We investigated the interaction of cytochrome P450 2C19 (CYP2C19) and CYP2D6 genotype on clinical outcome in tamoxifen treated breast cancer patients and evaluated the potential of modulation by menopausal status.
We investigated a cohort of breast cancer patients prescribed tamoxifen for 5 years that had been previously genotyped for CYP2D6 polymorphism. In the current study, we extended genetic analyses to encompass also CYP2C19 polymorphism by addressing presence of the defective CYP2C19*2 and the ultra-rapid CYP2C19*17 allele. Kaplan-Meier curves and Cox regression models of 306 patients on tamoxifen treatment for a minimum of 1 year were used to analyze the effect of genotype on relapse-free survival as the primary endpoint.
Through genotype-predicted CYP2C19 (increased↑, normal →, and decreased↓) and CYP2D6 (below↓ and above↑ 50%) phenotypes and combinations thereof, we show here that the group with worst outcome and highest risk of relapse is that of 2C19↑-2D6↓. The 2C19↑ -2D6↓ subgroup was found to be associated with a hazard ratio (HR) of 2.94 [98.75% multiplicity corrected confidence interval (CI); 1.24 - 6.97] when adjusting for age, Nottingham prognostic index and adjuvant chemotherapy, and compared to the remainder of patients. This can be related to the lower adjusted HR of 2.13 (95% CI; 1.23-3.68) when using CYP2D6 genotype alone. Furthermore, the higher risk of relapse for patients with the 2C19↑ -2D6↓ genotype-predicted phenotype is greatly enhanced in premenopausal patients (HR = 21.08, 98.75% multiplicity corrected CI; 2.60 - 171.27).
We hypothesize that poor bioactivation of tamoxifen in patients with low CYP2D6 activity and low levels of circulating female sex hormones through high CYP2C19 metabolism represent a tamoxifen treated breast cancer patient group that has worse clinical outcome. Furthermore, we found that the effect of this genotype-predicted phenotype combination on breast cancer relapse was enhanced in premenopausal women, where larger differences in circulating female sex hormones can be envisioned and thus genotype may have a greater effect.
Clinical trial identification
Legal entity responsible for the study
This work was supported by grants to Professor Jonas Bergh´s research group from the Swedish County Council, BRECT at the Karolinska Institute, the Swedish Cancer Society and The Knut and Alice Wallenberg Foundation.
All authors have declared no conflicts of interest.