Pharmacogenomics: the relevance of emerging genotyping technologies.
[FIGURE 1 OMITTED]
With cloning of DNA in the early 1980s, molecular genetics vastly expanded the opportunity to study variations in DNA and their effects on disease and drug reactions. From about 1992 to 2002, microsatellite markers were the "hot" technology used to study these disease- or drug-associated genes, particularly in linkage-analysis studies. In these time-consuming and costly studies, about 300 to 800 microsatellite markers were identified throughout a genomic region and studied in a single DNA sample. Between microsatellites and arrays was the five-year-long "re-sequence" era--expensive but useful for discovering single nucleotide polymorphisms (SNP).
More recently, high-density genotyping arrays have become available for the study of genetic variation. These genome-wide technologies use in situ hybridization for SNP detection and contain anywhere from 1,000 to 500,000 SNP markers, with about one marker per five kilobase, or 5 kb. These arrays, in conjunction with the abundance of genomic-sequence information now available, have lead to a new era in biology.
What will new genotyping technologies bring to pharmacogenomics, and how will this information eventually lead into the clinical arena to produce individualized drug therapy?
Where are we today
We are in an era of post-genome biology. The human genome has been sequenced, and genetic variation is being characterized. We have an understanding of what different parts of the genome vary; we know what certain variations mean, (i.e., silent versus functional SNPs), and it is estimated that there are more than 10 million SNPs where the less-common allele occurs in at least 1% of the population. (1) These SNPs can contribute directly to disease predisposition by modifying the function of a gene, or they can be used as a marker to find nearby disease-causing mutations through association or family-based studies.
There are three basic classes of SNPs:
* those that cause an alteration in gene function;
* those that cause more subtle gene function modifications and may predispose an individual to disease in conjunction with other genetic variants and environmental factors; and
* those that are silent with respect to gene function.
We are now able to measure genetic variation in DNA sequences expressed in phenotypes at the molecular, cellular, organ, and system levels. We can only learn about the consequences of these variants by studying these in parts.
There is a growing list of gene variants that contribute to aberrant drug responses, from drug-metabolizing enzymes to drug transporters to drug targets as well as variants that are disease-related. Before genome-wide analyses were available, most aberrant gene-variant correlations were monogenic, such as those found in Mendelian-inheritance patterns. Candidate gene approaches have been used for these tedious, costly studies. Single genes related to an adverse drug reaction were based on clinical pharmacological studies of proteins (receptors) and pathways known to be involved in a drug pharmaco-kinetic or -dynamic response. An example is thiopurine S-methyltransferase (TPMT) and its enhancement of the S-methylation of thiopurine drugs--an effect due to genetic variation that can be responsible for thiopurine toxicity. (2) It is clear, however, that monogenic relations are not the case for most diseases, such as diabetes, cancer, heart disease, and pharmacological deviants. With the completion of the Human Genome Project and the recent advances in genome technology, we now have the capabilities to analyze pharmacogenomic data in a timely and cost-efficient manner.
Types of studies being performed
Several different types of studies addressing genetic variation are being performed today on these emerging genome-wide technologies, depending on the marker density of a genotyping array. Broadly speaking, whole genome genotyping arrays containing 3,000 to 10,000 markers are sufficient for linkage-analysis studies; arrays containing 10,000 to 100,000 are sufficient in marker density to study loss of heterozygosity and comparative genomic hybridization. Association studies, however, require a large number of study subjects (cases and controls) and high-density arrays, such as those with 100,000 to 500,000 markers on a single array. (3)
Linkage-analysis studies with surrogate markers have been successfully used to track genomic regions as they co-segregate with a disease through a pedigree. Before high-density arrays were available, hundreds of extremely laborious, expensive microsatellite markers were individually analyzed to address Mendelian diseases and other monogenic phenotypes. Common practice for these studies now is to use genome-wide SNP arrays, which contain thousands of SNP markers in one assay; and, therefore, thousands of markers are tested on a single sample. SNP markers on genome arrays--used successfully in many linkage-analysis studies--are biallelic and evenly distributed throughout the genome. For example, in a matter of days, Puffenburger, et al, mapped sudden infant death with dysgenesis of the testes syndrome, or SIDDT, to a novel loss of function in the TSPYL gene using a 10K SNP genome array. (4) Another linkage-analysis study using a 10K SNP array showed that a Mendelian locus on chromosome 16 determines susceptibility to doxorubicin nephropathy in the mouse. (5) Both of these studies identified novel genes that can be used as possible drug targets for the related disease.
