Development of Two-Layer Hybrid Scaffolds Based on Oxidized Polyvinyl Alcohol and Bioactivated Chitosan Sponges for Tissue Engineering Purposes

Oxidized polyvinyl alcohol (OxPVA) is a new polymer for the fabrication of nerve conduits (NCs). Looking for OxPVA device optimization and coupling it with a natural sheath may boost bioactivity. Thus, OxPVA/chitosan sponges (ChS) as hybrid scaffolds were investigated to predict in the vivo behaviour of two-layered NCs. To encourage interaction with cells, ChS were functionalized with the self-assembling-peptide (SAP) EAK, without/with the laminin-derived sequences -IKVAV/-YIGSR. Thus, ChS and the hybrid scaffolds were characterized for mechanical properties, ultrastructure (Scanning Electron Microscopy, SEM), bioactivity, and biocompatibility. Regarding mechanical analysis, the peptide-free ChS showed the highest values of compressive modulus and maximum stress. However, among +EAK groups, ChS+EAK showed a significantly higher maximum stress than that found for ChS+EAK-IKVAV and ChS+EAK-YIGSR. Considering ultrastructure, microporous interconnections were tighter in both the OxPVA/ChS and +EAK groups than in the others; all the scaffolds induced SH-SY5Y cells’ adhesion/proliferation, with significant differences from day 7 and a higher total cell number for OxPVA/ChS+EAK scaffolds, in accordance with SEM. The scaffolds elicited only a slight inflammation after 14 days of subcutaneous implantation in Balb/c mice, proving biocompatibility. ChS porosity, EAK 3D features and neuro-friendly attitude (shared with IKVAV/YIGSR motifs) may confer to OxPVA certain bioactivity, laying the basis for future appealing NCs.


Introduction
Neurons' limited ability in self-regeneration is a significant issue in case of severe peripheral nerve injuries (PNIs). To date, despite advances in microsurgical techniques, peripheral nerve recovery remains a challenge and complete sensory/motor function restoration has never been achieved [1]. Hence, to support and guide a successful axonal nerve growth without resorting to autologous grafts (limited availability, comorbidities) or allografts/xenografts (possible immunological reactions), "on-the-bench" nerve conduits (NCs) have long been investigated [2]. Currently, the Food and Drug Administration (FDA) has approved 11 natural (collagen type I)/synthetic (polyglycolic acid, poly(D,L-lactideco-ε-caprolactone), polyvinyl alcohol) polymer-based conduits for PNIs repair but in vivo bound to the IKVAV (EAK-IKVAV) or YIGSR (EAK-YIGSR) laminin-derived adhesive sequences, were considered. ChS porosity (ultrastructure), together with EAK 3D intrinsic features and neuro-friendly attitude also shared with the IKVAV and YIGSR motifs, may confer to OxPVA a certain bioactivity, thus laying the basis for future appealing NCs to fully respond and satisfy both surgeons' and patients' expectations. Furthermore, heterotopic implant-based evidence also confirmed in vitro data, corroborating the great potential of these hybrid scaffolds for the development of innovative devices.

Mechanical Analysis of Chitosan-Based Sponges
Before the fabrication of the hybrid scaffolds, mechanical compression tests were performed on the four groups of chitosan-based sponges (ChS; ChS+EAK; ChS+EAK-IKVAV; ChS+EAK-YIGSR). The scaffolds showed a behaviour similar to that of a flexible foam, and the stress-strain curves are reported in Figure 1. The compressive modulus (E) was calculated as the slope of the initial linear region of the stress-strain curve. Table 1 shows the compressive modulus (E) and maximum stress (σ max ) as mean value ± standard deviation. With regard to the statistical analysis, the p-value corresponding to the F-statistic of ANOVA was much lower than 0.05 (and 0.01), suggesting that one or more pairs of treatments were significantly different. Thus, Tukey's HSD post hoc test was used to identify which of the pairs of treatments were significantly different from each other ( Table 2). Table 2. Statistical analysis on mechanical properties of ChS, ChS+EAK, ChS+EAK-IKVAV, and ChS+EAK-YIGSR scaffolds: Tukey's HSD results.
The compressive modulus and the maximum stress significantly decreased from 40.4 ± 3.1 kPa to 31.1 ± 2.1 kPa and from 16.8 ± 1.0 kPa to 11.9 ± 0.6 kPa, respectively, due to the inclusion of the EAK to ChS (Tables 1 and 2). Furthermore, if compared to ChS, a significant decrease in compressive modulus and maximum stress also occurred in the presence of the adhesive sequences IKVAV and YIGSR bonded to EAK (Tables 1 and 2). In particular, ChS+EAK-IKVAV and ChS+EAK-YIGSR showed the values of a compressive modulus (30.9 ± 1.9 kPa and 30.6 ± 2.0 kPa, respectively) and maximum stress (7.3 ± 0.4 kPa and 7.2 ± 0.4 kPa, respectively), which were significantly lower than those observed for ChS (Tables 1 and 2). However, as regards compressive modulus and maximum stress, no statistically significant differences were found between ChS+EAK-IKVAV and ChS+EAK-YIGSR (Tables 1 and 2). Furthermore, with regard to +EAK groups, even though the maximum stress found in the case of ChS+EAK (11.9 ± 0.6 kPa) was significantly higher than that obtained for ChS+EAK-IKVAV (7.3 ± 0.4 kPa) and ChS+EAK-YIGSR (7.2 ± 0.4 kPa), in terms of compressive modulus no statistically significant differences were observed among these three groups (Tables 1 and 2). Accordingly, the adhesive sequences IKVAV and YIGSR bonded to EAK negatively affected the maximum stress of ChS+EAK without significantly altering the compressive modulus.

