Author response:

We sincerely thank all reviewers for their thoughtful, detailed, and supportive evaluations of our manuscript. We are very pleased that the reviewers appreciated the integrative approach of our study, the quality of the imaging and analyses, and the insights provided into the parallel evolution of biomineralization mechanisms in sponges and corals.

We are carefully considering all the suggestions made, including those regarding the improvement of figure clarity and the clarification of certain image interpretations. These comments are extremely valuable, and we are preparing a detailed point-by-point reply to accompany our revised manuscript.

It was also brought to our attention that the links to the Zenodo repository were incorrect. We apologize for this oversight and any inconvenience it may have caused and will updae the links in our revised manuscript. In the meantime, the correct Zenodo repositories can be accessed using the following links:

https://zenodo.org/records/14755899

https://zenodo.org/records/13847772

We again thank the reviewers for their constructive feedback, which will help us to further strengthen the manuscript.

Author response:

We thank the editors and reviewers for their thoughtful and constructive evaluation of our manuscript, “Krüppel Regulates Cell Cycle Exit and Limits Adult Neurogenesis of Mushroom Body Neural Progenitors in Drosophila.” We are pleased that all reviewers recognised the novelty and significance of identifying Krüppel (Kr) as a key transcription factor promoting timely termination of mushroom body neuroblast (MBNB) proliferation, and the potential antagonistic function of Kr-h1.

We appreciate the helpful suggestions aimed at improving the mechanistic clarity and presentation of our findings. Below, we outline how we plan to address the major points raised in the full revision.

(1) Characterisation of the KrIf-1 allele and Kr expression

We agree that clarifying the nature of the KrIf-1 allele is important. In response to this concern, we will examine Kr expression in KrIf-1 mutant larval, pupal, and adult brains using immunostaining and available reporter lines. These experiments will help determine whether the observed neuroblast retention phenotype correlates with altered Kr expression in MBNBs.

(2) Regulatory relationships between Kr, Kr-h1, Imp, Syp, Chinmo, and E93

We are currently performing additional experiments to clarify the interactions among these temporal factors. For instance, we are testing whether Kr-h1 overexpression alters the expression of Imp, Syp, and E93. We have obtained a published E93 antibody from Dr Chris Doe (Syed et al., 2017) and will include E93 expression analysis in our revised manuscript.

While Chinmo is of interest, its expression is well established to be regulated downstream of Imp/Syp via mRNA stability (Liu et al., 2015; Ren et al., 2017). Given that we currently lack reliable tools to assess Chinmo levels, we will focus primarily on Imp, Syp, and E93 as readouts for Kr/Kr-h1 function. If we succeed in obtaining Chinmo antibodies or reporter lines in time, we will include corresponding data.

(3) Expression of Kr-h1 in MBNBs

We fully agree that direct evidence for Kr-h1 expression in MBNBs is important. To address this, we have obtained the Kr-h1::GFP BAC transgenic line (BDSC #96786) and are currently using it to assess Kr-h1 expression in MBNBs. We also tested an anti–Kr-h1 antibody previously reported by Kang et al. (2017), developed in the context of fat body studies, but it did not yield clear signals in larval MBNBs. However, previous work by Shi et al. (2007) clearly demonstrated Kr-h1 expression in the developing MB, including MBNBs, using a custom antibody developed by their lab. We also contacted the Lee lab to request this antibody, but unfortunately, it is no longer available. We will include the results obtained using the GFP BAC line in the revised manuscript and, if needed, pursue RNA in situ hybridisation to further validate Kr-h1 expression in MBNBs.

(4) Temporal Kr knockdown and MARCM analysis

We appreciate the suggestion to validate our RNAi-based temporal knockdown results using MARCM. We plan to perform MBNB-specific MARCM analysis following the strategy described by Rossi et al. (2020). However, this approach requires additional time due to the logistics of acquiring the necessary fly stocks, generating appropriate genetic combinations, and conducting clonal analyses. While we will make every effort to include these data, we note that RNAi-based knockdown offers the advantage of temporal reversibility and has been essential for assessing stage-specific requirements in our current study.

(5) Details of the targeted genetic screen

Kr was initially identified as part of a broader, ongoing effort to screen for candidate transcription factors and cell cycle regulators involved in neuroblast cell cycle exit and/or quiescence. As this screen is still preliminary and incomplete, we prefer not to include the full dataset at this stage. Instead, we will revise the manuscript to clarify that Kr was prioritised for further investigation based on the striking MBNB-specific phenotype observed upon RNAi-mediated knockdown and in the KrIf-1 mutant, rather than through a completed screening process.

(6) Clarifying the model (Figure 6D) and interactions

We will revise the proposed model to distinguish between experimentally supported interactions and speculative ones. As noted above, we will primarily focus on the Imp/Syp and E93 axis in relation to Kr and Kr-h1 activity. Chinmo will be omitted from the model unless further data become available to support its inclusion.

(7) Clarifications on figures and data presentation

We appreciate the feedback on figure clarity. We will revise figures such as 1B, 2C, and 3A to improve legibility and presentation. We will also correct typographical errors and figure references, and clarify the activity patterns of the GAL4 drivers. Specifically, while UASmCD8::GFP expression driven by OK107-GAL4 is markedly weaker in MBNBs than in their neuronal progeny (as seen, for example, in Figure S3C), the driver remains active and functionally relevant in MBNBs. We believe the weak expression in MBNBs likely explains the absence of a NB retention phenotype in OK107>KrIR adult brains (see main text, Lines 374–376). As suggested by the reviewer, we will clarify this point earlier in the manuscript and can include additional data showing OK107>GFP expression patterns in pupal MB lineages as supplementary material.

(8) Analysis of public datasets

We will include results from our analysis of publicly available datasets such as FlyAtlas2, modENCODE, and a time-course RNA-seq dataset specific to MBNBs (Liu et al., 2015). While the spatial resolution of FlyAtlas2 and modENCODE is limited, the MBNB dataset provides valuable temporal information up to 36 h after puparium formation (APF). From this dataset, we observe that Kr expression remains consistently low throughout development, with only a modest increase at 84 h ALH (mean TPM ~11) and 36 h APF (~7), suggesting it does not undergo strong transcriptional regulation in MBNBs. In contrast, Kr-h1 is highly expressed during early larval stages (24–84 h ALH; mean TPM ~55–60) and shows a marked suppression by 36 h APF (mean TPM ~2), consistent with its proposed role in promoting MBNB proliferation. Importantly, Eip93F (E93) exhibits a reciprocal pattern to Kr-h1—with minimal expression until 84 h ALH (mean TPM ~24), followed by a substantial induction at 36 h APF (mean TPM ~104), aligning with its known role in triggering neuroblast termination. These temporal expression dynamics support our model that Kr-h1 and E93 function in opposition during the transition from proliferative to terminating neuroblast states. We will summarise these findings in the revised manuscript, along with appropriate discussion of dataset limitations.

We hope this provisional response conveys our strong commitment to thoroughly addressing the reviewers’ concerns and improving the manuscript. We are currently carrying out additional experiments and will submit a revised version with new data and enhanced clarity in due course.

References:

Kang et al., 2017. Sci Rep. 7(1):16369. doi: 10.1038/s41598-017-16638-1.

Shi et al., 2007. Dev Neurobiol. 67(11):1614–1626. doi: 10.1002/dneu.20537.

Rossi et al., 2020. eLife. 9:e58880. doi: 10.7554/eLife.58880.

Liu et al., 2015. Science. 350(6258):317–320. doi: 10.1126/science.aad1886.

Ren et al., 2017. Curr Biol. 27(9):1303–1313. doi: 10.1016/j.cub.2017.03.018. Syed et al., 2017. eLife. 6:e26287. doi: 10.7554/eLife.26287.

Author response:

Reviewer #1 (Public review):

Summary:

The manuscript by Garcia et al. describes how the expression of a respiratory chain alternative oxidase (AOX) from the tunicate Ciona intestinalis, capable of transferring electrons directly from reduced coenzyme Q (CoQ) to oxygen, is able to induce an increase in the mass of Drosophila melanogaster larvae and an accelerated development, especially when the larvae are kept at low temperatures. In order to explain this phenomenon, the paper addresses the modifications in the activity and levels of the 'canonical' electron transfer system (ETS), i.e., complexes I-IV and of the ATP synthase. In addition, the abundance of different metabolites as well as the NAD+/NADH ratios are measured, finding significant differences between the larvae.

Strengths:

The observations of differences in growth, body mass and food intake in the wt D. melanogaster larvae vs. those expressing the AOX transgene are solid. The evidence that mild uncoupling of the ETS might accelerate development of the fly larvae is convincing."

We appreciate the reviewer’s attention to our results and hope we can improve the manuscript to address all criticism appropriately.

Weaknesses:

Some of the observations, especially those concerning the origin of the metabolic remodelling in AOX-expressing larvae, are left unexplained, and the argumentation is somewhat speculative. What the authors mean by "reconfiguration" of the mitochondrial electron transfer system is not clear. If this implies that there is an actual change in ETS function and/or structural organisation in the presence of AOX, this conclusion is not supported by the experimental data. In addition, the influence of AOX activity in the mitochondrial ETS system is tested in vitro in the presence of saturating concentrations of substrates. The real degree to which AOX activity is actually influencing ETS activity in vivo remains unknown.

Indeed, the term “reconfiguration” may seem a little too strong. However, we do have preliminary structural data on larval mitochondria indicating that the term is adequate in this context. We plan to work on obtaining concrete data to sustain our claims that AOX imparts significant functional and structural remodeling of the organelle, which would be consistent with our respirometry and BN-PAGE data. If the data turns out not to be robust enough, we will consider replacing the term with one that better reflects our findings.

We also realize that the in vivo data we are presenting (body mass, mobility, food intake) are indirect measurements of metabolism and that a more direct approach is necessary to assess the real degree to which AOX influences ETS activity in vivo. To address this issue, we plan to expand our pharmacological treatments of the larval development and to measure whole larval oxygen consumption.

Reviewer #2 (Public review):

Summary:

This manuscript presents intriguing findings about the role of alternative oxidase (AOX) from the tunicate Ciona intestinalis in accelerating growth and development when expressed in Drosophila melanogaster.

Strengths:

The study is overall well-constructed, including appropriate analysis. Likewise, the manuscript is written clearly and supported by high-quality figures. The present study provides valuable insights into AOX's role in Drosophila development. The paper attempts to explore a unique mechanism by which AOX influences Drosophila development, providing insights into mitochondrial respiration and its physiological effects. This is relevant for understanding mitochondrial dysfunction and potential therapeutic applications. The study employs a variety of approaches, including calorimetry, infrared thermography, and genetic analyses, to investigate AOX's impact on metabolism and development.

We sincerely thank the reviewer for recognizing the strengths and acknowledging the novelty of our study.

Weaknesses:

There are a number of methodological limitations and substantial gaps in the interpretation of the data presented, which reduces the strength of its conclusions. For instance, there is a misunderstanding of the non-proton motive nature of the AOX - it does not uncouple respiration, merely decouple it as it neither contributes to nor dissipates the proton motive force, in contrast to chemical uncouplers or proton uncouplers such as UCPs. The authors need to reassess their data in light of the above.

The reviewer is absolutely right about the non-proton motive nature of AOX. We will reassess our data considering that AOX decouples respiration and, if necessary and possible, we will add new experiments to address the methodological limitations raised by the reviewer.

Author response:

We appreciate the reviewers' positive feedback on our paper. We especially thank them for their evaluation of the genetic analysis, which required a significant amount of timef time. We acknowledge that several aspects of our interpretation and description of the results need correction, as noted by both reviewers. Additionally, we recognize the importance of providing a more comprehensive overview of previous findings, including those conducted in mice, in the manuscript. In the revised version, we will thoroughly address the reviewers' concerns.

Both reviewers emphasized the need for further validation to ascertain whether the specific requirement of Hox genes in the Hoxba and Hoxbb clusters for pectoral fin bud formation is due to their expression patterns or the functional roles of Hox proteins. This consideration has been on our agenda for some time; however, our submitted paper does not sufficiently address this aspect. In the revised manuscript, we will conduct a comprehensive analysis of the expression patterns of Hox genes in zebrafish to draw informed conclusions on this matter.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

The authors investigate how the viscoelasticity of the fingertip skin can affect the firing of mechanoreceptive afferents and they find a clear effect of recent physical skin state (memory), which is different between afferents. The manuscript is extremely well-written and well-presented. It uses a large dataset of low threshold mechanoreceptive afferents in the fingertip, where it is particularly noteworthy that the SA-2s have been thoroughly analyzed and play an important role here. They point out in the introduction the importance of the non-linear dynamics of the event when an external stimulus contacts the skin, to the point at which this information is picked up by receptors. Although clearly correlated, these are different processes, and it has been very well-explained throughout. I have some comments and ideas that the authors could think about that could further improve their already very interesting paper. Overall, the authors have more than achieved their aims, where their results very much support the conclusions and provoke many further questions. This impact of the previous dynamics of the skin affecting the current state can be explored further in so many ways and may help us to better understand skin aging and the effects of anatomical changes of the skin.

At the beginning of the Results, it states that FA-2s were not considered as stimuli did not contain mechanical events with frequency components high enough to reliably excite them. Was this really the case, did the authors test any of the FA-2s from the larger dataset? If FA-2s were not at all activated, this is also relevant information for the brain to signal that it is not a relevant Pacinian stimulus (as they respond to everything). Further, afferent receptive fields that were more distant to the stimulus were included, which likely fired very little, like the FA-2s, so why not consider them even if their contribution was low?

Thank you for bringing this up, we have now clarified in the text that while FA-2s did respond at a low rate during the experiment, their responses were not reliably driven by the force stimuli. In the Methods section we have included the following text:

“Initially, 10 FA-2 neurons were also included in the analysis. But their responsiveness during the experiment was remarkably low, and unlike the other neuron types, their responses were rarely affected by force stimuli. Specifically, only one of the observed FA-2 neurons responded during the force protraction phases. Due to the lack of clear stimulus-driven responses, FA-2 neurons were subsequently excluded from further analysis.”

One question that I wondered throughout was whether you have looked at further past history in stimulation, i.e. not just the preceding stimulus, but 2 or 3 stimuli back? It would be interesting to know if there is any ongoing change that can be related back further. I do not think you would see anything as such here, but it would be interesting to test and/or explore in future work (e.g. especially with sticky, forceful, or sharp indentation touch). However, even here, it could be that certain directions gave more effects.

This is a very interesting question! A discernible effect from the previous stimulus could persist at the end of the current stimulation (see Figure 4C), potentially influencing the next one—a 2-stimuli-back effect. Unfortunately, our experimental design did not allow for rigorous testing of this effect. While all possible pairs of stimulus directions were included in immediately consecutive trials, this was not the case for pairs separated by additional trials. Hence, the combination of a likely weak effect and limited variation in history precluded a thorough analysis of a 2-stimuli-back effect. Future work should delve into the time course of the viscoelastic effect in greater detail.

Did the authors analyze or take into account the difference between receptive field locations? For example, did afferents more on the sides have lower responses and a lesser effect of history?

An investigation into the potential impact of the relationship between the receptive field location on the fingertip skin and the primary contact site of the stimulus surface revealed no discernible influence for SA-1 and SA-2 neurons. In contrast, FA-1 neurons, particularly those predominantly sensitive to the previous stimulation or displaying mixed sensitivity, exhibited a tendency to terminate near the primary stimulation site. We have added these observations to the text:

“We found no straightforward relationship between a neuron's sensitivity to current and previous stimulation and its termination site in fingertip skin. Specifically, there was no statistically significant effect of the distance between a neuron's receptive field center and the primary contact site of the stimulus surface on whether neurons signaled current, prior, or mixed information for SA-1 (Kruskal-Wallis test H(2)=3.86, p= 0.15) or SA-2 neurons (H(2)=0.75, p=0.69). However, a significant difference emerged for FA-1 neurons (H(2)=8.66, p=0.01), indicating that neurons terminating closer to the stimulation site on the flat part of the fingertip were more likely to signal past or mixed information.”

Was there anything different in the firing patterns between the spontaneous and non-spontaneously active SA-2s? For example, did the non-spontaneous show more dynamic responses?

The firing patterns of both spontaneously and non-spontaneously active SA-2 neurons shared similarities in terms of adaptation and range of firing rate modulation in response to force stimuli, i.e., ‘dynamic response’. The distinction lay in the pattern of modulation of the firing rate associated with stimulus presentations. For spontaneously active SA-2 neurons, this modulation occurred around a significant background discharge, implying that a force stimulus could either decrease or increase the firing rate, depending on how it deformed the fingertip. This characteristic is well illustrated by the firing pattern of the neuron depicted in the lower panels of Figure 3D. Conversely, in non-spontaneously active SA-2 neurons, a force stimulus could only induce an increase in the firing rate or no change. Although the neuron depicted in the upper panels of Figure 3D exhibited some background activity, it serves to exemplify this characteristic. In the text, we have elucidated the dynamics of the SA-2 neuron response by highlighting that force stimulation can either decrease or increase the firing rate in neurons with spontaneous activity through the following addition/change:

“This increased variability was most evident during the force protraction phase where most neurons exhibited the most intense responses. Increased variability was also observed in instances where the dynamic response to force stimulation involved a decrease in the firing rate (lower panels of Figure 3D). This phenomenon was observed in SA-2 neurons that maintained an ongoing discharge during intertrial periods (cf. Fig. 2A). In these cases, the response to a force stimulus constituted a modulation of the firing rate around the background discharge, signifying that a force stimulus could either decrease or increase the firing rate depending on the prevailing stimulus direction.”

Were the spontaneously active SA-2 afferents firing all the time or did they have periods of rest - and did this relate to recent stimulation? Were the spontaneously active SA-2s located in a certain part of the finger (e.g. nail) or were they randomly spread throughout the fingertip? Any distribution differences could indicate a more complicated role in skin sensing.

SA-2 neurons, in general, are well-known for undergoing significant post-stimulation depression (e.g., Knibestöl and Vallbo, 1970; Chambers et al., 1972; Burgess and Perl, 1973). In our force stimulations, this post-excitatory depression manifested as a reduced or absent response during the latter part of the stimulus retraction period for stimuli in directions that markedly excited the neuron. The excitability recovered when the fingertip relaxed during the subsequent intertrial period, and for "spontaneously active" neurons, the firing resumed (see examples in Figure 7A). Furthermore, some “spontaneously active” neurons could be silenced or exhibit a near-silent period during force stimulation for certain force directions, while the spontaneous firing returned during the upcoming intertrial period when the fingertip shape recovered (for example, see responses to stimulation in the proximal and especially ulnar directions in the top panel in Figure 7A).

Regarding the location of the receptive field centres of spontaneously active and non-spontaneously active SA-2 neurons on the fingertip we did not observe any obvious spatial segregation. To illustrate this, we have revised Figure 1A by color-marking SA-2 neurons that exhibited ongoing activity in intertrial periods, and the figure caption has been modified accordingly:

“Figure 1. Experimental setup. A. Receptive field center locations shown on a standardized fingertip for all first-order tactile neurons included in the study, categorized by neuron type. Purple symbols denote spontaneously active SA-2 neurons exhibiting ongoing activity without external stimulation.”

Did the authors look to see if the spontaneous firing in SA-2s between trials could predict the extent to which the type 1 afferents encode the proceeding stimulus? Basically, does the SA-2 state relate to how the type 1 units fire?

We found no clear indications that the responses of FA-1 and SA-1 could be readily anticipated based on the firing patterns of SA-2 neurons.

In the discussion, it is stated that "the viscoelastic memory of the preceding loading would have modulated the pattern of strain changes in the fingertip differently depending on where their receptor organs are situated in the fingertip". Can the authors expand on this or make any predictions about the size of the memory effect and the distance from the point of stimulation?

We have explored this topic further in the text, referring to recent studies modeling essential aspects of fingertip mechanics. However, in our view, current models lack the capability to predict the specific nature sought by the reviewer. These models should include a detailed understanding of the intricate networks of collagen fibers anchoring the pulp tissue at the distal phalangeal bone and the nail. They should also consider potential inherent directional preferences of the receptor organs, attributed to their microanatomy. The text modifications are as follows:

“In addition to the receptor organ locations, the variation in sensitivity among neurons to fingertip deformations in response to both previous and current loadings would stem from the fingertip’s geometry and its complex composite material properties. Possible inherent directional preferences of the receptor organs, attributed to their microanatomy, could also be significant. However, mechanical anisotropy, particularly within the viscoelastic subcutaneous tissue of the fingertip induced by intricately oriented collagen fiber strands forming fat columns in the pulp (Hauck et al., 2004), are likely to play a crucial role. This anisotropy would shape the dynamic pattern of strain changes at neurons' receptor sites, intricately influencing a neuron's sensitivity not only to current but also to preceding loadings. Indeed, recent modeling efforts suggest that such mechanical anisotropy strongly influences the spatiotemporal distribution of stresses and strains across the fingertip (Duprez et al., 2024).”

Relatedly, we have included additional text to provide a more comprehensive explanation of the “bulk deformation” of the fingertip that occurs during the loadings:

“As pressure increases in the pulp, the pulp tissue bulges at the end and sides of the fingertip. Simultaneously, the tangential force component amplifies the bulging in the direction of the force while stretching the skin on the opposite side.”

In the discussion, it would be good if the authors could briefly comment more on the diversity of the mechanoreceptive afferent firing and why this may be useful to the system.

The diversity in responses among neurons is instrumental in enhancing the information transmitted to the brain by averting redundancy in information acquisition. This diversity thereby contributes to an overall increase in information. We've included a brief statement, along with several references, underscoring this concept:

"The resulting diversity in the sensitivities of neurons might enhance the overall information collected and relayed to the brain by the neuronal population, facilitating the discrimination between tactile stimuli or mechanical states of the fingertip (see Rongala et al., 2024; Corniani et al., 2022; Tummala et al., 2023, for more extensive explorations of this idea)."

Also, the authors could briefly discuss why this memory (or recency) effect occurs - is it useful, does it serve a purpose, or it is just a by-product of our skin structure? There are examples of memory in the other senses where comparisons could be drawn. Is it like stimulus adaptation effects in the other senses (e.g. aftereffects of visual motion)?

We have expanded the concluding paragraph of the discussion, specifically delving into the question of whether the mechanical memory effect serves a deliberate purpose or is simply an incidental byproduct of our skin structure:

“In any case, the viscoelastic deformability of the fingertips plays a pivotal role in supporting the diverse functions of the fingers. For example, it allows for cushioned contact with objects featuring hard surfaces and allows the skin to conform to object shapes, enabling the extraction of tactile information about objects' 3D shapes and fine surface properties. Moreover, deformability is essential for the effective grasping and manipulation of objects. This is achieved, among other benefits, by expanding the contact surface, thereby reducing local pressure on the skin under stronger forces and enabling tactile signaling of friction conditions within the contact surface for control of grasp stability. Throughout, continuous acquisition of information about various aspects of the current state of the fingertip and its skin by tactile neurons is essential for the functional interaction between the brain and the fingers. In light of this, the viscoelastic memory effect on tactile signaling of fingertip forces can be perceived as a by-product of an overall optimization process within prevailing biological constraints.”

One point that would be nice to add to the discussion is the implications of the work for skin sensing. What would you predict for the time constant of relaxation of fingertip skin, how long could these skin memory effects last? Two main points to address here may be how the hydration of the skin and anatomical skin changes related to aging affect the results. If the skin is less viscoelastic, what would be the implications for the firing of mechanoreceptors?

It is likely that the time constant depends to some extent on mechanical factors of the skin, which will likely change due to age or environmental factors. However, while these questions are intriguing, they fall outside the scope of the current study and we are not aware of studies that have addressed these issues directly in experiments either.

How long does it take for the effect to end? Again, this will likely depend on the skin's viscoelasticity. However, could the authors use it in a psychophysical paradigm to predict whether participants would be more or less sensitive to future stimuli? In this way, it would be possible to test whether the direction modifies touch perception.

Time constants for tissue viscoelasticity have been estimated to extend up to several seconds (see citations in the introduction). While direct perceptual effects could indeed be explored through psychophysical experimental paradigms, we are currently unaware of any studies specifically addressing the type of effect described in this study. In addition to the statement that, concerning manipulation and haptic tasks, "to our knowledge, a possible influence of fingertip viscoelasticity on task performance has not been systematically investigated," we have now also addressed tactile psychophysical tasks conducted during passive touch with the following sentence in the text:

“Similarly, there is a lack of systematic investigation of potential effects of fingertip viscoelasticity on performance in tactile psychophysical tasks conducted during passive touch.”

Reviewer #2 (Public Review):

Summary:

The authors sought to identify the impact skin viscoelasticity has on neural signalling of contact forces that are representative of those experienced during normal tactile behaviour. The evidence presented in the analyses indicates there is a clear effect of viscoelasticity on the imposed skin movements from a force-controlled stimulus. Both skin mechanics and evoked afferent firing were affected based on prior stimulation, which has not previously been thoroughly explored. This study outlines that viscoelastic effects have an important impact on encoding in the tactile system, which should be considered in the design and interpretation of future studies. Viscoelasticity was shown to affect the mechanical skin deflections and stresses/strains imposed by previous and current interaction force, and also the resultant neuronal signalling. The result of this was an impaired coding of contact forces based on previous stimulation. The authors may be able to strengthen their findings, by using the existing data to further explore the link between skin mechanics and neural signalling, giving a clearer picture than demonstrating shared variability. This is not a critical addition, but I believe would strengthen the work and make it more generally applicable.

Strengths:

- Elegant design of the study. Direct measurements have been made from the tactile sensory neurons to give detailed information on touch encoding. Experiments have been well designed and the forces/displacements have been thoroughly controlled and measured to give accurate measurements of global skin mechanics during a set of controlled mechanical stimuli.

- Analytical techniques used. Analysis of fundamental information coding and information representation in the sensory afferents reveals dynamic coding properties to develop putative models of the neural representation of force. This advanced analysis method has been applied to a large dataset to study neural encoding of force, the temporal dynamics of this, and the variability in this.

Weaknesses:

- Lack of exploration of the variation in neural responses. Although there is a viscoelastic effect that produces variability in the stimulus effects based on prior stimulation, it is a shame that the variability in neural firing and force-induced skin displacements have been presented, and are similarly variable, but there has been no investigation of a link between the two. I believe with these data the authors can go beyond demonstrating shared variability. The force per se is clearly not faithfully represented in the neural signal, being masked by stimulation history, and it is of interest if the underlying resultant contact mechanics are.

Thank you for this suggestion. We have added a new section investigating the link between skin deformation and neural firing in more depth via a simple neural model. Please see our answer below in the ‘Recommendations’ section for further details.

Validity of conclusions:

The authors have succeeded in demonstrating skin viscoelasticity has an impact on skin contact mechanics with a given force and that this impacts the resultant neural coding of force. Their study has been well-designed and the results support their conclusions. The importance and scope of the work is adequately outlined for readers to interpret the results and significance.

Impact:

This study will have important implications for future studies performing tactile stimulation and evaluating tactile feedback during motor control tasks. In detailed studies of tactile function, it illustrates the necessity to measure skin contact dynamics to properly understand the effects of a force stimulus on the skin and mechanoreceptors.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

(Very) minor comments

- The authors say at the beginning of the Results that, "The fourth type of tactile neurons in the human glabrous skin, fast adapting type II neurons...". Although generally written that there are four types of afferent in the glabrous skin, it would be better to state that these are low-threshold A-beta myelinated mechanoreceptive afferents, at least one time, as there are other types of afferent in the glabrous skin that respond to mechanical stimulation (e.g. low and high threshold C-fibers).

This is now clarified at the start of the Results section:

“We recorded action potentials in the median nerve of individual low-threshold A-beta myelinated first-order human tactile neurons innervating the glabrous skin of the fingertip…”

- Fig. 3: Could you add '(N)' as the measurement of force for Fig. 3A for Fz, Fy, and Fz? Also, please change 'Data was recorded' to 'Data were recorded' in the legend.

Fixed.

- At the beginning of the Methods, you say that your study conforms to the Declaration of Helsinki, which actually requires pre-registration in a database. If you did not pre-register your study, please can you add '... in accordance with the Declaration of Helsinki, apart from pre-registration in a database'.

Thanks for making us aware of this. We have added the suggested qualifier to the ethics statement.

Reviewer #2 (Recommendations For The Authors):

The neural representation/encoding of the actual displacement vectors would be a useful addition to the analyses. These vectors have been demonstrated to systematically change with the condition in the irregular series (Figure 2E) and will thus significantly act on the dynamics of induced mechanical changes in the skin with a given interaction force. Thus, it could be examined how the neurons code the magnitude of displacements as well as their direction. An evaluation of the extent to which the imposed displacement magnitudes are encoded in the neural responses would be a useful addition in explaining the signalling of the force events and how the central nervous system decodes these. Evaluating an alternative displacement encoding for comparison to pure force encoding may reveal more about how contact events are represented in the tactile system, which must decode these variable afferent signals to reconstruct a percept of the interaction. It could then be explored how the central nervous system may then scale the dynamic afferent responses based on the background viscoelastic state likely to be present in the SA-II afferent signals (Figure 7) for a context in which to evaluate the dynamic contact forces. This may of course be a complex relationship for the type-I afferents, where the underlying mechanical events evoking the firing (microslips not represented in global forces) have not been measured here. Such a model could be more widely applicable, as the skin viscoelasticity and displacement magnitudes are a straightforward measurement metric and could perhaps be used as a better proxy for neural signalling. This would allow the investigation of a wider variety of forces, and the study of the timing of the viscoelastic effect, both of which have been fixed here. This would give the work a broader impact, rather than just highlighting that this effect produces variability, it could reveal if this mechanical feature is structured in the neural representation. The categorical encoding/decoding tested here is specific to the stimuli used (magnitudes, intervals), but there is the possibility that this may be more generally applicable (within the bounds of forces/speeds) if the underlying basis of the variability in the signalling produced by the viscoelasticity is identified. Since the time course of the viscoelasticity has not been measured here (fixed forces and intervals), further study is required to fully understand the implications this has for a wider variety of situations.

We agree that a better understanding of how the mechanical deformations are reflected in the resulting spike trains would be valuable. While ultimately a full understanding will need precise measurements of skin deformation across the whole fingertip to account for mechanical propagation to mechanoreceptor locations, relating the deformations at the contact location with neural firing patterns directly can provide useful hints into which aspects of deformation are encoded and how. To this end, we ran a new analysis that aimed to predict the time-varying neural responses directly from the recorded mechanical movements of the contactor.

Below we have reproduced the new results and methods text along with the additional figures for this analysis. Note that we have also added text in the Discussion to interpret these findings in the context of our other results.

New section in Results titled Predicting neural responses from contactor movements: “The similarity in the history-dependent variation in neural firing and fingertip deformation at a given force stimulus suggests that neuronal firing is determined by how the fingertip deforms rather than the applied force itself. However, this similarity does not clarify the relationship between fingertip deformation dynamics and neural signaling. To investigate further, we fit cross-validated multiple linear regression models to evaluate how well distinct aspects of contactor movement could predict the time-varying firing rates of individual neurons during the protraction phases of the irregular sequence. The models used predictors based on (1) the three-dimensional position of the contactor, (2) its three-dimensional velocity, (3) a combination of position and velocity signals, and, finally, (4) position and velocity signals along with all possible two-way interactions between them, capturing potentially complex relationship between fingertip deformations and neural signaling.

Comparing the variance explained (R²) by each regression model for each neuron type revealed clear differences between the models (Figure 5A). A two-way mixed design ANOVA, with regression model as within-group effects and neuron type as a between-group effect revealed a main effect of model on variance explained (F(3,462) = 815.5, p < 0.001, η_p² = 0.84). Model prediction accuracy overall increased with the number of predictors, with the two-way interaction model outperforming all others (p < 0.001 for all comparisons, Tukey’s HSD). Additionally, a significant main effect of neuron type (F(2,154) = 29.8, p < 0.001, η_p² = 0.28) and a significant interaction between regression model and neuron type were observed (F(6,462) = 50.8, p < 0.001, η_p² = 0.40).

For neuron type, model predictions were most accurate for SA-2 neurons, followed by SA-1 neurons, with FA-1 neurons showing the lowest accuracy (p < 0.003 for all comparisons, Tukey’s HSD). The interaction between model and neuron type revealed distinct patterns. For SA-1 and SA-2 neurons, position-only and velocity-only models had similar prediction accuracy (p ≥ 0.996, Tukey’s HSD) with no significant differences between these neuron types (p ≥ 0.552, Tukey’s HSD). FA-1 neurons performed poorly with the position-only model but showed higher accuracy with the velocity-only model (p < 0.001, Tukey’s HSD) and better than SA-1 neurons (p = 0.006, Tukey’s HSD). Models combining position and velocity predictors (without interactions) surpassed both position-only and velocity-only models for SA-1 and SA-2 neurons (p < 0.001, Tukey’s HSD). Overall, the differences between neuron types broadly match their tuning to static and dynamic stimulus properties.

The two-way interaction model, accounting for most variance in neural responses, produced mean R² values of 0.75 for FA-1, 0.88 for SA-1, and 0.91 for SA-2 neurons (Figure 5A). To evaluate the contribution of the different predictors, we ranked them using the permutation feature importance method, focusing on the six most important ones. Regression analyses using only these variables explained almost all of the variance explained by the full model, with a median R² reduction of just 0.055 across all neurons. Across all neuron types, at least half included all three velocity components (dPx, dPy, dPz) among the top six, with FA-1 neurons showing the highest prevalence (Figure 5B). Interactions between normal position (Pz) and each velocity component were also frequently observed, while interactions involving tangential position and velocity components were less common. Interactions among velocity components were relatively well represented, followed by interactions limited to position components. Position signals were generally less represented, except for normal position (Pz) in slowly adapting neurons, where it appeared in 50% of SA-1 and 68% of SA-2 neurons. Despite these broad trends, important predictors varied widely across ranks even within a given neuron class (see Figure 5-figure supplement 1), and even the most frequent variables appeared in only a subset of cases, suggesting broad variability in sensitivity across neurons.”

New methods paragraph titled Predicting time-varying firing rates from skin deformations:

“This analysis was conducted in Python (v3.13) with pandas for data handling, numpy for numerical operations, and scikit-learn for model fitting and evaluation.

To assess how well individual neurons' time-varying firing rates could be predicted from simultaneous contactor movements, we fitted multiple linear regression models (see Khamis et al., 2015, for a similar approach}. This analysis focused on the force protraction phase of the irregular sequence, where neurons were most responsive and sensitive to stimulation history. Data from 100 ms before to 100 ms after the protraction phase (between -0.100 s and 0.225 s relative to protraction onset) were included for each trial. Neurons were included if they fired at least two action potentials during the force protraction phase and the following 100 ms in at least five of the 25 trials. This ensured sufficient variability in firing rates for meaningful regression analysis, resulting in 68 SA-1, 38 SA-2, and 51 FA-1 neurons being included.

Contractor position signals digitized at 400 Hz were linearly interpolated to 1000 Hz. Instantaneous firing rates, derived from action potentials sampled at 12.8 kHz, were resampled at 1000 Hz to align with position signals. A Gaussian filter (σ = 10 ms, cutoff ~16 Hz) was applied to the firing rate as well as to the position signals before differentiation. To account for axonal conduction (8–15 ms) and sensory transduction delays (1–5 ms), firing rates were advanced by 15 ms to align approximately with independent variables.

Regressions were performed using scikit-learn's Ridge and RidgeCV regressors, which apply L2 regularization to mitigate overfitting. Hyperparameter tuning for the regularization parameter (alpha) was performed using GridSearchCV with a predefined range (0.001–1000.0), incorporating five-fold cross-validation to select the best value. To minimize overfitting risks, model performance was further validated with independent five-fold cross-validation (KFold), and R² scores were computed using cross_val_score.

We constructed four linear regression models with increasing complexity: (1) Position-only, using three-dimensional contactor positions (Px, Py, Pz); (2) Velocity-only, using three-dimensional velocities (dPx, dPy, dPz); (3) Combined, including all position and velocity signals (6 predictors); and (4) Interaction, including all signals and their two-way interactions (21 predictors). All features were standardized using StandardScaler to improve regularization and model convergence. PolynomialFeatures generated second-order interaction terms for the interaction model. Feature importance was evaluated with permutation_importance, and simpler models were built using the most important features. These models were validated through cross-validation to assess retained explanatory power.”

Minor:

- It would be useful to add a brief description of the material aspects of the contactor tip to the methods (as per Birznieks 2001).

We have added the following statement:

“To ensure that friction between the contactor and the skin was sufficiently high to prevent slips, the surface was coated with silicon carbide grains (50–100 μm), approximating the finish of smooth sandpaper.”

- The axes labelling on Figure 3A and legend description is ambiguous, probably placing the Px, Py, and Pz labels on the far left axes and the Fx, Fy, and Fz on the right side of the far right axes would make this clearer.

Label placement has been improved along with some other minor fixes.

- For the quasi-static phase analysis, the phrase "absence of loading" used in reference to the interstimulus period and SA-II afferents does not seem to be a correct description. The finger is still loaded (at least in the normal direction), with a magnitude of imposed displacement that counteracts the viscoelastic force exerted by the skin mechanics of the fingertip. Although there is a zero net-force load, a mechanical stimulus is still being actively applied to the skin.

We have changed the wording throughout the text and now consistently refer either to the “interstimulus period” directly or to an “absence of externally applied stimulation” to avoid confusion.

Author response:

The following is the authors’ response to the previous reviews

Reviewer #2 (Public review):

Summary:

The revised paper by Kim et al. reports two disease mutations in proBMP4, S91C and E93G, disrupt the FAM20C phosphorylation site at Ser91, blocking the activation of proBMP4 homodimers, while still allowing BMP4/7 heterodimers to function. Analysis of DMZ explants from Xenopus embryos expressing the proBMP4 S91C or E93G mutants showed reduced expression of pSmad1 and tbxt1. The expert amphibian tissue transplant studies were expanded to in vivo studies in Bmp4S91C/+ and Bmp4E93G/+ mice, highlighting the impact of these mutations on embryonic development, particularly in female mice, consistent with patient studies. Additionally, studies in mouse embryonic fibroblasts (MEFs) demonstrated that the mutations did not affect proBMP4 glycosylation or ER-to-Golgi transport but appeared to inhibit the furin-dependent cleavage of proBMP4 to BMP4. Based on these findings and AI modeling using AlphaFold of proBMP4, the authors speculate that pSer91 influences access of furin to its cleavage site at Arg289AlaLysArg292 in a new "Ideas and Speculation" section. Overall, the authors addressed the reviewers' comments, improving the presentation.

Strengths:

The strengths of this work continue to lie in the elegant Xenopus and mouse studies that elucidate the impact of the S91C and E93G disease mutations on BMP signaling and embryonic development. Including an "Ideas and Speculation" subsection for mechanistic ideas reduces some shortcomings regarding the analysis of the underlying mechanisms.

Weaknesses:

(1)  (Minor) In Figure S1 and lines 165-174 and 179-180, the authors should consider that, unlike the wild-type protein (Ser), which can be reversibly phosphorylated or dephosphorylated, phosphomimic mutations are locked into mimicking either the phosphorylated state (Asp) or the non-phosphorylated state (Ala). Consequently, if the S91D mutant exhibits lower activity than WT, it could imply that S91D interferes with other regulatory constraints, as the authors suggest. However, it may also be inhibiting activation. Therefore, caution is warranted when comparing S91D with S91C to conclude that Ser91 phosphorylation increases BMP4 activity. While additional experiments are not necessary, further consideration is essential.

(Minor) In lines 394-399, the authors cleverly speculate that pS91 interacts with Arg289-the essential P4 arginine for furin processing. If so, this interaction could hinder the cleavage of proBMP4, as indicated by the results in Figure S1. The discussion would benefit from considering that, contrary to their favored model, dephosphorylation at Ser91 might actually facilitate cleavage.

We have added a paragraph raising this possibility but explaining why it is unlikely and inconsistent with our in vivo data. The S91D construct was a simple control that was tested in ectopic expression assays and not in vivo.  We can make no conclusions about whether this construct resembles the phosphorylated state or whether it hinders or facilitates cleavage in vivo. The conclusion that dephosphorylation promotes BMP4 cleavage or activity is not compatible with the finding that two mutations associated with birth defects in humans (p.S91C or p.E93G) that are predicted to prevent FAM20C-mediated phosphorylation of the BMP4 prodomain lead to impaired proteolytic maturation of endogenous BMP4 and reduced BMP activity in vivo. 

(2)  In Figure 4, panels A, E, and I, the proBMP bands in the mouse embryonic lysates and MEFs expressing the mutations show a clear size shift. Are these shifts a cause or a consequence of the lack of cleavage? Regardless, the size shifts should be explicitly noted.

These intriguing shifts were observed in some but not all biological replicates.  When present, the shifts were not reversed by treatment with phosphatases or deglycosylases, and the shifts were never observed in epitope tagged wild type controls.  We have added a paragraph noting the shifts and our tests of whether they might be due to glycosylation, phosphorylation or epitope tags. 

(3)  (Minor) In line 314, the authors should consider modifying the wording to: "is required for modulating proprotein convertase..."

The original wording (“Collectively, our findings are consistent with a model in which FAM20C-mediated phosphorylation of the BMP4 prodomain is not required for folding or exit of the precursor protein from the ER, but is required for proprotein convertase recognition and/or for trafficking to post-TGN compartment(s) where BMP4 is cleaved”) more accurately reflects the model that is supported by our findings. Stating that “phosphorylation ……is required to modulate proprotein convertase recognition and/or trafficking” is vague and leaves open the possibility that it modulates in either direction, which our data do not support as described in point 1 above.

Author response:

The following is the authors’ response to the original reviews

Public Reviews: 

Reviewer #1 (Public Review): 

This study investigates the role of microtubules in regulating insulin secretion from pancreatic islet beta cells. This is of great importance considering that controlled secretion of insulin is essential to prevent diabetes. Previously, it has been shown that KIF5B plays an essential role in insulin secretion by transporting insulin granules to the plasma membrane. High glucose activates KIF5B to increase insulin secretion resulting in the cellular uptake of glucose. In order to prevent hypoglycemia, insulin secretion needs to be tightly controlled. Notably, it is known that KIF5B plays a role in microtubule sliding. This is important, as the authors described previously that beta cells establish a peripheral sub-membrane microtubule array, which is critical for the withdrawal of excessive insulin granules from the secretion sites. At high glucose, the sub-membrane microtubule array is destabilized to allow for robust insulin secretion. Here the authors aim to answer the question of how the peripheral array is formed. Based on the previously published data the authors hypothesize that KIF5B organizes the sub-membrane microtubule array via microtubule sliding. 

General comment: 

This manuscript provides data that indicate that KIF5B, like in many other cells, mediates microtubule sliding in beta cells. This study is limited to in vitro assays and one cell line. Furthermore, the authors provide no link to insulin secretion and glucose uptake and the overall effects described are moderate. Finally, the overall effect of microtubule sliding upon glucose stimulation is surprisingly low considering the tight regulation of insulin secretion. Moreover, the authors state "the amount of MT polymer on every glucose stimulation changes only slightly, often undetectable…. In fact, we observe a prominent effect of peripheral MT loss only after a long-term kinesin depletion (three-four days)". This challenges the view that a KIF5Bdependent mechanism regulating microtubule sliding plays a major role in controlling insulin secretion. 

(1) Our initial study was indeed done in a cell line, which is a normal approach to addressing molecular mechanisms of a phenomenon in a challenging cell model: primary pancreatic beta cells are prone to rapidly dedifferentiate outside of the organism and are hard to genetically modify. To address this reviewer’s comment, in the revised manuscript we now confirm the phenotype in beta cells within intact pancreatic islets from a KIF5B KO mouse model (New Figure 2 – Supplemental Figure 1).

(2) We agree that testing the effect of microtubule sliding on insulin secretion is an important question. Unfortunately, the experimental design needed to accomplish this task is not straighDorward. Importantly, besides microtubule sliding, KIF5B is heavily engaged in insulin granule transport, and GSIS deficiency upon KIF5B inactivation is well documented (e.g. Varadi et al 2002). In this study, we choose not to repeat this GSIS assay because of ample existing data. However, this reported GSIS deficiency could result from a combination of lack of insulin granule delivery to the periphery (previous data) and from the depletion of insulin granules from the periphery due to the loss of the submembrane MT bundle (this study and Bracey et al 2020).  In order to exclusively test the role of MT sliding in secretion, a significant investment in mutant tool development would be needed. Ideally, a new mutant mouse model where insulin granule transport is allowed by MT sliding in blocked must be developed to specifically address this question. To conclude, answering this question will be the subject for another, follow-up study. 

(3) We respecDully disagree with the reviewer’s opinion that the effect of MT sliding in beta cells is moderate. As MT networks go, even a slight change in MT configuration often has dramatic consequences. For example, in mitotic spindles, a tiny overgrowth of microtubule ends during metaphase, which causes them to attach to both kinetochores rather than just one, is very significant for the efficiency of chromosome segregation, causing aneuploidy and cancer. The changes in beta-cell MT networks that we are reporting are much stronger: the effect on the peripheral MT network accumulated over three days of KIF5B depletion is dramatic (Fig 2 B, C). Short-term gross MT network configurations after a single glucose stimulation are harder to detect, but MTs at the cell periphery are, in fact, destabilized and fragmented, as we and others have previously reported (Ho et al 2020, Mueller et al 2021). Preventing this MT rearrangement completely blocks GSIS (Zhu et al 2015, Ho et al 2020). 

One of the most fascinating features of insulin secretion regulation is that the amount of generated insulin granules significantly exceeds the normal physiological needs for insulin secretion (~100 times more than needed). At the same time, even slightly facilitated glucose depletion can be devastating. Accordingly, the excessive insulin content of a beta cell resulted in the development of multiple levels of control, preventing excessive secretion. Our previous data suggest that the peripheral MT array provides one of those mechanisms. This study indicates that microtubule sliding is necessary to form the proper peripheral network in the long term. Short-term glucose-induced changes in the peripheral MT array likely need to be subtle to prevent over-secretion. Thus, we are not surprised that a dramatic effect of sliding inhibition is only detectable by our approaches after the changes in the MT network accumulate over time. In the revised paper, we now discuss the potential impact of peripheral MT sliding on positive and negative regulation of secretion and add a schematic model illustrating these processes.

Specific comments: 

(1) Notably, the authors have previously reported that high glucose-induced remodeling of microtubule networks facilitates robust glucose-stimulated insulin secretion. This remodeling involves the disassembly of old microtubules and the nucleation of new microtubules. Using real-time imaging of photoconverted microtubules, they report that high levels of glucose induce rapid microtubule disassembly preferentially in the periphery of individual β-cells, and this process is mediated by the phosphorylation of microtubule-associated protein tau. Here, they state that the sub-membrane microtubule array is destabilized via microtubule sliding. What is the relevance of the different processes? 

In this comment, the summary of our previous conclusions is correct, but the conclusion of this current study is re-stated incorrectly. Indeed, we have previously shown that in high glucose, MTs are destabilized at the cell periphery and nucleated in the cell interior. However, this current paper does not state that “the sub-membrane microtubule array is destabilized via microtubule sliding”. To answer this reviewer’s question, our data support a model where, during glucose stimulation, MT sliding within the peripheral bundle might move fragments of MTs severed by other mechanisms. Importantly, we propose that MT sliding restores the partially destabilized peripheral bundle by delivery of MTs that are nucleated at the cell interior and incorporating them into that bundle. In our overall model, three processes (destabilization, nucleation, and sliding to restore the bundle) are coordinated to maintain beta cell fitness on each GSIS cycle.

(2) On one hand the authors describe how KIF5B depletion prevents sliding and the transport of microtubules to the plasma membrane to form the sub-membrane microtubule array. This indicates KIF5B is required to form this structure. On the other hand, they describe that at high glucose concentration, KIF5B promotes microtubule sliding to destabilize the sub-membrane microtubule array to allow robust insulin secretion. This appears contradictory. 

We never intended to make an impression that MT sliding destabilized the sub-membrane bundle. Apologies if there was a reason in our wording that caused this misunderstanding of our model. We propose that while the bundle is destabilized downstream of glucose signaling (e.g. due to tau phosphorylation, please see Ho et al Diabetes 2020), MT sliding remodels the bundle and thereafter rebuilds it to prevent over-secretion. In the revised manuscript, we have doublechecked the whole text to make sure that such misunderstanding is avoided. 

(3) Previously, it has been shown that KIF5B induces tubulin incorporation along the microtubule shaft in a concentration-dependent manner. Moreover, running KIF5B increases microtubule rescue frequency and unlimited growth of microtubules. Notably, KIF5B regulates microtubule network mass and organization in cells (PMID: 34883065). Consequently, it appears possible that the here observed phenomena of changes in the microtubule network might be due to alterations in these processes. 

We thank the reviewer for proposing this alternative explanation to the observed change in microtubule networks after KIF5B depletion. We have now directly tested this possibility. Namely, we have re-expressed the kinesin-1 motor domain in MIN6 cells depleted of KIF5B. This motor domain construct by itself is not capable of driving microtubule sliding because it lacks the tail domain. At the same time, it is known to move very efficiently at microtubules and should provide the effects as reported in the article cited by the reviewer. We found that the reexpression of the kinesin motor domain does not rescue microtubule network defects in beta cells (see new Figure 2 – Supplemental Figure 2). Thus, we conclude that the effects of kinesin depletion on the microtubule network in beta cells are due to the lack of microtubule sliding, as reported here.

(4) The authors provide data that indicate that microtubule sliding is enhanced upon glucose stimulation. They conclude that these data indicate that microtubule sliding is an integral part of glucose-triggered microtubule remodeling. Yet, the authors fail to provide any evidence that this process plays a role in insulin secretion or glucose uptake. 

We would like to point out that we do not “fail” but rather choose not to overload our study by repeating insulin secretion assays in KIF5B-inactivated cells because this would not have been very informative. It has been found previously that kinesin-1 inactivation or knockout significantly attenuates insulin secretion because kinesin-1 is actively transporting insulin granules and kinesin-1 activity is enhanced under high glucose conditions (e.g. Varadi et al 2002, Cui et al., 2011, Donelan et al, 2002). That said, our current finding is very much in line with these previous data. When kinesin is depleted, two things would be happening at the same time: in the absence of sub-membrane microtubule bundle pre-existing insulin granules would be over-secreted, and new insulin would not be delivered to the periphery, both decreasing GSIS. Unfortunately, we do not have tools yet that would allow us to dissect which part of the insulin secretion defect is due to prior over-secretion (the consequence of deficient MT sliding) and which part is due to the lack of new granule delivery. We plan to develop such tools in the future and elaborate on them in a follow-up study. Here, our goal is to understand microtubule organization principles in beta cells, and we choose not to extend the scope of the current study to metabolic assays.  

(5) The authors speculate that the sub-membrane microtubule array prevents the over-secretion of insulin. Would one not expect in this case a change in the distribution of insulin granules at the plasma membrane when this array is affected? Or after glucose stimulation? Notably, it has been reported that "the defects of β-cell function in KIF5B mutant mice were not coupled with observable changes in islet morphology, islet cell composition, or β-cell size" and "the subcellular localization of insulin vesicles was found to not be affected significantly by the decreased Kif5b level. The cytoplasm of both wild-type and mutant β-cells was filled with insulin vesicles. Insulin vesicle numbers per square μm were determined by counting all insulin vesicles in randomly photographed β-cells. More insulin granules were found in Kif5b knockout β-cells compared with control cells. This phenomenon is consistent with the observation that insulin secretion by β-cells is affected" whereby "Insulin vesicles (arrowheads) were distributed evenly in both mutant and control cells" (PMID: 20870970).  

Quantitative analyses in the study cited by the reviewer do not include assays that would be relevant to our study. Particularly, in that study neither the amount of insulin granules at the cell periphery nor the ratio between the number of granules at the periphery and the beta cell interior has been analyzed. In addition, in our preliminary observations not shown here, insulin content in beta cells in KIF5B KO mice is highly heterogeneous, with a subpopulation of cells severely depleted of insulin. This opens a new avenue of investigation into beta cell heterogeneity, which is out of the scope of this current study. Thus, we chose to restrict this current study to microtubule organization data.   

(6) Does the sub-membrane microtubule array exist in primary beta cells (in vitro and/or in vivo) and how it is affected in KIF5B knockout mice?  

Yes, it does exist. In fact, we have first reported it in mouse islets (Bracey et al 2020, Ho et al 2020). Now, we report that the sub-membrane bundle is defective, and microtubules are misaligned in KIF5B KO mice (new Figure 2 – Supplemental Figure 1).

Reviewer #2 (Public Review): 

In this article, Bracey et al. provide insights into the factors contributing to the distinct arrangement observed in sub-membrane microtubules (MTs) within mouse β-cells of the pancreas. Specifically, they propose that in clonal mouse pancreatic β-cells (MIN6), the motor protein KIF5B plays a role in sliding existing MTs towards the cell periphery and aligning them with each other along the plasma membrane. Furthermore, similar to other physiological features of β-cells, this process of MTs sliding is enhanced by a high glucose stimulus. Because a precise alignment of MTs beneath the cell membrane in β-cells is crucial for the regulated secretion of pancreatic enzymes and hormones, KIF5B assumes a significant role in pancreatic activity, both in healthy conditions and during diseases. 

The authors provide evidence in support of their model by demonstrating that the levels of KIF5B mRNA in MIN6 cells are higher compared to other known KIFs. They further show that when KIF5B is genetically silenced using two different shRNAs, the MT sliding becomes less efficient. Additionally, silencing of KIF5A in the same cells leads to a general reorganization of MTs throughout the cell. Specifically, while control cells exhibit a convoluted and non-radial arrangement of MTs near the cell membrane, KIF5B-depleted cells display a sparse and less dense sub-membrane array of MTs. Based on these findings, the Authors conclude that the loss of KIF5B strongly affects the localization of MTs to the periphery of the cell. Using a dominant-negative approach, the authors also demonstrate that KIF5B facilitates the sliding of MTs by binding to cargo MTs through the kinesin-1 tail binding domain. Additionally, they present evidence suggesting that KIF5B-mediated MT sliding is dependent on glucose, similar to the activity levels of kinesin-1, which increase in the presence of glucose. Notably, when the glucose concentrations in the culturing media of MIN6 cells are reduced from 20 mM to 5 mM, a significant decrease in MT sliding is observed. 

Strengths:

This study unveils a previously unexplained mechanism that regulates the specific rearrangement of MTs beneath the cell membrane in pancreatic β-cells. The findings of this research have implications and are of significant interest because the precise regulation of the MT array at the secretion zone plays a critical role in controlling pancreatic function in both healthy and diseased states. In general, the author's conclusions are substantiated by the provided data, and the study demonstrates the utilization of state-of-the-art methodologies including quantification techniques, and elegant dominant-negative experiments. 

Weaknesses:

A few relatively minor issues are present and related to data interpretation and the conclusions drawn in the study. Namely, some inconsistencies between what appears to be the overall and sub-membrane MT array in scramble vs. KIF5B-depleted cells, the lack of details about the sub-cellular localization of KIF5B in these cells and the physiological significance of the effect of glucose levels in beta-cells of the pancreas. 

We thank the reviewer for this insighDul review. In the revised version, we provided re-worded and extended interpretations and conclusions to prevent any issues or misunderstandings.  We trust that while some noted apparent inconsistencies may reflect the intrinsic heterogeneity of the beta cell population, all data presented here indicate the same trend in phenotypes.  In the revised version, we have provided additional cell views and, in places, alternative representative images and videos, to clear out any apparent inconsistencies. We also would like to point out that we in fact reported KIF5B localization: not surprisingly, KIF5B predominantly localized to insulin granules and the punctate staining fills the whole cytoplasm (Figure 2A, bottom panel). However, as pointed out in detail in our response to reviewer 1, we choose to leave out an extensive study of the physiological and metabolic consequences of the reported microtubule network dynamics to a follow-up study. 

Reviewer #3 (Public Review): 

Prior work from the Kaverina lab and others had determined that beta-cells build a microtubule network that differs from the canonical radial organization typical in most mammalian cell types and that this organization facilitates the regulated secretion of insulin-containing secretory granules (IGs). In this manuscript, the authors tested the hypothesis that kinesin-driven microtubule sliding is an underlying mechanism that establishes a sub-membranous microtubule array that regulates IG secretion. They employed knock-down and dominant-negative strategies to convincingly show microtubule sliding does, in fact, drive the assembly of the sub-membranous microtubule band. They also used live cell imaging assays to demonstrate that kinesin-mediated microtubule sliding in beta-cells is triggered by extracellular high glucose. Overall, this is an interesting and important study that relates microtubule dynamics to an important physiological process. The experiments were rigorous and well-controlled. 

We truly appreciate this reviewer’ opinion. 

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors): 

Figures: 

(1) Figure 1: 

a) Why can one not see here, and in most following images, the peripheral sub-membrane microtubule array? One can also not see an accumulation of microtubules in the cell interior. 

Microtubule pattern in beta cells is variable, and the sub-membrane array is seen in the whole population to a variable extent (see directionality histogram in Figure 2E for statistics). In fact, an array of peripheral MTs parallel to the cell border is present in the example shown in Figure 1 and in all following control images. To make it clearer, we now show the pre-bleach images in Figure 1 D-F at a lower magnification, so that the differences in MT density at the cell periphery and cell center are more clearly seen: MTs lack at the periphery in KF5B-depleted but not the control cells.  

b) 5 min appears to be a long time and enough time to polymerize a significant number of new microtubules. 

We interpret this comment as the reviewer’s concern that in FRAP assays, fluorescently-labeled MTs moving into the bleached area might be newly polymerizing MTs rather than preexisting MT relocated into that area. However, this is not the case because newly polymerized MTs contain predominantly quenched “dark” tubulin molecules and only a small percent of fluorescent tubulin. These dim MTs are not included in MT sliding assay analysis, where a threshold for bright MTs is introduced. Now, we added more details for the quantification of these data to Materials and Methods section.

c) The overall effects appear minor. It is unclear how Fig. 1-Suppl-Fig.1, where no significant difference is shown, is translated into Figure 1 J and K showing a significant difference. 

With all due respect, we do not agree that the effect is minor. Please see our response to the Public Review where we discuss the major consequences of MT defects in detail. 

To answer this specific comment, we show that there are significant differences in the number of rapidly moving MTs (5-sec displacement over 0.3 µm) and in the amount of stationary MTs (5sec displacement is below 0.15 µm). There is no significant difference in the amount of slightly displaced MTs (displacements between 0.15 and 0.3 µm; the central part of the histogram). This might indicate that these slight displacements do not depend on kinesin-1 motor but rather are caused by experimental noise, pushing by moving organelles, and/or myosin-dependent forces in the cell. In the revised manuscript, we have this quantification more clearly detailed in Methods and included in Figure legends.

d) The authors utilize single molecule tracking to further strengthen their conclusion that KIF5B promotes microtubule sliding. The observed effects are weaker than the data obtained from photobleaching experiments. The videos clearly show that there is still significant movement also in KIF5B-depleted cells. If K560RigorE236A binds irreversibly to a microtubule and this microtubule is growing (not only by the addition of tubulin dimers to the plus end; see PMID: 34883065) wouldn't that also result in movement of the tagged K560RigorE236A? As KIF5B is also required in the transport of insulin granules, it should also label "interior microtubules". And in Video 2 it appears that pretty much all "labeled" microtubules are moving. 

K560RigorE236A forms fiducial marks along the whole MTs lattice, as previously shown in (Tanenbaum et al., 2014). When it is bound to MT lattice, K560RigorE236A moves with the whole MT if it is being relocated. The mechanism described in (PMID: 34883065) appears to be absent or minor in beta cells (see Figure 2- Supplemental Figure 2), thus, even if this mechanism would displace already polymerized MTs, this is not happening in this cell type.

The reviewer is correct, K560RigorE236A does mark all MTs throughout a beta cell. All MTs are moving slightly in a living cell because they are pushed around by moving organelles, actin contractility, etc. MTs may also be slid by other MT-dependent motors (dynein against the membrane and such). So, it is not surprising that the MT network is “breezing,” and kinesindependent sliding is only a part of MT movement. What we show here is that the KIF5Bdependent MT sliding is responsible for a relatively “long-distance” relocation of MTs manifested in long, directional displacement of fiducial marks.  This does not exclude other movements. This makes extraction of kinesin-dependent MT movements somewhat challenging, of course, that is why we needed to do those extensive analyses. 

e) Figure 1 G to K is misleading, at least in the context of the provided videos. There are several microtubules that move extensively in shRNA#2-treated cells and overall there appears more movement in this cell as in the control cell. Figure 1I is clearly not representative of the movement shown in Video 2. 

We apologize if our selection of representative movies/figures for this experiment was imperfect. Indeed, in all depleted cells, SunTag puncta still move to a certain extent, either due to incomplete depletion or to alternative intracellular forces dislocating microtubules. However, there is a clear difference in the fraction of persistently moving puncta (please see Figure 1K and  histogram in Figure 1 - Supplemental Figure 1B). Unfortunately, when the number of SunTag puncta per a cell is variable, it sometimes prevents a good visual perception of the actual distribution of moving versus stationary microtubules. We now show an alternative representative movie for the Figure 1I and the corresponding Video 2, with a goal to compare cells with more consistent numbers of Sun-Tag puncta.

(2) Figure 2A. 

a) This is the only image that clearly shows the existence of a sub-membrane microtubule array and the concentration of microtubules in the cell interior. The differences are unclear between the experimental setups including the length of cultivation and knockdown of KIF5B or expression of mutants. 

We now provide a more detailed description of each image acquisition and processing in Materials and Methods. In brief, while the morphology of MT patterns is intrinsically variable in beta cells, all control cells have populated peripheral MTs that exhibit a more parallel configuration as compared to depletions and mutants.

b) The authors state "While control cells had convoluted non-radial MTs with a prominent sub-membrane array, typical for beta cells (Fig. 2A), KIF5B-depleted cells featured extra-dense MTs in the cell center and sparse reseeding MTs at the periphery (Fig. 2B, C)". Could that not be explained with the observation that "Kinesin-1 controls microtubule length" (PMID: 34883065)? 

Thank you for this interesting alternative idea. It does not appear to be the case for beta cells.

Please see Figure 2-Supplemental Figure 2  and our response to Public Review Comment #3.

Also, our apologies for the typo in the original manuscript: this is “receding” nor “reseeding”.

(3) Figure 3: 

a) This is an elegant way to determine whether KIF5B is involved in microtubule sliding independent of the fact that the effect appears very small. 

Thank you!            

b) The assay depends on ectopic expression of a dominant negative mutant. It appears important to show that KIFDNwt is high enough expressed to indeed block the binding of endogenous KIF5B. The authors need to provide a control for this. Furthermore, authors need to provide evidence that other functions of KIF5B are not impaired such as transport of insulin granules and tubulin incorporation or microtubule stability and length.

Expression of cargo-binding motor domains routinely causes a dominant-negative effect of their cargo transport. This exact construct has been used for the purpose of dominant-negative action previously (Ravindran et al., 2017). It does prevent the membrane cargo binding of KIF5B (Ravindran et al., 2017), thus the transport of insulin granules is also impaired in overexpression cells. Confirming this fact would not influence our study conclusions, so we chose not to repeat these assays for the sake of time.

c) N-numbers should be similar. The data for KIFDNmut are difficult to interpret with possibly 2 experiments showing little to no displacement and 3 showing displacement. 

In the revised manuscript, additional data have been added to increase N-numbers.

(4) Figure 4 and supplements: The morphology of the KIFDNwt cells is greatly affected and this makes it difficult to say whether the effect on microtubules at the cell periphery is a direct or indirect effect. 

Yes, these cells often have less spread appearance, obscuring visual perception of MT distribution. We have now replaced the image of KIFDNwt cell (Figure 4, Supplemental Figure 1 A) to a more visually representative example.

Things to do: 

(1) Notably, the authors have previously reported that high glucose-induced remodeling of microtubule networks facilitates robust glucose-stimulated insulin secretion. This remodeling involves the disassembly of old microtubules and the nucleation of new microtubules. Here, they state that the sub-membrane microtubule array is destabilized via microtubule sliding. What is the relevance of the different processes? Please discuss these in the manuscript. 

Thank you, we have now extended our discussion of these points and our prior findings. We have also added a schematic model figure for clarity (Figure 7).  

(2) 5 min appears to be a long time and enough time to polymerize a significant number of new microtubules. Do the authors have any information about the speed of MT formation in MIN6 cells? Can the authors repeat this experiment by preventing MT polymerization? Or repeat the experiment with EB1/EB3 reporter to visualize microtubule growth in the same experimental setting? 

While some MT polymerization will happen in this timeframe, newly polymerized MTs contain predominantly quenched “dark” tubulin molecules and only a small percent of fluorescent tubulin. These dim MTs are not included in MT sliding assay analysis, where a threshold for bright MTs is introduced. We apologize for initially omitting certain details from the FRAP assay analysis. Now these details have been added.   

Are the microtubules shown on the cell surface (TIRF microscopy) or do we see here all microtubules? 

Please see Materials and Methods for microscopy methods and image processing for each figure. Specifically, FRAP assays show a maximum intensity projection of spinning disk confocal stacks over 2.4µm in height (approximately the ventral half of a cell).

(3) Previously, it has been shown that KIF5B induces tubulin incorporation along the microtubule shaft in a concentration-dependent manner. Moreover, running KIF5B increases microtubule rescue frequency and unlimited growth of microtubules. Notably, KIF5B regulates microtubule network mass and organization in cells (PMID: 34883065). Consequently, it appears possible that the here observed phenomena of changes in the microtubule network might be due to alterations in these processes. Authors need to exclude these possibilities and discuss them. 

Thank you for this interesting alternative idea. It does not appear to be the case for beta cells. Please see Figure 2-Supplemental Figure 2  and our response to Public Review Comment #3.

(4) It is important that the authors describe in the text and possibly in the figure legends the differences between the experimental set-ups including the length of cultivation and knock down of KIF5B or expression of mutants. 

Thank you, please see these details in the text (Materials and Methods section).

(5) Figure 5: Does KIF5B depletion rescue the kinesore-induced defects 

Thank you for suggesting this control. We have now conducted corresponding experiments. The answer is yes, it does. Kinesore does not induce detectable changes in MT patterns in KIF5Bdepleted cells (new Figure 5-Supplemental Figure 2). 

(6) Can the authors block kinesin-1 resulting in microtubule accumulation in the cell center and then release the block, and best inhibiting microtubule formation, to see whether the microtubules accumulated in the cell center will be transported to the periphery? 

This proposed experiment would have been a nice illustration to the study, however it has proven to be too challenging. Unfortunately we have to leave it for the future studies. However,  the experiments already included in the paper are sufficient to prove our conclusions. 

Minor comments: 

(1) The English needs to be improved. Oaen it is unclear what the authors try to convey. The manuscript is difficult to read and contains several overstatements. 

The revised manuscript has been through several rounds of proof-reading for clarity.

(2) It is important to describe in more detail in the introduction what is known about KIF5B in beta cells. Previously, it has been demonstrated that silencing, or inactivation by a dominant negative form of KIF5B, blocks the sustained phase of glucose-stimulated insulin secretion (PMID: 9112396, PMID: 12356920, PMID: 20870970). 

Yes, this is of course very important and have been cited in the original manuscript. Now, we have expanded the discussion on the matter.

(3) Figure 1B and Fig. 1 Suppl Fig.1: Please provide band sizes and provide information on the size of KIF5B. 

We have replaced Fig. 1B and Suppl Fig 1A with quantitative analysis of KIF5B depletion, not found in new Fig. 1B and Suppl Fig. 1A-C. 

(4) It is important to state the used glucose concentrations in Figure 1D (based on the methods section it is probably 25 mM glucose) and all subsequent experiments. Is this correct and comparable to Figure 6A or B? For the non-specialized reader, more information should be provided on why initial glucose starvation is performed.  

Cell culture models of pancreatic beta cells are routinely maintained at glucose levels that at considered “high”, or stimulatory for secretion. This is needed to prevent the loss of cells’ capacity to respond to glucose stimulation over generations. In order to test GSIS, cells need to be equilibrated at low (fasting, standardly 2.8mM) glucose levels for several hours, so that they are capable of secreting insulin upon glucose addition. 25mM glucose is normally used to stimulate GSIS in cell culture models of beta cells, like MIN6. This is a higher concentration as compared to what is needed to stimulate primary beta cells in islets.

Reviewer #2 (Recommendations For The Authors): 

I have the following specific questions that pertain to data interpretation and the conclusions drawn.

(1) The morphology of the overall MT array before the bleach treatment in both control cells and KIF5B-KD cells depicted in Figure 1D-F and Figure 2A-C appears to be distinct. In Figure 1, it seems that the absence of KIF5B results in a general augmentation of MT mass, whereas the arrangement presented in Figure 2 indicates the contrary. Even in the sub-membrane areas, this phenomenon appears to hold true. However, the images used in this study, which depict entire cells or a significant portion of cells, may not be ideal for visualizing the sub-membrane regions.

It would be beneficial if the author could offer some explanations for this apparent inconsistency. 

While beta cell population is intrinsically heterogeneous, all data presented here indicate the same trend in phenotypes. Possibly, some apparent inconsistency between figure 1 and 2 appeared because in the original manuscript we did not show the pre-bleach whole-cell overview in Figure 1. In the revised version, we now show the whole cells for pre-bleach so that MT organization at the cell periphery can be assessed. Please note that in the control cell, MTs are more or less equally distributed over the cell, while in KIF5B depletions the cell periphery is significantly less populated than the cell center. Furthermore, we did not detect MT mass augmentation or increase in KIF5B depletions. One possible explanation for such reviewer’s impression from Figure 2 is that Figure 2 F-H shows thresholded images where threshold was adjusted to highlight peripheral MTs in each cell. Please note that this is not the same threshold for each cell (see Figure 2 - Supplemental Figure 2 and 3). Thus, KIF5B-depleted cells that have fewer MTs at the periphery appear brighter in these thresholded images. For the true comparison of MT intensity, please see Figure 2 A-C (grayscale image, not the threshold).

(2) It would be helpful if the author could provide a visual representation or comment on the sub-cellular localization of KIF5B in MIN6 cells. Is it predominantly localized in the submembrane region, or is it more evenly distributed throughout the cytoplasm? 

Please see Fig 2A, lower panel. KIF5B is seen across the cell as a punctate staining, in agreement with previous findings that it mostly localize at IGs.

(3) The alteration in microtubule (MT) organization and sliding in the absence of KIF5B seems to initiate in proximity to the apparent microtubule organizing center (MTOC) depicted in Figure 2A, and then "simply" extends towards the sub-membrane region. Although the authors acknowledge it, it would be advantageous for the readers to have a clearer indication that the sub-membrane microtubule (MT) reorganization in the absence of KIF5B is a result of a broader MT reorganization rather than a specific occurrence restricted to the sub-membrane regions. 

Thank you for this comment. We now extend our discussion to clearer state our conclusions and interpretations of this point. We also have added a schematic Figure 7 as an illustration. 

(4) Regarding the "glucose experiments," it is common to add 20-25 mM glucose to culture media, but physiological concentrations of glucose typically hover around 5 mM. Therefore, it is somewhat unclear what the implications are when investigating the impact of KIF5B depletion on MT sliding at 2.8 mM of glucose. It would be helpful if the authors could provide some commentary on this matter, particularly in relation to physiological and pathological conditions. 

2.8 mM glucose is a standard low glucose condition used to model glucose deprivation/fasting. For functional primary beta cells within pancreatic islets, GSIS can be triggered by glucose stimulation as low as 8-12 mM glucose. However, for glucose stimulation of cultured beta cells such as MIN6 used in this paper, 20-25 mM glucose is standardly used because these cell lines have a higher threshold of stimulation compared to primary beta cells and whole islets.

(5) In supplementary Figure 1A, it would be helpful if the lanes in the WB were marked indicating what is what. In my observation, it appears that Supplementary Figure 1A, particularly lanes #2, 3, and 4, display the GAPDH protein (MW 36 kDa) (or is it alpha-tubulin, as mentioned in the Material and Methods section and indicated in lane #409?) relative to Figure 1A. I am curious about KIF5B (MW 108 kDa). Is it represented by the upper band? Did the author probe the same membrane simultaneously with two different primary antibodies? This should be clarified, and the author should indicate the molecular weight of the ladder. 

Indeed, in the original WB two antibodies have been used together, due to a challenge in collecting a sufficient number of shRNA-expressing beta cells. It caused a confusion and improper interpretation of the loading control. We thank the reviewer for catching this.  We have now replaced old Fig. 1B and Suppl. Fig. 1A with quantitative analysis of KIF5B depletion based on single-cell immunofluorescent staining. It is now found in new Fig. 1B and Suppl Fig. 1A-C.  

Reviewer #3 (Recommendations For The Authors): 

In all of the figures that present microtubule orientations (e.g. Figure 2E) the error bars obscure the vertical bins making them difficult to read or interpret. If they were rendered at a larger scale, it would be easier to read and interpret these results. 

Thank you pointing this out. We now show these histograms with a different format of error bars and without outliers that obscure the view. A variant with outliers is now shown in the supplement. 

Some of the callouts to the videos in the paper are inaccurate. Perhaps the authors reordered sections of the paper but failed to correctly renumber the video citations? 

Thank you for this comment, we have corrected all callouts now.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

Summary:

This short report shows that the transcription factor gene mirror is specifically expressed in the posterior region of the butterfly wing imaginal disk, and uses CRISPR mosaic knock-outs to show it is necessary to specify the morphological features (scales, veins, and surface) of this area.

Strengths:

The data and figures support the conclusions. The article is swiftly written and makes an interesting evolutionary comparison to the function of this gene in Drosophila. Based on the data presented, it can now be established that mirror likely has a similar selector function for posterior-wing identity in a plethora of insects.

We thank the reviewer for their feedback.

Weaknesses:

This first version has minor terminological issues regarding the use of the terms "domains" and "compartment".

We acknowledge that the terminologies “domains” and “compartments” might lead to confusion. To avoid confusion we have removed the term “compartment” from the manuscript.

Reviewer #2 (Public Review):

This is a short and unpretentious paper. It is an interesting area and therefore, although much of this area of research was pioneered in flies, extending basic findings to butterflies would be worthwhile. Indeed, there is an intriguing observation but it is technically flawed and these flaws are serious.

The authors show that mirror is expressed at the back of the wing in butterflies (as in flies). They present some evidence that is required for the proper development of the back of the wing in butterflies (a region dubbed the vannus by the ancient guru Snodgrass). But there are problems with that evidence. First, concerning the method, using CRISP they treat embryos and the expectation is that the mirror gene will be damaged in groups of cell lineages, giving a mosaic animal in which some lines of cells are normal for mirror and others are not. We do not know where the clones or patches of cells that are defective for mirror are because they are not marked. Also, we do not know what part of the wing is wild type and what part is mutant for mirror. When the mirror mutant cells colonise the back of the wing and that butterfly survives (many butterflies fail to develop), the back of the wing is altered in some selected butterflies. This raises a second problem: we do not know whether the rear of the wing is missing or transformed. From the images, the appearance of the back of the wing is clearly different from the wild type, but is that due to transformation or not? And then I believe we need to know specifically what the difference is between the rear of the wing and the main part. What we see is a silvery look at the back that is not present in the main part, is it the structure of the scales? We are not told.

Thank you for this feedback. We appreciate that many readers may not accustomed to looking at mosaic knockouts. As discussed in a previous review article (Zhang & Reed 2017), we rely on a combination of contralateral asymmetry and replicates to infer mutant phenotypes. For many genes (e.g. pigmentation enzymes) mutant clones are obvious, but for other types of genes (e.g. ligands) clone boundaries are sometimes not directly diagnosable. It is simply a limitation of our study system. Nonetheless, you see for yourself that “the back of the wing is altered in some butterflies” – the effects of deleting mirror are clear and repeatable.

In terms of interpreting mutant phenotypes, we agree that that paper would benefit from a better description of the specific effects. Therefore, we have included an improved, more systematic description of phenotypes, along with better-annotated figures showing changes in wing shape and venation, scale coloration, and color pattern transformation (e.g. posterior elongation of the orange marginal stripes).

There are other problems. Mirror is only part of a group of genes in flies and in flies both iroquois and mirror are needed to make the back of the wing, the alula (Kehl et al). What is known about iro expression in butterflies?

In Drosophila mirror, araucan, and caupolican comprise the so-called Iroqouis Complex of genes. As denoted in Figure S4 and in Kerner et al (doi: https://doi.org/10.1186/1471-2148-9-74) the divergence of araucan and caupolican into two separate paralogs is restricted to Drosophila. As in most insects, butterflies have only two Iroquois Complex genes: araucan and mirror. We tested the role of araucan in Junonia coenia as shown in our pre-print: https://doi.org/10.1101/2023.11.21.568172. Its expression appears to be restricted to early pupal wings where it is transcribed in all scale-forming cells. Mosaic araucan KOs resulted in a change in scale iridescent coloration associated with changes in the laminar thickness of scale cells.  

In flies, mirror regulates a late and local expression of dpp that seems to be responsible for making the alula. What happens in butterflies? Would a study of the expression of Dpp in wildtype and mirror compromised wings be useful?

We thank the reviewer for the proposal and agree that a future study comparing Dpp in wild-type versus mirror KO butterflies would be useful to clarify the mechanism of Dpp signalling in wing development. It is not clear, however, that the results of a Dpp experiment would change the conclusions of our current study therefore we decided not to undertake these additional experiments for our revision.

Thus, I find the paper to be disappointing for a general journal as it does little more than claim what was discovered in Drosophila is at least partly true in butterflies. 

We respect that the reviewer does not have a strong interest in the comparative aspects of this study. Fair enough. This report is primarily aimed at biologists interested in the evolutionary history of insect wings.

Also, it fails to explain what the authors mean by "wing domains" and "domain specification". They are not alone, butterfly workers, in general, appear vague about these concepts, their vagueness allowing too much loose thinking.

A domain is “a region distinctively marked by some physical feature”. This term is used extensively in the developmental biology literature (e.g. “expression domain”, “embryonic domain”, “tissue domain”, “domain specification”) and is found throughout popular textbooks (e.g. Alberts et al. “The Cell”, Gilbert “Developmental Biology”). We prefer the term “domain” because of its association in the Drosophila literature with transcription factors that define fields of cells. We specifically avoided using the term “compartment” because of its association with cell lineage, which we have not tested. 

Since these matters are at the heart of the purpose and meaning of the work reported here, we readers need a paper containing more critical thought and information. I would like to have a better and more logical introduction and discussion.

We would like the very same thing, of course, and we hope the reviewer finds our revised manuscript to be more satisfying to read.

The authors do define what they mean by the vannus of the wing. In flies the definition of compartments is clear and abundantly demonstrated, with gene expression and requirement being limited precisely to sets of cells that display lineage boundaries. It is true that domains of gene expression in flies, for example of the iroquois complex, which includes mirror, can only be related to patterns with difficulty. Some recap of what is known plus the opinion of the authors on how they interpret papers on possible lineage domains in butterflies might also be useful as the reader, is no wiser about what the authors might mean at the end of it!

We thank the reviewer for this suggestion. However, our experiments have little to contribute to the topic of cell lineage compartmentalization. We have therefore opted to avoid speculating on this topic to prevent confusion and to keep the manuscript focused on our experimental results.

The references are sometimes inappropriate. The discovery of the AP compartments should not be referred to Guillen et al 1995, but to Morata and Lawrence 1975. Proofreading is required.

We thank the reviewer for suggesting this important reference. We have included it in our revision.

Reviewer #3 (Public Review):

Summary:

The manuscript by Chatterjee et al. examines the role of the mirror locus in patterning butterfly wings. The authors examine the pattern of mirror expression in the common buckeye butterfly, Junonia coenia, and then employ CRISPR mutagenesis to generate mosaic butterflies carrying clones of mirror mutant cells. They find that mirror is expressed in a well-defined posterior sector of final-instar wing discs from both hindwings and forewings and that CRISPR-injected larvae display a loss of adult wing structures presumably derived from the mirror expressing region of hindwing primordium (the case for forewings is a bit less clear since the mirror domain is narrower than in the hindwing, but there also do seem to be some anomalies in posterior regions of forewings in adults derived from CRISPR injected larvae). The authors conclude that the wings of these butterflies have at least three different fundamental wing compartments, the mirror domain, a posterior domain defined by engrailed expression, and an anterior domain expressing neither mirror nor engrailed. They speculate that this most posterior compartment has been reduced to a rudiment in Drosophila and thus has not been adequately recognized as such a primary regional specialization.

Critique:

This is a very straightforward study and the experimental results presented support the key claims that mirror is expressed in a restricted posterior section of the wing primordium and that mosaic wings from CRISPR-injected larvae display loss of adult wing structures presumably derived from cells expressing mirror (or at least nearby). The major issue I have with this paper is the strong interpretation of these findings that lead the authors to conclude that mirror is acting as a high-level gene akin to engrailed in defining a separate extreme posterior wing compartment. To place this claim in context, it is important in my view to consider what is known about engrailed, for which there is ample evidence to support the claim that this gene does play a very ancestral and conserved function in defining posterior compartments of all body segments (including the wing) across arthropods.

(1) Engrailed is expressed in a broad posterior domain with a sharp anterior border in all segments of virtually all arthropods examined (broad use of a very good panspecies anti-En antibody makes this case very strong).

(2) In Drosophila, marked clones of wing cells (generated during larval stages) strictly obey a straight anterior-posterior border indicating that cells in these two domains do not normally intermix, thus, supporting the claim that a clear A/P lineage compartment exists.

In my opinion, mirror does not seem to be in the same category of regulator as engrailed for the following reasons:

(1) There is no evidence that I am aware of, either from the current experiments, or others that the mirror expression domain corresponds to a clonal lineage compartment. It is also unclear from the data shown in this study whether engrailed is co-expressed with mirror in the posterior-most cells of J. coenia wing discs. If so, it does not seem justified to infer that mirror acts as an independent determinant of the region of the wing where it is expressed.

(2) Mirror is not only expressed in a posterior region of the wing in flies but also in the ventral region of the eye. In Drosophila, mirror mutants not only lack the alula (derived approximately from cells where mirror is expressed), but also lack tissue derived from the ventral region of the eye disc (although this ventral tissue loss phenotype may extend beyond the cells expressing mirror).

In summary, it seems most reasonable to me to think of mirror as a transcription factor that provides important development information for a diverse set of cells in which it can be expressed (posterior wing cells and ventral eye cells) but not that it acts as a high-level regulator as engrailed.

Recommendation:

While the data provided in this succinct study are solid and interesting, it is not clear to me that these findings support the major claim that mirror defines an extreme posterior compartment akin to that specified by engrailed. Minimally, the authors should address the points outlined above in their discussion section and greatly tone down their conclusion regarding mirror being a conserved selector-like gene dedicated to establishing posterior-most fates of the wing. They also should cite and discuss the original study in Drosophila describing the mirror expression pattern in the embryo and eye and the corresponding eye phenotype of mirror mutants: McNeill et al., Genes & Dev. 1997. 11: 1073-1082; doi:10.1101/gad.11.8.1073.

We thank the reviewer for their summary, critique, and recommendations. We agree with everything the reviewer says. Honestly, however, we were surprised by these comments because we took great care in the paper to never refer to mirror as a compartmentalization gene or claim it has a function in cell lineage compartmentalization like engrailed. As pointed out, we lack clonal analyses to test for compartmentalization. This is why we used the term “domain” instead of “compartment” in the title and throughout the manuscript. Nevertheless, we have recrafted the discussion in the manuscript, including completely removing the term “compartment”, to better avoid implications that mirror plays a role in cell lineage compartmentalization. 

We also thank the reviewer for recommending the paper about the role of mirror in eye development. For the sake of keeping the paper focused, however, we decided not to broach the topic of mirror functions outside the context of wing development.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

I have minor comments for improvement.

The abstract and introductions are terminologically problematic when they refer to the concept of compartment and compartment boundaries. Allegedly this confusion has previously propagated in several articles related to butterfly wing development, which keeps alienating this literature from being taken seriously by fly specialists, for example. So it is important to use the right terms. I will try to explain point by point here, but I would appreciate it if the authors could undertake a significant rewrite taking these comments into account. The authors use the terms compartment and compartment boundary. This has a very specific use in developmental genetics: mitotic clones never cross a boundary (or compartment). I think the authors can keep referring to the equivalent of the A-P boundary, which is situated somewhere between M1-M2 based on unpublished data from the Patel Lab, and is not very well defined (Engrailed expression moves a little bit during development in this area). Domain is a looser term and can be used more liberally to describe genetically defined regions.

- "Classical morphological work subdivides insect wings into several distinct domains along the antero-posterior (AP) axis, each of which can evolve relatively independently." Yes. This concept of domain and individuation seems important. You could make a proposed link to selector genes here.

- "There has been little molecular evidence, however, for AP subdivision beyond a single compartment boundary described from Drosophila melanogaster." Incorrect, and this conflates "domain" and "compartment".

Flies have wing AP domains too, that pattern their veins (see the cited Banerjee et al). 

- "Our results confirm that insect wings can have more than one posterior developmental domain, and support models of how selector genes may facilitate evolutionarily individuation of distinct AP domains in insect wings". Yes, and I like the second part of the sentence. Still, I would recommend simply deleting "confirm that insect wings can have more than one posterior developmental domain, and" because this is neglecting previous work on AP genetic regionalization in both flies (vein literature) and butterflies (e.g. McKenna and Nijhout, Banerjee et al).

- "Analyses of wing pattern diversity across butterflies, considering both natural variation and genetic mutants, suggest that wings can be subdivided into at least five AP domains, bounded by the M1, M3, Cu2, and 2A veins respectively, within each of which there are strong correlations in color pattern variation and wing morphology (Figure 1A)". Yes, and I would recommend emphasizing they correspond to welldefined gene expression domains as mentioned in Banerjee et al, or McKenna and Nijhout.

- "The anterior-most of these domains, bordered by the M1 vein, appears to correspond to an AP compartment boundary originally described by cell lineage tracing in Drosophila melanogaster, and later supported in butterfly wings by expression of the Engrailed transcription factor. Interestingly, however, D. melanogaster work has yet to reveal clear evidence for additional AP domain boundaries in the wing." Confusingly, because the first sentence is about compartments while the second is about AP domains. I also think the claim that Dmel has no other known AP domains is dubious because Spalt is highly regionalized in flies.

- "Previous authors have proposed the existence of such individuated domains, and speculated that they may be specified by selector genes.5,10 Our data provide experimental support for this model, and now motivate us to identify factors that specify other domain boundaries between the M1 and A2 veins." Yes, I completely agree with this way to emphasize the selector effect, and to link it to the concept of "individuated domain"

We cannot thank the reviewer enough for the time and thought they devoted to giving helpful suggestions to improve our manuscript. We have applied all of the above recommendations to the revision.

Fig. S1: the field needs to move away from Red/Green microscopy images, for accessibility reasons.

The easiest fix here would be to change the red channels to magenta.

Green/Magenta provides excellent contrast and accessibility in general in 2-channel images.

We thank the reviewer for this suggestion. We have improved the color accessibility of Fig. S1.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

The study addresses how faces and bodies are integrated in two STS face areas revealed by fMRI in the primate brain. It builds upon recordings and analysis of the responses of large populations of neurons to three sets of images, that vary face and body positions. These sets allowed the authors to thoroughly investigate invariance to position on the screen (MC HC), to pose (P1 P2), to rotation (0 45 90 135 180 225 270 315), to inversion, to possible and impossible postures (all vs straight), to the presentation of head and body together or in isolation. By analyzing neuronal responses, they found that different neurons showed preferences for body orientation, head orientation, or the interaction between the two. By using a linear support vector machine classifier, they show that the neuronal population can decode head-body angle presented across orientations, in the anterior aSTS patch (but not middle mSTS patch), except for mirror orientation.

Strengths:

These results extend prior work on the role of Anterior STS fundus face area in face-body integration and its invariance to mirror symmetry, with a rigorous set of stimuli revealing the workings of these neuronal populations in processing individuals as a whole, in an important series of carefully designed conditions.

Minor issues and questions that could be addressed by the authors:

(1) Methods. While monkeys certainly infer/recognize that individual pictures refer to the same pose with varying orientations based on prior studies (Wang et al.), I am wondering whether in this study monkeys saw a full rotation of each of the monkey poses as a video before seeing the individual pictures of the different orientations, during recordings.

The monkeys had not been exposed to videos of a rotating monkey pose before the recordings. However, they were reared and housed with other monkeys, providing them with ample experience of monkey poses from different viewpoints.

(2) Experiment 1. The authors mention that neurons are preselected as face-selective, body-selective, or both-selective. Do the Monkey Sum Index and ANOVA main effects change per Neuron type?

We have performed a new analysis to assess whether the Monkey Sum Index is related to the response strength for the face versus the body as measured in the Selectivity Test of Experiment 1. To do this we selected face- and body-category selective neurons, as well as neurons responding selectively to both faces and bodies. First, we selected those neurons that responded significantly to either faces, bodies, or the two control object categories, using a split-plot ANOVA for these 40 stimuli. From those neurons, we selected face-selective ones having at least a twofold larger mean net response to faces compared to bodies (faces > 2 * bodies) and the control objects for faces (faces  > 2* objects). Similarly, a body-selective neuron was defined by a twofold larger mean net response to bodies compared to faces and the control objects for bodies. A body-and-face selective neuron was defined as having a twofold larger net response to the faces compared to their control objects, and to bodies compared to their control objects, with the ratio between mean response to bodies and faces being less than twofold. Then, we compared the distribution of the Monkey Sum Index (MSI) for each region (aSTS; mSTS), pose (P1, P2), and centering (head- (HC) or monkey-centered (MC)) condition. Too few body-and-face selective neurons were present in each combination of region, pose, and centering (a maximum of 7) to allow a comparison of their MSI distribution with the other neuron types. The Figure below shows the distribution of the MSI for the different orientation-neuron combinations for the body- and face-selective neurons (same format as in Figure 3a, main text). The number of body-selective neurons, according to the employed criteria, varied from 21 to 29, whereas the number of face-selective neurons ranged from 14 to 24 (pooled across monkeys). The data of the two subjects are shown in a different color and the number of cases for each subject is indicated (n1: number of cases for M1; n2: number of cases for M2). The arrows indicate the medians for the data pooled across the monkey subjects. For the MC condition, the MSI tended to be more negative (i.e. relatively less response to the monkey compared to the sum of the body and face responses) for the face compared to the body cells, but this was significant only for mSTS and P1 (p = 0.043; Wilcoxon rank sum test; tested after averaging the indices per neuron to avoid dependence of indices within a neuron). No consistent, nor significant tendencies were observed for the HC stimuli. This absence of a consistent relationship between MSI and face- versus body-selectivity is in line with the absence of a correlation between the MSI and face- versus body-selectivity using natural images of monkeys in a previous study (Zafirova Y, Bognár A, Vogels R. Configuration-sensitive face-body interactions in primate visual cortex. Prog Neurobiol. 2024 Jan;232:102545).

We did not perform a similar analysis for the main effects of the two-way ANOVA because the very large majority of neurons showed a significant effect of body orientation and thus no meaningful difference between the two neuron types can be expected.

Author response image 1.

(3) I might have missed this information, but the correlation between P1 and P2 seems to not be tested although they carry similar behavioral relevance in terms of where attention is allocated and where the body is facing for each given head-body orientation.

Indeed, we did not compute this correlation between the responses to the sitting (P1) and standing (P2) pose avatar images. However, as pointed out by the reviewer, one might expect such correlations because of the same head orientations and body-facing directions. Thus, we computed the correlation between the 64 head-body orientation conditions of P1 and P2 for those neurons that were tested with both poses and showed a response for both poses (Split-plot ANOVA). This was performed for the Head-Centered and Monkey-Centered tests of Experiment 1 for each monkey and region. Note that not all neurons were tested with both poses (because of failure to maintain isolation of the single unit in both tests or the monkey stopped working) and not all neurons that were recorded in both tests showed a significant response for both poses, which is not unexpected since these neurons can be pose selective. The distribution of the Pearson correlation coefficients of the neurons with a significant response in both tests is shown in Figure S1. The median correlation coefficient was significantly larger than zero for each region, monkey, and centering condition (outcome of Wilcoxon tests, testing whether the median was different from zero (p1 = p-value for M1; p2: p-value for M2) in Figure), indicating that the effect of head and/or body orientation generalizes across pose. We have noted this now in the Results (page 12) and added the Figure (New Figure S1) in the Suppl. Material.

(4) Is the invariance for position HC-MC larger in aSTS neurons compared to mSTS neurons, as could be expected from their larger receptive fields?

Yes, the position tolerance of the interaction of body and head orientation was significantly larger for aSTS compared to mSTS neurons, as we described on pages 11 and 12 of the Results. This is in line with larger receptive fields in aSTS than in mSTS. However, we did not plot receptive fields in the present study.

(5) L492 "The body-inversion effect likely results from greater exposure to upright than inverted bodies during development". Monkeys display more hanging upside-down behavior than humans, however, does the head appear more tilted in these natural configurations?

Indeed, infant monkeys do spend some time hanging upside down from their mother's belly. While we lack quantitative data on this behavior, casual observations suggest that even young monkeys spend more time upright. The tilt of the head while hanging upside down can vary, just as it does in standing or sitting monkeys (as when they search for food or orient to other individuals). To our knowledge, no quantitative data exist on the frequency of head tilts in upright versus upside-down monkeys. Therefore, we refrain from further speculation on this interesting point, which warrants more attention.

(6) Methods in Experiment 1. SVM. How many neurons are sufficient to decode the orientation?

The number of neurons that are needed to decode the head-body orientation angle depends on which neurons are included, as we show in a novel analysis of the data of Experiment 1. We employed a neuron-dropping analysis, similar to Chiang et al. (Chiang FK, Wallis JD, Rich EL. Cognitive strategies shift information from single neurons to populations in prefrontal cortex. Neuron. 2022 Feb 16;110(4):709-721) to assess the positive (or negative) contribution of each neuron to the decoding performance. We performed cross-validated linear SVM decoding N times, each time leaving out a different neuron (using N-1 neurons; 2000 resamplings of pseudo-population vectors). We then ranked decoding accuracies from highest to lowest, identifying the ‘worst’ (rank 1) to ‘best’ (rank N) neurons. Next, we conducted N decodings, incrementally increasing the number of included neurons from 1 to N, starting with the worst-ranked neuron (rank 1) and sequentially adding the next (rank 2, rank 3, etc.). This analysis focused on zero versus straight angle decoding in the aSTS, as it yielded the highest accuracy. We applied it when training on MC and testing on HC for each pose. Plotting accuracy as a function of the number of included neurons suggested that less than half contributed positively to decoding. We show also the ten “best” neurons for each centering condition and pose. These have a variety of tuning patterns for head and body orientation suggesting that the decoding of head-body orientation angle depends on a population code. Notably, the best-ranked (rank N) neuron alone achieved above-chance accuracy. We have added this interesting and novel result to the Results (page 16) and Suppl. Material (new Figure S3).

(7) Figure 3D 3E. Could the authors please indicate for each of these neurons whether they show a main effect of face, body, or interaction, as well as their median corrected correlation to get a flavor of these numbers for these examples?

We have indicated these now in Figure 3.

(8) Methods and Figure 1A. It could be informative to precise whether the recordings are carried in the lateral part of the STS or in the fundus of the STS both for aSTS and mSTS for comparison to other studies that are using these distinctions (AF, AL, MF, ML).

In experiment 1, the recording locations were not as medial as the fundus. For experiments 2 and 3, the ventral part of the fundus was included, as described in the Methods. We have added this to the Methods now (page 31).

Wang, G., Obama, S., Yamashita, W. et al. Prior experience of rotation is not required for recognizing objects seen from different angles. Nat Neurosci 8, 1768-1775 (2005). https://doi-org.insb.bib.cnrs.fr/10.1038/nn1600

Reviewer #2 (Public review):

Summary:

This paper investigates the neuronal encoding of the relationship between head and body orientations in the brain. Specifically, the authors focus on the angular relationship between the head and body by employing virtual avatars. Neuronal responses were recorded electrophysiologically from two fMRI-defined areas in the superior temporal sulcus and analyzed using decoding methods. They found that: (1) anterior STS neurons encode head-body angle configurations; (2) these neurons distinguish aligned and opposite head-body configurations effectively, whereas mirror-symmetric configurations are more difficult to differentiate; and (3) an upside-down inversion diminishes the encoding of head-body angles. These findings advance our understanding of how visual perception of individuals is mediated, providing a fundamental clue as to how the primate brain processes the relationship between head and body - a process that is crucial for social communication.

Strengths:

The paper is clearly written, and the experimental design is thoughtfully constructed and detailed. The use of electrophysiological recordings from fMRI-defined areas elucidated the mechanism of head-body angle encoding at the level of local neuronal populations. Multiple experiments, control conditions, and detailed analyses thoroughly examined various factors that could affect the decoding results. The decoding methods effectively and consistently revealed the encoding of head-body angles in the anterior STS neurons. Consequently, this study offers valuable insights into the neuronal mechanisms underlying our capacity to integrate head and body cues for social cognition-a topic that is likely to captivate readers in this field.

Weaknesses:

I did not identify any major weaknesses in this paper; I only have a few minor comments and suggestions to enhance clarity and further strengthen the manuscript, as detailed in the Private Recommendations section.

Reviewer #3 (Public review):

Summary:

Zafirova et al. investigated the interaction of head and body orientation in the macaque superior temporal sulcus (STS). Combining fMRI and electrophysiology, they recorded responses of visual neurons to a monkey avatar with varying head and body orientations. They found that STS neurons integrate head and body information in a nonlinear way, showing selectivity for specific combinations of head-body orientations. Head-body configuration angles can be reliably decoded, particularly for neurons in the anterior STS. Furthermore, body inversion resulted in reduced decoding of head-body configuration angles. Compared to previous work that examined face or body alone, this study demonstrates how head and body information are integrated to compute a socially meaningful signal.

Strengths:

This work presents an elegant design of visual stimuli, with a monkey avatar of varying head and body orientations, making the analysis and interpretation straightforward. Together with several control experiments, the authors systematically investigated different aspects of head-body integration in the macaque STS. The results and analyses of the paper are mostly convincing.

Weaknesses:

(1) Using ANOVA, the authors demonstrate the existence of nonlinear interactions between head and body orientations. While this is a conventional way of identifying nonlinear interactions, it does not specify the exact type of the interaction. Although the computation of the head-body configuration angle requires some nonlinearity, it's unclear whether these interactions actually contribute. Figure 3 shows some example neurons, but a more detailed analysis is needed to reveal the diversity of the interactions. One suggestion would be to examine the relationship between the presence of an interaction and the neural encoding of the configuration angle.

This is an excellent suggestion. To do this, one needs to identify the neurons that contribute to the decoding of head-body orientation angles. For that, we employed a neuron-dropping analysis, similar to Chiang et al. (Chiang FK, Wallis JD, Rich EL. Cognitive strategies shift information from single neurons to populations in prefrontal cortex. Neuron. 2022 Feb 16;110(4):709-721.) to assess the positive (or negative) contribution of each neuron to the decoding performance. We performed cross-validated linear SVM decoding N times, each time leaving out a different neuron (using N-1 neurons; 2000 resamplings of pseudo-population vectors). We then ranked decoding accuracies from highest to lowest, identifying the ‘worst’ (rank 1) to ‘best’ (rank N) neurons. Next, we conducted N decodings, incrementally increasing the number of included neurons from 1 to N, starting with the worst-ranked neuron (rank 1) and sequentially adding the next (rank 2, rank 3, etc.). This analysis focused on zero versus straight angle decoding in the aSTS, as it yielded the highest accuracy. We applied it when training on MC and testing on HC for each pose. Plotting accuracy as a function of the number of included neurons suggested that less than half contributed positively to decoding (see Figure S3). We examined the tuning for head and body orientation of the 10 “best” neurons (Figure S3). For half or more of those the two-way ANOVA showed a significant interaction. These are indicated by the red color in the Figure. They showed a variety of tuning patterns for head and body orientation, suggesting that the decoding of the head-body orientation angle results from a combination of neurons with different tuning profiles. Based on a suggestion from reviewer 2, we performed for each neuron of experiment 1 a one-way ANOVA with as factor head-body orientation angle. To do that, we combined all 64 trials that had the same head-body orientation angle. The percentage of neurons (required to be responsive in the tested condition) for which this one-way ANOVA was significant was low but larger than the expected 5% (Type 1 error), with a median of 16.5% (range: 3 to 23%) in aSTS and 8% for mSTS (range: 0-19%). However, a higher percentage of the 10 best neurons for each pose (indicated by the star) showed a significant one-way ANOVA for angle (for P1, MC: 50% (95% confidence interval (CI): 19% – 81%); P1, HC: 70% (CI: 35% - 93%); P2, MC: 70% (CI: 35% – 93%); P2: HC: 50% (CI: 19%-81%)). These percentages were significantly higher than expected for a random sample from the population of neurons for each pose-centering combination (expected percentages listed in the same order as above: 16%, 13%, 16%, and 10%; all outside CI). Thus, for at least half of the “best” neurons, the response differed significantly among the head-orientation angles at the single neuron level. Nonetheless, the tuning profiles were diverse, suggesting a populationl code for head-body orientation angle. We have added this interesting and novel result to the Results (page 16) and Suppl. Material (Figure S3).

(2) Figure 4 of the paper shows a better decoding of the configuration angle in the anterior STS than in the middle STS. This is an interesting result, suggesting a transformation in the neural representation between these two areas. However, some control analyses are needed to further elucidate the nature of this transformation. For example, what about the decoding of head and body orientations - dose absolute orientation information decrease along the hierarchy, accompanying the increase in configuration information?

We have performed now two additional analyses, one in which we decoded the orientation of the head and another one in which we decoded the orientation of the body. We employed the responses to the avatar of experiment 1, using the same sample of neurons of which we decoded the head-body orientation angle. To decode the head orientation, the trials with identical head orientation, irrespective of their body orientation, were given the same label. For this, we employed only responses in the head-centered condition. To decode the body orientation, the trials with identical body orientation, irrespective of their head orientation, had the same label, and we employed only responses in the body-centered condition. The decoding was performed separately for each pose (P1 and P2) and region. We decoded either the responses of 20 neurons (10 randomly sampled from each monkey for each of the 1000 resamplings), 40 neurons (20 randomly sampled per monkey), or 60 neurons (30 neurons per monkey) since the sample of 60 neurons yielded close to ceiling performance for the body orientation decoding. For each pose, the body orientation decoding was worse for aSTS than for mSTS, although this difference reached significance only for P1 and for the 40 neurons sample of P2 (p < 0.025; two-tailed test; same procedure as employed for testing the significance of the decoding of whole-body orientation for upright versus inverted avatars (Experiment 3))). Face orientation decoding was significantly worse for aSTS compared to mSTS. These results are in line with the previously reported decreased decoding of face orientation in the anterior compared to mid-STS face patches (Meyers EM, Borzello M, Freiwald WA, Tsao D. Intelligent information loss: the coding of facial identity, head pose, and non-face information in the macaque face patch system. J Neurosci. 2015 May 6;35(18):7069-81), and decreased decoding of body orientation in anterior compared to mid-STS body patches (Kumar S, Popivanov ID, Vogels R. Transformation of Visual Representations Across Ventral Stream Body-selective Patches. Cereb Cortex. 2019 Jan 1;29(1):215-229). As mentioned by the reviewer, this contrasts with the decoding of the head-body orientation angle, which increases when moving more anteriorly. We mention this finding now in the Discussion (page 27) and present the new Figure S10 in the Suppl. Material.    

(3) While this work has characterized the neural integration of head and body information in detail, it's unclear how the neural representation relates to the animal's perception. Behavioural experiments using the same set of stimuli could help address this question, but I agree that these additional experiments may be beyond the scope of the current paper. I think the authors should at least discuss the potential outcomes of such experiments, which can be tested in future studies.

Unfortunately, we do not have behavioral data. One prediction would be that the discrimination of head-body orientation angle, irrespective of the viewpoint of the avatar, would be more accurate for zero versus straight angles compared to the right versus left angles. We have added this to the Discussion (page 28).

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) P22 L373. It should read Figure S5C instead of S4C.

Thanks; corrected.

(2) Figure 7B. All inverted decoding accuracies, although significantly lower than upright decoding accuracies, appear significantly above baseline. Should the title be amended accordingly?

Thanks for pointing this out. To avoid future misunderstanding we have changed the title to:

“Integration of head and body orientations in the macaque superior temporal sulcus is stronger for upright bodies”

(3) Discussion L432-33. "with some neurons being tuned to a particular orientation of both the head and the body". Wouldn't that be visible as a diagonal profile on the normalized net responses in Fig 3D? Or can the Anova evidence such a tuning?

We meant to say that some neurons were tuned to a particular combination of head and body orientation, like the third aSTS example neuron shown in Figure 3D. We have corrected the sentence.

Reviewer #2 (Recommendations for the authors):

Major comment:

This paper effectively demonstrates that the angular relationship between the head and body can be decoded from population responses in the anterior STS. In other words, these neurons encode information about the head-body angle. However, how exactly do these neurons encode this information? Given that the study employed electrophysiological recordings from a local population of neurons, it might be possible to provide additional data on the response patterns of individual neurons to shed light on the underlying encoding mechanisms.

