Microbial Interventions in Cancer Therapy: Molecular Mechanisms, Engineering Strategies and Translational Constraints
Microbial interventions are being developed as anticancer agents, local drug factories, replicating immunotherapies and modifiers of treatment response. Their appeal arises from properties that conventional medicines do not readily reproduce: active replication, environmental sensing, penetration of poorly perfused tissue, programmable payload production and the capacity to engage innate and adaptive immunity. This critical narrative review evaluates tumour-targeting bacteria, engineered bacterial therapeutics, oncolytic viruses and manipulation of the host microbiome from a molecular-biology perspective. Literature published from 1 January 1990 to 29 May 2026 was selected through a transparent search of PubMed/MEDLINE, reference chaining and bibliographic verification of article identity and Digital Object Identifiers. The evidence indicates that microbial therapies operate through partially shared molecular modules: selective localisation or replication, pattern-recognition receptor signalling, immunogenic cell injury, antigen release and cross-presentation, metabolic reprogramming, and spatially restricted expression of therapeutic cargo. Yet the maturity of evidence differs sharply across platforms. Bacillus Calmette–Guérin and talimogene laherparepvec demonstrate that live microbial products can achieve clinical utility, whereas most engineered bacteria remain supported mainly by murine models and early-phase studies. Oncolytic viruses have produced durable responses in selected settings, but apparently strong phase I or II signals have not always translated into improved survival in randomised trials. Gut-microbiome associations with immune-checkpoint inhibitor outcomes are reproducible at a broad ecological level but inconsistent at the level of individual taxa; early faecal microbiota transplantation studies provide proof of principle rather than definitive efficacy. Across platforms, the central translational problem is not simply potency but control: biodistribution, genetic stability, inflammatory dose, immune clearance, manufacturing consistency and reversibility must be treated as molecular design variables. Progress will depend on mechanism-linked biomarkers, standardised pharmacodynamic measurements, rational combination trials and containment architectures that are validated under clinically realistic conditions.