Spatiotemporal regulation of endogenous MSCs using a functional injectable hydrogel system for cartilage regeneration

Characterization of p-hydroxybenzene propanoic acid-modified chitosan

The route of PC synthesis is shown in Fig. 2A. The carboxyl groups in the PA structure were activated using EDC and NHS and reacted with the amino groups in CS to form amide bonds, thus producing PC. The PC was then subjected to ¹H-NMR spectroscopy (Fig. 2B) and infrared spectroscopy (Fig. 2C). ¹H-NMR showed that the chemical shifts of δ 7.11–7.13 and δ 6.79–6.81 in PA spectra were attributable to proton absorption peaks at intersite and ortho sites of the benzene ring, indicating that PA was successfully grafted into the CS molecular chain. The absorption peaks in the infrared spectrum of PC at 1643, 1550, and 1243 cm⁻¹ belonged to amide I, amide II, and amide III, further indicating that the carboxyl groups of PA reacted with the amino groups of CS, and PA was grafted to the molecular chain of CS to form PC (Fig. 2C).

PLGA microspheres with sustainable release of KGN were prepared by the single emulsion method. The PLGA microspheres had an obviously spherical morphology and smooth surface (Fig. 3A and B). The particle sizes of unloaded and loaded microspheres were 4.87 ± 1.68 μm and 4.53 ± 1.65 μm, respectively (Fig. 3C). The results showed that the loading of KGN had no significant effects on the diameter of the prepared microspheres. The loading efficiency and encapsulation efficiency of KGN in PLGA microspheres, as shown in Table 1, were 0.0175 ± 0.0009 and 69.87 ± 3.7%, respectively. The results shown in Fig. 3D indicated that the KGN release rates from microspheres loaded with KGN at concentrations of 1, 10, and 100 μM were 18.65 ± 2.78%, 6.99 ± 2.39%, and 4.86 ± 1.66% at 24 h, respectively, indicating marked burst release from microspheres loaded with 1 μM KGN, while those loaded with 10 μM and 100 μM KGN showed no obvious burst release within 24 h. The release rates of the microspheres with the three different KGN concentrations were 79.01 ± 7.44%, 71.89 ± 5.35%, and 48.35 ± 5.78%, respectively, after 28 days. Therefore, KGN could be released slowly by microspheres.

Characterizations of injectable hydrogels

The 1:5 mass ratio of PC/SF for hydrogels with sequential release of SDF-1 and KGN was determined by the preparation and characterization (Fig. S2–S6) of hydrogels with different ratios (mass ratios of 1:0, 5:1, 1:1, 1:5, and 1:10 for PC and SF). The results indicated that the PC/SF (mass ratios of 1:0, 5:1, 1:1, and 1:5) hydrogel had good gelling performance (Fig. S2), injectability (Fig. S3) and porous structures s(Fig. S4), and the PC/SF solutions with a mass ratio of 1:5 also had good stability and mechanical properties (Figs. S5-S6). The PC/SF mixed solutions with/without SDF-1 and microspheres were extruded into different shapes to evaluate their injectability. The results presented in Fig. 4A1–D1 show that the mixed solutions in all groups could form hydrogels with good injectability, which were not affected by the addition of microspheres and drugs. Microstructural observation (Fig. 4A2–D2, Fig. 4A3-D3, and Fig. S7) showed that the hydrogels had porous network structures, and the microspheres were evenly distributed in the hydrogel. The porous structures did not change significantly with increases in the sizes of microspheres and were similar to the structure of the ECM, which is beneficial for cell adhesion, proliferation and differentiation. The PSH, SPSH, SKPSH1, SKPSH10, and SKPSH100 groups had pore sizes of 135.74 ± 55.89, 136.04 ± 54.38, 93.48 ± 32.18, 96.91 ± 32.32, and 128.94 ± 56.7 μm, respectively (Fig. 4E). The drug and microspheres had no effects on the hydrogel pore sizes.

The release profiles of SDF-1 and KGN from the PC/SF injectable hydrogel in vitro were investigated, and the results are presented in Fig. 4F. SDF-1 showed an obvious burst and rapid release, with release rates of 40.68 ± 2.19% at 24 h; the release subsequently slowed, with ~60% of the total SDF-1 released after 7 d. In contrast, the overall release of KGN from the hydrogel was comparatively slow and sustained compared with that of SDF-1; the microspheres loaded with 1, 10, and 100 μM KGN showed sustained release rates of 8.99 ± 2.02%, 5.97 ± 1.85%, and 7.79 ± 1.57%, respectively, at 24 h (Fig. 4F and Fig. S8A) and of 34.45 ± 7.99%, 56.53 ± 6.29%, and 42.17 ± 3.62%, respectively, at 28 days (Fig. 4F and Fig. S8B) in all groups.

