Effects of PDMS culture on stem cell differentiation towards definitive endoderm and hepatocytes

Abstract

The generation of human induced pluripotent stem cell (hiPSC) derivatives for regenerative medicine applications holds tremendous promise in treating various disorders. One critical target includes liver disease, in which the primary curative treatment is a cellular transplant aimed to restore the lost function of hepatocytes. In an effort to improve the differentiation of hiPSC-derived liver tissue, we manipulated the mechanical conditions of endoderm specification through directed perturbation of the cytoskeleton and through 2D substrate culture on viscoelastic materials. Through a combination of qRT-PCR, immunofluorescence staining, and functional assays, we found that mechanical cues can bias endoderm specification in an actomyosin and Yes-associated protein (YAP) dependent manner, unveiling new insights into mechanotransduction in germ layer specification and downstream maturation toward parenchymal cells.

1. Introduction

Cell responsiveness to changes in the microenvironmental niche guides differentiation, the ability to build complex tissues, and the maintenance of homeostasis. Signal transduction from the extracellular milieu includes sensing soluble chemical cues, matrix mechanics, and cell-cell interactions. Soluble factors induce signal transduction that leads to differentiation processes [1]. For signaling between neighboring cells, direct membrane-bound protein contacts such as gap junctions can be used to convey information [2]. Cells are able to sense the mechanical niche through bi-directional integrin-mediated signaling, wherein subunits of membrane-bound proteins engage with specific residues of the extracellular matrix (ECM), exerting force on bound integrin proteins, resulting in intercellular signaling cascades modifying cell states [3]. In addition to interaction with the ECM, membrane-bound integrin proteins are anchored to the intracellular actin cytoskeletal network, which impacts cell shape, migration, and signaling processes [4]. For example, the Rho GTPases, including RhoA, Rac1, and Cdc42, are activated by external stimuli and mediated by actin cytoskeletal remodeling. In these systems, stress fibers and focal adhesions are controlled through active RhoA, while Rac1 and Cdc42 lead to lamellipodia and filopodia formation, respectively. These mechanical signals are also conveyed in cells through the Yes-associated protein (YAP), which has been shown to control a variety of cellular processes, including differentiation [5–9].

A variety of techniques for perturbing cellular interactions with the ECM across various length scales in vitro, including manipulation of the cytoskeleton, integrins, and culture substrate, have been leveraged to understand how mechanotransduction controls differentiation [10]. For example, limiting ECM presentation with micropatterns has been shown to control the spatial maturation of hiPSCs towards cardiovascular lineages [11–14] while tools, such as engineered extracellular mechanics, have been used to prime mesoderm induction for enriched endothelial specification [15–17], organoid maturation [18,19], and specification of mesenchymal stem cells to brain, muscle, adipose, or osteogenic lineages [8,20]. During post-endoderm pancreatic differentiation, mechanical signaling triggered the upregulation of various pancreatic progenitor markers like NEUROG3, NKX2.2, and NEROD1, along with mature beta cell markers such as insulin secretion, GCG, and SST gene upregulation. Additionally, there was a downregulation of Alpha-fetoprotein (AFP), HNF6, KRT19, PRSS1, and CDX2 genes. These improvements to the cellular differentiation allowed for transplantable beta cells that demonstrated enhancements in blood glucose regulation in diabetic mouse models over 36 weeks of exposure, similar to primary human islet cells [19]. This mechanical regulation on cellular differentiation has also been demonstrated, along with the maturation of pluripotent cells in endothelial populations, through a compliant substrate culture in the initial mesoderm induction [7].

Traditionally, hiPSCs have been differentiated into hepatic lineages using chemical factors that activate and suppress cell signaling pathways, promoting definitive endoderm and foregut endoderm specification and eventually enriching hepatoblast populations [21]. Modulating mechanical cues during hepatic specification has been employed to functionally mature stem cell derivatives into cholangiocytes and hepatic populations, utilizing various approaches such as micropatterning, manipulation of cell junctions, and substrate stiffness [22–30]. However, the temporal effects of these changes have been investigated only to a limited extent. In this work, we combined chemical perturbation of the cytoskeletal architecture with viscoelastic substrate culture to modulate endoderm induction from hiPSCs. We found that coordinated regulation involving YAP signaling, actin cytoskeletal integrity, and extracellular matrix stiffness impacts the expression of definitive endoderm markers. By priming the mechanical state of progenitor populations with physiologically relevant polydimethylsiloxane substrates (PDMS), we observed an enrichment in the expression of definitive endoderm markers, leading to functional changes in downstream hepatic differentiation.

