*3.4. The Role of Tension in EFM Sca*ff*olds*

To further explore the role of internal tension, suspended EFMs were modified to reduce internal residual tension by electrospinning onto non-stick aluminum foil, cutting the fibers and removing them from the foil, and fixing free-standing EFMs to PDMS annular rings with an inner gap diameter of 10 mm (Figure 1C). These "tension-released" EFMs produced a statistically significant increase of 9.7% in Feret diameter compared to suspended EFM without released tension at the same gap diameter (Figure 7A). In addition, increasing EFM thickness led to lower Feret diameter (Figure 7B). This is the same trend observed for suspended EFMs. In this case, we have empirically reduced the stiffness of the EFMs by releasing internal tension, yet EFM thickness remains inversely correlated with Feret diameter in contrast to results on supported EFMs (Figure 3B). Taken together, this data suggests that the mechanical environment of EFM scaffolds is complex and that cells may react to this environment in different ways on depending on the stiffness and presentation of the support.

**Figure 7.** Average cell Feret diameter observed on tension-released and suspended EFM constructs, pooled data across all EFM thicknesses (**A**) and between tension-released mats at different fiber mat thicknesses (**B**). A total of 1651 cells were analyzed across at least two independent experiments. Levels connected by a star (\*) are statistically significant from each other (*p* < 0.05).

#### **4. Discussion**

This work explores the impact of EFM mechanical parameters, beyond Young's modulus, on cell morphology as measured by the Feret diameter. In particular, the effect of the mechanics of the support, the presence or absence of the support, inherent tension in EFMs and EFM thickness were explored. For supported EFMs, changing the stiffness of the support material underlying the EFM caused a large change in cell morphology, but only if the support was much stiffer than the EFM (Figure 4A). This is not because cells could not sense differences in stiffness between the underlying substrates; indeed, Feret diameter was different when grown directly on softer flat PDMS supports without an EFM (Figure 4B). However, when the much stiffer EFM was added, these differences were abrogated. Although this could result from topography differences, EFMs are fibrous whereas PDMS supports are relatively smooth and flat, others have shown that cells can "feel" the underlying substrate when cultured on thin hydrogels. Experiments in which cells spread more on thin gels than on thicker gels demonstrate this interplay of support substrates [33].

We are the first to demonstrate such edge effects in EFMs. Buxboim et al. described a "threshold matrix thickness," the length scale at which cells respond not only to the stiffness of the matrix, but also to the rigidity of the underlying support for hydrogel models [21]. However, these length scales were on the order of 10–20 μm compared to our observed responses on 50–200 μm thick EFMs. Cells may be able to "feel" further when grown on EFMs than when grown on hydrogels because of the orders of magnitude difference in length scales of individual fibers. Electrospun fiber length can range upwards of 35–50 cm [34], whereas collagen fibers in hydrogels are only 0.5–3 μm long [35]. Other factors aside (e.g., fiber strength, fiber interconnectivity, etc.), this difference in length scale may lead to differences in how well the matrix transmits tension. This likely permits stress to be transferred over a larger distance.

Suspended EFMs produced perhaps the most intriguing results. From a design perspective, the 50 μm thick, 3 mm gap EFM scaffolds yielded a key finding; these substrates elicited the highest mean Feret diameter of all (Figure 5). Cells on a thin EFM suspended over a small gap spread more than cells on a thin EFM fixed to a rigid support with a higher indentation modulus (Table 1). In general, cells on stiffer substrates demonstrate higher contractility and spreading [36], i.e., Feret diameter. Engler et al. demonstrated cells grown on hydrogels show a bell curve-shaped response to ligand density, though ligand density and substrate stiffness are highly difficult to decouple in gels [37]. We have previously shown that this bell curve-shaped response also exists in EFMs for cell mechanics when ligand density and stiffness were decoupled through the use of core-shell electrospinning [19]. Here, we have extended upon that work by decoupling material properties from more complex composite and structural mechanics by suspending fiber mats across gaps of different diameter without changing surface ligand density or Young's modulus. We also interrogated focal adhesion kinase (FAK) expression through Western blotting and found no significant differences in expression (Supplementary Figure S3). Although FAK phosphorylation could play a role, this suggests that this mechanosensing mechanism may not be mediated by classical focal adhesion kinase machinery, which we have previously implicated in EFM mechanotransduction [19]. These data help to illustrate the complex interplay of substrate mechanics, cell adhesion molecules, and mechanosensing.

