1. Introduction

Fibromyalgia (FM) is a common chronic pain disorder characterized by widespread musculoskeletal pain, fatigue, sleep disturbances, and cognitive symptoms. The pathophysiology of fibromyalgia is not fully understood, but it is believed to involve a combination of genetic, environmental, and neuroimmune factors leading to central sensitization of pain. Recent research has pointed to an epigenetic component in chronic pain conditions, with DNA methylation changes observed in patients with various chronic pain syndromes ^[1]. In fibromyalgia, emerging evidence suggests that altered DNA methylation patterns may correlate with the condition and could potentially serve as biomarkers. For example, an epigenome-wide study identified several “methylation factors” associated with fibromyalgia, notably in genes like APBB2, AKAP12, and CD38, which showed promise as diagnostic markers ^[2]. However, many earlier studies have been limited by small sample sizes and have found only modest differences – in one case, out of several candidate CpG sites tested, only a single site (in the GCSAML gene) remained significant after multiple-comparison correction ^[3].

Given the need for objective biomarkers for fibromyalgia, we set out to investigate DNA methylation on a genome-wide scale in a relatively large cohort. In this study, we focused on saliva-derived DNA, an easily obtainable sample, to see if fibromyalgia patients exhibit a distinct methylation signature compared to healthy controls. Saliva contains genomic DNA from leukocytes and epithelial cells and has been successfully used in epigenetic studies as a proxy for systemic methylation patterns. We utilized the Illumina Infinium MethylationEPIC array, which covers >850,000 CpG sites across the genome, to perform an unbiased search for differentially methylated positions (DMPs) in fibromyalgia. Our analysis identified a set of five autosomal CpG sites with highly significant methylation differences between fibromyalgia patients and controls. We further explored whether the mean methylation level of these top five CpGs could serve as a combined epigenetic indicator distinguishing fibromyalgia, which might be more robust than any single CpG alone. Here, we report that this five-CpG panel – and particularly its average methylation level – is markedly different in fibromyalgia patients, suggesting a potential DNA methylation biomarker for the disorder.

2. Materials and Methods

Study Participants: This study included two groups of participants. The fibromyalgia focus group consisted of 53 individuals (78% female) who had been clinically diagnosed with fibromyalgia. Their mean age was 43 years (standard deviation ±2 years), and none had any other known pathology or chronic condition. The control complement group comprised 1619 individuals (58% female, mean age 44 ± 2.3 years) with no fibromyalgia diagnosis and no known pathological conditions (i.e., essentially healthy controls). All participants provided saliva samples for DNA analysis. Informed consent and ethical approvals were in place for the collection and use of these samples (details omitted for brevity).

Sample Collection and DNA Methylation Profiling: Unstimulated saliva was collected from each participant using standard collection kits, and genomic DNA was extracted from the saliva samples. DNA methylation was profiled using the Infinium MethylationEPIC BeadChip array (Illumina, San Diego, CA, USA), following the manufacturer’s protocols. This array interrogates approximately 850,000 CpG sites across the human genome, covering promoter regions, gene bodies, enhancers, and other regulatory regions. Array data were processed using appropriate bioinformatics pipelines: briefly, raw intensity data were normalized and beta values (proportion of DNA methylation at each CpG site, ranging from 0 to 1) were calculated for each site. Quality control steps were applied to remove any poor-quality probes or samples. Importantly, all CpG sites located on the X chromosome were excluded from analysis a priori to avoid sex-related methylation biases, given the unequal sex distribution between groups. Only autosomal CpG sites were considered in the downstream analysis.

Statistical Analysis: To identify differentially methylated CpG sites between fibromyalgia patients and controls, we compared the distribution of beta values for each site in the fibromyalgia group versus the control group. Because of the non-normal distribution of beta values and the large disparity in sample size between groups, we employed the Mann–Whitney U test (a non-parametric rank-sum test) to assess whether methylation levels at each CpG differed significantly between the two groups. The threshold for statistical significance was set at p < 0.0001 (0.01% significance level) for an initial screen of top hits. This stringent cutoff was chosen to highlight the most robust differences, given the large number of CpG sites tested. For those sites meeting the threshold, we also calculated false discovery rate (FDR) q-values to account for multiple testing across the genome. Only sites with p < 1×10^−4 that were also not located on sex chromosomes were retained as candidates.

After identifying the top differential CpG sites, we calculated a composite methylation score for each individual by taking the mean beta value across these top five CpG sites. This composite represents the average methylation level (fraction methylated) of the five significant CpGs in that individual. We then compared this composite score between fibromyalgia patients and controls using the Mann–Whitney U test to evaluate whether the combined measure improved the distinction between groups. Summary statistics (mean, median, interquartile range) for this composite in each group were computed. 

