CD47, an immune checkpoint protein expressed on malignant cells, acts as a “don’t eat me” signal by binding to signal regulatory protein alpha (SIRPα) on macrophages, enabling cells to evade phagocytosis. High CD47 expression in acute myeloid leukemia (AML) is associated with adverse outcomes. We have previously reported that the combination of venetoclax and azacitidine (VEN/AZA) increased macrophage phagocytosis in preclinical models, with or without CD47 antibody (5F9, magrolimab) (Jia et al., Blood 2021). In this study, we aimed to elucidate the underlying mechanisms of the VEN/AZA therapy on macrophage recognition using multi-omics approaches, including CRISPR screening and proteomics.
First, we optimized the screening conditions for VEN/AZA pre-treatment, selecting the dose of 10nM VEN with 1μM AZA. This dose induced 32.7±6.8% phagocytosis (calculated as the percentage of Calcein+CD206+ cells from total CD206+ cells) of MOLM-13 cells after 24-hour pre-treatment followed by 1-hour co-culture with macrophages, compared to 24.5±6.0% in DMSO-treated cells (p=0.03). CRISPR screening using the Brunello sgRNA library revealed complementary characteristics in cells treated with VEN/AZA and VEN/AZA/macrophages. Genes including TP53, POU4F2, AKT2, ADIPOR1, etc. were positively selected in VEN/AZA-treated cells (compared to DMSO), but negatively selected in VEN/AZA/macrophages-treated cells (compared to VEN/AZA). This indicated that silencing of these genes confers leukemia cells resistance to VEN/AZA but increased sensitivity to macrophage-mediated phagocytosis. These overlapping genes were enriched in pathway regulating TP53 activity through association with co-factors (such as TP53, POU4F2 and AKT2), which influence p53 stability and the transcription of cell cycle arrest and death genes. Collectively, these data suggest that the activity of VEN/AZA in the presence of macrophages modulates TP53-dependent pathways.
To further test the hypothesis that TP53-mutant AML might derive benefit from triple therapy, we utilized TP53-mutant MOLM13 cell line and TP53-mutant patient-derived xenograft (PDX) models. In MOLM-13 cell line model, 5F9/VEN/AZA induced 40.8±2.9% phagocytosis in TP53-mutant cells and 40.5±3.1% in TP53-wild type cells (p=0.85). In the TP53-mutant PDX MDAM19692 (FLT3-ITD, DNMT3A, NPM1, TP53), mice were randomized into four groups (6-7 mice per cohort) after confirming peripheral blood engraftment by flow cytometry (14 days post-transplant). After two weeks treatment, the 5F9 and 5F9/VEN/AZA demonstrated significantly reduced leukemia burden, with an average hCD45% level of 0.3% for the 5F9 and 0.1% for combination, compared to 33.4% for vehicle, 30.0% for the VEN/AZA. The combination cohort achieved significantly better overall survival (p=0.01) with a median overall survival of 82.5 days, compared to 32, 29, and 37 days for vehicle, VEN/AZA, and 5F9 groups, respectively.
To understand the mechanisms underlying VEN/AZA-induced phagocytosis, we employed the subcellular protein fractionation and enzymatic shaving techniques to isolate different cellular compartments (including cytoplasmic, nuclear, membrane, and cytoskeletal proteins) and the exposed portions of membrane proteins at various time points during VEN/AZA treatment, respectively. Mass spectrometry analysis of the resulting peptides, followed by clustering trend analysis, based on the fuzzy C-mean algorithm, and Gene Ontology analysis of membrane proteins, revealed that VEN/AZA pretreatment primarily increased the expression of proteins involved in cellular transport and localization pathways. The shaving assay demonstrated that many of these upregulated proteins originated from the nucleus.
In summary, our study uncovered the complementary signature as a putative mechanism for the combinatorial activity of VEN/AZA therapy of AML in contact with macrophages. Our data indicate that VEN/AZA activates the translocation of nuclear proteins to the cell surface. These translocated proteins may function as damage-associated molecular patterns (DAMPs), activating macrophage pattern recognition receptors and enhancing phagocytosis. These findings provide a rationale for combining VEN/AZA with macrophage-based therapies in AML treatment, particularly in TP53-mutant cases. Further validation of the identified pathways is ongoing and will be reported.
Recent Comments