Mechanical stress-induced release of mitochondrial double-stranded RNA activates RIG-I to drive osteogenic differentiation in aortic disease
Background: Gain-of-function mutations in RIG-I and MDA5 (RIG-I-like receptors, RLR) cause Singleton-Merten syndrome, an interferonopathy with early aortic calcification. Here we set out to study a possible involvement of RIG-I-like receptor signaling in the formation of aortic valve stenosis in general. Methods and results: We investigated whether mechanical stress-induced RLR activation promotes osteogenic differentiation and calcification of primary human aortic cells, whether such RLR activation drives the development of aortic valve stenosis in vivo, and whether inhibition of downstream JAK/STAT activation ameliorates aortic valve stenosis. We found that activation of primary human aortic interstitial cells by exogenous RIG-I- and MDA5 agonists induced a type I interferon response and an osteogenic program including increased expression of BMP-2 and BSP2, and promoted calcification. These effects were RIG-I/MDA5 or MAVS-dependent, as assessed by CRISPR/Cas9 knockouts. The stress-granule protein G3BP1 and the Wnt receptor FZD5 were induced downstream of RLR activation and were required for osteogenic gene expression and calcification. To assess a potential endogenous source of RIG-I activation, mechanical stretch mimicking valvular tissue stress was applied in cell cuture, which induced a significant RIG-I-dependent interferon response. Sequencing of highly purified preparations of double-stranded RNA from stretched cells demonstrated a predominant mitochondrial origin of double-stranded RNA, and depleting mitochondrial RNA abolished the stretch-dependent response. The JAK/STAT inhibitor ruxolitinib suppressed the RIG-I-induced response including the expression of osteogenic factors in vitro. In a murine wire-injury model of aortic stenosis, MAVS deficiency attenuated valvular morphological changes and improved hemodynamic parameters. Ruxolitinib attenuated hemodynamic progression, prevented aortic stenosis development, and improved survival. Conclusions: Altogether, our results demonstrate that the mechanical-stress-induced release of mitochondrial double-stranded RNA activates RIG-I signaling in aortic interstitial cells, thereby driving osteogenic differentiation and calcification via a MAVS- and G3BP1-dependent pathway. JAK/STAT inhibition with ruxolitinib mitigates this process in a mouse model of aortic valve stenosis, identifying a candidate therapeutic strategy.