Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2011 Apr 21;117(16):4262-72.
doi: 10.1182/blood-2010-07-299248. Epub 2011 Feb 4.

An anti-PR1/HLA-A2 T-cell Receptor-Like Antibody Mediates Complement-Dependent Cytotoxicity Against Acute Myeloid Leukemia Progenitor Cells

Affiliations
Free PMC article

An anti-PR1/HLA-A2 T-cell Receptor-Like Antibody Mediates Complement-Dependent Cytotoxicity Against Acute Myeloid Leukemia Progenitor Cells

Anna Sergeeva et al. Blood. .
Free PMC article

Abstract

PR1 (VLQELNVTV) is a human leukocyte antigen-A2 (HLA-A2)-restricted leukemia-associated peptide from proteinase 3 (P3) and neutrophil elastase (NE) that is recognized by PR1-specific cytotoxic T lymphocytes that contribute to cytogenetic remission of acute myeloid leukemia (AML). We report a novel T-cell receptor (TCR)-like immunoglobulin G2a (IgG2a) antibody (8F4) with high specific binding affinity (dissociation constant [K(D)] = 9.9nM) for a combined epitope of the PR1/HLA-A2 complex. Flow cytometry and confocal microscopy of 8F4-labeled cells showed significantly higher PR1/HLA-A2 expression on AML blasts compared with normal leukocytes (P = .046). 8F4 mediated complement-dependent cytolysis of AML blasts and Lin(-)CD34(+)CD38(-) leukemia stem cells (LSCs) but not normal leukocytes (P < .005). Although PR1 expression was similar on LSCs and hematopoietic stem cells, 8F4 inhibited AML progenitor cell growth, but not normal colony-forming units from healthy donors (P < .05). This study shows that 8F4, a novel TCR-like antibody, binds to a conformational epitope of the PR1/HLA-A2 complex on the cell surface and mediates specific lysis of AML, including LSCs. Therefore, this antibody warrants further study as a novel approach to targeting leukemia-initiating cells in patients with AML.

