GeLMA powder was prepared as previously described (Luo et al., 2021a). First, 20 g of gelatin (type A, Sigma-Aldrich, United States) was dissolved in 200 mL phosphate-buffered saline (PBS, Gibco, United States) under magnetic stirring at 50 °C. Then, the gelatin solution was mixed with 2 mL of methacrylic anhydride (MA, Sigma-Aldrich, United States), which was added at a rate of 0.5 mL/min. After 3 h, the suspension was placed in a 1-kDa dialysis tube and dialyzed at 50 °C for 5 days. The final GeLMA macromer was lyophilized and stored at room temperature for further use.

To fabricate GeLMA/bFGF hydrogels, GeLMA powder was dissolved in deionized water to create a solution with a 5% concentration. Then, lithium phenyl (2,4,6-trimethylbenzoyl) phosphinate (LAP, Tokyo Chemical Industries, Japan), a photo-initiator, was added to this solution at the ratio of 16:1, according to our previous report (He et al., 2024). The mixture was heated for 1 h at 50 °C until the solution was clear. Eventually, the GeLMA solution was poured on a specific silicone mold and exposed to LED light for 2 min to form GeLMA hydrogel. In addition, 20 ng/mL of bFGF was added to this mixed solution before light curing to construct the GeLMA/bFGF hydrogel.

Morphological properties of these hydrogels were observed by scanning electron microscope (SEM, HITACHI, Japan). The hydrogels were divided into three groups: the GeLMA group as the GeLMA hydrogel received no treatment, GeLMA-SQ group as the GeLMA hydrogel was injected by a 1-mL syringe, and GeLMA/bFGF-SQ group as the GeLMA/bFGF hydrogel was injected by a 1-mL syringe. All samples were flash frozen in liquid nitrogen and lyophilized by a vacuum freeze dryer for 2 days. As for SEM observation, the fracture surface was sputtered with platinum. The images were quantified by ImageJ (Bethesda, United States).

The degradation rate of hydrogels was determined by using a weight loss method, as previously described (Liu et al., 2017). In brief, the lyophilized hydrogels (M0) were added to 10 μg/mL of type-I collagenase and kept in the shaker incubator (150 rpm/min) for 2, 6, 12, and 24 h at 37 °C. Then, the hydrogels were lyophilized and weighed at predetermined intervals (M1). The degradation rate was calculated using the following equation: M L % = M 0 − M 1 M 0 × 100 % .

DPSCs were obtained and cultured as previously described (Luo et al., 2023). In brief, the pulp tissues were cut into small pieces and digested with the mixture of collagenase type I (Gibco, United States) and dispase (Sigma, Germany). After 1 h, the cell suspensions were cultured with α-modified Eagle’s medium (α-MEM, Gibco, United States) supplemented with 20% fetal bovine serum (FBS, Gibco, United States) and 1% penicillin/streptomycin. The culture medium was changed every 3 days. The viability and proliferation of DPSCs which were co-cultured with hydrogels (GeLMA, GeLMA-SQ, and GeLMA/bFGF-SQ) were investigated by a live/dead viability/cytotoxicity kit (Invitrogen, United States) and a CCK-8 assay kit (Dojindo, Japan), respectively. In brief, DPSCs were dispersed in GeLMA and GeLMA/bFGF solutions and photo-cross-linked to create GeLMA-DPSCs and GeLMA/bFGF-DPSCs hydrogels. The GeLMA-DPSCs hydrogel, which was directly plated into the 96-well plate, was labeled as the GeLMA group. The GeLMA-DPSCs and GeLMA/bFGF-DPSCs hydrogels were injected into the 96-well plate by 1-mL syringes, which were coded as GeLMA-SQ and GeLMA/bFGF-SQ groups, respectively. After 1, 3, and 5 days, 10% CCK-8 solution was added and incubated for another 1 h. An absorbance microplate reader was used to measure the OD values at 450 nm. As for live/dead staining, the hydrogels would be removed and stained according to the manufacturer’s instructions. All samples were examined and captured by a fluorescent microscope (Carl Zeiss, Germany), and ImageJ was used to determine the number of living cells.

After 70%–80% confluence, the growth medium of DPSCs was removed and replaced with serum-free α-MEM. After 2 days, the supernatants were collected and centrifuged at 3,000 rpm for 10 min. Then, the supernatants, which were referred to as DPSC conditioned medium (DPSCs-CM), were filtered by a 0.22-μm filter unit and stored at −80°C for further experiments.

