Steering cell orientation through light-based spatiotemporal modulation of the mechanical environment

The anisotropic organization of cells and the extracellular matrix (ECM) is essential for the physiological function of numerous biological tissues, including the myocardium. This organization changes gradually in space and time, during disease progression such as myocardial infarction. The role of mechanical stimuli has been demonstrated to be essential in obtaining, maintaining and de-railing this organization, but the underlying mechanisms are scarcely known. To enable the study of the mechanobiological mechanisms involved, in vitro techniques able to spatiotemporally control the multiscale tissue mechanical environment are thus necessary. Here, by using light-sensitive materials combined with light-illumination techniques, we fabricated 2D and 3D in vitro model systems exposing cells to multiscale, spatiotemporally resolved stiffness anisotropies. Specifically, spatial stiffness anisotropies spanning from micron-sized (cellular) to millimeter-sized (tissue) were achieved. Moreover, the light-sensitive materials allowed to introduce the stiffness anisotropies at defined timepoints (hours) after cell seeding, facilitating the study of their temporal effects on cell and tissue orientation. The systems were tested using cardiac fibroblasts (cFBs), which are known to be crucial for the remodeling of anisotropic cardiac tissue. We observed that 2D stiffness micropatterns induced cFBs anisotropic alignment, independent of the stimulus timing, but dependent on the micropattern spacing. cFBs exhibited organized alignment also in response to 3D stiffness macropatterns, dependent on the stimulus timing and temporally followed by (slower) ECM co-alignment. In conclusion, the developed model systems allow improved fundamental understanding of the underlying mechanobiological factors that steer cell and ECM orientation, such as stiffness guidance and boundary constraints.


Introduction
The orientation and interaction of cells and extracellular matrix (ECM) components are strong determinants of biological tissue functionality [1,2].In the heart, the anisotropy of the ECM present in the myocardial microenvironment is essential for coordinated cardiac contractions and force transmission through the multiple length scales of the myocardial tissue [3,4].At the cellular level, the anisotropy is tightly linked to the orientation of cardiac fibroblasts (cFBs), which produce the ECM and mechanically remodel it via traction forces.Given their fundamental role for ECM organization, incorrect organization of cFBs can exacerbate pathologies.During cardiac diseases, e.g.hypertrophic cardiomyopathy or myocardial infarction, pathological signals induce proliferation and activation of the cFBs, which in turn produce disorganized fibrotic scar tissue that matures over time, hampering cardiac mechanical performance [5].The development of material-based strategies aimed to restore the physiological cFB orientation, including cardiac patches and hydrogel delivery, could improve ECM organization and, thus, cardiac contractility [6].However, these strategies are currently hampered by the lack of in vitro systems that could be employed not only to investigate the biological processes driving cell (re)orientation, but also to induce such cell response at different time and length scales relevant during disease progression [7].
To date, two-dimensional (2D) in vitro model systems have been fundamental to uncover stimuli that can lead to cell orientation.For example, it has been shown that linear cues given by either (micro)topographies [8][9][10] or ligand patterns [11] can induce cells to align in the direction of the lines, a phenomenon called contact guidance [12].More recently, it has been observed that cells also align in response to linear stiffness cues present on their substrate [13][14][15][16][17][18].Although insightful, these studies have concentrated solely on cell orientation in response to constant stimuli, limiting our understanding of how dynamic changes in ECM stiffness can influence cellular orientation.Moreover, fabrication techniques employed previously aimed at fabrication stiffness cues also elicit topographical changes that could be a confounding factor in understanding the cell orientation response.Other studies have shown that light-sensitive crosslinking of hydrogels allows to spatiotemporally tune the mechanical properties of 2D substrates [19][20][21].Consequently, these approaches could be utilized to manipulate cell orientation not only spatially, but also over time.
In 3D environments, cells exhibit different behaviours than in 2D, primarily due to the influence of ECM signals that shape their responses and functions [22].In 3D, the anisotropy of the environment can be imposed via constraining the construct uniaxially, generating a stiffness anisotropy throughout the tissue [23][24][25].Generally, these systems consist of microfabricated flexible posts separated by hundreds of micrometers with cells and ECM seeded in between [26,27].Upon exerting contractile forces on the surrounding ECM, cells induce displacements in anchoring posts.This results in the alignment of both cells and ECM fibers along the direction of applied tension.By changing the spatial configuration of the posts, the overall tissue orientation can be tuned on demand going from highly anisotropic to completely isotropic, mimicking different disease stages [28].These systems have provided a fundamental understanding of how cells respond to stiffness anisotropies in 3D systems and other mechanical stimuli, such as cyclic stretch [29][30][31].However, a major limitation of these systems is the constant and unmodifiable stiffness anisotropy generated by the pillars, which restricts temporal control of the constraints.
In the present study, we overcome the abovementioned limitations used to study cell-ECM anisotropy by creating 2D and 3D in vitro systems that allow spatiotemporal control of tissue stiffness anisotropies.To this end, we used a commercially available light-sensitive hydrogel, gelatin methacryolyl (GelMA), which crosslinks upon UV-illumination thereby increasing its stiffness.This approach allows the creation of spatial stiffness patterns with micrometer resolution (micropatterns).For the 2D system, GelMA was exposed to micrometer-resolution UV patterns via an illumination device consisting of an UV laser filtered through a digital micromirror device (DMD) [21].Inspired by studies showing that linear cues (e.g. protein patterns) direct cellular orientation in 2D [11], we designed linear stiffness patterns with similar micron-sized dimensions.The UV-mediated crosslinking of GelMA, allowed to impose the stiffness patterns on cFBs at a certain timepoint (6 h) after cell seeding.In this system, cFBs aligned in the direction of the stiffness patterns, depending on the pattern width but independent of the timepoint at which the stiffness patterns were created.The 3D system consisted of a cFB-containing collagenous hydrogel that was constrained with cast GelMA hydrogel to generate the stiffness anisotropies at the millimeter scale (macropatterns).Using GelMA to constrain the cFB-populated collagen hydrogel provides the flexibility to cast it with specific spacing and timing, thereby enabling on-demand control over the stiffness anisotropies.cFBs and ECM oriented in the direction of the stiffness constraints, with their orientation being influenced by both the cell density and the timing of the stiffness stimulus application.Therefore, in the present study, we developed 2D and 3D in vitro systems to study how stiffness anisotropies affect cell orientation.The use of light-sensitive hydrogels enabled us to introduce the anisotropies over time, providing novel methods for investigating how stiffness cues can be utilized to steer cellular and ECM orientation with direct application in tissue engineering and regeneration.

