Gelatin from porcine skin (Type A, ~300 g Bloom), methacrylate anhydride, polycaprolactone (PCL), Irgacure 2959 (I2959), triethanolamine, and SpectraPor 12–14 kDa molecular weight cutoff (MWCO) dialysis tubes were purchased from Millipore Sigma (St. Louis, MO, USA). Phosphate buffered saline (PBS), sodium chloride (NaCl), and dichloromethane (DCM) were purchased from VWR International (Wayne, PA, USA). Polydimethylsiloxane (Sylgard 184, PDMS) and 8 mm biopsy punches (Accu-Punch) were purchased from Electron Microscopy Sciences (Hatfield, PA, USA). Minimum Essential Medium (αMEM), fetal bovine serum (FBS), and penicillin/streptomycin (P/S) were purchased from Gibco (Waltham, MA). Stains for actin (phalloidin, Alexa Fluor 488) and double stranded DNA (Hoescht 33342) were purchased from Invitrogen (Carlsbad, CA, USA). Primary YAP antibody (SC-376830) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA) and secondary antibody (Alexa Fluor 568) was purchased from Invitrogen (Carlsbad, CA, USA).

GelMe was synthesized using a modified synthesis scheme previously described (19, 20). Briefly, 10 wt% gelatin was dissolved in a PBS stir bath (50°C, 1,200 RPM) for 20 min. 5 wt% of methacrylate anhydride was added to the gelatin solution at a rate of 0.5 ml/min while stirring at 50°C and was reacted for 1 h. The reaction was quenched by adding excess (4x) prewarmed PBS and GelMe was dialyzed (SpectraPor, 12–14 kDa MWCO) for 5 days at 40°C, frozen, and lyophilized. The degree of methacrylation was determined with ¹H NMR spectroscopy and was found to be 58.2% (Supplementary Figure 1).

GelMe solutions (5 wt% GelMe and 0.05 wt% photoinitiator I2959 in PBS) were pipetted into PDMS cylindrical molds (8 mm diameter, 2 mm height) and photopolymerized with UV light (10 mW/cm², 10 min). For compression testing, GelMe hydrogels were incubated in PBS overnight at 37°C and compressed with a Shimadzu EZ-SX Mechanical Tester up to 30% strain at a strain rate of 10%/min. The elastic modulus was calculated using the slope of the formed stress-strain curve between 10 and 20% strain.

Human bone marrow-derived MSCs (Lonza, PT-2501, Lot 684888, 33-year-old male) were cultured in 100 mm cell culture dishes in Growth Medium (αMEM supplemented with 10% FBS and 1% P/S) in a humidity-controlled environment under 5% CO₂ and 37°C. MSCs were passaged upon reaching 80% confluence and MSCs were used at passage 2 or passage 3 for all experiments.

MSC-laden GelMe solutions (1E6 MSCs/ml, 5 wt% GelMe, and 0.05 wt% I2959 in Growth Medium) were pipetted into PDMS cylindrical molds (8 mm diameter, 2 mm height) and photopolymerized with UV light (10 mW/cm², 10 min). Previous studies using comparable photopolymerization settings to crosslink hydrogels showed no impact on viability or cell function of encapsulated cells (6, 21). To evaluate cellular remodeling kinetics, free-swelling hydrogels were cultured in Growth Medium for 1, 3, 5, or 7 days and stained for actin (phalloidin 40 min, 1:100), YAP (primary 1:200, overnight; secondary 1:200, 2 h), and double stranded DNA (Hoescht, 5 min, 1:2,500). Samples were then imaged with a Nikon A1 confocal microscope and images were analyzed with ImageJ (Version 1.53j) software (NIH, Bethesda, MD, USA).

