• We provided the first evidence of a gene-environment interaction related to one of the single locus, high penetrance genetic causes of cancer, by showing that long-term use of oral contraceptives (OCs) impacts substantially on penetrance of BRCA1 and BRCA2. Because of the huge public health and scientific implications of this finding, Core members decided to pursue this preliminary finding with a more definitive study. However, the relative rarity of BRCA1 and BRCA2 mutations, combined with the strong possibility that a family history of breast cancer might substantially affect OC use patterns, led to major methodologic and study design challenges. The Study Design and Statistical Methodology Core led by Dr. Thomas created an entirely new, cost effective study design to address this issue (and similar gene x environment questions). Several agencies (NCI and CBCRP) are co-funding this important effort.

• Previous work by us in Shanghai demonstrated that aflatoxins are a major cause of liver cancer and that they interact strongly with hepatitis B virus infections in risk determination. This work was important because it was one of the early studies in which a molecular exposure biomarker was shown to predict cancer risk and also represents one of the few established synergisms between exposures in cancer causation. We have pursued similar work on isothiocyanates (ITCs, derived primarily from cruciferous vegetables) and lung cancer in this ongoing prospective study of diet and cancer in Shanghai. We have utilized a urinary biomarker of ITCs which we had previously validated as an index of dietary exposure in an independent study in Singapore. The work in Shanghai demonstrated a strong, inverse relationship between urinary ITCs and lung cancer risk. Moreover, as ITCs are metabolized by two glutathione S-transferases (GSTM1 and GSTT1) we predicted that these substantial chemopreventive effects of ITCs might be limited to those who are deficient (homozygous null) in these enzymes. This hypothesis was clearly supported. A manuscript describing these findings, the first study we know of demonstrating an interaction between genetic susceptibility and diet in relation to lung cancer risk, is "in review" in JNCI (London SJ, Yuan J-M, Chung F-L, Gao Y-T, Coetzee GA, Ross RK, and Yu MC. Isothiocyanates, glutathione S-transferase M1 and T1 polymorphisms and lung cancer risk: a prospective study of men in Shanghai, China). This work was presented at the first Center retreat and was the major stimulus for the formation of the research focus group on Phase I and Phase 2 enzymes.

• We have been conducting a large, molecular epidemiologic case-control study of bladder cancer in Los Angeles and Shanghai, China to determine why two populations with similar smoking habits, the main risk factor for bladder cancer, have more than a 3-fold difference in bladder cancer incidence. Our preliminary results have demonstrated that, as predicted, individuals with N-acetyltransferase 2 "fast" genotypes (but not NAT1) and GSTM1 null genotypes (but not GSTT1 or GSTM1) have an increase in bladder cancer risk, effects which are entirely limited to smokers. As these two enzymes are known to be involved in activation and deactivation of smoking-related arylamines, these findings are as we had predicted, and provide another strong example of a gene x environment interaction. A manuscript describing these findings is being prepared. These results were also presented at the Center retreat, and during discussion it was noted that for a specific NAT2 genotype, there was a wide range of "phenotype" as measured by caffeine metabolites. The idea was raised that this could be an epigenetic phenomenon, perhaps due to gene silencing by methylation, a suggestion which provided the impetus for the creation of a DNA Methylation research focus group. Methylation is an area of research in which several USC basic scientists have international reputations, so the Focus Group provides a mechanism for expanding the interdisciplinary nature of the Center.

• We published a detailed article in the "Perspectives" section of Cancer Research in which we described our body of work in creating a polygenic model of prostate cancer etiology (Ross et al., Cancer Res 58:4497-4504, 1998). This paper describes the methodology not only for building other such polygenic models for complex diseases like prostate cancer, but also describes how such models can readily incorporate environmental risk modifiers. In addition to providing a strategy that the Adult Cancer Research Core is now using for other cancers (breast, bladder, endometrium), this paper provided the model that the Center will be utilizing as it takes on asthma etiology and pathogenesis as a major research initiative and priority for the next five years. As implementation of the model (for cancer, asthma, or other diseases) will require SNP analysis of multiple genes in large numbers of research subjects, this work also has substantial implications for the Molecular Biology and Processing Utility Core.