Rheumatoid arthritis (RA) is a chronic, systemic autoimmune disease that leads to progressive disability and systemic manifestations, ultimately resulting in premature mortality (1, 2). The inflammatory process typically begins in smaller joints, presenting with discomfort, stiffness, and pain. As the disease progresses, it involves larger joints, impairs their function, and may lead to severe disability (3). In extreme cases, extra-articular manifestations may occur, affecting the skin, eyes, lungs, and heart, potentially contributing to premature death (1). RA affects approximately 1% of the global population (4), with a higher prevalence in women than men, and the peak incidence occurring between 40 and 50 years of age (5). The disease develops as a result of complex interactions among genetic, environmental, and immunological factors. Among the most significant genetic predispositions is carriage of the HLA-DRB1 allele, referred to as the “shared epitope” which markedly increases the risk of developing RA, although its presence alone is not sufficient to trigger disease onset. Environmental factors, such as tobacco smoking, infections, and oxidative stress, act as additional stimuli initiating the pathological inflammatory process (6). The pathogenesis of RA is driven by aberrant activation of T and B lymphocytes, resulting in the production of autoantibodies, including rheumatoid factor (RF) and anti-citrullinated protein antibodies (ACPA, anti-CCP). This is accompanied by increased production of pro-inflammatory cytokines, such as TNF-α, IL-6, and IL-1β, which mediate inflammation, joint tissue destruction, pain, and functional impairment (2). Despite advances in understanding the immunological mechanisms underlying RA, the phenotypic variability and heterogeneity of its clinical course suggest the presence of additional disease-modifying factors.

Currently, the diagnosis of RA is based on the 2010 classification criteria established by the American College of Rheumatology and the European League Against Rheumatism (ACR/EULAR). These criteria encompass four principal diagnostic domains: the number of affected joints, serology (including levels of RA-associated autoantibodies such as RF and ACPA), acute-phase reactants (APRs), including elevated C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR), and the duration of symptoms (7). The implementation of the 2010 criteria has significantly improved disease detection, particularly in early-stage arthritis, compared to previously used criteria (8–10).

The presence of autoantibodies, including ACPA and RF, provides a robust foundation for diagnostic classification. However, a substantial subset of RA patients remains ACPA-negative, indicating the existence of disease subtypes with distinct immunological profiles and potentially differing pathogenic mechanisms (11). Moreover, approximately 10% of all RA patients are not identified at early stages of the disease, of whom nearly 9% are seronegative, despite the critical importance of early diagnosis for both seropositive and seronegative patients (12–15). The integration of innovative diagnostic biomarkers may enhance detection efficacy, particularly in scenarios where conventional classification criteria are insufficient, such as in seronegative patients. Despite the growing body of evidence, there remains a lack of consistent and clinically validated epigenetic biomarkers derived from peripheral blood that could be implemented in routine clinical practice. Increasing attention is being directed toward epigenetic mechanisms, which act as a bridge between genotype and environmental influences. Epigenetics encompasses gene expression modifications independent of DNA sequence changes, including DNA methylation, histone modifications, and regulation by non-coding RNAs (ncRNAs) in RA (2, 16, 17).

Peripheral blood provides a practical and clinically relevant biological material for studying systemic molecular alterations associated with RA. It is easily accessible, minimally invasive, and suitable for repeated sampling, enabling longitudinal monitoring of disease-related molecular dynamics. Fractionation of blood into plasma, serum, and cellular components facilitates standardized analysis and supports the detection of circulating biomarkers, including stable extracellular ncRNAs protected within vesicles or ribonucleoprotein complexes (18–20).

Analyses of peripheral blood allow for the assessment of systemic molecular changes that may reflect early disease stages before the onset of clinical joint symptoms. The presence of ACPA in serum long before clinical manifestations suggests that RA pathogenesis may begin systemically prior to localization in the joints. This challenges the traditional view of RA as a disease initially confined to synovial tissue. Serum ACPA thus exemplifies a biomarker that can signal systemic alterations preceding local joint damage (21).

The aim of this review is to provide an integrated overview of epigenetic alterations in rheumatoid arthritis, encompassing DNA methylation, histone modifications, and non-coding RNA-mediated regulation, with a particular focus on blood-based biomarkers. We critically evaluate their clinical performance, reproducibility, and translational feasibility for diagnostic and prognostic applications. Special emphasis is placed on peripheral blood and its fractions as accessible biospecimens, as well as on emerging classes of non-coding RNAs that may expand current biomarker and therapeutic paradigms in RA.

Epigenetics is a field that investigates changes in gene expression that occur through chromosomal modifications rather than alterations in the DNA sequence itself (22). The main mechanisms include DNA methylation, histone modifications, and gene regulation by non-coding RNAs. The epigenome is highly influenced by environmental factors- aging, diet, and stress can induce epigenetic alterations that contribute to disease development (23, 24). These environmentally induced epigenetic modifications are increasingly recognized as key contributors to complex disease phenotypes. Current evidence indicates that epigenetic changes are involved in the pathogenesis of various disorders, including cancer (25), metabolic diseases such as type 2 diabetes (26), neurological disorders (27) and autoimmune diseases (28) including RA (29). Studies have demonstrated that patients with RA exhibit global DNA hypomethylation in immune cells, leading to excessive expression of pro-inflammatory genes (16). In addition, hypermethylation of anti-inflammatory genes has been observed, further sustaining chronic inflammation (30). Histone modifications also play an important role in RA by regulating genes involved in inflammatory responses, cell proliferation, and apoptosis of synovial cells (31). It is hypothesized that RA develops in genetically predisposed individuals through an interplay between genetic variability, epigenetic modifications, and environmental triggers, which may be initiated by stochastic events such as injury or infection (2, 32, 33).

DNA methylation is one of the principal epigenetic mechanisms. It primarily affects cytosine residues within CpG dinucleotides, which frequently cluster in CpG islands. In humans, this modification is catalyzed by DNA methyltransferases (DNMTs), which transfer a methyl group from S-adenosylmethionine to cytosine. Because approximately 60-70% of mammalian gene promoters contain CpG islands, methylation has a substantial impact on transcriptional regulation (34). Methylated DNA promotes heterochromatin formation, leading to nucleosome compaction and reduced accessibility for transcription factors, ultimately resulting in gene silencing (35). Conversely, active DNA demethylation removes methyl groups from cytosine residues, increases chromatin accessibility, and facilitates transcriptional activation (36). Together, these opposing processes help maintain stable gene-expression patterns, which is essential for understanding persistent inflammatory phenotypes observed in diseases such as RA.

