Stiffening hydrogels for investigating the dynamics of hepatic stellate cell mechanotransduction during myofibroblast activation

Tissue fibrosis contributes to nearly half of all deaths in the developed world and is characterized by progressive matrix stiffening. Despite this, nearly all in vitro disease models are mechanically static. Here, we used visible light-mediated stiffening hydrogels to investigate cell mechanotransduction in a disease-relevant system. Primary hepatic stellate cell-seeded hydrogels stiffened in situ at later time points (following a recovery phase post-isolation) displayed accelerated signaling kinetics of both early (Yes-associated protein/Transcriptional coactivator with PDZ-binding motif, YAP/TAZ) and late (alpha-smooth muscle actin, α-SMA) markers of myofibroblast differentiation, resulting in a time course similar to observed in vivo activation dynamics. We further validated this system by showing that α-SMA inhibition following substrate stiffening resulted in attenuated stellate cell activation, with reduced YAP/TAZ nuclear shuttling and traction force generation. Together, these data suggest that stiffening hydrogels may be more faithful models for studying myofibroblast activation than static substrates and could inform the development of disease therapeutics.

Tissue fibrosis contributes to nearly half of all deaths in the developed world 1 . Fibrosis develops in response to a wide range of injuries in many tissues, including the pancreas, lung, kidney, and liver 2,3 . The differentiation of pericytes and fibroblasts to fibrogenic myofibroblasts is central to the development of fibrosis and results in aberrant extracellular matrix (ECM) synthesis and remodeling, which in turn leads to tissue stiffening and loss of organ function. Hepatic stellate cells are the primary source of myofibroblasts in liver fibrosis 4 and primary liver cancer (often associated with fibrosis), which together account for over 1.7 million deaths annually worldwide 5 .
There is a growing appreciation for the role that local mechanical properties play in the initiation, development, and progression of pathologies such as liver fibrosis, especially in light of recent evidence that tissue stiffening, which is generally regarded as an outcome of disease, may be a significant contributing factor in the development of fibrosis and in tumor formation [6][7][8] . Hydrogel biomaterials are increasingly being applied as model microenvironments to probe cellular behavior with seminal studies defining the roles of substrate stiffness 9 , geometry 10 , and degradability 11 in regulating cell fate. The role of mechanics in stellate cell differentiation has been investigated using several hydrogel platforms to show that hepatic stellate cells cultured on compliant (elastic modulus (E) or stiffness < 3 kPa) substrates with stiffness levels similar to healthy liver remain phenotypically quiescent, while stellate cells cultured on stiffer substrates with stiffness levels similar to diseased liver (E > 20 kPa) undergo myofibroblast differentiation [12][13][14] .
Despite this, there is a paucity of in vitro model systems that can recapitulate the temporal dynamics of mechanics during disease; all previous studies of hepatic stellate cells during myofibroblast activation have occurred on mechanically static substrates. Unfortunately, these static materials may not replicate the changes observed in vivo, as primary cells are digested from tissue and directly seeded onto substrates rather than experiencing gradual or delayed stiffening. Although hepatic stellate cells in vivo up-regulate fibrogenic proteins such as connective tissue growth factor and β -platelet-derived growth factor receptor within several hours following fibrosis induction 15,16 , many freshly isolated primary cells (including stellate cells) show altered expression of integrins and other surface proteins 17,18 that may attenuate their response to early mechanical stimuli, highlighting the importance of the timing of mechanical signals for in vitro disease models. Additionally, as new molecular targets are identified, it is important to understand them in dynamic systems that are relevant to the in vivo setting. For example, the Hippo-related transcriptional co-activators Yes-associated protein/Transcriptional coactivator with PDZ-binding motif (YAP/TAZ) have been identified as important mechanosensing regulators involved in stem cell differentiation 19,20 and were recently shown to be responsive to dynamic mechanical loading 21 . However, the role of YAP/TAZ in the mechanically-driven progression of fibrosis has not been investigated in the context of changing mechanics 16,22 .
