Lipid metabolism dysregulation in rheumatoid arthritis: from multi-omics discovery to adjunctive therapy with conjugated linoleic acid

Abstract

Objective

This study aims to comprehensively analyze the role of metabolic signatures in rheumatoid arthritis (RA).

Methods

A total of 599 samples, including fecal (n = 206), plasma (n = 206), and peripheral blood mononuclear cell (PBMC) samples (n = 187), were collected from 443 individuals with RA, systemic lupus erythematosus, and controls. Fecal, plasma and PBMC samples were subjected to 16S rRNA sequencing, liquid chromatography-tandem mass spectrometry, RNA sequencing and 4D DIA proteomics, respectively. A collagen-induced arthritis mouse model was used to investigate the role of the metabolite conjugated linoleic acid (CLA) in experimental arthritis. An RA clinical cohort (n = 26) was recruited to validate the potential role of CLA as an adjunct to conventional synthetic disease-modifying anti-rheumatic drugs (csDMARDs) therapy.

Results

Multi-omics analysis revealed comprehensive alterations in metabolic signatures (metabolism-related molecules, metabolites, and metabolic pathways) in RA. Lipid metabolism pathways, particularly linoleic acid metabolism, appear to play a significant regulatory role in RA progression. Metabolites and molecules involved in linoleic acid metabolism were significantly associated with clinical phenotypes in RA patients. In animal experiments, CLA reduced arthritis scores, swollen joint counts, histopathological scores, and inflammatory levels. Finally, CLA supplementation accelerated the reduction of inflammatory levels in RA patients receiving csDMARDs therapy.

Conclusion

These findings reveal lipid metabolism as a druggable axis, position dietary CLA as an adjunct to improve outcomes in RA, and offer a new avenue to enhance remission rates.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13075-026-03831-9.

Introduction

Rheumatoid arthritis (RA) is a common chronic autoimmune disease that affects the synovial joints, leading to persistent inflammation, joint deterioration, and ultimately loss of mobility [1, 2]. Although therapeutic approaches for RA have significantly improved, the exact pathogenesis of RA remains unclear [3, 4]. Population-based surveys show that roughly 5 in every 1,000 adults worldwide have RA, and women are two- to three-times more likely to be affected than men [5]. Existing therapies do not eradicate the disease; their goal is to induce and maintain clinical remission. Yet present remission rates still fall short of clinical needs [6, 7]. With the widespread application and research of metabolomics in the field of autoimmune diseases, RA has also been characterized as a metabolic disorder [8]. However, the crucial role of metabolic signatures in RA has not yet been fully explored.

Abnormal metabolic signatures have been successively reported to be involved in the development of various diseases, including autoimmune diseases such as RA [9–11]. Multiple studies have demonstrated that dysregulation of glucose, glutamine, and lipid metabolism may contribute to the pathogenesis of RA [12, 13]. Qiu et al. identified metabolic abnormalities in glucose, amino acids, lactate, and citrate in the urine of RA patients through metabolomic analysis [14]. They suggested that increased glycolysis, disrupted citric acid cycle, oxidative stress, and proteolytic metabolism may be key features in RA patients. In our previous study, we identified significant alterations in the plasma metabolic profile of RA patients, with metabolism of glycine, serine, and threonine, as well as arginine and proline, playing critical roles in the disease. [15]. Importantly, changes in the metabolic profile are associated with disease activity [16]. We hypothesized that systemic metabolic disorders, rather than immune dysfunction, contribute to joint inflammation and contribute to inadequate clinical remission rates. Thus, investigating alterations in metabolic signatures associated with RA is essential for the effective clinical management of the disease.

In recent years, conjugated linoleic acid (CLA) has been identified as a metabolite with substantial potential to improve autoimmune diseases [17]. CLA is a fatty acid naturally found in the meat and dairy products of ruminant animals [18]. It has been demonstrated to provide protection in animal models of inflammatory bowel disease and beneficially modulate the immune response in patients with Crohn's disease [19]. The beneficial effects of CLA on immune and inflammatory responses have been reported in both animal models and clinical trials, including the reduction of colonic inflammation, decreased production of antigen-induced cytokines by immune cells, mitigation of adverse reactions to immune challenges, and regulation of inflammatory mediators [19–21]. Evidence suggests that CLA can alter immune responses by affecting the production of soluble factors and inflammatory molecules, thereby preventing immune-induced depletion [22]. In the autoimmune disease multiple sclerosis, dietary CLA supplementation for 6 months has been shown to significantly enhance the anti-inflammatory characteristics and functional profiles of circulating myeloid cells [17].

In this study, we comprehensively explored the role of metabolic signatures in the development of RA using a multi-omics strategy. Moreover, based on the results of the omics analysis, we evaluated the potential role of CLA in experimental arthritis mice and as an adjunct to conventional synthetic disease-modifying antirheumatic drugs (csDMARDs) therapy.

