Cytosolic escape of mitochondrial DNA triggers cGAS-STING-NLRP3 axis-dependent nucleus pulposus cell pyroptosis

The use of human medical records, imaging, and samples from human subjects was consistent with the Declaration of Helsinki. The Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology approved all experimental protocols related to human samples and animal experiments (No. S341) (No. S2394).

We included volunteers who underwent spine surgery because of idiopathic scoliosis or lumbar disc herniation without infection or tumor, immune or endocrine diseases in this study. In addition, we used the Pfirrmann MRI grading system to assess the degree of degeneration based on the patients’ symptoms, signs, and magnetic resonance images (MRI)^31,32. A total of 16 patients’ NP tissue samples were collected, 9 males and 7 females aged 12-58 years. We defined 4 tissues from volunteers as Grade I because of idiopathic scoliosis, while the other 12 tissues were defined as Grade II-IV because the participants had lumbar disc herniation in different segments. Further information on the participants is provided in Supplementary Table 1. All of the samples were separated into two parts, frozen using liquid nitrogen for protein expression analysis for western blotting and fixed with 4% formaldehyde buffer to assess the histological degree of degeneration. In particular, the NP tissues of volunteers with idiopathic scoliosis were regarded as healthy for human NP cell investigation.

The healthy samples described above were partly used to isolate human NP cells. Briefly, we used phosphate-buffered saline (PBS) to transport NP tissue, isolated cells, and plated the cells at 37 °C and 5% CO₂ in 1:1 DMEM/F12 in combination with 15% fetal bovine serum (FBS, Gibco, USA) and 1% penicillin/streptomycin (Servicebio). We used fluorescent-activated cell soring for NP cell marker determination (CD24, 311117; KRT18, 628404; Biolegend). The second passage cells were used for further in vitro experiment^33,34. NP cells were exposed to 100 μM or 200 μM tert-butyl-hydroperoxide (TBHP) in medium for 24 h in vitro, and vehicle treatment was used for the controls^14,35,36. Treatment with 1 μM H-151 (MedChemExpress) or 10 μM cyclosporine (CsA) (MedChemExpress) with 100 μM TBHP for 24 h was used to test the effects of STING activation inhibition and blockade of mPTP opening in NP cells under oxidative stress.

Frozen human tissues were homogenized, lysed, and extracted, and cells were lysed by RIPA (Beyotime) for 15 min and centrifuged to collect the proteins. Proteins (15–80 μg) were separated by 8–12% SDS-PAGE, transferred, blocked, washed, and incubated with specific primary antibodies and secondary antibodies. Information on the antibodies is presented in Supplementary Table. The reagents used for western blotting are presented in Supplementary Table. A ChemiDoc-It 610 Imaging System (UVP, USA) was used to detect and visualize the protein bands, and GAPDH was used for normalization. 2 3

We cut the tissue paraffin blocks into 4-μm slices and used hematoxylin and eosin (H&E) staining to evaluate the histological degree of degeneration of human samples. Furthermore, sections were labeled with the primary antibody at 4 °C overnight, washed, and probed with the secondary antibody at room temperature for 1 h for immunohistochemical analysis.

Formalin-fixed paraffin-embedded (FFPE) human NP sections as described above were stained with fluorescent multiplex immunohistochemistry for analysis of the expression levels of cGAS and STING^37,38. FFPE NP slides were deparaffinized by xylene, rehydrated by an ethanol gradient, and antigen retrieved by autoclaved Trilogy buffer (CellMarque) for 15 min. The sections were permeabilized by 0.5% Triton X-100 and blocked with 5% BSA. The sections were probed at 4 °C overnight with primary antibody, washed, and incubated for 1 h at 37 °C with the secondary antibody. Then, the sections were labeled with the tyramide-conjugated fluorophore for 10 min. Before the next primary antibody incubation, the sections were heated in 10 mM citric acid for 10 min to wash away the previous antibody. Antibodies paired with Opal tyramide-conjugated fluorophore were used as follows: cGAS and Opal 570, STING and Opal 520. The nuclei were labeled by 0.1 g/ml DAPI. Slices were captured under the Vectra 3.0 multispectral imaging system (PerkinElmer).

