Magnetically attracting hydrogel reshapes iron metabolism for tissue repair

Abstract

Ferroptosis, caused by disorders of iron metabolism, plays a critical role in various diseases, making the regulation of iron metabolism essential for tissue repair. In our analysis of degenerated intervertebral disc tissue, we observe a positive correlation between the concentration of extracellular iron ions (ex-iron) and the severity of ferroptosis in intervertebral disc degeneration (IVDD). Hence, inspired by magnets attracting metals, we combine polyether F127 diacrylate (FDA) with tannin (TA) to construct a magnetically attracting hydrogel (FDA-TA). This hydrogel demonstrates the capability to adsorb ex-iron and remodel the iron metabolism of cells. Furthermore, it exhibits good toughness and self-healing properties. Notably, it can activate the PI3K-AKT pathway to inhibit nuclear receptor coactivator 4–mediated ferritinophagy under ex-iron enrichment conditions. The curative effect and related mechanism are further confirmed in vivo. Consequently, on the basis of the pathological mechanism, a targeted hydrogel is designed to reshape iron metabolism, offering insights for tissue repair.

An iron ion–targeted magnetic hydrogel is designed to reshape iron metabolism, providing solutions for tissue repair.

INTRODUCTION

Ferroptosis, a newly discovered form of programmed cell death, has recently received increasing attention (1–3). When iron metabolism is disrupted, intracellular free iron (in-iron) overloading initiates the Fenton reaction, generating abundant reactive oxygen species (ROS). Polyunsaturated fatty acids (PUFAs) on the surface of cell and/or organelle membranes are oxidized to form peroxides, leading to cell death (4). Therefore, reshaping iron metabolism and inhibiting ferroptosis are of great significance for tissue repair. Presently, ferroptosis has been shown to be involved in intervertebral disc degeneration (IVDD) (5, 6). Excessive ROS triggers mitochondrial dysfunction in nucleus pulposus cells (NPCs) and causes an imbalance in extracellular matrix (ECM) synthesis and catabolism (7). Now, intracellular chelating system is developed to remove excess in-iron and inhibit ferroptosis (8). While these therapies have yielded specific results, the source of in-iron remains unexplored, posing challenges in achieving long-term therapeutic effects. The enrichment of extracellular iron ions (ex-iron) is a critical source of in-iron; its influx, under the action of a series of protein molecules such as transferrin receptor 1 (TFR1) and divalent metal ion transporter 1 (DMT1), triggers ferroptosis (9). Ni et al. (10) found that elevated serum iron levels caused ferroptosis in granulosa cells, significantly increasing the risk of infertility resulting from endometriosis. This underscores the importance of eliminating ex-iron accumulation for long-term inhibition of ferroptosis. Nevertheless, changes in ex-iron concentrations during IVDD and their roles in NPCs ferroptosis remain unclear.

In-iron overloading is confirmed to trigger ferroptosis during IVDD, and ex-iron enrichment is widely recognized as a crucial factor affecting in-iron concentrations (11). The enrichment of ex-iron in the nucleus pulposus (NP) during IVDD may subsequently cause in-iron overloading, leading to iron metabolism dysregulation. Preliminary experimental findings from this research unveiled heightened concentrations of ex-iron within degraded NPs. However, the expression of glutathione peroxidase 4 (GPX4) and ferritin heavy chain 1 (FTH1) decreased significantly. This decrease was closely correlated with the degree of degeneration, indicating an increase in ex-iron and the occurrence of ferroptosis in NPCs. Therefore, reducing the concentration of ex-iron could be a breakthrough in correcting iron metabolism and alleviating ferroptosis. Moreover, further exploration is warranted to elucidate the pathways involved in this process. High intervertebral disc pressure hinders the maintenance of the effective concentration of the chelating agents (12). Simultaneously, the microenvironment of NPCs struggles to recover without mechanical support. Hence, constructing a tissue-engineered scaffold capable of targeted adsorption of ex-iron and providing mechanical support would alleviate the dilemma for IVDD.

As a widely applied biomaterial, hydrogels have garnered substantial recognition in intervertebral disc research (13, 14). However, maintaining the injected volume poses a substantial challenge due to the high pressure within the intervertebral disc (15). Polyether F127, being a highly temperature-sensitive polymer material, can undergo solution-gel transformation at physiological conditions (16, 17). It could rapidly fill disc space while cross-linking at physiological temperatures immediately to avoid leakage after injected. However, the low mechanical strength makes it unsuitable for direct application in the treatment of IVDD (18). This research explored the grafting of diacrylate groups to F127 molecular chain [F127 diacrylate (FDA)], endowing it with ideal mechanical properties through photocuring process. Tannin (TA), as a natural compound containing phenolic hydroxyl groups, has the feature to coordinate with metal ions (19, 20). Therefore, it can effectively chelate iron ions and rescue NPCs from ferroptosis under the background of ex-iron overloading.

In this study, to explore changes in iron metabolism during IVDD, we first conducted a histological and molecular biological analysis based on human and rat-degraded NPs. Subsequently, drawing inspiration from the natural phenomenon of magnets attracting metals, we used diacrylate as a bridge to combine F127 and TA, constructing a magnetically attracting hydrogel (FDA-TA). The temperature-sensitive properties of this hydrogel facilitate its distribution and prevent leakage. The photocuring ability of the diacrylate group endows FDA-TA with excellent mechanical properties. Because of the stable grafting of TA through hydrogen bonds, the concentration of ex-iron in degenerative NPs is significantly ameliorated. By benefiting from the adsorption of ex-iron, FDA-TA can inhibit nuclear receptor coactivator 4 (NCOA4)–induced ferritinophagy and maintain in-iron homeostasis. Last, we verified the therapeutic effect of the iron-chelating system in vivo and provided a promising and attractive solution for tissue repair (Fig. 1).

Fig. 1. Schematic illustration of the preparation of magnetic absorbent hydrogel and reshaping iron metabolism in treating IVDD.

RESULTS

Iron ion concentration accumulation and ferroptosis in IVDD

As a new form of programmed cell death, ferroptosis plays a pivotal role in numerous diseases, including tumors, neurodegenerative disorders, and cardiovascular diseases (21, 22). While several studies have delved into its mechanisms, its involvement in IVDD remains incompletely elucidated. As an iron ion–dependent form of programmed cell death, excessive ROS produced by the Fenton reaction is an important part of the process of ferroptosis. During the conversion of Fe²⁺ to Fe³⁺ through the Fenton reaction, excess ROS is generated, which leads to oxidative stress and ferroptosis (8). Although ferroptosis has been shown to be associated with IVDD, the significance of iron overloading in disc degeneration remains a subject of debate. Therefore, investigating the correlation between their concentrations and ferroptosis seems especially notable (23). This study speculated that during IVDD, enriched ex-iron triggers in-iron overloading, inducing ferroptosis in NPCs. Here, we first verified the correlation among ex-iron concentration, degree of IVDD, and ferroptosis in human and rat degenerative NPs (fig. S1A).

On the basis of preoperative magnetic resonance imaging (MRI) examination, patients with Pfirrmann grades II and III were defined as having moderate degeneration (Fig. 2A), whereas those with grades IV and V were defined as having severe degeneration (Fig. 2B) (24). The NP removed during surgery was collected and stained with Prussian blue. We found significantly more brown granules in severely degenerated tissues than in moderately degenerated tissues (Fig. 2C), with a significant difference between the two groups (Fig. 2D). Brown granules in Prussian blue–enhanced staining revealed that iron ion deposits were present in the degenerated NP, and a higher degree of degeneration corresponded to a more significant deposition. GPX4, an essential intracellular antioxidant protein, is a crucial inhibitor of ferroptosis and an indicator of the degree of ferroptosis (25). FTH1, an essential storage protein for in-iron, indirectly reflects changes in concentration (26). Immunofluorescence staining of the NP tissue revealed significantly higher fluorescence intensity in the moderately degenerated group than that in the severely degenerated group (Fig. 2, C and D).

