Listen to this article on the DDW Podcast:
Slavé Petrovski and Andrew Harper, Centre for Genomics Research, AstraZeneca, outline the implications and insight offered by the recent publication of an AstraZeneca-authored paper “Rare variant contribution to human disease in 281,104 UK Biobank exomes”.
In August 2021, Nature published the largest and most comprehensive evaluation of the role rare genetic variants have across human disease. Performed in ~300,000 UK Biobank participants, and considering over 18,000 medical conditions, thousands of known and novel gene-phenotype relationships were described that collectively suggest rare genetic variation underpins some of the most common human diseases. This insight is important. For many years, rare variants were known to cause rare disease, but for common diseases, genetic susceptibility was generally attributed to the inheritance of hundreds of common variants, each of small effect, and the interplay they had with environmental risk factors. While it had been previously hypothesised that rare variants may also contribute to common diseases, sequencing costs were prohibitively expensive to formally address the hypothesis.
The formation of the UK Biobank Exome Sequencing Consortium (UKB-ESC), a private-public consortium that funded the exome sequencing (ie. sequencing the protein coding regions) of 500,000 UK Biobank participants, enabled assessment of the contribution of rare variants towards rare disease. The UK Biobank is a large, population-based, longitudinal study of 500,000 individuals from across the United Kingdom. AstraZeneca is part of the UKB-ESC and the team used the exome sequencing data, electronic health records and available biomarkers collected on 269,171 UK Biobank participants to assess the impact rare variants have on 18,780 phenotypes of interest. While the impact individual rare variants made towards a phenotype was assessed, most insights were derived from a statistical test that aggregated multiple rare variants, each with similar pre-defined characteristics together (eg. protein truncating variants), across a gene. In total, 46,837 variant-level and 1,703 gene-level statistically significant relationships were reported. These gene-phenotype associations included both well-established relationships and new findings that could help uncover novel disease biology and identify new therapeutic targets validated by human genetics.
The importance of genetic architecture
Understanding the genetic architecture of a disease is important when considering underlying disease biology and when evaluating tractable drug targets. An era of genome-wide association studies indicated that for most common diseases, an individual’s genetic susceptibility towards disease was influenced by the presence of many common genetic variants, each exerting a modest overall effect on their propensity towards disease. While hundreds of common variants have now been described for many common diseases, the process of translating common variant association signals into biological insights has proven challenging. And, turning these modest effect signals into medicines has been even more challenging. Several potential reasons contribute towards this, including the well-recognised fact that most common variant signals occupy non-coding genomic regions. Furthermore, although common variants are likely to contribute towards the regulation of gene expression, the specific genes under regulation and the cell types in which they operate often remain unclear. In contrast, rare variants tend to confer large biological and thus clinical effects with pronounced functional consequences and have long been recognised as the cause for many monogenic diseases, such as hypertrophic cardiomyopathy or familial hypercholesterolaemia. More recently it has been recognised that rare loss-of-function variants can also protect against disease, such as the now well-established example between PCSK9 protein truncating variants and low levels of LDL cholesterol. Given that it is generally easier to inhibit than activate a target, such examples of naturally occurring genetic inhibitors of drug targets have been considered to provide human validation for what an inhibitor or degrader may achieve, in terms of both safety and efficacy, and have consequently proven attractive for clinical development activities.
The study found that significant gene-phenotype relationships from this biobank population setting were seven-fold enriched for targets of FDA-approved medicines. This information can be used to further drug discovery and development and create a beneficial impact on the clinical success of human-validated drug targets. We know from AstraZeneca’s 5Rs framework (right target, right tissue, right safety, right patient population, right commercial potential) that selecting the right target is one of the most important decisions that needs to be made in drug discovery. Research has shown that drug candidates targeting genes clearly linked to human disease are much more likely to demonstrate clinically meaningful efficacy outcomes. This study, based on the in-depth analysis of rare variants in common diseases, suggests that clinically relevant gene-phenotype relationships were enriched seven-fold for targets of FDA-approved medicines. This is higher than previous estimates traditionally looking at either rare variants in rare diseases (Mendelian genetics) or common variants in common diseases (genome-wide association studies), underscoring the importance of human genetics in target identification and drug discovery. Whilst incorporating human genetics into target selection is anticipated to improve the overall probability of success for a molecule, human genetics can also be used to help inform precision medicine strategies and influence lifecycle management opportunities.
The value of a pan-ancestry based approach
To date, the vast majority of genomic research has been conducted using DNA from individuals of European ancestry, and as such the findings tend not to be representative of global populations and may result in future health inequalities. In this study, rare variant based gene-phenotype relationships were explored using multiple ancestral groups available within the UK Biobank, rather than solely focussing on individuals of European ancestry. By combining genetic analyses from across multiple populations, including European, African, East Asian and South Asian individuals, additional gene-phenotype relationships were uncovered, highlighting the scientific value of adopting a pan-ancestry based approach. However, it is also apparent that further work is required to rebalance ancestral diversity within large-scale genomic studies before global populations are appropriately represented – something AstraZeneca is committed to focus on. In this study, individuals of European descent, accounted for 96% of those included in the pan-ancestry UK Biobank analysis, yet reflect only 16% of the global population.
Volume 23, Issue 1 – Winter 2021/22
About the authors
Slavé Petrovski, Vice President, Head of Centre for Genomics Research, Discovery Sciences, R&D
Andrew Harper, Medical Director & Respiratory & Immunology Therapy Area Lead, Centre for Genomics Research, Discovery Sciences, R&D