FRI0208 Prediction of treatment response of tocilizumab for rheumatoid arthritis with comprehensive gene expression analysis in peripheral blood mononuclear cells. (23rd January 2014)

• Record Type: Journal Article
• Authors: Sanayama, Y.; Ikeda, K.; Kagami, S.; Furuta, S.; Kashiwakuma, D.; Matsuura, I.; Yamagata, M.; Iwamoto, I.; Umibe, T.; Matsumura, R.; Sugiyama, T.; Sueishi, M.; Nawata, Y.; Hiraguri, M.; Nonaka, K.; Ohara, O.; Nakajima, H.
• Is Part Of: Annals of the rheumatic diseases. Volume 71(2012)Supplement 3
• Page Start: 384
• Page End: 384
• Publication Date: 2014-01-23
• Subjects: Rheumatism -- Periodicals; 616.723005
• DOI: 10.1136/annrheumdis-2012-eular.2665
• Languages: English
• ISSNs: 0003-4967

Abstract

Abstract : Background: Tocilizumab (TCZ) is a biological agent which is highly efficacious for rheumatoid arthritis (RA). Although predicting treatment response to TCZ can be especially beneficial since TCZ takes longer time to demonstrate its efficacy as compared with anti-TNFa agents, such methods have not been established. On the other hand, previous reports showed usefulness of DNA array analysis of peripheral blood to predict clinical response to infliximab in RA. Objectives: To establish methods to predict clinical response to TCZ therapy with comprehensive gene expression analysis of peripheral blood mononuclear cells (PBMCs) in patients with RA. Methods: A total of 18 patients who received TCZ for inadequately controlled RA (CDAI >10) were analyzed as a training cohort. RNA was extracted from PBMCs before the first administration of TCZ and was analyzed for comprehensive gene expression using Human Whole Genome 4×44K format. Clinical response was assessed over six months using CDAI category improvement and physician's global assessment. Results: At six months of TCZ treatment, responders and non-responders were identified with two different measures (13 responders and three non-responders for CDAI category improvement; 14 responders and three non-responders for physician's global assessment). For each response measure, 20 probes were extracted which demonstrated lowest t-test P values for difference in normalized signals between responders and non-responders and also demonstrated >1.5 fold difference between responders and non-responders. 19 common probes between these two response measures were identified, all of which represented a distinctive gene. Clustering analysis of these genes in the training cohort showed a characteristic pattern for non-responders (figure ). A prediction model using these genes with a cut-off signal for each probe completely discriminated non-responders from responders in the training cohort. Conclusions: Comprehensive gene expression analysis of PBMCs provided a promising result in developing methods to predict clinical response to TCZ therapy for RA. Disclosure of Interest: None Declared