Impact of Metabolizing Enzymes on Drug Response of Endocrine Therapy in Breast Cancer

Pilar H Saladores; Jana C Precht; Werner Schroth; Hiltrud Brauch; Matthias Schwab

Expert Rev Mol Diagn. 2013;13(4):349-65.

Expert Commentary

The evidence of CYP2D6 genotype–phenotype correlation associated with tamoxifen pharmacokinetics in early breast cancer is compelling. However, the scientific community is currently divided on the clinical relevance of CYP2D6 genotype predicting tamoxifen response. Conflicting outcome studies have led to confusion about its utility in treatment decision making. The major aim of molecular diagnostics is to provide an accurate determination of phenotype in relation to outcome; however, outcome may be confounded by several factors that include inaccurate phenotype, CYP2D6 phenocopying and unsuitable sources of DNA for genetic testing. In addition, other DMEs involved in tamoxifen biotransformation may also be influential in prediction of treatment response but further studies are warranted to understand their clinical value. It must be noted that there is a subset of patients, mainly postmenopausal women, who may benefit from genetic testing if found to have impaired CYP2D6 metabolism. So far, most pharmacogenetic results have been based on retrospective studies with differing study designs; therefore, a final decision must await appropriately designed prospective clinical trials.

The current evidence suggests different roles and strengths for pharmacogenetic diagnostics within the context of tamoxifen and AI treatment. CYP2D6 was shown to be a promising predictor for tamoxifen outcome while AI pharmacogenetics is still in its infancy, suggesting a possible role for the prediction of the onset of adverse drug reactions.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Next Section

Table 1. Major *CYP2D6* variant alleles.
Variant alleles	Genetic variants	Predicted enzyme activity	Allele frequency (%)			Ref.
			European^†	Asian	African
*3	Frameshift	Nonfunctional	2	0.8	0.2–0.6	[38,40]
*4	Splicing defect	Nonfunctional	22	0.5–2.8	0.9–8.5	[38,40]
*5	Gene deletion	Absent	1.8	4.5–6.2	0.6–6.9	[38,40]
*6	Frameshift	Nonfunctional	0.8		0.5	[38,40]
*9	Deletion	Decreased	2.5	3.3	0.3	[38,40]
*10	Base substitution	Decreased	1.7	38.1–70.0	2.5–8.6	[38,40]
*17	Base substitution	Decreased		0.5	9–34	[38,40]
*41	Splicing defect	Decreased	9.2	2.3	2.8	[37,40]
*1xN	Amplification	Increased	0.8	0.5	1.6–3.3	[38,40]
*2xN	Amplification	Increased	0.5	0.5	1.3–2.5	[38,40]

^†All allelic frequencies derived from the German population of European descent (n = 492) in a study by Schroth et al. based on AmpliChip® genotyping [40].

Table 2. Genotyping assays for *CYP2D6*.
Assays	Methods	Comments	Selected references
Classic assays	PCR-RFLP	Mainly replaced by newer techniques	[47,49]
	Sanger sequencing	Gene duplication and deletion not detectable	[148]
Real-time PCR assays	LightCycler® (Roche)	Melting curve analysis	[149]
	TaqMan® (Applied Biosystems)	Separate assay for each SNP; determination of CNV possible	[19,33,47,66]
Microarray assays	AmpliChip® CYP450 Test^† (Roche)	CNVs detectable; high costs of instrument and arrays; 33 alleles in CYP2D6 including deletions and duplications	[40,64]
	Infinity® CYP450 2D6I Assay (AutoGenomics)	Covers 15 variants, including deletions and duplications; CE marked^‡	[43]
	xTAG® assay CYP2D6 kit^† (Luminex)	Covers 15 variants	[150]
Low multiplexed assays	Invader® assay, duplex Invader assay (Third Wave Technologies, Hologic)	Fast; universal fluorophore probes usable; each SNP requires a different reaction	[66]
	Pyrosequencing® (QIAGEN)	CNVs detectable; costly instrument and biotinylated primers; release of pyrophosphate	[151]
Other platforms	DMET™ Plus (Affymetrix)	1936 variants in 231 genes; CNV analysis included	[152]
	VeraCode® ADME-Core-Panel (Illumina)	300,000 genotypes in 6 h; covers 34 genes; BeadXPress Reader system	[153]
	MassARRAY® system, iPLEX® ADME PGx Panel (Sequenom)	MALDI-TOF MS technology; customizable for low-plex assays of highly similar genes	[41]
	Fluidigm	High-throughput: 9216 simultaneous reactions; compatible with TaqMan® genotyping assays	[154,155]

^†Approved by the US FDA.
^‡CE marking refers to conformity marking for products in the European Economic Area.
CNV: Copy number variation; MS: Mass spectrometry; RFLP: Restriction fragment length polymorphism.

