pharmacogenomics of tamoxifen therapy

Pharmacogenomics of Tamoxifen TherapyHiltrud Brauch,1,2* Thomas E. Murdter,1,2 Michel Eichelbaum,1,2 and Matthias Schwab1,2,3

BACKGROUND: Tamoxifen is a standard endocrine ther-apy for the prevention and treatment of steroid hor-mone receptor–positive breast cancer.

CONTENT: Tamoxifen requires enzymatic activation bycytochrome P450 (CYP) enzymes for the formation ofactive metabolites 4-hydroxytamoxifen and endoxifen.As compared with the parent drug, both metaboliteshave an approximately 100-fold greater affinity for theestrogen receptor and the ability to inhibit cell prolif-eration. The polymorphic CYP2D6 is the key enzyme inthis biotransformation, and recent mechanistic, pharma-cologic, and clinical evidence suggests that genetic vari-ants and drug interaction by CYP2D6 inhibitors influencethe plasma concentrations of active tamoxifen metabo-lites and the outcomes of tamoxifen-treated patients. Inparticular, nonfunctional (poor metabolizer) and se-verely impaired (intermediate metabolizer) CYP2D6 al-leles are associated with higher recurrence rates.

SUMMARY: Accordingly, CYP2D6 (cytochrome P450,family 2, subfamily D, polypeptide 6) genotyping be-fore treatment to predict metabolizer status may opennew avenues for individualizing endocrine treatment,with the maximum benefit being expected for exten-sive metabolizers. Moreover, strong CYP2D6 inhibi-tors such as the selective serotonin reuptake inhibitorsparoxetine and fluoxetine, which are used to treat hotflashes, should be avoided because they severely impairformation of the active metabolites.© 2009 American Association for Clinical Chemistry

The pharmacogenomics of drug-metabolizing en-zymes involved in the biotransformation of tamoxifenhas become a major area of interest, owing to its poten-tial to predict a breast cancer patient’s treatment out-come before the initiation of treatment. If the tamox-ifen pharmacogenomic paradigm were to be borne outin proof of principle, patients eligible for endocrine

treatment would be able to exploit it by opting for theirpersonally most powerful medication. Most breastcancers, particularly those of postmenopausal women,are hormone receptor positive; therefore, hundreds ofthousands of women worldwide initiate endocrinetreatment each year. On the basis of results of the EarlyBreast Cancer Trialist Collaborative Group, the stan-dard recommendation has been 5 years of therapy withthe selective estrogen receptor (ER)4 modulator ta-moxifen (1 ). Tamoxifen is currently prescribed in�120 countries worldwide as a component of standardadjuvant therapy in early breast cancer and in the met-astatic setting for patients with steroid hormonereceptor–positive breast tumors. In primary breastcancer, adjuvant tamoxifen significantly decreases re-lapse rates and mortality in pre- and postmenopausalpatients, and the therapy benefit from 5 years of adju-vant tamoxifen is maintained, even �10 years after pri-mary diagnosis (1 ). In postmenopausal women withendocrine-responsive disease, tamoxifen is a validtherapy option, along with aromatase inhibitors (AIs)(2 ), and is considered the standard care for the preven-tion of invasive breast cancer in premenopausalwomen at high risk, including those who have had duc-tal carcinoma in situ (3 ), and for the treatment of malebreast cancer (4 ). Tamoxifen is generally well toler-ated, and menopausal symptoms, including hotflashes, are the most common side effects. Severe sideeffects, such as thromboembolic events or endometrialcarcinoma, are rare (1 ). The clinical benefit of tamox-ifen has been evident for more than 3 decades; how-ever, up to 50% of patients who receive adjuvanttamoxifen relapse or die from tumor-specific resis-tance or host genome–associated factors.

The field of tamoxifen pharmacogenomics gainedsubstantial impetus from the elucidation of tamoxifenmetabolism and metabolite pharmacology throughstudies that identified major active metabolites formedby cytochrome P450 (CYP) enzymes, particularlyCYP2D6, which exhibit substantial genetic and pheno-typic polymorphism. Several clinical studies have re-ported on the relationship of genotype and/or pheno-type variants with the clinical outcome of tamoxifen

1 Dr. Margarete Fischer-Bosch Institute of Clinical Pharmacology, Stuttgart, Ger-many; 2 University Tubingen, Tubingen, Germany; 3 Department of ClinicalPharmacology, University Hospital Tubingen, Tubingen, Germany.

* Address correspondence to this author at: Dr. Margarete Fischer-Bosch Instituteof Clinical Pharmacology, Auerbachstrasse 112, 70376 Stuttgart, Germany. Fax�49-(0)711-859295; e-mail [email protected].

Received November 28, 2008; accepted June 8, 2009.Previously published online at DOI: 10.1373/clinchem.2008.121756

4 Nonstandard abbreviations: ER, estrogen receptor; CYP, cytochrome P450; UGT,UDP-glucuronosyltransferase; SULT, sulfotransferase; EM, extensive metabo-lizer; PM, poor metabolizer; IM, intermediate metabolizer; UM, ultrarapidmetabolizer; SSRI, serotonin reuptake inhibitor; AI, aromatase inhibitor.

Clinical Chemistry 55:101770–1782 (2009) Review

1770

therapy, and international efforts are currently underway to clarify this relationship.

In light of the potential for a future translation oftamoxifen pharmacogenomics into clinical practice,this review seeks to impart the underlying pharmaco-logic, genetic, and phenotypic principles for a mecha-nistic explanation of tamoxifen efficacy. It highlightsthe biotransformation of tamoxifen into primary andsecondary metabolites with an emphasis on the clini-cally active metabolites 4-hydroxytamoxifen and4-hydroxy-N-desmethyltamoxifen (endoxifen). Ow-ing to the key role of CYP2D6, this review focuses onthe relationships between the CYP2D65 (cytochromeP450, family 2, subfamily D, polypeptide 6) genotypeand phenotype. This discussion also includes the phe-nocopying effect of CYP2D6 inhibitors, which are fre-quently coadministered to alleviate hot flashes in post-menopausal women treated with tamoxifen. Thesebasic research findings provide the scientific back-ground for a thorough discussion of the currentlyavailable literature on tamoxifen pharmacogeneticstudies. Finally, there is a possibility that other drug-metabolizing enzymes and even nonmetabolic factorscan influence tamoxifen efficacy. In considering thesetopics, this review provides an overview of the princi-ples of the emerging practice of personalized medicinefor the improvement of the outcomes of endocrinedrug treatment in breast cancer.

Tamoxifen Metabolism and Active Metabolites

trans-Tamoxifen {(Z)-2-[4-(1,2-diphenylbut-1-enyl)-phenoxy]-N,N-dimethyl-ethanamine} undergoes ex-tensive phase I and phase II metabolism in the hu-man liver (Fig. 1). The bioconversion of tamoxifeninvolves N-oxidation, N-demethylation, and hy-droxylation. Formation of the major metaboliteN-desmethyltamoxifen is primarily catalyzed byCYP3A4 and 3A5, with minor contributions byCYP2D6, 1A1, 1A2, 2C19, and 2B6 (5–7 ). The steady-state plasma concentration of N-desmethyltamoxifenafter 20 mg tamoxifen is administered daily for at least3 months is approximately twice as high as that of the

parent drug (100 –290 �g/L vs 72–160 �g/L) (8 –14 ).This fact is of utmost clinical importance becauseN-desmethyltamoxifen is subject to hydroxylation,predominantly at the para position, to produce the ma-jor clinically active metabolite endoxifen. Importantly,the conversion of N-desmethyltamoxifen into endox-ifen is catalyzed almost exclusively by CYP2D6(15, 16 ). Plasma concentrations of endoxifen havebeen observed to range from a mean of 8.1 �g/L (n � 51)for patients with 2 variant CYP2D6 alleles to 20.7 �g/L(n � 55) for patients with 2 wild-type alleles (17). In ad-dition, N-desmethyltamoxifen can also be desmethylatedby CYP3A4 to form N,N-didesmethyltamoxifen.

Another clinically active metabolite is 4-hydroxytamoxifen, which is formed by 4-hydroxylation,also at the para position of the phenyl ring of the parentdrug. This conversion is catalyzed by a number of CYPs,including CYP2D6, 3A4, 2C9, 2B6, and 2C19 (7, 18 –21 ). Compared with endoxifen, the steady-state con-centrations of 4-hydroxytamoxifen are lower, rangingfrom 1.15 �g/L to 6.4 �g/L (11, 14, 22 ). With the ex-ception of endoxifen and 4-hydroxytamoxifen, noother highly active metabolites have been describedthus far.

Further hydroxylation also occurs at the 4� posi-tion of the other phenyl ring system, leading to 4�-hydroxytamoxifen, which is mainly mediated byCYP2B6 and 2D6 (7 ), and to 4�-hydroxy-N-desmethyltamoxifen. Another hydroxylated metabo-lite, �-hydroxytamoxifen, is produced mainly byCYP3A4 (5, 6, 23, 24 ).

4-Hydroxylated metabolites undergo in vitrochemical isomerization into the respective E or cis iso-mers (25 ), which are weak ER antagonists. In addition,isomerization of 4-hydroxytamoxifen is catalyzed byCYP1B1, 2B6, and 2C19 (7 ). Of note, an accumulationof cis-4-hydroxytamoxifen was observed in tumor tis-sues of patients whose tumors showed resistance to ta-moxifen treatment (26 ); however, because data on theplasma concentrations of cis isomers are sparse, thisobservation may be regarded as preliminary. Addi-tional hydroxylation of 4-hydroxytamoxifen byCYP3A4 and 2D6 at the phenyl moiety leads to 3,4-dihydroxytamoxifen (27 ), a compound that is capableof binding covalently to protein and to DNA, therebycontributing to the reported toxic and carcinogenic ef-fects associated with tamoxifen treatment (28, 29 ).

