Activating SIRT6 may improve aging spine disc health

To investigate whether SIRT6 activation improves disc health in the context of age‐dependent degeneration, we examined the spinal phenotype of B6 mice treated with MDL‐800 for 6 months starting at 18 months of age (Figure 1A). MDL‐800 treatment significantly decreased H3K9 acetylation levels in the disc tissues, confirming increased SIRT6 activity (Figure 1B). When disc degeneration was assessed histologically in the lumbar spine, treatment with MDL‐800 lead to a marked decline in the age‐dependent degenerative features, including NP fibrosis, loss of NP‐AF compartment demarcation, lamellar defects, and clefts indicative of structural disruptions (Figure 1C). Notably, NP cells showed a healthy morphology as evident from preservation of vacuolated cells in the cell band. Modified Thompson grading of MDL‐800‐treated mice showed significantly lower cumulative degeneration scores across the lumbar spine in both the NP and AF compartments compared to vehicle‐treated controls (Figure 1C–E). Notably, level‐by‐level analysis of grading data suggested that L3/4 and L4/5 discs were more responsive to the MDL‐800 treatment than L5/6 and L6/S1 discs, implying that outcomes were likely driven by location‐specific biomechanics (Figure S1). Since cell death is observed during disc degeneration, we performed TUNEL labeling, which showed a significant decrease in TUNEL‐positive cells in MDL‐800‐treated mice, but a comparable total cell number, indicating that the elevated apoptosis was not the primary driver of later stages of degeneration (Figure 1F,G). Despite an overall improvement in disc health parameters, MDL‐800‐treated mice did not exhibit altered disc height (DH) or disc height index (DHI) when compared to vehicle‐treated controls (Figure 1H–K).

We then performed plasma metabolic analyses to determine the systemic effects of long‐term MDL‐800 treatment in old mice using LC/MS (Figure) and NMR (Figuresand) assays. LC/MS analysis identified 184 plasma metabolites with high confidence but did not reveal any treatment‐dependent PCA or hierarchical clustering (Figure). Out of 6 metabolites that showed significant differences, four metabolites, including palmitoyl sphingomyelin, 5‐aminolevulinic acid, 2‐oxindole, and jasmonic acid, showed a significant increase, while nicotinamide and hypotaurine showed a decrease in MDL‐treated mice (Figure). Plasma NMR analysis showed no changes in any of the measured analytes, including indicators of inflammation and metabolic regulation: GlycA, total ketone bodies, acetoacetate, acetone, beta hydroxybutyrate total branched chained amino acids (BCAA), alanine, leucine, isoleucine, valine, citrate, ApoA‐1, ApoB, total triglyceride, total cholesterol, total TRLP, small TRLP, total calibrated low‐density lipoprotein particle (cLDLP), total calibrated high‐density lipoprotein particle (cHDLP), medium LDLP, H4P, TRL size, LDLC, HDLC, HDL size, TRL TG (Figuresand). In conclusion, these plasma analyses revealed minimal metabolic effects of long‐term MDL‐800 treatment in B6 mice, supporting a favorable safety profile under the conditions tested. S2 S3 S4 S2B,C S2D–I S3B–M S4

Asactivation improved histological scores of disc health, with a notable effect oncell morphology, we determined whether‐800 treatment preserved cell phenotype and matrix constituents using Picrosirius red staining and quantitative immunohistochemistry. Polarized imaging of picrosirius red‐stained disc sections showed diminished presence of collagen fibers in thecompartment and an increased abundance of small‐diameter fibers and a decrease in thick fibers in theof MDL‐800‐treated mice (Figure). These findings suggest an overall increase in collagen turnover and amelioration of age‐associatedfibrosis resulting from sustainedactivation. To determine the integrity of the disc, we stained the disc sections for Aggrecan (), a predominant constituent of thematrix, and noted a significant increase in abundance (Figure). Likewise, NP cells showed increased GLUT1 abundance, a known phenotypic marker, indicating preservation of NP tissue health (Figure). There were little changes in the abundance of FCHP, a marker of denatured collagen, and COL1 in the AF (Figure). Similarly, both groups exhibited comparable COLX abundance in the NP (Figure), suggesting thatactivation did not modulate the age‐dependent acquisition of chondrocyte‐like characteristics. These findings suggested thatis critical in maintaining a healthy cell phenotype and disc tissue matrix. SIRT6 NP MDL NP AF NP SIRT6 ECM ACAN NP SIRT6 SIRT6 S5A–D S5E,F S5G,H S6A,B S6E

The senescence burden and SASP have been causally linked to age‐dependent disc degeneration (Novais et al. 2021). We therefore investigated the senescence status of MDL‐800‐treated mice. We noted that levels of p19, an upstream regulator of senescence, and p21, a CDK inhibitor important for senescence initiation, were significantly decreased in the discs of MDL‐800‐treated mice (Figure 2A–D) (Stein et al. 1999). Additionally, we determined levels of IL‐6 (Figure 2E,F), TGF‐β (Figure 2G,H), MMP13 (Figure S7A,B), CCL2 (Figure S7C,D), important markers of SASP in the disc (Novais et al. 2021), which were lower in discs of MDL‐800‐treated mice, suggesting an attenuation of the overall burden of SASP and profibrotic factors.

