Disclosure(s): No financial relationships with ineligible companies to disclose
Background/Purpose: The aetiopathogenesis of SLE encompasses genetic and epigenetic factors, including hypomethylation of type I interferon (IFN) regulated genes (1) and the HLA-DRB1*03:01 haplotype, linked to C4A copy numbers and associated with anti-SSA/SSB antibodies (2). Further, a non-HLA SLE polygenic risk score (SLE-PRS) has been linked to a more severe disease (3). This study aimed to investigate the relationship between a methylation-based SLE risk score (MRS), an SLE-PRS, an HLA-DRB1*03:01 tag SNP, serum IFN-α levels and clinical phenotype. Methods: DNA methylation levels in whole blood from patients fulfilling ≥4 ACR-82 criteria (n=547) and controls matched for age and sex (n=587) were investigated by the Illumina HM450K methylation array. Significant differentially methylated CpGs with a Δβ of ≥0.1 between cases and controls and located at independent genetic loci (n=17) were selected for calculation of the MRS. Clinical data were collected from patient charts. Serum IFN-α2 levels were measured in a subgroup of patients (n=85) using Simoa.
We analysed if the MRS was associated with the SLE-PRS including 57 non-HLA SLE risk loci (3), with the HLA-DRB1*03:01 tag SNP rs1269852 and with levels of IFN-α using linear regression models. Associations between the MRS, SLE-PRS and HLA-DRB1*03:01 tag SNP and clinical variables were assessed applying logistic regression. Models were adjusted for age and sex, p< 0.05 was considered significant. Results: The MRS was associated with high disease activity (SLEDAI >4 or SLAM >6 ) (OR=1.01, p=2.0×10-4), higher IFNα levels (Exp(B)=1.03, p=8.1x10-13) and low complement (OR=1.01, p=2.2×10-3), while the SLE-PRS was not (all p >0.05). Further, no significant correlation between the MRS and the SLE-PRS was observed (p >0.05).
The MRS was associated with discoid rash (OR=1.02, p=3.5x10-4), hematologic criteria (OR=1.02, 1.6x10-7) and immunologic criteria (OR=1.01, p=1.4x10-4), while the SLE-PRS was associated with the nephritis (OR=1.19, p=0.04) and the immunologic criteria (OR=1.40, p=1.1x10-4). Further, the MRS, but not the SLE-PRS, associated with anti-SSA and anti-RNP antibodies (OR=1.03, p=1.1x10-12 and OR=1.02, p=8.2x10-11).
The HLA-DRB1*03:01 tag SNP was significantly associated with higher MRS (B=7.22, p=8.6x10-4) and with anti-SSA and anti-SSB antibodies (OR 3.09 and 4.75, both p< 1x10-10). Conclusion: The lack of correlation between the MRS and SLE-PRS suggests that they reflect distinct, yet complementary, aspects of SLE pathogenesis. The association between the MRS, high disease activity, higher levels of IFN-α2 in serum, anti-SSA antibodies and the HLA-DRB1*03:01 tag SNP implies that an epigenetic predisposition for IFN-α signalling might be of particular importance in HLA-DRB1*03:01 positive patients. These results underscore the relevance of integrating epigenetic profiling with genetic data to better understand disease heterogeneity and mechanisms in SLE.
