Pharmacogenomics:
Changing the health care paradigm

Understanding the role of genetic polymorphisms in drug response will facilitate drug efficacy and decrease adverse effects by helping scientists to tailor medication to a patient’s genetic makeup. Pharmacogenomics should profoundly change the way companies develop and market medicines. Pharmacogenomics can be leveraged throughout the research and health care community to aid in target discovery and validation, prioritize and optimize lead compounds, evaluate preclinical efficacy and safety, stratify patients enrolled in clinical trials, and create predictive and diagnostic tests.

Genetic variations
It is estimated that most commonly used drugs will be effective only in 30–60% of patients with the same disease. Additionally, a subset of these patients may suffer severe side effects. It has been reported that adverse drug reactions were the fourth to sixth leading cause of death in the United States in 1994 and that they have a $100 billion annual economic impact (1). For patients who survive, adverse drug effects negatively impact their health and quality of life and add a substantial financial burden to society. Severe adverse effects led to the withdrawal of several potential blockbuster drugs from the market, including Rezulin, Seldane, Redux, and Pondimin. More effectively predicting drug efficacy and toxicity offers significant benefits to society.

SNP-based genetic association studies have been extensively applied for common disease gene identification, such as in the case of Type 2 diabetes (2). A high density of markers will be necessary to identify detectable levels of SNPs with allelic association to the disease mutation. Statistical analysis based on haplotypes can provide additional information with respect to tests of significance and fine localization of complex disease genes such as those that causeAlzheimer’s disease (3). Genotyping large numbers of SNPs in linkage and association studies will assist in the understanding of complex disease traits, including many common human diseases and individual variations in drug responses.Identifying disease predisposition genes offers the potential for early diagnosis and development of therapies targeting them.

Like complex diseases, individual variations in drug response often can be linked to genetically based differences in drug metabolism and drug actions. A dose that produces the desired therapeutic response in one individual may not be efficacious in another or may even be toxic. Alterations in drug-metabolizing enzymes and genetic variations in receptors and transporters can produce variations in drug response (4). Pharmacokinetic variations are well-characterized genetic traits responsible for different drug responses. Recent studies have identified important variations in almost all drug-metabolizing enzymes, which can lead to variable metabolic profiles among individuals (4).

Different metabolizers

Figure 1. Plasma drug concentrations in different patients after receiving the same doses of a drug metabolized by a polymorphic enzyme.

An example of pharmacokinetic outcomes of a hypothetical drug metabolized by a polymorphic enzyme is illustrated in Figure 1. Normal metabolizers (wild-type) have plasma concentrations within the therapeutic range and benefit from the therapy. Poor metabolizers have plasma drug concentrations exceeding the maximal no-effect concentrations and are at risk of experiencing drug side effects and toxicity. Ultrafast metabolizers do not reach therapeutic concentrations and fail to achieve the desired therapeutic effect at regular dose regimens. When a new compound has a narrow therapeutic window, slight variations in plasma drug concentration can point to the drug’s success or failure for a particular treatment.

One example of a genetic mutation in certain patient populations that can produce a deleterious effect in drug responsiveness has been shown in thiopurine methyltransferase (TPMT), a drug-metabolizing enzyme that catalyzes the S-methylation of thiopurine drugs such as 6-mercaptopurine and azathioprine. Patients who are homozygous mutants have lower enzyme activity. These patients are at high risk for life-threatening toxicity such as myelosuppression at standard doses. Conversely, subjects with genetically high TPMT activity may be undertreated with standard doses of those drugs. Therefore, conducting genotyping or phenotyping for TPMT before giving the thiopurine drugs increasingly has been adopted in the clinical community (5). Genotyping of patients for drug-metabolizing enzyme polymorphisms has been commonly applied in clinical trials in many pharmaceutical companies (6).

In addition to drug-metabolizing enzymes, other genetic variations such as receptors and transporters could be attributed to the different drug responses. For example, the selective serotonin-reuptake inhibitor fluvoxamine is a commonly prescribed drug for treatment of delusional depression. The prime target of this drug is the serotonin transporter 5-HTT, which plays an important role in the termination of 5-HT neurotransmission. A 44-base pair insertion polymorphism in the 5-HTT promoter region has been associated with increased transcriptional activity of this gene. Individuals who are homozygous for the long variant of 5-HTT promoter respond better to fluvoxamine treatment than those who are either heterozygous or homozygous for the short variant.

Growing evidence links HIV-1 resistance mutations with drug failure.Costs of antiretroviral therapy for HIV-infected patients have increased because of drug resistance associated with virologic failure and a subsequent shift to more complex and costly therapies. In a prospective controlled random clinical trial, the virological and immunological impacts of genotypic-resistance testing on HIV viral loads were assessed (7). It was clearly demonstrated that when choosing a therapeutic alternative, genotypic-resistance testing has a significant benefit for the virological response.Also from the same study, genotypic-guided treatment proved to be cost-effective. The additional expense of genotyping appeared to be offset by the savings obtained in drug costs (8). As a result of this study and others, there is a rapid adoption in clinical practice to routinely conduct HIV genotyping for drug resistance before prescribing antiviral therapies. When used properly, pharmacogenomics can clearly deliver improved health benefits to patients while realizing cost savings. Table 1 lists examples of drugs and the pharmacogenomic markers that cause or are recognized by variable drug responses.

