BCL6 is required for a basic level of Imatinib-resistance in CML cells. (A) BCL6^−/− and BCL6^+/+ CML-like cells were treated with Imatinib at various concentrations for 3 d, and cell metabolism was measured in a Resazurin assay. Mean values ± SD of three experiments are depicted. (B) BCL6^+/+ and BCL6^−/− CML cells were treated with Imatinib as in A, and apoptosis was assessed by flow cytometry analysis of 7AAD and annexin V. Mean values ± SD of three experiments are depicted. (C) BCL6^−/− CML-like cells were transduced with 4-OHT–inducible BCL6 (BCL6-ER^T2/GFP) or an ER^T2/GFP empty vector control. BCL6^−/− CML-like cells were then treated with 1 μmol/liter Imatinib for the times indicated and in the presence of 4-OHT–mediated induction of BCL6-ER^T2 or ER^T2. Percentage of GFP⁺ cells were measured by flow cytometry as an indication of a BCL6-ER^T2– or ER^T2-mediated survival advantage. A time course of mean values ± SD of three experiments is depicted in C, and examples of the flow cytometry plots are shown in D. (E) Human CML cell lines (KCL22, JURL, LAMA84, KYO1, and KU812) were incubated in the presence or absence of 1 µmol/liter Imatinib, 5 µmol/liter RI-BPI, or a combination of both for 3 d. Viability was measured by flow cytometry. Mean values, SD, and p-values from three experiments are indicated.

BCL6-mediated transcriptional repression of p53 enables colony formation and proliferation of CML cells. p53^+/+ and p53^−/− CML-like cells were transduced with a 4-OHT–inducible dominant-negative BCL6 (DN-BCL6-ER^T2) or an empty vector control and 100,000 cells each were plated in semisolid agar (A). Photomicrographs of methylcellulose plates and statistical analysis are shown. (B) p53^+/+ and p53^−/− CML-like cells transduced with a 4-OHT–inducible dominant-negative BCL6 (DN-BCL6-ER^T2) or an empty vector control (ER^T2) and cell cycle was analyzed by flow cytometry (BrdU and 7AAD staining). FACS plots and statistical analysis are shown. (C and D) p53^+/+ and p53^−/− hematopoietic progenitor cells were transformed with BCR-ABL1 to generate CML-like leukemia and then transduced with DN-BCL6-ER^T2 or a GFP empty control vector. CML-like cells expressing GFP, were incubated in vitro and relative changes of GFP⁺ percentages were plotted against time (days) after 4-OHT–mediated induction (C). In D, one example of flow cytometry measurements over 9 d is shown. Mean values of three experiments ± SD are indicated.

BCL6-inhibition results in depletion and cell cycle arrest of CD34⁺ CD38⁻ CML cells. Patient-derived leukapheresis samples from 5 patients with CML in CP and 1 patient in myeloid blast crisis CML were incubated in the presence of RI-BPI or vehicle for 2 h, and then washed and incubated with cytokines (100 ng/ml SCF, 10 ng/ml G-CSF, 20 ng/ml FLT3, 20 ng/ml IL-3, and 20 ng/ml IL-6). (A) Cells were stained with CD34 and CD38 cell surface antibodies and analyzed by FACS. (B) Cells were stained with CFSE to track cell divisions over time. The analysis was gated on CD34⁺ CD38⁻ LICs. (C) Synopsis of the effect of 2 h RI-BPI treatment on CD34⁺ CD38⁻ LICs (top) and CD34⁻ transient amplifying CML cells (bottom) in 5 cases of CML-CP (green lines) and 1 case of CML-BC (red line).

