Immune-defensive microspheres promote regeneration of the nucleus pulposus by targeted entrapment of the inflammatory cascade during intervertebral disc degeneration

Abstract

Sustained and intense inflammation is the pathological basis for intervertebral disc degeneration (IVDD). Effective antagonism or reduction of local inflammatory factors may help regulate the IVDD microenvironment and reshape the extracellular matrix of the disc. This study reports an immunomodulatory hydrogel microsphere system combining cell membrane-coated mimic technology and surface chemical modification methods by grafting neutrophil membrane-coated polylactic-glycolic acid copolymer nanoparticles loaded with transforming growth factor-beta 1 (TGF-β1) (T-NNPs) onto the surface of methacrylic acid gelatin anhydride microspheres (GM) via amide bonds. The nanoparticle-microsphere complex (GM@T-NNPs) sustained the long-term release of T-NNPs with excellent cell-like functions, effectively bound to pro-inflammatory cytokines, and improved the release kinetics of TGF-β1, maintaining a 36 day-acting release. GM@T-NNPs significantly inhibited lipopolysaccharide-induced inflammation in nucleus pulposus cells in vitro, downregulated the expression of inflammatory factors and matrix metalloproteinase, and upregulated the expression of collagen-II and aggrecan. GM@T-NNPs effectively restored intervertebral disc height and significantly improved the structure and biomechanical function of the nucleus pulposus in a rat IVDD model. The integration of biomimetic technology and nano-drug delivery systems expands the application of biomimetic cell membrane-coated materials and provides a new treatment strategy for IVDD.

Graphical abstract

Highlights

• •GM@T-NNPs eliminate inflammatory cytokines during intervertebral disc degeneration.
• •GM@T-NNPs inhibit the PI3K-AKT signaling pathway and repair the extracellular matrix in vitro.
• •GM@T-NNPs improve the structure and biomechanical function in vivo.

1. Introduction

Long-term chronic inflammation is an important feature of intervertebral disc degeneration (IVDD). High levels of inflammatory cytokines, such as tumor necrosis factor-alpha (TNF-α) and interleukin (IL) family members (e.g., IL-1, IL-6, and IL-17), secreted by the nucleus pulposus and annulus fibrosus cells play major roles as initiators and promotors of IVDD [1]. Long-term inflammation leads to an imbalance between synthesis and catabolism in the nucleus pulposus tissue, resulting in disruption of the normal tissue repair process that occurs in IVDD [2,3]. Inflammation stimulates phenotypic changes in nucleus pulposus cells, which promotes the secretion of proteins, such as disintegrin, a matrix metalloproteinase (MMPs) with thrombospondin motifs, and other proteases, followed by inhibition of the synthesis of type II collagen (COL-II) and aggrecan, thereby affecting the accumulation of extracellular matrix (ECM). Consequently, the nucleus pulposus tissue forms a fibrotic structure that is less hydrated and less tolerant to biomechanical damage [4]. During the process of degeneration, the annulus fibrosus along with the endplates is prone to structural defects or nucleus pulposus prolapse, which disrupts the mechanical barrier around the nucleus pulposus, enabling the infiltration of immune cells and thereby altering the immune microenvironment of the intervertebral disc [5]. Simultaneously, microvascular infiltration stimulated by inflammation provides an additional channel for immune cells, including neutrophils, macrophages, and T cells, to migrate to the intervertebral discs, followed by the activation of the downstream inflammatory cascade and pain aggravation. Numerous studies have shown that during the process of IVDD, the expression levels of C–C motif chemokine ligand 3 in nucleus pulposus cells increase under the stimulation of TNF-α, IL-1β, and other inflammatory factors [6].

We previously constructed microspheres functionalized with cytokine antagonist materials to reduce inflammation in the nucleus pulposus and maintain synthesis and catabolic homeostasis [7]. However, most biological antagonists that function as protein-based drugs have short in vivo half-lives and are rapidly eliminated from the body. Moreover, the use of a single cytokine antagonist is insufficient to address the complex and unpredictable immune microenvironment of the intervertebral disc. Therefore, the design of functional nanomaterials that can reduce the inflammatory effects of infiltrating immune cells and reshape intervertebral disc ECM remains to be further explored.

Cell membrane-coated mimic (CMC) technology has recently been applied in various biomedical fields [8], simplifying the complex processes of surface modification and chemical treatment by combining the sophisticated bionic properties of cell membranes with nanocores that can function as scaffolds, resulting in a new generation of biological material delivery systems. Compared to traditional nano-drug delivery systems, CMC nanoparticles demonstrate excellent biocompatibility, longer circulation retention, and the ability to retain cellular properties through receptor–ligand interactions [9]. Biomimetic nanoparticles inheriting the characteristics of immune cells can facilitate molecular imaging and achieve targeted drug delivery to inflammatory sites, providing a new strategy for the clinical treatment of inflammatory diseases [10]. In particular, biomimetic camouflaged nanoparticles can inhibit excessive inflammatory progression in the local microenvironment, act as immune cell decoys to reduce the recruitment of immune cells, and exert anti-inflammatory effects. However, whether CMC materials can effectively regulate the inflammatory response in a chronic inflammatory environment, such as in IVDD, remains unclear.

Transforming growth factor-beta 1 (TGF-β1) can regulate cell proliferation, promote ECM accumulation, and play a key role in tissue repair and regeneration [11], such as fibrous ring repair [12]. Polylactic-glycolic acid copolymer (PLGA) is a suitable polymer core for CMC materials, including nanodrug delivery systems [13], owing to its superior biocompatibility and degradability [9]^, relatively slow degradation rate, and long-term release in vivo. Thus, PLGA may be an ideal TGF-β1 carrier. Neutrophils are a subgroup of white blood cells with the greatest sensitivity to inflammatory stimuli and are usually the first group of immune cells recruited to the site of inflammation [14]. Neutrophils can specifically migrate to inflammation sites via the expression of surface antibodies, such as integrin-β, macrophage-1 antigen, and leukocyte function-associated antigen 1 [15]. Nanoparticles camouflaged with a neutrophil cell membrane can dynamically regulate homeostasis by actively homing to the inflammatory site [16,17]. Therefore, considering the pathological characteristics of IVDD immune cell infiltration, the local inflammatory response may be alleviated by the application of neutrophil cell membrane-coated CMC materials. However, the intervertebral disc represents a relatively exclusive environment, and obstruction of nutrient transport is an important factor hindering ECM remodeling. Accordingly, a strategy focused only on reducing the local inflammatory effect may help alleviate the rate of degeneration but is unlikely to promote the repair and regeneration of the intervertebral disc.

As intact intervertebral disc tissue has minimal vascular infiltration, systemic drug administration is inefficient because there is insufficient drug residue in the intervertebral disc, resulting in a poor therapeutic effect [18]. Therefore, direct local injection is a common strategy for the treatment of IVDD. However, repeated local injections are usually required to maintain the concentration of the biologics in the intervertebral disc environment and achieve enduring therapeutic effects [19], and these punctures and injections themselves are likely to damage the annulus fibrosus and aggravate IVDD [20]. The intervertebral disc separates the two adjacent vertebrae and evenly distributes pressure to the upper and lower cartilage endplates, which are important load-bearing structures in the spine. Owing to the structural characteristics of this long-term, high-pressure environment, the injected nanoparticles and their loaded drugs easily leak through the fibrous annulus during local injection, affecting the therapeutic effect of biomaterials [21,22]. Therefore, carrier materials are currently used to improve the injection efficiency and stability of therapeutic nanoparticles and achieve long-term therapeutic effects in vivo.

Hydrogels and organic materials are commonly used as biomaterial vectors for the treatment of IVDD [23], owing to their excellent biocompatibility; however, they usually have poor mechanical strength [24] and cannot maintain their integrity in the high-pressure environment of the intervertebral disc. The hydrogel fragments formed after compression lose their mechanical properties, influencing the sustained release of loaded biological agents [25]. Although organic materials usually have good mechanical properties as biological carriers, the majority cannot be completely degraded in vivo, and the decomposition products further affect the environmental metabolism of intervertebral discs [26]. Therefore, there is an urgent need to design biocompatible materials with sufficient mechanical properties for therapeutic nanoparticle injection. We have previously reported that methacrylic acid gelatin anhydride (GelMA) microspheres prepared using microfluidic technology are excellent biological carrier materials [3,27]. Through surface modification, nanoparticles can be grafted onto the porous surface of GelMA microspheres to form a nanoparticle–microsphere complex, which could provide a more suitable material for the intervertebral disc tissue environment.

Based on the imbalance between anabolism and catabolism induced by excessive inflammation in IVDD, which recruits immune cells and activates the inflammatory cascade, we constructed injectable hydrogel microspheres functionalized with neutrophil membranes. TGF-β1-loaded PLGA nanoparticles coated with neutrophil membranes (T-NNPs) and GelMA were combined with amide bonds, and the microsphere–nanoparticle “inflammation camouflage” complex (GM@T-NNPs) was constructed with the nanoparticles evenly distributed on the surface of the microsphere carrier. With direct local injection using microsyringes, GM@T-NNPs exert their immune defense functions, regulate the local intervertebral disc microenvironment, and continuously release T-NNPs as broad-spectrum inflammatory factor inhibitors. As a competitive inhibitor of immune cells, GM@T-NNPs can not only reduce the local excessive inflammation and slow down the local recruitment of immune cells but can also release TGF-β1 for a long period, thereby blocking downstream inflammatory pathways and restoring the balance between anabolism and catabolism in the intervertebral disc microenvironment. Here, we used physical and chemical methods to construct microsphere–nanoparticle complexes and characterize neutrophil membrane coating technology. The immunomodulatory effect and biological mechanism of the complex were evaluated using a co-culture system of nucleus pulposus cells and materials, followed by RNA sequencing to understand the changes in gene expression and associated pathways. A rat IVDD model was established to verify the therapeutic strategy of using GM@T-NNPs to remodel the ECM of the intervertebral disc via inflammation regulation, thereby providing a potential new approach to effectively coordinate inflammation and regeneration in the treatment of IVDD (Scheme 1).

Scheme 1.

