Application of cerium dioxide nanoparticles in the treatment of intervertebral disc degeneration induced by oxidative stress Authors: Ruichao Cao, Pengyu Wang, Shuaichi Guo, Zhengyang Yang, Dongxu Li, Xiaohan Pan, Hai Mou, Zeyu Liu, Liping Wu, Jie Hao Published: 2026 ABSTRACT BackgroundOxidative stress (OS) induced by local reactive oxygen species (ROS) accumulation plays a vital role in initiating and accelerating intervertebral disc degeneration (IDD). Cerium oxide nanoparticles (CeNPs) have specific and efficient superoxide dismutase (SOD)-mimicking activity for ROS degradation. Their nanoscale particle size and nonstringent catalytic reaction conditions can potentially address challenges faced by traditional antioxidants in terms of remaining stable when used in degenerative avascular intervertebral disc (IVD).MethodsWe cocultured three kinds of IVD cells (cells of annulus fibrosus (AF), cartilage endplate (CEP) and nucleus pulposus (NP)) with CeNPs to determine the safe range of CeNPs concentrations. Additionally, Cell Counting Kit-8 (CCK-8) assay, ROS level assay and calcein/PI staining were performed to detect the protective effect of CeNPs on IVD cells against OS in vitro. Transmission electron microscopy (TEM), inductively coupled plasma mass spectrometry (ICP-MS) and blood analysis were used to describe its distribution pattern and the possible effects CeNPs may have on the body. Histological and immunohistochemical studies were implemented to analyse the capacity of CeNPs to prevent IDD induced by OS. Then western blotting was used to explore the corresponding mechanisms.ResultsThe results showed that ≤ 40 µM CeNPs was safe for the three kinds of IVD cells. CeNPs alleviated the inhibition of cell proliferation and reduced the degree of cell death caused by OS to different degrees. Additionally, the OS-induced degeneration of each part of the IVD was partially relieved after pretreatment with CeNPs. These improvements might be achieved through reducing OS-mediated cellular inflammatory response, extracellular matrix (ECM) catabolism and apoptosis.ConclusionThis study proposes that the local application of CeNPs to the IVD could mitigate OS-induced IDD. Our findings indicate that experimental designs incorporating multiple IVD structures, IVD cells, and surrounding tissues are essential for evaluating drug effect and safety in IDD therapeutic development.Graphical Abstract Supplementary InformationThe online version contains supplementary material available at 10.1186/s12967-026-08421-6. Introduction Low back pain (LBP), a common musculoskeletal symptom, affects many people at each income level and all ages [1]. Approximately 60% of LBP cases have not been effectively relieved even one year later [2]. Approximately 15−40% of LBP cases develop into chronic or recurrent LBP [3], which eventually results in disability for 10% of patients [4]. LBP can be caused by disparate factors, and at least 40% of LBP cases are considered directly related to intervertebral disc degeneration (IDD) [5]. The abnormal distribution of spinal stress caused by IDD is the basis of many spinal diseases, including disc herniation and nerve compression. Although the pathogenesis of IDD is complex and has not yet been fully elucidated, some studies have confirmed that the occurrence and progression of IDD are closely related to oxidative stress (OS) [6]. Reactive oxygen species (ROS) are a group of highly oxidatively active and unstable chemical substances, including superoxide anion (O2−), hydroxyl radical (·OH), hydrogen peroxide (H2O2), hypochlorite ion (OCl−) and nitric oxide (NO) [7]. They are the inevitable byproducts of aerobic metabolism in intervertebral disc (IVD) cells [8]. The production and removal of ROS are in a dynamic balance due to the presence of superoxide dismutase (SOD), catalase, glutathione, vitamin C, vitamin E and other antioxidants in cells [9]. However, when the IVD degenerates, ROS are overproduced, and the ability of IVD cells to remove ROS decreases. Known factors, including inflammation [10], nutritional deprivation [11], high glucose [12], hyperosmolarity [13] and aging [14], can lead to excessive production of ROS in IVD cells due to mitochondrial dysfunction. The overproduction of ROS, which cannot be eliminated in time, directly damage the nucleic acids, lipids and proteins of IVD cells [15]. Moreover, the cellular OS state induced by ROS transforms the anabolic extracellular matrix (ECM) phenotype of IVD cells into a catabolic [16] and proinflammatory [17] ECM phenotype. IVD cells with proinflammatory phenotypes and catabolic ECM phenotypes overexpress matrix metalloproteinase-13 (MMP-13) and matrix metalloproteinase-3 (MMP-3), as well as interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α). MMP-13 and MMP-3 are the key factors involved in the degradation of the IVD ECM [18], while IL-1β and TNF-α are representative proinflammatory cytokines that accelerate the degradation of the IVD ECM, promote the ageing of IVD cells, and initiate programmed cell death, including apoptosis, autophagy, necroptosis, pyroptosis and ferroptosis [19]. Inflammatory factors can also accelerate the production of ROS and further aggravate OS [20]. Considering the above mechanisms, the microenvironment of the IVD deteriorates continuously, eventually leading to irreversible IDD. To achieve ROS clearance and OS pathway blockade, researchers have tried diverse medicines, including glutathione [21], N-acetylcysteine [22], polyphenolic compounds [23], pyrroloquinoline quinone [24] and growth factors [25], for OS-induced IDD. Drug delivery