Table 3. Summary of pharmacokinetics studies of tamoxifen in pre- and post-menopausal women.
Study (year)	Ethnicity	Menopausal status	Genotyping method	CYP2D6 alleles	Major results	Ref.
Stearns et al. (2003)	Unknown	Unknown	PCR-RFLP	1, 4, 6, 8	Significant decrease of endoxifen levels after paroxetine coadministration (p = 0.004); decrease of endoxifen levels more pronounced in patients with CYP2D6 WT (64%) vs variant CYP2D6 patients (p = 0.03)	[49]
Jin et al. (2005)	Mainly European descent	Premenopausal (32.5%) Perimenopausal (8.8%) Postmenopausal (58.8%)	PCR-RFLP, TaqMan®	1, 3, 4, 5, *6	Significantly lower endoxifen levels in homozygous variant and heterozygous patients vs CYP2D6 WT patients (p = 0.003); under comedication with CYP2D6 inhibitors, decrease of endoxifen levels more pronounced in patients with CYP2D6 WT (58%) vs variant CYP2D6 patients (p = 0.0025)	[47]
Borges et al. (2006)	Mainly European descent	Mixed	PCR-RFLP, Taqman, AmpliChip® CYP450	33 alleles	Endoxifen/NDM plasma ratios significantly increased in accordance with number of functional CYP2D6 alleles; ratios were significantly different among the groups (p < 0.001); significantly lower endoxifen levels in CYP2D6 EM patients using CYP2D6 inhibitors compared with nonusers (p < 0.001)	[30]
Lim et al. (2007)	Asian descent	Unknown	Long PCR	5, 10, *2xN	CYP2D610/10 carriers displayed lower endoxifen levels than those with other genotypes (p < 0.0001)	[156]
Lim et al. (2011)	Asian descent	Premenopausal (83%) Postmenopausal (17%)	INFINITI CYP450 2D6	15 alleles	CYP2D6 WT carriers displayed 2.4- to 2.6-fold higher endoxifen levels compared with either 10/10 or 5/10 carriers (p < 0.001)	[43]
Mürdter et al. (2011)	European descent	Postmenopausal	Taqman	3, 4, 5, 6, 7, 8, 9, 10, *41	NDM/Z-endoxifen^† metabolic ratio and Z-endoxifen levels follow a strong CYP2D6 gene-dose effect (p < 10⁻¹⁶)	[19]
Madlensky et al. (2011)	Mainly European descent	Unknown	AmpliChip® CYP450	33 alleles	Endoxifen levels in CYP2D6 EM patients were 2- and 2.8-fold higher compared with IM and PM patients (p < 0.001)	[44]
Irvin et al. (2011)	Mainly European descent	Premenopausal (56%) Postmenopausal (44%)	AmpliChip® CYP450	33 alleles	Endoxifen levels significantly higher in CYP2D6 EM patients compared with PM (p < 0.001); tamoxifen dose adjustment from 20 to 40 mg/day in CYP2D6 IM resulted in a significant increase of endoxifen levels (p < 0.001)	[64]
Barginear et al. (2011)	European descent (51%) Mixed descent (49%)	Premenopausal (56%) Postmenopausal (44%)	Taqman	3, 4, 5, 6, 7, 8, 9, 10, *41	Z-endoxifen levels increased in direct proportion with the MPA score^‡ (p < 0.0001); after adjustment of tamoxifen dose to 30 mg/day, 90% of eligible individuals showed an increase in Z-endoxifen levels correlating with MPA score	[65]
Kiyotani et al. (2010); Kiyotani et al. (2012)^§	Asian descent	Premenopausal (85%) Postmenopausal (5%) Unknown (10%)	Invader® Assay, Taqman, other PCR-based assays	4, 5, 6, 10, 14, 18, 21, 36, *41	Patients carrying CYP2D6 variants were associated with lower plasma levels of endoxifen (p = 0.0000043); after adjustment of tamoxifen dose to 40 mg/day the endoxifen levels increased 1.4-fold in 1/10 and 1.7-fold in 10/10 carriers (p < 0.001)	[66,157]

^†Z-endoxifen is the active isomer of the endoxifen metabolite.
^‡MPA score is based on the CYP2D6 genotype in which values reflect expected enzymatic activity (i.e., normal function allele: 1; reduced function: 0.5; null: 0).
MPA score is the sum of both CYP2D6 alleles.
^§Overlap between study populations presented in Kiyotani et al. (2010) and Kiyotani et al. (2012).
EM: Extensive metabolizer; IM: Intermediate metabolizer; MPA: Metabolizer phenotype activity score; NDM: N-desmethyltamoxifen; PM: Poor metabolizer; RFLP: Restriction fragment length polymorphism; WT: Wild-type.

Table 4. Confounding factors for correct prediction of CYP2D6 phenotypes.
Confounding factors	Description of evidence	Ref.
CYP2D6 phenocopying	Use of CYP2D6 inhibitor (e.g., SSRI) during tamoxifen therapy impairs CYP2D6 enzyme function and may lead to decreased tamoxifen efficacy and inferior clinical outcome	[30,47,49–51]
DNA source	Observed genotype frequencies of PM alleles displayed exaggerated HWE deviation in DNA samples extracted from tumors possibly due to genotype misclassifications and may be attributable to well-known hemizygous deletions of CYP2D6 in breast cancer tissue (LOH)	[54,57–59]
Genotyping coverage	Insufficient CYP2D6 genotyping coverage can lead to phenotype misclassification affecting its ability to predict outcome	[40,61]