Another form of epigenetic modification is RNA methylation, which occurs post-transcriptionally, influencing RNA splicing, translation, stability, and degradation (37). Because these modifications are reversible, they dynamically modulate the expression of mRNAs, cytokines, and regulatory genes (38). Several types of RNA methylation have been identified, including m6A, m1A, m5C, and m7G, with m6A being the most abundant and best characterized (39). m6A marks are deposited by a methyltransferase complex primarily composed of METTL3 and METTL14, together with regulatory cofactors such as WTAP. The modification is removed by demethylases FTO and ALKBH5, whereas ‘reader’ proteins, most prominently the YTH domain family that bind to m6A-modified transcripts and determine their fate (40–42). Through these coordinated activities, RNA methylation fine-tunes gene-expression programs relevant to immune activation and chronic inflammation.

Histone modifications constitute another major layer of epigenetic regulation. Histones, which package DNA into nucleosomes, possess N-terminal tails that undergo multiple enzyme-driven post-translational modifications, including acetylation, methylation, phosphorylation, and ubiquitination (43). Among these, acetylation and methylation are the best characterized. Histone acetylation involves the addition of acetyl groups to lysine residues, reducing histone-DNA affinity, relaxing chromatin structure, and promoting transcriptional activation. This modification is highly dynamic and is controlled by the opposing activities of histone acetyltransferases (HATs) and histone deacetylases (HDACs) (44). Histone methylation, catalyzed by histone methyltransferases (HMTs), can either activate or repress gene expression depending on the specific residue and degree of methylation (45). Phosphorylation of histones also influences chromatin dynamics, although its biological roles are less well characterized. Histone ubiquitination, involving the attachment of ubiquitin to lysine residues, is most commonly linked to transcriptional repression, although context-dependent activating functions have also been described (46).

Non-coding RNAs represent another major epigenetic regulatory mechanism. ncRNAs encompass a broad group of transcripts with diverse roles, some of which participate in epigenetic regulation. Based on their length, ncRNAs can be classified into diverse groups - from short species such as miRNAs (micro RNA), siRNAs and piRNAs to longer forms like lncRNAs (long non-coding RNAs) and circRNAs (circular RNAs) - some of which exert regulatory influence on epigenetic processes including DNA methylation, chromatin remodeling and mRNA stability (47). Accumulating evidence demonstrates that ncRNAs regulate key epigenetic processes, including DNA methylation, histone modifications, and sequence-specific control of mRNA translation and stability (48–50). In rheumatoid arthritis, ncRNAs function as both regulators and effectors of inflammatory signaling, shaping the activity of lymphocytes, macrophages, and synovial fibroblasts and thereby contributing to chronic synovial inflammation (51).

The following subsections summarize the best-characterized classes of ncRNAs, including miRNAs, lncRNAs and circRNAs, and introduce additional RNA species that have received comparatively little attention to date, such as siRNAs, piRNAs, Y RNAs (Ro60-Y RNA complexes), snRNAs, snoRNAs and tRNAs. Expanding the scope beyond the traditionally studied ncRNA groups provides a more complete understanding of RNA-mediated epigenetic regulation and highlights regulatory mechanisms that may have been overlooked in rheumatoid arthritis research. By integrating emerging evidence on these lesser-known ncRNA classes, this review aims to open new research avenues, clarify their potential contribution to RA pathogenesis, and support the identification of novel diagnostic biomarkers and therapeutic targets.

DNA methylation in RA has been studied for several decades, whereas RNA methylation has only recently emerged as an additional epigenetic layer. However, the majority of current evidence still focuses predominantly on DNA methylation abnormalities. Early on, researchers observed global hypomethylation, particularly in rheumatoid arthritis synovial fibroblasts (RASFs), although the extent of this phenomenon in peripheral immune cells remains more heterogeneous across studies (52). Additionally, it has been shown that these changes lead to altered expression of genes regulating DNA methylation dynamics, including DNMTs and TET enzymes that intensify the pathogenesis of the disease (16, 53). In the course of RA, hypermethylation and decreased expression of anti-inflammatory genes also occur, which further sustains and amplifies inflammation (54). Various studies have been conducted to identify diagnostically and therapeutically useful biomarkers of RA, particularly in patients who do not produce RF or ACPA. Subsequently, selected biomarkers are discussed, which, according to the literature, may have potential diagnostic or therapeutic significance.