3. Results

Differentially Methylated CpG Sites in Fibromyalgia: Out of ~850,000 CpG sites analyzed on the EPIC array, five autosomal CpGs showed highly significant differences in DNA methylation between the fibromyalgia and control groups at the p < 0.0001 threshold. Table 1 summarizes these five CpG sites, their genomic annotation, and the methylation levels in each group. All five sites displayed lower methylation (hypomethylation) in fibromyalgia patients compared to controls. The mean beta values in the fibromyalgia group ranged from ~0.05 to 0.32 at these loci, whereas in controls the means ranged from ~0.09 to 0.39, with differences (Control minus Fibromyalgia) on the order of 3.5% to 7.2% absolute methylation (beta) units. Despite these modest magnitude differences, the large control sample size provided high statistical power, and p-values were on the order of 10^−5 or smaller for all five sites. After adjusting for multiple testing, each of these CpGs remained significant with FDR q-values below 0.01, indicating that they represent true positive differential methylation signals.

Table 1. Five CpG sites with significantly different DNA methylation in fibromyalgia vs. control groups (saliva DNA). Mean beta values (proportion methylated) in each group are shown with the absolute difference. P-values are from Mann–Whitney U tests comparing the two groups. q-values are FDR-adjusted p-values.

Several of these differentially methylated CpGs map to genes of potential interest. cg15957394 is located in the promoter region of AFAP1 (Actin Filament Associated Protein 1). cg13150977 lies in the promoter of UBE2Q2P1 (also known as UBE2QP1), which is a pseudogene related to the ubiquitin-conjugating enzyme E2 Q2. cg25294185 is in the promoter of RNASEH2C, a gene encoding the C subunit of RNase H2 (an enzyme involved in nucleic acid metabolism and immune DNA sensing). The remaining two sites (cg12691488 on chromosome 1 and cg13348458 on chromosome 6) are located in intergenic regions with no annotated gene in their immediate vicinity (they may lie in distal regulatory elements or non-coding RNA regions). All five CpGs were hypomethylated in the fibromyalgia group relative to controls, as evidenced by the negative differences (Fibro < Control in beta value for each). Among these, the largest methylation gap was observed at cg15957394 in AFAP1 (approximately 7.2% lower in fibromyalgia), and cg12691488 (intergenic on chr1, ~7.0% lower in fibromyalgia). The smallest difference was at cg25294185 in RNASEH2C (~3.5% lower in fibromyalgia), though even this was highly significant given the sample size. Notably, each of these five CpGs passed the stringent significance threshold; no other autosomal CpGs outside these five met the p < 1×10^−4 criterion in our dataset, underscoring that widespread large methylation differences are not present – rather, only a few specific genomic loci show detectable differential methylation associated with fibromyalgia.

Composite 5-CpG Methylation Index: We next examined whether combining the information from these five top CpGs could enhance the distinction between fibromyalgia patients and controls. For each individual, we calculated the mean methylation (beta value) across the five significant CpG sites. This composite methylation index was markedly different between the two groups. Fibromyalgia patients had a mean 5-CpG methylation of 0.189, compared to 0.241 in controls – an absolute difference of 0.052 (5.2 percentage points). The median of the composite in the fibromyalgia group was 0.178 (with interquartile range 0.144–0.221), whereas the median in controls was 0.239 (IQR 0.191–0.282). Figure 1 illustrates the distribution of this 5-CpG mean methylation in each group, highlighting that although there is some overlap, the fibromyalgia patients tend to cluster at lower methylation values while controls are generally higher. A statistical comparison confirmed that this composite index differentiates the groups extremely well (p = 3.1×10^−9 by Mann–Whitney U test). In fact, the p-value for the composite was even more significant than for any single CpG site alone, suggesting that aggregating multiple CpG signals improved the signal-to-noise ratio.

Overall, the results indicate that fibromyalgia patients, as a group, exhibit a consistent pattern of DNA methylation differences at this set of five CpG loci, and that a summary measure of these loci can serve as a strong distinguishing feature between fibromyalgia cases and non-cases. This finding raises the possibility of developing a methylation-based biomarker or epigenetic “signature” for fibromyalgia diagnosis.

