Overlooked stretches of DNA could hide cancer-causing mutations

Our genes encode our proteins, and genetic mutations that change the amino acid sequence of our proteins can have major health consequences, including cancer. But other areas of your DNA, once dismissed as “junk,” can play critical roles in switching a gene on and off — and play a role in tumor development and treatment response.

Now, scientists at Fred Hutchinson Cancer Research Center show that the often-overlooked regions that flank our genes (called non-coding regions) harbor mutations that can disrupt normal protein production and influence tumor spread and response to treatment. In work published today in the journal Nature Communications, the team report these findings and describe a high-throughput method they developed that can determine whether cancer-associated mutations in this region have functional consequences.

Though the team first described the effects of these mutations in prostate tumors, “these findings aren’t limited to just prostate cancer,” said Dr. Andrew Hsieh, the physician-scientist and prostate cancer expert who led the study. “There's a language here that is applicable to all cancers that have mutations [in these regions].”

The new work reveals an unappreciated aspect of cancer biology and could become the focus of future drug discovery efforts, he said.

Overlooked, understudied regions

Our genes may get the most attention, but these protein-coding regions make up a minority of the DNA in our cells. Research over the past few decades has revealed that a lot of important instruction that regulate how genes are turned into proteins lurk in these regions.

Hsieh and Dr. Yiting Lim, a postdoctoral scientist on his team, wanted to understand how mutations in these regions could influence cancer development. They focused on one type of non-coding DNA, a region that sits adjacent to every gene and helps two steps of the multistep protein-creation process. In the first step, a process known as transcription, cells create temporary copies of genes in order to ferry information within DNA to the cell’s protein-building factories elsewhere. The second step — the process of turning these gene copies, called messenger or mRNA, into strings of amino acids — is called translation.

mRNA includes copies of non-protein coding regions, which will never be translated into protein, but are critical for that process. The region the scientists focused on is called the 5-prime untranslated region, or 5’ UTR. 5’ UTRs carry information that can regulate protein production at two different steps of the process.

“The 5’ UTR interfaces between translation and transcription — it’s like a bridge that can connect these two processes together,” Hsieh said.

A gene’s 5’ UTR holds genetic instructions that help regulate when and how strongly it gets turned on, or transcribed. When it’s turned into part of an mRNA, the 5’ UTR is where the protein-building machinery docks for translation.

Hsieh’s previous work on prostate and bladder cancer had demonstrated that dysfunction in the protein-building process can promote cancer. More comprehensive DNA-sequencing studies, which included both coding and non-coding regions of DNA, had demonstrated that cancer-associated mutations can occur outside genes as well.

“But those studies were done one by one,” looking at each 5’ UTR mutation on its own, said Lim, who spearheaded the work in Hsieh’s lab before joining a Seattle-area biotech company as a senior scientist. No one had done a comprehensive survey of 5’ UTRs in any cancer type to see if they harbored any cancer-promoting mutations, she said.

PLUMAGE: A new technique for a new question

Lim and Hsieh started out with ambitious goals. First off, they wanted to get a big-picture view of the role that mutations in the 5’ UTR may play in cancer. This would require using a high-throughput method that could survey the effects of 5’ UTR mutations from hundreds of genes at once.

For every mutation they identified, Hsieh and Lim also wanted to be able to get a sense of how it might (or might not) affect its gene’s transcript levels and translation — that is, whether the mutation has functional consequences. And, they wanted to be able to view each mutation in the context of its full 5’ UTR, which can range in length from just 18 to over 3,000 genetic “letters.” This would require sequencing longer stretches of mRNA than most high-throughput methods can handle.

So Lim set about developing her own method, which she and Hsieh dubbed PLUMAGE, for pooled full-length UTR multiplex assay on gene expression. PLUMAGE enables her to simultaneously survey hundreds of 5’ UTR mutations and test their effect on both transcription and translation at the same time.

Many mutations in the 5’ UTR have functional consequences

Lim and Hsieh started their survey by looking at prostate tumors. Not only is Hsieh an expert, but Fred Hutch and the University of Washington, where he sees patients, share a large biobank of prostate cancer tissue samples generously donated by patients. This tissue bank represents everything from early-stage, localized disease to late-stage tumors that resist treatment and have spread through the body.