2. Methods

2.1. Stem cell culture

Induced pluripotent stem cells (ATCC—CYS0105) were maintained in either mTeSR Plus or mTeSR 1 Media (STEMCELL Technologies) following the manufacturer’s protocols. Briefly, cells were passaged <70 % confluence with a thin coating of ReLeSR (STEMCELL Technologies) for 7–10 min to allow for colony detachment. The cell solution was then neutralized with one standard well volume of mTeSR Plus or mTeSR 1 before cell count and subsequent seeding ranging between 20–50,000 cell/cm², depending on the experiment, on Vitronectin XF (STEMCELL Technologies) coated tissue culture plates (Cell Treat). Cells were seeded in 10 nM Y-27,632 dihydrochloride and maintained at 37 °C in 5 % CO₂ for growth.

2.2. PDMS cell culture

CytoSoft Discovery Kit (Advanced Biomatrix) PDMS cell culture plates were prepared for cell culture following the manufacturer’s protocol. First, the plates were washed three times in PBS before being coated with 3 mL of 40 μL/mL Vitronectin XF (Stem Cell Technologies) for one hour. Alternatively, plates were treated with oxygen plasma in a Harrick Plasma Cleaner on high for 45 s at 0.5 Torr before 1 mL of Vitronectin XF (Stem Cell Technologies) at 40 μL/mL was applied for one hour. Plates were then washed three times with PBS before being seeded with stem cells in mTeSR plus. Cells were then maintained in mTeSR plus until a confluence of ≈ 80 % following the manufacturer’s protocol using a definitive endoderm differentiation kit (Stem Cell Technologies). For all gene expression studies done with these plates, cells were lifted using a thin coating of TrypLE (Thermo Fisher) for 10 min before being neutralized in 1 mL mTeSR Plus and spun down at 300 × g for 5 min before the supernatant was aspirated and the cells lysed with TRIzol and stored at −80 °C until use.

The manufacturer has characterized these plates for their rheological properties using the following protocol. A layer of gold particles was deposited on a prepared slide before a layer of PDMS was added on top, followed by a second layer of gold particles. The slides were then centrifuged at various speeds and measured for the relative displacement of the top and bottom beads. This data was then tabulated, as discussed in Gutierrez et al. 2011, to provide Young’s modulus for each PDMS formulation.

2.3. Endoderm differentiation

Endoderm differentiation was carried out following the manufacturer’s protocol for the STEMdiff Definitive Endoderm kit (Stem Cell Technologies). When culture confluence reached ≈ 80 %, cells were washed with 1 standard well volume of PBS before being covered with one standard well volume of media 1 (basal media with supplements CJ and MR) overnight in a 37 °C 5 % CO2 incubator. For the next three days, the media was replaced with media 2 (basal media with supplement CJ) before the differentiation ended on day five. Differentiation was monitored daily to monitor cellular survival and morphology changes to ensure proper differentiation, then verified by immunofluorescence or qRT-PCR.

2.4. Hepatic differentiation

Following day five of endoderm differentiation, cells were passaged using one wash of PBS followed by a 5-minute treatment of TrypLE (Thermo Fisher) at 37 °C to detach cells; TrypLE was then neutralized with Media A (DMEM F-12 (Stem Cell Technologies), 10 % KnockOut Serum Replacement (Thermo Fisher), 1 % DMSO (Sigma-Aldrich), 0.5 % GlutaMAX (Gibco), 0.5 % MEM Non-essential amino acids (Gibco), and 100 ng/mL HGF (PeproTech)). Cells were then spun down at 300 × g for 5 min to pellet before the supernatant was aspirated, and the pellet was resuspended in Media A to count on a hemocytometer. Cells were then seeded on a Vitronectin XF-coated tissue culture plate (Cell Treat) at a density of 100,000 cells/cm² and maintained with daily media changes of Media A until day 13. On day 13 the media was replaced with Media B (DMEM F-12 (Stem Cell Technologies), 10 % KnockOut Serum Replacement (Thermo Fisher), 40 ng/mL Dexamethasone (Thermo Fisher), 0.5 % GlutaMAX (Gibco), 0.5 % MEM Non-essential amino acids (Gibco), and 100 ng/mL HGF (PeproTech)), which was used for days 13–15 with daily media changes. From days 16–21, cells were given HCM (Lonza HCM BulletKit) without rhEGF, supplemented with 20 ng/mL Oncostatin-M (PeperoTech). Past day 21, for the extended culture experiments, cells were maintained in HCM (Lonza HCM BulletKit) without rhEGF and with daily media changes.

For the full 21-day PDMS differentiation experiments, cells were passaged from the PDMS that they were on for endoderm differentiation, to fresh Vitronectin XF-coated PDMS plates on day 5. Cells were then continued through the differentiation as described above.

