therefore be expected to adjust their frequency tuning in opposite directions depending on whether they are tuned to frontal or rearward locations.

An extensive explanation was added to the methods detailing why we do not believe the neurons recorded in this study are tuned to the rear. Namely, studies mapping the barn owl’s ICx and optic tectum have not reported neurons tuned to locations >120°, with the number of neurons representing a given spatial location decreasing with eccentricity (Knudsen, 1982; Knudsen, 1984; Cazettes et al., 2014). While we agree that there does seem to be a change in ITD reliability at ~160° following ruff-removal, the result is largely similar to the change that occurs in frontal space (Fig 1b), which is consistent with the ruff-removed head functioning as a sphere. Thus, we wouldn’t expect rearwardly-tuned neurons, if they could be readily found, to adjust their frequency tuning to higher frequencies. Finally, we want to clarify that we focused our analyses on frontally-tuned neurons because frontal space is where we observed the largest change in ITD reliability. Text was added to the Discussion section to clarify this point (Lines 313-321).

2) The study suggests that information about high-frequency ITDs is not passed on to the ICX if the ICX does not contain neurons that have a high best frequency. However, neurons might be sensitive to ITDs at frequencies other than the best frequency, particularly if their frequency tuning is broader. It is also unclear whether the best frequency of a neuron always corresponds to the frequency that provides the most reliable ITD information, which the study implicitly assumes.

The concern about ITD sensitivity at non-preferred frequencies was addressed under the essential revision #3, as well as under Reviewer 1’s concerns.

Author Response

Reviewer #1 (Public Review):

This manuscript reports a systematic study of the cortical propagation patterns of human beta bursts (~13-35Hz) generated around simple finger movements (index and middle finger button presses).

The authors deployed a sophisticated and original methodology to measure the anatomical and dynamical characteristics of the cortical propagation of these transient events. MEG data from another study (visual discrimination task) was repurposed for the present investigation. The data sample is small (8 participants). However, beta bursts were extracted over a +/- 2s time window about each button press, from single trials, yielding the detection and analysis of hundreds of such events of interest. The main finding consists of the demonstration that the cortical activity at the source of movement related beta bursts follows two main propagation patterns: one along an anteroposterior directions (predominantly originating from pre central motor regions), and the other along a medio- lateral (i.e., dorso lateral) direction (predominantly originating from post central sensory regions). Some differences are reported, post-hoc, in terms of amplitude/cortical spread/propagation velocity between pre and post-movement beta bursts. Several control tests are conducted to ascertain the veracity of those findings, accounting for expected variations of signal-to-noise ration across participants and sessions, cortical mesh characteristics and signal leakage expected from MEG source imaging.

One major perceived weakness is the purely descriptive nature of the reported findings: no meaningful difference was found between bursts traveling along the two different principal modes of propagation, and importantly, no relation with behavior (response time) was found. The same stands for pre vs. post motor bursts, except for the expected finding that post-motor bursts are more frequent and tend to be of greater amplitude (yielding the observation of a so-called beta rebound, on average across trials).

Overall, and despite substantial methodological explorations and the description of two modes of propagation, the study falls short of advancing our understanding of the functional role of movement related beta bursts.

For these reasons, the expected impact of the study on the field may be limited. The data is also relatively limited (simple button presses), in terms of behavioral features that could be related to the neurophysiological observations. One missed opportunity to explain the functional role of the distinct propagation patterns reports would have been, for instance, to measure the cortical "destination" of their respective trajectories.

In response to this comment, we would like to highlight two important points.

First, our work constitutes the first non-invasive human confirmation of invasive work in animals (Balasubramanian et al., 2020; Roberts et al., 2019; Rule et al., 2018; (Balasubramanian et al., 2020; Best et al., 2016; Rubino et al., 2006; Takahashi et al., 2011, 2015) and patients (Takahashi et al., 2011). Thus, these results bridges between recordings limited to the size of multielectrode arrays (roughly 0.16 cm2; Balasubramanian et al., 2020; Best et al., 2016; Rubino et al., 2006; Takahashi et al., 2011, 2015) and human EEG recordings spanning across large areas of the cortex and several functionally distinct regions (Alexander et al., 2016; Stolk et al., 2019). The ability to access these neural signatures non- invasively is important for cross-species comparison. This further enables us, to provide an in-depth analysis of the spatiotemporal diversity of human MEG signals and a detailed characterisation of the two propagation directions, which significantly extends previous reports. We note that their functional role remains undetermined also in these animal studies, but being able to identify these signals now in humans can provide a steppingstone for identifying their role.

Second, and related, the reviewers are correct that we did not observe distinct propagation directions between pre- and post-movement bursts, nor a relationship with reaction time. However, such a null result would be relevant, in our view, towards understanding what the functional relevance of these signals, if any, might be. Recent work in macaques indicates that the spatiotemporal patterns of high-gamma activity carry kinematic information about the upcoming movement (Liang et al 2023). The functional role of beta may therefore be more complex and not relate to reaction times or kinematics in a straightforward manner. We believe this is a relevant observation, and in keeping with the continued efforts to identify how sensorimotor beta relates to behaviour. It is increasingly clear that spatiotemporal diversity in animal recordings and human E/MEG and intracranial recordings can constitute a substantial proportion of the measured dynamics. As such, our report is relevant in narrowing down what these signals may reflect.

Together, we think that our work provides new insights into the multidimensional and propagating features of burst activity. This is important for the entire electrophysiology community, as it transforms how we commonly analyse and interpret these important brain signals. We anticipate that our work will guide and inspire future work on the mechanistic underpinnings of these dominant neural signals. We are confident that our article has the scope to reach out to the diverse readership of eLife.

Reviewer #2 (Public Review):

The authors devised novel and interesting experiments using high precision human MEG to demonstrate the propagation of beta oscillation events along two axes in the brain. Using careful analysis, they show different properties of beta events pre- and post movement, including changes in amplitude. Due to beta's prominent role in motor system dynamics, these changes are therefore linked to behavior and offer insights into the mechanisms leading to movement. The linking of wave-like phenomena and transient dynamics in the brain offers new insight into two paradigms about neural dynamics, offering new ways to think about each phenomena on its own.

Although there is a substantial, and recent, body of literature supporting the conclusions that beta and other neural oscillations are transient, care must be taken when analyzing the data and the resulting conclusions about beta properties in both time and space. For example, modifying the threshold at which beta events are detected could alter their reported properties and expression in space and time. The authors should therefore performing parameter sweeps on e.g. the thresholds for detection of oscillation bursts to determine whether their conclusions on beta properties and propagation hold. If this additional analysis does not change their story, it would lend confidence in the results/conclusions.

We thank the reviewing team for this comment. As suggested, we evaluated the effect of different burst thresholds on the burst parameters.

The threshold in the main analysis was determined empirically from the data, as in previous work (Little et al., 2019). Specifically, trial-wise power was correlated with the burst probability across a range of different threshold values (from median to median plus seven standard deviations (std), in steps of 0.25, see Figure 6-figure supplement 1). The threshold value that retained the highest correlation between trial-wise power and burst probability was used to binarize the data.

We repeated our original analysis using four additional thresholds, i.e., original threshold - 0.5 std, -0.25 std, +0.25 std, +0.5 std. As one would expect, burst threshold is negatively related to the number of bursts (i.e., higher thresholds yield fewer bursts, Figure R4a [top]), and positively related to burst amplitude (i.e., higher thresholds yield higher burst amplitudes, Figure R4a [bottom]).

Similarly, the temporal duration of bursts and apparent spatial width are modulated by the burst threshold: lowering the threshold leads to longer temporal duration and larger apparent spatial width while increasing the threshold leads to shorter temporal duration and smaller apparent spatial width Figure R4b. Note that for the temporal and spectral burst characteristics, the difference to the original threshold can be numerically zero, i.e., changing the burst threshold did not lead to changes exceeding the temporal and spectral resolution of the applied time-frequency transformation (i.e., 200ms and 1Hz respectively).

Importantly, across these threshold values, the propagation direction and propagation speed remain comparable.

We now include this result as Figure 6-figure supplement 2and refer to this analysis in the manuscript (page 28 line 717).

“To explore the robustness of the results analyses were repeated using a range of thresholds (Figure 6-figure supplement 2).”

Determining the generators of beta events at different locations is a tricky issue. The authors mentioned a single generator that is responsible for propagating beta along the two axes described. However, it is not clear through what mechanism the beta events could travel along the neural substrate without additional local generators along the way. Previous work on beta events examined how a sequence of synaptic inputs to supra and infragranular layers would contribute to a typical beta event waveform. Although it is possible other mechanisms exist, how might this work as the beta events propagate through space? Some further explanation/investigation on these issues is therefore warranted.

Based on this and other comments (i.e., comments 7 and 8) we re-evaluated the use of the term ‘generator’ in this manuscript.

While the term generator can be used across scales, from micro- to macroscale, ifor the purpose of the present paper, we believe one should differentiate at least two concepts: a) generator of beta bursts, and b) generator of travelling waves.

We realised that in the previous version of the manuscript the term ‘generator’ was at times used without context. We removed the term where no longer necessary.

Further, the previous version of the manuscript discussed putative generators of travelling waves (page 19f.) but not generators of beta bursts. We now address this as follows:

“Studies using biophysical modelling have proposed that beta bursts are generated by a broad infragranular excitatory synaptic drive temporally aligned with a strong supragranular synaptic drive (Law et al., 2022; Neymotin et al., 2020; Sherman et al., 2016; Shin et al., 2017) whereby layer specific inhibition acts to stabilise beta bursts in the temporal domain (West et al., 2023). The supragranular drive is thought to originate in the thalamus (E. G. Jones, 1998, 2001; Mo & Sherman, 2019; Seedat et al., 2020), indicating thalamocortical mechanisms (page 22f).”

Once the mechanisms have been better understood, a question of how much the results generalize to other oscillation frequencies and other brain areas. On the first question of other oscillation frequencies, the authors could easily test whether nearby frequency bands (alpha and low gamma) have similar properties. This would help to determine whether the observations/conclusions are unique to beta, or more generally applicable to transient bursts/waves in the brain. On the second issue of applicability to other brain areas, the authors could relate their work to transient bursts and waves recorded using ECoG and/or iEEG. Some recent work on traveling waves at the brain-wide level would be relevant for such comparisons.

We appreciate the enthusiasm and the suggestions. To comment on the frequency specificity of the observed effects we conducted the same analysis focusing on the gamma frequency range (60-90 Hz). For computational reasons, we limited this analysis to one subject. Figure R1 shows the polar probability histogram for the beta frequency range (left) and the gamma frequency range (right). In contrast to the beta frequency range, no dominant directions were observed for the gamma range and von Mises functions did not converge. These preliminary results suggest some frequency specificity of the spatiotemporal pattern in sensorimotor beta activity. We believe this paves the way for future analysis mapping propagation direction across frequency and space.

Here we did not investigate the spatial specificity of the effects, as the beta frequency range is dominant in sensorimotor areas. Investigating beta bursts in other cortical areas would have likely resulted in very few bursts. We discuss our results across spatial scales in the section: Distinct anatomical propagation axes of sensorimotor beta activity. However, please note that most of the previous literature operates on a different spatial scale (roughly 4mm; Balasubramanian et al., 2020; Best et al., 2016; Rubino et al., 2006; Rule et al., 2018; Takahashi et al., 2011, 2015) and different species (e.g., non-human primates). Non-invasive recordings in humans capture temporospatial patterns of a very different scale, i.e., often across the whole cortex (Alexander et al., 2016; Roberts et al., 2019). Comparing spatiotemporal patterns, across different spatial scales is inherently difficult. Work

investigating different spatial scales simultaneously, such as Sreekumar et al. 2020, is required to fully unpack the relationship between mesoscopic and macroscopic spatiotemporal patterns.

Figure R1: Spatiotemporal organisation for the beta (β, 13-30Hz) and gamma (γ, 60-90) frequency range for one exemplar subject. Same as Figure 4a, but for one exemplar subject.

If the source code could be provided on github along with documentation and a standard "notebook" on use other researchers would benefit greatly.

All analyses are performed using freely available tools in MATLAB. The code carrying out the analysis in this paper can be found here: [link provided upon acceptance]. The 3D burst analyses can be very computationally intensive even on a modern computer system. The analyses in this paper were computed on a MacBook Pro with a 2.6 GHz 6-Core Intel Core i7 and 32 Gb of RAM. Details on the installation and setup of the dependencies can be found in the README.md file in the main study repository.

This information has been added to the paper in the methods section on page 35.

Author Response

Reviewer #1 (Public Review):

This manuscript provides a comprehensive investigation of the effects of the genetic ablation of three different transcription factors (Srf, Mrtfa, and Mrtfb) in the inner ear hair cells. Based on the published data, the authors hypothesized that these transcription factors may be involved in the regulation of the genes essential for building the actin-rich structures at the apex of hair cells, the mechanosensory stereocilia and their mechanical support - the cuticular plate. Indeed, the authors found that two of these transcription factors (Srf and Mrtfb) are essential for the proper formation and/or maintenance of these structures in the auditory hair cells. Surprisingly, Srf- and Mrtfb- deficient hair cells exhibited somewhat similar abnormalities in the stereocilia and in the cuticular plates even though these transcription factors have very different effects on the hair cell transcriptome. Another interesting finding of this study is that the hair cell abnormalities in Srfdeficient mice could be rescued by AAV-mediated delivery of Cnn2, one of the downstream targets of Srf. However, despite a rather comprehensive assessment of the novel mouse models, the authors do not have yet any experimentally testable mechanistic model of how exactly Srf and Mrtfb contribute to the formation of actin cytoskeleton in the hair cells. The lack of any specific working model linking Srf and/or Mrtfb with stereocilia formation decreases the potential impact of this study.

Major comments:

Figures 1 & 3: The conclusion on abnormalities in the actin meshwork of the cuticular plate was based largely on the comparison of the intensities of phalloidin staining in separate samples from different groups. In general, any comparison of the intensity of fluorescence between different samples is unreliable, no matter how carefully one could try matching sample preparation and imaging conditions. In this case, two other techniques would be more convincing: 1) quantification of the volume of the cuticular plates from fluorescent images; and 2) direct examination of the cuticular plates by transmission electron microscopy (TEM).

In fact, the manuscript provides no single TEM image of the F-actin abnormalities either in the cuticular plate or in the stereocilia, even though these abnormalities seem to be the major focus of the study. Overall, it is still unclear what exactly Srf or Mrtfb deficiencies do with F-actin in the hair cells.

Yes, we agree. As suggested by the reviewer, to directly examine the defects in F-actin organization within the cuticular plate of mutant mice, we conducted Transmission Electron Microscopy (TEM) analyses. The results, as presented in the revised Figures 1 and 4 (panels F, G, and E, F, respectively), provide crucial insights into the structural changes in the cuticular plate. Meanwhile, the comparison of the volume of the phalloidin labeled cuticular plate after 3-D reconstruction using Imaris software was conducted and shown in Author response image 1. The results of the cuticular plate (CP) volume were consistent with the relative F-actin intensity change of the cuticular plate in the revised Figures 1B and 4B. For the TEM analysis of the stereocilia, we regret that due to time constraints, we were unable to collect TEM images of stereocilia with sufficient quality for a meaningful comparison. However, we believe that the data we have presented sufficiently addresses the primary concerns, and we appreciate the reviewers’ understanding of these limitations.

Author response image 1.

Figures 2 & 4 represent another example of how deceiving could be a simple comparison of the intensity of fluorescence between the genotypes. It is not clear whether the reduced immunofluorescence of the investigated molecules (ESPN1, EPS8, GNAI3, or FSCN2) results from their mis-localization or represents a simple consequence of the fact that a thinner stereocilium would always have a smaller signal of the protein of interest, even though the ratio of this protein to the number of actin filaments remains unchanged. According to my examination of the representative images of these figures, loss of Srf produces mis-localization of the investigated proteins and irregular labeling in different stereocilia of the same bundle, while loss of Mrtfb does not. Obviously, a simple quantification of the intensity of fluorescence conceals these important differences.

Yes, we agree. In addition to the quantification of tip protein intensity, we have added a few more analyses in the revised Figure 3 and Figure 6, such as the percentage of row 1 tip stereocilia with tip protein staining and the percentage of IHCs with tip protein staining on row 2 tip. Using the results mentioned above, the differences in the expression level, the row-specific distribution and the irregular labeling of tip proteins between the control and the mutants can be analyzed more thoroughly.

Reviewer #2 (Public Review):

The analysis of bundle morphology using both confocal and SEM imaging is a strength of the paper and the authors have some nice images, especially with SEM. Still, the main weakness is that it is unclear how significant their findings are in terms of understanding bundle development; the mouse phenotypes are not distinct enough to make it clear that they serve different functions so the reader is left wondering what the main takeaway is.

Based on the reviewer’s comments, in this revised manuscript, we put more emphasis on describing the effects of SRF and MRTFB on key tip proteins’ localization pattern during stereocilia development, represented by ESPN1, EPS8 and GNAI3, as well as the effects of SRF and MRTFB on the F-actin organization of cuticular plate using TEM. We have made substantial efforts to interpret the mechanistic underpinnings of the roles of SRF and MRTFB in hair cells. This is reflected in the revised Figures 1, 3, 4, 6, and 10, where we provide more comprehensive insights into the mechanisms at play.

We interpret our data in a way that both SRF and MRTF regulate the development and maintenance of the hair cell’s actin cytoskeleton in a complementary manner. Deletion of either gene thus results in somewhat similar phenotypes in hair cell morphology, despite the surprising lack of overlap of SRF and MRTFB downstream targets in the hair cell.

In Figure 1 and 3, changes in bundle morphology clearly don't occur until after P5. Widening still occurs to some extent but lengthening does not and instead the stereocilia appear to shrink in length. EPS8 levels appear to be the most reduced of all the tip proteins (Srf mutants) so I wonder if these mutants are just similar to an EPS8 KO if the loss of EPS8 occurred postnatally (P0-P5).

To address this question, we performed EPS8 staining on the control and Srf cKO hair cells at P4 and P10. We found that the dramatic decrease of the row 1 tip signal for EPS8 started since P4 in Srf cKO IHCs. Although the major hair bundle phenotype of Eps8 KO, including the defects of row 1 stereocilia lengthening and additional rows of short stereocilia also appeared in Srf cKO IHCs, there are still some bundle morphology differences between Eps8 KO and Srf cKO. For example, firstly, both Eps8 KO OHCs and IHCs showed additional rows of short stereocilia, but we only observed additional rows of short stereocilia in Srf cKO IHCs. Secondly, in Valeria Zampini’s study, SEM and TEM images did not show an obvious reduction of row 2 stereocilia widening (P18-P35), while our analysis of SEM images confirmed that the width of row 2 IHC stereocilia was drastically reduced by 40% in Srf cKO (P15). Generally, we think although Srf cKO hair bundles are somewhat similar to Eps8 KO, the Srf cKO hair bundle phenotype might be governed by multiple candidate genes cooperatively.

Reference:

Valeria Zampini, et al. Eps8 regulates hair bundle length and functional maturation of mammalian auditory hair cells. PLoS Biol. 2011 Apr;9(4): e1001048.

A major shortcoming is that there are few details on how the image analyses were done. Were SEM images corrected for shrinkage? How was each of the immunocytochemistry quantitation (e.g., cuticular plates for phalloidin and tip staining for antibodies) done? There are multiple ways of doing this but there are few indications in the manuscript.

We apologize for not making the description of the procedure of images analyses clear enough. As described in Nicolas Grillet group’s study, live and mildly-fixed IHC stereocilia have similar dimensions, while SEM preparation results in a hair bundle at a 2:3 scale compared to the live preparation. In our study, the hair cells selected for SEM imaging and measurements were located in the basal turn (30-32kHz), while the hair cells selected for fluorescence-based imaging and measurements were located in the middle turn (20-24kHz) or the basal turn (32-36kHz). Although our SEM imaging and fluorescence-based imaging of basal turn’s hair bundles were not from the same area exactly, the control hair bundles with SEM imaging have reduced row 1 stereocilia length by 10%-20%, compared to the control hair bundles with fluorescence-based imaging (revised Figure 2 and Figure 5). Generally, our stereocilia dimensions data showed appropriate shrinkage caused by the SEM preparation.

Recognizing the need for clarity, we have provided a detailed description of our image quantification and analysis procedures in the “Materials and Methods” section, specifically under “Immunocytochemistry.” This will aid readers in understanding our methodologies and ensure transparency in our approach.

Reference:

Katharine K Miller, et al. Dimensions of a Living Cochlear Hair Bundle. Front Cell Dev Biol. 2021 Nov 25:9:742529.

The tip protein analysis in Figs 2 and 4 is nice but it would be nice for the authors to show the protein staining separately from the phalloidin so you could see how restricted to the tips it is (each in grayscale). This is especially true for the CNN2 labeling in Fig 7 as it does not look particularly tip specific in the x-y panels. It would be especially important to see the antibody staining in the reslices separate from phalloidin.

Thank you for the suggestions. We have shown tip proteins staining in grayscale separately from the phalloidin in the revised Figure 3 and Figure 6. To clearly show the tip-specific localization of CNN2, we conducted CNN2 staining at different ages during hair bundle development and showed CNN2 labeling in grayscale and in reslices in revised Figure 9-figure supplement 1B.

In Fig 6, why was the transcriptome analysis at P2 given that the phenotype in these mice occurs much later? While redoing the transcriptome analysis is probably not an option, an alternative would be to show more examples of EPS8/GNAI/CNN2 staining in the KO, but at younger ages closer to the time of PCR analysis, such as at P5. Pinpointing when the tip protein intensities start to decrease in the KOs would be useful rather than just showing one age (P10).

We agree with the reviewer. To address this question, we have performed ESPN1, EPS8 and GNAI3 staining on the control and the mutant’s hair cells at P4, P10 and P15 (the revised Figures 3 and 6). According to the new results, we found that the dramatic decreases of the row 1 tip signal for ESPN1 and EPS8 started since P4 in Srf cKO IHCs, is consistent with the appearance of the mild reduction of row 1 stereocilia length in P5 Srf cKO IHCs. For Mrtfb cKO hair cells, the obvious reduction of the row 1 tip signal for ESPN1 was observed until P10. However, a few genes related to cell adhesion and regulation of actin cytoskeleton were significantly down-regulated in P2 Mrtfb deficient hair cell transcriptome. We think that in hair cells the MRTFB may not play a major role in the regulation of stereocilia development, so the morphological defects of stereocilia happened much later in the Mrtfb mutant than in the Srf mutant.

While it is certainly interesting if it turns out CNN2 is indeed at tips in this phase, the experiments do not tell us that much about what role CNN2 may be playing. It is notable that in Fig 7E in the control+GFP panel, CNN2 does not appear to be at the tips. Those images are at P11 whereas the images in panel A are at P6 so perhaps CNN2 decreases after the widening phase. An important missing control is the Anc80L65-Cnn2 AAV in a wild-type cochlea.

We agree with the reviewer. We have conducted more immunostaining experiments to confirm the expression pattern of CNN2 during the stereocilia development, from P0 to P11. The results were included in the revised Figure 9-figure supplement 1B. As the reviewer suggested, CNN2 expression pattern in control cochlea injected with Anc80L65-Cnn2 AAV has also been provided in revised Figure 9E.

Author Response

Reviewer #1 (Public Review):

This is an awesome comprehensive manuscript. Authors start by sorting putative stromal cellcontaining BM non-hematopoietic (CD235a-/CD45-) plus additional CD271+/CD235a/CD45- populations to identify nine individual stromal identities by scRNA-seq. The dual sorting strategy is a clever trick as it enriches for rare stromal (progenitor) cell signals but may suffer a certain bias towards CD271+ stromal progenitors. The lack of readable signatures already among CD45-/CD45- sorts might argue against this fear. This reviewer would appreciate a brief discussion on number & phenotype of putative additional MSSC phenotypes in light of the fact that the majority of 'blood lineage(s)'-negative scRNA-seq signatures identified blood cell progenitor identities (glycophorin A-negative & leukocyte common antigen-negative). The nine stromal cell entities share the CXCL12, VCAN, LEPR main signature. Perhaps the authors could speculate if future studies using VCAN or LEPRbased sort strategies could identify additional stromal progenitor identities?

We would like to thank the reviewer for critically evaluating our work and for the generally positive evaluation of the paper. We apologize for delayed resubmission as it took a long time for a specific antibody to arrive to complete the confocal microscopy analyses.

The reviewer asks for a brief discussion on the cell numbers and phenotypes of MSSC phenotypes. The cell numbers and percentages of MSSC in sorted CD45low/-CD235a- and CD45low/-CD235a-CD271+ cells can be found in Supplementary File 3 and we have added a summary of the phenotypes of MSSC in the new Supplementary File 7.

Due to the extremely low frequency of stromal cells in human bone marrow, we chose a sorting strategy that also included CD45low cells (Fig 1A) to ensure that no stromal cells were excluded from the analysis. Although stromal elements are certainly enriched using this approach, the CD45low population contains several different hematopoietic cell types. These include CD34+ HSPCs which are characterized by low CD45 expression2, as well as the CD45low-expressing fractions of other hematopoietic cell populations such as B cells, T cells, NK cells, megakaryocytes, monocytes, dendritic cells, and granulocytes. Furthermore, CD235a- late-stage erythroid progenitors, which are negative for CD45, are represented as well. Of note, our data are consistent with previously reported murine studies showing the presence of a number of hematopoietic populations in CD45- cells, which accounted for the majority of CD45-Ter119-CD31- murine BM cells3,4. However, despite a certain enrichment of stromal elements in the CD45low cell fraction, frequencies were still too low to allow for a detailed analysis of this important bone marrow compartment. This prompted us to adopt the stromal cell-enrichment strategy as described in the manuscript to achieve a better resolution of the stromal compartment. In fact, sorting based on CD45low/-CD235a-CD271+ allowed us to sufficiently enrich bone marrow stromal cells to be clearly detectable in scRNAseq analysis. According to the reviewer’s suggestion, a brief discussion on this issue is now included in the Discussion (page 28, lines 10-15).

The reviewer also suggested using VCAN or LEPR-based sorting strategy to identify additional stromal identities in future studies.

However, as an extracellular matrix protein, FACS analysis of cellular VCAN expression can only be achieved based on its intracellular expression after fixation and permeabilization5,6. Additionally, while VCAN is highly and ubiquitously expressed by stromal clusters, VCAN is also expressed by monocytes (cluster 36). Therefore, VCAN is not an optimal marker to isolate viable stromal cells.

LEPR is the marker that was reported to identify the majority of colony-forming cells in adult murine bone marrow7. We have previously reported that the majority of human adult bone marrow CFU-Fs is contained in the LEPR+ fraction 8. In our current scRNAseq surface marker profiling analysis, group A cells showed high expression of several canonical stromal markers including VCAM1, PDGFRB, ENG (CD73), as well as LEPR (Fig. 4A). However, the four stromal clusters in Group A could not be separated based on the expression of LEPR. Therefore, we chose not to use LEPR as a marker to prospectively isolate the different stromal cell types.

The authors furthermore localized CD271+, CD81+ and NCAM/CD56+ cells in BM sections in situ. Finally, referring to the strong background of the group in HSC research, in silico prediction by CellPhoneDB identified a wide range of interactions between stromal cells and hematopoietic cells. Evidence for functional interdependence of FCU-F forming cells is completing the novel and more clear bone marrow stromal cell picture.

We thank the reviewer for the positive comments.

An illustrative abstract naming the top9 stromal identities in their top4 clusters by their "top10 markers" + functions would be highly appreciated.

We thank the reviewer for the suggestion. A summary of the characteristics of stromal clusters is now shown in the new Supplementary File 7, which we hope matches the reviewer’s expectations.

Reviewer #2 (Public Review):

Knowledge about composition and function of the different subpopulations of the hematopoietic niche of the BM is limited. Although such knowledge about the mouse BM has been accumulating in recent years, a thorough study of the human BM still needs to be performed. The present manuscript of Li and coworkers fills this gap by performing single cell RNA sequencing (scRNAseq) on control BM as well as CD271+ BM cells enriched for non-hematopoietic niche cells.

We apologize for delayed resubmission as it took a long time for a specific antibody to arrive to complete the confocal microscopy analyses. We thank the reviewer for the critical expert review and overall positive comments.

Based on their scRNAseq, the authors propose 41 different BM cell populations, ten of which represented non-hematopoietic cells, including one endothelial cell cluster. The nine remaining skeletal subpopulations were subdivided into multipotent stromal stem cells (MSSC), four distinct populations of osteoprogenitors, one cluster of osteoblasts and three clusters of pre-fibroblasts. Using bioinformatic tools, the authors then compare their results and divisions of subpopulations to some previously published work from others and attempt to delineate lineage relationships using RNA velocity analyses. From these, they propose different paths from which MSSC enter the progenitor stages, and might differentiate into pre-osteoblasts and -fibroblasts.

It is of interest to note, that apparently adipo-primed cells may also differentiate into osteolineage cells, something that should be further explored or validated. Furthermore, although this analysis yields a large adipo-primed populations, pre-adipocytes and mature adipocytes appear not to be included in the data set the authors used, which should also be explained.

We thank the reviewer for this comment. We chose to annotate Cluster 5 as adipoprimed cluster based on the higher expression of adipogenic differentiation markers as well as a group of stress-related transcription factors (FOS, FOSB, JUNB, EGR1) (Fig. 2B-C, Figure 2-figure supplement 1C) some of which had been shown to mark bone marrow adipogenic progenitors1. Although at considerably lower levels compared to adipogenic genes, osteogenic genes were also expressed in cluster 5 cells (Fig. 2B and D), indicating the multi-potent potential of this cluster. Therefore, our initial annotation of these cells as adipoprimed progenitors was too narrow as it did not include the possible osteogenic differentiation potential. We apologize for the confusion caused by the inappropriate annotation and, in order to avoid any further confusion, cluster 5 has now been re-annotated as ‘highly adipocytic gene-expressing progenitors (HAGEPs), which we believe is a better representation of the cells. We furthermore agree with the reviewer that in-vivo differentiation needs to be performed to address potential differentiation capacities in future studies.

With regard to the lack of adipocytes in our data set, we described in the Materials and Methods section that human bone marrow cells were isolated based on density gradient centrifugation. After centrifugation, the mononuclear cell-containing monolayers were harvested for further analysis. However, the resulting supernatant containing mature adipocytic cells was discarded14. Therefore, adipocyte clusters were not identified in our dataset. We have amended the manuscript accordingly (page 5, line 7).

Regarding the pre-adipocytes, we are not aware of any specific markers for pre-adipocytes in the bone marrow. We examined the only known markers (ICAM1, PPARG, FABP4) that have been shown to mark committed pre-adipocytes in human adipose tissue15. As illustrated in Fig. R1 (below), low expression of all three markers was not restricted to a single distinct cluster but could be found in almost all stromal clusters. These data thus allow us to neither confirm nor exclude the presence of pre-adipocytes in the dataset. Due to the lack of specific markers for pre-adipocytes and the absence of mature adipocytes in the current dataset, it is therefore difficult to identify a well-defined pre-adipocytes cluster.

Figure R1. UMAP illustration of the normalized expression of the markers for pre-adipocytes in stromal clusters.

In addition, based on a separate analysis of surface molecules, the authors propose new markers that could be used to prospectively isolate different human subpopulations of BM niche cells by using CD52, CD81 and NCAM1 (=CD56). Indeed, these analyses yield six different populations with differential abilities to form fibroblast-like colonies and differentiate into adipo-, osteo-, and chondrogenic lineages. To explore how the scRNAseq data may help to understand regulatory processes within the BM, the authors predict possible interactions between hematopoietic and non-hematopoietic subpopulations in the BM. These should be further validated, to support statements as the suggestion in the abstract that separate CXCL12- and SPP1-regulated BM niches might exist.

We agree with the reviewer that functional validation of the CellPhoneDB results using for example in vivo humanized mouse models would be needed to demonstrate the presence of different niches in the bone marrow. At this point of time we only put forward the hypothesis that different niche types exist while we will work on providing experimental proof in our future studies.

The scRNAseq analysis is indeed a strong and important resource, also for later studies meant to increase knowledge about the hematopoietic niche of the BM. Although the analyses using different bioinformatic tools is very helpful, they remain mostly speculative, since validatory experiments, as already mentioned, are missing. As such, I feel the authors did not succeed in achieving their goals of understanding how non-hematopoietic cells of the BM regulate the different hematopoietic processes within the BM. Nevertheless, they have created valuable resources, both in the scRNAseq data they generated, as well as the different predictions about different cell populations, their lineage relationships, and how they might interact with hematopoietic cells.

We thank the reviewer for the appreciation of the value of this dataset. We agree with the reviewer that it is of great importance to validate the contribution of potential driver genes for stromal cell differentiation and verify the in vitro data and in-silico prediction using in-vivo models. As the main goal of the current study was to formulate hypotheses based on the scRNAseq data for future studies, we believe that in vivo validation experiments using engineered human bone marrow models or humanized bone marrow ossicles are out of the scope of the current study, but certainly need to be performed in the future.

The impact of this work is difficult to envision, since validations still need to be performed. Also, it has the born in mind that humans are not mice, which can be studied in neat homogeneous inbred populations. Human populations on the other hand, are quite diverse, so that the data generated in this manuscript and others will probably have to be combined to extrapolate data relevant to the whole of the human population. However, as it is equally difficult to generate reliable scRNAseq data from human BM, it seems likely that the data will indeed an important resource, when more data from different donors become available.

We thank the reviewer for the generally positive evaluation of this study.

Taken at point value, the authors provide evidence that human counterparts exist to several BM populations described in mice. In my opinion, the lineage relationships predicted using the RNA velocity analyses need more substance, as it seems the differentiation-paths may diverge from what is known from mice. If so, this issue should be studied more stringently. Similarly, the paper would have been strengthened considerably if a relevant experimental validation would have been attempted, perhaps by using genetically modified (knockdown) MSSC, similar to Battula et al. (doi: 10.1182/blood-2012-06-437988).

In the study from Welner’s group, stromal differentiation trajectory was inferred based on scRNAseq analysis of murine bone marrow cells using Velocyto16. Velocyto identified MSCs as the ‘source’ cell state with pre-adipocytes, pro-osteoblasts, and prochondrocytes being end states. In our study, the MSSC population was predicted to be at the apex of the trajectory and the pre-osteoblast cluster was placed close to the terminal state of differentiation, which is consistent with the murine study. However, different stromal cell types were identified in mice compared with humans. For example, we have identified prefibroblasts in our dataset which are absent in the murine study, while a well-defined murine pre-adipocyte population was not identified in our human dataset. Therefore, it is not surprising to find some discrepancies between human and murine stromal differentiation trajectories. Of course and as mentioned before, critical in-vivo functional validations need to be carried out to address these important issues in the future.

In summary, this is a very interesting but also descriptive paper with highly important resources. However, to prospectively identify or isolate human non-hematopoietic/nonendothelial niche populations, more stringent validations should have been performed to strengthen the validity of the different analyses that have been performed. As such, it remains an open question which niche subpopulations has the most impact on the different hematopoietic processes important for normal and stress hematopoiesis, as well as malignancies.

Thank you for this comment. We completely agree that more stringent validations are necessary but are outside of the aim of our current hypothesis-generating study. Accordingly, we are planning functional verification studies using genetically manipulated stromal cells in combination with in-vivo humanized ossicles. Furthermore, other groups will hopefully use our database and contribute with functional studies in model systems that are currently not available to us, e.g. iPS-derived bone marrow in-vitro proxies.

Specific remarks

• Since CD45, CD235a, and CD271 are used as distinguishing markers in the sample preparation of the scRNAseq, it would be helpful to highlight these markers in the different analyses (Figures 1D, 2B, 2C-F, and 4A), and restrict the analyses to those cells that also not express CD45, CD235a (why use CD71?) and highly express CD271.

Thank you for this comment. As shown in Fig. R2, we have modified figures Fig. 1D, 2B, and 4A showing now also the expression of PTPRC (CD45), GYPA (CD235a), and NGFR (CD271) on the top (Fig. 1D and 2B) or right (Fig. 4A) panel of the figures. To complement Fig. 2C-F, we have generated new stacked violin plots showing the expression level of three markers by all 9 stromal clusters (Fig. R2B). As we believe that including these three markers in the figures does not provide a better strategy to improve the analyses, we decided to leave the original figures unchanged in this respect.

Figure R2. (A) Modified Fig. 1D, 2B and 4A with PTPRC (CD45), GYPA (CD235a) and NGFR (CD271) expression. (B) Stacked violin plots of PTPRC, GYPA and NGFR expressed by stromal clusters to complement Fig. 2C-F.

With regard to cell exclusion based on CD45, as shown in the modified Figure corresponding to Fig 1A in the manuscript (Fig R2A), CD45 gene expression is observed also in the endothelial cluster, basal cluster, and neuronal cluster (Fig. R2A). These clusters represent non-hematopoietic clusters that we would like to keep in our dataset for further analysis, such as cell-cell interaction. Therefore, we choose to not restrict the analysis to solely CD45 nonexpressing cells.

With regard to CD235a (GYPA), expression of CD235a is not detected in any of the nonhematopoietic clusters. Thus, CD235a-expressing cell exclusion is not necessary.

For CD271, according to our previous results (own unpublished data, belonging to a dataset of which only significantly expressed genes were reported in Li et al.8), protein expression of CD271 is not necessarily reflected by gene expression. In the other words, stromal cells with CD271 protein expression do not always have high mRNA expression. A significant fraction of stromal cells would be excluded if we restrict the analyses only to those cells that show high CD271 gene expression, which would not reflect the real cellular composition of human bone marrow stroma. In order to not risk losing stromal cells, we therefore kept our previous analyses which included stromal cells with various CD271 expression levels.

With regard to using CD71 as an exclusion marker, please see also the comments to reviewer 1. Briefly, according to our data, CD71 (TFRC)-expressing erythroid precursors could still be found after excluding CD45 and CD235a positive cells (Figure 1-figure supplement 1B and R3). As furthermore shown in Figure 1-figure supplement 1G and R2, CD71 expression in the stromal clusters is negligible. Therefore, we believe that this justifies the use of CD71 as an additional marker to exclude erythroid cells. We have amended the discussion to address this issue (page 19, lines 7-8).

Figure R3. FACS plots illustrating the expression of (A) CD71 (TFRC) vs CD271 in CD45- CD235a- cells and (B) FSC-A vs CD81 in CD45-CD235a-CD271+CD71+ cells following exclusion of doublets and dead cells.

• Despite a distinct neuronal cluster (39), there does not seem to be a distinctive marker for these cells. Is this true?

Yes, the reviewer is correct that there is no significantly-expressed distinctive marker for neuronal cells. Multiple markers indicating the presence of different cell types were identified in cluster 39 (Supplementary File 4). Among them, several neuronal markers (NEUROD1, CHGB, ELAVL2, ELAVL3, ELAVL4, STMN2, INSM1, ZIC2, NNAT) were found to be enriched in this cluster (Supplementary File 4 and Fig. 1D) with higher fold changes compared to other identified genes. However, the expression of these genes was not statistically significant, which is mainly due to the heterogeneity of the cluster and thus does not allow us to draw any firm conclusions.

Several genes including MALAT1, HNRNPH1, AC010970.1, and AD000090.1 were identified to be statistically highly expressed by cluster 39 (Supplementary File 4). The expression of these genes is not restricted to any specific cell type. It is therefore impossible to annotate the cluster based on this and our data thus indicated that cluster 39 is a heterogeneous population containing multiple cell types. Based on the expression of neuronal markers, we nevertheless chose to annotate Cluster 39 as “neuronal” as the prominent expression of neuronal markers indicated the presence of neurons in this cluster. To be more accurate, the annotation of cluster 39 has been changed to ‘neuronal cell-containing cluster’ to correctly reflect the presence of non-neuronal gene expressing cells as well (page 29, lines 3-8).

• Since based on 2C and 2D, the authors are unable to distinguish adipo- from osteogenic cells, would the authors use the same molecules to distinguish different populations of 2C-D, or would they use other markers, if so which and why.

We agree with the reviewer that at the first glance adipo-primed (cluster 5, now annotated as “highly adipocytic gene-expressing progenitors”, HAGEPs), balanced progenitors (cluster 16), and pre-osteoblasts (cluster 38) shared a similar expression pattern according to the violin plots in Fig. 2C and 2D. However, as illustrated in the heatmap (Fig. 2B), the expression patterns of adipo-primed (HAGEP) and balanced progenitors were quite different in terms of their expression of adipogenic and osteogenic markers. Both adipogenic and osteogenic marker expression was detected in HAGEPs, balanced progenitors, and preosteoblasts. Thus, as violin plots are summarizing the overall expression levels of a certain marker in a certain cluster, these plots tend to make it more difficult to detect differential expression patterns between different clusters. In this case, the heatmap shown in Fig. 2B is a good complement to the violin plots as it is demonstrating the different expression patterns of every cell in the different stromal clusters.

Additionally, cluster 5 showed the expression of a group of stress-related transcription factors (FOS, FOSB, JUNB, EGR1) (Fig. 2B and Figure 2-figure supplement 1C), some of which had been shown to mark bone marrow adipogenic progenitors1. The expression of the abovementioned stress-related transcription factors (putative adipogenic progenitor markers) was generally lower in cluster 38 compared to cluster 5, further demonstrating that clusters were different.

Furthermore, there was a gradual upregulation of more mature osteogenic markers such as RUNX1, CDH11, EBF1, and EBF3 from cluster 5 to cluster 16 and finally cluster 38. As shown in Fig. 2D, the expression of these markers was higher in cluster 38 compared to cluster 5. Therefore, cluster 38 was annotated as pre-osteoblasts.

Most of the stromal clusters form a continuum (Fig. 2A), which correlates very well with the gradual transition of different cellular states during stromal cell development. It is highly unlikely that abrupt and dramatic gene expression changes would occur during the cellular state transition of cells of the same lineage. Therefore, it is not surprising to find the differences in gene expression profiles between stromal clusters share a certain level of similarities.

In summary, we rely on several factors to distinguish different stromal clusters, which include canonical adipo-, osteo- and chondrogenic markers, stress markers, heatmap, violin plots, and the gradual up-regulation of certain lineage-specific markers.

To directly answer the reviewer’s question, we believe that we are able to distinguish different stromal clusters based on our data.

• In de Jong et al., an inflammatory MSC population (iMSC) is defined. Since the Schneider group showed that inflammatory S100A8 and A9 are expressed by inflamed MSC, is it possible that the some of the designated pre-fibroblasts actually correspond to these S100A8/A9-expressing iMSC?

We thank the reviewer for raising this interesting question.

First of all, we would like to point out that scRNAseq was performed using viably frozen bone marrow aspirates in de Jong’s study while freshly isolated bone marrows were used in our study. There might be discrepancies between frozen and fresh bone marrow samples in terms of cellular composition including stromal composition and, importantly, processinginduced stress-related gene expression profiles.

To investigate if designated pre-fibroblasts actually correspond to iMSCs as suggested by the reviewer, we have re-examined the expression of some of the key iMSC genes as reported by de Jong et al 17. As shown in Fig. R6, the markers that can distinguish iMSC from other MSC clusters in de Jong et al. study were not exclusively expressed by pre-fibroblasts, but also by other stromal cell types including HAGEPs, balanced progenitors, and pre-osteoblasts.

In the study by R. Schneider’s group18, significant upregulation of S100A8/S100A9 was observed in stromal cells from patients with myelofibrosis. Furthermore, base-line expression of S100A8/A9 was also observed in the fibroblast clusters in the control group, which correlates very well with our data of S100A8/9 expression in pre-fibroblasts in normal donors (Fig. 2F). Our data thus indicate – in line with Schneider’s findings - that there is a baseline level expression of S100A8/9 in fibroblasts in hematologically normal samples and that the expression of S100A8/9 is not restricted to inflamed MSC.

In summary, the gene expression profiles observed in our study do not indicate the presence of iMSC in the healthy bone marrow.

• Figure 3A: Do human adipo-primed cells (cluster 5) indeed differentiate into osteogenic cells (clusters 6, 38, and 39). This would be highly unexpected. Can the authors substantiate this "reliable outcome of the RNA velocity analysis"?

Please refer to our previous responses regarding this topic. Briefly, as shown in Fig. 2B and D, both osteogenic and adipogenic genes are expressed in cluster 5, indicating the multi-potent potentials of this cluster. Although the cluster was initially annotated as adipo-primed progenitors, this was not intended to exclude the osteogenic differentiation potential of these progenitors. Nevertheless, this annotation did not correctly reflect the differentiation potential and might thus have caused confusion, for which we apologize. In order to more correctly describe the characteristics of these cells, cluster 5 has now been reannotated as ‘highly adipocytic gene-expressing progenitors (HAGEPs)’.

In general, the outcome of the RNA velocity analysis needs to be corroborated by in-vivo differentiation experiments. But we believe that functional verification, which would be extensive, is out of the scope of the current study and we will address these questions in future studies.

• How statistically certain are the authors, that the populations in Figure 4B as defined by flow cytometry, correspond to MSSC, adipo-primed cells, osteoprogenitors, etc., as defined by scRNAseq?

To address this question, we sorted the A1-A4 populations and performed RT- PCR to examine the CD81 expression level in each cluster. As shown in Figure 4-figure supplement 1B, CD81 expression levels were higher in A1 and A2 compared with A3 and A4, which is consistent with the scRNAseq data that showed the highest CD81 expression in MSSCs compared to other clusters (Supplementary File 4).

The phenotypes defined in this study allowed us to isolate different stromal cell types which demonstrated significant functional differences as described in the manuscript (page 19, lines 17-25; page 20, lines 1-11). These results, in combination with the quantitative real-time PCR results (Figure 4-figure supplement 1B), demonstrated that the A1-A4 subsets in FACS are functionally distinct populations and are likely to be – at least in large parts – identical or equivalent to the transcriptionally identified clusters in group A stromal cells. However, at this point, we do not have performed the required experiments (scRNAseq of sorted cells) that would provide sufficient proof to confirm this statement statistically.

• The immunohistochemistry results shown do not allow distinct conclusions as the colors give unequivocal mix-colors, and surface expression cannot be distinguished from intracellular expression. Please use a 3D (confocal) method for such statements.

We thank the reviewer for the suggestion and we have performed additional confocal microscopy analysis of human bone marrow biopsies as suggested by the reviewer. Representative confocal images are now presented in the middle and right panel of Fig. 6E. We also include a separate file (Supplemental confocal image file). Here, confocal scans of all maker combinations are shown as ortho views in addition to detailed intensity profile analyses of the cells of interest clearly distinguishing surface staining from intracellular staining.

Confocal analysis of bone marrow biopsies confirmed our findings presented in the manuscript. As observed in the scanning images, CD271-expressing cells were negative for CD45 and were located in perivascular, endosteal, and peri-adipocytic regions. CD271/CD81double positive cells could be found either in the peri-adipocytic regions or perivascular regions while CD271/NCAM1 double-positive cells were exclusively situated at the bone-lining endosteal regions. The results of the confocal analysis have been added to the revised manuscript (page 21, lines 15-17).

• Figure 5A: as all cells seem to interact with all other cells, this figure does not convey relevant information about BM regions using for instance CXCL12 or SPP1. Please reanalyze to show specificity of the interactions of the single clusters. Also, since it is unlikely the CellPhoneDB2-predicted interactions are restricted to hematopoietic responders, please also describe the possible interactions between non-hematopoietic cells.

Fig. 5A was used to demonstrate the complexity of the interactions between hematopoietic cells and stromal cells.

To gain a more detailed understanding of the interactions, we also performed an analysis with the top-listed ligand-receptor pairs as shown in Fig. 5B-C and Figure 5-figure supplement 1B. Here, each dot represents the interaction of a specific ligand-receptor pair listed on the x-axis between the two individual clusters indicated in the y-axis, which we believe shows what the reviewer is asking for.

The specificity of the interactions between single clusters were shown in Fig. 5B-C and Figure 5-figure supplement 1B. The CXCL12- and SPP1-mediated interactions between MSSC/OC and hematopoietic clusters clearly suggested stromal cell type-specific interactions.

Regarding non-hematopoietic cells, both inter- and intra-stromal interactions were identified to be operative between different stromal subsets as well as within the same stromal cell population as shown in Figure 5-figure supplement 3B. In addition, we have also analyzed the interaction pattern between endothelial cells and hematopoietic cells as shown in Fig. 7A, and thus we believe that we have sufficiently described these interactions as requested by the reviewer.

Author Response

Reviewer #2 (Public Review):

Point 1: The transcriptomic analysis of E12.5 endocardial cushion cells in the various mouse models is informative in the extraction of Igf2- and H19-specific gene functions. In Fig. 6D, a huge sex effect is obvious with many more DEGs in female embryos compared to males. How can this be explained given that Igf2/H19 reside on Chr7 and do not primarily affect gene expression on the X chromosome? Is any chromosomal bias observed in the genomic distribution of DEGs?

We examined chromosomal distribution of DEGs between WT and +/hIC1 (Supplemental Figure 6D) and did not see any bias on X chromosome. We described this result on lines 278-280: “Although the number of +/hIC1-specific DEGs largely differed between males and females, there was no sex-specific bias on the X chromosome (Supplemental Figure 6D).” Additionally, we agree with the reviewer that it is noteworthy that the dysregulated H19/Igf2 expression affected transcriptome in a sex-specific manner, especially when the mutation is located on a somatic chromosome. Although investigating the role of hormones versus sex chromosome in these effects would be quite interesting, it is beyond the scope of current study.

Point 2: A separate issue is raised by Fig. 6E that shows a most dramatic dysregulation of a single gene in the delta3.8/hIC1 "rescue" model. Interestingly, this gene is Shh. Hence, these embryos should exhibit some dramatic skeletal abnormalities or other defects linked to sonic hedgehog function.

The reason why Shh appeared to be differentially expressed between wild-type and d3.8/hIC1 samples was that Shh expression was 0 across all the samples except for two wild-type samples. In order to detect all the DEGs that might be lowly expressed, we did not want to filter DEGs based on the level of total expression. As a result, Shh was represented as significantly differently expressed in d3.8/hIC1 samples, although its expression in our samples appears to be too low to have any significant effect on development. This explanation was added to lines 310-312. To confirm that this was an exceptional case, we analyzed the expression of DEGs obtained from other pairwise comparisons. In the volcano plots below, genes of which expression is not statistically different between two groups are marked grey. Genes of which expression is statistically different and detected in both groups are marked red. Genes with statistically different but not detected in one group at all, such as Shh, are marked blue (Figure G). It is clear that that almost all of our DEGs are expressed consistently across the groups, and genes with no expression detected in one group are very rare.

Point 3: The placental analysis needs to be strengthened. Placentas should be consistently positioned with the decidua facing up, and the chorionic plate down. The placentas in Fig. 3F are sectioned at an angle and the chorionic plate is missing. These images must be replaced with better histological sections.

As requested, we have replaced placental images with better representative sections (Figure 3F and 4E). In addition, we have improved alignment of placental histology figures.

Point 4: The CD34 staining has not worked and does not show any fetal vasculature, in particular not in the WT sample.

As requested, we have replaced the CD34 vascular stained images with those that better represent fetal vasculature (Figure 3G).

Point 5: The "thrombi" highlighted in Fig. 4E are well within the normal range, to make the point that these are persistent abnormalities more thorough measurements would need to be performed (number, size, etc).

As requested, we measured the number and relative size of the thrombi that are found in dH19/hIC1 placentas with lesions. No thrombi were found in wild-type placentas whereas an average of 1.3 thrombi were found in six dH19/hIC1 placentas. The size of the thrombi widely varied, but occupied average of 2.58% of the labyrinth zone where these lesions were found (Supplemental Figure 4D). Additionally, we replaced the image in Figure 4E into the section that better represents the lesion.

Point 6: The statement that H19 is disproportionately contributing to the labyrinth phenotype (lines 154/155) is not warranted as Igf2 expression is reduced to virtually nothing in these mice. Even though there is more H19 in the labyrinth than in the junctional zone, the phenotype may still be driven by a loss of Igf2. Given the quasi Igf2-null situation in +/hIC1 mice, is the glycogen cell type phenotype recapitulated in these mice, and how do glycogen numbers compare in the other mouse models?

The sentence was edited in line 157. We performed Periodic acid Schiff (PAS) staining on +/hIC1 placentas to address if glycogen cells are affected by abnormal H19/Igf2 expression (Supplemental Figure 1E). In contrary to previous reports where Igf2-null mice had lower placental glycogen concentration (Lopez et al., 1996) and H19 deletion led to increased placental glycogen storage (Esquiliano et al., 2009), our quantification on PAS-stained images showed that the glycogen content is not significantly different between wild-type and +/hIC1 placentas. We have described this result in lines 166-168.

Point 7: How do delta3.8/+ and delta3.8/hIC1 mice with a VSD survive? Is it resolved some time after birth such that heart function is compatible with postnatal viability? And more importantly, do H19 expression levels correlate with phenotype severity on an individual basis?

Our study was limited to phenotypes prior to birth, thus postnatal/adult phenotypes were not examined. Because the VSD showed only partial penetrance in these mice, we cannot state that the d3.8/+ or d3.8/hlC1 mice with VSDs survive. It has also been previously reported in another mouse model with incomplete penetrance of a VSD that the mice which survived to adulthood did not have the VSDs (Sakata et al., 2002). We find it highly unlikely that either mouse model would survive significantly past the postnatal timepoint with a VSD. We have examined two PN0 d3.8/hIC1 neonates, and both did not have VSD.

Regarding the second point, the only way to quantitatively address this question would be to do qPCR or RNA-seq on individual hearts, which then makes it impossible for those hearts to be examined for histology to confirm the VSD. Thus, hearts used to identify VSDs via histology could not also be used for quantitative H19 measurements. One thing to note is that the H19/Igf2 expression in independent replicates of d3.8/hIC1 cardiac ECs used in our RNA-seq experiment is quite variable, not clustering together in contrast to other mouse models used in this study (Fig. 6A). Such wide range of variability in the extent of H19/Igf2 dysregulation suggests that H19/Igf2 levels could have an impact on the penetrance or the severity of the VSD phenotype in d3.8/hIC1 embryos.

Author Response

Reviewer #2 (Public Review):

Weaknesses:

1) The relevance of the LPS-induced calvarial osteolysis model is not clear. Calvaria is mostly composed of cortical bone-like structures lacking marrow space, though small marrow space exists near the suture. Osteolysis appears to occur in areas apart from where marrow is located. The authors did not show in the manuscript which cells Adipoq-Cre marks in the calvaria.

We have shown in a recent publication that MALPs exist in the calvarial bone marrow (2). As shown in Fig. R1A, Td+ cells are layer of cortical bone (Fig. R1B, blue arrows). In WT mice, after LPS injection, the normal bone structure, including suture and cortical bone, were mostly eroded, and filled with inflammatory cells (green arrows). Thus, osteolysis does occur at the area where bone marrow is originally located. On the contrary, calvarial bone structure was preserved in the CKO mice, demonstrating that Csf1 deficiency in MALPs suppresses LPS-induced osteolysis. We included the H&E staining data in the revised manuscript:

"H&E staining showed that calvarial bone marrow is surrounded by a thin layer of cortical bone (Fig. 5C). After the LPS injection, normal calvarial structure, including suture and cortical bone, were mostly eroded and filled with inflammatory cells in WT mice, but unaltered in CKO mice."

Figure R1. Calvarial bone marrow structure. (A) Representative coronal section of 1.5-month-old Adipoq/Td mouse calvaria. Bone surfaces are outlined by dashed lines. Boxed areas in the low magnification image (top) are enlarged to show periosteum (bottom left), suture (bottom middle), and bone marrow (BM, bottom right) regions. Red: Td; Blue: DAPI. Adopted from our previous publication (2). (B) H&E staining of coronal sections of WT and Csf1 CKOAdipoq mice after LPS injection. Blue arrows point to bone marrow space close to suture (indicated by *). Green arrows point to the osteolytic lesion where cortical bone was eroded, and the space were filled with inflammatory cells.

2) Although the contrast between the two Csf1 conditional deletion models (Adipoq-Cre and Prx1-Cre) is very interesting, the relationship between these two cell populations are not well described. The authors did not clarify if MALPs are also targeted by Prx1-Cre, or these two cell types are from different cell lineages. "Other mesenchymal lineage cells" in the subtitle is not extremely helpful to place this finding in context.

We thank the Reviewer for this comment. The original article constructing Prx1-Cre mouse line demonstrates that Prx1-Cre targets all mesenchymal cells in the limb bud at early as 10.5 dpc (10). This early expression pattern ensures that all bone marrow mesenchymal lineage cells, including MALPs, are targeted by Prx1-Cre. In addition, based on our scRNA-seq data (1), Adipoq is mainly expressed in MALPs, while Prrx1 (Prx1) is highly expressed not only in MALPs but also in EMPs, IMPs, LMPs, LCPs, and OBs (Fig. R2). Thus, the fact that Prx1-Cre driven CKO mice have much more severer bone phenotypes than AdipoqCre driven CKO mice indicates that mesenchymal lineage cells other than MALPs also contribute Csf1 to regulate bone resorption. To avoid confusion, we changed the title and the first sentence in the Result session about Prx1 mice to the following:

"Csf1 from mesenchymal lineage cells other than MALPs regulate bone structure.

To explore whether Csf1 from MALPs plays a dominant role in regulating bone structure, we generated Prx1-Cre Csf1flox/flox (Csf1 CKOPrx1) mice to knockout Csf1 in all mesenchymal lineage cells in bone (10), including MALPs."

Figure R2. Dotplot of Prrx1 and Adipoq expression in bone marrow mesenchymal lineage cells based on our scRNA-seq analysis of 1-month-old mice.

3) The data supporting defective bone marrow hematopoiesis in Csf1 CKO mice are not particularly strong. They observed a reduction in bone marrow cellularity, but this was only associated with an expected reduction in macrophages and a mild reduction in overall HSPC populations. More in-depth analyses might be required to define mechanisms underlying reduced bone marrow cellularity in CKO mice.

We thank the Reviewer for this constructive comment. Accordingly, we performed a thorough analysis of bone marrow hematopoietic compartments and observed significant decreases of monocytes and erythroid progenitors in CKO mice compared to WT mice. These results are now included as Fig. 6E.

4) Some of the phenotypic analyses are still incomplete. The authors did not report whether CHet (Adipoq-Cre Csf1(flox/+)) showed any bone phenotype. Further, the authors did not report whether Csf1 mRNA or M-Csf protein is indeed expressed by MALPs, with current evidence solely reliant on scRNAseq and qPCR data of bulk-isolated cells. More specific histological methods will be helpful to support the premise of the study.

A pilot microCT study revealed the same femoral trabecular bone structure in WT and Adipoq-Cre Csf1flox/+ (Csf1 Het) mice at 3 months of age (Fig. R3). While the sample number for Het is low, we are confident about this conclusion.

Figure R3. MicroCT measurement of trabecular bone structural parameters from WT and Csf1 Het mice. BV/TV: bone volume fraction; BMD: bone mineral density; Tb.N: trabecular number; Tb.Th: trabecular thickness; Tb.Sp: trabecular separation; SMI: structural model index. n=3-8 mice/group.

Author response:

Reviewer #1 (Public Review):

In this paper, Tompary & Davachi present work looking at how memories become integrated over time in the brain, and relating those mechanisms to responses on a priming task as a behavioral measure of memory linkage. They find that remotely but not recently formed memories are behaviorally linked and that this is associated with a change in the neural representation in mPFC. They also find that the same behavioral outcomes are associated with the increased coupling of the posterior hippocampus with category-sensitive parts of the neocortex (LOC) during a post-learning rest period-again only for remotely learned information. There was also correspondence in rest connectivity (posterior hippocampus-LOC) and representational change (mPFC) such that for remote memories specifically, the initial post-learning connectivity enhancement during rest related to longer-term mPFC representational change.

This work has many strengths. The topic of this paper is very interesting, and the data provide a really nice package in terms of providing a mechanistic account of how memories become integrated over a delay. The paper is also exceptionally well-written and a pleasure to read. There are two studies, including one large behavioral study, and the findings replicate in the smaller fMRI sample. I do however have two fairly substantive concerns about the analytic approach, where more data will be required before we can know whether the interpretations are an appropriate reflection of the findings. These and other concerns are described below.

Thank you for the positive comments! We are proud of this work, and we feel that the paper is greatly strengthened by the revisions we made in response to your feedback. Please see below for specific changes that we’ve made.

1) One major concern relates to the lack of a pre-encoding baseline scan prior to recent learning.

a) First, I think it would be helpful if the authors could clarify why there was no pre-learning rest scan dedicated to the recent condition. Was this simply a feasibility consideration, or were there theoretical reasons why this would be less "clean"? Including this information in the paper would be helpful for context. Apologies if I missed this detail in the paper.

This is a great point and something that we struggled with when developing this experiment. We considered several factors when deciding whether to include a pre-learning baseline on day two. First, the day 2 scan session was longer than that of day 1 because it included the recognition priming and explicit memory tasks, and the addition of a baseline scan would have made the length of the session longer than a typical scan session – about 2 hours in the scanner in total – and we were concerned that participant engagement would be difficult to sustain across a longer session. Second, we anticipated that the pre-learning scan would not have been a ‘clean’ measure of baseline processing, but rather would include signal related to post-learning processing of the day 1 sequences, as multi-variate reactivation of learned stimuli have been observed in rest scans collected 24-hours after learning (Schlichting & Preston, 2014). We have added these considerations to the Discussion (page 39, lines 1047-1070).

b) Second, I was hoping the authors could speak to what they think is reflected in the post-encoding "recent" scan. Is it possible that these data could also reflect the processing of the remote memories? I think, though am not positive, that the authors may be alluding to this in the penultimate paragraph of the discussion (p. 33) when noting the LOC-mPFC connectivity findings. Could there be the reinstatement of the old memories due to being back in the same experimental context and so forth? I wonder the extent to which the authors think the data from this scan can be reflected as strictly reflecting recent memories, particularly given it is relative to the pre-encoding baseline from before the remote memories, as well (and therefore in theory could reflect both the remote + recent). (I should also acknowledge that, if it is the case that the authors think there might be some remote memory processing during the recent learning session in general, a pre-learning rest scan might not have been "clean" either, in that it could have reflected some processing of the remote memories-i.e., perhaps a clean pre-learning scan for the recent learning session related to point 1a is simply not possible.)

We propose that theoretically, the post-learning recent scan could indeed reflect mixture of remote and recent sequences. This is one of the drawbacks of splitting encoding into two sessions rather than combining encoding into one session and splitting retrieval into an immediate and delayed session; any rest scans that are collected on Day 2 may have signal that relates to processing of the Day 1 remote sequences, which is why we decided against the pre-learning baseline for Day 2, as you had noted.

You are correct that we alluded to in our original submission when discussing the LOC-mPFC coupling result, and we have taken steps to discuss this more explicitly. In Brief, we find greater LOC-mPFC connectivity only after recent learning relative to the pre-learning baseline, and cortical-cortical connectivity could be indicative of processing memories that already have undergone some consolidation (Takashima et al., 2009; Smith et al., 2010). From another vantage point, the mPFC representation of Day 1 learning may have led to increased connectivity with LOC on Day 2 due to Day 1 learning beginning to resemble consolidated prior knowledge (van Kesteren et al., 2010). While this effect is consistent with prior literature and theory, it's unclear why we would find evidence of processing of the remote memories and not the recent memories. Furthermore, the change in LOC-mPFC connectivity in this scan did not correlate with memory behaviors from either learning session, which could be because signal from this scan reflects a mix of processing of the two different learning sessions. With these ideas in mind, we have fleshed out the discussion of the post-encoding ‘recent’ scan in the Discussion (page 38-39, lines 1039-1044).

c) Third, I am thinking about how both of the above issues might relate to the authors' findings, and would love to see more added to the paper to address this point. Specifically, I assume there are fluctuations in baseline connectivity profile across days within a person, such that the pre-learning connectivity on day 1 might be different from on day 2. Given that, and the lack of a pre-learning connectivity measure on day 2, it would logically follow that the measure of connectivity change from pre- to post-learning is going to be cleaner for the remote memories. In other words, could the lack of connectivity change observed for the recent scan simply be due to the lack of a within-day baseline? Given that otherwise, the post-learning rest should be the same in that it is an immediate reflection of how connectivity changes as a function of learning (depending on whether the authors think that the "recent" scan is actually reflecting "recent + remote"), it seems odd that they both don't show the same corresponding increase in connectivity-which makes me think it may be a baseline difference. I am not sure if this is what the authors are implying when they talk about how day 1 is most similar to prior investigation on p. 20, but if so it might be helpful to state that directly.

We agree that it is puzzling that we don’t see that hippocampal-LOC connectivity does not also increase after recent learning, equivalently to what we see after remote learning. However, the fact that there is an increase from baseline rest to post-recent rest in mPFC – LOC connectivity suggests that it’s not an issue with baseline, but rather that the post-recent learning scan is reflecting processing of the remote memories (although as a caveat, there is no relationship with priming).

On what is now page 23, we were referring to the notion that the Day 1 procedure (baseline rest, learning, post-learning rest) is the most straightforward replication of past work that finds a relationship between hippocampal-cortical coupling and later memory. In contrast, the Day 2 learning and rest scan are less ‘clean’ of a replication in that they are taking place in the shadow of Day 1 learning. We have clarified this in the Results (page 23, lines 597-598).

d) Fourth and very related to my point 1c, I wonder if the lack of correlations for the recent scan with behavior is interpretable, or if it might just be that this is a noisy measure due to imperfect baseline correction. Do the authors have any data or logic they might be able to provide that could speak to these points? One thing that comes to mind is seeing whether the raw post-learning connectivity values (separately for both recent and remote) show the same pattern as the different scores. However, the authors may come up with other clever ways to address this point. If not, it might be worth acknowledging this interpretive challenge in the Discussion.

We thought of three different approaches that could help us to understand whether the lack of correlations in between coupling and behavior in the recent scan was due to noise. First, we correlated recognition priming with raw hippocampal-LOC coupling separately for pre- and post-learning scans, as in Author response image 1:

Author response image 1.

Note that the post-learning chart depicts the relationship between post-remote coupling and remote priming and between post-recent coupling and recent priming (middle). Essentially, post-recent learning coupling did not relate to priming of recently learned sequences (middle; green) while there remains a trend for a relationship between post-remote coupling and priming for remotely learned sequences (middle; blue). However, the significant relationship between coupling and priming that we reported in the paper (right, blue) is driven both by the initial negative relationship that is observed in the pre-learning scan and the positive relationship in the post-remote learning scan. This highlights the importance of using a change score, as there may be spurious initial relationships between connectivity profiles and to-be-learned information that would then mask any learning- and consolidation-related changes.

We also reasoned that if comparisons between the post-recent learning scan and the baseline scan are noisier than between the post-remote learning and baseline scan, there may be differences in the variance of the change scores across participants, such that changes in coupling from baseline to post-recent rest may be more variable than coupling from baseline to post-remote rest. We conducted F-tests to compare the variance of the change in these two hippocampal-LO correlations and found no reliable difference (ratio of difference: F(22, 22) = 0.811, p = .63).

Finally, we explored whether hippocampal-LOC coupling is more stable across participants if compared across two rest scans within the same imaging session (baseline and post-remote) versus across two scans across two separate sessions (baseline and post-recent). Interestingly, coupling was not reliably correlated across scans in either case (baseline/post-remote: r = 0.03, p = 0.89 Baseline/post-recent: r = 0.07, p = .74).

Finally, we evaluated whether hippocampal-LOC coupling was correlated across different rest scans (see Author response image 2). We reasoned that if such coupling was more correlated across baseline and post-remote scans relative to baseline and post-recent scans, that would indicate a within-session stability of participants’ connectivity profiles. At the same time, less correlation of coupling across baseline and post-recent scans would be an indication of a noisier change measure as the measure would additionally include a change in individuals’ connectivity profile over time. We found that there was no difference in the correlation of hipp-LO coupling is across sessions, and the correlation was not reliably significant for either session (baseline/post-remote: r = 0.03, p = 0.89; baseline/post-recent: r = 0.07, p = .74; difference: Steiger’s t = 0.12, p = 0.9).

Author response image 2.

We have included the raw correlations with priming (page 25, lines 654-661, Supplemental Figure 6) as well as text describing the comparison of variances (page 25, lines 642-653). We did not add the comparison of hippocampal-LOC coupling across scans to the current manuscript, as an evaluation of stability of such coupling in the context of learning and reactivation seems out of scope of the current focus of the experiment, but we find this result to be worthy of follow-up in future work.

In summary, further analysis of our data did not reveal any indication that a comparison of rest connectivity across scan sessions inserted noise into the change score between baseline and post-recent learning scans. However, these analyses cannot fully rule that possibility out, and the current analyses do not provide concrete evidence that the post-recent learning scan comprises signals that are a mixture of processing of recent and remote sequences. We discuss these drawbacks in the Discussion (page 39, lines 1047-1070).

2) My second major concern is how the authors have operationalized integration and differentiation. The pattern similarity analysis uses an overall correspondence between the neural similarity and a predicted model as the main metric. In the predicted model, C items that are indirectly associated are more similar to one another than they are C items that are entirely unrelated. The authors are then looking at a change in correspondence (correlation) between the neural data and that prediction model from pre- to post-learning. However, a change in the degree of correspondence with the predicted matrix could be driven by either the unrelated items becoming less similar or the related ones becoming more similar (or both!). Since the interpretation in the paper focuses on change to indirectly related C items, it would be important to report those values directly. For instance, as evidence of differentiation, it would be important to show that there is a greater decrease in similarity for indirectly associated C items than it is for unrelated C items (or even a smaller increase) from pre to post, or that C items that are indirectly related are less similar than are unrelated C items post but not pre-learning. Performing this analysis would confirm that the pattern of results matches the authors' interpretation. This would also impact the interpretation of the subsequent analyses that involve the neural integration measures (e.g., correlation analyses like those on p. 16, which may or may not be driven by increased similarity among overlapping C pairs). I should add that given the specificity to the remote learning in mPFC versus recent in LOC and anterior hippocampus, it is clearly the case that something interesting is going on. However, I think we need more data to understand fully what that "something" is.

We recognize the importance of understanding whether model fits (and changes to them) are driven by similarity of overlapping pairs or non-overlapping pairs. We have modified all figures that visualize model fits to the neural integration model to separately show fits for pre- and post-learning (Figure 3 for mPFC, Supp. Figure 5 for LOC, Supp. Figure 9 for AB similarity in anterior hippocampus & LOC). We have additionally added supplemental figures to show the complete breakdown of similarity each region in a 2 (pre/post) x 2 (overlapping/non-overlapping sequence) x 2 (recent/remote) chart. We decided against including only these latter charts rather than the model fits since the model fits strike a good balance between information and readability. We have also modified text in various sections to focus on these new results.

In brief, the decrease in model fit for mPFC for the remote sequences was driven primarily by a decrease in similarity for the overlapping C items and not the non-overlapping ones (Supplementary Figure 3, page 18, lines 468-472).

Interestingly, in LOC, all C items grew more similar after learning, regardless of their overlap or learning session, but the increase in model fit for C items in the recent condition was driven by a larger increase in similarity for overlapping pairs relative to non-overlapping ones (Supp. Figure 5, page 21, lines 533-536).

We also visualized AB similarity in the anterior hippocampus and LOC in a similar fashion (Supplementary Figure 9).

We have also edited the Methods sections with updated details of these analyses (page 52, lines 1392-1397). We think that including these results considerably strengthen our claims and we are pleased to have them included.

3) The priming task occurred before the post-learning exposure phase and could have impacted the representations. More consideration of this in the paper would be useful. Most critically, since the priming task involves seeing the related C items back-to-back, it would be important to consider whether this experience could have conceivably impacted the neural integration indices. I believe it never would have been the case that unrelated C items were presented sequentially during the priming task, i.e., that related C items always appeared together in this task. I think again the specificity of the remote condition is key and perhaps the authors can leverage this to support their interpretation. Can the authors consider this possibility in the Discussion?

It's true that only C items from the same sequence were presented back-to-back during the priming task, and that this presentation may interfere with observations from the post-learning exposure scan that followed it. We agree that it is worth considering this caveat and have added language in the Discussion (page 40, lines 1071-1086). When designing the study, we reasoned that it was more important for the behavioral priming task to come before the exposure scans, as all items were shown only once in that task, whereas they were shown 4-5 times in a random order in the post-learning exposure phase. Because of this difference in presentation times, and because behavioral priming findings tend to be very sensitive, we concluded that it was more important to protect the priming task from the exposure scan instead of the reverse.

We reasoned, however, that the additional presentation of the C items in the recognition priming task would not substantially override the sequence learning, as C items were each presented 16 times in their sequence (ABC1 and ABC2 16 times each). Furthermore, as this reviewer suggests, the order of C items during recognition was the same for recent and remote conditions, so the fact that we find a selective change in neural representation for the remote condition and don’t also see that change for the recent condition is additional assurance that the recognition priming order did not substantially impact the representations.

4) For the priming task, based on the Figure 2A caption it seems as though every sequence contributes to both the control and primed conditions, but (I believe) this means that the control transition always happens first (and they are always back-to-back). Is this a concern? If RTs are changing over time (getting faster), it would be helpful to know whether the priming effects hold after controlling for trial numbers. I do not think this is a big issue because if it were, you would not expect to see the specificity of the remotely learned information. However, it would be helpful to know given the order of these conditions has to be fixed in their design.

This is a correct understanding of the trial orders in the recognition priming task. We chose to involve the baseline items in the control condition to boost power – this way, priming of each sequence could be tested, while only presenting each item once in this task, as repetition in the recognition phase would have further facilitated response times and potentially masked any priming effects. We agree that accounting for trial order would be useful here, so we ran a mixed-effects linear model to examine responses times both as a function of trial number and of priming condition (primed/control). While there is indeed a large effect of trial number such that participants got faster over time, the priming effect originally observed in the remote condition still holds at the same time. We now report this analysis in the Results section (page 14, lines 337-349 for Expt 1 and pages 14-15, lines 360-362 for Expt 2).

5) The authors should be cautious about the general conclusion that memories with overlapping temporal regularities become neurally integrated - given their findings in MPFC are more consistent with overall differentiation (though as noted above, I think we need more data on this to know for sure what is going on).

We realize this conclusion was overly simplistic and, in several places, have revised the general conclusions to be more specific about the nuanced similarity findings.

6) It would be worth stating a few more details and perhaps providing additional logic or justification in the main text about the pre- and post-exposure phases were set up and why. How many times each object was presented pre and post, and how the sequencing was determined (were any constraints put in place e.g., such that C1 and C2 did not appear close in time?). What was the cover task (I think this is important to the interpretation & so belongs in the main paper)? Were there considerations involving the fact that this is a different sequence of the same objects the participants would later be learning - e.g., interference, etc.?

These details can be found in the Methods section (pages 50-51, lines 1337-1353) and we’ve added a new summary of that section in the Results (page 17, lines 424- 425 and 432-435). In brief, a visual hash tag appeared on a small subset of images and participants pressed a button when this occurred, and C1 and C2 objects were presented in separate scans (as were A and B objects) to minimize inflated neural similarity due to temporal proximity.

Reviewer #2 (Public Review):

The manuscript by Tompary & Davachi presents results from two experiments, one behavior only and one fMRI plus behavior. They examine the important question of how to separate object memories (C1 and C2) that are never experienced together in time and become linked by shared predictive cues in a sequence (A followed by B followed by one of the C items). The authors developed an implicit priming task that provides a novel behavioral metric for such integration. They find significant C1-C2 priming for sequences that were learned 24h prior to the test, but not for recently learned sequences, suggesting that associative links between the two originally separate memories emerge over an extended period of consolidation. The fMRI study relates this behavioral integration effect to two neural metrics: pattern similarity changes in the medial prefrontal cortex (mPFC) as a measure of neural integration, and changes in hippocampal-LOC connectivity as a measure of post-learning consolidation. While fMRI patterns in mPFC overall show differentiation rather than integration (i.e., C1-C2 representational distances become larger), the authors find a robust correlation such that increasing pattern similarity in mPFC relates to stronger integration in the priming test, and this relationship is again specific to remote memories. Moreover, connectivity between the posterior hippocampus and LOC during post-learning rest is positively related to the behavioral integration effect as well as the mPFC neural similarity index, again specifically for remote memories. Overall, this is a coherent set of findings with interesting theoretical implications for consolidation theories, which will be of broad interest to the memory, learning, and predictive coding communities.

Strengths:

1) The implicit associative priming task designed for this study provides a promising new tool for assessing the formation of mnemonic links that influence behavior without explicit retrieval demands. The authors find an interesting dissociation between this implicit measure of memory integration and more commonly used explicit inference measures: a priming effect on the implicit task only evolved after a 24h consolidation period, while the ability to explicitly link the two critical object memories is present immediately after learning. While speculative at this point, these two measures thus appear to tap into neocortical and hippocampal learning processes, respectively, and this potential dissociation will be of interest to future studies investigating time-dependent integration processes in memory.

2) The experimental task is well designed for isolating pre- vs post-learning changes in neural similarity and connectivity, including important controls of baseline neural similarity and connectivity.

3) The main claim of a consolidation-dependent effect is supported by a coherent set of findings that relate behavioral integration to neural changes. The specificity of the effects on remote memories makes the results particularly interesting and compelling.

4) The authors are transparent about unexpected results, for example, the finding that overall similarity in mPFC is consistent with a differentiation rather than an integration model.

Thank you for the positive comments!

Weaknesses:

1) The sequence learning and recognition priming tasks are cleverly designed to isolate the effects of interest while controlling for potential order effects. However, due to the complex nature of the task, it is difficult for the reader to infer all the transition probabilities between item types and how they may influence the behavioral priming results. For example, baseline items (BL) are interspersed between repeated sequences during learning, and thus presumably can only occur before an A item or after a C item. This seems to create non-random predictive relationships such that C is often followed by BL, and BL by A items. If this relationship is reversed during the recognition priming task, where the sequence is always BL-C1-C2, this violation of expectations might slow down reaction times and deflate the baseline measure. It would be helpful if the manuscript explicitly reported transition probabilities for each relevant item type in the priming task relative to the sequence learning task and discussed how a match vs mismatch may influence the observed priming effects.

We have added a table of transition probabilities across the learning, recognition priming, and exposure scans (now Table 1, page 48). We have also included some additional description of the change in transition probabilities across different tasks in the Methods section. Specifically, if participants are indeed learning item types and rules about their order, then both the control and the primed conditions would violate that order. Since C1 and C2 items never appeared together, viewing C1 would give rise to an expectation of seeing a BL item, which would also be violated. This suggests that our priming effects are driven by sequence-specific relationships rather than learning of the probabilities of different item types. We’ve added this consideration to the Methods section (page 45, lines 1212-1221).

Another critical point to consider (and that the transition probabilities do not reflect) is that during learning, while C is followed either by A or BL, they are followed by different A or BL items. In contrast, a given A is always followed by the same B object, which is always followed by one of two C objects. While the order of item types is semi-predictable, the order of objects (specific items) themselves are not. This can be seen in the response times during learning, such that response times for A and BL items are always slower than for B and C items. We have explained this nuance in the figure text for Table 1.

2) The choice of what regions of interest to include in the different sets of analyses could be better motivated. For example, even though briefly discussed in the intro, it remains unclear why the posterior but not the anterior hippocampus is of interest for the connectivity analyses, and why the main target is LOC, not mPFC, given past results including from this group (Tompary & Davachi, 2017). Moreover, for readers not familiar with this literature, it would help if references were provided to suggest that a predictable > unpredictable contrast is well suited for functionally defining mPFC, as done in the present study.

We have clarified our reasoning for each of these choices throughout the manuscript and believe that our logic is now much more transparent. For an expanded reasoning of why we were motivated to look at posterior and not anterior hippocampus, see pages 6-7, lines 135-159, and our response to R2. In brief, past research focusing on post-encoding connectivity with the hippocampus suggests that posterior aspect is more likely to couple with category-selective cortex after learning neutral, non-rewarded objects much like the stimuli used in the present study.

We also clarify our reasoning for LOC over mPFC. While theoretically, mPFC is thought to be a candidate region for coupling with the hippocampus during consolidation, the bulk of empirical work to date has revealed post-encoding connectivity between the hippocampus and category-selective cortex in the ventral and occipital lobes (page 6, lines 123-134).

As for the use of the predictable > unpredictable contrast for functionally defining cortical regions, we reasoned that cortical regions that were sensitive to the temporal regularities generated by the sequences may be further involved in their offline consolidation and long-term storage (Danker & Anderson, 2010; Davachi & Danker, 2013; McClelland et al., 1995). We have added this justification to the Methods section (page 18, lines 454-460).

3) Relatedly, multiple comparison corrections should be applied in the fMRI integration and connectivity analyses whenever the same contrast is performed on multiple regions in an exploratory manner.

We now correct for multiple comparisons using Bonferroni correction, and this correction depends on the number of regions in which each analysis is conducted. Please see page 55, lines 1483-1490, in the Methods section for details of each analysis.

Reviewer #3 (Public Review):

The authors of this manuscript sought to illuminate a link between a behavioral measure of integration and neural markers of cortical integration associated with systems consolidation (post-encoding connectivity, change in representational neural overlap). To that aim, participants incidentally encoded sequences of objects in the fMRI scanner. Unbeknownst to participants, the first two objects of the presented ABC triplet sequences overlapped for a given pair of sequences. This allowed the authors to probe the integration of unique C objects that were never directly presented in the same sequence, but which shared the same preceding A and B objects. They encoded one set of objects on Day 1 (remote condition), another set of objects 24 hours later (recent condition) and tested implicit and explicit memory for the learned sequences on Day 2. They additionally collected baseline and post-encoding resting-state scans. As their measure of behavioral integration, the authors examined reaction time during an Old/New judgement task for C objects depending on if they were preceded by a C object from an overlapping sequence (primed condition) versus a baseline object. They found faster reaction times for the primed objects compared to the control condition for remote but not recently learned objects, suggesting that the C objects from overlapping sequences became integrated over time. They then examined pattern similarity in a priori ROIs as a measure of neural integration and found that participants showing evidence of integration of C objects from overlapping sequences in the medial prefrontal cortex for remotely learned objects also showed a stronger implicit priming effect between those C objects over time. When they examined the change in connectivity between their ROIs after encoding, they also found that connectivity between the posterior hippocampus and lateral occipital cortex correlated with larger priming effects for remotely learned objects, and that lateral occipital connectivity with the medial prefrontal cortex was related to neural integration of remote objects from overlapping sequences.

The authors aim to provide evidence of a relationship between behavioral and neural measures of integration with consolidation is interesting, important, and difficult to achieve given the longitudinal nature of studies required to answer this question. Strengths of this study include a creative behavioral task, and solid modelling approaches for fMRI data with careful control for several known confounds such as bold activation on pattern analysis results, motion, and physiological noise. The authors replicate their behavioral observations across two separate experiments, one of which included a large sample size, and found similar results that speak to the reliability of the observed behavioral phenomenon. In addition, they document several correlations between neural measures and task performance, lending functional significance to their neural findings.

Thank you for this positive assessment of our study!

However, this study is not without notable weaknesses that limit the strength of the manuscript. The authors report a behavioral priming effect suggestive of integration of remote but not recent memories, leading to the interpretation that the priming effect emerges with consolidation. However, they did not observe a reliable interaction between the priming condition and learning session (recent/remote) on reaction times, meaning that the priming effect for remote memories was not reliably greater than that observed for recent. In addition, the emergence of a priming effect for remote memories does not appear to be due to faster reaction times for primed targets over time (the condition of interest), but rather, slower reaction times for control items in the remote condition compared to recent. These issues limit the strength of the claim that the priming effect observed is due to C items of interest being integrated in a consolidation-dependent manner.

We acknowledge that the lack of a day by condition interaction in the behavioral priming effect should discussed and now discuss this data in a more nuanced manner. While it’s true that the priming effect emerges due to a slowing of the control items over time, this slowing is consistent with classic time-dependent effects demonstrating slower response times for more delayed memories. The fact that the response times in the primed condition does not show this slowing can be interpreted as a protection against this slowing that would otherwise occur. Please see page 29, lines 758-766, for this added discussion.

Similarly, the interactions between neural variables of interest and learning session needed to strongly show a significant consolidation-related effect in the brain were sometimes tenuous. There was no reliable difference in neural representational pattern analysis fit to a model of neural integration between the short and long delays in the medial prefrontal cortex or lateral occipital cortex, nor was the posterior hippocampus-lateral occipital cortex post-encoding connectivity correlation with subsequent priming significantly different for recent and remote memories. While the relationship between integration model fit in the medial prefrontal cortex and subsequent priming (which was significantly different from that occurring for recent memories) was one of the stronger findings of the paper in favor of a consolidation-related effect on behavior, is it possible that lack of a behavioral priming effect for recent memories due to possible issues with the control condition could mask a correlation between neural and behavioral integration in the recent memory condition?

While we acknowledge that lack of a statistically reliable interaction between neural measures and behavioral priming in many cases, we are heartened by the reliable difference in the relationship between mPFC similarity and priming over time, which was our main planned prediction. In addition to adding caveats in the discussion about the neural measures and behavioral findings in the recent condition (see our response to R1.1 and R1.4 for more details), we have added language throughout the manuscript noting the need to interpret these data with caution.

These limitations are especially notable when one considers that priming does not classically require a period of prolonged consolidation to occur, and prominent models of systems consolidation rather pertain to explicit memory. While the authors have provided evidence that neural integration in the medial prefrontal cortex, as well as post-encoding coupling between the lateral occipital cortex and posterior hippocampus, are related to faster reaction times for primed objects of overlapping sequences compared to their control condition, more work is needed to verify that the observed findings indeed reflect consolidation dependent integration as proposed.

We agree that more work is needed to provide converging evidence for these novel findings. However, we wish to counter the notion that systems consolidation models are relevant only for explicit memories. Although models of systems consolidation often mention transformations from episodic to semantic memory, the critical mechanisms that define the models involve changes in the neural ensembles of a memory that is initially laid down in the hippocampus and is taught to cortex over time. This transformation of neural traces is not specific to explicit/declarative forms of memory. For example, implicit statistical learning initially depends on intact hippocampal function (Schapiro et al., 2014) and improves over consolidation (Durrant et al., 2011, 2013; Kóbor et al., 2017).

Second, while there are many classical findings of priming during or immediately after learning, there are several instances of priming used to measure consolidation-related changes to newly learned information. For instance, priming has been used as a measure of lexical integration, demonstrating that new word learning benefits from a night of sleep (Wang et al., 2017; Gaskell et al., 2019) or a 1-week delay (Tamminen & Gaskell, 2013). The issue is not whether priming can occur immediately, it is whether priming increases with a delay.

Finally, it is helpful to think about models of memory systems that divide memory representations not by their explicit/implicit nature, but along other important dimensions such as their neural bases, their flexibility vs rigidity, and their capacity for rapid vs slow learning (Henke, 2010). Considering this evidence, we suggest that systems consolidation models are most useful when considering how transformations in the underlying neural memory representation affects its behavioral expression, rather than focusing on the extent that the memory representation is explicit or implicit.

With all this said, we have added text to the discussion reminding the reader that there was no statistically significant difference in priming as a function of the delay (page 29, lines 764 - 766). However, we are encouraged by the fact that the relationship between priming and mPFC neural similarity was significantly stronger for remotely learned objects relative to recently learned ones, as this is directly in line with systems consolidation theories.

References

Abolghasem, Z., Teng, T. H.-T., Nexha, E., Zhu, C., Jean, C. S., Castrillon, M., Che, E., Di Nallo, E. V., & Schlichting, M. L. (2023). Learning strategy differentially impacts memory connections in children and adults. Developmental Science, 26(4), e13371. https://doi.org/10.1111/desc.13371

Dobbins, I. G., Schnyer, D. M., Verfaellie, M., & Schacter, D. L. (2004). Cortical activity reductions during repetition priming can result from rapid response learning. Nature, 428(6980), 316–319. https://doi.org/10.1038/nature02400

Durrant, S. J., Cairney, S. A., & Lewis, P. A. (2013). Overnight consolidation aids the transfer of statistical knowledge from the medial temporal lobe to the striatum. Cerebral Cortex, 23(10), 2467–2478. https://doi.org/10.1093/cercor/bhs244

Durrant, S. J., Taylor, C., Cairney, S., & Lewis, P. A. (2011). Sleep-dependent consolidation of statistical learning. Neuropsychologia, 49(5), 1322–1331. https://doi.org/10.1016/j.neuropsychologia.2011.02.015

Gaskell, M. G., Cairney, S. A., & Rodd, J. M. (2019). Contextual priming of word meanings is stabilized over sleep. Cognition, 182, 109–126. https://doi.org/10.1016/j.cognition.2018.09.007

Henke, K. (2010). A model for memory systems based on processing modes rather than consciousness. Nature Reviews Neuroscience, 11(7), 523–532. https://doi.org/10.1038/nrn2850

Kóbor, A., Janacsek, K., Takács, Á., & Nemeth, D. (2017). Statistical learning leads to persistent memory: Evidence for one-year consolidation. Scientific Reports, 7(1), 760. https://doi.org/10.1038/s41598-017-00807-3

Kuhl, B. A., & Chun, M. M. (2014). Successful remembering elicits event-specific activity patterns in lateral parietal cortex. The Journal of Neuroscience, 34(23), 8051–8060. https://doi.org/10.1523/JNEUROSCI.4328-13.2014

Richter, F. R., Chanales, A. J. H., & Kuhl, B. A. (2016). Predicting the integration of overlapping memories by decoding mnemonic processing states during learning. NeuroImage, 124, Part A, 323–335. https://doi.org/10.1016/j.neuroimage.2015.08.051

Schapiro, A. C., Gregory, E., Landau, B., McCloskey, M., & Turk-Browne, N. B. (2014). The necessity of the medial-temporal lobe for statistical learning. Journal of Cognitive Neuroscience, 1–12. https://doi.org/10.1162/jocn_a_00578

Schlichting, M. L., & Preston, A. R. (2014). Memory reactivation during rest supports upcoming learning of related content. Proceedings of the National Academy of Sciences, 111(44), 15845–15850. https://doi.org/10.1073/pnas.1404396111

Smith, J. F., Alexander, G. E., Chen, K., Husain, F. T., Kim, J., Pajor, N., & Horwitz, B. (2010). Imaging systems level consolidation of novel associate memories: A longitudinal neuroimaging study. NeuroImage, 50(2), 826–836. https://doi.org/10.1016/j.neuroimage.2009.11.053

Takashima, A., Nieuwenhuis, I. L. C., Jensen, O., Talamini, L. M., Rijpkema, M., & Fernández, G. (2009). Shift from hippocampal to neocortical centered retrieval network with consolidation. The Journal of Neuroscience, 29(32), 10087–10093. https://doi.org/10.1523/JNEUROSCI.0799-09.2009

Tamminen, J., & Gaskell, M. G. (2013). Novel word integration in the mental lexicon: Evidence from unmasked and masked semantic priming. The Quarterly Journal of Experimental Psychology, 66(5), 1001–1025. https://doi.org/10.1080/17470218.2012.724694

van Kesteren, M. T. R. van, Fernández, G., Norris, D. G., & Hermans, E. J. (2010). Persistent schema-dependent hippocampal-neocortical connectivity during memory encoding and postencoding rest in humans. Proceedings of the National Academy of Sciences, 107(16), 7550–7555. https://doi.org/10.1073/pnas.0914892107

Wang, H.-C., Savage, G., Gaskell, M. G., Paulin, T., Robidoux, S., & Castles, A. (2017). Bedding down new words: Sleep promotes the emergence of lexical competition in visual word recognition. Psychonomic Bulletin & Review, 24(4), 1186–1193. https://doi.org/10.3758/s13423-016-1182-7

Author Response

Reviewer #1 (Public Review):

This study used a multi-day learning paradigm combined with fMRI to reveal neural changes reflecting the learning of new (arbitrary) shape-sound associations. In the scanner, the shapes and sounds are presented separately and together, both before and after learning. When they are presented together, they can be either consistent or inconsistent with the learned associations. The analyses focus on auditory and visual cortices, as well as the object-selective cortex (LOC) and anterior temporal lobe regions (temporal pole (TP) and perirhinal cortex (PRC)). Results revealed several learning-induced changes, particularly in the anterior temporal lobe regions. First, the LOC and PRC showed a reduced bias to shapes vs sounds (presented separately) after learning. Second, the TP responded more strongly to incongruent than congruent shape-sound pairs after learning. Third, the similarity of TP activity patterns to sounds and shapes (presented separately) was increased for non-matching shape-sound comparisons after learning. Fourth, when comparing the pattern similarity of individual features to combined shape-sound stimuli, the PRC showed a reduced bias towards visual features after learning. Finally, comparing patterns to combined shape-sound stimuli before and after learning revealed a reduced (and negative) similarity for incongruent combinations in PRC. These results are all interpreted as evidence for an explicit integrative code of newly learned multimodal objects, in which the whole is different from the sum of the parts.

The study has many strengths. It addresses a fundamental question that is of broad interest, the learning paradigm is well-designed and controlled, and the stimuli are real 3D stimuli that participants interact with. The manuscript is well written and the figures are very informative, clearly illustrating the analyses performed.

There are also some weaknesses. The sample size (N=17) is small for detecting the subtle effects of learning. Most of the statistical analyses are not corrected for multiple comparisons (ROIs), and the specificity of the key results to specific regions is also not tested. Furthermore, the evidence for an integrative representation is rather indirect, and alternative interpretations for these results are not considered.

We thank the reviewer for their careful reading and the positive comments on our manuscript. As suggested, we have conducted additional analyses of theoretically-motivated ROIs and have found that temporal pole and perirhinal cortex are the only regions to show the key experience-dependent transformations. We are much more cautious with respect to multiple comparisons, and have removed a series of post hoc across-ROI comparisons that were irrelevant to the key questions of the present manuscript. The revised manuscript now includes much more discussion about alternative interpretations as suggested by the reviewer (and also by the other reviewers).

Additionally, we looked into scanning more participants, but our scanner has since had a full upgrade and the sequence used in the current study is no longer supported by our scanner. However, we note that while most analyses contain 17 participants, we employed a within-subject learning design that is not typically used in fMRI experiments and increases our power to detect an effect. This is supported by the robust effect size of the behavioural data, whereby 17 out of 18 participants revealed a learning effect (Cohen’s D = 1.28) and which was replicated in a follow-up experiment with a larger sample size.

We address the other reviewer comments point-by-point in the below.

Reviewer #2 (Public Review):

Li et al. used a four-day fMRI design to investigate how unimodal feature information is combined, integrated, or abstracted to form a multimodal object representation. The experimental question is of great interest and understanding how the human brain combines featural information to form complex representations is relevant for a wide range of researchers in neuroscience, cognitive science, and AI. While most fMRI research on object representations is limited to visual information, the authors examined how visual and auditory information is integrated to form a multimodal object representation. The experimental design is elegant and clever. Three visual shapes and three auditory sounds were used as the unimodal features; the visual shapes were used to create 3D-printed objects. On Day 1, the participants interacted with the 3D objects to learn the visual features, but the objects were not paired with the auditory features, which were played separately. On Day 2, participants were scanned with fMRI while they were exposed to the unimodal visual and auditory features as well as pairs of visual-auditory cues. On Day 3, participants again interacted with the 3D objects but now each was paired with one of the three sounds that played from an internal speaker. On Day 4, participants completed the same fMRI scanning runs they completed on Day 2, except now some visual-auditory feature pairs corresponded with Congruent (learned) objects, and some with Incongruent (unlearned) objects. Using the same fMRI design on Days 2 and 4 enables a well-controlled comparison between feature- and object-evoked neural representations before and after learning. The notable results corresponded to findings in the perirhinal cortex and temporal pole. The authors report (1) that a visual bias on Day 2 for unimodal features in the perirhinal cortex was attenuated after learning on Day 4, (2) a decreased univariate response to congruent vs. incongruent visual-auditory objects in the temporal pole on Day 4, (3) decreased pattern similarity between congruent vs. incongruent pairs of visual and auditory unimodal features in the temporal pole on Day 4, (4) in the perirhinal cortex, visual unimodal features on Day 2 do not correlate with their respective visual-auditory objects on Day 4, and (5) in the perirhinal cortex, multimodal object representations across Days 2 and 4 are uncorrelated for congruent objects and anticorrelated for incongruent. The authors claim that each of these results supports the theory that multimodal objects are represented in an "explicit integrative" code separate from feature representations. While these data are valuable and the results are interesting, the authors' claims are not well supported by their findings.

We thank the reviewer for the careful reading of our manuscript and positive comments. Overall, we now stay closer to the data when describing the results and provide our interpretation of these results in the discussion section while remaining open to alternative interpretations (as also suggested by Reviewer 1).

(1) In the introduction, the authors contrast two theories: (a) multimodal objects are represented in the co-activation of unimodal features, and (b) multimodal objects are represented in an explicit integrative code such that the whole is different than the sum of its parts. However, the distinction between these two theories is not straightforward. An explanation of what is precisely meant by "explicit" and "integrative" would clarify the authors' theoretical stance. Perhaps we can assume that an "explicit" representation is a new representation that is created to represent a multimodal object. What is meant by "integrative" is more ambiguous-unimodal features could be integrated within a representation in a manner that preserves the decodability of the unimodal features, or alternatively the multimodal representation could be completely abstracted away from the constituent features such that the features are no longer decodable. Even if the object representation is "explicit" and distinct from the unimodal feature representations, it can in theory still contain featural information, though perhaps warped or transformed. The authors do not clearly commit to a degree of featural abstraction in their theory of "explicit integrative" multimodal object representations which makes it difficult to assess the validity of their claims.

Due to its ambiguity, we removed the term “explicit” and now make it clear that our central question was whether crossmodal object representations require only unimodal feature-level representations (e.g., frogs are created from only the combination of shape and sound) or whether crossmodal object representations also rely on an integrative code distinct from the unimodal features (e.g., there is something more to “frog” than its original shape and sound). We now clarify this in the revised manuscript.

“One theoretical view from the cognitive sciences suggests that crossmodal objects are built from component unimodal features represented across distributed sensory regions.8 Under this view, when a child thinks about “frog”, the visual cortex represents the appearance of the shape of the frog whereas the auditory cortex represents the croaking sound. Alternatively, other theoretical views predict that multisensory objects are not only built from their component unimodal sensory features, but that there is also a crossmodal integrative code that is different from the sum of these parts.9,10,11,12,13 These latter views propose that anterior temporal lobe structures can act as a polymodal “hub” that combines separate features into integrated wholes.9,11,14,15” – pg. 4

For this reason, we designed our paradigm to equate the unimodal representations, such that neural differences between the congruent and incongruent conditions provide evidence for a crossmodal integrative code different from the unimodal features (because the unimodal features are equated by default in the design).

“Critically, our four-day learning task allowed us to isolate any neural activity associated with integrative coding in anterior temporal lobe structures that emerges with experience and differs from the neural patterns recorded at baseline. The learned and non-learned crossmodal objects were constructed from the same set of three validated shape and sound features, ensuring that factors such as familiarity with the unimodal features, subjective similarity, and feature identity were tightly controlled (Figure 2). If the mind represented crossmodal objects entirely as the reactivation of unimodal shapes and sounds (i.e., objects are constructed from their parts), then there should be no difference between the learned and non-learned objects (because they were created from the same three shapes and sounds). By contrast, if the mind represented crossmodal objects as something over and above their component features (i.e., representations for crossmodal objects rely on integrative coding that is different from the sum of their parts), then there should be behavioral and neural differences between learned and non-learned crossmodal objects (because the only difference across the objects is the learned relationship between the parts). Furthermore, this design allowed us to determine the relationship between the object representation acquired after crossmodal learning and the unimodal feature representations acquired before crossmodal learning. That is, we could examine whether learning led to abstraction of the object representations such that it no longer resembled the unimodal feature representations.” – pg. 5

Furthermore, we agree with the reviewer that our definition and methodological design does not directly capture the structure of the integrative code. With experience, the unimodal feature representations may be completely abstracted away, warped, or changed in a nonlinear transformation. We suggest that crossmodal learning forms an integrative code that is different from the original unimodal representations in the anterior temporal lobes, however, we agree that future work is needed to more directly capture the structure of the integrative code that emerges with experience.

“In our task, participants had to differentiate congruent and incongruent objects constructed from the same three shape and sound features (Figure 2). An efficient way to solve this task would be to form distinct object-level outputs from the overlapping unimodal feature-level inputs such that congruent objects are made to be orthogonal from the representations before learning (i.e., measured as pattern similarity equal to 0 in the perirhinal cortex; Figure 5b, 6, Supplemental Figure S5), whereas non-learned incongruent objects could be made to be dissimilar from the representations before learning (i.e., anticorrelation, measured as patten similarity less than 0 in the perirhinal cortex; Figure 6). Because our paradigm could decouple neural responses to the learned object representations (on Day 4) from the original component unimodal features at baseline (on Day 2), these results could be taken as evidence of pattern separation in the human perirhinal cortex.11,12 However, our pattern of results could also be explained by other types of crossmodal integrative coding. For example, incongruent object representations may be less stable than congruent object representations, such that incongruent objects representation are warped to a greater extent than congruent objects (Figure 6).” – pg. 18

“As one solution to the crossmodal binding problem, we suggest that the temporal pole and perirhinal cortex form unique crossmodal object representations that are different from the distributed features in sensory cortex (Figure 4, 5, 6, Supplemental Figure S5). However, the nature by which the integrative code is structured and formed in the temporal pole and perirhinal cortex following crossmodal experience – such as through transformations, warping, or other factors – is an open question and an important area for future investigation.” – pg. 18

(2) After participants learned the multimodal objects, the authors report a decreased univariate response to congruent visual-auditory objects relative to incongruent objects in the temporal pole. This is claimed to support the existence of an explicit, integrative code for multimodal objects. Given the number of alternative explanations for this finding, this claim seems unwarranted. A simpler interpretation of these results is that the temporal pole is responding to the novelty of the incongruent visual-auditory objects. If there is in fact an explicit, integrative multimodal object representation in the temporal pole, it is unclear why this would manifest in a decreased univariate response.

We thank the reviewer for identifying this issue. Our behavioural design controls unimodal feature-level novelty but allows object-level novelty to differ. Thus, neural differences between the congruent and incongruent conditions reflects sensitivity to the object-level differences between the combination of shape and sound. However, we agree that there are multiple interpretations regarding the nature of how the integrative code is structured in the temporal pole and perirhinal cortex. We have removed the interpretation highlighted by the reviewer from the results. Instead, we now provide our preferred interpretation in the discussion, while acknowledging the other possibilities that the reviewer mentions.

As one possibility, these results in temporal pole may reflect “conceptual combination”. “hummingbird” – a congruent pairing – may require less neural resources than an incongruent pairing such as “bark-frog”.

