Enhancing the Promise of Drug Repositioning through Genetics.

Jayne-Louise E Pritchard,Tracy A O’Mara,Dylan M Glubb

doi:10.3389/fphar.2017.00896

Jayne-Louise E Pritchard, Tracy A O’Mara + Show 1 more

Open Access

https://doi.org/10.3389/fphar.2017.00896

Copy DOI

Abstract

The development of new drugs has become challenging as the necessary investments in time and money have increased while drug approval rates have decreased. A potential solution to this problem is drug repositioning which aims to use existing drugs to treat conditions for which they were not originally intended. One approach that may enhance the likelihood of success is to reposition drugs against a target that has a genetic basis. The multitude of genome-wide association studies (GWASs) conducted in recent years represents a large potential pool of novel targets for drug repositioning. Although trait-associated variants identified from GWAS still need to be causally linked to a target gene, recently developed functional genomic techniques, databases, and workflows are helping to remove this bottleneck. The pre-clinical validation of repositioning against these targets also needs to be carefully performed to ensure that findings are not confounded by off-target effects or limitations of the techniques used. Nevertheless, the approaches described in this review have the potential to provide a faster, cheaper and more certain route to clinical approval.

Highlights

Over 6,000 human medical conditions have defined molecular phenotypes (Johns Hopkins University, 2017) but only ∼500 conditions have approved therapies (National Institutes of Health, 2015)
The linking of gene expression and genotype data can be applied at a multi-variant or gene-based level by combining genotype data to determine the cumulative effect of genetic variants on expression (Gamazon et al, 2015; Gusev et al, 2016)
These data are used to predict gene expression levels in cohorts of genotyped individuals, allowing case-control transcriptomewide association studies to examine whether the predicted gene expression associates with clinical phenotypes and the potentially causal genes identified could provide further targets for drug repositioning

Summary

INTRODUCTION

Over 6,000 human medical conditions have defined molecular phenotypes (Johns Hopkins University, 2017) but only ∼500 conditions have approved therapies (National Institutes of Health, 2015). As of November 2017, the GWAS catalog contains ∼53,000 unique variant-trait associations for more than 800 human traits and diseases (MacArthur et al, 2017), likely representing a large number of genes that could provide targets for drug repositioning studies. The linking of gene expression and genotype data can be applied at a multi-variant or gene-based level by combining genotype data to determine the cumulative effect of genetic variants on expression (Gamazon et al, 2015; Gusev et al, 2016) These data are used to predict gene expression levels in cohorts of genotyped individuals, allowing case-control transcriptomewide association studies to examine whether the predicted gene expression associates with clinical phenotypes and the potentially causal genes identified could provide further targets for drug repositioning. The ExSNP database can be used to query genetic variants and genes for associations with gene expression using 16 publicly available human eQTL studies These data allow tissue- and population-specific eQTLs to be identified. An additional approach would be to use genetic correlation analyses, such as LD-score regression, that use GWAS data to identify genetic similarities between diseases which could provide an avenue for further repositioning

LIMITATIONS

CONCLUSION