Fabrication and Morphology of OxPVA/ChS Hybrid Scaffolds
A weighted amount of chitosan, dissolved in acetic acid solution without/with the adhesive sequences (+EAK; +EAK-IKVAV; +EAK-YIGSR) was prepared to fabricate the sponges by freeze-drying and transferring to moulds (24-well plate).
Lyophilized chitosan-based sponges resulted in discoidal, white matrices with a diameter of 1.5 cm, compatible with the dimension of the mould used for their fabrication. Macroscopically, the ChS+EAK-IKVAV showed a less tight network than that of the other experimental groups, in accordance with the investigated mechanical properties (see Section 2.1).
After fabrication, the sponges were carefully positioned over the OxPVA polymer solution and the physical cross-linking occurred by six freezing-thawing (FT) cycles (−80 • C for 1 h/RT for 1 h), thus leading to the stable coupling of the two layers (Figure 2A-D). The protocol for the scaffolds' development preserved sponges' integrity: OxPVA solution did not impregnate and/or cover the chitosan seeding surfaces. Additionally, the synthetic polymer was guaranteed for the high manipulability of the chitosan matrices. (I) Pore density, calculated as total pores/100 µm 2 area (*** p < 0.001; **** p < 0.0001) and (J) mean pore area [µm 2 ] (** p < 0.01; **** p < 0.0001), by ImageJ software.

Bioactive Potential of the Hybrid Scaffolds
The enhanced bioactivity of the OxPVA/chitosan scaffolds was assessed by testing their interaction with SH-SY5Y cells. After cell seeding on the hybrid scaffolds, the MTT assay was performed at 3 and 7 days to evaluate cell viability and growth on the supports. Hence, eventual differences in cells adhesion/proliferation induced by ChS bioactivation (+EAK, +EAK-IKVAV, or +EAK-YIGSR) were verified vs. the peptide-free group.
As showed in Figure 3A,B, the chitosan-based layers without/with biochemical cues were all able to induce cell adhesion and proliferation over OxPVA supports. This is interesting evidence; in fact, as broadly demonstrated, smooth OxPVA scaffolds did not trigger cell adhesion and proliferation in vitro [10,11,14]. Initially, at 72 h from seeding, a certain growth trend was detected in the whole cohort, but no specific differences were identifiable when comparing the groups ( Figure 3A). The safeness of the scaffolds' preparation phases was also corroborated by these data. At day 7 from seeding, some statistical differences emerged ( Figure 3B). Considering the proliferation data, the higher total cells number was observed on OxPVA/ChS+EAK (p < 0.05 vs. OxPVA/ChS+EAK-IKVAV; p < 0.01 vs. OxPVA/ChS+EAK-YIGSR) followed by OxPVA/ChS (p < 0.05 vs. OxPVA/ChS +EAK-YIGSR).

SEM Analysis of Cells Distribution on OxPVA/Chitosan-Based Scaffolds
To qualitatively support the MTT assay results, the presence of SH-SY5Y cells on OxPVA/Chitosan-based supports was also investigated by SEM at 72 h and 7 days from seeding ( Figure 3C-J). The samples, properly fixed (2.5% glutaraldehyde in a 0.1 M cacodylate buffer) and dehydrated (graded ethanol series) allowed for the analysis.
Adherent cells were observed in the whole cohort at each endpoint, confirming the MTT data. Focusing on cell organization over the scaffolds' surfaces, colonies were detected after 72 h from seeding. After 7 days, thicker clusters of cells were visible. Extensive neurite outgrowths, suggesting differentiation into neuronal-like cells, were not observed; conversely, as a consequence of massive proliferation, cells tended to form clumps appearing as overlapping cells, growing on top of one another in the central region of a cell mass. Cells appeared to acquire a roundish morphology up to detachment (see Figure 3C,D,F,J).

