HCB101: a novel potent ligand-trap Fc-fusion protein targeting the CD47-SIRPα pathway with high safety and preclinical efficacy for hematological and solid tumors

Abstract

Cluster of differentiation 47 (CD47) delivers an inhibitory signal that suppresses phagocytosis and prevents immune clearance of tumor cells by interacting with signal regulatory protein alpha (SIRPα) on myeloid cells. Although blockade of the CD47-SIRPα axis is a promising immunotherapeutic strategy, clinical development has been hindered by on-target toxicities (e.g., severe anemia) and insufficient potency. Herein we report a third generation CD47-SIRPα inhibitor HCB101, a rationally designed SIRPα-Fc fusion protein generated from a large-scale screening of a structure-guided SIRPα extracellular domain (ECD) mutant library and fused to a human IgG4 Fc. HCB101 demonstrates high-affinity binding to CD47, robustly promotes macrophage-mediated phagocytosis of tumor cells without affecting red blood cells and exhibits unique advantages over current CD47-targeting agents, including Hu5F9-G4, TTI-622, and ALX148. In multiple xenograft cancer models, HCB101 induced significant inhibition of tumor growth as a single agent and showed synergistic anti-tumor effects when combined with anti-HER2 or anti-EGFR monoclonal antibodies. Additionally, HCB101 treatment increased the M1/M2 macrophage ratio in the tumor microenvironment, suggesting repolarization of tumor-associated macrophages (TAMs) toward a pro-inflammatory phenotype. No dose-limiting toxicities or hematologic adverse effects were observed in murine or non-human primate studies.

Keywords: CD47; Cancer immunotherapy.; HCB101; Macrophage; Phagocytosis; SIRP-alpha.

Fig. 1

HCB101 schematic structure and in vitro activities. Structure of HCB101 is elucidated in (A). Top shows a schematic structure of HCB101, and the bottom is the simulated 3D structure of HCB101 which was visualized using PyMOL software for structural analysis, and the mutant sites on SIRPα ECD were labelled. HCB101 preferentially binds with high affinity to CD47-expressing cancer cells, but not to red blood cells. Serially diluted test articles were incubated with CD47-expressing Raji (B), FaDu (C), NCI-H82 (D), and hCD47 KO NCI-H82 (E) cells at 4 °C for 30 min. The binding was detected by PE-conjugated goat anti-human IgG Fcγ. Hu5F9-G4 and/or TTI-622 analogs are used as comparators. A PD1_ECD_IgG4 was used as negative control. HCB101-mediated blocking of human CD47 with SIRPα was assessed by flow cytometry (F-G): Serially diluted HCB101 or control molecules were incubated simultaneously with Raji (F) or FaDu (G) in the presence of the biotin-labeled human SIRPα-Fc. Bound ligands were detected by streptavidin-PE and analyzed by flow cytometry. The Hu5F9-G4 and TTI-622 analogs were used as comparators. PD1_ECD_IgG4 was used as a negative control; HCB101-induced macrophage-mediated phagocytosis of CD47-expressing NCI-H82 cells and human red blood cells were accessed (H–M): test articles were incubated simultaneously with macrophage cells (RAW264.7) and Celltrace™ Violet-labeled target cells, NCI-H82 (H) or CD47 KO NCI-H82 (I) or Raji (J) or RBCs isolated from three human donors (K-M) for two hours. The phagocytic activity was calculated as the percentage of CellTrace™ Violet⁺ F4/80⁺ cells within F4/80⁺ macrophages. The Hu5F9-G4 and TTI-622 analogs were used as comparators. PD1_ECD_IgG4 were used as negative controls, respectively. All data points are shown as mean ± SD for the triplicate determinations

Fig. 2

Anti-tumor activity of HCB101 in various xenograft models. (A) NOD/SCID mice were engrafted with Raji cells and treated with HCB101 at doses of 4.5, 1.5, and 0.5 mg/kg once weekly for two weeks (n = 10 per group). (B) NOD/SCID mice were intraperitoneally injected with HCB101 or TTI-622 analog five times per week for four weeks, beginning on day3 following Daudi tumor inoculation (n = 8 per group). (C) KG-1a was intravenously administrated into NPG™ mice (n = 12 per group). Mice were treated with HCB101 or TTI-622 analog five times per week for two weeks. The mouse survival was monitored daily. (D & E) NCI-H82 xenograft-bearing NOD/SCID mice were treated with HCB101, TTI-622 analog, or Hu5F9-G4 analog on the indicated days (n = 10 per group). (F) NOD/SCID mice were inoculated with WiDr cells and then treated with HCB101, TTI-622 analog, ALX148 analog, or Hu5F9-G4 analog when tumor volumes reached 100–200 mm³. Treatments were administered twice weekly for 3.5 weeks (n = 8 per group). (G) MDA-MB-453 cells were inoculated into NOD/SCID mice. Upon reaching a tumor volume of 100–200 mm³, mice received intraperitoneal injections of HCB101 at various doses twice weekly for three weeks and once weekly for an additional week (n = 5 per group). (H) NOD/SCID mice engrafted with SNU-C1 cells were treated with HCB101 at doses of 20 and 4 mg/kg, administered twice weekly for four weeks (n = 5 per group). (I) In NCI-N87 gastric cancer-bearing mice, treatment of 3 mg/kg for HCB101 or ALX148, with or without trastuzumab at 3 mg/kg, was initiated when tumor size reached 100–200 mm³. Treatments were administered biweekly for four weeks (n = 8 per group). (J) NOD/SCID mice bearing SW48 xenografts were treated with 10 mg/kg of HCB101, cetuximab, or a combination of both agents twice weekly for four weeks (n = 8 per group). All tumor volume data are presented as mean ± SEM. (K) Hematology in cynomolgus monkeys. WBC, RBC, HGB, and platelet counts in monkeys treated weekly with HCB101 (10–150 mg/kg) or placebo for 4 weeks. Data are mean ± SEM