Activating YAP1 in a subset of mouse liver cells produced tumours with no detectable canonical driver mutations — the oncogenes were switched on by stripping methyl marks off DNA instead.

The standard account of cancer is a mutation story. Genes that drive proliferation acquire activating changes, genes that restrain it are broken, and enough of these accumulate in one cell lineage for growth to escape control. Cancer genomics has been built on finding those driver mutations, and treatment increasingly depends on knowing which ones a tumour carries.

A study in Communications Biology from the Institute of Science Tokyo describes liver tumours that formed without them. The genes were activated — but by having chemical marks removed from DNA rather than by any change to the sequence.

The work was led by Professor Hiroshi Nishina of the Medical Research Laboratory at Science Tokyo’s Institute of Integrated Research, with Yoshimi Okamoto-Uchida, Misaki Kosaka and Teppei Shimamura at Science Tokyo, and collaborators at the National Cancer Center Research Institute, the University of Tokyo, Kobe University and Showa Medical University.

A mosaic mouse

The model activated YAP1 in only a subset of liver cells. YAP1 is a transcriptional coactivator and the output of the Hippo pathway, which controls organ size and is disrupted in a range of cancers; its relevance to liver cancer is well established.

The mosaic design is a meaningful choice. Activating an oncogene in every cell of an organ produces overgrowth that is hard to distinguish from a developmental abnormality, whereas activating it in scattered cells leaves normal neighbours in place — which is closer to how a human tumour arises, in a tissue that is otherwise behaving itself.

The tumours that developed showed no detectable canonical driver gene mutations.

That absence is the study’s starting point rather than its conclusion, and the obvious next question is what was switching the cancer genes on.

Taking the marks off

Methylation analysis found widespread DNA demethylation, linked to the activation of multiple oncogenes.

DNA methylation is the best-understood epigenetic mark: a methyl group attached to cytosine, concentrated at gene promoters, where its presence generally keeps a gene quiet. Removing it from the wrong promoters turns genes on without altering a single base of sequence — which produces the same functional outcome as an activating mutation and leaves no trace in a mutation screen.

The mechanism the team identified runs through a specific enzyme. YAP1–TEAD activation increased expression of TET1, one of the dioxygenases that actively strip methyl marks from DNA, and the resulting demethylation activated cancer-related genes.

“We observed that the expression of TET1 increased, promoting the activation of multiple cancer-related genes,” said Okamoto-Uchida, an assistant professor at Science Tokyo.

The team’s hypothesis had been that the YAP1–TEAD complex itself was reshaping the epigenome, and TET1 is the intermediary that makes the account concrete — a transcription factor complex that induces a demethylating enzyme, which then opens up genes the complex does not need to bind directly.

Suppressing TET1 function reduced tumour formation, which is the test that matters. A correlation between demethylation and tumour growth could run either way; blocking the enzyme and getting fewer tumours establishes that the demethylation is doing work rather than following along.

Where the TET1-high cells sit

Spatial transcriptomics — sequencing that preserves where in the tissue each measurement came from — identified a subpopulation of hepatocellular carcinoma cells with high TET1 expression, interacting with other hepatic cell types.

Location is the information a conventional sequencing experiment destroys. A tumour ground up and sequenced gives an average over every cell in it; keeping the geometry shows that the TET1-high cells form a distinct population with particular neighbours, which is the difference between a tumour being uniformly driven by this mechanism and containing a specific compartment that is.

The human thread

The clinical connection comes through an unusual patient group. Elevated YAP1–TEAD–TET1-associated signatures were found in hepatocytes from patients with Fontan-associated liver disease.

The Fontan procedure is the operation that lets children born with a single functional heart ventricle survive, by routing venous blood directly to the lungs. It works, and it leaves the liver under chronically elevated venous pressure for decades, producing fibrosis and an elevated risk of hepatocellular carcinoma. Those cancers are known to carry canonical driver mutations less often than liver cancers generally do.

Which makes it the right population in which to look for a mutation-independent mechanism — a group with elevated cancer risk, a clear mechanical cause, and a known shortage of the usual mutations. Finding the signature there connects a mouse experiment to a human disease through a prediction rather than a coincidence.

What the release does not contain

The most conspicuous gap is quantitative. The release reports no numbers at all: no tumour incidence in the model or in TET1-suppressed animals, no counts of animals, no measure of how much demethylation occurred, and no indication of how many Fontan patients’ hepatocytes were examined. Every claim above is therefore directional — increased, reduced, elevated — with no magnitude attached.

“No detectable” driver mutations also depends entirely on how hard anyone looked. Whether the screen was whole-genome or a targeted panel, and at what sequencing depth, determines whether a subclonal mutation in a minority of tumour cells would have been seen; the release does not say.

This is also a mouse model in which an oncogene was switched on deliberately, which is a long way from a tumour that arises on its own. And the human component is a transcriptional signature in tissue, not an outcome: it shows that the pathway is active in a group at elevated risk, not that it caused anyone’s cancer.

The therapeutic implication is real but distant. TET1 inhibition reducing tumour formation in mice points at a target, and TET enzymes are involved in normal development and in immune cell function, which makes systemic inhibition a prospect requiring considerable care. No drug is reported here.

What the work offers is a mechanism for a category of tumour that mutation-based cancer genomics is structurally unable to explain, together with a human population in which to test it. The paper was published on 17 September 2026. Funding came from the Japan Society for the Promotion of Science, the Japan Agency for Medical Research and Development, the Secom Science and Technology Foundation and several Science Tokyo research programmes. The release states no conflicts of interest.

---

Notes

1. “YAP1 induces hepatocellular carcinoma via DNA demethylation rather than by canonical driver gene mutations.” Communications Biology, volume 9, 2026. DOI: 10.1038/s42003-026-10882-w.
2. Institute of Science Tokyo. “Beyond mutations: New study reveals alternative route to cancer.” EurekAlert!, 22 October 2026. https://www.eurekalert.org/news-releases/1148827.

Photo by Edward Jenner on Pexels.com