Scientists studying a hard-to-treat form of blood cancer called acute myeloid leukemia (AML) have found that a type of treatment — immunotherapy — may help change the environment where cancer cells live, possibly helping the immune system respond more effectively.

In a new study published in July in Science Advances, a team of researchers, including scientists with Virginia Tech’s Fralin Biomedical Research Institute Cancer Research Center in Washington, D.C., examined bone marrow samples from adult patients with relapsed or refractory AML, a serious and often aggressive form of the disease that is difficult to treat and associated with poor outcomes.

In these patients, the cancer had either returned or failed to respond to earlier treatment.

The subjects in the study were treated with two drugs: pembrolizumab, which helps the immune system attack cancer cells, and decitabine, which affects how certain genes are switched on or off.

While the treatment didn’t work for everyone, some patients showed signs that immune cells were mobilizing in the bone marrow — and researchers wanted to understand why.

To explore this, a large team of scientists from multiple institutions used high-powered tools to examine the patients’ bone marrow, including an analytical technique called single-cell spatial transcriptomics to understand where and how genes were active in the bone marrow.

This method, combined with advanced computer analysis, can examine individual cells in a biopsy sample and identify which RNA molecules are present in each cell, while keeping track of exactly where each cell is located.

This gave researchers a much clearer picture of how the immune system was responding to treatment and how it was interacting with leukemia cells. With this approach, the team found that certain immune cells moved closer to leukemia cells after treatment for some patients.

This change in cellular neighborhoods could reflect an immune system trying to fight back. The researchers also noticed changes in how cells were communicating — possibly a clue about how the treatment affects cancer’s ability to hide from the immune system.

“Our findings show how immunotherapy may shift the types of cells found in the neighborhood around leukemia cells,” said Gege Gui, the study’s first author and a research scientist with the Cancer Research Center in Washington, D.C., who was also a doctoral student with the Johns Hopkins University when the research was conducted.

“That gives us clues about how the immune system and cancer interact and how we might help patients by advancing our understanding of underlying biological mechanisms.”

Christopher Hourigan, director of the Cancer Research Center in Washington, D.C., and one of the senior authors of this work, said this kind of detailed, cell-by-cell analysis can reveal patterns that aren’t visible through traditional methods.

“I am impressed by the potential of the careful work Dr. Gui has done integrating powerful computational approaches with these novel genomic tools,” Hourigan said. “Too often, cancer therapy doesn’t work as well as we would like for patients with AML, but research like this is getting us to a stage where we can start understanding why that may be so that we can hopefully design better treatments in the future.”

Hourigan, professor at the Fralin Biomedical Research Institute and the Virginia Tech Carilion School of Medicine, is an oncologist and physician-scientist who focuses on research in translational medicine and precision oncology. Laura Dillon, research associate professor at the Fralin Biomedical Research Institute Cancer Research Center in Washington, D.C., also contributed to this work.

The study was a collaborative effort across several major research centers.

Kasper Hansen of Johns Hopkins University, contributed expertise in statistical genomics and computational analysis of high-throughput genomic data; Chen Zhao, from the National Cancer Institute of the National Institutes of Health, provided insights into tumor immunology and advanced tissue imaging techniques, including spatial transcriptomics.

Original study: DOI 10.1126/sciadv.adw4871