Long-term colonization of the stomach by Helicobacter pylori represents the highest known risk for the development of gastric adenocarcinoma, yet only a fraction of the many colonized individuals go on to develop cancer. Richard Peek and colleagues used a rodent model to show that bacterial adhesion to the host cells might be responsible for the increased risk.

The authors infected a gerbil with a human clinical isolate of H. pylori and allowed the bacteria to adapt to the host environment for 3 weeks. This produced a highly oncogenic strain, ideal for further study of the carcinogenic process. When a cohort of gerbils was infected with this strain, 75% developed gastric adenocarcinoma after 8 weeks, whereas no carcinoma developed following infection with the parental human isolate.

Using gastric epithelial cells in vitro, the authors found that their oncogenic strain increased the amount of β-catenin in the nucleus. In addition, a luciferase assay showed an induction of β-catenin-dependent transcriptional activation. These are key phenomena in the progression of many tumours, and can be caused by other carcinogenic agents. However, the authors were surprised to find that the mechanism by which H. pylori induces these changes is not the usual one of blocking the phosphorylation and ubiquitylation of β-catenin. Instead, they found that the cag set of pathogenicity genes from H. pylori was involved. These genes encode the molecular apparatus for transferring the CagA protein into the host cell, where CagA activates the host phosphatase SHP-2 to cause morphological changes. A series of genetic knockout experiments confirmed that CagA was indeed responsible for the oncogenic strain's pathological properties.

So, how does the oncogenic strain differ from its parent, given that both possess the cag genes and no other significant genetic gains or deletions were seen in a microarray comparison? Whereas both strains expressed CagA at similar levels, the oncogenic strain transferred it into the host more efficiently. This was shown to be the result of better adhesion of the bacteria to the host cell.

Some of these in vitro findings were tested back in the rodent model. Here, the authors found that nuclear localization of β-catenin was increased only in the early stages of infection by the oncogenic strain, implicating this as the crucial time for carcinogenesis. They also confirmed that β-catenin was found in the nucleus more often in epithelial cells harvested from humans infected with cag+ strains than those infected with cag− strains or no H. pylori infection at all.

The results indicate that the increased risk of cancer in long-term H. pylori infections might be a consequence of selection pressure on the bacteria to adhere to the host cells in order to remain in the stomach. However, the precise genetic change involved remains to be identified.