Background/Purpose: In early rheumatoid arthritis (RA), it is not possible to predict response to first line DMARDs (e.g. methotrexate (MTX)) and as such treatment decisions rely primarily on clinical algorithms. The capacity to classify drug naïve patients into those that will not respond to first line DMARDs would be an invaluable tool for patient stratification. Here we report that chromosome conformational signatures (highly informative and stable epigenetic modifications that have not previously been described in RA) in blood leukocytes of early RA patients can predict non-responsiveness to MTX treatment.

Methods: Peripheral blood mononuclear cells were obtained from DMARD naïve early RA patients recruited in the Scottish early rheumatoid arthritis inception cohort. Inclusion in this study was based on diagnosis of RA (fulfilling the 2010 ACR/EULAR Criteria) with moderate to high disease activity (DAS28 ≥ 3.2) and subsequent monotherapy with MTX. MTX responsiveness was defined at 6 months using the following criteria: Responders – DAS28 remission (DAS28 <2.6) or a good response (DAS28 improvement of >1.2 and DAS28 ≤3.2). Non-responders – no improvement in DAS28 (≤0.6). Initial analysis of chromosome conformational signatures (CCS) in 4 MTX responders, 4 MTX non-responders and 4 healthy controls was undertaken using an EpiSwitch™ array containing 13,322 unique probes covering 309 RA-related genetic loci. Differentiating CCS were defined by LIMMA linear modeling, subsequent binary filtering and cluster analysis. A validation cohort of 30 MTX responders and 30 non-responders were screened for the differentiating CCS using the EpiSwitch™ PCR platform. The differentiating signature was further refined using binary scores and logistical regression modeling, and the accuracy and robustness of the model determined by ROC analysis.

Results: CCS EpiSwitch™ array analysis identified 30 CCS markers with the potential to stratify early RA patients into responders and non-responders. Subsequent evaluation of this signature in our validation cohort refined this to a CCS signature that was able to discriminate responders and non-responders. Prediction modeling provided a probability score for responders and non-responders, ranging from 0.0098 to 0.99 (0 = responder, 1 = non-responder). There was a true positive rate of 92% (95% confidence interval [95% CI] 75-99%) for responders and a true negative rate of 93% (95% CI 76-99%) for non-responders. Importantly, ROC analysis to validate this stratification model demonstrated that the signature had a predictive power of sensitivity at 92% for non-responders to MTX.

Conclusion: We have identified a highly informative systemic epigenetic state in the peripheral blood of DMARD naïve early RA patients that has the power to stratify patients at the time of diagnosis. The capacity to differentiate patients a priori into non-responders, using a blood-based clinical test, would be an invaluable clinical tool; paving the way towards stratified medicine and justifying more aggressive treatment regimes in early RA clinics.

« Back to 2015 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/epigenetic-chromosome-conformations-predict-mtx-responsiveness-in-early-rheumatoid-arthritis-patients/