The KGN was released slowly for as long as 4 weeks. SDF-1 was released rapidly at the initial stage and throughout the release process, while KGN was released slowly. Therefore, SDF-1 and KGN have obvious sequential release behavior, indicating the spatiotemporally different release behavior of the two bioactive molecules from the hydrogels.

SDF-1 and KGN were loaded into hydrogels by physical blending and microsphere encapsulation strategies, resulting in sequential release behaviors. Moreover, the SKPSH100 group showed obviously reduced mechanical properties with the excessive addition of microspheres (Fig. 4G and Fig. S9). SKPSH100 was not suitable for subsequent experiments. Furthermore, the renewal of body fluids may reduce the concentration of the released drug, so we should select a higher concentration for evaluation in vivo to ensure that it retains good biological function. Therefore, the hydrogel containing 10 μM KGN was more suitable for practical applications.

Evaluation of cellular compatibility of hydrogels

The biocompatibility of hydrogel formulations was studied by evaluating the adhesion and viability of rBMSCs. The results of SEM observations indicated that the hydrogel formulations were conducive to rBMSC adhesion (Fig. S10). The viability of cells seeded onto hydrogel was evaluated by live/dead cell staining (Fig. 5A and Fig. S11) and CCK-8 (Fig. 5B and C). The live/dead cell staining results showed that the cells in the hydrogel were stained green, which indicated that hydrogels are good for cell survival. The quantitative results further revealed that the cells in the hydrogel had high activity, and there was no significant difference among the groups, but the SPSH group and SKPSH group showed a trend of increasing activity at 7 days. This may be related to the function of SDF-1 in promoting cell proliferation. These results indicated that the hydrogel had good cytocompatibility.

The rBMSCs migration in vitro

A Transwell system was used to evaluate the activity of SDF-1 released from hydrogel to recruit rBMSCs in vitro. The transmembrane migration of cells was determined in the control, PSH, SPSH, KPSH10, and SKPSH10 groups (Fig. 6). The numbers of migrating rBMSCs in the SDF-1 groups (SPSH and SKPSH10 groups) were greater than those in the control, PSH and KPSH10 groups; although the average number of migrating cells of SKPSH10 was slightly greater than that of SPSH, there was no significant difference between the SPSH and SKPSH10 groups. This may be related to the function of KGN in promoting cell proliferation to some extent⁴¹. Thus, the results showed that the SDF-1 released from the hydrogel had favorable bioactivity and could promote rBMSC recruitment without the influence of KGN.

Chondrogenesis of rBMSCs in vitro

The chondrogenic differentiation of rBMSCs in hydrogels was evaluated at different incubation times (Fig. 7). H&E staining (Fig. 7A) showed that the hydrogel was beneficial for cell infiltration and proliferation. Positive staining was more obvious in the KPSH10 and SKPSH10 groups, indicating that the ECM content was significantly higher in these two groups than in the control, PSH and SPSH groups after 21 days. With increasing culture time, KPSH10 and SKPSH10 showed increased positive staining with toluidine blue and safranin-O as well as immunohistological staining (Fig. 7A) for COL-II, indicating that hydrogels containing KGN promoted ECM formation. Furthermore, the results of quantitative analysis (Fig. 7B and C) indicated that the hydrogels containing KGN (KPSH10 and SKPSH10 groups) could enhance GAG and COL-II formation. Thus, the KGN released from hydrogel in vitro had effective bioactivity and could promote the chondrogenic differentiation of rBMSCs. The qRT–PCR results indicated the upregulation of chondrogenic genes at 21 days due to the release of KGN in the KPSH10 and SKPSH10 groups compared to the PSH and SPSH groups (Fig. 8).Specifically, the COL-II, ACAN, and SOX-9 mRNA levels were 27.66 ± 4.19, 11.99 ± 0.94, and 12.85 ± 2.95-fold higher in the SKPSH10 group than in the PSH, SPSH, and KPSH10 groups, respectively. Therefore, the hydrogel loaded with SDF-1 and KGN successfully promoted chondrogenic differentiation.

Macroscopic evaluation of regenerated cartilage

The macromorphology of regenerated cartilage was evaluated by optical microscopy after the operation (Fig. 9A). At 4 and 12 weeks, none of the groups showed any obvious signs of infection. After 4 weeks, all materials had been degraded. The MF and PSH groups showed obvious cavities, small amounts of new tissue and an obvious boundary with the surrounding normal tissue. However, there was a great deal of new tissue formation in the defect area with surface roughness and obvious boundaries in the SPSH and KPSH10 groups. However, the defect area was covered by hyaline tissue, and the surface with slight cracks did not show an obvious boundary with surrounding native tissue. At 12 weeks after the operation, the regenerated tissue filled the defect area with obviously cracked surfaces in the MF, PSH, SPSH, and KPSH10 groups. The new tissue completely filled defects with a smooth surface and was tightly combined with the surrounding tissue in the SKPSH10 group. The macroscopic score results in Fig. 9B clearly show a better regeneration effect in the SKPSH10 group than in the other groups.

