BMI1 polycomb group protein acts as a master switch for growth and death of tumor cells: regulates TCF4-transcriptional factor-induced BCL2 signaling

Abstract

For advanced prostate cancer (CaP), the progression of tumors to the state of chemoresistance and paucity of knowledge about the mechanism of chemoresistance are major stumbling blocks in the management of this disease. Here, we provide compelling evidence that BMI1 polycomb group protein and a stem cell factor plays a crucial role in determining the fate of tumors vis-à-vis chemotherapy. We show that progressive increase in the levels of BMI1 occurs during the progression of CaP disease in humans. We show that BMI1-rich tumor cells are non-responsive to chemotherapy whereas BMI1-silenced tumor cells are responsive to therapy. By employing microarray, ChIP, immunoblot and Luciferase reporter assays, we identified a unique mechanism through which BMI1 rescues tumor cells from chemotherapy. We found that BMI1 regulates (i) activity of TCF4 transcriptional factor and (ii) binding of TCF4 to the promoter region of anti-apoptotic BCL2 gene. Notably, an increased TCF4 occupancy on BCL2 gene was observed in prostatic tissues exhibiting high BMI1 levels. Using tumor cells other than CaP, we also showed that regulation of TCF4-mediated BCL2 by BMI1 is universal. It is noteworthy that forced expression of BMI1 was observed to drive normal cells to hyperproliferative mode. We show that targeting BMI1 improves the outcome of docetaxel therapy in animal models bearing chemoresistant prostatic tumors. We suggest that BMI1 could be exploited as a potential molecular target for therapeutics to treat chemoresistant tumors.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Figure 1. BMI1 protein levels are increased during the progression of CaP disease in human patients and BMI1 induces CaP cell proliferation.

(A) Immunoblot represents BMI1 protein levels in normal, dysplasia and tumor prostatic tissues as assessed by immunoblotting (B) in representative photomicrographs point to BMI1-positive immunostaining in neoplastic and non-neoplastic regions of prostatic specimens. Magnification 40X. (C) Immunoblot represents the effect of androgen on BMI1 expression in cells assessed by immunoblotting. (Di–Dii; Ei–Eii and Fi–Fii) Histograms represent the growth, rate of proliferation and clonogenic proliferation of BMI1-silenced and -overexpressed CaP cells measured by MTT, 3[H]thymidine uptake and soft-agar colony formation assays. Each histogram represents mean ± S.E. of three independent experiments, * indicates p<0.05. Equal loading of protein for immunoblotting was confirmed by ß-actin.

Figure 2. BMI1 induces growth of normal primary prostate cells (PrEC) by abolishing senescence and regulates the expression of proliferation-associated genes in CaP cells.

(A–B) While PrEC replicated for 5 passages and entered into senescence, BMI1-rich counterparts replicated and avoided senescence upto 8th passages. (A) Inset 400X of magnified areas show senescent morphology features i.e. globular shape and (B) indicate staining for ß-galactosidase. (C) Scattered Plot for qPCR array. The dots indicate gene expression on a log-scale representing the change in BMI1 silenced-LNCaP cells. Fold change (2∧- Delta Ct) is the normalized expression (2∧-Delta Ct) in the BMI-silenced cells divided by the normalized expression of Control. (D and E) Immunoblots represents the effect of BMI1-silencing and BMI1-overexpression on the expression of Cyclin-D1, BCL2 and p16 proteins in cells. The equal loading of protein was confirmed by ß-actin.

Figure 3. BMI1 confers chemoresistance to tumor cells.

Rate of proliferation and apoptosis in cells were determined by 3[H]thymidine uptake and flow cytometery respectively. (A–D) Histograms represents the rate of proliferation in (A–B) LNCaP and (C–D) PC3 cells harboring varied BMI1 levels and treated with chemotherapeutic agents. Vehicle treated cells served as control. (Ei and Fi) immunoblots represent the levels of BMI1, Cyclin-D1 and BCL2 proteins in docetaxel-resistant, and BMI1-silenced docetaxel-resistant cells. (Eii and Fii) Histograms showing the rate of proliferation in docetaxel-resistant, and BMI1-silenced docetaxel-resistant cells. For immunoblot analyses (Figure Ei and Fi), equal loading of proteins was confirmed by ß-actin. (A–D, Eii and Fii) Each bar represents mean ± SE of three independent experiments, * represents P<0.05. (G–H) represents quantitative estimation of apoptosis in BMI1-silencing chemoresistant cells. The lower right quadrant of the FL1/FL2 plot (Annexin V-FITC) represent early apoptosis and the upper right quadrant (labeled with AnnexinV-FITC and PI) represent late apoptosis.

Figure 4. BMI1 regulates BCL2 expression through activation of TCF-transcriptional activity in tumor cells.

(A–B; C–D) Histograms represent the effect of BMI1-overexpression and BM1-silencing on the transcriptional activation of TCF-responsive element in CaP and HT29 cells cells as assessed by luciferase-reporter assays. (E–F) representative immunoblots showing the effect of BMI1-silencing and -overexpression on the levels of BCL2 and Cyclin-D1 proteins in (E) HT29 cells, and (Fi–Fii) CaP cells treated with Cyclopamine (Shh inhibitor) for 12 h. Control cells were treated with DMSO. (G–H) Histogram represents the effect of (G) cyclopamine treatment and (H) TCF silencing on the transcriptional activity of BCL2 promoter in LNCaP, PC3 and HT29 cells. (A–D; G–H), relative luciferase activities were calculated with the values from vector group, and each bar represents mean ± SE of three independent experiments, *represents p<0.05. (E–F) Equal loading of proteins was confirmed by testing immunoblots for ß-actin.

Figure 5. (A–B) BMI1 induces TCF4 binding to promoter region of BCL2 gene. (C) TCF4 occupancy on BCL2 is elevated in malignant prostatic tissues. (D) BMI1 confers chemoresistance to tumors in a mouse model.

(A–B) Histogram represents effect of BMI1 expression on TCF4-occupancy on promoter regions of BCL2 in PC-3 and HT29 cells as assessed by ChIP assay. (C–D) immunoblot and histogram represents the BMI1 protein expression, and TCF4-occupancy on BCl2 gene in normal and malignant human prostate tissues as assessed by immunoblotting and ChIP assays. Equal loading of proteins was confirmed by ß-actin for immunoblotting. (A–B, D). Each bar represents mean ± SE of three independent experiments. (E–F) The line graph represents average volume of BMI1-overexpressing and BMI1-suppressed tumors as a function of time vis-à-vis docetaxel therapy on in nude mice. (G–H) The line graph shows the number of mice with tumor volumes <1000 mm3 for indicated weeks. Data is represented as mean±SE; * indicates p<0.05.