“Furthermore, these distinct anterior temporal lobe structures may be involved with integrative coding in different ways. For example, the crossmodal object representations measured after learning were found to be related to the component unimodal feature representations measured before learning in the temporal pole but not the perirhinal cortex (Figure 5, 6, Supplemental Figure S5). Moreover, pattern similarity for congruent shape-sound pairs were lower than the pattern similarity for incongruent shape-sound pairs after crossmodal learning in the temporal pole but not the perirhinal cortex (Figure 4b, Supplemental Figure S3a). As one interpretation of this pattern of results, the temporal pole may represent new crossmodal objects by combining previously learned knowledge. 8,9,10,11,13,14,15,33 Specifically, research into conceptual combination has linked the anterior temporal lobes to compound object concepts such as “hummingbird”.34,35,36 For example, participants during our task may have represented the sound-based “humming” concept and visually-based “bird” concept on Day 1, forming the crossmodal “hummingbird” concept on Day 3; Figure 1, 2, which may recruit less activity in temporal pole than an incongruent pairing such as “barking-frog”. For these reasons, the temporal pole may form a crossmodal object code based on pre-existing knowledge, resulting in reduced neural activity (Figure 3d) and pattern similarity towards features associated with learned objects (Figure 4b).”– pg. 18

(3) The authors ran a neural pattern similarity analysis on the unimodal features before and after multimodal object learning. They found that the similarity between visual and auditory features that composed congruent objects decreased in the temporal pole after multimodal object learning. This was interpreted to reflect an explicit integrative code for multimodal objects, though it is not clear why. First, behavioral data show that participants reported increased similarity between the visual and auditory unimodal features within congruent objects after learning, the opposite of what was found in the temporal pole. Second, it is unclear why an analysis of the unimodal features would be interpreted to reflect the nature of the multimodal object representations. Since the same features corresponded with both congruent and incongruent objects, the nature of the feature representations cannot be interpreted to reflect the nature of the object representations per se. Third, using unimodal feature representations to make claims about object representations seems to contradict the theoretical claim that explicit, integrative object representations are distinct from unimodal features. If the learned multimodal object representation exists separately from the unimodal feature representations, there is no reason why the unimodal features themselves would be influenced by the formation of the object representation. Instead, these results seem to more strongly support the theory that multimodal object learning results in a transformation or warping of feature space.

We apologize for the lack of clarity. We have now overhauled this aspect of our manuscript in an attempt to better highlight key aspects of our experimental design. In particular, because the unimodal features composing the congruent and incongruent objects were equated, neural differences between these conditions would provide evidence for an experience-dependent crossmodal integrative code that is different from its component unimodal features.

Related to the second and third points, we were looking at the extent to which the original unimodal representations change with crossmodal learning. Before crossmodal learning, we found that the perirhinal cortex tracked the similarity between the individual visual shape features and the crossmodal objects that were composed of those visual shapes – however, there was no evidence that perirhinal cortex was tracking the unimodal sound features on those crossmodal objects. After crossmodal learning, we see that this visual shape bias in perirhinal cortex was no longer present – that is, the representation in perirhinal cortex started to look less like the visual features that comprise the objects. Thus, crossmodal learning transformed the perirhinal representations so that they were no longer predominantly grounded in a single visual modality, which may be a mechanism by which object concepts gain their abstraction. We have now tried to be clearer about this interpretation throughout the paper.

Notably, we suggest that experience may change both the crossmodal object representations, as well as the unimodal feature representations. For example, we have previously shown that unimodal visual features are influenced by experience in parallel with the representation of the conjunction (e.g., Liang et al., 2020; Cerebral Cortex). Nevertheless, we remain open to the myriad possible structures of the integrative code that might emerge with experience.

We now clarify these points throughout the manuscript. For example:

“We then examined whether the original representations would change after participants learned how the features were paired together to make specific crossmodal objects, conducting the same analysis described above after crossmodal learning had taken place (Figure 5b). With this analysis, we sought to measure the relationship between the representation for the learned crossmodal object and the original baseline representation for the unimodal features. More specifically, the voxel-wise activity for unimodal feature runs before crossmodal learning was correlated to the voxel-wise activity for crossmodal object runs after crossmodal learning (Figure 5b). Another linear mixed model which included modality as a fixed factor within each ROI revealed that the perirhinal cortex was no longer biased towards visual shape after crossmodal learning (F1,32 = 0.12, p = 0.73), whereas the temporal pole, LOC, V1, and A1 remained biased towards either visual shape or sound (F1,30-32 between 16.20 and 73.42, all p < 0.001, η2 between 0.35 and 0.70).” – pg. 14

“To investigate this effect in perirhinal cortex more specifically, we conducted a linear mixed model to directly compare the change in the visual bias of perirhinal representations from before crossmodal learning to after crossmodal learning (green regions in Figure 5a vs. 5b). Specifically, the linear mixed model included learning day (before vs. after crossmodal learning) and modality (visual feature match to crossmodal object vs. sound feature match to crossmodal object). Results revealed a significant interaction between learning day and modality in the perirhinal cortex (F1,775 = 5.56, p = 0.019, η2 = 0.071), meaning that the baseline visual shape bias observed in perirhinal cortex (green region of Figure 5a) was significantly attenuated with experience (green region of Figure 5b). After crossmodal learning, a given shape no longer invoked significant pattern similarity between objects that had the same shape but differed in terms of what they sounded like. Taken together, these results suggest that prior to learning the crossmodal objects, the perirhinal cortex had a default bias toward representing the visual shape information and was not representing sound information of the crossmodal objects. After crossmodal learning, however, the visual shape bias in perirhinal cortex was no longer present. That is, with crossmodal learning, the representations within perirhinal cortex started to look less like the visual features that comprised the crossmodal objects, providing evidence that the perirhinal representations were no longer predominantly grounded in the visual modality.” – pg. 13

“Importantly, the initial visual shape bias observed in the perirhinal cortex was attenuated by experience (Figure 5, Supplemental Figure S5), suggesting that the perirhinal representations had become abstracted and were no longer predominantly grounded in a single modality after crossmodal learning. One possibility may be that the perirhinal cortex is by default visually driven as an extension to the ventral visual stream,10,11,12 but can act as a polymodal “hub” region for additional crossmodal input following learning.” – pg. 19

(4) The most compelling evidence the authors provide for their theoretical claims is the finding that, in the perirhinal cortex, the unimodal feature representations on Day 2 do not correlate with the multimodal objects they comprise on Day 4. This suggests that the learned multimodal object representations are not combinations of their unimodal features. If unimodal features are not decodable within the congruent object representations, this would support the authors' explicit integrative hypothesis. However, the analyses provided do not go all the way in convincing the reader of this claim. First, the analyses reported do not differentiate between congruent and incongruent objects. If this result in the perirhinal cortex reflects the formation of new multimodal object representations, it should only be true for congruent objects but not incongruent objects. Since the analyses combine congruent and incongruent objects it is not possible to know whether this was the case. Second, just because feature representations on Day 2 do not correlate with multimodal object patterns on Day 4 does not mean that the object representations on Day 4 do not contain featural information. This could be directly tested by correlating feature representations on Day 4 with congruent vs. incongruent object representations on Day 4. It could be that representations in the perirhinal cortex are not stable over time and all representations-including unimodal feature representations-shift between sessions, which could explain these results yet not entail the existence of abstracted object representations.

We thank the reviewer for this suggestion and have conducted the two additional analyses. Specifically, we split the congruent and incongruent conditions and also investigated correlations between unimodal representations on Day 4 with crossmodal object representations on Day 4. There was no significant interaction between modality and congruency in any ROI across or within learning days. One possible explanation for these findings is that both congruent and incongruent crossmodal objects are represented differently from their underlying unimodal features, and all of these representations can transform with experience.

However, the new analyses also revealed that perirhinal cortex was the only region without a modality-specific bias after crossmodal learning (e.g., Day 4 Unimodal Feature runs x Day 4 Crossmodal Object runs; now shown in Supplemental Figure S5). Overall, these results are consistent with the notion of a crossmodal integrative code in perirhinal cortex that has changed with experience and is different from the component unimodal features. Nevertheless, we explore alternative interpretations for how the crossmodal code emerges with experience in the discussion.

“To examine whether these results differed by congruency (i.e., whether any modality-specific biases differed as a function of whether the object was congruent or incongruent), we conducted exploratory linear mixed models for each of the five a priori ROIs across learning days. More specifically, we correlated: 1) the voxel-wise activity for Unimodal Feature Runs before crossmodal learning to the voxel-wise activity for Crossmodal Object Runs before crossmodal learning (Day 2 vs. Day 2), 2) the voxel-wise activity for Unimodal Feature Runs before crossmodal learning to the voxel-wise activity for Crossmodal Object Runs after crossmodal learning (Day 2 vs Day 4), and 3) the voxel-wise activity for Unimodal Feature Runs after crossmodal learning to the voxel-wise activity for Crossmodal Object Runs after crossmodal learning (Day 4 vs Day 4). For each of the three analyses described, we then conducted separate linear mixed models which included modality (visual feature match to crossmodal object vs. sound feature match to crossmodal object) and congruency (congruent vs. incongruent)….There was no significant relationship between modality and congruency in any ROI between Day 2 and Day 2 (F1,346-368 between 0.00 and 1.06, p between 0.30 and 0.99), between Day 2 and Day 4 (F1,346-368 between 0.021 and 0.91, p between 0.34 and 0.89), or between Day 4 and Day 4 (F1,346-368 between 0.01 and 3.05, p between 0.082 and 0.93). However, exploratory analyses revealed that perirhinal cortex was the only region without a modality-specific bias and where the unimodal feature runs were not significantly correlated to the crossmodal object runs after crossmodal learning (Supplemental Figure S5).” – pg. 14

“Taken together, the overall pattern of results suggests that representations of the crossmodal objects in perirhinal cortex were heavily influenced by their consistent visual features before crossmodal learning. However, the crossmodal object representations were no longer influenced by the component visual features after crossmodal learning (Figure 5, Supplemental Figure S5). Additional exploratory analyses did not find evidence of experience-dependent changes in the hippocampus or inferior parietal lobes (Supplemental Figure S4c-e).” – pg. 14

“The voxel-wise matrix for Unimodal Feature runs on Day 4 were correlated to the voxel-wise matrix for Crossmodal Object runs on Day 4 (see Figure 5 in the main text for an example). We compared the average pattern similarity (z-transformed Pearson correlation) between shape (blue) and sound (orange) features specifically after crossmodal learning. Consistent with Figure 5b, perirhinal cortex was the only region without a modality-specific bias. Furthermore, perirhinal cortex was the only region where the representations of both the visual and sound features were not significantly correlated to the crossmodal objects. By contrast, every other region maintained a modality-specific bias for either the visual or sound features. These results suggest that perirhinal cortex representations were transformed with experience, such that the initial visual shape representations (Figure 5a) were no longer grounded in a single modality after crossmodal learning. Furthermore, these results suggest that crossmodal learning formed an integrative code different from the unimodal features in perirhinal cortex, as the visual and sound features were not significantly correlated with the crossmodal objects. * p < 0.05, ** p < 0.01, *** p < 0.001. Horizontal lines within brain regions indicate a significant main effect of modality. Vertical asterisks denote pattern similarity comparisons relative to 0.” – Supplemental Figure S5

“We found that the temporal pole and perirhinal cortex – two anterior temporal lobe structures – came to represent new crossmodal object concepts with learning, such that the acquired crossmodal object representations were different from the representation of the constituent unimodal features (Figure 5, 6). Intriguingly, the perirhinal cortex was by default biased towards visual shape, but that this initial visual bias was attenuated with experience (Figure 3c, 5, Supplemental Figure S5). Within the perirhinal cortex, the acquired crossmodal object concepts (measured after crossmodal learning) became less similar to their original component unimodal features (measured at baseline before crossmodal learning); Figure 5, 6, Supplemental Figure S5. This is consistent with the idea that object representations in perirhinal cortex integrate the component sensory features into a whole that is different from the sum of the component parts, which might be a mechanism by which object concepts obtain their abstraction…. As one solution to the crossmodal binding problem, we suggest that the temporal pole and perirhinal cortex form unique crossmodal object representations that are different from the distributed features in sensory cortex (Figure 4, 5, 6, Supplemental Figure S5). However, the nature by which the integrative code is structured and formed in the temporal pole and perirhinal cortex following crossmodal experience – such as through transformations, warping, or other factors – is an open question and an important area for future investigation.” – pg. 18

In sum, the authors have collected a fantastic dataset that has the potential to answer questions about the formation of multimodal object representations in the brain. A more precise delineation of different theoretical accounts and additional analyses are needed to provide convincing support for the theory that “explicit integrative” multimodal object representations are formed during learning.

We thank the reviewer for the positive comments and helpful feedback. We hope that our changes to our wording and clarifications to our methodology now more clearly supports the central goal of our study: to find evidence of crossmodal integrative coding different from the original unimodal feature parts in anterior temporal lobe structures. We furthermore agree that future research is needed to delineate the structure of the integrative code that emerges with experience in the anterior temporal lobes.

Reviewer #3 (Public Review):

This paper uses behavior and functional brain imaging to understand how neural and cognitive representations of visual and auditory stimuli change as participants learn associations among them. Prior work suggests that areas in the anterior temporal (ATL) and perirhinal cortex play an important role in learning/representing cross-modal associations, but the hypothesis has not been directly tested by evaluating behavior and functional imaging before and after learning cross- modal associations. The results show that such learning changes both the perceived similarities amongst stimuli and the neural responses generated within ATL and perirhinal regions, providing novel support for the view that cross-modal learning leads to a representational change in these regions.

This work has several strengths. It tackles an important question for current theories of object representation in the mind and brain in a novel and quite direct fashion, by studying how these representations change with cross-modal learning. As the authors note, little work has directly assessed representational change in ATL following such learning, despite the widespread view that ATL is critical for such representation. Indeed, such direct assessment poses several methodological challenges, which the authors have met with an ingenious experimental design. The experiment allows the authors to maintain tight control over both the familiarity and the perceived similarities amongst the shapes and sounds that comprise their stimuli so that the observed changes across sessions must reflect learned cross-modal associations among these. I especially appreciated the creation of physical objects that participants can explore and the approach to learning in which shapes and sounds are initially experienced independently and later in an associated fashion. In using multi-echo MRI to resolve signals in ventral ATL, the authors have minimized a key challenge facing much work in this area (namely the poor SNR yielded by standard acquisition sequences in ventral ATL). The use of both univariate and multivariate techniques was well-motivated and helpful in testing the central questions. The manuscript is, for the most part, clearly written, and nicely connects the current work to important questions in two literatures, specifically (1) the hypothesized role of the perirhinal cortex in representing/learning complex conjunctions of features and (2) the tension between purely embodied approaches to semantic representation vs the view that ATL regions encode important amodal/crossmodal structure.

There are some places in the manuscript that would benefit from further explanation and methodological detail. I also had some questions about the results themselves and what they signify about the roles of ATL and the perirhinal cortex in object representation.

We thank the reviewer for their positive feedback and address the comments in the below point-by-point responses.

(A) I found the terms "features" and "objects" to be confusing as used throughout the manuscript, and sometimes inconsistent. I think by "features" the authors mean the shape and sound stimuli in their experiment. I think by "object" the authors usually mean the conjunction of a shape with a sound---for instance, when a shape and sound are simultaneously experienced in the scanner, or when the participant presses a button on the shape and hears the sound. The confusion comes partly because shapes are often described as being composed of features, not features in and of themselves. (The same is sometimes true of sounds). So when reading "features" I kept thinking the paper referred to the elements that went together to comprise a shape. It also comes from ambiguous use of the word object, which might refer to (a) the 3D- printed item that people play with, which is an object, or (b) a visually-presented shape (for instance, the localizer involved comparing an "object" to a "phase-scrambled" stimulus---here I assume "object" refers to an intact visual stimulus and not the joint presentation of visual and auditory items). I think the design, stimuli, and results would be easier for a naive reader to follow if the authors used the terms "unimodal representation" to refer to cases where only visual or auditory input is presented, and "cross-modal" or "conjoint" representation when both are present.

We thank the reviewer for this suggestion and agree. We have replaced the terms “features” and “objects” with “unimodal” and “crossmodal” in the title, text, and figures throughout the manuscript for consistency (i.e., “crossmodal binding problem”). To simplify the terminology, we have also removed the localizer results.

(B) There are a few places where I wasn't sure what exactly was done, and where the methods lacked sufficient detail for another scientist to replicate what was done. Specifically:

(1) The behavioral study assessing perceptual similarity between visual and auditory stimuli was unclear. The procedure, stimuli, number of trials, etc, should be explained in sufficient detail in methods to allow replication. The results of the study should also minimally be reported in the supplementary information. Without an understanding of how these studies were carried out, it was very difficult to understand the observed pattern of behavioral change. For instance, I initially thought separate behavioral blocks were carried out for visual versus auditory stimuli, each presented in isolation; however, the effects contrast congruent and incongruent stimuli, which suggests these decisions must have been made for the conjoint presentation of both modalities. I'm still not sure how this worked. Additionally, the manuscript makes a brief mention that similarity judgments were made in the context of "all stimuli," but I didn't understand what that meant. Similarity ratings are hugely sensitive to the contrast set with which items appear, so clarity on these points is pretty important. A strength of the design is the contention that shape and sound stimuli were psychophysically matched, so it is important to show the reader how this was done and what the results were.

We agree and apologize for the lack of sufficient detail in the original manuscript. We now include much more detail about the similarity rating task. The methodology and results of the behavioral rating experiments are now shown in Supplemental Figure S1. In Figure S1a, the similarity ratings are visualized on a multidimensional scaling plot. The triangular geometry for shape (blue) and sound (red) indicate that the subjective similarity was equated within each unimodal feature across individual participants. Quantitatively, there was no difference in similarity between the congruent and incongruent pairings in Figure S1b and Figure S1c prior to crossmodal learning. In addition to providing more information on these methods in the Supplemental Information, we also now provide a more detailed description of the task in the manuscript itself. For convenience, we reproduce these sections below.

“Pairwise Similarity Task. Using the same task as the stimulus validation procedure (Supplemental Figure S1a), participants provided similarity ratings for all combinations of the 3 validated shapes and 3 validated sounds (each of the six features were rated in the context of every other feature in the set, with 4 repeats of the same feature, for a total of 72 trials). More specifically, three stimuli were displayed on each trial, with one at the top and two at the bottom of the screen in the same procedure as we have used previously27. The 3D shapes were visually displayed as a photo, whereas sounds were displayed on screen in a box that could be played over headphones when clicked with the mouse. The participant made an initial judgment by selecting the more similar stimulus on the bottom relative to the stimulus on the top. Afterwards, the participant made a similarity rating between each bottom stimulus with the top stimulus from 0 being no similarity to 5 being identical. This procedure ensured that ratings were made relative to all other stimuli in the set.”– pg. 28

“Pairwise similarity task and results. In the initial stimulus validation experiment, participants provided pairwise ratings for 5 sounds and 3 shapes. The shapes were equated in their subjective similarity that had been selected from a well-characterized perceptually uniform stimulus space27 and the pairwise ratings followed the same procedure as described in ref 27. Based on this initial experiment, we then selected the 3 sounds from the that were most closely equated in their subjective similarity. (a) 3D-printed shapes were displayed as images, whereas sounds were displayed in a box that could be played when clicked by the participant. Ratings were averaged to produce a similarity matrix for each participant, and then averaged to produce a group-level similarity matrix. Shown as triangular representational geometries recovered from multidimensional scaling in the above, shapes (blue) and sounds (orange) were approximately equated in their subjective similarity. These features were then used in the four-day crossmodal learning task. (b) Behavioral results from the four-day crossmodal learning task paired with multi-echo fMRI described in the main text. Before crossmodal learning, there was no difference in similarity between shape and sound features associated with congruent objects compared to incongruent objects – indicating that similarity was controlled at the unimodal feature-level. After crossmodal learning, we observed a robust shift in the magnitude of similarity. The shape and sound features associated with congruent objects were now significantly more similar than the same shape and sound features associated with incongruent objects (p < 0.001), evidence that crossmodal learning changed how participants experienced the unimodal features (observed in 17/18 participants). (c) We replicated this learning-related shift in pattern similarity with a larger sample size (n = 44; observed in 38/44 participants). *** denotes p < 0.001. Horizontal lines denote the comparison of congruent vs. incongruent conditions. – Supplemental Figure S1

(2) The experiences through which participants learned/experienced the shapes and sounds were unclear. The methods mention that they had one minute to explore/palpate each shape and that these experiences were interleaved with other tasks, but it is not clear what the other tasks were, how many such exploration experiences occurred, or how long the total learning time was. The manuscript also mentions that participants learn the shape-sound associations with 100% accuracy but it isn't clear how that was assessed. These details are important partly b/c it seems like very minimal experience to change neural representations in the cortex.

We apologize for the lack of detail and agree with the reviewer’s suggestions – we now include much more information in the methods section. Each behavioral day required about 1 hour of total time to complete, and indeed, participants rapidly learned their associations with minimal experience. For example:

“Behavioral Tasks. On each behavioral day (Day 1 and Day 3; Figure 2), participants completed the following tasks, in this order: Exploration Phase, one Unimodal Feature 1-back run (26 trials), Exploration Phase, one Crossmodal 1-back run (26 trials), Exploration Phase, Pairwise Similarity Task (24 trials), Exploration Phase, Pairwise Similarity Task (24 trials), Exploration Phase, Pairwise Similarity Task (24 trials), and finally, Exploration Phase. To verify learning on Day 3, participants also additionally completed a Learning Verification Task at the end of the session. – pg. 27

“The overall procedure ensured that participants extensively explored the unimodal features on Day 1 and the crossmodal objects on Day 3. The Unimodal Feature and the Crossmodal Object 1-back runs administered on Day 1 and Day 3 served as practice for the neuroimaging sessions on Day 2 and Day 4, during which these 1-back tasks were completed. Each behavioral session required less than 1 hour of total time to complete.” – pg. 27

“Learning Verification Task (Day 3 only). As the final task on Day 3, participants completed a task to ensure that participants successfully formed their crossmodal pairing. All three shapes and sounds were randomly displayed in 6 boxes on a display. Photos of the 3D shapes were shown, and sounds were played by clicking the box with the mouse cursor. The participant was cued with either a shape or sound, and then selected the corresponding paired feature. At the end of Day 3, we found that all participants reached 100% accuracy on this task (10 trials).” – pg. 29

(3) I didn't understand the similarity metric used in the multivariate imaging analyses. The manuscript mentions Z-scored Pearson's r, but I didn't know if this meant (a) many Pearson coefficients were computed and these were then Z-scored, so that 0 indicates a value equal to the mean Pearson correlation and 1 is equal to the standard deviation of the correlations, or (b) whether a Fisher Z transform was applied to each r (so that 0 means r was also around 0). From the interpretation of some results, I think the latter is the approach taken, but in general, it would be helpful to see, in Methods or Supplementary information, exactly how similarity scores were computed, and why that approach was adopted. This is particularly important since it is hard to understand the direction of some key effects.

The reviewer is correct that the Fisher Z transform was applied to each individual r before averaging the correlations. This approach is generally recommended when averaging correlations (see Corey, Dunlap, & Burke, 1998). We are now clearer on this point in the manuscript:

“The z-transformed Pearson’s correlation coefficient was used as the distance metric for all pattern similarity analyses. More specifically, each individual Pearson correlation was Fisher z-transformed and then averaged (see 61).” – pg. 32

(C) From Figure 3D, the temporal pole mask appears to exclude the anterior fusiform cortex (or the ventral surface of the ATL generally). If so, this is a shame, since that appears to be the locus most important to cross-modal integration in the "hub and spokes" model of semantic representation in the brain. The observation in the paper that the perirhinal cortex seems initially biased toward visual structure while more superior ATL is biased toward auditory structure appears generally consistent with the "graded hub" view expressed, for instance, in our group's 2017 review paper (Lambon Ralph et al., Nature Reviews Neuroscience). The balance of visual- versus auditory-sensitivity in that work appears balanced in the anterior fusiform, just a little lateral to the anterior perirhinal cortex. It would be helpful to know if the same pattern is observed for this area specifically in the current dataset.

We thank the reviewer for this suggestion. After close inspection of Lambon Ralph et al. (2017), we believe that our perirhinal cortex mask appears to be overlapping with the ventral ATL/anterior fusiform region that the reviewer mentions. See Author response image 1 for a visual comparison:

Author response image 1.

The top four figures are sampled from Lambon Ralph et al (2017), whereas the bottom two figures visualize our perirhinal cortex mask (white) and temporal pole mask (dark green) relative to the fusiform cortex. The ROIs visualized were defined from the Harvard-Oxford atlas.

We now mention this area of overlap in our manuscript and link it to the hub and spokes model:

“Notably, our perirhinal cortex mask overlaps with a key region of the ventral anterior temporal lobe thought to be the central locus of crossmodal integration in the “hub and spokes” model of semantic representations.9,50 – pg. 20

(D) While most effects seem robust from the information presented, I'm not so sure about the analysis of the perirhinal cortex shown in Figure 5. This compares (I think) the neural similarity evoked by a unimodal stimulus ("feature") to that evoked by the same stimulus when paired with its congruent stimulus in the other modality ("object"). These similarities show an interaction with modality prior to cross-modal association, but no interaction afterward, leading the authors to suggest that the perirhinal cortex has become less biased toward visual structure following learning. But the plots in Figures 4a and b are shown against different scales on the y-axes, obscuring the fact that all of the similarities are smaller in the after-learning comparison. Since the perirhinal interaction was already the smallest effect in the pre-learning analysis, it isn't really surprising that it drops below significance when all the effects diminish in the second comparison. A more rigorous test would assess the reliability of the interaction of comparison (pre- or post-learning) with modality. The possibility that perirhinal representations become less "visual" following cross-modal learning is potentially important so a post hoc contrast of that kind would be helpful.

We apologize for the lack of clarity. We conducted a linear mixed model to assess the interaction between modality and crossmodal learning day (before and after crossmodal learning) in the perirhinal cortex as described by the reviewer. The critical interaction was significant, which is now clarified in the text as well as in the rescaled figure plots.

“To investigate this effect in perirhinal cortex more specifically, we conducted a linear mixed model to directly compare the change in the visual bias of perirhinal representations from before crossmodal learning to after crossmodal learning (green regions in Figure 5a vs. 5b). Specifically, the linear mixed model included learning day (before vs. after crossmodal learning) and modality (visual feature match to crossmodal object vs. sound feature match to crossmodal object). Results revealed a significant interaction between learning day and modality in the perirhinal cortex (F1,775 = 5.56, p = 0.019, η2 = 0.071), meaning that the baseline visual shape bias observed in perirhinal cortex (green region of Figure 5a) was significantly attenuated with experience (green region of Figure 5b). After crossmodal learning, a given shape no longer invoked significant pattern similarity between objects that had the same shape but differed in terms of what they sounded like. Taken together, these results suggest that prior to learning the crossmodal objects, the perirhinal cortex had a default bias toward representing the visual shape information and was not representing sound information of the crossmodal objects. After crossmodal learning, however, the visual shape bias in perirhinal cortex was no longer present. That is, with crossmodal learning, the representations within perirhinal cortex started to look less like the visual features that comprised the crossmodal objects, providing evidence that the perirhinal representations were no longer predominantly grounded in the visual modality.” – pg. 13

We note that not all effects drop in Figure 5b (even in regions with a similar numerical pattern similarity to PRC, like the hippocampus – also see Supplemental Figure S5 for a comparison for patterns only on Day 4), suggesting that the change in visual bias in PRC is not simply due to noise.

“Importantly, the change in pattern similarity in the perirhinal cortex across learning days (Figure 5) is unlikely to be driven by noise, poor alignment of patterns across sessions, or generally reduced responses. Other regions with numerically similar pattern similarity to perirhinal cortex did not change across learning days (e.g., visual features x crossmodal objects in A1 in Figure 5; the exploratory ROI hippocampus with numerically similar pattern similarity to perirhinal cortex also did not change in Supplemental Figure S4c-d).” – pg. 14

(E) Is there a reason the authors did not look at representation and change in the hippocampus? As a rapid-learning, widely-connected feature-binding mechanism, and given the fairly minimal amount of learning experience, it seems like the hippocampus would be a key area of potential import for the cross-modal association. It also looks as though the hippocampus is implicated in the localizer scan (Figure 3c).

We thank the reviewer for this suggestion and now include additional analyses for the hippocampus. We found no evidence of crossmodal integrative coding different from the unimodal features. Rather, the hippocampus seems to represent the convergence of unimodal features, as evidenced by …[can you give some pithy description for what is meant by “convergence” vs “integration”?]. We provide these results in the Supplemental Information and describe them in the main text:

“Analyses for the hippocampus (HPC) and inferior parietal lobe (IPL). (a) In the visual vs. auditory univariate analysis, there was no visual or sound bias in HPC, but there was a bias towards sounds that increased numerically after crossmodal learning in the IPL. (b) Pattern similarity analyses between unimodal features associated with congruent objects and incongruent objects. Similar to Supplemental Figure S3, there was no main effect of congruency in either region. (c) When we looked at the pattern similarity between Unimodal Feature runs on Day 2 to Crossmodal Object runs on Day 2, we found that there was significant pattern similarity when there was a match between the unimodal feature and the crossmodal object (e.g., pattern similarity > 0). This pattern of results held when (d) correlating the Unimodal Feature runs on Day 2 to Crossmodal Object runs on Day 4, and (e) correlating the Unimodal Feature runs on Day 4 to Crossmodal Object runs on Day 4. Finally, (f) there was no significant pattern similarity between Crossmodal Object runs before learning correlated to Crossmodal Object after learning in HPC, but there was significant pattern similarity in IPL (p < 0.001). Taken together, these results suggest that both HPC and IPL are sensitive to visual and sound content, as the (c, d, e) unimodal feature-level representations were correlated to the crossmodal object representations irrespective of learning day. However, there was no difference between congruent and incongruent pairings in any analysis, suggesting that HPC and IPL did not represent crossmodal objects differently from the component unimodal features. For these reasons, HPC and IPL may represent the convergence of unimodal feature representations (i.e., because HPC and IPL were sensitive to both visual and sound features), but our results do not seem to support these regions in forming crossmodal integrative coding distinct from the unimodal features (i.e., because representations in HPC and IPL did not differentiate the congruent and incongruent conditions and did not change with experience). * p < 0.05, ** p < 0.01, *** p < 0.001. Asterisks above or below bars indicate a significant difference from zero. Horizontal lines within brain regions in (a) reflect an interaction between modality and learning day, whereas horizontal lines within brain regions in reflect main effects of (b) learning day, (c-e) modality, or (f) congruency.” – Supplemental Figure S4.

“Notably, our perirhinal cortex mask overlaps with a key region of the ventral anterior temporal lobe thought to be the central locus of crossmodal integration in the “hub and spokes” model of semantic representations.9,50 However, additional work has also linked other brain regions to the convergence of unimodal representations, such as the hippocampus51,52,53 and inferior parietal lobes.54,55 This past work on the hippocampus and inferior parietal lobe does not necessarily address the crossmodal binding problem that was the main focus of our present study, as previous findings often do not differentiate between crossmodal integrative coding and the convergence of unimodal feature representations per se. Furthermore, previous studies in the literature typically do not control for stimulus-based factors such as experience with unimodal features, subjective similarity, or feature identity that may complicate the interpretation of results when determining regions important for crossmodal integration. Indeed, we found evidence consistent with the convergence of unimodal feature-based representations in both the hippocampus and inferior parietal lobes (Supplemental Figure S4), but no evidence of crossmodal integrative coding different from the unimodal features. The hippocampus and inferior parietal lobes were both sensitive to visual and sound features before and after crossmodal learning (see Supplemental Figure S4c-e). Yet the hippocampus and inferior parietal lobes did not differentiate between the congruent and incongruent conditions or change with experience (see Supplemental Figure S4).” – pg. 20

(F) The direction of the neural effects was difficult to track and understand. I think the key observation is that TP and PRh both show changes related to cross-modal congruency - but still it would be helpful if the authors could articulate, perhaps via a schematic illustration, how they think representations in each key area are changing with the cross-modal association. Why does the temporal pole come to activate less for congruent than incongruent stimuli (Figure 3)? And why do TP responses grow less similar to one another for congruent relative to incongruent stimuli after learning (Figure 4)? Why are incongruent stimulus similarities anticorrelated in their perirhinal responses following cross-modal learning (Figure 6)?

We thank the author for identifying this issue, which was also raised by the other reviewers. The reviewer is correct that the key observation is that the TP and PRC both show changes related to crossmodal congruency (given that the unimodal features were equated in the methodological design). However, the structure of the integrative code is less clear, which we now emphasize in the main text. Our findings provide evidence of a crossmodal integrative code that is different from the unimodal features, and future studies are needed to better understand the structure of how such a code might emerge. We now more clearly highlight this distinction throughout the paper:

“By contrast, perirhinal cortex may be involved in pattern separation following crossmodal experience. In our task, participants had to differentiate congruent and incongruent objects constructed from the same three shape and sound features (Figure 2). An efficient way to solve this task would be to form distinct object-level outputs from the overlapping unimodal feature-level inputs such that congruent objects are made to be orthogonal from the representations before learning (i.e., measured as pattern similarity equal to 0 in the perirhinal cortex; Figure 5b, 6, Supplemental Figure S5), whereas non-learned incongruent objects could be made to be dissimilar from the representations before learning (i.e., anticorrelation, measured as patten similarity less than 0 in the perirhinal cortex; Figure 6). Because our paradigm could decouple neural responses to the learned object representations (on Day 4) from the original component unimodal features at baseline (on Day 2), these results could be taken as evidence of pattern separation in the human perirhinal cortex.11,12 However, our pattern of results could also be explained by other types of crossmodal integrative coding. For example, incongruent object representations may be less stable than congruent object representations, such that incongruent objects representation are warped to a greater extent than congruent objects (Figure 6).” – pg. 18

“As one solution to the crossmodal binding problem, we suggest that the temporal pole and perirhinal cortex form unique crossmodal object representations that are different from the distributed features in sensory cortex (Figure 4, 5, 6, Supplemental Figure S5). However, the nature by which the integrative code is structured and formed in the temporal pole and perirhinal cortex following crossmodal experience – such as through transformations, warping, or other factors – is an open question and an important area for future investigation. Furthermore, these anterior temporal lobe structures may be involved with integrative coding in different ways. For example, the crossmodal object representations measured after learning were found to be related to the component unimodal feature representations measured before learning in the temporal pole but not the perirhinal cortex (Figure 5, 6, Supplemental Figure S5). Moreover, pattern similarity for congruent shape-sound pairs were lower than the pattern similarity for incongruent shape-sound pairs after crossmodal learning in the temporal pole but not the perirhinal cortex (Figure 4b, Supplemental Figure S3a). As one interpretation of this pattern of results, the temporal pole may represent new crossmodal objects by combining previously learned knowledge. 8,9,10,11,13,14,15,33 Specifically, research into conceptual combination has linked the anterior temporal lobes to compound object concepts such as “hummingbird”.34,35,36 For example, participants during our task may have represented the sound-based “humming” concept and visually-based “bird” concept on Day 1, forming the crossmodal “hummingbird” concept on Day 3; Figure 1, 2, which may recruit less activity in temporal pole than an incongruent pairing such as “barking-frog”. For these reasons, the temporal pole may form a crossmodal object code based on pre-existing knowledge, resulting in reduced neural activity (Figure 3d) and pattern similarity towards features associated with learned objects (Figure 4b).” – pg. 18

This work represents a key step in our advancing understanding of object representations in the brain. The experimental design provides a useful template for studying neural change related to the cross-modal association that may prove useful to others in the field. Given the broad variety of open questions and potential alternative analyses, an open dataset from this study would also likely be a considerable contribution to the field.

Author Response

Reviewer #1 (Public Review):

This is a well performed study to demonstrate the antiviral function and viral antagonism of the dynein activating adapter NINL. The results are clearly presented to support the conclusions.

This reviewer has only one minor suggestion to improve the manuscript.

Add a discussion (1) why the folds of reduction among VSV, SinV and CVB3 were different in the NINL KO cells and (2) why the folds of reduction of VSV in the NINL KO A549 and U-2 OS cells.

Thank you for this suggestion. We have amended the results section to include additional information about these observations and possible explanations for these results.

Reviewer #2 (Public Review):

This manuscript is of interest to readers for host-viral co-evolution. This study has identified a novel human-virus interaction point NINL-viral 3C protease, where NINL is actively evolving upon the selection pressure against viral infect and viral 3Cpro cleavage. This study demonstrates that the viral 3Cpros-mediated cleavage of host NINL disrupts its adaptor function in dynein motor-mediated cargo transportation to the centrosome, and this disruption is both host- and virus-specific. In addition, this paper indicates the role of NINL in the IFN signaling pathway. Data shown in this manuscript support the major claims.

In this paper, the authors have identified a novel host-viral interaction, where viral 3C proteases (3Cpro) cleave at specific sites on a host activating adaptor of dynein intracellular transportation machinery, ninein-like protein (NINL or NLP in short) and inhibit its role in the antiviral innate immune response.

The authors firstly found that, unlike other activating adaptors of dynein intracellular transportation machinery, NINL (or NLP) is rapidly evolving. Thus, the authors hypothesized that this rapid evolution of NINL was caused by its interaction with viral infection. The authors found that viruses replicated higher in NINL knock-out (KO) cells than in wild-type (WT) cells and the replication level was not attenuated upon IFNa treatment in NINL KO cells, unlike in WT cells. Next, the authors investigated the role of NINL in type I IFN-mediated immune response and found that the induction of Janus kinase/signal transducer and activation of transcription (JAK/STAT) genes were attenuated in NINL KO cells upon IFNa treatment. The author further showed that the reduction of replication IFNa sensitive Vaccinia virus mutant upon IFNa treatment was decreased in NINL KO A549 cells compared to WT cells. The authors further showed that the virus antagonized NINL function by cleaving it with viral 3Cpro at its specific cleavage sites. NINL-peroxisome ligation-based cargo trafficking visualization assay showed that the redistribution of immobile membrane-bound peroxisome was disrupted by cleavage of NINL or viral infection.

This paper has revealed a novel host-virus interaction, and an antiviral function of a rapidly evolving activating adaptor of dynein intracellular transportation machinery, NINL. The major conclusions of this paper are well supported by data, but several aspects can be improved.

1) It would be necessary to include a couple of other pathways involved in innate immune response besides JAK/STAT pathway.

We are very interested in this question as well. Our RNAseq data (Supplementary file 4 and Figure 3 – Figure supplement 4) suggest that there are several transcriptional changes that result from NINL KO. Our goal in this manuscript was to focus on IFN signaling in order to understand this specific effect of NINL KO since it might have wide-ranging consequences on viral replication. While we agree that broadening our studies to other signaling pathways, including other pathways involved in innate immune response, is a good idea, we feel that those experiments would take longer than two months to perform and therefore fall outside of the scope of this paper.

2) The in-cell cleavages of NINL by viral 3Cpros were well demonstrated and supported by data of high quality. A direct biochemical demonstration of the cleavage is needed with purified proteins.

We agree with the reviewer that a direct biochemical cleavage assay would further demonstrate that viral 3Cpros cleave NINL specifically. However, our attempts to purify full-length NINL have been unsuccessful due to solubility issues (see example gel below), which is not surprising given that NINL is a >150 kDa human protein that has multiple surfaces that bind to other human proteins. As such, we focused our efforts on in-cell cleavage assays using specificity controls for cleavage. Specifically, we used catalytically inactive CVB3 3Cpro to show a dependence on protease catalytic activity and a variety of NINL constructs in which the glutamine in the P1 position is replaced by an arginine to show site specificity of cleavage. Notably, the cleavage sites in NINL that we mapped using this mutagenesis were predicted bioinformatically from known sites of 3Cpro cleavage in viral polyproteins, further indicating that cleavage is 3Cpro-dependent. We believe these results thus demonstrate that cleavage of NINL is dependent on viral protease activity and occurs in a sequence-specific manner. In light of the difficulty of purifying full-length NINL that would make biochemical experiments very challenging and likely take longer than two months to perform, we believe that our in cell data should be sufficient to demonstrate activity-dependent site-specific cleavage of NINL by viral 3Cpros.

Sypro stained SDS-PAGE gel showing supernatant (S) and insoluble pellet (P) fractions across multiple purifications with altered buffer conditions.

3) The author used different cell types in different assays. Explain the rationale with a sentence for each assay.

Throughout this work, we choose to use a variety of cell lines for specific purposes. A549 cells were chosen as our main cell line as they are widely used in virology, are susceptible to the viruses we used, are responsive to interferon, and express both NINL and our control NIN at moderate levels. In the case of our virology and ISG expression data, we performed the same experiments with NINL KOs in other cell lines confirm that the phenotypes we observed in A549 cells could be attributed to the absence of NINL rather than off-target CRISPR perturbations or cell-line specific effects. All cleavage experiments were performed in HEK293T for their ease of transfection and protein expression. The inducible peroxisome trafficking assays were performed in U-2 OS cells as their morphology is ideal for observing the spatial organization of peroxisomes via confocal microscopy, and based on the fact that we had recapitulated the virology results and ISG expression results in those cells. At the suggestion of the reviewer, we have amended the text to include rationales where appropriate.

4) While cell-based assays well support the conclusions in this paper, further demonstration in vivo would be helpful to provide an implication on the pathogenicity impact of NINL.

We agree. However, we believe that examining the impact of the loss of or antagonism of NINL on the pathogenesis of infectious diseases in an in vivo model is outside the scope of this study.

In summary, this manuscript contributes to a novel antiviral target. In addition, it is important to understand the host-virus co-evolution. The use of the evolution signatures to identify the "conflict point" between host and virus is novel.

Author Response

Reviewer #1 (Public Review):

In the article "Neuroendocrinology of the lung revealed by single cell RNA sequencing", Kuo et. al. described various aspects of pulmonary neuroendocrine cells (PNECs) including the scRNA-seq profile of one human lung carcinoid sample. Overall, although this manuscript does not have any specific storyline, it is informative and would be an asset for researchers exploring various new roles of PNECs.

Thank you for appreciating the significance of the data presented. Our storyline focuses on the newly uncovered molecular diversity of PNECs and the extraordinary repertoire of peptidergic signals they express and cell types these signals can directly target in (and outside) the lung, in mice and human, and in health and disease (human carcinoid tumor).

Major comments:

The major concern about the work is most results are preliminary, and at a descriptive level, conclusions or sub-conclusions are derived from scRNA-seq analysis only, lacking in-depth functional analysis and validation in other methods or systems. There are many open-end results that have been predicted by the authors based on their scRNA-seq data analysis without functional validation. In order to give them a constructive roadmap, it would be better to investigate literature and put them in a potential or probable hypothesis by citing the available literature. This should be done in each section of the result part. The paper lacks a main theme or specific biology question to address. In addition, the description about the human lung carcinoid by scRNA-seq is somehow disconnected from the main study line. Also, these results are derived from the study on only one single patient, lacking statistical power.

We agree that much of the data and analysis presented in the paper is descriptive and hypothesis-generating for PNECs, however we do not consider it preliminary. We focused on validating two key conclusions from the scRNA-seq analysis: PNECs are extraordinarily diverse molecularly (as validated by multiplex in situ hybridization and immunostaining) and they express many different combinations of peptidergic signals (and appear to package them in separate vesicles). From the lung expression profiles of the cognate receptors, we also predicted the direct lung targets of the dozens of new PNEC peptidergic signals we uncovered, and validated the cell target (PSN4, a recently identified subtype of pulmonary sensory neuron) of one of the newly identified PNEC signals (the classic hormone angiotensin) by confirming expression of the cognate receptor gene in PSN4 neurons that innervate PNECs and showing that the hormone can directly activate PSN4 neurons. The characterized human carcinoid provided evidence that during tumorigenesis, the amplified PNECs retain a memory (albeit imperfect) of the molecular subtype of PNEC from which they originated. As suggested by the Reviewer, we have provided more background in Results by adding additional citations from the literature to clarify the rationale for each analysis and what was known prior to the analysis. We feel that our paper provides a broad foundation for exploring the diversity and signaling functions of PNECs, and although each molecular type of PNEC and new PNEC peptidergic signal we uncovered and potential target cell in (and outside) the lung warrants follow up (as do the sensory and other properties of PNECs we inferred from their expression profiles), such studies will require the effort of many individuals in many labs studying both normal and disease physiology in mouse and human, and exploiting the data, hypotheses, approaches, and framework we provide.

Reviewer #2 (Public Review):

Pulmonary neuroendocrine cells (PNECs) are known to monitor oxygen levels in the airway and can serve as stem cells that repair the lung epithelium after injury. Due to their rarity, however, their functions are still poorly understood. To identify potential sensory functions of PNECs, the authors have used single-cell RNA-sequencing (scRNA-seq) to profile hundreds of mouse and human PNECs. They report that PNECs express over 40 distinct peptidergic genes, and over 150 distinct combinations of these genes can be detected. Receptors for these neuropeptides and peptide hormones are expressed in a wide range of lung cell types, suggesting that PNECs may have mechanical, thermal, acid, and oxygen sensory roles, among others. However, since some of these cognate receptors are not expressed in the lung, PNECs may also have systemic endocrine functions. Although these data are largely descriptive, the results represent a significant resource for understanding the potential roles of PNECs in normal biology as well as in pulmonary diseases and cancer and are likely to be relevant for understanding neuroendocrine cells in other tissue contexts.

However, there are several aspects of the data analysis that are unclear and require clarification, most notably the definition of a neuroendocrine cell (points #1 and #2 below).

1) Figure S1 shows the sorting strategy used for isolation of putative PNECs from Ascl1CreER/+; Rosa26ZsGreen/+ mice, and distinguishes neuroendocrine cells defined as ZsGreen+ EpCAM+ and "neural" cells defined as ZsGreen+ EpCAM-; the figure legend also refers to the ZsGreen+ EpCAM- cells as "control" cells. However, the table shown in panel D indicates that the NE population combines 112 ZsGreen+ EpCAM+ cells together with 64 ZsGreen+ EpCAM- cells to generate the 176 cells used for subsequent analyses. Why are these ZsGreen+ EpCAM- cells initially labeled as neural or control, but are then defined as neuroendocrine? If these do not express an epithelial marker, can they be rigorously considered as neuroendocrine?

As explained above in the response to Essential Revision point 1, we define pulmonary neuroendocrine cells (PNECs) throughout the paper by their transcriptomic clustering and signatures, which includes the dozens of newly identified PNEC markers as well as the few extant marker genes available before this study (listed in Table S2). The confusion here arises from the two previously known markers (Ascl1 lineage marker ZsGreen, EpCAM) we used for flow sorting to enrich for these rare cells for transcriptomic profiling (Fig. S1). Although most of the cells with PNEC transcriptomic profiles were from the ZsGreenhi EpCAMhi sorted population (as expected), some were from the ZsGreenhi EpCAMlo sorted population. The latter resulted from the high EpCAM gating threshold we used during flow sorting, which excluded some PNECs with intermediate levels of surface EpCAM. Indeed, nearly all PNECs (> 95%) expressed EpCAM by scRNAseq, and there was no difference in EpCAM transcript levels or transcriptomic clustering of PNECs that were from the ZsGreenhi EpCAMhi vs. ZsGreenhi EpCAMlo sorted populations, as we now show in the new panels (C', C'') added to Fig S1C. This point is now clarified in the legend to Fig. S1C, and it nicely demonstrates that transcriptomic profiling is a more robust method of identifying PNECs than flow sorting based on two classical markers.

2) Similarly, in the human scRNA-seq analysis, how were PNECs defined? The methods description states that these cells were identified by their expression of CALCA and ASCL1, but does not indicate whether they also expressed epithelial markers.

Human PNECs were identified in the single cell transcriptomic analysis by the same strategy described above for mouse PNECs: by their transcriptomic clustering and signatures, which includes the dozens of newly identified PNEC markers as well as the few extant marker genes available before this study (listed in Table S2). In addition to expression of classic and new markers, the human PNEC cluster defined by scRNA-seq indeed showed the expected expressed of epithelial markers (e.g, EPCAM, see dotplot below), like other epithelial cells.

3) The presentation of sensitivity and specificity in Figure 1 is confusing and potentially misleading. According to Figure 1B, Psck1 and Nov are two of the top-ranked differentially expressed genes in PNECs with respect to both sensitivity and specificity. However, the specificity of these two genes appears to be lower than that of Scg5, Chgb, and several other genes, as suggested in Figure 1C and Figure S1E. In contrast, Chgb appears to have higher specificity and sensitivity than Psck1 in Figures 1C and E but is not shown in the list of markers in Figure 1B.

As explained above in the response to Essential Revision point 2, because different marker features are important for different applications, we have provided several different graphical formats (Figs. 1B,C, Fig. S1E) and a table (Table S1) to aid in selection of the optimal markers for each application. Fig. 1B shows the most sensitive and specific PNEC markers identified by ratio of the natural logs of the average expression of the marker in PNECs vs. non-PNEC epithelial cells (Table S1), and we have added a two-dimensional plot of this sensitivity and specificity for a large set of PNEC markers (new panel E of Fig. S1). The violin plots in Fig. 1C allow visual comparison of expression of selected markers across PNECs and 40 other lung cell types including non-epithelial cells (from our extensive mouse lung atlas in Travaglini, Nabhan et al, Nature 2020). Pcsk1 and Nov score high in the analysis of Fig. 1B because they are highly sensitive and specific markers within the pulmonary epithelium, and they are also valuable markers because they are highly expressed in PNECs. However, they appear slightly less specific in the violon plots of Fig. 1C (Pcsk1) and Fig. S1F (Nov) because of expression (though at much lower levels) in individual lung cell types outside the epithelium: Pcsk1 is expressed also at low levels in some Alox5+ lymphocytes, and Nov is expressed at low levels in some smooth muscle cells. Chgb is a new PNEC marker that did not make the cutoff for the list in Fig. 1B because it is expressed in a slightly higher percentage of non-PNEC epithelial cells than the markers shown, which ranked slightly above it by this metric (see Table S1).

4) The expression of serotonin biosynthetic genes in mouse versus human PNECs deserves some comment. The authors fail to detect the expression of Tph1 and Tph2 in any of the mouse PNECs analyzed, but TPH1 is expressed in 76% of the human PNECs (Table S8). Is it possible that Tph1 and Tph2 are not detected in the mouse scRNA-seq data due to gene drop-out? If serotonin signaling by mouse PNECs is due to protein reuptake, as implied on p. 5, is there a discrepancy between serotonin expression as detected by smFISH versus immunostaining?

It is always possible that the failure to detect expression of Tph1 and Tph2 in the mouse scRNA-seq dataset is due to technical dropout, however when we analyzed this in our other mouse PNEC scRNA-seq dataset obtained using a microfluidic platform and also deeply-sequenced (Ouadah et al, Cell 2019), we found similar values as in the previously analyzed dataset: no Tph2 expression was detected and only 3% (3 of 92) of PNECs had detected Tph1 expression, whereas 24% (22 of 92) had detected expression of serotonin re-uptake transporter Slc6a4. Because our mouse and human scRNA-seq datasets were prepared similarly and sequenced to a similar depth (105 to 106 reads/cell), the difference observed in Tph1/TPH1 expression between mouse (0-3% PNECs) and human (76% PNECs) is more likely a true biological difference. We also analyzed serotonin levels in mouse PNECs by immunohistochemistry (not shown) and detected serotonin in nearly all (~90%) embryonic PNECs but only ~10% of adult PNECs. Systematic follow up studies will be necessary to resolve the mechanism of serotonin biogenesis and uptake in PNECs, and the potential stage and species-specific differences in these processes suggested by this initial data.

5) The smFISH and immunostaining analyses are often presented without any indication of the number of independent replicate samples analyzed (e.g., Figure 2B, Figure 3F, G).

The number of samples analyzed have been added (the values for Fig. 2B are given in legend to Fig. 2C, the quantification of Fig. 2B).

6) It would be helpful to provide a statistical analysis of the similarities and differences shown in the graphs in Figures 1E and G.

We added a statistical analysis (Fisher's exact test, two-sided) of Fig. 1E comparing expression of each examined gene in the two scRNA-seq datasets (Table S4). We added a similar statistical analysis of Fig. 1G comparing the expression values of each examined gene by scRNA-seq vs smFISH (see Fig. 1G legend).

Author Response

Reviewer #1 (Public Review):

This paper tests the hypothesis that 1/f exponent of LFP power spectrum reflects E-I balance in a rodent model and Parkinson's patients. The authors suggest that their findings fit with this hypothesis, but there are concerns about confirmation bias (elaborated on below) and potential methodological issues, despite the strength of incorporating data from both animal model and neurological patients.

First, the frequency band used to fit the 1/f exponent varies between experiments and analyses, inviting concerns about potentially cherry-picking the data to fit with the prior hypothesis. The frequency band used for fitting the exponent was 30-100 Hz in Experiment 1 (rodent model), 40-90 Hz in Experiment 2 (PD, levodopa), and 10-50 Hz in Experiment 3 (PD, DBS). Ad-hoc reasons were given to justify these choices, such as " to avoid a spectral plateau starting > 50 Hz" in Experiment 3. However, at least in Experiment 3 (Fig. 3), if the frequency range was shifted to 1-10 Hz, the authors would have uncovered the opposite effect, where the exponent is smaller for DBS-on condition.

We agree that parameter choice is crucial, in particular, choice of the fitting range. In addition to the 40-90 Hz range (Figure 2C), we have performed aperiodic fitting for five other frequency ranges to test to what extent the reported results are sensitive to the selected frequency range (Figure S2A). This analysis showed that the results are robust when a broad frequency range from 30 to 95 Hz was chosen, which is consistent with what has been suggested by Gao et al., 2017 to make inferences on the E/I ratio.

Accordingly, we have now repeated the analyses for the animal data with the same fitting range used for the ON-OFF medication comparison in humans. Along with Figure S2A where different frequency ranges were tested for data used in Figure 2, this shows that the results in Figure 1 and 2 hold up with higher aperiodic exponents when STN spiking is low and vice versa. Therefore, a broad fitting range from 30 to 90 Hz (excluding harmonics of mains interference) generates consistent results for both human and animal data.

We opted against a fitting range from 1-10 Hz because of two restraints highlighted in Gerster et al., 2022. First, a fitting range starting at 1 Hz could have a larger y-intercept due to the presence of low-frequency oscillations. This could lead to a larger aperiodic exponent and could be misinterpreted as stronger neural inhibition. Therefore, the lower fitting bound should be chosen to best avoid known oscillations in the delta/theta range (Gerster et al., 2022). Second, frequencies should be chosen to avoid oscillations crossing fitting range limits. In Figure 3A, oscillations in the theta/alpha band both ON and OFF stimulation would complicate parameterisation and would likely result in spurious fits.

We also tested the effect of changing the peak threshold, peak width limits and the aperiodic fitting mode on FOOOF parameterisation. Increasing and decreasing the peak threshold from its default value (at 2 standard deviations) did not change results (Figure S2B). Similarly, adapting the peak width limits did not affect the exponent difference between medication states (Figure S2C). Finally, choosing the ‘knee’ mode instead of ‘fixed’ resulted in fundamentally different aperiodic fits that did not differ anymore with medication (Figure S2D). This is most likely a consequence of the near linear PSD in log-log space from 40 to 90 Hz (Figure 2B). If there is no bend in the PSD, the FOOOF algorithm will be forced to assign a ‘random’ knee and the aperiodic fit will then mostly reflect the slope of the spectrum above the knee point.

Second, there are important, fine-grained features in the spectra that are ignored in the analyses, which confounds the interpretation.

One salient example of this is Fig. 2, where based on the plots in B, one would expect that the power of beta-band oscillations to be higher in the Med-On condition, as the oscillatory peaks rise higher above the 1/f floor and reach the same amplitude level as the Med-OFF condition (in other words, similar total power is subtracted by a smaller 1/f power in the Med-ON condition). But this impression is opposite to the model-fitting results in C, where beta power is lower in the Med-ON condition.

We agree that PSDs over a broad frequency range (e.g. 5-90 Hz) typically do not have a single 1/f property. Instead, there can be multiple oscillatory peaks and ‘knees/bends’ in the aperiodic component. For these cases, fitting should be performed using the knee mode. To extract periodic beta power, we parameterise the PSD between 5 and 90 Hz and select the largest oscillatory component between 8 and 35 Hz (this range was extended to include the large oscillatory peaks in hemispheres 27 and 28 at ~ 10 Hz, see Figure R1). We now use the knee mode, to model the aperiodic component between 5 and 90 Hz when periodic beta power is calculated (see our previous comments). Figure R1 provides an overview of all PSDs ON and OFF medication, the aperiodic fits (5-90 Hz (knee) and 40-90 Hz (fixed)) and the detected beta peaks. In spite of this modification in our pipeline, periodic beta power is still larger OFF medication (Figure 2C), in keeping with previous studies (Kim et al., 2022; Kühn et al., 2006; Neumann et al., 2017; Ray et al., 2008). We acknowledge the reviewer’s point that the average spectra in Figure 2B are misleading in that respect and for clarity provide here all 30 spectra in both conditions. Note that the calculation of aperiodic exponents between 40 and 90 Hz is not affected by this change in our pipeline. Figures 2B, D+E were revised accordingly.

We have repeated the analysis of our animal data using the ‘knee mode’ with a fitting range from 30 to 100 Hz. However, using the knee mode did not improve the goodness of fit or fitting error and, in fact, made them slightly worse (Figure S5). Based on this, we think the fixed mode would provide a more holistic model for the PSDs used in this analysis. We have now added this comparison in Figure S5 to justify the choice of the fixed mode.

Figure R1. PSDs from all 30 hemispheres ON and OFF medication. Aperiodic fits are shown between 5-90 Hz (knee mode), which was used to calculate the power of beta peaks, and between 40-90 Hz (fixed mode), which was used to estimate the aperiodic exponent of the spectrum.

Another example is Fig. 1C, where the spectra for high and low STN spiking epochs are identical between 10 and 20 Hz, and the difference in higher frequency range could be well-explained by an overall increase of broadband gamma power (e.g. as observed in Manning et al., J Neurosci 2012, Ray & Maunsell PLoS Biol 2011). This increase of broadband gamma power is trivially expected, as broadband gamma power is tightly coupled with population spiking rate, which was used to define the two conditions.

We agree with the reviewer that in Figure 1C, high and low STN spiking states could well be separated by average gamma power (Figure 1E), too. However, the difference of aperiodic exponents is more prominent between both conditions (Figure 1D+E, based on p-values). What is more, in human LFP data recorded from clinical macroelectrodes, medication states can be reasonably well distinguished using the aperiodic exponent between 40-90 Hz (Figure 2C), but average gamma power does not separate both states (Figure S3A). This suggests that the aperiodic exponent reflects more than just power differences in the high gamma regions. In addition, power changes do not inevitably change the aperiodic exponent and vice versa as elaborated in (Donoghue et al., 2020).

Manning et al., 2009 show that the power spectrum is shifted to higher power values at all observed frequencies (2-150 Hz) as firing rates increase. As the reviewer points out, power spectra of our data are almost identical between 10-20 Hz (despite the marked spiking differences) and only drift apart from > 20 Hz (Figure 1C). This is a relevant difference between our study and Manning et al., 2009 and suggests that power differences in the gamma range are not solely explained by differences in spiking. This is confirmed when cortical activity at different spikes/sec is modelled (Miller et al., 2009). The entire spectrum is shifted to higher power values if spiking rates increase.

Ray & Maunsell, 2011 reported low (30-80 Hz) and high (> 80 Hz) gamma activity in the macaque visual cortex, with a positive correlation between spiking activity and high gamma activity. However, activities in the low gamma range (30-80 Hz), which largely overlaps with the frequency range in our study, does not necessarily correlate with firing rates.

In conclusion, the link between gamma power and spiking activity is not as strong as alluded. Even if the change in spiking activities can lead to changes of both gamma power and the aperiodic exponent, the aperiodic exponent would still constitute a measure to separate E/I levels and medication states.

The above consideration also speaks to a major weakness of the general approach of considering the 1/f spectrum a monolithic spectrum that can be captured by a single exponent. As the authors' Fig. 1C shows, there are distinct frequency regions within the 1/f spectrum that have different slopes. Indeed, this tripartite shape of the 1/f spectrum, including a "knee" feature around 40-70 Hz which is well visible here, was described in multiple previous papers (Miller et al., PLoS Comput Biol 2009; He et al., Neuron 2010), and have been successfully modeled with a neural network model using biologically plausible mechanisms (Chaudhuri et al., Cereb Cortex, 2017). The neglect of these fine-grained features confounds the authors' model fitting, because an overall increase in the broadband gamma power - which can be explained straightforwardly by the change in population firing rates - can result in the exponent, fit over a larger spectral frequency region, to decrease. However, this is not due to the exponent actually changing, but the overall increase of power in a specific sub-frequency-region of the broadband 1/f activity.

We have now used the knee mode for aperiodic fits between 5 and 90 Hz when periodic beta power is calculated. We agree that this broad frequency range is unlikely to have a single 1/f component.

We have also repeated the analysis of our animal data using the knee mode for aperiodic fits between 30 and 100 Hz (Figure S5). However, the goodness of fits had barely changed. In fact, the R2 and error become slightly worse. In addition, the knee parameter complicates interpretation of the aperiodic exponent and has to be considered along with the knee frequency. What is more, we do not see this bend around 40-70 Hz in all subjects. We show PSDs of representative LFP channels in Figure R2 and need to assert that the knee around 40-70 Hz is not a robust finding in our data set. Therefore, we chose the fixed mode for parameterisation within this frequency band.

Please see our answer to the previous comment regarding the link between broad gamma power and changes in population firing rates.

Figure R2. PSDs of representative PSD channels for each animal (data used in Figure 1C). The knee around 40-70 Hz is not a robust finding in all PSDs.

Author Response

Reviewer #1 (Public Review):

Iyer et al. address the problem of how cells exposed to a graded but noisy morphogen concentration are able to infer their position reliably, in other words how the positional information of a realistic morphogen gradient is decoded through cell-autonomous ligand processing. The authors introduce a model of a ligand processing network involving multiple ”branches” (receptor types) and ”tiers” (compartments where ligand-bound receptors can be located). Receptor levels are allowed to vary with distance from the source independently of the morphogen concentration. All rates, except for the ligand binding and unbinding rates, are potentially under feedback control. The authors assume that the cells can infer their position from the output of the signalling network in an optimal way. The resulting parameter space is then explored to identify optimal ”network architectures” and parameters, i.e. those that maximise the fidelity of the positional inference. The analysis shows how the presence of both specific and non-specific receptors, graded receptor expression and feedback loops can contribute to improving positional inference. These results are compared with known features of the Wnt signalling system in Drosophila wing imaginal disc.

The authors are doing an interesting study of how feedback control of the signalling network reading a morphogen gradient can influence the precision of the read-out. The main strength of this work is the attention to the development of the mathematical framework. While the family of network architectures introduced here is not completely generic, there is enough flexibility to explore various features of realistic signalling systems. It is exciting to find that some network topologies are particularly efficient at reducing the noise in the morphogen gradient. The comparison with the Wnt system in Drosophila is also promising.

Major comments:

1) The authors assume that the cell estimates its position through the maximum a posteriori estimate, Eq.(5), which is a well-defined mathematical object; it seems to us however that whether the cell is actually capable of performing this measurement is uncertain (it is an optimal measurement in some sense, but there is no guarantee that the cell is optimal in that respect). Notably, this entails evaluating p(theta), which is a probability distribution over the entire tissue, so this estimate can not be done with purely local measurements. Can the authors comment on this and how the conclusions would change if a different position measurement was performed?

This is indeed an important question. Our viewpoint is that if the cells were to use a maximum a posteriori (MAP) estimate (Eq. 5) to decode their positions, then what features of the channel architecture would lead to small errors in positional inference. Whether the maximum a posteriori estimate is employed by the cell, or some other estimate, is an important but difficult question to address. Our choice has been motivated by how this estimate has allowed the precise determination of developmental fates in the context of gap gene expression in Drosophila embryo [1, 2, 3]. We had earlier computed the inference error with a different estimate i.e.

which computes the mean squared deviations of the inferred positions from the true position for each x, taking into account the entire distribution p(x∗|x). While the qualitative results are the same, the inference errors showed spurious jitters from outliers in sampling the noisy morphogen input distribution. This consistency might suggest that our qualitative results are insensitive to the choice of the estimate.

Further, when evaluating the MAP estimate, the term p(θ) in the denominator serves as a normalisation factor to ensure p(x|θ) is a probability density. This is not strictly necessary for MAP estimation. Since p(θ) does not depend on x, the MAP estimate can be written as follows

without the need for evaluating p(θ). In the case of a uniform prior, it would be equivalent to maximum likelihood estimate (MLE) i.e.

2) One of the features of the signalling networks studied in the manuscript is the ability of the system to form a complex (termed a conjugated state, Q) made of two ligands L, one receptor and one nonsignalling receptor. While there are clear examples of a single ligand binding to two signalling receptors (e.g. Bmps), are there also known situations where such a complex with two ligands, one receptor, and one non-signalling receptor can form? In the Wnt example (Fig. 10a), it is not clear what this complex would be? In general, it would be great to have a more extended discussion of how the model hypothesis for the signalling networks could relate to real systems.

This is a good suggestion. We have now added a discussion on the various possible realisations of the “conjugate state” Q in Section 3.6. We have also explored the various states in the context of different signalling contexts such as Dpp, Hh, Fgf in the Discussion section.

The conjugated state ‘Q’ represents a combination of the readings from the two branches i.e. receptor types. This could be realised through processes like ligand exchange or complex formation, both in a shared spatial location such as a compartment. As discussed in the original manuscript (Section 3.6 of the revised manuscript), the ligand Wg in the Wg signalling pathway is internalised through two separate endocytic pathways associated with the receptor types - signalling receptor Frizzled (via Clathrin-mediated endocytosis (CME)) and non-signalling receptor HSPGs (via the CLIC/GEEC pathway (CLIC - (clathrin-independent carriers, GEEC - GPI-anchored protein-enriched early endosomal compartments)). Both pathways meet in a common early endosomal compartment where the ligands may be exchanged between the two receptors [4]. In a previous work by Hemalatha et al [4], we had shown that there are more Wg-DFz2 interactions in the endosomal compartment (measured through FRET) than on the cell surface. Therefore, the non-signalling receptors directing Wg through the CLIC/GEEC pathway titrate the amount of Wg interaction with the signalling receptor, DFz2.

As mentioned in the original manuscript (Section 3.3 and subsection 4.2 of the Discussion in the revised manuscript), apart from Wg signalling, non-signalling receptors such as the HSPGs have also been proposed to act as co-receptors for Dpp, Hh, FGF (reviewed in [5, 6]). Although some ligands bind to the core protein of HSPG, the majority of the ligands bind to the negatively charged HS chains [7, 8]. Here, the coreceptors HSPGs aid in capturing diffusible ligands and presenting the same to signalling receptors (either on the cell surface or within endosomes).

3) The authors consider feedback on reaction rates - it would seem natural to also consider feedback on the total number of receptors; notably, since there are known examples of receptors transcriptionally down-regulated by their ligands (e.g. Dpp/Tkv)? Also it is not clear in insets such as in Fig. 7b, if the concentration plotted corresponds to the concentration of receptors bound to ligands?

As mentioned in the original manuscript (Section 2.2 of the revised manuscript), we have indeed considered control on reaction rates and receptors, although the control on the latter is done with the constraint of receptor profiles being monotonic. Further, while the control on reaction rates is considered via feedbacks explicitly, the control on receptors is done via an approach akin to the openloop control used in control theory. In reality, cellular control on receptors will involve transcriptional up- or down-regulation of receptor and thus warrant a feedback control approach – however, the timescales involved in such a control are different from the binding-unbinding and signalling timescales.

Therefore, in the current work, we take the morphogen profile to be given i.e. independent of receptor concentrations, and we ask for the receptor concentrations that would help reduce the inference errors.

Our predictions of increasing signalling receptor and decreasing non-signalling receptors in a twobranch channel architecture are consistent with the known transcriptional up-regulation of Dally/Dlp and down-regulation of Fz by Wg signalling [9].

In a future work, we will extend the control on receptors to include feedbacks explicitly. Furthermore, the explicit feedback control on receptors may need to be considered concomitantly with the effect of receptors on morphogen dynamics (i.e. morphogen sculpting by receptors) along with the possibility of spatial correlations in receptor concentrations through neighbouring cell-cell interactions.

As mentioned in the original manuscript (Section 2.2 of the revised manuscript), the variables ψ and φ stand for the total (bound + unbound) surface receptor concentrations of the signalling and the non-signalling receptors respectively. Therefore, the insets showing receptor profiles such as in Fig. 6b, 7b, and Appendix H Fig.8b,e correspond to the total surface receptor concentrations.

4) The authors are clear about the fact that they consider the morphogen gradient to be fixed independently of the reaction network; however, that seems like a very strong assumption; in the Dpp morphogen gradient for instance over expression of the Tkv receptor leads to gradient shortening. Can the authors comment on this?

This point is related to the earlier question 4. As discussed in the Discussion of the original manuscript (subsection 4.3 of the revised manuscript), we focus on finding the optimal receptor concentration profiles and reaction networks that enable precision and robustness in positional information from a given noisy morphogen profile. The framework and the optimisation scheme within it will prescribe different receptor profiles and reaction networks for different monotonically behaving, noisy morphogen profiles. It is possible that cells may achieve the optimal receptor concentrations via feedback control on production of the receptors.

Broadly, morphogen dynamics depends on cell surface receptors, which could participate in both the inference and the sculpting of the morphogen profile, and factors independent of them such as extracellular degradation, transport and production, etc. In our present work, we have taken the receptors involved in sculpting and inference as being independent.

In a more general case, feedback control on receptors will change the receptor concentrations as well as the morphogen profile. We are currently working on realising such a feedback control on receptors within the same broader information theoretic framework proposed in the current work.

5) Fig. 10f is showing an exciting result on the change in endocytic gradient CV in the WT and in DN mutant of Garz. Can the authors check that the Wg morphogen gradient is not changing in these two conditions? And can they also show the original gradient, and not only its CV?

The reviewer raises a legitimate concern – could the observed changes in CV upon perturbation of endocytic machinery be attributed to a systematic change in the mean levels of the endocytosed Wg alone? In the original manuscript (Appendix O Fig.17b,c of the revised manuscript), we show the normalised profiles of endocytic Wg in control and myr-Garz-DN cases. Here, in Fig.1 below, we show a comparison between the mean Wg concentrations (measured as fluorescence intensity) in control wing discs and discs wherein CLIC/GEEC endocytic pathway is removed using UAS-myr-Garz-DN. For clarity, we show the discs with largest and smallest fluorescence intensities from the control and myr-Garz-DN discs. It is hard to conclude that the mean concentrations are significantly different in the two cases.

Reviewer #2 (Public Review):

The work of Iyer et al. uses a computational approach to investigate how cells using multiple tiers of processing and multiple parallel receptor types allow more accurate reading of position from a noisy signal. Authors find that combining signaling and non-signaling types of receptors together with additional feedback increases the accuracy of positional readout against extrinsic noise that is conveyed in the morphogen signal. Further, extending the number of layers of signal processing counteracts the intrinsic stochasticity of the signal reading and processing steps. The mathematical formulation of the model is general but comprehensive in the way it handles the difference between branches and tiers for the processing of channels with feedbacks. The results of the model are presented from simple one-branch and one-tier architecture to two-branch and two-tier architecture with feedbacks. Interestingly authors find that adding more tiers results in only very small improvements in the accuracy of positional readout. The model is tested against a perturbation experiment that impairs one of the signaling branches in the Drosophila wing disc, but the comparison is only qualitative as further experiment-oriented work is planned in a separate paper.

Strengths

There is a clear statement of objectives, model, and how the model is evaluated. In particular, the objective is to find what number of receptor types and their concentrations for a given number of tiers and feedback types is resulting in the most accurate positional readout. The employed optimization procedure is capable to find signalling architectures that result in one cell diameter positional precision for most of the tissue with 3-4 cells at the tissue end that is most distant to the morphogen source. This demonstrates that employing additional complexity in signal processing results in a very accurate positional readout, which is comparable with estimates of positional precision obtained in other developmental systems (Petkova et al., Cell 2019, Zagorski et al., Science 2017).

The optimal signalling architectures indicate that both signalling (specific) and non-signalling (nonspecific) receptors affect the precision of positional readout, but the contributions of each type of these receptors are qualitatively different. Even slight perturbation of signalling receptors drives the system out of optimum, resulting in a decrease in positional precision. In contrast, the non-signalling receptors could accommodate much larger perturbations. This observation could provide a biophysical explanation for how cross-talk between different morphogen species could be realized in a way that positional precision is kept at the optimum when morphogen signaling undergoes extrinsic and intrinsic perturbations.

Last, the model formulation allows to specifically address perturbations of signalling and feedbacks, that could be explored to validate model predictions experimentally in Drosophila wing disc, but also in other developmental tissues. The authors present a proof-of-concept by obtaining consistent results of variation of output profiles in two-tier two-branch architectures with non-signaling branch removed and intensity profiles of Wg in wing disc where the CLIC/GEEC endocytic pathway was perturbed.

Weaknesses

The list of model parameters is long including more than 20 entries for two-tier two-branch architectures. This is expected, as the aim of the model is to describe the sophisticated signalling architecture mimicking the biological system. However, this also makes it very challenging or impossible to provide guiding principles or understanding of the system behaviour for the complete space of signalling architectures that optimize positional readout. Although, the employed optimization procedure finds solutions that exhibit very high positional accuracy, there is only very limited notion how these solutions depend on variation of different parameters. The authors do not address the following question, whether these solutions correspond to broad global optima in the space of all solutions, or were rather fine-tuned by the optimization procedure and are quite rare.

It is unclear how contributions from the intrinsic noise affect the system behaviour compared to contributions from extrinsic noise. In principle, the two-branch one-tier architecture results in an already very accurate positional readout across the tissue. The adding of another tier seems to provide only a very weak improvement over a one-tier solution. It is possible that contributions from intrinsic noise for the investigated signalling architectures are only mildly affecting the system compared with contributions from extrinsic noise. Hence, it is difficult to assess whether the claim of reducing intrinsic noise by adding another tier is supported by the presented data, as the contributions from intrinsic noise could overall very weakly affect the positional readout.

The optimal response of the channel to extrinsic and intrinsic noises is very distinct. As noted correctly by the reviewer, an additional tier provides only a marginal improvement in inference error due extrinsic noise (compare Fig.7 and Fig.8 in the revised manuscript). However, as shown in Fig.9c of the revised manuscript (same as in the original manuscript), adding an extra tier provides a substantial improvement in inference errors due to intrinsic noise.

References

[1] Gasper Tkacik, Julien O Dubuis, Mariela D Petkova, and Thomas Gregor. Positional information, positional error, and readout precision in morphogenesis: a mathematical framework. Genetics, 199:39– 59, 2015.

[2] Mariela D Petkova, Gasper Tkacik, William Bialek, Eric F Wieschaus, and Thomas Gregor. Optimal decoding of cellular identities in a genetic network. Cell, 176:844–855, 2019.

[3] Julien O Dubuis, Gaˇsper Tkaˇcik, Eric F Wieschaus, Thomas Gregor, and William Bialek. Positional information, in bits. Proceedings of the National Academy of Sciences, 110:16301–16308, 2013.

[4] Anupama Hemalatha, Chaitra Prabhakara, and Satyajit Mayor. Endocytosis of wingless via a dynaminindependent pathway is necessary for signaling in drosophila wing discs. Proceedings of the National Academy of Sciences, 113:E6993–E7002, 2016.

[5] Xinhua Lin. Functions of heparan sulfate proteoglycans in cell signaling during development. Development, 131:6009–6021, 2004.

[6] Stephane Sarrazin, William C Lamanna, and Jeffrey D Esko. Heparan sulfate proteoglycans. Cold Spring Harbor perspectives in biology, 3(7):a004952, 2011.

[7] Catherine A Kirkpatrick, Sarah M Knox, William D Staatz, Bethany Fox, Daniel M Lercher, and Scott B Selleck. The function of a drosophila glypican does not depend entirely on heparan sulfate modification. Developmental biology, 300(2):570–582, 2006.

[8] Mariana I Capurro, Ping Xu, Wen Shi, Fuchuan Li, Angela Jia, and Jorge Filmus. Glypican-3 inhibits hedgehog signaling during development by competing with patched for hedgehog binding. Developmental cell, 14(5):700–711, 2008.

[9] Kenneth M Cadigan, Matthew P Fish, Eric J Rulifson, and Roel Nusse. Wingless repression of drosophila frizzled 2 expression shapes the wingless morphogen gradient in the wing. Cell, 93(5):767–777, 1998.

Author Response

Reviewer #1 (Public Review):

Strength: The study is summarizing a large cohort of human samples of blood, nasal swabs and nasopharyngeal aspirates. This is very uncommon as most of the time studies focus on the blood and serum of patients. Within the study, 3 monocyte and 3 DC subsets have been followed in healthy and Influenza A virus-infected persons. The study also includes functional data on the responsiveness of Influenza A virus-infected DC and monocyte populations. The authors achieved their aims in that they were able to show that the tissue microenvironment is important to understand subset specific migration and activation behavior in Influenza A virus infection and in addition that it matters with which kind of agent a person is infected. Thus, this study also impacts a better understanding of vaccine design for respiratory viruses.

We thank Reviewer 1 for highlighting what we believe to be the greatest strengths of our study. The key feature of this study was to generate a comprehensive description of monocytes and dendritic cells (DC) in the human nasopharynx during influenza A virus infection, and to provide a comparison with healthy and convalescent individuals. Further, we wished to emphasize the value of studying the nasopharynx during respiratory viral infections, particularly in light of the ongoing COVID-19 pandemic. We describe a non-invasive method to (longitudinally) sample this anatomical compartment that allows retrieval of intact immune cells as well as mucosal fluid for soluble marker analysis. We also believe that the addition of proteomic profiles in the different compartments (new Figure 7) further highlights the importance of the tissue microenvironment.

Weakness: In the described study, the authors used a different nomenclature to introduce the DC subsets. This is confusing and the authors should stick to the nomenclature introduced by Guilliams et al., 2014 (doi.org/10.1038/nri3712) and commented in Ginhoux et al., 2022 (DOI: 10.1038/s41577-022-00675-7 ) or at least should introduce the alternative names (cDC1, cDC2, expression markers XCR1, CD172a/Sirpa). Further, Segura et al., 2013 (doi: 10.1084/jem.20121103) showed that all three DC subpopulations were able to perform cross-presentation when directly isolated. Overall, a more up-to-date introduction would be useful.

Reviewer 1 commented on the DC nomenclature used in the manuscript. We agree that our manuscript would benefit from appropriately updating the DC nomenclature. We therefore revised the text, and now we refer to the subsets previously described as CD1c+ and CD141+ myeloid DCs (MDC) as cDC2 and CDC1 subsets, respectively. We have also modified the text in the Introduction of the revised manuscript to reflect the same and give a more up-to-date introduction of DC subsets (marked-up version lines 75-81).

As the data of this was already obtained in 2016-2018 it is clear that the FACS panel was not developed to study DC3. If possible, the authors might be able to speculate about the role of this subset in their data set. Moreover, there were other studies on SARS-CoV-2 infection and DC subset analyses in blood (line 87, and line 489) e.g. Winheim et al., (DOI: 10.1371/journal.ppat.1009742 ), which the authors should introduce and discuss in regard to their own data.

As reviewer 1 accurately pointed out, the flow cytometry panel used in this study was indeed not developed to study the DC3 subset. The data was obtained in 2016-2018, and lack the typical markers used to identify the DC3 subset, such as CD163, BTLA and CD5 (Cytlak et al, https://doi.org/10.1016/j.immuni.2020.07.003, Villani et al, https://doi.org/10.1126/science.aah4573). Due to the constraints of the panel, we would not be able to accurately identify DC3s. However, in an attempt to dig deeper into the data that is available, we re-analyzed the data to identify CD14+CD1c+ cells among the lineage–HLADR+CD16–CD14+ cells, here collectively called “mo-DC”. This population is likely a combination of monocytes upregulating CD1c and bona fide DC3 expressing CD14. Accordingly, the gating strategy was updated in Supplementary figure 1 (marked-up version lines 192-194), and new data plot in Figure 2H (marked-up version lines 208-220) summarizes the changes observed in mo-DC numbers in IAV patients between blood and the nasopharynx. Parallel to the pattern seen in other DC subsets, mo-DC frequencies are reduced in blood and we observed an increase (not significant) in the nasopharynx.

As CD88 was not included in the original panel, it was not possible to discriminate between bona fide monocytes and DC3s. We performed a staining of PBMCs (buffy coat) with CD88 (FITC) added to the original flow panel used in the study, to assess if CD88 can be helpful for future studies (Reviewer figure 1). The staining showed that some cells in the mo-DC population are CD88 positive, indicating a bona fide monocyte origin, whereas some are negative, indicating that they are bona fide DC3 expressing CD14. (Bourdely et al, https://doi.org/10.1016/j.immuni.2020.06.002).

Reviewer figure 1. Expression of CD88 in the “mo-DC” population. Cells from a buffy coat were stained with the flow cytometry panel used in the manuscript, with the addition of CD88 (FITC). Within the CD14+CD1c+ population, the “mo-DC” population, we identified both CD88+ and CD88- cells.

Reviewer 1 also suggested citing Winheim et al (https://doi.org/10.1371/journal.ppat.1009742), and we thank them for their suggestion. We have now cited Winheim et al, and two additional reports (Kvedaraite et al, https://doi.org/10.1073/pnas.2018587118 and Affandi et al, https://doi.org/10.3389/fimmu.2021.697840) describing a depletion of DC3s (and other DC subsets) from circulation, and functional impairment of DCs following SARS-CoV-2 infection. Further, Winheim et al observed an increased frequency of a CD163+CD14+ subpopulation within the DC3s, which correlated with systemic inflammatory responses in SARS-CoV-2 infection. We speculate that perhaps in IAV infection too, DC3s may follow the trend of other DC subsets and be found in increased numbers in the nasopharynx (marked-up version lines 75-81 and 543-552).

Taken together, although the data are very important and very interesting, my overall impression of the manuscript is that in the era of RNA seq and scRNA seq analyses the study lacks a bit of comprehensiveness.

The final comment from reviewer 1 is well taken, in that our study does not include RNA-seq analyses. Again, we ask Reviewer 1 to take into consideration the challenging material we worked with in our study in combination with the COVID-19 pandemic that subsequently has excluded recruitment of new influenza patients to the study. The cell numbers and viability in the nasopharyngeal aspirates limit what experimental approaches can be done simultaneously, and flow cytometry seemed to be the best approach for the study. However, we agree that in future studies, both our own and those of others in the field, will greatly benefit from single cell analysis of nasopharyngeal immune cells, and from generating transcriptomic or epigenetic profiles of these cells. Unfortunately, it is a limitation that we are currently unable to overcome within the scope of this revision. Despite this weakness, we agree with Reviewer 1 that the methods we developed and the data we generated are important and interesting.

Moreover, we have added additional proteomics data from both NPA and plasma from influenza and COVID-19 patients, using the SomaScan platform (new Figure 7) (marked-up version lines 472-511, 738-755 and 768-792). We also included a supplementary table listing enriched pathway data from gProfiler. Briefly, our data showed sizeable changes within the blood and nasopharyngeal proteome during respiratory virus infection (IAV or SARS-CoV-2), as compared to healthy controls. Importantly, we found several differentially expressed proteins unique to the nasopharynx that were not seen in blood, and pathway analysis highlighted “host immune responses” and “innate immunity” pathways, containing TNF, IL-6, ISG15, IL-18R, CCL7, CXCL10 (IP-10), CXCL11, GZMB, SEMA4A, S100A8, S100A9. These findings are in line with our flow cytometry data, and support our hypothesis that the immunological response to viral infection in the upper airways differ from that in matching plasma samples. One of the main messages in this manuscript is the importance of looking at the site of infection, and not only at systemic immune responses to better understand respiratory viral infections in humans. We believe that the addition of the proteomics data serves to further highlight this point.

Reviewer #2 (Public Review):

This study aims to describe the distribution and functional status of monocytes and dendritic cells in the blood and nasopharyngeal aspirate (NPA) after respiratory viral infection in more than 50 patients affected by influenza A, B, RSV and SARS-CoV2. The authors use flow cytometry to define HLA-DR+ lineage negative cells, and within this gate, classical, intermediate and non-classical monocytes and CD1c+, CD141+, and CD123+ dendritic cells (DC). They show a large increase in classical monocytes in NPA and an increase in intermediate monocytes in blood and NPA, with more subtle changes in non-classical monocytes. Changes in intermediate monocytes were age-dependent and resolution was seen with convalescence. While blood monocytes tended to increase in blood and NPA, DC frequency was reduced in blood but also increased in NPA. There were signs of maturation in monocytes and DC in NPA compared with blood as judged by expression of HLA-DR and CD86. Cytokine levels in NPA were increased in infection in association with enrichment of cytokine-producing cells. Various patterns were observed in different viral infections suggesting some specificity of pathogen response. The work did not fully document the diversity of human myeloid cells that have arisen from single-cell transcriptomics over the last 5 years, notably the classification of monocytes which shows only two distinct subsets (intermediate cannot be distinguished from classical), distinct populations of DC1, DC2 and DC3 (DC2 and 3 both having CD1c, but different levels of monocyte antigens), and the lack of distinction provided by CD123 which also includes a precursor population of AXL+SIGLEC6+ myeloid cells in addition to plasmacytoid DC. Furthermore, some greater precision of the gating could have been achieved for the subsets presented. Specifically, CD34+ cells were not excluded from the HLA-DR+ lineage- gate, and the threshold of CD11c may have excluded some DC1 owing to the low expression of this antigen. Overall, the work shows that interesting results can be obtained by comparing myeloid populations of blood and NPA during viral infection and that lineage, viral and age-specific patterns are observed. However, the mechanistic insights for host defense provided by these observations remain relatively modest.

We thank Reviewer 2 for their assessment of our manuscript and summarizing our key findings in their public review. As reviewer 2 noted, our study describes changes in frequencies of monocytes and DCs during acute IAV infection, in blood and in the nasopharynx. Additionally, we also demonstrate pathogen-specific changes in both compartments. Reviewer 2 also highlighted a drawback of our study- that the approach did not fully capture the breadth of monocyte and DC diversity as it currently stands. Despite this, the findings we presented here laid the groundwork for continued research and led to significant progress, including mechanistic insights (Falck-Jones et al, https://doi.org/10.1172/JCI144734 and Cagigi et al, https://doi.org/10.1172/jci.insight.151463, Havervall et al. https://doi.org/10.1056/nejmc2209651 and Marking et al. Lancet Infectious Diseases in press), in understanding the role of myeloid cells in the human airways during viral infections.

Author Response

We thank the reviewers for their positive feedback and thoughtful suggestions that will improve our manuscript. Here we summarise our plan for immediate action. We will resubmit our manuscript once additional experiments have been performed to clarify all the major and minor concerns of the reviewers and the manuscript has been revised. At that point, we will respond to all reviewer’s points and highlight the changes made in the text.

Reviewer #1 (Public Review):

The authors have tried to correlate changes in the cellular environment by means of altering temperature, the expression of key cellular factors involved in the viral replication cycle, and small molecules known to affect key viral protein-protein interactions with some physical properties of the liquid condensates of viral origin. The ideas and experiments are extremely interesting as they provide a framework to study viral replication and assembly from a thermodynamic point of view in live cells.

The major strengths of this article are the extremely thoughtful and detailed experimental approach; although this data collection and analysis are most likely extremely time-consuming, the techniques used here are so simple that the main goal and idea of the article become elegant. A second major strength is that in other to understand some of the physicochemical properties of the viral liquid inclusion, they used stimuli that have been very well studied, and thus one can really focus on a relatively easy interpretation of most of the data presented here.

There are three major weaknesses in this article. The way it is written, especially at the beginning, is extremely confusing. First, I would suggest authors should check and review extensively for improvements to the use of English. In particular, the abstract and introduction are extremely hard to understand. Second, in the abstract and introduction, the authors use terms such as "hardening", "perturbing the type/strength of interactions", "stabilization", and "material properties", for just citing some terms. It is clear that the authors do know exactly what they are referring to, but the definitions come so late in the text that it all becomes confusing. The second major weakness is that there is a lack of deep discussion of the physical meaning of some of the measured parameters like "C dense vs inclusion", and "nuclear density and supersaturation". There is a need to explain further the physical consequences of all the graphs. Most of them are discussed in a very superficial manner. The third major weakness is a lack of analysis of phase separations. Some of their data suggest phase transition and/or phase separation, thus, a more in-deep analysis is required. For example, could they calculate the change of entropy and enthalpy of some of these processes? Could they find some boundaries for these transitions between the "hard" (whatever that means) and the liquid?

The authors have achieved almost all their goals, with the caveat of the third weakness I mentioned before. Their work presented in this article is of significant interest and can become extremely important if a more detailed analysis of the thermodynamics parameters is assessed and a better description of the physical phenomenon is provided.

We thank reviewer 1 for the comments and, in particular, for being so positive regarding the strengths of our manuscript and for raising concerns that will surely improve the manuscript. At this point, we propose the following actions to address the concerns of Reviewer 1:

1) We will extensively revise the use of English, particularly, in the abstract and introduction, defining key terms as they come along in the text to make the argument clearer.

2) We acknowledge the importance of discussing our data in more detail and we propose the following. We will discuss the graphs and what they mean as exemplified in the paragraph below.

Regarding Figure 3 - As the concentration of vRNPs increases, we observe an increase in supersaturation until 12hpi. This means that contrary to what is observed in a binary mixture, in which the Cdilute is constant (Klosin et al., 2020), the Cdilute in our system increases with concentration. It has been reported that Cdilute increases in a multi-component system with bulk concentration (Riback et al., 2020). Our findings have important implications for how we think about the condensates formed during influenza infection. As the 8 different genomic vRNPs have a similar overall structure, they could, in theory, behave as a binary system between units of vRNPs and Rab11a. However, a change in Cdilute with concentration shows that our system behaves as a multi-component system. This means that the differences in length, RNA sequence and valency that each vRNP have are key for the integrity of condensates.

3) The reviewer calls our attention to the lack of analysis of phase separations. We think that phase separation (or percolation coupled to phase separation) governs the formation of influenza A virus condensates. However, we think we ought to exert caution at this point as the condensates we are working with are very complex and that the physics of our system in cells may not be sufficient to claim phase separation without an in vitro reconstitution system. In fact, IAV inclusions contain cellular membranes, different vRNPs and Rab11a. So far, we can only speculate that the liquid character of IAV inclusions may arise from a network of interacting vRNPs that bridge several cognate vRNP-Rab11 units on flexible membranes, similarly to what happens in phase separated vesicles in neurological synapses. However, the speculative model for our system, although being supported by correlative light and electron microscopy, currently lacks formal experimental validation.

For this reason, we thought of developing the current work as an alternative to explore the importance of the liquid material properties of IAV inclusions. By finding an efficient method to alter the material properties of IAV inclusions, we provide proof of principle that it is possible to impose controlled phase transitions that reduce the dynamics of vRNPs in cells and negatively impact progeny virion production. Despite having discussed these issues in the limitations of the study, we will make our point clearer.

We are currently establishing an in vitro reconstitution system to formally demonstrate, in an independent publication, that IAV inclusions are formed by phase separation. For this future work, we teamed up with Pablo Sartori, a theorical physicist to derive in- depth analysis of the thermodynamics of the viral liquid condensates. Collectively, we think that cells have too many variables to derive meaningful physics parameters (such as entropy and enthalpy) as well as models and need to be complemented by in vitro systems. For example, increasing the concentration inside a cell is not a simple endeavour as it relies on cellular pathways to deliver material to a specific place. At the same time, the 8 vRNPs, as mentioned above, have different size, valency and RNA sequence and can behave very differently in the formation of condensates and maintenance of their material properties. Ideally, they should be analysed individually or in selected combinations. For the future, we will combine data from in vitro reconstitution systems and cells to address this very important point raised by the reviewer.

From the paper on the section Limitations of the study: “Understanding condensate biology in living cells is physiologically relevant but complex because the systems are heterotypic and away from equilibria. This is especially challenging for influenza A liquid inclusions that are formed by 8 different vRNP complexes, which although sharing the same structure, vary in length, valency, and RNA sequence. In addition, liquid inclusions result from an incompletely understood interactome where vRNPs engage in multiple and distinct intersegment interactions bridging cognate vRNP-Rab11 units on flexible membranes (Chou et al., 2013; Gavazzi et al., 2013; Haralampiev et al., 2020; Le Sage et al., 2020; Shafiuddin & Boon, 2019; Sugita, Sagara, Noda, & Kawaoka, 2013). At present, we lack an in vitro reconstitution system to understand the underlying mechanism governing demixing of vRNP-Rab11a-host membranes from the cytosol. This in vitro system would be useful to explore how the different segments independently modulate the material properties of inclusions, explore if condensates are sites of IAV genome assembly, determine thermodynamic values, thresholds accurately, perform rheological measurements for viscosity and elasticity and validate our findings”.

Reviewer #2 (Public Review):

During Influenza virus infection, newly synthesized viral ribonucleoproteins (vRNPs) form cytosolic condensates, postulated as viral genome assembly sites and having liquid properties. vRNP accumulation in liquid viral inclusions requires its association with the cellular protein Rab11a directly via the viral polymerase subunit PB2. Etibor et al. investigate and compare the contributions of entropy, concentration, and valency/strength/type of interactions, on the properties of the vRNP condensates. For this, they subjected infected cells to the following perturbations: temperature variation (4, 37, and 42{degree sign}C), the concentration of viral inclusion drivers (vRNPs and Rab11a), and the number or strength of interactions between vRNPs using nucleozin a well-characterized vRNP sticker. Lowering the temperature (i.e. decreasing the entropic contribution) leads to a mild growth of condensates that does not significantly impact their stability. Altering the concentration of drivers of IAV inclusions impact their size but not their material properties. The most spectacular effect on condensates was observed using nucleozin. The drug dramatically stabilizes vRNP inclusions acting as a condensate hardener. Using a mouse model of influenza infection, the authors provide evidence that the activity of nucleozin is retained in vivo. Finally, using a mass spectrometry approach, they show that the drug affects vRNP solubility in a Rab11a-dependent manner without altering the host proteome profile.

The data are compelling and support the idea that drugs that affect the material properties of viral condensates could constitute a new family of antiviral molecules as already described for the respiratory syncytial virus (Risso Ballester et al. Nature. 2021).

Nevertheless, there are some limitations in the study. Several of them are mentioned in a dedicated paragraph at the end of a discussion. This includes the heterogeneity of the system (vRNP of different sizes, interactions between viral and cellular partners far from being understood), which is far from equilibrium, and the absence of minimal in vitro systems that would be useful to further characterize the thermodynamic and the material properties of the condensates.

We thank reviewer 2 for highlighting specific details that need improving and raising such interesting questions to validate our findings. We will address all the minor comments of Reviewer 2. To address the comments of Reviewer 2, we propose the actions described in blue below each point raised that is written in italics.

1) The concentrations are mostly evaluated using antibodies. This may be correct for Cdilute. However, measurement of Cdense should be viewed with caution as the antibodies may have some difficulty accessing the inner of the condensates (as already shown in other systems), and this access may depend on some condensate properties (which may evolve along the infection). This might induce artifactual trends in some graphs (as seen in panel 2c), which could, in turn, affect the calculation of some thermodynamic parameters.

The concern of using antibodies to calculate Cdense is valid. We will address this concern by validating our results using a fluorescent tagged virus that has mNeon Green fused to the viral polymerase PA (PA-mNeonGreen PR8 virus). Like NP, PA is a component of vRNPs and labels viral inclusions, colocalising with Rab11 when vRNPs are in the cytosol without the need of using antibodies.

This virus would be the best to evaluate inclusion thermodynamics, where it not an attenuated virus (Figure 1A below) with a delayed infection as demonstrated by the reduced levels of viral proteins (Figure 1B below). Consistently, it shows differences in the accumulation of vRNPs in the cytosol and viral inclusions form later in infection. After their emergence, inclusions behave as in the wild-type virus (PR8-WT), fusing and dividing (Figure 1C below) and displaying liquid properties. The differences in concentration may shift or alter thermodynamic parameters such as time of nucleation, nucleation density, inclusion maturation rate, Cdense, Cdilute. This is the reason why we performed the thermodynamics profiling using antibodies upon PR8-WT infection. For validating our results, and taking into account a possible delayed kinetics, and differenced that may occur because of reduced vRNP accumulation in the cytosol, this virus will be useful and therefore we will repeat the thermodynamics using it.

As a side note, vRNPs are composed of viral RNA coated with several molecules of NP and each vRNP also contains 1 copy of the trimeric RNA dependent RNA polymerase formed by PA, PB1 and PB2. It is well documented that in the cytosol the vast majority of PA (and other components of the polymerase) is in the form of vRNPs (Avilov, Moisy, Munier, et al., 2012; Avilov, Moisy, Naffakh, & Cusack, 2012; Bhagwat et al., 2020; Lakdawala et al., 2014), and thus we can use this virus to label vRNPs on condensates to corroborate our studies using antibodies.

Figure 1 – The PA- mNeonGreen virus is attenuated in comparison to the WT virus. A. Cells (A549) were infected or mock-infected with PR8 WT or PA- mNeonGreen (PA-mNG) viruses, at a multiplicity of infection (MOI) of 3, for the indicated times. Viral production was determined by plaque assay and plotted as plaque forming units (PFU) per milliliter (mL) ± standard error of the mean (SEM). Data are a pool from 2 independent experiments. B. The levels of viral PA, NP and M2 proteins and actin in cell lysates at the indicated time points were determined by western blotting. C. Cells (A549) were transfected with a plasmid encoding mCherry-NP and co-infected with PA-mNeonGreen virus for 16h, at an MOI of 10. Cells were imaged under time-lapse conditions starting at 16 hpi. White boxes highlight vRNPs/viral inclusions in the cytoplasm in the individual frames. The dashed white and yellow lines mark the cell nucleus and the cell periphery, respectively. The yellow arrows indicate the fission/fusion events and movement of vRNPs/ viral inclusions. Bar = 10 µm. Bar in insets = 2 µm.

2) Although the authors have demonstrated that vRNP condensates exhibit several key characteristics of liquid condensates (they fuse and divide, they dissolve upon hypotonic shock or upon incubation with 1,6-hexanediol, FRAP experiments are consistent with a liquid nature), their aspect ratio (with a median above 1.4) is much higher than the aspect ratio observed for other cellular or viral liquid compartments. This is intriguing and might be discussed.

IAV inclusions have been shown to interact with microtubules and the endoplasmic reticulum, that confers movement, and also undergo fusion and fission events. We propose that these interactions and movement impose strength and deform inclusions making them less spherical. To validate this assumption, we compared the aspect ratio of viral inclusions in the absence and presence of nocodazole (that abrogates microtubule-based movement). The data in figure 2 shows that in the presence of nocodazole, the aspect ratio decreases from 1.42±0.36 to 1.26 ±0.17, supporting our assumption.

Figure 2 – Treatment with nocodazole reduces the aspect ratio of influenza A virus inclusions. Cells (A549) were infected PR8 WT and treated with nocodazole (10 µg/mL) for 2h time after which the movement of influenza A virus inclusions was captured by live cell imaging. Viral inclusions were segmented, and the aspect ratio measured by imageJ, analysed and plotted in R.

3) Similarly, the fusion event presented at the bottom of figure 3I is dubious. It might as well be an aggregation of condensates without fusion.

We will change this, thank you for the suggestion.

4) The authors could have more systematically performed FRAP/FLAPh experiments on cells expressing fluorescent versions of both NP and Rab11a to investigate the influence of condensate size, time after infection, or global concentrations of Rab11a in the cell (using the total fluorescence of overexpressed GFP-Rab11a as a proxy) on condensate properties.

We will try our best to be able to comply with this suggestion as we think it is important.

Reviewer #3 (Public Review):

This study aims to define the factors that regulate the material properties of the viral inclusion bodies of influenza A virus (IAV). In a cellular model, it shows that the material properties were not affected by lowering the temperature nor by altering the concentration of the factors that drive their formation. Impressively, the study shows that IAV inclusions may be hardened by targeting vRNP interactions via the known pharmacological modulator (also an IAV antiviral), nucleozin, both in vitro and in vivo. The study employs current state-of-the-art methodology in both influenza virology and condensate biology, and the conclusions are well-supported by data and proper data analysis. This study is an important starting point for understanding how to pharmacologically modulate the material properties of IAV viral inclusion bodies.

We thank this reviewer for all the positive comments. We will address the minor issues brought to our attention entirely, including changing the tittle of the manuscript and we will investigate the formation and material properties of IAV inclusions in the presence and absence of nucleozin for the nucleozin escape mutant NP-Y289H.

References

Avilov, S. V., Moisy, D., Munier, S., Schraidt, O., Naffakh, N., & Cusack, S. (2012). Replication- competent influenza A virus that encodes a split-green fluorescent protein-tagged PB2 polymerase subunit allows live-cell imaging of the virus life cycle. J Virol, 86(3), 1433- 1448. doi:10.1128/JVI.05820-11

Avilov, S. V., Moisy, D., Naffakh, N., & Cusack, S. (2012). Influenza A virus progeny vRNP trafficking in live infected cells studied with the virus-encoded fluorescently tagged PB2 protein. Vaccine, 30(51), 7411-7417. doi:10.1016/j.vaccine.2012.09.077

Bhagwat, A. R., Le Sage, V., Nturibi, E., Kulej, K., Jones, J., Guo, M., . . . Lakdawala, S. S. (2020). Quantitative live cell imaging reveals influenza virus manipulation of Rab11A transport through reduced dynein association. Nat Commun, 11(1), 23. doi:10.1038/s41467-019-13838-3

Chou, Y. Y., Heaton, N. S., Gao, Q., Palese, P., Singer, R. H., & Lionnet, T. (2013). Colocalization of different influenza viral RNA segments in the cytoplasm before viral budding as shown by single-molecule sensitivity FISH analysis. PLoS Pathog, 9(5), e1003358. doi:10.1371/journal.ppat.1003358

Gavazzi, C., Yver, M., Isel, C., Smyth, R. P., Rosa-Calatrava, M., Lina, B., . . . Marquet, R. (2013). A functional sequence-specific interaction between influenza A virus genomic RNA segments. Proc Natl Acad Sci U S A, 110(41), 16604-16609. doi:10.1073/pnas.1314419110

Haralampiev, I., Prisner, S., Nitzan, M., Schade, M., Jolmes, F., Schreiber, M., . . . Herrmann, A. (2020). Selective flexible packaging pathways of the segmented genome of influenza A virus. Nat Commun, 11(1), 4355. doi:10.1038/s41467-020-18108-1

Klosin, A., Oltsch, F., Harmon, T., Honigmann, A., Julicher, F., Hyman, A. A., & Zechner, C. (2020). Phase separation provides a mechanism to reduce noise in cells. Science, 367(6476), 464-468. doi:10.1126/science.aav6691

Lakdawala, S. S., Wu, Y., Wawrzusin, P., Kabat, J., Broadbent, A. J., Lamirande, E. W., . . . Subbarao, K. (2014). Influenza a virus assembly intermediates fuse in the cytoplasm. PLoS Pathog, 10(3), e1003971. doi:10.1371/journal.ppat.1003971

Le Sage, V., Kanarek, J. P., Snyder, D. J., Cooper, V. S., Lakdawala, S. S., & Lee, N. (2020). Mapping of Influenza Virus RNA-RNA Interactions Reveals a Flexible Network. Cell Rep, 31(13), 107823. doi:10.1016/j.celrep.2020.107823

Riback, J. A., Zhu, L., Ferrolino, M. C., Tolbert, M., Mitrea, D. M., Sanders, D. W., . . . Brangwynne, C. P. (2020). Composition-dependent thermodynamics of intracellular phase separation. Nature, 581(7807), 209-214. doi:10.1038/s41586-020-2256-2

Shafiuddin, M., & Boon, A. C. M. (2019). RNA Sequence Features Are at the Core of Influenza a Virus Genome Packaging. J Mol Biol. doi:10.1016/j.jmb.2019.03.018

Sugita, Y., Sagara, H., Noda, T., & Kawaoka, Y. (2013). Configuration of viral ribonucleoprotein complexes within the influenza A virion. J Virol, 87(23), 12879- 12884. doi:10.1128/JVI.02096-13

Author Response

Reviewer #1 (Public Review):

The authors have tried to correlate changes in the cellular environment by means of altering temperature, the expression of key cellular factors involved in the viral replication cycle, and small molecules known to affect key viral protein-protein interactions with some physical properties of the liquid condensates of viral origin. The ideas and experiments are extremely interesting as they provide a framework to study viral replication and assembly from a thermodynamic point of view in live cells.