Various stiffening systems have been used to investigate stem cell fate decisions 23 , cell spreading 24 , cardiomyocyte maturation 25 , and myofibroblast phenotype in heart disease 26 . Collectively, these studies suggest that both the magnitude and the timing of mechanical signals are important for investigating mechanotransduction signaling dynamics in complex cell behaviors such as stem cell differentiation and myofibroblast activation. However, previous approaches do not offer precise temporal control of mechanical signals 25 or are limited by their use of ultraviolet (UV) light 23,26 , which may be harmful to some cell types (including stellate cells 27 ). In this work, we aimed to develop a hydrogel system amenable to in situ stiffening in the presence of sensitive primary cell types like hepatic stellate cells to investigate how stiffening events influence the dynamics of both early (YAP/TAZ nuclear translocation) and late (actin cytoskeleton maturation) events in stellate cell myofibroblast differentiation.

Results
Visible light-mediated secondary crosslinking enables stiffening of MeHA hydrogels. We observed that UV irradiation significantly reduced hepatic stellate cell viability and increased nuclear p53 accumulation when compared to visible (blue) light or no irradiation ( Supplementary Fig. 1). Based on these results, we conceived of a scheme to fabricate stiffening hydrogels from methacrylated hyaluronic acid (MeHA) (Fig. 1a). This approach combined an initial Michael-type addition reaction to form a soft gel followed by visible light-mediated radical polymerization of unreacted methacrylates at a user-defined time point to stiffen the substrate (Fig. 1b). Although we considered several initiators (e.g., Eosin Y, Supplementary Fig. 2), lithium acylphosphinate (LAP) 28 was identified as the most promising since there was no reduction in hepatic stellate cell viability even at high LAP concentrations (6.6 mM) or with combined LAP (6.6 mM) and visible blue light exposure (5 min, 10 mW cm −2 , peak wavelength ~470 nm) ( Supplementary Fig. 3). The combination of blue light exposure and LAP for secondary crosslinking enabled significant substrate stiffening as measured by rheometry (Fig. 1c) and atomic force microscopy (Fig. 1d). Using this approach, we fabricated both "soft" substrates (E = 1.75 ± 0.02 kPa, corresponding to healthy liver stiffness) and secondary crosslinked "stiff " substrates (E = 33.0 ± 3.2 kPa, fibrotic liver stiffness).
Hepatic stellate cells spread and assume myofibroblast morphology in response to in situ stiffening. Using this dynamic hydrogel system, we investigated stellate cell response to in situ secondary crosslinking (Fig. 2a). Hepatic stellate cells seeded on soft substrates remained rounded and retained lipid droplets, whereas stellate cells cultured on stiff substrates spread gradually over several days, lost their lipid droplets, and assumed a myofibroblast-like morphology (Fig. 2b). Meanwhile, stellate cells initially cultured on soft gels and subjected to in situ secondary crosslinking responded rapidly (~18-24 h) to the change in substrate stiffness, resembling the morphology of cells cultured on the stiff static substrates (Fig. 2b,c, Supplementary Movie 1). To ensure that the observed cell morphology results were due to changes in substrate stiffness and not generation of free radicals 29 , soft MeHA gels were fabricated where the remaining methacrylates were capped with a thiol, so that light exposure in the presence of a photoinitiator led to free radical generation but no increased crosslinking. No differences in hydrogel elastic modulus or cell spread area were observed when compared to cells on soft hydrogels, indicating that radical generation did not induce stellate cell spreading (Fig. 2d,e, Supplementary Fig. 4).
Delayed stiffening promotes more rapid YAP/TAZ nuclear translocation and α-SMA stress fiber assembly. We hypothesized that the gradual cell spreading observed over several days on stiff substrates ( Supplementary Fig. 5) was due to a post-isolation recovery phase where stellate cell mechanosensing was attenuated. To investigate this, we stiffened gels at either an early pre-recovery (day 1) or later post-recovery (day 6) time point and assessed stellate cell spreading over 14 total days of culture compared to soft and stiff static gels. Stellate cell spread area for early stiffening (day 1) largely mirrored trends observed for the static stiff hydrogels. Interestingly, stellate cells were observed to spread more rapidly following later stiffening (day 6) and "catch up" to early stiffening groups within 48 h ( Supplementary Figs 5 and 6). AFM assessment indicated that all stiffened substrates had similar elastic moduli after 14 days, regardless of stiffening time point (Supplementary Fig. 7). While previous studies showed that hepatic stellate cells spread gradually over several days when plated onto stiff culture substrates 14,30 , here we show that following a culture period of several days in a soft environment that allows recovered expression of surface proteins such as platelet-derived growth factor receptor beta (β -PDGFR) and β 1 integrin ( Supplementary Fig. 8), stellate cells spread and activate much more rapidly in response to a step increase in stiffness.