The great benefit of using these new technologies is that these studies are being produced in days instead of months and years. The saved time alone reduces the cost of the overall experiment. Additionally, several comparison studies have been performed between traditional microsatellite markers and SNP array markers, and have shown that the overall information content is increased by at least 20%, with a mean SNP call rate of 95% and a false-positive rate of 0.05% to 0.08% using the SNP arrays. (6)
Genome-wide molecular biological analyses, such as loss of heterozygosity (LOH) profiling and comparative genomic hybridization (CGH) have significantly enhanced our ability to detect chromosomal aberrations in cancer cells and assess their role in tumorigenesis; certainly larger genome-wide SNP arrays have been useful. In these genome-wide studies, whole regions of the chromosome that are amplified carry candidate oncogenes, and whole regions that are deleted carry potential tumor suppressor genes--giving a better understanding of the complex events that occur during tumorgenesis. LOH studies are useful in identifying new tumor-suppressor genes based on the model that one inherited allele is mutated and the other is lost somatically, which is often the case in cancer tissue.
For example, Garraway, et al, ran 100,000 genome-wide SNP arrays on NCI60 cell lines. They first looked at chromosome copy number changes and then added the loss of heterozygosity information to further analyze certain regions of amplification. From their analyses, they found a novel therapeutic pathway into the treatment of melanoma targeting a novel gene, microphthalmia-associated transcription factor, or MITF, which was identified in one of the amplified regions on the genome-wide SNP array. (7) This study and other like studies demonstrate that the use of genome-wide arrays to analyze LOH and CGH gives greater insight into the underlying genetic alterations in cancer cells with identification of complex events, including loss and reduplication of loci.
Most diseases are influenced by both genetic and environmental factors. To study these complex diseases, a large sample of well-placed markers must be taken throughout the genome in either a case-control association study or a transmission/disequilibrium study. Using high-density SNP arrays is ideal, as the approximately 500,000 markers on the array have been strategically placed and expected to fall within close proximity to true disease-causing variants. It is not expected that these 500,000 SNPs genotyped on the array are the actual disease-causing variant, but that they are in linkage disequilibrium with a disease-causing mutant. These technologies have proven fruitful in several association studies. For example, a recent association study on a Japanese population yielded several new candidate loci for Type 2 diabetes, (8) a complex disease that is the result of many environmental factors as well as several expected genetic variants.
There are still many challenges in this field, and we still do not fully understand the contribution of all genetic variants to individual drug responses or how to make this information useful at the clinical level. Most SNPs that are disease-causing are very rare events and, therefore, difficult to capture in a case-control study. SNPs that enhance an individual's risk for disease are more abundant, but the relationship between disease and variant is harder to define. Also, there are SNPs that fall in intergenic regions; thus, they would not result in an amino-acid change, but could affect alternate splicing, regulation, and transcript stability. The importance of these high-density genotyping technologies does not necessarily come from the functional relevance of the SNPs genotyped, but rather the proximity of the SNP to disease-causing genetic variants. It is clear, however, that these emerging technologies will re-shape our current way of administering drugs and prognosing disease. It will soon no longer be possible to do these things without accounting for an individual's genotype and taking into account specific gene variants.
With high-density genotyping technologies, the way that an individual reacts to a drug and personal disease prognosis are in the future for all scientific fields. Disease and drug pathways are becoming readily available to the clinician, along with aberrant drug responses due to genetic variation. Individual drug therapy is closer today than ever before. Drug effects to known variants, such as those in TPMT and thiopurine drug toxicity or BRAC1/BRAC2, increase patient risk to cancers (9,10) are well documented, so it is perplexing that these data are not yet fully integrated into the laboratory and clinical setting.
Whole genome analyses that look at 500,000 individual markers are possible, and the costs of these arrays are reasonable. So, what does a genotype strategy mean to the laboratory and clinician in the future? As the costs are driven down, genotype strategies will have an enormous affect on drug development. Novel genes are being identified in linkage-analysis studies, and cancer-therapy targets are being discovered on a frequent basis from larger SNP arrays looking at amplifications and deletions in patient genomes. This data will allow researchers to narrow drug targets and, therefore, maximize drug efficacy and minimize drug toxicity. In addition, genome-wide arrays allow the study of not a single gene that is affected by the drug but entire pathways of drug effects on gene products. These analyses call for genome strategies that are now readily available to most researchers
The next and more powerful step that genotyping strategies will help develop is personalized medicine and individualization of diagnosis and therapy. Genotype studies have a significant value to medicine, since a better understanding of a patient's genetic makeup can allow physicians to avoid drugs with deleterious effects. If a patient is found to have a particular variant in a drug gene target, dosage can be modified or medication can be individualized to avoid the aberrant drug effects that could possibly occur due to this gene variant.
Obvious benefits from pharmacogenomic information are still not apparent in a clinical setting; however, before this translation can occur, several factors need to be changed on the healthcare side, as outlined recently by Evans, et al. (13) Here, Evans points out that, first, healthcare professions will need to be re-educated in the way that they prescribe medication, such as dosing based in genotype, ethnicity, or age of the patient, rather than a mean dose for an entire population. Additional genotyping tests will need to be performed before the proper medication can be administered; this is not common practice today. Both the clinician and the patient will have to accept these extra tests that will be required to obtain one's genotype before a drug can be administered. In addition, clinical trials will also need to take genotyping into consideration, and perhaps this would save certain drugs from abandonment. For example, drug X has adverse affects in patients with genotype Y but is effective for patients who have genotype Z. The deleterious drug effects are only seen in patients with the genotype Y.
