Breaking the RAS Ceiling
After more than 30 years of research, there is still no effective RAS inhibitor, but OSUCCC – James researchers may help change that
BY DARRELL E. WARD
In 2014, Ohio State cancer researchers were deep into a study investigating a question related to microRNAs in melanoma when they made an astonishing discovery. Analysis of their data revealed that an infamous oncogene called NRAS encodes five proteins, not just one as was thought.
“RAS oncogenes are of immense importance in many cancers,” says principal investigator Albert de la Chapelle, MD, PhD, Distinguished University Professor in the Molecular Biology and Cancer Genetics Program at The Ohio State University Comprehensive Cancer Center – Arthur G. James Cancer Hospital and Richard J. Solove Research Institute (OSUCCC – James). “Virtually thousands of labs are working on various aspects of RAS, so our findings could have important implications.”
The three-member family of RAS genes—KRAS, HRAS and NRAS—has been notoriously difficult to understand and to target. After more than 30 years of research, there is still no effective RAS inhibitor, earning these proteins a reputation as “undruggable.” Agents that are approved for RAS-mutant tumors target downstream molecules, but these inhibitors have been largely ineffective, especially as single agents.
“RAS mutations drive some of the most deadly types of cancer, and they do so by activating many signaling pathways that facilitate cancer-cell growth and progression,” says William Carson III, MD, professor of Surgery and associate director for clinical research at the OSUCCC – James. “Targeting one downstream molecule isn’t enough to block the action of this cancer-causing gene because cancer cells often develop alternate pathways and escape the block. Inhibiting RAS directly would be much more effective in halting cancers driven by these mutations.”
In fact, nearly one-third of cancers contain a RAS mutation, but the gene that is mutated varies with cancer type. For example, RAS mutations are present in:
- 95 percent of pancreatic cancers (KRAS)
- 45 percent of colorectal cancers (KRAS)
- 35 percent of lung cancers (KRAS)
- 15 percent of AML (NRAS)
- 15 percent of melanoma (NRAS)
- 10 percent of bladder cancers (HRAS)
In some cases, RAS mutations make cells resistant to therapy, and their presence in tumor cells can influence treatment decisions. OSUCCC – James researchers are studying NRAS and KRAS in particular to better understand and treat these cancers.
In healthy cells, RAS proteins help regulate the transduction of signals triggered by hormones, cytokines and growth factors. All three proteins have GTPase activity and are involved in transmitting signals involved in cell growth, proliferation and migration.
Normally, RAS proteins receive signals that cause them to switch between an active state (bound to guanine triphosphate, or GTP) and an inactive state (bound to guanine diphosphate, or GDP). In cancer cells, mutated RAS is locked into its active state and continuously promotes proliferation, survival and invasion.
“All three RAS proteins are widely expressed throughout the body, so it is unclear why certain RAS genes are preferentially mutated in each cancer type,” says OSUCCC – James researcher Christin E. Burd, PhD, assistant professor of Molecular Genetics and of Cancer Biology and Genetics at Ohio State. “Certain lung-cancer-associated KRAS mutations have been attributed to the carcinogens in cigarette smoke, but the origins of other mutations are completely unknown.”
RAS was one of the earliest oncogenes to be discovered, and its family closet bulges with oddities and surprises. For example:
- The three genes are located on three different chromosomes;
- All three RAS proteins share over 80 percent of their amino acid sequence;
- Mutations in all three genes occur predominantly in only three positions in the protein: codons 12, 13 and 61.
ANOTHER RAS SURPRISE
The research that led to the discovery of the four new NRAS isomers began with a study led by de la Chapelle, Carson, Ann-Kathrin Eisfeld, MD, who is part of Ohio State’s Medical Scientist Training Program, and OSUCCC – James researcher Clara D. Bloomfield, MD, a Distinguished University Professor at Ohio State who also serves as cancer scholar and senior adviser to the OSUCCC – James.
“We discovered the four NRAS isoforms ‘by accident’ when looking at the impact of NRAS mutations and differential NRAS expression in AML patients,” Eisfeld says. “We were surprised that the existence of four isoforms of NRAS had gone unnoticed by researchers for 30 years, even though the RAS genes are so central to oncogenic signaling and so intensely studied worldwide.”
