Clinical pharmacogenetic models of treatment response to methotrexate monotherapy in Slovenian and Serbian rheumatoid arthritis patients: Differences in patient's management may preclude generalization of the models

Abstract

Objectives: Methotrexate (MTX) is the first line treatment for rheumatoid arthritis (RA), but nevertheless 30% of patients experience MTX inefficacy. Our aim was to develop a clinical pharmacogenetic model to predict which RA patients will not respond to MTX monotherapy. We also assessed whether this model can be generalized to other populations by validating it on a group of Serbian RA patients. Methods: In 110 RA Slovenian patients, data on clinical factors and 34 polymorphisms in MTX pathway were analyzed by Least Absolute Shrinkage and Selection Operator (LASSO) penalized regression to select variables associated with the disease activity as measured by Disease Activity Score (DAS28) score after 6 months of MTX monotherapy. A clinical pharmacogenetic index was constructed from penalized regression coefficients with absolute value above 0.05. This index was cross-validated and also independently validated on 133 Serbian RA patients. Results: A clinical pharmacogenetic index for prediction of DAS28 after 6 months of MTX monotherapy in Slovenian RA patients consisted of DAS28 score at diagnosis, presence of erosions, MTX dose, Solute Carrier Family 19 Member 1 (SLC19A1) rs1051266, Solute Carrier Organic Anion Transporter Family Member 1B1 (SLCO1B1) rs2306283, Thymidylate Synthase (TYMS), and Adenosine Monophosphate Deaminase 1 (AMPD1) rs17602729. It correctly classified 69% of Slovenian patients as responders or nonresponders and explained 30% of variability in DAS28 after 6 months of MTX monotherapy. Testing for validity in another population showed that it classified correctly only 22.5% of Serbian RA patients. Conclusions: We developed a clinical pharmacogenetic model for DAS28 after 6 months of MTX monotherapy in Slovenian RA patients by combining clinical and genetic variables. The clinical pharmacogenetic index developed for Slovenian patients did not perform well on Serbian patients, presumably due to the differences in patients' characteristics and clinical management between the two groups. © 2018 Jenko, Tomšič, Jekić, Milić, Dolžan and Praprotnik.