Effects of Genetic Polymorphisms of Drug Metabolizing Enzymes and co-Medications on Tamoxifen Metabolism in Black South African Women with Breast Cancer

Abstract

Clinical outcomes of tamoxifen (TAM) treatment show wide interindividual variability. Comedications and genetic polymorphisms of enzymes involved in TAM metabolism contributes to this variability. Drug–drug and drug–gene interactions have seldom been studied in African Black populations. We evaluated the effects of commonly co-administered medicines on TAM pharmacokinetics in a cohort of 229 South African Black female patients with hormone-receptor positive breast cancer. We also investigated the pharmacokinetic effects of genetic polymorphism in enzymes involved in TAM metabolism, including the variants CYP2D6*17 and *29, which have been mainly reported in people of African descent. TAM and its major metabolites, N-desmethyltamoxifen (NDM), 4-OH-tamoxifen, and endoxifen (ENDO), were quantified in plasma using the liquid chromatography-mass spectrometry. The GenoPharm open array was used to genotype CYP2D6, CYP3A5, CYP3A4, CYP2B6, CYP2C9, and CYP2C19. Results showed that CYP2D6 diplotype and CYP2D6 phenotype significantly affected endoxifen concentration (P < 0.001 and P < 0.001). CYP2D6*17 and CYP2D6*29 significantly reduced the metabolism of NDM to ENDO. Antiretroviral therapy had a significant effect on NDM levels and the TAM/NDM and NDM/ENDO metabolic ratios but did not result in significant effects on ENDO levels. In conclusion, CYP2D6 polymorphisms affected endoxifen concentration and the variants CYP2D6*17 and CYP2D6*29 significantly contributed to low exposure levels of ENDO. This study also suggests a low risk of drug–drug interaction in patients with breast cancer on TAM.

Tamoxifen (TAM) has been in use worldwide for > 30 years as an adjuvant therapy for patients with estrogen receptor positive breast cancer (BC).¹ TAM inhibits tumor progression and promotes apoptosis in ER-positive tumors² reducing the risk of recurrence by 30% when taken for 5 or more years.³ TAM is a prodrug that is metabolized by cytochrome P450 liver enzymes through 2 parallel pathways into the primary metabolites N-desmethyltamoxifen (NDM) and 4-hydroxy-tamoxifen (4-OH-TAM).⁴ NDM and 4-OH-TAM are further metabolized to the active metabolite endoxifen (ENDO). Of those major metabolites, 4-OH-TAM and ENDO exhibit 30–100-fold more potent anti-estrogenic effects than the prodrug.^5–7 ENDO is the most important metabolite, its steady-state concentration is 5–10-fold higher than that of 4-OH-TAM.⁶ The conversion of TAM by NDM via CYP3A4/5 to NDM represents ~ 92% of TAM metabolism, whereas the 4-OH-TAM pathway represents ~ 8%.⁷

Despite TAM’s high efficacy, the clinical outcomes of treatment with TAM at its standard dose are affected by factors such as age,⁸ body mass index (BMI),⁹ patient compliance variability,¹⁰ drug–drug interaction (DDI),¹¹ drug–food interaction,¹² liver and renal function,¹³ drug resistance,¹⁴ and genetic polymorphisms¹⁵ of enzymes involved in the metabolic pathway. Genetic polymorphisms of CYP450 enzymes can lead to adverse drug reactions or inadequate responses to commonly prescribed doses of some therapeutic agents.¹⁶ Genetic polymorphisms may also result in nonfunctional, reduced function, or increased function alleles whose genotypes are associated with predicted phenotype classifications as poor (PM), intermediate (IM), normal (NM), or ultra-rapid metabolizers (UMs).¹⁷