The major strengths of this article are the extremely thoughtful and detailed experimental approach; although this data collection and analysis are most likely extremely time-consuming, the techniques used here are so simple that the main goal and idea of the article become elegant. A second major strength is that in other to understand some of the physicochemical properties of the viral liquid inclusion, they used stimuli that have been very well studied, and thus one can really focus on a relatively easy interpretation of most of the data presented here.

There are three major weaknesses in this article. The way it is written, especially at the beginning, is extremely confusing. First, I would suggest authors should check and review extensively for improvements to the use of English. In particular, the abstract and introduction are extremely hard to understand. Second, in the abstract and introduction, the authors use terms such as "hardening", "perturbing the type/strength of interactions", "stabilization", and "material properties", for just citing some terms. It is clear that the authors do know exactly what they are referring to, but the definitions come so late in the text that it all becomes confusing. The second major weakness is that there is a lack of deep discussion of the physical meaning of some of the measured parameters like "C dense vs inclusion", and "nuclear density and supersaturation". There is a need to explain further the physical consequences of all the graphs. Most of them are discussed in a very superficial manner. The third major weakness is a lack of analysis of phase separations. Some of their data suggest phase transition and/or phase separation, thus, a more in-deep analysis is required. For example, could they calculate the change of entropy and enthalpy of some of these processes? Could they find some boundaries for these transitions between the "hard" (whatever that means) and the liquid?

The authors have achieved almost all their goals, with the caveat of the third weakness I mentioned before. Their work presented in this article is of significant interest and can become extremely important if a more detailed analysis of the thermodynamics parameters is assessed and a better description of the physical phenomenon is provided.

We thank you for the comments and, in particular, for being so positive regarding the strengths of our manuscript and for raising concerns that will surely improve it. We have taken the following actions to address your concerns:

1) Extensive revisions have been made to the use of English, particularly in the abstract and introduction. Key terms are defined as they are introduced in the text to enhance the clarity of the argument. This is a significant revision that is highlighted within the text, but it is too extensive to detail here.

2) In the results section, we improved and extended the discussion of our graphs to the extent possible. However, we found that attempting to explain the graphs' meanings more thoroughly would detract from our manuscript's main focus: identifying thermodynamic changes that could potentially lead to alterations in material properties, specifically aspect ratio, size, and Gibbs free energy. As a result, we introduced the type of information we could obtain from our analyses in the introduction (Lines 112-125) and briefly commented on it in the ‘results’ section (Lines 304-306, sentences below).

From introduction – lines 112-125:

“In addition, other parameters like nucleation density determine how many viral condensates are formed per area of cytosol. Overall, the data will inform us if changing one parameter, e.g. the concentration, drives the system towards larger condensates with the same or more stable properties, or more abundant condensates that are forced to maintain the initial or a different size on account of available nucleation centres (Riback et al., 2020:Snead, 2022 #1152). It will also inform us if liquid viral inclusions behave like a binary or a multi-component system. In a binary mixture, Cdilute is constant (Klosin et al., 2020). However, in multi-component systems, Cdilute increases with bulk concentration (Riback et al., 2020). This type of information could have direct implications about the condensates formed during influenza infection. As the 8 different genomic vRNPs have a similar overall structure, they could, in theory, behave as a binary system between units of vRNPs and Rab11a. However, a change in Cdilute with concentration would mean that the system behaves as a multi-component system. This could raise the hypothesis that the differences in length, RNA sequence and valency that each vRNP has may be relevant for the integrity and behaviour of condensates.”.

From results lines 304-306:

This indicates that the liquid inclusions behave as a multi-component system and allow us to speculate that the differences in length, RNA sequence and valency that each vRNP may be key for the integrity and behaviour of condensates.

3) The reviewer has drawn our attention to the absence of phase separation analysis in our study. We believe that the formation of influenza A virus condensates is governed by phase separation (or percolation coupled to phase separation). However, we must exercise caution at this point because the condensates we are studying are highly complex, and the physics of our cellular system may not be adequate to claim phase separation without being validated by an in vitro reconstitution system. IAV inclusions contain a variety of cellular membranes, different vRNPs, and Rab11a. While we have robust data to propose a model in which the liquid-like properties of IAV inclusions arise from a network of interacting vRNPs that bridge multiple cognate vRNP-Rab11 units on flexible membranes, similar to what occurs in phase-separated vesicles in neurological synapses, our model for this system still lacks formal experimental validation. As a note, the data supporting our model includes: the demonstration of the liquid properties of our liquid inclusions (Alenquer et al. 2019, Nature Communications, 10, 1629); and impairment of recycling endocytic activity during IAV infection Bhagwat et al. 2020, Nat Commun, 11, 23; Kawaguchi et al. 2012, J Virol, 86, 11086-95; Vale-costa et al. 2016, J Cell Sci, 129, 1697-710. This leads to aggregated vesicles seen by correlative light and electron microscopy (Vale-Costa et al., 2016 JCS, 129, 1697-710) and by immunofluorescence and FISH (Amorim et al. 2011,. J Virol 85, 4143-4156; Avilov et al. 2012, Vaccine 30, 7411-7417; Chou et al. 2013, PLoS Pathog 9, e1003358; Eisfeld et al. 2011, J Virol 85, 6117-6126 and Lakdawala et al. 2014, PLoS Pathog 10, e1003971.

To be able to explore the significance of the liquid material properties of IAV inclusions, we used the strategy described in this current work. By developing an effective method to manipulate the material properties of IAV inclusions, we provide evidence that controlled phase transitions can be induced, resulting in decreased vRNP dynamics in cells and a negative impact on progeny virion production. This suggests that the liquid character of liquid inclusions is important for their function in IAV infection. We have improved our explanation addressing this concern in the limitations of our study (as outlined below in the box and in manuscript in lines 857-872).

We are currently establishing an in vitro reconstitution system to formally demonstrate, in an independent publication, that IAV inclusions are formed by phase separation (or percolation coupled to phase separation). For this future work, we teamed up with Pablo Sartori, a theorical physicist to derive in-depth analysis of the thermodynamics of the viral liquid condensates in the in vitro reconstituted system and compare it to results obtained in the cell. This will provide means to establish comparisons. We think that cells have too many variables to derive meaningful physics parameters (such as entropy and enthalpy) and models that need to be complemented by in vitro systems. For example, increasing the concentration inside a cell is not a simple endeavour as it relies on cellular pathways to deliver material to a specific place. At the same time, the 8 vRNPs, as mentioned above, have different size, valency and RNA sequence and can behave very differently in the formation of condensates and maintenance of their material properties. Ideally, they should be analysed individually or in selected combinations. For the future, we will combine data from in vitro reconstitution systems and cells to address this very important point raised by the reviewer.

From the paper on the section ‘Limitations of the study’:

“Understanding condensate biology in living cells is physiological relevant but complex because the systems are heterotypic and away from equilibria. This is especially challenging for influenza A liquid inclusions that are formed by 8 different vRNP complexes, which although sharing the same structure, vary in length, valency, and RNA sequence. In addition, liquid inclusions result from an incompletely understood interactome where vRNPs engage in multiple and distinct intersegment interactions bridging cognate vRNP-Rab11 units on flexible membranes (Chou et al., 2013, Gavazzi et al., 2013, Sugita et al., 2013, Shafiuddin and Boon, 2019, Haralampiev et al., 2020, Le Sage et al., 2020). At present, we lack an in vitro reconstitution system to understand the underlying mechanism governing demixing of vRNP-Rab11a-host membranes from the cytosol. This in vitro system would be useful to explore how the different segments independently modulate the material properties of inclusions, explore if condensates are sites of IAV genome assembly, determine thermodynamic values, thresholds accurately, perform rheological measurements for viscosity and elasticity and validate our findings. The results could be compared to those obtained in cell systems to derive thermodynamic principles happening in a complex system away from equilibrium. Using cells to map how liquid inclusions respond to different perturbations provide the answer of how the system adapts in vivo, but has limitations.

Reviewer #2 (Public Review):

During Influenza virus infection, newly synthesized viral ribonucleoproteins (vRNPs) form cytosolic condensates, postulated as viral genome assembly sites and having liquid properties. vRNP accumulation in liquid viral inclusions requires its association with the cellular protein Rab11a directly via the viral polymerase subunit PB2. Etibor et al. investigate and compare the contributions of entropy, concentration, and valency/strength/type of interactions, on the properties of the vRNP condensates. For this, they subjected infected cells to the following perturbations: temperature variation (4, 37, and 42{degree sign}C), the concentration of viral inclusion drivers (vRNPs and Rab11a), and the number or strength of interactions between vRNPs using nucleozin a well-characterized vRNP sticker. Lowering the temperature (i.e. decreasing the entropic contribution) leads to a mild growth of condensates that does not significantly impact their stability. Altering the concentration of drivers of IAV inclusions impact their size but not their material properties. The most spectacular effect on condensates was observed using nucleozin. The drug dramatically stabilizes vRNP inclusions acting as a condensate hardener. Using a mouse model of influenza infection, the authors provide evidence that the activity of nucleozin is retained in vivo. Finally, using a mass spectrometry approach, they show that the drug affects vRNP solubility in a Rab11a-dependent manner without altering the host proteome profile

The data are compelling and support the idea that drugs that affect the material properties of viral condensates could constitute a new family of antiviral molecules as already described for the respiratory syncytial virus (Risso Ballester et al. Nature. 2021)

Nevertheless, there are some limitations in the study. Several of them are mentioned in a dedicated paragraph at the end of a discussion. This includes the heterogeneity of the system (vRNP of different sizes, interactions between viral and cellular partners far from being understood), which is far from equilibrium, and the absence of minimal in vitro systems that would be useful to further characterize the thermodynamic and the material properties of the condensates.

There are other ones.

We thank reviewer 2 for highlighting specific details that need improving and raising such interesting questions to validate our findings. We have addressed the comments of Reviewer 2, we performed the experiments as described (in blue) below each point raised.

1) The concentrations are mostly evaluated using antibodies. This may be correct for Cdilute. However, measurement of Cdense should be viewed with caution as the antibodies may have some difficulty accessing the inner of the condensates (as already shown in other systems), and this access may depend on some condensate properties (which may evolve along the infection). This might induce artifactual trends in some graphs (as seen in panel 2c), which could, in turn, affect the calculation of some thermodynamic parameters.

The concern of using antibodies to calculate Cdense is valid, and we thought it was very important. We addressed this concern by performing the same analyses using a fluorescent tagged virus that has mNeon Green fused to the viral polymerase PA (PA-mNeonGreen PR8 virus). Like NP, PA is a component of vRNPs and labels viral inclusions, colocalising with Rab11 when vRNPs are in the cytosol. However, per vRNP there is only one molecule of PA, whilst of NP there are 37-96 depending on the size of vRNPs. As predicted, we did observe changes in the Cdilute, Cdense and nucleation density. However, the measurements and values obtained for Gibbs free energy, size, aspect ratio detecting viral inclusions with fluorescently tagged vRNPs or antibody staining followed the same trend and allow us to validate our conclusion that major changes in Gibbs free energy occur solely when there is a change in the valency/strength of interactions but not in temperature or concentration (Figure 1 below). Given the extent of these data, we show here the results but, in the manuscript, we will describe the limitations of using antibodies in our study within the section ‘Limitations of the study’ from lines 881-894. Given the importance of the question regarding the pros and cons of the different systems for analysing thermodynamic parameters, we have decided to systematically assess and explore these differences in detail in a future manuscript.

For more information. This reviewer may be asking why we did not use the PA-fluorescent virus in the first place to evaluate inclusion thermodynamics and avoid problems in accessibility that antibodies may have to get deep into large inclusions. Our answer is that no system is perfect. In the case of the PA-fluorescent virus, the caveats revolve around the fact that the virus is attenuated (Figure 1a below), exhibiting a delayed infection as demonstrated by reduced levels of viral proteins (Figure 1b below). Consistently, it shows differences in the accumulation of vRNPs in the cytosol and viral inclusions form later in infection and the amount of vRNPs in the cytosol does not reach the levels observed in PR8-WT virus. After their emergence, inclusions behave as in the wild-type virus (PR8-WT), fusing and dividing (Figure 1c below) and displaying liquid properties.

As the overarching goal of this manuscript is to evaluate the best strategies to harden liquid IAV inclusions and given that one of the parameters we were testing is concentration, we reasoned that using PR8-WT virus for our analyses would be reasonable.

In conclusions, both systems have caveats that are important to systematically assess, and these differences may shift or alter thermodynamic parameters such as nucleation density, inclusion maturation rate, Cdense, Cdilute in particular by varying the total concentration. As a note, to validate all our results using the PA-mNeonGreen PR8 virus, we considered the delayed kinetics and applied our thermodynamic analyses up to 20 hpi rather than 16 hpi.

However, because of the question raised by this reviewer, on which is the best solution for mitigating errors induced by using antibodies, we re-checked all our data. Not only have we compared the data originated from attenuated fluorescently tagged virus with our data, but also made comparisons with images acquired from Z stacks (as used for concentration and for type/strength of interactions) with those acquired from 2D images. Our analysis revealed that there is a very good match using images acquired with Z-stacks and analysed as Z projections with between antibody staining and vRNP fluorescent virus. Therefore, we re-analysed all our thermodynamic data done with temperature using images acquired from Z stacks and altered entirely Figure 2. We believe that all these comparisons and analyses have greatly improved the manuscript and hence we thank all reviewers for their input.

Figure 1 – The PA-mNeonGreen virus is attenuated in comparison to the WT virus and data obtained is consistent for Gibbs free energy with analyses done with images processed with antibody fluorescent vRNPs. A. Representation of the PA-mNeonGreen virus (PA-mNG; Abbreviations: NCR: non coding region). B. Cells (A549) were transfected with a plasmid encoding mCherry-NP and co-infected with PA-mNeonGreen virus for 16h, at an MOI of 10. Cells were imaged under time-lapse conditions starting at 16 hpi. White boxes highlight vRNPs/viral inclusions in the cytoplasm in the individual frames. The dashed white and yellow lines mark the cell nucleus and the cell periphery, respectively. The yellow arrows indicate the fission/fusion events and movement of vRNPs/ viral inclusions. Bar = 10 µm. Bar in insets = 2 µm. C-D. Cells (A549) were infected or mock-infected with PR8 WT or PA-mNG viruses, at a multiplicity of infection (MOI) of 3, for the indicated times. C. Viral production was determined by plaque assay and plotted as plaque forming units (PFU) per milliliter (mL) ± standard error of the mean (SEM). Data are a pool from 2 independent experiments. D. The levels of viral PA, NP and M2 proteins and actin in cell lysates at the indicated time points were determined by western blotting. (E-G) Biophysical calculations in cells infected with the PA-mNeonGreen virus upon altering temperature (at 10 hpi, evaluating the concentration of vRNPs (over a time course) in conditions expressing native amounts of Rab11a or overexpressing low levels of Rab11a and upon altering the type/strength of vRNP interactions by adding nucleozin at 10 hpi during the indicated time periods. All data: Ccytoplasm/Cnucleus; Cdense, Cdilute, area aspect ratio and Gibbs free energy are represented as boxplots. Above each boxplot, same letters indicate no significant difference between them, while different letters indicate a statistical significance at α = 0.05 using one-way ANOVA, followed by Tukey multiple comparisons of means for parametric analysis, or Kruskal-Wallis Bonferroni treatment for non-parametric analysis.

2) Although the authors have demonstrated that vRNP condensates exhibit several key characteristics of liquid condensates (they fuse and divide, they dissolve upon hypotonic shock or upon incubation with 1,6-hexanediol, FRAP experiments are consistent with a liquid nature), their aspect ratio (with a median above 1.4) is much higher than the aspect ratio observed for other cellular or viral liquid compartments. This is intriguing and might be discussed.

IAV inclusions have been shown to interact with microtubules and the endoplasmic reticulum, that confers movement, and undergo fusion and fission events. We propose that these interactions and movement impose strength and deform inclusions making them less spherical. To validate this assumption, we compared the aspect ratio of viral inclusions in the absence and presence of nocodazole (that abrogates microtubule-based movement). The data in figure 2 shows that in the presence of nocodazole, the aspect ratio decreases from 1.42±0.36 to 1.26 ±0.17, supporting our assumption.

Figure 2 – Treatment with nocodazole reduces the aspect ratio of influenza A virus inclusions. Cells (A549) were infected with PR8 WT for 8 h and treated with nocodazole (10 µg/mL) for 2h, after which the movement of influenza A virus inclusions was captured by live cell imaging. Viral inclusions were segmented, and the aspect ratio measured by imageJ, analysed and plotted in R.

3) Similarly, the fusion event presented at the bottom of figure 3I is dubious. It might as well be an aggregation of condensates without fusion.

We have changed this (check Fig 5A and B in the manuscript), thank you for the suggestion.

4) The authors could have more systematically performed FRAP/FLAPh experiments on cells expressing fluorescent versions of both NP and Rab11a to investigate the influence of condensate size, time after infection, or global concentrations of Rab11a in the cell (using the total fluorescence of overexpressed GFP-Rab11a as a proxy) on condensate properties.

We have included a new figure, figure 5 with the suggested data.

Author Response

Reviewer #2 (Public Review):

1) The main limitation of this study is that the results are primarily descriptive in nature, and thus, do not provide mechanistic insight into how Ryr1 disease mutations lead to the muscle-specific changes observed in the EDL, soleus and EOM proteomes.

An intrinsic feature of the high-throughput proteomic analysis technology is the generation of lists of differentially expressed proteins (DEP) in different muscles from WT and mutated mice. Although the definition of mechanistic insights related to changes of dozens of proteins is very interesting, it is a difficult task to accomplish and goes beyond the goal of the high-throughput proteomic analysis presented here. Nevertheless, the analysis of DEPs may indeed provide arguments to speculate on the pathogenesis of the phenotype linked to recessive RyR1 mutations. In the unrevised manuscript, we pointed out that the fiber type I predominance observed in congenital myopathies linked to recessive Ryr1 mutation are consistent with the high expression level of heat shock proteins in slow twitch muscles. However, as suggested by Reviewer 3, we have removed "vague statements" from the text of the revised manuscript, concerning major insights into pathophysiological mechanisms, since we are aware that the mechanistic information, if any, that we can extract from the data set, cannot go over the intrinsic limitation of the high-throughput proteomic technology.

b) Results comparing fast twitch (EDL) and slow twitch (soleus) muscles from WT mice confirmed several known differences between the two muscle types. Similar analyses between EOM/EDL and EOM/soleus muscles from WT mice were not conducted.

We agree with the point raised by the Reviewer. In the revised manuscript we have changed Figure 2. The new Figure 2 shows the analysis of differentially expressed proteins in EDL, soleus and EOMs from WT mice. We have also added 2 new Tables (new Supplementary Table 2 and 3) and have inserted our findings in the revised Results section (page, 7, lines 157-176, pages 8 and 9).

c) While a reactome pathway analysis for proteins changes observed in EDL is shown in Supplemental Figure 1, the authors do not fully discuss the nature of the proteins and corresponding pathways impacted in the other two muscle groups analyzed.

We have now included in the revised manuscript a new Figure 2 which includes the Reactome pathway analysis comparing EDL with soleus, EDL with EOM and soleus with EOM (panels C, F and I, respectively). We have also inserted into the revised manuscript a brief description of the pathways showing the greatest changes in protein content (page 7 line 156-175, pages 8 and 9). We agree that the data showing changes in protein content between the 3 muscle groups of the WT mice are important also because they validate the results of the proteomic approach. Indeed, the present results confirm that many proteins including MyHCIIb, calsequestrin 1, SERCA1, parvalbumin etc are more abundantly expressed in fast twitch EDL muscles compared to soleus. Similarly, our results confirm that EOMs are enriched in MyHC-EO as well as cardiac isoforms of ECC proteins. This point has been clarified in the revised version of the manuscript (page 8, lines 198-213; page 9 lines 214-228). Nevertheless, we would like to point out that the main focus of our study is to compare the changes of protein content induced by the presence of recessive RyR1 mutations.

Reviewer #3 (Public Review):

a) it would be useful to determine whether changes in protein levels correlated with changes in mRNA levels …….

We performed qPCR analysis of Stac3 and Cacna1s in EDL, Soleus and EOM from WT mice (see Figure 1 below). The expression of transcripts encoding Cacna1s and Stac3 is approximately 9-fold higher in EDL compared to Soleus. The fold change of Stac3 and Cacna1s transcripts in EDL muscles is higher compared to the differences we observed by Mass spectrometry at the protein level between EDL and Soleus. Indeed, we found that the content of the Stac3 protein in EDL is 3-fold higher compared to that in soleus. Although there is no apparent linear correlation between mRNA and protein levels, we believe that a few plausible conclusions can be drawn, namely: (i) the expression level of both transcripts and proteins is higher EDL compared to EOM and soleus muscles, respectively, (ii) the expression level of transcripts encoding Stac3 correlate with those encoding Cacan1s and confirm proteomic data. In addition, the level of Stac3 transcript does not changes between WT and dHT, confirming our proteomic data which show that Stac3 protein content in muscles from dHT is similar to that found in WT littermates. Altogether these results support the concept that the differences in Stac3 content between EDL and soleus occur at both the protein and transcript levels, namely high Stac3 mRNA level correlates with higher protein content (EDL) and low mRNA levels correlated with low Stac3 protein content in Soleus muscles (see Figure 1 below).

Figure 2: qPCR of Cacna1s and Stac3 in muscles from WT mice. The expression levels of the transcripts encoding Cacna1s and Stac3 are the highest in EDL muscles and the lowest in soleus muscles (top panels). There are no significant changes in their relative expression levels in dHT vs WT. Each symbol represents the value from of a single mouse. * p=0.028 Mann Whitney test qPCR was performed as described in Elbaz et al., 2019 (Hum Mol Genet 28, 2987-2999).

….and whether or not the protein present was functional, and whether Stac3 was in fact stoichiometrically depleted in relation to Cacna1s.

We thought about this point but think that there are no plausible arguments to believe that Stac3 is not functional, one simple reason being that our WT mice do not have a phenotype which would be associated with the absence of Stac3 (Reinholt et al., PLoS One 8, e62760 2013, Nelson et al. Proc. Natl. Acad. Sci. USA 110:11881 2013).

b) In the abstract, the authors stated that skeletal muscle is responsible for voluntary movement. It is also responsible for non-voluntary. The abstract needs to be refocused on the mutation and on what we learn from this study. Please avoid vague statements like "we provide important insights to the pathophysiological mechanisms..." mainly when the study is descriptive and not mechanistic.

The abstract of the revised manuscript has been rewritten. In particular, we removed statements referring to important “pathophysiological mechanistic insight”.

c) The author should bring up the mutation name, location and phenotype early in the introduction.

In the revised manuscript we provide the information requested by the Reviewer (page 2 lines 36-38 and page 4, lines 98-102).

d) This reviewer also suggests that the authors refocus the introduction on the mutation location in the 3D RyR1 structure (available cryo-EM structure), if there is any nearby ligand binding site, protomers junction or any other known interacting protein partners. This will help the reader to understand how this mutation could be important for the channel's function

The residue Ala4329 is present inside the TMx (Auxiliary transmembrane helices) domain which spans from residue 4322 to 4370 and interposes structurally (des Georges A et al. 2016 Cell 167,145-57; Chen W, et al. 2020 EMBO Rep. 21, e49891). Although the structural resolution of the region has been improved (des Georges et al, 2016), parts of the domain still remain with no defined atomic coordinates, especially the region encompassing a.a. E4253 – F4540. Because of such undefined atomic coordinates of the region E4253-F4540, we are not able to determine the real orientation and the disposition of the amino acids in this region, including the A4329 residue. As reference, structure PDB: 5TAL of des Georges et al, 2016 was analyzed with UCSF Chimera (production version 1.16) (Pettersen et al. J. Comput. Chem. 25: 1605-1612. doi: 10.1002/jcc.20084).

Author Response

Public Evaluation Summary:

The authors re-analyzed a previously published dataset and identify patterns suggestive of increased bacterial biodiversity in the gut may creating new niches that lead to gene loss in a focal species and promote generation of more diversity. Two limitations are (i) that sequencing depth may not be sufficient to analyze strain-level diversity and (ii) that the evidence is exclusively based on correlations, and the observed patterns could also be explained by other eco-evolutionary processes. The claims should be supported by a more detailed analysis, and alternative hypotheses that the results do not fully exclude should be discussed. Understanding drivers of diversity in natural microbial communities is an important question that is of central interest to biomedically oriented microbiome scientists, microbial ecologists and evolutionary biologists.

We agree that understanding the drivers of diversity in natural communities is an important and challenging question to address. We believe that our analysis of metagenomes from the gut microbiomes is complementary to controlled laboratory experiments and modeling studies. While these other studies are better able to establish causal relationships, we rely on correlations – a caveat which we make clear, and offer different mechanistic explanations for the patterns we observe.

We also mention the caveat that we are only able to measure sub-species genetic diversity in relatively abundant species with high sequencing depth in metagenomes. These relatively abundant species include dozens of species in two metagenomic datasets, and we see no reason why they would not generalize to other members of the microbiome. Nonetheless, further work will be required to extend our results to rarer species.

Our revised manuscript includes two major new analyses. First, we extend the analysis of within-species nucleotide diversity to non-synonymous sites, with generally similar results. This suggests that evolutionarily older, less selectively constrained synonymous mutations and more recent non-synonymous mutations that affect protein structure both track similarly with measures of community diversity – with some subtle differences described in the manuscript.

Second, we extend our analysis of dense time series data from one individual stool donor and one deeply covered species (B. vulgatus) to four donors and 15 species. This allowed us to reinforce the pattern of gene loss in more diverse communities with greater statistical support. Our correlational results are broadly consistent with the predictions of DBD from modeling and experimental studies, and they open up new lines of inquiry for microbiome scientists, ecologists, and evolutionary biologists.

Reviewer #1 (Public Review):

This paper makes an important contribution to the current debate on whether the diversity of a microbial community has a positive or negative effect on its own diversity at a later time point. In my view, the main contribution is linking the diversity-begets-diversity patterns, already observed by the same authors and others, to genomic signatures of gene loss that would be expected from the Black Queen Hypothesis, establishing an eco-evolutionary link. In addition, they test this hypothesis at a more fine-grained scale (strain-level variation and SNP) and do so in human microbiome data, which adds relevance from the biomedical standpoint. The paper is a well-written and rigorous analysis using state-of-the-art methods, and the results suggest multiple new experiments and testable hypotheses (see below), which is a very valuable contribution.

We thank the reviewer for their generous comments.

That being said, I do have some concerns that I believe should be addressed. First of all, I am wondering whether gene loss could also occur because of environmental selection that is independent of other organisms or the diversity of the community. An alternative hypothesis to the Black Queen is that there might have been a migration of new species from outside and then loss of genes could have occurred because of the nature of the abiotic environment in the new host, without relationship to the community diversity. Telling the difference between these two hypotheses is hard and would require extensive additional experiments, which I don't think is necessary. But I do think the authors should acknowledge and discuss this alternative possibility and adjust the wording of their claims accordingly.

We concur with the reviewer that the drivers of the correlation between community diversity and gene loss are unclear. Therefore, we have now added the following text to the Discussion:

“Here we report that genome reduction in the gut is higher in more diverse gut communities. This could be due to de novo gene loss, preferential establishment of migrant strains encoding fewer genes, or a combination of the two. The mechanisms underlying this correlation remain unclear and could be due to biotic interactions – including metabolic cross-feeding as posited by some models (Estrela et al., 2022; San Roman and Wagner, 2021, 2018) but not others (Good and Rosenfeld, 2022) – or due to unknown abiotic drivers of both community diversity and gene loss.”

Additionally, we have revised Figure 1 to show that strain invasions/replacements, in addition to evolutionary change, could be an important driver of changes in intra-species diversity in the microbiome.

Another issue is that gene loss is happening in some of the most abundant species in the gut. Under Black Queen though, we would expect these species to be most likely "donors" in cross-feeding interactions. Authors should also discuss the implications, limitations, and possible alternative hypotheses of this result, which I think also stimulates future work and experiments.

We thank the reviewer for raising this point. It is unclear to us whether the more abundant species would be donors in cross-feeding interactions. If we understand correctly, the reviewer is suggesting that more abundant donors will contribute more total biomass of shared metabolites to the community. This idea makes sense under the assumption that the abundant species are involved in cross-feeding interactions in the first place, which may or may not be the case. As our work heavily relies on a dataset that we previously analyzed (HMP), we wish to cite Figure S20 in Garud, Good et al. 2019 PLoS Biology in which we found there are comparable rates of gene changes across the ~30 most abundant species analyzed in the HMP. This suggests that among the most abundant species analyzed, there is no relationship between their abundance and gene change rate.

That being said, we acknowledge that our study is limited to the relatively abundant focal species and state now in the Discussion: “Deeper or more targeted sequencing may permit us to determine whether the same patterns hold for rarer members of the microbiome.”

Regarding Figure 5B, there is a couple of questions I believe the authors should clarify. First, How is it possible that many species have close to 0 pathways? Second, besides the overall negative correlation, the data shows some very conspicuous regularities, e.g. many different "lines" of points with identical linear negative slope but different intercept. My guess is that this is due to some constraints in the pathway detection methods, but I struggle to understand it. I think the authors should discuss these patterns more in detail.

We sincerely thank the reviewer for raising this issue, as it prompted us to investigate more deeply the patterns observed at the pathway level. In short, we decided to remove this analysis from the paper because of a number of bioinformatics issues that we realized were contributing to the signal. However, in support of BQH-like mechanisms at play, we do find evidence for gene loss in more diverse communities across multiple species in both the HMP and Poyet datasets. Below we detail our investigation into Figure 5b and how we arrived at the conclusion that is should be removed:

(1) Regarding data points in Figure 5B where many focal species have “zero pathways”,we firstly clarify how we compute pathway presence and richness. Pathway abundance data per species were downloaded from the HMP1-2 database, and these pathway abundances were computed using HUMAnN (HMP Unified Metabolic Analysis Network). According to HUMAnN documentation, pathway abundance is proportional to the number of complete copies of the pathway in the community; this means that if at least one component reaction in a certain pathway is missing coverage (for a sample-species pair), the pathway abundance may be zero (note that HUMAnN also employs “gap filling” to allow no more than one required reaction to have zero abundance). As such, it is likely that insufficient coverage, especially for low-abundance species, causes many pathways to report zero abundance in many species in many samples. Indeed, 556 of the 649 species considered had zero “present” pathways (i.e. having nonzero abundance) in at least 400 of the 469 samples (see figure below).

(2) We thank the reviewer for pointing out the “conspicuous regularities” in Figure 5B,particularly “parallel lines” of data points that we discovered are an artifact of the flawed way in which we computed “community pathway richness [excluding the focal species].” Each diagonal line of points corresponds to different species in the same sample, and because community pathway richness is computed as the total number of pathways [across all species in the sample] minus the number of pathways in the focal species, the current Figure 5B is really plotting y against X-y for each sample (where X is a sample’s total community pathway richness, and y is the pathway richness of an individual species in that sample). This computation fails to account for the possibility that a pathway in an excluded focal species will still be present in the community due to redundancy, and indeed BQH tests for whether this redundancy is kept low in diverse communities due to mechanisms such as gene loss.

We attempted to instead plot community pathway richness defined as the number of unique pathways covered by all species other than the focal species. This is equivalent to [number of unique pathways across all species in a sample] minus the [number of pathways that are ONLY present in the focal species and not any other species in the sample]. However, when we recomputed community pathway richness this way, it is rare that a pathway is present in only one species in a sample. Moreover, we find that with the exception of E. coli, focal species pathway richness tended to be very similar across the 469 samples, often reaching an upper limit of focal species pathway richness observed. (It is unclear to what extent lower pathway richnesses are due to low species abundance/low sample coverage versus gene loss). This new plot reveals even more regularities and is difficult to interpret with respect to BQH. (Note that points are colored by species; the cluster of black dots with outlying high focal pathway richness corresponds to the “unclassified” stratum which can be considered a group of many different species.)

Overall, because community pathway richness (excluding a focal species) seems to primarily vary with sample rather than focal species in this dataset when using the most simple/strict definition of community pathway richness as described above, it is difficult to probe the Black Queen Hypothesis using a plot like Figure 5B. As pointed out by reviewers, lack of sequencing depth to analyze strain-level diversity and accurately quantify pathway abundance, irrespective of species abundance, seems to be a major barrier to this analysis. As such, we have decided to remove Figure 5B from the paper and rewrite some of our conclusions accordingly.

Finally, I also have some conceptual concerns regarding the genomic analysis. Namely, genes can be used for biosynthesis of e.g. building blocks, but also for consumption of nutrients. Under the Black Queen Hypothesis, we would expect the adaptive loss of biosynthetic genes, as those nutrients become provided by the community. However, for catabolic genes or pathways, I would expect the opposite pattern, i.e. the gain of catabolic genes that would allow taking advantage of a more rich environment resulting from a more diverse community (or at least, the absence of pathway loss). These two opposing forces for catabolic and biosynthetic genes/pathways might obscure the trends if all genes are pooled together for the analysis. I believe this can be easily checked with the data the authors already have, and could allow the authors to discuss more in detail the functional implications of the trends they see and possibly even make a stronger case for their claims.

We thank the reviewer for their suggestion. As explained above, we have removed the pathway analysis from the paper due to technical reasons. However, we did investigate catabolic and biosynthetic pathways separately as suggested by the reviewer as we describe below:

We obtained subsets of biosynthetic pathways and catabolic pathways by searching for keywords (such as “degradation” for catabolic) in the MetaCyc pathway database. After excluding the “unclassified” species stratum, we observe a total of 279 biosynthetic and 167 catabolic pathways present in the HMP1-2 pathway abundance dataset. Using the corrected definition of community pathway richness excluding a focal species, for each pathway type—either biosynthetic or catabolic—we plotted focal species pathway richness against community pathway richness including all pathways regardless of type:

We observe the same problem where, within a sample, community pathway richness excluding the focal species hardly varies no matter which focal species it is, due to nearly all of its detected pathways being present in at least one other species; this makes the plots difficult to interpret.

Reviewer #2 (Public Review):

The authors re-analysed two previously published metagenomic datasets to test how diversity at the community level is associated with diversity at the strain level in the human gut microbiota. The overall idea was to test if the observed patterns would be in agreement with the "diversity begets diversity" (DBD) model, which states that more diversity creates more niches and thereby promotes further increase of diversity (here measured at the strain-level). The authors have previously shown evidence for DBD in microbiomes using a similar approach but focusing on 16S rRNA level diversity (which does not provide strain-level insights) and on microbiomes from diverse environments.

One of the datasets analysed here is a subset of a cross-sectional cohort from the Human Microbiome Project. The other dataset comes from a single individual sampled longitudinally over 18 months. This second dataset allowed the authors to not only assess the links between different levels of diversity at single timepoints, but test if high diversity at a given timepoint is associated with increased strain-level diversity at future timepoints.

Understanding eco-evolutionary dynamics of diversity in natural microbial communities is an important question that remains challenging to address. The paper is well-written and the detailed description of the methodological approaches and statistical analyses is exemplary. Most of the analyses carried out in this study seem to be technically sound.

We thank the reviewer for their kind words, comments, and suggestions.

The major limitation of this study comes with the fact that only correlations are presented, some of which are rather weak, contrast each other, or are based on a small number of data points. In addition, finding that diversity at a given taxonomic rank is associated with diversity within a given taxon is a pattern that can be explained by many different underlying processes, e.g. species-area relationships, nutrient (diet) diversity, stressor diversity, immigration rate, and niche creation by other microbes (i.e. DBD). Without experiments, it remains vague if DBD is the underlying process that acts in these communities based on the observed patterns.

We thank the reviewer for their comments. First, regarding the issue of this being a correlative study, we now more clearly acknowledge that mechanistic studies (perhaps in experimental settings) are required to fully elucidate DBD and BQH dynamics. However, we note that our correlational study from natural communities is complementary to experimental and modeling studies, to test the extent to which their predictions hold in more complex, realistic settings. This is now mentioned throughout the manuscript, most explicitly at the end of the Introduction:

“Although such analyses of natural diversity cannot fully control for unmeasured confounding environmental factors, they are an important complement to controlled experimental and theoretical studies which lack real-world complexity.”

Second, to increase the number of data points analyzed in the Poyet study, we now include 15 species and four different hosts (new Figure 5). The association between community diversity and gene loss is now much more statistically robust, and consistent across the Poyet and HMP time series.

Third, we acknowledge more clearly in the Discussion that other processes, including diet and other environmental factors can generate the DBD pattern. We also now stress more prominently the possibility that strain migration across hosts may be responsible for the patterns observed. For example, in Figure 1, we illustrate the possibility of strain migration generating the patterns we observe.

Below we quote a paragraph that we have now added in the Discussion:

"Second, we cannot establish causal relationships without controlled experiments. We are therefore careful to conclude that positive diversity slopes are consistent with the predictions of DBD, and negative slopes with EC, but unmeasured environmental drivers could be at play. For example, increased dietary diversity could simultaneously select for higher community diversity and also higher intra-species diversity. In our previous study, we found that positive diversity slopes persisted even after controlling for potential abiotic drivers such as pH and temperature (Madi et al., 2020), but a similar analysis was not possible here due to a lack of metadata. Neutral processes can account for several ecological patterns such as species-area relationships (Hubbell, 2001), and must be rejected in favor of niche-centric models like DBD or EC. Using neutral models without DBD or EC, we found generally flat or negative diversity slopes due to sampling processes alone and that positive slopes were hard to explain with a neutral model (Madi et al., 2020). These models were intended mainly for 16S rRNA gene sequence data, but we expect the general conclusions to extend to metagenomic data. Nevertheless, further modeling and experimental work will be required to fully exclude a neutral explanation for the diversity slopes we report in the human gut microbiome.”

Finally, we now put more emphasis on the importance of migration (strain invasion) as a non-exclusive alternative to de novo mutation and gene gain/loss. This is mentioned in the Abstract and is also illustrated in the revised Figure 1.

Another limitation is that the total number of reads (5 mio for the longitudinal dataset and 20 mio for the cross-sectional dataset) is low for assessing strain-level diversity in complex communities such as the human gut microbiota. This is probably the reason why the authors only looked at one species with sufficient coverage in the longitudinal dataset.

Indeed, this is a caveat which means we can only consider sub-species diversity in relatively abundant species. Nevertheless, this allows us to study dozens of species in the HMP and 15 in the more frequent Poyet time series. As more deeply sequenced metagenomes become available, future studies will be able to access the rarer species to test whether the same patterns hold or not. This is now mentioned prominently as a caveat our study in the second Discussion paragraph:

“First, using metagenomic data from human microbiomes allowed us to study genetic diversity, but limited us to considering only relatively abundant species with genomes that were well-covered by short sequence reads. Deeper or more targeted sequencing may permit us to determine whether the same patterns hold for rarer members of the microbiome. However, it is notable that the majority of the dozens of species across the two datasets analyzed support DBD, suggesting that the phenomenon may generalize.”

We also note that rarefaction was only applied to calculate community richness, not to estimate sub-species diversity. We apologize for this confusion, which is now clarified in the Methods as follows:

“SNV and gene content variation within a focal species were ascertained only from the full dataset and not the rarefied dataset.”

Analyzing the effect of diversity at a given timepoint on strain-level diversity at a later timepoint adds an important new dimension to this study which was not assessed in the previous study about the DBD in microbiomes by some of the authors. However, only a single species was analysed in the longitudinal dataset and comparisons of diversity were only done between two consecutive timepoints. This dataset could be further exploited to provide more insights into the prevailing patterns of diversity.

We thank the reviewer for raising this point. We now have considered all 15 species for which there was sufficient coverage from the Poyet dataset, which included four different stool donors. Additionally, in the HMP dataset, we analyze 54 species across 154 hosts, with both datasets showing the same correlation between community diversity and gene loss.

Additionally, we followed the suggestion of the reviewer of examining additional time lags, and in Figure 5 we do observe a dependency on time. This is now described in the Results as follows:

“Using the Poyet dataset, we asked whether community diversity in the gut microbiome at one time point could predict polymorphism change at a future time point by fitting GAMs with the change in polymorphism rate as a function of the interaction between community diversity at the first time point and the number of days between the two time points. Shannon diversity at the earlier time point was correlated with increases in polymorphism (consistent with DBD) up to ~150 days (~4.5 months) into the future (Figure S4), but this relationship became weaker and then inverted (consistent with EC) at longer time lags (Fig 5A, Table S8, GAM, P=0.023, Chi-square test). The diversity slope is approximately flat for time lags between four and six months, which could explain why no significant relationship was found in HMP, where samples were collected every ~6 months. No relationship was observed between community richness and changes in polymorphism (Table S8, GAM, P>0.05).”

Finally, the evidence that gene loss follows increase in diversity is weak, as very few genes were found to be lost between two consecutive timepoints, and the analysis is based on only a single species. Moreover, while positive correlation were found between overall community diversity and gene family diversity in single species, the opposite trend was observed when focusing on pathway diversity. A more detailed analysis (of e.g. the functions of the genes and pathways lost/gained) to explain these seemingly contrasting results and a more critical discussion of the limitations of this study would be desirable.

We agree that our previous analysis of one species in one host provided weak support for gene loss following increases in diversity. As described in the response above, we have now expanded this analysis to 15 focal species and 4 independent hosts with extensive time series. We now analyze this larger dataset and report the more statistically robust results as follows:

“We found that community Shannon diversity predicted future gene loss in a focal species, and this effect became stronger with longer time lags (Fig 5B, Table S9, GLMM, P=0.006, LRT for the effect of the interaction between the initial Shannon diversity and time lag on the number of genes lost). The model predicts that increasing Shannon diversity from its minimum to its maximum would result in the loss of 0.075 genes from a focal species after 250 days. In other words, about one of the 15 focal species considered would be expected to lose a gene in this time frame.

Higher Shannon diversity was also associated with fewer gene gains, and this relationship also became stronger over time (Fig 5C, Table S9, GLMM, P=1.11e-09, LRT). We found a similar relationship between community species richness and gene gains, although the relationship was slightly positive at shorter time lags (Fig 5D, Table S9, GLMM, P=3.41e-04, LRT). No significant relationship was observed between richness and gene loss (Table S9, GLMM, P>0.05). Taken together with the HMP results (Fig 4), these longer time series reveal how the sign of the diversity slope can vary over time and how community diversity is generally predictive of reduced focal species gene content.”

As described in detail in the response to Reviewer 1 above, we found that the HUMAnN2 pathway analyses previously described suffered from technical challenges and we deemed them inconclusive. We have therefore removed the pathway results from the manuscript.

Reviewer #3 (Public Review):

This work provides a series of tests of hypothesis, which are not mutually exclusive, on how genomic diversity is structured within human microbiomes and how community diversity may influence the evolution of a focal species.

Strengths:

The paper leverages on existing metagenomic data to look at many focal species at the same time to test for the importance of broad eco-evolutionary hypothesis, which is a novelty in the field.

Thank you for the succinct summary and recognition of the strengths of our work.

Weaknesses:

It is not very clear if the existing metagenomic data has sufficient power to test these models.

It is not clear, neither in the introduction nor in the analysis what precise mechanisms are expected to lead to DBD.

The conclusion that data support DBD appears to depend on which statistics to measure of community diversity are used. Also, performing a test to reject a null neutral model would have been welcome either in the results or in the discussion.

In our revised manuscript, we emphasize several caveats – including that we only have power to test these hypotheses in focal species with sufficient metagenomic coverage to measure sub-species diversity. We also describe more in the Introduction how the processes of competition and niche construction can lead to DBD. We also acknowledge that unmeasured abiotic drivers of both community diversity and sub-species diversity could also lead to the observed patterns. Throughout the manuscript, we attempt to describe the results and acknowledge multiple possible interpretations, including DBD and EC acting with different strengths on different species and time scales. Our previous manuscript assessing the evidence for DBD using 16S rRNA gene amplicon data from the Earth Microbiome Project (Madi et al., eLife 2020) assessed null models based on neutral ecological theory, and found it difficult to explain the observation of generally positive diversity slopes without invoking a non-neutral mechanism like DBD. While a new null model tailored to metagenomic data might provide additional nuance, we think developing one is beyond the scope of the manuscript – which is in the format of a short ‘Research Advance’ to expand on our previous eLife paper, and we expect that the general results of our previously reported null model provide a reasonable intuition for our new metagenomic analysis. This is now mentioned in the Discussion as follows:

“In our previous study, we found that positive diversity slopes persisted even after controlling for potential abiotic drivers such as pH and temperature (Madi et al., 2020), but a similar analysis was not possible here due to a lack of metadata. Neutral processes can account for several ecological patterns such as species-area relationships (Hubbell, 2001), and must be rejected in favor of niche-centric models like DBD or EC. Using neutral models without DBD or EC, we found generally flat or negative diversity slopes due to sampling processes alone and that positive slopes were hard to explain with a neutral model (Madi et al., 2020). These models were intended mainly for 16S rRNA gene sequence data, but we expect the general conclusions to extend to metagenomic data. Nevertheless, further modeling and experimental work will be required to fully exclude a neutral explanation for the diversity slopes we report in the human gut microbiome.”

Author Response

Reviewer #2 (Public Review):

Zou et al. presented a comprehensive study where they generated single-cell RNA profiling of 138,982 cells from 13 samples of six patients including AK, squamous cell carcinoma in situ (SCCIS), cSCC, and their matched normal tissues, covering comprehensive clinical courses of cSCC. Using bioinformatics analysis, they identified keratinocytes, CAFs, immune cells, and their subpopulations. The authors further compared signatures within subpopulations of keratinocytes along with the clinical progression, especially basal cells, and identified many interesting genes. They also further validate some of the markers in an independent cohort using IHC, followed by some knockdown experiments using cSCC cell lines.

The strength of this study is the unique data set they have created, providing the community with invaluable resources to study and validate their findings. However, a lot of analyses were not robust enough to support the claims and conclusions in the paper. More clarification and cross-comparison with polished data are needed to further strengthen the study and claims.

1) Stemness markers were used. The authors used COL17A1, TP63, ITGB1, and ITGA3 to represent stemness markers. However, these were not common classic stemness markers used in cSCC. What is the source claiming these genes were stemness markers in cSCC? TP63 is a master regulator and early driver event in SCC, while COL17A1, ITGB1, and ITGA3 are all ECM genes. The authors need to use commonly well-known stem cell markers in cSCC, e.g., LGR5, to mark stem-like cells.

Thanks for raising this good point. We may not have provided a clear description of the markers COL17A1, TP63, ITGB1, and ITGA3 in the previous texts. We would like to clarify that these genes were used as the markers of epidermal stem cells in normal skin samples rather than tumor stem cells in cSCC. To avoid any possible misunderstanding, we revised the main text accordingly and added the references [4-11].

2) Cell proportion analysis. The authors used the mean proportions to compare different clinical groups for subpopulations of keratinocytes, e.g., Figure 2B, and Figure 5B. This is not robust, as no statistics can be derived from this. For example, from Fig 2A, it is clearly shown there is a high level of heterogeneity of cellular compositions for normal samples. One cannot say which group is higher or lower simply based on mean not variance as well.

We replotted the proportion analysis with statistics and presented the new graphs in Figure 2-figure supplement 1 for Figure 2B and Figure 5-figure supplement 1 for Figure 5B.

3) Basal tumour cells in SCCIS and SCC. To make the findings valid, authors need to compare these cells/populations with the keratinocyte cell populations defined by Ji et al. Cell 2020. Do basal-SCCIS-tumours cells, also in SCC samples, resemble any of the population defined in Ji et al. Ji et al. also had 10 match normal, thus the authors need to validate their findings of SCC vs normal analysis using the Ji et al. dataset.

Thanks for this valuable suggestion. We compared basal tumor cell in our study with the cell populations defined in Ji et al. Cell 2020 data using SingleCellNet [1]. The results showed that both the basal-SCCIS-tumor cells of SCCIS and basal tumor cells of cSCC in our study closely resemble the Tumor_KC_Basal subcluster defined in Ji et al’s paper (Figure 4-figure supplement 4, C and D). Tumor_KC_Basal highly expressed CCL2, CXCL14, FTH1, MT2A, which is consistent with our findings in basal tumor cells.

4) Copy number analysis. Authors used inferCNV to perform copy number analysis using scRNA-seq data and identified CNVs in subpopulations of keratinocytes in SCCIS and SCC. To ensure these CNVs were not artefacts, were some of the CNVs identified by inferCNV well-known copy number changes previously reported in cSCC?

In poorly-differentiated cSCC sample, the significant gains in chromosome 7, 9 and deletion in chromosome 10 were reported in previous study, indicating the reliability of the CNV analysis results (Figure 5-figure supplement 2) [12].

5) Pseudotime analysis lines 308-313. Not sure the pseudotime analysis added much as, as it is unclear two distinct subgroups were identified from this analysis. Suggest removing this to keep it neater

Thank you for this suggestion. We have deleted the result of pseudotime analysis.

6) Selection of candidate genes for validation using IHC and cell line work. For example, lines 205-206, lines 352-356 and lines 437-441, authors selected several genes associated with AK and SCC to further validate using IHC and cell line knockdown work. What are the criteria for selecting those genes for validation? It is unclear to readers how these were selected. It reads like a fishing experiment, then followed by a knockdown. Clear rationale/criteria need to be elaborated.

The first consideration of candidate gene selection is the fold change of expression. We have provided the statistical results of DEGs in Supplementary file 1b, 1h, 1j-1m. Then we selected top changed genes and conducted an extensive literature search on these genes. We prioritized genes that, although not directly associated with cSCC development, have a close relationship with related pathways, as determined through functional enrichment analysis. These genes were arranged for further verification experiments. We have added more details in main text and methods section.

7) TME. Compared to keratinocytes populations, the investigation of TME cells was weak. (a) can authors produce UMAP files just for T cells, DC cells, and fibroblasts separately? Figure 7B is not easy to see those subclusters. (b) similar to what was done for keratinocytes, can authors find differentially expressed clusters and genes among the different clinical groups, associated with disease progression? (c) where are the myeloid cell populations, also B cells?

Thank you for your suggestions. (a) We have added the UMAP files for T cells, DC cells and stromal cells separately in new Figure 7A. (b) We identified DEGs in TME cells among the different groups. Several key genes showed monotonically changing trends associated with disease progression. For example, with the increase of malignancy, FOS shows down-regulation while S100A8 and S100A9 monotonically increase in all three types of TME cells (Figure 7C). (c) We identified two types of myeloid cell populations, macrophage and monocyte derived DCs (MoDC). We didn’t find other myeloid cells, such as neutrophil. For B cells, there were only 28 B cells in poorly-differentiated cSCC sample, which didn’t meet the threshold for further cell-cell communication analysis.

8) Heat shock protein genes line 327-329. HSP signature was well-known to be induced via tissue dissociation and library prep during the scRNA experiment. How could the authors be sure these were not artefacts induced by the experiment? If authors regress their gene expression against HSP gene signatures, would this cluster still be identified?

Thank you for this valuable suggestion. It is important to note that the Basal-SCCIS-tumor cluster was identified through CNV analysis, rather than the HSP signature. To address this concern and further validate this result, “AddModuleScore” function in Seurat package was used to regress gene expression against HSP gene signatures for retrieved basal cells. Our result showed that Basal_SCCIS tumor population still can be identified after regression, even more clearly (Author response image 1).

Author response image 1.

The identity of Basal-SCCIS-tumor cluster considering regression against HSP signatures.

9) Cell-cell communication analysis. The authors claimed that that cell-to-cell interaction was significantly enhanced in poorly-differentiated cSCC, and multiple interaction pathways were significantly active. How was this kind of analysis carried out? How did the authors define significance? what statistical method was used? these were all unclear. Furthermore, it is difficult to judge the robustness of the cell-cell communication analysis. Were these findings also supported by another method, such as celltalker, and cellphoneDB?

To determine the significance of the increased overall cell-to-cell interaction strength between two groups, we utilized CellChat to obtain the communication strength in different samples. We combined the communication strength based on cell type pairs, where missing values were set to 0. We performed a paired Wilcoxon test to determine whether the enhancement of cell-to-cell interaction between samples was significant.

For the comparison of outgoing or incoming interaction strength of the same cell types between two groups, we first extracted the communication strength of each signal pathway contributing to outgoing or incoming strength, and then merged the strengths of signal pathways among samples, where the strength of non-shared pathways with missing value was determined to be 0. Subsequently, we performed a paired Wilcoxon test to define the significance.

For multiple groups comparisons, the Kruskal-Wallis rank sum test was first performed. If the p-value is less than 0.1, the pairwise Wilcoxon test was used for subsequent pairwise comparisons. The comparison of individual signaling pathways between groups is similar to the above. We defined p-value < 0.1 as significance threshold. We have added the significance test method in figure legend for Figure 7 and Figure 8 as well as and detailed statistical data in new Supplementary file 1q-1u.

As suggested, we also used the approach of CellPhoneDB based on CellChatDB database to verify our cell-cell communication results. There are 55-58% of the ligand-receptor interactions predicted by CellChat were also predicted by CellPhoneDB (Author response image 2). The enhancement of cell interaction through MHC-II, Laminin and TNF signaling pathways in poorly-differentiated cSCC sample compare to normal sample were consistent in both CellChat and CellPhoneDB (Figure 8C and Figure 8-figure supplement 1B).

Author response image 2.

The overlap of the predicted ligand-receptor interactions between CellChat and CellPhoneDB.

10) Statistics and significance. In general, the detail of statistics and significance was lacking throughout the paper. Authors need to specify what statistical tests were used, and the p-values. It is difficult to judge the correctness of the test, and robustness without seeing the stats.

We have included all statistics and significance values in the figure legend and supplemental tables, and described the statistical tests in the methods section. In this revision, we have added the necessary details of statistics and significance in the main text and figures.

11) Overall, this manuscript needs a lot of re-writing. A lot of discussion was also included in the results, making it really difficult to read overall. The authors should simplify the results sections, remove the discussion bits, and further highlight and streamline with the key results of this paper.

Thanks a lot for this advice. We have revised the paper thoroughly, removed discussion in results section to make the manuscript easier to read.

Author Response

Reviewer #3 (Public Review):

This manuscript by Pendse et al aimed to identify the role of the complement component C1q in intestinal homeostasis, expecting to find a role in mucosal immunity. Instead, however, they discovered an unexpected role for C1qa in regulating gut motility. First, using RNA-Seq and qPCR of cell populations isolated either by mechanical separation or flow cytometry, the authors found that the genes encoding the subunits of C1q are expressed predominantly in a sub-epithelial population of cells in the gut that Cd11b+MHCII+F4/80high, presumably macrophages. They support this conclusion by analyzing mice in which intestinal macrophages are depleted with anti-CSF1R antibody treatment and show substantial loss of C1qa, b and c transcripts. Then, they generate Lyz2Cre-C1qaflx/flx mice to genetically deplete C1qa in macrophages and assess the consequences on the fecal microbiome, transcript levels of cytokines, macromolecular permeability of the epithelial barrier, and immune cell populations, finding no major effects. Furthermore, provoking intestinal injury with chemical colitis or infection (Citrobacter) did not reveal macrophage C1qa-dependent changes in body weight or pathogen burden.

Then, they analyzed C1q expression by IHC of cross-sections of small and large intestine and find that C1q immunoreactivity is detectable adjacent to, but not colocalizing with, TUBB3+ nerve fibers and CD169+ cells in the submucosa. Interestingly, they find little C1q immunoreactivity in the muscularis externa. Nevertheless, they perform RNA-sequencing of LMMP preparations (longitudinal muscle with adherent myenteric plexus) and find a number of changes in gene ontology pathways associates with neuronal function. Finally, they perform GI motility testing on the conditional knockout mice and find that they have accelerated GI transit times manifesting with subtle changes in small intestinal transit and more profound changes in measures of colonic motility.

Overall, the manuscript is very well-written and the observation that macrophages are the major source of C1q in the intestine is well supported by the data, derived from multiple approaches. The observations on C1q localization in tissue and the strength of the conclusions that can be drawn from their conditional genetic model of C1qa depletion, however, would benefit from more rigorous validation.

1) Interpretation of the majority of the findings in the paper rest on the specificity of the Lyz2 Cre for macrophages. While the specificity of this Cre to macrophages and some dendritic cells has been characterized in the literature in circulating immune cells, it is not clear if this has been characterized at the tissue level in the gut. Evidence demonstrating the selectivity of Cre activity in the gut would strengthen the conclusions that can be drawn.

As indicated by the reviewer, Cre expression driven by the Lyz2 promoter is restricted to macrophages and some myeloid cells in the circulation (Clausen et al., 1999). To better understand intestinal Lyz2 expression at a cellular level, we analyzed Lyz2 transcripts from a published single cell RNAseq analysis of intestinal cells (Xu et al., 2019; see Figure below). These data show that intestinal Lyz2 is also predominantly expressed in gut macrophages with limited expression in dendritic cells and neutrophils.

Figure. Lyz2 expression from single cell RNAseq analysis of mouse intestinal cells. Data are from Xu et al., Immunity 51, 696-708 (2019). Analysis was done through the Single Cell Portal, a repository of scRNAseq data at the Broad Institute.

Additionally, our study shows that intestinal C1q expression is restricted to macrophages (CD11b+MHCII+F4/80hi) and is absent from other gut myeloid cell lineages (Figure 1E-H). This conclusion is supported by our finding that macrophage depletion via anti-CSF1R treatment also depletes most intestinal C1q (Figure 2A-C). Importantly, we found that the C1qaDMf mice retain C1q expression in the central nervous system (Figure 2 – figure supplement 1). Thus, the C1qaDMf mice allow us to assess the function of macrophage C1q in the gut and uncouple the functions of macrophage C1q from those of C1q in the central nervous system.

2) Infectious and inflammatory colitis models were used to suggest that C1qa depletion in Lyz2+ lineage cells does not alter gut mucosal inflammation or immune response. However, the phenotyping of the mice in these models was somewhat cursory. For example, in DSS only body weight was shown without other typical and informative read-outs including colon length, histological changes, and disease activity scoring. Similarly, in Citrobacter only fecal cfu were measured. Especially if GI motility is accelerated in the KO mice, pathogen burden may not reflect efficiency of immune-mediated clearance alone.

We have added additional results which support our conclusion that C1qaDMf mice do not show a heightened sensitivity to acute chemically induced colitis. In Figure 3 – figure supplement 1 we now show a histological analysis of the small intestines of DSS-treated C1qafl/fl and C1qaΔMφ mice. This analysis shows that C1qaDMf mice have similar histopathology, colon lengths, and histopathology scores following DSS treatment. Likewise, our revised manuscript includes histological images of the colons of Citrobacter rodentium-infected C1qafl/fl and C1qaΔMφ mice showing similar pathology (Figure 3 – figure supplement 2).

3) The evidence for C1q expression being restricted to nerve-associated macrophages in the submucosal plexus was insufficient. Localization was shown at low magnification on merged single-planar images taken from cross-sections. The data shown in Figure 4C is not of sufficient resolution to support the claims made - C1q immunoreactivity, for example, is very difficult to even see. Furthermore, nerve fibers closely approximate virtually type of macrophage in the gut, from those in the lamina propria to those in the muscularis….Finally, the resolution is too low to rule out C1q immunoreactivity in the muscularis externa.

Similar points were raised by Reviewer 2. Our original manuscript claimed that C1q-expressing macrophages were mostly located near enteric neurons in the submucosal plexus but were largely absent from the myenteric plexus. However, as both Reviewers have pointed out, this conclusion was based solely on our immunofluorescence analysis of tissue cross-sections.

To address this concern we further characterized C1q+ macrophage localization by performing a flow cytometry analysis on macrophages isolated from the mucosa (encompassing both the lamina propria and submucosa) and the muscularis, finding similar levels of C1q expression in macrophages from both tissues (Figure 4 – figure supplement 1 in the revised manuscript). Although the mucosal macrophage fraction encompasses both lamina propria and submucosal macrophages, our immunofluorescence analysis (Figure 4 B and C) suggests that the mucosal C1q-expressing macrophages are mostly from the submucosal plexus. This observation is consistent with the immunofluorescence studies of CD169+ macrophages shown in Asano et al., which suggest that most C169+ macrophages are located in or near the submucosal region, with fewer near the villus tips (Fig. 1e, Nat. Commun. 6, 7802).

Most importantly, our flow cytometry analysis indicates that the muscularis/myenteric plexus harbors C1q-expressing macrophages. To further characterize C1q expression in the muscularis, we performed RNAscope analysis by confocal microscopy of the myenteric plexus from mouse small intestine and colon (Figure 4D). The results show numerous C1q-expressing macrophages positioned close to myenteric plexus neurons, thus supporting the flow cytometry analysis. We note that although the majority of C1q immunofluorescence in our tissue cross-sections was observed in the submucosal plexus, we did observe some C1q expression in the muscularis by immunofluorescence (Figure 4B and C). We have rewritten the Results section to take these new findings into account.

Is the 5um average on the proximity analysis any different for other macrophage populations to support the idea of a special relationship between C1q-expressing macrophages and neurons?

We agree that the proximity analysis lacks context and have therefore removed it from the figure. The other data in the figure better support the idea that C1q+ macrophages are found predominantly in the submucosal and myenteric plexuses and that they are closely associated with neurons at these tissue sites.

There are many vessels in the submucosa and many associated perivascular nerve fibers - could the proximity simply reflect that both cell types are near vessels containing C1q in circulation?

Our revised manuscript includes RNAscope analysis showing C1q transcript expression by macrophages that are closely associated with enteric neurons (Figure 4D). These findings support the idea that the C1q close to enteric neurons is derived from macrophages rather than from the circulation.

4) A major disconnect was between the observation that C1q expression is in the submucosa and the performance of RNA-seq studies on LMMP preparations. This makes it challenging to draw conclusions from the RNA-Seq data, and makes it particularly important to clarify the specificity of Lyz2-Cre activity.

Our revised manuscript provides flow cytometry data (Figure 4 – figure supplement 1) and RNAscope analysis (Figure 4D) showing that C1q is expressed in macrophages localized to the myenteric plexus. This accords with the results of our RNAseq analysis, which indicates altered LMMP neuronal function in C1qa∆Mφ mice (Figure 6A and B). Since neurons in the myenteric plexus are known to govern gut motility, it also helps to explain our finding that gut motility is accelerated in C1qa∆Mφ mice.

Finally, the pathways identified could reflect a loss of neurons or nerve fibers. No assessment of ENS health in terms of neuronal number or nerve fiber density is provided in either plexus.

Reviewers 1 and 2 also raised this point. Our revised manuscript includes a comparison of the numbers of enteric neurons in C1qafl/fl and C1qaΔMφ mice. There were no marked differences in neuron numbers in C1qaDMf mice when compared to C1qafl/fl controls (Figure 5A and B). There were also similar numbers of inhibitory (nitrergic) and excitatory (cholinergic) neuronal subsets and a similar enteric glial network (Figure 5C-E). Thus, our data suggest that the altered gut motility in the C1qaΔMφ mice arises from altered neuronal function rather than from an overt loss of neurons or nerve fibers. This conclusion is further supported by increased neurogenic activity of peristalsis (Figure 6H and I), and the expression of the C1q receptor BAI1 on enteric neurons (Figure 6 – figure supplement 4).

5) To my knowledge, there is limited evidence that the submucosal plexus has an effect on GI motility. A recent publication suggests that even when mice lack 90% of their submucosal neurons, they are well-appearing without overt deficits (PMID: 29666241). Submucosal neurons, however, are well known to be involved in the secretomotor reflex and fluid flux across the epithelium. Assessment of these ENS functions in the knockout mice would be important and valuable.

Our revised manuscript provides new data showing C1q expression by muscularis macrophages in the myenteric plexus. We analyzed muscularis macrophages by flow cytometry and found that they express C1q (Figure 4 – figure supplement 1). These findings are further supported by RNAscope analysis of C1q expression in wholemounts of LMMP from small intestine and colon (Figure 4D and E). These results are thus consistent with the increased CMMC activity and accelerated gut motility in the C1qaDMf mice. As suggested by the reviewer, our finding of C1q+ macrophages in the submucosal plexus indicates that C1q may also have a role controlling the function of submucosal plexus neurons. We are further exploring this idea through extensive additional experimentation. Given the expanded scope of these studies, we are planning to include them in a follow-up manuscript.

6) Immune function and GI motility can be highly sex-dependent - in all experiments mice of both sexes were reportedly used but it is not clear if sex effects were assessed.

This is a great point, and as suggested by the reviewer we indeed did encounter differences between male and female mice in our preliminary assays of gut motility. We therefore conducted our quantitative comparisons of gut motility between C1qafl/fl and C1qaDMf mice in male mice and now clearly indicate this point in the Materials and Methods.

Author Response

Reviewer #1 (Public Review):

The manuscript by Royall et al. builds on previous work in the mouse that indicates that neural progenitor cells (NPCs) undergo asymmetric inheritance of centrosomes and provides evidence that a similar process occurs in human NPCs, which was previously unknown.

The authors use hESC-derived forebrain organoids and develop a novel recombination tag-induced genetic tool to birthdate and track the segregation of centrosomes in NPCs over multiple divisions. The thoughtful experiments yield data that are concise and well-controlled, and the data support the asymmetric segregation of centrosomes in NPCs. These data indicate that at least apical NPCs in humans undergo asymmetric centrosome inheritance. The authors attempt to disrupt the process and present some data that there may be differences in cell fate, but this conclusion would be better supported by a better assessment of the fate of these different NPCs (e.g. NPCs versus new neurons) and would support the conclusion that younger centriole is inherited by new neurons.

We thank the reviewer for their supportive comments (“…thoughtful experiments yield data that are concise and well-controlled…”).

Reviewer #2 (Public Review):

Royall et al. examine the asymmetric inheritance of centrosomes during human brain development. In agreement with previous studies in mice, their data suggest that the older centrosome is inherited by the self-renewing daughter cell, whereas the younger centrosome is inherited by the differentiating daughter cell. The key importance of this study is to show that this phenomenon takes place during human brain development, which the authors achieved by utilizing forebrain organoids as a model system and applying the recombination-induced tag exchange (RITE) technology to birthdate and track the centrosomes.

Overall, the study is well executed and brings new insights of general interest for cell and developmental biology with particular relevance to developmental neurobiology. The Discussion is excellent, it brings this study into the context of previous work and proposes very appealing suggestions on the evolutionary relevance and underlying mechanisms of the asymmetric inheritance of centrosomes. The main weakness of the study is that it tackles asymmetric inheritance only using fixed organoid samples. Although the authors developed a reasonable mode to assign the clonal relationships in their images, this study would be much stronger if the authors could apply time-lapse microscopy to show the asymmetric inheritance of centrosomes.

We thank the reviewer for their constructive and supportive comments (“…the study is well executed and brings new insights of general interest for cell and developmental biology with particular relevance to developmental neurobiology….”). We understand the request for clonal data or dynamic analyses in organoids (e.g., using time-lapse microscopy). We also agree that such data would certainly strengthen our findings. However, as outlined above (please refer to point #1 of the editorial summary), this is unfortunately currently not feasible. However, we have explicitly discussed this shortcoming in our revised manuscript and why future experiments (with advanced methodology) will have to do these experiments.

Reviewer #3 (Public Review):

In this manuscript, the authors report that human cortical radial glia asymmetrically segregates newly produced or old centrosomes after mitosis, depending on the fate of the daughter cell, similar to what was previously demonstrated for mouse neocortical radial glia (Wang et al. 2009). To do this, the authors develop a novel centrosome labelling strategy in human ESCs that allows recombination-dependent switching of tagged fluorescent reporters from old to newly produced centrosome protein, centriolin. The authors then generate human cortical organoids from these hESCs to show that radial glia in the ventricular zone retains older centrosomes whereas differentiated cells, i.e. neurons, inherit the newly produced centrosome after mitosis. The authors then knock down a critical regulator of asymmetric centrosome inheritance called Ninein, which leads to a randomization of this process, similar to what was observed in mouse cortical radial glia.

A major strength of the study is the combined use of the centrosome labelling strategy with human cortical organoids to address an important biological question in human tissue. This study is similarly presented as the one performed in mice (Wang et al. 2009) and the existence of the asymmetric inheritance mechanism of centrosomes in another species grants strength to the main claim proposed by the authors. It is a well-written, concise article, and the experiments are well-designed. The authors achieve the aims they set out in the beginning, and this is one of the perfect examples of the right use of human cortical organoids to study an important phenomenon. However, there are some key controls that would elevate the main conclusions considerably.

We thank the reviewer for their overall support of our findings (“..authors achieve the aims they set out in the beginning, and this is one of the perfect examples of the right use of human cortical organoids to study an important phenomenon…”). We also understand the reviewer’s request for additional experiments/controls that “…would elevate the main conclusions considerably.”

1) The lack of clonal resolution or timelapse imaging makes it hard to assess whether the inheritance of centrosomes occurs as the authors claim. The authors show that there is an increase in newly made non-ventricular centrosomes at a population level but without labelling clones and demonstrating that a new or old centrosome is inherited asymmetrically in a dividing radial glia would grant additional credence to the central conclusion of the paper. These experiments will put away any doubt about the existence of this mechanism in human radial glia, especially if it is demonstrated using timelapse imaging. Additionally, knowing the proportions of symmetric vs asymmetrically dividing cells generating old/new centrosomes will provide important insights pertinent to the conclusions of the paper. Alternatively, the authors could soften their conclusions, especially for Fig 2.

We understand the reviewer’s request. As outlined above (please refer to point #1 of the editorial summary), we had tried previously to add data using single cell timelapse imaging. However, due to the size and therefore weakness of the fluorescent signal we had failed despite extensive efforts. According to the reviewer’s suggestion we have now explicitly discussed this shortcoming and softened our conclusions.

2) Some critical controls are missing. In Fig. 1B, there is a green dot that does not colocalize with Pericentrin. This is worrying and providing rigorous quantifications of the number of green and tdTom dots with Pericentrin would be very helpful to validate the labelling strategy. Quantifications would put these doubts to rest. Additionally, an example pericentrin staining with the GFP/TdTom signal in figure 4 would also give confidence to the reader. For figure 4, having a control for the retroviral infection is important. Although the authors show a convincing phenotype, the effect might be underestimated due to the incomplete infection of all the analyzed cells.

We have included more rigorous quantifications in our revised manuscript.

For Figure 1: There are indeed some green speckles that might be misinterpreted as a green centrosome. However, the speckles are usually smaller and by applying a strict size requirement we exclude speckles. To check whether the classifier might interpret any speckles as centrosomes, we manually checked 60 green “dots” that were annotated as centrosome. From these images all green spots detected as centrosome co-localized with Pericentrin signal (Images shown in Author response image 1).

For Figure 4: as we are comparing cells that were either infected with a retrovirus expressing scrambled or Ninein-targeting shRNA we compare cells that experienced a similar treatment. Besides that, only cells infected with the virus express Cre-ERT2 whereby only the centrosomes of targeted cells were analyzed. Accordingly, we only compare cells expressing scrambled or Ninein-targeting shRNA, all surrounding “wt” cells are not considered.

Author response image 1.

Pictures used to test the classifier. Each of the green “dots” recognized by the classifier as a Centriolin-NeonGreen-containing centrosome (green) co-localized with Pericentrin signal (white).

3) It would be helpful if the authors expand on the presence of old centrosomes in apical radial glia vs outer radial glia. Currently, in figure 3, the authors only focus on Sox2+ cells but this could be complemented with the inclusion of markers for outer radial glia and whether older centrosomes are also inherited by oRGCs. This would have important implications on whether symmetric/asymmetric division influences the segregation of new/old centrosomes.

That is an interesting question and we do agree that additional analyses, stratified by ventricular vs. oRGCs would be interesting. However, at the time points analysed there are only very few oRGCs present (if any) in human ESC-derived organoids (Qian et al., Cell, 2016). However, we have now added this point for future experiments to our discussion.

Author Response

Reviewer #1 (Public Review):

The study by Akter et al demonstrates that astrocyte-derived L-lactate plays a key role in schema memory formation and promotes mitochondrial biogenesis in the Anterior Cingulate Cortex (ACC).

The main tool used by the authors is the DREADD technology that allows to pharmacologically activate receptors in a cell-specific manner. In the study, the authors used the DREADD technique to activate appropriately transfected astrocytes, a subtype of muscarinic receptor that is not normally present in cells. This receptor being coupled to a Gi-mediated signal transduction pathway inhibiting cAMP formation, the authors could demonstrate cell-(astrocyte) specific decreases in cAMP levels that result in decreased L-lactate production by astrocytes.

Behaviorally this pharmacological manipulation results in impairments of schema memory formation and retrieval in the ACC in flavor-place paired associate paradigms. Such impairments are prevented by co-administration of L-lactate.

The authors also show that activation of Gi signaling resulting in L-lactate decreased release by astrocytes impairs mitochondrial biogenesis in neurons in an L-lactate reversible manner.

By using MCT 2 inhibitors and an NMDAR antagonist the authors conclude that the molecular mechanisms underlying the observed effects are mediated by L-lactate entering neurons through MCT2 transporters and involve NMDAR.

Overall, the article's conclusions are warranted by the experimental evidence, but some weak points could be addressed which would make the conclusions even stronger.

The number of animals in some of the experiments is on the low side (4 to 6).

In the revised manuscript, we have increased the animal numbers in two key experimental groups (hM4Di-CNO and Control groups) of behavioral experiments. Now the animal numbers in different groups are as follows:

• 15 rats in hM4Di-CNO group

o Further divided into two subgroups for probe tests (PT1-4) conducted during flavor-place paired associate training; 8 rats in the hM4Di-CNO (saline) and 7 rats in the hM4Di-CNO (CNO) subgroups receiving I.P. saline or I.P. CNO, respectively, before these PTs.

• 8 rats in the Control group

• 7 rats in the Rescue group (hM4Di-CNO+L-lactate)

• 4 rats in the Control-CNO group. Animal number in this group was not increased as it was apparent from these 4 rats that CNO alone was not impairing the PA learning and memory retrieval in these rats (AAV8-GFAP-mCherry injected). Their result was very similar to the control group. Additionally, in a previous study (Liu et al., 2022), we showed that CNO administration in the rats injected with AAV8-GFAP-mCherry into the hippocampus does not show any impairments in schema.

Also, in the newly added open field test experiments to investigate the locomotor activity as suggested by the Reviewer #2, 8 rats were used in each group.

The use of CIN to inhibit MCT2 is not optimal. Authors may want to decrease MCT2 expression by using antisense oligonucleotides.

In the revised manuscript, we have conducted the experiment using MCT2 antisense oligodeoxynucleotide (ODN) as suggested.

To test whether the L-lactate-induced neuronal mitochondrial biogenesis is dependent on MCT2, we bilaterally injected MCT2 antisense oligodeoxynucleotide (MCT2-ODN, n=8 rats, 2 nmol in 1 μl PBS per ACC) or scrambled ODN (SC-ODN, n=8 rats, 2 nmol in 1 μl PBS per ACC) into the ACC. After 11 hours, bilateral infusion of L-lactate (10 nmol, 1 μl) or ACSF (1 μl) was given into the ACC and the rats were kept in the PA event arena. After 60 mins (12 hours from MCT2-ODN or SC-ODN administration), the rats were sacrificed. As shown in Author response image 1B, SC-ODN+L-lactate group showed significantly increased relative mtDNA copy number compared to the SC-ODN+ACSF group (p<0.001, ANOVA followed by Tukey's multiple comparisons test). However, this effect was completely abolished in MCT2-ODN+L-lactate group, suggesting that MCT2 is required for the L-lactate-induced mitochondrial biogenesis in the ACC.

We have integrated this new data and results in the revised manuscript.

Author response image 1.

Mitochondrial biogenesis by L-lactate is dependent on MCT2 and NMDAR. A. Experimental design to investigate whether MCT2 and NMDAR activity are required for L-lactate-induced mitochondrial biogenesis. B and C. mtDNA copy number abundance in the ACC of different rat groups relative to nDNA. Data shown as mean ± SD (n=4 rats in each group). ***p<0.001, ANOVA followed by Tukey's multiple comparisons test.

The experiment using AVP to block NMDAR only partially supports the conclusions. Indeed, blocking NMDAR will knock down any response that involves these receptors, whether L-lactate is necessary or not.

In the current study we found that Astrocytic Gi activation in the ACC reduced L-lactate level in the ECF of ACC which was also associated with decreased PGC-1α/SIRT3/ATPB/mtDNA abundance suggesting downregulation of mitochondrial biogenesis pathway. We also found that exogenous administration of L-lactate into the ACC of astrocytic Gi-activated rats rescued this downregulation. In line with this, in a recently published study (Akter et al., 2023), we found upregulation of mitochondrial biogenesis pathway in the hippocampus neurons of exogenous L-lactate-treated anesthetized rats. Another recent study has demonstrated that exercise-induced L-lactate release from skeletal muscle or I.P. injection of L-lactate can induce hippocampal PGC-1α (which is a master regulator of mitochondrial biogenesis) expression and mitochondrial biogenesis in mice (Park et al., 2021). Together, these results provide compelling evidence that L-lactate promotes mitochondrial biogenesis.

L-lactate is known to promote expression of synaptic plasticity genes like Arc, c-Fos, and Zif268 in neurons (Yang et al., 2014). After entry into the neuronal cytoplasm, mainly through MCT2, it is converted into pyruvate by lactate dehydrogenase 1 (LDH1). This conversion also produces NADH, affecting the redox state of the neuron. NADH positively modulates the activity of NMDAR resulting in enhanced Ca2+ currents, the activation of intracellular signaling cascades, and the induction of the expression of plasticity-associated genes (Yang et al., 2014; Magistretti & Allaman, 2018). The study demonstrated that L-lactate–induced plasticity gene expression was abolished in the presence of NMDAR antagonists including D-APV (Yang et al., 2014). These results suggested that the MCT2 and NMDAR are key players in the regulation of L-lactate induced plasticity gene expression.

In the current study, we investigated whether similar mechanisms might be involved in L-lactate-induced neuronal mitochondrial biogenesis. We now used MCT2 antisense oligodeoxynucleotide to decrease the expression of MCT2 (as mentioned in the previous response and Author response image 1B) and showed that MCT2 is necessary for L-lactate-induced mitochondrial biogenesis to manifest, indicating that L-lactate’s entry into the neuron is required. As mentioned before, after entry into neuron, L-lactate is converted into pyruvate by LDH, which also produce NADH, which in turn potentiates NMDAR activity. Therefore, we investigated whether NMDAR activity is required for L-lactate-induced mitochondrial biogenesis. We used D-APV to inhibit NMDAR (Author response image 1C) and found that L-lactate does not increase mtDNA copy number abundance if D-APV is given, suggesting that NMDAR activity is required for L-lactate to promote mitochondrial biogenesis.

NMDAR serves diverse functions. Therefore, as mentioned by the reviewer, blocking NMDAR may knock down many such functions. While our current data only suggests the involvement of MCT2 and NMDAR in the upregulation of mitochondrial biogenesis by L-lactate, we have not investigated other mechanisms and pathways modulating mitochondrial biogenesis that are either dependent or independent of MCT2 and NMDAR activity. Further studies are needed in future to dissect and better understand this interesting observation. We have now clarified this in the discussion section of the manuscript.

Is inhibition of glycogenolysis involved in the observed effects mediated by Gi signaling? Indeed, L-lactate is formed both by glycolysis and glycogenolysis. The authors could test whether the glycogen metabolism-inhibiting drug DAB would mimic the effects of Gi activation.

In this study we have shown that astrocytic Gi activation in the ACC leads to a decrease in the cAMP and L-lactate. L-lactate is produced by glycogenolysis and glycolysis. cAMP in astrocytes acts as a trigger for L-lactate production (Choi et al., 2012; Horvat, Muhič, et al., 2021; Horvat, Zorec, et al., 2021; Zhou et al., 2021) by promoting glycogenolysis and glycolysis (Vardjan et al., 2018; Horvat, Muhič, et al., 2021; Horvat, Zorec, et al., 2021). Therefore, one promising explanation of reduced L-lactate level observed in our study is the reduction of L-lactate production in the astrocyte due to decreased glycogen metabolism as a result of decreased cAMP. We have now mentioned this in the discussion.

DAB is an inhibitor of glycogen phosphorylase that suppresses L-lactate production. It was shown to impair memory by decreasing L-lactate (Newman et al., 2011; Suzuki et al., 2011; Iqbal et al., 2023). As we found that the impairment in the schema memory and mitochondrial biogenesis was associated with decreased L-lactate level in the ACC and that the exogenous L-lactate administration can rescue the impairments, it is likely that DAB will mimic the effect of Gi activation in terms of schema memory and mitochondrial biogenesis. However, further study is needed to confirm this.  

Reviewer #2 (Public Review):

The manuscript of Akter et al is an important study that investigates the role of astrocytic Gi signaling in the anterior cingulate cortex in the modulation of extracellular L-lactate level and consequently impairment in flavor-place associates (PA) learning. However, whereas some of the behavioral observations and signaling mechanism data are compelling, the conclusions about the effect on memory are inadequate as they rely on an experimental design that does not allow to differentiate acute or learning effect from the effect outlasting pharmacological treatments, i.e. effect on memory retention. With the addition of a few experiments, this paper would be of interest to the larger group of researchers interested in neuron-glia interactions during complex behavior.

• Largely, I agree with the authors' conclusion that activating Gi signaling in astrocytes impairs PA learning, however, the effect on memory retrieval is not that obvious. All behavioral and molecular signaling effects described in this study are obtained with the continuous presence of CNO, therefore it is not possible to exclude the acute effect of Gi pathway activation in astrocytes. What will happen with memory on retrieval test when CNO is omitted selectively during early, middle, or late session blocks of PA learning?

We have now added 8 more rats to the hM4Di-CNO group (i.e., the group with astrocytic Gi activation) to clarify the memory retrieval. These rats underwent flavor-place paired associate (PA) training similar to the previously described rats (n=7) of this group, that is they received CNO 30 minutes before and 30 minutes after the PA training sessions (S1-2, S4-8, S10-17). However, contrasting to the previous rats of this group which received CNO before PTs (PT1, PT2, PT3), we omitted the CNO (instead administered I.P. saline) selectively on these PTs conducted at the early, middle, and late stage of PA training, as suggested by the reviewer. These newly added rats did not show memory retrieval in these PTs, suggesting that the rats were not learning the PAs from the PA training sessions. See Author response image 2C-E, where this subgroup is denoted as hM4Di-CNO (Saline).

We then continued more PA training sessions (S21 onwards, Author response image 2B) for these rats without CNO. They gradually learned the PAs. PTs (PT5, PT6, PT7; Author response image 2G-I) were done during this continuation phase of PA training; once without CNO (i.e., with I.P. saline instead), and another one with CNO. As seen in the Author response image 2H and 2I, they retrieved the memory when PT6 and PT7 were done without CNO. However, if these PTs were done with CNO, they could not retrieve the memory. Together these results suggest that ACC astrocytic Gi activation by CNO during PT can impair memory retrieval in rats which have already learned the PAs.

As shown in the Author response image 2B, we replaced two original PAs with two new PAs (NPA 9 and 10) at S34. This was followed by PT8 (S35). As seen in Author response image 2J, these rats retrieved the NPA memory if the PT is done without CNO. However, they could not retrieve the NPA memory if the PT was done with CNO. This result suggests that ACC astrocytic Gi activation by CNO during PT can impair NPA memory retrieval.

In summary, these data show that astrocytic Gi activation in the ACC can impair PA memory retrieval. We have integrated this new data and results in the revised manuscript.

Author response image 2.

A. PI (mean ± SD) during the acquisition of the six original PAs (OPAs) (S1-2, 4-8, 10-17) and new PAs (NPAs) (S19) of the control (n=8), hM4Di-CNO (n=15), and rescue (hM4Di-CNO+L-lactate) (n=7) groups. From S6 onwards, hM4Di-CNO group consistently showed lower PI compared to control. However, concurrent L-lactate administration into the ACC (rescue group) can rescue this impairment. B. PI (mean ± SD) of hM4Di-CNO group (n=8) from S21 onwards showing gradual increase in PI when CNO was withdrawn. C, D, and E. Non-rewarded PTs (PT1, PT2, and PT3 conducted on S3, S9, and S18, respectively) to test memory retrieval of OPAs for the control, hM4Di-CNO, and rescue groups. The percentage of digging time at the cued location relative to that at the non-cued locations are shown (mean ± SD). In both PT2 and PT3, the control group spent significantly more time digging the cued sand well above the chance level, indicating that the rats learned OPAs and could retrieve it. Contrasting to this, hM4Di-CNO group did not spend more time digging the cued sand well above the chance level irrespective of CNO administration before the PTs. The rescue group showed results similar to the hM4Di-CNO group if CNO is given without L-lactate. On the other hand, they showed results similar to the control group if L-lactate is concurrently given with CNO, indicating that this group learned OPAs and could retrieve it. p < 0.05, p < 0.01, p < 0.001, one-sample t-test comparing the proportion of digging time at the cued sand well with the chance level of 16.67%. F. Non-rewarded PT4 (S20) which was conducted after replacing two OPAs with two NPAs (NPA 7 & 8) in S19 for the control, hM4Di-CNO, and rescue groups. Results show that the control group spent significantly more time digging the new cued sand well above the chance level indicating that the rats learned the NPAs from S19 and could retrieve it in this PT. Contrasting to this, hM4Di-CNO group did not spend more time digging the new-cued sand well above the chance level irrespective of CNO administration before the PT. The rescue group showed results similar to the hM4Di-CNO group if CNO is given without L-lactate. On the other hand, they showed results similar to the control group if L-lactate is concurrently given with CNO indicating that this group learned NPAs from S19 and could retrieve it. p < 0.001, one-sample t-test comparing the proportion of digging time at the new cued sand well with the chance level of 16.67%. G, H, and I. Non-rewarded PTs (PT5, PT6, and PT7 conducted on S23, S27, and S33, respectively) to test memory retrieval of OPAs for the hM4Di-CNO group. In both PT6 and PT7, the rats spent significantly more time digging the cued sand well above the chance level if the tests are done without CNO, indicating that the rats learned the OPAs and could retrieve it. However, CNO prevented memory retrieval during these PTs. p < 0.001, one-sample t-test comparing the proportion of digging time at the cued sand well with the chance level of 16.67%. J. Non-rewarded PT4 (S35) which was conducted after replacing two OPAs with two NPAs (NPA 9 & 10) in S34 for the hM4Di-CNO group. Results show that the rats spent significantly more time digging the new cued sand well above the chance level if CNO was not given before the PT, indicating that the rats learned the NPAs from S34 and could retrieve it in this PT. However, if CNO is given before the PT, the retrieval is impaired. *p < 0.001, one-sample t-test comparing the proportion of digging time at the new cued sand well with the chance level of 16.67%.

• I found it truly exciting that the administration of exogenous L-lactate is capable to rescue CNO-induced PA learning impairment, when co-applied. Would it be possible that this treatment has a sensitivity to a particular stage of learning (acquisition, consolidation, or memory retrieval) when L-lactate administration would be the most efficacious?

The hM4Di-CNO group, when continued with PA training without CNO (S21-S32) (Author response image 2B), was able to learn the six original PAs (OPAs). In the PT7 done at S33 (Author response image 2I), this group of rats was able to retrieve the memory if the test was done without CNO but could not retrieve the memory if CNO was given. Similarly, the Rescue group (hM4Di-CNO+L-lactate) (Author response image 2A), which received both CNO and L-lactate during PA training sessions (S1-S17), they were able to learn the OPAs. And at PT3 done at S18 (Author response image 2E), these rats were able to retrieve the memory when the test was done with CNO+L-lactate but not if the test is done with only CNO. Together, these results clearly show that ACC astrocytic Gi activation with CNO impairs memory retrieval and exogenous L-lactate can rescue the impairment. Therefore, it can be concluded that the memory retrieval is sensitive to L-lactate.

The PA learning is hippocampus-dependent. Over the course of repeated PA training, systems consolidation occurs in the ACC, after which the already learned PA memory (schema) becomes hippocampus-independent (Tse et al., 2007; Tse et al., 2011). A higher activation (indicated by expression of c-Fos) in the hippocampus relative to the ACC during the early period of schema development, and the reverse at the late stage was observed in our previous study (Liu et al., 2022). However, rapid assimilation of new PA into the ACC requires simultaneous activation/retrieval of previous schema from ACC and hippocampus dependent new PA learning (Tse et al., 2007; Tse et al., 2011). During new PA learning, increase of c-Fos neurons in both CA1 and ACC was detected (Liu et al., 2022).

Our hM4Di-CNO group received CNO 30 mins before and after each PA training session in S1-S17 (Author response image 2A). Also, the Rescue group similarly received CNO+L-lactate before and after each PA training session in S1-S17. Therefore, while this study design allowed us to conclude that ACC astrocytic Gi activation impairs PA learning and that exogenous L-lactate can rescue the impairment, it does not allow clear differentiation of the effects of these treatments on memory acquisition and consolidation. Further studies are needed to investigate this.

• The hypothesis that observed learning impairments could be associated with diminished mitochondrial biogenesis caused by decreased l-lactate in the result of astrocytic Gi-DREADDS stimulation is very appealing, but a few key pieces of evidence are missing. So far, the hypothesis is supported by experiments demonstrating reduced expression of several components of mitochondrial membrane ATP synthase and a decrease in relative mtDNA copy numbers in ACC of rats injected with Gi-DREADDs. L-lactate injections into ACC restored and even further increased the expression of the above-mentioned markers. Co-administration of NMDAR antagonist D-APV or MCT-2 (mostly neuronal) blocker 4-CIN with L-lactate, prevented L-lactate-induced increase in relative mtDNA copy. I am wondering how the interference with mitochondrial biogenesis is affecting neuronal physiology and if it would result in impaired PA learning or schema memory.

The observation of diminished mitochondrial biogenesis in the astrocytic Gi-activated rats that showed impaired PA learning is exciting. However, our study does not provide experimental data on how mitochondrial biogenesis could be associated with impaired PA learning and schema memory. Results from several previous studies linked mitochondrial biogenesis and its regulators such as PGC-1α and SIRT3 to diverse neuronal and cognitive functions as described in the discussion section of the manuscript. In the revised manuscript, we have provided further discussion as follows to discuss potential mechanisms:

“In this study, we have demonstrated that ACC astrocytic Gi activation impairs PA learning and schema formation, PA memory retrieval, and NPA learning and retrieval by decreasing L-lactate level in the ACC. Although we have shown that these impairments are associated with diminished expression of proteins of mitochondrial biogenesis, the precise mechanisms of how astrocytic Gi activation affects neuronal functions and schema memory remain to be elucidated. We previously demonstrated that neuronal inhibition in either the hippocampus or the ACC impairs PA learning and schema formation (Hasan et al., 2019). In another recent study (Liu et al., 2022), we showed that astrocytic Gi activation in the CA1 impaired PA training-associated CA1-ACC projecting neuronal activation. Yao et al. recently showed that reduction of astrocytic lactate dehydrogenase A (an enzyme that reversibly catalyze L-lactate production from pyruvate) in the dorsomedial prefrontal cortex reduces L-lactate levels and neuronal firing frequencies, promoting depressive-like behaviors in mice (Yao et al., 2023). These impairments could be rescued by L-lactate infusion. It is possible that the impairment in PA learning and schema observed in our study might have involved a similar functional consequence of reduced neuronal activity in the ACC neurons upon astrocytic Gi activation.

Schema consolidation is associated with synaptic plasticity-related gene expression (such as Zif268, Arc) in the ACC (Tse et al., 2011). L-lactate, after entry into neurons, can be converted to pyruvate during which NADH is also produced, promoting synaptic plasticity-related gene expression by potentiating NMDA signaling in neurons (Yang et al., 2014; Margineanu et al., 2018). Furthermore, L-lactate acts as an energy substrate to fuel learning-induced de novo neuronal translation critical for long-term memory (Descalzi et al., 2019). On the other hand, mitochondria play crucial role in fueling local translation during synaptic plasticity (Rangaraju et al., 2019). Therefore, it could be hypothesized that the rescue of astrocytic Gi activation-mediated impairment of schema by exogenous L-lactate could have been mediated by facilitating synaptic plasticity-related gene expression by directly fueling the protein translation, potentiating NMDA signaling, as well as increasing mitochondrial capacity for ATP production by promoting mitochondrial biogenesis. Furthermore, the potential involvement of HCAR1, a receptor for L-lactate that may regulate neuronal activity (Bozzo et al., 2013; Tang et al., 2014; Herrera-López & Galván, 2018; Abrantes et al., 2019), cannot be excluded. Future research could explore these potential mechanisms, examining the interactions among them, and determining their relative contributions to schema. Our previous study also showed that ACC myelination is necessary for PA learning and schema formation, and that repeated PA training is associated with oligodendrogenesis in the ACC (Hasan et al., 2019). Oligodendrocytes facilitate fast, synchronized, and energy efficient transfer of information by wrapping axons in myelin sheath. Furthermore, they supply axons with glycolysis products, such as L-lactate, to offer metabolic support (Fünfschilling et al., 2012; Lee et al., 2012). The association of oligodendrogenesis and myelination with schema memory may suggest an adaptive response of oligodendrocytes to enhance metabolic support and neuronal energy efficiency during PA learning. Given the impairments in PA learning observed in the ACC astrocytic Gi-activated rats in the current study, it is reasonable to conclude that the direct metabolic support to axons provided by oligodendrocytes is not sufficient to rescue the schema impairments caused by decreased L-lactate levels upon astrocytic Gi activation. On the other hand, L-lactate was shown to be important for oligodendrogenesis and myelination (Sánchez-Abarca et al., 2001; Rinholm et al., 2011; Ichihara et al., 2017). Therefore, it is tempting to speculate that a decrease in L-lactate level may also impede oligodendrogenesis and myelination, consequently preventing the enhanced axonal support provided by oligodendrocytes and myelin during schema learning. Recently, a study has demonstrated that upon demyelination, mitochondria move from the neuronal cell body to the demyelinated axon (Licht-Mayer et al., 2020). Enhancement of this axonal response of mitochondria to demyelination, by targeting mitochondrial biogenesis and mitochondrial transport from the cell body to axon, protects acutely demyelinated axons from degeneration. Given the connection between schema and increased myelination, it remains an open question whether L-lactate-induced mitochondrial biogenesis plays a beneficial role in schema through a similar mechanism. Nevertheless, our results contribute to the mounting evidence of the glial role in cognitive functions and underscores the new paradigm in which glial cells are considered as integral players in cognitive functions alongside neurons. Disruption of neurons, myelin, or astrocytes in the ACC can disrupt PA learning and schema memory.”

Reviewer #3 (Public Review):

Akter et al. investigated how the astroglial Gi signaling pathway in the rat anterior cingulate cortex (ACC) affects cognitive functions, in particular schema memory formation. Using a stereotactic approach they intracranially introduced AAV8 vectors carrying mCherry-tagged hM4Di DREADD (Designer Receptor Exclusively Activated by Designer Drugs) under astrocyte selective GFAP promotor (AAV8-GFAP-hM4Di-mCherry) into the AAC region of the rat brain. hM4Di DREADD is a genetically modified form of the human M4 muscarinic (hM4) receptor insensitive to endogenous acetylcholine but is activated by the inert clozapine metabolite clozapine-N-oxide (CNO), triggering the Gi signaling pathway. The authors confirmed that hM4Di DREADD is selectively expressed in astrocytes after the application of the AAV8 vector by analysing the mCherry signals and immunolabeling of astrocytes and neurons in the ACC region of the rat brain. They activated hM4Di DREADD (Gi signalling) in astrocytes by intraperitoneal administration of CNO and measured cognitive functions in animals after CNO administration. Activation of Gi signaling in astrocytes by CNO application decreased paired-associate (PA) learning, schema formation, and memory retrieval in tested animals. This was associated with a decrease in cAMP in astrocytes and L-lactate in extracellular fluid as measured by immunohistochemistry in situ and in awake rats by microdialysis, respectively. Administration of exogenous L-lactate rescued the astroglial Gi-mediated deficits in PA learning, memory retrieval, and schema formation, suggesting that activation of astroglial Gi signalling downregulates L-lactate production in astrocytes and its transport to neurons affecting memory formation. Authors also show that expression level of proteins involved in mitochondrial biogenesis, which is associated with cognitive functions, is decreased in neurons, when Gi signalling is activated in astrocytes, and rescued when exogenous L-lactate is applied, suggesting the implication of astrocyte-derived L-lactate in the maintenance of mitochondrial biogenesis in neurons. The latter depended on lactate MCT2 transporter activity and glutamate NMDA receptor activity.

The paper is very well written and discussed. The conclusions of this paper are well supported by the data. Although this is a study that uses established and previously published methodologies, it provides new insights into L-lactate signalling in the brain, particularly in AAC, and further confirms the role of astroglial L-lactate in learning and memory formation. It also raises new questions about the molecular mechanisms underlying astrocyte-derived L-lactate-mediated mitochondrial biogenesis in neurons and its contribution to schema memory formation.

• The authors discuss astrocytic L-lactate signalling without considering the recently discovered L-lactate-sensitive Gs and Gi protein-coupled receptors in the brain, which are present in both astrocytes and neurons. The use of nonendogenous L-lactate receptor agonists (Compound 2, 3-chloro-5-hydroxybenzoic acid) would clarify the implication of L-lactate receptor signalling in schema memory formation.

In the revised manuscript, we have included this point in the discussion section to mention the potential role of HCAR1 in schema memory as follows:

“Schema consolidation is associated with synaptic plasticity-related gene expression (such as Zif268, Arc) in the ACC (Tse et al., 2011). L-lactate, after entry into neurons, can be converted to pyruvate during which NADH is also produced, promoting synaptic plasticity-related gene expression by potentiating NMDA signaling in neurons (Yang et al., 2014; Margineanu et al., 2018). Furthermore, L-lactate acts as an energy substrate to fuel learning-induced de novo neuronal translation critical for long-term memory (Descalzi et al., 2019). On the other hand, mitochondria play crucial role in fueling local translation during synaptic plasticity (Rangaraju et al., 2019). Therefore, it could be hypothesized that the rescue of astrocytic Gi activation-mediated impairment of schema by exogenous L-lactate could have been mediated by facilitating synaptic plasticity-related gene expression by directly fueling the protein translation, potentiating NMDA signaling, as well as increasing mitochondrial capacity for ATP production by promoting mitochondrial biogenesis. Furthermore, the potential involvement of HCAR1, a receptor for L-lactate that may regulate neuronal activity (Bozzo et al., 2013; Tang et al., 2014; Herrera-López & Galván, 2018; Abrantes et al., 2019), cannot be excluded. Future research could explore these potential mechanisms, examining the interactions among them, and determining their relative contributions to schema.”

• The use of control animals transduced with an "empty" AAV9 vector (AAV8-GFAP-mCherry) compared with animals transduced with AAV8-GFAP-hM4Di-mCherry throughout the study would strengthen the results of this study, since transfection itself, as well as overexpression of the mCherry protein, may affect cell function.

We thank the reviewer for pointing this. The schema experiment includes a control group (Control-CNO group) of rats injected with AAV8-GFAP-mCherry bilaterally into the ACC. As shown in Author response image 3, after habituation and pretraining, these rats were trained for PA learning similarly to the other groups. Before 30 mins and after 30 mins of each PA training session, they received I.P. CNO. The PA learning, schema formation, memory retrieval, NPA learning and retrieval, and latency (time needed to commence digging at the correct well) were similar to the control group of rats. This result is consistent with our previous study where rats bilaterally injected with AAV8-GFAP-mCherry into CA1 of hippocampus did not show impairments in PA learning and schema formation upon CNO treatment (Liu et al., 2022).

Author response image 3.

A. PI (mean ± SD) during the acquisition of the original six PAs (OPAs) (S1-2, 4-8, 10-17) and new PAs (NPAs) (S19) of the control (n=6) and control-CNO (n=4) groups. B. Non-rewarded PTs (PT1, PT2, and PT3 done on S3, S9, and S18, respectively) to test memory retrieval of OPAs for the control-CNO group. C. Non-rewarded PT4 (S20) which was done after replacing two OPAs with two NPAs (NPA 7 & 8) in S19 for the control-CNO group. D. Latency (in seconds) before commencing digging at the correct well for control and control-CNO groups. Data shown as mean ± SD.