Fig. 2. Verifying the correlation between iron ion concentration, ferroptosis, and degeneration degree in human degenerative NP tissue.

(A) MRI images of moderate degeneration of the intervertebral disc. (B) MRI images of severe degeneration of the intervertebral disc. (C) Prussian blue enhanced staining of the NP tissue with different degrees of degeneration; GPX4 and FTH1 immunofluorescence double-labeled staining. (D) Quantitative analysis of the Prussian blue enhanced staining positive area, semi-quantitative analysis of immunofluorescence intensity (n = 6). (E) WB analysis of GPX4 and FTH1 in different degrees of degeneration (n = 3). (F to I) Correlation analysis between iron ion concentration and GPX4 mRNA expression, FTH1 mRNA expression, GPX4 immunohistochemical positive cell number, and FTH1 immunohistochemical positive cell number. (J) Correlation analysis between MRI grade and tissue iron ion concentration. Statistical analysis was performed using Tukey’s test; statistical significance was set at P < 0.05.

Similarly, the immunohistochemical results showed that the number of GPX4- or FTH1-positive cells in the moderate-degeneration group was significantly higher than that in the severe-degeneration group (fig. S1, B and C). This shows that the degree of ferroptosis was lower in the moderately degenerated group than in the severely degenerated group. We extracted proteins from degenerated tissues to verify the above results at the protein expression level. The Western blot (WB) results showed that GPX4 and FTH1 expressions in the moderate-degeneration group were significantly higher than those in the severe-degeneration group by 1.54 and 1.82 times, respectively (Fig. 2E). Last, we determined the concentration of iron ions in the NP of different patients. After correlation analysis, we discovered that the concentration of ex-iron was significantly negatively correlated with the GPX4 and FTH1 RNA expression levels and the number of positive cells by immunohistochemistry (Fig. 2, F to I) and significantly positively correlated with the degree of degeneration (Fig. 2J). Therefore, in human NP, the ex-iron concentration significantly increases with degeneration, inducing ferroptosis in NPCs.

We constructed an IVDD model for rats’ caudal spine acupuncture to verify this conclusion in rat NP. The MRI results showed that the signal intensity of the NP on T2-weighted images, namely, the “black intervertebral disc,” gradually weakened as degeneration progressed (Fig. 3A) (27). In addition, WB results demonstrated that the expression of GPX4 and FTH1 also gradually decreased with increasing degree of degeneration (Fig. 3B). Prussian blue–enhanced staining showed that the NP in the severe-degeneration group disappeared and was replaced with the annulus fibrosus. In the severe-degeneration group, the dark brown staining area of the intervertebral disc was significantly larger than that in the control and moderate-degeneration groups, indicating a significantly increased ex-iron concentration in the degenerated tissue (Fig. 3, C and D). In addition, the expression of GPX4 and FTH1 was significantly weaker in the degenerated group (Fig. 3, C, E and F). This trend was also reflected in GPX4 and FTH1 immunohistochemistry, with statistically significant differences among the three groups (fig. S1, D and E). Last, correlation analysis between tissue concentration of ex-iron and various indicators was performed, which showed a consistent trend with that of human NP (Fig. 3, G to K). Therefore, at the macro level, we analyzed the changes in ex-iron concentration in human and rat-degenerated NP, revealing a significant increase in concentration as degeneration progressed. Concurrently, ferroptosis in NP was induced during this period. However, at the microscopic level, whether ferroptosis induced by high ex-iron concentrations has the same effect on the structure and function of NPCs, as reported in previous studies (28), requires further investigation. In the forthcoming cellular experiments, we intend to increase the ex-iron concentration using a ferric chloride solution and assess the effect of iron ions on NPCs in vitro.

Fig. 3. Verifying the correlation between iron ion concentration, ferroptosis, and degeneration degree in rat-degenerated NP tissue.

(A) MRI images of rat-degenerated intervertebral discs in different degrees. (B) WB analysis of GPX4 and FTH1 in the degenerated NP tissue (n = 3). (C) Prussian blue enhanced staining in different degrees of degeneration; GPX4 and FTH1 immunofluorescence analysis. (D) Quantitative analysis of Prussian blue enhanced staining positive area (n = 6). (E) Semi-quantitative analysis of FTH1 immunofluorescence intensity (n = 6). (F) Semi-quantitative analysis of GPX4 immunofluorescence intensity (n = 6). (G to J) Correlation analysis between tissue iron ion concentration and GPX4 mRNA expression, FTH1 mRNA expression, GPX4 immunohistochemical positive cell number, and FTH1 immunohistochemical positive cell number. (K) Correlation analysis between MRI grade and tissue iron ion concentration. Statistical analysis was performed using a one-way analysis of variance (ANOVA) with Tukey’s post hoc test. Statistical significance was set at P < 0.05.

Extracellular induction of ferroptosis in NPCs by iron ions

We verified the correlation among ex-iron concentration, ferroptosis, and IVDD in vivo. In addition, we conducted in vitro experiments to further elucidate the impact of ex-iron on NPCs. On the basis of existing research, we stimulated this part of the experiment using a 100 μM ferric chloride solution (29). Tert-butyl hydrogen peroxide (TBHP), a common inducing agent for ferroptosis, was used as a positive control at a concentration of 100 μM (7). We collected NPCs 24 hours after the intervention, and crystal violet staining showed that cell proliferation in the TBHP and Fe³⁺ groups was significantly inhibited, with different cell morphology from that of the control group (fig. S2). WB showed that in the TBHP and Fe³⁺ groups, GPX4 and FTH1 expressions were significantly lower than those in the control group (Fig. 4A), consistent with the immunofluorescence results (Fig. 4, B, C, and F). As an important protein located on the cell membrane facilitating the entry of ex-iron into the cell, TFR1 protein expression is intricately linked to the concentration of ex-iron. When ferric chloride solution was applied, the concentration of iron ions in cells increased, thereby increasing TFR1 expression. In contrast, TBHP induces ferroptosis by increasing intracellular ROS and triggering lipid peroxidation, a mechanism distinct from that of the ferric chloride solution. Consequently, in the WB results presented in Fig. 4A, the TFR1 expression in Fe³⁺ group was higher than that in the TBHP group. DMT1 releases reduced Fe²⁺ from endosomes into unstable iron pools in the cytoplasm. Its expression in the Fe³⁺ group was slightly higher than that in the TBHP group; however, the difference was not statistically significant, and both were significantly higher than that in the control group. Ferroportin (FPN), an essential protein that transports Fe²⁺ outside the cell, has a relatively reduced expression during ferroptosis, aggravating in-iron overloading to a certain extent (9). Therefore, in the TBHP and Fe³⁺ group, FPN expression was significantly lower than that in the control group (Fig. 4A).

Fig. 4. In vitro verifications of ferroptosis induced by iron ions in NPCs.

