Microbial defence against cancer

Identifying the contribution of the microbiome to cancer initiation and progression could transform patient prevention and treatment strategies. Dysregulation of the microbiome has been established to promote intestinal tumorigenesis with distinct microbial species, such as Fusobacterium spp., reported to exert pro-tumorigenic effects during colorectal cancer (CRC) development. Yet, a non-mutually exclusive possibility is that tumorigenesis proceeds following a reduction in endogenous anti-tumorigenic bacterial species. Zagato et al. set out to explore this alternative explanation.

Credit: Lara Crow/Springer Nature Limited

Using a spontaneous mouse model of CRC in which the adenomatous polyposis coli (APC) gene is mutated (Apc^Min/+ mice), the authors longitudinally tracked changes in the microbiota composition compared with wild-type (WT) mice. While microbial diversity between the two groups of mice was unaffected at any age, differences in the abundance of the genus Faecalibaculum were noted between 8 to 12 weeks of age, coinciding with the development of tumours. In particular, the bacterial taxon Faecalibaculum rodentium was strongly under-represented in Apc^Min/+ mice compared with WT mice in this same timeframe. The specific strain within this taxon was then narrowed down to the isolate F. PB1, found in the mucus of the intestines.

The mucus layer constitutes a unique microbial niche, and alterations to its composition are likely to perturb the balance of bacterial species. Reasoning that tumorigenesis might cause changes to the production of mucins — the structural components of mucus — resulting from transformation of epithelial cells, the authors profiled mucin levels in ileal and tumour tissues. This revealed expression changes in a range of mucins in Apc^Min/+ mice compared with WT mice. Next, to test if the Apc^Min/+ gut microenvironment was non-conducive to F. PB1 colonization, both WT and Apc^Min/+ mice were treated with antibiotics to deplete the microbiota and then reconstituted with F. PB1 by gavage. At 48 h, the levels of F. PB1 in mucus were decreased in Apc^Min/+ mice relative to WT mice, consistent with F. PB1 loss co-occurring with mucus changes. However, reintroducing F. PB1 every other day to maintain constant levels of F. PB1 demonstrated that F. PB1 could protect against tumour progression once the tumours had started to develop, as tumour growth was reduced in 8 to 12 week old Apc^Min/+ mice.

Signals derived from the microbiota have been shown to shape the tumour immune microenvironment. To assess if F. PB1 imparts its anti-tumour activity by influencing the immune system, germ-free Apc^Min/+ mice were administered F. PB1. No major differences in the numbers of adaptive immune cells and only minor differences in innate cells were noted. However, tumour cell proliferation was decreased under these conditions, indicating that F. PB1 affects tumour cell growth directly.

The gut microbiota promote colonic health through the production of short-chain fatty acids (SCFAs), which serve as an energy source as well as having anti-inflammatory properties. Upon monocolonization of germ-free Apc^Min/+ mice with F. PB1, increased faecal levels of SCFAs — butyrate, propionate and acetate — were observed, suggesting that F. PB1 releases metabolites that might control cell proliferation.

By investigating this mechanism in more detail, it was noted that F. PB1 grown in anaerobic conditions in vitro was capable of producing both butyrate and lactate. Furthermore, the spent medium alone from F. PB1 culture was sufficient to prevent proliferation of mouse CRC cell lines. CRC cells are known not to use butyrate unlike their normal epithelial counterparts, which leads to accumulation of this SCFA in tumour tissues where it can act as an inhibitor of histone deacetylation. Correspondingly, incubation with F. PB1 spent medium increased the acetylation of histone H3 in CRC cell lines, and this was associated with a decrease in activation of the calcineurin–NFATC3 signalling pathway, known to be involved in cell proliferation.

To assess if the anti-proliferative activity of F. PB1 also applied to the in vivo situation, spent medium was administered to 11-week-old tumour-bearing Apc^Min/+ mice who were pre-treated with antibiotics to remove confounding microbiota. This resulted in a decrease in the size of tumour lesions with reduced NFATC3 activation and increased histone H3 acetylation within these lesions. A similar phenotype was observed in a chemically induced mouse model of colitis-associated cancer.

Faecalibaculum rodentium was strongly under-represented in Apc^Min/+ mice

Attempting to identify a human homologue to F. PB1, metagenomic analysis of tissue samples from patients with advanced colon adenomas revealed that Holdemanella biformis, which is phylogenetically close to F. PB1, was under-represented in these benign tumours. Importantly, spent medium from H. biformis contained SCFAs and could reduce the number of tumour lesions in antibiotic-treated Apc^Min/+ mice.

This study points towards SCFA-generating bacterial species having anti-tumoural properties, with their reduction upon early tumorigenesis possibly serving as a biomarker.