Lactate-Driven Reprogramming of Monocyte Bridges Bone Loss in Inflammatory Comorbidities

Abstract

Inflammatory bone loss is a shared pathological feature of chronic diseases such as periodontitis (PD) and rheumatoid arthritis (RA). Despite affecting distinct tissues, these diseases exhibit a bidirectional association and converge on common immune-mediated mechanisms of bone resorption. To uncover the molecular drivers underlying bone destruction across inflammatory comorbidities, we combined bioinformatic analyses with experimental validation, using PD and RA as clinically relevant models of inflammatory disease comorbidities. Elevated blood lactate levels were observed in murine models of PD and RA and correlated positively with disease severity. Single-cell RNA sequencing data from PD and RA cohorts revealed upregulation of lactate metabolism-related genes in specific monocyte subsets, accompanied by enhanced pro-inflammatory signaling and osteoclastogenic programs. Using multiple machine learning approaches, SAT1, TET2 and HIF1A were identified as core lactate-related genes with strong diagnostic potential for both diseases. In vivo and in vitro experiments further validated that lactate-driven reprogramming of monocytes, marked by activation of core lactate-related genes in circulating monocytes and local macrophages, functionally connects immune activation with exacerbated bone resorption in comorbid PD and RA. Together, these findings define a lactate-driven immunometabolic axis connecting immune responses and bone remodeling and identify SAT1, TET2 and HIF1A as potential biomarkers for inflammation-related bone loss.

1. Introduction

Inflammatory bone loss represents a common pathological hallmark across multiple chronic diseases, including periodontitis (PD), inflammatory arthritis, inflammatory bowel diseases, and systemic autoimmune disorders. Despite affecting distinct tissues, these diseases often exhibit systemic inflammatory interactions in comorbidity, which can synergistically exacerbate bone resorption [1,2,3]. With the global aging population and increasing prevalence of metabolic and inflammatory comorbidities, understanding the shared molecular drivers of inflammatory bone loss has become an urgent scientific and clinical priority [4,5,6].

In this context, PD and rheumatoid arthritis (RA) provide a clinically relevant and mechanistically informative model for exploring common mechanisms underlying inflammatory bone destruction across distinct tissue environments. Clinical and epidemiological studies have demonstrated a bidirectional relationship between the two diseases [7,8]. Individuals with PD have a greater likelihood of developing RA, while RA patients frequently present with more severe periodontal destruction compared with non-RA individuals, independent of demographic and behavioral factors [9,10,11]. PD and RA share striking pathogenic similarities, each driven by chronic inflammation that not only degrades connective tissues and promotes bone erosion, and both are associated with other systemic inflammatory diseases [12,13]. In PD, local inflammation extends from the gingiva into the alveolar bone and can elicit systemic inflammatory responses contributing to diabetes, adverse pregnancy outcomes and cardiovascular diseases [14,15,16,17]. Although RA primarily affects the synovial membranes and bones within joints, it can also involve multiple organs and increase the risk of cardiovascular and other chronic inflammatory conditions [18]. In addition, accumulating evidence suggests that PD and RA may also exacerbate each other through circulating inflammatory mediators, immune cell activation, and metabolic dysregulation [13,19]. However, the molecular mechanisms connecting inflammatory bone loss in periodontal tissues and peripheral joints remain poorly understood [20,21].

Metabolic reprogramming is a key determinant of inflammatory bone loss, governing the differentiation and functional states of bone-resident cells, such as osteoblasts and osteoclasts, together with diverse immune cell populations [22]. Chronic inflammation induces alterations in the metabolic states of these cells, thereby disrupting the balance between bone resorption and bone formation [23]. As the terminal product of glycolysis, lactate is a central component of cellular metabolic reprogramming. Accumulating evidence indicates that lactate functions as a bioactive metabolite with broad immunomodulatory effects in diverse pathological settings, including cancer, infectious diseases, cardiovascular disorders, and autoimmune conditions. Consequently, lactate metabolism has attracted growing interest as a key metabolic pathway shaping inflammatory responses [24,25]. Although recent evidence links aberrant lactate metabolism and lactylation to the pathogenesis of both PD and RA, whether lactate serves as a shared molecule driver of bone loss in these chronic inflammatory diseases remains unclear [26,27,28].