2.5. Immunofluorescence staining

Cells were washed with PBS before being fixed under 4 % PFA (Santa Cruz Biotechnology) for 10 min, followed by three washes of PBS. Cells were then blocked and permeabilized for 30 min in a blocking buffer solution of 0.1 % Triton X-100 (Sigma) and 0.5 % BSA (Thermo Fisher). Cells were then incubated at either room temperature for one hour or 4 °C overnight under primary antibody (Supplementary Table 1). Cells were then washed three times with PBS before incubating with secondary antibodies for one hour at room temperature with Phalloidin (1:500) and DAPI (1:1000) (Thermo Fisher) in a blocking buffer. Immunolabeled cells were imaged on a Keyence (BZ-X).

2.6. Quantitative RT-PCR

Total transcriptome from cell samples was collected by lysing cells with TRIzol reagent (Invitrogen) at various points in the differentiations at time points indicated in the text. RNA from this lysate was isolated with a PureLink RNA Mini Kit (Invitrogen) before being quantified with a Nano Quant (Tecan Spark). RNA was converted to complementary DNA (cDNA) using either the High-Capacity cDNA Reverse Transcription Kit or SuperScript IV VILO Master Mix with ezDNase Enzyme (Applied Biosystems) and used with TaqMan Fast Advanced Master Mix for qPCR (Applied Biosystems). Primers used in this study were used following Applied Biosystems instructions and compared against endogenous controls GAPDH and TBP for all samples (Supplementary Table 2). PCR was carried out using the Applied Biosystems One-Step Real-Time PCR Quant Studio 3 or 7. The relative expression of all genes was normalized in each sample to the amount of endogenous controls GAPDH and TBP. The values for these experiments were plotted as fold changes against a control sample and displayed with SEM bars of technical and biological replicant readings.

2.7. Albumin ELISA

Media was collected from days 16–21 of the hepatic differentiation and stored at −20 °C until use. A sandwich ELISA method was used with the coating antibody (Thermo Fisher A80–129A) diluted in coating buffer (0.175 g NaHCO₃ (Sigma-Aldrich), 0.02 g Na₂CO₃ (Thermo Fisher), 50 mL water, pH 9.6) to a final concentration of 1:100 then added to an ELISA plate (Stellar Scientific, NST-504,201) and left to incubate for 1 hour at room temperature. Plates were then washed three times with wash buffer (3.02 g Tris–HCL (Thermo Fisher), 4.06 g NaCl (VWR), 0.25 mL Tween 20 (Sigma-Aldrich) in 500 mL water, pH 8.0), next the plates were blocked for room temperature in blocking buffer (0.302 g Tris–HCL (Thermo Fisher), 0.406 g NaCl (VWR), 0.5 g BSA, 50 mL water, pH 8.0) before being stored overnight at 4 °C. Plates were then washed three times with wash buffer before standards, HSA (Sigma Aldrich) dilutions, and samples prepared in sample dilutant (0.302 g Tris–HCL (Thermo Fisher), 0.406 g NaCl (VWR), 0.5 g BSA, 0.025 mL Tween 20 (Sigma-Aldrich), 50 mL water, pH 8.0) and placed into the plate for 1 hr room temperature shaking at 100 rpm. Plates were then washed three times before the HRP-conjugated antibody (Thermo Fisher A80–129P) was added at 1:100,000 in sample diluent to be incubated dark for 45 min with 100 rpm shaking. Still in the dark, the plates were washed five times with wash buffer before TBM substrate (Thermo Fisher) was added and quenched after 7 min, based on the standard curve reaction point, with 0.5 N HCL (VWR). Absorbance at 450 nm was read using a Spark Multimode Microplate Reader (Tecan).

2.8. AFP ELISA

Media was collected from days 16–21 of the hepatic differentiation and stored at −20 °C until use. An AFP ELISA Kit (RayBiotech) was used following the manufacturer’s protocol. Briefly, media and standard samples were diluted before addition to a supplied coated ELISA plate specific for AFP and then left to shake for 2.5 h. before washing. Samples were then covered with a biotin-bound antibody and left to react for 1 hour, followed by washes. Next, the HRP-streptavidin solution was added for 45 min, then washed and replaced with a TBM reagent solution for 30 min. The reaction was then halted with a stop solution and read immediately at 450 nm using a Spark Multimode Microplate Reader (Tecan).

2.9. Urea colorimetric assay

Media was collected from hepatic differentiation from day 16 to the end of the experiment and stored at −20 °C. The urea concentration in the cellular supernatant was measured using the Stanbio^™ BUN Liquid Reagent for Diagnostic Set (EKF Diagnostics) following the manufacturer’s protocol. Concentration calculations were always based on standards read from the same plate as the samples to be calculated.

2.10. CYP3A4 activity

At day 21 of the hepatic differentiation, HCM media (Lonza HCM BulletKit) without rhEGF was supplemented with either vehicle control or 50 μM Dexamethasone for 48 hours and replaced 24 hours later. CYP3A4 activity was then determined using the P450-GloTM CYP3A4 Assay (Promega) following the manufacturer’s protocol. Sample wells were individually standardized for cell viability using the Cell-Titer-Glo Luminescent Cell Viability Assay (Promega) to remove variability. The assay was performed following the manufacturer’s protocol.

2.11. Graphs and statistics

All analyses were conducted in at least technical triplicate with each biological replicate (n) mentioned in the figure legend. qRT-PCR was analyzed using biological duplicate and triplicate readings. One-way and two-way ANOVA with Tukey post hoc tests or Student’s t-tests were performed to determine significance as listed in figure legends (GraphPad Prism 10.0.2).

3. Results

3.1. Generation of definitive endoderm progenitor cells

We first differentiated hiPSCs towards a definitive endoderm (DE) state using polystyrene-coated tissue culture plates (TCP) coated with vitronectin, which served as a baseline for later comparison to compliant viscoelastic matrices (Fig. 1A). To confirm successful specification to a DE state, cells were fixed each day of the differentiation to monitor protein and gene expression levels during the maturation process (Fig. 1B and C). Through immunofluorescence staining, we observed that on day 3 of the differentiation, the definitive endoderm marker forkhead box protein A2 (FOXA2) was increased in a portion of cells and became the predominant cell type by day 5 (Fig. 1B). Following this trend, there was a clear decrease in SRY-box 2 (SOX2) positive cells from day 3 onward, depicting an overall loss of pluripotency in the population. To monitor the efficacy of the differentiation, mRNA levels were measured, demonstrating downregulation of pluripotency markers, NANOG, and SOX2, and upregulation of definitive endoderm markers FOXA2, transcription factors GATA4, and SRY-box 17 (SOX17) (Fig. 1C).

Fig. 1. Inducing Definitive Endoderm Specification in hiPSCs.

(A) Stepwise schematic of endoderm induction. (B) Immunofluorescence images of the day-by-day induction of FOXA2 (red) and downregulation of SOX2 expression (green) along five days of differentiation, n = 3. (C) Gene expression of key definitive endoderm markers (GATA4, SOX17, and FOXA2) and pluripotency markers (NANOG and SOX2), n = 2. (D) Representative YAP localization as a function of time, n = 3. (E) Quantification of YAP nuclear to cytoplasmic expression through endoderm induction, n = 3. (F) Changes in YAP-mediated CTGF expression during endoderm specification, n = 2. Scale bar in B is 200 μm; scale bar in D is 100 μm. Data are shown as mean ± SEM, one-way ANOVA results shown are compared to day 1 with *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, if no * is shown P > 0.05. n refers to the number of biological replicates.

The pleiotropic transcriptional regulator and main effector of the Hippo pathway, YAP, was measured along DE specification in our hiPSCs. In addition to its role in mechanotransduction, YAP is a primary facilitator of proliferation, cell fate, apoptosis, and pluripotency and has previously been investigated in several cellular maturation events, such as in mesenchymal stem cell differentiation and mesodermal induction [7,20,31]. To this end, we measured YAP dynamics using immunofluorescence microscopy. Throughout the five-day differentiation, YAP was differentially shuttled into or out of the nucleus, likely as a control of cellular transcriptional gene regulation related to the loss of pluripotency and induction to a definitive endoderm state (Fig. 1D). In this process, a trend of decreased nuclear/cytoplasmic YAP localization was observed in the initial portions of the differentiation before localization returned to near-baseline levels at the end of the differentiation on day 5, suggesting that its transcriptionally regulated genes required downregulation for initial endoderm induction to take place (Fig. 1E). The function of YAP as a regulatory controller of gene expression was verified through mRNA levels of a known downstream target, the connective tissue growth factor (CTGF). We found that at time points where YAP was nuclear localized, there was increased transcription activity of CTGF (Fig. 1F). These findings of YAP localization activity during endoderm induction led us to hypothesize that this cell transition could be mechanically sensitive.

