Breast cancer is one of the most common types of cancer and the leading cause of cancer related death in women worldwide (Torre et al., 2015). More than 70% of affected patients are ER-positive and therefore eligible for adjuvant endocrine therapy with tamoxifen (Early Breast Cancer Trialists’ Collaborative Group, 2005; Howlader et al., 2014). This drug has been the gold standard treatment of ER-positive breast cancer patients for more than three decades (Jordan, 2014). Tamoxifen acts as an estrogen antagonist and thus inhibits ER-dependent cell proliferation in breast cancer cells (Jordan, 1976). Tamoxifen decreases the risk of recurrence and mortality by 50 and 30%, respectively. Even so, the efficacy of tamoxifen shows a high degree of variation, with 30–50% of patients exhibiting drug resistance (Early Breast Cancer Trialists’ Collaborative Group, 2005).

Tamoxifen is extensively metabolized by several members of the cytochrome P450 enzymes (Desta et al., 2004; Brauch et al., 2009). Some of the resulting metabolites have higher anti-estrogenic effects in breast cancer tissue than tamoxifen itself. Among these, 4-hydroxytamoxifen (4-OH-Tam) and 4-hydroxy-N-desmethyl-tamoxifen (endoxifen) show the highest anti-estrogenic potency (Lim et al., 2006). As the steady state plasma concentration of endoxifen is approx. 10-times higher compared with levels of 4-OH-Tam, endoxifen is considered to be the major active tamoxifen metabolite (Stearns et al., 2003).

The bioactivation of tamoxifen to its main active metabolite endoxifen comprises two permutable steps, namely N-desmethylation and 4-hydroxylation (Figure 1; Brauch et al., 2009). The rate limiting step in the major route is the 4-hydroxylation via CYP2D6 (Desta et al., 2004).

CYP2D6 is encoded by a highly polymorphic gene with more than 100 different alleles (Sim and Ingelman-Sundberg, 2010). It has been shown that a number of its genetic variants affect the activity of the enzyme and may decrease the levels of the active metabolites and thus result in decreased therapeutic response to tamoxifen (Goetz et al., 2005; Nowell et al., 2005; Schroth et al., 2009). According to the enzyme activity, four different phenotypic alleles including PM, IM, EM, and UM have been categorized (Bradford, 2002; Daly, 2003; Zanger et al., 2004; van Schaik, 2005; McGraw and Waller, 2012). There is a body of evidence that subjects carrying dysfunctional alleles of CYP2D6 have less benefit from tamoxifen therapy because of lower plasma concentrations of the active metabolite endoxifen (Jin et al., 2005; Lim et al., 2007; Madlensky et al., 2011; Karle et al., 2013; Saladores et al., 2015).

The frequency of CYP2D6 alleles varies in different populations. Various studies indicate that the prevalence of normal alleles (^∗1 and ^∗2) in the European population is around 70%. Twenty-six percent of the population in Europe is carrying one PM (^∗3, ^∗4, ^∗5, and ^∗6) allele and the most common PM allele is CYP2D6^∗4. However, the frequency of PM alleles in the Asian population is less than 1%. In contrast, in the Asian population between the reduced function alleles (^∗10, ^∗17, and ^∗41), CYP2D6^∗10 is significantly more frequent by about 40–50% compared to Europeans (Bradford, 2002).

Regarding the Iranian population and the ethnic background few studies indicate the most relevant CYP2D6 alleles (Kouhi et al., 2009; Hashemi-Soteh et al., 2011; Bagheri et al., 2015). Accordingly, we decided to genotype the breast cancer patients of our study for the CYP2D6 alleles ^∗1, ^∗2, ^∗4, ^∗5, ^∗6, ^∗10, ^∗17, and ^∗41 as well as the gene duplication to comprehensively cover CYP2D6 alleles representative of the Iran population (Isfahan province). Based on their CYP2D6 alleles, the patients were assigned to different genotype groups discriminating for CYP2D6 phenotypes by CYP2D6 activity scores (AS) (Gaedigk et al., 2008). Plasma concentrations of tamoxifen and its metabolites following a standard tamoxifen dose of 20 mg/day were correlated with CYP2D6 phenotype. In patients with an activity score of one or less, the standard tamoxifen dose was adjusted accordingly to 40 mg. Follow-up after 4 and 8 months after dose adjustment was performed and plasma levels of tamoxifen and its metabolites as well as adverse events were assessed.

