Pharmacogenomics: Getting Personal with Your Medicine


The age of personalized medicine is upon us. Well….almost.

by Grace Stafford

The scene is your doctor's office. You're in the exam room for your annual physical. During the “open wide and say ahh” routine, the doctor takes a swab and rubs it along the inside of your cheek. You don't have a sore throat; there's no need to culture for bacteria. So what's with the swab? Those few cells that have been scraped off your cheek contain all your genetic information. The next time you are sick, that information may be a major factor in your doctor prescribing a particular drug. Or, perhaps more importantly, it will be used in deciding which particular drug not to prescribe. The age of personalized medicine is upon us. Well….almost.

In June of 2000, United States President Bill Clinton and British Prime Minister Tony Blair together announced the completion of the first draft of the human genome. This was a major milestone in the Human Genome Project, a thirteen-year-long task to determine the sequences of all the DNA strands that make up the genetic material of a human. The U.S. Department of Energy's Office of Science (1) predicts many benefits from the Human Genome Project. Among those are the ability of Pharmacogenomics to give us:

•  More Powerful Medicines

•  Better, Safer Drugs the First Time

•  More Accurate Methods of Determining Appropriate Drug Dosages

•  Advanced Screening for Disease

•  Better Vaccines

•  Improvements in the Drug Discovery and Approval Process

•  Decrease in the Overall Cost of Health Care

Pharmacogenomics and pharmacogenetics refer to the study of the interaction of genes (or their encoded products) and drugs. Pharmacogenomics is the newer term, arising from the science of genomics, the study of the complete genetic composition of an organism. The National Center for Biotechnology Information (NCBI) is a division of the United States National Institutes of Health (NIH) devoted to developing information technologies to help understand the underlying molecular and genetic processes that control health and disease. NCBI defines pharmacogenomics as “a science that examines the inherited variations in genes that dictate drug response and explores the ways these variations can be used to predict” a patient's response (or lack of response) to a particular drug (2). Some people distinguish between the two terms, for example, the pharmaceutical company GlaxoSmithKline differentiates pharmacogenetics (the relationship between genetic variation and drug response) from pharmacogenomics (the study of the genome and its products - RNA and proteins - as related to an individual's response to medicines)(3). Others use the terms interchangeably, and NCBI considers the distinction arbitrary.

We speak of “the” human genome, but, except for identical twins, each person has a unique set of genetic material. Genes come in many variations, called alleles. Frequently, these differences arise from the alteration of a single base pair along the gene. These are called single nucleotide polymorphisms, or SNPs (pronounced “snips”). To be considered a SNP, rather than a rare mutation, the variation must occur in at least 1% of the population. SNPs occur every 100 to 300 bases in the human genome, both in the coding (genes) and non-coding regions. A SNP may have no effect on the function of the gene product, or it might significantly alter some property. A SNP within a gene can alter the amino acid sequence, leading to changes in the function of the protein. A SNP in a non-coding area might occur within a gene-regulation segment, altering the amount or timing of protein expression.




Although the Human Genome Project placed the spotlight on pharmacogenomics and the promise of personalized medicine, pharmacogenetics has been a field of study since at least the 1950s. Motulsky (4), a pioneer in the field, credits Friedrich Vogel for coining the term in 1959 to describe the genetic role in drug response. Motulsky was among the first to recognize that heredity could explain the differences among individual's responses to drug efficacy and toxicity(5). One famous example of an inherited response to a drug was discovered when soldiers in World War II were treated with the antimalarial drug primaquine. Some developed a severe anemia due to the destruction of red blood cells. Doctors noticed that there was a strong racial correlation, with blacks being far more susceptible to the anemia. Researchers in Chicago determined that the biochemical reason for the anemia was a reduction in the level of a particular enzyme in red blood cells, Glucose-6-phosphate dehydrogenase (G-6-PD)(6). This reduction caused no visible signs of defect in the red blood cells. The carrier was perfectly healthy until he took primaquine, when the consequences were very severe. G-6-PD has over a hundred different known variations. The variant(s) that are responsible for the sensitivity to primaquine also convey resistance to malaria, and hence are more frequent in populations that derived from areas where malaria has long been endemic.