Lim began by examining genetic sequencing data from 149 patients with localized prostate cancer and 80 patients with late-stage metastatic disease. She found 2,200 differences in the genetic code of 5’ UTRs in 1,878 different genes. About 500 cropped up more than once, often in the UTRs of known cancer-associated genes. Just over half the genes Lim studied had mutations only in their 5’ UTRs, but not in their coding regions. She also saw that genes in late-stage tumors averaged more mutations in their UTRs than those in early-stage tumors.

Strikingly, Lim saw that mutations in these two regions often affected genes involved in different cellular processes. 5’ UTR mutations often arose in genes involved in the cell’s cycle of growth and division, including a group of known cancer-promoting genes grouped together in what’s known as the MAP kinase pathway.

The factors that transcribe DNA into RNA land on the 5’ UTR, and Lim saw that more than 300 5’ UTR mutations were of types that could transform a gene’s regulation and connect it to previously unrelated cellular processes. In the case of a gene for a cellular growth factor, a mutation in its UTR created a new binding site for a well-known cancer driver called Myc.

“It’s an example where a 5’ UTR mutation put a gene under the control of an oncogenic [cancer-driving] protein that it wasn’t previously controlled by,” Lim said.

This means that 5’ UTR mutations can create new ways to control cancer genes and ultimately affect tumor-cell growth.

Mutations in the 5’ UTR also altered sites where the translation machinery docks, potentially changing the rate at which a gene is turned into protein once its mRNA is produced. All told, Lim found that just over a third of the recurrent 5’ UTR mutations affected gene transcript levels, translation, or both.

Notably, the mutations that affected both didn’t always affect them in the same way — some turned transcription down while toggling up translation, or vice versa. This shows that researchers can’t make assumptions about how changes in one process affect the other, Hsieh said.

“Instead, you need to go all the way to the end [of the protein production process],” he said. “It’s a cautionary lesson for how we look at gene expression. We need to keep in mind that there are multiple players we need to consider together.”

5’ UTR mutations can influence patient outcomes

Lim did more than look at protein levels in her samples. She also looked at whether any 5’ UTR mutations correlated with clinical outcomes, such as response to treatment or disease stage.

She found such a link in the UTR mutations that affect genes in the growth-promoting MAP kinase pathway. Patients whose tumors had UTR mutations that affected transcript levels or translation of these genes were more likely to have aggressive disease and twice as likely to have metastases, or areas of cancer spread — but also more likely to benefit from chemotherapy for longer than patients whose tumors did not have these mutations.

“It shows really for the first time that these [5’ UTR mutations] are not just bystanders, and they're not just mutations that are ‘silent,’” Hsieh said. “We know they’re not all going to count, but about a third of them do. … I do think that this technological breakthrough, which allows us to have these molecular findings, as well as patient-based findings. It really opened up a new way of looking at cancer.”

A potential new horizon for drug development

Hsieh and Lim are excited for the potential of PLUMAGE to reveal new processes and mechanisms that may promote cancer or change its behavior. Especially as tumor sequencing becomes more common and more complete, oncologists and researchers need technologies that shed light on a mutation’s consequences. PLUMAGE can do that, they said.

“Since we’re sequencing everything, now the big question is, ‘What's the function of all these mutations?’” Hsieh said. “And that's why we call our technology a functional genomics technology.”

mRNA doesn’t float around the cell as a tidy string — instead it folds back on itself in complex structures, which can be altered by mutations.

The relationship between a 5’ UTR’s structure and its function, and how mutations may alter both, is first on Hsieh’s list of important questions to explore next. He also wants to study whether 5’ UTR mutations may be vulnerabilities that could be the targets of future drug development. Though the idea is in its infancy, some biotech companies are examining the possibility of creating drugs that target new mRNA structures created by mutations, Hsieh said.

In the meantime, he and Lim hope that their findings inspire other researchers to begin studying how gene transcription and mRNA translation can interact to promote disease.

“I think this is a new area of cancer biology to get excited about,” Hsieh said. “And I do think this paper opens new doors for cancer biologists.”

This research was supported by the National Institutes of Health, a U.S. Department of Defense Congressionally Directed Medical Research Program, a Prostate Cancer Foundation Challenge Award, a Burroughs Wellcome Fund Career Award for Medical Scientists, the Robert J. Kleberg Jr. and Helen C. Kleberg Foundation and an American Association for Cancer Research-Bristol-Myer Squibb Oncology Fellowship.