Our primary area of research focuses on the relationship of hormones to cancer, particularly breast cancer and how health behaviors and genetics may affect cancer risk by modifying hormone levels. Breast cancer is a hormonally dependent cancer. Our prospective study of endogenous serum hormones and breast cancer risk in the Columbia, MO Serum Bank cohort (Dorgan JF et al, Cancer Epidemiol Biomarkers Prev, 1996;5:533) was among the first of several prospective epidemiologic studies to show an increased risk of breast cancer among postmenopausal women with elevated serum estrogen and androgen concentrations. In the pooled analysis of the world-wide data (Breast Cancer Collaborative Group, J Natl Cancer Inst 94:606, 2002), in which we participated, postmenopausal women with elevated serum estrogen and androgen levels were at an approximate 2-fold excess risk of developing breast cancer. Biomarker studies not only increase our understanding of cancer etiology, but also could identify women at increased risk. Furthermore, biomarkers associated with risk can be used as endpoints in trials to evaluate effectiveness of interventions to prevent cancer and improve outcomes. We continue to conduct research aimed at identifying biomarkers associated with breast cancer risk.

Numerous health behaviors including diet, alcohol ingestion, and physical activity could influence breast cancer risk through effects on serum hormones. In earlier cross-sectional analyses, we found that diet (Dorgan JF et al, Am J Clin Nutr 1996;64:25), alcohol ingestion (Dorgan JF et al, Cancer Causes Control, 1994;5:53; Garcia Closas M et al. Int J Cancer, 2002;102:172) and physical activity (Madigan MP et al. Cancer Causes Control, 1998;9:199) are associated with serum hormone levels in women. We more recently conducted a controlled feeding study of alcohol’s effects on serum biomarkers and found that consumption of 1-2 drinks per day increased serum levels of estrogens and androgens in postmenopausal women (Dorgan JF et al, J. Natl. Cancer Inst, 2001;93:710). Alcohol ingestion altered serum levels of several other serum biomarkers including insulin-like growth factor binding protein 3 (Lavigne JA et al. Am J Clin Nutr, 2005;81:503), leptin (Roth M et al, J Natl Cancer Inst 2003;95:1722), isoprostanes (Hartman TJ et al. Eur J Clin Nutr, 2005;59:161) and vitamin B12 (Laufer EM et al. Eur J Clin Nutr, 2004;58:1518). Thus, health behaviors likely contribute, at least in part, to breast cancer risk via effects on metabolism.

The association of body weight with breast cancer risk is complex. Whereas obesity decreases risk of breast cancer before menopause, it increases risk after menopause. The increased risk associated with breast cancer is partly explained by higher serum estrogens in heavier postmenopausal women. In a pooled analysis of prospective studies (Endogenous Hormones and Breast Cancer Collaborative Group, J Natl Cancer Inst, 2003;95:1218), we found that adjustment for serum estradiol concentration attenuated the association of body mass index with breast cancer risk. The decreased risk of breast cancer in obese premenopausal women has been attributed to a higher frequency of anovulatory menstrual cycles in these women. However, this would not explain the lower risk of premenopausal breast cancer associated with obesity in young prepubescent children. In our longitudinal analyses (Baer HJ et al, Cancer Epidemiol Biomarkers Prev, 2007;16:1880), pre-pubertal obesity was associated with serum dehydroepiandrosterone sulfate (DHEAS) and sex hormone binding globulin (SHBG) levels during puberty but not with estrogens or progesterone. Whether the differences observed affect breast cancer development is unclear.

Because most breast development occurs during adolescence, our studies focus not only on adult exposures but also on adolescent exposures. We previously found that a reduced fat diet lowered serum estrogens and progesterone in adolescent girls (Dorgan JF, J Natl Cancer Inst, 2003;95:132). With funding from NCI, we currently are following-up this finding and are investigating the long-term effects of adolescent diet on biomarkers in young women that are consistently associated with breast cancer risk in adults including serum hormones, bone mineral density and breast density.

The adrenal androgens DHEA and DHEAS are consistently positively associated with breast cancer risk. We reported increased risks associated with elevated levels of these hormones in premenopausal women (Sturgeon SR et al, Cancer Causes Controlm 2004;15:45) and in postmenopausal women (Dorgan JF et al, Cancer Epidemiol Biomarkers Prev, 1996;5:533; Breast Cancer Collaborative Group, J Natl Cancer Inst, 2002;94:606). Furthermore, we found that both the adrenals and ovaries contribute to serum androstenedione levels that are associated with breast cancer risk in postmenopausal women (Dorgan JF et al, Cancer Epidemiol Biomarkers Prev, 2001;10:407). Adrenal steroid production is relatively stable in individual women over time suggesting that genetics likely plays a major role in regulating adrenal steroidogenesis. In a DOD funded pilot study, we found that response to corticotrophin (ACTH) stimulation varied by genotype of several genes involved in adrenal steroidogenesis (Dorgan JF et al, Endocrine submitted 2009). One participant who was the sole carrier of specific haplotypes in the ACTH receptor gene (MC2R) and in the steroid acute regulatory protein (StAR) gene exhibited a particularly exuberant response as evidenced by rapid increases in cortisol, DHEA and progesterone to higher levels than observed for other participants. Three other participants carried the putative functional T-34C SNP in CYP17. All three exhibited enhanced cortisol responsivity to ACTH, while 2 participants also had an increased DHEA response.

Adrenal steroidogenesis is regulated by the hypothalamic-pituitary-adrenal (HPA) axis. The HPA axis regulates homeostasis under basal conditions. The HPA axis also is the primary effectors of the physiologic response to psychological, social, metabolic, and physical stressors. Thus the HPA is a key determinant of physiological responses to diverse environmental exposures. In an NIH funded pilot study, we are beginning to evaluate the effects of psychosocial, behavioral and genetic characteristics on stress responsivity.