Biomaterials 35 (2014) 4827e4834

Contents lists available at ScienceDirect

Biomaterials journal homepage: www.elsevier.com/locate/biomaterials

Cell adhesion mechanisms on laterally mobile polymer films Andreas P. Kourouklis a, Ronald V. Lerum b, Harry Bermudez b, * a b

Department of Chemical Engineering, University of Massachusetts, Amherst, MA, USA Department of Polymer Science and Engineering, University of Massachusetts, Amherst, MA, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 6 February 2014 Accepted 25 February 2014 Available online 18 March 2014

In contrast with the majority of substrates used to study cell adhesion, the natural extracellular matrix (ECM) is dynamic and remodeled over time. Here we use amphiphilic block copolymers to create selfassembled supported films with tunable lateral mobility. These films are intended to serve as partial mimics of the ECM in order to better understand cell adhesion responses, specifically in the context of dynamic substrates. Block copolymers are end-labeled with RGD peptide ligands to allow for integrinmediated cell adhesion, and the addition of a trace hydrophobic homopolymer is used to control the film lateral mobility. We find that NIH 3T3 fibroblasts cultured on these biomimetic films exhibit nonlinear spreading behavior in response to substrate mobility. In the absence of RGD ligands, however, fibroblasts do not spread. Employing quantitative analysis of focal adhesions (FA) and integrin ligation, we discover the presence of FA-dependent and FA-independent mechanisms responsible for the biphasic cell spreading behavior. The use of designed biomimetic platforms therefore yields insight into ECM mechanosensing by revealing that cells can engage distinct mechanisms to promote adhesion onto substrates with different time-dependent properties. Ó 2014 Elsevier Ltd. All rights reserved.

Keywords: Block copolymers Self-assembly Bilayers Lateral mobility Focal adhesions

1. Introduction Regulation of cellular function from a genetic or biochemical perspective has been appreciated and studied for many decades. Yet only recently have the biophysical effects on cellular function gained more attention. Towards understanding such effects, a large effort has been dedicated to the development of artificial materials [1,2] that mimic different characteristics of the native extracellular matrix (ECM) [3]. These artificial materials are designed to present cell-adhesive ligands or proteins, while displaying a range of physical properties such as texture [4], geometry [5], and stiffness [6e9]. In turn such materials allow the examination of processes including cell motility, differentiation and tumor progression [6,10e12]. However, with very few exceptions [13], artificial ECM materials involve the static display of signals and therefore are insufficient to mimic the dynamics of the ECM [14,15]. To explore the role of dynamics of the cellematerial interface, previous works have either used degradable hydrogels [16e19] or supported phospholipid bilayers [20e22]. In the former case the artificial material is provisional and intended for replacement with native ECM, whereas in the latter case substrates are generally unable to

* Corresponding author. Tel.: þ1 413 577 1413. E-mail address: [email protected] (H. Bermudez). http://dx.doi.org/10.1016/j.biomaterials.2014.02.052 0142-9612/Ó 2014 Elsevier Ltd. All rights reserved.

promote cell adhesion and spreading [20e23]. Towards overcoming these drawbacks, patterning techniques have been used to partition lipid bilayers with periodic barriers [24,25]. Nevertheless, both barriers and patterned substrates induce mechanotransduction in response to the static pattern density and/or geometry [22,26e28]. Therefore there is an unmet challenge in development of artifical materials that mimic native ECM characteristics such as: the dynamic display of ligands, cell-induced remodeling, no predefined spatial patterns, and the support of cell adhesion/ spreading. This challenge motivated us to examine the directed self-assembly of amphiphilic block copolymers as a potential platform. Such block copolymers share the amphiphilicity and mobility of lipids, while forming more stable structures due to their larger molecular weight [29e32]. These characteristics of amphiphilic block copolymer systems make them suitable candidates to mimic aspects of the native ECM. Indeed, related polymer systems have been used to study the role of ligand clustering [33] and ligand tether spacing [34] on cell migration and spreading, respectively. Towards achieving partial mimicry of the dynamic character of the ECM, here we fabricate ultrathin supported block copolymer films with independently tunable lateral mobility and ligand spacing. The lateral mobility is tuned by varying the amount of a “lubricating” homopolymer; a strategy inspired by the role of cholesterol in cell membranes [35,36]. Due to the self-assembly

