Nuclear eNOS interacts with andS-nitrosates ADAR1 to modulate type I interferon signaling and endothelial function

Abstract

Background: Nitric oxide (NO) generated by the endothelial NO synthase (eNOS) regulates vascular tone and endothelial homeostasis to counteract vascular inflammation. Most eNOS is localized at the cell membrane or in the Golgi apparatus, but the enzyme is also present in the endothelial cell nucleus. Here we assessed the relevance of nuclear eNOS and NO signaling for endothelial cell function. Methods: eNOS gain- and loss-of-function approaches were combined with confocal microscopy, biochemical, histological and multi-omics analyses. The pathophysiological relevance of the findings was assessed in murine models of atherogenesis (ApoE-/- and partial carotid ligation) as well as in samples from patients with atherosclerosis. Results: eNOS was present in the nucleus of unstimulated human and murine endothelial cells (in vitro and ex vivo) and stimulation with vascular endothelial growth factor (VEGF) enhanced its nuclear localization. Co-immunoprecipitation studies coupled with proteomics revealed the association of nuclear eNOS with 81 proteins involved in RNA binding and processing. Among the latter was double-stranded RNA-specific adenosine deaminase (ADAR1), an enzyme involved in editing double-stranded RNA (dsRNA) via the deamination of adenosine (A) to inosine (I). ADAR1 was S-nitrosated in human endothelial cells, and the knockdown of eNOS resulted in altered ADAR1-mediated A-to-I editing and an increase in dsRNA. Mitochondrial antiviral signaling protein (MAVS) was the transcript most affected by eNOS depletion, and endothelial cells lacking eNOS accumulated dsRNA as well as MAVS protein aggregates. These changes resulted in activation of the type I interferon (IFN) signaling pathway and a marked downregulation of cell cycle-related genes. ADAR1 depletion elicited similar effects on the activation of type I IFN signaling. As a result, growth factor-stimulated cell proliferation was abrogated and basal as well as stimulated cell death were increased. Endothelial dysfunction in mice as well as in subjects with atherosclerosis was accompanied by the accumulation of dsRNA and the activation of the type I IFN signaling. Preserving NO bioavailability in vivo in an eNOS gain-of-function mouse model (Tyr657Phe eNOS) prevented these effects. Conclusions: Our findings have uncovered a novel mechanism linking nuclear eNOS-derived NO with the activity of ADAR1 to maintain vascular homeostasis. Reduced NO bioavailability results in the previously unrecognized activation of a type I IFN response in the endothelium, which contributes to atherogenesis.