(A) WB analysis of ferroptosis-related indicators in different groups (n = 6). (B) Semi-quantitative analysis of GPX4 fluorescence intensity (n = 6). (C) GPX4 and FTH1 immunofluorescence staining. (D) FerroOrange probe staining and TEM observation of mitochondrial morphology of NPCs (yellow arrow indicates mitochondria). (E) JC-1 staining to detect mitochondrial membrane potential. (F) Semi-quantitative analysis for FTH1 fluorescence intensity (n = 6). (G) Semi-quantitative analysis for FerroOrange fluorescence intensity (n = 6). (H) Semi-quantitative analysis of JC-1 staining fluorescence intensity (n = 6). (I) Quantitative analysis of flow cytometry of H₂DCFDA and C11-BODIPY581/591 (n = 6). (J) Flow cytometry analysis of H₂DCFDA and C11-BODIPY581/591. (K) Scratch test to evaluate the activity of NPCs in different groups. (L) Quantitative analysis for C11-BODIPY581/591 flow cytometry (n = 6). (M) Scratch test healing quantitative analysis (n = 6). Statistical analysis was performed using a ANOVA with Tukey’s post hoc test. Statistical significance was set at P < 0.05.

Next, we measured the intracellular Fe²⁺ concentration using the FerroOrange probe. The fluorescence intensity in the TBHP and Fe³⁺ groups was significantly stronger. Moreover, the intensity in the Fe³⁺ group was stronger than that in the TBHP group (Fig. 4, D and G), indicating that the intracellular Fe²⁺ concentration was significantly increased in the presence of Fe³⁺ and that the iron metabolism was disturbed. Ferroptosis significantly affects mitochondrial morphology and function. Transmission electron microscope showed that compared with the control group, the mitochondria in the TBHP group were significantly denser, with the cristae disappearing (7). However, in the Fe³⁺ group, the mitochondrial membrane ruptured, resulting in the loss of its normal structure (Fig. 4D).

Regarding mitochondrial function, we used the JC-1 kit to detect changes in mitochondrial membrane potential (30). Following exposure to TBHP and Fe³⁺, the green fluorescence of NP mitochondria was significantly enhanced, indicating a decreased mitochondrial membrane potential and an abnormal function (Fig. 4, E and H). During ferroptosis, the Fenton reaction involving iron ions generates a large amount of ROS, consequently initiating lipid peroxidation of PUFAs in cell or organelle membranes. Thus, we detected ROS generation and lipid peroxidation in the NPCs (31). Flow cytometry and immunofluorescence analyses indicated that ROS generation and lipid peroxidation in the TBHP and Fe³⁺ groups were significantly higher than those in the control group (Fig. 4, I, J, and L, and fig. S2). Notably, in the scratch test, the migration performance of the two NPCs groups was poor after 24 hours, indicating a significant decrease in cell viability (Fig. 4, K and M). Thus, we comprehensively assessed the effect of Fe³⁺ on NPCs in vitro with TBHP as a positive control. The results showed that high ex-iron can cause in-iron overloading, inducing NPCs ferroptosis. On the basis of the in vivo experiments results, the timely removal of ex-iron and the reduction of iron ion concentration significantly inhibit NPCs ferroptosis and delay IVDD. Therefore, in the subsequent experiment, we designed a targeted magnetic hydrogel to adsorb iron ions, thereby inhibiting ferroptosis and reshaping iron metabolism.

Preparation, characterization, and biocompatibility evaluation of targeted magnetic absorbent hydrogel

According to the above findings, ex-iron concentration in the ECM of degenerated intervertebral discs increases, inducing ferroptosis and accelerating IVDD progression. Hence, it is imperative to design a tissue-engineering scaffold capable of magnetically absorbing iron ions for the treatment of IVDD. As polyether F127 has a high temperature sensitivity and can complete solution-gel transformations at physiological temperatures, it is widely used as a drug-loading material (32). However, its poor mechanical properties after gelation limit further development. As the intervertebral disc is a load-bearing body tissue, the histological repair scaffold must have good mechanical properties to ensure structural stability and continuous repair activity. Accordingly, this study introduced propylene groups (diacrylate) into the F127 molecular chain to construct FDA hydrogels, thereby imparting them with photocuring capabilities (33). The exceptional mechanical properties of post-light crosslinking can mimic the mechanical strength of NP. In ¹H nuclear magnetic resonance (NMR) analysis, FDA exhibited a propylene proton peak at δ = 5.8 to 6.4 parts per million compared with F127, indicating successful synthesis (fig. S3A) (34). TA endows the hydrogel with iron adsorption capacity. As a substance naturally rich in phenolic hydroxyl groups, TA has excellent antioxidant capacity and can effectively neutralize ROS. TA has a high-efficiency chelation effect on metal ions, which is vital in the construction of functional composite scaffolds that inhibit ferroptosis in the inner-outer phase (35). FDA readily forms hydrogen bonds, allowing it to connect with TA (17). Fourier transform infrared spectroscopy (FTIR) results showed that FDA-TA had characteristic peaks of FDA and TA hydrogen bonds at 3257 to 3455 cm⁻¹ (fig. S3B) (36). It can be generally observed that FDA and FDA-TA form relatively homogeneous colloids after ultraviolet (UV) light crosslinking. The FDA-TA colloid appeared light yellow after adding TA (Fig. 5A). We first screened out the optimal ratio between FDA and TA because the functional hydrogel must exhibit both iron ion–targeted adsorption and mechanical properties. We combined the two at volume ratios of 2:1, 3:1, and 4:1 (denoted as G2, G3, and G4, respectively) (fig. S4A). Subsequently, we observed the structure of the hydrogel using a scanning electron microscope. The hydrogels had relatively uniform pores (Fig. 5B). The pores of the G2, G3, and G4 groups measured 282.7 ± 77.97 μm, 188.4 ± 76.06 μm, and 184.4 ± 97.68 μm, respectively. No statistically significant difference was observed between the G3 and G4 groups (Fig. 5D). Furthermore, we placed the three colloids in 100 μM ferric chloride solution to evaluate the ability of hydrogels with different proportions to adsorb iron ions in vitro. The overall colloids were black owing to the high proportion of TA, large pores, and strong adsorption capacity in the G2 group. The colloid surface of the G3 group was black, and the iron ion content inside the colloid was lower than that in the G2 group; therefore, the transparency was slightly higher. However, owing to the low TA content and small pores in group G4, most iron ions were adsorbed onto the surface (fig. S4B). The mapping test showed that because of the relatively high TA content, the adsorption capacity of the G2 group was significantly higher than that of the G3 and G4 groups (Fig. 5C). The semi-quantitative analysis results also reflected the same trend (Fig. 5E). We calculated the crosslinking time of different groups to detect the effect of TA content on the crosslinking time. Because of the low TA content, the crosslinking of the G3 and G4 groups was significantly faster than that of the G2 group. The hydrogel strength significantly improved after ultraviolet crosslinking (Fig. 5F and fig. S5A). Simultaneously, we performed degradation and swelling tests on different hydrogel groups in vitro. The hydrogels in all three groups degraded quickly, with <10% of the mass remaining after approximately 14 days, and the overall degradation rate was relatively uniform (fig. S5B). As a foreign material, the long-term retention of hydrogel can have detrimental effects on tissue repair. Group G2 exhibited faster swelling owing to its abundant pores, with its mass approximately 5.5 times greater than at the initial stage after 48 hours (fig. S5C). Various factors are involved in the occurrence and evolution of IVDD (37). The discs’ ability to withstand stress weakens during the process, leading to endplate calcification. Subsequently, the blood vessels shrink, making it challenging for ex-iron to be discharged from the NPs (38). Providing timely mechanical support and enhancing the discs’ stress-bearing capacity could alleviate ex-iron accumulation. This further implies that the mechanical properties of hydrogels are important indicators in treating IVDD. Compressive and tensile modulus tests were conducted separately to assess these properties. Through compressive modulus detection, we found that the compressive moduli of the G2, G3, and G4 groups were 34.61 ± 3.95 kPa, 51.07 ± 1.77 kPa, and 55.07 ± 3.62 kPa, respectively, with no significant difference between G3 and G4 groups (Fig. 5, G and H). The trend in tensile modulus was consistent with this observation, indicating that a lower TA content can better maintain colloidal homogeneity and exhibit superior mechanical properties (Fig. 5, I and J). Thus, in subsequent experiments, we combined FDA and TA at a 3:1 volume ratio to simultaneously achieve iron ion adsorption capacity.

