Risankizumab differentially modulates circulating T-cell populations in psoriasis according to autoreactivity status

Background: In a recent paper, our group described that the presence of double autoreactivity to both LL37 and ADAMTSL5 autoantigens in psoriatic patients decreased the clinical responses to risankizumab, but how this influences the changes in the peripheral inflammatory T-cell populations is still unknown. Objective: This study aims to evaluate how risankizumab modulates the circulating inflammatory T-cell populations in psoriatic patients and, specifically, in autoreactive subjects. Methods: The presence of LL37- and ADAMTSL5-reactive circulating T-cells was assessed in a cohort of 142 psoriatic patients, and 87 demonstrated autoreactivity at baseline. Patients were treated with risankizumab for 52 weeks, and specific T-cell populations were analyzed at different timepoints. Results: The frequency of Ki67+ CD4+ , Ki67+ CD8+ T-cells, CD8+ IL-17+ and CD8+ IL-22+ T-cells showed a posi- tive correlation with baseline PASI and decreased with treatment. Notably, CD8+ IL-17+ T-cells decreased both in single-LL37 and single-ADAMTSL5-reactive subjects, but not in subjects that showed autoreactivity to both autoantigens. LL37 autoreactivity of CD4+ and CD8+ T-cells decreased with treatment, but not for CD4+ in double-reactive subjects. While Treg frequency negatively correlated with baseline PASI and in- creased within 16 weeks of treatment, significantly decreasing the IL-17+ CD4+ /Treg ratio over time, Treg modulation was not evident in double-reactive subjects. Interestingly, the subpopulations of CD8+ MAIT IL-17+ and CD3+ MAIT IL-22+ cells, involved also in psoriatic arthritis, decreased in treated subjects following IL-23 inhibition. Conclusion: Rizankizumab efficiently decreases the circulating inflammatory T-cell populations and modulates Tregs’ plasticity in single-LL37- or single-ADAMTSL5-reactive subjects, but not in double-reactive subjects. © 2025 Japanese Society for Investigative Dermatology. Published by Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.