Encoding metabolic stress into tumor evolution through ECM–integrin-dependent transcriptional plasticity

Tumor evolution is driven not only by genetic alterations but also by the capacity of cancer cells to adapt to dynamic microenvironmental stresses. Emerging evidence indicates that metabolic constraints can remodel the extracellular matrix (ECM), extending stress adaptation beyond cell-intrinsic programs. Importantly, stress-induced changes in ECM composition, organization and mechanics may persist beyond the initiating stimulus and influence subsequent cell–matrix signaling. We refer to this proposed persistence of stress-associated extracellular changes as“ECM encoding”, a form of physicochemical information storage that is mechanistically distinct from genetic memory or cell-intrinsic epigenetic stabilization. Here, we critically review the ECM–integrin axis as a signaling interface linking metabolic stress, redox adaptation and transcriptional plasticity. Metabolic pressures, including hypoxia, nutrient limitation and isolation stress, reshape ECM composition, mechanics and ligand availability, thereby reprogramming integrin signaling. Through focal adhesions, cytoskeletal tension and nuclear mechanotransduction, integrin-dependent signals can influence chromatin accessibility and converge with stress-responsive networks involving YAP/TAZ, ATF4, HIF, NRF2, STAT3 and NF-κB. In parallel, matrix-dependent regulation of mitochondrial reactive oxygen species production and NRF2-mediated antioxidant responses may provide mechanisms linking extracellular cues to redox adaptation. These interconnected processes can promote phenotypic plasticity associated with stemness, dormancy, metastasis, immune evasion and therapeutic resistance, although their relative contributions vary across tumor types and experimental contexts. Finally, we evaluate the therapeutic potential of targeting the ECM–integrin–redox network while considering signaling redundancy, previous clinical setbacks and the need for biomarker-guided patient selection.