The novelty of this study lies in our identification of lactate-driven reprogramming of monocytes as a shared pathogenic mechanism underlying inflammatory bone loss across distinct inflammatory diseases and comorbidities, using PD and RA as clinically relevant models. By integrating single-cell RNA sequencing (scRNA-seq), machine learning, and multi-level in vivo and in vitro validation, we demonstrated that elevated systemic lactate was associated with disease severity in both PD and RA. Monocyte subset with upregulation of lactate metabolism-related genes were enriched for pro-inflammatory signaling and osteoclastogenic programs. Using multiple machine learning approaches, we further identified SAT1, TET2, and HIF1A as core lactate-related genes with shared diagnostic relevance across both diseases. Experimental validation further demonstrated that elevated lactate levels drive immunometabolic reprogramming of monocytes and macrophages, characterized by the activation of key lactate-related genes. This reprogramming contributes to the mutual exacerbation of inflammatory bone loss in PD-RA comorbidity through enhanced osteoclastogenesis and amplification of inflammation. Collectively, this study highlights lactate as a pivotal immunometabolic mediator and lactate-driven monocyte reprogramming as a key link connecting inflammatory bone loss across comorbid chronic inflammatory diseases, while identifying potential biomarkers for translational intervention.

2. Materials and Methods

2.1. Animal Models

Animal experiments were conducted with C57BL/6J male mice that were 6–8 weeks old from Guangdong Medical Laboratory Animal Center. Mice were housed under specific pathogen-free conditions and had free access to food and water. The mice were divided randomly into four groups, each containing five animals (n = 5): a control group, mice with ligature-induced periodontitis (PD), mice with adjuvant-induced rheumatoid arthritis (RA), and mice with comorbid PD and RA. Sample size was determined by references and pre-experimental data. Animals were monitored every three days to exclude abnormal weight loss. The study was reported in accordance with the ARRIVE guidelines 2.0. Approval for all animal procedures was granted by the Animal Ethical and Welfare Committee of Sun Yat-sen University (SYSU-IACUC-2025-001568).

2.2. Induction of Periodontitis

PD was induced in five male C57BL/6J mice by ligating the maxillary second molars with 6-0 silk sutures under anesthesia by alfentanil [13,29]. The ligature was maintained for 14 days to generate a local biofilm-retentive milieu leading to localized inflammation and alveolar bone loss.

2.3. Induction of Adjuvant-Induced Rheumatoid Arthritis

RA was induced in five male C57BL/6J mice by subcutaneous injection of complete Freund’s adjuvant (Sigma-Aldrich, Saint Louis, MO, USA) into the footpad using a microinjector as previously described [30,31]. Briefly, 10 μL of complete Freund’s adjuvant was injected in two separate injections to the inner and outer sides of the tibiotarsal joint under anesthesia. The mice were monitored every day for indications of inflammation. Arthritis scores and ankle thickness were assessed daily based on joint swelling and erythema. Specifically, arthritis scores were rated with a semiquantitative scoring system, with scores ranging from 0 to 4 for each paw: 0 for normal; 1 for mild redness or slight swelling of the ankle or wrist; 2 for moderate swelling of the ankle or wrist; 3 for severe swelling, involving some digits, the ankle, and foot; and 4 for maximal joint inflammation. The total arthritis score for each mouse was the sum of the scores for all four paws, with a maximum possible score of 16 [32]. Ankle thickness was measured using a thickness gauge.

2.4. Induction of PD-RA Comorbidity

PD-RA comorbidity was established in five male C57BL/6J mice by inducing both PD and RA. PD was induced by ligating the maxillary second molars for 14 days, while RA was simultaneously induced in the same mice by subcutaneous injection of complete Freund’s adjuvant into the footpad. The mice were monitored daily for severity of arthritis as described above.

2.5. Micro-Computed Tomography (Micro-CT)

After euthanasia, the maxillae and hind limbs were harvested, preserved in 4% paraformaldehyde, and scanned using the VENUS^® Micro-CT scanner (PingSheng Medical Technology Co., Suzhou, China). Group allocation was recorded by one investigator, and samples were coded with anonymized IDs before imaging. Micro-CT reconstruction, ROI selection, and quantitative analyses were performed in a blinded manner using coded IDs. Bone morphometric parameters such as bone volume fraction (BV/TV), trabecular separation (Tb.Sp) and trabecular number (Tb.N) were analyzed. Statistical analyses were conducted on de-identified datasets, and group identities were revealed only after analyses were completed.

