Scale of Carbon Nanomaterials Affects Neural Outgrowth and Adhesion.

HHS Public Access Author manuscript Author Manuscript

IEEE Trans Nanobioscience. Author manuscript; available in PMC 2017 January 25. Published in final edited form as: IEEE Trans Nanobioscience. 2016 January ; 15(1): 11–18. doi:10.1109/TNB.2016.2519505.

Scale of Carbon Nanomaterials Affects Neural Outgrowth and Adhesion Eric Franca1, PitFee Jao2, Sheng-Po Fang2, Sankaraleengam Alagapan1, Liangbin Pan1, Jung Hae Yoon2, Yong-Kyu ‘YK’ Yoon2, and Bruce C Wheeler1 1J.

Crayton Pruitt Family Department of Biomedical Engineering, University of Florida, Gainesville, FL, USA

Author Manuscript

2Department

of Electrical and Computer Engineering, University of Florida, Gainesville, FL, USA

Abstract Carbon nanomaterials have become increasingly popular microelectrode materials for neuroscience applications. Here we study how the scale of carbon nanotubes and carbon nanofibers affect neural viability, outgrowth, and adhesion.

Author Manuscript

Carbon nanotubes were deposited on glass coverslips via a layer-by-layer method with polyethylenimine (PEI). Carbonized nanofibers were fabricated by electrospinning SU-8 and pyrolyzing the nanofiber depositions. Additional substrates tested were carbonized and SU-8 thin films and SU-8 nanofibers. Surfaces were O2-plasma treated, coated with varying concentrations of PEI, seeded with E18 rat cortical cells, and examined at 3, 4, and 7 days in vitro (DIV). Neural adhesion was examined at 4 DIV utilizing a parallel plate flow chamber. At 3 DIV, neural viability was lower on the nanofiber and thin film depositions treated with higher PEI concentrations which corresponded with significantly higher zeta potentials (surface charge); this significance was drastically higher on the nanofibers suggesting that the nanostructure may collect more PEI molecules, causing increased toxicity. At 7 DIV, significantly higher neurite outgrowth was observed on SU-8 nanofiber substrates with nanofibers a significant fraction of a neuron’s size. No differences were detected for carbonized nanofibers or carbon nanotubes. Both carbonized and SU-8 nanofibers had significantly higher cellular adhesion post-flow in comparison to controls whereas the carbon nanotubes were statistically similar to control substrates.

Author Manuscript

These data suggest a neural cell preference for larger-scale nanomaterials with specific surface treatments. These characteristics could be taken advantage of in the future design and fabrication of neural microelectrodes.

Keywords carbon nanotubes; carbon nanofibers; neural growth; neural nanotechnology

Correspondence to: Bruce C Wheeler.

Franca et al.

Page 2

Author Manuscript

I. Introduction The development and design of electronics and devices that interface with the nervous system is a hotly studied field in biomedicine. Microelectrodes have been utilized for treating the symptoms of patients suffering from ailments ranging from Alzheimer’s and Parkinson’s disease to treatment-resistant depression [1–3]. Though there has been great improvement in the development and application of neural prosthetics, particularly neural microelectrodes, there remain several serious challenges including the loss of physical and electrical contact between microelectrodes and neural tissue over time [4].

Author Manuscript

For over 50 years, sharpened metal electrodes have been used as neural probes [5]. More recently, different materials ranging from silicon to flexible polymer electrodes have been utilized [6, 7]. Innovations in neural electrode materials and designs have focused on more effectively integrating electrodes with neural tissue, improving the electrical characteristics, and optimizing the long-term electrode viability [8–10]. Novel electrode materials, as they are developed, require biocompatibility assessments prior to implantation in vivo. These neural microelectrode materials are commonly tested in vitro using dissociated neurons, such as cortical [11] or hippocampal [12] neural networks, to observe any adverse effects the materials have on the biological tissue. Materials with different physical topographies, conductivities, and surface chemistries have been shown to have a significant effect on the viability and outgrowth of neurons in vitro [13–26].

Author Manuscript

Carbon nanotubes (CNTs) and carbonized nanofibers (CNFs) are two popular microelectrode nanomaterials. These materials have been shown to be biocompatible when immobilized and are capable of improving the impedance and charge transfer of metal electrodes with a unique, nano-scale topography [27–31]. Interestingly, some research has shown that CNTs and CNFs not only improve electrode characteristics, but can also have some effects on neural and/or glial outgrowth and electrophysiology [14, 22, 32–39]. We are able to deposit these nanomaterials onto various substrates in patterned, uniform films on a large scale. Due to limitations with other deposition methods, such as electrodeposition, it is more difficult to produce these large, patterned depositions with other electroconductive materials, such as PEDOT [9]. This flexibility allows for the larger-scale experimentation with neural networks proposed in this work.

Author Manuscript

This study focuses on how variations in the scale of nanomaterial topography affects the viability, outgrowth, and adhesion of cortical neural cultures. CNTs deposited via a simple, layer-by-layer method form thin, conductive CNT-polyelectrolyte films. In contrast, CNFs and carbonized thin films (CTFs) deposited via an electrospinning/photo-patterning/ carbonization method form a thicker film of conductive, carbon depositions. Additional, non-conductive nanofibers and thin films (made from the photopatternable epoxy SU-8) were fabricated using the same methods. The substrates were treated with O2-plasma or the cationic polymer polyethylenimine (PEI). Though PEI has been demonstrated to be toxic when in solution [40], adsorbed it is a popular surface treatment to facilitate neural outgrowth [17, 41]. PEI is a cheap material with a simple implementation that has been used for decades to functionalize surfaces to support neural viability and outgrowth [17]. PEI has

IEEE Trans Nanobioscience. Author manuscript; available in PMC 2017 January 25.

Franca et al.

Page 3

Author Manuscript

also been demonstrated to be an effective cationic polymer to utilize in layer-by-layer assembly of various depositions [42, 43].Thus PEI is an attractive surface treatment for the purposes of this study. We are interested in how neural tissue integrates with these carbon nanomaterials as these characteristics may have implications in their utilization as microelectrode materials in the future. We demonstrate here the effects of higher concentrations of the PEI surface treatment and rougher surface topographies (e.g. nanofibers) on the viability of neurons. Additionally, we demonstrate the extent to which neurons grow and integrate into different topographies by measuring neurite outgrowth and performing neural adhesion experimentation with parallel plate flow chambers.

