*3.1. Fiber Mat Characterization*

Correlating to standard EFM characterization methods, we first characterized EFM morphology and stiffness. Aligned EFMs were spun directly onto PS or PDMS to create supported (solid) or suspended (annulus) scaffolds (Figure 1A,B). Alternatively, EFMs were spun onto non-stick foil, cut and removed from foil to release internal tension arising from the electrospinning process, and fixed to a PDMS annulus to create tension-released scaffolds (Figure 1C). As-spun EFMs demonstrated excellent alignment (Figure 2A), some of which was preserved in the tension-released EFMs, though a clear non-linearity is introduced to the fibers (Figure 2B). In these representative EFM images, individual fiber diameter was consistent at 0.97 ± 0.04 μm and 1.00 ± 0.04 μm for 100 μm thick supported and tension-released EFMs, respectively (*n* = 127 total individual fibers). Fiber diameter distribution was normal (Gaussian) for PS-supported EFMs, but not for the tension-released EFMs (Supplementary Figure S2). Fiber densities in these EFMs were statistically different at 595.8 <sup>±</sup> 17.6 mm−<sup>1</sup> and 539.3 <sup>±</sup> 11.7 mm−<sup>1</sup> for 100 <sup>μ</sup>m thick supported and tension-released EFMs, respectively (*p* = 0.0318, two-sided *t*-test). At the cellular level (~10 μm across), this is a difference between cells interacting with 5.96 fibers or 5.39 fibers. This may be negligible when considering that out-of-plane fibers were quantified, while cells largely interact with the top layer of fibers and do not penetrate the EFM. These fiber diameters and densities are very similar to our previously published results from the same electrospinning process [19].

**Figure 2.** Representative scanning electron microscope (SEM) images of EFMs. Aligned (**A**) and tension-released (**B**) 100 μm thick EFMs showing differences in fiber morphology following release of internal tension. Insets: Fourier transform of the figure, indicating degree of alignment. Scale bar = 50 μm.

Mechanical properties of EFMs were first characterized by calculating an FEM-generated indentation modulus (Table 1). The indentation modulus is based on an axisymmetric model simulating a rigid steel indenter impressing an EFM at its center. The indentation modulus for PS-supported EFMs was up to ~500×, or 8.233 MPa, higher than the indentation modulus for suspended EFMs of the same thickness. However, moduli for supported EFMs of different thicknesses were similar (e.g., 8.250 vs. 7.443 MPa). In contrast for suspended EFMs, the indentation modulus increased by ~50×, or 0.79 MPa, as the thickness of the suspended EFM increased from 50 to 200 μm at constant gap diameter (i.e., 10 mm). Finally, indentation modulus decreased by ~10×, or 0.145 MPa, as the gap diameter increased from 3 to 10 mm. Thus, the weight of each factor on indentation modulus from greatest to least was supported vs. suspended, EFM thickness (suspended), and gap diameter. Based on this model, supported presentations, increasing EFM thickness (suspended), and decreasing gap diameter all lead to a higher indentation modulus. Supported EFMs are generally stiffer than suspended EFMs. This suggests that decreasing gap diameter leads to a mechanical response more similar to a stiffer, supported EFM.


**Table 1.** Indentation modulus (MPa) for electrospun fiber mat constructs.

Indentation modulus generated by finite element modeling.

A uniaxial tensile test was then performed on the EFM support materials (i.e., PS and PDMS) and on the unsupported EFMs. The PDMS 100/0 was made of 100% SylGard 184 and PDMS 50/50 was mixed with 50% SylGard 527 to make it less stiff while keeping surface charge and chemistry constant [27]. As expected, PDMS 50/50 was less stiff, by about 1.5×, than PDMS 100/0 (Table 2). For EFMs with different thicknesses, the mean Young's modulus increased with thickness, although results were not statistically significant. The 200 μm EFM was 1.4× stiffer than the 50 μm EFM (Table 2). These results are similar to indentation modulus calculations for supported EFMs (Table 1), which show that modulus is not a strong function of thickness. Thus, we conclude that EFM thickness does not play a strong role in the mechanical response as measured by Young's modulus.

**Table 2.** Young's modulus of electrospun fiber mats and support materials.


Data are displayed as mean ± standard error.

### *3.2. The Role of the Support in EFM Sca*ff*olds*

In our previous work [20], we showed that the mechanics of an underlying support material can influence the response of cells cultured at close proximity by inducing edge effects in an interfacial region. To investigate how such edge effects can potentially be utilized in EFM scaffolds, cells were grown on EFMs of varying thickness and with differing support materials (Figure 1A). Cell morphology was assessed by the Feret diameter (Figure 3A, red lines), which is the longest distance between any two points of an object. We have previously shown that cells exhibit polarized morphologies on aligned EFMs and that Feret diameter correlates with migration speed for glioblastoma cells [19], the model system employed here. Increasing EFM thickness from 50 to 200 μm led to a statistically significant increase of ~19% in cell Feret diameter for PS-supported EFMs (Figure 3B). Thus, although a statistically significant difference in Young's modulus could not be detected, statistically significant cell responses were still observed.

**Figure 3.** Glioblastoma cell Feret diameter (**A**) as a function of polystyrene (PS)-supported EFM thickness (**B**). Feret diameter is indicated for a representative cell (red lines). A total of 979 cells were analyzed across at least two independent experiments. Levels connected by a star (\*) are statistically significant from each other (*p* < 0.05).

Cells were also cultured on EFMs at a thickness of 50 μm using support materials of decreasing Young's modulus (i.e., PS, PDMS 100/0 and PDMS 50/50) (Figure 4A). PS was selected because it exhibits a Young's modulus several orders of magnitude higher than that of the EFMs employed (i.e., ~2200 vs. ~15 MPa for EFMs), similar to our previous work examining Matrigel (e.g., 1 kPa) supported on glass (e.g., GPa) [20]. For comparison, two types of softer PDMS were employed; both with Young's moduli lower than that of the EFMs (3.13 and 1.29 MPa, respectively vs. 15.00 MPa for EFMs). We observed a ~26% decline in Feret diameter between the PS and the PDMS 50/50-supported 50 μm EFMs. In contrast, no difference in Feret diameter was detected between the two PDMS-supported models. However, when the cells were cultured directly on PDMS 100/0 and PDMS 50/50 support materials without an EFM, a significant difference in Feret diameter was seen (Figure 4B). This suggests that the mechanical response of the EFM scaffold is a combination of the support material and the EFM itself. Thus, tuning the mechanobiological response of cells to EFM scaffolds may be achieved either by altering the EFM thickness or by altering the stiffness of the underlying support material and that cells are able to sense the stiffer material in the interfacial region of composites.

**Figure 4.** Glioblastoma cell Feret diameter in response to culture on 50 μm EFMs supported on substrates of declining stiffness (**A**) and on polydimethylsiloxane (PDMS) supports of declining stiffness with no EFM present (**B**). A total of 1623 cells were analyzed across at least two independent experiments. Levels connected by a star (\*) are statistically significant.
