Background In rheumatoid arthritis (RA), clinical phenotype is often heterogeneous and response to DMARD therapy can be difficult to predict. We hypothesised that specific peripheral blood (PB) gene expression patterns would associate with subgroups of RA phenotype and/or treatment response and could ultimately provide a means of estimating prognosis

Objectives To determine whether clinically recognizable RA phenotypic and treatment response subgroups are associated with specific PB gene expression signatures

Methods Outcome data and blood samples were collated from 79 newly diagnosed RA patients meeting ACR/EULAR 2010 criteria who were participating in a study of step-up DMARD escalation strategies. Baseline and 3 month blood samples was collected into PAXgene RNA tubes and analysed using the Illumina HumanHT12v4 beadchip platform. Microarray data was quantile normalized and evidence of batch effect was corrected by ComBat transformation

Comparator groups were formed using established descriptors of phenotype, disease activity and treatment response. Evidence for differential gene expression between the groups was investigated using the Linear Models for Microarray Data technique and an adjusted p value (adjp) threshold of 0.05

Results 66 genes were differentially expressed when all males where compared to all females. 2 genes (Sorcin, NLR family apoptosis inhibitory protein) were up-regulated in rheumatoid factor positive patients (adjp=0.04 for both). The Leucine rich repeat neuronal 3 gene was up-regulated in current smokers compared to non and ex-smokers (adjp=0.002 and 0.003 respectively). Groupings based on anti-CCP and erosion status were not associated with differential gene expression

Disease Activity: groups based on baseline and 3 months quartiles of DAS28ESR, and high/ moderate/low/remission thresholds, did not exhibit differential expression in corresponding samples

Predictive ability: groups based on DAS28ESR thresholds and quartiles after 3 months and quartiles of mean DAS28ESR over 12 months did not exhibit differential expression in baseline samples

Change from baseline: In the whole cohort, 19 genes exhibited significant changes in expression between baseline and 3 months. However, there were no significant changes in gene expression between baseline and 3 month samples in the 34 patients who achieved DAS28ESR remission at 3 months (i.e. the biggest clinical change). Correcting for gender did not identify new evidence of differential gene expression in any of the preceding comparisons

Conclusions Herein we did not identify peripheral blood gene expression signatures that associate with either established descriptors of disease activity, treatment response or disease course. Apparent evidence of differential gene expression based on rheumatoid factor (but not anti-CCP) and smoking status requires validation. Consideration should be given as to whether peripheral blood is the most appropriate target tissue for identification of clinically useful gene expression signatures and to whether future studies are sufficiently large to account for all potential external confounders

Disclosure of Interest J. Dale Grant/research support from: Pfizer Inc, Paid instructor for: Abbvie, J. Nijjar: None declared, J. McClure: None declared, M. McBride: None declared, D. Porter Grant/research support from: Pfizer Inc, I. McInnes Grant/research support from: Pfizer Inc