Researchers uncover novel dual-target mechanism in treatment-resistant ovarian cancer

Media contact: Nicole Fawcett, 734-764-2220 | Patients may contact Cancer AnswerLine™ 800-865-1125

AI-driven discovery demonstrates platform's capability to identify new disease biology and therapeutic approaches

Photo credit: Michigan Medicine

ANN ARBOR, Michigan – Researchers from the University of Michigan Rogel Cancer Center and BenevolentAI have identified a promising new therapeutic approach for platinum-resistant ovarian cancer.

High-grade serous ovarian cancer affects over 300,000 women globally each year, with up to 90% of patients developing resistance to standard platinum-based chemotherapy. The disease has been particularly challenging to treat due to a lack of consistently targetable proteins.

The research, published in Molecular Cancer Therapeutics, leveraged technology from BenevolentAI called the Benevolent Platform, to analyze extensive biomedical data, including scientific literature and genomic databases. The platform connected seemingly disparate data points, which identified targets that had not previously been investigated in ovarian cancer, despite being present in clinical data.

The analysis prioritized 74 promising targets for experimental testing, leading to the identification of two key molecular drivers of the disease: TRAF2 and NCK interacting kinase, or TNIK, and cyclin-dependent kinase 9, or CDK9.

The study demonstrated that patients with elevated levels of both TNIK and CDK9 have significantly shorter survival rates. Further molecular analysis also revealed a novel interaction between these proteins, with CDK9 playing a previously unknown role in activating the Wnt signaling pathway, a crucial molecular mechanism that enables cancer cells to resist chemotherapy.

"This research represents a significant advancement in our understanding of treatment resistance in ovarian cancer," said study author Analisa DiFeo, Ph.D., professor of pathology and obstetrics and gynecology at Michigan Medicine. "By uncovering the interaction between TNIK and CDK9, and validating their combined role in treatment resistance across multiple patient-derived models, we have identified a promising new therapeutic approach. The discovery that CDK9 regulates the Wnt signalling pathway could have far-reaching implications for treating other cancers where these mechanisms are active."

The researchers validated these findings using NCB-0846, a documented TNIK inhibitor identified through the platform's analysis. Testing in patient-derived 3D models showed NCB-0846 effectively targeted both TNIK and CDK9, demonstrating superior efficacy compared to targeting either protein individually. These highly disease-relevant models, including ex vivo organoids and patient-derived cell lines, provide some of the most clinically translatable pre-clinical evidence possible, offering stronger validation than traditional cell line approaches.

"This dual-targeting approach highlights our AI platform’s ability to uncover new disease biology. This discovery exemplifies the power of combining AI-driven predictions with experimental validation in patient-derived models," said Ivan Griffin, D.Phil., co-founder and chief business officer of BenevolentAI.

Additional authors: Noah Puleo, Harini Ram, Michele L. Dziubinski, Dylan Carvette, Jessica Teitel, Sreeja C. Sekhar, Karan Bedi, Aaron Robida, Michael Nakashima, Sadaf Farsinejad, Marcin Iwanicki, Wojciech Senkowski, Arpita Ray, TJ Bollerman, James Dunbar, Peter Richardson, Andrea Taddei, Chantelle Hudson

Funding for this work is from National Cancer Institute grant P30 CA046592, the Silver Family Foundation, the Barbara Robson Foundation, and the Debra A. and Dean A. Frick Ovarian Cancer Research Fund

This work was supported by these Rogel Cancer Center Shared Resources: Flow Cytometry, Proteomics, Structure and Drug Screening, Cancer Data Science

Disclosure: None

Paper cited: “Identification of a TNIK-CDK9 axis as a targetable strategy for platinum-resistant ovarian cancer,” Molecular Cancer Therapeutics. DOI: 10.1158/1535-7163.MCT-24-0785