TMET-57. Integrating scRNA-seq and13C-tracing delineates metabolic heterogeneity in the glioblastoma tumor microenvironment

Abstract Although there has been a recent surge in the use of in vivo metabolite isotope tracing for metabolic flux estimations, these studies are limited to bulk tissues in preclinical models, posing a significant challenge in capturing the heterogeneous metabolic signatures of neoplastic and non-neoplastic cell types within tumor microenvironments (TME). Standard metabolomic techniques require tissue homogenization, which obstructs understanding the metabolic contributions of individual cell populations. On the other hand, single-cell flux balance analysis tools, which rely on scRNA-seq data, also face limitations when gene expression does not accurately reflect cellular metabolic activities. To address these challenges and provide mechanistic insights at the single-cell level, we developed 13C-scMFA—a novel integrated framework that combines scRNA-seq and 13C-metabolic tracing data—enabling the estimation of inter- and intracellular metabolic fluxes in various TME cell types for both human and mouse models. We validated this framework by comparing metabolic fluxes across different cell types in glioblastoma patients and xenografts, using paired scRNA-seq and isotope tracing data to predict responses to two treatments: (1) pharmacologic inhibition of de novo GMP synthesis and (2) environmental serine restriction. Our GMP synthesis model revealed purine metabolic heterogeneity across patients’ neoplastic and myeloid cells, which allow prediction of potential responders to mycophenolate mofetil (MMF), an inhibitor of de novo GMP synthesis. Furthermore, we found that astrocytes have a high glucose-derived serine synthesis flux, which lead to serine secretion into the TME. While de novo serine synthesis, circulating and microenvironmental-derived serine could account for serine sources in neoplastic cells, our model quantifies the contribution of these sources, hence potentially predict response to dietary serine depletion. In conclusion, our model potentially identifies glioblastoma patients who may benefit from MMF treatment or a serine- and glycine-restricted diet, opening pathways for personalized metabolic therapies.