Cell source
Human cFBs were derived from fetal epicardium.Human fetal cardiac tissue was anonymously collected with informed consent from elective abortion material of fetuses.Fetal epicardial layers were isolated by separating the epicardium from the underlying myocardium of human hearts aged 14-19 weeks post-gestation [32].This research was carried out according to the official guidelines of the Leiden University Medical Center and approved by the local Medical Ethics Committee (No. P08.087).cFBs were cultured in high-glucose Dulbecco's modified Eagle's medium (Invitrogen, Breda, the Netherlands) supplemented with 10% fetal bovine serum (Greiner Bio-one) and 1% Penicillin-Streptomycin (culture medium).To promote cell attachment, flasks were coated with 0.1% gelatin derived from porcine skin in PBS (Sigma-Aldrich).Passaging of the cFBs occurred when they reached 80% confluency.The experimental procedures were performed using cFBs with passage numbers ranging from 8 to 13.

Two-dimensional (2D
) spatiotemporal stiffness anisotropy in vitro system 2.2.1.Mold fabrication 4 mm thick polydimethylsiloxane (PDMS; Sylgard 184 ® Silicone Elastomer Kit, Dow Corning) blocks were fabricated by mixing the base and curing agent (10:1 ratio).The mixed solution was poured in a 100 mm in diameter petri dish and subsequently vacuumed at room temperature (RT) to remove air bubbles from the mixture.Finally, the PDMS was cured at 65 • C for 2 h in a dry oven.Hollow cylindric molds of 5 and 8 mm in inner and outer diameter, respectively, were created cutting the PDMS with biopsy punches.
To prepare the solutions, freeze dried GelMA and HAMA were resuspended in PBS 1× and incubated in a 37 • C water bath for 2 h, vortexing the mixture every 15 min for 2 min.Freeze dried ColMA was resuspended in 20 mM acetic acid and neutralized with 0.2 M NaOH just before use.Right before crosslinking, the photoinitatior lithium phenyl(2,4,6trimethylbenzoyl) phosphinate (LAP; Sigma-Aldrich: 900 889) was added to the gel mixture to achieve a working concentration of 1 mg ml −1 .The prepolymer solutions were sterilized before hydrogel fabrication.Stock solutions were always prepared the same day of the experiments.

Preparation of stiffness micropatterns
PDMS molds were sterilized by submerging them in 70% ethanol solution in MiliQ water for at least 1 h.Molds were dried at RT and placed inside a glassbottom 6-wellplates (MatTek).GelMA and HAMA physically gelate at RT. Therefore, 50 µl of hydrogel stock solution was pipetted inside the molds and was allowed to physically polymerize at RT for 15 min.Since ColMA polymerizes at 37 • C, ColMA hydrogels were allowed to polymerize for 30 min at 37 • C.After that, hydrogels were homogenously crosslinked to stabilize their structure (stabilization UV), using a UV-lamp (365 nm, Analytik Jena UVP XX-15L) with a total dose of 0.5 mJ mm −2 .Stabilized hydrogels were photopatterned using a high-precision and maskless UV illumination system projector, PRIMO ® (Alvéole).The optical setup, connected to a DMi8 inverted microscope (Leica), uses a UV laser diode (λ = 375 nm; corresponding to LAP's absorbance peak [36]) and a DMD to project user-defined light patterns onto a focused surface (hydrogel) with high spatial resolution (2 µm).To characterize the mechanics of different ECM hydrogel types under homogenous UV-exposure, hydrogels were illuminated without any pattern (full DMD) with a dose of 2 mJ mm −2 .The DMD also allows generation of UV-light exposed gradients, controlling gradually the UV transmission and therefore the degree of crosslinking.
High-resolution spatial distributions were designed in Adobe Illustrator (Adobe) and loaded into the PRIMO software (Leonardo, Alvéole).Stripes of 20, 50 and 200 µm in width were separated by 50 µm interline space.As a control, a full exposed hydrogel (full DMD: ∼1.9 × 1.1 mm) was created.The focal plane was focused on the middle of the gels.The hydrogels were exposed with a dosage of 2 mJ mm −2 .

Assessment of hydrogel crosslinking
To evaluate the resolution of the photocrosslinking, methacrylated rhodamine (RhoMA; Polysciences) was used.RhoMA is a fluorescent probe that will crosslink with the MA groups of the modified ECM components under exposure to UV light.That allows the visualization of the crosslinked pattern via fluorescent microscopy.The 200 µM RhoMA was incorporated in the prepolymer solution.Hydrogel fabrication, gelation and photopatterning was performed as described above.After exposure, hydrogels were thoroughly rinsed using PBS 1X to remove the excess of RhoMA.Fluorescence images were obtained using a Leica DMI8 epifluorescence microscope with an objective of 10× with NA of 0.3.Dimensions of the patterns and the fluorescence intensity profile were evaluated using ImageJ.For characterization of the micropattern stiffness, refer to the methods detailed in section 2.4.

Atomic force microscopy (AFM) imaging
To assess topography characteristics of the 2D system before and after projection-based photopatterning, we used an atomic force microscope (Bioscope Catalyst, Bruker, USA) in imaging contact mode.Surface of the hydrogel was scanned in liquid conditions and RT using a sharp pyramidal tip SNL-10 (Bruker) with a nominal tip radius of 2 nm.The quantification of the height and roughness of the surface was done with the NanoScope Analysis 1.5 software (Bruker).

2D stiffness micropatterns cell experiments
Two main experiments were conducted to explore the response of cFB on heterogenous stiffness distributions in the 2D system: (1) effect on cFB orientation produced by the spatial distribution of heterogenous stiffness micropattern anisotropies and (2) the effect of the introduction of the stiffness anisotropy at a determined time (T 0 ) on cFB orientation.In the first experiment, cFBs were seeded on top of the projection-based exposed hydrogels at a cell density of 1500 cells mm −2 .After 24 h (analysis time, T A ), the cells were fixed for immunostaining to quantify cellular orientation.In the second experiment, cFBs were initially seeded on stabilized hydrogels and allowed to spread for T 0 = 6 h.Following this, the hydrogels were projection-based exposed with a 20 µm spacing pattern.The cFBs were given time to adapt to the new mechanical environment and were fixed 24 h after light exposure.All groups and N are displayed in table S1.

System design and mold fabrication
In order to cast the hydrogel in the 3D system, an initial mold is 3D printed using gray polylactic acid (PLA) extruded with a conventional 3D printer (Ultimaker).The mold consists of three rectangular compartments arranged in parallel and separated by thin (0.5 mm) PLA walls (figure S1).The primary purpose of the middle compartment is to cast the hydrogel with embedded cFBs.The lateral compartments serve to add the GelMA-based hydrogel, allowing for the creation of a stiffness anisotropy throughout the entire system.The presence of the thin walls is crucial as they confine the hydrogel with embedded cells, preventing unwanted mixing with the GelMA-based hydrogel during the fabrication process.Additionally, the thin walls facilitate the controlled addition of the GelMA-based hydrogel at the desired timing (T 0 ).Two designs were created with different widths of 2 and 5 mm (with length of 12 mm) to create different spatial stiffness anisotropies.