References

Abrantes, H. d. C., Briquet, M., Schmuziger, C., Restivo, L., Puyal, J., Rosenberg, N., Rocher, A.-B., Offermanns, S., & Chatton, J.-Y. (2019). The Lactate Receptor HCAR1 Modulates Neuronal Network Activity through the Activation of Gα and Gβγ Subunits. The Journal of Neuroscience, 39(23), 4422-4433. https://doi.org/10.1523/jneurosci.2092-18.2019

Akter, M., Ma, H., Hasan, M., Karim, A., Zhu, X., Zhang, L., & Li, Y. (2023). Exogenous L-lactate administration in rat hippocampus increases expression of key regulators of mitochondrial biogenesis and antioxidant defense [Original Research]. Frontiers in Molecular Neuroscience, 16. https://doi.org/10.3389/fnmol.2023.1117146

Bozzo, L., Puyal, J., & Chatton, J.-Y. (2013). Lactate Modulates the Activity of Primary Cortical Neurons through a Receptor-Mediated Pathway. PLoS One, 8(8), e71721. https://doi.org/10.1371/journal.pone.0071721

Choi, H. B., Gordon, G. R., Zhou, N., Tai, C., Rungta, R. L., Martinez, J., Milner, T. A., Ryu, J. K., McLarnon, J. G., Tresguerres, M., Levin, L. R., Buck, J., & MacVicar, B. A. (2012). Metabolic communication between astrocytes and neurons via bicarbonate-responsive soluble adenylyl cyclase. Neuron, 75(6), 1094-1104. https://doi.org/10.1016/j.neuron.2012.08.032

Covelo, A., Eraso-Pichot, A., Fernández-Moncada, I., Serrat, R., & Marsicano, G. (2021). CB1R-dependent regulation of astrocyte physiology and astrocyte-neuron interactions. Neuropharmacology, 195, 108678. https://doi.org/https://doi.org/10.1016/j.neuropharm.2021.108678

Descalzi, G., Gao, V., Steinman, M. Q., Suzuki, A., & Alberini, C. M. (2019). Lactate from astrocytes fuels learning-induced mRNA translation in excitatory and inhibitory neurons. Communications Biology, 2(1), 247. https://doi.org/10.1038/s42003-019-0495-2

Endo, F., Kasai, A., Soto, J. S., Yu, X., Qu, Z., Hashimoto, H., Gradinaru, V., Kawaguchi, R., & Khakh, B. S. (2022). Molecular basis of astrocyte diversity and morphology across the CNS in health and disease. Science, 378(6619), eadc9020. https://doi.org/10.1126/science.adc9020

Fünfschilling, U., Supplie, L. M., Mahad, D., Boretius, S., Saab, A. S., Edgar, J., Brinkmann, B. G., Kassmann, C. M., Tzvetanova, I. D., Möbius, W., Diaz, F., Meijer, D., Suter, U., Hamprecht, B., Sereda, M. W., Moraes, C. T., Frahm, J., Goebbels, S., & Nave, K.-A. (2012). Glycolytic oligodendrocytes maintain myelin and long-term axonal integrity. Nature, 485(7399), 517-521. https://doi.org/10.1038/nature11007

Harris, R. A., Lone, A., Lim, H., Martinez, F., Frame, A. K., Scholl, T. J., & Cumming, R. C. (2019). Aerobic Glycolysis Is Required for Spatial Memory Acquisition But Not Memory Retrieval in Mice. eNeuro, 6(1). https://doi.org/10.1523/ENEURO.0389-18.2019

Hasan, M., Kanna, M. S., Jun, W., Ramkrishnan, A. S., Iqbal, Z., Lee, Y., & Li, Y. (2019). Schema-like learning and memory consolidation acting through myelination. FASEB J, 33(11), 11758-11775. https://doi.org/10.1096/fj.201900910R

Herrera-López, G., & Galván, E. J. (2018). Modulation of hippocampal excitability via the hydroxycarboxylic acid receptor 1. Hippocampus, 28(8), 557-567. https://doi.org/https://doi.org/10.1002/hipo.22958

Horvat, A., Muhič, M., Smolič, T., Begić, E., Zorec, R., Kreft, M., & Vardjan, N. (2021). Ca2+ as the prime trigger of aerobic glycolysis in astrocytes. Cell Calcium, 95, 102368. https://doi.org/https://doi.org/10.1016/j.ceca.2021.102368

Horvat, A., Zorec, R., & Vardjan, N. (2021). Lactate as an Astroglial Signal Augmenting Aerobic Glycolysis and Lipid Metabolism [Review]. Frontiers in Physiology, 12. https://doi.org/10.3389/fphys.2021.735532

Ichihara, Y., Doi, T., Ryu, Y., Nagao, M., Sawada, Y., & Ogata, T. (2017). Oligodendrocyte Progenitor Cells Directly Utilize Lactate for Promoting Cell Cycling and Differentiation. J Cell Physiol, 232(5), 986-995. https://doi.org/10.1002/jcp.25690

Iqbal, Z., Liu, S., Lei, Z., Ramkrishnan, A. S., Akter, M., & Li, Y. (2023). Astrocyte L-Lactate Signaling in the ACC Regulates Visceral Pain Aversive Memory in Rats. Cells, 12(1), 26. https://www.mdpi.com/2073-4409/12/1/26

Jourdain, P., Rothenfusser, K., Ben-Adiba, C., Allaman, I., Marquet, P., & Magistretti, P. J. (2018). Dual action of L-Lactate on the activity of NR2B-containing NMDA receptors: from potentiation to neuroprotection. Sci Rep, 8(1), 13472. https://doi.org/10.1038/s41598-018-31534-y

Kofuji, P., & Araque, A. (2021). G-Protein-Coupled Receptors in Astrocyte-Neuron Communication. Neuroscience, 456, 71-84. https://doi.org/10.1016/j.neuroscience.2020.03.025

Lee, Y., Morrison, B. M., Li, Y., Lengacher, S., Farah, M. H., Hoffman, P. N., Liu, Y., Tsingalia, A., Jin, L., Zhang, P. W., Pellerin, L., Magistretti, P. J., & Rothstein, J. D. (2012). Oligodendroglia metabolically support axons and contribute to neurodegeneration. Nature, 487(7408), 443-448. https://doi.org/10.1038/nature11314

Licht-Mayer, S., Campbell, G. R., Canizares, M., Mehta, A. R., Gane, A. B., McGill, K., Ghosh, A., Fullerton, A., Menezes, N., Dean, J., Dunham, J., Al-Azki, S., Pryce, G., Zandee, S., Zhao, C., Kipp, M., Smith, K. J., Baker, D., Altmann, D., Anderton, S. M., Kap, Y. S., Laman, J. D., Hart, B. A. t., Rodriguez, M., Watzlawick, R., Schwab, J. M., Carter, R., Morton, N., Zagnoni, M., Franklin, R. J. M., Mitchell, R., Fleetwood-Walker, S., Lyons, D. A., Chandran, S., Lassmann, H., Trapp, B. D., & Mahad, D. J. (2020). Enhanced axonal response of mitochondria to demyelination offers neuroprotection: implications for multiple sclerosis. Acta Neuropathologica, 140(2), 143-167. https://doi.org/10.1007/s00401-020-02179-x

Liu, S., Wong, H. Y., Xie, L., Iqbal, Z., Lei, Z., Fu, Z., Lam, Y. Y., Ramkrishnan, A. S., & Li, Y. (2022). Astrocytes in CA1 modulate schema establishment in the hippocampal-cortical neuron network. BMC Biol, 20(1), 250. https://doi.org/10.1186/s12915-022-01445-6

Magistretti, P. J., & Allaman, I. (2018). Lactate in the brain: from metabolic end-product to signalling molecule. Nat Rev Neurosci, 19(4), 235-249. https://doi.org/10.1038/nrn.2018.19

Margineanu, M. B., Mahmood, H., Fiumelli, H., & Magistretti, P. J. (2018). L-Lactate Regulates the Expression of Synaptic Plasticity and Neuroprotection Genes in Cortical Neurons: A Transcriptome Analysis. Front Mol Neurosci, 11, 375. https://doi.org/10.3389/fnmol.2018.00375

Netzahualcoyotzi, C., & Pellerin, L. (2020). Neuronal and astroglial monocarboxylate transporters play key but distinct roles in hippocampus-dependent learning and memory formation. Progress in Neurobiology, 194, 101888. https://doi.org/https://doi.org/10.1016/j.pneurobio.2020.101888

Newman, L. A., Korol, D. L., & Gold, P. E. (2011). Lactate produced by glycogenolysis in astrocytes regulates memory processing. PLoS One, 6(12), e28427. https://doi.org/10.1371/journal.pone.0028427

Park, J., Kim, J., & Mikami, T. (2021). Exercise-Induced Lactate Release Mediates Mitochondrial Biogenesis in the Hippocampus of Mice via Monocarboxylate Transporters. Front Physiol, 12, 736905. https://doi.org/10.3389/fphys.2021.736905

Peterson, S. M., Pack, T. F., & Caron, M. G. (2015). Receptor, Ligand and Transducer Contributions to Dopamine D2 Receptor Functional Selectivity. PLoS One, 10(10), e0141637. https://doi.org/10.1371/journal.pone.0141637

Rangaraju, V., Lauterbach, M., & Schuman, E. M. (2019). Spatially Stable Mitochondrial Compartments Fuel Local Translation during Plasticity. Cell, 176(1), 73-84.e15. https://doi.org/10.1016/j.cell.2018.12.013

Rinholm, J. E., Hamilton, N. B., Kessaris, N., Richardson, W. D., Bergersen, L. H., & Attwell, D. (2011). Regulation of oligodendrocyte development and myelination by glucose and lactate. J Neurosci, 31(2), 538-548. https://doi.org/10.1523/JNEUROSCI.3516-10.2011

Sánchez-Abarca, L. I., Tabernero, A., & Medina, J. M. (2001). Oligodendrocytes use lactate as a source of energy and as a precursor of lipids. Glia, 36(3), 321-329. https://doi.org/10.1002/glia.1119

Suzuki, A., Stern, S. A., Bozdagi, O., Huntley, G. W., Walker, R. H., Magistretti, P. J., & Alberini, C. M. (2011). Astrocyte-neuron lactate transport is required for long-term memory formation. Cell, 144(5), 810-823.

Tang, F., Lane, S., Korsak, A., Paton, J. F. R., Gourine, A. V., Kasparov, S., & Teschemacher, A. G. (2014). Lactate-mediated glia-neuronal signalling in the mammalian brain. Nature Communications, 5(1), 3284. https://doi.org/10.1038/ncomms4284

Tauffenberger, A., Fiumelli, H., Almustafa, S., & Magistretti, P. J. (2019). Lactate and pyruvate promote oxidative stress resistance through hormetic ROS signaling. Cell Death Dis, 10(9), 653. https://doi.org/10.1038/s41419-019-1877-6

Tse, D., Langston, R. F., Kakeyama, M., Bethus, I., Spooner, P. A., Wood, E. R., Witter, M. P., & Morris, R. G. (2007). Schemas and memory consolidation. Science, 316(5821), 76-82. https://doi.org/10.1126/science.1135935

Tse, D., Takeuchi, T., Kakeyama, M., Kajii, Y., Okuno, H., Tohyama, C., Bito, H., & Morris, R. G. (2011). Schema-dependent gene activation and memory encoding in neocortex. Science, 333(6044), 891-895. https://doi.org/10.1126/science.1205274

Vardjan, N., Chowdhury, H. H., Horvat, A., Velebit, J., Malnar, M., Muhič, M., Kreft, M., Krivec, Š. G., Bobnar, S. T., Miš, K., Pirkmajer, S., Offermanns, S., Henriksen, G., Storm-Mathisen, J., Bergersen, L. H., & Zorec, R. (2018). Enhancement of Astroglial Aerobic Glycolysis by Extracellular Lactate-Mediated Increase in cAMP [Original Research]. Frontiers in Molecular Neuroscience, 11. https://doi.org/10.3389/fnmol.2018.00148

Vezzoli, E., Cali, C., De Roo, M., Ponzoni, L., Sogne, E., Gagnon, N., Francolini, M., Braida, D., Sala, M., Muller, D., Falqui, A., & Magistretti, P. J. (2020). Ultrastructural Evidence for a Role of Astrocytes and Glycogen-Derived Lactate in Learning-Dependent Synaptic Stabilization. Cereb Cortex, 30(4), 2114-2127. https://doi.org/10.1093/cercor/bhz226

Wang, J., Tu, J., Cao, B., Mu, L., Yang, X., Cong, M., Ramkrishnan, A. S., Chan, R. H. M., Wang, L., & Li, Y. (2017). Astrocytic l-Lactate Signaling Facilitates Amygdala-Anterior Cingulate Cortex Synchrony and Decision Making in Rats. Cell Rep, 21(9), 2407-2418. https://doi.org/10.1016/j.celrep.2017.11.012

Yang, J., Ruchti, E., Petit, J. M., Jourdain, P., Grenningloh, G., Allaman, I., & Magistretti, P. J. (2014). Lactate promotes plasticity gene expression by potentiating NMDA signaling in neurons. Proc Natl Acad Sci U S A, 111(33), 12228-12233. https://doi.org/10.1073/pnas.1322912111

Yao, S., Xu, M.-D., Wang, Y., Zhao, S.-T., Wang, J., Chen, G.-F., Chen, W.-B., Liu, J., Huang, G.-B., Sun, W.-J., Zhang, Y.-Y., Hou, H.-L., Li, L., & Sun, X.-D. (2023). Astrocytic lactate dehydrogenase A regulates neuronal excitability and depressive-like behaviors through lactate homeostasis in mice. Nature Communications, 14(1), 729. https://doi.org/10.1038/s41467-023-36209-5

Yu, X., Zhang, R., Wei, C., Gao, Y., Yu, Y., Wang, L., Jiang, J., Zhang, X., Li, J., & Chen, X. (2021). MCT2 overexpression promotes recovery of cognitive function by increasing mitochondrial biogenesis in a rat model of stroke. Anim Cells Syst (Seoul), 25(2), 93-101. https://doi.org/10.1080/19768354.2021.1915379

Zhou, Z., Okamoto, K., Onodera, J., Hiragi, T., Andoh, M., Ikawa, M., Tanaka, K. F., Ikegaya, Y., & Koyama, R. (2021). Astrocytic cAMP modulates memory via synaptic plasticity. Proc Natl Acad Sci U S A, 118(3), e2016584118. https://doi.org/10.1073/pnas.2016584118

Zhu, J., Hu, Z., Han, X., Wang, D., Jiang, Q., Ding, J., Xiao, M., Wang, C., Lu, M., & Hu, G. (2018). Dopamine D2 receptor restricts astrocytic NLRP3 inflammasome activation via enhancing the interaction of β-arrestin2 and NLRP3. Cell Death Differ, 25(11), 2037-2049. https://doi.org/10.1038/s41418-018-0127-2

Author Response

Reviewer #1 (Public Review):

This thorough study expands our understanding of BMP signaling, a conserved developmental pathway, involved in processes diverse such as body patterning and neurogenesis. The authors applied multiple, state-of-art strategies to the anthozoan Nematostella vectensis in order to first identify the direct BMP signaling targets - bound by the activated pSMAD1/5 protein - and then dissect the role of a novel pSMAD1/5 gradient modulator, zwim4-6. The list of target genes features multiple developmental regulators, many of which are bilaterally expressed, and which are notably shared between Drosophila and Xenopus. The analysis identified in particular zswim4-6 a novel nuclear modulator of the BMP pathway conserved also in vertebrates. A combination of both loss-of-function (injection of antisense morpholino oligonucleotide, CRISPR/Cas9 knockout, expression of dominant negative) and gain-of-function assays, and of transcriptome sequencing identified that zwim acts as a transcriptional repression of BMP signaling. Functional manipulation of zswim5 in zebrafish shows a conserved role in modulating BMP signaling in a vertebrate.

The particular strength of the study lies in the careful and thorough analysis performed. This is solid developmental work, where one clear biological question is progressively dissected, with the most appropriate tools. The functional results are further validated by alternative approaches. Data is clearly presented and methods are detailed. I have a couple of comments.

1) I was intrigued - as the authors - by the fact that the ChiP-Seq did not identify any known BMP ligand bound by pSMAD1/5. Are these genes found in the published ChiP-Seq data of the other species used for the comparative analysis? One hypothesis could be that there is a change in the regulatory interactions and that the initial set-up of the gradient requires indeed a feedback loop, which is then turned off at later gastrula. In this case, immunoprecipitation at early gastrula, prior to the set-up of the pSMAD1/5 gradient, could reveal a different scenario. Alternately, the regulation could be indirect, for example, through RGM, an additional regulator of BMP signaling expressed on the side of lower BMP activity, which is among the targets of the ChiP-Seq. This aspect could be discussed. Additionally, even if this is perhaps outside the scope of this study, I think it would be informative to further assess the effect of ZSWIM manipulation on RGM (and vice versa).

Indeed, BMP genes are direct BMP signaling targets in Drosophila (dpp) (Deignan et al., 2016, https://doi.org/10.1371/journal.pgen.1006164) and frog (bmp2, bmp4, bmp5, bmp7) (Stevens et al., 2021, https://doi.org/10.1242/dev.145789). Of all these ligands, only the dorsally expressed Xenopus bmp2 is repressed by BMP signaling, while another dorsally expressed Xenopus BMP gene admp is not among the direct targets. All other BMP genes listed here are expressed in the pMad/pSMAD1/5/8-positive domain and are activated by BMP signaling.

In Nematostella, we do not find BMP genes among the ChIP-Seq targets, but this is not that surprising considering the dynamics of the bmp2/4, bmp5-8 and chordin expression, as well as the location of the pSMAD1/5-positive cells. In late gastrulae/early planulae, Chordin appears to be shuttling BMP2/4 and BMP5-8 away from their production source and over to the gdf5-like side of the directive axis (Genikhovich et al., 2015; Leclere and Rentsch, 2014). By 4 dpf, chordin expression stops, and BMP2/4 and BMP5-8 start to be both expressed AND signal in the mesenteries. If bmp2/4 and bmp5-8 expression were directly suppressed by pSMAD1/5 (as is the case chordin or rgm expression), this mesenterial expression would not be possible. Therefore, in our opinion, it is most likely that at late gastrula and early planula the regulation of bmp2/4 and bmp5-8 expression by BMP signaling is indirect. We do not have an explanation for why gdf5-like (another BMP gene expressed on the “high pSMAD1/5” side) is not retrieved as a direct BMP target in our ChIP data. Since we do not understand well enough how BMP gene expression is regulated, we do not discuss this at length in the manuscript.

As the Reviewer suggested, we analyzed the effect of ZSWIM4-6 KD on the expression of rgm. Expectedly, since it is expressed on the “low BMP side”, its expression was strongly expanded (Figure 6 - Figure Supplement 4)

2) I do not fully understand the rationale behind the choice of performing the comparative assays in zebrafish: as the conservation was initially identified in Xenopus, I would have expected the experiment to be performed in frog. Furthermore, reading the phylogeny (Figure 4A), it is not obvious to me why ZSWIM5 was chosen for the assay (over the other paralog ZSWIM6). Could the Authors comment on this experiment further?

The comparison was done in zebrafish because we were planning to generate zswim5 mutants, whose analysis is currently in progress. ZSWIM6 is not expressed at the developmental stages we were interested in, while ZSWIM5 was, based on available zebrafish expression data (White et al., 2017):

Reviewer #2 (Public Review):

The authors provide a nice resource of putative direct BMP target genes in Nematostella vectensis by performing ChIP-seq with an anti-pSmad1/5 antibody, while also performing bulk RNA-seq with BMP2/4 or GDF5 knockdown embryos. Genes that exhibit pSmad1/5 binding and have changes in transcription levels after BMP signaling loss were further annotated to identify those with conserved BMP response elements (BREs). Further characterization of one of the direct BMP target genes (zswim4-6) was performed by examining how expression changed following BMP receptor or ligand loss of function, as well as how loss or gain of function of zswim4-6 affected development and BMP signaling. The authors concluded that zswim4-6 modulates BMP signaling activity and likely acts as a pSMAD1/5 dependent co-repressor. However, the mechanism by which zswim4-6 affects the BMP gradient or interacts with pSMAD1/5 to repress target genes is not clear. The authors test the activity of a zswim4-6 homologue in zebrafish (zswim5) by over-expressing mRNA and find that pSMAD1/5/9 labeling is reduced and that embryos have a phenotype suggesting loss of BMP signaling, and conclude that zswim4-6 is a conserved regulator of BMP signaling. This conclusion needs further support to confirm BMP loss of function phenotypes in zswim5 over-expression embryos.

Major comments

1) The BMP direct target comparison was performed between Nematostella, Drosophila, and Xenopus, but not with existing data from zebrafish (Greenfeld 2021, Plos Biol). Given the functional analysis with zebrafish later in the paper it would be nice to see if there are conserved direct target genes in zebrafish, and in particular, is zswim5 (or other zswim genes) are direct targets. Since conservation of zswim4-6 as a direct BMP target between Nematostella and Xenopus seemed to be part of the rationale for further functional analysis, it would also be nice to know if this is a conserved target in zebrafish.

Thank you for the suggestion. In the paper by Greenfeld et al., 2021, zebrafish zswim5 was downregulated approximately 2.4x in the bmp7 mutant at 6 hpf, while zswim6 was barely expressed and not affected at this stage. We added this information to the text of the manuscript. Expression of several other zebrafish zswim genes was also affected in the bmp7 mutant, but these genes do not appear relevant for our study since their corresponding orthologs are not identified as pSMAD1/5 ChIP-Seq targets in Nematostella. Notably, zebrafish zzswim5 is not clearly differentially expressed in BMP or Chd overexpression conditions (See Supplementary file 1 in Rogers et al. 2020). Importantly, in the paper, we wanted to compare ChiP-Seq data with ChIP-Seq data, however, unfortunately, no ChIP-Seq data for pSMAD1/5/8 is currently available for zebrafish, thus precluding comparisons.

Related to this, in the discussion it is mentioned that zswim4/6 is also a direct BMP target in mouse hair follicle cells, but it wasn't obvious from looking at the supplemental data in that paper where this was drawn from.

Please see Supplementary Table 1, second Excel sheet labeled “Mx ChIP_Seq” in Genander et al., 2014, https://doi.org/10.1016/j.stem.2014.09.009. Zswim4 has a single pSMAD1 peak associated with it, Zswim6 has two.

2) The loss of zswim4-6 function via MO injection results in changes to pSmad1/5 staining, including a reduction in intensity in the endoderm and gain of intensity in the ectoderm, while over-expression results in a loss of intensity in the ectoderm and no apparent change in the endoderm. While this is interesting, it is not clear how zswim4-6 is functioning to modify BMP signaling, and how this might explain differential effects in ectoderm vs. endoderm. Is the assumption that the mechanism involves repression of chordin? And if so one could test the double knockdown of zswim4-6 and chordin and look for the rescue of pSad1/5 levels or morphological phenotype.

We do not think that the mechanism of the ZSWIM4-6 action is via repression of Chordin. As loss of chordin leads to the loss of pSMAD1/5 in Nematostella (Genikhovich et al., 2015), the proposed experiment is, unfortunately, not feasible to test this hypothesis. Currently, we see two distinct effects of the modulation of zswim4-6 expression. First, it affects the pSMAD1/5 gradient, possibly by destabilizing nuclear SMAD1/5, as has been proposed by Wang et al., 2022 for the vertebrate Zswim4. This is in line with our results shown on Fig. 6C-F’ and Fig. 6-Figure supplement 3. In our opinion, the reaction of the genes expressed on the “high BMP” side of the directive axis to the overexpression or KD of ZSWIM4-6 (Fig. 6I-K’, 6N-P’) can be explained by these changes in the pSMAD1/5 signaling intensity. Secondly, zswim4-6 appears to promote pSMAD1/5-mediated gene repression. This is in line with the reaction of the genes expressed on the “low BMP” side of the directive axis (Fig. 6G-H’, 6L-M’, Fig. 6-Figure Supplement 4). These genes are repressed by BMP signaling, but they expand their expression upon zswim4-6 KD in spite of the increased pSMAD1/5. Our ChiP experiment (Fig. 6Q) supports this view.

3) Several experiments are done to determine how zswim4-6 expression responds to the loss of function of different BMP ligands and receptors, with the conclusion being that swim4-6 is a BMP2/4 target but not a GDF5 target, with a lot of the discussion dedicated to this as well. However, the authors show a binary response to the loss of BMP2/4 function, where zswim4-6 is expressed normally until pSmad1/5 levels drop low enough, at which point expression is lost. Since the authors also show that GDF5 morphants do not have as strong a reduction in pSmad1/5 levels compared to BMP2/4 morphants, perhaps GDF5 plays a positive but redundant role in swim4-6 expression. To test this possibility the authors could inject suboptimal doses of BMP2/4 MO with GDF5 MO and look for synergy in the loss of zswim4-6 expression.

Thanks for this great suggestion! We performed this experiment (Fig. 5H’’-L) and indeed, a suboptimal dose of BMP2/4MO + GDF5lMO results in a complete radialization of the embryo and abolished zswim4–6, similar to the effect of a high dose of BMP2/4. This result suggests that rather than being a ligand-specific signaling function, GDF5-like signaling alone still provides sufficiently high pSmad1/5 levels to activate zswim4-6 expression to apparent wildtype levels, demonstrating the sensitivity of this gene to even very low amounts of BMP signaling.

4) The zswim4-6 morphant embryos show increased expression of zswim4-6 mRNA, which is said to indicate that zswim4-6 negatively regulates its own expression. However in zebrafish translation blocking MOs can sometimes stabilize target transcripts, causing an artifact that can be mistakenly assumed to be increased transcription (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7162184/). Some additional controls here would be warranted for making this conclusion.

Thanks for raising this important experimental consideration. To-date, we do not have any evidence for MO-mediated transcript stabilization in Nematostella, and we have not found such data in the literature on models other than zebrafish. mRNA stabilization by the MO also seemed unlikely because we were unable to KD zswim4-6 using several independent shRNAs - an effect we frequently observe with genes, whose activity negatively regulates their own expression. However, to test the possibility that zswim4-6MO binding stabilizes zswim4-6 mRNA, we injected mRNA containing the zswim4-6MO recognition sequence followed by the mCherry coding sequence (zswim4-6MO-mCherrry) with either zswim4-6MO or control MO. We could clearly detect mCherry fluorescence at 1 dpf if control MO was co-injected with the mRNA, but not if zswim4-6MO was coninjected with the mRNA. At 2 dpf (the stage at which we showed upregulation of zswim4-6 upon zswim4-6MO injection on Fig. 6I-I’), zswim4-6MO-mCherrry mRNA was undetectable by in situ hybridization with our standard FITC-labeled mCherry probe independent of whether zswim4-6MO-mCherrry mRNA was co-injected with the control MO or ZSWIM4-6MO, while hybridization with the FITC-labeled FoxA probe worked perfectly.

Author response image 1.

We are currently offering two alternative hypothesis for the observed increase in zswim4-6 levels in the paper rather than stating explicitly that ZSWIM4-6 negatively regulates its own expression: “The KD of zswim4-6 translation resulted in a strong upregulation of zswim4-6 transcription, especially in the ectoderm, suggesting that ZSWIM4-6 might either act as its own transcriptional repressor or that zswim4-6 transcription reacts to the increased ectodermal pSMAD1/5 (Fig. 6I-I’).” Given the sensitivity of zswim4-6 to even the weakest pSMAD1/5 signal (zswim4/6 is expressed upon GDF5-like KD, which drastically reduces pSMAD1/5 signaling intensity (see Fig. 1 and 2 in Genikhovich et al., 2015, http://doi.org/10.1016/j.celrep.2015.02.035 and Fig. 6-Figure supplement 3 of this paper), the latter option (that it reacts to the increased ectodermal pSMAD1/5) is, in our opinion, clearly the more probable one.

5) Zswim4-6 is proposed to be a co-repressor of pSmad1/5 targets based on the occupancy of zswim4-6 at the chordin BRE (which is normally repressed by BMP signaling) and lack of occupancy at the gremlin BRE (normally activated by BMP signaling). This is a promising preliminary result but is based only on the analysis of two genes. Since the authors identified BREs in other direct target genes, examining more genes would better support the model.

We suggest that ZSWIM4-6 may be a co-repressor of pSMAD1/5 targets because it is a nuclear protein (Fig. 4G), whose knockdown results in the expansion of the ectodermal expression of several genes repressed by pSMAD1/5 in spite of the expansion of pSMAD1/5 itself (Fig. 6G-H’, 6L-M’, Fig. 6-Figure Supplement 4). Our limited ChIP analysis supports this idea by showing that ZSWIM4-6 is bound to the pSMAD1/5 site of chordin (repressed by pSMAD1/5) but not on gremlin (activated by pSMAD1/5). We agree that adding the analysis of more targets in order to challenge our hypothesis would be good. However, given technical limitations (having to inject many thousands of eggs with the EF1a::ZSWIM4-6-GFP plasmid in order to get enough nuclei to extract sufficient immunoprecipitated chromatin for qPCR on 3 genes (chordin, gremlin, GAPDH) for each biological replicate, it is currently unfortunately not feasible to test more genes. It will be of great interest for follow up studies to generate a knock-in line with tagged zswim4-6 to analyze target binding on a genome-wide scale. We stress in the discussion that currently the power of our conclusion is low.

6) The rationale for further examination of zswim4-6 function in Nematostella was based in part on it being a conserved direct BMP target in Nematostella and Xenopus. The analysis of zebrafish zswim5 function however does not examine whether zswim5 is a BMP target gene (direct or indirect). BMP inhibition followed by an in situ hybridization for zswim5 would establish whether its expression is activated downstream of BMP.

In the paper by Greenfeld et al., 2021, zebrafish zswim5 was downregulated approximately 2.4x in the bmp7 mutant at 6 hpf. However, this gene was not among the 57 genes, which were considered to be direct BMP targets because their expression was affected by bmp7 mRNA injection into cycloheximide-treated bmp7 mutants (Greenfeld et al., 2021). We added this information to the text of the manuscript.

7) Although there is a reduction in pSmad1/5/9 staining in zebrafish injected with zswim5 mRNA, it is difficult to tell whether the resulting morphological phenotypes closely resemble zebrafish with BMP pathway mutations (such as bmp2b). More analysis is warranted here to determine whether stereotypical BMP loss of function phenotypes are observed, such as dorsalization of the mesoderm and loss of ventral tail fin.

We agree, and we have tuned down all zebrafish arguments. Analyses of zswim5 mutants are currently ongoing.

Author Response

Reviewer #3 (Public Review):

1) Validation of reagents: The authors generated a pY1230 Afadin antibody claiming that (page 6) "this new antibody is specific to tyrosine phosphorylated Afadin, and that pY1230 is targeted for dephosphorylation by PTPRK, in a D2-domain dependent manner". The WB in Fig 1B shows a lot of background, two main bands are visible which both diminish in intensity in ICT WT pervanadate-treated MCF10A cell lysates. The claim that the developed peptide antibody is selective for pY1230 in Afadin would need to be substantiated, for instance by pull down studies analysed by pY-MS to substantiate a claim of antibody specificity for this site. However, for the current study it would be sufficient to demonstrate that pY1230 is indeed the dephosphorylated site. I suggest therefore including a site directed mutant (Y1230F) that would confirm dephosphorylation at this site and the ability of the antibody recognizing the phosphorylation state at this position.

We would like this antibody to be a useful and freely accessible tool in the field and have taken on board the request for additional validation. To this end we have significantly expanded Supplementary Figure 2 (now Figure 1 - figure supplement 2) and included a dedicated section of the results as follows: 1. We have now included information about all of the Afadin antibodies used in this study, since Afadin(BD) appears to be sensitive to phosphorylation (Figure 1 - figure supplement 2A). 2. We have demonstrated that the Afadin pY1230 antibody detects an upregulated band in PTPRK KO MCF10A cells, consistent with our previous tyrosine phosphoproteomics (Figure 1 - figure supplement 2B). This indicates that the antibody can be used to detect endogenous Afadin phosphorylation. 3. We have included two new knock down experiments demonstrating the recognition of Afadin by our antibody (Figure 1 - figure supplement 2C). There appear to be two Afadin isoforms recognised in HEK293T cells by both the BD and pY1230 antibody, consistent with previous reports (Umeda et al. MBoC, 2015). We have highlighted these in the figure. 4. We have performed mutagenesis to demonstrate the specificity of the antibody. We tagged Afadin with a fluorescent protein tag, reasoning that it would cause a shift in molecular weight that could be resolved by SDS PAGE, as is the case. We noted that the phosphopeptide used spans an additional tyrosine, Y1226, which has been detected as phosphorylated (although to a much lower extent than Y1230) on Phosphosite plus. The data clearly show that Afadin cannot be phosphorylated when Y1230 is mutated to a phenylalanine (compared to CIP control), indicating that this is the predominant site recognised by the antibody. In addition, the endogenous pervanadate-stimulated signal is completely abolished by CIP treatment (Figure 1 - figure supplement 2D). 5. We have included densitometric quantification of the dephosphorylation assay shown in Figure 1B, which was part of a time course and shows preferential dephosphorylation by the PTPRK ICD compared to the PTPRK D1. The signal stops declining with time, which could indicate antibody background, or an inaccessible pool of Afadin-pY1230 (Figure 1 - figure supplement 2E). 6. To further demonstrate that this site is modulated by PTPRK in post-confluent cells, we have used doxycycline (dox)-inducible cell lines generated in Fearnley et al, 2019. Upon treatment with 500 ng/ml Dox for 48 hours PTPRK is induced to lower levels than wildtype, however, normalized quantification of the Afadin pY1230 against the Afadin (CST) signal clearly indicates downregulation by PTPRK WT, but not the catalytically inactive mutant (Figure 1 - figure supplement 2F and 2G). Together these data strengthen our assertion that this antibody recognises endogenously phosphorylated Afadin at site Y1230, which is modulated in vitro and in cells by PTPRK phosphatase activity. For clarity, we have highlighted and annotated the relevant bands in figures. We have also included identifiers for each Afadin total antibody was used in particular experiments.

2) The authors claim that a short, 63-residue predicted coiled coil (CC) region, is both necessary and sufficient for binding to the PTPRK-ICD. The region is predicted to have alpha-helical structure and as a consequence, a helical structure has been used in the docking model. Considering that the authors recombinantly expressed this region in bacteria, it would be experimentally simple confirming the alpha-helical structure of the segment by CD or NMR spectroscopy.

To clarify, the helical structure in the docking model was independently predicted by several sequence and structural analysis programmes including AlphaFold2, RobettaFold, NetSurfP and as annotated in Uniprot (as a coiled coil). We did not stipulate prior to the AF2 prediction that it was helical. Isolated short peptides frequently adopt helical structure, therefore prediction of a helix within the context of the full Afadin sequence is, in our opinion, stronger evidence than CD of an isolated fragment.

3) Only two mutants have been introduced into PTPRK-ICD to map the Afadin interaction site. One of the mutations changes a possibly structurally important residues (glycine) into a histidine. Even though this residue is present in PTPRM, it does not exclude that the D2 domain no longer functionally folds. Also the second mutation represents a large change in chemical properties and the other 2 predicted residues have not been investigated.

The residues that were selected for mutation are all localised to the protein surface and therefore are unlikely to be involved in stable folding of PTPRK. In support of the correct folding of the mutated PTPRK, we include in Figure 1 below SEC elution traces for wild-type and mutant D2 showing that they elute as single symmetric peaks at the same elution volume as the WT protein. This is consistent with them having a similar shape and size, and not being aggregated or unfolded.

Figure 1. PTPRK-D2 wild-type and mutant preparative SEC elution profiles. A280nm has been normalised to help illustrate that the different proteins elute at the same volume. The main peak from these samples was used for binding assays in the main paper.

Furthermore, the yield for the double mutant was very high (4 mg of pure protein from a 2 L culture, see A280 value in graph below), whereas poorly folded proteins tend to have significantly reduced yields. This protein was also very stable over time whereas unfolded proteins tend to degrade during or following purification.

Figure 2. Analytical SEC elution profile for the PTPRK-D2 DM construct showing the very high yield consistent with a well-folded, stable protein.

Finally, we have carried out thermal melt curves of the WT and mutant PTPRK D2 domains showing that they all possess melting temperatures between 39.3°C and 41.7°C, supporting that they are all equivalently folded. We include these data as an additional Supplementary Figure (Figure 4 - figure supplement 3) in the paper.

4) The interface on the Afadin substrate has not been investigated apart from deleting the entire CC or a central charge cluster. Based on the docking model the authors must have identified key positions of this interaction that could be mutated to confirm the proposed interaction site.

We have now made and tested several additional mutations within both the Afadin-CC and PTPRK-D2 domains to further validate the AF2 predicted model of the complex.

For Afadin-CC we introduced several single and double mutations along the helix including residues predicted to be in the interface and residues distal from the interface. These mutations and the pulldown with PTPRK are described in the text and are included as additional panels to a modified Figure 3. All mutations have the expected effect on the interaction based on the predicted complex structure. To help illustrate the positions of these mutations we have also included a figure of the interface with the residues highlighted.

For the PTPRK-D2 we have also introduced two new mutations, one buried in the interface (F1225A) and one on the edge of the interface encompassing a loop that is different in PTPRM (labelled the M-loop). GST-Afadin WT protein was bound to GSH beads and tested for their ability to pulldown WT and mutated PTPRK. These new mutations (illustrated in the new Figure 4 – figure supplement 2) further support the model prediction. F1225A almost completely abolishes binding as predicted, while the M-loop retains binding. These mutations and their effects are now described in the main text and the pull-down data, including controls and retesting of the original DM mutant, are included as panel H in a newly modified Figure 4 focussed solely on the PTPRK interface.

5) A minor point is that ITC experiments have not been run long enough to determine the baseline of interaction heats. In addition, as large and polar proteins were used in this experiment, a blank titration would be required to rule out that dilution heats effect the determined affinities.

All control experiments including buffer into buffer, Afadin into buffer and buffer into PTPRK were carried out at the same time as the main binding experiment and are shown below overlaid with the binding curve. These demonstrate the very small dilution heats consistent with excellent buffer matching of the samples.

We were able to obtain excellent fits to the titration curves by fitting 1:1 binding with a calculated linear baseline (see Figure 2B,D). Very similar results were obtained by fitting to the sum (‘composite’) of fitted linear baselines obtained for the three control experiments for each titration.

Author Response

Reviewer #1 (Public Review):

The authors convincingly show in this study the effects of the fas5 gene on changes in the CHC profile and the importance of these changes toward sexual attractiveness.

The main strength of this study lies in its holistic approach (from genes to behaviour) showing a full and convincing picture of the stated conclusions. The authors succeeded in putting a very interdisciplinary set of experiments together to support the main claims of this manuscript.

We appreciate the kind comments from the reviewer.

The main weakness stems from the lack of transparency behind the statistical analyses conducted in the study. Detailed statistical results are never mentioned in the text, nor is it always clear what was compared to what. I also believe that some tests that were conducted are not adequate for the given data. I am therefore unable to properly assess the significance of the results from the presented information. Nevertheless, the graphical representations are convincing enough for me to believe that a revision of the statistics would not significantly affect the main conclusions of this manuscript.

We apologize for neglecting a detailed description of statistical tests that were performed. We wrote additional paragraphs in the method part specifically explaining the statistical analyses (line 435-445; 489-502; 559-561; 586-591).

The second major problem I had with the study was how it brushes over the somewhat contradicting results they found in males (Fig S2). These are only mentioned twice in the main text and in both cases as being "similarly affected", even though their own stats seem to indicate otherwise for many of the analysed compound groups. This also should affect the main conclusion concerning the effects of fas5 genes in the discussion, a more careful wording when interpreting the results is therefore necessary.

Thank you for pointing this out. Though our focus clearly lay on the female CHC profiles as a function in sexual signaling has only been described thus far for them, we now elaborated the result and discussion for the fas5 RNAi male part (line 167-178; 258-268).

Reviewer #2 (Public Review):

Insects have long been known to use cuticular hydrocarbons for communication. While the general pathways for hydrocarbon synthesis have been worked out, their specificity and in particular the specificity of the different enzymes involved is surprisingly little understood. Here, the authors convincingly demonstrate that a single fatty acid synthase gene is responsible for a shift in the positions of methyl groups across the entire alkane spectrum of a wasp, and that the wasps males recognize females specifically based on these methyl group positions. The strength of the study is the combination of gene expression manipulations with behavioural observations evaluating the effect of the associated changes in the cuticular hydrocarbon profiles. The authors make sure that the behavioural effect is indeed due to the chemical changes by not only testing life animals, but also dead animals and corpses with manipulated cuticular hydrocarbons.

I find the evidence that the hydrocarbon changes do not affect survival and desiccation resistance less convincing (due to the limited set of conditions and relatively small sample size), but the data presented are certainly congruent with the idea that the methyl alkane changes do not have large effects on desiccation.

We appreciate the kind comments from the reviewer.

Reviewer #3 (Public Review):

In this manuscript, the authors are aiming to demonstrate that a fatty-acyl synthase gene (fas5) is involved in the composition of the blend of surface hydrocarbons of a parasitoid wasp and that it affects the sexual attractiveness of females for males. Overall, the manuscript reads very well, it is very streamlined, and the authors' claims are mostly supported by their experiments and observations.

We appreciate the kind comments from the reviewer.

However, I find that some experiments, information and/or discussion are absent to assess how the effects they observe are, at least in part, not due to other factors than fas5 and the methyl-branched (MB) alkanes. I'm also wondering if what the authors observe is only a change in the sexual attractiveness of females and not related to species recognition as well.

We appreciate the interesting point that the reviewer raises in sexual attractiveness and species recognition and now expand upon this potential aspect in the discussion (lines 327-330). However, in this manuscript, we very much focused on the effect of fas5 knockdown on the conveyance of female sexual attractiveness in a single species (Nasonia vitripennis). Therefore, we argue that species recognition constitutes a different communication modality here, and we currently cannot infer whether and how species recognition is exactly encoded in Nasonia CHC profiles despite some circumstantial evidence for species-specificity (Buellesbach et al. 2013; Mair et al. 2017). Thus, we would like to refrain from any further speculation on species recognition before this can be unambiguously demonstrated, and remain within the mechanism of sexual attractiveness within a single species which we clearly show is mediated by the female MB-alkane fraction governed by the fatty acid synthase genes. We however still consider potential alternative explanations (e.g., n-alkenes acting as a deterrent of homosexual mating attempts).

The authors explore the function of cuticular hydrocarbons (CHCs) and a fatty-acyl synthase in Nasonia vitripennis, a parasitic wasp. Using RNAi, they successfully knockdown the expression of the fas5 gene in wasps. The authors do not justify their choice of fatty-acyl synthase candidate gene. It would have been interesting to know if that is one of many genes they studied or if there was some evidence that drove them to focus their interest in fas5.

In a previous study, 5 fas candidate genes orthologous to Drosophila melanogaster fas genes were identified and mapped in the genome of Nasonia vitripennis (Buellesbach et al. 2022). We actually investigated the effects of all of these fas genes on CHC variation, but only fas5 led to such a striking, traceable pattern shift. We are currently preparing another manuscript discussing the effects of the other fas genes, but decided to focus exclusively on fas5 here, due to its significance for revealing how sexual attractiveness can be encoded and conveyed in complex chemical profiles, maintained and governed by a surprisingly simple genetic basis.

The authors observe large changes in the cuticular hydrocarbons (CHC) profile of male and females. These changes are mostly a reduction of some MB alkanes and an increase in others as well as an increase of n-alkene in fas5 knockdown females. For males fas5 knockdowns, the overall quantity of CHC is increased and consequently, multiple types of compounds are increased compared to wild-type, with only one compound appearing to decrease compared to wild-type. Insects are known to rely on ratios of compounds in blends to recognize odors. Authors address this by showing a plot of the relative ratios, but it seems to me that they do show statistical tests of those changes in the proportions of the different types of compounds. In the results section, the authors give percentages while referring to figures showing the absolute amount of CHCs. They should also test if the ratios are significantly different or not between experimental conditions. Similar data should be displayed for the males as well.

We appreciate your suggestions. We kindly refer you to our response to reviewer 1, where we addressed the statistical tests. Specifically, we generated separate subplots to display the proportions of different compound classes and performed statistical tests to compare these proportions between different treatments for both males and females. Additionally, we have revised the results section to replace relative abundances with absolute quantity, as depicted in Figure 2C-G.

Furthermore, the authors didn't use an internal standard to measure the quantity of CHCs in the extracts, which, to me, is the gold standard in the field. If I understood correctly, the authors check the abundance measured for known quantities of n-alkanes. I'm sure this method is fine, but I would have liked to be reassured that the quantities measured through this method are good by either testing some samples with an internal standard, or referring to work that demonstrates that this method is always accurate to assess the quantities of CHC in extracts of known volumes.

We actually did include 7,5 ng/μl dodecane (C12) as an “internal” standard in the hexane resuspensions of all of our processed samples (line 456, Materials and Methods). This was primarily done to allow for visually inspecting and comparing the congruence of all chromatograms in the subsequent data analysis and immediately detect any variation from sample preparation, injection process and instrument fluctuation. In our study, we have a very elaborate and standardized CHC extraction method that the volume of solvent and duration for extraction are strictly controlled to minimize the variation from sample preparation steps. Furthermore, we calibrated each individual CHC compound quantity with a dilution series of external standards (C21-C40) of known concentration. By constructing a calibration curve based on this dilution series, we achieved the most accurate compound quantification, also taking into account and counteracting the generally diminishing quantities of compounds with higher chain lengths.

The authors provide a sensible control for their RNAi experiments: targeting an unrelated gene, absent in N. vitripennis (the GFP). This allows us to see if the injection of RNAi might affect CHC profiles, which it appears to do in some cases in males, but not in females. The authors also show to the reader that their RNAi experiments do reduce the expression of the target gene. However, one of the caveats of their experiments, is that the authors don't provide evidence or information to allow the (non-expert) reader to assess whether the fas5 RNAi experiments did affect the expression of other fatty-acyl synthase genes. I'm not an expert in RNAi, so maybe this suggestion is not relevant, but it should, at least, be addressed somewhere in the manuscript that such off-target effects are very unlikely or impossible, in that case, or more generally.

We acknowledge the reviewer’s concern about potential off-target effect of the fas5 knockdown. We actually did check initially for off-target effects on the other four previously published fas genes in N. vitripennis (Lammers et al. 2019; Buellesbach et al. 2022) and did not find any effects on their respective expressions. We now include these results as supplementary data (Figure 2-figure supplement 1). However, as mentioned in the cover letter to the editor, we discovered a previously uncharacterized fas gene in the most recent N. vitripennis genome assembly (NC_045761.1), fas6, most likely constituting a tandem gene duplication of fas5. These two genes turned out to have such high sequence similarity (> 90 %, Figure 2-figure supplement 2) that both were simultaneously downregulated by our fas5 dsRNAi construct, which we confirmed with qPCR and now incorporated into our manuscript (Fig. 2H). Therefore, we now explicitly mention that the knockdown affects both genes, and either one or both could have the observed phenotypic effects. Recognizing this RNAi off-target effect, we have now also incorporated a discussion of this issue in the appropriate section of the manuscript (line 364-377), as well as the potential off-target effects of our GFP dsRNAi controls (line 262-274).

The authors observe that the modified CHCs profiles of RNAi females reduce courtship and copulation attempts, but not antennation, by males toward live and (dead) dummy females. They show that the MB alkanes of the CHC profile are sufficient to elicit sexual behaviors from males towards dummy females and that the same fraction from extracts of fas5 knockdown females does so significantly less. From the previous data, it seems that dummy females with fas5 female's MB alkanes profile elicit more antennation than CHC-cleared dummy females, but the authors do not display data for this type of target on the figure for MB alkane behavioral experiments.

Actually similar proportions of males performed antennation behavior towards female dummies with MB alkane fraction of fas5 RNAi females and CHC-cleared female dummies (55% and 50%, respectively, see Author response image 1 for the corresponding parts of the sub-figures 3 E and 4 D). We did not deem it necessary to show the same data on CHC-cleared female dummies in Figure 3 as well.

Author response image 1.

Unfortunately, the authors don't present experiments testing the effect of the non-MB alkanes fractions of the CHC extracts on male behavior toward females. As such, they are not able to (and didn't) conclude that the MB-alkane is necessary to trigger the sexual behaviors of males. I believe testing this would have significantly enhanced the significance of this work. I would also have found it interesting for the authors to comment on whether they observe aggressive behavior of males towards females (live or dead) and/or whether such behavior is expected or not in inter-individual interactions in parasitoids wasps.

In our experiment, we focus on the function of the MB-alkane fraction in female CHC profiles, and we comprehensibly demonstrate in figure 4 that the MB-alkane fraction from WT females alone is sufficient to trigger mating behavior coherent with that on alive and untreated female dummies. Therefore, we do not completely understand the reviewer’s concern about us not being ” able to (and didn't) conclude that the MB-alkane is necessary to trigger the sexual behaviors of males”. We appreciate the suggestion from the reviewer of testing the non-MB alkanes (n-alkanes and n-alkenes). However, due to the experimental procedure of separating the CHC compound class fractions through elution with molecular sieves, it was not possible for us to retrieve either the whole n-alkane or n-alkene fraction remaining bound to the sieves after separation). The role of n-alkenes in N. vitripennis is however considered in the discussion, as a deterrent for homosexual interactions between males (Wang et al. 2022a). Moreover, we did not observe aggressive behavior of males towards live or dead females.

CHCs are used by insects to signal and/or recognize various traits of targets of interest, including species or groups of origin, fertility, etc. The authors claim that their experiments show the sexual attractiveness of females can be encoded in the specific ratio of MB alkanes. While I understand how they come to this conclusion, I am somewhat concerned. The authors very quickly discuss their results in light of the literature about the role of CHCs (and notably MB alkanes) in various recognition behaviors in Hymenoptera, including conspecific recognition. Previous work (cited by the authors) has shown that males recognize males from females using an alkene (Z9C31). As such, it remains possible that the "sexual attractiveness" of N. vitripennis females for males relies on them not being males and being from the right species as well. The authors do not address the question of whether the CHCs (and the MB alkanes in particular) of females signal their sex or their species. While I acknowledge that responding to this question is beyond the scope of this work, I also strongly believe that it should be discussed in the manuscript. Otherwise, non-specialist readers would not be able to understand what I believe is one of the points that could temper the conclusions from this work.

We acknowledge the reviewer’s insight about the MB alkanes in signaling sex or species in N. vitripennis, and now include this aspect in our revised discussion (line 324-330). Moreover, we clearly demonstrate that n-alkenes have been reduced to minute trace components after our compound class separation, and the males still do not display courtship and copulation behaviors similar to WT females, thus strongly indicating that the n-alkenes do not play a role when relying solely on the changed MB-alkane patterns, further strengthening our main argument.

References

Benjamini, Y. and D. Yekutieli. 2001. The control of the false discovery rate in multiple testing under dependency. Ann. Stat. 29:1165-1188.

Buellesbach, J., J. Gadau, L. W. Beukeboom, F. Echinger, R. Raychoudhury, J. H. Werren, and T. Schmitt. 2013. Cuticular hydrocarbon divergence in the jewel wasp Nasonia: Evolutionary shifts in chemical communication channels? J. Evol. Biol. 26:2467-2478.

Buellesbach, J., C. Greim, and T. Schmitt. 2014. Asymmetric interspecific mating behavior reflects incomplete prezygotic isolation in the jewel wasp genus Nasonia. Ethology 120:834-843.

Buellesbach, J., H. Holze, L. Schrader, J. Liebig, T. Schmitt, J. Gadau, and O. Niehuis. 2022. Genetic and genomic architecture of species-specific cuticular hydrocarbon variation in parasitoid wasps. Proc. R. Soc. B 289:20220336.

Engl, T., N. Eberl, C. Gorse, T. Krüger, T. H. P. Schmidt, R. Plarre, C. Adler, and M. Kaltenpoth. 2018. Ancient symbiosis confers desiccation resistance to stored grain pest beetles. Mol. Ecol. 27:2095-2108.

Ferveur, J. F., J. Cortot, K. Rihani, M. Cobb, and C. Everaerts. 2018. Desiccation resistance: effect of cuticular hydrocarbons and water content in Drosophila melanogaster adults. Peerj 6.

Lammers, M., K. Kraaijeveld, J. Mariën, and J. Ellers. 2019. Gene expression changes associated with the evolutionary loss of a metabolic trait: lack of lipogenesis in parasitoids. BMC Genom. 20:309.

Mair, M. M., V. Kmezic, S. Huber, B. A. Pannebakker, and J. Ruther. 2017. The chemical basis of mate recognition in two parasitoid wasp species of the genus Nasonia. Entomol. Exp. Appl. 164:1-15.

Wang, Y., W. Sun, S. Fleischmann, J. G. Millar, J. Ruther, and E. C. Verhulst. 2022a. Silencing Doublesex expression triggers three-level pheromonal feminization in Nasonia vitripennis males. Proc. R. Soc. B 289:20212002.

Wang, Z., J. P. Receveur, J. Pu, H. Cong, C. Richards, M. Liang, and H. Chung. 2022b. Desiccation resistance differences in Drosophila species can be largely explained by variations in cuticular hydrocarbons. eLife 11:e80859.

Author Response

Reviewer #1 (Public Review):

In this manuscript, the authors investigate the genes involved in the retention of eggs in Aedes aegypti females. They do so by identifying two candidate genes that are differentially expressed across the different reproductive phases and also show that the transcripts of those two genes are present in ovaries and in the proteome. Overall, I think this is interesting and impressive work that characterizes the function of those two specific protein-coding genes thoroughly. I also really enjoyed the figures. Although they were a bit packed, the visuals made it easy to follow the authors' arguments. I have a few concerns and suggested changes, listed below.

1) These two genes/loci are definitely rapidly evolving. However, that does not automatically imply that positive selection has occurred in these genes. Clearly, you have demonstrated that these gene sequences might be important for fitness in Aedes aegypti. However, if these happen to be disordered proteins, then they would evolve rapidly, i.e., under fewer sequence constraints. In such a scenario, dN/dS values are likely to be high. Another possibility is that as these are expressed only in one tissue and most likely not expressed constitutively, they could be under relaxed constraints relative to all other genes in the genome. For instance, we know that average expression levels of protein-coding genes are highly correlated with their rate of molecular evolution (Drummond et al., 2005). Moreover, there have clearly been genome rearrangements and/or insertion/deletions in the studied gene sequences between closely- related species (as you have nicely shown), thus again dN/dS values will naturally be high. Thus, high values of dN/dS are neither surprising nor do they directly imply positive selection in this case. If the authors really want to investigate this further, they can use the McDonald Kreitman test (McDonald and Kreitman 1991) to ask if non- synonymous divergence is higher than expected. However, this test would require population-level data. Alternatively, the authors can simply discuss adaptation as a possibility along with the others suggested above. A discussion of alternative hypotheses is extremely important and must be clearly laid out.

We agree with the reviewer’s point that rapid evolution is not the same as positive selection. We also agree with the reviewer’s point that McDonald-Kreitman test (MK test) is more powerful than dN/dS analysis. We took advantage of a large population dataset from Rose et al. 2020. After filtering the data, we kept 454 genomes for MK tests. We found both genes are marginally significant or insignificant (tweedledee p = 0.068; tweedledum p = 0.048), despite that these are small genes and have low Pn values. This suggests that it is likely the genes evolve under positive selection.

In line with the reviewer’s suggestion, we performed another analysis using a large amount of population data. We asked if the SNP frequencies of tweedledee and tweedledum are correlated with environmental variables. We found that when compared to a distribution of 10,000 simulated genes with randomly-sampled genetic variants, both tweedledee and tweedledum showed significant correlation to multiple ecological variables reflecting climate variability, such as mean diurnal range, temperature seasonality, and precipitation seasonality (p<0.05). These results are now incorporated into the manuscript in Figure 5 and Figure 5 – Figure supplement 1.

2) The authors show that the two genes under study are important for the retention of viable eggs. However, as these genes are close to two other conserved genes (scratch and peritrophin-like gene), it is unclear to me how it is possible to rule out the contribution of the conserved genes to the same phenotype. Is it possible that the CRISPR deletion leads to the disruption of expression of one of the other important genes nearby (i.e., in a scratch or peritrophin-like gene) as the deleted region could have included a promoter region for instance, which is causing the phenotype you observe? Since all of these genes are so close to each other, it is possible that they are co-regulated and that tweedledee and tweedledum and expressed and translated along with the scratch and peritrophin-like gene. Do we know whether their expression patterns diverge and that scratch and peritrophin-like genes do not play a role in the retention of viable eggs?

This is a fair criticism; however, we think the chance that the phenotypes are caused by interrupting nearby genes is very low. First, peritrophin-like acts in the immune response, and scratch is a brain-biased transcription factor. Neither of the genes show expression in the ovary before or after blood feeding (TPM <1 or 2 are generally considered unexpressed, while scratch and peritrophin-like expression levels are overall lower than 0.1 TPM).

This suggests that peritrophin-like and scratch are not likely to function in the ovary. Thus, although we cannot completely rule out the gene knockout impacts regulation of very distant genes, it is unlikely. Since the mounting evidence we show in this manuscript that tweedledee and tweedledum are highly translated in the ovary after blooding feeding, under the principle of parsimony, we expect the phenotypes came from knocking out the highly expressed and translated genes.

Reviewer #2 (Public Review):

This manuscript is overall quite convincing, presenting a well- thought-out approach to candidate gene detection and systemic follow- ups on two genes that meet their candidate gene criteria. There are several major claims made by the authors, and some have more compelling evidence than others, but in general, the conclusions are quite sound. My main issues stem from how the strategy to identify genes playing a role in egg retention success has led to very particular genes being examined, and so I question some of the elements of the discussion focusing on the rapid evolution and taxon- uniqueness of the identified genes. In short, while I believe the authors have demonstrated that tweedledee and tweedledum play an important role in egg retention, I'm not sure whether this study should be taken as evidence that taxon-specific or rapidly evolving genes, in general, are responsible for this adaptation, or simply play an important role in it.

We have revised the paper to make it clearer that the focus is indeed on these two genes on not on the greater question of taxon-specific or rapidly-evolving genes.

First, the authors present evidence that Aedes aegypti females can retain eggs when a source of fresh water is lacking, confirming that females are not attracted to human forearms while retaining eggs and that up to 70% of the retained eggs hatch after retaining them for nearly a week. This ability is likely an important adaptation that allows Aedes aegypti to thrive in a broad range of conditions. The data here seem fairly compelling.

Based on this observation, the authors reason that genes responsible for the ability to retain eggs must: 1) be highly expressed in ovaries during retention, but not before or after. 2) be taxon-specific (as this behavior seems limited to Aedes aegypti). While this approach to enriching candidate genes has proven fruitful in this particular case, I'm not sure I agree with the authors' rationale. First, even genes at a low expression in the ovaries may be crucial to egg retention. Second, while egg-laying behavior is vastly varied in insects, I'm not sure focusing on taxon-restricted genes is necessary. It is entirely possible that many of the genes identified in Figure 2E play a crucial role in egg retention evolution. These are minor issues, but they are relevant to some later points made by the authors.

We regret framing the discovery of tweedledee and tweedledum in the original submission using this somewhat artificial set of filtering criteria. The reality is that the genes caught our attention for their novel sequence, tight genetic linkage, and interesting expression profile. That really is the focus of the paper, not these other peripheral questions that have been the focus of attention of the reviews. We really do apologize for all of the confusion about what this paper is about.

Nonetheless, the authors provide very compelling evidence that the two genes meeting their criteria - tweedledee and tweedledum, play an important role in egg retention. The genes seem to be expressed primarily in ovaries during egg retention (some observed expression in brain/testes is expected for any gene), and the proteins they code seem to be found in elevated quantities in both ovaries and hemolymph during and immediately after egg retention. RNA for the genes is detected in follicles within the ovary, and CRISPR knockouts of both the genes lead to a large decrease in egg viability post retention.

My earlier qualms about their search strategy relay into some issues with Figure 4, which describes how the two genes are 1) taxon- restricted and 2) have evolved very rapidly. Neither of the two statements is unexpected given the authors' search strategy. Of course, the genes examined precisely for their lack of homologs do not have any homologs. Similarly, by limiting themselves to genes that show a lack of homology (i.e. low sequence similarity) to other genes as well as genes with high expression levels in the ovaries, a higher rate of evolution is almost inevitable to infer (as ovary expressed genes tend to evolve more rapidly in mosquitoes). I agree with the authors that inferences of the evolutionary history of these genes are quite difficult because of their uniqueness, and I especially appreciate their attempts to identify homologs (although I really dislike the term "conceptualog").

We have removed our term “conceptualog” and replaced with the mor conventional “putative ortholog”

This leads to my main (fairly minor) issue of the paper - the discussion on the evolutionary history of these genes and its implications (sections "Taxon-restricted genes underlie tailored adaptations in a diverse world" and "Evolutionary histories and catering to different natural histories"). As noted, inferring this history is very difficult because the authors have focused on two rapidly evolving, taxon-restricted genes. The analyses they have performed here definitely demonstrate that the genes play an important role in egg retention, however, they do not show that taxon-restricted genes play a disproportionate role in egg retention evolution. Indeed, the only data relevant to this point would be the proportion of genes in Figure 2E that are taxon-restricted (3/9), but I'm not sure what the null expectation for this proportion for highly expressed ovary genes is to begin with. Furthermore, the extremely rapid evolution of this gene makes it hard to judge how truly taxon-restricted it is. My own search of tweedle homologs identified multiple as previously having been predicted to be "Knr4/Smi1-like", and while no similar genes are located in a similar location in melanogaster, there is generally little synteny conservation in Drosophila (for instance Bhutkar et al 2008), so I'm unsure what can really be said about their evolutionary origins/lack of homologs in Drosophila.

In short - the manuscript makes clear that tweedledee and tweedledum play an important role in egg retention in A. aegypti, nonetheless, it is not clear that this is a demonstration of how important taxon- restricted genes are to understanding the evolution of life-history strategies.

Again, we should have never framed the paper the way we did in the original version. We make no claims whatsoever that taxon-restricted genes in general should play a role in this biology, only that the two candidate genes under study influence egg viability after extended retention. We hope that the framing is clearer in this revision.

Author Response

Reviewer #1 (Public Review):

The authors evaluate the involvement of the hippocampus in a fast-paced time-to-contact estimation task. They find that the hippocampus is sensitive to feedback received about accuracy on each trial and has activity that tracks behavioral improvement from trial to trial. Its activity is also related to a tendency for time estimation behavior to regress to the mean. This is a novel paradigm to explore hippocampal activity and the results are thus novel and important, but the framing as well as discussion about the meaning of the findings obscures the details of the results or stretches beyond them in many places, as detailed below.

We thank the reviewer for their constructive feedback and were happy to read that s/he considered our approach and results as novel and important. The comments led us to conduct new fMRI analyses, to clarify various unclear phrasings regarding our methods, and to carefully assess our framing of the interpretation and scope of our results. Please find our responses to the individual points below.

1) Some of the results appear in the posterior hippocampus and others in the anteriorhippocampus. The authors do not motivate predictions for anterior vs. posterior hippocampus, and they do not discuss differences found between these areas in the Discussion. The hippocampus is treated as a unitary structure carrying out learning and updating in this task, but the distinct areas involved motivate a more nuanced picture that acknowledges that the same populations of cells may not be carrying out the various discussed functions.

We thank the reviewer for pointing this out. We split the hippocampus into anterior and posterior sections because prior work suggested a different whole-brain connectivity and function of the two. This was mentioned in the methods section (page 15) in the initial submission but unfortunately not in the main text. Moreover, when discussing the results, we did indeed refer mostly to the hippocampus as a unitary structure for simplicity and readability, and because statements about subcomponents are true for the whole. However, we agree with the reviewer that the differences between anterior and posterior sections are very interesting, and that describing these effects in more detail might help to guide future work more precisely.

In response to the reviewer's comment, we therefore clarified at various locations throughout the manuscript whether the respective results were observed in the posterior or anterior section of the hippocampus, and we extended our discussion to reflect the idea that different functions may be carried out by distinct populations of hippocampal cells. In addition, we also now motivate the split into the different sections better in the main text. We made the following changes.

Page 3: “Second, we demonstrate that anterior hippocampal fMRI activity and functional connectivity tracks the behavioral feedback participants received in each trial, revealing a link between hippocampal processing and timing-task performance.

Page 3: “Fourth, we show that these updating signals in the posterior hippocampus were independent of the specific interval that was tested and activity in the anterior hippocampus reflected the magnitude of the behavioral regression effect in each trial.”

Page 5: “We performed both whole-brain voxel-wise analyses as well as regions-of-interest (ROI) analysis for anterior and posterior hippocampus separately, for which prior work suggested functional differences with respect to their contributions to memory-guided behavior (Poppenk et al., 2013, Strange et al. 2014).”

Page 9: “Because anterior and posterior sections of the hippocampus differ in whole-brain connectivity as well as in their contributions to memory-guided behavior (Strange et al. 2014), we analyzed the two sections separately. “

Page 9: “We found that anterior hippocampal activity as well as functional connectivity reflected the feedback participants received during this task, and its activity followed the performance improvements in a temporal-context-dependent manner. Its activity reflected trial-wise behavioral biases towards the mean of the sampled intervals, and activity in the posterior hippocampus signaled sensorimotor updating independent of the specific intervals tested.”

Page 10: “Intriguingly, the mechanisms at play may build on similar temporal coding principles as those discussed for motor timing (Yin & Troger, 2011; Eichenbaum, 2014; Howard, 2017; Palombo & Verfaellie, 2017; Nobre & van Ede, 2018; Paton & Buonomano, 2018; Bellmund et al., 2020, 2021; Shikano et al., 2021; Shimbo et al., 2021), with differential contributions of the anterior and posterior hippocampus. Note that our observation of distinct activity modulations in the anterior and posterior hippocampus suggests that the functions and coding principles discussed here may be mediated by at least partially distinct populations of hippocampal cells.”

Page 11: Interestingly, we observed that functional connectivity of the anterior hippocampus scaled negatively (Fig. 2C) with feedback valence [...]

2) Hippocampal activity is stronger for smaller errors, which makes the interpretationmore complex than the authors acknowledge. If the hippocampus is updating sensorimotor representations, why would its activity be lower when more updating is needed?

Indeed, we found that absolute (univariate) activity of the hippocampus scaled with feedback valence, the inverse of error (Fig. 2A). We see multiple possibilities for why this might be the case, and we discussed some of them in a dedicated discussion section (“The role of feedback in timed motor actions”). For example, prior work showed that hippocampal activity reflects behavioral feedback also in other tasks, which has been linked to learning (e.g. Schönberg et al., 2007; Cohen & Ranganath, 2007; Shohamy & Wagner, 2008; Foerde & Shohamy, 2011; Wimmer et al., 2012). In our understanding, sensorimotor updating is a form of ‘learning’ in an immediate and behaviorally adaptive manner, and we therefore consider our results well consistent with this earlier work. We agree with the reviewer that in principle activity should be stronger if there was stronger sensorimotor updating, but we acknowledge that this intuition builds on an assumption about the relationship between hippocampal neural activity and the BOLD signal, which is not entirely clear. For example, prior work revealed spatially informative negative BOLD responses in the hippocampus as a function of visual stimulation (e.g. Szinte & Knapen 2020), and the effects of inhibitory activity - a leading motif in the hippocampal circuitry - on fMRI data are not fully understood. This raises the possibility that the feedback modulation we observed might also involve negative BOLD responses, which would then translate to the observed negative correlation between feedback valence and the hippocampal fMRI signal, even if the magnitude of the underlying updating mechanism was positively correlated with error. This complicates the interpretation of the direction of the effect, which is why we chose to avoid making strong conclusions about it in our manuscript. Instead, we tried discussing our results in a way that was agnostic to the direction of the feedback modulation. Importantly, hippocampal connectivity with other regions did scale positively with error (Fig. 2B), which we again discussed in the dedicated discussion section.

In response to the reviewer’s comment, we revisited this section of our manuscript and felt the latter result deserved a better discussion. We therefore took this opportunity to extend our discussion of the connectivity results (including their relationship to the univariate-activity results as well as the direction of these effects), all while still avoiding strong conclusions about directionality. Following changes were made to the manuscript.

Page 11: Interestingly, we observed that functional connectivity of the anterior hippocampus scaled negatively (Fig. 2C) with feedback valence, unlike its absolute activity, which scaled positively with feedback valence (Fig. 2A,B), suggesting that the two measures may be sensitive to related but distinct processes.

Page 11: Such network-wide receptive-field re-scaling likely builds on a re-weighting of functional connections between neurons and regions, which may explain why anterior hippocampal connectivity correlated negatively with feedback valence in our data. Larger errors may have led to stronger re-scaling, which may be grounded in a corresponding change in functional connectivity.

3) Some tests were one-tailed without justification, which reduces confidence in the robustness of the results.

We thank the reviewer for pointing us to the fact that our choice of statistical tests was not always clear in the manuscript. In the analysis the reviewer is referring to, we predicted that stronger sensorimotor updating should lead to stronger activity as well as larger behavioral improvements across the respective trials. This is because a stronger update should translate to a more accurate “internal model” of the task and therefore to a better performance. We tested this one-sided hypothesis using the appropriate test statistic (contrasting trials in which behavioral performance did improve versus trials in which it did not improve), but we did not motivate our reasoning well enough in the manuscript. The revised manuscript therefore includes the two new statements shown below to motivate our choice of test statistic more clearly.

Page 7: [...] we contrasted trials in which participants had improved versus the ones in which they had not improved or got worse (see methods for details). Because stronger sensorimotor updating should lead to larger performance improvements, we predicted to find stronger activity for improvements vs. no improvements in these tests (one-tailed hypothesis).

Page 18: These two regressors reflect the tests for target-TTC-independent and target-TTC-specific updating, respectively. Because we predicted to find stronger activity for improvements vs. no improvements in behavioral performance, we here performed one-tailed statistical tests, consistent with the direction of this hypothesis. Improvement in performance was defined as receiving feedback of higher valence than in the corresponding previous trial.

4) The introduction motivates the novelty of this study based on the idea that thehippocampus has traditionally been thought to be involved in memory at the scale of days and weeks. However, as is partially acknowledged later in the Discussion, there is an enormous literature on hippocampal involvement in memory at a much shorter timescale (on the order of seconds). The novelty of this study is not in the timescale as much as in the sensorimotor nature of the task.

We thank the reviewer for this helpful suggestion. We agree that a key part of the novelty of this study is the use of the task that is typically used to study sensorimotor integration and timing rather than hippocampal processing, along with the new insights this task enabled about the role of the hippocampus in sensorimotor updating. As mentioned in the discussion, we also agree with the reviewer that there is prior literature linking hippocampal activity to mnemonic processing on short time scales. We therefore rephrased the corresponding section in the introduction to put more weight on the sensorimotor nature of our task instead of the time scales.

Note that the new statement still includes the time scale of the effects, but that it is less at the center of the argument anymore. We chose to keep it in because we do think that the majority of studies on hippocampal-dependent memory functions focus on longer time scales than our study does, and we expect that many readers will be surprised about the immediacy of how hippocampal activity relates to ongoing behavioral performance (on ultrashort time scales).

We changed the introduction to the following.

Page 2: Here, we approach this question with a new perspective by converging two parallel lines of research centered on sensorimotor timing and hippocampal-dependent cognitive mapping. Specifically, we test how the human hippocampus, an area often implicated in episodic-memory formation (Schiller et al., 2015; Eichenbaum, 2017), may support the flexible updating of sensorimotor representations in real time and in concert with other regions. Importantly, the hippocampus is not traditionally thought to support sensorimotor functions, and its contributions to memory formation are typically discussed for longer time scales (hours, days, weeks). Here, however, we characterize in detail the relationship between hippocampal activity and real-time behavioral performance in a fast-paced timing task, which is traditionally believed to be hippocampal-independent. We propose that the capacity of the hippocampus to encode statistical regularities of our environment (Doeller et al. 2005, Shapiro et al. 2017, Behrens et al., 2018; Momennejad, 2020; Whittington et al., 2020) situates it at the core of a brain-wide network balancing specificity vs. regularization in real time as the relevant behavior is performed.

5) The authors used three different regressors for the three feedback levels, asopposed to a parametric regressor indexing the level of feedback. The predictions are parametric, so a parametric regressor would be a better match, and would allow for the use of all the medium-accuracy data.

The reviewer raises a good point that overlaps with question 3 by reviewer 2. In the current analysis, we model the three feedback levels with three independent regressors (high, medium, low accuracy). We then contrast high vs. low accuracy feedback, obtaining the results shown in Fig. 2AB. The beta estimates obtained for medium-accuracy feedback are being ignored in this contrast. Following the reviewer’s feedback, we therefore re-run the model, this time modeling all three feedback levels in one parametric regressor. All other regressors in the model stayed the same. Instead of contrasting high vs. low accuracy feedback, we then performed voxel-wise t-tests on the beta estimates obtained for the parametric feedback regressor.

The results we observed were highly consistent across the two analyses, and all conclusions presented in the initial manuscript remain unchanged. While the exact t-scores differ slightly, we replicated the effects for all clusters on the voxel-wise map (on whole-brain FWE-corrected levels) as well as for the regions-of-interest analysis for anterior and posterior hippocampus. These results are presented in a new Supplementary Figure 3C.

Note that the new Supplementary Figure 3B shows another related new analyses we conducted in response to question 4 of reviewer 2. Here, we re-ran the initial analysis with three feedback regressors, but without modeling the inter-trial interval (ITI) and the inter-session interval (ISI, i.e. the breaks participants took) to avoid model over-specification. Again, we replicated the results for all clusters and the ROI analysis, showing that the initial results we presented are robust.

The following additions were made to the manuscript.

Page 5: Note that these results were robust even when fewer nuisance regressors were included to control for model over-specification (Fig. S3B; two-tailed one-sample t tests: anterior HPC, t(33) = -3.65, p = 8.9x10-4, pfwe = 0.002, d=-0.63, CI: [-1.01, -0.26]; posterior HPC, t(33) = -1.43, p = 0.161, pfwe = 0.322, d=-0.25, CI: [-0.59, 0.10]), and when all three feedback levels were modeled with one parametric regressors (Fig. S3C; two-tailed one-sample t tests: anterior HPC, t(33) = -3.59, p = 0.002, pfwe = 0.005, d=-0.56, CI: [-0.93, -0.20]; posterior HPC, t(33) = -0.99, p = 0.329, pfwe = 0.659, d=-0.17, CI: [-0.51, 0.17]). Further, there was no systematic relationship between subsequent trials on a behavioral level [...]

Page 17: Moreover, instead of modeling the three feedback levels with three independent regressors, we repeated the analysis modeling the three feedback levels as one parametric regressor with three levels. All other regressors remained unchanged, and the model included the regressors for ITIs and ISIs. We then conducted t-tests implemented in SPM12 using the beta estimates obtained for the parametric feedback regressor (Fig. 2C). Compared to the initial analyses presented above, this has the advantage that medium-accuracy feedback trials are considered for the statistics as well.

6) The authors claim that the results support the idea that the hippocampus is findingan "optimal trade-off between specificity and regularization". This seems overly speculative given the results presented.

We understand the reviewer's skepticism about this statement and agree that the manuscript does not show that the hippocampus is finding the trade-off between specificity and regularization. However, this is also not exactly what the manuscript claims. Instead, it suggests that the hippocampus “may contribute” to solving this trade-off (page 3) as part of a “brain-wide network“ (pages 2,3,9,12). We also state that “Our [...] results suggest that this trade-off [...] is governed by many regions, updating different types of task information in parallel” (Page 11). To us, these phrasings are not equivalent, because we do not think that the role of the hippocampus in sensorimotor updating (or in any process really) can be understood independently from the rest of the brain. We do however think that our results are in line with the idea that the hippocampus contributes to solving this trade-off, and that this is exciting and surprising given the sensorimotor nature of our task, the ultrashort time scale of the underlying process, and the relationship to behavioral performance. We tried expressing that some of the points discussed remain speculation, but it seems that we were not always successful in doing so in the initial submission. We apologize for the misunderstanding, adapted corresponding statements in the manuscript, and we express even more carefully that these ideas are speculation.

Following changes were made to the introduction and discussion.

Page 2: Here, we approach this question with a new perspective by converging two parallel lines of research centered on sensorimotor timing and hippocampal-dependent cognitive mapping. Specifically, we test how the human hippocampus, an area often implicated in episodic-memory formation (Schiller et al., 2015; Eichenbaum, 2017), may support the flexible updating of sensorimotor representations in real time and in concert with other regions.

Page 12: Because hippocampal activity (Julian & Doeller, 2020) and the regression effect (Jazayeri & Shadlen, 2010) were previously linked to the encoding of (temporal) context, we reasoned that hippocampal activity should also be related to the regression effect directly. This may explain why hippocampal activity reflected the magnitude of the regression effect as well as behavioral improvements independently from TTC, and why it reflected feedback, which informed the updating of the internal prior.

Page 12: This is in line with our behavioral results, showing that TTC-task performance became more optimal in the face of both of these two objectives. Over time, behavioral responses clustered more closely between the diagonal and the average line in the behavioral response profile (Fig. 1B, S1G), and the TTC error decreased over time. While different participants approached these optimal performance levels from different directions, either starting with good performance or strong regularization, the group approached overall optimal performance levels over the course of the experiment.

Page 13: This is in line with the notion that the hippocampus [...] supports finding an optimal trade off between specificity and regularization along with other regions. [...] Our results show that the hippocampus supports rapid and feedback-dependent updating of sensorimotor representations, suggesting that it is a central component of a brain-wide network balancing task specificity vs. regularization for flexible behavior in humans.

Note that in response to comment 1 by reviewer 2, the revised manuscript now reports the results of additional behavioral analyses that support the notion that participants find an optimal trade-off between specificity and regularization over time (independent of whether the hippocampus was involved or not).

7) The authors find that hippocampal activity is related to behavioral improvement fromthe prior trial. This seems to be a simple learning effect (participants can learn plenty about this task from a prior trial that does not have the exact same timing as the current trial) but is interpreted as sensitivity to temporal context. The temporal context framing seems too far removed from the analyses performed.

We agree with the reviewer that our observation that hippocampal activity reflects TTC-independent behavioral improvements across trials could have multiple explanations. Critically, i) one of them is that the hippocampus encodes temporal context, ii) it is only one of multiple observations that we build our interpretation on, and iii) our interpretation builds on multiple earlier reports

Interval estimates regress toward the mean of the sampled intervals, an effect that is often referred to as the “regression effect”. This effect, which we observed in our data too (Fig. 1B), has been proposed to reflect the encoding of temporal context (e.g. Jazayeri & Shadlen 2010). Moreover, there is a large body of literature on how the hippocampus may support the encoding of spatial and temporal context (e.g. see Bellmund, Polti & Doeller 2020 for review).

Because both hippocampal activity and the regression effect were linked to the encoding of (temporal) context, we reasoned that hippocampal activity should also be related to the regression effect directly. If so, one would expect that hippocampal activity should reflect behavioral improvements independently from TTC, it should reflect the magnitude of the regression effect, and it should generally reflect feedback, because it is the feedback that informs the updating of the internal prior.

All three observations may have independent explanations indeed, but they are all also in line with the idea that the hippocampus does encode temporal context and that this explains the relationship between hippocampal activity and the regression effect. It therefore reflects a sparse and reasonable explanation in our opinion, even though it necessarily remains an interpretation. Of course, we want to be clear on what our results are and what our interpretations are.

In response to the reviewer’s comment, we therefore toned down two of the statements that mention temporal context in the manuscript, and we removed an overly speculative statement from the result section. In addition, the discussion now describes more clearly how our results are in line with this interpretation.

Abstract: This is in line with the idea that the hippocampus supports the rapid encoding of temporal context even on short time scales in a behavior-dependent manner.

Page 13: This is in line with the notion that the hippocampus encodes temporal context in a behavior-dependent manner, and that it supports finding an optimal trade off between specificity and regularization along with other regions.

Page 12: Because hippocampal activity (Julian & Doeller, 2020) and the regression effect (Jazayeri & Shadlen, 2010) were previously linked to the encoding of (temporal) context, we reasoned that hippocampal activity should also be related to the regression effect directly. This may explain why hippocampal activity reflected the magnitude of the regression effect as well as behavioral improvements independently from TTC, and why it reflected feedback, which informed the updating of the internal prior.

The following statement was removed, overlapping with comment 2 by Reviewer 3:

Instead, these results are consistent with the notion that hippocampal activity signals the updating of task-relevant sensorimotor representations in real-time.

8) I am not sure the term "extraction of statistical regularities" is appropriate. The termis typically used for more complex forms of statistical relationships.

We agree with the reviewer that this expression may be interpreted differently by different readers and are grateful to be pointed to this fact. We therefore removed it and instead added the following (hopefully less ambiguous) statement to the manuscript.

Page 9: This study investigated how the human brain flexibly updates sensorimotor representations in a feedback-dependent manner in the service of timing behavior.

Reviewer #2 (Public Review):

The authors conducted a study involving functional magnetic resonance imaging and a time-to-contact estimation paradigm to investigate the contribution of the human hippocampus (HPC) to sensorimotor timing, with a particular focus on the involvement of this structure in specific vs. generalized learning. Suggestive of the former, it was found that HPC activity reflected time interval-specific improvements in performance while in support of the latter, HPC activity was also found to signal improvements in performance, which were not specific to the individual time intervals tested. Based on these findings, the authors suggest that the human HPC plays a key role in the statistical learning of temporal information as required in sensorimotor behaviour.

By considering two established functions of the HPC (i.e., temporal memory and generalization) in the context of a domain that is not typically associated with this structure (i.e., sensorimotor timing), this study is potentially important, offering novel insight into the involvement of the HPC in everyday behaviour. There is much to like about this submission: the manuscript is clearly written and well-crafted, the paradigm and analyses are well thought out and creative, the methodology is generally sound, and the reported findings push us to consider HPC function from a fresh perspective. A relative weakness of the paper is that it is not entirely clear to what extent the data, at least as currently reported, reflects the involvement of the HPC in specific and generalized learning. Since the authors' conclusions centre around this observation, clarifying this issue is, in my opinion, of primary importance.

We thank the reviewer for these positive and extremely helpful comments, which we will address in detail below. In response to these comments, the revised manuscript clarifies why the observed performance improvements are not at odds with the idea that an optimal trade-off between specificity and regularization is found, and how the time course of learning relates to those reported in previous literature. In addition, we conducted two new fMRI analyses, ensuring that our conclusions remain unchanged even if feedback is modeled with one parametric regressor, and if the number or nuisance regressors is reduced to control for overparameterization of the model. Please find our responses underneath each individual point below.

1) Throughout the manuscript, the authors discuss the trade-off between specific and generalized learning, and point towards Figure S1D as evidence for this (i.e., participants with higher TTC accuracy exhibited a weaker regression effect). What appears to be slightly at odds with this, however, is the observation that the deviation from true TTC decreased with time (Fig S1F) as the regression line slope approached 0.5 (Fig S1E) - one would have perhaps expected the opposite i.e., for deviation from true TTC to increase as generalization increases. To gain further insight into this, it would be helpful to see the deviation from true TTC plotted for each of the four TTC intervals separately and as a signed percentage of the target TTC interval (i.e., (+) or (-) deviation) rather than the absolute value.

We thank the reviewer for raising this important question and for the opportunity to elaborate on the relationship between the TTC error and the magnitude of the regression effect in behavior. Indeed, we see that the regression slopes approach 0.5 and that the TTC error decreases over the course of the experiment. We do not think that these two observations are at odds with each other for the following reasons:

First, while the reviewer is correct in pointing out that the deviation from the TTC should increase as “generalization increases”, that is not what we found. It was not the magnitude of the regularization per se that increased over time, but the overall task performance became more optimal in the face of both objectives: specificity and generalization. This optimum is at a regression-line slope of 0.5. Generalization (or regularization how we refer to it in the present manuscript), therefore did not increase per se on group level.

Second, the regression slopes approached 0.5 on the group-level, but the individual participants approached this level from different directions: Some of them started with a slope value close to 1 (high accuracy), whereas others started with a slope value close to 0 (near full regression to the mean). Irrespective of which slope value they started with, over time, they got closer to 0.5 (Rebuttal Figure 1A). This can also be seen in the fact that the group-level standard deviation in regression slopes becomes smaller over the course of the experiment (Rebuttal Figure 1B, SFig 1G). It is therefore not generally the case that the regression effect becomes stronger over time, but that it becomes more optimal for longer-term behavioral performance, which is then also reflected in an overall decrease in TTC error. Please see our response to the reviewer’s second comment for more discussion on this.

Third, the development of task performance is a function of two behavioral factors: a) the accuracy and b) the precision in TTC estimation. Accuracy describes how similar the participant’s TTC estimates were to the true TTC, whereas precision describes how similar the participant’s TTC estimates were relative to each other (across trials). Our results are a reflection of the fact that participants became both more accurate over time on average, but also more precise. To demonstrate this point visually, we now plotted the Precision and the Accuracy for the 8 task segments below (Rebuttal Figure 1C, SFig 1H), showing that both measures increased as the time progressed and more trials were performed. This was the case for all target durations.

In response to the reviewer’s comment, we clarified in the main text that these findings are not at odds with each other. Furthermore, we made clear that regularization per se did not increase over time on group level. We added additional supporting figures to the supplementary material to make this point. Note that in our view, these new analyses and changes more directly address the overall question the reviewer raised than the figure that was suggested, which is why we prioritized those in the manuscript.

However, we appreciated the suggestion a lot and added the corresponding figure for the sake of completeness.

Following additions were made.

Page 5: In support of this, participants' regression slopes converged over time towards the optimal value of 0.5, i.e. the slope value between veridical performance and the grand mean (Fig. S1F; linear mixed-effects model with task segment as a predictor and participants as the error term, F(1) = 8.172, p = 0.005, ε2=0.08, CI: [0.01, 0.18]), and participants' slope values became more similar (Fig. S1G; linear regression with task segment as predictor, F(1) = 6.283, p = 0.046, ε2 = 0.43, CI: [0, 1]). Consequently, this also led to an improvement in task performance over time on group level (i.e. task accuracy and precision increased (Fig. S1I), and the relationship between accuracy and precision became stronger (Fig. S1H), linear mixed-effect model results for accuracy: F(1) = 15.127, p = 1.3x10-4, ε2=0.06, CI: [0.02, 0.11], precision: F(1) = 20.189, p = 6.1x10-5, ε2 = 0.32, CI: [0.13, 1]), accuracy-precision relationship: F(1) = 8.288, p =0.036, ε2 = 0.56, CI: [0, 1], see methods for model details).

Page 12: This suggests that different regions encode distinct task regularities in parallel to form optimal sensorimotor representations to balance specificity and regularization. This is in line with our behavioral results, showing that TTC-task performance became more optimal in the face of both of these two objectives. Over time, behavioral responses clustered more closely between the diagonal and the average line in the behavioral response profile (Fig. 1B, S1G), and the TTC error decreased over time. While different participants approached these optimal performance levels from different directions, either starting with good performance or strong regularization, the group approached overall optimal performance levels over the course of the experiment.

Page 15: We also corroborated this effect by measuring the dispersion of slope values between participants across task segments using a linear regression model with task segment as a predictor and the standard deviation of slope values across participants as the dependent variable (Fig. S1G). As a measure of behavioral performance, we computed two variables for each target-TTC level: sensorimotor timing accuracy, defined as the absolute difference in estimated and true TTC, and sensorimotor timing precision, defined as coefficient of variation (standard deviation of estimated TTCs divided by the average estimated TTC). To study the interaction between these two variables for each target TTC over time, we first normalized accuracy by the average estimated TTC in order to make both variables comparable. We then used a linear mixed-effects model with precision as the dependent variable, task segment and normalized accuracy as predictors and target TTC as the error term. In addition, we tested whether accuracy and precision increased over the course of the experiment using separate linear mixed-effects models with task segment as predictor and participants as the error term.

2) Generalization relies on prior experience and can be relatively slow to develop as is the case with statistical learning. In Jazayeri and Shadlen (2010), for instance, learning a prior distribution of 11-time intervals demarcated by two briefly flashed cues (compared to 4 intervals associated with 24 possible movement trajectories in the current study) required ~500 trials. I find it somewhat surprising, therefore, that the regression line slope was already relatively close to 0.5 in the very first segment of the task. To what extent did the participants have exposure to the task and the target intervals prior to entering the scanner?

We thank the reviewer for raising the important question about the time course of learning in our task and how our results relate to prior work on this issue. Addressing the specific reviewer question first, participants practiced the task for 2-3 minutes prior to scanning. During the practice, they were not specifically instructed to perform the task as well as they could nor to encode the intervals, but rather to familiarize themselves with the general experimental setup and to ask potential questions outside the MRI machine. While they might have indeed started encoding the prior distribution of intervals during the practice already, we have no way of knowing, and we expect the contribution of this practice on the time course of learning during scanning to be negligible (for the reasons outlined above).

However, in addition to the specific question the reviewer asked, we feel that the comment raises two more general points: 1) How long does it take to learn the prior distribution of a set of intervals as a function of the number of intervals tested, and 2) Why are the learning slopes we report quite shallow already in the beginning of the scan?

Regarding (1), we are not aware of published reports that answer this question directly, and we expect that this will depend on the task that is used. Regarding the comparison to Jazayeri & Shadlen (2010), we believe the learning time course is difficult to compare between our study and theirs. As the reviewer mentioned, our study featured only 4 intervals compared to 11 in their work, based on which we would expect much faster learning in our task than in theirs. We did indeed sample 24 movement directions, but these were irrelevant in terms of learning the interval distribution. Moreover, unlike Jazayeri & Shadlen (2010), our task featured moving stimuli, which may have added additional sensory, motor and proprioceptive information in our study which the participants of the prior study could not rely on.

Regarding (2), and overlapping with the reviewer’s previous comment, the average learning slope in our study is indeed close to 0.5 already in the first task segment, but we would like to highlight that this is a group-level measure. The learning slopes of some subjects were closer to 1 (i.e. the diagonal in Fig 1B), and the one of others was closer to 0 (i.e. the mean) in the beginning of the experiment. The median slope was close to 0.65. Importantly, the slopes of most participants still approached 0.5 in the course of the experiment, and so did even the group-level slope the reviewer is referring to. This also means that participants’ slopes became more similar in the course of the experiment, and they approached 0.5, which we think reflects the optimal trade-off between regressing towards the mean and regressing towards the diagonal (in the data shown in Fig. 1B). This convergence onto the optimal trade-off value can be seen in many measures, including the mean slope (Rebuttal Figure 1A, SFig 1F), the standard deviation in slopes (Rebuttal Figure 1B, SFig 1G) as well as the Precision vs. Accuracy tradeoff (Rebuttal Figure 1C, SFig 1H). We therefore think that our results are well in line with prior literature, even though a direct comparison remains difficult due to differences in the task.

In response to the reviewer’s comment, and related to their first comment, we made the following addition to the discussion section.

Page 12: This suggests that different regions encode distinct task regularities in parallel to form optimal sensorimotor representations to balance specificity and regularization. This is well in line with our behavioral results, showing that TTC-task performance became more optimal in the face of both of these two objectives. Over time, behavioral responses clustered more closely between the diagonal and the average line in the behavioral response profile (Fig. 1B, S1G), and the TTC error decreased over time. While different participants approached these optimal performance levels from different directions, either starting with good performance or strong regularization, the group approached overall optimal performance levels over the course of the experiment.

3) I am curious to know whether differences between high-accuracy andmedium-accuracy feedback as well as between medium-accuracy and low-accuracy feedback predicted hippocampal activity in the first GLM analysis (middle page 5). Currently, the authors only present the findings for the contrast between high-accuracy and low-accuracy feedback. Examining all feedback levels may provide additional insight into the nature of hippocampal involvement and is perhaps more consistent with the subsequent GLM analysis (bottom page 6) in which, according to my understanding, all improvements across subsequent trials were considered (i.e., from low-accuracy to medium-accuracy; medium-accuracy to high-accuracy; as well as low-accuracy to high-accuracy).

We thank the reviewer for this thoughtful question, which relates to questions 5 by reviewer 1. The reviewer is correct that the contrast shown in Fig 2 does not consider the medium-accuracy feedback levels, and that the model in itself is slightly different from the one used in the subsequent analysis presented in Fig. 3. To reply to this comment as well as to a related one by reviewer 1 together, we therefore repeated the full analysis while modeling the three feedback levels in one parametric regressor, which includes the medium-accuracy feedback trials, and is consistent with the analysis shown in Fig. 3. The results of this new analysis are presented in the new Supplementary Fig. 3B.

In short, the model included one parametric regressor with three levels reflecting the three types of feedback, and all nuisance regressors remained unchanged. Instead of contrasting high vs. low accuracy feedback, we then performed voxel-wise t-tests on the beta estimates obtained for the parametric feedback regressor. We found that our results presented initially were very robust: Both the observed clusters in the voxel-wise analysis (on whole-brain FWE-corrected levels) as well as the ROI results replicated across the two analyses, and our conclusions therefore remain unchanged.

We made multiple textual additions to the manuscript to include this new analysis, and we present the results of the analysis including a direct comparison to our initial results in the new Supplementary Fig. 3. Following textual additions were.

Page 5: Note that these results were robust even when fewer nuisance regressors were included to control for model over-specification (Fig. S3B; two-tailed one-sample t tests: anterior HPC, t(33) = -3.65, p = 8.9x10-4, pfwe = 0.002, d=-0.63, CI: [-1.01, -0.26]; posterior HPC, t(33) = -1.43, p = 0.161, pfwe = 0.322, d=-0.25, CI: [-0.59, 0.10]), and when all three feedback levels were modeled with one parametric regressors (Fig. S3C; two-tailed one-sample t tests: anterior HPC, t(33) = -3.59, p = 0.002, pfwe = 0.005, d=-0.56, CI: [-0.93, -0.20]; posterior HPC, t(33) = -0.99, p = 0.329, pfwe = 0.659, d=-0.17, CI: [-0.51, 0.17]). Further, there was no systematic relationship between subsequent trials on a behavioral level [...]

Page 17: Moreover, instead of modeling the three feedback levels with three independent regressors, we repeated the analysis modeling the three feedback levels as one parametric regressor with three levels. All other regressors remained unchanged, and the model included the regressors for ITIs and ISIs. We then conducted t-tests implemented in SPM12 using thebeta estimates obtained for the parametric feedback regressor (Fig. S2C). Compared to the initial analyses presented above, this has the advantage that medium-accuracy feedback trials are considered for the statistics as well.

4) The authors modeled the inter-trial intervals and periods of rest in their univariateGLMs. This approach of modelling all 'down time' can lead to model over-specification and inaccurate parameter estimation (e.g. Pernet, 2014). A comment on this approach as well as consideration of not modelling the inter-trial intervals would be useful.

This is an important issue that we did not address in our initial manuscript. We are aware and agree with the reviewer’s general concern about model over-specification, which can be a big problem in regression as it leads to biased estimates. We did examine whether our model was overspecified before running it, but we did not report a formal test of it in the manuscript. We are grateful to be given the opportunity to do so now.

In response to the reviewer’s comment, we repeated the full analysis shown in Fig. 2 while excluding the nuisance regressors for inter-trial intervals (ISI) and breaks (or inter-session intervals, ISI). All other regressors and analysis steps stayed unchanged relative to the one reported in Fig. 2. The new results are presented in a new Supplementary Figure 3B.

Like for our previous analysis, we again see that the results we initially presented were extremely robust even on whole-brain FWE corrected levels, as well as on ROI level. Our conclusions therefore remain unchanged, and the results we presented initially are not affected by potential model overspecification. In addition to the new Supplementary Figure 3B, we made multiple textual changes to the manuscript to describe this new analysis and its implications. Note that we used the same nuisance regressors in all other GLM analyses too, meaning that it is also very unlikely that model overspecification affects any of the other results presented. We thank the reviewer for suggesting this analysis, and we feel including it in the manuscript has further strengthened the points we initially made.

Following additions were made to the manuscript.

Page 16: The GLM included three boxcar regressors modeling the feedback levels, one for ITIs, one for button presses and one for periods of rest (inter-session interval, ISI) [...]

Page 16: ITIs and ISIs were modeled to reduce task-unrelated noise, but to ensure that this did not lead to over-specification of the above-described GLM, we repeated the full analysis without modeling the two. All other regressors including the main feedback regressors of interest remained unchanged, and we repeated both the voxel-wise and ROI-wise statistical tests as described above (Fig. S2B).

Page 17: Note that these results were robust even when fewer nuisance regressors were included to control for model over-specification (Fig. S3B; two-tailed one-sample t tests: anterior HPC, t(33) = -3.65, p = 8.9x10-4, pfwe = 0.002, d=-0.63, CI: [-1.01, -0.26]; posterior HPC, t(33) = -1.43, p = 0.161, pfwe = 0.322, d=-0.25, CI: [-0.59, 0.10]), and when all three feedback levels were modeled with one parametric regressors (Fig. S3C; two-tailed one-sample t tests: anterior HPC, t(33) = -3.59, p = 0.002, pfwe = 0.005, d=-0.56, CI: [-0.93, -0.20]; posterior HPC, t(33) = -0.99, p = 0.329, pfwe = 0.659, d=-0.17, CI: [-0.51, 0.17]). Further, there was no systematic relationship between subsequent trials on a behavioral level [...]

Reviewer #3 (Public Review):

This paper reports the results of an interesting fMRI study examining the neural correlates of time estimation with an elegant design and a sensorimotor timing task. Results show that hippocampal activity and connectivity are modulated by performance on the task as well as the valence of the feedback provided. This study addresses a very important question in the field which relates to the function of the hippocampus in sensorimotor timing. However, a lack of clarity in the description of the MRI results (and associated methods) currently prevents the evaluation of the results and the interpretations made by the authors. Specifically, the model testing for timing-specific/timing-independent effects is questionable and needs to be clarified. In the current form, several conclusions appear to not be fully supported by the data.

We thank the reviewer for pointing us to many methodological points that needed clarification. We apologize for the confusion about our methods, which we clarify in the revised manuscript. Please find our responses to the individual points below.

Major points

Some methodological points lack clarity which makes it difficult to evaluate the results and the interpretation of the data.

We really appreciate the many constructive comments below. We feel that clarifying these points improved our manuscript immensely.

1) It is unclear how the 3 levels of accuracy and feedback (high, medium, and lowperformance) were computed. Please provide the performance range used for this classification. Was this adjusted to the participants' performance?

The formula that describes how the response window was computed for the different speed levels was reported in the methods section of the original manuscript on page 13. It reads as follows:

“The following formula was used to scale the response window width: d ± ((k ∗ d)/2) where d is the target TTC and k is a constant proportional to 0.3 and 0.6 for high and medium accuracy, respectively.“

In response to the reviewer’s comment, we now additionally report the exact ranges of the different response windows in a new Supplementary Table 1 and refer to it in the Methods section as follows.

Page 10: To calibrate performance feedback across different TTC durations, the precise response window widths of each feedback level scaled with the speed of the fixation target (Table S1).

2) The description of the MRI results lacks details. It is not always clear in the resultssection which models were used and whether parametric modulators were included or not in the model. This makes the results section difficult to follow. For example,

a) Figure 2: According to the description in the text, it appears that panels A and B report the results of a model with 3 regressors, ie one for each accuracy/feedback level (high, medium, low) without parametric modulators included. However, the figure legend for panel B mentions a parametric modulator suggesting that feedback was modelled for each trial as a parametric modulator. The distinction between these 2 models must be clarified in the result section.

We thank the reviewer very much for spotting this discrepancy. Indeed, Figure 2 shows the results obtained for a GLM in which we modeled the three feedback levels with separate regressors, not with one parametric regressor. Instead, the latter was the case for Figure 3. We apologize for the confusion and corrected the description in the figure caption, which now reads as follows. The description in the main text and the methods remain unchanged.

Caption Fig. 2: We plot the beta estimates obtained for the contrast between high vs. low feedback.

Moreover, note that in response to comment 5 by reviewer 1 and comment 3 by reviewer 2, the revised manuscript now additionally reports the results obtained for the parametric regressor in the new Supplementary Figure 3C. All conclusions remain unchanged.

Additionally, it is unclear how Figure 2A supports the following statement: "Moreover, the voxel-wise analysis revealed similar feedback-related activity in the thalamus and the striatum (Fig. 2A), and in the hippocampus when the feedback of the current trial was modeled (Fig. S3)." This is confusing as Figure 2A reports an opposite pattern of results between the striatum/thalamus and the hippocampus. It appears that the statement highlighted above is supported by results from a model including current trial feedback as a parametric modulator (reported in Figure S3).

We agree with the reviewer that our result description was confusing and changed it. It now reads as follows.

Page 5: Moreover, the voxel-wise analysis revealed feedback-related activity also in the thalamus and the striatum (Fig. 2A) [...]

Also, note that it is unclear from Figure 2A what is the direction of the contrast highlighting the hippocampal cluster (high vs. low according to the text but the figure shows negative values in the hippocampus and positive values in the thalamus). These discrepancies need to be addressed and the models used to support the statements made in the results sections need to be explicitly described.

The description of the contrast is correct. Negative values indicate smaller errors and therefore better feedback, which is mentioned in the caption of Fig. 2 as follows:

“Negative values indicate that smaller errors, and higher-accuracy feedback, led to stronger activity.”

Note that the timing error determined the feedback, and that we predicted stronger updating and therefore stronger activity for larger errors (similar to a prediction error). We found the opposite. We mention the reasoning behind this analysis at various locations in the manuscript e.g. when talking about the connectivity analysis:

“We reasoned that larger timing errors and therefore low-accuracy feedback would result in stronger updating compared to smaller timing errors and high-accuracy feedback”

In response to the reviewer’s remark, we clarified this further by adding the following statement to the result section.

Page 5: “Using a mass-univariate general linear model (GLM), we modeled the three feedback levels with one regressor each plus additional nuisance regressors (see methods for details). The three feedback levels (high, medium and low accuracy) corresponded to small, medium and large timing errors, respectively. We then contrasted the beta weights estimated for high-accuracy vs. low-accuracy feedback and examined the effects on group-level averaged across runs.”

b) Connectivity analyses: It is also unclear here which model was used in the PPIanalyses presented in Figure 2. As it appears that the seed region was extracted from a high vs. low contrast (without modulators), the PPI should be built using the same model. I assume this was the case as the authors mentioned "These co-fluctuations were stronger when participants performed poorly in the previous trial and therefore when they received low-accuracy feedback." if this refers to low vs. high contrast. Please clarify.

Yes, the PPI model was built using the same model. We clarified this in the methods section by adding the following statement to the PPI description.

Page 17: “The PPI model was built using the same model that revealed the main effects used to define the HPC sphere “

Yes, the reviewer is correct in thinking that the contrast shows the difference between low vs. high-accuracy feedback. We clarified this in the main text as well as in the caption of Fig. 2.

Caption Fig 2: [...] We plot results of a psychophysiological interactions (PPI) analysis conducted using the hippocampal peak effects in (A) as a seed for low vs. high-accuracy feedback. [...]

Page 17: The estimated beta weight corresponding to the interaction term was then tested against zero on the group-level using a t-test implemented in SPM12 (Fig. 2C). The contrast reflects the difference between low vs. high-accuracy feedback. This revealed brain areas whose activity was co-varying with the hippocampus seed ROI as a function of past-trial performance (n-1).

c) It is unclear why the model testing TTC-specific / TTC-independent effects (resultspresented in Figure 3) used 2 parametric modulators (as opposed to building two separate models with a different modulator each). I wonder how the authors dealt with the orthogonalization between parametric modulators with such a model. In SPM, the orthogonalization of parametric modulators is based on the order of the modulators in the design matrix. In this case, parametric modulator #2 would be orthogonalized to the preceding modulator so that a contrast focusing on the parametric modulator #2 would highlight any modulation that is above and beyond that explained by modulator #1. In this case, modulation of brain activity that is TTC-specific would have to be above and beyond a modulation that is TTC-independent to be highlighted. I am unsure that this is what the authors wanted to test here (or whether this is how the MRI design was built). Importantly, this might bias the interpretation of their results as - by design - it is less likely to observe TTC-specific modulations in the hippocampus as there is significant TTC-independent modulation. In other words, switching the order of the modulators in the model (or building two separate models) might yield different results. This is an important point to address as this might challenge the TTC-specific/TTC-independent results described in the manuscript.