We next performed a series of experiments to elucidate the temporal dynamics of mechanotransduction events immediately following stiffening at either an early (day 1) or later (day 6) post-isolation recovery time point. Spread area, YAP/TAZ translocation (early marker), α -SMA expression (late marker), and F-actin organization were evaluated 1-72 h following stiffening. Results were compared to static soft gel controls (dashed lines, Fig. 3). No significant differences were observed between the static soft day 1 and day 6 controls (dashed lines) for any of the measured outcomes. YAP/TAZ nuclear intensity gradually increased following both early and late stiffening, although this change was more rapid following later stiffening with YAP/TAZ nuclear intensity 1 h after later stiffening roughly equivalent to 72 h following earlier stiffening ( Fig. 3c,g, Supplementary Figs 9 and 10). While significantly higher YAP/TAZ nuclear intensity and F-actin stress fiber organization were observed following later stiffening within 1 h (Fig. 3c,e,g, Supplementary Figs 9 and 10), myofibroblast maturation as measured by α -SMA stress fiber organization lagged behind both of these metrics (Fig. 3d,g, Supplementary Figs 9 and 10). Most stellate cells displayed diffuse α -SMA staining even 72 h following early stiffening, while α -SMA stress fiber organization gradually increased after later stiffening and was significantly higher at the 24-72 h time points compared to earlier stiffening. When these experiments were extended to 14 days of total culture time, stellate cells on all stiffened substrates (static stiff, early (day 1) and late (day 6) stiffening) showed significantly elevated levels of YAP/TAZ nuclear intensity, F-actin (and more specifically α -SMA) stress fiber formation, and fibrogenic gene expression of Acta2 and Col1a1 (encoding for α -SMA and type I collagen, respectively) compared to stellate cells on soft gels ( Supplementary Figs 11 and 12).
Specific inhibition of α-SMA polymerization results in reduced YAP/TAZ nuclear shuttling and traction force generation. We used our dynamic stiffening system to study the connection between early YAP/TAZ -mediated mechanosensing events and later markers of myofibroblast maturation, such as α -SMA organization. We cultured stellate cells on hydrogels that were stiffened after three days of recovery and then supplemented with an α -SMA blocking peptide (BP(+ )) 31 and compared outcomes to a non-treated control (BP(− )). Inhibiting α -SMA resulted in nearly complete abrogation of α -SMA staining without altering stress fiber formation of other F-actin isoforms (Fig. 4a,b). However, α -SMA blockade resulted in significantly reduced YAP/TAZ nuclear intensity (P < 0.001, Fig 4b,c). We hypothesized that the decrease in YAP/TAZ nuclear translocation was due to reduced mechanical feedback from α -SMA-mediated contractility. Although hepatic stellate cells are known to be sensitive to mechanical signals and are highly contractile during disease progression, to our knowledge no previous studies have quantified the traction forces exerted by these cells. Here, we used traction force microscopy (TFM) to show that treatment of stellate cells with the blocking peptide did not alter spread

Discussion
In this work we developed a cytocompatible method to fabricate in situ stiffening hydrogels in order to recapitulate changes in mechanical signaling that occur at the onset and during the progression of diseases such as liver fibrosis. This is an irreversible technique, where the magnitude and timing of the stiffening process are altered with light exposure parameters. Hydrogels were modified with an RGD peptide for cell adhesion; however, a range of ECM proteins and peptides could be investigated in this system. We then showed that hepatic stellate cells activated more quickly in response to delayed post-recovery stiffening in a manner reminiscent of in vivo activation compared to early stiffening. We further demonstrated the unique power of this system to show that inhibiting α -SMA polymerization following substrate stiffening resulted in reduced YAP/TAZ nuclear localization and traction force generation. Collectively, these data suggest that stiffening hydrogels may be superior to static culture substrates for studying the dynamics of mechanotransduction in disease and could inform the identification of targets for therapeutic intervention.
