Growing Neural PC-12 Cell on Crosslinked Silica Aerogels Increases Neurite Extension in the Presence of an Electric Field

Externally applied electrical stimulation (ES) has been shown to enhance the nerve regeneration process and to influence the directionality of neurite outgrowth. In addition, the physical and chemical properties of the substrate used for nerve-cell regeneration is critical in fostering regeneration. Previously, we have shown that polyurea-crosslinked silica aerogels (PCSA) exert a positive influence on the extension of neurites by PC-12 cells, a cell-line model widely used to study neurite extension and electrical excitability. In this work, we have examined how an externally applied electric field (EF) influences the extension of neurites in PC-12 cells grown on two substrates: collagen-coated dishes versus collagen-coated crosslinked silica aerogels. The externally applied direct current (DC) bias was applied in vitro using a custom-designed chamber containing polydimethysiloxane (PDMS) embedded copper electrodes to create an electric field across the substrate for the cultured PC-12 cells. Results suggest orientation preference towards the anode, and, on average, longer neurites in the presence of the applied DC bias than with 0 V DC bias. In addition, neurite length was increased in cells grown on silica-crosslinked aerogel when compared to cells grown on regular petri-dishes. These results further support the notion that PCSA is a promising material for nerve regeneration.


Introduction
Aerogels are gaining attention as biomaterials because of their low bulk density, high surface area to volume ratio, high mechanical strength and ability to serve as a "host" to therapeutic molecules and cells [1,2]. Recent investigations of polymer-crosslinked silica aerogels (PCSA) have established their biocompatibility and biostability [3,4], the ability to detect them within the body using standard clinical imaging techniques [5][6][7], and their capacity to enhance the extension of neurites by PC-12 neural cells and dorsal root ganglia neurons when compared to conventional cell culture substrates [8][9][10]. The latter results suggest that crosslinked silica aerogels have the potential to enhance regeneration after nerve injury. Electrical stimulation (ES) in the form of electric fields (EF) is another method that has been proven to positively influence the extension of neurites by neural cells both in tissue-culture systems and in animal models [11][12][13][14][15][16]. This applies to both static and pulsed electric fields [17] which have been shown to accelerate neurite outgrowth in vitro [16][17][18][19], nerve regeneration in vivo [20][21][22][23], and in some cases influence the direction and orientation of neurite outgrowth [14]. Collectively, these studies suggest that the stimulation of severed nerves by electrical fields has the potential to enhance both nerve regeneration and the prospects of recovery after nerve injury [24]. The majority of the studies on the role of electric fields on neuronal cells have plated those cells on conductive substrates such as a tissue-culture plastic dish covered by polypyrole [25] or thin gold layers [26]. It remains to be seen whether electric fields also increase neurite length when neural cells are plated on PCSA. We have examined this possibility by using the neural cell line PC-12 cells and by examining how these cells respond to electric fields when placed on traditional tissue-culture substrates and on PCSA for comparison. PC-12 cells are a well-established model for peripheral neurons as they express neuronal markers such as the intermediate filament protein peripherin [27] and grow axon-like neurites in response to nerve growth factor (NGF) [28]. Moreover, these cells have been used extensively as a model to understand the influence of electrical stimulation [29][30][31][32]. Additionally, in our previous work [9], we have demonstrated the biocompatibility of PCSA and the ability of this material to foster the extension of neurites longer than those produced by PC-12 cells plated on conventional tissue culture substrates. In this work, the authors report, for the first time, the effect of direct current (DC) applied bias on PC-12 cells grown on PCSA as well as tissue-culture polystyrene (TCPS) serving as the control. The DC electrical stimulation was represented by an applied electrical field generated by a 15 V electrical potential difference. The influence of the applied electrical potential difference on PC-12 cells behavior was evaluated by monitoring (a) average neurite length, (b) directionality of neurite growth with respect to the location of the anode and cathode, and (c) neurite density. The results of three separate trials show that applied DC bias accelerates the growth rate of neurite extension for PC-12 cells plated on PCSA and that this difference was greater than for cells plated on TCPS. In addition, the preferred orientation and greater length of neurites toward the anode which was observed in previous studies for cells on TCPS was also found here for cells plated on PCSA.

