CDKN2A-independent role of BMI1 in promoting growth and survival of Ph+ acute lymphoblastic leukemia

Abstract

BMI1 is a key component of the PRC1 (polycomb repressive complex-1) complex required for maintenance of normal and cancer stem cells. Its aberrant expression is detected in chronic myeloid leukemia and Ph+ acute lymphoblastic leukemia (ALL), but no data exist on BMI1 requirement in ALL cells. We show here that BMI1 expression is important for proliferation and survival of Ph+ ALL cells and for leukemogenesis of Ph+ cells in vivo. Levels of BIM, interferon-α (IFNα)-regulated genes and E2F7 were upregulated in BMI1-silenced cells, suggesting that repressing their expression is important for BMI1 biological effects. Consistent with this hypothesis, we found that: (i) downregulation of BIM or E2F7 abrogated apoptosis or rescued, in part, the reduced proliferation and colony formation of BMI1 silenced BV173 cells; (ii) BIM/E2F7 double silencing further enhanced colony formation and in vivo leukemogenesis of BMI1-silenced cells; (iii) overexpression of BIM and E2F7 mimicked the effect of BMI1 silencing in BV173 and SUP-B15 cells; and (iv) treatment with IFNα suppressed proliferation and colony formation of Ph+ ALL cells. These studies indicate that the growth-promoting effects of BMI1 in Ph+ ALL cells depend on suppression of multiple pathways and support the use of IFNα in the therapy of Ph+ ALL.

Fig. 1. BMI1 downregulation inhibits colony formation of Ph+ ALL primary cells

(A) BMI1 levels in ALL# 674 cells transduced with the shBMI1 #1 lentivirus; (B) methylcellulose colony formation (mean + SD of four independent experiments) of untreated or Doxy-treated (2.5μg/ml) parental, shBMI1#1- or shBMI1#2-transduced peripheral blood (n=2) or bone marrow (n=2) derived Ph+ ALL cells. Colonies were counted 10 days after seeding 4 × 10⁴ cells/plate.

Fig. 2. BMI1 silencing enhances imatinib-induced inhibition of cell proliferation and colony formation of BV173 cells

(A) Western blot shows BMI1 expression in imatinib (IM)- or Doxy-treated shBMI1 BV173 cells; (B) cell counts at 168h post treatment or (C) colony formation of untreated shBMI1 BV173 cells or cells treated with Doxy (2.5μg/ml), IM (1μM), or the Doxy/IM combination. Data represent the mean + SD of three independent experiments. Statistical analysis: one way Anova with Bonferroni’s correction. *<0.05.

Fig. 3. BMI1 inhibition suppresses growth of BV173 or SUP-B15 cells in NSG mice

(A, C) Kaplan-Meier plots show survival of NSG mice injected with 2×10⁶ parental, scramble, or shBMI1 silenced (shBMI1 #1 and #2) BV173 (A) or SUP-B15 (C) cells. Mice injected with parental cells were left untreated while those injected with scramble, shBMI1 #1 and #2 BV173 cells or shBMI1 SUP-B15 cells were continuously treated with Doxy in the drinking water (2 g/L) starting three days post-cell injection. (B) Expression of BMI1 in CD10-cALLA+ cells sorted from the bone marrow of 3 leukemic NSG mice injected with shScr BV173 (ShScr) or with shBmi-1 BV173 cells (ShBmi-1). GRB2 expression was used as loading control.

Fig. 4. BIM silencing rescues apoptosis but not the BMI1-mediated inhibition of cell proliferation and colony formation of BV173 cells

(A) Western blot shows expression of BIM in untreated or Doxy-treated (2.5μg/ml; 72 h) shScr or shBMI1-transduced BV173 cells (A); HSP90 expression was examined as loading control. (B) Cell counts, (C) methylcellulose colonies, and (D) % apoptosis (mean+ SD of three independent experiments) of untreated or Doxy-treated shBMI1/ShBIM BV173 cells. Statistical analysis: one way Anova with Bonferroni’s correction. **<0.01, ***<0.001.

Fig. 5. Effect of E2F7 or E2F7/BIM silencing on BMI1-dependent inhibition of proliferation and survival of BV173 cells

Cell counts (A, C) and colony formation (B) (mean SD of three independent experiments) of untreated or doxycycline-treated shBMI1, double-transduced-shBMI1/shE2F7, or triple-transduced shBMI1/shBIM/shE2F7 BV173 cells. Statistical analysis: one way Anova with Bonferroni’s correction. *<0.05, **<0.01, ***<0.001.

Fig. 6. E2F7 and BIM are BMI1 targets in Ph+ ALL primary cells and E2F7/BIM silencing rescues, in part, the leukemogenesis of BMI1-silenced BV173 cells in NSG mice

(A) Western blot shows the increased expression of BIM and E2F7 in Ph+ ALL#1222 patient derived primary cells. βACTIN is shown as loading control. (B) Kaplan-Meier plot shows survival of NSG mice injected with 2×10⁶ untreated or Doxy-treated BIM/E2F7-silenced shBMI1-BV173 cells. In the Doxy-treated group, mice were continuously treated with Doxy in the drinking water (2 g/L) starting three days post-cell injection. Survival curves of Doxy treated mice injected with shBMI1-BV173 cells are shown as control. Statistical analysis: log rank test..

Fig. 7. Overexpression of BIM or E2F7 suppresses colony formation of BV173 cells

(A) Western blots show overexpression of BIM or E2F7 in pUltra-BIM- or pUltra-E2F7-transduced BV173 cells. βACTIN expression is shown as loading control; (B, C) Colony formation of pUltra-BIM (B) or pUltra-E2F7 (C)-transduced BV173 cells. Data represent the mean + SD of two independent experiments. Statistical analysis: 2-tailed Student t test. ***<0.001

Fig. 8. Effect of IFNα- and/or imatinib or dasatinib on colony formation of Ph+ ALL

Methylcellulose colony formation of BV173 (mean+ SD of three independent experiments) or primary bone marrow Ph+ ALL cells (mean+ SD of triplicate plates) treated with IFNα and/or imatinib (A and B, respectively) or with the IFNα and/or dasatinib (C). In (B), four samples were treated with IFNα only while three samples were treated with imatinib alone or the IFNα/imatinib combination. Results are expressed as % inhibition of colony formation of untreated vs treated cells. Statistical analysis: one way Anova with Bonferroni’s correction. **<0.01, ***<0.001.