In Vivo Biocompatibility of OxPVA/Chitosan-Based Hybrid Scaffolds
The biocompatibility of the hybrid scaffolds was assessed by an in vivo implant into the subcutaneous region of BALB/c mice in order to evaluate the immunological and foreign-body reactions, as well as the fibrous encapsulation of the implanted biomaterials.
When anchored to the latissimus dorsi muscle of the animals, the four types of scaffolds exhibited adequate suture retention capacity ( Figure 4A-D). At the time of sample retrieval, 14 days from surgery ( Figure 4E-L), all scaffolds were well identifiable at the implant site and no macroscopic signs of inflammation were visible within the tissues surrounding the graft. A thin connective capsule enveloping the scaffolds was macroscopically recognizable, compatible with grade 1 soft tissue adhesion.  Explanted samples were properly fixed in formalin solution and paraffin-embedded and processed for subsequent histological and immunohistochemical analyses. The histological investigation by haematoxylin and eosin (H&E) staining allowed to preliminarily evaluate the graft and the surrounding host tissues then analysed for their typical microscopic morphology ( Figure 5A-D). At the superficial aspect of the implants, the multiple skin tissue layers were appreciable in all samples, including epidermis, dermis, white adipose tissue, and the panniculus carnosus (i.e., the thin striated muscle layer, localized between the adipose tissue and the interstitial connective tissue) ( Figure 5E-H). At the deep aspect of the implants, the latissimus dorsi muscle was visible, which served for scaffold anchorage with the chitosan layer in contact with the muscular tissue ( Figure 5I-L). At the microscopic analysis, a fibrous reactive capsule was confirmed to be completely surrounding the implant in all the experimental groups, being consistent with a normal host reaction towards biomaterial implantation. No evident differences were observed regarding this reactive capsule at the subcutaneous side vs. the latissimus dorsi muscle layer.
In parallel, Masson's Trichrome staining corroborated the presence of a moderate fibrotic reaction towards the hybrid scaffolds, as indicated by the fibroconnective capsule composed of dense, compacted collagen (green colour), which surrounds all the grafts at the interface with the host tissue at both the superficial ( Figure 6A-D) and deep ( Figure 6E-H) aspects of the implants.    Finally, ultrastructural analysis by SEM allowed the appreciation of the microporous aspects of the chitosan-based sponges cross-linked on the surface of OxPVA hydrogels ( Figure 9). Due to exposure to the physiological environment, a slight modification of the scaffolds surface ultrastructure was detected, compared to the pre-implant appearance; however, a certain fine porosity was still recognizable. Remarkably, host cells were observed to adhere on the chitosan layer of all the hybrid supports, confirming their adhesive properties also in the in vivo setting.

