PROJECT SUMMARY/ABSTRACT Basal-like carcinoma is a rapidly-progressing and highly variable subtype of breast cancer that arises spontaneously (often in African Americans) or in genetically predisposed women. Such tumors are believed to arise from the functional loss of the BRCA1 and TP53 tumor suppressors in uncommitted basoluminal progenitors of the breast. However, it has been challenging to dissect the origins of the disease for lack of appropriate tools, making it difficult to conceive of how the cell-state variability of premalignant mutants gives rise to basal-like breast cancer. The long-term goal of this work is to identify the critical cellular and molecular transitions underlying basal-like breast cancer genetics. The current application deploys a novel genetically engineered mouse model called mosaic analysis of double markers (MADM), which randomly deletes murine Brca1–Trp53 in transit-amplifying progenitors of the mammary gland. In MADM, stochastic deletion is genetically defined by coexpression of GFP, allowing locally expanded premalignant lesions to be visualized within the gland before the onset of basal-like disease. We found that premalignant expansion is accompanied by extensive recruitment of specific immune subsets, suggesting they play crucial roles in tumorigenesis. Our objective is to combine MADM with innovative methods for dissociation-free transcriptomics that will identify cell-state variabilities within mutant epithelial cells and the infiltrating immune lineages of a premalignancy. The hypothesis is that epithelial-cell plasticity and the stromal microenvironment coordinately diversify mutant lesions, revealing premalignant transcriptional states that ultimately progress to basal-like cancer in the breast and mammary gland. The aims of the proposal are: 1) To define shared premalignant trajectories of basoluminal diversification triggered by BRCA1–TP53 deficiency in mice and humans. 2) To deconvolve the immune heterogeneities that are locally paired with specific premalignant ecosystems for basal-like breast cancer. 3) To functionally validate cell states important for progression by using genetic, pharmacologic, and paracrine perturbations that homogenize intrinsic or extrinsic variability of premalignant cells ex vivo. Co-PIs Janes and Zong are thought leaders in their respective fields of intratumor cell-state heterogeneity and genetically engineered mouse modeling with a multi-year track record of collaboration. Together with a pair of senior clinicians, the team is poised to have a significant overall impact on our understanding of basal-like breast tumorigenesis.