Preparation of the stiffness macropattern
Similar to the 2D system, two main experiments were conducted: (1) effect on cFB orientation produced by the spatial distribution of heterogenous stiffness macropattern and (2) the effect of the timing (T 0 ) of the stiffness signal on cFB orientation.All experimental groups are listed in table S2.For the first experiment, cFBs were seeded inside a collagen I hydrogel (3 mg ml −1 ) in the middle compartment of the mold (2 or 5 mm width).cFBs were allowed to spread inside the hydrogel for 1 h (T 0 = 1 h) at 37 • C. To create the stiffness micropattern, the lateral compartments of the mold were filled with GelMA-based hydrogel composed by 10% (w/v) GelMA and 2 mg ml −1 collagen I (1:1 ratio in volume).After filling the lateral compartments, thin walls of the mold were cut with a scalpel to allow the GelMA-collagen hydrogel to be in contact with the cFBs-laden hydrogel in the middle.Directly afterwards, the system was UVilluminated with a UV-lamp of 5 mW cm −2 with a total dose of 6 mJ mm −2 to crosslink the GelMA.cFBladen hydrogels were cultured for 24 h and 72 h (analysis time, TA) post-UV illumination, after which they were fixed for further analysis.
The second experiment was designed to explore the effect of the initial cFBs spreading inside the hydrogel before the application of the stiffness macropatterns.The process of creation of the tissues is the same as described before but the stiffness anisotropies were created at 3 h, 6 h and overnight after casting the cFBs-laden hydrogel.
Finally, two more experiments were conducted to explore the versatility of the system (table S2) by (3) exploring the effect of stiffness anisotropy magnitude on cFB orientation.For this experiment 5% GelMA and 15% GelMA was used to create softer and stiffer anisotropies (with same illumination dose); and (4) exploring the effect of pharmacological manipulation of important mechanoresponsive element of the cells.For that, we added ROCK inhibitor as a mediator of cell contractility at a concentration of 15 µM during the whole culture period (24 h) and studied its effect on cell orientation.

Mechanical characterization of the stiffness patterns (nanoindentation)
We conducted the characterization of the elastic modulus (Young's modulus) of the stiffness microand macropatterns using nanoindentation with the Piuma Nano-indenter (Optics 11, Amsterdam, The Netherlands).To perform this analysis, we placed the 2D and 3D stiffness patterned systems on the nanoindenter stage, and force-indentation (F-δ) curves were acquired on the surface of the hydrogels with separation steps covering areas of 100 × 250 µm and 1 × 6 mm (or 15 mm for the 5 mm) for the 2D and 3D systems, respectively.The force-indentation curves were acquired using a spherical tip of 10 µm in diameter attached to a rectangular cantilever with an elastic constant, k, of 0.025 N m −1 .To compute for the Young's modulus (E) we fitted the forceindentation curves Hertz contact mechanical model for a sphere indenting a semi-infinite half-space [37]: where R is the sphere radius of the tip and ν the Poisson's ratio of the sample (assumed to be 0.5).The Young's modulus of the sample (E) was computed for each F-δ curve using the Optics 11 data analysis software.The fitting was performed for the loading curve.

Immunofluorescence staining
The 2D and 3D systems containing cFBs were washed with PBS three times before and after fixation and fixed in 3.7% formaldehyde (Merck, Darmstadt, Germany) for 15 min and 45 min at RT, respectively.Subsequently, the samples were washed three times with PBS 1× for 5 min.The cFBs loaded systems were permeabilized with 0.5% Triton-X-100 (Merck) in PBS for 10 min at RT and blocked for nonspecific antibody binding with 10% horse serum (Sigma-Aldrich) in PBS for 40 min.The cFBs in the 2D system were incubated 1.5 h with the primary antibody (anti-vinculin or anti-YAP) at a 1:200 dilution in PBS with 4% goat serum and 0.05% Tween at RT. Subsequently 2D samples were washed three times with PBS for 5 min before incubating with the secondary antibody in PBS and phalloidin for 1.5 h at RT.The samples were washed two times with PBS before incubating with DAPI in PBS for 5 min to visualize the cell nuclei.Finally, samples were washed four times with PBS for 5 min, and thereafter, the membranes were mounted to glass slides with Mowiol (Sigma-Aldrich) and stored protected from light at 4 • C. For the 3D system, samples were stained with phalloidin and DAPI only using the same protocol as the 2D system.Additionally, collagen fibers were stained by incubating the 3D samples with CNA35-OG488 probe [38] at 10% in PBS for 1 h at RT. Finally, both 2D and 3D tissues were washed three times with PBS for 5 min.All immunofluorescent samples were analyzed with a confocal fluorescence microscope (Leica SP5X) using the 10× and 20× objectives.

cFB and collagen orientation analysis
The orientation of the F-actin fibers (representing cFBs orientation) [39] and collagen was determined from triplicate of three independent experiments (n = 3) per condition (table S2).Immunofluorescent samples were analyzed with a confocal fluorescence microscope (Leica SP5X).The cFBs and collagen orientation were quantified with MATLAB using the in-house developed Fiber Orientation Analysis tool, where F-actin and collagen fibers are identified using Frangi vesselness [40].The angle distributions (h(θ)) obtained were further processed to calculate an order parameter to quantify for the degree of alignment [30,41,42]: S ranges from S = 1, all cFBs or collagen fibers oriented completely at 0 • or 180 • direction, to S = −1, all cFBs or collagen fibers oriented completely aligned at 90 • , with S = 0 representing random orientation (see figures 2 and 5 for angle definition).

Assessment of cFB viability
To assess the viability of cFB after UV-light exposure, cFBs cultures (photopatterned, 2D or cell-laden hydrogels, 3D) were carefully washed three times with PBS 1× for 5 min.Subsequently, samples were incubated with 1 µg ml −1 calcein AM and 750 nM propidium iodide (Invitrogen) for 30 min at 37 • C protected from light.After incubation, cFB cultures were washed three times with PBS 1× for 5 min and immediately imaged with a confocal microscope (Leica SP5, Leica, Mannheim, Germany) using the 20× objective.

qPCR
qPCR was performed as previously described [43].Briefly, total RNA from 1 × 10 5 cells or pooled 3 wells of 1 × 10 4 cells (totaling 3 × 10 4 cells) were isolated by the TRIzol RNA isolation protocol.QuantiTect RT kit (Qiagen) was then used to make cDNA from the RNA samples, following manufacturer's instructions.Relative gene expression was determined by quantitative real-time PCR (qRT-PCR) on a BioRad CFX384 real time system using ABsolute QPCR SYBR Green mix (Thermo Fisher Scientific).Gene expression was determined by correcting for reference gene values (GAPDH), and the calculated values were expressed relative to the control group per experiment by means of the delta-delta Ct method.For a list of used primers see table S3.

