Extracellular matrix viscoelasticity regulates TGFβ1‐induced epithelial‐mesenchymal transition and apoptosis via integrin linked kinase

Transforming growth factor (TGF)‐β1 is a multifunctional cytokine that plays important roles in health and disease. Previous studies have revealed that TGFβ1 activation, signaling, and downstream cell responses including epithelial‐mesenchymal transition (EMT) and apoptosis are regulated by the elasticity or stiffness of the extracellular matrix. However, tissues within the body are not purely elastic, rather they are viscoelastic. How matrix viscoelasticity impacts cell fate decisions downstream of TGFβ1 remains unknown. Here, we synthesized polyacrylamide hydrogels that mimic the viscoelastic properties of breast tumor tissue. We found that increasing matrix viscous dissipation reduces TGFβ1‐induced cell spreading, F‐actin stress fiber formation, and EMT‐associated gene expression changes, and promotes TGFβ1‐induced apoptosis in mammary epithelial cells. Furthermore, TGFβ1‐induced expression of integrin linked kinase (ILK) and colocalization of ILK with vinculin at cell adhesions is attenuated in mammary epithelial cells cultured on viscoelastic substrata in comparison to cells cultured on nearly elastic substrata. Overexpression of ILK promotes TGFβ1‐induced EMT and reduces apoptosis in cells cultured on viscoelastic substrata, suggesting that ILK plays an important role in regulating cell fate downstream of TGFβ1 in response to matrix viscoelasticity.

Factors that regulate the shift of TGFβ1 from a tumor suppressor to a tumor promoter role are not well understood.
Integrin linked kinase (ILK) plays an important role in cell mechanosensing and cell response to TGFβ1.ILK is an intracellular protein that localizes to focal adhesions and interacts with the actin cytoskeleton to facilitate communication between the cell and extracellular matrix (Li et al., 1999).Matrix stiffness, TGFβ signaling, and ILK have been shown to cooperatively influence cell behavior.

TGFβ1 upregulates ILK expression and ILK inhibition attenuates
TGFβ1-induced EMT (Serrano et al., 2013).Mammary carcinoma cells exhibit increased ILK expression as matrix stiffness increases (Pang et al., 2016).Furthermore, ILK mediates a matrix stiffnessinduced switch in TGFβ1 function; ILK knockdown results in TGFβ1induced apoptosis in cells cultured on both soft and stiff substrata (Kilinc et al., 2021).However, how matrix viscoelasticity regulates ILK signaling in the context of TGFβ1-induced EMT is not known.
Here, we synthesized a series of viscoelastic polyacrylamide hydrogels to examine how viscous dissipation impacts mammary epithelial cell response to TGFβ1.We held the storage modulus constant at 2000 Pa to mimic the elastic properties of breast tumors and varied the loss modulus to have substrata with varying levels of viscous dissipation.We find that culture of mammary epithelial cells on viscoelastic substrata reduces TGFβ1-induced cell spreading, elongation, stress fiber formation, and gene expression changes associated with EMT in comparison to what is observed in cells cultured on nearly elastic substrata with the same storage modulus.
In addition, when cells are cultured on viscoelastic substrata, TGFβ1 induces higher levels of apoptosis in comparison to when cells are cultured on nearly elastic substrata.Furthermore, TGFβ1 induces increased expression of ILK in cells cultured on nearly elastic substrata but not in cells cultured on viscoelastic substrata.Overexpression of ILK promotes TGFβ1-induced EMT and reduces TGFβ1-induced apoptosis in cells cultured on viscoelastic substrata, suggesting that ILK plays an important role in mediating cell response to TGFβ1 and substratum viscoelasticity.These findings provide evidence that the viscous properties of the cellular microenvironment regulate cell fate downstream of TGFβ1 and that ILK plays a critical role in sensing the mechanical properties of the matrix.

| Viscoelastic polyacrylamide hydrogel synthesis
Polyacrylamide hydrogels with varying loss moduli were synthesized using an adapted protocol (Charrier et al., 2020).Briefly, the hydrogels were composed of an acrylamide monomer (Sigma Aldrich) and bis-acrylamide crosslinker (Sigma Aldrich) and the loss moduli were varied by trapping strands of linear acrylamide within the crosslinked polyacrylamide network.To make linear polyacrylamide, a solution of 5% acrylamide in water was mixed, degassed, and polymerized by adding 0.05% N,N,Nʹ,Nʹ-tetramethylethylenediamine (TEMED) and 0.025% ammonium persulfate (APS).The solution was polymerized for 1 h at 37°C then stored at 4°C in the dark.The concentrations of TEMED and APS were low to ensure that long linear chains of polyacrylamide were formed.
To form the activated crosslinked network, water, acrylamide, bis-acrylamide, and 4% N-hydroxysulfosuccinimide (NHS) dissolved in dimethyl sulfoxide (DMSO) were mixed according to the compositions described in Supporting Information Table S1 and the solution was degassed.A 1:10 v/v mixture of 4% NHS dissolved in DMSO and gel solution were mixed, and the amount of water in the solution was adjusted to compensate for the added NHS.To polymerize the network, 0.075% APS and 0.25% TEMED were added to the solution and thoroughly mixed.The solution was deposited onto functionalized glass slides treated sequentially with 0.1 N NaOH (Sigma Aldrich), 2% v/v aminopropyltrimethoxysilane (APTMS; Sigma Aldrich) diluted in acetone, and 0.5% glutaraldehyde (Sigma Aldrich) diluted in 1× phosphate buffered saline (PBS).The solution was then covered with a Rain-X treated coverslip to obtain a flat hydrogel surface once polymerized.Hydrogels were polymerized for 15 min at room temperature.Then, the top coverslip was removed, and the hydrogels were covered in 0.5 M HEPES (pH 8.5) and sterilized in a CL-1000 Ultraviolet Crosslinker for 15 min.Immediately after sterilization, the hydrogels were covered with 0.1 µg/mL human fibronectin (BD Biosciences) and incubated overnight at 4°C.Hydrogels were made fresh for each experiment and not stored.