There is increasing evidence that personalized therapy in oncology contributes significantly to better treatment outcome with a tolerable safety (Meyer et al., 2013; Relling and Evans, 2015). Recently, genotyping of CYP2D6 in patients undergoing tamoxifen therapy for premenopausal and postmenopausal breast cancer patients has been proposed to improve treatment outcome (Province et al., 2014;Goetz et al., 2018). Previous clinical trials in breast cancer showed first evidence that CYP2D6 dose-adjusted treatment of tamoxifen resulted in significantly increased plasma concentrations of endoxifen, the major active metabolite of tamoxifen (Barginear et al., 2011; Irvin et al., 2011; Kiyotani et al., 2012; de Duenas et al., 2014; Dezentjé et al., 2015; Fox et al., 2016; Hertz et al., 2016).

To the best of our knowledge currently no data are available demonstrating the feasibility of individualized CYP2D6 guided tamoxifen therapy in an Iranian population of breast cancer patients. Therefore, we initiated a prospective clinical pilot study to adjust tamoxifen dosage in ER⁺ breast cancer patients with Iranian ethnic background. Considering major differences of CYP2D6 allele frequencies world-wide (Zanger and Schwab, 2013), genotyping of CYP2D6 was performed covering CYP2D6 alleles relevant for the Middle East population (Sistonen et al., 2007; Bagheri et al., 2015). Based on our study protocol and in agreement with previous studies elucidating the impact of CYP2D6-guided tamoxifen dose adjustment in patients with compromised CYP2D6 activity (PM and IM), plasma concentrations of tamoxifen and selected metabolites were quantified before, after 4 and 8 months of dose escalation. In addition to DM-Tam, (Z)-endoxifen, and 4-OH-Tam which are suggested to be clinically relevant, we comprehensively measured other tamoxifen metabolites by a previously established and validated sensitive and highly specific LC-MS/MS method (Mürdter et al., 2011). This method includes various hydroxylated metabolites and glucuronides (Figure 1; Brauch et al., 2009).

At baseline, the inter-individual variability of tamoxifen and selected metabolites in the Iranian population was similar to data in other ethnic populations of breast cancer patients (e.g., Europe, America and Asians) (Saladores et al., 2015). Regarding (Z)-endoxifen, our data confirms the close relationship between CYP2D6 genotypes and plasma concentrations proposed by other studies (Lim et al., 2011; Antunes et al., 2015; Schroth et al., 2017; Woo et al., 2017). Due to the small number of patients who were enrolled in our pilot study to evaluate dose escalation of tamoxifen in Iranian breast cancer, data analyses have been performed by combining IM and PM patients treated with 30 and 40 mg tamoxifen per day, respectively. Generally our data demonstrate a significant increase in plasma concentrations of tamoxifen and its metabolites following dose escalation for 4 and 8 months. Of note, (Z)-endoxifen plasma concentrations in patients with compromised CYP2D6 activity and dose escalation were similar to those in patients with CYP2D6 EM phenotype at standard dosage.

These data are in line with previous studies demonstrating that dose escalation of tamoxifen in CYP2D6 compromised breast cancer patients result in higher plasma levels of the active metabolite endoxifen. In the study by Irvin et al. (2011) in a US American population the concentration of endoxifen was increased in both, the IM and PM patients, after doubling the tamoxifen dosage to 40 mg/day for 4 months. The endoxifen plasma levels in the IM group receiving 40 mg/day tamoxifen were similar to the EM group treated with 20 mg/day (21.8 vs. 29.2 ng/ml; p = 0.84). In contrast, patients in the PM group failed to reach similar endoxifen plasma levels (12.9 vs. 29.2 ng/ml; p = 0.016) despite dose escalation. Barginear et al. (2011) increased the tamoxifen dosage from 20 to 30 mg/day based on plasma endoxifen concentration (<40 nmol/l) in US patients with various ethnic background and IM CYP2D6 phenotype. Dose adjustment resulted in elevated levels of (Z)-endoxifen, (Z)-4′-OH-DM-Tam, (Z)-4-OH-Tam, and (Z)-4′-OH-Tam. Noteworthy, in line with our data lower plasma concentrations of (Z)-4′-OH-DM-Tam in patients with an EM CYP2D6 activity score compared to IM/PM patients were reported, indicating a metabolic shift from (Z)-4′-OH-DM-Tam to endoxifen. In an Asian population Kiyotani et al. (2012) investigated dose escalation of tamoxifen to 30 and 40 mg/day in patients who were heterozygous and homozygous for CYP2D6 IM alleles, respectively. After 8 weeks, concentrations of tamoxifen, N-DM-Tam, 4-OH-Tam and endoxifen were increased in IM patients resulting in endoxifen and 4-OH-Tam levels similar to those detected in EM patients. After enrollment of the patients in the CYPTAM study, tamoxifen dose escalation for IM and PM patients was calculated by the formula: 20 mg × (average endoxifen serum concentration in EMs divided by the patient’s baseline endoxifen serum concentration) and led to tamoxifen doses between 30–100 mg for the IMs and 60–120 mg for PMs (Dezentjé et al., 2015). Dose adjustment resulted in elevated levels of endoxifen in both PMs and IMs. There was no significant difference with the mean endoxifen level in EMs (=33.7 nM), and IMs after dose escalation (=30.3 nM; p = 0.20), although the mean endoxifen level in PMs after dose escalation was significantly lower (=27.3 nM; p = 0.03).