For STING knockdown, we used Lipofectamine 2000 (Invitrogen) to transfect NP cells for 72 h with 100 nM STING small-interfering RNA (si-STING, GenePharma) or the control and immediately exposed them to 100 μM TBHP. The siRNA sequences are presented in Supplementary Table. 4

Cellular RNA was collected by TRIzol reagent (Invitrogen), and RNA reverse transcription to cDNA was performed. General PCR was performed, followed by digital imaging of agar gelatin electrophoresis to visualize the DNA abundance. RT-qPCR was conducted to quantitatively analyze the DNA content. Expression levels were normalized against that of GAPDH. The set of deltaCq replicates for the control and tested samples normalized against the geometric means of the reference genes was used for statistical testing and estimation of the P values. The primers used for PCR or RT-qPCR are presented in Supplementary Table 4.

After induction of TBHP for the indicated time and dose, cytosolic mtDNA was analyzed by general PCR or RT-qPCR^18,23. Briefly, half of the cells were lysed by mild lysis buffer, and the other half was lysed by strong lysis buffer. Cells were cytoplasmically lysed with 0.1% NP-40 for 20 min on ice and centrifuged at 14,000 × g for 20 min at 4 °C. Cytosolic mtDNA from the supernatant cytosolic fraction and total mtDNA from the total lysis were purified by a TIANamp Genomic DNA Kit. General PCR followed by agar gelatin electrophoresis and RT-qPCR as described above were used for DNA analysis normalized by the mtDNA in the total lysate.

DCFH-DA (Beyotime) and MitoSOX (Yeasen Biotech) were used as markers to measure cytosolic and mitochondrial ROS. After the treatment, the samples were stained according to the protocols described. The fluorescence intensities of DCFH-DA were detected by flow cytometry, while those of MitoSOX were visualized by fluorescence microscopy.

MMP was measured by a JC-1 Assay Kit (Beyotime) according to the protocol described, detected by a flow cytometer, and analyzed by the ratio of JC-1 polymers to monomers. mPTP was detected by an mPTP Assay Kit (Beyotime). Briefly, the treated cells were washed with PBS and incubated with calcein AM plus Co²⁺ quencher at 37 °C for 30 min. Then, the dye was replaced by culture medium, and the slides were cultured at 37 °C for 30 min in the dark and observed by a fluorescence microscope.

Cultured NP cells were fixed, permeabilized, blocked, incubated with primary antibodies at 4 °C overnight, washed thoroughly, and labeled with secondary antibodies at 37 °C for 1 h protected from light. Information on the antibodies is presented in Supplementary Table. Nuclei and mitochondria were stained with 0.1 g/ml DAPI (Beyotime) and 50 mM MitoTracker Red CMXRos (Invitrogen). 2

Adherent cells were washed thoroughly after treatment and probed with 5 μg/ml Hoechst 33342 (Invitrogen) and 5 μg/ml propidium iodide (PI, Invitrogen) for 20 min at 37 °C protected from light. After three washes, the adherent cell slices were captured by fluorescence microscopy.

After the treatment, cell culture supernatants were collected, secretory IL-1β was detected by an ELISA Kit (Elabscience), and the absorbance at 450 nm was measured by SoftMax Pro version 5 Software (Molecular Devices).

A CASP-1 Assay Kit (Beyotime) was used to detect the enzymatic activity of CASP-1. After treatment, cellular proteins were extracted and coincubated with 2 mM acetyl-Tyr-Val-Ala-Asp p-nitroanilide (Ac-YVAD-pNA) at 37 °C for 24 h. The amount of p-nitroanilide (pNA) cleaved by CASP-1 was detected by a spectrophotometer at 450 nm, and the enzyme activity was calculated according to Chemicon’s unit normalized to the cellular total protein amount.

