| dc.description.abstract | Rheumatoid arthritis (RA) is a chronic, systemic, and progressive autoimmune disease primarily affecting synovial joints. It is characterised by immune cell-infiltration, severe synovitis, and angiogenesis which leads to the erosion of bone and cartilage. Due to the considerable diversity exhibited by the disease at the level of synovial pathology, and in terms of variable response to treatment, it is a prime candidate for stratified medicine, whereby patients receive targeted treatment, personalised to their individual version of the disease. The aim of this work is to uncover biomarkers in genomic and transcriptomic data which can characterize this heterogeneity and determine subgroups or ‘endotypes’ of patients with shared patterns of disease at the molecular and therapeutic level. To achieve this, RNA-sequenced data from two RA patient cohorts is explored: PEAC (Pathobiology of Early Arthritis Cohort) which investigated early-stage treatment-naïve RA; and R4RA, a biopsy-driven randomised clinical trial, which compared the biologic therapies rituximab and tocilizumab in late-stage RA. Dual analysis of both cohorts, including expression quantitative trait loci (eQTL) analysis in the PEAC study, enabled genome-wide and transcriptome-wide associations to be made between RA phenotypes over a range of disease stages, treatments, and severity. Transcriptomic data from both studies was leveraged to determine differentially expressed genes between histological pathotypes which effectively describe different disease states. Several genetic variants were also identified as associated with RA disease severity and systemic inflammation. Based on pre-treatment synovial gene expression, separate machine learning models were able to predict response to individual biologic drugs in late-stage RA, as well as predict drug-refractory patients. Subsequently, this research has provided a path towards future bedside diagnostics which will be able to stratify patients into optimal treatment groups based on prediction of their likelihood to respond to specific treatments and ensure they are provided with optimal care. | en_US |