Aerogel Preparation
Polyurea-crosslinked silica aerogel (PCSA) substrates were prepared by means of the sol-gel method and supercritical drying as described in detail previously [3,4]. The as-prepared samples were weighed and evaluated for integrity by means of optical microscopy prior to cell-culture steps.

Electrostimulation Device
Electrodes used in this study were constructed in-house by soldering 30 AWG (American wire gauge) wires to two copper rectangular plates cut to the final desired size from a full sheet. The copper plates were then individually embedded in a uniform manner in Sylgard 184 silicone elastomer (Dow Corning) such that all sides of the copper plates were isolated from their surroundings with a 1 mm thick layer of Sylgard 184. Small holes were drilled into the lid of a 3.5 cm petri-dish and the wires were fed through the holes. Next the silicone-embedded copper plates (hereon referred to as electrodes) were attached to the interior of the lid with silicone adhesive (Permatex) 10 mm apart from one another (Figure 1a). PCSA or TCPS (control) substrates of (8 mm × 8 mm × 3 mm) (w × l × h) were then glued to the bottom of the petri dish with silicone adhesive (Figure 1b). The lid was then placed back on top of the petri-dish bottom now containing PCSA or TCPS substrates attached to the bottom. In its final form, when the lid was placed back on the petri-dish bottom, the two electrodes sat on either side of the PCSA or TCPS substrate extending beyond the limits of the sample (Figure 1c) such that the entire sample area was positioned within the electric field region. Prior to placing the lids back on the petri-dishes, PCSA and TCPS substrates were first coated with 4 µg/cm 2 collagen (Invitrogen) (discussed in detail below) in a sterile biological hood prior to plating with cells ( Figure 1d). All dimensions and distances were carefully selected and measured to ensure that the entire surface of the substrate (TCPS or PCSA) was uniformly exposed to the applied DC bias to the best of our ability.

PC-12 Cell Culture
PC-12 pheochromocytoma cells obtained from the American Type Culture Collection (ATCC) were grown in Roswell Park Memorial Institute (RPMI) 1640 medium supplemented with GlutaMax, hydroxyethyl piperazineethanesulfonic acid (HEPES) buffer, 10% horse serum, 5% fetal bovine serum, and penicillin/streptomycin. Before each experiment, PC-12 cells plated on a collagen 1 coated TCPS petri-dish cells were "primed" for neural differentiation with 50 ng/mL NGF for 8 days in low serum medium (RPMI 1640 supplemented with GlutaMax, HEPES buffer, 1% horse serum, and penicillin/streptomycin). This medium was replaced every 2-3 days with fresh low serum medium with NGF. After 8 days, the cells were harvested from the dish and frozen in liquid nitrogen [33] until used for electrostimulation experiments. Details of the method have been described previously [9]. Both 3-day and 1-day cultures were prepared for PCSA and TCPS for comparison.

PC-12 Cell Culture
PC-12 pheochromocytoma cells obtained from the American Type Culture Collection (ATCC) were grown in Roswell Park Memorial Institute (RPMI) 1640 medium supplemented with GlutaMax, hydroxyethyl piperazineethanesulfonic acid (HEPES) buffer, 10% horse serum, 5% fetal bovine serum, and penicillin/streptomycin. Before each experiment, PC-12 cells plated on a collagen 1 coated TCPS petri-dish cells were "primed" for neural differentiation with 50 ng/mL NGF for 8 days in low serum medium (RPMI 1640 supplemented with GlutaMax, HEPES buffer, 1% horse serum, and penicillin/streptomycin). This medium was replaced every 2-3 days with fresh low serum medium with NGF. After 8 days, the cells were harvested from the dish and frozen in liquid nitrogen [33] until used for electrostimulation experiments. Details of the method have been described previously [9]. Both 3-day and 1-day cultures were prepared for PCSA and TCPS for comparison.

