Addressing the challenges of the clinical application of pharmacogenetic testing

Addressing the Challenges
in particular, is referred to as the *1 or wild-type allele. The (TA)7 and (TA)8 of the Clinical Application
repeats are associated with low UGT enzymatic activity.3 For the purposes of of Pharmacogenetic Testing
this article, we focus on the variant (TA)7, or *28, allele because it has received the most attention.
ON Ikediobi,1 J Shin,1 RL Nussbaum2 and KA Phillips1 and the
The *28 allele is associated with inher- UCSF Center for Translational and Policy Research on Personalized
ited forms of unconjugated hyperbiliru- binemia such as Gilbert syndrome and Medicine (group authors: JM Walsh,2 U Ladabaum2 and
Crigler–Naj ar syndrome.4 With regard to D Marshall3)
drug therapy, retrospective and prospec- tive studies have demonstrated an asso- Pharmacogenomics aims to use molecular genetic markers to predict ciation between the presence of reduced
UGT1A1 enzymatic activity (especially treatment outcome. Indeed, within the past decade there has been
with the *28 allele) and increased inci- a rapid emergence of pharmacogenetic tests to aid clinicians in
dence of toxicity, most notably severe predicting efficacy or toxicity for some drugs. Despite this major
neutropenia in patients treated with advance in therapeutic drug management, there remain challenges
irinotecan.5,6 The association between the *28 al ele and severe diarrhea is not as sta- to the appropriate use of pharmacogenetic tests. We discuss UGT1A1
tistically significant as is the association pharmacogenetic testing to illustrate the knowledge gaps impeding
with neutropenia. Therefore, there may be widespread use of pharmacogenetic tests in the clinical setting.
other, yet unknown genetic factors that are more predictive of irinotecan-induced diarrhea. In 2005, the US Food and Drug Pharmacogenetic tests are potential y use- use the UGT1A1 pharmacogenetic test as Administration (FDA) amended the ful tools for making a therapeutic deci- an example to il ustrate some of the major package insert for irinotecan to include a sion by identifying patients who should challenges to the clinical application of recommendation, but not a requirement, or should not receive a particular drug, as pharmacogenetic tests: incomplete knowl- to test for the *28 UGT1A1 variant to pre- well as for guiding individual drug dos- edge of the extent of human genetic varia- dict those at risk for neutropenia.7 Shortly ing. Pharmacogenetic test information is tion, availability of alternative biomarkers, thereafter, the Invader Molecular assay for currently included in more than 200 drug and the lack of a model of delivery for UGT1A1 genotyping appeared on the labels in the United States.1 The informa- market as an FDA-approved test, with an tion is classified into three categories that overall correct-call rate of 98.8%.8 guide the clinical use of pharmacogenetic UGT1A1 pharmacogenetic testing
A recent meta-analysis reported that tests for reaching a therapeutic decision2: UGT1A1 is a hepatic enzyme involved the risk of severe neutropenia for patients (i) test required, (ii) test recommended, in the glucuronidation of bilirubin and homozygous for the *28 allele may be a and (iii) for information only . many drugs, such as the active metabo- function of the administered dose of Thus far, only four drugs—cetuximab, lite of the anticancer drug irinotecan. Its irinotecan.9 The study found that the trastuzumab, maraviroc, and dasatinib— enzymatic activity differs among indi- risk of toxicity was higher for patients require a pharmacogenetic test before they viduals and can be indirectly affected homozygous for *28 who were on chemo- are prescribed; the majority of drugs with by