Integrative Molecular Cancer Epidemiology 56-POS Glutathione S-transferase M1 genetic polymorphism is associated with salivary 17-β estradiol levels. A study based on hormonal profiles from entire menstrual cycles A.-S. Furberg1,2, A. McTiernan3, P.T. Ellison4, K.W. Makar3, A. Iversen2, A. Emaus5, I. Thune5 1 Department of Microbiology and Infection Control, University Hospital North Norway, Tromsø, Norway; 2 Institute of Community Medicine, University of Tromsø, Tromsø, Norway; 3 Cancer Prevention Program, Cancer Fred Hutchinson Cancer Research Center, Seattle, WA, USA; 4 Department of Anthropology, Harvard University, Cambridge, MA, USA; 5 The Cancer Center, Ullevål University Hospital, Oslo, Norway Introduction: Studies suggest that enzymes involved in the estrogen metabolic pathway are susceptibility factors for breast cancer; however, there is limited direct evidence in relation to estradiol levels. The glutathione-S-transferase M1 enzyme (GSTM1) enzyme is involved in estrogen peroxidation, and a shortage seems to enhance estrogen exposure. A homozygous deletion of the GSTM1 gene leads to total absence of enzyme activity, and studies indicate that an association with breast cancer risk may be modified by body mass. So far, parallel studies on intermediate biomarkers (i.e. estradiol) are largely missing. Thus, in the present study, we assessed the association between a GSTM1 genetic polymorphism (null/rs1065411) and daily salivary levels of 17-β estradiol throughout one entire menstrual cycle and evaluated modification of genotype effects according to metabolic profile in 190 premenopausal women aged 25-35 years. Methods: The women participated in the Energy Balance and Breast Cancer Aspects study (EBBA-I, 2000-2002) in Tromsø, North Norway (inclusion criteria: healthy, non-pregnant, non-lactating, no current use of exogenous hormones). Salivary levels of 17-β estradiol were assessed by radioimmunoassay (The Reproductive Ecology Laboratory, Harvard University). Height and weight were measured. Genomic DNA was extracted from EDTA whole blood. GSTM1 genotyping was done on the ABI PRISM® 3100 Genetic Analyzer (Fred Hutchinson Cancer Research Center). The GSTM1 assay distinguishes between the GSTM1*A allele (G nucleotide, Lys at aminoacid 173), the GSTM1*B allele (C nucleotide, Asn at aminoacid 173), and the homozygous gene deletion. Regression models were used to study the relation between genetic (null genotype versus all other alleles combined) and metabolic susceptibility factors, and levels of 17-β estradiol. Results: Mean body mass index (BMI) was 24.4 kg/m2. The homozygous gene deletion of GSTM1 had a prevalence of 52%, and the frequency of the null genotype was similar in heavy women (BMI ≥ 26.3 kg/m2, upper quartile) as compared to other women (p = .18). The null genotype was associated with lower levels of 17-β estradiol by cycle day compared to other alleles (p = .04). However, the association was confined to women that were normal weight to slightly overweight (p for interaction = .01). In women with BMI < 26.3 kg/m2, the null genotype was associated with significantly lower levels of 17-β estradiol by cycle day compared to other alleles (p = .02) and an 18% reduction in overall average level of 17-β estradiol throughout the cycle (p = .02), while in heavy women no association with estradiol levels was observed. Conclusion: In our study, GSTM1 genotype was a significant determinant of circulating estradiol levels in premenopausal women without marked adiposity. Thus, interestingly our results strongly support that the effect of body mass on estradiol levels and subsequent breast cancer development may be modified by GSTM1 genotype. 57-POS Prognostic significance of matrix metalloproteinases 1 and 3 polymorphisms in squamous cell carcinoma of the head and neck R.V.M. López1, M.A. Zago2, J. Eluf-Neto3, W.A. Silva-Jr2, D.L. Zanette2, J.E. Levi4, M.B. Carvalho5, M.P. Curado6,7, P. Boffetta7, V. Wünsch-Filho1 1 Department of Epidemiology, School of Public Health, University of São Paulo, São Paulo; 2Molecular Genetics Laboratory, School of Medicine, University of São Paulo, Ribeirão Preto; 3Department of Preventive Medicine, School of Medicine, University of São Paulo, São Paulo; 4 Laboratory of Virology, Institute of Tropical Medicine, University of São Paulo, São Paulo; 5Department of Head and Neck, Heliópolis Hospital, São Paulo; 6Araújo Jorge Hospital and Cancer Registry of Goiania, Goiânia; 7International Agency for Research on Cancer, Lyon Background: Matrix metalloproteinases 1 (MMP-1) and 3 (MMP-3) have been associated with occurrence and progression of several types of cancer. We examined the effect of MMP1 -1607, MMP1 -519, and MMP3 1171 polymorphisms on the prognosis of squamous cell carcinoma of the head and neck (SCCHN). Materials and Methods: A total of 515 SCCHN patients were enrolled in a hospital based case-control study conducted in three Brazilian cities from November 1998 to December 2002. They were followed until the end of June 2005. Information on patient vital status was