Fig. 5. Preparation and characterization of targeted magnetic hydrogel.

(A) General view of FDA and FDA-TA hydrogels before and after crosslinking. (B) SEM analysis of hydrogels with different volume ratios (G2 = 2:1, G3 = 3:1, and G4 = 4:1). (C) Mapping analysis of different groups of hydrogel adsorption iron ion in vitro. (D) Analysis of hydrogel pore size (n = 15). (E) Semi-quantitative analysis of fluorescence intensity of mapping in different groups (n = 4). (F) UV crosslinking time determination of different groups of hydrogels (n = 3). (G and H) Compression modulus curves and analysis of different groups of hydrogels (n = 3). (I to J) Tensile modulus curves and analysis of different groups of hydrogels (n = 4). (K) Hydrogel rheological testing to assess self-healing. (L) Demonstration of hydrogel toughness. Statistical analysis was performed using an ANOVA with Tukey’s post hoc test. Statistical significance was set at P < 0.05.

The disadvantage of previous hydrogel treatments for IVDD was the easy breakage of colloids, which does not guarantee a therapeutic effect (15). Broken colloids can easily damage the disc and pose a risk of inducing intervertebral disc fibrosis (39). However, this risk can be effectively mitigated if the hydrogel has self-healing ability and good toughness. We performed rheological testing on the FDA-TA, and the hydrogel was assessed by dynamic oscillation sweep frequency within the strain range of 1 to 200%. We observed that the G′ and G″ were periodically changed, and the hydrogel showed excellent self-healing properties (Fig. 5K). Furthermore, the self-healing ability of FDA-TA was visually demonstrated by cutting two complete pieces of colloid in half and reassembling them (fig. S6). Upon pressing the hydrogel, it contracted significantly and rapidly recovered upon release, showcasing its robust toughness (Fig. 5L), as evidenced in our video recording (movie S1).

Biocompatibility is a critical indicator for evaluating the safety of tissue-engineered scaffolds. We performed cell counting kit (CCK-8) assays to assess the effect of the targeted hydrogels on NPCs activity in vitro. The CCK-8 results showed that with the extension of culture time, NPCs continued to proliferate in the hydrogel (fig. S7A), and no dead cells were observed on LIVE/DEAD staining (fig. S7C). Metal ions mediate cell adhesion and enhance activity (40). Therefore, NPCs were cocultured with the hydrogel-adsorbed iron ions. The immunofluorescence results showed that, compared with the Control and FDA groups, the expression of vinculin in the FDA-TA group with adsorbed iron ions was significantly enhanced, and the number of focal adhesions was significantly increased (fig. S7, B and D). The hydrogel block was then placed in the subcutaneous tissue of the rats’ back skins, and hematoxylin and eosin (H&E) staining was performed 3 and 12 days after implantation. We observed that the hydrogel was gradually degraded, and there was no significant inflammatory cell infiltration around the colloid (fig. S8A). Simultaneously, we performed tumor necrosis factor–α (TNF-α) and CD68 immunofluorescence double-labeled staining to evaluate inflammation around the colloid. The overall fluorescence intensity was similar to that of the control group. The fluorescence intensity on day 12 after implantation was weaker than that on day 3 and similar to that of the control group, indicating that no inflammatory reaction occurred after hydrogel implantation (fig. S8B). This suggests that we had successfully prepared a hydrogel scaffold with excellent mechanical properties, good biocompatibility, and the ability to adsorb iron ions. In subsequent experiments, we aim to conduct in vitro and in vivo experiments to evaluate its therapeutic effects comprehensively.

Targeted hydrogel inhibits NPCs ferroptosis in vitro

In the context of tissue dysfunction, abnormal iron metabolism serves as an important cause to irregular iron ion distribution (26). Phenotypic changes in the NPCs contribute to iron metabolism disorders during IVDD (41). The increased concentration of ex-iron exacerbates cellular dysfunction, perpetuating a vicious cycle. Thus, restoring NPCs function and reshaping iron metabolism, as the core components of the intervertebral disc, hold paramount importance in IVDD treatment. On the basis of this understanding, a targeted adsorption hydrogel was innovatively developed. To evaluate the therapeutic effect of the targeted hydrogels in vitro, we cocultured the hydrogels with NPCs (Fig. 6A). On the basis of the coculture components, we divided them into the control, Fe³⁺, FDA, deferoxamine (DFO), and FDA-TA groups. As a widely used iron chelator, DFO was administrated as a positive control to evaluate the ability of targeted hydrogels to chelate iron ions in this study. Following stimulation, a comprehensive evaluation was performed using 100 μM ferric chloride solution for 24 hours. Crystal violet staining showed that in the Fe³⁺ and FDA groups, NPCs proliferated significantly slower, and the cell morphology differed from that of the control group (fig. S9A). WB results showed that GPX4, FTH1, and FPN expressions were significantly weaker in the Fe³⁺ and FDA groups, and the results of immunofluorescence also showed that GPX4 and FTH1 expressions were significantly stronger in the DFO and FDA-TA groups. However, the expression of TNF-α was significantly higher than that in the FDA-TA group (fig. S9, C to G). DFO can chelate intracellular Fe²⁺, and FDA-TA can chelate ex-iron and reshape iron metabolism, thus inhibiting ferroptosis. The expression of these two proteins in the treatment groups was comparable to that observed in the control group.

Fig. 6. The targeted hydrogel reshapes iron metabolism in vitro.

(A) Schematic illustration of coculture experiment. (B to D) WB results and statistical analysis of ferroptosis-related indicators and ECM synthesis-related indicators (n = 6). (E) Semi-quantitative analysis of FerroOrange fluorescence intensity (n = 6). (F) FerroOrange probe staining and TEM observation of mitochondrial morphology in NPCs (yellow arrows indicate mitochondria). (G) Semi-quantitative analysis of fluorescence intensity of JC-1 staining (n = 6). (H) Quantitative analysis of H₂DCFDA by flow cytometry (n = 6). (I) Quantitative analysis of C11-BODIPY581/591 by flow cytometry analysis (n = 6). (J) JC-1 staining of NPCs. (K) H₂DCFDA and C11-BODIPY581/591 flow cytometry analysis. Statistical analysis was performed using an ANOVA with Tukey’s post hoc test. Statistical significance was set at P < 0.05.

In contrast, DMT1 and TFR1 expression was significantly higher in the Fe³⁺ and FDA groups, whereas the expression of ACAN and COL-II, the main components of the ECM, was significantly higher in the DFO and FDA-TA groups (Fig. 6, B to D). Next, we determined intracellular Fe²⁺ concentration. The fluorescence intensity of the probes in the Fe³⁺ and FDA groups was significantly higher than that in the remaining three groups (Fig. 6, E and F). Intracellular Fe²⁺ accumulation was ameliorated to a certain extent because the DFO and FDA-TA groups could chelate in- and ex-iron, respectively. Simultaneously, we observed changes in the mitochondrial structure using TEM. The mitochondrial structure was dense in the Fe³⁺ and FDA groups, and the cristae disappeared. In contrast, the mitochondrial structure in the DFO and FDA-TA groups was relatively normal (Fig. 6F). JC-1 staining was applied to detect changes in mitochondrial membrane potential. In the Fe³⁺ and FDA groups, the green fluorescence intensity was significantly stronger, and the membrane potential decreased (Fig. 6, G and J). Last, ROS generation and lipid peroxidation were evaluated in NPCs. The results showed that the DFO and FDA-TA groups exhibited a significant reduction in ROS and lipid peroxidation, whereas the Fe³⁺ and FDA groups showed significantly increased ROS generation and lipid peroxidation (Fig. 6, H, I, and K), consistent with the immunofluorescence results (fig. S9B). This was also reflected in the scratch test, where NPCs migration and proliferation activities in the Fe³⁺ and FDA groups were significantly lower than those in the other three groups (fig. S10). The results showed that targeting hydrogels could reduce in-iron accumulation, inhibit NPCs ferroptosis in vitro, and maintain cellular activity. To explain the underlying mechanism involved, we aim to conduct further research on deep mining.

