Onion rings (Photo by Nik on Unsplash)
A Diabetes Drug Improved Memory in Mice Fed Fatty Diets
In A Nutshell
- One week of fatty eating cut a key brain protein, PQBP1, in mice.
- That protein edits genes that build brain-cell connections, which weakened.
- Gene therapy and a diabetes drug improved memory in the mice.
- No people were studied, so human effects remain unknown.
It took just one week of fatty, high-calorie eating for mice to lose a substantial share of a brain protein called PQBP1, according to new research published in Molecular Psychiatry. That protein acts like an editor, making sure the instructions for building synapses, the tiny junctions where brain cells pass signals to each other, come out correct. Without enough of it, those instructions get garbled, disrupting the connections brain cells use to communicate and potentially contributing to memory problems.
Mice on the fatty diet lost connections between brain cells and struggled on memory tests, and the researchers traced much of the problem back to that faulty editing.
Researchers did not stop at describing the damage. They followed the chain backward to PPARγ, a switch protein that normally keeps PQBP1 production up, and forward to the genes that get misread as a result. Then gene therapy delivered through engineered viruses, along with a diabetes drug that flips that switch back on, improved memory and brain-cell connections in the mice.
High-Fat Diet Insulin Spikes Appear to Silence a Brain Protein
Researchers worked with male lab mice that ate either standard chow or a high-fat, high-calorie diet for one week or six weeks, with eight mice per group in the main tests. By 12 weeks of age, the fatty-diet mice showed classic signs of poor metabolic health: weight gain, high blood sugar and elevated insulin.
Protein tests and tissue imaging of the cortex, the brain’s outer layer, showed PQBP1 dropped inside neurons in both the one-week and six-week groups. Levels held steady in two other cell types, the brain’s immune cells and its support cells. A repeat experiment in female mice found similar results, so the drop was not limited to males. Live imaging of the mouse brain added a visual: the tiny bumps on brain cells where connections form disappeared faster in the fatty-diet mice, leaving fewer connections overall.
To find out why PQBP1 fell, the team turned to PPARγ, a protein that flips other genes on and off. In lab-grown cells, including neurons from mouse brains, PPARγ boosted PQBP1 production, but adding insulin, which runs high on a fatty diet, canceled that boost. Silencing PPARγ also cut the boost, supporting the link.
Misread Synapse Genes Weaken Brain Cell Signaling in High-Fat Diet Mice
Next, the team hunted for genes that get misread when PQBP1 runs low. They compared the fatty-diet mice with a separate group bred to lack PQBP1 in their cortical neurons. Hundreds of genes showed the same scrambled instructions in both groups, and many were tied to synapse function.
A ranking of which genes mattered most narrowed the list to three standouts, Cask, Cacnb1 and Cyfip2, all tied to how synapses release chemical messengers. A fourth gene showed a very large but shorter-lived disruption, appearing only in the one-week group.
When misread, these genes produce proteins with missing pieces. Two of the three, Cask and Cacnb1, lost the ability to bind a partner protein needed to release chemical messengers, and in cultured mouse neurons, the faulty versions of all three disrupted that release.
Gene Therapy and a Diabetes Drug Improved Memory in High-Fat Diet Mice
Rescue experiments tested whether restoring normal protein levels could improve the damage tied to the diet. Pioglitazone, a drug already used for type 2 diabetes that switches PPARγ back on, was given for three or six weeks. Treated mice did better on tests of long-term spatial memory (a water maze), short-term spatial memory (a three-armed maze) and fear-based memory, and their brain tissue showed signs of restored connections. Treatment began alongside or partway through the diet, so the results show improvement, not a full cure.
Next, researchers delivered PQBP1 itself around the brain’s surface using a modified virus, a form of gene therapy. Memory and connections improved again, and electrical recordings agreed: neurons from mice fed the fatty diet for 10 weeks received fewer tiny signals, and the PQBP1 treatment brought those counts back up. Supplying any one of the three downstream genes alone helped less, with Cacnb1 helping most, which hints that PQBP1 manages several genes at once.
Pioglitazone did nothing for mice genetically missing PQBP1 in brain neurons. That result strengthened the case that restoring PQBP1 is central to the drug’s memory benefits in these mice.
Diet-Driven Brain Changes in Mice Do Not Yet Show Dementia Risk in People
Nothing here shows cheeseburgers cause Alzheimer’s disease, and no people were studied. PQBP1 has separately been found at reduced levels in neurons of people with Alzheimer’s, so diet-related damage and certain forms of dementia may converge on the same vulnerable protein. That overlap is speculative, and it is the detail worth following as the research moves forward.
A week of fatty eating was enough to disturb a key brain protein in mice, which makes diet-driven memory trouble a question worth testing in people.
Disclaimer: This article is for information only and is not medical advice. Findings come from mice, not people. Ask a doctor before changing diet or medication.
Paper Notes
Limitations
Every experiment was done in mice or lab-grown cells, so it is not yet known whether the same chain of events occurs in the human brain. The main diet experiments used male mice, with female mice described only in a repeat of the PQBP1 measurements. In most treatment experiments, therapy started before, alongside or partway through the diet rather than after long-term damage. The authors note that longer studies of PQBP1 in animals fed a high-fat diet would be useful, and that animal models better suited to studying communication between the brain and fat tissue, ideally primates, would help clarify open questions. They also write that pioglitazone may not work in people whose PQBP1 problems have already advanced, such as in advanced Alzheimer’s-related brain changes, and they point to earlier human trials of the drug for dementia that were negative in one case and positive in another. The authors state that further assessment of pioglitazone’s potential benefits for diet-associated dementia is necessary.
Funding and Disclosures
The study was supported by multiple grants to the senior author, including funding from the Japanese Agency for Medical Research and Development (AMED), a Grant-in-Aid for Scientific Research on Innovative Areas from Japan’s Ministry of Education, Culture, Sports, Science, and Technology (MEXT), and Grant-in-Aid for Scientific Research A funding from the Japan Society for the Promotion of Science (JSPS). One co-author’s funding support was provided by the National Institutes of Health. The authors declared no competing interests.
Publication Details
This article is based on the paper “PQBP1-dependent alternative RNA splicing underlies high calorie diet-induced cognitive impairment,” published in Molecular Psychiatry. The author list includes Yong Huang, Hidenori Homma, Yuki Yoshioka, Takaki Watanabe, Shuhei Fujino, Kyoko Matsuyama, Xigui Chen, Kyota Fujita, Hiroki Shiwaku, Albert R. La Spada, Masanobu Kano, Hikari Tanaka, and Hitoshi Okazawa, with affiliations including the Institute of Science Tokyo, the National Institute of Advanced Industrial Science and Technology, the University of Tokyo, Teikyo University, Kanazawa University, and the University of California Irvine. Corresponding author: Hitoshi Okazawa. DOI: https://doi.org/10.1038/s41380-026-03868-x. The paper was received July 26, 2024, revised August 12, 2026, accepted September 2, 2026, and published online September 12, 2026.
