Jacksonville, Fla. — Mayo Clinic researchers have used artificial intelligence to identify an experimental small-molecule drug that targets a protein region long considered difficult to treat. The compound is designed to block the PDZ domain of GIPC1, a protein that supports the growth, survival and treatment resistance of several cancers, including pancreatic ductal adenocarcinoma. In laboratory studies, inhibiting this molecular interaction slowed tumor growth, extended survival in experimental models and strengthened the activity of the chemotherapy drug gemcitabine.
The findings, published in Cell Reports, offer an example of how computational drug discovery can be applied to targets that have resisted conventional pharmaceutical approaches. Rather than searching only for compounds that bind to the most obvious catalytic sites on a protein, the research focused on the PDZ domain, a structural region involved in protein–protein interactions. These interactions help organize signaling networks inside cells, but they are often broad, shallow or chemically difficult to occupy with conventional drugs.
GIPC1, or GAIP-interacting protein C terminus 1, functions as a scaffold that helps bring signaling molecules and membrane-associated proteins into coordinated complexes. Its PDZ domain can influence the stability, localization and activity of proteins involved in tumor biology. By disrupting this domain, the researchers sought to interfere with signaling that cancer cells use to proliferate, survive stressful conditions and resist therapy. The strategy is particularly relevant to pancreatic cancer, in which malignant cells frequently adapt to treatment and grow within a highly protective tumor environment.
To search for a suitable inhibitor, the team worked with Sravathi AI Technology for IP Sharing, a company based in Bangalore, India. Their computational screening program evaluated nearly 40,000 candidate compounds, prioritizing molecules predicted to interact with the GIPC1 PDZ domain. The process was intended to reduce the time and resources required to examine a large chemical space experimentally. After the initial selection, the researchers performed laboratory testing to determine whether the leading compound could engage GIPC1 and interfere with its biological function.
The resulting inhibitor showed activity in models of pancreatic ductal adenocarcinoma. Blocking GIPC1 reduced cancer-related growth signals and limited the ability of tumor cells to maintain aggressive behavior. In animal studies, treatment with the compound was associated with slower tumor progression and longer survival compared with untreated controls. The investigators also reported that the inhibitor enhanced the effects of gemcitabine, a commonly used chemotherapy drug for pancreatic cancer, suggesting that GIPC1 blockade may eventually become part of a combination-treatment strategy.
Pancreatic cancer remains one of the most challenging malignancies because it is often detected only after it has reached an advanced stage. Tumors can grow around blood vessels and other critical structures, while dense connective tissue can restrict the delivery of medicines. Cancer cells also display extensive molecular diversity, allowing some populations to survive chemotherapy and repopulate the tumor. These features contribute to the disease’s poor prognosis and have created an urgent need for treatments that attack pancreatic cancer through mechanisms distinct from established chemotherapy.
The study also produced early evidence that the experimental drug may affect the tumor microenvironment, the network of noncancerous cells, extracellular matrix components and signaling molecules surrounding a tumor. This environment can shelter malignant cells, suppress immune activity and reduce the penetration or effectiveness of drugs. Although the precise changes induced by GIPC1 inhibition require further investigation, the findings raise the possibility that targeting the protein could make tumors more responsive to chemotherapy or other future treatments.
Artificial intelligence is increasingly being used in drug discovery to predict how molecules may fit into protein structures, estimate binding properties and prioritize candidates for laboratory testing. Its value is especially apparent when researchers are addressing proteins without conventional enzyme pockets or other easily targeted sites. However, computational predictions do not establish that a compound will be effective or safe in a living organism. Candidate molecules must still undergo biochemical validation, cell-based testing, animal studies and, eventually, carefully controlled clinical trials.
“Our study demonstrates that AI can help us identify entirely new therapeutic opportunities against targets that have historically been considered undruggable,” says Debabrata (Dev) Mukhopadhyay, Ph.D., senior author and a cancer researcher at Mayo Clinic in Florida. He cautions that the findings remain preclinical, but says they provide a foundation for the next phase of research. The team must now determine how the inhibitor behaves in the body, establish an effective dose, assess potential toxicity and clarify which patients might benefit from treatment.
The experimental therapy is not approved for use in people and has not yet entered clinical trials. Additional studies will be needed to confirm its selectivity for GIPC1, evaluate possible effects on healthy tissues and test whether its activity can be reproduced across additional pancreatic cancer models. If those investigations are successful, the work could demonstrate that AI-guided targeting of protein–protein interaction domains is a practical route toward new cancer medicines, particularly for tumors that have remained resistant to conventional therapeutic design.
Subject of Research: AI-guided discovery of a small-molecule inhibitor targeting the GIPC1 PDZ domain for pancreatic ductal adenocarcinoma.
Article Title: AI-driven discovery and validation of a GIPC1 PDZ domain inhibitor for pancreatic ductal adenocarcinoma
Web References: Mayo Clinic; Cell Reports study; Mayo Clinic pancreatic cancer information
References: Cell Reports, “AI-driven discovery and validation of a GIPC1 PDZ domain inhibitor for pancreatic ductal adenocarcinoma,” published 31 July 2026.
Keywords: pancreatic cancer, pancreatic ductal adenocarcinoma, GIPC1, PDZ domain, artificial intelligence, AI drug discovery, cancer therapy, gemcitabine, small-molecule inhibitor, tumor microenvironment, Mayo Clinic, preclinical research
