By Monica Heger
Despite being the third best-selling drug in 2010, as many as one-third of cardiology patients do not respond to Bristol-Myers Squibb's blood-thinning drug Plavix (clopidogrel).
While recent studies have uncovered one genetic variant highly predictive of response — the CYP2C19*2 allele, it doesn't explain all non-responders.
Now, a team from the Scripps Translational Science Institute is using exome sequencing to try to uncover other variants that predict whether a patient will respond to Plavix.
Sam Levy, director of genomic sciences at Scripps, led a study to sequence the exomes of 192 patients from a cohort of patients enrolled in a separate Scripps-led clinical trial of Plavix response, and presented preliminary results at this month's Future of Genomic Medicine meeting in San Diego, Calif.
The results pointed to already known genes — CYP219C and CYP218C — and also identified two novel genes — TIAM2 and ATP2B2 — that were associated with response.
Now, he is following up by genotyping the variants in more than 1,000 patients. If the variants are confirmed in a larger population, Levy said they could be used to create an improved test for Plavix response.
He is working with the molecular diagnostic company Molecular Response, which provided funding for the exome sequencing study, to design such a test. He is also using a similar exome sequencing strategy to look for variants that are predictive of response to other cardiology drugs.
The goal is to combine the variants identified in this study with the existing variants to "produce a better test because you would be able to characterize a larger fraction of individuals in relation to their response," Levy told Clinical Sequencing News.
Patients for the exome sequencing study were selected from a larger cohort of patients involved in a clinical trial led by Matthew Price at Scripps known as the GIFT cohort.
Those patients were evaluated for Plavix response at 12 hours, 24 hours, and 30 days, and given a score correlating with platelet aggregation that is used to determine whether a patient is responding to the drug.
Then Levy and his team selected a subset of those patients, which spanned a range of response including both responders and non-responders, as well as those dubbed partial responders because their platelet score decreased over the 30 days, but not below a set threshold used to define response.
The group performed exome sequencing on the 192-patient subset using Agilent's SureSelect and the Illumina HiSeq 2000. Aside from capturing all the exons, the team also targeted flanking sequences around the exons for about 100 megabases of total sequence. After doing some initial quality control and variant filtering to remove outliers, they analyzed 147 samples.
Three gene loci were associated with response, including the known CYP2C18 and CYP2C19 genes, and two novel genes TIAM2 and ATP2B2.
Variants to CYP2C18 and 19 affect an individual's ability to metabolize the drug. One allele, *2, is already tested for before prescribing the medication, because having that allele is a strong predictor of not responding. In Levy's exome sequencing study, around 22 percent of the non-responders had the *2 allele and only 2 percent of responders had the allele.
TIAM2 and ATP2B2 do not impact drug metabolism directly, but instead affect the downstream targets of Plavix, said Levy.
The team discovered three intronic SNPs in the ATP2B2 gene in regions thought to be involved in chromatin regulation. The ATP2B2 gene functions in the calcium channel pathway and helps control calcium homeostasis, which is necessary for platelet response.
The ATP2B2 variants were found in around two-thirds of non-responders and about one-third of responders. Looking at 1,000 Genomes data found that the variants are present in around 26 to 30 percent of the overall human population.
The team also identified more than 10 SNPs in the TIAM2 gene, most of which were only weakly associated with Plavix response. The most significant variant was found in 22 percent of non-responders and in 39 percent of responders. The TIAM2 gene helps regulate the RAC1 signaling pathway, which plays a role in platelet aggregation.
While the variants still need to be confirmed, Levy said that the gene function makes sense. For instance, TIAM2 activates the RAC1 gene, which "plays a central role in platelet aggregation," he said. The most significantly-associated variant in that gene was predicted by the SIFT prediction tool to be damaging, causing inactivation of TIAM2.
So, "you've knocked down the function of TIAM2, which knocks down the function of RAC1, which knocks down platelet aggregation," he said. So it makes sense that this variant would be elevated in responders.
Similarly, while the function of the variants in ATP2B2 are unknown and occur in intronic regions of the gene, they're found in a region that's thought to be involved with chromatin regulation. The ATP2B2 gene controls calcium levels in the cell, which is important for platelet aggregation.
The two genes aren't "directly associated with Plavix metabolism, but they are directly associated with drug response," and impact targets of Plavix, said Levy. So, the variants "may have relevance for other drugs that target platelet response," he said.
In addition, he said, it was important to analyze the variants only in patients without the *2 allele, to get a better understanding of how predictive they are of drug response. Those patients would be more likely to be able to metabolize the drug in the first place.
Around 103 of the 147 sequenced patients did not have the *2 allele, and of those about 60 percent were Plavix responders and 40 percent were non-responders.
Of patients without the *2 allele, variants to ATP2B2 were found in 88.9 percent of non-responders and in 67.5 percent of responders, while TIAM2 variants were found in 65 percent of responders and in 36.5 percent of non-responders.
The next step is to confirm the findings in additional patients. Levy also wants to figure out what percentage of patients' response cannot be explained by all three loci, which would give a clearer picture of how much genetic variation is still unaccounted for.
Nevertheless, he said that if the variants hold up in additional testing, they could be incorporated into the existing diagnostic test for the *2 allele to create a better test for predicting Plavix response.
He is working with the molecular diagnostic company Molecular Response to develop such a test. One critical requirement will be rapid turnaround time, and ultimately the test would need to be given at point of care so that when patients receive a stent procedure, they could be given the test and a drug prescription immediately, he said. Doctors want to give the patient medication the same day, said Levy, so it's "no good having a test come back five days later."
Have topics you'd like to see covered by Clinical Sequencing News? Contact the editor at mheger [at] genomeweb [.] com.