4828

A.P. Kourouklis et al. / Biomaterials 35 (2014) 4827e4834

nature of the fabrication process, the average ligand spacing is easily controlled by the fraction of RGD-labeled polymer. Most importantly, the self-assembly fabrication process means that these films are susceptible to cell-induced remodeling. The effects of substrate lateral mobility on murine fibroblast responses are quantified by cell spreading and adhesion strength. At constant ligand spacing, we find that fibroblasts respond non-linearly to substrate mobility, indicating that cell spreading is not a simple function of typical static properties such as ligand density and substrate elasticity. Analysis of focal adhesions (FA) and integrin ligation leads us to propose that cell spreading can be realized by FA-dependent and FA-independent mechanisms (in the presence of sufficient ligand density). Our results reveal that the dynamic display of ligands, as in the native ECM, plays an important role in cellular responses. Thus the strategic design of biomaterials has the potential to provide critical insight on mechanosensing within the ECM.

whole field of view Tot(t). The prefactor Tot(0)/ROI(0) accounts for heterogeneities of fluorescence intensity at the starting point of the experiment,

2. Materials and methods

2.4. Cell projection area and adhesion strength

2.1. Materials The polymers 1,2-polybutadiene-b-poly(ethylene oxide) (1,2-PBd-PEO) of Mw ¼ 10 kg/mol (PDI ¼ 1.15) and wEO ¼ 0.40, and poly(isobutylene) (PIB) of Mw ¼ 0.9 kg/mol (PDI ¼ 1.3) were obtained from Polymer Source, Inc., (Canada), and were used as received. Bovine serum albumin (BSA) was purchased from Sigma and used as received. Silicon wafers were purchased from International Wafer Service, Inc., (California, USA) and glass coverslips were purchased from Fisher. TrypsinEDTA solution, Dulbecco’s modified Eagle’s medium (DMEM), penicillinestreptomycin solution, and calf bovine serum were supplied from ATCC. Y-27632 dihydrochloride was purchased from SigmaeAldrich. 2.2. Fabrication of supported block copolymer films Silicon wafers or glass coverslips were rinsed with EtOH and RO water, subjected to oxygen plasma treatment, and submerged in the RO water subphase of a Langmuir trough. Chloroform solutions of polymers were applied dropwise at the air/ water interface and left quiescent for 15 min before compression. The initial surface pressure after the addition of polymer solution and before compression was between 20 and 22 mN/m. The interfacial films were compressed at a rate of 10 mm/ min up to a surface pressure of 39 mN/m [37]. For the fabrication of supported monolayers, we used chloroform solutions of PB-PEO or its mixture with PIB homopolymer. Interfacial films were transferred from the air/water interface to the silicon wafers or glass coverslips at a constant deposition pressure and rate (39 mN/m, 1e2 mm/min) using LangmuireBlodgett (LB) deposition. Within an hour post-fabrication, the supported monolayers were used to create a supported bilayer by the Langmuir-Schaefer (LS) technique. LS deposition was allowed a contact time of 1 min between the supported monolayer and the interfacial film of PB-PEO. The dry thickness of silicon-supported films (Table S1) was determined by ellipsometry (LSE Stokes Ellipsometer 7109-C370, Gaertner). To allow for fluorescence recovery after photobleaching (FRAP), chloroform solutions of PB-PEO (z90 vol%) and PB-PEO-FITC (z10 vol%) were premixed and applied at the air/water interface. The fluorescent interfacial film was introduced as the topmost layer through LS deposition onto neat or PIB-doped PB-PEO monolayers. It has been previously shown that labeling a fraction of the polymer chains up to z25 vol% does not alter their diffusion characteristics [38]. Herein, a minimum concentration of 10 vol% of PB-PEO-FITC was necessary to gain the required contrast for the FRAP experiment. FRAP studies were initiated within 1 h after bilayer formation. For the cell adhesion studies, chloroform solutions of PB-PEO and PB-PEO-RGDS were premixed in a stoichiometric ratio that resulted in the desired RGD spacing. The calculation of the RGD spacing assumes ideal mixing between the polymer chains and employs the deposition surface density. The interfacial film containing PB-PEO-RGDS was introduced as the topmost layer through LS deposition onto neat and PIB-doped PB-PEO monolayers.

ROIðt Þcorr ¼

ROIðt Þ Totð0Þ  Totðt Þ ROIð0Þ

(1)

The corrected fluorescence intensity ROI(t)corr was normalized to span between 0 and 1, the ideal limits for no and full recovery respectively,

NðtÞ ¼

ROIðtÞcorr  ROIð0Þcorr : ROIðNÞcorr  ROIð0Þcorr

(2)

This normalized intensity N(t) was fitted to the fractional recovery curve (MATLAB, R2011a) defined in Soumpasis et al. [40] NðtÞ ¼ expð2s=tÞ½Io ð2s=tÞ þ I1 ð2s=tÞ;

(3)

to extract the characteristic time s of polymer diffusion at the topmost layer of the films. Using this value of the characteristic time we calculate the diffusion coefficient through D ¼ A/s, where A is the area of the bleaching spot. The diffusion coefficients obtained are the mean values from independent circular bleaching spots for the corresponding films (Figs S1 and S2).