Another route of tamoxifen metabolism is the for-mation of tamoxifen-N-oxide by flavin-containingmonooxygenases 1 and 3, with a chance for tamoxifen-N-oxide to be reduced back to tamoxifen by a numberof different CYPs, including CYP2A6, 1A1, 3A4, andothers (30, 31 ). From an analytical point of view, how-ever, this metabolite cannot be ignored because of thelikelihood of chemical reduction of the N-oxide during

5 Human genes: CYP2D6, cytochrome P450, family 2, subfamily D, polypeptide 6;CYP2D7P1, cytochrome P450, family 2, subfamily D, polypeptide 7 pseudogene1; CYP2D7P2, cytochrome P450, family 2, subfamily D, polypeptide 7 pseudo-gene 2; CYP2D8P1, cytochrome P450, family 2, subfamily D, polypeptide 8pseudogene 1; CYP2D8P2, cytochrome P450, family 2, subfamily D, polypeptide8 pseudogene 2; CYP2C9, cytochrome P450, family 2, subfamily C, polypeptide9; CYP2C19, cytochrome P450, family 2, subfamily C, polypeptide 19; CYP2B6,cytochrome P450, family 2, subfamily B, polypeptide 6; CYP3A4, cytochromeP450, family 2, subfamily A, polypeptide 4; CYP3A5, cytochrome P450, family2, subfamily A, polypeptide 5; SULT1A1, sulfotransferase family, cytosolic, 1A,phenol-preferring, member 1; BRCA1, breast cancer 1, early onset; BRCA2,breast cancer 2, early onset.

Tamoxifen Pharmacogenomics Review

Clinical Chemistry 55:10 (2009) 1771

Fig. 1. Metabolic transformation of tamoxifen in humans.

Major metabolic pathways are highlighted with bold arrows. Enzymes preferentially catalyzing a distinct metabolic step areindicated in bold. Hb, hemoglobin; FMO1, flavin-containing monooxygenase 1.

Review

1772 Clinical Chemistry 55:10 (2009)

sample preparation, a reason why the quantification oftamoxifen-N-oxide may be regarded as a problem inthe analysis of tamoxifen metabolites. Thus far, fewdata on this issue are available, suggesting that theN-oxide in a patient’s plasma accounts for �15% oftamoxifen (32 ).

At the level of phase II tamoxifen metabolism, sul-fation and glucuronidation are major mechanisms.O-glucuronidation of 4-hydroxytamoxifen is mainlymediated by UDP-glucuronosyltransferases (UGTs)UGT1A4, 2B15, 2B7, 1A8, and various others to produce4-hydroxytamoxifen-O-glucuronide (33–35). Endox-ifen is predominantly glucuronidated by UGT1A10 and1A8 to the corresponding O-glucuronide. Of note isthat in addition to hydroxylated metabolites that un-dergo phase II metabolism at the hydroxyl moiety,tamoxifen itself is conjugated by UGT1A4 to the cor-responding N�-glucuronide (36, 37 ). In contrast toendoxifen, which does not form any N�-glucuronide,4-hydroxytamoxifen is glucuronidated by UGT1A4 atthe amino group to produce 4-hydroxytamoxifen-N�-glucuronide (34, 37 ). The formation of sulfatesof 4-hydroxytamoxifen and endoxifen is catalyzedby sulfotransferases (SULTs) SULT1E1, 1A1, and 2A1(32, 38 ). E isomers of both 4-hydroxytamoxifen andendoxifen are also substrates for these conjugationreactions but seem to have different affinities for differ-ent isoenzymes (33). �-Hydroxytamoxifen is sulfatized

by SULT2A1 (39 ); the resulting �-hydroxytamoxifensulfate is suspected to exert carcinogenic effects aftercovalently binding to DNA (40, 41 ).

Although the number of tamoxifen metabolitesthat have been identified in vitro is large (Fig. 1), invivo analytical measurements of plasma samplesfrom tamoxifen-treated patients have quantified fewmetabolites, including N-desmethyltamoxifen, endox-ifen, 4-hydroxytamoxifen, N,N-didesmethyltamoxifen,�-hydroxytamoxifen, and tamoxifen-N-oxide (Table 1).Therefore, there may be other, yet-unidentified tamox-ifen metabolites present at relevant concentrations in pa-tients’ plasma.

CYP2D6 Biochemistry and Genetics

CYP2D6 is a member of CYP enzyme family 2, which inhumans constitutes one third of all CYPs and is one of thelargest and best studied of isoenzyme families. HumanCYPs are heme-containing monooxygenases, and the hu-man genome contains 57 CYP genes and about the samenumber of pseudogenes grouped into 18 families and44 subfamilies according to sequence similarities (http://drnelson.utmem.edu/CytochromeP450.html). CYP2D6is involved in the metabolism of many clinically im-portant drugs, including �-blockers, antiarrhythmics,antihypertensives, antipsychotics, antidepressants,opioids, and others. A recent analysis of the routes of

Table 1. Tamoxifen and metabolites.

Compounds

Mean plasmaconcentrations,

nmol/La

Effect on ER/affinity for ER(estradiol � 100%)

Involvement ofCYP2D6

Tamoxifen 190–420 Weak antagonist/2%b

N-Desmethyltamoxifen 280–800 Weak antagonist/1%b Minor

N,N-Didesmethyltamoxifen 90–120 Weak antagonist No

Endoxifen 14–130 Strong antagonist/equal to 4-hydroxytamoxifen Almost exclusively

4-Hydroxytamoxifen 3–17c Strong antagonist/188%b Among others

�-Hydroxytamoxifen 1 None No

3,4-Dihydroxytamoxifen No data available Weak antagonist/high affinity Together with CYP3A4

Tamoxifen-N-oxide 15–24 Weak antagonistd No

4-Hydroxytamoxifen-O-glucuronide No data available No antagoniste See 4-hydroxy-tamoxifen

4-Hydroxytamoxifen-N�-glucuronide No data available No antagoniste See 4-hydroxy-tamoxifen

Endoxifen-O-glucuronide No data available No antagoniste See endoxifen

�-Hydroxytamoxifen sulfate No data available No data available No

a Range of mean plasma concentrations according to different investigators [Dowsett et al. (9 ), Hutson et al. (10 ), Jin et al. (11 ), Lee et al. (12 ), Sheth et al. (14 ),Lim et al. (17 ), Stearns et al. (22 ), Gjerde et al. (32 ), Langan-Fahey et al. (108 )].

b According to Wakeling and Slater (109 ).c MacCallum et al. (13 ) reported much higher concentrations (67 nmol/L).d Might be due to reduction to tamoxifen.e According to Lazarus et al. (110 ).



elimination for the “top 200 drugs” in the US (http://www.rxlist.com; most frequently prescribed 200 drugs,April 2008) showed that 15% were drugs that areCYP2D6 substrates, compared with subfamiliesCYP3A (37%) and CYP2C (33%) (42 ).

The human CYP2D6 locus on chromosome 22 in-cludes the CYP2D6 gene and pseudogenes CYP2D7P1(cytochrome P450, family 2, subfamily D, polypep-tide 7 pseudogene 1), CYP2D7P2 (cytochrome P450,family 2, subfamily D, polypeptide 7 pseudogene 2),CYP2D8P1 (cytochrome P450, family 2, subfamily D,polypeptide 8 pseudogene 1), and CYP2D8P2 (cyto-chrome P450, family 2, subfamily D, polypeptide 8pseudogene 2) originally described as pseudogenesCYP2D7 and CYP2D8 (43 ). The CYP2D6 gene consistsof 9 exons and 8 introns, and the sequence is highlypolymorphic. By way of clinical observation (i.e., ad-ministration of the antiarrhythmic and oxocytic drugsparteine (44 ) and the antihypertensive agent debriso-quine (45 )), the first CYP2D6 phenotypic variant(sparteine/debrisoquine polymorphism) distinct froman extensive metabolizer (EM) phenotype was identi-fied more than 30 years ago and was termed a “poormetabolizer” (PM) phenotype. Currently, 4 CYP2D6phenotypes are commonly observed in Caucasian pop-ulations on the basis of their drug-oxidation capacities:EM, intermediate metabolizer (IM), PM, and ultra-rapid metabolizer (UM) (46 – 48 ). Among Caucasians,about 7%–10% of individuals are PMs, 10%–15% areIMs, and, at the opposite end of the activity spectrum,up to 10%–15% are UMs.