Recruitment of rBMSCs in vivo

The recruitment of rBMSCs by hydrogel was evaluated by staining for CD44 and CD90 at 7 days after implantation, as shown in Fig. 10. The results shown in Fig. 10A indicated that CD44 (green) and CD90 (red) were positively expressed in the defect area. The fluorescence intensities for CD44 and CD90 on rBMSCs in the SDF-1 groups (SPSH and SKPSH10 groups) were significantly more obvious than those in other groups (i.e., the MF, PSH, and KPSH10 groups). The quantitative results in Fig. 10B and C clearly show that the number of CD44- and CD90-positive cells in the SPSH group and the SKPSH10 group was greater than that in the other groups. However, there was no significant difference between the SPSH group and the SKPSH10 group. These results further indicated that SDF-1 effectively promoted the recruitment of MSCs in vivo.

Restoration of cartilage defects in vivo

The results of H&E staining are shown in Fig. 11. At 4 weeks after the operation, hydrogels in all groups were completely degraded in the defect area. In MF, PSH, SPSH, and KPSH10, there were still defects in the subchondral bone, and fibrous tissues filled the defect area. However, the SKPSH10 group formed continuous subchondral bone, and the defect in subchondral bone caused by microfracture was reconstructed. In the MF and PSH groups, the regenerated cartilage tissues were less abundant and were discontinuous. The new tissues filling the defect areas in the SPSH and KPSH10 groups were more continuous but were still not smooth. The SKPSH10 group showed continuous new tissues in the cartilage defect area with a regular cell arrangement. Even so, the new tissues and natural tissue structure were still significantly different and had clear boundaries in all groups. At 12 weeks, the subchondral bones of all groups were basically reconstructed, but the reconstruction of subchondral bones in the KPSH10 group and SKPSH10 group showed more formation of trabeculae, which was more similar to the natural subchondral bone. Furthermore, the cartilage defect areas of every group were further filled with new tissues, especially in the SPSH, KPSH10, and SKPSH10 groups, which showed obvious regeneration of cartilage tissues. More continuous cartilage structure and mature chondrocyte morphology were observed in the SKPSH10 group than in the other groups. Some chondrocytes were inlaid in the cartilage lacuna, indicating a good regeneration effect.

Staining with safranin-O/fast green was performed to further visualize the regenerated cartilage tissues in the various groups, as shown in Fig. 12. Safranin-O/fast green stained the normal cartilage and subchondral bone light red and light green, respectively. At 4 weeks, no red staining was observed in the defect area in the MF, PSH, or SPSH groups. Fibrous tissues were formed, while the newly formed tissue surfaces in the KPSH10 and SKPSH10 groups showed partial red staining, which indicated the generation of cartilage extracellular matrix (CECM) rich in glycosaminoglycan (GAG), but the new CECM was more continuous in the SKPSH group. At 12 weeks, defect repair was significantly improved in all groups compared with that at 4 weeks, but the regenerated tissue in the MF group was filled with fibrous tissues without GAG formation, indicating that the animals had insufficient self-healing capacity to repair the cartilage in the defect sites. In the PSH, SPSH, KPSH10, and SKPSH10 groups, there were large amounts of CECM deposition containing GAG. In particular, the cartilage showed a continuous structure with a smooth surface in the SPSH, KPSH10, and SKPSH10 groups, similar to the surrounding natural tissues. Moreover, chondrocytes in the SKPSH10 group were closer in morphology, arrangement and distribution to the natural tissue. The results showed that the SKPSH10 group had advantages in the promotion of cartilage defect repair.

The results of immunohistochemical analysis at 4 and 12 weeks showed that positivity for COL-II expression was more continuous in the SKPSH10 group than in the other groups (Fig. 13), indicating an advantage with regard to the promotion of ECM generation and cartilage regeneration in this group. The regenerated cartilage tissue in the SKPSH10 group with uniform features showed no boundary with the surrounding native cartilage tissue, as well as a smooth surface with abundant COL-II. The chondrocytes in the regenerated cartilage showed a typical lacunar structure.

With respect to the histological scores for cartilage evaluation in Fig. 14, the SPSH, KPSH, and SKPSH groups at 4 weeks and at 12 weeks had dramatically higher scores than the MF group at both time points. In particular, the SKPSH group had a higher score than the other groups, which showed that the spatiotemporal release of SDF-1 and KGN could shorten the process of cartilage regeneration. There was no significant difference among the SPSH, KPSH, and SKPSH groups at 12 weeks. The results further and intuitively indicated that sequentially releasing SDF-1 and KGN in vivo had good bioactivity and therapeutic effects. Thus, the hydrogel with the sequential release of SDF-1 and KGN could enhance cartilage regeneration.