Design of immune-defensive microsphere to block inflammatory cascade and promote regeneration of the nucleus pulposus A) and B) Preparation of NMVs, T-NPs and T-NNPs. C) Preparation and grafting of GelMA and T-NNPs nanoparticles. D) In situ injection of GM@T-NNPs into rat IVDD model for local inflammation regulation and ECM regeneration.

2. Results and discussion

2.1. Preparation and characterization of T-NNPs

The preparation of CMC materials is usually divided into three steps: selection of cell membrane materials, selection and preparation of nanoparticle cores, and assembly of the nanoparticle core and cell membrane. Therefore, T-NNPs were prepared by (1) isolating and purifying neutrophil membranes, (2) preparing and characterizing PLGA nanoparticles loaded with TGF-β1 (T-NPs), and (3) assembling and characterizing the nanoparticle core and cell membrane [28].

Rat neutrophils were isolated and purified to obtain the membranes. As neutrophils have a short survival time in vitro, neutrophil extracts from rats should be administered as soon as possible [29]. In this study, neutrophils were isolated from a rat bone marrow cell suspension using density centrifugation (Fig. 1A), and cell purity was determined using flow cytometry. CD11b⁺/HIS48⁺ cells accounted for 78.9% of all extracted cells (Fig. S1) and were used to collect cell membranes via ultrasound, differential centrifugation, and purification, followed by analysis of neutrophil membrane vesicles (NMVs) [30].

Fig. 1.

Preparation and characterization of T-NPs and T-NNPs A) Schematic diagram of neutrophil purification. B) Image of T-NPs and T-NNPs. C) Morphology of NMVs, T-NPs, and T-NNPs under transmission electron microscopy. D) Particle size analysis of T-NPs and T-NNPs. E) Zeta potential analysis of NMVs, T-NPs, and T-NNPs. F) Glycoprotein content analysis of NMVs, T-NPs, and T-NNPs. G) Release curve of TGF-β1 from T-NPs and T-NNPs. H) Characteristic protein bands of neutrophils, NMVs, T-NPs, and T-NNPs resolved by western blotting. I) and J) Binding capacity of T-NPs and T-NNPs with TNF-α and IL-1β.

Nanoprecipitation was used to prepare T-NPs [31], resulting in a milky-white solution (Fig. 1B). The membrane materials and prepared nanoparticles were then assembled. Currently, common methods for assembling CMC materials include co-extrusion, ultrasonic polymerization, microfluidic electroporation, and in situ polymerization [9]. Transmission electron microscopy (TEM) showed that the average size of the NMVs after separation and purification, ultrasound, and co-extrusion was 130.3 ± 35.8 nm, and the nanoparticles of T-NNPs had a vivid “core-shell” structure (Fig. 1C, Fig. S2). T-NNPs were prepared using both co-extrusion and ultrasonic polymerization methods. Compared with the T-NPs, the T-NNP solution appeared transparent and clear. Dynamic light scattering (Fig. 1D) showed that the hydrodynamic diameter of T-NPs before cell membrane coating was 128.2 ± 4.08 nm, and the polydispersity index (PDI) was 0.1117, indicating that T-NPs had good stability in aqueous solution and that the size distribution was concentrated. The hydrodynamic diameter of T-NNPs was 146.7 ± 1.88 nm, and the PDI was 0.0156. The diameter distribution of the T-NNPs was more concentrated than that of the T-NPs, which may have been due to the elimination of some particles with larger sizes through the polycarbonate filter membrane. Compared with that of the T-NPs, the average particle size of the T-NNPs increased by approximately 18 nm, which can be attributed to the thickness of the cell membrane on the surface of the nanoparticle cores. The thickness of the monolayer was approximately 9 nm, which was consistent with previous reports [15,32]. Zeta potential analysis (Fig. 1E) showed that the average surface potential of T-NPs before coating in aqueous solution was −23.2 mV, while that of T-NNPs and the pure NMVs was −8.64 mV and −8.1 mV, respectively. This finding indicated that the surface charge of the coated nanoparticles was closer to that of the NMVs. These results confirm the existence of a cell membrane coating on the surface of the T-NNPs, indicating their successful construction.

Owing to the unilateral distribution of glycoproteins on the surface of cell membranes, quantitative glycoprotein methods are commonly used to detect the laterality of cell membranes during the construction of CMC materials [28]. Enzyme-linked immunosorbent assay (ELISA) showed that the glycoprotein content in T-NNPs was 93%, which was similar to that of NMVs, whereas almost no glycoprotein was detected in T-NPs (Fig. 1F), proving that the cell membrane surface glycoprotein was well preserved during the preparation of T-NNPs. During cell membrane coating, the “right-side-out” property was maintained, indicating that the side with the protein receptor was exposed to the outer surface of the nanoparticles. The glycoprotein content of T-NNPs at different times was measured to reflect the long-term preservation of the “right-side-out” property. The glycoprotein content of T-NNPs hardly fluctuated over time, indicating sustained stability of T-NNPs in vitro (Fig. S3). This property forms the structural basis for the biological efficacy of CMC materials, further verifying that T-NNPs exhibit neutrophil-like characteristics. ELISA confirmed the steady release of TGF-β1 in both T-NPs and T-NNPs. T-NPs maintained the cumulative release of TGF-β1 throughout the observation period (Fig. 1G), reaching 63.8 ± 3.87% on day 18, followed by a decrease, and the sustained release reached 79.1 ± 2.12% on day 36. The release rate of TGF-β1 for T-NNPs was slightly slower than that for T-NPs, with a cumulative release of 60.7 ± 2.62% on day 18 and 73.2 ± 2.95% on day 36, proving that after coating with the cell membrane, the nanoparticles possessed a better drug encapsulation effect.

Retention of membrane proteins is key to the functional preservation of CMC materials. Therefore, we verified the membrane protein content of the constructed T-NNPs using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Fig. S4) and western blotting. The inflammatory chemotactic ability of CMC materials is largely mediated by membrane surface inflammatory cytokine receptors [[32], [33], [34]], including the TNF-α and IL-1β receptors and CD11b molecules. Western blotting (Fig. 1H) of NMVs, T-NPs, and T-NNPs confirmed that these important membrane protein receptors on the surface of neutrophils were well-preserved during the separation and purification of cell membranes and the assembly of cell membranes and cores. To confirm the potential anti-inflammatory effect of T-NNPs, we determined the adsorption kinetics of TNF-α and IL-1β as important pro-inflammatory cytokines in IVDD. The adsorption curve obtained with ELISA showed that the concentrations of TNF-α and IL-1β remaining in the supernatant exhibited a clear downward trend with high concentrations of T-NNPs, indicating a significant dose-response relationship (Fig. 1I and J). The adsorption curve of T-NPs was used as a control group. Compared with T-NNPs, the cytokines in the supernatant remained at the same level. Based on the Hill equation, the neutrophil–nucleus pulposus had a half-maximal inhibitory concentration of 249 μg mL⁻¹ for IL-1β binding and 555 μg mL⁻¹ for TNF-α binding. This finding indicated that the T-NNPs acquired a leukocyte-like function and potential broad-spectrum inflammatory factor-binding and antagonistic effects.

2.2. Preparation and characterization of GelMA microspheres and GM@T-NNPs nanoparticle–microsphere complex

GelMA microspheres are widely used as biomaterial carriers in various applications. Hydrogels are considered appropriate encapsulation carriers for drugs, nanoparticles, and cells because of their excellent biocompatibility, which offers advantageous biomaterial properties for various applications [[35], [36], [37]]. In our previous study [3], GelMA microspheres were found to be ideal carriers for local intervertebral disc injections. Therefore, in this study, the primary amino group on the surface of the cell membrane protein was connected to the carboxyl group obtained from the activated surface of GelMA microspheres, and then T-NNPs were grafted onto the surface of the GelMA microspheres through an amide bond to obtain a GM@T-NNPs nanoparticle–microsphere complex, which could release nanoparticles uniformly and achieve better stability and continuous release in vivo. GelMA microspheres were prepared using previously described microfluidic technology [27]. To prepare GM@T-NNPs, freeze-dried porous hydrogel microspheres and the prepared T-NNPs were covalently linked to 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC)/N-hydroxysuccinimide (NHS).

Under bright-field microscopy, the average particle size of GM@T-NNPs before freeze-drying was 84.41 ± 26.59 μm (Fig. 2A), and particle size analysis of GM@T-NNPs is shown in Fig. 2B. After the freeze-dried GelMA microspheres and GM@TNNPs were sputtered with heavy metal ions, the morphology of the freeze-dried microspheres was observed using scanning electron microscopy (SEM) (Fig. 2C). Local magnification under SEM showed that, unlike the smooth surface of the GelMA microspheres, dense nanoparticles were distributed on the folded surface of the GM@T-NNPs hydrogel microspheres (Fig. 2D), which confirmed the successful preparation of the nanoparticle–microsphere complexes. Uniform pores were observed on the surfaces of the freeze-dried GM@T-NNPs microspheres, with an average pore size of 12.27 ± 3.34 μm (Fig. 2E). The freeze-dried porous structure of microspheres is the morphological basis of their carrier properties for loading nanomaterials and drugs.

Fig. 2.

Preparation and characterization of GM and GM@T-NNPs. A) Morphology of oil-phased GM@T-NNPs under bright-field microscopy. B) Particle sizes of the GM@T-NNP microspheres. C) Morphology of GelMA microspheres under a scanning electron microscope (SEM). D) Morphology of GM@T-NNP microspheres under SEM. E) Pore sizes of the GM@T-NNP microspheres. F) and G) Energy dispersive spectroscopy analysis of GelMA microspheres and GM@T-NNPs. H) X-ray photoelectron spectroscopy analysis of GelMA microspheres and GM@T-NNPs. I) P-element high-resolution energy spectrum scan of GelMA microspheres and GM@T-NNPs. J) Fourier-transform infrared analysis. K) Atomic force microscopy analysis of surface roughness of GM. L) Atomic force microscopy analysis of surface roughness of GM@T-NNPs.