To evaluate the effects of compression on spatial differences of MSC morphology and YAP signaling, a custom compression device was developed (Supplementary Figure 2). The device supports a standard 24-well plate within it, and it can maintain up to 24 hydrogel samples under varying degrees of compression using nylon bolts with a cylindrical head cap. To compress, MSC-laden GelMe hydrogels were affixed to thiolated coverslips which were then placed in wells of a 24-well plate. A nylon bolt then was then lowered from above until it contacted the top of the hydrogel and was rotated (1/6 rotation = 325 μm) until 25 or 50% compression was achieved. Hydrogels were held under compression for 3 days prior to fixation and staining. To determine the Edge and Center regions for imaging and analysis, stained hydrogels were placed on a thin glass coversip and the edge of the hydrogel along the x-axis (left to right) was identified. From there, the hydrogel was separated into Edge regions (distance 0–2 mm and 6–8 mm from the edge in the x-direction) and a Center region (distance 2–6 mm from the edge in the x-direction) for imaging and analysis.

Porous PCL scaffolds were formed using a salt-leaching technique outlined in Supplementary Figure 3. Briefly, 30 wt% PCL was dissolved in 70 wt% DCM. Salt crystals (NaCl) with diameters ranging between 212 and 500 μm were acquired using salt sieves. PCL solutions were mixed with either 8:1, 12:1, or 15:1 NaCl-to-PCL weight ratio representing Low, Med, and High porosity, respectively. The NaCl/PCL mixture was poured into a mold, flattened, and DCM was allowed to evaporate for 24 h in a fume hood. Post-DCM evaporation, a biopsy punch (8 mm diameter) was used to create cylindrical scaffolds, which were then placed in conical tubes containing purified water in a revolver lab rotator for 5 days (water changed 2x/day). This last step forms PCL scaffold porosity due to salt crystals leaching out of the bulk PCL scaffold.

GelMe solutions (1E6 MSCs/ml, 5 wt% GelMe, and 0.05 wt% I2959 in Growth Medium) were perfused into porous PCL scaffolds (Supplementary Figure 4). Briefly, GelMe solution (500 μl) was added to a 3 ml Luer lock syringe and loaded onto a syringe pump. The needle was inserted into a segment of sterile Tygon tubing (8 mm inner diameter), containing the porous scaffold. All connections were secured with parafilm. GelMe solution was then perfused into the porous scaffold at a rate of 1 ml/min, the porous scaffold-hydrogel composite was carefully removed from the Tygon tubing, and the GelMe within the construct was photopolymerized with UV light (10 mW/cm², 5 min per side). Samples were then placed in wells of a 24-well plate, washed in Growth Medium 3 times, and incubated in Growth Medium for 3 days. Lastly, samples were stained for actin, YAP, and double stranded DNA, imaged with a Nikon A1 confocal microscope, and images were analyzed with ImageJ software.

Image analysis was performed for single cells using ImageJ software. For each cell, three metrics were calculated: Cellular Volume, Sphericity, and Nuclear YAP. Cellular Volume and Sphericity (line and sphere have values of 0 and 1, respectively) for each cell were determined by converting 3D image stacks of the actin channel into binary objects using Otsu's intensity-based thresholding method, followed by ImageJ's 3D Objects Counter function. Nuclear YAP was determined by converting 3D image stacks of the actin and nuclei channels into binary objects, which were then superimposed with the YAP channel to get 3D image stacks of cytosolic and nuclear YAP. Using ImageJ's 3D Objects Counter function, the sum of pixel intensity values for the nuclear and cytosolic YAP was determined and normalized to nuclear and cytosolic volumes. Then, nuclear YAP was calculated by taking the ratio of normalized nuclear to cytosolic YAP.

All MSC experiments were carried out in at least three independent hydrogel (Figures 1C–E, 2D–F) and porous scaffold-hydrogel composite (Figures 3D–F) replicates. All single cell analyses were performed with at least 50 cells per group. For compression testing of hydrogel and porous scaffold-hydrogel composites (Supplementary Figure 5), at least 6 constructs were measured per group. Pore diameter measurements (Figure 3B) were from at least 6 porous scaffolds, and at least 50 individual pores were measured per group. Bar graphs represent mean ± standard error of the mean and values reported in the text are average ± standard deviation (SD). GraphPad Prism 9 was used for all statistical analysis (GraphPad Software Inc. San Diego, CA, USA). Analysis of variance (ANOVA) followed by Tukey's post-hoc test was performed on all data sets. Significance is indicated by ns (no significance, p > 0.05), ^*p < 0.05, ^**p < 0.01, and ^***p < 0.001.

