NEW YORK (GenomeWeb) – Seeking to exploit the extensive genomic data and electronic health records at its fingertips, a team of researchers at the Estonian Genome Center recently embarked on a pharmacogenomic study of the country's population.
According to investigators, the effort is already bearing fruit, as a screen of the country's e-health databases combined with an assessment of genetic data from the center's 52,000-sample biorepository enabled their team to validate several variants known to be connected with adverse drug reactions. They were also able to identify a new gene associated with adverse reactions to a class of anti-inflammatory drugs.
The researchers presented their work at the European Society of Human Genetics annual meeting, held in May in Copenhagen, Denmark. Lili Milani, senior scientist and head of pharmacogenomics at the center, along with Kristi Krebs, a PhD student in her group, met recently to discuss their work with GenomeWeb at the University of Tartu, of which the Estonian Genome Center is part.
According to Milani, the pharmacogenomics study has its roots in the creation of a reference panel for the Estonian population, based on 2,500 whole genomes sequenced at 30x and 2,500 whole exomes sequenced at 80x in partnership with the Broad Institute and Nestlé Health, respectively.
The center also recently finished genotyping all the samples in the biobank using the Illumina Global Screening Array, a deal they announced last year. Scientists from the center described their use of the genomic data to impute genotypes in a paper published in the European Journal of Human Genetics in April.
Using these resources, the center is working toward introducing genomic medicine via the country's e-health infrastructure by feeding medically relevant genetic information into the online medical records system. The endeavor prompted Eric Lander, founding director at the Broad, to say that Estonia is "at the cutting edge of what is possible" in genomic medicine, and call it a "model system" during a visit to Tartu in 2014.
Milani said that with the reference panel and array data in hand, various teams within the center have decided to embark on different projects. As her team works with pharmacogenomics, it decided to work on adverse drug reactions in the population.
The researchers first began to search the country's e-health database for instances of adverse drug reactions denoted by ICD-10 medical classification codes. They also obtained information on prescribed drugs by looking at accompanying records in the Estonian Health Insurance Fund. At the same time, the researchers looked at an ADME panel of 64 genes related to the absorption, distribution, metabolism, and excretion of drugs across the whole genome data in the reference panel data set, plus 14,000 imputed whole genomes using the array data that had been generated at the time. In total, they looked at 16,500 individuals, linking their recorded medical and genomic data.
According to Milani, they observed 622 variants in the coding regions and 388 variants in the regulatory regions of the genes in question and were able to validate several previously documented genetic variants associated with drug-induced adverse drug reactions. The researchers also identified a new gene,CTNNA3, that was associated with adverse drug reactions among individuals treated with oxicams, Milani said. They replicated the finding in an extended cohort from the biobank.
The findings led the researchers to state at ESHG that combining EHRs with genomic data and validation of adverse drug reactions is a way to identify people at risk for unexpected drug responses, but has also raised questions about how to best report that information back to patients.
Krebs said that the best means for communicating pharmacogenomics data to patients and physicians is still under discussion, though she is hopeful it will be accomplished in Estonia. "We are thinking of providing feedback to doctors, so that if a patient has a variant, and he or she is prescribed a certain drug, the doctor can advise on dosage," she said. Krebs noted that the reporting of ADRs is still infrequent in Estonia and that a lot of cases most likely go unreported, meaning it's hard to gauge how significant a problem it is in Estonia, which has a population of 1.3 million people. She said 400 instances had been reported by the Estonian Agency of Medicines in the past year.
Milani said that the country's ongoing personalized medicine program, which is administered by its Ministry of Social Affairs, could eventually lead to the creation of a national data infrastructure for reporting pharmacogenomic data back to doctors and patients. An important phase of the Estonian personalized medicine program, which includes extensive training of doctors in genomic medicine, commenced last year and is stated to run through 2020. The government is also ready to spend €1.5 million ($1.7 million) on consortiums that will drive two initial pilot projects focused on the implementation of genomics-based risk prediction for breast cancer and cardiovascular diseases, and Milani suggested that pharmacogenomics could be a third area of focus. The Estonian Genome Center is a partner in the new program.
Milani's team would also like to follow up on some of the findings of its study. "For example, we would like to contact and phenotype the individuals carrying the loss-of-function variants we identified in the drug metabolism genes," Milani said. She suggested that those identified could be provided with low doses of the drugs to monitor their response, by measuring the drug metabolite in blood samples taken over time.
While the Estonian researchers determine the best course forward for making pharmacogenomics available and actionable for patients in the country, the same center is already piloting the communication of genetic risk for hereditary cancers to biobank participants.
Researchers at the center decided to report back data related to pathogenic variants in the BRCA1 and BRCA2 genes upon request, citing American College of Medical Genetics and Genomics recommendations on 59 genes that should be reported back if found mutated via clinical sequencing, as well the Estonian Human Genes Research Act, which stipulates that participants have a right to know about their genetic data.
Neeme Tõnisson, head of clinical genetics at the center, reported on the effort in a second presentation at ESHG, noting that they had handled requests from nine participants to date who had either been diagnosed with cancer or had a family history of it. They also noted that, based on the reference panel, the prevalence of BRCA1 and BRCA2 high-risk variants in Estonia is approximately one in 125.
"It will be extremely important to establish a close collaboration with clinicians for a proper follow-up," the researchers said in their ESHG abstract. "Giving feedback on high-risk genetic variants to currently healthy individuals may have a large psychosocial impact," they noted. "This needs to be further studied, together with the individual's compliance to management recommendations."
"This is the first time someone has returned findings from the biobank," said Andres Metspalu, director of the Estonian Genome Center. "We are doing it in a pilot phase, trying to find the best way to return results and it looks like people appreciate it," he said. "The question is how to best transfer this to the medical system."
Milani noted that the reporting of mutations related to hereditary cancers like breast cancer, and the reporting of pharmacogenomics information is likely to differ going forward, as variants related to drug metabolism are not seen as imminently life threatening as the BRCA1 and BRCA2 mutations. However, starting in September, the center will provide biobank participants with pharmacogenomic data upon request, similar to what has been done with the hereditary cancer data.
"People can register and come in for counseling and they will be informed about their risk for various diseases, including cardiovascular diseases, type 2 diabetes, as well as breast cancer, and their pharmacogenetic profile," said Milani. "Then they will be informed if they should avoid certain drugs, or if they will need higher or lower dosages of them."