Targeted hydrogel inhibits NCOA4-mediated ferritinophagy by activating the PI3K-AKT pathway

The abovementioned analyses suggested that the targeted hydrogel could significantly inhibit NPCs ferroptosis. We collected cells from the Fe³⁺ and FDA-TA groups for transcriptome sequencing analysis to explore the underlying mechanisms. The volcano plot indicated that there were significantly differentially expressed genes between the two groups (fig. S11A), and principal components analysis (PCA) showed that the samples of the two groups had good repeatability (fig. S11B). We performed the Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis to determine possible signaling pathways for the hydrogel’s therapeutic effect. The results indicated significant up-regulation of the phosphatidylinositol 3-kinase (PI3K)–AKT pathway among the top 20 up-regulated pathways (Fig. 7A). In contrast, ferroptosis and autophagy were significantly inhibited among the down-regulated biological changes (Fig. 7B). We performed gene set enrichment analysis (GSEA). The results were consistent with those of the KEGG analysis (Fig. 7C and fig. S11C). Simultaneously, Gene Ontology (GO) enrichment analysis was performed to analyze changes in cell biological behavior (fig. S12). Therefore, we speculated that the targeted hydrogel inhibited specific types of cell autophagy by up-regulating the PI3K-AKT pathway, thereby inhibiting ferroptosis. Ferritin, a vital protein in iron storage, plays a crucial role in maintaining iron homeostasis by dynamically regulating the uptake, storage, and excretion of iron ions (42). Ferritinophagy, coordinated by NCOA4, is an important mechanism for the release of iron ions from ferritin (43). Physiologically, when the in-iron concentration decreases or the demand increases, ferritinophagy is activated under the mediation of NCOA4. This process releases free iron ions to maintain iron homeostasis. However, in pathological conditions where the intracellular iron concentration rises sharply, ferritinophagy is also activated, leading to iron overload and subsequent ferroptosis (26). As reported by Zhou et al. (11), ionizing radiation caused an increase in in-iron, initiating NCOA4-mediated ferritinophagy and leading to ferroptosis in intestinal epithelial cells. Therefore, NCOA-mediated ferritinophagy, as a double-edged sword, has a significant effect on cellular iron homeostasis. The PI3K-AKT pathway is a primary regulator of autophagy (44, 45) and is also involved in ferritinophagy. In a study based on Alzheimer’s disease, apolipoprotein E inhibited ferritinophagy and blocked iron-dependent lipid peroxidation by activating the PI3K-AKT pathway (46, 47). However, whether NCOA4-mediated ferritinophagy occurs during IVDD and whether targeted hydrogels inhibit autophagy through the PI3K-AKT pathway remain unknown. Therefore, on the basis of the sequencing results, we speculate that PI3K-AKT may be activated during the targeted hydrogel treatment of IVDD, thereby inhibiting ferritinophagy caused by in-iron overloading and inhibiting NPCs ferroptosis.

Fig. 7. The targeted hydrogel activates the PI3K-AKT pathway to inhibit ferritinophagy in NPCs.

(A) KEGG enrichment analysis of up-regulated signaling pathways. (B) KEGG enrichment analysis of down-regulated biological behaviors. (C) PI3K-AKT, ferroptosis, autophagy-animal GSEA enrichment analysis. (D) LC3b and NCOA4 protein expression in human degenerative NP tissue. (E) p-PI3K and p-AKT immunohistochemical staining of human-degenerated NP tissue. (F) LC3b and NCOA4 protein expression in rat-degenerated NP tissue. (G) p-PI3K and p-AKT immunohistochemical staining of rat-degenerated NP tissue. (H) LC3b immunofluorescence staining of NPCs. (I) Semi-quantitative analysis of LC3b immunofluorescence intensity (n = 5). Statistical analysis was performed using an ANOVA with Tukey’s post hoc test. Statistical significance was set at P < 0.05. FDR, false discovery rate; NES, normalized enrichment score.

Human and rat degenerative NP tissues were analyzed again to test this hypothesis. WB assay showed that light chain 3 beta (LC3b) II/I and NCOA4 expressions were significantly increased in patients with severe degeneration compared with those with moderate degeneration (Fig. 7D and fig. S13A). However, immunohistochemical analysis indicated that the numbers of p-PI3K–and p-AKT–positive cells in the severe-degeneration group were lower than those in the moderate-degeneration group (Fig. 7E and fig. S14A). This trend was also observed in the rat NP (Fig. 7, F and G, and figs. S13B and S14B). For in vitro experiments, LC3b immunofluorescence analyses were performed on the NPCs. LC3b spots were more frequent, and the fluorescence intensity was significantly stronger in the Fe³⁺ and FDA groups (Fig. 7, H and I). WB analysis showed that the LC3bII/I and NCOA4 expressions were significantly stronger in the Fe³⁺ and FDA groups, whereas the p-PI3K/PI3K and p-AKT/AKT expressions were significantly enhanced in the FDA-TA group (fig. S15). As a chelating agent for iron ions, DFO reduces the concentration of in-iron, thereby inhibiting ferritinophagy. Consequently, LC3bII/I and NCOA4 expressions were significantly decreased in the DFO group. Thus, FDA-TA and DFO ameliorated the concentrations of ex- and in-iron, respectively, and inhibited ferritinophagy.

Furthermore, in this part of the experiment, we hypothesized the mechanism of targeted hydrogel therapy through transcriptome sequencing. We conducted both in vivo and in vitro experiments to validate this hypothesis. Ultimately, we assessed the efficacy of the targeted hydrogel therapy using a rat IVDD model.

Targeted hydrogel mitigated ferritinophagy in vivo, inhibited ferroptosis, and treated rat IVDD

To assess the effects of targeted hydrogel therapy in vivo, we first observed the degradation of the FDA and FDA-TA hydrogels by in vivo imaging. The degradation rates of the two groups were similar, and the fluorescence intensity decreased significantly 3 days after the injection. This may be attributed to the increase in inflammatory factors caused by intervertebral disc injury after modeling and accelerated colloidal degradation. Subsequently, the degradation rate slowed down, and the fluorescence signal persisted until day 12 (fig. S16). We then established the rat caudal vertebral acupuncture IVDD model by puncturing the coccygeal vertebrae Co6/7, Co7/8, Co8/9, and Co9/10 and injected solutions of ferric chloride, FDA, DFO, and FDA-TA, respectively, with Co5/6 serving as the positive control group (Fig. 8A). Radiological examination was performed 4 and 8 weeks after surgery. The main indicators were the MRI T2-weighted image signal intensity of the intervertebral disc and the vertebral space height (Fig. 8B).

Fig. 8. The targeted hydrogel reshapes iron metabolism and IVDD in vivo.