Pharmacogenetics holds promise as a form of individualized medicine, but it is not yet a routine component in patient diagnoses. (14) In this post-genome era, many informatics challenges require the marriage of bioinformatics and clinical informatics. We are one step closer to a more unified laboratory and clinical setting, merging pharmacogenomic information into the clinical world.
Tina Hernandez-Boussard, PhD, is a graduate of Yale University and the University Claude Bernard in Lyon, France, and is a senior computational biologist at the Stanford Microarray Database. Teri E. Klein, PhD, is a graduate of the University of California at San Diego/Santa Cruz/San Francisco, and is currently a senior research scientist at Stanford. Russ B. Altman, MD, PhD, is a graduate of Harvard College and Stanford University and a professor of genetics, bioengineering, and medicine at Stanford.
Acknowledgement: This work is supported by NIH UO1GM61374.
1. Carlson, CS, et al. Mapping complex disease loci in whole-genome association studies. Nature. 2004;429(6990):446-452
2. Weinshilboum, R. Inheritance and drug response. N Engl J Med. 2003;348(6):529-537.
3. Syvanen, AC. Toward genome-wide SNP genotyping. Nat Genet. 2005;37 Suppl: S5-10.
4. Puffenberger, EG, et al. Mapping of sudden infant death with dysgenesis of the testes syndrome (SIDDT) by a SNP genome scan and identification of TSPYL loss of function. Proc Natl Acad Sci USA. 2004;101(32):11689-11694.
5. Zheng, Z, et al. A Mendelian locus on chromosome 16 determines susceptibility to doxorubicin nephropathy in the mouse. Proc Natl Acad Sci USA. 2005;102(7):2502-2507.
6. Schaid DJ, et al. Comparison of microsatellites versus single-nucleotide polymorphisms in a genome linkage screen for prostate cancer-susceptibility Loci. Am J Hum Genet. 2004;Dec;75(6):948-965.
7. Garraway, LA, et al. Integrative genomic analyses identify MITF as a lineage survival oncogene amplified in malignant melanoma. Nature. 2005;436(7047):117-122.
8. Nelson MR, Marnellos G, Kammerer S, Hoyal CR, Shi MM, Cantor CR, Braun A. Large-scale validation of single nucleotide polymorphisms in gene regions. Genome Res. 2004;August;14(8):1664-1668.
9. Majdak, EJ, et al. Prognostic impact of BRCA1 pathogenic and BRCA1/BRCA2 unclassified variant mutations in patients with ovarian carcinoma. Cancer. 2005;104(5):1004-1012.
10. Robson, M, et al. Appropriateness of breast-conserving treatment of breast carcinoma in women with germline mutations in BRCA1 or BRCA2: a clinic-based series. Cancer. 2005;103(1):44-51.
11. Evans, WE, et al. Moving towards individualized medicine with pharmacogenomics. Nature. 2004;May27;429(6990):464-468.
12. The Royal Society. Personalised medicines: hopes and realities. 21 September 2005.
13. Thorn, CF, et al. PharmGKB: the pharmacogenetics and pharmacogenomics knowledge base. Methods Mol Biol. 2005;311:179-191.
14. Klein TE, et al. Integrating genotype and phenotype information: an overview of the PharmGKB project. Pharmacogenetics Research Network and Knowledge Base. Pharmacogenomics J. 2001;1(3):167-170.
By Tina Hernandez-Boussard, PhD; Teri E. Klein, PhD; and Russ B. Altman, MD, PhD
RELATED ARTICLE: PharmGKB catalyzes pharmacogenomics research and deployment
The Pharmacogenetics and Pharmacogenomics Knowledge Base (PharmGKB) is a public resource that promotes research into the relationships between human genotypes, phenotypes, and clinical outcomes by publishing, linking, and annotating primary datasets from ongoing research and established data from the literature. (13,14) PharmGKB has worked with others to develop a number of disease and drug-centered pathways to identify target genes and investigate functional polymorphisms. PharmGKB pathways are generated by collaboration of investigators to link data, either novel or in the public domain, centered on a particular drug [the thiopurine pathway is shown in Figure 2], and the representation is a consensus of the opinions of the authors. PharmGKB and its associated network have developed a public database to catalyze pharmacogenetics research and deployment, as outlined in the accompanying article.
[FIGURE 2 OMITTED]
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||medical research|
|Author:||Hernandez-Boussard, Tina; Klein, Teri E.; Altman, Russ B.|
|Publication:||Medical Laboratory Observer|
|Date:||Mar 1, 2006|
|Previous Article:||Specific IgE testing: objective laboratory evidence supports allergy diagnosis and treatment.|
|Next Article:||What's in that pot o'gold?|