October 1, 2008 (Vol. 28, No. 17)

Pharmacogenomics Is Increasingly Being Employed to Improve Drug Safety Profiles

It is common knowledge that genetic influences are often significant in predicting how patients will respond to drugs. As a result, companies are now beginning to integrate the genetic component into drug discovery and development to improve safety profiles.

The terms pharmacogenetics and pharmacogenomics are often used loosely and even interchangeabley. According to Dan Burns, Ph.D., senior vp for pharmacogenetics at GlaxoSmithKline, pharmacogenetics can be seen as variation in a patient’s inherited genome, while pharmacogenomics is a broader term, covering anything to do with gene expression that might include, for instance, the genome of a patient’s tumor.

“It is accepted that genetics plays a role in adverse events for all kind of drugs,” Dr. Burns reported at the recent European Science Foundation and University of Barcelona’s “Conference on Pharmacogenetics and Pharmacogenomics.”

Dr. Burns presented research pertaining to GSK’s abacavir, a nucleoside reverse-transcriptase inhibitor with activity against HIV that is typically used in combination with other antiretroviral drugs. The most significant adverse effect of abacavir is a hypersensitivity reaction, affecting between 5 to 8% of patients during the first six weeks of treatment. The hypersensitivity reaction includes gastrointestinal, skin, or respiratory symptoms. It can be hard to distinguish from concomitant infection, reaction to other drugs, or inflammatory disease. Withdrawal of abacavir results in rapid resolution of the hypersensitivity reaction, but rechallenge with the drug can result in a renewed reaction that may be more rapid and severe.

There is a known association between a diagnosis of hypersensitivity reaction to abacavir and carriage of the major histocompatibility complex allele HLA-B*5701. In the Predict-1 study, a group of HIV positive patients was screened for HLA-B*5701 and not given abacavir if they tested positive. In a control group, no screening was performed, and all patients were given the drug.

Screening eliminated immunologically confirmed hypersensitivity reactions, which were found in 2.7% of the control group. The hypersensitivity reaction was clinically diagnosed in 3.4% of the prospectively screened group, compared to 7.8% of the control group. The findings show that a pharmacogenetic test can be used to prevent a specific toxic effect of a drug.

“This is one of the first examples where a genetic marker has been used prospectively to demonstrate its ability in predicting adverse events,” commented Dr. Burns.

Following discussions with regulatory agencies, changes were made in the labeling of abacavir to reflect the utility of pharmacogenetic testing for HLA-B*5701. Currently the test is available through clinical labs rather than at point of care. “Physicians will use their clinical judgement over whether to offer a patient a pharmacogenetic test,” he noted.

Human Ether-A-Go-Go

“We are trying to use genetics more and more in our preclinical screening strategies to mitigate against risk,” said Martin Amstrong, Ph.D., principal scientist in the R&D genetics group at AstraZeneca. Dr. Armstrong presented an example involving the pharmacological characterization of human ether-a-go-go (hERG) gene-encoded potassium channel variants in a preclinical screen. A number of drugs block hERG and can sometimes cause QT prolongation, a feature in the cardiac cycle that can lead to arrhythmias.

In this study, AstraZeneca researchers wanted to know if inclusion of some of the known genetic variants of hERG in the preclinical screen could help eliminate compounds likely to have adverse effects related to QT prolongation in the general population.

“We are trying to turn the paradigm of translational medicine around to ensure that we can screen out as much risk as possible early on,” Dr. Armstrong explained. “One approach we take is to use information obtained in clinical work and back translate it into something we can use preclinically.”

The AstraZeneca team selected nine hERG gene variants from literature and from population screening. Normally, the preclinical screen would use just wild type hERG expressed in a heterologous system as the target. This time, the gene variants were also included in the preclinical screen in an attempt to gain an idea of the risk of a compound blocking hERG, as might occur in a real clinical population. “We found that there was no difference in sensitivity of the variants against our 48 reference compounds,” reported Dr. Armstrong.