2.6. Immune Cell Isolation

Peripheral blood was collected in ethylenediaminetetraacetic acid-coated tubes and diluted with PBS. Peripheral blood mononuclear cells (PBMCs) were isolated using a density gradient separation kit (P5230, Solarbio, Beijing, China) according to the manufacturer’s instructions. Briefly, the diluted blood was layered onto the separation solution and centrifuged at 700× g for 25 min. The mononuclear cell layer was then collected and further purified to remove platelets and residual solution. For monocyte enrichment, PBMCs were plated and incubated at 37 °C for 2 h to allow adherence, after which non-adherent cells were removed. The adherent monocytes were subsequently harvested for downstream experiments. Neutrophils and lymphocytes were isolated from peripheral blood using cell-type-specific isolation kits according to the protocol in kits (P9201 and P8620, Solarbio, Beijing, China).

2.7. Lactate Measurement

Lactate levels were measured using an assay kit based on the WST-8 method (Beyotime, Shanghai, China), following the manufacturer’s directions. Briefly, peripheral blood serum and cell culture supernatants were collected after 24 h of incubation, and appropriate dilutions were prepared to ensure readings were within the linear range. For each sample, supernatant was incubated with freshly prepared working solution containing water-soluble tetrazolium salt-8 (WST-8) and enzyme mix at 37 °C for 30 min. Absorbance was recorded at 450 nm, with reagent-only wells serving as blanks. Lactate concentrations were calculated from a standard curve generated with serially diluted lactate standards and were normalized to cell number.

2.8. ScRNA-Seq Data Processing

Publicly available scRNA-seq datasets of PBMCs from cohorts with PD or RA were retrieved from the Gene Expression Omnibus (GEO) database (GSE244515) and from CellxGene Discover resource [33,34]. Raw gene–barcode matrices were processed using the Seurat package (v4.3) in R. Cells were filtered to remove low-quality entries. Data were log-normalized using the “LogNormalize” method, followed by identification of highly variable genes. Uniform Manifold Approximation and Projection (UMAP) was conducted for dimensionality reduction. Batch effects between datasets were adjusted with the Harmony (v1.2.3) algorithm [35,36].

2.9. Cell Clustering and Annotation

Cells were clustered using the Louvain algorithm, and cell types were annotated and visualized by dot plots based on canonical markers and reference-based markers. Major immune cell types were identified for downstream analysis.

2.10. Metabolic Pathway Activity Analysis

To evaluate the metabolic landscape of immune cells in different disease contexts, the scMetabolism package (v0.2.1) was utilized to calculate the scores of metabolism-related pathways in single cells [37]. The analysis utilized the Kyoto Encyclopedia of Genes and Genomes (KEGG) metabolic gene sets as the basis, with scores computed using the AUCell algorithm. Lactate metabolism-related pathways were particularly examined across immune cells.

2.11. Differential Expression and Enrichment Analysis

Differentially expressed genes (DEGs) were identified in monocytes with high lactate scores from periodontitis and rheumatoid arthritis datasets using FindMarkers. Identified DEGs underwent Gene Ontology (GO) and KEGG enrichment analyses using the clusterProfiler package (v4.8.3) [38]. Enriched biological processes and signaling pathways were visualized using dot plots.

2.12. Identification of Core Lactate-Related Genes and Machine Learning Analysis

The set of core lactate-related genes was obtained from the Gene Set Enrichment Analysis (GSEA) database and previously published studies. Public transcriptomic datasets from patients with PD (GSE16134) and RA (GSE89408) were retrieved from the GEO database and processed for downstream analysis [39,40]. DEGs were identified in monocytes with high lactate scores in both PD and RA cohorts. To refine these candidates, three machine learning algorithms—least absolute shrinkage and selection operator (LASSO) regression, random forest (RF) and support vector machine–recursive feature elimination (SVM-RFE)—were used. The overlapping results from these models were defined as core lactate-related genes. Their diagnostic performance was then evaluated by calculating the area under the receiver operating characteristic (ROC) curve using the pROC package (v1.18.4).

2.13. Histological Analysis

Following fixation, the tissues underwent decalcification and dehydration, were embedded in paraffin, and were sliced into 5 μm thick sections. Inflammatory infiltration and alveolar bone resorption were assessed by hematoxylin and eosin (H&E) staining.

