5
http://tpx.sagepub.com/ Toxicologic Pathology http://tpx.sagepub.com/content/32/1_suppl/9 The online version of this article can be found at: DOI: 10.1080/01926230490424743 2004 32: 9 Toxicol Pathol David A. Hosford, Eric H. Lai, John H. Riley, Chun-Fang Xu, Theodore M. Danoff and Allen D. Roses Pharmacogenetics to Predict Drug-Related Adverse Events Published by: http://www.sagepublications.com On behalf of: Society of Toxicologic Pathology can be found at: Toxicologic Pathology Additional services and information for http://tpx.sagepub.com/cgi/alerts Email Alerts: http://tpx.sagepub.com/subscriptions Subscriptions: http://www.sagepub.com/journalsReprints.nav Reprints: http://www.sagepub.com/journalsPermissions.nav Permissions: http://tpx.sagepub.com/content/32/1_suppl/9.refs.html Citations: What is This? - Jan 1, 2004 Version of Record >> by guest on August 12, 2013 tpx.sagepub.com Downloaded from

Toxicologic Pathology - GBC Adversos...Society of Toxicologic Pathology Additional services and information for Toxicologic Pathology can be found at:

Embed Size (px)

Citation preview

http://tpx.sagepub.com/Toxicologic Pathology

http://tpx.sagepub.com/content/32/1_suppl/9The online version of this article can be found at:

 DOI: 10.1080/01926230490424743

2004 32: 9Toxicol PatholDavid A. Hosford, Eric H. Lai, John H. Riley, Chun-Fang Xu, Theodore M. Danoff and Allen D. Roses

Pharmacogenetics to Predict Drug-Related Adverse Events  

Published by:

http://www.sagepublications.com

On behalf of: 

  Society of Toxicologic Pathology

can be found at:Toxicologic PathologyAdditional services and information for    

  http://tpx.sagepub.com/cgi/alertsEmail Alerts:

 

http://tpx.sagepub.com/subscriptionsSubscriptions:  

http://www.sagepub.com/journalsReprints.navReprints:  

http://www.sagepub.com/journalsPermissions.navPermissions:  

http://tpx.sagepub.com/content/32/1_suppl/9.refs.htmlCitations:  

What is This? 

- Jan 1, 2004Version of Record >>

by guest on August 12, 2013tpx.sagepub.comDownloaded from

Toxicologic Pathology, 32(Suppl. 1):9–12, 2004Copyright C© by the Society of Toxicologic PathologyISSN: 0192-6233 print / 1533-1601 onlineDOI: 10.1080/01926230490424743

Pharmacogenetics to Predict Drug-Related Adverse Events

DAVID A. HOSFORD,1 ERIC H. LAI,1 JOHN H. RILEY,1 CHUN-FANG XU,1 THEODORE M. DANOFF,2AND ALLEN D. ROSES1

1Department of Genetics Research and2Department of Clinical Pharmacology and Discovery Medicine,

GlaxoSmithKline R&D, Research Triangle Park, North Carolina 27709, USA

ABSTRACT

Identification of reliable markers to predict drug-related adverse events (DRAEs) is an important goal of the pharmaceutical industry and otherswithin the healthcare community. We have used genetic polymorphisms, including the most frequent source of variation (single nucleotide poly-morphisms, SNPs) in the human genome, in pharmacogenetic approaches designed to predict DRAEs. Three studies exemplify the principles ofusing polymorphisms to identify associations in progressively larger genomic regions: polymorphic repeats within the UDP-glucuronysltransferase I(UGT1A1) gene in patients experiencing hyperbilirubinemia after administration of tranilast, an experimental drug to prevent re-stenosis followingcoronary revascularization; high linkage disequilibrium within the Apolipoprotein E (ApoE) gene in patients with Alzheimer Disease (AD); and thepolymorphic variant HLA-B57 in patients with hypersensitivity reaction after administration of abacavir, a nucleoside reverse transcriptase inhibitorfor the treatment of HIV. Together, these studies demonstrate in a stepwise manner the feasibility of using pharmacogenetic approaches to predictDRAEs.

