The authors observed that patients with CRC (but not patients with breast cancer) who received antibiotics against anaerobic bacteria before resection experienced a longer disease-free survival. The same protective effect, however, was not seen in patients with CRC that took the antibiotics after surgery, suggesting a crucial role for bacteria in an active tumour microenvironment. Prompted by the clinical data, Wang et al. created a well-defined, bacteria-infected CRC mouse model by using gram-negative, anaerobic Fusobacterium nucleatum (present in human CRC) or probiotics (E. coli Nissle). They found that bacterial cells preferentially invade hypoxic tumour cells and confirmed their promoting role in tumour growth and metastasis. Next, the authors developed liposomes loaded with a silver-tinidazole complex (LipoAgTNZ) that deliver the antibiotic nitroimidazole cargo to bacteria-infected orthotopic CRC tumours in mice, successfully eliminating the bacteria from the tumour cells. Tumour-bearing mice treated with LipoAgTNZ also showed an improved long-term survival compared to the control groups. As LipoAgTNZ itself did not show any cytotoxicity to tumour cells the authors concluded that the antibiotics treatment must have induced anti-tumour immunity in these immunocompetent mice. Indeed, they observed an infiltration of cytotoxic CD8+ T cells and other immune cells into the bacteria-bearing, mouse tumours owing to specific targeting of the microbial neoepitopes by the host T cells after LipoAgTNZ treatment. Finally, Wang et al. were able to show that removal of tumour-associated bacteria by LipoAgTNZ turned the immune-suppressive microenvironment into an anti-tumoral immune-activated state.