Multi-cellular interactions defining the human brain metastatic niche

Brain metastasis (BM) occurs in up to 40% of patients with advanced cancers, most frequently arising from non-small cell lung cancer (NSCLC). Patients with BM frequently suffer from debilitating symptoms, have worse response rates to modern cancer therapies and are excluded from most clinical trials, resulting in an overall poor prognosis. While the clinical significance of BM is broadly recognized, our understanding of underlying molecular, cellular and microenvironmental mechanisms remains rudimentary. Here, through several innovations, we overcome experimental, technical, and analytical barriers to gain unprecedented insight into cellular and microenvironmental features of human BM. In an integrative analysis of multi-modal single-cell RNA, T cell receptor and spatial transcriptomics of primary tumors and BM from patients with NSCLC, coupled with analyses of public data sets, we identify chromosomal instability (CIN), and CIN-induced molecular adaptations as key driver of brain-metastatic organotropism. The brain metastatic ecosystem is enriched with a pro-tumorigenic myeloid and dysfunctional T cell compartment. In this proposal, we aim to define the mechanistic underpinnings of these observations. In Aim 1, we will employ a fundamentally novel analytical tool, ContactTracing, to map at a systems level all cell-cell interactions in the BM ecosystem based on multi-modal single-cell sequencing. We assembled a large validation cohort of NSCLC specimen, including primary tumors, brain and extracranial metastases, that underwent whole-exome sequencing (WES) and RNA-seq which will be deconvolved to validate pre- dictions derived from single-cell data. Furthermore, we assembled additional NSCLC for multiplexed immunofluorescence with established antibody panels to measure the rate of CIN, CIN-adaption, and their spatial association with myeloid and lymphoid immune infiltrates. In Aim 2, through functional imaging of human and murine models we identify CIN as a defining feature of LKB1-deficient tumors. LKB1 loss (or deleterious mutation) is a common genomic subtype of NSCLC (along with KRAS co-mutation) that frequently metastasizes to the brain and is characterized by treatment resistance and poor prognosis. We find that CIN results in tonic activation of the cGAS-STING pathway to promote BM. We will dissect underlying mechanisms of CIN-induced cellular adaptations that confer brain-metastatic organotropism through modulation of the brain-metastatic niche through cGAMP hydrolysis to adenosine, which permeabilizes the blood-brain-barrier and creates an immunosuppressive environment. For this purpose, we will use in situ niche-labeling of in vivo BM models, coupled with single-cell/spatial transcriptomics to track the evolution of brain-metastatic ecosystem, and dynamics of cell interaction networks using ContactTracing. Upon completion of this work, we will provide mechanistic and clinical insights into genomic and multi-cellular features of the evolving brain-metastatic ecosystem. These insights are the first step towards development of more effective therapies of affected patients.