Tumors vanish in 11 days as scientists flip a cancer protein

A Stanford-led team has made human lymphoma tumors disappear in mice within 11 days using a molecule that turns a cancer-driving protein into a trigger for cell death. ^{[1] [2]}

The experimental compound, TCIP3, brings that protein, BCL6, together with proteins capable of switching genes on. The aim is to activate the cell-death machinery that BCL6 normally keeps suppressed, exploiting a dependency that helps the cancer survive. ^{[1] [2]}

The study, published in the peer-reviewed journal Cell on July 20, 2026, includes experiments in laboratory cells and mice, not people. Researchers say the molecule needs chemical refinement and tests in additional animal species before human trials could be considered. ^{[2] [1]}

Even at this early stage, the result gives cancer researchers something worth pursuing: a way to redirect a cancer driver toward the destruction of the cells that depend on it. The strongest evidence so far is a rapid tumor response in one experimental setting. It does not establish a treatment for patients.

The work also contains a warning about how difficult development may be. Alongside the tumors, TCIP3 eliminated healthy immune-cell clusters that depend heavily on BCL6. The same biology that makes the molecule effective could limit how it can be used. ^[1]

Recruiting the cancer's own machinery

BCL6 helps control the behavior of healthy immune cells. It temporarily silences genes involved in stopping cell growth or initiating cell death. In many cases of diffuse large B-cell lymphoma, a form of lymphoma, BCL6 remains active and keeps those protective genes suppressed. ^[1]

That gives a cancer cell an advantage. Genes that could halt its growth or order its destruction remain quiet. The researchers designed TCIP3 to intervene at that point, where the cancer's survival depends on keeping a set of genes switched off. ^{[1] [2]}

The molecule has two binding parts. One attaches to BCL6, while the other recruits either P300 or CBP, proteins that add chemical marks called acetyl groups to other proteins. Bringing them together places machinery capable of activating genes next to machinery that normally silences them. ^{[1] [2]}

The proposed effect has two parts. Acetyl marks on BCL6 weaken its ability to suppress cell-death genes. Marks on nearby histones, the proteins around which DNA is packaged, help make that DNA accessible so the genes can be activated. ^{[1] [2]}

The intended endpoint is apoptosis, the controlled process through which a cell destroys itself. TCIP3 is designed to make the cell issue that order using proteins already involved in regulating its genes. ^{[1] [2]}

A conventional attempt to block BCL6 would seek to stop it doing its job. Here, the researchers seek to change the job. Their approach uses BCL6 as the point at which gene-activating proteins are recruited to a network of suppressed cell-death genes. ^{[1] [2]}

This is the central scientific advance. A protein that supports a tumor becomes a means of attacking it. That is a more ambitious intervention than simply disabling one component, because it aims to drive an active response inside the cancer cell.

The approach belongs to a broader field called chemically induced proximity: using a small molecule to bring proteins together so they perform an action they would rarely undertake on their own. The Cell paper notes that researchers have used this idea in other settings, including protein degradation and gene regulation. ^[2]

TCIP3 applies that principle to a cancer cell's gene controls. Its success depends on what the recruited proteins do once they meet, rather than on attachment alone.

Why the proteins stay together

The team examined the induced protein complex using crystallography and X-rays, methods that reveal molecular structure. The structure showed additional contacts between the recruited proteins that helped hold the complex together. That led the researchers to describe TCIP3 as a molecular glue. ^{[1] [2]}

Those contacts offer an explanation for the compound's potency. The two ends of TCIP3 bring the proteins into position, and interactions between the proteins add stability beyond the molecule's own attachments. The structural work therefore provides evidence for how the design operates, alongside the experiments measuring its effects. ^{[1] [2]}

In laboratory cultures, TCIP3 killed lymphoma cells at very low concentrations. In mice, the researchers implanted human lymphoma cells, waited for tumors to form and then administered the compound twice a day. They reported that treated tumors were completely gone by day 11, while tumors remained in the control animals. ^{[1] [2]}

The experiment tested an existing tumor, rather than merely whether the compound could prevent one from forming. That makes the disappearance of the treated tumors a particularly useful proof of concept. ^{[1] [2]}

But disappearance within 11 days is a measure of the initial response. The material available for this report does not establish how long the mice remained tumor-free after treatment ended or whether the tumors returned. It also does not provide the number of mice in each group. Those gaps prevent a fuller assessment of the result's durability and statistical strength. ^{[1] [2]}

The paper reports another finding relevant to future designs: compounds that recruited other gene-activating proteins produced different responses across the genome. Changing the recruited partner changed the biological response. ^[2]

That gives the strategy scope for further development, but also makes clear why each new version needs testing on its own merits. Bringing two proteins together is the starting point. The resulting pattern of gene activity determines what the cell actually does.

The safety question is already visible

According to Stanford's account, the treated mice showed no obvious signs of toxicity, and blood tests found no increase in inflammatory signals. Those are encouraging observations from the reported experiment. They do not establish that TCIP3 would be safe in people or during longer treatment. ^[1]

The compound also eliminated germinal centers, clusters of rapidly dividing immune cells that rely heavily on BCL6. These are healthy structures, and their loss shows that the treatment's effects extended beyond the implanted lymphoma. ^[1]

For development, that is a concrete problem to solve. The target is involved in both cancer and normal immune biology. A successful treatment would need to produce a useful anti-cancer effect while preserving enough normal immune function.

The researchers identify autoimmune diseases, including rheumatoid arthritis and myasthenia gravis, as possible future areas of investigation because germinal-center cells play a role in those conditions. These are proposed applications, not demonstrated treatments in the study. ^[1]

The same observation thus creates two research paths. Removing those cells might be useful in a different disease, while their loss raises safety questions in lymphoma treatment. Neither path can be judged from the tumor result alone.

There are other practical gaps. The material reviewed here does not establish the dose used, how it was delivered or how the compound behaved in the animals over time. Those details would matter when evaluating whether the response could be reproduced with an acceptable margin between benefit and harm. ^{[1] [2]}

The right conclusion is that TCIP3 has earned further testing. Calling it ready for patients would outrun the evidence. Dismissing it because it is early would overlook an unusually direct demonstration that a cancer-driving protein can be redirected toward cell death.

From an experiment to a candidate treatment

The Cell paper was written by Meredith Nix and colleagues. The project included researchers from Stanford, MD Anderson Cancer Center and Deep Origin, according to Stanford's account and the paper's author affiliations. ^{[2] [1]}

There is also a commercial connection. Stanford reports that senior authors Gerald Crabtree and Nathanael Gray have leadership or advisory roles at Shenandoah Therapeutics, which holds a Stanford license for the TCIP technology described in the study. ^[1]

That relationship belongs in the account of how this research might move toward development. It is not evidence against the experimental findings, but readers should know that some of the scientists involved also have ties to a company licensed to develop the technology.

The researchers are looking for other cancer-driving proteins that might be susceptible to the same approach. For now, that wider ambition remains a research program. The reported tumor result concerns TCIP3 and BCL6 in the lymphoma experiments described here. ^{[1] [2]}

The next steps are more exacting than finding another promising target. Chemical refinement must produce a compound suitable for further development, and additional animal testing must examine whether its activity holds up alongside an acceptable safety profile. The loss of healthy germinal centers makes immune function a central question, not a footnote. ^[1]

There is no established timetable for human testing in the material reviewed. The researchers' stated next steps are to refine TCIP3 and test it in additional animal species before considering trials in people. ^{[1] [2]}