Although the paper already presents example response patterns (Figures 3D, E) and shows that STS neurons encode interactions between head and body orientations (Figure 3B), it remains unclear whether the angle difference between the head and body has a systematic effect on neuronal responses. For instance, a description of whether some neurons preferentially encode specific head-body angle differences (e.g., a "45-degree angle neuron"), or additional population analyses such as a one-way ANOVA with angle difference as the main effect (or two-way ANOVA with angle difference as one of the main effect), would be very informative. Such data could offer valuable insights into how individual neurons contribute to the encoding of head-body angle differences-a detail that may also be reflected in the decoding results. Alternatively, it is possible that the encoding of head-body angle is inherently complex and only discernible via decoding methods applied to population activity. Either scenario would provide interesting and useful information to the field.

We have performed two additional analyses which are relevant to this comment. First, we attempted to relate the tuning for body and head orientation with the decoding of the head-body orientation angle. To do this, one needs to identify the neurons that contribute to the decoding of head-body orientation angles. For that, we employed a neuron-dropping analysis, similar to Chiang et al. (Chiang FK, Wallis JD, Rich EL. Cognitive strategies shift information from single neurons to populations in prefrontal cortex. Neuron. 2022 Feb 16;110(4):709-721.) to assess the positive (or negative) contribution of each neuron to the decoding performance. We performed cross-validated linear SVM decoding N times, each time leaving out a different neuron (using N-1 neurons; 2000 resamplings of pseudo-population vectors). We then ranked decoding accuracies from highest to lowest, identifying the ‘worst’ (rank 1) to ‘best’ (rank N) neurons. Next, we conducted N decodings, incrementally increasing the number of included neurons from 1 to N, starting with the worst-ranked neuron (rank 1) and sequentially adding the next (rank 2, rank 3, etc.). This analysis focused on zero versus straight angle decoding in the aSTS, as it yielded the highest accuracy. We applied it when training on MC and testing on HC for each pose. Plotting accuracy as a function of the number of included neurons suggested that less than half contributed positively to decoding (see Figure S3). We examined the tuning for head and body orientation of the 10 “best” neurons (Figure S3). For half or more of those the two-way ANOVA showed a significant interaction. These are indicated by the red color in the Figure. They showed a variety of tuning patterns for head and body orientation, suggesting that the decoding of the head-body orientation angle results from a combination of neurons with different tuning profiles.

Second, we have followed the suggestion of the reviewer to perform for each neuron of experiment 1 a one-way ANOVA with as factor head-body orientation angle. To do that, we combined all 64 trials that had the same head-body orientation angle. The percentage of neurons (required to be responsive in the tested condition) for which this one-way ANOVA was significant is shown in the Tables below for each region, separately for each pose (P1, P2), centering condition (MC = monkey-centered; HC = head-centered) and monkey subject (M1, M2). The percentages were low but larger than the expected 5% (Type 1 error), with a median of 16.5% (range: 3 to 23%) in aSTS and 8% for mSTS (range: 0-19%).

Author response table 1.

Interestingly, a higher percentage of the 10 best neurons for each pose (indicated by the star in the Figure above) showed a significant one-way ANOVA for angle (for P1, MC: 50% (95% confidence interval (CI): 19% – 81%); P1, HC: 70% (CI: 35% - 93%); P2, MC: 70% (CI: 35% – 93%); P2: HC: 50% (CI: 19%-81%)). These percentages were significantly higher than expected for a random sample from the population of neurons for each pose-centering combination (expected percentages listed in the same order as above: 16%, 13%, 16%, and 10%; all outside CI). Thus, for at least half of the “best” neurons, the response differed significantly among the head-orientation angles at the single neuron level. Nonetheless, the tuning profiles were quite diverse, suggesting population coding of head-body orientation angle. We have added this interesting and novel result to the Results (page 16) and Suppl. Material (Figure S3).    

Minor comments:

(1) Figure 4A, Fourth Row Example (Zero Angle vs. Straight Angle, Bottom of the P2 Examples): The order of the example stimuli might be incorrect- the 0{degree sign} head with 180{degree sign} body stimulus (leftmost) might be swapped with the 180{degree sign} head with 0{degree sign} body stimulus (5th from the left). While this ordering may be acceptable, please double-check whether it reflects the authors' intended arrangement.

We have changed the order of the two stimuli in Figure 4A, following the suggestion of the reviewer.

(2) Page 12, Lines 192-194: The text states, "Interestingly, some neurons (e.g. Figure 3D) were tuned to a particular combination of a head and body irrespective of centering." However, Figure 3D displays data for a total of 10 neurons. Could you please specify which of these neurons are being referred to in this context?

The wording was not optimal. We meant to say that some neurons were tuned to a particular combination of head and body orientation, like the third aSTS example neuron of Figure 3D. We have rephrased the sentence and clarified which example neuron we referred to.

(3) Page 28, Lines 470-471: The text states, "We observed no difference in response strength between anatomically possible and impossible configurations." Please clarify which data were compared for response strength, as I could not locate the corresponding analyses.

The anatomically possible and impossible configurations differ in the head-body orientation angle. However, as we reported before in the Results, there was no effect of head-body orientation angle on mean response strength across poses (Friedman ANOVA; all p-values for both poses and centerings > 0.1). We have clarified this now in the Discussion (page 28).

(4) Pages 40-43, Decoding Analyses: In experiments 2 and 3, were the decoding analyses performed on simultaneously recorded neurons? If so, such analyses might leverage trial-by-trial correlations and thus avoid confounds from trial-to-trial variability. In contrast, experiment 1, which used single-shank electrodes, would lack this temporal information. Please clarify how trial numbers were assigned to neurons in each experiment and how this assignment may have influenced the decoding performance.

For the decoding analyses of experiments 2 and 3, we combined data from different daily penetrations, with only units from the same penetration being recorded simultaneously. In the decoding analyses of each experiment, the trials were assigned randomly to the pseudo-population vectors, shuffling on each resampling the trial order per neuron. This shuffling abolishes noise correlations in the analysis of each experiment.

(5) Page 41, Lines 792-802: The authors state that "To assess the significance of the differences in classification scores between pairs of angles ... we computed the difference in classification score between the two pairs for each resampling and the percentile of 0 difference corresponded to the p-value." In a two-sided test under the null hypothesis of no difference between the distributions, the conventional approach would be to compute the p-value as the proportion of resampled differences that are as extreme or more extreme than the observed difference. Since a zero difference might be relatively rare, relying solely on its percentile could potentially misrepresent the tail probabilities relevant to a two-sided test. Could you clarify how their method addresses this issue?

This test is based on the computation of the distribution of the difference between classification accuracies across resamplings. This is similar to the computation of the confidence interval of a  difference. Thus, we assess whether the theoretical zero value (= no difference; = null hypothesis) is outside the 2.5 and 97.5 percentile interval of the computed distribution of the empirically observed differences. We clarified now in the Methods (page 41) that for a two-tailed test the computed p-value (the percentile of the zero value) should be smaller than 0.025.

(6) Page 43, Lines 829-834: The manuscript explains: "The mean of 10 classification accuracies (i.e., of 10 resamplings) was employed to obtain a distribution (n=100) of the differences in classification accuracy ... The reported standard deviations of the classification accuracies are computed using also the means of 10 resamplings." I am unfamiliar with this type of analysis and am unclear about the rationale for calculating distributions and standard deviations based on the means of 10 resamplings rather than using the original distribution of classification accuracies. This resampling procedure appears to yield a narrower distribution and smaller standard deviations than the original data. Could you please justify this approach?

The logic of the analysis is to reduce the noise in the data, by averaging across 10 randomly selected resamplings, but still keeping a sufficient number of data (100 values) for a test.

Reviewer #3 (Recommendations for the authors):

(1) Some sentences are too long and difficult to parse. For example, in line 177: "the correlations between the responses to the 64 head-body orientation conditions of the two centerings for the neuron and pose combinations showing significant head-body interactions for the two centerings were similar to those observed for the whole population."

We have modified this sentence: For neuron and pose combinations with significant head-body interactions in both centerings, the correlations between responses to the 64 head-body orientation conditions were similar to those observed in the whole population.

(2) The authors argue in line 485: "in our study, a search bias cannot explain the body-inversion effect since we selected responsive units using both upright and inverted images." However, the body-selective patches were localized using upright images, correct?

The monkey-selective patches were localized using upright images indeed. However, we recorded in experiment 3 (and 2) also outside the localized patches (as we noted before in the Methods:  “In experiments 2 and 3 we recorded from a wider region, which overlapped with the two monkey patches and the recording locations of experiment 1”). Furthermore, the preference for upright monkey images is not an all-or-nothing phenomenon: most units still responded to inverted monkeys. Also, we believe it is likely that the mean responses to the inverted bodies in the monkey patches, defined by upright bodies versus objects, would be larger than those to objects and we would be surprised to learn that there is a patch selective for inverted bodies that we would have missed with our localizer.

(3) Typo: line 447, "this independent"->"is independent"?

Corrected.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

Summary:

Kv2 subfamily potassium channels contribute to delayed rectifier currents in virtually all mammalian neurons and are encoded by two distinct types of subunits: Kv2 alpha subunits that have the capacity to form homomeric channels (Kv2.1 and Kv2.2), and KvS or silent subunits (Kv5,6,8.9) that can assemble with Kv2.1 or Kv2.2 to form heteromeric channels with novel biophysical properties. Many neurons express both types of subunits and therefore have the capacity to make both homomeric Kv2 channels and heteromeric Kv2/KvS channels. Determining the contributions of each of these channel types to native potassium currents has been very difficult because the differences in biophysical properties are modest and there are no Kv2/KvS-specific pharmacological tools. The authors set out to design a strategy to separate Kv2 and Kv2/KvS currents in native neurons based on their observation that Kv2/KvS channels have little sensitivity to the Kv2 pore blocker RY785 but are blocked by the Kv2 VSD blocker GxTx. They clearly demonstrate that Kv2/KvS currents can be differentiated from Kv2 currents in native neurons using a two-step strategy to first selectively block Kv2 with RY785, and then block both with GxTx. The manuscript is beautifully written; takes a very complex problem and strategy and breaks it down so both channel experts and the broad neuroscience community can understand it.

Strengths:

The compounds the authors use are highly selective and unlikely to have significant confounding cross-reactivity to other channel types. The authors provide strong evidence that all Kv2/KvS channels are resistant to RY785. This is a strength of the strategy - it can likely identify Kv2/KvS channels containing any of the 10 mammalian KvS subunits and thus be used as a general reagent on all types of neurons. The limitation then of course is that it can't differentiate the subtypes, but at this stage, the field really just needs to know how much Kv2/KvS channels contribute to native currents and this strategy provides a sound way to do so.

Weaknesses:

The authors are very clear about the limitations of their strategy, the most important of which is that they can't differentiate different subunit combinations of Kv2/KvS heteromers. This study is meant to be a start to understanding the roles of Kv2/KvS channels in vivo. As such, this is a minor weakness, far outweighed by the potential of the strategy to move the field through a roadblock that has existed since its inception.

The study accomplishes exactly what it set out to do: provide a means to determine the relative contributions of homomeric Kv2 and heteromeric Kv2/KvS channels to native delayed rectifier K+ currents in neurons. It also does a fabulous job laying out the case for why this is important to do.

Reviewer #2 (Public Review):

Summary:

Silent Kv subunits and the channels containing these Kv subunits (Kv2/KvS heteromers) are in the process of discovery. It is believed that these channels fine-tune the voltage-activated K+ currents that repolarize the membrane potential during action potentials, with a direct effect on cell excitability, mostly by determining action potentials firing frequency.

Strengths:

What makes silent Kv subunits even more important is that, by being expressed in specific tissues and cell types, different silent Kv subunits may have the ability to fine-tune the delayed rectifying voltage-activated K+ currents that are one of the currents that crucially determine cell excitability in these cells. The present manuscript introduces a pharmacological method to dissect the voltage-activated K+ currents mediated by Kv2/KvS heteromers as a means of starting to unveil their importance, together with Kv2-only channels, to the cells where they are expressed.

Weaknesses:

While the method is effective in quantifying these currents in any isolated cell under an electric voltage clamp, it is ineffective as a modulating maneuver to perhaps address these currents in an in vivo experimental setting. This is an important point but is not a claim made by the authors.

We agree. We have now stated in the introduction that this study does not address the roles of Kv2/KvS currents in an in vivo setting.

Manuscript revisions:

While this study does not address the impact of GxTX or RY785 on action potentials or in vivo, the distinct pharmacology of Kv2/KvS heteromers presented here suggests that KvS conductances could be targeted to selectively modulate discrete subsets of cell types.  

There are other caveats with the methods and data:

(i) The need for a 'cocktail' of blockers to supposedly isolate Kv2 homomers and Kv2/KvS heteromers' currents from others may introduce errors in the quantification Kv2/KvS heteromers-mediated K+ currents and that is due to possible blockers off targets.

We now point out that is possible that off target effects of blockers may introduce errors, include references that identify the selectivity of the blockers used in the cocktail, and specifically note that 4-aminopyridine in the cocktail is expected to block 2% of Kv2 homomers yet have a lesser impact Kv2/KvS heteromers. Additionally, to test whether the KvS isolation strategy requires the cocktail in neurons, we performed new experiments on a different subclass of nociceptors without the blocker cocktail and identified a substantial KvS-like component (new Fig 7 Supplement 3).

Manuscript revisions:

“After whole-cell voltage clamp was established, non-Kv2/KvS conductances were suppressed by changing to an external solution containing a cocktail of inhibitors: 100 nM alpha-dendrotoxin (Alomone) to block Kv1 (Harvey and Robertson, 2004), 3 μM AmmTX3 (Alomone) to block Kv4 (Maffie et al., 2013; Pathak et al., 2016), 100 μM 4-aminopyridine to block Kv3 (Coetzee et al., 1999; Gutman et al., 2005), 1 μM TTX to block TTX sensitive Nav channels, and 10 μM A803467 (Tocris) to block Nav1.8 (Jarvis et al., 2007). It is possible that off target effects of blockers may introduce errors in the quantification Kv2/KvS heteromer-mediated K⁺ currents. For example, 4-aminopyridine is expected to block a small fraction, 2%, of Kv2 homomers and have a lesser impact on Kv2/KvS heteromers (Post et al., 1996; Thorneloe and Nelson, 2003; Stas et al., 2015) which could result in a slight overestimation of the ratio of Kv2/KvS heteromers to Kv2 homomers.”

“We also tested the other major mouse C-fiber nociceptor population, peptidergic nociceptors, to determine if this subpopulation also has conductances resistant to RY785 yet sensitive to GxTX. We voltage clamped DRG neurons from a CGRP^GFP mouse line that expresses GFP in peptidergic nociceptors (Gong et al., 2003). Deep sequencing has identified mRNA transcripts for Kv6.2, Kv6.3, Kv8.1 and Kv9.3 present in GFP+ neurons, an overlapping but distinct set of KvS subunits from the Mrgprd^GFP non-peptidergic population (Zheng et al., 2019). In GFP+ neurons from CGRP^GFP mice, we found that a fraction of outward current was inhibited by 1 µM RY785 and additional current inhibited by 100 nM GxTX (Fig 7 Supplement 3 A-C). In these experiments, 58 ± 2% (mean ± SEM) was KvS-like (Fig 7 Supplement 3 D) identifying that KvSlike conductances are present in these peptidergic nociceptors. For CGRP^GFP neurons we did not include the Kv1, Kv3, Kv4, Nav and Cav channel inhibitor cocktail used for other neuron experiments, indicating that the cocktail of inhibitors is not required to identify KvS-like conductances.”

(ii) During the electrophysiology experiments, the authors use a holding potential that is not as negative as it is needed for the recording of the full population of the Kv2/KvS channels. Depolarized holding potentials lead to a certain level of inactivation of the channels, that vary according to the KvS involved/present in that specific population of channels. As a reminder, some KvS promote inactivation and others prevent inactivation. Therefore, the data must be interpreted as such.

We agree. We now point out that the physiological holding potentials used are insufficiently negative to relieve inactivation from all Kv2/KvS heteromeric channels. We also note that the ratio of Kv2-like to KvS-like conductance is expected to vary with voltage protocols.

Manuscript revisions:

“Neurons were held at a membrane potential of –74 mV to mimic a physiological resting potential. KvS subunits can profoundly shift the voltage-inactivation relation (Salinas et al., 1997a; Kramer et al., 1998; Kerschensteiner and Stocker, 1999) and this potential is likely insufficiently negative to relieve inactivation from all Kv2/KvS heteromeric channels. Also, the activation membrane potential is close to the half-maximal point of Kv2/KvS conductances. Thus the ratio of Kv2-like to KvS-like conductance is expected to vary with voltage protocols.”

(iii) The analysis of conductance activation by using tail currents is only accurate when dealing with non-inactivating conductances. Also, in dealing with a heterogenous population of Kv2/KvS heteromers, heterogenous K+ conductance deactivation kinetics is a must. Indeed, different KvS may significantly relate to different deactivation kinetics as well.

We now discuss that the bi-exponential fit of tail currents is likely inadequate to capture the deactivation kinetics of all underlying components of a heterogenous population of Kv2/KvS heteromers.

Manuscript revisions:

“We note that the analysis of conductance activation by using tail currents is only accurate when dealing with non-inactivating conductances. We expect that inactivation of Kv2/KvS conductances during the 200 ms pre-pulse is minimal (Salinas et al., 1997a; Kramer et al., 1998; Kerschensteiner and Stocker, 1999) and did not notice inactivation during the activation pulse. Also, deactivation kinetics can vary in a heterogenous population of Kv2/KvS heteromers. While analysis of tail currents could skew the quantification of total Kv2 like and KvS-like conductances, our data supports that mouse nociceptors and human neurons have tail currents that are resistant to RY785 and sensitive to GxTX consistent with the presence of Kv2/KvS heteromers.”

(iv) Silent Kv subunits may be retained in the ER, in heterologous systems like CHO cells. This aspect may subestimate their expression in these systems. Nevertheless, the authors show similar data in CHO cells and in primary neurons.

We agree. We now note that in heterologous systems, including CHO cells, transfection of KvS subunits can result in KvS subunits that are retained intracellularly.

Manuscript revisions:

“While a fraction of KvS subunits appear to be retained intracellularly, immunofluorescence for Kv5.1, Kv9.3 and Kv2.1 also appeared localized to the perimeter of transfected Kv2.1-CHO cells (Figure 1 Supplement).”

(v) The hallmark of silent Kv subunits is their effect on the time inactivation of K+ currents. As such, data should be shown throughout, preferably, from this perspective, but it was only done so in Figure 4G.

Indeed, effects on inactivation are a hallmark of KvS subunits. However, quantifying inactivation of Kv2/KvS channels requires steps to positive voltages for approximately 10 seconds. In neurons steps this long usually resulted in irreversible changes in leak currents/input resistance that degraded the accuracy of RY785/GxTX subtraction currents. Consequently, we did not acquire inactivation data in neurons, and we now explain in the manuscript why such data was not obtained.

Manuscript revisions:

“While changes in inactivation are prominent with KvS subunits, we did not investigate inactivation in neurons because the lengthy depolarizations required often resulted in irreversible leak current increases that degraded the accuracy of RY785/GxTX subtraction current quantification.”

(vi) Functional characterization of currents only, as suggested by the authors as a bona fide of Kv2 and Kv2/KvS currents, should not be solely trusted to classify the currents and their channel mediators.

We agree, and now state explicitly that functional characterization cannot be trusted to classify their channel mediators of conductances, and we try to be clear about this throughout the manuscript by using soft terms such as "KvS-like" when identity is uncertain.

Manuscript revisions:

“As functional characterization alone cannot be trusted to classify their channel mediators of conductances, we define conductances consistent with Kv2/KvS heteromers as 'KvS-like' and conductances consistent with Kv2 homomers as 'Kv2-like'.”

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

There is not a lot to do here - this was a real pleasure to read and very easy to understand, as written. Here are a few minor things to consider:

(1) The naming of the KvS subunits has always been confusing - it is not clear that Kv5,6,8,9 are members of the Kv2 subfamily from the names. KvS does a good job of differentiating them by assembly phenotype and has been used a lot in the literature, but it doesn't solve the misconception of what subfamily they belong to. This might not matter so much for mammals, where all KvS channels are in the Kv2 subfamily, but it makes it impossible to extend the naming system to other animals where subunits requiring heteromeric assembly are common in most subfamilies. How about trying the name Kv2S? It would have continuity with KvS in the reader's mind, make it clear that they are Kv2 subfamily, and make a naming system that could be extended beyond vertebrates. This is not a problem the authors created - just a completely optional suggestion on how to solve it if so inclined.

We agree that naming conventions for these subunits are problematic, and agonized quite a bit about nomenclature. In the end we chose to stick with the precedent of KvS.

(2) Another naming issue they should definitely change is the use of "subfamily" for the different KvS subtypes (Kv5, Kv6, Kv8, and Kv9). This really creates confusion with the higher-order subfamilies that have a very clear functional definition: a subfamily of Kv genes is a group of related genes that have assembly compatibility. Those are Kv1, Kv2, Kv3 and Kv4. KvS genes are assembly compatible with Kv2, evolutionarily derived from the Kv2 lineage, and thus clearly a part of the Kv2 subfamily. Using a subfamily for the next lower level of the naming hierarchy confuses this. The authors should use different terms like sub-type or class or subgroups for the divisions within KvS.

Thank you. We have standardized to Kv2/KvS as a subfamily; Kv5, Kv6, Kv8, and Kv9 as subtypes; and individual proteins, e.g. Kv8.1, as subunits.

(3) When you discuss whether the KvS subunit directly disrupts Ry785 binding in the pore or works allosterically and you said you know which KvS residues point into the pore from models, I thought that maybe you could tell from a sequence alignment whether the KvS channels you didn't test look the same in the conduction pathway as the ones you did test. If so, you could mention that if the binding site is the pore, they should all be resistant. Alternatively, if one you didn't test looks fundamentally more similar to the Kv2s in this region, then maybe it could be fingered as a possible exception that needs to be tested later.

Great ideas. We now assess sequence KvS variability near the proposed RY785 binding site in all KvS subunits. We generated structural models of RY785 docking to Kv2.1 and Kv2.1/Kv8.1 and found that residues near RY785 are different in all KvS subunits.

Manuscript revisions:

“We analyzed computational structural models of RY785 docked to a Kv2.1 homomer and a 3:1 Kv2.1:Kv8.1 heteromer (Fig 9) to gain structural insight into how KvS subunits might interfere with RY785 binding. We used Rosetta to dock RY785 to a cryo-EM structure of a Kv2.1 homomer in an apparently open state (Fernández-Mariño et al., 2023). The top-scoring docking pose has RY785 positioned below the selectivity filter and off-axis of the pore (Fig 9 A), similar to a stable pose observed in molecular dynamic simulations (Zhang et al., 2024). In this pose, RY785 contacts a collection of Kv2.1 residues that vary in every KvS subtype (Fig 9 B,D,E). Notably, RY785 bound similarly to a 3:1 model of Kv2.1/Kv8.1, in contact with the three Kv2.1 subunits, yet avoided the Kv8.1 subunit (Fig 9C). This is consistent with RY785 binding less well to Kv2.1/Kv8.1 heteromers, and also suggests that a 3:1 Kv2:KvS channel could retain a RY785 binding site when open.”

(4) Future suggestion or tip - not for this paper. Your data shows your isolation strategy works really well on Kv6 channels, and these are also the Kv2/KvS channels that have the most pronounced biophysical changes. Working on neurons that have a prominent Kv2/Kv6 component would really show how well the strategy outlined here works to describe the physiology of native neurons. The highest KvS expression I have seen in public data in a wellstudied cell type is Kv6.4 in spinal motor neurons.

Wonderful tip, thank you. We are indeed very interested in Kv6.4 in spinal motor neurons.

Reviewer #2 (Recommendations For The Authors):

The manuscript makes a good contribution to the identification of Kv2/KvS channels in primary cells. The pharmacological method proposed by the authors to dissect the currents in an experimental setting seems proper. Although meritorious in themselves, the findings are heavily phenomenological in the opinion of this reviewer. The manuscript should be improved with some level of mechanistic data and/or the demonstration of different levels of expression in different cell types.

Thank you for the suggestions. This manuscript now demonstrates strikingly higher levels of the KvS-like component of Kv2 currents in somatosensory (DRG nonpeptidergic and peptidergic nociceptor) versus autonomic (SCG) neuron types. The mechanistic question of what electrophysiological properties the KvS subunits are providing to the neuronal circuit is an exciting one that we are pursuing separately.

Manuscript revisions:

“While we found only RY785-sensitive Kv2-like conductances in SCG neurons, Kv2/KvS heteromer-like conductances were dominant in DRG neurons.”

At present, the manuscript says that the combination of RY785 and guangxitoxin-1E can be used to define Kv2/KvS-mediated K+ currents. Importantly, this method cannot be used in a way that one can functionally determine the function of Kv2/KvS channels, since it depends on the pre-blocking of Kv2-mediated K+ currents prior. In the opinion of this reviewer, this fact decreases the attention of a potential reader.

Indeed, our study is focused on revealing KvS heteromers by voltage clamp, and we now clarify in the introduction that we do not determine the function of Kv2/KvS channels in this study, so as not to lead the reader to expect studies of neuronal signaling.

However, the selective pharmacology we identify suggests RY785 application could reveal the function of Kv2 homomers, and for RY785-insensitive signaling, GxTX application of could reveal the function of Kv2/KvS heteromers. We now mention these possible applications in the Discussion.

Manuscript revisions:

“While this study does not address the impact of GxTX or RY785 on action potentials or in vivo, the distinct pharmacology of Kv2/KvS heteromers presented here suggests that KvS conductances could be targeted to selectively modulate discrete subsets of cell types.”

Please find below suggestions for improving the manuscript:

(1) The term "Kv2/KvS heteromers" should be used throughout instead of variations such as "Kv2/KvS channels", "Kv2/KvS" and others. Standardization of the term to refer to heteromers would make the manuscript easier to read.

Thank you. We have standardized terms to consistently refer to Kv2/KvS heteromers.

(2) Confusing terms like KvS conductances, KvS-like conductances, KvS-like (RY785-resistant, GxTX-sensitive) currents, and KvS channels should be avoided because they disregard the current belief that KvS cannot form functional homomeric channels. The term KvS-containing channels, and Kv2/KvS channels, seem more accurate. Uniformization in this regard will also make the manuscript more easily readable.

Thank you. We have standardized terms to Kv2/KvS heteromers and KvS-containing channels when channel subunits are known and the use terms KvS-like and Kv2-like for functionally identified endogenous conductances with unknown channel subunits.

(3) Referring to KvS as a regulatory subunit is inaccurate. It is clear that KvS is part of, and it makes up the alpha pore. KvS therefore is a part of the conductive pathway and not a regulatory (suggesting accessory) subunit. KvS take part in selectivity filter (fully conserved), but they also make up an important part of the conducting pathway with non-conserved amino acid residues.

We felt it important to include the descriptor “regulatory” to connect our nomenclature with prior use of the descriptor in the literature, and now only use the term at the start of the introduction.

Manuscript revisions:

“A potential source of molecular diversity for Kv2 channels are a group of Kv2-related proteins which have been referred to as regulatory, silent, or KvS subunits.”

(4) The use of a cocktail of channel inhibitors may affect the quantification of Kv2/KvS heteromers-mediated K+ currents because they may interact with RY785 and/or GxTx or they may even interact with the sites for these two drugs on Kv2-containing channels.

This is an interesting point worth considering, thank you. We now alert readers to this possibility in the discussion when considering the limitations of our approach.

Manuscript revisions:

“Also, the cocktail of inhibitors used in most neuron experiments here could potentially alter RY785 or GxTX action against KvS/Kv2 channels.”

(5) The graphical representation of fractional blocking and other parameters (e.g., Fig 1D), is hard to read in these slim plots. In my opinion, tall bars would be more meaningfully visualized.

Thank you for pointing out that the graphs were hard to read, we have made the graph easier to read and added tall bars.

(6) Vehicle control for IHC and electrophysiology. Please state what is the vehicle used in the electrophysiology experiments.

Thank you. The composition of vehicle has now been stated in the methods.

Manuscript revisions:

“All RY785 solutions contained 0.1% DMSO. Vehicle control solutions also contained 0.1% DMSO but lacked RY785.”

“Sections were incubated in vehicle solution (4% milk, 0.2% triton diluted in PB) for 1 hr at RT.”

(7) The reference Trapani & Korn, 2003 (?) is not included in the list. This reference is important since it sets what are the Kv2.1-CHO cells. In this regard it is also important to mention, even better to address, the expressing qualities of this system in the face of a co-expression with a plasmid-based expression of silent Kv subunits. Are these two ways of expressing Kv subunits, meant to come together (or not) in heteromers, balanced? This question is critical here. Still, in regard to Kv2.1-CHO cells, it was not clear in the manuscript if the term "transfection" refers only to the plasmids used to temporarily induce the expression of silent Kv subunits and potentially Kv channels accessory subunits.

We now include the Trapani & Korn, 2003 reference (thank you for pointing out this accidental omission), and better explain expression methods. The benefit of the inducible Kv2.1 expression is control of Kv conductance densities which can otherwise become so large as to be refractory to voltage clamp. The beauty of the expression system is that cells recently transfected with KvS subunits can be induced to express just enough Kv2.1 to get a substantial but not clampoverwhelming RY785-resistant Kv2/KvS conductance. We also discuss that our expression methods are distinct from past studies. We stop short of comparing the expression systems, as this is beyond the scope of what we set out to study.

Manuscript revisions: See next response

(8) Kv2.1-CHO cells transfection procedures, induction, and validation are unclear. This validation is important here.

We have clarified transfection procedures, induction, and validation in the methods section.

Manuscript revisions:

“The CHO-K1 cell line transfected with a tetracycline-inducible rat Kv2.1 construct (Kv2.1-CHO) (Trapani and Korn, 2003) was cultured as described previously (Tilley et al., 2014).”

Transfections were achieved with Lipofectamine 3000 (Life Technologies, L3000001). 1 μl Lipofectamine was diluted, mixed, and incubated in 25 μl of Opti-MEM (Gibco, 31985062).”

“Concurrently, 0.5 μg of KvS or AMIGO1 or Navβ2, 0.5 μg of pEGFP, 2 μl of P3000 reagent and 25 μl of Opti-MEM were mixed. DNA and Lipofectamine 3000 mixtures were mixed and incubated at room temperature for 15 min. This transfection cocktail was added to 1 ml of culture media in a 24 well cell culture dish containing Kv2.1-CHO cells and incubated at 37 °C in 5% CO2 for 6 h before the media was replaced. Immediately after media was replaced, Kv2.1 expression was induced in Kv2.1-CHO cells with 1 μg/ml minocycline (Enzo Life Sciences, ALX380-109-M050), prepared in 70% ethanol at 2 mg/ml. Voltage clamp recordings were performed 12-24 hours later. We note that the expression method of Kv2/KvS heteromers used here is distinct from previous studies which show that the KvS:Kv2 mRNA ratio can affect the expression of functional Kv2/KvS heteromers (Salinas et al., 1997b; Pisupati et al., 2018). We validated the functional Kv2/KvS heteromer expression using voltage clamp to establish distinct channel kinetics and the presence of RY785-resistant conductance in KvS-transfected cells and using immunohistochemistry to label apparent surface localization of KvS subunits (Figure 4, Figure 1 Supplement, Figure 1 and Figure 5).”

(9) It is important for readers to add some context to Kv2.1/Kv8.1 channels (and other Kv2/KvS heteromers) used to test the combination of RY785 and GxTx. In my opinion, this enriches the discussion.

Good idea. We have added context about each of the KvS subunits we test.

Manuscript revisions:

“To test the pharmacological response of KvS we began with Kv8.1, a subunit that creates heteromers with biophysical properties distinct from Kv2 homomers (Salinas et al., 1997a), and modulates motor neuron vulnerability to cell death (Huang et al., 2024).

Each of these KvS subunits create Kv2/KvS heteromers that have distinct biophysical properties (Kramer et al., 1998; Kerschensteiner and Stocker, 1999; Bocksteins et al., 2012). Kv5.1/Kv2.1 heteromers play an important role in controlling the excitability of mouse urinary bladder smooth muscle (Malysz and Petkov, 2020), mutations in Kv6.4 have been shown to influence human labor pain (Lee et al., 2020b), and deficiency of Kv9.3 disrupts parvalbumin interneuron physiology in mouse prefrontal cortex (Miyamae et al., 2021).”

(10) In general, the membrane potential used to activate Kv2 only channels and Kv2/KvS channels is too close to the activation V1/2. In case the comparing curves are displaced in their relative voltage dependence and voltage sensitivity, using that range of membrane potential may introduce a crucial error in the estimation of the conductance's relative amplitudes.

We now note that the relative conductances of Kv2-only vs Kv2/KvS channels are expected to vary with voltage protocol, as KvS inclusion results in channels with altered voltage responses.

Manuscript revisions:

“…the activation membrane potential is close to the half-maximal point of Kv2/KvS conductances. Thus the ratio of Kv2-like to KvS-like conductance is expected to vary with voltage protocols.”

(11) The use of tail currents to estimate conductance is problematic if i) lack of current inactivation is not assured, and ii) if the different currents, with possible different deactivation kinetics at the used membrane potential (e.g., mV), are not assured. Why was the activation peak used at times, and at different elapsed times the tail currents were used instead? These aspects of conductance's amplitude estimation methods should be well defined.

In CHO cells peak currents were analyzed because outward currents seem to offer the best signal/noise. In neurons, we restricted analysis to tail currents at elapsed times to minimize complications from non-Kv2 endogenous voltage-gated channels which deactivate more quickly. We have clarified this analysis in the methods section.

Manuscript revisions:

“In CHO cells peak currents were analyzed because outward currents seem to offer the best signal/noise. In neurons, we restricted analysis to tail currents at elapsed times to minimize complications from non-Kv2 endogenous voltage-gated channels which deactivate more quickly. In neurons, voltage gated currents remained in the toxin cocktail + RY785 and GxTX, that were sometimes unstable. To minimize complications from these currents, we restricted analysis of RY785 and GxTX subtraction experiments to tail currents at elapsed times to minimize complications from non-Kv2 endogenous voltage-gated channels which deactivate more quickly. We note that the analysis of conductance activation by using tail currents is only accurate when dealing with non-inactivating conductances. We expect that inactivation of Kv2/KvS conductances during the 200 ms pre-pulse is minimal (Salinas et al., 1997a; Kramer et al., 1998; Kerschensteiner and Stocker, 1999) and did not notice inactivation during the activation pulse. Also, deactivation kinetics can vary in a heterogenous population of Kv2/KvS heteromers. While analysis of tail currents could skew the quantification of total Kv2 like and KvS-like conductances, our data supports that mouse nociceptors and human neurons have tail currents that are resistant to RY785 and sensitive to GxTX consistent with the presence of Kv2/KvS heteromers.”

(12) Were the experiments including different conditions such as control, RY, and RY+GxTx done pair-wised? This could potentially better the statistics and strengthen the data and the conclusions drawn from them.

The control, RY, and RY+GxTX in neurons were done pairwise and the statistical tests performed for these experiments were pairwise tests. We have clarified this in the figure legends.

Manuscript revisions:

“Wilcoxon rank tests were paired, except the comparison of RY785 to vehicle which was unpaired.”

(13) The holding potential of the experiments, mostly -89 mV, may be biasing the estimation of Kv2 only channels vs. Kv2/KvS channels conductances. Figure 4I exemplifies this concern.

We agree. Figure 4I reveals that a holding potential of -89 mV vs -129 mV reduces conductance of Kv2.1/Kv8.1 heteromers vs Kv2.1 homomers in CHO cells by ~20%. We have now alerted readers that the ratio of Kv2 only channels vs. Kv2/KvS conductances can vary with holding voltage.

Manuscript revisions:

“Under these conditions, 58 ± 3 % (mean ± SEM) of the delayed rectifier conductance was resistant to RY785 yet sensitive to GxTX (KvS-like) (Fig 7 F). We note that the ratio of KvS- to Kv2-like conductances is expected to vary with holding potential, as KvS subunits can change the degree and voltage-dependence of steady state inactivation (e.g. Fig 4I).”

(14) It is possible that Figure 6A (control trace) and Figure 6C ("Kv2-like" trace) are the same, by mistake, since their noise pattern looks too similar.

Indeed the noise pattern of the Figure 6A (control trace) and Figure 6C ("Kv2-like" trace) are related, as they have inputs from the same trace, with Figure 6C ("Kv2-like" trace) being a subtraction of Figure 6A (+RY trace) from Figure 6A (control trace).

(15) For example, in Figure 7A, what is the identity of the current remaining after the RY+GxTx application? In Figure 7B, a supposed outlier in the group of data referring to "veh" in the right panel is what possibly is making this group different from +RY in the left panel (p=0.02, Wilcoxon rank test). I would recommend parametric tests only since the data is essentially quantitative.

In Figure 7A, we do not know the identity of the current remaining after the RY+GxTX application, the kinetics of the residual current appeared distinct from the Kv2/KvS-like currents blocked by RY or GxTX, but we did not analyze these.

The date in Figure 7B, was indeed the positive outlier in the group of data referring to "veh" in the right panel and contributes to the p-value, but we saw no reason to exclude it. We have now replaced the representative trace in 7B with a non-outlier trace. We respectfully disagree with the suggestion to use parametric statistical tests as we do not know the distribution underlying the variance our data.

Manuscript revisions:

“Subsequent application of 100 nM GxTX decreased tail currents by 68 ± 5% (mean ± SEM) of their original amplitude before RY785. We do not know the identity of the outward current that remains in the cocktail of inhibitors + RY785 + GxTX.”

(16) Please state the importance of using nonpeptidergic neurons to study silent Kv5.1 and Kv9.1 subunits. RNA data may not necessarily work to probe function or protein abundance, which is crucial in heteromeric complexes.

We have now more thoroughly explained our rationale for choosing the nonpeptidergic neurons.

RNA is not predictive of protein abundance, and we have not yet been successful in measuring KvS protein abundance in these neurons, so we've probed KvS abundance by assessing RY785 resistance.

Manuscript revisions:

“Mouse dorsal root ganglion (DRG) somatosensory neurons express Kv2 proteins (Stewart et al., 2024), have GxTX-sensitive conductances (Zheng et al., 2019), and express a variety of KvS transcripts (Bocksteins et al., 2009; Zheng et al., 2019), yet transcript abundance does not necessarily correlate with functional protein abundance. To record from a consistent subpopulation of mouse somatosensory neurons which has been shown to contain GxTXsensitive currents and have abundant expression of KvS mRNA transcripts (Zheng et al., 2019), we used a Mrgprd^GFP transgenic mouse line which expresses GFP in nonpeptidergic nociceptors (Zylka et al., 2005; Zheng et al., 2019). Deep sequencing identified that mRNA transcripts for Kv5.1, Kv6.2, Kv6.3, and Kv9.1 are present in GFP+ neurons of this mouse line (Zheng et al., 2019) and we confirmed the presence of Kv5.1 and Kv9.1 transcripts in GFP+ neurons from Mrgprd^GFP mice using RNAscope (Fig 7 Supplement 1).”

(17) In Figure 8B, were +RY data different from veh data? The figure shows no Wilcoxon (nonparametric) comparison and this is important to be stated. What conductance(s) is the vehicle solution blocking or promoting? What is RY dissolved in, DMSO? What is the DMSO final concentration?

We now state that in Figure 8B, +RY amplitudes were not statistically different from veh data in this limited data set. However, the RY-subtraction currents always had Kv2-like biophysical properties, whereas vehicle-subtraction currents had variable properties precluding biophysical analysis for Fig 8D.

In Figure 8B, we do not know what conductance(s) the vehicle solution is affecting, we think the changes observed are likely merely time dependent or due to the solution exchange itself. RY stock is in DMSO. All recording solutions have 0.1% DMSO final concentration, this is now noted in methods.

Manuscript revisions:

“Unlike mouse neurons, we did not detect a significant difference in tail currents of RY785 versus vehicle controls. However, RY785-subtracted currents always had Kv2-like biophysical properties whereas vehicle-subtraction currents had variable properties that precluded the same biophysical analysis. Overall, these results show that human DRG neurons can produce endogenous voltage-gated currents with pharmacology and gating consistent with Kv2/KvS heteromeric channels.”

“All RY785 solutions contained 0.1% DMSO. Vehicle control solutions also contained 0.1% DMSO but lacked RY785.”

(18) METHODS. The electrophysiology approach should be unified in all aspects as applicable and possible.

We have unified the mouse dorsal root ganglion and mouse superior cervical ganglion methods sections. We have kept CHO cells and mouse/human neurons section separate because the methods were substantially different.

(19) DISCUSSION. The discussion section spends half of its space trying to elaborate on possible blocking/inhibiting/modulating mechanisms for RY785. The present manuscript shows no data, at least not that I have noticed, that would evoke such discussion.

We have shortened this section, and enhance the discussion with structural models (new Fig 9), and our functional data indicating perturbed RY785 interaction with Kv2.1/8.1.

Manuscript revisions:

“In this pose, RY785 contacts a collection of Kv2.1 residues that vary in every KvS subtype (Fig 9 B,D,E). Notably, RY785 bound similarly to a 3:1 model of Kv2.1/Kv8.1, in contact with the three Kv2.1 subunits, yet avoided the Kv8.1 subunit (Fig 9C). This is consistent with RY785 binding less well to Kv2.1/Kv8.1 heteromers, and also suggests that a 3:1 Kv2:KvS channel could retain a RY785 binding site when open. However, the RY785 resistance of Kv2/KvS heteromers may primarily arise from perturbed interactions with the constricted central cavity of closed channels. In homomeric Kv2.1, RY785 becomes trapped in closed channels and prevents their voltage sensors from fully activating, indicating that RY785 must interact differently with closed channels (Marquis and Sack, 2022). Here we found that Kv2.1/Kv8.1 current rapidly recovers following washout of RY785, suggesting that Kv2.1/Kv8.1 heteromers do not readily trap RY785 (Figure 2 Supplement). Overall, the structural modeling suggests that KvS subunits sterically interfere with RY785 binding to the central cavity, while functional data suggest KvS subunits disrupt RY785 trapping in closed states.”

(20) DISCUSSION. Topics like ER retention and release upon certain conditions would be a better enrichment for the manuscript in my opinion.

ER retention of KvS subunits is indeed an important topic! However, we have opted not to delve into it here.

(21) DISCUSSION. Speculation about the binding site for RY on Kv2/KvS channels is also not touched by the data shown in the manuscript.

We have shortened this section of discussion, and now present this with structural models of RY785 docked to a Kv2.1 homomer and 3:1 Kv2.1: Kv8.1 heteromer (new Fig 9) to ground speculations. See manuscript changes noted in response to comment (19) above.

(22) DISCUSSION. An important reference is missing in regard to stoichiometry: Bocksteins et al., 2017. This work is the only one using a non-optical technique to add knowledge to that question.

Good point, and an excellent study we didn’t realize we’d not included before. We now include Bocksteins et al. 2017 as a reference in the Introduction.

(23) In my opinion, allosterism and orthosterism are concepts not yet useful for the discussion of RY binding sites without even a general piece of data.

We now include structural models of RY785 docked to a Kv2.1 homomer and 3:1 Kv2.1: Kv8.1 heteromer (new Fig 9) to ground blocking speculations. See manuscript changes noted in response to comment (19).

(24) The term "homogeneously susceptible" associated with a Hill slope close to 1 needs to be more elaborated.

Thank you, we have elaborated.

Manuscript revisions:

“Also, the degree of resistance to RY785 may vary if Kv2:KvS subunit stoichiometry varies. With high doses of RY785, we found that the concentration-response characteristics of Kv2.1/Kv8.1 in CHO cells revealed hallmarks of a homogenous channel population with a Hill slope close to 1 (Fig 2B). However, other KvS subunits might assemble in multiple stoichiometries and result in pharmacologically-distinct heteromer populations.”

(25) Stating the KvS are resistant to RY785 is not proper in my opinion. This opinion relates to the fact that the RY binding site in the channels is certainly not restricted to a binding site residing only on the Kv subunit.

Good point. We have now changed phrasing to convey that KvS subunits are a component of a heteromer that imbues RY785 resistance.

Manuscript revisions:

“These results show that voltage-gated outward currents in cells transfected with members from each KvS subtype have decreased sensitivity to RY785 but remain sensitive to GxTX. While we did not test every KvS subunit, the ubiquitous resistance suggests that all KvS subunits may provide resistance to 1 μM RY785 yet remain sensitive to GxTX, and that RY785 resistance is a hallmark of KvS-containing channels.”

Author response:

The following is the authors’ response to the previous reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

The authors investigate the role of the melanocortin system in puberty onset. They conclude that POMC neurons within the arcuate nucleus of the hypothalamus provide important but differing input to kisspeptin neurons in the arcuate or rostral hypothalamus.

Strengths:

Innovative and novel

Technically sound

Well-designed

Thorough

Weaknesses:

There were no major weaknesses identified.

Reviewer #2 (Public review):

Summary:

This interesting manuscript describes a study investigating the role of MC4R signalling on kisspeptin neurons. The initial question is a good one. Infertility associated with MC4 mutations in humans has typically been ascribed to the consequent obesity and impaired metabolic regulation. Whether there is a direct role for MC4 in regulating the HPG axis has not been thoroughly examined. Here, the researchers have assembled an elegant combination of targetted loss of function and gain of function in vivo experiments, specifically targetting MC4 expression in kisspeptin neurons. This excellent experimental design should provide compelling evidence for whether melanocortin signalling dirently affects arcuate kisspeptin neurons to support normal reproductive function. There were definite effects on reproductive function (irregular estrous cycle, reduced magnitude of LH surge induced by exogenous estradiol). However, the magnitude of these responses and the overall effect on fertility were relatively minor. The mice lacking MC4R in kisspeptin neurons remained fertile despite these irregularities. The second part of the manuscript describes a series of electrophysiological studies evaluating the pharmacological effects of melanocortin signalling in kisspeptin cells in ex-vivo brain slides. These studies characterised interesting differential actions of melanocortins in two different populations of kisspeptin neurons. Collectively, the study provides some novel insights into how direct actions of melanocortin signalling via the MC4 receptor in kisspeptin neurons contribute to the metabolic regulation of the reproductive system. Importantly, however, it is clear that other mechanisms are also at play.

Strengths:

The loss of function/gain of function experiments provides a conceptually simple but hugely informative experimental design. This is the key strength of the current paper - especially the knock-in study that showed improved reproductive function even in the presence of ongoing obesity. This is a very convincing result that documents that reproductive deficits in MC4R knockout animals (and humans with deleterious MC4R gene variants) can be ascribed to impaired signalling in the hypothalamic kisspeptin neurons and not necessarily caused as a consequence of obesity. As concluded by the authors: "reproductive impairments observed in MC4R deficient mice, which replicate many of the conditions described in humans, are largely mediated by the direct action of melanocortins via MC4R on Kiss1 neurons and not to their obese phenotype." This is important, as it might change how such fertility problems are treated.

I would like to see the validation experiments for the genetic manipulation studies given greater prominence in the manuscript because they are critical to interpretation. Presently, only single unquantified images are shown, and a much more comprehensive analysis should be provided.

Weaknesses:

(1) Given that mice lacking MC4R in kisspeptin neurons remained fertile despite some reproductive irregularities, this can be described as a contributing pathway, but other mechanisms must also be involved in conveying metabolic information to the reproductive system. This is now appropriately covered in the discussion.

(2) The mechanistic studies evaluating melanocortin signalling in kisspeptin neurons were all completed in ovariectomised animals (with and without exogenous hormones) that do not experience cyclical hormone changes. Such cyclical changes are fundamental to how these neurons function in vivo and may dynamically alter how they respond to hormones and neuropeptides. Eliminating this variable makes interpretation difficult, but the authors have justified this as a reductionist approach to evaluate estradiol actions specifically. However, this does not reflect the actual complexity of reproductive function.

For example, the authors focus on a reduced LH response to exogenous estradiol in ovariectomised mice as evidence that there might be a sub-optimal preovulatory LH surge. However, the preovulatory LH sure (in intact animals) was not measured.

They have not assessed why some follicles ovulated, but most did not. They have focused on the possibility that the ovulation signal (LH surge) was insufficient rather than asking why some follicles responded and others did not. This suggests some issue with follicular development, likely due to changes in gonadotropin secretion during the cycle and not simply due to an insufficient LH surge.

Reviewer #3 (Public review):

The manuscript by Talbi R et al. generated transgenic mice to assess the reproduction function of MC4R in Kiss1 neurons in vivo and used electrophysiology to test how MC4R activation regulated Kiss1 neuronal firing in ARH and AVPV/PeN. This timely study is highly significant in neuroendocrinology research for the following reasons.

(1) The authors' findings are significant in the field of reproductive research. Despite the known presence of MC4R signaling in Kiss1 neurons, the exact mechanisms of how MC4R signaling regulates different Kiss1 neuronal populations in the context of sex hormone fluctuations are not entirely understood. The authors reported that knocking out Mc4r from Kiss1 neurons replicates the reproductive impairment of MC4RKO mice, and Mc4r expression in Kiss1 neurons in the MC4R null background partially restored the reproductive impairment. MC4R activation excites Kiss1 ARH neurons and inhibits Kiss1 AVPV/PeN neurons (except for elevated estradiol).

(2) Reproduction dysfunction is one of obesity comorbidities. MC4R loss-of-function mutations cause obesity phenotype and impaired reproduction. However, it is hard to determine the causality. The authors carefully measured the body weight of the different mouse models (Figure 1C, Figure 2A, Figure 3B). For example, the Kiss1-MC4RKO females showed no body weight difference at puberty onset. This clearly demonstrated the direct function of MC4R signaling in reproduction but was not a consequence of excessive adiposity.

(3) Gene expression findings in the "KNDy" system align with the reproduction phenotype.

(4) The electrophysiology results reported in this manuscript are innovative and provide more details of MC4R activation and Kiss1 neuronal activation.

Overall, the authors have presented sufficient background in a clear, logical, and organized structure, clearly stated the key question to be addressed, used the appropriate methodology, produced significant and innovative main findings, and made a justified conclusion.

Comments on revisions:

The authors have addressed my comments.

Recommendations for the authors:

The reviewers noted that they received comments in response to their concerns, and some improvements have been made to the manuscript. However, as described below, in some cases, a rebuttal was provided, but changes were not made to the manuscript. It is suggested that these issues be addressed to improve the quality of the manuscript.

We thank the reviewers and editor for the assessment of the manuscript and recommendations for its improvement. We have addressed the remaining comments from reviewer #2 below, and hope that they find our revisions satisfactory.

Reviewer #2 (Recommendations for the authors):

The manuscript convincingly shows that MC4R in kisspeptin-producing cells can influence reproductive function. This suggests that fertility problems associated with melanocortin mutations are likely due to direct effects on the reproductive systems rather than simply being side effects of the resultant obesity.

We are pleased that this reviewer finds the data convincing and thank them for the careful review of the manuscript, which has helped to improve its published version.

The authors have responded to the reviewer's comments and made several improvements to the manuscript.

The authors are correct in pointing out that the POMC-Cre animals should be fine for studies involving the administration of AAVs to adult animals. I have misinterpreted how these mice were being used, and this concern is fully addressed.

Unfortunately, in some cases, the authors rebutted the reviewer's comments but did not change the manuscript. I suggest addressing several issues in the manuscript (after all, it is not the reviewer's opinion that counts; this process is about improving the manuscript).

(1) Validation of the KO is insufficiently reported. From the methods, it appears that this was done thoroughly, but currently, only a single image of the arcuate nucleus is shown, and no image of the AVPV is shown. There is no quantitative information provided. The authors can keep these data as supplementary material, but they should be comprehensive and convincing, as so much depends on the degree of knockout in this model. One cannot assume complete KO based simply on the relevant genetics, as there are examples in this system where different Cre lines produce different outcomes with various floxed genes in the two major populations of kisspeptin neurons. This figure should show the quantitation of the RNAscope analysis from each of the two regions regarding the percentage of kisspeptin cells showing expression of MC4R mRNA. In addition, the lack of MC4 labelling in the arcuate nucleus, outside of kisspeptin neurons, is a concern. One would expect to see AgRP or POMC cells at this level, but are they still showing expression of MC4? A single image is insufficient to be convinced of the model's efficacy.

We appreciate the reviewer’s concerns regarding the validation of the MC4RKO model. Below, we provide clarification and additional justification for our approach.

(1) Quantification of MC4R in the Arcuate Nucleus (ARC): As noted by the reviewer, we were unable to detect sufficient MC4R signal in the ARC of KO mice to perform meaningful quantification. This is consistent with the expected outcome of a successful MC4R deletion. Given the low endogenous expression levels of MC4R in this region, even in control animals, and the technical limitations of RNAscope in detecting very low-abundance transcripts, especially for receptors, the absence of MC4R signal in the ARC of KO mice strongly supports effective deletion. Moreover, the MC4R loxP mouse has been published and validated by many labs including Brad Lowell’s lab who’s done extensive work using these mice for selective deletion of Mc4r from various neuronal populations such as Sim1 and Vglut2 neurons (Shah et al., 2014, de Souza Cordeiro et al., 2020). To further strengthen our validation, we provide additional images from another animal (Fig_S1) to illustrate the consistency of the MC4R KO in the ARC. These will be included as supplementary material, as suggested.Regarding AgRP and POMC neurons, MC4R is not highly expressed in these neurons (as per previous literature, e.g., Garfield et al., Nat Neurosci. 2015; Padilla SL et al, Endocrinology 2012; Henry et al, Nature, 2015). Instead, MC4R is predominantly found in downstream neurons in the paraventricular nucleus (PVN) and other hypothalamic regions (which is intact in our KO mice as shown in our validation figure). Thus, the absence of MC4R labeling in AgRP or POMC cells in our images aligns with known expression patterns and does not contradict the validity of our model.

(2) MC4R Expression in the AVPV and OVX Effect on Kiss1 Expression: We acknowledge the reviewer’s request for MC4R expression analysis in the anteroventral periventricular nucleus (AVPV). However, due to the timing of tissue collection after ovariectomy (OVX), Kiss1 expression in the AVPV is significantly suppressed, making it technically unfeasible to perform co-staining of MC4R with Kiss1 in this region. This is a well-documented effect of estrogen depletion following OVX (Smith et al., 2005; Lehman et al., 2010). While we acknowledge that an ideal validation would include AVPV co-labeling, the experimental constraints related to OVX preclude this analysis in our dataset.

Given these considerations and validations, we are confident that the KO is effective and specific.

(2) Line 88: "... however, conflicting reports exist". Expand on this sentence to describe what these conflicting reports show. The authors responded to my comment but made no changes to the introduction. As a reader, I dislike being told there are conflicting reports, but then I have to go and look up the reference to see what that actual point of conflict is.

By conflicting reports we meant that other studies have shown no association between MC4R and reproductive disorders, this has now been included in the revised manuscript (Line 89).

(3) Could the authors explain how a decrease in AgRP would be interpreted as a "decrease in hypothalamic melanocortin tone" in line 142 and line 364? These overly simplistic interpretations of qPCR data detract from the overall quality of the paper.

The reference to a decrease in melanocortin tone referred to the decrease in the expression of melanocortin receptor signaling, this has been clarified in the revised manuscript (lines 142 and 360).

(4) Please show the individual cycle patterns for all animals, as in Figure 2B. This can be a supplemental figure, but the current bar charts are not informative.

We respectfully disagree that the bar charts are not informative as they include the critical statistical analysis. We have now included all individual estrous cycle data in new separate supplemental figure (Sup. Figure 3). Therefore, we have excluded the representative cycles from the main figures as they are now in the new Supplemental. We have changed the orders of the figures in the text accordingly.

(5) In their rebuttal, the authors state: "Mice lack true follicular and luteal phases, and therefore, it is impossible to separate estrogen-mediated changes from progesterone-mediated changes (e.g., in a proestrous female). Therefore, we use an ovariectomized female model in which we can generate an LH surge with an E2-replacement regimen [1]. This model enables us to focus on estrogen effects, exclude progesterone effects, and minimize variability. Inclusion of cycling females would make interpretation much more difficult." I disagree, but the authors can take this position if they wish. However, they should not report the responses to exogenous estradiol in an ovariectomised mouse as a "preovulatory LH surge" (line 380). An ovariectomised mouse cannot ovulate, and the estrogen-induced LH surge is significantly different in magnitude and timing from the endogenous preovulatory LH surge (likely due to the actions of progesterone). One goal of these studies is to understand why the ovulation rate appears to be low in the MC4-KO animals. Hence, evaluating whether the preovulatory LH surge is typical is important. This has not been done. The authors have shown that the response to exogenous estradiol is sub-normal. Such an effect might lead to a reduced preovulatory LH surge, but this has not been measured.

We appreciate this reviewer’s concern about the nature of the preovulatory LH surge. We have clarified this in the revised manuscript and described it as “an induced LH surge” throughout the text (Lines 163, 533, 6560).

(6) I believe that the ovulation process should be considered "all or none," and I do not quite understand the rebuttal discussion. The authors describe that "numerous follicles mature at the same time....". That is not disputed. My point was that each mature follicle will receive the identical endocrine ovulatory signal (correct? Or do the authors believe something different?). If it were sufficient for one follicle to ovulate, then all of those mature follicles (the number of which will be variable between animals and between cycles) would be expected to undergo ovulation. The fact that they do not raise several possibilities. One that the authors favor is that an insufficient ovulatory signal might approach a threshold where some follicles ovulate and others do not. This possibility is supported by the apparent increase in cystic follicles, which might be preovulatory follicles that did not complete the ovulation process. Such variation might be stochastic, within normal variation for sensitivity to LH. However, it is also possible that the follicles have not matured at the same rate, perhaps influenced by abnormal secretion of LH or FSH during earlier phases of the cycle, and hence are not in the appropriate condition to respond to the ovulation signal when it arrives. Some may even have matured prematurely due to the elevated gonadotropins reported in this study. Given the data and the partial fertility, the most likely explanation is that the genetic manipulation has resulted in fewer follicles being available for ovulation due to changes in follicular development rather than a deficit of the ovulation signal, although the latter mechanism might also contribute. A third possibility is that genetic manipulation has directly affected the ovary. The authors did not answer whether Kiss1 and MC4 are co-expressed in the ovary. I think the authors might want to rule this out by showing no change in MC4R expression in the ovary.

We thank the reviewer for this thoughtful comment and agree that these are possible outcomes. We have now acknowledged them in the Discussion.

To answer the reviewer’s question, we have not investigated the co-expression of Kiss1 and Mc4r in the ovary. While MC4R has indeed been documented in the ovary (Chen et al. Reproduction, 2017), the changes in gonadotropin release and supporting in vitro data included in this manuscript clearly document a central effect, however, an additional effect at the level of the ovary cannot be completely ruled out. This has now been added to the discussion (Line 378-387).

(7) Lines 390, 454 " impaired LH pulse" What was the evidence for impaired LH pulse (see figure 2D)?

Thank you for pointing this out. This comment referred to augmented LH release. This has been corrected in the revised manuscript (Line 394).

The paper's strengths remain, as outlined in my original review. The authors have addressed what I perceived to be weaknesses, predominantly by changing the tone of discussion and interpretation of the data. This is appropriate. I consider the focus on the LH surge as the primary mechanism too narrow, and the authors should be considering how other changes during the cycle might influence ovarian function.

We sincerely appreciate the reviewer’s thoughtful evaluation of our manuscript and their constructive feedback. We are pleased that our revisions have addressed the perceived weaknesses and that the adjustments to the discussion and interpretation were deemed appropriate.

We acknowledge the reviewer’s perspective on broadening the discussion beyond the LH surge to consider additional cycle-dependent influences on ovarian function. While our current study focuses on this specific mechanism, we recognize that ovarian function is influenced by multiple physiological changes throughout the cycle. We have refined our discussion to reflect this broader context and appreciate the suggestion to consider these additional factors in future studies.

We have addressed all of the reviewer’s comments to the best of our ability and hope they find the revised manuscript satisfactory.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

The authors track the motion of multiple consortia of Multicellular Magnetotactic Bacteria moving through an artificial network of pores and report a discovery of a simple strategy for such consortia to move fast through the network: an optimum drift speed is attained for consortia that swim a distance comparable to the pore size in the time it takes to align the with an external magnetic field. The authors rationalize their observations using dimensional analysis and numerical simulations. Finally, they argue that the proposed strategy could generalize to other species by demonstrating the positive correlation between the swimming speed and alignment time based on parameters derived from literature.

Strengths:

The underlying dimensional analysis and model convincingly rationalize the experimental observation of an optimal drift velocity: the optimum balances the competition between the trapping in pores at large magnetic fields and random pore exploration for weak magnetic fields.

Weaknesses:

The convex pore geometry studied here creates convex traps for cells, which I expect enhances their trapping. The more natural concave geometries, resulting from random packing of spheres, would create no such traps. In this case, whether a non-monotonic dependence of the drift velocity on the Scattering number would persist is unclear.

We agree that convex walls increase the time that consortia remain trapped in pores at high magnetic fields. Since the non-monotonic behavior of the drift velocity with the Scattering number arises largely due to these long trapping times, we agree that experiments using concave pores are likely to show a peak drift velocity that is diminished or erased.

However, we disagree that a random packing of spheres or similar particles provides an appropriate model for natural sediment, which is not composed exclusively of hard particles in a pure fluid. Pore geometry is also influenced by clogging. Biofilms growing within a network of convex pillars in two-dimensional microfluidic devices have been observed to connect neighboring pillars, thereby forming convex pores. Similar pore structures appear in simulations of biofilm growth between spherical particles in three dimensions. Moreover, the salt marsh sediment in which MMB live is more complex than simple sand grains, as cohesive organic particles are abundant. Experiments in microfluidic channels show that cohesive particles clog narrow passageways and form pores similar to those analyzed here. Thus, we expect convex pores to be present and even common in natural sediment where clogging plays a role.

The concentration of convex pores in the experiments presented here is almost certainly much higher than in nature. Nonetheless, since magnetotactic bacteria continuously swim through the pore space, they are likely to regularly encounter such convexities. Efficient navigation of the pore space thus requires that magnetotactic bacteria be able to escape these traps. In the original version of this manuscript, this reasoning was reduced to only one or two sentences. That was a mistake, and we thank the reviewer for prompting us to expand on this point. As the reviewer notes, this reasoning is central to the analysis and should have been featured more prominently. In the final version, we will devote considerable space to this hypothesis and provide references to support the claims made above.

The reviewer suggests that the generality of this work depends on our finding a ”positive correlation between the swimming speed and alignment [rate] based on parameters derived from literature.” We wish to emphasize that, in addition to predicting this correlation, our theory also predicts the function that describes it. The black line in Figure 3 is not fitted to the parameters found in the literature review; it is a pure prediction.

Reviewer #2 (Public review):

The authors have made microfluidic arrays of pores and obstacles with a complex shape and studied the swimming of multicellular magnetotactic bacteria through this system. They provide a comprehensive discussion of the relevant parameters of this system and identify one dimensionless parameter, which they call the scattering number and which depends on the swimming speed and magnetic moment of the bacteria as well as the magnetic field and the size of the pores, as the most relevant. They measure the effective speed through the array of pores and obstacles as a function of that parameter, both in their microfluidic experiments and in simulations, and find an optimal scattering number, which they estimate to reflect the parameters of the studied multicellular bacteria in their natural environment. They finally use this knowledge to compare different species to test the generality of this idea.

Strengths:

This is a beautiful experimental approach and the observation of an optimal scattering number (likely reflecting an optimal magnetic moment) is very convincing. The results here improve on similar previous work in two respects: On the one hand, the tracking of bacteria does not have the limitations of previous work, and on the other hand, the effective motility is quantified. Both features are enabled by choices of the experimental system: the use the multicellular bacteria which are larger than the usual single-celled magnetotactic bacteria and the design of the obstacle array which allows the quantification of transition rates due to the regular organization as well as the controlled release of bacteria into this array through a clever mechanism.

Weaknesses:

Some of the reported results are not as new as the authors suggest, specifically trapping by obstacles and the detrimental effect of a strong magnetic field have been reported before as has the hypothesis that the magnetic moment may be optimized for swimming in a sediment environment where there is a competition of directed swimming and trapping. Other than that, some of the key experimental choices on which the strength of the approach is based also come at a price and impose some limitations, namely the use of a non-culturable organism and the regular, somewhat unrealistic artificial obstacle array.

In the “Recommendations for the Authors,” this reviewer drew our attention to a manuscript that absolutely should have been prominently cited. As the reviewer notes, our manuscript meaningfully expands upon this work. We are pleased to learn that the phenomena discussed here are more general than we initially understood. It was an oversight not to have found this paper earlier. The final version will better contextualize our work and give due credit to the authors. We sincerely appreciate the reviewer for bringing this work to our attention.

We disagree that the use of non-culturable organisms and our unrealistic array should be considered serious weaknesses. While any methodological choice comes with trade-offs, we believe these choices best advance our aims. First, the goal of our research, both within and beyond this manuscript, is to understand the phenotypes of magnetotactic bacteria in nature. While using pure cultures enables many useful techniques, phenotypic traits may drift as strains undergo domestication. We therefore prioritize studying environmental enrichments.

Clearly, an array of obstacles does not fully represent natural heterogeneity. However, using regular pore shapes allows us to average over enough consortium-wall collisions to enable a parameter-free comparison between theory and experiment. Conducting an analysis like this with randomly arranged obstacles would require averaging over an ensemble of random environments, which is practically challenging given the experimental constraints. Since we find good agreement between theory and experiment in simple geometries, we are now in a position to justify extending our theory to more realistic geometries. Additionally, we note that a microfluidic device composed of a random arrangement of obstacles would also be a poor representation of environmental heterogeneity, as pore shape and network topology differ between two and three dimensions.

Recommendations for the Authors: 

Reviewer #1 (Recommendations for the authors):

My main suggestion is for the authors to describe the limitations of their approach in the case of concave pores.

As we noted in our public comments, this was a very useful comment to hear from you and one that has been repeated as we have spoken about these results to colleagues. Convexities here represent an experimentally simple way to force bacteria to back track through the maze, as they must through natural sediment. We have greatly expanded this discussion to clarify this reasoning (lines 84–105). We provide references to three types of physical processes that may lead to such traps. First, as in figure 1 of Kurz et al, biofilm (white) can fill the spaces between convex pillars to create covexities. Additionally, clogging by cohesive particles can make narrow passageways between convex particles impassible. An example of clogging is shown in figure 6 of Dressaire & Sauret 2017. Finally, air bubbles trapped in the sediment can create pore-scale dead ends that require bacteria to backtrack. The full references are provided in the main text.

Small points:

(1) How many trajectories were used to produce Figures 2 b and c?

We have modified the caption to note that these data represent the measured transition rates of a total 938 consortia at various Scattering numbers. Each consortium may pass between pores many times.

(2) Can the authors describe in more detail how Equation (3) is derived? Why doesn’t it depend on the gap size between the pores?

We have provided a derivation of this equation in Appendix 2 of the new version. This derivation shows that the drift velocity U_drift is proportional to the pore diameter and difference between the transition rates.

The proportionality constant α depends on how the pores are connected together in space. In the original version, we wanted to highlight the role of the asymmetry of the transition rates, so we imagined a one dimensional network of pores without gaps. In this case, α \= 1. This reasoning was poorly explained in the previous version and we thank the reviewer for pointing this deficiency out. In the new version, we include the gap size and use the layout of pores in a square lattice with gaps, which is shown in figure 1. The proportionality constant for a square lattice in the absence of gaps√ would be 1/2. The limitations of photolithography require some gap that increase the proportionality constant to α \= 0.8344.

We have updated the text, equation (3), and the figures to account for the finite gap sizes.

(3) I found the second part of the abstract, related to the comparison between diverse bacteria, to be slightly misleading. Upon first reading, my expectation was that the authors carried out experiments with different species.

We have modified the abstract to make clear that we rely on values taken from a literature review.

(4) More information is needed on how many trajectories were used to produce the probability densities in Figures 1b-d. How were the densities computed?

The probability distributions give the probability that a pixel in a pore is covered by a consortium. They reflect between 1.2 and 7 million measurements (depending on the panel) of the instantaneous positions of consortia. We have added a section (Lines 453–469) to Materials and Methods that describes exactly how these distributions were calculated.

Reviewer #2 (Recommendations for the authors):

(1) As mentioned under Weaknesses in the Public review, some results are less new than claimed here. The existence of an optimal magnetic moment has been shown by Codutti et al eLife eLife13:RP98001 in very similar experiments, where it was also proposed that this may be an evolutionary adaptation to the sediment habitat. The paper here provides additional evidence for this, and with better tracking and quantification, but previous work should be discussed. Likewise, the work by Dekharghani et al. that is mentioned rather suddenly in the Results section appears to be a crucial previous state of the art and could already be mentioned in the introduction.

We thank the reviewer for bringing this paper, which came out as we were writing this manuscript, to our attention. The hypothesis that there is an optimal phenotype that balances magnetotaxis with obstacle avoidance—and that natural selection could guide organisms to this optimum—goes back to at least 2022. It seems that Codutti et al independently came up with this same hypothesis and provided the first test.

We have substantively rewritten the introduction (Lines 46–58) to better contextualize our work and give due attention to Dekharghani et al.

(2) The first paragraph of Results also contains background information and could be moved into the introduction.

As part of the rewrite to better contextualize our work, we moved the first two paragraphs of results to the introduction.

(3) I found Figure 1 a bit confusing and it took me some time to understand the geometry. I think the black obstacles are very dominant to the viewer’s eye and draw attention away from the essentially circular shape of the pores. Likewise, I am not sure that cutting the neighboring pores off in a circular fashion in Figures 1b-d was the best choice. The authors should think about whether the presentation can be improved. Likewise, when describing the direction of the field in the text, I would suggest adding that it is along the horizontal direction in Figure 1.

We have modified the figure and the text as the reviewer suggests.

(4) That collisions with a pore wall are an important mechanism of changing direction is clear and it is nice to see the paper demonstrate that this mechanism is dominant over rotational diffusion. However, this may not be universal, as (i) rotational diffusion is more important for smaller cells and (ii) interaction with walls can result in all kinds of different behaviors than complete randomization (e.g. swimming along the walls as shown in microfluidic chambers, Ostapenko et al. Phys Rev Lett 2018, Codutti et al. eLife 2022, or reversals, Kuhn et al PNAS 2017). Here, it appears that complete randomization of the direction is an assumption, but this could be tested/quantified by analyzing the trajectories.

This is an excellent point. We have modified the text to describe qualitatively how these tendencies would shift the Critical Scattering number. We also note in the text that there is evidence of these differences in Fig 3. The Desulfobacterota are shifted upwards in Fig 3 relative to the α-proteobacteria. This shift indicates that Desulfobacterota tend to live at slightly greater scattering numbers of 0.9±0.3 than the α-proteobacteria, which live at scattering number 0.37 ± 0.03. It is likely that this difference reflects taxonomic differences in rotational diffusion and cell-wall interactions.

It is true that total randomization of the direction is indeed an assumption, and it is stated as such in line 189. We performed all of the numerics to find the solid curves in Fig 2 before we got any experimental data and so, at the time, total randomization seemed like a fair choice. Looking at Fig 2b, it is clear that these numerics systematically overestimate k₋. We believe that this error is do to the assumption of total randomization.

As this effect is small and does not change any of the conclusions of the paper and Codutti et al were able to publish their paper in the time that we were writing ours, we feel some urgency to move forward.

(5) From the manuscript it is not fully clear to what extent experiments and simulations are or can be quantitatively compared. For example: is the curve (“fit”) in Figure 2c based on the simulations? Is there an explicit expression or is this just a spline or something like that? Why does Figure 5 (simulation) show the velocity as a function of Sc⁻¹and Figure 2 (experiment) as a function of Sc? It looks to me as if a quantitative comparison could be achieved.

The original version of Figure 2 shows a quantitative comparison between theory and experiment with no fit parameters. The data points are the result of experiments in which consortia are tracked as they as they move between connected pores. The solid line is a found by interpolating a smooth curve through the data from simulations. As we make clear in the new version (Lines 537–551), this blue curve is the most probable smooth curve that explains the simulations.

We have added the simulations to figure 2 so that a single panel includes the data, the simulations, and the smooth curve. To further make clear that this comparison is quantitative and parameter free, we have added a panel to Figure 2. This panel directly compares the prediction to observation and is independent of the blue curve.

As was noted (deep within the methods section) in the original version, our numerics can exactly simulate Sc = ∞. Consequently, it was reasonable to simulate parameters that are uniformly spaced in Sc⁻¹.

(6) While I like the idea behind Figure 3, the data shown here is not as convincing as suggested. If one looks at the data without the black line, I think one gets a weaker dependence. The correlation between U₀ and γ_geo is likely not as strong as it seems. Calculating a correlation coefficient might be helpful here. In any case, the assumptions going into this figure should be discussed more explicitly and the results should in my opinion be phrased more cautiously (I tend to believe what the authors claim, but I don’t think the evidence for this point is very strong).

We appreciate the reviewer’s skepticism. However, we believe that the data are stronger than one might understand from the previous text. We have rewritten the text (Lines 219–291) and included new analysis, figures, and explanation to make three points clear.

(a) It is surprising that speed, magnetic moment, and mobility all vary tremendously(between one and three orders of magnitude) across taxa and environment, however, their dimensionless combination Sc is narrowly distributed. We have added a panel to Fig. 3 to show the measured Scattering numbers.

It is notable that there are no adjusted parameters in the calculation of the Scattering numbers: it is a simple dimensionless combination of phenotypic and environmental parameters. All but one of these parameters (the pore size) is measured either by us or by other authors. The pore radius is likely narrowly distributed. We measure it at our field site and, when it is not reported, we use a value typical of the geological and fluvial environment. Just as the size of sand grains does not vary greatly between the beaches of Australia, Africa, and California, it is a good assumption that the pore spaces that host these magnetotactic bacteria do not vary tremendously in size.

(b) In the new version we compare the Scattering number statistics to a parameterfree null model of phenotypic diversity. We argue in the text that it is appropriate to bootstrap over the phenotypic diversity of species. This null model provides the correct method to calculate p-values as the variability in the Scattering numbers is neither identically distributed nor normally distributed.

We use this null model to show that—given the measured phenotypic diversity across species—the probability that fifteen random species would fall within the measured range of Scattering numbers that is consistent with optimal navigation is ∼ 10⁻⁶. This result is strong evidence that the phenotypic variables exhibit the correlations that are predicted by our analysis.

(c) The correlation between U₀/r and γ_geo is reasonably strong. I think that our choice of axes in Fig 3, which were chosen to fit the legend, make the data look flatter than then they actually are. Here are the same data plotted without the line with tighter axes:

Author response image 1.

With the exception of the very first point and the very last point, the data appear to our eyes to be pretty correlated. This impression is born out by a calculation of the correlation coefficient which gives 0.77. The p-value is 4 × 10⁻⁴. We have included these values in the main text to clarify that this correlation is both statistically significant and of primary importance.

(7) There is a comment at the end of the discussion that the evolutionary hypothesis could be tested by transferring the magnetotaxis genes to nonmagnetotactic organisms. This would indeed be highly desirable, but this is very difficult as indicated by the successful efforts in that direction (which often are only moderately magnetic/magnetotactic), see Kolinko et al Nature Nanotech 2014, Dziuba et al Nature Nanotech 2024.

Thank you for highlighting these references, which we have included. We agree that these experiments will be challenging. Our results make a prediction about the evolution of these strains, so it seems worth mentioning this fact. We feel that this manuscript is not the correct space for a detailed description of challenges that we will encounter should we pursue this direction of study.

(8) A section on how the bacterial samples were obtained could be added in Methods.

We have done so.

Additional Changes

(1) In the original version, we feared that the consortia in the microfluidic device arepoorly representative of the natural population. Consequently, we used the values from previous experiments, which we performed using consortia taken from the same pond. Since submitting this manuscript we have undertaken new experiments that allowed us to measure the Scattering number of individual consortia. It turns out the effect is smaller than we worried. We have included these measurements in the new version. We find that even as the most common phenotypes vary over the course of time, the Scattering number remains constant. This result is additional evidence that there is strong selective pressure to optimally navigate.

As a result of these additions, we have added an author, Julia Hernandez, who contributed to these experiments and analysis.

(2) We have expanded the table of phenotypic variable in Appendix 1 to make it easier forother researchers to reproduce our analysis.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

This manuscript reports the investigation of PriC activity during DNA replication initiation in Escherichia coli. It is reported that PriC is necessary for the growth and control of DNA replication initiation under diverse conditions where helicase loading is perturbed at the chromosome origin oriC. A model is proposed where PriC loads helicase onto ssDNA at the open complex formed by DnaA at oriC. Reconstituted helicase loading assays in vitro support the model. The manuscript is well-written and has a logical narrative.

Thank you for understanding this study.

Major Questions/Comments:

An important observation here is that a ΔpriC mutant alone displays under-replication, suggesting that this helicase loading pathway is physiologically relevant. Has this PriC phenotype been reported previously? If not, would it be possible to confirm this result using an independent experimental approach (e.g. marker frequency analysis or fluorescent reporter-operator systems)?

We thank Reviewer 1 for this comment. This study provides the first direct evidence for PriC’s role in initiation of chromosome replication. Given the change of the oriC copy number of ∆priC cells in non-stressed conditions is only slight, resolution of the suggested methods is clearly not high enough to distinguish the differences in the oriC copy number between priC⁺ (WT) and ∆priC cells. Thus, to corroborate the ∆priC phenotype, we additionally analyzed using flow cytometry priC⁺ and ∆priC cells growing under various nutrition and thermal conditions.

As shown in Figure 2-figure supplement 1 of the revised version, the fraction of cells with non-2ⁿ oriC copies was slightly higher in ∆priC cells compared to priC⁺ cells. Furthermore, when grown in M9 minimal medium at 37˚C, ∆priC mutant cells exhibited slightly reduced ori/mass values. These are supportive to the idea that inhibition of replication initiation occurs at low frequency even in the WT dnaA and dnaC background, and that PriC function is necessary to ensure normal replication initiation. Related descriptions have been revised accordingly.

Is PriA necessary for the observed PriC activity at oriC? Is there evidence that PriC functions independently of PriA in vivo?

As described in Introduction of the original manuscript, PriA is a 3’-to-5’ helicase which specifically binds to the forked DNA with the 3’-end of the nascent DNA strand. Thus, structural specificity of target DNA is essentially different between PriA and PriC. Consistent with this, our in vitro data indicate that PriC alone is sufficient to rescue the abortive helicase loading at oriC (Figure 7), indicating that PriA is principally unnecessary for PriC activity at oriC. Consistently, as described in Introduction, PriC can interact with ssDNA to reload DnaB (Figure 1E). Nevertheless, a possibility that PriA might participate in the PriC-dependent DnaB loading rescue at oriC in vivo can not be completely excluded. However, elucidation of this possibility is clearly beyond the scope of the present study and should be analyzed in the future. An additional explanation has been included in Discussion of the revised version.

Is PriC helicase loading activity in vivo at the origin direct (the genetic analysis leaves other possibilities tenable)? Could PriC enrichment at oriC be detected using chromatin immunoprecipitation?

These are advanced questions about genomic dynamics of PriC. Given that PriC facilitates DnaB reloading at stalled replication forks (Figure 1E) (Heller and Marians, Mol Cell., 2005; Wessel et al., J Biol Chem, 2013; Wessel et al., J Biol Chem, 2016), PriC might interact with the whole genome and its localization might not necessarily exhibit a preference for oriC in growing cells. Analysis about these advanced questions is interesting but is beyond the scope of the present study and should be analyzed in the future study.

Reviewer #2 (Public review):

This is a great paper. Yoshida et al. convincingly show that DnaA does not exclusively do loading of the replicative helicase at the E. coli oriC, but that PriC can also perform this function. Importantly, PriC seems to contribute to helicase loading even in wt cells albeit to a much lesser degree than DnaA. On the other hand, PriC takes a larger role in helicase loading during aberrant initiation, i.e. when the origin sequence is truncated or when the properties of initiation proteins are suboptimal. Here highlighted by mutations in dnaA or dnaC.

This is a major finding because it clearly demonstrates that the two roles of DnaA in the initiation process can be separated into initially forming an open complex at the DUE region by binding/nucleation onto DnaA-boxes and second by loading of the helicase. Whereas these two functions are normally assumed to be coupled, the present data clearly show that they can be separated and that PriC can perform at least part of the helicase loading provided that an area of duplex opening is formed by DnaA. This puts into question the interpretation of a large body of previous work on mutagenesis of oriC and dnaA to find a minimal oriC/DnaA complex in many bacteria. In other words, mutants in which oriC is truncated/mutated may support the initiation of replication and cell viability only in the presence of PriC. Such mutants are capable of generating single-strand openings but may fail to load the helicase in the absence of PriC. Similarly, dnaA mutants may generate an aberrant complex on oriC that trigger strand opening but are incapable of loading DnaB unless PriC is present.

We would like to thank Revierwer#2 for the very positive comments about our work.

In the present work, the sequence of experiments presented is logical and the manuscript is clearly written and easy to follow. The very last part regarding PriC in cSDR replication does not add much to the story and may be omitted.

Given that the role PriC in stimulating cSDR was unclear, we believe that our finding that PriC has little or no role in cSDR, despite being a negative result, is valuable for the general readership of eLife. To further assess impact of PriC on cSDR and as recommended by Referee #1, we carried out the chromosome loci copy-number analysis by the whole-genome sequencing. As shown in Figure 8-supplement 1 of the revised version, the results support our conclusion from the original version.

Reviewer #3 (Public review):

Summary:

At the abandoned replication fork, loading of DnaB helicase requires assistance from PriABC, repA, and other protein partners, but it does not require replication initiator protein, DnaA. In contrast, nucleotide-dependent DnaA binding at the specific functional elements is fundamental for helicase loading, leading to the DUE region's opening. However, the authors questioned in this study that in case of impeding replication at the bacterial chromosomal origins, oriC, a strategy similar to an abandoned replication fork for loading DnaB via bypassing the DnaA interaction step could be functional. The study by Yoshida et al. suggests that PriC could promote DnaB helicase loading on the chromosomal oriC ssDNA without interacting with the DnaA protein. However, the conclusions drawn from the primarily qualitative data presented in the study could be slightly overwhelming and need supportive evidence.

Thank you for your understanding and careful comments.

Strengths:

Understanding the mechanism of how DNA replication restarts via reloading the replisomes onto abandoned DNA replication forks is crucial. Notably, this knowledge becomes crucial to understanding how bacterial cells maintain DNA replication from a stalled replication fork when challenging or non-permissive conditions prevail. This critical study combines experiments to address a fundamental question of how DnaB helicase loading could occur when replication initiation impedes at the chromosomal origin, leading to replication restart.

Thank you for your understanding.

Weaknesses:

The term colony formation used for a spotting assay could be misleading for apparent reasons. Both assess cell viability and growth; while colony formation is quantitative, spotting is qualitative. Particularly in this study, where differences appear minor but draw significant conclusions, the colony formation assays representing growth versus moderate or severe inhibition are a more precise measure of viability.

We used serial dilutions of the cell culture for the spotting assay and thus this assay should be referred as semi-quantitative rather than simply qualitative. For more quantitative assessment of viability, we analyzed the growth rates of cells and the chromosome replication activity using flow cytometry.

Figure 2

The reduced number of two oriC copies per cell in the dnaA46priC-deficient strain was considered moderate inhibition. When combined with the data suggested by the dnaAC2priC-deficient strain containing two origins in cells with or without PriC (indicating no inhibition)-the conclusion was drawn that PriC rescue blocked replication via assisting DnaC-dependent DnaB loading step at oriC ssDNA.

The results provided by Saifi B, Ferat JL. PLoS One. 2012;7(3):e33613 suggests the idea that in an asynchronous DnaA46 ts culture, the rate by which dividing cells start accumulating arrested replication forks might differ (indicated by the two subpopulations, one with single oriC and the other with two oriC). DnaA46 protein has significantly reduced ATP binding at 42C, and growing the strain at 42C for 40-80 minutes before releasing them at 30 C for 5 minutes has the probability that the two subpopulations may have differences in the active ATP-DnaA. The above could be why only 50% of cells contain two oriC. Releasing cells for more time before adding rifampicin and cephalexin could increase the number of cells with two oriCs. In contrast, DnaC2 cells have inactive helicase loader at 42 C but intact DnaA-ATP population (WT-DnaA at 42 or 30 C should not differ in ATP-binding). Once released at 30 C, the reduced but active DnaC population could assist in loading DnaB to DnaA, engaged in normal replication initiation, and thus should appear with two oriC in a PriC-independent manner.

This is a question about dnaA46 Δ_priC_ mutant cells. Inhibition of the replication forks causes inhibition of RIDA (the DNA-clamp complex-dependent DnaA-ATP hydrolysis) system, resulting in the increase of ATP-DnaA molecules (Kurokawa et al. (1999) EMBO J.). Thus, if Δ_priC_ inhibits the replication forks significantly, the ATP-DnaA level should increase and initiation should be stimulated. However, the results of Figure 2BC are opposite, indicating inhibition of initiation by Δ_priC_. Thus, we infer that the inhibition of initiation in the Δ_priC_ cells is not related to possible changes in the ATP-DnaA level. Even if the ATP-DnaA levels are different in subpopulations in dnaA46 cells, Δ_priC_ mutation should not affect the ATP-DnaA levels significantly. Thus, we infer that even in dnaA46 Δ_priC_ mutant cells, Δ_priC_ mutation directly affect initiation mechanisms, rather than indirectly through the ATP-DnaA levels.

Broadly, the evidence provided by the authors may support the primary hypothesis. Still, it could call for an alternative hypothesis: PriC involvement in stabilizing the DnaA-DnaB complex (this possibility could exist here). To prove that the conclusions made from the set of experiments in Figures 2 and 3, which laid the foundations for supporting the primary hypothesis, require insights using on/off rates of DnaB loading onto DnaA and the stability of the complexes in the presence or absence of PriC, I have a few other reasons to consider the latter arguments.

This is a very careful consideration. However, we infer that stabilization of the DnaA-DnaB interaction by PriC, even if present, does not always result in stimulation of DnaB loading to oriC. Given that interactions between DnaA and DnaB during DnaB loading to oriC are highly dynamic and complicated with multiple steps, stabilization of the DnaA-DnaB interaction by PriC, even if it occurs, has a considerable risk of inhibiting the DnaB loading by constructing abortive complexes. In addition, DnaA-DiaA binding is very tight and stable (Keyamura et al., 2007, 2009). Even if WT DnaA and WT DnaB are present, PriC can rescue the initiation defects of oriC mutants. Based on these facts and the known characteristics of PriC as explained in Introduction, it is more reasonable to infer that PriC provides a bypass of DnaB loading even at oriC, as proposed for the mechanism at the stalled replication fork. However, we cannot completely rule out the indicated possibility and these explanations are included in the revised version.

Figure 3

One should consider the fact that dnA46 is present in these cells. Overexpressing pdnaAFH could produce mixed multimers containing subunits of DnaA46 (reduced ATP binding) and DnaAFH (reduced DnaB binding). Both have intact DnaA-DnaA oligomerization ability. The cooperativity between the two functions by a subpopulation of two DnaA variants may compensate for the individual deficiencies, making a population of an active protein, which in the presence of PriC could lead to the promotion of the stable DnaA: DnaBC complexes, able to initiate replication. In the light of results presented in Hayashi et al. and J Biol Chem. 2020 Aug 7;295(32):11131-11143, where mutant DnaBL160A identified was shown to be impaired in DnaA binding but contained an active helicase function and still inhibited for growth; how one could explain the hypothesis presented in this manuscript. If PriC-assisted helicase loading could bypass DnaA interaction, then how growth inhibition in a strain carrying DnaBL160A should be described. However, seeing the results in light of the alternative possibility that PriC assists in stabilizing the DnaA: DnaBC complex is more compatible with the previously published data.

Unfortunately, in this comment, there is a crucial misunderstanding in the growth of cells bearing DnaA L160A. Hayashi et al. reported that the dnaB(Ts) cells bearing the dnaB L160A allele grew slowly and formed colonies even at 42°C. This feature is similar to the growth of dnaA46 cells bearing dnaA F46A H136A allele (Figure 2). Thus, the results of dnaB L160A cells are consistent with our model and support the idea that PriC partially rescues the growth inhibition of cells bearing the DnaB L160A allele by bypassing the strict requirement for the DnaA-DnaB interaction. Nevertheless, we have to be careful about a possibility that DnaB L160A could affect interaction with PriC, which we are going to investigate for a future paper.

As suggested, if mixed complexes of DnaA46 and DnaA F46A H136A proteins are formed, those might retain partial activities in oriC unwinding and DnaB interaction although those cells are inviable at 42°C without PriC. It is noteworthy that in the specific oriC mutants which are impaired in DnaB loading (e.g., Left-oriC), PriC effectively rescues the initiation and cell growth. In these cells, both DnaA and DnaB are intact. Thus, the idea that only mutant DnaA (or DnaB) protein is simulated specifically via PriC interaction is invalid. Even in cells bearing wild-type oriC, DnaA and DnaB, contribution of PriC for initiation is detected.

In addition, as described in the above response, given that interactions between DnaA and DnaB during DnaB loading to oriC are very dynamic and complicated with multiple steps, stabilization of the DnaA-DnaB interaction by PriC, even if present, would not simply result in stimulation of DnaB loading to oriC; rather we think a probability that it would inhibit the DnaB loading by constructing abortive complexes. Based on the known characteristics of PriC as explained in Introduction, it is more reasonable to infer that PriC provides a bypass of DnaB loading even at oriC, as proposed for the mechanism at the stalled replication fork.

However, we cannot completely rule out the indicated possibility and this explanation has been described in the revised version as noted in response to the above question.

Figure 4

Overexpression of DiaA could contribute to removing a higher number of DnaA populations. This could be more aggravated in the absence of PriC (DiaA could titrate out more DnaA)-the complex formed between DnaA: DnaBC is not stable, therefore reduced DUE opening and replication initiation leading to growth inhibition (Fig. 4A ∆priC-pNA135). Figure 7C: Again, in the absence of PriC, the reduced stability of DnaA: DnaBC complex leaves more DnaA to titrate out by DiaA, and thus less Form I*. However, adding PriC stabilizes the DnaA: DnaBC hetero-complexes, with reduced DnaA titration by DiaA, producing additional Form I*. Adding a panel with DnaBL160A that does not interact with DnaA but contains helicase activity could be helpful. Would the inclusion of PriC increase the ability of mutant helicase to produce additional Form I*?

Unfortunately, the proposed idea is biased disregarding the fact that DiaA effectively stimulates assembling processes of DnaA molecules at oriC. As oriC contains multiple DnaA boxes and multiple DnaA molecules are recruited there, DiaA will efficiently facilitate assembling of DnaA molecules on oriC. Even DnaA molecules of DnaA-DiaA complexes can efficiently bind to oriC. This is consistent with in vitro experiments showing that higher levels of DiaA stimulate assembly of DnaA molecules and oriC unwinding (i.e., DUE opening) but even excessive levels of DiaA do not inhibit those reactions (Keyamura et al., J. Biol. Chem. (2009) 284, 25038-25050). However, as shown in Figure 9, DiaA tightly binds to the specific site of DnaA which is the same as the DnaB L160-binding site, which causes inhibition of DnaA-DnaB binding (ibid). These are consistent with in vivo experiments, and concordantly consistent with the idea that the excessive DiaA level inhibits interaction and loading of DnaB by the DnaA-oriC complexes, but not oriC unwinding (i.e., DUE opening) in vivo. Also, as mentioned above, we do not consider that stabilization of DnaA-DnaBC complex simply results in stimulation of DnaB loading to oriC. Based on the known characteristics of PriC, it is more reasonable to infer that PriC provides a bypass of DnaB loading even at oriC, as proposed for the mechanism at the stalled replication fork (Figure 1E), as described in the above response.

As for DnaB L160A, as mentioned above, we are currently investigating interaction modes between DnaB and PriC. While investigating DnaB L160A could further support our model, we believe its contribution to the present manuscript would be incremental. In addition, there is a possibility that DnaA L160A could affect interaction with PriC. Thus, analysis of DnaB mutants in this PriC rescue mechanisms should be addressed in future study.

Figure 5

The interpretation is that colony formation of the Left-oriC ∆priC double mutant was markedly compromised at 37˚C (Figure 5B), and 256 the growth defects of the Left-oriC mutant at 25{degree sign}C and 30{degree sign}C were aggravated. However, prima facia, the relative differences in the growth of cells containing and lacking PriC are similar. Quantitative colony-forming data is required to claim these results. Otherwise, it is slightly confusing.

The indicated concern was raised due to our typing error lacking ∆priC. In the revised manuscript, we have amended as follows: the cell growth of the Left-oriC ∆priC double mutant was markedly compromised at 37˚C and moderately reduced at 25°C and 30°C (Figure 5B).

A minor suggestion is to include cells expressing PriC using plasmid DNA to show that adding PriC should reverse the growth defect of dnaA46 and dnaC2 strains at non-permissive temperatures. The same should be added at other appropriate places.

Even in the presence of PriC, unwinding of oriC and DnaB helicase loading to the wound oriC require DnaA and DnaC activities as indicated by previous studies (see for a review, Windgassen et al., (2018) Nucleic Acids Res. 46, 504-519). Thus, dnaA46 cells and dnaC2 cells bearing pBR322-priC can not grow at 42°C and 37°C (as follows). These are reasonable results. However, at semi-permissive temperatures (37°C for dnaA46 and 35°C for dnaC2), slight stimulation of the cell growth by pBR322-priC might be barely observed (Figure 2-supplement 1 of the revised version). These suggest that the intrinsic level of PriC is functionally nearly sufficient. This explanation has been included in the revised version.

Author response image 1.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Line 38. "in assembly of the replisome".

Corrected.

Line 137. "specifically" rather than specificity.

Corrected.

Line 139. "at" rather than by.

Corrected.

The DnaA46 protein variant contains two amino acid substitutions (A184V and H252Y) within the AAA+ motif. H136 appears to reside adjacent to A184 in structure. Is A184V mutation causative?

The DnaA H136A and A184V alleles are responsible for different defects. Indeed, the DnaA A184V variant is thermolabile and defective in ATP binding whereas the H136A variant retains ATP binding but impairs DnaB loading (Carr and Kaguni, Mol. Microbiol., 1996; Sakiyama et al., Front. Microbiol., 2018). These observations strongly support the view that the phenotype of the DnaA H136A allele is independent of that of the DnaA A184V allele.

Figure 2A. Regarding the dnaA46 allele grown at 37°C.

Individual colonies cannot be resolved. Is an image from a later time-point available?

We have replaced the original image with one from another replicate that provides better resolution. Please see Figure 2A in the revised version.

Figure 2C. Quantification of the number of cells with more than one chromosome equivalent in the dnaC2 ΔpriC strain. The plot from flow cytometry appears to show >20% of cells with only 1 genome. Are these numbers correct?

Thank you for this careful comment. We quantified the peaks more strictly, but the percentages were noy largely changed. To improve resolution of the DNA profiles, we have changed the range of the x-axis in panels B and C of Figure 2 in the revised version.

Figure 3. Are both F46A and H136A mutations in the plasmid-encoded dnaA necessary?

Yes. The related explanation is included in the Discussion section (the third paragraph) of the original manuscript. As described there, dnaA46 cells expressing the DnaA H136A single mutant exhibited severe defects in cell growth even in the presence of PriC (Sakiyama et al., 2018). The His136 residue is located within the weak, secondary DnaB interaction region in DnaA, and is crucial for DnaB loading onto oriC ssDNA. Given domain I in DnaA H136A can stably tether DnaB-DnaC complexes to DnaA complexes on oriC (Sakiyama et al., 2018), we infer that oriC-DnaA complexes including DnaA H136A stably bind DnaB via DnaA domain I as an abortive complex, which inhibits functional interaction between PriC and DnaB as well as DnaB loading to oriC DNA.

As for DnaA F46A mutant, our previous studies show that DnaA F46A has a limited residual activity in vivo (unlike in vitro), and allows slow growth of cells. As the stable DnaA-DnaB binding is partially impaired in vivo in DnaA F46A, this feature is consistent with the above ideas. Thus, both F46A and H136A mutations are required for severer inhibition of DnaB loading. This is additionally described in the revised Discussion.

Figure 3. Is the DnaA variant carrying F46A and H136A substitutions stably expressed in vivo?

We have performed western blotting, demonstrating that the DnaA variant carrying F46A and H136A substitutions is stable in vivo. In the revised version, we have added new data to Figure 3-figure supplement 1 and relevant description to the main text as follows:

Western blotting demonstrated that the expression levels were comparable between WT DnaA and DnaA F46A H136A double mutant (Figure 3-figure supplement 1).

Figure 5A. Should the dashed line extending down from I2 reach the R4Tma construct?

We have amended the indicated line appropriately.

Figure 6C. It was surprising that the strain combining the subATL mutant with ΔpriC displayed a pronounced under-initiation profile by flow cytometry, and yet there was no growth defect observed (see Figure 6B). This seems to contrast with results using the R4Tma origin, where the ΔpriC mutant produced a relatively modest change to the flow cytometry profile, and yet growth was perturbed (Figure 5C-D). How might these observations be interpreted? Is the absolute frequency of DNA replication initiation critical?

Please note that, in E. coli, initiation activity corelates closely with the numbers of oriC copies per cell mass (ori/mass), rather than the apparent DNA profiles measured by flow cytometer. When cells were grown in LB at 30˚C, the mean ori/mass values were as follows: 0.34 for R4Tma priC, 0.51 for R4Tma, 0.82 for DATL priC, 0.99 for DATL (Figures 5 & 6 in the original manuscript). These values closely correspond to the cell growth ability shown in Figure 5C in the original manuscript.

In the revised manuscript, we have cited appropriate references for introduction of the ori/mass values as follows.

To estimate the number of oriC copies per unit cell mass (ori/mass) as a proxy for initiation activity (Sakiyama et al., 2017, 2022),

Line 295. Reference for Form I* assay should cite the original publication.

Done. The following paper is additionally cited.

Baker, T. A., Sekimizu, K., Funnell, B. E., and Kornberg, A. (1986). Extensive unwinding of the plasmid template during staged enzymatic initiation of DNA replication from the origin of the Escherichia coli chromosome. Cell 45, 53–64.doi: 10.1016/0092-8674(86)90537-4

Reviewer #2 (Recommendations for the authors):

The partial complementation of the dnaC2 strain by PriC seems quite straightforward since this particular mutation leads to initiation arrest at the open complex stage and this sets the stage for PriC to load the helicase. The situation is somewhat different for dnaA46. Why is this mutation partly complemented by PriC at 37C? DnaA46 binds neither ATP nor ADP, yet it functions in initiation at permissive temperature. At nonpermissive temperature, it binds oriC as well but does not lead to initiation. Does the present data imply that the true initiation defect of DnaA46 lies in helicase loading? The authors need to comment on this in the text.

Given the thermolabile propensity of the DnaA46 protein, it is presumable that DnaA46 protein becomes partially denatured at the sub-permissive temperature of 37˚C. This partial denaturation should impair both origin unwinding and helicase loading, though not to the extent that cell viability is lost. The priC deletion should further exacerbate helicase loading defects by inhibiting the bypass mechanism, resulting in the lethality of dnaA46 cells at this temperature. This explanation is included in the revised Discussion section.

Relating to the above. In Figure 3 it is shown that the pFH plasmid partly complements dnaA46 in a PriC-dependent manner. Again, it would be nice to know the nature of the DnaA46 protein defect. It would be interesting to see how a pING1-dnaA46 plasmid performs in the experiment presented in Figure 3.

A previous paper showed that multicopy supply of DnaA46 can suppress temperature sensitivity of the dnaA46 cells (Rao and Kuzminov, G3, 2022). This is reasonable in that DnaA46 has a rapid degradation rate unlike wild-type DnaA. As DnaA46 preserves the intact sequences in DnaB binding sites such as G21, F46 and H136, the suppression would not depend on PriC but would be due to the dosage effect.

Figure 8 B: The authors should either remove the data or show a genome coverage: it is not clear that yapB is a good reference. A genome coverage would be nice, and show whether initiation can occur at oriC even if it is not the major place of initiation in a rnhA mutant.

As suggested, we carried out the chromosome loci copy-number analysis by whole-genome sequencing to assess impact of PriC on cSDR. The new data are shown in Figure 8-supplement 1 with relevant descriptions of the main text of the revised version as shown below. Briefly, results of the chromosome loci copy-number analysis are consistent with those of real-time qPCR (Figure 8B). Given that the role PriC in stimulating cSDR was unclear, we believe that our finding that PriC has little or no role in cSDR, despite being a negative result, is valuable for the general readership of eLife.

Line 38-39: .....resulting in replisome assembly.

Corrected.

Line 48: Something is wrong with the Michel reference. Also in the reference list.

Corrected

Line 156: replace retarded with reduced.

Corrected.

Line 171 and elsewhere: WT priC cells is somewhat misleading. Isn't this simply PriC+ cells?

Yes. We have revised the wording to “priC⁺” for clarity.

Line 349-350: "the oriC copy number ratio of the dnaA46 DpriC double mutant was lower than that of the dnaA46 single mutant....". This is only provided growth rate of the strains is the same.

These strains exhibited similar growth rates. This is included in the Result section of the revised manuscript as follows: At the permissive temperature, despite having similar growth rates, the oriC copy number ratio of the dnaA46 ∆priC double mutant strain was lower than that of the dnaA46 single mutant.

Reviewer #3 (Recommendations for the authors):

I would suggest improved or additional experiments, data, or analyses.

The revised version includes improved or additional experiments, data, or analyses.

Author response:

The following is the authors’ response to the original reviews

ANALYTICAL

(1) A key claim made here is that the same relationship (including the same parameter) describes data from pigeons by Gibbon and Balsam (1981; Figure 1) and the rats in this study (Figure 3). The evidence for this claim, as presented here, is not as strong as it could be. This is because the measure used for identifying trials to criterion in Figure 1 appears to differ from any of the criteria used in Figure 3, and the exact measure used for identifying trials to criterion influences the interpretation of Figure 3***. To make the claim that the quantitative relationship is one and the same in the Gibbon-Balsam and present datasets, one would need to use the same measure of learning on both datasets and show that the resultant plots are statistically indistinguishable, rather than simply plotting the dots from both data sets and spotlighting their visual similarity. In terms of their visual characteristics, it is worth noting that the plots are in log-log axis and, as such, slight visual changes can mean a big difference in actual numbers. For instance, between Figure 3B and 3C, the highest information group moves up only "slightly" on the y-axis but the difference is a factor of 5 in the real numbers. Thus, in order to support the strong claim that the quantitative relationships obtained in the Gibbon-Balsam and present datasets are identical, a more rigorous approach is needed for the comparisons.

***The measure of acquisition in Figure 3A is based on a previously established metric, whereas the measure in Figure 3B employs the relatively novel nDKL measure that is argued to be a better and theoretically based metric. Surprisingly, when r and r2 values are converted to the same metric across analyses, it appears that this new metric (Figure 3B) does well but not as well as the approach in Figure 3A. This raises questions about why a theoretically derived measure might not be performing as well on this analysis, and whether the more effective measure is either more reliable or tapping into some aspect of the processes that underlie acquisition that is not accounted for by the nDKL metric.

Figure 3 shows that the relationship between learning rate and informativeness for our rats was very similar to that shown with pigeons by Gibbon and Balsam (1981). We have used multiple criteria to establish the number of trials to learn in our data, with the goal of demonstrating that the correspondence between the data sets was robust. In the revised Figure 3, specifically 3C and 3D, we have plotted trials to acquisition using decision criterion equivalent to those used by Gibbon and Balsam. The criterion they used—at least one peck at the response key on at least 3 out of 4 consecutive trials—cannot be directly applied to our magazine entry data because rats make magazine entries during the inter-trial interval (whereas pigeons do not peck at the response key in the inter-trial interval). Therefore, evidence for conditioning in our paradigm must involve comparison between the response rate during CS and the baseline response rate, rather than just counting responses during the CS. We have used two approaches to adapt the Gibbon and Balsam criterion to our data. One approach, plotted in Figure 3C, uses a non-parametric signed rank test for evidence that the CS response rate exceeds the pre-CS response rate, and adopting a statistical criterion equivalent to Gibbon and Balsam’s 3-out-of-4 consecutive trials (p<.3125). The second method (Figure 3D) estimates the nDkl for the criterion used by Gibbon and Balsam and then applies this criterion to the nDkl for our data. To estimate the nDkl of Gibbon and Balsam’s data, we have assumed there are no responses in the inter-trial interval and the response probability during the CS must be at least 0.75 (their criterion of at least 3 responses out of 4 trials). The nDkl for this difference is 2.2 (odds ratio 27:1). We have then applied this criterion to the nDkl obtained from our data to identify when the distribution of CS response rates has diverged by an equivalent amount from the distribution of pre-CS response rates. These two analyses have been added to the manuscript to replace those previously shown in Figures 3B and 3C.

(2) Another interesting claim here is that the rates of responding during ITI and the cue are proportional to the corresponding reward rates with the same proportionality constant. This too requires more quantification and conceptual explanation. For quantification, it would be more convincing to calculate the regression slope for the ITI data and the cue data separately and then show that the corresponding slopes are not statistically distinguishable from each other. Conceptually, it is not clear why the data used to test the ITI proportionality came from the last 5 conditioning sessions. What were the decision criteria used to decide on averaging the final 5 sessions as terminal responses for the analyses in Figure 5? Was this based on consistency with previous work, or based on the greatest number of sessions where stable data for all animals could be extracted?

