Biomimetic surface delivery of NGF and BDNF to enhance neurite outgrowth

Treatment for peripheral nerve injuries includes the use of autografts and nerve guide conduits (NGCs). However, outcomes are limited, and full recovery is rarely achieved. The use of nerve scaffolds as a platform to surface immobilize neurotrophic factors and deliver locally is a promising approach to support neurite and nerve outgrowth after injury. We report on a bioactive surface using functional amine groups, to which heparin binds electrostatically. X‐ray photoelectron spectroscopy analysis was used to characterize the presence of nitrogen and sulfur. Nerve growth factor (NGF) and brain‐derived neurotrophic factor (BDNF) were bound by electrostatic interaction to heparin, and the release profile evaluated by enzyme‐linked immunosorbent assay, which showed that ca. 1% of NGF was released from each of the bioactive surface within 7 days. Furthermore, each surface showed a maximum release of 97% of BDNF. Neurotrophin release on neurite outgrowth was evaluated by primary dorsal root ganglion with a maximum neurite growth response in vitro of 1,075 µm detected for surfaces immobilized with NGF at 1 ng/ml. In summary, the study reports on the design and construction of a biomimetic platform to deliver NGF and BDNF using physiologically low concentrations of neurotrophin. The platform is directly applicable and scalable for improving the regenerative ability of existing NGCs and scaffolds.

Hence, for NGC functionality, it is important to design a sustained delivery system for neurotrophic factors (Achyuta, Cieri, Unger, & Murthy, 2009;Gomez & Schmidt, 2007). Electrochemistry, microfluidics, lithography, and microspheres have been used to immobilize neurotrophins in different delivery systems, including NGCs. These techniques can be disadvantageous, because they require complex experimental design and fabrication, leading to long preparation periods, limiting clinical application (Tang et al., 2013) and effectiveness on nerve regeneration (Madduri et al., 2010;Moran & Graeber, 2004;Sumner, 1990;Tang et al., 2013). Alternatively, development of bioactive surfaces in direct contact with regenerating axons will target the delivery of neurotrophic factors more precisely, with direct stimulation for neurite outgrowth.
Biomaterial surfaces are essential for interaction with the environment (Ratner, 2013), and several approaches exist to stimulate biological response (Jones, 2005). Bioactive surfaces as a delivery platform have advantages of controlled, targeted delivery, sustained release, and low dosage. A range of examples are reported, for example, Puleo, Kissling, and Sheu (2002) immobilized bone morphogenetic protein onto titanium alloy to promote osteoblastic activity. Ito, Chen, and Imanishi (1998) encouraged fibroblast growth by immobilizing epidermal growth factor onto polystyrene.
We have developed a glycosaminoglycan surface platform for growth factor delivery using heparin, through transient charge attraction of negative functional groups and local release (WO 2017(WO / 017425 A1, 2017. For this particular study, we employed a commercial source of amine functionalized surfaces that arose directly from this prior study (Robinson et al., 2012). As a method to permit charge attraction, positive surface amine functional groups (NH 2 + ) are created as a basis to attract and bind negatively charged heparin by adsorption to a biomaterial surface of interest (Cao & Li, 2011).
We have not previously explored this system for neurotrophin binding and release; however, it is reported that neurotrophins bind to heparin (Casu, Naggi, & Torri, 2015;D. Yang, Moon, & Lee, 2017).
The advantage of this method is that it would build the necessary electrostatic positive charge to immobilize neurotrophins, by means of electrostatic interactions between NH 2 + groups and negatively charged heparin, and by the heparin binding site of neurotrophins.
Investigation of a neurotrophin delivery platform is logical to explore for neurite outgrowth by the immobilized of NGFs onto NH 2 + + heparin surfaces for local and sustained release.

| Statistical analysis
One-way analysis of variance with a Tukey multiple comparison posttest was performed to identify statistical differences between conditions, using GraphPad Prism 8.2.0. p < .05 was considered significant.

| Characterization of surfaces by water contact angle
Water contact angles were measured by sessile drop as a measure of relative hydrophilicity (Liu et al., 2014). The mean contact angles of polystyrene (PS), TCP, TCP + NH 2 + , and TCP + NH 2 + + heparin were 98, 70, 43, and 39°, respectively ( Figure 2). The decrease in the contact angle confirmed the presence of charged functional groups at the surface, such as amine (NH 2 + ) and sulfate. The lowest contact angle was observed for TCP + NH 2 + + heparin. PS control angles were significantly different compared to TCP + NH 2 + and TCP + NH 2 + + heparin surfaces (****p < .0001). TCP angles were also significantly different when compared with TCP + NH 2 + and TCP + NH 2 + + heparin surfaces (***p < .001 and ****p < .0001, respectively). There was no difference between the contact angles for TCP + NH 2 + and TCP + NH 2 + + heparin.