(A) Overview of animal experiments. (B) MRI and radiographic analysis of rat degenerated intervertebral disc. (C) DHI% analysis of the intervertebral disc space. (D) H&E staining 4 and 8 weeks after treatment. DHI, disc height index. (E) S&O staining 4 and 8 weeks after treatment. (F) LC3b and NCOA4 immunohistochemical staining after 8 weeks of treatment. (G) Statistical analysis of MRI classification 4 weeks after treatment (n = 5). (H) Statistical analysis of MRI classification 8 weeks after treatment (n = 5). (I) Statistical analysis of histological findings 4 weeks after treatment (n = 5). (J) Statistical analysis of histological findings 8 weeks after treatment (n = 5). Statistical analysis was performed using a one-way or two-way ANOVA with Tukey’s post hoc test. Statistical significance was set at P < 0.05.

By calculating the 4- and 8-week disc height index, we observed a decreased vertebral space height in all experimental groups. However, the decline was slower in the FDA-TA group and significantly different from that in the control group at week 8, indicating superior results compared to the other experimental groups (Fig. 8C). MRI findings indicated that the intervertebral disc signal intensity decreased to a certain extent in all experimental groups as the degeneration time progressed. However, the signal strength of the FDA-TA group was closest to that of the control group (Fig. 8, G and H). As an in-iron chelator, DFO could prevent iron overloading. However, it is prone to leakage, making it difficult to maintain volume. In addition, DFO cannot provide mechanical support for degenerative intervertebral discs, thereby failing to enhance the mechanical environment necessary for subsequent cell proliferation. Therefore, intervertebral disc recovery is limited (48). We performed H&E and Safranin-O (S&O) fast green staining (Fig. 8, D and E) to observe the tissue volume of the NP and structural boundaries of the intervertebral discs. The results showed that the NP in the Fe³⁺ and FDA groups gradually decreased, and by the fourth to eighth week, it was replaced by the fibrous annulus tissue with an unclear structure. In the DFO group, while some NP persisted, the boundary between the NP and annulus fibrosus tissues was blurred. In contrast, in the FDA-TA group, the NP volume increased with a clear boundary, and the histological score was similar to that of the control group (Fig. 8, I and J).

As mentioned in the previous sections, we explored the mechanism of action of the targeted hydrogel in detail; however, in vivo verification was still required to evaluate its therapeutic effect comprehensively. In the eighth week, we performed Prussian blue–enhanced staining of different groups of NPs to determine the concentration of ex-iron. The results showed that the positive staining area of the FDA-TA group was significantly smaller, similar to that of the control group, indicating excellent iron chelating ability (fig. S17, A and B). The LC3b and NCOA4 immunohistochemistry results suggested that the number of positive cells in the Fe³⁺ and FDA groups was significantly higher than that in the DFO and FDA-TA groups, indicating that the autophagy process involving NCOA4 was inhibited in the DFO and FDA-TA groups (Fig. 8F and fig. S18). Simultaneously, we performed immunofluorescence staining and observed that the fluorescence intensities of GPX4 and FTH1 in the DFO and FDA-TA groups were significantly stronger (fig. S17, A and C). Last, we performed WB to analyze the expression of ferroptosis-related proteins after treatment (fig. S19) and observed that GPX4, FTH1, p-PI3K, and p-AKT expression was weaker in the Fe³⁺ and FDA groups. In addition, LC3B and NCOA4 expression was significantly higher in the Fe³⁺ and FDA groups.

Therefore, in this study, we thoroughly verified the effect of targeted hydrogels in weakening ferritinophagy, reshaping iron metabolism, and treating IVDD in vivo. Targeted hydrogel exerts dual regulation based on intervertebral disc structure and NPCs function, effectively solving the problem of ex-iron accumulation. On the one hand, its photocrosslinking properties provide mechanical support, restore the physiological structure of the discs, and accelerate the excretion of ex-iron. FDA-TA acts as an “ artificial iron pool” by chelating iron ions, thereby reducing ex-iron concentration in degenerative conditions. As the hydrogel degrades, the adsorbed iron ions are released in a controlled manner, facilitating the restoration of cellular iron metabolism and maintaining spatiotemporal regulation of iron homeostasis. Consequently, with the recovery of disc structure and function, ex-iron is either excreted through substance exchange or reintegrated into normal iron metabolism, significantly reducing ex-iron concentration.

DISCUSSION

This study demonstrated that the concentration of ex-iron significantly increased in the degenerated NP tissues and was significantly positively correlated with the degree of degeneration. Accordingly, a magnetic FDA-TA hydrogel targeting ex-iron was designed. In vitro studies showed that the magnetic hydrogel significantly scavenged ROS and iron metabolism disorders induced by accumulating ex-iron. Transcriptome sequencing revealed that the targeted hydrogel activated the PI3K-AKT pathway, thereby inhibiting NCOA4-mediated ferritinophagy. The in-iron homeostasis was maintained through “internal and external linkages.” Last, the therapeutic effects and mechanism of the targeted hydrogel were confirmed in vivo. On the basis of the pathophysiological changes observed in IVDD, we developed an ex-iron–targeted magnetic hydrogel to reshape iron metabolism, offering a promising clinical solution for tissue repair.

MATERIALS AND METHODS

Reagents and antibodies

F127, diacrylate, TA, TBHP, and ferric chloride were purchased from Shanghai Aladdin Company. A JC-1 kit (Abmole, M1045), FerroOrange reagent (Dojindo, F374), H₂DCFDA (Abmole, M9096), C11-BODIPY581/591 kit (Thermo Fisher Scientific, D3861), Iron Microplate Assay Kit (Elabscience, E-BC-K772-M), Live and Dead staining Kit (Yeasen), GPX4 (SAB, 32506), FTH1 (SAB, 32180), DMT1 (Proteintech, 20507-1-AP), FPN (Abcam, ab92477), TFR1 (Proteintech, 26601-1-AP), NCOA4 (SAB, 32981), LC3b (SAB, 46604), PI3K (SAB, 41338), p-PI3K (SAB, 11508), AKT (SAB, 21054), p-AKT (SAB, 11054), COL-II (Abcam, ab188570), ACAN (Abcam, ab52141), vinculin (SAB, 41534), CD68 (Abcam, ab283654), TNF-α (Abclonal, a11534), and Dulbecco’s modified Eagle’s medium/F12 culture medium (Gibco).

Tissue collection of degraded NP in humans and rats

On the basis of preoperative MRI findings, we collected NPs from patients with grades II to V degeneration following the Pfirrmann scale of disc degeneration (24). Notably, all NPs (30 cases) were obtained from patients undergoing percutaneous endoscopic lumbar discectomy or posterior lumbar interbody fusion for IVDD. This study was approved by the Ethics Committee of the First Affiliated Hospital of Soochow University (approval number: 2023180). Written informed consent was obtained from patients or their relatives before tissue collection. The relevant clinical information of the patients is presented in table S1. We defined grades II to III as moderate degeneration and grades IV to V as severe degeneration. The Sprague Dawley (SD) rats were purchased from the Experimental Animal Center of Soochow University, and the Ethics Committee of Soochow University approved the experiments (approval number: SUDA20230619A01). NPs with different degrees of degeneration were obtained by setting different puncture depths based on a previously research (49). Specifically, to achieve severe degeneration, the needle was inserted into the center of the NP (5 mm in depth) and rotated it for 5 s and maintained it for 30 s. As for moderate degeneration, we inserted the needle into the edge of NP (3 mm in depth), pulled it out after 5-s rotation, and kept stationary for 30 s. After model construction, the rats were divided into normal-, moderate-degeneration, and severe-degeneration groups based on MRI findings. NPs from 10 rats in each group were collected for follow-up experiments.

