Humanized Anti-PR1/HLA-A2 T Cell Receptor-like Monoclonal Antibody, h8F4, Has Multiple Effector Mechanisms

Previously we developed a T cell receptor (TCR) -like murine IgG2a monoclonal antibody, m8F4, which binds to the PR1 peptide/HLA-A2 complex, and targets both HLA-A2+ leukemia cell lines and primary HLA-A2+ patient blasts of various AML subtypes. We reported that m8F4 mediated complement dependent cytotoxicity (CDC) of AML in vitro and actively depleted AML in vivoin AML patient-derived xenograft (PDX) models. Due to the therapeutic potential of 8F4, we humanized 8F4 for clinical development. Here we report the specificity of this humanized IgG1κ 8F4 (h8F4) to PR1/HLA-A2 and document the activity of h8F4 against AML.

Both ELISA and surface plasmon resonance (SPR) were used to study h8F4 specificity to recombinant monomer, consisting of PR1 peptide, HLA-A2 and β₂ microglobulin. Like mouse 8F4, human 8F4 bound PR1/HLA-A2 monomer with high affinity (K_D= 6.5 nM), while no binding was detected to control monomers pp65/HLA-A2 and WT1/HLA-A2. Moreover, FACS of peptide-pulsed T2 cells showed that h8F4 bound PR1- pulsed T2, but did not bind T2 cells loaded with other HLA-A2-restricted peptides, including MART1, pp65, WT1 and gp100, thus confirming h8F4 specificity to PR1/HLA-A2. To test whether h8F4 retains the in vivoanti-leukemia activity seen with m8F4, we used an NSG mouse model. Similar to m8F4, h8F4 treatment reduced leukemia growth in NSG mice with established disease in 3 out of 4 AML PDX models tested.

To determine the mechanism of action (MOA) of humanized 8F4, we first tested h8F4 activity in the presence of complement because we had previously demonstrated the primary MOA of m8F4 to be CDC. However, unlike m8F4, even when tested at high concentrations up to 20 µg/ml, h8F4 did not mediate CDC of HLA-A2 transduced U937, HL60, or PR1-loaded T2 cells. To further examine h8F4 MOA, we tested antibody dependent cellular cytotoxicity (ADCC) of h8F4 using an lactase dehydrogenase (LDH) assay with pooled human PBMC as effectors. H8F4 induced ADCC of U937-A2 at E:T ratio 50 (EC₅₀=0.95 µg/ml), thus indicating improved ADCC anti-AML activity of h8F4. Introduction of alanine substitutions in positions 234-235 (AAh8F4), which disrupted the binding of h8F4 to Fc receptor (FcR), abolished ADCC activity against U937-A2 in vitro. In contrast, AA8F4 still retained partial in vivo activity against the same target cells in the NSG mouse model, suggesting that ADCC is important, but not the exclusive mechanism of h8F4 in vivo. To investigate whether h8F4 had direct activity against AML, as possible alternative MOA, we measured apoptosis of target cells by h8F4 with and without crosslinking anti-Fc antibody fragments (CLA). Indeed, we found, by using annexin V and propidium iodide staining, that a 24 hours incubation with h8F4 (1µg/ml) in the presence of CLA induced apoptosis of 55.5 ± 3.5% of U937-A2 cells, as compared to 23.5 ± 2.1% in CLA-only control (p < 0.05). Moreover, caspase 3 activation was observed as early as 10 minutes after adding h8F4+CLA, and was time- and h8F4 concentration- dependent.

Our data demonstrate that specificity of humanized anti-PR1/HLA-A2 TCR-like antibody, h8F4, is preserved and that h8F4 has a similar affinity compared with m8F4. In addition to inducing apoptosis, h8F4 mediates greater ADCC compared with m8F4, but lacks CDC activity in vitro. Together, our data is consistent with multiple mechanisms of action in vivo, including ADCC. Our results support further testing of h8F4 for patients with AML.