| Mechanical characterization of hydrogels
Rheology was performed using an RFS-3 rheometer (Rheometric Scientific Inc., Model number ARES-3).Immediately after polymerization initiation, the hydrogel solution was deposited into the cup of the rheometer.Time sweeps were performed at a frequency of 1 rad/s and 2% strain for about 30 min or until polymerization was complete, as indicated by a plateau in Gʹ and G″ values.Frequency sweeps were performed from 0.3 to 100 rad/s with 2% strain.The Gʹ and G″ values reported in Supporting Information Table S1 were measured at the frequency 1 rad/s, as it has been reported that cell mechanosensing occurs at this frequency (Charrier et al., 2018).

| Cell culture
Normal murine mammary gland (NMuMG) epithelial cells (American Type Culture Collection ATCC CRL-1636) were authenticated by Genetica Cell Line Testing, a LabCorp brand, and cells tested negative for mycoplasma.NMuMG cells were grown in Dulbecco's Modified Eagle Medium (DMEM) with 10% (v/v) fetal bovine serum (FBS, Atlanta Biologicals), 0.01% gentamicin (Gibco), and 10 μg/mL insulin (Sigma Aldrich).Cells were cultured in medium in a 37°C incubator with 5% CO 2 .Before seeding cells onto the polyacrylamide hydrogels, media was added to the hydrogels so that hydrogels were hydrated.Following incubation of the hydrogels at 37°C for 30 min, cells were then seeded onto the hydrogel surfaces at a density of 75,000 cells/cm 2 .Cells were allowed to adhere to the hydrogel surfaces for 24 h after seeding, followed by treatment with 10 ng/mL of recombinant human TGFβ1 (R&D Systems) or a vehicle control (1 mg/mL bovine serum albumin in 4 mM HCl) for 48 h.For cleaved caspase-3 staining and caspase-3 activity experiments, TGFβ1 treatment was for 24 h.

| Plasmids and transfections
A plasmid encoding mouse EGFP-tagged integrin linked kinase (mILK-EGFP-N1) (Martin et al., 2022) was obtained from Sara Wickström (Addgene plasmid #176896).The plasmid encoding the empty vector pEGFP-N1 was obtained from Gerard Elberg (Department of Pediatrics, The University of Oklahoma Health Sciences Center) (Elberg et al., 2008).Cells were transfected with the plasmids using Lipofectamine with Plus Reagent (Thermo) according to the manufacturer's protocol.Cells were seeded onto hydrogels 24 h after transfection.

| Immunofluorescence staining
For α smooth muscle actin (αSMA) staining, cells were fixed with ice cold 1:1 methanol/acetone for 10 min at −20°C.For staining of E-cadherin and cleaved caspase-3, cells were fixed with 4% paraformaldehyde at room temperature for 15 min.Samples were permeabilized with 0.5% IGEPAL and 0.1% Triton X-100 diluted in 1× PBS for 10 min each at room temperature.The samples were incubated with a blocking buffer composed of 10% goat serum (Sigma Aldrich) diluted in 1× PBS for 1.5 h at room temperature.
For visualization of F-actin, cells were incubated with Alexa-Fluor 594 phalloidin (Thermo) according to the manufacturer's protocol.
The cell nuclei were visualized by treatment with Hoechst 33342 (1:10,000; Life Technologies) at room temperature for 15 min.Samples were rinsed thoroughly with 1× PBS, and then mounted onto microscope slides with Fluoromount-G (Invitrogen) to prepare them for imaging.

| Microscopy and image analysis
Samples mounted on microscope slides were imaged using a 20× or 40× objective on a Nikon Eclipse Ti-E inverted fluorescence microscope equipped with a Photometrics CoolSNAP HQ 2 CCD camera.ImageJ software was used to measure cell spread area and aspect ratio (major/minor axis) of the cells.Vinculin and ILK images were overlayed in ImageJ to visualize colocalization.Vinculin and ILK organization was then categorized as punctate, mixed, or diffuse and the percentage of cells showing these staining patterns is reported as a function of substratum viscoelasticity and treatment with TGFβ1.

| Caspase-3 activity
Caspase-3 activity was determined using the EnzChek Caspase-3 Assay Kit #1 (Thermo).Briefly, cells were harvested from the hydrogels using trypsin then washed with 1× PBS.A hemocytometer was used to count the cells to ensure equal amounts of cells from each sample were used in the assay.The assay was carried out according to the manufacturer's instructions.The fluorescence was measured using a TECAN Infinite 200 Pro microplate reader with excitation/emission wavelengths of 342/441 nm.