| Characterization surfaces by XPS
XPS analysis was used to determine elemental composition of surfaces after each modification step. Chemical survey scans of each are shown in Figure 1. TCP + NH 2 + surfaces and TCP + NH 2 + + heparin surfaces identified a peak at 400 eV, confirming the presence of nitrogen with 12.6% and 9.15% of atomic composition (Figure 1c and 1g and Figure S1A). Furthermore, N1s scan revealed a single peak, which was attributed to the NH 2 + content on the TCP + NH 2 + surface and TCP + NH 2 + + heparin surface ( Figure S1B and S1C), as TCP has a negligible nitrogen content (Koller, Palsson, Manchel, Maher, & Palsson, 1998). Moreover, the decreased nitrogen content in the TCP + NH 2 + + heparin surface could possibly be due to the loss of nitrogen after the incubation step of heparin, during the rinsing step (Dehili, Lee, Shakesheff, & Alexander, 2006). An S2p peak was detected for the TCP + NH 2 + + heparin surface at 168 eV ( Figure 1g). Figure S1D shows the S2p scan with S2p 3/2 and S2p 1/2 . This confirmed the presence of sulfur, and hence heparin.
3.3 | Release of NGF and BDNF from NH 2 + + heparin surfaces NGF release from NH 2 + + heparin surfaces was measured over 168 hr by ELISA (Figure 1j and    Tang et al., 2013). In these studies, neurotrophins were associated with biomaterial surfaces using different approaches. Interestingly, the concentration of neurotrophin required was higher than those used in the present study. Gomez and Schmidt (2007) immobilized NGF at 1 ng/mm 2 on polypyrrole by a photochemical technique with average neurite lengths of 20 µm after 2 days of rat DRG culture. Tang et al. (2013) used a silk fibroin coating to immobilize NGF as a gradient on poly(ɛcaprolactone)-poly(L-lactic) acid. The cumulative release of NGF ranged from 3.6 to 11.35 ng/ml within 7 days. Neurite lengths of DRGs after 3 days in culture varied according to the gradient zone, from 600 to 1,500 µm (Tang et al., 2013). In contrast, surfaces created herein formed a platform to deliver immobilized neurotrophins after 7 days release using less than 1 ng/ml from 1 to 168 hr. Maximum average neurite length of 1,075 µm was observed when surfaces were immobilized with NGF at 1 ng/ml. This suggests that neurotrophin delivery is more efficient by using growth factors at less than 1 ng/ml for stimulating neurite lengths greater than 1,000 µm in vitro.
Evidence in support of functional surface characterization included water contact angle, initially conducted to evaluate relative wettability of the modified material. Surface wettability will change due to the presence of positively charged amine groups (Li, Li, Yu, & Sun, 2017) or negatively charged heparin (Cao & Li, 2011;Mascotti & Lohman, 1995;T. Yang et al., 2006 Note: This concentration is also presented as a percentage calculated from the initial immobilized neurotrophin load. Abbreviations: BDNF, brain-derived neurotrophic factor; NGF, nerve growth factor. Three experiments were performed independently; three DRGs were seeded per condition (N = 3 and n = 3). One-way ANOVA with Tukey multiple comparison test was performed (*p < .05, **p < .01, ***p < .001, and ****p < .0001). ANOVA, analysis of variance; BDNF, brain-derived neurotrophic factor; NGF, nerve growth factor; TCP, tissue culture polystyrene We hypothesized that heparin binds electrostatically to surface amine groups through attractive electrostatic force and have developed a glycosaminoglycan surface platform for growth factor delivery using heparin, through transient charge attraction of negative functional groups and local release (WO 2017/017425 A1, 2017).
For this particular study, we employed a commercial BD PureCoat™ source of amine functional surfaces that arose from a prior study (Robinson et al., 2012). This study precedes and underpins the application reported herein. The amine coated plates were used to fabricate the heparin + neurotrophin layers. This was the first step to study the effectiveness of this delivery platform on neurite outgrowth. XPS analysis confirmed the presence of amine groups, and separately the presence of sulfur for heparin at the modified surface.
This was paramount because to scale and apply this bioactive surface to other biomaterials, we need to replicate the nitrogen content on the commercially amine coated plates to replicate the results presented in this study. Therefore, we will aim to incorporate NH 2 + groups by plasma polymerization, as this process has effectively deposited amine groups in different materials (Beck et al., 2005;Dehili et al., 2006;Smith et al., 2016). The formation of an NH 2 + + heparin surface was used as a basis to permit NGF and BDNF immobilization in the heparin layer, and a platform the study of enhanced neurite outgrowth when in direct contact.
