Fred Hutch researchers discover molecular switch that helps prostate cancer quickly change its biology to evade therapy

Many cultures tell stories about shapeshifters: gods becoming animals, animals becoming humans, demons switching into forms both beguiling and terrifying.

In the past six years, cancer biologists at Fred Hutch Cancer Center and other research institutions have begun telling their own scientific stories about a different kind of shapeshifting called lineage plasticity.

The study, published in The Journal of Clinical Investigation, reveals a surprising trick prostate cancer uses to quickly shapeshift and escape therapy without going through the cumbersome process of reprogramming its DNA.

Instead, cancer hijacks a process called translation, which happens much further down the line in the production of proteins, the complex molecules that do the cell’s work.

Hsieh and Mishra discovered that the same mechanisms that aggressive prostate cancer cells manipulate to elude therapy can be used against them, forcing them to shapeshift back into their more treatable form.

“This is the first time anyone's ever shown that the identity of the cell can be regulated by inhibiting translation,” said Hsieh, who holds the Larry and Virginia Gordon Endowed Chair in Prostate and Bladder Cancer Research.

The discovery opens a new line of attack against aggressive prostate cancer, as well as breast cancer and possibly other cancers.

But as often happens in science, the shapeshifting discovery wasn’t the story they set out to tell at all.

Staging a level playing field

Hsieh’s lab in the Human Biology Division of Fred Hutch focuses on the role of translation in cancers of the prostate and bladder and how it affects runaway tumor growth.

Translation involves messenger RNA, a go-between molecule that copies DNA codes for building proteins and then delivers those work orders to the cell’s many protein factories, which are called ribosomes.

The ribosomes read the mRNA transcripts and translate genetic sequences into amino acid sequences. The resulting amino acid chains eventually fold into a multitude of proteins that are essential for cell survival.

When Hsieh came to Fred Hutch in 2014, he was among the few researchers in the world who studied the role of mRNA translation in cancer.

Most research at the time focused on earlier stages of the protein-making process such as cancer-causing mutations in the genes themselves and cancer’s exploitation of transcription to stamp out work orders favorable to cancer’s growth, spread and survival.

Until recently, translation of those work orders has been considered a relatively straightforward, generic process that warranted little attention because the important action driving cancer usually occurs much further up the line.

Hsieh’s research, as well as others, has shown that mRNA translation isn’t so straightforward, especially in aggressive prostate cancer that evolves to evade therapy.

But figuring out how translation matters for treating cancer has generated conflicting explanations.

“Our work has shown that if you inhibit translation, that has an anti-cancer effect in prostate cancer,” Hsieh said. “Other people have shown that if you ramp up translation, that also kills prostate cancer.”

Some drugs for aggressive prostate cancer inhibit translation and some promote it, but these results come from different labs using different cell lines or other preclinical models.

“We're making these conclusions, but we're comparing apples to oranges in terms of the models,” Hsieh said.

Mishra, one of the postdoctoral researchers in the Hsieh Lab, devised a way to stage a head-to-head competition among known inhibitors and promoters of translation using the same preclinical model and method to ensure apples-to-apples comparisons.

“Rashmi came up with a clever screen to level the playing field,” Hsieh said. “That's where the whole thing started.”

It took about a year to gather all the contestants for her screen: a handful of drugs or compounds in development that either hit the brakes on translation or floor the gas pedal. They applied the contestants to a variety of tumor cells that varied by how responsive they were to standard therapy.

The clear winner turned out to be a brake that inhibited translation, but the results of the screen were so surprising that it spurred them to figure out exactly how that brake works.

Understanding how the clear winner in the screen works

The mRNA translation process requires the assembly of the ribosome — a molecular complex that is separated into a small piece and a big piece when it’s in stand-by mode waiting for an mRNA to translate.

First, the small piece latches on to a handle tucked into the protective cap at the end of an mRNA strand. After a quick scan of the code to find the right starting place, the big piece joins the small piece, and the two halves lock onto the strand like pursed lips slurping a noodle.

All the grabbing, scanning, assembling and slurping that initiates translation comes together with helper proteins called eukaryote initiation factors, or eIFs.

One of those helper proteins, eIF4E, kicks the whole thing off by grabbing the handle at the business end of a messenger RNA.

When this helper’s activity is elevated in cancer, it acts like a gatekeeper, selectively grabbing mRNA transcripts for translation that will help the cancer grow and spread.

The winning brake in Mishra’s screen is a cap-inhibitor drug that interrupts the process by lodging itself in the grabby area of eIF4E, causing it to grab the drug instead of an mRNA.