Extraction of rat NPs and induction of ferroptosis in vitro

After the rats were euthanized under excessive anesthesia, their caudal vertebrae were removed under sterile conditions. All the paravertebral muscles were separated, annulus fibrosus was incised using a sharp knife blade, and NPs were removed using eye forceps. The collected NP tissues were digested in 0.25% type II collagenase for 2 hours, filtered, and centrifuged, and the supernatant was discarded. The complete medium was added after three washes in sterile phosphate-buffered saline (PBS). The extracted cells were cultured in a cell incubator, the medium was changed every 3 days, and the first three generations of cells were used for subsequent experiments. In this study, we separately induced ferroptosis with 100 μM TBHP and FeCl₃ solution for 24 hours.

Western blotting

Proteins were extracted from the NPs or cells using the radioimmunoprecipitation assay reagent containing protease and phosphatase inhibitors. After determining the protein concentration, the proteins were separated via electrophoresis with a 10 or 20% gradient gel at 150 V. The target proteins were then transferred to a polyvinylidene difluoride membrane at 400 mA. After isolation, they were incubated overnight at 4°C with a diluted primary antibody comprising GPX4 (1:1000; SAB, 32506), FTH1 (1:1000; SAB, 32180), DMT1 (1:1000; Proteintech, 20507-1-AP), FPN (1:2000; Abcam, ab92477), TFR1 (1:1000; Proteintech, 26601-1-AP), NCOA4 (1:1000; SAB, 32981), LC3b (1:1000; SAB, 46604), PI3K (1:1000; SAB, 41338), p-PI3K (1:1000; SAB, 11508), AKT (1:1000; SAB, 21054), p-AKT (1:1000; SAB, 11054), COL-II (1:1000; Abcam, ab188570), and ACAN (1:1000; Abcam, ab52141). After washing with Tris-Buffered Saline with Tween (TBST), proteins were incubated with a diluted secondary antibody for 2 hours at room temperature. After the bands were captured using the imaging system, β-actin was used as an internal reference to analyze the gray-value protein expression.

Mitochondrial membrane potential detection

Changes in cell membrane potential in different groups were detected using the JC-1 reagent. Following the instructions, the treated cells were washed with PBS, and 3 μM JC-1 reagent was added and incubated at 37°C for 30 min. After washing with PBS, the membrane potential was observed under a confocal laser microscope.

Lipid peroxidation detection

We evaluated lipid peroxidation in different cell groups using the BODIPY 581/591 C11 Kit. The cells were treated as described above, washed with PBS, and digested with pancreatic enzymes, and the suspended cells were collected. Furthermore, 10 μM BODIPY 581/591 C11 solution was added for suspension and incubated at 37°C for 30 min away from the light. The supernatant was centrifuged and washed with PBS for flow cytometry analysis.

Crystal violet staining

After cell intervention, the medium was removed, and the cells were washed thrice with PBS. Cells were fixed with 4% paraformaldehyde for 10 min and washed with PBS for 2 min. Crystal violet dye (100 ml; Beyotime, C0121) was then added for 10 min. After washing with PBS, the cells were observed and photographed under a light microscope.

ROS detection

H₂DCFDA was used to detect ROS. After the cell intervention, the suspension cells were digested with pancreatic enzymes. The cells were thoroughly washed with PBS, and a 5 μM H₂DCFDA working solution was added and incubated at 37°C for 30 min away from light. The supernatant was centrifuged and washed with PBS for flow cytometry analysis.

Intracellular ferrous ion detection

The FerroOrange probe was used to detect intracellular ferrous ion concentration. After the intervention, the cells were washed with PBS, and FerroOrange working liquid at 1 μM was added, cultured at 37°C and 5% CO₂ for 30 min, and then observed under a fluorescence microscope.

Transmission electron microscopy

Different groups of cells washed with PBS were collected using a cell spatula and centrifuged. The supernatant was discarded, slowly injected into 2.5% glutaraldehyde, fixed for 1 hour, and then fixed in 2% osmium tetroxide for 3 hours. After washing, the cells were stained with 0.5% uranyl acetate for 12 hours. After dehydration and polymerization, the samples were sliced into 70- to 90-nm ultrathin sections using an ultramicrotome (EMUC7, Leica), and the mitochondrial morphology of different NP groups was assessed via TEM (FEI, USA).

Scratch test

Scratch tests are often applied to assess cell migration ability and NPCs activity during IVDD (50, 51). The cells were seeded on a six-well plate, and the corresponding intervention was performed on the basis of the group. When cell fusion reached 100%, the pore plate surface was scratched perpendicularly using a 200-μl pipette tip, and the cells were cultured for 24 hours. After scratching, images were captured at 0 and 24 hours, and the NP migration ability was analyzed and compared.

Immunofluorescence staining

After the intervention of different cell groups, the culture medium was aspirated. Furthermore, 4% paraformaldehyde was added after washing thrice with PBS, and the cells were fixed at room temperature for 30 min. After drilling with Triton X-100, 5% bovine serum albumin solution was added to the block overnight at 4°C. After washing with PBS, the diluted primary antibody was added and incubated overnight at 4°C. The primary antibody was removed via suction; cells were washed with PBS thrice and incubated with a secondary antibody at room temperature for 2 hours in the dark. After incubation with rhodamine and 4′,6-diamidino-2-phenylindole staining solutions at room temperature, the cells were observed and photographed under a laser confocal microscope. Semi-quantitative fluorescence intensity was analyzed using the ImageJ software.

qRT-PCR analysis

Total RNA was extracted from cells using the TRIzol method based on previously described experimental procedures. Reverse transcription and quantitative amplification of cDNA were performed using a PrimeScript RT Kit (Takara, Japan) and SYBR Premix Ex TaqTM (Takara, Japan). Using glyceraldehyde-3-phosphate dehydrogenase as an internal reference gene, rat primers for quantitative reverse transcription polymerase chain reaction (qRT-PCR) were synthesized and provided by Jinweizhi Technology Co. Ltd. The primer sequence was as follows: GPX4 (5′-3′: GAGCCCATTCCCGAGCCTTTC, 3′-5′: CTCGTGGCTGTTGCTGGTCTG) and FTH1 (5′-3′: CGCTTTGAGCCTGAGCCCTTTG, 3′-5′: CCTCCGAGTCCTGGTGGTAGTTC). The 2^-ΔΔCt method was used to analyze target gene expression level and perform differential analysis.

Transcriptome sequencing

We cocultured the NPCs with the FDA-TA hydrogel, stimulated the NPCs with ferric chloride solution for 24 hours, extracted the total protein with TRIzol reagent, and measured and quantified the RNA purity using a NanoDrop 2000 spectrophotometer. Transcriptome sequencing and analyses were performed by OE Biotech Co. Ltd. (Shanghai, China). Clean reads were obtained using Trimmomtic software, and hisat2 was compared with reference genes. The fragments per kilobase of exons per million reads mapped were quantitatively obtained using Cufflinks software. GO and KEGG enrichment analysis of differential transcripts were performed on differential genes with a P value < 0.05 and a multiple of difference > 1.5 to determine the biological functions or pathways mainly affected by differential transcripts. Simultaneously, unsupervised hierarchical clustering of differential transcripts was performed, and the expression patterns of differential transcripts in different samples were displayed as heatmaps.

Iron ion detection in the NPs

A standard curve was developed using the standard solution following the manufacturer’s instructions. Therefore, 0.1 g of human or rat NP tissue from each group was used for detection. After the operation, absorbance at 593 nm was detected using a microplate reader, and the iron ion concentration of each sample was calculated on the basis of the standard curve.

Correlation analysis

On the basis of the correlation analysis between iron ion concentration and the expression levels of GPX4 and FTH1 mRNA in the tissue, the number of positive cells, and MRI T2-weighted images, the correlation between iron ion concentration and the degree of ferroptosis and degeneration was analyzed.

