Background The overall response rate (defined as low disease activity [LDA] related to DAS28[CRP]) to abatacept associated with MTX was 57.1% at 6 months in the 6-month open-label abatacept Power Doppler Ultrasonography (PDUS; APPRAISE) study in patients with RA and inadequate response to MTX [1–4].
Objectives The objective of this exploratory analysis was to identify potential predictors of response to the abatacept–MTX combination by whole blood gene expression profiling in order to optimize treatment choice.
Methods 104 patients with active RA and inadequate response to MTX were treated with abatacept + MTX at the approved doses. Whole blood in Paxgene tubes was collected for each RA patient at baseline. At 6 months, RA patients were categorized as responders if DAS28 was ≤3.2 (LDA) and non-responders if DAS28 was >3.2. Baseline RNAs from a first set of RA patients (n=44) were hybridized to a whole human genome 4 x 44K microarray Agilent slide to identify a gene combination able to separate responders and non-responders with the GeneSpring GX software. A t-test with false discovery rate correction for multiple testing (p<0.05) was used to determine mRNAs differentially regulated between responders and non-responders. This gene combination was validated with a second set of RA patients (n=32) by qRT-PCR.
Results Among these 104 RA patients, 28 were excluded from the following analysis because of missing clinical data (n=13), poor integrity of whole blood mRNA (RIN<7) (n=6), low concentration rate for reverse transcription (n=7), qRT-PCR failures (n=2). Responders (n=45) and non-responders (n=31) had similar baseline characteristics . In a first set of 44 enrolled RA patients (25 responders and 19 non-responders), 87 mRNAs identified by microarray analysis were expressed as a function of the response to treatment and an unsupervised hierarchical clustering almost perfectly separated these responders from non-responders. The informativeness of 12 of these 87 transcripts (selected with the lowest p-value), as measured by qRT-PCR, was re-assessed in a second set of 32 RA patients (20 responders and 12 non-responders). The combined levels of these 12 transcripts properly classified 23 out of 32 patients with a sensitivity of 80%, a specificity of 66.7%, a positive predictive value of 80%, and a negative predictive value of 66.8%. This combination enriched with more genes is in progress in order to better classify the second set of RA patients.
Conclusions Our gene profiling results obtained by a non-invasive procedure have generated a list of 12 candidate genes that were associated with a response to abatacept combined with MTX in this study. This combination of genes needs to be further validated in other RA studies to support their utility as a potential diagnostic assay for predicting abatacept response in an effort to personalize treatment for RA.
D'Agostino MA et al. Ann Rheum Dis 2012;71(Suppl 3):86.
D'Agostino MA, et al. Arthritis Rheum 2012;64(Suppl):S352.
D'Agostino MA, et al. Arthritis Rheum 2012;64(Suppl):S353.
Lequerre T, et al Arthritis Rheum 2013;65(Suppl):S811.
Disclosure of Interest : T. Lequerré Grant/research support: BMS, Consultant for: BMS, C. Derambure: None declared, M. D'Agostino: None declared, M. Hiron: None declared, P. Gaudin Consultant for: BMS, C. Gaillez Shareholder of: BMS, Employee of: BMS, M. Le Bars Shareholder of: BMS, Employee of: BMS, O. Vittecoq Consultant for: BMS Advisory Board