II. Materials and Methods Author Manuscript

A. Preparation of SU-8 and Carbon Nanomaterial Substrates

Author Manuscript

Carbonized nanofibers (CNFs) were fabricated by our collaborators Jao et al. via their previously described methods [44, 45]. In short, an SU-8 (2025, Microchem Inc.) solution was used to electrospin the nanofibers onto a polished silicon substrate. The nanofibers were then patterned by UV-lithography to produce patterned SU-8 nanofibers (NFs). These depositions can be carbonized in a tube furnace under forming gas at 1000 °C to produce carbonized nanofibers (CNFs); this pyrolysis causes the nanofibers to lose a significant portion of volume resulting in smaller, conductive, nearly pure carbon nanofibers. Patterned thin film substrates were prepared by spin-coating SU-8 onto a polished silicon wafer, followed by UV-lithography to produce SU-8 thin-films (TFs) which were carbonized to produce carbonized thin-films (CTFs). These four physically contrasting materials (CNFs, NFs, CTFs, and TFs) allow us to compare the effects of roughness, nano-feature scale, and the resulting carbonized nanomaterials on neural cell growth.

Author Manuscript

Carboxylated multi-walled carbon nanotubes (MWNTs; www.cheaptubes.com) were dispersed at 0.2 mg/mL in dimethyl formamide (DMF) by utilizing a pulsing (2 s on, 1 s off) ultrasonicator probe for 40 min. The solution was then centrifuged at 3000·g for 5 min and the supernatant was extracted. The thin films of CNTs were prepared on glass coverslips using the technique of layer-by-layer (LBL) assembly in conjunction with layers of poly(ethylenimine) (PEI). The nanotubes are electrostatically attracted to the positively charged PEI molecules on the substrate via their carboxyl groups resulting in a stable nanotube films. Briefly described, the glass coverslips were submerged fully into a solution of 0.1 % PEI (w/v) in 18 MΩ de-ionized (DI) water followed by submerging the substrate halfway into dispersed MWNTs to coat half of the substrate with the deposition; this process, once completed, results in a single bilayer coating on half of the coverslip. This process was repeated to add additional bilayers. Each layer-by-layer sample is designated as CNT-n where n is the number of bilayers. For this study, we utilized CNT-3 and CNT-9, denoting three and nine bilayers, respectively.


Franca et al.

Page 4

B. Surface Characterization

Author Manuscript

Surfaces were characterized via atomic force microscopy (AFM; VEECO Dimension 3100) and scanning electron microscopy (SEM; Hitachi S-4000 FE-SEM). Samples were sputter coated with a thin layer of gold palladium (AU-PD) with a Denton DeskV Sputter coater prior to SEM imaging. AFM micrographs were analyzed with Gwyddion 2.31, open-source scanning probe microscopy software [46] which provided the average roughness values. Ten, two-dimensional average roughness (Ra) profiles were randomly chosen from each micrograph and averaged for the reported Ra values. Nanofiber and nanotube widths (or feature widths) were measured from the SEM micrographs with ImageJ [47]. These values were compared to the average width of cortical neurites as measured using a 40× fluorescent image of a neural culture. C. Surface Preparation and Cell Culture

Author Manuscript

The nanofiber (CNF and NF) and thin-film (CTF and TF) surfaces were treated with O2plasma for 30 s followed by exposure to UV in a biosafety cabinet overnight. Surfaces were then treated with a sterilized 0.1 % PEI, 0.001 % PEI, 0.00001 % PEI (w/v), or 0 % PEI solution (denoted as the Plasma group) in 18 MΩ DI water. The surfaces were rinsed five times with sterile 18 MΩ DI water and dried via aspiration. CNT samples were sterilized in 70 % ethanol solution for 15 min, rinsed with 18 MΩ DI water, and dried via aspiration.

Author Manuscript

E18 rat cortex (Brainbits, LLC) is dissociated according to the vendor’s protocol. Specifically, the cortices were digested in sterile, 2 mg/mL papain in Hibernate E (Brainbits, LLC) at 35°C for five minutes. The tissue clump was then removed from the papain solution and dispersed via trituration in the original shipping solution (Hibernate E with B27). The cell suspension was centrifuged at 1000 rpm for one minute, the supernatant removed, and the cells were re-suspended in Neurobasal media supplemented with B27, GlutaMAX, and Penicillin/Streptomycin. The cells were plated at 200 cells/mm2. A control surface coated with 0.1 % PEI and seeded with > 600 cells/mm2 was used to condition the media of the Plasma groups of CNFs, CTFs, NFs, and TFs in order to support long-term growth; cells on plasma-only surfaces (without media conditioning) become non-adherent within one week. The cultures are kept at in an incubator at 37 °C and 5 % CO2. D. Surface Charge

Author Manuscript

Prior to cell culture, surface charge was investigated by measuring the zeta potential (Zp) of Plasma, 0.001 % PEI, and 0.1 % PEI surface treatments for NFs, TFs, and polished silicon. This measurement was performed using the clamping cell of a commercial electrokinetic analyzer with 1 mM KCl (EKA; Anton Paar). It was not possible to manufacture a stable CTF or CNF film large enough to fit the clamping cell required for this measurement; as a result, these samples have been excluded. Eight measurements were taken (n=8) from each sample and averaged to produce the Zp value from each sample at an average pH level of 7.1. E. Cell Viability and Outgrowth Measurements Cells were stained with a Live/Dead Cell Viability Assay utilizing the vendor's protocol (Life Technologies; L-3224). In short, 5 µL of calcein AM (live cells - green) and 20 µL of IEEE Trans Nanobioscience. Author manuscript; available in PMC 2017 January 25.

Franca et al.

