The PCL pellets (CAPA6800, Perstorp Ltd, Sweden, Mw = 80,000 kDa) used in this experiment are medical grade (approved by the United States Food and Drug Administration), and melt at 60°C. Lithium fluoride (LiF ≥99%) and 2,2,2-trifluoroethanol (TFE) were purchased from Sigma (St. Louis, MO). Hydrochloric acid (HCl), absolute ethyl alcohol, and absolute methyl alcohol were purchased from Aladdin Reagents (Shanghai) Co., Ltd., China. The Ti₃AlC₂ powder was purchased from Jilin 11 technology Co., Ltd. All reagents were analytical grade.

The d-Ti₃C₂T_x was synthesized by a typical chemical exfoliation process as we described previously with some modifications (Cao W.-T. et al., 2019). Typically, 1 g of Ti₃AlC₂ powder was dissolved in a mixture of 9 M HCl (20 ml) and 1 g LiF. Then, the obtained suspension was allowed to process for 48 h at 35°C in a water bath for etching the aluminum (Al) layer. The resultant suspension was washed repeatedly using deionized water until its pH reached about 5. Finally, the obtained sediment was further delaminated by ultrasonication in an ice bath and centrifuged to obtain uniform d-Ti₃C₂ MXene nanosheets suspension. The obtained MXene nanosheets suspension was further stored in a refrigerator at 4°C to prevent oxidation.

For the preparation of pure PCL NGCs, 2 g of PCL pellets was added into 12 ml of trifluoroethanol to form 17% (w/v) transparent adhesive solution with stirring for 24 h at room temperature. The NGCs were prepared by using electrospinning equipment (Changsha Nayi Instrument Technology Co., Ltd, China). Briefly, the obtained solutions were transferred into a 5-ml syringe and electrospun at 0.8 ml/h under 12-kV high DC voltage. The distance from the rotational collector (200 rpm, a tungsten steel rod with a diameter of 2 mm) to the metal needle is 25 cm. Then, the obtained PCL NGCs were gently detached. To prepare MXene-PCL NGCs, MXene solution (an aqueous solution containing 40 mg ml−¹ MXene) was uniformly coated on the outer surface of PCL NGC by spraying technology, and the thin conductive sheath was formed by spontaneous adhesion. In addition, we tried to coat the MXene on a variety of other materials (bacterial cellulose membrane, textiles, paper, Latex, and melt-blown fabric).

The morphology and surface structure of MXene-PCL and PCL NGCs were characterized by a digital camera and scanning electron microscope separately, and the size and morphology of MXene nanosheets were examined using a transmission electron microscope (TEM, JEOL JEM-2010 (HT)). Fourier-transform infrared (FTIR) spectroscopy spectra was obtained using a FTIR spectrometer (Nicolet iS 10). The conductivity capability of the nano scaffolds was measured using four-point probes (ST2258C, China). The tensile stress–strain mechanical property was evaluated using an electronic tensile test machine (HY-940FS, Shanghai, China). To measure the porosity (%) of PCL scaffolds, we obtained a certain volume of PCL scaffolds and weighed the mass to calculate the density ρ₁. Furthermore, the density of PCL particles is measured as ρ₂. Then the porosity (%) of the PCL scaffolds = (1 − (ρ₁/ρ₂)) × 100%. Similarly, the density of MXene-PCL NGC was measured as ρ₃, the bulk density of MXene and PCL was calculated as ρ₄, and the porosity (%) of the MXene-PCL scaffolds = (1 − (ρ₃/ρ₄)) × 100%.

Rat Schwann cells (RSCs) were provided by the Chinese Academy of Sciences (Shanghai, China). The MXene-PCL and PCL film were soaked in 75% alcohol and rinsed repeatedly with PBS at least three times. Then, they were placed in a sterile Petri dish, dried in a fume hood, and irradiated on the front and back sides for 2 h separately to achieve sterilization by an ultraviolet lamp. We seeded the cells on the PCL side of MXene-PCL scaffolds, PCL scaffolds, and TCP at a density of 2 × 10⁴ cm⁻². After incubation for 24 h, a LIVE/DEAD kit (Beyotime, China) was used for cell viability analysis according to the standard protocols. Finally, cells in each group were photographed under a fluorescence microscope.

