Integrating genomics with biomarkers and therapeutic targets to invigorate cardiovascular drug development

• Published in: Nature reviews cardiology 2021-06, Vol.18 (6), p.435-453
• Main Authors: Holmes, Michael V., Richardson, Tom G., Ference, Brian A., Davies, Neil M., Davey Smith, George
• Format: Article
• Language: English
• Subjects: 631/154/53; 631/208; 692/53; Biological markers; Biomarkers; Cancer; Cardiac Imaging; Cardiac Surgery; Cardiology; Cardiovascular agents; Cardiovascular Agents - chemistry; Cardiovascular diseases; Cardiovascular Diseases - drug therapy; Cardiovascular Diseases - genetics; Causality; Cholesterol; Clinical trials; Disease; Drug Development; Drug dosages; Drug therapy; Epidemiologic methods; Epidemiology; Genetics; Genomics; Humans; Medicine; Medicine & Public Health; Metabolites; Methods; Pharmaceutical research; Proteins; Public health; Review Article; Risk factors
• ISSN: 1759-5002; 1759-5010; 1759-5010
• DOI: 10.1038/s41569-020-00493-1

Drug development in cardiovascular disease is stagnating, with lack of efficacy and adverse effects being barriers to innovation. Human genetics can provide compelling evidence of causation through approaches such as Mendelian randomization, with genetic support for causation increasing the probability of a clinical trial succeeding. Mendelian randomization applied to quantitative traits can identify risk factors for disease that are both causal and amenable to therapeutic modification. However, important differences exist between genetic investigations of a biomarker (such as HDL cholesterol) and a drug target aimed at modifying the same biomarker of interest (such as cholesteryl ester transfer protein), with implications for the methodology, interpretation and application of Mendelian randomization to drug development. Differences include the comparative nature of the genetic architecture — that is, biomarkers are typically polygenic, whereas protein drug targets are influenced by either cis -acting or trans -acting genetic variants — and the potential for drug targets to show disease associations that might differ from those of the biomarker that they are intended to modify (target-mediated pleiotropy). In this Review, we compare and contrast the use of Mendelian randomization to evaluate potential drug targets versus quantitative traits. We explain how genetic epidemiological studies can be used to assess the aetiological roles of biomarkers in disease and to prioritize drug targets, including designing their evaluation in clinical trials. In this Review, Holmes and colleagues compare and contrast the use of Mendelian randomization to evaluate potential drug targets versus quantitative traits and explain how genetic epidemiological studies can be used to assess the aetiological roles of biomarkers in disease and to prioritize drug targets, including designing their evaluation in clinical trials. Key points Mendelian randomization offers unique opportunities to explore the causal role of biomarkers and drug targets in disease aetiology. A biomarker is typically a complex trait that has a polygenic architecture, whereas a drug target is usually a protein that has a distinct genetic architecture consisting of cis -acting and trans -acting variants. The motivation, application and interpretation of Mendelian randomization analysis applied to complex biomarkers differ from those of Mendelian randomization analysis applied to drug targets. These differences