Materials and methods

Patients and samples

Patient samples in this study were collected at Dazhou Central Hospital. All patients with RA met the 2010 American College of Rheumatology criteria [23]. Additionally, patients with systemic lupus erythematosus (SLE) were recruited as disease controls, and all SLE patients fulfilled the Systemic Lupus International Collaborating Clinics classification criteria [24]. The RA patients included were either newly diagnosed or had discontinued medication for more than 2 months, to minimize the potential impact of drug use. Patients under the age of 18, those with psoriatic arthritis, malignancies, or other autoimmune diseases were excluded. Blood and fecal samples from the control (CON) group were obtained from volunteers. All participants provided written informed consent, and the study was conducted with the approval of the Ethics Committee of Dazhou Central Hospital (approval number: 2023–047). A total of 599 samples were collected from 443 participants. The clinical baseline characteristics of the study population are detailed in Supplementary Tables 1 and 2. Blood samples (collected in EDTA-K2 anticoagulant tubes) were immediately transported to the laboratory, processed into plasma and peripheral blood mononuclear cell (PBMC) according to standard protocols, and stored at −80 °C. Similarly, fresh fecal samples (1 mL) were placed in 1.5-mL EP tubes, snap-frozen in liquid nitrogen for 1 min, and stored at −80 °C. Clinical information of all participants was collected, including age, sex, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), and disease activity based on 28 joints (DAS28). The detailed sample composition and analysis workflow were shown in Fig. 1.

Fig. 1.

Study Overview. Step 1. Recruitment of study population, including collection of different sample types for omics sequencing. Step 2. Multi-omics analysis to comprehensively explore the central role of lipid metabolism signatures in RA, including Transcriptomics, Proteomics, Gut microbiome, and Metabolomics. Step 3. Mouse model experiments. Step 4. Clinical cohort studies. CON, Control; RA, Rheumatoid arthritis; SLE, Systemic lupus erythematosus; PBMC, Peripheral blood mononuclear cell; LC/MC, Liquid chromatography-mass spectrometry; DIA, Data independent acquisition; PsA, Psoriatic arthritis; OA, Osteoarthritis; CLA, Conjugated linoleic acid; MTX, Methotrexate; csDMARDs, Conventional synthetic disease-modifying anti-rheumatic drugs; CRP, C-reactive protein; ESR, Erythrocyte sedimentation rate; DAS28, Disease activity based on 28 joints; SJC28, Swollen joint count in 28 joints; TJC28, Tender joint count in 28 joints

Clinical cohort patient recruitment and design

This study is a prospective cohort study aimed at evaluating the impact of CLA, a lipid metabolite, as an adjunct to csDMARDs therapy on disease control in RA patients. The study subjects were RA patients who visited Dazhou Central Hospital from June to December 2023. Inclusion criteria were age between 18 and 65 years, disease activity with DAS28-ESR greater than 3.2, receiving at least one csDMARD (including methotrexate (MTX), hydroxychloroquine sulfate, leflunomide, auranofin, etc.), and meeting the 2010 American College of Rheumatology criteria for RA [23]. Glucocorticoids (prednisone ≤ 10 mg/day) and NSAIDs (1–2 tablets/day) could be used in combination. Exclusion criteria included a history of biologic agent treatment, presence of other severe diseases, and refusal to participate. Patients were randomly divided into two groups: the csDMARDs group received only csDMARDs therapy; the csDMARDs + CLA group received CLA capsules (4 capsules of 1.25 g/day, containing 80% CLA, provided by AMIX company, Czech Republic) in addition to csDMARDs therapy for 6 weeks, with follow-up every two weeks. Blood samples and clinical assessments were collected at baseline and at 6-week follow-up. Data were derived from medical records, including patients' clinical information, laboratory indicators, and medication use. The trial design is shown in Fig. 1 (Fig. 1). This study has been registered with the Chinese Clinical Trial Registry, with the registration number ChiCTR2400091743.

Nontargeted metabolite profiling and 16S rRNA sequencing

As previously described [16], non-targeted metabolite profiling and 16S rRNA sequencing were performed by Majorbio Bio-pharm Technology Co., Ltd. (Shanghai, China). Plasma metabolites and gut flora were analyzed using the Majorbio cloud platform (www.majorbio.com), including principal component analysis (PCA), Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis, principal coordinates analysis (PCoA), and Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt2) function prediction analysis.

Transcriptome sequencing and metabolism-related gene analysis

The PBMC of all subjects were collected with Trizol reagent, and the transcriptome sequencing was performed by Novogene Co., Ltd. Using standard extraction methods, RNA was extracted and tested for RNA integrity and total amount (Bioanalvzer 2100, Aailent, USA). Sequencing was performed with Illumina NovaSeq 6000. Metabolism-related genes (MRGs) were extracted based on the MSigDB database and the reference gene set "C2.cp.kegg.v7.5.1.symbols.gmt". A total of 938 MRGs were extracted for further analysis. The "limma" R package was applied for differentially expressed gene analysis and false discovery rate (FDR) < 0.05 was used to define significant MRGs. The significant MRGs were analyzed for Gene ontology (GO) enrichment function (http://www.geneontology.org/) and KEGG (http://www.kegg.jp) pathways. A "GSVA" script was used to estimate the relative activity of metabolic pathways in each sample based on 938 MRGs transcriptomic data. Correlations between MRGs and clinical characteristics were calculated using the Spearman algorithm.

