AML Immunotherapy Using a Novel Tcrm-Based Bispecific Antibody That Targets a Leader Sequence Peptide Derived from Cathepsin G

Introduction: Myeloid azurophil granules provide a rich source of intracellular leukemia antigens. Cathepsin G (CG) is a serine protease that has higher expression in AML blasts in comparison to normal myeloid progenitors. Based on the unique biology of HLA-A*0201 (HLA-A2), in which presentation of leader sequence (LS)-derived peptides is favored, we focused on the LS-CG-derived peptide CG1 (FLLPTGAEA). LS peptides are naturally processed and loaded on HLA-A2, providing an abundant source of surface peptide/HLA (pHLA) targets. We eluted CG1 from the surface of primary HLA-A2 ⁺ AML blasts and AML cell lines, and demonstrated CG1-targeting immunity in leukemia patients following allogeneic stem cell transplant, hence highlighting the potential for CG1 to be a promising immunotherapeutic target in AML.

Numerous platforms have been developed to target cell surface pHLA complexes. T cell receptor (TCR) mimic (m) antibodies are immunotherapeutic antibodies that target pHLA, which are the natural ligands for the TCR. TCRm antibodies have been a breakthrough in immunotherapy, as they: (1) target intracellular antigens expressed on HLA; (2) are engineered to have high affinity for HLA; (3) can elicit their cytotoxic effect independent of the TCR; and (4) can be produced in large quantities providing an off-the shelf-product for patient use. Here we report on the engineering, preclinical efficacy, and safety evaluation of a novel CG1-targeting, T cell-engager, bispecific antibody (CG1xCD3e), which incorporates a novel TCRm construct that recognizes cell surface CG1/HLA-A2 complexes.

Methods: To generate the CG1/HLA-A2-binding monoclonal antibodies, H2L2 human transgenic mice were immunized with CG1/HLA-A2 monomers subcutaneously. The resulting paired heavy and light chains were cloned from the antigen-sorted memory B cells into the single pcDNA3.1(+) vector with both human (h) immunoglobulin (Ig)G1 and hKappa for subsequent ExpiCHO high-throughput transient expression. Antibody binding was assessed by Bio-Layer Interferometry (BLI) and T2 cellular binding assays. Xencor anti-CD3e scFv and Knobs-in-holes technology was applied for the bispecific antibody engineering for xMab format. In vitro activity of CG1xCD3e and T cell cytokine secretion was confirmed using flow cytometry-based T cell-dependent cellular cytotoxicity (TDCC).

The specificity of CG1xCD3e for leukemia was assessed using colony forming unit assays. In vivo efficacy was evaluated in NSG mice engrafted with AML cell lines and treated with CG1xCD3e with normal donor peripheral blood mononuclear cells (ND-PBMC).

Results: Using BLI, binding avidity of the CG1xCD3e to CD3-Fc, KD= 7.52 x 10-10 M, and affinity to CG1/HLA-A2, KD= 1.28 x 10-10 M were calculated. Using flow cytometry, we confirmed high affinity binding of CG1xCD3e to CG1/HLA on the CG1- expressing EM2 HLA-A2 ⁺ cell line, and also demonstrated binding of CG1xCD3e to CD3 using jurkat T cell line.

After co-culturing GFP/Luciferase ⁺ HLA-A2 ⁺ AML (U937-A2 and ML-2) or CML (EM2) cell lines with ND-PBMC and different concentrations of CG1xCD3e bispecific antibody for 24 or 48 hours, flow cytometry analysis demonstrated highly specific CG1/HLA-A2 dependent killing of AML cells by T cells cultured with CG1xCD3e bispecific antibody, in comparison with control bispecific antibody with a disabled CG1 binding arm. This correlated with both tumor- and bispecific antibody- dependent T cell activation and cytokine secretion. To study the in vivo anti-leukemia activity of CG1xCD3e, we engrafted NSG mice with the HLA-A2 ⁺ ML-2 and U937-A2 AML cell lines, and treated mice after tumor engraftment with ND-PBMC (1 x 107 cells) and weekly intraperitoneal injections of CG1xCD3e bispecific antibody (0.01 mg/Kg, 0.05 mg/Kg, and 0.1 mg/Kg) or PBS. Bioluminescence imaging demonstrated that the mice treated with ND-PBMC + CG1xCD3e had significantly lower levels of leukemia burden compared to the mice treated with ND-PBMC alone (Figure 1). Lastly CFU assays using HLA-A2 ⁺ normal donor marrow in a semi-solid matrix of methylcellulose in the presence of T cells and either CG1xCD3e bispecific antibody or isotype antibody, confirmed specificity of CG1xCD3e bispecific antibody for leukemia.

Conclusion: Our study provides strong pre-clinical evidence supporting the targeting of LS- derived peptides, specifically CG1, in the setting of AML using a novel TCRm-based bispecific antibody.