Discussion
In case of severe PNI, guaranteeing a satisfactory nerve regeneration still represents an unmet clinical need, thus revealing as an intriguing challenge still facing the tissue engineering field. In this context, research on new biomaterials and/or biomaterial combinations is particularly fervent, aiming to fabricate a NC able to support the adhesion, proliferation, and migration of nerve and glial cells. In particular, combining natural and synthetic polymers for a synergistic effect appears highly appealing, allowing the optimization of the advantages of both in terms of mechanical properties and biological functions [21]. This strategy was attempted here: within this study, discoidal hybrid scaffolds based on differently functionalized chitosan sponges + the new polymer OxPVA were developed and characterized both in vitro/in vivo, guiding toward future hybrid tubular devices.
Chitosan is a polysaccharide broadly used for supporting peripheral nerve regeneration, fulfilling the recommendations for a nerve conduit material [22]. In fact, chitosanbased devices show adequate biodegradability, biocompatibility, neglectable toxicity, and suitable biological absorptivity; excellent antibacterial activity is also reported [22][23][24][25]. In addition, chitosan proved beneficial in guiding Schwann cells' orientation, preserving the survival and differentiation of neuronal cells (commercial device, Reaxon ® ). All this evidence highlights its potential to serve as a valid alternative to autografts [22,24,25]. However, although distinguished as a highly promising material for NCs, positive outcomes are mainly related to short gaps, whereas its contribution in long gap nerve injury repair still needs to be clarified due to conflicting results [22,26]. To predict the in vivo crosstalk of the supports considered here, the human neuroblastoma cell line SH-SY5Y was used for in vitro studies, as these cells share certain properties with primary neurons [17].
To date, different modifications have been reported in the literature, proving the necessity of intense research to enhance chitosan in vivo performances. These include grafting with bioactive peptides, the alteration of the surface ultrastructure and combination with other polymers [27]. Within this study, the aim was to guide future research in NCs development but also in tissue engineering approaches, based on chitosan.
Cell-adhesive molecules derived from ECM (collagen, fibronectin, vitronectin, and laminin) can be effectively adsorbed on chitosan substrates [28][29][30]. However, despite encouraging evidence, easy desorption is likely to occur, prompting research into chitosan chemistry for more stable alternatives [29]. Surface chemical modification with immobilized, bioactive peptides likely assures a higher control and selectivity over cell-biomaterial interactions than other strategies [31]. As reviewed by Hozumi and Nomizu [32], several reagents/protocols have been reported for binding peptides to chitosan, intended for different destinations. Among these, 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide and N-hydroxy succinimide (NHS) were adopted for the immobilization of RGDS (Arg-Gly-Asp-Ser) carboxyl residue to the amine groups of chitosan (end use: bone) [33]. NHS and suberic acid bis (N-hydroxy-succinimide ester) were used for chitosan amine end group activation and conjugation with the amino terminal of the RGD sequence or epidermal growth factor (EGF) (end use: cartilage) [34]. Masuko et al. [35] introduced 2-iminothiolane to the chitosan amine residue and immobilized the cysteine-containing peptide through the disulfide bond formation (end use: synthetic ECM development). Hozumi and Nomizu [36] introduced the N-(m-maleimidobenzoyloxy) succinimide to the amine residue of chitosan, and then the cysteine-containing peptides were immobilized in the maleimidobenzoyl group through the mercapto group (end use: synthetic ECM development). Interestingly, in order to specifically modify neural cell adhesion to chitosan-based substrates and support neurite outgrowth, the bioactive laminin sequences YIGSR (eventually, with a glycine spacer, Wang et al. [37] bonded IKVAV to thiolated chitosan, forming chitosan-S-S-laminin peptide structures [29,38]. All these literature-based examples clearly highlight that chitosan is an excellent platform for tissue engineering purposes and is highly prone to functionalization due to its high content in various hydroxyl and amine groups [39]. However, utilizing chemical agents to stably decorate chitosan supports implies possible undesired effects connected to toxic remnants entrapment. To bypass this possibility, a different strategy was pursued here, examining SAPs dispersion in chitosan; furthermore, a focus over specifically bioactive sequences for neural regeneration was maintained because of IKVAV and YIGSR sequence inclusion. The SAPs are protein fragments that in solution self-organize in very stable amphiphilicbeta sheets conformation by hydrogen bonds; thus, they have great potential in tissue engineering, including peripheral nerve regeneration [20,40,41] due to a 3D self-arrangement mimicking ECM ultrastructural organization [41,42]. Furthermore, they also display high stability against thermal, chemical, and proteolytic attacks [43]. Avoiding adsorption (less stable) and covalent binding (possible chemical residues entrapment), the SAP EAK was here dissolved into a chitosan solution obtaining a complex ECM-like ultrastructure stabilized by secondary bonds; a similar approach was recently reported, for the first time in the literature, in order to enrich OxPVA hydrogels [17]. Interestingly, as SAPs do not have specific motifs promoting cell adhesion/growth likely associated with the nanometric fibrous scaffold or intrinsic mechanical properties [42], IKVAV (promotes adhesion/differentiation/neurons' axons growth/stimulates adequate microenvironment) or YIGSR (promotes cell adhesion) was grafted to EAK. The development of conjugates is expected to sustain the chitosan sponges' bio-effect [44,45].
Together with a favourable chemistry allowing bioactivation, chitosan also displays high mouldability, allowing for the fabrication of different scaffolds according to end-use destination (e.g., nanoparticles, hydrogels) [46]; here, sponge fabrication was chosen. In addition, the combination with other materials can easily occur. Specifically, regarding hybridization with polymers, the literature reports about several hybrids and/or composites combining the chitosan advantages with that, for instance, of collagen, polyglycolic acid (PGA), polylactic acid (PLA), poly(Lactic-Co-glycolic Acid) (PLGA), and polyethylene glycol [27]. The hybrid approach mainly revealed the superior results of axonal recovery compared to chitosan alone. OxPVA, a new material with an established promising attitude in supporting peripheral nerve regeneration [14][15][16][17]47], was here coupled with the ChS. The aim of this coupling was to improve OxPVA biological behaviour; in fact, as broadly reported, cell adherence was inhibited by the highly hydrophilic nature of the OxPVA hydrogel [14,17]. In the past, hybridization (SAPs, ECM derivatives) [9,10,17] and/or ultrastructural modification (patterning) [17] was revealed to be fundamental in triggering cell/biomaterial interactions in vitro.
As it is possible to achieve OxPVA-based scaffolds through physical crosslinking, Ox-PVA/ChS, OxPVA/ChS+EAK, OxPVA/ChS+EAK-YIGSR, and OxPVA/ChS+EAK-IKVAV supports were obtained through a completely "safe" strategy (freeze-thawing) not involving chemical agents (likely highly adsorbed due to a certain hydrogel swelling) and assuring the maintenance of OxPVA specific features (e.g., mechanical characteristics, permeability, biodegradation rate). In accordance with previous in vitro characterization studies [12,13,16] and preclinical evidence [14,15], OxPVA has mechanical characteristics that well match the requirements of the surgeon (no ruptures during NCs grafting), end-use destination (no NCs translocations), and implant site (no NCs kinking). Additionally, good in vivo results suggest that OxPVA-based devices display an adequate permeability, which supports the establishment of a favourable physico-chemical/biological environment during nerve regeneration. Possibly, this balance could be affected if a mixed hybrid scaffold OxPVA/ChS were to be prepared, requiring intense research for the identification of the optimal volume ratio. This is, for instance, seen in the study by Xie et al. [48], considering PLA/chitosan NCs fabrication, or in the research by Nawrotek et al. [49], fabricating chitosan-hydroxyapatite + collagen and/or hyaluronic acid grafts.
The mechanical properties of the chitosan-based sponges were investigated in order to appreciate eventual differences likely amenable to the presence and/or type of bioactive cues. Experimental data show a difference among the samples, highlighting a reduction in compressive modulus and maximum stress along with both a presence of SAPs and chain length. As regards the enriched chitosan matrices, EAK-IKVAV and EAK-YIGSR provided a contribution in reducing mechanical strength, the results of which were significantly lower than those of matrices with only EAK, without negatively affecting the compressive modulus. Despite this, these values do not affect the overall OxPVA/ChS-based scaffolds' mechanical properties (mainly dictated by the OxPVA polymer [9,12,13]), and they may correlate with ChS ultrastructure and crosstalk with cells and thus influence cell behaviour after seeding.
A preliminary consideration may associate the mechanical properties of the sponges with their ultrastructure. Lower mechanical properties likely imply a less tight organization of the ChS-based matrices; accordingly, larger mesh networks were evident in ChS+EAK-IKVAV and ChS+EAK-YIGSR (however, comparable among them) rather than in ChS+EAK and ChS alone; that being said, with regard to ultrastructure, image analyses also provided further information (higher pore density in ChS+EAK and ChS alone than ChS+EAK-IKVAV/-YIGSR; higher pore area in ChS+EAK-IKVAV/-YIGSR than ChS and ChS+EAK). Possibly, this may be ascribed to different chains steric hindrance, leading to a distinct spatial arrangement. In turn, the lower modulus of ChS+EAK-IKVAV and ChS+EAK-YIGSR (which was, however, not significantly different from that of ChS+EAK) may also correlate with cells penetration within the sponges, which was generally favoured [42]. Thus, the synergistic combination of a larger mesh network with a lower modulus may play a crucial role in cell behaviour. Overall, the obtained results showed the potential of tailoring the mechanical properties of the ChS-based scaffolds according to the specific application.
After seeding (72 h), all the matrices showed promising behaviour in supporting interaction with SH-SY5Y cells which, in turn, displayed a typical cluster-like organization [50]. At day 7, these clumps-like growths resulted in a certain number of cell reductions because of their possible detach due to overlapping and no surface available for adhesion. It is likely that, despite the ultrastructural differences discussed, all the chitosan-based networks were complex enough to induce cell adhesion and proliferation [50]. The ability to support interaction with cells is an interesting achievement, especially when considering that chitosan alone has no cell attachment activity, thus requiring functionalization/surface modification, in turn behaving like a modifiable platform [32]. Regarding this, an in vitro study by Luna et al. [51] is enlightening. The authors demonstrated that, whether untreated or treated (by argon or nitrogen-plasma), smooth chitosan membranes do not induce either cell adherence nor proliferation with increasing culturing times. This reveals that chitosan ultrastructural modification is an essential prerequisite to confer a certain bioactivity to chitosan and trigger its interaction with cells. Furthermore, as SAPs hydrogels can be designed to affect the network elasticity (specific peptide length, bioactive motifs), they can be exploited as a physiochemical regulator of cellular fate. As occurs in vivo, integrin-binding mediated events communicate the mechanics of the ECM to the cells and direct their fate through intracellular signalling pathways [52]. Within this scenario, a mechanical contribution from the OxPVA layer, above which the sponges are crosslinked, cannot be excluded.
The possibility of fabricating biocompatible scaffolds by means of bio-safe and reliable approaches is often a critical step towards the manufacture of tissue-engineered devices and their future translation into the market [53]. Biocompatibility is an important requisite for the success of a biomaterial for biomedical use, which will be in direct contact with living tissues; hence, this feature can be preliminarily assessed evaluating the intensity of an inflammatory reaction of the tissues adjacent to the implant (orthotopic implant) by histological/immunohistochemical analyses. Typically, lymphocytes and macrophage invasions are considered when examining eventual chronic inflammation and, in particular, macrophages are recognized as key mediators of immune reactions towards biomaterials [54]. According to experimental evidence, the tissues' reaction to the hybrid OxPVA/ChS-based membranes appeared to be favourable; only moderate inflammatory infiltration and a tolerable fibrous capsule were detected, without severe foreign body reaction signs. These data are a further confirmation regarding the "safe" fabrication methods of both the two layers and their coupling. In addition, it was also possible to recognize the two coupled sheaths' integrity (histological analysis and SEM), as well as the maintenance of the ChS-derived strata ultrastructure, confirming the adequacy of OxPVA/ChS crosslinking and suggesting sample stability not only towards the in vivo environment but also in surgery manipulation/suture for positioning.