To further verify the successful grafting of nanoparticles onto the surface of the microspheres, the hydrogel microspheres before and after grafting were analyzed using energy dispersive spectroscopy (EDS) and X-ray photoelectron spectroscopy (XPS). The EDS spectrometer can determine and quantitatively analyze all elements between Be and U. Therefore, in this study, we used the point analysis method for GelMA microspheres and GM@T-NNPs for the EDS analysis (Fig. 2F and G). For C, N, and O, owing to the presence of PLGA nanoparticles in GM@T-NNPs, the elemental distribution was slightly different from that of simple GelMA microspheres, showing a significant increase in the proportion of C and O elements. The phospholipid bilayer is the basic structure of the cell membrane in which saccharides and proteins are interspersed. Thus, the element with the highest content in the cell membrane material was P, and the expected results were obtained using EDS analysis. A peak of elemental P was found in the GM@T-NNPs, which accounted for 2.03% of all elements, confirming the existence of cell membrane-coated particles on the surface of the microspheres. XPS analysis showed similar results (Fig. 2H and I). Both GelMA microspheres and GM@T-NNPs exhibited C (284.8 eV), N (400.0 eV), and O (532.0 eV) peaks. Analysis of the high-resolution P-element energy spectrum revealed no P peak in the GelMA microsphere group; however, a characteristic P peak (133.4 eV) was observed in the GM@T-NNP group.

The primary amine group of the lysine residue on the surface of the cell membrane forms the structural basis of T-NNPs grafted onto the surface of GelMA microspheres [[38], [39], [40]]. Covalent binding using the primary amine groups of cell membrane surface proteins can provide a more stable and lasting connection than non-covalent binding methods, such as electrostatic adsorption. The accumulation of lactic acid in a degenerative intervertebral disc environment reduces the pH of the local microenvironment, which can induce the hydrolysis of amide bonds [3] and initiate the long-term release of nanoparticles. Fourier-transform infrared spectroscopy was performed to verify the existence of amide bonds in the nanoparticle–microsphere complexes (Fig. 2J). Compared with that of the GelMA group, the peak value of GM@T-NNPs decreased significantly at 3290 cm⁻¹, indicating an obvious absorption peak of the –OH group. At 1750 cm⁻¹, there was a C O peak in the GM@T-NNPs, indicating the presence of PLGA; however, there was no C O peak in the GelMA group. A new weak absorption peak at 1640 cm⁻¹ was observed in the T-NNP group compared with that in the T-NP group. A possible explanation is that owing to the presence of the primary aliphatic amine group retained in the lysine residue of the surface membrane protein after coating, the bending vibration of the -NH2 group could be detected in T-NNPs. Atomic force microscopy of the nanoparticle–microsphere composite (Fig. 2K and L) showed that the surface roughness of the microspheres increased significantly after grafting.

These results showed that T-NNPs can covalently connect with the carboxyl groups on the surface of the GelMA microspheres through the primary amine group on its surface via the EDC/NHS system, which verified the successful construction of the nanoparticle–microsphere complex. To investigate the mechanical properties of GM@T-NNPs, AFM indentation measurements were performed, and force-separation curves were recorded to reflect the small deformation of the blocky GelMA hydrogel, GM, and GM@T-NNPs. The Young's moduli of both GM and GM@T-NNPs improved significantly compared with the blocky GelMA hydrogel, indicating better mechanical properties of the microspheres over the blocky hydrogel (Fig. S5).

To trace the content and distribution of neutrophil membrane components after coating and chemical grafting in vivo, we prepared DiR-labeled neutrophil membranes. T-NNPs-DiR and GM@T-NNPs-DiR were obtained. Both T-NNPs-DiR and GM@T-NNPs-DiR were assessed for fluorescence intensity using an optical imaging system (IVIS Lumina LT, PerkinElmer, USA) (Fig. S6A). After Sprague–Dawley (SD) rats were punctured and injected with T-NNPs-DiR and GM@T-NNPs-DiR using a microsyringe, in vivo fluorescence images were taken on days 0, 1, 3, 7, and 14 (Fig. S6B). Quantitative analysis of radiant efficiency revealed that the fluorescence signal of GM@T-NNPs-DiR could still be detected 14 days after injection, indicating that the neutrophil membrane components were stable in vivo (Fig. S6C). However, in the T-NNP-DiR group, the rate of cancellation was faster than that in the GM@T-NNP-DiR group, which may be attributed to the more frequent leakage of the nanoparticle solution.

2.3. In vitro bioactivity and biocompatibility of GM@T-NNPs

Chemical grafting may compromise the activity of protein receptors on cell membranes. To investigate the bioactivity of various protein receptors on the surface of T-NNPs after grafting through amide bonds, the targeting ability of the nanoparticles was verified using immunofluorescence. According to a previous report, the expression of pro-inflammatory cytokines increases in nucleus pulposus cells during IVDD, and neutrophils participate in receptor-mediated migration and adhesion to nucleus pulposus cells [41]. Compared with neutrophil membranes, owing to the lack of inflammation-related surface receptors, erythrocyte membranes have poor plasticity for treating inflammation-related diseases and poor directed migration ability for inflammation [32]. Therefore, we used a Transwell culture system to explore the release of T-NNPs and the migration of nanoparticles to the inflammation site (Fig. 3A). We employed the frequently used cell membrane probe 1,1-dioctadecyl-3,3,3,3-tetramethylindotricarbocyaine iodide (DiR) to label T-NNPs and Phalloidin to label the F-actin cytoskeleton. Immunofluorescence results (Fig. 3B) showed that a few nanoparticles adhered to the surface of nucleus cells after inflammatory stimulation with lipopolysaccharide (LPS) in cells cultured with T-NNPs and GM@T-NNPs, whereas adhesion and phagocytosis of nanoparticles were observed on the surface of nucleus pulposus cells after inflammatory stimulation. In the PLGA nanoparticles coated with erythrocyte membranes (RBC-NP) and T-NP groups, almost no DiR intensities were detected owing to a lack of inflammation-related surface receptors. These findings demonstrated that chemically grafted T-NNPs had the chemotactic ability to respond to inflammation, thus verifying the retention of inflammatory receptor protein activity on the cell membrane surface. Semi-quantitative analysis of immunofluorescence staining images showed that the T-NNP group had the highest fluorescence intensity at 6 h. However, at 72 h, the fluorescence intensity in the GM@T-NNP group increased significantly, indicating that the long-term stable release of T-NNPs was maintained (Fig. 3C). The covalent connection between the nanoparticles and microspheres enabled the long-term release of T-NNPs in vitro and prevented their sudden release, allowing them to function as broad-spectrum anti-inflammatory agents.

Fig. 3.

Function and biocompatibility verification of GM@T-NNPs. A) Schematic diagram of the transwell assay to measure the targeting ability of the nanoparticles. B) Fluorescence images of nucleus pulposus (NP) Cell uptake of fluorescently tagged nanoparticles in the T-NP group, RBC-NP group, T-NNP group and GM@T-NNP group for 6 and 72 h. C) Semi-quantitative analysis of the uptake of fluorescently tagged nanoparticles by NP cells co-cultured with T-NPs, RBC-NPs, T-NNPs, and GM@T-NNPs D) and E) TEM images of NP cells after co-culture with RBC-NPs and T-NNPs. F) and G) Ability of T-NPs, GM, RBC-NPs, NMVs, T-NNPs and GM@T-NNPs to inhibit NP cell apoptosis induced by LPS, measured by immunofluorescence of BAX and cleaved Caspase-3. (n = 3). H) Semi-quantitative analysis of BAX and cCaspase-3. I) CCK-8 assay (n = 3). J) Flow cytometry test of Ki-67. K) Quantitative analysis of Ki-67⁺ proportion. *p < 0.05; **p < 0.01; ***p < 0.005; #p < 0.001; ns, not significant (one-way analysis of variance with Tukey's test).

To further verify the interaction between nanoparticles and nucleus pulposus cells in vitro, nucleus pulposus cells co-cultured with RBC-NPs and T-NNPs for 24 h were cryopreserved and observed using TEM. The results showed that endocytosis occurred in both groups, but the number of endocytosed nanoparticles in the RBC-NP group was less than that in the T-NNP group (Fig. 3D and E). This observation may be explained by the fact that RBC-NPs rarely exhibit chemotaxis to the surface of nucleus pulposus cells after inflammatory stimulation due to the lack of inflammatory receptors on the surface, resulting in few nanoparticles phagocytizing the cells detected by TEM.

To further test the retention of the membrane protein activity of GM@T-NNPs, we evaluated the effect of various groups of nanoparticles on the inhibition of cytokine-induced apoptosis of nucleus pulposus cells. LPS was used to induce apoptosis in nucleus pulposus cells cultured in vitro [[42], [43], [44]]. The degree of apoptosis was evaluated by the expression of two important proteins: Bcl-2-associated X protein (BAX) and cysteinyl aspartate-specific proteinase 3 (caspase-3). BAX is a perforating protein belonging to the Bcl-2 family, and when the mitochondrial outer membrane is perforated or permeated by BAX or Bcl-2 homologous antagonist/killer (BAK) proteins, the cells begin apoptosis [45]. Caspase-3 is the most important protein in apoptosis. Cleaved caspase-3 (cCaspase-3) cleaves fibronectin, actin, and other cytoskeletal proteins, causing significant morphological changes in the cells [46]. Nucleus pulposus cells stimulated with LPS (200 ng mL⁻¹) were treated with T-NPs, GM, RBC-NPs, NMVs, T-NNPs, and GM@T-NNPs in the Transwell system for 24 h. Immunofluorescence staining showed no significant difference in the expression of BAX and cCaspase-3 proteins among the NC, T-NP, GM, and RBC-NP groups, indicating that T-NPs, GM, and RBC-NPs had almost no effect on reversing apoptosis (Fig. 3F and G). However, the fluorescence intensity of BAX and cCaspase-3 significantly declined in the NMV, T-NNP, and GM@T-NNP groups (Fig. 3H). These results suggest that neutrophil membranes can absorb significantly more of the overexpressed pro-inflammatory cytokines, thereby downregulating the excessive inflammatory levels in nucleus pulposus cells, reducing the expression of BAX and cCaspase-3, suppressing their shearing effect on the mitochondrial membrane, and ultimately inhibiting apoptosis. The effect of GM@T-NNPs was similar to that of T-NNPs but with even better inflammation control, which also reflected that the activity of surface protein receptors of T-NNPs was retained after chemical grafting and that T-NNPs were released stably and evenly after grafting through amide bonds using GelMA as a biological carrier.

