Ulocuplumab (BMS-936564 / MDX1338): a fully human anti-CXCR4 antibody induces cell death in chronic lymphocytic leukemia mediated through a reactive oxygen species-dependent pathway

Abstract

The CXCR4 receptor (Chemokine C-X-C motif receptor 4) is highly expressed in different hematological malignancies including chronic lymphocytic leukemia (CLL). The CXCR4 ligand (CXCL12) stimulates CXCR4 promoting cell survival and proliferation, and may contribute to the tropism of leukemia cells towards lymphoid tissues. Therefore, strategies targeting CXCR4 may constitute an effective therapeutic approach for CLL. To address that question, we studied the effect of Ulocuplumab (BMS-936564), a fully human IgG4 anti-CXCR4 antibody, using a stroma--CLL cells co-culture model. We found that Ulocuplumab (BMS-936564) inhibited CXCL12 mediated CXCR4 activation-migration of CLL cells at nanomolar concentrations. This effect was comparable to AMD3100 (Plerixafor--Mozobil), a small molecule CXCR4 inhibitor. However, Ulocuplumab (BMS-936564) but not AMD3100 induced apoptosis in CLL at nanomolar concentrations in the presence or absence of stromal cell support. This pro-apoptotic effect was independent of CLL high-risk prognostic markers, was associated with production of reactive oxygen species and did not require caspase activation. Overall, these findings are evidence that Ulocuplumab (BMS-936564) has biological activity in CLL, highlight the relevance of the CXCR4-CXCL12 pathway as a therapeutic target in CLL, and provide biological rationale for ongoing clinical trials in CLL and other hematological malignancies.

Keywords: BMS-936564; CXCR4; Ulocuplumab; chronic lymphocytic leukemia; reactive oxygen species.

Figure 1. CXCR4 and CXCL12 expression in CLL, normal B and stroma cells

A. Expression of CXCR4 after surface staining using an anti-CXCR4 antibody in B cells derived from CLL patients, healthy volunteers, and stroma-NK-tert cells, as compared to the respective isotype. B. Expression of CXCL12 after intracellular staining using an anti-CXCL12 antibody in B cells derived from CLL patients, and normal PBMCs from healthy volunteers, and stroma-NK-tert cells, as compared to the isotype. C. Panel shows the CXCR4 expression in samples from CLL patients with high risk and low risk characteristics and normal B cells. The line indicates the mean of each group.

Figure 2. Scatchard analysis of Ulocuplumab (BMS-936564) binding to Ramos cells, human PBMCs and ADCC & CDC activity in Ramos cell line (Burkitt's lymphoma)

