Geometric guidance of integrin mediated traction stress during stem cell differentiation

2.1. Materials

All materials were purchased from Sigma unless otherwise noted. Tissue culture plastic ware and glass coverslips (18-mm circular) were purchased from Fisher Scientific. Cell culture media and reagents were purchased from Gibco. Rabbit anti-Runx2 (ab23981) and anti-Osteopontin (ab8448) were purchased from Abcam. Mouse anti-MyoD (MAB3878) Mouse anti-α5β1 (MAB1969) and αVβ3 (MAB1976Z) were purchased from Millipore. Blebbistatin, Y-27632, FR180204 (ERK inhibitor), SP600125 (JNK inhibitor), and SB202190 (p38 inhibitor) were purchased from Calbiochem. Tetramethylrhodamine-conjugated anti-rabbit IgG antibody, Alexa Fluor 647-conjugated anti-mouse IgG antibody, Alexa Fluor 555-conjugated anti-rabbit IgG antibody, Alexa488-phalloidin and 4,6-diamidino-2-phenylindole (DAPI) were purchased from Invitrogen.

2.2. Surface preparation

Polyacrylamide substrates were prepared as previously described [29]. Briefly, 10 to 40 kPa stiffness gels were made by using mixtures of acrylamide/bis-acrylamide according to the desired stiffness [35]. For the polymerization, 0.1% ammonium persulfate (APS) and 0.1% of tetramethylethylenediamine (TEMED) were mixed in the gel solutions and 20 μL of the mixture was pipetted onto hydrophobically treated glass slides. After polymerization, the gels on the coverslips were detached and treated with hydrazine hydrate 55% for 2 hours with rocking [36]. 5% glacial acetic acid for 1 hour and then distilled water for 1 hour were used to rinse hydrazine and glacial acid, respectively. Polydimethylsiloxane (PDMS, Polysciences, Inc.) stamps were produced by conventional polymerization methods. Sodium periodate (~3.5 mg/ml) was used to generate free aldehydes on matrix proteins and added to 25 μg/mL of fibronectin in PBS for at least 45 min. The protein solution was pipetted onto patterned stamps for 30 min and dried with air. Free aldehydes in proteins were chemically conjugated with reactive hydrazide groups on the gels, resulting in transferring the protein residue on the stamps to the gel surfaces [36].

2.3. Cell source and culture

Human MSCs were purchased from Lonza. The MSCs were harvested and cultured from normal bone marrow. Cells were positive for CD105, CD166, CD29, and CD44 and negative for CD14, CD34 and CD45 by flow cytometry (http://www.lonza.com). Purchased MSCs from bone marrow were cultured and then expanded cells were frozen in cryopreservation (10% DMSO) with passage 2. Cells were thawed and cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) low glucose (1 g/mL) media supplemented with 10% fetal bovine serum (MSC approved FBS; Invitrogen), and 1% penicillin/streptomycin (p/s). Media was changed every 3 or 4 days. Passage 4–8 MSCs were seeded on patterned surfaces at a cell density of ~5000 cells/cm². MSCs were cultured for 10 days before analysis.

2.4. Immunocytochemistry and histology

Cells on surfaces were fixed with 4% paraformaldehyde (Alfa Aesar) for 20 min. To permeabilize cells, 0.1% Triton X-100 in PBS was employed for 30 min. Cells were blocked with 1% bovine serum albumin (BSA) for 15 min and labeled with primary antibody in 1% BSA in PBS for 2 h at room temperature (20°C) with mouse anti-MyoD, α5β1, or αVβ3 and rabbit anti-Runx2 or Osteopontin (1:500 dilution). Secondary antibody labeling was performed by the same procedure with Tetramethylrhodamine-conjugated anti-rabbit IgG antibody, Alexa Fluor 488-phalloidin (1:200 dilution), Alexa647-conjugated anti-mouse IgG antibody, and 4,6-diamidino-2-phenylindole (DAPI, 1:5000 dilution) for 20 min in a humid chamber (37°C). Immunofluorescence microscopy was conducted using a Zeiss Axiovert 200 M inverted research-grade microscope (Carl Zeiss, Inc.). Immunofluorescent images were analyzed using ImageJ; the fluorescence intensity of single cells (over 20 cells) for each condition was measured to compare different levels of marker expression. To stain for alkaline phosphatase, surfaces were rinsed with distilled water and incubated for 30 min in BCIP/NBT solution, rinsed well in PBS and imaged in bright field using a Motic trinocular inverted microscope. All experiments were repeated at least three times. Only single cells that were captured in patterns were used in the analysis. The relative intensity of the fluorescence was determined by comparing each intensity value to the average intensity of one condition. For Figure 2, 3, and 6, average marker intensities of circular cells in 5000 μm² patterned stiff (10 kPa) substrates were selected. For Figure 5, average mRNA expressions of cells in 5000 μm² circle patterned (10 kPa) substrates were selected. The intensity value for single cells was obtained from nuclei (Runx2 and MyoD) or cytoplasmic (Osteopontin) staining intensity minus backgrounds.

