Abstract Cancer cells orchestrate a highly dynamic and interconnected network spanning metabolic, epigenetic, and immune mechanisms to drive adaptive plasticity and continuous development. This review synthesizes emerging insights into the coevolutionary strategies employed by malignant and stromal cells—particularly tumor cells and immune populations—across the continuum of tumorigenesis, metastasis, and treatment resistance. During tumor initiation, cancer cells rewire metabolism and generate oncometabolites that reshape the chromatin architecture to support immune evasion. Concurrently, metabolic competition in the tumor microenvironment (TME) induces epigenetic exhaustion of cytotoxic T cells, whereas tumor-associated myeloid cells adopt immunosuppressive and angiogenic phenotypes via metabolite-dependent histone modifications to promote carcinogenesis. At metastatic frontiers, under the local metabolic pressure of target organs, tumor cells undergo epigenetic reprogramming to evade immune attacks and support colonization. Premetastatic niches are preconditioned through exosome-mediated transfer of metabolic enzymes and noncoding RNAs that reprogram resident cells before tumor cells arrive. In cancer immunotherapy, tumors often exploit metabolic adaptative strategies to inhibit cell death signaling pathways or the compensatory activation of self-protective mechanisms to circumvent immune-mediated cytotoxicity and develop resistance to immunotherapy. By mapping these dynamic interactions, we propose a novel conceptual framework of the “metabolic-epigenetic-immune axis” that transcends traditional compartmentalized approaches and helps to identify nodal convergence points for therapeutic co-targeting. This review also prioritizes multitarget inhibitors arising from the convergence of metabolic reprogramming, epigenetic plasticity, and immune evasion networks. An integrated approach to these pathways advances next-generation precision oncology strategies aimed at circumventing the evolutionary resilience of cancer.
Wang et al. (Mon,) studied this question.