How the bone marrow microenvironment responds to immunotherapy

A new study has given leukemia researchers a powerful new dataset and collection of analytical methods—and a starting point for future research.

One common treatment for leukemia is hematopoietic stem cell transplant, or HSCT. This transplant replaces bone marrow lost to disease or chemotherapy, and it can induce a powerful immunologic response that combats leukemia cells. But some leukemia cells can remain after the transplant, and in that case, the next option is for a donor to provide lymphocytes, a process called donor lymphocyte infusion (DLI). While the percentage of people who positively respond positively is relatively high for chronic myeloid leukemia (CML) at about 70–80%, the response rate for acute myeloid leukemia (AML) is much lower at about 15–20%.

Now, researchers from Columbia University and the Dana-Farber Cancer Institute may have found a reason why some people with AML respond better to DLI. Their results were published in Science Immunology Jan. 24 (DOI: 10.1126/sciimmunol.adr0782).

The study uses transcriptomics data from bone marrow samples taken from people with relapsed leukemia, some of whom received and responded to DLI, some of whom received and didn’t respond to DLI, and a group that didn’t get DLI. This data can be used to identify a specific cell and what cells it may be communicating with by analyzing characteristic genes the cell is expressing and making inferences about what those gene products do.

But, study author Catherine Wu from the Dana Farber Cancer Institute says that this transcriptomics technique only captures what the cell is doing at one time, like a photograph. To see how cells’ gene expression changed, Wu’s team collaborated with Elham Azizi’s computational biology team at Columbia University to use DIISCO, a tool that characterizes cell-cell interactions over time. This tool allowed the researchers to see how cells were organizing with each other. Finally, to validate the computational findings, Wu’s team used microscopy and staining techniques to visualize these interactions.

People who responded displayed more activity from T cells known as ZNF683+ CD8+ cytotoxic T cells. Crucially, these T cells originated from the donor infusion, but they were able to coordinate with an existing, diverse microenvironment to eliminate leukemia cells. While nonresponders also received infusions of these cytotoxic T cells, a lack of diversity in the bone marrow microenvironment meant that these T cells did not coordinate an anticancer response.

Marie Bleakley, the chair for Leukemia Research at Fred Hutchinson Cancer Center and who was not involved with the research, says this paper introduces a robust dataset and new, cutting-edge analysis methods that could be applied to the area.

“How donor lymphocyte infusion impacts the leukemia and the marrow has been quite a black box, and this paper is shedding a bit more light on what’s actually happening in the marrow in response to the lymphocyte infusion,” she says.