Rxivist logo

A factorial Mendelian randomization study to systematically prioritize genetic targets for the treatment of cardiovascular disease

By Genevieve M. Leyden, Tom R Gaunt, Tom G Richardson

Posted 20 Feb 2020
medRxiv DOI: 10.1101/2020.02.16.20023010

ImportanceNew drugs which provide benefit alongside statin mono-therapy are warranted to reduce risk of cardiovascular disease. ObjectiveTo systematically evaluate the genetically predicted effects of 8,851 genes and cardiovascular disease risk factors using data from the UK Biobank and subsequently prioritize their potential to reduce cardiovascular disease in addition to statin therapy. Design, Setting, and ParticipantsA factorial Mendelian randomization study using individual level data from the UK Biobank study. This population-based cohort includes a total of 502,602 individuals aged between 40 and 96 years old at baseline who were recruited between 2006 to 2010. ExposuresGenetic variants robustly associated with the expression of target genes in whole blood (based on P<5x10-08) were used to construct gene scores using findings from the eQTLGen consortium (n=31,684). Main Outcomes and MeasuresGenetically predicted effects for each of the 8,851 genes were firstly evaluated with 5 measured outcomes from the UK Biobank in turn (body mass index, diastolic blood pressure, systolic blood pressure, low-density lipoproteins and triglycerides). Effects surviving multiple comparisons from this initial analysis were subsequently analyzed using factorial Mendelian randomization to evaluate evidence of an additive beneficial effect on cardiovascular disease risk compared to a HMGCR genetic score acting as a proxy for statin inhibition. Finally, a phenome-wide analysis was undertaken to evaluate predicted effects between prioritized targets and 569 outcomes in the UK Biobank to highlight potential adverse side-effects. Results377 genetically predicted effects on cardiovascular disease risk factors were identified by Mendelian randomization (based on P<1.13x10-6). Of the 68 druggable genes, 20 candidate genes were predicted to lower cardiovascular disease risk in combination with statin treatment (P<7.35x10-4). Genes such as FDFT1 were predicted to have added benefit with statin therapy (OR=0.93; 95% CI, 0.91-0.95; P=2.21x10-10) but are unlikely to make safe targets due to their predicted effects on adverse side effects. In contrast, PRKCE provided evidence of a putative added benefit in combination with statins (OR=0.94; CI, 0.91-0.96; P=1.72x10-9) with no predicted adverse effects. Conclusions and RelevanceThrough the application of a systematic factorial Mendelian randomization analysis, we have prioritized (and deprioritized) potential drug targets predicted to reduce cardiovascular disease risk in addition to statin therapy. Key pointsO_ST_ABSQuestionC_ST_ABSCan naturally occurring genetic variation in a population help us highlight and prioritize novel therapeutic targets for the treatment of cardiovascular disease? FindingsIn this factorial Mendelian randomization study of 334,915 individuals, we found that a genetically predicted 0.09 mmol/L decrease in LDL cholesterol attributed to statin inhibition results in 4.1% lower risk of cardiovascular disease. We then highlighted various genetic targets which were genetically predicted to further reduce disease risk without evidence of adverse side effects, such as PRKCE which is involved in the development of cardiac hypertrophy and reduced risk of cardiovascular disease by 6.4% in addition to statin therapy. MeaningEvidence from genetic analyses can improve the likelihood of success for therapeutic targets and findings from this study have prioritized several promising candidates for the treatment of cardiovascular disease.

Download data

  • Downloaded 511 times
  • Download rankings, all-time:
    • Site-wide: 65,128
    • In genetic and genomic medicine: 285
  • Year to date:
    • Site-wide: 52,627
  • Since beginning of last month:
    • Site-wide: 49,228

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide