The CD47-Blocking Immune Checkpoint Inhibitor Maplirpacept Does Not Interfere with Blood Transfusion Compatibility Testing

CD47 is a transmembrane glycoprotein that delivers an anti-phagocytic (“do-not-eat”) signal by binding to signal regulatory protein α (SIRPα) on the surface of macrophages. Many tumors express high levels of CD47 as a means to escape macrophage-mediated immune surveillance and blockade of the CD47-SIRPα pathway has emerged as a novel therapeutic strategy with demonstrated anti-tumor activity both in vitro and in vivo (Petrova et al. 2017; Weiskopf 2017). Maplirpacept (PF-07901801, TTI-622) is a soluble human SIRPαFc fusion protein with an IgG4 domain which acts as an innate immune checkpoint inhibitor that prevents CD47-SIRPα binding, engages activating Fcγ receptors on macrophages, and promotes phagocytosis of malignant cells (Lin et al. AACR 2018). Maplirpacept is currently in clinical development to treat patients with ovarian cancer, multiple myeloma, acute myeloid leukemia and lymphoma.

Additionally, CD47 is expressed on red blood cells (RBCs) and plays a role in the natural clearance of aged RBCs by splenic macrophages in vivo (Oldenborg et al. 2000). CD47 expression on RBCs was shown to interfere with pre-transfusion laboratory testing in patients treated with the anti-CD47 mAb, Hu5F9-G4 (Velliquette et al. 2019). Unlike anti-CD47 mAbs, minimal binding to human RBCs is observed with Maplirpacept (Lin et al. AACR 2018). In this study we investigated whether the minimal binding of Maplirpacept to human RBCs precludes interference with routine pre-transfusion laboratory tests.

RBC suspensions from healthy donors representing all 8 ABO-Rh blood groups were spiked with Maplirpacept at 0.6, 1.2 or 2 mg/ml or anti-CD47 (clone Hu5F9). Blood typing was performed according to the manufacture's protocol (Seraclone, Biorad) and observed for the presence or absence of agglutination. For cross matching experiments, plasma was prepared from blood of healthy donors and spiked with Maplirpacept or anti-CD47 at the same concentrations used for the blood typing experiments. RBCs from two healthy donors were mixed with plasma and an indirect antiglobulin test (IAT) was performed according to the manufacture's protocol (Coombs Anti-IgG, Biorad).

Using RBCs from healthy donors representing the 8 major ABO-Rh blood groups, we now show that the addition of Maplirpacept does not cause agglutination of test RBCs and produces the expected blood typing result in a standard forward blood typing test. No interference was observed at the highest concentration tested. However, addition of an anti-CD47 mAb caused pan-agglutination as well as false positive blood typing results (Table 1). In a standard IAT that mimics the in vitro compatibility cross matching test performed prior to a blood transfusion, the anti-CD47 mAb gave false positive results while Maplirpacept produced results identical to the control serum condition (Table 2). This data is consistent with a study demonstrating interference by Hu5F9 in routine pre-transfusion blood testing (Velliquette et al. 2019). Additionally, no interference was observed when Maplirpacept was added to conditions which produced agglutination in a standard IAT across a range of blood types.

Interference with RBC panel testing has important consequences in transfusion medicine as the presence of irregular antibodies can complicate the selection of suitable RBC units for treated patients. Here we demonstrate that interference with blood compatibility testing is dependent upon the type of CD47-blocking agent. Decoy receptors such as Maplirpacept, unlike anti-CD47 mAb, do not interfere with blood group serologic testing, which may confer a significant clinical advantage by negating the need for additional approaches to prevent interference.