Originally Published IVD Technology November/December 2003
MOLECULAR DIAGNOSTICS
A higher-throughput technique for clinical pharmacogenetic testingA pilot laboratory study of denaturing high-performance liquid chromatography demonstrates its potential suitability for routine genomics-based patient analysis.
Bonny L. Bukaveckas, Mark W. Linder, and Roland Valdes Jr.
 |
| Using the WAVE System 3500HT by Transgenomic Inc. (Omaha, NE), genetic variants are detected by differential retention of heteroduplex DNA on a small high-performance liquid chromatography column. |
The field of pharmacogenetics concerns the way different people respond differently to medicines owing to their unique genetic inheritance. The word has been pieced together from the roots pharmaco-, referring to medicinal drugs, and genetics, the study of heredity and the variation in traits.
The terms pharmacogenetics and pharmacogenomics are often used interchangeably. However, they differ conceptually. Pharmacogenetics is the study of variations between individuals in the DNA sequence related to drug response. Pharmacogenomics is the study of the variability in expression of individual genes relevant to disease susceptibility as well as to drug response at a cellular, tissue, individual, or population level. The latter term is broadly applicable to drug design, drug discovery, and clinical development. The practical difference for clinicians is that patient genetic testing falls within the realm of pharmacogenetics.
Pharmacogenetic testing links the underlying genotype with phenotypic differences in the pharmacokinetics or pharmacodynamics of medications.1–3 Eventually, pharmacogenetic testing will become routine, with demand for high-throughput testing increasing. With that end in view, laboratories are striving to reduce the cost and turnaround-time requirements of such testing.
The authors' own clinical pharmacogenetic testing laboratory has the mission of reducing or eliminating adverse drug reactions (ADRs) caused by pharmacogenetic variation. Achieving the goal of greatly reducing genetic ADRs will require offering pharmacogenetic testing at low cost and with rapid turnaround.
This article begins with a brief review of pharmacogenetics, focusing on prospective uses in individualized medicine. Then it describes and reports the results of a pilot study of the suitability of employing denaturing high-performance liquid chromatography in clinical pharmacogenetic testing, using a single variant test, CYP2C9*2, as an example.
Genomics and Modern Medicine
In the 50 years since publication of the structure of deoxyribonucleic acid (DNA), the science of genetics has come an amazing distance. The first descriptions of DNA caused wonderment that such a seemingly simple string of four molecules could hold all the information needed to make a marigold or a portabello mushroom, let alone a human being. Now, since early 2003, the complete sequence of the human genome is available.
That the total number of human genes is much smaller than once hypothesized (estimated to be as few as 34,000) suggests that human individuality may be due not to how many genes people have but to the variations in their DNA. These individual differences are called genetic variants. They are quite common, occurring at a frequency of one single nucleotide polymorphism (SNP) approximately every 1000 base pairs. How many of the variant single base-pair positions need to be looked for is still an open question.
Clinical Testing Potential. Some of these DNA variants are related to human physiological response to medications. Clinical pharmacogenetic testing translates basic genetics research into laboratory practice.4 The future of medicine is likely to include more-individualized treatment, and that new era of therapy will require knowledge of the individual patient's pharmacogenetic profile. Such knowledge will enable physicians to approach the ideal of administering the right medication in the right dosage the first time, with every patient.
Before the goal of individualized medicine based on pharmacogenetics can be reached, however, three major hurdles must be overcome.5
The first challenge is a lack of prospective clinical trials to investigate the utility of pharmacogenetic testing for guiding medication therapy.6 Some studies are under way, however, so knowledge should be forthcoming.
The second hurdle is the relatively high cost of genetic testing combined with slow turnaround.7 These analytical constraints reflect the fact that the field is in its infancy; clinical pharmacogenetic testing is performed at only a few sites.
These first two obstacles are related in that test costs and turnaround time are prohibitive factors for conducting clinical trials, while the paucity of trials has meant that adoption of the testing by clinicians has been slow so far. The result is a low demand for services and thus a lack of competition. Competition, when it comes, will bring down prices and reduce turnaround times, enabling pharmacogenetic testing to become routine.
