Tff2 identifies a gastric progenitor pool that can feed both metaplasia and dysplasia
In the gastric corpus, injury does not simply erase mature cells and trigger a uniform repair response. It can redirect a defined progenitor population into two very different outcomes: transient spasmolytic polypeptide-expressing metaplasia and, under oncogenic pressure, dysplasia. The study places Tff2-expressing corpus progenitors at the center of that decision-making process, linking tissue repair, metaplastic change, and early neoplastic progression in a way that challenges the classic stepwise view of gastric carcinogenesis.
The work focused on the corpus isthmus, where highly proliferative Tff2+ progenitors were localized. These cells generated multiple secretory lineages, including chief cells, but they did not behave like a durable stem-cell reservoir with long-term self-renewal. Their behavior after injury was rapid and adaptive. When the mucosa was damaged, or when chief cells were lost, the Tff2+ pool expanded quickly and gave rise to transient SPEM, a metaplastic state long associated with oxyntic atrophy and gastric cancer risk.
Injury drives a rapid metaplastic response from isthmus progenitors
The lineage-tracing experiments in Tff2-CreERT2 knock-in mice mapped a clear sequence. After acute injury, the Tff2+ progenitors did not sit idle in reserve. They proliferated and produced the cellular material required for a metaplastic repair program. The biology is consistent with a tissue that has lost its mature acid-secreting architecture and is trying to reconstruct a protective mucous phenotype before normal differentiation can resume.
One sentence captures the core finding: Tff2+ progenitors were required for SPEM formation.
Genetic ablation of these cells abolished SPEM, confirming that the metaplasia was not a generic bystander response from any surviving epithelial cell. The reverse experiment was just as informative. When chief cells were ablated using either Lgr5-DTR or Gif-rtTA-labeled systems, SPEM derived from Tff2+ progenitors became more prominent. In practical terms, chief-cell loss appears to open a niche that amplifies the contribution of the isthmus progenitor compartment rather than replacing it.
This arrangement helps explain why SPEM emerges so readily after oxyntic injury. The stomach is not recruiting a random epithelial population to fill the gap. It is using a pre-existing proliferative compartment with the capacity to shift toward a protective, secretory metaplastic program when the corpus is stressed. The metaplasia is therefore not just a marker of damage; it is an active lineage choice.
Oncogenic pressure pushes the same cells toward dysplasia
The story changes when injury intersects with H. pylori infection or KrasG12D activation. Under those conditions, the same Tff2+ progenitors moved beyond metaplasia and into dysplasia. The study showed that Kras activation in Tff2+ progenitors promoted direct progression to dysplasia through acquisition of stem cell-like properties. In other words, the oncogenic signal did not merely accelerate metaplastic change; it reprogrammed these cells toward a more plastic, tumor-prone state.
By contrast, Kras-mutant SPEM and chief cells did not progress to dysplasia. That asymmetry matters biologically. It suggests that dysplasia is not a simple inevitable sequel to metaplasia once Kras is activated somewhere in the epithelium. Cellular context governs fate. The Tff2+ progenitor state appears permissive for malignant conversion, whereas differentiated SPEM or mature chief cells, even when carrying Kras mutation, do not show the same progression in this model.
The implication is direct: the cell of origin shapes the trajectory of early gastric neoplasia as much as the mutation itself.
Human tissue data mirror the mouse lineage map
Mouse lineage tracing alone would not be enough to recast gastric carcinogenesis, so the human validation is important. Single-cell RNA sequencing and spatial transcriptomics on human gastric tissues revealed distinct differentiation trajectories from isthmus proliferating cells toward either SPEM or gastric cancer. Those human data align with the experimental mouse findings and support the idea that the corpus isthmus harbors a progenitor population capable of bifurcating into repair-associated metaplasia or neoplastic evolution.
The spatial data are especially persuasive because they anchor transcriptional states to tissue architecture. They show that the progenitor compartment is not dispersed randomly across the corpus; it occupies a defined niche where injury-related cues and oncogenic signals can shape lineage output. In a chronically inflamed or infected stomach, that niche could become the site where repair becomes metaplasia and metaplasia becomes dysplasia.
Why this reframes gastric carcinogenesis
For years, SPEM has been treated as part of a broader atrophic sequence that may eventually precede cancer. This study narrows the field to a specific epithelial origin and a specific anatomical zone. Tff2-positive corpus progenitors are not just a transient wound-healing population. They are a common source for both metaplastic and dysplastic programs, depending on the molecular context of the injury.
That creates a more precise model of gastric tumor initiation. Acute epithelial loss, chief-cell depletion, H. pylori infection, and Kras activation do not all provoke the same lineage response. Some drive a reversible metaplastic state. Others push the same progenitor compartment into stem-like remodeling and dysplasia. The difference lies in whether the tissue is simply repairing damage or being pushed, by infection or oncogenic signaling, into a state of malignant plasticity.
The practical consequence is an emphasis on early epithelial state rather than late histology alone. If Tff2+ progenitors are the point at which repair programs can divert into cancer-related pathways, then they become a biologically relevant target for future mechanistic work and potentially for risk stratification in atrophy-associated gastric disease. The study does not claim a clinical biomarker or a therapeutic intervention, and it does not establish human predictive value on its own. It does, however, define a lineage framework that makes the early steps of gastric carcinogenesis far more legible.
What remains unresolved is how these progenitors are normally restrained, which inflammatory cues widen their plasticity, and whether similar lineage control exists in human premalignant lesions outside the tissue contexts studied here. Those questions will determine whether this progenitor map becomes a tool for intervention or remains a detailed description of a dangerous repair pathway.