Regulation of BCL6 expression in CML cells. (A) To identify TKI-regulated genes in human CML cells, three human CML cell lines (KCL22, KU812, and JURL) were treated with (IM) or without (Ctrl) 1 μmol/liter Imatinib for 16 h and studied in an Affymetrix GeneChip analysis. Genes were sorted based on gene expression differences between TKI-treated and untreated CML cells. Likewise, BCR-ABL1–transformed leukemia cells from Stat5^fl/fl bone marrow were transduced with Cre or an empty vector control. Gene expression values for Stat5^fl/fl (Ctrl) and Stat5-deleted (Cre) leukemia cells are indicated for the genes identified in TKI-treated cells. (B) Affymetrix GeneChip data for BCL6 was validated by quantitative RT-PCR on three cases of human CML (with or without 1 μmol/liter Imatinib overnight). (C) BCL6 gene expression values for CML cells from six patients before and after 7 d of Imatinib-treatment are shown (meta-analysis of data from Bruennert et al. [2009]). (D) Human CML cells were treated with TKI (10 µM for 24 h) and BCL6 expression was evaluated by Western blotting using β-actin as a loading control. (E and F) Leukapheresis samples from two patients with CML-CP (CP21 and CP22) were sorted into four subpopulations based on CD34 and CD38 surface expression as depicted in the flow cytometry dot plots. Subpopulations were incubated overnight in the presence and absence of 10 µmol/liter Imatinib and then subjected to quantitative RT-PCR for BCL6 mRNA levels using COX6B as a reference gene. For each subpopulation, fold-induction of BCL6 mRNA levels are shown (triplicate measurements were performed; *, P < 0.05; **, P < 0.01). (G) Pten^fl/fl mouse CML cells were transduced with 4-OHT–inducible Cre-ERT2 or ER^T2 empty vectors. BCL6 protein levels are shown after treatment with 4-OHT in the presence or absence of Imatinib. Pten deletion in Cre-ER^T2–transduced cells was verified by Western blot. (H) Human CML cells (KCL22) were transduced with a retroviral vector encoding a constitutive active form of FoxO3A (FoxO3a^CA CD90) or an empty control vector (CD90 EV). CD90⁺ cells were sorted (sort gate indicated) and studied for BCL6 mRNA levels by quantitative RT-PCR using COX6B as a reference gene. Mean values of three experiments ± SD and p-value are indicated.

BCL6 is required for the maintenance of LSK⁺ cells in CML. (A and B) BCR-ABL1 transformed CML-like cells from BCL6^+/+ and BCL6^−/− bone marrow at 1 and 3 wk after transduction. Cells were gated on Lin⁻ phenotype and surface expression of Sca-1 and c-Kit (LSK; A) and CD44 and c-kit (B) is shown (n = 3). (C) Human CML cells (JURL cell line) were subjected to one round of ChIP-seq analysis for a genome-wide mapping analysis of recruitment of the BCL6 transcription factor. Overlays of input (green) and BCL6 ChIP (red) are shown for BCL6 (positive control; binding to its own promoter), HPRT (negative control), DNA damage response and cell cycle checkpoint genes including CHEK, CDKN2AIP, TP53, CDKN1A (p21), GADD45A, and CDKN2A (Arf). Peaks of significant enrichment of BCL6 in promoter regions relative to input were identified by ChIPSeeqer (black bars). (D) BCR-ABL1–transformed CML-like cells from BCL6^+/+ and BCL6^−/− bone marrow were analyzed by Western blot for Arf and p53 protein levels using β-actin as loading control (two experiments are shown). (E) 100,000 BCL6^+/+ and BCL6^−/− CML-like cells were plated in semisolid methylcellulose agar and colonies were counted after 22 d. Chart shows mean values ± SD and p-value of 5 experiments. (F) Cell cycle analysis of BCL6^+/+ and BCL6^−/− CML cells was performed studying BrdU incorporation in combination with 7AAD staining. Annotations indicate distribution of CML-like cells to G0/1, S, and G2/M phases of the cell cycle. One representative experiment of three is shown.

BCL6 is required for self-renewal and leukemia initiation in CML. (A) BCL6^−/− and BCL6^+/+ CML-like cells were labeled with firefly luciferase and 500,000 cells were injected into sublethally irradiated NOD/SCID mice (7 mice per group; two independent experiments). Engraftment and leukemic growth was measured by luciferase bioimaging at the times indicated. (B and C) Human CML cells (KCL22) were transduced with a 4-hydroxy-tamoxifen (4-OHT)-inducible dominant-negative mutant of BCL6 (DN-BCL6-ER^T2; Shaffer et al., 2000) or an ER^T2 empty vector control (B). Human CML cells were labeled with lentiviral firefly luciferase and 3 × 10⁶ cells were injected into sublethally irradiated NOD/SCID mice (7 mice per group). NOD/SCID recipients were treated 30 times (3 sets of 10 injections) with tamoxifen (20 mg/kg) after leukemia cell were injected (second set of treatment was performed after 34 d, third set of treatment was performed after 55 d). Leukemia burden (B) and overall survival (Kaplan-Meier analysis; C) are shown (two independent experiments). (D) 10,000 cells from three CML cases were plated in semisolid methylcellulose agar with or without RI-BPI (5 µmol/liter). Colonies were counted after 10 d. The chart shows mean values ± SD and p-value of five experiments. (E) 3 × 10⁶ human CML cells (KCL22) labeled with firefly luciferase were treated with or without 5 µmol/liter RI-BPI ex vivo and injected into sublethally irradiated NOD/SCID mice. Overall survival, engraftment, and leukemia progression were monitored by Kaplan-Meier analysis (E) and luciferase bioimaging (F), respectively. Two separate injections and treatments were performed and 10 mice per group were studied.