A Novel T-Cell Engaging Bi-specific Antibody Targeting the Leukemia Antigen PR1/HLA-A2

Abstract

Despite substantial advances in the treatment of acute myeloid leukemia (AML), only 30% of patients survive more than 5 years. Therefore, new therapeutics are much needed. Here, we present a novel therapeutic strategy targeting PR1, an HLA-A2 restricted myeloid leukemia antigen. Previously, we have developed and characterized a novel T-cell receptor-like monoclonal antibody (8F4) that targets PR1/HLA-A2 and eliminates AML xenografts by antibody-dependent cellular cytotoxicity (ADCC). To improve the potency of 8F4, we adopted a strategy to link T-cell cytotoxicity with a bi-specific T-cell-engaging antibody that binds PR1/HLA-A2 on leukemia and CD3 on neighboring T-cells. The 8F4 bi-specific antibody maintained high affinity and specific binding to PR1/HLA-A2 comparable to parent 8F4 antibody, shown by flow cytometry and Bio-Layer Interferometry. In addition, 8F4 bi-specific antibody activated donor T-cells in the presence of HLA-A2⁺ primary AML blasts and cell lines in a dose dependent manner. Importantly, activated T-cells lysed HLA-A2⁺ primary AML blasts and cell lines after addition of 8F4 bi-specific antibody. In conclusion, our studies demonstrate the therapeutic potential of a novel bi-specific antibody targeting the PR1/HLA-A2 leukemia-associated antigen, justifying further clinical development of this strategy.

Keywords: PR1; acute myeloid leukemia; bi-specific antibody; cancer immunotherapy; re-directed cytotoxicity.

Figure 1

Construction, expression, and purification of 8F4 bi-specific antibody. (A) Schematic diagram of the soluble 8F4 bi-specific antibody. Light chain variable fragments (VL) and heavy chain variable fragments (VH) of 8F4 and murine OKT3 are linked in tandem via flexible glycine-serine linkers of either 10 or 15 amino acids. A His₆Tag was included at the C terminus for purification. (B) SDS-PAGE of elution fractions of 8F4 bi-specific antibody from immobilized Nickel affinity chromatography. (C) Western blot analysis of 8F4 bi-specific antibody from SDS-PAGE shown in part B, probed with anti-HisTag antibody. (D) Fast protein liquid chromatography of soluble 8F4 bi-specific antibody following immobilized Nickel affinity chromatography. (E) SDS-PAGE of purified monomeric 8F4 bi-specific antibody. 2, 1.5, and 0.5 μg of 8F4 bi-specific antibody was electrophoresed in a 10% polyacrylamide gel under reducing conditions, and detected using coomassie die. (F) Western blot of monomeric 8F4 bi-specific antibody electrophoresed in a 10% polyacrylamide gel under reducing conditions probed with anti-HisTag antibody. Representative images from the multiple independent experiments were shown in (B–F).

Figure 2

Flow cytometry analysis of target binding specificity by 8F4 bi-specific antibody. Each data point represents the mean of triplicate measures and error bars represent SEM. (A) Flow cytometry analysis of 8F4 bi-specific antibody binding of CD3⁺ Jurkat and CD3⁻ (J.RT3) cell lines and (B) CD5⁺ and CD5⁻ normal healthy donor peripheral blood lymphocytes in a dose dependent manner. Mean fluorescent intensity is reported. (C–F) Flow cytometry analysis of 8F4 bi-specific antibody binding of different PR1/HLA-A2⁺ (T2 PR1, THP1, U937 A2⁺, K562 A2⁺) and control (T2 CMV, U937 WT, K652 WT) cell lines, detected with anti-HisTag PE. Mean fluorescent intensity is reported. These results indicate target-specific PR1/HLA-A2 and human CD3 cell surface binding. Data combined from 2 independent experiments in triplicate or quadruplicate were shown.

