Association of the response to tumor necrosis factor antagonists with plasma type I interferon activity and interferon-beta/alpha ratios in rheumatoid arthritis patients: a post hoc analysis of a predominantly Hispanic cohort

Abstract

Objective: Despite the substantial clinical efficacy of tumor necrosis factor alpha (TNFalpha) antagonist therapy in patients with rheumatoid arthritis (RA), some patients respond poorly to such agents. Since an interferon (IFN) signature is variably expressed among RA patients, we investigated whether plasma type I IFN activity might predict the response to TNF antagonist therapy.

Methods: RA patients (n = 35), the majority of whom were Hispanic, from a single center were evaluated before and after initiation of TNF antagonist therapy. As controls, 12 RA patients from the same center who were not treated with a TNF antagonist were studied. Plasma type I IFN activity was measured using a reporter cell assay, and disease status was assessed using the Disease Activity Score in 28 joints (DAS28). Levels of interleukin-1 receptor antagonist (IL-1Ra) were determined in baseline plasma samples using a commercial enzyme-linked immunosorbent assay. The clinical response was classified according to the European League Against Rheumatism criteria for improvement in RA.

Results: Plasma type I IFN activity at baseline was significantly associated with clinical response (odds ratio 1.36 [95% confidence interval 1.05-1.76], P = 0.020), with high baseline IFN activity associated with a good response. Changes in DAS28 scores were greater among patients with a baseline plasma IFNbeta/alpha ratio >0.8 (indicating elevated plasma IFNbeta levels). Consistent with the capacity of IFNbeta to induce IL-1Ra, elevated baseline IL-1Ra levels were associated with better therapeutic outcomes (odds ratio 1.82 [95% confidence interval 1.1-3.29], P = 0.027).

Conclusion: The plasma type I IFN activity, the IFNbeta/alpha ratio, and the IL-1Ra level were predictive of the therapeutic response in TNF antagonist-treated RA patients, indicating that these parameters might define clinically meaningful subgroups of RA patients with distinct responses to therapeutic agents.

Figures

Figure 1. Type I interferon (IFN) plasma activity in patients with rheumatoid arthritis (RA) and healthy control donors (HD)

Plasma type I IFN activity was quantified using the WISH epithelial cell line assay as described in Patients and Methods. The dotted line shows the cut-off point between high and low levels, defined as the type I IFN level above the mean + 2SD of a pool of healthy controls tested in the same assay.

Figure 2. Baseline plasma type I IFN activity in TNF-antagonist-treated RA patients as a function of clinical response

A. Plasma type I IFN activity was determined at baseline (prior to initiation of therapy) in RA patients with non-response (n = 11), moderate response (n = 17), or good response (n = 7) after 5.6±1.4 months of TNF-antagonist therapy. Each symbol indicates an individual subject. B. Plasma type I IFN activity was compared between non responders and the group comprising moderate and good responders. C. Plasma type I IFN activity was compared between good responders and the group comprising non and moderate responders.

Figure 3. Association of increased plasma IFNβ/α ratio with response to TNF antagonists

A. Antibody specific for IFNα or antibody specific for IFNβ was added to RA plasma prior to establishing cultures with WISH cells and IFN activity measured as described in the Patients and Methods section. B. Changes of DAS28 levels between baseline and follow-up in TNF-antagonist-treated patients are demonstrated according to the baseline IFNβ/α ratio (the median value, 0.8, of the distribution was used to differentiate two groups). Results are presented as in Figure 2.

Figure 4. Plasma IL-1ra levels are increased in TNF-antagonist-treated RA patients with a good clinical response

A. Baseline IL-1ra levels (pg/100ml) were determined at baseline (prior to initiation of therapy) in RA patients with non-response (n=11), moderate response (n = 17), or good response (n = 6). B. Baseline IL-1ra levels (pg/100ml) were compared between non responders and the group comprising moderate and good responders. C. Baseline IL-1ra levels (pg/100ml) were compared between good responders and the group comprising non and moderate responders. Results are presented as in Figure 2.