Despite improvement in screening and early detection, breast cancer remains the second leading cause of cancer related deaths among women according to the American Cancer Society. These deaths can be almost entirely attributed to metastatic disease, which is far more difficult to treat than local disease. It is therefore critical to gain a deeper understanding of what drives breast cancer metastasis and how the cells that surround metastatic tumors, known as the metastatic niche, respond to breast cancer cells to facilitate or inhibit their outgrowth. In this work, we use single-cell RNA-sequencing (scRNA-seq) and custom analytical pipelines to characterize breast cancer metastasis from multiple angles. We start by looking at pre-neoplastic breast epithelial cells and investigate the conserved lineage relationships between each epithelial cell state using a pseudotime analysis pipeline. We next look at matched primary tumor and metastatic cells from triple-negative breast cancer patient-derived xenograft models and identify biomarkers of micrometastasis (very small, early stage metastatic tumors) using generalized linear models which we validate are prognostically useful for relapse-free survival. And finally, we characterize the response heterogeneity of microglia, the brain resident macrophage, to breast cancer brain metastasis, demonstrate that these responses are conserved in human microglia, and show that microglia facilitate tumor regression using a genetic depletion model. In all of these projects, we find that our newly defined cell states and tissue heterogeneity can be generalized across patients and models, suggesting that much of the noise seemingly inherent to breast cancer metastasis can be understood by asking the right questions. Further, by observing these systems at the single-cell level, we demonstrate the plasticity of breast epithelial cells in homeostasis, identify novel biomarkers and patient stratification opportunities for early breast cancer metastasis detection, and suggest new therapeutic routes for breast cancer brain metastasis patients.

Although metastasis remains the cause of most cancer-related mortality, mechanisms governing seeding in distal tissues are poorly understood. I hypothesized the initial stages of seeding at distal tissues (micrometastases) contain a unique transcriptome compared to primary tumor cells, enabling the identification of novel molecular targets to specifically block micrometastatic establishment. Here I established a robust method for identification of global transcriptomic changes in rare metastatic cells during seeding using single-cell RNA-sequencing and patient-derived xenograft (PDX) models of breast cancer. I found that both primary tumors and micrometastases display transcriptional heterogeneity, but micrometastases harbor a distinct transcriptome program conserved across PDX models that is highly predictive of poor survival in patients. Pathway analysis revealed mitochondrial oxidative phosphorylation (OXPHOS) as the top pathway upregulated in micrometastases, in contrast to higher levels of glycolytic enzymes in primary tumor cells, which was corroborated by flow cytometric and metabolomic analyses. Pharmacological inhibition of OXPHOS dramatically attenuated metastatic seeding in the lungs, which demonstrates the functional importance of OXPHOS in metastasis and highlights its potential as a therapeutic target to prevent metastatic spread in breast cancer patients.

Breast cancer brain metastasis (BCBM) is an increasingly prevalent clinical problem, due largely to improvements in treatment of primary tumors. Incidence of BCBM is rapidly fatal and difficult to treat, owing to the relative impermeability of the blood-brain barrier (BBB) to standard-of-care treatments and restrictions on the influx of peripheral immune cells during tumor initiation. As the resident macrophage and predominant immune cell of the central nervous system (CNS), microglia are poised to be the first responders to metastatic tumor infiltration, however an accumulating body of literature implies that microglia facilitate tumor growth in the CNS and brain parenchyma. In the following studies, we employ single-cell RNA sequencing (scRNAseq), spatial immunophenotying and flow cytometry on murine models of BCBM to interrogate the CNS immune microenvironment during tumor initiation to evaluate the tumoricidal potential of microglia and evaluate modalities for treatment of intracranial metastases. We identify the emergence of a pro-inflammatory program in microglia upon incidence of BCBM consistent with the canonical activity of a tumoricidal macrophage. We subsequently demonstrate that microglia upregulate antigen presentation (AP) machinery, and this activity is largely dependent on the infiltration of lymphocytes from the periphery. Finally, we evaluate the impact of agonistic anti-CD40 antibody on AP activity and the capacity to clear BCBM lesions.

Breast cancer brain metastasis (BCBM) is a lethal disease with no effective treatment options and limited experimental models. BCBMs are densely infiltrated by activated macrophages, but their origin, function and potential for therapeutic targeting are controversial. We used single-cell RNA-sequencing to investigate the role of brain-resident microglia in mouse and humanized models of BCBM. We find that mouse and human microglia directly interface with metastatic cells and mount a robust pro-inflammatory response. This is associated with upregulation of type I and II interferon responses, antigen presentation machinery cytokine and exome secretion and expansion of T and natural killer (NK) lymphocytes that control tumor outgrowth. Functional studies show that microglia depleted animals display decreased survival, increased metastasis, and impaired T and NK cell response to BCBM. This demonstrates the important tumor suppressive function of microglia during BCBM initiation.