3.2. Actin depolymerization improves endoderm specification

One of the ways in which cells interpret the mechanical context of their niche is through integrin-mediated ECM sensing and cytoskeletal structure. Cells rely on their cytoskeleton for mechanical sensing of their growth environment through several pathways, such as Rho GTPase and Notch, where components of the cytoskeleton act as a scaffolding to give a semi-rigid structure for mechanical signal transduction [32,33]. To investigate the effects of cytoskeletal integrity during endoderm specification, we added cytoskeletal modifying drugs, namely Cytochalasin D and Latrunculin A, which inhibits actin polymerization, and Nocodazole to prevent microtubule polymerization at the onset of differentiation for 24 hours, returning to standard differentiation media for the remainder of the 5-day differentiation assay (Fig. 2A). The result from this initial priming at the point of differentiation induction was that following an initial softening, pluripotent markers at both the protein (SOX2) and gene (NANOG) levels were significantly decreased compared to control differentiation. Along with this change, DE markers at the protein (FOXA2) and gene (GATA4) levels were significantly upregulated at the terminal day 5 timepoint of the differentiation in Cytochalasin D and Latrunculin A treated cells compared to control differentiation conditions (Fig. 2B and C.i). In contrast to these findings, cells primed with Nocodazole, leading to increased contractility [34,35], had increased levels of pluripotent markers and decreased levels of DE markers at the end of the differentiation assay (Supplemental Figure 1). Taken together, our results indicate that priming hiPSCs to a softer state elicits improved endoderm specification. To gain additional mechanistic insight into the role of intracellular mechanosensing processes during endoderm specification, we investigated the role of YAP in this cytoskeletal modulated differentiation. Following adding Nocodazole during day one of the differentiation, we found that the gene expression of YAP-regulated genes CTGF and CYR61 were significantly upregulated on day three (Fig. 2C.ii). In contrast, there was not a significant upregulation of these genes under culture conditions with Latrunculin A or Cytochalasin D. To further elucidate YAP’s influence in this differentiation, we added verteporfin (Vert) to the differentiation media at separate time points along the differentiation, assay. Verteporfin, a photosensitizer drug, sequesters YAP in the cytoplasm through upregulating chaperone protein 14–3–3σ [36]. This, in turn, disrupts the YAP nuclear-shuttling program, implicated in cellular mechanotransduction in stiff environments. Verteporfin addition in the first 24 hours of differentiation led to cell death; however, adding at later stages of differentiation allowed for sustained cellular viability over a 48-hour period during the end of the differentiation (Fig. 2D). When we compared gene expression levels for control and endpoint verteporfin-treated cells, a significant modification to YAP transcriptional activity could be observed through decreases in YAP-dependent CTGF expression. DE genes SOX17, GATA4, and FOXA2 were significantly upregulated following verteporfin addition to the differentiation (Fig. 2F). Along with this, we observed a strong decrease in the gene expression of pluripotency markers NANOG and SOX2 (Fig. 2E). These results further demonstrate that the manipulation of known mechanosensitive pathways improves endoderm commitment and loss of pluripotency.

Fig. 2. Cytoskeletal Modulation Impacts Definitive Endoderm Induction.

(A) Stepwise schematic of endoderm induction and time points where small molecule drugs were introduced to affect the cytoskeletal tension of differentiating cells in the first 24 hours of the differentiation. (B) Immunofluorescence images of the cells after five days of differentiation after treatment with 100 nM Nocodazole (Noco) or actin depolymerization with either 100 nM Cytochalasin D (Cyto D) or 10 nM Latrunculin A (Lat A) during the first 24 hours of differentiation, with FOXA2 (red) being more prevalent in cytoskeletal softened conditions and SOX2 (green) being more highly expressed in cytoskeletal stiffened conditions, n = 3. (C.i) Gene expression at the terminal point of the differentiation after priming with cytoskeletal stiffening and softening for definitive endoderm markers GATA4, FOXA2, and SOX17 and pluripotency marker NANOG, n = 2. (C.ii) Gene expression at day 3 of the differentiation after priming with cytoskeletal drugs for YAP-regulated genes CTGF and CYR61, n = 4. (D) Stepwise schematic of endoderm induction with 250 nM verteporfin (Vert) addition during the final 48 hours of differentiation. Gene expression of (E) pluripotency markers (NANOG and SOX2) and (F) key endoderm markers (GATA4, SOX17, FOXA2) and YAP downstream markers CTGF after 250 nM verteporfin or control (+ or −) exposure at days 3 and 4 of the differentiation, n = 2. Scale bar is 100 μm. Data are shown as mean ± SEM, one-way ANOVA results are shown in C, t-test results are shown in E and F with *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, if no * is shown P > 0.05. n refers to the number of biological replicates.

3.3. Soft viscoelastic matrices increase endoderm specification in a cytoskeleton-dependent manner

To further isolate the mechanical impacts of this differentiation, we seeded cells on substrates with known stiffness without cytoskeletal modulation before differentiation without soluble factor-based cytoskeletal perturbation. In addition to TCP (~1–3 GPa)[10,37,38], viscoelastic polydimethylsiloxane (PDMS) substrates were used to simulate various tissue compliance ranging from 64, 8, and 0.2 kilopascals (kPa) mimicking tumor, liver, and soft (mucus/brain) tissues, respectively (Supplemental Fig. 2). Across the conditions, significant decreases in pluripotent markers NANOG and SOX2 were observed with decreasing stiffness (Fig. 3A and Supplemental Fig. 3). Similarly, definitive endoderm markers GATA4, FOXA2, and SOX17 were upregulated significantly as the substrate stiffness decreased, linking the degree of endoderm induction to the substrate stiffness. These results were consistent with the findings from cytoskeletal modulation; to this end, we next investigated if Nocodazole treatment could ablate the improvements seen with soft PDMS culture conditions. For these experiments, an initial 24-hour priming of Nocodazole was conducted during the beginning of the differentiation, which led to dramatic changes in the rate of definitive endoderm gene expression at the endpoint of the culture. Following the initial priming with Nocodazole, nearly all definitive endoderm marker expression improvements with decreased substrate stiffness were ablated, showing that even temporary changes to cellular mechanosensitive measurements could disrupt this differentiation process (Fig. 3B and Supplemental Fig. 4).

