Dunca, Diana; (2025) Leveraging the genetic characteristics of diverse populations to guide drug development. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text Dunca_thesis.pdf - Accepted Version Download (12MB) | Preview

Abstract

Individuals of diverse ancestries are under-represented or excluded from genome-wide association studies (GWAS) conducted predominantly in individuals of European ancestries. Consequently, a lack of population diversity in genetic studies has led to a limited understanding on the potential presence of difference in disease aetiology and genetically driven drug target identification. Understanding genetic differences between population groups would enable the full spectrum of potential druggable targets to be explored through genetically guided drug development. Therefore, the main objective of this thesis is to leverage the genetic differences in linkage disequilibrium patterns, allele frequencies and genetic variant effect sizes across populations to compare drug target validation approaches and drug target effect estimates between non-European and European populations in order to better inform drug development. Mendelian randomization (MR) is a powerful tool for causal inference that leverages genetic variants from genome-wide association studies (GWAS) as instruments to estimate the causal effect of an exposure on an outcome. This method has also been extended to evaluate the causal effect of drug targets on diseases. Integrating non-European population-based cohort studies and multi-ancestry biobanks in MR analyses can inform ancestry specific risk to cardiometabolic diseases, and allow effect estimates comparison between population groups. In addition, leveraging the genetic architecture of diverse populations can provide evidence of differences in association of disease risk with prescription between populations and the opportunity to perform ancestry specific drug target validation and drug target effect comparisons across population groups. The first chapter of this thesis aims to identify and compare risk factors (plasma lipids, blood pressure, and body mass index: BMI) for coronary artery diseases (CAD) between the European and South Asian populations, using transferable genetic variants across population groups as instruments in a trans-ancestry MR analysis. Different strategies for trans-ancestry MR were compared to assess the causal effect of cardiometabolic risk factors on the risk of CAD in the European populations and British Pakistani and Bangladeshi individuals from the Genes & Health cohort. Although most of the associations were not significant in the ancestry matched MR, a risk increasing effect for LDL-C and risk decreasing effect for HDL-C was seen when using the variants from the large European GWAS as instruments, and for the subset of loci that were transferable. The association of BMI with CAD risk in G&H was significant only for transferable loci. Therefore, this study highlighted the importance of considering transferability of risk factors loci to ensure causal inference and demonstrated that incorporating findings from large European GWAS can increase power for MR in other ancestry groups. The second chapter focuses on comparing the effect of cholesterol ester transfer protein (CETP), a lipid drug target under development for coronary heart disease (CHD), on biomarkers, cardiometabolic and immune mediated safety outcomes between individuals of European and East Asian populations using cross-ancestry colocalization and ancestry-specific drug target cis-MR. Previous drug target MR studies conducted in East Asians failed to show a CHD effect for CETP, which has been interpreted as lack of effectiveness of CETP inhibition for CHD prevention in this population. Drug target MR, scaled to a standard deviation increase in HDL-C, showed that lower CETP was associated with systolic blood pressure and pulse pressure in both population groups. The protective effect of lower CETP against CHD, angina, and intracerebral haemorrhage in both ancestries was also replicated using a European centric GWAS on plasma CETP concentration. Overall, on-target inhibition of CETP was anticipated to decrease cardiovascular disease in individuals of both European and East Asian ancestries. The third chapter leverages GWAS of medication prescription in a genome-wide MR in the European and East Asian ancestries in UK Biobank and Biobank Japan to compare the association between plasma biomarker levels and disease risk (exposure) and cardiovascular (CVD) medication prescription (outcome) between populations, and consequently assess whether drug prescriptions may be an important environmental factor mediating the genetic effects of disease or biomarker levels. Higher genetic susceptibility to cardiovascular diseases, and elevated plasma biomarker levels were found to be associated with high CVD medication prescription, including 3-hydroxyl-3methylglutaryl coenzyme A reductase (HMG-CoA reductase or HMGCR) inhibitors, calcium and beta channel blockers, in both population groups. Genome-wide MR was additionally employed to assess whether genetic predisposition to non-cardiovascular immune mediated diseases was associated with CVD medication. A higher risk of immune mediated comorbidities such as COPD, asthma and chronic sinusitis were associated with low CVD drug prescriptions, highlighting CVD drug contraindications that could lead to adverse drug reactions, such as atopic dermatitis, predominantly in the East Asian population. Finally, unanticipated prescription-drug associations, in the case of immune-mediated comorbidities, like COPD, could lead to a better understanding of the shared mechanistic pathways between cardiovascular diseases and associated comorbidities for which CVD medication is prescribed. The last chapter employed drug target MR of variants located within and around the LPA gene that determines approximately 90% of the plasma lipoprotein [a] (Lp[a]) concentration, to explore the association of low Lp[a] concentration with dementia and CVD related outcomes in individuals of European ancestry in UK Biobank (UKB). This analysis was restricted to the European population and could not be performed in the non-European individuals in UK Biobank due to the low number of dementia and CVD cases. Although the association of Lp[a] with CHD risk is well characterized and RNAs therapies targeting Lp[a] are now being evaluated in clinical trials for treatment of CHD, observational studies have previously linked Lp[a] concentration to cognitive decline and dementia, promoted by lipid dysregulation, atherosclerosis, and impaired vascular health. The cis-MR analysis showed that lower Lp[a] concentration was associated with reduced cardiometabolic biomarker levels and replicated previously found association between low Lp[a] and CHD, peripheral artery disease and stroke. Finally, lower Lp[a] was also associated with a protective effect on vascular dementia, but not associated with other types of dementia, such as all cause dementia, Alzheimer’s disease or frontotemporal dementia. Therefore, Lp[a] reduction is a promising therapeutic strategy for the management of CHD events, including AAA, PAD, and stroke, and reduce cardiometabolic biomarkers. Moreover, patients with vascular dementia could also benefit from Lp[a] reduction. In conclusion, integrating non-European population-based cohort studies and multi-ancestry biobanks in genome-wide MR analyses can inform ancestry specific risk to cardiometabolic diseases, and allow effect estimates comparison between population groups. Leveraging the genetic architecture of diverse populations can provide increased evidence of ancestry specific drug target effects on cardiometabolic diseases and safety outcomes. Finally, population specific drug target MR can successfully assess the validity and generalizability of drug targets in cardiometabolic diseases in non-European populations. Therefore, the concept of genetically driven drug development could potentially expand to ancestry-informed drug development to consequently reduce the burden of health disparities in diverse populations.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as XMLCSVJSON

Downloads by country - last 12 months

Export as JSONCSVXML

Archive Staff Only

View Item