Ph⁺/VE-cadherin⁺ leukemic cells form hematopoietic colonies/cords on bone marrow niche stromal cells and differentiate into endothelial cells in vitro. (A) Phase-contrast micrographs of Ph⁺ ALL Sup-B15 (top) and Nalm27 (bottom) cells cultured in medium alone (long-term medium culture, LTMC) or long-term cocultured (LTCC) on murine stromal cell line S10 (top and bottom right). Images were collected by using a Leica DMIL microscope and 10×/0.22 objective (Houston, TX). (B) Confocal LSM microphotographs of hematopoietic cords formed on stromal cells. Cells were surface stained with anti-CD34, CD19, VE-cadherin, and CD133, respectively. Nucleic DNA dyes include Sytox (green), PI (red), and TO-PRO-3 (blue). Z stacks were 3-D reconstructed with LSM510 software (version 3.2; Zeiss, Jena, Germany). The turning axis is the y-axis, and turning angle for the displayed static pictures is 45° and confocal images were obtained using 20×/0.75 Fluar objective (rows 1, 3) and 10×/0.50 Fluar objective (rows 2, 4). (C) Confocal micrographs of hematopoietic colonies (“hemospheres”) stained with a panel of early stem cell (ESC) markers. Stromal monolayer cells underneath hemospheres were counterstained with anti–VCAM-1. Cell nuclei were counterstained with DNA dyes PI (red fluorescence) or SYTOX Green (green fluorescence) to be compatible with the fluorochrome tag of the secondary antibodies of anti-ESC antibody. Merged images of 3-channel colors are shown in the final column and images were acquired using a 10×/0.5 Fluar objective. (D) Three-dimensional reconstruction of z-stacks showing the distribution of transcriptional factor Oct-4 within a hematopoietic cord. Cell nuclei were counterstained with PI (red), and localization of Oct-4 was labeled with green fluorescence. The turning axis is the y-axis. The tip of the hematopoietic cord was pressed flat by the cover slip and images were acquired using a 10×/0.5 Fluar objective. (E) DsRed red fluorescent protein-–labeled Sup-B15 cells (top left) or eGFP-labeled Nalm27 cells (bottom left) were cultured in standard medium or in EndoCult endothelial-defined medium (top and bottom right). White arrows denote individual endothelial sprouting observed for individual Ph⁺/VE-cadherin⁺ ALL cells.

Stromal cells up-regulate VE-cadherin expression and stabilize β-catenin in Ph⁺/VE-cadherin⁺ leukemic cells. (A) Western blot analysis of Sup-B15 cells in LTMC (Med) or in LTCC with either S10 (S10) or PatX (PatX) stromal cells. The same membrane was probed with anti–VE-cadherin, anti–β-catenin, and anti-Src, respectively. GAPDH was used as the lane loading control. (B) Western blot of cell lysates isolated from Sup-B15 cells exposed to recombinant human cytokines, chemokines, or growth factors compared with untreated control cells (Cont). (C) Reciprocal coimmunoprecipitation of Sup-B15 cell lysates from LTMC (Med) or LTCC (S10 and PatX) with specific antibodies for anti–VE-cadherin or anti–β-catenin. Membranes were reciprocally probed with anti–β-catenin and anti–VE-cadherin, respectively. IgG heavy chains served as the loading control. Input denotes sample with antibody alone, lacking cell lysate. (D) Western blot analysis of cell lysates from 293FT (top) or Sup-B15 cells (bottom) transduced with lentiviruses carrying the wild-type VE-cadherin (WT) or VE-cadherin lacking either the cytoplasmic domain (Δcyto) or β-catenin binding domain (Δbcat) CDSs (coding DNA sequences). Vect denotes the empty lentiviral vector encoding the blasticidin resistance protein without a CDS insert. (E) Western blot analysis of Sup-B15 cells transiently transfected with VE-cadherin siRNA or scrambled (Scr) control siRNA sequence. Time-course (top) and dose-response (bottom) of VE-cadherin siRNA-transfected samples were probed with specific antibodies for anti–VE-cadherin and β-catenin.

Bcr-abl fusion protein is essential for maintaining the VE-cadherin/β-catenin axis in Ph⁺ ALL. (A top) Western blot analysis of Bcr-abl expression by Sup-B15 in LTMC (Med) or LTCC (S10 or PatX). (Bottom) DLR assays of LTMC (Med) and LTCC (S10 and PatX) Sup-B15 cells transiently cotransfected with the Bcr-abl promoter reporters phRL-pp210L or phRL-pp210S with an internal control pGL4.13-CMV-Luc2 firefly luciferase reporter. Data were shown as mean plus or minus SD (n = 3). The * denotes statistically significant differences compared with values from cells in media alone (P < .01) based on the Student t test. (B) Western blot analysis of Bcr-abl expression by Sup-B15 cells transduced with retroviruses containing p185-eGFP or p210-eGFP forms of Bcr-abl, or the empty vector control (MSCV-IRES-eGFP, MiG) alone (top Western blot). Dual-color flow cytometric analysis of VE-cadherin expression in Sup-B15 cells overexpressing the exogenous Bcr-abl fusion proteins or empty retroviral control vector (bottom flow cytometry). Baseline VE-cadherin was set to zero for comparison of the net increases after retroviral transduction. (C) DLR assay of Sup-B15 cells infected with p185-eGFP, p210-eGFP or MiG vector control retroviruses. Data are shown as mean plus or minus SD (n = 3). The * denotes significant differences compared with MSCV empty vector controls (P < .01) based on the Student t test. (D) Western blot analysis of cell lysates from Sup-B15 cells treated with the Bcr-abl kinase inhibitors Imatinib Mesylate or AG957 at various concentrations. (E) Western blot of Sup-B15 or Nalm27 leukemic cells treated with 20 μM imatinib mesylate, 10 μM proteosome inhibitor MG132, or a combination of both inhibitors for 16 hours.

Identification and characterization of a leukemic stem cell–like population that expresses both endothelial and hematopoietic progenitor cell surface markers. (A) Confocal laser scanning microphotographs (LSMs) of Ph⁺ ALL Sup-B15 cells stained with anti–VE-cadherin in the presence of 10 mM EGTA and 5 mM CaCl₂. Cell nuclei were counterstained with propidium iodide (PI) and 1024 × 1024 12-bit confocal images were collected using 40×/0.75 Plan -Neofluar objective. (B) Cell lysates prepared from Ph⁻ acute leukemia cell lines JM1, RS4;11, REH, Jurkat, HL60, and Ph⁺ acute leukemia cell lines K562, Sup-B15, OP-1, Nalm20, Nalm27, and Nalm29 were Western blotted with anti–VE-cadherin. The same membranes were stripped and reprobed with anti–β-tubulin as a loading control. (C) Immunophenotyping of Ph⁺/VE-cadherin⁺ Sup-B15 cells by flow cytometry. Cells were surface stained to evaluate hematopoietic stem/progenitor cell and classic endothelial markers.

LTCC with bone marrow stromal cells promotes self-renewal activity of Ph⁺/VE-cadherin⁺ leukemic cells independent of the Wnt/β-catenin signaling pathway. Clustering analysis of gene expression profiles in Sup-B15 cells after LTCC with S10 or PatX stromal cells. The hierarchical clusters were created with GEASuite software based on the similarity of gene expression. The green color at the farthest left end of the color scale corresponds to the minimal value; the red color at the farthest right end of the color scale corresponds to the maximum value, black corresponds to the average value of expression. BAS2C, HSPCB, ACTB, and GAPDH served as housekeeping gene controls. Analysis of the representative Wnt pathway is shown.

β-Catenin is constitutively activated in leukemic stem-like cells overexpressing VE-cadherin during LTCC. (A) Dual-luciferase reporter (DLR) assays of Sup-B15 LTMC (Med) or LTCC cells (S10 or PatX) transiently cotransfected with phRL-TOP1 or phRL-TOP2 Renilla luciferase reporters with an internal control pGL4.13-CMV-Luc2 firefly luciferase reporter. Experimental Renilla luciferase activity was normalized to that of control firefly luciferase and is shown as relative luciferase units (RLUs). Data were presented as mean plus or minus SD (n = 3). (B) DLR assays of cell lysates from Sup-B15 cells transduced with wild-type or the 2 truncated forms of VE-cadherin. Data were presented as mean plus or minus SD (n = 3). The * denotes significant differences compared with Vect controls (P < .01) based on the Student t test. (C) Sup-B15 cells transfected with 100 nM VE-cadherin siRNA for 8 to 72 hours (top) or at a concentration of 50 to 200 nM for 24 hours (bottom). Data are shown as the mean plus or minus SD (n = 3). (D) Confocal micrographs of LTMC or LTCC Sup-B15 cells stained with anti-active β-catenin antibody and were acquired using a 63×/1.2 water C-Apochromat objective. Cell nuclei were counterstained with SYTOX Green. The photograph of LTMC was made from cytospin preparation, and the photograph of LTCC Sup-B15 cells was taken from a portion of 1 hematopoietic cord.

Tyrosine phosphorylation of β-catenin mediated by Bcr-abl kinase diminishes β-catenin ubiquitination-proteosome–dependent degradation. (A) Western blot analysis of cell lysates from 293FT cells cotransduced with lentiviral vector (Lenti-Vect), lentiviruses carrying WT VE-cadherin (Lenti-VE-cad), retroviral vector (Retro-MiG), or retroviruses encoding p210-eGFP (Retro-MiG-p210). β-Catenin was densitometrically normalized to GAPDH with the control value set at 1.0, and all other values are shown relative to the control. (B) Immunoprecipitation of β-catenin was done using the above-mentioned cell lysates, and Western blots were probed with anti–P-Tyr (Phospho-Tyr102), anti–β-catenin, anti–phospho-Src, and anti-Src antibodies. IgG heavy chain served as the loading control. (C) DLR assays from the above 293FT cells transiently cotransfected with the phRL-TOP1 and phRL-TOP2 reporters and an internal pGL4.13-CMV-Luc2 firefly luciferase reporter. Data are shown as mean plus or minus SD (n = 3). The * denotes significant differences compared with vector controls (P < .01) based on the Student t test. (D) Western blot analysis of β-catenin stabilization from 293FT cells coexpressing VE-cadherin/eGFP-p210 or VE-cadherin/eGFP treated with 10 μM Src kinase inhibitor PP2, 10 μM Bcr-abl inhibitor imatinib mesylate, or both inhibitors for 16 hours. β-Catenin was densitometrically normalized to VE-cadherin with the untreated control value of expressing MiG vector alone set to 1.0. All other ratios are normalized to this control value. (E) VE-cadherin was immunoprecipitated from 293FT cells coexpressing VE-cadherin/eGFP-p210 or VE-cadherin/eGFP. The immunoprecipitates and cell lysates were loaded in parallel and probed with anti–phospho-Tyrosine (4G10) and VE-cadherin, c-Abl and eGFP antibodies. (F) β-Catenin immunoprecipitated either from 293FT expressing eGFP-210 or eGFP alone, or from Ph⁺ Sup-B15 or Ph⁻ REH leukemic cells was in vitro labeled with recombinant ubiquitin and subjected to in vitro degradation assay as described in “In vitro poly-ubiquitination of β-catenin.”