High-fat diet lowers brain protein PQBP1, impairs cognition in mice

In the cerebral cortex of mice fed a high-fat diet, the team led by Hitoshi Okazawa found less of a single protein: PQBP1. The cerebral cortex is the brain region that supports cognitive functions, and the drop in this protein was enough to disturb how neurons communicate.

Calorie-rich diets are known to contribute to obesity and metabolic disorders such as diabetes. Growing evidence suggests they also affect the brain, and studies have linked them to poor cognitive performance and an increased risk of neurodegenerative conditions such as Alzheimer's disease. Earlier research had pointed to inflammation and insulin resistance, but the changes inside neurons behind this damage remained poorly understood. PQBP1 helps explain part of that gap: it regulates alternative RNA splicing, a process by which one gene can produce different RNA transcripts and potentially different protein versions.

To see what the lower level of PQBP1 did to neurons, the researchers used RNA sequencing to compare mice fed a high-fat diet for one or six weeks with mice lacking PQBP1 in their neurons. Both groups showed shared splicing changes in specific genes involved in synaptic function. Further experiments showed that this abnormal splicing led to abnormalities in synaptic structure and function, alongside cognitive impairment. The study links the drop in PQBP1 to suppressed PPARγ-mediated transcription of the PQBP1 gene, and it implicates insulin-related signaling as a possible upstream mechanism.

The team then tested whether the damage could be reversed. Pioglitazone, a diabetes drug that activates PPARγ, restored PQBP1 levels in high-fat-fed mice, corrected the abnormal splicing of key synapse genes, and improved synaptic and cognitive outcomes. Gene therapy to restore PQBP1 also rescued synaptic and cognitive deficits, and restoring some downstream genes had beneficial effects as well. Everything reported here comes from mice, and the study was published in Molecular Psychiatry on Sept. 12, 2026.

For people, the study does not yet change anything at the bedside. What it offers is a specific target: PQBP1-dependent splicing, which the authors describe as a potential route to interventions for conditions involving cognitive dysfunction. Okazawa's team says further research is needed to translate these findings into therapeutic strategies, and it gives no timeline and reports no human trial. Pioglitazone is already used for diabetes, but these experiments tested it only in mice fed a high-fat diet.