After treatment, the cells were lysed with moderate IP buffer (NP-40, Beyotime) with 1% protease inhibitors (Servicebio) and 1.5% phosphatase inhibitor (MedChemExpress) at 4 °C for 20 min and centrifuged at 14,000 × g for 15 min at 4 °C to separate the soluble fraction. The resultant supernatants were incubated with anti-dsDNA or IgG isotype control monoclonal antibody at 4 °C for 8 h. Then, 1 mg protein was incubated with 5 μg antibodies. Protein A/G magnetic beads (MedChemExpress) were coincubated with the antigen-antibody mixture for 2 h at 4 °C. The beads were washed and divided into two parts. Half of the beads were eluted with SDS loading buffer (Boster) and then boiled at 95 °C for 10 min for western blotting analysis, while the other half were eluted with chromatin immunoprecipitation (ChIP) elution buffer (5 mM Tris-HCl, 10 mM EDTA and 1% SDS). After proteinase K treatment, DNA was collected using a TIANamp Genomic DNA Kit (Tiangen), followed by general PCR analysis as described above^18,39.

After treatment, cells were fixed at room temperature for 15 min and washed twice with PBS. Then, the Duolink® In situ PLA ® kit (Sigma-Aldrich) for mouse/rabbit (with red detection) was used to show the physical interaction of dsDNA (double-stranded DNA) and cGAS according to the protocol.

After treatment, cultured cells were successively fixed with 2.5% glutaraldehyde (Sigma-Aldrich) for 1 h and 2% osmium tetraoxide for 2 h. After washes with double distilled water, the cell samples were stained with 0.5% uranyl acetate for 12 h, dehydrated, polymerized, and cut into 70-90 nm ultrathin sections. Images were observed and captured by a Tecnai G2 TWIN TEM (FEI, USA).

We used Sprague-Dawley rats (three months old, 200 ± 20 g) from the Laboratory Animal Center of Huazhong University of Science and Technology and housed the rats with a 12 h:12 h light:dark cycle at 21 °C. Pentobarbital (2%, 40 mg/kg) was used to anesthetize the rats (w/v). As previously described, the needle puncture-induced IVD degeneration model was established by using 29-gauge needle-to-percutaneous disc puncture on the Co8/9 and Co9/10 discs of each rats^32,40,41. The needle vertical to the tail was inserted into the disc, punctured into at least 5 mm depth of disc, rotated through 360°, and kept in the disc for 60 s to ensure the degenerative effects. To avoid individual differences, we defined the punctured Co8/9 disc of each rats as the degenerated + PBS group (n = 7), the punctured Co9/10 disc of the same rat as the degenerated + CsA group (n = 7) or degenerated + C-176 group (n = 7), and the nonpunctured Co7/8 disc of the same rat as the negative control group (n = 7). For evaluation of the effects of cytosolic mtDNA and the cGAS-STING axis on IVD degeneration and the therapeutic potential of CsA and C-176 administration in vivo, punctured Co9/10 discs were immediately intradiscally injected with CsA (100 μM) or C-176 (10 μM) every week for one month, while other punctured Co8/9 discs were injected with the same amount of PBS every week for one month. To eliminate the influences of the injected volume, we injected 2 μl of the solution of interest into the center of the NP⁴².

After one month of surgery, X-ray, micro-CT, and MRI analyses were performed on all rats before sacrifice. After anesthetization, the rats were kept in a supine position with their tail placed in a straight line. The X-ray, micro-CT, and MRI parameters are presented in Supplementary Table. We used NRecon to create three-dimensional reconstructions of micro-CT scans. The disc height index (DHI) and MRI degree of degeneration according to the Pfirrmann grade were determined as described previously. 5

After radiological examination, the rat samples were collected, fixed, decalcified, dehydrated, embedded in paraffin, and cut into 4 μm sections. H&E staining, safranin O-fast green staining, and Masson staining were conducted to assess the histological degree of degeneration of rat discs⁴³. Immunofluorescence staining was carried out to analyze the expression levels of cGAS, STING, and NLRP3.

Data are shown as the mean ± SEM. The number of experimental replicates and the level of significance are presented in the figure legends. GraphPad Prism software 8.0 was used to generate charts and perform statistical tests. We used Student’s t test and two-way ANOVA to determine significant differences between two groups and multiple groups. The P value was indicated by stars: n.s. no significance, *P < 0.05, **P < 0.01, ***P < 0.001.