2.14. Quantitative Real-Time PCR (qRT-PCR)

The Ultrapure RNA kit (CWBIO, Taizhou, China) was used for the extraction of total RNA. Reverse transcription was performed with StarScript III RTase (Genstar, Beijing, China). The primer sequences used for polymerase chain reaction were recorded in Table S1. The 2^−ΔΔCt method was employed to evaluate the relative levels of gene expression.

2.15. Immunofluorescence Staining

The slides underwent deparaffinization, rehydration, and antigen retrieval. After blocking, sections were incubated with primary antibodies against SAT1, TET2, HIF1A and F4/80, followed by fluorophore-conjugated secondary antibodies, with DAPI applied to stain the nuclei. Images were obtained using a confocal microscope (LSM980, Zeiss, Oberkochen, Germany), and fluorescence intensity was analyzed using ImageJ software (1.54g).

2.16. Isolation of Bone Marrow-Derived Macrophages (BMDMs) and TRAP Staining

Femurs and tibias were aseptically harvested from mice. Bone marrow cells were flushed out using cold DMEM medium. The cell suspension was filtered through a 70 µm strainer and centrifuged at 300× g for 5 min. Red blood cells were lysed for 5 min at room temperature. Bone marrow cells were seeded and cultured with 20 ng/mL M-CSF for 3 days to promote macrophage adherence and differentiation. Floating cells were discarded, and adherent macrophages were collected as BMDMs. Osteoclasts were induced using 2 ng/µL M-CSF and 40 ng/µL of RANKL and stained using the TRAP staining kit (AMIZONA, Vestavia Hills, AL, USA).

2.17. Statistical Analysis

Statistical analyses were conducted with R software (v4.3.1) and GraphPad Prism 9.0. For association analysis, Pearson’s correlation coefficient was calculated in R software. Comparisons between groups were measured and assessed with Student’s t-test and one-way ANOVA.

3. Results

3.1. Elevated Blood Lactate Level Correlates with Inflammatory Bone Loss Severity and Tissue Damage in PD and RA Murine Models

We first established murine models of PD and RA to investigate whether blood lactate levels are elevated in both diseases. After 14 days of ligation around the maxillary second molars, micro-CT scans and 3D reconstructions showed marked alveolar bone loss in PD mice compared with healthy controls (Figure 1A). This was further reflected by decreased BV/TV and Tb.N, along with increased Tb.Sp and greater distances from the cementoenamel junction to the alveolar bone crest (CEJ-ABC) (Figure 1B). In RA models, the first symptom of hind paw swelling was observable on Day 6 (Figure 1C). The increases in the arthritis score and ankle thickness afterwards indicated that the arthritic conditions were successfully induced (Figure 1D). We measured lactate levels in the blood and discovered a significant increase in both PD and RA groups compared to healthy controls (Figure 1E). To explore whether elevated lactate level contributes to disease severity, we conducted a correlation analysis between blood lactate, CEJ-ABC distances and arthritis scores. As shown in Figure 1F, higher lactate concentrations were closely associated with greater tissue destruction in both conditions, as indicated by increased CEJ-ABC distances and arthritis scores. These findings suggest that increased accumulation of blood lactate is a shared trait of PD and RA and is linked to the severity of inflammatory bone loss and tissue damage.

3.2. Monocytes Exhibit Enhanced Lactate Metabolic Activity in PD and RA

To investigate whether immune cell activation contributes to systemic lactate elevation in PD and RA, we analyzed scRNA-seq datasets from PBMCs of PD patients [33] and RA patients [34]. Following quality control and integration of scRNA-seq data (Figure S1A–D), we identified 207947 and 108717 cells in the peripheral blood of PD and RA patients, respectively. Next, UMAP was utilized to reduce dimensions, dividing cells into ten distinct clusters for PD and eleven clusters for RA (Figure 2A,B) according to the expression level of markers (Figure 2C,D). These clusters were annotated according to markers reported in the previous literature as follows: T cells (CD4/CD8A), monocytes (CD68), NK cells (NCAM1), B cells (CD79A) and plasma cells (IGHA1), respectively (Figure S2A–C). Since increased lactate production is the result of enhanced glycolysis, the changes in glycolytic activity in various PBMCs were further assessed. As shown in Figure 2E,F, glycolysis-related gene sets were markedly upregulated in monocyte clusters (highlighted in yellow). Bar plot analysis of glycolysis scores across PBMC subsets showed that monocytes significantly scored higher than other subsets in both PD and RA groups (Figure S3A). Dot plot analysis further confirmed that glycolytic gene expression was most elevated in monocytes (Figure 2G,H). To validate these transcriptomic findings, we measured lactate concentrations in sorted peripheral neutrophils, lymphocytes and monocytes of healthy, PD and RA mice after incubation for 24 h. Consistent with the gene expression data, monocytes from disease models exhibited significantly higher lactate levels compared to those from healthy controls. This trend was not observed in neutrophils or lymphocytes, suggesting that lactate accumulation is most prominent in monocytes (Figure 2I). Together, these results indicate that circulating monocytes exhibit enhanced lactate metabolic activity in both PD and RA, in parallel with elevated systemic lactate levels.