A. CXCR4 Expression profiling was done using an anti-CXCR4 antibody for surface staining in K562 and B. Ramos cell lines followed by analysis of samples using flow cytometry. The CXCR4 expression is presented in form of ΔMFI. C. The affinity of Ulocuplumab (BMS-936564) was measured using endogenously expressed CXCR4 positive Ramos cells. The binding capacity of ¹²⁵I-Ulocuplumab (BMS-936564) to CXCR4⁺ cells in presence or absence of excess unlabeled antibody was determined and a K_d of 2.9 nM and approximately 1.5×10⁵ receptors/cell was calculated. D. Normal peripheral blood cells were incubated with increasing concentrations of ¹²⁵I-Ulocuplumab (BMS-936564). The counts per minute bound (CPM) in the presence or absence of unlabeled antibody was measured. Scatchard analysis of saturation binding curves was performed with Prism 4 GraphPad software (GraphPad software, San Diego, CA) using nonlinear regression analysis. The symbol , and represent total binding (TB), non-saturable binding (NSB), and saturable binding (SB), respectively. E. Ramos cells (target) were labeled with bis(acetoxymethyl) 2,2′:6′,2″-terpyridine-6,6″-dicarboxylate (BADTA). Freshly isolated human peripheral blood mononuclear cells PBMCs (effector) were used for allotyping of the Ulocuplumab (BMS-936564). Human PBMCs were cultured with labeled Ramos cells at 1:50 ratio of target/effector cell (T/E) ratios in the presence or absence of rituximab, Ulocuplumab, or isotype (BMS-936564) for 1 h at 37°C. Ramos cells alone served as spontaneous release (SR) and Ramos cells lysed with 1% Triton X-100 served as total release (TR). The lysis was measured by using europium (Eu)-based detection. BMS-936564, Rituximab and their respective isotypes with varied concentrations were tested using Ramos cell line. F. Cell-based CDC assay of Ulocuplumab (BMS-936564) /Rituximab and their respective isotypes as controls: lysis of Ramos cells in the presence of human complement was measured by using Alamar Blue release. G. The stable transfected Jurkat cell line expressing FcgRIIIa and NFAT-RE luc was used as effector E. in an ADCC Reporter Bioassay from Promega. CLL cells (target, T) were plated in ratio of 1:1 with the effector cells. The effector: target cells were incubated in the presence or absence of rituximab or Ulocuplumab (BMS-936564), or obinutuzumab between 0.001-30 ug/ml of concentrations for 6 hrs at 37°C. Rituximab and obinutuzumab were used as positive controls. The reaction was developed by incubating the cells with Bio-Glo^TM reagent for 30 minutes at room temperature in dark. The plates were read on luminometer and following the background subtraction, relative-light units (RLU) were calculated for different antibodies/isotypes using GraphPad Prism software. H. Complement dependent cytotoxicity (CDC) for 10 μg/ml of either Ulocuplumab (BMS-936564), rituximab, obinutuzumab or isotype was tested in CLL cells after incubation with either 5% fresh human or heat inactivated serum (to denature complement) was measured by using CD19/CD5/Annexin V staining followed by flow cytomerty analysis. Rituximab and obinutuzumab were used as positive controls. The data are the mean and SD of triplicate cultures. The statistical data was analyzed using Bonferroni correction test in GraphPad Prism software.

Figure 3. % SIA induction by Ulocuplumab (BMS-936564) is CXCR4 specific as observed in Ramos and primary CLL cells but not in K562

A. Ramos and K562 cell lines were treated with different concentrations of Ulocuplumab (BMS-936564) and AMD3100 for 48 hrs followed by running the samples on flow cytometer to study the % SIA. Treatment with AMD3100 in Ramos/K562/CLL did not induce significant % SIA as compared to Ulocuplumab (BMS-936564). In contrast, Ulocuplumab (BMS-936564) induced significant % SIA in Ramos and CLL cells (CXCR4+), but not in K562 (CXCR4-) as evaluated by flow cytometry). B. The CLL cells cultured either alone or with stromal cell support were treated with Ulocuplumab (BMS-936564) or AMD3100 for 48 hrs at 37°C followed by flow cytometer for measurement of % SIA. AMD3100 does not induce significant % SIA in CLL as compared to Ulocuplumab (BMS-936564). Ulocuplumab (BMS-936564) induced significant amount of % SIA in CLL cells alone or CLL co-cultured with stromal cell support.

Figure 4. Ulocuplumab (BMS-936564) antibody induces apoptosis in CLL cells independent of TP53 status

A. The primary leukemia B-cells were obtained from CLL patients with or without TP53 mutation or Del(17p) [TP53mut /Del(17p)]. Normal B and T lymphocytes were obtained from healthy donors. The % SIA was measured in normal B, T cells, CLL cells, TP53mut /Del(17p) CLL cells, alone or co-cultured with stroma-NK-tert cells after treatment with 200 nM of Ulocuplumab (BMS-936564) for 48 hrs of incubation at 37°C. % SIA of Normal B and T cells was significantly less when compared to CLL-LR, CLL-HR and TP53mut /Del(17p) patients, showing that Ulocuplumab (BMS-936564) has a good therapeutic index. Ulocuplumab (BMS-936564) induced apoptosis in CLL cells regardless of TP53mut /Del(17p) mutation status, while normal B or T cells showed statistically-lower level of apoptosis. The data shows the results of samples analyzed in duplicate with the mean and its respective standard deviations. B. Representative CLL samples from each subgroup - TP53wt or TP53mut /Del(17p) and Ramos TP53 mutant were incubated for 48 hrs with the Ulocuplumab (BMS-936564) and Fludarabine (F-ara-A). Apoptosis was measured by flow cytometry.