The PM status can be deduced with �99% cer-tainty from the presence of 2 nonfunctional alleles,with �20 null alleles having been identified (43 ).Therefore, it is possible to exactly predict the CYP2D6PM phenotype (i.e., lack of catalytic function of theenzyme) by genotyping the patient’s DNA without theneed to phenotype (42, 46, 48, 49 ). The EM phenotypeis due to the presence of 1 or 2 allelic variants withwild-type function, such as *1 or *2. This phenotypecan be separated by genotype into homozygous orheterozygous EMs, depending on whether they carry1 or 2 functional alleles. Because heterozygous EMswho carry one *1 or *2 allele in combination with anIM or PM allele have somewhat impaired enzymeproduction and function, they have been classified asIMs, assuming a gene-dosage effect such that het-erozygous EMs would have only 50% of the enzymeamount and catalytic activity of homozygous EMs.This assumption is not correct, however, and there issubstantial overlap between homozygous and het-erozygous EMs in both enzyme content and activity.Consequently, the genotype has a rather poor pre-dictive value. Of note is that the IM has a phenotypeand genotype distinct from the heterozygous EM

(47, 50 –52 ) that involves impaired gene expressionand enzyme function (these variants include *9, *10,and *41) and/or nonfunctional variants (47, 52 ).Within the German population, 2%–3% are carriersof a duplicated/multiplied CYP2D6 gene and there-fore have very high enzyme activity (UM). These dif-ferences in enzyme activity can have profound con-sequences on the plasma concentrations of drugmetabolites, as has been observed for the tricyclicantidepressant nortriptyline. A �30-fold differencebetween PMs and UMs in steady-state plasma con-centrations of nortriptyline was observed when nor-triptyline was prescribed as a standard daily dose of100 –150 mg (53, 54 ). With respect to UM pheno-type, however, only 20%–30% of UM phenotypesobserved in the Caucasian population are identifi-able through genotyping (46, 48, 55 ).

Thus far, systematic genetic analyses of largenumbers of individuals have led to the discovery of�100 different alleles [http://www.cypalleles.ki.se,(56 )]. At least 15 of these alleles encode nonfunc-tional gene products caused by aberrant splicing,nonsense codons, mutations of single base pairs,small insertions/deletions, larger chromosomal de-letions of the entire CYP2D6 gene, CYP2D6/CYP2D7hybrid genes, or mutations that cause lack of hemeincorporation or otherwise produce nonfunctionalfull-length proteins.

There are significant ethnic differences with re-spect to PM, IM, and UM frequencies, heralding thepossibility that different ethnic groups vary with re-spect to the clinical outcomes of drug therapy withCYP2D6 substrates. Within this context it is importantto appreciate that the frequency of gene duplication ismuch higher in northeastern African populations[e.g., 29% in Ethiopia (57 )] and in Saudi Arabia[21% (58 )] compared with populations of Europeandescent (59, 60 ). In Asian populations, however, theCYP2D6*10-associated IM is prevalent (61 ), withthe frequency in Han Chinese being 57% and the PMplaying a minor role (59 ).

Overall, an awareness of the CYP2D6 genotype–phenotype relationship may influence treatment deci-sions, particularly in cases for which an effective alter-native drug is available. As in the case of orallyadministered codeine, which in the 10% of Caucasianswho are PMs is not metabolized efficiently to mor-phine and therefore provides little analgesic effect,there is a chance that women with a CYP2D6 PM or IMgenotype/phenotype also will not benefit from the an-tiestrogenic effects of tamoxifen, owing to insufficientproduction of active metabolites. With respect to UMs,who in cases of codeine treatment develop severe opi-oid side effects due to rapid morphine formation(62, 63 ), it is important to note that such women pa-

Review


tients may be more susceptible to hot flashes duringtamoxifen therapy.

CYP2C9, 2C19, 2B6, 3A4, and 3A5 Genetics

Other important CYP isoenzymes of subfamily 2 thatare involved in the bioactivation of tamoxifen areCYP2C9, 2C19, and 2B6 (15, 18 ); these enzymes arealso polymorphic. Of the �30 variant alleles ofCYP2C9 (cytochrome P450, family 2, subfamily C,polypeptide 9), the *2 and *3 alleles have been thor-oughly investigated and found to be associated withsignificant but highly variable reductions in intrinsicclearance, depending on the substrate (64 ). The *3 al-lele is more strongly affected than *2, with a reductionin enzyme activity of up to 90% for some specific drugs(65 ). Both alleles are present in approximately 35% ofCaucasians but are less prevalent in black and Asianpopulations (42, 66 ). About 2% and 24% of individu-als in the Caucasian population are homozygous andheterozygous for the variants, respectively (67 ). Nu-merous clinical studies have demonstrated the clinicalsignificance of CYP2C9 genetics with respect to an as-sociation with higher incidences of adverse drug reac-tions. The most prominent example is warfarin, an an-ticoagulant, and several retrospective and prospectivestudies have confirmed that CYP2C9 genetics is clini-cally useful for adjusting warfarin dosage to reduce se-rious warfarin-related bleeding events (68, 69 ). Theanticoagulant response also depends on the genetics ofvitamin K epoxide reductase (68 ). Moreover, gastroin-testinal bleeding from nonsteroidal antiinflammatorydrugs (70 ) and such side effects as hypoglycemiacaused by sulfonylureas (71 ) have also been attributedto CYP2C9 polymorphisms.

For the CYP2C19 gene (cytochrome P450, family2, subfamily C, polypeptide 19), the known null alleles(CYP2C19*2, *3, *4, *5, *6, *7, and *8) have noCYP2C19 enzyme activity (PM); the *2 allele is preva-lent in Caucasians. These null alleles are due to a splicedefect (*2), a premature stop codon (*3), or an alter-ation in CYP2C19 structure and/or stability (72 )(http://www.cypalleles.ki.se/). Recently, several newCYP2C19 alleles have been identified (*9–*25) in indi-viduals from different racial groups; however, whetherthese mutations produce significant alterations in en-zyme activity in vivo is not clear. CYP2C19*2 and *3are the most frequent variants. According to geno-typing and phenotyping results and in analogy toCYP2D6, the distribution of PMs shows wide inter-ethnic differences. In Caucasian Europeans, the meanfrequency of PM individuals is 3%, whereas PM fre-quencies as high as 23% have been identified in Asian/Oceanian populations (72, 73 ). Carriers of heterozy-gous variants constitute 32% of Caucasians, however

(74 ). A promoter variant of CYP2C19*17 has recentlybeen identified and shown to be associated with in-creased CYP2C19 activity in vivo (UM) with theCYP2C19 substrate omeprazole [a proton pump in-hibitor (75 )] and the antidepressant escitalopram(76 ). Differences in CYP2C19*17 allele frequency havebeen reported: 18% in both a Swedish and an Ethiopianpopulation (75 ), 25% in a German population (77 ),and 27% in a Polish population (78 ). A lower fre-quency (4%) has been reported for Chinese individuals(75 ). Given these genotype/phenotype relationships,there is a possibility that the CYP2C19 UM may play arole in tamoxifen metabolism and clinical outcome, aswe have reported for our breast cancer tamoxifen phar-macogenetic study (79 ).

With respect to CYP2B6 (cytochrome P450, family2, subfamily B, polypeptide 6), the most common vari-ant allele, *6, occurs at frequencies of 15%– 60% acrossdifferent populations (80 ). Genotyping of CYP2B6*6predicted increased plasma concentrations of efavirenzand nevirapine and efavirenz-related neurotoxicity inHIV-infected individuals (81, 82 ), and the results sug-gested reducing the dose by 35% in African patientswho were homozygous for CYP2B6*6 (83 ). These find-ings are in agreement with the lower activities ofCYP2B6*6 isoenzyme, which may be substrate depen-dent, however. At present, any contribution of CYP2B6variants to tamoxifen outcome is unknown.

The most important CYP isoenzyme subfamiliesinvolved in human drug metabolism are CYP3A4 and3A5, which participate in the metabolism of 40% of thedrugs that are most frequently prescribed (42 ). There islittle evidence for a relevant contribution of CYP3A4(cytochrome P450, family 2, subfamily A, polypeptide4) gene expression and enzyme function, although de-fective CYP3A4 mutants may account for toxicity invery rare cases (84 ). In contrast, genetic polymor-phisms define much of the variation in CYP3A5 (cyto-chrome P450, family 2, subfamily A, polypeptide 5)expression. The higher incidence of the inactiveCYP3A5*3 variant in Caucasians (85%–95%) vs Afri-can Americans (30%–50%) causes the lower CYP3A5protein level seen in Caucasians compared with AfricanAmericans (�30% vs 50%). CYP3A5*6 and *7 lack anyfunctional activity and occur solely in individuals ofAfrican origin. Apart from a clear effect on the immu-nosuppressant tacrolimus (85 ), the contribution of thepolymorphic CYP3A5 enzyme to CYP3A-mediatedmetabolism remains controversial. It is difficult to de-lineate the relative contributions of CYP3A4 andCYP3A5 because their protein structures, functions,and substrates are so similar. In fact, one of these en-zymes may functionally compensate for the lack of theother. Whether CYP3A4 and/or CYP3A5 variants con-tribute to tamoxifen outcome is unknown.