3.3. Lactate-Associated Reprogramming of Monocytes Is Linked to PD and RA Progression

To explore the pathogenic role of lactate metabolism, we investigated the expression of lactate metabolism-related genes in PBMCs from patients with PD and RA. Dot plot revealed widespread upregulation of these genes in monocytes from both diseases groups compared with healthy controls, with a more evident pattern observed in the PD samples (Figure 3A). Using the AddModuleScore method, scores for lactate metabolism were evaluated across immune cell populations. Monocytes consistently showed significantly higher lactate metabolism scores than other immune cell types (Figure 3B). Moreover, lactate metabolism scores of monocytes were markedly elevated in PD and RA patients compared with controls, aligning with the observed increase in lactate metabolism-related gene expression (Figure S3B). These findings indicate enhanced lactate-associated metabolic programs in monocytes, which may contribute to functional reprogramming and disease progression in PD and RA. To further assess the functional implications of lactate metabolism, monocytes were subsequently stratified into high- and low-lactate metabolism score subsets. We next investigated intercellular communication among PBMC subsets using CellChat analysis. Compared with healthy controls, monocytes with high lactate metabolism scores displayed markedly increased outgoing and incoming signaling strength in both PD and RA cohorts (Figure 3C,D), indicating enhanced communication activity. Consistent with a globally activated immune microenvironment in PD and RA, CellChat analysis revealed a pronounced upregulation of the VISFATIN signaling pathway across PBMC populations in both disease conditions (Figure 3E,F). VISFATIN (also known as nicotinamide phosphoribosyltransferase, NAMPT) is an adipokine secreted by various cell types, with increasing evidence pointing to its role as an inflammatory cytokine in several inflammatory conditions [41]. Its levels are notably elevated in diseases associated with bone resorption, such as PD and osteoporosis [42,43]. Mechanistically, VISFATIN signaling is emerging as a key mediator linking RANKL to osteoclastogenesis, suggesting its potential contribution to inflammatory bone loss [44]. Notably, detailed analysis of pathway-specific signaling networks demonstrated that VISFATIN signaling was selectively enriched in monocytes with high lactate metabolism scores, compared with their low-score counterparts, in both PD and RA samples, while this pattern was not observed in healthy controls (Figure 3G,H). Given the role of VISFATIN signaling in RANKL-induced osteoclastogenesis and inflammatory bone loss, these results suggest that monocytes characterized by enhanced lactate metabolism exhibit distinct functional communication profiles in PD and RA. To further characterize the intrinsic transcriptional programs underlyingthe altered communication patterns, we next compared the transcriptomic landscapes of the monocyte subsets with high and low lactate metabolism scores. Volcano plot analysis revealed distinct transcriptomic patterns between the two subsets in both PD and RA cohorts (Figure 3I), which was further supported by heatmap visualization (Figure S3C). Differential gene expression analysis revealed functional divergence between the subsets. KEGG and GO enrichment analyses showed that monocyte subsets with higher lactate metabolism scores were enriched for pathways associated with inflammatory signaling, oxidative stress, and immune activation (Figure 3J,K), supporting the notion that enhanced lactate metabolism is linked to an immune-activated monocyte phenotype and may accelerate disease progression in both PD and RA. We identified 1770 DEGs in monocytes from the PD dataset and 785 from the RA dataset. To identify lactate-related transcriptional alterations shared between PD and RA, rather than disease-specific changes, we identified nine overlapping genes in monocytes from both PD and RA by intersecting these DEGs with lactate metabolism-related genes, as shown in the Venn diagram (Figure 3L). The subsequent analyses focused on the characterization of these shared lactate metabolism-related genes.