They reported the 2014 findings in the journal Proceedings of the National Academy of Sciences. The researchers found that the NRAS gene can produce five naturally occurring NRAS proteins, not just the one known protein, now called isoform 1. Isoforms 2 to 5 are produced through alternative splicing of the NRAS messenger RNA (see box, Alternative Splicing).
The five NRAS proteins differ in size, in expression levels in different human malignancies and in their intracellular location and effects (see box, Characteristics of the Five NRAS Splice-Variant Proteins). Notably, isoform 5 consists of just 20 amino acids and triggers highly aggressive behavior in melanoma cells. The researchers published its structure in the journal Protein Science. “We are hopeful that our discovery will go beyond being a surprising biological observation and have implications for overcoming the ‘undruggable’ attribute of NRAS,” Eisfeld adds.
The researchers’ efforts to elucidate the functions of each isoform are also yielding insights into isoform 2.
Megan Duggan, PhD, a graduate fellow in Carson's lab, examined BRAF-mutated melanoma cells exposed to vemurafenib, a drug approved for the treatment of BRAF-mutated melanoma.
“Patients using the drug have fantastic responses, but then a few weeks later the disease returns in full,” Duggan says.
She found that after exposing BRAF-mutant melanoma cells to the drug, expression of isoform 2 shoots up, when formerly it had been undetectable. “We hypothesize that this is an escape mechanism for the cells, and it may explain why the disease recurs after a great response to the anti-BRAF agent.
“When we knocked down isoform 2 in a mouse model, the animals became responsive again to vemurafenib. These results in the animal model were very encouraging,” she says.
UNLOCKING RAS MUTATIONS
Christin Burd wants to learn why particular RAS mutants are common in some cancers and not others. For example, why NRAS is by far the most frequently mutated RAS gene in melanoma, and why these mutations seem to always appear in location 61. This is contrary to other cancer types like acute myeloid leukemia, where NRAS is also selectively mutated, but typically at codon 12 or 13.
“There appears to be evolutionary selection for particular RAS mutations in each cancer type,” Burd says. “The question is, can we better understand why these specific mutations evolve in each tumor type such that drugs can be made to target only their cancer-inducing functions?”
To begin her work, she is focusing on melanoma, a cancer type for which more than 10 new therapeutic regimens have been FDA-approved since 2011. Unfortunately, not one of these drugs directly targets mutant NRAS, the second most common genetic alteration in melanoma, after BRAF mutations, Burd explains.
“We have drugs that directly target BRAF-mutant tumors, but for melanoma patients with an NRAS-mutant tumor, therapeutic options are limited.“A person can do most everything right and still be diagnosed with highly metastatic disease,” she says. “We know that sunlight contributes to melanoma, but we don’t know the fundamental cause of the disease, which makes it very hard to prevent.”
The idea that evolutionary selection might be the key to identifying the Achilles’ heel of NRAS-melanoma began with another surprising Ohio State observation. Burd and her colleagues designed a mouse model in which they can switch on an NRAS gene mutation in just the melanocytes of mice. They generated one mouse model containing an inducible NRAS mutation at codon 12 and another with a mutation at codon 61.
“We were shocked to learn that, of the mice with an inducible codon-12 NRAS mutation, none developed melanoma, but when we activated the codon 61 mutation, 80 percent of the mice spontaneously developed tumors,” Burd says. This experiment, published in the journal Cancer Discovery, was the first to suggest that the cancer-causing potential of NRAS codons 12, 13 and 61 mutants were different, and it led to an “Aha!” moment for Burd, who reasoned that the findings might offer a way to discover vulnerabilities specific to each NRAS-mutant cancer type. She hypothesized that forcing an NRAS codon-61 mutant, which causes melanoma, to behave like a codon 12 mutant, which doesn’t cause melanoma, could prevent cancer formation or shut down tumor growth.
“This approach was drastically different from ongoing tactics in the field,” she says. While others aimed to completely shut down mutant NRAS activity, Burd hypothesized that inhibiting only a small subset of NRAS functions would be sufficient to stop melanoma in its tracks.
