To capture a previously poorly understood but clinically significant mechanism of cancer drug resistance, a team of researchers, led by Ludwig Cancer Research scientist Paul Mischel and James Heath of the California Institute of Technology, has probed biochemical signaling cascades within individual cancer cells.
Published in the current issue of Cancer Cell, their paper shows that cells of the invariably lethal brain cancer glioblastoma multiforme (GBM) begin adapting to resist therapy within as little as three days of its initiation. The researchers further show that their technology and analytical approach has the potential to be harnessed by clinicians to anticipate, and even overcome, drug resistance to GBM.
‘Cancers in general, and glioblastoma multiforme (GBM) in particular, are marked by staggering cellular and molecular heterogeneity.’
"Cancers in general, and GBM in particular, are marked by staggering cellular and molecular heterogeneity," says Mischel, a member of the Ludwig Institute for Cancer Research, San Diego. "It is what makes them so difficult to treat. We risk missing important elements of this heterogeneity if we can't collect and analyze data at the level of individual cells."
Drug resistance is a case in point. Many targeted therapies bind and inactivate specific proteins known to be essential nodes in the signaling circuits that drive cancer cell growth. But, almost invariably, some cells within tumors randomly develop mutations that either disrupt the binding of a drug or otherwise obviate its effects. Those cells tend to resist treatment and can grow to dominate a tumor. This classical mechanism of drug resistance is known as clonal selection.
But tumors can also adapt to targeted therapies in other ways. Their cells might, for example, skip the genetic changes and alter the coordination of their internal signaling circuitry-in effect sending critical growth signals down alternative circuits. To capture such changes, the researchers used a micromechanical device developed by Heath's lab known as the Single Cell Barcode Chip (SCBC) to study GBM cells taken from a patient-derived mouse model of the cancer. The mice were treated with drugs that target key molecular nodes in GBM signaling circuits, mTORC1 and mTORC2.
Mischel, Heath and their colleagues first eliminated the possibility that clonal selection accounted for the inevitable resistance that developed against the drugs. They then used SCBC to track a selected subset of biochemical reactions that expose the activation of distinct signaling pathways in individual or small groups of cells.
"Drug resistance was marked by the emergence of signaling networks that were previously in the background and that now were newly coordinated to drive the growth of these tumors," says Mischel, who is also a professor of Pathology at the University of California, San Diego. "This was a key insight because it led to a series of testable predictions."
Specifically, after pinning down the precise alternative circuits used by GBM cells under treatment, they showed that by concurrently targeting both the original and the new signaling pathways, they could durably suppress the growth of tumors in their mouse model. Conversely, targeting any one pathway alone, even with more than one drug, had no such effect. The findings were replicated in tumor samples obtained from GBM patients at various stages of treatment.
"Our findings do not overturn the idea of clonal selection," says Beatrice Gini, a former post-doctoral researcher in Mischel's lab who conducted much of the study with Wei Wei, an assistant professor of Molecular and Medical Pharmacology at UCLA and former graduate student in Heath's laboratory. "In fact, genetically encoded resistance and this kind of adaptive resistance are probably occurring at the same time in a GBM tumor. But the rewiring of signaling circuits occurs much faster-apparently in the first few days of treatment, even when the tumor seems to be responding to a targeted therapy."
The findings also explain why that response has typically proved so minor and fleeting in GBM, an extremely heterogeneous tumor. "This study builds on ten years of collaborative work with Jim Heath's lab and really highlights how heterogeneity is linked to a cancer's clinical behavior," says Mischel. "The cancer cell has a vast repertoire of responses at its disposal when it is challenged by a drug."
Support for this study was provided by Ludwig Cancer Research, Ben and Catherine Ivy Foundation Fund, the National Institutes of Health, the Phelps Family Foundation and the National Brain Tumor Society.