MSCs encapsulated in GelMe hydrogels are able to spread via MMP degradation (19, 20), and this is a time-dependent process (Figure 1A). For MMP-mediated hydrogel remodeling, the rate and extent of 3D cellular spreading is dependent on the concentration of degradable material. For instance, MSCs encapsulated in hyaluronic acid hydrogels crosslinked with a low concentration of MMP-peptides spread after 1 week in culture, whereas MSC spreading is restricted at higher MMP-peptide concentrations (7). To determine how long it takes encapsulated MSCs to degrade and remodel GelMe hydrogels, MSCs (1E6 cells/ml) were encapsulated in GelMe hydrogels (5 wt%) for 1, 3, 5, or 7 days. At each time-point, resident MSCs were evaluated for changes in morphology and YAP signaling. At 1- and 3-day time-points, MSCs retained a spherical morphology. However, between 3 and 5 days, MSCs began to extend protrusions indicative that local cellular remodeling was taking place (Figure 1B). After 1 day in culture, MSCs encapsulated in GelMe hydrogels had an average volume of 5,100 ± 2,500 μm³, and at 3 days there was a non-significant increase in MSC volume to 8,200 ± 3,800 μm³. Between 3 and 5 days, there was a significant increase in cell volume to 14,200 ± 8,900 μm³, and between 5 and 7 days the volume increased even further, to 18,300 ± 12,200 μm³ (Figure 1C). It is interesting to note that not only did the average cell volumes increase over time, but so did the standard deviation. This suggests that initially, MSC remodeling is highly consistent on a per-cell basis, and once cells surpass a threshold of remodeling, the ability (or inclination) for cells to continue spreading varies significantly on a single-cell level with increasing cell culture time.

Between 1 and 3 days in culture, there was a non-significant change in sphericity (0.65 ± 0.02 vs. 0.66 ± 0.04). However, between 3 and 5 days in culture, there was a significant decrease in sphericity, to 0.50 ± 0.12. As cells continued to spread, there was slight drop in sphericity, with 7-day average measured as 0.47 ± 0.10 (Figure 1D). Sphericity did not significantly change, with average values of 0.65 and 0.66 after 1 and 3 days in culture. As expected, nuclear YAP also increased with time, with a significant change observed between 1 and 5 days and 1 and 7 days in culture (Figure 1E). Taken together, these findings show that in unconfined GelMe hydrogels it takes MSCs at least 3 days to significantly remodel their surrounding gelatin material, leading to significant changes in cell morphology and YAP signaling. Further, the range of MSC metrics measured here (Volume, Sphericity, Nuclear YAP) increased with increasing culture time.

Cells in compressed hydrogels can sense local changes in hydrogel stress induced by an applied force (22). For example, chondrocytes encapsulated in hydrogels under dynamic compression produced more cartilage extracellular matrix in comparison to chondrocytes in unconfined hydrogels (23). To determine the extent in which uniaxial compression causes regional changes in intra-hydrogel stress, a finite element analysis model of an elastic hydrogel under compression was developed using FEBio software (methods and model parameters in Supplementary Material). At 0% compression, the cylindrical hydrogel render did not show any stress, as expected. At 25% compression, a concentric gradient of stress was observed, with higher levels of stress in the edge region (~1 kPa) vs. the center region (~0.5 kPa). At 50% compression, the same hydrogel stress trend was observed, however, the values were significantly higher. The center (low stress) of the 50% compressed hydrogel had stress of 1 kPa, which was the same value of the edge (high stress) of the 25% compressed hydrogel model (Figure 2A). Based on these findings, it was hypothesized that encapsulated MSCs would be larger, less spherical, and feature higher YAP signaling in edge regions of compressed hydrogels, due to increased hydrogel stress from an applied compressive force.