The CYP2D6 enzyme plays a major role in the formation of ENDO and the gene coding the enzyme is highly polymorphic contributing to the large interindividual differences in enzyme activity.¹⁷ It has been shown that ENDO concentrations significantly correlate with CYP2D6 genotype.^10,18 The commonly reported alleles are categorized as functional (CYP2D6*1 and *2), decreased function (CYP2D6*9, *10, *17, *29, and *41), and nonfunctional (CYP2D6*3, *4, *5, and *6).^18,19 Genetic variation in the CYP2D6 gene has been associated with variation in plasma concentrations of ENDO, accounting for up to 50% of the variability in ENDO concentrations.¹⁸ Genetic and drug interactions increase or decrease systemic ENDO concentrations. Patients with low CYP2D6 enzyme activity as a result of CYP2D6 genetic polymorphisms exhibit lower plasma ENDO concentrations than patients without those polymorphisms.^10,18 However, whether or not genetic polymorphisms of TAM metabolizing enzymes predict clinical outcomes, such as survival and disease-free survival, is as yet unclear.²⁰

CYP3A4/5 plays a significant role in the transformations of TAM to NDM and of 4-OH-TAM to ENDO whereas CYP2C9 has been shown to play a minor role in the transformation of TAM to NDM and TAM to 4-OH-TAM.^10,21 The clinical effects of genetic polymorphisms in CYP3A4 and CYP3A5 on TAM metabolism are not well-understood. Significant associations of CYP3A4, CYP3A5, and CYP2C9 variations with metabolic ratios (MRs) have been reported, although there were no associations with plasma concentrations of the metabolites, showing their modest effects on TAM bioactivation.²¹ Other studies have shown that CYP2C9 and CYP2C19 activity may also be associated with TAM metabolism.^22,23

The DDIs are a significant confounding factor in determining drug efficacy and safety. For example, in studies where TAM was co-prescribed with a CYP2D6 inhibitor, paroxetine,²⁴ and a CYP3A4 inducer, rifampicin,¹¹ the efficacy of TAM for BC management was compromised as ENDO plasma concentrations dropped below the efficacious concentration.^25,26

Most reported drug-gene and DDI studies have been performed in European and Asian populations, with a few in African populations. It has been shown that CYP2D6*17 allele can significantly influence the generation of ENDO from TAM in African healthy volunteers.²⁷ We therefore conducted this observational study in patients with BC to see if the effects of genetic variations of CYP2D6 and other involved enzymes could account for observed variations in ENDO exposure levels.

The primary aim of this study was to determine the effects of genetic polymorphism of CYPs on the pharmacokinetics of TAM and its major metabolites in African patients with BC. The secondary aim was to evaluate the effect of co-administered medications on TAM metabolism and pharmacokinetics.

MATERIALS AND METHODS

Study population

Black African women with hormone-responsive BC enrolled in the South Africa Breast Cancer Health Outcomes Study (SABCHO)²⁸ were invited to participate in the study. Patients had to be prescribed a daily dose of 20 mg TAM for at least 3 months and verbally confirmed to be on their daily treatment to ensure that TAM and metabolite levels had reached a steady state. Study participants were recruited from Sefako Makgatho Health Sciences University, located in rural Ga-Rankuwa, North West Province and Chris Hani Baragwanath and Charlotte Maxeke Johannesburg Academic Hospitals located in urban Johannesburg, Gauteng, South Africa. There were some patients who were concurrently on antiretroviral therapy (ART) for HIV and had taken it for at least 3 months. Moreover, patient’s intake of antihypertensive and antidiabetic drugs was recorded. Patients with BC not on TAM, non-Black were excluded from participation and noncompliant (with TAM concentration < 60 ng/mL) were excluded from further data analysis.

Bioanalysis of TAM and metabolites

Samples were collected from all patients about 12–15 hours since their last TAM dose. TAM was taken before going to bed (times ranging from 8 to 10 PM) and samples were collected the morning of the next day (between 10 and 12 midday). Single blood samples were collected from all the patients. TAM and its metabolites were extracted from plasma samples using protein precipitation with ice-cold acetonitrile as the extraction solvent. Briefly, 200 μL of plasma was spiked with 10 μL of 2 μg/mL propranolol (internal standard) followed by addition of 590 μL of ice-cold acetonitrile. The mixture was vortexed for 30 seconds, and sonicated for 2 minutes before centrifugation at 16,000 g for 10 minutes. Aliquots of 700 μL of supernatant were collected and evaporated to dryness under a gentle stream of nitrogen. The dried residue was reconstituted in 50 μL of mobile phase and 10 μL was injected into the liquid-chromatography tandem mass spectrometry (LC-MS–MS) system for analysis (a 3200 Q TRAP Series triple quadrupole (AB Sciex) LC-MS/MS system coupled to an Agilent 1,100 series HPLC system with Analyst software version 1.6 (AB Sciex) was used to carry out all the analysis). The compounds of interest were separated using Zorbax C18 2.1 X100 mm, 3.5 μm column (Zorbax Agilent, Santa Clara, CA). The mobile phase consisted of 0.1% formic acid in 10 mM ammonium formate solution as mobile phase A and 0.1% formic acid in acetonitrile as mobile phase B delivered using a gradient elution: 0–9 minutes, B 30%, 9.01–9.5 minutes, B 52%, and 9.5–13 minutes B 30%. The column was maintained at a temperature of 40°C.