The Cell Counting Kit 8 (CCK-8, Beyotime, China) was used to assess cytotoxicity. The cells were cultured on the PCL side of MXene-PCL scaffolds and PCL scaffolds in 24-well plates for 6, 12, 24, 72, and 168 h. Subsequently, a CCK-8 detection solution was added to each well to ensure that the concentration of CCK-8 in the medium solution was 10%. After incubation at 37°C for 3 h, the absorbance of the supernatant was detected by a microplate reader (Thermo 3,001, Thermo Fischer Scientific, United States) at a wavelength of 450 nm. TCP was used as a control. This was repeated at least three times for each group.

RSCs were seeded onto the PCL side of MXene-PCL nanofiber membrane, and PCL nano scaffolds separately. After 4 days, the morphology of the cells on the different scaffolds were observed by scanning electron microscopy (ZEISS Gemini 300). First, the medium was discarded, washed with PBS, and fixed with 2.5% glutaraldehyde at room temperature for 2 h. We added 1% osmium acid and incubated at 4°C for 2 h after discarding the fixation solution, followed by alcohol gradient (10, 30, 50, 70, 80, 90, 95, 100%) dehydration for 15 min each. After drying at room temperature, SEM pictures were taken to evaluate the adhesion and morphology of cells on the different scaffolds.

In addition, phalloidin staining was also used to observe the adhesion ability and morphology of RSCs on the different scaffolds. After the cells were cultured on MXene-PCL and PCL nanoscaffolds for 4 days, the cell culture medium was discarded and washed with PBS twice. Then, the cells were fixed in 3.7% formaldehyde solution prepared by PBS at room temperature for about 20 min and washed with PBS containing 0.1% Triton X-100 three times, 5 min each. Actin-tracker Red (Beyotime, China) was diluted with PBS containing 5% BSA (Beyotime, China) and 0.1% Triton X-100 (Beyotime, China) at a ratio of 1:200. The solution was added to each slide in a ratio of 200 µl and incubated in a slide dyeing box at room temperature away from light for 60 min. Then, we washed the cells with PBS containing 0.1% Triton X-100 three times, 5 min each. Finally, a fluorescence microscope was used for observation.

The NGCs were dried to a constant weight, and the sample mass W₀ was recorded. Then, we placed the samples in a quantitative PBS solution at 37°C. At fixed time points, each sample was dried and weighed W₁. The formula of mass loss rate W% is as follows: W% =(W₀−W₁)/W₀×100%.

SD rats (male, weighing 150–200 g) were used to evaluate the nerve regeneration capacity. The modal rats were assigned into three groups randomly: PCL, autograft, and MXene-PCL groups. Animals were deeply anesthetized with 2% pentobarbital sodium. The sciatic nerve and its main branches were exposed, and a 15-mm long nerve defect was created under aseptic conditions. In the autograft group, the nerve was sutured end to end after rotating the nerve stump around 180°. In the PCL or MXene-PCL groups, the nerve defect was sutured with a nerve conduit. After that, the soft tissue and skin were sequentially sutured. To prevent infection, antibiotic (ceftriaxone sodium, 0.1 g/kg) was injected for three consecutive days. Postoperative assessments took place at weeks 4, 8, and 12. Ethical approval for animal studies was obtained from the Institutional Animal Care and Use Committee of Shanghai Tenth People’s Hospital (SHDSYY-2021–2,305).

The rats’ footprints were acquired to calculate the sciatic functional index (SFI) at weeks 4, 8, and 12 postoperative. The formula used to measure nerve function was as follows: SFI = (13.3 × (EIT − NIT)/NIT) + (−38.3 × (EPL − NPL)/NPL) + (109.5 × (ETS − NTS)/NTS) − 8.8. In this formula, the IT stands for the distance from the second to the fourth toes; the PL represents the length from the distal of the third toe to the heel; TS is the length from the first to the fifth toe, N means the normal side, and E means the experimental side. The SFI is scored from 0 (normal function) to 100 (the function is completely impaired). Electrical signals such as compound motor action potential (CMAP) and nerve conduction velocity (NCV) were recorded by electrophysiology analysis at 12 weeks postoperative.