Keywords. Pharmacogenetics; adverse drug reaction; drug-related adverse event; single nucleotide polymorphism; Gilbert’s syndrome; Tranilast;Apolipoprotein E; Alzheimer’s Disease; abacavir; hypersensitivity reaction.

INTRODUCTION

Adverse drug reactions are a significant cause of morbid-ity and mortality in the United States (Schenkel, 2000), oc-curring in more than 3% of patients admitted into hospi-tals in the US and other developed countries (Thomas et al.,2000). Drug-related adverse events (DRAEs) prolong hos-pital stays and hospital costs dramatically (Rigby and Litt,2000; Schioler et al., 2001), and they produce significant med-ical and economic consequences within the community evenwhen they do not result in hospitalization (Haramburu et al.,2000). Identification of reliable markers to predict DRAEsis a continuing goal of the pharmaceutical industry and ofothers within healthcare, particularly if the marker(s) has(have) a high negative predictive value, rapid turn-around,and affordable cost. Pharmacogenetic approaches to identifypredictive genetic markers for DRAEs are now feasible be-cause of 4 factors: 1) incremental advances and decreasingcosts of high throughput-genotyping; 2) the availability oflarge numbers of single nucleotide polymorphisms (SNPs)to serve as potential predictive markers; 3) new statisticalgenetics approaches to analyze SNP data and detect associa-tions with DRAEs in patient cohorts; and 4) the availabilityof positional information from the human genome sequenc-ing project (Roses, 2002). Application of these pharmacoge-netic approaches to DRAEs is being pursued in clinical trialsand studies within GlaxoSmithKline R&D (GSK), and thefindings from these studies demonstrate that this approach isfeasible.

Address correspondence to: David A. Hosford, GlaxoSmithKlineR&D, 5.5820, 5 Moore Drive, Research Triangle Park, North Carolina27709, USA; e-mail: [email protected]

Three studies will be reviewed here, to demonstrate thefeasibility of pharmacogenetic approaches that use polymor-phisms including SNPs in a high-throughput manner. Thesestudies were chosen from the many performed in GSK, be-cause they illustrate the progressively higher-throughput useof SNPs: from candidate gene approach; to mapping acrossa larger multigene region; to screening across the wholegenome. In the first study, polymorphic repeats in a can-didate gene were used to demonstrate the pharmacogeneticbasis of a DRAE (Roses, 2002; Xu et al. [unpublished ob-servations]). In the second study, high-density SNP-mappingacross a larger genomic region was used to detect associationsof a gene to disease, illustrating the principle of high-densitySNP approaches (Lai et al., 1998; Roses, 2002). In the thirdstudy, whole-genome SNP mapping is being applied to a co-hort with a DRAE to confirm and extend initial observationsthat were made on the basis of polymorphisms within a casecontrol study (Hetherington et al., 2002).

Polymorphic Repeats in the UGT1A1 Gene in PatientsAdministered Tranilast

Tranilast is a compound that was in phase III trials todetermine if the drug prevented restenosis following coro-nary revascularization. In the trial, approximately 12% ofsubjects developed transient, reversible hyperbilirubinemia(levels greater than 2 mg per dL). The nature of this DRAEwas strikingly similar to the transient benign hyperbilirubine-mia that characterizes Gilbert’s syndrome, and this similar-ity led to the hypothesis that subjects experiencing tranilast-induced hyperbilirubinemia may carry genetic variants whichhave been associated with Gilbert’s syndrome. For example,Gilbert’s syndrome has been associated with homozygosityfor a (TA) 7-repeat element within the promoter region of the

9

by guest on August 12, 2013tpx.sagepub.comDownloaded from

10 HOSFORD ET AL TOXICOLOGIC PATHOLOGY

FIGURE 1.—Pre- and posttranilast bilirubin levels in 1231 individuals from the PRESTO trial. Cases (n = 146) were individuals with serum bilirubin levels greaterthan 2.0 mg/dl after tranilast administration; controls (n = 1,085) had bilirubin levels equal to or less than 2 mg/dl. Mean and standard error are shown.