More supplementary materials and methods are available in Supplementary Table. 6

To examine the roles of the cGAS-STING axis and NLRP3 inflammasome in IVD degenerative progression, we collected tissue samples from patients who underwent open spine surgery or lumbar discectomy surgery because of idiopathic scoliosis or lumbar disc herniation. According to the Pfirrmann MRI grading system, healthy tissues or disc tissues with different degrees of degeneration exhibited varied signal intensities by T2-weighted MRI. This result indicates a decrease in the water content of NP tissues with the progression of IVD degeneration³¹. Specifically, NP tissues with higher degrees of degeneration had atrophic volumes and decreased elasticity compared to those of tissues with lower degrees of degeneration. In addition, H&E staining showed that degenerated NP tissues exhibited clustered cellularity and chondroid nests with loss of the well-organized structure of the nucleus matrix, especially in Grade III and Grade IV tissues^43,44 (Fig. 1e). Western blot analysis indicated that the protein expression levels of cGAS, STING, and NLRP3 were higher in the degenerated NP tissues (Fig. 1a, b). Linear regression analysis of cGAS, STING, or NLRP3 protein expression levels and degenerative IVD grades revealed that higher cGAS, STING, and NLRP3 protein levels were associated with worsened IVD degeneration (Fig. 1c). We further analyzed the relationship between cGAS or STING and NLRP3 protein levels, and a significant positive correlation was observed (Fig. 1d). Immunohistochemistry (IHC) showed that the protein expression levels of NLRP3 increased with the progression of IVD degeneration (Fig. 1f). Tyramide signal amplification immunofluorescence (TSA-IF) further revealed that cGAS and STING were increased and colocalized in the cytoplasm of degenerative NP tissues (Fig. 1g). These results revealed that IVD degeneration was accompanied by the activation of the cGAS-STING signaling axis and NLRP3 inflammasome.

Previous studies have revealed that cellular oxidative stress plays a critical role in cell death and the microenvironmental imbalance in the progression of IVD degeneration^13,32. To further investigate the cellular response under oxidative stress, we cultured human NP cells with different concentrations of TBHP for 24 h. Here, the cytosolic and mitochondrial ROS levels were probed by DCFH-DA and MitoSOX. We observed an increase in intracellular and mitochondrial ROS in the NP cells cocultured with TBHP for 24 h (Fig. 2a, b).

To determine whether NP cells activated the cGAS-STING axis under oxidative stress, we performed western blotting and showed a dose-dependent increase in cGAS and STING protein expression in the NP cells after TBHP treatment (Fig. 2c, d). Once bound to the second messenger cGAMP, the STING homodimer will close, release the C-terminal tail and expose the polymerization interface, which contributes to the formation of polymers via the cysteine-148 residue disulfide linkage⁴⁵. Thus, immunofluorescence staining showed increased protein levels of STING with different degrees of polymerization and increased STING localization in the cytoplasm of NP cells after TBHP treatment (Fig. 2e), and quantitative analysis of the fluorescence intensity is shown in Supplementary Fig. 1a. Furthermore, to determine whether the interaction between dsDNA and cGAS increased in NP cells under oxidative stress, we immunoprecipitated DNA followed by western blotting to measure the coprecipitated proteins, and the results showed significant dose-dependent enrichment of cGAS in the TBHP-treated groups (Fig. 2f, g). In situ PLA was used to demonstrate the physical interaction between dsDNA and cGAS in NP cells, and PLA signals showed a significant dose-dependent increase after TBHP treatment (Fig. 2h). These data indicated that the cGAS-STING axis was activated in NP cells under oxidative stress.

