Actionable druggable genome-wide Mendelian randomization identifies repurposing opportunities for COVID-19

Articles https://doi.org/10.1038/s41591-021-01310-z Actionable druggable genome-wide Mendelian randomization identifies repurposing opportunities for COVID-19 Liam Gaziano1,2, Claudia Giambartolomei 3,4, Alexandre C. Pereira5,6, Anna Gaulton 7, Daniel C. Posner 1, Sonja A. Swanson8, Yuk-Lam Ho1, Sudha K. Iyengar9,10, Nicole M. Kosik 1, Marijana Vujkovic 11,12, David R. Gagnon 1,13, A. Patrícia Bento 7, Inigo Barrio-Hernandez14, Lars Rönnblom 15, Niklas Hagberg 15, Christian Lundtoft 15, Claudia Langenberg 16,17, Maik Pietzner 17, Dennis Valentine18,19, Stefano Gustincich 3, Gian Gaetano Tartaglia 3, Elias Allara 2, Praveen Surendran2,20,21,22, Stephen Burgess 2,23, Jing Hua Zhao2, James E. Peters Bram P. Prins 2,21, Emanuele Di Angelantonio2,20,21,25,26, Poornima Devineni1, Yunling Shi1, Kristine E. Lynch27,28, Scott L. DuVall27,28, Helene Garcon1, Lauren O. Thomann1, Jin J. Zhou29,30, Bryan R. Gorman1, Jennifer E. Huffman 31, Christopher J. O’Donnell 32,33, Philip S. Tsao34,35, Jean C. Beckham36,37, Saiju Pyarajan1, Sumitra Muralidhar38, Grant D. Huang38, Rachel Ramoni38, Pedro Beltrao 14, John Danesh2,20,21,25,26, Adriana M. Hung39,40, Kyong-Mi Chang 12,41, Yan V. Sun 42,43, Jacob Joseph1,44, Andrew R. Leach7, Todd L. Edwards45,46, Kelly Cho1,47, J. Michael Gaziano1,47, Adam S. Butterworth 2,20,21,25,26 ✉, Juan P. Casas1,47 ✉ and VA Million Veteran Program COVID-19 Science Initiative* , 21,24 Drug repurposing provides a rapid approach to meet the urgent need for therapeutics to address COVID-19. To identify therapeutic targets relevant to COVID-19, we conducted Mendelian randomization analyses, deriving genetic instruments based on transcriptomic and proteomic data for 1,263 actionable proteins that are targeted by approved drugs or in clinical phase of drug development. Using summary statistics from the Host Genetics Initiative and the Million Veteran Program, we studied 7,554 patients hospitalized with COVID-19 and >1 million controls. We found significant Mendelian randomization results for three proteins (ACE2, P = 1.6 × 10−6; IFNAR2, P = 9.8 × 10−11 and IL-10RB, P = 2.3 × 10−14) using cis-expression quantitative trait loci genetic instruments that also had strong evidence for colocalization with COVID-19 hospitalization. To disentangle the shared expression quantitative trait loci signal for IL10RB and IFNAR2, we conducted phenome-wide association scans and pathway enrichment analysis, which suggested that IFNAR2 is more likely to play a role in COVID-19 hospitalization. Our findings prioritize trials of drugs targeting IFNAR2 and ACE2 for early management of COVID-19. T he global COVID-19 pandemic is responsible for substantial mortality, morbidity and economic hardship. Even with efficacious vaccines against the SARS-CoV-2 virus, it unknown how long it will take to achieve herd immunity, to what extent protection will diminish over time or if future mutations will enable SARS-CoV-2 to evade immune responses stimulated by current vaccines. Hence, there is a need to rapidly identify drugs that can minimize the burden of COVID-19. Although large randomized trials have begun to successfully identify drugs that can be repurposed to address COVID-19 (refs. 1–3), most drugs evaluated so far have failed to show efficacy and have been largely confined to hospitalized or critically ill patients. It is a priority, therefore, to identify additional drugs that can be repurposed for early management in COVID-19. Large-scale human genetic studies are now widely used to inform drug development programs. Drug–target disease pairs supported by human genetics have a greater odds of success in drug discovery pipelines4. For example, identification of variants in PCSK9 associated with lower risk of coronary disease led to the successful development of PCSK9 inhibitors, which are now licensed for prevention of cardiovascular events5. The value of human genetics for drug discovery and development has also been realized for infectious diseases. Human genetic studies showed that genetic variation in the CCR5 gene provides protection against infection by human immunodeficiency virus (HIV) type 1. These findings were key for the development of Maraviroc, an antagonist of CCR5, approved by the US Food and Drug Administration (FDA) for the treatment of patients with HIV-1 (ref. 6). Genetic variants acting in ‘cis’ on druggable protein levels or gene expression that encode druggable proteins can provide powerful tools for informing therapeutic targeting, as they mimic the on-target A full list of affiliations appears at the end of the paper. 668 Nature Medicine | VOL 27 | April 2021 | 668–676 | www.nature.com/naturemedicine Articles Nature Medicine (beneficial or harmful) effects observed by pharmacological modification7. Such Mendelian randomization (MR) analyses have been used to suggest repurposing opportunities for licensed drugs8. MR analysis that focuses on actionable druggable genes, defined as genes that encode the protein targets of drugs that are licensed or in the clinical phase of drug development, could therefore serve as a swift and robust strategy to identify drug-repurposing opportunities to prevent the complications and mortality due to COVID-19. To identify further potential repurposing opportunities to inform trials of patients with COVID-19, we conducted large-scale MR and colocalization analyses using gene expression and soluble protein data for 1,263 actionable druggable genes that encode protein targets for approved drugs or drugs in clinical development. By combining transancestry genetic data from 7,554 hospitalized patients with COVID-19 and more than 1 million population-based controls from the COVID-19 Host Genetics Initiative9 (HGI) and the Million Veteran Program10 (MVP), we provide support for two therapeutic strategies. ~20,000 proteins encoded by human DNA Identifying actionable proteins 1,263 actionable proteins Instrument selection Instruments using conditionally independent cis-eQTLs from GTEx in 49 tissues Mendelian randomization Six genes with significant MR results (P < 3.96 × 10–5) in at least one tissue Actionable druggable proteins. Using data available in ChEMBL v.26, we identified 1,263 human proteins as ‘actionable’ (therapeutic targets of approved or clinical-stage drugs; Supplementary Table 2). Of these, we noted 700 proteins that are targets for drugs with potential relevance to COVID-19 from cell-based screening, registers of clinical trials against COVID-19 or approved immunomodulatory/anticoagulant drugs (given the clear role of these pathways in COVID-19 outcomes) or have biological evidence for the role of the protein in SARS-CoV-2 infection (Supplementary Table 3). Genetic proposed instruments for actionable druggable proteins. Using GTEx v.8 (ref. 12), we identified all conditionally independent expression quantitative trait loci (eQTLs) in 49 tissues that act in cis (within 1 Mb on either side of the encoded gene), wh (...truncated)