Page 5

Author Manuscript

ethidium homodimer-1 (EthD; dead cells - red) were added to 10 mL of Dulbecco’s phosphate buffered saline 1X (D-PBS). The samples were washed with D-PBS once, and submerged in the staining solution for 30 minutes. Pictures were taken at 10× magnification. As the calcein (green) also stains neurites, the images were utilized to quantify cellular and neurite characteristics on and off of each patterned material.

Author Manuscript

Neural viability was calculated by counting the living cells and dead cells at 3 days in vitro (DIV) using the aforementioned Live/Dead Cell Viability Assay and reported as the percentage of live cells over total cell count (sum of live and dead cells). Both on-pattern (experimental substrate) and off-pattern (control silicon or glass) viabilities are reported. Cellular coverage was measured at 7 DIV by setting a threshold in ImageJ highlighting the total fluorescent cell area and quantified as a percentage of the total area of substrate; this measurement was further subdivided into neurite and soma coverage by limiting the mask threshold. Neurite length measurements were performed utilizing the NeuriteQuant toolkit in ImageJ [48]. Reported data are cell density (cells per mm2) and total neurite length per area at 7 DIV. F. Neural Adhesion Assay

Author Manuscript

A parallel plate flow chamber was utilized to study the adhesion of neurons at 4 DIV to the nanotube and nanofiber substrates (both carbon and SU-8). Specifically, a perforated silicon gasket was placed over the vacuum channel of the flow chamber. The sample was then placed on top of the gasket, cell side downwards, carefully so as to avoid bubbles. A vacuum was applied to create a seal and hold the sample in place. An area in the center of the sample was chosen randomly and photographed for cell counting. Three 60 mL syringes were filled with 1× D-PBS at 37 °C and loaded onto a syringe pump. A flow rate of 180 mL/minute (251 dynes/cm2) was applied for 30 s per sample. The same area was photographed once more after the experiment. The cells were counted before and after to quantify what percentage of cells remained following the flow trial; this value was taken as a representation of neural adhesion, or neural resistance to shear flow. G. Statistical Analyses One-way ANOVAs were utilized for the physical and surface charge measurements of the substrates. Repeated-measures ANOVAs were utilized for all viability, neurite, and adhesion measurements. Where significance was reported (p < 0.05), Newman-Keuls post hoc tests were performed.

III. Results Author Manuscript

A. Surface Characterization SEM micrographs for nanotube and nanofiber surfaces are shown in Figure 1. This figure clearly shows the contrasting scale between the carbon nanotube and nanofiber substrates while also demonstrating the shrinkage of nanofibers caused by the carbonization process in comparing Figure 1C and 1D. Average roughness (Ra) values are presented in Table 1 and range from 0.26 nm to 164 nm for carbonized thin film and carbonized nanofibers, respectively. Additionally, the average size of the topographical features for each substrate


Franca et al.

Page 6

Author Manuscript

is reported (Table 1) which further demonstrates the contrasting scales between the nanofibers and the carbon nanotubes. There are no values reported for the thin films as there are no significant nano-topographical features to measure (e.g. a nanofiber or nanotube). We found that the average width of neurites was statistically larger than all features with the exception of the SU-8 nanofibers (n = 15; p < 0.0005). B. Surface Charge

Author Manuscript

We investigated how surface charge might correlate to the toxicity seen in our 0.1 % PEI viability results. At an average pH of 7.1, the surface charge (as measured by Zp) varied dependent on the physical structure of the sample as well as the specific surface treatment (Figure 2). The PEI treatment yielded positive Zp values, whereas the Plasma-only treatment yielded negative Zp values. SU-8 nanofibers and thin films treated with 0.1 % PEI had significantly higher Zp values in comparison to polished silicon under the same conditions (n = 8; *p < 0.0005). Diluting the PEI solution to 0.001 % resulted in similar Zp values across the measured groups. The data suggest that the nanofiber substrates are especially attractive to PEI in comparison to controls or flat substrates. C. Neural Viability

Author Manuscript

The neural viability assay was performed across several concentrations of PEI at 3 DIV (Figure 3). For both CNT substrates, the neural cell viability showed no significant difference to each paired control; both of these groups utilize 0.1 % PEI in their deposition procedure. This concentration of PEI is an industry standard which supports neural outgrowth on many substrates including glass and culture plastic. However, with both the SU-8 and carbonized nanofibers and thin films, the concentration of PEI applied to the surface significantly impacted the neural viability in comparison to off-pattern controls. Neural viability on the substrates treated with the standard 0.1 % PEI concentration was significantly lower than most other treatments on- and off-pattern, with the exception of the CTF group (Figure 3; n = 4; *p < 0.001 and †p < 0.05). Upon dilution of PEI to 0.001 %, the on- and off-pattern cell viability was statistically comparable for all groups. Treatment of the substrates with 0.00001 % PEI and without any PEI (Plasma) yielded significantly higher cell viability on-pattern. As the 0.001 % PEI concentration yielded comparable results between on- and off-pattern areas, this concentration was utilized for all further experiments with the patterned nanofiber (CNF-PEI, NF-PEI) and thin film substrates (CTF-PEI, TFPEI). D. Neural Outgrowth

Author Manuscript

The contrasting neural outgrowth on patterned substrates (nanofibers and thin films) treated with PEI are visualized in Figure 4. The density (per mm2) of adhering cell bodies was quantified at 7 DIV across paired experimental (on-pattern) and control background (offpattern) substrates (Figure 5A). No significant differences were observed for this metric. On the carbon nanotube and TF-PEI substrates, there were fewer cells per area compared to controls, though this was not statistically significant. This demonstrates that the cells had positive growth across all of the tested substrates with no significant preference for a surface topography in comparison to the flat, silicon or glass backgrounds.


Franca et al.