UDP-glucuronysltransferase I (UGT1A1) gene (Bosma et al.,1992, 1995). To test the hypothesis that tranilast-induced hy-perbilirubinemia is expressed in patients with this geneticvariant, a candidate gene association study was performed.The results showed that in tranilast-treated subjects who werehomozygotes for the (TA)7 repeat element, 40% developedhyperbilirubinemia, compared to none of of the (TA)7 ho-mozygotes who were treated with placebo. (Figure 1) (Roses,2002; Xu et al., unpublished observations).

Unfortunately, tranilast failed to demonstrate efficacy inthis phase III trial, and so there is no reason to prospectivelyvalidate this observation and to determine if polymorphicvariants of the UGT1A1 gene (or other measures) will be auseful measure in patient treatment regimens using tranilast.Nevertheless, this result illustrates the principle of identify-ing polymorphic variants at a candidate gene level to predictDRAEs.

High-Density SNP-Mapping Within a Multigene Regionin Subjects with Alzheimer’s Disease

Clearly the above approach was successful because ofan educated hypothesis about variants within an appropri-ate candidate gene. However, application of a pharmaco-genetic approach to detect the basis for DRAEs will oftenrequire nonhypothesis driven approaches, using polymor-phisms that are common and frequent within the humangenome. SNPs exhibit these properties and hence are suit-able for nonhypothesis-driven approaches in larger genomicregions (Melton, 2003). A second study illustrates the prin-ciple of extending SNP-based approaches beyond candidategenes to larger, multigene regions.

The apolipoprotein E4 (ApoE4) allele is a susceptibilitygene variant that accounts for a large proportion of late-onsetAlzheimer Disease (AD; Saunders et al., 1993; Strittmatteret al., 1993). To test the feasibility of using a high-densitySNP association study to detect associations in specified co-horts, a high-density SNP map was made by identifying SNPswithin a large, 4-Mb region of chromosome 19 containingthe ApoE gene (Lai et al., 1998). When these SNPs weregenotyped in subjects with AD, the study showed high link-age disequilibrium (LD; i.e., high strength of association)

within the ApoE gene (Lai et al., 1998) and the surround-ing area (Figure 2). This result demonstrated the feasibil-ity of using high-density SNP genotyping to detect signif-icant associations within a case control cohort of subjects.Although the trait defining the cohort in this case was a dis-ease (i.e., AD) rather than a DRAE, this finding demonstratesthe feasibility of a SNP-based approach in larger genomicregions.

Whole-Genome SNP-Mapping to Detect Genes Associatedwith Abacavir Hypersensitivity Reaction

The third example demonstrates application of SNP-basedpharmacogenetics to the whole genome, in a stepwise ap-proach. The cohort chosen for this study was subjects whodeveloped hypersensitivity reactions after administration ofabacavir (Hetherington et al., 2002). Hypersensitivity reac-tion is a DRAE that occurs in approximately 4% of subjectsadministered this drug (Hetherington et al., 2001). Becausethe human leukocyte antigen (HLA) complex on chromo-some 6 encodes many genes whose products participate inhypersensitivity reactions, a study was performed to exam-ine polymorphisms within the HLA region in abacavir-treatedsubjects. Polymorphisms within the HLA-B region occurredwith significantly greater frequency in cases who devel-oped hypersensitivity after abacavir, compared to controlswho did not develop hypersensitivity (Hetherington et al.,2001). In particular, the polymorphic variants HLA-B57 andTNF-alpha were identified in 46% and 43% (respectively) incases with the DRAE of hypersensitivity, compared to 4%and 7% (respectively) of controls without this DRAE ( p <.0001) (Figure 3). Similar results were reported in anotherstudy (Mallal et al., 2002), and these findings have now beenreplicated by GSK in a second, larger cohort (E Lai et al.,unpublished observations).