Live/dead staining and Cell Counting Kit-8 (CCK-8) cell viability tests were performed to determine the biocompatibility of the GM@T-NNPs nanoparticle–microsphere complex. During IVDD, the number of nucleus pulposus cells usually decreases due to chronic inflammation; therefore, appropriate biological carrier materials need to have superior biocompatibility and low biological toxicity, serving as a scaffold for nucleus pulposus cells to climb and grow. Nucleus pulposus cells from SD rats directly co-cultured with GM@T-NNPs adhered to the microspheres, and most of the cells grew from the edge of the microspheres to the center (Fig. S7). On day 5 of co-culture, nucleus pulposus cells were uniformly distributed on the surface of the microspheres, and few dead cells were observed. These results showed that GM@T-NNPs exhibited good biocompatibility and no obvious cytotoxicity. The CCK-8 assay showed that nucleus pulposus cells intervened with LPS and treated with the NMVs, T-NNPs and GM@T-NNPs showed faster cell proliferation on the third and fifth days of culture than the NC group (Fig. 3I). This result may be attributed to the anti-apoptosis ability of the neutrophil membranes in all three groups. Moreover, T-NNPs and GM@T-NNPs revealed superior cell proliferation ability compared with the NMVs after 5 days of culture because they both maintained stable release of TGF-β1 which promotes cell differentiation, and matrix accumulation in vitro [12,47]. However, there was no significant proliferative effect in the T-NP, GM and RBC-NP groups. Cell growth (based on the optical density at 450 nm) of the GM@T-NNP group was 1.31 and 1.41 times higher than that of the NC group on the third and fifth day of culture, respectively. Flow cytometry of Ki-67 protein was also conducted to detect the proliferation state of nucleus pulposus cells co-cultured with biomaterials and with intervention of LPS (Fig. 3J). Ki-67 positive proportion analysis (Fig. 3K) revealed that nucleus pulposus cells in the NC, T-NP, GM and RBC-NP group possessed almost the same proportion of Ki-67⁺ cells, which is consistent with the CCK-8 results, whereas in the NMV, T-NNP, and GM@T-NNP groups, the number of Ki-67⁺ cells was higher, indicating a better proliferation state because of the existence of the neutrophil membranes.

2.4. Effect of GM@T-NNPs in the control of inflammation

The progression of IVDD is characterized by increased secretion of pro-inflammatory cytokines, a decreased number of nucleus pulposus cells, and acceleration of ECM degeneration. Increased levels of IL-1β, TNF-α, and other inflammatory mediators contribute greatly to the degenerative process [[48], [49], [50]], followed by the chemical attraction of neutrophils in the intervertebral disc environment [41]. Therefore, the application of biomaterials to downregulate the expression of various inflammatory mediators, block immune cell infiltration, and restore the balance between ECM synthesis and catabolism is an important strategy for the treatment of IVDD.

To evaluate the therapeutic effect of GM@T-NNPs as a broad-spectrum anti-inflammatory agent in regulating the local inflammatory microenvironment, immunofluorescence staining was performed to determine the protein expression of TNF-α and IL-1β in each group after direct co-culture, as shown in Fig. 4A. The protein expression of both inflammatory factors in nucleus pulposus cells was very limited without a conditioned culture medium. However, after stimulation with LPS (200 ng mL⁻¹), abundant pro-inflammatory factors were secreted, accompanied by morphological changes in cells (Fig. 4B and C). The spreading area of the nucleus pulposus cells in the inflammatory environment was reduced. Semi-quantitative fluorescence analysis revealed that normalized TNF-α and IL-1β fluorescence intensities in the NC, T-NP, GM, and RBC-NP groups showed no statistical difference. The normalized TNF-α fluorescence intensity of the NMV, T-NNP, and GM@T-NNP groups decreased by 62.6%, 64.2%, and 78.9%, respectively, and the normalized IL-1β fluorescence intensity decreased by 31.6%, 30.8%, and 64.6%, respectively (Fig. 4D and E), compared to that in the NC group. This suggests that GM@T-NNPs mediated the negative regulation of both pro-inflammatory cytokines, confirming the high adsorption efficiency of inflammatory factors and the downregulation of the excessive inflammatory response.

Fig. 4.

Inflammation control effect of GM@T-NNPs A) Schematic diagram of nucleus pulposus cells (NP) co-cultured with microspheres after lipopolysaccharide (LPS) stimulation for 7 days. B) and C) Immunofluorescence of TNF-α and IL-1β. D) and E) Semi-quantitative analysis of the immunofluorescent staining of TNF-α and IL-1β.

COL-II is an important component of the ECM and can be used as a marker to indicate the decomposition of the ECM and the degree of degeneration of the nucleus pulposus. Immunofluorescence staining showed that after GM@T-NNPs effectively inhibited excessive local inflammation, the protein expression of COL-II was significantly upregulated compared with that of other groups, and its normalized fluorescence intensity was 7.50 times that of the NC group (Fig. 5A and B). Comparing the normalized COL-II fluorescence intensity of the NMV group with the T-NNP and GM@T-NNP groups, although the expression of COL-II protein increased in the NMV group due to the remission of local inflammation, as shown in Fig. 4, the T-NNP and GM@T-NNP groups still had a superior function in maintaining the secretion of COL-II protein because of the appropriate release of TGF-β1. Based on the results of immunofluorescence experiments, we preliminarily verified the anti-apoptotic and anti-inflammatory functions in each group. The results in Fig. 4, Fig. 5 demonstrate that the therapeutic effect of RBC-NPs was similar to that of the NC group, revealing limited inflammation regulation ability due to the lack of inflammatory surface receptors. NMVs demonstrated regulation of apoptosis and inflammation, confirming the function of neutrophil membranes. The poor performance of NMVs in inducing cell proliferation and regeneration, as shown in Fig. 3I, K, and 5A, limits their further applications. Based on the abovementioned reasons, we decided to curtail the experimental groups in subsequent experiments by ruling out the RBC-NP and NMV groups.

Fig. 5.

Protein expression analysis of nucleus pulposus (NP) cells cultured in an inflammatory microenvironment in vitro for 7 days. A) Immunofluorescence of COL-Ⅱ. B) Semi-quantitative analysis of the immunofluorescent staining of COL-Ⅱ. C) Western blotting of TNF-α, pro-IL-1β, IL-6, MMP-13, COL-II, and aggrecan in NP cells after co-culture with different nanoparticles. D–I) Semi-quantitative analysis of TNF-α, pro-IL-1β, IL-6, MMP-13, COL-II, and aggrecan protein expression (n = 3); the results were compared with the control group by one-way analysis of variance and Tukey's post-hoc test: *p < 0.05; **p < 0.01; ***p < 0.001.

Western blotting was used to verify the anti-apoptotic and anti-inflammatory effects of T-NPs, GM, T-NNPs, and GM@T-NNPs. To investigate the expression of BAX and cleaved Caspase-3 to confirm the anti-apoptotic effect of the biomaterials, western blotting was conducted in each group. Quantitative analysis of the western blot bands (Fig. S8) revealed that both cleaved Caspase-3 and BAX expression in the T-NP and GM groups was similar to that in the NC group. T-NNPs and GM@T-NNPs downregulated the expression of these two proteins and inhibited LPS-induced apoptosis. Anti-inflammatory experiments showed that the expression of pro-inflammatory cytokines (IL-1β, TNF-α, and IL-6) induced by LPS decreased significantly after treatment with GM@T-NNPs (Fig. 5C). Compared to the NC group, the average gray values of the IL-1β, TNF-α, and IL-6 bands decreased by 50%, 72%, and 70%, respectively (Fig. 5D–I). The absorption of pro-inflammatory cytokines can downregulate the expression of MMP-13, which effectively reduces ECM degradation in nucleus pulposus cells. The main components of the ECM in nucleus pulposus cells are COL-II and proteoglycan (aggrecan), which maintain the hydrophilic and mechanical structure of the nucleus pulposus. The expression levels of COL-II and aggrecan in GM@T-NNPs were 2.36 and 2.27 times higher than those in the control group, respectively, which further proved that GM@T-NNPs effectively restrained the chronic inflammation of the nucleus pulposus, improved the expression of protease, and alleviated degradation of the ECM.

The expression levels of downstream genes were further verified using real-time quantitative polymerase chain reaction (RT-qPCR). The mRNA expression levels of inflammatory factors, such as IL-1β, TNF-α, and IL-6, were significantly downregulated after GM@T-NNP treatment (Fig. 6). The mRNA expression of the MMP-13 gene was downregulated, and the expression levels of COL-II and aggrecan genes in the GM@T-NNP group were 4.86 and 2.43 times higher than those in the NC group, respectively, further indicating that GM@T-NNPs improved the inflammation-induced disorder of anabolism and catabolism. Overall, these results demonstrate that GM@T-NNPs can regulate the level of inflammation, inhibit the activation of nucleus pulposus cells, and upregulate the expression of ECM proteins, revealing their potential to repair IVDD. Therefore, we next explored the molecular mechanism by which GM@T-NNP nanoparticle–microsphere complexes promote the repair of nucleus pulposus cells.

Fig. 6.

mRNA expression analysis A–F) Relative mRNA expression levels of IL-1β, TNF-α, IL-6, MMP-13, COL-Ⅱ, and aggrecan genes (n = 3). All data are presented as mean ± standard deviation: *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant (one-way analysis of variance and post-hoc Tukey test compared with the control group).

2.5. Potential signaling pathways underlying the immune-defensive effects of microspheres

To explore the molecular mechanism by which GM@T-NNPs alleviate IVDD, we co-cultured nucleus pulposus cells with GM@T-NNPs and GM for seven days of LPS stimulation and extracted the total RNA of cells for transcriptome sequencing. Compared with the GM group, the expression of 1944 genes in the GM@TNNP nanoparticle–microsphere complex group was upregulated, whereas the expression of 1610 genes was downregulated (Fig. 7A). Heatmap analysis of the differentially expressed genes (DEGs) showed a significant difference between the two groups (Fig. 7B), which proved that the biomaterials had an important effect on gene expression in nucleus pulposus cells. Principal component analysis showed that the two groups of samples met the requirements for further analysis of sequencing data (Fig. S9), demonstrating obvious differences between the two groups. Gene Ontology enrichment analysis of the DEGs showed significant downregulation in cell adhesion and the immune response and upregulation in ECM structural constituents (Fig. S10), indicating that after T-NNPs treatment, nucleus pulposus cells showed a reduced response to immune cells and expressed more proteins associated with the ECM structure. These results confirmed that the GM@T-NNP biomaterials inhibited the chemotactic recruitment of immune cells by nucleus pulposus cells to some extent and alleviated the degeneration process of the ECM, which was consistent with the expected results.