environmental, physiological, and therapeutic regimens of irinotecan (alone labels containing pharmacogenetic test epigenetic changes. However, UGT1A1 or in combination with other myelo- information do not require such testing. enzyme function is genetically deter- toxic drugs) involving a moderate dose For example, the label for the anticancer mined by inherited sequence variation in (150–250 mg/m2, odds ratio (OR) = 3.22, drug irinotecan recommends testing for the coding and noncoding regions, most 95% confidence interval (CI) = 1.52–6.81; the presence of a variant of UDP-glucuron- notably a two-base-pair insertion (TA)n P = 0.008) or a high dose of irinotecan osyl transferase 1A1 (UGT1A1) to prevent in its promoter region.3 The most com- (>250 mg/m2, OR = 27.8, 95% CI =4.0– drug toxicity. However, pharmacogenetic mon promoter variants are (TA)5, (TA)6, 195; P = 0.005) as compared with patients testing for UGT1A1, like most pharmaco- (TA)7, and (TA)8 (ref. 3). The (TA)5 and heterozygous for *28 and homozygous for genetic tests, has many challenges for its (TA)6 variants are associated with high *1 (wild type).9 For this reason, the study's appropriate clinical use. In this article, we UGT enzymatic activity; the (TA)6 allele, authors have asserted, contrary to the 1Department of Clinical Pharmacy, School of Pharmacy, University of California, San Francisco, California, USA; 2Department of Medicine, School of Medicine, University of California, San Francisco, California, USA; 3Department of Community Health Sciences, Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada. Correspondence: KA Phillips ( VOLUME 86 NUMBER 1 JULY 2009 Table 1 Selected pharmacogenetic tests available for clinical application
FDA classification
Test example
EGFR expressiona Colorectal cancer Dako EGFR pharmDx Presence of Philadelphia Acute lymphoblastic CYP2C9 variants; VKORC1 Verigene Warfarin Metabolism Nucleic Acid Test TPMT variants Azathioprine, 6-MP, Acute lymphocytic Prometheus Therapeutics & Diagnostics and othersc Colorectal cancer Invader UGT1A1 Molecular Assay Glucose-6-phosphate dehydrogenase screening For information only c-KIT expressiona Gastrointestinal stromal Dako Cytomation c-Kit CYP2C19 variants Roche AmpliChip CYP450 CYP2D6 variants Atomoxetine, tamoxifen, Attention-deficit Roche AmpliChip CYP450 hyperactivity disease DPD deficiencya Capecitabine, 5-FU Colorectal cancer Genelex and othersc EGFR expressiona Non-small-cell lung cancer Dako EGFR pharmDx G6PD deficiencya Glucose-6-phosphate dehydrogenase screening NAT variants Isoniazid, rifampin Genelex and othersc Absence of Philadelphia Chronic myelogenous PML/RAR gene expressiona Acute promyelocystic PML/RARα quantitative Chronic myelogenous Invader UGT1A1 Molecular 5-FU, 5-fluorouracil; 6-MP, 6-mercaptopurine; ABL, Abelson; BCR, breakpoint cluster region; CCR-5, chemokine C-C motif receptor; c-KIT, v-kit Hardy-Zuckerman 4 feline sarcoma viral oncogene homolog; CYP2C9, cytochrome P450 2C9; CYP2C19, cytochrome P450 2C19; CYP2D6, cytochrome P450 2D6; DPD, dihydropyrimidine dehydrogenase; EGFR, epidermal growth factor receptor; FDA, US Food and Drug Administration; G6PD, glucose 6-phosphate dehydrogenase; HER2/NEU, v-erb-b2 erythroblastic leukemia viral oncogene homolog 2; HLA, human leukocyte antigen; NAT, N-acetyltransferase; PCR, polymerase chain reaction; PML/RAR, retinoic acid receptor; TPMT, thiopurine S-methyltransferase; UGT1A1, uridine diphosphate glucuronosyltransferase 1A1; VKORC1, vitamin K epoxide reductase complex subunit 1.
aThese tests determine whether a particular gene is expressed. The other tests determine the sequence variability of a particular gene. bFor patients with an Asian ancestry. cA home-brew test offered by a commercial clinical laboratory.
