AI helps find promising therapeutic target for rare salivary gland cancer

Researchers have uncovered a promising therapeutic target for adenoid cystic carcinoma (ACC), a rare cancer of the salivary glands with few treatment options. The study, led by researchers from the University of Chicago Medicine Comprehensive Cancer Center, in collaboration with scientists from InSilico Medicine and Prelude Therapeutics, found that blocking the activity of protein arginine methyl transferase 5 (PRMT5) is a potential treatment strategy against ACC.

ACC accounts for only 1-5% of head and neck cancers and 25-35% of salivary gland neoplasms.

“The disease itself is very, very, very rare, which makes it very difficult to study,” explained Evgeny Izumchenko, PhD, Assistant Professor of Medicine at UChicago. He clarified that, in addition to being rare, it is also hard to detect early because patients tend to not show symptoms until it has progressed significantly.

“Not much is known about this disease; not much is known about how to treat it; you don't have the rich history of treated patients that you can look back to and define what would be the best approach,” Izumchenko said.

Given the lack of targeted therapies for ACC, the team turned to artificial intelligence (AI), which has been gaining traction in the discovery of novel therapeutic targets. Using an AI-based predictive discovery tool, the team analyzed gene expression data from 87 ACC tumor samples and 35 matched normal controls to identify potential drug targets. Among the top-ranked candidates was PRMT5, an enzyme implicated in epigenetic regulation (altering gene expression and protein activity without changing the DNA sequence itself) and known to play a role in cancer development.

Once they identified PRMT5 as a promising target, the researchers collaborated with Prelude Therapeutics, a company that had developed a highly selective PRMT5 inhibitor called PRT543. The researchers evaluated PRT543 in cellular and animal models, including ACC cell lines (cells derived from cancer tissues), organoids (3D tumor models derived from patient samples), and patient-derived xenografts (PDXs), which are human tumors implanted in mice.

“Organoids are better for assessing the drug response because they better represent the genetic composition of the cancer compared to cell lines. It gives you more confidence that your compound is working because you inhibit the same cells that drive the progression of the cancer,” explained Izumchenko.

The results demonstrated that PRMT5 inhibition significantly suppressed tumor growth across multiple preclinical models, downregulating key ACC-associated genes such as MYB and MYC. The treatment also appeared to be effective regardless of whether the tumors carried mutations in NOTCH1, a gene associated with more aggressive ACC.

While early clinical trials have found that PRMT5 inhibitors show anti-tumor activity in various cancers, their effectiveness in ACC had not been fully explored until now.

“We found that while the drug is effective, it is not a miracle drug. It inhibits and shrinks tumors, showing effects on cell lines, organoids, and PDX models, but it doesn't cure the disease,” Izumchenko said.

To get closer to a real “cure,” the researchers explored potential combination treatments to enhance the effectiveness of PRT543.

“Right now, in cancer treatment, ‘combination’ is pretty much the keyword. You combine several drugs together to inhibit key pathways, and you hope that the combination of the two or three drugs together will have a better effect than each drug alone,” explained Izumchenko.

The team explored drugs already approved for solid tumors and identified Lenvatinib, a multi-kinase inhibitor, as a potential partner. The combination treatment led to a stronger inhibitory effect on tumor growth in vitro.

Notably, a subset of UChicago Medicine patients used in the researchers’ analysis exhibited high PRMT5 expression alongside elevated levels of MYC, MYB, and Lenvatinib target genes. These findings suggest that targeting PRMT5 signaling in combination with Lenvatinib could be a promising strategy for patients with this specific molecular profile.

“We think that patients that have this molecular signature are potential candidates for the combination treatment,” Izumchenko noted. “The patients that have high levels of PRMT5, MYB, and MYC, but not Lenvatinib targets, are the patients most likely to benefit from monotherapy with a PRMT5 blocker without requiring a combination therapy.”

This study highlights a shift toward more personalized and targeted therapies for treating cancers.

“Developing something that can specifically inhibit the tumors in a targeted fashion, that’s exciting. Even if the drug will not work perfectly as a monotherapy, by combining it with a lower dosage of chemotherapy, the patient can benefit while reducing the amount of side effects,” Izumchenko said.

The study, “PRMT5 inhibition has a potent anti-tumor activity against adenoid cystic carcinoma of salivary glands,” was supported by grants from the National Institutes of Health. The study was also made possible with the support of the ACC Research Foundation and the UChicago Pathology, Genomic and Organoid and Primary Culture Research Cores.

Additional authors included Vasudha Mishra, Alka Singh, Xiangying Cheng, Claudia Wing, Manu Sundaresan, Anna Trzcinska, Everett E. Vokes, Grayson Cole, Le Shen, Yuxuan Miao, Alexander T. Pearson, Mark W. Lingen, Ari J. Rosenberg and Nishant Agrawal from The University of Chicago; Alex Zhavoronkov, Michael Korzinkin, Viktoria Sarkisova, Ivan Ozerov, Alexandra Pogorelskava, Oksana Glushchenko and Frank W. Pun from InSilico Medicine; Ashwin L. Koppayi from Northwestern University; Venkat Thodima, Jack Carter, Koichi Ito, Peggy Scherle and Bruce Ruggeri from Prelude Therapeutics.