Fig. 7.

Transcriptome sequencing analysis of the mechanism by which composite microspheres down-regulate the inflammatory microenvironment of nucleus pulposus (NP) cells. A) Volcano map of differentially expressed genes (DEGs) in the GM@T-NNP group compared to the GM group. (B) Heatmap of DEGs. (C) KEGG enrichment analysis of DEGs. D) and E) GSEA analysis of TNF signaling and chemokine receptor-binding chemokine pathways.

Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis showed significant inhibition of signaling pathways related to inflammatory factors, such as the TNF signaling pathway, IL-17 signaling pathway, and cytokine–cytokine receptor interaction, in the GM@T-NNP-treated group (Fig. 7C). Gene set enrichment analysis showed that the expression levels of genes related to the TNF signaling pathway, chemokine binding pathway, and inflammatory migration decreased (Fig. 7D and E), revealing inflammatory inhibition and potential downregulation of chemotaxis after treatment with nanoparticle–microsphere complexes. In addition, the gene enrichment heat map showed that phosphatidylinositol 3-kinase (PI3K) and RAC-alpha serine/threonine-protein kinase (AKT) related pathways, which are associated with inflammatory cascade amplification, were significantly inhibited by treatment with GM@T-NNPs (Fig. 8A and B). To further clarify the relationship between inflammatory stimulation and this pathway, we performed western blotting analysis (Fig. 8C) to evaluate the expression of AKT and phosphorylated AKT (p-AKT) proteins in nucleus pulposus cells after each intervention. The expression of p-AKT was significantly upregulated under LPS stimulation and was alleviated after treatment with T-NNPs and GM@T-NNPs (Fig. 8D), indicating that the therapeutic effect of biomaterials may be related to the inhibition of the PI3K-AKT signaling pathway.

Fig. 8.

Transcriptome sequencing analysis of PI3K/AKT pathway. A) GSEA of the PI3K-AKT pathway. B) 1-Phosphatidylinositol-3-kinase regulator activity heatmap of the analyzed genes. C) Western blot analysis of the PI3K-AKT pathway. D) Semi-quantitative protein analysis of the PI3K-AKT pathway. E) Relative mRNA expression levels of PI3K and AKT genes.

The PI3K-AKT pathway is located upstream of the NF-κB pathway and has been previously reported to be associated with the inflammatory activation of nucleus pulposus cells [51,52]. In addition, the interaction between p-AKT and phosphatidylinositol-3-phosphate (PIP3) on the cell membrane stimulates the phosphorylation of downstream targets, thus regulating a variety of inflammatory and immune functions, including macrophage and neutrophil recruitment and T-cell activation. Therefore, inhibition of the PI3K pathway is considered a feasible therapeutic strategy for inflammatory diseases [53]. RT-qPCR analysis showed that the expression of AKT and PI3K genes in the GM@T-NNP treatment group was downregulated compared with that in the NC group (Fig. 8E). Collectively, these results suggest that the molecular mechanism by which GM@T-NNPs reduce the degeneration of nuclear pulposus cells may involve inhibition of the PI3K-AKT pathway, local inflammatory reactions, and further infiltration of immune cells in the nucleus pulposus tissue.

2.6. Therapeutic effect of the nanoparticles–microsphere complex in vivo

The ultimate goal of this study was to develop biomaterials to inhibit excessive inflammation, block immune cell recruitment to intervertebral discs, and reconstruct the ECM. Therefore, we established a rat model of caudal vertebral IVDD [3] to verify the therapeutic effects of T-NNPs and GM@T-NNPs in vivo that were observed in vitro. The caudal vertebrae of the rats were punctured with a 21 G needle tip and then injected with phosphate-buffered saline (PBS), T-NPs, GM, T-NNPs, or GM@T-NNPs. Significant differences in intervertebral disc height, according to the disc height index (DHI), were observed between the control and PBS groups at 4 and 8 weeks after injection, and the DHI of each group slightly decreased at week 8 compared to that at week 4 (Fig. 9B). Analysis of the histological grade using magnetic resonance imaging (MRI) led to a similar conclusion, which proved that rat IVDD models obtained by fine-needle puncture were successfully established. The DHI values in the T-NP injection group and the simple GM microsphere group were similar to those in the PBS group, with no significant difference at either 4 or 8 weeks, proving that T-NPs and GM have minimal therapeutic effects in vivo. Although there was a significant difference in the DHI between the GM@T-NNP and PBS groups, the intervertebral disc height was significantly restored at 8 weeks, and the difference was more obvious at 8 weeks than at 4 weeks. No significant difference was observed between the therapeutic effects in the T-NNP and GM@T-NNP groups at 4 weeks. However, at 8 weeks, the therapeutic effect in the GM@T-NNP group was better than that in the simple T-NNP group. This may indicate that fewer nanoparticles were retained at the injection site following T-NNP injection. In contrast, injected nanoparticles tend to move with changes in the body position because of the fluidity of the solution. Leakage may occur from the injection channel, leading to a failure to exert a long-term therapeutic effect in vivo. However, after the chemical grafting of microspheres, the nanoparticles exhibited good stability in vivo and showed a stable release effect [7].

Fig. 9.

Therapeutic effect of the nanoparticles–microsphere complex in vivo. A) Overview of the rat IVDD model. B) X-ray images. C) MRI T2 scans. D) H&E staining at 4 and 8 weeks. E) Safranin O-Fast Green staining at 4 and 8 weeks. F) COL-II immunohistochemical staining. G) and H) Changes in disc height index (DHI) at 4 and 8 weeks. I) MRI grading 4 and 8 weeks after surgery. J) Histological grading at 4 weeks and 8 weeks. K) Semi-quantitative analysis of COL-II (n = 3); all data are presented as mean ± standard deviation: *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant (one-way analysis of variance and Tukey's post-hoc test).

MRI T2-weighted imaging (T2WI) is a simple and intuitive visual evaluation method for various diseases and has been widely used in the diagnosis and grading of IVDD. Healthy intervertebral disc tissue showed a high signal on T2 images, indicating a higher level of hydration in the nucleus pulposus tissue. During degeneration, with a decrease in collagen and aggrecan, the number of nucleus pulposus cells decreases, and the hydration of the nucleus pulposus tissue decreases gradually. The Thomson classification is commonly used to evaluate and grade degenerative intervertebral discs [54]. Compared with the control group, there was a significant loss of the T2WI signal in the PBS group at both 4 and 8 weeks (Fig. 9C), indicating irreversible damage after disc puncture. The MRI grades of the T-NP and GM groups were similar to those of the PBS group. However, the MRI scores of the T-NNP and GM@TNNP groups were significantly decreased. Injection of GM@T-NNPs resulted in a significant improvement in the MRI signal, confirming the repair of the nucleus pulposus tissue.

Hematoxylin and eosin (H&E) and Safranin O/Fast green staining were performed on tissue samples for histological analysis, demonstrating a certain degree of atrophy of the nucleus pulposus and annulus fibrosus in all groups at 8 weeks compared with that at 4 weeks (Fig. 9D), which was consistent with the imaging findings. This trend indicated that the IVDD process continued over time. The staining results further showed that the boundary between the nucleus pulposus and annulus fibrosus was vague in the T-NP, GM, and PBS groups (Fig. 9E), indicating that after intervertebral disc puncture, the broken fibrous tissue of the nucleus pulposus proliferated over time, which interfered with its function and structure of the nucleus pulposus. Therefore, the therapeutic effects of GM or T-NPs are limited. At 4 weeks, there was a certain degree of nucleus pulposus tissue regeneration in the T-NNP and GM@T-NNP groups, and the boundary of fibrotic tissue gradually appeared to be slightly restored. In the T-NNPs group, the tissue boundary disappeared again at 8 weeks, indicating that the material failed to maintain repair function. In contrast, a significant therapeutic effect was maintained in the GM@T-NNP group at 8 weeks. Owing to the high pressure of the intervertebral disc tissue, the injected nanoparticles had fewer opportunities to reach the targeted region, thus failing to attain long-term treatment and further validating the superior therapeutic effect of GM@T-NNPs in vivo. A similar conclusion was obtained from immunohistochemical analysis of COL-II (Fig. 9F), which is generally considered an important component of the ECM. Quantitative analysis of the percentage of COL-II positive areas showed that GM@T-NNPs had a better inhibitory effect on ECM degradation than T-NNPs at 4 weeks, and the COL-II positive area was approximately 1.6 times higher than that of the T-NNP group at 4 weeks, with an even more significant difference detected at 8 weeks (p < 0.001).

To compare the retention efficiencies of T-NNPs and GM@T-NNPs, Cy5-labeled T-NNPs and GM@T-NNPs were injected into rat caudal vertebral discs and observed using an in vivo imaging system (Fig. 10A). On day 3, the radiant efficiency signal of the T-NNP group significantly decreased compared with that on day 0, and the fluorescence signal was hardly observed on day 14. In the GM@T-NNP group, a faint radiant efficiency signal could still be detected 14 days after injection, indicating a better retention efficiency of GM@T-NNPs in vivo.

Fig. 10.

Inflammation evaluation of the IVDD model. A) Representative fluorescence image after injection of a Cy5-labeled free T-NNPs solution and Cy5-labeled GM@T-NNPs into the disc sites over time. B) and C) Semi-quantitative analysis of MPO and TNF-α immunofluorescence. D) Representative image of TNF-α immunofluorescence. B) Representative image of MPO immunofluorescence.

Finally, to verify the inflammatory control effect in the local microenvironment of the intervertebral disc, we performed histological immunofluorescence staining to observe the protein expression of TNF-α and IL-1β in the nucleus pulposus (Fig. 10D and Fig. S11). At 4 and 8 weeks, the fluorescence intensity of IL-1β in the GM@T-NNP group decreased by 74.4% and 64.2%, respectively, and the intensity of TNF-α decreased by 70.9% and 71.4%, respectively, compared to those in the PBS group, revealing a strong therapeutic effect in vivo. In addition, GM@T-NNPs significantly improved the immune microenvironment owing to excessive inflammation of the nucleus pulposus in vivo. By downregulating the expression of pro-inflammatory cytokines, ECM degradation was alleviated, and the nanoparticle–microsphere complex showed excellent stability and long-term release performance in vivo.