Peptides' Synthesis and Purification
Three different peptides were synthetized for the subsequent development of functionalized ChS. Specifically, the bioactive sequences included the SAPs EAK, EAK-IKVAV, and EAK-YIGSR.
The peptide EAK (sequence: H-Ala-Glu-Ala-Glu-Ala-Lys-Ala-Lys-Ala-Glu-Ala-Glu-Ala-Lys-Ala-Lys-NH 2 ) was synthesized on Rink Amide MBHA resin (0.52 mmol/g) using Fmoc chemistry by a Syro I synthesizer (Multisyntech, Witten, Germany). The side-chain protecting groups were OtBu, Glu, and Boc, Lys. All the couplings were double (for each coupling 5 equivalents of Fmoc-amino acid; 5 eq. HBTU; 5 eq. oxyma pure; and 10 eq. DIEA were used). After the Fmoc-deprotection of the last inserted amino acid, the peptide was deblocked from the resin and deprotected from side-chain protecting groups using the mixture 4.75 mL TFA, 0.125 mL TES, and 0.125 mL H 2 O, for 1.5 h. The resin was filtered off and the solution was concentrated and added with cold diethyl ether. The product was precipitated and filtered. The identity of the crude peptide was determined by mass spectrometry (expected mass = 1614.49 Da; theoretical mass = 1614.79 Da; AB-SCIEX TOF-TOF 4800 instrument, Figure S1A). The peptide EAK was isolated by RP-HPLC and characterized by analytical RP-HPLC (conditions: Vydac C 18 column (5 µm, 300 Å, 4.6 × 250 mm, Grace), eluent A: 0.05% TFA in H 2 O; eluent B: 0.05% TFA in CH 3 CN; gradient: from 5 to 20% di B in 30 min, flow rate: 1 mL/min; detector: 214 nm. t R = 21.62 min, Figure S1B). The integration of the chromatogram confirmed a 94% purity grade.