If the model is that animals produce response rates during the ITI (a period with no possible rewards) based on the overall rate of rewards in the context, wouldn't it be better to test this before the cue learning has occurred? Before cue learning, the animals would presumably only have attributed rewards in the context to the context and thus, produce overall response rates in proportion to the contextual reward rate. After cue learning, the animals could technically know that the rate of rewards during ITI is zero. Why wouldn't it be better to test the plotted relationship for ITI before cue learning has occurred? Further, based on Figure 1, it seems that the overall ITI response rate reduces considerably with cue learning. What is the expected ITI response rate prior to learning based on the authors' conceptual model? Why does this rate differ from pre and post-cue learning? Finally, if the authors' conceptual framework predicts that ITI response rate after cue learning should be proportional to contextual reward rate, why should the cue response rate be proportional to the cue reward rate instead of the cue reward rate plus the contextual reward rate?

A single regression line, as shown in Figure 5, is the simplest possible model of the relationship between response rate and reinforcement rate and it explains approximately 80% of the variance in response rate. Fixing the log-log slope at 1 yields the maximally simple model. (This regression is done in the logarithmic domain to satisfy the homoscedasticity assumption.) When transformed into the linear domain, this model assumes a truly scalar relation (linear, intercept at the origin) and assumes the same scale factor and the same scalar variability in response rates for both sets of data (ITI and CS). Our plot supports such a model. Its simplicity is its own motivation (Occam’s razor).

If separate regression lines are fitted to the CS and ITI data, there is a small increase in explained variance (R₂ = 0.82). These regression lines have been added to the plot in the revised manuscript (Figure 5). We leave it to further research to determine whether such a complex model, with 4 parameters, is required. However, we do not think the present data warrant comparing the simplest possible model, with one parameter, to any more complex model for the following reasons:

· When a brain—or any other machine—maps an observed (input) rate to a rate it produces (output rate), there is always an implicit scalar. In the special case where the produced rate equals the observed rate, the implicit scalar has value 1. Thus, there cannot be a simpler model than the one we propose, which is, in and of itself, interesting.

· The present case is an intuitively accessible example of why the MDL (Minimum Description Length) approach to model complexity (Barron, Rissanen, & Yu, 1998; Grünwald, Myung, & Pitt, 2005; Rissanen, 1999) can yield a very different conclusion from the conclusion reached using the Bayesian Information Criterion (BIC) approach. The MDL approach measures the complexity of a model when given N data specified with precision of B bits per datum by computing (or approximating) the sum of the maximum-likelihoods of the model’s fits to all possible sets of N data with B precision per datum. The greater the sum over the maximum likelihoods, the more complex the model, that is, the greater its measured wiggle room, it’s capacity to fit data. Recall that von Neuman remarked to Fermi that with 4 parameters he could fit an elephant. His deeper point was that multi-parameter models bring neither insight nor predictive power; they explain only post-hoc, after one has adjusted their parameters in the light of the data. For realistic data sets like ours, the sums of maximum likelihoods are finite but astronomical. However, just as the Sterling approximation allows one to work with astronomical factorials, it has proved possible to develop readily computable approximations to these sums, which can be used to take model complexity into account when comparing models. Proponents of the MDL approach point out that the BIC is inadequate because models with the same number of parameters can have very different amounts of wiggle room. A standard illustration of this point is the contrast between logarithmic model and power-function model. Log regressions must be concave; whereas power function regressions can be concave, linear, or convex—yet they have the same number of parameters (one or two, depending on whether one counts the scale parameter that is always implicit). The MDL approach captures this difference in complexity because it measures wiggle room; the BIC approach does not, because it only counts parameters.

· In the present case, one is comparing a model with no pivot and no vertical displacement at the boundary between the black dots and the red dots (the 1-parameter unilinear model) to a bilinear model that allows both a change in slope and a vertical displacement for both lines. The 4-parameter model is superior if we use the BIC to take model complexity into account. However, 4-parameter has ludicrously more wiggle room. It will provide excellent fits—high maximum likelihood—to data sets in which the red points have slope > 1, slope 0, or slope < 0 and in which it is also true that the intercept for the red points lies well below or well above the black points (non-overlap in the marginal distribution of the red and black data). The 1-parameter model, on the other hand, will provide terrible fits to all such data (very low maximum likelihoods). Thus, we believe the BIC does not properly capture the immense actual difference in the complexity between the 1-parameter model (unilinear with slope 1) to the 4-parameter model (bilinear with neither the slope nor the intercept fixed in the linear domain).

· In any event, because the pivot (change in slope between black and red data sets), if any, is small and likewise for the displacement (vertical change), it suffices for now to know that the variance captured by the 1-parameter model is only marginally improved by adding three more parameters. Researchers using the properly corrected measured rate of head poking to measure the rate of reinforcement a subject expects can therefore assume that they have an approximately scalar measure of the subject’s expectation. Given our data, they won’t be far wrong even near the extremes of the values commonly used for rates of reinforcement. That is a major advance in current thinking, with strong implications for formal models of associative learning. It implies that the performance function that maps from the neurobiological realization of the subject’s expectation is not an unknown function. On the contrary, it’s the simplest possible function, the scalar function. That is a powerful constraint on brain-behavior linkage hypotheses, such as the many hypothesized relations between mesolimbic dopamine activity and the expectation that drives responding in Pavlovian conditioning (Berridge, 2012; Jeong et al., 2022; Y.  Niv, Daw, Joel, & Dayan, 2007; Y. Niv & Schoenbaum, 2008).

The data in Figures 4 and 5 are taken from the last 5 sessions of training. The exact number of sessions was somewhat arbitrary but was chosen to meet two goals: (1) to capture asymptotic responding, which is why we restricted this to the end of the training, and (2) to obtain a sufficiently large sample of data to estimate reliably each rat’s response rate. We have checked what the data look like using the last 10 sessions, and can confirm it makes very little difference to the results. We now note this in the revised manuscript. The data for terminal responding by all rats, averaged over both the last 5 sessions and last 10 sessions, can be downloaded from https://osf.io/vmwzr/

Finally, as noted by the reviews, the relationship between the contextual rate of reinforcement and ITI responding should also be evident if we had measured context responding prior to introducing the CS. However, there was no period in our experiment when rats were given unsignalled reinforcement (such as is done during “magazine training” in some experiments). Therefore, we could not measure responding based on contextual conditioning prior to the introduction of the CS. This is a question for future experiments that use an extended period of magazine training or “poor positive” protocols in which there are reinforcements during the ITIs as well as during the CSs. The learning rate equation has been shown to predict reinforcements to acquisition in the poor-positive case (Balsam, Fairhurst, & Gallistel, 2006).

(3) There is a disconnect between the gradual nature of learning shown in Figures 7 and 8 and the information-theoretic model proposed by the authors. To the extent that we understand the model, the animals should simply learn the association once the evidence crosses a threshold (nDKL > threshold) and then produce behavior in proportion to the expected reward rate. If so, why should there be a gradual component of learning as shown in these figures? In terms of the proportional response rule to the rate of rewards, why is it changing as animals go from 10% to 90% of peak response? The manuscript would be greatly strengthened if these results were explained within the authors' conceptual framework. If these results are not anticipated by the authors' conceptual framework, this should be explicitly stated in the manuscript.

One of us (CRG) has earlier suggested that responding appears abruptly when the accumulated evidence that the CS reinforcement rate is greater than the contextual rate exceeds a decision threshold (C.R.  Gallistel, Balsam, & Fairhurst, 2004). The new more extensive data require a more nuanced view. Evidence about the manner in which responding changes over the course of training is to some extent dependent on the analytic method used to track those changes. We presented two different approaches. The approach shown in Figures 7 and 8 (now 6 and 7), extending on that developed by Harris (2022), assumes a monotonic increase in response rate and uses the slope of the cumulative response rate to identify when responding exceeds particular milestones (percentiles of the asymptotic response rate). This analysis suggests a steady rise in responding over trials. Within our theoretical model, this might reflect an increase in the animal’s certainty about the CS reinforcement rate with accumulated evidence from each trial. While this method should be able to distinguish between a gradual change and a single abrupt change in responding (Harris, 2022) it may not distinguish between a gradual change and multiple step-like changes in responding and cannot account for decreases in response rate.

The other analytic method we used relies on the information theoretic measure of divergence, the nDkl (Gallistel & Latham, 2023), to identify each point of change (up or down) in the response record. With that method, we discern three trends. First, the onset tends to be abrupt in that the initial step up is often large (an increase in response rate by 50% or more of the difference between its initial value and its terminal value is common and there are instances where the initial step is to the terminal rate or higher). Second, there is marked within-subject variability in the response rate, characterized by large steps up and down in the parsed response rates following the initial step up, but this variability tends to decrease with further training (there tend to be fewer and smaller steps in both the ITI response rates and the CS response rate as training progresses). Third, the overall trend, seen most clearly when one averages across subjects within groups is to a moderately higher rate of responding later in training than after the initial rise. We think that the first tendency reflects an underlying decision process whose latency is controlled by diminishing uncertainty about the two reinforcement rates and hence about their ratio. We think that decreasing uncertainty about the true values of the estimated rates of reinforcement is also likely to be an important part of the explanation for the second tendency (decreasing within-subject variation in response rates). It is less clear whether diminishing uncertainty can explain the trend toward a somewhat greater difference in the two response rates as conditioning progresses. It is perhaps worth noting that the distribution of the estimates of the informativeness ratio is likely to be heavy tailed and have peculiar properties (as witness, for example, the distribution of the ratio of two gamma distributions with arbitrary shape and scale parameters) but we are unable at this time to propound an explanation of the third trend.

(4) Page 27, Procedure, final sentence: The magazine responding during the ITI is defined as the 20 s period immediately before CS onset. The range of ITI values (Table 1) always starts as low as 15 s in all 14 groups. Even in the case of an ITI on a trial that was exactly 20 s, this would also mean that the start of this period overlaps with the termination of the CS from the previous trial and delivery (and presumably consumption) of a pellet. It should be indicated whether the definition of the ITI period was modified on trials where the preceding ITI was < 20 s, and if any other criteria were used to define the ITI. Were the rats exposed to the reinforcers/pellets in their home cage prior to acquisition?

There was an error in the description provided in the original text. The pre-CS period used to measure the ITI responding was 10 s rather than 20 s. There was always at least a 5-s gap between the end of the previous trial and the start of the pre-CS period. The statement about the pre-CS measure has been corrected in the revised manuscript.

(5) For all the analyses, the exact models that were fit and the software used should be provided. For example, it is not necessarily clear to the reader (particularly in the absence of degrees of freedom) that the model discussed in Figure 3 fits on the individual subject data points or the group medians. Similarly, in Figure 6 there is no indication of whether a single regression model was fit to all the plotted data or whether tests of different slopes for each of the conditions were compared. With regards to the statistics in Figure 6, depending on how this was run, it is also a potential problem that the analyses do not correct for the potentially highly correlated multiple measurements from the same subjects, i.e. each rat provides 4 data points which are very unlikely to be independent observations.

Details about model fitting have been added to the revision. The question about fitting a single model or multiple models to the data in Figure 6 (now 5) is addressed in response 2 above. In Figure 5, each rat provides 2 behavioural data points (ITI response rate and CS response rate) and 2 values for reinforcement rate (1/C and 1/T). There is a weak but significant correlation between the ITI and CS response rates (r = 0.28, p < 0.01; log transformed to correct for heteroscedasticity). By design, there is no correlation between the log reinforcement rates (r = 0.06, p = .404).

CONCEPTUAL

(1) We take the point that where traditional theories (e.g., Rescorla-Wagner) and rate estimation theory (RET) both explain some phenomenon, the explanation in terms of RET may be preferred as it will be grounded in aspects of an animal's experience rather than a hypothetical construct. However, like traditional theories, RET does not explain a range of phenomena - notably, those that require some sort of expectancy/representation as part of their explanation. This being said, traditional theories have been incorporated within models that have the representational power to explain a broader array of phenomena, which makes me wonder: Can rate estimation be incorporated in models that have representational power; and, if so, what might this look like? Alternatively, do the authors intend to claim that expectancy and/or representation - which follow from probabilistic theories in the RW mould - are unnecessary for explanations of animal behaviour?***

It is important for the field to realize that the RW model cannot be used to explain the results of Rescorla’s (Rescorla, 1966; Rescorla, 1968, 1969) contingency-not-pairing experiments, despite what was claimed by Rescorla and Wagner (Rescorla & Wagner, 1972; Wagner & Rescorla, 1972) and has subsequently been claimed in many modelling papers and in most textbooks and reviews (Dayan & Niv, 2008; Y. Niv & Montague, 2008). Rescorla programmed reinforcements with a Poisson process. The defining property of a Poisson process is its flat hazard function; the reinforcements were equally likely at every moment in time when the process was running. This makes it impossible to say when non-reinforcements occurred and, a fortiori, to count them. The non-reinforcements are causal events in RW algorithm and subsequent versions of it. Their effects on associative strength are essential to the explanations proffered by these models. Non-reinforcements—failures to occur, updates when reinforcement is set to 0, hence also the lambda parameter—can have causal efficacy only when the successes may be predicted to occur at specified times (during “trials”). When reinforcements are programmed by a Poisson process, there are no such times. Attempts to apply the RW formula to reinforcement learning soon foundered on this problem (Gibbon, 1981; Gibbon, Berryman, & Thompson, 1974; Hallam, Grahame, & Miller, 1992; L.J. Hammond, 1980; L. J. Hammond & Paynter, 1983; Scott & Platt, 1985). The enduring popularity of the delta-rule updating equation in reinforcement learning depends on “big-concept” papers that don’t fit models to real data and discretize time into states while claiming to be real-time models (Y. Niv, 2009; Y. Niv, Daw, & Dayan, 2005).

The information-theoretic approach to associative learning, which sometimes historically travels as RET (rate estimation theory), is unabashedly and inescapably representational. It assumes a temporal map and arithmetic machinery capable in principle of implementing any implementable computation. In short, it assumes a Turing-complete brain. It assumes that whatever the material basis of memory may be, it must make sense to ask of it how many bits can be stored in a given volume of material. This question is seldom posed in associative models of learning, nor by neurobiologists committed to the hypothesis that the Hebbian synapse is the material basis of memory. Many—including the new Nobelist, Geoffrey Hinton— would agree that the question makes no sense. When you assume that brains learn by rewiring themselves rather than by acquiring and storing information, it makes no sense.

When a subject learns a rate of reinforcement, it bases its behavior on that expectation, and it alters its behavior when that expectation is disappointed. Subjects also learn probabilities when they are defined. They base some aspects of their behavior on those expectations, making computationally sophisticated use of their representation of the uncertainties (Balci, Freestone, & Gallistel, 2009; Chan & Harris, 2019; J. A. Harris, 2019; J.A. Harris & Andrew, 2017; J. A. Harris & Bouton, 2020; J. A. Harris, Kwok, & Gottlieb, 2019; Kheifets, Freestone, & Gallistel, 2017; Kheifets & Gallistel, 2012; Mallea, Schulhof, Gallistel, & Balsam, 2024 in press).

(2) The discussion of Rescorla's (1967) and Kamin's (1968) findings needs some elaboration. These findings are already taken to mean that the target CS in each design is not informative about the occurrence of the US - hence, learning about this CS fails. In the case of blocking, we also know that changes in the rate of reinforcement across the shift from stage 1 to stage 2 of the protocol can produce unblocking. Perhaps more interesting from a rate estimation perspective, unblocking can also be achieved in a protocol that maintains the rate of reinforcement while varying the sensory properties of the US (Wagner). How does rate estimation theory account for these findings and/or the demonstrations of trans-reinforcer blocking (Pearce-Ganesan)? Are there other ways that the rate estimation account can be distinguished from traditional explanations of blocking and contingency effects? If so, these would be worth citing in the discussion. More generally, if one is going to highlight seminal findings (such as those by Rescorla and Kamin) that can be explained by rate estimation, it would be appropriate to acknowledge findings that challenge the theory - even if only to note that the theory, in its present form, is not all-encompassing. For example, it appears to me that the theory should not predict one-trial overshadowing or the overtraining reversal effect - both of which are amenable to discussion in terms of rates.

I assume that the signature characteristics of latent inhibition and extinction would also pose a challenge to rate estimation theory, just as they pose a challenge to Rescorla-Wagner and other probability-based theories. Is this correct?

The seemingly contradictory evidence of unblocking and trans-reinforcer blocking by Wagner and by Pearce and Ganesan cited above will be hard for any theory to accommodate. It will likely depend on what features of the US are represented in the conditioned response.

RET predicts one-trial overshadowing, as anyone may verify in a scientific programming language because it has no free parameters; hence, no wiggle room. Overtraining reversal effects appear to depend on aspects of the subjects’ experience other than the rate of reinforcement. It seems unlikely that it can proffer an explanation.

Various information-theoretic calculations give pretty good quantitative fits to the relatively few parametric studies of extinction and the partial-reinforcement extinction effect (see Gallistel (2012, Figs 3 & 4); Wilkes & Gallistel (2016, Fig 6) and Gallistel (2025, under review, Fig 6). It has not been applied to latent inhibition, in part for want of parametric data. However, clearly one should not attribute a negative rate to a context in which the subject had never been reinforced. An explanation, if it exists, would have to turn on the effect of that long period on initial rate estimates AND on evidence of a change in rate, as of the first reinforcement.

Recommendations for authors:

MINOR POINTS

(1) It is not clear why Figure 3C is presented but not analyzed, and why the data presented in Figure 4 to clarify the spread of the distribution of the data observed across the plots in Figure 3 uses the data from Figure 3C. This would seem like the least representative data to illustrate the point of Figure 4. It also appears that the data plotted in Figure 4 corresponds to Figure 3A and 3B rather than the odds 10:1 data indicated in the text.

Figures 3 has changed as already described. The data previously plotted in Figure 4 are now shown in 3B and corresponds to that plotted in Figure 3A.

(2) Log(T) was not correlated with trials to criterion. If trials to criterion is inversely proportional to log(C/T) and C is uncorrelated with T, shouldn't trials to criterion be correlated with log(T)? Is this merely a matter of low statistical power?

Yes. There is a small, but statistically non-significant, correlation between log(T) and trials to criterion, r = 0.35, p = .22. That correlation drops to .08 (p = .8) after factoring out log(C/T), which demonstrates that the weak correlation between log(T) and trials to criterion is based on the correlation between log(t) and log(C/T).

(3) The rationale for the removal of the high information condition samples in the Fig 8 "Slope" plot to be weak. Can the authors justify this choice better? If all data are included, the relationship is clearly different from that shown in the plot.

We have now reported correlations that include those 3 groups but noted that the correlations are largely driven by the much lower slope values of those 3 groups which is likely an artefact of their smaller number of trials. We use this to justify a second set of correlations that excludes those 3 groups.

(4) The discussion states that there is at most one free parameter constrained by the data - the constant of proportionality for response rate. However, there is also another free parameter constrained by data-the informativeness at which expected trials to acquisition is 1.

I think this comment is referring to two different sets of data. The constant of proportionality of the response rate refers to the scalar relationship between reinforcement rate and terminal response rate shown in Figure 5. The other parameter, the informativeness when trials to acquisition equals 1, describes the intercept of the regression line in Figure 1 (and 3).

(5) The authors state that the measurement of available information is not often clear. Given this, how is contingency measurable based on the authors' framework?

(6) Based on the variables provided in Supplementary File 3, containing the acquisition data, we were unable to reproduce the values reported in the analysis of Figure 3.

Figure 3 has changed, using new criteria for trials to acquisition that attempt to match the criterion used by Gibbon and Balsam. The data on which these figures are based has been uploaded into OSF.

GRAPHICAL AND TYPOGRAPHICAL

(1) Y-axis labels in Figure 1 are not appropriately placed. 0 is sitting next to 0.1. 0 should sit at the bottom of the y-axis.

If this comment refers to the 0 sitting above an arrow in the top right corner of the plot, this is not misaligned. The arrow pointing to zero is used to indicate that this axis approaches zero in the upward direction. 0 should not be aligned to a value on the axis since a learning rate of zero would indicate an infinite number of learning trials. The caption has been edited to explain this more clearly.

(2) Typo, Page 6, Final Paragraph, line 4. "Fourteen groups of rats were trained with for 42 session"

Corrected. Thank you.

(3) Figure 3 caption: Typo, should probably be "Number of trials to acquisition"?

This change has now been made. The axis shows reinforcements to acquisition to be consistent with Gibbon and Balsam, but trials and number of reinforcements are identical in our 100% reinforcement schedule.

(4) Typo Page 17 Line 1: "Important pieces evidence about".

Correct. Thank you.

(5) Consider consistent usage of symbols/terms throughout the manuscript (e.g. Page 22, final paragraph: "iota = 2" is used instead of the corresponding symbol that has been used throughout).

Changed.

(6) Typo Page 28, Paragraph 1, Line 9: "We used a one-sample t-test using to identify when this".

This section of text has been changed to reflect the new analysis used for the data in Figure 3.

(7) Typo Page 29, Paragraph 1, Line 2: "problematic in cases where one of both rates are undefined" either typo or unclear phrasing.

“of” has been corrected to “or”

(8) Typo Page 30: Equation 3 appears to have an error and is not consistent with the initial printing of Equation 3 in the manuscript.

The typo in initial expression of Eq 3 (page 23) has been corrected.

(9) Typo Page 33, Line 5: "Figures 12".

Corrected.

(10) Typo Page 34, Line 10: "and the 5 the increasingly"? Should this be "the 5 points that"?

Corrected.

(11) Typo Page 35, Paragraph 2: "estimate of the onset of conditioned is the trial after which".

Corrected.

(12) Clarify: Page 35, final paragraph: it is stated that four-panel figures are included for each subject in the Supplementary files, but each subject has a six-panel figure in the Supplementary file.

The text now clarifies that the 4-panel figures are included within the 6-panel figures in the Supplementary materials.

(13) It is hard to identify the different groups in Figure 2 (Plot 15).

The figure is simply intended to show that responding across seconds within the trial is relatively flat for each group. Individuation of specific groups is not particularly important.

(14) It appears that the numbering on the y-axis is misaligned in Figure 2 relative to the corresponding points on the scale (unless I have misunderstood these values and the response rate measure to the ITI can drop below 0?).

The numbers on the Y axes had become misaligned. That has now been corrected.

(15) Please include the data from Figure 3A in the spreadsheet supplementary file 3. If it has already been included as one of the columns of data, please consider a clearer/consistent description of the relevant column variable in Supplementary File 1.

The data from Figure 3 are now available from the linked OSF site, referenced in the manuscript.

(16) Errors in supplementary data spreadsheets such that the C/T values are not consistent with those provided in Table 1 (C/T values of 4.5, 54, 180, and 300 are slightly different values in these spreadsheets). A similar error/mismatch appears to have occurred in the C/T labels for Figures (e.g. Figure 10) and the individual supplementary figures.

The C/T values on the figures in the supplementary materials have been corrected and are now consistent with those in Table 1.

(17) Currently the analysis and code provided at https://osf.io/vmwzr/ are not accessible without requesting access from the author. Please consider making these openly available without requiring a request for authorization. As such, a number of recommendations made here may already have been addressed by the data and code deposited on OSF. Apologies for any redundant recommendations.

Data and code are now available in at the OSF site which has been made public without requiring request.

(18) Please consider a clearer and more specific reference to supplementary materials. Currently, the reader is required to search through 4 separate supplementary files to identify what is being discussed/referenced in the text (e.g. Page 18, final line: "see Supplementary Materials" could simply be "see Figure S1").

We have added specific page numbers in references to the Supplementary Materials.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

This manuscript describes a novel magnetic steering technique to target human adipose derived mesenchymal stem cells (hAMSC) or induce pluripotent stem cells to the TM (iPSC-TM). The authors show that delivery of the stem cells lowered IOP, increased outflow facility, and increased TM cellularity.

Strengths:

The technique is novel and shows promise as a novel therapeutic to lower IOP in glaucoma. hAMSC are able to lower IOP below the baseline as well as increase outflow facility above baseline with no tumorigenicity. These data will have a positive impact on the field and will guide further research using hAMSC in glaucoma models.

Weaknesses:

The transgenic mouse model of glaucoma the authors used did not show ocular hypertensive phenotypes at 6-7 months of age as previously reported. Therefore, if there is no pathology in these animals the authors did not show a restoration of function, but rather a decrease in pressure below normal IOP.

We appreciate the reviewer’s feedback and agree with the statement of weakness. Accordingly, we have revised the language to improve clarity. Specifically, all references to "restoration of IOP" or "restoration of conventional outflow function" have been replaced with more precise phrases, in the following locations: 

• lines 2-3 (title): Magnetically steered cell therapy for reduction of intraocular pressure  as a treatment strategy for open-angle glaucoma

• lines 36-8 (abstract): We observed a 4.5 [3.1, 6.0] mmHg or 27% reduction in intraocular pressure (IOP) for nine months after a single dose of only 1500 magnetically-steered hAMSCs, explained by increased conventional outflow facility and associated with higher TM cellularity.

• lines 45-6 (one-sentence summary): A novel magnetic cell therapy provided effective intraocular pressure reduction in mice, motivating future translational studies.

• lines 123-4 (introduction): Despite the absence of ocular hypertension in our MYOC^Y437H mice, our data demonstrate sustained IOP lowering and a significant benefit of magnetic cell steering in the eye, particularly for hAMSCs, strongly indicating further translational potential.

• line 207 (results): The observed reductions in IOP and increases in outflow facility after delivery of both cell types suggested functional changes in the conventional outflow pathway.

• line 509-10 (discussion): In summary, this work shows the effectiveness of our novel magnetic TM cell therapy approach for long-term IOP reduction through functional changes in the conventional outflow pathway.

It is very important to note that at the 23rd annual Trabecular Meshwork Study Club meeting (San Diego, December 2024), Dr. Zode, the lead author of reference 26 originally describing the transgenic myocilin mouse model, announced during his talk that this model no longer demonstrates the glaucomatous phenotype in his hands, which incidentally has motivated him to create a new, CRISPR MYOC mouse model. Dr. Zode also stated that he was uncertain of the reason for this loss of phenotype. His observation is consistent with our report. However, other investigators continue to observe the desired phenotype in their colonies of this mouse (Dr. Wei Zhu, personal communication). Continued use of this mouse model should therefore be approached with caution. 

Reviewer #2 (Public review):

Summary:

This observational study investigates the efficacy of intracameral injected human stem cells as a means to re-functionalize the trabecular meshwork for the restoration of intraocular pressure homeostasis. Using a murine model of glaucoma, human adiposederived mesenchymal stem cells are shown to be biologically safer and functionally superior at eliciting a sustained reduction in intraocular pressure (IOP). The authors conclude that the use of human adipose-derived mesenchymal stem cells has the potential for long-term treatment of ocular hypertension in glaucoma.

Strengths:

A noted strength is the use of a magnetic steering technique to direct injected stem cells to the iridocorneal angle. An additional strength is the comparison of efficacy between two distinct sources of stem cells: human adipose-derived mesenchymal vs. induced pluripotent cell derivatives. Utilizing both in vivo and ex vivo methodology coupled with histological evidence of introduced stem cell localization provides a consistent and compelling argument for a sustainable impact exogenous stem cells may have on the refunctionalization of a pathologically compromised TM.

Weaknesses:

A noted weakness of the study, as pointed out by the authors, includes the unanticipated failure of the genetic model to develop glaucoma-related pathology (elevated IOP, TM cell changes). While this is most unfortunate, it does temper the conclusion that exogenous human adipose derived mesenchymal stem cells may restore TM cell function. Given that TM cell function was not altered in their genetic model, it is difficult to say with any certainty that the introduced stem cells would be capable of restoring pathologically altered TM function. A restoration effect remains to be seen. 

We acknowledge that the phrase “restoration of TM function” is not fully supported by our results, given the absence of ocular hypertension in our animal model. Accordingly, we have revised the language to more precisely describe our findings. For specific details regarding these changes, please refer to our response to Reviewer 1’s public comments above.

Another noted complication to these findings is the observation that sham intracameralinjected saline control animals all showed elevated IOP and reduced outflow facility, compared to WT or Tg untreated animals, which allowed for more robust statistically significant outcomes. Additional comments/concerns that the authors may wish to address are elaborated in the Private Review section.

We agree that sham-injected animals tended to have higher average IOPs than transgenic animals in our study. However, these differences did not reach statistical significance and therefore remain inconclusive. Further, an increase in IOP following placebo injection has been previously reported (Zhu et al., 2016). 

Prompted by the Referee’s comments and also a private comment from Referee 1, we further investigated this effect by analyzing IOP in uninjected contralateral eyes at the mid-term time point and comparing the IOPs in these eyes to other cohorts, as now presented as additional data in Supplementary Tables 1 and 2 and Supplementary Figure 4 (see below). In brief, the uninjected contralateral transgenic eyes (10 months old) showed an IOP of 16.5 [15.9, 17.1] mmHg, which was intermediate between the IOP levels of the 6–7-month-old Tg group (15.4 [14.7, 16.1] mmHg) and the sham group (16.9 [15.5, 18.2] mmHg). However, none of these differences reached statistical significance. Additionally, we cannot rule out potential contralateral effects induced by the injections.

Regarding the best way to assess the effect of cell treatment, we feel very strongly that the most relevant IOP comparison is between cell-injected eyes and control (vehicle)-injected eyes, since this provides the most direct accounting for the effects of injection itself on IOP. Other comparisons, such as WT or untreated Tg eyes vs. cell-treated eyes, are interesting but harder to interpret. However, in response to the referee’s comment, we have added comparisons between cell-treated groups and untreated Tg eyes to Table 2, adjusting the post-hoc corrections accordingly. All hAMSC treated groups show statistically significant decrease in IOP even compared to Tg untreated eyes, while iPSC-TMs fail to reach such significance.

The following changes were made to the manuscript:

Lines 326 et seq.: Eyes subjected to saline injection exhibited marginally higher IOPs and lower outflow facilities on average, in comparison to the transgenic animals at baseline. However, due to the lack of statistical significance in these differences and the inherent age difference between the saline-injected animals and the non-injected controls at baseline, no conclusive inference can be drawn regarding the effect of saline injection. To investigate this phenomenon further, we also analyzed IOPs in uninjected contralateral eyes at the midterm time point (Supplementary Tables 1 and 2, Supplementary Figure 4). The uninjected contralateral transgenic eyes (10 months old) showed an IOP of 16.5 [15.9, 17.1] mmHg, which was intermediate between the IOP levels of the 6–7-month-old Tg group (15.4 [14.7, 16.1] mmHg) and the sham-injected group (16.9 [15.5, 18.2] mmHg). However, none of these differences reached statistical significance. Of note, contralateral hypertension has been previously reported after subconjunctival and periocular injection of dexamethasoneloaded nanoparticles (34), and we similarly cannot definitively rule out potential contralateral effects induced by our stem cell injections. Thus, we cannot draw any definite conclusions from these additional IOP comparisons at this time.

Reviewer #3 (Public review):

Summary:

The purpose of the current manuscript was to investigate a magnetic cell steering technique for efficiency and tissue-specific targeting, using two types of stem cells, in a mouse model of glaucoma. As the authors point out, trabecular meshwork (TM) cell therapy is an active area of research for treating elevated intraocular pressure as observed in glaucoma. Thus, further studies determining the ideal cell choice for TM cell therapy is warranted. The experimental protocol of the manuscript involved the injection of either human adipose derived mesenchymal stem cells (hAMSCs) or induced pluripotent cell derivatives (iPSC-TM cells) into a previously reported mouse glaucoma model, the transgenic MYOCY437H mice and wild-type littermates followed by the magnetic cell steering. Numerous outcome measures were assessed and quantified including IOP, outflow facility, TM cellularity, retention of stem cells, and the inner wall BM of Schlemm's canal.

Strengths:

All of these analyses were carefully carried out and appropriate statistical methods were employed. The study has clearly shown that the hAMSCs are the cells of choice over the iPSC-TM cells, the latter of which caused tumors in the anterior chamber. The hAMSCs were shown to be retained in the anterior segment over time and this resulted in increased cellular density in the TM region and a reduction in IOP and outflow facility. These are all interesting findings and there is substantial data to support it.

Weaknesses:

However, where the study falls short is in the MYOCY437H mouse model of glaucoma that was employed. The authors clearly state that a major limitation of the study is that this model, in their hands, did not exhibit glaucomatous features as previously reported, such as a significant increase in IOP, which was part of the overall purpose of the study. The authors state that it is possible that "the transgene was silenced in the original breeders". The authors did not show PCR, western blot, or immuno of angle tissue of the tg to determine transgenic expression (increased expression of MYOC was shown in the angle tissue of the transgenics in the original paper by Zode et al, 2011). This should be investigated given that these mice were rederived. Thus, it is clearly possible that these are not transgenic mice.

All MYOC mice that were used in this study were genotyped and confirmed to carry the transgene as noted in the original version of the paper (see lines 590-2). However, the transgene seems not to have been active, based on the lack of ocular hypertension as well as the lack of differences in supporting endpoints such as outflow facility and TM cellularity. While it would have been possible to carry out their recommended assays to investigate the root cause of this loss of phenotype this was not an objective of our study. Thus we instead here focus simply on communicating the observed loss of phenotype to readers. We also refer the referee to the final paragraph of our response to Referee 1. 

If indeed they are transgenics, the authors may want to consider the fact that in the Zode paper, the most significant IOP elevation in the mutant mice was observed at night and thus this could be examined by the authors. 

This is a good point. However, while the dark-phase IOP does exhibit a distinctly larger elevation (as previously observed in hypertonic saline sclerosis), Zode et al. also reported a notable 3 mmHg IOP increase during the light phase. The complete absence of such daytime (light phase) IOP elevation in our animals diminished our enthusiasm for pursuing darkphase IOP measurements. 

Other glaucomatous features of these mice could also have been investigated such as loss of RGCs, to further determine their transgenic phenotype. 

We agree that these other phenotypes could be studied, but in the absence of any detectable IOP elevation (and thus lack of mechanical insult on RGC axons), loss of RGC is extremely unlikely. We also note that the loss of retinal ganglion cells (RGCs) in the Myocilin model remains a subject of controversy. For example, despite a significant increase in IOP (>10 mmHg) in this model across four mouse strains, three, including C57BL6/J, did not exhibit any signs of optic nerve damage (McDowell et al., 2012). In contrast, Zhu et al. observed considerable nerve damage in this model, which was reversed following iPSC-TM cell transplantation (Zhu et al., 2016). Given these conflicting findings, we directed our efforts toward outcome measures directly related to aqueous humor dynamics.

Finally, while increased cellular density in the TM region was observed, proliferative markers could be employed to determine if the transplanted cells are proliferating.

We agree that identifying the source of the increased trabecular meshwork (TM) cellularity we observed is interesting and we plan to pursue that in future studies. 

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

The sham-injected transgenic animals showed elevated IOP 3-4 weeks after the baseline measurements in the transgenic mice. The authors justify this may be due to the increase in age in these animals. However, this seems unlikely due to the short duration of time between measurement of the baseline IOP and the Short time point (3-4 weeks). The authors do not provide IOP data for any WT sham injected eyes or naïve Tg eyes at these time points. These data are essential to determine if the elevation is due to the sham injection, age, or the transgene. Could it be that the IOP in this cohort of Tg mice didn't increase until 7-8 months of age instead of 6-7 months of age? The methods state only unilateral injections of the stem cells were done so it is assumed the contralateral eye was uninjected. What was the IOP in these eyes? These data would clarify the confusion in the data from sham-injected animals compared to baseline (naive) measurements.

We agree that the average IOP in saline-injected groups is higher than in WT or non-treated Tg mice, although the difference is inconclusive due to a lack of statistical significance. It is important to note, however, that this difference is subtle and not comparable to the 3 mmHg light-phase IOP elevation previously observed in this model (Zode et al., 2011). 

We appreciate the reviewer’s suggestion to include IOP data from the contralateral uninjected eyes, and we have now provided this information along with the comparative statistics in the supplementary materials. Additional details can be found in our response to a similar comment from Reviewer 2’s public review. In summary, the IOP difference in contralateral non-injected ten-month-old transgenic eyes was even smaller than in the original Tg group. IOP elevation following saline injection in mice has been reported previously (Zhu et al., 2016). As a potential confounding factor, we highlight possible contralateral effects of the injection itself (which is why we initially did not analyze IOP in the contralateral eyes).

The hAMSC-treated eyes appear to lower IOP even from baseline (although stats were only provided compared to the sham-injected eyes, which as stated above appear to have increased).

However, the iPSC-TM-treated eyes had IOPs equal to that of the baseline measurements taken 3 weeks prior. The significance is coming from the "sham-treated" eyes which had elevated IOPs. The controls listed above should be included to make these conclusions.

The reviewer makes an astute observation. Please refer to our response to a similar observation by Reviewer 2 under public reviews, where we provide and discuss the comparative statistics noted by the reviewer. However, we feel very strongly that the most relevant IOP comparison is between cell-injected eyes and control-injected eyes. 

If the transgenic mouse model truly did not have a phenotype, then the authors are testing the ability of the stem cells to lower IOP from baseline normal pressures. Therefore, the authors are not "restoring function of the conventional outflow pathway" as there is no damage to begin with. The language in the manuscript should be corrected to reflect this if the transgenics have no phenotype.

We agree and have adjusted the language accordingly. For further details, please refer to our response to your public review.

The authors noted in the iPSC-TM-treated eyes there was a high rate of tumorigenicity. If the magnetic steering of these cells is specific and targeted to the TM, why do the tumors form near the central iris?

While magnetic steering is more specific to the trabecular meshwork (TM) than previouslyused approaches (Bahrani Fard et al., 2023), it is not perfect, and a modest amount of offtarget delivery to the iris, including its central portion, still occurs. Apparently, it took only a few mis-directed iPSC-TM cells to lead to tumors in this work, which is a serious concern for future translational approaches. 

Reviewer #2 (Recommendations for the authors):

(1) It appears that mice were injected unilaterally (Line 590). I may have missed this, but was the companion un-injected eye analyzed in this study? If not analyzed, was there a confounding concern or limitation that necessitated omitting this possible control option?

Contralateral effects, such as hypertension in the untreated eye after subconjunctival and periocular injection of dexamethasone-loaded nanoparticles, have previously been reported in the literature (Li et al., 2019) and also reported anecdotally by other leaders in the field to the senior authors, which is why we did not initially analyze contralateral eyes in this study. However, prompted by this comment and others, we have now included the IOP measurements for contralateral uninjected ten-month-old transgenic eyes in the supplementary materials. For further details, please refer to our response to your public review.

(2) Were all these mice the same gender? Would gender be expected to alter the findings of this study?

Animals of both sexes were randomly chosen and included in the study. We added the following statement to the Materials and Methods section (line 530): After breeding and genotyping, mice, regardless of sex, were maintained to age 6-7 months, when transgenic animals were expected to have developed a POAG phenotype.

(3) As noted in the public review, the use of PBS for a control seems to have resulted in a slight elevation in IOP (Figure 2) as well as a reduction in outflow facility (Figure 3B) when compared to WT or Tg mice. Was this difference statistically significant? 

The differences between the sham (saline)-injected groups at any time point and untreated Tg mice did not reach statistical significance for IOP, facility, or TM cellularity and for facility, did not even show clear trends. For example, WT mice had, on average, 0.2 mmHg higher IOP and 0.6 nl/min/mmHg greater facility than the Tg group. Meanwhile on a similar scale, the long-term sham group exhibited 0.4 nl/min/mmHg higher facility compared to the Tg group. As the statistical tests indicate, these differences should be interpreted more as noise than meaningful signal. 

If so, then it should be noted as to whether the observed decrease in IOP following stem cell injection remained statistically significant when compared to these un-injected control animals. If significance was lost, then this should be appropriately noted and discussed. It is not apparently obvious why sham controls should have elevated IOP. This is a design and statistical concern.

Please refer to our response to a similar observation by Reviewer 1. We believe that comparing the treatment (cell suspension in saline) with its age-matched vehicle (saline) is the appropriate approach which maintains rigor by most directly accounting for the effects of injection. 

(4) The tonicity of the PBS used as a vehicle control was not stated and I did not see within the methods whether the stem cells were suspended using this same PBS vehicle. I assume isotonic phosphate buffered saline was used and that the stem cells were resuspended using the same sterile PBS. 

Thanks for catching this. We added “sterile PBS (1X, Thermo Fisher Scientific, Waltham, MA)” to the Methods section of the manuscript (line 567). 

With regards to using PBS as an injection control, I wonder if a better comparable control might have been to use mesenchymal stem cells that were rendered incapable of proliferating prior to intracameral injection. This, of course, addresses the unexplained mechanism(s) by which mesenchymal stem cells elicit a decrease in IOP.

This is an interesting idea, and represents another level of control. However, we explicitly chose not to use non-proliferating hAMSCs as a control, for several reasons. Firstly, a saline injection is the simplest control and in this initial study with multiple groups, we did not feel another experimental group should be added. Second, this control would not rule out paracrine effects from injected cells, which our data suggested are an important effect. Third, rendering injected cells truly non-proliferative could introduce unwanted/unknown phenotypes in these cells that would need to be carefully characterized. That being said, if an efficient method could be developed to render an entire population of these cells irreversibly non-proliferating, the reviewer’s suggestion would be worth pursuing to better understand the mechanism of TM cell therapies. 

(5) As noted in Figure 4C, TM cellular density as quantified was not altered in the sham control, so a loss of cellular density can not explain the elevated IOP with this group. Injecting viable (not determined?) mesenchymal stem cells did show, over the short term, a noted increase in TM cellular density. 

Thank you for noting this. We agree that changes in cell density do not explain the mild IOP elevation in the sham group. As the referee certainly is aware, there are multiple reasons that IOP can be elevated (changes in trabecular meshwork extracellular matrix, changes in trabecular meshwork stiffness) that are not necessarily related to cell density.  Since we do not know definitively the cause of this mild elevation, we would prefer to not speculate about it in the manuscript. 

Thanks for pointing out our omission of a statement about injected cell viability. We have now included the following statement in the Materials and Methods section (564-566): “For all the experiments where animals received hAMSC, cell count and >90% viability was verified using a Countess II Automated Cell Counter (Thermo Fisher Scientific, Waltham, MA).”

I'm confused, as clearly stated (Lines 431-432), mesenchymal stem cells accumulated close to, but not within, the TM. How is it that TM cellular density increased if these stem cells did not enter the TM? The authors may wish to clarify this distinction. Given that mesenchymal stem cells did not increase the risk of tumorigenicity, do the authors have any evidence that these cells actually proliferated post-injection or did they undergo senesce thereby displaying senescence-associated secretory phenotype as a source of paracrine support?

As the reviewer correctly noted, our observations show that hAMSCs primarily accumulated close to, but outside, the TM (likely caught up in the pectinate ligaments). Based on observations of increased TM cellularity, we think that the most likely explanation of these findings is paracrine signaling, as the reviewer suggests and which was discussed at length in the original version of the manuscript (lines 453-477). 

We agree that, despite observing little signal from hAMSCs within the TM, labeling with proliferation markers (e.g., Ki-67) and searching for co-localization with exogenous cells, and/or labeling for senescence markers would have provided more mechanistic information. This is an excellent topic for future study, which we plan to pursue, but was outside the scope of this study. 

(6) As noted in the public review, I think it is a bit of a stretch to even suggest that the findings of this study support stem cell restoration of TM function given that the model apparently did not produce TM cell dysfunction as anticipated. A restoration effect remains to be seen.

We agree and have adjusted the language accordingly. For further details, please refer to our response to Reviewer 1’s public comment.

Reviewer #3 (Recommendations for the authors):

(1) Show PCR, western blot, or immuno of angle tissue of the MYOC tg to confirm transgenic expression.

(2) Examine the IOP of mice at night.

(3) Investigate other glaucomatous features in the mice to determine if they have any of the transgenic phenotypes previously reported.

(4) Examine proliferative markers in the TM region of angles injected with stem cells.

Please see our responses to all four of these comments in the public section.

Bibliography (for this response letter only)

Bahrani Fard, M.R., Chan, J., Sanchez Rodriguez, G., Yonk, M., Kuturu, S.R., Read, A.T., Emelianov, S.Y., Kuehn, M.H., Ethier, C.R., 2023. Improved magnetic delivery of cells to the trabecular meshwork in mice. Exp. Eye Res. 234, 109602. https://doi.org/10.1016/j.exer.2023.109602

Li, G., Lee, C., Agrahari, V., Wang, K., Navarro, I., Sherwood, J.M., Crews, K., Farsiu, S., Gonzalez, P., Lin, C.-W., Mitra, A.K., Ethier, C.R., Stamer, W.D., 2019. In vivo measurement of trabecular meshwork stiffness in a corticosteroid-induced ocular hypertensive mouse model. Proc. Natl. Acad. Sci. U. S. A. 116, 1714–1722.

https://doi.org/10.1073/pnas.1814889116

Zhu, W., Gramlich, O.W., Laboissonniere, L., Jain, A., Sheffield, V.C., Trimarchi, J.M., Tucker, B.A., Kuehn, M.H., 2016. Transplantation of iPSC-derived TM cells rescues glaucoma phenotypes in vivo. Proc. Natl. Acad. Sci. 113, E3492–E3500.

Zode, G.S., Kuehn, M.H., Nishimura, D.Y., Searby, C.C., Mohan, K., Grozdanic, S.D., Bugge, K., Anderson, M.G., Clark, A.F., Stone, E.M., Sheffield, V.C., 2011. Reduction of ER stress via a chemical chaperone prevents disease phenotypes in a mouse model of primary open angle glaucoma. J. Clin. Invest. 121, 3542–3553. https://doi.org/10.1172/JCI58183

Author response:

The following is the authors’ response to the original reviews

Public reviews:

Reviewer #1:

The authors attempted to replicate previous work showing that counterconditioning leads to more persistent reduction of threat responses, relative to extinction. They also aimed to examine the neural mechanisms underlying counterconditioning and extinction. They achieved both of these aims and were able to provide some additional information, such as how counterconditioning impacts memory consolidation. Having a better understanding of which neural networks are engaged during counterconditioning may provide novel pharmacological targets to aid in therapies for traumatic memories. It will be interesting to follow up by examining the impact of varying amounts of time between acquisition and counterconditioning phases, to enhance replicability to real-world therapeutic settings.

Major strengths

· This paper is very well written and attempts to comprehensively assess multiple aspects of counterconditioning and extinction processes. For instance, the addition of memory retrieval tests is not core to the primary hypotheses but provides additional mechanistic information on how episodic memory is impacted by counterconditioning. This methodical approach is commonly seen in animal literature, but less so in human studies.

· The Group x Cs-type x Phase repeated measure statistical tests with 'differentials' as outcome variables are quite complex, however, the authors have generally done a good job of teasing out significant F test findings with post hoc tests and presenting the data well visually. It is reassuring that there is a convergence between self-report data on arousal and valence and the pupil dilation response. Skin conductance is a notoriously challenging modality, so it is not too concerning that this was placed in the supplementary materials. Neural responses also occurred in logical regions with regard to reward learning.

· Strong methodology with regards to neuroimaging analysis, and physiological measures.

·The authors are very clear on documenting where there were discrepancies from their pre-registration and providing valid rationales for why.

We thank reviewer 1 for the positive feedback and for pointing out the strengths of our work. We agree that future research should investigate varying times between acquisition and counterconditioning to assess its success in real-life applications.

Major Weaknesses

(1) The statistics showing that counterconditioning prevents differential spontaneous recovery are the weakest p values of the paper (and using one-tailed tests, although this is valid due to directions being pre-hypothesized). This may be due to a relatively small number of participants and some variability in responses. It is difficult to see how many people were included in the final PDR and neuroimaging analyses, with exclusions not clearly documented. Based on Figure 3, there are relatively small numbers in the PDR analyses (n=14 and n=12 in counterconditioning and extinction, respectively). Of these, each group had 4 people with differential PDR results in the opposing direction to the group mean. This perhaps warrants mention as the reported effects may not hold in a subgroup of individuals, which could have clinical implications.

General exclusion criteria are described on page 17. We have added more detailed information on the reasons for exclusion (see page 17). All exclusions were in line with pre-registered criteria. For the analysis, the reviewer is referring to (PDR analysis that investigated whether CC can prevent the spontaneous recovery of differential conditioned threat responses), 18 participants were excluded from this analysis: 2 participants did not show evidence for successful threat acquisition as was already indicated on page 17, and 16 participants were excluded due to (partially) missing data. We now explicitly mention the exclusion of the additional 16 participants on page 7 and have updated Figure 3 to improve visibility of the individual data points. Therefore, for this analysis both experimental groups consisted of 15 participants (total N=30).

It is true that in both groups a few participants show the opposite pattern. Although this may also be due to measurement error, we agree that it is relevant to further investigate this in future studies with larger sample sizes. It will be crucial to identify who will respond to treatments based on the principles of standard extinction or counterconditioning. We have added this point in the discussion on page 14.

Reviewer #2:

Summary:

The present study sets out to examine the impact of counterconditioning (CC) and extinction on conditioned threat responses in humans, particularly looking at neural mechanisms involved in threat memory suppression. By combining behavioral, physiological, and neuroimaging (fMRI) data, the authors aim to provide a clear picture of how CC might engage unique neural circuits and coding dynamics, potentially offering a more robust reduction in threat responses compared to traditional extinction.

Strengths:

One major strength of this work lies in its thoughtful and unique design - integrating subjective, physiological, and neuroimaging measures to capture the various aspects of counterconditioning (CC) in humans. Additionally, the study is centered on a well-motivated hypothesis and the findings have the potential to improve the current understanding of pathways associated with emotional and cognitive control. The data presentation is systematic, and the results on behavioral and physiological measures fit well with the hypothesized outcomes. The neuroimaging results also provide strong support for distinct neural mechanisms underlying CC versus extinction.

We thank reviewer 2 for the feedback and for valuing the thoughtfulness that went into designing the study.

Weaknesses:

(1) Overall, this study is a well-conducted and thought-provoking investigation into counterconditioning, with strong potential to advance our understanding of threat modulation mechanisms. Two main weaknesses concern the scope and decisions regarding analysis choices. First, while the findings are solid, the topic of counterconditioning is relatively niche and may have limited appeal to a broader audience. Expanding the discussion to connect counterconditioning more explicitly to widely studied frameworks in emotional regulation or cognitive control would enhance the paper's accessibility and relevance to a wider range of readers. This broader framing could also underscore the generalizability and broader significance of the results. In addition, detailed steps in the statistical procedures and analysis parameters seem to be missing. This makes it challenging for readers to interpret the results in light of potential limitations given the data modality and/or analysis choices.

In this updated version of the manuscript, we included the notion that extinction has been interpreted as a form of implicit emotion regulation. In addition to our discussion on active coping (avoidance), we believe that our discussion has an important link to the more general framework of emotion regulation, while remaining within the scope of relevance. Please see pages 14 and 15 for the changes. In addition to being informative to theories of emotion regulation, our findings are also highly relevant for forms of psychotherapy that build on principles of counterconditioning (e.g. the use of positive reinforcement in cognitive behavioral therapy), as we point out in the introduction. We believe this relevance shows that counterconditioning is more than a niche topic. In line with the recommendation from reviewer 2, we added more details and explanations to the statistical procedures and analyses where needed (see responses to recommendations).

Reviewer #3:

Summary:

In this manuscript, Wirz et al use neuroimaging (fMRI) to show that counterconditioning produces a longer lasting reduction in fear conditioning relative to extinction and appears to rely on the nucleus accumbens rather than the ventromedial prefrontal cortex. These important findings are supported by convincing evidence and will be of interest to researchers across multiple subfields, including neuroscientists, cognitive theory researchers, and clinicians.

In large part, the authors achieved their aims of giving a qualitative assessment of the behavioural mechanisms of counterconditioning versus extinction, as well as investigating the brain mechanisms. The results support their conclusions and give interesting insights into the psychological and neurobiological mechanisms of the processes that underlie the unlearning, or counteracting, of threat conditioning.

Strengths:

· Mostly clearly written with interesting psychological insights

· Excellent behavioural design, well-controlled and tests for a number of different psychological phenomena (e.g. extinction, recovery, reinstatement, etc).

· Very interesting results regarding the neural mechanisms of each process.

· Good acknowledgement of the limitations of the study.

We thank reviewer 3 for the detailed feedback and suggestions.

Weaknesses:

(1) I think the acquisition data belongs in the main figure, so the reader can discern whether or not there are directional differences prior to CC and extinction training that could account for the differences observed. This is particularly important for the valence data which appears to differ at baseline (supplemental figure 2C).

Since our design is quite complex with a lot of results, we left the fear acquisition results as a successful manipulation check in the Supplementary Information to not overload the reader with information that is not the main focus of this manuscript. If the editor would like us to add the figure to the main text, we are happy to do so. During fear acquisition, both experimental groups showed comparable differential conditioned threat responses as measured by PDRs and SCRs. Subjective valence ratings indeed differed depending on CS category. Importantly, however, the groups only differed with respect to their rating to the CS- category, but not the CS+ category, which suggests that the strength of the acquired fear is similar between the groups. To make sure that these baseline differences cannot account for the differences in valence after CC/Ext, we ran an additional group comparison with differential valence ratings after fear acquisition added as a covariate. Results show that despite the baseline difference, the group difference in valence after CC/Ext is still significant (main effect Group: F_(1,43)=7.364, p=0.010, η²=0.146). We have added this analysis to the manuscript (see page 7).

(2) I was confused in several sections about the chronology of what was done and when. For instance, it appears that individuals went through re-extinction, but this is just called extinction in places.

We understand that the complexity of the design may require a clearer description. We therefore made some changes throughout the manuscript to improve understanding. Figure 1 is very helpful in understanding the design and we therefore refer to that figure more regularly (see pages 6-7). We also added the time between tasks where appropriate (e.g. see page 7). Re-extinction after reinstatement was indeed mentioned once in the manuscript. Given that the reinstatement procedure was not successful (see page 9), we could not investigate re-extinction and it is therefore indeed not relevant to explicitly mention and may cause confusion. We therefore removed it (see page 12).

(3) I was also confused about the data in Figure 3. It appears that the CC group maintained differential pupil dilation during CC, whereas extinction participants didn't, and the authors suggest that this is indicative of the anticipation of reward. Do reward-associated cues typically cause pupil dilation? Is this a general arousal response? If so, does this mean that the CSs become equally arousing over time for the CC group whereas the opposite occurs for the extinction group (i.e. Figure 3, bottom graphs)? It is then further confusing as to why the CC group lose differential responding on the spontaneous recovery test. I'm not sure this was adequately addressed.

Indeed, reward and reward anticipation also evoke an increase in pupil dilation. This was an important reason for including a separate valence-specific response characterization task. Independently from the conditioning task, this task revealed that both threat and reward-anticipation induced strong arousal-related PDRs and SCRs. This was also reflected in the explicit arousal ratings, which were stronger for both the shock-reinforced (negative valence) and reward-reinforced (positive valence) stimuli. Therefore, it is not surprising that reward anticipation leads to stronger PDRs for CS+ (which predict reward) compared to CS- stimuli (which do not predict reward) during CC, but is reduced during extinction due to a decrease in shock anticipation. During the spontaneous recovery test, a return of stronger PDRs for CS+ compared to CS- stimuli in the standard extinction group can only reflect a return of shock anticipation. Importantly, the CC group received no rewards during the spontaneous recovery task and was aware of this, so it is to be expected that the effect is weakened in the CC group. However, CS+ and CS- items were still rated of similar valence and PDRs did not differ between CS+ and CS- items in the CC group, whereas the Ext group rated the CS+ significantly more negative and threat responses to the CS+ did return. It therefore is reasonable to conclude that associating the CS+ with reward helps to prevent a return of threat responses. We have added some clarifications and conclusions to this section on page 8.

(4) I am not sure that the memories tested were truly episodic

In line with previous publications from Dunsmoor et al.[1-4], our task allows for the investigation of memory for elements of a specific episode. In the example of our task, retrieval of a picture probes retrieval of the specific episode, in which the picture was presented. In contrast, fear retrieval relies on the retrieval of the category-threat association, which does not rely on retrieval of these specific episodic elements, but could be semantic in nature, as retrieval takes place at a conceptual level. We have added a small note on what we mean with episodic in this context on page 4. We do agree that we cannot investigate other aspects of episodic memories here, such as context, as this was not manipulated in this experiment.

(5) Twice as many female participants than males

It is indeed unfortunate that there is no equal distribution between female and male participants. Investigating sex differences was not the goal of this study, but we do hope that future studies with the appropriate sample sizes are able to investigate this specifically. We have added this to the limitations of this study on page 17.

(6) No explanation as to why shocks were varied in intensity and how (pseudo-randomly?)

The shock determination procedure is explained on pages 18-19 (Peripheral stimulation). As is common in fear conditioning studies in humans (see references), an ascending staircase procedure was used. The goal of this procedure is to try and equalize the subjective experience of the electrical shocks to be “maximally uncomfortable but not painful”.

Recommendations for the authors:

Reviewer #1:

Very well written. No additional comments

We thank reviewer 1 for valuing our original manuscript version. To further improve the manuscript, we adapted the current version based on the reviewer’s public review (see response to reviewer #1 public review comment 1).

Reviewer #2:

(1) I feel that more justification/explanation is needed on why other regions highly relevant to different aspects of counterconditioning (e.g., threat, memory, reward processing) were not included in the analyses.

We first performed whole-brain analyses to get a general idea of the different neural mechanisms of CC compared to Ext. Clusters revealing significant group differences were then further investigated by means of preregistered ROI analyses. We included regions that have previously been shown to be most relevant for affective processing/threat responding (amygdala), memory (hippocampus), reward processing (NAcc) and regular extinction (vmPFC). We restricted our analyses to these most relevant ROIs as preregistered to prevent inflated or false-positive findings[5]. Beyond these preregistered ROIs, we applied appropriate whole-brain FEW corrections. The activated regions are listed in Supplementary Table 1 and include additional regions that were expected, such as the ACC and insula.

(2) Were there observed differences across participants in the experiment? Any information on variance in the data such as how individual differences might influence these findings would provide a richer understanding of counterconditioning and increase the depth of interpretation for a broad readership.

We agree that investigating individual differences is crucial to gain a better understanding of treatment efficacy in the framework of personalized medicine. Specifically, future research should aim to identify factors that help predict which treatment will be most effective for a particular patient. The results of this study provide a good basis for this, as we could show that the vmPFC in contrast to regular extinction, is not required in CC to improve the retention of safety memory. Therefore, this provides a viable option for patients who are not responding to treatments that rely on the vmPFC. In addition, as noted by Reviewer 1, in both groups a few participants show the opposite pattern (see Figure 3). It will be crucial to identify who will respond to treatments based on the principles of standard extinction or counterconditioning. We have added this point in the discussion on page 14.

(3) While most figures are informative and clear, Figure 3 would benefit from detailed axis labels and a more descriptive caption. Currently, it is challenging to navigate the results presented to support the findings related to differential PDRs. A supplementary figure consolidating key patterns across conditions might also further facilitate understanding of this rather complicated result.

We have made some changes to the figure to improve readability and understanding. Specifically, we changed the figure caption to “Change from last 2 trials CC/Ext to first 2 trials Spontaneous recovery test”, to give more details on what exactly is shown here. We also simplified the x-axis labels to “counterconditioning”, “recovery test” and “extinction”. With the addition of a clearer figure description, we hope to have improved understanding and do not think that another supplemental figure is needed.

(4) Additional details on the statistical tests are needed. For example, please clarify whether p-values reported were corrected across all experimental conditions. Also, it would be helpful for the authors to discuss why for example repeated measures ANOVA or mixed-effects conditions were not used in this study. Might those tests not capture variance across participants' PDRs and SCRs over time better?

We added that significant interactions were followed by Bonferroni-adjusted post-hoc tests where applicable (see page 21). We have used repeated measures ANOVAs to capture early versus late phases of acquisition and CC/extinction, as well as to compare late CC/extinction (last 2 trials) compared to early spontaneous recovery (first 2 trials) as is often done in the literature. A trial-level factor in a small sample would cost too many degrees of freedom and is not expected to provide more information. We have added this information and our reasoning to the methods section on page 21.

Reviewer #3:

(1) Suggest putting acquisition data into the main figures. In fact many of the supplemental figures could be integrated into the main figures in my opinion.

See response to reviewer #3 public review comment 1.

(2) Include explanations for why shock intensity was varied

See response to reviewer #3 public review comment 6.

(3) Include a better explanation for the change in differential responding from training to spontaneous recovery in the CC group (I think the loss of such responding in extinction makes more sense and is supported by the notion of spontaneous recovery, but I'm not sure about the loss in the CC group. There is some evidence from the rodent literature - which I am most familiar with - regarding a loss in contextual gradient across time which could account for some loss in specificity, could it be something like this?).

See response to reviewer #3 public review comment 3.

If we understand the reviewer correctly in that the we see a loss of differential responding due to a generalization to the CS-, this would imply an increase in responding to the CS-, which is not what we see. Our data should therefore be correctly interpreted as a loss of the specific response to the CS+ from the CC phase to the recovery test. Therefore, there is no spontaneous recovery in the CC group, and also not a non-specific recovery. To clarify this we relabeled Figure 3 by indicating “recovery test” instead of “spontaneous recovery”.

(4) Is there a possibility that baseline differences, particularly that in Supplemental Figure 2C, could account for later differences? If differences persist after some transformation (e.g. percentage of baseline responding) this would be convincing to suggest that it doesn't.

See response to reviewer #3 public review comment 1.

(5) As I mentioned, I got confused by the chronology as I read through. Maybe mention early on when reporting the spontaneous recovery results that testing occurred the next day and that participants were undergoing re-extinction when talking about it for the second time.

See response to reviewer #3 public review comment 2.

(6) Page 8 - I was confused as to why it is surprising that the CC group were more aroused than the extinction group, the latter have not had CSs paired with anything with any valence, so doesn't this make sense? Or perhaps I am misunderstanding the results - here in text the authors refer back to Figure 2B, but I'm not sure if this is showing data from the spontaneous recovery test or from CC/extinction. If it is the latter, as the caption suggests, why are the authors referring to it here?

Participants in the CC group showed increased differential self-reported arousal after CC, whereas arousal ratings did not differ between CS+ and CS- items after extinction. We interpret this in line with the valence and PDR results as an indication of reward-induced arousal. At the start of the next day, however, participants from the CC and extinction groups gave comparable ratings. It may therefore be surprising why participants in the CC group do not still show stronger ratings since nothing happened between these two ratings besides a night’s sleep (see design overview in Figure 1A). We removed the “suprisingly” to prevent any confusion.

(7) I suggest that the authors comment on whether there were any gender differences in their results.

See response to reviewer #3 public review comment 5.

(8) The study makes several claims about episodic memory, but how can the authors be sure that the memories they are tapping into are episodic? Episodic has a very specific meaning - a biographical, contextually-based memory, whereas the information being encoded here could be semantic. Perhaps a bit of clarification around this issue could be helpful.

See response to reviewer #3 public review comment 4.

References

(1) Dunsmoor, J. E. & Kroes, M. C. W. Episodic memory and Pavlovian conditioning: ships passing in the night. Curr Opin Behav Sci 26, 32-39 (2019). https://doi.org/10.1016/j.cobeha.2018.09.019

(2) Dunsmoor, J. E. et al. Event segmentation protects emotional memories from competing experiences encoded close in time. Nature Human Behaviour 2, 291-299 (2018). https://doi.org/10.1038/s41562-018-0317-4

(3) Dunsmoor, J. E., Murty, V. P., Clewett, D., Phelps, E. A. & Davachi, L. Tag and capture: how salient experiences target and rescue nearby events in memory. Trends Cogn Sci 26, 782-795 (2022). https://doi.org/10.1016/j.tics.2022.06.009

(4) Dunsmoor, J. E., Murty, V. P., Davachi, L. & Phelps, E. A. Emotional learning selectively and retroactively strengthens memories for related events. Nature 520, 345-348 (2015). https://doi.org/10.1038/nature14106

(5) Gentili, C., Cecchetti, L., Handjaras, G., Lettieri, G. & Cristea, I. A. The case for preregistering all region of interest (ROI) analyses in neuroimaging research. Eur J Neurosci 53, 357-361 (2021). https://doi.org/10.1111/ejn.14954

Author response:

The following is the authors’ response to the previous reviews

Reviewer #1 (Public review):

Summary:

Audio et al. measured cerebral blood volume (CBV) across cortical areas and layers using high-resolution MRI with contrast agents in non-human primates. While the non-invasive CBV MRI methodology is often used to enhance fMRI sensitivity in NHPs, its application for baseline CBV measurement is rare due to the complexities of susceptibility contrast mechanisms. The authors determined the number of large vessels and the areal and laminar variations of CBV in NHP, and compared those with various other metrics.

Strengths:

Noninvasive mapping of relative cerebral blood volume is novel for non-human primates. A key finding was the observation of variations in CBV across regions; primary sensory cortices had high CBV, whereas other higher areas had low CBV. The measured CBV values correlated with previously reported neuronal and receptor densities.

We appreciate your recognition of the novelty of our non-invasive relative cerebral blood volume (CBV) mapping in non-human primates, as well as the observed areal variations and their correlations with neuronal and receptor densities. However, we are concerned that key contributions of our work—such as cortical layer-specific vasculature mapping and benchmarking surface vessel density estimations against anatomical ground truth—are being framed as limitations rather than significant advances in the field pushing the boundaries of current neuroimaging capabilities and providing a valuable foundation for future research. Additionally, we would like to clarify that dynamic susceptibility contrast (DSC) MRI using gadolinium is the gold standard for CBV measurement in clinical settings and the argument that “baseline CBV measurements are rare due to the complexities of susceptibility contrast” is simply not true. The limited use of ferumoxytol for CBV imaging is primarily due to previous FDA regulatory restrictions, rather than inherent methodological shortcomings.

Changes in text:

Compared to clinically used gadolinium-based agents, ferumoxytol's substantially longer half-life and stronger R₂* effect allows for higher-resolution and more sensitive vascular volume measurements (Buch et al., 2022), albeit these methodologies are hampered by confounding factors such as vessel orientation relative to the magnetic field (B₀) direction (Ogawa et al., 1993).

Weaknesses:

A weakness of this manuscript is that the quantification of CBV with postprocessing approaches to remove susceptibility effects from pial and penetrating vessels is not fully validated, especially on a laminar scale. Further specific comments follow.

(1) Baseline CBV indices were determined using contrast agent-enhanced MRI (deltaR₂*). Although this approach is suitable for areal comparisons, its application at a laminar scale poses challenges due to significant contributions from large vessels including pial vessels. The primary concern is whether large-vessel contributions can be removed from the measured deltaR₂* through processing techniques.

Eliminating the contribution of large vessels completely is unlikely, and we agree with the reviewer that ΔR₂* results likely reflect a weighted combination of signals from both large vessels and capillaries. However, the distribution of ΔR₂* more closely aligns with capillary density in areas V1–V5 than with large vessel distributions (Weber et al., 2008), suggesting that our ΔR₂* results are more weighted toward capillaries. Moreover, we demonstrated that the pial vessel induced signal-intensity drop-outs are clearly limited to the superficial layers and exhibit smaller spatial extent than generally thought (Supp. Figs. 2 and 4).

(2) High-resolution MRI with a critical sampling frequency estimated from previous studies (Weber 2008, Zheng 1991) was performed to separate penetrating vessels. However, this approach is still insufficient to accurately identify the number of vessels due to the blooming effects of susceptibility and insufficient spatial resolution. The reported number of penetrating vessels is only applicable to the experimental and processing conditions used in this study, which cannot be generalized.

Our intention was not to suggest that our measurements provide a general estimate of vessel density across the macaque cerebral cortex. At 0.23 mm isotropic resolution, we successfully delineated approximately 30% of the penetrating vessels in V1. Our primary objective was to demonstrate a proof-of-concept quantifiable measurement rather than to establish a generalized vessel density metric for all brain regions. We have consistently emphasized this throughout the manuscript, but if there is a specific point of misunderstanding, we would be happy to consider revisions for clarity.

(3) Baseline R₂* is sensitive to baseline R₂, vascular volume, iron content, and susceptibility gradients. Additionally, it is sensitive to imaging parameters; higher spatial resolution tends to result in lower R₂* values (closer to the R₂ value). Thus, it is difficult to correlate baseline R₂* with physiological parameters.

The observed correlation between R₂* and neuron density is likely indirect, as R₂* is strongly influenced by iron, myelin, and deoxyhemoglobin densities. However, the robust correlation between R₂* and neuron density, peaking in the superficial layers (R = 0.86, p < 10^-10), is striking and difficult to ignore (revised Supp. Fig. 6D-E). Upon revision, we identified an error in Supp. Fig. 6D-E, where the previous version used single-subject R₂* and ΔR₂* maps instead of the group-averaged maps. The revised correlations are slightly stronger than in the earlier version.

Given that the correlation between neuron density and R₂* is strongest in the superficial layers, we suggest this relationship reflects an underlying association with tissue cytochrome oxidase (CO) activity and cumulative effect of deoxygenated venous blood drainage toward the pial network. The superficial cortical layers are also less influenced by myelin and iron densities, which are more concentrated in the deeper cortical layers. Additional factors may contribute to this relationship, including the iron dependence of mitochondrial CO activity, as iron is an essential component of CO’s heme groups. Moreover, myelin maintenance depends on iron, which is predominantly stored in oligodendrocytes. The presence of myelinated thin axons and a higher axonal surface density may, in turn, be a prerequisite for high neuron density.

In this context, it is also valuable to note the absolute range of superficial R₂* values (≈ 6 s^-1; Supp. Fig. 6D). This variation in cortical surface R₂* is about 12-30 times larger compared to the signal changes observed during task-based fMRI (6 vs. 0.2-0.5 s^-1). This relation seems reasonable because regional increases in absolute blood flow associated with imaging signals, as measured by PET, typically do not exceed 5%–10% of the brain's resting blood flow (Raichle and Mintum 2016; Brain work and brain imaging). The venous oxygenation level is typically 60%, with task-induced activation increasing it by only a few percent. We suggest that this is ~40% oxygen extraction is reflected in the superficial R₂*. Finally, the large intercept (≈ 14.5 1/s; Supp. Fig. 6D), which is not equivalent to the water R₂* (≈ 1 1/s), suggests that R₂* is influenced by substantial non-neuron density factors, such as receptor, myelin, iron, susceptibility gradients and spatial resolution.

The R₂* values are well known to be influenced by intra-voxel phase coherence and thus spatial resolution. However, our view is that the proposed methodology of acquiring cortical-layer thickness adjusted high-resolution (spin-echo) R₂ maps poses more methodological limitations and is less practical. Notwithstanding, to further corroborate the relationship between R₂* and neuron density, we investigated whether a similar correlation exists in non-quantitative T2w SPACE-FLAIR images (0.32 mm isotropic) signal-intensity and neuron density. Using B₁ bias-field and B₀ orientation bias corrected T2w SPACE-FLAIR images (N=7), we parcellated the equivolumetric surface maps using Vanderbilt sections. Our findings showed that signal intensity—where regions with high signal intensity correspond to low R₂ values, and areas with low signal intensity correspond to high R₂ values—was positively correlated with neuron density, particularly in the superficial layers (R = 0.77, p = 10^-11; Author response image 1).This analysis confirmed the correlation with neuron density and R₂ peaks at superficial layers. However, this correlation was slightly weaker compared to quantitative R₂* (Supp. Fig. 6D), suggesting the variable flip-angle spin-echo train refocused signal-phase coherence loss from large draining vessels or that non-quantitative T2w-FLAIR images may be confounded by other factors such as B₁ transmission field biases (Glasser et al., 2022). Notwithstanding, this non-quantitative fast spin-echo with variable flip-angles approach, which is in principle less dependent on image resolution and closer to R_2,intrinsic than R₂*, yields similar findings in comparison to quantitative gradient-echo.

Author response image 1.

(A) T2w-FLAIR SPACE normalized signal-intensity plotted vs neuron density. Note that low signal-intensity corresponds to high R₂ and high neuron density, consistent with findings using ME-GRE. (B) Correlation between T2w-FLAIR SPACE and neuron density across equivolumetric layers. Notably, a similar relationship with neuron density was observed using a variable spin-echo pulse sequence as with quantitative gradient-echo-based imaging.

Changes in text:

Results:

“Because the Julich cortical area atlas covers only a section of the cerebral cortex, and the neuron density estimates are interpolated maps, we extended our analysis using the original Collins sample borders encompassing the entire cerebral cortex (Supp. Fig. 6A-C). This analysis reaffirmed the positive correlation with ΔR₂* (peak at EL2, R = 0.80, p < 10^-11) and baseline R₂* (peak at EL2a, R = 0.86, p < 10^-13), yielding linear coefficients of ΔR₂* = 102 × 10³ neurons/s and R₂* = 41 × 10³ neurons/s (Supp. Fig. 6D-G). This suggests that the sensitivity of quantitative layer R₂* MRI in detecting neuronal loss is relatively weak, and the introduction of the Ferumoxytol contrast agent has the potential to enhance this sensitivity by a factor of 2.5.”

A new paragraph was added into discussion section 4.3 corroborating the relation between R₂* and neuron density:

“Another key finding of this study was the strong correlation between baseline R₂* and neuron density (Supp. Fig. 6D, E). While R₂* is well known to be influenced by iron, myelin, and deoxyhemoglobin densities, this correlation peaks in the superficial layers (Supp. Fig. 6E), suggesting a link to CO activity and the accumulation of deoxygenated venous blood draining from all cortical layers toward the pial network. Notably, the absolute range of superficial R₂* values (max - min ≈ 6 s^-1; Supp. Fig. 6D) is approximately 12-30 times larger than the ΔR₂* observed during task-based BOLD fMRI at 3T (0.2-0.5 1/s) (Yablonskiy and Haacke 1994). Since venous oxygenation is around 60% and task-induced changes in blood flow account for only 5%–10% of the brain's resting blood flow (Raichle & Mintun, 2006), these results suggest that superficial R₂* (Fig. 1D) may serve as a more accurate proxy for total deoxyhemoglobin content (and thus total oxygen consumption), which scales with the neuron density of the underlying cortical gray matter. Importantly, superficial layers may also provide a more specific measure of deoxyhemoglobin, as they are less influenced by myelin and iron, which are more concentrated in deeper cortical layers. Additionally, smaller but direct contributors, such as mitochondrial CO density—an iron-dependent factor—may also play a role in this relationship.”

References:

Raichle, M.E., Mintun, M.A., 2006. BRAIN WORK AND BRAIN IMAGING. Annu. Rev. Neurosci. 29, 449–476. https://doi.org/10.1146/annurev.neuro.29.051605.112819

(4) CBV-weighted deltaR₂* is correlated with various other metrics (cytoarchitectural parcellation, myelin/receptor density, cortical thickness, CO, cell-type specificity, etc.). While testing the correlation between deltaR₂* and these other metrics may be acceptable as an exploratory analysis, it is challenging for readers to discern a causal relationship between them. A critical question is whether CBV-weighted deltaR₂* can provide insights into other metrics in diseased or abnormal brain states.

We acknowledge that having multivariate analysis using dense histological maps would be valuable to establish causality among these several metrics:

“To comprehensively understand the factors contributing to the vascular organization of the brain, experimental disentanglement through multivariate analysis of laminar cell types and receptor densities is needed (Hayashi et al., 2021, Froudist-Walsh et al., 2023). Moreover, employing more advanced statistical modeling, including considerations for synapse-neuron interactions, may be important for refined evaluations.”

We think the primary contributors to the brain's energy budget are neurons and receptors, as shown in several references and stated in the manuscript. To investigate relationship between neuron density and CBV, we estimated the energy budget allocated to neurons and extrapolated the remaining CBV to other contributing factors:

Changes in text:

“However, this is a simplified estimation, and a more comprehensive assessment would need to account for an aggregate of biophysical factors such as neuron types, neuron membrane surface area, firing rates, dendritic and synaptic densities (Fig. 6F-G), neurotransmitter recycling, and other cell types (Kageyama 1982; Elston and Rose 1997; Perge et al., 2009; Harris et al., 2012). Indeed, the majority of the mitochondria reside in the dendrites and synaptic transmission is widely acknowledged to drive the majority of the energy consumption and blood flow (Wong-Riley, 1989; Attwell et al., 2001).

Extrapolating cortical ΔR₂* to zero neuron density results in a large intercept (~35 1/s), corresponding to 60% of the maximum cortical CBV (57 1/s; Supp. Fig. 6F). This supports the view that the majority of energy consumption occurs in the neuropil—comprising dendrites, synapses, and axons—which accounts for ~80–90% of cortical gray matter volume, whereas neuronal somata constitute only ~10–20% (Wong-Riley, 1989). Although neuronal cell bodies exhibit higher CO activity per unit volume due to their dense mitochondrial content, these results suggest their overall contribution to the total CBV per mm³ tissue remains lower than that of the neuropil, given the latter's substantially larger volume fraction in cortical tissue.

Contrary to our initial expectations, we observed a relatively smaller CBV in regions and layers with high receptor density (Fig. 6B, D, F). This relationship extends to other factors, such as number of spines (putative excitatory inputs) and dendrite tree size across the entire cerebral cortex (Supp. Fig. 7) (Froudist-Walsh et al., 2023, Elston 2007). These results align with the work of Weber and colleagues, who reported a similar negative correlation between vascular length density and synaptic density, as well as a positive correlation with neuron density in macaque V1 across cortical layers (Weber et al., 2008).”

Variations in neurons and receptors are reflected in cytoarchitecture, myelin (axon density likely scales with neuron density and myelin inhibits synaptic connections), and cell-type composition. For example, fast-spiking parvalbumin interneurons, which target the soma or axon hillock, are well-suited for regulating activity in regions with high neuron density, whereas bursting calretinin interneurons, which target distal dendrites, are more adapted to areas with high synaptic density. These factors in turn, gradually change along the cortical hierarchy level (higher levels have thinner cortical layer IV, more complex dendrite trees and more numerous inter-areal connectivity patterns). In our view, these factors are tightly interlinked and explain the strong correlations and metabolic demands observed across different metrics.

We also agree that cortical layer imaging of vasculature in diseased or abnormal brain states is an intriguing direction for future research; however, it falls beyond the scope of the present study.

Reviewer #2 (Public review):

Summary:

This manuscript presents a new approach for non-invasive, MRI-based, measurements of cerebral blood volume (CBV). Here, the authors use ferumoxytol, a high-contrast agent and apply specific sequences to infer CBV. The authors then move to statistically compare measured regional CBV with known distribution of different types of neurons, markers of metabolic load and others. While the presented methodology captures and estimated 30% of the vasculature, the authors corroborated previous findings regarding lack of vascular compartmentalization around functional neuronal units in the primary visual cortex.

Strengths:

Non invasive methodology geared to map vascular properties in vivo.

Implementation of a highly sensitive approach for measuring blood volume.

Ability to map vascular structural and functional vascular metrics to other types of published data.

Weaknesses:

The key issue here is the underlying assumption about the appropriate spatial sampling frequency needed to captures the architecture of the brain vasculature. Namely, ~7 penetrating vessels / mm2 as derived from Weber et al 2008 (Cer Cor). The cited work, begins by characterizing the spacing of penetrating arteries and ascending veins using vascular cast of 7 monkeys (Macaca mulatta, same as in the current paper). The ~7 penetrating vessels / mm2 is computed by dividing the total number of identified vessels by the area imaged. The problem here is that all measurements were made in a "non-volumetric" manner and only in V1. Extrapolating from here to the entire brain seems like an over-assumption, particularly given the region-dependent heterogeneity that the current paper reports.

We appreciate the reviewer’s concerns regarding spatial sampling frequency and its implications for characterizing brain vasculature, which we investigated in this study. To clarify, our analysis of surface vessel density was explicitly restricted to V1 precisely due to the limitations of our experimental precision. While we reported the total number of vessels identified in the cortex, we intentionally chose not to present density values across regions in this manuscript. Although these calculations are feasible, we focused on the data directly analyzed and avoided extrapolating density values beyond the scope of our findings. Thus, we are uncertain about the suggestion that we extrapolated vessel density values across the entire brain, as we have taken care to limit our conclusions of our vessel density precision to V1.

Regarding methodology, we conducted two independent analyses of vessel density specifically in V1. The first involved volumetric analysis using the Frangi filter, while the second used surface-based analysis of local signal-intensity gradients (as illustrated in Fig. 2E and Supp. Figs. 3 and 4), albeit the final surface density analysis is performed using the ultra-high resolution equivolumetric layers. Notably, these two approaches produced consistent and comparable vessel density estimates, supporting the reliability of our findings within the scope of V1 (we found 30% of the vessels relative to the ground-truth).

Comments on revisions:

I appreciate the effort made to improve the manuscript. That said, the direct validation of the underlying assumption about spatial resolution sampling remains unaddressed in the final version of this manuscript. With the only intention to further strengthen the methodology presented here, I would encourage again the authors to seek a direct validation of this assumption for other brain areas.

In their reply, the authors stated "... line scanning or single-plane sequences, at least on first impression, seem inadequate for whole-brain coverage and cortical surface mapping. ". This seems to emanate for a misunderstanding as the method could be used to validate the mapping, not to map per-se.

We apologize for any misunderstanding in our previous response and appreciate your clarification. We now understand that you were suggesting the use of line-scanning or single-plane sequences as a method to validate, rather than map, our spatial sampling assumptions.

We agree that single-plane sequences at very high in-plane resolution (e.g., 50 × 50 × 1000 µm) have great potential to detect penetrating vessels and even vessel branching patterns. These techniques could indeed provide valuable insights into region-specific vessel density variations which could then be used to validate whole brain 3D acquisitions. However, as noted above, we have refrained from reporting vessel densities outside V1 precisely due to sampling limitations (we only found 30% of the penetrating vessels in V1, or only 2 mm²/30mm² ≈ 7% of branching vessel ground-truth, see discussion).

We acknowledge the merit of incorporating such methods to validate regional vessel densities and agree that this would be an important avenue for future research. Thank you for suggesting this point, we have briefly mentioned the advantage of single-plane EPI at discussion.

Changes in text:

“4.1 Methodological considerations - vessel density informed MRI

…anatomical studies accounting for branching patterns have reported much higher vessel densities up to 30 vessels/mm² (Keller et al., 2011; Adams et al., 2015). Further investigations are warranted, taking into account critical sampling frequencies associated with vessel branching patterns (Duverney 1981), and achieving higher SNR through ultra-high B₀ MRI (Bolan et al., 2006; Harel et al., 2010; Kim et al., 2013) and utilize high-resolution single-plane sequences and prospective motion correction schemes to accurately characterize regional vessel densities. Such advancements hold promise for improving vessel quantification, classifications for veins and arteries and constructing detailed cortical surface maps of the vascular networks which may have diagnostic and neurosurgical utilities (Fig. 2A, B) (Iadecola, 2013; Qi and Roper, 2021; Sweeney et al., 2018).”

During the revision we found a typo and corrected it in Supp. Fig. 8: Dosal -> Dorsal.

Author response:

The following is the authors’ response to the original reviews

eLife Assessment

This study provides useful findings about the effects of heterozygosity for Trio variants linked to neurodevelopmental and psychiatric disorders in mice. However, the strength of the evidence is limited and incomplete mainly because the experimental flow is difficult to follow, raising concerns about the conclusions' robustness. Clearer connections between variables, such as sex, age, behavior, brain regions, and synaptic measures, and more methodological detail on breeding strategies, test timelines, electrophysiology, and analysis, are needed to support their claims.

We appreciate the opportunity to address the constructive feedback provided by eLife and the reviewers. Below, we respond to the overall assessment and individual reviewers' comments, clarifying our experimental approach, addressing concerns, and providing additional details where necessary.

We thank the editors for highlighting the significance of our findings regarding the effects of Trio variant heterozygosity in mice. We acknowledge the feedback concerning the experimental flow and agree that clarity is paramount. To address these concerns:

(1) Connections between variables: The word limit of the initial submission constrained our ability to provide adequate details and connections between variables. We have revised the manuscript to explicitly outline and extend explanations and the relationships between sex, age, behavior, brain regions, and synaptic measures, ensuring that the rationale for each experiment and its relevance to the overall conclusions are improved.

(2) Methodological details: The Methods section of our initial submission was condensed, with key details provided in the Supplemental Methods section. We have merged all into an extended section to improve clarity. We have expanded our description of breeding strategies, test timelines, electrophysiological protocols, and data analysis methods in the revised Methods section. We believe the additions have enhanced the transparency and reproducibility of our study and ensured full support of our conclusions.

(3) Experimental flow: We have revised and extended our results, methods, and discussion sections to clarify the rationale and experimental design to guide readers through the experimental sequence and rationale.

We are confident these revisions address the concerns raised and enhance the robustness and coherence of our findings.

Public Reviews:

Reviewer #1 (Public review):

Summary:

This study explores how heterozygosity for specific neurodevelopmental disorder-associated Trio variants affects mouse behavior, brain structure, and synaptic function, revealing distinct impacts on motor, social, and cognitive behaviors linked to clinical phenotypes. Findings demonstrate that Trio variants yield unique changes in synaptic plasticity and glutamate release, highlighting Trio's critical role in presynaptic function and the importance of examining variant heterozygosity in vivo.

Strengths:

This study generated multiple mouse lines to model each Trio variant, reflecting point mutations observed in human patients with developmental disorders. The authors employed various approaches to evaluate the resulting behavioral, neuronal morphology, synaptic function, and proteomic phenotypes.

Weaknesses:

While the authors present extensive results, the flow of experiments is challenging to follow, raising concerns about the strength of the experimental conclusions. Additionally, the connection between sex, age, behavioral data, brain regions, synaptic transmission, and plasticity lacks clarity, making it difficult to understand the rationale behind each experiment. Clearer explanations of the purpose and connections between experiments are recommended. Furthermore, the methodology requires more detail, particularly regarding mouse breeding strategies, timelines for behavioral tests, electrophysiology conditions, and data analysis procedures.

We appreciate the reviewer’s recognition of the novelty and comprehensiveness of our approach, particularly the generation of multiple mouse lines and our efforts to model Trio variant effects in vivo.

Weaknesses

(1) Experimental flow and rationale and connection between variables: We have expanded on the connections between behavioral data, neuronal morphology, synaptic function, and proteomics in the Results and Discussion sections to clarify how each experiment informs the reasoning and the conclusions and to highlight the relationships between sex, age, behavior, and synaptic measures.

(2) Methodological details: Our initial Methods section was formatted to be short to fulfill word limits on the submitted version, with additional details provided in the Supplemental Methods section. We have merged our Methods and Supplemental Methods sections and expanded on our breeding strategies, test timelines, electrophysiological protocols, and data analysis. We believe these additions enhance the transparency and reproducibility of our study.

(3) Recommendations for the authors: We thank Reviewer #1 for providing several recommendations to improve our manuscript. We have addressed their comments in the revision, as detailed below, adding key experiments that bolster our findings.

Reviewer #2 (Public review):

Summary:

The authors generated three mouse lines harboring ASD, Schizophrenia, and Bipolar-associated variants in the TRIO gene. Anatomical, behavioral, physiological, and biochemical assays were deployed to compare and contrast the impact of these mutations in these animals. In this undertaking, the authors sought to identify and characterize the cellular and molecular mechanisms responsible for ASD, Schizophrenia, and Bipolar disorder development.

Strengths:

The establishment of TRIO dysfunction in the development of ASD, Schizophrenia, and Bipolar disorder is very recent and of great interest. Disorder-specific variants have been identified in the TRIO gene, and this study is the first to compare and contrast the impact of these variants in vivo in preclinical models. The impact of these mutations was carefully examined using an impressive host of methods. The authors achieved their goal of identifying behavioral, physiological, and molecular alterations that are disorder/variant specific. The impact of this work is extremely high given the growing appreciation of TRIO dysfunction in a large number of brain-related disorders. This work is very interesting in that it begins to identify the unique and subtle ways brain function is altered in ASD, Schizophrenia, and Bipolar disorder.

Weaknesses:

(1) Most assays were performed in older animals and perhaps only capture alterations that result from homeostatic changes resulting from prodromal pathology that may look very different.

(2) Identification of upregulated (potentially compensating) genes in response to these disorder-specific Trio variants is extremely interesting. However, a functional demonstration of compensation is not provided.

(3) There are instances where data is not shown in the manuscript. See "data not shown". All data collected should be provided even if significant differences are not observed.

I consider weaknesses 1 and 2 minor. While they would be very interesting to explore, these experiments might be more appropriate for a follow-up study. I would recommend that the missing data in 3 should be provided in the supplemental material.

We are grateful for the reviewer’s recognition of our study’s significance and methodological rigor. The acknowledgment of Trio dysfunction as a novel and impactful area of research is deeply appreciated.

Weaknesses:

We agree that focusing on older animals limits insights into early-stage pathophysiology. However, our goal in this study was to examine the functional impacts of Trio heterozygosity at an adolescent stage and to reveal the ultimate impact of these alleles on synaptic function. Our choice of age aligns with our objectives. Future studies of earlier developmental stages will be beneficial and complement these findings.

Functional compensation:

We tested functional compensation through rescue experiments in +/K1431M brain slices using a Rac1-specific inhibitor, NSC23766, which prevents Rac1 activation by Trio or Tiam1. Our finding that direct Rac1 inhibition normalizes deficient neurotransmitter release in +/K1431M mice strongly suggests that increased Rac1 activity drives this phenotype.

Data not shown:

We will incorporate all previously shown data into the Supplemental Materials, even when results are nonsignificant. We agree that this ensures full transparency and facilitates a more comprehensive evaluation of our findings.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) In Figure 1K-N, the lack of observed differences in +/M2145T mice across all tests raises questions about its validity as a BPD model. Furthermore, the differences in female behavior data compared to males, as shown in the Supplemental section, lack clarification-specifically, whether these variations are due to sex differences or sample size disparities, which is not discussed. Additionally, it's unclear if the same mice were used in tests K through L-N, as the reported numbers differ without explanation; if relevant, any mortality should be reported. Given the observed body weight differences, it is important to display locomotor data, despite the mention of no change in open field results. Lastly, a detailed breeding strategy and timeline for behavioral testing would enhance clarity.

We thank Reviewer 1 for recognizing these confusing points in our behavioral data and seek to add clarification in our Revision as below:

(a) We have revised the text to emphasize our goal to evaluate the impact of NDD-related Trio alleles that have discrete and measurable effects on brain development and function, and not to model specific NDDs (e.g. ASD, SCZ, or BPD). The three specific Trio mutations were chosen based on strong evidence of these mutations impairing the biochemical functions of Trio. We reasoned our approach would reveal how impairing Trio in different ways – i.e. altering protein level or GEF1/GEF2 function – and under genetic conditions (heterozygosity) that mimic those found in individuals with Trio-related disorders impacts brain development and function. The lack of behavioral phenotypes in +/M2145T mice is indeed intriguing, especially given the alterations in electrophysiology and biochemistry experiments. It remains possible that further behavioral analyses of these mice will reveal behavioral phenotypes.

(b) Given that the prevalence and clinical presentation of individuals with various NDDs are influenced by sex, it is possible that the behavioral differences we see in male versus female Trio variant mice reflect human sex difference phenotypes. We have reorganized the Figure panels to clarify these sex differences in behaviors (new Fig. 2, Supp. Fig. 2). We focused on the most significant behavioral phenotypes shared by both sexes in the main text, or in males alone, as our anatomical and electrophysiological experiments were restricted to males to reduce variation due to estrus. The observed behavioral sex differences are not likely due to sample size disparities as power analyses were performed for all experimental results to ensure adequate sample size. A comprehensive study of the mechanisms underlying these behavioral findings merits examination but is outside the scope of this study.

(c) All mice were subjected to all behavioral tests described. No sudden mortality was observed during the behavioral experiments. Outliers in post-hoc statistical analyses were removed, which explains the apparent sample size differences between behavioral tests. We have revised the Data analysis section in our Methods to include these details (Lines 216-289, 450-457).

(d) Results of the open field test have been added to the Supplemental Data (new Supp. Fig. 2) and Results (Lines 532-537)

(e) The Methods section was expanded to include more detail on the breeding strategy (Lines 98-106). A timeline for behavioral testing has also been included in the Figures to enhance clarity (new Fig. 2A).

(2) In Figure 2A-E, head width and brain weight showed significant differences, but not body weight, how come the ratio does not change? Comparing with female results in Supplementary Figure 2A-E, it does show a difference between males and females. It is essential to clarify which sex authors use in all follow-up experiments, including synapse, transmission, and plasticity. Since the males and females have different phenotypes, why do the authors focus on males only? The E plot has no data points on the bar graph. In Figure 2I, it lacks example images for all four conditions.

We greatly appreciate this Reviewer’s attention to details in our brain and body weight data and revised the manuscript to address these concerns.

(a) The ratios of head width/body weight were calculated for each individual mouse. Hence the distribution of the ratio data (old Fig. 2D; new Fig. 3D) differs from the distribution of head width or body weight data alone (old Fig. 2A, 2C, resp.; now Fig. 3A, 3C), and therefore can affect the p-value for statistical significance. The body weight of +/M2145T males is 21.217 ±0.327 g, while for WT males is 21.745 ±0.224 g, a non-significant decrease of 0.528 g (adjusted p=0.3806). These values have been added to the Fig 3. figure legend (Lines 1020-1034) for clarity.

(b) Similar to the behavioral experiments in comment (1), we observed sex differences in head width, brain weight, and body weight in Trio heterozygous variant mice compared to WT counterparts. The differences in the ratios of head width/body weight or brain weight/body weight were the same for both males and females (i.e. head width/body weight ratio is decreased in +/K1431M mice compared to WT regardless of sex, and brain weight/body weight ratio is decreased in both +/K1431M and +/K1918X mice compared to WT regardless of sex). These findings affirm the impact of Trio mutations on these phenotypes across both sexes. We have modified the text to draw more attention to this key point (Lines 554-566 and 777-801).

(c) All experiments (excluding behavior and weight data) were performed in males only to minimize the variation in spine and synapse morphology and physiological activity that can occur due to estrus. We have clarified this in the ‘Animal Work’ section of the Methods (Lines 103-106) as well as in the Figure Legends.

(d) We thank the Reviewer for pointing out Fig. 3E lacks individual data points on the bar graph. Fig. 3E has been modified to now include the brain weight/body weight ratio for each individual mouse rather than across the population, to be consistent with the calculation of head width/body weight ratio (see point 2a).

On original submission, only a representative WT image was selected due to space constraints. The figure (new Fig. 3H and 3K) and figure legend have been revised to include representative traces for all genotypes examined.

(3) In lines 315-320, "None of the Trio variant heterozygotes exhibited altered dendritic spine density on M1 L5 pyramidal neurons compared to WT mice on either apical or basal arbors (Supplementary Figure 3L, M). Electron microscopy of cortical area M1 L5 revealed that synapse density was significantly increased in +/K1918X mice compared to WT (Figure 3A, B), possibly due to a net reduction in neuropil resulting from smaller dendritic arbors." The proposed explanation does not adequately address the observed discrepancy between spine density and synapse density reported in these two experiments. A more thorough analysis is needed to reconcile these conflicting findings and clarify how these distinct measurements may relate to each other in the context of the study's conclusions.

We acknowledge the apparent discrepancy between our dendritic spine density data, which is unchanged from WT for all three Trio variant heterozygotes, and our synapse density data, which showed an increase in +/K1918X M1 L5 compared to WT. We have expanded the explanation for this discrepancy below and added this to the Discussion (Lines 802-811):

a) Because spine density can vary by dendritic branch order and distance from the soma, only protrusions from secondary dendritic arbors of M1 L5 pyramidal neurons were quantified for consistency in analyses. However, all synapses meeting criteria were quantified in EM images, regardless of where they were located along an individual neuron’s arbors. It is possible that the density and distribution of spines along other arbors are different between genotypes but was not captured in our current data.

b) +/K1918X L5 pyramidal neurons are smaller and less complex than WT neurons, especially in the basal compartment corresponding to L5 where EM images were obtained, consistent with the smaller brain size and reduced cortical thickness of +/K1918X mice. We posit that due to their smaller dendritic field size, L5 neurons pack more densely contributing to the increased synapse density observed in +/K1918X M1 L5 cortex. Consistent with this hypothesis, we observed a trend toward increased DAPI+ cell density in M1 L5 of +/K1918X neurons (Supp. Fig. 3N).

(4) In Figure 4, one potential rationale for measuring AMPAR mEPSC frequency is to infer synapse density changes. However, the findings show no frequency change in +/K1431M and +/K1918X, with an increase only in +/M2145T, which contradicts Figure 3 results indicating a trend toward increased density across variants.

This inconsistency is confusing, especially since the authors claim to follow the methodology from the study "Trio Haploinsufficiency Causes Neurodevelopmental Disease-Associated Deficits"; yet, the observed mEPSC amplitude differs significantly from that study, while the frequency remains unaffected. Additionally, the NMDAR mEPSCs reflect combined AMPAR and NMDAR responses at positive holding potentials, with peak amplitude dominated by AMPAR. This inconsistency between holding potential results is unclear, as frequency should theoretically align across negative and positive potentials. For accurate NMDAR mEPSC measurement, it would be optimal to assess amplitude 50 ms post-initial peak and, if possible, increase the holding potential to enhance the driving force given the typically low signal of NMDAR response.

We thank the Reviewer for highlighting these important points.

a) Previous work from our lab and others demonstrate that Trio regulates synaptic AMPA receptor levels, which is why we chose to focus on AMPAR-mediated evoked and miniature EPSC frequencies and amplitudes in the current study. We acknowledge Reviewer 1’s comment on seemingly contradictory results regarding AMPAR mEPSC frequency and synapse density; however, the unchanged AMPAR mEPSC frequency in +/K1431M and +/K1918X mice is consistent with our finding of unaltered dendritic spine density in these mice compared to WT (Supp. Fig. 4L,M). The differences between dendritic spine counts and synapse density is addressed in Response (3) above.

b) While synapse density changes can be inferred from AMPAR mEPSC frequency, mEPSCs are also measures of spontaneous neurotransmitter release changes especially in the absence of changes in synaptic numbers. Notably, the increased mEPSC frequency in the +/M2145T variant is linked to enhanced spontaneous release, not to spine or synapse density changes. These findings are reinforced by increase in counts of synaptic vesicles, calculated PPR changes, and estimates of the Pr and RRP from HFS train analysis. We have included these points in the Discussion (Lines 861-863).

c) While it is tempting to compare the current study to our previously published conditional Trio haploinsufficiency model, we highlight key distinctions that may underlie phenotypic differences between these two mouse models. First, our prior model used a NEX-Cre transgene to ablate one Trio allele from excitatory neurons only beginning at embryonic day 11. In contrast, our Trio variants are expressed in all cell types throughout development, akin to the genetic variants found in individuals with TRIO-related disorders. Second, the Trio variant mice in this study are on a C57BL/6 background, while the Trio haploinsufficient mice were on a mixed 129Sv/J X C57BL/6 background. These differences in the current study may explain why some measures, such as mEPSC amplitude, may not align with those from the Trio conditional haploinsufficiency model.

d) Recordings were performed using specific inhibitors to isolate AMPA and NMDA mEPSCs; these missing methodological details have now been clarified in the updated Methods section (Lines 353-360).

(5) In Supplementary Figure 4, the sample traces indicate a higher NMDA/AMPA ratio, raising the question of whether the AMPA EPSC amplitude changes, as this could reflect PSD length. In Figure 4B, the increased AMPAR mEPSC amplitude in the +/K1918X condition compared to WT suggests an enhanced postsynaptic response, yet the PSD length is reduced in Figure 3C. Can the authors provide a potential hypothesis to explain this?

We appreciate the Reviewer’s feedback. Yes, both evoked and miniature recordings indicate increased AMPAR amplitudes in the +/K1918X variants compared to WT. While PSD length is often linked to synaptic strength, the observed reduction in PSD length in EM PSD length reduction in +/K1918X synapses is small (~6% of WT) and clearly does not correlate with significant changes in synaptic strength. We also note that the whole cell recordings of mEPSCs represent input from all active synapses on the neuron, while PSD length is measured only in synapses of the L5.

(6) In Figure 4, synaptic plasticity appears to decrease to around 50% of baseline; could this reduction be attributed to LTD, or might it result from changes in pipette resistance? Additionally, is the observed potentiation due to changes in presynaptic release probability? Measuring paired-pulse ratio (PPR) before and after induction would clarify this aspect.

We thank the Reviewer for highlighting these important points.

a) We used a well-established theta burst stimulation method for LTP induction in M1 L5 pyramidal neurons. This protocol reliably evokes LTP in WT neurons, as shown in Fig. 5J and K. Both +/K1431M and +/K1918X variants exhibit a slight but discernible increase in evoked excitatory postsynaptic currents (eEPSCs), indicative of the initiation of LTP. Although this increase is smaller compared to WT, the presence of potentiation indicates that long-term depression (LTD) is an unlikely explanation for the observed reduction.

b) To rule out the influence of technical artifacts, pipette resistance was carefully monitored before and after LTP induction. Any cells exhibiting resistance changes exceeding 20% during electrophysiological recordings were excluded from the analysis, ensuring that fluctuations in pipette resistance did not confound LTP measurements. These technical details are denoted in the Methods (Lines 344-346 and 364-366).

c) The potentiation in the +/M2145T variant may stem from increased release probability (Pr) and greater synaptic vesicle availability, but is beyond the scope of this work. We agree this is an intriguing question, not only for +/M2145T but also for +/K1431M mice. Future studies should address this, ideally using models where the Trio variant is selectively introduced into the presynaptic neuron.

(7) In lines 377-380, "The +/M2145T PPR curve was unusual, with significantly reduced PPF at short ISIs, yet clearly increased PPF at longer ISI (Figure 5A, B) compared to WT." The unusual PPR observed at the 100 ms ISI appears unexpected. Can the authors provide an explanation for this anomaly? This finding could suggest atypical presynaptic dynamics or modulation at this specific interval, which may differ from typical synaptic behavior. Further insights into possible mechanisms or experimental conditions affecting this result would be valuable.

"The decreased PPF at initial ISI in +/M2145T mice correlated with increased mEPSC frequency (Fig. 4A-C), suggestive of a possible increase in spontaneous glutamate Pr." If this is the case, it raises the question of why the increased PPR at the initial ISI in +/K1431M does not correspond to the result shown in Figure 4C. This discrepancy suggests that factors beyond initial presynaptic release probability might be influencing the observed synaptic response, or that compensatory mechanisms could be affecting PPR and mEPSC frequency differently in this variant. Further clarification on the interplay between these measurements would help resolve this inconsistency.

We appreciate the Reviewer’s critical reading and genuine interest on this phenotype in +/M2145T mice.

a) The unusual shift of the PPR in +/M2145T at ISI 100ms is fascinating and will require significant additional experimentation that lies beyond the scope of this report to address. We propose it results from altered presynaptic regulators, including increased Syt3 and reduced RhoA activity. Notably, Syt3 influences calcium-dependent SV replenishment, which can cause similar PPR defects (Weingarten DJ et al., 2022); this is now included in the Discussion. (Lines 915-918).

Weingarten DJ, Shrestha A, Juda-Nelson K, Kissiwaa SA, Spruston E, Jackman SL. Fast resupply of synaptic vesicles requires synaptotagmin-3. Nature. 2022 Nov;611(7935):320-325. doi: 10.1038/s41586-022-05337-1. Epub 2022 Oct 19. PMID: 36261524.

b) Thank you for raising the concern in clarity of this statement "The decreased PPF at initial ISI in +/M2145T mice correlated with increased mEPSC frequency (Fig. 4A-C), suggestive of a possible increase in spontaneous glutamate Pr." We have edited the sentence to be more clear (Lines 701-703). First, the K1431M and M2145T variants impact different TRIO catalytic activities disrupting distinct GTPase pathways and differentially affecting presynaptic regulators, which can lead to non-overlapping phenotypes. Also, we expand our discussion that +/K1431M variant data suggest increased AMPAR numbers and fewer silent synapses (Lines 850-855), potentially increasing AMPAR mEPSC frequency and masking the expected decrease in spontaneous release (Lines 905-910). Further experiments are needed, ideally using mixed cultures with TRIO variants in presynaptic neurons with synapses on WT neurons, as minimal stimulation variance analysis in slices would be inconclusive due to its reflection of both Pr and silent synapse changes, similar to mEPSC frequency.

(8) In Figure 5, there is no evidence demonstrating that the NSC inhibitor functions specifically in the +/K1431M condition without affecting other conditions. To verify its specificity, the authors should test the NSC inhibitor's effects across other conditions in parallel, including a control group. Additionally, cumulative RRP measurements should be provided for a more comprehensive assessment of the inhibitor's impact on synaptic function.

We appreciate the Reviewer’s feedback.

a) Previous studies have shown that Rac1 activity can bidirectionally regulate synchronous release probability (Pr). We used the Rac1-specific inhibitor NSC23766 (NSC) to test how Rac1 inhibition impacted the neurotransmitter release deficits observed in +/K1431M mice. We also added control experiments testing the impact of NSC on WT slices. These new experiments are now presented in new Fig. 8 of the revised manuscript, with expanded details in the Results (Lines 737-750) and Discussion (Lines 892-900).

b) To estimate Pr and the RRP, we employed the Decay method as described by (Ruiz et al., 2011), which does not rely on cumulative EPSC plots for RRP estimation. This approach was chosen to account for the initial facilitation in these synapses and fits are done using EPSCs plotted against stimulus number. Additional details have been provided in the Methods section  (Lines 367-373).