Protein extraction and digestion

The PBMCs of the subjects were retrieved from a −80 °C freezer and an appropriate amount of DB lysis buffer (8 M urea, 100 mM TEAB (Sigma/T7408-500ML), pH = 8.5) was added to ensure complete lysis. After centrifugation at 4 °C at 12,000 g for 15 min, the supernatant was collected and an adequate volume of 1 M DTT (Sigma/D9163-25G) was added and incubated at 56 °C for 1 h. Subsequently, in a dark room, a sufficient amount of IAM (Sigma/I6125-25G) was added and the reaction was allowed to proceed for another hour in the dark. The supernatant was then collected, and the protein concentration was determined using the Bradford assay for quality control. A 20 µg sample of protein was subjected to electrophoresis using a 12% SDS-PAGE gel, with no significant signs of degradation observed in any of the samples. One milliliter of protein lysate, containing 5 mg of protein sample, was mixed with 50 µg of trypsin and 10 µL of 50 mM TEAB, and incubated for 4 h. Subsequently, an additional 50 µg of trypsin and CaCl2 were added for overnight enzymatic digestion. The reaction was halted using an equal volume of 1% formic acid (HCOOH, Fisher Chemical, A117-50). The supernatant was then applied to a C18 column for desalination, and the filtrate was collected and freeze-dried.

Phosphorylation-modified peptide enrichment

The freeze-dried powder was reconstituted in binding buffer and centrifuged for 5 min at 4 °C at 12,000 g. The supernatant was then collected and added to a pre-treated IMAC-Fe column with binding buffer, and incubated at room temperature for 30 min. After centrifugation at 2,000 g for 30 s, the column was washed successively with washing buffer and water. Finally, the column was eluted with elution buffer, and the peptide eluate was collected and freeze-dried.

DIA quantitative protein

The mobile phase A (2% acetonitrile, 98% water, pH = 10) and B (98% acetonitrile, 2% water, pH = 10) were prepared. The freeze-dried samples were dissolved in mobile phase A and fractionated using a C18 chromatographic column (Waters BEH C18 4.6 × 250 mm, 5 µm) on the Rigol L-3000 HPLC system. The flow rate for both mobile phases A and B was set at 1 mL/min, and the elution gradient was programmed as specified. The nanoElute UHPLC system (Bruker) and tims TOF pro2 mass spectrometer (Bruker) were utilized to generate mass spectral raw data in Data-Dependent Acquisition (DDA) mode to construct a DDA spectral library. Standard peptide samples containing the iRT reagent (Biognosys) were analyzed under standard liquid chromatography elution conditions using the UHPLC system. The tims TOF pro2 mass spectrometer operated in Non-Data-Independent Acquisition (DIA) mode to generate mass spectral raw data. For DIA acquisition, the full scan range of the mass spectrometer was set from m/z 100 to 1700, with a ramp time of 100 ms, locking the duty cycle to 100%; the scan window size was 25 Da, and the number of scan windows was two. The raw files from DDA and DIA mass spectrometry were searched against the human reference proteome from the UniProt database (homo sapiens uniport 2023.3.13. fasta) using the database search software Spectronaut (Biognosys, v9.0). The search parameters were as follows: the mass tolerance for precursor ions was set to 10 ppm, and for fragment ions, it was 0.02 Da, allowing up to two missed cleavages for the peptides. Spectronaut-Pulsar retained peptides with a confidence level above 99% that included at least one unique peptide per protein and removed peptides and proteins with FDR greater than 1%.

Animal experiments

Male DBA/1 mice (Specific pathogen-free, SPF, Huafu Kang, China), aged 6 weeks and weighing 17 ~ 21 g, were housed in an SPF animal facility at 25 °C. All mice had free access to water and food and were acclimatized for 7 days before the experiment began. The DBA/1 mice were randomly divided into four groups (n = 10 per group): a CON group, a model group (Model), a CLA group, and a MTX treatment group (treatment began on 21 days, 2 mg/kg, twice weekly). CON, Model and MTX groups, fed with maintenance chow; CLA group, −7 days were fed a chow rich in CLA (Tonalin® TG80, supplied by BASF/BTC-Europe GmbH), that was, 12 g CLA per 1 kg of maintenance chow. The CON group received tail vein injections of saline, while the other three groups were used to establish collagen-induced arthritis (CIA) mouse models. For the first immunization, 2 mg/ml chicken type II collagen was emulsified in an equal volume of Freund’s complete adjuvant. On day 21, the chicken type II collagen was emulsified in an equal volume of Freund’s incomplete adjuvant and used for booster immunization in the same manner. At week 10, the mice were euthanized by cervical dislocation, and blood and knee joint tissues were collected. The knee joint tissues were used for hematoxylin and eosin (H&E) staining. Blood samples were analyzed for serum levels of IL-1β, IL-6, IL-17A, and TNF-α using a commercial Merck Millipore liquid bead array kit (MHSTCMAG-70 K), according to the manufacturer’s instructions. This study was conducted under the approval of the Experimental Animal Ethics Committee of West China Hospital, Sichuan University (approval number: 20220629002).