Fig. 3. Stiffness Dependent Endoderm Induction.

(A,) Control gene expression for endoderm markers (GATA4, SOX17, FOXA2) and pluripotency markers (NANOG, SOX2) at the endpoint of a differentiation carried out on PDMS gels of various stiffnesses, n = 2. (B) Nocodazole-treated differentiation as marked in Figure 2A, with 100 nM Nocodazole added for 24 hours at the start of the differentiation. Individual plots of these genes with one-way ANOVA plots can be found in Supplemental Figures 3 and 4 (C.i-iii) Gene expression of key definitive endoderm markers (GATA4, SOX17, and FOXA2) and pluripotency markers (NANOG and SOX2) at day 5 of the differentiation for cells grown on PDMS substrates at the listed stiffnesses with or without 250 nM verteporfin on days 3 and 4 of the differentiation, n = 2. Data are shown as mean ± SEM, two-way ANOVA results are shown with *P < 0.05, **P < 0.01,***P < 0.001, ****P < 0.0001, if no * is shown P > 0.05. n refers to the number of biological replicates.

After establishing that tuning substrate mechanics alters endoderm maturation kinetics, we next investigated if YAP inhibition on PDMS substrates would change this change in differentiation. At the highest stiffnesses tested in these conditions (i.e., 64 kPa), the inclusion of Vert led to improvements in definitive endoderm GATA4, showing significant expression upregulation, while FOXA2 and SOX17 were not significantly altered (Fig. 3C.i). At 8 kPa, there were again no significant changes in FOXA2 or SOX17 expression, but GATA4 levels decreased (Fig. 3C.ii). For the softest condition tested, FOXA2 was significantly increased, however GATA4 was decreased (Fig. 3C.ii). This change in gene expression was most substantial under the stiffest conditions tested, indicating that these mechanical responses vary in supraphysiologically stiff environments (Supplemental Figure 5). Improvements in the downregulation of pluripotency markers were observed across all stiffnesses, again demonstrating that mechanical perturbation along the endoderm differentiation away from overly stiff TCP culture improved the loss of stemness for all conditions.

3.4. Hepatic differentiation of PDMS-primed endoderm cells

With the improvements observed at the endoderm stage at both the protein and gene levels across cytoskeletal modulation and compliant substrate culture, we next investigated if these improvements would lead to later downstream improvements of an endoderm lineage. For this, we conducted a 21-day hepatocyte differentiation, with the first five days being an endoderm induction on substrate stiffness-controlled plates before sixteen days of hepatic induction on TCP plates (Figs. 4A and B). We first investigated the ability of cells initially cultured on PDMS during endoderm induction to express characteristic proteins of early and mature hepatocytes. We observed that across all initial priming conditions, immature hepatic marker alpha-fetoprotein (AFP), and the hepatic markers, hepatocyte nuclear factor 4 alpha (HNF4a), alpha-1 anti-trypsin (A1AT), and albumin showed that the initial modification to the endoderm induction did not prevent hepatic like cell maturation (Fig. 4C–G). At the protein level, immature hepatic expression of AFP was inconsistent across stiffnesses, demonstrating the highest rate of positive cells on plastic and the lowest levels on 8 and 0.2 kPa cultured cells (Fig. 4D). In addition, when testing for cholangiocyte marker SOX9, we found that there was significantly less SOX9 expression in cells primed on PDMS before hepatic maturation (Supplemental Fig. 6).

Fig. 4. Hepatic Differentiation After Mechanical Priming Endoderm Induction.

(A) Schematic of hepatic differentiation following endoderm differentiation on PDMS substrates of defined stiffness. (B) Day 21 images of mature hepatocyte-like cells prior to fixation. (C) Immunofluorescence images post hepatic differentiation demonstrate hepatic markers in red and green, with the nucleus in blue and the actin cytoskeleton in white. Immature hepatic markers alpha-fetoprotein (AFP) (top), mature hepatic marker HNF4a (middle), and mature hepatic markers alpha-1 antitrypsin (A1AT) and albumin (ALB) (red) (bottom). (D) Quantification of AFP-positive cells, n = 4. (E) Quantification of HNF4a positive cells, n = 4. (F) Quantification of A1AT mean fluorescence intensity (MFI) per cell, n = 4. (G) Quantification of ALB (MFI) per cell, n = 4. Scale bars are 50 μm. Data are shown as mean ± SEM, one-way ANOVA results shown with *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, and if no * is shown P > 0.05. n refers to the number of biological replicates.