Analytes were followed using multiple reaction monitoring (m/z 372.5 → 72.2, 374.4 → 58.1, 358.4 → 58.0, 388.4 → 72.2, and 260.3 → 183.3 for TAM, total ENDO, NDM, 4-OH-TAM and propranolol, respectively). MS/MS analyses were performed in positive ionization mode, ion source temperature 500°C, curtain gas 25, ion spray voltage 5,500, and GS1 and GS2 gas 50 and 30, respectively.

Genotype analysis

DNA isolation and genotyping.

TAM-metabolizing enzymes were tested for genetic variation. DNA was extracted from 200 μL of peripheral whole blood using the MagMAX DNA Multi-Sample Ultra 2.0 Kit on the ThermoFisher KingFisher Flex Purification System with the MagMAX Ultra 2.0—200 μL script for KingFisher Flex. Extracted DNA was quantified with the Qubit 4 fluorometer using the Qubit dsDNA BR Assay Kit and stored at −20°C short-term or −80°C long-term before analysis. Genotyping for drug metabolizing enzymes was performed on the GenoPharm open array panel as per manufacturers’ protocol. The arrays were run on the QuantStudio 12 K Flex Real-Time PCR System. Genotype calls were generated with TaqMan Genotyper Software. CYP2D6 copy number was determined using the Applied biosystems TaqMan copy number assays for exon 9 the primary copy number assay (Assay ID: Hs00010001_cn) to quantify CYP2D6 duplications or to identify CYP2D6 gene deletion (*5) in the samples. AlleleTyper software was used to convert sample genotype information for the CYP genes interrogated to the star (*) allele nomenclature, using a predefined allele translation table that maps a specified allele pattern to the star allele call. In total, we genotyped for variants of 6 genes (CYP2B6, CYP2C9, CYP2C19, CYP2D6, CYP3A4, and CYP3A5) involved in the metabolism of TAM.

Statistical analysis

Data were analyzed using Stata version 16 (Stata Corp). Categorical variables were described as frequencies and percentages. TAM plasma concentration was used as a measure of compliance,²⁹ with TAM steady-state concentrations of > 60 ng/mL defined as compliant to drug intake. Shapiro–Wilk’s test was performed to test for the normality of continuous variables. Continuous variables were summarized as medians and interquartile ranges (IQRs). The predicted phenotypes for the CYP2D6 gene were derived from the genotype activity scores and classified as UM, NM, IM and PM according to the Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines.³⁰ The frequency histogram of log (NDM/ENDO) was plotted and the presence of different categories of patients was indicated if the histogram was bimodal and revealed the presence of different metabolic categories of patients. On visual inspection of frequency histogram and probit plot, approximate antimode position was established as the point on graph where two different modes are separated. The Mann Whitney U test was used to test for the effects of comedications and BMI on MR, metabolites, and TAM concentrations. The Kruskal-Wallis test was performed to test for the effects of age, diplotypes, and predicted phenotypes on drugs, metabolite concentrations, and MRs. For all analyses, P < 0.05 was regarded as statistically significant.

Ethical approval

The study was approved by the University of the Witwatersrand Human Research Ethics Committee with Clearance Certificate Number M180865. Written informed consent was obtained from all individual participants in the study.

