Background Systemic lupus erythematosus (SLE) is a complex autoimmune disease. High monozygotic twin discordance implies a role for non-genetic effects in disease pathogenesis. Previous epigenome-wide association studies (EWAS) have demonstrated a role for DNA methylation in SLE, but different immune cell subsets have so far been insufficiently characterised. The disease discordant twin model is a powerful design for detecting SLE-associated DNA methylation variation.

Objectives To investigate genome-wide DNA methylation changes in sorted CD4+ T-cells, monocytes, granulocytes and B-cells from twin pairs with at least one SLE-affected twin.

Methods Altogether 15 SLE twin pairs participated in the study, of which 6 pairs were monozygotic (MZ) and 9 were dizygotic (DZ), including 2 concordant pairs (1 MZ, 1 DZ). Disease activity was evaluated using the SLE disease activity index (SLEDAI). Peripheral blood was processed using gradient density centrifugation for the granulocyte fraction and the mononuclear cell fraction was sorted serially for CD14+ monocytes, CD4+ T-cells and B-cell enrichment using a RoboSep device (Stemcell Technologies). DNA was extracted using the DNA/RNA/miRNA Universal kit (Qiagen). Genome-wide DNA methylation was evaluated using the Infinium HumanMethylation450K BeadChip (Illumina). Paired analyses were performed using a Wilcoxon test with p<0.01 and median differential methylation >7% considered as statistically significant. Top differentially methylated genes were validated using pyrosequencing.

Results In paired analyses of discordant SLE twins without restriction to probe category, we found 176, 510, 393 and 2882 differentially methylated CpGs in CD4+ T-cells, monocytes, granulocytes and B-cells, respectively. Restricted to the promoter and transcription start sites, there were 55, 327, 247 and 1628 genes in CD4+ T-cells, monocytes, granulocytes and B-cells with differentially methylated CpGs, respectively. In all cell types, there was a profound hypomethylation of interferon-regulated genes, including IFI44L, DTX3L, PARP9 and IFITM1, which was more pronounced in twins with recent flare within the past 2 years. In contrast to the other cell types, hypermethylation was predominantly observed in B-cells. Using Ingenuity Pathway Analysis, the top upstream regulators of hypermethylated genes in B-cells were TNF, miR-146a-5p and EP300. Hypomethylation of CpGs was validated in all cell types at the genes IFI44L, PARP9, IFITM1, LGALS3BP, LOXL1, MIR10A, PLSCR1 and RSAD1, in B-cells and granulocytes at TACSTD2 and at MIR146B in B-cells. Hypermethylation of CpGs was validated in B-cells alone for CXCR5, DDR1 and TRAF5.

Conclusions Robust hypomethylation of interferon-regulated genes is common to all major cellular compartments in SLE twins. The finding of hypermethylated CpGs in B-cells is novel and might be of interest to SLE pathogenesis.

Disclosure of Interest None declared