By Monica Heger
Just as sequencing is increasingly being used to identify disease-causing variants, researchers are increasingly turning to the technique to pinpoint the variants responsible for drug efficacy and adverse side effects.
"Looking at drug response genetics and looking at disease genetics are very parallel processes," said Ron Krauss, senior scientist and director of atherosclerosis research at Children's Hospital Oakland Research Institute in California. His team is using whole-exome sequencing to try to uncover why some people experience severe muscle toxicity in response to statins, a class of cholesterol-lowering drugs.
Krauss's team is part of the Pharmacogenomics Research Network, which the National Institutes of Health last year awarded $9.2 million over the next five years to implement next-gen sequencing resources (IS 10/7/2010).
The sequencing for Krauss's project is being conducted by Deborah Nickerson's team at the University of Washington School of Medicine.
Krauss identified 16 people who experienced severe muscle toxicity as a result of statins, and their exomes will be compared to exome sequencing data that the University of Washington team has stored from other sequencing projects.
The sequencing has been completed and Krauss said his team is now in the process of analyzing the data.
Richard Weinshilboum, a professor of pharmacology and medicine at the Mayo Clinic who heads a PGRN project that is looking at the pharmacogenetics of phase II drug metabolizing enzymes, said that the goal is for all the projects within the PGRN to eventually incorporate next-gen sequencing.
"The purpose is to find ways to expedite applying sequencing techniques to pharmacogenomics to understand the genetic role in drug response phenotypes," he said.
Weinshilboum's group at the Mayo Clinic recently applied for a grant to use next-gen sequencing resources at the Baylor Genome Sequencing Center to study tamoxifen response in breast cancer patients.
Aside from these PGRN projects, a team from Washington University is using sequencing to study drug response in a clinical trial of hormone therapy among breast cancer patients (IS 8/10/2010), researchers from the HudsonAlpha Biotechnology Institute are planning to use exome sequencing to identify variants responsible for adverse reactions to a drug used to treat Parkinson's disease (CSN 3/16/2011), and a team from the University of British Columbia is using exome sequencing to study adverse behavioral responses to statins.
Traditionally, researchers have used Sanger sequencing of candidate genes to study the genetics of drug response, and recently have moved to using genome-wide association studies. However, Krauss said that he thinks next-gen sequencing will become increasingly used as costs decrease, particularly for rare adverse events, since it will be able to detect rare variants.
Candidate gene sequencing is limited by its lower throughput, and the fact that researchers must first have an idea of which genes to study and therefore may miss variants in unexpected genes. Additionally, GWA studies, while higher throughput, will not detect rare variants.
Weinshilboum said his group at the Mayo Clinic began studying the genetics of drug response about 10 years ago using a candidate gene approach, targeting genes known to play a role in drug metabolism, and also hormones and neurotransmitters. "Now, we're identifying totally unanticipated genes that explain efficacy," he said.
He said that a particularly useful technique is to first do a GWAS to identify regions of interest and then do deep sequencing across the signals. Researchers can now pursue "sequencing across those genes to see if there are rare variants beyond the SNPs that might explain variation in drug response," he said.
The UBC team plans to sequence the exomes of 20 people who have experienced anti-social behavior and aggression as a result of statins. Neal Boerkoel, a senior clinician scientist in the department of medical genetics at UBC who is leading the project, said that epidemiologic literature has found a link between statins and antisocial behavior, plus additional evidence linking disorders of cholesterol metabolism to behavior. This led them to hypothesize that genetic variants could be responsible for the reaction to statins.
For the study, the team is first doing a biochemical analysis, which will help determine what types of variants might be involved, said Boerkoel. The biochemical analysis will be done on cells, studying what happens when they are deprived of cholesterol and also how they grow in the presence of statins.
Since exome sequencing typically yields between 100,000 to 200,000 variants per individual, the biochemical analysis will help the researchers hone in on functionally relevant variants, Boerkoel said.
Rather than "arbitrarily narrowing [the list of variants] to known enzymes of cholesterol synthesis," Boerkoel said the team is first looking at the biochemical disturbance, pinpointing the enzymes involved in that disturbance, and then looking for variants in those enzymes. This method gives an unbiased view, and allows for the potential discovery of variants in enzymes not previously thought to be involved in cholesterol synthesis. Additionally, the team will also use family information to narrow down the candidate list.
Boerkoel said they chose to go with exome sequencing rather than whole-genome sequencing because of cost and data interpretation considerations. "We're still not very smart about interpreting what's between the coding regions," he said.
The initial project is being funded for $250,000, about half of which is from Genome BC.
If the project yields interesting results, Boerkoel said there is the possibility of a larger, follow-up study that would be done either by his group or a separate group. Doing a larger study would pose additional challenges, particularly in identifying enough people with the correct phenotypic profile.
A major hurdle is "definitive phenotyping," he said. "In order to know someone had an adverse response attributable to a cholesterol-lowering drug, they have to have been put on it, gotten sick, taken off, and then gotten better," he said. "There are not many people with that much clinically documented data."
Krauss agreed that identifying appropriate participants is crucial, saying his team identified "with very careful scrutiny" 16 individuals who exhibited extreme muscle toxicity in response to statins.
Another challenge, he said, will be in determining which variants are "related to the trait we're actually interested in." While Boerkoel's group is using biochemical studies to narrow down the variants to include only those related to statins, Krauss's team is making use of the plethora of exome sequencing data that the University of Washington has compiled to narrow down the list of variants.
He also agreed with Boerkoel that the difficulty of data analysis made exome sequencing a more attractive approach than whole-genome sequencing. "The exome is tractable right now in a way that the genome is not," Krauss said.
However, he said that important variants are likely to fall in regions not captured by the exome. What he hopes to do is use exome sequencing to first identify the genes of interest, and then to sequence the region around that gene to "expand the number of functional, relevant SNPs that may have regulatory roles, or affect splicing or RNA stability."
"The approach we've taken, we're really just putting our toe in the water," Boerkoel said. Ideally, the study will lead to interesting data that could be used to generate hypotheses for further work, but "we're entering into it gingerly, because it is uncharted territory."
Have topics you'd like to see covered by Clinical Sequencing News? Contact the editor at mheger [at] genomeweb [.] com.