Studies have previously reported on chemically binding heparin to coatings and scaffolds (Kim, Kang, Huh, & Yoon, 2000;Liu et al., 2014;Zhou & Meyerhoff, 2005). Gigliobianco, Chong, and Macneil (2015) bound heparin to alternate layers of polyethyleneimine (positively charged) and heparin or polyethyleneimine and polyacrylic acid by repeated rounds of electrostatic adsorption using acrylic acid plasma deposition (known to have negatively charged carboxyl groups; Campbell & Farrel, 2008). Their results showed that heparin bound to alternate layers of polyethyleneimine and acrylic acid. Moreover, heparin bound to positively charged polyethyleneimine after seven layers of coating, whereas no heparin was bound after five layers of coating (Gigliobianco et al., 2015). This finding raises the question as to whether heparin could bind to other positively charge functional groups, such as NH 2 + . The atomic percentage of sulfur (found in heparin) present in seven layers of polyethyleneimine was 0.53% (Gigliobianco et al., 2015). In comparison, the present study demonstrated an atomic percentage of 0.69% for sulfur on TCP + NH 2 + + heparin surfaces. This supported the hypothesis that heparin was bound to the positively charged amine groups at the surface.
Heparin is also known to bind growth factors (Meneghetti et al., 2015;Sakiyama-Elbert & Hubbell, 2000), including NGF and BDNF (Sakiyama-Elbert & Hubbell, 2000), albeit with moderate affinity. Considering these studies, we hypothesized that NGF and BDNF would bind sufficiently through electrostatic and noncovalent interactions, and with overall surface energy to immobilize. ELISA analysis for each neurotrophin was used to investigate if the growth factors were eluted over a timeframe consistent with nerve cell interaction and growth at 1, 24, 48, and 168 hr. Results showed that for each time point, the release was <35% of the total growth factor immobilized onto the surface. Interestingly, surfaces immobilized with NGF at 1 ng/ml did not show any release until Day 7 (168 hr).
This kinetic profile was of note and was hypothesized as advantageous in regard to neural implant applications for repair, mapping broadly to the initial stages of nerve repair following injury.
According to Roach, Eglin, Rohde, and Perry (2007), after implantation of a biomaterial, water molecules form an intimate layer.
Proteins then adsorb on the material surface, and adherent cells that express functionally binding integrins reach the surface and start to interact with the protein layer. Thereafter, cells adhere, migrate, and differentiate. This last step occurs from a few hours to a few days after implantation. As functionalized surfaces containing immobilized NGF at 1 ng/ml did not show any growth factor release between 1 and 48 hr, therefore, such a delivery platform could permit time for neuronal cells to adhere to a regenerating guidance scaffold surface before receiving stimuli from NGF at the surface to stimulate growth.
Chick embryo DRGs were cultured on surfaces containing immobilized neurotrophins. Neurite development is shown in Figure 4.
NGF-TrkA complexes are internalized at the neurite growth cones and retrogradely transported to the soma, inducing microtubule and actin polymerization, calcium influx, and organelle recruitment (Kapur & Shoichet, 2004). However, when NGF-TrkA complexes are not internalized, activation of the phosphatidylinositol 3-kinase and Akt/protein kinase B (PI3K/Akt) pathways arise, which regulate cell survival and neurite extension (Gomez & Schmidt, 2007;Madduri et al., 2009;Tang et al., 2013). When BDNF binds to TrkB receptors, the PI3K/Akt pathway is triggered and microtubules are rearranged to encourage filopodial elongation and lamellipodial formation (Mcgregor & English, 2019 Moreover, high doses of BDNF would inhibit neurite elongation (Santos et al., 2016), whereas low doses of BDNF stimulate neurite outgrowth in sensory neurons (e.g., DRGs). Thus, when delivered with NGF, an equilibrium would be expected.
In summary, we report on the design, manufacture, physicochemical and biological characterization of surfaces immobilized with NGF and BDNF using commercially amine coated plates and heparin.
NGF loaded surfaces at 1 ng/ml encouraged longest neurite lengths due to (a) DRGs having sufficient time to attach to surfaces and (b) neurotrophin release being controlled over a similar timeframe. DRG neurites most likely formed via activation of PI3K through membrane activation of NGF-TrkA. The system reported is relatively simple to produce, and scalable for direct applications, using plasma polymerization, for enhancing the function of NGCs to enhance nerve growth following acute injury.

| CONCLUSION
This study reports on the design and construction of bioactive surfaces with heparin bound electrostatically to amine groups, that can be used to immobilize NGF and BDNF. The surface was used as a model platform for delivery of NGF and BDNF. Bioactive surfaces stimulated neurite outgrowth in DRGs using very low concentrations of neurotrophin, making this approach directly applicable and scalable for improving the function of existing NGCs.

ACKNOWLEDGMENTS
The authors thank Dr. Deborah Hammond for contribution on the XPS study and Dr. Gabriella Kakonyi for technical support of the contact angle study. Furthermore, they also thank CONACyT for financially supporting AMSC. All authors contributed equally to this study.