The primary purpose of this study was to examine the role of surface chemistry versus topography in directing cell attachment and proliferation in hydrogel-EFM composites for neural tissue engineering. It is therefore critical that the fibers employed display similar topographical properties after surface chemistry modification. Thus, in addition to water contact angle, which establishes hydrophobicity of the fiber surface, average fiber diameter, mean pore size, and porosity (Table 2, Figure 1) were determined for both PCL and PCL/PEGPCL core/shell randomly aligned EFM scaffolds. The water contact angles of PCL and PCL/PEGPCL EFMs were 135.2 ± 5.4° and ~0°, respectively, establishing these materials as example hydrophobic (PCL) and hydrophilic (core/shell PCL/PEGPCL) EFM substrates. Morphological characterization via scanning electron microscopy (SEM) (Figure 1) showed no significant difference in average fiber width and mean pore size between PCL and PCL/PEGPCL EFMs (p > 0.05). Similarly, no significant difference of porosity (p > 0.05), as measured via ethanol displacement, was found between PCL and PCL/PEGPCL EFMs. The “fractional contact area”, Φ, which is the fraction of fiber surface contacting other fibers, was used to establish the “available surface fraction”, (1 − Φ), open to cell adhesion. The available surface fraction was calculated from ε based on the theory of Eichhorn and Sampson [28] and assuming a mean coverage of ~10 to ∞ based on SEM data. These values suggest that these materials are suitable as tissue engineering constructs as >70% of the surface is available. Thus, the EFMs employed exhibited nearly identical morphologies, but altered surface properties (i.e., composition and hydrophobicity).

SK-N-SH neuroblastoma cell attachment was examined on two materials with differing topographical features: Gs and EFMs, and two materials with differing hydrophobicity in each group: PEG < PEGPCL for Gs and PEGPCL < PCL for EFMs. We also investigated composites of these two materials that were formed by in situ photopolymerization to secure EFMs to the outer hydrogel surface, thus presenting EFMs as the primary surface available for cell attachment (see Table 1 for all materials employed). Cell attachment and proliferation were qualitatively evaluated using fluorescently-labeled samples after 1 day of incubation, which is sufficient to permit attachment, but unlikely to result in substantial proliferation. Qualitatively, it is clear that cells preferred composite (Figure 2C, D) and EFM (Figure 2E, F) surfaces to G surfaces (Figure 2A, B), despite the fact that the PEGPCL G (Figure 2B), PEGPCL G-PCL/PEGPCL EFM composites (Figure 2D), and PCL/PEGPCL EFMs (Figure 2F) present identical PEGPCL surface compositions.

To further quantify these results, SK-N-SH cell proliferation was evaluated using the MTT assay (Figure 3). Interestingly, there were no significant differences in cell number as measured via normalized absorbance between Gs, EFMs, and composites of differing surface chemistry (i.e., more hydrophilic vs. more hydrophobic). However, significant differences were observed between any two of these materials (p < 0.05). Thus, the number of adherent cells was highest for materials that displayed more fibrous character (e.g., EFM > composites > hydrogels), regardless of the hydrophilicity of the component materials (e.g., PEG > PEGPCL > PCL); and the incorporation of EFMs increased cell attachment to hydrogels composed of the same materials. These results most likely occur because of differences in topography; however, the increased adhesive surface area of EFMs could also contribute. EFMs display pore sizes of ~1–2 µm (Table 2), in contrast, hydrogels have pore sizes on the order of a few nanometers [29], and would thus appear as more two-dimensional surfaces than the EFMs.

The morphology of adherent cells was also examined on EFM and composite surfaces using qualitative SEM imaging (Figure 4) and quantification of cell spreading and circularity using ImageJ image analysis software (Figure 5A, B). G surfaces were not examined because of the low number of adherent cells (e.g., Figure 2). In all SEM images, cells spread on EFM and composite surfaces, aligning to and interacting with multiple fibers. However, cells on composite materials (Figure 4C, D) appeared more rounded than those on EFM-only surfaces (Figure 4A, B), despite the similarities in surface composition between materials that displayed PCL (Figure 4A, C) and PCL/PEGPCL EFMs (Figure 4B, D), respectively.

Cell circularity and spreading values further describe these morphological observations. Circularity is measured on a scale of 0 to 1, with 1 representing a completely circular cell and 0 representing a completely polarized cell. Cell spreading is measured by quantifying the surface area coverage of cells and is measured in µm². Consistent with SEM observations (Figure 4), the circularity of SK-N-SH cells on EFM-only scaffolds was significantly lower than that of cells on composite scaffolds (p < 0.05) (Figure 5A), despite similarities in material surface composition (e.g., PCL EFM vs. PEGPCL G-PCL EFM composite and PCL/PEGPCL EFM vs. PEGPCL G-PCL/PEGPCL EFM composite). This suggests a topographical influence; however, the role of mechanical properties cannot be excluded. PCL EFMs have a modulus of ~7 MPa [30], whereas PEGPCL hydrogel moduli are on the order of ~0.1–1 MPa [29]. Mechanical properties, including increased stiffness, have been shown to influence cell spreading, proliferation, and migration [31,32].