Figures

Figure 1
Figure 1
8F4 binding to PR1 in the peptide-binding cleft of HLA-A2. Binding was determined by ELISA (A, B, E), flow cytometry (C,F), and surface plasmon resonance (D). (A) 8F4 binding to PR1/HLA-A2 monomer in ELISA depends on PR1/HLA-A2 monomer concentration (blue squares). 8F4 does not bind to control pp65/HLA-A2 monomer, (green squares), PR1 peptide alone (open squares), P3 (black X), or NE (black vertical dashes). In contrast, the BB7.2 antibody, which binds to defined residues of the HLA-A0201 molecule but not to specific peptides within the peptide-binding cleft, binds to both PR1/HLA-A2 (blue vertical dashes) and pp65/HLA-A2 (green X marks). (B) 8F4 binding to PR1/HLA-A2 monomer (blue circles) depends on the 8F4 concentration. 8F4 does not bind to peptide/HLA-A2 monomers containing control peptides WT1 (RMFPNAPYL; red diamonds), Flu (GILGFVFTL; green triangles), or HA-2 (YIGEVLVSV; black triangles). (C) 8F4 binding is dependent on PR1 occupancy of cell-surface HLA-A2 molecules. T2 cells were loaded with increasing concentrations of PR1 (filled bars) or pp65 control peptide (empty bars). Peptide-loaded cells were labeled with 8F4 followed by FITC goat anti–mouse IgG secondary antibody, and fluorescence was measured with FACS. Bars show MFI. (D) Kinetics and affinity of 8F4 mAb binding to the PR1/HLA-A2 complex measured by SPR. PR1/HLA-A2 monomer binds to 8F4 captured on anti–mouse surfaces with a calculated KD = 9.9nM. Measured response units (black and orange lines) show results from a 1:1 interaction model that was used to calculate KD. (E) Sensitivity of 8F4 binding to PR1 and peptide analogs synthesized with Ala substitutions at P1-P9 (ALA1-ALA9). Peptide/HLA-A2 monomers were tested for binding to 8F4 with ELISA. ALA substitution at P1 (ALA1) abrogated 8F4 binding at 50 μg/mL. (F) Epitope mapping shows 8F4 binding to a helical domain of HLA-A2 molecules. T2 cells were loaded with 20 μg/mL of peptide (PR1 or control pp65 peptide) and incubated with A647-conjugated 8F4 in the presence of increasing concentrations of the HLA-A2–specific mAbs W6/32 (left), MA2.1 (center), and BB7.2 (right).
Figure 2
Figure 2
PR1/HLA-A2 visualization on normal and leukemia cells. Visualization was with fluorochrome-conjugated 8F4 in confocal microscopy (A) and flow cytometry (B-C). (A) From top to bottom: T2 cells loaded with PR1, T2 cells loaded with pp65, leukocytes from patient AML2, and leukocytes (PBMCs and granulocytes) from normal donor ND1. Cells were costained with Alexa Fluor 488 (A488)–conjugated anti–HLA-A2 (green; left panels) and Alexa Fluor 647 (A647)–conjugated 8F4 (red) and DAPI (blue; middle panels). Images were viewed with a Leica Microsystems SP2 SE confocal microscope with 10×/25 air, 63×/1.4 oil objectives and Leica Type F immersion oil. Leica LCS software (Version 2.61) was used for image analysis. Scale bar on merged images = 10 μm. (B) PR1 peptide is presented on HLA-A2+ PBMCs from normal donor ND2. Fresh peripheral blood was purified from red cells by lysis, stained with PE-conjugated 8F4 and the phenotype markers CD14, CD3, CD19, CD16, and Live/Dead Aqua viability indicator, and analyzed by flow cytometry (top panel). Scatter profiles and lineage markers identified the indicated cell types. CD14+ monocytes from HLA-A2+ NDs consistently expressed more PR1/HLA-A2 than lymphocytes and granulocytes. Healthy donor ND10 bone marrow cells were labeled with Live/Dead Fixable Aqua, 8F4, HLA-A2, CD34, CD33, CD13, CD14, and a lineage “dump” cocktail composed of Pacific Blue–conjugated CD3, CD7, CD10, CD19, and CD20 (middle panel). Myeloblasts were identified as viable LinCD33+CD34+; monocytes were CD14+. Healthy donor ND10 bone marrow cells were labeled with CD45, CD33, CD11b, CD16, and HLA-A2, and 8F4 (bottom panel). Granulocytes were identified based on scatter characteristics and then examined for expression of CD11b and CD16. Promyelocytes were identified as CD11blo/CD16lo; immature granulocytes were CD11bhi/CD16lo; mature granulocytes stained brightly for both markers CD11b and CD16. (C) Histograms show representative labeling of AML samples (red) and fresh bone marrow cells (blue) with 8F4, mAb directed to lineage markers (the Lin cocktail was CD3, CD4, CD7, CD8, CD10, CD14, CD16, CD19, and CD20, all in Pacific Blue or V450 conjugates), CD38, CD34, and Live/Dead Fixable Aqua (Invitrogen; top panel). For each sample, filled histograms show live cells, and open histograms show LinCD34+CD38 stem cells. Normal LinCD34+CD38 cells show slightly higher 8F4 MFI than total Lin cells; in contrast, LSCs show lower PR1 expression compared with total blasts. Bottom panel combines 8F4 data from 3 different experiments. Each point represents one patient and is the mean value from 1 to 3 independent experiments. MFI for each sample was normalized and presented as a percentage of the MFI of the positive peak of Simply Cellular compensation beads labeled with 8F4. (B-C) The vertical line through the histograms represents background fluorescence established in the fluorescence-minus-one (FMO; gray) labeling control.
Figure 3
Figure 3