Synchronized and enzymatically recovered fibroblasts were centrifuged (125 g, 10 min, 2) and then resuspended in complete DMEM. The RO water phase above freshly prepared polymer films was exchanged with PBS solution (3, 5 mL). PBS was exchanged with BSA solution (1 mg/mL, pH 7.4) (3, 5 mL) and left quiescent for film passivation (T ¼ 20  C, t ¼ 30 min). Afterward, the BSA solution was exchanged with complete DMEM (3, 3 mL). The cell suspension was added above the bilayer films to an initial surface concentration of 3  103 cells/cm2 and incubated for 24 h (T ¼ 37  C and 5% CO2). Only the cells that did not participate in cellecell contacts were used for cell projection area measurements. After image acquisition, the incubation wells were filled with complete DMEM of adjusted temperature (T ¼ 37  C) and sealed with para-film to avoid bubbles. The sealed wells were centrifuged at 600g for 10 min Ref. [33]. Following centrifugation, we discarded the media and counted the fraction of cells that remained attached. 2.5. Immunofluorescence staining After a seeding period of 24 h, the cells underwent fixation by transferring the coverslips to wells containing 4% formaldehyde (Carson-Millonig Formulation; Fisher Scientific) in PBS containing Ca2þ and kept at ambient temperature for 15e 20 min. Following three rinses with PBS, free aldehydes were quenched with 0.3 M glycine in PBS (3, 15 min) and permeabilized with 0.1% Triton X-100 for 5 min. To block non-specific interaction, 2% bovine serum albumin (BSA) in PBS was added and incubated for 60 min at ambient temperature. Cells were rinsed with 0.1 M EDTA in PBS (3, 5 min) to remove trace metals. Anti-vinculin-FITC (1:50 dilution, SigmaAldrich) was added and left in the dark for 60 min at room temperature. After rinsing with 0.1% Triton-X100 in PBS (3, 2 min) and 0.1 M EDTA in PBS (3, 5 min), actinphalloidin-orange (1unit; Molecular Probes) (2% BSA) in PBS was added for 30 min. The coverslips were mounted on microscope slides with ProLong antifade reagent (Molecular Probes) and left to cure overnight in the dark prior to image acquisition. 2.6. Pharmacological experiments Fibroblasts and polymer films were prepared as described above. Contractility inhibitor (Y-27632) in PBS was added at a final concentration of 50 mM and allowed a 30 min incubation time at room temperature [41] before measuring the new cell projection area and performing the centrifugationeadhesion assay (600g, 10 min). The fraction of remaining cells was measured and compared with the corresponding untreated control cells. 2.7. Statistics Unless otherwise noted, data are reported as mean values and error bars as the standard error of the mean. If analysis by ANOVA (Kaleidagraph 4.1.2) detected significant differences, Student-NewmaneKeuls multiple comparison tests were performed for pair-wise comparisons. Because ANOVA did not detect a statistically significant difference for FA sizes on different films, we employed the Student’s t-test for pair-wise comparisons. The error bars for the normalized area and adhesion strength data are the propagation error [42].

2.3. Fluorescence recovery after photobleaching (FRAP) FRAP experiments were performed on a confocal microscope (Zeiss510) using 50% of argon laser (488 nm) intensity with a 40x oil immersion objective. FRAP experiments were conducted with the films immersed in reverse-osmosis water at T ¼ 20  C. The fluorescence intensity was doubly normalized according Phair et al. [39]. Specifically, we corrected for acquisition bleaching by division of the fluorescence intensity at the region of interest ROI(t) with the corresponding intensity of the

3. Results 3.1. Fabrication and characterization of laterally mobile films Our supported polymer films are created by LangmuireBlodgett/LangmuireSchaefer (LB/LS) self-assembly (Fig. 1), which allows

A.P. Kourouklis et al. / Biomaterials 35 (2014) 4827e4834

4829

Fig. 1. Schematic of supported block copolymer films. (a) Cross-section view illustrates the bilayer structure of the films, the RGD ligands at the topmost surface, and the location of the PIB homopolymer. (b) Top view has dashed lines depicting the diffusive motion of RGD ligands (for clarity only two paths are shown).