3.4. Identification of Core Lactate-Related Genes for PD and RA

Enhanced lactate-associated metabolic reprogramming in monocytes may contribute to the development and interconnection of PD and RA. However, not all lactate metabolism-related genes are expected to play equivalent roles in disease progression. To identify core lactate-related genes that are most informative across both diseases, we applied multiple machine learning algorithms to prioritize key genes from the nine shared lactate metabolism-related candidates. The LASSO identified a subset of genes with strong predictive value, as shown by the coefficient profiles (Figure 4A and Figure S4A) and cross-validation error curve (Figure 4B and Figure S4B), resulting in the selection of five genes in the PD dataset and eight in the RA dataset. The random forest algorithm ranked the genes by their importance, identifying five top contributors for PD (Figure 4C) and five for RA (Figure S4C). In parallel, the SVM-RFE algorithm iteratively selected the most informative features, demonstrating a model involving seven genes for PD (Figure 4D,E) and six genes for RA (Figure S4D,E) and achieving the highest accuracy. The overlap among genes selected by LASSO regression, random forest, and SVM-RFE was visualized using a Venn diagram, revealing SAT1, TET2 and HIF1A as robust core lactate-related genes shared by PD and RA (Figure 4F). At a single-cell resolution, all three genes were distinctly expressed in monocytes from both disease groups (Figure 4G). ROC curve analysis in independent validation datasets confirmed the diagnostic potential of these genes, with high corresponding area under the curve (AUC) values supporting their capability to distinguish PD and RA patients from healthy individuals (Figure 4H). Finally, to further elucidate the biological functions associated with core lactate-related genes, we performed correlation-based GSEA in monocytes from PD and RA datasets. Genes were ranked according to their Spearman correlation with SAT1, TET2 and HIF1A expression, and KEGG enrichment analysis was conducted. The top ten pathways with the most significant enrichment were visualized by dot plots (Figure S5A,B). SAT1-associated genes showed a notable enrichment in metabolic and immune-related pathways, including glycolysis/gluconeogenesis and pyruvate metabolism. TET2-associated transcripts were mainly linked to osteoclast differentiation and TNF signaling. HIF1A-correlated genes showed strong enrichment in pyruvate metabolism and osteoclast differentiation. Collectively, these analyses identified core lactate-related genes that robustly distinguished disease states and were potentially associated with bidirectional disease exacerbation and bone destruction in PD-RA comorbidity, and they prompted further in vivo investigations.

3.5. Upregulation of Core Lactate-Related Genes in Peripheral Monocytes Is Associated with Exacerbated Bone Loss in PD-RA Comorbidity

To assess the in vivo relevance of lactate-associated metabolic reprogramming and the core lactate-related genes identified by scRNA-seq and machine learning analyses, we established a murine model of PD-RA comorbidity. Micro-CT imaging of the maxillary alveolar bone demonstrated markedly increased bone resorption in mice with combined PD and RA compared with those subjected to either condition alone (Figure 5A), which was further supported by quantitative analysis of bone parameters (Figure 5B). Consistently, H&E staining revealed increased inflammatory infiltration and aggravated alveolar bone loss in the combined group (Figure 5C). In parallel, more pronounced hind paw swelling (Figure 5D) and exacerbated joint destruction (Figure 5E,F) shown in mice with PD-RA comorbidity indicated bidirectional exacerbation of periodontal and arthritic bone damage under comorbid conditions. Additionally, qRT-PCR analysis of peripheral monocytes isolated from all experimental groups revealed significantly elevated expression of the identified core lactate-related genes in the comorbid group (Figure 5G), suggesting systemic activation of lactate-associated reprograming in monocytes during PD-RA comorbidity.

3.6. Core Lactate-Related Genes Are Locally Upregulated in Bone Tissues in PD-RA Comorbidity

Given the systemic upregulation of core lactate-related genes in peripheral monocytes, we next examined their spatial expression and cellular localization in periodontal and arthritic tissues. We performed immunofluorescence staining for SAT1, TET2 and HIF1A with the macrophage marker F4/80 in the periodontal and ankle joint tissues across experimental groups. In the PD-RA comorbidity group, all three markers of lactate-related genes showed stronger fluorescence signals in the maxillary alveolar bone compared with the PD-only, RA-only, or healthy control groups (Figure 6A), which was further supported by quantitative analysis of fluorescence intensity (Figure 6B). A similar pattern was observed in the ankle joint tissues (Figure 6C,D). Notably, fluorescence signals from the core lactate-related genes were predominantly colocalized with F4/80+ cells in both periodontal and joint tissues, indicating that lactate-associated metabolic programs are preferentially activated in tissue macrophages at inflammatory lesion sites. This spatial distribution was consistent with the elevated expression of core lactate-related genes observed in circulating monocytes. These findings provide further evidence that lactate-associated metabolic reprogramming in monocytes is a shared feature of inflammation in periodontal and joint tissues and may contribute to exacerbated bone loss in PD-RA comorbidity.