In contrast to other observed parameters, cell spreading (Figure 5B) was primarily influenced by composition. The more hydrophobic PCL EFM-based materials (PCL EFM and PEGPCL G-PCL EFM composite) displayed significantly higher average spreading values than PCL/PEGPCL EFM-based samples (PCL/PEGPCL EFM and PEGPCL G-PCL/PEGPCL EFM composite) (p < 0.05), whereas there was no significant difference in cell spreading within groups potentially presenting different topographies but similar surface characteristics (PCL EFM vs. PEGPCL G-PCL EFM; or PCL/PEGPCL EFM vs. PEGPCL G-PCL/PEGPCL EFM composite) (p > 0.05).

To further evaluate composite materials (PEGPCL G-PCL EFM and PEGPCL G-PCL/PEGPCL EFM composites) in a more physiologically relevant system, we also explored the behavior of adherent primary rat cortical neurons. SEM images (Figure 6) show that cells on the more hydrophilic PEGPCL G-PCL/PEGPCL EFM composite samples exhibited greater spreading, indicating better interaction between cortical cells and this scaffold. This was further validated by quantification of cell spreading. Cells on more hydrophilic PEGPCL G-PCL/PEGPCL EFM composite surfaces showed significantly larger spread areas than those cultured on PEGPCL G-PCL EFM composite materials (p < 0.05) (Figure 7). Interestingly, although surface chemistry was still a driving factor in cell spreading, the trend observed with cortical cells was opposite to that of SK-N-SH cells. More hydrophilic surfaces (i.e., PEGPCL G-PCL/PEGPCL EFM composites) favored cell spreading of rat cortical cells (Figure 7), whereas SK-N-SH cells spread more on hydrophobic surfaces (Figure 5B). No significant difference was observed in cell circularity between these two samples (p > 0.05) (Figure 7) consistent with the results of SK-N-SH culture (Figure 5A).

All chemicals were purchased from Sigma-Aldrich, unless stated otherwise.

To evaluate EFM hydrophobicity and surface properties, EFM water contact angles were measured at room temperature with a contact angle goniometer (Ramé-hart Instrument co., Model 200, Succasunna, NJ, USA). At least 5 measurements were taken per condition. The morphologies (i.e., fiber width and pore size) of PCL and PCL/PEGPCL core/shell electrospun fiber mats were characterized by scanning electron microscopy (SEM, Quanta 200, FEI). All samples were secured to aluminum mounts by double-sided carbon tape and then sputter coated with gold before viewing. SEM images at 800×, 1500× and 3000× magnifications (N = 3) were analyzed using Image J (NIH) to measure the fiber width (ω) and pore size. For fiber width measurements, at least 75 fibers per sample and 3 samples per condition were examined. For pore size measurements, at least 50 pores per sample and 3 samples per condition were analyzed to obtain the average value of the major axis for these noncircular pores [34].

Network porosity (ε), was estimated using a liquid displacement method reported by Guan et al. [35]. Ethanol was selected as the displacement liquid because of its ease of penetration into open pores without inducing swelling or shrinkage. EFM samples were immersed in a known volume of ethanol (V₁) for 5 min, and then pressed to force air from open pores. The total volume of ethanol and immersed samples was recorded as V₂. After the removal of ethanol-impregnated samples, the volume of the residual ethanol was recorded as V₃. EFM porosity was calculated using the following Equation:

Diacryl-PEGPCL copolymer was synthesized as described previously [33], using a modified method derived from Hubbell et al. [36]. Briefly, in the presence of stannous octoate (Tin II-ethylhexanoate), new chemical bonds between ε-caprolactone and the hydroxyl ends of toluene-azeotropic distilled PEG (molecular weight M_w = 950–1,050) were formed through ring-opening polymerization at 130 °C under an argon atmosphere. The molar ratio of ε-caprolactone to PEG was 2 to 1. Then, after being precipitated in ice-cold hexane, filtered through a Buchner funnel, and dried in a vacuum oven for 1 hour at 40 °C, purified PEGPCL intermediate was acrylated using acryloyl chloride (ACCl). Under argon atmosphere, the mixture of ACCl and dichloromethane (DCM) was slowly added (0.2 mL/min) to the mixture of PEGPCL intermediate, triethylamine (TEA), and DCM in dry ice-acetone bath ~−30 °C. After the addition of ACCl, the reaction was completed through ~60 hours continuous stirring at room temperature. Byproduct TEA-HCl was removed by repeated solvent evaporation and filtration. Diacryl-PEGPCL product was precipitated in ice-cold diethyl ether, filtered through a Buchner funnel and dried in a vacuum oven in the presence of P₂O₅ overnight. Purified PEGPCL intermediate and diacryl-PEGPCL polymer could be stored in a −20 °C freezer for up to a month. Proton nuclear magnetic resonance (H¹ NMR, Bruker DPX400) and Fourier transform infrared resonance (FT-IR, Thermo Scientific) were applied to characterize copolymer samples. The yields of PEGPCL intermediate and diacryl-PEGPCL were 91.3% and 85.2% respectively, consistent with previously reported results [33].