Tamoxifen Pharmacogenomics

The rationale underlying the tamoxifen pharmaco-genomic principle is that variant DNA sequences ofdrug-metabolizing enzymes that encode proteinswith reduced or absent enzyme function may be as-sociated with lower plasma concentrations of activetamoxifen metabolites, which could have an impacton the efficacy of tamoxifen treatment. About 30years ago, Jordan et al. characterized the first potentantiestrogen metabolite, 4-hydroxytamoxifen, andreported a 100-fold greater affinity for the ER thanthe parent drug (86 ). This metabolite was latershown to be 30- to 100-fold more potent than ta-moxifen in suppressing estrogen-dependent cellproliferation (86 – 89 ). Despite its potency as an an-tiestrogen, the contribution of this metabolite to theoverall clinical effect of tamoxifen has remained un-clear, because its plasma concentrations are rela-tively low compared with those of tamoxifen andother metabolites (86 ). Our knowledge of the linkbetween tamoxifen metabolism and treatment re-sponse rapidly expanded after the characterizationof endoxifen (16, 22 ), which, although it had beenidentified in the late 1980s, initially remained ob-scure with respect to its biological activity. Finally, aseries of laboratory studies for the characterizationof its pharmacology established that endoxifen has apotency equivalent to 4-hydroxytamoxifen in termsof its binding affinity for ERs (16 ), suppression ofestrogen-dependent proliferation of breast cancercells (16, 89, 90 ), and modulation of estrogen-mediated global gene expression (91 ). A detailed invitro analysis showed that endoxifen is formedmainly by 4-hydroxylation of the primary metabo-lite N-desmethyltamoxifen, with the CYP2D6 en-zyme catalyzing this rate-limiting step (15 ). Owing tothe dominant role of CYP2D6 in the formation of en-doxifen, variation in the CYP2D6 genotype and pheno-type is at the heart of tamoxifen pharmacogenetics. Thecurrently available evidence for this notion is based onfindings obtained at 2 levels of clinical investigations,which addressed (a) the association between the con-centrations of active tamoxifen metabolites either withCYP2D6 genotype or by clinical outcome, and (b) theassociation between CYP2D6 genotype and clinicaloutcome. The latter approach has shown that patientswith 2 functional CYP2D6 alleles benefited the mostfrom tamoxifen treatment. Further elucidation of therelationship between plasma concentrations of en-doxifen in vivo and clinical outcomes will requireadditional detailed investigations with large patientcohorts.

Effects of Tamoxifen Metabolite Concentrations

Prospective cohort studies of adjuvant tamoxifen treat-ment have shown wide interindividual variation in theformation of tamoxifen metabolites and substantial re-ductions in the steady-state plasma concentrations ofendoxifen during tamoxifen treatment in women car-rying CYP2D6 gene variants (8, 11, 22 ). Moreover,convincing evidence have shown that selective seroto-nin reuptake inhibitors (SSRIs) such as paroxetine andfluoxetine, which are known to be strong CYP2D6 in-hibitors, reduce plasma endoxifen concentrations. Inparticular, the phenocopy of a significant reduction inendoxifen plasma concentrations induced by SSRIswas observed in breast cancer patients homozygous forthe wild-type CYP2D6 genotype, whereas the concen-trations of other metabolites remained unaffected bythe CYP2D6 genotype/phenotype. Although the rela-tionship between CYP2D6 variants and plasma endox-ifen concentrations was first shown for patients withthe PM CYP2D6*4 genotype (11 ), a quantitative ap-proach that included PM, IM, and UM genotypes sub-stantiated this relationship (8 ); however, endoxifenconcentrations overlap across genotypes. It followsthat other factors may modify plasma endoxifenconcentrations.

A relationship between CYP2D6 variants andhigher concentrations of N-desmethyltamoxifen (i.e.,the endoxifen precursor) has been reported at the levelof chemoprevention. Significantly higher plasma con-centrations of N-desmethyltamoxifen were reportedfor mutation carriers after 1 year of tamoxifen therapy,indicating that the conversion to clinically active en-doxifen may be impaired (92 ).

A more recent study addressed the relationship be-tween CYP2D6 and SULT1A1 (sulfotransferase family,cytosolic, 1A, phenol-preferring, member 1) geno-types, including the effect of SULT1A1 copy numberon the pharmacokinetics of tamoxifen during steady-state treatment (32 ). Whereas both CYP2D6 andSULT1A1 genotypes influenced the pharmacokinet-ics of tamoxifen metabolites, SULT1A1 copy num-ber did not. Lower metabolic ratios with respect tothe formation of endoxifen and 4-hydroxytamoxifenbut higher metabolic ratios for the formation ofN-desmethyltamoxifen (endoxifen precursor) wereobserved in carriers of CYP2D6 variant genotypes, aresult consistent with a gene-dosage effect. In contrast,patients carrying CYP2D6 alleles with high predictedenzymatic activity showed higher metabolic ratios forboth active metabolites. Whether such metabolic ratiosare of clinical relevance remains to be determined.

Similarly, a study of a prospective cohort of Ko-rean patients with early or metastatic breast cancerfound an association between the IM CYP2D6*10 ho-

Review


mozygous variant and lower steady-state plasma con-centrations of 4-hydroxytamoxifen and endoxifen(17 ), and a Chinese study found that patients homozy-gous for CYP2D6*10 had lower serum concentrationsof 4-hydroxytamoxifen (93 ). The high prevalence ofthe CYP2D6*10 allele in East Asia, together with the IMassociation of impaired formation of an active metabolite,confirms the CYP2D6 PM findings in Caucasians.

Clinical Outcome of Tamoxifen Therapy andPrediction

The first evidence linking CYP2D6 variants with treat-ment response was obtained from a prospective ran-domized phase III trial of postmenopausal womenwith ER-positive breast cancer (North Central CancerTreatment Group adjuvant breast cancer trial) for theinvestigation of the effect of adding the androgen flu-oxymesterone for 1 year to the standard regimen of 5years of adjuvant tamoxifen. The pharmacogenetic in-vestigation of patients from the tamoxifen-only armshowed that after a median follow-up of 11.4 years, theCYP2D6*4 variant allele was an independent predictorof a higher risk of relapse and a lower incidence of hotflashes (94 ). A follow-up study found that in additionto CYP2D6 genetics, the phenocopying due to thecoprescription of CYP2D6 inhibitors (SSRIs) was anindependent predictor of breast cancer outcome in

postmenopausal women taking tamoxifen (95 ). Re-cently, a robust association between CYP2D6 genotypeand treatment outcome was obtained from a nonran-domized retrospective cohort of ER-positive post-menopausal breast cancer patients undergoing adju-vant tamoxifen therapy (79 ). At a median follow-up of71 months, carriers of PM and IM genotypes (i.e., car-riers of CYP2D6*4, *5, *10, and *41 alleles) had signif-icantly more breast cancer recurrences, shorter relapse-free times, and worse event-free survival than carriersof functional alleles (Fig. 2). This association was notobserved in postmenopausal ER-positive patients nottreated with tamoxifen. Interestingly, the UMCYP2C19*17 variant also had a favorable effect on ta-moxifen treatment outcome. Patients with the ho-mozygous *17 genotype had significantly fewer breastcancer recurrences, longer relapse-free times, and bet-ter event-free survival than non-*17 carriers. Overall,this study suggested that genotyping for CYP2D6*4, *5,*10, and *41 could identify patients who would derivelittle benefit from adjuvant tamoxifen therapy. Al-though the CYP2D6 EM phenotype will identify thepatients likely to benefit from tamoxifen, accountingfor about 50% of all patients, the benefit will be maxi-mal for individuals with the combination of fully func-tional CYP2D6 alleles and the CYP2C19 UM. The lat-ter will apply to one third of all patients, indicating thatthe tamoxifen pharmacogenetics issue will be relevant

Fig. 2. Kaplan–Meier probabilities of relapse-free time (RFT) of breast cancer patients for CYP2D6-metabolizerphenotypes predicted from genotypes.

(A), Patients treated with adjuvant tamoxifen (TAM). EMs had a significantly more favorable RFT than patients with impairedphenotypes (PMs or IMs). (B), Patients without TAM showed no differences with respect to a relationship between the CYP2D6-predicted phenotype and RFT [Schroth et al. (79)]. hetEM, heterozygous EM. Originally published in Schroth, W et al.: J Clin Oncol 25(33), 2007: 5187–93. Reprinted with permission. © 2008 American Society of Clinical Oncology. All rights reserved.



for a substantial fraction of breast cancer patients re-ceiving endocrine treatment.

Clinical studies from Korea, China, and Japan alsohave linked poor clinical outcome with CYP2D6 genet-ics. As expected for populations with a high prevalenceof the IM CYP2D6*10 allele, the *10 homozygote geno-type was associated with a poor clinical outcome in aKorean cohort of metastatic breast cancer patients,whereas the *10 heterozygote and wild-type homozy-gote genotypes were not (17 ). Likewise, patients fromChina who were homozygous for the CYP2D6*10 allele(93 ) showed an association with unfavorable disease-free survival. The latter result was substantiatedthrough comparison with a control patient groupwithout tamoxifen treatment, in which no associationbetween clinical outcome and the CYP2D6*10 variantwas observed. Moreover, patients homozygous forCYP2D6*10 from a Japanese breast cancer cohort thatunderwent adjuvant tamoxifen monotherapy showeda significantly higher incidence of recurrence within 10years of follow-up, compared with patients with wild-type CYP2D6 (96 ). Although some of the sample sizeswere low in the Asian studies demonstrating the geno-type– efficacy correlation, the findings of the clinicalimplications of CYP2D6 genotypes predictive for ta-moxifen efficacy are in line with the findings of others.

On the other hand, a study from the US reportedno association between CYP2D6 genetics and tamox-ifen outcome (97 ), and contradictory results for thisrelationship were reported in a study from Sweden,which found the CYP2D6*4 variant to be associatedwith a better clinical outcome in tamoxifen-treated pa-tients (98 ). An extended study showed favorabledisease-free survival in CYP2D6*4 carriers comparedwith patients homozygous or heterozygous for thefunctional CYP2D6 allele (99 ).

