Background/Purpose: In the general population, an increase in chronological age is associated with differential DNA methylation of particular CpGs in a highly conserved fashion. Since age is a key risk factor for many rheumatic diseases, we tested whether DNA methylation (DNAm) patterns exhibited during normal aging occur prematurely in subjects with systemic autoimmune rheumatic diseases (SARD).

Methods: Peripheral blood CD4⁺T-cells and CD14⁺monocytes were sorted from incident treatment naïve rheumatoid arthritis (RA, n=13) and systemic sclerosis (SSc, n=17) subjects, as well as systemic lupus erythematosus subjects (SLE, n=12 – not all of whom were treatment naïve), and healthy controls (HC, n=8). Illumina HumanMethylation450 BeadChip and Illumina TruSeq stranded RNA-seq were used to perform genome-wide methylome and transcriptome analysis. First, we compared chronological and predicted age across SARDs using the Horvath calculator, a publicly available ‘epigenetic-aging-signature’ based on DNAm levels at 353 CpG sites (Genome Biology 2013). Second, using principal component of explained variance (PCEV), we examined whether the 353 CpGs were associated with chronological age in treatment-naïve RA and SSc, separately. Finally, we verified the findings from the PCEV analysis by examining the methylation levels in RA and SSc at sites that have been reported to correlate between chronological age and methylation levels (Lin and Wagner, PLOS Genetics2015).

Results: Across SARDs, the estimated age predicted using the Horvath calculator was younger than chronological age in both T-cells (average age acceleration -6.04 years, p=1.4E-12) and monocytes (average age acceleration -1.2 years, p=0.023). We found similar results in SSc [T-cells average age acceleration -7.91 years (p=7.36E-09), and monocytes average age acceleration -2.51 years (p=0.015)], and in RA T cells (average age acceleration -6.49 years, p=6.59E-0.4). We were unable to establish differences in predicted and chronological age in RA monocytes and SLE T cells and monocytes. PCEV analysis showed that the 353 CpGs in the Horvath model correlated with age in SSc (p value 0.002), but not in RA (p 0.268). We verified the findings of the PCEV analysis by examining methylation levels at CpGs reported to be correlated with aging (97 hyper- and 402 hypomethylated). Again, the CpG sites correlated with age as expected in SSc. However, while the correlations with hypermethylated CpGs were as expected in RA, this was not the case with hypomethylated CpGs, where evidence of more methylation than expected in both T cells and monocytes was found.

Conclusion: The models that predict aging-associated epigenetic drift in the general population did not predict premature aging in our SARDs patients. Moreover, age-associated DNAm patterns at hypomethylated CpG sites seem to be coherently modified in RA. This provides additional evidence supporting dysregulated aging in RA and novel mechanistic clues.

« Back to 2016 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/dysregulation-of-hypomethylated-age-related-cpg-sites-characterize-t-cells-and-monocytes-from-treatment-naive-rheumatoid-arthritis-subjects/