The PC12 cells were obtained from the American Type Culture Collection (ATCC) and cultured in advanced DMEM/F12 medium supplemented with 5% FBS and 1% penicillin/streptomycin. The PC12 cells were divided into three groups according to the culture conditions: the control group (DMEM/F12), NGF group (50 ng/mL NGF + DMEM/F12), and DPSC group (50% DPSCs-CM + DMEM/F12). The culture medium was replaced every 2 days. After 7 days, the surface marker MAP2 of PC12 cells was analyzed by immunofluorescence staining. ImageJ software was used to evaluate the average neurite length and the fluorescence intensity of MAP2.

A total of 24 female Sprague-Dawley rats, weighing 250–300 g, were purchased from the animal center of the Chinese Academy of Science (Shanghai, China). All rats were kept in specific pathogen-free (SPF) conditions at a room temperature of 22°C–24°C and 12 h light/dark cycle. All rats had water and food ad libitum. The experimental procedures were performed under the regulations of the Institutional Animal Care and the Ethics Committee of Wenzhou Medical University (wydw 2020–0897). All rats were randomly divided into three groups: control, GeLMA-SQ (injected with GeLMA hydrogel), and GeLMA/bFGF/DPSCs-SQ groups (injected with GeLMA/bFGF/DPSCs hydrogel). The left optic nerve was crushed intraorbitally, as previously described (Jacobi et al., 2022). After being anesthetized, a 1-cm incision extended from the temporal side of each left eye was made to expose the optic nerve. Then, the optic nerve was crushed at 1 mm for 10 s by blood vessel forceps. After the optic nerve tissue was extruded from the injured site, 5 μL of hydrogels was immediately injected into the crush site of the optic nerve using 1-mL syringes. The rats that had not undergone any kind of surgery were defined as the control group. The optic nerve was collected after 14 days for further analyses.

RGCs survival was determined by counting the number of βIII-tubulin-positive RGCs in the retina. The retinas were carefully removed from the rats after they were transcardially infused with cold PBS and 4% paraformaldehyde (PFA). The retinas were fixed with 4% PFA for 1 h and blocked with PBS containing 5% normal goat serum for 6 h. Then, the retinal whole mounts were incubated with the primary antibody βIII-tubulin overnight at 4°C. Following PBS washes, the retinas were incubated with the secondary antibody for another 1 h at room temperature. RGCs were counted at three different retinal views.

In order to trace the regenerated axons, cholera toxin B fragments (CTBs) were intraocularly injected 2 days before the animals were harvested, as previously described (Cameron et al., 2020). The quantity and length of CTB-labeled regenerated axons were counted on the optic nerve slices under a fluorescence microscope. The number of regenerated axons was counted at distances (d) of 100, 300, 500, and 700 μm from the original site to the injured site. The average number of axons was calculated from the three parts of each optic nerve. The number of axons per millimeter was calculated from the measurement (2r) of the cross-sectional width of the nerve. ∑ad was the total number of axons extending for a distance in the nerve of radius r. It was estimated by summing all cross sections of thickness t (10 mm) and was calculated as follows: ∑ ad = π r 2 × average   axons mm t .

The histological sections were derived from the optic nerve segments that were embedded in optimal cutting temperature (OCT, Sakura, American). The cryosections (10 μm) were stained with hematoxylin and eosin (HE) and observed using a light microscope (TS100, Nikon, Japan). In addition, after euthanization, the optic nerve segments were picked out and fixed with 2.5% glutaraldehyde for 2 h, further fixed with 1% osmium tetroxide for another 2 h, dehydrated by ethanol and embedded in Epon812, cut into 50–60 nm sections, and stained with lead–uranium for transmission electron microscopy (TEM, H-600, Hitachi, Japan) observation. ImageJ was used to measure the diameter of the myelin sheath at each location.

All the data were presented in mean ± standard deviation (SD). The unpaired two-tailed t-test was used for comparisons between two groups. One-way analysis of variance (ANOVA) was used, followed by Tukey’s multiple comparison test for comparisons among three or more groups. P < 0.05 was considered significantly statistical. Statistical analyses were calculated by SPSS 19.0 (SPSS, Chicago, IL).