The third challenge for clinical pharmacogenetics, derived from the first two, is to win acceptance from the medical community. Mere availability of test protocols will not translate into usage. This hurdle will become less of a barrier, of course, if testing is supported by solid clinical research and becomes rapid enough and inexpensive enough to be performed widely.
Inadequacy of Established Methodology.
| Gene | Phenotypes | |||
| Poor Metabolizer |
Intermediate Metabolizer |
Extensive Metabolizer |
Ultrarapid Metabolizer |
|
| CYP2C9 | 25% ± 13% | 58% ± 15% | 128% ± 10% | — |
| CYP2C19 | 58% ± 12% | 81% ± 6% | 105% ± 5% | — |
| CYP2D6 | 52% ± 21% | 83% ± 16% | 121% ± 10% | 263% ± 35% |
| Table I. Generalized dose adjustment for three polymorphic cytochrome P450 genes, based on pharmacogenetic genotype. Values indicate percentage of the standard dose recommended for the given phenotypes. | ||||
Other genes with variants also have pharmacogenetic implications, and this list is growing regularly. For some of these genetic variants, the information regarding clinical potential is compelling (see Table II).9–20 Most of the enzymes in both tables are involved in the metabolism of several therapeutic compounds.
| Gene | Affected
by Polymorphism |
| CYP3A49 | Pantoprazole |
| CYP2A69 | Nicotine |
| CYP2B69 | Cyclophosphamide |
| CYP2E19 | Halothane |
| CYP3A59 | Tacrolimus |
| UGT1A19 | Etoposide |
| ALDH110 | Acetaldehyde |
| ADH211 | Alcohol |
| DCP-1 (ACE)12 | Ramipril |
| ADRB213 | Albuterol |
| HTR2A14 | Clozapine |
| INPP115 | Lithium |
| SDF116 | HIV infection |
| TPMT17 | Azathioprine |
| DYPD18 | 5-Fluorouracil |
| NAT219 | Amonafide |
| MDR120 | Digoxin |
| Table II. Other genes reported by more than one author to have pharmacogenetic significance (superscripts refer to article citations). The representative compound or condition in the right column exhibits in metabolism or disease course a consequence of genetic variation. | |
However, the disadvantages of using restriction fragment length polymorphism (RFLP) are significant. They include low throughput and the limitation that not all polymorphisms are amenable to an RFLP protocol. These disadvantages have been addressed by some groups through the use of multiplexed PCR and by achieving faster resolution of fragments through such technologies as capillary electrophoresis. But when high throughput becomes a standard of practice, RFLP falls far short.
The Louisville medical school facility has faced the decision of choosing its next methodology as demand for services increases. There are many choices, as the list of available technologies expands rapidly.
The laboratory retains a research interest in gene polymorphism discovery as well. Therefore, it was looking for a technology that allowed rapid screening of gene regions for polymorphisms while at the same time having utility for basic research. The facility's clinical testing section also needed to be upgraded to a higher-throughput system. Consequently, the full search encompassed systems that could satisfy both needs. However, so far, no one technology has surfaced as a leading candidate for clinical pharmacogenetic testing.
This article presents comparison data generated from a study of RFLP and one of the new technologies, partially denaturing high-performance liquid chromatography (dHPLC), which shows potential.
Other Possibilities. Numerous other genotyping technologies besides dHPLC and RFLP are on the market or are very close to market release. FDA has published draft guidelines for the submission of DNA technologies for approval.21 These draft guidelines present information relevant for direct comparison of technologies, including their clinical utility.
A short list of emerging technologies for genotyping clinical samples must include luminescent technology and electrochemical detection. Other assays that show potential involve flow cytometry with fluorescence-labeled beads. Also being developed are DNA microarrays to scan for thousands of variations at once, though no sign of this technology being used in clinical pharmacogenetic testing is yet apparent.