Mature markers of hepatocytes A1AT (encoded by gene serpina1) and albumin both showed slight upregulation at the protein level (Fig. 4F and G). Hepatocyte identity was also measured through secreted urea and albumin, which showed trends toward soft condition primed cells having increased hepatic marker secretions (Fig. 5A and B). During the differentiation, both 8 and 64 kPa cultured cells had higher levels of albumin secretion compared to TCP cultured cells, and 0.2 kPa cells consistently had the lowest levels (Fig. 5A). At the terminal day 21 timepoint of the differentiation, high levels of urea above the TCP baseline were observed for cells initially cultured on 8 kPa PDMS, with 64 and 0.2 kPa cells having lower secretion (Fig. 5B and Supplemental Figure 7). In addition, comparisons of secreted mature hepatic marker ALB to immature hepatic marker AFP showed trends to a more mature cellular state on soft culture conditions (Fig. 5C). Mature CYP3A4 and immature CYP3A7 gene expression measurements at day 21 of the differentiation showed trends to a more mature cellular state in 64 and especially 8 kPa cultured cells (Fig. 5D and Supplemental Figures 9 and 11). Finally, albumin and serpina1 mRNA expression was highest 0.2 kPa substrates (Fig. 5E). CYP3A4 activity levels were measured following 48 hours of drug detoxification induction from Dexamethasone exposure for the day 21 hepatocytes to measure the functionality of the generated cells. In these trials, 8 kPa-grown cells were found to have the highest base and induced CYP3A4 activity levels, with 64 and 0.2 kPa-cultured cells having nonsignificant activation levels (Supplemental Fig. 10). Lastly, we maintained a population of cells primed from PDMS before they transitioned to TCP out to 31 days from the start of the differentiation. We found that while the 64 and 8 kPa culture cells did tend to be higher in ALB and urea secretion, both did decrease steadily over time (Supplemental Fig. 8).

Fig. 5. Functional Differences of Hepatic Specification Following Definite Endoderm Priming on PDMS.

(A) Secreted albumin levels from days 16 to 21 of differentiating hepatocytes. (B) Secreted urea levels from day 21 of differentiation, n = 4. (C) Secreted levels of albumin divided by secreted levels of AFP in cell media measured on day 21, n = 3. (D) Relative gene expression of mature hepatocyte marker CYP3A4 compared to immature hepatocyte marker CYP3A7, n = 2. (E) Gene expression for definitive hepatic markers serpina1 and albumin on day 21 of differentiation, n = 3. Data are shown as mean ± SEM, one-way ANOVA results are shown compared to plastic controls with *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, and if no * is shown P > 0.05. n refers to the number of biological replicates.

3.5. Hepatic differentiation on PDMS-maintained cells

Following the results showing that initial compliant substrate culture can prime cells at the endoderm stage for downstream hepatic differentiation, we next investigated how maintaining compliant substrate culture could affect the cells along the entire differentiation. For this culture, we first seeded hiPSCs on PDMS before taking them through a five-day endoderm induction; at the end of this differentiation, the cells were passaged to new PDMS wells of matched stiffness to continue the differentiation to hepatocytes (Fig. 6A). From day 16 to 21, cell media was collected daily and measured for albumin secretion. Cells cultured on 8 and 64 kPa substrates maintained a higher level of albumin secretion on average (Fig. 6B). For these cells, the ratio of albumin to AFP also showed increased trends for 8 and 0.2 kPa cultured cells above plastic and 64 kPa culture conditions (Fig. 6C). This change was in large part due to the change in AFP secretion, showing the highest levels for the 64 and 8 kPa gel conditions (Supplemental Figure 11). We found that the change in CYP3A4 to CYP3A7 gene expression for cells differentiated consistently on PDMS showed only minimal improvements for 64 kPa gels, with 8 and 0.2 being very close to unchanged from plastic cultured cells (Fig. 6D). Mature hepatic genes were also measured for their expression levels at the end of the differentiation on day 21, and we found that albumin was significantly upregulated for all soft substrates compared to plastic (Fig. 6E.i). This trend was similar for the mature hepatic markers CYP3A4, HNF4a, and serpina 1 and showed the highest levels of expression in 64 and 8 kPa culture conditions, but still were significantly upregulated in 0.2 kPa compared to traditional plastic culture (Figs. 6E.iii, iv, and Supplemental Figure 9). Immature hepatic markers AFP and CYP3A7 were also upregulated following PDMS culture, with 64 and 8 kPa displaying the highest expression, followed by 0.2 kPa, all of which were significantly higher than traditional plastic culture (Fig. 6E.ii and Supplemental Figure 9).