To test this hypothesis, a custom uniaxial compression device was developed (Supplementary Figure 2). MSC-laden GelMe hydrogels (1E6 cells/ml) were formed in cylindrical molds affixed onto thiolated glass coverslips, and cylindrical molds were carefully removed prior to transferring samples to wells of a 24-well plate. Culture wells were filled with Growth Medium and placed in the compression device. A nylon bolt was then lowered from above until it compressed the hydrogel 0, 25, or 50% of its free-swelling height for 3 days. After 3 days, samples were fixed, stained, and MSC morphology and YAP signaling was evaluated in edge or center regions of the hydrogel samples (Figure 2B). Representative silhouettes of single MSCs show that at 0% compression, MSCs retain a spherical morphology after 3 days in culture. Although a spherical morphology is also observed for the center (low stress) region of 25% compressed hydrogels, MSCs spread in edge (high stress) regions of these hydrogels. MSCs in 50% compressed hydrogels had the same highly elongated and spread morphology in center and edge regions (Figure 2C). This could be due to the larger applied compressive force, since the finite element analysis model of compression predicted that at 50% compression the lowest region of stress (center) would equal the highest region of stress (edge) for 25% compressed hydrogels.

An edge vs. center region comparison of MSC volume shows that MSCs in the edge region of 25% compressed hydrogels were significantly larger than those in the lower stress center region. MSC volume in unconfined (0%) and highly (50%) compressed hydrogels was the same in edge and center regions (Figure 2D). Sphericity was the largest for MSCs in uncompressed GelMe hydrogels, and lowest in 50% compressed hydrogels. Concomitant with volume, MSCs were significantly less round in the edge region when compared to the center region of 25% compressed hydrogels (Figure 2E). Nuclear YAP of MSCs in uncompressed GelMe hydrogels was also the lowest after 3 days in culture (1.04 ± 0.19), whereas regional differences were only observed in 25% compressed hydrogels. MSCs in 50% compressed hydrogels had the highest levels of nuclear YAP (2.30 ± 1.1 in edge, 2.10 ± 0.8 in center) (Figure 2F). Nuclear YAP in highly compressed hydrogels is over two-fold higher than in uncompressed hydrogels, which suggests that confining hydrogels into tightly packed regions will induce changes in mechanosensitive states comparable to those present in tissue interfaces.

Based on the findings presented and previous studies on cell-hydrogel mechanosensing (17, 22, 24), it is hypothesized that MSC-laden GelMe hydrogels inside rigid yet porous PCL scaffolds will feature a range of hydrogel stresses which will translate to diverse mechanosensitive states. To test this hypothesis, MSCs in GelMe solution were perfused into porous PCL scaffolds and cultured for 3 days (Figure 3A). A representative confocal image of GelMe-encapsulated MSCs (green) in a red-labeled porous PCL scaffold display spherical and spread MSCs based on their proximity to rigid PCL pores (Figure 3A, inset). Using a salt-leaching technique, PCL scaffolds with Low (L, 280 ± 100 μm), Med (M, 325 ± 210 μm), and High (H, 430 ± 145 μm) pore diameters were developed (Figure 3B). The composites had a significantly higher compressive modulus than GelMe alone. The compressive modulus of GelMe hydrogels is 7.15 ± 0.8 kPa, whereas composites made with Low pore diameter are 570 ± 60 kPa. Increasing porosity resulted in a decrease in compressive modulus, with values for Med and High pore diameter composites as 300 ± 40 kPa and 100 ± 20 kPa, respectively (Supplementary Figure 5). MSCs were highly viable in the composites, with at least 80% viability across all constructs after at least 7 days in culture (Supplementary Figure 6).

MSC mechanosensing and population heterogeneity can be controlled depending on pore diameter. On Low pore diameter composites, MSCs are in closest proximity to PCL, and a large fraction of cells elongate due to being within small pores that are always close to the rigid PCL (Figure 3C, red box). On Med pore diameter composites, the pores are slightly larger and although most cells are spread, cell morphology is more diverse (Figure 3C, green box). In contrast, on High pore diameter composites, MSCs are either in proximity (near PCL/hydrogel interface) or further away (e.g., suspended in GelMe at the center of the larger pores), resulting in round and spread morphologies (Figure 3C, blue box). Quantification of cell volume shows that cell volume decreases with increasing pore diameter (Figure 3D). Similarly, Nuclear YAP decreases with increasing pore size (Figure 3E), and Sphericity increases with increasing PCL porosity (Figure 3F).