Set-Up of the Chitosan-Based Sponges
Peptide-free ChS were prepared by dissolving 3.15 mg of chitosan in 472 µL of acetic acid solution 0.2 M under magnetic stirring. Of the solution, 375 mg were then poured within each 24-well plate, frozen with nitrogen liquid, and freeze-dried. After the lyophilization, cleavages were performed by filling each well three times with ethanol and later three times with MilliQ water. Finally, the scaffolds were once again frozen with nitrogen liquid and freeze-dried.
Regarding the development of the functionalized ChS, each scaffold was prepared by dissolving under magnetic stirring 0.945 mg of peptide (30% w/w peptide/chitosan), and 3.15 mg of chitosan in 472 µL of acetic acid solution 0.2 M. Of the solution, 375 mg were then poured within each 24-well plate, frozen with nitrogen liquid, and freeze-dried. After the lyophilization, cleavages were performed by filling each well three times with ethanol and later three times with MilliQ water. Finally, the scaffolds were once again frozen with nitrogen liquid and freeze-dried.

Mechanical Analysis of the Chitosan-Based Sponges
Compression tests were performed on peptide-free ChS and functionalized ChS (radius Ro = 6 mm, height Ho = 0.7 mm). Samples were immersed in physiological solution at 37 • C and tested using an INSTRON 5566 testing machine at a rate of 1 mm/min up to a strain of 50%. Considering the measured force F, the initial cross-sectional area of the specimen (A 0 =π R 0 2 ), the height variation (∆H), the initial height (H 0 ) of the specimen, the engineering stress (σ), and the engineering strain (ε) were calculated as follows: σ = F/A 0 and ε = ∆H/H 0 . ANOVA and Tukey's HSD test were used for statistical analysis (p < 0.05).