With the recent decoding of the human genome, the next goal of the industry is to rapidly uncover the disease genes and expedite drug discovery and development to create safer and more effective therapeutics.

SNP identification
In 1999, pharmaceutical companies and the Wellcome Trust created the SNP Consortium to develop a catalog of 300,000 SNPs for public use. In parallel, the International Human Genome Sequencing Consortium continues to contribute new SNPs to the public domain. These initiatives have generated 1.42 million SNPs distributed throughout the human genome, providing an average density on available sequence of one SNP every 1.9 kilobases (12). Private SNP databases provided by Celera Genomics (Rockville, MD), CuraGen (New Haven, CT), and Incyte Genomics (Palo Alto, CA) provide proprietary SNPs to clients on a paid subscription basis.

The large volume of expressed sequence tags (ESTs) and genomic DNA sequences in the public databases provides a rich source for in silico identification of SNPs. Through fragment clustering and multiple alignment of the sequences from redundant EST and bacterial artificial chromosome clones, potential SNPs can be identified without the initial effort to sequence multiple individuals. A computer SNP screening algorithm called POLBAYES is available at no cost for nonprofit use (http://genome.wustl.edu/). Another candidate SNP identification tool called SNPpipeline is free for public use (http://cgap.nci.nih.gov/GAI). By calculating the probability scores, putative SNPs could be identified. These putative SNPs can be confirmed by resequencing in multiple individuals, providing a fast and cost-efficient way to identify new SNPs.

Technologies for SNP analysis
Advanced technologies that rapidly, accurately, and economically validate and genotype SNPs across large populations are of significant value to pharmaceutical research and development. Recent developments have provided a foundation for large-scale SNP genotyping applications.

Mass spectrometry for SNP analysis. Currently, severalSNP analysis technologies use mass spectrometry (MS) through electrospray or matrix-assisted laser desorption ionization (MALDI) and ion-trap or time-of-flight detectors. This technique typically uses primer extension chemistry to generate allele-specific products that can delineate associated genotypes based on molecular weight. The technique requires purified samples free of ions and other impurities, thus increasing technical time and sample processing costs. A chip-based method using MS for SNP genotyping commercialized by Sequenom, Inc. (La Jolla, CA), has the potential for high-throughput, low-cost genotyping. Other companies using MS technologies for SNP genotyping are Orchid Biosciences (Princeton, NJ) and Qiagen Genomics (Bothell, WA).

High-throughput microarray technologies. The DNA microarray is a hybridization-based genotyping technique that offers simultaneous analysis of numerous SNPs. High-density microarrays are created by attaching hundreds of thousands of oligonucleotides to a solid surface in an ordered array. The DNA sample of interest is amplified using the polymerase chain reaction to incorporate fluorescent-labeled nucleotides and then hybridized to the array. The hybridization signals are quantitated by high-resolution fluorescent scanning and analyzed by computer software. DNA alterations such as SNPs, insertions, and deletions can be identified.

High-density microarray genotyping technologies are currently under development in several companies, including Affymetrix (Santa Clara, CA) and Incyte Genomics (Palo Alto, CA). Medium-density arrays complement many of the high-density products currently on the market. Genometrix, Inc. (The Woodlands, TX), has developed a low-cost, medium-density array format using a multiplex capillary printer and high-speed robotics. This technology can provide a cost-effective means of generating SNP or gene expression profiles for a statistically relevant number of samples.

Microsphere technology for SNP analysis. The LabMAP technology using microscopic beads called microspheres is well suited to address the needs of cost-effective SNP genotyping (Luminex Corp., Austin, TX). The LabMAP technology consists of fluorescently dyed microspheres that function as carriers of the molecule of interest. For SNP analysis, DNA probes are readily attached to the surface of the microsphere using a one-step coupling reaction. The microsphere with attached DNA probe is then analyzed by an instrument that uses lasers to illuminate the fluorescent dyes on the inside and surface of the molecule. Advanced optics capture and translate the illuminated signal into quantitative data. Direct hybridization, oligonucleotide ligation, and primer extension assay formats have been successfully used with the LabMAP technology.

Another bead-based technology platform, the BeadArray developed by Illumina (San Diego), combines fiber-optic bundles and specially coated beads that self-assemble into an array. Prepared samples bind to the molecules on the coated bead, allowing for simultaneous analysis of tens of thousands of quantitative measurements per sample.