Study Background
The present discussion addresses clinical genotyping of one SNP in the cytochrome P450 2C9 gene (C416T, Arg144Cys, CYP2C9*2), which is used as an example. This gene is one of a class of medication-metabolizing enzymes that has been reviewed extensively.22
A large number of therapeutic compounds are metabolized by CYP2C9, including S-warfarin, antecoumaral, tolbutamide, glyburide, glipizide, phenytoin, and celecoxib. CYP2C9 variant alleles exhibit virtually no enzymatic activity.
In vitro studies have shown that the protein product of the CYP2C9*3 variant is less than 5% as efficient as the CYP2C9*1 allozyme. (An allozyme is an allele that produces an expressed protein that is functionally different from the wild-type enzyme.) The CYP2C9*2 allozyme displays about 78% less activity in most assays. CYP2C9*2 and CYP2C9*3 result in poor metabolizer (PM) phenotypes in the homozygous state and intermediate metabolizer (IM) phenotypes when heterozygous. Proposed genotype-specific dose adjustments have been published for warfarin, glipizide, tolbutamide, and phenytoin.6,8
The anticoagulant activity of Coumadin (warfarin) is related to its concentration in the blood—too high and the patient may bleed, too low and inappropriate coagulation may occur. The dose ordinarily is adjusted in the first days of treatment according to the international normalized ratio (INR), an indirect measure of anticoagulant activity. The blood concentration of warfarin at any given time reflects a balance between ingestion (the dose taken) and metabolism by the liver and elimination, with the optimal serum concentration in most people being 0.68 µg/ml.23
Warfarin is metabolized in the liver by the enzyme CYP2C9. In 35% of individuals this enzyme is genetically variant. These variants with reduced enzymatic efficiency allow an accumulation of warfarin in these people if they are on a standard therapeutic dose.
Pharmacogenetic testing for CYP2C9 variation can predict which people will be sensitive to warfarin because of genetic variation. Test results provide clinical support for rational dose adjustment by physicians caring for people already on warfarin and who are experiencing clinical signs and symptoms of warfarin sensitivity. Administering a standard 5-mg dose of warfarin to individuals with CYP2C9 variants can lead to excessive anticoagulating responses and bleeding episodes if the dosages are not lowered to accommodate these patients' reduced-metabolizer phenotypes.
In a study of 185 patients, those with CYP2C9 variants on average required 95 more days to achieve a stable INR. Researchers also found that possession of a variant allele put a patient at increased risk of serious or life-threatening bleeding complications (HR, 2.39; 95% CI, 1.18–4.86). Genotype-adjusted dosing has the potential to prevent the over-anticoagulation resulting in bleeding events that is often seen in CYP2C9 IM and PM individuals.23,24
The Method-Comparison Study
The Louisville medical school laboratory study involved the following method protocols.
The RFLP Technique. The testing of clinical samples for CYP2C9*2 using RFLP has been done previously.25 The technique is based on the fact that changes in the nucleotide sequence of a gene may alter the ability of restriction enzymes to cut the DNA at or near the site of the polymorphism. After amplification and digestion, these different patterns of DNA fragments can be visualized, and the genomic sequence can be inferred from the pattern. Testing for the CYP2C9*2 allele involves isolating the genomic DNA from the buffy-coat of EDTA-anticoagulated whole blood samples.
For PCR amplification, the Louisville laboratory follows published protocols, using forward primer 5'-tacaaatacaatgaaaatatcatg-3', and reverse primer 5'-ctaacaaccagactcataatg-3'. The laboratory uses standard PCR conditions with PCR MasterMix 2X from Promega Inc. (Madison, WI), which containes Taq DNA polymerase, 400 µmol of dNTP, 3 mmol of MgCl2, and buffers, and adds approximately 1 µg of DNA and primers to 1-µmol final concentration of each. Amplification takes place over 45 cycles of 1 minute each at temperatures of 94°, 50°, and 72°C. These conditions produce a 691-base-pair amplicon from exon 26 of the CYP2C9 gene, genebank RefSeq D00173 and M61857. A no-DNA control, as well as sequencing verified CYP2C9*2-positive and -negative DNA samples, is included in each PCR event as quality control material.