OxPVA Solution Preparation
OxPVA solution was prepared in accordance with a protocol previously reported in Stocco et al. [8,11]. Preliminarily, a suspension of PVA powder (molecular weight (Mw) 146,000-186,000 Da, 99+% hydrolysed) (i.e., 10 g in MilliQ water) was heated in boiling bath under stirring to promote complete polymer solubilization. Thereafter, the solution was cooled at 37 • C before proceeding with partial oxidation and dialysis. Partial oxidation was performed by the addition to the PVA solution of 151 mg potassium permanganate (KMnO 4 ) in 10 mL of acidic MilliQ water (1.60 g of 70% HClO 4 (w/w)). The end of the oxidation phase corresponded to the complete discoloration of the system, which generally took 1 h at 37 • C. Hence, extensive dialysis occurred by means of 8000 Da cut-off membrane (Sigma-Aldrich). The OxPVA solution was then frozen at −20 • C overnight and lyophilized (Speed Vac Concentrator Savant, Instruments Inc., Farmingdale, NJ, USA) for long-term storage. For polymer recovering, 16 wt% OxPVA was weighted, suspended into MilliQ water, and then heated for 48 h at 100 • C.

Hybrid Scaffolds Set-Up
Hybrid scaffolds based on OxPVA/ChS were prepared by pouring an OxPVA solution within a 6-well plate (mould), assuring the formation of a thin polymer layer (2 mm in thickness). Hence, the lyophilized chitosan matrices +/− peptides (i.e., EAK, EAK-IKVAV, EAK-YIGSR), prepared as previously described, were carefully placed over the polymer film, and a freezing-thawing (FT) treatment was adopted to physically cross-link the hydrogel and to embed the lyophilized matrix upon it. In particular, 6 FT cycles (−80 • C for 1 h/RT for 1 h) were performed, and then the two-layered hybrid scaffolds were kept at 20 • C until use ( Figure 10).

Scaffolds Ultrastructure
The ultrastructure of the OxPVA/ChS scaffolds seeding surface was investigated by SEM. After fabrication, the scaffolds were fixed with 2.5% glutaraldehyde (Sigma-Aldrich) in a 0.1 M cacodylate buffer (Sigma-Aldrich) (pH = 7.2) for 24 h, and then dehydrated through a graded ethanol series (Vetrotecnica, Padova, Italy). After critical point drying and gold sputtering, observation occurred using a Stereoscan-205 S (Cambridge Instruments, Pine Brook, NJ, USA).
The hybrid scaffolds were preliminarily disinfected using a 2% antibiotic/antimycotic solution (penicillin/streptomycin, Life Technologies, Paisley, United Kingdom) for 1 h under UV light (30 min/side) and then they were washed in a DMEM/F-12 (1:1) basal medium (Life Technologies). Hence, the supports were located in 48-well plate and seeded with SH-SY5Y cells (10 × 10 4 cells/cm 2 ; European Collection of Cell Cultures, Porton Down, UK) cultured in a proliferative medium consisting of DMEM/F-12, added with 15% FBS (Fetal Bovine Serum, Sigma-Aldrich), 1% non-essential amino acids (Sigma-Aldrich) and 1% antibiotic solution; the medium was refreshed every 2 days. Cell proliferation was evaluated by MTT (0.5 mg/mL) at 72 h and 7 d from seeding; briefly, after scaffolds incubation in MTT solution (4 h at 37 • C), formazan precipitates were dissolved in 2-propanol acid (0.04 M HCl in 2-propanol) and a microplate autoreader EL 13 (BIO-TEK Instruments, Winooski, VT, USA) was adopted to measure the solution optical density at 570 nm. As previously described [57], the results of cell seeding were expressed as total cell number/scaffold, according to a MTT standard curve, preliminary prepared. To this purpose, 1000, 5000, 10,000, 20,000, and 100,000 SH-SY5Y cells/well were seeded in 96-well plates and left to adhere for 12 h. Thereafter, cell viability was measured by MTT assay, obtaining optical density values to associate to each point of the curve. To quantify cells grown on seeded scaffolds, optical density values measured for each sample were plotted on the standard curve, gaining the corresponding cell number.
For SEM analysis, the samples were processed as described in Section 4.2.3.

