Caspase-centred and Mitochondrial Apoptosis Signalling in Cancer: Pro-survival Rewiring, Therapeutic Vulnerabilities and Translational Constraints

Abstract

Apoptosis is a regulated cell-death programme in which upstream stress sensing is converted into mitochondrial outer membrane permeabilisation, caspase activation and ordered cellular dismantling. Cancer cells rarely abolish this machinery completely. Instead, they reconfigure its thresholds by altering death-receptor signalling, B-cell lymphoma 2 family interactions, apoptosome competence, inhibitor-of-apoptosis proteins and survival pathways. This critical narrative review examines how pro-apoptotic and anti-apoptotic signals converge on caspase proteases and the mitochondrial cascade, how these circuits are dysregulated during malignant evolution, and why apparently intact pathway components often fail to predict therapeutic response. Literature published principally from 1990 to 20 May 2026 was identified through live searches of accessible biomedical indexes and scholarly platforms, supplemented by citation searching and DOI-level verification. The evidence supports a threshold model in which initiator-caspase activation is shaped by signalling-platform stoichiometry, whereas mitochondrial commitment is governed by the dynamic balance among BH3-only proteins, BAX/BAK effectors and pro-survival BCL2 proteins. Cytochrome c–APAF1 apoptosome assembly and relief of XIAP-mediated restraint then determine the amplitude and duration of executioner-caspase activity. In cancer, oncogenic PI3K–AKT and nuclear factor-κB signalling, TP53 dysfunction, altered death-receptor complexes, BCL2-family dependence and IAP activity generate context-specific blocks rather than a single universal lesion. Clinical success with the BCL2-selective inhibitor venetoclax validates direct pharmacological activation of apoptosis, but lineage restriction, adaptive transfer to MCL1 or BCL-XL, normal-tissue toxicity and intratumour heterogeneity constrain generalisation. Sublethal mitochondrial permeabilisation and incomplete caspase activation further complicate the traditional binary model by promoting genomic instability, inflammatory signalling, persister states or tumour repopulation. The most informative translational strategy is therefore not static measurement of one protein, but integrated functional assessment of mitochondrial priming, pathway topology and temporal response. Future studies should combine single-cell perturbation, spatial profiling, pharmacodynamic biomarkers and clinically anchored resistance sampling to distinguish cells that are apoptosis-defective from those that are merely apoptosis-buffered.