Statistical analysis
Data are displayed as mean ± standard error of the mean (SEM), based on three distinct experiments with three replicates for each condition, unless specified differently.For two-group comparisons, the Student's t-test with a 95% confidence interval was employed, following normality and equal variance assessments using the Shapiro-Wilk and Brown-Forsythe tests, respectively.In the case of multiple group comparisons, a one-way ANOVA complemented by Tukey's post-hoc test was used after the Shapiro-Wilk and Brown-Forsythe tests confirmed normality and equal variances.To determine the combined influence of T 0 and cell density on cFBs and ECM orientation, a two-way ANOVA was applied.This was after heteroscedasticity and normality of residuals were verified using Spearman's and Anderson-Darling tests, respectively.For all comparisons, p < 0.05 was considered statistically significant.

Projection-based illumination increases hydrogel stiffness
To test the capabilities of the UV illumination projector system to crosslink methacryolyl modified natural hydrogels and choose the most suitable hydrogel to obtain stiffness cues, we illuminated ColMA, HAMA and GelMA hydrogels with different polymer concentrations.We observed that the stiffness of the hydrogels upon crosslinking stabilization was dependent on the hydrogel type and concentrations (figure S2(A)).The projection-mediated illumination could successfully increase the stiffness of each hydrogel group.In particular, after projection-mediated UV illumination, all hydrogels showed a ∼2.5 to ∼4-fold significant increase in their Young's modulus (figure S2(A)).Overall, GelMA 10% hydrogels showed the highest stiffness difference between the non-and UV-illuminated samples.The versatility to tune the mechanical properties of this hydrogel type, once combined with the projection-mediated photopolymerization, was also demonstrated by generating a stiffness gradient of GelMA 10% with a slope of ∼4kPa mm −1 (figure S2(B)).Given these observations, the GelMA 10% was chosen for the subsequent development of stiffness cues in 2D and 3D.

Creation of 2D linear stiffness micropatterns with cell-size resolution
Next, we aimed at exploiting the stiffness versatility exhibited by the projection-mediated crosslinked GelMA 10% to expose cells to stiffness patterns with high spatial resolution, to control cell orientation.Previous studies have shown that cells orient in response to topographical or ligand-based linear micropatterns [11,12,44,45].Inspired by those studies, we fabricated linear micropatterns of alternating stiffness.In particular, we exposed the hydrogels to UV patterns using the projection-mediated illumination (20 µm, 50 µm and 200 µm) separated by 50 µm non-illuminated areas (figures 1(A) and (B)).To verify the success of the technique in eliciting patterns of crosslinked GelMA, these hydrogels were mixed with RhoMA (figure 1(C)).Since RhoMA crosslinks with GelMA with the presence of photoinitiator and UV light, the observed linear patterns of higher RhoMA intensity confirmed the formation of crosslinked patterns in response to the UV treatment, with the size of the patterns well corresponding with the width of the illuminated areas (figures 1(C) and (D)).
Nanoindentation of the hydrogel surfaces also confirmed positive correlation between the observed crosslinking patterns and the stiffness patterns (figure 1(E)).In particular, the Young's modulus measurements showed a pattern very similar to the crosslinking intensity curves, demonstrating that a higher crosslinking degree with projection-mediated UV illumination corresponds to a higher stiffness.The values of the soft regions were ∼1 kPa while the stiff areas corresponded to ∼4-5 kPa, in agreement with the ∼4-fold increase obtained with homogeneous illumination (figure S2(A)).To further evaluate the agreement between the crosslinking and stiffness profiles, we performed a comparison of the width and distance of the two kinds of patterns, which showed a clear agreement (figures 1(F) and (G)).No significant difference was observed between the two groups, except for the pattern separation in the case of 50 µm (figure 1(G)).The width of the patterns was slightly larger than the designed theoretical values (figure 1(F)), although still in the same order of magnitude, probably due to light scattering inside the hydrogels.This increased pattern width was counterbalanced by a decreased pattern separation compared to the theoretical value of 50 µm (figure 1(G)).This highlights a practical lower limit for the spatial definition of the technique, as also shown via a test on a smaller theoretical pattern separation of 25 µm (figure S3), which exhibited a ∼50% decrease in amplitude of the crosslinking pattern values.
Another common limitation of photocrosslinking of hydrogels is caused by the surface topographical modifications, which can derive, for example, from swelling [46], and may topographically influence cell orientation next to stiffness cues [44].Hence, topographical measurements were conducted on photopatterned hydrogels 24 h postillumination, which were maintained in PBS at 37 • C throughout this period.The difference in height between the crosslinked and non-crosslinked areas was ∼4 nm, with a nano-roughness difference of ∼1.5 nm (figure S4), which excludes topography from the possible cues affecting cell orientation in our system [47].As such, our measurements confirmed the establishment of a biofabrication technique leveraging on photo-sensitive crosslinking hydrogels to obtain linear stiffness cues with cell size resolution.

2D stiffness micropatterns steer cFBs orientation
Given the previously demonstrated influence of topographical and ligand-based linear patterns on cell orientation [8][9][10][11], we hypothesized that our linear stiffness patterns could elicit an analogous orientation response.We decided to test our hypothesis using cFBs, which is the most important cell type that shapes the structure and gives support to the myocardial tissue [48].cFBs were seeded on top of the (differently sized) patterned substrates.After 24 h of culture (T A = 24 h), the cFB cytoskeleton showed a preferred directionality along the direction of the stiffness patterns (90 • , figures 2(A) and (B)) for line widths of 20 and 50 µm.The quantification of the order parameter (S), based on the angle direction of the actin cell fibers, confirmed that such response depends on the width of the lines (figure 2(C)).In particular, the S value was closest to −1 for the 20 µm pattern size, indicating a strong preferred directionality towards the pattern direction (90 • ).The cFB seeded on top of the 200 µm did not show a preferred directionality, similar to the non-patterned control and the homogenously illuminated sample, while the 50 µm samples exhibited an in-between response (figures 2(C) and S5).To further test the potential of our 2D system, while also corroborating the hypothesis that cFB orientation arises from mechanosensing, we quantified the YAP nuclei-tocytoplasm ratio present on the non-patterned substrate and compared it with the value present on the 20 µm-patterned samples, as cFBs showed the strongest alignment in this condition.The ratio was significantly higher in cells cultured on the patterned hydrogels (figure 2(D)), supporting our conclusion that cFBs are mechanosensitive to and orient in response to linear stiffness patterns, in a line widthdependent fashion.