Statistical analysis

Data management and analysis were conducted using IBM SPSS (version 25) and GraphPad Prism (version 8.0). Statistical analyses and visualization of results were performed using R software (version 4.1.1) and the OmicStudio tools (https://www.omicstudio.cn/tool). For normally distributed data, comparisons between two groups were analyzed using Student's t-test. For non-normally distributed data, the Mann–Whitney U test was used for comparisons between two groups. Differences were considered statistically significant at P < 0.05.

Results

Lipid metabolism was dysregulated in RA and was associated with clinical features

As previously described, 936 MRGs were extracted for subsequent analysis. PCA based on MRGs showed differences between the RA and CON groups (Fig. 2A). A total of 441 differentially expressed MRGs were identified, including 117 up-regulated genes and 324 down-regulated genes (Fig. 2B). KEGG enrichment analysis of differentially expressed MRGs revealed that multiple lipid metabolism pathways were significantly enriched, including glycerophospholipid metabolism, linoleic acid metabolism, and alpha-linolenic acid metabolism pathways (Fig. 2C). Notably, the majority of MRGs in these 14 significantly enriched lipid metabolism pathways were down-regulated in RA. Additionally, the same results were observed in GO enrichment analysis (Supplementary Fig. 1A). GSVA analysis showed that the relative activity of the linoleic acid metabolism pathway was significantly decreased in RA compared to the CON group (Fig. 2D, Supplementary Fig. 1B). Moreover, genes in lipid metabolism pathways were significantly correlated with clinical features, including DAS28, CRP, and ESR, especially PLA2 family members PLA2G4B and PLA2G6, which were significantly negatively correlated with DAS28, while PLA2G4A was significantly positively correlated with DAS28 (Fig. 2E, Supplementary Fig. 1C). Compared to the CON group, mRNA levels of PLA2G4B, PLA2G6, and PLA2G10 were reduced in RA, while mRNA levels of PLA2G4A were increased (Fig. 2F). Additionally, PTGS1 expression was increased in RA and was significantly positively correlated with DAS28 (Fig. 2G, Supplementary Fig. 1D).

Fig. 2.

Alterations of Lipid Metabolism Molecules in RA. A PCA analysis of metabolism-related genes. B Volcano plot of differentially expressed metabolism-related genes. Genes with FDR < 0.05 were defined as differentially expressed. Red indicates up-regulation, and blue indicates down-regulation. C Lipid metabolism-related pathways were significantly enriched. D GSVA analysis suggests decreased relative activity of linoleic acid metabolism. E Metabolism-related genes were significantly correlated with DAS28 scores. F mRNA expression levels of phospholipase A2 family molecules (PLA2G6, PLA2G4A, PLA2G10). G Elevated mRNA expression levels of PTGS1 in the arachidonic acid pathway. H Increased protein levels of phospholipase A2 and PTGS1. I Significant positive correlation between phospholipase A2 and PTGS1 at the protein level. J Elevated phosphorylation levels of phospholipase A2 at the S437 site. CON, Control; RA, Rheumatoid arthritis; PCA, Principal component analysis; FC, Fold change; FDR, False discovery rate; GSVA, Gene set variation analysis. mean ± standard error

Next, changes in metabolism-related molecules in RA were further explored at the protein level. Here, differences were also observed in metabolism-related molecules, with a total of 66 differentially expressed molecules identified (Supplementary Fig. 2A-B). Differentially expressed metabolism-related molecules were significantly correlated with DAS28 (Supplementary Fig. 2C). Additionally, the protein level of Phospholipase A2 was found to be up-regulated in RA (Fig. 2H). However, although PTGS1 expression was increased, it was not statistically significant (Fig. 2H). Moreover, a significant positive correlation was observed between Phospholipase A2 and PTGS1 (Fig. 2I). Finally, further analysis of the phosphorylation levels of Phospholipase A2 revealed that phosphorylation at the S437 site was significantly increased (Fig. 2J). In summary, lipid metabolism was dysregulated during the progression of RA and was associated with clinical features.

Changes in lipid metabolism signatures play an important role in the progression of RA