We previously developed an in situ stiffening hydrogel system that relied on UV light irradiation to introduce secondary crosslinking into hydrogels 23 . Although this system had many salient features towards our goals here, previous reports indicated that hepatic stellate cells are sensitive to UV light 27 , possibly due to the presence of vitamin A-containing lipid droplets 32 in quiescent stellate cells. As an alternative approach, we demonstrated that visible light-mediated secondary crosslinking resulted in a pathologically-relevant increase in hydrogel stiffness without harming stellate cell viability.
Harnessing the power of this mechanically dynamic model system, we hypothesized that not only the stiffening event itself, but also the timing of stiffening would affect stellate cell myofibroblast phenotype as manifested in both early (YAP/TAZ) and late (α -SMA) markers. YAP/TAZ are Hippo-related transcriptional co-activators that primarily interact with TEAD-related transcription factors to regulate gene expression in a wide range of downstream processes. Although the involvement of YAP/TAZ in the (dys)regulation of organ growth during development and disease (e.g., cancer) is well-characterized 33 , its prominent role in the progression of liver fibrosis has only recently been elucidated 16 . Numerous YAP/TAZ target genes such as connective tissue growth factor (Ctgf) promote fibrotic responses, including increased α -SMA expression 34 , a hallmark of the myofibroblast phenotype. We showed that later stiffening within the first week of culture resulted in accelerated stellate cell spreading and α -SMA stress fiber formation, mediated by quicker YAP/TAZ nuclear translocation.
The more rapid activation observed following later stiffening is consistent with hepatic stellate cell activation kinetics in vivo where YAP/TAZ target genes such as Ctgf as well as fibrogenic proteins like β -PDGFR are up-regulated within several hours of injury 15,16 . The difference in signaling kinetics observed between early and late stiffening might be due to a post-isolation recovery phase that initially retards the ability of stellate cells to respond to mechanical inputs in a pathologically-relevant manner. Hepatic stellate cell gene expression profiles in cultured activated cells (including various surface proteins and cytoskeletal components) are notably different than in vivo activated stellate cells 18 . Although the simplified microenvironment for in vitro activation likely accounts for some of these differences, expression of these markers may also be skewed by protease-dependent . While YAP/TAZ nuclear localization was more evident 72 h following earlier stiffening (day 1), α -SMA staining was still diffuse and did not co-localize with F-actin stress fibers. In contrast, within 1 h following later stiffening (day 6) YAP/TAZ nuclear translocation was present and 72 h following stiffening the majority of stellate cells displayed α -SMA stress fibers. Representative immunostaining for all experimental groups is presented in the Supplementary Information. Dashed lines represent levels on soft gels at Day 1 (light blue) or Day 6 (dark blue). *P < 0.05, **P < 0.01, ***P < 0.001 where comparisons shown in figure are between day 1 and day 6 results for each post-stiffening time point (e.g., in panel (c) the day 6 YAP/TAZ nuclear intensity ratio 1 h post-stiffening is significantly greater than day 1 YAP/ TAZ intensity 1 h post-stiffening, P < 0.05). Scale bars: 50 μ m. cell isolation procedures. Indeed, we observed significantly increased expression of β 1 integrin in stellate cells after 6 days culture on soft gels compared to freshly isolated cells, as well as a return to basal levels of β -PDGR following a recovery period of several days. Critically, the observed differences in signaling kinetics following cell isolation and subsequent substrate stiffening would not have been possible with conventional cell culture platforms. Previous work examining stellate cell activation on soft and stiff supports highlighted the role of mechanics and suggested that culture on soft and physiologically stiff substrates, rather than on plastic, is optimal for understanding normal stellate cell function and activation. The data presented here further suggest that culture on standard plastic supports misrepresents the kinetics of activation, and that optimal cell culture requires a post-isolation recovery period as well as mechanically physiologic substrates. Stiffening hydrogels, which address both requirements, could be ideal in vitro models for studying stiffness-mediated events in development and disease, including although not limited to the timing and relevance of α -SMA expression.
