Effect of prednisone on type I interferon signature in rheumatoid arthritis: consequences for response prediction to rituximab

Abstract

IntroductionElevated type I interferon (IFN) response gene (IRG) expression has proven clinical relevance in predicting rituximab non-response in rheumatoid arthritis (RA). Interference between glucocorticoids (GCs) and type I IFN signaling has been demonstrated in vitro. Since GC use and dose are highly variable among patients before rituximab treatment, the aim of this study was to determine the effect of GC use on IRG expression in relation to rituximab response prediction in RA.MethodsIn two independent cohorts of 32 and 182 biologic-free RA patients and a third cohort of 40 rituximab-starting RA patients, peripheral blood expression of selected IRGs was determined by microarray or quantitative real-time polymerase chain reaction (qPCR), and an IFN-score was calculated. The baseline IFN-score was tested for its predictive value towards rituximab response in relation to GC use using receiver operating characteristics (ROC) analysis in the rituximab cohort. Patients with a decrease in disease activity score (∆DAS28) >1.2 after 6 months of rituximab were considered responders.ResultsWe consistently observed suppression of IFN-score in prednisone users (PREDN+) compared to non-users (PREDN−). In the rituximab cohort, analysis on PREDN− patients (n = 13) alone revealed improved prediction of rituximab non-response based on baseline IFN-score, with an area under the curve (AUC) of 0.975 compared to 0.848 in all patients (n = 40). Using a group-specific IFN-score cut-off for all patients and PREDN− patients alone, sensitivity increased from 41% to 88%, respectively, combined with 100% specificity.ConclusionsBecause of prednisone-related suppression of IFN-score, higher accuracy of rituximab response prediction was achieved in PREDN− patients. These results suggest that the IFN-score-based rituximab response prediction model could be improved upon implementation of prednisone use.

Using the Immunophenotype to Predict Response to Biologic Drugs in Rheumatoid Arthritis

“…This was supported by a prospective analysis of 14 RA patients commenced on rituximab with further validation in 26 RA patients, with an AUC of 0.87 [51]. Furthermore, prednisolone treatment drives down type I interferon signatures and prediction of rituximab response was therefore highest in the prednisolone-negative group with an AUC of 0.97 [62]. These data suggest that those patients with a high baseline type I interferon signature respond well to anti-TNF and less well to rituximab, which may indicate differences in pathogenesis in the two groups.…”

Type I interferon response gene expression in established rheumatoid arthritis is not associated with clinical parameters

“…As previously described, suppression of the IFN score by certain treatment could affect the applicability of the IFN signature as a biomarker for therapy response, particularly to rituximab [5, 19]. That is, the treatment-related suppression of IFN score might impair the discriminative capacity of the biomarker, which would consequently lead to more false predictions.…”

Genotypes, phenotypes and treatment with immunomodulators in the rheumatic diseases

“…This work led to the development of a RTX response prediction model; more recent work by De Jong et al . has shown that because of glucocorticoid‐related suppression of the IFN‐score, the highest accuracy of rituximab response prediction is achieved in patients who are not treated with glucocorticoids .…”

“…Moreover, Vosslamber et al showed that a good clinical response to rituximab is associated with a selective drug-induced increase in type I IFN-response activity [41]. This work led to the development of a RTX response prediction model; more recent work by De Jong et al has shown that because of glucocorticoid-related suppression of the IFN-score, the highest accuracy of rituximab response prediction is achieved in patients who are not treated with glucocorticoids [42,43].…”