The issue of the role of CYP2D6 in tamoxifen ther-apy for breast cancer has also been addressed within thecontext of breast cancer prevention. For example, datafrom the Italian Tamoxifen Trial suggest that womenwith a CYP2D6*4/*4 genotype may be less likely to ben-efit from tamoxifen as a chemopreventive agent. Thisfinding supports the notion of CYP2D6 playing an im-portant role in tamoxifen’s metabolic activation andefficacy (100 ). Moreover, the “a priori” hypothesis thathot flashes may be an independent predictor of tamox-ifen efficacy has been addressed in the Women’sHealthy Eating and Living randomized trial (101 ). Ofthe 864 patients taking tamoxifen, 674 (78%) reportedhot flashes, and 12.9% of these patients had experi-enced recurrent breast cancer after 7.3 years of follow-up, whereas 21% of the patients who did not have hotflashes had recurrent breast cancer during this period.Because hot flashes were a stronger predictor of a breastcancer–specific outcome than age, hormone receptor

status, or tumor stage at diagnosis, the authors sug-gested an association between side effects, tamoxifenmetabolism, and efficacy. Finally, a small study of fa-milial breast cancer patients who were carriers of eitherBRCA1 (breast cancer 1, early onset) or BRCA2 (breastcancer 2, early onset) mutations and treated with ta-moxifen suggested a relationship between CYP2D6 PMstatus and a worse survival in familial breast cancer(102 ); however, because of the small numbers of pa-tients as well as the inclusion of ER-positive and ER-negative patients in this investigation, clarificationprovided by further studies will be needed to distin-guish a pharmacogenetic effect from a poor prognosticeffect in carriers of these BRCA mutations.

Given the current treatment practice of long-termestrogen deprivation in ER-positive postmenopausalbreast cancer patients with the use of AIs as a validoption, the question of the impact of pharmacogeneticvariation on the optimal choice for adjuvant endocrinetherapy has been addressed in a modeling analysis(103 ). A Markov model was created to examinewhether the optimal treatment strategy for patientswith the wild-type CYP2D6 gene differs from that forcarriers of the CYP2D6*4 mutation. The study usedpatients from the BIG1–98 trial, information from thistrial on relapse risk, and the corresponding genotypedata of Goetz et al. (94 ). Under the assumption that AImetabolism is independent from CYP2D6, the modelsuggests that the 5-year benefit of adjuvant tamoxifentherapy may exceed even that of up-front AI treatmentin postmenopausal CYP2D6 EM patients.

Conclusion: Clinical Relevance of CYP2D6 in BreastCancer

Strong mechanistic, pharmacologic, and clinical evi-dence, as well as modeling data, now indicate that ta-moxifen efficacy and clinical outcome depend onCYP2D6 metabolism controlled by CYP2D6 enzymepolymorphisms and on pharmacologic interactions.Data from international studies have consistently dem-onstrated that plasma concentrations of active tamox-ifen metabolites are linked with genetically determinedCYP2D6 metabolizer status, phenocopying by strongCYP2D6 inhibitors, and clinical outcome. The fewconflicting data may be explained by variation in thestudies with respect to patient-inclusion criteria, ta-moxifen doses, length of treatment, additional chemo-therapy regimens, or a lack of consistent ER testing.Importantly, most authors agree that CYP2D6 genevariants, as well as inhibition of CYP2D6 by prescribedcomedications such as SSRIs, may decrease tamoxifenmetabolism and thus negatively affect tamoxifen effi-cacy and treatment outcome.

Review


There are a number of potential clinical conse-quences from these emerging data on CYP2D6 and theoutcomes of tamoxifen treatment. First, potent SSRIssuch as paroxetine or fluoxetine should not be used torelieve hot flashes in breast cancer patients receiving ta-moxifen. Although SSRIs are one of the few evidence-based therapy options for menopausal vasomotor symp-toms (104), convincing data now indicate that thesedrugs may compromise tamoxifen efficacy via a pheno-copying effect due to interference with CYP2D6-dependent tamoxifen metabolism. Yet, differences in theplasma concentrations of tamoxifen metabolites havebeen observed, depending on the strength of the CYP2D6inhibitor (11, 105). If treatment of hot flashes is indicated,an SSRI such as citalopram or escitalopram or a selectivenorepinephrine reuptake inhibitor such as venlafaxineshould be used, because these compounds have shown noappreciable inhibition of CYP2D6.

Second, the relationship between CYP2D6 geno-type, phenotype, and treatment outcome points to apossible benefit of up-front CYP2D6 genotyping priorto a decision on an adjuvant endocrine treatment. Acomprehensive robust, standardized, and quality-controlled CYP2D6-genotyping assay will have to testfor genetic variants that could affect tamoxifen metab-olism. According to the data of Goetz et al. (94 ) andSchroth et al. (79 ), such assays should include testingfor common PM alleles (CYP2D6*3, *4, and *5) and forIM alleles, depending on the individual’s ethnic origin.Of note, *41 is the most frequent IM allele in Europe-ans, *17 is the principal IM allele in Africans, and *10dominates in Asians (*9 should also be considered)(59 ). Other areas of interest with respect to clinicalapplication are the measurement of plasma endoxifenconcentrations as a surrogate of CYP2D6 phenotype.

Given alternative treatment options (i.e., tamox-ifen vs AI), and considering the available scientific andclinical evidence, an individualized approach for en-docrine treatment of postmenopausal breast cancerpatients is desirable. One may speculate that tamoxi-fen alone is adequate for CYP2D6 EMs and EM carri-ers, whereas postmenopausal patients with variantCYP2D6 alleles may fare better with up-front AI ther-apy. Although this approach may be regarded asstraightforward for PM patients, the best treatmentmay be less clear for IM patients. IM is a common phe-notype among many ethnic groups, including Cauca-sians, African Americans, and Asians, so data on link-ing IM genotypes with therapeutic threshold andefficacy are in demand to adequately address the clini-cally important question of tamoxifen dose adjust-ment. Similarly, any impact of UM phenotypes onmetabolite concentrations, treatment efficacy, and tox-icity that have potential implications for dosing re-quires further investigations. Formal recommenda-

tions on the integration of CYP2D6 genotypes intotreatment decisions still must await validation of thesegenotypes in larger retrospective studies, as is beingattempted by the International Tamoxifen Pharmaco-genetics Consortium (http://www.pharmgkb.org/do/serve?objId�63&objCls�Project), or prospectiveclinical trials. Thus far, no study has addressed thequestion of whether genetically predisposed differ-ences in 4-hydroxytamoxifen and endoxifen concen-trations are associated with treatment response or dis-ease progression and with side effects such as hotflashes, including phenocopying effects; therefore,therapeutic drug monitoring as a useful surrogate iscurrently not available in the case of tamoxifen.Whether determination of the CYP2D6 genotype willbecome a diagnostic tool for selecting the appropriateadjuvant endocrine therapy for ER-positive postmeno-pausal breast cancer patients awaits validation in pro-spective clinical trials that randomize tamoxifen vs AItreatment according to CYP2D6 genotypes. Such pro-spective clinical trials are currently being planned.

Other open questions may address the clinical rel-evance of other drug-metabolizing enzymes and muta-tions, as well as ethnic variation, in the prevalence oftheir treatment outcome–relevant genotypes. Finally,there is the possibility that pharmacokinetic genes willonly partly explain the pharmacogenomics of tamox-ifen. It will therefore be important to also explore thecontribution of pharmacodynamic genes in evaluatingantiestrogen resistance as a feature of the tumor celland in addressing the role of genes associated withestrogen-mediated cell proliferation. Within this con-text, it will be interesting to learn whether genes encod-ing the ER, its coactivators, or its corepressors (106 ), aswell as antiestrogen resistance genes (107 ) and theirvariants, will affect the response to tamoxifen. Suchresults may increase the overall potential of tamoxifenpharmacogenomics.

To this end, it is important to appreciate that mostcancer therapies in current use have been establishedempirically. The recent progress in our understandingof the pharmacology and pharmacogenetics of tamox-ifen, however, holds promise for the improvement oftreatments through personalized medicine. Becausethe genome-based approach uses CYP2D6 genotypingto predict a patient’s metabolizer phenotype, ethicalissues need to be sufficiently addressed. In the light ofacceptable alternatives, an informed choice about ad-juvant endocrine treatment and, most importantly,avoiding a therapy that may lack efficacy must be ofprime interest. It will therefore be important to makepatients and their caregivers aware of these issues andto initiate discussions with regulatory authorities.



Author Contributions: All authors confirmed they have contributed tothe intellectual content of this paper and have met the following 3 re-quirements: (a) significant contributions to the conception and design,acquisition of data, or analysis and interpretation of data; (b) draftingor revising the article for intellectual content; and (c) final approval ofthe published article.

Authors’ Disclosures of Potential Conflicts of Interest: Uponmanuscript submission, all authors completed the Disclosures of Poten-tial Conflict of Interest form. Potential conflicts of interest:

Employment or Leadership: None declared.Consultant or Advisory Role: None declared.Stock Ownership: None declared.Honoraria: None declared.Research Funding: Robert Bosch Foundation, Stuttgart, Germany,and Bundesministerium fur Bildung und Forschung Grant No.01ZP0502.Expert Testimony: None declared.

Role of Sponsor: The funding organizations played no role in thedesign of study, choice of enrolled patients, review and interpretationof data, or preparation or approval of manuscript.

References

1. Early Breast Cancer Trialists’ CollaborativeGroup (EBCTCG). Effects of chemotherapy andhormonal therapy for early breast cancer onrecurrence and 15-year survival: an overview ofthe randomised trials. Lancet 2005;365:1687–717.

2. Goldhirsch A, Wood WC, Gelber RD, Coates AS,Thurlimann B, Senn HJ. Progress and promise:highlights of the international expert consensuson the primary therapy of early breast cancer2007. Ann Oncol 2007;18:1133–44.