difficulties caused by the avascular characteristics of IVDs and the calcification of degenerated CEPs [26] are worthy of attention. Additionally, the rapid consumption and inactivation of medicine caused by the harsh microenvironment of degenerated IVDs demands prompt solutions. Therefore, there is a need to develop medicines that can remain effective in degenerated IVDs after a single delivery. Cerium is the 58th element in the periodic table and can exist in two valence states: 3 + and 4+ [27]. Cerium oxide is a metal oxide with a typical fluorite structure and mixed valence states. It is used mainly in industrial catalysis and polishing. As the diameter of cerium oxide nanoparticles (CeNPs) decreases, their surface area to volume ratio increases, the number of 3 + sites increases, and oxygen atoms are lost to form oxygen vacancies [28]. These oxygen vacancies can provide redox binding sites for free radicals, and thus, free radicals are scavenged [29]. Cerium in CeNP crystals exists mainly in the form of cerium dioxide, and its high redox activity results from the fast and convenient transition in oxidation state from Ce3+ to Ce4+. During the reversible conversion of Ce3+ to Ce4+, O2−, ·OH and H2O2 are ultimately converted to H2O and O2 [30]. In recent years, CeNPs, which mimic the activities of SOD and catalase, have been shown to reduce the level of free radicals in the cellular microenvironment and can be applied to a variety of diseases caused by OS [31]. There have been reports of CeNPs being used in the treatment of ophthalmology [32], gastroenterology [33] and dermatology [34] related diseases. As inorganic catalysts, CeNPs can adapt to the microenvironment of degenerated IVDs, which inspired this study. Materials and methods Materials and experimental animals The solution of CeNPs was customized from XFNANO (China). Cell Counting Kit-8 (CCK-8) was purchased from APExBIO (USA). The Hematoxylin and Eosin (HE) Staining Kit, Calcein/PI Cell Viability/Cytotoxicity Assay Kit, ROS assay and Total Superoxide Dismutase (SOD) Assay Kit were purchased from Beyotime (China). The Western Blot Kit was purchased from Epizyme (China). The Safranin O-Fast Green Staining Kit was purchased from Leagene (China). Sprague Dawley rats (male, 5 weeks old, weighing 150 ± 10 g) were purchased from the Animal Experimental Center of Chongqing Medical University. Characterization of CeNPs The morphology of the CeNPs was characterized using transmission electron microscopy (TEM) (JEM-2100, Japan). The hydrodynamic particle size and zeta potential were measured with a Malvern laser particle size analyzer (Malvern Instruments, UK). The SOD-mimicking activity was detected by a Total SOD Assay Kit. Cell culture IVD cells were isolated and cultured as reported in previous studies [35]. Briefly, annulus fibrosus (AF), cartilage endplate (CEP) and nucleus pulposus (NP) tissue was harvested from the tails of rats under sterile conditions after euthanasia. Tissue was washed with phosphate-buffered saline (PBS) and cut into small pieces (approximately 1 mm3). Then, AF tissue was treated with 0.02% collagenase type I, while CEP and NP tissue was treated with 0.02% collagenase type II, and these procedures were carried out in a 37 °C water bath with shaking for 0.5 h. The cell suspension was filtered with a nylon mesh (70 μm) to remove the tissue debris and centrifuged (300 g, 5 min) to harvest the cells. AF, CEP and NP cells were cultured at 37 °C and 5% CO2 with complete medium, which consisted of 89% Dulbecco’s modified Eagle’s medium: nutrient mixture F-12 (DMEM/F12, Gibco, USA), 10% FBS (fetal bovine serum, Gibco, USA) and 1% penicillin–streptomycin (Gibco, USA). The third to fifth passages of cells were used for in vitro studies. CCK-8 assay To evaluate the biocompatibility of CeNPs, IVD cells were pretreated with CeNPs (0, 20, 40, 60, 80, 100 µM) for 24 h and then detected by CCK-8 assay at 24, 48 and 72 h. To assess the protective effect of CeNPs on IVD cells under OS, IVD cells were divided into 4 groups (sham group, H2O2 group, 1 µM group and 10 µM group). 1 µM group and 10 µM group were respectively pretreated with 1 µM and 10 µM CeNPs for 24 h and, then together with group H2O2, were cultured in complete medium containing 50 µM H2O2 for 24 h, while sham group did not receive any additional processing. After treatment, 4 groups were detected by CCK-8 assay at 24, 48 and 72 h. To quantify the toxicity of CeNPs on IVD cells, the cells were divided into groups and cultured for 48 h in complete medium containing CeNPs at concentrations of 4000, 2000, 1000, 500, 250, 125, 62.5, 32, or 0 µM. Following this treatment, cell viability was assessed using a CCK-8 assay. Calcein/PI staining The protective effect of CeNPs on IVD cells against OS was assayed using Calcein/PI staining. The grouping and processing of IVD cells was almost the same as above, the only difference was that the concentration of H2O2 was increased to 300 µM. Twenty-four hours later, protocols provided by manufacturing guidelines were performed. The results were observed under a fluorescence microscope, and the viable cells were green, while dead cells were red. The fluorescence intensity was quantified using ImageJ Ver. 1.53t (NIH, USA). Western blot assay The grouping and processing method of IVD cells referred to the treatment method of the Calcein/PI staining experiment. However, the treatment concentration of H2O2 was lowered to 100 µM to simulate the slow natural process of IDD. An equivalent amount of H2O2 was supplemented daily to maintain this concentration, and the processing time was