To determine whether the MDL‐800‐mediated decrease in SASP factors observed in mice is reproduced in humans, we conducted a pilot study using NP cells isolated from Grade IV and V degenerated discs from four patients (Figure 2I). The baseline senescence in these cells was high, as evidenced by significant β‐galactosidase positivity. When cells were subjected to a short proinflammatory stimulus of IL‐1β and IL‐6, we observed increases in IL6 (Figure 2J), TGFβ (Figure 2K), MMP13 (Figure S7E), and CCL2 (Figure S7F) but not in the p16 transcript (Figure S7G). Notably, MDL‐800 co‐treatment with stimulus significantly reduced expression of all these markers, with a strong trend toward decreasing p16 levels.

To determine the broader effects of long‐term, sustained SIRT6 activation on the transcriptomic landscape in NP and AF tissues, we performed RNA sequencing. While SIRT6 suppresses gene expression via limiting chromatin accessibility, long‐term effects were expected to reflect second‐order adaptive changes. Indeed, we observed that more genes were upregulated than downregulated in both NP and AF tissues by MDL‐800 treatment (Figure S8). In NP tissue, there were major upregulated thematic clusters related to (1) Histone acetylation/activity, and JNK/MAPK signaling (2) a large supercluster for circadian clock, cell polarity/notochord, follistatin‐inhibin‐activin signaling, connexins/Cx43, and matrix remodeling (3) lipid metabolism and PPAR signaling, and (4) microtubules anchoring at centrosome (Figure 3). Notably, when downregulated DEGs were analyzed, we observed a less prominent thematic clustering with major clustering centered around (1) follistatin‐myostatin and actin‐myosin, (2) connexins/Cx43 and Natriuretic peptide/myocardium (Figure S9). The upregulated DEGs in AF clustered into two major superclusters. The first cluster was related to skeletal tissues, with themes including articular cartilage development, endochondral processes, and collagen/MMPs. While the second supercluster was related to neuronal processes, it encompassed themes such as the hypothalamic nucleus, synaptic vesicle exocytosis, axonal guidance, sensory neurons, synapses, clathrin‐coated vesicles, and voltage‐gated Ca²⁺ channels (Figure S10). Again, compared with the extensive clustering observed in upregulated DEGs in AF, less clustering was observed in downregulated DEGs. Some of the enriched themes in these downregulated DEGs included immune cell function (CD8 effector T cells, granzyme secretion, MCP‐1 Production, PD‐1 secretion, macrophage response), caspase activation, myogenesis, Schwann cells, and oligodendrocytes and myelination (Figure S11).

Immune cells play a crucial role in regulating aging phenotypes (Delgado‐Pulido et al. 2025). To investigate the immune cell profile and its association with disc phenotype in MDL‐800 treated mice, we performed single‐cell RNA sequencing (scRNA‐seq) of two major lymphatic tissues, the spleen and bone marrow (Figure 4, Figures S12 and S13). After normalization, the splenocytes (Figure 4) and bone marrow cells (Figure S12) were grouped into 12 and 15 clusters, respectively, using the FindNeighbors and FindClusters functions in the Seurat single‐cell analysis package. Moreover, the data was visualized according to treatment, and cell populations were identified using the SingleR annotation tool, demonstrating differences in the proportion of B cells, T cells, and granulocytes, and a small change in monocytes or macrophages in the MDL‐800 treated group in lymphoid tissues (Figure 4B–E, Figure S12B–E). Previous studies have implied a relationship between immune cell heterogeneity with functional state and origin (Papalexi and Satija 2018; Sebastian et al. 2022). To further visualize cellular heterogeneity, T cell populations were subdivided into CD4⁺, CD8a⁺, and CD4⁻, CD8a⁻ T cells (Figure S13). We observed an overall decrease in T cell numbers in the MDL‐800‐treated mice (Figure S13A–D,G–J). When cell proportions were compared, there was a relative increase in the proportion of CD4⁺ and a decreased proportion of CD8a⁺, and CD4⁻, CD8a⁻ T cells in the spleen (Figure S13D′). However, there was no appreciable change in the proportion of T cell subpopulations in bone marrow (Figure S13J′).