2.5. Traction stress measurement

Polyacrylamide gels with desired stiffness (10 and 30 kPa) were fabricated on a glass cover slip (18 mm) as described above [29]. To obtain fluorescent bead-infused gels, the polyacrylamide solution was mixed with a 1 μm-bead suspension (Invitrogen, F-8821) at 1:250 and a small amount (1~2 μl) was applied to gel solutions. Upon the placement of the gel surface faced down, beads in a single layer at the same focal plane were imaged using a fluorescent microscope. Matrix proteins were patterned as described above. An Olympus IX81 fluorescent microscope and 20x objective was used to obtain the live cell images [37]. Throughout the experiment, temperature and carbon dioxide levels were maintained at 37°C and 5% respectively. Live cell images on gels embedded with fluorescent beads were captured. Bright field images were firstly taken of the cells to visualize cell shape and location, and then fluorescent images of beads were taken. In order to assess the displacement of beads under the null-force condition, cells were removed from the surface using sodium dodecyl sulfate (SDS, Fisher Inc.), resulting in the gel returning to its relaxed initial state without cells. To characterize the gel displacements, the images before and after cell removal were analyzed using Matlab digital image correlation programs published in [37] to obtain the 2D displacement field (ux, uy). The resolution of the algorithm is 1/10 of pixel size, i.e. ~ 33 nm, and signal-to-noise ratio reaches 40. The detailed procedures of cell traction computation using finite element method can be found in a previous report [37]. In brief, our computation employed a mixed boundary condition model, by prescribing zero traction at all nodes outside the cell (Fx=Fy=Fz=0) and the obtained 2D displacement field (ux, uy) as well as Fz=0 at the nodes within the cell boundaries. We did not measure uz during the experiments. Our theoretical derivation suggests that for elastic biomaterial substrates with Poisson’s ratio close to 0.5, such as PA gels, prescribing Fz=0 for all surface nodes results in an error of less than 2% in the calculation of in-plane forces Fx and Fy [37].

2.6. RNA isolation and RT-PCR

Adherent cells were lysed directly in TRIZOL reagent (Invitrogen). Chloroform extraction and ethanol precipitation were employed to isolate total RNA. Total RNA was reverse transcribed using Superscript III^® First Strand Synthesis System for RT-PCR (Invitrogen). RT-PCR was achieved linearly by cycle number for each primer set using SYBR^® Green Real-Time PCR Master Mix (Invitrogen) on an Eppendorf Realplex 4S Realtime PCR system. Primer sequences were as follows: α1—CTC CTCACTGTTGTTCTACGCT and ATCCAAACATGTCTTCCACCG, α3—CCCACCTGGTGTGACTTCTT and TCCCTGGAGGTGGGTAGC, α5—TGCCGAGTTCACCAAGACTG and TGCAATCTGCTCCTGAGTGG, αv—CATCTTAATGTTGTGCCGGATGT and TCCTTCCACAATCCCAGGCT, α6—CAACTTGGACACTCGGGAGG and ACGAGCAACAGCCGCTT, β1—CCGCGCGGAAAAGATGAATTT and AGCAAACACACAGCAAACTGA, β3—TTGGAGACACGGTGAGCTTC and GCCCACGGGCTTTATGGTAA, GAPDH—TGCCTCGATGGGTGGAGT and GCCCAATACGACCAAATCAGA. All reactions were performed linearly by cycle number for each set of primers.

2.7. Inhibition assays

Inhibitors were added to cell culture media at the following concentrations before and after cell seeding and with each media change: Blebbistatin (1 μM) and Y-27632 (2 μM) (Calbiochem). Integrin blocking antibodies (α5β1 and αvβ3) were added to cells in media prior to deposition at 1 μg/mL. MAP kinase inhibition was performed by adding supplemented media of the following molecules at 6 μM after cell seeding and with each media change: FR180204 (ERK1/2), SP600125 (JNK), and SB202190 (p38).