FDA's initial recommendation of a dose of irinotecan. However, recent reports UGT1A1 test into their clinical practice. reduction based on genotype alone, that show that a decrease in irinotecan dose It is unknown, however, how knowledge a reduced irinotecan dose or an alterna- for patients homozygous for the *28 al ele of the genotype influenced their care of tive treatment regimen be used only when may result in decreased tumor response, patients.8 On the other hand, there are the *28 allele is present and a high-dose further complicating determination of others who did not heed the FDA's rec- irinotecan regimen is prescribed.9 This appropriate drug dose to prevent irinote- ommendation for UGT1A1 genotyping, recent finding may help refine the FDA can toxicity.8 in part because of the uncertainty of its labeling of irinotecan as it pertains to the In light of the FDA's initial recom- clinical utility. What, then, are the chal- interpretation of the UGT1A1 genotype mendation, oncologists at several aca- lenges to the use of pharmacogenetic test- within the context of appropriate dosing demic health centers incorporated the ing for UGT1A1 in the clinic? CLInICAL PhARMACoLogy & TheRAPeuTICs VOLUME 86 NUMBER 1 JULY 2009
The first challenge is that the UGT1A1 Third, there is no model for the deliv- tion, many populations, such as African test may be limited in its general appli- ery and interpretation of the UGT1A1 Americans and Hispanics, are currently cability in diverse populations. There test. In the modified label for irinote- underrepresented in pharmacogenomic are at least 113 variants of the UGT1A1 can, the FDA recommends a reduced studies, and potentially important vari- gene, most of which are associated with initial dose of irinotecan for patients ants in those groups are not well known. reduced or inactive enzyme activity; a homozygous for the *28 allele. The FDA Although diversity in a study population few are associated with increased activ- recommendations go on to state, how- may complicate analysis, multiple phar- ity, and still others are of unknown sig- ever, that the precise dose reduction in macogenomic studies in different ethnic nificance.4 However, the most widely patients homozygous for the *28 allele groups will support the clinical use of used UGT1A1 test (the Invader Molec- is unknown and that subsequent dose the test in broader populations.
ular Assay) assesses only one promoter modifications should be tailored accord- Second, it is essential to evaluate the variant, the *28 allele. Homozygosity for ing to an individual patient's treatment clinical utility of pharmacogenetic the *28 allele occurs in 10% of the North tolerance.7 Unlike other genetic tests testing, including the risks, costs, and American population and at a similar for disease prognosis, the reports for benefits relative to use of the available frequency in Caucasians and African pharmacogenetic tests are currently not alternative biomarkers. Lack of clini- Americans but a lower frequency in interpreted for the ordering physicians. cal utility data is one of the reasons the Asians.4 The frequency of the *28 allele There is neither a detailed explanation of Evaluation of Genomic Applications in has not been thoroughly studied in other how and to what extent the initial dose Practice and Prevention Working Group ethnic groups, such as Hispanics living should be reduced, nor instructions on cannot recommend for or against the in North America. Even if one were to how long the patient should remain on clinical use of some pharmacogenetic comprehensively test for all the known the reduced dose. The onus is on the tests.9 Such data would be very useful genotypes of the UGT1A1 promoter ordering oncologist to consult with if third-party payers are to shoulder the polymorphisms, the mere fact that only other colleagues and the literature to cost of pharmacogenomic testing.
that portion of the gene is assessed and make sense of how to use the test result Third, successful integration of not the entire gene sequence will mean to manage irinotecan dosing. Without pharma cogenetic testing in the clinic that other potentially important variants clear instructions on how to interpret requires not only a clinical laboratory are missed, variants that may have an and use the test results, oncologists, who that establishes and validates the assay impact on clinical care. For example, the are generally not trained in interpreting but also a rapid reporting system that currently FDA-approved UGT1A1 test genetic tests, may be reluctant to order provides appropriate guidance about the does not include the reduced-activity *6 the test in the first place.
interpretation of the test results. Given allele in the coding region (211G→A, the complexity of the process, it will G71R), which is less common than the Discussion
probably require the coordinated efforts *28 allele but has been shown to be pre- Given the challenges for the appropriate of multiple health-care professionals, dictive of hyperbilirubinemia, especially use of the UGT1A1 pharmacogenetic test, including laboratory medicine specialists, it is clear that considerably more clarifi- physicians, nurses, and pharmacists.