Building on the success of the pilot study by Irvin et al. (2011) and Hertz et al. (2016) expanded enrollment to 500 patients and again they doubled the tamoxifen dosage from 20 to 40 mg/day in both, the IM and PM patients. After 4 months, the endoxifen plasma levels in the genotype-dosed IM patients, had risen by 48% to 10.74 ng/mL, which was no longer different from, and was in fact nominally greater than, the endoxifen concentrations in EM/UM patients (9.30 ng/mL, p = 0.08). Endoxifen concentrations in genotype-dosed PM patients also rose by approximately 61% to 5.52 ng/mL but remained significantly lower than in EM/ UM patients (9.30 ng/ml; p = 0.009). In the Australian TADE study patients with basal endoxifen concentrations <30 nmol/l underwent a sequential dose escalation scheme with a maximum tamoxifen dose of 60 mg/day resulting in endoxifen concentrations >30 nmol/l in 76% (93 of 122) of participants (Fox et al., 2016). Of note patients with a PM phenotype, i.e., carrying two non-functional CYP2D6 alleles were unable to achieve the target endoxifen concentration of 30 nmol/l despite dose escalation.

Independent from CYP2D6 genotype increased dosing of tamoxifen to 30 and 40 mg/day resulted in significantly higher plasma levels of tamoxifen compared to tamoxifen standard treatment (20 mg/day).

Since altered pharmacokinetics may result in a higher rate of ADR, we systematically monitored tamoxifen specific side effects as well as laboratory parameters indicating liver and kidney toxicity. No severe side effects (e.g., thromboembolism) occurred during dose escalation. Moreover there was no significant increase in the occurrence of hot flashes and only slightly increased frequencies of vomiting and agitation. These data corroborate previous studies demonstrating that doses escalation of tamoxifen in CYP2D6 compromised breast cancer patients is well tolerated (Barginear et al., 2011; Irvin et al., 2011; Kiyotani et al., 2012; Dezentjé et al., 2015; Fox et al., 2016; Hertz et al., 2016).

Besides CYP2D6, CYP2C9, and CYP2C19 as well as some phase 2 enzymes have also been reported to participate in tamoxifen pharmacokinetics, however, with far less impact than CYP2D6 (Mürdter et al., 2011; Saladores et al., 2015; Marcath et al., 2017; Sutiman et al., 2016). Therefore, this study was focused on CYP2D6.

A limitation of our study is the number of patients included. Previous studies have shown a low frequency of CYP2D6 PM subjects in the Iranian population (Kouhi et al., 2009; Hashemi-Soteh et al., 2011; Bagheri et al., 2015). In line with these findings, the number of PM patients was not high enough to analyze those patients separately. Moreover no outcome data are available.

Taken together, for the first time we showed that the concept of dose escalation of tamoxifen in CYP2D6 compromised breast cancer patients is feasible in an Iranian population resulting in significantly higher plasma concentrations of the active metabolite (Z)-endoxifen. Our study supports ongoing and future activities to comprehensively elucidate the impact of CYP2D6- guided dose escalation on clinical outcome in prospective clinical trials. The strategy of dose escalation in patients with compromised genetics for drug metabolizing enzymes regarding other drugs as tamoxifen appears to be a promising approach which warrants future trials.

This study was carried out in accordance with the recommendations of the Ethical Committee of Isfahan University of Medical Sciences with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the Ethical Committee of Isfahan University of Medical Sciences.

ZK, ZB, PN, HK, RS, and ManS designed the project and involved in doing the experiments. MatS and TM performed the measurements of tamoxifen metabolite concentrations. ES performed the CYP2D6 CNV analysis. ZK, ZB, MatS, TM, ManS, and PN analyzed the data and wrote the manuscript. FM was the clinician of the project, who helped us in designing the project, sample collection, and data analysis. All authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.