Gene set enrichment analysis (GSEA) was then performed on DEGs (p < 0.05) for cell populations that exhibited significant differences in proportion using the clusterProfiler package (Yu et al. 2012). In the MDL‐800 group, DEGs in splenic B cells showed activation pathways related to MHC Class I processing, plasma membrane protein folding, ribonucleoprotein complex and ribosome biogenesis, and small molecule metabolic process (Figure 4F,F′). In B cells, GSEA identified suppression of the response to corticosterone, B cell proliferation in immune response, plasma membrane L‐glutamate import, and regulation of humoral immune response, with top DEGs including Lifr, St7, Ffar2, Hspa1a/1b, responsible for this enrichment (Figure 4F,F′). In the MDL‐800 group, DEGs from splenic T cells showed activation of defense response–related pathways, with top DEGs including Tnfrsf9/18, Itgae, Il1r2, and Tshz2 (Figure 4G,G′). T cells also exhibited DEGs linked to the suppression of multiple pathways related to protein translation, ribosomes, and rRNA, with top DEGs that included Dapl1, Nedd4l, Prss12, and Ptgs2 (Figure 4G,G′). GSEA of granulocytes showed that MDL‐800 treatment led to activation of the reproductive hormone response and nucleoside metabolism, suppression of pathways related to translation, ischemic response, and an overall decrease in B‐ and T‐cell activation and differentiation (Figure 4H,H′). There was no appreciable enrichment of DEGs in GSEA pathways in bone marrow immune cells in MDL‐800 treated mice.

To investigate whether MDL‐800 dependent changes in immune cell profiles in primary lymphatic tissues were correlated with systemic inflammation status, we measured plasma levels of key cytokines and chemokines. We did not observe any significant alterations in levels of IFNγ, IL‐1β, IL‐2, IL‐4, IL‐5, IL‐6, IL‐10, IL‐12/p70, IL‐15, IL‐17A/F, IL‐27/p28/IL‐30, IL‐33, IP‐10, KC/GRO, MCP‐1, MIP‐1α, MIP‐2, and TNF‐α (Figure), suggesting an absence of systemic inflammation. S14A–R

To delineate mechanistic insights into how MDL‐800 treatment and subsequent SIRT6 activation affect disc cells, we performed in vitro experiments using primary rat NP cells (Figure 5A). Western blot analysis showed a significant decrease in H3K9ac, a known marker of SIRT6 activation, without changes in SIRT6 levels in MDL‐800 treated NP cells (Figure 5B,C). This finding is further supported by immunofluorescence staining, which shows decreased nuclear H3K9ac staining and comparable levels of H3K27ac in MDL‐800‐treated cells (Figure 5D).

To understand the broader consequences of SIRT6 activation, we measured various H3K modifications using ELISA. We again saw a significant decrease in H3K9 acetylation levels in MDL‐treated NP cells (Figure 5E). Moreover, in line with our earlier studies, acetylation status of H3K18 and H3K56 was unaffected (Figure 5E). MDL‐800 treatment also decreased methylation of H3K36me3 and H3K79me2 in NP cells without changes in H3K4me, H3K9me, H3K27me, H3K36me1, H3K36me2, H3K79me1 and H3K79me3 levels (Figure 5E, Figure S15), underscoring tissue‐type specificity in epigenetic changes following elevated SIRT6 action.

SIRT6 modulates autophagy in various tissues, including disc cells (Cheng et al. 2023; Ramteke et al. 2025), and plays a crucial role in maintaining disc health (Madhu et al. 2021). To investigate whether SIRT6 activation induces autophagy, we measured autophagic flux in the presence of MDL‐800 and Bafilomycin A1, a known inhibitor of autophagic flux. Treatment of cells with MDL‐800 resulted in a significant increase in LC3II levels and the LC3II/GAPDH ratio, which showed a trend of further increase in the presence of BafA1, suggesting that MDL promoted autophagic flux (Figure 6A,B). Furthermore, we observed a significant increase in the number of LC3‐positive autophagic puncta in MDL‐800‐treated NP cells (Figure 6C,D). Since SIRT6 is known to modulate glycolytic metabolism, we determined the effect of MDL treatment on the bioenergetics of NP cells using Seahorse assays (Sebastián et al. 2012). MDL‐800‐treated cells showed ATP production rates comparable to those of vehicle‐treated control cells, suggesting no short‐term effect on cellular bioenergetics (Figure 6E–G).