Interestingly, Caucasians but not African-Americansdemonstrated the association between these polymorphismswithin the HLA-B region (Hetherington et al., 2001). It is un-certain whether this lack of association within the African-American cohort was due to the smaller sample size of thatcohort, or to other factors that include ethnic differences in

by guest on August 12, 2013tpx.sagepub.comDownloaded from

Vol. 32(Suppl. 1), 2004 PHARMACOGENETICS OF ADVERSE EVENTS 11

FIGURE 2.—Localized SNP scan around apoE. The distance from apoE was determined by sequencing a cosmid. The p values were calculated for allelic differencesusing standard chi-squared tests in an affected Alzheimer population of 270 individuals and an age-matched control population of 278 individuals.

genes associated with specific DRAEs. This illustrates thepotential need to perform pharmacogenetic studies in multi-ple cohorts that have specific DRAEs but differ on the basisof ethnicity, in order to identify a complete profile of DRAE-associated markers that will be predictive for the entire humanpopulation taking a drug.

FIGURE 3.—Association of abacavir related hypersensitivity reactions with markers in the HLA locus. The p values were calculated using Fisher’s Exact teststatistic to determine whether there was an association between the presence of a given allele and the presence of hypersensitivity in 84 cases and 113 controls. Nop-values were adjusted for multiple testing. The chromosomal position of the markers is based on the sequence information at The Wellcome Trust Sanger Institute〈http://www.sanger.ac.uk/HGP/Chr6/〉.

Moreover, the predictiveness of HLA-B polymorphismsaccounts for only half of cases with hypersensitivity, even inthe Caucasian population (Hetherington et al., 2001). For thatreason, the study has now been extended to a larger cohortand a whole-genome SNP-based approach is being used. Theresults forthcoming should increase the predictive nature of

by guest on August 12, 2013tpx.sagepub.comDownloaded from

12 HOSFORD ET AL TOXICOLOGIC PATHOLOGY

associated SNPs to a larger proportion of subjects with hy-persensitivity to abacavir.

SUMMARY

Together, the results of these 3 studies illustrate that theprinciple of using pharmacogenetic markers to predict actualDRAEs is feasible. These studies also demonstrate the needto apply these approaches in cohorts of different ethnicities,and to consider a whole genome approach, in order to obtaina profile of markers that will be fully predictive in the globalclinical setting.

The goal of this approach will only be fully met, when thecost and speed of detecting predictive markers in an individ-ual subject is appropriate for routine use in a clinical setting;and when the negative predictive value of these markers ishigh enough to warrant their commonplace use. This goalappears closer with the results of each new study.

ACKNOWLEDGMENTS

GlaxoSmithKline scientists who contributed to the worksummarized in this review include: Claire Allan, KatherineBaker, Clive Bowman, Kirstie Davies, David Dow, MaryFling, Abigail Handley, Seth Hetherington, Arlene Hughes,David Kazierad, Karen Lewis, Linda McCarthy, MichaelMosteller, Ian Purvis, Mary Repasch, Denise Shortino,William Spreen, Nigel Spurr, Michael Stocum, LindaThurmond and numerous others within the Departmentsof Genetics Research, and Clinical Pharmacology andDiscovery Medicine.

REFERENCES

Bosma, P., Chowdhury, J. R., Bakker, C., Gantla, S., de Boer, A., Oostra, B.A., Lindhout, D., Tytgat, G. N., Jansen, P. L., and Oude Elferink, R. P.(1995). The genetic basis of the reduced expression of bilirubin UDP-glucuronosyltransferase 1 in Gilbert’s syndrome. N Eng J Med 333, 1171–5.