Pitaksalee et al. developed qMSP-based assays to identify blood-derived epigenetic markers for RA, identifying TNF-α methylation as the most promising candidate (55). The researchers compared methylation levels in peripheral blood mononuclear cells from RA patients (n = 64, IACON database; n = 126, RADAR database and from patients with other typical joint types: reactive, undifferentiated and psoriatic arthritis (PsA) (n = 63, IACON; n = 31, RADAR). TNF-α plays a fundamental role in RA pathogenesis, driving synovial inflammation, angiogenesis and immune cell activation (56, 57). Hypomethylation of the TNF-α gene has been observed in the course of RA. These changes have been shown to significantly distinguish patients progressing to RA from those with other forms of arthritis. Moreover, no correlations were found between TNF-α values and clinical or demographic variables, which grants this biomarker independent value as an indicator for RA detection. A DNA methylation threshold of 4.5% was established to classify patients into low- and high-risk groups for progression to RA. Based on this, the odds ratio in the high-risk group from the IACON database was 8.4, with a sensitivity of 68.7% and specificity of 79.4%. The positive predictive value (PPV) was 77.2%, the negative predictive value (NPV) 71.4%, and AUC of 0.741. Even better results were obtained for patients from the RADAR database, where the odds ratio for developing RA in the high-risk group was 27.6, with a sensitivity of 84.1%, specificity of 83.8%, PPV of 95.4%, NPV of 56.5%, and AUC of 0.840. The differences between the two groups are due to a revised patient classification system implemented after 2015, which was based solely on the 2010 EULAR classification criteria and excluded patients whose symptoms resolved within 1–3 months (55). Although these results are promising, they require validation in independent, prospectively collected cohorts before TNF-α methylation can be established as a reliable diagnostic biomarker. These findings indicate that assessment of DNA methylation changes in TNF-α has high diagnostic value. Moreover, adding this biomarker to the currently used classification criteria increased diagnostic accuracy from 0.917 to 0.954, and among ACPA (–)patients from 0.892 to 0.954. The introduction of the additional TNF-qMSP test appears to have significant added value and potential as a biomarker, particularly in the diagnosis of seronegative patients with early RA. Another biomarker that has recently attracted researchers’ attention is the CXC chemokine receptor 5 (CXCR5). A strong association of this surface protein with the pathogenesis of autoimmune diseases, including RA, has been observed. CXCR5 contributes to lymphocyte trafficking and ectopic lymphoid organization, processes highly relevant to RA pathogenesis (58). CXCR5 is upregulated in RA synovial tissue and expressed on multiple infiltrating immune cell types, including B cells, T cells, macrophages, and endothelial cells (59) In contrast, peripheral blood CXCR5⁺ B cells are decreased in RA patients compared with healthy controls, suggesting altered trafficking or receptor regulation during chronic inflammation (60). These findings likely reflect the migration of CXCR5⁺ B cells toward the inflamed synovium, rather than an overall decrease in CXCR5 expression. Significant methylation changes have also been identified within this gene specifically at site cg04537602 between patients with RA, osteoarthritis, and healthy individuals.

Moreover, methylation levels at this site correlate with the severity of inflammation in RA patients (61). Based on these findings, Shi et al. compared methylation levels at site cg19599951 of the CXCR5 gene in peripheral blood from patients with RA (n = 164), SLE (n = 30), ankylosing spondylitis (AS) (n = 30), PsA (n = 30), Sjögren’s syndrome (SS) (n = 24), and healthy controls (HC) (n = 30) (61). The study revealed significant differences in methylation levels between RA and AS patients, as well as between RA patients and HC. RA patients exhibited distinctly higher methylation levels than those with AS and HC. In other groups, no statistically significant differences in CXCR5 methylation were found. As in the previously discussed study, RA patients were subdivided according to the presence of autoantibodies. The researchers then performed ROC analysis to assess the ability of CXCR5 methylation to distinguish RA subgroups from healthy controls and AS patients. The results indicated that, for the entire RA group, measuring CXCR5 methylation changes has limited diagnostic value when compared to healthy controls (AUC = 0.658) and similarly when compared to AS patients (AUC = 0.662). For individual RA subgroups compared with AS patients, diagnostic performance varied (AUC ranging from 0.624 to 0.967), with the best results obtained for ACPA (–) patients. However, these results require replication in larger and independently collected AS cohorts, as the diagnostic confidence intervals were wide and subgroup analyses may be underpowered. Thus, this method shows potential for differentiating ACPA (–) RA patients from those with AS, particularly in cases of seronegative patients presenting with atypical symptoms. However, to better assess the clinical utility of this approach, studies involving a larger cohort of AS patients are necessary. The study also lacked comparative data between RA subgroups and healthy controls. Shi et al. additionally considered clinical features and patient data such as age, sex, ESR, and CRP values. When these variables were included alongside CXCR5 methylation levels in the analysis, the diagnostic value of the method improved markedly- the AUC increased from 0.662 (for CXCR5 methylation alone) to 0.934- allowing for a clear distinction between RA and AS patient groups. For similar study groups, Jiang et al. conducted research investigating the diagnostic potential of DNA methylation levels of the HIPK3 gene (62). HIPK3 encodes a kinase involved in transcriptional regulation and cellular stress responses (63). It has been demonstrated that HIPK3 exhibits significantly lower methylation levels in the peripheral blood of RA patients, which negatively correlates with CRP levels (64). Jiang et al. analyzed differences in methylation levels across several rheumatic diseases, with the most statistically significant differences observed in patients with RA, SLE, and gout. ROC curve analysis for mean HIPK3 methylation levels yielded AUC values of 0.742, 0.701, and 0.948, respectively. Nevertheless, the inclusion of multiple heterogeneous disease groups raises concerns about disease-specificity, and future studies should validate HIPK3 methylation in strictly defined early-RA cohorts and indicated an added diagnostic value of this method for the mentioned diseases. The RA patient group was further divided into four subgroups based on the presence of autoantibodies. The best diagnostic performance (AUC = 0.836) was observed in RF(+)/ACPA (–) patients, while the lowest was found in seronegative patients (AUC = 0.612). Despite this lower value, it was still sufficient to suggest potential improvement in the diagnosis of autoantibody-negative patients. ROC analyses were also conducted to differentiate RA from other inflammatory joint diseases- AS, gout, and PsA-yielding AUC values of 0.788, 0.643, and 0.696, respectively. These findings suggest that HIPK3 methylation levels have added value in distinguishing RA from other inflammatory joint conditions.

Another biomarker tested as a potential candidate for aiding RA detection is CDKN2A. CDKN2A encodes cell-cycle regulators implicated in dysregulated immune-cell proliferation in RA (65). In RA, CDKN2A undergoes hypomethylation and increased expression, which activates immune cells by inducing the production of pro-inflammatory mediators and modulating aberrant T and B lymphocytes, thereby promoting inflammation progression (66). Consistently with the role of CDKN2A in immune dysregulation, Li et al. performed a comprehensive analysis of autophagy-related genes (ARGs) in RA and identified CDKN2A as one of five key marker genes associated with disease susceptibility. Using a generalized linear model, the authors constructed a diagnostic nomogram integrating CDKN2A together with TP53, ATG16L2, FKBP1A and GABARAPL1, which robustly discriminated RA patients from healthy individuals. The model demonstrated high predictive performance across multiple independent datasets, with AUC values exceeding 0.85 in all validation cohorts, underscoring the diagnostic potential of CDKN2A within an autophagy-based biomarker panel (67). Moreover, Gravand et al. investigated CDKN2A methylation in peripheral blood mononuclear cells as a potential diagnostic biomarker for RA (66). The study included 75 RA patients and 75 healthy controls, and ROC analysis yielded an AUC of 0.705, with 72% sensitivity and 77.3% specificity. These results indicate that CDKN2A methylation possesses moderate discriminatory power and may serve as a promising component of a multi-marker diagnostic panel rather than a standalone test. Importantly, CDKN2A methylation levels showed no significant correlations with CRP or ESR, suggesting that this epigenetic alteration reflects disease susceptibility rather than inflammatory activity, which limits its potential utility for monitoring disease activity.