3. Fisher B, Costantino JP, Wickerham DL, CecchiniRS, Cronin WM, Robidoux A, et al. Tamoxifenfor the prevention of breast cancer: current sta-tus of the National Surgical Adjuvant Breast andBowel Project P-1 study. J Natl Cancer Inst2005;97:1652–62.

4. Fentiman IS, Fourquet A, Hortobagyi GN. Malebreast cancer. Lancet 2006;367:595–604.

5. Boocock DJ, Brown K, Gibbs AH, Sanchez E,Turteltaub KW, White IN. Identification of hu-man CYP forms involved in the activation oftamoxifen and irreversible binding to DNA. Car-cinogenesis 2002;23:1897–901.

6. Coller JK, Krebsfaenger N, Klein K, Wolbold R,Nussler A, Neuhaus P, et al. Large interindi-vidual variability in the in vitro formation oftamoxifen metabolites related to the develop-ment of genotoxicity. Br J Clin Pharmacol 2004;57:105–11.

7. Crewe HK, Notley LM, Wunsch RM, Lennard MS,Gillam EM. Metabolism of tamoxifen by recom-binant human cytochrome P450 enzymes:formation of the 4-hydroxy, 4�-hydroxy andN-desmethyl metabolites and isomerization oftrans-4-hydroxytamoxifen. Drug Metab Dispos2002;30:869–74.

8. Borges S, Desta Z, Li L, Skaar TC, Ward BA,Nguyen A, et al. Quantitative effect of CYP2D6genotype and inhibitors on tamoxifenmetabolism: implication for optimization ofbreast cancer treatment. Clin Pharmacol Ther2006;80:61–74.

9. Dowsett M, Cuzick J, Howell A, Jackson I, andthe ATAC Trialists’ Group. Pharmacokinetics ofanastrozole and tamoxifen alone, and in com-bination, during adjuvant endocrine therapy forearly breast cancer in postmenopausal women:a sub-protocol of the ‘Arimidex™ and Tamox-ifen Alone or in Combination’ (ATAC) trial. Br JCancer 2001;85:317–24.

10. Hutson PR, Love RR, Havighurst TC, Rogers E,Cleary JF. Effect of exemestane on tamoxifenpharmacokinetics in postmenopausal women

treated for breast cancer. Clin Cancer Res 2005;11:8722–7.

11. Jin Y, Desta Z, Stearns V, Ward B, Ho H, Lee KH,et al. CYP2D6 genotype, antidepressant use,and tamoxifen metabolism during adjuvantbreast cancer treatment. J Natl Cancer Inst2005;97:30–9.

12. Lee KH, Ward BA, Desta Z, Flockhart DA, JonesDR. Quantification of tamoxifen and three me-tabolites in plasma by high-performance liquidchromatography with fluorescence detection: ap-plication to a clinical trial. J Chromatogr B AnalytTechnol Biomed Life Sci 2003;791:245–53.

13. MacCallum J, Cummings J, Dixon JM, Miller WR.Concentrations of tamoxifen and its major me-tabolites in hormone responsive and resistantbreast tumours. Br J Cancer 2000;82:1629–35.

14. Sheth HR, Lord G, Tkaczuk K, Danton M, LewisLM, Langenberg P, et al. Aging may be associ-ated with concentrations of tamoxifen and itsmetabolites in breast cancer patients. J WomensHealth (Larchmt) 2003;12:799–808.

15. Desta Z, Ward BA, Soukhova NV, Flockhart DA.Comprehensive evaluation of tamoxifen sequen-tial biotransformation by the human cytochromeP450 system in vitro: prominent roles for CYP3Aand CYP2D6. J Pharmacol Exp Ther 2004;310:1062–75.

16. Johnson MD, Zuo H, Lee KH, Trebley JP, Rae JM,Weatherman RV, et al. Pharmacological charac-terization of 4-hydroxy-N-desmethyl tamoxifen,a novel active metabolite of tamoxifen. BreastCancer Res Treat 2004;85:151–9.

17. Lim HS, Ju LH, Seok LK, Sook LE, Jang IJ, RoJ. Clinical implications of CYP2D6 genotypespredictive of tamoxifen pharmacokinetics inmetastatic breast cancer. J Clin Oncol 2007;25:3837–45.

18. Coller JK, Krebsfaenger N, Klein K, Endrizzi K,Wolbold R, Lang T, et al. The influence ofCYP2B6, CYP2C9 and CYP2D6 genotypes on theformation of the potent antioestrogen Z-4-hydroxy-tamoxifen in human liver. Br J ClinPharmacol 2002;54:157–67.

19. Crewe HK, Ellis SW, Lennard MS, Tucker GT.Variable contribution of cytochromes P4502D6,2C9 and 3A4 to the 4-hydroxylation of tamox-ifen by human liver microsomes. Biochem Phar-macol 1997;53:171–8.

20. Dehal SS, Kupfer D. CYP2D6 catalyzes tamoxifen4-hydroxylation in human liver. Cancer Res1997;57:3402–6.

21. Mani C, Gelboin HV, Park SS, Pearce R, Parkin-son A, Kupfer D. Metabolism of the antimam-

mary cancer antiestrogenic agent tamoxifen. I.Cytochrome P-450-catalyzed N-demethylationand 4-hydroxylation. Drug Metab Dispos 1993;21:645–56.

22. Stearns V, Johnson MD, Rae JM, Morocho A,Novielli A, Bhargava P, et al. Active tamoxifenmetabolite plasma concentrations after coad-ministration of tamoxifen and the selective se-rotonin reuptake inhibitor paroxetine. J NatlCancer Inst 2003;95:1758–64.

23. Kim SY, Suzuki N, Santosh Laxmi YR, Rieger R,Shibutani S. Alpha-hydroxylation of tamoxifenand toremifene by human and rat cytochromeP450 3A subfamily enzymes. Chem Res Toxicol2003;16:1138–44.

24. Notley LM, Crewe KH, Taylor PJ, Lennard MS,Gillam EM. Characterization of the human cy-tochrome P450 forms involved in metabolism oftamoxifen to its �-hydroxy and �,4-dihydroxyderivatives. Chem Res Toxicol 2005;18:1611–8.

25. Katzenellenbogen JA, Carlson KE, Katzenellenbo-gen BS. Facile geometric isomerization of phenolicnon-steroidal estrogens and antiestrogens: limita-tions to the interpretation of experiments charac-terizing the activity of individual isomers. J SteroidBiochem 1985;22:589–96.

26. Osborne CK, Wiebe VJ, McGuire WL, Ciocca DR,DeGregorio MW. Tamoxifen and the isomers of4-hydroxytamoxifen in tamoxifen-resistant tu-mors from breast cancer patients. J Clin Oncol1992;10:304–10.

27. Dehal SS, Brodie AMH, Kupfer D. The aromataseinactivator 4-hydroxyandrostenedione (4-OH-A)inhibits tamoxifen metabolism by rat hepaticcytochrome P-450 3A: potential for drug-druginteraction of tamoxifen and 4-OH-A in com-bined anti-breast cancer therapy. Drug MetabDispos 1999;27:389–94.

28. Dehal SS, Kupfer D. Evidence that the catechol3,4-dihydroxytamoxifen is a proximate interme-diate to the reactive species binding covalentlyto proteins. Cancer Res 1995;56:1283–90.

29. Liu X, Pisha E, Tonetti DA, Yao D, Li Y, Yao J, etal. Antiestrogenic and DNA damaging effectsinduced by tamoxifen and toremifene metabo-lites. Chem Res Toxicol 2003;16:832–7.

30. Hodgson E, Rose RL, Cao Y, Dehal SS, Kupfer D.Flavin-containing monooxygenase isoform spec-ificity for the N-oxidation of tamoxifen deter-mined by product measurement and NADPHoxidation. J Biochem Mol Toxicol 2000;14:118–20.

31. Parte P, Kupfer D. Oxidation of tamoxifen byhuman flavin-containing monooxygenase (FMO)

Review


1 and FMO3 to tamoxifen-N-oxide and its novelreduction back to tamoxifen by human cyto-chromes P450 and hemoglobin. Drug MetabDispos 2005;33:1446–52.

32. Gjerde J, Hauglid M, Breilid H, Lundgren S,Varhaug JE, Kisanga ER, et al. Effects of CYP2D6and SULT1A1 genotypes including SULT1A1gene copy number on tamoxifen metabolism.Ann Oncol 2008;19:56–61.

33. Nishiyama T, Ogura K, Nakano H, Ohnuma T,Kaku T, Hiratsuka A, et al. Reverse geometricalselectivity in glucuronidation and sulfation ofcis- and trans-4-hydroxytamoxifens by humanliver UDP-glucuronosyltransferases and sulfo-transferases. Biochem Pharmacol 2002;63:1817–30.

34. Ogura K, Ishikawa Y, Kaku T, Nishiyama T,Ohnuma T, Muro K, Hiratsuka A. Quaternaryammonium-linked glucuronidation of trans-4-hydroxytamoxifen, an active metabolite of ta-moxifen, by human liver microsomes and UDP-glucuronosyltransferase 1A4. Biochem Pharmacol2006;71:1358–69.

35. Sun D, Sharma AK, Dellinger RW, Blevins-Primeau AS, Balliet RM, Chen G, et al. Glucu-ronidation of active tamoxifen metabolites bythe human UDP glucuronosyltransferases. DrugMetab Dispos 2007;35:2006–14.

36. Kaku T, Ogura K, Nishiyama T, Ohnuma T, MuroK, Hiratsuka A. Quaternary ammonium-linkedglucuronidation of tamoxifen by human livermicrosomes and UDP-glucuronosyltransferase1A4. Biochem Pharmacol 2004;67:2093–102.