Collectively, our findings indicate that while initial stellate cell culture on soft substrates suppresses numerous canonical indicators of myofibroblast phenotype, including spreading and α -SMA stress fiber organization, stellate cells may still be primed for quicker activation as they recover from the isolation procedure. Despite a previous report on the mechanical memory of lung fibroblasts that found extended (~2 week) culture on soft substrates protected against stiffness-mediated myofibroblast activation 35 , our data are supported by a study of hepatic stellate cells cultured on Matrigel that showed shorter-term culture (~1 week) did not prevent activation 36 . It is also important to note that in the aforementioned studies cells were treated with proteases (similar to procedures during cell isolation) and moved to different substrates, likely altering the mechanoresponsive timeline 35,36 . Our system is uniquely positioned to allow cells to recover post-isolation without any exposure to aberrant mechanical signals (e.g., tissue culture plastic), and then introduce a pathologically-relevant mechanical input at a user-defined time point. Our results suggest that this approach may more faithfully capture the dynamics of mechanotransduction signaling events in disease compared to conventional static culture substrates. More broadly, our results may have implications for other applications involving the use of soft and/or stiff culture substrates as tools to prevent or guide specific differentiation fates.
Finally, we applied our mechanically dynamic hydrogel system to demonstrate that blocking α -SMA polymerization following substrate stiffening results in reduced YAP/TAZ nuclear localization and traction force generation, indicating that α -SMA contractility is critical for YAP/TAZ -mediated mechanotransduction during myofibroblast maturation. Importantly, we believe that this could only be investigated with a system such as that developed here. In these experiments, the stiffening time was less important as cells undergo activation across the range of stiffening timepoints and we chose an intermediate recovery time (3 days) prior to stiffening for investigation. This was partly to ensure that the hydrogel properties were unchanged at the time traction studies were performed. While previous studies demonstrated that α -SMA inhibition resulted in reduced contractility, motility, and Erk1/2 activity in lung 31 and liver 37 myofibroblasts, to our knowledge the connection between α -SMA contractility and YAP/TAZ activity in myofibroblast activation has not been made. The observed reduction in traction force generation is functionally important since activated stellate cells in vivo display traction force-dependent increases in motility and contractility that lead to restricted portal blood flow and permit the release of latent ECM-sequestered fibrogenic cytokines such as TGF-β 38 . Well-defined cell culture systems such as the one developed here provide platforms for improving our understanding of how mechanosensitive cells like hepatic stellate cells respond to dynamic changes in their environment and could inform the development of disease therapeutics (e.g., inhibition of YAP/TAZ translocation or α -SMA organization).

Methods
MeHA synthesis. Hydrogels were formed from methacrylated hyaluronic acid as previously reported 39 .
2 wt% sodium hyaluronate (Lifecore, 75 kDa) in deionized water was reacted with methacrylic anhydride (Sigma, 5.6 mL per g HA) on ice at pH 8.0-9.5 for 6 h. The solution was dialyzed at room temperature in deionized water (SpectraPor, 6-8 kDa molecular weight cutoff) for 7 d, and then lyophilized. The degree of modification was ~90% as measured by 1 H NMR (Bruker, Supplementary Fig. 13).
