The successful application of pharmacogenetics requires a full understanding of the expression of genotype as phenotype. A large number of genes code for proteins that mediate response to medicines, and while it is clear that variations in one gene can alter the clinical response to a medicine significantly, this is rare. We propose to study multiple genetic influences on the clinical pharmacology of tamoxifen and aromatase inhibitors. Our goal is to define multiple genetic influences on the action of this drug as a model for other drugs in the class, where multiple genetic variations are likely to alter pharmacologic responses. An interdisciplinary group of investigators will use established pharmacogenetics and analytical core laboratories and the resources of the Indiana University School of Medicine and University of Michigan Cancer Centers to study genetic influences on the metabolism, pharmacokinetics, efficacy and toxicity of tamoxifen. While tamoxifen has been shown to be metabolized by genetically polymorphic cytochrome P450 enzymes in human liver microsomes in vitro, the hypothesis that mutations in the genes coding for these enzymes might alter the drug's metabolism, effects or toxicity has never been tested in vivo, either in normal volunteers or in women with breast cancer.

Our work involves the following broad specific aims:

1. To identify common genetic variants of the human estrogen receptors and important nuclear coactivators and repressors of these receptors using a combined bioinformatic and direct sequencing approach. We will resequence critical portions of the ERα gene and the ERβ gene and identify the haplotype blocks that encompass these clinically important SNPs. Haplotype-specific tagSNPs will be identified for use in the following specific aims designed to test functionality and association with phenotypic effects.



2. To test the hypothesis that ERα and ERβ variants alter gene expression or function using in vitro assays. The expression and processing of both ERs is highly variable, yet the mechanisms regulating this are poorly understood. We will use a variety of functional assays to test effects of important estrogen receptor variants in vitro. We will also test the association of the genetic variants with the expression of alternatively spliced variants in vivo. Results from these studies will be used to focus our interest on the most functionally important variants in our clinical studies of tamoxifen and aromatase inhibitors.


3. To test the contribution of ERα and ERβ variants identified during specific aim 1 and 2 to tamoxifen response in the clinical trial of tamoxifen pharmacogenetics already conducted. The objective of this aim will be to use the genotype/haplotype information from Aim 1 to more accurately define the association of tamoxifen responses with genetic variations in these receptors and coregulators. The working hypothesis is that haplotype blocks in these genes will be useful predictors of tamoxifen responses. This analysis will provide more detailed understanding of the role of genetics in phenotypic responses to tamoxifen.



4. To characterize the involvement of genetically polymorphic drug metabolizing enzymes in the human metabolism of the available aromatase inhibitors: letrozole, exemestane and anastrozole in vitro. Letrozole, anastrozole and exemestane are all predominantly cleared by metabolism. However, data indicating their specific metabolic catalysts are very limited at present. This has hindered identification of factors that might influence pharmacokinetics and response to aromatase inhibitors. Here, we will use cellular fractions of human livers and expressed enzymes to unequivocally identify the major metabolic routes and enzymes involved in the metabolism of letrozole, exemestane and anastrozole. This approach will allow us to pin point important metabolic candidate genes, and these data will be used to test focused associations between genetic variations in these genes and the phenotypic outcomes collected in Specific Aim 5.



5. To test the hypothesis that variants in candidate genes identified in aims 1 - 4 are associated with well curated phenotypic outcomes, including estrogen metabolite concentrations, pharmacokinetics, hot flashes, breast density, bone metabolism and serum lipid subfractions in breast cancer patients receiving exemestane and letrozole. We will enroll 500 women into a trial in which women with breast cancer are randomized to take exemestane or letrozole. The associations that we identify will be used to generate mechanistic hypotheses and to identify genotypic patterns to be examined in larger clinical trials.