Page 7

Author Manuscript Author Manuscript

The percentage of the total area covered by cells was calculated (Figure 5B) and subdivided into the area covered by neurites and somata. Only the SU-8 nanofiber groups had significantly higher cellular coverage on-pattern in comparison to the paired, off-pattern background (n = 5; *p < 0.01). There was also significantly higher cell coverage on SU-8 nanofibers than on all other on-pattern substrates (n = 5; *p < 0.05). Neurites account for the majority of this cellular coverage and the same statistical significances were observed for neurite coverage. The somata account for less than 10% of the total area covered in all cases and no statistical differences were observed. We also measured the total neurite length per area to quantify the length that neurites grew across each substrate with respect to total substrate area (Figure 5C). We found that the total neurite length per area is significantly higher on the SU-8 nanofibers with and without PEI (n = 5; p< 0.05). We also found that the TF-PEI group had significantly less neurite length per area in comparison to the respective off-pattern control (n = 5; p < 0.05). The results suggest that the size of the SU-8 nanofibers is ideal for more extensive neurite outgrowth; this shows that neurites prefer a growth substrate with physical features of a similar size. E. Neural Adhesion

Author Manuscript

Adhesion results revealed substantial differences among substrates (Figure 6). We tested CNF, NF, CNT-3, and CNT-9 within a parallel plate flow chamber to assess the ability of the neural network to resist shear flow on different substrates with highly contrasting roughnesses. We show that nearly half of all neurons on 0.1 % PEI control substrates remained post-shear stress. Both the CNT-3 and the CNT-9 were statistically indistinguishable from these controls. The CNF and NF samples treated both with and without 0.001 % PEI were also statistically similar to the 0.1 % control substrates. However, the neurons on the nanofibers were significantly more resistant to the shearing flow than the off-pattern controls, treated with the same PEI concentration (n = 4; *p < 0.005).

IV. Discussion A. Concentration-dependent PEI Toxicity on Patterned Substrates Cationic polymers, especially PEI, can be cytotoxic when exposed to cells in solution by way of disrupting the cellular membrane, causing membrane leakage, and the cessation of cellular metabolic activity, suggesting penetration through the cellular membrane [40]. This toxicity has also been shown to be dependent on the molecular weight of the PEI [49]. Though PEI in solution can be toxic, it is known to be an effective facilitator of neural cell growth at high and low molecular weights when immobilized on the culture surface following a thorough rinsing step [17, 41], as it is also evident in our study.

Author Manuscript

We found that all of our substrates supported viable neural cell outgrowth under most conditions tested. Surprisingly, some patterned samples treated with 0.1 % PEI (a common PEI concentration for supporting neural outgrowth) suffered from significantly decreased cell viability in contrast to off-pattern controls. Treated with 0.1 % PEI, patterned SU-8 nanofibers and thin films demonstrated cellular viabilities well below 20 %. These results suggest that SU-8 attracts more PEI molecules which can cause more cytotoxicity. To compare the levels of PEI on different substrates with varying PEI-treatment concentrations


Franca et al.

Page 8


we analyzed the zeta potential (Zp) of each substrate. We found that at 0.1 % PEI, SU-8 nanofibers and thin films have a significantly higher Zp than polished silicon with the Zp magnitude of the NFs over three times larger than the TFs. This effect is eliminated when reducing the PEI concentration to 0.001 %, resulting in similar Zp levels across all substrates. This means that there is a very strong positive charge on the NF and TF substrates when treated with higher concentrations of PEI, with an especially high charge on the nanofibers; this could be explained by more PEI bound to the NFs or TFs which would contribute to higher zeta potentials. The larger Zp values for the 0.1 % PEI-treated NFs suggests that the nanostructure of the nanofibers may trap or collect significantly more PEI molecules. The plasma-treated substrates demonstrated highly negative Zp values and with significant differences across all groups but will support neural growth for a limited period of time (without media conditioning) on all substrates. The short term cellular viability of the Plasma-treated substrates are more likely driven by increasing the hydrophilicity of the substrate. However, long-term viability likely requires a combination of increased hydrophilicity and positive surface charge which can be accomplished with a PEI treatment. These results suggest that there is an ideal range of surface charge which supports long-term neural viability, effectively a Zp near 0 mV. The results also suggest that a Zp which is very positive (> 60 mV) can result in a significantly more toxic surface.

Author Manuscript

Carbonized nanofibers also demonstrated a significantly lower cellular viability when treated with 0.1% PEI solution. However, the same effect was not observed on the carbonized thin film substrates as these substrates were similar to controls. Due to manufacturing limitations, we were unable to analyze the zeta potentials of the carbonized substrates. However, preliminary x-ray photoelectron spectroscopy (XPS) results (not shown) have suggested that carbonized nanofibers collect more PEI molecules when treated with higher PEI concentrations than off-pattern controls (determined by comparing nitrogen peaks which are exclusive to PEI molecules). These preliminary results and the lack of toxicity on carbonized thin films further suggest that the topography of the nanofibers (both SU-8 and carbon) has a significant effect on the level of PEI collected by the surface. Further experimentation is required to fully characterize the mechanism of this toxic phenomena as well as testing if similar results can be observed with other cationic polymers. B. Neural Growth Preference on Larger-scale Topography

Author Manuscript

No significant differences in cell densities after 7 DIV were reported (p = 0.058). Neural somata had positive outgrowth on all substrates without demonstrating a preference for any specific topography. This suggests that none of the nanomaterials strongly attracts or repels cell bodies. The neural cell bodies, at approximately 10 µm in diameter, are significantly larger than the topographical features of the nanofiber, nanotube, and thin film depositions. It is possible that somata growth is not significantly attracted or repelled by smaller topographical features, unlike what has been shown previously in neurites [15, 50–52]. However, there was a noticeable contrast between the on-pattern cell densities of the SU-8 thin film and all nanofiber groups (comparing Figure 4A and 4C to 4B). This contrast could prove to be useful in future experiments by pairing SU-8 thin film (as an insulator) and carbonized nanofibers (as an electrode material) to draw cells to adhere to the conductive, nano-fibrous CNF substrates; as the strongest extracellular signals are recorded from


Franca et al.

Page 9

Author Manuscript

electrodes in proximity to cell bodies, this could potentially yield an electrode capable of encouraging growth onto the electrode area and recording from more neurons.