Ruiz R, Cano R, Casañas JJ, Gaffield MA, Betz WJ, Tabares L. Active zones and the readily releasable pool of synaptic vesicles at the neuromuscular junction of the mouse. J Neurosci. 2011 Feb 9;31(6):2000-8. doi: 10.1523/JNEUROSCI.4663-10.2011. PMID: 21307238; PMCID: PMC6633039.

(9) Given the relevance to NDD, specifying the age window of the mice used is crucial. It is confusing that the synaptic function studies were conducted at P42, while the proteomic analysis was performed at P21. Could the authors clarify the rationale behind using different age points for these analyses? Consistency in age selection, or an explanation for this variation, would help in interpreting the developmental relevance of the findings.

P42 was chosen as the age as it represents young adulthood, by which time clinical features will have already presented in individuals with neurodevelopmental disorders. Our prior studies of NEX-Cre Trio^-/- mice found significant measurable differences from WT at this age, after neuronal migration, differentiation, synaptogenesis and pruning have occurred. An earlier developmental timepoint, P21, which coincides with juvenile age in mice, was chosen for proteomics studies to identify earlier changes and potentially targetable and modifiable mechanisms that could influence the phenotypes we observed in older mice. The experiments in P42 versus P21 mice were originally two independent lines of investigation that converged in the current study.

Author response:

The following is the authors’ response to the previous reviews

Recommendations for the authors:

Reviewer #1:

First, I thank the authors for clarifying some of the confusion I had in the previous comment and I appreciate the efforts the authors put into improving the quality of the manuscript. However, my concerns about the lack of novelty of the key findings are not perfectly addressed and there is no additional analysis done in this revision. Currently in this version of the manuscript, asserting that a p-value of 10-6 is close to genome-wide significance may be considered an overstatement. Further analysis focusing on finding novel and additional discovery is very necessary.

We thank the reviewer for their comments. Reviewer #2 also made a comment regarding the genomewide threshold, “However, it remains unclear why the authors found it appropriate to apply STEAM to the LAAA model, a joint test for both allele and ancestry effects, which does not benefit from the same reduction in testing burden.” The reviewers’ have correctly identified our oversight - we have amended the manuscript as follows:

(1) The abstract, “We identified a suggestive association peak (rs3117230, p-value = 5.292 x10-6, OR = 0.437, SE = 0.182) in the HLA-DPB1 gene originating from KhoeSan ancestry.”

(2) From line 233 to 239: “The R package STEAM (Significance Threshold Estimation for Admixture Mapping) (Grinde et al., 2019) was used to determine the admixture mapping significance threshold given the global ancestral proportions of each individual and the number of generations since admixture (g = 15). For the LA model, a genome-wide significance threshold of pvalue < 2.5 x 10-6 was deemed significant by STEAM. The traditional genome-wide significance threshold of 5 x 10-8 was used for the GA, APA and LAAA models, as recommended by the authors of the LAAA model (Duan et al., 2018).” 

(3) We excluded the results for the signal on chromosome 20, since this also did not reach the LAAA model genome-wide significance threshold.  

(4) From line 296 to 308: “LAAA models were successfully applied for all five contributing ancestries (KhoeSan, Bantu-speaking African, European, East Asian and Southeast Asian). However, no variants passed the threshold for statistical significance. Although no variants reached genome-wide significance, a suggestive peak was identified in the HLA-II region of chromosome 6 when using the LAAA model and adjusting for KhoeSan ancestry (Figure 3). The QQ-plot suggested minimal genomic inflation, which was verified by calculating the genomic inflation factor ( = 1.05289) (Supplementary Figure 1). The lead variants identified using the LAAA model whilst adjusting for KhoeSan ancestry in this region on chromosome 6 are summarised in Table 3. The suggestive peak encompasses the HLA-DPA1/B1 (major histocompatibility complex, class II, DP alpha 1/beta 1) genes (Figure 4). It is noteworthy that without the LAAA model, this suggestive peak would not have been observed for this cohort. This highlights the importance of utilising the LAAA model in future association studies when investigating disease susceptibility loci in admixed individuals, such as the SAC population.”

We acknowledge that our results are not statistically significant. However, our study advances this area of research by identifying suggestive African-specific ancestry associations with TB in the HLA-II region. These findings build upon the work of the ITHGC, which did not identify any significant associations or suggestive peaks in their African-specific analyses. We have included this argument in our manuscript (from lines 425 to 432):

“The ITHGC did not identify any significant associations or suggestive peaks in their African ancestryspecific analyses.  Notably, the suggestive peak in the HLA-DPB1 region was only captured in our cohort using the LAAA model whilst adjusting for KhoeSan local ancestry. This underscores the importance of incorporating global and local ancestry in association studies investigating complex multi-way admixed individuals, as the genetic heterogeneity present in admixed individuals (produced as a result of admixtureinduced and ancestral LD patterns) may cause association signals to be missed when using traditional association models (Duan et al., 2018; Swart, van Eeden, et al., 2022).”

We appreciate the comment regarding additional analyses. We acknowledge that we did not validate our SNP peak in the HLA-II region through fine-mapping due to the lack of a suitable reference panel (see lines 490 to 500). Our long-term goal is to develop a HLA-imputation reference panel incorporating KhoeSan ancestry; however, this is beyond the scope and funding allowances of this study.

Reviewer #2 (Recommendations for the authors):

The authors we think have done an excellent job with their responses and the manuscript has been substantially improved.

Thank you for taking the time to help us improve our manuscript.

Author response:

Public Reviews:

Reviewer #1 (Public review):

Summary:

This study reveals that TRPV1 signaling plays a key role in tympanic membrane (TM) healing by promoting macrophage recruitment and angiogenesis. Using a mouse TM perforation model, researchers found that blood-derived macrophages accumulated near the wound, driving angiogenesis and repair. TRPV1-expressing nerve fibers triggered neuroinflammatory responses, facilitating macrophage recruitment. Genetic Trpv1 mutation reduced macrophage infiltration, angiogenesis, and delayed healing. These findings suggest that targeting TRPV1 or stimulating sensory nerve fibers could enhance TM repair, improve blood flow, and prevent infections. This offers new therapeutic strategies for TM perforations and otitis media in clinical settings. This is an excellent and high-quality study that provides valuable insights into the mechanisms underlying TM wound healing.

Strengths:

The work is particularly important for elucidating the cellular and molecular processes involved in TM repair. However, there are several concerns about the current version.

We sincerely thank Reviewer #1 for their time and effort in evaluating and improving our study. Below, we are pleased to address the Reviewer's concerns point by point.

Weaknesses:

Major concerns

(1) The method of administration will be a critical factor when considering potential therapeutic strategies to promote TM healing. It would be beneficial if the authors could discuss possible delivery methods, such as topical application, transtympanic injection, or systemic administration, and their respective advantages and limitations for targeting TRPV1 signaling. For example, Dr. Kanemaru and his colleagues have proposed the use of Trafermin and Spongel to regenerate the eardrum.

We are grateful to the reviewer for raising this important point. While the present study primarily focuses on the mechanistic role of TRPV1 in TM repair, we agree that the mode of therapeutic delivery will be pivotal in translating these findings into clinical practice. In response, we will expand the discussion to explore possible delivery methods—such as topical application, transtympanic injection, and systemic routes—along with their respective benefits and challenges. We will also cite the work by Dr. Kanemaru and colleagues as an example of how local delivery systems may facilitate TM regeneration.

(2) The authors appear to have used surface imaging techniques to observe the TM. However, the TM consists of three distinct layers: the epithelial layer, the fibrous middle layer, and the inner mucosal layer. The authors should clarify whether the proposed mechanism involving TRPV1-mediated macrophage recruitment and angiogenesis is limited to the epithelial layer or if it extends to the deeper layers of the TM.

We apologize for any confusion caused by our previous description. In our study, we utilized Z-stack confocal imaging to capture the full thickness of the TM, as illustrated in Author response image 1 (reconstructed from the acquired Z-sections). This imaging technique allowed us to encompass all three layers of the TM entirely. Each sample was imaged using a 10X objective on an Olympus fluorescence microscope. Given the conical shape and size of the TM, we imaged it in four quadrants, acquiring approximately 30 optical sections (with a 3 µm step) per region. Each acquired images were projected and exported using FV10ASW 4.2 Viewer, then stitched together using Photoshop. The resulting Z-stack projections enabled us to visualize the distribution of macrophages, angiogenesis, and the localization of nerve fibers throughout the TM. We will include this detailed methodology in our revision to clarify any potential confusion.

Author response image 1.

Representative confocal images showing one quadrant of the TM collected from collected from CSR1F^EGFP bone marrow transplanted mouse at day 7 post-perforation. (A-B) 3D-rendered views from different angles reveal the close spatial relationship between CSF1R^EGFP cells (green) and blood vessels (red) within the TM. (C) Cross-sectional view highlights the depth-wise distribution of CSF1R^EGFP cells (green) and blood vessels (red) across the layered TM architecture. All images were processed using Imaris Viewer x64 (version 10.2.0).

Minor concerns

In Figure 8, the schematic illustration presents a coronal section of the TM. However, based on the data provided in the manuscript, it is unclear whether the authors directly obtained coronal images in their study. To enhance the clarity and impact of the schematic, it would be helpful to include representative images of coronal sections showing macrophage infiltration, angiogenesis, and nerve fiber distribution in the TM.

As noted above, we utilized Z-stack confocal imaging to capture the full thickness of the TM, enabling us to visualize structures across all three layers. This approach ensured that all layers were included in our analysis. Due to the thin and curved nature of the TM, traditional cross-sectional imaging often struggles to clearly depict the spatial relationships between macrophages, blood vessels, and nerve fibers, especially at low magnification as shown in Author response image 2. In response to the reviewer's suggestion, we will include representative coronal images in the revised manuscript to better illustrate the distribution of these structures at higher magnification.

Author response image 2.

Confocal images of eardrum cross-sections collected at day 1 (A), 3 (B), and 7 (C) post perforation to demonstrate the wound healing processes.

Reviewer #2 (Public review):

Summary:

This study examines the role of TRPV1 signaling in the recruitment of monocyte-derived macrophages and the promotion of angiogenesis during tympanic membrane (TM) wound healing. The authors use a combination of genetic mouse models, macrophage depletion, and transcriptomic approaches to suggest that neuronal TRPV1 activity contributes to macrophage-driven vascular responses necessary for tissue repair.

Strengths:

(1) The topic of neuroimmune interactions in tissue regeneration is of interest and underexplored in the context of the TM, which presents a unique model due to its anatomical features.

(2) The use of reporter mice and bone marrow chimeras allows for some dissection of immune cell origin.

(3) The authors incorporate transcriptomic data to contextualize inflammatory and angiogenic processes during wound healing.

We sincerely thank Reviewer #2 for their time and effort in improving our study and recognizing its strengths. Below, we are pleased to address the reviewer's concerns point by point.

Weaknesses:

(1) The primary claims of the manuscript are not convincingly supported by the evidence presented. Most of the data are correlative in nature, and no direct mechanistic experiments are included to establish causality between TRPV1 signaling and macrophage recruitment or function.

We appreciate Reviewer #2's perspective on the lack of molecular mechanisms linking TRPV1 signaling and macrophages. However, our data demonstrates that TRPV1 mutations significantly affect macrophage recruitment and angiogenesis. This initial study primarily focuses on the intriguing phenomenon of how sensory nerve fibers are involved in eardrum immunity and wound healing, an area that has not been clearly reported in the literature before. We believe that further research is necessary to explore this topic in greater depth.

(2) Functional validation of key molecular players (such as Tac1 or Spp1) is lacking, and their roles are inferred primarily from gene expression data rather than experimentally tested.

Although we have identified the TAC1 and SPP1 signals as potentially important for TM wound healing for the first time, we agree with the Reviewer's view regarding the lack of molecular mechanisms explored in this study. We have not yet tested the downstream signaling pathways, but we plan to investigate them in a series of future studies. As this is an early report, we will continue to explore these signals and their potential clinical applications based on our initial findings moving forward.

(3) The reuse of publicly available scRNA-seq data is not sufficiently integrated or extended to yield new biological insights, and it remains largely descriptive.

We appreciate Reviewer #2 for highlighting this point. Leveraging publicly available scRNA-seq databases and established analysis pipelines not only saves time and resources—my lab recently collected macrophages from the eardrums of postnatal P15 mice, with each trial requiring 20 eardrums from 10 animals to obtain a sufficient number of cells—but also allows researchers to build on previous work and focus on new biological questions without the need to repeat experiments. A prior study conducted by Dr. Tward and his team utilized scRNA-seq data to make initial discoveries related to eardrum wound healing, primarily focusing on epithelial cells rather than macrophages. We are building on their raw data to uncover new biological insights regarding macrophages, even though we have not yet tested the unidentified signals, which we believe will be valuable to our peers.

(4) The macrophage depletion model (CX3CR1CreER; iDTR) lacks specificity, and possible off-target or systemic effects are not addressed.

We agree with reviewer #2, although macrophage depletion model used in our study is a standard and well-used animal model (Shi, Hua et al. 2018), which has been used by many other laboratories, it is important to note that any macrophage depletion model may have potential issues. We will discuss this in our revision.

(5) Several interpretations of the data appear overstated, particularly regarding the necessity of TRPV1 for monocyte recruitment and wound healing.

We thank the reviewer for pointing this out. We will revise our manuscript where it is overstated accordingly.

(6) Overall, the study appears to apply known concepts - namely, TRPV1-mediated neurogenic inflammation and macrophage-driven angiogenesis - to a new anatomical site without providing new mechanistic insight or advancing the field substantially.

Although our study may not seem highly innovative at first glance, it reveals a previously unknown role of the TRPV1 pain signaling pathway in promoting eardrum healing for the first time. This healing process includes the recruitment of monocyte-derived macrophages and the formation of new blood vessels (angiogenesis). While this process has been documented in other organs, most research on macrophage-driven angiogenesis has been conducted using in vitro models, with very few studies demonstrating this process in vivo. Our findings could lead to new translational opportunities, especially considering that tympanic membrane perforation, along with damage-induced otitis media and conductive hearing loss, are common clinical issues affecting millions of people worldwide. Targeting TRPV1 signaling could enhance tympanic membrane immunity, improve blood circulation, promote the repair of damaged tympanic membranes, and ultimately prevent middle ear infections—an idea that has not been previously proposed.

Overall:

While the study addresses an interesting topic, the current version does not provide sufficiently strong or novel evidence to support its major conclusions. Additional mechanistic experiments and more rigorous validation would be necessary to substantiate the proposed model and clarify the relevance of the findings beyond this specific tissue context.

We greatly thank the two reviewers for their helpful critiques to improve our study. We especially thank the Section Editors for their insightful and constructive comments on this initial study.

References:

Shi, J., L. Hua, D. Harmer, P. Li and G. Ren (2018). "Cre Driver Mice Targeting Macrophages." Methods Mol Biol 1784: 263-275.

Author response:

We are grateful to the reviewers for their extensive and constructive feedback. In large the three reviewers noted the following main points:

(1) The overall evidence for any rhythmicity in this data is not ‘very strong’.

We do agree and will tone down the conclusions accordingly. However, as one of the reviewers noted, a qualitative interpretation of the specific statistical results remains somewhat vague and speculative by necessity.

(2) The differences between the results for the individual experiments are generally small. Yet, the same reviewer also asks for speculations as to how differences between experiments can be interpreted.

We will consider these, but also note that a clear demonstration of the robustness of specific effects requires the replication of individual experiments in a separate experiment.

(3) A clear-cut interpretation of the current experimental design in the context of continuous listening and true vigilance tasks remains difficult. This makes the interpretation and generalization of the results difficult.

We do agree in principle, but also note that task designs very widely in previous work, which may be one reason for why there is no clear consensus on the existence or absence of a rhythmic mode of listening. We will consider specific suggestions for future work to be included in the revision.

(4) The adjustment of task difficulty in the present task design may pose a challenge. Reviewers also suggest analyzing potential rhythmicity in this task difficulty parameter.

We will consider this for the revision.

(5) A more clear-cut interpretation of what potential differences in the rhythmicity of sensitivity and bias would mean should be included.

We will provide this in the revision.

(6) The study should provide a stronger conceptual framework both for the source of "rhythmic modes" and why one may expect differences between ears.

In large this has been put forward by many previous studies testing and reporting rhythmicity in auditory tasks.  Rhythmicity is pervasive in neural activity, but whether and how this relates to behavioral data remains less clear. These points will be clarified in a revision.

(7) Parallels to work in the visual domain by Fiebelkorn, Landau & Fries should be included.

We will discuss similarities and differences between studies on perceptual rhythmicity in the visual and auditory domains.

Author response:

The following is the authors’ response to the current reviews.

We again thank you for the positive and constructive feedback on our manuscript, and for highlighting its contributions to understanding the role of CARD8 in viral protease-triggered sensing of viral spread, and the potential impact of our findings on chronic inflammation and immune activation. We agree that it will be important for future work to address whether or not HIV-1 protease-triggered CARD8 inflammasome activation contributes to chronic inflammation in PLWH who are receiving ART.

In response to the question about the baseline level of IL-1β in Fig. 4D, the figure below shows the mock condition for the CD4+ T cell:MDM coculture. We had done this control in parallel with the data presented in the submitted figure. Levels of IL-1β during HIV-1 infection are increased over background (i.e., mock infection). We note that for donor G the IL-1β concentration is below the limit of detection for this assay. Thus, it remains possible that other inflammasomes contribute modestly during cell-to-cell transmission of HIV-1; however, incomplete knockout of CARD8 in a minority of cells may also contribute to the observed levels of IL-1β in response to HIV-1 infection. Nonetheless, collectively, our data strongly supports the role for CARD8 in HIV-1 protease-triggered inflammasome activation.

The following is the authors’ response to the original reviews.

Joint Public Review:

Following up on their previous work, the authors investigated whether cell-to-cell transmission of HIV-1 activates the CARD8 inflammasome in macrophages, an important question given that inflammasome activation in myeloid cells triggers proinflammatory cytokine release. The data support the idea that CARD8 is activated by the viral protease and promotes inflammation. However, time-course analyses in primary T cells and macrophages and further information on the specific inflammasome involved would further increase the significance of the study.

Strengths:

The manuscript is well-written and the data is of good quality. The evidence that CARD8 senses the HIV-1 protease in the context of cell-to-cell transmission is important since cell-to-cell transmission is thought to play a key role in viral spread in vivo, and inflammation is a major driver of disease progression. Clean knockout experiments in primary macrophages are a notable strength and the results clearly support the role of CARD8 in protease-dependent sensing of viral spread and the induction of IL1β release and cell death. The finding that HIV-1 strains are resistant to protease inhibitors differ in CARD8 activation and IL1β production is interesting and underscores the potential clinical relevance of these results.

Weaknesses:

One weakness is that the authors used T cell lines which might not faithfully reflect the efficiency of HIV-1 production and cell-cell transfer by primary T cells. To assess whether CARD8 is also activated by protease from incoming viral particles earlier time points should be analyzed. Finally, while the authors exclude the role of NLRP3 in IL-1b and the death of macrophages it would be interesting to know whether the effect is still Gasdermin D dependent.

Recommendations for the authors

(1) Co-culture assay should also be done between primary CD4 cells and primary MDMs, because T-cell lines produce much more viruses, and the efficiency of cell-tocell transmission might be dramatically different in primary cells compared to cell lines.

We have now added data from experiments using infected primary CD4 cells as the donor cells in cell-to-cell HIV-1 transmission to MDMs in new Figure 4. The results largely phenocopy the SUPT1:MDM coculture in that we observe inflammasome activation after co-culture of HIV-infected primary T cells with primary MDMs. We find that this inflammasome activity induced by the CD4:MDM cell-to-cell transmission is abrogated by knockout of CARD8 in the MDMs or treatment of HIV protease inhibitor lopinavir (LPV) or caspase 1 inhibitor VX765, suggesting that this activation is dependent on CARD8, HIV protease, and caspase 1. Additionally, the signal persists in the presence of reverse transcriptase inhibitor nevirapine (NVP), suggesting that the incoming protease is driving activation.

(2) For all co-culture experiments, supernatants were collected at 48 or 72 hours. Since CARD8 activation is expected to be driven by incoming viral particles without RT, they should measure cytokine production at much earlier time points. 2-3 days co-culture raises concerns. Ideally, the authors can provide a time-course.

We have now added a time course of the SUPT1:MDM coculture from 3 unique donors taken at 4, 24, 48, and 72 hours post coculture in the presence or absence of reverse transcriptase inhibitor (see new Figure 3B) as well as for the primary CD4 cells to MDM co-culture (see new Figure 4B). We detect IL-1β at the 24hour time point (and later), but not at the 4-hour time point which is slower than what was detected by direct cell-free infection (Kulsuptrakul et al., 2023). However, we still hypothesize that this is driven by active incoming viral protease because the signal is not abrogated by a reverse transcriptase inhibitor, which indicates that de novo protease production is not necessary. We also observed that IL-1β levels do not increase after plateauing 24h after establishing the co-culture, suggesting that secondary infection does not further amplify inflammasome activation. We now speculate on this in the Discussion.

(3) A potential confounder in the data in Figure 4 is that despite rightly including the cognate adaptations in the Gag cleavage sites with the PI-R protease mutants, some of these viruses still display Gag processing defects. Can the authors disentangle the potency of PR mutant cleavage with either reduced cell entry or reduced protease availability due to processing defects in the incoming virions?

The reviewer is correct that although the western blot with the p24^gag antibody suggests that Gag is processed, we cannot rule out that other variables do not contribute to the observed difference in CARD8 inflammasome activation. For example, PI-R clones relative to the LAI strain may have distinct protease substrate specificity, variable efficiency/kinetics in viral assembly, gag dimerization, and other factors may ultimately influence CARD8 inflammasome activation. We have updated the text to reflect these possibilities. Nonetheless, this argument does not change the conclusion that CARD8 inflammasome activation is affected by protease mutations acquired during drug resistance.

(4) There is considerable donor variation in the macrophages (unsurprising) but can the authors correlate this with CARD8 expression and are there any off-target effects on macrophage permissivity to HIV-1 infection?

We have now considerably increased the number of primary cell donors from the first submission (see Author response table 1 below). We find that the non-responsive donor presented in the first submission is aberrant since all others do respond to a greater or lesser degree (Figure 3, Figure 4). However, the reviewer may be correct that the particular aberrant donor MDMs were poorly infected. We also note that despite donor variability in the degree of activation (IL-1β secretion) from cocultures with HIV_BaL-infected SUPT1 cells, HIV-induced activation is comparable to the activation induced by VbP (see new Figure 3–figure supplement 1B). We do not see a notable difference in CARD8 expression between donors. Nonetheless, with the added number of primary cell donors, the data are consistent with a role of primary MDMs from nearly all donors in supporting a CARD8-dependent, HIV-protease dependent inflammasome response after co-culture with infected T cells. We have left in data from all of the donors so that readers can appreciate the variability among primary cells.

Author response table 1.

In addition, to address the reviewer concerns about off-target effects of the sgRNAs on macrophage permissivity, we assessed our CD4:MDM cocultures for percent infectivity via intracellular p24^gag in AAVS1 vs CARD8 KO MDMs and we observed no significant difference in infectivity in AAVS1 vs CARD8 KO MDMs (see Author response image 1 of MDMs after co-culture with T cells that is not affected any potential off-target effects of the sgRNAs.

Author response image 1.

Equivalent infection in AAVS1 vs CARD8 KOMDMs. AAVS1 or CARD8 KO from donor 12 were cocultured with mock or HIV infected CD4 T cells as described in Figure 4D for 72 hours then assessed for HIV infection of the MDMs by washing away CD4 T cells, harvesting MDMs, and staining attached MDMs for intracellular p24^gag for flow cytometry analysis. Datasets represent mean ± SD (n=2 technical replicates from one donor). One-way ANOVA with Dunnett’s test using GraphPad Prism 10. ns = not significant, *p<0.05,**p<0.01, ***p<0.001, ****p<0.0001.

(5) The authors suggest that NLRP3 is unlikely to be the mediator of IL-1b and cell death in the macrophages. Is this death still GSDMDdependent, what other NLRs are expressed in this system and does it make a difference what PAMP you use to prime the response?

We have now added additional data in support of the conclusion that NLRP3 is not a mediator of the IL-1β secretion in the infected SUPT1 cells to primary MDMs coculture. In addition to using an NLRP3 inhibitor, we have now also made NLRP3 KOs MDMs and used these in the coculture experiments which show that the IL-1β secretion after coculture of infected SUPT1 cells and primary MDMs is mediated by CARD8 and not NLRP3 because the signal is abrogated by CARD8 knockout, but not by NLRP3 knockout. This new data is shown in Figure 3C and D.

To assess the role of GSDMD, we treated SUPT1:MDM cocultures with disulfiram, a GSDMD inhibitor (Hu et al., 2020). Disulfiram treatment abrogated IL-1β secretion, suggesting that this activation is indeed GSDMD-mediated (see Author response image 2 below). We choose not to include the disulfiram result in the final manuscript since we have not ruled out cytotoxic effects of the drug.

There are likely other NLRs expressed in primary MDMs; however, since inflammasome activation is completely absent in the CARD8 KO MDMs, we infer that CARD8 is the main inflammasome-forming sensor in this system. However, we cannot rule out the possibility of other innate sensors being activated downstream of CARD8 or under different differentiation conditions.

To address the concern that alternative priming affects CARD8 activation, we compared pre-treatment of cells with Pam3CSK4 or lipopolysaccharide (LPS) in the presence or absence of HIV protease inhibitor and reverse transcriptase inhibitor. Regardless of the priming agent used, we observed HIV protease-dependent activation that persisted in the presence of reverse transcriptase inhibitor, suggesting that CARD8 is the main sensor under LPS and Pam3CSK4 priming (new Figure 3–figure supplement 1A).

Author response image 2.

Inflammasome activation following cell-to-cell HIV infection is mediated by GSDMD. SUPT1-CCR5 cells were either mock-infected or infected with HIV-1_NL4.3BaL for 20 hours before coculturing with MDMs in either the presence or absence of GSDMD inhibitor disulfarim (25μM). Cocultures were harvested 24 hours later to assess (left) IL-1β secretion via IL-1 reporter assay and (right) cell viability via CellTiter-Glo® assay. Viability was calculated by normalizing to relative luminescence units in the mock untreated control. Dotted line indicates limit of detection (LoD). Dashed line indicates 100% viability as determined by untreated mock control. Datasets represent mean ± SD (n=2 technical replicates for one donor). Two-way ANOVA with Sidak’s test (using GraphPad Prism 10. ns = not significant, *p<0.05,**p<0.01, ***p<0.001, ****p<0.0001.

Minor points

(1) In Figure 1, the authors should clarify whether LAI or LAI-VSV-G was used.

Wild-type virus (LAI strain) was used in Figure 1. This has now been clarified in the figure legend.

(2) In Figure 1, the fraction of infected cells without DEAE was ~20% in both WT and CARD8 KO THP-1, suggesting somewhat efficient viral entry even in the absence of DEAE. How do the authors reconcile this with the lack of IL-1β production? The increase in infection observed in WT THP-1 +DEAE was overall modest (from ~20% to 25-30%) compared to the dramatic difference in IL-1β production. Can they provide more evidence or discuss how DEAE might be impacting cytokine production? If differences in viral entry are the explanation for differences in inflammasome activation, then they should be able to overcome this by using virus at a higher MOI in the absence of DEAE. Experiments proposed in Figure 1 +/- DEAE should be repeated using a range of MOI for LAI and showing the corresponding percent infection in THP-1 cells (which is not shown in Figure S2 for LAI-VSVG).

We hypothesize that the lack of IL-1β production without DEAE is likely due to an insufficient amount of incoming viral protease to induce CARD8 activation. Though the increase in infection with DEAE is modest by intracellular p24^gag at 24 hours post infection, we infer that intracellular p24^gag may be largely underestimating the actual increase in viral efficiency achieved with DEAE (now in Supplemental Note). We have also updated Figure S2 (now Figure 2–figure supplement 1) legend to include the percent infection for HIV-1_LAI and HIV-1_LAI-VSVG infections. We agree that activation in the absence of DEAE could be overcome by infecting with a more concentrated viral stock to increase the MOI. Indeed, our decision to use the cell-to-cell transmission model achieves this in a more physiologic context.

(3) In Figure S1, the authors point out that RT-activity in the supernatants was similar in the cell-free vs. cell-to-cell model. While in the transwell system THP-1 cells are the only cells capable of producing new virions, how are they able to differentiate viral production from sup-T1 vs. THP-1 in the cell-to-cell system? At a minimum, they should provide some data on the observed RT activity in matching wells containing the same number of infected sup-T1 cells utilized in coculture experiments.

We think this may have been a misinterpretation. In Figure S1 (now Figure 1B, right), we compare the amount of virus available in the lower chamber of the transwell versus the cell-to-cell condition. We are not comparing cell-free to cell-to-cell infection. We have changed the text and figure title to clarify this point.

(4) Can the authors provide additional comments on the lack of IL-1β release in donor C in Figure 3? The donor did not produce IL-1β in response to VbP or HIV, although the WB for CARD8 appears similar to the other two donors.

We have now tested MDMs from additional donors and continue to find a range of IL-1β secretion after the coculture. However, donor C is aberrant since each of the other donors had detectable IL-1β secretion in response to VbP and HIV-1 to greater or lesser extents. Nonetheless, we have included additional donors summarized in the table above corresponding to major comment #4.

(5) For Figure 3, can the authors provide information on the fraction of MDMs that were infected after coculture with sup-T1 cells? Why didn't the authors measure cell death in MDMs?

It is difficult to measure the fraction of MDMs infected or dying in the cocultures since it is hard to separate signal from the T cells. Although it would be possible to do so, in this manuscript, we instead prefer to focus on the potential contribution of CARD8 inflammasome activation in exacerbating chronic inflammation in response to HIV rather than the depletion of macrophages.

(6) In Figure 4, did the authors introduce the mutations associated with PI resistance into the same LAI backbone? If not, this is not a fair comparison, as viral protein expression levels were not at the same level, indicated in Figure 4A. Additionally, such comparison will be further strengthened by using cells other than 293T cells for the coculture assay.

No, we did not introduce these mutations into LAI, since they were already in an NL4.3 backbone and NL4.3 and LAI differ by only 1 amino acid in protease. We have updated Table S1 to report this amino acid difference. We also note that in our previous manuscript we tested much more diverse proteases such as a clade A HIV-1, HIV-2, and SIVs and find comparable CARD8 cleavage to LAI.

Additions not requested by Reviewers:

THP-1 characterization

In our previous work, we noticed that different “wildtype” THP-1 lines behaved uniquely in response to DEAE-dextran. In particular, we observed inflammasome activation in response to DEAE-dextran alone at the concentration used for spinoculations (20μg/mL), whereas the other THP-1 line did not. Thus, we performed STR profiling on each THP-1 cell line and determined that the THP-1 cells used in our studies (JK THP1s) are distinct from THP-1 cells from ATCC at 3 different loci. This data is now included in the Supplemental Note (Figure A1). Please note that all data in this and the accompanying manuscript were performed in JK THP-1 cells.

Whole plasmid sequencing of the PI-resistant HIV clones

Since preprint submission, we have done whole plasmid Oxford Nanopore sequencing on the PI-resistant HIV clones obtained from the NIAID HIV/AIDS Specimen Repository Program. Of note, there were a handful of previously unreported mutations included in these plasmid stocks within protease. We have updated Table S1 to include an additional column titled “Additional amino acid changes in HIV^PR relative to NL4.3.”

References

Hu JJ, Liu X, Xia S, Zhang Z, Zhang Y, Zhao J, Ruan J, Luo X, Lou X, Bai Y, Wang J, Hollingsworth LR, Magupalli VG, Zhao L, Luo HR, Kim J, Lieberman J, Wu H. 2020. FDA-approved disulfiram inhibits pyroptosis by blocking gasdermin D pore formation. Nat Immunol 21:736–745. doi:10.1038/s41590-020-0669-6

Kulsuptrakul J, Turcotte EA, Emerman M, Mitchell PS. 2023. A human-specific motif facilitates CARD8 inflammasome activation after HIV-1 infection. eLife 12:e84108. doi:10.7554/eLife.84108

Author Response:

eLife assessment

This is a valuable initial study of cell type and spatially resolved gene expression in and around the locus coeruleus, the primary source of the neuromodulator norepinephrine in the human brain. The data are generated with cutting-edge techniques, and the work lays the foundation for future descriptive and experimental approaches to understand the contribution of the locus coeruleus to healthy brain function and disease. However, due to small sample size and the need for additional confirmatory data, the data only incompletely support the main conclusions presented here. With the strengthening of the analyses, this paper, and the associated web application, will be of great interest to neuroscientists working on arousal-based behaviors and neurological and neuropsychiatric phenotypes.

Thank you for the assessment and comments. Overall, the majority of the issues raised by the reviewers relate either directly or indirectly to limitations of the sample size that precluded further optimization of protocols and expansion of the dataset. We fully acknowledge the limited sample size in this dataset and aim to be transparent about the limitations of the study. This is the first report of snRNA-seq and spatially-resolved transcriptomics in the human locus coeruleus (LC). The LC is a very small nucleus, located deep within the brainstem, which is extremely challenging to study due to its small size, difficult to access location, and the very small number of norepinephrine (NE) neurons located within the nucleus, which were of prime interest for this study. We note that this study represents our initial attempt to molecularly and spatially characterize cell types within the human LC. We note that we did not have significant, established funding from extramural sources dedicated to this study, and tissue resources for the LC are difficult to ascertain, contributing to the small sample size in this initial study. We acknowledge that there are limitations in sample size as well as data quality. Findings from this study will be used to inform, improve, and optimize future and ongoing experimental design, as well as technical and analytical workflows for larger-scale studies. As brought up by one of the reviewers, this field is still in its infancy -- pilot experimentation in new brain regions is labor-intensive and these sequencing approaches remain costly. Moreover, due to the small size and difficulties in dissecting, tissue resources from the human brain in this area are a highly limited resource. Hence, notwithstanding limitations, in our view it is important to release the data for community access at this time. Specific responses to the reviewers’ comments are provided point-by-point in the following sections.

Reviewer #1 (Public Review):

Weber et al. collect locus coeruleus (LC) tissue blocks from 5 neurotypical European men, dissect the dorsal pons around the LC and prepare 2-3 tissue sections from each donor on a slide for 10X spatial transcriptomics. […] The authors transparently present limitations of their work in the discussion, but some points discussed below warrant further attention.

Specific comments:

1) snRNAseq:

a. Major concerns with the snRNAseq dataset are A) the low recovery rate of putative LC-neurons in the snRNAseq dataset, B) the fact that the LC neuron cluster is contaminated with mitochondrial RNA, and C) that a large fraction of the nuclei cannot be assigned to a clear cell type (presumably due to contamination or damaged nuclei). The authors chose to enrich for neurons using NeuN antibody staining and FACS. But it is difficult to assess the efficacy of this enrichment without images of the nuclear suspension obtained before FACS, and of the FACS results. As this field is in its infancy, more detail on preliminary experiments would help the reader to understand why the authors processed the tissue the way they did. It would be nice to know whether omitting the FACS procedure might in fact result in higher relative recovery of LC-neurons, or if the authors tried this and discovered other technical issues that prompted them to use FACS.

Thank you for these comments. We agree these are valid concerns in assessing the data quality and validity of the findings from the snRNA-seq dataset. We will respond to these concerns here to the best of our ability, but in some cases, we do not have definitive answers since comparison data are not yet available for this region. In particular, we were limited in resources for this initial study -- some of the results of the study and issues that we identified in attempting to molecularly profile cells in the human LC were surprising to us, and we intend to generate additional samples and troubleshoot these issues to improve data quality and increase recovery in future work. However, these experiments are (i) expensive, (ii) time- and labor-intensive, and (iii) the tissue for this region is limited and difficult to ascertain. Given the extremely small size of the LC, the tissue resource is quickly depleted. For this study, we had fixed resources and made best-guess decisions on how to proceed with the experimental design, based on our experience with snRNA-seq in other human brain regions (Tran and Maynard et al. 2021). However, the LC is a unique region, and our experiences with this dataset will guide us to make technical adjustments in future studies. Due to the limitations in the tissue resources and the lack of data currently available to the community, we wanted to share these results immediately while acknowledging the limitations of the study as we work to increase our resource availability to expand molecular and spatial profiling studies in this region of the human brain.

Regarding the reviewer’s concern that our choice to use FANS to enrich for neurons could have potentially led to more damage and contributed to the low recovery rate of LC-NE neurons and the mitochondrial contamination -- we do not have a definitive answer to this question, since we did not perform a direct comparison with non-sorted data. As noted above, our limited tissue resource dictated that we could not do both. We made the decision to enrich for neurons based on our previous experience with identifying relatively rare populations in other brain regions (e.g. nucleus accumbens and amygdala; Tran and Maynard et al. 2021). Based on this previous work, our rationale was that without neuronal enrichment, we could potentially miss the LC-NE population, given the relative scarcity of this neuronal population. The low recovery rate and relatively lower quality / contamination issues may be due to technical issues that lead to LC-NE neurons being more susceptible to damage during nuclear preparation and sorting. We agree that directly comparing to data prepared without NeuN labeling and sorting is reasonable, as the additional perturbations may indeed contribute to cell damage. As mentioned in the discussion, we do not have a definitive answer to the reasons for increased mitochondrial contamination and we suspect that multiple technical factors may contribute -- including the relatively large size and increased fragility of LC-NE neurons. We agree that systematically optimizing the preparation to attempt to increase recovery rate and decrease mitochondrial contamination are important avenues for future work.

b. It is unclear what percentage of cells that make up each cluster.

We will add this information in the clustering heatmaps or as a supplementary plot in a revised version of the manuscript.

c. The number of subjects used in each analysis was not always clear. Only 3 subjects were used for snRNAseq, and one of them only yielded 4 LC-nuclei. This means the results are essentially based on n=2. The authors report these numbers in the corresponding section, but the first sentence of the results section (and Figure 1C specifically!) create the impression that n=5 for all analyses. Even for spatial transcriptomics, if I understood it correctly, 1 sample had to be excluded (n=4).

This is correct. We will update the figures and text in a revised version of the manuscript to make this limitation (small sample size) more clear, and to further emphasize that the intention of this study is to provide initial data to help determine next steps and best practices for a larger scale and more comprehensive study on this region, especially given the limited availability of tissue resources and currently limited data resources available for this region.

2) Spatial transcriptomics:

a. It is not clear to me what the spatial transcriptomics provides beyond what can be shown with snRNAseq, nor how these two sets of results compare to each other. It would be more intuitive to start the story with snRNAseq and then try to provide spatial detail using spatial transcriptomics. The LC is not a homogeneous structure but can be divided into ensembles based on projection specificity. Spatial transcriptomics could - in theory - offer much-needed insights into the spatial variation of mRNA profiles across different ensembles, or as a first step across the spatial (rostral/caudal, ventral/dorsal) extent of the LC. The current analyses, however, cannot address this issue, as the orientation of the LC cannot be deduced from the slices analyzed.

We understand the point of the reviewer. However, we structured the manuscript in this format due to our aims of creating a data resource for the community as well as being transparent about the limitations of our study. Our experiments began with the spatial experiments on the tissue blocks because this (i) helped orient ourselves to the region, and (ii) provided guidance for how best to score the tissue blocks for the snRNA-seq experiments to maximize recovery of LC-NE neurons. Therefore, we also decided to present the results in this sequence.

The spatial data also provides more information in that the measurements are from nuclei, cytoplasm, and cell processes (instead of nuclei only). This is one of the main differences / advantages between the platforms at this level of spatial resolution. As noted above, we were also working with a finite tissue resource -- if we ran snRNA-seq first and captured no neurons, the tissue block would be depleted. Due to the logistics / thickness of the required tissue sections for Visium and snRNA-seq respectively, running Visium first allowed us to ensure that we could collect data from both assays.

Regarding a point raised below on why we only ran snRNA-seq on a subset of the donors -- this was due to resource depletion and not enough available tissue remaining on the tissue blocks to run the assay. We have conducted extensive piloting in other brain regions on the amount (mg) of tissue that is needed from various sized cryosections, and the LC is particularly difficult since these are small tissue blocks and the extent of the structure is small. Hence, in some of the subjects, we did not have sufficient tissue available for the snRNA-seq assay.

We agree with the reviewer that spatial studies could, in future work, offer needed and important information about expression profiles across the spatial axes (rostral/caudal, ventral/dorsal) of the LC. Our study provides us with insight about optimizing the dissections for spatial assays, as well as bringing to light a number of technical and logistical issues that we had not initially foreseen. For example, during the course of this study and parallel, ongoing work in other small, challenging brain regions, we have now developed a number of specialized technical and logistical strategies for keeping track of orientation and mounting serial sections from the same tissue block onto a single spatial array, which is extremely technically challenging. We are now well-prepared for addressing these issues in future studies with larger numbers of donors and samples, e.g. spaced serial sections across the extent of the LC to make these types of insights. Due to the rarity of the tissue, limited availability of information in this region, and high expense of conducting these studies, we want to share this initial data with the community immediately. We also note that in addition to the 10x Genomics Visium platform, which lacks cellular and sub-cellular resolution, many new and exciting spatial platforms are entering the market, which may be able to address questions in very small regions such as the LC at higher spatial resolution.

b. Unfortunately, spatial transcriptomics itself is plagued by sampling variability to a point where the RNAscope analyses the authors performed prove more powerful in addressing direct questions about gene expression patterns. Given that the authors compare their results to published datasets from rodent studies, it is surprising that a direct comparison of genes identified with spatial transcriptomics vs snRNAseq is lacking (unless this reviewer missed this comparison). Supplementary Figure 17 seems to be a first step in that direction, but this is not a gene-by-gene comparison of which analysis identifies which LC-enriched genes. Such an analysis should not compare numbers of enriched genes using artificial cutoffs for significance/fold-change, but rather use correlations to get a feeling for which genes appear to be enriched in the LC using both methods. This would result in one list of genes that can serve as a reference point for future work.

We agree this is a good suggestion, and will add additional computational analyses to address this point in a revised version of the manuscript.

c. Maybe the spatial transcriptomics could be useful to look at the peri-LC region, which has generated some excitement in rodent work recently, but remains largely unexplored in humans.

We agree this is an excellent suggestion -- assessing cross-species comparisons related to convergence, especially, of GABAergic cell populations in the human LC is of high interest. We note that these types of extensions are exactly the reason why we have provided the publicly accessible web app (R/Shiny app, which includes the ability to annotate regions). We hope that others will use these apps for specialized topics they are interested in. As discussed above, we note that our initial dissections precluded the ability to keep track of the exact orientation of our tissue sections on the Visium arrays with respect to their location within the brainstem, so definitive localization of this region across subjects is difficult in our current study. However, it is possible, for example, to investigate whether there is a putative peri-LC region that is densely GABAergic that is homologous with the GABAergic peri-LC region in rodents. We also raise attention to a recent preprint by Luskin and Li et al. (2022), who apply snRNA-seq and spatially-resolved transcriptomics to molecularly define both LC and peri-LC cell types in mice -- in a revised version of our manuscript, we will extend our computational analyses of inhibitory neuronal subtypes in our data (Supplementary Figures 13, 16) to directly compare with those identified in this study in more detail. As noted above, we we have now developed a number of specialized technical and logistical strategies for keeping track of orientation of sections from the tissue block onto a single spatial array, and we feel that combined with optimized dissection strategies for this region and the guide of RNAscope for GABAergic markers on serial sections, that annotating the peri-LC region on spatial arrays in future studies will be possible.

3) The comparison of snRNAseq data to published literature is laudable. Although the authors mention considerable methodological differences between the chosen rodent work and their own analyses, this needs to be further explained. The mouse dataset uses TRAPseq, which looks at translating mRNAs associated with ribosomes, very different from the nuclear RNA pool analyzed in the current work. The rat dataset used single-cell LC laser microdissection followed by microarray analyses, leading to major technical differences in terms of tissue processing and downstream analyses. The authors mention and reference a recent 10x mouse LC dataset (Luskin et al, 2022), however they only pick some neuropeptides from this study for their analysis of interneuron subtypes (Figure S13). Although this is a very interesting part of the manuscript, a more in-depth analysis of these two datasets would be very useful. It would likely allow for a better comparison between mouse and human, given that the technical approach is more similar (albeit without FACS), and Luskin et al have indicated that they are willing to share their data.

As noted above, we plan to extend our comparisons with the dataset from Luskin and Li et al. (2022) in a revised version of the manuscript, which will provide a more in-depth cross-species comparison. In addition, we also note that there are some additional recent studies using TRAPseq of LC-NE neurons in a functional context, i.e. treatment vs. control experiments or in model systems (e.g. Iannitelli et al. 2023), which provide new opportunities for understanding disease context using in-depth cross-species comparisons. By providing our dataset and reproducible code, we will enable others to adapt and extend these types of comparisons (i.e. TRAPseq of LC-NE neurons or LC snRNA-seq following functional manipulations or in the context of disease or behavioral models) in the future.

4) Statements in the manuscript about the unexpected identification of a 5-HT (serotonin) cell-cluster seem somewhat contradictory. Figure S14 suggests that 5-HT markers are expressed in the LC-regions just as much as anywhere else, but the RNAscope image in Figure S15 suggests spatial separation between these two populations. And Figure S17 again suggests almost perfect overlap between the LC and 5HT clusters. Maybe I misunderstood, in which case the authors should better clarify/explain these results.

In our view, the most likely scenario is that the 5-HT neurons come from contamination from the dorsal raphe nucleus based on spatial separation from the RNAscope images, which we agree are more definitive. As mentioned above, since we do not have definitive documentation for the tissue sections in terms of orientation, it is difficult to say with clarity that the regions are the dorsal raphe and which sub-portion of the dorsal raphe they are. This initial study has now allowed us to optimize and improve our dissection strategy and approaches for retaining documentation of the orientation of the tissue sections from their intact position within the brainstem as they move from cryosection to placement on the array, which will enable us to better annotate regions with definitive anatomical information with respect to the rostral/caudal and dorsal/ventral axes in future experiments. Given that there are reports in the rodent that 5-HT markers have been identified in LC-NE neurons (Iijima 1993; Iijima 1989), and taking into account the technical limitations in our study, we felt that it was premature to definitively conclude in the manuscript that we were sure these signals arose from the dorsal raphe. We will update this language in a revised version of the manuscript to ensure that these limitations are clear (referring to Supplementary Figures S14-15, S17).

Reviewer #2 (Public Review):

The data generated for this paper provides an important resource for the neuroscience community. The locus coeruleus (LC) is the known seed of noradrenergic cells in the brain. Due to its location and size, it remains scarcely profiled in humans. Despite the physically minute structure containing these cells, its impact is wide-reaching due to the known neuromodulatory function of norepinephrine (NE) in processes like attention and mood. As such, profiling NE cells has important implications for most neurological and neuropsychiatric disorders. This paper generates transcriptomic profiles that are not only cell-specific but which also maintain their spatial context, providing the field with a map for the cells within the region.

Strengths:

Using spatial transcriptomics in a morphologically distinct region is a very attractive way to generate a map. Overlaying macroscopic information, i.e. a region with greater pigmentation, with its corresponding molecular profile in an unbiased manner is an extremely powerful way to understand the specific cellular and molecular composition of that brain structure.

The technologies were used with an astute awareness of their limitations, as such, multiple technologies were leveraged to paint a more complete and resolved picture of the cellular composition of the region. For example, the lack of resolution in the spatial transcriptomic platform was compensated by complementary snRNA-seq and single molecule FISH.

This work has been made publicly available and accessible through a user-friendly application such that any interested researcher can investigate the level of expression of their gene of interest within this region.

Two important implications from this work are 1) the potential that the gene regulatory profiles of these cells are only partially conserved across species, humans, and rodents, and 2) that there may be other neuromodulatory cell types within the region that were otherwise not previously localized to the LC

Weaknesses:

Given that the markers used to identify cells are not as specific as they need to be to definitively qualify the desired cell type, the results may be over-interpreted. Specifically, TH is the primary marker used to qualify cells as noradrenergic, however, TH catalyzes the synthesis of L-DOPA, a precursor to dopamine, which in turn is a precursor for epinephrine and norepinephrine suggesting some of the cells in the region may be dopaminergic and not NE cells. Indeed, there are publications to support the presence of dopaminergic cells in the LC (see Kempadoo et al. 2016, Takeuchi et al., 2016, Devoto et al. 2005). This discrepancy is further highlighted by the apparent lack of overlap per given Visium spots with TH, SCL6A2, or DBH. While the single-nucleus FISH confirms that some of the cells in the region are noradrenergic, others very possibly represent a different catecholamine. As such it is suggested that the nomenclature for the cells be reconsidered.

We appreciate the reviewer’s comment, and are aware of the reports suggesting the potential presence of dopaminergic cells in the LC. We initially had the same thought as the reviewer when we observed Visium spots in the spatial data with lack of overlap between TH, SLC6A2, and DBH as well as single nuclei in the snRNA-seq data with lack of overlap between TH, SLC6A2, and DBH. This surprising result was exactly why we performed the smFISH/RNAscope experiment with these three marker genes. Given known issues with read depth and coverage in the 10x Genomics assays, we wanted to better understand if this was a technical limitation in the sequencing coverage, or rather a true biological finding. The RNAscope data showed very clearly that nearly every cell body we looked at had co-localization of these three marker genes. We included an image from a single capture array of one tissue section in Supplementary Figure 11, but could, in a revised version of the manuscript, provide additional examples to illustrate how conclusive the images were by visualization. As such, we were quite convinced that the lack of overlap on Visium spots and in single nuclei in the snRNA-seq data was more likely related to technical issues with sequencing coverage, rather than a biological finding. We also note that we checked for the presence of the dopamine transporter, SLC6A3, and as can be appreciated in the iSEE web app for the snRNA-seq data or the R/Shiny web app for the Visium data, there is virtually no expression of SLC6A3 in the dataset, which in our view provides additional evidence against the possibility that there are substantial quantities of dopaminergic cells in this human LC dataset. We will include supplementary plots showing the lack of SLC6A3 expression in a revised version of the manuscript.

The authors are unable to successfully implement unsupervised clustering with the spatial data, this greatly reduces the impact of the spatial technology as it implies that the transcriptomic data generated in the study did not have enough resolution to identify individual cell types.

The reviewer is correct -- this is a fundamental limitation of the 10x Genomics Visium platform, i.e. the spatial resolution captures multiple cells per spot (e.g. around 1-10 cells per spot in human brain tissue). We note that new spatial platforms now provide cellular resolution (e.g. Vizgen MERSCOPE, 10x Genomics Xenium, 10x Genomics Visium HD), which will help address this in future work. However, many of these cellular-resolution in situ sequencing platforms have the limitation that they do not quantify genome-wide expression, and instead require users to select a priori gene panels to investigate. This is a problem if no genome-wide reference datasets are available. Hence, despite the limited spatial resolution of the Visium platform, this dataset is useful precisely for helping investigators choose gene panels for higher-resolution platforms or higher-order smFISH multiplexing.

We also applied spatial clustering (using BayesSpace; Zhao et al. 2021) to attempt to segment the LC regions within the Visium samples in a data-driven manner as an alternative to the manual annotations, which was unsuccessful (and hence we relied on the manually annotated regions for downstream analyses) (Supplementary Figure S5). However, this is a different application of unsupervised clustering, which is separate from the task of identifying cell types.

The sample contribution to the results is highly unbalanced, which consequently, may result in ungeneralizable findings in terms of regional cellular composition, limiting the usefulness of the publicly available data.

We acknowledge the limitations of the work due to the small/unbalanced sample sizes. As mentioned above for Reviewer 1, this was an initial study in this region -- results of which will inform our (and hopefully others’) experimental design and approach to molecular profiling in this difficult to access brain region. Overall, this study was executed with finite tissue and financial resources and was intended to uncover limitations and help develop best practices and design workflows for future studies with larger numbers of donors and samples. Given the limited data availability for this brain region, we wanted to make this dataset available for the research community immediately. In addition, we note that making this genome-wide dataset available will help inform targeted gene panel design for higher-resolution platforms (e.g. 10x Genomics Xenium).

This study aimed to deeply profile the LC in humans and provide a resource to the community. The combination of data types (snRNA-seq, SRT, smFISH) does in fact represent this resource for the community. However, due to the limitations, of which, some were described in the manuscript, we should be cautious in the use of the data for secondary analysis. For example, some of the cellular annotations may lack precision, the cellular composition also may not reflect the general population, and the presence of unexpected cell types may represent the accidental inclusion of adjacent regions, in this case, serotonergic cells from the Raphe nucleus.

We agree, and have attempted to explain these limitations in the manuscript. We will clarify the language regarding the interpretation of the annotated cell populations and unexpected cell types, and the limited sample sizes, in a revised version of the manuscript.

Nonetheless having a well-developed app to query and visualize these data will be an enormous asset to the community especially given the lack of information regarding the region in general.

Reviewer #3 (Public Review):

[…] This study has many strengths. It is the first reported comprehensive map of the human LC transcriptome, and uses two independent but complementary approaches (spatial transcriptomics and snRNA-seq). Some of the key findings confirmed what has been described in the rodent LC, as well as some intriguing potential genes and modules identified that may be unique to humans and have the potential to explain LC-related disease states. The main limitations of the study were acknowledged by the authors and include the spatial resolution probably not being at the single cell level and the relatively small number of samples (and questionable quality) for the snRNA-seq data. Overall, the strengths greatly outweigh the limitations. This dataset will be a valuable resource for the neuroscience community, both in terms of methodology development and results that will no doubt enable important comparisons and follow-up studies.

Major comments:

Overall, the discovery of some cells in the LC region that express serotonergic markers is intriguing. However, no evidence is presented that these neurons actually produce 5-HT.

The reviewer is correct that we did not provide any additional evidence to show that these neurons actually produce 5-HT. As noted above in the response to Reviewer 1, in our view, the most likely explanation is that these neurons are from dorsal raphe contamination on the tissue section. However, due to technical and logistical limitations in this study, we could not definitively say this because we did not clearly track the orientation of the tissue sections, and we did not have remaining tissue sections from all donor tissue blocks to repeat RNAscope experiments. For some of the donors, where we had remaining tissue sections to go back to repeat RNAscope experiments after completion of the snRNA-seq and Visium assays, we could see clear separation of the LC region / LC-NE neuron core from where putative 5-HT neurons were located (Supplementary Figure 15). However, we did not have sufficient tissue resources to map this definitively in all donors, and the orientation and anatomy of each tissue block were not fully annotated.

Due to the lack of clarity, and the fact that there have been reports that LC-NE neurons express serotonergic markers (Iijima 1993; Iijima 1989), we felt that it was premature to definitively declare that these putative 5-HT neurons that we identified were definitively from the raphe. We will clarify the language around this discrepancy in a revised version of the manuscript to ensure that these limitations are clearly described.

Concerning the snRNA-seq experiments, it is unclear why only 3 of the 5 donors were used, particularly given the low number of LC-NE nuclear transcriptomes obtained, why those 3 were chosen, and how many 100 um sections were used from each donor. It is also unclear if the 295 nuclei obtained truly representative of the LC population or whether they are just the most "resilient" LC nuclei that survive the process.

As discussed above for Reviewer 1, the reason we included only 3 of the 5 donors for the snRNA-seq assays was due to the tissue availability on the tissue blocks. We will clarify the language in a revised version of the manuscript to make this limitation more clear. We will also include additional details in the Methods section on the number of 100 μm sections used for each donor (which varied between 10-15, approximating 60-80 mg of tissue).

The LC displays rostral/caudal and dorsal/ventral differences, including where they project, which functions they regulate, and which parts are vulnerable in neurodegenerative disease (e.g. Loughlin et al., Neuroscience 18:291-306, 1986; Dahl et al., Nat Hum Behav 3:1203-14, 2019; Beardmore et al., J Alzheimer's Dis 83:5-22, 2021; Gilvesy et al., Acta Neuropathol 144:651-76, 2022; Madelung et al., Mov Disord 37:479-89, 2022). It was not clear which part(s) of the LC was captured for the SRT and snRNAseq experiments.

As discussed above for Reviewer 1, a limitation of this study was that we did not record the orientation of the anatomy of the tissue sections, precluding our ability to annotate the tissue sections with the rostral/caudal and dorsal/ventral axis labels. We agree with the reviewer that additional spatial studies, in future work, could offer needed and important information about expression profiles across the spatial axes (rostral/caudal, ventral/dorsal) of the LC. Our study provides us with insight about optimizing the dissections for spatial assays, as well as bringing to light a number of technical and logistical issues that we had not initially foreseen. For example, during the course of this study and parallel, ongoing work in other, small, challenging regions, we have now developed a number of specialized technical and logistical strategies for keeping track of orientation and mounting serial sections from the same tissue block onto a single spatial array, which is extremely technically challenging. We are now well-prepared for addressing these issues in future studies with larger numbers of donors and samples in order to make these types of insights.

The authors mention that in other human SRT studies, there are typically between 1-10 cells per expression spot. I imagine that this depends heavily on the part of the brain being studied and neuronal density, but it was unclear how many LC cells were contained in each expression spot.

The reviewer is correct that we did not include this information in the manuscript. We attempted to apply a computational method to count nuclei contained in each gene expression spot based on analyzing the histological H&E images (VistoSeg; Tippani et al. 2022), which we have developed and previously applied in data from the dorsolateral prefrontal cortex (DLPFC) (Maynard and Collado-Torres et al. 2021). Based on the segmentation using this workflow we observe that the counts in this region are similar to what we observed in the DLPFC, i.e., typically between 1-10 LC cells per expression spot, with approximately 1-2 LC-NE neurons (which are characterized by their large size) per expression spot. However, these analyses had several technical issues related to the images themselves, the relatively large size and pigmentation of LC-NE neurons, and parameter settings that had been optimized for different brain regions. We are currently optimizing this analysis workflow for these images to provide more accurate estimates of cell counts per spot to give readers additional context on the number of nuclei per spot in the annotated LC regions and outside the LC regions in a revised version of the manuscript.

Regarding comparison of human LC-associated genes with rat or mouse LC-associated genes (Fig. 2D-F), the authors speculate that the modest degree of overlap may be due to species differences between rodents and human and/or methodological differences (SRT vs microarray vs TRAP). Was there greater overlap between mouse and rat than between mouse/rat and human? If so, that is evidence for the former. If not, that is evidence for the latter. Also would be useful for more in-depth comparison with snRNA-seq data from mouse LC: https://www.biorxiv.org/content/10.1101/2022.06.30.498327v1.

We will investigate this question and discuss this in updated results in a revised version of the manuscript.

The finding of ACHE expression in LC neurons is intriguing, especially in light of work from Susan Greenfield suggesting that ACHE has functions independent of ACH metabolism that contributes to cellular vulnerability in neurodegenerative disease.

We thank the reviewer for pointing this out. We were very surprised too by the observed expression of SLC5A7 and ACHE in the LC regions (Visium data) and within the LC-NE neuron cluster (snRNA-seq data), coupled with absence of other typical cholinergic marker genes (e.g. CHAT, SLC18A3), and we do not have a compelling explanation or theory for this. Hence, the work of Susan Greenfield and colleagues suggesting non-cholinergic actions of ACHE, particularly in other catecholaminergic neurons (e.g. dopaminergic neurons in the substantia nigra) is very interesting. We will include references to this work and how it could inform interpretation of this expression in a revised version of the manuscript (Greenfield 1991; Halliday and Greenfield 2012).

High mitochondrial reads from snRNA-seq can indicate lower quality. It was not clear why, given the mitochondrial read count, the authors are confident in the snRNA-seq data from presumptive LC-NE neurons.

We will include additional analyses to further investigate and/or confirm this finding (e.g. comparing sum of UMI counts / number of detected genes and mitochondrial percentage per nucleus for this population to confirm data quality) in additional supplementary figures in a revised version of the manuscript.

References

• Greenfield (1991), A noncholinergic action of acetylcholinesterase (AChE) in the brain: from neuronal secretion to the generation of movement, Cellular and Molecular Neurobiology, 11, 1, 55-77.

• Halliday and Greenfield (2012), From protein to peptides: a spectrum of non-hydrolytic functions of acetylcholinesterase, Protein & Peptide Letters, 19, 2, 165-172.

• Iannitelli et al. (2023), The neurotoxin DSP-4 dysregulates the locus coeruleus-norepinephrine system and recapitulates molecular and behavioral aspects of prodromal neurodegenerative disease, eNeuro, 10, 1, ENEURO.0483-22.2022.

• Iijima K. (1989), An immunocytochemical study on the GABA-ergic and serotonin-ergic neurons in rat locus ceruleus with special reference to possible existence of the masked indoleamine cells. Acta Histochema, 87, 1, 43-57.

• Iijima K. (1993), Chemocytoarchitecture of the rat locus ceruleus, Histology and Histopathology, 8, 3, 581-591.

• Luskin A.T., Li L. et al. (2022), A diverse network of pericoerulear neurons control arousal states, bioRxiv (preprint).

• Maynard and Collado-Torres et al. (2021), Transcriptome-scale spatial gene expression in the human dorsolateral prefrontal cortex, Nature Neuroscience, 24, 425-436.

• Tippani et al. (2022), VistoSeg: processing utilities for high-resolution Visium/Visium-IF images for spatial transcriptomics data, bioRxiv (preprint).

• Tran M.N., Maynard K.R. et al. (2021), Single-nucleus transcriptome analysis reveals cell-type-specific molecular signatures across reward circuitry in the human brain, Neuron, 109, 3088-3103.

• Zhao E. et al. (2021), Spatial transcriptomics at subspot resolution with BayesSpace, Nature Biotechnology, 39, 1375-1384.

Author response:

The following is the authors’ response to the current reviews.

Reviewer #2:

Line 295 – was the time post-infection, which varies considerably between groups and across samples, taken into consideration when comparison of response was between ChatCre mice (4-9 weeks post-infection) and WT mice (four to five weeks post-infection)?

Thank you for your comment. We did not originally assess the effects of time post-injection on DREADD response. Generally, AAV transgene expression has been demonstrated to be long-term and stable in the CNS of mice.[1] However, there is some variation in the reporting time of peak transgene expression[2], and this may potentially impact our results.

In investigating this issue further, we discovered an error in our reporting as we did have n = 1 wild-type mouse that underwent EMG recordings 62 days (~9 weeks) post-AAV injection. This has been corrected in the manuscript (lines 87-88).

Addressing this question is challenging due to the uneven distribution of time points within the 4–9-week windows for each group. Essentially, there were two groups per cohort, one studied at 4-5 weeks and one at 8-9 weeks. More specifically:

- Wild-type cohort: n = 10 animals were studied 28–33 days post-injection, and n = 1 at 62 days.

- ChAT-Cre cohort: n = 4 animals were studied 28–30 days post-injection, and n = 5 at 56–59 days.

We performed Pearson correlation analyses between time post-injection and diaphragm EMG response to DREADD activation (peak amplitude and area under the curve, AUC) for both cohorts (Author response image 1):

- ChAT-Cre: No significant correlations were found (peak amplitude: r² = -0.117, r = -0.1492, p = 0.702, Figure 1a-b; AUC:r² = -0.0883, r = 0.2184, p = 0.572, Figure 1c-d).

- Wild type: Initial analysis of all data showed significant correlations (peak amplitude:r² = 0.362, r = 0.6523, p = 0.0296, Figure 1a; AUC: r² = 0.347, r = 0.6424, p = 0.033, Figure 1c), suggesting a moderate positive correlation between time post-injection and EMG response. However, when the single 8–9-week wild-type mouse was excluded, these correlations were no longer significant (peak amplitude: r² = 0.172, r = 0.5142, p = 0.128, Figure 1b; AUC: r² = 0.23, r = 0.5614, p = 0.0913, Figure1d).

Comparing wild-type and ChAT-Cre groups directly was unreliable due to the single wild-type mouse studied at the later time point. We attempted to model time post-injection as a continuous variable (i.e., exact days post-injection) using a restricted maximum likelihood mixed linear model in JMP; however, the analysis could not be performed because there were not sufficient overlapping time points between the two cohorts (i.e., not all days post-injection were represented in both groups). To mitigate this, we binned animals into two groups: 4–5 weeks and 8–9 weeks post-injection. This analysis returned a significant interaction between cohort and time post-injection (p = 0.0391), however there were no significant multiple comparisons upon Tukey post hoc test (i.e., p > 0.05).

Based on these findings, we feel confident that time post-injection is unlikely to have a significant impact on diaphragm EMG response to DREADD activation in the ChAT-Cre cohort. However, in the wild-type cohort, it is difficult to draw definitive conclusions, as only one animal was studied at the 8–9-week time point. For similar reasons, it remains unclear whether the relationship between time post-AAV transduction and DREADD response differs between cohorts. Given the inconclusive nature of these results, we have elected not to include this analysis in the manuscript. Nevertheless, to ensure transparency, we have provided Author response image 1 below of peak amplitude and AUC plotted against time, allowing readers to evaluate the data independently.

Author response image 1.

Plots of diaphragm EMG peak amplitude (a-b) and area under the curve (c-d) vs. days post-AAV injection for wild-type (blue) and ChAT-Cre (orange) mice. Pearson correlation analyses were performed to assess the relationship between time post-AAV injection and diaphragm EMG DREADD response in wild-type and ChAT-Cre mouse cohorts. r², r, and p-values are shown in each panel for both cohorts. Panels a and c display peak amplitude and AUC, respectively, including all animals. Panels b and d present the same variables with the n = 1 wild-type mouse at the 9-week time point excluded; ChAT-Cre data is unchanged between corresponding panels. Scatter points represent data from individual animals. Polynomial trendlines are displayed for each cohort with wild-type in blue and ChAT-Cre in orange.

REFERENCES

(1) Kim, J. Y., Grunke, S. D., Levites, Y., Golde, T. E. & Jankowsky, J. L. Intracerebroventricular viral injection of the neonatal mouse brain for persistent and widespread neuronal transduction. J Vis Exp, 51863 (2014). https://doi.org/10.3791/51863

(2) Hollidge, B. S. et al. Kinetics and durability of transgene expression after intrastriatal injection of AAV9 vectors. Front Neurol 13, 1051559 (2022). https://doi.org/10.3389/fneur.2022.1051559

The following is the authors’ response to the original reviews.

Response to reviewer’s public reviews:

We chose the dose of J60 based on a prior publication that established that off-target effects were possible at relatively high doses[1]. The dose that we used (0.1 mg/kg) was 30-fold less than the dose that was reported in that paper to potentially have off-target responses (3 mg/kg). Further, Author response image 1 shows the results of experiments in which J60 was given to animals that did not have the excitatory DREADD expressed in the spinal cord. This includes a sample of mice (n = 2) and rats (n = 3), recorded from using the same diaphragm EMG procedure described in the manuscript. The figure shows that there was no consistent response to the J60 at 0.1 mg/kg in the “control experiment” in which the DREADD was not expressed in the spinal cord.

Author response image 1.

Diaphragm EMG response to J60 administrated to naïve rats and mice. Panel a-b show raw EMG values at baseline, following vehicle (saline) and J60 administration for the left and right hemidiaphragm. Panel c-d shows EMG values normalized to baseline. Neither One-way RM ANOVA (panel a-b) nor paired t-test (panel c-d) returned significant p values (p < 0.05).

Response to specific reviewer comments:

Reviewer #1:

How old were the animals at the time of AAV injection, and in subsequent experiments?

The wildtype cohort of mice were 7-9 weeks old at time of AAV injection and DREADD experiments took place 4-5 weeks after AAV injection. ChAT-Cre mice were 6-10 weeks old at time of AAV injection and DREADD experiments took place 4-9 weeks after AAV injection. ChAT-Cre rats were 2-5 months old at time of AAV spinal injection. These animals underwent plethysmography recordings 3-4 months post-AAV injection and subsequently phrenic nerve recording 3-8 weeks later. These details have been added to the Method section.

How many mice were excluded from electrophysiology experiments due to deteriorating electrode contact?

No mice were excluded from electrophysiology experiments due to deteriorating electrode contact. If you are referring to the n = 1 excluded ChAT-Cre mouse (line 368) this animal was excluded because it showed no histological evidence of DREADD expression (lines 200-206).

What was the urethane dose?

The urethane dose for phrenic nerve recordings was 2.1 g/kg. See methods section line 395.

A graphical timeline of the experimental progression for plethysmography and electrophysiology studies would enhance clarity.

A graphical timeline has been added. See Figure S6.

Significance indicators in the figures would greatly enhance clarity. It is a little awkward to have to refer to supplemental tables to figure out statistical differences.

Significance indicators have been added. See Figures 1, 2, 4, and 5

In Figures 1, 2, and 5, individual data points should be shown, as in Fig 4.

Thank you for this suggestion. We agree that, in general, it is best practice to scatter individual data points. However, when we drafted the new figures, it was apparent that including individual scatter points, in this case, created very “cluttered” figures that were very difficult to interpret.

More detail regarding the plethysmography studies is needed. Was saline/J60 infused via a tail vein catheter? Were animals handled during the infusion? How long is the "IV" period? What volume of fluid was delivered?

All IV infusions were delivered via a tail vein catheter. Animals were not handled during infusion nor at any point during the recording. An IV catheter was externalized via a port in the plethysmograph allowing for IV infusion without handling of the animal or opening the plethysmograph. The infusion period for both saline and J60 was standardized to 2 minutes. The volume of fluid of both saline and J60 was standardized to 0.6 mL. This information has been added to the methods section (lines 408-410, 415-16, 419-420).

Reviewer #2:

The abstract could be improved by briefly highlighting the rationale, scope, and novelty of the study - the intro does a great job of highlighting the scope of the study and the research questions.

A brief explanation of the rationale, scope, and novelty of the study has been added to the abstract. See lines 2-8.

Line 18, specifies that this was done under urethane anesthesia.

This detail has been added to the abstract (line 20).

The methods section should be moved to the end of the manuscript according to Journal policy.

The methods section has been moved to the end of the manuscript.

The authors mention the use of both female and male rats but it is not indicated if they tested for and observed any differences between sexes across experiments.

We included the use of both male and female animals in this study to improve the generalizability of the results. However, we were not adequately powered for sex comparisons and therefore did not perform any statistical analysis to assess differences between sexes across experiments. Text has been added to the methods section (lines 534-537) to clarify.

Line 40, since delivery of J60 was performed in both IV and IP, this general statement should be updated.

This detail has been revised to include both IV and IP. See line 43.

Line 42. "First, we determined if effective diaphragm activation requires focal DREADD expression targeting phrenic motor neurons, or if non-specific expression in the immediate vicinity of the phrenic motor nucleus would be sufficient...." I don't think that in the experiments with wild-type mice the authors can claim that they selectively targeted the cervical propriospinal network (in isolation from the motoneurons). Given the fact that the histological analysis did not quantify interneurons or motoneurons in the spinal cord, authors should be cautious in proposing which neuronal population is activated in the non-specific approach.

We agree, and this was a poorly worded statement in our original text. We agree that wild-type DREADD expression was not limited to the cervical propriospinal networks but likely a mix of interneurons and motoneurons. The text has been edited to reflect that (see lines 56-60).

AAV virus source is not described.

All AAVs were obtained from the UF Powell Gene Therapy Center. Details of virus source and production have been added to the methods section. See lines 336-347.

Line 108-125. Because the diaphragm EMG recordings are only described for mice here, I would suggest editing this methods section to clearly state mice instead of vaguely describing "animals" in the procedure.

“Animals” has been changed to “mice” to avoid ambiguity.

Line 120, add parenthesis.

Parenthesis has been added.

Line 126. Whole body plethysmography protocol. Three hypercapnic hypoxic challenges are a lot for a rat within a 3-hour recording session in freely behaving rats. Did the authors verify with control/ vehicle experiments that repeated challenges in the absence of J60 do not cause potentiation of the response? I understand that it is not possible to invert the order of the injections (due to likely long-term effects of J60) or it is too late to perform vehicle and J60 injections on different days, but controls for repeated challenges should be performed in this type of experiment, especially considering the great variability in the response observed in Figure 4 (in normoxic conditions).

We did not conduct control experiments to assess the impact of repeated hypercapnic hypoxic challenges on the naïve response (i.e., in the absence of J60). However, our experimental protocol was designed such that each experimental period (i.e., post-vehicle or post-J60 infusion) was normalized to baseline recordings taken immediately prior to the vehicle or J60 infusion. While repeated exposure to hypercapnic hypoxic challenges may have altered respiratory output, we are confident that normalizing each experimental period to its respective baseline effectively captures the impact of DREADD activation on ventilation, independent of any potential potentiation that may have occurred due to gas challenge exposure. We have included raw values for all plethysmography outcomes (see Figure 4, panels a-c) to ensure full data transparency. Still, we believe that the baseline-normalized values more accurately reflect the impact of DREADD activation on the components of ventilation.

Furthermore, why the response to the hypercapnic hypoxic challenges are not reported? These could be very interesting to determine the effects of DREADD stimulation on chemosensory responses and enhance the significance of the study.

Response to the hypercapnic hypoxic challenges has been added to the manuscript. See Figure S3 and results section lines 162-167. Briefly, there were no statistically significant (p < 0.05) differences in tidal volume, respiratory rate, or minute ventilation between J60 vs sham condition during hypercapnic-hypoxic ventilatory challenges.

Line 200 - what is the reason behind performing a qualitative analysis of mCherry in various quadrants? This limits the interpretation of the results. If the authors used Chat-cre rats, the virus should only be in Chat+ MN. Knowing how selective the virus is, and whether its expression was selective for Phrenic MN versus other MN pools, could address several technical questions.

We agree that detailed quantification of expression by motoneuron pool would be of value in future work.  However, for these initial proof-of-concept experiments, we performed the quadrant-based qualitative analysis of mCherry expression to provide a simple comparison of mCherry expression between groups (i.e., ChAT-Cre vs. wildtype mice). This analysis allowed us to: 1) show the reader that each animal included in the study showed evidence of mCherry expression and 2) give the reader an idea of patterns of mCherry expression throughout the mid-cervical spinal cord. Additionally, it is important to note that while ChAT is a marker of motoneurons some populations of interneurons also express ChAT(2-4).

Given the increased values of Dia EMG AUC and no changes in respiratory rate, did the authors determine if there was a change in the inspiratory time with J60 administration?

We did not assess inspiratory time.

High death rate in DREADD WT mice - was histological analysis performed on these mice? Could it be due to the large volume injected into the spinal cord that affects not only descending pathways but also ascending ones? Or caused by neuronal death due to the large volume of viral solution in injected in mice.

Histological analysis was performed on these animals to assess mCherry expression only (i.e., no staining for NeuN or other markers was performed). While the reviewer's speculations are reasonable, we feel these reasons are unlikely to explain the death rate in DREADD WT mice as ChAT-Cre mice received the same volume injected into their spine and lived up until and during diaphragm EMG recordings. Additionally, WT mice lived for 4-5 weeks post-injection which would be past the acute phase that a large immune response to the viral dose would have occurred.

Line 299-304. Can you please clarify whether these rats were tested under anesthesia?

These rats were assessed under anesthesia. This detail has been added (line 146).

Given some of the unexpected results on cardiovascular parameters in urethane anesthetized rats, did the authors test the effects of J60 in the absence of AAV construct infection?

A small cohort (n = 2) of urethane anesthetized naïve wildtype rats were given the J60 ligand (IV, 0.1 mg/kg dose). We did observe a sudden drop in blood pressure after J60 administration that was sustained for the duration of the recording. One animal showed a 12% decrease in mean arterial blood pressure following J60 administration while the other showed a 35% decrease. Thus, it does appear that in this preparation the J60 ligand is producing a drop in arterial blood pressure.

Line 393. I believe this comment is referred to the intrapleural and diaphragmatic injection. Maybe this should clarified in the sentence.

This sentence has been revised for clarity (see lines 248-250).

Figures 1 and 2. It would be informative to show raw traces of the Diaphragm EMG to demonstrate the increase in tonic EMG. It is not possible to determine that from the integrated traces in Figures 1A and B.

Thank you for bringing up this concern. While the mean data in Figures 1F and 2F do indicate that, on average, animals had tonic diaphragm EMG responses to DREADD activation, the examples given in Figures 1A and 2A show minimal responses. This makes it difficult to fully appreciate the tonic response from those particular traces. However, clear tonic activity can be appreciated from Figures 5A and S2. In these figures, tonic activity is evident from the integrated EMG signals, presenting as a sustained increase in baseline activity between bursts—essentially an upward shift from the zero point.

References

(1) Van Savage, J. & Avegno, E. M. High dose administration of DREADD agonist JHU37160 produces increases in anxiety-like behavior in male rats. Behav Brain Res 452, 114553 (2023). https://doi.org/10.1016/j.bbr.2023.114553

(2) Mesnage, B. et al. Morphological and functional characterization of cholinergic interneurons in the dorsal horn of the mouse spinal cord. J Comp Neurol 519, 3139-3158 (2011). https://doi.org/10.1002/cne.22668

(3) Gotts, J., Atkinson, L., Yanagawa, Y., Deuchars, J. & Deuchars, S. A. Co-expression of GAD67 and choline acetyltransferase in neurons in the mouse spinal cord: A focus on lamina X. Brain Res 1646, 570-579 (2016). https://doi.org/10.1016/j.brainres.2016.07.001

(4) Alkaslasi, M. R. et al. Single nucleus RNA-sequencing defines unexpected diversity of cholinergic neuron types in the adult mouse spinal cord. Nat Commun 12, 2471 (2021). https://doi.org/10.1038/s41467-021-22691-2

Author response:

The following is the authors’ response to the original reviews

We thank the public reviewers and editors for their insightful comments on the manuscript. We have made the following changes to address their concerns and think the resulting manuscript is stronger as a result. Specifically, we have 1) added RNA FISH data of specific STB-2 and STB-3 RNA markers to confirm their distribution changes between STBⁱⁿ and STB^out TOs, 2) removed language throughout the text that refer to STB-3 as a terminally differentiated nuclear subtype, and 3) generated CRISPR-mediated knock-outs of two genes identified by network analysis and validated their rolse in mediating STB nuclear subtype gene expression.

Reviewer #1 (Public review): 

Strengths: 

The study offers a comprehensive SC- and SN-based characterization of trophoblast organoid models, providing a thorough validation of these models against human placental tissues. By comparing the older STBⁱⁿ and newer STB^out models, the authors effectively demonstrate the improvements in the latter, particularly in the differentiation and gene expression profiles of STBs. This work serves as a critical resource for researchers, offering a clear delineation of the similarities and differences between TO-derived and primary STBs. The use of multiple advanced techniques, such as high-resolution sequencing and trajectory analysis, further enhances the study's contribution to the field. 

Thank you for your thoughtful review—we appreciate your recognition of our efforts to comprehensively validate trophoblast organoid models and highlight key advancements in STB differentiation and gene expression.

Weaknesses: 

While the study is robust, some areas could benefit from further clarification. 

(1) The importance of the TO model's orientation and its impact on outcomes could be emphasized more in the introduction. 

We agree that TO orientation may significantly influence STB nuclear subtype differentiation. As the STB is critical for both barrier formation and molecular transport in vivo, lack of exposure to the surrounding media in STBⁱⁿ TOs in vitro could compromise these functions and the associated environmental cues that influence STB nuclear differentiation. We have added text to the introduction to highlight this point (lines 117-120).

(2) The differences in cluster numbers/names between primary tissue and TO data need a clearer explanation, and consistent annotation could aid in comparison. 

Thank you for highlighting that the comparisions and cluster annotations need clarification. In Figure 1, we did not aim to directly compare CTB and STB nuclear subtypes between TOs and tissue. Each dataset was analyzed independently, with clusters determined separately and with different resolutions decided via a clustering algorithm (Zappia and Oshlack, 2018). For example, for the STB, this approach identified seven subtypes in tissue but only two in TOs, making direct comparison challenging. To address this challenge, we integrated the SN datasets from TOs and tissue in Figure 6. This integration allowed us to directly compare gene expression between the sample types and examine the proportions within each STB subtype. Similarly, in Figure 2, direct comparison of individual CTB or STB clusters across the separate datasets is challenging (Figures 2A-C) due to differences in clustering. To overcome this, we integrated the datasets to compare cluster gene expression and relative proportions (Figures 2D-E). Nonetheless, to address the reviewers concern we have added text to the results section to clarify that subclusters of CTB and STB between datasets should not be directly compared until the datasets are integrated in Figure 2D-E and Figure 6 (lines 166-167).

(3) The rationale for using SN sequencing over SC sequencing for TO evaluations should be clarified, especially regarding the potential underrepresentation of certain trophoblast subsets. 

This is an important point as the challenges of studying a giant syncytial cell are often underappreciated by researchers that study mononucleated cells. We have added text to the introduction to clarify why traditional single cell RNA sequencing techniques were inadequate to collect  and characterize the STB (lines 91-93).

(4) Additionally, more evidence could be provided to support the claims about STB differentiation in the STB^out model and to determine whether its differentiation trajectory is unique or simply more advanced than in STBⁱⁿ. 

Our original conclusion that STB^out nuclei are more terminally differentiated than STBⁱⁿ was based on two observations: (1) STB^out TOs exhibit increased expression of STB-specific pregnancy hormones and many classic STB marker genes and (2) STB^out nuclei show an enrichment of the STB-3 nuclear subtype, which appears at the end of the slingshot pseudotime trajectory. However, upon consideration of the reviewer comments, we agree that this evidence is not sufficient to definitively distinguish if STB^out nuclei are more advanced or follow a unique differentiation trajectory dependent on new environmental cues. Pseudotime analyses provided only a predictive framework for lineage tracing, and these predictions must be experimentally validated. Real-time tracking of STB nuclear subtypes in TOs would require a suite of genetic tools beyond the scope of this study. Therefore, to address the reviewers' concerns we have removed language suggesting that STB-3 is a terminally differentiated subtype or that STB^out nuclei are more differentiated than STBⁱⁿ nuclei throughout the text until the discussion. Therein we present both our original hypothesis (that STB nuclei are further differentiated in STB^out) and alternative explanations like changing trajectories due to local environmental cues (lines 619-625).

Reviewer #2 (Public review): 

Strengths: 

(1) The use of SN and SC RNA sequencing provides a detailed analysis of STB formation and differentiation. 

(2) The identification of distinct STB subtypes and novel gene markers such as RYBP offers new insights into STB development. 

Thank you for highlighting these strengths—we appreciate your recognition of our use of SN and SC RNA sequencing to analyze STB differentiation and the discovery of distinct STB subtypes and novel gene markers like RYBP.

Weaknesses: 

(1) Inconsistencies in data presentation. 

We address the individual comments of reviewer 2 later in this response.

(2) Questionable interpretation of lncRNA signals: The use of long non-coding RNA (lncRNA) signals as cell type-specific markers may represent sequencing noise rather than true markers. 

We appreciate the reviewer’s attention to detail in noticing the lncRNA signature seen in many STB nuclear subtypes. However, we disagree that these molecules simply represent sequencing noise. In fact, may studies have rigorously demonstrated that lncRNAs have both cell and tissue specific gene expression (e.g., Zhao et al 2022, Isakova et al 2021, Zheng et al 2020). Further, they have been shown to be useful markers of unique cell types during development (e.g., Morales-Vicente et al 2022, Zhou et al 2019, Kim et al 2015) and can enhance clustering interpretability in breast cancer (Malagoli et al 2024). Many lncRNAs have also been demonstrated to play a functional role in the human placenta, including H19, MEG3, and MEG8 (Adu-Gyamfi et al 2023) and differences are even seen in nuclear subtypes in trophoblast stem cells (Khan et al 2021). Therefore, we prefer to keep these lncRNA signatures included and let future researchers test their functional role.

To improve the study's validity and significance, it is crucial to address the inconsistencies and to provide additional evidence for the claims. Supplementing with immunofluorescence staining for validating the distribution of STB_in, STB_out, and EVT_enrich in the organoid models is recommended to strengthen the results and conclusions. 

Each general trophoblast cell type (CTB, STB, EVT) has been visualized by immunofluorescence by the Coyne laboratory in their initial papers characterizing the STBⁱⁿ, STB^out, and EVT^enrich models (Yang et al, 2022 and 2023). We agree that it is important to validate the STB nuclear subtypes found in our genomic study. However, one challenge in studying a syncytia is that immunofluorescence may not be a definitive method when the nuclei share a common cytoplasm. This is because protein products from mRNAs transcribed in one nucleus are translated in the cytoplasm and could diffuse beyond sites of transcription. Therefore, RNA fluorescence in situ hybridization (RNA-FISH) is instead needed. While a systematic characterization of the spatial distribution of the many marker genes found each subtype is outside the scope of this study, we include RNA-FISH of one STB-2 marker (PAPPA2) and one STB-3 marker (ADAMTS6) in Figure 3F-G and Supplemental Figure 3.3. This demonstrates there is an increase in STB-2 marker gene expression in STBⁱⁿ TOs and an increase in STB-3 marker gene expression in STB^out TOs. 

Reviewer #3 (Public review):  

The authors present outstanding progress toward their aim of identifying, "the underlying control of the syncytiotrophoblast". They identify the chromatin remodeler, RYBP, as well as other regulatory networks that they propose are critical to syncytiotrophoblast development. This study is limited in fully addressing the aim, however, as functional evidence for the contributions of the factors/pathways to syncytiotrophoblast cell development is needed. Future experimentation testing the hypotheses generated by this work will define the essentiality of the identified factors to syncytiotrophoblast development and function. 

We thank the reviewer for their thoughtful assessment, constructive feedback, and encouraging comments. We acknowledge that the initial manuscript primarily presented analyses suggesting correlations between RYBP and other factors identified in the gene network analysis and STB function. Understanding how gene networks in the STB are formed and regulated is a long-term goal that will require many experiments with collaborative efforts across multiple research groups.

Nonetheless, to address this concern we have knocked out two key genes, RYBP and AFF1, in TOs using CRISPR-Cas9-mediated gene targeting. Bulk RNA sequencing of STBⁱⁿ TOs from both wild-type (WT) and knockout strains revealed that deletion of either gene caused a statistically significant decrease in the expression of the pregnancy hormone human placental lactogen and an increase in the expression of several genes characteristic of the oxygen-sensing STB-2 subtype, including FLT-1, PAPPA2, SPON2, and SFXN3. These findings demonstrate that knocking out RYBP or AFF1 results in an increase in STB-2 marker gene expression and therefore play a role in inhibiting their expression in WT TOs (Figure 5D-E and supplemental Figure 5.2). We also note that this is the first application of CRISPR-mediated gene silencing in a TO model.

Future work will visualize the distribution of STB nuclear subtypes in these mutants and explore the mechanistic role of RYBP and AFF1 in STB nuclear subtype formation and maintenance. However, these investigations fall outside the scope of the current study.

Localization and validation of the identified factors within tissue and at the protein level will also provide further contextual evidence to address the hypotheses generated. 

We agree that visualizing STB nuclear subtype distribution is essential for testing the many hypotheses generated by our analysis. To address this, we have included RNA-FISH experiments for two STB subtype markers (PAPPA2 for STB-2 and ADAMTS6 for STB-3) in TOs. These experiments reveal an increase in PAPPA2 expression in STBⁱⁿ TOs and an increase in ADAMTS6 expression in STB^out TOs (Figure 3F-G and Supplemental Figure 3.3). Genomic studies serve as powerful hypothesis generators, and we look forward to future work—both our own and that of other researchers—to validate the markers and hypotheses presented from our analysis.

Recommendations for the authors: 

Reviewing Editor Comments: 

We strongly encourage the authors to further strengthen the study by addressing all reviewers' comments and recommendations, with particular attention to the following key aspects:

(1) Clarifying the uniqueness of the STB differentiation trajectory between STBⁱⁿ and STB^out, and determining whether STB^out represents a more advanced stage of differentiation compared to STBⁱⁿ. It is also important to specify which developmental stage of placental villi the STB^out and STBⁱⁿ are simulating. 

We have revised the manuscript to remove definitive language claiming that STB-3 represents a terminally differentiated subtype or that STB^out nuclei are more differentiated than STBⁱⁿ nuclei. Instead, we now present our hypothesis and alternative explanations in the discussion (lines 619-625), and emphasize the need for experimental validation of pseudotime predictions to test these hypotheses.

(2) Utilizing immunofluorescence to validate the distribution of cell types in the organoid models. 

The Coyne lab has previously performed immunofluorescence of CTB and STB markers in STBⁱⁿ and STB^out TOs (Yang et al 2023). The syncytial nature of STBs complicates immunofluorescence-based validation of the STB nuclear subtypes due translating proteins all sharing a single common cytoplasm and therefore being able to diffuse and mix. Instead, we performed RNA-FISH for two STB subtype markers (PAPPA2, STB-2 and ADAMTS6, STB-3), which showed subtype-specific nuclear enrichment in STBⁱⁿ and STB^out TOs, respectively (Figure 3F-G and Supplemental Figure 3.3).  

(3) Addressing concerns regarding the use of lncRNA as cell marker genes. Employing canonical markers alongside critical TFs involved in differentiation pathways to perform a more robust cell-type analysis and validation is recommended.  

As discussed in detail above, we maintain that lncRNAs are valuable markers, supported by their demonstrated roles in cell and tissue specificity and placental function. These signatures provide important insights and hypotheses for future research, and we have clarified this rationale in the revised manuscript.

Reviewer #1 (Recommendations for the authors): 

(1) The authors have presented an extensive SC- and SN-based characterization of their improved trophoblast TO model, including a comparison to human placental tissues and the previous TO iteration. In this way, the authors' work represents an invaluable resource for investigators by providing thorough validation of the TO model and a clear description of the similarities and differences between primary and TO-derived STBs. I would suggest that the authors reshape the study to further highlight and emphasize this aspect of the study. 

We thank the reviewer for their thoughtful recommendation and agree that our datasets will serve as an invaluable resource for comparing in vitro models to in vivo gene expression. However, extensive validation is required to make definitive conclusions about the extent to which these systems mirror one another and where they diverge. For this reason, in this manuscript, we have focused on characterizing STB subtypes to provide a foundational understanding of the model and this poorly characterized subtype.

(2) Introduction, Paragraph 3: What is the importance of orientation for the trophoblast TO model? The authors may consider removing some of the less important methodologic details from this paragraph and including more emphasis on why their TO model is an improvement. 

Text has been added to this paragraph to highlight the importance of outward facing STB orientation, which is essential to mirror the STB’s transport function in vitro (lines 118-120).

(3) Results, Figure 1: In addition to the primary placental tissue plots showing all cell populations, it may be useful to have side-by-side versions of similar plots showing only the trophoblast subsets, so that the primary and TO data could be more easily compared visually. 

This has been implemented and added to the Supplemental Figure 1.4.

(4) Results, Figure 1: In simple terms, what is the reason for ending up with different cluster numbers/names from the primary tissue and TO? Would it be possible to apply the same annotation to each (at least for trophoblast types) and thus allow direct comparison between the two? 

As described above, each dataset was separately analyzed and clusters determined with an algorithm to determine the optimal clustering resolution. Therefore, the number of clusters between each dataset cannot be directly compared until the SN TO and tissue datasets are integrated together in Figure 6. We have added text to the manuscript to make it clear that they should not be compared except for in bulk number until this point (230-232).

(5) Results, Figure 2: For subsequent evaluation of different in vitro TO conditions, did the authors use only SN sequencing because they wanted to focus on STB? Based on Figure 1, it seems some CTB subsets would be underrepresented if using only SN. Given that the authors look at both STB and CTB in their different TOs, is this an issue? 

The CTB clusters that showed the greatest divergence between SC and SN datasets were those associated with mitosis and the cell cycle, likely due to nuclear envelope breakdown interfering with capture by the 10x microfluidics pipeline. While cytoplasmic gene expression provides valuable insights into CTB function, our manuscript focuses on the STB starting from Figure 2. Since the STB is captured exclusively by the SN dataset, we concentrated on this approach to streamline our analysis.

(6) Results, Figure 3: What do the authors consider to be the primary contributing factors for why the STB subsets display differential gene expression between STBⁱⁿ and STB^out? Is this due primarily to the cultural conditions and/or a result of the differing spatial arrangement with CTBs? 

This is an intriguing question that is challenging to disentangle because the culture conditions are integral to flipping the orientation. The two primary factors that differ between STBⁱⁿ and STB^out TOs are the presence of extracellular matrix in STBⁱⁿ and direct exposure to the surrounding media in STB^out. We believe these environmental cues play a significant role in shaping the gene expression of STB subsets. Fully disentangling this relationship would require a method to alter the TO orientation without changing the culture conditions. While this is an exciting direction for future research, it falls outside the scope of the present study.

(7) Results, Figure 4: The authors' analysis indicates that the STB nuclei from the STB^out TO are likely "more differentiated" than those in STBⁱⁿ TO. Could the authors provide some qualitative or quantitative support for this? Is the STB^out differentiated phenotype closer to what would be observed in a fully formed placenta? 

As discussed earlier, we agree with the reviewers that this claim should be removed from the text outside of the discussion.

(8) Results, Figure 5: Based on the trajectory analysis, do the authors consider that the STB from STB^out TO are simply further along the differentiation pathway compared to those from STBⁱⁿ TO, or do the STB from STB^out TO follow a differentiation pathway that is intrinsically distinct from STBⁱⁿ TO? 

We think the idea of an intrinsically distinct pathway is a fascinating alternative hypothesis and have added it into the discussion. We do not find the pseudotime currently allows us to answer this question without additional experiments, so we have removed claims that the STB^out STB nuclei are further along the differentiation pathway.

(9) Results, Figure 6: A notable difference between the STB^out TO and the term tissue is that the CTB subsets are much more prevalent. Is this simply a scale difference, i.e. due to the size of the human placenta compared to the limited STB nuclei available in the STB^out TO? Or are there other contributing factors? 

The proportion of CTB to STB nuclei in our term tissue (9:1) aligns with expectations based on stereological estimates. We believe the relatively low number of CTB nuclei in our dataset is due to the need for a larger sample size to capture more of this less abundant cell type. Since the primary focus of this paper is on STB, and we analyzed over 4,000 STB nuclei, we do not view this as a limitation. However, future studies utilizing SN to investigate term tissue should account for the abundance of STB nuclei and plan their sampling carefully to ensure sufficient representation of CTB nuclei if this is a desired focus.

Reviewer #2 (Recommendations for the authors): 

(1) The color annotations for cell types in Figure 2 are inconsistent between the different panels, and the term "Prolif" in Figure 2E is not explained by the authors. 

We chose colors to enhance visibility on the UMAP. We do not wish readers to make direct comparisons between the different CTB or STB subtypes of the sample types until the datasets are integrated in Figure 2D. This is because an algorithm for the clustering resolution has been chosen independently for each dataset. Cluster proportions are better compared in the integrated datasets in Figure 2D. We have added text to the results section to make this clear to the reader (lines 166-167).

(2) In Figure 3 and Supplementary Figures 1.3, the authors frequently present long non-coding RNA (lncRNA) signals as cell type-specific markers in the bubble plots. These signals are likely sequencing noise and may not accurately represent true markers for those cell types. It is recommended to revise this interpretation. 

As referenced above, there are many examples of lncRNAs that have biological and pathological significance in the placenta (H19, Meg3, Meg8) and lncRNAs often have cell type specific expression that can enhance clustering. We prefer to keep these signatures included and let future researchers determine their biological significance.

(3) In Figure 3C, the authors performed pathway enrichment analysis on the STB subtypes after integrating STB_in and STB_out organoids. The enrichment of the "transport across the blood-brain barrier" pathway in the STB-3 subtype does not align with the current understanding of STB cell function. Please provide corresponding supporting evidence. Additionally, please verify whether the other functional pathways represent functions specific to the STB subtypes. 

Interestingly, many of the genes categorized under “transport across the blood-brain barrier” are transporters shared with “vascular transport.” These include genes involved in the transport of amino acids (SLC7A1, SLC38A1, SLC38A3, SLC7A8), molecules essential for lipid metabolism (SLC27A4, SLC44A1), and small molecule exchange (SLC4A4, SLC5A6). Given that the vasculature, the STB, and the blood-brain barrier all perform critical barrier functions, it is unsurprising that molecules associated with these GO terms are enriched in the STB-3 subtype, which expresses numerous transporter proteins. Since the transport of materials across the STB is a well-established function, we have not included additional supporting evidence but have clarified the genes associated with this GO term in the text (lines 392-394 and supplemental Table 9).

(4) The pseudotime heatmap in Figure 4B is not properly arranged and is inconsistent with the differentiation relationships shown in Figure 4A. It is recommended to revise this. 

We are uncertain which aspect of the heatmap in Figure 4A is perceived as inconsistent with Figure 4B. One distinction is that pseudotime in Figure 4A is normalized from 0 to 100 to fit the blue-to-yellow-to-red color scale, whereas in Figure 4B, the color scale is not normalized and the color bar ranging from white to red. This difference reflects our intent to simplify Figure 4B-C, as the abundance of color between cell types and gene expression changes required a streamlined representation to ensure the figure remained clear and easy to interpret. This is classically done in the field and consistent with the default code in the slingshot package.

(5) In Figures 4C and 4D, although RYBP is highly expressed in STB, it is difficult to support the conclusion that RYBP shows the most significant expression changes. It is recommended to provide additional evidence. 

The claim that RYBP exhibits the most significant expression changes was based on p-value ordering of genes associated with pseudotime via the associationTest function in slingshot and not with immunofluorescence data. The text has been revised to make this distinction clear (lines 390-393).

(6) In Figure 4E, staining for CTB marker genes is missing, and in Figure 4F, CYTO is difficult to use as a classical STB marker. It is recommended to use the CGBs antibody from Figure 4E as a STB marker for staining to provide evidence.  

We have revised the Figure 5B-C to use e-Cadherin as a CTB marker gene in TOs and CGB antibody as a marker of STB.

In tissue, however, obtaining a good STB marker that does not overlap with the RYBP antibody (rabbit) in term tissue is difficult as the STB downregulates hCG expression closer to term to initiate contractions. SDC1 is often used but only labels the plasma membrane so does not help in distinguishing the STB cytoplasm. We have added an image of cytokeratin, e-Cadherin, and the STB marker ENDOU to validate that our current approach with e-Cadherin and cytokeratin allows us to accurately distinguish between CTB and STB cells.

(7) The velocity results in Figure 5A do not align with the differentiation relationships between cells and contradict the pseudotime results presented in Figure 4 by the authors. 

The reviewer raises an interesting observation regarding the velocity map in Figure 5A, which appears to show a bifurcation into two STB subtypes. This observation aligns with similar findings reported in tissue by our colleagues (Wang et al., 2024). However, given the low number of CTB cells in our tissue dataset, we were cautious about making definitive conclusions about pseudotime without a larger sample size. Notably, the RNA velocity map closely resembles the pseudotime trajectory in TOs, with CTB transitioning into the CTB-pf subtype and subsequently into the STB. One potential explanation for discrepancies between tissue and TOs is the difference in nuclear age: nuclei in tissue can be up to nine months old, whereas those in TOs are only hours or days old. It is possible that the lineage in TOs could bifurcate if cultured for longer than 48 hours, but our current dataset captures only the early stages of the STB differentiation process. While exploring these hypotheses is fascinating, they are beyond the scope of this current study.  

Reviewer #3 (Recommendations for the authors): 

Amazing work - I greatly enjoyed reading the manuscript. Here are a few questions and suggestions for consideration: 

Evidence presented throughout the results sections hints that the organoids may represent an earlier stage of placental development compared to the term. Increased hCG gene expression is observed, but as noted expression is decreased in term STB. STB:CTB ratios are also higher at term compared to the first trimester, etc. It was difficult to conclude definitively based on how data is presented in Fig 6 and discussed. Maybe there is no clear answer. Perhaps the altered cell type ratios in the organoid models (e.g., few STB in EVT enrich conditions) impact recapitulation of the in vivo local microenvironment signaling. As such, can the authors speculate on whether cell ratios could be strategically leveraged to model different gestational time points? 

Along these same lines, syncytiotrophoblast in early implantation (before proper villi development) is often described as invasive and later at the tertiary villi stage defined by hormone production, barrier function, and nutrient/gas exchange. Do the authors think the different STB subtypes captured in the organoid models represent different stages/functions of syncytiotrophoblast in placental development? 

Minor Comments 

(1) Please clarify what the third number represents in the STB:CTB ratio (e.g., 1:3:1 and 2:5:1). EVT? 

The first number is a decimal point and not a colon (ie 1.3 and 2.5). Therefore these numbers are to be read as the STB:CTB ratio is 1.3 to 1 or 2.5 to 1.

(2) Could consider co-localizing RYBP in term tissue with a syncytio-specific marker like CGB used for organoids (Fig 4F). 

We addressed this concern in comment 6 to reviewer 2.

(3) Recommend defining colors-which colors represent which module in Figure 5C in the legend and main body text. I see the labels surrounding the heatmap in 5B, but defining colors in text (e.g. cyan, magenta, etc.) would be helpful. Do the gray circles represent targets that don't belong to a specific module? Are the bolded factor names based on a certain statistical cutoff/defining criteria or were they manually selected? 

The text of both the results and figure legends has been revised to clarify these points.

(4) Data Availability: It would be helpful to provide supplemental table files for analyses (e.g., 5C to list the overlapping relationships in TGs for each TF/CR (5C) and 3E/6F to list DEG genes in comparisons). 

Supplemental files for each analysis have been added (Supplemental Table 8-14). In addition, the raw and processed data is available on GEO and we have created an interactive Shiny App so people without coding experience can interact with each dataset (lines 917-919).

(5) “...and found that each sample expressed these markers (Figure 6D), suggesting..." Consider clarifying "these". 

Text has been added to refer to a few of these marker genes within the text (line 540).

Citations

(1) Zappia L, Oshlack A. Clustering trees: a visualization for evaluating clusterings at multiple resolutions. GigaScience. 2018;7(7):giy083. PMCID: PMC6057528

(2) Zhou J, Xu J, Zhang L, Liu S, Ma Y, Wen X, Hao J, Li Z, Ni Y, Li X, Zhou F, Li Q, Wang F, Wang X, Si Y, Zhang P, Liu C, Bartolomei M, Tang F, Liu B, Yu J, Lan Y. Combined Single-Cell Profiling of lncRNAs and Functional Screening Reveals that H19 Is Pivotal for Embryonic Hematopoietic Stem Cell Development. Cell Stem Cell. 2019;24(2):285-298.e5. PMID: 30639035

(3) Malagoli G, Valle F, Barillot E, Caselle M, Martignetti L. Identification of Interpretable Clusters and Associated Signatures in Breast Cancer Single-Cell Data: A Topic Modeling Approach. Cancers. 2024;16(7):1350. PMCID: PMC11011054

(4) Adu-Gyamfi EA, Cheeran EA, Salamah J, Enabulele DB, Tahir A, Lee BK. Long non-coding RNAs: a summary of their roles in placenta development and pathology†. Biol Reprod. 2023;110(3):431–449. PMID: 38134961

(5) Zheng M, Hu Y, Gou R, Nie X, Li X, Liu J, Lin B. Identification three LncRNA prognostic signature of ovarian cancer based on genome-wide copy number variation. Biomed Pharmacother. 2020;124:109810. PMID: 32000042

(6) Khan T, Seetharam AS, Zhou J, Bivens NJ, Schust DJ, Ezashi T, Tuteja G, Roberts RM. Single Nucleus RNA Sequence (snRNAseq) Analysis of the Spectrum of Trophoblast Lineages Generated From Human Pluripotent Stem Cells in vitro. Front Cell Dev Biol. 2021;9:695248. PMCID: PMC8334858

(7) Isakova A, Neff N, Quake SR. Single-cell quantification of a broad RNA spectrum reveals unique noncoding patterns associated with cell types and states. Proc Natl Acad Sci United States Am. 2021;118(51):e2113568118. PMCID: PMC8713755

(8) Morales-Vicente DA, Zhao L, Silveira GO, Tahira AC, Amaral MS, Collins JJ, Verjovski-Almeida S. Singlecell RNA-seq analyses show that long non-coding RNAs are conspicuously expressed in Schistosoma mansoni gamete and tegument progenitor cell populations. Front Genet. 2022;13:924877. PMCID: PMC9531161

(9) Kim DH, Marinov GK, Pepke S, Singer ZS, He P, Williams B, Schroth GP, Elowitz MB, Wold BJ. Single-Cell

Transcriptome Analysis Reveals Dynamic Changes in lncRNA Expression during Reprogramming. Cell Stem Cell. 2015;16(1):88–101. PMCID: PMC4291542

(10) Yang L, Liang P, Yang H, Coyne CB. Trophoblast organoids with physiological polarity model placental structure and function. bioRxiv. 2023;2023.01.12.523752. PMCID: PMC9882188

Author response:

General Statements

In our manuscript, we demonstrate for the first time that RNA Polymerase I (Pol I) can prematurely release nascent transcripts at the 5' end of ribosomal DNA transcription units in vivo. This achievement was made possible by comparing wild-type Pol I with a mutant form of Pol I, hereafter called SuperPol previously isolated in our lab (Darrière at al., 2019). By combining in vivo analysis of rRNA synthesis (using pulse-labelling of nascent transcript and cross-linking of nascent transcript - CRAC) with in vitro analysis, we could show that Superpol reduced premature transcript release due to altered elongation dynamics and reduced RNA cleavage activity. Such premature release could reflect regulatory mechanisms controlling rRNA synthesis. Importantly, This increased processivity of SuperPol is correlated with resistance with BMH-21, a novel anticancer drugs inhibiting Pol I, showing the relevance of targeting Pol I during transcriptional pauses to kill cancer cells. This work offers critical insights into Pol I dynamics, rRNA transcription regulation, and implications for cancer therapeutics.

