The Cells That Aren't Infected Decide What the Infection Does

Helicobacter pylori lives on the surface of the stomach lining, causes chronic gastritis, and is an established cause of gastric cancer. It does not invade the tissue beneath. Yet the tissue beneath changes, and a group at Charité in Berlin has worked out that those uninfected cells are where the decision actually gets made.

The lining renews itself constantly. Under normal conditions that is governed by signals from neighbouring support cells. After injury, a different program takes over: a pathway called YAP, active in embryonic development and switched off in adults, comes back on and returns the tissue to a fetal-like state that grows fast and repairs quickly.

Why it matters: That regenerative state is useful after damage and dangerous if it persists. YAP is elevated in gastric cancer and in the lesions that precede it, so what turns it on during a decades-long infection is a question about how this cancer starts.

The missing trigger

This group had already shown that H. pylori makes the lining proliferate and that YAP is needed for the regenerative state. What did not work was the obvious next step: driving that signal on its own was not enough to reproduce the tissue changes. Something else had to be happening at the same time, and what it was had stayed unidentified.

The answer runs through three cell types, none of which is the bacterium.

It starts with a loss. Differentiated cells near the stomach surface normally carry active BMP signalling, and blocking it in the epithelium alone was enough to set the whole program in motion. Those BMP-deficient cells release chemokines, which draw in immune cells carrying IL-1β. The IL-1β does not act back on the epithelium; it acts on the mesenchymal support cells underneath, which carry the most receptor for it and respond by switching on COX2 and producing prostaglandin E2. That prostaglandin then activates YAP in the epithelial cells next door.

So the signal leaves the epithelium, passes through immune and stromal cells, and comes back. As Beccaceci and colleagues write in Nature Communications, the stromal niche is not only required to provide the growth factors necessary to guide the stem cell fate during homeostasis, but that it is remodeled upon tissue inflammation.

Two tests that make it a mechanism

A chain reconstructed from correlations would not be worth much. Two experiments hold it together.

The first isolates the middle. Growing stomach epithelium and stromal cells together as assembloids, adding IL-1β produced prostaglandin from the stroma and YAP activation in the epithelium, with no bacteria and no immune system present. The relay works on its own.

The second removes it. Deleting the IL-1 receptor specifically from stromal cells abolished both the reprogramming and the tissue damage caused by infection. Not reduced it, abolished it. That places the stroma on the necessary path rather than alongside it.

A smaller detail sharpens the specificity: the closely related IL-1α did not produce the effect, so this is not general inflammatory signalling but one particular cytokine acting on one particular cell type.

An old drug observation, explained

The COX2 step connects to something clinicians have noticed for years. COX2 is elevated in gastric cancer and in premalignant lesions, treating the infection brings it down, and anti-inflammatory drugs that block COX have been linked to reduced tumour risk in the gut. Those were separate observations with no shared mechanism. Here they become steps in one pathway.

There is also a tidy explanation for a puzzle about tolerance. The stomach surface meets microbes constantly without mounting an inflammatory response, and BMP signalling appears to be why: it holds down inflammatory signalling in differentiated cells. Lose BMP and those cells stop tolerating what they previously ignored.

The wider point is about where a pathogen's damage is done. H. pylori stays on the surface for decades, and the temptation is to look for what it does to the cells it touches. What this describes is closer to a hijack of the tissue's normal repair response: the bacterium changes conditions at the surface, and the stomach's own machinery for healing itself carries out the remodelling. That distinction matters for where an intervention would sit, since the stromal receptor and the prostaglandin step are both reachable without targeting the bacterium at all.

What the study can't say yet

What causes the initial BMP loss during real infection is not established. The experiments show that removing BMP is sufficient to start the cascade, which is not the same as showing that H. pylori removes it, or how.

The work is in mice and in cultured assembloids. The human relevance rests on the pathway's components being known to matter in human disease rather than on the cascade having been observed in patients.

And the link to cancer stays an inference. This explains how infection produces a fetal-like regenerative state, and that state is associated with transformation. Whether it is a step toward cancer or a parallel consequence is not tested here, and the timescales differ by decades.

Quick questions

What is a fetal-like state? Adult tissue reverting to a gene expression pattern resembling the developing organ, which grows and repairs quickly. Useful briefly after injury, riskier if it persists.

Why does it matter that the stroma is involved? Because the bacterium never touches those cells. It means the damage is routed through the tissue's own repair machinery rather than inflicted directly.

What's the one-line takeaway? Loss of a signal in the stomach lining recruits immune cells whose IL-1β acts on support cells underneath, and their prostaglandin switches the lining into a fetal-like regenerative state, with removing the receptor from those support cells alone enough to prevent the damage.