Researchers have shown that glioblastomas can develop resistance to a drug by rewiring their signaling networks to counteract its effects. Combining drugs that target two network changes produced remissions in mice implanted with human glioblastomas.

Tumors can develop drug resistance through a number of mechanisms, including the acquisition of mutations, natural selection of existing resistant cell populations, and adaptively rewiring their signaling networks. Analyzing changes to these networks may help researchers identify drug combinations to overcome resistance, according to a recent Cancer Cellstudy.

Although genome sequencing of glioblastomas has identified mutated genes in several pathways, including the receptor tyrosine kinase, p53, and RB pathways, clinical trials of targeted therapies have not extended patients' lives. In part, these disappointing results are likely due to the rapid development of resistance, which may have occurred because cells in the tumors already had mutations that allowed them to survive therapy, referred to as clonal selection. However, the tumors might also have relied on an adaptive, nongenetic mechanism that involves rewiring signaling networks to circumvent the drugs' effects.

To determine whether glioblastomas take advantage of this alternative resistance mechanism, a team led by Paul Mischel, MD, of the Ludwig Institute for Cancer Research San Diego and the University of California, San Diego, and James Heath, PhD, of the California Institute of Technology in Pasadena, used a patient-derived xenograft mouse model of glioblastoma. The researchers treated tumor-bearing animals with a compound that inhibits the mTOR pathway, which is overactive in almost all glioblastomas. Tumor growth initially slowed, but resistance quickly set in. Although the tumors had accrued some mutations, they didn't occur in genes that enable cells to overcome mTOR inhibitors, suggesting that clonal selection didn't cause resistance.

Next, the scientists used single-cell proteomic analysis to measure the abundance and phosphorylation status—an indicator of activation—of key proteins, including AKT, Src, and ERK. They also determined which protein levels changed in concert, a sign that the proteins are working together. Their results suggest that the tumors activated compensatory signaling pathways in response to mTOR inhibition, including the Src and ERK pathways.

“We are not saying that the clonal selection model is wrong, but there's another type of resistance that can develop quite quickly,” says Mischel. In fact, cells had rewired their signaling networks within 3 days of treatment, long before tumors showed any outward signs of resistance.

Attempting to prevent resistance, the researchers treated the mice with the mTOR inhibitor and either the Src inhibitor dasatinib (Sprycel; Bristol-Myers Squibb) or U0126, which blocks the ERK pathway, or both. Tumor growth halted in the animals that received either or both of the additional compounds. Growth resumed, however, when the researchers stopped treatment.

“We aren't the first to see [nongenetic] resistance, but the rapid response that we uncovered, coupled with our signaling analysis, may permit timely clinical interventions,” says Heath. The researchers are analyzing glioblastoma tissue from additional patients to find other drugs that might prevent resistance, and they say changes in protein signaling networks could also suggest treatments for other cancer types.

“It's exciting work,” says Cameron Brennan, MD, of the Memorial Sloan Kettering Cancer Center in New York, NY, who wasn't connected to the research. “It opens the door for analyzing previously hidden mechanisms of resistance and potential drug combinations” for patients with glioblastoma, he says. –Mitch Leslie