RESULTS

Patient demographic characteristics

Of the 478 patients recruited in the SABCHO cohort, 229 patients had plasma TAM concentrations > 60 ng/mL and were deemed compliant to treatment and included in the descriptive (Table 1) and subsequent analyses. Most (57.2%) of the included participants were recruited from Chris Hani Baragwanath Academic Hospital. Median age was 52.1 years (IQR 44.9–60.8 years) and median BMI was 32.5 kg/m² (range 27.6–38.1 kg/m²). Median duration on TAM was 12.6 months (IQR 6.1–26.3 months), with a minimum and maximum duration of 3 months and 58 months, respectively. Forty patients (17.5%) were HIV-positive and were mainly prescribed a cocktail of tenofovir, emtricitabine, and efavirenz as ART to be taken once daily. A total of 12 (5.2%) and 98 (42.8%) patients were diabetic and hypertensive and they were on antidiabetics and antihypertensives, respectively.

Table 1.

Demographic characteristics of patients with breast cancer

Variable	Frequency (%), median (IQR) N = 229
Site
CHBAH	131 (57.2)
CMJAH	66 (28.8)
SMU	32 (14.0)
Age, years	52.1 (44.9–60.8)
BMI, kg/m2	32.5 (27.6–38.1)
Duration on tamoxifen (months)	12.6 (6.1–26.3)
Radiotherapy
Yes	117 (51.3)
No	111 (48.7)
Menopausal status
Postmenopausal	124 (54.2)
Premenopausal	105 (45.9)
HIV patients on ART
Yes	40 (17.5)
No	189 (82.5)
Diabetic patients on antidiabetics
Yes	12 (5.2)
No	217 (94.8)
Hypertensive patients on antihypertensives
Yes	98 (42.8)
No	131 (57.2)

Missing: Radiotherapy (n = 1).

ART, anti-retroviral therapy; BMI, body mass index; CHBAH, Chris Hani Baragwanath Academic Hospital; CMJAH, Charlotte Maxeke Johannesburg Academic Hospital; IQR, interquartile range; SMU, Sefako Makgatho Health Sciences University.

Plasma concentrations of TAM and its metabolites

Table 2 shows plasma concentrations of TAM, metabolites, and MRs. NDM had a much higher median concentration (183 ng/mL) than ENDO (6.03 ng/mL) or 4-OH-TAM (1.98 ng/mL). The median concentration of ENDO was 3.05 times higher than that of 4-OH-TAM. The highest interindividual variation was found in 4-OH-TAM plasma concentration, with a coefficient of variation of 70.7%.

Table 2.

Plasma concentration of TAM and its metabolites

Plasma concentrations and MRs	Median (IQR)	CV (%)
Plasma concentrations, ng/mL
TAM	106 (82.1–154.2)	54.7
ENDO	6.03 (3.68–9.36)	67.8
NDM	183 (113.19–283)	69.5
Hydroxytamoxifen	1.98 (1.12–2.92)	70.7
MRs
NDM/ENDO	27.59 (18.99–42.07)	86.9
4-OH-TAM/ENDO	0.31 (0.21–0.48)	80.3
TAM/4-OH-TAM	58.99 (37.77–92.39)	71.7
TAM/NDM	0.62 (0.42–0.89)	68.3

4-OH-TAM, 4-hydroxytamoxifen; CV, coefficient of variation; ENDO, endoxifen; IQR, interquartile range; MR, metabolic ratio; NDM, N-desmethyltamoxifen; TAM, tamoxifen.

Patients’ subgroups by probit analysis

The frequency histogram of the logarithm of the MR (NDM/ENDO; Figure 1a) revealed the presence of different metabolic categories of patients and a bimodal distribution of CYP2D6 activity in the population with respect to their MR. The antimode (Figure 1b) was calculated to be 56.23 (log MR = 1.75). Patients above the antimode log MR of 1.75 had low concentrations of ENDO and their genotypes were predictive of either IM or PM. All patients (n = 3) with genotypes predictive of the PM phenotype where in this subgroup above the log MR of 1.75.

Figure 1.

(a) Histogram of the phenotyping data showing demarcation of metabolic groups based on antimode calculations. (b) Patient’s probit plot obtained by fitting the log MR data. ENDO, endoxifen; MR, metabolic ratio; NDM, N-desmethyltamoxifen.