The anemia due to primaquine is what the pharmaceutical companies describe as an adverse drug reaction. Perhaps the best publicized occurrence of an adverse drug reaction was the increased incidence of heart attacks and strokes attributed to the use of the arthritis drug Vioxx. Merck had to withdraw its blockbuster drug on September 30, 2004 (7). The genetic basis of how Vioxx increases the risk of heart attack is not known, but if it were to be discovered, it might be possible for Merck to bring the drug back to market for specific populations of those not at risk. There is precedence for such an action. The United States Food and Drug Administration (FDA) allowed the drug Lotronex back on the market under strict guidelines for prescription, as patients and physicians argued that it was frequently the only drug available to treat intractable irritable bowel syndrome (8).

The standard parameters your doctor uses today to determine what dose of drug to prescribe are your weight and, possibly, age. This assumes that there is an “average” dosage range that benefits most people. This often works, but, as we see from the above examples, it can sometimes result in devastating consequences. Figure 2 shows how differing efficacy of a drug in different types can affect the likelihood of an adverse effect.


The red line indicates the drug dose at which adverse effects occur, increasing frequency with increasing dose. If a patient is a high responder, the drug in question has good efficacy at a dose safely below that where it induces adverse effects. The intermediate responder doesn't get as much benefit from the drug before adverse effects may kick in. The poor low responder doesn't receive much, if any benefit from the drug. If the doctor knew the responder level of the patient, she could determine if this were an appropriate drug and prescribe the best dosage.

Pharmacogenomics offers the hope of reducing adverse drug reactions and shortening the time needed to bring a new drug to market (9). It may also encourage the development of drugs that don't fit the “blockbuster” paradigm that has driven the pharmaceutical companies in the traditional mode of drug development. Blockbuster drugs are those that produce sales of one million dollars or more per year. Drugs with a more limited market can't approach those sales, but drugs proven efficacious and safe for specific populations could make up the difference, both in reduced costs as well as increased sales. This could be of benefit to the less developed countries, as Daar and Singer point out in their opinion piece (10). Pharmaceutical companies will need to look beyond North America and Europe, both for markets and, from the scientific viewpoint, for the genetic diversity necessary to provide the data for pharmacogenetics to work.

So after a fifty plus year history, and many promises, has anything been delivered yet? We certainly do not understand the many complex ways our genes interact with our environment, be it the food we eat or the drugs we take. But on a smaller scale, pharmacogenomics has delivered life altering and even life saving results. There are a few genes where some specific alleles are known to be involved in drug metabolism or disease progression.



Azathioprine (AZA) is a drug used to treat rheumatoid arthritis. It has also been used to suppress rejection of transplanted organs, to treat inflammatory bowel disease and to treat multiple sclerosis. However, some patients taking AZA will develop severe myelosuppression, that is, the bone marrow is unable to produce blood cells. This is a potentially fatal adverse drug response. The majority of people (about 90% of whites) have high activity of the enzyme thiopurine methyltransferase (TPMT), which metabolizes AZA to inactive products. These people can generally tolerate a fairly high dose of AZA. Some people (~11%) have an intermediate level of TPMT, meaning they may be able to use AZA, but at a reduced dose. Finally, about 0.03% lack TPMT activity and should not take AZA, as they will metabolize AZA to thioguanine nucleotides, which accumulate and kill the bone marrow (11). The therapeutics and diagnostics company Prometheus Laboratories Inc. markets a genotyping test to determine if patients carry one of two the most common TPMT gene variations that would put them at risk with AZA therapy (12). These variations are called TPM*2 and TPM*3A. TPM*2 has one SNP that changes an amino acid in the protein, resulting in a 100-fold reduction in enzymatic activity. TPM* 3A has two SNPs, causing a 200-fold drop in activity (13).

People who are treated with AZA still must have their blood count and enzyme profile monitored. While the three SNPs account for much of the adverse drug reaction, some patients without any of these variations will still develop liver abnormalities, indicating additional complexity to the pharmacogenetics of AZA that have yet to be resolved (14).



Although Iressa showed great promise in animal tests, in clinical trials most patients with non-small cell lung cancer (NSCLC) did not respond to the drug. However, about 10% of the patients showed dramatic improvement. Further investigation demonstrated a correlation between certain mutations in a cell surface protein called EGFR (for Epidermal Growth Factor Receptor) and a positive response to Iressa (15,16). EGFR is what is known as a tyrosine kinase. That is, it attaches a phosphate group to a protein, specifically on the amino acid tyrosine. This starts a cascade of signals through the cell that stimulate growth and cell division. These mutations caused the kinase to be continuously on. Iressa specifically inactivates this kinase. Thus it was effective only against cancers that expressed these constantly active proteins. Doctors can now test cancer patients for the presence of mutated and/or overexpressed EGFR.