Cellular stress-mediated NLRP3 inflammasome activation initiates pyroptotic cell death, triggering GSDMD-mediated transmembrane pore formation and the release of interleukin-1β (IL-1β) to form an inflammatory microenvironment^25,46. To determine whether NLRP3 inflammasome activation and pyroptosis were responsible for NP cell death under oxidative stress, we used western blotting to examine the expression levels of key pyroptosis molecules and observed that NLRP3, ASC, pro-CASP-1, and GSDMD levels were dose-dependently upregulated in the TBHP-treated NP cells (Fig. 2i, j). Next, we measured the downstream protein expression levels and found that cleaved CASP-1 and cleaved GSDMD levels were upregulated after TBHP treatment. The ratios of cleaved CASP-1 to pro-CASP-1 and cleaved GSDMD to GSDMD increased dose-dependently after TBHP treatment, quantitatively revealing the increased activation of the NLRP3 inflammasome under oxidative stress (Fig. 2k, l). The enzymatic activity of CASP-1 was determined by measuring the cleavage rate of Ac-YVAD-pNA. We observed a dose-dependent increase in CASP-1 enzymatic activity induced by TBHP (Fig. 2m). In addition, ELISA results indicated that the level of IL-1β in cell culture supernatants were elevated significantly, revealing that oxidative stress promoted inflammatory factor levels in the extracellular microenvironment (Fig. 2n). We used TEM to examine the ultrastructure of NP cells and found cellular swelling with large bubbles in the TBHP-treated NP cells, which is characteristic of pyroptosis induced by the N-terminus of GSDMD⁴⁷ (Fig. 2o). Cell viability was assessed with Hoechst and propidium iodide (PI) staining, which showed significant cell death after TBHP treatment (Fig. 2p). The above data revealed the involvement of NLRP3 inflammasome activation and pyroptosis in the NP cell death under TBHP-induced oxidative stress.

Previous studies have shown that the cGAS-STING axis increases NLRP3 inflammasome activation in response to pathogenic or damage-associated stress^48,49. To verify that oxidative stress-induced STING activation was responsible for NLRP3 inflammasome formation and NP cell pyroptotic death, we used a selective human STING antagonist (H-151) to pharmacologically block the activity of STING and treated NP cells with 100 μM TBHP for 24 h⁵⁰. Western blotting results showed that the expression levels of core proteins (NLRP3, ASC, pro-CASP-1, and GSDMD) and effector proteins (cleaved CASP-1 and cleaved GSDMD) of pyroptosis were significantly downregulated in the group treated with the STING antagonist (Fig. 3a, b). In addition, the cleaved CASP-1:pro-CASP-1 ratio, the cleaved GSDMD:GSDMD ratio, and the CASP-1 enzymatic activity were reduced, which indicated reduced activation of the NLRP3 inflammasome (Fig. 3c–e). In addition, the ELISA results indicated that the level of secreted IL-1β in NP cell culture supernatants decreased after STING inhibition (Fig. 3f). TEM showed that inhibition of STING could alleviate TBHP-induced cell swelling and bubble formation (Fig. 3g). The STING antagonist reduced the number of PI/Hoechst double-positive cells and blocked TBHP-induced NP cell death (Fig. 3h). Furthermore, we knocked down STING using small-interfering RNA in NP cells, which were then treated with 100 μM TBHP. We used agar gelatin electrophoresis for visualization and RT-qPCR to quantitatively verify the knockdown efficiency of siRNA against STING at the transcript level (Fig. 3i, k). In addition, western blotting and immunofluorescence showed that the protein expression level of STING was downregulated significantly and that STING knockdown was efficient (Fig. 3j, l, m and Supplementary Fig. 1b). STING knockdown reduced TBHP-induced NLRP3 inflammasome activation and blocked pyroptotic NP cell death (Fig. 3n–u). Collectively, the above data indicated that inhibition of STING could alleviate NLRP3 inflammasome-mediated pyroptosis in NP cells under oxidative stress.