37. Sun D, Chen G, Dellinger RW, Duncan K, FangJL, Lazarus P. Characterization of tamoxifen and4-hydroxytamoxifen glucuronidation by humanUGT1A4 variants. Breast Cancer Res 2006;8:R50.

38. Falany JL, Pilloff DE, Leyh TS, Falany CN. Sulfa-tion of raloxifene and 4-hydroxytamoxifen byhuman cytosolic sulfotransferases. Drug MetabDispos 2006;34:361–8.

39. Apak TI, Duffel MW. Interactions of the stereo-isomers of alpha-hydroxytamoxifen with humanhydroxysteroid sulfotransferase SULT2A1 andrat hydroxysteroid sulfotransferase STa. DrugMetab Dispos 2004;32:1501–8.

40. Kim SY, Laxmi YR, Suzuki N, Ogura K, Watabe T,Duffel MW, Shibutani S. Formation oftamoxifen-DNA adducts via O-sulfonation, notO-acetylation, of alpha-hydroxytamoxifen in ratand human livers. Drug Metab Dispos 2005;33:1673–8.

41. Osborne MR, Hewer A, Phillips DH. Resolutionof alpha-hydroxytamoxifen; R-isomer formsmore DNA adducts in rat liver cells. Chem ResToxicol 2001;14:888–93.

42. Zanger UM, Turpeinen M, Klein K, Schwab M.Functional pharmacogenetics/genomics of hu-man cytochromes P450 involved in drug bio-transformation. Anal Bioanal Chem 2008;392:1093–108.

43. Zanger UM. The CYP2D subfamily. In: IoannidesC, ed. Cytochromes P450: role in the metabo-lism and toxicity of drugs and other xenobiotics.London: Royal Chemical Society; 2008. p 241–75.

44. Eichelbaum M, Spannbrucker N, Dengler HJ.Proceedings: N-oxidation of sparteine in manand its interindividual differences. Naunyn

Schmiedebergs Arch Pharmacol 1975;287(Suppl):R94.

45. Mahgoub A, Idle JR, Dring LG, Lancaster R,Smith RL. Polymorphic hydroxylation of debriso-quine in man. Lancet 1977;2:584–6.

46. Griese EU, Zanger UM, Brudermanns U, GaedigkA, Mikus G, Morike K, et al. Assessment of thepredictive power of genotypes for the in-vivocatalytic function of CYP2D6 in a German pop-ulation. Pharmacogenetics 1998;8:15–26.

47. Raimundo S, Toscano C, Klein K, Fischer J,Griese EU, Eichelbaum M, et al. A novel intronicmutation, 2988G�A, with high predictivity forimpaired function of cytochrome P450 2D6 inwhite subjects. Clin Pharmacol Ther 2004;76:128–38.

48. Sachse C, Brockmoller J, Bauer S, Roots I. Cyto-chrome P450 2D6 variants in a Caucasianpopulation: allele frequencies and phenotypicconsequences. Am J Hum Genet 1997;60:284–95.

49. Marez D, Legrand M, Sabbagh N, Guidice JM,Spire C, Lafitte JJ, et al. Polymorphism of thecytochrome P450 CYP2D6 gene in a Europeanpopulation: characterization of 48 mutationsand 53 alleles, their frequencies and evolution.Pharmacogenetics 1997;7:193–202.

50. Raimundo S, Fischer J, Eichelbaum M, GrieseEU, Schwab M, Zanger UM. Elucidation of thegenetic basis of the common ‘intermediate me-tabolizer’ phenotype for drug oxidation byCYP2D6. Pharmacogenetics 2000;10:577–81.

51. Toscano C, Klein K, Blievernicht J, Schaeffeler E,Saussele T, Raimundo S, et al. Impaired expres-sion of CYP2D6 in intermediate metabolizerscarrying the *41 allele caused by the intronicSNP 2988G�A: evidence for modulation ofsplicing events. Pharmacogenet Genomics 2006;16:755–66.

52. Zanger UM, Fischer J, Raimundo S, Stuven T,Evert BO, Schwab M, Eichelbaum M. Compre-hensive analysis of the genetic factors determin-ing expression and function of hepatic CYP2D6.Pharmacogenetics 2001;11:573–85.

53. Bertilsson L, Dahl ML, Sjoqvist F, Aberg-WistedtA, Humble M, Johansson I, et al. Molecular basisfor rational megaprescribing in ultrarapid hy-droxylators of debrisoquine. Lancet 1993;341:63.

54. Dalen P, Dahl ML, Bernal Ruiz ML, Nordin J,Bertilsson L. 10-Hydroxylation of nortriptyline inwhite persons with 0, 1, 2, 3, and 13 functionalCYP2D6 genes. Clin Pharmacol Ther 1998;63:444–52.

55. Lovlie R, Daly AK, Matre GE, Molven A, SteenVM. Polymorphisms in CYP2D6 duplication-negative individuals with the ultrarapid metabo-lizer phenotype: a role for the CYP2D6*35 allelein ultrarapid metabolism? Pharmacogenetics2001;11:45–55.

56. Algeciras-Schimnich A, O’Kane DJ, Snozek CL.Pharmacogenomics of tamoxifen and irinotecantherapies. Clin Lab Med 2008;28:553–67.

57. Aklillu E, Persson I, Bertilsson L, Johansson I,Rodrigues F, Ingelman-Sundberg M. Frequentdistribution of ultrarapid metabolizers of de-brisoquine in an Ethiopian population carryingduplicated and multiduplicated functionalCYP2D6 alleles. J Pharmacol Exp Ther 1996;278:441–6.

58. McLellan RA, Oscarson M, Seidegard J, EvansDA, Ingelman-Sundberg M. Frequent occurrenceof CYP2D6 gene duplication in Saudi Arabians.Pharmacogenetics 1997;7:187–91.

59. Sistonen J, Sajantila A, Lao O, Corander J, Bar-bujani G, Fuselli S. CYP2D6 worldwide geneticvariation shows high frequency of altered activ-ity variants and no continental structure. Phar-macogenet Genomics 2007;17:93–101.

60. Ingelman-Sundberg M. Genetic polymorphismsof cytochrome P450 2D6 (CYP2D6): clinical con-sequences, evolutionary aspects and functionaldiversity. Pharmacogenomics J 2005;5:6–13.

61. Bradford LD. CYP2D6 allele frequency in Euro-pean Caucasians, Asians, Africans and their de-scendants. Pharmacogenomics 2002;3:229–43.

62. Gasche Y, Daali Y, Fathi M, Chiappe A, Cottini S,Dayer P, Desmeules J. Codeine intoxication as-sociated with ultrarapid CYP2D6 metabolism.N Engl J Med 2004;351:2827–31.

63. Koren G, Cairns J, Chitayat D, Gaedigk A, LeederSJ. Pharmacogenetics of morphine poisoning ina breastfed neonate of a codeine-prescribedmother. Lancet 2006;368:704.

64. Lee CR, Goldstein JA, Pieper JA. CytochromeP450 2C9 polymorphisms: a comprehensive re-view of the in-vitro and human data. Pharma-cogenetics 2002;12:251–63.

65. King BP, Khan TI, Aithal GP, Kamali F, DalyAK. Upstream and coding region CYP2C9polymorphisms: correlation with warfarin doseand metabolism. Pharmacogenetics 2004;14:813–22.

66. Garcia-Martin E, Martinez C, Ladero JM, Agun-dez JA. Interethnic and intraethnic variability ofCYP2C8 and CYP2C9 polymorphisms in healthyindividuals. Mol Diagn Ther 2006;10:29–40.

67. Xie HG, Kim RB, Wood AJ, Stein CM. Molecularbasis of ethnic differences in drug dispositionand response. Annu Rev Pharmacol Toxicol2001;41:815–50.

68. Flockhart DA, O’Kane D, Williams MS, WatsonMS, Flockhart DA, Gage B, et al. Pharmacoge-netic testing of CYP2C9 and VKORC1 alleles forwarfarin. Genet Med 2008;10:139–50.

69. Limdi NA, Veenstra DL. Warfarin pharmacoge-netics. Pharmacotherapy 2008;28:1084–97.

70. Pilotto A, Seripa D, Franceschi M, Scarcelli C, Co-laizzo D, Grandone E, et al. Genetic susceptibilityto nonsteroidal anti-inflammatory drug-relatedgastroduodenal bleeding: role of cytochrome P4502C9 polymorphisms. Gastroenterology 2007;133:465–71.

71. Holstein A, Plaschke A, Ptak M, Egberts EH,El-Din J, Brockmoller J, Kirchheiner J. Associa-tion between CYP2C9 slow metabolizer geno-types and severe hypoglycaemia on medicationwith sulphonylurea hypoglycaemic agents. Br JClin Pharmacol 2005;60:103–6.

72. Desta Z, Zhao X, Shin JG, Flockhart DA. Clinicalsignificance of the cytochrome P450 2C19 ge-netic polymorphism. Clin Pharmacokinet 2002;41:913–58.

73. Xie HG, Stein CM, Kim RB, Wilkinson GR, Flock-hart DA, Wood AJ. Allelic, genotypic and phe-notypic distributions of S-mephenytoin 4�-hydroxylase (CYP2C19) in healthy Caucasianpopulations of European descent throughoutthe world. Pharmacogenetics 1999;9:539–49.

