Brain Metastasis Organotropism and Therapy Response in Small Cell Lung Cancer - 1 PROJECT SUMMARY 2 Small cell lung cancer (SCLC) afflicts more than 30,000 patients per year and is rapidly fatal in 94% of cases, 3 with median survival of less than one year. SCLC is almost always metastatic at diagnosis, and in most cases 4 spreads to the brain, often causing severe neurologic morbidity in addition to mortality. Importantly, most brain 5 metastases (BrMs) emerge after diagnosis, and for SCLC patients with BrMs, prophylactic cranial irradiation 6 (PCI) can reduce their incidence. However, PCI is highly toxic, causing chronic headaches and cognitive decline, 7 and has a negligible impact on patient survival. There is an urgent need for therapies to prevent or treat BrMs 8 that are more effective and less toxic that PCI. However, there are no reported models of SCLC that metastasize 9 spontaneously to the brain. Most SCLC BrM research has been performed by intracranial injection of cancer 10 cells, which skips so much of the metastatic cascade that it may not distinguish between SCLC that is 11 predisposed to form BrMs (brain organotropic) and SCLC that would normally remain extracranial. 12 We have developed a large panel of over 80 patient-derived xenografts (PDXs) of SCLC, more than half from 13 relapsed patients. We have characterized these models thoroughly, with full clinical and molecular annotation 14 (genomic, transcriptomic, proteomic and metabolomic), and in vivo response profiles to the main chemotherapy 15 regimens that we give patients. These studies have revealed that PDX models are faithful to their corresponding 16 patient tumors and clinical outcomes. However, they do not metastasize from their subcutaneous compartment, 17 failing to capture one of the cardinal features of the disease. 18 We recently optimized a surgical protocol for orthotopic transplantation of dissociated SCLC PDX cells into the 19 left lungs of mice. Across four models, all produced extensive extracranial metastases (EMs), and genomic 20 analyses confirmed these EMs arose from the primary lung tumor rather than the initial injection. Notably, two 21 models developed spontaneous, recurrent brain metastases (BrMs) that caused distinct neurologic symptoms, 22 and in one model, 70% of mice developed hydrocephalus associated with BrMs. Transcriptional analysis showed 23 that these BrMs, but not their primary lung tumors or other metastases, demonstrated an epithelial-to-neuronal 24 (ENT) transition. Here, we will leverage these models to test whether ENT drives brain metastasis in SCLC and 25 to uncover additional molecular determinants of brain organotropism. We will compare BrM-specific changes 26 with clinical tumor samples to determine if similar molecular alterations arise during brain metastasis in patients. 27 In parallel, we will evaluate clinical anti-cancer therapies with known effects on BrMs to assess whether our 28 models recapitulate patient responses, to establish a preclinical platform for testing BrM prevention and 29 treatment strategies. Ultimately, our goal is to develop new therapies to protect SCLC patients from the 30 devastating impact of brain metastases.