Author Manuscript

We also wanted to observe how these materials affected the outgrowth of neurites. We found that the SU-8 nanofibers had a positive effect on the degree of neurite integration with the nanofiber matrix. After 7 DIV, neurites covered a significant percentage of the total nanofiber matrix, which also was accounted for in the total length of neurites per unit area of nanofiber matrix. Interestingly, this result was not replicated for any other substrate tested, including carbonized nanofibers and carbon nanotubes. This result suggests that neurites prefer to follow and grow on like-sized topographical cues on the surface. It has been shown previously that neurites like to follow large surface cues, such as walls, and features similar in size to neurites [15, 50–52]. However, it has also been shown that substrates with smallscale features (e.g. carbon nanotubes) can have a positive effect on cellular outgrowth [14, 53]. Considering these studies, we had expected to find an increase in neurite outgrowth on all of our tested nanofibers and nanotubes. However, the results show that neurite growth is most greatly enhanced by topographical features of a similar scale to the neurites; unfortunately this was only achieved on the non-conductive, SU-8 nanofibers.


Alternatively, it is possible that potential chemical (e.g. surface chemistry) or mechanical differences (e.g. modulus of elasticity) of the materials have a more significant effect on the neurite outgrowth. Prior to cell culture there are significant changes to the surface chemistry of the SU-8 and CNF substrates due to oxygen plasma treatment which includes increased wettability and increased acidic and carbon-based functional groups (e.g. carboxyl, carbonyl, etc.) [54, 55]. However, it has also been shown that, when exposed to a cell culture environment (e.g. culture media), the substrate continually accumulates a biofilm consisting of proteins and cell matter [56]. As a result, the unique surface chemistries on each of the substrates in this study are likely increasingly obfuscated by the buildup of a biofilm beginning with exposure to culture media. This may result in similar chemical profiles across all substrates (nanofibers and thin films) over time. Though this chemical profile will have an effect on cellular viability and outgrowth, it is likely that the large topographical differences between the substrates may have a stronger impact on the outgrowth observed in this study. Future studies will fully characterize the chemical profile of these nanofiber substrates in a cell culture environment at different time points. Further, it has been shown that the stiffness of a substrate can affect neural outgrowth and the differentiation of neural stem cells [57]. However, it has also been shown that embryonic cortical neural growth (like the neurons utilized in this work) is insensitive to substrate stiffness suggesting that this phenomena is cell-specific [58, 59]. Given that cortical neural growth has been shown to be relatively unaffected by the mechanical properties of a substrate, it is probable that, as suggested in previous literature [50–52], the size of the nanofibers have a stronger effect on the neurite outgrowth. This result suggests that increasing the size of the carbonized nanofibers (after shrinkage due to carbonization) would enhance the potential for electrodes to attract neurons and neurites.


Franca et al.

Page 10

C. Neurons Resist Shearing Flow on Rougher Substrates


The measurement of neural adhesion, by way of resistance to shear stress, has not been investigated extensively or at all. However, this assay can provide a unique method for observing how well neurons integrate with these substrates. In both PEI and Plasma samples, neurons located on both carbonized and SU-8 nanofiber substrates were significantly more resistant to the shear flow when compared to their paired controls. The adhesion of neurons on these samples treated with 0.001 % PEI is not significantly different from the 0.1 % PEI controls. However, the decrease in the concentration of PEI negatively affects cell adhesion on paired control substrates. This suggests that the nanofiber substrates are simply more effective at attracting the PEI molecules, thus resulting in cell adhesion similar to 0.1 % PEI controls. However, the adhesion of the neurons on the plasma-treated NF substrates was also statistically comparable to the 0.1 % controls. This suggests that the topography of the nanofiber substrates has some effect on the resistance of the neurons to the shearing flow.

Author Manuscript

These results demonstrate that the cells are able to resist shear stress more effectively on the nanofibers regardless of surface treatment. It is possible that the surface topography can act as a substitute for charge-based adhesion. By introducing a physical topography, less adhesion factor (e.g. PEI, PDL, PLL, etc.) is required to achieve similar results potentially providing more flexibility for in vitro and in vivo experimental designs. The improved adhesion on the nanofiber topography could be attributed to a couple of factors: 1) the neurons are able to anchor to a larger surface area resulting in stronger resistance to shear stress or 2) the neurons sink into the CNF matrix in which they avoid more of the shearing flow. Though this method has not demonstrated the mechanism of the neural resistance to shearing flow, either of the aforementioned factors contributing to the adhesion are positive in terms of the nanofibers’ usefulness as neural microelectrode materials. An ideal neural microelectrode, in addition to having attractive impedance and conductivity characteristics, would attract neural integration with the electrode. In either case, each neuron is in contact with a larger surface area on the nanofiber matrices as opposed to flat controls. This would potentially result in tighter coupling at the cell-electrode interface or more neural contact with electrode materials which would increase an electrode’s ability to collect neural signals and improve the signal-to-noise ratio. This cellular integration means that these nanofibers are of particular interest for designing future microelectrodes.

V. Conclusion

Author Manuscript

We have established that our substrates are viable for neural growth, the efficacy of which depends upon surface treatment and surface topography. We have shown that there is a significant correlation (Figure 5B–C) between the topographical scale of substrate features and neurite outgrowth. Correspondingly, the number of neurons resistant to shear stress on the rougher, larger scale nanofiber substrates was significantly higher than that of controls regardless of surface treatment. All of the results in this study suggest that some feature sizes and surface charges of rough substrates have a direct effect on neural outgrowth. In future experiments, a focus on increasing the size of the carbonized nanofibers will be emphasized to potentially replicate the extensive neurite outgrowth on the SU-8 nanofibers


Franca et al.

Page 11

Author Manuscript

(Figure 4A and Figure 5B–C). There may also be an opportunity to further explore the method of toxicity at specific concentrations of PEI; it is possible that these nanofibers slowly release PEI from within the interstices of the matrix, which may be utilized in controlled molecule or drug release applications. Ultimately, the carbonized nanofibers are most attractive as a microelectrode material and we have begun to analyze the electrical properties of the material in microelectrode form [45, 60]. The facilitation of neural outgrowth on microelectrodes consisting of carbonized nanofibers, with an optimized surface treatment and nanofiber scale, could result in more effective neural signal detection and generally improved microelectrodes. The neural outgrowth onto the microelectrodes could be further accentuated by utilizing SU-8 thin film as an insulator; this design would support neural outgrowth while encouraging more neural growth onto the nanofiber microelectrodes.