To further explain the relationship between IVDD and immune cell infiltration, we used myeloperoxidase (MPO) staining to locate the main distribution of immune cells in the nucleus pulposus. As shown in Fig. 10E, in the control group, there was almost no expression of MPO in the nucleus pulposus tissue at 4 and 8 weeks, which verified the relative immune privilege of the nucleus pulposus in the physiological state. Abundant MPO expression was detected in intervertebral disc tissue after puncture, indicating the intrusion of immune cells. However, MPO expression was significantly downregulated after treatment with GM@T-NNPs, with a decrease in fluorescence intensity of 81.6% and 78.9% at 4 and 8 weeks, respectively, compared with that in the PBS group. In addition, a decline in MPO fluorescence intensity was observed in each group at 8 weeks compared with that at 4 weeks. To further explore the immune-regulated function of GM@T-NNPs, tissue immunofluorescence of HIS48 and F4/80 was performed to clarify the immune infiltration of neutrophils and macrophages. Almost no fluorescence signals of HIS48 and F4/80 were detected in the control group at 4 and 8 weeks (Figs. S12A and S12B), implying that the mechanical barrier of the intact intervertebral disc prevents immune infiltration. Following barrier disruption and ingrowth of blood vessels, the populations of both neutrophils (HIS48⁺) and macrophages (F4/80⁺) increased in the PBS group. While the T-NP and GM groups revealed no significant difference from the PBS group, T-NNPs and GM@T-NNPs restrained the migration of both neutrophils and macrophages. Semi-quantitative analysis (Fig. S12C) further confirmed the therapeutic effect of GM@T-NNPs, downregulating the fluorescence intensity of HIS48 by 70.2% and 78.1% at 4 and 8 weeks, respectively, compared with the PBS group. Meanwhile, the expression of F4/80 GM@T-NNPs decreased by 58.9% and 67.4% at 4 and 8 weeks, respectively, compared to the PBS group. Notably, based on the in vivo immunofluorescence results, the population of HIS48⁺ neutrophils decreased at 8 weeks, while the population of F4/80⁺ macrophages remained the same over time. This phenomenon suggests that neutrophils tend to infiltrate the intervertebral disc at a relatively early stage, while the infiltration of macrophages is a long-term process.

Western blotting was performed to clarify changes in the PI3K-AKT pathway in vivo (Fig. S13). After puncture, the phosphorylation levels of PI3K and AKT increased. Similar to the in vitro experiments, T-NPs and GM had a limited inhibitory effect on the phosphorylation process. However, in the T-NNP and GM@T-NNP groups, there was no significant inhibition, indicating the ability of neutrophil membranes to downregulate the phosphorylation levels of PI3K and AKT proteins in vivo.

3. Conclusion

We developed novel functionalized microspheres based on the pathology of IVDD, in which neutrophil-coated PLGA nanoparticles loaded with TGF-β1 were modified to graft with GelMA microspheres as a broad-spectrum inflammatory factor antagonist, enabling targeting and maintenance in the immune environment of intervertebral discs while simultaneously improving the release kinetics of TGF-β1. After being chemically grafted onto the surface of microspheres, T-NNPs retained their biological activity, reversed the inflammatory microenvironment induced by LPS in vitro, absorbed pro-inflammatory cytokines (such as IL-1β and TNF-α), and achieved the long-term release of TGF-β1, thus maintaining the accumulation of ECM in the nucleus pulposus. The therapeutic effect was further confirmed by RNA sequencing, which demonstrated that treatment with GM@T-NNPs may be related to inhibition of the PI3K-AKT pathway. In vivo experiments using a rat IVDD model showed that the nanoparticle–microsphere complex achieved even better inflammation control than that found in vitro, maintained the height of the intervertebral disc, and restored the biological structure of the nucleus pulposus. We further explored the relationship between the biological function of GM@T-NNPs and innate immune cell infiltration, revealing that GM@T-NNPs could function as immunological bait to bind with nucleus pulposus cells and compete with neutrophils and macrophages, thereby reducing the infiltration of immune cells. Inspired by this study, more designs should be explored to enable CMC nanoparticles to fit the pathology of chronic inflammatory diseases and balance local excessive inflammation. In summary, this study expands the application of CMC materials using microspheres as carriers to achieve the effective application of membrane-coated nanoparticles through chemical grafting for the treatment of IVDD and to provide an individualized and precise solution for the treatment of long-term chronic inflammation of the disc environment.

4. Experimental section

4.1. Materials

PLGA (MV 8000, 50:50) was purchased from Jinan Daigang Bioengineering Co. Ltd (Jinan, China). TGF-β1, IL-1β, TNF-α, 1,1-dioctadecyl-3,3,3,3-tetramethylindotricarbocyaine iodide (DiR), and LPS were purchased from AbMole Biology Co., Ltd (Delaware, US). The rat bone marrow neutrophil extraction kit was purchased from Beijing Solarbio Technology Co., Ltd (Beijing, China). and the protease/phosphatase inhibitor mixture was purchased from Beyotime Biotechnology Co., Ltd (Shanghai, China). Rat TGF-β1, IL-1β, and TNF-α ELISA kits were purchased from Signalway Antibody Company (Maryland, US), and the rat GP130 glycoprotein ELISA kit was purchased from Wuhan JIYINMEI Biotechnology Co., Ltd (Wuhan, China). Dulbecco's modified Eagle medium (DMEM), fetal bovine serum (FBS), trypsin, and penicillin/streptomycin were purchased from Gibco (Carlsbad, CA, USA). Cy5 was purchased from TargetMol Biotechnology Co., Ltd (Massachusetts, US). All other chemical reagents were purchased from Aladdin Reagent Co., Ltd (Beijing, China).

4.2. Extraction of the neutrophil membrane

Six hours before sampling, the rats were intraperitoneally injected with LPS (1.5 mg kg⁻¹) to stimulate the neutrophils. After the animals were sacrificed with excess pentobarbital (100 mg kg⁻¹), the femurs and tibiae were collected, and bone marrow cell suspensions were prepared and counted. Cells were purified using the Solarbio Rat Bone Marrow Neutrophil Isolation Kit. Purified cells were immediately labeled with CD11b (Abcam) and HIS48 (Abcam). After washing with PBS, the cells were analyzed by flow cytometry (Merck Millipore, Germany). After the purification procedure, the cells were washed with PBS for three times and then centrifuged at 800×g. The cells were then resuspended in hypotonic cell fluid consisting of 30 mM Tris-HCl (pH 7.5), 225 mM d-mannitol, 75 mM sucrose, 0.2 mM EGTA, and 4% protease phosphatase. The cell suspension was homogenized on ice for 8 min. The homogenized solution was centrifuged at 4 °C for 20 min, the precipitate was discarded, and the supernatant was centrifuged at 100,000×g and 4 °C for 35 min. After centrifugation, the neutrophil membrane material was obtained and the precipitate was collected and washed twice with 0.2 mM EDTA aqueous solution. A BCA quantitative kit (Solarbio) was applied to quantify the membrane protein content.

To obtain NMVs, the extracted neutrophil membrane was first treated with ultrasound for 15 min and then extruded through an Avestin micro-extruder through 450 nm and 220 nm polycarbonate filters approximately 10 times. The prepared cell membrane vesicles were stored at −80 °C for subsequent use. For long-term restoration, membrane materials were preserved by freeze-drying.

Erythrocyte membranes were extracted as non-nucleated cells. The washed RBCs were resuspended in 0.25 × PBS on ice for 20 min, centrifuged at 800×g for 5 min, and the supernatant was discarded and resuspended in 0.25 × PBS on ice for 20 min. This process was repeated until the hemoglobin was completely removed. The erythrocyte membranes were resuscitated and quantified at a protein concentration of 4 mg mL⁻¹ using a BCA quantitative kit (Solarbio). The obtained erythrocyte membrane vesicles were stored at −80 °C until use.

4.3. Preparation of T-NPs

T-NPs were prepared using the nanoprecipitation method as previously described [31]. In brief, 10 mg PLGA (MW: 8000 mg, 50:50) was added to 10 mL acetone solution and then TGF-β1 was added to the solution, maintaining the concentration at 2 ng mL⁻¹. The solution was stirred at room temperature for more than 2 h until the PLGA and TGF-β1 were fully dissolved and mixed. The obtained solution was added dropwise to 30 mL ddH₂O using microfluidic equipment and then stirred for more than 12 h to fully evaporate the acetone. The solution was dialyzed in a 3500 kDa dialysis bag for more than 12 h to obtain PLGA nanoparticles. The nanoparticles solution was stored at 4 °C. To prepare DiR fluorescence-labeled T-NPs, 1 wt% DiR was added to a mixture of PLGA and acetone solution for fluorescence labeling. For Cy5 fluorescence-labeled nanoparticles, when preparing PLGA nanoparticles, 9 mL PLGA and 1 mg Cy5 fluorescent dye powder were mixed [55] and added to 10 mL acetone. The subsequent preparation process was the same as the procedure described above.

4.4. Preparation of T-NNPs and RBC-NPs

Nanoparticle cores and cell membrane materials were assembled using ultrasonic and co-extrusion methods [56]. The cell membrane material and the T-NPs core were mixed at a protein mass ratio of 2:1 and the mixture were placed in an ultrasonic water bath for more than 3 min. After passing through a polycarbonate membrane at 450 nm and 220 nm, the prepared T-NNPs were extruded back and forth more than 10 times. The prepared T-NNPs were placed in 1 × PBS at 4 °C for subsequent use. The RBC-NPs were prepared using the same method.

