Therapeutic Targeting of the Macrophage Immune Checkpoint CD47 in Myeloid Malignancies

Abstract

In recent years, immunotherapies have been clinically investigated in AML and other myeloid malignancies. While most of these are focused on stimulating the adaptive immune system (including T cell checkpoint inhibitors), several key approaches targeting the innate immune system have been identified. Macrophages are a key cell type in the innate immune response with CD47 being identified as a dominant macrophage checkpoint. CD47 is a "do not eat me" signal, overexpressed in myeloid malignancies that leads to tumor evasion of phagocytosis by macrophages. Blockade of CD47 leads to engulfment of leukemic cells and therapeutic elimination. Pre-clinical data has demonstrated robust anti-cancer activity in multiple hematologic malignancies including AML and myelodysplastic syndrome (MDS). In addition, clinical studies have been underway with CD47 targeting agents in both AML and MDS as monotherapy and in combination. This review will describe the role of CD47 in myeloid malignancies and pre-clinical data supporting CD47 targeting. In addition, initial clinical data of CD47 targeting in AML/MDS will be reviewed, and including the first-in-class anti-CD47 antibody magrolimab.

Keywords: AML; CD47; MDS; immunotherapy; leukemia stem cell (LSC); macrophage.

Figure 1

Mechanism of action of CD47 blocking antibodies. (A) Under normal conditions, normal and cancer cells evade macrophage phagocytosis by expressing CD47. In cancer cells CD47 is overexpressed to protect against the expression of eat me/pro-phagocytic signals. (B) With CD47 blockade (with an anti-CD47 antibody), cancer cells are phagocytosed due to CD47 blockade and resulting unmasking of the “eat me” signal. In contrast, normal cells are spared given the lack of expression of pro-phagocytic signals.

Figure 2

Magrolimab combination with azacitidine enhances therapeutic phagocytosis and pre-clinical efficacy in AML. (A) Calreticulin cell surface binding sites were assessed by flow cytometry on HL60 AML cells in the presence of increasing concentrations of azacitidine that are comparable to human exposure. (B) in vitro phagocytosis by human macrophages of HL60 cells with two different macrophage donors was evaluated in the presence of IgG4 control, 5F9/magrolimab, azacitidine (AZA), or the combination. Triplicate experiments were conducted. (C) HL60 mice were engrafted into immunodeficient NOD/SCID/IL2-R-gamma knockout (NSG) mice intravenously with engraftment assessed by bioluminescence imaging. Post-engraftment, mice (n = 10 each) were treated with PBS control, 5F9, AZA, or the combination with treatment initiated on either Day 4 (D4) or Day 7 (D7) post-engraftment as indicated. No leukemic disease was detected in mice treated with the combination that exhibited long-term survival.