| Western blotting
Whole cell lysates were collected using RIPA buffer (Thermo) with protease and phosphatase inhibitors.Protein concentration was quantified using a Pierce BCA Protein Assay Kit (Thermo).Equal amounts of protein were separated on NuPAGE 4%-12% Bis-Tris (Thermo) or NuPAGE 3%-8% Tris-Acetate (Thermo) gels and transferred to polyvinyl difluoride (PVDF) or nitrocellulose membranes using an XCell SureLock Mini-Cell (Invitrogen).Membranes were blocked with 5% nonfat milk for an hour then incubated with the following primary antibodies overnight at 4°C: E-cadherin

| Statistical analysis
At least three trials were performed for each experiment.Data are presented as mean ± standard error of the mean (s.e.m.).An analysis of variance followed by a Tukey-Kramer post hoc test using MATLAB was used to determine statistical differences.Differences were considered significant for p < 0.05.

| Synthesis of viscoelastic hydrogels to mimic breast tumor mechanical properties
Soft tissues such as the mammary gland are viscoelastic and mechanical properties of these tissues can be described by a storage modulus (Gʹ), a parameter indicative of material stiffness, and a loss modulus (G″), a parameter that gives information about the viscosity of the tissue (Chaudhuri et al., 2020).Healthy tissue has a loss modulus that is approximately 10% of its storage modulus (Chaudhuri et al., 2020) and it has been reported that breast tumors are more viscous than surrounding healthy breast tissue (Siegmann et al., 2010;Sinkus et al., 2007).We designed a series of polyacrylamide hydrogels with controlled elasticity and viscous dissipation to mimic breast tumor viscoelastic properties.Crosslinker and linear acrylamide concentrations were varied following a modified protocol (Charrier et al., 2020), to create hydrogels with defined Gʹ and G″ (Figure 1 and Supporting Information Table S1).Rheometry was used to characterize the mechanical properties of the hydrogels.During polymerization, Gʹ and G″ increase over time and then begin to plateau as the hydrogels polymerize (Figure 1a).The frequency dependence of Gʹ and G″ was measured over a range of 0.3-100 rad/s, and in this range Gʹ is nearly constant (Figure 1b).
Previous studies suggest that cell mechanosensing occurs at a frequency of 1 rad/s (Charrier et al., 2018), therefore, we report values of Gʹ and G" measured at 1 rad/s (0.16 Hz).For this study, we utilize hydrogels with G' of approximately 2000 Pa (Figure 1c and the storage modulus using E G′ = 2 (1 + ϑ) where ϑ is the Poisson ratio.Assuming a Poisson ratio of 0.5, the Young's modulus for the 0% linear acrylamide composition, which exhibits nearly elastic behavior, is 5700 Pa, which is comparable to the stiffness of malignant human breast tumors (Plodinec et al., 2012).Fibronectin was coupled to the hydrogels using NHS to enable cell adhesion to the hydrogels.Immunofluorescence microscopy revealed that fibronectin couples uniformly to hydrogel surfaces and the relative level of fibronectin attached and the number of cells attached to different hydrogel compositions is not statistically different (Supporting Information Figure S1a-c).This hydrogel system enables studying how the loss modulus impacts mammary cell response to TGFβ1.In addition to inducing cell shape changes, TGFβ1 promotes rearrangements of the cytoskeleton in epithelial cells, including stress fiber formation (Bhowmick et al., 2001;Haynes et al., 2011;Nalluri et al., 2015).Previous studies have shown that matrix viscoelasticity can impact cytoskeletal organization and stress fiber formation (Charrier et al., 2018;Chaudhuri et al., 2015;Hui et al., 2021;Marozas et al., 2019).Thus, we hypothesized that matrix viscoelasticity may regulate cytoskeletal structure in response to TGFβ1.

| Increasing substratum viscous dissipation attenuates TGFβ1-induced changes in EMTassociated gene expression
TGFβ1-induced EMT is characterized by a downregulation in epithelial genes, such as E-cadherin, and an upregulation in mesenchymal genes, such as α smooth muscle actin (αSMA) (Kalluri & Weinberg, 2010).We examined the effect of substratum viscous dissipation on gene expression by monitoring E-cadherin and αSMA levels using immunofluorescence staining and western blotting.E-cadherin localizes to cell-cell junctions in control and TGFβ1treated cells cultured on viscoelastic substrata and in control-treated cells cultured on nearly elastic substrata (Figure 3a).TGFβ1 treatment reduces E-cadherin expression and significantly increases the percentage of cells expressing αSMA on nearly elastic substrata (Figure 3a,b).In contrast, cells cultured on viscoelastic substrata do not show increased expression of αSMA in response to TGFβ1.
Consistent with immunofluorescence staining, western blotting revealed that TGFβ1 treatment significantly reduces E-cadherin protein levels and significantly increases αSMA protein levels in cells cultured on nearly elastic substrata (Figure 3c-e).Conversely, cells cultured on viscoelastic substrata do not show gains in αSMA expression in response to TGFβ1 (Figure 3e).E-cadherin and αSMA protein levels do not change significantly in control-treated cells cultured on substrata as a function of varying loss moduli (Figure 3d,e).Furthermore, expression of E-cadherin is not downregulated in cells cultured on hydrogels with a loss modulus of 410 Pa following treatment with TGFβ1 (Figure 3d).These findings suggest that increasing viscous dissipation blocks TGFβ1-induced changes in the expression of some epithelial and mesenchymal markers.