“This validated our current preclinical screening strategy.” That is, if there is a population risk associated with these hERG variants, then it is not detectable preclinically within the sensitivity of the present assay systems (although the variants could still be important on an individual basis in vivo). AstraZeneca is now extending its pharmacogenomic approach to other preclinical screens.


Nadya Oks, senior product marketing manager at Affymetrix, described work the company has carried out on a genetic variant that alters the dosage of warfarin. Affymetrix, in collaboration with Marshfield Clinic, has used the firm’s Drug Metabolizing Enzymes and Transporters (DMET) solution to discover a genetic variation that is associated with an individual’s response to warfarin.

The researchers, led by Michael D. Caldwell, M.D., Ph.D., of Marshfield Clinic, discovered a genetic variation in CYP4F2, a member of the cytochrome CYP450 drug-metabolizing family, which explains approximately 8% of a patient’s variable response to warfarin. The findings, published in Blood Online, could help reduce the serious adverse effects caused by incorrect warfarin dosage and enable doctors to treat patients with a more effective, personalized form of medicine, Oks said.

“This discovery should enable us to better predict a stable therapeutic dose and hopefully reduce the overall complications of warfarin therapy,” said Dr. Caldwell. “Understanding the factors and genetic variations that affect the metabolism of drugs gives doctors a much more accurate idea of how best to treat patients. The nation’s entire healthcare system can benefit from these types of advances.”

The use of warfarin is complicated due to its small therapeutic dosage window. If a dose is too high, there is a risk of bleeding; if a dose is too low, the drug will be ineffective and there may be a risk of clotting. The optimum dosage is affected by factors such as genetics, age, body mass index, and gender.

In August 2007, the FDA announced that it would be updating the labeling of warfarin with known valid markers in the CYP2C9 and VKORC1 genes to provide information on how an individual’s genetics may impact their response to the drug.

In January, Affymetrix launched its DMET Early Access solution, an updated version of the assay used in the Marshfield warfarin study. The Early Access version of the product profiles more than 1,000 drug metabolism markers including 172 core genetic markers, according to the company. Data is automatically interpreted into a common format that can be integrated into clinical trial workflows.

Whole Genome Association Studies

Whole genome association studies can be more fruitful than a candidate gene approach in pharmacogenomics, according to Nicole Soranzo, Ph.D., staff scientist in the genetics of common diseases team at the Wellcome Trust Sanger Institute.

The Sanger team is employing commercial chips that allow measurement of up to one million SNPs at a time in an individual sample. The SNPS can be tested either individually or in groups. “The great power of the whole genome approach is that it does not make assumptions about where gene variants will be found, unlike the candidate gene approach,” she said.

“The whole-genome approach is a powerful tool for looking at novel biological mechanisms underlying diseases and how parts of the genome interact with one another.” In pharmacogenetic studies, it is important to know the mode of action of a drug and look for gene variants in either the drug’s target or on its metabolic pathways.

The team has yet to publish whole-genome work on pharmacogenomics, but it has built on candidate gene work on warfarin, where it has published four studies on gene variants that, together, explain up to 60% of the variance in response to the drug, according to Dr. Soranzo. Warfarin is widely used in pharmacogenetics because the phenotypes involved are clear cut and well understood.

“Having good predictors for warfarin upfront saves months of trial and error prescribing and can save the patient from the adverse effects of exposure,” Dr. Soranzo explained.

In pharmacogenetics, the main challenges are assuring QC of the data in both the SNP and the sample itself and having an understanding of the heterogeneity of the population. There is also a need to have sufficient power in studies to pick up rare adverse effect events.

Susan Aldridge, Ph.D. ([email protected]), is a freelance science and medical writer specializing in biotechnology, pharmaceuticals, chemistry, medicine, and health.

Previous articleScientists Unravel How a Gene Encourages Metastasis in Human Cancer Cells
Next articleWyeth Inks $100M Pact for Advanced Life Sciences’ Late-Stage Antibiotic in Asia Pacific Region