4. Discussion

In this study, we identified a distinct epigenetic signature in fibromyalgia patients’ saliva DNA. Five specific CpG sites were found to be significantly hypomethylated in fibromyalgia cases compared to controls, and a combined average of these sites provides a clear separation between the groups. To our knowledge, this is one of the first demonstrations of a saliva-based DNA methylation panel that differentiates fibromyalgia patients from healthy individuals with such a high level of statistical significance. These results support the notion that fibromyalgia involves measurable biological changes, in this case at the level of DNA methylation, that could potentially be harnessed as biomarkers.

The implicated CpG sites and their nearby genes offer clues to possible biological pathways involved in fibromyalgia. Notably, AFAP1 (actin filament-associated protein 1) showed the largest methylation difference. AFAP1 is an adaptor protein that binds to actin filaments and interacts with Src-family kinases, acting as a modulator of cytoskeletal dynamics in response to cellular signals. Intriguingly, AFAP1 has been identified as a mediator in inflammatory signalling pathways – for example, it was shown to be a key player in TNF-α induced signalling events in endothelial cells ^[4]. Hypomethylation in the AFAP1 promoter in fibromyalgia patients could imply upregulation of AFAP1 expression (since lower DNA methylation in promoter regions is often associated with increased gene transcription). If AFAP1 is upregulated, it might influence inflammatory signal transduction or cytoskeletal organization in immune cells or other cell types, which could in turn relate to fibromyalgia’s pathophysiology (potentially affecting immune responses or neural signalling given actin’s role in synaptic function). While speculative, this aligns with growing evidence that inflammatory and immune pathways are altered in fibromyalgia and related central sensitivity syndromes.

Another gene of interest is RNASEH2C, encoding a subunit of RNase H2. This enzyme complex is involved in clearing RNA-DNA hybrids and in DNA repair. Importantly, RNase H2 plays a role in preventing inappropriate immune activation: it helps remove nucleic acid by-products that, if accumulated, can trigger the innate immune system ^[5]. Mutations in RNASEH2C are known to cause Aicardi–Goutières syndrome, a severe autoimmune-inflammatory condition, due to accumulation of nucleic acids that activate interferon responses. In our data, fibromyalgia patients had significantly lower methylation in the RNASEH2C promoter (cg25294185), which could lead to higher expression of RNASEH2C. One could hypothesize that increased RNase H2 activity might alter the handling of nucleic acid debris and immune signalling in fibromyalgia. Although fibromyalgia is not classically considered an autoimmune disease, there is evidence of immune system dysregulation and chronic low-level inflammation in some patients. The epigenetic upregulation of a gene involved in nucleic acid immune sensing fits into a broader theme that fibromyalgia may involve subtle immune/inflammatory component or response to cellular stress. This warrants further functional investigation.

The site cg13150977 in the UBE2Q2P1 gene (also called UBE2QP1) is also of interest. UBE2Q2P1 is a pseudogene related to the ubiquitin-conjugating enzyme E2 Q2. Pseudogenes do not code for functional proteins, but they can have regulatory roles, such as producing non-coding RNAs that regulate their protein-coding counterparts or other genes. The hypomethylation of the UBE2Q2P1 promoter might indicate active transcription of this pseudogene. While the direct impact of UBE2Q2P1 in fibromyalgia is unclear, the ubiquitin-proteasome system is fundamental to protein turnover and cellular stress responses. It is conceivable that changes in ubiquitin pathway regulation (even via pseudogene transcripts) could influence muscle or neuronal cell function in fibromyalgia. Further research would be needed to clarify this link.

The remaining two significant CpGs are intergenic (cg12691488 on chromosome 1 and cg13348458 on chromosome 6). These sites might lie in regulatory regions such as enhancers or within non-coding RNAs that were not annotated in our dataset. It is possible they exert regulatory effects on nearby genes or have structural roles in chromatin. Interestingly, despite being intergenic, these sites showed some of the largest methylation differences (around 7% lower in FM for the chr1 site). This suggests that they are non-random hits and could mark important regulatory hotspots. Future work, such as examining chromatin state or gene expression quantitative trait loci (eQTLs) in those regions, could illuminate whether these intergenic methylation changes have downstream effects on gene activity.