| Increasing substratum viscous dissipation promotes TGFβ1-induced apoptosis
Previous studies have shown that TGFβ1 induces EMT in mammary epithelial cells cultured on stiff elastic substrata but induces apoptosis in cells cultured on soft elastic substrata (Leight et al., 2012).Given that ILK is an important component of focal adhesions and facilitates cell-extracellular matrix signaling (Li et al., 1999), we sought to determine whether matrix viscoelasticity and TGFβ1 mediate localization of ILK to cell-matrix adhesions.Immunofluorescence staining for vinculin and ILK was used to monitor focal adhesion organization as a function of matrix viscous dissipation (Figure 5c).In cells cultured on nearly elastic substrata, vinculin and ILK localize to puncta in response to TGFβ1 treatment (Figure 5c,d).In contrast, cells treated with control vehicle show diffuse staining or a mixture of punctate and diffuse staining for vinculin and ILK.In cells cultured on substrata with increasing viscous dissipation, vinculin and ILK staining is mainly diffuse with a small percentage of cells showing mixed punctate and diffuse organization, regardless of treatment condition.
For substrata with a loss modulus of 410 Pa, there is a slightly, though not significant, higher percentage of cells with diffuse vinculin and ILK staining compared to cells cultured on substrata with a loss modulus of 190 Pa, regardless of treatment condition (Figure 5c,d).
Importantly, vinculin and ILK are colocalized, indicated by the yellow color in the merged images in Figure 5c.These results suggest that cell-matrix adhesive structures are regulated by TGFβ1 and matrix viscoelasticity.
To determine whether ILK expression is sufficient to promote EMT and block apoptosis in cells cultured on viscoelastic substrata in Matrix elasticity and ILK regulate a switch between TGFβ1induced EMT and apoptosis (Kilinc et al., 2021).We determined that ILK overexpression rescues the ability of TGFβ1 to induce EMT on viscoelastic substrata, thus we hypothesized that ILK overexpression would also reduce TGFβ1-induced apoptosis on viscoelastic substrata.We find that caspase-3 activity remains low in cells cultured on nearly elastic substrata across all treatment conditions (Figure 6g).