We thank the reviewer for raising this important issue. When running the respective analysis, we made sure that the regressors were not collinear and we therefore did not expect substantial overlap in shared variance between them. However, we agree with the reviewer that orthogonalizing one regressor with respect to the other could still affect the results. To make sure that our expectations were indeed met, we therefore repeated the main analysis twice: 1) switching the order of the modulators and 2) turning orthogonalization off (which is possible in SPM12 unlike in previous versions). In all cases, our key results and conclusions remained unchanged, including the central results of the hippocampus analyses.

Anterior (ant.) / Posterior (post.) Hippocampus ROI analysis with A) original order of modulators, B) switching the order of the modulators and C) turning orthogonalization of modulators off. ABC) Orange color corresponds to the TTC-independent condition whereas light-blue color corresponds to the TTC-specific condition. Statistics reflect p<0.05 at Bonferroni corrected levels () obtained using a group-level one-tailed one-sample t-test against zero; A) pfwe = 0.017, B) pfwe = 0.039, C) pfwe = 0.039.*

Because orthogonalization did not affect the conclusions, the new manuscript simply reports the analysis for which it was turned off. Note that these new figures are extremely similar to the original figures we presented, which can be seen in the exemplary figure below showing our key results at a liberal threshold for transparency. In addition, we clarified that orthogonalization was turned off in the methods section as follows.

Page 18: These two regressors reflect the tests for target-TTC-independent and target-TTC-specific updating, respectively, and they were not orthogonalized to each other.

Comparison of old & new results: also see Fig. 3 and Fig. S5 in manuscript

d) It is also unclear how the behavioral improvement was coded/classified "wecontrasted trials in which participants had improved versus the ones in which they had not improved or got worse"- It appears that improvement computation was based on the change of feedback valence (between high, medium and low). It is unclear why performance wasn't used instead? This would provide a finer-grained modulation?

We thank the reviewer for the opportunity to clarify this important point. First, we chose to model feedback because it is the feedback that determines whether participants update their “internal model” or not. Without feedback, they would not know how well they performed, and we would not expect to find activity related to sensorimotor updating. Second, behavioral performance and received feedback are tightly correlated, because the former determines the latter. We therefore do not expect to see major differences in results obtained between the two. Third, we did in fact model both feedback and performance in two independent GLMs, even though the way the results were reported in the initial submission made it difficult to compare the two.

Figure 4 shows the results obtained when modeling behavioral performance in the current trial as an F-contrast, and Supplementary Fig 4 shows the results when modeling the feedback received in the current trial as a t-contrast. While the voxel-wise t-maps/F-maps are also quite similar, we now additionally report the t-contrast for the behavioral-performance GLM in a new Supplementary Figure 4C. The t-maps obtained for these two different analyses are extremely similar, confirming that the direction of the effects as well as their interpretation remain independent of whether feedback or performance is modeled.

The revised manuscript refers to the new Supplementary Figure 4C as follows.

Page 17: In two independent GLMs, we analyzed the time courses of all voxels in the brain as a function of behavioral performance (i.e. TTC error) in each trial, and as a function of feedback received at the end of each trial. The models included one mean-centered parametric regressor per run, modeling either the TTC error or the three feedback levels in each trial, respectively. Note that the feedback itself was a function of TTC error in each trial [...] We estimated weights for all regressors and conducted a t-test against zero using SPM12 for our feedback and performance regressors of interest on the group level (Fig. S4A). [...]

Page 17: In addition to the voxel-wise whole-brain analyses described above, we conducted independent ROI analyses for the anterior and posterior sections of the hippocampus (Fig. S2A). Here, we tested the beta estimates obtained in our first-level analysis for the feedback and performance regressors of interest (Fig. S4B; two-tailed one-sample t tests: anterior HPC, t(33) = -5.92, p = 1.2x10-6, pfwe = 2.4x10-6, d=-1.02, CI: [-1.45, -0.6]; posterior HPC, t(33) = -4.07, p = 2.7x10-4, pfwe = 5.4x10-4, d=-0.7, CI: [-1.09, -0.32]). See section "Regions of interest definition and analysis" for more details.

If the feedback valence was used to classify trials as improved or not, how was this modelled (one regressor for improved, one for no improvement? As opposed to a parametric modulator with performance improvement?).

We apologize for the lack of clarity regarding our regressor design. In response to this comment, we adapted the corresponding paragraph in the methods to express more clearly that improvement trials and no-improvement trials were modeled with two separate parametric regressors - in line with the reviewer’s understanding. The new paragraph reads as follows.

Page 18: One regressor modeled the main effect of the trial and two parametric regressors modeled the following contrasts: Parametric regressor 1: trials in which behavioral performance improved \textit{vs}. parametric regressor 2: trials in which behavioral performance did not improve or got worse relative to the previous trial.

Last, it is also unclear how ITI was modelled as a regressor. Did the authors mean a parametric modulator here? Some clarification on the events modelled would also be helpful. What was the onset of a trial in the MRI design? The start of the trial? Then end? The onset of the prediction time?

The Inter-trial intervals (ITIs) were modeled as a boxcar regressor convolved with the hemodynamic response function. They describe the time after the feedback-phase offset and the subsequent trial onset. Moreover, the start of the trial was the moment when the visual-tracking target started moving after the ITI, whereas the trial end was the offset of the feedback phase (i.e. the moment in which the feedback disappeared from the screen). The onset of the “prediction time” was the moment in which the visual-tracking target stopped moving, prompting participants to estimate the time-to-contact. We now explain this more clearly in the methods as shown below.

Page 16: The GLM included three boxcar regressors modeling the feedback levels, one for ITIs, one for button presses and one for periods of rest (inter-session interval, ISI), which were all convolved with the canonical hemodynamic response function of SPM12. The start of the trial was considered as the trial onsets for modeling (i.e. the time when the visual-tracking target started moving). The trial end was the offset of the feedback phase (i.e. the moment in which the feedback disappeared from the screen). The ITI was the time between the offset of the feedback-phase and the subsequent trial onset.

On a related note, in response to question 4 by reviewer 2, we now repeated one of the main analyses (Fig. 2) without modeling the ITI (as well as the Inter-session interval, ISI). We found that our key results and conclusions are independent of whether or not these time points were modeled. These new results are presented in the new Supplementary Figure 3B.

Page 16: ITIs and ISIs were modeled to reduce task-unrelated noise, but to ensure that this did not lead to over-specification of the above-described GLM, we repeated the full analysis without modeling the two. [...]

1. Perhaps as a result of a lack of clarity in the result section and the MRI methods, it appears that some conclusions presented in the result section are not supported by the data. E.g. "Instead, these results are consistent with the notion that hippocampal activity signals the updating of task-relevant sensorimotor representations in real-time." The data show that hippocampal activity is higher during and after an accurate trial. This pattern of results could be attributed to various processes such as e.g. reward or learning etc. I would recommend not providing such interpretations in the result section and addressing these points in the discussion.

Similar to above, statements like "These results suggest that the hippocampus updates information that is independent of the target TTC". The data show that higher hippocampal activity is linked to greater improvement across trials independent of the timing of the trial. The point about updating is rather speculative and should be presented in the discussion instead of the result section.

The reviewer is referring to two statements in the results section that reflect our interpretation rather than a description of the results. In response to the reviewer’s comment, we therefore removed the following statement from the results.

Instead, these results are consistent with the notion that hippocampal activity signals the updating of task-relevant sensorimotor representations in real-time.

In addition, we replaced the remaining statement by the following. We feel this new statement makes clear why we conducted the analysis that is described without offering an interpretation of the results that were presented before.

Page 8: We reasoned that updating TTC-independent information may support generalization performance by means of regularizing the encoded intervals based on the temporal context in which they were encoded.

Author Response

Reviewer #2 (Public Review):

In this MEG work employing two types of bistable perception test and unique regression analyses, the authors identified different neural frequencies to different components of visual perception: its content and stability.

Strengths:

This study has a nice set of three different experiments to clarify neural differences between content, memory and stability of visual perception.

The state space analysis appears to be powerful to identify such different neural signatures for different cognitive components as well.

Weaknesses:

Despite such strengths, this work may have the somewhat critical weakness specified in the recommendations for the authors.

First, in the analysis to identify content-specific neural frequency, the authors concluded that the SCP is more relevant to the visual perceptual content compared to the neural activity in the alpha and beta-band frequencies. In my impression, to claim this, it would be necessary to show statistically significant differences in the prediction accuracy between the SCP and the other frequencies. Given the not-so-high prediction accuracy seen in the SCP-based analysis, such statistical supports appear essential.

We have now directly compared decoding accuracy for SCP and alpha/beta oscillations, which showed statistically significant differences in both the ambiguous and unambiguous conditions for both ambiguous images. We have added these results as a supplementary figure (new Figure 2—figure supplement 1).

Second, two behavioural metrics in the neural state space analysis-i.e., Switch and Direction-may be too arbitrary. As suggested by the power-law distribution of the percept duration, the neural dynamics during seemingly stable percept may not be able to be described in linear functions. Instead, the brain may go back and forth between several neural states even when we are thinking we're experiencing stable visual consciousness. If so, the current definition of the Switch metric and Direction index, which seems to be based on the behaviour of the Switch index, may be arbitrary. In other words, I feel the authors may have to elaborate the rationale for the definitions of such metrics.

First, we note it is generally accepted in the field that the distribution of percept durations follows a gamma distribution instead of a power-law distribution (e.g., Sterzer et al., TiCS 2009; Blake & Logothetis Nature Rev. Neurosci 2002; Kleinschmidt et al., 1998; Leopold et al., TiCS 1999), and microswitches have not been reported either using the more classic task as that employed here or the more recently developed ‘no-report’ task of using eye-tracking statistics to deduce perceptual switches without overt report (e.g., Frassle et al., J Neurosci 2014).

Second, while brain activity may fluctuate during these time periods, it never crosses the threshold of evoking a conscious report, and thus we would expect that such fluctuations, if they do occur, would be of a lower magnitude than those that do produce a conscious report.

Most importantly, our goal here is to define behavioral metrics in order to identify components of neural dynamics underpinning the relevant aspect of behavior. As such, our definition of the behavioral metric should not be directly informed by observed spontaneous dynamics of brain activity (especially those that may be observed in the data but are of unclear relevance to perceptual switching); otherwise the analysis would be prone to circularity and spurious correlations (i.e., using observed brain dynamics to inform construction of behavioral metrics might pick up aspect of brain dynamics not really relevant to behavior in the analysis results).

Finally, the timing characteristics of ‘Switch’ and ‘Direction’ behavioral metrics are not arbitrary; instead they are the simplest behavioral functions that allow a comparison of pre- and post-switching periods (or when the percepts might be in the ‘stabilizing’ phase vs. the ‘destabilizing’ phase). Nevertheless, the regression analysis can pick up on other temporal patterns of changes not exactly the same as our defined behavioral metric. This can be seen for SCP and beta activity projected onto the Direction axis, where it has the lowest value at ~20th percentile of the trial (not 50th percentile as assumed by the behavioral metric). To confirm that the analysis is not highly dependent on the precise timing definition of the behavioral metrics, we ran a control analysis, where the switching point was set at 30%tile (rather than 50%tile as in the original analysis). This control analysis resulted in a similar pattern of neural results (Figure R1).

Figure R1: Changing temporal behavior definition (switching point moved from 50th percentile to 30th percentile of percept duration) does not significantly alter the neural results. Compare to Figure 4—figure supplement 1, ‘Switch’ and “Direction’ Columns.

Author Response

Public Evaluation Summary

The authors aim to tackle a fundamental question with their study: whether there is a direct age-associated increase of transcriptional noise. To investigate this question, they develop tools to analyze single-cell sequencing data from mouse and human aging datasets. Ultimately, application of their novel tool (Scallop) suggests that transcriptional noise does not change with age, changes in transcriptional noise can be attributed to other sources such as subtle shifts in cell identity. This study is in principle of broad interest, but it currently lacks a definitive demonstration of the robustness of Scallop. Systematic testing of this new package would ultimately strengthen the key conclusion of the work and give additional users more confidence when using the tool to estimate expression noise.

We have now attempted to further demonstrate the robustness of Scallop by performing a more systematic analysis and a side-by-side comparison to other existing methods using a set of artificially generated datasets. These analyses have resulted in the inclusion of six supplementary figures that are presented in the subsections Scallop membership score accurately identifies transcriptionally noisy cells, Ability to detect noisy cells within cell types, Effect of cellular composition, Effect of dataset size, Effect of feature expression and Effect of cell type marker expression within the Results section of the revised manuscript.

We have also included a supplementary figure showing an in-depth analysis of a dataset where ageassociated increase in transcriptional noise was detected using alternative methods, but whose closer dissection has revealed that the difference in noise is due to a single donor and to the choice of methods. We discuss this is in the subsection Distance-to-centroid methods detect transcriptionally stable cell subtypes as transcriptional noise within the Results section.

Finally, we have revised the manuscript to clarify the main points raised by the reviewers: the definition of transcriptional noise, the reasoning behind the choice of the single-cell aging datasets and Leiden’s rationale. Also, we have expanded the description of the method to make the definition of membership score more clear to the readers, and discussed the implications of our main findings (a lack of evidence for age-related transcriptional noise) in the broader context of theories of aging.

Reviewer #1 (Public Review):

In the present study, Ibanez-Sole et al evaluate transcriptional noise across aging and tissues in several publicly available mouse and human datasets. Initially, the authors compare 4 generalized approaches to quantify transcriptional noise across cell types and later implement a new approach which uses iterative clustering to assess cellular noise. Based on implementation of this approach (scallop), the authors survey noise across seven sc-seq datasets relevant for aging. Here, the authors conclude that enhanced transcriptional noise is not a hallmark of aging, rather changes in cell identity and abundances, namely immune and endothelial cells. The development of new tools to quantify transcriptional noise from sc-seq data presents appeal, as these datasets are increasing exponentially. Further, the conclusion that increased transcriptional noise is not a defined aspect of aging is clearly an important contribution; however, given the provocative nature of this claim, more comprehensive and systematic analyses should be performed. In particular, the robustness and appeal of scallop is still not sufficiently demonstrated and given the complexity (multiple tissues, species and diverse relative age ranges) of datasets analyzed, a more thorough comparison should be performed. I list a few thoughts below:

Initially, the authors develop Decibel, which centralizes noise quantification methods. The authors provide schematics shown in Fig 1, and compare noise estimates with aging in Fig 2 - Supplement 2. Since the authors emphasize the necessary use of scallop as a ”better” pipeline, more systematic comparisons to the other methods should be made side-by-side.

We thank the reviewer for their positive assessment of the manuscript and their suggestions. We agree that side-by-side benchmarking of Scallop with the methods implemented in Decibel, as well as a more thorough analysis on the effect of different features such as dataset size, cellular composition, etc. might have on the output of Scallop will reinforce the main points of the manuscript. To experimentally respond to these requests, we took advantage of a set of four artificial datasets previously generated by us with the R package splatter (v1.10.1; as described in Ascensión et al. [1]). In the present work, we first run a side-by-side comparison between Scallop and two distance-to-centroid (DTC) methods on the four artificial datasets with increasing degrees of transcriptional noise present in them (the novel data are included as Figure 1 – Figure supplement 1 in the revised manuscript). Then, we compared Scallop to one DTC method regarding their ability to detect noisy cells in different cell types (Figure 1 – Figure supplement 2). Finally, we implemented four simulations to test the effect of the following features on the performance of Scallop: cellular composition (Figure 1 – Figure supplement 3), dataset size (Figure 1 – Figure supplement 4), number of genes (Figure 1 – Figure supplement 5) and marker gene expression (Figure 1 – Figure supplement 6). A summary of these results follows.

Side-by-side comparison of Scallop vs DTC methods

Each of the four artificial datasets used consists of 10K cells, from 9 populations, named Group1 to Group9, with the following relative abundances: 25, 20, 15, 10, 10, 7, 5.5, 4, and 3.5%, respectively. The four datasets only differ in the de.prob parameter used in their generation. The de.prob parameter determines the probability that a gene is differentially expressed between subpopulations within the dataset. The greater the de.prob value, the more differentially expressed genes there will be between clusters, meaning that the different cell types present in the dataset will cluster in a more robust way. Decreasing the value of de.prob results in datasets with noisy cells, with populations that do not have such a strong transcriptional signature. In order to study how Scallop can capture the degree of robustness with which cells of the same cell type cluster together, we selected four de.prob values (0.05, 0.016, 0.01 and 0.005) and measured transcriptional noise using Scallop and two DTC methods, the whole transcriptome-based Euclidean distance to cell type mean and the invariant gene-based Euclidean distance to tissue mean expression. These two methods were selected because GCL does not yield a transcriptional noise measure per cell, so no comparisons can be made with respect to the amount of noisy cells the method is able to detect within a cluster. Similarly, comparing Scallop to the ERCC spike in-based method was not possible for artificial datasets. Importantly, these analyses showed that Scallop, unlike DTC methods, was able to discern between the core transcriptionally stable cells within each cell type cluster from the more noisy cells that lie in between clusters (provided in the Figure 1 - Supplement 1 of revised manuscript).

Effect of dataset features on the performance of Scallop

We simulated five artificial datasets with the same nine cell type populations but whose relative abundances were different between datasets. We used the imbalance degree (ID) to measure class imbalance in each of them and to make sure that the selected cell compositions represented a wide range of imbalance degrees (to this end, we explored ID values between 1.2 and 5.3). The ID provides a normalized summary of the extent of class imbalance in a dataset in so-called ”multiclass” settings, that is to say, where more than two classes are present. It was specifically developed to improve the commonly used imbalance ratio (IR) measurement, whose calculation only considers the abundance of the most and the least popular classes and which gives the same summary for datasets with different numbers of minority classes. The presence of multiple minority classes is not uncommon in single-cell RNAseq datasets, as tissues might contain several rare cell types. We observed that the transcriptional noise measurements provided by Scallop were very robust to changes in imbalance degree (see Figure 1 - Supplement 3), both in qualitative and in quantitative terms. For instance, Group2 and Group8 were always detected as the most stable and noisiest cell types, respectively, regardless of their relative abundance in the dataset, and their average percentage of noise had little variation between different ID values: it ranged between 0-0.14% (Group2) and 16-18% (Group8).

The effect of dataset size (number of cells) and the number of genes was evaluated by generating versions of an artificial dataset where cells/genes had been subsampled from an original artificial dataset (the one generated with de.prob=0.001). We tested datasets sized 1,000-10,000 cells and with a number of genes between 5,000 and 14,000. Dataset size had nearly no impact on the transcriptional noise measurements provided by Scallop (Figure 1 - Supplement 4 of the revised manuscript). The average percentage of transcriptional noise per cell type remained within a narrow range as we implemented a ten-fold increase in dataset size. Perhaps more strikingly, removing the expression of most genes did not substantially impact transcriptional noise measurements per cell type (Figure 1 - Supplement 5). The variation when removing half of the genes (7,000 genes) was minimal, and we did not see important changes in transcriptional noise measurements unless over 60% of the genes from the original dataset were removed. For example, Figure 1 - Supplement 5C shows that noise measurements suffer important variations when removing 8,000 and 9,000 genes (and therefore keeping 6,000 and 5,000 genes, respectively), but only some cell types (Groups 4, 7, 8 and 9) were affected by these variations.

In order to measure the effect marker gene expression has on the membership with which cells are assigned to their cell type cluster, we ran a simulation where the top 10 markers for a cell type were removed from the dataset one by one, so that the first simulation lacked the expression of the Top1 marker, the second simulation had the effect of the first 2 markers removed (Top1 and Top2), and so on. Then, we ran Scallop on each of the resulting datasets and observed a steady increase in transcriptional noise associated with that cell type. This provided evidence that the strength of cell type marker expression in a cluster is directly related to its transcriptional stability (or lack of transcriptional noise). We included the result of this experiment in the revised version of the manuscript (Figure 1 - Supplement 6).

In conclusion, by using artificially generated datasets where the ground truth (cell type labels, degree of noise, etc) was known, the newly provided systematic analyses showed that Scallop had a remarkably robust response to said changes in dataset features, further reinforcing the manuscript conclusions.

For example, scallop noise estimates (Fig 2) compared to other euclidean distance-based measures (Fig 2 supplement 2) looks fairly similar.

It is true that some datasets show similar trends regardless of the transcriptional noise quantification method. For instance, the murine brain dataset by Ximerakis et al. shows no overall change in noise between the age groups across different methods. However, we do observe important differences in other examples. This is the case of the human pancreas dataset by Enge et al. and the human skin dataset by Solé-Boldo et al., where not only the magnitude but also the directionality of the trend are different depending on the method used to measure noise. In the former, three methods (Scallop, invariant gene-based Euclidean distance to average tissue expression and GCL) show an age-related increase in noise, whereas one method (whole transcriptome-based Euclidean distance to the cell type mean) shows a decrease in noise. In the latter, two methods (Scallop and GCL) yield a decrease in noise and the two DTC methods measure a mild increase in noise. These inconsistencies can now be reconciled with our proposed explanation that said ”noise” may actually be referring to substantially different biology in the diverse experimental settings.

Are downstream observations (ex lung immune composition changes more than noise) supported from these methods as well? If so, this would strengthen the overall conclusion on noise with age, but if not, it would be relevant to understand why.

Studying changes in cell type composition in the lung and other aged tissues would be highly pertinent. Nevertheless, we have measured changes in cell type composition using only one method that is based on Generalized Linear Models, covered in the subsection Age-related cell type enrichment of the Methods. The methods that we have compared in our study (DTC methods, ERCC-based methods, GCL, etc.) were all designed to measure transcriptional noise, but not changes in cell type composition.

Whether the effects of cell type composition changes are bigger than changes in noise for the rest of the methods used to measure noise was probably not clear enough in the original manuscript. We found no evidence for an increase in noise associated with aging, regardless of the method used. Although not included in the manuscript, we did generate heatmaps similar to the one shown in Figure 3B for each of the noise quantification methods. However, as the heatmap on the right side (the one showing cell type enrichment) was identical in each figure, we considered them to be redundant and decided not to include them, since they did not provide any additional insight besides giving more examples of lack of evidence for transcriptional noise, this time at the cell type level. We consider that the lack of evidence was already well demonstrated in the previous analyses (Figure 2 and Figure 2 - Supplement 2.

Similarly, the ’validation of scallop seems mostly based on the ability to localize noisy vs stable cells in Fig 1 supplement 1 and relative robustness within dataset to input parameters (Fig 1 supplement 2). A more systematic analysis should be performed to robustly establish this method. For example, noise cell clustering comparisons across the 7 datasets used. In addition, the Levy et all 2020 implemented a pathway-based approach to validate. Specifically, surrogate genes were derived from GCL value where KEGG preservation was used as an output. Similar additional types of analyses should be performed in scallop.

We believe that this legitimate concern is now solved with the newly included data. In particular, with the systematic comparison between Scallop and DTC methods on three artificially generated datasets with different degrees of transcriptional noise provided in Figure 1 - Supplement 2. The ability of Scallop to detect cells that are particularly noisy within a cell type, or cells that lie between cell types, may represent its biggest advantage with respect to other methods. DTC methods fail to discern between stable and noisy cells within cell types. Also, in our analysis, DTC methods were unable to distinguish between cell types that have a marked transcriptional program (which systematically cluster together) and those that have a less clear transcriptomic identity (which have at least part of their cells be assigned to other cell types across bootstrap iterations). However, comparing the performance of Scallop on the same datasets showed that our method was able distinguish between the two cases.

The conclusion that immune and endothelial cell transcriptional shifts associate more with age than noise are quite compelling, but seem entirely restricted to the mouse and human lung datasets. It would be interesting to know if pan-tissues these same cell types enrich age-related effects or whether this phenomenon is localized.

We agree with the reviewer that it would be very interesting to see whether a change in cell type composition (and particularly, an increase in abundance of immune cell types) is observed in aged tissues other than the lung. Qualitative cell type composition changes in the aging lung have been described in the literature [5]. Specifically, the higher abundance of immune cell types was observed in a single-nucleus RNAseq dataset of cardiopulmonary cells in Macaca fascicularis [6]. However, we believe that trying to answer the question whether this phenomenon holds in other tissues would require a systematic analysis of several datasets for each tissue with a sufficient number of donors/individuals in each of them. This is because our approach to measure age-associated cell type enrichment using generalized linear models relies heavily on having multiple biological replicates for each age group. Unfortunately, this is not the case for most published single-cell RNAseq datasets of aging. In any case, we have toned down the last sentence in the subsection Changes in the abundance of the immune and endothelial cell repertoires characterize the human aging lung by making it more clear that our claim regarding changes in the cellular composition of aged tissues is based on lung datasets (the text in italics represents what was added in the revised version of the manuscript):

"Even though the evidence for changes in tissue composition are based on a single tissue, we hypothesize that these facts may have influenced previous analyses of transcriptional noise associated with aging."

As discussed in the original manuscript, there is evidence published by other groups pointing out to pantissue changes in cellular composition with age, which undoubtedly will influence those analyses that did not pay attention to cellular composition changes in the datasets that they compared. Cellular composition is in fact a very important aspect that has been greatly overlooked. In fact, only one [7] out of the seven articles that had measured transcriptional noise in aging (the datasets used in Figure 2) had attempted to remove its effect by subsampling cells to balance compositions between age groups prior to their noise analysis. In any case, we do not believe this is the only phenomenon underlying the purported increase in transcriptional noise associated with age. Each dataset will most probably have different issues that the authors originally misread as an increase in noise or loss of cellular identity of a particular organ or tissue. As an additional example of such phenomena, we have now included a re-analysis of the data by Enge et al. [3] on ”noisy” β-cells in the aged human pancreas (Figure 5–Figure supplement 2 of the revised manuscript). In this case, rather than observing an age-dependent pattern, the 21-year-old donor presents much lower transcriptional noise values than the rest of the donors. However, there is no significant difference between the 22-year-old donor and the rest of the donors. We conclude that the statistically significant differences between the ”young” and ”old” age categories can be attributed to the abnormal noise values obtained for the 21-year-old donor, of uncertain origin. Finding out all causes of apparent transcriptional noise in other organs and tissues would be too lengthy, and certainly out of scope for the present manuscript.

Related to these, there does not seem to be a specific rationale for why these datasets (the seven used in total or the lung for deep-dive), were selected. Clearly, many mouse and human sc-RNA-seq datasets exist with large variations in age so expanding the datasets analyzed and/or providing sufficient rationale as to why these ones are appearing for noise analyses would be helpful. For example, querying ”aging” across sc-seq datasets in Single cell portal yields 79 available datasets: https://singlecell.broadinstitute. org/single_cell?type=study&page=1&terms=aging&facets=organism_age%3A0%7C103%7Cyears.

We now realize that the reasoning behind our selection of aging datasets was not sufficiently clear in the original manuscript. We thank the reviewer for pointing out this omission. We have made a more explicit reference to Appendices 2, 3, 4 and 6 in the revised manuscript. The seven selected scRNAseq datasets are those where transcriptional noise had originally been measured by the authors, using the computational methods that we later implemented in Decibel. Our aim was to first recapitulate previous reports of transcriptional noise using our novel method (Scallop). Thus, we downloaded all publicly available scRNAseq datasets of aged tissues where transcriptional noise had explicitly been measured. Some of them had reported an increase in transcriptional noise only in some cell types (for instance, the human aged pancreas dataset by Enge et al. [3]), whereas others found an increase in most cell types [7]. Appendix 2 summarizes the main features of those seven datasets (tissue, organism and number of cells) and provides information on whether an increase in transcriptional noise was observed in the original article where they were published. Additionally, the ”scope” column indicates where that increase was found (in which cell types), and the ”Method” column briefly describes the computational method used to measure transcriptional noise in that article. Appendix 3 provides information on the final datasets that were used in our analysis (Figure 2). Not every sample from the original dataset was included, so the inclusion criteria are specified there, as well as the number of cells, individuals and age of each of the cohorts. Appendix 4 shows the abnormal count distribution of two samples that were discarded from the Kimmel lung dataset. As for the selection of lung for the deep dive, the reason was that this was the organ with most datasets available, both for mouse and human. Appendix 6 provides information on the number of cells and donors per age cohort in the human lung datasets included in this study.

We have included the following sentence in the Increased transcriptional noise is not a universal hallmark of aging subsection in the Results:

"We provide a summary of the main characteristics of each dataset, as well as the findings regarding transcriptional noise obtained in each of the original studies, whether changes in transcriptional noise were restricted to particular cell types, and the computational method used to measure noise (see Appendix 2)."

The analysis that noise is indistinguishable from cell fate shifts is compelling, but again relies on one specific example where alternative surfactant genes are used as markers. The same question arises if this observation holds up to other cell types within other organs. For example the human cell atlas contains over dozens of tissue with large variations in age (https://www.science.org/doi/10.1126/science. abl4290).

We sympathize with this comment but hope that the reviewer will agree with us that providing an additional example of different phenomena originally reported as ”transcriptional noise” (in this case in aged human pancreas; see Figure 5 – Figure supplement 2), but actually reflecting something else, may be sufficient to prevent interested readers. In our opinion, it is likely that diverse phenomena will underlie the purported increases in transcriptional noise, and a re-analysis should be made case-by-case. We can only hope that researchers in the field re-analyze the available aging datasets in this new light.

Reviewer #2 (Public Review):

In this manuscript, Ibanez-Sole et al. focus on an important open question in ageing research; ”how does transcriptional noise increase at the cellular level?”. They developed two python toolkits, one for comparison of previously described methods to measure transcriptional noise, Decibel, and another one implementing a new method of variability measure based on cluster memberships, Scallop. Using published datasets and comparing multiple methods, they suggest that increased transcriptional noise is not a fundamental property of ageing, but instead, previous reports might have been driven by age-related changes in cell type compositions.

I would like to congratulate the authors on openly providing all code and data associated with the manuscript. The authors did not restrict their paper to one dataset or one approach but instead provided a comprehensive analysis of diverse biology across murine and human tissues.

While the results support their main conclusions, the lack of robustness/sensitivity measures for the methods used makes it difficult to judge the biology.The authors use real data to compare between methods but using synthetic data with known artificial ’variability’ across cell clusters can first establish the methods, which would make the results more convincing and easier to interpret. Despite the comprehensive analysis of biological data, a detailed prior description of how the methods behave against e.g. the number of cells in each cell type cluster, the number of cell types in the dataset, and % feature expression, would make the paper more convincing. Once the details of the method is provided, the python toolkit can be widely used, not limited to the ageing research community. I am also concerned that a definition of ’transcriptional noise’ (e.g. genome-wide noise, transcriptional dysregulation in cell-type-specific genes, noise in certain pathways) and its interpretation with regard to the biology of ageing is missing. Differences in different methods could be explained by the different biology they capture. Moreover, the interpretation of a lack of different types of variability may not be the same for the biology of ageing.

Increased transcriptional noise is compatible with genomic instability, loss of proteostasis and epigenetic regulation. Showing a lack of consistent transcriptional noise can challenge the widespread assumptions about how these hallmarks affect the organism. Overall, I found the paper very interesting and central to the field of ageing biology. However, I believe it requires a more detailed description of the methods and interpretations in the context of biology and theories of ageing.

We thank the reviewer for their positive assessment of the manuscript and their suggestions. We respond to each of the specific comments below.

Major comments

1) The concept of transcriptional noise is central to the manuscript; however, what the authors consider as transcriptional noise and why is not clear. Genome-wide vs. function or cell-type specific noise could have different implications for the biology of ageing. In line with this, a discussion of the findings in the context of theories of ageing is necessary to understand its implications.

We thank the reviewer for pointing out the lack of clarity in this key point. The use of the ”transcriptional noise” term in the literature is quite heterogeneous, and we agree that the lack of a consensus definition may be confusing to the reader. For this reason, we adopted in the introduction the definition by Raser and O’Shea [8] as ”the measured level of variation in gene expression among cells supposed to be identical”, i.e. the sum of both intrinsic and extrinsic noise as previously defined by Swain and colleagues [9, 10]. In our opinion, this is generally what the literature of age-associated transcriptional noise is referring to.

With Scallop, we aimed to translate this concept to the context of single-cell RNAseq datasets, where clusters obtained using a community detection algorithm are typically annotated as distinct cell types.

Therefore, we aimed to measure transcriptional noise here defined as ”lack of membership to cell type clusters”. When running a clustering algorithm iteratively, if a cell is not unambiguously assigned to the same cluster, we consider it to be noisy. Conversely, when a cell consistently clusters with the same group of cells, we consider it to be stable. The membership score we use as a measure of stability is the frequency with which any given cell was assigned to the same cluster across all iterations.

We have included in the Results section an explicit reference to the Methods subsection that explains how Scallop works in detail, so that the readers can easily find that information:

"A detailed description of the three steps of the method (bootstrapping, cluster relabeling and computation of the membership score) is provided in the Scallop subsection in the Methods."

Additionally, we have now realized that the formula to compute the membership score might be more easily understood if we renamed the freq_score as freq_score(c), to make it clear that each cell is assigned a score. Also, we have used n and m instead of i and j in this notation, to avoid confusing the readers with the notation used in the previous section, where i and j represented the i-th and j-th bootstrap iterations. Finally, we have included a small paragraph to clarify what each component of the formula refers to. Below we show the formula and text included in the Methods section of the revised manuscript:

"Where |cn| is the number of times cell c was assigned to the n-th cluster, and Pm∈clusters |cm| is the sum of all assignments made on cell c, which is the same as the number of times cell c was clustered across bootstrap iterations."

Thus, and in order to accommodate this reviewer’s concerns, we have now included this exact definition of how we measure noise plus a statement making clear that we refer to the sum of both intrinsic and extrinsic noise aspects, with no distinction among them.

Similarly, we had discussed our findings in the framework of different theories of aging, such as their potential relationship to some of the established hallmarks of aging (genomic instability, epigenetic deregulation and loss of proteostasis), as well as with more recent theories of aging such as cell type imbalance in aged organs [11] and inter-tissue convergence [12]. However, it is now clear to us that this was not enough so we have now expanded these paragraphs to make our understanding of the work implications better understood. More specifically:

"Our results suggest that transcriptional noise is not a bona fide hallmark of aging. Instead, we posit that previous analyses of noise in aging scRNAseq datasets have been confounded by a number of factors, including both computational methods used for analysis as well as other biology-driven sources of variability."

2) While I found the suggested method, Scallop, quite exciting and valuable, I would suggest including a number of performance/robustness measures (primarily based on simulations) on how sensitive the method is to the number of cells in each cell type (cellular composition), misannotations, % feature expression (number of 0s) etc.:

We have analyzed the effect of cellular composition and the percentage of feature expression by using artificially generated datasets (see Figure 1 - Supplements 3 and 5, respectively; and section Effect of dataset features on the performance of Scallop in the response to reviewer #1). Although studying the effect of misannotations on downstream analysis is important, we believe that Scallop was already designed so that its effects could be avoided, since the membership is measured for each cluster (and not for each cell type label). That is to say, a reference clustering is obtained at the beginning of the pipeline and memberships are computed using that output as a reference, which means Scallop noise values attributed to each cell are not affected by the original labeling of the dataset.

The output of these analyses reinforced our original conclusions, and it is now included in the Results section:

"In order to characterize and validate our method for transcriptional noise quantification, we conducted three types of analyses. First, we used artificially generated datasets containing various degrees of transcriptional noise to compare the performance of Scallop and DTC methods side-by-side, regarding their ability to measure transcriptional noise and detect noisy cells within cell types. Next, we ran simulations using artificial datasets in order to study the effect of a number of dataset features on the performance of Scallop: cellular composition, dataset size, number of genes and marker expression. Finally, we graphically evaluated the output of Scallop on a dataset of human T cells, we analyzed its robustness to its input parameters, and we studied the relationship between membership and robust marker expression, using a PBMC dataset."

2.1) Most importantly, knowing that cell-type composition changes with age, it is important to know how sensitive community detection is to the number of cells in each cell type. While the average can be robust, I wonder if the size of the cell-type cluster affects membership (voting).

We have included an analysis on a set of artificial datasets with different cellular compositions to evaluate the performance of Scallop in the presence of different degrees of class imbalance (see Figure 1 - Supplement 3). We explain the output of this analysis, which reinforces the algorithm’s robustness, in the Results section:

"Next, we ran a series of simulations on artificially generated datasets to evaluate the performance of Scallop in the presence of different levels of class imbalance, dataset size, number of genes, and different degrees of expression of cell type markers. Our analysis showed that Scallop was remarkably robust to changes in cellular composition (see Figure 1 - Supplement 3). Both the average percentage of noise and the distribution remained unchanged for a wide range of class imbalance degrees. Similarly, altering the dataset size (number of cells) and the number of genes of an artificial dataset did not cause any major changes on the transcriptional noise values attributed to each cell type (see Figure 1 - Supplements 4 and 5). Additionally, we conducted an analysis where we identified the 10 most differentially expressed gene markers for a cell type and measured the transcriptional noise associated with that cell type as we removed the expression of those genes from the dataset (Figure 1 - Supplement 5). Transcriptional noise steadily increased as we removed the effect of the top marker genes that defined the cell type under study (see Figure 1 - Supplement 5B). This experiment provides further evidence on how strong marker expression is related to robust cell type identity and how the lack of it results in transcriptional noise."

3) Although the Leiden algorithm is widely used by many single-cell clustering methods, since the proposed methodology is heavily dependent on clustering, I suggest including a description of the Leiden algorithm.

We agree that understanding how community detection algorithms in general –and Leiden in particular– work is crucial to understand the core of the paper, so we have included a brief introduction to these methods in the Methods section, at the beginning of the Scallop subsection:

Leiden is a graph-based community detection algorithm that was designed to improve the popular Louvain method [13]. Graph-community detection methods take a graph representation of a dataset. In the context of single-cell RNAseq data, shared nearest neighbor (SNN) graphs are commonly used. These are graphs whose nodes represent individual cells and edges connect pairs of cells that are part of the K-nearest neighbors of each other by some distance metric. The aim of community detection algorithms like Leiden is to find groups of nodes that are densely connected between them, by optimizing modularity. For a graph with C communities, the modularity (Q) is computed by taking, for each community (group of cells), the difference between the actual number of edges in that community (ei) and the number of expected edges in that community ( K2/1/2m).

Where r is a resolution parameter (r > 0) that controls for the amount of communities: a greater resolution parameter gives more communities whereas a low resolution parameter fewer clusters. Since maximizing the modularity of a graph is an NP-hard problem, different heuristics are used, and Leiden has shown to outperform Louvain in this task both in terms of quality and speed [14]. However, users can choose to run the Louvain method instead by setting the parameter clustering="louvain" in the initialization of the Bootstrap object.

3.1) Most importantly, the authors comment that they found stronger expression of cell-type specific markers in the cells with high membership values - is it already a product of the Leiden algorithm that it weighs highly variable (thus cell-type specific) features higher - resulting in better prediction of cell-types for cells with strong cell-marker expression? It is important to make a description of transcriptional noise at this stage as it could be genome-wide or more specific to cell-type markers. Can authors provide any support that their method can capture both?

We agree with the reviewer that finding a stronger expression of cell-type markers in cells with high membership values is indeed something we expected. The graph representation of the dataset taken as input by Leiden is built after running highly variable gene detection and PCA. The neighbors of each cell are detected based on the expression of genes that are highly variable, as the reviewer pointed out, so genes that are differentially expressed between cells are more likely to contribute to the clusters found by Leiden.

Whether Scallop measures genome-wide or cell type-specific noise (or a mixture of both) is a very interesting question. Clusters in single-cell RNA sequencing datasets are often mainly driven by the presence/absence of a few cell type markers, rather than changes in expression levels of broader sets of genes. Moreover, it has been shown that single-cell RNAseq datasets generally preserve the same population structure even after data binarization [15]. This is a consequence of the sparsity of single-cell RNAseq datasets. In our case, any difference in expression between one cluster vs the rest of the cells in the dataset –be it the expression of a gene that was not detected in the rest of the cells or a higher expression of a gene whose presence is weaker in other clusters– will certainly have an impact on the output of every downstream analysis, from clustering to dimensionality reduction. The influence of the expression of cell type-specific markers on Scallop membership has been demonstrated in several analyses. First, the simulation where we measured the impact of removing the 10 most defining markers for a particular cell type on transcriptional noise measurements (included in the Figure 1 - Supplement 6 of the revised manuscript). Also, Figure 5 provides evidence that the differential expression of a handful of genes (in this case, genes coding for surfactant proteins) can have an impact on the clustering solutions obtained for a set of human alveolar macrophages, and this in turn influences the membership scores obtained with Scallop. In essence, Scallop merely provides a measure of the robustness of clustering at the single-cell level, so any type of transcriptional noise might have an impact on Scallop memberships, provided it is sufficiently strong to influence the output of the clustering algorithm used. In other words, the fact Scallop membership captures a mixture of both types of noise (genome-wide and that associated with cell type-specific markers) is a consequence of the influence both types of noise have on clustering.

4) The authors conclude that Scallop outperforms other methods through the analysis of biological data, where there is no positive and negative control. I suggest creating synthetic datasets (which could be based on real data), introducing different levels of noise artificially (considering biological constraints like max/min expression levels) and then testing the performance where the truth about each dataset is known. Otherwise, the definitions of noisy and stable cells, regardless of the method, are arbitrary.

Our initial focus was on biological datasets, were no positive and negative controls regarding transcriptional noise could be used, but we agree in the need of including an analysis using simulations on artificial datasets. We analyzed artificially generated datasets with known degrees of transcriptional noise in order to evaluate the performance of Scallop on a setting where the ground truth is known beforehand. The way we modeled transcriptional noise was by tuning the de.prob parameter, which determines the probability that a gene will be differentially expressed between clusters. The creation of these datasets is explained in detail in the Methods section of the revised manuscript, and specifically in the subsections Performance of Scallop and two DTC methods on four artificial datasets with increasing transcriptional noise. and Ability to detect noisy cells within cell types.

We have now included the following section in the Results:

"We compared the output of Scallop and two DTC methods (the whole transcriptome-based Euclidean distance to average cell type expression and the invariant gene-based Euclidean distance to average tissue expression) on four artificially generated datasets containing various levels of transcriptional noise. The analysis showed that Scallop, unlike DTC methods, was able to discern between the core transcriptionally stable cells within each cell type cluster from the more noisy cells that lie in between clusters (see Figure 1 - Supplement 1). We then compared one of the DTC methods to Scallop regarding their ability to detect noisy cells within each of the cell types, by plotting the top 10% noisiest and top 10% most stable cells and (see Figure 1 - Supplement 2A). Analyzing the distribution of noise values for each cell type separately revealed that Scallop can distinguish between clusters that mainly consist of transcriptionally stable cells from noisier clusters that do not have such a distinct transcriptional signature (Figure 1 - Supplement 2B."

Reviewer #3 (Public Review):

In this manuscript, Ibáñez-Solé et al aim to clarify the answer to a very basic and important question that has gained a lot of attention in the past ∼5 years due to fast-increasing pace of research in the aging field and development/optimization of single-cell gene expression quantification techniques: how does noise in gene expression change during the course of cellular/tissue aging? As the authors clearly describe, there have been multiple datasets available in the literature but one could not say the same for the number of available analysis pipelines, especially a pipeline that quantifies membership of single cells to their assigned cell type cluster. To address these needs, Ibáñez-Solé et al developed: 1. a toolkit (named Decibel) to implement the common methods for the quantification of age-related noise in scRNAseq data; and 2. a method (named Scallop) for obtaining membership information for single-cells regarding their assigned celltype cluster. Their analyses showed that previously-published aging datasets had large variability between tissues and datasets, and importantly the author’s results show that noise-increase in aging could not be claimed as a universal phenotype (as previously suggested by various studies).

We thank the reviewer for their positive assessment of the manuscript and their suggestions.

Comments:

1) In two relevant papers (doi.org/10.1038/s41467-017-00752-9anddoi.org/10.1016/j.isci. 2018.08.011), previous work had already shown what haploid/diploid genetic backgrounds could show in terms of intercellular/intracellular noise. Due to the direct nature of age/noise quantification in these papers, one cannot blame any computational pipeline-related issues for the ”unconventional” results. The authors should cite and sufficiently discuss the noise-related results of these papers in their Discussion section. These two papers collectively show how the specific gene, its protein half-life and ploidy can lead to similar/different noise outcomes.

We agree that we have failed to mention and sufficiently discuss the effects of measuring transcriptional noise from data generated via destructive experimentation, where no longitudinal analyses are possible. As aforementioned in the response to other reviewers, the body of literature on transcriptional noise is quite wide and based on heterogeneous assumptions. We have focused our efforts in measuring actual noise in scRNAseq aging datasets, which by definition imply sampling of different cells and thus make assumptions at the population level. We believe our results provide a different and interesting perspective into transcriptional noise and aging, but we agree with this reviewer in the need to discuss our findings in the context of other attempts to measure transcriptional noise in a more direct way. We have now included a brief discussion of the work by Sarnoski et al. and Liu et al.. This point is explained in more detail later in the letter.

2) While the authors correctly put a lot of emphasis on studying the same cell type or tissue for a faithful interpretation of noise-related results, they ignore another important factor: tracking the same cell over time instead of calculating noise from single-cell populations at supposedly-different age points. Obviously, scRNAseq cannot analyze the same cell twice, but inability to assess noise-in-aging in the same cell over time is still an important concern. Noise could/does affect the generation durations and therefore neighboring cells in the same cluster may not have experienced the same amount of mitotic aging, for example. Also, perhaps a cell has already entered senescence at early age in the same tissue. This caveat should be properly discussed.

The distinction between intrinsic and extrinsic noise and the impossibility to discern between the two in destructive experiments is a relevant point that we have now included in the Discussion (the newly added text is shown in italics):

"Transcriptional noise could be related to genomic instability [18], epigenetic deregulation [19, 20] or loss of proteostasis [21], all established hallmarks of aging. Some authors consider transcriptional noise to be a hallmark of aging in and of itself [22]. In any case, the origin of transcriptional noise is unclear, as it could arise from many different sources. Most importantly, it not possible to distinguish between intrinsic and extrinsic noise from a snapshot of cellular states, i.e., one cannot tell whether the observed differences between cells in a single-cell RNA experiment reflect time-dependent variations in gene expression or differences between cells across a population [23]. Interestingly, recent work by Liu et al. measuring intrinsic noise in S. cerevisiae showed that aging is associated with a steady decrease in noise, with a sudden increase in soon-to-die cells. Another longitudinal study found an increase extrinsic noise and a lack of change in intrinsic noise in diploid yeast [16]."

Regarding the caveat of cells of individuals in the Young groups showing signs of aging, we can only agree that this is correct: there will be cells sampled that already show signs of cellular damage in the absence of chronological aging. However this applies to every study of aging that samples cells in a destructive manner and it is generally assumed by the field that this is a discrete phenomenon that does not affect the overall results in a meaningful way.

3) Another weakness of this study is that the authors did not show the source/cause of decreasing/stable/increasing noise during aging. Understanding the source of loss of cell type identity is also important but this manuscript was about noise in aging, so it would have been nice if there could be some attempts to explain why noise is having this/that trend in differentially aged cell types in specific tissues.

The reviewer raises here a very important point that we would like to discuss in detail. The papers that we have re-analyzed generally assume that an increase in transcriptional noise and a loss in cell type identity are equivalent terms. However, as this reviewer points out, you could theoretically have cells that lose their cell type identity without a concomitant increase in transcriptional noise, for instance by a sharp decrease in a limited number of marker genes that collectively define that cell within a given cell type/cluster. Thus, transcriptional noise can certainly arise from different sources and several mechanisms have been proposed to explain its presence in the context of cellular aging. We agree with the reviewer that discussing how transcriptional noise could be related to aging is of interest to the readers. However, as pointed out in the responses to similar concerns by the other reviewers, our main finding is that we don’t detect meaningful and reliable increases in transcriptional noise associated with cell aging. Instead, what we see is a number of different technical and biological issues/phenomena that have been interpreted as transcriptional noise. We hope this reviewer will agree that the manuscript now presents a full and robust story and that finding the causes of up/down ”noise” trends in the different datasets may be more appropriately tackled by follow up studies.

4) In the discussion section, the authors say that ”Most importantly, Scallop measures transcriptional noise by membership to cell type-specific clusters which is a re-definition of the original formulation of noise by Raser and O’Shea.” It is not clear what the authors refer to by ”the original formulation of noise by Raser and O’Shea”. Intrinsic/extrinsic noise formulations?? Please be more specific.

We thank the reviewer for pointing this out, since we agree that the sentence needed to be reformulated for the sake of clarity. What we meant by the definition by Raser and O’Shea was ”the measured level of variation in gene expression among cells supposed to be identical”, which does not make any distinction between intrinsic and extrinsic noise. Since their definition is previous to the development of single-cell technologies, we meant to state our attempt to bring this classic concept to the context of single-cell RNAseq. Nowadays, cell clusters produced by a community detection algorithm are given cell type annotations depending on their expression of known cell type markers. What Scallop aims to measure is the extent of membership each individual cell has for their cluster as evidence of its transcriptional stability. In order to make this point more clear, we have now rewritten the paragraph as follows:

Most importantly, Scallop measures transcriptional noise by membership to cell type-specific clusters which is a re-definition of the original formulation of noise by Raser and O’Shea: measurable variation among cells that should share the same transcriptome. This is in stark contrast to measurements of noise including other phenomena (as demonstrated in Figure 5) by the distance-to-centroid methods prevalent in the literature.

References

[1] M. Alex Ascensión, Olga Ibáñez-Solé, Iñaki Inza, Ander Izeta, and Marcos J Araúzo-Bravo. Triku: A feature selection method based on nearest neighbors for single-cell data. GigaScience, 11, 2022. doi: 10.1093/gigascience/giac017.

[2] M. Ximerakis, S. L. Lipnick, B. T. Innes, S. K. Simmons, X. Adiconis, D. Dionne, B. A. Mayweather, L. Nguyen, Z. Niziolek, C. Ozek, V. L. Butty, R. Isserlin, S. M. Buchanan, S. S. Levine, A. Regev, G. D. Bader, J. Z. Levin, and L. L. Rubin. Single-cell transcriptomic profiling of the aging mouse brain. Nat Neurosci, 22(10), 2019. doi: https://doi:10.1038/s41593-019-0491-3.

[3] M. Enge, H. E. Arda, M. Mignardi, J. Beausang, R. Bottino, S. K. Kim, and S. R. Quake. Single-cell analysis of human pancreas reveals transcriptional signatures of aging and somatic mutation patterns. Cell, 171(2), 2017. doi: https://doi:10.1016/j.cell.2017.09.004.

[4] L. Solé-Boldo, G. Raddatz, and S. et al. Schütz. Single-cell transcriptomes of the human skin reveal age-related loss of fibroblast priming. Commun Biol, 3(188), 2020. doi: https://doi.org/10.1038/ s42003-020-0922-4.

[5] Jaime L. Schneider, Jared H. Rowe, Carolina Garcia-de Alba, Carla F. Kim, Arlene H. Sharpe, and Marcia C. Haigis. The aging lung: Physiology, disease, and immunity. Cell, 184(8):1990–2019, 2021. doi: 10.1016/j.cell.2021.03.005.

[6] Shuai Ma, Shuhui Sun, Jiaming Li, Yanling Fan, Jing Qu, Liang Sun, Si Wang, Yiyuan Zhang, Shanshan Yang, Zunpeng Liu, and et al. Single-cell transcriptomic atlas of primate cardiopulmonary aging. Cell Research, 31(4):415–432, 2020. doi: 10.1038/s41422-020-00412-6.

[7] I. Angelidis, L. M. Simon, and I. E. et al. Fernandez. An atlas of the aging lung mapped by single cell transcriptomics and deep tissue proteomics. Nature Communications, 2019. doi: https://doi.org/10. 1038/s41467-019-08831-9.

[8] Jonathan M. Raser and Erin K. O’Shea. Noise in gene expression: origins, consequences, and control. Science, 309(5743):2010–2013, 2005. doi: 10.1126/science.1105891.

[9] Michael B. Elowitz, Arnold J. Levine, Eric D. Siggia, and Peter S. Swain. Stochastic gene expression in a single cell. Science, 297:1183– 1186, 2002. doi: 10.1126/science.1070919.

[10] Peter S. Swain, Michael B. Elowitz, and Eric D. Siggia. Intrinsic and extrinsic contributions to stochasticity in gene expression. Proc Natl Acad Sci U S A., 99:12795–12800, 2002. doi: 10.1073/pnas.162041399.

[11] Alex Cagan, Adrian Baez-Ortega, Natalia Brzozowska, Federico Abascal, Tim H. H. Coorens, Mathijs A. Sanders, Andrew R. J. Lawson, Luke M. R. Harvey, Shriram Bhosle, David Jones, Raul E. Alcantara, Timothy M. Butler, Yvette Hooks, Kirsty Roberts, Elizabeth Anderson, Sharna Lunn, Edmund Flach, Simon Spiro, Inez Januszczak, Ethan Wrigglesworth, Hannah Jenkins, Tilly Dallas, Nic Masters, Matthew W. Perkins, Robert Deaville, Megan Druce, Ruzhica Bogeska, Michael D. Milsom, Björn Neumann, Frank Gorman, Fernando Constantino-Casas, Laura Peachey, Diana Bochynska, Ewan St. John Smith, Moritz Gerstung, Peter J. Campbell, Elizabeth P. Murchison, Michael R. Stratton, and Iñigo Martincorena. Somatic mutation rates scale with lifespan across mammals. Nature, 604: 517–524, 2022. doi: 10.1038/s41586-022-04618-z.

[12] Hamit Izgi, Dingding Han, Ulas Isildak, Shuyun Huang, Ece Kocabiyik, Philipp Khaitovich, Mehmet Somel, and Handan Melike Dönertas. Inter-tissue convergence of gene expression during ageing suggests age-related loss of tissue and cellular identity. eLife, 11, 2022. doi: 10.7554/eLife.68048.

[13] Vincent D Blondel, Jean-Loup Guillaume, Renaud Lambiotte, and Etienne Lefebvre. Fast unfolding of communities in large networks. Journal of Statistical Mechanics: Theory and Experiment, 2008(10): P10008, oct 2008. doi: 10.1088/1742-5468/2008/10/p10008. URL https://doi.org/10.1088/ 1742-5468/2008/10/p10008.

[14] V. A. Traag, L. Waltman, and N. J. van Eck. From louvain to leiden: guaranteeing well-connected communities. Scientific Reports, 9, 2019. doi: https://doi.org/10.1038/s41598-019-41695-z.

[15] Peng Qiu. Embracing the dropouts in single-cell rna-seq analysis. Nature Communications, 11(1), 2020. doi: 10.1038/s41467-020-14976-9.

[16] Ethan A. Sarnoski, Ruijie Song, Ege Ertekin, Noelle Koonce, and Murat Acar. Fundamental characteristics of single-cell aging in diploid yeast. iScience, 7:96–109, 2018. doi: 10.1016/j.isci.2018.08.011.

[17] Ping Liu, Ruijie Song, Gregory L. Elison, Weilin Peng, and Murat Acar. Noise reduction as an emergent property of single-cell aging. Nature Communications, 8(1), 2017. doi: 10.1038/s41467-017-00752-9.

[18] Jan Vijg. From dna damage to mutations: All roads lead to aging. Ageing Res Rev., 68(101316), 2021. doi: 10.1016/j.arr.2021.101316.

[19] Yuancheng Lu, Benedikt Brommer, Xiao Tian, Anitha Krishnan, Margarita Meer, Chen Wang, Daniel L. Vera, Qiurui Zeng, Doudou Yu, Michael S. Bonkowski, Jae-Hyun Yang, Songlin Zhou, Emma M. Hoffmann, Margarete M. Karg, Michael B. Schultz, Alice E. Kane, Noah Davidsohn, Ekaterina Korobkina, Karolina Chwalek, Luis A. Rajman, George M. Church, Konrad Hochedlinger, Vadim N. Gladyshev, Steve Horvath, Morgan E. Levine, Meredith S. Gregory-Ksander, Bruce R. Ksander, Zhigang He, and David A. Sinclair. Reprogramming to recover youthful epigenetic information and restore vision. Nature, 588(7836):124–129, 2020. doi: 10.1038/s41586-020-2975-4.

[20] Giorgio Oliviero, Sergey Kovalchuk, Adelina Rogowska-Wrzesinska, Veit Schwämmle, and Ole N. Jensen. Distinct and diverse chromatin proteomes of ageing mouse organs reveal protein signatures that correlate with physiological functions. eLife, 11(e73524), 2022. doi: 10.7554/eLife.73524.

[21] Jingyi Li, Yuxuan Zheng, Pengze Yan, Moshi Song, Si Wang, Liang Sun, Zunpeng Liu, Shuai Ma, Juan Carlos Izpisua Belmonte, Piu Chan, Qi Zhou, Weiqi Zhang, Guang-Hui Liu, Fuchou Tang, and Jing Qu. A single-cell transcriptomic atlas of primate pancreatic islet aging. Natl Sci Rev., 8(2): nwaa127, 2020. doi: 10.1093/nsr/nwaa127.

[22] Alexander R. Mendenhall, George M. Martin, Matt Kaeberlein, and Rozalyn M. Anderson. Cellto-cell variation in gene expression and the aging process. Geroscience, 43(1):181–196, 2021. doi: 10.1007/s11357-021-00339-9.

[23] Lucy Ham, Marcel Jackson, and Michael PH Stumpf. Pathway dynamics can delineate the sources of transcriptional noise in gene expression. eLife, 10, 2021. doi: 10.7554/elife.69324.

Author Response

Reviewer #1 (Public Review):

It is now widely accepted that the age of the brain can differ from the person's chronological age and neuroimaging methods are ideally suited to analyze the brain age and associated biomarkers. Preclinical studies of rodent models with appropriate neuroimaging do attest that lifestyle-related prevention approaches may help to slow down brain aging and the potential of BrainAGE as a predictor of age-related health outcomes. However, there is a paucity of data on this in humans. It is in this context the present manuscript receives its due attention.

Comments:

1) Lifestyle intervention benefits need to be analyzed using robust biomarkers which should be profiled non-invasively in a clinical setting. There is increasing evidence of the role of telomere length in brain aging. Gampawar et al (2020) have proposed a hypothesis on the effect of telomeres on brain structure and function over the life span and named it as the "Telomere Brain Axis". In this context, if the authors could measure telomere length before and after lifestyle intervention, this will give a strong biomarker utility and value addition for the lifestyle modification benefits. 2) Authors should also consider measuring BDNF levels before and after lifestyle intervention.

Response to comments 1+2: we agree that associating both telomere length and BDNF level with brain age would be interesting and relevant. However, we did not measure these two variables. We would certainly consider adding these in future work. Regarding telomere length, we now include a short discussion of brain age in relation to other bodily ages, such as telomere length (Discussion section):

“Studying changes in functional brain aging is part of a broader field that examines changes in various biological ages, such as telomere length1, DNA methylation2, and arterial stiffness3. Evaluating changes in these bodily systems over time allows us to capture health and lifestyle-related factors that affect overall aging and may guide the development of targeted interventions to reduce age-related decline. For example, in the CENTRAL cohort, we recently reported that reducing body weight and intrahepatic fat following a lifestyle intervention was related to methylation age attenuation4. In the current work, we used RSFC for brain age estimation, which resulted in a MAE of ~8 years, which was larger than the intervention period. Nevertheless, we found that brain age attenuation was associated with changes in multiple health factors. The precision of an age prediction model based on RSFC is typically lower than a model based on structural brain imaging5. However, a higher model precision may result in a lower sensitivity to detect clinical effects6,7. Better tools for data harmonization among dataset6 and larger training sample size5 may improve the accuracy of such models in the future. We also suggest that examining the dynamics of multiple bodily ages and their interactions would enhance our understanding of the complex aging process8,9. “

And

“These findings complement the growing interest in bodily aging indicated, for example, by DNA methylation4 as health biomarkers and interventions that may affect them.”

Reviewer #2 (Public Review):

In this study, Levakov et al. investigated brain age based on resting-state functional connectivity (RSFC) in a group of obese participants following an 18-month lifestyle intervention. The study benefits from various sophisticated measurements of overall health, including body MRI and blood biomarkers. Although the data is leveraged from a solid randomized control set-up, the lack of control groups in the current study means that the results cannot be attributed to the lifestyle intervention with certainty. However, the study does show a relationship between general weight loss and RSFC-based brain age estimations over the course of the intervention. While this may represent an important contribution to the literature, the RSFC-based brain age prediction shows low model performance, making it difficult to interpret the validity of the derived estimates and the scale of change. The study would benefit from more rigorous analyses and a more critical discussion of findings. If incorporated, the study contributes to the growing field of literature indicating that weight-reduction in obese subjects may attenuate the detrimental effect of obesity on the brain.

The following points may be addressed to improve the study:

Brain age / model performance:

1) Figure 2: In the test set, the correlation between true and predicted age is 0.244. The fitted slope looks like it would be approximately 0.11 (55-50)/(80-35); change in y divided by change in x. This means that for a chronological age change of 12 months, the brain age changes by 0.11*12 = 1.3 months. I.e., due to the relatively poor model performance, an 80-year-old participant in the plot (fig 2) has a predicted age of ~55. Hence, although the age prediction step can generate a summary score for all the RSFC data, it can be difficult to interpret the meaning of these brain age estimates and the 'expected change' since the scale is in years.

2) In Figure 2 it could also help to add the x = y line to get a better overview of the prediction variance. The estimates are likely clustered around the mean/median age of the training dataset, and age is overestimated in younger subs and overestimated in older subs (usually referred to as "age bias"). It is important to inspect the data points here to understand what the estimates represent, i.e., is variation in RSFC potentially lost by wrapping the data in this summary measure, since the age prediction is not particularly accurate, and should age bias in the predictions be accounted for by adjusting the test data for the bias observed in the training data?

Response to comment 1+2: we agree with the reviewer that due to the relatively moderate correlation between the predicted and observed age, a large change in the observed age corresponds to a small change in the predicted age. We now state this limitation in Results section 2.1:

“Despite being significant and reproducible, we note that the correlations between the observed and predicted age were relatively moderate.”

And discuss this point in the Discussion section:

“In the current work, we used RSFC for brain age estimation, which resulted in a MAE of ~8 years, which was larger than the intervention period. Nevertheless, we found that brain age attenuation was associated with changes in multiple health factors. The precision of an age prediction model based on RSFC is typically lower than a model based on structural brain imaging5. However, a higher model precision may result in a lower sensitivity to detect clinical effects6,7. Better tools for data harmonization among dataset6 and larger training sample size5 may improve the accuracy of such models in the future.”

Moreover, , we now add the x=y line to Fig. 2, so the readers can better assess the prediction variance as suggested by the reviewer:

We prefer to avoid using different scales (year/month) in the x and y axes to avoid misleading the readers, but the list of observed and predicted ages are available as SI files with a precision of 2 decimals point (~3 days).

We note that despite the moderate precision accuracy, we replicated these results in three separate cohorts.

Regarding the effect of “age bias” (also known as “regression attenuation” or “regression dilution” 10), we are aware of this phenomenon and agree that it must be accounted for. In fact, the “age bias” is one of the reasons we chose to use the difference between the expected and observed ages as the primary outcome of the study, as this measure already takes this bias into account. To demonstrate this effect we now compute brain age attenuation in two ways: 1. As described and used in the current study (Methods 4.9); and 2. By regressing out the effect of age on the predicted brain age at both times separately, then subtracting the adjusted predicted age at T18 from the adjusted predicted age at T0. The second method is the standard method to account for age bias as described in a previous work 11. Below is a scatter plot of both measures across all participants:

The x-axis represents the first method, used in the current study, and the y-axis represents the second method, described in Smith et al., (2019). Across all subjects, we found a nearly perfect 1:1 correspondence between the two methods (r=.998, p<0.001; MAE=0.45), as the two are mathematically identical. The small gap between the two is because the brain age attenuation model also takes into account the difference in the exact time that passed between the two scans for each participant (mean=21.36m, std = 1.68m).

We now note this in Methods section 4.9:

“We note that the result of computing the difference between the bias-corrected brain age gap at both times was nearly identical to the brain age attenuation measure (r=.99, p<0.001; MAE=0.45). The difference between the two is because the brain age attenuation model takes into account the difference in the exact time that passed between the two scans for each participant (mean=21.36m, std = 1.68m).”

3) In Figure 3, some of the changes observed between time points are very large. For example, one subject with a chronological age of 62 shows a ten-year increase in brain age over 18 months. This change is twice as large as the full range of age variation in the brain age estimates (average brain age increases from 50 to 55 across the full chronological age span). This makes it difficult to interpret RSFC change in units of brain age. E.g., is it reasonable that a person's brain ages by ten years, either up or down, in 18 months? The colour scale goes from -12 years to 14 years, so some of the observed changes are 14 / 1.5 = 9 times larger than the actual time from baseline to follow-up.

• The questions above should be investigated and addressed in the context of potential challenges with using brain age as a marker (see e.g., https://onlinelibrary.wiley.com/doi/full/10.1002/hbm.25837, https://onlinelibrary.wiley.com/doi/full/10.1002/hbm.26144).

We agree that our model precision was relatively low, especially compared to the period of the intervention, as also stated by reviewer #1. We now discuss this issue in light of the studies pointed out by the reviewer (Discussion section):

“In the current work, we used RSFC for brain age estimation, which resulted in a MAE of ~8 years, which was larger than the intervention period. Nevertheless, we found that brain age attenuation was associated with changes in multiple health factors. The precision of an age prediction model based on RSFC is typically lower than a model based on structural brain imaging5. However, a higher model precision may result in a lower sensitivity to detect clinical effects6,7. Better tools for data harmonization among datasets6 and larger training sample size5 may improve the accuracy of such models in the future.”

Again, we note that despite the moderate precision accuracy, we replicated these results in three separate cohorts and found that both the correlation and the MAE between the predicted and observed age were significant in all of them.

RSFC for age prediction:

1) Several studies show better age prediction accuracy with structural MRI features compared to RSFC. If the focus of the study is to use an accurate estimate of brain ageing rather than specifically looking at changes in RSFC, adding structural MRI data could be helpful.

We focused on brain structural changes in a previous work, and the focus of the current work was assessing age-related functional connectivity alterations. We now added a few sentences in the Introduction section that would hopefully better motivate our choice:

“We previously found that weight loss, glycemic control, lowering of blood pressure, and increment in polyphenols-rich food were associated with an attenuation in brain atrophy 12. Obesity is also manifested in age-related changes in the brain’s functional organization as assessed with resting-state functional connectivity (RSFC). These changes are dynamic13 and can be observed in short time scales14 and thus of relevance when studying lifestyle intervention.”

2) If changes in RSFC are the main focus, using brain age adds a complicated layer that is not necessarily helpful. It could be easier to simply assess RSFC change from baseline to follow up, and correlate potential changes with changes in e.g., BMI.

We are specifically interested in age-related changes as we described a-priori in the registration of the study: https://clinicaltrials.gov/ct2/show/NCT03020186

Moreover, age-related changes in RSFC are complex, multivariate and dependent upon the choice of theoretical network measures. We think that a data-driven brain age prediction approach might better capture these multifaceted changes and their relation to aging. We now state this in the Introduction section:

“Studies have linked obesity with decreased connectivity within the default mode network15,16 and increased connectivity with the lateral orbitofrontal cortex17, which are also seen in normal aging18,19. Longitudinal trials have reported changes in these connectivity patterns following weight reduction20,21, indicating that they can be altered. However, findings regarding functional changes are less consistent than those related to anatomical changes due to the multiple measures22 and scales23 used to quantify RSFC. Hence, focusing on a single measure, the functional brain age, may better capture these complex, multivariant changes and their relation to aging. “

The lack of control groups

1) If no control group data is available, it is important to clarify this in the manuscript, and evaluate which conclusions can and cannot be drawn based on the data and study design.

We agree that this point should be made more clear, and we now state this in the limitation section of the Discussion:

“We also note that the lack of a no-intervention control group limits our ability to directly relate our findings to the intervention. Hence, we can only relate brain age attenuation to the observed changes in health biomarkers.”

Also, following reviewers’ #2 and #3 comments, we refer to the weight loss following 18 months of lifestyle intervention instead of to the intervention itself. This is now made clear in the title, abstract, and the main text.

Reviewer #3 (Public Review):

The authors report on an interesting study that addresses the effects of a physical and dietary intervention on accelerated/decelerated brain ageing in obese individuals. More specifically, the authors examined potential associations between reductions in Body-Mass-Index (BMI) and a decrease in relative brain-predicted age after an 18-months period in N = 102 individuals. Brain age models were based on resting-state functional connectivity data. In addition to change in BMI, the authors also tested for associations between change in relative brain age and change in waist circumference, six liver markers, three glycemic markers, four lipid markers, and four MRI fat deposition measures. Moreover, change in self-reported consumption of food, stratified by categories such as 'processed food' and 'sweets and beverages', was tested for an association with change in relative brain age. Their analysis revealed no evidence for a general reduction in relative brain age in the tested sample. However, changes in BMI, as well as changes in several liver, glycemic, lipid, and fat-deposition markers showed significant covariation with changes in relative brain age. Three markers remained significant after additionally controlling for BMI, indicating an incremental contribution of these markers to change in relative brain age. Further associations were found for variables of subjective food consumption. The authors conclude that lifestyle interventions may have beneficial effects on brain aging.

Overall, the writing is concise and straightforward, and the langue and style are appropriate. A strength of the study is the longitudinal design that allows for addressing individual accelerations or decelerations in brain aging. Research on biological aging parameters has often been limited to cross-sectional analyses so inferences about intra-individual variation have frequently been drawn from inter-individual variation. The presented study allows, in fact, investigating within-person differences. Moreover, I very much appreciate that the authors seek to publish their code and materials online, although the respective GitHub project page did not appear to be set to 'public' at the time (error 404). Another strength of the study is that brain age models have been trained and validated in external samples. One further strength of this study is that it is based on a registered trial, which allows for the evaluation of the aims and motivation of the investigators and provides further insights into the primary and secondary outcomes measures (see the clinical trial identification code).

One weakness of the study is that no comparison between the active control group and the two experimental groups has been carried out, which would have enabled causal inferences on the potential effects of different types of interventions on changes in relative brain age. In this regard, it should also be noted that all groups underwent a lifestyle intervention. Hence, from an experimenter's perspective, it is problematic to conclude that lifestyle interventions may modulate brain age, given the lack of a control group without lifestyle intervention. This issue is fueled by the study title, which suggests a strong focus on the effects of lifestyle intervention. Technically, however, this study rather constitutes an investigation of the effects of successful weight loss/body fat reduction on brain age among participants who have taken part in a lifestyle intervention. In keeping with this, the provided information on the main effect of time on brain age is scarce, essentially limited to a sign test comparing the proportions of participants with an increase vs. decrease in relative brain age. Interestingly, this analysis did not suggest that the proportion of participants who benefit from the intervention (regarding brain age) significantly exceeds the number of participants who do not benefit. So strictly speaking, the data rather indicates that it's not the lifestyle intervention per sé that contributes to changes in brain age, but successful weight loss/body fat reduction. In sum, I feel that the authors' claims on the effects of the intervention cannot be underscored very well given the lack of a control group without lifestyle intervention.

We agree that this point, also raised by reviewer #2, should be made clear, and we now state this in the limitation section of the Discussion:

“We also note that the lack of a no-intervention control group limits our ability to directly relate our findings to the intervention. Hence, we can only relate brain age attenuation to the observed changes in health biomarkers.”

Also, following reviewers #2 and #3, we refer to the weight loss following 18 months of lifestyle intervention instead of to the intervention itself. This is now explicitly mentioned in the title, abstract, and within the text:

Title: “The effect of weight loss following 18 months of lifestyle intervention on brain age assessed with resting-state functional connectivity”

Abstract: “…, we tested the effect of weight loss following 18 months of lifestyle intervention on predicted brain age, based on MRI-assessed resting-state functional connectivity (RSFC).”

Another major weakness is that no rationale is provided for why the authors use functional connectivity data instead of structural scans for their age estimation models. This gets even more evident in view of the relatively low prediction accuracies achieved in both the validation and test sets. My notion of the literature is that the vast majority of studies in this field implicate brain age models that were trained on structural MRI data, and these models have achieved way higher prediction accuracies. Along with the missing rationale, I feel that the low model performances require some more elaboration in the discussion section. To be clear, low prediction accuracies may be seen as a study result and, as such, they should not be considered as a quality criterion of the study. Nevertheless, the choice of functional MRI data and the relevance of the achieved model performances for subsequent association analysis needs to be addressed more thoroughly.

We agree that age estimation from structural compared to functional imaging yields a higher prediction accuracy. In a previous publication using the same dataset12, we demonstrated that weight loss was associated with an attenuation in brain atrophy, as we describe in the introduction:

“We previously found that weight loss, glycemic control and lowering of blood pressure, as well as increment in polyphenols rich food, were associated with an attenuation in brain atrophy 12.”

Here we were specifically interested in age-related functional alterations that are associated with successful weight reduction. Compared to structural brain changes aging effect on functional connectivity is more complex and multifaced. Hence, we decided to utilize a data-driven or prediction-driven approach for assessing age-related changes in functional connectivity by predicting participants’ functional brain age. We now describe this rationale in the introduction section:

“Studies have linked obesity with decreased connectivity within the default mode network15,16 and increased connectivity with the lateral orbitofrontal cortex17, which are also seen in normal aging18,19. Longitudinal trials have reported changes in these connectivity patterns following weight reduction20,21, indicating that they can be altered. However, findings regarding functional changes are less consistent than those related to anatomical changes due to the multiple measures22 and scales23 used to quantify RSFC. Hence, focusing on a single measure, the functional brain age, may better capture these complex changes and their relation to aging.”

We address the point regarding the low model performance in response to reviewer #2, comment #2.

Author Response

Reviewer #1 (Public Review):

This study investigates how pathogens might shape animal societies by driving the evolution of different social movement rules. The authors find that higher disease costs induce shifts away from positive social movement (preference to move towards others) to negative social movement (avoidance from others). This then has repercussions on social structure and pathogen spread.

Overall, the study comprises a good mixture of intuitive and less intuitive results. One major weakness of the work, however, is that the model is constructed around one pathogen that repeatedly enters a population across hundreds of generations. While the authors provide some justification for this, it does not capture any biological realism in terms of the evolution of the pathogen itself, which would be expected. The lack of co-evolution in the model substantially limits the generality of the results. For example, a number of recent studies have reported that animals might be expected to become very social when pathogens are very infectious, because if the pathogen is unavoidable they may as well gain the benefits of being social. The authors make some arguments about being focused on introduction events, but this does not really align well with their study design that carries through many generations after the introduction. Given the rapid evolutionary dynamics, perhaps the study could have a more focused period immediately after the initial introduction of the pathogen to look at rapid evolutionary responses (albeit this may need some sensitivity analyses around the parameters such as the mutation rates).

We appreciate the reviewer’s evaluation of our work, and acknowledge that we have not currently included evolutionary dynamics for the pathogen.

One conceptual impediment to such inclusion is knowing how pathogen traits could be modelled in a mechanistic way. For example, it is widely held that there is a trade-off between infection cost and transmissibility, with a quadratic relationship between them, but this is a pattern and not a process per se. We are unsure which mechanisms could be modelled that impinge upon both infection cost and transmissibility.

On the practical side, we feel that a mechanistic, individual-based model that includes both pathogen and host evolution would become very challenging to interpret. It might be more tractable to begin with a mechanistic, spatial model that examines pathogen trait evolution with an unchanging host (such as an adaptation of Lion and Boots, 2010). We would be happy to take this on in future work, with a view to combining models thereafter.

We have taken the suggestion to focus on the period immediately after the introduction, and we now focus on the following 500 generations. While 500 generations is still a long time, we would note that our model dynamics typically stabilise within 200 generations. We show the following generations primarily to check that some stability in the dynamics has indeed been reached (but see our new scenario 2).

We also appreciate the point regarding mutation rates. Our mutation rates are relatively high to account for the small size of our population. We have found that with smaller mutation rates (0.001 rather than 0.01), evolutionary shifts in our population do not occur within the first 500 generations. This is primarily because prior to pathogen introduction, the ‘agent avoiding’ strategy that becomes common later is actually quite rare. Whether a rapid transition takes place thus depends on whether there are any agent avoiding individuals in the population at the moment of pathogen introduction, or on whether such individuals emerge rapidly thereafter through mutations on the social weights. We expect that with larger population sizes, we would be able to recover our results with smaller mutation rates as well.

A final, and much more minor comment is whether this is really a paper about movement. The model does not really look at evolutionary changes in how animals move, but rather at where they move. How important is the actual movement process under this model? For example, would the results change if the model was constructed without explicit consideration of space and resources, but instead simply modelled individuals' decisions to form and break ties? (Similar to the recent paper by Ashby & Farine https://onlinelibrary.wiley.com/doi/full/10.1111/evo.14491 ). It might help to provide more information about how putting social decisions into a spatially explicit framework is expected to extend studies that have not done so (e.g.., because they are analytical).

This paper is indeed about movement, as where to move is a key part of the movement ecology paradigm (Nathan et al. 2008). That said, we appreciate the advice to emphasise the importance of social decisions in a spatial context, we have added these to the Introduction (L. 79 – 81) and Discussion (L. 559 – 562). In brief, we do expect different dynamics that result from the explicit spatial context, as compared to a model in which social associations are probabilistic and could occur with any individual in the population.

In our models, individual social tendency (whether they are prefer moving towards others) is separated from individual sociality (whether they actually associate with other individuals). This can be seen from our (new) Fig. 3D, in which individuals of each of the social strategies can sometimes have similar numbers of associations (although modulated by movement). This separation of the pattern from the underlying process is possible, we believe, due to the heterogeneity in the social landscape created by the explicit spatial context.

Reviewer #2 (Public Review):

This theoretical study looks at individuals' strategies to acquire information before and after the introduction of pathogens into the system. The manuscript is well-written and gives a good summary of the previous literature. I enjoyed reading it and the authors present several interesting findings about the development of social movement strategies. The authors successfully present a model to look at the costs and benefits of sociality.

I have a couple of major comments about the work in its current form that I think are very important for the authors to address. That said, I think this is a promising start and that with some revisions, this could be a valuable contribution to the literature on behavioral ecology.

We appreciate the reviewer’s kind words.

Before starting, I would like to be precise that, given the scope of the models and the number of parameter choices that were necessary, I am going to avoid criticisms of the decisions made when designing the models. However, there are a few assumptions I rather find problematic and would like to give proper attention to.

The first regards social vs. personal information. Most of the model argumentation is based on the reliance on social information (considering four, but to me overlapping, social strategies that are somehow static and heritable) but in fact, individuals may oscillate between relying on their personal information and/or on social information -- which may depend on the availability of resources, population density, stochastic factors, among others (Dall et al. 2005 Trends Ecol. Evol., Duboscq et al. 2016 Frontiers in Psychology). In my opinion, ignoring the influence of personal and social information decreases the significance of this work. I am aware that the authors consider the detection of food present in the model, but this is considered to a much smaller extent (as seen in their weight on individual decisions) than the social information cues.

We appreciate the point that individuals can switch between relying on social and personal information. However, we would point out that in our model, the social strategies are not static. The social strategy is a convenient way of representing individuals’ position in behavioural trait-space (the ‘behavioural hypervolume’ of Bastille-Rousseau and Wittemeyer 2019). This essentially means that the importance assigned to each of the three cues available in our model varies among individuals. There are indeed individuals that are primarily guided by the density of food items, and this is the commonest ‘overall’ movement strategy before the pathogen is introduced. We represent this by showing how the importance of social information is low before pathogen introduction (Fig. 2B).

While we primarily focus on the importance of social information, this is because the population quite understandably evolves a persistent preference for moving towards food items (i.e., using personal information if available). We have made this clearer in the text on lines 367 – 371.

Critically, it is also unclear how, if at all, the information and pathogen traits are related to each other. If a handler gets sick, how does this affect its foraging activity (does it stop foraging, slow its activities, or does it show signs of sickness)? Perhaps this model is attempting to explore the emergence of social movement strategies only, but how they disentangle an individual's sickness status and behavioral response is unclear.

We appreciate that infection may lead to physiological effects (e.g. altered metabolic rates, reduction in cognitive capacity) that may then influence behaviour. Our model aims to be relatively simple and general one, and does not consider the explicit mechanisms by which infection imposes a cost on fitness. Thus we do not include any behavioural modifications due to infection, as we feel that these would be much too complex to include in such a model. We would be happy to explore, in future work, phenomena such as the evolution of self-isolation and infection detection which is common among animals such as social insects (Stroeymeyt et al. 2018, Pusceddu et al. 2021).

However, we have considered an alternative implementation of our model’s scenario 1 which could be interpreted as the infection reducing foraging efficiency by a certain percentage (other interpretations of the redirection of energy away from reproduction are also possible). We show how this implementation leads to very similar outcomes as those seen in our

Very little is presented about the virulence of the pathogens and how they could affect the emergence of social strategies. The authors keep their main argumentation based on the introduction of novel pathogens (without distinctions on their pathogenicity), but a behavioral response is rather influenced by how fast individuals are infected and which are their chances of recovering. Besides, they consider that only one or two social interactions would be enough for pathogen transmission to occur.

We have indeed considered a fixed transmission probability of 0.05, a relatively modest attack rate. Setting transmission probability to two other values (0.025, 0.1), we find that our general results are recovered - there is an evolutionary transition away from sociality, with the proportion of agent avoidance evolved increasing with the transmission probability. While we do not show these results in the main text, we have included figures showing the proportions of each social movement strategy here for the reviewers’ reference.

Figures showing the proportion of social movement strategies in two simulation runs of our default implementation of scenario 1 (dE = 0.25, R = 2, pathogen introduction begins from G = 500). Top: Probability of transmission = 0.025 (half of the default). Bottom: Probability of transmission = 0.10 (double the default). Overall, the proportion of agent avoidance evolved (purple) increases with the probability of transmission. Each figure shows a single replicate of each parameter combination, for only 1,000 generations.

Another important component is that individuals do not die, and it seems that they always have a chance (even if it is small) to reproduce. So, how the authors consider unsuccessful strategies in the model outputs or how these social strategies would be potentially "dismissed" by natural selection are not considered.

We appreciate the point that our simulation does not include mortality effects, and that all individuals have some small chance of reproducing. There are a few practical and conceptual challenges when incorporating this level of realism in a general model. Including mortality effects could allow for the emergence of more complex density-dependent dynamics, as dead individuals would not be able to transmit the pathogen to other foragers (although for some pathogens, this could be a valid choice), nor would they be sources of social information. This would make the model much more challenging to interpret, and we have tried to keep this model as simple as possible.

We have also sought to keep the model’s focus on the evolutionary dynamics, and to not focus on mortality. In order to balance this aim with the reviewer's suggestion, we have included a new implementation of the model’s scenario 1 which has a threshold on reproduction. That means that only individuals with a positive energy balance (intake > infection costs) are allowed to reproduce. We show a potentially counter-intuitive result, that the more social ‘handler tracking’ strategy persists at a higher frequency than in our default implementation, despite having a higher infection rate than the ‘agent avoiding’ strategy. We suggest that this is because the ‘agent avoiding’ individuals have very low or no intake. This is sufficient in our default implementation to have relatively higher fitness than the more frequently infected handler tracking individuals.

Reviewer #3 (Public Review):

Gupte and colleagues develop an individual-based model to examine how the introduction of a novel pathogen influences the evolution of social cue use in a population of agents for which social cues can both facilitate more efficient foraging, but also expose individuals to infection. In their simulations, individuals move across a landscape in search of food, and their movements are guided by a combination of cues related to food patches, individuals that are currently handling food items, and individuals that are not actively handling food. The latter two cues can provide indirect information about the likely presence of food due to the patchiness of food across the landscape.

The authors find that prior to introducing the novel pathogen, selection favors strategies that home in on agents, regardless of whether those agents are currently handling food items. The overall contribution of these social cues to movement decisions, however, tends to be relatively small. After pathogen introduction, agents evolve to rely more heavily on social information and to either be more selective in their use of it (attending to other agents that are currently handling food and avoiding non-handlers) or avoiding other agents altogether. Gupte and colleagues further examine the ecological consequences of these shifts in social decision-making in terms of individuals' overall movement, food consumption, and infection risk. Relative to pre-introduction conditions, individuals move more, consume less food, and are less likely to be infected due to reduced contact with others. Epidemiological models on emergent social networks confirm that evolved behavioral changes generate networks that impede the spread of disease.

The introduction of novel pathogens into wild populations is expected to be increasingly common due to climate change and increasing global connectedness. The approach taken here by the authors is a potentially worthwhile avenue to explore the potential eco-evolutionary consequences of such introductions. A major strength of this study is how it couples ecological and evolutionary timescales. Dominant behavioral strategies evolve over time in response to changing environmental conditions and impact social, foraging, and epidemiological dynamics within generations. I imagine there are many further questions that could be fruitfully explored using the authors' framework. There are, however, important caveats that impact the interpretation of the authors' findings.

First, reproduction bears no cost in this model. Individuals produce offspring in proportion to their lifetime net energy intake, which is increased by consuming food and decreased by a set amount per turn once infected. However, prior to reproduction, net energy intake is normalized (0-1) according to the lowest individual value within the generation. This means that individuals need not maintain a positive energy balance nor even consume food at all to successfully reproduce, so long as they perform reasonably well relative to other members of the population. Since consuming food is not necessary to reproduce, declining per capita intake due to evolved social avoidance (Fig. 1d) likely decreases the importance of food to an individual's reproductive success relative to simply avoiding infection. This dynamic could explain the delayed emergence of the 'agent avoiding' strategy (Fig. 1a), as this strategy potentially is only viable once per capita intake reaches a sufficiently low level across the population (Fig. 1d). I am curious to know what the results would be if reproduction required some minimal positive net energy, such that individuals must risk food patches in order to reproduce. It would also be useful for the authors to provide information on how net energy intake changes across generations, as well as whether (and if so, how) attraction to the food itself may change over time.

We thank the reviewer for their assessment of our work, and appreciate the point raised here (and in an earlier review) about individuals potentially reproducing without any intake. We have addressed this by running our default model [repeated introductions, R = 2, dE = 0.25], with a threshold on reproduction such that only individuals with a positive energy balance can reproduce. We mention these results in the text (L. 495 – 500), and show related figures in the SI Appendix. In brief, as the reviewer suggests, agent avoiding is less common for our default parameter combination, but becomes as common as the default combination when the infection cost is doubled (to dE = 0.5).

We appreciate the reviewer’s suggestion about decreasing per-capita intake being a precondition for the proliferation of the agent avoiding strategy. With our new results, we now show that there is no overall decrease in intake, but the agent avoiding strategy still becomes a common strategy after pathogen introduction. As the reviewer suggests, this is because these individuals have an equivalent net energy as handler tracking individuals, as they are less frequently infected.

We suggest that the delayed emergence of the agent avoiding strategy is primarily due to mutation limitations – such individuals are uncommon or non-existent in the simulation before pathogen introduction, and random mutations are required for them to emerge. As we have noted in response to an earlier comment, this becomes clear when the mutation rate is reduced from 0.01 to 0.001 – agent avoidance usually does not evolve at all.

A second important caveat is that the evolutionary responses observed in the model only appear when novel pathogen introductions are extremely frequent. The model assumes no pathogen co-evolution, but rather that the same (or a functionally identical) pathogen is re-introduced every generation (spillover rate = 1.0). When the authors considered whether evolutionary responses were robust to less frequent introductions, however, they found that even with a per-generation spillover rate of 0.5, there was no impact on social movement strategies. The authors do discuss this caveat, but it is worth highlighting here as it bears on how general the study's conclusions may be.

We appreciate the reviewer’s point entirely. We would point out that current knowledge about pathogen introductions across species and populations in the wild is very poor. However, the ongoing highly pathogenic avian influenza outbreak (Wille and Barr 2022), the spread of multiple strains of SARS-CoV-2 to wild deer in several different human-to-wildlife transmission events, and recent work on the potential for coronavirus spillovers from bats to humans, all suggest that at least some generalist pathogens must circulate quite widely among wildlife, often crossing into novel host species or populations. We have added these considerations to the text on lines 218 – 231.

We have also added, in order to confront this point more squarely, a new scenario of our model in which the pathogen is introduced just once, and then transmits vertically and horizontally among individuals (lines 519 – 557). This scenario more clearly suggests when evolutionary responses to pathogen introductions are likely to occur, and what their consequences might be for a pathogen becoming endemic in a population. This scenario also serves as a potential starting point for models of host-pathogen trait co-evolution, and we have added this consideration to the text on lines 613 – 623.

References

● Albery, G. F. et al. 2021. Multiple spatial behaviours govern social network positions in a wild ungulate. - Ecology Letters 24: 676–686.

● Bastille-Rousseau, G. and Wittemyer, G. 2019. Leveraging multidimensional heterogeneity in resource selection to define movement tactics of animals. - Ecology Letters 22: 1417–1427.

● Gupte, P. R. et al. 2021. The joint evolution of animal movement and competition strategies. - bioRxiv in press.

● Lion, S. and Boots, M. 2010. Are parasites ‘“prudent”’ in space? - Ecology Letters 13: 1245–1255.

● Lloyd-Smith, J. O. et al. 2005. Superspreading and the effect of individual variation on disease emergence. - Nature 438: 355–359.

● Nathan, R. et al. 2008. A movement ecology paradigm for unifying organismal movement research. - PNAS 105: 19052–19059.

● Pusceddu, M. et al. 2021. Honey bees increase social distancing when facing the ectoparasite varroa destructor. - Science Advances 7: eabj1398.

● Sánchez, C. A. et al. 2022. A strategy to assess spillover risk of bat SARS-related coronaviruses in Southeast Asia. - Nat Commun 13: 4380.

● Stroeymeyt, N. et al. 2018. Social network plasticity decreases disease transmission in a eusocial insect. - Science 362: 941–945.

● Wilber, M. Q. et al. 2022. A model for leveraging animal movement to understand spatio-temporal disease dynamics. - Ecology Letters in press.

● Wille, M. and Barr, I. G. 2022. Resurgence of avian influenza virus. - Science 376: 459–460.

Author Response

Reviewer #1 (Public Review):

This work sheds light on the adverse effects of Bacillus thuringiensis, a strong pathogenic bacteria used as a microbial pesticide to kill lepidopteran larvae that threaten crops, on gut homeostasis of non-susceptible organisms. By using the Drosophila melanogaster as a non-susceptible organism model, this paper reveals the mechanisms by which the bacteria disrupt gut homeostasis. Authors combined the use of different genetic tools and Western blot experiments to successfully demonstrate that bacterial protoxins are released and activated throughout the fly gut after ingestion and influence intestinal stem cell proliferation and intestinal cell differentiation. This phenomenon relies on the interaction of activated protoxins with specific components of adherens junctions within the intestinal epithelium. Due to conserved mechanisms governing intestinal cell differentiation, this work could be the starting point for further studies in mammals.

The conclusions proposed by the authors are in general well supported by the data. However, some improvements in data representation, as well as additional key control experiments, would be needed to further reinforce some key points of the paper.

We thank reviewer1 for her appreciation of the work and in depth analysis of the data. We agree with all her comments and believe the suggestions significantly improved the manuscript.

1) Figure 1 and others: Several graphs in the manuscript show the number of cells/20000µm2. How is the shape of the gut in the different conditions studied in this manuscript? The gut shape (shrunk gut versus normal gut for example) could influence the number of cells seen in a small area. For example, the number of total cells quantified in a small area (here 20000µm2) of a shrunk gut can be increased while their size decrease. As a result, the quantification of a specific cell type in a small region (here 20000µm2) can be biased and not represent the real number of cells present in the whole posterior part of the R4 region. Would it make sense to calculate a ratio "number of X cells/number of DAPI positive cells per 20000µm2"?

We provided a suitable answer in the "Essential Revisions point 1" corresponding to this reviewer's concern. To summarize, we have now added whole posterior midgut images in the different conditions to highlight the intestinal morphology (Figure 1-figure supplement 1A). The whole gut morphology was not affected by the different challenges we performed. Indeed, we used low doses of spores and/or toxins in order to mimic "natural" amounts of spores/toxins the fly can eat in the environment and in order to avoid drastic gut lining disturbances.

We have also added the cell type ratio in figure 1- figure supplement 2.

2) Figure 4: Is it possible that Arm staining is less intense between ISC and progenitors after ingestion of the bacteria due to the fact there is a high rate of stem cell proliferation? Could it be an indirect effect of stem cell proliferation rather than the binding of the toxins to Cadherins?

We thank the reviewer for this pertinent comment. Indeed, for this reason, we compared the intensity of Arm expression at the junction between neighboring progenitors with the Arm intensity around the rest of the cellular membranes and calculated the ratio between both values (see Figure 4-figure supplement 1F-G for an illustration of how we proceeded and the new section in the Material and Methods 736-742). Using this method, even if the whole Arm staining intensity is different (in all the midgut), the ratio reflects the internal cell-cell interaction changes between the two neighboring cells. Moreover, we have observed that Arm staining (using the usual monoclonal antibody N2 7A1 from the DSHB) was very variable from one midgut to another in the same feeding/intoxication condition. So, we do not want to draw conclusion about the whole Arm intensity due to this variability whatever are the intoxication conditions. Finally, the challenged guts always displayed a more disorganized epithelium due to cell proliferation and differentiation. Consequently, Arm staining in ECs and progenitor cells are found in the same focal plane while in unchallenged and well-organized guts, Arm staining in ECs is above the focal plane of Arm staining in progenitor cells. This likely leads to the impression that Arm staining is more intense in challenged midguts. This method description is now added in the Material and methods section (lines 736-742).

Could the authors use the ReDDM system to distinguish between "old" and newly formed cells? This could be a good control to make sure that the signal is quantified in similar cells between the control and the different conditions.

We have analyzed intensity of Arm expression between pairs of GFP cells. Most of these pairs arose from de novo divisions. Indeed, as shown in control conditions (water) with Dl-ReDDM (for example see figure 1-figure supplement 1D), pairs of GFP cells (ISC-ISC) are rare. Most pairs correspond to ISC-EB or ISC-EEP pairs with the progenitor marked by the RFP, meaning that it just arises from the GFP+ mother ISC. Therefore we assume, that in the esg>GFP genotype, pairs of GFP+ cells correspond to one ISC and one progenitor (see Figure 4 – figure supplement 1A-A'). Therefore, when we analyzed the Arm intensity between pairs of GFP cells after intoxication, these cells are very likely "newborn" cells. Even if we suppose there are ISCs and progenitors that remain stuck together for a long time (for instance several days), Cry1A toxins can also be able to disrupt their cell junction. In the context of Cry1A toxin activity, it seems important to analyze the whole impact on cell-cell junctions without discriminating old and new cell-cell interactions.

We tried to use anti-Arm and anti-Pros double staining to mark new EEPs. Unfortunately, anti-Arm and anti-Prospero antibodies were both raised in mice. Co-staining with both antibodies give rise to bad labelling either for Arm or for Prospero or for both. Our first author spent lot of energy trying to set up good conditions but unfortunately this was unsuccessful.

Here is an example of what we got (this was the best image we got) with esg>GFP flies fed with water (control) and labelled for Arm and Pros in red. White arrows point two EEPs. Red arrows points the Arm staining between two precursors (ISC/ISC or ISC/EB or EB/EB). It was extremely hard to identify junctions marked by Arm between EEPs and ISCs because the Pros staining was too strong.

Another example with flies fed with spores of SA11 (increasing the number of EEs). In green is the esg>GFP and in Red Arm and Prospero. The right panel correspond to the single red channel (Arm/Prospero).

Nevertheless, we have now performed a similar analysis in an esg>GFP, Shg::RFP background and analyzed Shg::RFP (Tomato::DE-Cadherin) labelling intensity. We found similar results that are presented in the new Figure 4 (data we Arm have been moved in Figure 4-figure supplement 1). This last analysis have been included in the text lines 285-299.

Figure 4E' and 4G': Arm staining seems more intense when looking at the whole membrane levels of cells compared to control. Is it possible that the measured ratio contact intensity/membrane intensity presented in Figure 4I could be impacted and not reflect the real contact intensity between ISC and progenitor cells?

Please check our answer just above: "…//… we have observed that Arm staining (using the usual monoclonal antibody N2 7A1 from the DSHB) was very variable from one midgut to another in the same feeding/intoxication condition. So, we do not want to draw conclusion about the whole Arm intensity due to this variability whatever are the intoxication conditions".

See also our intensity measurement method described above to avoid bias: "…//… we compared the intensity of Arm expression at the junction between neighboring progenitors with the Arm intensity around the rest of the cellular membranes and calculated the ratio between both values (see Figure 4-figure supplement 1F-G for an illustration of how we proceeded and the new section in the Material and Methods 736-742). Using this method, even if the whole Arm staining intensity is different (in all the midgut), the ratio reflects the internal cell-cell interaction changes between the two neighboring cells."

What is the hypothesis of the authors about the decrease of Arm or DE-Cad seen after bacterial/crystal ingestion? Does the interaction between the toxins and DE-Cad induce a relocation of DE-Cad?

It has been shown that E-Cadherin could be recycled when adherens junctions are destabilized both in Drosophila and mammals(Buchon et al., 2010; O'Keefe et al., 2007; Tiwari et al., 2018). To investigate this possibility, we tried to analyze DE-Cad cytoplasmic relocalization using anti-DE-Cad immunostaining (DCAD2 antibody from DSHB) as well as Shg::RFP (Bloomington stock #58789) or Shg::GFP (Bloomington stock #60584) endogenous fusion. Unfortunately, we did not see obvious differences. Nevertheless, we have now added the split channels of the Shg::RFP labelling in the different conditions in Figure 4A-D'. Nevertheless, we are still interested in the behavior of the DE-cadherin (and signaling, see (Liang et al., 2017)) upon binding of the Cry1A toxin. N. Zucchini-Pascal (author in this article) are currently investigating this question.

The authors should add more details about the way to quantify in the Material and methods section. How many cells have been quantified per intestine? How did they choose the cells where they quantified the contact intensity?..etc

These details were missing in the methods and we thank the reviewer for highlighting this issue. We added these information to the methods (lines 725-742). The number of cell pairs analyzed was present in the raw data related to figure 4 but absent from the main figure and legend. It is now rectified. We only measured the intensity in isolated pairs of cells.

Figure 4B, D, F and H: How did the authors recognize the ISCs?

We agree with the reviewer comment. We cannot recognize ICS per se. Green cells correspond either to ISCs or to EBs. We modified the text accordingly (lines 285-287).

Could the authors do quantifications of DE-Cad signal?

This has been done. It is shown now in figure 4E and in Table 1. We also adapted the text (lines 289-299) to fine-tune our interpretation in light of this new analysis. Indeed, what we have now defined as "mild" adherens junction intensity is between the ratio 1.4 and 1.6 instead of the previous ratio (1.3 to 1.6), because we observed most of the EEP progenitors arising from cell displaying a junction intensity with their mother cells below the 1.4 ratio (see Table 1).

Like Arm staining, the staining seems stronger at the whole membrane level in F and H compared to the control.

As we described above for Arm staining, the intensity of Tomato::DE-Cad labelling can differ from one posterior midgut to another one. One simple explanation would be related to changes in the structure of midgut epithelium which is well organized in unchallenged conditions, while in challenged midguts the epithelial cells are not well-arranged anymore due to rapid cell proliferation and differentiation. Consequently, DE-Cad labelling in ECs is at the same level as that in ISC/progenitors cells, giving the impression that the labelling is stronger.

3) Figure 5: How is the stem cell proliferation upon overexpression of DE-Cad in control or upon bacteria/crystals ingestion? Do the authors think that the decrease of Pros+RFP+ new cells upon overexpression of DE-Cad could result from a decrease of stem cell proliferation?

Great suggestion. Thereby, we chose to count the progenitor cells (GFP+ cells) reflecting the ISC division during the last 3 days. Moreover, this also has the advantage of working on the same pictures (samples) used for all the analyzes shown in figure 5 and Figure 5-figure supplement 1. Hence, If we consider the number of GFP+ cells (esg expressing cells corresponding to ISC, EB or EEP) in challenged midguts, the overexpression of the DE-Cad did not seem to alter ISC division. In addition, we still observed more GFP+ cells when the midguts were challenged with SA11 or crystals than with BtkCry, in agreement with the rate of ISC division observed in the WT genetic background shown in figure 1B.

We have now added the counting of GFP+ cells in Figure 5-figure supplement 1E. The text has been modified to integrate this results (lines 306-308).

Did the authors quantify the % of new ECs in the context of overexpression of DE-Cad?

The data has been added in figure 5F. The text has been modified to integrate this result lines 312-313.

Figure 5F: As asked before, did the authors distinguish the signal between newly born cells and the signal between older cells?

In the new figure 5G: we used the esg-ReDDM system that is very efficient. Almost all ISC and progenitors express the GFP. The counting have been done between cell pairs that express both the GFP and RFP. It is specified in the text lines 310-311. Nevertheless, we cannot distinguish between new and old cells here. Indeed, the esg-ReDDM system induce both the GFP and the RFP in all esg+ cells (the old ones and the new ones). Hence, if a division has occurred just before the induction of the system to give birth for instance to an ISC and an EB, both cells will express the GFP and the RFP. But should we consider those pairs of cells as old cells or new cells? Noteworthy, as we analyzed the intensity of junctions 3 days after intoxication and induction of the ReDDM system, we assume that the pairs of GFP+/RFP+ cells arose after the induction of the system. Indeed, to our knowledge, nobody has shown in the posterior midgut, that a progenitor remains stuck to its mother ISC as long as 3 days. Even if we assume that this event can occur, Cry1A toxins can also be able to disrupt their cell junction.

We now have removed the DAPI channel and added the RFP+ channel in Figure 5-figure supplement 1A-D' (previously the Figure S4A-D) to illustrate this explanation and to facilitate the interpretation by the reader.

It would be interesting to compare the junction intensity between mother ISCs and their daughter progenitors before and after intoxication in a same intestine. But we think that this event is quite rare because of the experimental conditions we used (i.e. analyses 3 days after the induction of the ReDDM/intoxication).

The same experiments (stem cell proliferation + quantification of the % of new ECs) could be also done when authors overexpress of the Connectin, supplemental figure 5. This would be another control to conclude that the effects on cell differentiation are specific due to the interaction between DE-Cad and the toxins.

We have added the analyses in Figure 5 - figure supplement 2J and K.

The text has been completed lines 317-320.

In the "crystals" condition, the overexpression of Connection seems to partially rescue the increase % of new Pros+RFP+ new cells observed in Figure 3F (Figure S5I compared to Figure 3F).

