The breast lobules are the structures that produce milk in lactating women, and are also the primary site of origin for breast cancer. Age-related lobular involution (LI) is a physiological process in which the epithelial tissue of the breast lobules gradually regresses with age, corresponding with elimination of the need for milk production beyond the child-bearing years. Analysis of our cohort of more than 14,000 women with benign breast disease (BBD) revealed that LI is strongly associated with reduced risk of subsequent breast cancer, and thus LI constitutes a natural mechanism for prevention of breast cancer. However, more than 40% of postmenopausal women with BBD have not completed LI, and these women are at substantially greater risk of developing breast cancer compared to women of similar age whose involution process has completed. In this application, we propose experiments to identify critical effectors of LI and to determine how these can be used to improve breast cancer risk assessment for women who have not undergone LI. In Aim 1, we will define the mediators that underlie the link between delayed LI and a high risk cellular phenotype. We will use a newlygenerated panel of very early passage primary human mammary epithelial cells (HMECs) derived from women with BBD and varying degrees of LI, in combination with humanized mouse xenograft models in which the LI status of the HMECs is manifest. In Aim 2, we will identify mediators associated with ongoing vs delayed LI in a newly-defined clinical cohort of peri- and post-menopausal women for whom multiple, sequential benign biopsies are available. In Aim 3, biomarkers discovered in Aims 1 and 2 (plus other promising biomarkers) will be evaluated for association with subsequent breast cancer in a patient cohort of postmenopausal women. The cohorts interrogated in Aims 2 and 3 will be drawn both from the Mayo Clinic BBD cohort, which is primarily composed of White/Caucasian women, and the Detroit BBD cohort, which is composed of Black/African-American women, which will allow us to evaluate race as a biological variable in our analyses. By providing insight into the processes that control LI and by developing tractable experimental models in which to evaluate the effects of modulating those processes, our work will provide a better understanding of LI delay and its link to breast cancer and will point toward methods for inducing LI in women for whom this process is stalled or delayed, as a physiologically-derived method for breast cancer prevention.