Objective: Methotrexate (MTX) is the drug of first choice for the treatment of rheumatoid arthritis (RA), but is effective only in around 60% of the patients. Identification of genetic markers to predict response is essential for effective treatment within a critical window period of 6 months after diagnosis, but have been hitherto elusive. In this study, we used genome-wide genotype data to identify the potential risk variants associated with MTX (poor)response in a north Indian RA cohort.
Materials and methods: Genome-wide genotyping data for a total of 457 RA patients [297 good (DAS28-3≤3.2) and 160 poor (DAS28-3≥5.1) responders] on MTX monotherapy were tested for association using an additive model. Support vector machine and genome-wide pathway analysis were used to identify additional risk variants and pathways. All risk loci were imputed to fine-map the association signals and identify causal variant(s) of therapeutic/diagnostic relevance.
Results: Seven novel suggestive loci from genome-wide (P≤5×10(-5)) and three from support vector machine analysis were associated with MTX (poor)response. The associations of published candidate genes namely DHFR (P=0.014), FPGS (P=0.035), and TYMS (P=0.005) and purine and nucleotide metabolism pathways were reconfirmed. Imputation, followed by bioinformatic analysis indicated possible interaction between two reversely oriented overlapping genes namely ENOSF1 and TYMS at the post-transcriptional level.
Conclusion: In this first ever genome-wide analysis on MTX treatment response in RA patients, 10 new risk loci were identified. These preliminary findings warrant replication in independent studies. Further, TYMS expression at the post-transcriptional level seems to be probably regulated through an antisense-RNA involving the 6-bp ins/del marker in the overlapping segment at 3'UTR of TYMS-ENOSF1, a finding with impending pharmacogenetic applications.