We performed RNA‐seq and ATAC‐seq to understand the direct consequence of SIRT6 activation on the transcriptomic landscape of MDL‐800‐treated NP cells. CompBio analysis of DEGs from RNAseq (Log₂FC ≥ 1.5, FDR ≤ 0.05) revealed upregulation of three major interconnected super thematic clusters related to (i) actin‐myosin cytoskeleton and cell junctions, (ii) skeletal tissues, TGFβ/actin/BMP signaling, and extracellular matrix, (iii) glycolysis, purinergic signaling, and aryl hydrocarbon receptor activity (Figure S16). The downregulated clusters in MDL‐treated cells included two superclusters with themes related to: (1) fatty acid β‐oxidation, PPAR signaling, mevalonate pathway, dioxygenases and diacylglycerol formation, and (2) Hsp70/Hsc70 chaperons, Jagged1‐Notch1 signaling, Cx36 gap junctions, and primary cilia (Figure S17). These processes have been correlated with disc health (Hiyama et al. 2011; Zhang and Lin 2008). We then used an advanced feature of the CompBio tool, which identifies significantly affected shared pathways between two datasets, to further identify the pathways commonly affected in NP tissue from MDL‐800 treated mice and in primary NP cells in vitro, regardless of the duration of SIRT6 activation. A significant overlap in upregulated themes was noted between the two datasets for ECM proteins, collagens, ECM turnover, MMPs, myofibroblasts, gap junctions, cell–cell and cell‐matrix adhesion, actin cytoskeleton, microtubules, activin/BMP/FGF, and eNOS signaling (Figure S18A). Within downregulated datasets, themes related to ECM, collagens and fibronectin, chemokine and cytokines (interferon, IL12), coagulation and clotting (fibrinogen, fibrinolysis, and plasmin) were shared (Figure S18B).

Next, using ATAC‐seq data, we identified chromatin accessibility peaks and performed differential chromatin accessibility analysis between the groups. Chromatin accessibility signal profiles for each sample were visualized as heatmaps, and the average relative distribution of ATAC‐seq peaks assigned to genomic features, including promoters, introns, exons, and distal regions, in vehicle control and MDL‐800‐treated NP cells was computed (Figure S19A,B). We identified 251 upregulated (enriched ATAC peaks) and 21 downregulated (diminished ATAC peaks) shared genes between ATAC‐ and RNA‐seq datasets (FC ≥ 2, FDR < 0.05), which are shown in a heatmap (top genes sorted based on ATAC‐seq fold‐changes) and a quadrant plot (Figure S19C,D), demonstrating a strong correlation between chromatin accessibility and gene expression. We performed CompBio analysis on these shared genes and identified enriched and highly connected themes in upregulated genes related to HIF‐1 signaling, Glut1‐dependent glucose import, glycolytic metabolism, actin cytoskeletal organization, chondrocyte morphogenesis‐aggrecan, ECM‐collagens, and MAPK signaling, suggesting favorable transcriptomic outcomes. On the other hand, downregulated genes showed enrichment in themes related to multiple aspects of lipid biology (Figure S20). Additionally, a set of seven genes Serpine1, Evpl, Kcnh2, Gadd45g, Dlk2, Csrnp1, and Ptk7 were identified as consistently upregulated by MDL‐800 treatment in both in vivo and in vitro NP RNA‐seq datasets and shared with ATAC‐seq data. Serpine1 and Evpl are shown as average chromatin accessibility maps (Figure S19E). No genes were consistently downregulated across all three datasets.

While there were no discernable differences in EP morphology between treatment groups (Figure 1c), in the MDL‐800 treated cohort, we noted a slight decrease in vertebral trabecular bone structural parameters (Figure S21A), including BV/TV, trabecular number (Tb.N), and thickness (Tb. Th), without changes in trabecular spacing (Figure S21B–E), suggesting a minimal effect on bone turnover. MDL‐800 treated mice also exhibited a slight decrease in cortical bone parameters, including % closed porosity, cortical thickness (Cs. Th); again, notably, without changes in bone perimeter (B.Pm) (Figure S21F–I). Notably higher lumbar bone mineral density, both integral and trabecular, has been correlated with an increased risk of disc height loss and diminished disc area in humans (Harada et al. 1998; Kaiser et al. 2018), suggesting that a small decrease in bone structural parameters could be protective.