4.5. Preparation of GelMA microspheres and GM@T-NNPs

The GelMA hydrogel microspheres were prepared using a water-in-oil microfluidic method. A coaxial electrospinning needle (with an external diameter of the spinning needle of 30 G and an internal diameter of 21 G) was connected to the microfluidic thrust pump via a silicone tube. The outer needle of the coaxial electrospinning nozzle was used as a continuous phase for co-flow shearing and connected to an isopropyl myristate solution containing 10% (w/w) Span80. The inner needle was connected to the dispersed phase using a 7% (w/v) GelMA aqueous solution and 0.5% (w/v) I2595 photoinitiator. By adjusting the flow velocity of the aqueous (GelMA) phase to that of the oil phase at 15:500, continuous monodispersed spherical droplets were formed at the interface of the main phase. After confirming stable production of the GelMA droplets, the droplets were placed in a beaker filled with the oil phase to receive the microspheres under a dark field. The beaker was removed every 10–20 min, and after full irradiation with ultraviolet light and stirring at low speed, the microsphere droplets tended to spill up as individual droplets rather than adhering to each other. The collected cross-linked hydrogel microspheres were washed three times with 75% ethanol and washed with PBS five times. The PBS was changed every 3 h to remove the photoinitiator and oil. The purified microspheres were frozen overnight at −80 °C and then immediately freeze-dried for at least 48 h to obtain porous GelMA hydrogel microspheres. To prepare GM@T-NNPs, freeze-dried GelMA microspheres, 16 mg EDC, and 24 mg NHS were added to 2 mL MES buffer. The prepared T-NNPs were centrifuged at 20,000×g and resuspended in ddH₂O. The T-NPPs solution was added to the microsphere solution and incubated overnight in a constant-temperature shaker coat at 37 °C.

4.6. Preparation of GelMA

Gelatin (20 g) was mixed with 200 mL PBS and placed in a water bath and stirred at 60 °C for 1 h until the gelatin was completely dissolved to obtain a 10% (w/w) gelatin solution. Under dark-field conditions, 16 mL of methacrylic acid was dripped into the gelatin solution at a rate of 0.25 mL min⁻¹ by a microfluidic device. The mixture was stirred continuously at 60 °C for 2 h. Preheated PBS (800 mL) was added to the system and stirred continuously for 15 min. The mixture was dialyzed in a 12–14 kDa dialysis bag for 2 weeks. The PBS was changed every 2 days to obtain the GelMA solution and the GelMA hydrogel was obtained after freeze-drying the solution.

4.7. Characterization of nanoparticles and microspheres

The particle size and zeta potential of T-NPs and T-NNPs were measured using a Zetasizer (Malvern Zetasizer Nano zs90, UK) for DLS analysis. The morphologies of the NMVs vesicles, T-NPs, and T-NNPs were examined using TEM (Hitachi HT7700, Japan) at 200 kV. The GelMA and GM@T-NNP complexes were immobilized using a conductive adhesive, plated with gold for 45 s (Quorum Technologies, SC7620, UK), and observed under SEM (Hitachi S4800); the acceleration voltage was gradually adjusted to 10 kV. Elemental analyses of the GM and GM@T-NNP surfaces were conducted using EDS and XPS (Thermo Scientific Escalab, USA). The force-separation curve of freeze-dried GelMA hydrogel, GM and GM@T-NNPs was measured and analyzed by atomic force microscopy (AFM, Bruker, USA) under indentation model. Young's moduli was calculated according to each force-separation curve based on Hertzian model using the NanoScope Analysis software (1.7 version, Bruker, USA).

The neutrophil membrane release efficiency from T-NNPs and GM@T-NNPs was explored through in vivo fluorescence imaging. The DiR labeled neutrophil membranes were prepared following the procedure described in previous study [57]. Generally, after being collected and purified from bone marrow, neutrophils were labeled with 5 μl mL⁻¹ DiR (Abmole) cellular membrane labeling-solution for 20 min at 37 °C. DiR-labeled neutrophil membranes, T-NNPs-DiR and GM@T-NNPs-DiR were obtained after subsequent procedure as we described above. After injection of 20 μl of T-NNPs-DiR and GM@T-NNPs-DiR respectively, in vivo fluorescence images were photographed at day 0, day 1, day 3, day 7 and day 14 using optical imaging system (IVIS Lumina LT).

4.8. Determination of release and absorption curves of nanoparticles

For determination of the release curve of TGF-β1, T-NPs and T-NNPs (2 mL), theoretically containing 8 ng TGF-β1, were placed in a 35 kDa dialysis bag, immersed in a 50 mL centrifuge tube containing 10 mL PBS solution, and then placed on a thermostatic shaker at 37 °C. PBS in the centrifuge tube was collected after 0. 5, 1, 2, 4, 6, 8, 10, 14, 18, 24, 30, and 36 days. The fresh buffer solution in the centrifuge tube was changed after each collection. The cytokines released from the nanoparticles were detected by a rat TGF-β1 ELISA kit (SAB) and the cumulative absorption curve was drawn. To determine the inflammatory adsorption effect of the nanoparticles, recombinant TNF-α and IL-1β (initial amount of 8 ng) were mixed with different concentrations of T-NNPs and T-NPs, incubated at 37 °C for 2 h, centrifuged at 14,000×g for more than 10 min, and the supernatant was diluted 10 times after discarding the precipitate. The concentrations of residual cytokines in the supernatant were determined by rat TNF-α and IL-1β ELISA kits according to the manufacturer instructions, and the cytokine binding curve was drawn using Origin 2021 software.

4.9. Cell membrane laterality and surface protein receptor assays

Cell membrane laterality was examined using a rat glycoprotein assay. NMVs, T-NNPs with the same protein concentration and T-NPs were mixed with trypsin to initiate trypsin digestion. The samples were centrifuged at 3000×g for 5 min and the supernatant was collected. The glycoprotein content was quantified using a rat gp130 glycoprotein ELISA kit (Wuhan Jiyinmei Science and Technology Co., Ltd.). To examinate cell membrane laterality over time, T-NNPs solution with the same protein concentration were each incubated in PBS for 0, 10, 20, 30, 40, 50, 60 and 70 h. NMVs solution with the same protein concertation at 0 h was considered to contain 100 percentage of glycoprotein. After ELISA procedure we described above, glycoprotein content percentages at different time were measured. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was used to detect the content of receptor proteins on the cell membrane surface. Membrane proteins of NMVs vesicles and T-NNPs were extracted using a total protein extraction kit. The extracted membrane protein was run at 75 V for 0.5 h on a 10% pore microgel with a BioRad electrophoresis system and then at 140 V for 1 h. Finally, the polyacrylamide gel was stained overnight with Coomassie brilliant blue for visualization, followed by immersion in a Coomassie brilliant blue decolorizing solution for 5 h, during which the decolorizing solution was replaced two to three times until the protein gel had no obvious background color [58]. Western blotting was performed to detect the expression of receptor proteins in the nucleus pulposus cells, NMVs, T-NNPs, and T-NPs. Total protein extraction kits and BCA protein quantitative kits were used to extract and quantify the proteins in each group. The protein samples were subjected to electrophoresis on a 10% SDS-PAGE gel and transferred to a polyvinylidene fluoride film, followed by blocking and treatment with TNF-α receptor, IL-1β receptor, and CD11b primary antibodies (all purchased from SAB). The signals were detected after incubation with horseradish peroxide-labeled goat/rabbit secondary antibody (Beyotime) and measured using a chemiluminescence imaging system (Bio-Rad, Hercules, CA, USA).

4.10. Isolation and culture of nucleus pulposus cells

Nucleus pulposus cells were extracted from the caudal vertebrae of SD rats. The nucleus pulposus tissue of the rats was collected from the region between the two vertebrae in each segment using ophthalmic forceps under aseptic conditions. Type II collagenase (0.25%) was added to digest the tissue at 37 °C for 2 h. The cell suspension was filtered with 70 μm cell filters. The suspension was then centrifuged and the supernatant was discarded. The precipitates were collected and washed three times with aseptic PBS with the addition of DMEM/F12 containing 10% FBS. The cells were incubated at 37 °C in 5% CO₂ cell incubator. The culture medium was changed every 2–3 days during the cell culture process.

4.11. Migration and cell adhesion assays of T-NNPs

The nucleus pulposus cells were placed in the lower chamber of the transwell system with conditioned medium containing LPS (200 ng mL⁻¹), maintaining a cell density of approximately 1 × 10⁵ cells. Approximately 100 GM@T-NNPs nanoparticle–microsphere complexes were placed in the upper chamber, and the cells were cultured with the same amount of DiR fluorescence-labeled T-NPs, T-NNPs and RBC-NPs. The cell culture was terminated after incubation for 6 h and 72 h. The nucleus pulposus cells in the lower chamber were washed with PBS for five times, fixed with paraformaldehyde at room temperature for 30 min, and treated with 0.3% Triton X-100 for 30 min. After washing in PBS, the cells were blocked overnight at 4 °C with 5% bovine serum albumin (BSA) and then stained with phalloidine and DAPI (Abclonal, China). The entire process was performed in the darkness, and images were captured under an inverted fluorescent microscope (Zeiss). Semi-quantitative fluorescence analysis was performed using ImageJ software. The adhesion of T-NNPs and RBC-NPs to nucleus pulposus cells was observed by cryopreservation TEM. After culturing in the transwell system for 24 h, the nucleus pulposus cells in the lower chamber were collected, treated with 2.5% glutaraldehyde, and dehydrated with an alcohol gradient. The alcohol in the samples was then replaced with anhydrous propylene oxide or acetone and the samples were infiltrated with an epoxy resin. Acetone was then added to the epoxy resin at a ratio of 1.5%–2% and the samples were reacted for 12 h at 35 °C, 12 h at 45 °C, and 24 h at 60 °C. Finally, the samples were cut into ultrathin sections and observed using TEM (HT7700; Tokyo, Japan).

4.12. Determination of the apoptosis of nucleus pulposus cells

After being cultured with LPS-conditioned medium for 24 h in the transwell system (expect that the control group was cultured with DMEM/F12 without LPS), the cells of NC group, T-NP group, GM group, RBC-NP group, NMV group, T-NNP group and GM@T-NNP group were collected, fixed with paraformaldehyde at room temperature for 30 min, treated with 0.3% Triton X-100 for 30 min, washed with PBS three times, blocked with 5% BSA at 4 °C overnight, incubated with BAX and caspase-3 antibodies (SAB) overnight, and then stained with phalloidine and DAPI. The images were captured under an inverted fluorescent microscope (Zeiss), and fluorescence semi-quantitative analysis was performed as described above.