The most recent and promising research on improving RA detection through DNA methylation changes was conducted by Zhao et al. (68). They analyzed DNA methylation of the HTR2A gene (serotonin receptor 5-hydroxytryptamine 2A) in peripheral blood samples from RA patients and healthy controls. Previous studies link HTR2A polymorphisms to heightened pro-inflammatory cytokine responses in RA (69). Additionally, gene-gene interactions between a protective HTR2A haplotype and the shared epitope allele of HLA-DRB1 were correlated with positive RA autoantibody status (70). Zhao et al. demonstrated significantly higher methylation levels at HTR2A cg15692052 in RA patients. Furthermore, methylation levels positively correlated with ESR and CRP levels. Incorporating HTR2A methylation into standard clinical assessments increased the AUC for distinguishing RA patients from healthy controls to 0.757. Comparable results were observed for RA subgroups stratified by autoantibody status: double-negative RF/ACPA patients achieved AUC = 0.966, double-positive RF/ACPA AUC = 0.846, and RF-negative patients AUC = 0.932. These findings highlight the potential of HTR2A methylation as a highly effective biomarker for early RA detection, particularly in seronegative patients, where diagnostic performance was the highest among all studied groups (68). However, because subgroup analyses involved smaller sample sizes, these high AUC values should be interpreted with caution. However, as the analysis was based on a single CpG site, further multi-site and multi-cohort validation is required to confirm the robustness of this biomarker. In a separate study, Zhao J. et al. expanded the investigation of methylation markers by analyzing 50 CpG sites and evaluating their diagnostic potential in relation to clinical parameters. Epigenetic markers may correlate with clinical measures of RA activity, such as the number of tender and swollen joints, CRP levels, and disease-specific autoantibodies. Zhao J. et al. investigated 50 CpG methylation sites and, using logistic regression, assessed their diagnostic potential in relation to clinical parameters (71). Parameters included swollen joint count, tender joint count, ESR, CRP, RF, ACPA, VAS, and DAS28-CRP/DAS28-ESR. Patients were categorized into high- and low-value groups for each clinical parameter, based on median, mean, or clinically relevant cutoffs. AUC values ranged from 0.64 to 0.78, sensitivity from 0.57 to 0.88, and specificity from 0.42 to 0.77. The highest diagnostic performance was observed for patients with a high number of tender joints (AUC = 0.78, sensitivity: 0.75, specificity: 0.72). The results indicated that DNA methylation levels at 7 CpG sites exhibited statistically significant correlations with the assessed clinical parameters. Given the large number of CpG sites tested, the risk of false-positive associations is non-negligible, and correction for multiple testing is essential in future analyses. Additionally, specific CpG sites showed significant correlations with clinical measures, such as: chr1:235998714 positively correlated with the number of tender joints; chr1:211500151 positively correlated with the number of swollen joints; chr15:67357339, chr3:13895991, chr14:105055171 negatively correlated with RF; chr10:70231628 positively correlated with RF; chr3:13895991 negatively correlated with ACPA. The study also assessed whether patient-specific traits such as age, weight, and height influenced methylation. Several CpG sites showed significant positive or negative correlations with age and height, indicating that some selected biomarkers are affected by non-modifiable patient factors. This highlights the need to identify CpG markers that remain stable across demographic variables. Despite this, the study demonstrates that genome-wide methylation markers can aid in identifying disease susceptibility and progression trends. Future research should aim to reduce the number of CpG sites required for analysis and identify biomarkers that are independent of immutable patient characteristics such as age or height. Overall, these results highlight the diagnostic potential of DNA methylation changes in RA, particularly in seronegative patients. A summary of the main DNA methylation biomarkers investigated in blood and its fractions is presented in Table 1.

Luo et al. conducted a study on the presence of m6A methylation in peripheral blood and the mRNA expression of METTL3, METTL14, WTAP, ALKBH5, FTO, and YTHDF2 in peripheral blood cells (72). They demonstrated a statistically significant increase in m6A levels accompanied by a decrease in the mRNA expression of ALKBH5, FTO, and YTHDF2 compared to the control group. Additionally, the reduction in FTO mRNA expression correlated with DAS28-ESR, DAS28-CRP, IgG antibody titers, and the lymphocyte-to-monocyte ratio, whereas YTHDF2 mRNA expression correlated with RBC count, the percentages of lymphocytes and neutrophils, the neutrophil-to-lymphocyte ratio, and the lymphocyte-to-monocyte ratio. No correlations were found between ALKBH5 mRNA expression and clinical symptoms of RA. Similar results for YTHDF2 protein were reported by Yao et al. (73). Notably, Luo et al. examined total RNA from whole peripheral blood, which may partly account for discrepancies with studies analyzing PBMCs rather than whole blood.

Completely different results were obtained by Tang et al., who also examined m6A levels and METTL14 in peripheral blood mononuclear cells (74). They reported a decrease in METTL14 concentration in patients with active RA compared to patients in remission and healthy controls. This decrease led to reduced m6A expression. Furthermore, m6A levels negatively correlated with DAS28. The reduced m6A expression induced the secretion of proinflammatory cytokines by PBMCs, thereby exacerbating joint inflammation. These findings contrast with Luo et al., likely reflecting differences in cell populations analyzed and disease-activity stratification. In addition, Wang et al. observed significantly increased METTL3 expression in PBMCs and monocytes compared to healthy controls (75). Given the heterogeneity of sample types, disease activity states and methodological approaches across studies, these findings should be interpreted cautiously. These studies suggest that changes in m6A levels are likely due to alterations in the concentrations of methylation-related enzymes. Taken together, the studies underscore the relevance of RNA methylation in RA pathogenesis and indicate its potential yet still preliminary diagnostic utility, pending multi-cohort validation. The key findings are summarized in Table 2.