extended to 72 h. After being treated, cells were washed with ice-cold PBS and lysed with RIPA lysis buffer containing 1% PMSF for 30 min at 4 °C. The lysates were collected and decomposed by sonication. Then, the samples were centrifuged (12,000 g, 15 min) at 4 °C. After the protein concentration was determined by a Bradford Kit (Beyotime, China), the quantified protein samples (20 mg total protein) were separated by SDS-polyacrylamide gel electrophoresis and transferred to PVDF membranes. The membranes were further blocked with TBST containing 5% skim milk and incubated with primary rabbit antibodies against GAPDH (1:1000, GB11002, Servicebio), IL-1β (1:500, WL00891, Wanleibio), TNF-α (1:1000, PY19810, Abmart), MMP-3 (1:1000, 380816, zenbio), MMP-13 (1:1000, 381531, zenbio), Bcl-2 (1:1000, T40056, Abmart) and cytochrome-c (1:1000, T55734, Abmart) overnight at 4 °C. After washing, the membranes were incubated with secondary antibody for 1 h at room temperature. Then, an enhanced chemiluminescence system (Fusion Fx, France) was used to visualize the specific protein bands. Images were quantified using ImageJ Ver. 1.53t (NIH, USA). IVD dissection and organ culture IVDs were harvested under sterile conditions from the tails of SD rats [36] within 2 h after sacrifice. To remove blood clots, IVDs were washed under agitation in PBS supplemented with 1% penicillin‒streptomycin-amphotericin B and 20 mM Na-citrate in sealed sterile centrifuge tubes (30 min, three times). After washing, IVDs were cultured in 12-well plates with 2 ml of complete medium (DMEM/F12 containing 10% fetal bovine serum, 1% penicillin‒streptomycin-amphotericin B and 20 mM Na-citrate) (Fig. S1). Histological and immunohistochemical studies IVDs were randomly assigned to 4 groups (sham group, H2O2 group, 1 µM group and 10 µM group) and cultured with complete medium containing different concentrations of CeNPs (0, 0, 1, and 10 µM) for 24 h, followed by being cultured with complete medium containing various doses of H2O2 (0, 100, 100, and 100 µM) for 48 h. The above treatments were one cycle. Seven cycles later, the IVDs were sequentially washed with PBS (30 min, three times), fixed with 4% buffered paraformaldehyde for 24 h, decalcified in 10% EDTA under agitation for 1 month, dehydrated in graded ethanol solutions, embedded in paraffin and cut into serial slices (4 μm thick). HE staining, safranin O-fast green staining and IHC were used to compare morphological changes and the degree of degeneration in each group. Images were scanned with KF-viewer (KFBIO, China) and quantified using ImageJ Ver. 1.53t (NIH, USA). The degeneration grading system for rat IVDs was introduced to grade the degree of intervertebral disc degeneration in rats [37]. The scoring criteria were based on the NP morphology, NP cellularity, NP-AF border, AF morphology and endplate in safranin O-fast green staining images (Table. S1). Statistical analysis Each experiment involved in this study was independently repeated at least 3 times. All experimental data were presented as the mean ± SD, n ≥ 3 and were analyzed using GraphPad Prism (Ver.9.5.0). Differences between groups were analyzed by one-way ANOVA followed by Tukey’s post hoc test. P values < 0.05 indicated differences with statistical significance. In the statistical chart, ∗, ∗∗, ∗∗∗ and ∗∗∗∗ respectively represent p < 0.05, p < 0.01, p < 0.001 and p < 0.0001. Results Characterization of CeNPs The aqueous solution of CeNPs showed a light-yellow color at room temperature (Fig. S2A). The topographical features of CeNPs captured by TEM showed that CeNPs gathered into clumps with diameters of approximately 100 nm (Fig. S2B). Dynamic light scattering (DLS) studies revealed that the average hydration particle size of CeNPs was 103.19 ± 19.57 nm (polydispersity = 0.838) (Fig. S2C). The zeta potential was − 42.71 ± 2.79 mV (Fig. S2D). In the total SOD assay, the inhibition ratio of the reaction increased with increasing concentrations of CeNPs, indicating that the SOD-mimicking activity increased with the concentration of CeNPs (Fig. S2E&F). Biocompatibility of CeNPs The results of the CCK-8 assay revealed that during the 3 days of culture, the cells in each group exhibited sustained proliferation, but the concentration of CeNPs affected the rate of proliferation of the three kinds of cells to some extent. AF cells treated with 100 µM CeNPs exhibited significantly inhibited proliferation at 48 h and 72 h, and AF cells treated with 80 µM CeNPs presented the same trend at 72 h (Fig. 1A). Moreover, CEP cells treated with 100 µM CeNPs and NP cells treated with 60, 80, or 100 µM CeNPs presented similar proliferation inhibition at different time points (Fig. 1B&C). In contrast, the proliferation of AF cells and CEP cells treated with 20 µM CeNPs significantly increased at 72 h (Fig. 1A&B). The cell viability of AF, CEP, and NP cells decreased as the CeNPs concentration increased. The half-inhibitory concentrations (IC50) of CeNPs for AF, CEP, and NP cells were 606.1 µM, 617.0 µM, and 317.4 µM, respectively (Fig. S5). Fig. 1In vitro evaluation of biocompatibility of CeNPs. (A), (B) and (C) Cell proliferation of AF, CEP and NP cells incubated with different concentrations of CeNPs In vitro evaluation of biocompatibility of CeNPs. (A), (B) and (C) Cell proliferation of AF, CEP and NP cells incubated with different concentrations of CeNPs In vitro protective effect of CeNPs against OS The results of the CCK-8 assay revealed that during the 3-day culture, the proliferation of the three kinds of IVD cells in the H2O2 group was significantly inhibited by H2O2-induced