Author Manuscript

Acknowledgments This work was supported in part by the National Science Foundation (Grant No. 1132413) and the National Institutes of Health (research grant NS 052233). The authors would also like to thank the Department of Biomedical Engineering and the University of Florida for their continued support throughout the study. Additionally, the authors would like to extend special thanks to Dr. Edward Keefer, Dr. Thomas DeMarse, Valentina Rapazzetti, and Elizabeth Ankudowich for their valuable input and support during the design and implementation of this project.

References


1. Laxton AW, et al. A phase I trial of deep brain stimulation of memory circuits in Alzheimer’s disease. Annals of neurology. 2010; 68(4):521–534. [PubMed: 20687206] 2. Deuschl G, et al. A randomized trial of deep-brain stimulation for Parkinson’s disease. New England Journal of Medicine. 2006; 355(9):896–908. [PubMed: 16943402] 3. Mayberg HS, et al. Deep brain stimulation for treatment-resistant depression. Neuron. 2005; 45(5): 651–660. [PubMed: 15748841] 4. Grill WM, et al. Implanted neural interfaces: biochallenges and engineered solutions. Annual Review of Biomedical Engineering. 2009; 11:1–24. 5. Hubel DH, et al. Tungsten microelectrode for recording from single units. Science. 1957; 125(3247):549–550. [PubMed: 17793797] 6. Kozai TDY, Kipke DR. Insertion shuttle with carboxyl terminated self-assembled monolayer coatings for implanting flexible polymer neural probes in the brain. Journal of neuroscience methods. 2009; 184(2):199–205. [PubMed: 19666051] 7. Turner J, et al. Cerebral astrocyte response to micromachined silicon implants. Experimental neurology. 1999; 156(1):33–49. [PubMed: 10192775] 8. Lacour SP, et al. Flexible and stretchable micro-electrodes for in vitro and in vivo neural interfaces. Medical & biological engineering & computing. 2010; 48(10):945–954. [PubMed: 20535574] 9. Richardson-Burns SM, et al. Polymerization of the conducting polymer poly (3, 4ethylenedioxythiophene)(PEDOT) around living neural cells. Biomaterials. 2007; 28(8):1539–1552. [PubMed: 17169420] 10. Winter JO, et al. Neurotrophin-eluting hydrogel coatings for neural stimulating electrodes. Journal of Biomedical Materials Research Part B: Applied Biomaterials. 2007; 81(2):551–563. 11. Yoon I, et al. Intracellular Neural Recording with Pure Carbon Nanotube Probes. PLOS ONE. 2013; 8(6):e65715. [PubMed: 23840357] 12. Cyster L, et al. The effect of surface chemistry and nanotopography of titanium nitride (TiN) films on primary hippocampal neurones. Biomaterials. 2004; 25(1):97–107. [PubMed: 14580913] 13. Li S, et al. Aligned single-walled carbon nanotube patterns with nanoscale width, micron-scale length and controllable pitch. Nanotechnology. 2007; 18:455302.


Franca et al.

Page 12

Author Manuscript Author Manuscript Author Manuscript Author Manuscript

14. Zhang X, et al. Guided neurite growth on patterned carbon nanotubes. Sensors and Actuators B: Chemical. 2005; 106(2):843–850. 15. Dowell-Mesfin N, et al. Topographically modified surfaces affect orientation and growth of hippocampal neurons. Journal of neural engineering. 2004; 1:78. [PubMed: 15876626] 16. Fan Y, et al. Culture of neural cells on silicon wafers with nano-scale surface topograph. Journal of neuroscience methods. 2002; 120(1):17–23. [PubMed: 12351203] 17. Lelong IH, et al. Neuronal cells mature faster on polyethyleneimine coated plates than on polylysine coated plates. Journal of neuroscience research. 1992; 32(4):562–568. [PubMed: 1527802] 18. Rajnicek AM, et al. The direction of neurite growth in a weak DC electric field depends on the substratum: contributions of adhesivity and net surface charge. Developmental biology. 1998; 203(2):412–423. [PubMed: 9808790] 19. Johansson F, et al. Axonal outgrowth on nano-imprinted patterns. Biomaterials. 2006; 27(8):1251– 1258. [PubMed: 16143385] 20. Fan Y, et al. Adhesion of neural cells on silicon wafer with nano-topographic surface. Applied surface science. 2002; 187(3):313–318. 21. Khan SP, et al. Influence of nanoscale surface roughness on neural cell attachment on silicon. Nanomedicine: Nanotechnology, Biology and Medicine. 2005; 1(2):125–129. 22. Malarkey EB, et al. Conductive single-walled carbon nanotube substrates modulate neuronal growth. Nano letters. 2008; 9(1):264–268. [PubMed: 19143503] 23. Xie J, et al. Neurite outgrowth on nanofiber scaffolds with different orders, structures, and surface properties. ACS nano. 2009; 3(5):1151–1159. [PubMed: 19397333] 24. Christopherson GT, et al. The influence of fiber diameter of electrospun substrates on neural stem cell differentiation and proliferation. Biomaterials. 2009; 30(4):556–564. [PubMed: 18977025] 25. Mitra J, et al. Patterned growth and differentiation of neural cells on polymer derived carbon substrates with micro/nano structures in vitro. Carbon. 2013; 65:140–155. 26. Jang MJ, et al. Directional neurite growth using carbon nanotube patterned substrates as a biomimetic cue. Nanotechnology. 2010; 21:235102. [PubMed: 20463384] 27. Gabay T, et al. Electro-chemical and biological properties of carbon nanotube based multielectrode arrays. Nanotechnology. 2007; 18:035201. [PubMed: 19636111] 28. Keefer EW, et al. Carbon nanotube coating improves neuronal recordings. Nature nanotechnology. 2008; 3(7):434–439. 29. Gabriel G, et al. Single-walled carbon nanotubes deposited on surface electrodes to improve interface impedance. Physiological measurement. 2008; 29:S203. [PubMed: 18544808] 30. Webster TJ, et al. Nano-biotechnology: carbon nanofibres as improved neural and orthopaedic implants. Nanotechnology. 2004; 15(1):48. 31. Jan E, et al. Layered carbon nanotube-polyelectrolyte electrodes outperform traditional neural interface materials. Nano letters. 2009; 9(12):4012–4018. [PubMed: 19785391] 32. Fabbro A, et al. Spinal Cord Explants Use Carbon Nanotube Interfaces to Enhance Neurite Outgrowth and to Fortify Synaptic Inputs. ACS nano. 2012 33. Cellot G, et al. Carbon Nanotube Scaffolds Tune Synaptic Strength in Cultured Neural Circuits: Novel Frontiers in Nanomaterial-Tissue Interactions. The Journal of Neuroscience. 2011; 31(36): 12945–12953. [PubMed: 21900573] 34. Lovat V, et al. Carbon nanotube substrates boost neuronal electrical signaling. Nano letters. 2005; 5(6):1107–1110. [PubMed: 15943451] 35. Mazzatenta A, et al. Interfacing neurons with carbon nanotubes: electrical signal transfer and synaptic stimulation in cultured brain circuits. The Journal of neuroscience. 2007; 27(26):6931– 6936. [PubMed: 17596441] 36. Malarkey EB, et al. Water soluble single-walled carbon nanotubes inhibit stimulated endocytosis in neurons. Nano letters. 2008; 8(10):3538–3542. [PubMed: 18759491] 37. McKenzie JL, et al. Decreased functions of astrocytes on carbon nanofiber materials. Biomaterials. 2004; 25(7):1309–1317. [PubMed: 14643605]


