Multi-scale Network Topology Analysis Reveals Metabolic Reprogramming and Therapeutic Vulnerabilities in the Tumor Microenvironment

The tumor microenvironment comprises diverse cell populations that coordinate metabolic activities to sustain malignant growth, yet the systems-level organization of these interactions remains poorly understood. Here, we present an integrated computational framework combining single-cell transcriptomics, genome-scale metabolic modeling, and multi-scale network geometry to systematically decode metabolic coordination in cancer ecosystems. Analyzing colorectal cancer stromal populations, we demonstrate that FAP ⁺ cancer-associated fibroblasts and MARCO ⁺ tumor-associated macrophages undergo extensive metabolic reprogramming establishing a division of labor: fibroblasts specialize in amino acid and fatty acid metabolism, while macrophages adopt cancer-like biosynthetic programs. Systematic in silico knockout analysis identified tumor-selective vulnerabilities that aligned with the cells’ specialized metabolism, and we validated these targets using patient survival data. To reveal architectural organization, we applied metabolite role transition analysis and multifractal geometric characterization to metabolic networks. Critically, while conventional network metrics failed to distinguish tumor from normal phenotypes, our multifractal geometric analysis successfully separated tissue states through coordinated architectural reorganization across hierarchical scales. Role transition analysis revealed that 20-25% of metabolites undergo functional reorganization, with prostaglandin and bile acid derivatives emerging as critical communication hubs. Ollivier-Ricci curvature analysis identified pathway-specific geometric remodeling in fatty acid, xenobiotic, and leukotriene metabolism. Our findings demonstrate that metabolic adaptation represents ecosystem-level network reorganization rather than isolated pathway changes, and that flux-based analysis alone cannot capture architectural vulnerabilities emerging from multicellular coordination. Our integrated framework, generalizable across cancer types, provides a roadmap for identifying therapeutic strategies targeting metabolic cooperation networks and topological bottlenecks.