3.7. Elevated Lactate Induces Increased Expression of Core Lactate-Related Genes in Monocytes and Enhanced Osteoclast Differentiation

Building on the in vivo evidence of enhanced lactate metabolic-associated activation in PD-RA comorbidity, we next assessed whether elevated lactate was merely a metabolic byproduct or actively contributedto monocyte reprogramming and disease exacerbation. Blood lactate levels were therefore re-evaluated across all experimental groups, incorporating the PD-RA comorbidity group. Consistent with disease severity, lactate concentrations were highest in the PD-RA comorbidity group and positively correlated with the extent of periodontal and joint tissue destruction (Figure 7A–C). Although increased blood lactate levels paralleled the upregulation of core lactate-related genes in peripheral monocytes and tissue macrophages, it remained unclear whether this elevation simply reflected enhanced aerobic glycolysis in activated immune cells or actively participated in monocyte functional reprogramming. To address this question, PBMCs from healthy control mice and PD-RA comorbid mice were cultured for 24 h, and the resulting supernatants were collected as conditioned medium (CM). Circulating monocytes isolated from healthy mice were subsequently exposed to CM derived from each group (Figure 7D). qRT-PCR analyses revealed that CM from the PD-RA comorbidity group significantly increased the expression of the three identified core lactate-related genes and pro-inflammatory cytokines (TNF-α, IL-1β, and IL-6) in monocytes compared with CM from healthy controls. Notably, these effects were largely attenuated by AZD3965, a well-characterized pharmacological inhibitor that preferentially blocks MCT1-mediated lactate transport (Figure 7E,F) [45,46,47]. These results support the notion that monocytes not only generate elevated amounts of lactate during inflammatory activation but also actively respond to extracellular lactate-related cues, thereby linking lactate-rich inflammatory milieus to monocyte reprogramming and inflammation regulating. To further evaluate the functional consequences of lactate-driven monocyte reprogramming, we examined osteoclast differentiation under CM treatment. TRAP staining demonstrated that CM derived from the comorbidity group markedly enhanced osteoclast formation, as evidenced by in-creased numbers and areas of TRAP-positive multinucleated cells, whereas inhibition of lactate transport by AZD3965 significantly suppressed this effect (Figure 7G,H). Together, these findings suggest that elevated lactate contributes to transcriptional reprogramming of monocytes and promote osteoclast differentiation, thereby providing a mechanistic link between lactate metabolism and exacerbated bone resorption in PD-RA comorbidity.

4. Discussion

In this study, by integrating single-cell transcriptomic analysis, machine learning-based feature selection, and in vivo and in vitro validation, we demonstrate that lactate-associated immunometabolic dysregulation represents a shared feature of PD and RA. Moreover, our findings provide evidence supporting the lactate-driven reprogramming of monocytes as a mechanistic link underlying exacerbated inflammatory bone loss in the context of PD-RA comorbidity.

Traditionally regarded as a byproduct of glycolysis, lactate is now increasingly recognized as an active regulator of immune responses across diverse pathological contexts, including chronic inflammation and autoimmune diseases [48,49]. This includes modulation of immune cell fate and function by lactate accumulation in inflammatory microenvironments [50]. In PD, local lactate accumulation has been shown to delay junctional epithelial healing and facilitate bacterial invasion, thereby exacerbating periodontal lesions [51]. In RA, elevated lactate levels impair macrophage migratory capacity and promote the activation of CD4⁺ T cells within joint tissues, leading to immune cell accumulation and sustained inflammation [21,25,47,52]. Lactate also activates intracellular signaling pathways, such as MAPK and NF-κB in synovial fibroblasts, contributing to enhanced production of pro-inflammatory cytokines such as IL-6 and IL-8, further amplifying joint inflammation [53]. In line with these studies, our findings show a strong correlation between lactate levels and inflammatory bone loss, a common pathological outcome in both PD and RA. We further expand on this by demonstrating lactate as a systemic metabolic signal that links inflammatory bone loss in the oral cavity with that in distal joints through monocyte reprogramming, which also contributes to exacerbated bone destruction in comorbidity.

