Mapping the Divergence and Convergence of Mouse and Human Tumor Immunity

Mouse models remain the premier preclinical model organisms for cancer immunotherapy, yet how they faithfully recapitulate the human tumor microenvironment (TME) has remained poorly defined. In a recent Nature Immunology study, Courau and colleagues addressed this gap through a systematic, cross-species immune profiling spanning 15 widely used mouse models and corresponding human cohorts. Their analysis revealed that commonly used mouse tumors only capture a portion of human disease-predominantly macrophage-rich, T-cell-poor microenvironments, but largely miss immune-rich, CXCL13-organized human tumor types, which are more responsive to immune checkpoint blockade in clinic. Beyond cellular composition, the authors also uncovered species-specific chemokine networks and cell-cell interactions that may account for the unique immune profiles in mouse and human. Despite the differences, conserved transcriptional modules were detected by consensus gene expression profile (GEP) analysis. Notably, a strong association between interferon-responsive myeloid cells and T-cell cytotoxicity marks a key transcriptional program where mouse and human biology converge, which is able to predict patient survival. Taken together, this work serves as a cautionary guide to the limitations of mouse models; more importantly, it offers a queryable atlas for precisely aligning preclinical models with the human conditions they credibly emulate.