The wet weight of the gastrocnemius muscle was measured at week 12 after implantation. The gastrocnemius muscles from both legs were dissected, and the fat tissues attached were carefully removed with ophthalmic micro-scissors. The normalized wet weight of gastrocnemius muscle (%) = Wo/Wn, where Wo is the gastrocnemius muscle weight of the operated leg and Wn is the gastrocnemius muscle weight of the nonoperated leg.

Regenerated nerves were collected on completion of the electrophysiological experiments at week 12 postoperatively. The cross-sectional morphologies of the nerves (in the middle of the specimens) were visualized by TB staining, hematoxylin-eosin (HE) staining, and TEM. For HE staining, the samples were fixed with 4% paraformaldehyde and then embedded in paraffin, cut into sections by using a microtome. For TEM and TB staining, the tissues were fixed at 4°C in 2.5% glutaraldehyde, followed by embedding in Epon812, cut into ultrathin sections. The sections were observed using an immunofluorescence microscope (Leica, United States ). The superfine microstructure of the regenerated myelin sheath was observed via a TEM (China Titan) at a voltage of 80 kV. The diameter and density of the axon and the thickness of the myelin sheath were quantitatively analyzed by ImageJ software. The S100/MBP and NF200/Tuj1 triple immunofluorescence staining were used to assess nerve myelin protein and nerve axon protein of the regenerated nerve, respectively. Tissues were fixed with 1% tetraoxide solution, dehydrated, and embedded in Epon812 resin. The cross-section was cut to a thickness of 4 μm and mounted on 2% gelatin-coated slides. The primary antibodies (Abcam, United States) included anti-Tuj1 (1:200), anti-MBP (1:200), anti-NF 200 (1:200), and anti-S100 (1:200). All slides were evaluated using an immunofluorescence microscope (Leica, United States). At week 12 after surgery, the paraffin sections of regenerated sciatic nerve sections were prepared, respectively, for immunofluorescence staining with CD34 antibody and immunohistochemistry (ICH) staining with CD31 antibody as described above. The primary antibodies included rabbit anti-CD34 antibody (1:200, ABclonal, A7429) and rabbit anti-CD31 antibody (1:200, Bioss, bs-0195R). The microvessel density (MVD) and CD 31 areas were measured by ImageJ software. The morphology of muscle fibers (from the gastrocnemius muscle of the operated leg) and major organs (heart, liver, spleen, lung, and kidney) were evaluated by HE staining.

Images from immunofluorescence staining, TEM, HE, and immunohistochemistry staining were analyzed using Prism 8 (GraphPad, United States) and ImageJ software. All measurements were performed five times, and the results are presented as the mean ± s.d. Differences between the values were analyzed with one-way analysis with Tukey’s post hoc with GraphPad Prism. For all figures, NS: p > .05; *p < .05.