Figure 1 shows the morphology and degradation of GeLMA-SQ and GeLMA/bFGF-SQ hydrogels. As shown in Figure 1A, the results showed that the large pore size in GeLMA hydrogel was similar to that in GeLMA-SQ hydrogel, whereas the pore size in GeLMA/bFGF-SQ hydrogel was significantly decreased with the addition of bFGF. In addition, the median pore diameter of GelMA and GeLMA-SQ hydrogels was nearly 45 μm, but it reduced to 20 μm in GeLMA/bFGF-SQ hydrogel (Figure 1B). Furthermore, there was a statistical difference in the average pore size between the GeLMA and GeLMA/bFGF-SQ hydrogels, as well as the GeLMA-SQ and GeLMA/bFGF-SQ hydrogels (Figure 1C). In addition, the degradation rate was the fastest in GeLMA/bFGF-SQ hydrogel, followed by the GeLMA-SQ hydrogel, while the GeLMA hydrogel degraded slowly. Specifically, the GeLMA hydrogel displayed a reduction of nearly 40% at 24 h, while the GeLMA-SQ and GeLMA/bFGF-SQ hydrogels reached degradation levels of approximately 65% and 80%, respectively (Figure 1D).

As shown in Figure 2, the results revealed that the hydrogels had great cytocompatibility with DPSCs after being co-cultured for 1, 3, and 5 days. In addition, the morphology of DPSCs gradually elongated from day 3 to day 5, especially in the GeLMA/bFGF-SQ hydrogel (Figure 2A). However, the number of living cells had no significant difference among the three hydrogels (Figure 2B). Moreover, the CCK-8 results suggested that DPSCs also had good proliferation in all the hydrogels, and the trend of data was consistent with live/dead staining (Figure 2C).

The PC12 cells were stained with MAP2 to identify the axons’ length. According to immunofluorescence staining, PC12 cells treated with NGF and DPSCs-CM had better growth of axons than in the control group (Figure 3A). Quantification of the MAP2 fluorescence intensity showed a statistically significant difference between the control and NGF groups. However, there was no statistically significant difference between the DPSC-CM and control groups, as well as the NGF group (Figure 3B). However, there was a statistically significant difference in the length of axon growth between the control and NGF groups, as well as the DPSC-CM group (Figure 3C). These results indicated that DPSCs-CM might be able to encourage the growth of the axons but not the proliferation of PC12 cells.

The survival of RGCs was analyzed by counting the number of βIII-tubulin-positive labeled RGCs in the retina. From center to periphery sites, RGCs were positively labeled with βIII-tubulin, and the results suggested that the optic nerve axons were the thickest in the control group, followed by those in the GeLMA/bFGF/DPSCs-SQ group, and the axons were the thinnest in the GeLMA-SQ group (Figures 4A, B). In addition, the number of RGCs in the control group was significantly greater than that in the GeLMA-SQ and GeLMA/bFGF/DPSCs-SQ groups. Compared to the GeLMA-SQ group, the number of RGCs was much higher in the GeLMA/bFGF/DPSCs-SQ group, and the difference was statistically significant (Figure 4C).

Figure 5 displays the outcomes of anterograde tracing with CTB on the optic nerve. In the GeLMA/bFGF/DPSCs-SQ group, more regenerated axons were found in the proximal crush site, but there was little regeneration of axons in the lesion site of the GeLMA-SQ group (Figure 5A). Compared with the GeLMA-SQ group, the number of CTB-positive-labeled axons was significantly increased at distances of 100, 300, and 500 µm in the GeLMA/bFGF/DPSCs-SQ group (Figure 5B). In addition, the average length of regenerated axons in the GeLMA/bFGF/DPSCs-SQ group was longer than that in the GeLMA-SQ group, with a statistically significant difference (Figure 5C).

As shown in Figure 6A, the results of HE staining indicated that there were obviously regenerated nerve tissue in the lesion site of the GeLMA/bFGF/DPSCs-SQ group, whereas a large number of vacuoles existed in the injured site of the GeLMA-SQ group. TEM results revealed that the optic nerve axons were arranged regularly and the myelin structure was intact in the control and GeLMA/bFGF/DPSCs-SQ groups. In contrast, the axons of the optic nerve were disorganized and several of the myelin sheaths were thin in the GeLMA-SQ group (Figure 6B). Compared to the GeLMA-SQ group, the thickness of the myelin sheath was thicker than that in the control and GeLMA/bFGF/DPSCs-SQ groups but with no significant differences. Furthermore, the diameter and number of axons in the GeLMA/bFGF/DPSCs-SQ group were higher than those in the GeLMA-SQ group, and the differences were statistically significant (Figure 6C).