| DISCUSSION
In this study, we found that substratum viscous dissipation regulates cell response downstream of TGFβ1 via ILK.Using an established protocol (Charrier et al., 2020), we synthesized polyacrylamide hydrogels with storage and loss moduli similar to diseased breast tissue and utilized the hydrogels to examine the impact of viscous dissipation on cell response to TGFβ1.In NMuMG epithelial cells cultured on nearly elastic substrata, TGFβ1 treatment induces cell spreading and elongation, F-actin stress fiber formation, and a decrease in E-cadherin and a gain in αSMA expression, suggesting that these cells undergo EMT.In contrast, cells cultured on viscoelastic substrata and treated with TGFβ1 maintain a rounded morphology, high E-cadherin levels, and low αSMA levels.Furthermore, TGFβ1 treatment increases cleaved caspase-3 levels and caspase-3 activity in cells cultured on viscoelastic substrata, suggesting that increasing substratum viscous dissipation blocks TGFβ1-induced EMT and promotes TGFβ1-induced apoptosis.
TGFβ1 treatment also increases ILK levels in cells cultured on nearly  Several recent studies have examined the impact of matrix viscous dissipation on cell morphology and actin polymerization.U2OS human bone osteosarcoma epithelial cells, LX-2 hepatic stellate cells, human lung fibroblasts, and NIH3T3 fibroblasts exhibit a decrease in cell spreading on viscoelastic hydrogels in comparison to on nearly elastic hydrogels when the storage moduli or Young's moduli are high (Chaudhuri et al., 2015;Hui et al., 2019Hui et al., , 2021;;Marozas et al., 2019).Computational modeling results suggest that substratum viscosity suppresses cell spreading when cells are cultured on substrata with high stiffness, as observed in the abovementioned experiments (Gong et al., 2018).Our results for control vehicle treated cells are consistent with these previous studies as we find that mammary epithelial cells spread less when cultured on viscoelastic hydrogels than on nearly elastic hydrogels when the storage modulus is fixed to 2000 Pa.In addition, for hepatic stellate cells and fibroblasts, more actin stress fibers have been observed in cells cultured on nearly elastic substrata compared to viscoelastic substrata when the Young's modulus and storage modulus are high (E = 9 kPa and Gʹ = 5 kPa, respectively) (Chaudhuri et al., 2015;Hui et al., 2021).Here, we find that mammary epithelial cells cultured on hydrogels with a storage modulus of 2000 Pa have decreased F-actin stress fiber formation as viscous dissipation increases.Moreover, we find that mammary epithelial cells exhibit changes in cell area, aspect ratio, and cytoskeletal organization in response to TGFβ1 when cultured on nearly elastic hydrogels but not when cultured on viscoelastic hydrogels.These experiments provide new insights into how viscous dissipation impacts cell morphology and cytoskeletal organization in response to growth factors.While many of the aforementioned studies found that increasing substratum viscoelasticity decreases cell spreading and elongation at higher stiffness, the opposite trend has been observed in some studies when cells are cultured on hydrogels with lower stiffness.For osteosarcoma and fibroblasts, cell spreading, actin stress fiber formation, and focal adhesion formation increase on viscoelastic substrata compared to nearly elastic substrata when the Young's modulus is low (1.4 kPa) (Chaudhuri et al., 2015;Gong et al., 2018).
Spreading was higher in LX-2 hepatic stellate cells cultured on nearly elastic hydrogels compared to viscoelastic hydrogels of the same "stiff" storage modulus of 5000 Pa, but cell spreading and elongation did not significantly change in cells cultured on viscoelastic hydrogels in comparison to nearly elastic hydrogels of the same "soft" storage modulus of 500 Pa (Hui et al., 2019).Furthermore, it has been suggested that fibroblast migration and correlated movement of epithelial sheets depend on the ratio of Gʹ to G″ (Chester et al., 2018;Murrell et al., 2011).However, another study found no correlation between the ratio of Gʹ to G″ and cell spreading (Chaudhuri et al., 2015).Future work should aim to understand how viscous dissipation impacts cell behaviors, including EMT and apoptosis, over a range of elastic moduli.These efforts would help to elucidate whether the balance between elastic and viscous properties of the matrix is an important factor in regulating epithelial cell response to growth factors.
EMT is characterized by the downregulation of epithelial markers such as E-cadherin and the upregulation of mesenchymal markers including αSMA, vimentin, and N-cadherin.Recent studies have examined αSMA stress fiber organization in response to substratum viscoelasticity, though not in response to TGFβ1.The organization of αSMA stress fibers was reduced in stellate cells when cultured on viscoelastic polyacrylamide or hyaluronic acid hydrogels compared to when cultured on nearly elastic hydrogels of the same storage modulus (Gʹ = 5000 Pa) (Charrier et al., 2018;Hui et al., 2019).
Madin-Darby canine kidney (MDCK) epithelial cells cultured on P(CL-co-DLLA) hydrogels with an elastic modulus of 5 MPa and varying relaxation times had significantly higher mRNA levels of the mesenchymal marker N-cadherin as relaxation time increased and had no significant change in the levels of the epithelial marker E-cadherin (Chang et al., 2022).In spheroid models, cells grown in viscoelastic matrices have higher expression of mesenchymal genes compared to when grown in nearly elastic matrices.In one study, PANC-1 human pancreatic cancer cells encapsulated in viscoelastic phenylboronic acid containing polymers had an increase (all less than 2-fold) in N-cadherin, vimentin, Snail, and Zeb1 mRNA levels compared to cells cultured within nearly elastic hydrogels of the same storage modulus (Gʹ = 6000 Pa) (Lin et al., 2023).Interestingly, there was an almost four-fold increase in E-cadherin mRNA levels in cells cultured in viscoelastic compared to nearly elastic hydrogels (Lin et al., 2023).In another recent study, MCF10A human mammary epithelial cell spheroids were able to branch when grown in 3D viscoelastic alginate hydrogels and cells at the ends of branches had  Microenvironmental cues, such as matrix elasticity, have also been shown to regulate apoptosis.For example, A549 lung cancer cells, MDA-MB-231 breast cancer cells, and NIH3T3 fibroblasts undergo apoptosis more readily on low elastic modulus hydrogels than when cultured on hydrogels with a higher elastic modulus (Tilghman et al., 2010;Wang et al., 2000).Previous studies have also demonstrated that matrix elasticity can regulate apoptosis in response to TGFβ1.Upon TGFβ1 stimulation, NMuMG and MDCK epithelial cells undergo apoptosis when cultured on polyacrylamide hydrogels with a Young's modulus of 400 Pa but undergo EMT when cultured on stiffer polyacrylamide hydrogels with a Young's modulus greater than 8000 Pa (Leight et al., 2012).Regulation of apoptosis by matrix viscoelasticity is less investigated, but some recent studies demonstrate that matrix stress relaxation can impact apoptosis in stem cells.Mesenchymal stem cells cultured in 3D viscoelastic collagen matrices with an initial elastic modulus of 3000 Pa and varying stress relaxation times undergo increased apoptosis when cultured in slow-relaxing matrices compared to cells cultured in fastrelaxing matrices (Huang et al., 2023).In another study, human induced pluripotent stem cells (hiPSCs) were encapsulated in 3D viscoelastic alginate matrices with varying stress relaxation times (Indana et al., 2021).Cells cultured in slow-relaxing hydrogels with low ligand density had higher levels of apoptosis than cells cultured in fast-relaxing hydrogels.Surprisingly, in this system, there was no difference in apoptosis levels in cells cultured in viscoelastic hydrogels with an initial elastic modulus of 3000 Pa compared to 20,000 Pa with similar relaxation times (Indana et al., 2021).In our studies, we find that when the loss modulus of the matrix is increased by embedding linear polyacrylamide within the hydrogel, cells cultured on the hydrogels exhibit higher levels of apoptosis in response to TGFβ1.Future work should aim to further understand how matrix viscous dissipation influences apoptosis in response to growth factors.
To elucidate mechanistically how matrix viscoelasticity mediates cell response to TGFβ1, we investigated the impact of viscous dissipation on ILK expression.ILK is a critical regulator of focal adhesion dynamics and stability and it provides a link between the actin cytoskeleton and integrin-mediated adhesion sites (Zervas et al., 2001).Previous studies have demonstrated that deficiency of ILK can lead to disrupted focal adhesion maturation and actin remodeling (Stanchi et al., 2009) and point mutations to the ATP binding domain of ILK reduce cell traction and migration (Martin et al., 2022).Furthermore, downregulation of ILK has been shown to reduce spreading in mammary epithelial cells cultured on stiff elastic substrata (Han et al., 2018) and ILK knockdown promotes TGFβ1induced apoptosis instead of EMT in cells cultured on stiff elastic substrata (Kilinc et al., 2021).In our study, we found that mammary epithelial cells increase the expression of ILK in response to TGFβ1 treatment when cultured on nearly elastic hydrogels but not when cultured on viscoelastic hydrogels with the same storage modulus.
We found that in cells cultured on nearly elastic substrata, TGFβ1