OS (sham group vs. H2O2 group). However, when protected by CeNPs, the inhibition of AF, CEP and NP cell proliferation was significantly alleviated (H2O2 group vs. 1 µM group and 10 µM group). (Fig. 2A-C). Fig. 2In vitro evaluation of the protective effect of CeNPs against OS. (A), (B) and (C) Cell proliferation of AF, CEP and NP cells protected by CeNPs under H2O2 induced OS. (D), (E) and (F) Viability of AF, CEP and NP cells protected by CeNPs under H2O2 induced OS. (G), (H) and (I) Semiquantitative statistical analysis of the fluorescence intensity of (D), (E) and (F) In vitro evaluation of the protective effect of CeNPs against OS. (A), (B) and (C) Cell proliferation of AF, CEP and NP cells protected by CeNPs under H2O2 induced OS. (D), (E) and (F) Viability of AF, CEP and NP cells protected by CeNPs under H2O2 induced OS. (G), (H) and (I) Semiquantitative statistical analysis of the fluorescence intensity of (D), (E) and (F) Consistent with the results of the CCK-8 assay, the results of calcein/PI staining revealed that H2O2 caused the death of IVD cells (sham group vs. H2O2 group) (Fig. 2D-F). Under the protection of CeNPs, cell death was significantly reduced (H2O2 group vs. 1 µM group and 10 µM group). (Fig. 2G-I). Statistical analysis (Fig. S4) of each group revealed that H2O2 treatment significantly increased ROS levels in all component cells of the IVD and in the IVDs overall. CeNPs treatment inhibited this increase in ROS levels to varying degrees across the groups. Protective effect of CeNPs against OS-induced IDD in an ex vivo model From the perspective of the entire IVD area, the HE and safranin O-fast green staining images of the H2O2 group (sham group vs. H2O2 group) showed obvious morphological changes. In contrast, fewer morphological changes were observed in the 1 µM group and the 10 µM group (sham group vs. the 1 µM and 10 µM groups) (Fig. 3A (Scale bar: 1000 μm)). Fig. 3(A), (B), (C) and (D) Representative histology and IHC images of IVD, AF, CEP and NP cultured under OS and protection of CeNPs. (E), (F) and (G) are the semiquantitative statistical analysis of Safranin O-fast green staining images of (B), (C) and (D). (H), (I) and (J) are the semiquantitative statistical analysis of IL-1β IHC images of (B), (C) and (D) (A), (B), (C) and (D) Representative histology and IHC images of IVD, AF, CEP and NP cultured under OS and protection of CeNPs. (E), (F) and (G) are the semiquantitative statistical analysis of Safranin O-fast green staining images of (B), (C) and (D). (H), (I) and (J) are the semiquantitative statistical analysis of IL-1β IHC images of (B), (C) and (D) The healthy AF in the sham group had a dense structure, and the fibres were arranged in an orderly manner. The outer AF was stained blue‒green, and the inner AF, which is the transition to the NP and contains more collagen II and aggrecan, was stained the same red as was the NP. (Fig. 3B a&e(Scale bar: 200 μm)). Owing to OS-induced damage, the structure of the AF in the H2O2 group became loose and disordered, and the colour resulting from Safranin O-Fast Green staining changed significantly (sham group vs. H2O2 group) (Fig. 3B b&f). The H2O2 group had the highest disordered ratio of AF (The width of AF with disordered structure / The total width of AF * 100%) (Fig. 3E). With CeNP protection, the degree of AF degeneration in the 1 µM group and the 10 µM group was mild (H2O2 group vs. 1 µM and 10 µM groups) (Fig. 3B c&d, g&h). The disordered ratio of AFs in the 1 µM and 10 µM groups was lower than that in the H2O2 group (Fig. 3E). The healthy CEP in the sham group had a complete structure, and almost every cartilage lacuna contained CEP cells (Fig. 3C a (Scale bar: 150 μm)). In the Safranin O-Fast Green staining image, the extracartilage matrix was stained red (Fig. 3C e). In response to OS, the number of CEP cells in the H2O2 group was significantly lower. Thus, many cartilage lacunae were empty (Fig. 3C b). OS can cause cartilage calcification. The calcified CEP was stained blue with Safranin O-Fast Green. The calcium nodules in the H2O2 group were the most visible and were stained dark blue (Fig. 3C f). The H2O2 group had the highest calcification ratio of CEPs (the area of calcified CEPs/the total area of CEPs * 100%) (Fig. 3F). Less calcification occurred in the 1 µM group and the 10 µM group. Therefore, the CEPs in these two groups were stained lighter blue (Fig. 3C g&h). The calcification ratio of CEPs in the 1 µM group and the 10 µM group was lower than that in the H2O2 group, whereas the calcification ratio of CEPs in the 1 µM group was greater than that in the 10 µM group (Fig. 3F). In the sham group, the healthy NPs almost filled the entire space enclosed by the AFs and CEPs, and many NP cells were evenly distributed in the ECM. The ECM of the NP cells was very dense and stained red with Safranin O-Fast Green (Fig. 3D e (Scale bar: 200 μm)). Owing to OS, the ECM in the H2O2 group was degraded and almost invisible (Fig. 3D b&f). The H2O2 group had the smallest NP area ratio (area of NP/area enclosed by AF and CEP * 100%) (Fig. 3G). The ECM in the 1 µM group and the 10 µM group also degraded to a certain extent and became sparse (Fig. 3D g&h). However, the areas of the NP in the 1 µM group and the 10 µM group were still greater than those in the H2O2 group. Moreover, the area of NP in the 1 µM group was less than that in the 10 µM group (Fig. 3G). HE and Safranin O-Fast Green staining images also revealed that the number of NP cells roughly matched the area of the NP in each group (Fig. 3D a-d). The IL-1β IHC results were consistent with the