Diacryl- PEG or PEGPCL polymer was dissolved in Dulbecco’s phosphate buffered saline (D-PBS) at 22 wt %, and then mixed with initiators (0.1 wt % Irgacure 2959 and 0.5 wt % 1-vinyl-2-pyrrolidone). Precursor solution was sterilized via filtration (0.22 µm filter with 0.8 µm pre-filter, Millipore) before UV photopolymerization.

SK-N-SH cells were purchased from American Type Culture Collection (ATCC, CRL-1721, Manassas, VA, USA). Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS, Gibco) and 1% penicillin-streptomycin (Gibco) as per the manufacturer’s instructions. PEG and PEGPCL Gs, PEGPCL G-PCL EFM and PEGPCL G-PCL/PEGPCL EFM composites, and PCL and PCL/PEGPCL EFMs were deposited in a 96-well tissue culture plate as described previously (N = 6 for each material). SK-N-SH cells were seeded at a density of 5 × 10⁴ cells/cm² into each well and cultured overnight in an incubator (37 °C, 5% CO₂).

SK-N-SH cell attachment and viability on these materials was evaluated using fluorescent cell tracking (N = 3), the MTT cell proliferation assay (N = 3), and SEM imaging (N = 3, as described in Section 3.8). For fluorescent cell tracking, SK-N-SH cells were pre-stained using the Celltracker^TM fluorescent imaging probe (Green CMFDA, Invitrogen). After 1 day, fluorescently-labeled SK-N-SH cells were imaged using a reflected differential interference contrast (DIC) microscope (Olympus BX41). Five images per magnification (10×, 20×) were taken at random spots for each sample (N = 3). The MTT cell proliferation assay was conducted following standard protocols (Sigma-Aldrich). Sample absorbance was observed using a Versamax UV-visible micro-plate reader in triplicate.

Rat cortical neurons (from E18 rat cortex) were purchased from BrainBits LLC. PEGPCL G-PCL EFM and PEGPCL G-PCL/PEGPCL EFM composites were examined in cortical culture. Hydrogel-EFM composite materials (N = 3 for each sample) were sterilized and secured to the bottom of wells in sterile 24-well tissue culture plates as described previously. Rat cortical neurons were seeded at 5 × 10⁵ cells/cm², cultured in neurobasal growth medium with 2% B-27, 0.5mM Glutamax added (all Invitrogen), and incubated at 37 °C, 5% CO₂ for 4 days before morphological evaluation. Cortical cultures were examined 4 days after seeding, as opposed to the 24 hour period used for SK-N-SH experiments, to permit maturation of the embryonic cells to the neural phenotype [38,39].

SK-N-SH cells were fixed after 1 day of incubation. After rinsing residual proteins with a PBS wash, cells were fixed in PBS containing 4% paraformaldehyde and 4% sucrose for 1 hour at room temperature. Samples were then rinsed three times with PBS containing 4% sucrose, and dehydrated in a graded ethanol series (50, 70, 80, 95 and 100% for 5, 5, 10, 10 and 10 × 2 min, respectively) and ethanol:hexamethyldisilazane (HMDS, Sigma-Aldrich) series (75, 50, 25 and 0% Ethanol remainder HMDS for 15, 15, 15 and 15 × 2 plus overnight, respectively). After completely dehydrated, scaffolds were mounted on specimen holders, sputter-coated with gold, and observed using SEM (Quanta 200, FEI). Images were collected at 800×, 1,500× and 3,000× (N = 6) magnifications and analyzed using Image J software (NIH, Bethesda, MD, USA). Cell area and cell circularity were measured by selecting individual cells and obtained by setting the measurements to report area and shape descriptors. Similarly, after a 4 days incubation period, the same fixation, dehydration, and SEM imaging procedures were applied to characterize rat cortical neuron seeded scaffolds.

Statistical analyses were performed using one way analysis of variance (ANOVA) with Tukey HSD testing (JMP 8.0). p values of <0.05 were considered to be significant.