To assess if MDL‐800 treatment caused any behavioral changes, mechanical (Von Frey filament) and temperature (Hargreaves thermal; cold plantar assay) sensitization were measured in MDL‐800 and vehicle‐treated mice (Figure S22A) (Novais et al. 2025). While MDL‐800 treatment did not alter the threshold to mechanical stimulation, we observed decreased latencies to cold and hot stimuli compared with vehicle controls (Figure S22B–E), which may arise from the drug's action on peripheral nerves independent of its action on the disc. The open field test, however, showed comparable performance between cohorts, suggesting no adverse effect on overall ambulation in MDL‐800‐treated mice (Figure S22F). Likewise, in both treatment cohorts, we observed a similar response to the grip strength test, an important metric used to assess frailty in humans and to measure axial pain in mice (Figure S22G) (Dudzińska‐Griszek et al. 2017; Millecamps et al. 2012). To further study the locomotion‐related parameters, gait analysis was performed. We did not observe changes in swing‐, stance‐, and stride durations, as well as stride length and frequency, suggesting the absence of impaired mobility or gait asymmetry (Figure S22H–K). Accordingly, while these measurements suggested a small effect on vertebral structural parameters and mild hyperalgesia following long‐term MDL‐800 treatment, the effect sizes were too small to cause vertebral fractures, and overall motor function and animal behavior remained unaffected, raising no safety concerns.

Animal studies were approved by the Institutional Animal Care and Use Committee of Thomas Jefferson University following guidelines from the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Eighteen‐month‐old male C57BL/6 mice (NIA colony, Charles River) were housed in groups of 4 per cage and had access to water and food ad libitum. Mice were randomly divided into two groups, and except for 2 doses/week in the first month, received weekly intraperitoneal injections of either DMSO (vehicle control group) or 75 mg/kg of MDL‐800 (SIRT6 activator) in DMSO for 6 months, that is, up to 24 months of age, n = 13 control, 15 MDL‐800 mice. We chose to use male mice based on our recent study, which showed enhanced disc degeneration in male C57BL/6J mice following SIRT6 deletion, as well as previous studies demonstrating larger effect sizes on lifespan and protective effects against age‐associated changes in male mice (Ramteke et al. 2025; Roichman et al. 2021).

Mice were acclimated to the behavioral testing room for 1 h before all behavioral tests. Forelimb grip strength to measure the frailty and axial pain, open field test to assess the general locomotion, Von Frey test for mechanical allodynia, Hargreaves thermal and cold plantar assays to measure hyperalgesia were performed (n = 7 vehicle control, n = 9 MDL mice) (Novais et al. 2025; Ottone et al. 2025). Gate analysis was performed to understand any changes in the ambulation (Novais et al. 2025) (see Supporting Information Methods for details in Data S1).

Spines were dissected and fixed in 4% paraformaldehyde (PFA) for either 6 h or 48 h, followed by decalcification in 20% EDTA at 4°C before embedding in OCT or paraffin for sectioning. 7 μm midcoronal sections from four lumbar levels (L3‐S1) of each mouse (n = 7 vehicle control, n = 9 MDL mice; 2–4 discs/mouse) were stained with Safranin‐O/Fast Green/Hematoxylin for histological assessment using a modified Thompson grading scale for NP and AF by at least three blind observers. Similarly, Picrosirius Red staining was performed to characterize collagen fibers in tissue sections. Safranin‐O staining was visualized using an Axio Imager 2 microscope (Carl Zeiss, Germany) using 5×/0.15 N‐Achroplan or 20×/0.5 EC Plan‐Neofluar objectives and Zen2TM software (Carl Zeiss). The heterogeneity of collagen organization was evaluated using a polarizing light microscope, Eclipse LV100 POL (Nikon, Tokyo, Japan), equipped with a 10×/0.25 Pol/WD 7.0 objective and a DS‐Fi2 camera. Images were analyzed using NIS Elements AR 4.50.00 (Nikon, Tokyo, Japan). Under polarized light, as the thickness of collagen fibers increases, their birefringence color changes from green to yellow to orange to red, that is, from shorter to longer wavelengths (Dayan et al. 1989; Montes et al. 1989), allowing quantitative assessment of compositional information in tissue sections. Color threshold levels were maintained constant between all analyzed images.