Bosma, P., Chowdhury, N. R., Goldhoorn, B. G., Hofker, M. H., Oude Elferink,R. P., Jansen, P. L., and Chowdhury, J. R. (1992). Sequence of exons and theflanking regions of human bilirubin-UDP-glucuronosyltransferase genecomplex and identification of a genetic mutation in a patient with Crigler-Najjar syndrome, type I. Hepatology 15: 941–7.

Haramburu, F., Pouyanne, P., Imbs, J. L., Blayac, J. P., and Begaud, B. (2000).Incidence and prevalence of adverse drug reactions. Presse Medicale 29,111–4.

Hetherington, S., Hughes, A. R., Mosteller, M., Shortino, D., Baker, K. L.,Spreen, W., Lai, E., Davies, K., Handley, A., Dow, D. J., Fling, M. E.,Stocum, M., Bowman, C., Thurmond, L. M., and Roses, A. D. (2002). Ge-netic variations in HLA-B region and hypersensitivity reactions to abacavir.Lancet 359, 1121–2.

Hetherington, S., McGuirk, S., Powell, G., Cutrell, A., Naderer, O., Spreen,B., Lafon, S., Pearce, G., and Steel, H. (2001). Hypersensitivity reactionsduring therapy with the nucleoside reverse transcriptase inhibitor abacavir.Clin Ther 23, 1603–14.

Lai, E., Riley, J., Purvis, I., and Roses, A. (1998). A 4-Mb high-density sin-gle nucleotide polymorphism-based map around ApoE. Genomics 54,31–8.

Mallal, S., Nolan, D., Witt, C., Masel, G., Martin, A. M., Moore, C., Sayer, D.,Castley, A., Mamotte, C., Maxwell, D., James, I., and Christiansen, F. T.(2002). Lancet 359, 727–32.

Melton, L. (2003). On the trail of SNPs. Nature 422, 917–23.Rigby, K. D., and Litt, J. C. (2000). Errors in health care management: what do

they cost? Qual Health Care 9, 216–21.Roses, A. D. (2002). Genome-based pharmacogenetics and the pharmaceutical

industry. Nature Rev 1, 541–9.Saunders, A. M., Strittmatter, W. J., Schmechel, D., St George-Hyslop, P. H.,

Pericak-Vance, M. A., Joo, S. H., Rosi, B. L., Gusella, J. F., Crapper-MacLachlan, D. R., Alberts, M. J., and Roses, A. D. (1993). Association ofapolipoprotein E allele E4 with late-onset familial and sporadic alzheimerdisease. Neurology 43, 1467–72.

Schenkel, S. (2000). Promoting patient safety and preventing medical error inemergency departments. Acad Emer Med 7, 1204–22.

Schioler, T., Lipczak, H., Pedersen, B. L., Mogensen, T. C., Bach, K. B.,Stockmarr, A., Svenning, A. R., and Frolich, A. (2001). Incidence of ad-verse events in hospitals. A retrospective study of medical records. ULaeger 163, 5370–8.

Strittmatter, W. J., Saunders, A. M., Schmechel, D., Pericak-Vance, M.,Enghild, J., Salvesen, G. S., and Roses, A. D. (1993). ApolipoproteinE: high-avidity binding to beta-amyloid and increased frequency of type 4allele in late-onset familial alzheimer disease. Proc Natl Acad Sci USA 90,1977–81.

Thomas E. J., Studdert, D. M., Runciman, W. B., Webb, R. K., Sexton, E. J.,Wislon, R. M., Gibberd, R. W., Harrison, B. T., and Brennan, T. A. (2000).A comparison of iatrogenic injury studies in australia and the USA. I. Con-text, methods, casemix, population, patient and hospital characteristics. IntJ Qual Health Care 12, 371–8.

by guest on August 12, 2013tpx.sagepub.comDownloaded from