western blot results. There was a significant increase in the proportion of IL-1β-positive cells in each structure of the IVD in the H2O2 group (sham group vs. H2O2 group) (Fig. 3B-D, H-J). This proportion decreased when CeNPs were introduced (H2O2 group vs. 1 µM group and 10 µM group for AF, CEP and NP cells). It was demonstrated in the results of rat IVD degeneration grading system that the sham group consistently received the lowest score of 0 for nearly all features, whereas the H2O2 group received the highest score of 2 for most features. Treatment with CeNPs resulted in a reduction in scores across all features, though the degree of improvement varied (Table. S1). Possible mechanism for the protective effect of CeNPs The western blot images revealed that OS upregulated the expression of MMP-13, MMP-3, IL-1β, TNF-α and cytochrome C and downregulated the expression of Bcl-2 in AF, CEP, and NP cells (sham group vs. H2O2 group) (Fig. 4A-C). However, CeNPs downregulated the expression of MMP-13 (Fig. 4D, F&H), MMP-3 (Fig. 4E, G&I), IL-1β (Fig. 4J, L&N), TNF-α (Fig. 4K, M&O) and cytochrome C (Fig. 4Q, S&U). Additionally, CeNPs upregulated the expression of Bcl-2 (Fig. 4P, R&T). Fig. 4(A), (B) and (C) Western blot assay results of key proteins expressed by AF, CEP and NP cells under OS and protection of CeNPs. (D) to (U) are the semiquantitative statistical analysis of western blot images of these proteins (A), (B) and (C) Western blot assay results of key proteins expressed by AF, CEP and NP cells under OS and protection of CeNPs. (D) to (U) are the semiquantitative statistical analysis of western blot images of these proteins Discussion Since the relationship between OS and IDD was revealed, various antioxidants have been applied for the prevention and treatment of IDD. However, the avascular characteristic of IVDs, the harsh microenvironment inside degenerated IVDs [38] and the difficulty of material exchange caused by calcification of CEP [34] are inescapable challenges for the clinical application of IDD drugs. Due to the nanoscale particle size and ROS-scavenging activity of CeNPs, they are able to disperse stably and alleviate the progression of OS-induced IDD to a certain extent. Compared to traditional ROS scavengers, CeNPs offer several distinct advantages. Their nanoscale dimension facilitates higher catalytic efficiency resulted from a larger surface area to volume ratio. As catalytic agents, CeNPs promote ROS degradation during oxidative stress alleviation without a theoretical decrease in mass, enabling continuous function once inside the disc. Furthermore, being inorganic catalysts, CeNPs impose less stringent requirements on the reaction environment than biological enzymes, resulting in greater stability and efficacy within the harsh microenvironment of degenerative IVDs. The catalytic activity of CeNPs increases with their concentration theoretically, which contrasts with the behavior of traditional drugs. However, the CCK-8 assay results constrain the permissible concentration of CeNPs. The biocompatibility of CeNPs was inconsistent across different IVD cell types. While prior work has established that the IVD’s components are structurally and functionally integrated, with cell types closely linked in both spatial location and metabolism, many studies have nevertheless selected a single cell type or structural component for experimental observation. Our study included the three major IVD cell and structure types to circumvent the limitations inherent in using a single structure or cell type. The half-inhibitory concentrations (IC50) of CeNPs against AF and NP cells, which may directly contact the CeNPs, are 606.1 µM and 317.4 µM, respectively. Although these values are tens of times higher than the concentrations used in this study, the dosage and total amount of CeNPs delivered to the IVD must still be carefully considered. The IVD’s relatively closed structure and limited material exchange, together with the entrapment of CeNPs by collagen fibers in the NP, hinder their escape following introduction. Furthermore, CeNPs are not uniformly distributed within the IVD, creating potential local concentration hotspots. NP cells can also phagocytose CeNPs, resulting in intracellular accumulation. Finally, the fragile nature and slow proliferation of IVD cells imply that any factor affecting cell proliferation could be devastating. The application of CeNPs significantly reduces ROS levels throughout the entire IVD and within its constituent cell populations. CeNPs exhibit diverse biomedical effects in previous studies, including antioxidant therapy, anti-tumor activity, antibacterial action, and neuroprotection [39]. In IDD, oxidative stress driven by excessive ROS production is a key pathophysiological process affecting all three IVD cell types, a process concurrently linked to the biological activity of CeNPs. Therefore, the primary mechanism by which CeNPs mitigate IDD is likely their antioxidant function. The bio-toxicity of CeNPs depends not only on their concentration and exposure duration, but also on their particle size, shape, and surface modifications. Furthermore, their effects vary across different organs [40–44]. Marzi et al. reported that CeNPs exhibit high long-term toxicity but minimal short-term toxicity in A549, CaCo2, and HepG2 cells, a human hepatic carcinoma cell line [45]. To investigate potential systemic effects from the local intradiscal injection of CeNPs, we analyzed multiple blood indicators and measured cerium content in major organs. The blood analysis revealed no significant short-term changes in