“The idea of targeting mutation-specific pathways to stop cancer growth is completely new,” Burd says. “Furthermore, we believe that the critical NRAS-driven pathways in melanoma may be different from those in AML. Thus, the drug for a patient with NRAS-mutant melanoma may be completely different than that needed to treat NRAS-mutant AML.”
To identify these critical pathways, Burd proposed her novel idea to the Damon Runyon Cancer Research Foundation and was selected as the first ever Ohioan to receive the foundation’s prestigious Innovator Award. Using these funds, she is developing nine mouse models containing distinct, inducible NRAS mutations localized to codons 12, 13and 61. She and her team will then examine and compare the properties of these mutants, determining why each can or cannot drive melanoma.
Furthermore, because these mutations can be activated in any tissue, she plans to share these mice with colleagues around the world aiming to understand the role of RAS in other tumor types.
“It’s incredibly exciting to have so many possibilities. These models will help us better understand RAS as a cancer driver not only in melanoma, but in the brain, the bone marrow, the colon and many other tissues,” Burd says. “RAS is such a big problem that we can’t expect to tackle it on our own, and I am extremely hopeful for what will come from sharing these models with the scientific community. After all, our ultimate goal is not just to end melanoma, but to provide answers for all RAS-mutant cancer types.”
Terence M. Williams, MD, PhD, associate professor of Radiation Oncology, and his laboratory team are working to understand how mutations in KRAS cause resistance to radiation and other DNA-damaging agents.
About half of patients with cancer undergo radiation therapy during treatment, and about 95 percent of pancreatic cancer patients, 45 percent of colorectal cancer patients and 25 percent of lung cancer patients have an active KRAS mutation.
“Evidence from laboratory studies and emerging data in humans suggest that KRAS mutations foster radioresistance in these tumors, but the mechanism is not well understood,” Williams says. “Our work suggests that it happens because KRAS mutations heighten the repair of DNA double-strand breaks by a mechanism called non-homologous end-joining.”
The researchers presented their initial findings at the 2016 American Society for Radiation Oncology meeting. More specifically, their work implicates a well-characterized protein involved in double-strand break DNA repair in KRAS-mutant cells called 53BP1.
“Cells hit with radiation sustain DNA damage and then go into cell-cycle arrest,” Williams explains. “During that period, DNA repair occurs, and mutated KRAS seems to hijack the 53BP1 pathway to accelerate that process. After proper repair, tumor cells can resume cell division and growth, enabling them to survive therapy.”
KRAS mutations also activate the MEK kinase signaling pathway to facilitate DNA repair, Williams says. “Inhibiting MEK kinase has been shown to help make radiation work effectively, especially in KRAS-mutated cancer cells.” Overall, their findings suggest how KRAS-mutated cells rescue themselves from radiation-induced DNA damage and provide insight into potential pathways to target downstream of KRAS.
WHY NOT TARGET KRAS ITSELF?
The KRAS protein is so difficult to target because of its small size and because the mutant form binds so strongly to GTP, Williams says. “No one has succeeded yet in finding a molecule that kicks the GTP out of the binding pocket, which would likely make an effective inhibitor.”
The preclinical data from Williams’ lab suggesting that MEK inhibition would block the DNA pathways that sustain KRAS-mutant cancers after radiation helped lead to a phase I clinical trial (ClinicalTrials.gov NCT01740648) evaluating the combination of preoperative 5-fluorouracil (5-FU), radiation and the MEK inhibitor trametinib in patients with locally advanced rectal cancer.
“That trial is nearly done, and then we will apply to continue the trial in a larger multi-institutional phase II trial,” he says. “Our results so far show that the combination is safe and tolerable, and that there are initial signs of activity.”
WILL RAS REMAIN UNDRUGGABLE?
“All these areas of RAS research are building upon one another,” Burd says. “And that, along with new data emerging in the field, is changing how we think about targeting RAS. From understanding how expression of the newly discovered isoforms changes during therapy, to defining cancer-specific dependencies and testing drugs to overcome RAS-induced radiation resistance, investigators at the OSUCCC – James are helping to move the field forward.
“It will take creativity, hard work and collaboration,” Burd adds, “but I am confident that the field will ultimately devise novel strategies that target these proteins and improve outcomes for patients with RAS-mutant tumors.”