(i) control over the distance between polymer chains [37], (ii) independent composition of the upper and lower layers, and (iii) the physical introduction of distinct polymers during film fabrication (Fig. 1). With regard to the last point, the incorporation of trace polyisobutylene (PIB) homopolymer alters the block copolymer chain dynamics within the film, without any significant degree of swelling normal to the substrate (Table S1). To determine the lateral (i.e., two-dimensional) diffusion coefficient of the block copolymers in the upper layer of the film, we performed fluorescence recovery after photobleaching (FRAP) [43]. FRAP yields a characteristic time s from which the lateral diffusion coefficient is calculated (Figs. S1 and S2). The film diffusion coefficients D increase logarithmically with the concentration of PIB homopolymer (Fig. 2). Thus by the addition of homopolymer during film fabrication, the film lateral mobility can be readily tuned. The changes in D are statistically significant (p < 0.05), with the exception of the films with PIB volume fraction F corresponding to 105 and 108.

any differences in cell behavior can be attributed solely to differences in the presentation of adhesive ligands. The projected cell area A of fibroblasts undergoes a sharp initial decrease but thereafter increases with mobility D (Fig. 3). This biphasic cell spreading response suggests that more than one mechanism is responsible for the observed behavior, dominating at different extremes of substrate mobility. In the absence of any

3.2. Cell spreading response to substrate mobility Our supported films allow for independent control over ligand spacing and lateral mobility, providing unique substrates to interrogate RGD-mediated [44e47] cell adhesion and spreading. We fabricated films with different substrate mobilities at an average inter-ligand spacing of d ¼ 50 nm (Fig. 1). This spacing was chosen based on works that established the critical ligand spacing dc capable of supporting fibroblast spreading: dc ¼ 58e140 nm [48,49]. Because our ligand spacing is below the critical value, adherent cells are not limited with respect to ligand availability and

Fig. 2. Lateral diffusion coefficient D versus volume fraction of PIB homopolymer added to supported block copolymer films.

4830

A.P. Kourouklis et al. / Biomaterials 35 (2014) 4827e4834

ligand (i.e., infinite spacing), there is no significant change in spreading behavior with D (dashed line in Fig. 3), explained by a lack of cell interactions with our bare polymer films [37]. 3.3. Focal adhesion responses to substrate mobility

Fig. 3. Projected cell area A of NIH 3T3 fibroblasts versus substrate mobility D with ligand spacing d ¼ 50 nm. The dashed line corresponds to films without RGD (d ¼ N).

Because they are correlated with cell spreading and features such as adhesion strength [50e52], we monitored the organization of intracellular proteins, specifically F-actin and vinculin (Fig. 4). Analysis of the FA size distribution obtained by vinculin immunostaining reveals that the FA density also shows a biphasic response to substrate mobility (Fig. 5a), strongly suggesting the importance of FAs to the cell spreading response. The FA size strictly decreases with substrate mobility (Fig. 5b), which agrees with previous works showing that FA size is proportional to the magnitude of cell-generated lateral forces applied to the substrate [53]. In other words, these data support the notion that cells transmit progressively less force on films with higher substrate mobility. From this view alone, contractility-dependent cell spreading [54] should decrease with substrate mobility, suggesting the possibility of other mechanisms contributing to the biphasic response in Fig. 3.

Fig. 4. Immunostaining images of NIH 3T3 fibroblasts seeded for 24 h under complete serum conditions on films with different substrate mobility. Red and green (left and right) images correspond to the immunostaining of actin and vinculin, respectively. Panels A and B correspond to substrate mobility of D ¼ 1010 cm2/sec, panels C and D correspond to D ¼ 1.75  1010 cm2/sec, and panels E and F correspond to D ¼ 3.5  1010 cm2/sec. Scale bar is 20 mm.

A.P. Kourouklis et al. / Biomaterials 35 (2014) 4827e4834

4831

Fig. 6. Normalized change in adhesion strength d(W)/W versus substrate mobility D following administration of the contractility inhibitor Y-27632.

Fig. 5. Focal adhesion (a) density and (b) size versus substrate mobility D, as determined by vinculin immunostaining.

3.4. Contractility-independent mechanisms of cell adhesion We therefore examined if there are other contributions to cell adhesion and spreading on our films besides those directly mediated by cell contractility. To this end we suppressed cellular force generation by treatment with the Rho kinase inhibitor Y-27632 [41,55]. Dumbauld et al. [41] demonstrated that Y-27632 treatment of NIH 3T3 fibroblasts results in the dissolution of FAs. Importantly, this dissolution of FAs by Y-27632 does not affect integrin ligation. We determined cell adhesion strength following Y-27632 treatment using a detachment assay [33]. In this assay the adhesion strength is proportional to the percentage of cells that remain attached. We normalize the adhesion strength relative to untreated cells d(W)/W ¼ Ntreated  Nuntreated/Nuntreated in order to isolate the effect of Y-27632. Fig. 6 shows that d(W)/W < 0 for low mobility films, indicating a loss of adherent cells relative to the untreated control and a strong role for FAs in adhesion strength on these substrates. On the contrary, at higher substrate mobility d(W)/ W > 0, meaning that the number of cells attached remains the same or even increases. These results indicate that Y-27632 does not affect adhesion strength on high mobility substrates. We therefore suggest that on high mobility substrates, clusters of ligated

Fig. 7. (a) Normalized change in cell area d(A)/A and (b) difference in normalized change in cell area 1  d(A)/A versus substrate mobility D, after administration of the contractility inhibitor Y-27632.