To explore the shared and unique metabolic signatures of RA, a disease control group with SLE was included. PCA showed that the principal components of RA and SLE were the same under different ion modes and differed from those of the CON group (Supplementary Fig. 3A-B). Compared to the CON group, RA and SLE had 219 and 205 differentially expressed metabolites, respectively (Fig. 3A). A total of 141 differentially expressed metabolites were shared between the two comparison groups. Notably, 41 unique differentially expressed metabolites were identified in the RA vs. CON comparison. Annotation based on the Human Metabolome Database revealed that changes in metabolites Lipids and lipid-like molecules and Organic acids and derivatives were predominant (Fig. 3B). Differentially expressed metabolites in the two comparison groups were enriched in 45 and 40 metabolic pathways, respectively, with 10 shared pathways. However, the linoleic acid metabolism pathway was unique to RA (Fig. 3C, Supplementary Fig. 3C). Correlation analysis between differentially expressed metabolites in two lipid metabolism pathways (linoleic acid metabolism and glycerophospholipid metabolism) and clinical features of RA showed that metabolites linoleic acid and glycerophospholipid were significantly negatively correlated with CRP, ESR, DAS28, and IL-6 (Fig. 3D). Compared to the CON group, the relative abundance of glycerophospholipid was decreased in both the RA and SLE groups, while linoleic acid was only different in the RA vs. CON group (Supplementary Fig. 3D). Gut microbiota analysis revealed differences in the composition of gut microbiota (at both phylum and genus levels) among the three study groups (Supplementary Fig. 4A-B). Among the top 10 genera with differential abundance, several have been previously reported to be associated with RA. Notably, changes in the abundance of Bacteroides and Faecalibaculum were also linked to lipid metabolism (Supplementary Fig. 4C). Differences in β-diversity were observed among the three study groups, while no significant differences were found in α-diversity (Chao and Shannon indices) (Fig. 3E, Supplementary Fig. 4D). Importantly, PICRUSt2 functional prediction analysis also suggested differences in lipid metabolism pathways in RA, including linoleic acid metabolism (Supplementary Fig. 4E). Additionally, gut microbiota genera significantly enriched in RA and CON were associated with lipid metabolism pathways (Supplementary Fig. 4F). In summary, the metabolic profile of RA was altered, with lipid metabolism being the primary focus. Figure 3F summarizes the changes in metabolic signatures of four lipid metabolism pathways in RA (Fig. 3F). Therefore, changes in lipid metabolism signatures, especially linoleic acid metabolism, may play an important role in the progression of RA.

Fig. 3.

RA showed disrupted linoleic acid metabolism negatively correlated with clinical features. A Upset plot showing significantly different plasma metabolites across three study populations. Metabolites with p < 0.05 and VIP > 1 were defined as significantly different. B Sankey diagram showing the classification of different metabolites in HMDB between RA, SLE, and CON groups. C KEGG enrichment analysis of different metabolites between RA and CON groups. Pathways with adjusted p < 0.05 were considered significantly enriched; pathway categories are indicated by different colored fonts. D Spearman correlation analysis between differentially expressed metabolites in glycerophospholipid and linoleic acid metabolism pathways and clinical features of RA; color bars represent the magnitude of correlation; metabolites in red font indicate significant up-regulation in RA, while those in green font indicate significant down-regulation. E Differences in β-diversity of gut microbiota at the genus level among RA, SLE, and CON groups based on PCoA analysis. R and P values were derived based on the distance algorithm: Bray–Curtis, 999 permutation, ANOSIM analysis method. F Comprehensive landscape of lipid metabolism signatures in RA progression. Circles represent metabolites, and rectangles represent molecules. Red font indicates up-regulation, green font indicates down-regulation, and black font indicates no observed change. CON, Control; RA, Rheumatoid arthritis; SLE, Systemic lupus erythematosus; HMDB, Human metabolome database; PCoA, Principal coordinates analysis; CRP, C-reactive protein; ESR, Erythrocyte sedimentation rate; DAS28, Disease activity based on 28 joints; RF, Rheumatoid factor; IL6, Interleukin 6; KEGG; Kyoto encyclopedia of genes and genomes

The lipid metabolite CLA significantly alleviates experimental arthritis

Omics analysis has implicated the linoleic acid metabolism pathway as significantly associated with the disease progression of RA. Therefore, we established a CIA model to assess the impact of the lipid metabolite CLA on RA. The animal experimental design was as previously described (Fig. 4A). CLA or MTX treatment groups exhibited milder symptoms, as assessed by arthritis scoring and swelling counts (Fig. 4B, Supplementary Fig. 5A). Compared to the model group, mice in the CLA or MTX treatment groups had significantly improved knee joint pathological scores (Fig. 4C). Additionally, compared to the model group, the pro-inflammatory cytokine IL-6 was significantly reduced in the CLA or MTX treatment groups (Fig. 4D). IL-1β and IL-10 were different in the MTX group, while TNF-α and TGF-β1 showed no differences between the two groups (Supplementary Fig. 5B). Moreover, CLA also significantly reduced the expression levels of IL-17A (Fig. 4D). Analysis of the gut microbiota in RA patients suggests an association between gut microbiota genera and lipid metabolism pathways. Next, to explore whether CLA alleviates experimental arthritis by affecting the composition of the gut microbiota, feces from mice in four study groups were collected for 16S rRNA sequencing. Analysis of the gut microbiota revealed profound changes in the overall microbiome composition with CLA treatment, as shown by PCoA and relative abundance at the genus level (Fig. 4E-F). The results showed that CLA supplementation helped to restore gut microbiome composition. In summary, additional CLA supplementation significantly alleviates the inflammatory features of experimental arthritis.

Fig. 4.