One aspect these examples have in common is that a single gene is involved in the primary interaction of the drug with the body. These are, in a sense, the low hanging fruit of pharmacogenomics. Most drug effects are multifactorial. Treatment outcomes and risk of adverse reaction are the result of interactions among several genes, plus the environment, including lifestyle choices, such as smoking and drinking.



So while some forms of personalized medicine are showing up in the clinics, much work has yet to be done before your genetic makeup becomes a routine part of your doctor's diagnostic toolbox.

First of all, more data needs to be gathered to determine what genes (or genetic profiles) are associated with clinical outcomes. Presenters at the 2006 Beyond Genome conference called for pharmaceutical companies to share their data on pharmacogenomic studies in order to get momentum going in the drug development world (17).

As the data are gathered and utilized, pharmacogenomics should reduce the time and cost of developing a drug. Modeling studies demonstrate that utilizing pharmacogenomic information could increase the positive response rate for a test drug from less than 30% to greater than 60% of the patients in simulated clinical trials (18). These simulations used scenarios where only a single gene with two alleles determined the rate of drug metabolism. However, similar results were obtained modeling more complex scenarios where two or three different alleles had independent effects on drug metabolism.

There are, of course, ethical issues to gathering and using these data. Proper safeguards must be implemented to ensure the patient's rights are maintained (19). But the potential benefits to society, as well as to individuals, argue for the necessity of implementing pharmacogenomic information in the drug development and clinical settings.

Finally, while the drug companies wrestle with the implementation of pharmacogenomics, the technology continues to advance so that the diagnostics will likely be available when needed. Right now, the cost to sequence any one person's genome would be prohibitive. But companies are working to improve the technology to bring the cost down (20) for that day when your doctor swabs those cheek cells.



1) Human Genome Project Information, “Pharmacogenomics”

2) National Center for Biotechnology Information “One Size Does Not Fit All: The Promise Of Pharmacogenomics”

3) GlaxoSmithKline “Pharmacogenetics and Pharmacogenomics”

4) Motulsky, A.G. and Qi, M. “Pharmacogenetics, pharmacogenomics and ecogenetics” Journal of Zhejiang University - Science B 7(2): 169-170 (2006)

5) Motulsky, A. “Drug reactions, enzymes and biochemical genetics” JAMA 165: 835–837 (1957)

6) Beutler E. “The Hemolytic Effect of Primaquine and Related Compounds: a Review” Blood 14: 103-139 (1959)

7) Kaufman, M. “Merck Withdraws Arthritis Medication” Washington Post October 1, 2004 pA.01

8) Center for Drug Evaluation and Research “Lotronex Tablets (alosetron hydrochloride): Questions and Answers”

9) Guo Y, Shafer S, Weller P, Usuka J, Peltz G. “Pharmacogenomics and drug development” Pharmacogenomics 6(8):857-64 (2005)

10) Daar AS and Singer PA “Pharmacogenetics and geographical ancestry: implications for drug development and global health” Nat Rev Genet. 6(3):241-6 (2005)

11) Black AJ, et al. “Thiopurine methyltransferase genotype predicts therapy-limiting severe toxicity from azathioprine” Ann Intern Med. 129(9):716-8 (1998)

12) PRO-PredictRx TPMT®

13) Yates CR et al. "Molecular diagnosis of thiopurine S-methyltransferase deficiency: genetic basis for azathioprine and mercaptopurine intolerance" Ann Intern Med 126(8):608-14 (1997)

14) Black AJ et al. “Thiopurine methyltransferase genotype predicts therapy-limiting severe toxicity from azathioprine” Ann Intern Med 129(9):716-8 (1998)

15) Petsko, G.A. "Pharmacogenomics arrives" Genome Biology 5:108 2004

16) Paez, J.G."EGFR Mutations in Lung Cancer: Correlation with Clinical Response to Gefitinib Therapy" Science 304:1497-1500 (2004)

17) Lakhman, K. "Pharmas Must Share PGx Outcomes Data In Order for Targeted Medicine to Catch On" GenomeWeb Daily News June 20, 2006

18) Guo, Y. et al. "Pharmacogenomics and Drug Development" Pharmacogenomics 6(8): 857–864 (2005)

19) Nuffield Council on Bioethics “Pharmacogenetics: ethical issues” (2003)

20) Wade, N. “The Quest for the $1,000 Human Genome” New York Times New York, N.Y. Jul 18, 2006. pg. F.1

You must be logged in to post a comment.