Deparaffinized sections following antigen retrieval or frozen sections were blocked in 5% normal serum in PBS‐T and incubated with antibodies against H3K9Ac (1:50, Sigma, 06‐942), Collagen I (1:100, Abcam ab34710), Collagen X (1:500, Abcam ab58632), Aggrecan (1:300, Abcam ab11570); GLUT1 (1:200, Abcam, ab40084), F‐CHP (1:100, 3‐Helix), p21 (1:200, Novus NB100‐1941), IL‐6 (1:50, Novus NB600‐1131), TGF‐β (1:200, Abcam; ab92486) (n = 4–6 mice/group, 2–3 lumbar discs/mouse were analyzed). For mouse antibodies, a MOM kit (Vector Laboratories, BMK‐2202) was used for blocking and primary antibody incubation. Tissue sections were washed and incubated with species‐appropriate Alexa Fluor‐594 conjugated secondary antibodies (Jackson ImmunoResearch, 1:700). TUNEL staining was performed using the In situ cell death detection kit (Roche Diagnostics). Briefly, sections were deparaffinized and permeabilized using Proteinase K (20 μg/mL), and the TUNEL assay was performed according to the manufacturer's protocol. The sections were mounted with ProLong Gold Antifade Mountant with DAPI (Fisher Scientific, P36934↗), visualized using an Axio Imager 2 microscope with 5×/0.15 N‐Achroplan or 20×/0.5 EC Plan‐Neofluar objectives, and images were captured with an Axiocam MRm monochromatic camera (Carl Zeiss) and Zen2 software (Carl Zeiss AG, Germany). Both caudal and lumbar discs were used for the analysis. Staining area and cell number quantification were performed using the ImageJ software, v1.53e, (http://rsb.info.nih.gov/ij/↗).

Human cervical and lumbar NP cells isolated from discs with Pfirrmann scores of IV and V were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% FBS and antibiotic‐antimycotic solution and grown at 37°C under 1% O₂. Cells in passage 2 were divided into three treatment groups. Vehicle (0.1% DMSO), stimulus only: IL‐ β (10 ng/mL) and IL‐6 (70 ng/mL), and MDL‐800 (25 μM, MedKoo #565267) with stimulus for a period of 24 h (Novais et al. 2026). Following treatments, total RNA was isolated as previously described (Novais et al. 2026), and qRT‐PCR was performed using SYBR Green chemistry. Target gene expression was normalized to Hprt1.

NP and AF tissues were dissected from vehicle control and MDL‐800‐treated mice (n = 5 mice/group). Samples were homogenized, and DNA‐free, total RNA was extracted using the RNeasy Mini kit (Qiagen). RNA with RIN > 7 was used for RNA‐sequencing analysis. Significantly up‐ and downregulated DEGs (FC > 1.5, p < 0.05) were analyzed using the GTAC‐CompBio Analysis Tool (https://gtac‐compbio‐ex.wustl.edu↗). CompBio, a multiomic analysis platform built on an episodic and semantic memory‐based intelligence engine, is used to analyze DEGs. This engine stores PubMed abstracts and full‐text articles as memories, referencing the input DEGs to identify relevant processes and pathways. Conditional probability analysis is used to calculate the statistical enrichment score for biological concepts (processes/pathways) relative to those expected by random sampling. The scores are then normalized for significance empirically over a large, randomized query group. Related concepts derived from DEGs are grouped into broader, higher level themes, which represent overarching biological patterns, such as pathways, processes, cell types, and structures as described previously (Barve et al. 2025; Madhu et al. 2025).

Spleens were collected from vehicle control and MDL‐treated mice. Splenocytes were dissociated from the tissue matrix by gentle compression through a 100‐μm strainer (StemCell Technologies, 27217). Similarly, bone marrow cells were isolated by gently flushing the femur and humerus. Thereafter, RBC lysis (BioLegend, 420302) was performed, followed by two PBS + 3% FBS washes. ScRNA‐sequencing was performed on freshly isolated cells using the standard 10X protocol (see Supplemental Methods in Data S1). Raw sequencing data were aligned to the mouse reference transcriptome (mm10‐2020‐A) and processed with Cell Ranger Multi version 7.0.1. The count matrix was analyzed using Seurat version 4.3.0. Briefly, data were converted into Seurat Objects using the CreateSeuratObject function, and all cells were analyzed for their unique molecular identifier and mitochondrial gene fractions; cells were included if 200 < nfeat < 2500, and mitochondrial genes accounted for < 5%. Data were integrated with logarithmic normalization (LogNormalize) and organized according to variation in the data using FindVariableFeatures. Dimensional reduction was performed using principal components analysis, and the FindNeighbors (dims = 1:20) function, followed by the FindClusters function, was used to calculate the number of cell clusters using the Louvain algorithm, with the resolution parameter set to 0.5. DimPlot was then used to visualize the data as UMAP. Cell types for each condition were determined using SingleR, with the MouseRNAseqData reference dataset and main labels. The clusterProfiler package was used to conduct GSEA on DEGs (FDR < 0.05) identified in various tissue and treatment‐based comparisons.