Franca et al.

Page 13


38. Fabbro A, et al. Adhesion to carbon nanotube conductive scaffolds forces action-potential appearance in immature rat spinal neurons. PloS one. 2013; 8(8):e73621. [PubMed: 23951361] 39. Cellot G, et al. Carbon nanotubes might improve neuronal performance by favouring electrical shortcuts. Nature nanotechnology. 2008; 4(2):126–133. 40. Fischer D, et al. In vitro cytotoxicity testing of polycations: influence of polymer structure on cell viability and hemolysis. Biomaterials. 2003; 24(7):1121–1131. [PubMed: 12527253] 41. Lakard S, et al. Adhesion and proliferation of cells on new polymers modified biomaterials. Bioelectrochemistry. 2004; 62(1):19–27. [PubMed: 14990322] 42. Pei R, et al. Assembly of alternating polycation and DNA multilayer films by electrostatic layerby-layer adsorption. Biomacromolecules. 2001; 2(2):463–468. [PubMed: 11749207] 43. Brunot C, et al. Cytotoxicity of polyethyleneimine (PEI), precursor base layer of polyelectrolyte multilayer films. Biomaterials. 2007; 28(4):632–640. [PubMed: 17049374] 44. Jao PF, et al. Fabrication of an all SU-8 electrospun nanofiber based supercapacitor. Journal of Micromechanics and Microengineering. 2013; 23(11):114011. 45. Jao PF, et al. Immersion lithographic patterning of electrospun nanofibers for carbon nanofibrous microelectrode arrays. IEEE Journal of Microelectromechanical Systems. 2015; 24(3):703–715. 46. Necas D, Klapetek P, et al. Gwyddion: an open-source software for SPM data analysis. Central European Journal of Physics. 2012; 10(1):181–188. 47. Schneider CA, et al. NIH Image to ImageJ: 25 years of image analysis. Nat Methods. 2012; 9(7): 671–675. [PubMed: 22930834] 48. Dehmelt L, et al. NeuriteQuant: An open source toolkit for high content screens of neuronal Morphogenesis. BMC neuroscience. 2011; 12(1):100. [PubMed: 21989414] 49. Morimoto K, et al. Molecular weight-dependent gene transfection activity of unmodified and galactosylated polyethyleneimine on hepatoma cells and mouse liver. Molecular Therapy. 2003; 7(2):254–261. [PubMed: 12597914] 50. Rajnicek A, et al. Contact guidance of CNS neurites on grooved quartz: influence of groove dimensions, neuronal age and cell type. Journal of cell science. 1997; 110(23):2905–2913. [PubMed: 9359873] 51. Sørensen A, et al. Long-term neurite orientation on astrocyte monolayers aligned by microtopography. Biomaterials. 2007; 28(36):5498–5508. [PubMed: 17905429] 52. Nisbet D, et al. Interaction of embryonic cortical neurons on nanofibrous scaffolds for neural tissue engineering. Journal of neural engineering. 2007; 4:35. [PubMed: 17409478] 53. Clark P, et al. Cell guidance by ultrafine topography in vitro. Journal of cell science. 1991; 99(1): 73–77. [PubMed: 1757503] 54. Bubert H, et al. Basic analytical investigation of plasma-chemically modified carbon fibers. Spectrochimica Acta Part B: Atomic Spectroscopy. 2002; 57(10):1601–1610. 55. Walther F, et al. Surface hydrophilization of SU-8 by plasma and wet chemical processes. Surface and Interface Analysis. 2010; 42(12–13):1735–1744. 56. Branch DW, et al. Long-term stability of grafted polyethylene glycol surfaces for use with microstamped substrates in neuronal cell culture. Biomaterials. 2001; 22(10):1035–1047. [PubMed: 11352085] 57. Georges PC, et al. Matrices with compliance comparable to that of brain tissue select neuronal over glial growth in mixed cortical cultures. Biophysical journal. 2006; 90(8):3012–3018. [PubMed: 16461391] 58. Palchesko RN, et al. Development of polydimethylsiloxane substrates with tunable elastic modulus to study cell mechanobiology in muscle and nerve. PLoS ONE. 2012; 7(12):e51499. [PubMed: 23240031] 59. Moore SW, Sheetz MP. Biophysics of substrate interaction: influence on neural motility, differentiation, and repair. Dev Neurobiol. 2011; 71(11):1090–1101. [PubMed: 21739614] 60. Jao, PF., et al. Fabrication of carbon nanofibrous microelectrode array (CNF-MEA) using nanofiber immersion photolithography; Micro Electro Mechanical Systems (MEMS), 2014 IEEE 27th International Conference on; 2014. p. 498-501.


