[Adobe Firefly]
Researchers at the University of California San Diego believe they have found an early pinch point in the biology of Alzheimer’s disease, one that can be blocked without disturbing the brain’s normal chemistry. The scientists credit AI for guiding them there.
The study, published April 23 in Cell, focused on phosphoglycerate dehydrogenase, or PHGDH, an enzyme best known for producing the amino acid serine. AI tools later proved crucial in analyzing the protein’s structure after initial hypotheses failed. So Zhong and his team took a closer look at PHGDH, which they had previously discovered as a potential blood biomarker for early detection of Alzheimer’s disease through analyses of human blood and brain-tissue datasets.
PHGDH’s conventional role is straightforward: it creates an enzyme key for the production of serine, an essential amino acid and a neurotransmitter. But the models suggested the protein had a secondary role. Instead of simply making serine, PHGDH appeared to be influencing the activity of downstream genes (specifically IKKa and HMGB1) involved in pathways that suppress autophagy and accelerate amyloid pathology, processes linked to memory loss and anxiety in Alzheimer’s patients. That upstream-intervention logic set the team on a drug-discovery path distinct from most Alzheimer’s programs which tend to focus on treating the abnormal buildup of beta-amyloid plaques.
Because the researchers knew of PHGDH’s enzymatic activity, they hypothesized that its metabolic function could be tied to Alzheimer’s outcome, yet their first experiments to show as much failed, as the press release notes. “At that time, our study hit a wall, and we didn’t have a clue of what mechanism it is,” said study senior author Sheng Zhong, a professor in the Shu Chien-Gene Lay Department of Bioengineering at the UC San Diego Jacobs School of Engineering, in the announcement.
Building an in silico activity map was only the first step towards unraveling PHGDH’s deeper role. The group then screened small-molecule libraries, again with an AI boost, for compounds that could block PHGDH’s regulatory antics without sabotaging its enzymatic output. One small molecule, known as NCT-503, stood out because it spares PHGDH’s serine-synthesizing activity while targeting its non-canonical gene regulatory function.
Dampening abnormal gene-expression patterns
Mouse work filled in the next set of blanks. When NCT-503 was administered to mouse models of familial Alzheimer’s disease (5XFAD and APP-KI), the compound dampened aberrant gene-expression patterns without altering brain serine levels. Behavioral assays showed reduced anxiety-like activity and improved memory performance. These behavioral improvements correlated with reduced amyloid-beta plaque load in the brain. That suggests that targeting PHGDH’s upstream regulatory role can mitigate downstream pathology
In the long run, the researchers’ discovery could pave the way for a new treatment for a disease that afflicts about one in nine people aged 65 and older, a population for whom treatment options remain limited despite FDA approvals since 2021 including aducanumab (now withdrawn from the market), lecanemab (approved in 2023), and donanemab (approved in 2024). The authors suggest that therapeutic strategies targeting PHGDH’s newly discovered transcriptional regulation function, potentially via the PHGDH–IKKa–HMGB1 axis, could hold promise in managing late-onset AD. The authors specifically note limitations such as testing the inhibitor NCT-503 in familial AD mouse models (though the mechanism was linked to LOAD models) and the use of only two behavioral assays to assess cognitive improvements..
Why chasing PHGDH holds promise
With the mechanistic link established and promising mouse efficacy demonstrated, the team is now optimizing analogs of NCT-503. The researchers selected this initial candidate partly owing to its ability to interact with PHGDH’s regulatory domain while largely sparing its essential enzymatic function, and its known ability to cross the blood-brain barrier. As is always the case, further work aims to enhance properties like pharmacokinetics ahead of potential investigational new drug (IND) application filings. The researchers have not yet shared a specific FDA timeline, but scale-up chemistry and toxicology studies are underway.
By targeting PHGDH’s upstream influence on gene expression, specifically its non-enzymatic role in transcriptional regulation identified in this study, rather than mopping up plaques downstream, the approach could theoretically prevent or delay both cognitive and neuropsychiatric symptoms. Because the mechanism appears independent of familial AD mutations, it might benefit a broader patient population. If validated, such small-molecule inhibitors, which could potentially be administered orally, might complement existing therapies by combining upstream and downstream interventions in Alzheimer’s disease.
Financing for the project came in part from the National Institutes of Health. Zhong is also a founder and shareholder of Genemo, Inc. and Neurospan, LLC, disclosures noted in the Cell paper.
Filed Under: Neurological Disease