F
I G U R E 1 Mechanical characterization of viscoelastic polyacrylamide hydrogels.(a) Time sweep showing the storage (Gʹ) and loss (G″) moduli as a hydrogel composed of 5% acrylamide, 0.1% bis-acrylamide, and 2.0% linear acrylamide polymerizes.(b) Gʹ and G″ as a function of frequency.Frequency sweep performed from 0.3 to 100 rad/s for the 2.0% linear acrylamide hydrogel.(c) Gʹ and (d) G″ as a function of w/v % linear acrylamide.Mechanical properties were measured at a frequency of 1 rad/s and 2% strain.Data represent the mean ± s.e.m. for n ≥ 4.

3. 2 |
Substratum viscous dissipation regulates TGFβ1-induced changes in cell morphology and cytoskeletal organization During TGFβ1-induced EMT, cells undergo significant morphological changes including increased spreading and elongation (Nelson et al., 2008; O'Connor & Gomez, 2013).Furthermore, studies have shown that blocking the ability of cells to spread attenuates TGFβ1-induced expression of some mesenchymal protein markers (O'Connor & Gomez, 2013).To examine the impact of matrix viscous dissipation on cell morphology in response to TGFβ1, normal murine mammary gland (NMuMG) epithelial cells were cultured on fibronectin-coated viscoelastic hydrogels with fixed Gʹ (2000 Pa) and varying G″ (16, 190, and 410 Pa) and treated with TGFβ1 or a control vehicle.Phase contrast microscopy images reveal that TGFβ1 treatment significantly increases cell spread area and aspect ratio compared to control vehicle treatment in cells cultured on nearly elastic substrata (G″ = 16 Pa) (Figure 2a-d).Cells cultured on the viscoelastic substrata (G″ = 190 or 410 Pa) have reduced cell spreading compared to cells cultured on the nearly elastic substrata following treatment with TGFβ1.These data suggest that matrix viscous dissipation impacts the ability of cells to undergo morphological changes in response to TGFβ1.
Cytoskeletal remodeling was monitored using fluorescently tagged F I G U R E 2 Substratum viscoelasticity regulates cell morphology and cytoskeletal organization in response to TGFβ1.(a) Phase contrast microscopy images of NMuMG cells cultured on hydrogels with Gʹ = 2000 Pa and varying G″ and treated with TGFβ1 or a vehicle control.Scale bars: 25 μm.(b) Filamentous (F)-actin visualized with fluorescently tagged phalloidin.Scale bars: 25 μm.Quantification of cell (c) spread area and (d) aspect ratio from phase contrast images.Data are displayed as individual points with error bars representing mean ± s.e.m. with at least 150 cells analyzed over three different experiments; *p < 0.05, **p < 0.01, ***p < 0.001.phalloidin to visualize filamentous (F)-actin (Figure 2b, Supporting Information Figure S2).In cells cultured on the nearly elastic substrata, TGFβ1 treatment promotes F-actin stress fiber formation, while cells treated with the control vehicle exhibit a cortical actin structure.Cells cultured on the viscoelastic substrata, regardless of treatment condition, have reduced formation of stress fibers compared to cells treated with TGFβ1 that are cultured on nearly elastic substrata.These results suggest that substratum viscous dissipation impacts cytoskeletal reorganization in response to TGFβ1.
To examine the impact of matrix viscoelasticity on TGFβ1-induced apoptosis, we monitored apoptosis by staining for cleaved caspase-3 and measuring caspase-3 activity.Immunofluorescence images reveal that in control vehicle treated cells, the percentage of cells staining positive for cleaved caspase-3 increases with increasing substratum viscous dissipation, though not significantly (Figure 4a,b).TGFβ1 treatment slightly increases the percentage of cells expressing cleaved caspase-3 when cells are cultured on nearly elastic substrata.In contrast, TGFβ1 treatment significantly increases the percentage of cells staining positive for cleaved caspase-3 when cells are cultured on viscoelastic substrata compared to control vehicle treated cells cultured on all substrata and compared to TGFβ1 treated cells cultured on nearly elastic substrata (Figure 4a,b).Consistent with the cleaved caspase-3 staining results, TGFβ1 treatment significantly increases caspase-3 activity in cells cultured on viscoelastic substrata compared to cells cultured on nearly elastic substrata (Figure 4c).Caspase-3 activity does not vary in control vehicle treated cells as a function of loss moduli.Taken together, these observations demonstrate that TGFβ1 promotes apoptosis in cells cultured on substrata with a storage modulus of 2000 Pa when the substrata also exhibit viscous dissipation.

