TAMI-18. DIFFERENTIAL LIPID METABOLISM IN CANCER MICROENVIRONMENTS LEADS TO A REQUIREMENT FOR FATTY ACID DESATURASES FADS1 AND FADS2 IN GBM CANCER STEM CELL MAINTENANCE

Glioblastoma (GBM) is marked by cellular heterogeneity through microenvironments of a tumor, including metabolic heterogeneity. While altered cellular metabolism in cancer is well-known, how lipid metabolism is altered in different GBM microenvironmental conditions and cancer stem cell (CSC) states within a tumor remains an open question. We developed 3-dimensional GBM organoid models that mimic the transition zone between nutrient-rich cellular tumor and nutrient-poor psuedopalisading/perinecrotic tumor regions and performed spatially defined RNA-sequencing to investigate lipid metabolism. Spatial analysis revealed striking differences in metabolism between diverse cell populations from the same patient, with lipid enrichment in the hypoxic organoid cores and the pseudopalisading regions of patient tumors. This was accompanied by regionally restricted upregulation of lipid droplets and Hypoxia Inducible Lipid Droplet Associated gene expression in organoid cores and in the pseudopalisading regions of clinical GBM tumors. Using targeted lipidomic analysis, we assessed differences in acutely enriched CSC and non-CSCs from patient-derived models to explore the link between stem cell state and lipid metabolism. CSCs have low lipid droplet accumulation compared to non-CSCs in organoids and xenograft tumors, and prospectively sorted lipid-low GBM cells are functionally enriched for stem cell activity. This suggests lipid metabolism may not be simply a product of the microenvironment but also may be a reflection of cellular state. CSCs had decreased levels of major classes of neutral lipids compared to non-CSCs, but had significantly increased polyunsaturated fatty acid production due to increased expression of fatty acid desaturases FADS1 and FADS2. FADS1 and FADS2 expression are both essential to maintain CSC viability and self-renewal. Our data demonstrate that spatially and hierarchically distinct lipid metabolism phenotypes occur clinically in the majority of patients, can be recapitulated in laboratory models, and these altered lipid metabolic pathways may represent therapeutic targets for GBM.