RBC, WBC, PLT, ALT, AST, or Scr following the local application of CeNPs to the rat IVDs. These values remained within established reference ranges for rats [46], indicating that the local injection did not cause acute damage to the blood system or function of liver and kidneys in the short term. ICP-MS analysis showed that the average proportion of the injected CeNPs dose redistributed to other organs ranged from 0.1135 ± 0.1965% to 0.9884 ± 1.711%. Given the very low concentration of CeNPs used, the redistributed quantity likely exerted a limited impact on organs outside the IVDs. The measured recovery rate of cerium indicates that nearly all injected CeNPs were detected. The recovery rate below 100% may be attributed to several factors: the prepared CeNPs injection concentration was below 10 µM, the precision of the injection tools and procedure was limited, cerium may have been metabolized and excreted (as the excrement of rats was not measured), or cerium may reside in other undetected tissues. Although statistical analysis indicated a difference in the distribution of cerium in IVDs between the one-hour and one-week time points, cerium was detected in extraspinal organs in only one rat from the 1-week group and one from the 2-week group. This sporadic detection might result from accidental intravascular injection caused by animal movement during the procedure, suggesting that the injection process itself significantly influences CeNPs redistribution. The two-week experimental observation confirmed that only a minimal fraction of CeNPs injected into IVDs redistributed in the short term. Longer-term observation remains necessary, however. This finding does not preclude the possibility that under extreme conditions—such as the severe degeneration persisting after CeNPs application, where extracellular matrix degradation frees CeNPs from constraint of the collagen fiber, or during disc herniation, where the nucleus pulposus is directly exposed outside the disc—CeNPs could still be released. The research conducted by Kalyanaraman et al. indicates that after the implantation of CeNPs low levels of particle migration to the liver, lungs, spleen and kidneys were detectable within 28 days, but no significant systemic toxicity or genetic toxicity was observed [47]. Ernst and coworkers’ research was the first to report the 100-day excretion dynamics of 3 nm CeNPs in mice, demonstrating the excellent biological safety of CeNPs [48]. Natalia et al. conducted the CeNPs exposure research which covered the entire life cycle of drosophila melanogaster. The results showed that CeNPs have good biocompatibility and little effect on the lifespan of drosophila melanogaster. But the changes occur in adult size and larval crawling [49]. Although IVD is a relatively enclosed environment, especially during its degeneration process, the long-term in vivo biocompatibility of CeNPs when applied in IVD still requires further observation. A systematic toxicity assessment of CeNPs across the entire rat life cycle was not conducted, which is a limitation of this study. CeNPs, when taken up by human cells, have been described as a double-edged sword in some studies because of their innate radical scavenging activity and biotoxicity. Watson and coworkers reported that CeNPs at a concentration of 20 µg/mL (116 µM) significantly inhibited proliferation, whereas CeNPs at a concentration of 5 µg/mL (29 µM) significantly decreased the viability of human lymphoblastic TK6 cells. However, the same dose of CeNPs did not induce a similar decrease in proliferation or viability in H9T3 Chinese hamster ovary cells [50]. Auffan and coworkers reported that CeNPs were genotoxic to human fibroblasts when the concentration of CeNPs was greater than 6 × 10 –5 g/L (0.35 µM) [42]. These two examples suggest that the minimum toxic concentration of CeNPs for human cells may be lower than that for rodent cells. Therefore, future experiments should incorporate human IVD cell lines and tissues to establish the balance between the minimum toxic and effective concentrations of CeNPs for these cells. The reduction in IVD height during degeneration decreases the overall IVD volume, thereby increasing the concentration of CeNPs. To prevent dose-related toxicity in clinical applications, the minimum effective concentration of CeNPs should be employed. We therefore selected a low CeNP concentration (≤ 10 µM) for this study to mitigate potential side effects. To address the limitations of the needle puncture and H2O2 injection models, which fail to accurately replicate the pathophysiological process of OS-induced IDD, we developed an in vitro IVD culture system for precise control over both CeNP concentration and oxidative stress intensity. The following IHC analysis demonstrated that CeNPs effectively mitigated inflammation in both IVD cells and their microenvironment. HE staining verified the protective effect of CeNPs on IVD cell morphology. Furthermore, Safranin O-Fast Green staining confirmed that CeNPs ameliorated ECM degradation and structural damage across all IVD compartments. The ex vivo experiments not only confirmed the validity of the CeNPs but also standardized the experimental procedure and simplified the experimental process, providing a new option for relevant research that needs accurate drug concentration. In addition, each SD rat can provide multiple IVDs for experiments, which can reduce the use of experimental animals to a certain extent. Notably, ex vivo experiments cannot completely replace experiments in vivo, especially when pathophysiological processes involve systemic