3. 5 |
ILK mediates cell response to TGFβ1 and substratum viscoelasticity ILK plays an important role in transducing mechanical signals from the cellular microenvironment into biophysical responses.Previous findings show that ILK knockdown results in TGFβ1-induced apoptosis in cells cultured on both soft and stiff elastic substrata, suggesting that ILK may regulate cell response to TGFβ1 and matrix elasticity (Kilinc et al., 2021).Thus, we hypothesized that ILK may mediate the differential EMT and apoptotic responses of cells to TGFβ1 and substratum viscoelasticity.Western blotting revealed that TGFβ1 treatment significantly increases ILK levels in cells cultured on nearly elastic substrata, while TGFβ1 treatment does not significantly increase ILK levels in cells cultured on substrata with increased viscous dissipation (Figure 5a,b).ILK protein levels do not change in control vehicle treated cells cultured on substrata with a storage modulus of 2000 Pa as a function of loss moduli.These findings suggest that the expression of ILK is regulated by TGFβ1 and matrix viscoelasticity.

F
I G U R E 3 Increasing substratum loss modulus attenuates EMT-associated gene expression changes in response to TGFβ1.(a) Immunofluorescence staining images for the epithelial marker E-cadherin (red) and mesenchymal marker αSMA (green) in NMuMG cells cultured on hydrogels with Gʹ = 2000 Pa and varying G″ and treated with TGFβ1 or a vehicle control for 48 h.Cell nuclei are shown in blue.Scale bars: 25 μm.(b) Percentage of cells expressing αSMA.Data represent the mean ± s.e.m. for n = 3 trials; ***p < 0.001.(c) Western blot for E-cadherin and αSMA in cells cultured on viscoelastic substrata and treated with TGFβ1 or a vehicle control.Densitometric analysis quantifying relative (d) E-cadherin and (e) αSMA levels from western blots.Data represent the mean ± s.e.m. for n = 5 trials; *p < 0.05, **p < 0.01, ***p < 0.001.response to TGFβ1, we transfected cells with an ILK-GFP construct or with a GFP plasmid as a control.Overexpression of ILK-GFP was confirmed by fluorescence microscopy and western blotting (Supporting Information Figure S3a-c).In cells cultured on viscoelastic substrata with a storage modulus of 2000 Pa and loss modulus of 410 Pa, ILK overexpression increases cell spreading compared to cells transfected with the GFP control plasmid, and TGFβ1 significantly amplifies this increase (Figure 6a,b).Cell aspect ratio also increases in ILK-overexpressing cells compared to cells transfected with the GFP control plasmid when cultured on the viscoelastic substrata with a storage modulus of 2000 Pa and a loss modulus of 410 Pa, and TGFβ1treatment augments this increase (Figure 6a,c).Given that ILK plays an important role in TGFβ1-induced EMT (Serrano et al., 2013) and overexpression of ILK restores TGFβ1-induced cell spreading and elongation on viscoelastic substrata, we posited that ILK overexpression may enable TGFβ1 to induce EMT when cells are cultured on viscoelastic substrata.Western blotting revealed that for cells cultured on nearly elastic substrata (G″ = 16 Pa), TGFβ1 treatment significantly reduces E-cadherin protein levels and increases αSMA protein levels compared to control vehicle treatment in both GFP and ILK-GFP transfected cells (Figure 6d-f).For cells cultured on viscoelastic substrata (G″ = 410 Pa), TGFβ1 treatment decreases E-cadherin protein levels and significantly increases αSMA protein levels in ILK-GFP transfected cells, but GFP transfected cells are refractive to TGFβ1-induced changes in EMT markers.Together, these results suggest that ILK overexpression promotes changes in some EMT associated proteins in response to TGFβ1 when cells are cultured on viscoelastic substrata.

For
cells cultured on viscoelastic substrata, TGFβ1 treatment increases caspase-3 activity in GFP-transfected cells compared to control vehicle-treated cells and compared to cells cultured on nearly elastic substrata across all treatment conditions.ILK overexpressing cells cultured on viscoelastic substrata do not exhibit an increase in caspase-3 activity levels with TGFβ1 treatment, but rather show similar caspase-3 activity to cells cultured on nearly elastic substrata for both TGFβ1 and control vehicle treatments.These findings provide support that ILK overexpression attenuates TGFβ1-induced apoptosis in cells cultured on viscoelastic substrata mimicking breast tumor mechanical properties.