Long gap nerve defect still is a common clinical disease with poor prognosis, and the conductive NGCs may have clinical utility in curing this disease. MXene shows good electrical conductivity, could promote tissue regeneration, and has potential application in neural tissue engineering. There is extensive literature supporting that the optimal porosity of the NGC is a critical feature for achieving neovascularization, modulating the immune response, and nerve growth patency (Shen et al., 2021; Yu et al., 2021). In this study, we fabricated PCL NGCs by electrospinning (Figure 1), which had a porous structure and appropriate mechanical properties, and size (Figure 2). To prepare MXene-PCL NGCs, MXene solution was uniformly coated on the outer surface of PCL NGC by spraying technology, and the thin conductive sheath was formed by spontaneous adhesion. Thus, the MXene-PCL NGC with PCL inner cavity and MXene outside sheath was obtained finally. The morphologies of MXene-PCL and PCL NGCs were first characterized using an optical microscope (OM) and SEM (Figure 2). The obtained PCL NGC has a milky white color, and the MXene-PCL NGC showed a white inner cavity and a black shell. The infrared absorption peak of MXene-PCL composite film (MXene side as the front side) was completely consistent with those of pure MXene film, indicating that MXene formed dense films on the surface of PCL through self-adhesion (Supplementary Figure S1, Supporting Information). Further tests indicated that the MXene is also easy to coat on other materials (including bacterial cellulose membrane, textiles, paper, latex, and melt-blown fabric) (Supplementary Figure S2, Supporting Information). SEM images obtained from the PCL side of MXene-PCL film showed a loose porous appearance formed by PCL fibers interlaced. Meanwhile, the MXene sheath is relatively smooth. TEM analysis further revealed the MXene had a lamellar nanostructure. We further estimated the mechanical and electrical properties of the MXene-PCL and PCL NGCs. The scaffold thickness, average elastic modulus, and the elongation at break, and porosity results were similar for both materials (Figure 2 and Supplementary Figure S3, Supporting Information). Therefore, the addition of the MXene sheath is too thin to influence these properties. The electrical conductivity changed from 0 S/cm (PCL NGC) to 2.91 × 10^–2 S/cm (MXene-PCL NGC) after coating of MXene sheath. According to the results presented above, the MXene-PCL NGC exhibited excellent mechanical and topological properties, ideal rigidity and flexibility, microporosity for cell adhesion and proliferation in the PCL side, and relatively high electrical conductivity on the outer surface.

Biosafety is always a considerable concern for potential clinical applications of electroactive nanoparticles, such as MXene. Previous research claims that MXene-contained materials were noncytotoxic, healable, and degradable (Xu et al., 2016; Wang et al., 2019; Basara et al., 2020; Wang et al., 2021; Wei et al., 2021; Zhou et al., 2021). The design in which the MXene was coated outside the conduit rather than inside could avoid the potential toxicity caused by direct contact between the MXene particles and newly regenerated nerves to some extent. For this study, we used Schwann cells to evaluate the potential toxicity of PCL or MXene-PCL scaffold at different time points. According to CCK-8 results, Schwann cells on both scaffolds were active and showed no statistically significant differences in cell proliferation after culturing for 6, 12, 24, 72, and 168 h compared with the TCP controls (Figure 3). The LIVE/DEAD test results revealed negligible differences among the TCP, MXene-PCL, and TCP groups (Figure 3). Cell attachment on the scaffolds is a crucial factor for cell viability. The morphology and attachment of RSCs after seeding on the scaffolds for 72 h are examined by SEM. The results show that the RSCs adhered to the surface of PCL fiber and extended many pseudopodia (Figure 4). Similar results could be obtained by staining actin cytoskeleton with phalloidin (Supplementary Figure S4, Supporting Information). Furthermore, the heart, liver, spleen, lung, and kidney were harvested from the rats at week 12 postoperatively. There were no abnormalities observed from the major organs’ histological assessment (Figure 5). Taken as a whole, these results indicate that the MXene-PCL NGCs are biocompatible and could afford a beneficial microenvironment for cell growth and support nerve regeneration.

The modal rats were assigned into three groups randomly: PCL, autograft, and MXene-PCL groups. Postoperative assessments took place at weeks 4, 8, and 12. All rats were raised in an SPF-level environment; none was infected postoperatively. No animals showed any signs of operative complications, and the wounds healed well. At week 12, all the structures of the NGCs in animals remained intact, but the surface was degraded to a certain extent, and this can be confirmed to some extent by comparing Figure 2 and Supplementary Figure S5, Supporting Information. Although the degradation rate of PCL is very slow (Supplementary Figure S5, Supporting Information) (Lam et al., 2009), it does not affect the biocompatibility or nerve regeneration in this experiment. The pictures present the morphology of MXene-PCL NGC before and at implantation and the macroscopic view of NGCs and the regenerated nerves from different groups at 12 weeks after implantation (Figure 6, and Supplementary Figure S5, Supporting Information).

