The Leukemic Microenvironment Represses Blinatumomab-Dependent Killing in Pediatric B-Cell Precursor Acute Lymphoblastic Leukemia

Background and aim

Blinatumomab is a bispecific T-cell engaging antibody that binds to CD19 on B-cells and CD3 on T-cells, thereby initiating an interaction between these cells and leading to the subsequent lysis of B-cells. Blinatumomab has demonstrated promising results in pediatric B-cell precursor acute lymphoblastic leukemia (BCP-ALL) patients. However, not all patients show an initial or lasting response to blinatumomab, and the mechanism behind this is not clearly understood.

T-cells are highly sensitive to modulation by factors such as cytokines and co-stimulatory signals, which can impact T-cell activity, and subsequently the working of blinatumomab. In BCP-ALL, leukemic cells exploit the bone marrow (BM) microenvironment and actively modulate the supportive function of stromal cells, such as mesenchymal stromal cells (MSCs). We showed that this leukemic BM microenvironment is markedly different from its healthy counterpart, facilitates chemotherapy resistance and secretes specific cyto- and chemokines (Polak et al., Blood 2015; Smeets et al, Haematologica 2024). B-ALL-derived MSCs have been observed to suppress T-cell proliferation and reduce pro-inflammatory cytokine secretion (Zanetti et al., JITC 2020). This study aims to identify the effect of the leukemic BM microenvironment on the blinatumomab-dependent killing of BCP-ALL cells.

Results

An ex vivo co-culture of patient derived xenograft (PDX) BCP-ALL cells and pre-activated allogeneic T-cells (with IL-2, IL-7, IL-15) was created with or without patient-derived MSCs. Forty hours after initiation, the co-culture was exposed to 1nM blinatumomab for 48 hours. Cell survival was assessed using flow cytometry. A lower blinatumomab-dependent killing was observed in 3 out of 4 BCP-ALL samples in the presence of MSCs (median reduction of 20.5%), which was independent of T-cell donor (n=2).

Next, BM-derived samples of newly diagnosed pediatric BCP-ALL patients were used to confirm the observed immunosuppressive effect of the leukemic microenvironment. These samples contain both BCP-ALL cells and varying numbers of autologous T-cells that have been exposed to cytokines and chemokines in the leukemic microenvironment of the patient. Nineteen primary BCP-ALL samples were ex vivo exposed to 1 nM blinatumomab for 48 hours. Fourteen (~75%) BCP-ALL samples did not reach a blinatumomab-dependent killing of leukemic cells greater than 45% and were considered ex vivo blinatumomab non-responders. Following CD3/CD28 activation of autologous T-cells, there was a 30% increase in BCP-ALL cell kill after 48 hours blinatumomab in 5 out of 6 tested samples which were non-responding without this activation. These data suggests that T-cell function is often repressed in the BM at the time of the diagnosis of BCP-ALL.

The effector-to-target (E:T) ratio, the CD8+:CD4+ T-cell ratio and CD19 surface expression before and after treatment did not differ between responding and non-responding cases, nor between activated and non-activated patient samples. Furthermore, the differential ex vivo response of BCP-ALL cells to 1nM blinatumomab could not be attributed to aberrations in commonly deleted genes in BCP-ALL (BTG1, CDKN2A/B, EBF1, ETV6, IKZF1, PAX5, PAR1, RB1). The blinatumomab killing assay included BCP-ALL samples with ETV6::RUNX1, high hyperdiploidy, near haploidy, DUX4 rearrangement and B-other subtype. Interestingly, only 2 out of 10 samples with the generally favorable subtype ETV6::RUNX1 were ex vivo responsive to blinatumomab. Furthermore, ex vivo responses to blinatumomab were not predictive for event-free survival (EFS) of patients receiving 2 years of chemotherapy.

Conclusion

We found that T-cells in the leukemic BM microenvironment were suppressed at diagnosis of pediatric BCP-ALL. Activation of these T-cells restored the effector function and resulted in a blinatumomab-dependent killing of BCP-ALL cells. We noticed that blinatumomab was less effective in 14 out of 19 BCP-ALL samples taken from BM at the time of diagnosis. T-cell function could be recovered in 5 out of 6 of these non-responding samples by activation. The limited ex vivo blinatumomab-dependent killing of BCP-ALL cells in diagnosis samples warrants a careful implementation of blinatumomab in upfront full-blown leukemia.