Abstract Essential nutrients, other dietary constituents, and various foods have been shown to be associated with cancer risk in epidemiologic studies and in animal models of carcinogenesis. Modulation of phase I and phase II biotransformation enzymes is one mechanism by which diet is hypothesized to influence cancer risk. These enzymes are involved in the metabolism of exogenous compounds, including drugs, carcinogens, and dietary constituents, and endogenous compounds, such as steroid hormones. Many phytochemicals in plant foods are metabolized by biotransformation enzymes, and they also influence expression and activities of these enzymes. Particular botanical families of plant foods are often unique, rich sources of specific classes of phytochemicals. For example the allium or onion family contains high amounts of allyl sulfides; apiaceous or carrot-family vegetables are sources of furanocoumarins; certain legumes are major sources of isoflavones; and cruciferous or broccoli-family vegetables are the primary dietary sources of precursors of isothiocyanates and indoles. This presentation will focus on epidemiologic and intervention studies that have investigated associations between plant food, particularly cruciferous vegetable, intakes and cancer risk or biotransformation enzyme activity and will examine the impact of genetic variation on these responses. Many of the genes that encode for biotransformation enzymes are highly polymorphic and variants in both phase I and phase II enzymes have been shown to modify response to dietary exposures. Phase I metabolic activation of carcinogens is typically catalyzed by the cytochrome P450 (CYP) family of oxidation enzymes and other classes of enzymes, such as peroxidases (e.g., cyclooxygenases), and transferases (e.g., N-acetyltransferase). CYP1A2 activates several procarcinogens, including heterocyclic amines, and is also the main enzyme involved in caffeine metabolism. In a study of BRCA1 mutation carriers, a polymorphism in CYP1A2 had no overall effect on breast cancer risk; however, among women with at least one variant C allele (AC or CC), those who consumed caffeinated coffee had a 64% reduction in breast cancer risk, compared with women who never consumed coffee (OR, 0.36; 95% CI, 0.18-0.73) (1). In controlled feeding studies, cruciferous vegetable consumption increases CYP1A2 activity in humans in a dose-dependent manner, whereas apiaceous vegetables decreases activity of this enzyme (2,3), due to the inhibitory effects of the furanocoumarins (4). When participants were stratified by CYP1A2 genotype, induction of CYP1A2 activity by cruciferous vegetables was most pronounced in the CC genotype (3). Interestingly, when cruciferous vegetables are fed concurrently with apiaceous vegetables, the effect is similar to when eating a fruit- and vegetable-free diet. Polymorphisms in the GSTM1 and GSTT1 genes result in complete lack of GSTM1-1 and GSTT1-1 proteins, respectively. Studies show effect modification of GSTM1 and GSTT1 genotypes on the association between cruciferous vegetable intake (or isothiocyanate exposure) and the risk of several cancers, including lung, gastric, prostate, bladder, and colon. Investigators have hypothesized that GSTM1-null or GSTT1-null individuals have reduced metabolism of isothiocyanates and therefore greater and longer exposure to these compounds. Interestingly, two studies show that GSTM1-null individuals tend to have more rapid and higher percentages of sulforaphane excretion than individuals who have at least one intact allele (i.e., GSTM1-positive) (5,6) and another shows no differences in isothiocyanate excretion (7). Despite this apparent lack of a metabolic difference, several short-term interventions with cruciferous vegetables suggest that GSTM1-null individuals have a greater intermediate biomarker response to isothiocyanates. For example, two levels of cruciferous vegetables supplementation significantly increased serum GST-alpha, primarily in GSTM1-null/GSTT1-null men (8). Broccoli supplementation also conferred increased protection against hydrogen peroxide-induced DNA strand breaks and lower levels of oxidized DNA bases in peripheral blood mononuclear cells from smokers (9). We have observed similar genotype differences in relation to changes in biomarkers of inflammation with cruciferous vegetable feeding. In conclusion, results of observational and intervention studies point to the complex interactions among types of foods consumed, genetic variation, and modulation of xenobiotic metabolism. Effects of genotype must be considered within the context of diet and other exposures in order to understand the impact of genetic variation on cancer risk and opportunity for prevention.