Researchers can now identify patients likely to respond to cancer drug therapy

Nov 13 2005

Researchers at UCLA's Jonsson Cancer have identified key characteristics in certain deadly brain tumors that make them 51 times more likely to respond to a specific class of drugs than tumors in which the molecular signature is absent.

The discovery of the telltale molecular signature - the expression of a mutant protein and the presence of a tumor suppressor protein called PTEN - will allow researchers to identify patients who are likely to respond to the drug treatment before they undergo therapies that are not likely to work, said Dr. Paul Mischel, an associate professor of pathology and laboratory medicine and a Jonsson Cancer Center researcher.

Mischel and his colleagues say in an article in the Nov. 10 issue of the New England Journal of Medicine that the discovery could change the way doctors treat glioblastomas, the most common type of malignant brain tumor and one of the those lethal forms of cancer.

"In a biologically aggressive disease like glioblastoma, it's vital to be able to stratify patients up front so we can treat them with drugs that they are more likely to respond to," Mischel said. "This will help prevent patients from having therapies that are much more toxic and less beneficial. With the short survival times associated with glioblastoma, that is critical."

Between 8,000 and 10,000 new cases of glioblastoma will be diagnosed in Americans this year. Average survival is less than a year, according to the American Cancer Society. Although treatment may prolong life, most malignant brain tumors are not curable, making the search for better treatments even more urgent, Mischel said.

A protein called epidermal growth factor receptor (EGFR) is commonly amplified in glioblastoma, making it a prime focus for therapies. Drugs such as Tarceva and Iressa target EGFR, blocking the cell signals that drive amplification of the protein and speed cancer growth. A subset of glioblastoma patients responded to Tarceva and Iressa, but it was not clear what characteristics made them respond to these drugs. There had to be critical molecular factors that determined response, Mischel said.

He and his team set out to find the molecular determinants that indicated which patients would respond best to EGFR blockers. Previous UCLA research in brain and other cancers suggested that the key might be the interaction of the PTEN protein and a mutant protein called EGFRvIII. About half of patients with amplified EGFR also have this mutant protein.

The UCLA team and their collaborators studied a subset of 26 glioblastoma patients who either responded very well or very poorly to EGFR blocking drugs and developed a way to test their brain tumor tissue for the presence of both the mutant and PTEN proteins. Mischel's team found that patients with both genetic variations were 51 times more likely to respond to EGFR blockers. They also lived five times longer after initiating therapy than those without the variation, surviving 253 days versus 50 days.

To confirm their promising work, Mischel and his team obtained tissue samples from 33 brain cancer patients treated at another facility without knowing who the responders were. They were able to independently replicate their results, confirming that those with both genetic variations were more likely to respond to EGFR blocking drugs.

The study shows that glioblastoma patients can respond to targeted agents, and suggests that patients likely to benefit from treatment can be identified by molecular testing. The study also raised the possibility that patients whose tumors lack the genetic variations in the molecular signature could possibly be treated with drugs to make them more sensitive to EGFR blockers.

Of the 8,000 to 10,000 glioblastoma patients diagnosed each year, about 10 to 20 percent have the combination of the mutant and PTEN proteins, Mischel said. The next step is a prospective study, determining the molecular signature of patients' tumors and directing those with the right protein combination to EGFR blocking therapies. Mischel's team also is working to uncover the molecular signatures in the tumors of non-responders so they can determine what therapies might be most effective for them.

"This is a much more hopeful period now in cancer research," Mischel said. "Genomic and proteomic technologies are helping us begin to understand the underlying molecular features of disease, and new drugs are making it possible to safely and specifically target pathways that are altered in cancer cells. This was impossible five years ago. Glioblastoma is still a difficult disease, but the idea that it may be possible to induce long-term disease suppression gives us reason for hope."

The study, Mischel said, also may have important implications in other cancers.

"Many cancers have a similar combination of a mutant cancer-causing protein and either the expression or loss of the PTEN protein," Mischel said. "The interactions of the two may be important in determining response to targeted agents."