F
I G U R E 4 Increasing substratum viscoelasticity promotes TGFβ1-induced apoptosis.(a) Immunofluorescence staining for cleaved caspase-3 (red) and nuclei (blue) in NMuMG cells cultured on hydrogels with Gʹ = 2000 Pa and varying G″ and treated with TGFβ1 or a vehicle control for 24 h.Scale bars: 25 μm.(b) Percentage of cells staining positive for cleaved caspase-3.Data represent the mean ± s. e.m. for n = 4 trials; *p < 0.05, **p < 0.01.(c) Cleaved caspase-3 activity in NMuMG cells cultured on hydrogels with Gʹ = 2000 Pa and varying G″ and treated with TGFβ1 or a vehicle control for 24 h.Data represent the mean ± s.e.m. for n = 3; *p < 0.05, **p < 0.01.F I G U R E 5 (See caption on next page).
does not affect ILK levels in cells cultured on viscoelastic substrata.Furthermore, TGFβ1 promotes punctate focal adhesion organization in cells cultured on nearly elastic substrata, where vinculin and ILK are colocalized.ILK overexpression reduces TGFβ1-induced apoptosis and promotes TGFβ1-induced EMT in cells cultured on viscoelastic substrata, indicating that ILK plays an important role in mediating cell response to substratum viscoelasticity downstream of TGFβ1.
increased vimentin expression compared to cells grown in nearly F I G U R E 5 Matrix viscoelasticity and TGFβ1 regulate integrin linked kinase (ILK) expression and focal adhesion organization.(a) Western blot for ILK in NMuMG cells cultured on hydrogels with Gʹ = 2000 Pa and varying G″ and treated with TGFβ1 or a vehicle control for 48 h.(b) Densitometric analysis of western blot for relative ILK levels.Data represent the mean ± s.e.m. for n = 3; *p < 0.05; **p < 0.01.(c) Immunofluorescence staining for vinculin (green) and ILK (red) in NMuMG cells cultured on hydrogels with Gʹ = 2000 Pa and varying G″ and treated with TGFβ1 or a vehicle control for 48 h.Cell nuclei are shown in blue.Merged images show colocalization of vinculin and ILK.Scale bars: 25 μm.(d) Percentage of cells showing vinculin and ILK organization characterized as punctate, mixed, or diffuse as indicated by the representative images.Data represent the mean ± s.e.m. for n = 3; **p < 0.001 compared to all other samples characterized as punctate.FI G U R E 6 (See caption on next page).
SACCO ET AL.| 11 of 15 elastic hydrogels of the same storage modulus (Gʹ = 5000 Pa)(Elosegui-Artola et al., 2022).In our study, we find that TGFβ1 promotes a decrease in the expression of the epithelial marker E-cadherin and an increase in the expression of the mesenchymal marker αSMA in cells cultured on nearly elastic hydrogels, but cells cultured on viscoelastic hydrogels are refractive to TGFβ1-induced changes in these markers.Differences in the regulation of epithelial and mesenchymal proteins in these studies may arise from differences in cell type, growth factor treatment conditions, the elastic modulus of the matrix, or matrix dimensionality.Here, we use an elastic modulus that mimics that of a breast tumor (Gʹ = 2000 Pa), while the previous studies highlighted utilize matrices with elastic moduli ranging from 5000 Pa to 5 MPa, all of which are more than two-fold greater in magnitude than that of normal mammary tissue and breast tumors.More research is needed to understand how matrix viscous dissipation regulates gene expression downstream of TGFβ1 and how cell type, matrix viscoelasticity, and matrix dimension play a role in this regulation.
promotes punctate focal adhesion organization with colocalization of vinculin and ILK at focal adhesion sites.In contrast, when viscous dissipation of the hydrogel is increased cells show a mixture of punctate and diffuse staining for vinculin and ILK.In addition, we found that overexpression of ILK increases cell spreading and promotes TGFβ1-induced EMT and abrogates TGFβ1-induced apoptosis in cells cultured on viscoelastic substrata with a storage modulus of 2000 Pa.These results suggest that by increasing ILK expression, focal adhesion formation and maturation may allow for cell spreading and EMT induction by TGFβ1.Future work should aim to understand how matrix viscoelasticity influences integrin engagement and focal adhesion properties downstream of TGFβ1 to regulate cell spreading and cell fate.

F
I G U R E 6 Integrin linked kinase (ILK) mediates cell response to TGFβ1 and viscoelasticity.(a) Phase contrast microscopy images of NMuMG cells cultured on hydrogels with Gʹ = 2000 Pa and varying G″, transfected with a plasmid to overexpress ILK-GFP or a control GFP plasmid, and treated with TGFβ1 or a vehicle control for 48 h.Scale bars: 25 μm.Quantification of cell (b) spread area and (c) aspect ratio from phase images shown in panel (a).Data are displayed as individual points with error bars representing mean ± s.e.m. with at least 100 cells analyzed over three different experiments; *p < 0.05, **p < 0.01, ***p < 0.001.(d) Western blot for E-cadherin and αSMA in NMuMG cells cultured on hydrogels with Gʹ = 2000 Pa and varying G″, transfected with a plasmid to overexpress ILK or a control GFP plasmid, and treated with TGFβ1 or a vehicle control for 48 h.Densitometric analysis of western blots shown in panel d for relative levels of (e) E-cadherin and (f) αSMA.Data represent the mean ± s.e.m. for n = 3; **p < 0.01; ***p < 0.001.(g) Cleaved caspase-3 activity in NMuMG cells cultured on hydrogels with Gʹ = 2000 Pa and varying G″, transfected with a plasmid to overexpress ILK or a control GFP plasmid, and treated with TGFβ1 or a vehicle control for 24 h.Data represent the mean ± s.e.m. for n = 3; *p < 0.05.