Temporal control of 2D stiffness micropatterns and cFBs orientation
One of the main consequences of myocardial infarction is the progressive loss of the characteristic anisotropy of the cardiac tissue organization [7].The signals stimulating cells to switch from an isotropic organization back to an anisotropic still need to be better understood.Therefore, we tested whether our system can be used in this context by imposing stiffness cues on an initially isotropic cell culture.cFBs were seeded on top of homogeneous GelMA hydrogels directly after stabilization crosslinking and were allowed to attach for T 0 = 6 h to obtain an initially isotropic cell organization (figure 3(A)).Thereafter, these systems were illuminated with the projection system to elicit 20 × 50 µm stiffness patterns.cFBs were allowed to adapt to the change of stiffness environment for additional T A = 24 h before analysis (total culture time of 30 h).The stiffness profile of the micropatterned hydrogel at 6 h (figure 3(B)) was similar to the one obtained with illumination at 0 h (figure 1(E)), with a range of stiffness going from ∼1.5 to ∼4 kPa (figure 3(B)).In agreement with that, 24 h after exposure to the stiffness patterns, cFBs showed a clear orientation towards the direction of the stiffness micropatterns (figures 3(C) and (D)), similar to the cFBs seeded on pre-patterned hydrogels (figure 2).Moreover, viability assays confirmed that the cells retained their viability post-illumination (figure S6).Therefore, our system allows for the investigation of cell (re-)orientation in response to 2D linear stiffness cues not only with spatial, but also temporal resolution.

Generation of stiffness macropatterns in 3D cell cultures
Since cells in vivo are surrounded by 3D ECM, whose architecture is influenced and dynamically changed by the cells in an inter-dependent fashion [49], we next aimed at developing an analogous system allowing for a similar spatiotemporal control of cell and ECM orientation in 3D.In particular, knowing that uniaxially constrained static tissues can provide stiffness cues aligning cells and ECM [29,30,42,50], we leveraged the crosslinking properties of GelMA to develop uniaxially constrained and temporally dynamic 3D tissues.The obtained 3D in vitro system is based on cFBs embedded in collagen I gels that are uniaxially constrained by GelMA-collagen photosensitive hydrogels (figures 4(A) and (B)).The bio-fabrication strategy is summarized as follows: (i) collagen I hydrogel is mixed with cFBs and casted in a 3D printed mold (figures 4(A) and (B), first column); (ii) cFBs then spread and grow inside the collagen gel for a time (T 0 ); (iii) finally the GelMA-ColI hydrogels are added on both sides of the cFBs laden collagen hydrogel, and illuminated with UV-light to crosslink the GelMA (figures 4(A) and (B), second column), thereby increasing their stiffness and providing uniaxial stiffness constrains to the adjacent collagen hydrogel (figures 4(A) and (B), third column).Notably, this bio-fabrication strategy provided a handle on the time T 0 , chosen for initiation of the stiffness cue provided by the stiffening of the GelMA.
We first characterized the resulting stiffness in the GelMA, the collagen gel, and their interface region.The stiffness of the GelMA constraint was 19.9 ± 1.2 kPa, while the stiffness of the collagen I gel was 3.9 ± 1.1 kPa (figure 4(C)).Accordingly, the longitudinal stiffness profile of the entire system showed a stiff area where the GelMA-ColI was crosslinked, next to a soft area where the collagen I hydrogel was present (figure 4(D)).The interface area presented a stiffness gradient with a slope of 11.8 ± 3.6 kPa mm −1 .Nanoindentation data showed that incubation time of the system (24 h versus 72 h at 37 • C) does not have a major impact on the stiffness profile: no statistical difference was observed, although a decreasing trend is present for the stiffness of the GelMA region.Therefore, our bio-fabrication strategy successfully developed a collagenous construct where cells are exposed to stiffness anisotropy that can be imposed at a specific timepoint.
Changing the size of the casts also enabled changing the stiffness profile in space, as demonstrated by analyzing analogous systems created with 5 × 12 mm casts (figure S7).The stiffness profile of this differently sized in vitro system resulted similar to the original counterpart: the GelMA-ColI crosslinked area was 22.9 ± 0.7 kPa; the collagen hydrogel area was 1.9 ± 0.6 kPa; and the gradient in the transition zone was 8.1 ± 1.1 kPa mm −1 (figure S7).As a result of the larger size of this system, combined with the similar slope of the stiffness gradient, a much larger central soft area was observed compared to the 2 × 12 mm systems.This mechanical characterization thus confirms that our strategy can be employed to obtain uniaxially constrained cell-populated 3D systems with different temporal and spatial features.

3D stiffness macropatterns of different spatial dimensions to steer cFB orientation
Cells embedded in uniaxially constrained ECM constructs usually align along the constrained direction due to stiffness anisotropy [26,28].Furthermore, cells can align in response to stiffness gradients [51,52].However, other studies show that cells also have a length limit for stiffness mechanosensitivity [53,54].Therefore, we hypothesized that cFBs will align in response to our stiffness anisotropy, with this response affected by the width of the construct.
The PLA frame primarily serves to delineate the hydrogel's geometry.Our findings, supported by brightfield microscopy images, show no significant attachment between the hydrogel and the PLA frame, indicating that the mechanical environment of the cells is unaffected by the PLA.In addition, subsequent addition of GelMA results in a fusion with the collagen hydrogel indicating that GelMA rather than PLA plays a significant role in providing the mechanical cues (figure S8).
To test our hypothesis, we measured the cFBs orientation seeded in 2 mm and 5 mm 3D systems UVtreated after T 0 = 1 h (figure 5(A)).As expected, after T A = 24 h, 3D-cFBs at the center of the 2 mm construct exhibited alignment in the constrained direction (figure 5).Cells in the 5 mm counterpart were analogously oriented, but to a significantly lower extent (figure 5), thereby confirming that the size of the constructs can affect cell responses.Inspired by the impact of the constraint distance on cell behavior, we checked cell orientation in the interface regions (figures 5(B) and (C)).Surprisingly, while the cFBs populating the 2 mm constructs still showed alignment in the constrained direction, the 5 mm counterparts were more randomly oriented, exhibiting alignment heterogeneity previously observed also in other studies [50,55].Interestingly, cFBs showed their alignment mainly restricted to the XY planes (planes of stiffness anisotropies) (figure S9).Taken together, our 3D system can expose cells to different profiles of stiffness throughout collagenous gels, which induce a profile-dependent and spatially heterogeneous cFBs alignment.The versatility in size also allows for future investigation of the length limit at which cells can sense the stiffness cue provided by uniaxial constraints.