) |
| Figure 1. Photograph of 3% agarose gel after electrophoresis of CYP2C9*2 RFLP digestion products. (click to enlarge) |
PCR products are digested with Ava II at 37°C for 2 hours using standard procedures. Digestion products are visualized via 3% agarose electrophoresis with ethidium bromide staining, and photodocumented for interpretation by a medical technologist (see Figure 1). Wild-type DNA will be cut by the restriction enzyme Ava II; the CYP2C9*2 allele will remain uncut.
Lanes 1 and 2 in the figure contain CYP2C9*1*1 control DNA. In lane 1 is undigested DNA. Lanes 3 and 4 contain undigested and digested CYP2C9*1*2 DNA, respectively. Lane 5 is a 100-base-pair DNA ladder, the brighter band being 500 base pairs. Lanes 6 and 7 contain CYP2C9*2*2 DNA, respectively undigested and digested. Lane 12 is a no-DNA control used to detect DNA carryover or contamination in the PCR and restriction-enzyme digest reactions. Lanes 8 to 11 contain two unknown DNA pairs; undigested and digested samples are shown. Unknown number 1 is shown to be CYP2C9*1*2, and unknown number 2 is CYP2C9*1*1.
Final results are reviewed by a Diplomate of the American Board of Clinical Chemistry– certified clinical chemist and then are released to the ordering physician electronically.
The dHPLC Alternative. The instrument evaluated by the University of Louisville laboratory for its suitability as an upgrade to its pharmacogenetic testing procedures was a Wave 3500HT system from Transgenomic Inc. (Omaha, NE), which uses dHPLC. The concept of using this form of chromatography for detection of DNA variants was developed at Stanford University in the early 1990s; the Stanford patent is licensed to Transgenomic.
The dHPLC method partially denatures the amplified DNA fragment by means of a heated column. The degree of helicity and the percentage of acetonitrile determine the retention time on a small column for each section of DNA.
The DNA is generally detected through use of either the UV absorbance or fluorescence of labeled products. For its method evaluation, the laboratory used UV detection alone. Fluorescence would increase sensitivity, but it would also increase reagent costs because dye-labeled primers cost more. Software for predicting optimal separation conditions is available with the Wave 3500HT.
DNA variants are detected through post-PCR formation of heteroduplexes with samples of known genotype. This process involves mixing the sample, heat-denaturing the DNA, then allowing a slow reannealing to occur, which enables the heteroduplexes to form. The heteroduplexes resolve as separate peaks on the chromatogram.
Resolution of genetic variants is reported by the manufacturer to be 1% from 20 to 2000 base pairs, meaning that a single-base-pair change can be resolved completely in a 100-base-pair amplicon and a two-base-pair difference in a 200-base-pair amplicon. The manufacturer claims that its system is nearly 100% sensitive within the analytical range of the instrument.
Denaturing HPLC has been employed to measure variation in medication-metabolizing enzymes in the CYP2A6 and CYP2B6 genes.26,27 (The dHPLC technique is widely cited for genetic analysis of other genes; an April 2003 PubMed search using the term dHPLC resulted in 210 hits. The Louisville laboratory was primarily interested in the technique's potentially fast analysis times, which could reduce labor expenses and shorten turnaround time for clinical pharmacogenetic testing. Its agarose gels typically run for more than 40 minutes, whereas dHPLC rapid DNA analysis involves runs of typically 2–5 minutes.
At this time, two columns are available for DNA analysis using dHPLC. The original column is packed with alkylated polystyrene divinylbenzene particles.28 The material is commercially available from Transgenomic as the DNASep Column, part number 621 5046. The Helix analytical column from Varian Inc. (Palo Alto, CA), part number 0392613001, is a column for DNA analysis and dHPLC packed with 1000-Å alkylated silica. It comes configured with a 3.0-mm inner diameter and a 75-mm length, and offers the advantages of low buffer consumption and strong signal intensity. This column is generally available; it is not restricted to users of the Helix chromatography system.
