Background and Objectives An early prediction of the outcome of particular DMARD therapies is necessary to treat the patients individually. This study aimed at defining predictive mRNA biomarkers in rheumatoid arthritis (RA) patients for the response to future treatment with methotrexate (MTX).
Materials and Methods For this purpose whole blood samples from 52 active RA patients before treatment were collected in PAXgene® Blood RNA Tubes (PreAnalytiX). Extraction of intracellular RNA including miRNA was performed according to PAXgene® Blood miRNA. Kit protocol. Extracted RNA samples were amplified employing the GeneChip® 3′IVT Express Kit and hybridised onto Affymetrix GeneChip® 133 Plus 2.0 Arrays. Labelling was carried out with the GeneChip® Hybridisation, Wash and Stain Kit in a GeneChip® Fluidics Station 450. Generated signal data was normalised and evaluated using the Affymetrix Expression Console (MAS5.0) and the BioRetis online database for candidate gene selection (BioRetis GmbH, Berlin). Classification of RA patients into good, moderate and non-responders was performed based on the DAS28 and EULAR response criteria after 3 months of MTX treatment. Prior to further expression analysis all patients were divided into two genetically more homogeneous subsets – one group of 29 donors expressing a specific HLA allele mRNA and one group of 23 not expressing this specific HLA mRNA. Hierarchical clustering of discriminating candidate genes was performed using the Genesis software v1.7.6 (IGB-TU, Graz) including and excluding 9 medium responders in the HLA positive cluster and 4 medium responders in the HLA negative cluster.
Results Gene comparison analysis was improved by separation of the patient collective into positive (n = 29) or negative (n = 23) groups expressing a specific HLA mRNA. Differential mRNA expression before treatment was determined between 14 good responders and 6 non-responders in the HLA positive and between 12 good responders and 7 non-responders in the HLA negative group, calculating expression change calls and fold changes. A clear discrimination between responders and non-responders to future MTX treatment was achieved with 16 distinct candidate genes for each group. These mRNA candidates will be validated in independent qPCR analyses and with further statistical cross validation algorithms.
Conclusions Early prediction of response to MTX monotherapy using microarray analyses is an opportunity for effective individual medication and therefore allows preventing side effects.
Employing specific response marker genes such as CD11c for anti-TNF monotherapy it is also of interest to define predictive biomarkers for the commonly used anti-TNF/MTX combination therapy not only preventing side-effects but also reducing costs.