Histone acetylation of lysine residues particularly on histones H3 and H4 has been highlighted as a relevant epigenetic mechanism in RA. Acetylation by HATs increases chromatin accessibility, whereas HDAC-mediated deacetylation promotes chromatin compaction and transcriptional repression; accordingly, the HDAC/HAT activity balance may serve as a proxy for overall transcriptional dynamics (76, 77). In general, histone acetylation is associated with enhanced transcriptional activity, whereas deacetylation suppresses gene expression by promoting chromatin condensation (78). HDACs are classified into four main classes: class I (HDAC1, 2, 3, and 8), class IIa (HDAC4, 5, 7, and 9), class IIb (HDAC6 and 10), class III (SIRT1-7), and class IV (HDAC11) (76, 79).

Most studies have focused on joint tissue, particularly fibroblast-like synoviocytes (FLS), where dysregulation of the HAT/HDAC balance and increased expression of proinflammatory genes have been observed (80–82). Although early studies of histone modifications in rheumatoid arthritis focused predominantly on synovial tissue and fibroblast-like synoviocytes, increasing attention has been directed toward peripheral blood-based material, particularly PBMCs, owing to their accessibility, minimal invasiveness, and suitability for biomarker development and longitudinal monitoring. At the same time, mechanistic studies in synoviocytes and macrophages have demonstrated that aberrant histone acetylation promotes excessive production of proinflammatory cytokines (TNF-α, IL-6, IL-1β) and matrix-degrading enzymes (MMPs), supporting both the pathogenic relevance of histone modifications and the therapeutic potential of HDAC inhibition in RA (83). A seminal mechanistic study by Li et al. demonstrated that PBMCs from RA patients exhibit reduced HDAC3 expression and activity, increased histone acetyltransferase activity, and significantly elevated total histone H3 acetylation compared with healthy controls, supporting the contribution of altered HDAC/HAT balance to epigenetic dysregulation in RA (84). Importantly, these epigenetic alterations were associated with enhanced expression of pro-inflammatory genes and increased activation of circulating T lymphocytes. Although this study predates the primary 2020–2025 literature window, it is repeatedly referenced by subsequent blood-based epigenetic studies and provides essential mechanistic context for interpreting later findings. For this reason, it is included as a foundational reference in the present review. Subsequent studies published within the 2020–2025 timeframe have reinforced and extended these observations. Furthermore, modulation of HAT and HDAC activity in PBMCs has been shown to downregulate chemokine receptor expression and proinflammatory cytokine production, resulting in attenuated T-cell activation and highlighting the therapeutic potential of targeting histone acetylation pathways (85).

Beyond bulk PBMC analyses, cell-type-specific epigenetic profiling has revealed additional layers of complexity in rheumatoid arthritis. Using ATAC-seq across major peripheral immune cell populations, Zong et al. demonstrated that monocytes from RA patients display substantially increased chromatin accessibility compared with lymphocyte subsets, with preferential enrichment at regulatory regions associated with antigen presentation pathways and inflammatory cytokine genes (86). This epigenetically permissive chromatin landscape likely facilitates sustained innate immune activation and enhanced cytokine output, thereby shaping an inflammatory milieu that influences downstream adaptive immune responses. Importantly, epigenetic dysregulation is not confined to innate immune compartments but also critically affects T-cell fate decisions, contributing to imbalances between regulatory T cells and Th17 effector populations that are characteristic of RA. In line with this, a prospective study of 45 RA patients at baseline, 16 RA patients post-treatment, and 17 healthy controls demonstrated using flow cytometry that dysregulated acetylation and methylation correlate with disturbed Treg/Th17 balance and increased inflammatory responses, although heterogeneous treatment regimens and limited stratification by therapy may confound these findings (87).

Finally, emerging data indicate that histone acetylation is dynamically modulated by metabolic cues. Chandrasekaran et al. (preprint) reported that insulin signaling alters histone H3 acetylation in PBMCs from RA patients, linking immune activation with metabolic regulation. Although conceptually innovative, the study remains exploratory, as detailed cohort characterization and peer-reviewed validation are not yet available, limiting its immediate clinical interpretability (88).

Collectively, studies published between 2020 and 2025, summarized in Table 3, demonstrate that dysregulated histone acetylation is a shared feature across multiple immune cell types in RA. These markers differ substantially in procedural feasibility, sensitivity, and translational readiness. PBMC-based H3 acetylation offers accessibility but lacks longitudinal validation; monocyte ATAC-seq provides mechanistic depth but limited practicality; Treg/Th17-associated epigenetic markers capture treatment responsiveness but are confounded by therapeutic effects; and metabolic modulation of histone acetylation remains exploratory. A critical limitation across many studies is insufficient control for treatment status. In several cohorts, patients were sampled while receiving disease-modifying antirheumatic drugs, raising the possibility that observed epigenetic signatures partially reflect pharmacodynamic effects rather than primary disease pathology. This potential bias should be explicitly acknowledged as a limitation of the field and underscores the need for studies in treatment-naïve patients and standardized longitudinal designs.

In summary, histone modifications represent both mechanistic drivers of immune dysregulation and promising biomarker candidates in rheumatoid arthritis. Future translational progress will depend on harmonized analytical methodologies, rigorous control of clinical confounders including treatment exposure and integration of epigenetic profiling with longitudinal clinical outcomes and therapeutic response metrics. Such approaches will be essential to enable reliable disease monitoring and the development of personalized therapeutic strategies.

Synovial tissue data are included exclusively to provide mechanistic context for circulating biomarkers and are not presented as diagnostic evidence in this blood-focused review. Several limitations should be considered when interpreting this review. First, the current evidence base on epigenetic mechanisms and biomarkers in rheumatoid arthritis is still dominated by cross-sectional studies and relatively small observational cohorts, whereas clinical trials explicitly addressing epigenetic biomarkers with hard clinical endpoints (e.g., long-term structural damage, disability, or survival) remain scarce. Accordingly, conclusions regarding diagnostic performance, prognostic value, and treatment response prediction should be regarded as hypothesis-generating and require validation in large, prospective, and clinically well-phenotyped cohorts.