Figure 5. Inhibition of CXCL12-induced response and migration of primary CLL cells in a transwell assay by Ulocuplumab (BMS-936564) and AMD3100

A. Intracellular F-Actin was measured using FITC-labeled phalloidin in CD19/CD5- pre-labeled CLL cells after stimulation with CXCL12 (90 nM) for 15 seconds. CLL cells from 10 patients (5 HR & 5 LR) were treated with no compounds (control), AMD3100 (4-40 μM) or Ulocuplumab (BMS-936564) (20-2000 nM) prior to stimulation and the representative example of F-actin polymerization is shown. All samples are plotted relative to the mean fluorescence of the sample without any addition of the chemokine CXCL12. B. A total of 5×10⁵ cells per well were plated overnight before the experiment. The cells were incubated with Ulocuplumab (BMS-936564) and AMD3100 (using indicated concentrations) for one hour and loaded on to the transwell chamber and incubated for two hrs in the presence of CXCL12 (12.5 nM) or media control. After that, cells that migrated to the lower chamber were counted using flow cytometry.

Figure 6. Effect of Ulocuplumab (BMS-936564)-mediated apoptosis in CLL cells is caspase-independent

A. Cells were treated for 6 hrs with either 200 nM of Ulocuplumab (BMS-936564) or isotype control (IC). 100 μg of protein was taken from each treatment for caspase analysis. Caspase 2, 3, 8 and 9 were activated upon treatment with the Ulocuplumab (BMS-936564) antibody. There was a significant difference between normal versus CLL cells for caspase 2, 3, 8, and 9. Caspase activity of cells incubated in media only was used as a baseline control. Each value is expressed as mean ± S.D. of two independent experiments. B. CLL cells were incubated with Ulocuplumab (BMS-936564) (200 nM), F-ara-A (10 μM), and Etoposide (30 μM) for 48 hrs either alone or in combination with different concentrations of a pan-caspase inhibitor, Z-VAD-FMK (10, 30, 90 μM). CLL cells viability was analyzed by CD19⁺/CD5⁺/Annexin-V staining followed by flow cytometry. Statistical significance was determined by using Dunnett's multiple comparison test (* p < 0.05; ** p<0.01, *** p < 0.001). Z-VAD inhibited apoptosis in a dose dependent manner when apoptosis was induced by chemotherapy controls but not by Ulocuplumab (BMS-936564) (untreated media control vs. 0-90 μM Z-VAD, p < 0.0001).

Figure 7. Reactive oxygen species (ROS) mediated mechanism of cell death by BMS-936564 in patients' derived CLL cells

A. Primary CLL cells derived from patients were treated for 4 hrs either with 0.1% H₂O₂, 10 ug/ml of obinutuzumab, Rituximab, Ulocuplumab (BMS-936564) antibody or F-ara-A at the indicated concentrations. ROS production and cell death were assessed by flow cytometry. ROS⁺/Annexin-V⁺ double positive cells indicate the population of cell undergoing cell death with concurrent release of ROS (Upper right quadrant). Cells incubated in media only serve as base line control. This experiment represents typical data from three separate experiments conducted with duplicates per each incubation condition. B. The cells were incubated alone with Tiron (30 mM) or 10 μg/ml of either rituximab, obinutuzumab, Ulocuplumab (BMS-936564), 0.1% of H₂O₂ and 10 μM Fludarabine (F-ara-A) with 30 mM of Tiron for 4 hrs at 37°C. To measure ROS production, following CD19/CD5 labeling, cells were labeled with hydroxyethidium dye and incubated for 20 minutes at room temperature followed by flow cytometry analysis. The data presented represent mean ROS production± S.D. C. The cells were incubated alone with Tiron (30 mM) or 10 μg/ml of either rituximab, obinutuzumab, Ulocuplumab (BMS-936564), 0.1% of H₂O₂ and 10 μM Fludarabine (F-ara-A) with 30 mM of Tiron for 4 hrs at 37°C. TheThe percentage (%) cell death was measured using CD19/CD5/Annexin-V labeling followed by flow cytometry analysis. The representative data for ROS and % cell death has been shown from the same patient. The data presented represent mean ROS production± S.D.