Mayo Clinic scientists say AI led them to an experimental drug candidate for GIPC1. The protein was long considered undruggable, and lab results showed slower tumor growth in pancreatic cancer models.
Working with Sravathi AI Technology for IP Sharing, a Bangalore-based company, the team screened nearly 40,000 candidate compounds and identified a small molecule that blocks the PDZ domain of GIPC1, a protein that supports the growth of several cancers, including pancreatic cancer, and helps them resist treatment.
Lab experiments found the compound slowed tumor growth and began reshaping the tissue around tumors, a change researchers think could enable combination strategies. Pairing the compound with gemcitabine produced a stronger effect than the drug alone.
The senior author, Debabrata (Dev) Mukhopadhyay, works at Mayo Clinic Florida. He noted that pancreatic tumors adapt fast and defeat most available treatments. That is why the field keeps hunting new strategies, he said. His take: the research demonstrates AI can reach targets that resisted drug design for years. The findings remain preclinical.
The drug has not been tested in people, and the team says further studies are needed on safety and whether the approach can ultimately benefit patients.
