Tumors More Complicated, Harder to Treat Than Expected, Gene Studies Show

Results suggest that cancer is ‘a disease of pathways’

by Aalok Mehta

October 10, 2008

Exhaustive genetic analyses of two of the deadliest types of cancer—including the most common kind of brain tumor—suggest that the disease is far more complicated and varied than previously believed.

The findings predict that current techniques for developing cancer drugs would be largely unsuccessful and that new approaches are needed, experts say. But the results, which appeared Sept. 4 in two papers in Science, also suggest potentially powerful new diagnostic tests for detecting cancers early.

In the studies, researchers looked at nearly every protein-coding gene in tumor samples from 24 people with pancreatic cancer and 22 with glioblastoma multiforme (GBM), the type of brain tumor affecting U.S. Sen. Edward Kennedy. These cancers have only about a 5 percent survival rate after five years.

On average, each tumor contained at least 60 different genetic changes, including dozens of small alterations in DNA sequence as well as several whole-scale rearrangements of the genome. But unexpectedly, the scientists found that most of the specific changes were rare—instead of affecting the same proteins, the cancers seemed to be linked by similar changes in overall genetic function.

“The results of these analyses show that cancer is a disease of pathways, where many different individual genes can be broken at the genetic level but where the genes considered together affect a much smaller number of common underlying mechanisms,” study author Kenneth Kinzler, a professor of oncology at Johns Hopkins Medical School, said at a press conference to announce the results.

In pancreatic cancer, there was "a core set of 12 altered pathways that each individually affected over two-thirds of the tumors analyzed," Kinzler said.

In GBM, several of these same pathways showed changes, as did others that seemed to affect processes specific to the development of brain cancers, study author Victor Velculescu of Johns Hopkins said at the conference.

New ways to treat cancer

The absence of many repeated mutations dims hopes that researchers can repeat the enormous success of targeted therapies such as imatinib (Gleevec), the study authors say. Gleevec, used to treat chronic myelogenous leukemia (CML) and a few other types of cancer, inhibits a specific enzyme known to prevent cell death.

"Our work suggests that most solid tumors—particularly those of the brain and pancreas, but others as well—are nothing like CML," said study author Bert Vogelstein. "It is extremely unlikely that drugs which target single genes, like Gleevec, will be active against a major fraction of solid tumors."

The authors suggest that targeting pathways instead of specific proteins might be more fruitful. One way to do this is functional screening, Vogelstein said. In this method, scientists look for compounds that affect a cultured cell line with a disrupted pathway but not a similar cell line with a normal pathway.

In the meantime, the new studies may lead to better cancer diagnostics. Researchers have found mutated genetic material "just floating in the plasma," even in early stages of cancer, Vogelstein said. Nearly all cancers are curable if caught in their initial stages, he said.

In addition, the researchers did discover at least one significant single gene, at least for brain cancer. Approximately 12 percent of GBM patients in the study had an altered form of IDH1, and these patients seemed to develop the disease earlier and live longer than those with a normal copy. Scientists hope to use this finding to develop better tests and treatments for the cancer type.

"GBMs used to be thought of as one disease. It is now clear that they are two," Velculescu said.

More info on the way

The Science papers are "very, very interesting," particularly for turning up genes more-focused surveys would have missed, says Ronald DePinho, a professor of medicine and genetics at Harvard Medical School and chair of the external steering committee of the Cancer Genome Atlas (TCGA) project.

But he cautions against drawing too many conclusions from the data given the small number of samples. It's too early to confirm whether cancer is truly as complex as the papers seem to indicate or to suggest new approaches for developing drugs, he says.

But that may change very soon, he adds. Interim results from the genome atlas also appeared on Sept. 4 in the journal Nature, offering a complementary analysis of GBM. It's just the "first of many companion papers," says DePinho, who was not an author on the study.

TCGA, a massive program to catalog the genetic mutations involved in cancer, has been controversial, with critics charging that the anticipated results don't justify the project's enormous cost. DePinho, however, says "almost everybody gets converted" by the results to date, along with the fact that they are immediately deposited into a public database.

In the Nature paper, TCGA scientists outline changes in 601 genes from more than 200 GBM patients. The researchers identified mutations in three genes—NF1, ERBB2 and PIK3R1—not previously known to be important in GBM. Pathways regulating growth, cell division, and DNA damage were also disrupted in more than two-thirds of the samples, the scientists found.