Franca et al.

Page 14

Author Manuscript Author Manuscript Author Manuscript

Figure 1.

SEM micrographs of nanomaterials. Variations in scale can be seen between (A) CNT-9, (B) CNT-3, (C) carbonized nanofibers, and (D) SU-8 nanofibers (scale bar = 3 µm). Insets of (C–D) are patterned carbonized and SU-8 nanofibers, respectively (scale bar = 200 µm). All nanofiber and thin film samples are patterned similarly.

Author Manuscript IEEE Trans Nanobioscience. Author manuscript; available in PMC 2017 January 25.

Franca et al.

Page 15

Author Manuscript Author Manuscript Author Manuscript Figure 2.

Author Manuscript

Nanofibers and thin films treated with 0.1 % PEI have a significantly higher positively charged zeta potential values. The measured zeta potentials for varying surface treatments on SU-8 nanofibers, thin films, and silicon controls. In the plasma and 0.1 % PEI groups, all surfaces tested were significantly different from each other (*p < 0.001). All surfaces were treated with O2 plasma prior to any PEI treatment.


Franca et al.

Page 16

Author Manuscript Author Manuscript Figure 3.

Author Manuscript

The concentration of PEI directly affects the viability of neurons on- and off-pattern across substrates. The cell viabilities for CNT-3 and CNT-9 are comparable to their respective controls. The on-pattern and off-pattern cell viabilities of the CNF, NF, CTF, and TF groups depend greatly on the concentration of the PEI treatment (n = 4; *p < 0.01, †p < 0.05).


Franca et al.

Page 17

Author Manuscript Author Manuscript Author Manuscript

Figure 4.

Neurite outgrowth on 0.001 % PEI-treated substrates varies depending on substrate topography. Contrasting neural outgrowth can be observed between (A) SU-8 nanofibers (NF), (B) SU-8 thin film (TF), (C) carbonized nanofibers (CNF), and (D) carbon thin film (CTF). Silicon (Si) control areas are indicated next to the nanofiber and thin film patterns. No significant differences were detected in neural cells per unit area on- and off-pattern across all substrates. The neurite coverage on the (A) SU-8 nanofibers is significantly higher than off-pattern controls and most other substrates


Franca et al.

Page 18


Figure 5.

Neurite outgrowth is significantly higher on SU-8 nanofibers with and without PEI treatment (*p < 0.05). (A) There was no significance detected for the number of cell bodies per mm2. (B) The percentage of cell coverage shows significantly higher outgrowth on SU-8 nanofibers. (C) The total length of neurites per unit area shows that the neurite outgrowth on SU-8 nanofibers was significantly higher than off-pattern controls; additionally, there was significantly less neurite length per area on PEI-treated SU-8 thin films.


Franca et al.

Page 19

Author Manuscript Author Manuscript Author Manuscript Figure 6.

Adhesion was significantly higher on all nanofiber substrates in comparison to paired controls regardless of surface treatment (*p < 0.005). The CNT groups were statistically similar to their 0.1 % PEI controls.


Franca et al.

Page 20

TABLE I

Author Manuscript

Average Roughness and Feature Width for Each Substrate Sample

Average Roughness (nm)

Average Feature Width (nm)

Carbon Nanotubes (3)

12.4 ± 1.12

34.2 ± 4.15

Carbon Nanotubes (9)

16.9 ± 1.89

34.0 ± 4.25

Carbonized Nanofibers

164 ± 19.6

286 ± 58.0

SU-8 Nanofibers

83.7 ± 11.7

781 ± 187

Carbon Thin Film

0.27 ± 0.02

N/A

SU-8 Thin Film

0.72 ± 0.07

N/A

Average Neurite†

N/A

850 ± 101

*

Carbon Nanotubes (3) and (9) refer to carbon nanotube samples with 3 and 9 bilayers, respectively.

Author Manuscript

†

For a biological size reference, the average width of a cortical neurite was measured with fluorescence images in ImageJ.

Author Manuscript Author Manuscript IEEE Trans Nanobioscience. Author manuscript; available in PMC 2017 January 25.

Carbon nanomaterials.

Ganglioside modulation of neural cell adhesion molecule and N-cadherin-dependent neurite outgrowth.

Synergic interaction between amyloid precursor protein and neural cell adhesion molecule promotes neurite outgrowth.

Enhanced neural cell adhesion and neurite outgrowth on graphene-based biomimetic substrates.

Organometallic and coordination chemistry of carbon nanomaterials.

Carbon and fullerene nanomaterials in plant system.

Ice Nucleation Properties of Oxidized Carbon Nanomaterials.

Films of Carbon Nanomaterials for Transparent Conductors.

Advanced glycation endproducts interfere with adhesion and neurite outgrowth.

Amine functionalized nanodiamond promotes cellular adhesion, proliferation and neurite outgrowth.

Fluorescent carbon nanomaterials: "quantum dots" or nanoclusters?

Calcium-Mediated Adhesion of Nanomaterials in Reservoir Fluids.

Recent applications of carbon nanomaterials in fluorescence biosensing and bioimaging.

Synthesis and Modification of Carbon Nanomaterials utilizing Microwave Heating.

Comment on 'Carbon and fullerene nanomaterials in plant system'.

Carbon-based smart nanomaterials in biomedicine and neuroengineering.

Carbon nanomaterials for nerve tissue stimulation and regeneration.

Superior piezoelectric composite films: taking advantage of carbon nanomaterials.

Biophysical influence of airborne carbon nanomaterials on natural pulmonary surfactant.

Guiding osteogenesis of mesenchymal stem cells using carbon-based nanomaterials.

Carcinogenic Potential of High Aspect Ratio Carbon Nanomaterials.

Laser-Plasma Driven Synthesis of Carbon-Based Nanomaterials.

Dynamic In-Situ Experimentation on Nanomaterials at the Atomic Scale.

Surface Curvature Relation to Protein Adsorption for Carbon-based Nanomaterials.