Fewer mRNA transcripts get grabbed and translated into proteins, which disrupts cell growth.

But Mishra discovered that the cap-inhibiting drug doesn’t work equally well on all kinds of tumor cells. For example, it doesn’t work on cells that are still vulnerable to standard therapy. It only slams the brakes on the more lethal kind that have become drug-resistant.

Why only those?

To answer that question, Mishra needed to figure out precisely which proteins are affected by the cap inhibitor.

Developing a technique to identify newly made proteins

Conventional ways of measuring protein synthesis don’t discriminate between newly made proteins and all the proteins already floating around in the cell before the drug is administered.

So, Mishra coupled a new biochemistry technique that isolates newly produced proteins with mass spectrometry, providing a more accurate snapshot of translational activity at specific times to measure the drug’s effect on protein production.

“We can quantify how much active protein synthesis is happening in whatever time frame we want to look at,” Mishra said.

The analysis revealed something astonishing. The cap inhibitor drove down protein synthesis by about 40% overall, but not uniformly across all proteins.

It was disproportionately effective at reducing the production of key structural proteins called keratins, which affect a cell’s shape and strength.

Keratins vary depending on whether the cell originates in an outer, basal layer of the prostate gland or in the luminal layer that lines the inside of the tiny hollow tubes in the prostate gland that produce and secrete prostate fluid.

Prostate cancer usually originates in the luminal cells and while it remains in that luminal state, it’s susceptible to therapy. When those cells shapeshift into a basal form, expressing basal keratins, the cells become drug-resistant.

Mishra’s innovative approach showed that when the cells are treated with cap inhibitors, fewer basal keratin proteins get made.

But crucially, it’s not because the genes for basal keratin proteins are kicking out fewer mRNA transcripts. That part of the protein-making process doesn’t change.

Whatever is going on to drive down the production of these basal keratins happens after transcription but before the resulting mRNA transcripts get grabbed and translated into proteins.

Mishra discovered that basal keratins — but not luminal keratins — share a common signature, or motif, embedded in the part of the transcript that doesn’t get translated because it includes the set-up instructions telling the ribosome where to begin.

That untranslated stretch of code is right next to the cap, the part that gets grabbed by the helper protein eIF4E.

Mishra deleted the motif and observed how the altered sequence responded to the cap inhibitor.

“We reasoned that maybe this is a motif that is important for the drug's ability to work,” Hsieh said. “The drug did not work when we took out this motif.”

That explains why cells treated with the cap inhibitor disproportionately lose more basal keratins and therefore lose their basal identity.

The results confirmed that these basal keratin proteins are not only markers of the basal subtype, they’re also necessary for the basal subtype’s survival.

But it also suggested something more profound.

It was known that cancer can hijack the helper protein eIF4E and use its selective powers to grab more cancer-promoting transcripts, which is why the cap-inhibiting drug was developed in the first place.

But Hsieh and Mishra showed that blocking eIF4E can change a prostate cancer cell’s whole identity in a clinically relevant way.

Nothing they could find in the prostate cancer literature suggested that inhibiting translation in this way could make such a big difference in the destiny of a cell.

This was a new story in shapeshifting lore.

Tracking the transition to a more treatable form

The signature Mishra had discovered makes it possible to better identify proteins essential for basal identity and block their translation with a drug.

But that was only half the story because the cap inhibitor didn’t significantly change the production of luminal keratins.

To find out if the cap inhibitor also makes cells more luminal — and therefore more vulnerable to standard therapy — they needed to count something other than keratins.

Luminal cells also produce high levels of androgen receptors, the proteins that enable cancer cells to respond to hormonal signals that drive growth.

Drugs such as enzalutamide attack those androgen receptors so they can no longer channel growth-promoting hormones into the cell.

“Enzalutamide is a very common drug that we give to prostate cancer patients,” Hsieh said. “Now, we actually give this type of therapy right away with androgen deprivation therapy if patients come in with metastatic disease.  Essentially, we're trying to crush the activity of the androgen receptor in prostate cancer patients.”

But usually within a few years, most patients with metastatic disease become resistant to the treatment.

Their cancer has shapeshifted from a luminal state where there’s a lot of androgen receptors to a basal state, which produces far fewer androgen receptors and therefore fewer targets for the drug to hit.

Misha confirmed that cells treated with cap inhibitors also produce more androgen receptors in addition to producing fewer basal keratins.

They become less basal and more luminal at the same time.