The lipid metabolite CLA alleviates experimental arthritis inflammation. A Experimental workflow for mice. B Arthritis scores and swelling scores in mice. CON (n = 10), MTX (n = 9), CLA (n = 10), Model (n = 9). Comparisons with the Model group. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant. C H&E staining and histological scores of knee joint tissues in mice; n = 8 for each group; scale bar = 50 µm. ‌a, Control group (CON); ‌b, Model group (Model); ‌c, treated with methotrexate (MTX); ‌d, treated with CLA. * indicates comparison between the CON group and the Model group. ***, p < 0.001. # indicates comparison between the MTX and CLA groups and the Model group. #, p < 0.05. D Expression levels of IL-6 and IL-17A in mouse serum samples; n = 8 for each group. * indicates comparison between the CON group and the Model group. *, p < 0.05; **, p < 0.01. # indicates comparison between the MTX and CLA groups and the Model group. #, p < 0.05; ##, p < 0.01. E PCA analysis of gut microbiota at the genus level in mice; n = 5 for each group. R and P values were derived based on the distance algorithm: Bray–Curtis, 999 permutation, ANOSIM analysis method. F Bar charts showing the relative abundance of gut bacterial genera with abundance > 0.01 at the genus level in the four groups of mice. CON, Control; CLA, Conjugated linoleic acid; MTX, Methotrexate; IL-6, Interleukin 6; IL-17A, Interleukin 17 A; PCA, Principal component analysis. mean ± standard error

The lipid metabolite CLA can be used as an adjunct to csDMARDs therapy to accelerate the reduction of inflammatory levels in RA

Furthermore, we recruited individuals with RA to assess the potential of CLA supplements in disease treatment modulation. After adjusting for the use of conventional antirheumatic drugs, a total of 26 individuals with RA were included for analysis (Fig. 5A). The csDMARDs and csDMARDs + CLA groups had similar clinical characteristics, except for differences in SJC28 and rheumatoid factor between the two groups (Fig. 5B). Our results indicated that, compared with the csDMARDs group, the csDMARDs + CLA group had a significantly lower fold change in ESR (Fig. 5C). Both the csDMARDs and csDMARDs + CLA groups significantly reduced CRP and DAS28 scores (Figs. 5D-E). However, the reduction in DAS28 score was more pronounced in the csDMARDs + CLA group (Fig. 5F). Interestingly, compared with the csDMARDs group, the csDMARDs + CLA group had significantly lower levels of white blood cells, including total white blood cell count, neutrophils, and lymphocytes (Supplementary Fig. 6A-C). However, the Neutrophil-to-Lymphocyte Ratio (NLR), an emerging inflammatory marker, was not significantly changed before and after treatment in either group (Supplementary Fig. 6D). Therefore, CLA supplements can assist in the clinical treatment of RA and accelerate the reduction of inflammatory levels in individuals undergoing conventional RA treatment.

Fig. 5.

Lipid metabolite CLA as an adjunct to csDMARDs therapy response. A Details of medication treatment for participants. The csDMARDs treatment group included Methotrexate, Leflunomide, Hydroxychloroquine, Prednisone, and NSAIDs. The csDMARDs + CLA group included CLA in addition to the csDMARDs regimen. B Baseline clinical characteristics of patients in both groups. Values were shown as Q50 (Q25-Q75) or percentages. C, D Impact on ESR (C) and CRP (D) after 6 weeks of treatment; paired t-test. E Impact on DAS28 after 6 weeks of treatment; paired t-test. F Change in DAS28 from baseline to 6 weeks (∆DAS28 BL-6w) in patients treated with csDMARDs and csDMARDs + CLA; Mann–Whitney test. csDMARDs, Conventional synthetic disease-modifying antirheumatic drugs; CLA, Conjugated linoleic acid; NSAIDs, Nonsteroidal anti-inflammatory drugs; CRP, C-reactive protein; ESR, Erythrocyte sedimentation rate; DAS28, Disease activity based on 28 joints; SJC28, Swollen joint count in 28 joints; TJC28, Tender joint count in 28 joints; VAS, Visual analog scale; HAQ, Health assessment questionnaire; SF-36, Short form-36 health survey; WBC, White blood cell; NEUT, Neutrophil; LYM, Lymphocyte; MONO, Monocyte. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant. mean ± standard error

Discussion

In this study, we employed non-targeted metabolomics, 16S rRNA sequencing, transcriptomics, and proteomics to analyze the alterations in metabolic features, including metabolites, MRGs, and metabolic pathways, in RA from multiple dimensions. Our multi-omics analysis highlighted the importance of lipid metabolism, particularly the linoleic acid metabolic pathway, in the progression of RA. Subsequently, we selected CLA, an isomer of linoleic acid, from the linoleic acid metabolic pathway for animal experiments and clinical cohort studies to evaluate its potential role in the disease course and treatment of RA. The study confirmed that CLA significantly alleviated the inflammatory features of experimental arthritis. Additionally, supplementation with CLA could serve as an adjunct to clinical treatment for RA, accelerating the reduction of inflammation in patients.

In recent years, metabolomics studies have identified RA as a complex disease potentially caused by dysregulation of multiple metabolic pathways, including glucose metabolism, amino acid metabolism, glutamine metabolism, and lipid metabolism [8]. James R Anderson, Srivastava NK, and others have highlighted the potential roles of glycolysis, the tricarboxylic acid cycle, the pentose phosphate pathway, arachidonic acid metabolism, and amino acid metabolism in RA through metabolomics analyses [12, 25]. Some researchers suggest that abnormal glucose metabolism in RA may be associated with increased glycolysis [14]. Notably, a study related to cancer found that glycolysis-derived metabolites were important for immune cell infiltration and the activation of inflammatory pathways [26]. Our previous research indicated that dysregulation of glucose, glutamine, and lipid metabolism may contribute to the development of RA [27]. The metabolic profile in RA was abnormal and associated with disease progression. Therefore, comprehensive analysis of the metabolic features in RA is crucial for understanding disease changes and clinical management.

