CGP-35269

Severe viral pneumonia caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is characterized by a hyperinflammatory state typified by elevated circulating pro-inflammatory cytokines, frequently leading to potentially lethal vascular complications including thromboembolism, disseminated intracellular coagulopathy and vasculitis. Though endothelial infection and subsequent endothelial damage have been described in patients with fatal COVID-19, the mechanism by which this occurs remains elusive, particularly given that, under naïve conditions, pulmonary endothelial cells demonstrate minimal cell surface expression of the SARS-CoV-2 binding receptor ACE2. Herein we describe SARS-CoV-2 infection of the pulmonary endothelium in postmortem lung samples from individuals who died of COVID-19, demonstrating both heterogeneous ACE2 expression and endothelial damage. In primary endothelial cell cultures, we show that SARS-CoV-2 infection is dependent on the induction of ACE2 protein expression and that this process is facilitated by type 1 interferon-alpha (IFNα) or -beta(β)—two of the main anti-viral cytokines induced in severe SARS-CoV-2 infection—but not significantly by other cytokines (including interleukin 6 and interferon γ/λ). Our findings suggest that the stereotypical anti-viral interferon response may paradoxically facilitate the propagation of COVID-19 from the respiratory epithelium to the vasculature, raising concerns regarding the use of exogenous IFNα/β in the treatment of patients with COVID-19.

As in bacterial infections and endotoxin shock, type I interferons (IFNs) also have complex and often opposing effects in various models of autoimmune disease. We have shown that type I IFN paradoxically inhibits autoimmune diabetes in the nonobese diabetic mouse (NOD) and biobreeding (BB) rat. We hypothesize that type I IFN activity differs by IFN subtype and interaction with IFN-gamma. We examined the structure-function relationship of the type I IFN molecule and the mechanism of its diabetes-sparing activity in the NOD mouse. While both recombinant human IFN-alpha A/D (bgl 11) (rHuIFN-alphaA/D) and ovine IFN-tauImod (ovIFN-tau) potently inhibited the development of diabetes (P < 0.01), neither recombinant human IFN-alpha B/D (rHuIFN-alphaB/D) nor recombinant human IFN-alpha consensus (CIFN) were efficacious. The activity of IFN subtypes correlate with their NH3-terminal amino acid sequences. All type I IFN save CIFN, which has no diabetes-sparing activity, inhibited the accessory cell function. IFN-tau administration decreased the expression of Fas and ICAM on total cells, class II MHC expression on B cells, and CD40L expression on T cells by 39%, 45%, 45%, and 60%, respectively. In addition, IFN-tau inhibited the development of diabetes in the NOD.IL4(null) but not the NOD.IFN-gamma(null) mice, suggesting a coordinated interaction between type I and type II IFNs to suppress diabetes development. Thus, the amino terminal portion of the type I IFN molecule influences its ability to inhibit the development of autoimmune diabetes in NOD mice. These data also support the contention that IFN-gamma may have a role in mediating the diabetes-sparing effect of high-dose type I IFNs by the inhibition of the IFN-gamma-inducible immune modulators, class II MHC, Fas, ICAM, and CD40L.