Simulating Brain Tumor Heterogeneity with a Multiscale Agent-Based Model: Linking Molecular Signatures, Phenotypes and Expansion Rate

摘要

We have extended our previously developed 3D multi-scale agent-based braintumor model to simulate cancer heterogeneity and to analyze its impact acrossthe scales of interest. While our algorithm continues to employ an epidermalgrowth factor receptor (EGFR) gene-protein interaction network to determine thecells' phenotype, it now adds an explicit treatment of tumor cell adhesionrelated to the model's biochemical microenvironment. We simulate a simplifiedtumor progression pathway that leads to the emergence of five distinct gliomacell clones with different EGFR density and cell 'search precisions'. The insilico results show that microscopic tumor heterogeneity can impact the tumorsystem's multicellular growth patterns. Our findings further confirm that EGFRdensity results in the more aggressive clonal populations switching earlierfrom proliferation-dominated to a more migratory phenotype. Moreover, analyzingthe dynamic molecular profile that triggers the phenotypic switch betweenproliferation and migration, our in silico oncogenomics data display spatialand temporal diversity in documenting the regional impact of tumorigenesis, andthus support the added value of multi-site and repeated assessments in vitroand in vivo. Potential implications from this in silico work for experimentaland computational studies are discussed.