regulation. The factors that affect the development and progression of IVD diseases, such as the spinal anatomy, biomechanics, physiology, immune system and inflammatory response, differ between animals and humans [51]. Compared with that of humans, the rodent NP retains more notochordal cells, which can result in relatively slower degeneration [52, 53]. These findings indicate that drugs that are effective for treating IDD in rats may not be effective in humans. The average lifespan of a rat ranges from 2 to 4 years. This limits the study of the effects of long-term treatment with CeNPs and the long-term toxicity of CeNPs on IVDs. Fortunately, some other animal models can overcome these shortcomings [54]. Therefore, combining a variety of animal models and exploiting their respective advantages to simulate the pathophysiological process of IDD as much as possible is needed in subsequent studies. In this study, OS significantly inhibited the proliferation of IVD cells, leading to their death and causing IDD at the structural level. The above negative changes were particularly pronounced in NP and cells of NP. This finding indicates that each part of the IVD is not uniformly damaged under OS, and compared with other parts, the NP is most significantly affected by OS. The NP is the weakest part of the IVD system in this study. The above inference implies that the IDD could be solved by resolving the principal contradiction, the degeneration of the NP. However, histological studies have shown that the degeneration of different IVD structures has different characteristics. The degenerated AF is characterized mainly by structural disorder, CEP degeneration is reflected primarily as calcification of the cartilage matrix, and the major characteristics of the degenerated NP are the loss of active functional cells and degradation of the ECM. The lost cells and the degraded ECM of the NP can be replaced and repaired by cell proliferation and ECM synthesis, respectively. However, repairing the disordered AF structure and reversing CEP calcification can be challenging. From this perspective, the AF and CEP appear more susceptible to damage than does the NP. This fact also confirms the necessity of integrating the main structures and cells of IVDs into research simultaneously. Western blot analysis showed that the relative expression levels of MMP-13, MMP-3, IL-1β, TNF-α, and cytochrome C were significantly higher in the H2O2 group than in the sham group, while Bcl-2 expression exhibited the opposite pattern. Pretreatment with 1 µM or 10 µM CeNPs reduced the expression of MMP-13, MMP-3, IL-1β, TNF-α, and cytochrome C relative to the H2O2 group. Conversely, Bcl-2 expression was elevated in both CeNPs-treated groups compared to the H2O2 group. The results suggests that CeNPs exert their protective effect by attenuating cellular inflammation, downregulating the expression of ECM-degrading enzymes, and inhibiting apoptosis. Specifically, in the three types of IVD cells, there was no significant difference in most indicators between the 1 µM group and the 10 µM group. Two factors likely explain this observation. First, cellular uptake of CeNPs by the three IVD cell types may have reached saturation at the 1 µM concentration. Second, the level of ROS induced by OS in this study probably did not exceed the maximum catalytic capacity of 1 µM CeNPs. These interpretations are supported by semiquantitative results from ex vivo experiments. When using whole IVDs with more complex three-dimensional structures instead of isolated cells, concentration gradients likely led to non-uniform distribution and retention of CeNPs across different tissue regions within each treatment group. Furthermore, the sustained generation of ROS under OS conditions may have exceeded the catalytic capacity of the available CeNPs. A limitation of this study is its lack of direct investigation into the potential signaling pathways involved. This study confirms that CeNPs alleviate oxidative stress in IVD cells by scavenging ROS. Previous studies have shown that ROS can activate the signalling molecules ERK, JNK, and p38 to regulate matrix metabolism, the proinflammatory phenotype and cell cycle progression in rat IVD cells [55]. In addition to ERK, JNK, and p38, other signalling molecules, such as Akt, ATM, Chk2, Nrf2, p53, and p65, are involved in the regulation of corresponding pathophysiological processes in human IVD cells [20]. Due to current experimental constraints and there are certain differences in metabolism and degeneration between the IVD of rodents and humans., we have verified only the downstream effector molecules of these pathways. This verification indirectly suggests the underlying mechanism. We acknowledge that a full exploration of the mechanistic pathway would strengthen the article’s completeness, and we intend to prioritize the experimental validation of these pathways by including human IVD cells and tissue in future work. In our study, OS-induced damage was not fully reversed or blocked in any treatment group. The preventive application of CeNPs only delayed and alleviated the progression of OS-induced IDD to a limited extent. Consequently, CeNPs are not perfect therapeutic agents for IDD. Since IDD can initiate as early as adolescence and OS-induced degeneration is a cumulative, age-progressive process, the preventive use of CeNPs represents a context in which their utility is more effectively realized. One thing that previous researchers rarely noticed is that as IDD progresses, when a herniated disc occurs, the CeNPs that accumulate in the IVD can spill out