2.8. Statistical analysis

Error bars represent standard deviation and N value is the number of experimental replicates. For statistical analysis one-way ANOVA for comparing multiple groups and two-tailed p-values from unpaired t-test for comparing two groups were employed and values of P < 0.05 were considered statistically significant.

3.1. Single cell patterning with adhesive proteins on polyacrylamide hydrogels of different stiffness

We used microcontact printing of adhesive proteins (fibronectin, laminin and collagen) on polyacrylamide (PAAm) hydrogels as a flexible platform to investigate the combinatorial effects of substrate elasticity, matrix composition and cell shape in controlling osteogenesis and myogenesis on protein-coated hydrogels (Fig. 1a). Polydimethylsiloxane (PDMS) stamps were fabricated using photolithography and used to pattern adhesive islands of proteins on the surface of chemically modified hydrogels. We employed two different shapes of identical area for patterning the hydrogels, one a simple circle and one approximating a star, where the cell body is expected to span non-adhesive regions [19]. PAAm hydrogels with a range of stiffness (10 to 40 kPa) were prepared as previously reported [29]. This range of stiffness is physiologically relevant with 10 and 30 kPa stiffness mimicking the rigidity of muscle or pre-calcified bone tissue, respectively [17]. The surfaces of PAAm gels were chemically modified with hydrazine hydrate [36], which allows for covalent immobilization between the treated gel surface and an oxidized glycoprotein solution via microcontact printing. After seeding, the mesenchymal stem cells (MSCs) attached and conformed to the patterned regions. To explore the influence of cell shape on the distribution of the cytoskeleton, we fixed and stained the patterned cultures for filamentous actin. Figure 1b shows actin stains and heatmaps of > 30 cells per shape which demonstrates classical cortical actin pattern for MSCs in circles, while MSCs in a star shape show pentagonally organized regions of actin stress fibers. Morphological analysis reveals that the patterned cells that adhere to the printed area show a comparable size to the defined patterns (5000 μm²) (Fig. 1c). Patterned cells stayed viable and maintained adhesion to the islands for 10 days in culture, but a higher number of cells on stiffer substrates and patterns with higher actomyosin contractility escaped from geometric confinement and proliferated (Fig. 1d).

3.2. The influence of cell shape, matrix stiffness and composition during mesenchymal stem cell differentiation

First we investigated osteogenic and myogenic marker expressions of MSCs cultured in the different shapes on fibronectin-coated hydrogels of varying stiffness (~10 to 40 kPa). We used three different osteogenic markers (Runx2 and ALP as early osteogenic markers; Osteopontin as a late osteogenic marker) and a myogenic marker (MyoD) to compare the degree of osteogenesis and myogenesis specification depending on matrix stiffness and cell shape after 10 days in culture (Fig. 2). Cells cultured on substrates with different stiffness express markers associated with osteogenesis and myogenesis in a stiffness dependent manner with a maximum at ~30 to 40 kPa. In addition, cells in star shapes show higher levels of osteogenic and myogenic marker expressions compared to those cultured in circular shapes. We also explored alternative shapes previously shown to influence actomyosin contractility: oval shapes with different aspect ratios (1:1, 2:1, 4:1, 8:1, and 12:1, 5000 μm²) (Fig. S1). Similar to cells on circle and star shapes, those on shapes that promote higher contractility express higher levels of osteogenic markers. Next we patterned our two shapes using different matrix proteins (fibronectin, laminin and collagen) across surfaces with different stiffness to explore how these cues influence lineage specification when presented in combination (Fig. 3). We used representative osteogenic (Runx2) and myogenic (MyoD) transcription factors to assess early differentiation to these lineages. Cells cultured on fibronectin or collagen matrices show increased Runx2 expression as substrate stiffness is increased, while MSCs cultured on laminin did not show a trend in differentiation on account of substrate stiffness. In contrast, MSC myogenesis was shown to be sensitive to substrate stiffness across all matrices. Changing the geometry of single MSCs from a circular shape to that approximating a star led to increased expression of Runx2 (fibronectin and collagen) and MyoD (fibronectin). However, similar to stiffness the shape of single cells on laminin did not influence osteogenesis.