Collagen Coating of PC-12 Substrates
8 mm × 8 mm × 3 mm (l × w × h) PCSA and TCPS substrates previously attached to the petri-dish bottom ( Figure 1) were sterilized as previously described [3,4] through immersion in 100% ethanol followed by 1 h exposure to 254 nm ultraviolet (UV) light. Rat tail type I collagen (Invitrogen, Carlsbad, CA, USA) was diluted to a final concentration of 0.05 mg/mL in 20 mM acetic acid [9]. Collagen solution was then laid on the substrate to reach a protein density of 4 µg/cm 2 and was allowed to sit for 1 h at room temperature in a sterile cabinet. Next, the collagen solution was removed with a vacuum line, and the samples were rinsed twice with serum free RPMI 1640.

Electrical Stimulation
NGF-primed PC-12 cells were plated on the collagen I coated TCPS and PCSA electrical devices at a density of 1-1.5 × 10 4 cells/cm 2 in the presence of low serum medium containing 50 ng/mL NGF. Samples that were to be electrically stimulated were connected to a calibrated voltage source (Korad). The voltage source was set to the desired voltage setting and calibrated prior to commencing the experiment. Cells were plated on each substrate type and allowed to grow for 24 h, in the presence of an applied DC bias set to 15 V. The DC voltage was applied to the substrates containing cells approximately 3 h after the cell-plating stage. Cells plated on TCPS and PCSA devices with no applied potential source (setting at 0 V) served as a control. No electrophoresis was observed at 15 V as judged by the lack of bubble formation in the medium. For each condition, three separate trials were performed and results were expressed as means ± SEM (standard error of the mean).

COMSOL Computation of Electric Field
The magnitude of the electric field over the substrates was modeled by using the equations which describe the relations between the electric field E with the gradient of the potential V and the divergence of the displacement field D with the free-charge density ρ f . Finite element analysis (FEA) was performed using COMSOL Multiphysics Modeling software version 5.1. A 2-D model was created consisting of 2 rectangles representing the copper anode and cathode at a distance of 10 mm apart encapsulated by larger rectangles representing Sylgard 184 silicone with a relative permittivity of 2.72 (Dow Corning, Midland, MI, USA). A 3.5 cm diameter circle, encircling the encapsulated electrodes was produced to mimic a petri-dish containing cell-culture medium. The cell-culture medium was assigned a relative permittivity of 80 [34]. After setting a potential difference of 15 V on the copper electrodes, a time independent study was initiated. A maximum mesh size of 3.5 × 10 −4 and an absolute tolerance of 1 × 10 −6 were used for this study.

Scanning Electron Microscopy (SEM)
Samples were fixed in 2.5% glutaraldehyde (Tousimis, Rockville, MD, USA) in 0.1 M sodium cacodylate buffer for 2 h followed by 2 × 10 min washes with 0.1 M sodium cacodylate buffer. Samples were then immersed in an aqueous solution of 1% osmium tetroxide (Electron Microscopy Sciences, Hatfield, PA, USA) for 1 h followed by 2 × 10 min washes in sodium cacodylate buffer and then dehydrated by a series of 10 min each washes in 10%, 30%, 50%, 70%, 90%, and 100% ethanol. Finally, all substrates were allowed to fully air dry before sputter deposition of 10 nm gold/palladium film and imaging with Nova NanoSEM 650 Field Emission Scanning Electron Microscope (FEI Co., Hillsboro, OR, USA).