The reinnervation of the gastrocnemius muscle was evaluated 12 weeks after implantation. The gastrocnemius muscle is an important target organ innervated by the sciatic nerve. Once the nerve damage occurs, the innervated muscles atrophy. To some extent, muscle recovery is positively correlated with nerve injury. The results show that the MXene-PCL group restored the muscle weight compared with the PCL group and showed no notable difference to that in the autograft groups. The diameters of muscle fibers in the MXene-PCL group were notably longer than that of the PCL group and showed no significant difference from that in the autograft group (Figure 6).

The SFI and electrophysiological assessment methods were conducted to evaluate the functional recovery of the sciatic nerve. The step length of each paw and the length and width of the paw prints were measured according to the footprints (Supplementary Figure S6, Supporting Information). We found that the SFI of the MXene-PCL group (−34.9) was obviously better than the PCL group (−48.8, p < .05) and showed no obvious difference to that of the autograft group (−33.8, p > .05) at postoperative 12 weeks. Electrophysiological analysis shows that the regenerated nerve CMAP and NCV of the MXene-PCL group (38.9 m s⁻¹ and 3.0 mV) were notably higher than that of the PCL group (21.4 m s⁻¹ and 1.6 mV, p < .05) and showed no notable difference from that in the autograft groups (41.28 m s⁻¹ and 3.6 mV, p > .05) (Figure 6) at 12 weeks after implantation.

For validating the morphological improvement, TEM, HE, and toluidine blue staining were performed. The results reveal that the MXene-PCL contributed to the regeneration of nerve fibers and myelination compared with the PCL group. The myelin wrapped around the axon in a compact, multilayered spiral in the MXene-PCL and autograft groups (Figure 7, and Supplementary Figure S7, Supporting Information).

We found that the diameter and density of the myelinated axon, the thickness of the myelin sheaths of MXene-PCL group (3.91 μm, 2.71 × 10⁴ and 0.64 μm) showed no notable difference to that in the autograft groups (4.09 μm, 2.98 × 10⁴ and 0.685 μm), respectively (Figure 7). In contrast, only a few nerve fibers and many vacuoles appeared in the PCL group, and the diameter and density of myelinated axon myelin thickness were significantly less than that in the MXene-PCL and autograft groups.

To evaluate the effect of the electroconductive MXene-PCL NGC on nerve regeneration, the Schwann cell marker protein 100 (S100) and neuron-specific protein (MBP) and the typical axon protein of neurofilament protein 200 (NF200) and β-III-tubulin (Tuj 1) were evaluated by immunofluorescence staining. The results reveal that the expression level of MBP and S100 in the MXene-PCL group was 3.12 and 2.93 times higher than that in the PCL group, respectively (p < .05) while 3.48 and 2.93 times higher in the autograft group than that in the PCL group. The expression level of NF200 in the MXene-PCL and autograft groups was 2.16 and 2.90 times higher than that in PCL group. The expression level of Tuj 1 in the MXene-PCL and autograft groups was 3.40 and 4.22 times of the PCL group (p < .05). The above results confirmed that the MXene-PCL NGC effectively improved the myelination and regeneration of the nerve fibers with a similar effect to the autograft nerve (Figure 8, and Supplementary Figure S8, Supporting Information). Taking a step further, we evaluated the neovascularization by immunohistochemistry of specific protein markers of endothelial cells (CD31) and immunofluorescence of hematopoietic progenitor cell markers (CD34). It was found that the MXene-PCL and autograft groups showed more vascular-like structures than that of the PCL group by nerve cross-section staining. The average microvessel density calculated from CD34 and CD31 areas of the MXene-PCL group (43.67 mm⁻² and 2.44 mm²) were notably higher than that of the PCL group (26.00 mm⁻² and 1.64 mm², p < .05) and showed no notable difference from that in the autograft groups (45.67 mm⁻² and 2.71 mm², p > .05) at 12 weeks postoperative (Figure 9).