These bioinks comprise two mis-cible polymer networks with complementary gelation mechanismsthat regulate different stages of the fabrication process. A thermo-responsive gelatin network provides excellent extrusion and struc-tural stability during 3D printing, while a photocrosslinkable networkallows the printed structure to be stabilized by covalent crosslink-ing.

complementarynetwork bioinks naturally release the majority of gelatin at 37°C.Critically, the inclusion of gelatin did not notably affect themechanical properties of the hydrogels after thermal release, nor

Another important and enduring challenge of hydrogel-based musculoskeletal tissue engineering is replicating the mechani-cal properties of the native tissues. The mechanical properties of thetissue constructs could potentially be improved by adjusting thebioink (e.g., through the use of double-network bioinks), coprintingsolid polymeric scaffolds (e.g., by melt electrowriting), incorporatingmechanical stimulation into the tissue culture setup, or extendingthe maturation period

polymer type, photocrosslinking kinetics, and bioink concentrationcan all be selected in the absence of any constraints imposed by theprintability of the formulation. In turn, this offers a more effectivemethodological approach to 3D printing living tissue constructs, wherebythe bioink can be readily tuned to meet the biological requirementsof the encapsulated cells.

concept of complementary networkbioinks

While highly specific bioink formulations (12) or customizedprinting methodologies (13) can address the printability of certainbiomaterial systems, these do not provide a standardized method thatcan be applied across different 3D printing scenarios. This consider-ation highlights the motivation for a more generalizable strategy thatextends the biofabrication window across a broad palette of print-able, cytocompatible bioinks

rarelyhave both the physicochemical properties required for the 3D print-ing process and the physicochemical cues to meet the biological needsof the encapsulated cells

We have also used a very standard thermalextrusion–based approach that is compatible with existing 3D bio-printing methods (11, 23, 29). Together, these design factors shouldenable complementary network bioinks to find broad applicabilityacross different biomaterial systems and 3D printing protocols, andprovide new opportunities for biofabrication and tissue engineering.

it remains to be examined how inputs from differentsignaling pathways converge on Sox9 to promote PanIN initia-tion. Our findings now pave the way for future studies exploringwhether the inhibition of acinar cell plasticity could have thera-peutic applications in the prophylaxis of PDA in high-riskindividuals.

therapeutic targeting of signaling path-ways involved in ductal reprogramming of acinar cells couldprevent PDA initiation

reducing ADMformation through inhibition of the EGFR pathway or overexpres-sion of pro-acinar genes reduces Kras G12D-induced PanINs

both ADM and PanIN lesions arise from acinar cells

a direct lineage relationship between ADM and PanINshas yet to be formally demonstrated and the possibility remainsthat ADM lesions do not directly transition into PanINs.

risk factorsfor PDA, such as chronic pancreatitis or genetic variants ofacinar-specific genes (Li et al., 2012; Lowenfels et al., 1993,1997), may increase PDA risk by rendering acinar cells moreplastic and reducing the threshold for ADM.

PanINs predominantly originate from acinar cellsidentifies the transition of acinar cells into a duct-like state as animportant early event in tumor initiation.

transcriptome analysis

In addition to activating mutations inKRAS, loss-of-function mutations in the tumor suppressorsTP53, CDKN2A, or PTEN have also been found in PDA

However, multipo-tent progenitor markers, such as Hnf1b and Nkx6.1 (Schafferet al., 2010; Solar et al., 2009), are not present in acinar-derivedduct-like cells (Jensen et al., 2005; Prevot et al., 2012) indicatingthat these cells may resemble an acinar-committed yet stillimmature Nestin-positive progenitor cell

Sox9, Hnf6expression is not maintained in PanINs and PDA (Prevot et al.,2012). Therefore, only Sox9 expression is sustained in PanINs,providing a possible explanation for why Sox9 is absolutelyrequired for PanIN formation

uggests that other factors cancompensate for Sox9 during injury-induced ADM. One candi-date is the duct-specific transcription factor Hnf6

specific induction ofdifferent oncogenic mutations in mice may define morphologi-cally and molecularly distinct tumors, which may help identifyhuman PDA subtypes that respond differently to therapeuticintervention

biological roleof pancreatic duct glands, but their possible involvement inPDA initiation warrants further investigation.

CACsexpress some markers of embryonic pancreatic progenitorsand exhibit features of tissue stem cells in vitro (

ox9 deletion in the presence of oncogenic Krasabrogated caerulein-induced PanIN formation, while someSox9
CK19 + Cpa1
duct-like lesions persisted

Sox9 is critically required for reprogram-ming of acini into PanINs

ADM were replaced by normal acinar tissue after 7 days

considerable latency, which is significantly shortened inthe presence of pancreatitis

Sox9 expression in acinar cells destabilizes theacinar cell state and promotes expression of ductal genes, butis not sufficient to induce complete ductal reprogramming.