4832

A.P. Kourouklis et al. / Biomaterials 35 (2014) 4827e4834

integrins collectively strengthen cell adhesion to the same extent as untreated controls, even though they do not efficiently promote FA growth. Having established roles for ligated integrins both inside and outside FAs in cell adhesion, we proceeded to examine their relative contributions to cell spreading. Because of the effect of Y27632 on FAs, we can decouple the contributions from ligated integrins both inside and outside FAs as follows: d(A)/A z AFA/Ai and 1  d(A)/A z Aoutside/Ai, where the change in cell area d(A)/A is normalized to the initial cell area Ai. The quantity d(A)/A z AFA/Ai accounts for contributions of ligated integrins inside FAs to cell adhesion. This quantity decreases with substrate mobility (Fig. 7a), indicating that Y-27632 has a reduced influence on cell spreading for high mobility substrates. This response can also be partly understood by our earlier result that FA size decreases with substrate mobility (see Fig. 5b). The quantity 1  d(A)/A z Aoutside/ Ai, accounts for contributions of ligated integrins outside FAs to cell adhesion. It is seen that the importance of ligated integrins outside FAs is enhanced as substrate mobility is increased (Fig. 7b). The biphasic response of cell area with substrate mobility in Fig. 3 is therefore due to a combination of separate effects. The first effect is at low mobility and is due to ligated integrins inside FAs and cell contractility. This effect causes cell area to decrease with mobility. The second effect is at high mobility and is due to ligated

integrins outside FAs and their mobility-enabled clustering. This effect causes cell area to increase with mobility. The two effects together give rise to a biphasic response.

4. Discussion 4.1. Substate mobility as a step towards mimicry of the ECM We created films with tunable lateral mobility so as to examine the nature of cell responses on “mobile” substrates. This is in contrast to a great deal of previous work with “static” substrates that are largely based on highly crosslinked networks [6e8,56]. Although our polymer films are two-dimensional in nature, they capture a key feature of cell adhesion to the ECM: the ability to induce changes in the substrate over time. The present films are composed of a brush-like block copolymer bilayer [37] with trace amounts of hydrophobic homopolymer. The concentration of the latter directly modulates the lateral mobility D (Fig. 2). Because we have previously shown that these films are nonfouling [37], the cellefilm interactions reported here are specifically due to integrin-RGD binding. The mobility of our films can be tuned within a range that lies between polymersomes (D z 1010 cm2/s [57]), protein-tethered lipids in supported lipid bilayers (D z 109 cm2/s [25]), and integrin transmembrane proteins (D z 109 cm2/s [58]).

Fig. 8. Schematic of different spreading mechanisms on low and high mobility substrates. (a) Initial state with few bound integrins (light green ligands) and predominantly unbound integrins (white ligands). (b) Intermediate state with clustering of ligated integrins (green ligands) favored on high mobility substrates. (c) Final state with mature FAs (dark green ligands) favored on low mobility substrates and a contribution by ligated integrin clusters on high mobility substrates. For clarity, the schematic is not drawn to scale. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

A.P. Kourouklis et al. / Biomaterials 35 (2014) 4827e4834

4833

4.2. Distinct modes of cell spreading due to substrate mobility

Appendix A. Supplementary data

It is not our aim here to determine the precise critical spacing required for cell spreading on “mobile” substrates. Rather, we utilize ligand spacing that are sufficient to establish adhesion and spreading so that we can focus attention on the role of film mobility. Specifically, our ligand spacing of d ¼ 50 nm is below previously reported critical ligand spacings dc of 140 nm and 58 nm [48,49]. The variation in dc is due to differences in ligand chemistry, substrates, and other factors. Our observation that actin stress fiber formation is irrespective of film mobility (Fig. 4), is supportive of a ligand density sufficient to engage integrin receptors. As expected, the limit of infinite spacing (i.e., no RGD ligand) leads to the absence of cell spreading (dashed line in Fig. 3). Importantly, Fig. 3 shows that ligand density does not guarantee a cell spreading response; the mobility of the ligands plays a critical role that has been previously unexplored. The observation of biphasic cell spreading with substrate mobility (Fig. 3) and subsequent experiments lead us to propose that there are two distinct contributions by which cells can adhere and spread (Fig. 8). At low substrate mobility, cellular contractile forces are efficiently sustained and induce the formation of mature FAs (Figs. 5 and 8c). These FAs provide the anchorage points at the cell’s leading edge to sustain actomyosin-driven cell adhesion and spreading [54]. At high substrate mobility, ligated integrins are able to more efficiently find each other and form clusters (Fig. 8b) [26,59]. A fraction of clusters forms mature FAs (Fig. 5a), although the size of these FAs is limited by the force they can sustain (Fig. 5b). The other fraction of clusters remains outside of FAs but can collectively support cell adhesion (Fig. 6) even though they are lacking in cytoplasmic reinforcement. By monitoring changes in cell area following treatment of cells with Y-27632, we are able to discriminate between the roles of ligated integrins inside and outside FAs. The contribution of ligated integrins inside FAs on cell spreading progressively diminishes with substrate mobility (Fig. 7a), indicating that ligated integrins outside FAs compensate. The overall result is cell spreading at both extremes of substrate mobility (Figs. 3 and 8c).

Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.biomaterials.2014.02.052.

5. Conclusions We have studied cell adhesion and spreading on supported polymer films created by an interfacial self-assembly process. These substrates are laterally mobile and display cell-adhesive peptides, towards partially mimicry of the native ECM character. Using NIH 3T3 fibroblasts, we find that substrate mobility causes a previously unreported biphasic spreading response. Subsequent experiments decouple the mobility-dependent contributions of FAs and integrin clustering to cell adhesion strength and spreading. These separate contributions to cell spreading are most prominent at the extremes of substrate mobility. Our results show that cells sense substrate mobility, and in response, they use more than one mechanism to promote spreading. Such dynamic cell behavior reflects the inherent ability of cells to sense and adapt to cues from the ECM, and furthermore is in sharp contrast to previous works with cells on immobile substrates (e.g., crosslinked hydrogels). The present work highlights the need for future studies on cell-substrate interactions to consider the time-dependent mechanical properties of the ECM in addition to factors such as adhesive patterns, gradients, and biochemical cues. Acknowledgment We thank H. Aranda-Espinoza, T. H. Barker, and D. E. Discher for helpful discussions. We acknowledge the NSF for financial support (DMR-0847558) and the MRSEC at UMass e Amherst (DMR0820506) for use of their facilities.

References [1] Stevens MM, George JH. Exploring and engineering the cell surface interface. Science 2005;310:1135e8. [2] Lutolf MP, Hubbell JA. Synthetic biomaterials as instructive extracellular microenvironments for morphogenesis in tissue engineering. Nat Biotechnol 2005;23:47e55. [3] Christian Frantz KMS, Weaver VM. The extracellular matrix at a glance. J Cell Sci 2010;24:4195e200. [4] Dalby MJ, Gadegaard N, Tare R, Andar A, Riehle MO, Herzyk P, et al. The control of human mesenchymal cell differentiation using nanoscale symmetry and disorder. Nat Mater 2007;6:997e1003. [5] Nelson CM, Vanduijn MM, Inman JL, Fletcher DA, Bissell MJ. Tissue geometry determines sites of mammary branching morphogenesis in organotypic cultures. Science 2006;314:298e300. [6] Engler A, Bacakova L, Newman C, Hategan A, Griffin M, Discher D. Substrate compliance versus ligand density in cell on gel responses. Biophys J 2004;86: 617e28. [7] Yeung T, Georges P, Flanagan L, Marg B, Ortiz M, Funaki M, et al. Effects of substrate stiffness on cell morphology, cytoskeletal structure, and adhesion. Cell Motil Cytoskelet 2005;60:24e34. [8] Solon J, Levental I, Sengupta K, Georges PC, Janmey PA. Fibroblast adaptation and stiffness matching to soft elastic substrates. Biophys J 2007;93:4453e61. [9] Ghosh K, Pan Z, Guan E, Ge S, Liu Y, Nakamura T, et al. Cell adaptation to a physiologically relevant ECM mimic with different viscoelastic properties. Biomaterials 2007;28:671e9. [10] Lo CM, Wang HB, Dembo M, Wang YL. Cell movement is guided by the rigidity of the substrate. Biophys J 2000;79:144e52. [11] Cukierman E, Pankov R, Stevens DR, Yamada KM. Taking cell-matrix adhesions to the third dimension. Science 2001;294:1708e12. [12] Levental KR, Yu H, Kass L, Lakins JN, Egeblad M, Erler JT, et al. Matrix crosslinking forces tumor progression by enhancing integrin signaling. Cell 2009;139:891e906. [13] Jiang FX, Yurke B, Schloss RS, Firestein BL, Langrana NA. The relationship between fibroblast growth and the dynamic stiffnesses of a DNA crosslinked hydrogel. Biomaterials 2010;31:1199e212. [14] Daley WP, Peters SB, Larsen M. Extracellular matrix dynamics in development and regenerative medicine. J Cell Sci 2008;121:255e64. [15] Antia M, Baneyx G, Kubow KE, Vogel V. Fibronectin in aging extracellular matrix fibrils is progressively unfolded by cells and elicits an enhanced rigidity response. Faraday Discuss 2008;139:229e49. discussion 309e25, 419e20. [16] Halstenberg S, Panitch A, Rizzi S, Hall H, Hubbell JA. Biologically engineered protein-graft-poly(ethylene glycol) hydrogels: a cell adhesive and plasmindegradable biosynthetic material for tissue repair. Biomacromolecules 2002;3:710e23. [17] Holland TA, Tabata Y, Mikos AG. In vitro release of transforming growth factor-beta 1 from gelatin microparticles encapsulated in biodegradable, injectable oligo(poly(ethylene glycol) fumarate) hydrogels. J Control Release 2003;91:299e313. [18] Rizzi SC, Hubbell JA. Recombinant protein-co-peg networks as cell-adhesive and proteolytically degradable hydrogel matrixes. Part I: development and physicochemical characteristics. Biomacromolecules 2005;6:1226e38. [19] Kloxin AM, Kasko AM, Salinas CN, Anseth KS. Photodegradable hydrogels for dynamic tuning of physical and chemical properties. Science 2009;324: 59e63. [20] Garcia AS, Dellatore SM, Messersmith PB, Miller WM. Effects of supported lipid monolayer fluidity on the adhesion of hematopoietic progenitor cell lines to fibronectin-derived peptide ligands for alpha 5 beta 1 and alpha 4 beta 1 integrins. Langmuir 2009;25:2994e3002. [21] McBeath R, Pirone DM, Nelson CM, Bhadriraju K, Chen CS. Cell shape, cytoskeletal tension, and RhoA regulate stem cell lineage commitment. Dev Cell 2004;6:483e95. [22] Chen C, Mrksich M, Huang S, Whitesides G, Ingber D. Geometric control of cell life and death. Science 1997;276:1425e8. [23] Ananthanarayanan B, Little L, Schaffer DV, Healy KE, Tirrell M. Neural stem cell adhesion and proliferation on phospholipid bilayers functionalized with RGD peptides. Biomaterials 2010;31:8706e15. [24] Salaita K, Nair PM, Petit RS, Neve RM, Das D, Gray JW, et al. Restriction of receptor movement alters cellular response: physical force sensing by EphA2. Science 2010;327:1380e5. [25] Perez T, Nelson W, Boxer S, Kam L. E-cadherin tethered to micropatterned supported lipid bilayers as a model for cell adhesion. Langmuir 2005;21: 11963e8. [26] Ch Yu, Law JBK, Suryana M, Low HY, Sheetz MP. Early integrin binding to ArgGly-Asp peptide activates actin polymerization and contractile movement that stimulates outward translocation. Proc Natl Acad Sci U S A 2011;108: 20585e90.

