Intestine Bacteria and Breast Cancer Risk

Background/Significance: Commensal microbes (the microbiota), particularly in the gut, are required for human health and are postulated to affect cancer risk through several mechanisms. This proposed study builds upon a pilot within the NCI Division of Cancer Epidemiology and Genetics (DCEG) that identified highly acceptable methods for collecting fecal specimens and significant correlation between fecal microbial beta-glucuronidase activity and a marker of breast cancer risk, urine levels of estrogens. The proposed study will determine the correlation between levels of fecal enzyme activities and systemic estrogens in a random sample of postmenopausal women at Kaiser Permanente Colorado (KPCO). Demonstration of an association between the fecal microbiota and systemic estrogens would help to motivate future studies of how microbes affect cancer risk.

Objectives: The main objective is to characterize the fecal microbiota and its association with levels of systemic estrogens among randomly sampled postmenopausal female members of a health maintenance organization (HMO). A secondary objective will determine the proportions of women who consent and then provide questionnaire data, a fecal specimen to characterize the microbiota, and a urine specimen to quantify systemic estrogens; and differences between participants (N=60) and non-participants. A third objective will determine whether consecutive, newly diagnosed postmenopausal breast cancer cases have similar participation rates, fecal microbiota characteristics and urine estrogen levels compared to the randomly sampled women.

Methods: We will randomly sample from the KPCO population of approximately 50,000 women ages 55-69 who have recently had a negative screening mammogram. An invitation packet (letter, information pamphlet, consent form, eligibility questionnaire, and opt-out postcard) will be sent in batches of 100, up to a maximum of 500 women. The consenting women will receive a second packet with a cancer risk-factor questionnaire, a link to the on-line Block brief food frequency questionnaire, and a specimen collection kit (with illustrated instructions) for collecting a fecal and urine specimen. Participants (minimum 60, maximum 80) will ship the specimens to the DCEG repository. As amended, consecutive newly diagnosed breast cancer cases (target N=60) will be enrolled in the oncology clinic prior to definitive surgery, with the same eligibility criteria as for the random sample. For contemporaneous controls, participants in the randomly sampled, mammogram-negative population will be recruited back (target N=60). Urine estrogens will be quantified by DCEG at NCI Frederick. Fecal microbial classification and enzymatic expression and activity will be performed at the University of Maryland Medical School (UMMS).

Analyses: Data will be analyzed and summarized for publication with representatives from KPCO, DCEG and UMMS. Simple proportions will be used to estimate participation rates. Extant electronic records at KPCO will be used to compare participants to non-participants, overall and by pre-specified subgroups (5-year age groups, race/ethnicity, length of KPCO enrollment), with descriptive statistics and logistic regression. For the primary objective, fecal microbial 16S rRNA pyrosequences will be classified by phylogenetic and principal components analyses, while estrogen levels and Beta-glucuronidase RNA expression and enzymatic activity levels will be categorized using log standard deviations. The study will have 80% power to detect a 17% increase in estrogen level per 2.4-fold increase in glucuronidase activity and 80% power to detect a 3-fold case-control difference in above-median microbiome alpha diversity.