NEW YORK – By investigating the genetic factors influencing stroke risk across multiple human ancestry groups, an international group of researchers has come up with proposed genetic predictors of the disease and tracked down potential drug targets.
"[O]ur genomic findings derived from [more than] 200,000 patients who had a stroke worldwide provide critical insights to inform future biological research on stroke pathogenesis, highlight potential drug targets for intervention, and provide tools for genetic risk prediction across ancestries," co-senior and co-corresponding authors Stéphanie Debette and Martin Dichgans, along with their colleagues, reported in Nature on Friday.
Debette is codirector of the University of Bordeaux-INSERM Bordeaux Population Health Research Center in France, while Dichgans is a stroke and dementia researcher affiliated with Ludwig Maximilian University and other centers in Munich.
The team noted that stroke genetic risk has mainly been characterized in individuals of European ancestry, despite the overrepresentation of certain cerebral ischemic stroke forms, such as small-vessel stroke or intracerebral hemorrhage, in Asian and African populations.
For their study, the researchers performed a multi-ancestry genome-wide association study meta-analysis that included nearly 110,200 stroke patients and more than 1.5 million unaffected controls, enrolled through dozens of prior population genetic efforts, biobanks, or clinical studies. The participants included individuals with European, African, East Asian, South Asian, or Hispanic ancestry.
The team's search led to dozens of new as well as known loci linked to any type of stroke, ischemic stroke, or specific ischemic stroke subtypes, along with variants found through secondary analyses on ancestry-specific risk loci or loci with ties to additional traits such as heart disease or brain white matter features.
The candidate loci were subsequently tested in more than 1.1 million additional cross-ancestry stroke cases or controls to validate 87 percent of risk loci from the primary analysis and roughly 60 percent from secondary analyses.
After distinguishing between cross-ancestry and ancestry-specific risk loci, the investigators used a quantitative machine learning method called MENTR to assess mutation effects in noncoding regions, bringing in transcriptomic and proteomic clues to narrow in on potential causal genes such as FURIN or SH3PXD2A, as well as candidate causal variants within genes.
From there, the researchers relied on a combination of gene enrichment analyses, negative correlation testing, and protein quantitative trait locus profiling, using data from the Estonian Biobank, to uncover a handful of potential drug target genes, ranging from KLKB1, GP1BA, or VCAM1 to genes such as F11 or PROC that are targeted by compounds already being studied.
"Overall, combining evidence from genomics-driven drug discovery approaches, characterization of stroke risk loci … , and previous knowledge from monogenic disease models and experimental data, we found evidence for the potential functional implication of 56 genes that should be prioritized for further functional follow-up, with evidence from multiple approaches for 20 genes," the authors noted.
With a computational method called elastic-net logistic regression, meanwhile, the team brought together summary statistic data from several GWAS to come up with integrative polygenic risk scores (iPGS) within and across the ancestry groups considered — an approach that appeared to outperform existing PRS models for predicting stroke in individuals with European and East Asian ancestry.
In particular, the investigators touted the stroke prediction performance of a cross-ancestry risk score tested in around 52,600 individuals participating in five clinical trials that focused on cardiovascular and metabolic conditions.
"Our results highlight the importance of ancestry-specific and cross-ancestry genomic studies for the transferability of genomic risk predictions across populations, and the urgent need to substantially increase participant diversity in genomic studies, especially from the most underrepresented regions such as Africa, to avoid exacerbation of health disparities in the era of precision medicine and precision public health, " the authors wrote.