Effects of CYP2D6, CYP3A4, CYP3A5, CYP2C9, CYP2C19, and CYP2B6 polymorphisms on tamoxifen metabolism

The CYP2D6 diplotypes were NM/NM (CYP2D6*1/*1, CYP2D6*1/*2, and CYP2D6*2/*2), NM/IM (CYP2D6*1/*17, CYP2D6*1/*10, CYPDD6*2/*17, CYP2D6*2/*29, CYP2D6*1/*29, CYP2D6*2/*41, and CYP2D6*1/*12), IM/IM (CYP2D6*17/*29, CYP2D6*17/*41, CYP2D6*17/*17, and CYP2D6*10/*17), NM/PM (CYP2D6*1/*5, CYP2D6*2/*5, and CYP2D6*1/*12), PM/IM (CYP2D6*5/*29, CYP2D6*5/*17, and CYP2D6*7/*10), PM/PM (CYP2D6*5/*5), NM/UM (CYP2D6*1/*1X2 and CYP2D6*1/*2X2), and UM/UM(CYP2D6*2AX2/*2X2, CYP2D6* 1X2/*2X2, and CYP2D6*2X2/*2X2), respectively. The median ENDO concentration in carriers of heterozygous NM and UM alleles (NM/UM) and two UM alleles were the highest (19.1 ng/mL and 9.3 ng/mL; Figure 2a). For NM/NM patients, the median and range of ENDO concentrations were 8.7 (6.4–12.0) ng/mL (Figure 2a). ENDO concentrations were lowest in carriers of heterozygous IM and PM alleles (3.2 ng/mL), respectively (Figure 2a). The low levels of ENDO were detected in patients classified as PM/IM, NM/PM, and IM/IM with medians and IQRs of 3.2 (2.3–7.6) ng/mL, 5.5 (3.7–10.4) ng/mL, and 3.7 (2.8–4.5) ng/mL, respectively (Figure 2a). The median plasma ENDO concentration differed significantly by CYP2D6 diplotype (Figure 2a) and phenotype (Figure 2b; P < 0.001).

Figure 2.

Box plot of plasma endoxifen concentration by CYP2D6 diplotype (a) and phenotype (b), NDM (c), hydroxytamoxifen (d), plasma concentrations, and MRs (e–h) by CYP2D6 phenotype. IM, intermediate metabolizer; MR, metabolic ratio; NM, normal metabolizer; PM, poor metabolizer; UM, ultra-rapid metabolizer.

Median levels of NDM (Figure 2c) and 4-OH-TAM (Figure 2d) significantly differed by CYP2D6 phenotype (P = 0.012 and P < 0.001). The median MR (TAM/4-OH-TAM; Figure 2e) and MR (NDM/ENDO; Figure 2f) differed by CYP2D6 phenotype (P = 0.032 and 0.013, respectively), but MR (TAM/NDM; Figure 2g) and MR (4-OH-TAM/ENDO; Figure 2h) did not (P = 0.300 and P = 0.320). No significant differences were observed in TAM, ENDO, NDM, hydroxytamoxifen, and MRs by CYP3A4, CYP3A5, CYP2C9, CYP2C19, and CYP2B6 phenotypes.

Influence of CY2D6*17 and CYP2D6*29 variants on endoxifen concentration

The effect of CYP2D6*17 and CYP2D6*29 variants on ENDO concentration and MR (NDM/ENDO) was assessed for CYP2D6*1/*1, CYP2D6*1/*2, and CYP2D6*2/*2, then. for the heterozygous genotypes CYP2D6*1/*17, CYP2D6*2/*17, CYP2D6*1/*29, and CYP2D6*2/*29, and the low enzyme activity alleles (CYP2D6*17/*17, CYP2D6*29/*29, and CYP2D6*29/*17), as shown in Figure 3. There were significant differences in median ENDO levels by CYP2D6 genotype groups. The genotype groups for CYP2D6*1/*1, CYP2D6*1/*2, and CYP2D6*2/*2 had the highest concentrations (8.67 ng/mL), whereas the lowest concentrations (3.65 ng/mL) were observed for the CYP2D6*17/*17, CYP2D6*17*/29, and CYP2D6*29*29 genotype groups (Figure 3a). There were significant differences in median MR (NDM/ENDO) between the CYP2D6 genotype groups (P = 0.035; Figure 3b). The higher the MRs we observed, the less the enzyme activity for the pathway.