4.13. Live/dead staining and cell viability assay

To evaluate the biocompatibility of the GM@T-NNPs nanoparticle–microsphere complex, GM@T-NNPs were directly co-cultured with nucleus pulposus cells for live/dead staining to test their biotoxicity. The GM@T-NNPs were washed in PBS for five times and then laid at the bottom of a 24-well plate. Approximately 100 microspheres were placed in each well and inoculated with nucleus pulposus cells. The process was terminated on days 1, 3, and 5 of the co-culture. After the culture medium was removed, the Live/Dead stain was added and incubated at 37 °C for 30 min. The cell morphology was observed using an inverted fluorescence microscope. To determine the cell viability of each group, RBC-NPs, NMVs, T-NPs, GM, T-NNPs, and GM@T-NNPs were co-cultured with nucleus pulposus cells and intervened with 200  ng mL⁻¹ LPS. At 1, 3, and 5 days of co-culture, 10% CCK-8 staining solution was added and incubated at 37 °C for 4 h. The supernatants were centrifuged under 14,000×g to remove suspended nanoparticles and 100 μL medium was transferred to a 96-well plate. The absorbance of each group was measured by an enzyme-labeling instrument at 450 nm.

4.14. Ki-67 cell proliferation test

Nucleus pulposus cells were co-cultured with T-NPs, GM, RBC-NPs, NMVs, T-NNPs and GM@T-NNPs for five days with the intervention of 200  ng mL⁻¹ LPS. After 5 days of culture, cells were digested with trypsin and went through 70 μm cell filter to rule out the influence of GM and GM@T-NNPs. Then the cells were incubated in 70% ethyl alcohol for 1 h at 4 °C. After labelling with Ki-67 antibody (Abcam), the cells were washed and analyzed by flow cytometry (Merck Millipore, Germany).

4.15. Immunofluorescence

Nucleus pulposus cells were co-cultured with each group of materials as described above. After 7 days of culture, the cells were fixed, infiltrated, blocked, washed and incubated with IL-1β, TNF-α, and COL-II primary antibodies overnight, followed by incubation with secondary antibodies at room temperature for 2 h. After staining with phalloidin and DAPI, the samples were observed under a confocal microscope. Cell counting was conducted with ImageJ to detect the number of cells in each immune fluorescence images. Semi-quantitative fluorescence analysis was performed using ImageJ software. Integrated fluorescence intensity was divided by the cell count in each image and normalized immune fluorescence intensity was acquired.

4.16. Western blot

Based on the co-culture system described above, cells were stimulated with LPS (200 ng mL⁻¹) and cultured for 7 days. Proteins of each group were extracted by adding RIPA lysate containing a protease inhibitor. The protein concentration was quantified using a BCA protein quantification kit (Solarbio, China). Next, the protein samples were subjected to electrophoresis on SDS-PAGE gel and transferred to a polyvinylidene fluoride film. After blocking with QuickBlock western blocking buffer (Beyotime, China) for 15 min, the film was treated with the BAX, cleaved Caspase-3, IL-1β, IL-6, TNF-α, MMP-13, COL-II, aggrecan, PI3K, p-PI3K, AKT, p-AKT and β-actin antibodies, followed by the incubation of secondary antibody. The film was washed and scanned using a gel imaging system and the images were assessed using the ImageJ software for gray value analysis of the bands.

4.17. RT-qPCR analysis

Based on the above co-culture system, after LPS (200 ng mL⁻¹) stimulation, the relative mRNA expression levels of the IL-1β, TNF-α, IL-6, MMP-13, COL- Ⅱ, aggrecan, and β-actin genes in each group were detected with RT-qPCR. The relative expression levels of the PI3K and AKT genes were quantified in each group in the same manner. Primer sequences were designed by Sangon Biotech (Shanghai, China). Total RNA was extracted with TRIzol Reagent (Vazyme, China). A total of 0.5 μg isolated RNA was reverse transcribed into cDNA with the cDNA synthesis kit (Vazyme, China). qPCR was performed according to the procedure specified by the reagent manufacturer (Vazyme, China). Samples were normalized to β-actin. All primer sequences are shown in Supplementary Table 1.

4.18. RNA-sequencing and differential gene expression analysis

Nucleus pulposus cells were co-cultured with GM and GM@T-NNPs. After seven days of 200  ng mL⁻¹ LPS stimulation, the cells were digested with trypsin, collected, and RNA was extracted using TRIzol reagent (Invitrogen). The RNA purity was determined and quantified on a NanoDrop2000 spectrophotometer. Transcriptome sequencing and analysis were performed by OE Biotech Company on the illumina NovaseqTM 6000 platform. Original clean reads were obtained using Trimmomatichisat2 software. The base fragments obtained were quantified using Cufflinks software. GO and KEGG enrichment analysis was carried out to determine the biological functions or pathways mainly affected by differential transcripts. Q value < 0.05 and foldchange >2 or foldchange <0.5 was set as the threshold for significantly differential expression gene (DEGs). The heatmap was used to cluster the differential transcripts and evaluate the expression patterns between different samples.

4.19. Establishment of the rat IVDD model

Male SD rats with an average weight of 300–350 g were purchased from the Experimental Animal Center of Suzhou University. This study was approved by the Ethics Committee of First Affiliated Hospital of Suzhou University (Approval No. SUDA20230827A03) and good surgical and therapeutic outcomes were achieved. Rats were anesthetized with an intraperitoneal injection of 4% pentobarbital (25 mg kg⁻¹) and their tails were sterilized. Degeneration of the 7th–10th caudal vertebrae was induced with a 21 G needle puncture. To ensure the induction of trauma, the needle was inserted into the center of the intervertebral disc, rotated for 5 s, and kept still for 30 s. Microspheres, nanoparticles, or PBS (20 μL) were injected into each intervertebral disc with 26G needle connected to a 100-μL micro-syringe (Hamilton). After the operation, the rats were placed in a warm, ventilated position.

4.20. Radiographic evaluation of animal models

At 4 and 8 weeks postoperatively, three rats were randomly selected from each group and examined using digital radiography and MRI before sacrifice. Each rat was placed in the supine position and its tail was placed on a molybdenum target X-ray device. The X-ray images were evaluated using ImageJ software to calculate the DHI (%). MRI was performed with a 1.5T system (GE). T2-weighted images were obtained by coronal scanning. A radiologist who was blinded to the grouping information assessed the T2-weighted signal strength of each group, and the image was scored on a scale of I to IV to assess the pathological damage according to the modified Thomson classification as shown in Supplementary Table 2.

4.21. Histological evaluation, immunofluorescence, and immunohistochemistry

Four and eight weeks after the operation, the rats in each group were sacrificed. Intervertebral disc samples from each group were collected and immersed in formalin. After decalcification in 10% EDTA for 30 days and embedding in paraffin wax, the specimens were sectioned at a thickness of approximately 5 μm. Changes in tissue structure and collagen content in the intervertebral discs were observed by H&E and Safranin O/Fast green staining, respectively. Histological grading was performed using the scale established by Masuda et al. as shown in Supplementary Table 3. The score ranges from 4 (normal) to 12 (severe). COL-II, IL-1β, TNF-α, MPO, HIS48 and F4/80 immunofluorescence staining was performed in each group as described above.

4.22. Retention efficiency of T-NNPs and GM@T-NNPs In vivo

The preparation process for the Cy5 fluorescence-labeled T-NNPs and GM@T-NNPs is described above. After puncture of the C9–10 caudal vertebrae, SD rats were injected with T-NNPs and GM@T-NNP microspheres (20 μL) using a microsyringe. At days 0, 3, 7, and 14, images of the two groups of rats were obtained using a small-animal in vivo optical imaging system (IVIS Lumina LT), and the images were analyzed using Living Image software.

4.23. Verification of the PI3K/AKT pathway in vivo

The nucleus pulposus tissue of the rats was collected from the region between the two vertebrae in each segment using ophthalmic forceps after 4 weeks of animal surgeries and injection of materials in each group. The nucleus pulposus tissues were cut into small pieces and then lysed with RIPA lysis buffer containing proteinase and phosphatase inhibitor. The mixture was then grinded with glass homogenizer for 10 min. The protein concentrations were measured by the BCA protein assay kit (Solarbio, China). Equal amounts of total proteins were subjected and separated to 10% SDS-polyacrylamide gel electrophoresis (PAGE) and transferred to polyvinylidene fluoride film. After blocking for 15 min, the film was incubated with the β-actin, PI3K, p-PI3K, AKT and p-AKT primary antibodies overnight at 4 °C and then secondary antibodies. Finally, the film was washed and scanned using a gel imaging system and the images were assessed using the ImageJ software for gray value analysis of the bands.

4.24. Statistical analysis

All data are expressed as the mean ± standard deviation. Statistical analyses and images were processed using Origin software and assessed using one- or two-way analysis of variance. Tukey's multiple comparison test was used to further assess differences between groups (n = 3); p < 0.05 was considered to indicate a statistically significant difference.

Ethics approval and consent to participate

This study was approved by the Ethics Committee of First Affiliated Hospital of Suzhou University (Approval No. SUDA20230827A03) and good surgical and therapeutic outcomes were achieved.

Data availability

The authors declare that the data supporting the findings of this study are available within the paper and its Supplementary Material. Raw data generated for this study are available from the corresponding author on reasonable request.

CRediT authorship contribution statement

Liang Zhou: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Feng Cai: Resources, Methodology, Investigation, Formal analysis. Hongyi Zhu: Writing – review & editing, Methodology, Formal analysis, Conceptualization. Yichang Xu: Writing – review & editing, Writing – original draft, Methodology, Formal analysis, Data curation, Conceptualization. Jincheng Tang: Supervision, Formal analysis. Wei Wang: Investigation, Formal analysis. Ziang Li: Formal analysis, Data curation. Jie Wu: Investigation, Formal analysis. Zhouye Ding: Software, Methodology. Kun Xi: Writing – review & editing, Supervision, Funding acquisition, Formal analysis. Liang Chen: Writing – review & editing, Supervision, Methodology, Funding acquisition, Conceptualization. Yong Gu: Supervision, Methodology, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare no conflict of interest.

Appendix A. Supplementary data

The following is the supplementary data to this article: