"We already knew that genomic mutations or transcription factor–dependent gene expression are altered during cancer progression and can lead to the development of treatment resistance," Kim explained. "But we wanted to know more about how mRNA translation — also known as protein synthesis — is involved when the disease becomes more aggressive."

What did the researchers find?

Kim focused on one step of protein synthesis: the moment when tRNA reads genetic instructions from mRNA and delivers amino acids, the building blocks that cells assemble into proteins.

To test its role, she used:

• prostate cancer cell lines grown in the lab;
• mouse models of the disease;
• tumor samples from patients.

Across all three systems, the pattern held. Levels of tRNA1Arg(UCU) were high in androgen receptor–dependent prostate cancer cells but dropped in treatment-resistant tumors that relied less on the receptor.

The decisive experiment came next. When the researchers delivered this specific tRNA to treatment-resistant tumors, the tumors regained sensitivity to therapies that target the androgen receptor.

"We found that we can shift the cell state between an androgen receptor–dependent state and an androgen receptor–independent state with a single tRNA," Kim said. "This is important because changes in cell identity are a major reason prostate cancers become resistant to treatment. We found that this tRNA can influence whether cells remain in a drug-sensitive state or transition to one that is more aggressive."

Could this apply to other cancers?

The researchers see prostate cancer as a proof of principle rather than an isolated case.

"In this study, we used prostate cancer as an archetype to study tRNA-dependent state changes, but we think it's just the beginning," said Kim, who plans to pursue this line of research in her own lab. "We can apply this approach to any type of disease model or other types of cancers that undergo identity switches after treatment, such as lung and breast cancers."

Kim and Hsieh are now investigating whether tRNA could serve as a biomarker — a measurable signal that helps doctors predict how a disease will behave — or as a therapeutic target for aggressive prostate cancer. Either use could eventually help guide more precise treatment decisions or point the way to new therapies.

What are the limitations?

The results come from cell lines, mouse models and patient tumor samples, not yet from clinical trials in people. Much work remains before a tRNA-based approach could reach patients, including studies on how to deliver such a molecule safely to tumors in the body.

Still, the study establishes a previously unrecognized mechanism: small RNA molecules once considered passive participants can actively decide whether cancer cells stay treatable or turn aggressive.

The paper's co-corresponding authors are Andrew Hsieh, who holds the Larry and Virginia Gordon Endowed Chair in Prostate and Bladder Cancer Research at Fred Hutch and is a professor in the Division of Hematology and Oncology at the University of Washington School of Medicine, and Tao Pan of the University of Chicago. The study appears in Nature (2026), DOI: 10.1038/s41586-026-11153-8.