FDA-TA hydrogel preparation

The FDA synthesis was based on previous studies (52). We dissolved 10 g of F127 and 10 mol of excess ethylamine in 100 ml of anhydrous dichloromethane under nitrogen and then introduced 10 mol of excess propylene chloride into a three-neck flask through a drop funnel. The mixture was stirred at 25°C nitrogen for 24 hours, the precipitated trichloroethylene amine was filtered out, and the anhydrous ether was added to the filtrate. The product was dried at 30°C for 24 hours under vacuum. The obtained FDA and TA were mixed in 2:1, 3:1, and 4:1 volume ratios to prepare the hydrogel and placed in a ferric chloride solution to evaluate the adsorption effect.

Hydrogel characterization

Proton NMR (¹H NMR, Bruker Advance III, USA) and FTIRs (Nicolet 6700, Thermo Fisher Scientific, USA)

¹H NMR and FTIR were used to analyze the allylation and FDA hydrogel chemical bond formation.

Scanning electron microscopy

Hydrogels with different proportions were freeze-dried and glued onto the surface of the stage using a conductive adhesive. After gold spraying, the morphologies and pore sizes of the hydrogels in each group were observed and statistically analyzed (Crossbeam 300, Zeiss). Hydrogels that adsorbed iron ions in the different groups were analyzed by mapping.

Compression and tensile moduli testing

The macroscopic mechanical properties of the hydrogels were tested using a general mechanical testing machine (Hengyi HY-1080, Shanghai, China). The hydrogels were prepared in columns or dumbbells, pretreated in PBS for 12 hours, and subjected to compression/tension tests. The compressive modulus of the hydrogel was obtained by calculating the slope of the stress-strain curve passing through the origin, and the elongation of the hydrogel before failure was calculated from the tensile displacement (E1) and original length (E0) of the hydrogel as following formula

$$\text{Stretch rate}=\frac{E1}{E0}*100\%$$

Detection of rheological properties

Rheological properties of the hydrogels were measured using an HR-2 rheological instrument (TA Instrument). The hydrogel sample was transferred to the middle of a parallel plate with 15 mm in diameter, and an appropriate gap was left. A dynamic oscillatory sweep frequency was measured in the 1 to 200% strain range. The frequency of the strain-scanning measurements was set to 10 rad/s. The self-healing capacity was measured by repeating the 1 and 200% dynamic strain tests. Simultaneously, we crosslinked hydrogels with different concentrations of UV light and measured the dynamic changes of G′ and G″ during the crosslinking process.

Self-healing performance

The self-healing ability was further verified by cutting two fully formed scaffolds in half, dyeing them with different colors, and rejoining them. Tensile stress was applied manually with tweezers after allowing healing for 15 min.

Swelling test

The swelling rate was measured using lyophilized hydrogel samples. First, the dry weight (W1) of the sample was recorded. After the sample was placed in PBS to expand for different times at room temperature, its wet weight (W2) was measured, and the expansion rate was calculated on the basis of the following formula

$$\text{Swelling ratio}=\frac{W2-W1}{W1}*100\%$$

Degradation test

Three groups of hydrogels with the same mass (W0) were taken, 10 ml of PBS was added to each group and incubated at 37°C with slow stirring. The hydrogels were lyophilized and weighed (Wt) at 0, 2, 6, 10, and 14 days. The residual mass was calculated on the basis of the following formula

$$\text{Mass remaining} (\%)=\frac{W0-\mathit{Wt}}{W0}*100\%$$

In vitro biocompatibility evaluation

Targeted hydrogel co-cultured with NPCs

Slides were placed in a well plate, and then the hydrogel solution was spread on the surface and immediately crosslinked with UV light. After washing three times with PBS, the cell suspension was seeded, and intervention was performed for 24 hours after adhered, and then experimental evaluation was performed.

Cell proliferation experiment

A cell adherent culture was used as a control, and the CCK-8 Kit (Dojindo, Japan) was used to detect cell proliferation among different groups. At appropriate time points (1, 3, 5, and 7 days), a medium containing a 10% volume ratio of CCK-8 reagent was added. After incubating for 4 hours, 100 μl of the culture medium was added to each well of a new 96-well plate, and the absorbance at 450 nm was measured using a microplate reader.

Cell viability assay

To assess NPCs viability after inoculation, a LIVE/DEAD staining assay was performed. On the first, third, fifth, and seventh day after cell culture, LIVE/DEAD staining reagent (Yeasen, 40747ES80) was added to the wells and incubated at room temperature for 30 min. The cells were then observed under a fluorescence microscope.

In vivo experiment

Hydrogel subcutaneous implantation model

SD rats weighing 300 to 350 g were used to conduct the subcutaneous implantation experiments using different hydrogels. After the rats were anesthetized with 2% pentobarbital (2.5 ml/kg), the back skin was shaved and disinfected. Two longitudinal incisions were made on both sides to make a subcutaneous pocket, and a 100 μl of gel block was implanted into the pocket. Tissues were removed 3 and 12 days postoperatively for histological staining.

Animal model establishment and treatment

A rat acupuncture model of IVDD was established to evaluate the therapeutic effects of the targeted hydrogel. First, the animals were anesthetized using an intraperitoneal injection of 2% pentobarbital sodium at a concentration of 2.5 ml/kg. After complete anesthesia, 21-gauge needles were used to puncture the Co6-7, Co7-8, Co8-9, and Co9-10 intervertebral discs, and Co5-6 was used as the control. When puncturing the center of the intervertebral disc (5 mm in depth), it was rotated for 5 s and then kept stationary for 30 s. According to previous research, to obtain the best therapeutic effect while avoiding detrimental effect to the intervertebral disc, we chose an injection volume of 20 μl (15, 53). After the puncture, 20 μl of ferric chloride solution, F127-DA, DFO, and F127DA-TA solution were injected, respectively, and the hydrogel group was injected with an UV point light source for in vivo crosslinking. Simultaneously, to evaluate the degradation rate of the hydrogel in the intervertebral disc, we injected Cy5 (TargetMol, T15025L) hydrogel into Co6-7 and measured and analyzed the fluorescence intensity using a small animal in vivo imaging system (IVIS Lumina Xrms) at 0, 3, 6, and 12 days postoperatively.

Imaging analysis

Radiography and MRI were conducted on rats at 4 and 8 weeks after surgery to observe intervertebral space height and disc signal intensity changes. The disc height index (DHI) was calculated on radiography, and the ratio of DHI of the experimental group to that of the control group was recorded as the DHI%. Each disc was graded on the basis of MRI T2-weighted findings with reference to modified Thomson grading.

Histological evaluation

After an anesthesia overdose was used to euthanize the rats, the intervertebral disc tissue was removed, fixed with formalin together with the human NP tissue, and then placed in 10% EDTA solution for decalcification. After embedding into wax blocks, the tissue specimens were sliced into 5-μm sections, including endplate, NP, and annulus fibrosus. After H&E and S&O staining, the intervertebral discs in different groups were assessed histologically as described previously (12). Simultaneously, to observe indicators such as ferroptosis, ferritin autophagy, and related signaling pathways, enhanced Prussian blue staining, immunohistochemistry, and immunofluorescence staining were performed.

Statistical analysis

All experimental data are expressed as means ± SDs. Statistical analysis (Origin 9.1 or GraphPad Prism 7.0 software) was performed by a one- or two-way analysis (ANOVA) of variance using Tukey’s multiple comparison test to evaluate the differences among groups unless otherwise stated. Statistical significance was set at P < 0.05.

Supplementary Materials

The PDF file includes:

Figs. S1 to S19

Table S1

Legend for movie S1

Other Supplementary Material for this manuscript includes the following:

Movie S1