Lipid metabolism, which encompasses the digestion, absorption, breakdown, and utilization of lipids, plays a crucial role in energy supply and signaling processes [28]. Recently, the relationship between lipid metabolism and RA has garnered increasing attention [29, 30]. Multiple studies have identified associations between plasma lipid changes in RA patients and inflammatory markers [31–33]. With the advent of lipid metabolomics, alterations in lipid small molecules and their derived metabolites, as well as metabolic pathways in RA, have been highlighted. Ahn et al. found a close link between lipid metabolism and RA, with increased levels of fatty acids, including palmitoleic acid, linoleic acid, arachidic acid, and stearic acid, in RA patients [34]. Therefore, they speculated that the inflammatory response might be related to increased lipolysis [35, 36]. Similarly, another study suggested that fatty acids could serve as markers of inflammatory arthritis [13]. In our study, four lipid metabolic pathways—linoleic acid metabolism, alpha-linolenic acid metabolism, glycerophospholipid metabolism, and arachidonic acid metabolism—were significantly enriched. The metabolites linoleic acid and glycerophosphocholine, which are involved in these pathways, showed significant correlations with inflammatory markers in RA patients, including CRP, ESR, DAS28, and IL-6. Linoleic acid, an unsaturated fatty acid, has been shown to reduce IL-1 and IL-6 concentrations when administered orally, indicating its anti-inflammatory potential [37]. However, in our study, linoleic acid levels were decreased in the RA group, which was inconsistent with previous findings [34]. This discrepancy may be due to the different study subjects. In vitro synovial cells reflect adaptive metabolic alterations in the local inflammatory microenvironment, whereas patient plasma metabolites represent systemic metabolic states. Another unsaturated fatty acid, arachidonic acid (AA), and its derivatives, such as prostaglandins (PGs), thromboxanes, and leukotrienes, are primarily considered pro-inflammatory mediators [38]. Although our non-targeted metabolomics analysis did not detect metabolites in the arachidonic acid metabolic pathway, both gut microbiota functional prediction and MRGs analysis highlighted this pathway. Importantly, the mRNA levels of PTGS1, which catalyzes the conversion of AA to PGs, were elevated in the RA group and significantly correlated with DAS28. PTGS1 has been confirmed to be involved in inflammatory response and plays an important role in a variety of physiological and pathological processes [39].

More studies have linked RA to the gut microbiota [40, 41]. We analyzed the gut microbiota of patients using 16S rRNA sequencing, and the results demonstrated altered gut microbial composition in RA populations. Several differential genera were found to be associated with RA, including Faecalibaculum, Bacteroides, Collinsella, and Lactobacillus. Importantly, changes in the abundance of the differential genera Bacteroides and Faecalibaculum were also linked to lipid metabolism. They influence fat absorption and distribution through multiple pathways, directly affecting the host's lipid metabolic balance [42–44]. This suggests that gut dysbiosis may participate in the development and progression of RA by regulating host lipid metabolism. It also provides a new avenue to explain the altered lipid metabolism characteristics in RA—the combined effects of gut microbiota dysregulation and host lipid metabolic dysfunction may be a key mechanism driving inflammatory progression in RA. In animal experiments, the gut microbiota composition also changed in the Model group. The relative abundance of Odoribacter was increased in the gut of mice in the arthritis model group, while it was relatively decreased in the MTX and CLA groups. Odoribacter belongs to the phylum Bacteroidetes, and its role in the context of autoimmune diseases remains controversial. We noted that a study by Liu X et al. also found higher abundance of Odoribacter in the gut microbiota of arthritic mice compared to untreated CIA-susceptible mice, which is consistent with our findings [45]. However, the relatively decreased abundance of Odoribacter in the MTX and CLA groups is inconsistent with the study by Chen Y et al., which reported that CLA increased Odoribacter abundance [46]. This discrepancy may be related to the distinct immunopathological mechanisms of arthritis and colitis. In our CIA model, the increase in Odoribacter may reflect the overall inflammatory microenvironment or state of microbial imbalance. Romboutsia belongs to the class Clostridia, and studies have linked it to gut health, butyrate production, and anti-inflammatory responses [47]. In some studies, increased abundance of Romboutsia is considered a marker of a shift toward a healthier gut microbiota state [48]. In our results, the abundance of Romboutsia in the microbiota of CLA-treated mice was higher than that in the MTX-treated group, which may indicate distinct or complementary pathways by which CLA and MTX regulate arthritis. CLA may focus more on restoring gut ecology, while MTX primarily exerts its effects through systemic immunosuppression.

