Precision Medicine: Using Genomic Data to Predict Drug Side Effects and Benefits

Gene Variant and Corornary Heart DiseasePeople with type 2 diabetes are at increased risk for heart attacks, stroke, and other forms of cardiovascular disease, and at an earlier age than other people. Several years ago, the Food and Drug Administration (FDA) recommended that drug developers take special care to show that potential drugs to treat diabetes don’t adversely affect the cardiovascular system [1]. The challenge in implementing that laudable exhortation is that a drug’s long-term health risks may not become clear until thousands or even tens of thousands of people have received it over the course of many years, sometimes even decades.

Now, a large international study, partly funded by NIH, offers some good news: proof-of-principle that “Big Data” tools can help to identify a drug’s potential side effects much earlier in the drug development process [2]. The study, which analyzed vast troves of genomic and clinical data collected over many years from more than 50,000 people with and without diabetes, indicates that anti-diabetes therapies that lower glucose by targeting the product of a specific gene, called GLP1R, are unlikely to boost the risk of cardiovascular disease. In fact, the evidence suggests that such drugs might even offer some protection against heart disease.

Genetic approaches have increasingly been used to identify potentially promising new drug targets. In the study reported in Science Translational Medicine, researchers led by Robert Scott and Nick Wareham from the University of Cambridge, England, and Dawn Waterworth from GlaxoSmithKline, King of Prussia, PA, also wanted to explore whether genomic data could yield important clues about the potential side effects of drugs targeting particular genes.

They began with the results of a previous study of 202 genes encoding potential drug targets in more than 14,000 people [3]. That study had uncovered many examples in which people carried rare genetic variants expected to influence the function of those important genes. The effect on function occurs because a varied gene sequence can give rise to a slightly tweaked protein (either overactive or underactive) that interacts differently with other proteins, in much the same way that a specially designed, or tweaked, drug manipulates interactions among the proteins that it targets.

Scott and colleagues focused on six genes that encode potential drug targets licensed or in development by GlaxoSmithKline for the treatment of obesity or diabetes. Of the 121 variants identified in those six genes, the researchers looked to see whether any had effects on diabetes, obesity, body mass index, and fasting glucose levels that are similar to those of drugs that target the genes.

The search led them to an intriguing variant in the GLP1R gene. The gene encodes the glucagon-like peptide-1 receptor that, when triggered by the GLP-1 hormone, stimulates the release of insulin in response to glucose.

The researchers found the variant is associated with lower fasting glucose and a reduced risk of type 2 diabetes. That suggests that the variant actually increases the effectiveness of GLP1R. To confirm this initial finding, they turned to existing genomic and clinical data from another 40,000 people. Their original finding was correct. The variant lowers glucose levels to protect against diabetes, just as GLP1R-targeted anti-diabetes drugs are designed to do.

The next question was whether this GLP1R variant might also influence a person’s risk of cardiovascular disease or other health outcomes. To find out, the researchers took advantage of genomic and clinical data for almost 60,000 people with heart disease and 160,000 healthy controls available from international data-sharing consortia.

This kind of analysis is currently referred to as “Mendelian randomization”—because the inheritance of genetic variants follows strict Mendelian rules and is not confounded by any other environmental or lifestyle choices. The researchers found that people with the glucose-lowering GLP1R variant were actually less likely to develop heart disease. In fact, the reduced risk of heart disease was greater than would be expected based solely on the gene variant’s known glucose-lowering effects, suggesting there is more to the story than just fasting glucose.

Waterworth notes that the findings come on the heels of early results released in March from a clinical trial testing the cardiovascular safety of the GLP1R-targeted drug Victoza® (liraglutide), which confirm the newly reported genetic findings [4]. While the full trial results have yet to be published, preliminary findings reportedly show a significant reduction in major cardiac events experienced by people with diabetes who took Victoza® for up to five years.

The new study outlines a genomic approach that makes use of several research cohorts of more than 50,000 participants. The hope is that this teaming of genomic and clinical Big Data will help to streamline the drug development process, helping to avoid late-stage failures attributable to lack of efficacy or adverse safety profiles. That’s critical considering that just 1 in 10 drug candidates entering human clinical trials successfully goes on to receive FDA approval [5].

As the Precision Medicine Initiative (PMI) Cohort Program enlists 1 million or more Americans in a participant-engaged effort to assemble genomic data together with electronic health records, behavioral data, and more, these types of data-driven analyses will become more common. The study of rare variants in potential drug targets will undoubtedly uncover many intriguing leads, much like the one reported here.


[1] Guidance for industry diabetes mellitus—Evaluating cardiovascular risk in new antidiabetic therapies to treat type 2 diabetes. Food and Drug Administration. December 2008.

[2] A genomic approach to therapeutic target validation identifies a glucose-lowering GLP1R variant protective for coronary heart disease. Scott RA, Freitag DF, Li L, Chu AY, Ehm MG, Wareham NJ, Waterworth DM, et al. Sci Transl Med. 2016 Jun 1;8(341):341ra76.

[3] An abundance of rare functional variants in 202 drug target genes sequenced in 14,002 people. Nelson MR, Wegmann D, Ehm MG, Kessner D, St Jean P, Verzilli C, Shen J, Tang Z, Bacanu SA, Fraser D, Warren L, Aponte J, Zawistowski M, Liu X, Zhang H, Zhang Y, Li J, Li Y, Li L, Woollard P, Topp S, Hall MD, Nangle K, Wang J, Abecasis G, Cardon LR, Zöllner S, Whittaker JC, Chissoe SL, Novembre J, Mooser V. Science. 2012 Jul 6;337(6090):100-104.

[4] Victoza® significantly reduced the risk of major adverse cardiovascular events in the LEADER Trial. Novo Nordisk. 2016 March 4.

[5] Clinical development success rates for investigational drugs. Hay M, Thomas DW, Craighead JL, Economides C, Rosenthal J. Nat Biotechnol. 2014 Jan;32(1):40-51.


Diabetes, Heart Disease and Stroke (National Institute of Diabetes and Digestive and Kidney Diseases/NIH)

Aetiology of Diabetes and Related Metabolic Disorders (University of Cambridge, England)

Precision Medicine Initiative Cohort Program (NIH)

NIH Support: National Institute on Aging; National Heart, Lung, and Blood Institute; National Human Genome Research Institute; National Cancer Institute; National Institute for Diabetes and Digestive and Kidney Diseases

3 thoughts on “Precision Medicine: Using Genomic Data to Predict Drug Side Effects and Benefits

Comments are closed.