Challenging the notion of personalized medicine

By Dr. Sabrina Richards / Fred Hutch News Service

New brain cancer findings by scientists at Fred Hutchinson Cancer Research Center fly in the face of personalized medicine, which tailors treatment to a tumor’s unique genetic or molecular profile. Instead, the study suggests that effective brain cancer therapies should target the similarities between tumors rather than their differences.

Published online in Cancer Cell, lead author Dr. Eric Holland and colleagues suggest that several subtypes of a brain cancer called glioblastoma that previously were thought to sprout independently from one another are actually branches of the same developmental tree. The data show that the same genomic changes consistently initiate the transition of healthy brain cells toward various types of glioblastoma.

"These tumors really are different ‘flavors’ of the same thing,” said Holland, who directs the Hutch’s Human Biology Division and Solid Tumor Translational Research group. He also directs the University of Washington’s Alvord Brain Tumor Center.

The results also suggest that subtype-specific genetic mutations that occur late in the cancer-development process are not good drug targets.

Almost half of brain tumors are glioblastomas, which arise from the glial cells that produce the myelin that sheathes our neurons. It’s a deadly disease. “Outcomes are particularly bad,” noted Holland — and most therapies designed to hone in on specific mutations in the defined particular subsets have not been successful.

Glioblastomas come in many varieties but can be broadly separated into two categories based on a certain type of DNA modification known as methylation. Holland and his team focused on the broad category with lower levels of methylation. Such glioblastomas can be categorized into four subtypes based on characteristic sets of molecular changes: proneural, mesenchymal, classical and neural.

‘It was a little heretical’

The fact that specific mutations are associated with particular tumor types suggested to many researchers that each type arose independently and then followed its own developmental path. And even though clinical trials of drugs designed to combat specific mutations failed to show dramatic results, “It was a little heretical then to say that no, these subtypes are basically the same thing,” Holland said. “The characteristic mutations are late events in the evolutionary trajectory.”

To chart the sequences of cancer-causing changes underlying each tumor type, Holland and his team drew on publicly available genetic and molecular tumor data housed in The Cancer Genome Atlas — and discovered the same two major genomic alterations in virtually all glioblastomas. In one fell cancerous swoop, glial cells lost one copy of chromosome 10 and simultaneously gained several copies of chromosome 7. According to the group’s analysis, these early cancer cells looked very similar to proneural tumor cells, and only later did specific mutations transform them into other subtypes.

The researchers also used a mouse model of proneural tumors and found they could transform them into the mesenchymal subtype by adding a characteristic mutation, which supported the hypothesis that other brain tumor subtypes can evolve from proneural-like tumors.

A novel finding

It’s unusual that the first noteworthy step in the march toward cancer occurs via a dramatic change in chromosome number instead of a small localized mutation, Holland noted. “That’s a novel finding. It’s not what you usually think of in cancer.” The discovery sheds new light on the biology of glioblastomas and their early resistance to therapy. “A whole collection of things have gone wrong at the beginning,” he said.

The results also have implications for researchers trying to design brain cancer therapies. The loss and gain of whole chromosomes affect many genes, making it difficult to determine which to target. Though Holland and his collaborators identified a few key players that drive the transformation of glial cells into cancerous proneural cells, these genes are not alone in affecting the cancer’s aggressiveness and response to therapy. Rather than tailoring treatments to the unique mutations that define each subtype, Holland instead advocates for a more comprehensive approach.

“The cells were cancerous before those late mutations occurred. Targeting them is unlikely to help,” he said. Such a strategy would be akin to tackling weeds by snipping at their leaves rather than wrenching up their roots. In this vein, Holland’s team is “marching along” chromosome 7 to see if any of its genes affect the resistance of glial cells to radiation, one of the few generally effective treatments against glioblastoma.

Potentially, targeting these genes could make cells from any glioblastoma subtype sensitive to radiation. The best strategy to treat brain cancer would be to optimize a therapy that works against all subtypes. “It’s common sense, and you’re not shooting in the dark,” he said.

The National Institutes of Health funded the research, which also involved investigators at Memorial Sloan Kettering Cancer Center, Dana-Farber Cancer Institute, Harvard School of Public Health and the Spanish National Research Centre.

Author: Dr. Sabrina Richards is a staff writer at Fred Hutchinson Cancer Research Center. She has written about scientific research and the environment for The Scientist and OnEarth Magazine. She has a Ph.D. in immunology from the University of Washington, an M.A. in journalism and an advanced certificate from the Science, Health and Environmental Reporting Program at New York University. Reach her at srichar2@fredhutch.org.

Photo: Dr. Eric Holland, director of the Human Biology Division at Fred Hutch, is the lead author of the Cancer Cell paper that found subtypes of a brain cancer called glioblastoma are actually branches of the same developmental tree. Photo by Robert Hood / Fred Hutch News Service