Second, this work represents a structured narrative review rather than a fully systematic review with meta-analysis; no formal standardized risk-of-bias assessment or quantitative synthesis was performed. As a result, the strength of evidence attributed to individual epigenetic biomarkers depends on the design, sample size, and methodological rigor of the original studies, which are heterogeneous and not directly comparable.

Third, the included studies employed a broad range of methodological approaches, including different epigenetic platforms (from targeted assays such as quantitative methylation-specific PCR to genome-wide arrays and sequencing technologies), diverse analytical pipelines, and various biological fractions (whole blood, plasma, serum, peripheral blood mononuclear cells, and specific immune cell subsets). This heterogeneity limits the feasibility of quantitative integration and may contribute to inconsistencies in reported effect sizes, cut-off values, and directions of change for some biomarkers.

Fourth, an important limitation of the current literature is the potential confounding effect of ongoing pharmacological treatment. Many primary studies included patients receiving disease-modifying therapies, and treatment status was not always consistently reported or controlled for. As a consequence, some reported epigenetic signatures may primarily reflect treatment-induced modulation rather than intrinsic disease-associated regulation, underscoring the need for treatment-naïve and longitudinal cohorts.

Finally, the literature search was restricted to articles written in English, indexed in PubMed/MEDLINE, Web of Science Core Collection, and Scopus, and published between 2020 and 2025. Relevant studies published in other languages, outside this time frame, or indexed in other databases may therefore have been missed. Moreover, most studies were not fully balanced in terms of sex, reflecting the higher prevalence of rheumatoid arthritis in women. Although patient and control groups were generally matched for key demographic variables, residual confounding related to age, sex, smoking status, therapy, and blood cell composition cannot be excluded and may limit the generalizability of the findings.

To identify ongoing or completed clinical studies related to epigenetic mechanisms in RA, we searched the ClinicalTrials.gov registry using the keyword “rheumatoid arthritis” (accessed in 2025) (191). The search was limited to studies registered between 2015 and 2025 and initially retrieved 1718 records tagged with “rheumatoid arthritis”. Titles and brief study descriptions were screened to identify trials that explicitly mentioned epigenetic modifications, epigenetic regulators or epigenetic biomarkers (e.g. DNA or RNA methylation, histone modifications, miRNAs or other ncRNAs). Trials not directly related to RA, studies withdrawn before enrolment and studies without any epigenetic endpoints were excluded. This process yielded seven clinical trials that focused on epigenetic alterations in RA: most evaluated changes in miRNA expression in response to therapeutic interventions, while one investigated a selective HDAC6 inhibitor. An overview of these trials is presented in Table 8. The limited availability of published results underscores that epigenetics in RA remains a promising yet still emerging research area requiring further comprehensive investigation.

The clinical study conducted by Adami et al. aimed, among other objectives, to determine changes in miRNA expression in PBMCs from RA patients in response to Filgotinib (192). Filgotinib is an orally administered JAK1 inhibitor that acts by suppressing JAK phosphorylation and inhibiting STAT3 activation. It is currently approved in Europe and Japan for the treatment of RA. Filgotinib demonstrates selective inhibition of JAK1-dependent cytokine signaling in vitro studies. It has been evaluated in phase 2 (DARWIN) and phase 3 (FINCH) clinical trials in patients with moderate to severe RA. In September 2020, the drug received marketing authorization in Japan and Europe. However, in August 2020, the U.S. Food and Drug Administration (FDA) requested additional data regarding the potential effects of filgotinib on semen parameters (193–195). To date, no data have been reported in the scientific literature concerning Filgotinib-induced alterations in miRNA expression levels in RA.

Conversely, the clinical study by Zissman et al. examined the impact of Tocilizumab treatment on miRNA expression (196). Tocilizumab is an interleukin-6 receptor antagonist that inhibits IL-6-dependent signaling, thereby reducing the expression of proinflammatory cytokines and enhancing the expression of genes associated with tissue repair in synovial fluid (197, 198). This drug has been extensively evaluated in clinical trials as part of combination therapies, administered both subcutaneously and intravenously, in patients with moderate to severe RA who exhibited an inadequate response to csDMARDs and/or bDMARDs. The studies confirmed that Tocilizumab represents an effective therapeutic option for adult patients with severe, active, and progressive RA who have not been previously treated with methotrexate. Moreover, the drug demonstrates high efficacy as a first-line biological therapy or in subsequent treatment lines for patients with moderate to severely active RA who have had an insufficient response or poor tolerance to at least one conventional synthetic DMARD (csDMARD) or TNF inhibitor (199–201). Additionally, a reduction in EMMPRIN/CD147 levels and an upregulation of circulating miR-146a-5p and miR-150-5p in plasma have been observed, resulting in a decreased angiogenic potential (202). However, no significant differences were found in the relative expression levels of miR-16, -21, -132, -146a, -150, -155, -203, -221, and -323. Only after stratifying patients into responders and non-responders based on CDAI and DAS28 scores was a significant change in miR-203 expression observed in the group treated with Tocilizumab for four months compared with the control group (203).

The only recent clinical study addressing histone modifications was a phase 2a trial sponsored by Chong Kun Dang Pharmaceutical investigating CKD-506, a selective HDAC6 inhibitor (204). It has been demonstrated that in RA, this compound suppresses the production of TNF-α and IL-6 in PBMCs derived from RA patients, as well as inhibits the production of MMP-1, MMP-3, IL-6, and IL-8 in FLS, thereby attenuating inflammatory responses and reducing arthritis severity (205).

Notably, to date, no large-scale clinical studies have been identified that specifically investigate DNA or RNA methylation in RA. An exception is the clinical trial NCT05808309 (206), which focuses on the identification of genomic biomarkers, including non-coding RNAs and DNA methylation profiles. However, this study remains at an early stage and involves several patient groups- individuals with RA with disease onset before 60 years of age, individuals over 65 years, patients with osteoarthritis without RA, and relatives of RA patients (both affected and unaffected). Consequently, it does not yet provide clinical data on the role of methylation in disease progression or treatment response.