Neurite Measurement
Neurite lengths and angles were measured using NIH Image J software. Neurite selection was performed as described in detail in previous published work by the authors [9]. The overlaid scale bar from scanning electron microscope (SEM) images was used to calibrate the measurement tools for length measurement. Neurites were treated as vectors and as such were measured with a series of straight line segments using the line segment tool, as shown in Figure 2a where Image J traces have been superimposed on an SEM image. Neurites were defined as the longest cytoplasmic extensions (outlined in red) originating from the cell body and short branches (outlined in green) originating from these extensions were not considered for measurement. Yellow lines show traced neurites used for this study. If the neurite branched, only the longest branch was considered. For all experiments, the anode corresponded to 270 • (a negative y value) and the cathode corresponded to 90 • (a positive y value). In all cases the positive X-axis was set at 0, as indicated on the polar plot in Figure 2b inset. Angles were determined by drawing a line connecting the cell soma to the end of the neurite using the straight line tool creating a resultant vector. Neurite length and density were analyzed with respect to orientation. To produce a neurite density plot, the Rayleigh test of uniformity was used to assess the orientation of the neurites [35]. To determine a preferred direction of growth, vector components were calculated by the use of: where N represents the total number of neurites measured per trial per sample, L represents the individual neurite length, and θ represents the individual neurite angle. Additionally, neurites were broken into 180 • sections and labeled as cathodal (those orientated toward the cathode) and anodal (those orientated towards the anode). Length and density values were compared amongst these groupings. (outlined in red) originating from the cell body and short branches (outlined in green) originating from these extensions were not considered for measurement. Yellow lines show traced neurites used for this study. If the neurite branched, only the longest branch was considered. For all experiments, the anode corresponded to 270° (a negative y value) and the cathode corresponded to 90° (a positive y value). In all cases the positive X-axis was set at 0, as indicated on the polar plot in Figure 2b inset. Angles were determined by drawing a line connecting the cell soma to the end of the neurite using the straight line tool creating a resultant vector. Neurite length and density were analyzed with respect to orientation. To produce a neurite density plot, the Rayleigh test of uniformity was used to assess the orientation of the neurites [35]. To determine a preferred direction of growth, vector components were calculated by the use of: where N represents the total number of neurites measured per trial per sample, L represents the individual neurite length, and ɵ represents the individual neurite angle. Additionally, neurites were broken into 180° sections and labeled as cathodal (those orientated toward the cathode) and anodal (those orientated towards the anode). Length and density values were compared amongst these groupings.

Statistical Analysis
Statistical analysis of neurite length, neurite density, and directional growth bias was performed using a two-tailed student's t-test. Significance (*, **) is defined as p < 0.05, and p < 0.005, respectively. In all cases error bars represent standard error of mean.

Polyurea-Crosslinked Silica Aerogel (PCSA) Surface Effects on PC-12 Neurite Length
The effect of the culture period on neurite density and proliferation was compared for both substrates and shown in Figure 3. Figure 3 shows a 3-day culture period on collagen-coated PCSA (Figure 3a,b) and TCPS (Figure 3c,d) in the absence of applied bias. For both substrates, cells demonstrated neurite outgrowth. The large number and density of neurites for a 3-day culture made it challenging to distinguish the origin and growth cone of neurites, thereby hampering length measurements. Reducing the number of days in culture from 3 to 1 substantially reduced the connectivity of neurites, thereby enabling reliable length measurements. Hence, the electrical stimulation study was performed on 1-day cultures only and select SEM images from different locations are shown in Figure 4. J. Funct. Biomater. 2018, 9, x FOR PEER REVIEW 6 of 12

Statistical Analysis
Statistical analysis of neurite length, neurite density, and directional growth bias was performed using a two-tailed student's t-test. Significance (*, **) is defined as p < 0.05, and p < 0.005, respectively. In all cases error bars represent standard error of mean.

Polyurea-Crosslinked Silica Aerogel (PCSA) Surface Effects on PC-12 Neurite Length
The effect of the culture period on neurite density and proliferation was compared for both substrates and shown in Figure 3. Figure 3 shows a 3-day culture period on collagen-coated PCSA (Figure 3a,b) and TCPS (Figure 3c,d) in the absence of applied bias. For both substrates, cells demonstrated neurite outgrowth. The large number and density of neurites for a 3-day culture made it challenging to distinguish the origin and growth cone of neurites, thereby hampering length measurements. Reducing the number of days in culture from 3 to 1 substantially reduced the connectivity of neurites, thereby enabling reliable length measurements. Hence, the electrical stimulation study was performed on 1-day cultures only and select SEM images from different locations are shown in Figure 4.