Yes, we agree with the reviewer comment. In an esg-ReDDM background (figure 3F), crystals induced a much greater increase in EE numbers than did SA11 spores. However, in a WT or esg>GFP background, crystals induced a similar increase in EE/EEP to that induced by SA11 spores. So we do not yet have explanation excepted the genetic background of the esg-ReDDM.

Author Response

Reviewer #1 (Public Review):

The authors start the study with an interesting clinical observation, found in a small subset of prostate cancers: FOXP2-CPED1 fusion. They describe how this fusion results in enhanced FOXP2 protein levels, and further describe how FOXP2 increases anchorageindependent growth in vitro, and results in pre-malignant lesions in vivo. Intrinsically, this is an interesting observation. However, the mechanistic insights are relatively limited as it stands, and the main issues are described below.

Main issues:

1) While the study starts off with the FOXP2 fusion, the vast majority of the paper is actually about enhanced FOXP2 expression in tumorigenesis. Wouldn't it be more logical to remove the FOXP2 fusion data? These data seem quite interesting and novel but they are underdeveloped within the current manuscript design, which is a shame for such an exciting novel finding. Along the same lines, for a study that centres on the prostate lineage, it's not clear why the oncogenic potential of FOXP2 in mouse 3T3 fibroblasts was tested.

We thank the reviewer very much for the comment. We followed the suggestion and added a set of data regarding the newly identified FOXP2 fusion in Figure 1 to make our manuscript more informative. We tested the oncogenic potential of FOXP2 in NIH3T3 fibroblasts because NIH3T3 cells are a widely used model to demonstrate the presence of transformed oncogenes2,3. In our study, we observed that when NIH3T3 cells acquired the exogenous FOXP2 gene, the cells lost the characteristic contact inhibition response, continued to proliferate and eventually formed clonal colonies. Please refer to "Answer to Essential Revisions #1 from the Editors” for details.

2) While the FOXP2 data are compelling and convincing, it is not clear yet whether this effect is specific, or if FOXP2 is e.g. universally relevant for cell viability. Targeting FOXP2 by siRNA/shRNA in a non-transformed cell line would address this issue.

We appreciate these helpful comments. Please refer to the "Answer to Essential Revisions #1 from the Editors” for details.

3) Unfortunately, not a single chemical inhibitor is truly 100% specific. Therefore, the Foretinib and MK2206 experiments should be confirmed using shRNAs/KOs targeting MEK and AKT. With the inclusion of such data, the authors would make a very compelling argument that indeed MEK/AKT signalling is driving the phenotype.

We thank the reviewer for highlighting this point and we agree with the reviewer’s point that no chemical inhibitor is 100% specific. In this study, we used chemical inhibitors to provide further supportive data indicating that FOXP2 confers oncogenic effects by activating MET signaling. We characterized a FOXP2-binding fragment located in MET and HGF in LNCaP prostate cancer cells by utilizing the CUT&Tag method. We also found that MET restoration partially reversed oncogenic phenotypes in FOXP2-KD prostate cancer cells. All these data consistently supported that FOXP2 activates MET signaling in prostate cancer. Please refer to the "Answer to Essential Revisions #2 from the Editors” and to the "Answer to Essential Revisions #7 from the Editors” for details.

4) With the FOXP2-CPED1 fusion being more stable as compared to wild-type transcripts, wouldn't one expect the fusion to have a more severe phenotype? This is a very exciting aspect of the start of the study, but it is not explored further in the manuscript. The authors would ideally elaborate on why the effects of the FOXP2-CPED1 fusion seem comparable to the FOXP2 wildtype, in their studies.

We thank the reviewer very much for the comment. We had quantified the number of colonies of FOXP2- and FOXP2-CPED1-overexpressing cells, and we found that both wildtype FOXP2 and FOXP2-CPED1 had a comparable putative functional influence on the transformation of human prostate epithelial cells RWPE-1 and mouse primary fibroblasts NIH3T3 (P = 0.69, by Fisher’s exact test for RWPE-1; P = 0.23, by Fisher’s exact test for NIH3T3). We added the corresponding description to the Results section in Line 487 on Page 22 in the tracked changes version of the revised manuscript. Please refer to the "Answer to Essential Revisions #5 from the Editors” for details.

5) The authors claim that FOXP2 functions as an oncogene, but the most-severe phenotype that is observed in vivo, is PIN lesions, not tumors. While this is an exciting observation, it is not the full story of an oncogene. Can the authors justifiably claim that FOXP2 is an oncogene, based on these results?

We appreciate the comment, and we made the corresponding revision in the revised manuscript. Please refer to the "Answer to Essential Revisions #3 from the Editors” for details.

6) The clinical and phenotypic observations are exciting and relevant. The mechanistic insights of the study are quite limited in the current stage. How does FOXP2 give its phenotype, and result in increased MET phosphorylation? The association is there, but it is unclear how this happens.

We appreciate this valuable suggestion. In the current study, we used the CUT&Tag method to explore how FOXP2 activated MET signaling in LNCaP prostate cancer cells, and we identified potential FOXP2-binding fragments in MET and HGF. Therefore, we proposed that FOXP2 activates MET signaling in prostate cancer through its binding to MET and METassociated gene. Please refer to the "Answer to Essential Revisions #2 from the Editors” for details.

Reviewer #2 (Public Review):

1) The manuscript entitled "FOXP2 confers oncogenic effects in prostate cancer through activating MET signalling" by Zhu et al describes the identification of a novel FOXP2CPED1 gene fusion in 2 out of 100 primary prostate cancers. A byproduct of this gene fusion is the increased expression of FOXP2, which has been shown to be increased in prostate cancer relative to benign tissue. These data nominated FOXP2 as a potential oncogene. Accordingly, overexpression of FOXP2 in nontransformed mouse fibroblast NIH-3T3 and human prostate RWPE-1 cells induced transforming capabilities in both cell models. Mechanistically, convincing data were provided that indicate that FOXP2 promotes the expression and/or activity of the receptor tyrosine kinase MET, which has previously been shown to have oncogenic functions in prostate cancer. Notably, the authors create a new genetically engineered mouse model in which FOXP2 is overexpressed in the prostatic luminal epithelial cells. Overexpression of FOXP2 was sufficient to promote the development of prostatic intraepithelial neoplasia (PIN) a suspected precursor to prostate adenocarcinoma and activate MET signaling.

Strengths:

This study makes a convincing case for FOXP2 as 1) a promoter of prostate cancer initiation and 2) an upstream regulator of pro-cancer MET signaling. This was done using both overexpression and knockdown models in cell lines and corroborated in new genetically engineered mouse models (GEMMs) of FOXP2 or FOXP2-CPED1 overexpression in prostate luminal epithelial cells as well as publicly available clinical cohort data.

Major strengths of the study are the demonstration that FOXP2 or FOXP2-CPED1 overexpression transforms RWPE-1 cells to now grow in soft agar (hallmark of malignant transformation) and the creation of new genetically engineered mouse models (GEMMs) of FOXP2 or FOXP2-CPED1 overexpression in prostate luminal epithelial cells. In both mouse models, FOXP2 overexpression increased the incidence of PIN lesions, which are thought to be a precursor to prostate cancer. While FOXP2 alone was not sufficient to cause prostate cancer in mice, it is acknowledged that single gene alterations causing prostate cancer in mice are rare. Future studies will undoubtedly want to cross these GEMMs with established, relatively benign models of prostate cancer such as Hi-Myc or Pb-Pten mice to see if FOXP2 accelerates cancer progression (beyond the scope of this study).

We appreciate these positive comments from the reviewer. We agree with the suggestion from the reviewer that it is worth exploring whether FOXP2 is able to cooperate with a known disease driver to accelerate the progression of prostate cancer. Therefore, we are going to cross Pb-FOXP2 transgenic mice with Pb-Pten KO mice to assess if FOXP2 is able to accelerate malignant progression.

2) Weaknesses: It is unclear why the authors decided to use mouse fibroblast NIH3T3 cells for their transformation studies. In this regard, it appears likely that FOXP2 could function as an oncogene across diverse cell types. Given the focus on prostate cancer, it would have been preferable to corroborate the RWPE-1 data with another prostate cell model and test FOXP2's transforming ability in RWPE-1 xenograft models. To that end, there is no direct evidence that FOXP2 can cause cancer in vivo. The GEMM data, while compelling, only shows that FOXP2 can promote PIN in mice and the lone xenograft model chosen was for fibroblast NIH-3T3 cells.

To determine the oncogenic activity of FOXP2 and the FOXP2-CPDE1 fusion, we initially used mouse primary fibroblast NIH3T3 for transformation experiments, because NIH3T3 cells are a widely used cell model to discover novel oncogenes2,3,10,11. Subsequently, we observed that overexpression of FOXP2 and its fusion variant drove RWPE-1 cells to lose the characteristic contact inhibition response, led to their anchorage-independent growth in vitro, and promoted PIN in the transgenic mice. During preparation of the revised manuscript, we tested the transformation ability of FOXP2 and FOXP2-CPED1 in RWPE1 xenograft models. We subcutaneously injected 2 × 106 RWPE-1 cells into the flanks of NOD-SCID mice. The NODSCID mice were divided into five groups (n = 5 mice in each group): control, FOXP2overexpressing (two stable cell lines) and FOXP2-CPED1- overexpressing (two cell lines) groups. The experiment lasted for 4 months. We observed that no RWPE-1 cell-injected mice developed tumor masses. We propose that FOXP2 and its fusion alone are not sufficient to generate the microenvironment suitable for RWPE-1-xenograft growth. Collectively, our data suggest that FOXP2 has oncogenic potential in prostate cancer, but is not sufficient to act alone as an oncogene.

3) There is a limited mechanism of action. While the authors provide correlative data suggesting that FOXP2 could increase the expression of MET signaling components, it is not clear how FOXP2 controls MET levels. It would be of interest to search for and validate the importance of potential FOXP2 binding sites in or around MET and the genes of METassociated proteins. At a minimum, it should be confirmed whether MET is a primary or secondary target of FOXP2. The authors should also report on what happened to the 4-gene MET signature in the FOXP2 knockdown cell models. It would be equally significant to test if overexpression of MET can rescue the anti-growth effects of FOXP2 knockdown in prostate cancer cells (positive or negative results would be informative).

We appreciate all the valuable comments. As suggested, we performed corresponding experiments, please refer to the " Answers to Essential Revisions #2 from the Editors”, to the "Answer to Essential Revisions #6 from the Editors”, and to the "Answer to Essential Revisions #7 from the Editors” for details.

Reviewer #3 (Public Review):

1) In this manuscript, the authors present data supporting FOXP2 as an oncogene in PCa. They show that FOXP2 is overexpressed in PCa patient tissue and is necessary and sufficient for PCa transformation/tumorigenesis depending on the model system. Overexpression and knock-down of FOXP2 lead to an increase/decrease in MET/PI3K/AKT transcripts and signaling and sensitizes cells to PI3K/AKT inhibition.

Key strengths of the paper include multiple endpoints and model systems, an over-expression and knock-down approach to address sufficiency and necessity, a new mouse knock-in model, analysis of primary PCa patient tumors, and benchmarking finding against publicly available data. The central discovery that FOXP2 is an oncogene in PCa will be of interest to the field. However, there are several critically unanswered questions.

1) No data are presented for how FOXP2 regulates MET signaling. ChIP would easily address if it is direct regulation of MET and analysis of FOXP2 ChIP-seq could provide insights.

2) Beyond the 2 fusions in the 100 PCa patient cohort it is unclear how FOXP2 is overexpressed in PCa. In the discussion and in FS5 some data are presented indicating amplification and CNAs, however, these are not directly linked to FOXP2 expression.

3) There are some hints that full-length FOXP2 and the FOXP2-CPED1 function differently. In SF2E the size/number of colonies between full-length FOXP2 and fusion are different. If the assay was run for the same length of time, then it indicates different biologies of the overexpressed FOXP2 and FOXP2-CPED1 fusion. Additionally, in F3E the sensitization is different depending on the transgene.

We appreciate these valuable comments and constructive remarks. As suggested, we performed the CUT&Tag experiments to detect the binding of FOXP2 to MET, and to examine the association of CNAs of FOXP2 with its expression. Please refer to the " Answer to Essential Revisions #2 from the Editors" and the " Answer to Essential Revisions #4 from the Editors" for details. We also added detailed information to show the resemblance observed between FOXP2 fusion- and wild-type FOXP2-overexpressing cells. We added the corresponding description to the Results section in Line 487 on Page 22 in the tracked changes version of the revised manuscript. Please refer to the “Answer to Essential Revisions #5 from the Editors” for details.

2) The relationship between FOXP2 and AR is not explored, which is important given 1) the critical role of the AR in PCa; and 2) the existing relationship between the AR and FOXP2 and other FOX gene members.

We thank the reviewer very much for highlighting this point. We agree that it is important to examine the relationship between FOXP2 and AR. We therefore analyzed the expression dataset of 255 primary prostate tumors from TCGA and observed that the expression of FOXP2 was significantly correlated with the expression of AR (Spearman's ρ = 0.48, P < 0.001) (Figure 1. a). Next, we observed that both FOXP2- and FOXP2-CPED1overexpressing 293T cells had a higher AR protein abundance than control cells (Figure 1. b). In addition, shRNA-mediated FOXP2 knockdown in LNCaP cells resulted in a decreased AR protein level compared to that in control cells (Figure 1. c). However, we analyzed our CUT&Tag data and observed no binding of FOXP2 to AR (Figure 1. d). Our data suggest that FOXP2 might be associated with AR expression.

Figure 1. a. AR expression in a human prostate cancer dataset (TCGA, Prostate Adenocarcinoma, Provisional; n = 493) classified by FOXP2 expression level (bottom 25%, low expression, n = 120; top 25%, high expression, n = 120; negative expression, n = 15). P values were calculated by the MannWhitney U test. The correlation between FOXP2 and AR expression was evaluated by determining the Spearman's rank correlation coefficient. b. Immunoblot analysis of the expression levels of AR in 293T cells with overexpression of FOXP2 or FOXP2-CPED1. c. Immunoblot analysis of the expression levels of AR in LNCaP cells with stable expression of the scrambled vector or FOXP2 shRNA. d. CUT&Tag analysis of FOXP2 association with the promoter of AR. Representative track of FOXP2 at the AR gene locus is shown.

Reference

1. Mayr C, Bartel DP. Widespread shortening of 3'UTRs by alternative cleavage and polyadenylation activates oncogenes in cancer cells. Cell. 2009 Aug 21;138(4):673-84.
2. Gara SK, Jia L, Merino MJ, Agarwal SK, Zhang L, Cam M et al., Germline HABP2 Mutation Causing Familial Nonmedullary Thyroid Cancer. N Engl J Med. 2015 Jul 30;373(5):448-55.
3. Kohno T, Ichikawa H, Totoki Y, Yasuda K, Hiramoto M, Nammo T et al., KIF5B-RET fusions in lung adenocarcinoma. Nat Med. 2012 Feb 12;18(3):375-7.
4. Chen F, Byrd AL, Liu J, Flight RM, DuCote TJ, Naughton KJ et al., Polycomb deficiency drives a FOXP2-high aggressive state targetable by epigenetic inhibitors. Nat Commun. 2023 Jan 20;14(1):336.
5. Kaya-Okur HS, Wu SJ, Codomo CA, Pledger ES, Bryson TD, Henikoff JG et al., CUT&Tag for efficient epigenomic profiling of small samples and single cells. Nat Commun. 2019 Apr 29;10(1):1930.
6. Spiteri E, Konopka G, Coppola G, Bomar J, Oldham M, Ou J et al., Identification of the transcriptional targets of FOXP2, a gene linked to speech and language, in developing human brain. Am J Hum Genet. 2007 Dec;81(6):1144-57.
7. Lai CS, Fisher SE, Hurst JA, Vargha-Khadem F, Monaco AP. A forkhead-domain gene is mutated in a severe speech and language disorder. Nature. 2001 Oct 4;413(6855):519-23.
8. Hannenhalli S, Kaestner KH. The evolution of Fox genes and their role in development and disease. Nat Rev Genet. 2009 Apr;10(4):233-40.
9. Shu W, Yang H, Zhang L, Lu MM, Morrisey EE. Characterization of a new subfamily of winged-helix/forkhead (Fox) genes that are expressed in the lung and act as transcriptional repressors. J Biol Chem. 2001 Jul 20;276(29):27488-97.
10. Wang C, Liu H, Qiu Q, Zhang Z, Gu Y, He Z. TCRP1 promotes NIH/3T3 cell transformation by over-activating PDK1 and AKT1. Oncogenesis. 2017 Apr 24;6(4):e323.
11. Suh YA, Arnold RS, Lassegue B, Shi J, Xu X, Sorescu D et al., Cell transformation by the superoxide-generating oxidase Mox1. Nature. 1999 Sep 2;401(6748):79-82.

Author Response

Reviewer #1 (Public Review):

The present study by Zander et al. aims at improving our understanding of CD4+ T cell heterogeneity in response to chronic viral infections. The authors utilize the murine LCMV c13 infection model and perform single cell RNA seq analysis on day 10 post infection to identify multiple, previously unappreciated, T cell subsets. The authors then go on and verify these analyses using multi-color flow cytometry before comparing the transcriptome of CD4 T cells from chronic infection to a previously generated data set of CD4 T cells obtained from acutely-resolved LCMV infection.

The analyses are very well done and provide some interesting novel insights. In particular, the comparison of CD4 T cell subsets across acute and chronic infections is very exciting as they provide a very valuable platform that can answer a long-standing question: do CD4 T cells in chronic infection undergo exhaustion similar to CD8 T cells. While this has been proposed for an extended period, this new dataset by Zander et al. can provide some novel insights by comparing individual cell subsets cross-infection. The manuscript would, however, benefit from a more extensive analysis and focus on this interesting point.

We thank the reviewer for their time and careful assessment of our manuscript. We were happy to hear that the reviewer found our work interesting.

On that note, the authors should take advantage of more accurate and present gene datasets to compare the 'dysfunctional' state of CD4 T cells in chronic infection vs acute infection. Also, a different illustration to demonstrate the module score analyses would be more intuitive.

We have now included T cell “exhaustion” genesets from recently published data (Zander et. al 2019 Immunity), and we have also displayed the relative expression of select signature genes from these genesets in an updated supplemental figure 3.

Also, at multiple sections in the manuscript, the authors are missing the accurate citations as they are still mentioned as '(Ref)'.

We apologize for this oversight and have corrected these citations.

Nevertheless, this study does not require major revisions.

Reviewer #2 (Public Review):

In their study "Delineating the transcriptional landscape and clonal diversity of virus-specific CD4+ T cells during chronic viral infection" Zander and co-workers analyze the phenotypic and clonotypic distributions of T cells specific to a LCMV epitope following infection with a chronic LCMV strain in mice. The paper largely follows an earlier study from the same group (Khatun JEM 2021) that has used a similar experimental strategy to analyze T cells responding to an LCMV strain establishing acute infection, and it adds a scTCRseq component to another earlier study of chronic LCMV (Zander Immunity 2022). The main contributions of the paper are to demonstrate that interesting differences between gene expression profiles between chronic and acute LCMV exist, and to identify a new T cell subset (of unknown functional significance).

While the paper is framed around differences between T cell responses to acute and chronic infections, all analysis is done on T cells at day 10 post primary infection. At such an early time point even the acute LCMV strain virus is likely not completely cleared, or at the very least viral antigens are still presented. The relevance of the presented phenotypic differences to other settings with long-term chronic infection is thus questionable. Additionally, there are a number of methodological concerns regarding the robustness of the statistical and bioinformatic analyses that put in doubt some of the conclusions. Most notably, the analysis of fate biases needs to be substantiated by tests against baseline expectations from random assortment to test for statistical significance.

We thank the reviewer for their careful review of our manuscript as well as their helpful comments.

Regarding the day 10 time point-post LCMV Armstrong infection, several groups have previously reported that LCMV viral load is undetectable by day 10 post-infection (see one published example below), although we completely agree with the reviewer that there is still likely to be viral antigens being presented at this time point, as well as ongoing inflammation, which we believe (and as discussed further below) is actually a strength of the study as it allows for a more fair comparison of the transcriptional state of recently stimulated virus-specific CD4 T cells under different contexts (acute vs chronic LCMV infection) . We chose day 10 post LCMV Cl13 and LCMV Armstrong infections as the timepoint for analysis, as this is approximately the peak of the endogenous Gp66-77 CD4+ T cell response (see previously published data below), and is also when there is a more balanced distribution of Th1, Tfh, and T central memory precursor (Tcmp)/ or memory-like cells in these settings, thereby allowing for sufficient numbers of cells/cluster to conduct an in-depth analysis and high-resolution comparison of these subsets between the two different infections. Further, as some degree of TCR stimulation is still likely being experienced at this timepoint during LCMV Armstrong infection, we believe that this is a more useful comparison than at a memory time point (when CD4 T cells are in a quiescent state) as it gives us a better picture of the differentially expressed genes at the peak of the CD4 T cell response, and also provides insight into how chronic viral infection perturbs the transcriptional program of CD4 T cells.

Author Response

Reviewer #2 (Public Review):

This study evaluates the causal relationship between childhood obesity on the one hand, and childhood emotional and behavioral problems on the other. It applies Mendelian Randomization (MR), a family of methods in statistical genetics that uses genetic markers to break the symmetry between correlated traits, allowing inference of causation rather than mere correlation. The authors argue convincingly that previous studies of these traits, both those using non-genetic observational epidemiology methods and those using standard MR methods, may be confounded by demographic effects and familial effects. One possible example of this kind of confounding is that the idea that obesity in parents may contribute to emotional and behavioral problems in children; another is the idea that adults with emotional and behavioral issues may be more likely to have children with partners who are obese, and vice-versa. They then make use of a recently proposed "within-family" MR method, which should effectively control for these confounders, at the cost of higher uncertainty in the estimated effect size, and therefore lower power to detect small effects. They report that none of the previously reported associations of childhood BMI with anxiety, depression, or ADHD are replicated using the within-family MR method, and that in the case of depression the primary association appears to be with maternal BMI rather than the child's own BMI.

This argument that these confounders may affect these phenotypes is fairly sound, and within-family MR should indeed do a good job of controlling for them. I do not see any major issues with the cohort itself or the choice of genetic instruments. I also do not see any major issues with the definitions or ascertainment of the phenotypes studied, though I am not an expert on any of these phenotypes in particular. I am especially satisfied with the series of analyses demonstrating that the results are robust to many variations of MR methodology. Overall, I think the positive result this study reports is very credible: that the known association between childhood BMI and depression is likely primarily due to an effect of maternal BMI rather than the child's own BMI (though given that paternal BMI has a similar effect size with only a slightly wider confidence interval, I would instead say that the effect is from parental BMI generally, not specifically maternal.)

In the updated results based on the larger genetic data release, the estimates for the association of maternal BMI and paternal BMI with the child’s depressive symptoms are more clearly different than they were in the smaller dataset (for maternal BMI, beta= 0.11, CI:0.02,0.19, p=0.01; for paternal BMI, beta=0.02, CI:-0.09,0.12, p=0.71). Therefore, in this version, it makes sense to note an association with maternal BMI specifically.

The main weakness of the study comes from its negative results, which the authors emphasize as their primary conclusion: that previously reported associations of childhood BMI with anxiety, depression, and ADHD are not replicated using within-family MR methods. These claims do not seem justified by the evidence presented in this study. In fact, in every panel of figures 2 and 3, the error bars for the within-family MR analysis encompass the estimates for both the regression analysis and the traditional MR analysis, suggesting that the within-family analysis provides no evidence one way or another about which of these analyses is more accurate. More generally, in order to convincingly claim that there is no causal relationship between two traits, an MR study must argue that the study would be powered to detect a relationship if one existed. Within-family MR methods are known to have less power to detect associations and less precision to estimate effect sizes than traditional MR methods or traditional observational epidemiology methods, so it is not sufficient to show that these other methods have power to detect the association. To make this kind of claim, it is necessary to include some kind of power analysis, such as a simulation study or analytic power calculations, and likely also a positive control to show that this method does have power to detect known effects in this cohort.

We agree that it is imperative that negative (i.e. “non-significant”) results are correctly interpreted - it is just as important to discover what is unlikely to affect emotional and behavioural outcomes as what does affect them. Negative results (non-significant estimates) are neither a weakness nor strength of the study, but simply reflect the estimation error in our analysis of the data. The key question is whether our within-family MR estimates are sufficiently powered to detect effect sizes of interest or rule out clinically meaningful effect sizes – or are they simply too imprecise to draw any conclusions? As the reviewer suggests, one way to address this is via a post-hoc power calculation. We consider post-hoc power calculations redundant, since all the information about the power of our analysis is reflected in the standard errors and reported confidence intervals. Moreover, any post-hoc power calculation will be necessarily approximate compared to using the standard errors and confidence intervals which we report.

Despite these methodological reservations, we have conducted simulations to estimate the power of our within-family models (the R code is included at the end of this document). These simulations indicate that we do have sufficient power to detect the size of effects seen for depressive symptoms and ADHD in models using the adult BMI PGS. They also indicate that we cannot rule out smaller effects for non-significant associations (e.g., for the impact of the child’s BMI on anxiety). Naturally, this is entirely consistent with the width of the confidence intervals reported in results tables and in Figures 1 and 2. However, although power calculations are important when planning a study, they make little contribution to interpretation once a study has been conducted and confidence intervals are available (e.g., https://psyarxiv.com/tcqrn/). For this reason, we comment on these simulations in this response to reviewers but do not include them in the manuscript or supplementary materials. At the same time, we have changed the language used in the manuscript to be clearer that the results were imprecise and that values contained within the confidence limits cannot be ruled out.

For example, the discussion now includes the following:

‘However, within-family MR estimates using the childhood body size PGS are still consistent with small effects of the child’s BMI on all outcomes, with upper confidence limits around a 0.2 standard-deviation increase in the outcome per 5kg/m2 increase in BMI.’

And the conclusion of the paper now reads:

‘Our results suggest that genetic variation associated with BMI in adulthood affects a child’s depressive and ADHD symptoms, but genetic variation associated with recalled childhood body size does not substantially affect these outcomes. There was little evidence that BMI affects anxiety. However, our estimates were imprecise, and these differences may be due to estimation error. There was little evidence that parental BMI affects a child’s ADHD or anxiety symptoms, but factors associated with maternal BMI may independently influence a child’s depressive symptoms. Genetic studies using unrelated individuals, or polygenic scores for adult BMI, may have overestimated the causal effects of a child’s own BMI.’

Regarding a positive control: for analyses of BMI in adults, suitable positive controls would include directly measured biomarkers such as fat mass or blood pressure or reported medical outcomes like type 2 diabetes. In adolescents and younger adults, age at menarche or other measures of puberty can be used, as these are reliably influenced by BMI. However, the age of the participants for whom within-family effects are being estimated (8 years), together with the lack of any biomarkers such as fat mass (due to the questionnaire-based survey design) mean no suitable measures are available.

Reviewer #3 (Public Review):

Higher BMI in childhood is correlated with behavioral problems (e.g. depression and ADHD) and some studies have shown that this relationship may be causal using Mendelian Randomization (MR). However, traditional MR is susceptible to bias due to population stratification, assortative mating, and indirect effects (dynastic effects). To address this issue, Hughes et al. use within-family MR, which should be immune to the above-listed problems. They were unable to find a causal relationship between children's BMI and depression, anxiety, or ADHD. They do, however, report a causal effect of mother's BMI on depression in their children. They conclude that the causal effect of children's BMI on behavioral phenotypes such as depression and anxiety, if present, is very small, and may have been overestimated in previous studies. The analyses have been carried out carefully in a large sample and the paper is presented clearly. Overall, their assertions are justified but given that the conclusions mostly rest on an absence of an effect, I would like to see more discussion on statistical power.

1) The authors show that the estimates of within-family MR are imprecise. It would be helpful to know how much power they have for estimating effect sizes reported previously given their sample size.

As discussed in response to a comment from reviewer 2, the power of our results is already indicated by our standard errors and confidence intervals. Nevertheless, we conducted simulations to estimate the size of effects which we had 80% power to detect. Results, presented below, are consistent with our main results. As discussed in response to a comment from reviewer 2, we consider post-hoc power calculations redundant when standard errors and confidence intervals are reported; for this reason, we include this information in the response to reviewers but not the manuscript itself.

2) They used the correlation between PGS and BMI to support the assertion that the former is a strong instrument. Were the reported correlations calculated across all individuals? Since we know that stratification, assortative mating, and indirect effects can inflate these correlations, perhaps a more unbiased estimate would be the proportion of children's BMI variance explained by their PGS conditioned on the parents' PGS. This should also be the estimate used in power calculations.

The manuscript has been updated to quote Sanderson-Windmeijer conditional R2 values: the proportion of BMI variance explained by the BMI PGS for each member of a trio, conditional on the PGS of the other members of the trio, and all genetic covariates included in within-family models. Similarly, we now show Sanderson-Windmeijer conditional F-statistics for a model including the child, mother, and father’s BMI instrumented by the child, mother, and father’s PGS.

3) In testing the association of mothers' and fathers' BMI with children's symptoms, the authors used a multivariable linear regression conditioning on the child's own BMI. Was the other parent's BMI (either by itself or using the polygenic score) included as a covariate in the multivariable and MR models? This was not entirely clear from the text or from Fig. 2. I suspect that if there were assortative mating on BMI in the parent's generation, the effect of any one parent's BMI on the child's symptoms might be inflated unless the other parent's BMI was included as a covariate (assuming both mother's and father's BMI affect the child's symptoms).

Non-genetic models include both the mother and father’s phenotypic BMI as well as the child’s, allowing estimation of conditional effects of all three. This controls for assortative mating as noted by the reviewer. This was not previously clear - all relevant text and figure captions have been updated to clarify this.

4) They report no evidence of cross-trait assortative mating in the parents generation. The power to detect cross-trait assortative mating in the parents' generation using PGS would depend on the actual strength of assortative mating and the respective proportions of trait variance explained by PGS. Could the authors provide an estimate of the power for this test in their sample?

We have updated the discussion of assortative mating (in both the results and the discussion section) to note possible limitations of power and clarify that that this approach to examining assortment may not capture its full extent.

The relevant part of the results section now reads:

“In the parents’ generation, phenotypes were associated within parental pairs, consistent with assortative mating on these traits (Appendix 1 – Table 5). Adjusted for ancestry and other genetic covariates, maternal and paternal BMI were positively associated (beta: 0.23, 95%CI: 0.22,0.25, p<0.001), as were maternal and paternal depressive symptoms (beta: 0.18, 95%CI: 0.16,0.20, p<0.001), and maternal and paternal ADHD symptoms (beta: 0.11, 95%CI: 0.09,0.13, p<0.001). Consistent with cross-trait assortative mating, there was an association of mother’s BMI with father’s ADHD symptoms (beta: 0.03, 95%CI: 0.02,0.05, p<0.001) and mother’s ADHD symptoms with father’s depressive symptoms (beta: 0.05,95%CI: 0.05,0.06, p<0.001). Phenotypic associations can reflect the influence of one partner on another as well as selection into partnerships, but regression models of paternal polygenic scores on maternal polygenic scores also pointed to a degree of assortative mating. Adjusted for ancestry and genotyping covariates, there were small associations between parents’ BMI polygenic scores (beta: 0.01, 95%CI: 0.00,0.02, p=0.02 for the adult BMI PGS, and beta: 0.01, 95%CI: 0.00,0.02, p=0.008 for the childhood body size PGS), and of the mother’s childhood body size PGS with the father’s ADHD PGS (beta: 0.01, 95%CI: 0.00,0.02, p=0.03). We did not detect associations with pairs of other polygenic scores, which may be due to insufficient statistical power.”

And the relevant part of the discussion section now reads:

“We found some genomic evidence of assortative mating for BMI, and cross-trait assortative mating between BMI and ADHD, but not between other traits. However, associations between polygenic scores, which only capture some of the genetic variation associated with these phenotypes, may not capture the full extent of genetic assortment on these traits.”

5) Are the actual phenotypes (BMI, depression or ADHD) correlated between the parents? If so, would this not suffice as evidence of cross-trait assortative mating? It is known that the genetic correlation between parents as a result of assortative mating is a function of the correlation in their phenotypes and the heritabilities underlying the two traits (e.g., see Yengo and Visscher 2018). An alternative way to estimate the genetic correlation between parents without using PGS (which is noisy and therefore underpowered) would be to use the phenotypic correlation and heritability estimated using GREML or LDSC. Perhaps this is outside the scope of the paper but I would like to hear the author's thoughts on this.

Associations between maternal and paternal phenotypes are consistent with a degree of assortative mating (shown below). These results have added to Appendix 1 - Table 5, which also shows associations between maternal and paternal polygenic scores, and methods and results updated accordingly (see quoted text in response to the comment above). For comparability, both sets of results are based on regression models adjusting for the mother’s and father’s ancestry PCs and genotyping covariates. We agree that analysis of assortative mating using GREML or LDSC is out of scope for this paper. As noted above, we have updated the discussion to acknowledge the limitations of the approach taken:

‘We found some genomic evidence of assortative mating for BMI, and cross-trait assortative mating between BMI and ADHD, but not between other traits. However, associations between polygenic scores, which only capture some of the genetic variation associated with these phenotypes, may not capture the full extent of genetic assortment on these traits.’

6) It would be helpful to include power calculations for the MR-Egger intercept estimates.

As with our response to the comments above, post-hoc power calculations are redundant, as all the information about the power of our analysis, including the MR-Egger is indicated by the standard errors and confidence intervals. MR-Egger is less precise than other estimators, as is made clear from the wide confidence intervals reported in the relevant tables (Appendix 1 - Tables 8 and 9). However, we have now updated the discussion to give more weight to this as a limitation. The discussion of pleiotropy in the final paragraph of the discussion now reads:

‘While robustness checks found little evidence of pleiotropy, these methods rely on assumptions. Moreover, MR-Egger is known to give imprecise estimates (Burgess and Thompson 2017), and confidence intervals from MR-Egger models were wide. Thus, pleiotropy cannot be ruled out.’

Similarly, we have updated the relevant line of the results section, which now reads:

‘MR-Egger models found little evidence of horizontal pleiotropy, although MR-Egger estimates were imprecise (Appendix 1 - Tables 8 and 9).’

7) Finally, what is the correlation between PGS and genetic PCs/geography in their sample? A correlation might provide evidence to support the point that classic MR effects are inflated due to stratification.

Figures presenting the association of the child’s BMI polygenic scores and their PCs have been added to the supplementary information as Appendix 1 - Figure 2 and Appendix 1 - Figure 3. Consistent with an influence of residual stratification, a regression of the child’s BMI polygenic scores against their ancestry PCs (adjusting for genotyping centre and chip) found that 7 of the 20 PCs were associated at p<0.05 with the adult BMI PGS, and 8 of 20 with the childhood body size PGS (under the null hypothesis, we would expect one association in each case). When parental polygenic scores were added to the models, these associations attenuated towards to null.

Author Response

Reviewer #1 (Public Review):

This manuscript seeks to identify the mechanism underlying priority effects in a plantmicrobe-pollinator model system and to explore its evolutionary and functional consequences. The manuscript first documents alternative community states in the wild: flowers tend to be strongly dominated by either bacteria or yeast but not both. Then lab experiments are used to show that bacteria lower the nectar pH, which inhibits yeast - thereby identifying a mechanism for the observed priority effect. The authors then perform an experimental evolution unfortunately experiment which shows that yeast can evolve tolerance to a lower pH. Finally, the authors show that low-pH nectar reduces pollinator consumption, suggesting a functional impact on the plant-pollinator system. Together, these multiple lines of evidence build a strong case that pH has far-reaching effects on the microbial community and beyond.

The paper is notable for the diverse approaches taken, including field observations, lab microbial competition and evolution experiments, genome resequencing of evolved strains, and field experiments with artificial flowers and nectar. This breadth can sometimes seem a bit overwhelming. The model system has been well developed by this group and is simple enough to dissect but also relevant and realistic. Whether the mechanism and interactions observed in this system can be extrapolated to other systems remains to be seen. The experimental design is generally sound. In terms of methods, the abundance of bacteria and yeast is measured using colony counts, and given that most microbes are uncultivable, it is important to show that these colony counts reflect true cell abundance in the nectar.

We have revised the text to address the relationship between cell counts and colony counts with nectar microbes. Specifically, we point out that our previous work (Peay et al. 2012) established a close correlation between CFUs and cell densities (r2 = 0.76) for six species of nectar yeasts isolated from D. aurantiacus nectar at Jasper Ridge, including M. reukaufii.

As for A. nectaris, we used a flow cytometric sorting technique to examine the relationship between cell density and CFU (figure supplement 1). This result should be viewed as preliminary given the low level of replication, but this relationship also appears to be linear, as shown below, indicating that colony counts likely reflect true cell abundance of this species in nectar.

It remains uncertain how closely CFU reflects total cell abundance of the entire bacterial and fungal community in nectar. However, a close association is possible and may be even likely given the data above, showing a close correlation between CFU and total cell count for several yeast species and A. nectaris, which are indicated by our data to be dominant species in nectar.

We have added the above points in the manuscript (lines 263-264, 938-932).

The genome resequencing to identify pH-driven mutations is, in my mind, the least connected and developed part of the manuscript, and could be removed to sharpen and shorten the manuscript.

We appreciate this perspective. However, given the disagreement between this perspective and reviewer 2’s, which asks for a more expanded section, we have decided to add a few additional lines (lines 628-637), briefly expanding on the genomic differences between strains evolved in bacteria-conditioned nectar and those evolved in low-pH nectar.

Overall, I think the authors achieve their aims of identifying a mechanism (pH) for the priority effect of early-colonizing bacteria on later-arriving yeast. The evolution and pollinator experiments show that pH has the potential for broader effects too. It is surprising that the authors do not discuss the inverse priority effect of early-arriving yeast on later-arriving bacteria, beyond a supplemental figure. Understandably this part of the story may warrant a separate manuscript.

We would like to point out that, in our original manuscript, we did discuss the inverse priority effects, referring to relevant findings that we previously reported (Tucker and Fukami 2014, Dhami et al. 2016 and 2018, Vannette and Fukami 2018). Specifically, we wrote that: “when yeast arrive first to nectar, they deplete nutrients such as amino acids and limit subsequent bacterial growth, thereby avoiding pH-driven suppression that would happen if bacteria were initially more abundant (Tucker and Fukami 2014; Vannette and Fukami 2018)” (lines 385-388). However, we now realize that this brief mention of the inverse priority effects was not sufficiently linked to our motivation for focusing mainly on the priority effects of bacteria on yeast in the present paper. Accordingly, we added the following sentences: “Since our previous papers sought to elucidate priority effects of early-arriving yeast, here we focus primarily on the other side of the priority effects, where initial dominance of bacteria inhibits yeast growth.” (lines 398-401).

I anticipate this paper will have a significant impact because it is a nice model for how one might identify and validate a mechanism for community-level interactions. I suspect it will be cited as a rare example of the mechanistic basis of priority effects, even across many systems (not just pollinator-microbe systems). It illustrates nicely a more general ecological phenomenon and is presented in a way that is accessible to a broader audience.

Thank you for this positive assessment.

Reviewer #2 (Public Review):

The manuscript "pH as an eco-evolutionary driver of priority effects" by Chappell et al illustrates how a single driver-microbial-induced pH change can affect multiple levels of species interactions including microbial community structure, microbial evolutionary change, and hummingbird nectar consumption (potentially influencing both microbial dispersal and plant reproduction). It is an elegant study with different interacting parts: from laboratory to field experiments addressing mechanism, condition, evolution, and functional consequences. It will likely be of interest to a wide audience and has implications for microbial, plant, and animal ecology and evolution.

This is a well-written manuscript, with generally clear and informative figures. It represents a large body and variety of work that is novel and relevant (all major strengths).

We appreciate this positive assessment.

Overall, the authors' claims and conclusions are justified by the data. There are a few things that could be addressed in more detail in the manuscript. The most important weakness in terms of lack of information/discussion is that it looks like there are just as many or more genomic differences between the bacterial-conditioned evolved strains and the low-pH evolved strains than there are between these and the normal nectar media evolved strains. I don't think this negates the main conclusion that pH is the primary driver of priority effects in this system, but it does open the question of what you are missing when you focus only on pH. I would like to see a discussion of the differences between bacteria-conditioned vs. low-pH evolved strains.

We agree with the reviewer and have included an expanded discussion in the revised manuscript [lines 628-637]. Specifically, to show overall genomic variation between treatments, we calculated genome-wide Fst comparing the various nectar conditions. We found that Fst was 0.0013, 0.0014, and 0.0015 for the low-pH vs. normal, low pH vs. bacteria-conditioned, and bacteria-conditioned vs. normal comparisons, respectively. The similarity between all treatments suggests that the differences between bacteria-conditioned and low pH are comparable to each treatment compared to normal. This result highlights that, although our phenotypic data suggest alterations to pH as the most important factor for this priority effect, it still may be one of many affecting the coevolutionary dynamics of wild yeast in the microbial communities they are part of. In the full community context in which these microbes grow in the field, multi-species interactions, environmental microclimates, etc. likely also play a role in rapid adaptation of these microbes which was not investigated in the current study.

Based on this overall picture, we have included additional discussion focusing on the effect of pH on evolution of stronger resistance to priority effects. We compared genomic differences between bacteria-conditioned and low-pH evolved strains, drawing the reader’s attention to specific differences in source data 14-15. Loci that varied between the low pH and bacteria-conditioned treatments occurred in genes associated with protein folding, amino acid biosynthesis, and metabolism.

Reviewer #3 (Public Review):

This work seeks to identify a common factor governing priority effects, including mechanism, condition, evolution, and functional consequences. It is suggested that environmental pH is the main factor that explains various aspects of priority effects across levels of biological organization. Building upon this well-studied nectar microbiome system, it is suggested that pH-mediated priority effects give rise to bacterial and yeast dominance as alternative community states. Furthermore, pH determines both the strengths and limits of priority effects through rapid evolution, with functional consequences for the host plant's reproduction. These data contribute to ongoing discussions of deterministic and stochastic drivers of community assembly processes.

Strengths:

Provides multiple lines of field and laboratory evidence to show that pH is the main factor shaping priority effects in the nectar microbiome. Field surveys characterize the distribution of microbial communities with flowers frequently dominated by either bacteria or yeast, suggesting that inhibitory priority effects explain these patterns. Microcosm experiments showed that A. nectaris (bacteria) showed negative inhibitory priority effects against M. reukaffi (yeast). Furthermore, high densities of bacteria were correlated with lower pH potentially due to bacteria-induced reduction in nectar pH. Experimental evolution showed that yeast evolved in low-pH and bacteria-conditioned treatments were less affected by priority effects as compared to ancestral yeast populations. This potentially explains the variation of bacteria-dominated flowers observed in the field, as yeast rapidly evolves resistance to bacterial priority effects. Genome sequencing further reveals that phenotypic changes in low-pH and bacteriaconditioned nectar treatments corresponded to genomic variation. Lastly, a field experiment showed that low nectar pH reduced flower visitation by hummingbirds. pH not only affected microbial priority effects but also has functional consequences for host plants.

We appreciate this positive assessment.

Weaknesses:

The conclusions of this paper are generally well-supported by the data, but some aspects of the experiments and analysis need to be clarified and expanded.

The authors imply that in their field surveys flowers were frequently dominated by bacteria or yeast, but rarely together. The authors argue that the distributional patterns of bacteria and yeast are therefore indicative of alternative states. In each of the 12 sites, 96 flowers were sampled for nectar microbes. However, it's unclear to what degree the spatial proximity of flowers within each of the sampled sites biased the observed distribution patterns. Furthermore, seasonal patterns may also influence microbial distribution patterns, especially in the case of co-dominated flowers. Temperature and moisture might influence the dominance patterns of bacteria and yeast.

We agree that these factors could potentially explain the presented results. Accordingly, we conducted spatial and seasonal analyses of the data, which we detail below and include in two new paragraphs in the manuscript [lines 290-309].

First, to determine whether spatial proximity influenced yeast and bacterial CFUs, we regressed the geographic distance between all possible pairs of plants to the difference in bacterial or fungal abundance between the paired plants. If plant location affected microbial abundance, one should see a positive relationship between distance and the difference in microbial abundance between a given pair of plants: a pair of plants that were more distantly located from each other should be, on average, more different in microbial abundance. Contrary to this expectation, we found no significant relationship between distance and the difference in bacterial colonization (A, p=0.07, R2=0.0003) and a small negative association between distance and the difference in fungal colonization (B, p<0.05, R2=0.004). Thus, there was no obvious overall spatial pattern in whether flowers were dominated by yeast or bacteria.

Next, to determine whether climatic factors or seasonality affected the colonization of bacteria and yeast per plant, we used a linear mixed model predicting the average bacteria and yeast density per plant from average annual temperature, temperature seasonality, and annual precipitation at each site, the date the site was sampled, and the site location and plant as nested random effects. We found that none of these variables were significantly associated with the density of bacteria and yeast in each plant.

To look at seasonality, we also re-ordered Fig 2C, which shows the abundance of bacteria- and yeast-dominated flowers at each site, so that the sites are now listed in order of sampling dates. In this re-ordered figure, there is no obvious trend in the number of flowers dominated by yeast throughout the period sampled (6.23 to 7/9), giving additional indication that seasonality was unlikely to affect the results.

Additionally, sampling date does not seem to strongly predict bacterial or fungal density within each flower when plotted.

These additional analyses, now included (figure supplements 2-4) and described (lines 290-309) in the manuscript, indicate that the observed microbial distribution patterns are unlikely to have been strongly influenced by spatial proximity, temperature, moisture, or seasonality, reinforcing the possibility that the distribution patterns instead indicate bacterial and yeast dominance as alternative stable states.

The authors exposed yeast to nectar treatments varying in pH levels. Using experimental evolution approaches, the authors determined that yeast grown in low pH nectar treatments were more resistant to priority effects by bacteria. The metric used to determine the bacteria's priority effect strength on yeast does not seem to take into account factors that limit growth, such as the environmental carrying capacity. In addition, yeast evolves in normal (pH =6) and low pH (3) nectar treatments, but it's unclear how resistance differs across a range of pH levels (ranging from low to high pH) and affects the cost of yeast resistance to bacteria priority effects. The cost of resistance may influence yeast life-history traits.

The strength of bacterial priority effects on yeast was calculated using the metric we previously published in Vannette and Fukami (2014): PE = log(BY/(-Y)) - log(YB/(Y-)), where BY and YB represent the final yeast density when early arrival (day 0 of the experiment) was by bacteria or yeast, followed by late arrival by yeast or bacteria (day 2), respectively, and -Y and Y- represent the final density of yeast in monoculture when they were introduced late or early, respectively. This metric does not incorporate carrying capacity. However, it does compare how each microbial species grows alone, relative to growth before or after a competitor. In this way, our metric compares environmental differences between treatments while also taking into account growth differences between strains.

Here we also present additional growth data to address the reviewer’s point about carrying capacity. Our experiments that compared ancestral and evolved yeast were conducted over the course of two days of growth. In preliminary monoculture growth experiments of each evolved strain, we found that yeast populations did reach carrying capacity over the course of the two-day experiment and population size declined or stayed constant after three and four days of growth.

However, we found no significant difference in monoculture growth between the ancestral stains and any of the evolved strains, as shown in Figure supplement 12B. This lack of significant difference in monoculture suggests that differences in intrinsic growth rate do not fully explain the priority effects results we present. Instead, differences in growth were specific to yeast’s response to early arrival by bacteria.

We also appreciate the reviewer’s comment about how yeast evolves resistance across a range of pH levels, as well as the effect of pH on yeast life-history traits. In fact, reviewer #2 pointed out an interesting trade-off in life history traits between growth and resistance to priority effects that we now include in the discussion (lines 535-551) as well as a figure in the manuscript (Figure 8).

Author Response

The following is the authors’ response to the original reviews.

Reviewer #1 (Recommendations For The Authors):

The brain-machine interface used in this study differs from typical BMIs in that it's not intended to give subjects voluntary control over their environment. However, it is possible that rats may become aware of their ability to manipulate trial start times using their neural activity. Is there any evidence that the time required to initiate trials on high-coherence or low-coherence trials decreases with experience?

This is a great question. First, we designed the experiment to avoid this possibility. Rats were experienced on the sequence of the automatic maze both pre and post implantation (totaling to weeks of pre-training and habituation). As such, the majority of the trials ever experienced by the rat were not controlled by their neural activity. During BMI experimentation, only 10% of trials were triggered during high coherence states and 10% for low coherence states, leaving ~80% of trials not controlled by their neural activity. We also implemented a pseudo-randomized trial sequence. When considered together, we specifically designed this experiment to avoid the possibility that rats would actively use their neural activity to control the maze.

Second, we had a similar question when collecting data for this manuscript and so we conducted a pilot experiment. We took 3 rats from experiment #1 (after its completion) and we required them to perform “forced-runs” over the course of 3-4 days, a task where rats navigate to a reward zone and are rewarded with a chocolate pellet. The trajectory on “forced-runs” is predetermined and rats were always rewarded for navigating along the predetermined route. Every trial was initiated by strong mPFC-hippocampal theta coherence. We were curious as to whether time-to-trial-onset would decrease if we repeatedly paired trial onset to strong mPFC-hippocampal theta coherence. 1 out of 3 rats (rat 21-35) showed a significant correlation between time-to-trial onset and trial number, indicating that our threshold for strong mPFC-hippocampal theta coherence was being met more quickly with experience (Figure R1A). When looking over sessions and rats, there was considerable variability in the magnitude of this correlation and sometimes even the direction (Figure R1B). As such, the degree to which rat 21-35 was aware of controlling the environment by reaching strong mPFC-hippocampal theta coherence is unclear, but this question requires future experimentation.

Author response image 1.

Strong mPFC-hippocampal theta coherence was used to control trial onset for the entirety of forced-navigation sessions. Time-to-trial onset is a measurement of how long it took for strong coherence to be met. A) Time-to-trial onset was averaged across sessions for each rat, then plotted as a function of trial number (within-session experience on the forced-runs task). Rat 21-35 showed a significant negative correlation between time-to-trial onset and trial number, indicating that time-to-coherence reduced with experience. The rest of the rats did not display this effect. B) Correlation between trial-onset and trial number (y-axis; see A) across sessions (x-axis). A majority of sessions showed a negative correlation between time-to-trial onset and trial number, like what was seen in (A), but the magnitude and sometimes direction of this effect varied considerably even within an animal.

Is there any evidence that rats display better performance on trials with random delays in which HPC-PFC coherence was naturally elevated?

This question is now addressed in Extended Figure 5 and discussed in the section titled “strong prefrontal-hippocampal theta coherence leads to correct choices on a spatial working memory task”.

The introduction frames this study as a test of the "communication through coherence" hypothesis. In its strongest form, this hypothesis states that oscillatory synchronization is a pre-requisite for inter-areal communication, i.e. if two areas are not synchronized, they cannot transfer information. Recent experimental evidence shows this relationship is more likely inverted-coherence is a consequence of inter-areal interactions, rather than a cause. See Schneider et al. (DOI: 10.1016/j.neuron.2021.09.037) and Vinck et al. (10.1016/j.neuron.2023.03.015) for a more in-depth explanation of this distinction. The authors should expand their treatment of this hypothesis in light of these findings.

Our introduction and discussions have sections dedicated to these studies now.

Figure 6 - It would be much more intuitive to use the labels "Rat 1", "Rat 2", and "Rat 3"; the "21-4X" identifiers are confusing.

This was corrected in the paper.

Figure 6C - The sub-plots within this figure are rather small and difficult to interpret. The figure would be easier to parse if the data were presented as a heatmap of the ratio of theta power during blue vs. red stim, with each pixel corresponding to one channel.

This suggestion was implemented in the paper. See Fig 6C. Extended Fig. 8 now shows the power spectra as a function of recording shank and channel.

Ext. Figure 2B - What happens during an acquisition failure? Instead of "Amount of LFP data," consider using "Buffer size".

Corrected.

Ext. Figure 2D-E - Instead of "Amount of data," consider using "Window size"

Referred to as buffer size.

Ext. Figure 2E - y-axis should extend down to 4 Hz. Are all of the last four values exactly at 8 Hz?

Yes. Values plateau at 8Hz. These data represent an average over ~50 samples.

Ext. Figure 2F - consider moving this before D/E, since those panels are summaries of panel F

Corrected.

Ext. Figure 4A - ANOVA tells you that accuracy is impacted by delay duration, but not what that impact is. A post-hoc test is required to show that long delays lead to lower accuracy than short ones. Alternatively, one could compute the correlation between delay duration and proportion correctly for each mouse, and look for significant negative values.

We included supplemental analyses in Extended Fig. 4

Reviewer #2 (Recommendations For The Authors):

The authors should replace terms that suggest a causal relationship between PFC-HPC synchrony and behavior, such as 'leads to', 'biases', and 'enhances' with more neutral terms.

Causal implications were toned down and wherever “leads” or “led” remains, we specifically mean in the context of coherence being detected prior to a choice being made.

The rationale for the analysis described in the paragraph starting on line 324, and how it fits with the preceding results, was not clear to me. The authors also write at the start of this paragraph "Given that mPFC-hippocampal theta coherence fluctuated in a periodical manner (Extended Fig. 5B)", but this figure only shows example data from 2 trials.

The reviewer is correct. While we point towards 3 examples in the manuscript now, we focused this section on the autocorrelation analysis, which did not support our observation as we noticed a rather linear decay in correlation over time. As such, the periodicity observed was almost certainly a consequence of overlapping data in the epochs used to calculate coherence rather than intrinsic periodicity.

Shortly after the start of the results section (line 112), the authors go into a very detailed description of how they validated their BMI without first describing what the BMI actually does. This made this and the subsequent paragraphs difficult to follow. I suggest the authors start with a general description of the BMI (and the general experiment) before going into the details.

Corrected. See first paragraph of “Development of a closed-loop…”.

In Figure 2C, as expected, around the onset of 'high' coherence trials, there is an increase in theta coherence but this appears to be very transient. However, it is unclear what the heatmap represents: is it a single trial, single session, an average across animals, or something else? In Figure 3F, however, the increase appears to be much more sustained.

The sample size was rats for every panel in this figure. This was clarified at the end of Fig. 3.

In Figure 2D, it was not clear to me what units of measurement are used when the averages and error bars are calculated. What is the 'n' here? Animals or sessions? This should be made clear in this figure as well as in other figures.

The sample size is rats. This is now clarified at the end of Fig 2.

Describing the study of Jones and Wilson (2005), the authors write: "While foundational, this study treated the dependent variable (choice accuracy) as independent to test the effect of choice outcome on task performance." (line 83) It was not clear to me what is meant by "dependent" and "independent" here. Explaining this more clearly might clarify how the authors' study goes beyond this and other previous studies.

The reviewer is correct. A discussion on independent/dependent variables in the context of rationale for our experiment was removed.

Reviewer #3 (Recommendations For The Authors):

As explained in the public review, my comments mainly concern the interpretation of the experimental paradigm and its link with previous findings. I think modifying these in order to target the specific advance allowed by the paradigm would really improve the match between the experimental and analytical data that is very solid and the author's conclusions.

Concerning the paradigm, I recommend that the authors focus more on their novel ability to clearly dissociate the functional role of theta coherence prior to the choice as opposed to induced by the choice. Currently, they explain by contrasting previous studies based on dependent variables whereas their approach uses an independent variable. I was a bit confused by this, particularly because the task variable is not really independent given that it's based on a brain-driven loop. Since theta coherence remains correlated with many other neurophysiological variables, the results cannot go beyond showing that leading up to the decision it correlates with good choice accuracy, without providing evidence that it is theta coherence itself that enhances this accuracy as they suggest in lines 93-94.

The reviewer is correct. A discussion on independent/dependent variables in the context of rationale for our experiment was removed.

Regarding previous results with muscimol inactivation, I recommend that the authors expand their discussion on this point. I think that their correlative data is not sufficient to conclude as they do that despite "these structures being deemed unnecessary" (based on causal muscimol experiments), they "can still contribute rather significantly" since their findings do not show a contribution, merely a correlation. This extra discussion could include possible explanations of the apparent, and thought-provoking discrepancies that they uncover such as: theta coherence may be a correlate of good accuracy without an underlying causal relation, theta coherence may always correlate with good accuracy but only be causally important in some tasks related to spatial working memory or, since muscimol experiments leave the brain time to adapt to the inactivation, redundancy between brain areas may mask their implication in the physiological context in certain tasks (see Goshen et al 2011).

The second paragraph of the discussion is now dedicated to this.

Possible further analysis :

• In Extended 4A the authors show that performance drops with delay duration. It would be very interesting to see this graph with the high coherence / low coherence / yoked trials to see if the theta coherence is most important for longer trials for example.

This is a great suggestion. Due to 10% of trials being triggered by high coherence states, our sample size precludes a robust analysis as suggested. Given that we found an enhancement effect on a task with minimal spatial working memory requirements (Fig. 4), it seems that coherence may be a general benefit or consequence of choice processes. Nonetheless, this remains an important question to address in a future study.

• Figure 6: The authors explain in the text that although the effect of stimulation of VMT is variable, overall VMT activation increased PFC-HPC coherence. I think in the figure the results are only shown for one rat and session per panel. It would be interesting to add a figure including their whole data set to show the overall effect as well as the variability.

The reviewer is correct and this comment promoted significant addition of detail to the manuscript. We have added an extended figure (Ext. Fig. 9) showing our VMT stimulation recording sessions. We originally did not include these because we were performing a parameter search to understanding if VMT stimulation could increase mPFC-hippocampal theta coherence. The results section was expanded accordingly.

Changes to writing / figures :

• The paper by Eliav et al, 2018 is cited to illustrate the universality of coupling between hippocampal rhythms and spikes whereas the main finding of this paper is that spikes lock to non-rhythmic LFP in the bat hippocampus. It seems inappropriate to cite this paper in the sentence on line 65.

We agree with the reviewer and this citation was removed.

• Line 180 when explaining the protocol, it would help comprehension if the authors clearly stated that "trial initiation" means opening the door to allow the rat to make its choice. I was initially unfamiliar with the paradigm and didn't figure this out immediately.

We added a description to the second paragraph of our first results section.

• Lines 324 and following: the analysis shows that there is a slow decay over around 2s of the theta coherence but not that it is periodical (as in regularly occurring in time), this would require the auto-correlation to show another bump at the timescale corresponding to the period of the signal. I recommend the authors use a different terminology.

This comment is now addressed above in our response to Reviewer #2.

• Lines 344: I am not sure why the stable theta coherence levels during the fixed delay phase show that the link with task performance is "through mechanisms specific to choice". Could the authors elaborate on this?

We elaborated on this point further at the end of “Trials initiated by strong prefrontal-hippocampal theta coherence are characterized by prominent prefrontal theta rhythms and heightened pre-choice prefrontal-hippocampal synchrony”

• Line 85: "independent to test the effect of choice outcome on task performance." I think there is a typo here and "choice outcome" should be "theta coherence".

The sentence was removed in the updated draft.

Author Response

Reviewer #2 (Public Review):

Activation of TEAD-dependent transcription by YAP/TAZ has been implicated in the development and progression of a significant number of malignancies. For example, loss of function mutations in NF2 or LATS1/2 (known upstream regulators that promote YAP phosphorylation and its retention and degradation in the cytoplasm) promote YAP nuclear entry and association with TEAD to drive oncogenic gene transcription and occurs in >70% of mesothelioma patients. High levels of nuclear YAP have also been reported for a number of other cancer cell types. As such, the YAP-TEAD complex represents a promising target for drug discovery and therapeutic intervention. Based on the recently reported essential functional role for TEAD palmitoylation at a conserved cysteine site, several groups have successfully targeted this site using both reversible binding non-covalent TEAD inhibitors (i.e., flufenamic acid (FA), MGH-CP1, compound 2 and VT101~107), as well as covalent TEAD inhibitors (i.e., TED-347, DC-TEADin02, and K-975), which have been demonstrated to inhibit YAP-TEAD function and display antitumor activity in cells and in vivo.

Here, Fan et al. disclose the development of covalent TEAD inhibitors and report on the therapeutic potential of this class of agents in the treatment of TEAD-YAP-driven cancers (e.g., malignant pleural mesothelioma (MPM)). Optimized derivatives of a previously reported flufenamic acid-based acrylamide electrophilic warhead-containing TEAD inhibitor (MYF-01-37, Kurppa et al. 2020 Cancer Cell), which display improved biochemical- and cell-based potency or mouse pharmacokinetic profiles (MYF-03-69 and MYP-03-176) are described and characterized.

Strengths:

All of the authors' claims and conclusions are very well supported and justified by the data that is provided. Clear improvements in biochemical- and cell-based potencies have been made within the compound series. Cell-based selective activities in the HIPPO pathway defective versus normal/control cell types are established. Transcriptional effects and the regulation of BMF proapoptotic mRNA levels are characterized. A 1.68 A X-Ray co-crystal structure of MYF-03-69 covalently bound to TEAD1 via Cys359 is provided. In vivo efficacy in a relevant xenograft is demonstrated, using a 30 mg/kg, BID PO dose.

We thank the reviewer for appreciating and highlighting the strengths of our study.

Weaknesses:

Beyond the impact on BMF gene regulation, new biological insights reported here for this compound series are moderate. Progress and differentiation with respect to activity and/or ADME PK profiles relative to the very closely related and previously described (Keneda et al. 2020 Am J Cancer Res 10:4399. PMID 33415007) acrylamide-based covalent TEAD inhibitor K-975 (identical 11 nM cell-based potencies when compared head-to-head and identical reported in vivo efficacy doses of 30 mg/kg) is not entirely clear. Demonstration of on-target in vivo activity is lacking (e.g., impact on BMF gene expression at the evaluated exposure levels).

We thank the reviewer’s question. We have compared mouse liver microsome stability and hepatocyte stability of K-975 and MYF-03-176 and found that K-975 is metabolically less stable.

Consistently, when NCI-H226 cells derived xenograft mice were dosed with 30 mg/kg K-975 twice daily, the tumors kept growing and reach more than 1.5-fold volume on 14th day. While with the same dosage, MYF-03-176 showed a significant tumor regression. K-975 did not reach such efficacy even with 100 or 300 mg/kg twice daily, either in NCI-H226 or MSTO-211H CDX mouse model according to the paper (Keneda et al. 2020 Am J Cancer Res 10:4399).

To demonstrate the on-target in vivo activity, we tested expression of the TEAD downstream genes and BMF in tumor sample after 3-day BID treatment (PD study) and we observed reduction of CTGF, CYR61, ANKRD1 and an increase of BMF, which indicates an on-target activity in vivo.

Author Response

Reviewer #2 (Public Review):

This paper by Angueyra, et al., adds to the field’s current understanding of photoreceptor specification and factors regulating opsin expression in vertebrates. Current models of specification of vertebrate photoreceptors are largely based on studies of mammals. However, a great number of animals including teleosts express a wider array of photoreceptor subtypes. Zebrafish for example have 4 distinct cone subtypes and rods. The approach is sound and the data are quite convincing. The only minor weaknesses are that the statistical analyses need to be revisited and the discussion should be a bit more focused.

To identify differentially expressed transcription factors, the authors performed bulk RNA-seq of pooled, hand-sorted photoreceptors. The selection criterion was tightly controlled to limit unhealthy cells and cellular debris from other photoreceptors subtypes. The pooling of cells provided a considerable depth of sequencing, orders of magnitude better than scSeq. The authors identified known transcription factors and several that appear to be novel or their role has not been determined. The data are made available on the PIs website as is a program to access and compare the gene expression data.

The authors then used CRISPR/Cas9 gene targeting of two known and several novel factors identified in their analysis for effects on cell fate decisions and opsin expression. Phenotyping performed on the injected larvae is possible, and the target genes were applied and sequenced to demonstrate the efficiency of the gene targeting. Targeting of 2 genes with know functions in photoreceptor specification in zebrafish, Tbx2b and Foxq2 resulted in the anticipated changes in cell fate, albeit, the strength of the alterations in cell fate in the F0 larvae appears to be less than the published phenotypes for the inherited alleles. Interestingly, the authors also identified the expression of an RH2 opsin in the SWS2 another cone type. The changes are subtle but important.

The authors then targeted tbx2a, the function of which was not known. The result is quite interesting as it matches the increase of rods and decrease of UV cones observed in tbx2b mutants. However, the injected animals also showed RH2 opsin expression but are now in the LWS cone subtype. These data suggest that Tbx2 transcription factors repress misexpression of opsins in the wrong cell type.

The authors also show that targeting additional differentially expressed factors does not affect photoreceptor fate or survival in the time frame investigated. These are important data to present. For these or any of the other targeted genes above, did the authors test for changes in photoreceptor number or survival?

We have attempted to address this point, but the answer is not clear cut. We used activated caspase-3 inmmunolabeling as a marker of apoptosis (Lusk and Kwan 2022). At 5 dpf, the age we chose to make quantifications, we don’t see an increase in activated caspase-3 positive cells when we compare control and tbx2a F0 mutants (Reviewer Figure 1A-B). Labeled cells are very rare and located near the ciliary marginal zone irrespective of genotype. This suggests that there is no detectable active death at this late stage of development in tbx2 F0 mutants. Earlier in development, at 3 dpf, when photoreceptor subtypes first appear, there is also a normal wave of apoptosis in the retina (Blume et al. 2020; Biehlmaier, Neuhauss, and Kohler 2001), resulting in many cells positive for activated caspase-3; our preliminary quantifications don’t show a marked increase in the number of labeled cells in tbx2a F0 mutants, but we consider that it’s likely that subtle effects might be obscured by the physiological wave of apoptosis (Reviewer Figure 1C-D).

Reviewer Figure 1 - Assessment of apoptosis in tbx2a F0 mutants. (A-B) Confocal images of 5 dpf larval eyes of control (A and A’) and tbx2a F0 mutants (B and B’) counterstained with DAPI (grey) and immunolabeled against activated Caspase 3 (yellow) show sparse and dim labeling, restricted to cells located in the ciliary marginal zone, without clear differences between groups. (C-D) Confocal images of 3 dpf larval eyes of control (C and C’) and tbx2a F0 mutants (D and D’) immunolabeled against activated Caspase 3 show many positive cells, located in all retinal layers, as expected from physiological apoptosis at this stage of development and without clear differences between groups.

Furthermore, the additional single-cell RNA-seq datasets we have reanalyzed suggest that tbx2a and tbx2b are expressed by other retinal neurons and progenitors and not just photoreceptors (Reviewer Figure 2), further confounding attempts at the quantification of apoptosis specifically in photoreceptor progenitors.

Reviewer Figure 2 – Expression of tbx2 paralogues across retinal cell types. The transcription factors tbx2a and tbx2b are expressed by many retinal cells. Plots show average counts across clusters in RNA-seq data obtained by Hoang et al. (2020).

At this stage, we consider that fully resolving this issue is important and will require considerably more work, which we will pursue in the future using full germline mutants and live-imaging experiments.

Reviewer #3 (Public Review):

Angueyra et al. tried to establish the method to identify key factors regulating fate decisions in the retinal visual photoreceptor cells by combining transcriptomic and fast genome editing approaches. First, they isolated and pooled five subtypes of photoreceptor cells from the transgenic lines in each of which a specific subtype of photoreceptor cells are labeled by fluorescence protein, and then subjected them to RNA-seq analyses. Second, by comparing the transcriptome data, they extracted the list of the transcription factor genes enriched in the pooled samples. Third, they applied CRISPR-based F0 knockout to functionally identify transcription factor genes involved in cell fate decisions of photoreceptor subtypes. To benchmark this approach, they initially targeted foxq2 and nr2e3 genes, which have been previously shown to regulate S-opsin expression and S-cone cell fate (foxq2) and to regulate rhodopsin expression and rod fate (nr2e3). They then targeted other transcription factor genes in the candidate list and found that tbx2a and tbx2b are independently required for UV-cone specification. They also found that tbx2a expressed in the L-cone subtype and tbx2b expressed in L-cones inhibit M-opsin gene expression in the respective cone subtypes. From these data, the authors concluded that the transcription factors Tbx2a and Tbx2b play a central role in controlling the identity of all photoreceptor subtypes within the retina.

Overall, the contents of this manuscript are well organized and technically sound. The authors presented convincing data, and carefully analyzed and interpreted them. It includes an evaluation of the presented data on cell-type specific transcriptome by comparing it with previously published ones. I think the current transcriptomic data will be a valuable platform to identify the genes regulating cell-type specific functions, especially in combination with the fast CRISPR-based in vivo screening methods provided here. I hope that the following points would be helpful for the authors to improve the manuscript appropriately.

1) The manuscript uses the word “FØ” quite often without any proper definition. I wonder how “Ø” should be pronounced - zero or phi? This word is not common and has not been used in previous publications. I feel the phrase “F0 knockout,” which was used in the paper cited by the authors (Kroll et al 2021), is more straightforward. If it is to be used in the manuscript, please define “FØ” and “CRISPR-FØ screening” appropriately, especially in the abstract.

We have made changes to replace “FØ” to “F0.” In our other citation (Hoshijima et al., 2019), “F0 embryo” was used throughout the paper. Following our references and Dr Kojima’s suggestion, we adopted “F0 mutant larva” as the most straightforward and less confusing term. We have also made changes in the abstract to define our approach more clearly and made appropriate changes throughout the manuscript.

2) Figure 1-supplement 1 shows that opn1mw4 has quite high (normalized) FPKM in one of the S-cone samples in contrast to the least (or no) expression in the M-cone samples, in which opn1mw4 is expected to be detected. The authors should address a possible origin of this inconsistent result for opn1mw4 expression as well as a technical limitation of using the Tg(opn1mw2:egfp) line for detection of opn1mw4 expression in the GFP-positive cells.

In Figure 1 - Supplement 1, we had attempted to provide a summarized figure of all phototransduction genes, but the big differences in expression levels — in particular, the high expression of opsins genes — forced us to use gene-by-gene normalization for display. Without normalization, the expression of opn1mw4 is very low across all samples, and its detection in that sole S-cone sample can likely be attributed to some degree of inherent noise in our methods. We have revised Figure 1 - Supplement 1: we find that we can avoid gene-by-gene normalization and still provide a good summary of the expression of phototransduction genes if the heatmap is broken down by gene families, which have more similar expression levels. In addition, we have added caveats to the use of the Tg(opn1mw2:egfp) line as our sole M-cone marker in the results section describing our RNA-seq approach, including our inability to provide data on Opn1mw4-expressing M cones.

3) The manuscript lacks a description of the sampling time point. It is well known that many genes are expressed with daily (or circadian) fluctuation (cf. Doherty & Kay, 2010 Annu. Rev. Genet.). For example, the cone-specific gene list in Fig.2C includes a circadian clock gene, per3, whose expression was reported to fluctuate in a circadian manner in many tissues of zebrafish including the retina (Kaneko et al. 2006 PNAS). It appears to be cone-specific at this time point of sample collection as shown in Fig.2, but might be expressed in a different pattern at other time points (eg, rod expression). The authors should add, at least, a clear description of the sampling time points so as to make their data more informative.

We have included this information in the materials and methods. We collected all our samples during the most active peak of the zebrafish circadian rhythm between 11am and 2pm (3h to 6h after light onset) to avoid the influence of circadian fluctuations in our analysis.

Author Response

Reviewer #1 (Public Review):

The authors use a newly developed object-space memory task comprising of a "Stable" version and "Overlapping" version where two objects are presented in two locations per trial in a square open field. Each version consists of 5 training trials of 5-min presentations of an object-space configuration, with both object locations staying constant across training trials in the Stable condition, and only one object location staying fixed in the Overlapping condition. Memory is tested in a test trial 24 hours later where the opposite configuration is presented - overlapping configuration presented for the Stable condition and stable configuration presented for the Overlapping condition - with the thesis that memory in this test trial for the Overlapping condition will depend on the accumulated memory of spatial patterns over the training trials, whereas memory for the test trial in the Stable condition can be due to episodic memory of last trial or accumulated memory. Memory is quantified using a Discrimination Index (DI), comparing the amount of time animals spend exploring the two object locations.

Here, animals in other groups are also presented with an interference trial equivalent to the test trial, to test if the memory of the Overlapping condition can be disrupted. The behavioral data show that for RGS14 over-expressing animals, memory in the Overlapping condition is diminished compared to controls with no interference or controls where over-expression is inhibited, whereas memory in the Stable condition is enhanced. This is interpreted as interference in semantic-like memory formation, whereas one-shot episodic memory is improved. The authors speculate that increased cortical plasticity should lead to increased and larger delta waves according to the sleep homeostasis hypothesis, and observe that instead increased cortical plasticity leads to less non-REM sleep and smaller delta waves, with more prefrontal neurons with slower firing rates (presumably more plastic neurons). They further report increased hippocampal-cortical theta coherence during task and REM sleep, increased NonREM oscillatory coupling, and changes in hippocampal ripples in RGS14 over-expressing animals.

While these results are interesting, there are several issues that need to be addressed, and the link between physiology and behavioral results is unclear.

1) The behavioral results rely on the interpretation that the Overlapping condition corresponds to semantic-like memory and the Stable condition corresponds to episodic-like memory. While the dissociation in memory performance due to interference seen in these two conditions is intriguing, the Stable condition can correspond not just to the memory of the previous trial but also accumulated memory of a stable spatial pattern over the 5 testing trials, similar to accumulated memory of a changing spatial pattern in the Overlapping pattern.

Yes! We completely agree on this. We do not claim the stable condition corresponds to episodic-like memory, instead we refer to it as simple memory, since it can be solved either way (one trial memory or cumulative memory). We now expanded this in the discussion to make it clearer.

Here, it is puzzling that in the behavioral control with no interference (Figure 1D), memory in the Stable and Overlapping condition is unchanged in the test trial, with the DI statistically at 0 in the test trial. In the original description of the Object Space task by the authors in the referenced paper, the measure of memory was a Discrimination Index significantly higher than 0 in both the Stable and Overlapping conditions. This discrepancy needs to be reconciled. Is the DI for the interference trial shown in Fig. S1 significantly different than 0? No statistics or description is provided in the figure legend here.

As mentioned above, we apologize that we oversimplified the description. The 24h interference trial would be what corresponds to the original test trial. We added a clarifying figure for comparison in S1 (bar graph in addition to the violin plot) and stats. Performance was for all groups and conditions above chance, replicating our previous results.

2) The physiology experiments compare Home cage (HC) conditions to the Object Space task (OS) throughout the manuscript. While some differences are seen in the control and RGS14 over-expressing animals, there is no comparison of the Stable vs. Overlapping condition in the physiology experiments. This precludes making explicit links between physiological observations and behavioral effects.

As also mentioned above, we have now added analysis exploring the detailed OS conditions. We would like to thank the reviewers for giving us the opportunity of doing so.

3) The authors speculate that learning will result in larger and more delta waves as per the synaptic homeostasis hypothesis. It should be noted here that an alternative hypothesis is that there should also be a selective increase in synaptic plasticity for learning and consolidation. The authors do observe that control animals show more frequent and higher-amplitude delta waves, but rather than enhancing this process, RGS14 animals with increased plasticity show the opposite effect. How can this be reconciled and linked with the behavioral data in the Stable and Overlapping condition?

In the context of the Object Space Task, we would expect all behavioural conditions (Stable and Overlapping) to induce synaptic changes since learning does occur also in the Stable condition (see also performance on 24h trial). Thus, especially homeostatic responses such as increase in delta amplitude, we would expect for all experiences independent if subtle statistical rules are presented or not. In contrast, detailed processing, extracting underlying regularities is rather proposed by the Sleep for Active Systems Consolidation Hypothesis to occur during hippocampal-cortical interactions in form of delta/ripple/spindle interactions (with different theories emphasising different types of interactions). As mentioned above, we now add a more specific analysis in this regards, where we can show that the two OS conditions that involve moving objects (where thus potentially statistical regularities can be extracted) show a higher percentage of ripples occurring after large slow oscillations in comparison to home cage or the simple learning condition Stable. In contrast, RGS14 already has higher participation in both control conditions, emphasising that in these animals all experiences are treated by the brain as significant learning condition, explaining the behavioural effect (increased interference due to better memory for the interference). Further, we expanded in the discussion how in RGS we sometimes see an enhancement of learning effects but sometimes see a more complex interaction of what we would expect from physiological learning.

Similarly, there is an increase in slower-firing neurons in RGS14 over-expressing animals. Slower-firing neurons have been proposed to be more plastic in the hippocampus based on their participation in learned hippocampal sequences, but appropriate references or data are needed to support the assertion that slower-firing neurons in the prefrontal cortex are more plastic.

As described above, we have expanded the discussion including other citations that also consider the cortex. We can show that our changes would be expected if one turns the cortex as plastic as the hippocampus.

4) It is noted that changing cortical plasticity influences hippocampal-cortical coupling and hippocampal ripples, suggesting a cortical influence on hippocampal physiological patterns. It has been previously shown that disrupting prefrontal cortical activity does alter hippocampal ripples and hippocampal theta sequences (Schmidt et al., 2019; Schmidt and Redish, 2021). The current results should be discussed in this context.

We would like to thank the reviewer for these suggestions, they are now incorporated in the manuscript.

Reviewer #2 (Public Review):

In this paper, the authors provide evidence to support the longstanding proposition that a dual-learning system/systems-level consolidation (hippocampus attains memories at a fast pace which are eventually transmitted to the slow-learning neocortex) allows rapid acquisition of new memories while protecting pre-existing memories. The authors leverage many techniques (behavior, pharmacology, electrophysiology, modelling) and report a host of behavioral and electrophysiological changes on induction of increased medial prefrontal cortex (mPFC) plasticity which are interesting and will be of significant interest to the broad readership.

The experimental design and analyses are convincing (barring some instances which are discussed below). The following recommendations will bolster the strength/quality of the manuscript:

1) Certain concerns regarding the interpretation and analysis of the behavioral data remain. The authors need to clarify if increased mPFC plasticity leads to only an increase in one-shot memory or 'also' interference of previous information. It seems that the behavioral results could also be explained by the more parsimonious explanation that one-shot memory is improved. Do the current controls tease apart these two scenarios?

We agree we cannot disentangle if one memory is just stronger than the other or if its an overwriting effect. We added this now to the discussion. Of note, we do not think it actually would be possible to distinguish these two effects behaviourally in rodents, or at least we cannot think of a fitting study design that would enable the contrast.

Additionally, the authors need to clarify why the 'no trial' and 'anisomycin' controls for the stable task perform at chance levels on exposure to a new object-place association on test day (Fig 1D).

Violin plots are sometimes hard to see. Here simple bar plots where you can see that the animals are not at chance at the 72h test in the control conditions.

Finally, further description of how the discrimination index (exploration time of novel-exploration time of familiar/sum of both) is recommended i.e., in the stable condition, which 'object' is chosen as 'novel' (as both are in the same locations) for computing the index (Fig 1). Do negative DI values imply a neophobia to novel objects (and thus are a form of memory; this is also crucial because the modelling results (Fig 1E) use both neophilia and neophobia while negative discrimination indexes are considered similar to 0 for interpreting the behavioral results, as stated on page 3, lines 84-86?

We added this now to the methods (For Overlapping it is moved location – stable location, for Stable it is location-to-be-moved-at-test – stable location and for random which is assigned as moved and stable is random, and then for each divided by total time). We agree that neophilia/neophobia (especially changes in the distribution) can be an issue and have discussed it in detail in Schut et al NLM 2020 where we see difference in absolute beta values (thus controlling for philia/phobia differences). We also discuss there why it is difficult to control for this in the DI in more detail. In short, one could use absolute values but then it is difficult to determine what a group chance-level would look like. However, luckily here there is not issue since we did not observe difference in neophilic or phobic tendencies while running the experiments. Critically the interference trial (that can also function as simple test trial) confirms that as a group animals show positive DI and neophilia.

2) The authors report lower firing rates in RGS14414 animals during the task in Fig 2F. It is indeed remarkable how large the reported differences are. The authors need to rule out any differences in the behavioral state of the animals in the two groups during the task, i.e., rest vs. active exploration/movement dynamics. Are only epochs during the task while the animals interact with the objects used for computing the firing rates (same epochs as Fig 1)? If not, doing so will provide a useful comparison with Fig 1. Additionally, although the authors make the case for slow firing rate neurons being important for plasticity (based on Grosmark and Buzsaki, 2016), it is crucial to note that the firing rate dynamic (slow vs. fast) in that study for the hippocampus is defined based on the whole recorded session (predominated by sleep), indeed the firing rates of the two groups (slow vs. fast/plastic vs. rigid) during the task/maze-running do not differ in that study. Therefore, the results here seem incongruent with the Grosmark and Buzsaki paper. Since this finding is central to the main claim of the authors, it either warrants further investigation or a re-interpretation of their results.

As mentioned in the main points, we now added the firing rate analysis (including new groups splits) for wake in the sleep box, NREM and REM separately. Each time the same results are obtained. Currently, we do not yet have the tracking and video synchronization set-up, therefore we cannot split the task for specific behaviours.

However, we now also cite Buzsaki’s original log-normal brain review, where he first proposed the idea. There he also shows same effects as we do, in that the general firing rate distribution is the same for task and different sleep stages, just overall shifted. The analysis from Grosmark included more strigent subselection of neurons to be able to also argue that incorporation into run/replay-sequences could not have been biased by firing rate per se (instead of plasticity). However, the original proposition from Buzsaki does fit to our results. He further presents hippocampus vs cortex firing rates, which also confirm the idea (hippocampus more plastic and has slower firing rates). We included this figure above in the general comments. Further, we now expanded the discussion in this point.

3) A concern remains as to how many of the electrophysiological changes they observe (firing rate differences, LFP differences including coupling, sleep state differences, Figs. 2-4) support their main hypothesis or are a by-product of injection of RGS14414 (for instance, one might argue that an increased 'capability' to learn new information/more plasticity might lead to more NREM sleep for consolidation, etc.). The authors need to carefully interpret all their data in light of their main hypothesis, which will substantially improve the quality/strength of the manuscript.

We now expanded the discussion, included more structure and also include that we cannot disentangle if the cellular changes or sleep oscillation changes or an interaction of both is the cause of the result. Furthermore, we added that we cannot distinguish if the interference memory is stronger or actually overwrites the original training memory.

Reviewer #3 (Public Review):

The authors set out to test the idea that memories involve a fast process (for the acquisition of new information) and a slow process (where these memories are progressively transferred/integrated into more-long term storage). The former process involves the hippocampus and the latter the cerebral cortex. This 'dual-learning' system theoretically allows for new learning without causing interference in the consolidation of older memories. They test this idea by artificially increasing plasticity in the pre-limbic cortex and measuring changes in different learning/memory tasks. They also examined electrophysiological changes in sleep, as sleep is linked to memory formation and synaptic plasticity.

The strengths of the study include a) meticulous analyses of a variety of electrophysiological measurements b) a combination of neurobiological and computational tools c) a largely comprehensive analysis of sleep-based changes. Some weaknesses include questions about the technique for increasing cortical plasticity (is this physiological?) and the absence of some additional experiments that would strengthen the conclusions. However, overall, the findings appear to support the general idea under examination.

This study is likely to be very impactful as it provides some really new information about these important neural processes, as well as data that challenges popular ideas about sleep and synaptic plasticity.

We would like to thank the reviewer for these positive comments. Answers to the weaknesses are presented below in the recommendations for the authors.

Author Response

Reviewer 1 (Public Review):

To me, the strengths of the paper are predominantly in the experimental work, there's a huge amount of data generated through mutagenesis, screening, and DMS. This is likely to constitute a valuable dataset for future work.

We are grateful to the reviewer for their generous comment.

Scientifically, I think what is perhaps missing, and I don't want this to be misconstrued as a request for additional work, is a deeper analysis of the structural and dynamic molecular basis for the observations. In some ways, the ML is used to replace this and I think it doesn't do as good a job. It is clear for example that there are common mechanisms underpinning the allostery between these proteins, but they are left hanging to some degree. It should be possible to work out what these are with further biophysical analysis…. Actually testing that hypothesis experimentally/computationally would be nice (rather than relying on inference from ML).

We agree with the reviewer that this study should motivate a deeper biophysical analysis of molecular mechanisms. However, in our view, the ML portion of our work was not intended as a replacement for mechanistic analysis, nor could it serve as one. We treated ML as a hypothesis-generating tool. We hypothesized that distant homologs are likely to have similar allosteric mechanisms which may not be evident from visual analysis of DMS maps. We used ML to (a) extract underlying similarities between homologs (b) make cross predictions across homologs. In fact, the chief conclusion of our work is that while common patterns exist across homologs, the molecular details differ. ML provides tantalizing evidence to this effect. The conclusive evidence will require, as the reviewer rightly suggests, detailed experimental or molecular dynamics characterization. Along this line, we note that we have recently reported our atomistic MD analysis of allostery hotspots in TetR (JACS, 2022, 144, 10870). See ref. 41.

Changes to manuscript:<br /> “Detailed biophysical or molecular dynamics characterization will be required to further validate our conclusions(38).”

Reviewer 3 (Public Review):

However - at least in the manuscript's present form - the paper suffers from key conceptual difficulties and a lack of rigor in data analysis that substantially limits one's confidence in the authors' interpretations.

We hope the responses below address and allay the reviewer’s concerns.

A key conceptual challenge shaping the interpretation of this work lies in the definition of allostery, and allosteric hotspot. The authors define allosteric mutations as those that abrogate the response of a given aTF to a small molecule effector (inducer). Thus, the results focus on mutations that are "allosterically dead". However, this assay would seem to miss other types of allosteric mutations: for example, mutations that enhance the allosteric response to ligand would not be captured, and neither would mutations that more subtly tune the dynamic range between uninduced ("off) and induced ("on") states (without wholesale breaking the observed allostery). Prior work has even indicated the presence of TetR mutations that reverse the activity of the effector, causing it to act as a co-repressor rather than an inducer (Scholz et al (2004) PMID: 15255892). Because the work focuses only on allosterically dead mutations, it is unclear how the outcome of the experiments would change if a broader (and in our view more complete) definition of allostery were considered.

We agree with the reviewer that mutations that impact allostery manifest in many different ways. Furthermore, the effect size of these mutations runs the full gamut from subtle changes in dynamic range to drastic reversal of function. To unpack allostery further, allostery of aTF can be described, not just by the dynamic range, but by the actual basal and induced expression levels of the reporter, EC50 and Hill coefficient. Given the systemic nature of allostery, a substantial fraction of aTF mutations may have some subtle impact on one or more of these metrics. To take the reviewer’s argument one step further, one would have to accurately quantify the effect size of every single amino acid mutation on all the above properties to have a comprehensive sequence-function landscape of allostery. Needless to say, this is extremely hard! Resolution of small effect sizes is very difficult, even at high sequencing depth. To the best of our knowledge, a heroic effort approaching such comprehensive analysis has been accomplished so far only once (PMID: 3491352).

Our focus, therefore, was to screen for the strongest phenotypic impact on allostery i.e., loss of function. Mutations leading to loss of function can be relatively easily identified by cell-sorting. Because our goal was to compare hotspots across homologs, we surmised that loss of function mutations, given their strong phenotypic impact, are likely to provide the clearest evidence of whether allosteric hotspots are conserved across remote homologs.

The reviewer raised the point of activity-reversing mutations. Yes, there are activity reversing mutations in TetR. However, they represent an insignificant fraction. In the paper cited by the reviewer, there are 15 activity-reversing mutations among 4000 screened. Furthermore, the paper shows that activity-reversing in TetR requires two-tofour mutations, while our library is exclusively single amino acid substitutions. For these reasons, we did not screen for activity-reversing mutations. Nonetheless, we agree with the reviewer that screening for activity-reversing mutations across homologs would be very interesting.

The separation in fluorescence between the uninduced and induced states (the assay dynamic range, or fold induction) varies substantially amongst the four aTF homologs. Most concerningly, the fluorescence distributions for the uninduced and induced populations of the RolR single mutant library overlap almost completely (Figure 1, supplement 1), making it unclear if the authors can truly detect meaningful variation in regulation for this homolog.

Yes, the reviewer is correct that the fold induction ratio varies among the four aTF homologs. However, we note that such differences are common among natural aTFs. Depending on the native downstream gene regulated by the aTF, some aTFs show higher ligand-induced activation, and others are lower. While this is not a hard and fast rule, aTFs that regulate efflux pumps tend to have higher fold induction than those that regulate metabolic enzymes. In summary, the variation in fold induction among the four aTFs is not a flaw in experimental design nor indicates experimental inconsistency but is instead just an inherent property of protein-DNA interaction strength and the allosteric response of each aTF.

Among the four aTFs, wildtype RolR has the weakest fold induction (15-fold) which makes sorting the RolR library particularly challenging. To minimize false positives as much as possible, we require that dead mutant be present in (a) non-fluorescent cells after ligandinduction (b) non-fluorescent cells before ligand-induction (c) at least two out of the three replicates for both sorts. Additionally, for RolR specifically, we adjusted the nonfluorescent gate to be far more stringent than the other three aTFs (Fig. 1 – figure supplement 1). Furthermore, we assign residues as allosteric hotspots, not individual dead mutations. This buffers against false strong signals from stray individual dead mutations. Finally, the top interquartile range winnows them to residues showing strong consistent dead phenotype. As a result of these “safeguards” we have built in, the number of allosteric hotspots of RolR (57) is comparable to the other three aTFs (51, 53 and 48). This suggests that we are not overestimating the number of hotspots despite the weaker fold induction of RolR. We highlight in a new supplementary figure (Figure 1 – figure supplement 4) that changing the read count threshold from 5X to 10X produces near identical patterns of mutations suggesting that our results are also robust to changes in ready depth stringency.

Changes to manuscript: In response to the reviewer's comment, we have added the following sentence.

“We note that the lower fold induction (dynamic range) of RolR makes it particularly challenging to separate the dead variants from the rest.”

The methods state that "variants with at least 5 reads in both the presence and absence of ligand in at least two replicates were identified as dead". However, the use of a single threshold (5 reads) to define allosterically dead mutations across all mutations in all four homologs overlooks several important factors:

Depending on the starting number of reads for a given mutation in the population (which may differ in orders of magnitude), the observation of 5 reads in the gated nonfluorescent region might be highly significant, or not significant at all. Often this is handled by considering a relative enrichment (say in the induced vs uninduced population) rather than a flat threshold across all variants.

We regret the lack of clarity in our presentation. We wish to better explain the rationale behind our approach. First, we understand the reviewer’s point on considering relative enrichment to define a threshold. This approach works well in DMS experiments involving genetic selections, which is commonly the case, because activity scales well with selection stringency. One can then pick enrichment/depletion relative to the middle of the read count distribution as a measure of gain or loss of function.

Second, this strategy does not, in practice, work well for cell-sorting screens. While it may be tempting to think of cell sorting as comparably activity-scaled as genetic selections, in reality, the fidelity of fluorescent-activated cell sorters is much lower. Making quantitative claims of activity based on cell sorting enrichment can be risky. It is wiser to treat cell sorting results as yes/no binary i.e., does the mutation disrupt allostery or not. More importantly, the yes/no binary classification suffices for our need to identify if a certain mutation adversely impacts allosteric activity or not.

Third, the above argument does not imply that all mutations have the same effect size on allostery. They don’t. We capture the effect size on individual residues, not individual mutations, by counting the number of dead mutations at a residue position. This is an important consideration because it safeguards us from minor inconsistencies that inevitably arise from cell sorting.

Fourth, a variant to be classified as allosterically dead, it must be present both in uninduced and induced DNA-bound populations in at least two out of three replicates (four conditions total). This is a stringent criterion for selecting dead variants resulting in highly consistent regions of importance in the protein even upon varying read count thresholds. To the extent possible, we have minimized the possibility of false positive bleed-through.

Finally, two separate normalizations were performed on the total sequence reads to be able to draw a common read count threshold 1) between experimental conditions & replicates and 2) across proteins. First, total sequencing reads were normalized to 200k total across all sample conditions (presorted, -inducer, and +inducer) and replicates for each homolog, allowing comparisons within a single protein. Next, reads were normalized again to account for differences in the theoretical size of each protein’s single-mutant library, allowing for comparisons across proteins by drawing a commont readcount cutoff. For example, total sequencing reads of RolR (4,332 possible mutants) increased by 1.18x relative to MphR (3,667 possible mutants) for a total of 236k reads.

Changes to manuscript: We have provided substantial additional details in the Fluorescence-activated cell sorting and NGS preparation and analysis sections.

We also added the following in the main text.

“In other words, we use cell sorting as a binary classifier i.e., does the mutation disrupt allostery or not. We capture the effect size on individual residues, not individual mutations, by counting the number of dead mutations at a residue position. This is an important consideration because it safeguards us from minor inconsistencies that inevitably arise from cell sorting.”

Depending on the noise in the data (as captured in the nucleotide-specific q-scores) and the number of nucleotides changed relative to the WT (anywhere between 1-3 for a given amino acid mutation) one might have more or less chance of observing five reads for a given mutation simply due to sequencing noise.

All the reads considered in our analyses pass the Illumina quality threshold of Q-score ≥ 30 which as per Illumina represent “perfect reads with no errors or ambiguities”. This translates into a probability of 1 in 1000 incorrect base call or 99.9% base call accuracy.

We use chip-based oligonucleotides to build our DMS library, which allows us to prespecify the exact codon that encodes a point mutation. This means the nucleotide count and protein count are the same. The scenario referred to by the reviewer i.e., “anywhere between 1-3 for a given amino acid mutation” only applies to codon randomized or errorprone PCR library generation. We regret if the chip-based library assembly part was unclear.

Depending on the shape and separation of the induced (fluorescent) and uninduced (non-fluorescent) population distributions, one might have more or less chance of observing five reads by chance in the gated non-fluorescent region. The current single threshold does not account for variation in the dynamic range of the assay across homologs.

We have addressed the concern raised by the reviewer on fluorescent population distributions in answers to questions 10 and 11.

The reviewer makes an important point about the choice of sequencing threshold. We use the sequencing threshold to simply make a binary choice for whether a certain variant exists in the sorted population or not. We do not use the sequencing reads as to scale the activity of the variant. To address the reviewer's comment, we have included a new supplementary figure (Fig 1 – figure supplement 4) where we compare the data by adjust the threshold two levels – 5 and 10 reads. As is evident in the new figure, the fundamental pattern of allosteric hotspots and the overall data interpretation does not change.

TetR: 5x – 53 hotspots, 10x – 51 hotspots

TtgR: 5x – 51 hotspots, 10x – 51 hotspots

MphR: 5x – 48 hotspots, 10x – 48 hotspots

RolR: 5x – 57 hotspots, 10x – 60 hotspots

In other words, changing the threshold to be more or less strict may have a modest impact on the overall number of hotspots in the dataset. Still, the regions of functional importance are consistent across different thresholds. We have expanded the discussion in the manuscript to address this point.

Changes to manuscript: We have now included a new supplementary comparing hotspot data at two thresholds: Figure 1 – figure supplement 4.

We also added the following in the main text.

“To assess the robustness of our classification of hotspots, we determined the number of hotspots at two different sequencing thresholds – 5x and 10x. At 5x and 10x, the number of hotspots are – TetR: 53, 51; TtgR: 51, 51; MphR: 48, 48 and RolR: 57,60, respectively. Changing the threshold has a modest impact on the overall number of hotspots and the regions of functional importance are consistent at both thresholds”

The authors provide a brief written description of the "weighted score" used to define allosteric hotspots (see y-axis for figure 1B), but without an equation, it is not clear what was calculated. Nonetheless, understanding this weighted score seems central to their definition of allosteric hotspots.

We regret the lack of clarity in our presentation. The weighted score was used to quantify the “deadness” of every residue position in the protein. At each position in the protein, the number of mutations that inhibited activity was summed up and the ‘deadness’ of each mutation was weighted based on how many replicates is appeared to inactivate the protein. Weighted score at each residue position is given by

Where at position x in the protein, D1 is the number of mutations dead in one replicate only, D2 is the number of mutations dead in 2 replicates, D3 is the number of mutations dead in 3 replicates, and Total is the total number of variants present in the data set (based on sequencing data). Any dead mutation that is seen in only one replicate is discarded and does not contribute to the “deadness” of the residue. Mutations seen in two and three replicates contribute to the score. We have included a new supplementary figure (Fig. 1 – figure supplement 2) to give the reader a detailed heatmap of all mutations and their impact for each protein.

Changes to manuscript: The weighted scoring scheme is now described in greater detail under Materials and Methods in the “NGS preparation and analysis” section.

The authors do not provide some of the standard "controls" often used to assess deep mutational scanning data. For example, one might expect that synonymous mutations are not categorized as allosterically dead using their methods (because they should still respond to ligand) and that most nonsense mutations are also not allosterically dead (because they should no longer repress GFP under either condition). In general, it is not clear how the authors validated the assay/confirmed that it is giving the expected results.

As we state in response to question 12, we use chip-based oligonucleotides to build our DMS library, which allows us to pre-specify the exact codon that encodes a point mutation. We have no synonymous or nonsense mutations in our DMS library. Each protein mutation is encoded by a single unique codon. The only stop codon is at 3’end of the gene.

The authors performed three replicates of the experiment, but reproducibility across replicates and noise in the assay is not presented/discussed.

Changes to manuscript: A new supplementary table (Table 1) is now provided with the pairwise correlation coefficients between all replicates for each protein.

In the analysis of long-range interactions, the authors assert that "hotspot interactions are more likely to be long-range than those of non-hotspots", but this was not accompanied by a statistical test (Figure 2 - figure supplement 1).

In response to the reviewer's comment, we now include a paired t-test comparing nonhotspots and hotspots with long-range interactions in the main text.

Changes to manuscript: In all four aTFs, hotspots constituted a higher fraction of LRIs than non-hotspots (Figure 2 – figure supplement 1; P = 0.07).

Author Response

Reviewer #1 (Public Review):

In this study, the authors describe an elegant genetic screen for mutants that suppress defects of MCT1 deletions which are deficient in mitochondrial fatty acid synthesis. This screen identified many genes, including that for Sit4. In addition, genes for retrograde signaling factors (Rtg1, Rtg2 and Rtg3), proteins influencing proteasomal degradation (Rpn4, Ubc4) or ribosomal proteins (Rps17A, Rps29A) were found. From this mix of components, the authors selected Sit4 for further analysis. In the first part of the study, they analyzed the effect of Sit4 in context of MCT1 mutant suppression. This more specific part is very detailed and thorough, the experiments are well controlled and convincing. The second, more general part of the study focused on the effect of Sit4 on the level of the mitochondrial membrane potential. This part is of high general interest, but less well developed. Nevertheless, this study is very interesting as it shows for the first time that phosphate export from mitochondrial is of general relevance for the membrane potential even in wild type cells (as long as they live from fermentation), that the Sit4 phosphatase is critical for this process and that the modulation of Sit4 activity influences processes relying on the membrane potential, such as the import of proteins into mitochondria. However, some aspects should be further clarified.

1) It is not clear whether Sit4 is only relevant under fermentative conditions. Does Sit4 also influence the membrane potential in respiring cells? Fig. S2D shows the membrane potential in glucose and raffinose. Both carbon sources lead to fermentative growths. The authors should also test whether Sit4 levels influence the membrane potential when cells are grown under respirative conditions, such in ethanol, lactate or glycerol. Even if deletions of Sit4 affect respiration, mutants with altered activity can be easily analyzed.

sit4Δ cells fail to grow on nonfermentable media as shown by us (Figure 2—figure supplement 1C) and others (Arndt et al., 1989; Dimmer et al., 2002; Jablonka et al., 2006). In our opinion, the exact reason is unclear, but there is an interesting observation that addition of aspartate can partially restore growth on ethanol (Jablonka et al., 2006). Despite the lack of thorough investigation on this sit4Δ defect, an early study speculated that this defect could be related to the cAMP-PKA pathway (Sutton et al., 1991). This study pointed out genetic interactions of SIT4 with multiple genes in cAMP-PKA (Sutton et al., 1991). In addition, sit4Δ cells have similar phenotypes as those cAMP-PKA null mutants, such as glycogen accumulation, caffeine resistant, and failure to grow on nonfermentable media (Sutton et al., 1991). We have not found sit4Δ mutants that could grow on nonfermentable media based on literature search.

2) The authors should give a name to the pathway shown in Fig. 4D. This would make it easier to follow the text in the results and the discussion. This pathway was proposed and characterized in the 90s by George Clark-Walker and others, but never carefully studied on a mechanistic level. Even if the flux through this pathway cannot be measured in this study, the regulatory role of Sit4 for this process is the most important aspect of this manuscript.

We now refer this mechanism as the mitochondrial ATP hydrolysis pathway.

3) To further support their hypothesis, the authors should show that deletion of Pic1 or Atp1 wipes out the effect of a Sit4 deletion. In these petite-negative mutants, the phosphate export cycle cannot be carried out and thus, Sit4, should have no effect.

The mitochondrial phosphate transport activity is electroneutral as it also pumps a proton together with inorganic phosphate. The F1 subunit of the ATP synthase (Atp1 and Atp2) is suggested among many literatures to be responsible for the ATP hydrolysis. We performed tetrad dissection to generate atp1Δ or atp2Δ in pho85Δ background. After streaking the single colony to a fresh plate, we noticed that atp1Δ mct1Δ and atp2Δ mct1Δ cells are lethal, and knocking out PHO85 rescued this synthetic lethality. It is not surprising that atp1Δ mct1Δ or atp2Δ mct1 Δ cells are lethal since the F1 subunit is important to generate a minimum of MMP in mct1 Δ cells when the ETC is absent (i.e., rho0 cells). However, knocking out PHO85 can generate MMP independent of F1 subunit of ATP synthase, which is suggested by the viable atp1Δ mct1Δ pho85Δ and atp2Δ mct1Δ pho85Δ cells. There are many ATPases in the mitochondrial matrix that could hydrolyze ATP for ADP/ATP carrier to generate MMP theoretically. However, we do not currently know exactly which ATPase(s) is activated by phosphate starvation. This data is now included as Figure 5—figure supplement 1F-G.

4) What is the relevance of Sit4 for the Hap complex which regulates OXPHOS gene expression in yeast? The supplemental table suggests that Hap4 is strongly influenced by Sit4. Is this downstream of the proposed role in phosphate metabolism or a parallel Sit4 activity? This is a crucial point that should be addressed experimentally.

To investigate the role of the Hap complex in MMP generation in sit4Δ cells, we overexpressed and knocked out HAP4, the catalytic subunit of the Hap complex, separately in wild-type and sit4Δ cells. We confirmed the HAP4 overexpression by the enriched abundance of ETC complexes as shown in the BN-PAGE (Figure 2—figure supplement 1E). However, we did not observe any rescue of ETC or ATP synthase in mct1Δ cells when HAP4 was overexpressed. The enriched level of ETC complexes by HAP4 overexpress is not sufficient to rescue the MMP (Figure 2—figure supplement 1F).

Next, we knocked out HAP4 in sit4Δ cells. Knocking out SIT4 could still increase MMP in hap4Δ cells with a much-reduced magnitude, which phenocopied ETC subunit and RPO41 deletion in sit4Δ cells (Figure 2—figure supplement 1G).