Overall, while the topic of methylation in RA is being actively explored at the level of experimental and publication-based research, it is not yet well represented in clinical trials. In particular, there is a lack of clinical studies evaluating the effects of pharmacological treatments on methylation changes, which could yield valuable insights into therapeutic mechanisms and potential biomarkers of treatment response. This gap highlights a critical need for further targeted investigations into the epigenetic mechanisms underlying RA.

All reported AUC values are discussed descriptively and should not be interpreted as directly comparable across studies due to heterogeneity in analytical platforms, normalization strategies, and cohort composition. Across the epigenetic layers reviewed, reported AUC values provide a pragmatic, though incomplete, framework for assessing the translational relevance of blood-based biomarkers in RA. While AUC alone is insufficient to establish clinical utility, consistent performance above predefined thresholds, together with validation status and feasibility of blood-based assessment, allows stratification of candidates according to translational readiness.

Only a limited subset of biomarkers consistently reaches high diagnostic performance (AUC ≥ 0.85) and is supported by independent or multi-stage validation. Within DNA methylation studies, TNF-α, HIPK3, and selected CpG-based signatures involving CXCR5 and HTR2A demonstrate the most robust performance, including clinically meaningful gains in seronegative RA. Among non-coding RNAs, serum miR-223, miR-155, composite miRNA panels, and circRNA-based combined models, particularly hsa_circ_0000175/hsa_circ_0008410, show the strongest evidence of translational potential. Within the lncRNA layer, only NORAD and multi-lncRNA PBMC panels achieve comparable performance, indicating that panel-based strategies are generally superior to single-transcript approaches. A substantially larger group of epigenetic markers exhibits moderate discriminatory power (AUC ~0.70-0.85). These candidates-including individual miRNAs, selected circRNAs, and DNA methylation markers-are unlikely to function as standalone diagnostics but may provide incremental value when integrated into multivariate molecular-clinical models. Such hybrid approaches appear particularly relevant in seronegative RA, where conventional serological markers perform suboptimally. Most remaining epigenetic candidates fall within a discovery-phase category characterized by low AUC values, absence of ROC analysis, or reliance on small, cross-sectional cohorts. This group includes the majority of individual lncRNAs, piRNAs, Y-RNA-derived fragments, snoRNAs, snRNAs, and tRNA-derived fragments. While mechanistically informative, their current clinical relevance remains limited.

Interpretation of reported AUC values across studies is further constrained by methodological heterogeneity, lack of independent validation, cross-sectional designs, and insufficient control for treatment exposure. Notably, apparently high AUC values derived from small subgroup analyses-particularly in seronegative populations-should be interpreted with caution. Overall, these findings indicate that high AUC values alone do not ensure translational readiness. Epigenetic biomarkers most likely to progress toward clinical implementation are those that combine strong and reproducible diagnostic performance, feasibility in peripheral blood, and demonstrable added value beyond established serological tests. In RA, such biomarkers are currently confined to a narrow subset and are most likely to succeed as adjunctive tools addressing unmet needs, including seronegative disease, early detection, and treatment response stratification.

Recent studies confirm that the economic burden of RA remains substantial in the biologics and targeted-therapy era, with pharmacotherapy representing the dominant cost component in many health systems. US national survey data (MEPS 2018-2020) demonstrate higher annual healthcare expenditures among patients with RA, largely driven by prescription medication use (207). Consistently, a systematic review and meta-analysis estimated total direct medical costs of approximately $12,509 per year for RA overall and $36,053 per year among bDMARD users (208). Importantly, cost-of-illness estimates vary considerably across countries due to differences in pricing, reimbursement policies, wage levels, and social security systems, limiting cross-study comparability (208–212).

From a clinical perspective, standardization of diagnostic methods has direct implications for both patient management and healthcare resource utilization. Harmonized laboratory assays, imaging protocols, and uniform clinical assessment tools can reduce diagnostic variability and unnecessary repeat testing, thereby supporting earlier and more accurate treatment decisions (212). Such standardization is particularly important for reliable patient stratification prior to escalation to high-cost biologic therapies (213) and for the clinical implementation of novel biomarkers, including blood-based epigenetic markers (214).

Evidence from real-world models of early rheumatology access further demonstrates that turnaround time between symptom onset, initial assessment, and specialist consultation is a critical determinant of timely diagnosis and treatment initiation. Structured telephone triage and rapid access rheumatology clinics have been shown to shorten waiting times while maintaining diagnostic accuracy, enabling earlier identification of patients with inflammatory rheumatic disease (215, 216).

Beyond these organizational strategies, the feasibility of implementing novel diagnostic and stratification tools in routine rheumatology practice depends on system-level readiness and workflow integration. Hybrid implementation-effectiveness research highlights the importance of standardized data capture, stakeholder engagement, and alignment with existing care pathways to ensure successful translation beyond specialized centers (217).

Within this translational framework, cost-effectiveness represents a critical prerequisite for the clinical adoption of any novel biomarker. In RA, established serological tests such as RF and ACPA are inexpensive, widely available, and highly standardized, and they remain integral components of current diagnostic and prognostic algorithms. In particular, ACPA testing offers high diagnostic specificity and meaningful prognostic information at relatively low cost, setting a demanding benchmark for emerging biomarkers (7, 218, 219).

In contrast, most blood-based epigenetic biomarkers rely on complex laboratory workflows, including nucleic acid extraction, specialized molecular platforms, and often computationally intensive bioinformatic analyses. These requirements substantially increase per-sample costs and limit scalability compared with routine immunoassays such as RF or ACPA (220, 221). At present, there is no evidence demonstrating that epigenetic assays provide a cost-effective alternative to established serological tests in routine diagnostic pathways. Systematic reviews of health economic evaluations of diagnostic biomarkers highlight persistent gaps in economic evidence and methodological challenges in evaluating the cost-effectiveness of biomarker tests, including limited data on long-term health outcomes and resource use (222). This review does not perform formal cost-effectiveness modelling, as such analyses would require prospective health-economic inputs and standardized clinical endpoints that are currently unavailable for most blood-based epigenetic assays in rheumatoid arthritis.