Second, an inexpensive biomarker, cation of the role of the UGT1A1 test in Finally, clinicians' knowledge of phar- total serum bilirubin, is available as the management of irinotecan toxicity is macogenetic testing may also influ- a clinical predictor of liver function necessary. Importantly, these challenges ence its successful integration into the and can serve as a surrogate marker of are not unique to the UGT1A1 test, and clinic. A recent systematic review sug- UGT1A1 enzyme function and severe therefore solutions to redress knowledge gests that clinicians are not generally neutropenia before administering iri- gaps in the use of UGT1A1 testing can confident in providing genetic services notecan.6 Some oncologists argue be extended to other pharmaco genetic to patients because of lack of training against the use of the UGT1A1 geno- tests. First, pharmacogenetic tests could and knowledge.10 Therefore, efforts are type because they think bilirubin levels, be more clinically applicable if they needed to improve clinicians' knowl- with which they are more comfortable, included a comprehensive survey of edge of pharmaco genetic tests in order may be as reliable an indicator as the variation in the human genome. Many to facilitate their successful integration genotype in predicting the appropriate current pharmacogenetic tests evalu- into clinical practice.
irinotecan dose (anecdotal evidence). ate one or a few candidate genes with ACKNOWLEDGMENTS
However, there are no studies showing a biologically plausible link to drug Drs Ikediobi and Shin contributed equally to how total bilirubin levels may be used to responses. Although successful for some this work. This study was partially supported by guide irinotecan dose modification or a drugs, this approach may miss impor- Program Project Grant P01CA130818 from the National Cancer Institute to K.A.P.
priori selection. Thus, a second challenge tant contributions of variation in other is uncertainty over the clinical utility of genes, thereby reducing the predictive CONFLICT OF INTEREST
UGT1A1 genotype and serum bilirubin value of the test. It is promising that The authors declared no conflict of interest.
in guiding the appropriate dose selection many recent pharmacogenomic studies for irinotecan.
use a genome-wide approach. In addi- VOLUME 86 NUMBER 1 JULY 2009 1. Pollack, A. Patient's DNA may be signal syndrome. Pharmacogenomics 9, 703–715
Practice and Prevention (EGAPP) Working to tailor medication. New York Times, 29 Group. Recommendations from the EGAPP December 2008.
5. Ando, Y. et al. Polymorphisms of UDP- Working Group: can UGT1A1 genotyping 2. US Food and Drug Administration. Table of glucuronosyltransferase gene and Irinotecan reduce morbidity and mortality in patients Valid Genomic Biomarkers in the Context of toxicity: a pharmacogenetic analysis. Cancer with metastatic colorectal cancer treated with Approved Drug Labels < Res. 60, 6921–6926 (2000).
Irinotecan? Genet. Med. 11, 15–20 (2009).
6. Innocenti, F. et al. Genetic variants in the UDP- 9. Hoskins, J.M., Goldberg, R.M., Qu, P., Ibrahim, (updated 10 September 2008).
glucuronosyltransferase 1A1 gene predicts J.G. & McLeod, H.L. UGT1A1*28 genotype 3. Beutler, E., Gelbart, T. & Demina, A. Racial the risk of severe neutropenia of Irinotecan. J. and irinotecan-induced neutropenia: dose variability in the UDP-glucuronosyltransferase Clin. Oncol. 8, 1382–1388 (2004).
matters. J. Natl. Cancer Inst. 99, 1290–1295
1 (UGT1A1) promoter: a balanced 7. Camptosar full prescribing information. Pfizer polymorphism for regulation of bilirubin 10. Scheuner, M.T., Sieverding, P. & Shekelle, P.G. metabolism? Proc. Natl. Acad. Sci. USA 95,
> (July 2008). Accessed 4 Delivery of genomic medicine for common 8170–8174 (1998).
January 2009.
chronic adult diseases: a systematic review. 4. Strassburg, C.P. Pharmacogenetics of Gilbert's 8. Evaluation of Genomic Applications in JAMA 299, 1320–1334 (2008).
CLInICAL PhARMACoLogy & TheRAPeuTICs VOLUME 86 NUMBER 1 JULY 2009

Source: http://www.uc.pt/en/fmuc/phdhs/Courses/genetics/Pharmacogenomics.pdf