Disturbed mitochondrial homeostasis is a characteristic of IVD degeneration, and mitochondrial damage intrinsically triggers mPTP opening and the subsequent leakage of DAMPs into the cytoplasm, regulating cell death events^33,51,52. Recent studies have shown increased mPTP opening, cytosolic escape of mtDNA and cGAS-STING axis activation under mitochondrial stress^18,21–24. To evaluate whether oxidative stress triggers mPTP opening in TBHP-treated NP cells, we performed fluorescence microscopy, confirming a reduction in the amount of calcein-AM retained in the mitochondria, indicating the opening of the mPTP (Fig. 4a, c). In addition, mitochondrial membrane potential (MMP) is maintained by the electrochemical gradient across two mitochondrial membranes, which is reduced by mPTP opening. JC-1 staining revealed a decreased ratio of JC-1 polymers to monomers after TBHP treatment, functionally indicating that oxidative stress enhanced mPTP opening in NP cells (Fig. 4b, d). TEM indicated swelling and vacuolation of mitochondria in the TBHP-exposed NP cells, which is the most prominent characteristic of mPTP opening (Fig. 4e). To further verify the leakage of mtDNA, cytosolic DNA was extracted, visualized by agar gelatin electrophoresis and quantitatively analyzed by RT-qPCR. The results indicated that the levels of ND1 and ND2, which are genes located in mtDNA, were higher in the cytosol of the NP cells exposed to TBHP, while there were no increases in the total levels^18,23 (Fig. 4f–i). In addition, the immunofluorescence assay indicated increased dsDNA fluorescence intensity outside of mitochondria and nuclei in the NP cells cocultured with TBHP (Fig. 4j). These data suggested that oxidative stress triggered mPTP opening and the excessive accumulation of cytosolic mtDNA, which might contribute to the activation of cGAS-STING-NLRP3 axis-dependent NP cell pyroptosis.

To further confirm that mPTP opening and mtDNA escape into the cytosol were essential for cGAS-STING-NLRP3 axis activation and NP cell pyroptosis, we treated NP cells with cyclosporine (CsA), a pharmacological inhibitor of mPTP opening, which significantly abrogated TBHP-induced mPTP opening and decreased MMP, morphological mitochondrial damage and cytosolic mtDNA accumulation (Fig. 5a–f). In addition, CsA decreased the protein expression levels of cGAS and STING, the formation of STING polymers, the enrichment of cGAS in DNA immunoprecipitates, and PLA signals between dsDNA and cGAS (Fig. 5g–l and Supplementary Fig. 1c). Moreover, TBHP-induced NLRP3 inflammasome activation and NP cell pyroptosis were inhibited by CsA (Fig. 5m–t). These data suggested that oxidative stress-induced mPTP opening and mtDNA leakage were responsible for cGAS-STING-NLRP3 axis activation and NP cell pyroptosis.

To further evaluate the therapeutic efficacy of pharmacological inhibition of mPTP opening and STING activation in vivo, we established a disc percutaneous needle puncture animal model of IVD degeneration using Sprague-Dawley rats. After one month, X-ray and micro-CT analyses confirmed that the disc heights in the needle puncture group that was treated with PBS collapsed more than those in the group that was treated with CsA or C-176 (a selective rat STING antagonist) (Figs. 6a, b, 7a, b). MRI examination showed that the signal intensities of the IVD after PBS injection were not homogeneous and that the T2-weighted signals were lower than those in the group that received CsA or STING antagonist injection (Figs. 6a, c, 7a, c). H&E staining, safranin O-fast green staining and Masson staining showed a loose well-organized NP tissue structure with star-shaped cells, increased tissue fibrillation in the NP area, and no distinguishable boundary between NP and AF tissue. Conversely, CsA or STING antagonist administration reduced IVD histological degeneration compared with PBS injection (Figs. 6e, 7e). In addition, the histological scores in the IVD degeneration with needle puncture group were significantly increased compared with those of the CsA or STING antagonist injection group (Figs. 6d, 7d). Immunofluorescence staining showed that CsA reduced the fluorescence intensities of cGAS, STING and NLRP3, and the STING antagonist decreased the fluorescence intensity of NLRP3 but not cGAS or STING compared to PBS injection (Figs. 6f–h, 7f–h). These data demonstrated that pharmacological inhibition of mPTP opening and STING activation could ameliorate damage in the needle puncture rat model of IVD degeneration.

The datasets generated during the current study are not publicly available due to the data also forming part of our ongoing study but are available from the corresponding author on reasonable request.