We sincerely thank the three reviewers for their insightful comments and recognition of the strengths and weaknesses of our study. Their acknowledgment of our rigorous methodology, the relevance of our findings on rRNA transcription regulation, and the significant enzymatic properties of the SuperPol mutant is highly appreciated. We are particularly grateful for their appreciation of the potential scientific impact of this work. Additionally, we value the reviewer’s suggestion that this article could address a broad scientific community, including in transcription biology and cancer therapy research. These encouraging remarks motivate us to refine and expand upon our findings further.

All three reviewers acknowledged the increased processivity of SuperPol compared to its wildtype counterpart. However, two out of three questions our claims that premature termination of transcription can regulate ribosomal RNA transcription. This conclusion is based on SuperPol mutant increasing rRNA production. Proving that modulation of early transcription termination is used to regulate rRNA production under physiological conditions is beyond the scope of this study. Therefore, we propose to change the title of this manuscript to focus on what we have unambiguously demonstrated:

“Ribosomal RNA synthesis by RNA polymerase I is subjected to premature termination of transcription”.

Reviewer 1 main criticisms centers on the use of the CRAC technique in our study. While we address this point in detail below, we would like to emphasize that, although we agree with the reviewer’s comments regarding its application to Pol II studies, by limiting contamination with mature rRNA, CRAC remains the only suitable method for studying Pol I elongation over the entire transcription units. All other methods are massively contaminated with fragments of mature RNA which prevents any quantitative analysis of read distribution within rDNA.  This perspective is widely accepted within the Pol I research community, as CRAC provides a robust approach to capturing transcriptional dynamics specific to Pol I activity. 

We hope that these findings will resonate with the readership of your journal and contribute significantly to advancing discussions in transcription biology and related fields.

(1) Description of the planned revisions

Despite numerous text modification (see below), we agree that one major point of discussion is the consequence of increased processivity in SuperPol mutant on the “quality” of produced rRNA. Reviewer 3 suggested comparisons with other processive alleles, such as the rpb1-E1103G mutant of the RNAPII subunit (Malagon et al., 2006). This comparison has already been addressed by the Schneider lab (Viktorovskaya OV, Cell Rep., 2013 - PMID: 23994471), which explored Pol II (rpb1-E1103G) and Pol I (rpa190-E1224G). The rpa190-E1224G mutant revealed enhanced pausing in vitro, highlighting key differences between Pol I and Pol II catalytic ratelimiting steps (see David Schneider's review on this topic for further details).

Reviewer 2 and 3 suggested that a decreased efficiency of cleavage upon backtracking might imply an increased error rate in SuperPol compared to the wild-type enzyme. Pol I mutant with decreased rRNA cleavage have been characterized previously, and resulted in increased errorrate. We already started to address this point. Preliminary results from in vitro experiments suggest that SuperPol mutants exhibit an elevated error rate during transcription. However, these findings remain preliminary and require further experimental validation to confirm their reproducibility and robustness. We propose to consolidate these data and incorporate into the manuscript to address this question comprehensively. This could provide valuable insights into the mechanistic differences between SuperPol and the wild-type enzyme. SuperPol is the first pol I mutant described with an increased processivity in vitro and in vivo, and we agree that this might be at the cost of a decreased fidelity.

Regulatory aspect of the process:

To address the reviewer’s remarks, we propose to test our model by performing experiments that would evaluate PTT levels in Pol I mutant’s or under different growth conditions. These experiments would provide crucial data to support our model, which suggests that PTT is a regulatory element of Pol I transcription. By demonstrating how PTT varies with environmental factors, we aim to strengthen the hypothesis that premature termination plays an important role in regulating Pol I activity.

We propose revising the title and conclusions of the manuscript. The updated version will better reflect the study's focus and temper claims regarding the regulatory aspects of termination events, while maintaining the value of our proposed model.

(2) Description of the revisions that have already been incorporated in the transferred manuscript

Some very important modifications have now been incorporated:

Statistical Analyses and CRAC Replicates:

Unlike reviewers 2 and 3, reviewer 1 suggests that we did not analyze the results statistically. In fact, the CRAC analyses were conducted in biological triplicate, ensuring robustness and reproducibility. The statistical analyses are presented in Figure 2C, which highlights significant findings supporting the fact WT Pol I and SuperPol distribution profiles are different. We CRAC replicates exhibit a high correlation and we confirmed significant effect in each region of interest (5’ETS, 18S.2, 25S.1 and 3’ ETS, Figure 1) to confirm consistency across experiments. We finally took care not to overinterpret the results, maintaining a rigorous and cautious approach in our analysis to ensure accurate conclusions.

CRAC vs. Net-seq:

Reviewer 1 ask to comment differences between CRAC and Net-seq. Both methods complement each other but serve different purposes depending on the biological question on the context of transcription analysis. Net-seq has originally been designed for Pol II analysis. It captures nascent RNAs but does not eliminate mature ribosomal RNAs (rRNAs), leading to high levels of contamination. While this is manageable for Pol II analysis (in silico elimination of reads corresponding to rRNAs), it poses a significant problem for Pol I due to the dominance of rRNAs (60% of total RNAs in yeast), which share sequences with nascent Pol I transcripts. As a result, large Net-seq peaks are observed at mature rRNA extremities (Clarke 2018, Jacobs 2022). This limits the interpretation of the results to the short lived pre-rRNA species. In contrast, CRAC has been specifically adapted by the laboratory of David Tollervey to map Pol I distribution while minimizing contamination from mature rRNAs (The CRAC protocol used exclusively recovers RNAs with 3′ hydroxyl groups that represent endogenous 3′ ends of nascent transcripts, thus removing RNAs with 3’-Phosphate, found in mature rRNAs). This makes CRAC more suitable for studying Pol I transcription, including polymerase pausing and distribution along rDNA, providing quantitative dataset for the entire rDNA gene.

CRAC vs. Other Methods:

Reviewer 1 suggests using GRO-seq or TT-seq, but the experiments in Figure 2 aim to assess the distribution profile of Pol I along the rDNA, which requires a method optimized for this specific purpose. While GRO-seq and TT-seq are excellent for measuring RNA synthesis and cotranscriptional processing, they rely on Sarkosyl treatment to permeabilize cellular and nuclear membranes. Sarkosyl is known to artificially induces polymerase pausing and inhibits RNase activities which are involved in the process. To avoid these artifacts, CRAC analysis is a direct and fully in vivo approach. In CRAC experiment, cells are grown exponentially in rich media and arrested via rapid cross-linking, providing precise and artifact-free data on Pol I activity and pausing.

Pol I ChIP Signal Comparison:

The ChIP experiments previously published in Darrière et al. lack the statistical depth and resolution offered by our CRAC analyses. The detailed results obtained through CRAC would have been impossible to detect using classical ChIP. The current study provides a more refined and precise understanding of Pol I distribution and dynamics, highlighting the advantages of CRAC over traditional methods in addressing these complex transcriptional processes.

BMH-21 Effects:

As highlighted by Reviewer 1, the effects of BMH-21 observed in our study differ slightly from those reported in earlier work (Ref Schneider 2022), likely due to variations in experimental conditions, such as methodologies (CRAC vs. Net-seq), as discussed earlier. We also identified variations in the response to BMH-21 treatment associated with differences in cell growth phases and/or cell density. These factors likely contribute to the observed discrepancies, offering a potential explanation for the variations between our findings and those reported in previous studies. In our approach, we prioritized reproducibility by carefully controlling BMH-21 experimental conditions to mitigate these factors. These variables can significantly influence results, potentially leading to subtle discrepancies. Nevertheless, the overall conclusions regarding BMH-21's effects on WT Pol I are largely consistent across studies, with differences primarily observed at the nucleotide resolution. This is a strength of our CRAC-based analysis, which provides precise insights into Pol I activity.

We will address these nuances in the revised manuscript to clarify how such differences may impact results and provide context for interpreting our findings in light of previous studies.

Minor points:

Reviewer #1:

•  In general, the writing style is not clear, and there are some word mistakes or poor descriptions of the results, for example: 

•  On page 14: "SuperPol accumulation is decreased (compared to Pol I)". 

•  On page 16: "Compared to WT Pol I, the cumulative distribution of SuperPol is indeed shifted on the right of the graph." 

We clarified and increased the global writing style according to reviewer comment.

•  There are also issues with the literature, for example: Turowski et al, 2020a and Turowski et al, 2020b are the same article (preprint and peer-reviewed). Is there any reason to include both references? Please, double-check the references.  

This was corrected in this version of the manuscript.

•  In the manuscript, 5S rRNA is mentioned as an internal control for TMA normalisation. Why are Figure 1C data normalised to 18S rRNA instead of 5S rRNA? 

Data are effectively normalized relative to the 5S rRNA, but the value for the 18S rRNA is arbitrarily set to 100%.

•  Figure 4 should be a supplementary figure, and Figure 7D doesn't have a y-axis labelling. 

The presence of all Pol I specific subunits (Rpa12, Rpa34 and Rpa49) is crucial for the enzymatic activity we performed. In the absence of these subunits (which can vary depending on the purification batch), Pol I pausing, cleavage and elongation are known to be affected. To strengthen our conclusion, we really wanted to show the subunit composition of the purified enzyme. This important control should be shown, but can indeed be shown in a supplementary figure if desired.

Y-axis is figure 7D is now correctly labelled

•  In Figure 7C, BMH-21 treatment causes the accumulation of ~140bp rRNA transcripts only in SuperPol-expressing cells that are Rrp6-sensitive (line 6 vs line 8), suggesting that BHM-21 treatment does affect SuperPol. Could the author comment on the interpretation of this result? 

The 140 nt product is a degradation fragment resulting from trimming, which explains its lower accumulation in the absence of Rrp6. BMH21 significantly affects WT Pol I transcription but has also a mild effect on SuperPol transcription. As a result, the 140 nt product accumulates under these conditions.

Reviewer #2:

•  pp. 14-15: The authors note local differences in peak detection in the 5'-ETS among replicates, preventing a nucleotide-resolution analysis of pausing sites. Still, they report consistent global differences between wild-type and SuperPol CRAC signals in the 5'ETS (and other regions of the rDNA). These global differences are clear in the quantification shown in Figures 2B-C. A simpler statement might be less confusing, avoiding references to a "first and second set of replicates" 

According to reviewer, statement has been simplified in this version of the manuscript.

•  Figures 2A and 2C: Based on these data and quantification, it appears that SuperPol signals in the body and 3' end of the rDNA unit are higher than those in the wild type. This finding supports the conclusion that reduced pausing (and termination) in the 5'ETS leads to an increased Pol I signal downstream. Since the average increase in the SuperPol signal is distributed over a larger region, this might also explain why even a relatively modest decrease in 5'ETS pausing results in higher rRNA production. This point merits discussion by the authors. 

We agree that this is a very important discussion of our results. Transcription is a very dynamic process in which paused polymerase is easily detected using the CRAC assay. Elongated polymerases are distributed over a much larger gene body, and even a small amount of polymerase detected in the gene body can represent a very large rRNA synthesis. This point is of paramount importance and, as suggested by the reviewer, is now discussed in detail.

•  A decreased efficiency of cleavage upon backtracking might imply an increased error rate in SuperPol compared to the wild-type enzyme. Have the authors observed any evidence supporting this possibility? 

Reviewer suggested that a decreased efficiency of cleavage upon backtracking might imply an increased error rate in SuperPol compared to the wild-type enzyme. We already started to address this point. Preliminary results from in vitro experiments suggest that SuperPol mutants exhibit an elevated error rate during transcription. However, these findings remain preliminary and require further experimental validation to confirm their reproducibility and robustness. We propose to consolidate these data and incorporate into the manuscript to address this question comprehensively.

•  pp. 15 and 22: Premature transcription termination as a regulator of gene expression is welldocumented in yeast, with significant contributions from the Corden, Brow, Libri, and Tollervey labs. These studies should be referenced along with relevant bacterial and mammalian research. 

According to reviewer suggestion, we referenced these studies.

•  p. 23: "SuperPol and Rpa190-KR have a synergistic effect on BMH-21 resistance." A citation should be added for this statement. 

This represents some unpublished data from our lab. KR and SuperPol are the only two known mutants resistant to BMH-21. We observed that resistance between both alleles is synergistic, with a much higher resistance to BMH-21 in the double mutant than in each single mutant (data not shown). Comparing their resistance mechanisms is a very important point that we could provide upon request. This was added to the statement.

•  p. 23: "The released of the premature transcript" - this phrase contains a typo 

This is now corrected.

Reviewer #3:

•  Figure 1B: it would be opportune to separate the technique's schematic representation from the actual data. Concerning the data, would the authors consider adding an experiment with rrp6D cells? Some RNAs could be degraded even in such short period of time, as even stated by the authors, so maybe an exosome depleted background could provide a more complete picture. Could also the authors explain why the increase is only observed at the level of 18S and 25S? To further prove the robustness of the Pol I TMA method could be good to add already characterized mutations or other drugs to show that the technique can readily detect also well-known and expected changes. 

The precise objective of this experiment is to avoid the use of the Rrp6 mutant. Under these conditions, we prevent the accumulation of transcripts that would result from a maturation defect. While it is possible to conduct the experiment with the Rrp6 mutant, it would be impossible to draw reliable conclusions due to this artificial accumulation of transcripts.

•  Figure 1C: the NTS1 probe signal is missing (it is referenced in Figure 1A but not listed in the Methods section or the oligo table). If this probe was unused, please correct Figure 1A accordingly. 

We corrected Figure 1A.  

•  Figure 2A: the RNAPI occupancy map by CRAC is hard to interpret. The red color (SuperPol) is stacked on top of the blue line, and we are not able to observe the signal of the WT for most of the position along the rDNA unit. It would be preferable to use some kind of opacity that allows to visualize both curves. Moreover, the analysis of the behavior of the polymerase is always restricted to the 5'ETS region in the rest of the manuscript. We are thus not able to observe whether termination events also occur in other regions of the rDNA unit. A Northern blot analysis displaying higher sizes would provide a more complete picture. 

We addressed this point to make the figure more visually informative. In Northern Blot analysis, we use a TSS (Transcription Start Site) probe, which detects only transcripts containing the 5' extremity. Due to co-transcriptional processing, most of the rRNA undergoing transcription lacks its 5' extremity and is not detectable using this technique. We have the data, but it does not show any difference between Pol I and SuperPol. This information could be included in the supplementary data if asked.

•  "Importantly, despite some local variations, we could reproducibly observe an increased occupancy of WT Pol I in 5'-ETS compared to SuperPol (Figure 1C)." should be Figure 2C. 

Thanks for pointing out this mistake. it has been corrected.

•  Figure 3D: most of the difference in the cumulative proportion of CRAC reads is observed in the region ~750 to 3000. In line with my previous point, I think it would be worth exploring also termination events beyond the 5'-ETS region. 

We agree that such an analysis would have been interesting. However, with the exception of the pre-rRNA starting at the transcription start site (TSS) studied here, any cleaved rRNA at its 5' end could result from premature termination and/or abnormal processing events. Exploring the production of other abnormal rRNAs produced by premature termination is a project in itself, beyond this initial work aimed at demonstrating the existence of premature termination events in ribosomal RNA production.

•  Figure 4: should probably be provided as supplementary material. 

As l mentioned earlier (see comments), the presence of all Pol I specific subunits (Rpa12, Rpa34 and Rpa49) is crucial for the enzymatic activity we performed. This important control should be shown, but can indeed be shown in a supplementary figure if desired.

•  "While the growth of cells expressing SuperPol appeared unaffected, the fitness of WT cells was severely reduced under the same conditions." I think the growth of cells expressing SuperPol is slightly affected. 

We agree with this comment and we modified the text accordingly.

•  Figure 7D: the legend of the y-axis is missing as well as the title of the plot. 

Legend of the y-axis and title of the plot are now present.

•  The statements concerning BMH-21, SuperPol and Rpa190-KR in the Discussion section should be removed, or data should be provided.

This was discussed previously. See comment above.

•  Some references are missing from the Bibliography, for example Merkl et al., 2020; Pilsl et al., 2016a, 2016b. 

Bibliography is now fixed

(3) Description of analyses that authors prefer not to carry out

Does SuperPol mutant produces more functional rRNAs ?

As Reviewer 1 requested, we agree that this point requires clarification.. In cells expressing SuperPol, a higher steady state of (pre)-rRNAs is only observed in absence of degradation machinery suggesting that overproduced rRNAs are rapidly eliminated. We know that (pre)rRNas are unable to accumulate in absence of ribosomal proteins and/or Assembly Factors (AF). In consequence, overproducing rRNAs would not be sufficient to increase ribosome content. This specific point is further address in our lab but is beyond the scope of this article.

Is premature termination coupled with rRNA processing 

We appreciate the reviewer’s insightful comments. The suggested experiments regarding the UTP-A complex's regulatory potential are valuable and ongoing in our lab, but they extend beyond the scope of this study and are not suitable for inclusion in the current manuscript.

Author response:

The following is the authors’ response to the original reviews

Reviewer #1 (Public Review):

Summary:

This study investigates the hypoxia rescue mechanisms of neurons by non-neuronal cells in the brain from the perspective of exosomal communication between brain cells. Through multi-omics combined analysis, the authors revealed this phenomenon and logically validated this intercellular rescue mechanism under hypoxic conditions through experiments. The study proposed a novel finding that hemoglobin maintains mitochondrial function, expanding the conventional understanding of hemoglobin. This research is highly innovative, providing new insights for the treatment of hypoxic encephalopathy.

Overall, the manuscript is well organized and written, however, there are some minor/major points that need to be revised before this manuscript is accepted.

We thank the reviewer for the detailed analysis of our study. Please find our answers to the points raised by the reviewer below.

Major points:

(1) Hypoxia can induce endothelial cells to release exosomes carrying hemoglobin, however, how neurons are able to actively take up these exosomes? It is possible for other cells to take up these exosomes also? This point needs to be clarified in this study.

We sincerely appreciate the reviewer’s valuable comments. Regarding the question of how neurons actively uptake extracellular vesicles (EVs) carrying hemoglobin mRNA, existing studies suggest that EVs can enter cells via three main pathways: direct fusion, receptor-mediated endocytosis, and phagocytosis (PMID: 25288114). Our experimental results show that neurons are able to actively uptake EVs from endothelial cells without any treatment, and hypoxic conditions did not significantly increase the uptake of endothelial EVs by neurons (Fig. 5A and I). As for the specific uptake mechanism, there is currently no definitive conclusion. Some studies have found that hypoxic-ischemic injury may induce neurons to upregulate Cav-1, which could enhance the uptake of endothelial-derived EVs via Cav-1-mediated endocytosis (PMID: 31740664), but this mechanism still requires further validation.

Regarding whether other cell types also take up these EVs, we focused on neurons based on existing literature and our own data, which show that the increased hemoglobin in the brain under hypoxic conditions is primarily found in neurons (Fig. 4H-J, PMID: 19116637). Moreover, we observed that, under hypoxic conditions, almost all non-neuronal supporting cells in the brain transcribe hemoglobin in large amounts and release it via EVs (Fig. 3J). Furthermore, we would like to emphasize that although neurons do not transcribe hemoglobin, we observed substantial expression of hemoglobin within neurons. This suggests that it may serve as an important protective mechanism for the brain. Therefore, the focus of our study is on the protective effect of EVs carrying hemoglobin mRNA on neurons, and the uptake by other cell types was not explored. We greatly appreciate the reviewer’s question, and we believe this is an intriguing avenue for further investigation. This could provide new insights for interventions in hypoxic brain injury, and we plan to delve into this topic in future studies.

(2) The expression of hemoglobin in neurons is important for mitochondrial homeostasis, but its relationship with mitochondrial homeostasis needs to be further elucidated in the study.

We sincerely appreciate the reviewer’s valuable comments. We fully agree with the importance of hemoglobin expression in neurons for mitochondrial homeostasis. In this study, we have confirmed through in vitro experiments that when neurons are treated with conditioned medium from endothelial cells, they exhibit increased hemoglobin expression. This, in turn, enhances their resistance to hypoxia by restoring mitochondrial membrane potential and increasing mitochondrial numbers, thereby effectively improving neuronal viability. Notably, this protective effect disappears when EVs are removed from the endothelial-conditioned medium or when hemoglobin in endothelial cells is disrupted, further supporting the notion that endothelial cells transfer hemoglobin via EVs, helping neurons express hemoglobin under hypoxic conditions and exert protective effects.

In summary, hemoglobin primarily helps maintain mitochondrial membrane potential, thereby supporting the restoration of energy metabolism and production under hypoxic conditions, which effectively improves the neuronal resistance to hypoxia. Although we were unable to explore the specific mechanisms of hemoglobin’s role in mitochondrial homeostasis in detail within this study, we recognize the importance of this aspect and plan to further investigate how hemoglobin regulates mitochondrial homeostasis and function in neurons in future research.

Once again, we greatly appreciate the reviewer’s insightful comments. We will continue to optimize our research direction and look forward to further elucidating these important biological mechanisms in future studies.

Minor points:

(1) In Figures 1-3, the authors use "Endo" to represent endothelial cells, while in Figures 4-7, the abbreviation "EC" is used. Please standardize the format.

Thank you for the reviewer’s suggestion. We will use “EC” consistently to refer to endothelial cells throughout the manuscript to ensure uniformity.

(2) In all qPCR statistical results, please italicize the gene names on the axis.

Thank you for the reviewer’s valuable suggestion. We will make sure to italicize the gene names on the axis in all qPCR statistical results to adhere to the formatting requirements.

(3) In the Western blot result of Figure 3C, what type of cell-derived exosomes does the Control group represent, and why can it be used as a control group for brain-derived exosomes?

Thank you for the reviewer’s insightful question. In Fig. 3C, the control group (Control) represents the cell lysate sample, which serves as a positive control in the EVs Western blot analysis. In this experiment, the positive control is primarily used to validate the specificity of the antibody and the accuracy of the experimental procedure. We used cell lysate as the control to confirm that the antibody can detect EV-associated markers in the cell lysates, thus providing a comparative basis for the identification of brain-derived EVs.

(4) In Figure 4F, the morphology of hemoglobin in the Con group and the H28d group is not entirely consistent with Figure 4H. Is this difference due to different experimental batches?

Thank you for the reviewer’s careful observation. The observed difference may indeed be due to variations between different experimental batches. To ensure consistency of the results, we have updated the representative immunofluorescence images, which are now presented in Fig. 4H.

(5) Supplement the transcription and expression levels of hemoglobin in neurons under different treatment conditions after medium exchange with exosome removal and medium exchange after HBA1 interference.

Thank you for the reviewer’s valuable suggestions. We have added the experimental data regarding the exchange of culture medium after the removal of EVs. As shown in Fig. S6, the endothelial-derived medium without EVs does not enhance the hemoglobin levels in neurons under hypoxic conditions. Additionally, we have included the detection results of hemoglobin expression in neurons after HBA1 interference, as shown in Fig. S7E-F. The results indicate that the culture medium derived from HBA1-interfered endothelial cells also fails to help neurons increase hemoglobin expression under hypoxic conditions.

(6) Figure S3 should be split to separately explain the increased exosome release induced by hypoxia, the non-toxic effect of endothelial cell culture medium on neurons, and the successful screening of the HBA1 interference plasmid.

Thank you for the reviewer’s suggestions. Based on your feedback, we have split the original Fig. S3 into multiple parts to more clearly present the different experimental results. Specifically, the results of hypoxia-induced EVs release increase have been updated in Fig. S4, the non-toxic effects of endothelial cell culture medium on neurons are shown in Fig. S5, and the successful screening of the HBA1 interference plasmid is presented in Fig. S7.

(7) Regarding the extracellular vesicles/exosomes, it should be expressed consistently in the whole manuscript.

Thank you for the reviewer’s reminder. We will ensure that the term “extracellular vesicles” is used consistently throughout the manuscript.

(8) In lines 70 and 80, the O2 should be changed to "O₂".

Thank you for the reviewer’s careful observation. We have corrected the formatting of “O2” to “O₂” in lines 70 and 80.

We would like to thank the Reviewer for taking the time to thoroughly examine our work, for their helpful feedback that has significantly contributed to improving our manuscript, and for their kind and encouraging words.

Reviewer #2 (Public Review):

Summary:

This is an interesting study with a lot of data. Some of these ideas are intriguing. But a few major points require further consideration.

We thank the reviewer for the detailed assessment of our study and pinpointing its current weaknesses. Please find our answers to all comments below.

Major points:

(1) What disease is this model of whole animal hypoxia supposed to mimic? If one is focused on the brain, can one just use a model of focal or global cerebral ischemia?

Thank you for the reviewer’s insightful question. The chronic hypoxia model we employed is designed to mimic the multi-organ damage caused by systemic hypoxia, which is relevant to clinical conditions such as high-altitude hypoxia, chronic obstructive pulmonary disease, and acute hypoxic brain injury. In contrast to focal or global cerebral ischemia models, the focus of our study is on how the brain, under extreme systemic hypoxia, utilizes endothelial cell-derived extracellular vesicles (EVs) to transfer hemoglobin mRNA, thereby protecting neurons and aiding the brain’s response to hypoxia-induced damage.

We understand the reviewer’s concern that focal or global ischemia models are typically used to simulate localized brain hypoxia or ischemic injury. However, the core of our research is to explore the brain’s overall adaptive mechanisms under systemic hypoxic conditions. By using a systemic hypoxia model, we can more comprehensively simulate the effects of global hypoxia on the brain and uncover how the brain engages specific molecular mechanisms for self-protection. This approach offers a novel perspective on brain hypoxic-ischemic diseases and holds potential clinical applications, particularly in the study of stroke, vascular cognitive impairment and dementia (VCID), and related conditions.

Additionally, we have observed that hemoglobin significantly increases in the brain in an animal model of focal ischemia (as shown in Author response image 1 below). This finding further supports the idea that hemoglobin upregulation may be a universal protective mechanism for the brain’s response to hypoxic damage. While this part of the research is still ongoing, preliminary results suggest that both systemic hypoxia and focal ischemia might trigger protective effects through hemoglobin regulation.

Author response image 1.

The expression level of Hba-a1 in the brain of VCID mouse.

Therefore, the core of our study is to elucidate the brain’s self-protection mechanisms under systemic hypoxia, rather than focusing solely on cerebral ischemia models. We believe this approach provides new insights into the prevention and treatment of brain hypoxic-ischemic diseases, with significant clinical application potential.

In light of this, we have added a related discussion to the manuscript, clearly explaining the rationale for choosing the systemic hypoxia model. The updated content can be found on P11, Line 13-21 as follows: “To investigate this phenomenon, we employed a chronic hypoxia model in which mice were exposed to 7% oxygen for 28 days. This model aims to mimic systemic hypoxia-induced multi-organ damage, a condition observed in diseases such as high-altitude hypoxia, chronic obstructive pulmonary disease, and acute hypoxic brain injury. The primary goal of this model is to explore how the brain adapts under extreme low-oxygen conditions and employs specific mechanisms to protect itself from hypoxia-induced damage. This approach provides valuable insight into diseases related to hypoxic-ischemic injury in the brain, including stroke and vascular dementia, offering a novel perspective for potential clinical applications.”

(2) If this model subjects the entire animal to hypoxia, then other organs will also be hypoxic. Should one also detect endothelial upregulation and release of extracellular vesicles containing hemoglobin mRNA in non-CNS organs? Where do these vesicles go? Into blood?

Thank you for the reviewer’s valuable feedback. Indeed, in a whole-body hypoxia model, other organs are also affected by hypoxia. Therefore, future research may need to investigate the upregulation of endothelial cells in organs other than the central nervous system, as well as the release of EVs containing hemoglobin mRNA from these organs. However, in this study, we isolated EVs from the brain tissue in situ following perfusion with physiological saline, a method that effectively eliminates the influence of EVs from blood or other organs. As a result, our primary focus was on studying how EVs released by brain endothelial cells are actively taken up by neurons to exert neuroprotective effects. The potential for these EVs to enter the bloodstream and their subsequent fate is indeed a topic worthy of further investigation. Future research could offer new insights into the cross-organ effects of systemic hypoxia.

(3) What other mRNA are contained in the vesicles released from brain endothelial cells?

Thank you for the reviewer’s valuable suggestions. We have further analyzed EVs derived from brain endothelial cells, and in addition to hemoglobin mRNA, these EVs also contain a variety of other mRNAs, including Vwf, Hbb-bt, Hba-a1, Hbb-bs, Hba-a2, Acer2, Angpt2, Ldha, Gm42418, Slc16a1, Cxcl12, B2m, Ctla2a, Ccnd1, and Hmgcs2 (Log2FC > 1.2). The biological processes associated with these mRNAs primarily involve: cell-substrate adhesion, regulation of cellular amide metabolic process, negative regulation of cell migration, negative regulation of cell motility, and negative regulation of cellular component movement. These processes may be closely related to the neuroprotective effects of endothelial cell EVs in a hypoxic environment, especially in terms of regulating cell behavior and maintaining cell structure and function. Additionally, these EVs contain multiple key factors associated with intracellular metabolism, movement, and migration, which may collectively influence neuronal function and survival. Notably, our study also found that mRNA of various hemoglobin subunits ranks among the top five in terms of abundance in the mRNA secreted by hypoxic endothelial EVs, further emphasizing the importance of hemoglobin mRNA in endothelial-derived EVs. Therefore, future research may explore the functions of these mRNAs and reveal how they act in concert to protect neurons from hypoxia-induced damage.

We have updated and added these results in Fig. S4, and have further elaborated on the findings in the revised figure. Once again, we thank the reviewer for the attention and valuable suggestions regarding our work.

(4) Where do the endothelial vesicles go? Only to neurons? Or to other cells as well?

Thank you for the reviewer’s important question. As previously mentioned, the focus of this study is to investigate how EVs carrying hemoglobin mRNA influence neuronal function. Through a combined analysis of single-cell transcriptomics and EV transcriptomics from brain tissue, we found that, besides neurons, almost all types of supportive cells in the brain and their secreted EVs contain a significant amount of hemoglobin mRNA (Fig. 3J, 4B). Notably, although neurons do not transcribe hemoglobin mRNA themselves, under hypoxic conditions, neurons significantly increase hemoglobin expression, resulting in a phenomenon where the transcription and expression levels of hemoglobin in neurons are inconsistent. This phenomenon has been observed both in our study and others (Fig. 4H-J, PMID: 19116637). This observation led us to focus on the active uptake of EVs by neurons and the potential neuroprotective effects they might bring.

Regarding whether other cell types uptake these EVs and their potential functions, although our current research is focused on neurons, this is indeed an important area for further investigation. Given that non-neuronal supportive cells may also transfer hemoglobin mRNA via EVs under hypoxic conditions, future research will further explore the uptake of EVs by different cell types and their roles in hypoxic adaptation.

We are particularly interested in the hemoglobin expression in neurons under hypoxic conditions and consider neurons to be the primary expressers of hemoglobin, providing a new perspective for exploring the neuroprotective role of hemoglobin. We plan to delve deeper into these issues in future studies.

(5) Neurons can express endogenous hemoglobin. Is it useful to subject neurons to hypoxia and then see how much the endogenous mRNA goes up? How large is the magnitude of endogenous hemoglobin gene upregulation compared to the hypothesized exogenous mRNA that is supposed to be donated from endothelial vesicles?

Thank you for the reviewer’s valuable question. We have observed that, in the absence of treatment with endothelial cell-derived conditioned medium, there is no significant change in the transcription and expression levels of endogenous hemoglobin in neurons under hypoxic conditions (Fig. 5I, 6C-D). However, when neurons were treated with endothelial cell-conditioned medium, under the same hypoxic conditions, the transcription levels of hemoglobin increased by approximately 1.2-fold, and the expression levels increased by approximately 3.8-fold (Fig. 6B-D). Additionally, we have added pre-treatment experiments involving EVs depletion from the endothelial culture medium and HBA interference. The results show that, after these two pre-treatments, the conditioned medium lost its ability to enhance the transcription and expression of hemoglobin in neurons under hypoxic conditions (Fig. S6, S7D-F), further emphasizing the important role of endothelial EVs in this process. This finding indicates that endothelial-derived EVs significantly promote hemoglobin expression in neurons, and this effect is far greater than the upregulation of endogenous hemoglobin in neurons. Therefore, while neurons can express endogenous hemoglobin, exogenous hemoglobin significantly enhances its expression, which may help neurons tolerate the hypoxic environment and provide additional protection.

(6) Finally, it may be useful to provide more information and data to explain how the expression of this exogenous endothelial-derived hemoglobin binds to neuronal mitochondria to alter function.

Thank you for the reviewer’s valuable suggestion. As we previously mentioned, hemoglobin plays a protective role in neurons by maintaining mitochondrial membrane potential, helping neurons restore energy metabolism and energy production under hypoxic conditions. We fully agree on the importance of this research direction. Several studies have shown that when hemoglobin is expressed in neurons, it predominantly localizes to mitochondria, which aligns with the physiological process of heme synthesis within mitochondria (PMID: 23187133). Furthermore, in the brains of Parkinson’s disease patients, the localization of hemoglobin in neuronal mitochondria is altered compared to normal conditions (PMID: 27181046). Therefore, the interaction between hemoglobin and mitochondria plays a crucial role in neuronal function.

Although existing research indicates the role of hemoglobin in neuronal mitochondria, studies in this area remain limited. We plan to further investigate how hemoglobin binds to mitochondria and its specific effects on mitochondrial function in our future work. We believe that a deeper understanding of this mechanism will provide essential theoretical insights into the effects of hypoxia on neurons and offer new potential strategies for neuroprotective therapies.

We would like to thank the Reviewer for taking the time to thoroughly examine our work, for their helpful feedback that has significantly contributed to improving our manuscript, and for their kind and encouraging words.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

In this study, the authors introduced an essential role of AARS2 in maintaining cardiac function. They also investigated the underlying mechanism that through regulating alanine and PKM2 translation are regulated by AARS2. Accordingly, a therapeutic strategy for cardiomyopathy and MI was provided. Several points need to be addressed to make this article more comprehensive:

Thank this reviewer for the overall supports on our manuscript.

(1) Include apoptotic caspases in Figure 2B, and Figure 4 B and E as well.

This is a good point for further investigating the role of apoptosis signaling in cardiac-specific AARS2 knockout hearts. Since we are focusing on cardiomyocyte phenotypes, immunostaining on TUNEL and anti-cTnT directly evaluated the level of cardiomyocyte apoptosis, which was supported by Western blots with anti-Bcl-2 and anti-BAX of control and mutant hearts. TUNEL data accurately represents biochemical and morphological characteristics of apoptotic cells, and is more sensitive than the conventional histochemical and biochemical methods. Future studies are needed to address how apoptosis components including apoptotic caspases are involved in cardiomyocyte apoptosis in AARS2 mutant hearts.

(2) It would be better to show the change of apoptosis-related proteins upon the knocking down of AARS2 by small interfering RNA (siRNA).

Since primary culture of neonatal cardiomyocytes also contained non-cardiomyocytes, using Western blots with anti-apoptosis proteins cannot directly assess cardiomyocytes phenotypes. In this work, our data on the elevation of cTnT⁺/TUNEL⁺ cardiomyocytes and cardiac fibrosis in AARS2 mutant hearts suggest that AARS2 deficiency induced cardiomyocyte death.

(3) In Figure 5, the authors performed Mass Spectrometry to assess metabolites of homogenates. I was wondering if the change of other metabolites could be provided in the form of a heatmap.

Indeed, we assessed other metabolites by mass spectrometry as shown below, we found that overexpression of AARS2 in either transgenic mouse hearts or neonatal cardiomyocytes had no consistent changes on the level of fumarate, succinate, malate, alpha-ketoglutarate (alpha-KG), citrate, oxaloacetate (OAA), ATP, and ADP, thus suggesting that AARS2 overexpression has more specific effect on the level of lactate, pyruvate, and acetyl-CoA.

Author response image 1.

(4) The amounts of lactate should be assessed using a lactate assay kit to validate the Mass Spectrometry results.

We carried out several rounds of mass spectrometry experiments, suggesting that lactate is consistently elevated after AARS2 overexpression in neonatal cardiomyocytes as shown below. We will establish other lactate assays in future studies.

Author response image 2.

(5) How about the expression pattern of PKM2 before and after mouse MI. Furtherly, the correlation between AARS2 and PKM2?

Previous studies have shown that the expression level of PKM2 in mice is significantly increased after cardiac surgery at different time points, which may be related to cardiometabolic changes [1]. Our co-IP experiments showed no direct interactions between AARS2 and PKM2 (Figure 6K), while both AARS2 proteins and mRNA decreased on the 3 days (Figure 1A-B) and 7 days (Author response image 3)after myocardial infarction in mice. Thus, the level of AARS2 is reversely related to PKM2 after myocardial infarction.

Author response image 3.

(6) In Figure 5, how about the change of apoptosis-related proteins after administration of PKM2 activator TEPP-46?

It has been shown that TEPP-46 treatment decreased cardiomyocyte death in different models that induced cardiomyocyte apoptosis [2, 3]. We would like to refer these published works that TEPP-46 treatment improves heart function by inhibiting cardiac injury-induced cardiomyocyte death.

Reviewer #2 (Public Review):

Summary:

The authors aimed to elucidate the role of AARS2, an alanyl-tRNA synthase, in mouse hearts, specifically its impact on cardiac function, fibrosis, apoptosis, and metabolic pathways under conditions of myocardial infarction (MI). By investigating the effects of both deletion and overexpression of AARS2 in cardiomyocytes, the study aims to determine how AARS2 influences cardiac health and survival during ischemic stress.

The authors successfully achieved their aims by demonstrating the critical role of AARS2 in maintaining cardiomyocyte function under ischemic conditions. The evidence presented, including genetic manipulation results, functional assays, and mechanistic studies, robustly supports the conclusion that AARS2 facilitates cardiomyocyte survival through PKM2-mediated metabolic reprogramming. The study convincingly links AARS2 overexpression to improved cardiac outcomes post-MI, validating the proposed protective AARS2-PKM2 signaling pathway.

This work may have a significant impact on the field of cardiac biology and ischemia research. By identifying AARS2 as a key player in cardiomyocyte survival and metabolic regulation, the study opens new avenues for therapeutic interventions targeting this pathway. The methods used, particularly the cardiomyocyte-specific genetic models and ribosome profiling, are valuable tools that can be employed by other researchers to investigate similar questions in cardiac physiology and pathology.

Understanding the metabolic adaptations in cardiomyocytes during ischemia is crucial for developing effective treatments for MI. This study highlights the importance of metabolic flexibility and the role of specific enzymes like AARS2 in facilitating such adaptations. The identification of the AARS2-PKM2 axis adds a new layer to our understanding of cardiac metabolism, suggesting that enhancing glycolysis can be a viable strategy to protect the heart from ischemic damage.

We thank this reviewer for his/her supports on our manuscript.

Strengths:

(1) Comprehensive Genetic Models: The use of cardiomyocyte-specific AARS2 knockout and overexpression mouse models allowed for precise assessment of AARS2's role in cardiac cells.

(2) Functional Assays: Detailed phenotypic analyses, including measurements of cardiac function, fibrosis, and apoptosis, provided evidence for the physiological impact of AARS2 manipulation.

(3) Mechanistic Insights: This study used ribosome profiling (Ribo-Seq) to uncover changes in protein translation, specifically highlighting the role of PKM2 in metabolic reprogramming.

(4) Therapeutic Relevance: The use of the PKM2 activator TEPP-46 to reverse the effects of AARS2 deficiency presents a potential therapeutic avenue, underscoring the practical implications of the findings.

Weaknesses:

(1) Species Limitation: The study is limited to mouse and rat models, and while these are highly informative, further validation in human cells or tissues would strengthen the translational relevance.

We fully agree with this reviewer that this study is limited to mouse and rat models. It would certainly be important to address how AARS2-PKM2 is related myocardial infarction patients in the future.

(2) Temporal Dynamics: The study does not extensively address the temporal dynamics of AARS2 expression and PKM2 activity during the progression of MI and recovery, which could offer deeper insights into the timing and regulation of these processes.

Thanks for this critical point. Indeed, we found that both AARS2 proteins and mRNA decreased on 3 days (Figure 1A-B) and 7 days (Author response image 3) after myocardial infarction in mice as shown below. Others have reported PKM2 proteins increased after heart surgery in mice at different time points [1]. Thus, the level of AARS2 is reversely related to PKM2 after myocardial infarction.

Reviewer #3 (Public Review):

In the present study, the author revealed that cardiomyocyte-specific deletion of mouse AARS2 exhibited evident cardiomyopathy with impaired cardiac function, notable cardiac fibrosis, and cardiomyocyte apoptosis. Cardiomyocyte-specific AARS2 overexpression in mice improved cardiac function and reduced cardiac fibrosis after myocardial infarction (MI), without affecting cardiomyocyte proliferation and coronary angiogenesis. Mechanistically, AARS2 overexpression suppressed cardiomyocyte apoptosis and mitochondrial reactive oxide species production, and changed cellular metabolism from oxidative phosphorylation toward glycolysis in cardiomyocytes, thus leading to cardiomyocyte survival from ischemia and hypoxia stress. Ribo-Seq revealed that AARS2 overexpression increased pyruvate kinase M2 (PKM2) protein translation and the ratio of PKM2 dimers to tetramers that promote glycolysis. Additionally, PKM2 activator TEPP-46 reversed cardiomyocyte apoptosis and cardiac fibrosis caused by AARS2 deficiency. Thus, this study demonstrates that AARS2 plays an essential role in protecting cardiomyocytes from ischemic pressure via fine-tuning PKM2-mediated energy metabolism, and presents a novel cardiac protective AARS2-PKM2 signaling during the pathogenesis of MI. This study provides some new knowledge in the field, and there are still some questions that need to be addressed in order to better support the authors' views.

We thank this reviewer for his/her overall supports on our manuscript.

(1) WGA staining showed obvious cardiomyocyte hypertrophy in the AARS2 cKO heart. Whether AARS affects cardiac hypertrophy needs to be further tested.

WGA staining is widely used to measure the size of cardiomyocytes in the literature. Here, we found that the size of mutant cardiomyocytes increased by ~20% after AARS2 knockout. In addition, we also measured and found that the ratio of heart to body weight increased in AARS2 mutant mice compared with control siblings as shown below.

Author response image 4.

(2) The authors observed that AARS2 can improve myocardial infarction, and whether AARS2 has an effect on other heart diseases.

Thanks for this critical point. We agree with this reviewer that it will be important to address whether overexpression of AARS2 has cardiac protection in other heart diseases such as transverse aortic constriction in the future.

(3) Studies have shown that hypoxia conditions can lead to mitochondrial dysfunction, including abnormal division and fusion. AARS2 also affects mitochondrial division and fusion and interacts with mitochondrial proteins, including FIS and DRP1, the authors are suggested to verify.

This is a good point. Mitochondrial dysfunction occurs when cardiomyocytes are subjected to hypoxia conditions such as myocardial infarction. Our ribosome sequencing data suggested that overexpression of AARS2 had no effect on the level of FIS1 and DRP2 as shown below. We agree with this reviewer that future studies are needed to clarify potential interactions between AARS2 and FIS/DRP1 proteins.

Author response image 5.

(4) The authors only examined the role of AARS2 in cardiomyocytes, and fibroblasts are also an important cell type in the heart. Authors should examine the expression and function of AARS2 in fibroblasts.

We fully agree with this reviewer that AARS2 may also function in cardiac fibroblasts since it is expressed in fibroblasts and cardiomyocyte-specific AARS2 knockout led to more fibrosis after myocardial infarction, which certainly warrant future investigations.

(5) Overexpression of AARS2 can inhibit the production of mtROS, and has a protective effect on myocardial ischemia and H/ R-induced injury, and the occurrence of iron death is also closely related to ROS, whether AARS protects myocardial by regulating the occurrence of iron death?

Thank this reviewer for his/her critical point. Our current data cannot rule out whether iron-mediated death is involved in AARS2 function in cardiac protection, which warrant future investigations.

(6) Please revise the English grammar and writing style of the manuscript, spelling and grammatical errors should be excluded.

Sorry for spelling and grammatical errors. We have carefully revised this manuscript now.

(7) Recent studies have shown that a decrease in oxygen levels leads to an increase in AARS2, and lactic acid rises rapidly without being oxidized. Both of these factors inhibit oxidative phosphorylation and muscle ATP production by increasing mitochondrial lactate acylation, thereby inhibiting exercise capacity and preventing the accumulation of reactive oxygen species ROS. The key role of protein lactate acylation modification in regulating oxidative phosphorylation of mitochondria, and the importance of metabolites such as lactate regulating cell function through feedback mechanisms, i.e. cells adapt to low oxygen through metabolic regulation to reduce ROS production and oxidative damage, and therefore whether AARS2 in the heart also acts in this way.

This is an interesting question. Since overexpression of AARS2 in muscles has previously been reported to increase PDHA1 lactylation and decrease its activity [4]. Actually, we initially examined whether overexpression of AARS2 in cardiomyocytes has similar effect on PDHA1 lactylation. However, our results showed that overexpression of AARS2 had no evident increases of lactylated PDHA1 in cardiomyocytes as shown below. However, future studies are needed to explore whether other proteins lactylation by AARS2 are involved in its cardiac protection function.

Author response image 6.

Reviewer #2 (Recommendations For The Authors):

Suggestions for Improved or Additional Experiments, Data, or Analyses:

(1) Validation in Human Models: It would be great if, in the future, the authors could conduct experiments with human cardiomyocytes derived from induced pluripotent stem cells (iPSCs) to validate the findings in a human context. This would strengthen the translational relevance of the results.

We fully agree with this reviewer that this study is limited to mouse and rat models. It would certainly be important to address how AARS2-PKM2 is related myocardial infarction patients and/or human iPSC-derived cardiomyocytes in the future.

(2) Broader Metabolic Analysis: To perform comprehensive metabolic profiling (e.g., metabolomics) to identify other metabolic pathways influenced by AARS2 overexpression or deficiency. This could provide a more holistic view of the metabolic changes and potential compensatory mechanisms.

As noted above, we indeed assessed other metabolites by mass spectrometry, we found that overexpression of AARS2 in either transgenic mouse hearts or neonatal cardiomyocytes had no consistent changes on the level of fumarate, succinate, malate, alpha-ketoglutarate (alpha-KG), citrate, oxaloacetic acid (OAA), ATP, and ADP, thus suggesting that AARS2 overexpression has more specific effect on the level of lactate, pyruvate, and acetyl-CoA.

(3) Temporal Dynamics: Investigate the temporal expression and activity of AARS2 and PKM2 during the progression and recovery phases of myocardial infarction. Time-course studies could elucidate the dynamics and regulatory mechanisms involved.

As noted above, we found that both AARS2 proteins and mRNA decreased on the third and seventh day after myocardial infarction in mice. Others have reported PKM2 proteins increased after heart surgery in mice at different time points [1]. Thus, the level of AARS2 is reversely related to PKM2 after myocardial infarction.

(4) Investigate Additional Pathways: Explore the involvement of other signaling pathways and tRNA synthetases that might interact with or complement the AARS2-PKM2 axis. This could uncover broader regulatory networks affecting cardiomyocyte survival and function.

Thank this reviewer for his/her critical point. This certainly warrants future investigations.

(5) Mitochondrial Function Assays: Perform detailed mitochondrial function assays, including measurements of mitochondrial respiration and membrane potential, to further elucidate the role of AARS2 in mitochondrial health and function under stress conditions.

We fully agree with this reviewer that future studies are needed to address how AARS2 is involved in mitochondrial function.

(6) Single-Cell Analysis: Utilize single-cell RNA sequencing to examine the heterogeneity in cardiomyocyte responses to AARS2 manipulation, providing insights into cell-specific adaptations and potential differential effects within the heart tissue.

We fully agree with this reviewer that it is important to address how AARS2 (cKO or overexpression) regulate cardiomyocyte heterogeneity and function in the future. 

Recommendations for Improving the Writing and Presentation:

(1) Visual Aids: Include more schematic diagrams to illustrate the proposed mechanisms, especially the AARS2-PKM2 signaling pathway and its impact on metabolic reprogramming. This can help readers better understand complex interactions.

Below is our working hypothesis on the role of AARS2 in cardiac protection. AARS2 deficiency caused mitochondrial dysfunction due to increasing ROS production and apoptosis while decreasing PKM2 function and glycolysis, thus leading to cardiomyopathy in mutant mice.  On the other hand, overexpression of AARS2 in mice activates PKM2 and glycolysis while decreases ROS production and apoptosis, thus improving heart function after myocardial infarction.

Author response image 7.

(2) Discussion: Shorten the Discussion and systematically address the significance of the findings, limitations of the study, and potential future directions. This will provide a clearer narrative and context for the results.

We have now made revisions on the Discussion part to highlight the significance of this work and brief perspective of future direction.

(3) Minor corrections to the text and figures.

We have now revised the full text carefully.

(4) Typographical Errors: Carefully proofread the manuscript to correct any typographical errors and ensure consistent use of terminology and abbreviations throughout the text.

Thanks. Based on the reviewer’s suggestions, we have carefully revised the manuscript and have done proof-reading on the whole manuscript.

Availability of data, code, reagents, research ethics, or other issues:

(1) Data Presentation: Ensure that all graphs and charts are clearly labeled with appropriate units, scales, and legends. Use color schemes that are accessible to color-blind readers.

We followed these rules to present the data.

(2) Supplementary Information: Provide detailed supplementary information, including raw data, experimental protocols, and analysis scripts, to enhance the reproducibility of the study.

We provided the raw data, experimental protocols, and analysis scripts in the manuscript.

(3) Data and Code Availability. Data Sharing: Authors should ensure that all raw data, processed data, and relevant metadata are deposited in publicly accessible repositories. Provide clear instructions on how to access these data. Code Availability: Make all analysis code available in a public repository, such as GitHub, with adequate documentation to allow other researchers to replicate the analyses.

We have deposited RNA-Seq data at ArrayExpress (E-MTAB-13767). We have also uploaded the original data in the supplementary file.

(4) Research Ethics and Compliance. Ethics Statement: Include a detailed statement on the ethical approval obtained for animal experiments, specifying the institution and ethical review board that granted approval. Conflict of Interest: Clearly state any potential conflicts of interest and funding sources that supported the research to ensure transparency.

Thanks. In the manuscript we made an ethical statement, stating conflicts of interest and sources of funding.

References:

(1) Y. Tang, M. Feng, Y. Su, T. Ma, H. Zhang, H. Wu, X. Wang, S. Shi, Y. Zhang, Y. Xu, S. Hu, K. Wei, D. Xu, Jmjd4 Facilitates Pkm2 Degradation in Cardiomyocytes and Is Protective Against Dilated Cardiomyopathy, Circulation, 147 (2023) 1684-1704.

(2) L. Guo, L. Wang, G. Qin, J. Zhang, J. Peng, L. Li, X. Chen, D. Wang, J. Qiu, E. Wang, M-type pyruvate kinase 2 (PKM2) tetramerization alleviates the progression of right ventricle failure by regulating oxidative stress and mitochondrial dynamics, Journal of translational medicine, 21 (2023) 888.

(3) B. Saleme, V. Gurtu, Y. Zhang, A. Kinnaird, A.E. Boukouris, K. Gopal, J.R. Ussher, G. Sutendra, Tissue-specific regulation of p53 by PKM2 is redox dependent and provides a therapeutic target for anthracycline-induced cardiotoxicity, Science translational medicine, 11 (2019).

(4) Y. Mao, J. Zhang, Q. Zhou, X. He, Z. Zheng, Y. Wei, K. Zhou, Y. Lin, H. Yu, H. Zhang, Y. Zhou, P. Lin, B. Wu, Y. Yuan, J. Zhao, W. Xu, S. Zhao, Hypoxia induces mitochondrial protein lactylation to limit oxidative phosphorylation, Cell research, 34 (2024) 13-30.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

Summary:

In this paper, Manley and Vaziri investigate whole-brain neural activity underlying behavioural variability in zebrafish larvae. They combine whole brain (single cell level) calcium imaging during the presentation of visual stimuli, triggering either approach or avoidance, and carry out whole brain population analyses to identify whole brain population patterns responsible for behavioural variability. They show that similar visual inputs can trigger large variability in behavioural responses. Though visual neurons are also variable across trials, they demonstrate that this neural variability does not degrade population stimulus decodability. Instead, they find that the neural variability across trials is in orthogonal population dimensions to stimulus encoding and is correlated with motor output (e.g. tail vigor). They then show that behavioural variability across trials is largely captured by a brain-wide population state prior to the trial beginning, which biases choice - especially on ambiguous stimulus trials. This study suggests that parts of stimulus-driven behaviour can be captured by brain-wide population states that bias choice, independently of stimulus encoding.

Strengths:

-The strength of the paper principally resides in the whole brain cellular level imaging in a well-known but variable behaviour.

- The analyses are reasonable and largely answer the questions the authors ask.

- Overall the conclusions are well warranted.

Weaknesses:

A more in-depth exploration of some of the findings could be provided, such as:

- Given that thousands of neurons are recorded across the brain a more detailed parcelation of where the neurons contribute to different population coding dimensions would be useful to better understand the circuits involved in different computations.

We thank the reviewer for noting the strengths of our study and agree that these findings have raised a number of additional avenues which we intend to explore in depth in future studies. In response to the reviewer’s comment above, we have added a number of additional figure panels (new Figures S1E, S3F-G, 4I(i), 4K(i), and S5F-G) and updated panels (Figures 4I(ii) and 4K(ii) in the revised manuscript) to show a more detailed parcellation of the visually-evoked neurons, noise modes, turn direction bias population, and responsiveness bias population. To do so. we have aligned our recordings to the Z-Brain atlas (Randlett et al., 2015) as shown in new Figure S1E. In addition, we provided a more detailed parcellation of the neuronal ensembles by providing projections of the full 3D volume along the xy and yz axes, in addition to the unregistered xy projection shown in Figures 4H and 4J in the revised manuscript. We also found that the distribution of neurons across our huc:h2b-gcamp6s recordings is very similar to the distribution of labeling in the huc:h2b-rfp reference image from the Z-Brain atlas (Figure S1E), which further supports our whole-brain imaging results.

Overall, we find that this more detailed quantification and visualization is consistent with our interpretations. In particular, we show that the optimal visual decoding population (w_opt) and the largest noise mode (e1) are localized to the midbrain (Figures S3F-G). This is expected, as in Figure 3 we first extracted a low-dimensional subspace of whole-brain neural activity that optimally preserved visual information. Additionally, we provide new evidence that the populations correlated with the turn bias and responsiveness bias are distributed throughout the brain, including a relatively dense localization to the cerebellum, telencephalon, and dorsal diencephalon (habenula, new Figures 4H-K and S5F-G).

- Given that the behaviour on average can be predicted by stimulus type, how does the stimulus override the brain-wide choice bias on some trials? In other words, a better link between the findings in Figures 2 and 3 would be useful for better understanding how the behaviour ultimately arises.

We agree with the reviewer that one of the most fundamental questions that this study has raised is how the identified neuronal populations predictive of decision variables (which we describe as an internal “bias”) interact with the well-studied, visually-evoked circuitry. A major limitation of our study is that the slow dynamics of the NL-GCaMP6s prevent clearly distinguishing any potential difference in the onset time of various neurons during the short trials, which might provide clues into which neurons drive versus later reflect the motor output. However, given that these ensembles were also found to be correlated with spontaneous turns, our hypothesis is that these populations reflect brain-wide drives that enable efficient exploration of the local environment (Dunn et al. 2016, doi.org/10.7554/eLife.12741). Further, we suspect that a sufficiently strong stimulus drive (e.g., large, looming stimuli) overrides these ongoing biases, which would explain the higher average pre-stimulus predictability in trials with small to intermediate-sized stimuli. An important follow-up line of experimentation could involve comparing the neuronal dynamics of specific components of the visual circuitry at distinct internal bias states, ideally utilizing emerging voltage indicators to maximize spatiotemporal specificity. For example, what is the difference between trials with a large looming stimulus in the left visual fields when the turn direction bias indicates a leftward versus rightward drive?

- What other motor outputs do the noise dimensions correlate with?

To better demonstrate the relationship between neural noise modes and motor activity that we described, we have provided a more detailed correlation analysis in new Figure S4A. We extracted additional features related to the larva’s tail kinematics, including tail vigor, curvature, principal components of curvature, angular velocity, and angular acceleration (S4A(i)). Some of these behavioral features were correlated with one another; for example, in the example traces, PC1 appears to capture nearly the same behavioral feature as tail vigor. The largest noise modes showed stronger correlations with motor output than the smaller noise modes, which is reminiscent recent work in the mouse showing that some of the neural dimensions with highest variance were correlated with various behavioral features (Musall et al. 2019; Stringer et al. 2019; Manley et al. 2024). We anticipate additional motor outputs would exhibit correlations with neural noise modes, such as pectoral fin movements (not possible to capture in our preparation due to immobilization) and eye movements.

The dataset that the authors have collected is immensely valuable to the field, and the initial insights they have drawn are interesting and provide a good starting ground for a more expanded understanding of why a particular action is determined outside of the parameters experimenters set for their subjects.

We thank the reviewer for noting the value of our dataset and look forward to future efforts motivated by the observations in our study.

Reviewer #2 (Public Review):

Overview

In this work, Manley and Vaziri investigate the neural basis for variability in the way an animal responds to visual stimuli evoking prey-capture or predator-avoidance decisions. This is an interesting problem and the authors have generated a potentially rich and relevant data set. To do so, the authors deployed Fourier light field microscopy (Flfm) of larval zebrafish, improving upon prior designs and image processing schemes to enable volumetric imaging of calcium signals in the brain at up to 10 Hz. They then examined associations between neural activity and tail movement to identify populations primarily related to the visual stimulus, responsiveness, or turn direction - moreover, they found that the activity of the latter two populations appears to predict upcoming responsiveness or turn direction even before the stimulus is presented. While these findings may be valuable for future more mechanistic studies, issues with resolution, rigor of analysis, clarity of presentation, and depth of connection to the prior literature significantly dampen enthusiasm.

Imaging

- Resolution: It is difficult to tell from the displayed images how good the imaging resolution is in the brain. Given scattering and lensing, it is important for data interpretation to have an understanding of how much PSF degrades with depth.

We thank the reviewer for their comments and agree that the dependence of the PSF and resolution as a function of depth is an important consideration in light field imaging. To quantify this, we measured the lateral resolution of the fLFM as a function of distance from the native image plane (NIP) using a USAF target. The USAF target was positioned at various depths using an automated z-stage, and the slice of the reconstructed volume corresponding to that depth was analyzed. An element was considered resolved if the modulation transfer function (MTF) was greater than 30%.

In new Figure S1A, we plot the resolution measurements of the fLFM as compared to the conventional LFM (Prevedel et al., 2014), which shows the increase in resolution across the axial extent of imaging. In particular, the fLFM does not exhibit the dramatic drop in lateral resolution near the NIP which is seen in conventional LFM. In addition, the expanded range of high-resolution imaging motivates our increase from an axial range of 200 microns in previous studies to 280 microns in this study.

- Depth: In the methods it is indicated that the imaging depth was 280 microns, but from the images of Figure 1 it appears data was collected only up to 150 microns. This suggests regions like the hypothalamus, which may be important for controlling variation in internal states relevant to the behaviors being studied, were not included.

The full axial range of imaging was 280 microns, i.e. spanning from 140 microns below to 140 microns above the native imaging plane. After aligning our recordings to the Z-Brain dataset, we have compared the 3D distribution of neurons in our data (new Figure S1E(i)) to the labeling of the reference brain (Figure S1E(ii)). This provides evidence that our imaging preparation largely captures the labeling seen in a dense, high-resolution reference image within the indicated 280 microns range.

- Flfm data processing: It is important for data interpretation that the authors are clearer about how the raw images were processed. The de-noising process specifically needs to be explained in greater detail. What are the characteristics of the noise being removed? How is time-varying signal being distinguished from noise? Please provide a supplemental with images and algorithm specifics for each key step.

We thank the reviewer for their comment. To address the reviewer’s point regarding the data processing pipeline utilized in our study, in our revised manuscript we have added a number of additional figure panels in Figure S1B-E to quantify and describe the various steps of the pipeline in greater depth.

First, the raw fLFM images are denoised. The denoising approach utilized in the fLFM data processing pipeline is not novel, but rather a custom-trained variant of Lecoq et al.’s (2021) DeepInterpolation method. In our original manuscript, we also described the specific architecture and parameters utilized to train our specific variation of DeepInterpolation model. To make this procedure clearer, we have added the following details to the methods:

“DeepInterpolation is a self-supervised approach to denoising, which denoises the data by learning to predict a given frame from a set of frames before and after it. Time-varying signal can be distinguished from shot noise because shot noise is independent across frames, but signal is not. Therefore, only the signal is able to be predicted from adjacent frames. This has been shown to provide a highly effective and efficient denoising method (Lecoq et al., 2021).”

Therefore, time-varying signal is distinguished from noise based on the correlations of pixel intensity across consecutive imaging frames. To better visualize this process, in new Figure S1B we show example images and fluorescence traces before and after denoising.

- Merging: It is noted that nearby pixels with a correlation greater than 0.7 were merged. Why was this done? Is this largely due to cross-contamination due to a drop in resolution? How common was this occurrence? What was the distribution of pixel volumes after aggregation? Should we interpret this to mean that a 'neuron' in this data set is really a small cluster of 10-20 neurons? This of course has great bearing on how we think about variability in the response shown later.

First, to be clear, nearby pixels were not merged; instead neuronal ROIs identified by CNMF-E were merged, as we had described: “the CNMF-E algorithm was applied to each plane in parallel, after which the putative neuronal ROIs from each plane were collated and duplicate neurons across planes were merged.” If this merging was not performed, the number of neurons would be overestimated due to the relatively dense 3D reconstruction with voxels of 4 m axially. Therefore, this merging is a requisite component of the pipeline to avoid double counting of neurons, regardless of the resolution of the data.

However, we agree with the reviewer that the practical consequences of this merging were not previously described in sufficient detail. Therefore, in our revision we have added additional quantification of the two critical components of the merging procedure: the number of putative neuronal ROIs merged and the volume of the final 3D neuronal ROIs, which demonstrate that a neuron in our data should not be interpreted as a cluster of 10-20 neurons.

In new Figure S1C(i), we summarize the rate of occurrence of merging by assessing the number of putative 2D ROIs which were merged to form each final 3D neuronal ROI. Across n=10 recordings, approximately 75% of the final 3D neuronal ROIs involved no merging at all, and few instances involved merging more than 5 putative ROIs. Next, in Figure S1C(ii), we quantify the volume of the final 3D ROIs. To do so, we counted the number of voxels contributing to each final 3D neuronal ROI and multiplied that by the volume of a single voxel (2.4 x 2.4 x 4 µm³). The majority of neurons had a volume of less than 1000 µm<up>3</sup>, which corresponds to a spherical volume with a radius of roughly 6.2 m. In summary, both the merging statistics and volume distribution demonstrate that few neuronal ROIs could be consistent with “a small cluster of 10-20 neurons”.

- Bleaching: Please give the time constants used in the fit for assessing bleaching.

As described in the Methods, the photobleaching correction was performed by fitting a bi-exponential function to the mean fluorescence across all neurons. We have provided the time constants determined by these fits for n=10 recordings in new Figure S1D(i). In addition, we provided an example of raw mean activity, the corresponding bi-exponential fit, and the mean activity after correction in Figure S1D(ii). These data demonstrate that the dominant photobleaching effect is a steep decrease in mean signal at the beginning of the recording (represented by the estimated time constant τ₁), followed by a slow decay (τ₂).

Analysis

- Slow calcium dynamics: It does not appear that the authors properly account for the slow dynamics of calcium-sensing in their analysis. Nuclear-localized GCaMP6s will likely have a kernel with a multiple-second decay time constant for many of the cells being studied. The value used needs to be given and the authors should account for variability in this kernel time across cell types. Moreover, by not deconvolving their signals, the authors allow for contamination of their signal at any given time with a signal from multiple seconds prior. For example, in Figure 4A (left turns), it appears that much of the activity in the first half of the time-warped stimulus window began before stimulus presentation - without properly accounting for the kernel, we don't know if the stimulus-associated activity reported is really stimulus-associated firing or a mix of stimulus and pre-stimulus firing. This also suggests that in some cases the signals from the prior trial may contaminate the current trial.

We would like to respond to each of the points raised here by the reviewer individually.

(1) “It does not appear that the authors properly account for the slow dynamics of calcium-sensing in their analysis. Nuclear-localized GCaMP6s will likely have a kernel with a multiple-second decay time constant for many of the cells being studied. The value used needs to be given…”

We disagree with the reviewer’s claim that the slow dynamics of the calcium indicator GCaMP were not accounted for. While we did not deconvolve the neuronal traces with the GCaMP response kernel, in every step in which we correlated neural activity with sensory or motor variables, we convolved the stimulus or motor timeseries with the GCaMP kernel, as described in the Methods. Therefore, the expected delay and smoothing effects were accounted for when analyzing the correlation structure between neural and behavioral or stimulus variables, as well as during our various classification approaches. To better describe this, we have added the following description of the kernel to our Methods:

“The NL-GCaMP6s kernel was estimated empirically by aligning and averaging a number of calcium events. This kernel corresponds to a half-rise time of 400 ms and half-decay time of 4910 ms.”

This approach accounts for the GCaMP kernel when relating the neuronal dynamics to stimuli and behavior, while avoiding any artifacts that could be introduced from improper deconvolution or other corrections directly to the calcium dynamics. Deconvolution of calcium imaging data, and in particular nuclear-localized (NL) GCaMP6s, is not always a robust procedure. In particular, GCaMP6s has a much more nonlinear response profile than newer GCaMP variants such as jGCaMP8 (Zhang et al. 2023, doi:10.1038/s41586-023-05828-9), as the reviewer notes later in their comments. The nuclear-localized nature of the indicator used in our study also provides an additional nonlinear effect. Accounting for a nonlinear relationship between calcium concentration and fluorescence readout is significantly more difficult because such nonlinearities remove the guarantee that the optimization approaches generally used in deconvolution will converge to global extrema. This means that deconvolution assuming nonlinearities is far less robust than deconvolution using the linear approximation (Vogelstein et al. 2010, doi: 10.1152/jn.01073.2009). Therefore, we argue that we are not currently aware of any appropriate methods for deconvolving our NL-GCaMP6s data, and take a more conservative approach in our study.

We also argue that the natural smoothness of calcium imaging data is important for the analyses utilized in our study (Shen et al., 2022, doi:10.1016/j.jneumeth.2021.109431). Even if our data were deconvolved in order to estimate spike trains or more point-like activity patterns, such data are generally smoothed (e.g., by estimating firing rates) before dimensionality reduction, which is a core component of our neuronal population analyses. Further, Wei et al. (2020, doi:10.1371/journal.pcbi.1008198) showed in detail that deconvolved calcium data resulted in less accurate population decoding, whereas binned electrophysiological data and raw calcium data were equally accurate. When using other techniques, such as clustering of neuronal activity patterns (a method we do not employ in this study), spike and deconvolved calcium data were instead shown to be more accurate than raw calcium data. Therefore, we do not believe deconvolution of the neuronal traces is appropriate in this case without a better understanding of the NL-GCaMP6s response, and do not rely on the properties of deconvolution for our analyses. Still, we agree with the reviewer that one must be mindful of the GCaMP kernel when analyzing and interpreting these data, and therefore have noted the delayed and slow kinematics of the NL-GCaMP within our manuscript, for example: “To visualize the neuronal activity during a given trial while accounting for the delay and kinematics of the nuclear-localized GCaMP (NL-GCaMP) sensor, a duration of approximately 15 seconds is extracted beginning at the onset of the 3-second visual stimulus period.”

(2) “… and the authors should account for variability in this kernel time across cell types.”

In addition to the points raised above, we are not aware of any deconvolution procedures which have successfully shown the ability to account for variability in the response kernel across cell types in whole-brain imaging data when cell type is unknown a priori. Pachitariu et al. (2018, doi:10.1523/JNEUROSCI.3339-17.2018) showed that the best deconvolution procedures for calcium imaging data rely on a simple algorithm with a fixed kernel. Further, more complicated approaches either utilize either explicit priors about the calcium kernel or learn implicit priors using supervised learning, neither of which we would be able to confirm are appropriate for our dataset without ground truth electrophysiological spike data.

However, we agree with the reviewer that we must interpret the data while being mindful that there could be variability in this kernel across neurons, which is not accounted for in our fixed calcium kernel. We have added the following sentence to our revised manuscript to highlight this limitation:

“The used of a fixed calcium kernel does not account for any variability in the GCaMP response across cells, which could be due to differences such as cell type or expression level. Therefore, this analysis approach may not capture the full set of neurons which exhibit stimulus correlations but exhibit a different GCaMP response.”

(3) “without properly accounting for the kernel, we don't know if the stimulus-associated activity reported is really stimulus-associated firing or a mix of stimulus and pre-stimulus firing”

While we agree with the reviewer that the slow dynamics of the indicator will cause a delay and smoothing of the signal over time, we would like to point out that this effect is highly directional. In particular, we can be confident that pre-stimulus activity is not contaminated by the stimulus given the data we describe in the next point regarding the timing of visual stimuli relative to the GCaMP kernel. The reviewer is correct that post-stimulus firing can be mixed with pre-stimulus firing due to the GCaMP kernel. However, our key claims in Figure 4 center around turn direction and responsiveness biases, which are present even before the onset of the stimulus. Still, we have highlighted this delay and smoothing to readers in the updated version of our manuscript.

(4) “This also suggests that in some cases the signals from the prior trial may contaminate the current trial”

We have carefully chosen the inter-stimulus interval for maximum efficiency of stimulation, while ensuring that contamination from the previous stimulus is negligible. The inter-stimulus interval was chosen by empirically analyzing preliminary data of visual stimulation with our preparation. New Figure S3C shows the delay and slow kinematics due to our indicator; indeed, visually-evoked activity peaks after the end of the short stimulus period. Importantly, however, the visually-evoked activity is at or near baseline at the start of the next trial.

Finally, we would like to note that our stimulation protocol is randomized, as described in the Methods. Therefore, the previous stimulus has no correlation with the current stimulus, which would prevent any contamination from providing predictive power that could be identified by our visual decoding methods.

- Partial Least Squares (PLS) regression: The steps taken to identify stimulus coding and noise dimensions are not sufficiently clear. Please provide a mathematical description.

We have updated the Results and Methods sections of our revised manuscript to describe in more mathematical detail the approach taken to identify the relevant dimensions of neuronal activity:

“The comparison of the neural dimensions encoding visual stimuli versus trial-to-trial noise was modeled after Rumyantsev et al. (2020). Partial least squares (PLS) regression was used to find a low-dimensional space that optimally predicted the visual stimuli, which we refer to as the visually-evoked neuronal activity patterns. To perform regression, a visual stimulus kernel was constructed by summing the timeseries of each individual stimulus type, weighted by the stimulus size and negated for trials on the right visual field, thus providing a single response variable encoding both the location, size, and timing of all the stimulus presentations. This stimulus kernel was the convolved with the temporal response kernel of our calcium indicator (NL-GCaMP6s).

PLS regression identifies the normalized dimensions and that maximize the covariance between paired observations and , respectively. In our case, the visual stimulus is represented by a single variable , simplifying the problem to identifying the subspace of neural activity that optimally preserves information about the visual stimulus (sometimes referred to as PLS1 regression). That is, the N x T neural time series matrix X is reduced to a d x T matrix spanned by a set of orthonormal vectors. PLS1 regression is performed as follows:

PLS1 algorithm

Let X_i = X and . For i = 1…d,

(1) 

(2) 

(3) 

(4) 

(5)  (note this is scalar)

(6) 

The projections of the neural data {p_i} thus span a subspace that maximally preserves information about the visual stimulus . Stacking these projections into the N x d matrix P that represents the transform from the whole-brain neural state space to the visually-evoked subspace, the optimal decoding direction is given by the linear least squares solution . The dimensionality d of PLS regression was optimized using 6-fold cross-validation with 3 repeats and choosing the dimensionality between d = 1 and 20 with the lowest cross-validated mean squared error for each larva. Then, was computed using all time points.

For each stimulus type, the noise covariance matrix  was computed in the low-dimensional PLS space, given that direct estimation of the noise covariances across many thousands of neurons would likely be unreliable. A noise covariance matrix was calculated separately for each stimulus, and then averaged across all stimuli. As before, the mean activity µ_i for each neuron  was computed over each stimulus presentation period. The noise covariance then describes the correlated fluctuations δ_i around this mean response for each pair of neurons i and j, where

The noise modes for α = 1 …d were subsequently identified by eigendecomposition of the mean noise covariance matrix across all stimuli, . The angle between the optimal stimulus decoding direction and the noise modes is thus given by .”

- No response: It is not clear from the methods description if cases where the animal has no tail response are being lumped with cases where the animal decides to swim forward and thus has a large absolute but small mean tail curvature. These should be treated separately. 

We thank the reviewer for raising the potential for this confusion and agree that forward-motion trials should not treated the same as motionless trials. While these types of trial were indeed treated separately in our original manuscript, we have updated the Methods section of our revised manuscript to make this clear:

“Left and right turn trials were extracted as described previously. Response trials included both left and right turn trials (i.e., the absolute value of mean tail curvature > σ_active), whereas nonresponse trials were motionless (absolute mean tail curvature < σ_active). In particular, forward-motion trials were excluded from these analyses.”

While our study has focused specifically on left and right turns, we hypothesize that the responsiveness bias ensemble may also be involved in forward movements and look forward to future work exploring the relationship between whole-brain dynamics and the full range of motor outputs.

- Behavioral variability: Related to Figure 2, within- and across-subject variability are confounded. Please disambiguate. It may also be informative on a per-fish basis to examine associations between reaction time and body movement.

The reviewer is correct that our previously reported summary statistics in Figure 2D-F were aggregated across trials from multiple larvae. Following the reviewer’s suggestion to make the magnitudes of across-larvae and within-larva variability clear, in our revised manuscript we have added two additional figure panels to Figure S2.

New Figure S2A highlights the across-larvae variability in mean head-directed behavioral responses to stimuli of various sizes. Overall, the relationship between stimulus size and the mean tail curvature across trials is largely consistent across larvae; however, the crossing-over point between leftward (positive curvature) and rightward (negative curvature) turns for a given side of the visual field exhibits some variability across larvae.

New Figure S2B shows examples of within-larva variability by plotting the mean tail curvature during single trials for two example larvae. Consistent with Figure 2G which also demonstrates within-larva variability, responses to a given stimulus are variable across trials in both examples. However, this degree of within-larva variability can appear different across larvae. For example, the larva shown on the left of Figure S2B exhibits greater overlap between responses to stimuli presented on opposite visual fields, whereas the larva shown on the right exhibits greater distinction between responses.

- Data presentation clarity: All figure panels need scale bars - for example, in Figure 3A there is no indication of timescale (or time of stimulus presentation). Figure 3I should also show the time series of the w_opt projection.

We appreciate the reviewer’s attention to detail in this regard. We have added scalebars to Figures 3A, 3H-I, S4B(ii), 4H, 4J in the revised manuscript, and all new figure panels where relevant. In addition, the caption of Figure 3A has been updated to include a description of the time period plotted relative to the onset of the visual stimulus.

Additionally, we appreciate the reviewer’s idea to show w_opt in Figure 3J of the revised manuscript (previously Figure 3I). This clearly shows that the visual decoding project is inactive during the short baseline period before visual stimulation begins, whereas the noise mode is correlated with motor output throughout the recording.

- Pixel locations: Given the poor quality of the brain images, it is difficult to tell the location of highlighted pixels relative to brain anatomy. In addition, given that the midbrain consists of much more than the tectum, it is not appropriate to put all highlighted pixels from the midbrain under the category of tectum. To aid in data interpretation and better connect this work with the literature, it is recommended that the authors register their data sets to standard brain atlases and determine if there is any clustering of relevant pixels in regions previously associated with prey-capture or predator-avoidance behavior.

We agree with the reviewer that registration of our datasets to a standard brain atlas is a highly useful addition. While the dense, pan-neuronal labeling makes the isolation of highly specific circuit components difficult, we have shown in more detail the specific brain regions contributing to these populations by aligning our recordings to the Z-Brain atlas (Randlett et al., 2015) as shown in new Figures S1E, S3F-G, 4I, 4K, and S5F-G. In addition, we provided a more detailed parcellation of the neuronal ensembles by providing projections of the full 3D volume along the xy and yz axes, in addition to the unregistered xy projection shown in new Figures 4H and 4J. We also found that the distribution of neurons in our huc:H2B-GCaMP6s recordings is very similar to the distribution of labeling in the huc:H2B-RFP reference image from the Z-Brain atlas (new Figure S1E), which further supports our whole-brain imaging results.

Overall, we find that this more detailed quantification and visualization is consistent with the interpretations in the previous version of our manuscript. In particular, we show that optimal visual decoding population (w_opt) and largest noise mode (e1) are localized to the midbrain (new Figures S3F-G), which is expected since in Figure 3 we first extracted a low-dimensional subspace of whole-brain neural activity that optimally preserved visual information. Additionally, we provide additional evidence that the populations correlated with the turn bias and responsiveness bias are distributed throughout the brain, including a relatively dense localization to the cerebellum, telencephalon, and dorsal diencephalon (habenula, new Figures 4H-K and S5F-G).

Finally, the reviewer is correct that our original label of “tectum” was a misnomer; the region analyzed corresponded to the midbrain, including the tegmentum, torus longitudinalis, and torus semicicularis in addition to the tectum. We have updated the brain regions shown and labels throughout the manuscript.

Interpretation

- W_opt and e_1 orthogonality: The statement that these two vectors, determined from analysis of the fluorescence data, are orthogonal, actually brings into question the idea that true signal and leading noise vectors in firing-rate state-space are orthogonal. First, the current analysis is confounding signals across different time periods - one could assume linearity all the way through the transformations, but this would only work if earlier sources of activation were being accounted for. Second, the transformation between firing rate and fluorescence is most likely not linear for GCaMP6s in most of the cells recorded. Thus, one would expect a change in the relationship between these vectors as one maps from fluorescence to firing rate.

Unfortunately, we are not entirely sure we have understood the reviewer’s argument. We are assuming that the reviewer’s first sentence is suggesting that the observation of orthogonality in the neural state space measured in calcium imaging precludes the possibility (“actually brings into question”, as the reviewer states) that the same neural ensembles could be orthogonal in firing rate state space measured by electrophysiological data. If this is the reviewer’s conjecture, we respectfully disagree with it. Consider a toy example of a neural network containing N ensembles of neurons, where the neurons within an ensemble all fire simultaneously, and two populations never fire at the same time. As long as the “switching” of firing between ensembles is not fast relative to the resolution of the GCaMP kernel, the largest principal components would represent orthogonal dimensions differentiating the various ensembles, both when observing firing rates or observing timeseries convolved by the GCaMP kernel. This is a simple example where the observed orthogonality would appear similar in both calcium imaging and electrophysical data, demonstrating that we should not allow conclusions from fluorescence data to “bring into question” that the same result could be observed in firing rate data.

We also disagree with the reviewer’s argument that we are “confounding signals across time periods”. Indeed, we must interpret the data in light of the GCaMP response kernel. However, all of the analyses presented here are performed on instantaneous measurements of population activity patterns. These activity patterns do represent a smoothed, likely nonlinear integration of recent neuronal activity, but unless the variability in the GCaMP response kernel (discussed above) is widely different across these populations (which has not been observed in the literature), we do not expect that the GCaMP transformations would artificially induce orthogonality in our analysis approach. Such smoothing operations tend to instead increase correlations across neurons and population decoding approaches generally benefit from this smoothness, as we have argued above. However, a much more problematic situation would be if we were comparing the activity of two neuronal populations at different points in time (which we do not include in this study), in which case the nonlinearities could overaccentuate orthogonality between non-time-matched activity patterns.

Finally, we agree with the reviewer that the transformation between firing rate and fluorescence is very likely nonlinear and that these vectors of population activity do not perfectly represent what would be observed if one had access to whole-brain, cellular-resolution electrophysiology spike data. However, similar observations regarding the brain-wide, distributed encoding of behavior have been confirmed across recording modalities in the mouse (Stringer et al., 2019; Steinmetz et al., 2019), where large-scale electrophysiology utilizing highly invasive probes (e.g., Neuropixels) is more feasible than in the larval zebrafish. With the advent of whole-brain voltage imaging in the larval zebrafish, we expect any differences between calcium and voltage dynamics will be better understood, yet such techniques will likely continue to suffer to some extent from the nonlinearities described here.

- Sources of variability: The authors do not take into account a fairly obvious source of variability in trial-to-trial response - eye position. We know that prey capture responsiveness is dependent on eye position during stimulus (see Figure 4 of PMID: 22203793). We also expect that neurons fairly early in the visual pathway with relatively narrow receptive fields will show variable responses to visual stimuli as the degree of overlap with the receptive field varies with eye movement. There can also be small eye-tracking movements ahead of the decision to engage in prey capture (Figure 1D, PMID: 31591961) that can serve as a drive to initiate movements in a particular direction. Given these possibilities indicating that the behavioral measure of interest is gaze, and the fact that eye movements were apparently monitored, it is surprising that the authors did not include eye movements in the analysis and interpretation of their data.

We agree with the reviewer that eye movements, such as saccades and convergence, are important motor outputs that are well-known to play a role in the sequence of motor actions during prey capture and other behaviors. Therefore, we have added the following new eye tracking results to our revised manuscript:

“In order to confirm that the observed neural variability in the visually-evoked populations was not predominantly due to eye movements, such as saccades or convergence, we tracked the angle of each eye. We utilized DeepLabCut, a deep learning tool for animal pose estimation (Mathis et al., 2018), to track keypoints on the eye which are visible in the raw fLFM images, including the retina and pigmentation (Figure S3D(i)). This approach enabled identification of various eye movements, such as convergence and the optokinetic reflex (Figure S3D(ii-iii)). Next, we extracted a number of various eye states, including those based on position (more leftward vs. rightward angles) and speed (high angular velocity vs. low or no motion). Figure S3E(i) provides example stimulus response profiles across trials of the same visual stimulus in each of these eye states, similar to a single column of traces in Figure 3A broken out into more detail. These data demonstrate that the magnitude and temporal dynamics of the stimulus-evoked responses show apparently similar levels of variability across eye states. If neural variability was driven by eye movement during the stimulus presentation, for example, one would expect to see much more variability during the high angular velocity trials than low, which is not apparent. Next, we asked whether the dominant neural noise modes vary across eye states, which would suggest that the geometry of neuronal variability is influenced by eye movements or states. To do so, the dominant noise modes were estimated in each of the individual eye conditions, as well as bootstrapped trials from across all eye conditions. The similarity of these noise modes estimated from different eye conditions (Figure S3E(ii), right)) was not significantly different from the similarity of noise modes estimated from bootstrapped random samples across all eye conditions (Figure S3E(ii), left)). Therefore, while movements of the eye likely contribute to aspects of the observed neural variability, they do not dominate the observed neural variability here, particularly given our observation that the largest noise mode represents a considerable fraction of the observed neural variance (Figure 3E).”

While these results provide an important control in our study, we anticipate further study of the relationship between eye movements or states, visually-evoked neural activity, and neural noise modes would identify the additional neural ensembles which are correlated with and drive this additional motor output.

Reviewer #3 (Public Review):

Summary:

In this study, Manley and Vaziri designed and built a Fourier light-field microscope (fLFM) inspired by previous implementations but improved and exclusively from commercially available components so others can more easily reproduce the design. They combined this with the design of novel algorithms to efficiently extract whole-brain activity from larval zebrafish brains.

This new microscope was applied to the question of the origin of behavioral variability. In an assay in which larval zebrafish are exposed to visual dots of various sizes, the fish respond by turning left or right or not responding at all. Neural activity was decomposed into an activity that encodes the stimulus reliably across trials, a 'noise' mode that varies across trials, and a mode that predicts tail movements. A series of analyses showed that trial-to-trial variability was largely orthogonal to activity patterns that encoded the stimulus and that these noise modes were related to the larvae's behavior.

To identify the origins of behavioral variability, classifiers were fit to the neural data to predict whether the larvae turned left or right or did not respond. A set of neurons that were highly distributed across the brain could be used to classify and predict behavior. These neurons could also predict spontaneous behavior that was not induced by stimuli above chance levels. The work concludes with findings on the distributed nature of single-trial decision-making and behavioral variability.

Strengths:

The design of the new fLFM microscope is a significant advance in light-field and computational microscopy, and the open-source design and software are promising to bring this technology into the hands of many neuroscientists.

The study addresses a series of important questions in systems neuroscience related to sensory coding, trial-to-trial variability in sensory responses, and trial-to-trial variability in behavior. The study combines microscopy, behavior, dynamics, and analysis and produces a well-integrated analysis of brain dynamics for visual processing and behavior. The analyses are generally thoughtful and of high quality. This study also produces many follow-up questions and opportunities, such as using the methods to look at individual brain regions more carefully, applying multiple stimuli, investigating finer tail movements and how these are encoded in the brain, and the connectivity that gives rise to the observed activity. Answering questions about variability in neural activity in the entire brain and its relationship to behavior is important to neuroscience and this study has done that to an interesting and rigorous degree.

Points of improvement and weaknesses:

The results on noise modes may be a bit less surprising than they are portrayed. The orthogonality between neural activity patterns encoding the sensory stimulus and the noise modes should be interpreted within the confounds of orthogonality in high-dimensional spaces. In higher dimensional spaces, it becomes more likely that two random vectors are almost orthogonal. Since the neural activity measurements performed in this study are quite high dimensional, a more explicit discussion is warranted about the small chance that the modes are not almost orthogonal.

We agree with the reviewer that orthogonality is less “surprising” in high-dimensional spaces, and we have added this important point of interpretation to our revised manuscript. Still, it is important to remember that while the full neural state space is very high-dimensional (we record that activity of up to tens of thousands of neurons simultaneously), our analyses regarding the relationship between the trial-to-trial noise modes and decoding dimensions were performed in a low-dimensional subspace (up to 20 dimensions) identified by PLS regression to that optimally preserved visual information. This is a key step in our analysis which serves two purposes: 1. it removes some of the confound described the reviewer regarding the dimensionality of the neural state space analyzed; and 2. it ensures that the noise modes we analyze are even relevant to sensorimotor processing. It would certainly not be surprising or interesting if we identified a neural dimension outside the midbrain which was orthogonal to the optimal visual decoding dimension. 

Regardless, in order to better control for this confound, we estimated the distribution of angles between random vectors in this subspace. As we describe in the revised manuscript:

“However, in high-dimensional spaces, it becomes increasingly common that two random vectors could appear orthogonal. While this is particularly a concern when analyzing a neural state space spanned by tens of thousands of neurons, our application of PLS regression to identify a low-dimensional subspace of relevant neuronal activity partially mitigates this concern. In order to control for this confound, we compared the angles between w_opt and e1 across larvae to that computed with shuffled versions of w_opt,shuff estimated by randomly shuffling the stimulus labels before identifying the optimal decoding direction. While it is possible to observe shuffled vectors which are nearly orthogonal to e₁, the shuffled distribution spans a significantly greater range of angles than the observed data, demonstrating that this orthogonality is not simply a consequence of analyzing multi-dimensional activity patterns.”

The conclusion that sparsely distributed sets of neurons produce behavioral variability needs more investigation because the way the results are shown could lead to some misinterpretations. The prediction of behavior from classifiers applied to neural activity is interesting, but the results are insufficiently presented for two reasons.

(1) The neurons that contribute to the classifiers (Figures 4H and J) form a sufficient set of neurons that predict behavior, but this does not mean that neurons outside of that set cannot be used to predict behavior. Lasso regularization was used to create the classifiers and this induces sparsity. This means that if many neurons predict behavior but they do so similarly, the classifier may select only a few of them. This is not a problem in itself but it means that the distributions of neurons across the brain (Figures 4H and J) may appear sparser and more distributed than the full set of neurons that contribute to producing the behavior. This ought to be discussed better to avoid misinterpretation of the brain distribution results, and an alternative analysis that avoids the confound could help clarify.

We thank the reviewer for raising this point, which we agree should be discussed in the manuscript. Lasso regularization was a key ingredient in our analysis; l2 regularization alone was not sufficient to prevent overfitting to the training trials, particularly when decoding turn direction and responsiveness. Previous studies have also found that sparse subsets of neurons better predict behavior than single neuron or non-sparse populations, for example Scholz et al. (2018).

While showing l2 regularization would not be a fair comparison given the poor performance of the l2-regularized classifiers, we opted to identify a potentially “fuller” set of neurons correlated with these biases based on the correlation between each neuron’s activity over the recording and the projection along the turn direction or responsiveness dimension identified using l1 regularization. This procedure has the potential to identify all neurons correlated with the final ensemble dynamics, rather than just a “sufficient set” for lasso regression. In new Figures S5F-G, we show the 3D distribution of all neurons significantly correlated with these biases, which appear similar to those in Figures 4H-K and widely distributed across practically the entire labeled area of the brain.

(2) The distribution of neurons is shown in an overly coarse manner in only a flattened brain seen from the top, and the brain is divided into four coarse regions (telencephalon, tectum, cerebellum, hindbrain). This makes it difficult to assess where the neurons are and whether those four coarse divisions are representative or whether the neurons are in other non-labeled deeper regions. For these two reasons, some of the statements about the distribution of neurons across the brain would benefit from a more thorough investigation.

We agree with the reviewer that a more thorough description and visualization of these distributed populations is warranted.

While the dense, pan-neuronal labeling makes the isolation of highly specific circuit components difficult, we have shown in more detail the specific brain regions contributing to these populations by aligning our recordings to the Z-Brain atlas (Randlett et al., 2015) as shown in new Figures S1E, S3F-G, 4I, 4K, and S5F-G. In addition, we provided a more detailed parcellation of the neuronal ensembles by providing projections of the full 3D volume along the xy and yz axes, in addition to the unregistered xy projection shown in new Figures 4H and 4J. We also found that the distribution of neurons in our huc:H2B-GCaMP6s recordings is very similar to the distribution of labeling in the huc:H2B-RFP reference image from the Z-Brain atlas (new Figure S1E), which further supports our whole-brain imaging results.

Overall, we find that this more detailed quantification and visualization is consistent with the interpretations in the previous version of our manuscript. In particular, we show that optimal visual decoding population (w_opt) and largest noise mode (e1) are localized to the midbrain (new Figures S3F-G), which is expected since in Figure 3 we first extracted a low-dimensional subspace of whole-brain neural activity that optimally preserved visual information. Additionally, we provide additional evidence that the populations correlated with the turn bias and responsiveness bias are distributed throughout the brain, including a relatively dense localization to the cerebellum, telencephalon, and dorsal diencephalon (habenula, new Figures 4H-K and S5F-G).

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

In addition to the overall strengths and weaknesses above, I have a few specific comments that I think could improve the study:

(1) In lines 334-335 you write that 'We proceeded to build various logistic regression classifiers to decode'. Do you mean you tested this with other classifier types as well (e.g. SVM, Naive Bayes) or do you mean various because you trained the classifier described in the methods on each animal? This is not clear. If it is the first, more information is needed about what other classifiers you used.

We appreciate the reviewer raising this point of clarification. Here, we simply meant that we fit the multiclass logistic regression classifier in the one-vs-rest scheme. In this sense, a single multiclass logistic regression classifier was fit for each larva. We have updated our revised manuscript with this clarification: “The visual stimuli were decoded using a one-versus-rest, multiclass logistic regression classifier with lasso regularization.”

(2) In Figure 3 you train the decoder on all visually responsive cells identified across the brain. Does this reliability of stimulus decoding also hold for neurons sampled from specific brain regions? For example, does this reliable decoding come from stronger and more reliable responses in the optic tectum, whereas stimulus decodability is not as good in visual encoding neurons identified in other structures?

In new Figure S5B, we show the performance of stimulus decoding from various brain regions. We find that stimulus classification is possible from the midbrain and cerebellum, very poor from the hindbrain, and not possible from the telencephalon during the period between stimulus onset and the decision.

(3) In relation to point 2, it would be good to show in which brain areas the visually responsive neurons are located, and maybe the average coefficients per brain area. Plots like Figures 3G, and H would benefit from a quantification into areas. Similarly, a parcellation into more specific brain areas in Figure 4 would also be valuable.

In addition to providing a more detailed parcellation of the turn direction and responsiveness bias populations in Figure 4, we have provided a similar visualization and quantification of the optimal stimulus decoding population and the dominant noise mode in new Figures S3F-G, respectively.

(4) In Figure 3f, it is not clear to me how this shows that w_opt and e1 are orthogonal. They appear correlated.

The orthogonality we quantify is related to the pattern of coefficients across neurons, not necessarily the timeseries of their projections. The slight shift in the noise mode activations as you move from stimuli on the left visual field to the right actually comes from the motor outputs. Large left stimuli tend to evoke a rightward turn and vice versa, and the example noise mode shown encodes the directionality and vigor of tail movements, resulting in the slight shifts observed.

(5) I think the wording of this conclusion is too strong for the results and a bit illogical:

'Thus, our data suggest that the neural dynamics underlying single-trial action selection are the result of a widely-distributed circuit that contains subpopulations encoding internal time-varying biases related to both the larva's responsiveness and turn direction, yet distinct from the sensory encoding circuitry.'

If that is the case, how is it even possible that the larvae can do a visually guided behaviour?

Especially given Suppl Fig 4C it would be more appropriate to say something along the lines of: 'When stimuli are highly ambiguous, single trial action selection is dominated by widely-distributed circuit that contains subpopulations encoding internal time-varying biases related to both the larva's responsiveness and turn direction, that encode choice distinctly from the sensory encoding circuitry'.

We appreciate the reviewer’s suggestion and have re-worded this line in the discussion in order to clarify that these time-varying biases are predominant in the case of ambiguous stimuli, as shown in Figure S5C in our revised manuscript (corresponding to Figure S4C in our original submission).

(6) Line 599: typo: trial-to-trail

We thank the reviewer for noting this error, which has been corrected in the revised text of the manuscript.

Author response:

The following is the authors’ response to the original reviews

Reviewer #1 (Public review):

Weaknesses:  

(1) The heatmaps (for example, Figure 3A, B) are challenging to read and interpret due to their size. Is there a way to alter the visualization to improve interpretability? Perhaps coloring the heatmap by general anatomical region could help? We feel that these heatmaps are critical to the utility of the registration strategy, and hence, clear visualization is necessary. 

We thank the reviewers for this point on aesthetic improvement, and we agree that clearer visualization of our correlation heatmaps is important. To address this point, we have incorporated the capability of grouping “child” subregions in anatomical order by their more general “parent” region into the package function, plot_correlation_heatmaps(). Parent regions will be can now be plotted as smaller sub-facets in the heatmaps. We have also rearranged our figures to fit enlarged heatmaps in Figures 3-5, and Supplementary Figure 10 for easier visualization. 

(2) Additional context in the Introduction on the use of immediate early genes to label ensembles of neurons that are specifically activated during the various behavioral manipulations would enable the manuscript and methodology to be better appreciated by a broad audience. 

We thank the reviewers for this suggestion and have revised the first part of our Introduction to reflect the broader use and appeal of immediate early genes (IEGs) for studying neural changes underlying behavior.

(3) The authors mention that their segmentation strategies are optimized for the particular staining pattern exhibited by each reporter and demonstrate that the manually annotated cell counts match the automated analysis. They mention that alternative strategies are compatible, but don't show this data. 

We thank the reviewers for this comment. We also appreciate that integration with alternative strategies is a major point of interest to readers, given that others may be interested in compatibility with our analysis and software package, rather than completely revising their own pre-existing pipelines. 

Generally, we have validated the ability to import datasets generated from completely different workflows for segmentation and registration. We have since released documentation on our package website with step-by-step instructions on how to do so (https://mjin1812.github.io/SMARTTR/articles/Part5.ImportingExternalDatasets). We believe this tutorial is a major entry point to taking advantage of our analysis package, without adopting our entire workflow.

This specific point on segmentation refers to the import_segmentation_custom()function in the package. As there is currently not a standard cell segmentation export format adopted by the field, this function still requires some data wrangling into an import format saved as a .txt file. However, we chose not to visually demonstrate this capability in the paper for a few reasons.  

i) A figure showing the broad testing of many different segmentation algorithms, (e.g., Cellpose, Vaa3d, Trainable Weka Segmentation) would better demonstrate the efficacy of segmentation of these alternative approaches, which have already been well-documented. However, demonstrating importation compatibility is more of a demonstration of API interface, which is better shown in website documentation and tutorial notebooks.

ii) Additionally, showing importation with one well-established segmentation approach is still a demonstration of a single use case. There would be a major burden-of-proof in establishing importation compatibility with all potential alternative platforms, their specific export formats, which may be slightly different depending on post-processing choices, and the needs of the experimenters (e.g., exporting one versus many channels, having different naming conventions, having different export formats). For example, output from Cellpose can take the form of a NumPy file (_seg.npy file), a .png, or Native ImageJ ROI archive output, and users can have chosen up to four channels. Until the field adopts a standardized file format, one flexible enough to account for all the variables of experimental interest, we currently believe it is more efficient to advise external groups on how to transform their specific data to be compatible with our generic import function.  

(4) The authors provided highly detailed information for their segmentation strategy, but the same level of detail was not provided for the registration algorithms. Additional details would help users achieve optimal alignment.

We apologize for this lack of detail. The registration strategy depends upon the WholeBrain (Fürth et al., 2018) package for registration to the Allen Mouse Common Coordinate Framework. While this strategy has been published and documented elsewhere, we have substantially revised our methods section on the registration process to better incorporate details of this approach.

(5) The authors illustrate registration to the Allen atlas. Can they comment on whether the algorithm is compatible with other atlases or with alternative sectioning planes (horizontal/sagittal)? 

Since the current registration workflow integrates WholeBrain (Fürth et al., 2018), any limitations of WholeBrain apply to our approach, which means limited support for registering non-coronal sectioning planes and reliance on the Allen Mouse Atlas (Dong, 2008). However, network analysis and plotting functions are currently compatible with the Allen Mouse Brain Atlas and the Kim Unified Mouse Brain Atlas version (2019) (Chon et al., 2019). Therefore, current limitations in registration do not preclude the usefulness of the SMARTTR software in generating valuable insights from network analysis of externally imported datasets. 

There are a number of alternative workflows, such as the QUINT workflow (Yates et al., 2019), that support multiple different mouse atlases, and registration of arbitrarily sectioned angles. We have plans to support and a facilitate an entry point for this workflow in a future iteration of SMARTTR, but believe it is of benefit to the wider community to release and support SMARTTR in its current state.

(6) Supplemental Figures S10-13 do not have a legend panel to define the bar graphs. 

We apologize for this omission and have fixed our legends in our resubmission. Our supplement figure orders have changed and the corresponding figures are now Supplemental Figures S11-14.

(7) When images in a z-stack were collapsed, was this a max intensity projection or average? Assuming this question is in regards to our manual cell counting validation approach, the zstacks were collapsed as a maximum intensity projection.  

Reviewer #2 (Public review): 

Weaknesses: 

(1) While I was able to install the SMARTR package, after trying for the better part of one hour, I could not install the "mjin1812/wholebrain" R package as instructed in OSF. I also could not find a function to load an example dataset to easily test SMARTR. So, unfortunately, I was unable to test out any of the packages for myself. Along with the currently broken "tractatus/wholebrain" package, this is a good example of why I would strongly encourage the authors to publish SMARTR on either Bioconductor or CRAN in the future. The high standards set by Bioc/CRAN will ensure that SMARTR is able to be easily installed and used across major operating systems for the long term. 

We greatly thank the reviewer for pointing out this weakness; long-term maintenance of this package is certainly a mutual goal. Loading an .RDATA file is accomplished by either doubleclicking directly on the file in a directory window, after specifying this file type should be opened in RStudio or by using the load() function, (e.g., load("directory/example.RData")). We have now explicitly outlined these directions in the online documentation. 

Moreover, we have recently submitted our package to CRAN and are currently working on revisions following comments. This has required a package rebranding to “SMARTTR”, as there were naming conflicts with a previously archived repository on CRAN. Currently, SMARTTR is not dependent on the WholeBrain package, which remains optional for the registration portion of our workflow. Ultimately, this independence will allow us to maintain the analysis and visualization portion of the package independently.

In the meantime, we have fully revised our installation instructions (https://mjin1812.github.io/SMARTTR/articles/SMARTTR). SMARTTR is now downloadable from a CRAN-like repository as a bundled .tar.gz file, which should ease the burden of installation significantly. Installation has been verified on a number of different versions of R on different platforms. Again, we hope these changes are sufficient and improve the process of installation. 

(2) The package is quite large (several thousand lines include comments and space). While impressive, this does inherently make the package more difficult to maintain - and the authors currently have not included any unit tests. The authors should add unit tests to cover a large percentage of the package to ensure code stability. 

We have added unit testing to improve the reliability of our package. Unit tests now cover over 71% of our source code base and are available for evaluation on our github website (https://github.com/mjin1812/SMARTTR). We focused on coverage of the most front-facing functions. We appreciate this feedback, which has ultimately enhanced the longevity of our software.

(3) Why do the authors choose to perform image segmentation outside of the SMARTTR package using ImageJ macros? Leading segmentation algorithms such as CellPose and StarMap have well-documented APIs that would be easy to wrap in R. They would likely be faster as well. As noted in the discussion, making SMARTTR a one-stop shop for multi-ensemble analyses would be more appealing to a user. 

We appreciate this feedback. We believe parts of our response to Reviewer 1, Comment 3, are relevant to this point. Interfaces for CellPose and ClusterMap (which processes in situ transcriptomic approaches, like STARmap) are both in python, and currently there are ways to call python from within R (https://rstudio.github.io/reticulate/index.html). We will certainly explore incorporating these APIs from R. However, we would anticipate this capability is more similar to “translation” between programming languages, but would not currently preclude users from the issue of needing some familiarity with the capabilities of these python packages, and thus with python syntax.

(4) Given the small number of observations for correlation analyses (n=6 per group), Pearson correlations would be highly susceptible to outliers. The authors chose to deal with potential outliers by dropping any subject per region that was> 2 SDs from the group mean. Another way to get at this would be using Spearman correlation. How do these analyses change if you use Spearman correlation instead of Pearson? It would be a valuable addition for the author to include Spearman correlations as an option in SMARTTR. 

We thank reviewers for this suggestion and we have updated our code base to include the possibility for using Spearman’s correlation coefficient as opposed to Pearson’s correlation coefficient for heatmaps in the get_correlations() function. Users can now use the `method` parameter, set to either “pearson” or “spearman” and results will propagate throughout the rest of the analysis using these results.

Below, in Author response image 1 we show a visual comparison of the correlation heat maps for active eYFP⁺ ensembles in the CT and IS groups using both Pearson and Spearman correlations. We see a strongly qualitative similarity between the heat maps. Of course, since the statistical assumptions underlying the relationship between variables using Pearson correlation (linear) vs Spearman correlation (monotonic) are different, users should take this into account when interpreting results using different approaches.

Author response image 1.

Pearson and Spearmen regional correlations of eYFP+ ensembles activity in the CT and IS groups.

(5) I see the authors have incorporated the ability to adjust p-values in many of the analysis functions (and recommend the BH procedure) but did not use adjusted p-values for any of the analyses in the manuscript. Why is this? This is particularly relevant for the differential correlation analyses between groups (Figures 3P and 4P). Based on the un-adjusted pvalues, I assume few if any data points will still be significant after adjusting. While it's logical to highlight the regional correlations that strongly change between groups, the authors should caution which correlations are "significant" without adjusting for multiple comparisons. As this package now makes this analysis easily usable for all researchers, the authors should also provide better explanations for when and why to use adjusted p-values in the online documentation for new users. 

We appreciate the feedback note that our dataset is presented as a more demonstrative and exploratory resource for readers and, as such, we accept a high tolerance for false positives, while decreasing risk of missing possible interesting findings. As noted by Reviewer #2, it is still “logical to highlight the regional correlations that strongly change between groups.” We have clarified in our methods that we chose to present uncorrected p-values when speaking of significance. 

We have also removed any previous recommendations for preferred methods for multiple comparisons adjustment in our function documentations, as some previous documentation was outdated. Moreover, the standard multiple comparisons adjustment approaches assume complete independence between tests, whereas this assumption is violated in our differential correlational analysis (i.e., a region with one significantly altered connection is more likely than another to have another significantly altered connection).

Ultimately, the decision to correct for multiple comparisons with standard FDR, and choice of significance threshold, should still be informed by standard statistical theory and user-defined tolerance for inclusion of false-positives and missing of false-negatives. This will be influenced by factors, such as the nature and purpose of the study, and quality of the dataset.  

(6) The package was developed in R3.6.3. This is several years and one major version behind the current R version (4.4.3). Have the authors tested if this package runs on modern R versions? If not, this could be a significant hurdle for potential users. 