Figure 3.

Effect of CYP2D6*17 and *29 on ENDO concentration and on the metabolism of NDM to ENDO. 1 (CYP2D6*1/*1, CYP2D6*1/*2, and CYP2D6*2/*2), 2 (CYP2D6*1/*17 and CYP2D6*2/*17), 3 (CYP2D6*1/*29 and CYP2D6*2/*29), and 4 (CYP2D6*17/*29 and CYP2D6*17/*17). ART, antiretroviral therapy; ENDO, endoxifen, MR, metabolic ratio; NDM, N-desmethyltamoxifen; TAM, tamoxifen.

Effect of efavirenz-based ART on tamoxifen, metabolites, and metabolic ratios

Patients on efavirenz, emtricitabine, and tenofovir-based ART (n = 31) differed from patients not on ART (n = 198) in median NDM (P = 0.018) concentration (Figure 4c), TAM/NDM (P = 0.004; Figure 4e) and NDM/ENDO (P = 0.012; Figure 4g).

Figure 4.

TAM (a), ENDO (b), NDM (c), hydroxytamoxifen (d), and MRs (e–h) by efavirenz-based ART. ART, antiretroviral therapy; ENDO, endoxifen, MR, metabolic ratio; NDM, N-desmethyltamoxifen; TAM, tamoxifen.

DISCUSSION

In this observational study, we observed a high level of interindividual variability in exposure levels of TAM and its major metabolites. These observations are in line with several studies evaluating TAM metabolism that have also observed high interindividual variability of TAM concentration.^10,31 Of the 478 patients recruited in the SABCHO cohort, only 229 (47.9%) of the patients taking TAM were compliant with their treatment and the rest had plasma concentrations below 60 ng/mL. We showed that HIV treatment was associated with nonadherence in this cohort.³² With respect to the active metabolite, ENDO, of the 229 patients, 111 (48.5%) had concentrations below the putative threshold of 5.97 ng/mL.³³ This finding is in contrast to lower numbers reported in other populations of patients below this minimum effective concentration (MEC) of 32% in White patients³⁴ and 35.8% in Northern African patients.³¹ It is worth noting that the north African populations of Tunisia had low frequencies of low activity variants CYP2D6*17 and *29.³⁵ The allele frequencies of CYP2D6 *17 and CYP2D6 *29 were 21% and 7.9%, respectively (as described in Supplementary Material S1, Table S1), and high in this cohort compared with other world populations. It can therefore be postulated that the variants CYP2D6*17 and CYP2D6*29 significantly contributed to low exposure levels of ENDO as compared with the most common reduced-function allele in White patients, CYP2D6*41.³⁴

We identified 13 patients with discordant predicted vs. observed CYP2D6 activity. Their diplotypes were predictive of NMs but their ENDO levels were indicative of IMs and PMs. It is possible that these 13 patients have other variants on the “backbones” of the alleles genotyped for. Whole Genome Sequencing is required to investigate this further. It could also be that these patients had ingested unreported drugs or herbs that inhibit the activity of CYP2D6, thus phenocopying the IM status. Most of the IM (70%) had ENDO concentrations below the putative threshold, similar to a previous study.³⁶ Such patients may benefit from dose escalation, as reported in other studies,^37,38 in order to reach high concentrations of ENDO.

Given the large variation in plasma concentrations of TAM and its metabolites, we evaluated the effects of previously reported factors that influence TAM metabolism, such as BMI, age, polypharmacy, and genetics. To evaluate their effects, we used both plasma concentrations of TAM and its metabolites and MRs of the two parallel pathway of TAM metabolism. The MRs enabled us to interrogate enhancement or reduction effects on the enzymatic pathways involved in the variables evaluated. We did not find associations of BMI and age with exposure levels of TAM and its primary metabolites. However, a study among patients of Asian descent found a significant difference in ENDO²⁹ levels between patients with BMI > 30 compared with those with BMI ≤ 30.³⁰ Our findings of no impact of age and BMI on plasma concentration of TAM and its metabolites (Table S2, Supplementary Material S2) are in agreement with reports in Ethiopia and elsewhere.^31,39,40 Median concentrations of TAM, ENDO, and NDM were higher in patients with BC who had the radiotherapy exposure as compared with patients without the exposure. This is in agreement with findings reported elsewhere.⁴¹ TAM and metabolite concentrations were comparable between postmenopausal and premenopausal patients (all P > 0.05), also in agreement with other findings in an African population.³¹ The median concentrations of TAM and its metabolites were higher in premenopausal patients as compared with postmenopausal patients, which is in contrast with other findings.^12,42

In this study, the relative contributions of the metabolic pathways TAM to 4-OH-TAM and TAM to NDM were ~ 2% and 98%, respectively. This is within the range of reported values of 8% and 92% relative contributions.⁴ Average NDM to ENDO metabolic activity was assessed for the effects of CYP2D6*1/*1, CYP2D6*1/*2, and CYP2D6*2/*2 for the NMs, then for the heterozygous genotypes CYP2D6*1/*17, CYP2D6*2/*17, CYP2D6*1/*29, and CYP2D6*2/*29, and the low enzyme activity alleles (CYP2D6*17/*17, CYP2D6*29/*29, and CYP2D6*17/*29). The higher the MR observed, the less enzyme activity there is in the pathway. The variants showed a significant effect on ENDO levels. The patients with diplotypes predictive of IM/IM had concentrations below the putative threshold of 5.97 ng/mL.³³ The effects of CYP2D6*17 and *29 in the IM/IM diplotype were all associated with subtherapeutic levels of ENDO. Those with diplotypes of UM/NM had very high concentrations of ENDO above 5.97 ng/mL. Those with the IM/PM and IM/NM mutations had ENDO concentrations that ranged from below to above the therapeutic range. In our analysis, 3 patients with the CYP2D6 PM/PM diplotypes reported low levels of ENDO (6.6 ng/mL, 4.1 ng/mL, and 8.7 ng/mL), which is consistent with an inability to metabolize NDM. The observed ENDO may have been produced via the 4-OH-TAM route. Although two of the values are above the MEC, it could be that this MEC value is not hard set as it is only from one report.⁴³ The probit analysis (Figure 1), might be a better tool to categorize those at high risk of subtherapeutic ENDO exposure levels at log MR above 1.75, a category all those predicted to be PM fall ins.

The prevalence of HIV among our patients with BC was 17.5%. Many of the patients were receiving multiple comedications. This polypharmacy could put them at risk of DDIs either through inhibition or induction of enzymes involved in the metabolism of TAM. Antihypertensive (mainly hydrochlorothiazide, enalapril, and amlodipine drugs) and antidiabetic (mainly metformin, glibenclamide, and insulin) drugs have not been reported in literature to have any inhibition or induction effect on enzymes involved in TAM metabolism, hence unlikely to result in DDI. The effect of efavirenz-containing ART could, however, result in DDI because efavirenz is a potent inducer of enzymes involved in TAM metabolism, such as CYP3A, CYP2B6, and the UGTs.⁴⁴ Effects of ART on TAM and metabolite levels was evaluated. There was a significant difference in median NDM (P = 0.018; Figure 4c), MR (TAM/NDM, P = 0.004; Figure 4e), and MR (NDM/ENDO, P = 0.012; Figure 4g) by efavirenz-based ART. We did find a lower plasma concentration of NDM and a higher MR for the TAM/NDM pathway in ART patients, which was surprising because efavirenz is a known inducer of CYP3A4/5,²⁵ and we expected it to result in more NDM production. Based on this mechanism, even more ENDO was expected through enhanced 4-OH-TAM metabolism to ENDO and through availability of more NDM substrate for the CYP2D6 to metabolize to ENDO. Rifampicin, an anti-tuberculosis medication that induces ABCB1 and UDP-glucuronosyltransferases and CYP3A, was shown to lower the levels of TAM and result in subtherapeutic levels of ENDO in patients who were on TAM.¹¹ This was explained by the possible increased activity of UGT which then removed ENDO from the system.⁴⁵ Our findings suggest little risk for DDI in patients with BC taking TAM and efavirenz-based ART, which are routinely prescribed. It must, however, be noted that observational studies might not be optimal to detect some DDIs as many factors are uncertain, such as time the various medicines where taken, compliance to comedications, and time on the various medications.