cFB orientation in response to 3D stiffness macropatterns is determined by cell density and T 0
Having confirmed the potential of our newly developed system to align cFBs, we further tested its versatility in modifying other parameters relevant for cell (re)orientation.The response of cells within 3D collagenous constructs has been shown to depend on both the timing of the mechanical stimulus and the density of cells [29].The different responses seem to arise from different cell-cell and cell-ECM interactions [56].While these previous studies focused on varying the timing of cyclic strain initiation [29], our system can be exploited to vary the start of the stiffness anisotropy stimulus instead (T 0 = 1 h, 3 h, ON/6 h).T 0 tunability allows the cells to initially adapt and remodel their surroundings before the stiffness stimulus is applied.To further investigate the capability of the stiffness macropatterns to direct cFBs organization in 3D, we thus studied two fundamental parameters: cell density, and the stiffness cue initiation time T 0 .Information over the temporal dynamics of the cFBs orientation response was obtained by analyzing different timepoints upon UV-illumination (T A = 0 h, 24 h, 72 h) (figure 6(A)).The orientation of cells was quantified by the order parameter S, which was derived from the angular distribution histograms, h(θ) (figure S10).Furthermore, cFB viability was evaluated following UV exposure, demonstrating high cell viability immediately after illumination and at 24 h (figure S11).
The alignment of cFBs depended on both cell density and the timing of the constraints, T 0 .Applying the stimulus relatively early (T 0 = 1 h) was found to be the most effective strategy to influence cell orientation, progressively directing the initially isotropic cell organization (figure 5(B)) towards an anisotropic one at 24 h and 72 h (figures 6(C), (D) and S12), with little influence from the cell density.An analogous response could be observed for the T 0 = 3 h samples, but only in the case of high cell density (1 M cells ml −1 ); in fact, for relatively low cell density (100 k cells ml −1 ), the delayed temporal stimulus was not able to reorient the cells, suggesting a possible involvement of cell-ECM interactions, perhaps provided by contact guidance or steric hindrance [56].Finally, surprisingly, the T 0 = ON/6 h samples exhibited an initial orientation perpendicular to the constrained direction, independent from cell density.This initial orientation was successfully perturbed by the initiation of the stiffness cue, which led cFBs to attain a finally anisotropic organization parallel to the constrained direction (figures 6(C), (D) and S12).The 5 mm systems showed a behavior similar to the 2 mm ones, although with a lower extent, with cFBs showing an initial (T A = 0 h) random alignment at any T 0 (1 h, 3 h), which was then directed parallel to the stiffness anisotropy (figures S13 and S14).Summarizing, we observed that the stiffness cues are able to progressively direct cFBs towards the constrained direction, with a temporal dynamic profile that is dependent on both the cell density and the timing of the stimulus initiation.

Stiffness magnitude and pharmacological manipulation of the 3D in vitro system
Compared to most of previous 3D systems, our strategy also allows for a relatively easy modification of the impact of the mechanical constraint.In vitro systems are also employed to expose cells to pharmacological treatments [29,30].To appreciate the possible effects that these modifications might have on the cell alignment, we perturbed the conditions that exhibited maximum (1 M cells ml −1 , T 0 = 3 h, T A = 24 h) and minimum (100 k cells ml −1 , T 0 = 3 h, T A = 24 h) alignment (figure 5).Reducing the stiffness of the GelMA by 75%, corresponding to a decrease in stiffness to ∼5-7 kPa, caused a loss of cell alignment for the high cell density samples.This confirms the importance of the stiffness cue for cell alignment in our system.Nevertheless, increasing the stiffness of the low cell density samples to 40 kPa did not elicit a major effect on cells, which still exhibited a random orientation.This suggests that cell-to-cell distance might play a role in cell (re)organization, possibly mediated by cell contractility [26].Therefore, to assess the effect of pharmacological treatment on the system, we explored the use of ROCK inhibitor (added when constraints were created at T 0 = 3 h), directly related to cell contractility, actin dynamics, and cell (re)orientation.In agreement with a previous  study [29], ROCK treatment caused a significant decrease in the mechano-responsive cell alignment (figure 7(D)).Overall, these perturbations showcase the versatility of the designed 3D system, which can be employed to explore different mechanisms influencing the (re)orientation response of cells to stiffness cues.

3D in vitro system allows for the study of ECM remodeling
The dependency of cell orientation on T 0 (figure 6) suggests a possible cell-ECM interaction interfering with the cell (re)orientation response to stiffness.To validate such interaction and further test the versatility of the 3D system, we imaged collagen I in the tissues that showed consistent alignment response (1 M cells ml −1 , 2 mm distance, figure 6(D)).Overall, collagen matrix exhibited a trend in orientation similar to cFBs, although to a much lower extent.At T A = 0 h, the ECM showed an organization independent of T 0 , with the orientation slightly perpendicular to the constrained direction (figures 8(A) and (B)).Over time, similar to what happened for cFBs, the initiation of the stiffness constraint induced the ECM to align parallel to the constrained direction (figures 8(A), (B) and 6).However, such response was much lower and slower compared to cFBs, with an isotropic organization observed at T A = 24 h for T 0 = 1 h and 3 h.This delayed response of the ECM suggests that, rather than being orientationally directed by the ECM, cFBs responded first to the stiffness pattern and subsequently aligned and remodeled the surrounding ECM.These results were complemented by longitudinal qPCR analysis (figure 8(C)), showing that cFBs gene expression of ECM proteins (collagen I and fibronectin) decrease with cultured time, while the expression of ECM degradation related genes (MMP9, TIMP1) increase with time.Thus, cFBs most likely remodel the matrix via degradationmediated mechanisms.Furthermore, we thus showed that our 3D system also allows for the imaging and quantification of the ECM network organization, with the aim to gain fundamental understanding on the mechanisms and causes of the stiffness guided tissues on ECM remodeling.