For its study, the laboratory developed novel primers for a 103-base-pair region centered on the CYP2C9*2 site, using forward primer 5'-atccggcgtttctccctcat-3', and reverse primer 5'-aactcctccacaaggcagcg-3'. The same Promega master mix was used as with the RFLP technique, for direct comparison. Transgenomic, however, does recommend that a proofreading enzyme be used. The CYP2C9*2 protocol optimization with positive and negative quality control samples took two days.
Next, using the optimized protocol, 79 DNA samples that had been previously genotyped by means of RFLP were chosen for the evaluation. The amplification took nearly the same amount of time for the two techniques: 2.5 hours for dHPLC versus 2 hours for RFLP. The former was longer owing to the inclusion of a touch-down PCR protocol and the heteroduplex formation.
) |
| Figure 2. A composite CYP2C9*2 chromatogram generated by monitoring UV absorption of positive control DNA fragments eluting from the dHPLC column. (click to enlarge) |
Allele-specific peaks were generated through the formation of heteroduplex DNA, which resulted in reduced DNA helicity and altered retention times (see Figure 2). PCR products were injected directly onto the column, with no post-PCR cleanup required. Retention time was related to DNA fragment size and percent helicity.
In Figure 2, the chromatogram resulting from the CYP2C9*1*1 DNA (blue line) is composed mainly of the longest-retention-time component, while the CYP2C9*2*2 DNA (red line) is primarily composed of the short-retention-time moiety. The CYP2C9*1*2 DNA (black line) is a roughly equal mix of short and long retention times.
As seen in the figure, shoulder chromatographic peaks also may form as a result of multiple melt domains within the same DNA fragment. The patterns are highly reproducible, however, and are clearly distinguishable from alternative genotypes. The shape of the peaks is not predictable a priori, but once elution conditions are optimized, the shape of the chromatogram is highly specific for a given allele. Therefore, unknown samples can be genotyped on the basis of comparison with positive control samples.
In its preliminary evaluative study, the Louisville medical school laboratory used dHPLC in a blinded manner to determine the genotype of 79 DNA samples and 4 quality controls (no-DNA, CYP2C9*2*1, CYP2C9*1*1, and CYP2C9*2*2). Analysis and interpretation of the total of 83 samples took 4.5 hours.
A peak height threshold value of 2 mV for genotype calling was established. Owing to the fact that a 1000-mV peak area is about equal to 1 ng of DNA, the minimum amount of amplified PCR product for this test was approximately 0.6 pg of DNA. The ethidium bromide staining detection limit is approximately 1500 pg of DNA. The more-sensitive DNA dyes such as SYBR Green (Molecular Probes Inc.; Eugene, OR) have reported sensitivities of 60 pg of DNA. Therefore, dHPLC is 100 times more sensitive than the gel-based analyses. This should allow an adequate signal to be generated from fewer amplification cycles, thereby reducing turnaround time.
Three of the 79 samples did not achieve the 2-mV threshold, indicating low or no PCR efficacy. These samples were reamplified and reanalyzed using the same protocol, which required another round of PCR and an additional 15 minutes of dHPLC time. Ultimately, the genotypes of all of the samples agreed with those determined by RFLP, including those that required a second amplification. There were no false negatives or false positives.
PCR failure is a common analytical problem with all techniques that require amplification prior to interrogation. If the pilot study reported here can be taken as predictive of routine performance, a repeat amplification rate of 3.7% could be expected. This amplification failure rate is consistent with that of the RFLP method the Louisville laboratory currently uses. With both methods, repetition of amplification adds to the overall time and cost of analysis, and a calculation of it should be included in the final budget. During a future evaluation of the full clinical testing method, these variables can be determined more precisely.
RFLP versus dHPLC. The RFLP and dHPLC methods demonstrated 100% concordance in the pharmacogenetic test application. The dHPLC testing protocols are straightforward to design. Because of the short analysis times and the predictive software, the protocol was optimized for the study in two days. As with RFLP, interpretation of results is straightforward. Unlike RFLP, analysis can be semiautomated by means of mutation-detection software. The dHPLC technique allows for simultaneous clinical testing and discovery research, an advantage for laboratories with a hybrid basic research and clinical service function. This option is not available with RFLP.
