OA10 Systematic review and independent validation of genetic factors of radiographic outcome in rheumatoid arthritis identifies a genome-wide association with CARD9

HLA polymorphisms associated with rheumatoid arthritis (RA) susceptibility are also associated with severity. However, emerging evidence suggests that HLA-independent genetic mechanisms underpinning susceptibility to disease and its progression are different. Despite a plethora of reports on genetic associations with radiographic outcome in RA, the characterisation of biological mechanisms leading to poor outcome remains elusive, as most studies fail replication. Our goal was to identify non-HLA loci that warrant prioritization for future research and clinical application through 1) systematic review of all reports on radiographic severity in RA across ethnicities, 2) independent validation in a large cohort, 3) gaining biological insights into mechanisms of disease prognosis.

Systematic review was conducted in line with STROBE guidelines. The Norfolk Arthritis Register (NOAR) recruited 1884 RA patients and followed them prospectively for up to 20 years with a total of 2198 longitudinal radiographs of hands and feet. Genome-wide genotyping was performed followed by imputation. Association testing with longitudinal outcome was performed with generalized estimating equation models for the presence of erosions and with generalised linear latent and mixed model with discrete random effects and 3 latent classes for Larsen score. Fine mapping of regions of interest was performed, together with functional annotations and estimation of functional impact (FUMA, PolyPhen2 and RegulomeDB).

Systematic review identified 243 publications reporting a total of 142 independent non-HLA associations with radiographic outcome. In NOAR, 10 independent SNPs (∼7%) replicated with consistent direction of effect to the original publication. Among those, rs59902911, located in exon 5 of CARD9, a locus not previously reported to be involved in RA susceptibility, was part of a 95% credible SNP set comprising 30 SNPs. The lead SNP (rs78892335) reached genome-wide significance with an effect size more than twice the minimal clinically important difference for each copy of the minor allele (4.78 Larsen unit/copy, confidence interval 3.15;6.41, p = 9.01 × 10^⁻9) and this association was independent of anti-CCP status. This intronic SNP located in the nearby ENTR1 gene is an expression quantitative trait locus (eQTL) for expression of CARD9 in B, CD4⁺ and CD8⁺ T cells, monocytes and neutrophils. CARD9 plays an integral role in the immune system through upregulation of inflammatory cytokines such as IL-6, TNF-α and IFN-γ, and processes such as maturation of antigen presenting cells, activation of T cells, neutrophil recruitment and polarisation of macrophages.

We provide a comprehensive list of validated genetic associations with RA outcome and demonstrate that non-HLA severity polymorphisms, despite lack of association with disease susceptibility, can reach genome-wide significance and point towards genetic mechanisms of radiographic outcome independent of the HLA and anti-CCP status. This highlights the importance of dedicated outcome studies in identifying genetic markers for patient stratification in precision medicine for RA.

S.D. Sharma: None. R. Hum: None. N. Nair: None. L. Marshall: None. J. Bowes: None. A. MacGregor: None. S.M. Verstappen: None. A. Barton: None. H. van Steenbergen: None. R. Knevel: None. A. van der Helm van Mil: None. S. Viatte: None.