Electric Stimulation Effect on Neurite Length, Density, and Orientation
Neurite length: Figure 4a

Electric Stimulation Effect on Neurite Length, Density, and Orientation
Neurite length: Figure 4a   The analysis of 3 separate trials (n = 3) for each substrate at 0 V and at 15 V has been summarized and presented in Figure 5. The effect of applied DC bias was first analyzed without consideration of neurite orientation. The graph presented in Figure 5 shows that, at 0 V, neurites of cells plated on PCSA are about 10% longer that those extended by cells plated on TCPS. For both cells plated on TCPS and PCSA, 15 V externally applied electrical stimulation increased neurite length by at least~12%. Because of the cumulative effects of substrate and of electrical stimulation, at 15 V cells plated on PCSA have extended neurites that are about 10% greater than cells plated on TCPS and subjected to the same electrical stimulation. Also, a qualitative assessment of the separate trials showed no migration of cell body for either substrate at the 15 V applied bias.
J. Funct. Biomater. 2018, 9, x FOR PEER REVIEW 8 of 12 The analysis of 3 separate trials (n = 3) for each substrate at 0 V and at 15 V has been summarized and presented in Figure 5. The effect of applied DC bias was first analyzed without consideration of neurite orientation. The graph presented in Figure 5 shows that, at 0 V, neurites of cells plated on PCSA are about 10% longer that those extended by cells plated on TCPS. For both cells plated on TCPS and PCSA, 15 V externally applied electrical stimulation increased neurite length by at least ~12%. Because of the cumulative effects of substrate and of electrical stimulation, at 15 V cells plated on PCSA have extended neurites that are about 10% greater than cells plated on TCPS and subjected to the same electrical stimulation. Also, a qualitative assessment of the separate trials showed no migration of cell body for either substrate at the 15 V applied bias. Neurite density: PC-12 neurites were divided into two 180° sections, anodal-and cathodal-facing (described in Section 2.8) for determining the neurite orientations. Neurites in the anodal-and cathodal-facing groups were then counted and normalized for the total of number of extensions counted per treatment group. Figure 6a compares the neurite counts for each treatment group. Electrically stimulated samples showed a significant increase in the number of neurites growing toward the anode compared to the cathodal-facing group. However, 0 V treatment showed no statistically significant orientation preference for either substrate. Neurites orientated towards the anode on average were longer than those growing towards the cathode. Neurite density: PC-12 neurites were divided into two 180 • sections, anodal-and cathodal-facing (described in Section 2.8) for determining the neurite orientations. Neurites in the anodal-and cathodal-facing groups were then counted and normalized for the total of number of extensions counted per treatment group. Figure 6a compares the neurite counts for each treatment group. Electrically stimulated samples showed a significant increase in the number of neurites growing toward the anode compared to the cathodal-facing group. However, 0 V treatment showed no statistically significant orientation preference for either substrate. Neurites orientated towards the anode on average were longer than those growing towards the cathode.

Directional Bias
Treating the neurites in each group as vectors, the neurites were vector summed in accordance with relationships 1 and 2 and projected perpendicular to the field direction (Lcosθ) or, parallel to the field direction (Lsinθ). Figure 6b shows the projection of neurites parallel to the externally applied electric field for both substrates, with the arrow indicating the field direction. 15 V groups experienced a significant bias towards the anode (negative y-direction), whereas the 0 V y-and x-components (not shown) of all groups showed no significant bias and no statistical difference between these growth rates. Composite tracings of the cells showed a roughly symmetrical pattern of growth along the x and y directions.

COMSOL Computation of Electric Field
To determine the strength of the electric field that cells were exposed to, a computational model was created with the use of COMSOL. A relatively uniform electric field of 210 mV/mm was computed within the cell-growth area between the two electrodes which agrees with hand calculations. Figure 7 shows the electric field distribution on a logarithmic scale. Figure 6. Effect of electrical stimulation (ES) on neurite density and orientation: (a) all neurites were summed as vectors indicating the bias or the preferred direction of growth on TCPS and PCSA with and without ES. (b) Average length of neurites in the anodal-and cathodal-facing groups of PC-12 neurites growing in the presence of an external EF on TCPS and PCSA substrates.

Directional Bias
Treating the neurites in each group as vectors, the neurites were vector summed in accordance with relationships 1 and 2 and projected perpendicular to the field direction (Lcosɵ) or, parallel to the field direction (Lsinɵ). Figure 6b shows the projection of neurites parallel to the externally applied electric field for both substrates, with the arrow indicating the field direction. 15 V groups experienced a significant bias towards the anode (negative y-direction), whereas the 0 V y-and xcomponents (not shown) of all groups showed no significant bias and no statistical difference between these growth rates. Composite tracings of the cells showed a roughly symmetrical pattern of growth along the x and y directions.