MeHA hydrogel fabrication. The adhesion ligand RGD (GCGYGRGDSPG, GenScript) was coupled to methacrylates on HA via a Michael-type addition reaction to enable cell attachment. MeHA and RGD peptide were incubated under shaking for 45 min at room temperature in 0.2 M triethanolamine (TEOA, Sigma) buffer at pH 9 at ratios to obtain a final RGD concentration in hydrogels of 1 mM. Soft MeHA hydrogels were also formed by Michael-type addition. 3 wt% RGD-modified MeHA was crosslinked with dithiothreitol (DTT, Sigma) at pH 9. Hydrogel precursor solution (70 μ L) was pipetted into polydimethylsiloxane molds (PDMS, Dow Corning, 18 × 18 × 0.15 mm) and covered with methacrylated glass coverslips 40 for 3 h at room temperature to form a hydrogel film attached to the functionalized coverslip. RGD functionalization and crosslinking of soft hydrogels only consumes ~20% of available methacrylate groups, leaving ample groups available for secondary crosslinking. For TFM experiments, 0.2 μ m diameter fluorescent beads (Invitrogen) were mixed with hydrogel precursor solution at 1% v/v prior to gelation. LAP synthesis. Lithium acylphosphinate (LAP) was synthesized as previously described 28,41 . Briefly, equimolar amounts of dimethyl phenylphosphonite (Sigma) and 2,4,6-trimethylbenzoyl chloride (Sigma) were reacted for 18 h under constant stirring and nitrogen at room temperature. A four-fold excess of lithium bromide (TCI America) dissolved in 2-butanone (Sigma) was added to the reaction mixture and heated at 50 °C for 10 min, forming a precipitate. The precipitate was washed several times in 2-butanone, dried overnight, and stored at 4 °C under nitrogen in the dark until use. The structure was confirmed by 1 H NMR (Bruker, Supplementary Fig. 14).
Peptide synthesis. The α -SMA blocking peptide (Ac-EEEDGRQIKIWFPNRRMKWKK) 31,42 was synthesized using standard solid supported Fmoc protected methods. The peptide was cleaved in trifluoroacetic acid under vigorous stirring for 4 h, precipitated in cold ether, dissolved in water, frozen at − 80 °C, and lyophilized. Successful synthesis was confirmed by MALDI ( Supplementary Fig. 15). In situ stiffening. MeHA hydrogels were stiffened at a user-defined time point either before or during cell culture. Soft hydrogels were incubated in culture media containing 6.6 mM LAP photoinitiator for 30 min at 37 °C followed by blue light irradiation for 5 min at 10 mW cm −2 using a dental curing lamp (ESPE Elipar 2500, 3 M). Following light exposure, hydrogels were rinsed twice in PBS to remove LAP and placed in fresh culture media.
Hydrogel mechanical characterization. Hydrogel mechanics were assessed using rheology and atomic force microscopy (AFM). Rheometer experiments were performed on an AR2000ex (TA Instruments) with a light guide connected to the blue light lamp. MeHA hydrogels were allowed to form via an initial addition reaction at pH 10 for 30 min and then were exposed to blue light for 15 min (3 mW cm −2 ). The elastic moduli of hydrogel thin films were obtained from AFM (Agilent 6000ILM). Force curves of 2 × 2 arrays were taken on at least 3 different areas per hydrogel with a 1 μ m SiO 2 spherical probe (0.06 N/m, Novascan). Elastic moduli were obtained from force curve data using the Sneddon approximation of the Hertz indentation model.
Hepatic stellate cell isolation. Hepatic stellate cells were isolated from Sprague-Dawley rats as previously described 43 . Rat livers were enzymatically digested in situ by treatment with 0.4% pronase (Roche Diagnostics) and 0.04% collagenase II (Worthington). The liver slurry was diluted in minimal essential media (MEM) and filtered through cheesecloth. The cell suspension was then washed twice in 0.002% DNase (Worthington). Stellate cells were isolated by density gradient centrifugation with a 9% Nycodenz (Sigma) solution at 1400 g for 25 min, washed in MEM, and stored on ice until seeding on hydrogels.
Hepatic stellate cell culture on MeHA hydrogels. Prior to cell seeding, hydrogel thin films were incubated in PBS overnight at 37 °C and then sterilized under germicidal UV irradiation for 2 h. Sterilized hydrogels were then incubated in culture media for at least 30 min prior to seeding. Culture media consisted of phenol red-free M199 media (Invitrogen) supplemented with 10 v/v% fetal bovine serum (Sigma), 2 v/v% penicillin streptomycin (Invitrogen), and 1 v/v% fungizone amphotericin B (Invitrogen). Cells were seeded on hydrogel thin films in 6-well plates at a density of 5 × 10 3 cells/cm 2 for all experiments except for α -SMA inhibition experiments, which were seeded at 3.5 × 103 cells/cm 2 . Media was replaced 2-3 times per week. For α -SMA inhibition studies, culture media was supplemented with α -SMA blocking peptide (25 μ g/mL) and changed daily.
