Bone marrow cancer leaves immune cells poorly prepared to defend

A previously underexplored population of immune cells in the bone marrow may help explain why some patients with multiple myeloma or acute myeloid leukemia respond strongly to immunotherapy while others do not. Researchers from the German Cancer Research Center (DKFZ), the HI-STEM Stem Cell Institute, and Medical Clinic V at Heidelberg University Hospital have identified a distinct group of tumor-reactive T cells that retains the molecular machinery needed to recognize malignant cells, yet appears to remain insufficiently activated inside the body. Their findings suggest that these cells could become valuable biomarkers for predicting treatment response and potential starting points for new therapies.

T cells are central components of the adaptive immune system. They identify abnormal cells through receptors that bind specific peptide fragments displayed by major histocompatibility complex molecules on the cell surface. In many solid tumors, persistent exposure to cancer antigens and suppressive signals drives T cells into a state commonly known as exhaustion. Exhausted T cells often show reduced proliferation and impaired effector functions, including the ability to release cytotoxic molecules. The Heidelberg study indicates that tumor-reactive T cells in bone marrow cancers may follow a different biological path.

The researchers conducted a combined molecular and functional analysis of immune cells collected from the bone marrow of 21 patients with multiple myeloma or acute myeloid leukemia. Multiple myeloma develops from malignant plasma cells, whereas AML arises from abnormal blood-forming precursor cells. Both diseases occupy the bone marrow, an immune environment with distinctive cellular interactions, nutrient conditions, and concentrations of signaling molecules. By integrating single-cell gene-expression analysis, T-cell receptor profiling, and laboratory experiments, the team examined which T cells were responding to tumor-associated antigens and how their behavior differed from that of other immune cells.

The analysis revealed that tumor-reactive T cells retained several features associated with functional immune responses. They expressed genes and cellular programs linked to activation, antigen recognition, and the potential to develop into effective killer cells. At the same time, they did not appear to be continuously engaged in eliminating cancer cells in the bone marrow. Instead, the researchers describe them as being in a state of “conditional preparedness.” These cells appear equipped for an immune response but do not receive, or do not sustain, enough stimulation to attack tumors effectively under natural conditions.

This distinction could help clarify why immunotherapies that activate T cells can work in blood cancers even when the patient’s immune system has not spontaneously controlled the disease. Bispecific antibodies, for example, are engineered to bind both a molecule on a cancer cell and a molecule on a T cell. By physically bringing the two cells together, these drugs can trigger T-cell activation, immune synapse formation, and the release of cytotoxic proteins such as perforin and granzymes. According to the study, this therapeutic mechanism may awaken bone marrow T cells that are biologically capable of responding but remain inadequately stimulated in their tumor environment.

The investigators also identified a molecular signature based on the activity of 15 genes. This signature distinguished tumor-reactive T cells from other T-cell populations and showed strong predictive performance in an independent patient group. The finding is important because conventional approaches often rely on individual surface markers, which may be shared by several functionally different immune-cell states. A multi-gene profile can capture a broader biological program, potentially providing a more reliable way to estimate the abundance and activity of tumor-reactive T cells in a patient sample.

In multiple myeloma, patients who later responded more favorably to treatment already had higher numbers of these signature-positive T cells before therapy began. During treatment with a bispecific antibody, tumor-reactive T-cell clones expanded preferentially, suggesting that the drug was recruiting the very population identified by the researchers. A greater baseline abundance of these cells was associated with a more favorable clinical course. A similar relationship was observed in AML among patients receiving immune-based treatment. The signature did not predict responses to conventional chemotherapy, supporting the idea that it reflects the state of antitumor immunity rather than simply indicating a less aggressive disease.

The study also explored the molecular targets recognized by these T cells. The researchers examined more than 17,000 distinct peptide fragments presented by cancer cells and found antigens that appeared in multiple patients. Some of these tumor-associated targets were detected in both multiple myeloma and AML, raising the possibility that shared antigenic features could support the design of immunotherapies with broader applicability. Such targets might eventually be used to develop vaccines, engineered T-cell therapies, or T-cell receptor-based treatments, although their safety and effectiveness would require extensive validation.

The findings come with important limitations. The work involved a relatively small patient cohort, and the 15-gene signature is not yet ready for routine clinical use. Larger prospective studies will be needed to determine whether it can reliably guide treatment decisions across different patient populations and therapies. In addition, several functional experiments were performed in vitro. The researchers demonstrated that the T cells recognize cancer cells and can be activated, but they have not yet established direct evidence that these cells consistently kill malignant cells inside patients. Even so, the study provides a detailed map of bone marrow immunity and suggests that the success of future immunotherapies may depend not only on the drug itself, but also on whether the patient possesses a sufficient reserve of tumor-reactive T cells capable of being switched on.

Web References: https://doi.org/10.1016/j.ccell.2026.07.011

References: Kehl et al., “Latent effector T cells mediate immunotherapy responses in the bone marrow microenvironment,” Cancer Cell (2026). DOI: 10.1016/j.ccell.2026.07.011