Blood was collected by intracardiac puncture in heparin‐coated syringes and centrifuged at 1500 relative centrifugal force (rcf) at 4°C for 15 min to isolate the plasma (n = 10 mice/group). Samples were stored at −80°C until analysis. According to the manufacturer's specifications, levels of pro‐inflammatory proteins, cytokines, and TH17 mediators were analyzed using the V‐PLEX Mouse Cytokine 29‐Plex Kit (Meso Scale Diagnostics, Rockville, MD).

Plasma LC/MS analysis was performed as described previously (n = 8 vehicle, 7 MDL mice) (Li et al. 2020). Briefly, polar metabolites were extracted from 50 μL plasma samples with 500 μL ice‐cold 80% methanol, and deproteinated supernatants were stored at −80°C prior to analysis. A quality control (QC) sample was generated by pooling equal volumes of all samples after extraction. LC–MS analysis was performed on a Thermo Scientific Q‐Exactive HF‐X mass spectrometer with HESI II probe and Vanquish Horizon UHPLC system. Hydrophilic interaction liquid chromatography (HILIC) was performed at a flow rate of 0.2 mL/min on a ZIC‐pHILIC column (150 × 2.1 mm, 5 μm particle size, EMD Millipore) with a ZIC‐pHILIC guard column (20 × 2.1 mm, EMD Millipore) at 45°C. Solvent A was 20 mM ammonium carbonate, 0.1% ammonium hydroxide, pH 9.2, and solvent B was acetonitrile. The gradient was 85% B for 2 min, 85% B to 20% B over 15 min, 20% B to 85% B over 0.1 min, and 85% B for 8.9 min. The autosampler was held at 4°C. For each analysis, 4 μL of the sample was injected. The following parameters were used for the MS analysis: sheath gas flow rate, 40; auxiliary gas flow rate, 10; sweep gas flow rate, 2; auxiliary gas heater temperature, 350°C; spray voltage, 3.5 kV for positive mode and 3.2 kV for negative mode; capillary temperature, 325°C; and funnel RF level, 40. All samples were analyzed by full MS with polarity switching. The QC sample was analyzed at the start of the sample sequence and after every 8–14 samples. The QC sample was also analyzed by data‐dependent MS/MS with separate runs for positive and negative ion modes. Full MS scans were acquired at 120,000 resolution with an automatic gain control (AGC) target of 1e6, maximum injection time (IT) of 100 ms, and scan range of 65–975 m/z. Data‐dependent MS/MS scans were acquired for the top 10 highest intensity ions at 15,000 resolution with an AGC target of 5e4, maximum IT of 50 ms, isolation width of 1.0 m/z, and stepped normalized collision energy (NCE) of 20, 40, 60. Data analysis was performed using Compound Discoverer 3.1 (ThermoFisher Scientific) with separate analyses for positive and negative polarities. Results were manually processed to remove entries with apparent peak misintegrations and correct commonly misannotated metabolites. Positive and negative data sets of identified compounds were merged, and the preferred polarity was selected for compounds identified in both polarities. Compound quantifications were normalized per volume of plasma injected, which was equivalent for all samples. Values from the study were further normalized to the summed area of identified metabolites in each sample. For compounds identified multiple times at different retention times, a single entry was selected with priority given to standards database matches, followed by greater mzCloud match factors and peak areas.

Additionally, compounds shown in Figures S2 and S3 were measured from plasma samples (n = 8 vehicle control, 7 MDL mice) at Labcorp (Labcorp, Morrisville, NC). NMR spectra were acquired on a 400 MHz NMR instrument, the Vantera Clinical Analyzer, from EDTA plasma samples as described for the NMR LipoProfile test. The data were transferred and processed using the LP4 lipoprotein profile deconvolution algorithm, as reported in the supplemental methods (Jeyarajah et al. 2006; Matyus et al. 2014).

Micro‐CT (μCT) scanning (Bruker SkyScan 1275) was performed on fixed spines using parameters of 50 kV (voltage) and 200 μA (current) at 15 μm resolution as previously described (n = 8 vehicle control, 12 MDL mice, 4 lumbar vertebrae/mouse) (Novais et al. 2021; Ramteke et al. 2025). Images were reconstructed using the nRecon program (Version: 1.7.1.0, Bruker) and analysis was performed using CTan (version 1.17.7.2, Bruker). Disc height and vertebral length were measured at three different points equidistant from the center of the bone on the sagittal plane and used to calculate the DHI.

