A seesaw between transcription and translation restrains bladder carcinogenesis

From the Hsieh Lab, Human Biology Division

What is it that makes a good story, one that catches your attention and refuses to let go? Does it include a peculiar mystery? A well-motivated conflict which craves resolution? A catchy hook? A recent study from the Hsieh Lab in the Human Biology Division at Fred Hutch, led by postdoctoral associate Dr. Sujata Jana, checks all of these boxes.

The Hsieh Lab broadly studies the interplay between transcriptional and translational regulation in cancer, with a particular focus on cancers of the genitourinary system. Their latest work, published in Cancer Cell, took aim at a mysterious protein called ARID1A—a chromatin-remodeling component that had recently been identified as a tumor suppressor commonly mutated in bladder cancers. As their name implies, tumor suppressors are genes which normally function to restrain cell growth and proliferation whose selective loss leads to cancer formation. Many notorious tumor suppressors—including TP53, PTEN, and APC—are thought to keep cells from becoming cancerous in part by negatively regulating translation—their loss leads to unrestrained translation of different oncogenic mRNAs, without which cancer would not be able to take hold. When Dr. Jana joined the Hsieh lab, it seemed that this would also hold true for ARID1A: loss of this gene had previously been shown to increase transcription of a collection of oncogenic genes in several tissues, which fit with the aforementioned paradigm of tumor suppression. In her own words, “when [Dr. Hsieh] pitched me this project, it seemed like a straightforward approach, like something which could be completed in a couple of years without too much difficulty.”

And so, Jana went to work creating a transgenic mouse model missing ARID1A in several tissues, verified that these mice did indeed have higher mRNA levels of a collection of oncogenic genes, and looked forward to analyzing the tumors of these mice to figure out how ARID1A did what it did. But then, the mystery: after growing the mice for over 400 days, not a single one developed a bladder tumor! While many would call it quits and find another project, Dr. Jana was determined to figure out why these mice didn’t develop any tumors. Turning back to the central dogma of biology, she reasoned that for tumors to develop, it wasn’t enough to have mRNA transcripts encoding oncogenic proteins present—those mRNAs needed to be translated into oncogenic proteins. So, Jana generated organoids from the basal urothelium of ARID1A-null and control mice and measured overall rates of protein synthesis in these organoids. Shockingly, she found that ARID1A-null organoids had lower protein synthesis rates than wild-type controls, indicating a striking departure from the standard model of tumor suppressor function. While ARID1A loss increased transcription of oncogenic genes, it concurrently decreased the translation of those same genes, presenting a plausible explanation for the lack of tumors in ARID1A-deficient mice. The concept of transcription and translation opposing each other to restrain tumor formation was so novel that Jana had to give this phenomenon a name: transcriptional-translational conflict.

It thus appeared that ARID1A—unlike most known tumor suppressors—functioned by promoting translation of oncogenic proteins (that is, its loss increased transcription but decreased translation of oncogenic proteins). But on a mechanistic level, how was it accomplishing this? Leveraging the strengths of her organoid model and the full force of genetic and pharmacological tools, Dr. Jana was able to zero in on a single suspect which stood out from the crowd: eEF2K, a kinase which regulates the elongation factor eEF2, a core component of cellular translation elongation machinery. ARID1A-null organoids had increased levels of active eEF2K (which decreases translation elongation rates), and inhibition or loss of eEF2K restored the protein synthesis rates of ARID1A-null organoids to match that of wild-type controls. Diving even deeper into large and complicated datasets of proteins, transcripts, and post-translational modifications in her organoids, Jana then found the crucial link between ARID1A and eEF2K: a small signaling protein called RASGRP1, whose expression the team showed was regulated at its promoter by ARID1A.

An artwork depicting two angry Greek gods: one representing transcription and one representing translation, fighting over a city on a hill representing the body.
An artistic interpretation of the conflict between transcription and translation serving a tumor-suppressive function in urothelium. Image credit: DrawImpacts

At this point, Jana and colleagues had a pretty convincing molecular map drawn which connected ARID1A loss to translation suppression: ARID1A modulates the expression of RASGRP1, which increases eEF2K activity, which then inhibits eEF2 to decrease translation elongation. However, the ultimate test of this pathway would be to disrupt it in vivo and see if doing so would promote tumor formation, which Jana then set out to do. Strikingly, mice lacking both ARID1A and eEF2K did form urothelial tumors; furthermore, treating cells from these mice with inhibitors against two of the oncogenic proteins modulated by ARID1A decreased their growth, suggesting de-repression of oncogenic proteins as the reason behind tumor formation in these mice. Altogether, these results describe transcriptional-translation conflict as a novel mechanism of tumor suppression and put ARID1A and eEF2K front and center as key regulators of this mechanism.

Faced with this novel tumor-suppressive molecular pathway, the team realized that their findings suggest a targetable vulnerability in ARID1A-deficient bladder cancers: if the loss of ARID1A causes a suppression of translation which must be overcome for tumors to form, then these tumors may be stopped by drugs which specifically inhibit translation. To put this speculation to the test, Jana treated organoids and mice harboring patient-derived xenografts (PDXs) with homoharringtonine (HHT), an FDA-approved translation inhibitor. Gratifyingly, this revealed that indeed, ARID1A-low organoids and tumors were much more sensitive to HHT (at physiological doses) than their ARID1A-high counterparts, capping off a satisfying result with a potentially novel therapeutic approach to boot.

All in all, what was initially expected to be a quick and straightforward project ended up taking nearly six years to complete, but Dr. Jana doesn’t regret this fact, and is enthusiastic to take her groundbreaking findings into the clinic to help patients. “With a challenging collection of in vivo systems and constant roadblocks or unexpected results, there were a lot of low points in this project, but it taught me that science is really a conversation with your data,” Jana notes. “I think it also illustrates the value of not giving up in the face of ‘weird’ results: I would encourage anyone to take the extra step and dig into results that tell them something they didn’t expect—that’s your system trying to tell you that there’s something interesting going on which we don’t understand.”

The spotlighted research was funded by the National Institutes of Health, the Kleberg Foundation, a Burroughs Wellcome Fund Career Award, the Emerson Collective, Nancy & Dick Bernheimer, the Matthews Family, Stinchcomb Family, and Thomas & Patricia Wright Memorial Funds, the Mark Foundation for Cancer Research, and the Department of Defense.

Fred Hutch/University of Washington/Seattle Children’s Cancer Consortium members Drs. Andrew Hsieh, Steven Henikoff, and Arvind Subramaniam contributed to this study.

Jana, S., Brahma, S., Arora, S., Wladyka, C. L., Hoang, P., Blinka, S., Hough, R., Horn, J. L., Liu, Y., Wang, L.-J., Depeille, P., Smith, E., Montgomery, R. B., Lee, J. K., Haffner, M. C., Vakar-Lopez, F., Grivas, P., Wright, J. L., Lam, H.-M., … Hsieh, A. C. 2023. Transcriptional-translational conflict is a barrier to cellular transformation and cancer progression. Cancer Cell. 41(5), 853-870.e13.