into the surrounding tissue like a flood. Under such conditions, whether the surrounding tissue is still safe has rarely been considered. Taking the CeNPs used in this study as an example, evaluating the biocompatibility of drugs on IVDs alone, including the tissues surrounding IVDs, is not comprehensive in the research and development of drugs that act on IVDs. When patients with IDD experience disc herniation during treatment, the drug (especially a drug delivered using intradiscal injection), which was originally limited to the IVD, can leak through disc herniation. Drug leakage-induced osmotic pressure changes and side effects of the drug itself combined with disc herniation may cause multiple strikes to the surrounding tissue, such as the spinal cord or nerve root. Luckily, in the study by Sun Lidi [56], a detailed assessment of the toxic effects of CeNPs on nerve cells was performed. According to their findings, when the accident described above occurs, not only will the concentration of CeNPs used in our study not cause toxic damage to nerves, but it can also alleviate the damage caused by compression of the spinal cord and nerve root. Based on results of degeneration grading, the ex vivo IVD culture model successfully induced severe IDD via oxidative stress, manifesting all features defined by the degeneration grading system for rat IVDs and confirming the model’s disease representativeness. These degeneration events occurred sequentially or simultaneously within a short period of 3 weeks. CeNPs application partially mitigated the overall degenerative progression. However, the influence of CeNPs varies in the specific features defined by the grading system. These differences may stem from: (1) The distinct oxidative stress tolerance of different IVD cells. (2) The interaction of CeNPs with these diverse cell types likely differs. (3) The inherent self-repair capacity of the IVD’s structures is not uniform. In TEM images, CeNPs within the NP are interwoven with collagen fibers, a configuration that restricts the mobility of the CeNPs entering the IVDs and promotes their retention inside the IVDs. This localization also limits the potential systemic toxicity associated with secondary distribution of CeNPs. CeNPs were further observed to be internalized by NP cells. These findings indicate that CeNPs are able to scavenge ROS both in the extracellular matrix and inside the cells of the NP. CeNPs were also detected within the lamellar structure of the AF, suggesting their potential activity in this region as well. However, without direct imaging of AF cells, their precise mode of action on these cells remains unclear. Only a very small number of CeNPs were observed within the CEP. This likely resulted from the denser tissue structure of CEP, which filtered most CeNPs and only allowed those with smaller particle sizes to be internalized. Histological results nevertheless demonstrate that CeNPs partially alleviate CEP degeneration. These improvements likely result primarily from the indirect protection conferred by CeNPs present in the surrounding environment of the CEP, rather than primarily from the direct action of the small fraction of CeNPs that are internalized by CEP. Since the CEP exchanges metabolites with the NP via diffusion, CeNPs within NP may consume ROS in the NP, creating a concentration gradient that draws ROS from the CEP to be decomposed. This process would indirectly protect CEP cells. Because this protection is indirect, the ameliorative effect of CeNPs on oxidative stress-induced CEP degeneration is histologically weaker than their effects on the AF and NP. These findings remind us that focusing on the sequence of IDD and which structure is more important while ignoring the integrity of the IVD contributes little to proposing a meaningful plan for treating IDD. Therefore, we recommend that subsequent researchers take the IVD as a whole and consider its surrounding tissue when they develop novel medicines for treating IDD. In every step of the experimental design and implementation, the AF, CEP and NP should be included. Conclusions This study revealed that the SOD-mimicking activity of CeNPs and the concentration of CeNPs are positively correlated. CeNPs at concentrations ≤ 10 µM show good biocompatibility with IVD cells. CeNPs can effectively alleviate the inhibition of proliferation and death of IVD cells caused by OS. IVD pretreated with CeNPs can effectively resist structural damage to the AF, calcification of the CEP, degradation of the NP and other structural changes in the IVD caused by OS. The above protective effects may be related to the downregulation of the expression of enzymes involved in the degradation of the ECM, the control of the inflammatory response and the restriction of apoptosis. Overall, this study suggests that locally applying CeNPs to IVDs might be a potentially promising strategy for alleviating OS-induced IDD. However, the high degree of heterogeneity between rodents and humans should not be neglected. Therefore, more relevant in vivo studies and pharmacokinetic evaluations in the human body must be carried out in the future. Taking CeNPs as an example, this study revealed that the same drug might have different effects on various IVD structures and types of IVD cells as well as the tissues around the IVD, indicating that incorporating as many of the IVD structures and IVD cells into the experimental design as possible could benefit drug safety for development of medicine against IDD. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1 Supplementary Material 2 Supplementary Material 2