In conclusion, the Hap complex is involved in the MMP increase when SIT4 is absent. However, it is not sufficient to increase MMP by overexpressing HAP4. The Hap complex discussion is now included in the manuscript, and the data is presented as Figure 2—figure supplement 1E-G.

5) The authors use the accumulation of Ilv2 precursors as proxy for mitochondrial protein import efficiency. Ilv2 was reported before as a protein which, if import into mitochondria is slow, is deviated into the nucleus in order to be degraded (Shakya,..., Hughes. 2021, Elife). Is it possible that the accumulation of the precursor is the result of a reduced degradation of pre-Ilv2 in the nucleus rather than an impaired mitochondrial import? Since a number of components of the ubiquitin-proteasome system were identified with Sit4 in the same screen, a role of Sit4 in proteasomal degradation seems possible. This should be tested.

We thank the reviewer for pointing out this potential caveat with our Ilv2-FLAG reporter. With limited search and tests, we could not find another reporter that behaves like Ilv2FLAG. The reason Ilv2-FLAG is a perfect reporter for this study is because in wild-type cells, Ilv2-FLAG is not 100% imported. Therefore, we could demonstrate that mitochondria with higher MMP import more efficiently. Unfortunately, all of the mitochondrial proteins that we tested could efficiently import in wild-type cells. To identify other suitable mitochondrial proteins that behave like Ilv2-FLAG, we would need to conduct a more comprehensive screen.

To address the concern of the involvement of protein degradation in obscuring the interpretation of Ilv2-FLAG import, we performed two experiments. First, we measured the proteasomal activity in wild-type and our mutants using a commercial kit (Cayman). We did not observe a statistically significant difference in 20S proteasomal activity between wild-type and sit4Δ cells.

In the second experiment, we reduced the MMP of sit4 cells using CCCP treatment and measured the Ilv2-FLAG import. We first treated sit4Δ cells with different dosage of CCCP for six hours and measured their MMP. sit4Δ cells treated with 75 µM CCCP had comparable MMP to wild-type cells. When we treated sit4Δ cells with higher concentrations of CCCP, most of the cells did not survive after six hours. Next, we performed the Ilv2-FLAG import assay. We observed similar level of unimported Ilv2FLAG (marked with *) in sit4Δ cells treated with 75 µM CCCP. This result confirms that sit4Δ cells have similar Ilv2-FLAG turnover mechanism and activity as the wild-type cells, because when we lower the MMP in sit4Δ background we observe a similar level of unimported Ilv2-FLAG. We thus feel confident in concluding that the Ilv2-FLAG import results are indeed an accurate proxy for MMP level. These data are now included as Figure 1—figure supplement 1H-J in the manuscript.

Author response image 1.

Reviewer #2 (Public Review):

This study reports interesting findings on the influence of a conserved phosphatase on mitochondrial biogenesis and function. In the absence of it, many nucleus-encoded mitochondrial proteins among which those involved in ATP generation are expressed much better than in normal cells. In addition to a better understanding of th mechanisms that regulate mitochondrial function, this work may help developing therapeutic strategies to diseases caused by mitochondrial dysfunction. However there are a number of issues that need clarification.

1) The rationale of the screening assay to identify genes required for the gene expression modifications observed in mct1 mutant is not clear. Indeed, after crossing with the gene deletion libray, the cells become heterozygote for the mct1 deletion and should no longer be deficient in mtFAS. Thank you for clarifying this and if needed adjust the figure S1D to indicate that the mated cells are heterozygous for the mct1 and xxx mutations.

We updated the methods section and the graphic for the genetic screen to clarify these points within the SGA workflow overview. After we created the heterozygote by mating mct1Δ cells with the individual KO cells in the collection, these diploids underwent sporulation and selection for the desired double KO haploid. As a result, the luciferase assay was performed in haploid cells with MCT1 and one additional non-essential gene deleted.

2) The tests shown in Fig. S1E should be repeated on individual subclones (at least 100) obtained after plating for single colonies a glucose culture of mct1 mutant, to determine the proportion of cells with functional (rho+) mtDNA in the mct1 glucose and raffinose cultures. With for instance a 50% proportion of rho- cells, this could substantially influence the results of the analyses made with these cells (including those aiming to evaluate the MMP).

We agree that this would provide a more confident estimate for population-level characterization of these colonies. It is important to note that we randomly chose 10 individual subclones, and 100% of these colonies were verified to be rho+. This suggests the population has functional mtDNA, and thus felt confident in the identity of our populations.

3) The mitochondria area in mct1 cells (Fig.S1G) does not seem to be consistent with the tests in Fig. 1C. that indicate a diminished mitochondrial content in mct1 cells vs wild-type yeast. A better estimate (by WB for instance) of the mitochondrial content in the analyzed strains would enable to better evaluate MMP changes monitored with Mitotracker since the amount of mitochondria in cells correlate with the intensity of the fluorescence signal.

As this reviewer pointed out, we quantified mitochondrial area based on Tom70-GFP signal. This measurement is quantified by mitochondrial area over cell size. Cell size is an important parameter when measuring organelle size as most of the organelles scale up and down with the cell size. mct1Δ cells generally have smaller cell size than WT cells. Therefore, the mitochondrial area of mct1Δ cells was not significantly different from WT cells when scaled to cell size. We believe this is the best method to compare mitochondrial area. As for quantifying MMP from these microscopy images, we measured the average MitoTracker Red fluorescence intensity of each mitochondria defined by Tom70-GFP. This method inherently normalizes to subtract the influence of mitochondria area when quantifying MMP.

4) Page 12: "These data demonstrate that loss of SIT4 results in a mitochondrial phenotype suggestive of an enhanced energetic state: higher membrane potential, hyper-tubulated morphology and more effective protein import." Furthermore, the sit4 mutant shows higher levels of OXPHOS complexes compared to WT yeast.

Despite these beneficial effects on mitochondria, the sit4 deletion strain fails to grow on respiratory substrates. It would be good to know whether the authors have some explanation for this apparent contradiction.

We agree that this was initially puzzling. We provide a more complete explanation above (see comments to reviewer #1 - major concern #1). Briefly, the growth deficiency in non-fermentable media with sit4Δ cells was reported and studied by multiple groups (Arndt et al., 1989; Dimmer et al., 2002; Jablonka et al., 2006). These seems to indicate that sit4Δ cells contain more ETC complexes and more OCR but cannot respire on nonfermentable carbon source. However, we do not think there is yet a clear explanation for this phenotype. One interesting observation reported is the addition of aspartate partly restoring cells’ growth on ethanol (Jablonka et al., 2006). One early study speculates that this defect could be related to the cAMP-PKA pathway. Sutton et al. pointed out genetic interactions with sit4 and multiple genes in cAMP-PKA (Sutton et al., 1991). In addition, sit4Δ cells have similar phenotypes as those cAMP-PKA null mutants, such as glycogen accumulation, caffeine resistance, and failure to grow on non-fermentable media. However, to keep this manuscript succinct, we opted to stay focused on MMP.

Reviewer #3 (Public Review):

In this study, the authors investigate the genetic and environmental causes of elevated Mitochondrial Membrane Potential (MMP) in yeast, and also some physiological effects correlated with increased MMP.

The study begins with a reanalysis of transcriptional data from a yeast mutant lacking the gene MCT1 whose deletion has been shown to cause defects in mitochondrial fatty acid synthesis. The authors note that in raffinose mct1del cells, unlike WT cells, fail to induce expression of many genes that code for subunits of the Electron Transport Chain (ETC) and ATP synthase. The deletion of MCT1 also causes induction of genes involved in acetyl-CoA production after exposure to raffinose. The authors therefore conduct a screen to identify mutants that suppress the induction of one of these acetylCoA genes, Cit2. They then validate the hits from this screen to see which of their suppressor mutants also reduce expression in four other genes induced in a mct1del strain. This yielded 17 genes that abolished induction of all 5 genes tested in an mct1del background during growth on raffinose.

The authors chose to focus on one of these hits, the gene coding for the phosphatase SIT4 (related to human PP6) which also caused an increase in expression of two respiratory chain genes. The authors then investigated MMP and mitochondrial morphology in strains containing SIT4 and MCT1 deletions and surprisingly saw that sit4del cells had highly elevated MMP, more reticular mitochondria, and were able to fully import the acetolactate synthase protein Ilv2p and form ETC and ATP synthase complexes, even in cells with an mct1del background, rescuing the low MMP, fragmented mitochondria, low import of Ilv2 and an inability to form ETC and ATP synthase complexes phenotypes of the mct1del strain. Surprisingly, the authors find that even though MMP is high and ETC subunits are present in the sit4del mct1del double deletion strain, that strain has low oxygen consumption and cannot grow under respiratory conditions, indicating that the elevated MMP cannot come from fully functional ETC subunits. The authors also observe that deleting key subunits of ETC complex III (QCR2) and IV (COX5) strongly reduced the MMP of the sit4del mutant, which would suggest that the majority of the increase in MMP of the sit4del mutant was dependant on a partially functional ETC. The authors note that there was still an increase in MMP in the qcr2del sit4del and cox4del sit4del strains relative to qcr2del and cox4del strains indicating that some part of the increase in MMP was not dependent on the ETC.

The authors dismiss the possibility that the increase in MMP could have been through the reversal of ATP synthase because they observe that inhibition of ATP synthase with oligomycin led to an increase of MMP in sit4del cells. Indicating that ATP synthase is operating in a forward direction in sit4del cells.

Noting that genes for phosphate starvation are induced in sit4del cells, the authors investigate the effects of phosphate starvation on MMP. They found that phosphate starvation caused an increase in MMP and increased Ilv2p import even in the absence of a mitochondrial genome. They find that inhibition of the ADP/ATP carrier (AAC) with bongkrekic acid (BKA) abolishes the increase of MMP in response to phosphate starvation. They speculate that phosphate starvation causes an increase in MMP through the import and conversion of ATP to ADP and subsequent pumping of ADP and inorganic phosphate out of the mitochondria.

They further show that MMP is also increased when the cyclin dependent kinase PHO85 which plays a role in phosphate signaling is deleted and argue that this indicates that it is not a decrease in phosphate which causes the increase in MMP under phosphate starvation, but rather the perception of a decrease in phosphate as signalled through PHO85. Unlike in the case of SIT4 deletion, the increase in MMP caused by the deletion of pho85 is abolished when MCT1 is deleted.

Finally they show an increase in MMP in immortalized human cell lines following phosphate starvation and treatment with the phosphate transporter inhibitor phosphonoformic acid (PFA). They also show an increase in MMP in primary hepatocytes and in midgut cells of flies treated with PFA.

The link between phosphate starvation and elevated MMP is an important and novel finding and the evidence is clear and compelling. Based on their experiments in various mammalian contexts, this link appears likely to be generalizable, and they propose and begin to test an interesting hypothesis for how MMP might occur in response to phosphate starvation in the absence of the Electron Transport Chain.

The link between phosphate starvation and deletion of the conserved phosphatase SIT4 is also interesting and important, and while the authors' experiments and analysis suggest some connection between the two observations, that connection is still unclear.

Major points

Mitotracker is great fluorescent dye, but it measures membrane potential only indirectly. There is a danger when cells change growth rates, ion concentrations, or when the pH changes, all MMP indicating dyes change in fluorescence: their signal is confounded Change in phosphate levels can possibly do both, alter pH and ion concentrations. Because all conclusions of the manuscript are based on a change in MMP, it would be a great precaution to use a dye-independent measure of membrane potential, and confirm at least some key results.

Mitochondrial MMP does strongly influence amino acid metabolism, and indeed the SIT4 knockout has a quite striking amino acid profile, with histidine, lysine, arginine, tyrosine being increased in concentration. http://ralser.charite.de/metabogenecards/Chr_04/YDL047W.html Could this amino acid profile support the conclusions of the authors? At least lysine and arginine are down in petites due to a lack of membrane potential and iron sulfur cluster export.- and here they are up. Along these lines, according to the same data resource, the knock-outs CSR2, ASF1, SSN8, YLR0358 and MRPL25 share the same metabolic profile. Due to limited time I did not re-analyse the data provided by the authors- but it would be worth checking if any of these genes did come up in the screens of the authors.

We tested the mutants within the same cluster as SIT4 shown in this paper from the deletion collection and measured their MMP. yrl358cΔ cells have similar high MMP as observed in sit4Δ cells. However, this gene has a yet undefined function. Beyond YRL358C, we did not observe similar MMP increases in other gene deletions from this panel, which does not support the notion that amino acids such as histidine, lysine, arginine, or tyrosine play a determining effect in driving MMP.

The media condition and strain used in the suggested paper is very different from what we used in our study. Instead of growing prototrophic cells in minimal media without any amino acids, we used auxotrophic yeast strains and grew them in media containing complete amino acids. So far, none of the other defects or signaling associated with SIT4 deletion could influence MMP as much as the phosphate signaling. We interpret these data to support the hypothesis that the MMP observation in sit4Δ cells is connected with the phosphate signaling as illustrated by the second half of the story in our manuscript.

Author reponse image 2.

One important claim in the manuscript attempts to explain a mechanism for the MMP increase in response to phosphate starvation which is independent of the ETC and ATP synthase.

It seems to me the only direct evidence to support this claim is that inhibition of the AAC with BKA stops the increase of mitotracker fluorescence in response to phosphate starvation in both WT and rho0 cells (Figs 4B and 4C). It would strengthen the paper if the authors could provide some orthogonal evidence.

This is a similar comment as raised by reviewer #1 - major concern #3. We refer the reviewer to our discussion and the new data above. Briefly, we do not think F1 subunit is responsible for the ATP hydrolysis activity to generate MMP in phosphate depleted situation. We believe there are additional ATPase(s) in the mitochondrial matrix that can be utilized to couple to ADP/ATP carrier for MMP generation during phosphate starvation. However, we have not identified the relevant ATPase(s) at this point, and it is likely that multiple ATPases could contribute to this activity.

Introduction/Discussion The author might want to make the reader of the article aware that the 'reversal' of the ATP synthase directionality -i.e. ATP hydrolysis by the ATP synthase as a mechanism to create a membrane potential (in petites), has always been a provocative idea - but one that thus far could never be fully substantiated. Indeed some people that are very familiar with the topic, are skeptical this indeed happens. For instance, Vowinckel et al 2021 (PMID: 34799698) measured precise carbon balances for peptide cells, and found no evidence for a futile cycle - peptides grow slower, but accumulate the same biomass from glucose as peptides that re-evolve at a fast growth rate . Perhaps the manuscript could be updated accordingly.

We thank the reviewer for pointing out this additional relevant study. We have rephased the referenced sentence in the introduction. The MMP generation in phosphate starvation is independent of the F1 portion of ATP synthase. Therefore, our data neither supports or refutes either of these arguments.

In the introduction and conclusion there is discussion of MMP set points. In particular the authors state:

"Critically, we find that cells often prioritize this MMP setpoint over other bioenergetic priorities, even in challenging environments, suggesting an important evolutionary benefit."

This does not seem to be consistent with the central finding of the manuscript that MMP changes under phosphate starvation. MMP doesn't seem so much to have a 'set point' but rather be an important physiological variable that reacts to stimuli such as phosphate starvation.

The reviewer raises a rational alternative hypothesis to the one that we have proposed. In reality, both of these are complete speculations to explain the data and we can’t think of any way to test the evolutionary basis for the mechanisms that we describe. We recognize that untested/untestable speculative arguments have limitations and there are viable alternative hypotheses. We have softened our language to ensure that it is clear that this is only a speculation.

The authors suggest that deletion of Pho85 causes an increase in MMP because of cellular signaling. However, they also state in the conclusion:

"Unlike phosphate starvation, the pho85D mutant has elevated intracellular phosphate concentrations. This suggests that the phosphate effect on MMP is likely to be elicited by cellular signaling downstream of phosphate sensing rather than some direct effect of environmental depletion of phosphate on mitochondrial energetics."

The authors should cite the study that shows deletion of PHO85 causes increased intracellular phosphate concentrations. It also seems possible that the 'cellular signaling' that causes the increase in MMP could be a result of this increase in intracellular phosphate concentrations, which could constitute a direct effect of an environmental overload of phosphate on mitochondrial energetics.

We now cited the literature that shows higher intracellular phosphate in pho85Δ cells (Gupta et al., 2019; Liu et al., 2017). Depleting phosphate in the media drastically reduced intracellular phosphate concentration, which is the opposing situation as pho85Δ cells. Nevertheless, we observed higher MMP in either situation. We concluded from these two observations that the increase in MMP is a response to the signaling activated by phosphate depletion rather than the intracellular phosphate abundance.

Related to this point, in the conclusion, the authors state:

"We now show that intracellular signaling can lead to an increased MMP even beyond the wild-type level in the absence of mitochondrial genome."

In sum, the data shows that signaling is important here- but signaling alone is only the message - not the biophysical process that creates a membrane potential. The authors then could revise this slightly.

We have rephrased this sentence as suggested, which now reads “We now show that intracellular signaling triggers a process that can lead to an increased MMP even beyond the wild-type level in the absence of mitochondrial genome”.

The authors state in the conclusion that

"We first made the observation that deletion of the SIT4 gene, which encodes the yeast homologue of the mammalian PP6 protein phosphatase, normalized many of the defects caused by loss of mtFAS, including gene expression programs, ETC complex assembly, mitochondrial morphology, and especially MMP (Fig. 1)"

The data shown though indicates that a defect in mtFAS in terms of MMP, deletion of SIT4 causes a huge increase (and departure away from normality) whether or not mct1 is present (Fig 1D)

We changed the word “normalized” to “reversed”. In the discussion section, we also emphasized that many of these increases are independent of mitochondrial dysfunction induced by loss of mtFAS.

The language "SIT4 is required for both the positive and negative transcriptional regulation elicited by mitochondrial dysfunction" feels strong. SIT4 seems to influence positive transcriptional regulation in response to mitochondrial dysfunction caused by MCT1 deletion (but may not be the only thing as there appears to be an increase in CIT2 expression in a sit4del background following a further deletion of MCT1). In terms of negative regulation, SIT4 deletion clearly affects the baseline, but MCT1 deletion still causes down regulation of both examples shown in Fig 1B, showing that negative transcriptional regulation can still occur in the absence of SIT4. The authors might consider showing fold change of expression as they do in later figures (Figs 4B and C) to help the reader evaluate the quantitative changes they demonstrate.

We now displayed the fold change as suggested. This sentence now reads “These data suggest that SIT4 positively and negatively influences transcriptional regulation elicited by mitochondrial dysfunction”.

The authors induce phosphate starvation by adding increasing amounts of potassium phosphate monobasic at a pH of 4.1 to phosphate dropout media supplemented with potassium. The authors did well to avoid confounding effects of removing potassium. The final pH of YNB is typically around 5.2. Is it possible that the authors are confounding a change in pH with phosphate starvation? One would expect the media in the phosphate starvation condition to have a higher pH than the phosphate replacement or control media. Is a change in pH possibly a confounding factor when interpreting phosphate starvation? Perhaps the authors could quantify the pH of the media they use for the experiment to understand how much of a factor that could be. One needs to be careful with Miotracker and any other fluorescent dye when pH changes. Albeit having constraints on its own, MitoLoc as a protein rather than small molecule marker of MMP might be a good complement.

We followed the protocol used by many other studies that depleted phosphate in the media. The reason we and others adjusted the media without inorganic phosphate to a pH of 4.1 is because that is the pH of phosphate monobasic. From there, we could add phosphate monobasic to create +Pi media without changing the media pH. Therefore, media containing different concentrations of phosphate all have the exact same pH. We now emphasize that all media containing different levels of inorganic phosphate have the same pH to the manuscript to eliminate such concern (see page 18).

Even though all media have the similar pH, we also provided complementary data using a parallel approach to measure the MMP by assessing mitochondrial protein import as demonstrated previously with Ilv2-FLAG, which shares the same principle as mitoLoc.

Reference

Arndt, K. T., Styles, C. A., & Fink, G. R. (1989). A suppressor of a HIS4 transcriptional defect encodes a protein with homology to the catalytic subunit of protein phosphatases. Cell, 56(4), 527–537. https://doi.org/10.1016/00928674(89)90576-X

Dimmer, K. S., Fritz, S., Fuchs, F., Messerschmitt, M., Weinbach, N., Neupert, W., & Westermann, B. (2002). Genetic basis of mitochondrial function and morphology in Saccharomyces cerevisiae. Molecular Biology of the Cell, 13(3), 847–853. https://doi.org/10.1091/mbc.01-12-0588

Gupta, R., Walvekar, A. S., Liang, S., Rashida, Z., Shah, P., & Laxman, S. (2019). A tRNA modification balances carbon and nitrogen metabolism by regulating phosphate homeostasis. ELife, 8, e44795. https://doi.org/10.7554/eLife.44795

Jablonka, W., Guzmán, S., Ramírez, J., & Montero-Lomelí, M. (2006). Deviation of carbohydrate metabolism by the SIT4 phosphatase in Saccharomyces cerevisiae. Biochimica et Biophysica Acta (BBA) - General Subjects, 1760(8), 1281–1291. https://doi.org/10.1016/j.bbagen.2006.02.014

Liu, N.-N., Flanagan, P. R., Zeng, J., Jani, N. M., Cardenas, M. E., Moran, G. P., & Köhler, J. R. (2017). Phosphate is the third nutrient monitored by TOR in Candida albicans and provides a target for fungal-specific indirect TOR inhibition. Proceedings of the National Academy of Sciences, 114(24), 6346–6351. https://doi.org/10.1073/pnas.1617799114

Sutton, A., Immanuel, D., & Arndt, K. T. (1991). The SIT4 protein phosphatase functions in late G1 for progression into S phase. Molecular and Cellular Biology, 11(4), 2133–2148.

Author Response

Reviewer #1 (Public Review):

This study provides further detailed analysis of recently published Fly Atlas datasets supplemented with newly generated single cell RNA-seq data obtained from 6,000 testis cells. Using these data, the authors define 43 germline cell clusters and 22 somatic cell clusters. This work confirms and extends previous observations regarding changing gene expression programs through the course of germ cell and somatic cell differentiation.

This study makes several interesting observations that will be of interest to the field. For example, the authors find that spermatocytes exhibit sex chromosome specific changes in gene expression. In addition, comparisons between the single nucleus and single cell data reveal differences in active transcription versus global mRNA levels. For example, previous results showed that (1) several mRNAs remain high in spermatids long after they are actively transcribed in spermatocytes and (2) defined a set of post-meiotic transcripts. The analysis presented here shows that these patterns of mRNA expression are shared by hundreds of genes in the developing germline. Moreover, variable patterns between the sn- and sc-RNAseq datasets reveals considerable complexity in the post-transcriptional regulation of gene expression.

Overall, this paper represents a significant contribution to the field. These findings will be of broad interest to developmental biologists and will establish an important foundation for future studies. However, several points should be addressed.

In figure 1, I am struck by the widespread expression of vasa outside of the germ cell lineage. Do the authors have a technical or biological explanation for this observation? This point should be addressed in the paper with new experiments or further explanation in the text.

Thank you for pointing this out. We found that our single cell dataset shows a similar (low) level of vasa expression outside the germline, suggesting that this is not due to single nucleus versus single cell RNA-seq (cluster 1, red in the lefthand umap).

Analyzing the single nucleus RNA-seq in more detail revealed that, compared to the germline, both the fraction of cells in a cluster expressing vasa and the level at which they express it are very low. This analysis is included in a new Figure 1 – figure supplement 1. It is likely that much of this is due to a technical artifact, such as ambient RNA. Finally, we note in the resubmission that vasa is in fact expressed in embryonic somatic cells, and thus some of the vasa expression we observe may be real (Renault. Biol Open 2012; https://doi.org/10.1242/bio.20121909).

Plots in the original submission drew undue attention to the few somatic cells that exhibited vasa signal, due to the fact that expressing cell points were forced to the front of the plot. Given our new analysis reporting the low levels and fraction of cells exhibiting vasa expression (Figure 1 – figure supplement 1), we have modified the panels of Figure 1, changing point size to more faithfully reflect the small proportion of somatic cells with some vasa expression.

The proposed bifurcation of the cyst cells into head and tail populations is interesting and worth further exploration/validation. While the presented in situ hybridization for Nep4, geko, and shg hint at differences between these populations, double fluorescent in situs or the use of additional markers would help make this point clearer. Higher magnification images would also help in this regard.

We thank the reviewer for their suggestions on clarifying the differences between HCC and TCC populations. As suggested, we have repeated the FISH experiments of Nep4 and geko with higher resolution, and included the additional marker Coracle that demarcates the junction between HCC and TCC (Figure 6O,Q,S,T). These panels replaced previous Nep4 and geko FISH images (see previous Figure 6Q,U,U’). FISH for Nep4 validated the split, and the enrichment of geko strongly suggests that this arm represents one cell type (HCCs). We have not yet identified a gene reciprocally enriched to the other arm. Therefore, in the revised submission, we call the assignment of TCC identity, and to a lesser extent, HCC identity ‘tentative’, but point out that genes predicted to be enriched to one or the other arm represent fertile candidates for the field to test.

Reviewer #2 (Public Review):

In this manuscript the authors explain in greater detail a recent testis snRNAseq dataset that many of these authors published earlier this year as part of the Fly Cell Atlas (FCA) Li et al. Science 2022. As part of the current effort additional collaborators were recruited and about 6,000 whole cell scRNAseq cells were added to the previous 42,000 nuclei dataset. The authors now describe 65 snRNseq clusters, each representing potential cell types or cell states, including 43 germline clusters and 22 somatic clusters. The authors state that this analysis confirms and extends previously knowledge of the testis in several important areas.

“However, in areas where testis biology is well studied, such as the development of germ cells from GSC to the onset of spermatocyte differentiation, the resolution seems less than current knowledge by considerable margins. No clusters correspond to GSCs, or specific mitotic spermatogonia, and even the major stages of meiotic prophase are not resolved. Instead, the transitions between one state and the next are broad and almost continuous, which could be an intrinsic characteristic of the testis compared to other tissues, of snRNAseq compared to scRNAseq, or of the particular experimental and software analysis choices that were used in this study.”

Note that the referee raises the same issue later in their review also. To respond succinctly, we placed the relevant sentence from a later portion of this referee’s comment here

“Support for the view that the problems are mostly technical, rather than a reflection of testis biology, comes from studies of scRNAseq in the mouse, where it has been possible to resolve a stem cell cluster, and germ cell pathways that follow known germ cell differentiation trajectories with much more discrete steps than were reported here (for example, Cao et al. 2021 cited by the authors).”

Respectfully, we have a different interpretation of other work as cited by this referee. Our data, as well as that from others, supports the notion that transitions are generally broad and continuous and are indeed a feature of testis biology. As we report here, data from both single cell and single nucleus RNAseq exhibit transitions from one cluster to the next. Thus, this feature cannot be due to the choice of method (single cell versus single nucleus).

In fact, prior scRNA-seq results on systems containing a continuously renewing cell population, such as is the case in the testis, do indeed exhibit a contiguous trajectory rather than discrete, well-separated cell states in gene expression space (that is, in a UMAP presentation). For example, this is the case from single-cell or single-nucleus sequencing from spermatogenesis in mouse (Cao et al 2021), human (Sohni et al 2019), and zebrafish (Qian et al 2022).

Along differentiation trajectories in these tissues, successive clusters are defined by their aggregate, transcript repertoire. Indeed, differentially-expressed genes can be identified for clusters, with expression enriched in a given cluster. However, expression is rarely restricted to a cluster. For instance, Cao et al. subcluster spermatogonia into four subgroups, termed SPG1-4. They state clearly that these SPG1-4 “follow a continuous differentiation trajectory,” as can be inferred by marker expression across cells in this lineage. Similar to our findings, while the spermatogonia can fall into discrete clusters, gene expression patterns are contiguous. For example, the “undifferentiated” marker used in Cao et al, Crabp1, clearly shows expression in SPG1-3, annotated as spermatogonial stem cells, undifferentiated spermatogonia, and early differentiated spermatogonia, respectively. Likewise, markers for the “SPG3” state spermatogonia have detectable expression in SPG2 and SPG4, and likewise for markers of the “SPG4” state (with expression found also in SPG3). <br /> Analogous study of human spermatogenesis arrives at a similar conclusion. In that work, although clusters are named as “spermatogonial stem cell (SSC)”, the authors are careful to specifically point out that, “…while we refer to the SSC-1 and SSC-2 cell clusters as ‘‘SSCs,’’ scRNA-seq is not a functional assay and thus we do not know the percentage of cells in these clusters with SSC activity. These subsets almost certainly contain other A-SPG cells [A type spermatogonia], including SPG progenitors that have committed to differentiate.” (Sohi et al 2019)

Thus, the work in several disparate systems, all involving renewing lineages, finds that discrete clusters, such as a “stem cell cluster” are not identified. In the Drosophila testis, germline differentiation flows in a continuous-like manner similar to spermatogenesis in several other organisms studied by scRNA-seq, and our finding is not a function of the methodology, but rather a facet of the biology of the organ.

Operating in parallel with continuous differentiation, we did find evidence of, and extensively discussed in concert with Figure 4, huge and dramatic shifts in transcriptional state in spermatocytes compared to spermatogonia, in early spermatids compared to spermatocytes, and in late spermatid elongation. Lastly, as we describe further below, new data in this resubmission identify four distinct genes with stage-selective expression as predicted by our analysis (new Figure 2 - figure supplement 1), illustrating the utility of our study for the field to find new markers and new genes to test for function.

A goal of the study was to identify new rare cell types, and the hub, a small apical somatic cell region, was mentioned as a target region, since it regulates both stem cell populations, GSCs and CySCs, is capable of regeneration, and other fascinating properties. However the analysis of the hub cluster revealed more problems of specificity. 41 or 120 cells in the cluster were discordant with the remaining 79 which did express markers consistent with previous studies. Why these cells co-clustered was not explained and one can only presume that similar problems may be found in other clusters.

Our writing seems not to have been clear enough on this point and we thank the reviewer. We have revised the section. In addition, we have added new data (Figure 7 - figure supplement 2). We had already stated that only 79 of these 120 nuclei were near to each other in 2D UMAP space, while other members of original cluster 90 were dispersed. Thus the 79 hub nuclei in fact clustered together on the UMAP. Other nuclei that mapped at dispersed positions were initially ‘called’ as part of this cluster in the original Fly Cell Atlas (FCA) paper (Li et al., 2022), making it obvious that a correction to that assignment was necessary, which we carried out. To our eye, no other called cluster was represented by such dispersed groupings. For the hub, we definitively established the 79 nuclei to represent hub cells by marker gene analysis, including the identification of a new maker, tup, that was included in the 79 annotated hub nuclei but excluded from the 41 other nuclei (Figure 7). In this resubmission, to independently verify the relationship of the 79 nuclei to each other, we subjected the 120 nuclei from the original cluster 90 defined by the FCA study to hierarchical clustering using only genes that are highly expressed and variable in these nuclei (Figure 7 - figure supplement 2). This computationally distinct approach strongly supported our identification of the 79 definitive hub nuclei.

Indeed, many other indications of specificity issues were described, including contamination of fat body with spermatocytes, the expression of germline genes such as Vasa in many somatic cell clusters like muscle, hemocytes, and male gonad epithelium, and the promiscuous expression of many genes, including 25% of somatic-specific transcription factors, in mid to late spermatocytes. The expression of only one such genes, Hml, was documented in tissue, and the authors for reasons not explained did not attempt to decisively address whether this phenomenon is biologically meaningful.

We discussed the question of vasa expression in somatic clusters in some detail above, in response to referee #1, and included new analysis in the resubmission.

With respect to the observation of ‘somatic gene’ expression in spermatocytes, we are also intrigued. We do not believe this is due to “contamination,” but rather a spermatocyte expression program that includes expression of somatic genes. First, these somatic markers were not observed in other germline clusters, which would be expected if this was due to general transcript contamination. Second, we observed expression of somatic markers in spermatocytes independently in the single-cell and single-nucleus data, making it unlikely to be an artifact of preparation of isolated nuclei. Finally, in the resubmission, in addition to Hml, we validated ‘somatic’ marker expression in spermatocytes by FISH of a somatic, tail cyst cell marker, Vsx1. Vsx1 is predicted to be expressed at low levels in spermatocytes in our dataset and is clearly visible in germline cells by FISH (Figure 3 – figure supplement 2G,H). We also refer the referee to Figure 6K, where the mRNA for the somatic cyst cell marker eya was observed by FISH at low levels in spermatocytes.

A truly interesting question mentioned by the authors is why the testis consistently ranks near the top of all tissues in the complexity of its gene expression. In the Li et al. (2022) paper it was suggested that this is due an inherently greater biological complexity of spermiogenesis than other tissues. It seems difficult to independently and rationally determine "biological complexity," but if a conserved characteristic of testis was to promiscuously express a wide range of (random?) genes, something not out of the question, this would be highly relevant and important.

We agree that the massive transcriptional program found in spermatocytes is, indeed, truly interesting. There are many speculations as to why spermatocytes are so highly transcriptional, including the possibility of “transcriptional scanning” (e.g., Xia et al. 2020) regulating the evolution of new genes. Testing such models is beyond the scope of this paper. However, one must also keep in mind that spermatogenesis involves one of the most dramatic cellular transformations in biology, where cellular components spanning from nuclei to chromatin to Golgi, cell cycle, extensive membrane addition, changes in cell shape, and building of a complex swimming organelle all must occur and be temporally coordinated. Small wonder that many genes must be expressed to accomplish these tasks.

Unfortunately, the most likely problems are simply technical. Drosophila cells are small and difficult to separate as intact cells. The use of nuclei was meant to overcome this inherent problem, but the effectiveness of this new approach is not yet well-documented. Support for the view that the problems are mostly technical, rather than a reflection of testis biology, comes from studies of scRNAseq in the mouse, where it has been possible to resolve a stem cell cluster, and germ cell pathways that follow known germ cell differentiation trajectories with much more discrete steps than were reported here (for example, Cao et al. 2021 cited by the authors).

We respectfully disagree with the referee about this collection of statements. First, the use of snRNASeq has been extensively characterized and compared to scRNA-seq in brain tissue by McLaughlin et al., 2021 (cited in the original submission) and was shown to be effective (McLaughlin, et al. eLife 2021;10:e63856. DOI: https://doi.org/10.7554/eLife.63856). snRNA-seq has a distinct advantage when dealing with long, thin cells, such as neurons or cyst cells (as featured in this work), where cytoplasm can easily be sheared off during cell isolation. Second, in a previous portion of our response to this referee, we discussed how our interpretation of Cao et al., 2021 differs from that expressed by this referee. Lastly, as requested in ‘Essential revision’ 2, we adjusted clustering methods and selected four genes, two predicted to be markers for early stage germline cells, and two for mid-spermatocyte stage development. FISH analysis demonstrates that expression for each of these maps to the appropriate stages (new Figure 2 - figure supplement 1). This confirms that the datasets we present in this manuscript can be mined to identify unique, diagnostic markers for various stages.

The conclusions that were made by the authors seem to either be facts that are already well known, such as the problem that transcriptional changes in spermatocytes will be obscured by the large stored mRNA pool, or promises of future utility. For example, "mining the snRNA-seq data for changes in gene expression as one cluster advances to the next should identify new sub-stage-specific markers." If worthwhile new markers could be identified from these data, surely this could have been accomplished and presented in a supplemental Table. As it currently stands, the manuscript presents the dataset including a fair description of its current limitations, but very little else of novel biological interest is to be found.

“In sum, this project represents an extremely worthwhile undertaking that will eventually pay off. However, some currently unappreciated technical issues, in cell/nuclear isolation, and certainly in the bioinformatic programs and procedures used that mis-clustered many different cells, has created the current difficulties.

Most scRNAseq software is written to meet the needs of mammalian researchers working with cultured cells, cellular giants compared to Drosophila and of generally similar size. Such software may not be ideal for much smaller cells, but which also include the much wider variation in cell size, properties and biological mechanisms that exist in the world of tissues.”

We appreciate the referee’s acknowledgement that this ‘undertaking will eventually pay off’. It was not our intention to address ‘function’ for this study, but rather to make the system accessible to the broadest community possible. We are uncertain if there is any remaining reservation held by this referee. A brief summary of what we covered in the manuscript may help allay any residual concern. Obviously, study of the Drosophila testis and spermatogenesis benefits from the knowledge of a large number of established cell-type and stage-selective markers. Thus, we extensively used the community’s accepted markers to assign identity to clusters in both the sn- and sc-RNA-seq UMAPs. We believe that effort well establishes the validity and reliability of the dataset . Furthermore, we identified upwards of a dozen new markers out of the cluster analysis, and verified their expression by FISH or reporter line in various figures throughout (tup, amph, piwi, geko, Nep4, CG3902, Akr1B, loqs, Vsx1, Drep2, Pxt, CG43317, Vha16-5, l(2)41Ab). To our mind, these contributions, coupled with annotation of the datasets, suggest strongly that they will serve the community well. This is especially true as we provide users with objects that they can feed into commonly used software algorithms such as Seurat and Monocle to explore the datasets to their purposes. Rather than simply relying on default settings within some of the applications, we also adjusted parameters for various clusterings as called for; some of which were in response to astute comments from referees, and included in the resubmission. Of course, it is possible that rare issues may arise in the datasets as these are further studied, but that is the case with all scRNA-seq data, and is not specific to work on this model organism.

Reviewer #3 (Public Review):

In this study, the authors use recently published single nucleus RNA sequencing data and a newly generated single cell RNA sequencing dataset to determine the transcriptional profiles of the different cell types in the Drosophila ovary. Their analysis of the data and experimental validation of key findings provide new insight into testis biology and create a resource for the community. The manuscript is clearly written, the data provide strong support for the conclusions, and the analysis is rigorous. Indeed, this manuscript serves as a case study demonstrating best practices in the analysis of this type of genomics data and the many types of predictions that can be made from a deep dive into the data. Researchers who are studying the testis will find many starting points for new projects suggested by this work, and the insightful comparison of methods, such as between slingshot and Monocle3 and single cell vs single nucleus sequencing will be of interest beyond the study of the Drosophila testis.

We greatly appreciate the reviewer’s comments.

Reviewer #4 (Public Review):

This is an extraordinary study that will serve as key resource for all researchers in the field of Drosophila testis development. The lineages that derive from the germline stem cells and somatic stem cells are described in a detail that has not been previously achieved. The RNAseq approaches have permitted the description of cell states that have not been inferred from morphological analyses, although it is the combination of RNAseq and morphological studies that makes this study exceptional. The field will now have a good understanding of interactions between specific cell states in the somatic lineage with specific states in the germ cell lineage. This resource will permit future studies on precise mechanisms of communication between these lineages during the differentiation process, and will serve as a model for studies of co-differentiation in other stem cell systems. The combination of snRNAseq and scRNAseq has conclusively shown differences in transcriptional activation and RNA storage at specific stages of germ cell differentiation and is a unique study that will inform other studies of cell differentiation.

Could the authors please describe whether genes on the Y chromosome are expressed outside of the male germline. For example, what is represented by the spots of expression within the seminal vesicle observed in Figure 3D?

Prior work demonstrated that proteins encoded by Y-linked genes are not expressed outside of the germline (Zhang et al. Genetics 2020. https://doi.org/10.1534/genetics.120.303324). In our snRNAseq dataset, we find that genes on the Y chromosome are not highly expressed outside of the male germline (on the order of ~100-fold lower in other tissues). In fact, we observe Y chromosome transcripts at this level in many nuclei across tissues collected for the Fly Cell Atlas project, including the ovary. Since we have not followed up on the Fly Cell Atlas observations directly using FISH to examine Y chromosome transcript expression outside the germline, we cannot rule out the possibility that such low level expression is real. However, the detection across several tissues argues that this is likely technical artifact. With regard to ‘spots of expression within the seminal vesicle’ (Figure 3D), a spot is colored red if the average expression level of genes on the Y chromosome is greater in that cell than in an average cell on our plot. These red spots are likely due to ambient RNA being carried over.

I would appreciate some discussion of the "somatic factors" that are observed to be upregulated in spermatocytes (e.g. Mhc, Hml, grh, Syt1). Is there any indication of functional significance of any of these factors in spermatocytes?

This is an excellent question. Although we validated expression for several (Hml, Vsx1 and eya), we did not test for their function here and this issue remains to be studied. This is now directly stated in the main text.

In the discussion of cyst cell lineage differentiation following cluster 74 the authors state that neither the HCC or TCC lineages were enriched for eya (Figure 6V). It seems in this panel that cluster 57 shows some enrichment for eya - is this regarded as too low expression to be considered enriched?

We thank the reviewer for their insightful comment and we agree with their conclusions. We have modified the text to reflect the low, but present, expression of eya in the HCC and TCC lineages. The text now reads as follows at line (insert line # here): “Enrichment of eya was dramatically reduced in the clusters along either late cyst cell branch compared to those of earlier lineage nuclei (Figure 6J,U).”

Author Response

Reviewer #2 (Public Review):

The work proposes a new computational rule for classifying synaptic plasticity outcome based on the geometry of synaptic enzyme dynamics. Specifically, the authors implement a multi-timescale model of hippocampal synaptic plasticity induction that takes into account the dynamics of the membrane potential, calcium concentration as well as CaMKII and calcineurin signalling pathways. They show that the proposed rule could be applied to reproduce the outcomes from nine published experimental studies involving different spike-timing and frequency-dependent plasticity induction protocols, animal ages, and experimental conditions. The model has been also used to generate predictions regarding the effect of spike-timing irregularity on plasticity outcomes. The proposed approach constitutes an interesting and original idea that contributes to the ongoing effort in discovering the rules of synaptic plasticity.

The conclusions of this paper are mostly well supported by data, but some model assumptions and interpretation of modelling results need to be clarified and extended.

1) The proposed model captures well the stochastic nature of the dendritic spine ion channels and receptors except for the calcium-sensitive potassium (SK) channel that has been modelled deterministically. Given that the same justification in terms of small number of channels present in the small dendritic spine compartment applies to the SK channels as well as to the voltage gated calcium channels and the AMPA and NMDA receptors, it is not clear why the authors have chosen a deterministic representation in the case of SK. The implications of this assumption needs to be investigated and discussed.

There are several stochastic models of AMPA and NMDA receptors based on single-channel recordings. Additionally, we had enough experimental data on single channel recordings to build a custom Markov chain model of VGCCs. For the SK channel, we could not find enough experimental data (age-dependence activity, temperature sensitivity, etc.) to custom-build a stochastic model. We thus decided to implement a deterministic model. Yet, we understand the reviewers’ comment that in theory, a stochastic model of SK channels could impact our results. We thus now provide a simulation with a stochastic model of SK, comparing it to the deterministic model implemented in the study.

We describe a minimal version of a stochastic model of SK compatible with the deterministic version. The deterministic model of SK channel fit at ~35C is described in the methods section.

Because of the factor ρ 𝑓𝑆𝐾 in the equation, which multiplies r(Ca) by ~2, the equation cannot be related to a 2-state Markov chain (MC). This could probably be possible with a 3-state MC but we used a different strategy. Noting that ρ 𝑆𝐾 ∼ 2 , we introduce a new equation

As 0 < r(Ca) < 1, it is straightforward to introduce a 2-state MC for which the above equation describes the probability of the open state. We then simulate two such independent (for a given Ca concentration) channels and approximate 𝑚 𝑆𝐾 as the sum (which belongs to [0,2Nsk]) of the open states for the 2 channels.

As the reviewer can see in the figure below, we do not find a major difference in the simulations of 3 protocols. Thus, we argue that adding a stochastic version of the SK channels in our current study would not fundamentally alter our main conclusions.

Figure Legend: a comparison using Tigaret et al. 2016 1Pre2Post10 and 1Pre2Post50 protocols, and 900 at 50 Hz protocol from Dudek and Bear 1992 (100 repetitions) between the model with the deterministic SK channel (original model - blue), and the modified model including the stochastic SK channel (stochastic SK - red). Deterministic vs stochastic SK channel does not significantly modify the model’s behaviour.

To explain our rationale of using a deterministic version of SK channel, we provide this sentence in the Methods when describing SK channel model: “"Due to a lack of single-channel recordings of SK channels, and a lack of published stochastic models of SK channels, we modelled SK channels deterministically. In tests we found that this assumption had only a negligible impact on the outcomes of plasticity protocols (data not shown)" (page 40).

2) Many of the model parameters have been set to values previously estimated from synaptic physiology and biochemistry experiments, However, a significant number of important parameter values have been tuned to reproduce the plasticity experiments targeted in this study. As such, it needs to be explained which of the plasticity outcomes have been reproduced because the parameters are chosen to do so. A clarification would have helped to substantiate the authors' conclusions.

Most parameters were set with values previously defined by experimental work. We referred to these publications where necessary throughout the Methods and Tables in our original manuscript. For the few free parameters that were adjusted, we now provide additional information wherever necessary for the Tables concerned.

● In the legend of Table 4 (neuron electrical properties), we explain which parameters are different from values obtained from the literature to fit experimental data (Golding et al. 2001; Buchanan et al. 2007).

● Parameters for the sodium and potassium conductance (Table 5) are labelled as generic since they are intentionally set to produce the BaP dynamics we have shown in the paper.

● Table 6 has no free parameters.

● Table 7 caption now includes a description saying ’Note that the buffer concentration, calcium diffusion coefficient, calcium diffusion time constant and calcium permeability were considered free parameters to adjust the calcium dynamics’.

● In Table 8 we had originally pointed out how we adapted the GluN2B rates from a published GluN2A model (Popescu et al. 2004; and Iacobucci and Popesco 2018). We now describe this adaptation in the Table 8 legend. In this Table, we now also better explain how we adjusted the NMDAr model to reflect the ratio between GluN2B and GluN2A, fitted from Sinclair et al. 2016; and the NMDAr conductance depending on calcium fitted from Maki and Popescu 2014.

● In Table 9 caption we now explain how the GABAr number and conductance were modified to fit GABAr currents as in Figures 15 b and e. The relevant parameters are indicated in the table.

● In Table 10 caption we now state the number of VGCCs per subtype that we used as a free parameter to reproduce the calcium dynamics (Figure 12).

3) Adding experimental testing of model predictions, for example, that firing variability can alter the rules of plasticity, in the sense that it is possible to add noise to cause LTP for protocols that did not otherwise induce plasticity would be needed to increase confidence in the presented modelling results.

We agree that it would be interesting in the future to test the many model predictions suggested in this work with biological experiments. This would however require a lot of work and will be the subject of further studies.

Reviewer #3 (Public Review):

This manuscript presents and analyzes a novel calcium-dependent model of synaptic plasticity combining both presynaptic and postsynaptic mechanisms, with the goal of reproducing a very broad set of available experimental studies of the induction of long-term potentiation (LTP) vs. long-term depression (LTD) in a single excitatory mammalian synapse in the hippocampus. The stated objective is to develop a model that is more comprehensive than the often-used simplified phenomenological models, but at the same time to avoid biochemical modeling of the complex molecular pathways involved in LTP and LTD, retaining only its most critical elements. The key part of this approach is the proposed "geometric readout" principle, which allows to predict the induction of LTP vs. LTD by examining the concentration time course of the two enzymes known to be critical for this process, namely (1) the Ca2+/calmodulin-bound calcineurin phosphatase (CaN), and (2) the Ca2+/calmodulin-bound protein kinase (CaMKII). This "geometric readout" approach bypasses the modeling of downstream pathways, implicitly assuming that no further biochemical information is required to determine whether LTP or LTD (or no synaptic change) will arise from a given stimulation protocol. Therefore, it is assumed that the modeling of downstream biochemical targets of CaN and CaMKII can be avoided without sacrificing the predictive power of the model. Finally, the authors propose a simplified phenomenological Markov chain model to show that such "geometric readout" can be implemented mechanistically and dynamically, at least in principle.

Importantly, the presented model has fully stochastic elements, including stochastic gating of all channels, stochastic neurotransmitter release and stochastic implementation of all biochemical reactions, which allows to address the important question of the effect of intrinsic and external noise on the induction of LTP and LTD, which is studied in detail in this manuscript.

Mathematically, this modeling approach resembles a continuous stochastic version of the "liquid computing" / "reservoir computing" approach: in this case the "hidden layer", or the reservoir, consists of the CaMKII and CaM concentration variables. In this approach, the parameters determining the dynamics of these intermediate ("hidden") variables are kept fixed (here, they are constrained by known biophysical studies), while the "readout" parameters are being trained to predict a target set of experimental observations.

Strengths:

1) This modeling effort is very ambitious in trying to match an extremely broad array of experimental studies of LTP/LTD induction, including the effect of several different pre- and post-synaptic spike sequence protocols, the effect of stimulation frequency, the sensitivity to extracellular Ca2+ and Mg2+ concentrations and temperature, the dependence of LTP/LTD induction on developmental state and age, and its noise dependence. The model is shown to match this large set of data quite well, in most cases.

2) The choice for stochastic implementation of all parts of the model allows to fully explore the effects of intrinsic and extrinsic noise on the induction of LTP/LTD. This is very important and commendable, since regular noise-less spike firing induction protocols are not very realistic, and not every relevant physiologically.

3) The modeling of the main players in the biochemical pathways involved in LTP/LTD, namely CaMKII and CaN, aims at sufficient biological realism, and as noted above, is fully stochastic, while other elements in the process are modeled phenomenologically to simplify the model and reveal more clearly the main mechanism underlying the LTP/LTD decision switch.

4) There are several experimentally verifiable predictions that are proposed based on an in-depth analysis of the model behavior.

We thank the reviewer for pointing out these strengths.

Weaknesses:

1) The stated explicit goal of this work is the construction of a model with an intermediate level of detail, as compared to simplified "one-dimensional" calcium-based phenomenological models on the one hand, and comprehensive biochemical pathway models on the other hand. However, the presented model comes across as extremely detailed nonetheless. Moreover, some of these details appear to be avoidable and not critical to this work. For instance, the treatment of presynaptic neurotransmitter release is both overly detailed and not sufficiently realistic: namely, the extracellular Ca2+ concentration directly affects vesicle release probability but has no effect on the presynaptic calcium concentration. I believe that the number of parameters and the complexity in the presynaptic model could be reduced without affecting the key features and findings of this work.

This point is largely answered in Essential Revisions point 4 where we argue the choices we made for the presynaptic model. We acknowledge, however, that in this current version, we did not incorporate all biophysical components, such as the modulation of presynaptic calcium concentration with external calcium variations and multivesicular release. The calcium-dependence of presynaptic release, as modeled currently, is however fitted in Figure 8e against data from Hardingham et al. 2006 and Tigaret et al. 2016. These current limitations could be addressed in a next version of our presynaptic model where we also plan to incorporate age and temperature influence.

2) The main hypotheses and assumptions underlying this work need to be stated more explicitly, to clarify the main conclusions and goals of this modeling work. For instance, following much prior work, the presented model assumes that a compartment-based (not spatially-resolved) model of calcium-triggered processes is sufficient to reproduce all known properties of LTP and LTD induction and that neither spatially-resolved elements nor calcium-independent processes are required to predict the observed synaptic change. This could be stated more explicitly. It could also be clarified that the principal assumption underlying the proposed "geometric readout" mechanisms is that all information determining the induction of LTP vs. LTP is contained in the time-dependent spine-averaged Ca2+/calmodulin-bound CaN and CaMKII concentrations, and that no extra elements are required. Further, since both CaN and CaMKII concentrations are uniquely determined by the time course of postsynaptic Ca2+ concentration, the model implicitly assumes that the LTP/LTD induction depends solely on spine-averaged Ca2+ concentration time course, as in many prior simplified models. This should be stated explicitly to clarify the nature of the presented model.

We thank the reviewer for the suggestions on how to clarify the main hypotheses and assumptions of our work. We slightly modified the sentences provided by the reviewer and added them in the main text (page 2, lines 82 and page 19, lines 593).

3) In the Discussion, the authors appear to be very careful in framing their work as a conceptual new approach in modeling STD/STP, rather than a final definitive model: for instance, they explicitly discuss the possibility of extending the "geometric readout" approach to more than two time-dependent variables, and comment on the potential non-uniqueness of key model parameters. However, this makes it hard to judge whether the presented concrete predictions on LTP/LTD induction are simply intended as illustrations of the presented approach, or whether the authors strongly expect these predictions to hold. The level of confidence in the concrete model predictions should be clarified in the Discussion. If this confidence level is low, that would call into question the very goal of such a modeling approach.

These are very good questions. Let us first comment on the parameter uniqueness. We believe, like in E. Marder’s work on ion channels expression in neurons, that the synapse has the possibility to adapt its internal parameters (proteins number, transition rates, etc) to provide a given functioning behaviour. As a by-product, there is non uniqueness of parameters associated with behavior. Additionally, since our model is able to reproduce 9 published experimental outcomes with a single set of parameters, it is a functioning synapse with adjusted parameters which output the expected behaviours. Thus by extrapolation, our confidence in the further predictions is high. We modified sentences in the discussion section to argue this point (page 21, line 707).

Let us comment now on increasing the complexity. To our best, we strived to design a plasticity readout as simple as possible yet providing a functioning synapse. Given our success to reproduce 9 published experimental outcomes with a single set of parameters, adding more complexity would be akin to overfitting.

4) The authors presented a simplified mechanistic dynamical Markov chain process to prove that the "geometric readout" step is implementable as a dynamical process, at least in principle. However, a more realistic biochemical implementation of the proposed "region indicator" variables may be complex and not guaranteed to be robust to noise. While the authors acknowledge and touch upon some of these issues in their discussion, it is important that the authors will prove in future work that the "geometric readout" is implementable as a biochemical reaction network. Barring such implementation, one must be extra careful when claiming advantages of this approach as compared to modeling work that attempts to reconstruct the entire biochemical pathways of LTP/LTD induction.

We acknowledge this issue and agree this would be an interesting subject for future work.

Author Response

Reviewer #1 (Public Review):

1) Comment: To determine the effect of diseased monocytes on retinal health, light-injured mouse retinas were injected with monocytes isolated from AMD patients (Figure 1 - figure supplement 1). This resulted in a reduction in photoreceptor number and ERG b-wave amplitude. However, the light-injured control eye was injected with PBS only, so no cells were present. The reasoning for using this control was not provided. The appropriate injection control would include monocytes isolated from non-AMD patients. This control should be performed side-by-side with cells from AMD patients.

We thank the reviewer for this important comment. The purpose of the current study was to identify the macrophage subtype that may be associated with cell death in aAMD. We have previously reported that macrophages from AMD patient demonstrate a different phenotype compared with healthy patient in the rodent model for laser induced CNV (Hagbi-Levi S et al, 2016). Per the reviewer comment, we have performed additional experiments to assess the effect of monocytes from healthy controls in the photic retinal injury model. Results showed that monocytes from AMD and healthy patients exert different impact on the retina in this rodent model for aAMD. Interestingly, we found that monocytes from healthy patients were more neurotoxic to photoreceptors compared with monocytes from AMD patients. These results are included in the revised ms. as Figure 1- figure supplement 1H. A possible explanation for these findings is discussed in lines 179-190 of the revised manuscript. This finding reinforces the idea that the use of monocytes from AMD patients in the experiments is required to obtain a comprehensive understanding of their involvement in the progression of the disease.

2) Comment: The authors hypothesize, from the experiments presented in Figure 1 - figure supplement 1, that the injected monocytes generated macrophages in the retina, which were responsible for the observed neurotoxicity (Lines 143-145). However, no direct evidence was presented. This idea should be tested in vivo. This could be done by injecting tracer-labeled human AMD-derived monocytes into light-injured mouse retinas. If the authors' hypothesis is true, collected retinas should contain tracer-labeled cells that express macrophage markers. Tracer-labeled M2a macrophage cells should be present since subsequent experiments identify this subclass as being associated with retinal cell death.

Thank you for this important comment. To address the reviewers comment, retinal section from mice exposed to photic-retinal injury and injected with Dio-tracer labelled monocytes were stained with two M2a macrophages markers, CD206 (mannose receptor) and VEGF (Kadomoto, S et al, 2022; Jayasingam SD et al, 2019). Interestingly, we found co-localization of Dio-tracer staining (representing the injected human macrophages) with CD206 and VEGF markers in monocytes localized in different retinal layers, but not in monocytes remaining in the vitreous cavity. These data indicate that M2a markers are expressed during the polarization of monocytes into M2a phenotype which is maintained only upon entry into the retina tissue. These results were included in Figure 1- figure supplement 1K-S and discussed in the revised manuscript in lines 179-182.

3) Comment: Photoreceptor number and b-wave amplitudes were measured in light-injured retinas injected with one of four macrophage cell types generated from human AMD-derived monocytes. The authors conclude that only injection of M2a cells reduced photoreceptor number and b-wave amplitudes (Figure 1C, E). This may be true, but it is difficult for the reader to make a conclusion (especially in Fig. 1E) due to the large error bars and five different traces overlapping each other. To make these results easier to interpret, graph control cells with only one experimental sample (cell type) at a time.

Thank you for this comment. Per the reviewer comment, the graphs were modified in the revised ms. (Figure 1, panel H-K).

4) Comment: Most injected macrophages were located in the vitreous. In the case of M2a cells, the authors note that "several of the cells migrated across the retinal layers reaching the subretinal space" (Lines 167,168). One possible explanation for why M0, M1, and M2c macrophages did not induce retinal degeneration is that they did not migrate to the subretinal space and around the optic nerve head. Supplementary figures should be added to demonstrate that this is not the case.

Thank you for this comment. To address the reviewer comment we compared the migration patterns of the different macrophage phenotypes following intravitreal injection in mice exposed to photic-injury. Our results indicated that M0, M1 and M2c macrophages, similarly to M2a macrophages, migrated to the subretinal space and around the optic nerve. Thus, the neurotoxic effect of M2a is not explained by their capacity to infiltrate the retinal tissues. These results was included in Figure 1- figure supplement 2 E-H of the revised manuscript. These results are supported by our ex-vivo experiments, showing that co-culture of M2a macrophages with a retinal explants was associated with increased photoreceptor cells death compared to M1 macrophages. The results are presented and discussed in the revised manuscript in lines 200-203.

5) Comment: Figure 1 - figure supplement 2: Panel A, B cells were stained with CD206 to demonstrate the presence of M2a macrophages (panel B). The authors conclude that panel A contains M1 and panel B contains M2a cells. The lack of CD206 expression illustrates that panel A cells are not M2a macrophages but do not demonstrate they are M1 macrophages. A control using an M1 cell marker is necessary to show that panel A cells are M1 and M1 cells are not detected in M2a cultures.

Thank you for this comment. We have validated the phenotype of each macrophages subtype by qPCR (Figure 1 panel A). To further address the reviewer comment, we have performed additional immunocytochemistry for M1 macrophages using anti-CD80 antibody which is utilized as M1 macrophages marker (Bertani FR et al.2017). Results of the staining confirmed the identity of the M1 macrophages. These new results were included in Figure 1- figure supplement 2A, and are discussed in lines 168-170.

6) Comment: Ex vivo, apoptotic photoreceptor and RPE cells are observed when cultured with M2a macrophages (Figure 2). Do injected M2a cells also induce apoptosis of RPE cells in vivo? This is important to establish that retinal explants are a good model for in vivo experiments.

Thank you for this comment. To address the reviewer comment, we assessed RPE apoptosis (using TUNEL, Caspase 3 staining and RPE65 marker) after M2A cells delivery, in the in-vivo photic injury model. We could not detect apoptotic signal in the RPE layers 7 days after photic injury and therefore could not evaluate the effect of M2a macrophages on the RPE cells in-vivo (see Author response image 1). One possible explanation is that RPE cells that have undergone apoptosis are rapidly removed from the damaged tissue and are no longer detectable unlike photoreceptors. Furthermore, a study that investigated the impact of bright light on RPE cells in-vivo, showed that although RPE cells undergone structural and chemical modifications after photic-injury, TUNEL signal was not detected because RPE cell die by necrosis mechanism and not apoptosis (Jaadane I et al, 2017). Other studies validated that blue light induces RPE necrosis (Song W et al, 2022; Mohamed A et al, 2022). Taken together, it seems that ex-vivo retinal explant and in-vivo photic injury both simulate the mechanism of retinal cell death. However, the use of ex-vivo model allows for establishing the direct impact of M2a macrophages on retina in non-inflammatory context.

Author responnse image 1.

7) Comment: Reactive oxygen species (ROS) production was measured to determine if M2a cell-mediated neurotoxicity was due to oxidative stress. It is concluded that a ROS increase is partly responsible (Line 218). The data do not support this conclusion. ROS was detected in cultured M2a macrophages. More importantly, however, there was no increase in oxidative damage in vivo. The in vivo and cell culture results contradict each other so no conclusion can be made. The lack of in vivo confirmation weakens the argument that ROS drives M2a neurotoxicity. Text suggesting a role for ROS in neurotoxicity should be appropriately edited (Lines including 218, 244, 401,406,481).

Thank you for this comment. The manuscript was revised according to the reviewer suggestion (Lines 250-256).

8) Comment: The authors ask if the photoreceptor cell death is cytokine-mediated. Multiple cytokines were enriched in M2a-conditioned media. Of particular interest were CCR1 ligands MPIF1 and MCP4. The implication is that these two ligands mediate the M2a macrophages to photoreceptor cell death through CCR1. However, there is no attempt to show that either MPIF1 or MCP4 are present in vivo, or are sufficient to induce the retinal response observed. This could be demonstrated by injection of MPIF1 or MCP4. Evidence that either ligand phenocopies M2a macrophage injection would be direct evidence that CCR1 ligands activate the retinal response. Furthermore, co-injection with BX174 should block the effect of these ligands if they work through CCR1.

Thank you for this comment. The identification of CCR1 ligands expression from M2a polarized macrophages directed our decision to study CCR1 in the context of atrophic AMD. We do not claim that these specific CCR1 ligands are sufficient to activate CCR1 and exert retinal injury. The mechanism is likely more complex. Yet, to address the reviewer comment, we have performed the experiments suggested by the reviewer. Mice were exposed to photic injury and immediately injected in one eye with MPIF1, MCP-4, or a combination of both and in second eye with PBS as vehicle. Intravitreal cytokines delivery was repeated two days later (following the half-life time of these cytokines) and ERG were recorded two days after the last injection. Injection of cytokines at a concentration of 300 ng per eye did not exacerbated photoreceptor death. Then, the same experiment was repeated with two higher concentrations of cytokine, 1.2 ug/eye and 2 ug/eye, but no changes are observed between the cytokines treated-eyes and the vehicle treated-eyes. Based on previous studies reporting the physiological concentration of different cytokines in eyes of un/healthy individuals and on experiments in which different cytokines are injected in rodent eye (Estevao C et al, 2021. Zeng Y et al, 2019; Roybal CN et al, 2018; Mugisho OO et al, 2018), the cytokine concentrations used in our experiment are in the range in which effect on the retina is expected.

It is likely that a synergistic effect of M2a-secreted proteins in a particular microenvironment is necessary to increase the level of retinal damage (Bartee E et al, 2013). It is also likely that in the photic retinal injury model there is upregulation of cytokines that may mask additional delivery of exogenous cytokines. Comprehensive understanding of the complex interactions of these cytokines during retinal degeneration is beyond the scope of the current manuscript which is not focus on identifying ligand-induced CCR1 activation and its consequences. Additionally, we suggest that due to cytokine redundancy (Nicola NA; 1994), demonstrating that MPIF-4 or MCP-3 can increase photoreceptor death is not required for proving CCR1 receptor involvement.

Author Response

Reviewer #1 (Public Review):

In this work George et al. describe RatInABox, a software system for generating surrogate locomotion trajectories and neural data to simulate the effects of a rodent moving about an arena. This work is aimed at researchers that study rodent navigation and its neural machinery.

Strengths:

• The software contains several helpful features. It has the ability to import existing movement traces and interpolate data with lower sampling rates. It allows varying the degree to which rodents stay near the walls of the arena. It appears to be able to simulate place cells, grid cells, and some other features.

• The architecture seems fine and the code is in a language that will be accessible to many labs.

• There is convincing validation of velocity statistics. There are examples shown of position data, which seem to generally match between data and simulation.

Weaknesses:

• There is little analysis of position statistics. I am not sure this is needed, but the software might end up more powerful and the paper higher impact if some position analysis was done. Based on the traces shown, it seems possible that some additional parameters might be needed to simulate position/occupancy traces whose statistics match the data.

Thank you for this suggestion. We have added a new panel to figure 2 showing a histogram of the time the agent spends at positions of increasing distance from the nearest wall. As you can see, RatInABox is a good fit to the real locomotion data: positions very near the wall are under-explored (in the real data this is probably because whiskers and physical body size block positions very close to the wall) and positions just away from but close to the wall are slightly over explored (an effect known as thigmotaxis, already discussed in the manuscript).

As you correctly suspected, fitting this warranted a new parameter which controls the strength of the wall repulsion, we call this “wall_repel_strength”. The motion model hasn’t mathematically changed, all we did was take a parameter which was originally a fixed constant 1, unavailable to the user, and made it a variable which can be changed (see methods section 6.1.3 for maths). The curves fit best when wall_repel_strength ~= 2. Methods and parameters table have been updated accordingly. See Fig. 2e.

• The overall impact of this work is somewhat limited. It is not completely clear how many labs might use this, or have a need for it. The introduction could have provided more specificity about examples of past work that would have been better done with this tool.

At the point of publication we, like yourself, also didn’t know to what extent there would be a market for this toolkit however we were pleased to find that there was. In its initial 11 months RatInABox has accumulated a growing, global user base, over 120 stars on Github and north of 17,000 downloads through PyPI. We have accumulated a list of testimonials[5] from users of the package vouching for its utility and ease of use, four of which are abridged below. These testimonials come from a diverse group of 9 researchers spanning 6 countries across 4 continents and varying career stages from pre-doctoral researchers with little computational exposure to tenured PIs. Finally, not only does the community use RatInABox they are also building it: at the time of writing RatInABx has received logged 20 GitHub “Issues” and 28 “pull requests” from external users (i.e. those who aren’t authors on this manuscript) ranging from small discussions and bug-fixes to significant new features, demos and wrappers.

Abridged testimonials:

● “As a medical graduate from Pakistan with little computational background…I found RatInABox to be a great learning and teaching tool, particularly for those who are underprivileged and new to computational neuroscience.” - Muhammad Kaleem, King Edward Medical University, Pakistan

● “RatInABox has been critical to the progress of my postdoctoral work. I believe it has the strong potential to become a cornerstone tool for realistic behavioural and neuronal modelling” - Dr. Colleen Gillon, Imperial College London, UK

● “As a student studying mathematics at the University of Ghana, I would recommend RatInABox to anyone looking to learn or teach concepts in computational neuroscience.” - Kojo Nketia, University of Ghana, Ghana

● “RatInABox has established a new foundation and common space for advances in cognitive mapping research.” - Dr. Quinn Lee, McGill, Canada

The introduction continues to include the following sentence highlighting examples of past work which relied of generating artificial movement and/or neural dat and which, by implication could have been done better (or at least accelerated and standardised) using our toolbox.

“Indeed, many past[13, 14, 15] and recent[16, 17, 18, 19, 6, 20, 21] models have relied on artificially generated movement trajectories and neural data.”

• Presentation: Some discussion of case studies in Introduction might address the above point on impact. It would be useful to have more discussion of how general the software is, and why the current feature set was chosen. For example, how well does RatInABox deal with environments of arbitrary shape? T-mazes? It might help illustrate the tool's generality to move some of the examples in supplementary figure to main text - or just summarize them in a main text figure/panel.

Thank you for this question. Since the initial submission of this manuscript RatInABox has been upgraded and environments have become substantially more “general”. Environments can now be of arbitrary shape (including T-mazes), boundaries can be curved, they can contain holes and can also contain objects (0-dimensional points which act as visual cues). A few examples are showcased in the updated figure 1 panel e.

To further illustrate the tools generality beyond the structure of the environment we continue to summarise the reinforcement learning example (Fig. 3e) and neural decoding example in section 3.1. In addition to this we have added three new panels into figure 3 highlighting new features which, we hope you will agree, make RatInABox significantly more powerful and general and satisfy your suggestion of clarifying utility and generality in the manuscript directly.

On the topic of generality, we wrote the manuscript in such a way as to demonstrate how the rich variety of ways RatInABox can be used without providing an exhaustive list of potential applications. For example, RatInABox can be used to study neural decoding and it can be used to study reinforcement learning but not because it was purpose built with these use-cases in mind. Rather because it contains a set of core tools designed to support spatial navigation and neural representations in general. For this reason we would rather keep the demonstrative examples as supplements and implement your suggestion of further raising attention to the large array of tutorials and demos provided on the GitHub repository by modifying the final paragraph of section 3.1 to read:

“Additional tutorials, not described here but available online, demonstrate how RatInABox can be used to model splitter cells, conjunctive grid cells, biologically plausible path integration, successor features, deep actor-critic RL, whisker cells and more. Despite including these examples we stress that they are not exhaustive. RatInABox provides the framework and primitive classes/functions from which highly advanced simulations such as these can be built.”

Reviewer #3 (Public Review):

George et al. present a convincing new Python toolbox that allows researchers to generate synthetic behavior and neural data specifically focusing on hippocampal functional cell types (place cells, grid cells, boundary vector cells, head direction cells). This is highly useful for theory-driven research where synthetic benchmarks should be used. Beyond just navigation, it can be highly useful for novel tool development that requires jointly modeling behavior and neural data. The code is well organized and written and it was easy for us to test.

We have a few constructive points that they might want to consider.

• Right now the code only supports X,Y movements, but Z is also critical and opens new questions in 3D coding of space (such as grid cells in bats, etc). Many animals effectively navigate in 2D, as a whole, but they certainly make a large number of 3D head movements, and modeling this will become increasingly important and the authors should consider how to support this.

Agents now have a dedicated head direction variable (before head direction was just assumed to be the normalised velocity vector). By default this just smoothes and normalises the velocity but, in theory, could be accessed and used to model more complex head direction dynamics. This is described in the updated methods section.

In general, we try to tread a careful line. For example we embrace certain aspects of physical and biological realism (e.g. modelling environments as continuous, or fitting motion to real behaviour) and avoid others (such as the biophysics/biochemisty of individual neurons, or the mechanical complexities of joint/muscle modelling). It is hard to decide where to draw but we have a few guiding principles:

1. RatInABox is most well suited for normative modelling and neuroAI-style probing questions at the level of behaviour and representations. We consciously avoid unnecessary complexities that do not directly contribute to these domains.

2. Compute: To best accelerate research we think the package should remain fast and lightweight. Certain features are ignored if computational cost outweighs their benefit.

3. Users: If, and as, users require complexities e.g. 3D head movements, we will consider adding them to the code base.

For now we believe proper 3D motion is out of scope for RatInABox. Calculating motion near walls is already surprisingly complex and to do this in 3D would be challenging. Furthermore all cell classes would need to be rewritten too. This would be a large undertaking probably requiring rewriting the package from scratch, or making a new package RatInABox3D (BatInABox?) altogether, something which we don’t intend to undertake right now. One option, if users really needed 3D trajectory data they could quite straightforwardly simulate a 2D Environment (X,Y) and a 1D Environment (Z) independently. With this method (X,Y) and (Z) motion would be entirely independent which is of unrealistic but, depending on the use case, may well be sufficient.

Alternatively, as you said that many agents effectively navigate in 2D but show complex 3D head and other body movements, RatInABox could interface with and feed data downstream to other softwares (for example Mujoco[11]) which specialise in joint/muscle modelling. This would be a very legitimate use-case for RatInABox.

We’ve flagged all of these assumptions and limitations in a new body of text added to the discussion:

“Our package is not the first to model neural data[37, 38, 39] or spatial behaviour[40, 41], yet it distinguishes itself by integrating these two aspects within a unified, lightweight framework. The modelling approach employed by RatInABox involves certain assumptions:

1. It does not engage in the detailed exploration of biophysical[37, 39] or biochemical[38] aspects of neural modelling, nor does it delve into the mechanical intricacies of joint and muscle modelling[40, 41]. While these elements are crucial in specific scenarios, they demand substantial computational resources and become less pertinent in studies focused on higher-level questions about behaviour and neural representations.

2. A focus of our package is modelling experimental paradigms commonly used to study spatially modulated neural activity and behaviour in rodents. Consequently, environments are currently restricted to being two-dimensional and planar, precluding the exploration of three-dimensional settings. However, in principle, these limitations can be relaxed in the future.

3. RatInABox avoids the oversimplifications commonly found in discrete modelling, predominant in reinforcement learning[22, 23], which we believe impede its relevance to neuroscience.

4. Currently, inputs from different sensory modalities, such as vision or olfaction, are not explicitly considered. Instead, sensory input is represented implicitly through efficient allocentric or egocentric representations. If necessary, one could use the RatInABox API in conjunction with a third-party computer graphics engine to circumvent this limitation.

5. Finally, focus has been given to generating synthetic data from steady-state systems. Hence, by default, agents and neurons do not explicitly include learning, plasticity or adaptation. Nevertheless we have shown that a minimal set of features such as parameterised function-approximator neurons and policy control enable a variety of experience-driven changes in behaviour the cell responses[42, 43] to be modelled within the framework.

• What about other environments that are not "Boxes" as in the name - can the environment only be a Box, what about a circular environment? Or Bat flight? This also has implications for the velocity of the agent, etc. What are the parameters for the motion model to simulate a bat, which likely has a higher velocity than a rat?

Thank you for this question. Since the initial submission of this manuscript RatInABox has been upgraded and environments have become substantially more “general”. Environments can now be of arbitrary shape (including circular), boundaries can be curved, they can contain holes and can also contain objects (0-dimensional points which act as visual cues). A few examples are showcased in the updated figure 1 panel e.

Whilst we don’t know the exact parameters for bat flight users could fairly straightforwardly figure these out themselves and set them using the motion parameters as shown in the table below. We would guess that bats have a higher average speed (speed_mean) and a longer decoherence time due to increased inertia (speed_coherence_time), so the following code might roughly simulate a bat flying around in a 10 x 10 m environment. Author response image 1 shows all Agent parameters which can be set to vary the random motion model.

Author response image 1.

• Semi-related, the name suggests limitations: why Rat? Why not Agent? (But its a personal choice)

We came up with the name “RatInABox” when we developed this software to study hippocampal representations of an artificial rat moving around a closed 2D world (a box). We also fitted the random motion model to open-field exploration data from rats. You’re right that it is not limited to rodents but for better or for worse it’s probably too late for a rebrand!

• A future extension (or now) could be the ability to interface with common trajectory estimation tools; for example, taking in the (X, Y, (Z), time) outputs of animal pose estimation tools (like DeepLabCut or such) would also allow experimentalists to generate neural synthetic data from other sources of real-behavior.

This is actually already possible via our “Agent.import_trajectory()” method. Users can pass an array of time stamps and an array of positions into the Agent class which will be loaded and smoothly interpolated along as shown here in Fig. 3a or demonstrated in these two new papers[9,10] who used RatInABox by loading in behavioural trajectories.

• What if a place cell is not encoding place but is influenced by reward or encodes a more abstract concept? Should a PlaceCell class inherit from an AbstractPlaceCell class, which could be used for encoding more conceptual spaces? How could their tool support this?

In fact PlaceCells already inherit from a more abstract class (Neurons) which contains basic infrastructure for initialisation, saving data, and plotting data etc. We prefer the solution that users can write their own cell classes which inherit from Neurons (or PlaceCells if they wish). Then, users need only write a new get_state() method which can be as simple or as complicated as they like. Here are two examples we’ve already made which can be found on the GitHub:

Author response image 2.

Phase precession: PhasePrecessingPlaceCells(PlaceCells)[12] inherit from PlaceCells and modulate their firing rate by multiplying it by a phase dependent factor causing them to “phase precess”.

Splitter cells: Perhaps users wish to model PlaceCells that are modulated by recent history of the Agent, for example which arm of a figure-8 maze it just came down. This is observed in hippocampal “splitter cell”. In this demo[1] SplitterCells(PlaceCells) inherit from PlaceCells and modulate their firing rate according to which arm was last travelled along.

• This a bit odd in the Discussion: "If there is a small contribution you would like to make, please open a pull request. If there is a larger contribution you are considering, please contact the corresponding author3" This should be left to the repo contribution guide, which ideally shows people how to contribute and your expectations (code formatting guide, how to use git, etc). Also this can be very off-putting to new contributors: what is small? What is big? we suggest use more inclusive language.

We’ve removed this line and left it to the GitHub repository to describe how contributions can be made.

• Could you expand on the run time for BoundaryVectorCells, namely, for how long of an exploration period? We found it was on the order of 1 min to simulate 30 min of exploration (which is of course fast, but mentioning relative times would be useful).

Absolutely. How long it takes to simulate BoundaryVectorCells will depend on the discretisation timestep and how many neurons you simulate. Assuming you used the default values (dt = 0.1, n = 10) then the motion model should dominate compute time. This is evident from our analysis in Figure 3f which shows that the update time for n = 100 BVCs is on par with the update time for the random motion model, therefore for only n = 10 BVCs, the motion model should dominate compute time.

So how long should this take? Fig. 3f shows the motion model takes ~10-3 s per update. One hour of simulation equals this will be 3600/dt = 36,000 updates, which would therefore take about 72,000*10-3 s = 36 seconds. So your estimate of 1 minute seems to be in the right ballpark and consistent with the data we show in the paper.

Interestingly this corroborates the results in a new inset panel where we calculated the total time for cell and motion model updates for a PlaceCell population of increasing size (from n = 10 to 1,000,000 cells). It shows that the motion model dominates compute time up to approximately n = 1000 PlaceCells (for BoundaryVectorCells it’s probably closer to n = 100) beyond which cell updates dominate and the time scales linearly.

These are useful and non-trivial insights as they tell us that the RatInABox neuron models are quite efficient relative to the RatInABox random motion model (something we hope to optimise further down the line). We’ve added the following sentence to the results:

“Our testing (Fig. 3f, inset) reveals that the combined time for updating the motion model and a population of PlaceCells scales sublinearly O(1) for small populations n > 1000 where updating the random motion model dominates compute time, and linearly for large populations n > 1000. PlaceCells, BoundaryVectorCells and the Agent motion model update times will be additionally affected by the number of walls/barriers in the Environment. 1D simulations are significantly quicker than 2D simulations due to the reduced computational load of the 1D geometry.”

And this sentence to section 2:

“RatInABox is fundamentally continuous in space and time. Position and velocity are never discretised but are instead stored as continuous values and used to determine cell activity online, as exploration occurs. This differs from other models which are either discrete (e.g. “gridworld” or Markov decision processes) or approximate continuous rate maps using a cached list of rates precalculated on a discretised grid of locations. Modelling time and space continuously more accurately reflects real-world physics, making simulations smooth and amenable to fast or dynamic neural processes which are not well accommodated by discretised motion simulators. Despite this, RatInABox is still fast; to simulate 100 PlaceCell for 10 minutes of random 2D motion (dt = 0.1 s) it takes about 2 seconds on a consumer grade CPU laptop (or 7 seconds for BoundaryVectorCells).”

• Regarding the Geometry and Boundary conditions, would supporting hyperbolic distance might be useful, given the interest in alternative geometry of representations (ie, https://www.nature.com/articles/s41593-022-01212-4)?

Whilst this would be very interesting it would likely represent quite a significant edit, requiring rewriting of almost all the geometry-handling code. We’re happy to consider changes like these according to (i) how simple they will be to implement, (ii) how disruptive they will be to the existing API, (iii) how many users would benefit from the change. If many users of the package request this we will consider ways to support it.

• In general, the set of default parameters might want to be included in the main text (vs in the supplement).

We also considered this but decided to leave them in the methods for now. The exact value of these parameters are subject to change in future versions of the software. Also, we’d prefer for the main text to provide a low-detail high-level description of the software and the methods to provide a place for keen readers to dive into the mathematical and coding specifics.

• It still says you can only simulate 4 velocity or head directions, which might be limiting.

Thanks for catching this. This constraint has been relaxed. Users can now simulate an arbitrary number of head direction cells with arbitrary tuning directions and tuning widths. The methods have been adjusted to reflect this (see section 6.3.4).

• The code license should be mentioned in the Methods.

We have added the following section to the methods:

6.6 License RatInABox is currently distributed under an MIT License, meaning users are permitted to use, copy, modify, merge publish, distribute, sublicense and sell copies of the software.

Author Response

Reviewer #2 (Public Review):

“The authors wish to relate beat-to-beat coordination of cardiac function (in this case as measured left ventricular pressure) to the activity of sympathetic neuron spiking within the stellate ganglion. A strength includes the challenging measurements from multiple stellate neuron activity over long durations in situ in the anesthetized pig.”

We thank the reviewer for their feedback.

“A major and overriding weakness is the founding assumption of the analysis that the underlying sympathetic neurons are all cardiac functioning in nature - an assumption that is overwhelmingly unlikely given the evidence in other species including humans that stellate postganglionic neurons are functionally mixed and have functional noncardiac targets. The use of broad and poorly explained/defined terms such as "event entropy" is difficult to follow and find meaning from. The manuscript is filled with difficult-to-follow text like "The neural specificity metric (Sudarshan et al., 2021). Fig. 5", is used to evaluate the degree to which neural activity is biased toward control target states taken here as LVP" and "The neural specificity is reduced from a multivariate signal to a univariate signal by computing the Shannon entropy at each timestamp of the mapped neural specificity metric". The figures are difficult to understand with axes that often bear no units or are quite compressed obscuring the intuitive meaning of the data trends. Fundamentally, cardiac pressure cycles with each heartbeat - roughly once per second - yet fluctuations in the depicted mean spike rate data with changes perhaps ten times in 25 minutes. Such plots are disorienting and difficult to associate with cardiac or neuron "functioning". Only 17 of the 38 references are not self-citations and thus the cited literature represents a narrow view of sympathetic regulation and sympathetic/stellate ganglion knowledge. Much of the foundations are self-professed in earlier publications by the present group and assumed to be accepted.”

“Fundamentally, cardiac pressure cycles with each heartbeat - roughly once per second - yet fluctuations in the depicted mean spike rate data with changes perhaps ten times in 25 minutes. Such plots are disorienting and difficult to associate with cardiac or neuron "functioning”

We would like to clarify this point with the understanding that the reviewer is referring to the time axis in Figure 3C in the manuscript.

The coactivity matrix constructed in Figure 3C computes the cross correlation in sliding mean/std spike activities for different pairs of channels. The mean spiking activities across channels, as the reviewer correctly pointed out, do indeed have a weak autocorrelation with the period of the heart rate. The weak correlation for the heart rate period, possibly due to slow firing rates, was seen across all channels of both control and HF animals. But, the cause of a large proportion of channel-pairs exhibiting high coactivity, termed as cofluctuation (Shown as red tracings in Fig 3D), is not known and cannot be directly associated with cardiac functioning.

The cofluctuation was also found to be aperiodic in nature approximating a lognormal distribution (Fig R1) with the HF animals containing heavy tails outside their confidence intervals (Fig R1B). The event rate computed from the cofluctuation time series (shown as blue steps in Fig 3E) for an animal is a measure of spatial coherence among SG neural populations and was developed as a novel metric to be used in future studies.

Figure R1: Cofluctuation histograms (calculated from mean or standard deviation of sliding spike rate, referred as Cofluctuation_MEAN and Cofluctuation_STD, respectively) and log-normal fits for each animal group. μF IT and σF IT are the respective mean and standard deviation (STD) of fitted distribution, used for 68% confidence interval bounds. A-B: Control animals have narrower bounds and represent a better fit to log-normal distribution. C-D: Heart failure (HF) animals display more heavily skewed distributions that indicate heavy tails.

“Only 17 of the 38 references are not self-citations and thus the cited literature represents a narrow view of sympathetic regulation and sympathetic/stellate ganglion knowledge. Much of the foundations are self-professed in earlier publications by the present group and assumed to be accepted.”

We thank the reviewer for pointing this out. We have added four additional citations that include methods such as neural population bias and spatiotemporal dynamics linkages to control targets in the neuroscience literature. We have added these citations to page 15 in the “Conclusion” section of the manuscript. In addition, it is our group’s specialty to carry these cardiac nervous system experiments, we are not aware of another group collecting multi-electrode array data from the cardiac nervous system and studying population dynamics of cardiac neurons. Hence we build on based on our previous learnings. The most relevant literature (not necessarily related to cardiac nervous system) can be found in the neuroscience references we cited that contain applications of neural population recordings for different brain areas, mainly in neuropsychiatry domain to understand disease dynamics.

“For the expert or even the uninformed reader, this report is broadly confused and confusing. The premises (beat to beat or whether LVP conveys cardiac function) are poorly supported. The conclusions are quite vague.”

Thank you for your feedback. To simplify the understanding, we moved all mathematical details to supplementary material, re-wrote the abstract and the conclusion from scratch, and splitted the methods figures that may be confusion. We believe that our novel metrics event rate and entropy capture non-trivial linkages between heart failure status, cardiac neural activity (spike activity), and peripheral activity (LVP). We have supported our metrics with 17 animals with state-of-the-art surgical techniques and technology, and reported our results with detailed statistical analyses. Our manuscript essentially highlights that event rate and entropy metrics are significantly different between control animals and animals with heart failure. These metrics can be used to design future studies with these animal models to provide a more quantitative approach to heart disease, rather than binary (yes or no) descriptions.

“Discussion: The abstract does not convey conclusions from the findings and contains broad statements such as "signatures based on linking neuronal population cofluctuation and examine differences in "neural specificity" of SG network" that have little substantive value or conclusion for the reader. Fundamentally what does the title "signatures based on linking neuronal population" cofluctuation mean to the reader? What changed in HF?”

Thank you for this comment. We completely revised the abstract and conclusion as detailed in our response to Essential Revision #1. Event rate is a metric related to neural activity recordings and entropy is related to the association of neural activity to left ventricular blood pressure. Our findings suggest that both the neural population activity itself (event rate) and its ability to pay attention to cycles of left ventricular pressure (neural specificity) are significantly higher in animals with HF compared to controls.

Author Response

Reviewer #1 (Public Review):

(1.1) The work by Porciello and colleagues provides scientific evidence that the acidic content of the stomach covaries with the experienced level of disgust and fear evoked by disgusting videos. The working of the inside of the gut during cognitive or emotional processes have remained elusive due to the invasiveness of the methods to study it. The major strength of the paper is the use of the non-invasive smart pill technology, which senses changes in Ph, pressure and temperature as it travels through the gut, allowing authors to investigate how different emotions induced with validated video clips modulate the state of the gut. The experimental paradigm used to evoke distinct emotions was also successful, as participants reported the expected emotions after each emotion block. While the reported evidence is correlational in nature, I believe these results open up new avenues for studying brain-body interactions during emotions in cognitive neuroscience, and future causal manipulations will shed more insight on this phenomena. Indeed, this is the first study to provide evidence for a link between gastric acidity and emotional experience beyond single patient studies, and it has major implications for the advancement of our understanding of disorders with psycho-somatic influences, such as stress and it's influence of gastritis.

1.1 First of all, we want to thank Reviewer#1 for his cogent comments and for highlighting that our findings may inspire future research on brain-body interactions. We took into the highest consideration all the remarks and changed the manuscripts accordingly.

(1.2) As for the limitations, little insight is provided on the mechanisms, time scales, and inter-individual variability of the link between gastric Ph and emotional induction. Since this is a novel phenomena, it would be important to further validate and characterize this finding. On this line, one of the most well known influences of disgust on the gut is tachygastria, the acceleration of the gastric rhythm. It would be important to understand how acid secretion by disgusting film is related to tachygastria, but authors only examine the influence of disgusting film on the normogastric frequency range.

1.2 We are aware that at the moment our data are mainly descriptive and do not provide a clear picture of the causal mechanisms. However, to deal with this outstanding issue we added a new series of analysis.

Most of the data on gastric activity come from analysis of the normogastric band. However, information about the EGG tachygastric rhythm in humans is of potential great importance. To deal with the reviewer’s comment and considering the previously published literature, we re-examined the EGG data focusing on the tachygastric rhythm. The methodology remained consistent with the process described for normogastric peak extraction but this time, we extracted the peak in the tachygastric band, specifically 0.067 to 0.167 Hz (i.e., 4–10 cpm). The ANOVA performed over the tachygastric cycle revealed a significant main effect of the type of video clip (F(4, 112) = 2.907, p = 0.025, Eta2 (partial) = 0.09). However, the Bonferroni corrected post hoc tests did not show any significant difference between the different type of emotional video clips and the neutral condition. The sole significant comparison was observed between participants viewing happy and fearful video clips, indicating that participants’ tachygastric cycles were faster when exposed to happy rather than fearful video clips (p = 0.038). For a visual representation of the outcomes, please see Fig S6.

We revised the main text (Page 17, lines: 472-482) to include this analysis. The revised text now reads as follows:

“Finally, we explored whether normogastric and/or tachygastric cycle changed in response to specific emotional experience. After checking that normogastric and tachygastric peak frequencies were normally distributed (all ps > 0.05), we ran two separate ANOVAs on the individual peak frequencies in the normogastric and tachygastric range. Each analysis had the type of video clip as within-subjects factor. The ANOVA performed on the normogastric rhythm was not significant (F(4, 44) = 1.037, p = 0.399) suggesting that the gastric rhythm did not change while participants observed the different emotional video clips. In contrast, the ANOVA performed on the tachygastric rhythm did show a significant main effect (F(4, 112) = 2.907, p = 0.025, Eta2 (partial) = 0.09). However, the only comparison that survived the Bonferroni correction was the one between happy and fearful video clips, namely participants’ tachygastric cycle was faster when they observed happy vs fearful video clips (p = 0.038) see Fig. S6 for a graphical representation of the results.”

To deal with the Reviewer’s comment, we also correlated the average pH value with the corresponding frequency of the tachygastric cycle recorded in the disgusting, happy and the fearful video clips, namely the emotions associated to changes in pH. The only significant correlation was the one found during the disgusting video clips (r= 0.435; p= 0.023, all the other rs ≤ 0.351, all the other ps ≥ 0.073). Differently from what we expected, we found a positive correlation suggesting that when participants were exposed to disgusting video clips the less acidic was the pH the higher was the frequency of the tachygastric cycle. Instead, we know from our pill data that disgusting video clips are associated to more acid values, and from literature (not replicated by us) to a faster gastric rhythm. Since we did not find strong support in the EGG analysis suggesting a relationship between the gastric rhythm and the emotional experience, we believe that additional evidence will help to clarify the relationship between pH and gastric rhythm.

(1.3) Additionally, only one channel of the electrogastrogram (EGG) was used to measure the gastric rhythm, and no information is provided on the quality of the recordings. With only one channel of EGG, it is often impossible to identify the gastric rhythm as the position of the stomach varies from person to person, yielding inaccurate estimates of the frequency of the gastric rhythm.

1.3 We agree with Reviewer 1 on this point. We acknowledge the potential limitation associated with one-channel EGG recording in our study. To deal with this remark, in a separate (ongoing) study (N# participants= 25) we recorded the electrogastrogram following the methodology outlined by Wolpert et al., 2020 published on Psychophysiology. Thus, in order to study the EGG in association to the emotional experience, we used a bipolar 4-channels montage while participants observed the same emotional video clips used in our current study (see picture below for the montage set-up).

Author response image 1 shows the 4-channels EGG bipolar recording montage reproducing the one proposed by Wolpert et al., 2020.

Author response image 1.

Then, we extracted the gastric cycle in both the normogastric and the tachygastric bands.

After checking that data were normally distributed (Kolmogorov-Smirnov ds > 0.10; ps> .20), in the case of the gastric cycle extracted in the normogastric band, we ran a repeated measures ANOVA with the type of video clip as the only within-subjects factor measured on the 5 levels (i.e. the five types of video clip: Disgusting, Fearful, Happy, Neutral, and Sad). The ANOVA shows that the gastric cycle recorded during the different video clips did not differ (F (4,96) = 0.39; p= 0.81), see the plot on Author response image 2.

Author response image 2.

Gastric cycle (normogastric band) recorded via multiple-channels electrogastrogram (EGG) during the emotional experience. The plot shows the gastric cycle extracted in the normogastric band while participants were observing the five categories of the video clips (i.e. those inducing disgust, fear, happiness, sadness and, as control, a neutral state).

We also extracted the gastric cycle in the tachygastric band, the distribution of the data was not normal in one condition (Kolmogorov-Smirnov ds > 0.27; p < 0.05), therefore we ran a Friedman ANOVA to compare the gastric cycle during the different emotional experiences. The Friedman ANOVA was not statistically significant (χ2 (4) = 2.88; p = 0.58), suggesting that, similarly to the gastric cycle extracted in the normogastric band, also the one extracted in the tachygastric band was not clearly associated to the investigated emotional states, see Author response image 3.

Author response image 3.

Gastric cycle (tachygastric band) recorded via multiple-channels electrogastrogram (EGG) during the emotional experience. The plot above shows the gastric cycle extracted in the tachygastric band while participants were observing the five categories of the video clips (i.e. those inducing disgust, fear, happiness, sadness and as control a neutral state).

Results from this control study seem to suggest that the non-significant effect of the gastric cycle was probably not due to the fact that we use a one-channel egg montage, at least for what concerns the gastric cycle extracted from the normogastric band.

For what concerns the tachygastric frequency associated to the emotional experience these results from a multi-channel EGG recording seem to go in the same direction of the normogastric one, namely no frequency of the gastric cycles recorded during the emotional video clips was different from the control condition.

The only significant difference that we found in our 1-channel EGG study was the one between the happy and the fearful video clips (see Fig. S6 contained in the supplementary materials and above). Specifically, we found that happy video clips were associated to higher gastric frequency compared to the fearful ones. However, we did not replicate these findings in our multi-channels EGG study.

Although suggestive, this evidence is not conclusive. Indeed, we are aware that a final word on the results of our multi-channel study can be said only when a larger sample is obtained.

(1.4) Finally, I believe that the results do not show evidence in favor of the discrete nature of emotions theory as they claim in the discussion. Authors chose to use stimuli inducing discrete emotions, and only asked subjective reports of these same discrete emotions, so these results shed no light on whether emotions are represented discretely vs continuously in the brain.

We revised the discussion in order to better describe our results and toned down the interpretation that the present findings directly support the discrete nature of emotions, as suggested by this Reviewer.

Now page 21&22 lines 622-631 reads as follow:

“Overall, and in line with theoretical and empirical evidence (Damasio, 1999; Harrison et al., 2010; James, 1994, Lettieri et al., 2019; Stephens et al., 2010), our findings may suggest that specific patterns of subjective, behavioural, and physiological measures are linked to unique emotional states...We acknowledge that our results, although novel, are restricted to a sample of male participants, and more importantly they need to be replicated. We also acknowledge that future studies should better investigate the mechanisms underlying the role of the pH in the emergence of specific emotion. For instance, pharmacologically manipulating stomach pH during emotional induction, not only for basic emotions but also for exploring complex emotions such as moral disgust (Rozin et al., 2009), would enable researchers to generalize these findings and examine the directionality of this relationship.”

Reviewer #2 (Public Review):

To measure the role of gastric state in emotion, the authors used an ingestible smart pill to measure pH, pressure, and temperature in the gastrointestinal tract (stomach, small bowel, large bowel) while participants watched videos that induced disgust, fear, happiness, sadness, or a control (neutral). The study has a number of strengths, including the novelty of the measurement (very few studies have ever measured these gut properties during emotion processing) and the apparent robustness of their main finding (that during disgusting video clips, participants who experienced more feelings of disgust (and to a lesser degree which might not survive more stringent multiple comparison correction, fear) had more acidic stomach measurements, while participants who experienced more happiness during the disgusting video clips had a less acidic (more basic) stomach pH. Although the study is correlational (which all discussion should carefully reflect) and is restricted to a moderately-sized, homogenous sample, the results support their general conclusion that stomach pH is related to emotion experience during disgust induction. There may be additional analyses to conduct in order for the authors to claim this effect is specific to the stomach. Nevertheless, this work is likely to have a large impact on the field, which currently tends to rely on noninvasive measures of gastric activity such as electrogastrography (which the authors also collect for comparison); the authors' minimally-invasive approach yields new and useful measurements of gastric state. These new measures could have relevance beyond emotion processing in understanding the role of gut pH (and perhaps temperature and pressure) in cognitive processes (e.g. interoception) as well as mental and physical health.

We are very grateful to Reviewer#2 for skilfully managing the paper and highlighting its strengths, particularly the innovative measurement approach and the potential implications these findings might offer for future research into the impact of gastric signals on emotional experiences and potentially on many other higher-order cognitive functions. Additionally, we would like to thank her for the highly valuable feedback. We have incorporated all the comments into the revised manuscript, aiming to enhance its quality.

Reviewer #3 (Public Review):

This study used novel ingestible pills to measure pH and other gastric signals, and related these measures to self-report ratings of emotions induced by video clips. The main finding was that when participants viewed videos of disgust, there was an association between gastric pH and feelings of disgust and fear, and (in the opposite direction) happiness. These findings may be the first to relate objective measures of gastric physiology to emotional experience. The methods open up many new questions that can be addressed by future studies and are thus likely to have an impact on the field.

We thank very much also Reviewer#3 for the accurate reading of our manuscript; for highlighting the strengths of our study; and for providing valuable feedback. Below, a point-by-point response to all the comments raised by this Reviewer. We have incorporated their comments, and we hope they are satisfied by the new version of the manuscript.

(3.1) My main concern is with the reliability of the results. The study associates many measures (pH, temperature, pressure, EGG) in stomach, small bowel, and large bowel with multiple emotion ratings. This amounts to many statistical tests. Only one of these measures (pH in the stomach) shows a significant effect. Furthermore, the key findings, as displayed in Figure 4 do not look particularly convincing. Perhaps this is a display issue, but the relations between stomach pH and Vas ratings of disgust, fear, and happiness were not apparent from the scatter plot and may be influenced by outliers (e.g., happiness).

3.1 We thank Reviewer#3 for raising this issue which was also raised by Reviewer#1 and #2, se replies above. As reported above we worked on the data analysis in order to provide more evidence supporting our claim, i.e. that pH plays a role in the emotional experience of disgust, happiness and fear. We modified Figure 4 (now 5) as also requested by Reviewer 1 and 2, and we now hope that it is clearer. We included a new analysis, in which we used all the datapoints recorded from the ingestible device and we performed a mixed models analysis with pH as dependent variable, type of video clips and number of datapoints (‘Time’) as fixed factors, and the by-subject intercepts as random effects. This analysis not only supported the results of the original one but provided evidence for a causal role of the emotional induction on the pH of the stomach. Results of this analysis are described in point 1.7 in the response to Reviewer#1 and results of the new analysis and the revised version of the main figure can be found in track change in the manuscript (Page 15&16, lines: 408-439) in the main text and copied and pasted below.

“To explore how the emotional induction could modulate the pH of the stomach and how the length of the exposure to that specific emotional induction could also play a role in modulating pH variations, we ran an additional model, Model 2. This model included all the pH datapoints registered using the Smartpill as dependent variable, the type of video clip and the number of the datapoints (“Time”) as fixed effects, and the by-subject intercepts as random effects (see Supplementary information for a detailed description of the model). Model 2 had a marginal R2 = 0.014 and a conditional R2 = 0.79. Visual inspection of the plots did reveal some small deviations from homoscedasticity, visual inspection of the residuals did not show important deviations from normality. As for collinearity (tested by means of vif function of car package), all independent variables had a GVIF^(1/(2*Df)))^2 < 10.

Type III analysis of variance of Model 2 showed a statistically significant main effect of the Time (F = 20.237, p < 0.001, Eta2 < 0.01) suggesting that independently from the type of video clip observed, the stomach pH significantly decreased as a function of the time of exposure to the induction. A significant main effect of the type of video clip was also found (F = 22.242, p < 0.001, Eta2 = 0.01) suggesting that pH of the stomach changes when participants experienced different types of emotions. In particular, post hoc analysis revealed that pH was more acidic when participants observed disgusting compared to fearful (t= -11.417; p < 0.001), happy (t= -15.510; p < 0.001) and neutral (t= -3.598; p = 0.003) video clips.

Also, pH was more acidic when participants observed fearful compared to happy (t= -4.064; p < 0.001), and less acidic compared to neutral (t= 7.835; p < 0.001) and sad scenarios (t= 9.743; p < 0.001). Finally, pH was less acidic when participants observed happy compared to neutral (t= 11.923; p < 0.001). and sad videoclips (t= 13.806; p < 0.001), see Fig.6, left panel. Interestingly, also the double interaction Time X Type of video clip was significant (F = 3.250, p = 0.0113, Eta2 < 0.01) suggesting that the time of the exposure to the induction differentially influenced the pH of the stomach depending on to the type of the observed video clip. Simple slope analysis showed that while pH did not change over time when observing disgusting (t= -1.2691; p = 0.2045) and happy (t= 0.4466; p = 0.6552) clips, it did significantly decrease over time when observing fearful (t= -4.4212; p < 0.001), sad (t= -2.0487; p = 0.0405) and neutral video clips (t= -2.7956; p = 0.0052), see Fig.6, right panel."

We believe that the new evidence reported provides support of our claims and we hope that the reviewer agrees with us. However, as we also mentioned in the paper, we are aware that replications are needed and we are already working on this.

Author Response

Reviewer #1 (Public Review):

This study provided evidence to interpret and understand the aging and developmental processes in children. The main strength of the study is it measures a set of biological age measures and a set of developmental measures, thus providing multi-faceted evidence to explain the associations between aging and development in children. The main weakness of this study is that how to measure and test the aging hypothesis of "a buildup of biological capital model" and "wear and tear" is not well-explained. Why the observed associations between biological age measures and developmental measures could support the aforementioned aging theories?

Thank you. On reflection we agree that how to test the aging hypotheses of "a buildup of biological capital model" and "wear and tear" is not well-explained in the manuscript. We have addressed this issue in the point-by-point responses below:

1) Abstract - conclusion: The aging hypothesis of "a buildup of biological capital model" and "wear and tear" were mentioned in the conclusion without an explanation of these theories in the previous section. Readers who are not experts in the field may not understand the logic.

We have replaced these phrases in the abstract with the following interpretation, which we hope will be more readily understood:

“Patterns of associations suggested that accelerated immunometabolic age may be beneficial for some aspects of child development while accelerated DNA methylation age and telomere attrition may reflect early detrimental aspects of biological ageing, apparent even in children.”

2) Result - Biological age marker performance: the correlation between transcriptome age and chronological age is very strong (r =0.94). I am afraid that very little age-independent information could be captured by the transcriptome age. Is it possible to down-regulate the age dependency of the transcriptome age in the training process?

Thank you for this important comment: We agree the high accuracy of this clock may in fact reduce its relevance as a biological age marker and note that this is a concern generally in the field. We have explored the possibility of using a less accurate transcriptome age model as follows: Instead of elastic net modelling we tested using the lasso penalisation only, which will result in more parsimonious (sparse) models as less important features are dropped as the strength of the lambda parameter is increased. Plotting the correlation in the test set against number of features in models, as the lambda is sequentially increased, we can see (as shown in Author response image 1 by the blue line) that after the inclusion of around 200 features, the gain in accuracy becomes less steep.