Invader assay. The Invader assay is an attractive FRET-based genotyping method with the potential to genotype SNPs without PCR amplification (Third Wave Technologies, Inc.; Madison, WI). The assay detects specific DNA and RNA sequences by using structure-specific Cleavase enzymes to cleave a complex formed by the hybridization of overlapping oligonucleotide probes. Each cleaved product then serves as an Invader oligonucleotide in a secondary reaction, where it directs the cleavage of a combined labeled FRET probe/template construct. Upon cleavage, the fluorescein-labeled product is detected with a fluorescence plate reader, and genotypes are assigned after determining the wild-type/variant signal ratio for each sample.

Bioinformatics tools for SNP analysis. Genomic studies generate an unprecedented amount of disparate data that must be assimilated and analyzed to reap the promise of pharmacogenomics. Bioinformatics tools such as Discovery Manager (Genomica; Boulder, CO), HelixTree (Golden Helix; Bozeman, MT), and the VistaLogic Information System (Genometrix) each provide a unique partial solution to the demand for bioinformatics tools. Combined, they enable researchers to quickly organize massive quantities of data, mine it for interesting relationships or test directed hypotheses, and visualize the results.

Discovery Manager contains a variety of visualization tools for exploring genotypes and haplotypes through pedigree structures, as well as tools for visualizing gene maps and probe designs. This software excels in the organization and visualization of data structures; however, it lacks many numerical analysis features. HelixTree consists of a recursive partitioning algorithm that, for example, can discern groups of individuals defined by genotypic and phenotypic values that are most suitable to a particular drug. This software provides a very powerful exploratory analysis tool.

The VistaLogic Information System is a suiteof data management, visualization, and numerical tools that enable the analysis of both gene expression and genotyping data. While VistaLogic provides a wealth of visualization tools, its greatest strength is its ability to analyze and correlate gene expression, genotyping, and, ultimately, proteomic data, all within the same environment.

The future
Pharmacogenomics holds the potential to allow physicians and researchers to precisely prescribe or design the most efficacious and least toxic therapeutics based on an individual’s genotype. The success of pharmacogenomics depends on the adoption of advanced technologies for SNP identification and bioinformatic tools for data analysis.

With the improvement of these enabling technologies, pharmacogenomics may fundamentally change the practice of medicine by providing physicians with essential information to precisely prescribe the appropriate drug at the correct dose for each patient and will provide enormous health benefits and cost savings to the public at large.

References

1. Lazarou, J.; Pomeranz, B. H.; Corey, P. N. Incidence of adverse drug reactions in hospitalized patients: A meta-analysis of prospective studies. J. Amer. Med. Assoc. 1998, 279, 1200–1205.
2. Altshuler, D.; et al. The common PPARgamma Pro12Ala polymorphism is associated with decreased risk of type 2 diabetes. Nat. Genet. 2000, 26 (1), 76–80.
3. Martin, E. R.; et al. SNPing away at complex diseases: Analysis of single-nucleotide polymorphisms around APOE inAlzheimer’s disease. Am. J. Hum. Genet. 2000, 67 (2), 383–394.
4. Weber, W. W. Pharmacogenomics; Oxford University Press: New York, 1997.
5. Krynetski, E. Y.; Evans, W. E. Pharmacogenetics of cancer therapy:Getting personal. Am. J. Hum. Genet. 1998, 63, 11–16.
6. Kleyn, P. W.; Vesell, E. S. Genetic variation as a guide to drug development. Science 1998, 281, 1820–1821.
7. Durant, J.; et al. Drug-resistance genotyping in HIV-1 therapy: The VIRADAPT randomised controlled trial. Lancet 1999, 353, 2195–2199.
8. Chaix-Couturier, C.; et al. HIV-1 drug resistance genotyping. A review of clinical and economic issues. Pharmacoeconomics 2000, 18 (5), 425–433.
9. Drazen, J. M.; et al. Pharmacogenetic association between ALOX5 promoter genotype and the response to anti-asthma treatment. Nat. Genet. 1999, 22, 168–170.
10. Kuivenhoven, J. A.; et al. The role of a common variant of the cholesteryl ester transfer protein gene in the progression of coronary atherosclerosis. N. Engl. J. Med. 1998, 338, 86–93.
11. Slamon, D. J.; et al. Use of chemotherapy plus a monoclonal antibody against Her2 for metastatic breast cancer that overexpresses Her2. N. Engl. J. Med. 2001, 344, 783–792.
12. Sachidanandam, R.; et al. A map of human genome sequence variation containing 1.42 million single nucleotide polymorphisms. Nature 2001, 409, 928–933.

Michael M. Shi, M.D., Ph.D., is in the department of applied genomics, and Dorothy Mehrens and Karon Dacus, Ph.D., are in the marketing department of Genometrix, Inc. Send your comments or questions regarding this article to mdd@acs.org or the Editorial Office by fax at 202-776-8166 or by post at 1155 16th Street, NW; Washington, DC 20036.