Several practical limitations on the use of dHPLC for clinical testing are well known. The initial capital investment is high; the high-throughput instrument costs almost $100,000. This cost might come down if more manufacturers competed in the dHPLC market. Also, dHPLC is not now a method generally cited in the U.S. clinical literature for mutation detection in pharmacogenetics. Therefore, there are few published protocols. Nor are any commercial assay kits or positive control materials using the technology available. These are materials that clinical testing laboratories with less-comprehensive research and development capabilities than academic facilities have would find necessary.
Conclusion
The first protocols used in the clinical pharmacogenetics testing laboratory of the University of Louisville Medical School employed RFLP. In order to decrease analysis turnaround time and cost, the laboratory searched for a high-throughput system with short analysis times and low reagent costs. An investigation of the applicability of a dHPLC system revealed that this system offered several advantages over RFLP.
The laboratory had experience with CYP2C9*2, the SNP described in this article, in connection with several assays: RFLP, PacMan assays, capillary electrophoresis–based sequencing, real-time PCR with the LightCycler (Roche Diagnostics Corp.; Indianapolis), and the luminescence-based Promega ReadIt.
The preliminary investigation of the use of dHPLC for clinical pharmacogenetic testing found it to compare favorably with the alternative techniques. The main drawback is the initial capital investment required. However, for laboratories already using RFLP, the anticipated increase in demand for pharmacogenetic testing should make possible a cost-neutral facility changeover from RFLP to dHPLC.
References
1. CP Cartwright, "Pharmacogenetics: The Dx Perspective," Expert Review of Molecular Diagnostics 1, no. 4 (2001): 371–376.
2. G Schmitz, C Aslanidis, and KJ Lackner, "Pharmacogenetics: Implications for Laboratory Medicine," Clinica Chimica Acta 308, no. 1-2 (2001): 43–53.
3. P Hess and D Cooper, "Impact of Pharmacogenetics on the Clinical Laboratory," Molecular Diagnostics 4, no. 4 (1999): 289–298.
4.AD Roses, "Pharmacogenetics and Future Drug Development and Delivery," Lancet 355 (2000): 1358–1361.
5. CI Barash, "Role of the Laboratory in Leveraging Adoption of Pharmacogenetics," American Clinical Laboratory 20, no. 8 (2001): 35–37.
6. J Brockmöller et al., "Pharmacogenetic Diagnostics of Cytochrome P450 Polymorphisms in Clinical Drug Development and in Drug Treatment," Pharmacogenomics 1, no. 2 (2000): 125–151.
7. S Wong, "Clinical Pharmacogenomics for Forensic Toxicology, Drug Addiction, and Pain Management," AACC Expert Access Live On-Line: October 2, 2001 [accessed 20 October 2003]; available from Internet: www.aacc.org/access/forensic/default.stm.
8. C Meisel et al., "How to Manage Individualized Medication Therapy: Application of Pharmacogenetic Knowledge of Medication Metabolism and Transport," Clinical Chemistry and Laboratory Medicine 38 (2000): 869–876.
9. Y Watanabe et al., "Glucuronidation of Etoposide in Human Liver Microsomes Is Specifically Catalyzed by UDP-Glucuronosyltransferase 1A1," Drug Metabolism and Disposition: The Biological Fate of Chemicals 31, no. 5 (2003): 589–595.
10. A Yoshida et al., "Cytosolic Aldehyde Dehydrogenase (ALDH1) Variants Found in Alcohol Flushers," Annals of Human Genetics 53, no. 1 (1989): 1–7.
11. DP Agarwal and HW Goedde, "Pharmacogenetics of Alcohol Metabolism and Alcoholism," Pharmacogenetics 2, no. 2 (1992): 48–62.
12. MJ Rieder et al., "Sequence Variation in the Human Angiotensin Converting Enzyme," Nature Genetics 22, no. 1 (1999): 59–62.
13. MD Littlejohn et al., "Determination of b2-Adrenergic Receptor (ADRB2) Haplotypes by a Multiplexed Polymerase Chain Reaction Assay," Human Mutation 20, no. 6 (2002): 479.