Discussion
To investigate the effect of stiffness anisotropies on cFBs (and their collagen production) orientation, we developed systems with tuneable spatiotemporal mechanical properties.Photosensitive hydrogel GelMA was selected for its versatility, mechanical tunability, and widespread use in the field.GelMA was combined with either projection-mediated UV illumination or cast in a 3D-printed mold to expose cFBs to 2D and 3D stiffness cues at different length scales, with high spatial and temporal resolution.These techniques enabled the development of in vitro systems with stiffness cues at the micron (cellular) and mm (tissue) scale, to systematically investigate, understand, and ultimately guide the dynamic (re)orientation of cFB in response to linear stiffness patterns in 2D or temporally varying constraints in 3D.The developed 2D system enabled the development of linear stiffness cues on flat surfaces, without insignificant topographical changes (figure 1).To develop this 2D system, we employed a technique based on a UV-laser, filtered by a DMD (projectionmediated illumination) that allowed for the precise positioning and timing of the light exposure in the GelMA hydrogel.Most of the previous studies that create heterogeneous stiffness patterns and anisotropies used photomasks to filter the light from a lamp source [57].Here, using a DMD avoids the complex and time-consuming process of fabricating physical photomasks.This novel technique allowed for the generation of stiffness lines in the GelMA hydrogel (figures 1, S3) as well as the creation of stiffness gradients (figure S2) [21,58,59].Moreover, here the stiffness lines can be created at different timepoints, which enabled the observation of the transition from isotropically organized cell populations to collectively aligned cells (figure 2).We demonstrate that a relatively low amount of UV dose was sufficient to generate ∼4-5-fold changes in stiffness (figures 1, S2), accompanied by negligible changes of the hydrogel topographical surface properties (figure S4).This constitutes a major advantage compared to previous approaches [46], since it enables the isolation of cell responses to stiffness cues, unaccompanied by topographical cues that might interfere with the cell response [44].
Our data demonstrate that cFBs sense and respond to 2D linear stiffness cues by aligning in the direction of the micropatterns, specially at 20 µm width, irrespective of when the stimulus is applied (figures 2 and 3).Other studies have developed heterogenous stiffness environment to influence and study cell behavior to understand cell-ECM interactions in complex 2D mechanical environments [15,46,60].To the best of our knowledge, for the first time, we show here that such stimuli can influence cell directionality over time (figure 2).Our YAP/TAZ analysis suggests that such alignment is a mechanoresponse to stiffness, since cFBs respond to the stiffness patterned environment by translocating YAP/TAZ transcription factor into the nucleus (figure 2).This orientation response seems similar to the contact guidance phenomenon, where cells interact with the underlying topographies or patterned ECM through receptors like integrins, aligning their cytoskeleton in the direction of the linear patterns [11].Similar to previous studies on contact guidance, we also observed a dependency of cell orientation on the size of the linear cues.Therefore, perhaps, similar mechanisms of action might play a role in determining cell orientation in response to contact guidance and linear stiffness cues.Overall, our 2D system allows for future investigation of the underlying mechanisms involved in temporally controlling cell directionality by stiffness cues, which holds promise investigating strategies aimed at restoring anisotropic tissue organization that are lost, for example, upon myocardial infarction.
The effect of the cellular microenvironment on cellular behavior can largely differ from 2D to 3D settings, where cells are surrounded by the ECM network.Most previous techniques to generate stiffness anisotropy in 3D do not allow for the temporal control of the stiffness cue because they depend on complex microfabricated systems such as polymeric posts [26,[28][29][30][31][61][62][63].One of the few studies that varied the stiffness cue over time was the study by Foolen et al [27], wherein they modified the posts constraints by detaching the tissue from the posts.Because of this detachment, there was a notable increase in tissue compaction, leading to a subsequent rise in cell and ECM density at these areas.Such an increase could potentially act as a confounding factor when assessing cell and ECM orientation [64].Here, we were able to overcome these limitations by constraining collagenous gels with surrounding photo-sensitive hydrogels that were casted inside different compartments of a 3D printed mold.The obtained 3D collagenous gels presented stiffness profiles that could be varied over space (figures 4 and 5), time (figure 6), and magnitude (figure 7).The developed 3D system offers a valuable tool to independently investigate the impact of these parameters on cellular stiffness guidance.The developed 3D system was able to transiently align cFBs and collagen fibers in response to the stiffness constraint.Initially, the collagenous hydrogel mixed with cFBs was cast allowing the cells to spread for a pre-stimulus time, T 0 (figure 6).Independently of the construct size studied, cFBs showed an isotropic organization for shorter T 0 and tended to align perpendicular to the constrained direction for longer T 0 .Interestingly, collagen presented a slightly perpendicular organization also for shorter T 0 .After applying the stiffness cue, cFBs aligned parallel to the stiffness anisotropy in a cell density dependent manner (figures 6, S10, S12 and S13).Importantly, cFBs predominantly aligned within the planes exhibiting stiffness anisotropy (figure S9).This observation suggests that the 3D system induces stiffness anisotropies primarily within 2D planes, akin to observations in previous models.This underscores the relevance of 3D models where cellular alignment is confined to multiple planes, highlighting the nuanced interplay between 3D structural complexity and directional cellular responses.
To further explore the effect of the stiffness cue on collective cellular alignment, we changed the stiffness magnitude of the constraints by varying the GelMA hydrogel concentration (figure 7).cFBs were sensitive to the magnitudes in a cell density dependent manner.These findings suggest a collective effect of cells within the tissue, whereby cell contraction through the ECM contributes to the stiffness mechanosensing.Cells exert forces on their surrounding ECM, creating a stiffening of the ECM network, which can be sensed by neighboring cells in close proximity [65].To validate the involvement of cell contractility in cellular alignment, we inhibited the ROCK pathway.ROCK inhibitor is a potent inhibitor of actomyosin complex, described to be essential for numerous cellular functions such as cell contractility [66].Our results showed that tissues treated with the ROCK inhibitor did not exhibit a preferred directionality.Similarly, Foolen et al showed that the inhibiting the ROCK pathway, fibroblasts were not mechanically responsive to either cyclic stretch or static constraints [29].Cell contractility did not only align cells, but also the ECM.Our results show that after applying the stiffness cue, cells align parallel to the constrained direction.Interestingly, the ECM data suggests that ECM has a delay in alignment compared to cells.Gene expression analysis showed that cFBs are degrading the ECM and not synthetizing new one.Therefore, the newly developed 3D system introduces an innovative approach that facilitates the manipulation of stiffness cues in space and time.This enables a comprehensive exploration of the impact of stiffness magnitude, cell density, and pharmacological treatments on cell-ECM alignment dynamics.
Finally, we acknowledge the limitations associated with our methodologies, which are pivotal for guiding future research directions.The use of UV light, while instrumental for modulating mechanical properties with precision, raises concerns about its impact on cell viability, particularly for other cell types.This necessitates further investigation into photocrosslinkers that operate in the visible light spectrum and the development of compatible projection systems [67,68].Additionally, our innovative 3D modeling approach is constrained by a spatial resolution limit of 2 mm, limiting our ability to explore cellular responses to microscale stiffness gradients.Advancements in bioprinting technologies that offer enhanced resolution and precision are essential for overcoming this limitation [69].Moreover, our models currently only allow for unidirectional mechanical stimulation, which does not fully replicate the dynamic mechanical environments cells encounter in vivo.Developing systems capable of providing multidirectional stimuli, mimicking the physiological and pathophysiological conditions cells face, is crucial for advancing our understanding of mechanobiology and improving the fidelity of our models to realworld biological systems [70].