For suspended EFMs, changing gap diameter caused significant changes in cell morphology (Figure 6). Releasing the internal tension in EFMs also altered mean Feret diameter (Figure 7A). These results highlight the interplay of substrate mechanics and mechanosensing beyond Young's modulus. It is generally believed that cells spread more on stiffer substrates [36]. Our investigation into the effect of support material stiffness (Figure 4) generally corroborates this belief; Feret diameter correlated positively with stiffness, as measured by Young's modulus. However, these results were inverted with EFM thickness for both supported and suspended EFM presentations (Figures 3B and 6). Feret diameter inversely correlated with stiffness, as measured by indentation modulus. In supported and suspended EFMs, increasing EFM thickness had opposite effects on indentation modulus. Increasing supported EFM thickness results in only a slight decrease in indentation modulus; however, increasing

suspended EFM thickness resulted in ~50× increase in indentation modulus (Table 1). For supported scaffolds, a given cell at the center of the EFM is only hundreds of microns from the rigid support (i.e., one EFM thickness away), whereas for suspended scaffolds the rigid support is millimeters away. This likely affects how the cell feels mechanics between these models. Suspending EFMs may also cause an impactful change in the curvature of the EFM. These complexities that arise from inverted cell response to stiffness and suspending EFMs are reflected in part by the bell curve-shaped response to indentation modulus when comparing suspended and supported presentations (Figure 5). Further experiments are warranted to decouple these variables to explain these differences.

In addition, improved characterization employing more advanced FEM models that consider individual fibers over different length scales or empirical data that recapitulates cell contractility would yield helpful insights. We employ an FEM-generated indentation modulus and an experimentally derived Young's modulus. Neither completely describes the results seen herein. Young's modulus is especially poorly-suited to provide insights because the most challenging data arise specifically from changes in structural mechanics, not changes in the materials themselves. Thus, systems have identical Young's moduli.

These data highlight a gap in understanding between cell sensing and response to the mechanical environment. Further, the large differences generated by small, nuanced changes in the mechanical environment, for example a ~37% change in morphology with a change in EFM thickness (Figure 6), suggest caution in interpreting data from mechanical studies that may lack scientific rigor. In addition, these data suggest new variables (i.e., support material, suspension diameter) that can be tuned to alter material mechanobiology effects. The field of biomaterials has historically progressed from early generation scaffolds that are bioinert to later generations that are bioactive or bioresorbable. For example, earlier generation bioinert alumina vs. more recent bioresorbable tri-calcium phosphate dental implants [38]. Whereas it is acceptable to "avoid" biological complexity early, this complexity also provides powerful opportunities to improve patient outcomes. Impending generations of biomedical scaffolds will need to harness the powerful, yet relatively untapped realm of nuanced mechanobiology.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2079-4991/9/4/644/s1, Figure S1: Boundary conditions and interactions for indentation FE models, Figure S2: Individual fiber diameter distributions. Frequency distributions for PS-supported (A) and tension-released (B) EFMs, Figure S3: Focal adhesion kinase expression.

**Author Contributions:** Conceptualization, M.A.C., R.B.D., J.O.W.; methodology, M.A.C., S.S.C., M.T.N., R.B.D.; formal analysis, M.A.C.; investigation, M.A.C.; resources, J.J.L., R.B.D., J.O.W.; data curation, M.A.C., R.B.D.; writing—original draft preparation, M.A.C.; writing—review and editing, M.A.C., M.T.N., J.J.L., R.B.D., J.O.W.; funding acquisition, M.A.C., M.T.N., J.J.L., R.B.D., J.O.W.

**Funding:** This research was funded by the National Science Foundation (DGE0221678, CMMI0747252), and the National Institutes of Health R01HL132355.

**Acknowledgments:** The authors acknowledge support of the Pelotonia Cancer Research Fellowship (MAC, MTN). These are the words and opinions of the authors only and do not reflect those of the Pelotonia program.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.