8F4 induces cytotoxicity in PR1-presenting cells. (A-D) For CDC, 5 × 104 target cells (A-B: T2 cells loaded with PR1 or control peptide; C-D: primary AML or ND cells) in 10-RPMI/HEPES (N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid) were incubated with 8F4 or control antibody in the presence of complement at 37°C for 90 minutes. Cytotoxicity was assessed with the Cyto Tox-Glo Cytotoxicity Assay (Promega). (A) 8F4-mediated lysis of T2 cells is PR1 specific, requires the presence of complement, and depends on 8F4 antibody concentration. (B) At a constant 8F4 concentration (10 μg/mL), CDC depends on the PR1 concentration. T2 cells were loaded with increasing amounts of PR1 or pp65 control peptide. (C) 8F4 induces CDC of HLA-A2+ cells from AML1 and AML5, but not HLA-A2- cells from AML6 or PBMCs from an HLA-A2+ normal donor (ND4). *P = .0019 AML5 compared with ND4; **P < .0001 AML1 compared with ND4. (D) CDC of leukemia cells from AML1 depends on 8F4 concentration. Mouse IgG2a (isotype control) and pooled human intravenous immunoglobulin were compared at the same concentration as 8F4. (E) 8F4 induces lysis of PR1-loaded T2 cells by ADCC. Target T2 cells were loaded with PR1 or control peptide. Fresh PBMCs from a healthy donor were activated with IL-2 (200 IU/mL) for 2 days and used as effector cells at an E:T ratio of 40:1. Cells were mixed and incubated for 15 hours at 37°C with or without 8F4 or control antibody BB7.2 (50 μg/mL). (A-E) Specific lysis from representative experiments is shown as mean ± SEM from 3 replicates. The negative values for specific lysis are due to background luminescence of cells in the presence of complement likely caused by the enzymatic cleavage of substrate by complement in the absence of an antibody-coated target.
Figure 4
Figure 4
8F4 mediated CDC of both AML blasts and AML stem cells. AML or ND bone marrow cells were incubated with no rabbit complement (C′), C′ only, isotype control antibody, or 8F4 (2.5 μg/mL) for 1 hour at 37°C. Cells were then washed, labeled with lineage-specific mAbs (Lin cocktail was CD3, CD4, CD7, CD8, CD10, CD14, CD16, CD19, and CD20), CD38, CD34, and Live/Dead Fixable Aqua for 30 minutes on ice, and resuspended in 200 μL of 1% PFA in PBS. Before analysis, 50 μL of Caltag Counting Beads was added to each sample for single-platform determination of absolute cell numbers. (A) Representative flow cytometric plots showing scatter distribution and counting beads (top panels) and phenotypes (bottom panels) of AML2 from CDC assay. Beads (FSClo/SSChi gate) were counted, and debris was excluded using the FSChi/SSClo gate. Gates for viable and Lin cells are not shown. LSCs were identified as viable LinCD34+CD38 cells, as shown in the bottom panels. The cytotoxicity of 8F4 increased the bead-to-cell ratio, and CD34+/CD38 LSCs constituted a reduced fraction of the few remaining viable cells in the 8F4-treated samples. (B) 8F4 mediates CDC of HLA-A2+ patient-derived AML blasts (●). Normal HLA-A2+ PBMCs (▵) and bone marrow cells (◊) were not affected (P = .0007). Leukemia cell line U937–transfected HLA-A2 was used as a positive control. Cells from an HLA-A2 AML patient (□) and bone marrow cells from an HLA-A2 normal donor (○) were used as negative controls. (C) Flow cytometric analysis of the same samples gated on live LinCD34+CD38 cells shows preferential CDC of LSCs. HSCs from 3 of 6 healthy donors are also affected by 8F4-mediated CDC. (B-C) CDC assay for each sample and enumeration of cell populations was performed as shown in panel A. For each sample, isotype-mediated background lysis was subtracted from the measured value. Each point is the calculated mean value from 1-3 independent experiments (based on available cells) from individual patients. Error bars show SEM for each group.
Figure 5
Figure 5
8F4 inhibits CFU-L from AML patients, but does not inhibit CFU-L from ALL patients or CFU-E, BFU-E, CFU-GM, and CFU-GEMM from umbilical cord blood and normal bone marrow. (A) CFU-L inhibition by 8F4 from patients AML1, AML7, AML9 (see patient characteristics in Table 1). 8F4 inhibited day-10 CFU-L from AML1 (AML-M1) by 33% compared with the isotype control (P = .004). Similarly, 8F4 inhibited day 10 CFU-L from AML7 (AML-M7) by 44% (P = .03) and AML9 (AML-M1) by 41% (P = .008), respectively. 8F4 did not inhibit CFU-L from patient ALL1. (B) 8F4 had no effect on CFU-E, BFU-E, CFU-GM, or CFU-GEMM from HLA-A2+ umbilical cord blood units. The day-14 CFU-E count was not significantly different between 8F4- and isotype-treated mononuclear cells; results were similar for BFU-E, CFU-GM, and CFU-GEMM. A representative experiment of 3 performed with independent cord blood units is shown. (C) 8F4 had no effect on CFU-E, BFU-E, CFU-GM, or CFU-GEMM from fresh HLA-A2+ normal bone marrow. The day-14 CFU-E count was not significantly different between 8F4- and isotype-treated mononuclear cells; results were similar for BFU-E, CFU-GM, and CFU-GEMM. (A-C) Data represent mean colony counts ± SEM. (D) 8F4 inhibits CFU-L from AML1, but not normal progenitors from HLA-A2+ individual cord blood units (n = 5) in the presence of added rabbit complement. Data represent mean percent colony inhibition calculated for 5 individual cord blood units ± SEM. (A-D) Assays were performed in duplicate.

Comment in

Similar articles

See all similar articles

Cited by 53 articles

See all "Cited by" articles

Publication types

Feedback