Background/Purpose: TNF-α inhibitors are among the most widely used biological-DMARDs in rheumatoid arthritis (RA). Means to predict response would allow for a more effective, targeted approach to therapy. Th17-related cytokines synergise with TNF-α to intensify joint inflammation. Evidence suggests poor response to TNF-α inhibitors reflects a TNF-α-independent, Th17-driven process. Our aim was to assess plasma concentrations of a panel of Th17-related cytokines to identify predictors of response to TNF-α inhibitors.

Methods: Ninety-three patients with RA as defined by ACR criteria were seen prior to and 4-6 months after commencing etanercept or adalimumab. Plasma concentrations of Th17-related cytokines (IL-1β, IL-4, IL-6, IL-10, IL-17A, IL-17F, IL-21, IL-22, IL-23, IL-25, IL-31, IL-33, sCD40L, IFN-γ, TNF-α) were measured at baseline. DAS28-CRP with 4 variables (DAS28) was calculated and EULAR cut-off for low disease activity was used to group patients into poor (DAS28 >3.2) and good (DAS28 ≤3.2) responders at follow up. IL17-IL23-pathway cytokines (IL-17A, IL-17F, IL-22, IL-23, IL-1b and IFN-γ) were grouped to form a pro-inflammatory cumulative score (presence of 0 to 6 cytokines), while anti-inflammatory cytokines, IL-4, IL-10 and IL-25, were grouped to form an anti-inflammatory cumulative score (presence of 0 to 3 cytokines). Multivariate logistic regression was used to identify predictors of response, and odds ratios (OR) and 95% confidence intervals (CI) were calculated.

Results: Patients with IL-23 present at baseline were more likely to be poor-responders [14/20 (70%) of IL-23⁺ patients versus 31/73 (42.5%) of IL-23^– patients; p<0.05]. Multivariate analysis suggests that presence of IL-23 predicts poor response following treatment with anti-TNF-α [OR (95% CI) for 6.55 (1.82-23.63), p<0.01], while presence of IL-25 predicts good response [OR (95% CI) = 4.50 (1.53-13.19), p<0.01]. These associations were independent of baseline DAS28. The remaining individual cytokines tested were not associated with response to anti-TNF-α therapy. Pro-inflammatory and anti-inflammatory cumulative scores were not associated with response to anti-TNF-α therapy in univariate analysis. However the anti-inflammatory score was associated with reduced risk of poor response in a multivariate model, including baseline DAS28 and the pro-inflammatory score [OR (95% CI) = 0.54 (0.30-0.97), p<0.05], suggesting the association was dependent on confounding effects of these variables.

Conclusion: IL-23, which promotes pathogenic activity of Th17 cells, may be a useful predictor of poor outcome following anti-TNF-α therapy. IL-25 which downregulates pathogenic activity of Th17 cells may be useful alongside IL-23 to predict a favourable outcome. A cytokine expression profile including IL-25 and additional anti-inflammatory cytokines, IL-4 and IL-10, may indicate increased likelihood of response.

« Back to 2018 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/plasma-il-23-and-il-25-predict-response-to-anti-tnf-%ce%b1-therapy-in-rheumatoid-arthritis/