COMSOL Computation of Electric Field
To determine the strength of the electric field that cells were exposed to, a computational model was created with the use of COMSOL. A relatively uniform electric field of 210 mV/mm was computed within the cell-growth area between the two electrodes which agrees with hand calculations. Figure 7 shows the electric field distribution on a logarithmic scale.

Discussion
PC-12 cells are a good in vitro model for studying the mechanisms of neurite outgrowth and identifying factors influencing this process [36]. Since directed neurite outgrowth is important for nerve regeneration, PC-12 cells are also a useful model for nerve regeneration. Previously, we showed that, compared to TCPS, PCSA represents a substrate that enhances the length of neurites extended by PC-12 [9]. The presence of the collagen layer is not expected to alter the response of the cells to the combined effect of electrical stimulation and substrate. Based on the study published by Willitz et al. [37] collagen layers with the concentration used here have an insignificant stiffness which means that

Discussion
PC-12 cells are a good in vitro model for studying the mechanisms of neurite outgrowth and identifying factors influencing this process [36]. Since directed neurite outgrowth is important for nerve regeneration, PC-12 cells are also a useful model for nerve regeneration. Previously, we showed that, compared to TCPS, PCSA represents a substrate that enhances the length of neurites extended by PC-12 [9]. The presence of the collagen layer is not expected to alter the response of the cells to the combined effect of electrical stimulation and substrate. Based on the study published by Willitz et al. [37] collagen layers with the concentration used here have an insignificant stiffness which means that cells are in fact recognizing the different substrates and their influence is not masked by the collagen layer. Electric fields have also been shown to increase the length of neurites of cells grown on TCPS. Our results show that electric fields also enhance neurite outgrowth by PC-12 cells plated on PCSA when compared to cells plated on PCSA but not stimulated by an electric field. Compared to unstimulated cells grown on TCPS, neurites extended by PC-12 cells on electrically stimulated PCSA are about 20% longer, and about 10% longer than neurites extended by PC-12 cells on electrically stimulated TCPS.
It was shown here that not only is the PCSA surface conducive to longer PC-12 neurite extension, but also the addition of externally-applied DC bias (of 200 mV/mm) selectively increased the growth rate of PC-12 neurites growing towards the anode rather than the cathode on both TCPS (control) and PCSA substrates. This corroborates previous work by Cork et al. [30] and Blumenthal et al. [38]. Under further analysis the data suggests that the external electric field stimulates longer extensions towards the anode, and does not appear to have a noticeable effect on extensions oriented towards the cathode. There was no statistically significant difference between TCPS and PCSA cathodal-facing neurite length and their respective 0 V controls. In addition to analyzing neurite length, neurite density in response to the electric field was also considered. An increase in the number of neurites facing the anode relative to the number of neurites facing the cathode was observed when considering the total number of neurites in the anodal and cathodal groups as a percentage, which also corroborates a previous study done by Cork et al. [30].
Endogenous electric fields of varying strengths occur naturally in vivo. With no injury present, endogenous field strengths are of the order of 0-10 mV/mm [39]. However, upon injury endogenous field strengths can increase 10-fold at the site of injury in mammalian tissue [26]. For example, in guinea pig spinal cord explants, endogenous field strengths of up to 200 mV/mm have been reported [40]. A potential difference of 15 V resulting in an EF strength of 200 mV/mm was chosen for this study because it is at the upper end of what is seen in mammalian tissue naturally and resulted in an EF strength comparable to those used in previous studies of this nature. The results indicate that it is possible to accelerate neurite growth and influence directionality. Previous studies have focused on conducting substrates such that a current is passed through the substrate [41]; the results presented here show similar behavior, but in the presence of an electric field applied to PCSA. Altogether, these results support the notion that PCSA is a promising material to support nerve regeneration as, alone, they enhance the outgrowth of neurites, while in the presence of an electric field this outgrowth is boosted and exhibits a definite polarity.