Loss of immune homeostasis in Type 1 diabetes (T1D) leads to a dysregulated and autoreactive immune response that destroys pancreatic β cells, causing absolute insulin deficiency. Nevertheless, current strategies for restoring immune homeostasis remain limited. Inspired by our earlier research, we leveraged the single-cell RNA sequencing data from T1D patients, and unexpectedly found that STAT1 overrepresentation is much more prominent in dendritic cells (DCs) rather in CD4 T cells. Subsequently, we repurposed the clinically-applied STAT1 inhibitor fludarabine in T1D setting. To resolve the underlying mechanism, we employed a multipronged approach in animal studies, incorporating FACS, RNA-seq, ChIP-qPCR, Co-IP/MS, CESTA, TEM, Seahorse assay and Conditional gene knockout model. Furthermore, we investigated the significance of our data in human autoimmune diabetes. Other than directly targeting CD4+ effector T cells, fludarabine also elevated regulatory T cell (Treg) frequency, and therefore, its administration markedly alleviated T1D pathogenesis. Interestingly, fludarabine did not show a direct effect on Treg cells but indirectly fosters Treg program via inducing tolerogenic DCs (tolDCs). The fludarabine-reprogrammed tolDCs are featured by the metabolic shift towards mitochondrial oxidative respiration and exert protective effects on the adoptive transfer studies. Since the emergence of tolDCs could not be fully explained by STAT1 itself, we further explored whether fludarabine alters the STAT1 interactome. Notably, fludarabine binds to STAT1 and disrupts its interaction with the aryl hydrocarbon receptor (AhR), thereby facilitating AhR nuclear translocation. Activated AhR transcriptionally upregulated the expression of anti-inflammatory, anti-ferroptotic, and mitochondrial respiration genes to uphold the tolerogenic DC phenotype. Our findings identified that fludarabine could be a promising immunometabolic therapeutic candidate to restore immune tolerance, which may be a viable approach against T1D in clinical settings.
Rong et al. (Thu,) studied this question.