How Artificial Intelligence Helped Reveal a Hidden Cause of Alzheimer’s

A fascinating breakthrough at UC San Diego has shown how artificial intelligence (AI) is not just reshaping technology—it’s reshaping medicine too. Researchers uncovered that a gene called PHGDH, long thought to be just a biomarker for Alzheimer’s disease, actually plays a causal role in triggering it. And surprisingly, it wasn’t the gene’s usual enzymatic function causing trouble—it was a hidden ability to disrupt how other genes in the brain are turned on and off.

The UCSD team, led by bioengineering professor Sheng Zhong, used advanced AI to build a precise 3D model of the PHGDH protein. In doing so, they discovered a “moonlighting” function of the protein: it could bind to DNA and throw off the brain’s delicate gene regulation process. This disturbance can kickstart the early stages of Alzheimer’s, offering fresh clues about why so many people develop the disease without the usual genetic mutations typically associated with it.

From Biomarker to Culprit: PHGDH’s Double Life

For years, PHGDH was considered just an indicator for Alzheimer’s, not a cause. Initial studies found higher levels of the PHGDH gene’s products—proteins and RNA—in Alzheimer’s patients. That alone raised eyebrows. But until now, nobody knew it was actually doing something harmful.

The research team took a bold step by altering PHGDH levels in mouse models and human brain organoids. What they found was striking: when PHGDH levels were lowered, Alzheimer’s progression slowed. When they were raised, the disease worsened. This confirmed that PHGDH isn’t just a marker—it’s a driver.

Digging deeper with the help of AI modeling, they identified a previously unknown structural feature of the PHGDH protein. It closely resembled a DNA-binding domain seen in transcription factors—proteins that control whether genes are active or inactive. That meant PHGDH could bind to DNA and mess with brain cells’ normal operations. And sure enough, it activated two critical genes in ways that promoted disease development.

A Promising New Treatment Candidate: NCT-503

Discovering a new cause of Alzheimer’s is thrilling on its own. But the UCSD team went even further: they may have found a way to stop it.

Enter NCT-503, a small molecule drug originally studied for other purposes. Importantly, NCT-503 doesn’t interfere with PHGDH’s regular role in making serine, an essential amino acid. Instead, it targets the newly discovered DNA-binding function that seems to trigger Alzheimer’s.

When tested in two different Alzheimer’s mouse models, NCT-503 dramatically reduced disease progression. Treated mice not only showed less memory loss but also experienced reduced anxiety—both hallmark symptoms of Alzheimer’s.

The hope now is to optimize NCT-503 and move toward human clinical trials. Because it’s a small molecule drug, it could even be taken orally in the future, unlike many current Alzheimer’s treatments that require infusions.

The Road Ahead: From Discovery to Cure

Of course, there’s still a long road before any new drug reaches pharmacy shelves. Mouse models can’t perfectly mimic the human experience of Alzheimer’s, and there’s always a risk that what works in animals won’t translate to people. But the findings are incredibly exciting.

Unlike treatments that target sticky amyloid plaques after they’ve already caused damage, this new approach strikes much earlier in the disease process. By stopping the harmful gene regulation before it spirals out of control, we may finally have a shot at truly slowing or even preventing Alzheimer’s, rather than just treating symptoms.

As Professor Zhong put it, “Now there is a therapeutic candidate with demonstrated efficacy that has the potential of being further developed into clinical tests.”

This discovery showcases the power of blending AI with medicine—and it brings new hope to millions of families affected by Alzheimer’s around the world.

To dive deeper into the full study, you can visit the original article here: AI Uncovers New Cause of Alzheimer’s.

My Take: AI, Alzheimer’s, and the Future of AI Robotics

This discovery shows just how powerful AI is becoming—not just in analyzing data, but in transforming how we understand life at a molecular level. By revealing hidden genetic mechanisms, AI is helping us move beyond symptom management and toward true prevention.

Looking ahead, the impact on AI robotics could be profound. Imagine robotic systems embedded with AI models that can assist doctors in real time, predict neurodegenerative risks earlier, and even design custom treatments. We are stepping into an era where AI-driven robotics won’t just care for aging populations—they’ll prolong healthy lives. This research is a glimpse into that future, and it’s incredibly exciting.

Keywords: AI and Alzheimer’s, Alzheimer’s genetics, Alzheimer’s treatment options

Source: Transcriptional regulation by PHGDH drives amyloid pathology in Alzheimer’s disease