Figure 3

Characterization of 8F4 bi-specific antibody target-specific binding. (A) Bio-layer interferometry analysis of PR1/HLA-A2, HLA-A2/CMV monomers binding to immobilized 8F4 bi-specific antibody and (B) 8F4 bi-specific antibody binding to immobilized CD3εδ fusion protein. The 8F4 bi-specific antibody showed binding affinity to PR1/HLA-A2 and CD3, with no detectable binding to HLA-A2/CMV. (C) Flow cytometry analysis of 8F4 bi-specific antibody and parent 8F4 monoclonal antibody staining of T2 cells pulsed with PR1 peptides containing sequential alanine substitutions, with PR1, CMV-pp65, WT1, MART1 and unpulsed T2 controls. Percent maximum PR1 specific binding is reported. The 8F4 bi-specific antibody showed similar binding pattern to PR1 alanine mutants presented by HLA-A2 compared to parent 8F4 antibody (D) Survey of conformational epitopes of antigen binding domain of 8F4 bi-specific antibody. ELISA detection of 8F4 bi-specific antibody and parent 8F4 antibody (with appropriate negative controls) binding to 8F4 anti-idiotype antibody clones. O.D. 450 is reported. Each data point represents the mean of triplicate measures and error bars represent SEM. One representative data of 3 independent experiments were shown.

Figure 4

T-cell activation and cytokine production by 8F4 bi-specific antibody. Eighteen hour co-culture experiments combining 8F4 bi-specific antibody with healthy donor lymphocyte effectors and target AML cell lines U937 A2+, U937 WT, and THP1 at an E:T ratio of 2:1 were conducted in triplicate. Bi-specific antibodies led to activation of human PBL in the presence of PR1/HLA-A2 in a dose dependent manner, and activated human PBL produced various inflammatory cytokines only in the presence of 8F4 bi-specific antibody and target AML cells. Flow cytometry assessed surface CD69 on activated T-cells in the presence of target cells as compared to control experiments lacking U937 (A) and THP1 (B) AML targets. Cytokines from co-culture experiments were then quantified by ELISA for TNFα (C), IFNγ (D), and IL-6 (E). Each data point represents the mean of triplicate measures and error bars represent SEM. One representative data of 2 independent experiments were shown.

Figure 5

Redirected cytotoxicity of T-cells against PR1/HLA-A2⁺ cell lines. Target AML cells were pre-stained with pacific blue dye and following co-incubation with 8F4 bi-specific antibody and healthy donor PBL at an E:T ratio of 2:1, cells were stained with a fixable live/dead stain. Cytotoxicity calculations were based on total live pacific blue positive cells counts. Calculated % cytotoxicity for co-culture with AML U937 A2⁺ or U937 WT (A) and THP1 (B) cell lines is presented, based on flow cytometry analysis. These data indicate the 8F4 bi-specific antibody initiated dose-dependent AML-specific cytotoxicity after only 18 h co-culture. Each data point represents the mean of triplicate measures and error bars represent SEM. One representative data of 2 independent experiments were shown.

Figure 6

Allogeneic T-cell activation and redirected cell lysis of primary AML patient samples by 8F4 bi-specific antibody. HLA-A2⁺ and HLA-A2⁻ AML patient samples were co-cultured with allogeneic healthy donor PBL at a 2:1 E:T ratio for 18 h in the presence or absence of 8F4 bi-specific antibody. Each bar represents the mean of triplicate measures and error bars represent SEM. (A) Summary of T-cell activation for primary patient sample co-culture experiments show increased T-cell activation following co-culture with HLA-A2⁺ patient samples. (B) A summary of redirected T-cell cytotoxicity for several HLA-A2⁺ and HLA-A2⁻ AML patient samples. These data indicate the 8F4 bi-specific antibody robustly activates allogeneic T-cells following co-culture in a dose and PR1/HLA-A2-specific manner, and these activated T-cells exert their cytotoxicity toward local AML blasts.