Time-lapse microscopy. Time-lapse microscopy (Olympus IX81 equipped with environmental chamber for live cell imaging) was used to track stellate cell spreading in real-time following hydrogel stiffening. Images of stellate cells cultured on soft and stiff static substrates, as well as in situ stiffened hydrogels were captured every 15 min for 24 h.
Cell imaging, staining, and quantification. Phase contrast images of stellate cells on hydrogels were acquired using a Zeiss Axiovert 200 inverted microscope (Hitech Instruments, Inc.). NIH ImageJ was used to measure cell spread areas. Hydrogels were prepared for immunocytochemistry by fixing cells in 10% buffered formalin for 15 min, permeabilizing in 0.1% Triton X-100 for 15 min, and blocking in 3% BSA in PBS for 1 h at room temperature. Cell-hydrogel constructs were then incubated with primary antibodies diluted in blocking buffer against α -SMA (mouse monoclonal anti− α -SMA clone 1A4 Ab, Sigma, 1:400), YAP (rabbit polyclonal anti-YAP, Santa Cruz Biotechnology, 1:200), p53 (mouse monoclonal anti-p53, Santa Cruz Biotechnology, 1:400), β -PDGFR (rabbit monoclonal anti-PDGF receptor beta, Cell Signaling Technology, 1:200), and/or β 1 integrin (rabbit monoclonal anti-integrin beta 1, Abcam, 1:200) overnight at 4 °C. For staining of the cell surface proteins β -PDGFR and β 1 integrin the permeabilization step was omitted. Hydrogels were washed three times in PBS and then incubated with appropriate secondary antibodies (AlexaFluor ® 488 goat anti-mouse IgG or AlexaFluor ® 568 goat anti-rabbit, Invitrogen) or rhodamine phalloidin to visualize F-actin (Invitrogen, 1:200) for 2 h at room temperature. Hydrogels were then washed twice with PBS, incubated with DAPI for 1 min (1:10000), rinsed again in PBS, and stored at 4 °C in the dark until imaging. Fluorescent images were acquired using an Olympus BX51 microscope (B&B Microscopes Limited). YAP/TAZ nuclear and cytoplasmic intensities as well as relative β -PDGFR and β 1 staining intensity were measured in NIH ImageJ.
Quantitative PCR. RNA was extracted from stellate cells on hydrogels using an RNeasy Plus Micro kit (Qiagen). RNA was reverse transcribed to cDNA using Super Script III Reverse Transcriptase (Invitrogen). Quantitative PCR was performed with a StepOnePlus instrument (Applied Biosystems) for targets including β 1 integrin (Itgb1), β -PDGFR (Pdgfrb), α -SMA (Acta2), and type I collagen (Col1a1). Ribosomal protein S12 (Rps12) and 18S ribosomal RNA (18S) were used for normalization. The sequences of the primers are presented in Supplementary Table 1.
Traction force microscopy. Traction forces exerted on hydrogels by hepatic stellate cells were measured as described previously 21 . Hydrogels embedded with fluorescent microspheres were fabricated as described earlier and fixed to culture plates with a UV-curable fixative (NOA68, Norland Products). Phase contrast and fluorescent images of multiple cells and embedded beads were captured at 20× magnification on a DeltaVision Deconvolution Microscope (GE Healthcare Life Sciences). Image sequences for each cell were taken before and after cell lysis with sodium dodecyl sulfate (SDS) buffer. All imaging was performed in an environmental chamber (37 °C, 5% CO 2 ). Data analysis (stack alignment, particle image velocitometry, and Fourier transform traction cytometry) was performed using a freely available plugin suite for ImageJ, created by Tseng et al. 44 , which was adapted from Dembo and Wang 45 . For Fourier transform traction cytometry, the Poisson's ratio of the hydrogel was assumed to be 0.45 and a regularization parameter of 10 −9 was used. Traction force vector maps were analyzed to determine the average force generation by each cell using a custom MATLAB script.