14. J Veenstra-VanderWeele, GM Anderson, and EH Cook Jr., "Pharmacogenetics and the Serotonin System: Initial Studies and Future Directions," European Journal of Pharmacology 410, no. 2-3 (2000): 165–181.
15. R Lovlie et al., "Genomic Structure and Sequence Analysis of a Human Inositol Polyphosphate 1-Phosphatase Gene (INPP1)," Pharmacogenetics 9, no. 4 (1999): 517–528.
16. A Brambilla et al., "Shorter Survival of SDF1-3'A/3'A Homozygotes Linked to CD4+ T Cell Decrease in Advanced Human Immunodeficiency Virus Type 1 Infection," Journal of Infectious Disease 182, no. 1 (2000): 311–315.
17. SA Coulthard and AG Hall, "Recent Advances in the Pharmacogenomics of Thiopurine Methyltransferase," The Pharmacogenomics Journal 1, no. 4 (2001): 254–261.
18. LK Mattison, R Soong, and RB Diasio, "Implications of Dihydropyrimidine Dehydrogenase on 5-Fluorouracil Pharmacogenetics and Pharmacogenomics," Pharmacogenomics 3, no. 4 (2002): 485–492.
19. F Innocenti, L Iyer, and MJ Ratain, "Pharmacogenetics of Anticancer Agents: Lessons from Amonafide and Irinotecan," Medication Metabolism and Disposition: The Biological Fate of Chemicals 29, no. 4, part 2 (2001): 596–600.
20. I Cascorbi et al., "Frequency of Single Nucleotide Polymorphisms in the P-Glycoprotein Drug Transporter MDR1 Gene in White Subjects," Clinical Pharmacology and Therapeutics 69, no. 3 (2001): 169–174.
21. "Medical Devices: Draft Guidance for Industry and FDA Reviewers; Multiplex Tests for Heritable DNA Markers, Mutations, and Expression Patterns; Availability," Federal Register, 68 FR:19549–19550, April 21, 2003.
22. TH Rushmore and AN Kong, "Pharmacogenomics, Regulation and Signaling Pathways of Phase I and II Medication Metabolizing Enzymes," Current Medication Metabolism 3, no. 5 (2002): 481–490.
23. MW Linder et al., "Warfarin Dose Adjustments Based on CYP2C9 Genetic Polymorphisms," Journal of Thrombosis and Thrombolysis 14, no. 3 (2002): 227–232.
24. MK Higashi et al., "Association between CYP2C9 Genetic Variants and Anticoagulation-Related Outcomes during Warfarin Therapy," Journal of the American Medical Association 287, no. 13 (2002): 1690–1698.
25. MJ Stubbins et al., "Genetic Analysis of the Human Cytochrome P450 CYP2C9 Locus," Pharmacogenetics 6, no. 5 (1996): 429–439.
26. M Pitarque et al., "Identification of a Single Nucleotide Polymorphism in the TATA Box of the CYP2A6 Gene: Impairment of Its Promoter Activity," Biochemical and Biophysical Research Communication 284, no. 2 (2001): 455–460.
27. UM Zanger et al., "Detection of Single Nucleotide Polymorphisms in CYP2B6 Gene," Methods in Enzymology 357 (2002): 45–53.
28. G Bonn, C Huber, and P Oefner, Nucleic acid separation on alkylated nonporous polymer beads, U.S. Pat. 5,585,236, 1996.
PHOTO COURTESY UNIVERSITY OF LOUISVILLE
Bonny L. Bukaveckas, PhD, is a clinical chemistry fellow at the University of Louisville Pharmacogenetics Diagnostic Laboratory (Louisville, KY); Mark W. Linder, PhD, is an assistant professor in the University of Louisville Department of Pathology and Laboratory Medicine and associate director of the Pharmacogenetics Diagnostic Laboratory; and Roland Valdes Jr., PhD, is a professor in the Department of Pathology and Laboratory Medicine as well as a professor of biochemistry and molecular biology at the University of Louisville. The authors can be reached through mwlind01@gwise.louisville.edu.
Copyright ©2003 IVD Technology