3.3. The role of biophysical and biochemical parameters in guiding mesenchymal stem cell traction stress

Micropatterning single cells allows precise control over adhesive structures, and we postulated that the way in which MSCs deform their matrices would be influenced by shape, stiffness and protein composition. First, to explore the relationship between substrate mechanics, adhesion and differentiation, we measured the traction stress exerted by circular and star-shaped MSCs on hydrogels of two different stiffness (10 and 30 kPa), across three different matrix proteins (fibronectin, laminin, and collagen) (Fig. 4). We observed that cells on star shapes on fibronectin matrix showed higher traction stress than those on laminin (2.5-fold on 10 kPa; 7.2-fold on 30 kPa) or on collagen (4.3-fold on 10 kPa; 10.3-fold on 30 kPa). For the same shape and adhesive proteins, matrix stiffness gave rise to different levels of traction stresses; star shaped cells on fibronectin coated 30 kPa substrates displayed 3.2-fold higher traction stresses than those on 10 kPa gels. In addition, MSCs tended to exert higher traction when they were cultured in star geometries on fibronectin matrix (6.4-fold or 7.5-fold higher than circular cells on 10 or 30 kPa, respectively). However, traction exerted by cells on laminin substrates displayed no significant difference (within the limitations of small sample size) even when cultured on different stiffness or in the contractile star geometry. While stiffness influenced the MSCs’ ability to exert traction on collagen coated gels, there was no discernible influence of cell shape.

3.4. The expression of integrin receptors in response to cell geometry and matrix stiffness

Since MSCs cultured on fibronectin show clear differences in both differentiation and traction stress as a function of matrix stiffness and cell shape, we analyzed the expression of common integrin receptors involved in fibronectin recognition. Cells were cultured for 1 day on matrices of different stiffness (10 or 30 kPa) and in different geometries (circle or star shape) followed by lysis, RNA isolation and RT-PCR. Interestingly, MSCs cultured in the star shape show higher expression than MSCs cultured in circular shapes on both 10 kPa hydrogels (2.3-fold α1, 3.1-fold α3, 2.1-fold α5, 73-fold αv, 3.3-fold α6, 2.1-fold β1, 5.5-fold β3; Fig. S2) and 30 kPa (1.2-fold α1, 2.7-fold α3, 2.7-fold α5, 261.3-fold αv, 2.5-fold α6, 2.1-fold β1, and 2.6-fold β3; Fig. 5). In general integrin expression is higher for cells cultured in the star shape, but in particular integrin αv shows an enormous increase in expression for culture in the star geometry compared to the circle in both 10 and 30 kPa fibronectin conjugated hydrogels.

To further verify the observed trends in integrin expression, we performed immunofluorescence staining of a focal adhesion marker (Paxillin) and two major integrin receptors in fibronectin (α5β1 and αvβ3) (Fig. 6 and S3). Protein expression by immunofluorescence showed the same trend as the RT-PCR study: higher levels of focal adhesion and integrin expression for MSCs cultured in star shapes compared to those in circular shapes. Since we cultured MSCs for 10 days to study lineage specification and differentiation, we also measured paxillin, α5β1, and αvβ3 at day 10. Similar to cells cultured for 1 day, MSCs cultured for 10 days on star shaped fibronectin substrates displayed higher levels of focal adhesion proteins and integrin receptors.

3.5. Blocking integrin receptors and downstream signaling during differentiation of mesenchymal stem cells

MSCs cultured in star shapes show enhanced traction stress, integrin expression, and lineage specification to both osteogenesis and myogenesis programs. To elucidate signal transduction pathways that are involved in linking extracellular recognition to differentiation, we treated our patterned cultures with mitogen activated protein kinase (MAPK) inhibitors (p38, ERK1/2, and JNK), the Rho-associated kinase inhibitor Y-27632, the non-muscle myosin inhibitor blebbistatin, and integrin blocking antibodies for α5β1 and αvβ3. MSCs were cultured in 5000 μm² star geometries with or without 6 μM p38, ERK1/2, and JNK, 2 mM Y-27632, 1 mM blebbistatin, or 1 μg/mL anti-α5β1 and anti-αvβ3 for 10 days. We employed early (Runx2) and late (Osteopontin) osteogenic markers to investigate the effects of inhibitors on different stages of differentiation (Fig. 7). The expression of Runx2 shows a modest decrease after treatment with pharmacological inhibitors and blocking antibodies; however, the later marker Ostopontin shows a decreases on account of both blocking integrins and inhibiting downstream signal transduction players. Blocking integrin α5β1 in particular shows decreased expression of both Runx2 and Osteopontin, which suggests that signaling through this integrin plays a significant role during osteogenesis on these matrices.