We thank reviewers for pointing out concerns regarding versioning. We have since updated our installation approach for SMARTTR, which is compatible with versions of R >= 3.6 and has been tested on Mac ARM-based (Apple silicon) architecture (R v4.4.2), and Windows 10 (R v3.6.3, v4.5.0 [devel]). 

The recommendation for users to install R 3.6.3 is primarily for those interested in using our full workflow, which requires installation of the WholeBrain package, which is currently a suggested package. We anticipate updating and supporting the visualization and network analysis capabilities, whilst maintaining previous versioning for the full workflow presented in this paper.  

(7) In the methods section: "Networks were constructed using igraph and tidygraph packages." - As this is a core functionality of the package, it would be informative to specify the exact package versions, functions, and parameters for network construction. 

We thank reviewers for pointing out the necessity for these details for code reproducibility. We have since clarified our language in the manuscript on the exact functions we use in our analysis and package versions, which we also fully document in our online tutorial. Additionally. We have printed our package development and analysis environment online at https://mjin1812.github.io/SMARTTR/articles/Part7.Development.

(8) On page 11, "Next, we examined the cross-correlations in IEG expression across brain regions, as strong co-activation or opposing activation can signify functional connectivity between two regions" - cross-correlation is a specific analysis in signal processing. To avoid confusion, the authors should simply change this to "correlations". 

We thank the reviewer for pointing out this potentially confusing phrasing. We have changed all instances of “cross-correlation” to “correlation”.

(9) Panels Q-V are missing in Figure 5 caption. 

We thank the reviewer for pointing out this oversight. We have now fixed this in our revision.

References

Chon, U., Vanselow, D. J., Cheng, K. C., & Kim, Y. (2019). Enhanced and unified anatomical labeling for a common mouse brain atlas. Nature Communications, 10(1), 5067. https://doi.org/10.1038/s41467-019-13057-w

Dong, H. W. (2008). The Allen reference atlas: A digital color brain atlas of the C57Bl/6J male mouse (pp. ix, 366). John Wiley & Sons Inc.

Fürth, D., Vaissière, T., Tzortzi, O., Xuan, Y., Märtin, A., Lazaridis, I., Spigolon, G., Fisone, G., Tomer, R., Deisseroth, K., Carlén, M., Miller, C. A., Rumbaugh, G., & Meletis, K. (2018). An interactive framework for whole-brain maps at cellular resolution. Nature Neuroscience, 21(1), 139–149. https://doi.org/10.1038/s41593-017-0027-7

Yates, S. C., Groeneboom, N. E., Coello, C., Lichtenthaler, S. F., Kuhn, P.-H., Demuth, H.-U., Hartlage-Rübsamen, M., Roßner, S., Leergaard, T., Kreshuk, A., Puchades, M. A., & Bjaalie, J. G. (2019). QUINT: Workflow for Quantification and Spatial Analysis of Features in Histological Images From Rodent Brain. Frontiers in Neuroinformatics, 13. https://www.frontiersin.org/articles/10.3389/fninf.2019.00075

Author response:

Public Reviews:

Reviewer #1 (Public review):  

Summary:  

This study provides new insights into the role of miR-19b, an oncogenic microRNA, in the developing chicken pallium. Dynamic expression pattern of miR-19b is associated with its role in regulating cell cycle progression in neural progenitor cells. Furthermore, miR-19b is involved in determining neuronal subtypes by regulating Fezf2 expression during pallial development. These findings suggest an important role for miR-19b in the coordinated spatio-temporal regulation of neural progenitor cell dynamics and its evolutionary conservation across vertebrate species.  

Strengths:  

The authors identified conserved roles of miR-19 in the regulation of neural progenitor maintenance between mouse and chick, and the latter is mediated by the repression of E2f8 and NeuroD1. Furthermore, the authors found that miR-19b-dependent cell cycle regulation is tightly associated with specification of Fezf1 or Mef2c-positive neurons, in spatio-temporal manners during chicken pallial development. These findings uncovered molecular mechanisms underlying microRNA-mediated neurogenic controls.  

Weaknesses:  

Although the authors in this study claimed striking similarities of miR-19a/b in neurogenesis between mouse and chick pallium, a previous study by Bian et al. revealed that miR-19a contributes the expansion of radial glial cells by suppressing PTEN expression in the developing mouse neocortex, while miR-19b maintains apical progenitors via inhibiting E2f2 and NeuroD1 in chicken pallium. Thus, it is still unclear whether the orthologous microRNAs regulate common or species-specific target genes.  

In this study, we have proposed that miR-19b regulates similar phenomena in both species using different targets, such as regulation of proliferation through PTEN in mouse and through E2f8 in the chicken.

The spatiotemporal expression patterns of miR-19b and several genes are not convincing. For example, the authors claim that NeuroD1 is initially expressed uniformly in the subventricular zone (SVZ) but disappears in the DVR region by HH29 and becomes detectable by HH35 (Figure 1). However, the in situ hybridization data revealed that NeuroD1 is highly expressed in the SVZ of the DVR at HH29 (Figure 4F). Thus, perhaps due to the problem of immunohistochemistry, the authors have not been able to detect NeuroD1 expression in Figure 1D, and the interpretation of the data may require significant modification.  

While Fig. 1B may suggest that NeuroD1 expression has disappeared from the DVR region by HH29, this is not true in general because we have observed NeuroD1 to be expressed in the DVR at HH29 in images of other sections. In the revised version, we will include improved images for panels of Fig. 1B which accurately show the expression pattern of NeuroD1 and miR19b at stages HH29 and HH35.  

It seems that miR-19b is also expressed in neurons (Figure 1), suggesting the role of miR19-b must be different in progenitors and differentiated neurons. The data on the gain- and loss-offunction analysis of miR-19b on the expression of Mef2c should be carefully considered, as it is possible that these experiments disturb the neuronal functions of miR19b rather than in the progenitors.

As pointed out by the reviewer, it is quite possible that upon manipulation of miR19b its neuronal functions are also perturbed in addition to its function in progenitor cells. After introducing gain-of-function construct in progenitor cells, we have observed changes in the morphology of these cells. These data will be included in the revised version.

The regions of chicken pallium were not consistent among figures: in Figure 1, they showed caudal parts of the pallium (HH29 and 35), while the data in Figure 4 corresponded to the rostral part of the pallium (Figure 4B).  

We will address this by providing images from a similar region of the pallium showing Fezf2 and Mef2c expression patterns.

The neurons expressing Fezf2 and Mef2 in the chicken pallium are not homologous neuronal subtypes to mammalian deep and superficial cortical neurons. The authors must understand that chicken pallial development proceeds in an outside-in manner. Thus, Mef2c-postive neurons in a superficial part are early-born neurons, while FezF2-positive neurons residing in deep areas are later-born neurons. It should be noted that the expression of a single marker gene does not support cell type homology, and the authors' description "the possibility of primitive pallial lamina formation in common ancestors of birds and mammals" is misleading.  

We appreciate this clarification and will modify or remove this statement regarding the “primitive pallial lamina formation” to avoid any confusion and misinterpretation. 

Overexpression of CDKN1A or Sponge-19b induced ectopic expression of Fezf2 in the ventricular zone (Figure 3C, E). Do these cells maintain progenitor statement or prematurely differentiate to neurons? In addition, the authors must explain that the induction of Fezf2 is also detected in GFP-negative cells.  

We propose to follow up on the fate of these cells by extending the observation period post-overexpression of CDKN1A or Sponge-19b to assess whether they retain progenitor characteristics or differentiate. The presence of Fezf2 in GFP-negative cells could be due to the non-cell-autonomous effects, and we will discuss this possibility in the revised manuscript.

Reviewer #2 (Public review):  

Summary:  

This paper investigates the general concept that avian and mammalian pallium specifications share similar mechanisms. To explore that idea, the authors focus their attention on the role of miR-19b as a key controlling factor in the neuronal proliferation/differentiation balance. To do so, the authors checked the expression and protein level of several genes involved in neuronal differentiation, such as NeuroD1 or E2f8, genes also expressed in mammals after conducting their functional gene manipulation experiments. The work also shows a dysregulation in the number of neurons from lower and upper layers when miR-19b expression is altered.  

To test it, the authors conducted a series of functional experiments of gain and loss of function (G&LoF) and enhancer-reporter assays. The enhancer-reporter assays demonstrate a direct relationship between miR-19b and NeuroD1 and E2f8 which is also validated by the G&LoF experiments. It´s also noteworthy to mention that the way miR-19b acts is maintaining the progenitor cells from the ventricular zone in an undifferentiated stage, thus promoting them into a stage of cellular division.  

Overall, the paper argues that the expression of miR-19b in the ventricular zone promotes the cells in a proliferative phase and inhibits the expression of differentiation genes such as E2f8 and NeurD1. The authors claim that a decrease in the progenitor cell pool leads to an increase and decrease in neurons in the lower and upper layers, respectively.  

Strengths:  

(1) Novelty Contribution  

The paper offers strong arguments to prove that the neurodevelopmental basis between mammals and birds is quite the same. Moreover, this work contributes to a better understanding of brain evolution along the animal evolutionary tree and will give us a clearer idea about the roots of how our brain has been developed. This stands in contrast to the conventional framing of mammal brain development as an independent subject unlinked to the "less evolved species". The authors also nicely show a concept that was previously restricted to mammals - the role of microRNAs in development.  

(2) Right experimental approach  

The authors perform a set of functional experiments correctly adjusted to answer the role of miR-19b in the control of neuronal stem cell proliferation and differentiation. Their histological, functional, and genetic approach gives us a clear idea about the relations between several genes involved in the differentiation of the neurons in the avian pallium. In this idea, they maintain the role of miR-19b as a hub controller, keeping the ventricular zone cells in an undifferentiated stage to perpetuate the cellular pool.  

(3) Future directions  

The findings open a door to future experiments, particularly to a better comprehension of the role of microRNAs and pallidal genetic connections. Furthermore, this work also proves the use of avians as a model to study cortical development due to the similarities with mammals.  

Weaknesses:  

While there are questions answered, there are still several that remain unsolved. The experiments analyzed here lead us to speculate that the early differentiation of the progenitor cells from the ventricular zone entails a reduction in the cellular pool, affecting thereafter the number of latter-born neurons (upper layers). The authors should explore that option by testing progenitor cell markers in the ventricular zone, such as Pax6. Even so, it remains possible that miR-19b is also changing the expression pattern of neurons that are going to populate the different layers, instead of their numbers, so the authors cannot rule that out or verify it. Since the paper focuses on the role of miR-19b in patterning, I think the authors should check the relationship and expression between progenitors (Pax6) and intermediate (Tbr2) cells when miR-19b is affected. Since neuronal expression markers change so fast within a few days (HH24HH35), I don't understand why the authors stop the functional experiments at different time points.  

To address this, we will examine the expression of Pax6 and Tbr2 following both gain-of-function and loss-of-function manipulations of miR-19b. We agree with the reviewer that miR-19b may influence not only the number of neurons but also the expression pattern of neuronal markers.  Due to the limitations of our experimental design, we acknowledge that this possibility cannot be ruled out. 

Regarding time points chosen for the functional experiments: We selected different stages based on the expression dynamics of specific markers. To detect possible ectopic induction, we analyzed developmental stages where the expression of a given marker is normally absent. Conversely, to detect loss of expression we examined stages in which the marker is typically expressed robustly. This approach allowed us to better interpret the functional consequences of miR-19b manipulation within relevant developmental windows. 

Reviewer #3 (Public review):  

Summary:  

This is a timely article that focuses on the molecular machinery in charge of the proliferation of pallial neural stem cells in chicks, and aims to compare them to what is known in mammals. miR19b is related to controlling the expression of E2f8 and NeuroD1, and this leads to a proper balance of division/differentiation, required for the generation of the right number of neurons and their subtype proportions. In my opinion, many experiments do reflect an interaction between all these genes and transcription factors, which likely supports the role of miR19b in participating in the proliferation/differentiation balance.  

Strengths:  

Most of the methodologies employed are suitable for the research question, and present data to support their conclusions.  

The authors were creative in their experimental design, in order to assess several aspects of pallial development.  

Weaknesses:  

However, there are several important issues that I think need to be addressed or clarified in order to provide a clearer main message for the article, as well as to clarify the tools employed. I consider it utterly important to review and reinterpret most of the anatomical concepts presented here. The way the are currently used is confusing and may mislead readers towards an understanding of the bird pallium that is no longer accepted by the community.  

Major Concerns:  

(1) Inaccurate use of neuroanatomy throughout the entire article. There are several aspects to it, that I will try to explain in the following paragraphs:  

Figure 1 shows a dynamic and variable expression pattern of miR19b and its relation to NeuroD1. Regardless of the terms used in this figure, it shows that miR19b may be acting differently in various parts of the pallium and developmental stages. However, all the rest of the experiments in the article (except a few cases) abolish these anatomical differences. It is not clear, but it is very important, where in the pallium the experiments are performed. I refer here, at least, to Figures 2C, E, F, H, I; 3D, E; 4C, D, G, I. Regarding time, all experiments were done at HH22, and the article does not show the native expression at this stage. The sacrifice timing is variable, and this variability is not always justified. But more importantly, we don't know where those images were taken, or what part of the pallium is represented in the images. Is it always the same? Do results reflect differences between DVR and Wulst gene expression modifications? The authors should include low magnification images of the regions where experiments were performed. And they should consider the variable expression of all genes when interpreting results.  

We agree that precise anatomical context is essential. In the revised version, we propose to: 

a) Include schematics of the regions of interest where experimental manipulations were performed.

b) Provide low-magnification panoramic images where appropriate, for anatomical reference.

c) Show the expression patterns of relevant marker genes to better justify stages and region selection. 

d) Provide the expression pattern of markers in panoramic view to show differential expression in the DVR and Wulst region and interpret our results accordingly.

b) SVZ is not a postmitotic zone (as stated in line 123, and wrongly assigned throughout the text and figures). On the contrary, the SVZ is a secondary proliferative zone, organized in a layer, located in a basal position to the VZ. Both (VZ and SVZ) are germinative zones, containing mostly progenitors. The only postmitotic neurons in VZ and SVZ occupy them transiently when moving to the mantle zone, which is closer to the meninges and is the postmitotic territory. Please refer to the original Boulder committee articles to revise the SVZ definition. The authors, however, misinterpret this concept, and label the whole mantle zone as it this would be the SVZ. Indeed, the term "mantle zone" does not appear in the article. Please, revise and change the whole text and figures, as SVZ statements and photographs are nearly always misinterpreted. Indeed, SVZ is only labelled well in Figure 4F.  

The two articles mentioning the expression of NeuroD1 in the SVZ (line 118) are research in Xenopus. Is there a proliferative SVZ in Xenopus?  

For the actual existence of the SVZ in the chick pallium, please refer to the recent Rueda-Alaña et al., 2025 article that presents PH3 stainings at different timepoints and pallial areas.  

We appreciate the correction suggested by the reviewer. In the revised manuscript: a) SVZ will be labeled correctly in all figures and descriptions b) The mantle zone terminology will be incorporated appropriately c) The two Xenopus-based references in line 118 will be removed as they are not directly relevant and d) We will refer to the Rueda-Alaña et al., (2025) to guide accurate anatomical labeling and interpretation of proliferative zones.

We also acknowledge that while some proliferative cells exist in the SVZ of the chicken, they are relatively few and do not express typical basal progenitor markers such as Tbr2 (Nomura et al., 2016, Development). We will ensure that this nuance is clearly reflected in the text. 

What is the Wulst, according to the authors of the article? In many figures, the Wulst includes the medial pallium and hippocampus, whereas sometimes it is used as a synonym of the hyperpallium (which excludes the medial pallium and hippocampus). Please make it clear, as the addition or not of the hippocampus definitely changes some interpretations.  

We propose to modify the text and figures to accurately represent the correct location of the Wulst in the chick pallium.

d) The authors compare the entirety of the chick pallium - including the hippocampus (see above), hyperpallium, mesopallium, nidopallium - to only the neocortex of mammals. This view - as shown in Suzuki et al., 2012 - forgets the specificity of pallial areas of the pallium and compares it to cortical cells. This is conceptually wrong, and leads to incorrect interpretations (please refer to Luis Puelles' commentaries on Suzuki et al results); there are incorrect conclusions about the existence of upper-layer-like and deep-layer-like neurons in the pallium of birds. The view is not only wrong according to the misinterpreted anatomical comparisons, but also according to novel scRNAseq data (Rueda-Alaña et al., 2025; Zaremba et al., 2025; Hecker et al., 2025). These articles show that many avian glutamatergic neurons of the pallium have highly diversified, and are not comparable to mammalian cortical cells. The authors should therefore avoid this incorrect use of terminology. There are not such upper-layer-like and deeplayer-like neurons in the pallium of birds.  

We acknowledge this conceptual oversight. In the manuscript: a) We will avoid direct comparisons between the entire chick pallium and the mammalian neocortex b) Terms like “upper-layer-like” and deep-layer-like” neurons will be removed or modified d) We will cite and integrate recent findings from Rueda-Alaña et al. (2025), Zaremba et al. (2025), and Hecker et al. (2025), which provide updated insights from scRNAseq analyses into the complexity of avian pallial neurons. Cell types will be described based on marker gene expression only, without unsupported evolutionary or homology claims.

(2) From introduction to discussion, the article uses misleading terms and outdated concepts of cell type homology and similarity between chick and pallial territories and cells. The authors must avoid this confusing terminology, as non-expert readers will come to evolutionary conclusions which are not supported by the data in this article; indeed, the article does not deal with those concepts.  

We agree with the reviewer. In the revised version, we will remove the misleading terms and outdated concepts and avoid speculative evolutionary conclusions.  

a) Recent articles published in Science (Rueda-Alaña et al., 2025; Zaremba et al., 2025; Hecker et al., 2025) directly contradict some views presented in this article. These articles should be presented in the introduction as they are utterly important for the subject of this article and their results should be discussed in the light of the new findings of this article. Accordingly, the authors should avoid claiming any homology that is not currently supported. The expression of a single gene is not enough anymore to claim the homology of neuronal populations.  

In the revised version, these above-mentioned articles (Rueda-Alaña et al., 2025; Zaremba et al., 2025; Hecker et al., 2025) will be included in the introduction and discussion.  Our interpretations will be updated to reflect these new insights into neuronal diversity and regionalization in the chick pallium. 

Auditory cortex is not an appropriate term, as there is no cortex in the pallium of birds. Cortical areas require the existence of neuronal arrangements in laminae that appear parallel to the ventricular surface. It is not the case of either hyperpallium or auditory DVR. The accepted term, according to the Avian Nomenclature forum, is Field L.  

We will replace all instances of “auditory cortex” with “Field L”, as per the accepted terminology in the Avian Nomenclature Forum.

c) Forebrain, a term overused in the article, is very unspecific. It includes vast areas of the brain, from the pretectum and thalamus to the olfactory bulb. However, the authors are not researching most of the forebrain here. They should be more specific throughout the text and title.  

In the revised version, we will replace “forebrain” with “Pallium” throughout the manuscript to more accurately reflect the regions studied.

(3) In the last part of the results, the authors claim miR19b has a role in patterning the avian pallium. What they see is that modifying its expression induces changes in gene expression in certain neurons. Accordingly, the altered neurons would differentiate into other subtypes, not similar to the wild type example. In this sense, miR19b may have a role in cell specification or neuronal differentiation. However, patterning is a different developmental event, which refers to the determination of broad genetic areas and territories. I don't think miR19b has a role in patterning.  

We agree with the reviewers that an alteration in one marker for a particular cell type may not indicate a change in patterning. However, including the effect of miR-19b gain- and loss-of-function on Pax6 and Tbr2, may strengthen the idea that it affects patterning as suggested by reviewer #2. 

(4) Please add a scheme of the molecules described in this article and the suggested interaction between them.  

In the revised version, we propose to include a diagram to visually summarize the proposed interactions between miR-19b, E2f8, NeuroD1, and other key regulators.  

(5) The methods section is way too brief to allow for repeatability of the procedures. This may be due to an editorial policy but if possible, please extend the details of the experimental procedures.  

We will expand the Methods section to provide more detailed protocols and justifications for experimental design, in alignment with journal policy.

Author response:

The following is the authors’ response to the original reviews

Reviewer #1 (Public review):

Summary:

In this study, the authors aim to understand the neural basis of implicit causal inference, specifically how people infer causes of illness. They use fMRI to explore whether these inferences rely on content-specific semantic networks or broader, domain-general neurocognitive mechanisms. The study explores two key hypotheses: first, that causal inferences about illness rely on semantic networks specific to living things, such as the 'animacy network,' given that illnesses affect only animate beings; and second, that there might be a common brain network supporting causal inferences across various domains, including illness, mental states, and mechanical failures. By examining these hypotheses, the authors aim to determine whether causal inferences are supported by specialized or generalized neural systems.

The authors observed that inferring illness causes selectively engaged a portion of the precuneus (PC) associated with the semantic representation of animate entities, such as people and animals. They found no cortical areas that responded to causal inferences across different domains, including illness and mechanical failures. Based on these findings, the authors concluded that implicit causal inferences are supported by content-specific semantic networks, rather than a domain-general neural system, indicating that the neural basis of causal inference is closely tied to the semantic representation of the specific content involved.

Strengths:

(1) The inclusion of the four conditions in the design is well thought out, allowing for the examination of the unique contribution of causal inference of illness compared to either a different type of causal inference (mechanical) or non-causal conditions. This design also has the potential to identify regions involved in a shared representation of inference across general domains.

(2) The presence of the three localizers for language, logic, and mentalizing, along with the selection of specific regions of interest (ROIs), such as the precuneus and anterior ventral occipitotemporal cortex (antVOTC), is a strong feature that supports a hypothesis-driven approach (although see below for a critical point related to the ROI selection).

(3) The univariate analysis pipeline is solid and well-developed.

(4) The statistical analyses are a particularly strong aspect of the paper.

Weaknesses:

Based on the current analyses, it is not yet possible to rule out the hypothesis that inferring illness causes relies on neurocognitive mechanisms that support causal inferences irrespective of their content, neither in the precuneus nor in other parts of the brain.

(1) The authors, particularly in the multivariate analyses, do not thoroughly examine the similarity between the two conditions (illness-causal and mechanical-causal), as they are more focused on highlighting the differences between them. For instance, in the searchlight MVPA analysis, an interesting decoding analysis is conducted to identify brain regions that represent illness-causal and mechanical-causal conditions differently, yielding results consistent with the univariate analyses. However, to test for the presence of a shared network, the authors only perform the Causal vs. Non-causal analysis. This analysis is not very informative because it includes all conditions mixed together and does not clarify whether both the illness-causal and mechanical-causal conditions contribute to these results.

(2) To address this limitation, a useful additional step would be to use as ROIs the different regions that emerged in the Causal vs. Non-causal decoding analysis and to conduct four separate decoding analyses within these specific clusters:

(a) Illness-Causal vs. Non-causal - Illness First;

(b) Illness-Causal vs. Non-causal - Mechanical First;

(c) Mechanical-Causal vs. Non-causal - Illness First;

(d) Mechanical-Causal vs. Non-causal - Mechanical First.

This approach would allow the authors to determine whether any of these ROIs can decode both the illness-causal and mechanical-causal conditions against at least one non-causal condition.

(3) Another possible analysis to investigate the existence of a shared network would be to run the searchlight analysis for the mechanical-causal condition versus the two non-causal conditions, as was done for the illness-causal versus non-causal conditions, and then examine the conjunction between the two. Specifically, the goal would be to identify ROIs that show significant decoding accuracy in both analyses.

The hypothesis that a neural mechanism supports causal inference across domains predicts higher univariate responses when causal inferences occur than when they do not. This prediction was not generated by us ad hoc but rather has been made by almost all previous cognitive neuroscience papers on this topic (Ferstl & von Cramon, 2001; Satpute et al., 2005; Fugelsang & Dunbar, 2005; Kuperberg et al., 2006; Fenker et al., 2010; Kranjec et al., 2012; Pramod, Chomik-Morales, et al., 2023; Chow et al., 2008; Mason & Just, 2011; Prat et al., 2011). Contrary to this hypothesis, we find that the precuneus (PC) is most activated for illness inferences and most deactivated for mechanical inferences relative to rest, suggesting that the PC does not support domain-general causal inference. To further probe the selectivity of the PC for illness inferences, we created group overlap maps that compare PC responses to illness inferences and mechanical inferences across participants. The PC shows a strong preference for illness inferences and is therefore unlikely to support causal inferences irrespective of their content (Supplementary Figures 6 and 7). We also note that, in whole-cortex analysis, no shared regions responded more to causal inference than noncausal vignettes across domains. Therefore, the prediction made by the ‘domain-general causal engine’ proposal as it has been articulated in the literature is not supported in our data.

Taking a multivariate approach, the hypothesis that a neural mechanism supports causal inference across domains also predicts that relevant regions can decode between all possible pairs of causal vs. noncausal conditions (e.g., Illness-Causal vs. Noncausal-Illness First, Mechanical-Causal vs. Noncausal-Illness First, etc.). The analysis described by the reviewer in (2), in which the regions that distinguish between causal vs. noncausal conditions in searchlight MVPA are used as ROIs to test various causal vs. noncausal contrasts, is non-independent. Therefore, we cannot perform this analysis. In accordance with the reviewer’s suggestions in (3), now include searchlight MVPA results for the mechanical inference condition compared to the two noncausal conditions (Supplementary Figure 9). No regions are shared across the searchlight analyses comparing all possible pairs of causal and noncausal conditions, providing further evidence that there are no shared neural responses to causal inference in our dataset.

(4) Along the same lines, for the ROI MVPA analysis, it would be useful not only to include the illness-causal vs. mechanical-causal decoding but also to examine the illness-causal vs. non-causal conditions and the mechanical-causal vs. non-causal conditions. Additionally, it would be beneficial to report these data not just in a table (where only the mean accuracy is shown) but also using dot plots, allowing the readers to see not only the mean values but also the accuracy for each individual subject.

We have performed these analyses and now include a table of the results as well as figures displaying the dispersion across participants (Supplementary Tables 2 and 3, Supplementary Figures 10 and 11). In the left PC, the illness inference condition was decoded from one of the noncausal conditions, and the mechanical inference condition was decoded from the same noncausal condition. The language network did not decode between any causal/noncausal pairs. In the logic network, the illness inference condition was decoded from one of the noncausal conditions, and the mechanical inference condition was decoded from the other noncausal condition. Thus, no regions showed the predicted ‘domain-general’ pattern, i.e., significant decoding between all causal/noncausal pairs. 

Importantly, the decoding results must be interpreted in light of significant univariate differences across conditions (e.g., greater responses to illness inferences compared to noncausal vignettes in the PC). Linear classifiers are highly sensitive to univariate differences (Coutanche, 2013; Kragel et al., 2012; Hebart & Baker, 2018; Woolgar et al., 2014; Davis et al., 2014; Pakravan et al., 2022).

(5) The selection of Regions of Interest (ROIs) is not entirely straightforward:

In the introduction, the authors mention that recent literature identifies the precuneus (PC) as a region that responds preferentially to images and words related to living things across various tasks. While this may be accurate, we can all agree that other regions within the ventral occipital-temporal cortex also exhibit such preferences, particularly areas like the fusiform face area, the occipital face area, and the extrastriate body area. I believe that at least some parts of this network (e.g., the fusiform gyrus) should be included as ROIs in this study. This inclusion would make sense, especially because a complementary portion of the ventral stream known to prefer non-living items (i.e., anterior medial VOTC) has been selected as a control ROI to process information about the mechanical-causal condition. Given the main hypothesis of the study - that causal inferences about illness might depend on content-specific semantic representations in the 'animacy network' - it would be worthwhile to investigate these ROIs alongside the precuneus, as they may also yield interesting results.

We thank the reviewer for their suggestion to test the FFA region. We think this provides an interesting comparison to the PC and hypothesized that, in contrast to the PC, the FFA does not encode abstract causal information about animacy-specific processes (i.e., illness). As we mention in the Introduction, although the fusiform face area (FFA) also exhibits a preference for animates, it does so primarily for images in sighted people (Kanwisher et al., 1997; Kanwisher et al., 1997; Grill-Spector et al., 2004; Noppeney et al., 2006; Konkle & Caramazza, 2013; Connolly et al., 2016; Bi et al., 2016).

We did not select the FFA as a region of interest when preregistering the current study because we did not predict it would show sensitivity to causal knowledge. In accordance with the reviewer’s suggestions, we now include the FFA as an ROI in individual-subject univariate analysis (Supplementary Figure 8, Appendix 4). Because we did not run a separate FFA localizer task when collecting the data, we used FFA search spaces from a previous study investigating responses to face images (Julian et al., 2012). We followed the same analysis procedure that was used to investigate responses to illness inferences in the PC. Neither left nor right FFA exhibited a preference for illness inferences compared to mechanical inferences or to the noncausal conditions. This result is interesting and is now briefly discussed in the Discussion section.

(6) Visual representation of results:

In all the figures related to ROI analyses, only mean group values are reported (e.g., Figure 1A, Figure 3, Figure 4A, Supplementary Figure 6, Figure 7, Figure 8). To better capture the complexity of fMRI data and provide readers with a more comprehensive view of the results, it would be beneficial to include a dot plot for a specific time point in each graph. This could be a fixed time point (e.g., a certain number of seconds after stimulus presentation) or the time point showing the maximum difference between the conditions of interest. Adding this would allow for a clearer understanding of how the effect is distributed across the full sample, such as whether it is consistently present in every subject or if there is greater variability across individuals.

We thank the reviewer for this suggestion. We now include scattered box plots displaying the dispersion in average percent signal change across participants in Figures 1, 3, and 4, and Supplementary Figures 8, 12, and 14.

(7) Task selection:

(a) To improve the clarity of the paper, it would be helpful to explain the rationale behind the choice of the selected task, specifically addressing: (i) why an implicit inference task was chosen instead of an explicit inference task, and (ii) why the "magic detection" task was used, as it might shift participants' attention more towards coherence, surprise, or unexpected elements rather than the inference process itself.

(b) Additionally, the choice to include a large number of catch trials is unusual, especially since they are modeled as regressors of non-interest in the GLM. It would be beneficial to provide an explanation for this decision.

We chose an orthogonal foil detection task, rather than an explicit causal judgment task, to investigate automatic causal inferences during reading and to unconfound such processing as much as possible from explicit decision-making processes (see Kuperberg et al., 2006 for discussion). Analogous foil detection paradigms have been used to study sentence processing and word recognition (Pallier et al., 2011; Dehaene-Lambertz et al., 2018). We now clarify this in the Introduction. The “magical” element occurred both within and across sentences so that participants could not use coherence as a cue to complete the task. Approximately 1/5 (19%) of the trials were magical catch trials to ensure that participants remained attentive throughout the experiment.

Reviewer #2 (Public review):

Summary:

In this study, the authors hypothesize that "causal inferences about illness depend on content-specific semantic representations in the animacy network". They test this hypothesis in an fMRI task, by comparing brain activity elicited by participants' exposure to written situations suggesting a plausible cause of illness with brain activity in linguistically equivalent situations suggesting a plausible cause of mechanical failure or damage and non-causal situations. These contrasts identify PC as the main "culprit" in a whole-brain univariate analysis. Then the question arises of whether the content-specificity has to do with inferences about animates in general, or if there are some distinctions between reasoning about people's bodies versus mental states. To answer this question, the authors localize the mentalizing network and study the relation between brain activity elicited by Illness-Causal > Mech-Causal and Mentalizing > Physical stories. They conclude that inferring about the causes of illness partially differentiates from reasoning about people's states of mind. The authors finally test the alternative yet non-mutually exclusive hypothesis that both types of causal inferences (illness and mechanical) depend on shared neural machinery. Good candidates are language and logic, which justifies the use of a language/logic localizer. No evidence of commonalities across causal inferences versus non-causal situations is found.

Strengths:

(1) This study introduces a useful paradigm and well-designed set of stimuli to test for implicit causal inferences.

(2) Another important methodological advance is the addition of physical stories to the original mentalizing protocol.

(3) With these tools, or a variant of these tools, this study has the potential to pave the way for further investigation of naïve biology and causal inference.

Weaknesses:

(1) This study is missing a big-picture question. It is not clear whether the authors investigate the neural correlates of causal reasoning or of naïve biology. If the former, the choice of an orthogonal task, making causal reasoning implicit, is questionable. If the latter, the choice of mechanical and physical controls can be seen as reductive and problematic.

We have modified the Introduction to clarify that the primary goal of the current study is to test the claim that semantic networks encode causal knowledge – in this case, causal intuitive theories of biology. Most conceptions of intuitive biology, intuitive psychology, and intuitive physics describe them as causal frameworks (e.g., Wellman & Gelman, 1992; Simons & Keil, 1995; Keil et al., 1999; Tenenbaum, Griffiths, & Niyogi, 2007; Gopnik & Wellman, 2012; Gerstenberg & Tenenbaum, 2017). As noted above, we chose an implicit task to investigate automatic causal inferences during reading and to unconfound such processing as much as possible from explicit decision-making processes. We are not sure what the reviewer means when they say that mechanical and physical controls are reductive. This is the standard control condition in neural and behavioral paradigms that investigate intuitive psychology and intuitive biology (e.g., Saxe & Kanwisher, 2003; Gelman & Wellman, 1991).

(2) The rationale for focusing mostly on the precuneus is not clear and this choice could almost be seen as a post-hoc hypothesis.

This study is preregistered (https://osf.io/6pnqg). The preregistration states that the precuneus is a hypothesized area of interest, so this is not a post-hoc hypothesis. Our hypothesis was informed by multiple prior studies implicating the precuneus in the semantic representation of animates (e.g., people, animals) (Fairhall & Caramazza, 2013a, 2013b; Fairhall et al., 2014; Peer et al., 2015; Wang et al., 2016; Silson et al., 2019; Rabini, Ubaldi, & Fairhall, 2021; Deen & Freiwald, 2022; Aglinskas & Fairhall, 2023; Hauptman, Elli, et al., 2025). We also conducted a pilot experiment with separate participants prior to pre-registering the study. We now clarify our rationale for focusing on the precuneus in the Introduction:

“Illness affects living things (e.g., people and animals) rather than inanimate objects (e.g., rocks, machines, houses). Thinking about living things (animates) as opposed to non-living things (inanimate objects/places) recruits partially distinct neural systems (e.g., Warrington & Shallice, 1984; Hillis & Caramazza, 1991; Caramazza & Shelton, 1998; Farah & Rabinowitz, 2003). The precuneus (PC) is part of the ‘animacy’ semantic network and responds preferentially to living things (i.e., people and animals), whether presented as images or words (Devlin et al., 2002; Fairhall & Caramazza, 2013a, 2013b; Fairhall et al., 2014; Peer et al., 2015; Wang et al., 2016; Silson et al., 2019; Rabini, Ubaldi, & Fairhall, 2021; Deen & Freiwald, 2022; Aglinskas & Fairhall, 2023; Hauptman, Elli, et al., 2025). By contrast, parts of the visual system (e.g., fusiform face area) that respond preferentially to animates do so primarily for images (Kanwisher et al., 1997; Grill-Spector et al., 2004; Noppeney et al., 2006; Mahon et al., 2009; Konkle & Caramazza, 2013; Connolly et al., 2016; see Bi et al., 2016 for a review). We hypothesized that the PC represents causal knowledge relevant to animates and tested the prediction that it would be activated during implicit causal inferences about illness, which rely on such knowledge (preregistration: https://osf.io/6pnqg).”

(3) The choice of an orthogonal 'magic detection' task has three problematic consequences in this study:

(a) It differs in nature from the 'mentalizing' task that consists of evaluating a character's beliefs explicitly from the corresponding story, which complicates the study of the relation between both tasks. While the authors do not compare both tasks directly, it is unclear to what extent this intrinsic difference between implicit versus explicit judgments of people's body versus mental states could influence the results.

(b) The extent to which the failure to find shared neural machinery between both types of inferences (illness and mechanical) can be attributed to the implicit character of the task is not clear.

(c) The introduction of a category of non-interest that contains only 36 trials compared to 38 trials for all four categories of interest creates a design imbalance.

We disagree with the reviewer’s argument that our use of an implicit “magic detection” task is problematic. Indeed, we think it is one of the advances of the current study over prior work.

a) Prior work has shown that implicit mentalizing tasks (e.g., naturalistic movie watching) engages the theory of mind network, suggesting that the implicit/explicit nature of the task does not drive the activation of this network (Jacoby et al., 2016; Richardson et al., 2018). With these data in mind, it is unlikely that the implicit/explicit nature of the causal inference and theory of mind tasks in the present experiment can explain observed differences between them.

b) Explicit causal inferences introduce a collection of executive processes that potentially confound the results and make it difficult to know whether neural signatures are related to causal inference per se. The current study focuses on the neural basis of implicit causal inference, a type of inference that is made routinely during language comprehension. We do not claim to find neural signatures of all causal inferences, we do not think any study could claim to do so because causal inferences are a highly varied class.

c) Our findings do not exclude the possibility that content-invariant responses are elicited during explicit causality judgments. We clarify this point in the Results (e.g., “These results leave open the possibility that domain-general systems support the explicit search for causal connections”) and Discussion (e.g., “The discovery of novel causal relationships (e.g., ‘blicket detectors’; Gopnik et al., 2001) and the identification of complex causes, even in the case of illness, may depend in part on domain-general neural mechanisms”).

d) Because the magic trials are excluded from our analyses, it is unclear how the imbalance in the number of magic trials could influence the results and our interpretation of them. We note that the number of catch trials in standard target detection paradigms are sometimes much lower than the number of target trials in each condition (e.g., Pallier et al., 2011).

(4) Another imbalance is present in the design of this study: the number of trials per category is not the same in each run of the main task. This imbalance does not seem to be accounted for in the 1st-level GLM and renders a bit problematic the subsequent use of MVPA.

Each condition is shown either 6 or 7 times per run (maximum difference of 1 trial between conditions), and the number of trials per condition is equal across the whole experiment: each condition is shown 7 times in two of the runs and 6 times four of the runs. This minor design imbalance is typical of fMRI experiments and should not impact our interpretations of the data, particularly because we average responses from each condition within a run before submitting them to MVPA.

(5) The main claim of the authors, encapsulated by the title of the present manuscript, is not tested directly. While the authors included in their protocol independent localizers for mentalizing, language, and logic, they did not include an independent localizer for "animacy". As such, they cannot provide a within-subject evaluation of their claim, which is entirely based on the presence of a partial overlap in PC (which is also involved in a wide range of tasks) with previous results on animacy.

We respectfully disagree with this assertion. Our primary analysis uses a within-subject leave-one-run-out approach. This approach allows us to use part of the data itself to localize animacy-relevant causal responses in the PC without engaging in ‘double-dipping’ or statistical non-independence (Vul & Kanwisher, 2011). We also use the mentalizing network localizer as a partial localizer for animacy. This is because the control condition (physical reasoning) does not include references to people or any animate agents (Supplementary Figures 1 and 15). We now clarify this point in Methods section of the paper (see below).

From the Methods: “To test the relationship between neural responses to inferences about the body and the mind, and to localize animacy regions, we used a localizer task to identify the mentalizing network in each participant (Saxe & Kanwisher, 2003; Dodell-Feder et al., 2011; http://saxelab.mit.edu/use-our-efficient-false-belief-localizer)...Our physical stories incorporated more vivid descriptions of physical interactions and did not make any references to human agents, enabling us to use the mentalizing localizer as a localizer for animacy.”

Reviewer #3 (Public review):

Summary:

This study employed an implicit task, showing vignettes to participants while a bold signal was acquired. The aim was to capture automatic causal inferences that emerge during language processing and comprehension. In particular, the authors compared causal inferences about illness with two control conditions, causal inferences about mechanical failures and non-causal phrases related to illnesses. All phrases that were employed described contexts with people, to avoid animacy/inanimate confound in the results. The authors had a specific hypothesis concerning the role of the precuneus (PC) in being sensitive to causal inferences about illnesses.

These findings indicate that implicit causal inferences are facilitated by semantic networks specialized for encoding causal knowledge.

Strengths:

The major strength of the study is the clever design of the stimuli (which are nicely matched for a number of features) which can tease apart the role of the type of causal inference (illness-causal or mechanical-causal) and the use of two localizers (logic/language and mentalizing) to investigate the hypothesis that the language and/or logical reasoning networks preferentially respond to causal inference regardless of the content domain being tested (illnesses or mechanical).

Weaknesses:

I have identified the following main weaknesses:

(1) Precuneus (PC) and Temporo-Parietal junction (TPJ) show very similar patterns of results, and the manuscript is mostly focused on PC (also the abstract). To what extent does the fact that PC and TPJ show similar trends affect the inferences we can derive from the results of the paper? I wonder whether additional analyses (connectivity?) would help provide information about this network.

We thank the reviewer for this suggestion. While the PC shows the most robust univariate preference for illness inferences compared to both mechanical inferences and noncausal vignettes, the TPJ also shows a preference for illness inferences compared to mechanical inferences in individual-subject fROI analysis. However, as we mention in the Results section, the TPJ does not show a preference for illness inferences compared to noncausal vignettes, suggesting that the TPJ is selective for animacy but may not be as sensitive to causal knowledge about animacy-specific processes. When describing our results, we refer to the ‘animacy network’ (i.e., PC and TPJ) but also highlight that the PC exhibited the most robust responses to illness inferences (from the Results: “Inferring illness causes preferentially recruited the animacy semantic network, particularly the PC”; from the Discussion: “We find that a semantic network previously implicated in thinking about animates, particularly the precuneus (PC), is preferentially engaged when people infer causes of illness…”). We did not collect resting state data that would enable a connectivity analysis, as the reviewer suggests. This is an interesting direction for future work.

(2) Results are mainly supported by an univariate ROI approach, and the MVPA ROI approach is performed on a subregion of one of the ROI regions (left precuneus). Results could then have a limited impact on our understanding of brain functioning.

The original and current versions of the paper include results from multiple multivariate analyses, including whole-cortex searchlight MVPA and individual-subject fROI MVPA performed in multiple search spaces (see Supplementary Figures 10 and 11, Supplementary Tables 2 and 3).

We note that our preregistered predictions focused primarily on univariate differences. This is because the current study investigates neural responses to inferences, and univariate increases in activity is thought to reflect the processing of such inferences. We use multivariate analyses to complement our primary univariate analyses. However, given that we observe significant univariate effects and that multivariate analyses are heavily influenced by significant univariate effects (Coutanche, 2013; Kragel et al., 2012; Hebart & Baker, 2018; Woolgar et al., 2014; Davis et al., 2014; Pakravan et al., 2022), our univariate results constitute the main findings of the paper.

(3) In all figures: there are no measures of dispersion of the data across participants. The reader can only see aggregated (mean) data. E.g., percentage signal changes (PSC) do not report measures of dispersion of the data, nor do we have bold maps showing the overlap of the response across participants. Only in Figure 2, we see the data of 6 selected participants out of 20.

We thank the reviewer for this suggestion. We now include graphs depicting the dispersion of the data across participants in the following figures: Figures 1, 3, and 4, and Supplementary Figures 8, 12, and 14. We have also created 2 figures that display the overlap of univariate responses across participants (Supplementary Figures 6 and 7). These figures show that there is high overlap across participants in PC responses to illness inferences but not mechanical inferences. In addition, all participants’ results from the analysis depicted in Figure 2 are included in Supplementary Figure 3. 

(4) Sometimes acronyms are defined in the text after they appear for the first time.

We thank the reviewer for pointing this out. We now define all acronyms before using them.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) I was unable to access the pre-registration on OSF because special permission is required.

We apologize for this technical error. The preregistration is now publicly available: https://osf.io/6pnqg.

(2) The length of the MRI session is quite long (around 2 hours). It is generally discouraged to have such extended data acquisition periods, as this can affect the stability and cleanliness of the data. Did you observe any effects of fatigue or attention decline in your data?

The session was 2 hours long including 1-2 10-minute breaks. Without breaks, the scan would be approximately 1.5 hours. This is a standard length for MRI experiments. The main experiment (causal inference task) was always conducted first and lasted approximately 1 hour. Accuracy did not decrease across the 6 runs of this experiment (repeated measures ANOVA, F_(5,114) = 1.35, p = .25).

(3) The last sentence of the results states: "Although MVPA searchlight analysis identified several areas where patterns of activity distinguished between causal and non-causal vignettes, all of these regions showed a preference for non-causal vignettes in univariate analysis (Supplementary Figure 5)." This statement is not entirely accurate. As I previously pointed out, the MVPA searchlight analysis is not very informative and is difficult to interpret. However, as previously suggested, there are additional steps that could be taken to better understand and interpret these results. It is incorrect to conclude that because the brain regions identified in the MVPA analyses show a preference for non-causal vignettes in univariate analyses, the multivariate results lack value. While univariate analyses may show a preference for a specific condition, multivariate analyses can reveal more fine-grained representations of multiple conditions. For a notable example, consider the fusiform face area (FFA) that shows a clear preference for faces at the univariate level but can significantly decode other categories at the multivariate level, even when faces are not included in the analysis.

The decoding analysis that the reviewer is suggesting for the current study would be analogous to identifying univariate differences between faces and places in the FFA and then decoding between faces and places and claiming that the FFA represents places because the decoding is significant. The decoding analyses enabled by our design are not equivalent to decoding within a condition (e.g., among face identities, among types of illness inferences), as the reviewer suggests above. It is not that such multivariate analyses “lack value” but that they recapitulate established univariate differences. Multivariate analyses are useful for revealing more fine-grained representations when i) significant univariate differences are not observed, or ii) when it is possible to decode among categories within a condition (e.g., among face identities, among types of illness inferences). We are currently collecting data that will enable us to perform within-condition decoding analyses in future work, but the design of the current study does not allow for such a comparison.

We note that the original quotation from the manuscript has been removed because it is no longer accurate. When including participant response time as a covariate of no interest in the GLM, no regions are shared across the 4 searchlight analyses comparing causal and noncausal conditions, suggesting that there are no shared neural responses to causal inference in our dataset.

Reviewer #2 (Recommendations for the authors):

(1) Moderating the strength of some claims made to justify the main hypothesis (e.g., "people but not machines transmit diseases to each other through physical contact").

We changed this wording so that it now reads: “Illness affects living things (e.g., people and animals) rather than inanimate objects (e.g., rocks, machines, houses).” (Introduction)

(2) Expanding the paragraph introducing the sub-question about inferring people's "body states" vs "mental states". In addition, given the order in which the hypotheses are introduced, and the results are presented, I would suggest switching the order of presentation of both localizers in the methods section and adding a quick reminder of the hypotheses that justify using these localizers.

We thank the reviewer for these suggestions. In accordance their suggestions, we have expanded the paragraph Introduction that introduces the “body states” vs. “mental states” question (see below). We have also switched the order of the localizer descriptions in the Methods section and added a sentence at the start of each section describing the relevant hypotheses (see below).

From the Introduction: “We also compared neural responses to causal inferences about the body (i.e., illness) and inferences about the mind (i.e., mental states). Both types of inferences are about animate entities, and some developmental work suggests that children use the same set of causal principles to think about bodies and minds (Carey, 1985, 1988). Other evidence suggests that by early childhood, young children have distinct causal knowledge about the body and the mind (Springer & Keil, 1991; Callanan & Oakes, 1992; Wellman & Gelman, 1992; Inagaki & Hatano, 1993; 2004; Keil, 1994; Hickling & Wellman, 2001; Medin et al., 2010). For instance, preschoolers are more likely to view illness as a consequence of biological causes, such as contagion, rather than psychological causes, such as malicious intent (Springer & Ruckel, 1992; Raman & Winer, 2004; see also Legare & Gelman, 2008). The neural relationship between inferences about bodies and minds has not been fully described. The ‘mentalizing network’, including the PC, is engaged when people reason about agents’ beliefs (Saxe & Kanwisher, 2003; Saxe et al., 2006; Saxe & Powell, 2006; Dodell-Feder et al., 2011; Dufour et al., 2013). We localized this network in individual participants and measured its neuroanatomical relationship to the network activated by illness inferences.”

From the Methods, localizer descriptions: “To test the relationship between neural responses to inferences about the body and the mind, and to localize animacy regions, we used a localizer task to identify the mentalizing network in each participant… To test for the presence of domain-general responses to causal inference in the language and logic networks (e.g., Kuperberg et al., 2006; Operskalski & Barbey, 2017), we used an additional localizer task to identify both networks in each participant.”

(3) Adding a quick analysis of lateralization to support the corresponding claim of left lateralization of responses to causal inferences.

In accordance with the reviewer’s suggestion, we now include hemisphere as a factor in all ANOVAs comparing univariate responses across conditions.

From the Results: “In individual-subject fROI analysis (leave-one-run-out), we similarly found that inferring illness causes activated the PC more than inferring causes of mechanical breakdown (repeated measures ANOVA, condition (Illness-Causal, Mechanical-Causal) x hemisphere (left, right): main effect of condition, F_(1,19) = 19.18, p < .001, main effect of hemisphere, F_(1,19) = 0.3, p = .59, condition x hemisphere interaction, F_(1,19) = 27.48, p < .001; Figure 1A). This effect was larger in the left than in the right PC (paired samples t-tests; left PC: t₍₁₉₎ = 5.36, p < .001, right PC: t₍₁₉₎ = 2.27, p = .04)…In contrast to the animacy-responsive PC, the anterior PPA showed the opposite pattern, responding more to mechanical inferences than illness inferences (leave-one-run-out individual-subject fROI analysis; repeated measures ANOVA, condition (Mechanical-Causal, Illness-Causal) x hemisphere (left, right): main effect of condition, F_(1,19) = 17.93, p < .001, main effect of hemisphere, F_(1,19) = 1.33, p = .26, condition x hemisphere interaction, F_(1,19) = 7.8, p = .01; Figure 4A). This effect was significant only in the left anterior PPA (paired samples t-tests; left anterior PPA: t₍₁₉₎ = 4, p < .001, right anterior PPA: t₍₁₉₎ = 1.88, p = .08).”

(4) Making public and accessible the pre-registration OSF link.

We apologize for this technical error. The preregistration is now publicly available: https://osf.io/6pnqg.

Reviewer #3 (Recommendations for the authors):

In all figures: there are no measures of dispersion of the data across participants. The reader can only see aggregated (mean) data. E.g., percentage signal changes (PSC) do not report measures of dispersion of the data, nor do we have bold maps showing the overlap of the response across participants. Only in Figure 2, we see the data of 6 selected participants out of 20.

We thank the reviewer for this suggestion. We now include graphs depicting the dispersion of the data across participants in the following figures: Figures 1, 3, and 4, and Supplementary Figures 8, 12, and 14. We have also created 2 figures that display the overlap of univariate responses across participants (Supplementary Figures 6 and 7). In addition, all participants’ results from the analysis depicted in Figure 2 are included in Supplementary Figure 3.

Minor

(1) Figure 2: Spatial dissociation between responses to illness inferences and mental state inferences in the precuneus (PC). If the analysis is the result of the MVPA, the figure should report the fact that only the left precuneus was analyzed.

Figure 2 depicts the spatial dissociation in univariate responses to illness inferences and mental state inferences. We now clarify this in the figure legend.

(2) VOTC and PSC acronyms are defined in the text after they appear for the first time. TPJ is never defined.

We thank the reviewer for pointing this out. We now define all acronyms before using them.

Author response:

The following is the authors’ response to the original reviews

Reviewer #1 (Public review):

Summary:

The paper addresses the knowledge gap between the representation of goal direction in the central complex and how motor systems stabilize movement toward that goal. The authors focused on two descending neurons, DNa01 and 02, and showed that they play different roles in steering the fly toward a goal. They also explored the connectome data to propose a model to explain how these DNs could mediate response to lateralized sensory inputs. They finally used lateralized optogenetic activation/inactivation experiments to test the roles of these neurons in mediating turnings in freely walking flies.

Strengths:

The experiments are well-designed and controlled. The experiment in Figure 4 is elegant, and the authors put a lot of effort into ensuring that ATP puffs do not accidentally activate the DNs. They also have explained complex experiments well. I only have minor comments for the authors.

We are grateful for this positive feedback.

Weaknesses:

(1) I do not fully understand how the authors extracted the correlation functions from the population data in Figure 1. Since the ipsilateral DNs are anti-correlated with the contralateral ones, I expected that the average will drop to zero when they are pooled together (e.g., 1E-G). Of course, this will not be the case if all the data in Figure 1 are collected from the same brain hemisphere. It would be helpful if the authors could explain this.

We regret that this information was not easy to find in our initial submission. As noted in the Figure 1D legend, Here and elsewhere, ipsi and contra are defined relative to the recorded DN(s). We have now added a sentence to the Results (right after we introduce Figure 1D) that also makes this point.

(2) What constitutes the goal directions in Figures 1-3 and 8, as the authors could not use EPG activity as a proxy for goal directions? If these experiments were done in the dark, without landmarks, one would expect the fly's heading to drift randomly at times, and they would not engage the DNa01/02 for turning. Do the walking trajectories in these experiments qualify as menotactic bouts?

Published work (Green et al., 2019) has shown that, even in the dark, flies will often walk for extended periods while holding the bump of EPG activity at a fixed location. During these epochs, the brain is essentially estimating that the fly is walking in a straight line in a fixed direction. (The fact that the fly is actually rotating a bit on the spherical treadmill is not something the fly can know, in the dark.) Thus, epochs where the EPG bump is held fixed are treated as menotactic bouts, even in darkness.

Our results provide additional support for this interpretation. We find that, when flies are walking in darkness and holding the bump of EPG activity at a fixed location, they will make a corrective behavioral turning maneuver in response to an imposed bump-jump. This result argues that the flies are actually engaging in goal-directed straight-line walking, i.e. menotaxis, and it reproduces the findings of Green et al. (2019).

To clarify this point, we have adjusted the wording of the Results pertaining to Figure 4.

(3) In Figure 2B, the authors mentioned that DNa02 overpredicts and 01 underpredicts rapid turning and provided single examples. It would be nice to see more population-level quantification to support this claim.

In this revision, we have reorganized Figures 1 and 2 (and associated text) to improve clarity. As part of this reorganization, we have removed this passage from the text, as it was a minor point in any event.

Reviewer #2 (Public review):

The data is largely electrophysiological recordings coupled with behavioral measurements (technically impressive) and some gain-of-function experiments in freely walking flies. Loss-of-function was tested but had minimal effect, which is not surprising in a system with partially redundant control mechanisms. The data is also consistent with/complementary to subsequent manuscripts (Yang 2023, Feng 2024, and Ros 2024) showing additional descending neurons with contributions to steering in walking and flying.

The experiments are well executed, the results interesting, and the description clear. Some hypotheses based on connectome anatomy are tested: the insights on the pre-synaptic side - how sensory and central complex heading circuits converge onto these DNs are stronger than the suggestions about biomechanical mechanisms for how turning happens on the motor side.

Of particular interest is the idea that different sensory cues can converge on a common motor program. The turn-toward or turn-away mechanism is initiated by valence rather than whether the stimulus was odor or temperature or memory of heading. The idea that animals choose a direction based on external sensory information and then maintain that direction as a heading through a more internal, goal-based memory mechanism, is interesting but it is hard to separate conclusively.

To clarify, we mention the role of memory in connection with two places in the manuscript. First, we note that the EPG/head direction system relies on learning and memory to construct a map of directional cues in the environment. These cues are, in principle, inherently neutral, i.e. without valence. Second, we note that specific mushroom body output neurons rely on learning and memory to store the valence associated with an odor. This information is not necessarily associated with an allocentric direction: it is simply the association of odor with value. Both of these ideas are well-attested by previous work.

The reviewer may be suggesting a sequential scheme whereby the brain initializes an allocentric goal direction based on valence, and then maintains that goal direction in memory, based on that initialization. In other words, memory is used to associate valence with some allocentric direction. This seems plausible, but it is not a claim we make in our manuscript.

The "see-saw", where left-right symmetry is broken to allow a turn, presumably by excitation on one side and inhibition of the other leg motor modules, is interesting but not well explained here. How hyperpolarization affects motor outputs is not clear.

We have added several sentences to the Discussion to clarify this point. According to this see-saw model, steering can emerge from right/left asymmetries in excitation, or inhibition, or both. It may be nonintuitive to think that inhibitory input to a DN can produce an action. However, this becomes more plausible given our finding that DNa02 has a relatively high basal firing rate (Fig. 1D), and DNa02 hyperpolarization is associated with contraversive turning (Fig. 5A). It is also relevant to note that there are many inhibitory cell types that form strong unilateral connections onto DNa02 (e.g., AOTU019).

The statement near Figure 5B that "DNa02 activity was higher on the side ipsilateral to the attractive stimulus, but contralateral to the aversive stimulus" is really important - and only possible to see because of the dual recordings.

We thank the reviewer for this positive feedback.

Reviewer #3 (Public review):

Summary:

Rayshubskiy et al. performed whole-cell recordings from descending neurons (DNs) of fruit flies to characterize their role in steering. Two DNs implicated in "walking control" and "steering control" by previous studies (Namiki et al., 2018, Cande et al., 2018, Chen et al., 2018) were chosen by the authors for further characterization. In-vivo whole-cell recordings from DNa01 and DNa02 showed that their activity predicts spontaneous ipsilateral turning events. The recordings also showed that while DNa02 predicts transient turns DNa01 predicts slow sustained turns. However, optogenetic activation or inactivation showed relatively subtle phenotypes for both neurons (consistent with data in other recent preprints, Yang et al 2023 and Feng et al 2024). The authors also further characterized DNa02 with respect to its inputs and showed a functional connection with olfactory and thermosensory inputs as well as with the head-direction system. DNa01 is not characterized to this extent.

Strengths:

(1) In-vivo recordings and especially dual recordings are extremely challenging in Drosophila and provide a much higher resolution DN characterization than other recent studies that have relied on behavior or calcium imaging. Especially impressive are the simultaneous recordings from bilateral DNs (Figure 3). These bilateral recordings show clearly that DNa02 cells not only fire more during ipsilateral turning events but that they get inhibited during contralateral turns. In line with this observation, the difference between left and right DNa02 neuronal activity is a much better predictor of turning events compared to individual DNa02 activity.

(2) Another technical feat in this work is driving local excitation in the head-direction neuronal ensemble

(PEN-1 neurons), while simultaneously imaging its activity and performing whole-cell recordings from DNa02

(Figure 4). This impressive approach provided a way to causally relate changes in the head-direction system to DNa02 activity. Indeed, DNa02 activity could predict the rate at which an artificially triggered bump in the PEN-1 ring attractor returns to its previous stable point.

(3) The authors also support the above observations with connectomics analysis and provide circuit motifs that can explain how the head direction system (as well as external olfactory/thermal stimuli) communicated with DNa02. All these results unequivocally put DNa02 as an essential DN in steering control, both during exploratory navigation as well as stimulus-directed turns.

We are grateful for this detailed positive feedback.

Weaknesses:

(1) I understand that the first version of this preprint was already on biorxiv in 2020, and some of the "weaknesses" I list are likely a reflection of the fact that I'm tasked to review this manuscript in late 2024 (more than 4 years later). But given this is a 2024 updated version it suffers from laying out the results in contemporary terms. For instance, the manuscript lacks any reference to the DNp09 circuit implicated in object-directed turning and upstream to DNa02 even though the authors cite one of the papers where this was analyzed (Braun et al, 2024). More importantly, these studies (both Braun et al 2024 and Sapkal et al 2024) along with recent work from the authors' lab (Yang et al 2023) and other labs (Feng et al 2024) provide a view that the entire suite of leg kinematics changes required for turning are orchestrated by populations of heterogeneous interconnected DNs. Moreover, these studies also show that this DN-DN network has some degree of hierarchy with some DNs being upstream to other DNs. In this contemporary view of steering control, DNa02 (like DNg13 from Yang et al 2023) is a downstream DN that is recruited by hierarchically upstream DNs like DNa03, DNp09, etc. In this view, DNa02 is likely to be involved in most turning events, but by itself unable to drive all the motor outputs required for the said events. This reasoning could be used while discussing the lack of major phenotypes with DNa02 activation or inactivation observed in the current study, which is in stark contrast to strong phenotypes observed in the case of hierarchically upstream DNs like DNp09 or DNa03. In the section, "Contributions of single descending neuron types to steering behavior": the authors start off by asking if individual DNs can make measurable contributions to steering behavior. Once more, any citations to DNp09 or DNa03 - two DNs that are clearly shown to drive strong turning-on activation (Bidaye et al, 2020, Feng et al 2024) - are lacking. Besides misleading the reader, such statements also digress the results away from contemporary knowledge in the field. I appreciate that the brief discussion in the section titled "Ensemble codes for steering" tries to cover these recent updates. However, I think this would serve a better purpose in the introduction and help guide the results.

We apologize for these omissions of relevant citations, which we have now fixed. Specifically, in our revised Discussion, we now point out that:

- Braun et al. (2024) reported that bilateral optogenetic activation of either DNa02 or DNa01 can drive turning (in either direction). 

- Braun et al. (2024) also identified DNb02 as a steering-related DN.

- Bidaye et al. (2020), Sapkal et al. (2024), and Braun et al. (2024) all contributed to the identification of DNp09 as a broadcaster DN with the capacity to promote ipsiversive turning.

We have also revised the beginning of the Results section titled “Contributions of single descending neuron types to steering behavior”, as suggested by the Reviewer.

Finally, we agree with the Reviewer’s overall point that steering is influenced by multiple DNs. We have not claimed that any DN is solely responsible for steering. As we note in the Discussion: “We found that optogenetically inhibiting DNa01 produced only small defects in steering, and inhibiting DNa02 did not produce statistically significant effects on steering; these results make sense if DNa02 is just one of many steering DNs.”

(2) The second major weakness is the lack of any immunohistochemistry (IHC) images quantifying the expression of the genetic tools used in these studies. Even though the main split-Gal4 tools for DNa01 and DNa02 were previously reported by Namiki et al, 2018, it is important to document the expression with the effectors used in this work and explicitly mention the expression in any ectopic neurons. Similarly, for any experiments where drivers were combined together (double recordings, functional connectivity) or modified for stochastic expression (Figure 8), IHC images are absolutely necessary. Without this evidence, it is difficult to trust many of the results (especially in the case of behavioral experiments in Figure 8). For example, the DNa01 genetic driver used by the authors is also expressed in some neurons in the nerve cord (as shown on the Flylight webpage of Janelia Research Campus). One wonders if all or part of the results described in Figure 8 are due to DNa01 manipulation or manipulation of the nerve cord neurons. The same applies for optic lobe neurons in the DNa02 driver.

This is a reasonable request. We used DN split-Gal4 lines to express three types of UAS-linked transgenes:

(1) GFP

In these flies, we know that expression in DNs is restricted to the DN types in question, based on published work (Namki et al., 2018), as well as the fact that we see one labeled DN soma per hemisphere. When we label both cells with GFP, we use the spike waveform to identify DNa02 and DNa01, as described in Figure S1

(2) ReaChR

In these flies, expression patterns were different in different flies because ReaChR expression was stochastically sparsened using hs-FLP. Expression was validated in each fly after the experiment, as described in the Methods (“Stochastic ReaChR expression”). hs-FLP-mediated sparsening will necessarily produce stochastic patterns of expression in both DNa02 and off-target cells, and this is true of all the flies in this experiment. What makes the “unilateral” flies distinct from the “bilateral” flies is that unilateral flies express ReaChR in one copy of DNa02, whereas bilateral flies express ReaChR in both copies of DNa02. On average, off-target expression will be the same in both groups.

(3) GtACR1

In these flies, we initially assumed that GtACR1 expression was the same as GFP expression under control of the same driver. However, we agree with the reviewer’s point that these two expression patterns are not necessarily identical. Therefore, to address the reviewer’s question, we performed immunofluorescence microscopy to characterize GtACR1 patterns in the brain and VNC of both genotypes. These expression patterns are now shown in a new supplemental figure (Figure S8). This figure shows that, as it happens, expression of GtACR1 is indeed indistinguishable from the GFP expression patterns for the same lines (archived on the FlyLight website). Both DN split-Gal4 lines are largely selective for the DNs in question, with limited off-target labeling. We have now drawn attention to this off-target labeling in the last paragraph of the Results, where the GtACR1 results are discussed.

(3) The paper starts off with a comparative analysis of the roles of DNa01 and DNa02 during steering. Unfortunately, after this initial analysis, DNa01 is largely ignored for further characterization (e.g. with respect to inputs, connectomics, etc.), only to return in the final figure for behavioral characterization where DNa01 seems to have a stronger silencing phenotype compared to DNa02. I couldn't find an explanation for this imbalance in the characterization of DNa01 versus DNa02. Is this due to technical reasons? Or was it an informed decision due to some results? In addition to being a biased characterization, this also results in the manuscript lacking a coherent thread, which in turn makes it a bit inaccessible to the non-specialist.

Yes, the first portion of the manuscript focuses on DNa01 and DNa02. The latter part of the manuscript transitions to focus mainly on DNa02. 

Our rationale is noted at the point in the manuscript where we make this transition, with the section titled “Steering toward internal goals”: “Having identified steering-related DNs, we proceeded to investigate the brain circuits that provide input to these DNs. Here we decided to focus on DNa02, as this cell’s activity is predictive of larger steering maneuvers.” When we say that DNa02 is predictive of larger steering maneuvers, we are referring to several specific results:

- We obtain larger filter amplitudes for DNa02 versus DNa01 (Fig. 2A-C). This means that, just after a unit change in DN firing rate, we see on average a larger change in steering velocity for DNa02 versus DNa01.

- The linear filter for DNa02 has a higher variance explained, as compared to DNa01 (Fig. 2D). This means that DNa02 is more predictive of steering.

- The relationship between firing rate and rotational velocity (150 ms later) is steeper for DNa02 than for DNa01 (Fig. 2G). This means that, if we ignore dynamics and we just regress firing rate against subsequent rotational velocity, we see a higher-gain relationship for DNa02.

Our focus on DNa02 was also driven by connectivity considerations. In the same paragraph (the first paragraph in the section titled “Steering toward internal goals”). We note that “there are strong anatomical pathways from the central complex to DNa02”; the same is not true of DNa01. This point has also been noted by other investigators (Hulse et al. 2021).

We don’t think this focus on DNa02 makes our work biased or inaccessible. Any study must balance breadth with depth. A useful general way to balance these constraints is to begin a study with a somewhat broader scope, and then narrow the study’s focus to obtain more in-depth information. Here, we began with comparative study of two cell types, and we progressed to the cell type that we found more compelling.

(4) There seems to be a discrepancy with regard to what is emphasized in the main text and what is shown in Figures S3/S4 in relation to the role of these DNs in backward walking. There are only two sentences in the main text where these figures are cited.

a) "DNa01 and DNa02 firing rate increases were not consistently followed by large changes in forward velocity

(Figs. 1G and S3)."

b) "We found that rotational velocity was consistently related to the difference in right-left firing rates (Fig. 3B). This relationship was essentially linear through its entire dynamic range, and was consistent across paired recordings (Fig. 3C). It was also consistent during backward walking, as well as forward walking (Fig. S4)." These main text sentences imply the role of the difference between left and right DNa02 in turning. However, the actual plots in the Figures S3 and S4 and their respective legends seem to imply a role in "backward walking". For instance, see this sentence from the legend of Figure S3 "When (ΔvoltageDNa02>>ΔvoltageDNa01), the fly is typically moving backward. When (firing rateDNa02>>firing rateDNa01), the fly is also often moving backward, but forward movement is still more common overall, and so the net effect is that forward velocity is small but still positive when (firing rateDNa02>>firing rateDNa01). Note that when we condition our analysis on behavior rather than neural activity, we do see that backward walking is associated with a large firing rate differential (Fig. S4)." This sort of discrepancy in what is emphasized in the text, versus what is emphasized in the figures, ends up confusing the reader. More importantly, I do not agree with any of these conclusions regarding the implication of backward walking. Both Figures S3 and S4 are riddled with caveats, misinterpretations, and small sample sizes. As a result, I actually support the authors' decision to not infer too much from these figures in the "main text". In fact, I would recommend going one step further and removing/modifying these figures to focus on the role of "rotational velocity". Please find my concerns about these two figures below:

a) In Figures S3 and S4, every heat map has a different scale for the same parameter: forward velocity. S3A is -10 to +10mm/s. S3B is -6 to +6 S4B (left) is -12 to +12 and S4B (right) is -4 to +4. Since the authors are trying to depict results based on the color-coding this is highly problematic.

b) Figure S3A legend "When (ΔvoltageDNa02>>ΔvoltageDNa01), the fly is typically moving backward." There are also several instances when ΔvoltageDNa02= ΔvoltageDNa01 and both are low (lower left quadrant) when the fly is typically moving backwards. So in my opinion, this figure in fact suggests DNa02 has no role in backward velocity control.

c) Based on the example traces in S4A, every time the fly walks backwards it is also turning. Based on this it is important to show absolute rotational velocity in Figure S4C. It could be that the fly is turning around the backward peak which would change the interpretation from Figure S4C. Also, it is important to note that the backward velocities in S4A are unprecedentedly high. No previous reports show flies walking backwards at such high velocities (for example see Chen et al 2018, Nat Comm. for backward walking velocities on a similar setup).

d) In my opinion, Figure S4D showing that right-left DNa02 correlates with rotational velocity, regardless of whether the fly is in a forward or backward walking state, is the only important and conclusive result in Figures S3/S4. These figures should be rearranged to only emphasize this panel.

We agree that it is difficult to interpret some of the correlations between DN activity and forward velocity, given that forward velocity and rotational velocity are themselves correlated to some degree. This is why we did not make claims based on these results in the main text. In response to these comments, we have taken the Reviewer’s suggestion to preserve Figure S4D (now Figure S3). The other components of these supplemental figures have been removed.

(5) Figure 3 shows a really nice analysis of the bilateral DNa02 recordings data. While Figure S5 [now Figure S4] shows that authors have a similar dataset for DNa01, a similar level analysis (Figures 3D, E) is not done for DNa01 data. Is there a reason why this is not done?

The reason we did not do the same analysis for DNa01 is that we only have two paired DNa01-DNa01 recordings. It turned out to be substantially more difficult to perform DNa01-DNa01 recordings, as compared to DNa02-DNa02 recordings. For this reason, we were not able to get more than two of these recordings.

(6) In Figure 4 since the authors have trials where bump-jump led to turning in the opposite direction to the DNa02 being recorded, I wonder if the authors could quantify hyperpolarization in DNa02 as is predicted from connectomics data in Figure 7.

We agree this is an interesting question. However, DNa02 firing rate and membrane potential are variable, and stimulus-evoked hyperpolarizations in these DNs tend to be relatively small (on the order of 1 mV, in the case of a contralateral fictive olfactory stimulus, Figure 5A). In the case of our fictive olfactory stimuli, we could look carefully for these hyperpolarizations because we had a very large number of trials, and we could align these trials precisely to stimulus onset. By contrast, for the bump-jump experiments, we have a more limited number of trials, and turning onset is not so tightly time-locked to the chemogenetic stimuli; for these reasons, we are hesitant to make claims about any bump-jump-related hyperpolarization in these trials.

(7) Figure 6 suggests that DNa02 contains information about latent steering drives. This is really interesting. However, in order to unequivocally claim this, a higher-resolution postural analysis might be needed. Especially given that DNa02 activation does not reliably evoke ipsilateral turning, these "latent" steering events could actually contain significant postural changes driven by DNa02 (making them "not latent"). Without this information, at least the authors need to explicitly mention this caveat.

This is a good point. We cannot exclude the possibility that DNa02 is driving postural changes when the fly is stopped, and these postural changes are so small we cannot detect them. In this case, however, there would still be an interesting mismatch between the stimulus-evoked change in DNa02 firing rate (which is large) and the stimulus-evoked postural response (which would be very small). We have added language to the relevant Results section in order to make this explicit.

(8) Figure 7 would really benefit from connectome data with synapse numbers (or weighted arrows) and a corresponding analysis of DNa01.

In response to this comment, we have added synapses number information (represented by weighted arrows) to Figures 7C, E, and F. We also added information to the Methods to explain how cells were chosen for inclusion in this diagram. (In brief: we thresholded these connections so as to discard connections with small numbers of synapses.)

We did perform an analogous connectome circuit analysis for DNa01, but if we use the same thresholds as we do for DNa02, we obtain a much sparser connectivity graph. We now show this in a new supplemental figure (Figure S9). MBON32 makes no monosynaptic connections onto DNa01, and it only forms one disynaptic connection, via LAL018, which is relatively weak. PFL3 and PFL2 make no mono- or disynaptic connections onto DNa01 comparable in strength to what we find for DNa02. 

The sparser connectivity graph for DNa01 is partly due to the fact that fewer cell types converge onto DNa01 as compared to DNa02 (110 cell types, versus 287 cell types). Also, it seems that DNa01 is simply less closely connected to the central complex and mushroom body, as compared to DNa02.

(9) In Figure 8E, the most obvious neuronal silencing phenotype is decreased sideways velocity in the case of DNa01 optogenetic silencing. In Figure S2, the inverse filter for sideways velocity for DNa01 had a higher amplitude than the rotational velocity filter. Taken together, does this point at some role for DNa01 in sideways velocity specifically?

No. The forward filters describe the average velocity impulse response, given a brief step change in firing rate.

Figure 1 and Figure S2 show that the sideways velocity forward filter is actually smaller for DNa01 than for DNa02. This means that a brief step change in DNa01 firing rate is followed by only a very small sideways velocity response. Conversely, the reverse filters describe the average firing rate impulse response, given a brief step change in sideways velocity. Figure S2 shows that the sideways velocity reverse filter is larger for DNa01 than for DNa02, but this means that the relationship between DNa01 activity and sideways velocity is so weak that we would need to see a very large neural response in order to get a brief step change in sideways velocity. In other words, the reverse filter says that DNa01 likely has very little role in determining sideways velocity.

(10) In Figure 8G, the effect on inner hind leg stance prolongation is very weak, and given the huge sample size, hard to interpret. Also, it is not clear how this fits with the role of DNa01 in slow sustained turning based on recordings.

Yes, this effect is small in magnitude, which is not too surprising, given that many DNs seem to be involved in the control of steering in walking. To clarify the interpretation of these phenotypes, we have added a paragraph to the end of the Results:

“All these effects are weak, and so they should be interpreted with caution. Also, both DN split-Gal4 lines drive expression in a few off-target cell types, which is another reason for caution (Fig. S8). However, they suggest that both DNs can lengthen the stance phase of the ipsilateral back leg, which would cause ipsiversive turning. These results are also compatible with a scenario where both DNs decrease the step length in the ipsilateral legs, which would also cause ipsiversive turning. Step frequency does not normally change asymmetrically during turning, so the observed decrease in step frequency during optogenetic inhibition may just be a by-product of increasing step length when these DNs are inhibited.” We have also added caveats and clarifications in a new Discussion paragraph:

“Our study does not fully answer the question of how these DNs affect leg kinematics, because we were not able to simultaneously measure DN activity and leg movement. However, our optogenetic experiments suggest that both DNs can lengthen the stance phase of the ipsilateral back leg (Fig. 8G), and/or  decrease the step length in the ipsilateral legs (Fig. 8H), either of which would cause ipsiversive turning. If these DNs have similar qualitative effects on leg kinematics, then why does DNa02 precede larger and more rapid steering events? This may be due to the fact that DNa02 receives stronger and more direct input from key steering circuits in the brain (Fig. S9). It may also relate to the fact that DNa02 has more direct connections onto motor neurons (Fig. 1B).”

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) I found the sign conventions for rotational velocity particularly confusing. Figure 3 represents clockwise rotations as +ve values, but Figure 4H represents anticlockwise rotations as positive values. But for EPG bumps, anticlockwise rotations are given negative values. Please make them consistent unless I am missing something obvious.

Different fields use different conventions for yaw velocity. In aeronautics, a clockwise turn is generally positive. In robotics and engineering of terrestrial vehicles, a counterclockwise turn is generally positive. Historically, most Drosophila studies that quantified rotational (yaw) velocity were focused on the behavior of flying flies, and these studies generally used the convention from aeronautics, where a clockwise turn is defined as a positive turn. When we began working in the field, we adopted this convention, in order to conform to previous literature. It might be argued that walking flies are more like robots than airplanes, but it seemed to us that it was confusing to have different conventions for different behaviors of the same animal. Thus, all of the published studies from our lab define clockwise rotation as having positive rotational velocity.

Figure 4 focuses on the role of the central complex in steering. As the fly turns clockwise (rightward), the bump of activity in EPG neurons normally moves counterclockwise around the ellipsoid body, as viewed from the posterior side (Turner-Evans et al., 2017). The posterior view is the conventional way to represent these dynamics, because (1) we and others typically image the brain from the posterior side, not the anterior side, and (2) in a posterior view, the animal’s left is on the left side of the image, and vice versa. We have added a sentence to the Figure 4A legend to clarify these points.

Previous work has shown that, when an experimenter artificially “jumps” the EPG bump, this causes the fly to make a compensatory turn that returns the bump to (approximately) its original location (Green et al., 2019). Our work supports this observation. Specifically, we find that clockwise bump jumps are generally followed by rightward turns (which drive the bump to return to its approximate original location via a counterclockwise path), and vice versa. This is noted in the Figure 4D legend. Note that Figure 4D plots the fly’s rotational velocity during the bump return, plotted against the initial bump jump. 

Figure 4H shows that clockwise (blue) bump returns were typically preceded by leftward turning, counter-clockwise (green) bump returns were preceded by rightward turning, as expected. This is detailed in the Figure 4H legend, and it is consistent with the coordinate frame described above.

(2) It would be helpful to have images of the DNa01 and DNa02 split lines used in this paper, considering this paper would most likely be used widely to describe the functions of these neurons. Similarly, images of their reconstructions would be a useful addition.

High-quality three-dimensional confocal stacks of all the driver lines used in our study are publicly available. We have added this information to the Methods (under “Fly husbandry and genotypes”). Confocal images of the full morphologies of DNa01 and DNa02 have been previously published (Namiki et al., 2018). Figure 1A is a schematic that is intended to provide a quick visual summary of this information.

EM reconstructions of DNa01 and DNa02 are publicly accessible in a whole-brain dataset (https://codex.flywire.ai/) and a whole-VNC dataset (https://neuprint.janelia.org/). Both datasets are referenced in our study. As these datasets are easy to search and browse via user-friendly web-based tools, we expect that interested readers will have no difficulty accessing the underlying datasets directly.

Reviewer #2 (Recommendations for the authors):

(1) The description of the activity of the DNs that they "PREDICT steering during walking". This is an interesting word choice. Not causes, not correlates with, not encodes... does that mean the activity always precedes the action? Does that mean when you see activity, you will get behavior? This is important for assessing whether the DN activity is a cause or an effect. It is good to be cautious but it might be worth expanding on exactly what kind of connection is implied to justify the use of the word 'predict'.

Conventionally, “predict” means “to indicate in advance”. We write that DNs “predict” certain features of behavior. We use this term because (1) these DNs correlate with certain features of behavior, and (2) changes in DN activity precede changes in behavior.

The notion that neurons can “predict” behavior is not original to our study. Whenever neuroscientists summarize the relationship between neural activity and behavior by fitting a mathematical model (which may be as simple as a linear regression), the fitted model can be said to represent a “prediction” of behavior. These models are evaluated by comparing their predictions with measured behaviors. A good model is predictive, but it also implies that the underlying neural signal is also predictive (Levenstein et al., 2023 Journal of Neuroscience 43: 1074-1088; DOI: 10.1523/JNEUROSCI.1179-22.2022). Here, prediction simply means correlation, without necessarily implying causation. We also use “prediction” to imply correlation.

We do not think the term “prediction” implies determinism. Meteorologists are said to predict the weather, but it is understood that their predictions are probabilistic, not deterministic. Certainly, we would not claim that there is a deterministic relationship between DN activity and behavior. Figure 2D shows that neither DN type can explain all the variance in the fly’s rotational or sideways velocity. At the same time, both DNs have significant predictive power.

We might equally say that these DNs “encode” behavior. We have chosen to use the word “predict” rather than “encode” because we do not think it is necessary to use the framework of symbolic communication in connection with these DNs.

We agree with the Reviewer that it is helpful to test whether any neuron that “predicts” a behavior might also “cause” this behavior. In Figure 8, we show that directly perturbing these DNs can indeed alter locomotor behavior, which suggests a causal role. Connectome analyses also suggest a causal role for these DNs in locomotor behavior (Figure 1B, see especially also Cheong et al., 2024).

At the same time, it is clear from our results that these DNs are not “command neurons” for turning: they do not deterministically cause turning. Therefore, to avoid misunderstanding, we have generally been careful to summarize the results of our perturbation experiments by avoiding the statement that “this DN causes this behavior”. Rather, we have generally tried to say that “this DN influences this behavior”, or “this DN promotes this behavior”.

(2) There is some concern about how the linear filter models were developed and then used to predict the relationship between firing rate and steering behavior: how exactly were the build and test data separated to avoid re-extracting the input? It reads like a self-fulfilling prophecy/tautology.

We used conventional cross-validation for model fitting and evaluation. We apologize that this was not made explicit in our original submission; this was due to an oversight on our part. To be clear: linear filters were computed using the data from the first 20% of a given experiment. We then convolved each cell’s firing rate estimate with the computed Neuron→Behavior filter (the “forward filter”) using the data from the final 80% of the experiment, in order to generate behavioral predictions. Thus, when a model has high variance explained, this is not attributable to overfitting: rather, it quantifies the bona fide predictive power of the model. We have added this information to the Methods (under “Data analysis - Linear filter analysis”).

(3) Type-O right above Figure 2 [now Figure 1E]: I assume spike rate fluctuations in DNa02 precede DNa01?

Fixed. Thank you for reading the manuscript carefully.

(4) The description of the other manuscripts about neural control of the steering as "follow-up" papers is a bit diminishing. They were likely independent works on a similar theme that happened afterwards, rather than deliberate extensions of this paper, so "subsequent" might be a more accurate description.

We apologize, as we did not intend this to be diminishing. Given this request, we have revised “follow-up” to “subsequent”.

(5) The idea that DNa02 is high-gain because it is more directly connected to motor neurons is a hypothesis and this should be made clear. We really don't know the functional consequences of the directness of a path or the number of synapses, and which circuits you compare to would change this. DNa02 may be a higher gain than DNa01, but what about relative to the other DNs that enter pre-motor regions? How do you handle a few synapses and several neurons in a common class? All of these connectivity-based deductions await functional tests - like yours! I think it is better to make this clear so readers don't assume a higher level of certainty than we have.

The Reviewer asks how we handled few-synapse connections, and how we combined neurons in the same class. We apologize for not making this explicit in our original submission. We have now added this information to the Methods. Briefly, to select cell types for inclusion in Figures 7C, we identified all individual cells postsynaptic to PFL3 and presynaptic to DNa02, discarding any unitary connections with <5 synapses. We then grouped unitary connections by cell type, and then summed all synapse numbers within each connection group (e.g., summing all synapses in all PFL3→LAL126 connections). We then discarded connection groups having <200 synapses or <1% of a cell type’s pre- or postsynaptic total. Reported connection weights are per hemisphere, i.e. half of the total within each connection group. For Figure 7F we did the same, but now discarding connection groups having <70 synapses or <0.4% of a cell type’s pre- or postsynaptic total. In Figure S9, we used the same procedures for analyzing connections onto DNa01. 

We agree that it is tricky to infer function from connectome data, and this applies to motor neuron connectivity. We bring up DN connectivity onto motor neurons in two places. First, in the Results, we note that “steering filters (i.e., rotational and sideways velocity filters) were larger for DNa02 (Fig. 2A,B). This means that an impulse change in firing rate predicts a larger change in steering for this neuron. In other words, this result suggests that DNa02 operates with higher gain. This may be related to the fact that DNa02 makes more direct output synapses onto motor neurons (Fig. 1B) [emphasis added].” We feel this is a relatively conservative statement.

Subsequently, in the Discussion, we ask, “why does DNa02 precede larger and more rapid steering events? This may be due to the fact that DNa02 receives stronger and more direct input from key steering circuits in the brain (Fig. S9). It may also relate to the fact that DNa02 has more direct connections onto motor neurons (Fig. 1B) [emphasis added].” Again, we feel this is a relatively conservative statement.

To be sure, none of the motor neurons postsynaptic to DNa02 actually receive most of their synaptic input from DNa02 (or indeed any DN), and this is typical of motor neurons controlling leg muscles. Rather, leg motor neurons tend to get most of their input from interneurons rather than motor neurons (Cheong et al. 2024). Available data suggests that the walking rhythm originates with intrinsic VNC central pattern generators, and the DNs that influence walking do so, in large part, by acting on VNC interneurons. These points have been detailed in recent connectome analyses (see especially Cheong et al. 2024).

We are reluctant to broaden the scope of our connectome analyses to include other DNs for comparison, because we think these analyses are most appropriate to full-central-nervous-system-(CNS)-connectomes (brain and VNC together), which are currently under construction. Without a full-CNS-connectome, many of the DN axons in the VNC cannot be identified. In the future, we expect that full-CNS-connectomes will allow a systematic comparison of the input and output connectivity of all DN types, and probably also the tentative identification of new steering DNs. Those future analyses should generate new hypotheses about the specializations of DNa02, DNa01, and other DNs. Our study aims to help lay a conceptual foundation for that future work.

(6) Given the emphasis on the DNa02 to Motor Neuron connectivity shown (Figure 1B) and multiple text mentions, could you include more analyses of which motor neurons are downstream and how these might be expected to affect leg movements? I would like to see the synapse numbers (Figure 1B) as well as the fraction of total output synapses. These additions would help understand the evidence for the "see-saw" model.

We agree this is interesting. In follow-up work from our lab (Yang et al., 2023), we describe the detailed VNC connectivity linking DNa02 to motor neurons. We refer the Reviewer specifically to Figure 7 of that study (https://www.cell.com/cell/fulltext/S0092-8674(24)00962-0).

We regret that the see-saw model was perhaps not clear in our original submission. Briefly, this model proposes that an increase in excitatory synaptic input to one DN (and/or a disinhibition of that DN) is often accompanied by an increase in inhibitory synaptic input to the contralateral DN. This model is motivated by connectome data on the brain inputs to DNa02 (Figure 7), along with our observation that excitation of one DN is often accompanied by inhibition of the contralateral DN (Figure 5). We have now added text to the Results in several places in order to clarify these points. 

This model specifically pertains to the brain inputs to DNs, comparing the downstream targets of these DNs in the VNC would not be a test of this hypothesis. The Reviewer may be asking to see whether there is any connectivity in the brain from one DN to its contralateral partner. We do not find connections of this sort, aside from multisynaptic connections that rely on very weak links (~10 synapses per connection). Figure 7 depicts a much stronger basis for this hypothesis, involving feedforward see-saw connections from PFL3 and MBON32. 

(7) The conclusions from the data in Figure 8 could be explained more clearly. These seem like small effect sizes on subtle differences in leg movements - maybe like what was seen in granular control by Moonwalker's circuits? Measuring joint angles or step parameters might help clarify, but a summary description would help the reader.

We agree that these results were not explained very well in our original submission. 

In our revised manuscript, we have added a new paragraph to the end of this Results section providing some summary and interpretation:

“All these effects are weak, and so they should be interpreted with caution. However, they suggest that both DNs can lengthen the stance phase of the ipsilateral back leg, which would promote ipsiversive turning. These results are also compatible with a scenario where both DNs decrease the step length in the ipsilateral legs, which would also promote ipsiversive turning. Step frequency does not normally change asymmetrically during turning, so the observed decrease in step frequency during optogenetic inhibition may just be a by-product of increasing step length when these DNs are inhibited.”

Moreover, in the Discussion, we have also added a new paragraph that synthesizes these results with other results in our study, while also noting the limitations of our study:

“Our study does not fully answer the question of how these DNs affect leg kinematics, because we were not able to simultaneously measure DN activity and leg movement. However, our optogenetic experiments suggest that both DNs can lengthen the stance phase of the ipsilateral back leg (Fig. 8G), and/or  decrease the step length in the ipsilateral legs (Fig. 8H), either of which would promote ipsiversive turning. If these DNs have similar qualitative effects on leg kinematics, then why does DNa02 precede larger and more rapid steering events? This may be due to the fact that DNa02 receives stronger and more direct input from key steering circuits in the brain (Fig. S9). It may also relate to the fact that DNa02 has more direct connections onto motor neurons (Fig. 1B).”

In Figure 8D-H, we measure step parameters in freely walking flies during acute optogenetic inhibition of DNa01 and DNa02. In experiments measuring neural activity in flies walking on a spherical treadmill, we did not have a way to measure step parameters. Subsequently, this methodology was developed by Yang et al. (2023) and results for DNa02 are described in that study. 

Reviewer #3 (Recommendations for the authors):

Minor Points:

(1) If space allows, actual membrane potential should be mentioned when raw recordings are shown (for example Figure 1D).

We have now added absolute membrane potential information to Figure 1d.

(2) Typo in the sentence "To address this issue directly, we looked closely at the timing of each cell's recruitment in our dual recordings, and found that spike rate fluctuations in DNa02 typically preceded the spike rate fluctuations in DNa02 (Fig. 2A)." The final word should be "DNa01".

Fixed. Thank you for reading the manuscript carefully.

(3) Figure 2A - although there aren't direct connections between a01 and a02 in the connectome, the authors never rule out functional connectivity between these two. Given a02 precedes a01, shouldn't this be addressed?

In the full brain FAFB data set, there are two disynaptic connections from DNa02 onto the ipsilateral copy of DNa01. One connection is via CB0556 (which is GABAergic), and the other is via LAL018 (which is cholinergic). The relevant DNa02 output connections are very weak: each DNa02→CB0556 connection consists of 11 synapses, whereas each DNa02→LAL018 connection consists of 10 synapses (on average). Conversely, each CB0556→DNa01 connection consists of 29 synapses, whereas  each LAL018→DNa01 connection consists of 64 synapses. In short, LAL018 is a nontrivial source of excitatory input to DNa01, but DNa02 is not positioned to exert much influence over LAL018, and the two disynaptic connections from DNa02 onto DNa01 also have the opposite sign. Thus, it seems unlikely that DNa02 is a major driver of DNa01 activity. At the same time, it is difficult to completely exclude this possibility, because we do not understand the logic of the very complicated premotor inputs to these DNs in the brain. Thus, we are hesitant to make a strong statement on this point.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

Summary:

Sammons, Masserini et al. examine the connectivity of different types of CA3 pyramidal cells ("thorny" and "athorny"), and how their connectivity putatively contributes to their relative timing in sharp-wave-like activity. First, using patch-clamp recordings, they characterize the degree of connectivity within and between athorny and thorny cells. Based upon these experimental results, they compute a synaptic product matrix, and use this to inform a computational model of CA3 activity. This model finds that this differential connectivity between these populations, augmented by two different types of inhibitory neurons, can account for the relative timing of activity observed in sharp waves in vivo.

We thank the reviewer for reading our manuscript, as well as for their nice summary and constructive comments

Strengths:

The patch-clamp experiments are exceptionally thorough and well done. These are very challenging experiments and the authors should be commended for their in-depth characterization of CA3 connectivity.

Thank you for the recognition of our efforts.

Weaknesses:

(1) The computational elements of this study feel underdeveloped. Whereas the authors do a thorough job experimentally characterizing connections between excitatory neurons, the inhibitory neurons used in the model seem to be effectivity "fit neurons" and appear to have been tuned to produce the emergent properties of CA3 sharp wave-like activity. Although I appreciate the goal was to implicate CA3 connectivity contributions to activity timing, a stronger relationship seems like it could be examined. For example, did the authors try to "break" their model? It would be informative if they attempted different synaptic product matrices (say, the juxtaposition of their experimental product matrix) and see whether experimentally-derived sequential activity could not be elicited. It seems as though this spirit of analysis was examined in Figure 4C, but only insofar as individual connectivity parameters were changed in isolation.

Including the two interneuron types (B and C) in the model is, on the one hand, necessary to align our modeling framework to the state-of-the-art model by Evangelista et al. (2020), which assumes that these populations act as switchers between an SPW and a non-SPW state, and on the other hand, less straightforward because the connectivity involving these interneurons is largely unknown.

For B cells, the primary criterion to set their connections to and from excitatory cells was to balance the effect of the strong recurrent excitation and to achieve a mid-range firing rate for each population during sharp wave events. Our new simulations (Figure 5B) show that the initial suppression of population T (resulting in the long delay) indeed depends in equal proportions on the outlined excitatory connections and on how strongly each excitatory population is targeted by the B interneurons. However, these simulations demonstrate that there is a broad, clearly distinct, region of the parameter space that supports a long delay between the peaks, rather than a marginal set of finetuned parameters. In addition, the simulations show that B interneurons optimally contribute to the suppression of T when they primarily target T (Fig. 5B, panels 3,7,11,12,13) rather than A (panels 4,8,9,10,11). On the contrary, as reported in the parameter table, and now also displayed graphically in the new Figure 4A (included above, with arrow sizes proportional to the synaptic product between the parameters determining the total strength of each connection), we assume B to target A less weakly than T (to make up for the higher excitability of population A). Therefore, the long delay between the peaks in our model emerges in spite of the interneuron connectivity, rather than because of it, and it is an effect of the asymmetric connectivity between the two excitatory populations, in particular the extremely low connection from A to T.

(2) Additional explanations of how parameters for interneurons were incorporated in the model would be very helpful. As it stands, it is difficult to understand the degree to which the parameters of these neurons are biologically constrained versus used as fit parameters to produce different time windows of activity in types of CA3 pyramidal cells.

Response included in point (1).

Reviewer #2 (Public Review):

Sharp wave ripples are transient oscillations occurring in the hippocampus that are thought to play an important role in organising temporal sequences during the reactivation of neuronal activity. This study addresses the mechanism by which these temporal sequences are generated in the CA3 region focusing on two different subtypes of pyramidal neurons, thorny and athorny. Using high-quality electrophysiological recordings from up to 8 pyramidal neurons at a time the authors measure the connectivity rates between these pyramidal cell subtypes in a large dataset of 348 cells. This is a significant achievement and provides important data. The most striking finding is how similar connection characteristics are between cell types. There are no differences in synaptic strength or failure rates and some small differences in connectivity rates and short-term plasticity. Using model simulations, the authors explore the implications of the differences in connectivity rates for the temporal specificity of pyramidal cell firing within sharp-wave ripple events. The simulations show that the experimentally observed connectivity rates may contribute to the previously observed temporal sequence of pyramidal cell firing during sharp wave ripples.

Thank you very much for your careful review of our manuscript and the overall positive assessment.

The conclusions drawn from the simulations are not experimentally tested so remain theoretical. In the simple network model, the authors include basket cell and anti-SWR interneurons but the connectivity of these cell types is not measured experimentally and variations in interneuron parameters may also influence temporal specificity of firing.

As variations in some of these parameters can indeed influence the temporal specificity of firing, we have now performed additional simulations, the results of which are in the new Figures 5 and S5. Please also see response to Reviewer 1, point 1.

In addition, the influence of short-term plasticity measured in their experiments is not tested in the model.

We have now included short-term synaptic depression in all the excitatory-to-excitatory synapses and compensated for the weakened recurrent excitation by scaling some of the other parameters. The results of re-running our simulations in this alternative version of the model are reported in Figure S3 and are qualitatively analogous to those in Figure 4.

Interestingly, the experimental data reveal a large variability in many of the measured parameters. This may strongly influence the firing of pyramidal cells during SWRs but it is not represented within the model which uses the averaged data.

We have now incorporated variability in the following simulation parameters: the strength and latency of the four excitatory-to-excitatory connections as well as the reversal potential and leak conductance of both types of pyramidal cells, assuming variabilities similar to those observed experimentally (see Materials and Methods for details). Upon a slight re-balancing of some inhibitory connection strengths, in order to achieve comparable firing rates, we found that this version of the model also supports the generation of sharp waves with two pyramidal components (Figure S4B), and is, thus, fully analogous to our basic model. Varying the excitatory connectivities as in the original simulations (cf. Figure 4C and Figure S4C) reveals that increasing the athorny-toathorny or decreasing the athorny-to-thorny connectivity still increases the delay between the peaks, although for some connectivity values the peak of the athorny population appears more spread out in time.

Reviewer #3 (Public Review):

Summary:

The hippocampal CA3 region is generally considered to be the primary site of initiation of sharp wave ripples-highly synchronous population events involved in learning and memory although the precise mechanism remains elusive. A recent study revealed that CA3 comprises two distinct pyramidal cell populations: thorny cells that receive mossy fiber input from the dentate gyrus, and athorny cells that do not. That study also showed that it is athorny cells in particular that play a key role in sharp wave initiation. In the present work, Sammons, Masserini, and colleagues expand on this by examining the connectivity probabilities among and between thorny and athorny cells. First, using whole-cell patch clamp recordings, they find an asymmetrical connectivity pattern, with athorny cells receiving the most synaptic connections from both athorny and thorny cells, and thorny cells receiving fewer. They then demonstrate in spiking neural network simulations how this asymmetrical connectivity may underlie the preferential role of athorny cells in sharp wave initiation.

Strengths:

The authors provide independent validation of some of the findings by Hunt et al. (2018) concerning the distinction between thorny and athorny pyramidal cells in CA3 and advance our understanding of their differential integration in CA3 microcircuits. The properties of excitatory connections among and between thorny and athorny cells described by the authors will be key in understanding CA3 functions including, but not limited to, sharp wave initiation.

As stated in the paper, the modeling results lend support to the idea that the increased excitatory connectivity towards athorny cells plays a key role in causing them to fire before thorny cells in sharp waves. More generally, the model adds to an expanding pool of models of sharp wave ripples which should prove useful in guiding and interpreting experimental research.

Thank you very much for your careful review of our manuscript and this positive assessment.

Weaknesses:

The mechanism by which athorny cells initiate sharp waves in the model is somewhat confusingly described. As far as I understood, random fluctuations in the activities of A and B neurons provide windows of opportunity for pyramidal cells to fire if they have additionally recovered from adaptive currents. Thorny and athorny pyramidal cells are then set in a winner-takes-all competition which is quickly won by the athorny cells. The main thesis of the paper seems to be that athorny cells win this competition because they receive more inputs both from themselves and from thorny cells, hence, the connectivity "underlies the sequential activation". However, it is also stated that athorny cells activate first due to their lower rheobase and steeper f-I curve, and it is also indicated in the methods that athorny (but not thorny) cells fire in bursts. It seems that it is primarily these features that make them fire first, something which apparently happens even when the A to A connectivity is set to 0albeit with a very small lag. Perhaps the authors could further clarify the differential role of single cell and network parameters in determining the sequential activation of athorny and thorny cells. Is the role of asymmetric excitatory connectivity only to enhance the initial intrinsic advantage of athorny cells? If so, could this advantage also be enhanced in other ways?

Thank you for the time invested in the review of our manuscript. We especially thank you for pointing out that the description of these dynamics was unclear: we have now improved it in the main text and we provide here an additional summary. As correctly highlighted by Reviewer 3, athorny neurons (A) are more excitable than thorny (T) ones due to single-neuron parameters: therefore, if there is a winner-takes-all competition, they are going to win it. Whether there is a competition in the first place, however, depends on the excitatory (and inhibitory) connections. In particular, we should distinguish two questions: does the activity of populations A and B (PV baskets), without adaptation (so at the beginning of the sharp wave) suppress T? And does the activity of populations T and B suppress A?

The four possible combinations can be appreciated, for example, in the new Figure 5A5. If A can suppress T, but T cannot suppress A (low A-to-T, high T-to-A, bottom right corner, like in the data), A “wins” and T fires later, after a long delay. If both A and T can suppress each other (both cross-connections are low, bottom left corner), we still get the same outcome: A wins because of its earlier and sharper onset (due to single-neuron parameters). If neither population can suppress the other (high cross-connections, top right corner), then there is no competition and the populations reach the peak approximately at the same time. Only in the case in which T can suppress A, but A cannot suppress T (low T-to-A, high A-to-T, top left corner, opposite to the data), then A “loses” the competition. However, since A neurons nevertheless display some early activity (again, due to the single neuron parameters), this scenario is not as clean as the reversed one: rather, A cells have an initial, small peak, then T neurons quickly take over and grow to their own peak, and then, depending on how strongly T neurons suppress A neurons, there may or may not be a second peak for the A neurons. This is the reason why, in the top left corner of Figure 5B, the statistics show either a long positive or long negative delay, depending on whether the first (small) or second (absent, for some parameters) peak of A is taken into account. In summary, the experimentally measured connectivity does not only enhance the initial intrinsic advantage of A cells, but sets up the competitive dynamics in the first place, which are crucial for the emergence of two distinct peaks, rather than a single peak involving both populations.

Although a clear effort has been made to constrain the model with biological data, too many degrees of freedom remain that allow the modeler to make arbitrary decisions. This is not a problem in itself, but perhaps the authors could explain more of their reasoning and expand upon the differences between their modeling choices and those of others. For example, what are the conceptual or practical advantages of using adaptation in pyramidal neurons as opposed to short-term synaptic plasticity as in the model by Hunt et al.?

It should be pointed out that the model by Hunt et al. features adaptation in pyramidal neurons as well, as the neuronal units employed are also adaptive-exponential integrate-and-fire. In an early stage of this project, we obtained from Hunt et al. the code for their model, and ascertained that adaptation is the main mechanism governing the alternations between the sharp-wave and the non-sharp-wave states, to the extent that fully removing short-term plasticity from their model does not have any significant impact on the network dynamics. Therefore, our choices are, in this regard, fully consistent with theirs. In order to confirm that synaptic depression does not significantly impact the dynamics also in our model, we now performed additional simulations (Figure S3), addressed in the main text (lines 149-151) and in the response to Reviewer 1, who expressed similar concerns.

Relatedly, what experimental observations could validate or falsify the proposed mechanisms?

As sharp wave generation in this model relies on disinhibitory dynamics (suppression of the anti-sharp-wave interneurons C), the model could be validated/falsified by proving/disproving that a class of interneurons with anti-sharp-wave features exists. In addition, the mechanism we proposed for the long delay between the peaks of the athorny and thorny activity requires at least some connectivity from athorny to basket and from basket to thorny neurons.

In the data by Hunt et al., thorny cells have a higher baseline (non-SPW) firing rate, and it is claimed that it is actually stochastic correlations in their firing that are amplified by athorny cells to initiate sharp waves. However, in the current model, the firing of both types of pyramidal cells outside of ripples appears to be essentially zero. Can the model handle more realistic firing rates as described by Hunt et al., or as produced by e.g., walking around an environment tiled with place cells, or would that trigger SPWs continuously?

When building this model, we aimed at having two clearly distinct states the network could alternate between, so we picked a rather polarized connectivity to and from the anti-sharp wave cells (C), resulting in polarized states. As a result, we obtain a low, although non-zero, activity of pyramidal neurons in non-SPW states (0.4 spikes/s for athorny and 0.2 spikes/s for thorny). These assumptions can be partially relaxed, for example in the original model by Evangelista et al. (2020), where the background firing rate of pyramidal cells is ~2 spikes/s. It should also be noted that, when walking in an environment tiled with place cells, the hippocampus is subject to additional extra-hippocampal inputs (e.g. from the medial septum, resulting in theta oscillations) and to neuromodulation, which can alter the network in various ways that we have not included in our model. However, our results are not in contradiction to transient SPW-like activity states initiated at a certain phase of the theta oscillation, when the inhibition is weakest.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

(1) The manuscript reads like it was intended as a short-form manuscript for another journal. The introduction and discussion in particular are very brief and would benefit from being expanded and providing a bigger picture for the reader.

We had originally aimed to submit in the eLife “short report” format. However, also thanks to the suggestion of Reviewer 1, we realized that our text would be better supported by extended introduction and discussion sections, as well as additional figures.

(2) Graphs would benefit from including all datapoints, where appropriate.

All datapoints have now been added to boxplots in the main figures and supplement.

(3) The panels of Figure 4 are laid out strangely, it may be worthwhile to adjust.

We thank the reviewer for this suggestion. We have now adjusted the layout of Figure 4 and believe it is now easier to follow.

Reviewer #2 (Recommendations For The Authors):

Useful points to address include:

(1) Explore within the model the effect of altering interneuron connectivity. Are there other factors that can influence temporal specificity within SWRs?

The effects of varying the connectivity to and from B interneurons (the ones which are SPWactive and therefore relevant for temporal specificity) have now been investigated in the new Figure 5B, in which such parameters were varied in pairs or combined with the two most relevant excitatoryto-excitatory connections.

(2) Implement the experimentally observed short-term plasticity in the model to determine how this influences temporal specificity.

All the findings in Figure 4 have now been replicated in the new Figure S3, in which excitatory-to-excitatory synapses feature synaptic depression.

(3) Consider if it is possible to incorporate observed experimental variability in the model and explore the implications.

All the findings in Figure 4 have now been replicated in the new Figure S4, in which heterogeneity has been introduced in multiple neuronal and synaptic parameters of thorny and athorny neurons.

(4) Include the co-connectivity rates in the data. Ie how many of the recorded neurons are reciprocally connected? Does this change the model simulations?

We have now added the rates of reciprocal connections that we observed into the main text (lines 8688). We found 2 pairs of reciprocally connected athorny neurons and 2 pairs of reciprocally connected thorny neurons. These rates of reciprocity were not statistically significant. We did not observe reciprocal connections in other paired neuron combinations (i.e. athorny-thorny or vice-versa). Coconnectivity does not have any effect on the model simulations, as the model includes thousands of neurons grouped in populations without specific sub-structures. It might, however, be more relevant if the excitatory populations were further subdivided in assemblies.

