Increased DNA methylation variability in rheumatoid arthritis-discordant monozygotic twins

Abstract

Background: Rheumatoid arthritis is a common autoimmune disorder influenced by both genetic and environmental factors. Epigenome-wide association studies can identify environmentally mediated epigenetic changes such as altered DNA methylation, which may also be influenced by genetic factors. To investigate possible contributions of DNA methylation to the aetiology of rheumatoid arthritis with minimum confounding genetic heterogeneity, we investigated genome-wide DNA methylation in disease-discordant monozygotic twin pairs.

Methods: Genome-wide DNA methylation was assessed in 79 monozygotic twin pairs discordant for rheumatoid arthritis using the HumanMethylation450 BeadChip array (Illumina). Discordant twins were tested for both differential DNA methylation and methylation variability between rheumatoid arthritis and healthy twins. The methylation variability signature was then compared with methylation variants from studies of other autoimmune diseases and with an independent healthy population.

Results: We have identified a differentially variable DNA methylation signature that suggests multiple stress response pathways may be involved in the aetiology of the disease. This methylation variability signature also highlighted potential epigenetic disruption of multiple RUNX3 transcription factor binding sites as being associated with disease development. Comparison with previously performed epigenome-wide association studies of rheumatoid arthritis and type 1 diabetes identified shared pathways for autoimmune disorders, suggesting that epigenetics plays a role in autoimmunity and offering the possibility of identifying new targets for intervention.

Conclusions: Through genome-wide analysis of DNA methylation in disease-discordant monozygotic twins, we have identified a differentially variable DNA methylation signature, in the absence of differential methylation in rheumatoid arthritis. This finding supports the importance of epigenetic variability as an emerging component in autoimmune disorders.

Keywords: Autoimmune disease; DNA methylation; Epigenetics; Rheumatoid arthritis; Twins.

Fig. 1

Overview of study design. Rheumatoid arthritis-discordant twin pairs were recruited from the RA twins study in Manchester and TwinsUK in London, and genome-wide DNA methylation was investigated in the context of both differentially methylated positions and differentially variable positions

Fig. 2

CpG plots for six top ranked differentially variable positions in RA-discordant MZ twins. Cpg sites shown are cg11374732 (Bartlett’s test p value = 4.09E−06), cg01999539, cg23280983, cg20500144, cg26985354 and cg26827503. Hypervariability of differentially variable positions was enriched in RA twins. Boxplots indicating the mean methylation and range of methylation values are shown overlaid with scatterplots indicating DNA methylation measurements of individual samples

Fig. 3

Feature enrichment for RA-associated differentially variable positions which are hypervariable in healthy and RA-affected twins respectively. Differentially variable positions which are hypervariable in the healthy twins were enriched in CpG island shelves and non-CpG island-associated regions, and in the bodies of genes. Meanwhile, differentially variable positions enriched in RA-affected twins were enriched in 3′UTRs of genes, gene bodies and non-CpG island-associated regions

Fig. 4

Variance and range for DVPs which were hypervariable in RA co-twins. Variance and range were calculated for each of the 763 DVPs found to be hypervariable in RA co-twins, plotted for three comparison groups; RA co-twins, healthy co-twins and an independent cohort of healthy individuals

Fig. 5

Prevailing and suggested models of RA disease development. The current model of RA aetiology primarily involves the contribution of genetic and non-genetic factors including triggers such as infection, which mediate an immune response and ultimately contribute to the clinical manifestation of RA. Our suggested model of RA development incorporates the potential role of DNA methylation alongside the contributing factors of oxidative and cellular stress