Genetic Dissection of the Regulatory Mechanisms of Ace2 in the Infected Mouse Lung

Xu, Fuyi; Gao, Jun; Bergmann, Silke; Sims, Amy C.; Ashbrook, David G.; Baric, Ralph S.; Cui, Yan; Jonsson, Colleen B.; Li, Kui; Williams, Robert W.; Schughart, Klaus; Lu, Lu

Acute lung injury (ALI) is an important cause of morbidity and mortality after viral infections, including influenza A virus H1N1, SARS-CoV, MERS-CoV, and SARS-CoV-2. The angiotensin I converting enzyme 2 (ACE2) is a key host membrane-bound protein that modulates ALI induced by viral infection, pulmonary acid aspiration, and sepsis. However, the contributions of ACE2 sequence variants to individual differences in disease risk and severity after viral infection are not understood. In this study, we quantified H1N1 influenza-infected lung transcriptomes across a family of 41 BXD recombinant inbred strains of mice and both parents—C57BL/6J and DBA/2J. In response to infection Ace2 mRNA levels decreased significantly for both parental strains and the expression levels was associated with disease severity (body weight loss) and viral load (expression levels of viral NA segment) across the BXD family members. Pulmonary RNA-seq for 43 lines was analyzed using weighted gene co-expression network analysis (WGCNA) and Bayesian network approaches. Ace2 not only participated in virus-induced ALI by interacting with TNF, MAPK, and NOTCH signaling pathways, but was also linked with high confidence to gene products that have important functions in the pulmonary epithelium, including Rnf128, Muc5b, and Tmprss2. Comparable sets of transcripts were also highlighted in parallel studies of human SARS-CoV-infected primary human airway epithelial cells. Using conventional mapping methods, we determined that weight loss at two and three days after viral infection maps to chromosome X—the location of Ace2. This finding motivated the hierarchical Bayesian network analysis, which defined molecular endophenotypes of lung infection linked to Ace2 expression and to a key disease outcome. Core members of this Bayesian network include Ace2, Atf4, Csf2, Cxcl2, Lif, Maml3, Muc5b, Reg3g, Ripk3, and Traf3. Collectively, these findings define a causally-rooted Ace2 modulatory network relevant to host response to viral infection and identify potential therapeutic targets for virus-induced respiratory diseases, including those caused by influenza and coronaviruses.