AML Treatment By the AXL Inhibitor Bemcentinib in Combination with Cytarabine Shows Clinical Efficacy Related to TNFα and Cytotoxic Immune Cells: A Single-Cell Translational Study from the BGBC003 Trial

Acute myeloid leukemia can be treated by intensive chemotherapy. However, this treatment is unsuitable for a subset of patients because of age and/or co-morbidities. Such patients benefit from hypomethylating agents and venetoclax but frequently develop resistance to this treatment regimen. This calls for the development of novel drugs that can prolong survival in AML patients unfit for intensive chemotherapy. The receptor tyrosine kinase AXL is a negative prognostic factor in AML. Its activation promotes survival, chemoresistance and proliferation of AML blasts. In addition, it is also present on innate immune cells and contributes to the formation of immunosuppressive environments. Together, this makes AXL a promising target for the treatment of AML. Bemcentinib is a first-in-class, orally available, selective inhibitor of AXL. Here, we report a single-cell translational sub-study in a sub-cohort of patients unfit for intensive chemotherapy (B2+B5) from the phase Ib/II clinical trial BGBC003 (NCT02488408). This sub-cohort was treated by a combination of bemcentinib and low-dose cytarabine (LDAC). To gain first insights into the mechanisms underlying treatment response, we conducted a translational study evaluating 13 participating patients.

Among 32 patients evaluated for efficacy, 8 patients responded to their treatment (objective response rate: 25%). Among patients who received previous therapy ( i.e. relapsed/refractory patients), objective response rate was 18.5% (5/27). The median overall survival was 8.0 months, with a notable survival benefit (median overall survival of 24.8 months) in responders. Relapsed/refractory patients exhibited a median overall survival of 7.8 months. These data highlight that bemcentinib-LDAC benefits a considerable subset of AML patients unfit for intensive chemotherapy.

In order to gain mechanistic insights into the efficacy of bemcentinib-LDAC, we profiled cells from bone marrows of 13 participating patients (6 responders, 7 non-responders, according to best response) using single-cell transcriptomics and multi-omics (CITE-seq). Cell type annotation highlighted various immune cell populations next to AML blasts. Successful treatment was associated with stronger TNFα signaling in blasts before treatment. A tight link between TNFα and AXL could subsequently be established in vitro as several AML cell lines up-regulated expression of AXL upon exposure to TNFα. Inhibiting AXL in these cell lines using bemcentinib increased expression of TNFα. This indicates a potential negative feedback loop between these two players. Furthermore, cytotoxic immune cells (CD8 ⁺ effector T cells, γδ T cells and natural killer cells) from responders displayed evidence for increased pro-inflammatory signaling upon bemcentinib-LDAC treatment. This included TNFα signaling programs next to IFNα and IFNγ signaling programs. Thus, our data indicate that bemcentinib-LDAC promotes the activity of such cytotoxic cells. In line with this, we observed increasing crosstalk between immune cells upon bemcentinib-LDAC treatment in responders. Importantly, cells from non-responders generally exhibited the opposite trend, highlighting the aforementioned cell types and pathways as factors that distinguish responders from non-responders. Together, our results indicate a potential role of TNFα and cytotoxic immune cells in the successful application of bemcentinib-LDAC.

In conclusion, our findings warrant further clinical development of bemcentinib-LDAC for AML patients unfit for intensive chemotherapy. Additionally, extended research on TNFα and cytotoxic immune cells may lead to a better understanding of what mechanisms cause a treatment response. This may ultimately enable an accurate selection of patients who will benefit from bemcentinib-LDAC.