Subcutaneous Implant
Hybrid OxPVA scaffold biocompatibility was verified through a heterotopic implant in mice [58]. Preliminarily, OxPVA/ChS discoidal samples (8 mm diameter; 2.5 mm in thickness) were prepared using a biopsy punch and disinfected by soaking in a 2% penicillin/streptomycin solution in PBS under UV light (15 min/side). Hence, twelve female Balb/c mice (12 weeks old; mean weight 19.8 ± 1.3 g) were anesthetized by gaseous anaesthesia (isoflurane/oxygen, in mixture) and, after shaving and disinfecting the dorsal cutis, a median dorsal pouch (10 mm in length) was created by a No. 10 surgical blade (Becton-Dickinson, Franklin Lakes, NJ, USA). The scaffolds were carefully placed and fixed to the latissimus dorsi muscle using Tycron 4/0 sutures to avoid eventual dislocation/migration, while the skin was sutured by absorbable Novosyn 4/0 stitches. The animals were administered antibiotic (Bytril, 5 mg/kg) and anti-inflammatory (Rimadil, 5 mg/kg) therapy for 5 days after surgery. In the following period, they were housed in a temperature-controlled facility and were given laboratory rodent diet and water ad libitum.
Euthanasia occurred after 14 days from by carbon dioxide asphyxiation; tissues were dissected and the implant-site was preliminarily observed for possible evidence of inflammation-related signs. Thereafter, the retrieved scaffolds, together with the surrounding tissues. were adequately fixed and investigated by histology and immunohistochemistry to unravel the biological response triggered in vivo. The ultrastructural evaluation of the scaffolds surface was also performed by SEM.

Histological Analyses
Graft integration and the inflammatory infiltrate at the graft-host interface was assessed by Hematoxylin and Eosin staining, whereas scaffold encapsulation by fibroconnective tissue was evaluated using Masson's trichrome staining.
Samples were first fixed in 10% formalin for one week, paraffin-embedded, cut into 5 µm-thick sections, dewaxed, and rehydrated with a series of ethanol (Vetrotecnica) solutions (99%, 95%, 70%) before being stained. Regarding Haematoxylin and Eosin staining, the dewaxed and rehydrated sections were immersed in haematoxylin for 8 min, tap water for 5 min, and eosin for 1 min. The sections were then dehydrated in a series of ethanol (95% for 45 s and 100% for 3 min (×2)), and finally immersed in xylene. Each section was then mounted with Eukitt (Merck LifeScience, Darmstadt, Germany). Masson's trichrome staining occurred using the Masson trichrome staining kit (Bio-Optica, Milano, Italy), following the manufacturer's instructions. Briefly, the rehydrated sections were exposed to Weigert's iron haematoxylin (A solution) + Weigert's iron haematoxylin (B solution) (for 10 min); without washing, the slides were drained and picric acid alcoholic stable solution (C solution) was dripped on the sections and allowed to act (for 4 min). Thus, after a quick wash in distilled water (3-4 s), Ponceau acid fuchsin according to Masson (D solution) was added to the section (for 4 min). Following a wash in distilled water, phosphomolybdic acid solution was added to the section (E solution) (for 10 min). Without washing, the slides were drained and a light green solution according to Goldner (reagent F) was left to act (for 5 min). Finally, the sections were washed in distilled water, dehydrated, and cleared in xylene as previously described, and mounted with Eukitt (Merck LifeScience).

Ultrastructural Evaluation by SEM
After sample retrieval, the analysis of scaffold ultrastructure was performed by processing specimens as described in Section 4.2.3 for SEM investigation.

Statistical Analysis
Data were expressed as the mean ± standard deviation (SD) of six different replicates (five images for porosity analysis based on SEM photomicrographs). Statistical analysis was performed by the one-way analysis of variance (ANOVA) and Tukey's post hoc test for multiple comparisons. Differences among experimental groups were considered statistically significant at p < 0.05. Statistical analyses were performed by Prism 8.1.0 (GraphPad Software).

Conclusions
According to in vitro/in vivo experimental evidence, the topographic modulation of OxPVA scaffolds surfaces by coupling with a chitosan bioactive layer proved safe and effective; specifically, different chitosan sponges were here compared, including ChS as such, Ch+EAK, Ch+EAK-IKVAV, and Ch+EAK-YIGSR.
ChS, even without the presence of SAPs, showed a certain ability in supporting cell adhesion and proliferation; this attitude is mainly mediated by sponge ultrastructure, depending on the fabrication method. The presence of bioactive SAPs did not interfere with the cell colonization of chitosan sponges; moreover, it was demonstrated that SAPs chain lengths can modulate chitosan's fine organization: longer chains (+EAK-IKVAV, +EAK-YIGSR) lead to larger interconnected pores than shorter sequences (+EAK). Interestingly, modulating the length of bioactive peptides can tune the mechanical properties of chitosan. It will be enlightening for future pre-clinical studies to assess the bioactive role of the peptides regarding both the impact of different NC lumen ultrastructures and the specific bioactive role of the sequences.
OxPVA/ChS-based scaffolds may have great potential for biomedical applications; however, further research is required to identify the ideal equilibrium between stimuli (3D organization, bioactive sequences length/type). This study lays the basis for future van-guard hybrid scaffolds supporting structural and functional nerve recovery, including nerve conduits for long-gap reconjunction. Additionally, other possible end-use destinations cannot be excluded through both varying and combining different SAPs.