74. Wedlund PJ. The CYP2C19 enzyme polymor-



phism. Pharmacology 2000;61:174–83.75. Sim SC, Risinger C, Dahl ML, Aklillu E, Chris-

tensen M, Bertilsson L, Ingelman-Sundberg M. Acommon novel CYP2C19 gene variant causesultrarapid drug metabolism relevant for the drugresponse to proton pump inhibitors and antide-pressants. Clin Pharmacol Ther 2006;79:103–13.

76. Rudberg I, Mohebi B, Hermann M, Refsum H,Molden E. Impact of the ultrarapid CYP2C19*17allele on serum concentration of escitalopram inpsychiatric patients. Clin Pharmacol Ther 2008;83:322–7.

77. Justenhoven C, Hamann U, Pierl CB, Baisch C,Harth V, Rabstein S, et al. CYP2C19*17 is as-sociated with decreased breast cancer risk.Breast Cancer Res Treat 2008;115:391–6.

78. Kurzawski M, Gawronska-Szklarz B, Wrzes-niewska J, Siuda A, Starzynska T, Drozdzik M.Effect of CYP2C19*17 gene variant on Helico-bacter pylori eradication in peptic ulcer patients.Eur J Clin Pharmacol 2006;62:877–80.

79. Schroth W, Antoniadou L, Fritz P, Schwab M,Muerdter T, Zanger UM, et al. Breast cancertreatment outcome with adjuvant tamoxifen rel-ative to patient CYP2D6 and CYP2C19 geno-types. J Clin Oncol 2007;25:5187–93.

80. Zanger UM, Klein K, Saussele T, Blievernicht J,Hofmann MH, Schwab M. Polymorphic CYP2B6:molecular mechanisms and emerging clinicalsignificance. Pharmacogenomics 2007;8:743–59.

81. Rotger M, Colombo S, Furrer H, Bleiber G, BuclinT, Lee BL, et al. Influence of CYP2B6 polymor-phism on plasma and intracellular concentra-tions and toxicity of efavirenz and nevirapine inHIV-infected patients. Pharmacogenet Genom-ics 2005;15:1–5.

82. Rotger M, Tegude H, Colombo S, Cavassini M,Furrer H, Decosterd L, et al. Predictive value ofknown and novel alleles of CYP2B6 for efavirenzplasma concentrations in HIV-infected individu-als. Clin Pharmacol Ther 2007;81:557–66.

83. Nyakutira C, Roshammar D, Chigutsa E, ChonziP, Ashton M, Nhachi C, Masimirembwa C. Highprevalence of the CYP2B6 516G3T(*6) variantand effect on the population pharmacokineticsof efavirenz in HIV/AIDS outpatients in Zimba-bwe. Eur J Clin Pharmacol 2008;64:357–65.

84. Westlind-Johnsson A, Hermann R, HuennemeyerA, Hauns B, Lahu G, Nassr N, et al. Identificationand characterization of CYP3A4*20, a novelrare CYP3A4 allele without functional activity.Clin Pharmacol Ther 2006;79:339–49.

85. Anglicheau D, Legendre C, Beaune P, Thervet E.Cytochrome P450 3A polymorphisms and immu-nosuppressive drugs: an update. Pharmacog-enomics 2007;8:835–49.

86. Jordan VC. Metabolites of tamoxifen in animalsand man: identification, pharmacology, and sig-

nificance. Breast Cancer Res Treat 1982;2:123–38.

87. Coezy E, Borgna JL, Rochefort H. Tamoxifen andmetabolites in MCF7 cells: correlation betweenbinding to estrogen receptor and inhibition ofcell growth. Cancer Res 1982;42:317–23.

88. Robertson DW, Katzenellenbogen JA, Long DJ,Rorke EA, Katzenellenbogen BS. Tamoxifen an-tiestrogens. A comparison of the activity, phar-macokinetics, and metabolic activation of thecis and trans isomers of tamoxifen. J SteroidBiochem 1982;16:1–13.

89. Buck MB, Coller JK, Murdter TE, Eichelbaum M,Knabbe C. TGF�2 and T�RII are valid molecularbiomarkers for the antiproliferative effects oftamoxifen and tamoxifen metabolites in breastcancer cells. Breast Cancer Res Treat 2008;107:15–24.

90. Lim YC, Desta Z, Flockhart DA, Skaar TC. En-doxifen (4-hydroxy-N-desmethyl-tamoxifen) hasanti-estrogenic effects in breast cancer cellswith potency similar to 4-hydroxy-tamoxifen.Cancer Chemother Pharmacol 2005;55:471–8.

91. Lim YC, Li L, Desta Z, Zhao Q, Rae JM, FlockhartDA, Skaar TC. Endoxifen, a secondary metabo-lite of tamoxifen, and 4-OH-tamoxifen inducesimilar changes in global gene expression pat-terns in MCF-7 breast cancer cells. J PharmacolExp Ther 2006;318:503–12.

92. Decensi A, Gandini S, Serrano D, Cazzaniga M,Pizzamiglio M, Maffini F, et al. Randomizeddose-ranging trial of tamoxifen at low doses inhormone replacement therapy users. J Clin On-col 2007;25:4201–9.

93. Xu Y, Sun Y, Yao L, Shi L, Wu Y, Ouyang T, et al.Association between CYP2D6 *10 genotype andsurvival of breast cancer patients receiving ta-moxifen treatment. Ann Oncol 2008;19:1423–9.

94. Goetz MP, Rae JM, Suman VJ, Safgren SL, AmesMM, Visscher DW, et al. Pharmacogenetics oftamoxifen biotransformation is associated withclinical outcomes of efficacy and hot flashes.J Clin Oncol 2005;23:9312–8.

95. Goetz MP, Kamal A, Ames MM. Tamoxifenpharmacogenomics: the role of CYP2D6 as apredictor of drug response. Clin Pharmacol Ther2008;83:160–6.

96. Kiyotani K, Mushiroda T, Sasa M, Bando Y,Sumitomo I, Hosono N, et al. Impact ofCYP2D6*10 on recurrence-free survival in breastcancer patients receiving adjuvant tamoxifentherapy. Cancer Sci 2008;99:995–9.

97. Nowell SA, Ahn J, Rae JM, Scheys JO, Trovato A,Sweeney C, et al. Association of genetic varia-tion in tamoxifen-metabolizing enzymes withoverall survival and recurrence of disease inbreast cancer patients. Breast Cancer Res Treat2005;91:249–58.

98. Wegman P, Vainikka L, Stal O, Nordenskjold B,Skoog L, Rutqvist LE, Wingren S. Genotype of

metabolic enzymes and the benefit of tamoxifenin postmenopausal breast cancer patients.Breast Cancer Res 2005;7:R284–90.

99. Wegman P, Elingarami S, Carstensen J, Stal O,Nordenskjold B, Wingren S. Genetic variants ofCYP3A5, CYP2D6, SULT1A1, UGT2B15 and ta-moxifen response in postmenopausal patientswith breast cancer. Breast Cancer Res 2007;9:R7.

100. Bonanni B, Macis D, Maisonneuve P, JohanssonHA, Gucciardo G, Oliviero P, et al. Polymor-phism in the CYP2D6 tamoxifen-metabolizinggene influences clinical effect but not hotflashes: data from the Italian Tamoxifen Trial.J Clin Oncol 2006;24:3708–9.

101. Mortimer JE, Flatt SW, Parker BA, Gold EB,Wasserman L, Natarajan L, Pierce JP. Tamox-ifen, hot flashes and recurrence in breast cancer.Breast Cancer Res Treat 2008;108:421–6.

102. Newman WG, Hadfield KD, Latif A, Roberts SA,Shenton A, McHague C, et al. Impaired tamox-ifen metabolism reduces survival in familialbreast cancer patients. Clin Cancer Res 2008;14:5913–8.

103. Punglia RS, Burstein HJ, Winer EP, Weeks JC.Pharmacogenomic variation of CYP2D6 and thechoice of optimal adjuvant endocrine therapyfor postmenopausal breast cancer: a modelinganalysis. J Natl Cancer Inst 2008;100:642–8.

104. Carlson RW, Hudis CA, Pritchard KI. Adjuvantendocrine therapy in hormone receptor-positivepostmenopausal breast cancer: evolution ofNCCN, ASCO, and St Gallen recommendations.J Natl Compr Canc Netw 2006;4:971–9.

105. Lien EA, Solheim E, Kvinnsland S, Ueland PM.Identification of 4-hydroxy-N-desmethyltamoxifenas a metabolite of tamoxifen in human bile. Can-cer Res 1988;48:2304–8.

106. Jordan VC, O’Malley BW. Selective estrogen-receptor modulators and antihormonal resis-tance in breast cancer. J Clin Oncol 2007;25:5815–24.

107. van Agthoven T, Sieuwerts AM, Meijer-vanGelder ME, Look MP, Smid M, Veldscholte J, etal. Relevance of breast cancer antiestrogen re-sistance genes in human breast cancer progres-sion and tamoxifen resistance. J Clin Oncol2009;27:542–9.

108. Langan-Fahey SM, Tormey DC, Jordan VC. Ta-moxifen metabolites in patients on long-termadjuvant therapy for breast cancer. Eur J Cancer1990;26:883–8.

109. Wakeling AE, Slater SR. Estrogen-receptor bind-ing and biologic activity of tamoxifen and itsmetabolites. Cancer Treat Rep 1980;64:741–4.

110. Lazarus P, Blevins-Primeau AS, Zheng Y, Sun D.Potential role of UGT pharmacogenetics in can-cer treatment and prevention: focus on tamox-ifen. Ann N Y Acad Sci 2009;1155:99–111.

Review


pharmacogenomics of tamoxifen therapy

Documents