Together, these resultssuggest that Sox9 accelerates KrasG12D-mediated PanIN formation by suppress-ing a mature acinar cell program and/or promoting a duct-likestate

most cells hadalready transitioned from a CK19 + Cpa1+into a CK19 + Cpa1
state characteristicof ADR

very few cells retained Cpa1 in Ptf1a Cre ;Kras G12D;Sox9 OE mice and CK19+ ductal structures werepredominant (Figure 4AF)

concomitant misexpression of Sox9 andKras G12D rapidly induces transformation of acinar cells intoduct-like cells and subsequent PanIN formation.

ynergy between Sox9 and KrasG12D wasalso observed at a microscopic level, as evidenced byreplacement of normal pancreas parenchyma by large areas ofSox9 + ADM and Alcian blue + PanINs in Ptf1a Cre ;KrasG12D ;Sox9 OE mice

suggesting that Ptf1a Cre-mediatedSox9 misexpression has no overt effect on pancreaticdevelopment

Notably, PanINs that stillformed in Ptf1a CreER;Kras G12D;Sox9 f/f ;R26RYFP mice were YFP +and Sox9 + (Figures 4N and 4Q), indicating that these PanINsarose from acinar cells that recombined the R26R YFP andKras G12D alleles, but not the Sox9 flox allele. Together, these find-ings demonstrate that PanIN formation from Kras-active acinarcells requires Sox9 activity (Figure S3R)

53-fold reduction in the Alcian blue+ area in Sox9-deleted Ptf1a CreER;Kras G12D mice compared to mice with twofunctional Sox9 alleles

data indicate that Sox9 expression is initiated before Kras-activeacinar cells progress to a duct-like state and become PanINs.

To inducibly recombine the LSL-Kras G12D allele (hereafterreferred to as the Kras G12D allele) in ductal/CAC or acinar cells,we used Sox9CreER or Ptf1a CreER mice, respectively

As seen in PanINs originating from acinar cells (Figures 2F, S1G,and S1H), the duct-derived lesions were Alcian blue + (Figure 2H,inset), and expressed Muc5AC and Claudin18 (Figures S1I andS1J) and the lineage marker YFP

Alcian Blue/Eosin

A

4 days after tamoxifen injection. After accounting for the 4.6-fold difference (Figure S1L), Ptf1a CreER-induced Kras activationstill resulted in a 112-fold higher frequency of PanINs thanSox9CreER-mediated Kras activation (Figure 2M). These datashow that acinar cells have a much greater propensity thanductal/CACs to form PanINs in response to oncogenic Kras(Figure 2N)

In Ptf1a CreER;Kras G12D;R26R YFP mice, on average6.7 ± 1.5% of the pancreas was Alcian blue + . In contrast, only0.013 ± 0.004% of the pancreas exhibited Alcian blue stainingin Sox9CreER;Kras G12D;R26RYFP mice

unlikeacinar-derived PanINs, the duct-derived lesions were notrandomly distributed throughout the pancreas, but were moreoften associated with large ducts (Figure S1J, arrow).

all Ptf1a CreER;Kras G12D;R26R YFP mice displayed abundant lesions with histo-logic and molecular characteristics of PanINs, including high

assess the frequency of PanINs arising from acinar orductal/CACs after Kras activation, we examined pancreatafrom Ptf1a CreER;Kras G12D;R26R YFP and Sox9CreER;KrasG12D;R26RYFP mice 8 to 17 months after Kras G12D induction

FP expression was observed after8 months or longer (Figures 2I and 2K),confirming that acinar and ductal cellsdo not spontaneously convert into other pancreatic cell types

analyzed pancreatic labeling specificityand efficiency by quantifying the percentage of ductal and acinarcells expressing YFP

ow-grade PanINs wereuniformly SOX9 +, whereas higher-grade PanIN2/3 lesions andPDA displayed heterogeneous SOX9 expression (Figure 1L;72% of PanIN2/3 and 69% of PDA were SOX9+ ). These findingssuggest that a SOX9 + state is associated with PDA initiation.

SOX9 was expressed in chronic pancre-atitis, as well as premalignant and malignant lesions, includingMCNs, IPMNs, PanINs, and PDA.

tissue microarray for immunohisto-chemical analysis of SOX9 expression in different pancreaticlesions

Sox9 is induced during ADM and ex-pressed in PanINs and PDA

directly compared the propensity of ductal/CACs and acinar cells to form PanINs

nducers of ductal cell identity, such asSox9 (Delous et al., 2012; Shih et al., 2012), play a role in theinduction of PanINs from acinar cells.

directoncogenic transformation of ductal orCACs.

CK19 promoter-based alleles to activateoncogenic Kras in ductal cells, suggest that PanINs rarely arisefrom ducts

tumor suppressorPten, the formation of invasive pancreatic cancer is associatedwith increased proliferation of centroacinar cells (CACs) (Stangeret al., 2005), which reside at the tips of the ductal tree

revious studies have modeled PDA initiation mostlyby expressing oncogenic Kras in all cell types of the pancreas(Aguirre et al., 2003; Hingorani et al., 2003, 2005), little is knownabout its cell of origin

accel-erates KrasG12D-mediated PanIN and PDA formation in mice

PanINs),

ADM is also observed in pancreatitis, which is a significantrisk factor for PDA in humans

loss-and gain-of-function approaches we identify the ductal fate determinant Sox9 as a critical mediator of Kras-induced prema-lignant acinar cell reprogramming.

PanIN 2-3 2/7 [29%] 2/7 [ 29%] 3/7 [43%]PDA 6/19 [32%] 7/19 [ 37%] 6/19 [32%]

PanIN 1-2 1/26 [ 4%] 12/26 [ 46%] 13/26 [50%]

CK19 and theductal fate determinant Sox9

KrasG12D-expressing mice, PanINformation coincides with, or is preceded by, acinar-to-ductalmetaplasia (ADM

pancreatic intraepithelial neoplasias(PanINs)

spectrum of preneoplastic mucinous lesions withductal morphology

morphologic features do not necessarily predicta lineage relationship

the cell type responsible for tumor initiation has oftenbeen inferred based on the histologic appearance of the tumor

Defining tumor-initiating events

oncogenic Kras, Sox9 accelerates formation of premalignant lesions. These results provideinsight into the cellular origin of PDA and suggest that its precursors arise via induction of a duct-like state inacinar cells