Author response image 1.

We then tested the sensitivity of a model optimised for sparsity at the expense of some prediction accuracy, selected based on visual inspection (blue line, r in test set =0.87, number of features= 187) of the above plot, against developmental measures, compared to the most accurate model as presently included in the manuscript:

Author response image 2.

We find that, across all outcomes tested, the less accurate model, based on only the most important features, does not provide an improvement in sensitivity to developmental outcomes compared to the currently used model.

We therefore prefer to keep the more accurate model in this study. Especially as it is consistent with the methodology used in the Horvath and Immunometabolic age models and generally in the field, and otherwise it is not obvious how the biological clock should be trained (especially for children without mortality data) without altering the whole approach of the study. We have acknowledged and discussed this issue on page 15.

3) The study population comes from several cohorts, which might influence the results. How the cohort effects were controlled for in the analyses?

The possible influence of cohort is a limitation of the study which we have discussed on page 16. We did not include cohort as a predictor in any of the candidate biological clocks since this may reduce detection of some age -related features. Instead, we include a variable for cohort as a fixed effect in all analyses with risk factors and developmental outcomes and examined the performance of candidate biological clocks in predicting chronological age within each cohort. As a further check, we have added an additional sensitivity analysis (Figure 4-figure supplement 6), against developmental outcomes significant in the main analysis, stratified by cohort. We find generally consistent effects across cohorts.

4) Figure 3 only showed the number of p values. Can the author also provide the number of point estimates and 95% confidence intervals, perhaps in the supplemental table?

This information was originally provided in supplemental table 5 (now Supplementary file 7), combined with the sensitivity analyses. To make this information easier to find, we have made this a stand-alone table (table 3). We now direct readers to this information within the caption of Figure 4 (previously figure 2).

Reviewer #2 (Public Review):

The study had an especially relevant aim for aging research and utilized various data types in an especially interesting human population. Multi-omics perspective adds great value to the work. The researchers aimed to evaluate how different indicators of biological age (BA) behave in children during their developmental stage. In the analysis, relationships between indicators of BA, health risk factors, and developmental factors were assessed in cross-sectional data comprising children aged 5-12 years. The manuscript is well-written and easy to follow. The methodology is good. The authors succeeded to reach the aim in most parts.

In the study, previously known and unknown biological age indicators were used. Known indicators included telomere length and Horvath's epigenetic age. Unknown (novel) indicators, transcriptomic and immunometabolic clocks, were developed in the present study and they showed a strong correlation with calendar age in this population, also in the validation data set. Although the transcriptomic and immunometabolic clocks have the potential of being true indicators of biological age, they are still lacking scientific evidence of being such indicators in adults. That is, their associations with age-related diseases and mortality are yet to be shown. Thus, the major remark of the study relates to the phrasing: these novel transcriptomic and immunometabolic clocks should be presented as BA indicator candidates waiting for the needed evidence.

Thank you for this important observation. However, we still find that “biological age indicator” is a useful umbrella term in this manuscript and there is not an obvious alternative. We therefore have added the following sentence on page 8, and highlighted the difference between the markers at key points in the abstract, introduction, results and discussion.

“We note that since a common definition of markers of biological age is that they should be associated with age-related disease and mortality [69] these new clocks may only currently be considered “candidate” biological age markers. However, we have referred to both the established and candidate markers as biological age markers throughout to simplify presentation.”

Author Response

Reviewer #1 (Public Review)

[...] One potential issue is that the high myelination signal is associated with the compartment in V2 (pale stripes) which was not functionally defined itself but by the absence of specific functional activations. No difference was reported between those stripes that were defined functionally. Other explanations for the differential pattern of a qMRI signals, e.g. ROI distribution for presumed pale stripes is not evenly distributed (more foveal), ROIs with low activations due to some other factor show higher myelin-related signals, cannot be excluded based on the analysis presented.

Indeed, it would have been advantageous to directly functionally delineate pale stripes in V2. Since we were not able to achieve this by fMRI, we needed an indirect method to infer pale stripe contributions in the analysis. We also added a statement in the discussion section to emphasize this more (p. 9, lines 286–288).

Furthermore, different myelination between thin and thick stripes was not tested, since we did not have a concrete hypothesis on this. Despite the conflicting findings of stronger myelination in dark or pale CO stripes in the literature, no histological study stated myelination differences between dark CO thin and thick stripes. Therefore, our primary interest and hypothesis was lying in comparing the different myelination of thin/thick and pale stripes using MRI.

Thank you very much for this comment about potential other sources of differential qMRI parameter patterns. Indeed, based on the original analysis we could not exclude that the absence of functional activation around the foveal representation may have biased our analysis. We therefore added a supporting analysis, in which we excluded the region around the foveal representation from the analysis. The excluded cortical region was kept consistent between participants by excluding the same eccentricity range in all maps. We added more details in the results section of the revised manuscript (p. 8, lines 189–202). In Figure 5-Supplement 1 and Figure 5-Supplement 3, results from this supporting analysis are shown which reproduced the primary findings from the main analysis, particularly the relatively higher myelination of pale stripes.

ROI definitions solely based on fMRI activation amplitude have additional limitations. However, we find it unlikely that a small fMRI effect size and low contrast-to-noise ratio (i.e. stochastic cause of low statistical parameter values/”activation”) has impacted the results, since Figure 3 shows that we could achieve a high degree of reproducibility for each participant.

We would note that the fact that we found consistent differences across MPM and MP2RAGE sessions makes some potential artifacts driving the differences unlikely. We also find it unlikely that systematic cerebral blood volume differences between stripes would have driven the results. A higher local blood volume would lead to increased BOLD responses but also to a higher R1 value due to the deoxy-hemoglobin induced relaxation, which is opposite to the observation of higher activity in the thick/thin stripes but lower R1 values.

Further studies using other functional metrics (e.g. VASO, ASL etc.) may help us to even more clearly demonstrate specificity but were out of the scope of this already rather extensive study. Although we have added extensive further analyses in the revised manuscript such as controlling for foveal effects or registration performance, we did not see a possibility to fully exclude a systematic bias that might potentially be caused by unknown factors.

Another theoretical and practical issue is the question of "ground truth" for the non-invasive qMRI measures, as the authors - as their starting point - roundly dismiss direct histological tissue studies as conflicting, rather than take a critical look at the merit of the conflicting study results and provide a best hypothesis. If so, they need to explain better how they calibrate their non-invasive MR measurements of myelin.

We agree and have now further elaborated on the limits of specificity of the R1 and R2* signal as cortical myelin marker (p. 2, lines 68–88; p. 6, line 163; p. 8, line 216; p. 9, lines. 257–260). However, we still think that it is important for the reader to appreciate the conflicting results in histological studies using staining methods for myelin, which adds to the study’s background.

We did not intend to give the impression that MRI provides the missing ground-truth to adjudicate histological controversies, but that it provides an alternative and additional view on the open questions. We changed the introduction to better reflect the aspect that the study offers a unique view by providing myelination proxies and functional measures in the same individual, which allows for direct comparison and investigation of structure-function relationships (see p. 2, lines 68–70; p. 3, lines 93–95), which is not accessible to any other approach. Nevertheless, we would like to note that R1 has been well established as a myelin marker under particular conditions (Kirilina et al., 2020; Mancini et al., 2020; Lazari and Lipp, 2021). It has also been widely used for cortical myelin mapping across a variety of populations, systems and field strengths. We added this statement to the introduction (see p. 2, lines 82-85). We note that we excluded volunteers with pathologies or neurological disorders from the study and their mean age was about 28 years. Thus, we had conditions comparable to previous (validation) studies.

Because of the contradictory findings of histological studies, we could not further finesse the hypothesis beyond our previous a priori hypothesis that we expected differences in the myelin sensitive MRI metrics between the thin/thick versus pale stripes. To improve the contextual understanding, we added a paragraph in the discussion section covering in more depth how the MRI results relate to known histological findings (see pp. 8–9, lines 216–240).

While this paper makes an important contribution to the question of the association of specific myelination patterns defining the columnar architecture in V2, it is not entirely clear whether the authors can fully resolve it with the data presented.

Indeed, we agree that non invasive aggregate measures, such as the R1 metrics, offer limited specificity which precludes a fully conclusive inference about cortical myelination. We have further emphasized this on several occasions in the text (see p. 2, lines 68–88; p. 6, line 163; p. 8, line 216; p. 9, lines. 257–260). Since the correspondence of cortical myelin levels and R1 (and other metrics) is an active area of research, we expect that the understanding, sensitivity and specificity of R1 to cortical myelination will further improve. We note that the use of qMRI is a substantial advance over weighted MRI typically used, which suffers from lack of specificity due to instrumental idiosyncrasies and varying measurement conditions.

Reviewer #2 (Public Review)

[...] Unfortunately, this particular study seems to fall into an unhappy middle ground in terms of the conclusions that can be drawn: the relaxometry measures lack the specificity to be considered "ground truth", while the authors claim that the literature lacks consensus regarding the structures that are being studied. The authors propose that their results resolve whether or not stripes differ in their patterns of myelination, but R1 lacks the specificity to do this. While myelin is a primary driver of relaxation times in cortex, relaxometry cannot be considered to be specific to myelin. It is possible that the small observed changes in R1 are driven by myelin, but they could also reflect other tissue constituents, particularly given the small observed effect sizes. If the literature was clear on the pattern of myelination across stripes, this study could confirm that R1 measurements are sensitive to and consistent with this pattern. But the authors present the work as resolving the question of how myelination differs between stripes, which over-reaches what is possible with this method. As it stands, the measured differences in R1 between functionally-defined cortical regions are interesting, but require further validation (e.g., using invasive myelin staining).

We agree that we have inadvertently overstated the specificity of R1 at several occasions in the text. We therefore toned down the statements concerning the correspondence between R1 and myelin throughout the manuscript (e.g. see p. 2, lines 68–88; p. 6, line 163; p. 8, line 216; p. 9, lines. 257–260).

We also removed the phrase that gave the impression that MRI can conclusively resolve the conflicting results found in histological studies. In the Introduction, we changed the corresponding paragraph by emphasizing the alternative view, which can be obtained from MRI by the possibility to investigate structure-function relationships in the living human brain, which would not be possible by invasive myelin staining (see p. 2, lines 68–70; p. 3, lines 93–95).

We acknowledge that – perhaps aside from electron microscopy – all common markers have shortcomings, which limit their specificity. For example, classic histology is not quantitative and resulted in conflicting results. It even includes the very fundamental issue, that the composition of myelin varies across the brain and within brain areas significantly (e.g., its lipid composition (González de San Román et al., 2018)). Thus, we regard the different invasive/non-invasive measures as complementary. R1 adds to this arsenal of measures and can be acquired non invasively. It has been shown to be a reliable myelin marker under certain circumstances. It follows the known myeloarchitecture patterns of the human brain, which was also checked for the data of the present study (see Figure 4 and Appendix 2). It is responsive to traumatic changes (Freund et al., 2019), development (Whitaker et al., 2016; Carey et al., 2018; Natu et al., 2019) and plasticity (Lazari et al., 2022). Since we studied healthy volunteers with no known pathologies that were sampled randomly from the population, we believe that the previous results generally apply and suggest sufficient specificity of the R1 marker. Of course, we cannot fully exclude bias due to unknown factors that have not been investigated/discovered by validation studies yet. However, in this case we expect that the systematic differences between stripe types would remain an important result most likely pointing to another interesting biological difference between stripes.

While more research is needed to clarify the precise role of R1 for cortical myelin, we think that the meaningful determination of quantitative MR parameter within one cortical area is still interesting for the neuroscientific community.

Moreover, the results make clear that R1 differences are not sufficiently strong to provide an independent measure of this structure (e.g., for segmentation of stripe). As such, one would still require fMRI to localise stripes, making it unclear what role R1 measures would play in future studies.

Indeed, the observed small effect sizes in the present study still requires a functional localization with fMRI. We expected small effect sizes using R1 and R2* due to the known small inter-areal or intra-cortical differences of MRI myelin markers. Therefore, this study aimed at a proof-of-concept investigating whether intra-areal R1 differences at the spatial scale of columnar structures can be detected using non-invasive MRI. Our study shows that these differences can be seen but currently not at the single voxel level. We anticipate that with further improvements in sequence development and scanner hardware, high-resolution R1 estimates with sufficient SNR can be acquired making fMRI redundant (for this kind of investigations). Please see the reply to the next comment concerning the impact of using R1 in future studies.

The Introduction concludes with the statement that "Whereas recent studies have explored cortical myelination ... using non-quantitative, weighted MR images... we showed for the first time myelination differences using MRI on a quantitative basis". As written, this sentence implies that others have demonstrated that simpler non-quantitative imaging can achieve the same aims as qMRI. Simply showing that a given method is able to achieve an aim would not be sufficient: the authors should demonstrate that this constitutes an important advance.

Thank you for this comment. It goes to the heart of the concerns raised about specificity and sensitivity of MRI based myelin metrics. We elaborate here on the main advantage of using qMRI in our current study and why it is more specific than weighted MR imaging. However, we emphasize that a thorough comparison between qMRI and weighted MRI is highly complex and refer to our recent review paper on qMRI for further details (Weiskopf et al., 2021), which are beyond the scope of our paper. The signal in weighted MRI, even when optimally optimized to the tissue of interest, additionally depends on both inhomogeneities in the RF transmit and receive (bias) fields. Other methods like using a ratio image (T1w/T2w) can cancel out the receive field bias entirely (in the case of no subject movements between scans) but not the transmit field bias. This hampers the direct analysis and interpretation of signal differences between distant regions of the brain. For high resolution imaging applications, the usage of high magnetic fields such as 7 T is beneficial or even mandatory due to signal-to-noise (SNR) penalties. With increasing field strength, these inhomogeneities also apply to small regions as V2. For these cases, qMRI is advantageous since it provides metrics which are free from these technical biases, significantly improving the specificity. As high-field MRI has the potential to non invasively study the structure and function of the human brain at the spatial scale of cortical layers and cortical columns, we believe that the results of our current study, which successfully demonstrate the applicability of qMRI to robustly detect small differences at the level of columnar systems, is relevant for future studies in the field of neuroscience.

We emphasized these considerations in the revised manuscript (see. p. 9, lines 273–285).

The study includes a very small number of participants (n=4). The advantage of non-invasive in-vivo measurements, despite the fact that they are indirect measures, should be that one can study a reasonable number of subjects. So this low n seems to undermine that point. I rarely suggest additional data collection, but I do feel that a few more subjects would shore up the study's impact.

The present study was conducted in line with a deep phenotyping study approach. That is, we focused on acquiring highly reliable datasets on individuals. We did not intend to capture the population variance, which is often the goal of other group studies, since low level and basic features such as stripes in V2 are expected to be present in all healthy individuals. Thus we traded off and prioritized test-retest measurements for fMRI sessions and using an alternative MP2RAGE acquisition over a larger number of individuals. This resulted in 6–7 scanning sessions on different days for each individual, summing up to 26 long scanning session in total. We also note that the used sample size is not smaller than in other studies with a similar research question. For example, another fMRI study investigating V2 stripes in humans used the same sample size of n=4 (Dumoulin et al., 2017).

The paper overstates what can be concluded in a number of places. For example, the paper suggests that R1 and R2 are highly-specific to myelin in a number of places. For example, on p7 the text reads" "We tested whether different stripe types are differentially myelinated by comparing R1 and R2..." Relaxation times lack the specificity to definitively attribute these changes purely to myelin. Similarly, on p11: "Our study showed that pale stripes which exhibit lower oxidative metabolic activity according to staining with CO are stronger myelinated than surrounding gray matter in V2." This implies that the study directly links CO staining to myelination. In addition to using non-specific estimates of myelination, the study does not actually measure CO.

We agree that we did not clearly point out the limitations of R1 myelin mapping. Therefore, we toned down the statements about the connection between cortical myelin and R1. The mentioned statements in the reviewer’s comment were changed accordingly (see p. 6, line 163; p. 11, lines 353–354). We also included a small paragraph to clarify the used terminology (color-selective thin stripes, disparity-selective thick stripes) in the manuscript (see p. 4, lines 110–114) to avoid the inadvertent conflation of CO staining and actually measured brain activity.

I'm confused by the analysis in Figure 5. I can appreciate why the authors are keen to present a "tripartite" analysis (thick, thin, and pale stripes). But I find the gray curves confusing. As I understand it, the gray curves as generated include both the stripe of interest (red or blue plots) and the pale stripes. Why not just generate a three-way classification? Generating these plots in effect has already required hard classification of thin and thick stripes, so it is odd to create the gray plots, which mix two types of stripes. Alternatively, could you explicitly model the partial volume for a given cortical location (e.g., under the assumption that partial volume of thick and thin strips is indicated by the z-score) for the corresponding functional contrast? One could then estimate the relaxation times as a simple weighted sum of stripe-wise R1 or R2.

Figure on weighted average of stripe-wise R1 and R2. (a) shows the weighted sum of R1 (de-meaned and de-curved) over all V2 voxels. z-scores from color-selective thin stripe experiments and disparity-selective thick stripes were used as weights in the left and middle group of bars, respectively. An intermediate threshold of zmax=1.96 was used, i.e., final weights were defined as weights=(z-1.96). Weights with z<0 were set to 0. For pale stripes (right group of bars), we used the maximum z-score value from thin and thick stripe measurements. We then set all weights with z≥1.96 to 0 and used the inverse as final weights. i.e., weights = -1 * (max(z)-1.96). (b) shows the same analysis for R2. Error bars indicate 1 standard error of the mean.

(1) Yes, indeed. We agree that modeling the partial volume of each compartment (thin, thick and pale stripes) in each V2 voxel would be the most elegant approach. However, we note that z-scores between thin and thick stripe experiments may not reflect the voxel-wise partial volume effect, since they are a purely statistical measure and not a partial volume model. Having said this, we think that this general approach can give some additional insights and we provide results for a similar analysis here. We calculated the weighted sum of R1 and R2 values over all V2 voxels for each stripe compartment (thin, thick and pale stripes) independently (see above figure). For R1, we see the same pattern of R1 between stripe types as in the manuscript (Figure 5). Additionally, we show the differences here for each subject, which further demonstrates the reproducibility across subjects in our study. For R2, no clear pattern across subjects emerged, confirming the results in our manuscript. Since, this analysis did not add relavant new information to the manuscript, we refrained from adding this figure to the manuscript, in order not to overload it.

(2) In our current study, we were not primarily interested in investigating differences between thin/thick stripes and pale stripes. While histological analysis found differences (though not consistent) between CO dark stripes (more myelinated, (Tootell et al., 1983)) and CO pale stripes (more myelinated, Krubitzer and Kaas, 1989)), no study stated myelin differences between CO dark stripes. This does not fully exclude the possibility of myelination differences but suggests that if myelination differences between CO dark stripes existed, they would presumably be smaller than differences between CO dark and CO pale stripes. Thus, it would be even more difficult to demonstrate than the hypothesis of this manuscript.

Therefore, we decided to directly test two compartments against each other instead of modeling all three compartments within a single model. In our analysis, we thereby loosely followed the analysis methods described in Li et al. (2019), which compared myelin differences between thin/thick and pale stripes in macaques. We note that this demonstrates further consistency, since it is not trivial that both thick and thin stripes show lower R1 values than the pale stripes. For example, there may be no or opposite differences.

(3) Just for clarification, the plots in Figure 5 show the comparison of R1 (or R2*) between two compartments in V2. The red (blue) curve includes the thin (thick) stripe of interest. The gray curve includes everything in V2 minus contributions from thick (thin) stripes of interest. If we take the thin stripe comparison as example (Figure 5a), then red contains the thin stripes of interest while gray contains everything minus the thick stripes. Therefore, assuming a tripartite stripe arrangement, the gray curve contains both thin and pale stripe contributions.

References

Carey D, Caprini F, Allen M, Lutti A, Weiskopf N, Rees G, Callaghan MF, Dick F. Quantitative MRI provides markers of intra-, inter-regional, and age-related differences in young adult cortical microstructure. Neuroimage 2018; 182:429–440.

Dumoulin SO, Harvey BM, Fracasso A, Zuiderbaan W, Luijten PR, Wandell BA, Petridou N. In vivo evidence of functional and anatomical stripe-based subdivisions in human V2 and V3. Sci Rep 2017; 7:733.

Freund P, Seif M, Weiskopf N, Friston K, Fehlings MG, Thompson AJ, Curt A. MRI in traumatic spinal cord injury: from clinical assessment to neuroimaging biomarkers. Lancet Neurol 2019; 18:1123–1135.

González de San Román E, Bidmon H-J, Malisic M, Susnea I, Küppers A, Hübbers R, Wree A, Nischwitz V, Amunts K, Huesgen PF. Molecular composition of the human primary visual cortex profiled by multimodal mass spectrometry imaging. Brain Struct Func 2018; 223:2767–2783.

Kirilina E, Helbling S, Morawski M, Pine K, Reimann K, Jankuhn S, Dinse J, Deistung A, Reichenbach JR, Trampel R, Geyer S, Müller L, Jakubowski N, Arendt T, Bazin P-L, Weiskopf N. Superficial white matter imaging: Contrast mechanisms and whole-brain in vivo mapping. Sci Adv 2020; 6:eaaz9281.

Krubitzer LA, Kaas JH. Cortical integration of parallel pathways in the visual system of primates. Brain Res 1989; 478:161–165.

Lazari A, Lipp I. Can MRI measure myelin? Systematic review, qualitative assessment, and meta-analysis of studies validating microstructural imaging with myelin histology. Neuroimage 2021; 230:117744.

Lazari A, Salvan P, Cottaar M, Papp D, Rushworth MFS, Johansen-Berg H. Hebbian activity-dependent plasticity in white matter. Cell Rep 2022; 39:110951.

Li X, Zhu Q, Janssens T, Arsenault JT, Vanduffel W. In Vivo Identification of Thick, Thin, and Pale Stripes of Macaque Area V2 Using Submillimeter Resolution (f)MRI at 3 T. Cereb 2019; 29:544–560.

Mancini M, Karakuzu A, Cohen-Adad J, Cercignani M, Nichols TE, Stikov N. An interactive meta-analysis of MRI biomarkers of myelin. Elife 2020; 9:e61523.

Natu VS, Gomez J, Barnett M, Jeska B, Kirilina E, Jaeger C, Zhen Z, Cox S, Weiner KS, Weiskopf N, Grill-Spector K. Apparent thinning of human visual cortex during childhood is associated with myelination. PNAS 2019; 116:20750–20759.

Tootell RBH, Silverman MS, De Valois RL, Jacobs GH. Functional Organization of the Second Cortical Visual Area in Primates. Science 1983; 220:737–739.

Weiskopf N, Edwards LJ, Helms G, Mohammadi S, Kirilina E. Quantitative magnetic resonance imaging of brain anatomy and in vivo histology. Nat Rev Phys 2021; 3:570–588.

Whitaker KJ, Vértes PE, Romero-Garcia R, Váša F, Moutoussis M, Prabhu G, Weiskopf N, Callaghan MF, Wagstyl K, Rittman T, Tait R, Ooi C, Suckling J, Inkster B, Fonagy P, Dolan RJ, Jones PB, Goodyer IM, NSPN Consortium, Bullmore ET. Adolescence is associated with genomically patterned consolidation of the hubs of the human brain connectome. PNAS 2016; 113:9105–9110.

Author Response

Reviewer #1 (Public Review):

Determination of the biomechanical forces and downstream pathways that direct heart valve morphogenesis is an important area of research. In the current study, potential functions of localized Yap signaling in cardiac valve morphogenesis were examined. Extensive immunostainings were performed for Yap expression, but Yap activation status as indicated by nuclear versus cytoplasmic localization, Yap dephosphorylation, or expression of downstream target genes was not examined.

We thank the reviewer for appreciating the significance of this work, and we also thank the reviewer for the constructive suggestions. Following these suggestions, we have improved analysis of YAP activation status and used nuclear versus cytoplasmic localization to quantify YAP activation. To address the reviewer’s concerns, we have conducted extra qPCR analysis of YAP downstream target genes and YAP upstream genes in Hippo pathway. Please find the detailed revisions in our responses to the Recommendations for authors.

The goal of the work was to determine Yap activation status relative to different mechanical environments, but no biomechanical data on developing heart valves were provided in the study.

We appreciate the reviewer for raising this concern. We have previously published the biomechanical data of developing chick embryonic heart valves in the following study:

Buskohl PR, Gould RA, Butcher JT. Quantification of embryonic atrioventricular valve biomechanics during morphogenesis. Journal of Biomechanics. 2012;45(5):895-902.

In that study, we used micropipette aspiration to measure the nonlinear biomechanics (strain energy) of chick embryonic heart valves at different developmental stages. Here in this study, we used the same method to measure the strain energy of YAP activated/inhibited cushion explants and compared it to the data from our previous study. Our findings were summarized in the Results: “YAP inhibition elevated valve stiffness”, and the detailed measurements, including images and data, are presented in Figure S4.

There are several major weaknesses that diminish enthusiasm for the study.

1) The Hippo/Yap pathway activation leads to dephosphorylation of Yap, nuclear localization, and induced expression of downstream target genes. However, there are no data included in the study on Yap nuclear/cytoplasmic ratios, phosphorylation status, or activation of other Hippo pathway mediators. Analysis of Yap expression alone is insufficient to determine activation status since it is widely expressed in multiple cells throughout the valves. The specificity for activated Yap signaling is not apparent from the immunostainings.

We thank the reviewer for pointing out this weakness. We have now implemented nuclear versus cytoplasmic localization as recommended to quantify YAP activation. We have also conducted additional experiments to analyze via qPCR YAP downstream target genes and YAP upstream genes in Hippo pathway. Please see the detailed revisions in our responses to the Recommendations for authors.

2) The specific regionalized biomechanical forces acting on different regions of the valves were not measured directly or clearly compared with Yap activation status. In some cases, it seems that Yap is not present in the nuclei of endothelial cells surrounding the valve leaflets that are subject to different flow forces (Fig 1B) and the main expression is in valve interstitial subpopulations. Thus the data presented do not support differential Yap activation in endothelial cells subject to different fluid forces. There is extensive discussion of different forces acting on the valve leaflets, but the relationship to Yap signaling is not entirely clear.

We thank the reviewer for these important questions. The region-specific biomechanics have been well mapped and studied, thanks to the help from Computational Fluid Dynamics supported by ultrasound velocity and pressure measurements. For example:

Yalcin, H.C., Shekhar, A., McQuinn, T.C. and Butcher, J.T. (2011), Hemodynamic patterning of the avian atrioventricular valve. Dev. Dyn., 240: 23-35.

Bharadwaj KN, Spitz C, Shekhar A, Yalcin HC, Butcher JT. Computational fluid dynamics of developing avian outflow tract heart valves. Ann Biomed Eng. 2012 Oct;40(10):2212-27. doi: 10.1007/s10439-012-0574-8.

Ayoub S, Ferrari G, Gorman RC, Gorman JH, Schoen FJ, Sacks MS. Heart Valve Biomechanics and Underlying Mechanobiology. Compr Physiol. 2016 Sep 15;6(4):1743-1780.

Salman HE, Alser M, Shekhar A, Gould RA, Benslimane FM, Butcher JT, et al. Effect of left atrial ligation-driven altered inflow hemodynamics on embryonic heart development: clues for prenatal progression of hypoplastic left heart syndrome. Biomechanics and Modeling in Mechanobiology. 2021;20(2):733-50.

Ho S, Chan WX, Yap CH. Fluid mechanics of the left atrial ligation chick embryonic model of hypoplastic left heart syndrome. Biomechanics and Modeling in Mechanobiology. 2021;20(4):1337-51.

Those studies have shown that USS develops on the inflow surface of valves while OSS develops on the outflow surface of valves, CS develops in the tip region of valves while TS develops in the regions of elongation and compaction. Here in this study, we mimic those forces in our in-vitro and ex-vivo models. This allows us to study the direct effect of specific force on the YAP activity in different cell lineages. The results showed that OSS promoted YAP activation in VECs while USS inhibited it, CS promoted YAP activation in VICs while TS inhibited it. This result well explained the spatiotemporal distribution of YAP activation in Figure 1. For example, nuclear YAP was mostly found in VECs on the fibrosa side, where OSS develops, and YAP was not expressed in the nuclei in VECs of the atrialis/ventricularis side, where USS develops. It is also worth noting that formation of OSS on the outflow side is slower, and thus the side specific YAP activation in VECs was not in effect at the early stage, from E11.5 to E14.5.

3) The requirement for Yap signaling in heart valve remodeling as described in the title was not demonstrated through manipulation of Yap activity.

With respect, it is unclear what the reviewer is asking for given no experiments are suggested nor an elaboration of alternative interpretations of our results that emphasize against YAP requirement. It has been previously shown that YAP signaling is required for early EMT stages of valvulogenesis using conditional YAP deletion in mice:

Zhang H, von Gise A, Liu Q, Hu T, Tian X, He L, et al. Yap1 Is Required for Endothelial to Mesenchymal Transition of the Atrioventricular Cushion. Journal of Biological Chemistry. 2014;289(27):18681-92.

Signaling roles for early regulators at these later fetal stages are different, sometimes opposite early EndMT stages, thus contraindicating reliance on these early data to explain later events:

Bassen D, Wang M, Pham D, Sun S, Rao R, Singh R, et al. Hydrostatic mechanical stress regulates growth and maturation of the atrioventricular valve. Development. 2021;148(13).

However, embryos with YAP deletion failed to form endocardial cushions and could not survive long enough for the study of its roles in later cushion growth and remodeling into valve leaflets. In this work,

We first showed the localization of YAP activity and its direct link with local shear or pressure domains. Then we explicitly applied controlled gain and loss of function of YAP via specific molecules. We also applied critical mechanical gain or loss of function studies to demonstrate YAP mechanoactivation necessity and sufficiency to achieve growth and remodeling.

Reviewer #2 (Public Review)

This study by Wang et al. examines changes in YAP expression in embryonic avian cultured explants in response to high and low shear stress, as well as tensile and compressive stress. The authors show that YAP expression is increased in response to low, oscillatory shear stress, as well as high compressive stress conditions. Inhibition of YAP signaling prevents compressive stress-induced increases in circularity, decreased pHH3 expression, and increases VE-cadherin expression. On the other hand, YAP gain of function prevents tensile stress-induced decreases in pHH3 expression and VE-cadherin expansion. It also decreases the strain energy density of embryonic avian cushion explants. Finally, using an avian model of left atrial ligation, the authors demonstrate that unloaded regions within the primitive valve structures are associated with increased YAP expression, compared to regions of restricted flow where YAP expression is low. Overall, this study sheds light on the biomechanical regulation of YAP expression in developing valves.

We thank the reviewer for the accurate summary and their enthusiasm for this work.

Strengths of the manuscript include:

• Novel insights into the dynamic expression pattern of YAP in valve cell populations during post-EMT stages of embryonic valvulogenesis.

• Identify the positive regulation of YAP expression in response to low, oscillatory shear stress, as well as high compressive stress conditions.

• Identify a link between YAP signaling in regulating stress-induced cell proliferation and valve morphogenesis.

• The inclusion of the atrial left atrial ligation model is innovative, and the data showing distinguishable YAP expression levels between restricted, and non-restricted flow regions is insightful.

We thank the reviewer for appreciating the strengths of this work.

This is a descriptive study that focuses on changes in YAP expression following exposure to diverse stress conditions in embryonic avian cushion explants. Overall, the study currently lacks mechanistic insights, and conclusions based on data are highly over-interpreted, particularly given that the majority of experimental protocols rely on one method of readout.

We thank the reviewer for constructive suggestions.

Reviewer #3 (Public Review)

In this manuscript, Wang et al. assess the role of wall shear stress and hydrostatic pressure during valve morphogenesis at stages where the valve elongates and takes shape. The authors elegantly demonstrate that shear and pressure have different effects on cell proliferation by modulating YAP signaling. The authors use a combination of in vitro and in vivo approaches to show that YAP signaling is activated by hydrostatic pressure changes and inhibited by wall shear stress.

We thank the reviewer for their enthusiasm for the impact of our work.

There are a few elements that would require clarification:

1) The impact of YAP on valve stiffness was unclear to me. How is YAP signaling affecting stiffness? is it through cell proliferation changes? I was unclear about the model put forward:

• Is it cell proliferation (cell proliferation fluidity tissue while non-proliferating tissue is stiffer?)

• Is it through differential gene expression?

This needs clarification.

We thank the reviewer for raising this important question. Cell proliferation can affect valve stiffness but is a minor factor compared with ECM deposition and cell contractility Our micropipette aspiration data showed that the higher cell proliferation rate induced by YAP activation did lead to stiffer valves when compared to the controls. This may be because at the early stages, cells are more elastic than the viscous ECM. However, the stiffness of YAP activated valves were only about half of that of YAP inhibited valves, showing that the transcriptional level factor plays a more important role. This also suggests that YAP inhibited valves exhibited a more mature phenotype. An analogous role of YAP has also been found in cardiomyocytes. Many theories propose that in cardiomyocytes when YAP is activated the proliferation programs are turned on, while when YAP is inhibited the proliferation programs are turned off and maturation programs are released. Similarly, here we hypothesize that YAP works like a mechanobiological switch, converting mechanical signaling into the decision between growth and maturation. We have revised the Discussion to include this hypothesis.

2) The model proposes an early asymmetric growth of the cushion leading to different shear forces (oscillatory vs unidirectional shear stress). What triggers the initial asymmetry of the cushion shape? is YAP involved?

Although the initial geometry of the cushion model is symmetric, the force acting on it is asymmetric. The detailed numerical simulation of how the initial forces trigger the asymmetric morphogenesis can be found in our previous publication:

Buskohl PR, Jenkins JT, Butcher JT. Computational simulation of hemodynamic-driven growth and remodeling of embryonic atrioventricular valves. Biomechanics and Modeling in Mechanobiology. 2012;11(8):1205-17.

The color maps represent the dilatation rates when a) only pressure is applied, b) only shear stress is applied, and c) both pressure and shear stress are applied. It is such load that initiates an asymmetric morphological change, as shown in d). In addition, we believe YAP is involved during the initiation because it is directly nuclear activated by CS and OSS or cytoplasmically activated by TS and LSS.

3) The differential expression of YAP and its correlation to cell proliferation is a little hard to see in the data presented. Drawings highlighting the main areas would help the reader to visualise the results better.

We thank the reviewer for this helpful suggestion, we have improved the visualization of Figure 3C and Figure 4C with insets of higher magnification.

4) The origin of osmotic/hydrostatic pressure in vivo. While shear is clearly dependent upon blood flow, it is less clear that hydrostatic pressure is solely dependent upon blood flow. For example, it has been proposed that ECM accumulation such as hyaluronic acid could modify osmotic pressure (see for example Vignes et al.PMID: 35245444). Could the authors clarify the following questions:

• How blood flow affects osmotic pressure in vivo?

• Is ECM a factor that could affect osmotic pressure in this system?

We thank the reviewer for sharing this interesting study. The osmotic pressure plays a critical role in mechanotransduction and the development of many tissues including cardiovascular tissues and cartilage. As proposed in the reference, osmotic pressure is an interstitial force generated by cardiac contractility. Here in our study, the hydrostatic pressure is different, which is an external force applied by flowing blood. According to Bernoulli's law, when an incompressible fluid flows around a solid, the static pressure it applies on the solid is equal to its total pressure minus its dynamic pressure.

Despite the difference, the osmotic pressure can mimic the effect of hydrostatic pressure in-vitro. The in-vitro osmotic pressure model has been widely used in cartilage research, for example:

P. J. Basser, R. Schneiderman, R. A. Bank, E. Wachtel, and A. Maroudas, “Mechanical properties of the collagen network in human articular cartilage as measured by osmotic stress technique.,” Arch. Biochem. Biophys., vol. 351, no. 2, pp. 207–19, 1998.

D. a. Narmoneva, J. Y. Wang, and L. a. Setton, “Nonuniform swelling-induced residual strains in articular cartilage,” J. Biomech., vol. 32, no. 4, pp. 401–408, 1999.

C. L. Jablonski, S. Ferguson, A. Pozzi, and A. L. Clark, “Integrin α1β1 participates in chondrocyte transduction of osmotic stress,” Biochem. Biophys. Res. Commun., vol. 445, no. 1, pp. 184–190, 2014.

Z. I. Johnson, I. M. Shapiro, and M. V. Risbud, “Extracellular osmolarity regulates matrix homeostasis in the intervertebral disc and articular cartilage: Evolving role of TonEBP,” Matrix Biol., vol. 40, pp. 10–16, 2014.

When maturing cushions shift from GAGs dominated ECM to collagen dominated ECM, the water and ion retention capacity of the tissue would be greatly changed, and thus reducing the osmotic pressure. This could in turn accelerate the maturation of cushions. By contrast, the ECM of growing cushions remain GAGs dominated, which would delay maturation and prolong the growth.

The revised second section of Results is as follows:

Shear and hydrostatic stress regulate YAP activity

In addition to the co-effector of the Hippo pathway, YAP is also a key mediator in mechanotransduction. Indeed, the spatiotemporal activation of YAP correlated with the changes in the mechanical environment. During valve remodeling, unidirectional shear stress (USS) develops on the inflow surface of valves, where YAP is rarely expressed in the nuclei of VECs (Figure 2A). On the other side, OSS develops on the outflow surface, where VECs with nuclear YAP localized. The YAP activation in VICs also correlated with hydrostatic pressure. The pressure generated compressive stress (CS) in the tips of valves, where VICs with nuclear YAP localized (Figure 2B). Whereas tensile stress (TS) was created in the elongated regions, where YAP was absent in VIC nuclei.

To study the effect of shear stress on the YAP activity in VECs, we applied USS and OSS directly onto a monolayer of freshly isolated VECs. The VEC was obtained from AV cushions of chick embryonic hearts at HH25. The cushions were placed on collagen gels with endocardium adherent to the collagen and incubated to enable the VECs to migrate onto the gel. We then removed the cushions and immediately applied the shear flow to the monolayer for 24 hours. The low stress OSS (2 dyn/cm2) promoted YAP nuclear translocation in VEC (Figure 2C, E), while high stress USS (20 dyn/cm2) restrained YAP in cytoplasm.

To study the effect of hydrostatic stress on the YAP activation in VICs, we used media with different osmolarities to mimic the CS and TS. CS was induced by hypertonic condition while TS was created by hypotonic condition, and the Unloaded (U) condition refers to the osmotically balanced media. Notably, in-vivo hydrostatic pressure is generated by flowing blood, while in-vivo osmotic pressure is generated by cardiac contractility and plays a critical role in the mechanotransduction during valve development (30). Despite the different in-vivo origination, the osmotic pressure provides a reliable model to mimic the hydrostatic pressure in-vitro (31). We cultured HH34 AV cushion explants under different loading conditions for 24 hours and found that the trapezoidal cushions adopted a spherical shape (Figure 2D). TS loaded cushions significantly compacted, and the YAP activation in VICs of TS loaded cushions was significantly lower than that in CS loaded VICs (Figure 2F).

Author Response

Reviewer #1 (Public Review):

Huang et al. sought to study the cellular origin of Tuft cells and the molecular mechanisms that govern their specification in severe lung injury. First the authors show ectopic emergence of Tuft cells in airways and distal parenchyma following different injuries. The authors also used lineage tracing models and uncovered that p63-expressing cells and to some extent Scgb1a1-lineaged labeled cells contribute to tuft cells after injury. Further, the authors modulated multiple pathways and claim that Notch inhibition blocks tuft cells whereas Wnt inhibition enhances Tuft cell development in basal cell cultures. Finally, the authors used Trpm5 and Pou2f3 knock-out models to claim that tuft cells are indispensable for alveolar regeneration.

In summary, the findings described in this manuscript are somewhat preliminary. The claim that the cellular origin of Tuft cells in influenza infection was not determined is incorrect. Current data from pathway modulation is preliminary and this requires genetic modulation to support their claims.

We thank the reviewer for the comments and we have performed extensive experiments to address the reviewer’s comments. In the revised manuscript we provide additional data including genetic modulation findings to support our model.

Major comments:

1) The abstract sounds incomplete and does not cover all key aspects of this manuscript. Currently, it is mainly focusing on the cellular origin of Tuft cells and the role of Wnt and notch signaling. However, it completely omits the findings from Trpm5 and Pou2f3 knock-out mice. In fact, the title of the manuscript highlights the indispensable nature of tuft cells in alveolar regeneration.

We have modified the abstract and title accordingly.

2) In lines 93-94, the authors state that "It is also unknown what cells generate these tuft cells.....". This statement is incorrect. Rane et al., 2019 used the same p63-creER mouse line and demonstrated that all tuft cells that ectopically emerge following H1N1 infection originate from p63+ lineage labeled basal cells. Therefore, this claim is not new.

We thank the reviewer’s comment. Although Rane et al. reported the p63-expressing lineage-negative epithelial stem/progenitor cells (LNEPs) could contribute to the ectopic tuft cells after PR8 virus infection, it is still not clear whether the p63+ cells immediately give rise to tuft cells or though EBCs. Thus, we performed TMX injection after PR8 infection, different from Rane et al (Rane et al., 2019). who performed Tmx injection before viral infection to indicate the ectopic tuft cells are derived from EBCs, as shown in revised Figure 2.

3) Lines 152-153 state that "21.0% +/- 2.0 % tuft cells within EBCs are labeled with tdT when examined at 30 dpi...". It is not clear what the authors meant here ("within EBC's")? And also, the same sentence states that "......suggesting that club cell-derived EBCs generate a portion of tuft cells....". In this experiment, the authors used club cell lineage tracing mouse lines. So, how do the authors know that the club cell lineage-derived tuft cells came through intermediate EBC population? Current data do not show evidence for this claim. Is it possible that club cells can directly generate tuft cells?

We apologize for the confusion and revised the text accordingly. Here, “within EBCs” means within the “pods” area where p63+ basal cells are ectopically present. The sentence is revised as “21.0% +/- 2.0 % tuft cells that are ectopically present in the parenchyma are labeled by tdT. Notably, these lineage labeled tuft cells were co-localized with EBCs.” We don’t know whether the club cell lineage-derived tuft cells transit through intermediate EBCs and that is why we use “suggest”. It is also possible that club cells can directly generate tuft cells. To avoid the confusion, we delete the sentence.

4) Based on the data from Fig-3A, the authors claim that treatment with C59 significantly enhances tuft cell development in ALI cultures. Porcupine is known to facilitate Wnt secretion. So, which cells are producing Wnt in these cultures? It is important to determine which cells are producing Wnt and also which Wnt? Further, based on DBZ treatments, it appears that active Notch signaling is necessary to induce Tuft cell fate in basal cells. Where are Notch ligands expressed in these tissues? Is Notch active only in a small subset of basal cells (and hence generate rate tuft cells)? This is one of the key findings in this manuscript. Therefore, it is important to determine the expression pattern of Wnt and Notch pathway components.

We thank the reviewer’s interesting questions and agree the importance of identifying the specific ligands and receptors for relevant Wnt and Notch signaling during tuft cell derivation. That being said, we think the topic is beyond the scope of this study which is focused on the role of tuft cells in alveolar regeneration. The point is well taken and we will investigate the topic in our future study.

5) How do the authors explain different phenotypes observed in Trpm5 knockout and Pou2f3 mutants? Is it possible that Trpm5 knockout mice have a subset of tuft cells and that they might be something to do with the phenotypic discrepancy between two mutant models?

Again we thank the reviewer for the interesting question. As discussed in the discussion section, Trpm5 is also reported to be expressed in B lymphocytes (Sakaguchi et al., 2020). It is possible that loss of Trpm5 modulates the inflammatory responses following viral infection, which may contribute to improved alveolar regeneration. However, it is also possible that Trpm5-/- mice keep a subset of tuft cells that facilitate lung regeneration as suggested by the reviewer.

6) One of the key findings in this manuscript is that Wnt and Notch signaling play a role in Tuft cell specification. All current experiments are based on pharmacological modulation. These need to be substantiated using genetic gain loss of function models.

We have performed the genetic studies.

Reviewer #2 (Public Review):

In this manuscript, the authors describe the ectopic differentiation of tuft cells that were derived from lineage-tagged p63+ cells post influenza virus infection. These tuft cells do not appear to proliferate or give rise to other lineages. They then claim that Wnt inhibitors increase the number of tuft cells while inhibiting Notch signaling decreases the number of tuft cells within Krt5+ pods after infection in vitro and in vivo. The authors further show that genetic deletion of Trpm5 in p63+ cells post-infection results in an increase in AT2 and AT1 cells in p63 lineage-tagged cells compared to control. Lastly, they demonstrate that depletion of tuft cells caused by genetic deletion of Pou2f3 in p63+ cells has no effect on the expansion or resolution of Krt5+ pods after infection, implying that tuft cells play no functional role in this process.

Overall, in vivo and in vitro phenotypes of tuft cells and alveolar cells are clear, but the lack of detailed cellular characterization and molecular mechanisms underlying the cellular events limits the value of this study.

We thank the reviewer for the comments and acknowledging that our findings are clear. In the revised manuscript we provide more detailed characterization and genetic evidence to elucidate the role of tuft cells in lung regeneration.

1) Origin of tuft cells: Although the authors showed the emergence of ectopic tuft cells derived from labelled p63+ cells after infection, it cannot be ruled out that pre-existing p63+Krt5- intrapulmonary progenitors, as previously reported, can also contribute to tuft cell expansion (Rane et al. 2019; by labelling p63+ cells prior to infection, they showed that the majority of ectopic tuft cells are derived from p63+ cells after viral infection). It would be more informative if the authors show the differentiation of tuft cells derived from p63+Krt5+ cells by tracing Krt5+ cells after infection, which will tell us whether ectopic tuft cells are differentiated from ectopic basal cells within Krt5+ pods induced by virus infection.

We thank the reviewer for the helpful suggestion. We have performed the experiment accordingly.

2) Mechanisms of tuft cell differentiation: The authors tried to determine which signaling pathways regulate the differentiation of tuft cells from p63+ cells following infection. Although Wnt/Notch inhibitors affected the number of tuft cells derived from p63+ labelled cells, it remains unclear whether these signals directly modulate differentiation fate. The authors claimed that Wnt inhibition promotes tuft cell differentiation from ectopic basal cells. However, in Fig 3B, Wnt inhibition appears to trigger the expansion of p63+Krt5+ pod cells, resulting in increased tuft cell differentiation rather than directly enhancing tuft cell differentiation. Further, in Fig 3D, Notch inhibition appears to reduce p63+Krt5+ pod cells, resulting in decreased tuft cell differentiation. Importantly, a previous study has reported that Notch signalling is critical for Krt5+ pod expansion following influenza infection (Vaughan et al. 2015; Xi et al. 2017). Notch inhibition reduced Krt5+ pod expansion and induced their differentiation into Sftpc+ AT2 cells. In order to address the direct effect of Wnt/Notch signaling in the differentiation process of tuft cells from EBCs, the authors should provide a more detailed characterization of cellular composition (Krt5+ basal cells, club cells, ciliated cells, AT2 and AT1 cells, etc.) and activity (proliferation) within the pods with/without inhibitors/activators.

Again we thank the reviewer for the insightful suggestions. We agree that it will be interesting to further address the direct effect of Wnt/Notch signaling in the differentiation process of tuft cells from EBCs. In this revised manuscript we added new findings of EBC differentiation into tuft cells in mice with genetic deletion of Rbpjk.

3) Impact of Trpm5 deletion in p63+ cells: It is interesting that Trpm5 deletion promotes the expansion of AT2 and AT1 cells derived from labelled p63+ cells following infection. It would be informative to check whether Trpm5 regulates Hif1a and/or Notch activity which has been reported to induce AT2 differentiation from ectopic basal cells (Xi et al. 2017). Although the authors stated that there was no discernible reduction in the size of Krt5+ pods in mutant mice, it would be interesting to investigate the relationship between AT2/AT1 cell retaining pods and the severity of injury (e.g. large Krt5+ pods retain more/less AT2/AT1 cells compared to small pods. What about other cell types, such as club and goblet cells, in Trpm5 mutant pods? Again, it cannot be ruled out that pre-existing p63+Krt5- intrapulmonary progenitor cells can directly convert into AT2/AT1 cells upon Trpm5 deletion rather than p63+Krt5+ cells induced by infection.

We thank the reviewer for the comments and suggestions. Our new data using KRT5-CreER mouse line confirmed that pod cells (Krt5+) do not contribute to AT2/AT1 cells, consistent with previous studies (Kanegai et al., 2016; Vaughan et al., 2015). Our data also show that p63-CreER lineage labeled AT2/AT1 cells are separated from pod cell area, suggesting pod cells and these AT2/AT1 cells are generated from different cell of origin. We also checked the Notch activity in pod cells in Trpm5-/- mice, and some pod cell-derived cells are Hes1 positive, whereas some are Hes1 negative (RLFigure 1). As indicated in discussion we think that AT2/AT1 cells are possibly derived from pre-existing AT2 cells that transiently express p63 after PR8 infection. It will be interesting to test whether Trpm5 regulates Hif1a in this population (p63+,Krt5-), and this will be our next plan.

RLFigure 1. Representative area staining in Trpm5-/- mice at 30 dpi. Area 1: Notch signaling is active (Hes1+, arrows) in pod cells following viral infection. Area 2: pod cells exhibit reduced Notch activities. Note few Hes1+ cells in pods (arrows). Scale bar: 50 µm.

4) Ectopic tuft cells in COVID-19 lungs: The previous study by the authors' group revealed the presence of ectopic tuft cells in COVID-19 patient samples (Melms et al. 2021). There appears to be no additional information in this manuscript.

In Melms et al., Nature, 2021 (Melms et al., 2021), we showed tuft cell expansion in COVID-19 lungs but not the potential origin of tuft cells. In this manuscript we show some cells co-expressing POU2F3 and KRT5, suggesting a pod-to-tuft cell differentiation.

5) Quantification information and method: Overall, the quantification method should be clarified throughout the manuscript. Further, in the method section, the authors stated that the production of various airway epithelial cell types was counted and quantified on at least 5 "random" fields of view. However, virus infection causes spatially heterogeneous injury, resulting in a difficult to measure "blind test". The authors should address how they dealt with this issue.

We clarified that quantification method as suggested. For the in vitro cell culture assays on the signaling pathways, we took pictures from at least five random fields of view for quantification. For lung sections, we tile-scanned the lung sections including at least three lung lobes and performed quantification.

Reviewer #3 (Public Review):

In this manuscript Huang et al. study how the lung regenerates after severe injury due to viral infection. They focus on how tuft cells may affect regeneration of the lung by ectopic basal cells and come to the conclusion that they are not required. The manuscript is intriguing but also very puzzling. The authors claim they are specifically targeting ectopic basal progenitor cells and show that they can regenerate the alveolar epithelium in the lung following severe injury. However, it is not clear that the p63-CreERT2 line the authors are using only labels ectopic basal cells. The question is what is a basal cell? Is an ectopic basal progenitor cell only defined by Trp63 expression?

The accompanying manuscript by Barr et al. uses a Krt5-CreERT2 line to target ectopic basal cells and using that tool the authors do not see a signification contribution of ectopic basal cells towards alveolar epithelial regeneration. As such the claim that ectopic basal cell progenitors drive alveolar epithelial regeneration is not well-founded.

We appreciate the reviewer for the positive comments and agreeing that our findings are interesting.

The title itself is also not very informative and is a bit misleading. That being said I think the manuscript is still very interesting and can likely easily be improved through a better validation of which cells the p63-CreERT2 tool is targeting.

We have revised the title accordingly and performed extensive experiments to address the reviewer’s concerns.

I, therefore, suggest the following experiments.

1) Please analyze which cells p63-CreERT2 labels immediately after PR8 and tamoxifen treatment. Are all the tdTomato labeled cells also Krt5 and p63 positive or are some alveolar epithelial cells or other airway cell types also labeled?

We thank the reviewer for the question. To answer the reviewer’s question, we performed PR8 infection (250 pfu) on three Trp63-CreERT2;R26tdT mice and TMX treatment at days 5 and 7 post viral infection. We didn't perform TMX injection immediately as the mice were sick at a few days post infection. The lung samples were collected at 14 dpi. We observed that tdT+ cells are present in the airways (rebuttal letter RLFigure 2A, B), and it appears that the lineage labeled cells (tdT+) include club cells (CC10+) that are underlined by tdT+Krt5+ basal cells (RLFigure 2C). We think that these labeled basal cells give rise to club cells. However, we also noticed that rare club cells and ciliated cells (FoxJ1+) are labeled by tdT in the areas absent of surrounding tdT+ basal cells (RLFigure 2D). Moreover, a minor population of tdT+ SPC+ cells are present in the terminal airways that were disrupted by viral infection (RLFigure 2E and D). We did not see any pods formed in this experiment and we did not observe any tdT+ cells in the intact alveoli (uninjured area).

RLFigure 2. Trp63-CreERT2 lineage labeled cells in the airways but not alveoli when Tamoxifen was induced at day 5 and 7 after PR8 H1N1 viral infection. Trp63-CreERT2;R26-tdT mice were infected with PR8 at 250 pfu and Tmx were delivered at a dose of 0.25 mg/g bodyweight by oral gavage. Lung samples were collected and analyzed at 14 dpi. Stained antibodies are as indicated. Scale bar: 100 µm.

2) Please also show if p63-CreERT2 labels any cells in the adult lung parenchyma in the absence of injury after tamoxifen treatment.

Dr. Wellington Cardoso’s group demonstrated that Trp63-CreERT2 only labels very few cells in the airways but not the lung parenchyma in the absence of injury after tamoxifen treatment (Yang et al., 2018). Dr. Ying Yang has revisited the data and she did not observe any labeling in the lung parenchyma (n = 2).

3) Please analyze if p63-CreERT2 labels any cells with tdTomato in the absence of injury or after PR8 infection but without tamoxifen treatment.

We performed the experiment and didn't observe any labeled cells in the lung parenchyma without Tamoxifen treatment (n = 4).

4) Please analyze when after PR8 infection do the first p63-CreERT2 labeled tdTomato positive alveolar epithelial cells appear.

We administered tamoxifen at day 5 and 7 after PR8 infection and harvested lung tissues at day 14. As shown in Figure 1, we observed a few tdT+ SPC+ cells in the terminal airways that are disrupted by viral infection. Notably, we did not observe any lineage labeled cells in the intact alveoli (uninjured) in this experiment..

5) A clonal analysis of p63-CreERT2 labeled cells using a confetti reporter might also help interpret the origin of p63-CreERT2 labeled cells.

We thank the reviewer for the suggestion. Our new data demonstrate that a rare population of SPC+tdT+ cells are present in the disrupted terminal airways of Trp63-CreERT2;R26tdT mice. Our data in the original manuscript and the new data suggest that the initial SPC+;tdT+ cells are rare because we have to administrate multiple doses of Tamoxifen to label them. Given the less labeling efficiency of confetti than R26tdT mice, it is possible we will not be able to label these SPC+ cells. Moreover, our original manuscript clearly shows individual clones of SPC+tdT+ cells in the regenerated lung, and they do not seem to compose of multiple clones. Therefore we think that use of confetti mice may not add new information..

6) Lastly could the authors compare the single-cell RNAseq transcription profile of p63-CREERT2 labeled cells immediately after PR8 and tamoxifen treatment and also at 60dpi. A pseudotime analysis and trajectory interference analysis could help elucidate the identity of p63-CreERT2 labeled cells that are actually not ectopic basal progenitor cells.

We appreciated the reviewer’s suggestion and agree that single cell RNA sequencing with pseudotime analysis can provide further information regarding the origin of the lineage labeled alveolar cells of Trp63-CreERT2;R26tdT mice. That said, our new data clearly show that KRT5-CreER lineage labeled cells do not give rise to AT1/2 cells as previously described (Kanegai et al., 2016; Vaughan et al., 2015), suggesting that the ectopic basal progenitor cells do not generate alveolar cells. By contrast, Trp63-CreERT2 lineage labeled cells do give rise to AECs, suggesting that this p63+ cell population capable of generating AECs are different from Krt5+ ectopic basal progenitor cells. Our single cell core has an extremely long waiting list due to the pandemic and we hope that our new findings are enough to address the reviewer’s concern without the need of single cell analysis..

Author Response

Reviewer #1 (Public Review):

This manuscript applies the framework of information theory to study a subset of cellular receptors (called lectins) that bind to glycan molecules, with a specific focus on the kinds of glycans that are typical of fungal pathogens. The authors use the concentration of various types of ligands as the input to the signaling channel, and measure the "response" of individual cells using a GFP reporter whose expression is driven by a promoter that responds to NFκB. While this work is overall technically solid, I would suggest that readers keep several issues in mind while evaluating these results.

1) One of the largest potential limitations of the study is the reliance of the authors on exogenous expression of the relevant receptors in U937 cells. Using a cell-line system like this has several advantages, most notably the fact that the authors can engineer different reporters and different combinations of receptors easily into the same cells. This would be much more difficult with, say, primary cells extracted from a mouse or a human. While the ability to introduce different proteins into the cells is a benefit, the problem is that it is not clear how physiologically relevant the results are. To their credit, the authors perform several controls that suggest that differences in transfection efficiency are not the source of the differences in channel capacity between, say, dectin-1 and dectin-2. As the authors themselves clearly demonstrate, however, the differences in the properties of these signaling system are not based on receptor expression levels, but rather on some other property of the receptor. Now, it could be that the dectin-2 receptor is somehow just more "noisy" in terms of its activity compared to, say, dectin-1. This seems a somewhat less likely explanation, however, and so it is likely that downstream details of the signaling systems differ in some way between dectin-2 and the more "information efficient" receptors studied by the authors.

The channel capacity of a cell signaling network depends critically on the distributions of the downstream signaling molecules in question: see the original paper by Cheong et al. (2011, Science 334 (6054), 354-8) and subsequent papers (notably Selimkhanov et al. (2014) Science 346 (6215), 1370-3 and Suderman et al. (2018) Interface Focus 8 (6), 20180039). The U937 cells considered here clearly don't serve the physiological function of detecting the glycans considered by the authors; despite the fact that this is an artificial cell line, the fact the authors have to exogenously express the relevant receptors indicates that these cells are not necessarily a good model for the types of cells in the body that actually have evolved to sense these glycan molecules.

Signaling molecules readily exhibit cell-type-specific expression levels that influence cellular responses to external stimuli (Rowland et al.(2017) Nat Commun 8, 16009). So it is unclear that the distributions of downstream signaling molecules in U937 cells mirror those that would be observed in the immune cell types relevant to this response. As such, the physiological relevance of the differences between dectin-2 channel capacities and those exhibited by the other receptors are currently unclear.

We appreciate Reviewer #1’s in-depth comments related to physiological relevance of the U937 cell. A big benefit of using information theory to investigate a biological communication channel is the realization of quantitative measurement of information that the channel transmits without having detailed measurement of spatiotemporal dynamics of receptors and downstream signaling cascades. In addition, the quantity of measured information itself in turn gives us a decent prediction about detailed signaling mechanisms by comparing the information quantity difference. For example, we investigated how transmission of glycan information from dectin-2 is synergistically modulated in the presence of either dectin-1, DC-SIGN or mincle. Our approach allows to investigate how individual lectins on immune cells contribute to glycan information transmission and be integrated in the presence other type of lectins. Therefore, the findings describe how physiologically relevant lectins are integrating the extracellular signal in a more defined way. Furthermore, we found that our model cell line has one order of magnitude higher expression of dectin-2 compared with primary human monocytes and exhibits a similar zymosan binding pattern (will be described in Recommendations for the authors and Figure R8).

We fully agree that acquiring more information on the information transmission capability of primary immune cells would increase physiological relevance. In the revised manuscript we addressed this concern by comparing the receptor expression levels of our model cell lines with primary monocytes, for which we find an agreement of cellular heterogeneity. However, we would also like to point out that the very basic nature of our question, of how information stored in glycans is processed by lectins, is not tightly bound to these difference of primary cells and cell lines.

Line 382: Finally, it is important to take into consideration that our conclusions came from model cell lines, which were used as a surrogate for cell-type-specific lectin expression patterns of primary immune cells. Human monocytes and dectin-2 positive U937 cells have comparable receptor densities and respond similar to stimulation with zymosan particles (SI Fig. 6A and B).

2) Another issue that readers might want to keep in mind is that the details of the channel capacity calculation are a bit unclear as the manuscript is currently written. The authors indicate that their channel capacity calculations follow the approach of Cheong et al. (2011) Science 334 (6054), 354-8. However, the extent to which they follow that previous approach is not obvious. For instance, the calculations presented in the 2011 work use a combined bootstrapping/linear extrapolation approach to estimate the mutual information at infinite population size in order to deal with known inaccuracies in the calculation that arise from finite-size effects. The Cheong approach also deals with the question of how many bins to use in order to estimate the joint probability distribution across signal and response.

They do this by comparing the mutual information they calculate for the real data with that calculated for random data to ensure that they are not calculating spuriously high mutual information based on having too many bins. While the Cheong et al. paper does a great job explaining why these steps need to be undertaken, a subsequent paper by Suderman et al. (2017, PNAS 114 (22), 5755-60) explains the approach in even greater detail in the supporting information. Those authors also implemented several improvements to the general approach, including a bootstrap method for more accurately estimating the error in the mutual information and channel capacity estimates.

The problem here is that, while the authors claim to follow the approach of Cheong et al., it seems that they have re-implemented the calculation, and they do not provide sufficient detail to evaluate the extent to which they are performing the same exact calculation. Since estimates of mutual information are technically challenging, specific details of the steps in their approach would be helpful in order to understand how closely their results can be compared with the results of previous authors. For instance, Cheong et al. estimate the "channel capacity" by trying a set of likely unimodal and bimodal distributions for the input to the channel, and choosing the maximal value as the channel capacity. This is clearly a very approximate approach, since the channel capacity is defined as the supremum over an (uncountably infinite) set of input probability distributions. In any case, the authors of the current manuscript use a different approach to this maximization problem. Although it is a bit unclear how their approach works, it seems that they treat the probability of each input bin as an independent parameter (under the constraint that the probabilities sum to one) and then use an optimization algorithm implemented in Python to maximize the mutual information. In principle, this could be a better approach, since the set of input distributions considered is potentially much larger. The details of the optimization algorithm matter, however, and those are currently unclear as the paper is written.

We thank Reviewer #1’s recommendation for increasing the legitimacy of the calculation. In the revised manuscript we tried to explain channel capacity calculation procedures in more detail with statistical approaches that adopted from Cheong et al. (2011) and Suderman et al. (2018) (SI section 1 and 2). Furthermore, we decide the number of binning from not only random dataset but also the number of total samples as shown below:

Figure R1. A) Extrapolated channel capacity values of random dataset at infinitely subsampled distribution under various total number of samples and output binning. The white line in the heatmap represents the channel capacity value at 0.01 bit. B) Extrapolated channel capacity values at infinite subsample size of U937 cells’ input (TNF-a doses) and output (GFP reporter) response.

Figure R1 describes channel capacity values from random (A) and experimental dataset (B, TNFAR + TNF-a). The channel capacity values from random data indicates the dependence of channel capacity on the number of the output binning and total number sample. According to this heatmap, we decided the allowed bias as 0.01 bits as shown in contour line shown in Figure R1A. Since our minimum dataset that used for channel capacity calculation in the absence of labelled input is near 90,000, the expected bias in channel capacity calculation is therefore less than 0.01 bits in binning range from 10 to 1000 as shown in Figure R1A.

Furthermore, we demonstrated mutual information maximization procedure using predefined unibimodal input distribution and compared with the systematic method that we used in the work. We found that there is no noticeable difference in channel capacity value between two approaches (SI Figure 3M).

3) Another issue to be careful about when interpreting these findings is the fact that the authors use logarithmic bins when calculating the channel capacity estimates. This is equivalent to saying that the "output" of the cell signaling channel is not the amount of protein produced under the control of the NFκB promoter, but rather the log of the protein level. Essentially, the authors are considering a case where the relevant output of the system is not the amount of protein itself, but the fold change in the amount of protein. That might be a reasonable assumption, especially if the protein being produced is a transcription factor whose own promoters have evolved to detect fold changes. For many proteins, however, the cell is likely responsive to linear changes in protein concentration, not fold changes. And so choosing the log of the protein level as the output may not make sense in terms of understanding how much information is actually contained in this particular output variable. Regardless, choosing logarithmic bins is not purely a matter of convenience or arbitrary choice, but rather corresponds to a very strong statement about what the relevant output of the channel is.

We understand Reviewer #1’s concern regarding the choice of log binning. We found that if the number of binning is higher than 200, no matter the binning methods, including linear, logarithmic or equal frequency, the estimated channel capacities in each binning number are converged into the same value. The only difference is how quickly the values approach the converged channel capacity as increasing the binning number (shown in Figure R2). In the revised manuscript, we used linear binning to represent more relevant protein signaling as the Reviewer mentioned. Note that the channel capacity values calculated from linear binning do not show noticeable different from our previously calculated channel capacity values.

On the other hand, linear binning generates significant bias, if we consider labelled input (i.e., continuous input) into channel capacity calculation, due to the increase of binning in input region.

Figure R2. Output binning number and binning method dependence of channel capacity value for experimental dataset. The inset plots show the relative difference of channel capacity value to the maximum channel capacity value in the entire binning range (i.e., from 10 to 1000) of the corresponding binning method.

According to Reviewer #1’s comment we have changed the binning method from logarithmic binning to linear binning in the whole experimental dataset except in the presence of labelled input (i.e., dectin-2 antibody). If we consider channel capacity between labelled input and NF-kB reporter, equal frequency binning is used for every layer of the channel capacity (i.e., labelled input-binding, binding-GFP, labelled input-GFP)

Reviewer #2 (Public Review):

My expertise is more on the theoretical than the experimental aspects of this paper, so those will be the focus of these comments.

Signal transduction is an important area of study for mathematical biologists and biophysicists. This setting is a natural one for information-theoretic methods, and such methods are attracting increasing research interest. Experimental results that attempt to directly quantify the Shannon capacity of signal transduction are particularly interesting. This paper represents an important contribution to this emerging field.

My main comments are about the rigorousness and correctness of the theoretical results. More details about these results would improve the paper and help the reader understand the results.

We understand reviewer #2’s comment related with rigorousness and correctness of the theoretical results of this work. In the revised manuscript, we added following contents to help the reader to better understand the channel capacity calculation procedures.

• General illustrative introduction regarding how we measured input and output dataset and how we handle those data to prepare joint probability distribution shown in SI section 1.1 and 1.2.

• Exemplified mutual information maximization procedure using experimental and arbitrary dataset shown in SI section 1.3.

The calculation of channel capacity, given in the methods, is quite a standard calculation and appears to be correct. However, I was confused by the use of the "weighting value" w_i, which is not specified in the manuscript. The input distribution appears to be a product of the weight w_i and the input probability value p_i, and these appear always to occur together as a product w_i p_i. (In joint probabilities w_i p(i,j), the input probability can be extracted using Bayes' rule, leaving w_i p_i p(j|i).) This leads met wonder two things. First, what role does w_i play (is it even necessary)? Second, of particular interest here is the capacity-achieving input distribution p_i, but w_i obscures it; is the physical input distribution p_i equal to the capacity-achieving distribution? If not, what is the meaning of capacity?

We thank Reviewer #2’s comment regarding the arbitrariness of the weightings. We realize there was a lack of explanation on the weighting values in the original manuscript. 𝑃x(𝑖) is a marginal probability distribution of input from the original dataset and 𝑃x'(𝑖) is the marginal probability distribution of modified input that maximize the mutual information. In usual case 𝑃x(𝑖) is not equal to 𝑃x'(𝑖) and therefore one needs to find 𝑃x'(𝑖) from 𝑃x(𝑖). Because 𝑃x'(𝑖) is a linear combination of 𝑃x(𝑖), it can be expressed as 𝑤(𝑖)𝑃x(𝑖) , where 𝑤(𝑖) is the weightings, under constraint ∑input/i 𝑤(𝑖)𝑃x (𝑖) = 1 . The changed input distribution, in turn, modifies the joint probability distribution as 𝑃'xy (𝑖, 𝑗) = 𝑤(𝑖)𝑃xy)(𝑖, 𝑗). To help readers understand of this work we expanded the Appendix with illustrative descriptions.

A more minor but important point: the inputs and outputs of the communication channel are never explicitly defined, which makes the meaning of the results unclear. When evaluating the capacity of an information channel, the inputs X and outputs Y should be carefully defined, so that the mutual information I(X;Y) is meaningful; the mutual information is then maximized to obtain capacity. Although it can be inferred that the input X is the ligand concentration, and the output Y is the expression of GFP, it would be helpful if this were stated explicitly.

We agree with Reviewer’s suggestion for better description of input and output in the manuscript. Therefore, we have modified Figure 1 A and B and the main text to describe the source of input and output much clearly, as follows:

Line 92: Accounting for the stochastic behavior of cellular signaling, information theory provides robust and quantitative tools to analyze complex communication channels. A fundamental metric of information theory is entropy, which determines the amount of disorder or uncertainty of variables. In this respect, cellular signaling pathways having high variability of the initiating input signals (e.g. stimulants) and the corresponding highly variable output response (i.e. cellular signaling) can be characterized as a high entropy. Importantly, input and output can have mutual dependence and therefore knowing the input distribution can partly provide the information of output distribution. If noise is present in the communication channel, input and output have reduced mutual dependence. This mutual dependence between input and output is called mutual information. Mutual information is, therefore, a function of input distribution and the upper bound of mutual information is called channel capacity (SI section 1) (Cover and Thomas, 2012). In this report, a communication channel describes signal transduction pathway of C-type lectin receptor, which ultimately lead to NF-κB translocation and finally GFP expression in the reporter model (Fig. 1A). To quantify the signaling information of the communication channels, we used channel capacity. Importantly, the channel capacity isn’t merely describing the resulting maximum intensity of the reporter cells. The channel capacity takes cellular variation and activation across a whole range of incoming stimulus of single cell resolved data into account and quantifies all of that data into a single number.

Author Response

Reviewer #3 (Public Review):

The authors examine the role of secreted BAFF in senescence phenotypes in THP1 AML cells and primary human fibroblasts. In the former, BAFF is found to potentiate the inflammatory phenotype (SASP) and in the latter to potentiate cell cycle arrest. This is an important study because the SASP is still largely considered in generic and monolithic terms, and it is necessary to deconvolute the SASP and examine its many components individually and in different contexts.

Although the results show differences for BAFF in the two cell models, there are many places where key results are missing and the results over-interpreted and/or missing controls.

1) Figure 1. Test whether the upregulation of BAFF is specific to senescence, or also in reversible quiescence arrest.

We appreciate the Reviewer’s requests. We performed the experiments in fibroblasts and THP-1 cells to assess BAFF levels in quiescence. As shown below in the figure for Reviewers, we induced quiescence in fibroblasts by serum starvation (0.1%) for 96 h and confirmed the quiescent state by measuring two markers of quiescence (reduction of CCND1 mRNA and reduction of phopho-S6, when compared to cycling cells, following markers established previously (PMID 25483060) (panel A). In this case, the level of BAFF mRNA was increased upon quiescence (panel B).

In THP-1 cells, we tried to induce quiescence by serum starvation and glutamine depletion for 96 h. Unfortunately, however, inducing quiescence in THP-1 cells was rather challenging, likely because they are cancer cells. Thus, we observed a reduction of cell proliferation in both conditions, but we observed a reduction in phospho-S6 only in the samples without glutamine (panel C). We failed to see increased BAFF mRNA levels in quiescent THP-1 cells after either serum starvation or glutamine depletion (panel D).

In summary, further studies will be necessary to fully understand if the increased expression of BAFF seen in senescent cells is also observed in other conditions of growth suppression (such as quiescence or differentiation), as well as whether this effect is specific to different cell types.

2) Figure 1, Supplement 1G. Show negative control IgG for immunofluorescence.

We thank the Reviewer for this suggestion. Along with other changes during the revision, we decided to remove the immunofluorescence data in order to include more informative data.

3) All results with siRNA should be validated with at least 2 individual siRNAs to eliminate the possibility of off-target effects.

We agree with the Reviewer on the importance of testing individual siRNAs. For BAFF, we originally tested two independent siRNAs (BAFF#1 and BAFF#2) individually, but we also pooled them for additional analysis (and referred to simply as “BAFFsi” along the manuscript). In the revised version of our manuscript, we included the key experiments performed with these two individual BAFF siRNAs. Upon BAFF silencing in THP-1 cells, we observed a reduction of SASP factors and SA-β-Gal activity levels with each individual siRNA (Figure 4-Figure Supplement 1D-F) and with the pooled siRNAs (Figure 4C). For WI-38 cells, we observed a reduction of p53 levels with individual and pooled siRNAs (Figure 7-Figure Supplement 1A), as well as a reduction in IL6 levels and SA-β-Gal activity (Figure 6-Figure Supplement 1D,E). After IRF1 silencing, we observed a reduction in BAFF pre-mRNA with two different pairs of CTRLsi and IRF1si pools (Figure 2I and supplementary Figure 2E). For the data on BAFF receptors, we used SMARTpools from Dharmacon, which are combinations of 4 siRNAs designed by the company to minimize off-target effects. These additions and clarifications are indicated in the revised manuscript.

4) To confirm a role for IRF1 in the activation of BAFF, the authors should confirm the binding of IRF1 to the BAFF promoter by ChIP or ChIP-seq.

We thank the Reviewer for this suggestion. We performed ChIP-qPCR analysis in THP-1 cells that were either proliferating or rendered senescent after exposure to IR (Figure 2H, Materials and methods section), and we confirmed the binding of IRF1 to the proximal promoter region of BAFF. As anticipated, this interaction was stronger after inducing senescence.

5) Key antibodies should be validated by siRNA knockdown of their targets, for example, TACI, BCMA, and BAFF-R in Figure 5. Note that there is an apparent discrepancy between BCMA data in Figure 5B vs 5C.

We fully agree with the Reviewer on this point and we thank him/her for helping us to improve this part of our manuscript. To address the discrepancy regarding BCMA western blot analysis and flow cytometry data, we silenced BCMA in THP-1 cells and tested two different antibodies advertised to recognize BCMA. This experiment allowed us to identify the correct band for BCMA by western blot analysis. We then confirmed that BCMA is upregulated in senescence, as observed by both western blot and flow cytometry analyses. We have modified the manuscript to reflect these changes. Please find these data in Figure 5A,B and Figure 5-Figure Supplement 1A of the revised manuscript.

6) Figure 5E. Negative/specificity controls for this assay should be shown.

We thank the reviewer for this comment and regret that we were unable to provide a negative control. The kit only provides a competitive wild-type oligomer used to test the specificity of the binding. For each sample (CTRLsi, BAFFsi, CTRLsi IR, BAFFsi IR) and each antibody tested (p65, p50, p52, RelB and c-Rel), we evaluated the reductions in signal upon addition of excess competitive oligomer per well (20 pmol/well) compared to wells with an inactive oligomer. However, the negative control was performed only as single replicate, due to the limited quantity of nuclear extracts and the high number of samples and antibodies analyzed. We therefore considered this control as being ‘qualitative’ rather than fully ‘quantitative’.

7) Hybridization arrays such as Figure 5H, Figure 6 - Supplement 1I, and Figure 6H should be shown as quantitated, normalized data with statistics from replicates.

We appreciate this request. We have included the quantification and statistics to the phosphoarrays used for THP-1 and WI-38 cells, which had been performed in triplicate (Figure 7A, Figure 5-Figure Supplement 1D). The original arrays are shown in the respective Source Data Files. In the interest of space, we removed the cytokine array performed on IMR-90 cells and left instead the quantitative ELISA for IL6 (Figure 6-Figure Supplement 1F). The data obtained from the cytokine array analysis in Figure 4F and Figure 4-Supplemental Figure 1C are supported by quantitative multiplex ELISA measurements (Figure 4E and Figure 4C).

8) Figure 6B - Supplement 1. Controls to confirm fractionation (i.e., non-contamination by cytosolic and nuclear proteins) should be shown.

We thank the Reviewer for this suggestion. We tested the efficiency of fractionation and we did in fact observe some degree of contamination from cytosolic proteins using the earlier version of the kit (Pierce, cat. 89881). We therefore purchased an improved version of the kit (Pierce, cat. A44390) and repeated the surface fractionation assay, which this time showed improved fractionation (Figure 7-Figure Supplement 1B). Interestingly, with the improved fractionation strategy, we observed that BAFF receptors in fibroblasts were almost exclusively localized inside the cell and not on the surface, as we found in THP-1 cells. Further validation of BAFF receptor antibodies has been provided in Figure 5-Figure Supplement 1A. As described in the text, the intracellular localization of BAFF receptors was previously reported in other cell types and conditions (PMID 31137630, PMID 19258594, PMID 30333819, PMID 10903733), and thus it is possible that BAFF may act through non-canonical mechanisms in WI-38 cells. Nonetheless, we did detect a small amount of BAFFR on the cell surface, and furthermore, BAFFR silencing reduced the level of p53 in fibroblasts. Therefore, we propose that BAFFR may be the primary receptor involved in p53 regulation in fibroblasts (Figure 7-Figure Supplement 1B,C). Our data on BAFF receptors deserve deeper characterization in a future study of the functions of BAFF receptors in senescence.

9) Figure 6A. Knockdown of BAFF should be shown by western blot.

Yes, definitely. We appreciate this comment and have included BAFF knockdown data in fibroblasts by western blot analysis (Figure 7B).

10) Figure 6G. Although BAFF knockdown decreases the expression of p53, p21 increases. How do the authors explain this?

We thank the Reviewer for the interesting question. We too were surprised to observe that the p53-dependent transcripts regulated by BAFF did not include CDKN1A (p21) mRNA, as confirmed by western blot analysis. The accumulation of p21 in senescence can be also regulated by p53-independent pathways and in p53-/- cells, for example by p90RSK, SP1, and ZNF84 (PMID 24136223, PMID 25051367, PMID 33925586). Eventually, we removed the data relative to p21 and γ-H2AX in favor of other data and to streamline the content of this manuscript for the reader.

Author Response

Reviewer #1 (Public Review):

1-1. I do have some concerns that the differences in network clustering reported in Fig 6 may be due to noise and I think the comparisons against the HCP parcellation could be more robust. Specifically, with regard to the network clustering in Fig 6. The authors use a clustering algorithm (which is not explained) to cluster the parcels into different functional networks. They achieve this by estimating the mean time series for each parcel in each individual, which they then correlate between the n regions, to generate an nxn connectivity matrix. This they then binarise, before averaging across individuals within an age group. It strikes me that binarising before averaging will artificially reduce connections for which only a subset of individuals are set to zero. Therefore averaging should really occur before binarising. Then I think the stability of these clusters should be explored by creating random repeat and generation groups (as done for the original parcells) or just by bootstrapping the process. I would be interested to see whether after all this the observation that the posterior frontoparietal expands to include the parahippocampal gryus from 3-6 months and then disappears at 9 months - remains.

We thank the reviewer for this insightful comment on our clustering process. For the step of “binarizing before averaging”, we followed the method proposed by Yeo et al (1). In this method, all correlation matrices are binarized according to the individual-specific thresholds. Specifically, each individual-specific threshold is determined according to the percentile, and only 10% of connections are kept and set to 1, while all other connections are set to 0. Yeo et al. (1) explained their motivation for doing so as “the binarization of the correlation matrix leads to significantly better clustering results, although the algorithm appears robust to the particular choice of the threshold”. We consider that the possible reason is that the binarization of connectivity in each individual offers a certain level of normalization so that each subject can contribute the same number of connections. If averaging occurs before binarizing, the actual connectivity contributed by different subjects would be different, which leads to bias. Meanwhile, we tested the stability of ‘binarizing first’ and ‘averaging first’, and the result is shown in Fig. R1 below. This figure suggests a similar conclusion as (1), where binarizing first before averaging leads to better clustering stability. We added the motivation of binarizing before averaging in the revised manuscript between line 577 and line 581.

Fig. R1. The comparison of clustering stability of different methods. The red line refers to the clustering stability when binarizing the correlation matrices first and then averaging the matrices across individuals, while the blue line refers to the clustering stability when averaging the correlation matrices across individuals first and then binarizing the average matrix.

For the final clustering results, we performed our clustering method using bootstrapping 100 times, and the final result is a majority voting of each parcel. The comparison of these two results is shown in Fig. R2. Overall, we do observe good repeatability between these two results. However, we also observed that some parcels show different patterns between the two results, especially for those parcels that are spatially located around the boundaries of networks or the medial wall. The pattern of the observation that “the posterior frontoparietal expands to include the parahippocampal gyrus from 3-6 months and then disappears at 9 months – remains” was not repeated in the bootstrapped results. These results might suggest that the clustering method is quite robust, the discovered patterns are relatively stable, and the differences between our original results and bootstrapping results might be caused by noises or inter-subject variabilities.

Fig. R2. Top panel: the network clustering results using all data in the original manuscript. Bottom panel: the network clustering results using majority voting through 100 times of bootstrapping. Black circles and red arrows point to the parahippocampal gyrus, which was included in the posterior frontoparietal network, and is not well repeated in the bootstrapped results. (M: months)

1-2. Then with regard to the comparison against the HCP parcellation, this is only qualitative. The authors should see whether the comparison is quantitatively better relative to the null clusterings that they produce.

Thank you for this great suggestion! As suggested, we added this quantitative comparison using the Hausdorff distance. Similar to the comparison in parcel variance and homogeneity, the 1,000 null parcellations were created by randomly rotating our parcellation with small angles on the spherical surface 1,000 times. We compared our parcellation and the null parcellations by accordingly evaluating their Hausdorff distances to some specific areas of the HCP parcellation on the spherical space, including Brodmann's area 2, 3b, 4+3a, 44+45, V1, and MT+MST. The results are listed in Figure 4. From the results, we can observe that our parcellation generally shows statistically much lower Hausdorff distances to the HCP parcellation, suggesting that our parcellation generates parcel borders that are closer to HCP parcellations compared to the null parcellations.

However, we noticed very few null parcellations that show smaller Hausdorff distances compared to our parcellation. A possible reason comes from our surface registration process with the HCP template purely based on cortical folding, without using functional gradient density maps, which are not available in the HCP template. As a result, this does not ensure high-quality functional alignment between our infant data and the HCP space, thus inevitably increasing the Hausdorff distance between our parcellation and the HCP parcellation.

1-3. … not all individuals appear (from Fig 8) to be acquired exactly at the desired timepoints, so maybe the authors might comment on why they decided not to apply any kernel weighted or smoothing to their averaging? Pg. 8 'and parcel numbers show slight changes that follow a multi-peak fluctuation, with inflection ages of 9 and 18 months' explain - the parcels per age group vary - with age with peaks at 9 and 18 - could this be due to differences in the subject numbers, or the subjects that were scanned at that point?

We do agree with the reviewer that subjects are not scanned at similar time points. This is designed in the data acquisition protocol to seamlessly cover the early postnatal stage so that we will have a quasi-continuous observation of the dynamic early brain development.

We didn’t apply kernel weighted average or smoothing when generating the parcellation, as we would like each scan to contribute equally, and each parcellation map could be representative of the cohort of the covered age, instead of only part of them. Meanwhile, our final ‘age-common parcellation’ could be representative of all subjects from birth to 2 years of age. However, we do agree that the parcellation map that is only designed for the use of a specific age, e.g., 1-year-olds, kernel weighted average, or even a more restricted age range could be a more appropriate solution.

For the parcel number that likely shows fluctuations with subject numbers, we added an experiment, where we randomly selected 100 scans by considering the minimum scan number in each age group using bootstrapping and repeated this process 100 times. The average parcel number of each age is reported in the following Table R1. We didn’t observe strong changes in parcel numbers when reducing scan numbers, which further demonstrates that our parcel numbers do not show a strong relation to subject numbers. However, the parcel number does not increase greatly from 18M to 24M in the bootstrapping results, so we modified the statement in the manuscript about the parcel number to ‘… all parcel numbers fall between 461 to 493 per hemisphere, where the parcel number attains a maximum at around 9 months and then reduces slightly and remains relatively stable afterward. …’, which can be found between line 121 and line 122.

1-4. I also have some residual concerns over the number of parcels reported, specifically as to whether all of this represents fine-grained functional organisation, or whether some of it represents noise. The number of parcels reported is very high. While Glasser et al 2016 reports 360 as a lower bound, it seems unlikely that the number of parcels estimated by that method would greatly exceed 400. This would align with the previous work of Van Essen et al (which the authors cite as 53) which suggests a high bound of 400 regions. While accepting Eickhoff's argument that a more modular view of parcellation might be appropriate, these are infants with underdeveloped brain function.

We thank the reviewer for this insightful comment. We agree that there might be noises for some of the parcels, as noises exist in each step, such as data acquisition, image processing, surface reconstruction, and registration, especially considering functional MRI is noisier than structural MRI. Though our experiments show that our parcellation is fine-grained and is suitable for the study of the infant brain functional development, it is hard to directly quantitatively validate as there is no ground truth available.

Despite these, we are still motivated to create fine-grained parcellations, as with the increase of bigger and higher resolution imaging data and advanced computational methods, parcellations with more fine-grained regions are desired for downstream analyses, especially considering the hierarchical nature of the brain organization (2). And the main reason that our method generates much finer parcellation maps, is that both our registration and parcellation process is based on the functional gradient density, which characterizes a fine-grained feature map based on fMRI. This leads to both better inter-subject alignment in functional boundaries and finer region partitions. This strategy is different from Glasser et al (3), which jointly considers multimodal information for defining parcel boundaries, thus parcels revealed purely by functional MRI might be ignored in the HCP parcellation. We hope our parcellation framework can be a useful reference for this research direction. We added this discussion in the revised manuscript between line 268 and line 271.

For the parcel number, even without performing surface registration based on fine-grained functional features, recent adult fMRI-based parcellations greatly increased parcel numbers, such as up to 1,000 parcels in Schaefer et al. (4), 518 parcels in Peng et al. (5), and 1,600 parcels in Zhao et al. (6). For infants, we do agree that the infant functional connectivity might not be as strong as in adults. However, there are opinions (7-9) that the basic units of functional organization are likely to present in infant brains, and brain functional development gradually shapes the brain networks. Therefore, the functional parcel units in infants could be possibly on a comparable scale to adults. Even so, we do agree that more research needs to be performed on larger datasets for better evaluations. We added this discussion in the revised manuscript between line 275 and line 280.

1-5. Further comparisons across different subjects based on small parcels increases the chances of downstream analyses incorporating image registration noise, since as Glasser et al 2016 noted, there are many examples of topographic variation, which diffeomorphic registration cannot match. Therefore averaging across individuals would likely lose this granularity. I'm not sure how to test this beyond showing that the networks work well for downstream analyses but I think these issues should be discussed.

We agree with the reviewer that averaging across individuals inevitably brings some registration errors to the parcellation, especially for regions with high topographic variation across subjects, which would lead to loss of granularity in these regions. We believe this is an important issue that exists in most methods on group-level parcellations, and the eventual solution might be individualized parcellation, which will be our future work. We added this discussion in the revised manuscript between line 288 and line 292.

We also agree with the reviewer that downstream analyses are important evaluations for parcellations. We provided a beta version of our parcellation with 602 parcels (10) to our colleagues, and they tested our parcellation in the task of infant individual recognition across ages using functional connectivity, to explore infant functional connectome fingerprinting (10). We compared the performance of different parcellations with 602 ROIs (our beta version), 360 ROIs (HCP MMP parcellation (3)), and 68 ROIs (FreeSurfer parcellation (11)). The results (Fig. R3) show that our parcellation with a higher parcellation number yields better accuracy compared to other parcellations. We added a description of this downstream application in the discussion between line 284 and line 287.

Fig. R3. The comparison of different parcellations for infant individual recognition across age based on functional connectivity (figure source: Hu et al. (10)). The parcellation with 602 ROIs is the beta version of our parcellation, 360 ROIs stands for HCP MMP parcellation (3) and 68 ROIs stands for the FreeSurfer parcellation (11). This downstream task shows that a higher parcellation number does lead to better accuracy in the application.

1-6. Finally, I feel the methods lack clarity in some areas and that many key references are missing. In general I don't think that key methods should be described only through references to other papers. And there are many references, particular to FSL papers, that are missing.

We thank the reviewer for this great suggestion. We added related references for FLIRT, FSL, MCFLIRT, and TOPUP For the alignment to the HCP 32k_LR space, we first aligned all subjects to the fsaverage space using spherical demons, and then used part of the HCP pipeline (12) to map the surface from the fsaverage space to HCP 164k_LR space, and downsampled to 32k_LR space. We modified this citation by referencing the HCP pipeline by Glasser et al. (12) instead and detailed this registration process in the revised manuscript between line 434 to line 440 in the revised manuscript and as below:

“… The population-mean surface maps were mapped to the HCP 164k ‘fs_LR’ space using the deformation field that deforms the ‘fsaverage’ space to the ‘fs_LR’ space released by Van Essen et al. (13), which was obtained by landmark-based registration. By concatenating the three deformation fields of steps 1, 3, and 4, we directly warped all cortical surfaces from individual scan spaces to the HCP 164k_LR space and then resampled them to 32k_LR using the HCP pipeline (12), thus establishing vertex-to-vertex correspondences across individuals and ages …”

Reviewer #2 (Public Review):

2-1. Diminishing enthusiasm is the lack of focus in the result section, the frequent use of jargon, and figures that are often difficult to interpret. If those issues are addressed, the proposed atlas could have a high impact in the field especially as it is aligned with the template of the Human Connectome Project.

We’d like to thank Reviewer #2 for the appreciation of our atlas. According to the reviewer’s suggestion, we went through the manuscript again by focusing on correcting the use of jargon, clarity in the result section, as well as figures and figure captions. We hope our corrections can help explain our work to a broader community. Our revisions are accordingly detailed in the following. Meanwhile, our parcellation maps have been aligned with the templates in HCP and FreeSurfer and made available via NITRC at: https://www.nitrc.org/projects/infantsurfatlas/.

References

1. B. Thomas Yeo, F. M. Krienen, J. Sepulcre, M. R. Sabuncu, D. Lashkari, M. Hollinshead, J. L. Roffman, J. W. Smoller, L. Zöllei, J. R. Polimeni, The organization of the human cerebral cortex estimated by intrinsic functional connectivity. Journal of neurophysiology 106, 1125-1165 (2011).

2. S. B. Eickhoff, R. T. Constable, B. T. Yeo, Topographic organization of the cerebral cortex and brain cartography. NeuroImage 170, 332-347 (2018).

3. M. F. Glasser, T. S. Coalson, E. C. Robinson, C. D. Hacker, J. Harwell, E. Yacoub, K. Ugurbil, J. Andersson, C. F. Beckmann, M. Jenkinson, S. M. Smith, D. C. Van Essen, A multi-modal parcellation of human cerebral cortex. Nature 536, 171-178 (2016).

4. A. Schaefer, R. Kong, E. M. Gordon, T. O. Laumann, X.-N. Zuo, A. J. Holmes, S. B. Eickhoff, B. T. J. C. C. Yeo, Local-global parcellation of the human cerebral cortex from intrinsic functional connectivity MRI. 28, 3095-3114 (2018).

5. L. Peng, Z. Luo, L.-L. Zeng, C. Hou, H. Shen, Z. Zhou, D. Hu, Parcellating the human brain using resting-state dynamic functional connectivity. Cerebral Cortex, (2022).

6. J. Zhao, C. Tang, J. Nie, Functional parcellation of individual cerebral cortex based on functional mri. Neuroinformatics 18, 295-306 (2020).

7. W. Gao, S. Alcauter, J. K. Smith, J. H. Gilmore, W. Lin, Development of human brain cortical network architecture during infancy. Brain Structure and Function 220, 1173-1186 (2015).

8. W. Gao, H. Zhu, K. S. Giovanello, J. K. Smith, D. Shen, J. H. Gilmore, W. J. P. o. t. N. A. o. S. Lin, Evidence on the emergence of the brain's default network from 2-week-old to 2-year-old healthy pediatric subjects. 106, 6790-6795 (2009).

9. K. Keunen, S. J. Counsell, M. J. J. N. Benders, The emergence of functional architecture during early brain development. 160, 2-14 (2017).

10. D. Hu, F. Wang, H. Zhang, Z. Wu, Z. Zhou, G. Li, L. Wang, W. Lin, G. Li, U. U. B. C. P. Consortium, Existence of Functional Connectome Fingerprint during Infancy and Its Stability over Months. Journal of Neuroscience 42, 377-389 (2022).

11. R. S. Desikan, F. Ségonne, B. Fischl, B. T. Quinn, B. C. Dickerson, D. Blacker, R. L. Buckner, A. M. Dale, R. P. Maguire, B. T. Hyman, An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage 31, 968-980 (2006).

12. M. F. Glasser, S. N. Sotiropoulos, J. A. Wilson, T. S. Coalson, B. Fischl, J. L. Andersson, J. Xu, S. Jbabdi, M. Webster, J. R. Polimeni, The minimal preprocessing pipelines for the Human Connectome Project. NeuroImage 80, 105-124 (2013).

Author Response

Reviewer #1 (Public Review):

The authors present a strong set of experiments to uncover what type of role non-mutant stromal cells might be playing in the development of VM and AST, two vascular lesions that share some similarities.

Questions about experimental design.

1) For quantification of gene expression in VM and AST specimens in Figure 2, the methods say qPCR data were normalized to housekeeping genes, but it would be helpful to normalize to endothelial content. It might be that increased TGFa is due to increased endothelium.

We thank the Reviewer for this excellent suggestion. We have now added this new data as suggested with normalization of TGFA mRNA to the endothelial marker PECAM-1/CD31 mRNA. A trend towards an increased expression of TGFA mRNA was detected in VM/AST specimens in comparison to the control group. We also show in the manuscript that besides CD31-positive vascular structures, TGFA is expressed in intervascular areas, i.e. between the vessels, in the patients’ lesions (Fig.2) and in lesion-derived CD31negative intervascular stromal cells. These data altogether demonstrate that i) TGFA is expressed also in other cell types than endothelial cells and ii) indicates that the increased expression of TGFA in lesion samples is not only due to increased vasculature/endothelium in the patient samples.

The new RT-qPCR data has now been added to the manuscript as a new Fig. 2 - figure supplement 1.

2) The mutant allelic frequency for the HUVEC-PIK3CA WT versus HUVEC-PIK3CA H1047R should be provided. This is critically needed for the interpretation of the results.

Thank you for this valuable comment. To confirm that PIK3CA H1047R is still present in transduced HUVECs at the end-point of the mouse xenograft experiment, we performed a new ddPCR analysis detecting fractional abundance of PIK3CA p.H1047R from the matrigel plug-in samples. In this new data, mean fractional abundance of PIK3CA p.H1047R in fibroblast containing PIK3CA H1047R EC plugs was shown to be 27.1 % (variation 26.5-27.8 %; n=2 mice in duplicates). This corresponds to ~54 % of PIK3CA p.H1047R mutation positive cells in the plug, assuming a single copy of the mutation in each cell. As a control group, no positivity was detected in samples with fibroblast and in PIK3CAwt EC, as all the cells express the wildtype form of the PIK3CA gene. Please see Author Response Image 1 representative 2D amplification plots of the mutation analysis. Fractional abundances of PIK3CA mutations in the patient tissue samples and patient-derived CD31+ cells can also be seen in Table 1 and were in a range of 5-12 % (whole tissue) and 44-51 % (EC fraction).

3) From Figure 5, it appears that the human primary fibroblasts are not required for the mutant ECs to form perfused vessels (panel H).

We thank the Reviewer for the comment and agree that based on our H&E staining and erythrocyte analysis, perfused vessels are evident in PIK3CA mutant plugs containing ECs with fibroblasts but also in plugs containing ECs alone. This was expected as PIK3CA mutation in ECs alone has shown to be a driver of venous malformation. However, prior to our study the role of fibroblasts in PIK3CA-driven lesions had not been studied. To better understand the role of fibroblasts in lesion formation, we have now added new data to the manuscript containing example images of the PIK3CA H1047R plugs with or without fibroblasts, and added a new quantitation of their erythrocyte amount. Please see Author Response Image 2. Our data demonstrates that there are significantly: i) more CD31-positive vascular structures (Fig. 5E-G), ii) larger lumens (Fig. 5D-F) and iii) more erythrocyte-containing regions, indicative for perfused vessels (new Fig. 5H) in lesions with fibroblasts in comparison to plugs containing ECs alone. This implies that fibroblasts further induce PIK3CA-driven EC lesion formation.

Author Response Image 2. Vascular structures formed with PIK3CA H1047R ECs alone and PIK3CA H1047R ECs + FBs in mouse xenograft plugs. In the figure panel, H&E staining on each individual plug in these groups is presented. Equal size close-up images were taken from the middle of each plug covering > 50% of plug area (scale bar 250µm). More erythrocytes (red) are seen in the plugs with fibroblasts in comparison to ECs alone. Scanned images of the H&E stained whole tissue sections can be seen in the Fig. 5 – source file.

A new quantitative analysis of erythrocyte positive area in relation to whole plug area using SproutAngio quantification tool was additionally performed (). Analysis was done on a blinded manner and showed significantly increased erythrocyte amount in the plugs containing PIK3CA H1047R ECs and fibroblasts (in comparison to EC alone). Describtion of the analysis has now been added in the manuscript (p. 42, rows 839-843) Figures 5G and 5H in the manuscript were updated to show statistics and automated intensitybased quantification of the erythrocyte positive area/ plug instead of erythrocyte scoring (scale 0-3).

Is it possible that TGFa from the ECs is sufficient to drive vascular malformation?

Mutations in genes such as PIK3CA, TEK and KRAS have been shown to drive formation of vascular anomalies. Thus it is unlikely that a single growth factor, such as increased expression of TGFA, would drive this process alone. That being said, our data shows that TGFA is able to regulate proliferation of PIK3CA mutated ECs via secondary mechanism (Fig. 4F), and we show that inhibition of EGFR pathway is able to reduce PIK3CA-driven lesion growth in mice (Fig. 7). As our bulk RNA-sequencing data from patient-derived cells, showed expression of also other growth factors in lesion ECs (Table 3), it is likely that multiple angiogenic growth factors are involved in lesion formation similarly as in tumors and their expression is primarily driven by mutated cells and secondary by cell-cell crosstalk with other lesion cell types. Thus, targeting of multiple signalling pathways could be a beneficial treatment strategy in the future.

Reviewer #2 (Public Review):

In this manuscript, Ilmonen H. et al explored potential crosstalk between endothelial cells and fibroblasts in a context of sporadic vascular malformation (venous malformation and angiomatoses of soft tissue). With a high level of evidence, they found that mutated endothelial cells secrete TGFA that will activate surrounding fibroblasts, leading in turn to VEGFA secretion that will stimulate endothelial cell sprouting and vascular malformation development. Experiments are well-designed and support their hypothesis. Some controls are missing, particularly in Fig. 2. Indeed, it is mandatory to provide data from healthy skin biopsies (that are available in many laboratories): TGFa, CD31, P-EGFR staining.

We thank the Reviewer for the comments. Although it is common that VM presents in skin, in this work we solely focused on intramuscular and subcutaneous AST and VM patient samples and excluded the samples containing skin from this study. We did TGFA immunostainings from healthy skeletal muscle that can be seen Figure 2 – figure supplement 2B. CD31 staining of vessels in healthy skeletal muscle near the resection margin can be seen in Figure 1B. Please see below also tissue locations of all VM and AST samples in this study:

• Intramuscular, 42.1 % of lesions (n=16)

• Intramuscular and subcutaneous, 21.1 % of lesions (n=8)

• Intramuscular, subcutaneous and synovial membrane, 5.3 % of lesions (n=2)

• Intramuscular and synovial membrane, 2.6 % of lesions (n=1)

• Subcutaneous and synovial membrane, 2.6 % of lesions (n=1)

• Subcutaneous only, 26.3 % of lesions (n=10)

• Skin, none of the lesions

Author Response

Reviewer #1 (Public Review):

Redox signaling is a dynamic and concerted orchestra of inter-connected cellular pathways. There is always a debate whether ROS (reactive oxygen species) could be a friend or foe. Continued research is needed to dissect out how ROS generation and progression could diverge in physiological versus pathophysiological states. Similarly, there are several paradoxical studies (both animal and human) wherein exercise health benefits were reported to be accompanied by increases in ROS generation. It is in this context, that the present manuscript deserves attention.

Utilizing the in-vitro studies as well as mice model work, this manuscript illustrates the different regulatory mechanisms of exercise and antioxidant intervention on redox balance and blood glucose level in diabetes. The manuscript does have some limitations and might need additional experiments and explanation.

The authors should consider addressing the following comments with additional experiments.

1) Although hepatic AMPK activation appears to be a central signaling element for the benefits of moderate exercise and glucose control, additional signals (on hepatic tissue) related to hepatic gluconeogenesis such as Forkhead box O1 (FoxO1), phosphoenolpyruvate carboxykinase (PEPCK), and GLUT2 needs to be profiled to present a holistic approach. Authors should consider this and revise the manuscript.

We appreciate the constructive suggestion. Besides glycolysis, gluconeogenesis and glucose uptake are critical in maintaining liver and blood glucose homeostasis.

FoxO1 has been tightly linked with hepatic gluconeogenesis through inhibiting the transcription of gluconeogenesis-related PEPCK and G6Pase expression (1, 2). Herein, we found the expression of FoxO1 increased in the diabetic group but reduced in the CE, IE and EE groups (Fig. X1A, Fig.5E-F in manuscript). Meanwhile, the mRNA level of Pepck and G6PC (one of the three G6Pase catalytic-subunit-encoding genes) also decreased in the CE, IE, and EE groups (Fig. X1B-1C, Fig.5H-I in manuscript). These results indicates that these three modes of exercise all inhibited gluconeogenesis through down-regulating FoxO1.

For the glucose uptake, we detected the protein expression of GLUT2 in the liver tissue. Glut2 helps in the uptake of glucose by the hepatocytes for glycolysis and glycogenesis. Accordingly, we found GLUT2,a glucose sensor in liver, was up-regulated in diabetic rats, but down-regulated by the CE and IE intervention. However, GLUT2 didn’t decrease in the EE group, which is consistent with the results of the unimproved blood glucose by EE intervention (Figure X1A, Fig.5E and 5G in manuscript).

Taken together, moderate exercise could benefits glucose control through increasing glycolysis and decreasing gluconeogenesis. We added this part in Page 9 line 251-263 and Figure 5E-5I in this version.

Figure X1. A. Representative protein level and quantitative analysis of FOXO1 (82 kDa), GLUT2 (60-70 kDa) and Actin (45 kDa) in the rats in the Ctl, T2D, T2D + CE, T2D + IE and T2D + EE groups. C-D. Expression of hepatic Pepck and G6PC mRNA in the Ctl, T2D, T2D + CE, T2D + IE and T2D + EE groups were evaluated by real-time PCR analysis. Values represent mean ratios of Pepck and G6PC transcripts normalized to GAPDH transcript levels.