In silico modeling guides identification of novel JAK1 variants associated with immune dysregulation.

Characterization of primary immune dysregulations and deficiency disorders caused by hyperactivating variants of the JAK/STAT pathway highlighted its crucial role in immune cell development and response. To systematically evaluate pathogenic JAK1 variants, we developed a structure-based predictive framework adapting AlphaFold2, modeling both the active and inactive conformations of JAK1. Dual-state modeling of 21,926 JAK1 variants enabled discrimination between pathogenic and benign variants based on their impact on regulatory conformation. Applying this approach to a large cohort of patients with suspected primary immune dysregulation and deficiency led to the identification of five novel variants located in key cis-regulatory and catalytic domains, with predicted gain of function activity. Ectopic expression of these variants in cell line resulted in varying levels of hyperactivation of JAK1 and multiple STATs at baseline. Furthermore, treatment of two patients with Tofacitinib suppressed JAK1 hyperactivation, normalized plasma cytokine levels and interferon signatures, and significantly improved clinical symptoms. These findings reveal diverse mechanisms of JAK1 gain of function, expanding the clinical spectrum JAK1 GOF, and underscore the importance of precise variant characterization for effective personalized therapy.