4834

A.P. Kourouklis et al. / Biomaterials 35 (2014) 4827e4834

[27] Arnold M, Schwieder M, Blü mmel J, Cavalcanti-Adam EA, Ló pez-Garcia M, Kessler H, et al. Cell interactions with hierarchically structured nanopatterned adhesive surfaces. Soft Matter 2009;5:72e7. [28] Huang J, Grater SV, Corbellinl F, Rinck S, Bock E, Kemkemer R, et al. Impact of order and disorder in RGD nanopatterns on cell adhesion. Nano Lett 2009;9: 1111e6. [29] Discher D, Eisenberg A. Polymer vesicles. Science 2002;297:967e73. [30] Discher DE, Janmey P, Wang YL. Tissue cells feel and respond to the stiffness of their substrate. Science 2005;310:1139e43. [31] Bermudez H, Brannan A, Hammer D, Bates F, Discher D. Molecular weight dependence of polymersome membrane structure, elasticity, and stability. Macromolecules 2002;35:8203e8. [32] Lee J, Bermudez H, Discher B, Sheehan M, Won Y, Bates F, et al. Preparation, stability, and in vitro performance of vesicles made with diblock copolymers. Biotechnol Bioeng 2001;73:135e45. [33] Maheshwari G, Brown G, Lauffenburger D, Wells A, Griffith L. Cell adhesion and motility depend on nanoscale RGD clustering. J Cell Sci 2000;113: 1677e86. [34] Kuhlman W, Taniguchi I, Griffith LG, Mayes AM. Interplay between PEO tether length and ligand spacing governs cell spreading on RGD-modified PMMA-gPEO comb copolymers. Biomacromolecules 2007;8:3206e13. [35] Almeida PF, Vaz WL, Thompson TE. Lateral diffusion in the liquid phases of dimyristoylphosphatidylcholine/cholesterol lipid bilayers: a free volume analysis. Biochemistry 1992;31:6739e47. [36] Babayco CB, Turgut S, Smith AM, Sanii B, Land D, Parikh AN. A comparison of lateral diffusion in supported lipid monolayers and bilayers. Soft Matter 2010;6:5877e81. [37] Lerum RV, Bermudez H. Controlled interfacial assembly and transfer of brushlike copolymer films. Chemphyschem 2010;11:665e9. [38] Hamersky M, Tirrell M, Lodge T. Self-diffusion of a polystyrene-polyisoprene block copolymer. J Polym Sci B 1996;34:2899e909. [39] Phair RD, Gorski SA, Misteli T. Measurement of dynamic protein binding to chromatin in vivo, using photobleaching microscopy. Methods Enzymol 2004;375:393e414. [40] Soumpasis DM. Theoretical-analysis of fluorescence photobleaching recovery experiments. Biophys J 1983;41:95e7. [41] Dumbauld DW, Shin H, Gallant ND, Michael KE, Radhakrishna H, Garcia AJ. Contractility modulates cell adhesion strengthening through focal adhesion kinase and assembly of vinculin-containing focal adhesions. J Cell Physiol 2010;223:746e56. [42] Shoemaker DP, Garland CW, Nibler JW. Experiments in physical chemistry. 6th ed. New York: McGraw-Hill; 1996. [43] Axelrod D, Koppel DE, Schlessinger J, Elson E, Webb WW. Mobility measurement by analysis of fluorescence photobleaching recovery kinetics. Biophys J 1976;16:1055e69.