Fig. 6. iPSC-Derived Hepatocytes Through PDMS Culture.

(A) Schematic of hepatic differentiation on PDMS substrates of defined stiffness. (B) Secreted albumin levels from days 16 to 21 of differentiating hepatocytes, n = 3. (C) Secreted albumin/AFP on day 21 of differentiation, n = 2. (D) Gene expression of mature hepatic marker CYP3A4 compared to immature CYP3A7 gene expression on day 21 of differentiation, n = 2. (E) Relative gene expression of markers (i) albumin, (ii) AFP, (iii) HNF4a, and (iv) serpina1, n = 3. Data are shown as mean ± SEM, one-way ANOVA results are shown against the plastic with *P < 0.05, **P < 0.01, ****P < 0.0001, and if no * is shown P > 0.05. n refers to the number of biological replicates.

4. Discussion

Pluripotent stem cells can develop into any type of adult tissue, demonstrating immense potential. This encompasses studying early development, modeling patient-specific diseases, drug discovery, and the development of cell therapies [39]. Traditional approaches have leveraged stepwise addition of chemical factors to convert pluripotent populations to the desired tissue of choice. However, mechanical cues have been shown to significantly alter fate choice decisions [19,40,41]. This work expands on these studies and demonstrates that a low cellular tension state significantly improves definitive endoderm differentiation from hiPSCs. Here, we first investigated the role of cytoskeletal modulation and YAP inhibition in mediating the degree of pluripotency loss and endoderm induction. We found that Nocodazole treatment significantly impaired endoderm induction, while verteporfin improved definitive endoderm gene expression on TCP. This finding was further exemplified by seeding hiPSCs on defined PDMS substrates and investigating endoderm differentiation. We compared the differentiation potential between standard TCP plates and supraphysiologically stiff 64 kPa, hepatic tissue-relevant 8 kPa, and soft 0.2 kPa PDMS substrates [20, 42]. These tests showed clear impacts of external mechanical forces on the efficacy of endoderm induction through changes in definitive endoderm gene upregulation and downregulation of pluripotent markers. Further, when these compliant substrates were paired with Nocodazole addition, the improvements to endoderm gene expression were notably decreased. This was also further investigated by adding verteporfin to PDMS cultured cells. While differentiation efficiency was improved at the stiffest TCP condition, the improvement was minor in 8 kPa and 0.2 kPa experimental conditions. This would indicate that hiPSCs cultured on a stiff substrate benefit from transcriptional decreases in YAP target genes, evidenced by a reduction of CTGF expression. Next, we explored if the improved expression of endoderm markers could enhance downstream differentiation towards hepatocyte-like cells. This rationale has been previously supported by studies showing that transitioning iPSC-derived immature hepatocytes to soft culture conditions or maintaining mature hepatocytes on hydrogel substrates improves and maintains cellular hepatic function [22,25]. Here, TCP and PDMS cultured endoderm cells were passaged to standard TCP plates before continuing through a 21-day hepatocyte differentiation. Across all stiffness conditions, cells could attach and continue through the differentiation, verifying that the initial changes to the cell culture procedure did not lead to cellular inhibition of hepatic differentiation. The resulting hiPSC-derived hepatocyte like cells expressed several key hepatocyte markers, which varied depending on the substrate in which the endoderm was specified. With this approach, we find a trend of increased hepatic functionality when the definitive endoderm was specified on soft substrates. These results were also sustained by differentiating iPSCs to hepatocyte-like cells solely on PDMS, where we found an increase in the relative amounts of albumin to AFP via ELISA and transcriptional analysis. In addition, the mRNA level of hepatic markers was improved in the PDMS culture compared to TCP plates. Collectively, this study demonstrates that the mechanical context of differentiating hiPSCs can impact definitive endoderm induction and downstream hepatic specification.

Supplementary Material

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.actbio.2025.05.017.

Statement of significance:

The translational potential of using human induced pluripotent stem cell (hiPSC) derived hepatocytes to therapeutically improve impaired liver function holds great clinical promise. However, challenges remain in efficiently differentiating functional hepatocytes with mature marker expression. In an effort to improve the differentiation efficiency of hepatocytes, the role of early mechanosensing mechanisms was investigated in the specification of hiPSCs to definitive endoderm progenitor populations. Through a combination of cytoskeletal modulation, control of mechanoresponsive, yes-associated protein expression, and culture on physiologically compliant PDMS substrates, we found that soft environments not only improve progenitor specification but also impact the downstream functionality of differentiated hepatocytes. These results contribute to the collective appreciation that mechanical cues are critical in developmental processes.

Data availability

The data supporting this study’s findings are available upon request from the corresponding author Q.S.