KLRB1 and IL12RB1 are pretherapeutic predictors of time to relapse in patients with psoriasis treated with IL-17 blockade.

Abstract

Relapse of psoriasis after the cessation of treatment remains an unresolved clinical challenge. To identify predictors of risk of relapse in patients with psoriasis treated by interleukin (IL) 17 blockade (ixekizumab) and to investigate longitudinal dynamics driving relapse. Longitudinal skin transcriptomes were collected from 23 patients across baseline, psoriasis area and severity index 75 at week 8, and relapse phases. An iterative random forest regression model and SHapley Additive exPlanation values were used to determine key predictive genes of relapse. Profiling of immune-infiltrating cells and time-series-based clustering were performed to identify longitudinal transcriptomic signatures. KLRB1 and IL12RB1 were identified as key predictive genes by a machine learning model and validated by immunofluorescence on baseline skin biopsies. Multivariable Cox regression built by KLRB1 and IL12RB1 stratified patients with high risk of relapse (log-rank test P = .0013). Reactivation of CD4⁺ memory T cells, dendritic cells, and mast cells was associated with relapse; KLRB1 and IL12RB1 were positively correlated with the activated state of these immune-infiltrating cells. Patients were recruited from a single center, and all were treated with ixekizumab. Patients with psoriasis who exhibit elevated levels of KLRB1 and IL12RB1 at baseline have a higher risk of relapse after ixekizumab withdrawal.