207 obtained from medical records and through linkage with population mortality and cancer registry database. Tobacco smoking and alcohol consumption habits were obtained through face-to-face interviews using a structured questionnaire. Genotyping studies were performed using realtime polymerase chain reaction (PCR). Kaplan-Meier estimates and Cox proportional hazard models were used to examine overall and diseasespecific survival of SCCHN patients according to specific anatomical sites: oral cavity and oropharynx; hypopharynx; and larynx; and MMP1 -1607, MMP1 -519 and MMP3 -1171 polymorphisms. Results: MMP3 -1171 5A/6A and advanced clinical stage (CS IV) were predictors of poor survival (overall and disease-specific) in oral cavity and oropharynx cancers. Advanced clinical stage and high alcohol consumption were independent factors of poor overall and disease-specific survival in hypopharynx cancer. Advanced clinical stage, high alcohol consumption and low education were independent predictors of overall and disease-specific survival in larynx cancer. Conclusions: Of genetic polymorphisms studied, only MMP3 -1171 5A/6A was related with poor survival in oral cavity and oropharynx cancers. Keywords: MMP-1; MMP-3; head and neck cancer; survival analysis; cancer prognosis 58-POS Genetic polymorphisms in the cytochrome gene family and the risk of childhood acute lymphoblastic leukemia G. Scelo1, A.P. Chokkalingam1, C. Metayer1, N. Guha1, L.F. Barcellos1, J.L. Wiemels2, J.K. Wiencke2, P.A. Buffler1 1 School of Public Health, University of California at Berkeley, Berkeley, USA; 2Department of Epidemiology and Biostatistics, University of California at San Francisco, San Francisco, USA Although the most common of childhood cancers, leukemia remains a rare disease in 0-14 year-olds and little is known about genetic susceptibilities involved in the disease risk. Polymorphisms in cytochrome P450 gene family (CYPs) have been inconsistently associated with leukemia risk in children. The Northern California Childhood Leukemia Study (NCCLS) is a population-based, case-control study that has been recruiting childhood leukemia cases from 9 hospitals covering 35 counties in California since 1995. Controls matched to cases for sex, age, Hispanic status, and mother’s race have been randomly selected from the Californian birth registry and buccal cell specimens have been collected from both cases and controls. Genotyping of 60 tag single nucleotide polymorphisms (SNPs) in 12 CYP genes (CYP1A1, CYP1A2, CYP1B1, CYP2A6, CYP2B6, CYP2C19, CYP2C8, CYP2D6, CYP2E1, CYP2F1, CYP3A4, and CYP3A5) has been conducted so far in 385 cases of acute lymphoblastic leukemia (ALL) and 454 controls. Genotyping failed for 5 SNPs (8.3%) and eight children with call rates <80% were excluded, leaving 380 cases and 451 controls in the present analyses. Unconditional logistic regression analysis was used to analyze the 55 successful SNPs. All results were adjusted for matching variables and stratified for ethnicity (non-Hispanic Whites and Hispanics). Overall, gene variants in rs3786552 and rs3745274 (CYP2B6), in rs7077618, rs7073968, rs1934954 and rs10509681 (CYP2C8), and in rs12333983 (CYP3A4) were significantly associated with ALL risk in both ethnic groups (log-additive p= 0.044, 0.048, 0.016, 0.020, 0.009, 0.016, and 0.028, respectively). Significant associations were also found in nonHispanic Whites only with all three genotyped SNPs in CYP1A2 (rs11854147, rs11072508, rs2470890: p=0.032, 0.030, and 0.020, respectively), and in Hispanics only with two SNPs in CYP3A5 (rs776746, p=0.013; and rs10242455, p=0.048). Although subject to multiple testing issues, our analyses support an association between polymorphisms of cytochrome genes and ALL risk in children. Further analyses are warranted to investigate these associations, as well as the potential interaction of genetic susceptibilities with environmental exposures, specifically with polycyclic aromatic hydrocarbons, which are metabolized through the cytochrome P450 pathway. 59-POS Common breast cancer predisposition alleles are associated with breast cancer risk in BRCA1 and BRCA2 mutation carriers O.M. Sinilnikova1,2, A.C. Antoniou3, A.B. Spurdle4, S. Healey4, R.K. Schmutzler5, S.L. Neuhausen6, F.J. Couch7, D. Stoppa-Lyonnet8, G. Chenevix-Trench4, D.F. Easton3 on behalf of Consortium of Investigators of Modifiers in BRCA1/2 (CIMBA) 1 Unité Mixte de Génétique Constitutionnelle des Cancers Fréquents, Hospices Civils de Lyon / Centre Léon Bérard, Lyon, France; 2Laboratoire de Génétique Moléculaire, Signalisation et Cancer UMR5201 CNRS, Université Claude Bernard, Lyon, France; 3Cancer Research UK, Genetic Epidemiology Unit, Department of Public Health and Primary Care, University of Cambridge, United Kingdom; 4Queensland Institute of Medical Research, Brisbane, QLD, Australia; 5Department of Obstetricts Gynaecology, Division of Molecular Gynaeco-Oncology, University of Cologne, Germany; 6Department of Epidemiology, University of California