CLA is an isomer of linoleic acid, and its overall effects are mediated by the interaction of two major isomers: cis-9, trans-11 and trans-10, cis-12. In recent years, CLA has been identified as a metabolite with the potential to improve autoimmune diseases [49]. Multiple studies have reported the positive impacts of CLA on immune and inflammatory responses, including its roles in alleviating colonic inflammation, reducing antigen-specific cytokines produced by immune cells, and modulating inflammatory mediators [19, 20]. Mechanistically, the anti-inflammatory effects of CLA and its metabolites may be related to the activation of nuclear receptors within immune cells [50]. Additionally, CLA has been found to exert anti-inflammatory effects by reducing immune cell activity and modulating the innate immune response to prostaglandins and leukotrienes [51]. In studies related to RA, CLA has been shown to reduce TNF-α levels in RA patients, demonstrating anti-inflammatory effects in active RA [21]. Another study on experimental arthritis found that dietary CLA alleviated inflammatory features and delayed the onset of arthritis, which was consistent with our findings [20]. Moreover, dietary CLA exhibited antibody-dependent anti-inflammatory activity [20]. In our study, mice fed with CLA-rich diets not only exhibited lower inflammatory features but also helped restore gut microbiota composition, which may be related to CLA's ability to mitigate intestinal barrier dysfunction and inflammation [17]. More importantly, this study confirmed that CLA can assist RA patients receiving csDMARDs treatment in accelerating the reduction of inflammation levels. As a composite indicator reflecting systemic inflammatory status, an elevated NLR is associated with disease activity, progression of joint damage, and poor prognosis in RA [52]. Recent studies have also indicated that NLR is linked to the pathogenesis of RA and response to biologic/targeted synthetic disease-modifying antirheumatic drugs (b/tsDMARDs) [53]. Although we observed significant improvements in clinical symptoms and inflammatory markers (CRP, ESR) in both patient groups, no obvious change in NLR was detected. This may be attributed to the short 6-week treatment follow-up period of our study. Existing research has shown that a decrease in NLR is typically associated with long-term disease remission in RA patients [54]. Future studies could extend the follow-up duration to further investigate the relationship between NLR and long-term disease remission. Therefore, CLA offers a novel therapeutic approach for rheumatoid arthritis, CLA may hold promise for clinical applications in the treatment of RA.

This study has several limitations. First, limited generalizability of clinical cohort findings. ‌The prospective clinical cohort evaluating CLA as an adjunct therapy included only 26 RA patients. While statistically significant improvements in inflammatory markers (e.g., CRP, DAS28) were observed, the small sample size and single-center design may restrict the extrapolation of these results to broader populations. Second, incomplete mechanistic elucidation of CLA’s anti-Inflammatory effects.‌ Although CLA supplementation demonstrated efficacy in murine models and human cohorts, the precise molecular mechanisms underlying its anti-inflammatory actions—particularly its interactions with immune cell receptors (e.g., nuclear receptors) and downstream signaling pathways—remain partially characterized. The study primarily correlates CLA with restored gut microbiota composition and reduced pro-inflammatory cytokines (e.g., IL-6, IL-17A) without fully delineating causal relationships or CLA’s direct targets in RA pathogenesis. Lastly, short-term follow-up in clinical intervention.‌ The 6-week duration for assessing CLA supplementation in RA patients, though sufficient to capture acute inflammatory changes, falls short of evaluating long-term sustainability, safety, and impact on structural joint damage. These limitations highlight opportunities for future studies to validate CLA’s therapeutic potential in larger, multi-center trials and deepen mechanistic insights into lipid metabolism-driven inflammation in RA.

In summary, our integrated multi-omics analysis delineates a comprehensive landscape of altered metabolic features in patients with RA. We identified lipid metabolic pathways, most notably linoleic acid metabolism, as playing a pivotal role in RA pathogenesis. Furthermore, the lipid metabolite CLA demonstrates the potential to ameliorate inflammatory manifestations and function as a valuable adjunct to csDMARDs therapy, thereby paving the way for novel adjunctive therapeutic strategies in RA clinical management.

Authors’ contributions

Conceptualization: Q.Z., P.S., and F.Z. Methodology: J.C. and C.J. Investigation: T.W., Y.W., J.G., and J.Z (Jie Zhang). Visualization: C.J., X.Z., and J.H. Supervision: F.Z. Writing—original draft: J.C., Z.Y., and W.D. Writing—review and editing: Q.Z., P.S., and F.Z. Resources: F.Z. Data curation: J.C. and J.S. Validation: J.Z (Jing Zhu). and J.W. Formal analysis: J.C., Z.Y., and W.D. Software: C.J. Project administration: F.Z. Funding acquisition: F.Z.

Funding

This study was supported by Sichuan Science and Technology Program (2022YFS0250), the National Science and Technology Major Project of the Ministry of Science and Technology of China (2024ZD0520505), the Key Research and Development Program of Dazhou Science and Technology Bureau (24ZDYF0005), the Administration of Traditional Chinese Medicine of Sichuan Province Fund (25MSZX439), the Chengdu University of Traditional Chinese Medicine (YYZX2022178), and Tianjin Key Medical Discipline (Specialty) Construction Project (TJYXZDXK-009A).

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee Board of the Dazhou Central Hospital (2023–047). This study has been registered with the Chinese Clinical Trial Registry, with the registration number ChiCTR2400091743. All subjects provided written informed consent to participate in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.