부산대학교 도서관

The process of developing novel medicines is complex, expensive, time-consuming and fraught with failure. A major contributor to the high failure rate of drug development is the poor ability of existing preclinical models to predict drug efficacy and safety in humans. Human genetics enables an emerging set of investigations with utility in drug target identification and validation, akin to a "natural randomised trial". In this thesis, I applied these methods to examine the efficacy and safety of several drug targets. Using variants in the gene encoding angiotensinogen, I found that inhibition of this novel anti-hypertensive drug target is likely to safely reduce the risk of cardiovascular outcomes, with vascular effects that are comparable to those arising from other targets in the renin-angiotensin pathway, and blood pressure lowering more broadly. Next, I applied human genetics to examine several therapeutic questions relating to coronavirus disease 2019 (COVID-19). I found that genetic proxies for renin-angiotensin system modulating therapies show no strong evidence of association with COVID-19, suggesting that these therapies are unlikely to produce substantial benefit or harm in the context of COVID-19. I then investigated the use of human genetic data for potential repurposing of immunomodulatory therapies for the treatment of COVID-19. This work was the first to show that genetic proxies for interleukin-6 receptor (IL-6R) inhibition associate with a lower risk of severe COVID-19. Meta-analysis of IL-6R inhibitor clinical trial data showed these medicines reduced the risk of COVID-19 progression and mortality, therefore recapitulating the findings from the genetic analysis, and illustrating the value of applying human genetics to guide drug repurposing in the context of an emerging pandemic disease. I perform a further set of analysis to show that COVID-19 GWAS data can be leveraged to prioritise a further set of immunomodulatory drug targets. I next analysed clinical trial data of a sclerostin inhibitor (romosozumab) approved for the treatment of osteoporosis, and found evidence suggesting that this drug may increase cardiovascular risk. Genetic variants in the gene encoding sclerosting (SOST) were also associated with a higher risk of cardiovascular events, suggesting that this adverse effect is likely real and target-mediated. Finally, I performed a genome-wide association meta-analysis of erectile dysfunction, which led to the discovery of the first robust risk locus, and a potential drug target, for this common condition. The work presented in this thesis showcases the utility of human genetics in guiding drug target identification and validation, and its potential for ameliorating the high failure rate of drug development.