Conclusions
In conclusion, we developed in vitro model systems to manipulate and investigate the effect of heterogenous stiffness anisotropies at different dimensions, length scales, and times.The systems allow to test the effect of pharmacological treatments followed by standard cellular and ECM analysis, including (real-time) microscopy, immunostainings and qPCR.Overall, the 2D and 3D systems developed in this study hold great potential to gain important understanding of the mechanical basis of cellular alignment and could help design mechanotherapies for restoring cellular and ECM alignment to improve tissue functionality.

Figure 1 .
Figure 1.2D in vitro system to generate cell-size resolution stiffness micropatterns.(A) Schematic of the setup: GelMA hydrogel is prepared and crosslinked with a projection-mediated UV-light system with high spatial resolution, upon which cFBs are seeded on top (B) Phase contrast image of the GelMA gel with cultured cells on top of the stiffness pattern.(C) RhoMA fluorescent image.High intensity values indicate a higher crosslinking degree.Scale bar = 100 µm.(D) Fluorescence profile of the different illumination patterns (20 µm, 50 µm and 200 µm; separated 50 µm) (n = 3).(E) Stiffness profile of the different illumination patterns (n = 3).(F) Pattern width measured with the RhoMA intensity values from C (shaded color) and stiffness values from D (dark color).Dashed lines indicate the theoretical value of the pattern width corresponding to the different sizes of 20 µm, 50 µm and 200 µm.(G) Pattern separation measured with the RhoMA intensity values from (C) (shaded color) and stiffness values from (D) (dark color).The dashed line corresponds to the 50 µm theoretical separation value.

Figure 2 .
Figure 2. cFBs orient in the direction of the 2D stiffness micropatterns.(A) Immunofluorescence images of the cFBs on the different stiffness patterns (Blue: DAPI; green: actin; red: vinculin).Scale bar: 50 µm (B) representative polar plots of the cFBs on the stiffness patterns, showing a higher degree of anisotropy for narrower patterns.(C) Quantification of the cFBs orientation by the order parameter (S) (n = 3, 3 analyzed areas per n).(D) cFBs respond to the mechanical cue (left: no pattern; middle: 20 µm pattern) and translocate the mechanosensitive transcription factor YAP to the nucleus on the patterned substrates (n = 5, 3 analyzed areas per n).* shows statistical significance against 'No pattern' condition.Scale bar = 100 µm.

Figure 3 .
Figure 3. Temporal control of 2D stiffness micropatterns.(A) cFBs are seeded on top of the GelMA hydrogel for 6 h and afterwards projection-mediated UV illumination applied to crosslink the hydrogel.cFBs are left on the substrate for 24 h before fixation.(B) Stiffness profile for a 20 µm pattern separated 50 µm (n = 3).(C) Immunostaining image of cFB on top of the stiffness pattern (Blue: DAPI; Green: actin; Red: Vinculin) with radar plot of cFB actin directionality.Scale bar = 100 µm.(D) Order parameter quantification (n = 3, 3 analyzed areas per n).

Figure 4 .
Figure 4. 3D in vitro system to investigate cell and ECM orientation with stiffness anisotropies.(A) Schematics of the system: an initial collagen gel with cFBs inside is created.cFBs are allowed to spread in the collagen I hydrogel for a certain amount of time (T0: 1 h, 3 h, 6 h or ON) and afterwards a GelMA gel is added and UV-illuminated to create the stiffness pattern.cFBs and collagen orientation is analyzed at TA = 24 h and 72 h after the creation of the stiffness pattern.* * * indicates p < 0.001 between conditions.(B) Photos of the 2 × 12 mm system created to culture these gels.Scale bar = 5 mm.(C) Young's modulus of the constraint vs the cell-laden collagen hydrogel.(D) Young's modulus profile of the 2 mm construct (n = 3).(E) Effect of incubation time in the Young's modulus of the system (n = 3).

Figure 5 .
Figure 5. cFBs orient in 3D in the direction of the stiffness anisotropy 24 h after stiffness patterns are applied.(A) Schematics of the 3D system.(B) F-actin images of the cFBs on the different regions of the gel (interphase, close to the stiff region, and the center, far from the stiff region).Scale bar = 200 µm.(B) Polar plots of the cell alignment for the different regions.(C) Order parameter quantification of the cFBs alignment (n = 3, 3 analyzed areas per n).* , * * * means statistical significance compared to 2 mm.$ means statistical significance between the center and interphase regions.

Figure 6 .
Figure 6.Temporal control of the stiffness pattern in the 3D in vitro system (A) Schematics of the experimental design.The 2 mm hydrogels were incubated for a T0 time period to allow for cFB spreading.Subsequently, GelMA hydrogels were added surrounding the cFB-laden collagen hydrogel to create the stiffness constraint.Next, the cells at the center of the hydrogels were imaged and analyzed at 24 and 72 h, while adapting to their new mechanical environment.(B) Actin images of the cFBs spreading before adding the stiffness pattern (TA = 0 h).Order parameter, S, to quantify for cFBs orientation (n = 3, 3 analyzed areas per n).Scale bar = 200 µm.(C) Actin images of the cFBs spreading before adding the stiffness pattern (TA = 24 h).Order parameter, S, to quantify for cFBs alignment (n = 3, 3 analyzed areas per n).(D) Actin images of the cFBs spreading before adding the stiffness pattern (TA = 72 h).Order parameter, S, to quantify for cFBs alignment (n = 3, 3 analyzed areas per n).* , * * , * * * means statistical significance (p < 0.05, 0.01 and 0.001, respectively) compared to the T0 = 1 h group.

Figure 7 .
Figure 7. 3D in vitro system allow for the investigation of stiffness magnitude and pharmacological treatments.(A) 3D in vitro system allows for the manipulation of the stiffness micropattern magnitude.(B) Order parameter, S, of high and low cFB density tissues with different stiffness anisotropy magnitude (n = 3, 3 analyzed areas per n).(D) Actin images of control and cFBs treated with ROCK inhibitor with representative polar plots.Order parameter, S, to quantify cellular alignment (n = 3, 3 analyzed areas per n).Scale bar = 200 µm.* means the statistical significance of the ROCK inhibitor group compared to the control.

Table 1 .
Hydrogel ECM source and concentration used.