Reviewer #3 (Recommendations For The Authors):

(1) Specify which part of CA3 you are recording from.

We have added this information into our results section - we recorded from 20 cells in CA3a, 274 cells in CA3b and 54 cells in CA3c. This information can now be found in the text on lines 68-69.

(2) Comment on why you might observe a larger fraction of athorny cells than Hunt et al.

Hunt et al. cite a broad range for the fraction of athorny cells in their discussion (10-20%). It is unclear where these estimates originate from. In their study, Hunt et al. use the bursting and nonbursting phenotypes as proxies for athorny and thorny cells respectively, and report here numbers of 32 and 70 equating to 31% athorny and 69% thorny. This fraction of athorny cells is more or less in line with our own findings, albeit slightly lower (34% and 66%). However, we believe this difference falls within the range of experimental variability. One caveat is that our electrophysiological recordings likely represent a biased sample of cells. In particular, with multipatch recordings, placement of later electrodes is often restricted to the borders of the pyramidal layer so as not to disturb already patched cells. Thus, our recorded cells do not represent a fully random sample of CA3 pyramidal cells. We believe that, only once a reliable genetic marker for athorny cells has been established can the size of this cell population be properly estimated. Furthermore, the ratio of thorny and athorny cells varies along the proximal distal axis of the CA3 so differences in ratios seen between our study and Hunt et al. may arise from sampling differences along this axis.

(3) In Figure 3, Aiii (the cell fractions) could also be represented as a vector of two squares stacked one on top of the other, then you could add multiplication signs between Ai, Aii and Aiii, and an equal sign between Aiii and Aiv.

Thank you! We have implemented this very nice suggestion.

(4) In Figure 4A, it would be helpful to display the strength of the connections similar to how it is done in Figure 3B.

We thank the reviewer for this suggestion. We have now updated Fig 4A to include connection strengths.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

Summary:

Cognitive and brain development during the first two years of life is vast and determinant for later development. However, longitudinal infant studies are complicated and restricted to occidental high-income countries. This study uses fNIRS to investigate the developmental trajectories of functional connectivity networks in infants from a rural community in Gambia. In addition to resting-state data collected from 5 to 24 months, the authors collected growing measures from birth until 24 months and administrated an executive functioning task at 3 or 5 years old.

The results show left and right frontal-middle and right frontal-posterior negative connections at 5 months that increase with age (i.e., become less negative). Interestingly, contrary to previous findings in high-income countries, there was a decrease in frontal interhemispheric connectivity. Restricted growth during the first months of life was associated with stronger frontal interhemispheric connectivity and weaker right frontal-posterior connectivity at 24 months. Additionally, the study describes that some connectivity patterns related to better cognitive flexibility at pre-school age.

Strengths:

- The authors analyze data from 204 infants from a rural area of Gambia, already a big sample for most infant studies. The study might encourage more research on different underrepresented infant populations (i.e., infants not living in occidental high-income countries).

- The study shows that fNIRS is a feasible instrument to investigate cognitive development when access to fMRI is not possible or outside a lab setting.

- The fNIRS data preprocessing and analysis are well-planned, implemented, and carefully described. For example, the authors report how the choices in the parameters for the motion artifacts detection algorithm affect data rejection and show how connectivity stability varies with the length of the data segment to justify the threshold of at least 250 seconds free of artifacts for inclusion.

- The authors use proper statistical methods for analysis, considering the complexity of the dataset.

We thank the reviewer for highlighting the strengths of this work.

Weaknesses:

- No co-registration of the optodes is implemented. The authors checked for correct placement by looking at pictures taken during the testing session. However, head shape and size differences might affect the results, especially considering that the study involves infants from 5 months to 24 months and that the same fNIRS array was used at all ages.

The fNIRS array used in this work was co-registered onto age-appropriate MNI templates at every time point in a previous published work L. H. Collins-Jones, et al., Longitudinal infant fNIRS channel-space analyses are robust to variability parameters at the group-level: An image reconstruction investigation. Neuroimage 237, 118068 (2021). This is reference No. 68 in the manuscript.

As we mentioned in the section fNIRS preprocessing and data-analysis: ‘The sections were established via the 17 channels of each hemisphere which were grouped into front, middle and back (for a total of six regions) based on a previous co-registration of the BRIGHT fNIRS arrays onto age-appropriate templates’. The procedure mentioned by the reviewer, involving the examination of pictures showing the placement of headbands on participants, aimed to exclude infants with excessive cap displacement from further analysis.

- The authors regress the global signal to remove systemic physiological noise. While the authors also report the changes in connectivity without global signal regression, there are some critical differences. In particular, the apparent decrease in frontal inter-hemispheric connections is not present when global signal regression is omitted, even though it is present for deoxy-Hb. The authors use connectivity results obtained after applying global signal regression for further analysis. The choice of regressing the global signal is questionable since it has been shown to introduce anti-correlations in fMRI data (Murphy et al., 2009), and fNIRS in young infants does not seem to be highly affected by physiological noise (Emberson et al., 2016). Systemic physiological noise might change at different ages, which makes its remotion critical to investigate functional network development. However, global signal regression might also affect the data differently. The study would have benefited from having short separation channels to measure the systemic psychological component in the data.

The work of Emberson et. al (2016) mentioned by the reviewer highlights indeed the challenges of removing systemic changes from the infants’ haemodynamic signal with short-channel separation (SSC). In fact, even a SSC of 1 cm detected changes in the blood in the brain, therefore by regressing this signal from the recorded one, the authors removed both systemic changes AND haemodynamic signal. This paper from Emberson et. al (2016) is taken as a reference in the field to suggest that SSC might not be an ideal tool to remove systemic changes when collecting fNIRS data on young infants, as we did in this work.

We agree with the reviewer's observation that systemic physiological noise may vary with age and among infants. Therefore, for each infant at each age, we regressed the mean value calculated across all channels. This ensures that the regressed signal is not biased by averaged calculations at group levels.

We are aware of the criticisms directed towards global signal regression in the fMRI literature, although some other works showed anticorrelations in functional connectivity networks both with and without global signal regression (Chaia, 2012). Furthermore, Murphy himself revised his criticism on the use of global signal regression in functional connectivity analysis in one of his more recent works (Murphy et al, 2017). The fact that the decreased FC is significant in results from data pre-processed without global signal regression gives us confidence that this finding is statistically robust and not solely driven by this preprocessing choice in our pipeline.

An interesting study by Abdalmalak et al. (2022) demonstrated that failing to correct for systemic changes using any method is inappropriate when estimating FC with fNIRS, as it can lead to a high risk of elevated connectivity across the whole brain (see Figure 4 of the mentioned paper). Consequently, we strongly advocate for the implementation of global signal regression in our analysis pipeline as a fundamental step for accurate functional connectivity estimations.

References:

Emberson, L. L., Crosswhite, S. L., Goodwin, J. R., Berger, A. J., & Aslin, R. N. (2016). Isolating the effects of surface vasculature in infant neuroimaging using short-distance optical channels: a combination of local and global effects. Neurophotonics, 3(3), 031406-031406.

Chaia, X. J., Castañóna, A. N., Öngürb, D., & Whitfield-Gabrielia, S. (2012). Anticorrelations in resting state networks without global signal regression. NeuroImage, 59(2), 1420–1428. https://doi.org/10.1515/9783050076010-014

Murphy, K., & Fox, M. D. (2017). Towards a consensus regarding global signal regression for resting state functional connectivity MRI. NeuroImage, 154(November 2016), 169–173. https://doi.org/10.1016/j.neuroimage.2016.11.052

Abdalmalak, A., Novi, S. L., Kazazian, K., Norton, L., Benaglia, T., Slessarev, M., ... & Owen, A. M. (2022). Effects of systemic physiology on mapping resting-state networks using functional near-infrared spectroscopy. Frontiers in neuroscience, 16, 803297.

- I believe the authors bypass a fundamental point in their framing. When discussing the results, the authors compare the developmental trajectories of the infants tested in a rural area of Gambia with the trajectories reported in previous studies on infants growing in occidental high-income countries (likely in urban contexts) and attribute the differences to adverse effects (i.e., nutritional deficits). Differences in developmental trajectories might also derive from other environmental and cultural differences that do not necessarily lead to poor cognitive development.

We agree with the reviewer that other factors differing between low- and poor-resource settings might have an impact on FC trajectories. We therefore specified this in the discussion as follows: “We acknowledge that differences in FC could also be attributed to other environmental and cultural disparities between high-resource and low-resource settings, and future studies are needed to investigate this further” (line 238).

- While the study provides a solid description of the functional connectivity changes in the first two years of life at the group level, the evidence regarding the links between adverse situations, developmental trajectories, and later cognitive capacities is weaker. The authors find that early restricted growth predicts specific connectivity patterns at 24 months and that certain connectivity patterns at specific ages predict cognitive flexibility. However, the link between development trajectories (individual changes in connectivity) with growth and later cognitive capacities is missing. To address this question adequately, the study should have compared infants with different growing profiles or those who suffered or did not from undernutrition. However, as the authors discussed, they lacked statistical power.

We agree with the reviewer, and indeed we highlighted this as one of the main limitation of our work: “Even given the large sample in our study, we were underpowered to test for group comparisons between sets of infants with distinct undernutrition growth profiles, e.g., infants with early poor growth that later resolved and infants with standard growth early that had a poor growth later. We were also underpowered to test the associations between early growth and FC on clinically undernourished infants (defined as having DWLZ two standard deviations below the mean) (line 311, discussion section).

We believe this is an important point to consider for the field, as it addresses the sample size required for studies investigating brain development in clinically malnourished infants. We hope this will serve as a valuable reference for future studies in the field. For example, a new study led by Prof. Sophie Moore and other members of the BRIGHT team (INDiGO) is currently recruiting six-hundreds pregnant women with the aim of obtaining a broader distribution of infants’ growth measures (https://www.kcl.ac.uk/research/sophie-moore-research-group).

Reviewer #2 (Public Review):

Summary and strengths:

The article pertains to a topic of importance, specifically early life growth faltering, a marker of undernutrition, and how it influences brain functional connectivity and cognitive development. In addition, the data collection was laborious, and data preprocessing was quite rigorous to ensure data quality, utilizing cutting-edge preprocessing methods.

We thank the reviewer for highlighting the strengths of this work.

Weaknesses:

However, the subsequent analysis and explanations were not very thorough, which made some results and conclusions less convincing. For example, corrections for multiple tests need to be consistently maintained; if the results do not survive multiple corrections, they should not be discussed as significant results. Additionally, alternative plans for analysis strategies could be worth exploring, e.g., using ΔFC in addition to FC at a certain age. Lastly, some analysis plans lacked a strong theoretical foundation, such as the relationship between functional connectivity (FC) between certain ROIs and the development of cognitive flexibility.

Thus, as much as I admire the advanced analysis of connectivity that was conducted and the uniqueness of longitudinal fNIRS data from these samples (even the sheer effort to collect fNIRS longitudinally in a low-income country at such a scale!), I have reservations about the importance of this paper's contribution to the field in its present form. Major revisions are needed, in my opinion, to enhance the paper's quality. 

We acknowledge the reviewer’s concern regarding the reporting of results that do not survive multiple comparisons. However, considering the uniqueness of our dataset and the novelty of our work, we believe it is crucial to report all significant findings as well as hypothesis-generating findings that may not pass stringent significance thresholds. We have taken great care to transparently distinguish between results that survived multiple comparisons and those that did not in both the Results and Discussion sections, ensuring that readers are not misled. It is possible that future studies may replicate and further strengthen these associations. Therefore, by sharing these results with the research community, we provide valuable insights for future investigations.

The relationship between FC and cognitive flexibility (as well as the relationship between growth and FC) has been explored focusing on those FC that showed a significant change with age, as specified in the results sections: ‘To investigate the impact of early nutritional status on FC at 24 months, we used multiple regression with the infant growth trajectory [...] and FC at 24 months [...]. To maximise power, we considered only those FC that showed a statistically significant change with age’ (line 183) and ‘To investigate whether FC early in life predicted cognitive flexibility at preschool age, we used multiple regression of FC across the first two years of life against later cognitive flexibility in preschoolers at three and five years. As per the analysis above, we focused on only those FC that showed a statistically significant change with age’ (line 198).

We explored the possibility of investigating the relationship between changes in FC and changes in growth. However, the degrees of freedom in these analyses dropped dramatically (~25/30), thereby putting the significance and the meaning of the results at risk. We look forward to future longitudinal studies with less attrition across these time points to maintain the statistical power necessary to run such analyses.

Reviewer #3 (Public Review):

Summary:

This study aimed to investigate whether the development of functional connectivity (FC) is modulated by early physical growth and whether these might impact cognitive development in childhood. This question was investigated by studying a large group of infants (N=204) assessed in Gambia with fNIRS at 5 visits between 5 and 24 months of age. Given the complexity of data acquisition at these ages and following data processing, data could be analyzed for 53 to 97 infants per age group. FC was analyzed considering 6 ensembles of brain regions and thus 21 types of connections. Results suggested that: i) compared to previously studied groups, this group of Gambian infants have different FC trajectory, in particular with a change in frontal inter-hemispheric FC with age from positive to null values; ii) early physical growth, measured through weight-for-length z-scores from birth on, is associated with FC at 24 months. Some relationships were further observed between FC during the first two years and cognitive flexibility at 4-5 years of age, but results did not survive corrections for multiple comparisons.

Strengths:

The question investigated in this article is important for understanding the role of early growth and undernutrition on brain and behavioral development in infants and children. The longitudinal approach considered is highly relevant to investigate neurodevelopmental trajectories. Furthermore, this study targets a little-studied population from a low-/middle-income country, which was made possible by the use of fNIRS outside the lab environment. The collected dataset is thus impressive and it opens up a wide range of analytical possibilities.

We thank the reviewer for highlighting the strengths of this work.

Weaknesses:

- Analyzing such a huge amount of collected data at several ages is not an easy task to test developmental relationships between growth, FC, and behavioral capacities. In its present form, this study and the performed analyses lack clarity, unity and perhaps modeling, as it suggests that all possible associations were tested in an exploratory way without clear mechanistic hypotheses. Would it be possible to specify some hypotheses to reduce the number of tests performed? In particular, considering metrics at specific ages or changes in the metrics with age might allow us to test different hypotheses: the authors might clarify what they expect specifically for growth-FC-behaviour associations. Since some FC measures and changes might be related to one another, would it be reasonable to consider a dimensionality reduction approach (e.g., ICA) to select a few components for further correlation analyses?

We confirm that this work was motivated by a compelling theoretical question: whether neural mechanisms, specifically FC, can be influenced by early adversity, such as growth, and subsequently impact cognitive outcomes, such as cognitive flexibility. This aligns with the overarching goal of the BRIGHT project, established in 2015 (Lloyd-Fox, 2023). We believe this was evident throughout the manuscript in several instances, for example:

- “The goal of the study was to investigate early physical growth in infancy, developmental trajectories of brain FC across the first two years of life, and cognitive outcome at school age in a longitudinal cohort of infants and children from rural Gambia, an environment with high rates of maternal and child undernutrition. Specifically, we aimed to: (i) investigate whether differences in physical growth through the first two years of life are related to FC at 24 months, and (ii) investigate if trajectories of early FC have an impact on cognitive outcome at pre-school age in these children.” (page 4, introduction)

- “This study investigated how early adversity via undernutrition drives longitudinal changes in brain functional connectivity at five time points throughout the first two years of life and how these developmental trajectories are associated with cognitive flexibility at preschool age.” (page 6, discussion)

- We had a clear hypothesis regarding short-range connectivity decreasing with age and long-range connectivity increasing with age, as stated at the end of the introduction: We hypothesized that (i) long-range FC would increase and short-range FC would decrease throughout the first two years of life” (page 4, line 147). However, we were not able to formulate clear hypotheses about the localization of these connections due to the scarcity of previous studies conducted within this age range, particularly in low-resource settings. The ROI approach for analysis was chosen to mitigate this challenge by reducing the number of comparisons while still enabling us to estimate the developmental trajectories of all the connections from which we acquired data.

Regarding the use of dimensionality reduction approach, we have not considered the use of ICA in our analysis. These methods require selecting a fixed number of components to remove from all participants. However, due to the high variability of infant fNIRS data across the five timepoints, we considered it untenable to precisely determine the number of components to remove at the group level. Such a procedure carries the risk of over-cleaning the data for some participants while leaving noise in for others (Di Lorenzo, 2019). We also felt that using PCA in this initial study would be beyond the scope of the brain-region-specific hypotheses and would be more appropriate in a follow-up analysis of these important data.

References:

Lloyd-Fox, S., McCann, S., Milosavljevic, B., Katus, L., Blasi, A., Bulgarelli, C., Crespo-Llado, M., Ghillia, G., Fadera, T., Mbye, E., Mason, L., Njai, F., Njie, O., Perapoch-Amado, M., Rozhko, M., Sosseh, F., Saidykhan, M., Touray, E., Moore, S. E., … Team, and the B. S. (2023). The Brain Imaging for Global Health (BRIGHT) Study: Cohort Study Protocol. Gates Open Research, 7(126).

Di Lorenzo, R., Pirazzoli, L., Blasi, A., Bulgarelli, C., Hakuno, Y., Minagawa, Y., & Brigadoi, S. (2019). Recommendations for motion correction of infant fNIRS data applicable to multiple data sets and acquisition systems. NeuroImage, 200(April), 511–527.

- It seems that neurodevelopmental trajectories over the whole period (5-24 months) are little investigated, and considering more robust statistical analyses would be an important aspect to strengthen the results. The discussion mentions the potential use of structural equation modelling analyses, which would be a relevant way to better describe such complex data.

We appreciate the complexity of the dataset we are working with, which includes multiple measures and time points. Currently, our focus within the outputs from the BRIGHT project is on examining the relationship between selected measures. While this may not involve statistically advanced modelling at the moment, it is worth noting that most of the results presented in this work have survived correction for multiple comparisons, indicating their statistical robustness. We believe that more advanced statistical analyses are beyond the scope of this rich initial study. In the next phase of the project, known as BRIGHT IMPACT, our team is collaborating with statisticians and experts in statistical modelling to apply more sophisticated and advanced statistical techniques to the data.

- Given the number of analyses performed, only describing results that survive correction for multiple comparisons is required. Unifying the correction approach (FDR / Bonferroni) is also recommended. For the association between cognitive flexibility and FC, results are not significant, and one might wonder why FC at specific ages was considered rather than the change in FC with age. One of the relevant questions of such a study would be whether early growth and later cognitive flexibility are related through FC development, but testing this would require a mediation analysis that was not performed.

We acknowledge the reviewer’s concern regarding the reporting of results that do not survive multiple comparisons. However, considering the uniqueness of our dataset and the novelty of our work, we believe it is crucial to report all significant findings. We have taken great care to transparently distinguish between results that survived multiple comparisons and those that did not in both the Results and Discussion sections, ensuring that readers are not misled. It is possible that future studies may replicate and further strengthen these associations. Therefore, by sharing these results with the research community, we provide valuable insights for future investigations.

We did not perform a mediation analysis as i) ΔWLZ between birth and the subsequent time points positively predicted frontal interhemispheric FC at 24 months, ii) frontal interhemispheric FC at 18 months (and right fronto-posterior connectivity at 24 months) predicted cognitive flexibility at preschool age. Considering that the frontal interhemispheric FC at 24 months that was positively predicted by growth, did not significantly predicted cognitive outcome at preschool age, we did not perform mediation models.

The reviewer raised concerns about using different methods to correct for multiple comparisons throughout the work. Results showing changes in FC with age were Bonferroni corrected, while we used FDR correction for the regression analyses investigating the relationship between growth and FC, as well as FC and cognitive flexibility. Both methods have good control over Type I errors (false positives), but Bonferroni is very conservative, increasing the likelihood of Type II errors (false negatives). We considered Bonferroni an appropriate method for correcting results showing changes in FC with age, where we had a large sample with strong statistical power (i.e. linear mixed models with 132 participants who had at least 250 seconds of good data for 2 out of 5 visits). However, Bonferroni was too conservative for the regression analyses, with N between 57 and 78) (Acharya, 2014; Félix & Menezes, 2018; Narkevich et al., 2020; Narum, 2006; Olejnik et al., 1997).

References:

Acharya, A. (2014). A Complete Review of Controlling the FDR in a Multiple Comparison Problem Framework--The Benjamini-Hochberg Algorithm. ArXiv Preprint ArXiv:1406.7117.

Félix, V. B., & Menezes, A. F. B. (2018). Comparisons of ten corrections methods for t-test in multiple comparisons via Monte Carlo study. Electronic Journal of Applied Statistical Analysis, 11(1), 74–91.

Narkevich, A. N., Vinogradov, K. A., & Grjibovski, A. M. (2020). Multiple comparisons in biomedical research: the problem and its solutions. Ekologiya Cheloveka (Human Ecology), 27(10), 55–64.

Narum, S. R. (2006). Beyond Bonferroni: less conservative analyses for conservation genetics. Conservation Genetics, 7, 783–787.

Olejnik, S., Li, J., Supattathum, S., & Huberty, C. J. (1997). Multiple testing and statistical power with modified Bonferroni procedures. Journal of Educational and Behavioral Statistics, 22(4), 389–406.

- Growth is measured at different ages through different metrics. Justifying the use of weight-for-length z-scores would be welcome since weight-for-age z-scores might be a better marker of growth and possible undernutrition (this impacting potentially both weight and length). Showing the distributions of these z-scores at different ages would allow the reader to estimate the growth variability across infants.

We consistently used WLZ as the metric to measure growth throughout. Our analysis investigating the relationship between WLZ and growth included HCZ at 7/14 days to correct for head size at birth. When selecting the best growth measure for this paper, we opted for WLZ over WAZ, given extant evidence that infants in our sample are smaller and shorter compared to the reference WHO standard for the same age group (Nabwera et al., 2017). Therefore, using WLZ allows us to adjust each infant's weight for its own length.

References:

Nabwera, H. M., Fulford, A. J., Moore, S. E., & Prentice, A. M. (2017). Growth faltering in rural Gambian children after four decades of interventions: a retrospective cohort study. The Lancet Global Health, 5(2), e208–e216.

- Regarding FC, clarifications about the long-range vs short-range connections would be welcome, as well as drawing a summary of what is expected in terms of FC "typical" trajectory, for the different brain regions and connections, as a marker of typical development. For instance, the authors suggest that an increase in long-range connectivity vs a decrease in short-range is expected based on previous fNIRS studies. However anatomical studies of white matter growth and maturation would suggest the reverse pattern (short-range connections developing mostly after birth, contrarily to long-range connections prenatally).

We expected an increase in long-range functional connectivity with age, as discussed in the introduction:

- “Based on data from fMRI, current models hypothesize that FC patterns mature throughout early development (23–27), where in typically developing brains, adult-like networks emerge over the first years of life as long-range functional connections between pre-frontal, parietal, temporal, and occipital regions become stronger and more selective (28–31). This maturation in FC has been shown to be related to the cascading maturation of myelination and synaptogenesis (32, 33) - fundamental processes for healthy brain development (34)” (line 93, page 3, introduction);

- “Importantly, normative developmental patterns may be disrupted and even reversed in clinical conditions that impact development; e.g., increased short-range and reduced long-range FC have been observed in preterm infants (36) and in children with autism spectrum disorder (37, 38)” (line 103, page 3, introduction);

- “We hypothesized that (i) long-range FC would increase and short-range FC would decrease throughout the first two years of life” (line 147, page 4, introduction).

Since inferences about FC patterns recorded with fNIRS are highly limited by the number and locations of the optodes, it is challenging to make strong inferences about specific brain regions. Moreover, infant FC fNIRS studies are still limited, which is why we focused our inferences on long-range versus short-range connectivity, without specifically pinpointing particular brain regions.

Additionally, were unable to locate the works mentioned by the reviewer regarding an increase in short-range white matter connectivity immediately after birth. On the contrary, we found several studies documenting an increase in white-matter long-range connectivity after birth, which is consistent with the hypothesised increase in FC long-range connectivity, such as:

Yap, P. T., Fan, Y., Chen, Y., Gilmore, J. H., Lin, W., & Shen, D. (2011). Development trends of white matter connectivity in the first years of life. PloS one, 6(9), e24678.

Dubois, J., Dehaene-Lambertz, G., Kulikova, S., Poupon, C., Hüppi, P. S., & Hertz-Pannier, L. (2014). The early development of brain white matter: a review of imaging studies in fetuses, newborns and infants. Neuroscience, 276, 48-71.

Stephens, R. L., Langworthy, B. W., Short, S. J., Girault, J. B., Styner, M. A., & Gilmore, J. H. (2020). White matter development from birth to 6 years of age: a longitudinal study. Cerebral Cortex, 30(12), 6152-6168.

Hagmann, P., Sporns, O., Madan, N., Cammoun, L., Pienaar, R., Wedeen, V. J., ... & Grant, P. E. (2010). White matter maturation reshapes structural connectivity in the late developing human brain. Proceedings of the National Academy of Sciences, 107(44), 19067-19072.

Collin G, van den Heuvel MP. The ontogeny of the human connectome: development and dynamic changes of brain connectivity across the life span. Neuroscientist. 2013 Dec;19(6):616-28. doi: 10.1177/1073858413503712.

The authors test associations between FC and growth, but making sense of such modulation results is difficult without a clearer view of developmental changes per se (e.g., what does an early negative FC mean? Is it an increase in FC when the value gets close to 0? In particular, at 24m, it seems that most FC values are not significantly different from 0, Figure 2B). Observing positive vs negative association effects depending on age is quite puzzling. It is also questionable, for some correlation analyses with cognitive flexibility, to focus on FC that changes with age but to consider FC at a given age.

We thank the reviewer for bringing up this important point and understand that it requires some additional consideration. The negative FC values decreasing with age indicate that these regions go from being anti-correlated to becoming increasingly correlated. Hence, FC of these ROIs increased with age. The trajectory seems to suggest that this will keep increasing with age but of course further data need to be collected to assess this.

Unfortunately, when considering ΔFC to predict cognitive flexibility, the numbers of participants dropped significantly, with N=~15/20 infants per group of preschoolers, making it very challenging to interpret the results with meaningful statistical power.

- The manuscript uses inappropriate terms "to predict", "prediction" whereas the conducted analyses are not prediction analyses but correlational.

We thank the reviewer for giving us to opportunity to thoroughly revise the manuscript about this matter. In this work, we had clear hypotheses regarding which variables predicted which certain measures (such as growth predicting FC and FC predicting cognitive outcomes). Therefore, we performed regression analyses rather than correlational analyses to investigate these associations. Hence, we believe that using the term ‘predict and ‘prediction’ is appropriate

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

(1) In the introduction and discussion, the authors talk about the link between developmental trajectories and cognitive capacities, and undernutrition. However, they did not compare developmental trajectories but connectivity patterns at different ages with ΔWLZ and cognitive flexibility. I recommend that the authors rephrase the introduction and discussion.

We thank the reviewer for pointing out places requiring better clarity in the text. We made edits through the introduction to better match our investigations. In particular we changed:

- ‘our understanding of the relationships between early undernutrition, developmental trajectories of brain connectivity, and later cognitive outcomes is still very limited,’ to, ‘our understanding of the relationships between early undernutrition, brain connectivity, and later cognitive outcomes is still very limited’ (line 89, introduction);

- ‘(ii) investigate if trajectories of early FC have an impact on cognitive outcome at pre-school age in these children,’ to, ‘(ii) investigate if early FC has an impact on cognitive outcome at pre-school age in these children’ (line 137, introduction);

- ‘This study investigated how early adversity via undernutrition drives longitudinal changes in brain functional connectivity at five time points throughout the first two years of life and how these developmental trajectories are associated with cognitive flexibility at preschool age,’ to, ‘This study investigated how early adversity via undernutrition drives brain functional connectivity throughout the first two years of life and how these early functional connections are associated with cognitive flexibility at preschool age’ (line 215, discussion).

(2) Considering most research is done in occidental high-income countries, and this work is one of the few presenting research in another context, I think the authors should discuss in the manuscript that differences with previous studies might also be due to environmental and cultural differences. Since the study lacks the statistical power to perform a statistical analysis that directly establishes a link between developmental trajectories and restricted growth and cognitive flexibility, the authors cannot disentangle which differences are related to undernutrition and which might result from growing up in a different environment. I recommend that the authors avoid phrases like (lines 57-58): "We observed that early physical growth before the fifth month of life drove optimal developmental trajectories of FC..." or (lines 223-224) "...our cohort of Gambian infants exhibit atypical developmental trajectories of functional connectivity...".

We thank the reviewer for this observation, and we agree with the reviewer that other factors differing between low- and poor-resource settings might have an impact on FC trajectories. We therefore specified this in the discussion as follows: “We acknowledge that differences in FC could also be attributed to other environmental and cultural disparities between high-resource and low-resource settings, and future studies are needed to explore this further” (line 238). We revised the whole manuscript to reflect similar statements.

(3) To better interpret the results, it would be interesting to know if poor early growth predicts late cognitive flexibility in the tested sample and if the ΔWLZ distributions differ compared to a population in a high-income country where undernutrition is less frequent.

We explored the relationship between changes in growth and cognitive flexibility in the two preschooler group, but there were no significant associations.

Mean and SD values of WLZ are reported in Table 3. The values at every age are negative, indicating that the infants' weight-for-length is below the expected norm at all ages. To our knowledge, no other studies have assessed changes in growth in an infant sample with similar closely spaced age time points in high-income countries, making comparisons on growth changes challenging.

(4) It is unclear why WLZ at birth and HCZ at 7-14 days are included in the models. I imagine this is to ensure that differences are not due to growing restrictions before birth. It would be nice if the authors could explain this.

As the reviewer pointed out, HCZ at 7-14 days was included to ensure associations between growth and FC are not due to physical differences at birth. This case be considered as a 'baseline' measure for cerebral development, in the same way that WLZ at birth was used as a baseline for physical development. Therefore, we can more confidently  assume that the associations between growth and FC were specific to the impact of change in WLZ postnatally and not confounded by the size or maturity of the infant at birth. We specified this in the manuscript as follows: “These analyses were adjusted by WLZ at birth and HCZ at 7/14 days, to more confidently assume that the associations between growth and FC were specific to the impact of change in WLZ postnatally and not confounded by the size or maturity of the infant at birth” (line 520, statistical analysis section in the method section).

(5) Right frontal-posterior connections at 24 months negatively correlate with ΔWLZ. Thus, restricted growth results in stronger frontal-posterior connections at 24 months. However, the same connections at 24 months positively correlate with cognitive flexibility (stronger connections predict better cognitive flexibility). Do the authors have any interpretation of this? How could this relate to previous findings of the authors (Bulgarelli et al. 2020), showing first an increase and then a decrease in functional connectivity between frontal and parietal regions?

We acknowledge that interpreting the negative relationship between changes in growth and fronto-posterior FC at 24 months, alongside the positive association between the same connection and later cognitive flexibility, is challenging. We refrain from relating these findings to those published by Bulgarelli in 2020 due to differences in optode locations and because in that work the decrease in fronto-posterior FC was observed after 24 months (up to 36 months), whereas the endpoint in this study is right at 24 months.

(6) With the growth of the head, the frontal channels move to more temporal areas, right? Could this determine the decrease in frontal inter-hemisphere connections?

As shown in Nabwera (2017) head size does not increase that much in Gambian infants, or at least as expected by the WHO standard measures. We have added HCZ mean and SD values per age in Table 3.

Minor points

- HCZ is used in line 184 but not defined.

We thank the reviewer for spotting this, we have now specified HCZ at line 184 as follows: ‘head-circumference z-score (HCZ)’.

- Table SI2: NIRS not undertaken = the participant was assessed but did want or could not perform... I imagine there is a missing "not".

We thank the reviewer for spotting this, we have now modified the legend of Table SI2 as follows: ‘the participant was assessed but did not want or could not perform the NIRS assessments.’

- The authors should explain what weight-for-length is for those who are not familiar with it.

We have added an explanation of weight-for-length in the experimental design section, line 339 as follows: ‘We then tested for relationships between brain FC at age 24 months with measures of early growth, as indexed by changes in weight-for-length z-scores (reflecting body weight in proportion to attained growth in length) at one month of age, and at each of the four subsequent visits (details provided below).’

Reviewer #2 (Recommendations For The Authors):

(1) I am confused about the authors' interpretation that left and right front-middle and right front-back FC increased with age. It appears in Figure 2 that the negative FC among these ROIs should actually decrease with age. This means that as individuals grow older, the FC values between these regions and zero diminished, albeit starting with negative FC (anticorrelation values) in younger age groups.

Yes, the reviewer is correct. The negative values of the left and right front-middle and right front-back FC decreasing with age indicate that these regions go from being anti-correlated to becoming increasingly correlated. Hence, FC of these ROIs increased with age.

(2) Are these negative values mentioned above at 24 months still negative? Have t-tests been run to examine the differences from zero?

As suggested, we performed t-tests against zero for the mentioned FC at 24 months, and only the left and right fronto-middle FC are significantly different than zero (left fronto-middle FC: t(94) = 1.8, p = 0.036; right fronto-middle FC t(94) = 2.7, p = 0.003).

(3) With so many correlation analyses, have multiple comparisons been consistently controlled for? While I assume this was done according to the Methods section, could the authors clarify whether FDR adjustment was applied to all the p-values at once or to a group of p-values each time? I found the following way of reporting FDR-adjusted p-values quite informative, such as PFDR, 24 pairs of ROIs < 0.05.

We thank the reviewer for this insightful comment. P-values of regression analyses were FDR corrected per connection investigated, i.e. 21 possible ΔWLZ values per connection. We have specified this in the method section as follows: “To ensure statistical reliability, results from the regression analyses on each FC were corrected for multiple comparisons using false discovery rate (FDR)(Benjamini & Hochberg, 1995) per each connection investigated, i.e. 21 possible ΔWLZ values per each connection,” (page 12, Statistical Analyses section).

(4) Can early growth trajectories predict changes in FC? Why not use ΔWLZ to predict ΔFC?

Unfortunately, when considering ΔWLZ to predict ΔFC, the numbers of participants dropped significantly, with N=~30 infants, making it very challenging to interpret the results. We believe this emphasizes the importance of recruiting large samples when conducting longitudinal studies involving infants and employing multiple measures.

(5) I might have missed the rationale, but why weren't the growth changes after 5 months studied?

ΔWLZ between all time points were assessed as predictors of FC at 24 months. We have specified this at line 183 as follows: ‘we used multiple regression with the infant growth trajectory (delta weight for length z-score between all time points, DWLZ) and FC at 24 months’. As indicated in Table 2 and 3 the associations between ΔWLZ at all time points and FC at 24 months were tested, but only those with DWLZ calculated between birth and 1 month and the subsequent time points were significant. DWLZ between 5 months and the subsequent time points, DWLZ between 8 months and the subsequent time points, DWLZ between 12 months and the subsequent time points, DWLZ between 18 months and the subsequent time points did not significantly predict FC at 24 months. These are highlighted in Table 2 and Figure 3 in blue and marked as NS (non-significant).

(6) Once more, the advantage of longitudinal data is that it allows us to tap into developmental changes. Analyzing and predicting cognitive development based solely on FC values at a single age stage (i.e., 24 months) would overlook the benefits of a longitudinal design, which is regrettable. I suggest that the authors attempt to use ΔFC for predictions and observe the outcomes.

As mentioned to point (4) raised by the reviewer, unfortunately, when considering ΔWLZ to predict ΔFC, the numbers of participants dropped significantly, with N=~30 infants, making it very challenging to interpret the results. We believe this emphasizes the importance of recruiting large samples when conducting longitudinal studies involving infants and employing various measures.

(7) In the section "Early FC predicts cognitive flexibility at preschool age", the authors pointed out that "...,none of these survived FDR correction for multiple comparisons." However, the paper discussed the association between FC at 24 months of age and cognitive flexibility, as it was supported by the statistical analysis in the following sections. If FDR correction cannot be satisfied, I would rephrase the implication/conclusion of the results to suggest that early FC does not predict cognitive flexibility at preschool age.

We acknowledge the reviewer’s concern regarding the reporting of results that do not survive multiple comparisons. However, considering the uniqueness of our dataset and the novelty of our work, we believe it is crucial to report all significant findings, even those not passing multiple comparisons corrections, as they may motivate hypothesis-generation for future studies. We have taken great care to transparently distinguish between results that survived multiple comparisons and those that did not in both the Results and Discussion sections, ensuring that readers are not misled. It is possible that future studies may replicate and further support these associations. Therefore, by sharing these results with the research community, we provide valuable insights for future investigations.

Following the reviewer’ suggestion, we specified that results from regression analysis are significant but they did not survive multiple comparisons in the discussion as follows: ‘While our results are consistent with previous studies, we acknowledge that the significant association between early FC and later cognitive flexibility does not withstand multiple comparisons. Therefore, we encourage future studies that may replicate these findings with a larger sample. (line 290, discussion section).

(8) Have the authors assessed the impact of growth trajectories on cognitive flexibility?

We explored the relationship between changes in growth and cognitive flexibility in the two preschooler groups, but there were no significant associations.

(9) Are there no other cognitive or behavioural measures available? Cognitive flexibility is just one domain of cognitive development, and would the impact of undernutrition on cognitive development be domain-specific? There is a lack of theoretical support here. Why choose cognitive flexibility, and should the impact of undernutrition be domain-specific or domain-general?

We agree with the reviewer that in this work, we chose to focus on one specific cognitive outcome. While this does not imply that the impact of undernutrition is domain-specific, cognitive flexibility, being a core executive function, has been extensively studied in terms of its neural underpinnings using other neuroimaging modalities, especially fMRI (for example see Dajani, 2015; Uddin, 2021).

Moreover, other studies looking at the effect of adversity on cognitive outcomes focus on specific cognitive skills, such as working memory (Roberts, 2017), reading and arithmetic skills (Soni, 2021).

We did assess infants also with Mullen Scales of Early Learning (MSEL), although the cognitive flexibility task within the Early Years Toolbox has been specifically designed for preschoolers (Howard, 2015), and this set of tasks has recently been validated in our team in The Gambia (Milosavljevic, 2023).Future works from the BRIGHT team will investigate performance at the MSEL in relation to other variable of the project.

References:

D. R. Dajani, L. Q. Uddin, Demystifying cognitive flexibility: Implications for clinical and developmental neuroscience. Trends Neurosci. 38, 571–578 (2015).

L. Q. Uddin, Cognitive and behavioural flexibility: neural mechanisms and clinical considerations. Nat. Rev. Neurosci. 22, 167–179 (2021).

Roberts, S. B., Franceschini, M. A., Krauss, A., Lin, P. Y., de Sa, A. B., Có, R., ... & Muentener, P. (2017). A pilot randomized controlled trial of a new supplementary food designed to enhance cognitive performance during prevention and treatment of malnutrition in childhood. Current developments in nutrition, 1(11), e000885.

Soni, A., Fahey, N., Bhutta, Z. A., Li, W., Frazier, J. A., Moore Simas, T., ... & Allison, J. J. (2021). Early childhood undernutrition, preadolescent physical growth, and cognitive achievement in India: A population-based cohort study. PLoS Medicine, 18(10), e1003838.

Howard, S. J., & Melhuish, E. (2015). An Early Years Toolbox (EYT) for assessing early executive function, language, self-regulation, and social development: Validity, reliability, and preliminary norms. Journal of Psychoeducational Assessment, 35(3), 255-275.

Milosavljevic, B., Cook, C. J., Fadera, T., Ghillia, G., Howard, S. J., Makaula, H., ... & Lloyd‐Fox, S. (2023). Executive functioning skills and their environmental predictors among pre‐school aged children in South Africa and The Gambia. Developmental Science, e13407.

(10) I would review more previous fNIRS studies on infants if they exist (e.g., the work by S Lloyd-Fox, L Emberson, and others). These studies can help identify brain ROIs likely linked to undernutrition and cognitive flexibility. The current analysis methods lean towards exploratory research. This makes the paper more of a proof-of-concept report rather than a strongly theoretically-driven study.

We thank the reviewer for this important point. While we have reviewed existing fNIRS infant studies, there are no extant works that showed whether specific brain regions are related undernutrition. However, several fMRI studies assessed regions that do support cognitive flexibility, and we mentioned these in the manuscript (for example see Dajani, 2015; Uddin, 2021).

Other than the BRIGHT project, we are aware of two other projects that assessed the effect of undernutrition on brain development, assessing cognitive outcomes in poor-resource settings:

- the BEAN project in Bangladesh in which fNIRS data were recorded from the bilateral temporal cortex (i.e. Pirazzoli, 2022);

- a project in India in which fNIRS data were recorded from frontal, temporal and parietal cortex bilaterally (i.e. Delgado Reyes, 2020)

The brain regions recorded in these studies largely overlap with the brain regions we recorded from in this study.

Another aspect to consider is that infants underwent several fNIRS tasks as part of the BRIGHT project, focusing on social processing, deferred imitation, and habituation responses. Therefore, brain regions for data acquisition were chosen to maximize the likelihood of recording meaningful data for all tasks (Lloyd-Fox, 2023). To clarify the text, we specified this information in the methods section (line 383).

References:

D. R. Dajani, L. Q. Uddin, Demystifying cognitive flexibility: Implications for clinical and developmental neuroscience. Trends Neurosci. 38, 571–578 (2015).

Pirazzoli, L., Sullivan, E., Xie, W., Richards, J. E., Bulgarelli, C., Lloyd-Fox, S., ... & Nelson III, C. A. (2022). Association of psychosocial adversity and social information processing in children raised in a low-resource setting: an fNIRS study. Developmental Cognitive Neuroscience, 56, 101125.

Delgado Reyes, L., Wijeakumar, S., Magnotta, V. A., Forbes, S. H., & Spencer, J. P. (2020). The functional brain networks that underlie visual working memory in the first two years of life. NeuroImage, 219, Article 116971.

Lloyd-Fox, S., McCann, S., Milosavljevic, B., Katus, L., Blasi, A., Bulgarelli, C., Crespo-Llado, M., Ghillia, G., Fadera, T., Mbye, E., Mason, L., Njai, F., Njie, O., Perapoch-Amado, M., Rozhko, M., Sosseh, F., Saidykhan, M., Touray, E., Moore, S. E., … Team, and the B. S. (2023). The Brain Imaging for Global Health (BRIGHT) Study: Cohort Study Protocol. Gates Open Research, 7(126).

(11) Last but not least, in the paper, the authors mentioned that fNIRS offers better spatial resolution and anatomical specificity compared to EEG, thereby providing more precise and reliable localization of brain networks. While I partially agree with this perspective, it remains to be explored whether the current fNIRS analysis strategies indeed yield higher spatial resolution. It is hoped that the authors will delve deeper into this discussion in the paper.

The brain regions of focus were selected based on coregistration work previously conducted at each time point on the array used in this project (Collins-Jones, 2019). We deliberately avoided making claims about small brain regions, considering that head size might increase slightly less with age in The Gambia compared to Western countries (Nabwera, 2017) . However, we maintain that the conclusions drawn in this study offer higher brain-region specificity than could have been  identified with current common EEG methods alone.

References:

L. H. Collins-Jones, et al., Longitudinal infant fNIRS channel-space analyses are robust to variability parameters at the group-level: An image reconstruction investigation. Neuroimage 237, 118068 (2021).

Nabwera, H. M., Fulford, A. J., Moore, S. E., & Prentice, A. M. (2017). Growth faltering in rural Gambian children after four decades of interventions: a retrospective cohort study. The Lancet Global Health, 5(2), e208–e216.

Reviewer #3 (Recommendations For The Authors):

Introduction

- Among important developmental mechanisms to mention are the development of exuberant connections and the further selection/stabilization of the relevant ones according to environmental stimulation, vs the pruning of others.

We agree with the reviewer that the development of exuberant connections and subsequent pruning is a universal process of paramount importance during the first years of life. However, after revising our introduction, given the word limit of the journal, we maintained focus on neurodevelopment and early adversity.

Results

- Adding a few more information on the 6 sections and 21 connections would be welcome. In particular for within-section FC: how was this computed?

The 6 sections were created based on the co-registration of the array used in this study at each age in a previous published work L. H. Collins-Jones, et al., Longitudinal infant fNIRS channel-space analyses are robust to variability parameters at the group-level: An image reconstruction investigation. Neuroimage 237, 118068 (2021). This is reference No. 68 in the manuscript.

As we mentioned in the section fNIRS preprocessing and data-analysis: ‘The sections were established via the 17 channels of each hemisphere which were grouped into front, middle and back (for a total of six regions) based on a previous co-registration of the BRIGHT fNIRS arrays onto age-appropriate templates’.

The 21 connections were defined as all the possible links between the 6 regions, specifically: the interhemispheric homotopic connections (in orange in Figure SI1), which connect the same regions between hemispheres (i.e., front left with front right); the intrahemispheric connections (in green in Figure SI1), which correlate channels belonging to the same region; the fronto-posterior connections (in blue in Figure SI1), which link front and middle, middle and back, and front and back regions of the same hemisphere; and the crossing interhemispheric connections (non-homotopic interhemispheric, in yellow in Figure SI1), which link the front, middle, and back areas between left and right hemispheres. We added these specifications also in the legend of Figure SI1 for clarity.

- The denomination intrahemispheric vs fronto-posterior vs crossed connections is not clear. Maybe prefer intra-hemispheric vs inter-hemispheric homotopic vs inter-hemispheric non-homotopic (also in Figure SI1).

We appreciate the reviewer's suggestion regarding terminology. However, we believe that the term 'inter-hemispheric non-homotopic' could potentially refer to both connections within the same brain hemisphere from front to back and connections crossing between hemispheres, leading to increased confusion. Therefore, we have chosen not to include the term 'non-homotopic' and instead added 'homotopic' to 'interhemispheric' throughout the manuscript to emphasize that these functional connections occur between corresponding regions of the two hemispheres.

- with time -> with age.

We replaced “with time” with “with age” as suggested through the manuscript.

- The description of both HbO2 and HHb results overloads the main text: would it be relevant to present one of the two in Supplementary Information if the results are coherent?

We understand the reviewer’s concern regarding overloading the results section with reporting both chromophores. However, reporting results for both HbO and HHb is considered a crucial step for publications in the fNIRS field, as emphasized in recent formal guidance (Yücel et al., 2020). One of the strengths of fNIRS compared to fMRI is its ability to record from both chromophores, enabling a more precise characterization of brain activations and oscillations. Moreover, in FC studies like this one, ensuring that HbO and HHb results overlap is an important check that increases confidence in interpreting the findings.

References:

Yücel, M. A., von Lühmann, A., Scholkmann, F., Gervain, J., Dan, I., Ayaz, H., Boas, D., Cooper, R. J., Culver, J., Elwell, C. E., Eggebrecht, A. ., Franceschini, M. A., Grova, C., Homae, F., Lesage, F., Obrig, H., Tachtsidis, I., Tak, S., Tong, Y., … Wolf, M. (2020). Best Practices for fNIRS publications. Neurophotonics, 1–34. https://doi.org/10.1117/1.NPh.8.1.012101

- HCZ is not defined when first used.

We thank the reviewer for spotting this, we have now specified HCZ at line 184 as follows: ‘head-circumference z-score (HCZ)’.

- Choosing the analyzed measures to "maximize power" could be criticised.

We appreciate the reviewer’s concern. However, correlating all the FC values with all changes in growth would have raised an important issue for multiple comparisons. We therefore we made a priori decision to focus on investigating the relationship between changes in growth and those FC that showed a significant change with age, considering these as the most interesting ones from a developmental perspective in our sample.

Discussion

- I would recommend using the same order to synthesize results and further discuss them.

We agree with the reviewer that the suggested structure is optimal for a clear discussion section. We have indeed followed it, with each paragraph covering specific aspects:

- Recap of the study aims

- Results summary and discussion of developmental changes

- Results summary and discussion of the relationship between changes in growth and FC

- Results summary and discussion of the relationship between FC and cognitive flexibility

- Limitations

- Conclusion

Given the numerous results presented in this paper, we believe that readers will better digest them by first reading a summary of the results followed by their interpretations, rather than condensing all the interpretations together.

- Highlighting how "atypical" developmental trajectories are in Gambian infants would be welcome in the Results section. Other interpretations can be found than "The observed decrease in frontal inter-hemispheric FC with increasing age may be due to the exposure to early life undernutrition adversity".

We agree with the reviewer that other factors that differ between low- and high-resource settings might have an impact on FC trajectories. We therefore specified this in the discussion as follows: “We acknowledge that differences in FC could also be attributed to other environmental and cultural disparities between high-resource and low-resource settings, and future studies are needed to further investigate cultural, environmental, and genetic effects on brain FC” (line 238).

- Focusing on FC at 24m for the relationship with growth is questionable.

Correlating the FC values at 5 time points with all changes in growth would have raised an important issue for multiple comparisons. We therefore we made a decision a priori to focus on investigating the relationship between changes in growth and FC at 24 months as our final time point of data collection. We added this information in the methods section as follows: “To investigate the impact of undernutrition on FC development, we used DWLZ as independent variables in regression analyses on HbO2 (as the chromophore with the highest signal-to-noise ratio) FC at 24 months, our final time point of data collection” (line 517, method section).

- There is too much emphasis on the correlation between FC and cognitive flexibility, whereas results are not significant after correction for multiple comparisons.

Following the reviewer’ suggestion, we specified that results from regression analysis are significant but they did not survive multiple comparisons in the discussion as follows: While our results are consistent with previous studies, we acknowledge that the significant association between early FC and later cognitive flexibility does not withstand multiple comparisons. Therefore, we encourage future studies that may replicate these findings with a larger sample. (line 290, discussion section).

Methods

- I would recommend detailing how z-scores were computed in the paragraph "Anthropometric measures".

We specified how z-scores were computed in the statistical analysis section as follows: “Anthropometric measures were converted to age and sex adjusted z‐scores that are based on World Health Organization Child Growth Standards (93). Weight‐for‐Length (WLZ) and Head Circumference (HCZ) z-scores were computed” (line 509, method section). As transforming data is the first step of statistical analysis and is not directly related to data collection, we believe it is more appropriate to retain this description in the statistical analysis section.

- FC computation: the mention of "correlating the first and the last 250s" is not clear.

We specified this more clearly in the text as follows: We found that correlating the first and the last 250 seconds of valid data after pre-processing provided the highest percentage of infants with strong correlation between the first and the last portion of data (line 467).

- The manuscript mentions "age 3 years" for the younger preschoolers but ~48months rather corresponds to 4 years.

We revised the entire manuscript and the supplementary materials, but we could not find any instance in which preschoolers are referred with age in months rather than in years.

- Specify the number of children evaluated at 4 and 5 years. Is the test of cognitive flexibility normalized for age? If not, how were the 2 groups considered in the analyses? (age as a confounding factor).

We have added the number of children in the two preschooler groups as follows: younger preschoolers (age mean ± SD=47.96 ± 2.77 months, N=77) and older preschoolers (age mean ± SD=57.58 ± 2.11 months, N=84). (line 484).

The cognitive flexibility test was not normalized for age, as this task was specifically developed for preschoolers (Howard, 2015). As mentioned in ‘Cognitive flexibility at preschool age’ of the methods section, “data were collected in two ranges of preschool ages”, which guided our decision to perform regression analysis on the impact of FC on cognitive flexibility separately within these two age groups, rather than treating them as a single group of preschoolers.

References:

Howard, S. J., & Melhuish, E. (2015). An Early Years Toolbox (EYT) for assessing early executive function, language, self-regulation, and social development: Validity, reliability, and preliminary norms. Journal of Psychoeducational Assessment, 35(3), 255-275.

Figures and Tables

- Table 1 could highlight the significant results. It is not clear what the "baseline" results correspond to.

We have marked in bold the results that are statistically significant in Table 1. In the linear mixed model we performed, the first time point (i.e. 5 months) is chosen as ‘baseline’, i.e. the reference against which the other time points are compared to, and its statistical values refer to its significance against 0 (as it has been performed in Bulgarelli 2020).

- Figures 2 B and C seem redundant? What is SE vs SD?

We believe that both figures 2B and 2C are useful for the readers. While the first one shows the mean FC values at the group level, the second one highlights the individual variability of FC values (typical of infant neuroimaging data), which also why it is interesting to relate these measures to other variables of our dataset (i.e. growth and cognitive flexibility). Figure 2C also reports mean FC values per age, but these might be less visible considering that also one dot per infant is also plotted.

SE stands for standard error, and in the legend of the figure we specified this as follows: ‘Mean and standard error of the mean (SE)’. SD stands for standard deviation, and we have now specified this as follows: ‘mean ± standard deviation (SD)’ .

- Table 2: I would recommend removing results that don't survive corrections for multiple comparisons.

We acknowledge the reviewer’s concern regarding the reporting of results that do not survive multiple comparisons. However, considering the uniqueness of our dataset and the novelty of our work, we believe it is crucial to report all significant findings. We have taken great care to transparently distinguish between results that survived multiple comparisons and those that did not in both the Results and Discussion sections, ensuring that readers are not misled. It is possible that future studies may replicate and further strengthen these associations. Therefore, by sharing these results with the research community, we provide valuable insights for future investigations.

- Figure 3: the top is redundant with Table 2: to be merged? B: the statistical results might be shown in a Table.

We agree with the reviewer that the top part of Figure 3 and Table 2 report the same results. However, given the richness of these findings, we believe that the top part of Figure 3 serves as a useful summary for readers. Additionally, examining both the top and bottom parts of Figure 3 provides a comprehensive overview of the regression analysis conducted in this study.

- Figure SI6: Is it really a % in x-axis?

We thank the reviewer for spotting this typo, the percentage is relevant for the y-axis only. We removed the % symbol from ticks of the x-axis.

- Table SI1: the presented p-values don't seem to survive Bonferroni correction, contrary to what is written.

We thank the reviewer for spotting this mistake, we removed the reference to the Bonferroni correction for the p-values.

- Table SI2: For the proportion of children included in the analysis, maybe be precise that the proportion was computed based on the ones with acquired data. Maybe also add the proportion according to all children, to better show the high drop-out rate at certain ages?

We thank the reviewer for these useful suggestions. We have specified in the legend of the table how we calculated the proportion of infants included as follows: ‘The proportion of children included in the analysis was computed based on the infants with FC data’. We have also added a column in the table called ‘Inclusion rate (from the 204 infants recruited)’, following the reviewer’s suggestion. This will be a useful reference for future studies.

- A few typos should be corrected throughout the manuscript.

We thoroughly revised the main manuscript and the supplementary materials for typos.

Author response:

The following is the authors’ response to the original reviews

Reviewer #1 (Public review):

Summary:

Building on previous in vitro synaptic circuit work (Yamawaki et al., eLife 10, 2021), Piña Novo et al. utilize an in vivo optogenetic-electrophysiological approach to characterize sensory-evoked spiking activity in the mouse's forelimb primary somatosensory (S1) and motor (M1) areas. Using a combination of a novel "phototactile" somatosensory stimuli to the mouse's hand and simultaneous high-density linear array recordings in both S1 and M1, the authors report in awake mice that evoked cortical responses follow a triphasic peak-suppression-rebound pattern response. They also find that M1 responses are delayed and attenuated relative to S1. Further analysis revealed a 20-fold difference in subcortical versus corticocortical propagation speeds.

They also report that PV interneurons in S1 are strongly recruited by hand stimulation. Furthermore, they report that selective activation of PV cells can produce a suppression and rebound response similar to "phototactile" stimuli. Lastly, the authors demonstrate that silencing S1 through local PV cell activation reduces M1 response to hand stimulation, suggesting S1 may directly drive M1 responses.

Strengths:

The study was technically well done, with convincing results. The data presented are appropriately analyzed. The author's findings build on a growing body of both in vitro and in vivo work examining the synaptic circuits underlying the interactions between S1 and M1. The paper is well-written and illustrated. Overall, the study will be useful to those interested in forelimb S1-M1 interactions.

Weaknesses:

Although the results are clear and convincing, one weakness is that many results are consistent with previous studies in other sensorimotor systems, and thus not all that surprising. For example, the findings that sensory stimulation results in delayed and attenuated responses in M1 relative to S1 and that PV inhibitory cells in S1 are strongly recruited by sensory stimulation are not novel (e.g., Bruno et al., J Neurosci 22, 10966-10975, 2002; Swadlow, Philos Trans R Soc Lond B Biol Sci 357, 1717-1727, 2002; Gabernet et al., Neuron 48, 315-327, 2005; Cruikshank et al., Nat Neurosci 10, 462-468, 2007; Ferezou et al., Neuron 56, 907-923, 2007; Sreenivasan et al., Neuron 92, 1368-1382, 2016; Yu et al., Neuron 104, 412-427 e414, 2019). Furthermore, the observation that sensory processing in M1 depends upon activity in S1 is also not novel (e.g., Ferezou et al., Neuron 56, 907-923, 2007; Sreenivasan et al., Neuron 92, 1368-1382, 2016). The authors do a good job highlighting how their results are consistent with these previous studies.

We thank the reviewer for the close reading of the manuscript and the many constructive comments and critiques. As the reviewer notes, there have been many prior studies of related circuits in other sensorimotor systems forming an important context for our study and findings, as we have tried to highlight. We appreciate the suggestions for additional relevant articles to cite.

Perhaps a more significant weakness, in my opinion, was the missing analyses given the rich dataset collected. For example, why lump all responsive units and not break them down based on their depth? Given superficial and deep layers respond at different latencies and have different response magnitudes and durations to sensory stimuli (e.g., L2/3 is much more sparse) (e.g., Constantinople et al., Science 340, 1591-1594, 2013; Manita et al., Neuron 86, 1304-1316, 2015; Petersen, Nat Rev Neurosci 20, 533-546, 2019; Yu et al., Neuron 104, 412-427 e414, 2019), their conclusions could be biased toward more active layers (e.g., L4 and L5). These additional analyses could reveal interesting similarities or important differences, increasing the manuscript's impact. Given the authors use high-density linear arrays, they should have this data.

We have analyzed the activity patterns as a function of cortical depth, and now include these results in the manuscript as suggested. The key new finding is that the M1 responses are strongest in upper layers, consistent with expectations based on the excitatory corticocortical synaptic connectivity characterized previously. Changes to the manuscript include new figures (Figure 5; Figure 5 - figure supplement 1), which we explain (Methods: page 14, lines 618-621), describe (new Results section: pages 4-5, lines 183-189), comment on (Discussion: page 9, lines 378-391), and summarize the significance of (Abstract: page 1, lines 22-24). In addition, we incorporated the new laminar analysis into a summary schematic (Figure 9). We thank the reviewer for suggesting this analysis.

Similarly, why not isolate and compare PV versus non-PV units in M1? They did the photostimulation experiments and presumably have the data. Recent in vitro work suggests PV neurons in the upper layers (L2/3) of M1 are strongly recruited by S1 (e.g., Okoro et al., J Neurosci 42, 8095-8112, 2022; Martinetti et al., Cerebral cortex 32, 1932-1949, 2022). Does the author's data support these in vitro observations?

These experiments were relatively complex and M1 optotagging was not routinely included in the stimulus and acquisition protocol. Therefore, we don’t have sufficient data for this analysis. We plan to address this in future studies.

It would have also been interesting to suppress M1 while stimulating the hand to determine if any part of the S1 triphasic response depends on M1 feedback.

We agree that this is of interest but consider this to be outside the scope of the current study.

I appreciate the control experiment showing that optical hand stimulation did not evoke forelimb movement. However, this appears to be an N=1. How consistent was this result across animals, and how was this monitored in those animals? Can the authors say anything about digit movement?

We have performed additional experiments to address this point. A constraint with EMG is that it is limited to the muscle(s) one chooses to record from, and it is difficult to implant tiny muscles of the hand. Therefore, for this analysis, we used kilohertz videography as a high-sensitivity method for movement surveillance across the hand. Hand stimulation did not evoke any detectable movements. Changes in the manuscript include: revised Figure 1 - figure supplement 1; supplementary Figure 1 - video 1; and associated text edits in the Methods (page 13, line 557; page 14, lines 626-639) and Results sections (page 2, lines 84-85).

A light intensity of 5 mW was used to stimulate the hand, but it is unclear how or why the authors chose this intensity. Did S1 and M1 responses (e.g., amplitude and latency) change with lower or higher intensities? Was the triphasic response dependent on the intensity of the "phototactile" stimuli?

As we now say in the Methods > Optogenetic photostimulation of the hand section (page 13, lines 562-565), “This intensity was chosen based on pilot experiments in which we varied the LED power, which showed that this intensity was reliably above the threshold for evoking robust responses in both S1 and M1 without evoking any visually detectable movements (as subsequently confirmed by videography)”.

Reviewer #2 (Public review):

Summary:

Communication between sensory and motor cortices is likely to be important for many aspects of behavior, and in this study, the authors carefully analyse neuronal spiking activity in S1 and M1 evoked by peripheral paw stimulation finding clear evidence for sensory responses in both cortical regions

Strengths:

The experiments and data analyses appear to have been carefully carried out and clearly represented.

Weaknesses:

(1) Some studies have found evidence for excitatory projection neurons expressing PV and in particular some excitatory pyramidal cells can be labelled in PV-Cre mice. The authors might want to check if this is the case in their study, and if so, whether that might impact any conclusions.

Thank you for pointing this out. The prior studies suggest it is mainly a subset of layer 5B excitatory neurons that may express PV. We checked this in two ways. Anatomically, we did not find double-labeling. An electrophysiology assay showed that, although some evoked excitatory synaptic input could be detected in some neurons, these inputs were very weak. Results from these assays are shown in new Figure 6 - figure supplement 1, with associated text edits in the Methods (page 11, lines 469-471; page 15, lines 657-668) and Results (page 5, lines 198-199) sections.

(2) I think the analysis shown in Figure S1 apparently reporting the absence of movements evoked by the forepaw stimulation could be strengthened. It is unclear what is shown in the various panels. I would imagine that an average of many stimulus repetitions would be needed to indicate whether there is an evoked movement or not. This could also be state-dependent and perhaps more likely to happen early in a recording session. Videography could also be helpful.

As noted above, we have performed additional experiments to address this.

(3) Some similar aspects of the evoked responses, including triphasic dynamics, have been reported in whisker S1 and M1, and the authors might want to cite Sreenivasan et al., 2016.

Thank you for pointing this out; we now cite this article (page 1, line 46; page 10, line 415).

Reviewer #3 (Public review):

Summary:

This is a solid study of stimulus-evoked neural activity dynamics in the feedforward pathway from mouse hand/forelimb mechanoreceptor afferents to S1 and M1 cortex. The conclusions are generally well supported, and match expectations from previous studies of hand/forelimb circuits by this same group (Yamawaki et al., 2021), from the well-studied whisker tactile pathway to whisker S1 and M1, and from the corresponding pathway in primates. The study uses the novel approach of optogenetic stimulation of PV afferents in the periphery, which provides an impulselike volley of peripheral spikes, which is useful for studying feedforward circuit dynamics. These are primarily proprioceptors, so results could differ for specific mechanoreceptor populations, but this is a reasonable tool to probe basic circuit activation. Mice are awake but not engaged in a somatosensory task, which is sufficient for the study goals.

The main results are:

(1) brief peripheral activation drives brief sensory-evoked responses at ~ 15 ms latency in S1 and ~25 ms latency in M1, which is consistent with classical fast propagation on the subcortical pathway to S1, followed by slow propagation on the polysynaptic, non-myelinated pathway from S1 to M1;

(2) each peripheral impulse evokes a triphasic activation-suppression-rebound response in both S1 and M1;

(3) PV interneurons carry the major component of spike modulation for each of these phases; (4) activation of PV neurons in each area (M1 or S1) drives suppression and rebound both in the local area and in the other downstream area;

(5) peripheral-evoked neural activity in M1 is at least partially dependent on transmission through S1.

All conclusions are well-supported and reasonably interpreted. There are no major new findings that were not expected from standard models of somatosensory pathways or from prior work in the whisker system.

Strengths:

This is a well-conducted and analyzed study in which the findings are clearly presented. This will provide important baseline knowledge from which studies of more complex sensorimotor processing can build.

Weaknesses:

A few minor issues should be addressed to improve clarity of presentation and interpretation:

(1) It is critical for interpretation that the stimulus does not evoke a motor response, which could induce reafference-based activity that could drive, or mask, some of the triphasic response. Figure S1 shows that no motor response is evoked for one example session, but this would be stronger if results were analyzed over several mice.

As noted above, we have performed additional experiments to address this point.

(2) The recordings combine single and multi-units, which is fine for measures of response modulation, but not for absolute evoked firing rate, which is only interpretable for single units. For example, evoked firing rate in S1 could be higher than M1, if spike sorting were more difficult in S1, resulting in a higher fraction of multi-units relative to M1. Because of this, if reporting of absolute firing rates is an essential component of the paper, Figs 3D and 4E should be recalculated just for single units.

Thank you for noting this. Although the absolute firing rates are not essential for the main findings or conclusions (which as noted focus on response modulations and relative differences) we agree that analyzing the single-unit response amplitudes is useful. Therefore, changes in the manuscript now include: revised Figure 3, and associated text edits in the Methods (page 12, lines 543-545), Results (page 3, lines 115-119), and Discussion (page 7, lines 305-311) sections.

(3) In Figure 5B, the average light-evoked firing rate of PV neurons seems to come up before time 0, unlike the single-trial rasters above it. Presumably, this reflects binning for firing rate calculation. This should be corrected to avoid confusion.

Yes, this reflects the binning. We agree that this is potentially confusing and have removed these average plots below the raster plots, as the rasters alone suffice to demonstrate the result (i.e., that PV units are strongly activated and thus tagged by optogenetic stimulation). Changes are now reflected in revised Figure 6.

(4) In Figure 6A bottom, please clarify what legends "W. suppression" and "W. rebound" mean.

In the figure plot legends, the “W.” has been removed. Changes are now reflected in revised Figure 7 and Figure 7 – figure supplement 1.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) Did you filter the neural signals during acquisition? If so, please include these details in the results.

Signals were bandpass-filtered (2.5 Hz to 7.6 KHz) at the hardware level at acquisition (with no additional software filtering applied), as now clarified in the Methods Electrophysiological recordings section as requested (page 12, lines: 525-526).

Reviewer #2 (Recommendations for the authors):

(1) Some studies have found evidence for excitatory projection neurons expressing PV and in particular some excitatory pyramidal cells can be labelled in PV-Cre mice. The authors might want to check if this is the case in their study, and if so, whether that might impact any conclusions.

Please see above for our response to this issue.

(2) I think the analysis shown in Figure S1 apparently reporting the absence of movements evoked by the forepaw stimulation could be strengthened. It is unclear what is shown in the various panels. I would imagine that an average of many stimulus repetitions would be needed to indicate whether there is an evoked movement or not. This could also be state-dependent and perhaps more likely to happen early in a recording session. Videography could also be helpful.

Please see above for our response to this issue.

(3) Some similar aspects of the evoked responses, including triphasic dynamics, have been reported in whisker S1 and M1, and the authors might want to cite Sreenivasan et al., 2016.

As noted above, we now cite this study.

Author response:

The following is the authors’ response to the original reviews

Reviewer #1 (Public review):

Summary:

In this manuscript, the authors discovered MYL3 of marine medaka (Oryzias melastigma) as a novel NNV entry receptor, elucidating its facilitation of RGNNV entry into host cells through macropinocytosis, mediated by the IGF1R-Rac1/Cdc42 pathway.

Strengths:

In this manuscript, the authors have performed in vitro and in vivo experiments to prove that MnMYL3 may serve as a receptor for NNV via macropinocytosis pathway. These experiments with different methods include Co-IP, RNAi, pulldown, SPR, flow cytometry, immunofluorescence assays, and so on. In general, the results are clearly presented in the manuscript.

Weaknesses:

For the writing in the introduction and discussion sections, the author Yao et al mainly focus on the viral pathogens and fish in Aquaculture, the meaning and novelty of results provided in this manuscript are limited, and not broad in biology. The authors should improve the likely impact of their work on the viral infection field, maybe also in the evolutionary field with the fish model.

(1) Myosin is a big family, why did authors choose MYL3 as a candidate receptor for NNV?

We appreciate your insightful question. We selected MYL3 as a candidate receptor based on a combination of proteomic screening and literature evidence, and functional validation. Increasing evidence indicated that myosins have been implicated in viral infections. For instance, myosin heavy chain 9 plays a role in multiple viral infections (Li et al., 2018), and non-muscle myosin heavy chain IIA has been identified as an entry receptor for herpes simplex virus-1 (Arii et al., 2010). Furthermore, myosin II light chain activation is essential for influenza A virus entry via macropinocytosis (Banerjee et al., 2014). Our previous studies hinted at a potential interaction between MYL3 and CP (Zhang et al., 2020). Huang et al also reported that Epinephelus coioides MYL3 might interact with native NNV CP by proteomic analysis of immunoprecipitation (IP) assay (Huang et al., 2020). Our Co-IP and SPR analyses confirmed a direct interaction between MYL3 and the RGNNV CP. Based on these studies, we selected MYL3 as a candidate receptor for NNV.

References

Huang PY, Hsiao HC, Wang SW, Lo SF, Lu MW, Chen LL. 2020. Screening for the Proteins That Can Interact with Grouper Nervous Necrosis Virus Capsid Protein. Viruses 12:1–20.

Li L, Xue B, Sun W, Gu G, Hou G, Zhang L, Wu C, Zhao Q, Zhang Y, Zhang G, Hiscox JA, Nan Y, Zhou EM. 2018. Recombinant MYH9 protein C-terminal domain blocks porcine reproductive and respiratory syndrome virus internalization by direct interaction with viral glycoprotein 5. Antiviral Res 156:10–20.

Arii J, Goto H, Suenaga T, Oyama M, Kozuka-Hata H, Imai T, Minowa A, Akashi H, Arase H, Kawaoka Y, Kawaguchi Y. 2010. Non-muscle myosin IIA is a functional entry receptor for herpes simplex virus-1.

Banerjee I, Miyake Y, Philip Nobs S, Schneider C, Horvath P, Kopf M, Matthias P, Helenius A, Yamauchi Y. 2014. Influenza A virus uses the aggresome processing machinery for host cell entry. Science (80- ) 346:473–477.

(2) What is the relationship between MmMYL3 and MmHSP90ab1 and other known NNV receptors? Why does NNV have so many receptors? Which one is supposed to serve as the key entry receptor?

We acknowledge the functional diversity of receptors for NNV. MmHSP90ab1 and MmHSC70 have been identified as receptors involved in NNV entry through clathrin-mediated endocytosis (CME), whereas MYL3 facilitates entry via macropinocytosis. These pathways serve as complementary mechanisms for the virus to enter host cells, potentially enhancing infection efficiency. While HSP90ab1 facilitates CME, MYL3 promotes macropinocytosis, both of which are critical for viral internalization, but through distinct endocytic mechanisms.

NNV likely utilizes multiple receptors to increase its host range and infection efficiency. The diversity of receptors ensures that the virus can infect a wide variety of host species. By employing HSP90ab1, HSC70, and MYL3, NNV can exploit different cellular pathways for entry, making it more adaptable to various host environments.

Regarding the identification of a key entry receptor, we agree this is a critical unresolved question. While HSP90ab1/HSC70 appear essential for CME-mediated entry, our data suggest MYL3 plays a distinct role in macropinocytic uptake. To systematically evaluate receptor hierarchy, we initially proposed comparative knockout studies targeting these candidate genes. However, we must acknowledge that current technical limitations in marine fish models – particularly the extended generation time for stable knockout cell lines and challenges in maintaining viable cell cultures post-editing – have delayed these experiments. Nevertheless, we are actively exploring strategies to overcome these obstacles and will continue to refine our approach to address these questions in future research.

(3) In vivo knockout of MYL3 using CRISPR-Cas9 should be conducted to verify whether the absence of MYL3 really inhibits NNV infection. Although it might be difficult to do it in marine medaka as stated by the authors, the introduction of zebrafish is highly recommended, since it has already been reported that zebrafish could serve as a vertebrate model to study NNV (doi: 10.3389/fimmu.2022.863096).

As noted in our manuscript from line 374 to 384, marine medaka is a relatively new model for studying viral infections and is not yet optimized for CRISPR-Cas9-mediated gene knockout. The technical challenges related to precise embryo microinjection and off-target effects using CRISPR-Cas9 in marine medaka complicate the establishment of knockout lines. These limitations, including the time required for multiple breeding generations and molecular screening, currently make this approach difficult to implement.

We fully agree with your suggestion to consider zebrafish as an alternative model. Zebrafish have been well-established as a vertebrate model for studying NNV, and their genetic tractability and well-developed CRISPR-Cas9 protocols provide a more accessible and efficient platform for generating knockout models. In our future studies, we plan to conduct CRISPR-Cas9-mediated knockout experiments targeting multiple NNV receptors in zebrafish. This will allow us to systematically evaluate the role of different receptors in NNV infection and elucidate their potential interactions. The findings from these studies will be included in a future publication, which will provide a more comprehensive understanding of the molecular mechanisms underlying NNV infection in vertebrate models.

(4) The results shown in Figure 6 are not enough to support the conclusion that "RGNNV triggers macropinocytosis mediated by MmMYL3". Additional electron microscopy of macropinosomes (sizes, morphological characteristics, etc.) will be more direct evidence.

Previous study has reported that dragon grouper nervous necrosis virus (DGNNV) enters SSN-1 cells primarily through micropinocytosis and macropinocytosis pathways. Electron microscopy observations revealed several kinds of membrane ruffling and large disproportionate macropinosomes were observed in DGNNV infected cells, indicating NNV infection could triggers micropinocytosis (Liu et al., 2005). In our study, the data from inhibitor treatments, co-localization of MmMYL3 with RGNNV CP, and dextran uptake assays also provide compelling evidence for the involvement of macropinocytosis in RGNNV entry via MmMYL3. These methods are well-established in the literature and have been used extensively to study viral entry pathways (Lingemann et al., 2019). Specifically, the dextran uptake assay has been widely utilized as a marker for macropinocytosis and has provided clear evidence of RGNNV internalization via this pathway. The use of macropinocytosis inhibitors, such as EIPA and Rottlerin, significantly reduced RGNNV entry, further supporting our conclusion. Nonetheless, we acknowledge the potential value of additional electron microscopy studies and will consider this approach in our future research.

References

Liu W, Hsu CH, Hong YR, Wu SC, Wang CH, Wu YM, Chao CB, Lin CS. 2005. Early endocytosis pathways in SSN-1 cells infected by dragon grouper nervous necrosis virus, J Gen Virol.

Lingemann M, McCarty T, Liu X, Buchholz UJ, Surman S, Martin SE, Collins PL, Munir S. 2019. The alpha-1 subunit of the Na+,K+-ATPase (ATP1A1) is required for macropinocytic entry of respiratory syncytial virus (RSV) in human respiratory epithelial cells, PLoS Pathogens.

(5) MYL3 is "predominantly found in muscle tissues, particularly the heart and skeletal muscles". However, NNV is a virus that mainly causes necrosis of nervous tissues (brain and retina). If MYL3 really acts as a receptor for NNV, how does it balance this difference so that nervous tissues, rather than muscle tissues, have the highest viral titers?

While MYL3 is highly expressed in cardiac and skeletal muscles, studies have shown that MYL3, like other myosin light chains, can also be present in non-muscle tissues. Additionally, proteins involved in viral entry do not always need to be the most highly expressed in the final target tissue, as long as they facilitate the initial infection process. For instance, rabies virus is a rhabdovirus which exhibits a marked neuronotropism in infected animals. Transferrin receptor protein 1 can serve as a receptor for rabies virus through CME pathway, but TfR1 expressed most abundantly in liver tissue not nervous system (Wang et al., 2023).

Viral tropism is often determined not only by the presence of an entry receptor but also by co-receptors, cellular factors, and post-entry mechanisms. While MYL3 may act as a receptor for NNV, other factors, such as cell-specific proteases, signaling molecules, and intracellular trafficking pathways, likely contribute to NNV’s preferential replication in the brain and retina.

Reference

Wang Xinxin, Wen Z, Cao H, Luo J, Shuai L, Wang C, Ge J, Wang Xijun, Bu Z, Wang J. 2023. Transferrin Receptor Protein 1 Is an Entry Factor for Rabies Virus. J Virol 97. doi:10.1128/jvi.01612-22

Reviewer #2 (Public review):

Summary:

The manuscript offers an important contribution to the field of virology, especially concerning NNV entry mechanisms. The major strength of the study lies in the identification of MmMYL3 as a functional receptor for RGNNV and its role in macropinocytosis, mediated by the IGF1R-Rac1/Cdc42 signaling axis. This represents a significant advance in understanding NNV entry mechanisms beyond previously known receptors such as HSP90ab1 and HSC70. The data, supported by comprehensive in vitro and in vivo experiments, strongly justify the authors' claims about MYL3's role in NNV infection in marine medaka.

Strengths:

(1) The identification of MmMYL3 as a functional receptor for RGNNV is a significant contribution to the field. The study fills a crucial gap in understanding the molecular mechanisms governing NNV entry into host cells.

(2) The work highlights the involvement of IGF1R in macropinocytosis-mediated NNV entry and downstream Rac1/Cdc42 activation, thus providing a thorough mechanistic understanding of NNV internalization process. This could pave the way for further exploration of antiviral targets.

Thanks for your review.

Reviewer #3 (Public review):

Summary:

The manuscript presents a detailed study on the role of MmMYL3 in the viral entry of NNV, focusing on its function as a receptor that mediates viral internalization through the macropinocytosis pathway. The use of both in vitro assays (e.g., Co-IP, SPR, and GST pull-down) and in vivo experiments (such as infection assays in marine medaka) adds robustness to the evidence for MmMYL3 as a novel receptor for RGNNV. The findings have important implications for understanding NNV infection mechanisms, which could pave the way for new antiviral strategies in aquaculture.

Strengths:

The authors show that MmMYL3 directly binds the viral capsid protein, facilitates NNV entry via the IGF1R-Rac1/Cdc42 pathway, and can render otherwise resistant cells susceptible to infection. This multifaceted approach effectively demonstrates the central role of MmMYL3 in NNV entry.

Thanks for your review.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) Line94: SPR analysis? The full name should be provided when it first shows.

We have defined SPR when it first appears at line 97 in the revised manuscript.

(2) Moreover, is it too many for a manuscript to have a total of nine figures in the main text? Some of them might be moved to the supplementary file.

We have merged the previous Fig 4 and Fig 5 and combined Fig 8 and Fig 9, reducing the number of figures to seven. For the specific details of the figure adjustments, please refer to the corresponding figure legends.

Reviewer #2 (Recommendations for the authors):

(1) Expand on the potential therapeutic implications of targeting MYL3 or the IGF1R pathway in aquaculture settings. Including a discussion of how inhibitors could be developed or tested in future research would give practical context to the findings.

Thanks for your valuable suggestion to expand on the therapeutic implications of targeting MYL3 and the IGF1R pathway in aquaculture. In response, we have discussed potential strategies for developing inhibitors, such as small molecules, peptides, or monoclonal antibodies targeting MYL3 to block its interaction with the viral capsid, and IGF1R inhibitors to prevent macropinocytosis-mediated viral entry. We propose using virtual screening platforms to identify these inhibitors, followed by in vivo testing in aquaculture models. Additionally, combining MYL3 and IGF1R inhibitors could provide a synergistic approach to enhance antiviral efficacy. The relevant discussions have been supplemented at lines 358 to 368 in the revised manuscript.

(2) It is recommended to include the data regarding the lack of interaction between the CMNV CP and MmMYL3 as a supplementary figure.

We have included supplementary data demonstrating that CMNV CP does not interact with MmMYL3, highlighting the specificity of MYL3 for RGNNV. For detailed information, please refer to Fig. S4.

Reviewer #3 (Recommendations for the authors):

Consider discussing the broader implications of these findings, particularly whether MYL3 might serve as a receptor for other viruses.

We appreciate this suggestion. It is important to note that viral receptors typically exhibit specificity for specific types of viruses. Receptor recognition is typically highly specific, and the binding interactions between viral proteins and host receptors often depend on the structural compatibility between the viral capsid/ viral envelope and the host receptor. Our study demonstrates that MYL3 serves as a receptor for NNV based on its direct interaction with the NNV capsid protein (CP). However, when we tested whether MYL3 interacts with CMNV (Covert Mortality Nodavirus), which is phylogenetically closer to NNV, we found that CMNV CP does not bind to MYL3. Given the lack of interaction between MYL3 and CMNV, it is unlikely that MYL3 serves as a receptor for more distantly related viruses. Since MYL3 does not interact with CMNV, a virus more closely related to NNV, it is less likely to function as a receptor for viruses that are more distantly related to NNV. The relevant discussions have been supplemented at lines 306 to 310 in the revised manuscript.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

Diarrheal diseases represent an important public health issue. Among the many pathogens that contribute to this problem, Salmonella enterica serovar Typhimurium is an important one. Due to the rise in antimicrobial resistance and the problems associated with widespread antibiotic use, the discovery and development of new strategies to combat bacterial infections is urgently needed. The microbiome field is constantly providing us with various health-related properties elicited by the commensals that inhabit their mammalian hosts. Harnessing the potential of these commensals for knowledge about host-microbe interactions as well as useful properties with therapeutic implications will likely remain a fruitful field for decades to come. In this manuscript, Wang et al use various methods, encompassing classic microbiology, genomics, chemical biology, and immunology, to identify a potent probiotic strain that protects nematode and murine hosts from S. enterica infection. Additionally, authors identify gut metabolites that are correlated with protection, and show that a single metabolite can recapitulate the effects of probiotic administration.

We gratefully appreciate your positive and professional comments.

Strengths:

The utilization of varied methods by the authors, together with the impressive amount of data generated, to support the claims and conclusions made in the manuscript is a major strength of the work. Also, the ability to move beyond simple identification of the active probiotic, also identifying compounds that are at least partially responsible for the protective effects, is commendable.

We gratefully appreciate your positive and professional comments.

Weaknesses:

Although there is a sizeable amount of data reported in the manuscript, there seems to be a chronic issue of lack of details of how some experiments were performed. This is particularly true in the figure legends, which for the most part lack enough details to allow comprehension without constant return to the text. Additionally, 2 figures are missing. Figure 6 is a repetition of Figure 5, and Figure S4 is an identical replicate of Figure S3.

We gratefully appreciate your professional comments. Additional details to perform the related experiments had been added in Materials and methods section and figure legends (e.g., see Line 478-487, Line 996-1001, Line 1010-1012, Line 1019-1020, Line 1031-1033, Line 1041-1042, Line 1051-1053, Line 1082-1083, Line 1087-1088, Line 1093-1094, Line 1105-1107, Line 1113-1114,). Furthermore, we sincerely apologize for the mistakes and the inconvenience in the evaluating process of your review, and we have added the correct Figure 6 (see Line 1043-1053) and Figure S4 (see Line 1084-1088). We will carefully and thoroughly check the whole submitted manuscript along with supplementary information to avoid such mistakes in the future.

Reviewer #2 (Public review):

In this work, the investigators isolated one Lacticaseibacillus rhamnosus strain (P118), and determined this strain worked well against Salmonella Typhimurium infection. Then, further studies were performed to identify the mechanism of bacterial resistance, and a list of confirmatory assays was carried out to test the hypothesis.

We gratefully appreciate your positive and professional comments.

Strengths:

The authors provided details regarding all assays performed in this work, and this reviewer trusted that the conclusion in this manuscript is solid. I appreciate the efforts of the authors to perform different types of in vivo and in vitro studies to confirm the hypothesis.

We gratefully appreciate your positive and professional comments.

Weaknesses:

I have two main questions about this work.

(1) The authors provided the below information about the sources from which Lacticaseibacillus rhamnosus was isolated. More details are needed. What are the criteria to choose these samples? Where did these samples originate from? How many strains of bacteria were obtained from which types of samples?

Sorry for the ambiguous and limited information, more details had been added in Materials and methods section (see Line 480-496). We gratefully appreciate your professional comments.

Lines 486-488: Lactic acid bacteria (LAB) and Enterococcus strains were isolated from the fermented yoghurts collected from families in multiple cities of China and the intestinal contents from healthy piglets without pathogen infection and diarrhoea by our lab.

Sorry for the ambiguous and limited information, we had carefully revised this section and more details had been added in Materials and methods section (see Line 480-496). We gratefully appreciate your professional comments.

Lines 129-133: A total of 290 bacterial strains were isolated and identified from 32 samples of the fermented yoghurt and piglet rectal contents collected across diverse regions within China using MRS and BHI medium, which consist s of 63 Streptococcus strains, 158 Lactobacillus/ Lacticaseibacillus Limosilactobacillus strains, and 69 Enterococcus strains.

Sorry for the ambiguous information, we had carefully revised this section and more details had been added in this section (see Line 129-132). We gratefully appreciate your professional comments.

(2) As a probiotic, Lacticaseibacillus rhamnosus has been widely studied. In fact, there are many commercially available products, and Lacticaseibacillus rhamnosus is the main bacteria in these products. There are also ATCC type strains such as 53103.

I am sure the authors are also interested to know whether P118 is better as a probiotic candidate than other commercially available strains. Also, would the mechanism described for P118 apply to other Lacticaseibacillus rhamnosus strains?

It would be ideal if the authors could include one or two Lacticaseibacillus rhamnosus which are currently commercially used, or from the ATCC. Then, the authors can compare the efficacy and antibacterial mechanisms of their P118 with other strains. This would open the windows for future work.

We gratefully appreciate your professional comments and valuable suggestions. We deeply agree that it will be better and make more sense to include well-known/recognized/commercial probiotics as a positive control to comprehensively evaluate the isolated P118 strain as a probiotic candidate, particularly in comparison to other well-established probiotics, and also help assess whether the mechanisms described for P118 are applicable to other L. rhamnosus strains or lactic acid bacteria in general. Those issues will be fully taken into consideration and included in the further works.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) Line 28 - The sentence "with great probiotic properties" suggests that this strain was already known to have probiotic properties. Is that the case?

We gratefully appreciate your professional comments. This sentence "with great probiotic properties" in this context was intended as a summary of our findings, emphasizing that L. rhamnosus P118 exerts great probiotic properties after evaluating by traditional and C. elegans-infection screening strategies. We had revised this sentence (see Line27-30).

(2) Line 30 - What exactly do authors mean by "traditional"? They should add a bit more information here as to what these methods would be.

We gratefully appreciate your professional comments. By "traditional" methods, we refer to time-consuming and labor-intensive strategies for screening probiotic candidates with heavy works, which include bacterial isolation, culturing, phenotypic characterization, randomized controlled trials, and various in vitro and in vivo tests to assess probiotic properties (Sun et al., 2022). We had indicated this strategy in Line 91-94.

Reference:

Sun Y, Li HC, Zheng L, Li JZ, Hong Y, Liang PF, Kwok LY, Zuo YC, Zhang WY, Zhang HP. Iprobiotics: A machine learning platform for rapid identification of probiotic properties from whole-genome primary sequences. Briefings in Bioinformatics 2022;23.

(3) Line 37 - I believe "harmful microbes" is not the correct term here. I suggest authors use "potentially harmful".

Done as requested (see Line 36, 209, 212, 217, 381). We gratefully appreciate your valuable suggestions.

(4) Line 75 - What exactly do authors mean by "irregular dietary consumption"?

"irregular dietary consumption" means "irregular dietary habits" or " eating irregularly " or "abnormal eating behaviors". We had change to "irregular dietary habits" (see Line 76). We gratefully appreciate your professional comments.

(5) Line 85 - What exactly do authors mean by "without residues in raw food products"?

Here, "without residues in raw food products" means that probiotics barely remain in food animal products (e.g., meat, eggs, dairy) after dietary with probiotics in feeds by livestock and poultry. We gratefully appreciate your professional comments.

(6) Line 86 - Please, give a specific example of yeast.

Done as requested (see Line 85-86), “yeast (e.g., Saccharomyces boulardii, S. cerevisiae)”. We gratefully appreciate your valuable suggestions.

(7) Line 112 - Lactobacillus reuteri should be written out, since this is the first time the species name appears in the main text.

Done as requested (see Line 112). We gratefully appreciate your valuable suggestions.

(8) Lines 115-118 - Please, rewrite for clarity.

Done as requested (see Line 115-118). We gratefully appreciate your valuable suggestions.

(9) Line 118 -Lacticaseibacillus rhamnosus should be written out, since this is the first time the species name appears in the main text.

Done as requested (see Line 118). We gratefully appreciate your valuable suggestions.

(10) Line 119 - Throughout the text authors make it seem like strain P118 was previously known. Is that the case? If yes, how was it isolated again? This should be briefly mentioned in the introduction.

Sorry for the misunderstand caused by this statement, P118 strain was isolated and its probiotic properties were evaluated by our lab, not previously known, and we have revised this sentence (see Line 118-120). We gratefully appreciate your professional comments.

(11) Line 131 - How were strains identified?

Matrix-Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry (MALDI-TOF MS) method was employed to identify of bacterial species (He et al., 2022). This information was indicated in Materials and methods section (see Line 485-489). We gratefully appreciate your professional comments.

Reference

He D, Zeng W, Wang Y, Xing Y, Xiong K, Su N, Zhang C, Lu Y, Xing X. Isolation and characterization of novel peptides from fermented products of lactobacillus for ulcerative colitis prevention and treatment. Food Science and Human Wellness 2022;11:1464-74.

(12) Figure 1 - Legend needs a lot more info. Where are legends to panels PQ? Also, some of the text is too small to read.

Sorry for the limited info, we have revised Figure 1 legend and added more info (see Line 1000-1019), and we also provide vector graphic of Figure 1. We gratefully appreciate your professional comments.

(13) Line 136 - All strains were screened and 27 strains were positive, right?

Yes, all strains were screened and 27 strains were positive. We gratefully appreciate your professional comments.

(14) Figure 2 - What do authors mean by "spleen index" and "liver index"? This should be described in more detail. Also, p values for 'a', 'b', 'ab' should be given.

The organ index (spleen index, liver index) were calculated according to the formula: organ index = organ weight (g) / body weight (g) *1000, indicating in Materials and methods section (see Line 587-588). “Different lowercase letters ('a', 'b') indicate a significant difference (P < 0.05)” had been added in Line 1020-1029. We gratefully appreciate your professional comments.

(15) Line 212-214 - Again, I suggest authors use "potentially harmful" and "potentially beneficial".

Done as requested (see Line 36, 210, 213, 218, 383). We gratefully appreciate your valuable suggestions.

(16) Figure 3 - Which groups were tested in panels CD? Is this based on color? Legends should be restated in panels or clearly marked in the legend.

Sorry for this mistake, we have revised and added group info in Figure 3C-D (see Line 1013-1020). We gratefully appreciate your professional comments.

(17) Figure 4 - Lacks details.

Sorry for the mistakes, we have revised and added group info in Figure 4D-E and legend (see Line 1031-1037). We gratefully appreciate your professional comments.

(18) Figure 6 - This is a repetition of Figure 5.

Sorry for the mistakes, we have added the correct Figure 6 (see Line 1060-1070). We gratefully appreciate your professional comments.

(19) Lines 329-330 - C. elegans does not "mimic" animal intestinal physiology.

Sorry for the mistakes, we have revised this statement (see Line 139-142, 324-325). We gratefully appreciate your professional comments.

(20) Lines 358 and 418 - What do authors mean by "metabolic dysfunction" and "metabolic disorder"? I assume they mean changes in fecal metabolites. However, these are terms that may have different interpretations in the field of human metabolism. Therefore, I would suggest that the authors specify that they mean changes in fecal metabolite profiles when using these terms.

Sorry for the mistakes caused by this statement, we have revised this statement in the revised version (see Line 34-35, 122, 353-354, 413). We gratefully appreciate your professional comments.

(21) Line 475 - What do authors mean by "superficial effects"?

Sorry for the mistakes, we had change to “beneficial/protective effects” (see Line 469, Line 1074). We gratefully appreciate your professional comments.

(22) Line 486 - Were all yogurts artisanal? Where were piglets from? How were samples collected? Feces, rectal swabs? Does the ethics statement at the end of the manuscript also cover work with piglets?

Yes, all yogurts were artisanal. The 6 healthy piglet rectal content samples without pathogen infection and diarrhea were from a pig farm of Zhejiang province. Yes, the ethics statement at the end of the manuscript also cover the work with piglets.

(23) Line 490 - Which MALDI platform was used? The database used can have important implications for strain identification. What was the confidence of ID? This should be included.

Matrix-Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry (MALDI-TOF MS, Bruker Daltonik GmbH, Bremen, Germany) was employed to identify of bacterial species with a confidence level > 90%. This information was indicated in Materials and methods section (see Line 487-489). We gratefully appreciate your professional comments.

(24) Line 501 - Is this a widely used method to characterize probiotics? Please, add a reference.

Done as requested (see Line 498). Many probiotics or microbes can produce milk clotting enzyme to clot milk. It's an important measurement in the dairy industry, especially when making cheese (Zhang et al., 2023; Arbita et al., 2024; Shieh et al., 2009). The milk-clotting activity analysis is usually used for evaluating the potential ability of candidate probiotic isolates in clotting milk into cheeses.

Reference:

Zhang Y, Wang J, He J, Liu X, Sun J, Song X, Wu Y. Characteristics and application in cheese making of newly isolated milk-clotting enzyme from bacillus megaterium ly114. Food Res Int 2023;172:113202.

Arbita AA, Zhao J. Milk clotting enzymes from marine resources and their role in cheese-making: A mini review. Crit Rev Food Sci Nutr. 2024;64(27):10036-10047.

Chwen-Jen Shieh, Lan-Anh Phan Thi, Ing-Lung Shih. Milk-clotting enzymes produced by culture of Bacillus subtilis natto. Biochemical Engineering Journal. 2009;1(43): 85-91.

(25) Line 713 - How were fecal metabolites extracted?

Sorry for the missed information, the fecal metabolites extracted information had been added we have revised and added Materials and methods section (see Line 705-706). We gratefully appreciate your professional comments.

(26) Figure 7 - Please correct "macrophages".

Done as requested (see Figure 7, Line 1072). We gratefully appreciate your valuable suggestions.

(27) Table 1 - Should read "number of strains", not size.

Done as requested (see Line1084). We gratefully appreciate your valuable suggestions.

(28) Figure S1B - Is this data for P118?

Sorry for the mistakes, we have revised Figure S1 legend (see Line 1086-1088). We gratefully appreciate your professional comments.

(29) Figure S3 - Legends C, S, PS, P are not specified.

Sorry for the missed information, we have revised and added group info in Figure S3 legend (see Line 1095-1101). We gratefully appreciate your professional comments.

(30) Figure S3B - What is the "clinical symptom score"? How was this determined?

Sorry for the lack information, and the detailed information had been added in Materials and methods section (see Line 659-661, Table S7). We gratefully appreciate your professional comments.

(31) Figure S4 - This is an identical copy of Figure S3.

Sorry for the mistakes, we have added the correct Figure S4 (see Line 1103-1106). We gratefully appreciate your professional comments.

(32) Figure S5 - Legend lacks details.

Sorry for the missed information, we have revised and added group info in Figure S5 legend (see Line 1107-1112). We gratefully appreciate your professional comments.

(33) Figure S8 - What is "GM"? Since it inhibits growth to a greater extent than the highest metabolite concentration used, I imagine it must be an antibiotic (gentamycin?) as a positive control. This needs to be clearly stated.

Sorry for the missed information, GM: 100 μg/mL gentamicin (see Line 1134). We gratefully appreciate your professional comments.

(34) Figure S9 - Labels for panels are missing.

Sorry for the missed information, labels had been added (see Line 1135-1139). We gratefully appreciate your professional comments.

Reviewer #2 (Recommendations for the authors):

(1) This reviewer appreciates the efforts of the authors to provide the details related to this work. In the meantime, the manuscript shall be written in a way that is easy for the readers to follow.

We had tried our best to revise and make improve the whole manuscript to make it easy for the readers to follow (e.g., see Line 27-30, Line 115-120, Line 129-132, Line 480-496). We gratefully appreciate your valuable suggestions.

(2) For example, under the sections of Materials and Methods, there are 19 sub-titles. The authors could consider combining some sections, and/or citing other references for the standard procedures.

We gratefully appreciate your professional comments and valuable suggestions. Some sections had been combined according to the reviewer’s suggestions (see Line 497-530, Line 637-671).

(3) Another example: the figures have great resolution, but they are way too busy. Figures 1 and 2 have 14-18 panels. Figure 5 has 21 panels. Please consider separating into more figures, or condensing some panels.

We deeply agree with you that some submitted figures are way too busy, but it’s not easy to move some results into supplementary information sections, because all of them are essential for fully supporting our hypothesis and conclusions. Nonetheless, some panels had been combined or condensed according to the reviewer’s suggestions (see Line 1000-1020, Line 1052-1071). We gratefully appreciate your professional comments and valuable suggestions.

(4) Line 30: spell out "C." please.

Done as requested (see Line 31). We gratefully appreciate your valuable suggestions.

Author response:

The following is the authors’ response to the original reviews

eLife Assessment

This valuable work explores how synaptic activity encodes information during memory tasks. All reviewers agree that the quality of the work is high. Although experimental data do support the possibility that phospholipase diacylglycerol signaling and synaptotagmin 7 (Syt7) dynamically regulate the vesicle pool required for presynaptic release, concerns remain that the central finding of paired pulse depression at very short intervals was more likely caused by Ca²⁺ channel inactivation than pool depletion. Overall, this is a solid study with valuable findings, but the results warrant consideration of alternative interpretations.

We greatly appreciate invaluable and constructive comments from Editors and Reviewers. We also thank for their time and patience. We are pleased for our manuscript to have been assessed valuable and solid.

One of the most critical concerns was a possible involvement of Ca²⁺ channel inactivation in the strong paired pulse depression (PPD). Meanwhile, we have measured total (free plus buffered) calcium increments induced by each of first four APs in 40 Hz trains at axonal boutons of prelimbic layer 2/3 pyramidal cells. We found that first four Ca²⁺ increments were not different from one another, arguing against possible contribution of Ca²⁺ channel inactivation to PPD. Please see our reply to the 2nd issue in the Weakness section of Reviewer #3.

The second critical issue was on the definition of ‘vesicular probability’. Previously, vesicular probability (p_v) has been used with reference to the releasable vesicle pool which includes not only tightly docked vesicles but also reluctant vesicles. On the other hand, the meaning of p_v in the present study is the release probability of tightly docked vesicles. We clarified this point in our replies to the 1st issues in the Weakness sections of Reviewer #2 and Reviewer #3.

We below described our point-by-point replies to the Reviewers’ comments.

Public Reviews:

Reviewer #1 (Public review):

Shin et al. conduct extensive electrophysiological and behavioral experiments to study the mechanisms of short-term synaptic plasticity at excitatory synapses in layer 2/3 of the rat medial prefrontal cortex. The authors interestingly find that short-term facilitation is driven by progressive overfilling of the readily releasable pool, and that this process is mediated by phospholipase C/diacylglycerol signaling and synaptotagmin-7 (Syt7). Specifically, knockdown of Syt7 not only abolishes the refilling rate of vesicles with high fusion probability, but it also impairs the acquisition of trace fear memory. Overall, the authors offer novel insight to the field of synaptic plasticity through well-designed experiments that incorporate a range of techniques.

Reviewer #2 (Public review):

Summary:

Shin et al aim to identify in a very extensive piece of work a mechanism that contributes to dynamic regulation of synaptic output in the rat cortex at the second time scale. This mechanism is related to a new powerful model is well versed to test if the pool of SV ready for fusion is dynamically scaled to adjust supply demand aspects. The methods applied are state-of-the-art and both address quantitative aspects with high signal to noise. In addition, the authors examine both excitatory output onto glutamatergic and GABAergic neurons, which provides important information on how general the observed signals are in neural networks, The results are compellingly clear and show that pool regulation may be predominantly responsible. Their results suggests that a regulation of release probability, the alternative contender for regulation, is unlikely to be involved in the observed short term plasticity behavior (but see below). Besides providing a clear analysis pof the underlying physiology, they test two molecular contenders for the observed mechanism by showing that loss of Synaptotagmin7 function and the role of the Ca dependent phospholipase activity seems critical for the short term plasticity behavior. The authors go on to test the in vivo role of the mechanism by modulating Syt7 function and examining working memory tasks as well as overall changes in network activity using immediate early gene activity. Finally, they model their data, providing strong support for their interpretation of TS pool occupancy regulation.

Strengths:

This is a very thorough study, addressing the research question from many different angles and the experimental execution is superb. The impact of the work is high, as it applies recent models of short term plasticity behavior to in vivo circuits further providing insights how synapses provide dynamic control to enable working memory related behavior through nonpermanent changes in synaptic output.

Weaknesses:

(1) While this work is carefully examined and the results are presented and discussed in a detailed manner, the reviewer is still not fully convinced that regulation of release provability is not a putative contributor to the observed behavior. No additional work is needed but in the moment I am not convinced that changes in release probability are not in play. One solution may be to extend the discussion of changes in release probability as an alternative.

Quantal content (m) depends on n * p_v, where n = RRP size and p_v =vesicular release probability. The value for p_v critically depends on the definition of RRP size. Recent studies revealed that docked vesicles have differential priming states: loosely or tightly docked state (LS or TS, respectively). Because the RRP size estimated by hypertonic solution or long presynaptic depolarization is larger than that by back extrapolation of a cumulative EPSC plot (Moulder & Mennerick, 2005; Sakaba, 2006) in glutamatergic synapses, the former RRP (denoted as RRP_hyper) may encompass not only AP-evoked fast-releasing vesicles (TS vesicle) but also reluctant vesicles (LS vesicles). Because we measured p_v based on AP-evoked EPSCs such as strong paired pulse depression (PPD) and associated failure rates, p_v in the present study denotes vesicular fusion probability of TS vesicles, not that of LS plus TS vesicles.

Recent studies suggest that release sites are not fully occupied by TS vesicles in the baseline (Miki et al., 2016; Pulido and Marty, 2018; Malagon et al., 2020; Lin et al., 2022). Instead, the occupancy (p_occ) by TS vesicles is subject to dynamic regulation by reversible rate constants (denoted by k₁ and b₁, respectively). The number of TS vesicles (n) can be factored into the number of release sites (N) and p_occ, among which N is a fixed parameter but p_occ depends on k₁/(k₁+b₁) under the framework of the simple refilling model (see Methods). Because these refilling rate constants are regulated by Ca²⁺ (Hosoi, et al., 2008), p_occ is not a fixed parameter. Therefore, release probability should be re-defined as p_occ * p_v. Given that N is fixed, the increase in release probability is a major player in STF. Our study asserts that STF by 2.3 times can be attributed to an increase in p_occ rather than p_v, because p_v is close to unity (Fig. S8). Moreover, strong PPD was observed not only in the baseline but also at the early and in the middle of a train (Fig. 2 and 7) and during the recovery phase (Fig. 3), arguing against a gradual increase in p_v of reluctant vesicles.

We imagine that the Reviewer meant vesicular release or fusion probability (p_v) by ‘release probability’. If so, p_v (of TS vesicles) cannot be a major player in STF, because the baseline p_v is already higher than 0.8 even if it is most parsimoniously estimated (Fig. 2). Moreover, considering very high refilling rate (23/s), the high double failure rate cannot be explained without assuming that p_v is close to unity (Fig. S8).

Conventional models for facilitation assume a post-AP residual Ca²⁺-dependent step increase in p_v of RRP (Dittman et al., 2000) or reluctant vesicles (Turecek et al., 2016). Given that p_v of TS vesicles is close to one, an increase in p_v of TS vesicles cannot account for facilitation. The possibility for activity-dependent increase in fusion probability of LS vesicles (denoted as p_v,LS) should be considered in two ways depending on whether LS and TS vesicles reside in distinct pools or in the same pool. Notably, strong PPD at short ISI implies that p_v,LS is near zero at the resting state. Whereas LS vesicles do not contribute to baseline transmission, short-term facilitation (STF) may be mediated by cumulative increase in p_{v v,LS} that reside in a distinct pool. Because the increase in p_v,LS during facilitation recruits new release sites (increase in N), the variance of EPSCs should become larger as stimulation frequency increases, resulting in upward deviation from a parabola in the V-M plane, as shown in recent studies (Valera et al., 2012; Kobbersmed et al., 2020). This prediction is not compatible with our results of V-M analysis (Fig. 3), showing that EPSCs during STF fell on the same parabola regardless of stimulation frequencies. Therefore, it is unlikely that an increase in fusion probability of reluctant vesicles residing in a distinct release pool mediates STF in the present study.

For the latter case, in which LS and TS vesicles occupy in the same release sites, it is hard to distinguish a step increase in fusion probability of LS vesicles from a conversion of LS vesicles to TS. Nevertheless, our results do not support the possibility for gradual increase in p_v,LS that occurs in parallel with STF. Strong PPD, indicative of high p_v, was consistently found not only in the baseline (Fig. 2 and Fig. S6) but also during post-tetanic augmentation phase (Fig. 3D) and even during the early development of facilitation (Fig. 2D-E and Fig. 7), arguing against gradual increase in p_v,LS. One may argue that STF may be mediated by a drastic step increase of p_v,LS from zero to one, but it is not distinguishable from conversion of LS to TS vesicles.

To address the reviewer’s concern, we incorporated these perspectives into Discussion and further clarified the reasoning behind our conclusions.

References

Moulder KL, Mennerick S (2005) Reluctant vesicles contribute to the total readily releasable pool in glutamatergic hippocampal neurons. J Neurosci 25:3842–3850.

