Genetic Polymorphisms in Rheumatoid Arthritis: From GWAS-Identified Loci to Functional, Biomarker, and Pharmacogenomic Insights

Rheumatoid arthritis (RA) is a chronic autoimmune disease with a complex, multifactorial aetiology, in which both environmental factors and genetic predisposition play a significant role. This narrative review summarises the genetic polymorphisms associated with RA that have been identified by genome-wide association studies (GWAS). Its aim is to synthesise the loci and single-nucleotide polymorphisms (SNPs) reported to date, to relate them to the immunological mechanisms through which they are thought to act, and to assess their emerging diagnostic, prognostic, and pharmacogenomic applications. Its scope covers large-scale GWAS and multi-ancestry meta-analyses together with the candidate-gene, functional, and pharmacogenomic studies required for their interpretation. Three principal conclusions emerge from the literature reviewed. First, the genetic architecture of RA remains dominated by the HLA-DRB1 shared epitope, which is associated primarily with seropositive, anti-citrullinated protein antibody-positive disease, whereas the numerous non-HLA loci identified to date each confer only a modest individual effect. Second, these non-HLA variants converge on a limited number of biological processes—the activation thresholds of T and B lymphocytes, NF-κB and JAK–STAT cytokine signalling, and the invasive behaviour of fibroblast-like synoviocytes—so that their value lies less in the prediction of risk by individual variants than in identifying pathways amenable to therapeutic targeting. Third, the most readily translatable findings are pharmacogenomic. Variants affecting adenosine signalling and nucleotide metabolism modulate the response to methotrexate. Several polymorphisms predict the response to TNFα inhibitors. Polygenic risk scores are beginning to support the differential diagnosis of early inflammatory arthritis. Progress remains constrained by the over-representation of European-ancestry cohorts, by the difficulty of moving from an associated locus to a causal variant and its target gene, and by the limited replication of pharmacogenomic associations. Integration of genomic with transcriptomic, epigenomic, proteomic, and metabolomic data, supported by machine learning approaches, is therefore required before genotype-guided management of RA can enter routine clinical practice.