An important aspect of our findings is the use of a composite methylation score. By averaging the methylation levels of the five CpGs, we effectively created an epigenetic biomarker index for fibromyalgia. This composite was extremely significantly different between patients and controls (p ~10^−9). In practical terms, if one were to use this 5-CpG mean as a diagnostic indicator, fibromyalgia patients tended to have values below ~0.20, whereas most controls had values above ~0.20 (with medians ~0.18 vs ~0.24, respectively). There was still some overlap between the lowest-methylation controls and highest-methylation patients, which is expected given biological variability. However, the separation was sufficient that one could envision applying a threshold on this methylation score to classify new samples with a reasonable degree of accuracy. In a preliminary sense, this suggests potential diagnostic utility – for instance, a receiver operating characteristic (ROC) analysis could be performed in future studies to estimate the sensitivity and specificity of the methylation panel for fibromyalgia identification. We note that currently fibromyalgia diagnosis is clinical, based on symptom criteria and exclusion of other conditions; thus, an objective biomarker (or biomarker panel) would be highly valuable to support diagnosis or even to subclassify patients. Our results move in that direction by providing a set of candidate epigenetic markers.

It is also worth discussing the fact that we removed X-chromosome probes from analysis and the rationale behind it. The fibromyalgia group in our study had a higher proportion of females than the control group (78% vs 58%). Since DNA methylation patterns on the X chromosome differ greatly by sex (e.g., X-inactivation in females), including X-linked CpGs could lead to false positives driven by the sex imbalance rather than fibromyalgia status. By excluding X-linked sites, we ensured that the significant findings are not due to sex chromosome differences. 

Our study has several strengths, including the use of a genome-wide platform and a very large control sample which enhances statistical power. The saliva-based approach is non-invasive and could be easily reproduced or applied in clinical settings if validated. However, there are also limitations to consider. First, the fibromyalgia sample size (n=53) is moderate and from a single cohort; larger case numbers from diverse populations would bolster confidence in these findings. Second, saliva DNA methylation is a surrogate for potential changes in blood or tissues – the signal we detect likely originates from blood leukocytes present in saliva, as well as oral epithelial cells. These cell types may not fully reflect methylation changes in muscle or central nervous system tissues that are directly involved in fibromyalgia pathology. Nonetheless, systemic epigenetic alterations could mirror or contribute to the overall disease state. Third, while the association is clear, causality cannot be inferred. We do not know if the methylation differences are causing or resulting from fibromyalgia (or an upstream factor such as chronic stress or pain). Longitudinal studies could determine if these methylation marks precede disease onset or change with treatment, which would indicate a functional role. Finally, although the composite methylation score differentiates patients and controls on a group level, it is not a perfect diagnostic tool on its own – there is overlap and one would need to establish a formal cut off and validate its performance (sensitivity, specificity) in independent samples before considering clinical application.

Despite these caveats, our findings contribute to the growing body of evidence that fibromyalgia has an epigenetic fingerprint. Previous epigenetic studies in fibromyalgia and related disorders have reported differential methylation in genes related to the nervous system, immune function, and stress response . Our results add new candidate genes (e.g., AFAP1, RNASEH2C) to this list, highlighting pathways involving cytoskeletal signalling and nucleic acid metabolism/immune activation that have not been widely linked to fibromyalgia before. Interestingly, these pathways resonate with some hypothesized mechanisms in fibromyalgia – for instance, neuroinflammation and glial activation have been proposed in central sensitization, and here we find epigenetic changes in an inflammation-related gene (AFAP1) and an immune DNA-sensing enzyme (RNase H2). This opens up avenues for further research: investigating whether these epigenetic changes correspond to altered gene expression in patient immune cells, or if they correlate with clinical features (pain severity, fatigue levels, etc.). If confirmed, such methylation markers could also potentially be used to monitor disease progression or responses to therapy (epigenetic marks are generally stable, but certain interventions or lifestyle changes might normalize pathological methylation profiles).

5. Conclusion

In summary, our epigenome-wide analysis of saliva DNA identified a reproducible methylation signature associated with fibromyalgia. Five specific CpG sites across the genome showed significantly lower methylation in fibromyalgia patients compared to controls, and a composite average of these five sites yielded an even stronger separation between the groups. This suggests that fibromyalgia is accompanied by coordinated epigenetic alterations that can be detected peripherally. The involved genes – such as AFAP1, RNASEH2C, and a ubiquitin-pathway pseudogene – hint at underlying mechanisms related to inflammatory signalling and immune regulation in fibromyalgia. While preliminary, these findings raise the possibility of developing a methylation-based biomarker panel to assist in fibromyalgia diagnosis, a condition that currently lacks objective laboratory tests. Future studies should aim to replicate these results in independent cohorts (including blood-based methylation analyses), explore the functional consequences of these epigenetic changes, and evaluate the predictive power of the 5-CpG methylation index for clinical use. Ultimately, a better understanding of fibromyalgia’s epigenetic landscape may not only improve diagnosis but also uncover novel targets for therapeutic intervention in this challenging chronic pain disorder.