Sakaba, T (2006) Roles of the fast-releasing and the slowly releasing vesicles in synaptic transmission at the calyx of Held. J Neurosci 26(22): 5863-5871.

Please note that papers cited in the manuscript are not repeated here.

(2) Fig 3 I am confused about the interpretation of the Mean Variance analysis outcome. Since the data points follow the curve during induction of short term plasticity, aren't these suggesting that release probability and not the pool size increases? Related, to measure the absolute release probability and failure rate using the optogenetic stimulation technique is not trivial as the experimental paradigm bias the experiment to a given output strength, and therefore a change in release probability cannot be excluded.

Under the recent definition of release probability, it can be factored into p_v and p_occ, which are fusion probability of TS vesicles and the occupancy of release sites by TS vesicles, respectively. With this regard, our interpretation of the Variance-Mean results is consistent with conventional one: different data points along a parabola represent a change in release probability (= p_occ x p_v). Our novel finding is that the increase in release probability should be attributed to an increase in p_occ, not to that in p_v.

(3) Fig4B interprets the phorbol ester stimulation to be the result of pool overfilling, however, phorbol ester stimulation has also been shown to increase release probability without changing the size of the readily releasable pool. The high frequency of stimulation may occlude an increased paired pulse depression in presence of OAG, which others have interpreted in mammalian synapses as an increase in release probability.

To our experience in the calyx of Held synapses, OAG, a DAG analogue, increased the fast releasing vesicle pool (FRP) size (Lee JS et al., 2013), consistent with our interpretation (pool overfilling). Once the release sites are overfilled in the presence of OAG, it is expected that the maximal STF (ratio of facilitated to baseline EPSCs) becomes lower as long as the number of release sites (N) are limited. As aforementioned, the baseline p_v is already close to one, and thus it cannot be further increased by OAG. Instead, the baseline p_occ seems to be increased by OAG.

Reference

Lee JS, et al., Superpriming of synaptic vesicles after their recruitment to the readily releasable pool. Proc Natl Acad Sci U S A, 2013. 110(37): 15079-84.

(4) The literature on Syt7 function is still quite controversial. An observation in the literature that loss of Syt7 function in the fly synapse leads to an increase of release probability. Thus the observed changes in short term plasticity characteristics in the Syt7 KD experiments may contain a release probability component. Can the authors really exclude this possibility? Figure 5 shows for the Syt7 KD group a very prominent depression of the EPSC/IPSC with the second stimulus, particularly for the short interpulse intervals, usually a strong sign of increased release probability, as lack of pool refilling can unlikely explain the strong drop in synaptic output.

The reviewer raises an interesting point regarding the potential link between Syt7 KD and increased initial p_v, particularly in light of observations in Drosophila synapses (Guan et al., 2020; Fujii et al., 2021), in which Syt7 mutants exhibited elevated initial p_v. However, it is important to note that these findings markedly differ from those in mammalian systems, where the role of Syt7 in regulating initial p_v has been extensively studied. In rodents, consistent evidence indicates that Syt7 does not significantly affect initial p_v, as demonstrated in several studies (Jackman et al., 2016; Chen et al., 2017; Turecek and Regehr, 2018). Furthermore, in our study of excitatory synapses in the mPFC layer 2/3, we observed an initial p_v already near its maximal level, approaching a value of 1. Consequently, it is unlikely that the loss of Syt7 could further elevate the initial p_v. Instead, such effects are more plausibly explained by alternative mechanisms, such as alterations in vesicle replenishment dynamics, rather than a direct influence on p_v.

References

Chen, C., et al., Triple Function of Synaptotagmin 7 Ensures Efficiency of High-Frequency Transmission at Central GABAergic Synapses. Cell Rep, 2017. 21(8): 2082-2089.

Fujii, T., et al., Synaptotagmin 7 switches short-term synaptic plasticity from depression to facilitation by suppressing synaptic transmission. Scientific reports, 2021. 11(1): 4059.

Guan, Z., et al., Drosophila Synaptotagmin 7 negatively regulates synaptic vesicle release and replenishment in a dosage-dependent manner. Elife, 2020. 9: e55443.

Jackman, S.L., et al., The calcium sensor synaptotagmin 7 is required for synaptic facilitation. Nature, 2016. 529(7584): 88-91.

Turecek, J. and W.G. Regehr, Synaptotagmin 7 mediates both facilitation and asynchronous release at granule cell synapses. Journal of Neuroscience, 2018. 38(13): 3240-3251.

Reviewer #3 (Public review):

Summary:

The report by Shin, Lee, Kim, and Lee entitled "Progressive overfilling of readily releasable pool underlies short-term facilitation at recurrent excitatory synapses in layer 2/3 of the rat prefrontal cortex" describes electrophysiological experiments of short-term synaptic plasticity during repetitive presynaptic stimulation at synapses between layer 2/3 pyramidal neurons and nearby target neurons. Manipulations include pharmacological inhibition of PLC and actin polymerization, activation of DAG receptors, and shRNA knockdown of Syt7. The results are interpreted as support for the hypothesis that synaptic vesicle release sites are vacant most of the time at resting synapses (i.e., p_occ is low) and that facilitation (and augmentation) components of short-term enhancement are caused by an increase in occupancy, presumably because of acceleration of the transition from not-occupied to occupied. The report additionally describes behavioural experiments where trace fear conditioning is degraded by knocking down syt7 in the same synapses.

Strengths:

The strength of the study is in the new information about short-term plasticity at local synapses in layer 2/3, and the major disruption of a memory task after eliminating short-term enhancement at only 15% of excitatory synapses in a single layer of a small brain region. The local synapses in layer 2/3 were previously difficult to study, but the authors have overcome a number of challenges by combining channel rhodopsins with in vitro electroporation, which is an impressive technical advance.

Weaknesses:

(1) The question of whether or not short-term enhancement causes an increase in p_occ (i.e., "readily releasable pool overfilling") is important because it cuts to the heart of the ongoing debate about how to model short term synaptic plasticity in general. However, my opinion is that, in their current form, the results do not constitute strong support for an increase in p_occ, even though this is presented as the main conclusion. Instead, there are at least two alternative explanations for the results that both seem more likely. Neither alternative is acknowledged in the present version of the report.

The evidence presented to support overfilling is essentially two-fold. The first is strong paired pulse depression of synaptic strength when the interval between action potentials is 20 or 25 ms, but not when the interval is 50 ms. Subsequent stimuli at frequencies between 5 and 40 Hz then drive enhancement. The second is the observation that a slow component of recovery from depression after trains of action potentials is unveiled after eliminating enhancement by knocking down syt7. Of the two, the second is predicted by essentially all models where enhancement mechanisms operate independently of release site depletion - i.e., transient increases in p_occ, p_v, or even N - so isn't the sort of support that would distinguish the hypothesis from alternatives (Garcia-Perez and Wesseling, 2008, https://doi.org/10.1152/jn.01348.2007).

The apparent discrepancy in interpretation of post-tetanic augmentation between the present and previous papers [Sevens Wesseling (1999), Garcia-Perez and Wesseling (2008)] is an important issue that should be clarified. We noted that different meanings of ‘vesicular release probability’ in these papers are responsible for the discrepancy. We added an explanation to Discussion on the difference in the meaning of ‘vesicular release probability’ between the present study and previous studies [Sevens Wesseling (1999), Garcia-Perez and Wesseling (2008)]. In summary, the p_v in the present study was used for vesicular release probability of TS vesicles, while previous studies used it as vesicular release probability of vesicles in the RRP, which include LS and TS vesicles. Accordingly, p_occ in the present study is the occupancy of release sites by TS vesicles.

Not only double failure rate but also other failure rates upon paired pulse stimulation were best fitted at p_v close to 1 (Fig. S8 and associated text). Moreover, strong PPD, indicating release of vesicles with high p_v, was observed not only at the beginning of a train but also in the middle of a 5 Hz train (Fig. 2D), during the augmentation phase after a 40 Hz train (Fig 3D), and in the recovery phase after three pulse bursts (Fig. 7). Given that p_v is close to 1 throughout the EPSC trains and that N does not increase during a train (Fig. 3), synaptic facilitation can be attained only by the increase in p_occ (occupancy of release sites by TS vesicles). In addition, it should be noted that Fig. 7 demonstrates strong PPD during the recovery phase after depletion of TS vesicles by three pulse bursts, indicating that recovered vesicles after depletion display high p_v too. Knock-down of Syt7 slowed the recovery of TS vesicles after depletion of TS vesicles, highlighting that Syt7 accelerates the recovery of TS vesicles following their depletion.

As addressed in our reply to the first issue raised by Reviewer #2 and the third issue raised by Reviewer #3, our results do not support possibilities for recruitment of new release sites (increase in N) having low p_v or for a gradual increase in p_v of reluctant vesicles during short-term facilitation.  

Following statement was added to Discussion in the revised manuscript

“Previous studies suggested that an increase in p_v is responsible for post-tetanic augmentation (Stevens and Wesseling, 1999; Garcia-Perez and Wesseling, 2008) by observing invariance of the RRP size after tetanic stimulation. In these studies, the RRP size was estimated by hypertonic sucrose solution or as the sum of EPSCs evoked 20 Hz/60 pulses train (denoted as ‘RRP_hyper’). Because reluctant vesicles (called LS vesicles) can be quickly converted to TS vesicles (16/s) and are released during a train (Lee et al., 2012), it is likely that the RRP size measured by these methods encompasses both LS and TS vesicles. In contrast, we assert high p_v based on the observation of strong PPD and failure rates upon paired stimulations at ISI of 20 ms (Fig. 2 and Fig. S8). Given that single AP-induced vesicular release occurs from TS vesicles but not from LS vesicles, p_v in the present study indicates the fusion probability of TS vesicles. From the same reasons, p_occ denotes the occupancy of release sites by TS vesicles. Note that our study does not provide direct clue whether release sites are occupied by LS vesicles that are not tapped by a single AP, although an increase in the LS vesicle number may accelerate the recovery of TS vesicles. As suggested in Neher (2024), even if the number of LS plus TS vesicles are kept constant, an increase in p_occ (occupancy by TS vesicles) would be interpreted as an increase in ‘vesicular release probability’ as in the previous studies (Stevens and Wesseling (1999); Garcia-Perez and Wesseling (2008)) as long as it was measured based on RRP_hyper.”

(2) Regarding the paired pulse depression: The authors ascribe this to depletion of a homogeneous population of release sites, all with similar p_v. However, the details fit better with the alternative hypothesis that the depression is instead caused by quickly reversing inactivation of Ca²⁺ channels near release sites, as proposed by Dobrunz and Stevens to explain a similar phenomenon at a different type of synapse (1997, PNAS, https://doi.org/10.1073/pnas.94.26.14843). The details that fit better with Ca²⁺ channel inactivation include the combination of the sigmoid time course of the recovery from depression (plotted backwards in Fig1G,I) and observations that EGTA (Fig2B) increases the paired-pulse depression seen after 25 ms intervals. That is, the authors ascribe the sigmoid recovery to a delay in the activation of the facilitation mechanism, but the increased paired pulse depression after loading EGTA indicates, instead, that the facilitation mechanism has already caused p_r to double within the first 25 ms (relative to the value if the facilitation mechanism was not active). Meanwhile, Ca²⁺ channel inactivation would be expected to cause a sigmoidal recovery of synaptic strength because of the sigmoidal relationship between Ca²⁺-influx and exocytosis (Dodge and Rahamimoff, 1967, https://doi.org/10.1113/jphysiol.1967.sp008367).

The Ca²⁺-channel inactivation hypothesis could probably be ruled in or out with experiments analogous to the 1997 Dobrunz study, except after lowering extracellular Ca²⁺ to the point where synaptic transmission failures are frequent. However, a possible complication might be a large increase in facilitation in low Ca²⁺ (Fig2B of Stevens and Wesseling, 1999, https://doi.org/10.1016/s0896-6273(00)80685-6).

We appreciate the reviewer's thoughtful comment regarding the potential role of Ca²⁺ channel inactivation in the observed paired-pulse depression (PPD). As noted by the Reviewer, the Dobrunz and Stevens (1997) suggested that the high double failure rate at short ISIs in synapses exhibiting PPD can be attributed to Ca²⁺ channel inactivation. This interpretation seems to be based on a premise that the number of RRP vesicles are not varied trial-by-trial. The number of TS vesicles, however, can be dynamically regulated depending on the parameters k₁ and b₁, as shown in Fig. S8, implying that the high double failure rate at short ISIs cannot be solely attributed to Ca²⁺ channel inactivation. Nevertheless, we acknowledge the possibility that Ca²⁺ channel inactivation may contribute to PPD, and therefore, we have further investigated this possibility. Specifically, we measured action potential (AP)-evoked Ca²⁺ transients at individual axonal boutons of layer 2/3 pyramidal cells in the mPFC using two-dye ratiometry techniques. Our analysis revealed no evidence for Ca²⁺ channel inactivation during a 40 Hz train of APs. This finding indicates that voltage-gated Ca²⁺ channel inactivation is unlikely to contribute to the pronounced PPD.

Figure 2—figure supplement 2 shows how we measured the total Ca²⁺ increments at axonal boutons. First we estimated endogenous Ca²⁺-binding ratio from analyses of single AP-induced Ca²⁺ transients at different concentrations of Ca²⁺ indicator dye (panels A to E). And then, using the Ca²⁺ buffer properties, we converted free [Ca²⁺] amplitudes to total calcium increments for the first four AP-evoked Ca²⁺ transients in a 40 Hz train (panels G-I). We incorporated these results into the revised version of our manuscript to provide evidence against the Ca²⁺ channel inactivation.

(3) On the other hand, even if the paired pulse depression is caused by depletion of release sites rather than Ca²⁺-channel inactivation, there does not seem to be any support for the critical assumption that all of the release sites have similar p_v. And indeed, there seems to be substantial emerging evidence from other studies for multiple types of release sites with 5 to 20-fold differences in p_v at a wide variety of synapse types (Maschi and Klyachko, eLife, 2020, https://doi.org/10.7554/elife.55210; Rodriguez Gotor et al, eLife, 2024, https://doi.org/10.7554/elife.88212 and refs. therein). If so, the paired pulse depression could be caused by depletion of release sites with high p_v, whereas the facilitation could occur at sites with much lower p_v that are still occupied. It might be possible to address this by eliminating assumptions about the distribution of p_v across release sites from the variance-mean analysis, but this seems difficult; simply showing how a few selected distributions wouldn't work - such as in standard multiple probability fluctuation analyses - wouldn't add much.

We appreciate the reviewer’s insightful comments regarding the potential increase in p_fusion of reluctant vesicles. It should be noted, however, that Maschi and Klyachko (2020) showed a distribution of release probability (p_r) within a single active zone rather than a heterogeneity in p_fusion of individual docked vesicles. Therefore both p_occ and p_v of TS vesicles would contribute to the p_r distribution shown in Maschi and Klyachko (2020). 

The Reviewer’s concern aligns closely with the first issue raised by Reviewer #2, to which we addressed in detail. Briefly, new release site may not be recruited during facilitation or post-tetanic augmentation, because variance of EPSCs during and after a train fell on the same parabola (Fig. 3). Secondly, strong PPD was observed not only in the baseline but also during early and late phases of facilitation, indicating that vesicles with very high p_v contribute to EPSC throughout train stimulations (Fig. 2, 3, and 7). These findings argue against the possibilities for recruitment of new release sites harboring low p_v vesicles and for a gradual increase in fusion probability of reluctant vesicles.

To address the reviewers’ concern, we incorporated the perspectives into Discussion and further clarified the reasoning behind our conclusions.

(4) In any case, the large increase - often 10-fold or more - in enhancement seen after lowering Ca²⁺ below 0.25 mM at a broad range of synapses and neuro-muscular junctions noted above is a potent reason to be cautious about the LS/TS model. There is morphological evidence that the transitions from a loose to tight docking state (LS to TS) occur, and even that the timing is accelerated by activity. However, 10-fold enhancement would imply that at least 90 % of vesicles start off in the LS state, and this has not been reported. In addition, my understanding is that the reverse transition (TS to LS) is thought to occur within 10s of ms of the action potential, which is 10-fold too fast to account for the reversal of facilitation seen at the same synapses (Kusick et al, 2020, https://doi.org/10.1038/s41593-020-00716-1).

As the Reviewer suggested, low external Ca²⁺ concentration can lower release probability (p_r). Given that both p_v and p_occ are regulated by [Ca²⁺]_i, low external [Ca²⁺] may affect not only p_v but also p_occ, both of which would contribute to low p_r. Under such conditions, it would be plausible that the baseline p_r becomes much lower than 0.1 due to low p_v and p_occ (for instance, p_v decreases from 1 to 0.5, and p_occ from 0.3 to 0.1, then p_r = 0.05), and then p_r (= p_v x p_occ) has a room for an increase by a factor of ten (0.5, for example) by short-term facilitation as cytosolic [Ca²⁺] accumulates during a train.

If p_v is close to one, p_r depends p_occ, and thus facilitation depends on the number of TS vesicles just before arrival of each AP of a train. Thus, post-train recovery from facilitation would depend on restoration of equilibrium between TS and LS vesicles to the baseline. Even if transition between LS and TS vesicles is very fast (tens of ms), the equilibrium involved in de novo priming (reversible transitions between recycling vesicle pool and partially docked LS vesicles) seems to be much slower (13 s in Fig. 5A of Wu and Borst 1999). Thus, we can consider a two-step priming model (recycling pool -> LS -> TS), which is comprised of a slow 1st step (-> LS) and a fast 2nd step (-> TS). Under the framework of the two-step model, the slow 1st step (de novo priming step) is the rate limiting step regulating the development and recovery kinetics of facilitation. Given that on and off rate for Ca²⁺ binding to Syt7 is slow, it is plausible that Syt7 may contribute to short-term facilitation (STF) by Ca²⁺-dependent acceleration of the 1st step (as shown in Fig. 9). During train stimulation, the number of LS vesicles would slowly accumulate in a Syt7 and Ca²⁺-dependent manner, and this increase in LS vesicles would shift LS/TS equilibrium towards TS, resulting in STF. After tetanic stimulation, the recovery kinetics from facilitation would be limited by slow recovery of LS vesicles.

Reference

Wu, L.-G. and Borst J.G.G. (1999) The reduced release probability of releasable vesicles during recovery from short-term synaptic depression. Neuron, 23(4): 821-832.

Please note that papers cited in the manuscript are not repeated here.

Individual points:

(1) An additional problem with the overfilling hypothesis is that syt7 knockdown increases the estimate of p_occ extracted from the variance-mean analysis, which would imply a faster transition from unoccupied to occupied, and would consequently predict faster recovery from depression. However, recovery from depression seen in experiments was slower, not faster. Meanwhile, the apparent decrease in the estimate of N extracted from the mean-variance analysis is not anticipated by the authors' model, but fits well with alternatives where p_v varies extensively among release sites because release sites with low p_v would essentially be silent in the absence of facilitation.

Slower recovery from depression observed in the Syt7 knockdown (KD) synapses (Fig. 7) may results from a deficiency in activity-dependent acceleration of TS vesicle recovery. Although basal occupancy was higher in the Syt7 KD synapses, this does not indicate a faster activity-dependent recovery.

Higher baseline occupancy does not always imply faster recovery of PPR too. Actually PPR recovery was slower in Syt7 KD synapses than WT one (18.5 vs. 23/s). Under the framework of the simple refilling model (Fig. S8Aa), the baseline occupancy and PPR recovery rate are calculated as k₁ / (k₁ + b₁) and (k₁ + b₁), respectively. The baseline occupancy depends on k₁/b₁, while the PPR recovery on absolute values of k₁ and b₁. Based on p_occ and PPR recovery time constant of WT and KD synapses, we expect higher k₁/b₁ but lower values for (k₁ + b₁) in Syt7 KD synapses compared to WT ones.

Lower release sites (N) in Syt7-KD synapses was not anticipated. As you suggested, such low N might be ascribed to little recruitment of release sites during a train in KD synapses. But our results do not support this model. If silent release sites are recruited during a train, the variance should upwardly deviate from the parabola predicted under a fixed N (Valera et al., 2012; Kobbersmed et al. 2020). Our result was not the case (Fig. 3). In the first version of the manuscript, we have argued against this possibility in line 203-208.

As discussed in both the Results and Discussion sections, the baseline EPSC was unchanged by KD (Fig. S3) because of complementary changes in the number of docking sites and their baseline occupancy (Fig. 6). These findings suggest that Syt7 may be involved in maintaining additional vacant docking sites, which could be overfilled during facilitation. It remains to be determined whether the decrease in docking sites in Syt7 KD synapses is related to its specific localization of Syt7 at the plasma membrane of active zones, as proposed in previous studies (Sugita et al., 2001; Vevea et al., 2021).

(2) Figure S4A: I like the TTX part of this control, but the 4-AP part needs a positive control to be meaningful (e.g., absence of TTX).

The reason why we used 4-AP in the presence of TTX was to increase the length constant of axon fibers and to facilitate the conduction of local depolarization in the illumination area to axon terminals. The lack of EPSC in the presence of 4-AP and TTX indicates that illumination area is distant from axon terminals enough for optic stimulation-induced local depolarization not to evoke synaptic transmission. This methodology has been employed in previous studies including the work of Little and Carter (2013).

Reference

Little JP and Carter AG (2013) Synaptic mechanisms underlying strong reciprocal connectivity between the medial prefrontal cortex and basolateral amygdala. J Neurosci, 33(39): 15333-15342.

(3) Line 251: At least some of the previous studies that concluded these drugs affect vesicle dynamics used logic that was based on some of the same assumptions that are problematic for the present study, so the reasoning is a bit circular.

(4) Line 329 and Line 461: A similar problem with circularity for interpreting earlier syt7 studies.

(Reply to #3 and #4) We selected the target molecules as candidates based on their well-characterized roles in vesicle dynamics, and aimed to investigate what aspects of STP are affected by these molecules in our experimental context. For example, we could find that the baseline p_occ and short-term facilitation (STF) are enhanced by the baseline DAG level and train stimulation-induced PLC activation, respectively. Notably, the effect of dynasore informed us that slow site clearing is responsible for the late depression of 40 Hz train EPSC. The knock-down experiments also provided us with information on the critical role of Syt7 in replenishment of TS vesicles. These approaches do not deviate from standard scientific reasoning but rather builds upon prior knowledge to formulate and test hypotheses.

Importantly, our conclusions do not rely solely on the assumption that altering the target molecule impacts synaptic transmission. Instead, our conclusions are derived from a comprehensive analysis of diverse outcomes obtained through both pharmacological and genetic manipulations. These interpretations align closely with prior literature, further validating our conclusions.

Therefore, the use of established studies to guide candidate selection and the consistency of our findings with existing knowledge do not represent a logical circularity but rather a reinforcement of the proposed mechanism through converging lines of evidence.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Comments:

(1) While the authors claim that Syt7-mediated facilitation is connected to the behavioral deficits they observed, this link is still somewhat speculative. This manuscript could benefit from further discussions of other alternative mechanisms to consider.

We added following statement to Discussion of the revised manuscript:

“The acquisition of trace fear memory was impaired by inhibition of persistent activity in mPFC during trace period (Gilmartin et al., 2013). The similar deficit observed in Syt7 KD animals is consistent with the hypothesis that STF provides bi-stable ensemble activity in a recurrent network (Mongillo et al., 2012). Nevertheless, alternative mechanisms may be responsible for the behavioral deficit. Not only recurrent network but also long-range loop between the mPFC and the mediodorsal (MD) thalamus play a critical role in maintaining persistent activity within the mPFC especially for a delay period longer than 10 s (Bolkan et al., 2017). Prefrontal L2/3 is heavily innervated by MD thalamus, and L2/3-PCs subsequently relay signals to L5 cortico-thalamic (CT) neurons (Collins et al., 2018). Given that L2/3 is an essential component of the PFC-thalamic loop, loss of STF at recurrent synapses between L2/3 PCs may lead to insufficient L2/3 inputs to L5 CT neurons and failure in the reverberant PFC-MD thalamic feedback loop. Therefore, not only L2/3 recurrent network but also its output to downstream network should be considered as a possible network mechanism underlying behavioral deficit caused by Syt7 KD L2/3.”

(2) The authors mention that Syt7 contributes to persistent activity during working memory tasks but focus on using only a trace fear conditioning task. However, it would be interesting to see if their results are generalizable to other working memory tasks (i.e. a delayed alternation task).

We thank to Reviewer for the insightful suggestion. Trace fear conditioning (tFC) shares behavioral properties with working memory (WM) tasks in that tFC is vulnerable to attentional distraction and to the load of WM task. In general WM tasks including delayed alternation tasks such as a T-maze task need persistent activity of ensemble neurons representing target-specific information among multiple choices. Different from such WM tasks, tFC is not appropriate to examine target-specific ensemble activity. Because it is not trivial to examine in vivo recordings in KD animals during delayed alternation tasks, it will be appropriate to study the effect of Syt7 KD in a separate study. 

(3) The figure legend in Figure 6A and 6B mentions dotted lines and broken lines in the figure. However, this is confusing, and it is unclear as to what these lines are referring to in the figure.

To avoid the confusion in the figure legend for Figure 6A and 6B, we corrected “dotted line” to " vertical broken line", and “broken lines” to “dashed parabolas”.

(4) The manuscript can benefit from close reading and editing to catch typos and improve general readability (i.e. line 173: the word "are" is repeated twice).

We corrected typographical errors throughout the manuscript and carefully read the manuscript to improve readability. A revised version reflecting these corrections has been prepared and will be resubmitted for your consideration.

Reviewer #3 (Recommendations for the authors):

The points in this section are all minor.

(1) Line 44: Define release probability (p_r) more clearly. Authors use it to mean p_v*p_occ, but others routinely use it to mean p_v*p_occ*N.

We understand that the Reviewer meant “others routinely use it to mean p_v”. At this statement, we meant conventional definition of release probability, which is release probability among vesicles of RRP. We think that it is not appropriate to re-define release probability as p_v * p_occ in this first paragraph of Introduction. Therefore we clarified this issue in Discussion as we mentioned in our reply to the 1st weakness issue raised by Reviewer #3.   

(2) Line 82: For clarity, define better what recurrent excitatory synapses are. It seems that synapses between L2/3 PCs and local targets may all be recurrent?

Each of L2/3 and L5 of the prefrontal cortical layers harbors intralaminar recurrent excitatory synapses between pyramidal cells, called a recurrent network. Previous theoretical studies have proposed that a single layer recurrent network model can have bi-stable E/I balanced states (up- and down-states) if recurrent excitatory synapses display short-term facilitation (STF), and thus is able to temporally hold an information once external input shifts the network to the up-state. In this theory, synapses to local targets across layers are not considered and specific roles of L2/3 and L5 in working memory tasks are still elusive. For clarity, we added a statement at the beginning of the paragraph (line 82): “Each of layer 2/3 (L2/3) and layer 5 (L5) of neocortex displays intralaminar excitatory synapses between pyramidal cells comprising a recurrent network (Holmgren et al., 2003; Thomson and Lamy, 2007)”

(3) Cite earlier studies of short-term synaptic plasticity at synapses between L2/3 pyramidal neurons and local targets in mPFC. If there are none, take more explicit credit for being first.

As we mentioned in Introduction, previous studies on short-term plasticity (STP) at neocortical excitatory recurrent synapses have focused on synapses between L5 pyramidal cells (PCs) (Hemple et al. 2000; Wang et al. 2006; Morishima et al., 2011; Yoon et al., 2020). The local connectivity between L2/3 PCs in the somatosensory cortex has been elucidated by Homgren et al. (2003) and Ko et al. (2011). Although these study showed STP of EPSPs, it was at a fixed frequency or stimulus pattern at high external [Ca²⁺] (2 mM). There is a study on the frequency-dependence of STP of EPSP between L2/3-PCs (Feldmyer et al., 2006). Different from our study, Feldmyer et al., (2006) observed monotonous STD at all frequencies less than 50 Hz, but this study was done in the somatosensory cortex and at high external [Ca²⁺] (2 mM). To our knowledge, no previous study have investigated STP at recurrent excitatory synapses of L2/3 pyramidal cells of the mPFC especially at physiological external [Ca²⁺]. The present study, therefore, represents the first extensive investigation of STP at recurrent excitatory synapses in L2/3 of the mPFC under physiologically relevant external [Ca²⁺].

References

Feldmeyer D, Lubke J, Silver RA, Sakmann B (2002) Synaptic connections between layer 4 spiny neurone-layer 2/3 pyramidal cell pairs in juvenile rat barrel cortex: physiology and anatomy of interlaminar signalling within a cortical column. J Physiol 538:803-822.

Holmgren C, Harkany T, Svennenfors B, Zilberter Y (2003) Pyramidal cell communication within local networks in layer 2/3 of rat neocortex. J Physiol 551:139-153.

Ko H, Hofer SB, Pichler B, Buchanan KA, Sjöström PJ, Mrsic-Flogel TD (2011) Functional specificity of local synaptic connections in neocortical networks. Nature 473:87-91.

Morishima M, Morita K, Kubota Y, Kawaguchi Y (2011) Highly differentiated projection-specific cortical subnetworks. Journal of Neuroscience 31:10380-10391.

Wang Y, Markram H, Goodman PH, Berger TK, Ma J, Goldman-Rakic PS (2006) Heterogeneity in the pyramidal network of the medial prefrontal cortex. Nat Neurosci 9:534-542.

(4) I couldn't figure out the significance of Figure S3. Perhaps this could be explained better.

Optical minimal stimulation methods have not been previously documented in detail. This figure illustrates what parameters we should carefully examine in order to attain optical minimal stimulation, which hopefully stimulates a single afferent fiber. A single fiber stimulation by optical minimal stimulation is supported by the similarity of our estimate for the number of release sites (N) as the previous morphological estimate (Holler et al., 2021). For minimal stimulation, we used a collimated DMD-coupled LED was employed to restrict 470 nm illumination to a small and well-defined region within layer 2/3 of the prelimbic mPFC, and carefully adjusted the illumination radius such that one step smaller (by 1 μm) illumination results in failure to evoke EPSCs. Our typical illumination area ranged between 3–4 μm, as shown in Figure S3A. Under this minimal illumination area, we confirmed unimodal distributions for the EPSC parameters (amplitude, rise time, decay time and time to peak; Figure 3B-E). Otherwise, we excluded the recordings from analysis. We hope this explanation provides a clearer understanding of the figure's significance.

(5) Note that CTZ seems to alter p_r at some synapses.

We acknowledge that CTZ can increase release probability by blocking presynaptic K⁺ currents. Indeed, Ishikawa and Takahashi (2001) reported that CTZ slowed the repolarizing phase of presynaptic action potentials and the frequency of miniature EPSCs in the calyx synapses. Consistently, we observed a slight increase in the baseline EPSC amplitude, from 33.3 pA to 41.9 pA (p=0.045) following the application of 50 µM CTZ. However, given that vesicular release probability (p_v) is already close to 1 at the synapse of our interest, we believe that the observed effect is more likely attributed to an increase in release sites occupancy (p_occ), which would be reflected as an increase in miniature EPSC frequency in Ishikawa and Takahashi (2001). Given that PPR depends on p_v rather than p_occ, this increase in p_occ would not critically change our conclusion that AMPA receptor desensitization is not responsible for the strong PPD.

Reference

Ishikawa, T., & Takahashi, T. (2001). Mechanisms underlying presynaptic facilitatory effect of cyclothiazide at the calyx of Held of juvenile rats. The Journal of Physiology, 533(2), 423-431.

(6) Figure 8B. The result in Figure 8C seems important, but I couldn't figure out why behaviour was not altered during the acquisition phase summarized in Figure 8B. Perhaps this could be explained more clearly for non-experts.

Little difference in freezing behavior during acquisition has been also observed when prelimbic persistent firing was optogenetically inhibited (Gilmartin, 2013). Not only CS (tone) but also other sensory inputs (visual and olfactory etc.) and the spatial context could be a cue predicting US (shock). Moreover, during the acquisition phase, the presence of the electric shock inherently induces a freezing response as a natural defensive behavior, which may obscure specific behavioral changes related to the associative learning process. Therefore, the freezing behavior during acquisition cannot be regarded as a sign for specific association of CS and US. Instead, on the next day, we specifically evaluated the CS-US association of the conditioned animals by measuring freezing behavior in response to CS in a distinct context. We explicitly documented little difference between WT and KD animals during the acquisition phase in the relevant paragraph (line 397).

Author response:

The following is the authors’ response to the original reviews

Reviewer #1 (Public Review):

Summary:

It seems as if the main point of the paper is about the new data related to rat fish although your title is describing it as extant cartilaginous fishes and you bounce around between the little skate and ratfish. So here's an opportunity for you to adjust the title to emphasize ratfish is given the fact that leader you describe how this is your significant new data contribution. Either way, the organization of the paper can be adjusted so that the reader can follow along the same order for all sections so that it's very clear for comparative purposes of new data and what they mean. My opinion is that I want to read, for each subheading in the results, about the the ratfish first because this is your most interesting novel data. Then I want to know any confirmation about morphology in little skate. And then I want to know about any gaps you fill with the cat shark. (It is ok if you keep the order of "skate, ratfish, then shark, but I think it undersells the new data).

The main points of the paper are 1) to define terms for chondrichthyan skeletal features in order to unify research questions in the field, and 2) add novel data on how these features might be distributed among chondrichthyan clades. However, we agree with the reviewer that many readers might be more interested in the ratfish data, so we have adjusted the order of presentation to emphasize ratfish throughout the manuscript.

Strengths:

The imagery and new data availability for ratfish are valuable and may help to determine new phylogenetically informative characters for understanding the evolution of cartilaginous fishes. You also allude to the fossil record.

Thank you for the nice feedback.

Opportunities:

I am concerned about the statement of ratfish paedomorphism because stage 32 and 33 were not statistically significantly different from one another (figure and prior sentences). So, these ratfish TMDs overlap the range of both 32 and 33. I think you need more specimens and stages to state this definitely based on TMD. What else leads you to think these are paedomorphic? Right now they are different, but it's unclear why. You need more outgroups.

Sorry, but we had reported that the TMD of centra from little skate did significantly increase between stage 32 and 33. Supporting our argument that ratfish had features of little skate embryos, TMD of adult ratfish centra was significantly lower than TMD of adult skate centra (Fig1). Also, it was significantly higher than stage 33 skate centra, but it was statistically indistinguishable from that of stage 33 and juvenile stages of skate centra. While we do agree that more samples from these and additional groups would bolster these data, we feel they are sufficiently powered to support our conclusions for this current paper.

Your headings for the results subsection and figures are nice snapshots of your interpretations of the results and I think they would be better repurposed in your abstract, which needs more depth.

We have included more data summarized in results sub-heading in the abstract as suggested (lines 32-37).

Historical literature is more abundant than what you've listed. Your first sentence describes a long fascination and only goes back to 1990. But there are authors that have had this fascination for centuries and so I think you'll benefit from looking back. Especially because several of them have looked into histology and development of these fishes.

I agree that in the past 15 years or so a lot more work has been done because it can be done using newer technologies and I don't think your list is exhaustive. You need to expand this list and history which will help with your ultimate comparative analysis without you needed to sample too many new data yourself.

We have added additional recent and older references: Kölliker, 1860; Daniel, 1934; Wurmbach, 1932; Liem, 2001; Arratia et al., 2001.

I'd like to see modifications to figure 7 so that you can add more continuity between the characters, illustrated in figure 7 and the body of the text.

We address a similar comment from this reviewer in more detail below, hoping that any concerns about continuity have been addressed with inclusion of a summary of proposed characters in a new Table 1, re-writing of the Discussion, and modified Fig7 and re-written Fig7 legend.

Generally Holocephalans are the outgroup to elasmobranchs - right now they are presented as sister taxa with no ability to indicate derivation. Why isn't the catshark included in this diagram?

While a little unclear exactly what was requested, we restructured the branches to indicate that holocephalans diverged earlier from the ancestors that led to elasmobranchs. Also in response to this comment, we added catshark (S. canicula) and little skate (L. erinacea) specifically to the character matrix.

In the last paragraph of the introduction, you say that "the data argue" and I admit, I am confused. Whose data? Is this a prediction or results or summary of other people's work? Either way, could be clarified to emphasize the contribution you are about to present.

Sorry for this lack of clarity, and we have changed the wording in this revision to hopefully avoid this misunderstanding.

Reviewer #1 (Recommendations For The Authors):

Further Strengths and Opportunities:

Your headings for the results subsection and figures are nice snapshots of your interpretations of the results and I think they would be better repurposed in your abstract, which needs more depth. It's a little unusual to try and state an interpretation of results as the heading title in a results section and the figures so it feels out of place. You could also use the headings as the last statement of each section, after you've presented the results. In order I would change these results subheadings to:

Tissue Mineral Density (TMD)

Tissue Properties of Neural Arches

Trabecular mineralization

Cap zone and Body zone Mineralization Patterns

Areolar mineralization

Developmental Variation

Sorry, but we feel that summary Results sub-headings are the best way to effectively communicate to readers the story that the data tell, and this style has been consistently used in our previous publications. No changes were made.

You allude to the fossil record and that is great. That said historical literature is more abundant than what you've listed. Your first sentence describes a long fascination and only goes back to 1990. But there are authors that have had this fascination for centuries and so I think you'll benefit from looking back. Especially because several of them have looked into histology of these fishes. You even have one sentence citing Coates et al. 2018, Frey et al., 2019 and ørvig 1951 to talk about the potential that fossils displayed trabecular mineralization. That feels like you are burying the lead and may have actually been part of the story for where you came up with your hypothesis in the beginning... or the next step in future research. I feel like this is really worth spending some more time on in the intro and/or the discussion.

We’ve added older REFs as pointed out above. Regarding fossil evidence for trabecular mineralization, no, those studies did not lead to our research question. But after we discovered how widespread trabecular mineralization was in extant samples, we consulted these papers, which did not focus on the mineralization patterns per se, but certainly led us to emphasize how those patterns fit in the context of chondrichthyan evolution, which is how we discussed them.

I agree that in the past 15 years or so a lot more work has been done because it can be done using newer technologies. That said there's a lot more work by Mason Dean's lab starting in 2010 that you should take a look at related to tesserae structure... they're looking at additional taxa than what you did as well. It will be valuable for you to be able to make any sort of phylogenetic inference as part of your discussion and enhance the info your present in figure 7. Go further back in time... For example:

de Beer, G. R. 1932. On the skeleton of the hyoid arch in rays and skates. Quarterly Journal of Microscopical Science. 75: 307-319, pls. 19-21.

de Beer, G. R. 1937. The Development of the Vertebrate Skull. The University Press, Oxford.

Indeed, we have read all of Mason’s work, citing 9 of his papers, and where possible, we have incorporated their data on different species into our Discussion and Fig7. Thanks for the de Beer REFs. While they contain histology of developing chondrichthyan elements, they appear to refer principally to gross anatomical features, so were not included in our Intro/Discussion.

Most sections within the results, read more like a discussion than a presentation of the new data and you jump directly into using an argument of those data too early. Go back in and remove the references or save those paragraphs for the discussion section. Particularly because this journal has you skip the method section until the end, I think it's important to set up this section with a little bit more brevity and conciseness. For instance, in the first section about tissue mineral density, change that subheading to just say tissue mineral density. Then you can go into the presentation of what you see in the ratfish, and then what you see in the little skate, and then that's it. You save the discussion about what other elasmobranch's or mineralizing their neural arches, etc. for another section.

We dramatically reduced background-style writing and citations in each Results section (other than the first section of minor points about general features of the ratfish, compared to catshark and little skate), keeping only a few to briefly remind the general reader of the context of these skeletal features.

I like that your first sentence in the paragraph is describing why you are doing. a particular method and comparison because it shows me (the reader) where you're sampling from. Something else is that maybe as part of the first figure rather than having just each with the graph have a small sketch for little skate and catch shark to show where you sampled from for comparative purposes. That would relate back, then to clarifying other figures as well.

Done (also adding a phylogenetic tree).

Second instance is your section on trabecular mineralization. This has so many references in it. It does not read like results at all. It looks like a discussion. However, the trabecular mineralization is one of the most interesting aspect of this paper, and how you are describing it as a unique feature. I really just want a very clear description of what the definition of this trabecular mineralization is going to be.

In addition to adding Table 1 to define each proposed endoskeletal character state, we have changed the structure of this section and hope it better communicates our novel trabecular mineralization results. We also moved the topic of trabecular mineralization to the first detailed Discussion point (lines 347-363) to better emphasize this specific topic.

Carry this reformatting through for all subsections of the results.

As mentioned above, we significantly reduced background-style writing and citations in each Results section.

I'd like to see modifications to figure 7 so that you can add more continuity between the characters, illustrated in figure 7 and the body of the text. I think you can give the characters a number so that you can actually refer to them in each subsection of the results. They can even be numbered sequentially so that they are presented in a standard character matrix format, that future researchers can add directly to their own character matrices. You could actually turn it into a separate table so it doesn't taking up that entire space of the figure, because there need to be additional taxa referred to on the diagram. Namely, you don't have any out groups in figure 7 so it's hard to describe any state specifically as ancestral and wor derived. Generally Holocephalans are the outgroup to elasmobranchs - right now they are presented as sister taxa with no ability to indicate derivation. Why isn't the catshark included in this diagram?

The character matrix is a fantastic idea, and we should have included it in the first place! We created Table 1 summarizing the traits and terminology at the end of the Introduction, also adding the character matrix in Fig7 as suggested, including specific fossil and extant species. For the Fig7 branching and catshark inclusion, please see above.

You can repurpose the figure captions as narrative body text. Use less narrative in the figure captions. These are your results actually, so move that text to the results section as a way to truncate and get to the point faster.

By figure captions, we assume the reviewer refers to figure legends. We like to explain figures to some degree of sufficiency in the legends, since some people do not read the main text and simply skim a manuscript’s abstract, figures, and figure legends. That said, we did reduce the wording, as requested.

More specific comments about semantics are listed here:

The abstract starts negative and doesn't state a question although one is referenced. Potential revision - "Comprehensive examination of mineralized endoskeletal tissues warranted further exploration to understand the diversity of chondrichthyans... Evidence suggests for instance that trabecular structures are not common, however, this may be due to sampling (bring up fossil record.) We expand our understanding by characterizing the skate, cat shark, and ratfish... (Then add your current headings of the results section to the abstract, because those are the relevant takeaways.)"

We re-wrote much of the abstract, hoping that the points come across more effectively. For example, we started with “Specific character traits of mineralized endoskeletal tissues need to be clearly defined and comprehensively examined among extant chondrichthyans (elasmobranchs, such as sharks and skates, and holocephalans, such as chimaeras) to understand their evolution”. We also stated an objective for the experiments presented in the paper: “To clarify the distribution of specific endoskeletal features among extant chondrichthyans”.

In the last paragraph of the introduction, you say that "the data argue" and I admit, I am confused. Whose data? Is this a prediction or results or summary of other people's work? Either way, could be clarified to emphasize the contribution you are about to present.

Sorry for this lack of clarity, and we have changed the wording in this revision to hopefully avoid this misunderstanding.

In the second paragraph of the TMD section, you mention the synarcual comparison. I'm not sure I follow. These are results, not methods. Tell me what you are comparing directly. The non-centrum part of the synarcual separate from the centrum? They both have both parts... did you mean the comparison of those both to the cat shark? Just be specific about which taxon, which region, and which density. No need to go into reasons why you chose those regions here.. Put into methods and discussion for interpretation.

We hope that we have now clarified wording of that section.

Label the spokes somehow either in caption or on figure direction. I think I see it as part of figure 4E, I, and J, but maybe I'm misinterpreting.

Based upon histological features (e.g., regions of very low cellularity with Trichrome unstained matrix) and hypermineralization, spokes in Fig4 are labelled with * and segmented in blue. We detailed how spokes were identified in main text (lines 241-243; 252-254) and figure legend (lines 597-603).

Reviewer #2 (Public Review):

General comment:

This is a very valuable and unique comparative study. An excellent combination of scanning and histological data from three different species is presented. Obtaining the material for such a comparative study is never trivial. The study presents new data and thus provides the basis for an in-depth discussion about chondrichthyan mineralised skeletal tissues.

Many thanks for the kind words

I have, however, some comments. Some information is lacking and should be added to the manuscript text. I also suggest changes in the result and the discussion section of the manuscript.

Introduction:

The reader gets the impression almost no research on chondrichthyan skeletal tissues was done before the 2010 ("last 15 years", L45). I suggest to correct that and to cite also previous studies on chondrichthyan skeletal tissues, this includes studies from before 1900.

We have added additional older references, as detailed above.

Material and Methods:

Please complete L473-492: Three different Micro-CT scanners were used for three different species? ScyScan 117 for the skate samples. Catshark different scanner, please provide full details. Chimera Scncrotron Scan? Please provide full details for all scanning protocols.

We clarified exact scanners and settings for each micro-CT experiment in the Methods (lines 476-497).

TMD is established in the same way in all three scanners? Actually not possible. Or, all specimens were scanned with the same scanner to establish TMD? If so please provide the protocol.

Indeed, the same scanner was used for TMD comparisons, and we included exact details on how TMD was established and compared with internal controls in the Methods. (lines 486-488)

Please complete L494 ff: Tissue embedding medium and embedding protocol is missing. Specimens have been decalcified, if yes how? Have specimens been sectioned non-decalcified or decalcified?

Please complete L506 ff: Tissue embedding medium and embedding protocol is missing. Description of controls are missing.

Methods were updated to include these details (lines 500-503).

Results:

L147: It is valuable and interesting to compare the degree of mineralisation in individuals from the three different species. It appears, however, not possible to provide numerical data for Tissue Mineral Density (TMD). First requirement, all specimens must be scanned with the same scanner and the same calibration values. This in not stated in the M&M section. But even if this was the case, all specimens derive from different sample locations and have, been preserved differently. Type of fixation, extension of fixation time in formalin, frozen, unfrozen, conditions of sample storage, age of the samples, and many more parameters, all influence TMD values. Likewise the relative age of the animals (adult is not the same as adult) influences TMD. One must assume different sampling and storage conditions and different types of progression into adulthood. Thus, the observation of different degrees of mineralisation is very interesting but I suggest not to link this observation to numerical values.

These are very good points, but for the following reasons we feel that they were not sufficiently relevant to our study, so the quantitative data for TMD remain scientifically valid and critical for the field moving forward. Critically, 1) all of the samples used for TMD calculations underwent the same fixation protocols, and 2) most importantly, all samples for TMD were scanned on the same micro-CT scanner using the same calibration phantoms for each scanning session. Finally, while the exact age of each adult was not specified, we note for Fig1 that clear statistically significant differences in TMD were observed among various skeletal elements from ratfish, shark, and skate. Indeed, ratfish TMD was considerably lower than TMD reported for a variety of fishes and tetrapods (summarized in our paper about icefish skeletons, who actually have similar TMD to ratfish: https://doi.org/10.1111/joa.13537).

In response, however, we added a caveat to the paper’s Methods (lines 466-469), stating that adult ratfish were frozen within 1 or 2 hours of collection from the wild, staying frozen for several years prior to thawing and immediate fixation.

Parts of the results are mixed with discussion. Sometimes, a result chapter also needs a few references but this result chapter is full of references.

As mentioned above, we reduced background-style writing and citations in each Results section.

Based on different protocols, the staining characteristics of the tissue are analysed. This is very good and provides valuable additional data. The authors should inform the not only about the staining (positive of negative) abut also about the histochemical characters of the staining. L218: "fast green positive" means what? L234: "marked by Trichrome acid fuchsin" means what? And so on, see also L237, L289, L291

We included more details throughout the Results upon each dye’s first description on what is generally reflected by the specific dyes of the staining protocols. (lines 178, 180, 184, 223, 227, and 243-244)

Discussion

Please completely remove figure 7, please adjust and severely downsize the discussion related to figure 7. It is very interesting and valuable to compare three species from three different groups of elasmobranchs. Results of this comparison also validate an interesting discussion about possible phylogenetic aspects. This is, however, not the basis for claims about the skeletal tissue organisation of all extinct and extant members of the groups to which the three species belong. The discussion refers to "selected representatives" (L364), but how representative are the selected species? Can there be a extant species that represents the entire large group, all sharks, rays or chimeras? Are the three selected species basal representatives with a generalist life style?

These are good points, and yes, we certainly appreciate that the limited sampling in our data might lead to faulty general conclusions about these clades. In fact, we stated this limitation clearly in the Introduction (lines 126-128), and we removed “representative” from this revision. We also replaced general reference to chondrichthyans in the Title by listing the specific species sampled. However, in the Discussion, we also compare our data with previously published additional species evaluated with similar assays, which confirms the trend that we are concluding. We look forward to future papers specifically testing the hypotheses generated by our conclusions in this paper, which serves as a benchmark for identifying shared and derived features of the chondrichthyan endoskeleton.

Please completely remove the discussion about paedomorphosis in chimeras (already in the result section). This discussion is based on a wrong idea about the definition of paedomorphosis. Paedomorphosis can occur in members of the same group. Humans have paedormorphic characters within the primates, Ambystoma mexicanum is paedormorphic within the urodeals. Paedomorphosis does not extend to members of different vertebrate branches. That elasmobranchs have a developmental stage that resembles chimera vertebra mineralisation does not define chimera vertebra centra as paedomorphic. Teleost have a herocercal caudal fin anlage during development, that does not mean the heterocercal fins in sturgeons or elasmobranchs are paedomorphic characters.

We agree with the reviewer that discussion of paedomorphosis should apply to members of the same group. In our paper, we are examining paedomorphosis in a holocephalan, relative to elasmobranch fishes in the same group (Chrondrichthyes), so this is an appropriate application of paedomorphosis. In response to this comment, we clarified that our statement of paedomorphosis in ratfish was made with respect to elasmobranchs (lines 37-39; 418-420).

L432-435: In times of Gadow & Abott (1895) science had completely wrong ideas bout the phylogenic position of chondrichthyans within the gnathostomes. It is curious that Gadow & Abott (1895) are being cited in support of the paedomorphosis claim.

If paedomorphosis is being examined within Chondrichthyes, such as in our paper and in the Gadow and Abbott paper, then it is an appropriate reference, even if Gadow and Abbott (and many others) got the relative position of Chondrichthyes among other vertebrates incorrect.

The SCPP part of the discussion is unrelated to the data obtained by this study. Kawaki & WEISS (2003) describe a gene family (called SCPP) that control Ca-binding extracellular phosphoproteins in enamel, in bone and dentine, in saliva and in milk. It evolved by gene duplication and differentiation. They date it back to a first enamel matrix protein in conodonts (Reif 2006). Conodonts, a group of enigmatic invertebrates have mineralised structures but these structure are neither bone nor mineralised cartilage. Cat fish (6 % of all vertebrate species) on the other hand, have bone but do not have SCPP genes (Lui et al. 206). Other calcium binding proteins, such as osteocalcin, were initially believed to be required for mineralisation. It turned out that osteocalcin is rather a mineralisation inhibitor, at best it regulates the arrangement collagen fiber bundles. The osteocalcin -/- mouse has fully mineralised bone. As the function of the SCPP gene product for bone formation is unknown, there is no need to discuss SCPP genes. It would perhaps be better to finish the manuscript with summery that focuses on the subject and the methodology of this nice study.

We completely agree with the reviewer that many papers claim to associate the functions of SCPP genes with bone formation, or even mineralization generally. The Science paper with the elephant shark genome made it very popular to associate SCPP genes with bone formation, but we feel that this was a false comparison (for many reasons)! In response to the reviewer’s comments, however, we removed the SCPP discussion points, moving the previous general sentence about the genetic basis for reduced skeletal mineralization to the end of the previous paragraph (lines 435-439). We also added another brief Discussion paragraph afterwards, ending as suggested with a summary of our proposed shared and derived chondrichthyan endoskeletal traits (lines 440-453).

Reviewer #2 (Recommendations For The Authors):

Other comments

L40: remove paedomorphism

No change; see above

L53: down tune languish, remove "severely" and "major"

Done (lines 57-59)

L86: provide species and endoskeletal elements that are mineralized

No change; this paragraph was written generally, because the papers cited looked at cap zones of many different skeletal elements and neural arches in many different species

L130: remove TMD, replace by relative, descriptive, values

No change; see above

L135: What are "segmented vertebral neural arches and centra" ?

Changed to “neural arches and centra of segmented vertebrae” (lines 140-141)

L166: L168 "compact" vs. "irregular". Partial mineralisation is not necessarily irregular.

Thanks for pointing out this issue; we changed wording, instead contrasting “non-continuous” and “continuous” mineralization patterns (lines 171-174)

L192: "several endoskeletal regions". Provide all regions

All regions provided (lines 198-199)

L269: "has never been carefully characterized in chimeras". Carefully means what? Here, also only one chimera is analyses, not several species.

Sentence removed

302: Can't believe there is no better citation for elasmobranch vertebral centra development than Gadow and Abott (1895)

Added Arriata and Kolliker REFs here (lines 293-295)

L318 ff: remove discussion from result chapter

References to paedomorphism were removed from this Results section

L342: refer to the species studied, not to the entire group.

Sorry, the line numbering for the reviewer and our original manuscript have been a little off for some reason, and we were unclear exactly to which line of text this comment referred. Generally in this revision, however, we have tried to restrict our direct analyses to the species analyzed, but in the Discussion we do extrapolate a bit from our data when considering relevant published papers of other species.

346: "selected representative". Selection criteria are missing

“selected representative” removed

L348: down tune, remove "critical"

Done

L351: down tune, remove "critical"

Done

L 364: "Since stem chondrichthyans did not typically mineralize their centra". Means there are fossil stem chondrichthyans with full mineralised centra?

Re-worded to “Stem chondrichthyans did not appear to mineralize their centra” (lines 379)

L379: down tune and change to: "we propose the term "non-tesseral trabecular mineralization. Possibly a plesiomorphic (ancestral) character of chondrichthyans"

No change; sorry, but we feel this character state needs to be emphasized as we wrote in this paper, so that its evolutionary relationship to other chondrichthyan endoskeletal features, such as tesserae, can be clarified.

L407: suggests so far palaeontologist have not been "careful" enough?

Apologies; sentence re-worded, emphasizing that synchrotron imaging might increase details of these descriptions (lines 406-408)

414: down tune, remove "we propose". Replace by "possibly" or "it can be discussed if"

Sentence re-worded and “we propose” removed (lines 412-415)

L420: remove paragraph

No action; see above

L436: remove paragraph

No action; see above

L450: perhaps add summery of the discussion. A summery that focuses on the subject and the methodology of this nice study.

Yes, in response to the reviewer’s comment, we finished the discussion with a summary of the current study. (lines 440-453)

Author response:

The following is the authors’ response to the previous reviews

Public Reviews:

Reviewer #1 (Public review):

The manuscript investigates the role of the membrane-deforming cytoskeletal regulator protein Abba in cortical development and its potential implications for microcephaly. It is a valuable contribution to the understanding of Abba's role in cortical development. The strengths and weaknesses identified in the manuscript are outlined below:

Clinical Relevance:

The authors identified a patient with microcephaly and intellectual disability patient harboring a mutation in the Abba variant (R671W), adding a clinically relevant dimension to the study.

Mechanistic Insights:

The study offers valuable mechanistic insights into the development of microcephaly by elucidating the role of Abba in radial glial cell proliferation, radial fiber organization, and the migration of neuronal progenitors. The identification of Abba's involvement in the cleavage furrow during cell division, along with its interaction with Nedd9 and positive influence on RhoA activity, adds depth to our understanding of the molecular processes governing cortical development.

In Vivo Validation:

The overexpression of mutant Abba protein (R671W), which results in phenotypic similarities to Abba knockdown effects, supports the significance of Abba in cortical development.

Weaknesses:

The findings in the study suggest that heterozygous expression of the R671W variant may exert a dominant-negative effect on ABBA's role, disrupting normal brain development and leading to microcephaly and cognitive delay. However, evidence also points to a possible gain-of-function effect, as the mutation does not decrease RhoA activity or PH3 expression in vivo. Additionally, the impact of ABBA depletion on cell fate is not fully addressed. While abnormal progenitor accumulation in the ventricular and subventricular zones is observed, the transition of progenitors to neuroblasts and their ability to support neuroblast migration remains unclear. Impaired cleavage furrow ingression and disrupted Nedd9 and RhoA signaling could lead to structural abnormalities in radial glial progenitors, affecting their scaffold function and neuroblast progression.  The manuscript lacks an exploration of the loss or decrease in interaction between Abba and NEDD9 in the case of the pathogenic patient-derived mutation in Abba. Furthermore, addressing the changes in localization and ineraction in for NEDD9 following over-expression of the mutant are important to further mehcanistically characterizxe this interaction in future studies. These gaps suggest the need for further exploration of ABBA's role in progenitor cell fate and neuroblast migration to clarify its mechanistic contributions to cortical development.

(1) Response to statement on dominant-negative vs. gain-of-function effect of R671W variant:

We appreciate the reviewer’s thoughtful analysis of the potential mechanisms underlying the R671W variant. We agree that the heterozygous expression of the human R671W mutation may initially suggest a dominant-negative effect. However, our data indicate that this variant may instead exert a gain-of-function effect. As highlighted in the discussion section, overexpression of ABBA-R671W in cells that also express wild-type ABBA did not result in a dominant-negative decrease in RhoA activation nor affect PH3 expression in vivo. These findings suggest that the R671W mutation does not impair the canonical ABBA-mediated activation of RhoA, and instead, the resulting phenotype may involve post-mitotic processes, such as altered cell migration. This interpretation is further supported by previous clinical studies reporting additional patients with the same mutation and phenotypic outcomes.

(2) Response to statement on ABBA depletion and progenitor-to-neuroblast transition:

We agree that the question of how ABBA depletion affects cell fate and the progression of radial glial progenitors (RGPs) to neuroblasts is of significant importance. Our findings suggest that ABBA knockdown disrupts cleavage furrow ingression, which may block radial glial cells prior to abscission. This likely contributes to the observed accumulation of cells in the ventricular and subventricular zones, as seen in Figures 2A and 4D. Additionally, disrupted Nedd9 expression and impaired RhoA signaling appear to alter the structural integrity of RGPs, leading to detachment of apical and basal endfeet (Supplementary Figure 3). These structural abnormalities compromise the ability of RGPs to function as scaffolds for neuroblast migration. Although direct live imaging of neuroblast migration was beyond the scope of the current dataset, we believe our evidence strongly supports a model in which ABBA depletion disrupts progenitor structure and migration. Future studies will address these transitions more directly using live imaging and fate-mapping strategies

(3) Response to statement on loss of interaction between ABBA and NEDD9 with the R671W mutation:

We fully agree with the importance of investigating whether the R671W mutation alters ABBA’s interaction with NEDD9. While our study provides evidence for a role of NEDD9 in mediating ABBA function, we acknowledge that we did not directly test whether the R671W mutation disrupts this interaction. We apologize if our manuscript conveyed the impression that this point had been fully addressed. Due to technical limitations, particularly the poor performance of anti-NEDD9 antibodies in slice immunohistochemistry, we were unable to reliably assess the interaction or localization changes in vivo. Nevertheless, this remains a priority for future studies aimed at better understanding the mechanistic underpinnings of the R671W mutation.

(4) Response to statement on future directions for mechanistic characterization of NEDD9 localization and interaction:

We agree with the reviewer that further investigation into NEDD9 localization and its interaction with the ABBA R671W mutant is essential to better define the molecular consequences of this mutation. Unfortunately, as mentioned above, the current tools available to us did not permit reliable immunohistochemical detection of NEDD9 in tissue. We fully intend to pursue alternative approaches, such as tagging strategies or the use of more sensitive detection platforms, to determine whether the R671W mutation affects the subcellular localization or stability of the ABBA-NEDD9 interaction. These experiments will be critical to elucidate the pathway through which ABBA regulates progenitor cell behavior and cortical development.

Reviewer #2 (Public review):

Summary:

Carabalona and colleagues investigated the role of the membrane-deforming cytoskeletal regulator protein Abba (MTSS1L/MTSS2) in cortical development to better understand the mechanisms of abnormal neural stem cell mitosis. The authors used short hairpin RNA targeting Abba20 with a fluorescent reporter coupled with in utero electroporation of E14 mice to show changes to neural progenitors. They performed flow cytometry for in-depth cell cycle analysis of Abba-shRNA impact to neural progenitors and determined an accumulation in S phase. Using culture rat glioma cells and live imaging from cortical organotypic slides from mice in utero electroporated with Abba-shRNA, the authors found Abba played a prominent role in cytokinesis. They then used a yeast-two-hybrid screen to identify three high confidence interactors: Beta-Trcp2, Nedd9, and Otx2. They used immunoprecipitation experiments from E18 cortical tissue coupled with C6 cells to show Abba requirement for Nedd9 localization to the cleavage furrow/cytokinetic bridge. The authors performed an shRNA knockdown of Nedd9 by in utero electroporation of E14 mice and observed similar results as with the Abba-shRNA. They tested a human variant of Abba using in utero electroporation of cDNA and found disorganized radial glial fibers and misplaced, multipolar neurons, but lacked the impact of cell division seen in the shRNA-Abba model.

Strengths:

Fundamental question in biology about the mechanics of neural stem cell division.

Directly connecting effects in Abba protein to downstream regulation of RhoA via Nedd9.

Incorporation of human mutation in ABBA gene.

Use of novel technologies in neurodevelopment and imaging.

Weaknesses:

Unexplored components of the pathway (such as what neurogenic populations are impacted by Abba mutation) and unleveraged aspects of their data (such as the live imaging) limit the scope of their findings and left significant questions about the effect of ABBA on radial glia development.

(1) Claim of disorganized radial glial fibers lacks quantifications.

- On page 11, the authors claim that knockdown of Abba lead to changes in radial glial morphology observed with vimentin staining. Here they claim misoriented apical processes, detached end feet, and decreased number of RGP cells in the VZ. However, they no not provide quantification of process orientation to better support their first claim. Measurements of radial glia fiber morphology (directionality, length) and of angle of division would be metrics that can be applied to data. Some of these analysis could be done in their time-lapse microscopy images, such as to quantify the number of cell division during their period of analysis (though that is short-15 hours).

Response to: Lack of quantification of disorganized radial glial fibers and cell divisions in time-lapse data

We appreciate the reviewer’s insightful comment regarding the need for quantification of radial glial (RG) fiber morphology. In the revised manuscript, we have addressed this by providing new quantification of changes in vimentin staining, specifically measuring the dispersion of the signal as a proxy for fiber disorganization (see Supplementary Figure 1). These data support the observed morphological changes, including misoriented apical processes and detachment of endfeet, following Abba knockdown.

Regarding time-lapse analysis to track cell divisions, we attempted to follow individual cells during the 15-hour imaging window. However, due to the relatively short duration and limited number of cells that could be reliably tracked, the dataset did not allow for statistically meaningful conclusions. As an alternative approach, we performed live-cell imaging using Anillin-GFP, a reliable marker of mitotic progression. The distribution and accumulation of Anillin-GFP were analyzed in ABBA-shRNA3 and control conditions, and the results (now included in Supplementary Figure 3) indicate an increased number of cells arrested in late mitosis upon ABBA knockdown. This supports the notion of disrupted cytokinesis as a consequence of Abba depletion.

(2) Unclear where effect is:

- In RG or neuroblasts? Is it in cell cleavage that results in accumulation of cells at VZ (as sometimes indicated by their data like in Fig 2A or 4D)? Interrogation of cell death (such as by cleaved caspase 3) would also help. Given their time lapse, can they identify what is happening to the RG fiber? The authors describe a change in "migration" but do not show evidence for this for either progenitor or neuroblast populations. Given they have nice time-lapse imaging data, could they visualize progenitor versus young neuron migration? Analysis of neuroblasts (such as with doublecortin expression in the tissue) would also help understand any issues in migration (of neurons v stem cells).

- At cleaveage furrow? In abscission? There is high resolution data that highlights the cleavage furrow as the location of interest (fig 3A), however there is also data (fig 3B) to suggest Abba is expressed elsewhere as well and there is an overall soma decrease. More detail of the localization of Abba during the division process would be helpful-for example, could cleavage furrow proteins, such as Aurora B, co-localization (and potentially co-IP) help delineate subpopulations of Abba protein? Furthermore, the FRET imaging is unique way to connect their mutation with function-could they measure/quantify differences at furrow compared to rest of soma to further corroborate that Abba-associated RhoA effect was furrow-enriched?

- The data highlights nicely that a furrow doesn't clearly form when ABBA expression and subsequent RhoA activity are decreased (in Fig 3 or 5A). Does this lead to cells that can't divide because of poor abscission, especially since "rounding" still occurs? Or abnormal progenitors (with loss of fiber or inability to support neuroblast migration)? Or abnormal progression of progenitors to neuroblasts?

Response to: Unclear location of the effect (RG vs. neuroblasts; cleavage furrow/abscission; migration issues)

We thank the reviewer for this comprehensive and thought-provoking set of questions.

a) Site of the effect – Radial Glia vs. Neuroblasts:

Our data suggest that the primary effect of ABBA depletion occurs in radial glial progenitors (RGPs), specifically prior to abscission. We observed accumulation of electroporated cells in the ventricular zone (VZ), which we interpret as a result of cytokinetic failure (e.g., Figure 2A, 4D). We also documented detachment of apical and basal endfeet (see Supplementary Figure 3), further supporting structural disruption of RG fibers.

b) Cell death analysis:

We considered using cleaved caspase-3 as a marker for apoptosis, but due to its transient and non-specific activation during development, we opted to assess overall survival via quantification of RGP cell numbers and localization. This approach better reflects the developmental impact of ABBA knockdown (Supplementary Figure 3).

c) Migration defects:

We agree that distinguishing between progenitor and neuroblast migration would be highly informative. Although we did not perform doublecortin or similar staining to differentiate these populations in this dataset, the accumulation of electroporated cells in VZ/SVZ strongly suggests a migration deficit. Addressing this in detail will require new experiments using lineage-specific markers and longer time-lapse recordings, which we plan to explore in future studies.

d) Cleavage furrow and abscission:

Our high-resolution imaging of Anillin-GFP and FRET-based RhoA activity shows that ABBA localizes predominantly at the cleavage furrow. New quantifications of RhoA activity (now in Figure 5) show that the reduction in signaling is most pronounced at the furrow in ABBA knockdown cells. These findings align with the hypothesis that ABBA, through Nedd9 and RhoA, is essential for proper furrow formation and abscission.

e) Mechanistic implications:

As the reviewer notes, ABBA knockdown leads to cells that "round" but do not complete division, likely due to poor cleavage furrow ingression. This could generate abnormal progenitors that are structurally compromised (detached fibers) and thus unable to support neuroblast migration or proper differentiation. The cumulative result is disrupted progression from RGPs to neuroblasts, impaired structural scaffolding, and possibly reduced cell viability.

(3) Limited to a singular time point of mouse cortical development

On page 13, the authors outline the results of their Y2H screen with the identification of three high confidence interactors. Notably, they used a E10.5-E12.5 mouse brain embryo library rather than one that includes E14, the age of their in utero electroporation mice. Many of the authors' claims focus on in utero electroporation of shRNA-Abba of E14 mice that are then evaluated at E16-18. Justification for the focus on this age range should be included to support that their findings can then be applied to all of mouse corticogenesis.

Response to: Use of E10.5–E12.5 library for yeast-two-hybrid (Y2H) screen

We appreciate the reviewer’s concern regarding the developmental stage of the Y2H library. We chose the E10.5–E12.5 brain embryo library based on prior work demonstrating that ABBA expression is strongest during early cortical development, particularly in radial glia at these stages (see Saarikangas et al., J Cell Sci 2008). The radial glia-specific expression of ABBA was previously validated using RC2 and Tuj1 markers at E12.5. Thus, the library we used is well-suited for identifying interactors relevant to radial glial function, including Nedd9. We have clarified this rationale in the revised manuscript.

(4) Detail of the effect of the human variant of the ABBA mutation in mouse is lacking.

Their identification of the R671W mutation is interesting and the IUE model warrants more characterization, as they did with their original KD experiments.

- Could they show that Abba protein levels are decreased (in either cell lines or electroporated tissue)?

- While time-lapse morphology might not have been performed, more analysis on cell division phenotype (such as plane of division and radial glia morphology) would be helpful.

Response to: Lack of detail on R671W human variant effects

We thank the reviewer for encouraging further characterization of the R671W variant. In the revised manuscript, we now provide additional data on interkinetic nuclear migration (INM) defects resulting from R671W overexpression (see Supplementary Figure 3). These changes are consistent with disrupted radial glial organization and mirror aspects of the ABBA knockdown phenotype.

a) Protein levels:

We quantified ABBA expression in cells overexpressing the R671W variant (Supplementary Figure 5) and found no significant reduction compared to wild-type. This argues against a loss-of-function mechanism and supports a gain-of-function or dominant-interfering effect.

b) Morphological and division phenotyping:

While time-lapse imaging of R671W-expressing cells was not available in our dataset, we acknowledge that analyses such as division angle or radial glial morphology would be informative. Unfortunately, we were unable to perform these with the current data, but we agree these are important goals for future work.

Reviewer 2 conclusion:

The resubmission has addressed many of the questions raised.

I have a few comments that should be addressed:

(1) The authors maintain a deficit in "migration of immature neurons" which remains unsubstantiated. In their resonse, they state: "we believe that the data showing the accumulation of migrating electroporated cells in the ventricular (V) and subventricular (SV) zones provide compelling evidence of abnormal migration in ABBA-shRNA electroporated cells. "

- Firstly, they do not demonstrate that it's immature neurons, not RGs, that are affected. Secondly, accumulation of cells at the V-SVZ could be due to soley the inability for the RGC to undergo mitosis, therefore remaining stuck"

The commentary of migration, especially of neurons, should be modified.

We appreciate the reviewer’s careful reading and valid concern regarding our use of the term "migration of immature neurons." We fully agree that the current dataset does not definitively distinguish whether the accumulated cells in the ventricular (V) and subventricular (SV) zones are immature neurons or radial glial progenitors (RGPs) arrested in mitosis.

To clarify, our observations indicate that electroporated cells accumulate in the VZ/SVZ following ABBA knockdown (Figures 2A and 4D), and this was interpreted as evidence of impaired migration. However, we now recognize that this accumulation may primarily reflect a block in cell cycle progression—specifically, at the stage of cleavage furrow ingression and abscission—rather than a migratory defect per se. This is supported by our new data using Anillin-GFP (Supplementary Figure 3), which show increased accumulation of cells with persistent Anillin expression, consistent with mitotic arrest. Furthermore, the detachment of apical and basal processes (also shown in Supplementary Figure 3) suggests that ABBA knockdown affects the structural integrity of RGPs, potentially compromising their scaffold function.

In light of these points, we have revised the manuscript to temper our conclusions regarding “migration defects.” Specifically, we now refer to the phenotype as “abnormal accumulation of progenitor cells” and clarify that, while these findings are consistent with impaired cell progression or scaffolding required for migration, we do not directly demonstrate impaired migration of immature neurons. As suggested, addressing this would require additional analyses, such as time-lapse imaging of post-mitotic cells or staining with markers like Doublecortin, which are beyond the scope of the current dataset but will be a focus of future investigations.

We thank the reviewer again for encouraging a more precise interpretation of our findings

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Supplementary Fig 4B - The figure doesn't show an increase in percentage of PH3 positive cells in the NEDD9-shRNA condition. The control images are also missing for comparison. The figure legend needs to be corrected to match with the figure showing no significant changes.

Thank you for this comment. This has been amended in the revised manuscript in the form of a new revised Supplementary Fig 4.

Reviewer #2 (Recommendations for the authors):

Minor annotations for slice culture assay

The authors should make note of ages of slice cultures in text and have better annotations of slice cultures (for example, in Fig 4-where is mitosis?)

We are sorry for the mistake it's not mitosis, it's the cleavage furrow stage.  In addition, a new amended Figure 4 is provided. 

The effects are hard to see in lower mag slice images in Fig. 6. Would recommend focusing on higher mag to highlight RG differences.

Thank you for this comment. This has been amended in the revised manuscript in the form of a new revised Figure 6.

Author response:

The following is the authors’ response to the original reviews

Reviewer #1:

(1) The mechanism by which fenofibrate rescues memory loss in Kallistatin-transgenic mice is unclear. As a PPARalpha agonist, does fenofibrate target the Kallistatin pathway directly or indirectly? Please provide a discussion based on literature supporting either possibility.

Thank you for your important suggestion. Fenofibrate is indeed acting as a PPARα agonist. Fenofibrate has been shown to protect memory and cognitive function by downregulating α- and β-secretases[1]. Activation of PPARα can reduce Aβ plaques by upregulating ADAM10, thereby protecting memory and cognition[2]. Whereas, Fenofibrate can also act through a PPARα-independent pathway[3]. In our previous study, we proved that Fenofibrate can directly down-regulate the expression of Kallistatin in hepatocytes[4]. Here, our findings showed that Kallistatin induces cognitive memory deterioration by increasing amyloid-β plaques accumulation and tau protein hyperphosphorylation (Fig. 1-3), and Fenofibrate can directly down-regulate the serum level of Kallistatin (Fig. 8G). In addition, the expression of PPARα in the hippocampal tissue of Kallistatin (KAL-TG) mice showed no significant difference compared to the WT group (Author response image 1A-B). Therefore, we think Fenofibrate may improve memory and cognitive function at least in part through a PPARα-independent effect, which provides a new mechanism of Fenofibrate in AD with elevated Kallistatin levels.

Author response image 1.

(A-B) Protein levels of PPARα were tested by western blot analysis in hippocampal tissue, then statistically analyzed the above results.

(2) The current study exclusively investigated the hippocampus. What about other cognitive memory-related regions, such as the prefrontal cortex? Including data from these regions or discussing the possibility of their involvement could provide a more comprehensive understanding of the role of Kallistatin in memory impairment.

Thank you for your suggestion. In addition to hippocampal tissue analysis, we performed immunohistochemical detection of Aβ and phosphorylated Tau levels in the prefrontal cortex. Our findings revealed that KAL-TG mice exhibited significantly elevated Aβ and phosphorylated Tau levels in the prefrontal cortex compared to WT mice. These observations align with the pathological patterns observed in hippocampal tissues, demonstrating consistent neurodegenerative pathology across both the hippocampus and prefrontal cortex. The data for this part are seen as follows.

Author response image 2.

(A-B) Immunofluorescence staining of Aβ and phosphorylated tau (p-tau T231) was carried out in the prefrontal cortex tissue of KAL-TG and WT mice. Error bars represented the Standard Error of Mean (SEM); **p < 0.01. Scale bar, 100 μm.

(3) Fenofibrate rescued phenotypes in Kallistatin-transgenic mice while rosiglitazone, a PPARgamma agonist, did not. This result contradicts the manuscript's emphasis on a PPARgamma-associated mechanism. Please address this inconsistency.

Thank you for the reminder. In fact, our results showed a trend towards improved memory and cognitive function in KAL-TG mice treated with Rosiglitazone, although its effect is not as significant as that of Fenofibrate. Several studies have reported that Rosiglitazone has a beneficial effect on memory and cognitive function in mouse models of dementia, while these studies involve treatment periods of 3 to 4 months[5, 6], whereas our treatment period was only one month. Extending the treatment period with Rosiglitazone may result in a more pronounced improvement. In addition, Fenofibrate may have a PPAR-independent pathway by downregulating Kallistatin directly as discussed above and then show stronger effects.

(4) Most of the immunohistochemistry images are unclear. Inserts have similar magnification to the original representative images, making judgments difficult. Please provide larger inserts with higher resolution.

According to your suggestion, we provided larger inserts with higher resolution in Fig 3A and Fig 4B, as follows:

(5) The immunohistochemistry images in different figures were taken from different hippocampal subregions with different magnifications. Please maintain consistency, or explain why CA1, CA3, or DG was analyzed in each experiment.

Thank you for your advice. The trends of changes in different brain regions(including CA1, CA3, or DG) are consistent. Following your suggestion, we have now selected the DG region replaced the different hippocampal subregions with the DG area, and re-conducted the statistical analysis in Fig 5I & 6C, as follows. Due to the significant deposition of Aβ only in the CA1 region, Fig 2A was not replaced.

(6) Figure 5B is missing a title. Please add a title to maintain consistency with other graphs.

Thanks for your suggestion. We have added a title to Figure 5B, as follows:

(7) Please list statistical methods used in the figure legends, such as t-test or One-way ANOVA with post-hoc tests.

Thanks for your suggestion. We have listed the statistical methods used in the figure legends.

Reviewer #2:

(1) It was suggested that Kallistatin is primarily produced by the liver. The study demonstrates increased Kallistatin levels in the hippocampus tissue of AD mice. It would be valuable to clarify if Kallistatin is also increased in the liver of AD mice, providing a comprehensive understanding of its distribution in disease states.

Thank you for your suggestion. We extracted liver tissue from APP/PS1 mice, and the Western blot results indicated that the expression of Kallistatin in the liver of APP/PS1 mice was elevated, as follows:

Author response image 3.

(A-B) Protein levels of Kallistatin were tested by western blot analysis in the liver tissue, then statistically analyzed the above results. Error bars represented the Standard Error of Mean (SEM); **p < 0.01.

(2) Does Kallistatin interact directly with Notch1 ligands? Clarifying this interaction mechanism would enhance understanding of how Kallistatin influences Notch1 signaling in AD pathology.

Thank you for your suggestion. This study reveals that Kallistatin directly binds to Notch1 and contributes to the activation of the Noch1-HES1 signaling pathway. As for whether Kallistatin can bind to the ligands of Notch1, it needs to conduct further investigations in future studies. Our preliminary data showed that Jagged1 was upregulated in the hippocampal tissues of KAL-TG mice by qPCR and Western blot analyses.

Author response image 4.

Kallistatin promoted Notch ligand Jagged1 expression to activate Notch1 signaling. (A) QPCR analysis of Notch ligands (Dll1, Dll3, Jagged1, Jagged2) expression in the 9 months hippocampus tissue. (B) Western blotting analysis of Notch ligand Jagged1 expression in the hippocampus tissue. (C) Western blotting analysis of Notch ligand Jagged1 expression in the hippocampus primary neuron. β-actin served as the loading control. Error bars represented the Standard Error of Mean (SEM); *p < 0.05.

(3) Is there any observed difference in AD phenotype between male and female Kallistatin-transgenic (KAL-TG) mice? Including this information would address potential gender-specific effects on cognitive decline and pathology.

Thank you for your suggestion. Actually, we have previously used female mice for Morris Water Maze experiments, and the results showed that both male and female KAL-TG mice exhibited a phenotype of decreased memory and cognitive function compared to the gender-matched WT group, while there was no significant difference between male and female KAL-TG mice as follows:

Author response image 5.

(A-D) Behavioral performance was assessed through the Morris water maze test. (A) The escape latency time was presented during 1-5 days. (B-D) Cognitive functions were evaluated by spatial probe test on day 6, then analyzing each group of mice crossing platform times(B), time percent in the targeted area (C), and the path traces heatmap (D). Error bars represented the Standard Error of Mean (SEM); F represents Female, M represents Male, and TG refers to KAL-TG; *p < 0.05.

(4) It is recommended to include molecular size markers in Western blots for clarity and accuracy in protein size determination.

Thank you for your reminder. We have shown the molecular weight of each bolt.

(5) The language should be revised for enhanced readability and clarity, ensuring that complex scientific concepts are communicated effectively to a broader audience.

According to your suggestion, we have polished the article for enhancing readability and clarity.

Reviewer #3:

(1) The authors did not illustrate whether the protective effect of fenofibrate against AD depends on Kallistatin.

Thank you for your important suggestion. Fenofibrate is indeed acting as a PPARα agonist. Fenofibrate has been shown to protect memory and cognitive function by downregulating α- and β-secretases[1]. Activation of PPARα can reduce Aβ plaques by upregulating ADAM10, thereby protecting memory and cognition[2]. Whereas, Fenofibrate can also act through a PPARα-independent pathway[3]. In our previous study，we proved Fenofibrate can directly down-regulate the expression of KAL in hepatocytes[4]. Here, our findings showed that Kallistatin induces cognitive memory deterioration by increasing amyloid-β plaques accumulation and tau protein hyperphosphorylation (Fig. 1-3), and Fenofibrate can directly down-regulate the serum level of Kallistatin (Fig. 8G). In addition, the expression of PPARα in the hippocampal tissue of Kallistatin (KAL-TG) mice showed no significant difference compared to the WT group (Author response image 1-B). Therefore, we think Fenofibrate may improve memory and cognitive function at least in part through downregulatin Kallistatin. To conclusively determine whether fenofibrate’s therapeutic effects depend on Kallistatin, future studies should employ Kallistatin-knockout AD animal models to evaluate fenofibrate’s impact on cognitive and memory functions. These investigations will further clarify the mechanistic underpinnings of fenofibrate in AD therapy.

(2) The conclusions are supported by the results, but the quality of some results should be improved.

Thank you for your kind suggestion. We have updated the magnified images in the immunohistochemistry section of the article, ensuring that the fields of view for the immunohistochemistry are within the same brain region, and have shown the molecular weights in each bolt. Additionally, we have conducted a quantitative analysis of the protein levels in the Western blot results presented in Fig6&8.

(3) Figures 2c, 3c, and 4a present the Western blot results of p-tau from mice of different ages on one membrane, showing age-dependent expression. The authors analyzed the results of mice of different ages in one statistical chart, which will create ambiguity with the results of the representative images. For example, the expression of p-tau 396 in the blot was lower in the WT-12 M group than in the WT-9 M group (Figure 3c), which is contradictory to the statistical analysis.

Thank you for your reminder. The statistical presentation here does not match the figure. At that time, the WB experiments for the hippocampal tissue at each age group were conducted separately, and it was not appropriate to compare different age groups together. This graph cannot illustrate age dependency. We have replaced the statistical graph in Figure 3B&D, as follows:

(4) Figure 4b shows that KAL-TG-9 M had greater BACE1 expression than KAL-TG-12 M. Furthermore, the nuclei are not uniformly colored. Please provide more representative figures.

Thank you for your reminder. Due to the fact that these sets of data were not processed in a single batch, the ages in the graph are not comparable. Regarding the issue of inconsistent nuclear staining, we have provided another representative image from this group, as follows:

(5) Unclear why the BACE1 and Aβ levels seems less with KAL+shHES1 treatment than GFP+shNC treatment (Fig 6H)? This finding contradicts the conclusion.

Thank you for your reminder. This experiment was repeated three times, and here, we have represented the representative results along with the corresponding statistical data. There are no difference between KAL+shHES1 treatment and GFP+shNC treatment. We have updated the Fig. 6H.

(6) The Western blot results in figure 6e-h, 8h-i, and S3-S5 were not quantified.

Thank you for your reminder. We have added statistical graphs and original images of the pictures in figure 6e-h, 8h-i, and S3-S5.

(7) The authors did not provide the detection range of the Aβ42 ELISA kit.

Thank you for your suggestion. The Aβ42 ELISA kit is from the IBL, with the product number 27721. Its standard range is 1.56 - 100 pg/mL, and the sensitivity is 0.05 pg/mL.

(8)The authors did not specify the sex of the mice. This is important since sex could have had a dramatic impact on the results.

Thank you for your suggestion. The results we present in the text are all statistically obtained from male mice. Actually, we have previously used female mice for Morris Water Maze experiments, and the results showed that both male and female KAL-TG mice exhibited a phenotype of decreased memory and cognitive function compared to the gender-matched WT group, while there was no significant difference between male and female KAL-TG mice (Author response image 5).

Minor:

(1) In Figure 2b, there are no units for the vertical coordinates of the statistical graph.

Thank you for your reminder. We have added units for the vertical coordinates in Figure 2b.

(2) In Figure 2c, the left Y-axis title is lacking in the statistic chart.

Thank you for your reminder. We have added the left Y-axis title in the statistic chart.

Reference：

(1) Assaf N, El-Shamarka ME, Salem NA, Khadrawy YA, El Sayed NS. Neuroprotective effect of PPAR alpha and gamma agonists in a mouse model of amyloidogenesis through modulation of the Wnt/beta catenin pathway via targeting alpha- and beta-secretases. Progress in Neuro-Psychopharmacology and Biological Psychiatry 2020, 97: 109793.

(2) Rangasamy SB, Jana M, Dasarathi S, Kundu M, Pahan K. Treadmill workout activates PPARα in the hippocampus to upregulate ADAM10, decrease plaques and improve cognitive functions in 5XFAD mouse model of Alzheimer’s disease. Brain, Behavior, and Immunity 2023, 109: 204-218.

(3) Yuan J, Tan JTM, Rajamani K, Solly EL, King EJ, Lecce L, et al. Fenofibrate Rescues Diabetes-Related Impairment of Ischemia-Mediated Angiogenesis by PPARα-Independent Modulation of Thioredoxin-Interacting Protein. Diabetes 2019, 68(5): 1040-1053.

(4) Fang Z, Shen G, Wang Y, Hong F, Tang X, Zeng Y, et al. Elevated Kallistatin promotes the occurrence and progression of non-alcoholic fatty liver disease. Signal Transduct Target Ther 2024, 9(1): 66.

(5) Nelson ML, Pfeifer JA, Hickey JP, Collins AE, Kalisch BE. Exploring Rosiglitazone's Potential to Treat Alzheimer's Disease through the Modulation of Brain-Derived Neurotrophic Factor. Biology (Basel) 2023, 12(7).

(6) Pedersen WA, McMillan PJ, Kulstad JJ, Leverenz JB, Craft S, Haynatzki GR. Rosiglitazone attenuates learning and memory deficits in Tg2576 Alzheimer mice. Exp Neurol 2006, 199(2): 265-273.

Author response:

The following is the authors’ response to the previous reviews

Public Reviews: 

Reviewer #1 (Public review): 

The authors investigated the role of the C. elegans Flower protein, FLWR-1, in synaptic transmission, vesicle recycling, and neuronal excitability. They confirmed that FLWR-1 localizes to synaptic vesicles and the plasma membrane and facilitates synaptic vesicle recycling at neuromuscular junctions. They observed that hyperstimulation results in endosome accumulation in flwr-1 mutant synapses, suggesting that FLWR-1 facilitates the breakdown of endocytic endosomes. Using tissue-specific rescue experiments, the authors showed that expressing FLWR-1 in GABAergic neurons restored the aldicarb-resistant phenotype of flwr-1 mutants to wild-type levels. By contrast, cholinergic neuron expression did not rescue aldicarb sensitivity at all. They also showed that FLWR-1 removal leads to increased Ca²⁺ signaling in motor neurons upon photo-stimulation. From these findings, the authors conclude that FLWR-1 helps maintain the balance between excitation and inhibition (E/I) by preferentially regulating GABAergic neuronal excitability in a cell-autonomous manner. 

Overall, the work presents solid data and interesting findings, however the proposed cell-autonomous model of GABAergic FLWR-1 function may be overly simplified in my opinion. 

Most of my previous comments have been addressed; however, two issues remain. 

(1) I appreciate the authors' efforts conducting additional aldicarb sensitivity assays that combine muscle-specific rescue with either cholinergic or GABergic neuron-specific expression of FLWR-1. In the revised manuscript, they conclude, "This did not show any additive effects to the pure neuronal rescues, thus FLWR-1 effects on muscle cell responses to cholinergic agonists must be cellautonomous." However, I find this interpretation confusing for the reasons outlined below. 

Figure 1 - Figure Supplement 3B shows that muscle-specific FLWR-1 expression in flwr-1 mutants significantly restores aldicarb sensitivity. However, when FLWR-1 is co-expressed in both cholinergic neurons and muscle, the worms behave like flwr-1 mutants and no rescue is observed. Similarly, cholinergic FLWR-1 alone fails to restore aldicarb sensitivity (shown in the previous manuscript).

This data is still shown in the manuscript, Fig. 3D. We interpreted our finding in the muscle/cholinergic co-rescue experiment as meaning, that FLWR-1 in cholinergic neurons over-compensates, so worms should be resistant, and the rescuing effect of muscle FLWR-1 is therefore cancelled. But it is true, if this were the case, why does the pure cholinergic rescue not show over-compensation? We added a sentence to acknowledge this inconsistency and we added a sentence in the discussion (see also below, comment 1) of reviewer #2).

These observations indicate a non-cell-autonomous interaction between cholinergic neurons and muscle, rather than a strictly muscle cell-autonomous mechanism. In other words, FLWR-1 expressed in cholinergic neurons appears to negate or block the rescue effect of muscle-expressed FLWR-1. Therefore, FLWR-1 could play a more complex role in coordinating physiology across different tissues. This complexity may affect interpretations of Ca²⁺ dynamics and/or functional data, particularly in relation to E/I balance, and thus warrants careful discussion or further investigation. 

For the Ca²⁺ dynamics, we think the effects of flwr-1 are likely very immediate, as the imaging assay relies on a sensor expressed directly in the neurons or muscles under study, and not on indirect phenotypes as muscle contraction and behavior, that depend on an interplay of several cell types influencing each other.

(2) The revised manuscript includes new GCaMP analyses restricted to synaptic puncta. The authors mention that "we compared Ca²⁺ signals in synaptic puncta versus axon shafts, and did not find any differences," concluding that "FLWR-1's impact is local, in synaptic boutons." This is puzzling: the similarity of Ca²⁺ signals in synaptic regions and axon shafts seems to indicate a more global effect on Ca²⁺ dynamics or may simply reflect limited temporal resolution in distinguishing local from global signals due to rapid Ca²⁺ diffusion. The authors should clarify how they reached the conclusion that FLWR-1 has a localized impact at synaptic boutons, given that synaptic and axonal signals appear similar. Based on the presented data, the evidence supporting a local effect of FLWR-1 on Ca²⁺ dynamics appears limited.

We apologize, here we simply overlooked this misleading wording in our rebuttal letter. The data we mentioned, showing no obvious difference in axon vs. bouton, are shown below, including time constants for the onset and the offset of the stimulus (data is peak normalized for better visualization):

Author response image 1.

One can see that axonal Ca²⁺ signals may rise a bit slower than synaptic Ca²⁺ signals, as expected for Ca²⁺ entering the boutons, and then diffusing out into the axon. The loss of FLWR1 does not affect this. However, the temporal resolution of the used GCaMP6f sensor is ca. 200 ms to reach peak, and the decay time (to t1/2) is ca. 400 ms (PMID: 23868258). Thus, it would be difficult to see effects based on Ca²⁺ diffusion using this assay. For the decay, this is similar for both axon and synapse, while flwr-1 mutants do not reduce Ca²⁺ as much as wt. In the axon, there is a seemingly slightly slower reduction in flwr-1 mutants, however, given the kinetics of the sensor, this is likely not a meaningful difference. Therefore, we wrote we did not find differences. The interpretation should not have been that the impact of FLWR-1 is local. It may be true if one could image this at faster time scales, i.e. if there is more FLWR-1 localized in boutons (as indicated by our data showing FLWR-1 enrichment in boutons; Fig. 3), and when considering its possible effect on MCA-3 localization (and assuming that MCA-3 is the active player in Ca²⁺ removal), i.e. FLWR-1 recruiting MCA-3 to boutons (Fig. 9C, D).  

Reviewer #2 (Public review): 

Summary: 

The Flower protein is expressed in various cell types, including neurons. Previous studies in flies have proposed that Flower plays a role in neuronal endocytosis by functioning as a Ca²⁺ channel. However, its precise physiological roles and molecular mechanisms in neurons remain largely unclear. This study employs C. elegans as a model to explore the function and mechanism of FLWR-1, the C. elegans homolog of Flower. This study offers intriguing observations that could potentially challenge or expand our current understanding of the Flower protein. Nevertheless, further clarification or additional experiments are required to substantiate the study's conclusions. 

Strengths: 

A range of approaches was employed, including the use of a flwr-1 knockout strain, assessment of cholinergic synaptic activity via analyzing aldicarb (a cholinesterase inhibitor) sensitivity, imaging Ca²⁺ dynamics with GCaMP3, analyzing pHluorin fluorescence, examination of presynaptic ultrastructure by EM, and recording postsynaptic currents at the neuromuscular junction. The findings include notable observations on the effects of flwr-1 knockout, such as increased Ca²⁺ levels in motor neurons, changes in endosome numbers in motor neurons, altered aldicarb sensitivity, and potential involvement of a Ca²⁺-ATPase and PIP2 binding in FLWR-1's function. 

The authors have adequately addressed most of my previous concerns, however, I recommend minor revisions to further strengthen the study's rigor and interpretation: 

Major suggestions 

(1) This study relies heavily on aldicarb assays to support its conclusions. While these assays are valuable, their results may not fully align with direct assessment of neurotransmitter release from motor neurons. For instance, prior work has shown that two presynaptic modulators identified through aldicarb sensitivity assays exhibited no corresponding electrophysiological defects at the neuromuscular junction (Liu et al., J Neurosci 27: 10404-10413, 2007). Similarly, at least one study from the Kaplan lab has noted discrepancies between aldicarb assays and electrophysiological analyses. The authors should consider adding a few sentences in the Discussion to acknowledge this limitation and the potential caveats of using aldicarb assays, especially since some of the aldicarb assay results in this study are not easily interpretable. 

Aldicarb assays have been used very successfully in identifying mutants with defects in chemical synaptic transmission, and entire genetic screens have been conducted this way. The reviewer is right, one needs to realize that it is the balance of excitation and inhibition at the NMJ of C. elegans, which underlies the effects on the rate of aldicarb-induced paralysis, not just cholinergic transmission. I.e. if a given mutant affects cholinergic and GABAergic transmission differently, things become difficult to interpret, particularly if also muscle physiology is affected. Therefore, we combined mutant analyses with cell-type specific rescue. We acknowledge that results are nonetheless difficult to interpret. We thus added a sentence in the first paragraph of the discussion.

(2) The manuscript states, "Elevated Ca²⁺ levels were not further enhanced in a flwr-1;mca-3 double mutant." (lines 549-550). However, Figure 7C does not include statistical comparisons between the single and double mutants of flwr-1 and mca-3. Please add the necessary statistical analysis to support this statement. 

Because we only marked significant differences in that figure, and n.s. was not shown. This was stated in the figure legend.

(3) The term "Ca²⁺ influx" should be avoided, as this study does not provide direct evidence (e.g. voltage-clamp recordings of Ca²⁺ inward currents in motor neurons) for an effect of the flwr-1 mutation of Ca²⁺ influx. The observed increase in neuronal GCaMP signals in response to optogenetic activation of ChR2 may result from, or be influenced by, Ca²⁺ mobilization from of intracellular stores. For example, optogenetic stimulation could trigger ryanodine receptor-mediated Ca²⁺ release from the ER via calcium-induced calcium release (CICR) or depolarization-induced calcium release (DICR). It would be more appropriate to describe the observed increase in Ca²⁺ signal as "Ca²⁺ elevation" rather than increased "Ca²⁺ influx". 

Ok, yes, we can do this, we referred by ‘influx’ to cytosolic Ca²⁺, that fluxes into the cytosol, be it from the internal stores or the extracellular. Extracellular influx, more or less, inevitably will trigger further influx from internal stores, to our understanding. We changed this to “elevated Ca²⁺ levels” or “Ca²⁺ level rise” or “Ca²⁺ level increase”.

Recommendations for the authors: 

Reviewer #1 (Recommendations for the authors):

A thorough discussion on the impact of cell-autonomous versus non-cell-autonomous effects is necessary. 

Revise and clarify the distinction between local and global Ca²⁺ changes. 

see above.

Reviewer #2 (Recommendations for the authors): 

Minor suggestions 

(1) In "Few-Ubi was shown to facilitate recovery of neurons following intense synaptic activity (Yao et al.,....." (lines 283-284), please specify which aspects of neuronal recovery are influenced by the Flower protein. 

We added “refilling of SV pools”.

(2) The abbreviation "Few-Ubi" is used for the Drosophila Flower protein (e.g., line 283, Figure 1A, and Figure 8A). Please clarify what "Ubi" stands for and verify whether its inclusion in the protein name is appropriate.

This is inconsistent across the literature, sometimes Fwe-Ubi is also referred to as FweA. We now added this term. Ubi refers to ubiquitous (“Therefore, we named this isoform fweubi because it is expressed ubiquitously in imaginal discs“) (Rhiner 2010)

(3) The manuscript uses "pflwr-1" (line 303 and elsewhere) to denote the flwr-1 promoter. This notation could be misleading, as it may be interpreted as a gene name. Please consider using either "flwr-1p" or "Pflwr-1" instead. Additionally, ensure proper italicization of gene names throughout the manuscript. 

We changed this throughout. We will change to italicized at proof stage, it would be too timeconsuming to spot these incidents now.

(4) The authors tagged the C-terminus of FLWR-1 by GFP (lines 321). The fusion protein is referred to as "GFP::FLWR-1" throughout the manuscript. Please verify whether "FLWR-1::GFP" would be the more appropriate designation.

Thank you, yes, we changed this in the text, GFP is indeed N-terminal.

(5) In "This did not show any additive effects...." (line 363), please clarify what "This" refers to. 

Altered to “The combined rescues did not show any additive effects…”

(6) In "..., supporting our previous finding of increased neurotransmitter release in GABAergic neurons" (lines 412-413), please provide a citation for the referenced previous study.

This refers to our aldicarb data within this paper, just further up in the text. We removed “previous”.

(7) Figure 4C, D examines the effect of flwr-1 mutation on body length in the genetic background of the unc-29 mutation, which selectively disrupts the levamisole-sensitive acetylcholine receptor. Please comment on the rationale for implicating only the levamisole receptor rather than the nicotinic acetylcholine receptor in muscle cells. 

This was because we used a behavioral assay. Despite the fact that the homopentameric ACR16/N-AChR mediate about 2/3 of the peak currents in response to acute ACh application to the NMJ (e.g. Almedom et al., EMBO J, 2009), the acr-16 mutant has virtually no behavioral / locomotion phenotype. Likely, this is because the heteropentameric, UNC-29 containing LAChR, while only contributing 1/3 of the peak current, desensitizes much more slowly and thus unc-29 mutants show a severe behavioral phenotype (uncoordinated locomotion, etc.). We thus did not expect a major effect when performing the behavoral assay in acr-16 mutants and thus chose the unc-29 mutant background.

(8) In "we found no evidence ....insertion into the PM (Yao et al., 2009)", It appears that the cited paper was not authored by any of the current manuscript. Please confirm whether this citation is correctly attributed. 

This sentence was arranged in a misleading way, we did not mean that we authored this paper. It was change in the text: “While a facilitating role of Flower in endocytosis appears to be conserved in C. elegans, in contrast to previous findings from Drosophila (Yao et al., 2009), we found no evidence that FLWR-1 conducts Ca²⁺ upon insertion into the PM.”

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

In this study, Xiao et al. classified retroperitoneal liposarcoma (RPLS) patients into two subgroups based on whole transcriptome sequencing of 88 patients. The G1 group was characterized by active metabolism, while the G2 group exhibited high scores in cell cycle regulation and DNA damage repair. The G2 group also displayed more aggressive molecular features and had worse clinical outcomes compared to G1. Using a machine learning model, the authors simplified the classification system, identifying LEP and PTTG1 as the key molecular markers distinguishing the two RPLS subgroups. Finally, they validated these markers in a larger cohort of 241 RPLS patients using immunohistochemistry. Overall, the manuscript is clear and well-organized, with its significance rooted in the large sample size and the development of a classification method.

Thank you for your positive assessment of our study on classifying RPLS patients based on whole transcriptome sequencing. We appreciate your recognition of the distinct characteristics of the G1 and G2 groups, as well as the significance of our simplified classification system and the identification of LEP and PTTG1 as key molecular markers. Your acknowledgment of the clarity and organization of our manuscript, along with the importance of the large sample size, is greatly appreciated. We will continue to refine our work based on your feedback as we prepare for resubmission.

Weakness:

(1) While the authors suggest that LEP and PTTG1 serve as molecular markers for the two RPLS groups, the process through which these genes were selected remains unclear. The authors should provide a detailed explanation of the selection process.

The selection criteria for identifying LEP and PTTG1 as biomarkers involved selecting prognostic genes that were highly expressed in C1 and C2, respectively, and achieved the highest AUC value in distinguishing the two RPLS groups (Page17 lines 288-290).

(2) To ensure the broader applicability of LEP and PTTG1 as classification markers, the authors should validate their findings in one or two external datasets.

We sincerely appreciate your insightful suggestion regarding the external validation of LEP and PTTG1 as classification biomarkers. To address this concern, we performed an independent validation using an external liposarcoma cohort (GSE30929; Page 6, Lines 104-105)), which comprises 140 primary liposarcoma samples with annotated clinicopathological and survival data. This dataset was selected due to its relevance to RPLS (N=63, 45%) and the availability of distant recurrence-free survival (DRFS) outcomes, aligning with the clinical focus of our study. 

Applying our previously established prognostic model (Risk value = 2.182 × PTTG1 - 2.204 × LEP) to this cohort, we stratified patients into high- and low-risk groups using the median risk score as the cutoff. Consistent with our original findings, the high-risk group exhibited significantly worse DRFS compared to the low-risk group. The ROC curves based on the 1-, 3-, 5-year survival status of patients demonstrated that this model can effectively predict patient DRFS (log-rank P < 0.001, Figure S3A-B). Furthermore, the high-risk group demonstrated a higher proportion of high-grade histology (P < 0.001, Fisher’s exact test, Figure S3C-D).

These results validate the robustness and generalizability of our risk stratification model across distinct liposarcoma cohorts. The external dataset’s alignment with our findings underscores the potential of LEP and PTTG1 as reproducible biomarkers for prognosis and therapeutic stratification in liposarcoma. We have incorporated these validation results into the revised manuscript (Page 18, Lines 305-315) to strengthen the clinical applicability of our conclusions.

(3) Since molecular subtyping is often used to guide personalized treatment strategies, it is recommended that the authors evaluate therapeutic responses in the two distinct groups. Additionally, they should validate these predictions using cell lines or primary cells.

We sincerely appreciate your insightful comments and suggestions regarding the evaluation of therapeutic responses and the validation of our predictions using cell lines or primary cells. We would like to address these points in detail below:

(1) Purpose of the PTTG1- and LEP-based RPLS Classification Model

The primary objective of our study was to develop a molecular subtyping model based on PTTG1 and LEP to guide personalized treatment strategies for patients with RPLS, particularly those classified as low-grade by traditional histopathological criteria but exhibiting poor prognosis. This subgroup of patients may benefit from more aggressive surgical resection, which is a potentially curative approach for RPLS. Our model aims to identify these high-risk patients to ensure complete tumor resection, thereby improving their clinical outcomes.

(2) Therapeutic Response Evaluation in Distinct Groups

In both our validation cohort and external validation cohort, surgical resection was the primary treatment modality for RPLS. After stratifying patients using our model, we observed significant differences in surgical outcomes between the two groups: the high-risk group exhibited poor prognosis, while the low-risk group showed favorable outcomes (Figure 5D-E and Figure S3A-B). Importantly, our model successfully identified low-grade histopathological cases with poor prognosis, who might otherwise be undertreated (Figure 5G-I and Figure S3C-D). By advocating for more thorough surgical resection in these high-risk patients, we aim to improve their prognosis. This achievement aligns with the primary goal of our study, which is to provide a molecular tool for personalized treatment guidance.

(3) Future Validation and Functional Exploration of PTTG1 and LEP

Our study has identified PTTG1 and LEP as key biomarkers for RPLS classification, and we recognize the urgent need to elucidate their molecular functions in RPLS pathogenesis. Here, we are pleased to report that we have already initiated cellular and animal experiments to investigate the roles of PTTG1 and LEP in RPLS. These experiments aim to validate our predictions and explore the underlying mechanisms by which these biomarkers contribute to tumor behavior and treatment response. We anticipate that the results of these studies will provide further mechanistic insights and will be submitted for publication in a suitable journal in the near future.

Reviewer #2 (Public review):

Surgical resection remains the most effective treatment for retroperitoneal liposarcoma. However, postoperative recurrence is very common and is considered the main cause of disease-related death. Considering the importance and effectiveness of precision medicine, the identification of molecular characteristics is particularly important for the prognosis assessment and individualized treatment of RPLS. In this work, the authors described the gene expression map of RPLS and illustrated an innovative strategy of molecular classification. Through the pathway enrichment of differentially expressed genes, characteristic abnormal biological processes were identified, and RPLS patients were simply categorized based on the two major abnormal biological processes. Subsequently, the classification strategy was further simplified through nonnegative matrix factorization. The authors finally narrowed the classification indicators to two characteristic molecules LEP and PTTG1, and constructed novel molecular prognosis models that presented obviously a great area under the curve. A relatively interpretable logistic regression model was selected to obtain the risk scoring formula, and its clinical relevance and prognostic evaluation efficiency were verified by immunohistochemistry. Recently, prognostic model construction has been a hot topic in the field of oncology. The interesting point of this study is that it effectively screened characteristic molecules and practically simplified the typing strategy on the basis of ensuring high matching clinical relevance. Overall, the study is well-designed and will serve as a valuable resource for RPLS research.

Thank you for your insightful feedback on our manuscript. We appreciate your recognition of the importance of precision medicine and molecular characteristics in improving prognosis and individualized treatment for RPLS.

We are pleased that you found our gene expression mapping and innovative molecular classification strategy valuable. Your positive remarks on our pathway enrichment analysis and the categorization of RPLS patients based on abnormal biological processes affirm our approach.

We are also grateful for your acknowledgment of our focus on the characteristic molecules LEP and PTTG1, as well as the development of novel molecular prognosis models with significant predictive capability.