Moreover, reported diagnostic performance metrics, including AUC values, cannot be directly compared across studies or translated into a unified clinical protocol, consistent with current regulatory and biomarker qualification frameworks (223, 224). The potential value of epigenetic biomarkers may therefore lie not in replacing existing low-cost serological assays, but in addressing specific clinical niches where conventional markers perform suboptimally. Such scenarios may include seronegative RA, very early or preclinical disease stages, patient stratification prior to escalation to high-cost therapies, or prediction of treatment response. In these contexts, even relatively expensive biomarkers could become cost-effective if they lead to earlier diagnosis, improved treatment allocation, or avoidance of ineffective and costly therapies. However, formal cost-effectiveness analyses integrating analytical costs, clinical performance, and downstream healthcare utilization are currently lacking and should be incorporated into future prospective validation studies.

From a regulatory perspective, biomarkers intended for clinical use must comply with formal qualification and approval frameworks. Regulatory agencies such as the European Medicines Agency (EMA) and the U.S. Food and Drug Administration (FDA) distinguish between biomarker discovery for research purposes and biomarkers intended to support clinical or regulatory decision-making (“Qualification of novel methodologies for medicine development | European Medicines Agency (EMA),” 2020). In the European Union, the regulatory landscape has evolved substantially with the introduction of Regulation (EU) 2017/746 on in vitro diagnostic medical devices (IVDR), which defines requirements for analytical validity, clinical validity, and clinical utility of in vitro diagnostic tests, including genetic and molecular assays (Regulation (EU) 2017/746 of the European Parliament and of the Council of 5 April 2017 on in vitro diagnostic medical devices and repealing Directive 98/79/EC and Commission Decision 2010/227/EU (Text with EEA relevance.), 2017). Recent regulatory science literature emphasizes that epigenetic biomarkers, if intended for clinical application, must be supported by robust analytical performance data, standardized assay methodologies, and clear evidence of clinical relevance within a predefined context of use (224).

A critical limitation for blood-based epigenetic biomarkers in rheumatoid arthritis is the absence of systematic validation pipelines, Most published studies on blood-based epigenetic biomarkers in rheumatoid arthritis report statistically significant associations between specific DNA methylation patterns or non-coding RNA signatures and disease presence or activity; however, they rarely include independent validation cohorts, inter-laboratory reproducibility testing, or prospective study designs (183, 225). Regulatory frameworks clearly distinguish between biomarker discovery and biomarker qualification, the latter requiring a structured validation process linking the biomarker assay to clinically meaningful endpoints. Only a small fraction of candidate biomarkers across medical disciplines successfully complete this transition, largely due to insufficient analytical standardization, limited clinical validation, and lack of prospective evidence demonstrating clinical utility in real-world settings (223). Furthermore, the complexity of epigenetic assays poses additional challenges for regulatory acceptance. Differences in biological material (whole blood, plasma, serum, or isolated immune cell populations), pre-analytical handling, and analytical technologies (e.g., qMSP, sequencing-based approaches, or array-based platforms) significantly affect comparability between studies. These factors complicate the definition of standardized cutoff values and performance characteristics required for regulatory approval. As highlighted by regulatory guidance documents, early interaction with regulatory authorities and clearly defined validation strategies are essential to bridge the gap between exploratory epigenetic research and clinically implementable diagnostic assays (223, 224, 226).

At present, no published formal cost-effectiveness studies exist for blood-based epigenetic biomarkers in RA, although several authors note the substantial resource requirements for nucleic acid extraction, specialized molecular platforms, and computational bioinformatic analyses (220, 221). These requirements increase per-sample costs and limit scalability compared with standard serological assays such as RF or ACPA. Therefore, the potential value of epigenetic biomarkers may lie not in replacing existing low-cost serological tests, but in addressing specific clinical niches where conventional markers perform suboptimally-such as seronegative RA, very early or preclinical disease stages, patient stratification prior to escalation to high-cost therapies, or prediction of treatment response. In these contexts, even relatively expensive biomarkers could become cost-effective if they enable earlier diagnosis, improved treatment allocation, or avoidance of ineffective therapies.

Therefore, despite encouraging discovery-phase data, the translation of blood-based epigenetic biomarkers into routine clinical practice in rheumatoid arthritis remains substantially limited. This limitation reflects the absence of clearly defined regulatory and validation pathways, pronounced methodological heterogeneity, lack of standardized analytical and bioinformatic pipelines, insufficient prospective clinical studies, and unresolved regulatory and health-economic challenges. The current evidence base is dominated by observational, retrospective, or exploratory studies, frequently conducted in single centers with relatively small and heterogeneous cohorts, and rarely incorporating independent validation cohorts or prospective clinical endpoints. As a result, even epigenetic biomarkers demonstrating promising diagnostic performance metrics, such as high AUC values, sensitivity, or specificity in cross-sectional analyses, should at present be regarded as hypothesis-generating discovery-phase tools rather than clinically actionable diagnostic or prognostic assays. Bridging the substantial gap between biomarker discovery and routine clinical implementation will require coordinated, multidisciplinary efforts that integrate rigorous analytical and clinical validation, early regulatory alignment, implementation science frameworks, and formal health-economic evaluation to ensure feasibility, scalability, and clinical utility in real-world rheumatology practice.

Blood-based epigenetic biomarkers in rheumatoid arthritis represent a promising adjunct to current serological and inflammatory markers, as they are minimally invasive, repeatable, and mechanistically linked to immune dysregulation. In this structured narrative synthesis (63 publications, Tables 2-8) and registry review (7 clinical trials, Table 1), the most translationally plausible signals arise from a limited subset of candidates, notably selected DNA methylation markers (including TNF-α, CXCR5, and HIPK3) and circulating non-coding RNAs such as miR-223, miR-155, miR-146a, and the lncRNA MEG3, whereas the majority of reported biomarkers remain at the discovery stage. The principal barrier to clinical implementation is substantial heterogeneity across studies including biospecimen type, analytical platforms, normalization strategies, cohort design and frequent incomplete control for treatment exposure, which can confound epigenetic readouts. Future progress will require standardized pre-analytical procedures and reporting, multicenter prospective validation (particularly in treatment-naïve, early, or at-risk cohorts), and a clear definition of context of use (diagnosis, disease monitoring, or therapy response) to demonstrate incremental value and real-world feasibility in rheumatology care pathways.