We evaluated the effects of polymorphisms of CYP2C9, CYP2B6, CYP2C19, CYP3A4, CYP3A5, and CYP2D6 on the metabolism of TAM, with a special focus on the effects of the variants which have been commonly reported in people of African ancestry CYP2D6*17 and CYP2D6*29. Variants of CYP2C9, CYP2B6, CYP2C19, CYP3A4, and CYP3A5 did not show any effect on the metabolism of TAM to its various metabolites. Other studies in populations of Asian and European descent have also found that CPY2C9, CYP2C19, and CYP2B6 polymorphisms had no impact on the pharmacokinetics of TAM.^23,46 The CYP2D6 diplotype predictive of the IM phenotype had a significant effect on ENDO concentrations (P < 0.001 and P < 0.001). This is in agreement with the proposed role of CYP2D6 in these pathways from previous studies.^10,31

The PM and IM genotypes were expected to have lower ENDO due to effects on the NDM to ENDO pathway catalyzed by CYP2D6. PMs due to CYP2D6 *5/*4 and *4/*5 genotypes have been previously reported to be associated with low and subtherapeutic levels of ENDO⁴⁷ and IM genotypes, such as CYP2D6*10, have also been shown to be associated with low and sometimes subtherapeutic levels of ENDO.^46,47 These observations have led to CYP2D6 genotype dose escalation studies in Asians³⁷ and mixed populations³⁸ that have demonstrated that increasing the dose of TAM can raise ENDO to therapeutic levels. Our finding that patients carrying CYP2D6*17 and CYP2D6*29 diplotypes have ENDO concentrations below the therapeutic range implies that such patients might benefit from dose adjustments, as has been done for patients carrying the CYP2D6*10 in Asian populations³⁷ and in patients with diplotypes predictive of IMs in other populations.⁴⁸ Our recent study on ENDO exposure levels in patients of different CYP2D6 genotypes indicated that dose escalation from 20 mg/day to 40 mg/day could result in all those homozygous for the CYP2D6*17 reaching therapeutic levels of ENDO.²⁷

In this observational study, patients’ drug intakes were not actively monitored and we only know that all patients enrolled took their last dose of TAM within 15 hours before blood sample collection. Further studies in a controlled setting with respect to treatment compliance and specific drug intake and sampling times are recommended.

CONCLUSION

The effects of genetic polymorphisms of enzymes involved in the metabolism of TAM demonstrated the significance of CYP2D6 polymorphisms on tamoxifen metabolism. The common variants in African populations, CYP2D6*17 and CYP2D6*29, decreased NDM metabolism to ENDO, which also had a significant effect in reducing ENDO exposure levels. These findings suggest the need to evaluate the effectiveness of dose escalation in patients carrying the reduced function CYP2D6 variants, CYP2D6*17, and CYP2D6*29.

Study Highlights.

WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?

☑ The CYP2D6 enzyme, which plays an important role in the formation of endoxifen (ENDO), is highly polymorphic. The most commonly reported alleles are categorized as normal functional alleles (CYP2D6*1 and *2), which are associated with higher ENDO concentrations; decreased functional alleles (CYP2D6*9, *10, *17, *29, and *41), and nonfunctional alleles (CYP2D6*3, *4, *5, and *6) which are associated with lower ENDO concentrations.

WHAT QUESTION DID THIS STUDY ADDRESS?

☑ What impact do genetic polymorphism of tamoxifen (TAM) metabolizing enzymes and comedications for other chronic conditions have on ENDO exposure levels in patients with breast cancer of African descent?

WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?

☑ Reduced ENDO exposure levels in patients carrying the African-specific variants CYP2D6*17 and CYP2D6*29. There is minimum risk for drug–drug interactions (DDIs) due to effect of efavirenz containing antiretroviral therapy (ART) on ENDO levels.

HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE?

☑ The findings of this study demonstrate that among Black patients with breast cancer, the variants CYP2D6*17 and CYP2D6*29 mostly found in people of African ancestry are associated with significant reduction of ENDO levels and that dose escalation in patients with genotypes predictive of intermediate metabolizer phenotype could result in improved treatment outcomes. There is no clinically significant DDI involving ART among patients with breast cancer taking TAM.