But the increase isn’t because more androgen receptor mRNA transcripts get grabbed and translated.

Something else pushes the levels of androgen receptors higher.

To figure out what, they collaborated with a colleague in the Human Biology Division, Haolong Li, PhD, who tracks changes in androgen receptor levels dynamically in living tumor cells to find proteins that affect their signaling.

Collaboration leads to discovery of a second mechanism

Merging their data with Li’s data on proteins that affect androgen receptor levels, they discovered how the cap inhibitor indirectly boosts androgen receptors.

The drug turns up the production of two other proteins that help stabilize androgen receptors by preventing their degradation.

The gatekeeping switch — the helper protein eIF4E — allows more of these stabilizing proteins to get grabbed and translated, which in turn boosts the androgen receptor levels so they more closely resemble the luminal state of the cell.

Hsieh and Mishra reasoned that tumor cells threatened with a drug targeting androgen receptors exploit this gatekeeping switch to shapeshift from its vulnerable luminal form into a basal state that defeats the drug by taking away its targets.

Cancer cells evolve to make this shift quickly without circling back to DNA for new work orders.

Instead of wasting time turning on and off genes, making new mRNA transcripts and translating those transcripts into proteins to evade the drug, the tumor just flips this switch. The hijacked eIF4E grabs more of the transcripts needed for basal identity while simultaneously grabbing fewer of the transcripts needed to stabilize androgen receptors.

The cap inhibitor works because it flips the switch back, reprogramming basal cells toward a luminal state.

Mishra had followed the trail from a simple competition pitting translation brakes and gas pedals to discover not one but two mechanisms that can be blocked with a drug to push aggressive prostate cancer back into a more treatable form.

“She spearheaded all this work,” Hsieh said. “She kind of went down the rabbit hole of each one and then came up with these mechanisms that together demonstrate how translation can drive lineage transitions.”

Finding patterns in other cancers

The next step was to confirm if the switch also works in living organisms.

They tested the cap inhibitor on a variety of patient tumors grown in mice and found a decrease in tumor growth.  When they combined the cap inhibitor with enzalutamide in aggressive “basal” tumors, they got even more dramatic results.

“If I just did the cap inhibitor, I would see some progress,” Mishra said. “But when we combine the drug with enzalutamide, the tumors were not just reduced in growth, but they were shrinking.”

Many cancers have molecular subtypes that make some more treatable with drugs and others more drug-resistant, so Mishra tested some of those as well.

“Beyond prostate cancer, we tested breast cancer cell lines,” she said. “And there we also saw the same pattern, which was the basal breast cancer cells were more sensitive than the luminal ones.”

Pancreatic ductal adenocarcinoma, or PDAC, also has clinically relevant subtypes, but the cap inhibitor didn’t work on them.

“We also tried it in PDACs, but we don't see the same effects,” Mishra said.

That negative result shows that mRNA translation is much more complex and tissue-specific than generic housekeeping.

“Context really does matter,” Hsieh said.

Hsieh and Mishra’s work shows that mRNA translation — once considered the passive readout of changes that occur earlier in the process — also plays a role in transforming cancer into more aggressive and lethal forms as “an active engine of malignant identity,” according to a commentary accompanying the published study.

“The work sharpens a larger idea whose time has come,” according to the commentary. “Cancer does not simply hijack transcriptional programs; it hijacks the machinery that decides which RNAs become protein.”

Hsieh said his lab will have more to say about lineage plasticity in upcoming studies. It’s also what Mishra wants to study when she runs a lab of her own.

“This is our first story on lineage plasticity,” Hsieh said. “This is going to be the basis of the lab that she wants to start.”

This work was supported in part by DBG-25-1373505-01-RMC from the American Cancer Society, the National Institutes of Health, a Prostate Cancer Foundation Challenge Award, Seattle Translational Tumor Research, The Nancy & Dick Bernheimer Memorial Fund, The Matthews Family Memorial Fund, The Stinchcomb Family Memorial Fund, The Thomas & Patricia Wright Memorial Fund, the Larry and Virginia Gordon Endowed Chair in Prostate and Bladder Cancer Research, the Fred Hutch Interdisciplinary Training Grant, Genomics, Proteomics, Cellular Imaging, Histopathology, Comparative Medicine and Bioinformatics Shared Resources of Fred Hutch Cancer Center, The Institute for Prostate Cancer Research, Prostate Cancer Foundation Young Investigator Award, a pilot grant from the Mike Slive Foundation for Prostate Cancer Research and the National Science Foundation.