Primary NP cells from adult Sprague Dawley rats (3–6‐month‐old, Charles River) were isolated and cultured in antibiotic‐supplemented complete DMEM containing 5 mM glucose and 10% FBS. To explore the role of SIRT6 in vitro, cells were treated with either DMSO (1:2000) or MDL‐800 (25 μM in DMSO) in 1% hypoxia for 24–96 h.

In brief, NP cells were plated in a 24‐well Seahorse V7‐PS test plate, treated with MDL‐800 or vehicle for 4 days, and placed in hypoxia (1% O₂) 24 h before the experiment (n = 4 independent experiments, four replicates/experiment). On the day of the experiment, cells were washed three times with 700 μL of KRPH (Krebs‐Ringer phosphate HEPES) and incubated with KRPH + BSA for 1 h at 37°C. Seahorse XFe24 flux analyzer (Agilent Technologies) was used to determine maximum glycolytic capacity and ATP production rate using methods reported by Mookerjee et al. (Mookerjee et al. 2016) Experimental design for ATP‐consumption included sequential additions of 10 mM glucose, 1 μM rotenone plus 1 μM myxothiazol, 2 μg/mL oligomycin. To measure glycolytic capacity, sequential additions of 10 mM glucose, 1 μM rotenone plus 1 μM myxothiazol, and 200 μM monensin plus 1 μM FCCP were performed. The normalized traces for oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were used to calculate experimental parameters.

Vehicle control and MDL‐800‐treated (25 μM) NP cells (n = 4 independent experiments) were used to extract either histones (Abcam‐ab113476) or total cellular proteins. 10–40 μg of protein was electroblotted to nitrocellulose membranes (Amersham, GE, Burlington, MA, USA). The membranes were blocked and incubated overnight at 4°C with antibodies against H3K9ac (1:1000, 06–942S, Millipore), Histone 3 (1:1000, 9715), SIRT6 (1:1000, D8D12–12,486), LC3 (1:1000, 12,741), GAPDH (1:3000, 5174) from Cell Signaling Technology. Immunolabeling was detected on the Azure 300 system using an ECL reagent (Azure Biosystems, Dublin, CA), and densitometric analysis was performed with ImageJ.

The Histone Modification Multiplex Assay kit (ab185910, Abcam) was used to quantify Histone 3 modifications, including lysine acetylation and mono‐, di‐, and tri‐methylation, using histones isolated from NP cells treated with vehicle control and MDL‐800 for 24 h, according to the manufacturer's instructions (n = 3 independent experiments).

Following vehicle control and MDL‐800 treatment (25 μM) of rat NP cells for 24 h, total DNA‐free RNA was extracted using RNeasy mini columns (Qiagen) (n = 4 independent experiments). The extracted RNA with RIN > 7 was used for RNA sequencing, and DEGs (Log₂FC ≥ 1.5, FDR ≤ 0.05) were analyzed using the GTAC‐CompBio tool (https://gtac‐compbio‐ex.wustl.edu↗). For ATAC‐sequencing, vehicle control and MDL‐800‐treated NP cells were trypsinized and collected by centrifugation at 600–800 g for 5 min at 4°C. Cells were resuspended in 500 μL cryopreservation medium containing 50% serum and 10% DMSO and then cryopreserved in 2 mL cryopreservation tubes and shipped to Azenta for sequencing. Briefly, cells were thawed and treated with Tn5 transposase, and tagmented DNA fragments were purified using a DNA purification kit for library synthesis (Azenta). Purified sequencing libraries were used for paired‐end 150‐bp‐long reads on Illumina NovaSeq. Reads were mapped to the reference genome with default parameters. Aligned ATAC‐seq reads were used to remove duplicates, and peak calling was performed, followed by a multi‐omic analysis of ATAC‐seq and RNA‐seq (FC ≥ 2, FDR ≤ 0.05) (see Supplemental Methods in Data S1).

All statistical analyses were performed using Prism7 or above (GraphPad, La Jolla). Because interactions among genetic, biological, and biomechanical factors at individual spinal levels can result in distinct phenotypic outcomes, each intervertebral disc or vertebra was treated as an independent experimental sample, consistent with methodologies adopted by several other groups (Battié et al. 2008; Mohanty et al. 2019; Novais et al. 2021; Tessier et al. 2020). Data are represented as box and whisker plots with median, minimum, and maximum values or mean ± SEM. Data distribution was assessed using normality tests, and differences between the two groups were analyzed using an unpaired t‐test. The differences among the three groups were analyzed using ANOVA or the Kruskal–Wallis test for non‐normally distributed data. A chi‐square (χ²) or Fisher test, as appropriate, was used to analyze the differences between the distribution of percentages. p ≤ 0.05 was considered statistically significant.