Evasive Action

Just as species evolve through natural selection, individual cancer cells find ways to adapt to treatment and then evade it through an evolutionary process, notes Dr. Dan Landau, associate professor of medicine and a member of the Meyer Cancer Center at Weill Cornell Medicine, whose team is at the vanguard of this new frontier of cancer research. Their strategy: By identifying how individual cells evolve to mutate and proliferate, the scientists aim to develop tools to interrupt the evolutionary process — work key to detecting and treating cancer and avoiding unnecessary treatment.

The New Frontier: Single-Cell Multiomics

Data visualization: Matt Twombly

We are all born with 30 trillion or so cells that evolve throughout our lives. Even when we are healthy, these cells accrue mutations in their DNA that can then grow into “clones” or clusters of genetically distinct cells, Dr. Landau explains. “What is it that allows that single mutated cell, that is not yet cancer, to outcompete and outgrow its neighbors?” he asks. Since malignancy represents a “runaway process of evolution,” he says, it’s crucial to understand the behavior of that single misbehaving cell as an early harbinger of the malignancy process.

Dr. Landau and colleagues at Weill Cornell Medicine and the New York Genome Center are doing just that, using a set of advanced new techniques — called single-cell multi-omics — that allow them to profile thousands of individual cells in unprecedented detail.

In recent work reported in Nature Genetics, the scientists sampled individual cells from patients with glioma brain tumors. In addition to mapping distinct cell behaviors in gliomas, and identifying key programming marks on DNA (called “methylations”) that appear to shift glioma cells from one cellular behavior to another, the team devised techniques to reconstruct an “ancestral tree” for each cell.

“It’s like having a time machine,” says Dr. Landau. “We can take a sample from a patient’s tumor and infer many details of how that tumor has been developing” — going back to the very origin of the tumor.

The lineage trees revealed that glioblastoma cells had a higher degree of plasticity when compared to cells from lower-grade gliomas, Dr. Landau notes. This made it easier for them to switch back and forth between stem cell states and more mature states — potentially allowing the cells to survive stem cell-killing treatments.

Taken together, the findings have offered crucial new insights into the dynamics of gliomas — information that should be helpful, Dr. Landau says, in developing better methods for detecting, staging, monitoring and treating brain cancer. And in principle, he says, “This approach could be used to study the development of many types of cancer.”

Next Steps

Whole genome DNA sequencing, supported by machine learning, represents a “tectonic shift” in cancer research and treatment, Dr. Landau believes. Currently, without knowing whether residual cancer remains after surgery or other treatment, the paradigm is often to prescribe aggressive follow-up treatment that can carry significant side effects — and in most cases isn’t needed. And for some patients with residual cancer, the cells can adapt to treatment and evolve to become more aggressive — without doctors knowing.

“Imagine instead if you could have a simple blood test after surgery and find out yes, you do have residual disease and we can begin treatment and follow up to make sure it is working, or no, you are clear and don’t need aggressive treatment,” says Dr. Landau. His team will continue to develop their technique to be even more sensitive and specific and aim to confirm its effectiveness through clinical trials.