[44] Houseman B, Mrksich M. The microenvironment of immobilized Arg-Gly-Asp peptides is an important determinant of cell adhesion. Biomaterials 2001;22: 943e55. [45] Biesalski MA, Knaebel A, Tu R, Tirrell M. Cell adhesion on a polymerized peptide-amphiphile monolayer. Biomaterials 2006;27:1259e69. [46] Hersel U, Dahmen C, Kessler H. RGD modified polymers: biomaterials for stimulated cell adhesion and beyond. Biomaterials 2003;24:4385e415. [47] Chollet C, Chanseau C, Remy M, Guignandon A, Bareille R, Labrugère C, et al. The effect of RGD density on osteoblast and endothelial cell behavior on RGDgrafted polyethylene terephthalate surfaces. Biomaterials 2009;30:711e20. [48] Massia S, Hubbell J. An RGD spacing of 440nm is sufficient for integrin alphav- beta-3-mediated fibroblast spreading and 140nm for focal contact and stress fiber formation. J Cell Biol 1991;114:1089e100. [49] Arnold M, Cavalcanti-Adam E, Glass R, Blummel J, Eck W, Kantlehner M, et al. Activation of integrin function by nanopatterned adhesive interfaces. Chemphyschem 2004;5:383e8. [50] Gallant ND, Michael KE, García AJ. Cell adhesion strengthening: contributions of adhesive area, integrin binding, and focal adhesion assembly. Mol Biol Cell 2005;16:4329e40. [51] Elineni KK, Gallant ND. Regulation of cell adhesion strength by peripheral focal adhesion distribution. Biophys J 2011;101:2903e11. [52] Baxter LC, Frauchiger V, Textor M, ap Gwynn I, Richards RG. Fibroblast and osteoblast adhesion and morphology on calcium phosphate surfaces. Eur Cell Mater 2002;4:1e17. [53] Balaban NQ, Schwarz US, Riveline D, Goichberg P, Tzur G, Sabanay I, et al. Force and focal adhesion assembly: a close relationship studied using elastic micropatterned substrates. Nat Cell Biol 2001;3:466e72. [54] Giannone G, Dubin-Thaler BJ, Rossier O, Cai Y, Chaga O, Jiang G, et al. Lamellipodial actin mechanically links myosin activity with adhesion-site formation. Cell 2007;128:561e75. [55] Kernochan LE, Tran BN, Tangkijvanich P, Melton AC, Tam SP, Yee Jr HF. Endothelin-1 stimulates human colonic myofibroblast contraction and migration. Gut 2002;50:65e70. [56] Engler AJ, Sen S, Sweeney HL, Discher DE. Matrix elasticity directs stem cell lineage specification. Cell 2006;126:677e89. [57] Lee J, Santore M, Bates F, Discher D. From membranes to melts, rouse to reptation: diffusion in polymersome versus lipid bilayers. Macromolecules 2002;35:323e6. [58] Rossier O, Octeau V, Sibarita JB, Leduc C, Tessier B, Nair D, et al. Integrins 1 and 3 exhibit distinct dynamic nanoscale organizations inside focal adhesions. Nat Cell Biol 2012;14:1057e67. [59] Dustin ML, Ferguson LM, Chan PY, Springer TA, Golan DE. Visualization of CD2 interaction with LFA-3 and determination of the two-dimensional dissociation constant for adhesion receptors in a contact area. J Cell Biol 1996;132: 465e74.