Abstract
8F4, a T cell receptor (TCR) - like antibody which targets the leukemia-associated antigen PR1 (VLQELNVTV) when presented in the context of MHC class I (HLA-A2:01), is a mouse IgG2a monoclonal antibody (mAb) which mediates both complement-dependent cytotoxicity (CDC) and antibody-dependent cellular cytotoxicity (ADCC) in vitro and eliminates AML in PDX (patient derived xenograft) mouse models. Recognizing the potential therapeutic value of 8F4, we produced a humanized IgG1 (Hu8F4) as a first-in-class mAb for preclinical development and clinical use.
Characterization of Hu8F4 by ELISA and surface plasmon resonance showed high affinity (KD=6.5 nM) and specificity of Hu8F4 binding to PR1/HLA-A2 monomer. Staining of peptide-pulsed T2 cells confirmed Hu8F4 specificity to PR1, but not to control peptide MART1. To study anti-leukemia activity of Hu8F4 we used HLA-A2-transfected leukemia cell lines U937 (U937-A2) and MV411 (MV411-A2), as well as THP1 (endogenous HLA-A2+) as targets. In contrast to mouse 8F4, Hu8F4 did not induce CDC in vitro. Rather, Hu8F4 mediated both potent antibody-dependent cellular phagocytosis (ADCP) and ADCC of U937-A2 and THP1 cells in NFAT reporter assays, using FcgRI- (EC50 = 0.002 µg/ml), FcgRII- (EC50 = 1.49 µg/ml), and FcgRIII- (EC50 = 0.04 µg/ml) transfected Jurkat cells, as well as pooled human PBMC (EC50 = 0.95 µg/ml) and donor-derived macrophages (EC50 = 1.1 µg/ml), as effectors. Based on the ADCP activity of Hu8F4, we sought to examine the potential synergy of anti-CD47 antibody, a macrophage checkpoint inhibitor, with Hu8F4. Treatment of U937-A2 and THP1 target cells with anti-CD47 F(ab')2 neutralizing antibody made cells susceptible to Hu8F4-mediated phagocytosis by NSG mouse bone marrow derived macrophages (BMDM) in vitro. To confirm the critical role of CD47 in Hu8F4-mediated ADCP we created CD47 KO U937-A2 and THP1 AML lines. Within 96 hours of treatment with Hu8F4 both CD47 KO cell lines were completely phagocytosed by NSG BMDM, thus confirming the role of phagocytosis as a mechanism.
To test Hu8F4 activity in vivo, we established xenografts using luciferase-transfected U937-A2 and MV411-A2 cell lines in NSG mice. Hu8F4 was then administered starting on either day 3 (U937-A2) or day 7 (MV411-A2). Treatment with Hu8F4 resulted in elimination of leukemia as measured by FACS and BLI and increased median survival (from 33 days to >85 days for U937-A2, p=0.0075; and from 125 days to >225 days for MV411-A2, p=0.0033). To further test Hu8F4 efficacy, we established PDX in NSG mice using cells from HLA-A2+ refractory/relapsed AML patients. Hu8F4 treatment reduced leukemia by 90.3% to 99.9% when assessed by FACS in four out of five PDX and resulted in increased median survival of the mice (from 41 days to 49 days, p=0.0002) in one PDX tested.
To study the mechanism of action of Hu8F4, we produced a 3 aa mutant (Hu8F4 PG LALA), which completely abrogated binding to Fcg receptors. Hu8F4 PG LALA, as well as Hu8F4 F(ab')2, showed no in vivo activity against U937-A2 or MV411-A2 xenografts, confirming that the anti-leukemia activity of Hu8F4 is dependent on its ability to bind Fc receptors. Because FcgR may be expressed on AML we tested fratricide as a possible in vivo mechanism of Hu8F4. We utilized the CRISPR /Cas9 system to edit out both FcgRI and II in the U937-A2 cell line that is naturally deficient in FcgRIII (DKO U937-A2). We then used a retroviral system to stably express FcgRI or FcgRII in U937-A2 to create FcgRI-U937-A2 and FcgRII-U937-A2, respectively, and tested their susceptibility to Hu8F4 in the xenograft model. Compared to FcR-expressing wild-type U937-A2, Hu8F4 showed similar anti-leukemia activity in all three xenografts. There was no significant difference noted in the anti-leukemia potency of Hu8F4 between DKO U937-A2 or single FcR-expressing xenografts compared with FcgRI/FcgRII expressing wild-type U937-A2 xenografts, suggesting that fratricide is an unlikely mechanism in this model.
Here we demonstrate high specificity, affinity, and remarkable anti-AML activity of Hu8F4, a first-in-class humanized TCR mimic monoclonal antibody that is currently being tested in a phase I clinical trial for HLA-A2+ refractory/relapsed AML and MDS. This study provides evidence for the importance of FcgR-mediated mechanisms (ADCC and ADCP) in the activity of Hu8F4. In addition, our results support future combination trials with treatments that promote phagocytosis.
Disclosures
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.