Monocytes exposed to inflammatory stimuli undergo rapid transcriptional and metabolic reprogramming that primes them for migration and differentiation into downstream effector cells, indicating that metabolic rewiring of monocytes is a key feature of inflammation [54]. Emerging evidence from human and animal studies show that monocyte reprogramming plays a role in the pathogenesis of inflammatory bone loss. Monocytes from RA patients with significant bone erosions exhibit pronounced activation of immuno-inflammatory pathways and differential expression of genes involved in tissue remodeling and bone formation [55]. This altered phenotype has been observed in both circulating and local synovial monocytes in arthritis [56]. Our study builds on these findings by identifying lactate-driven reprogramming of monocytes and expanding it to inflammatory bone loss in both PD and RA, linking it functionally to osteoclastogenesis and increased pro-inflammatory cytokine production. Collectively, our results suggest a positive feedback loop where elevated lactate reinforces monocyte reprogramming, connecting chronic inflammation to bone destruction and contributing to the bidirectional exacerbation observed in PD-RA comorbidity.

A set of core lactate-related genes (SAT1, TET2, HIF1A) identified in our study were consistently associated with inflammatory bone loss across both PD and RA. Importantly, these genes were not derived from a single experimental model, but were robustly identified through the integration of cross-disease single-cell transcriptomic analyses, machine learning-based feature selection, and multilayered in vivo and in vitro validation. This integrative strategy minimizes model-specific bias and underscores the biological relevance of these genes in chronic inflammatory settings. Functionally, the identified core lactate-related genes converge on shared biological dimensions, including metabolic adaptation to inflammatory stress, epigenetic and transcriptional regulation, and the amplification or tolerance of inflammatory signaling [57,58,59,60]. Rather than acting in isolation, these genes appear to constitute a coordinated lactate-responsive transcriptional module within monocytes and macrophages, enabling immune cells to dynamically adapt to elevated lactate levels. The consistent association of this lactate-responsive gene module with disease severity and bone destruction highlights their potential utility as biomarkers for inflammatory bone loss.

While our study highlights lactate-driven reprogramming of monocytes as a mechanistic bridge across inflammatory bone loss in PD-RA comorbidity, several limitations of this study should be acknowledged. First, although the murine models of PD, RA, and PD-RA comorbidity recapitulate key features of inflammatory bone loss, species-specific differences in immune regulation and metabolism may limit the direct extrapolation of our findings to human disease. Second, while this study focuses on extracellular lactate as a signaling molecule and its role in reprogramming monocytes, further insights could be gained by investigating intracellular lactate. A more detailed exploration of intracellular lactate levels could offer a deeper understanding of the intracellular mechanisms that mediate lactate-induced effects, such as intracellular receptors or enzymes that link lactate to lactylation. Future research should focus on these intracellular mechanisms to further elucidate how lactate drives monocyte reprogramming. Lastly, while SAT1, TET2 and HIF1A were identified as core lactate-related genes through bioinformatic analyses and experimental validation, this study did not directly manipulate individual genes in vivo or in vitro. As such, the precise gene-specific contributions and potential hierarchical relationships among these factors remain to be fully elucidated. Future studies employing targeted genetic or pharmacological approaches will be essential to further explore the roles of these lactate-responsive genes in inflammatory bone loss.

5. Conclusions

In conclusion, this study integrates single-cell transcriptomic analysis, machine learning-based feature selection, and multi-level in vivo and in vitro validation to identify lactate-driven reprogramming of monocytes as a shared immunometabolic mechanism underlying bone loss in PD, RA and their comorbid condition. Elevated lactate actively reshapes transcriptional programs in monocytes, promoting pro-inflammatory signaling and osteoclastogenesis. SAT1, TET2, and HIF1A were identified as core lactate-related genes that form a lactate-responsive transcriptional module closely associated with disease severity and inflammatory bone destruction. Collectively, our findings propose lactate-driven immunometabolic reprograming as a novel framework for understanding and monitoring inflammatory bone loss in PD-RA comorbidity and potentially other chronic inflammatory diseases.