Amplification-driven BCL6-suppressed cytostasis is mediated by transrepression of FOXO3 and post-translational modifications of FOXO3 in urinary bladder urothelial carcinoma

Muscle-invasive urinary bladder urothelial carcinoma (UBUC) is a lethal disease for which effective prognostic markers and potential therapy targets are still lacking. Previous array comparative genomic hybridization identified that 3q27 is frequently amplified in muscle-invasive UBUCs, one candidate proto-oncogene, B-cell CLL/lymphoma 6 (BCL6), mapped to this region. We therefore aimed to explore its downstream targets and physiological roles in UBUC progression. Methods: Specimens from UBUC patients, NOD/SCID mice and several UBUC-derived cell lines were used to perform quantitative RT-PCR, fluorescence in situ hybridization immunohistochemistry, xenograft, gene stable overexpression/knockdown and a series of in vitro experiments. Results: Amplification of the BCL6 gene lead to upregulation of BCL6 mRNA and protein levels in a substantial set of advanced UBUCs. High BCL6 protein level significantly predicted poor disease-specific and metastasis-free survivals. Knockdown of the BCL6 gene in J82 cells inhibited tumor growth and enhanced apoptosis in the NOD/SCID xenograft model. In vitro experiments demonstrated that BCL6 inhibited cytostasis, induced cell migration, invasion along with alteration of the expression levels of several related regulators. At molecular level, BCL6 inhibited forkhead box O3 (FOXO3) transcription, subsequent translation and upregulation of phosphorylated/inactive FOXO3 through phosphoinositide 3-kinase (PI3K)/AKT serine/threonine kinase (AKT) and/or epidermal growth factor receptor (EGFR)/mitogen-activated protein kinase 1/2 (MAP2K1/2) signaling pathway(s). Two BCL6 binding sites on the proximal promoter region of the FOXO3 gene were confirmed. Conclusion: Overexpression of BCL6 served a poor prognostic factor in UBUC patients. In vivo and in vitro studies suggested that BCL6 functions as an oncogene through direct transrepression of the FOXO3 gene, downregulation and phosphorylation of the FOXO3 protein.

Lentivirus production and stable knockdown of the BCL6 and FOXO3 genes Clones were obtained from the National RNAi Core Facility, Institute of Molecular Biology, Academia Sinica, Taipei, Taiwan. A total of 4 plasmids targeting the BLC6 gene were preliminarily screened. The BCL6 mRNA levels could be effectively downregulated by only 2 clones. Another two shFOXO3 clones were also identified. Briefly, Phoenix-AMPHO cells (ATCC) were seeded in 6-cm tissue culture plate at a density of 3  10 6 in 5 mL medium with 10% FBS, 100 IU/mL penicillin and 100 g/mL streptomycin (Corning, Corning, NY, USA) overnight. PolyJet (15 L, #SL100688, SignaGen Laboratories, Rockville, MD, USA) was used to transfect the plasmid mixture [psPAX2 (2.25 g, Addgene), PMD2.G (0.25 g, Addgene) and 2.5 g of shLacZ (control), shBCL6#1 and shBCL6#2 plasmids], and the medium was changed after a 16 h incubation. Medium was collected and filtered (0.22 m) at 40 h and 64 h post-transfection, aliquots of 1 mL were stored at -80C for further infection. For stable shRNAi, lentiviral particles were produced and 1  10 6 BFTC905 and BFTC909 cells were next transduced with media containing lentiviral particles containing polybrene (8 g/mL) and incubated for another 24 h at 37C. Afterwards, media containing 4 g/mL puromycin were used to select positive cells for 3 days and subsequently maintained in media containing 2 g/mL puromycin for further experiments. The same protocol was performed in double knockdown of the BCL6 and FOXO3 genes in BFTC905 cells.

Immunoblot analysis
Cell lysates were prepared with RadioImmunoPrecipitation Assay buffer (Merck Millipore, Burlington, MA, USA). Lysates containing equal amounts of protein were separated by SDS-PAGE and electroblotted onto FluoroTrans PVDF Transfer Membrane (PALL). The filters were individually probed with specific primary antibody. Protein bands were detected by the Western Lightning Chemiluminescence Reagent Plus Kit (Perkin-Elmer Life Sciences, Cleveland, OH, USA) with horseradish peroxide labeled secondary antibody as suggested by the manufacturer and visualized on a ChemiDoc XRS+ System (Bio-Rad, Hercules, CA, USA). The intensity of bands was quantified by densitometry (ImageJ) and normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or pan-actin in each lane.

Cell-cycle, 5-bromo-2'-deoxyuridine, cell proliferation, soft agar/anchorage-independent growth
For cell cycle analysis, 1  10 6 cells were collected, washed with ice-cold PBS, fixed with 70% ethanol and stored at -20C after stable transfection of pCMV6-Entry, pCMV6-BCL6 plasmid, or infection with shBCL6#1, shBCL6#2 or shLacZ lentiviral particles. Before analysis, fixed cells were washed with ice-cold PBS for three times and treatments with 200 g/mL RNase A and 20 g/mL propidium iodide (PI). A total of 10,000 events were analyzed; cell cycle distribution was analyzed by a Beckman Coulter Epics XL Flow Cytometer and the Modfit LT software (BD Biosciences, Franklin Lakes, NJ, USA) (Kuo et al. 2011;Toxicol Appl Pharmacol 256:8-23).
To determine cell proliferation upon alternation of BCL6 expression levels, 3  10 3 cells were seeded on 96-well microplates for 5-bromo-2'-deoxyuridine (BrdU) or cell proliferation (fluorometric) (#K307-1000, Biovision, Milpitas, CA, USA) assays. After removing the medium, BrdU Cell Proliferation Assay Kit (QIA58, Merck Millipore) was used to perform cell proliferation test. BrdU label (1:2000 dilution) was incubated for 24 h. Plates were then washed, stained with anti-BrdU antibody, and peroxidase-conjugated goat anti-mouse IgG. 3,3',5,5'-tetramethylbenzidine substrate (0.1 mL in ethanol) was next added into the immunocomplex and the reaction was terminated via adding 100 L of sulfuric acid (2.5 N). Absorbances were afterward measured at wavelengths 450/540 nm using a Beckman Coulter PARADIGM™ Detection Platform. Percentages of proliferation rate (%) were calculated as 100  [(OD indicated time after transfection -OD 7d after transfection )/OD 7d after transfection ]. Experiments were triplicated and results are expressed as mean  SEM. On the other hand, Biovision's Cell Proliferation Assay Kit is based on a nuclear dye that specifically binds nucleic acid in the cell and generates green fluoresces. The generated fluorescent intensity is directly proportional to the cell number, which can be quantified by measuring fluorescence (Ex/Em = 480/538 nm).

Wound healing, transwell migration and transwell invasion
Wound healing assay and QCM ECMatrix Cell Invasion Kit (ECM554, Millipore) were used to measure cell migration and invasion. Before the cells reached confluence in 6-mm culture plates, a cell-free gap was created using a silicon Culture-Insert (ibidi GmbH) placed on the Petri dishes. After removing the silicon insert from the surface, a clean gap was formed. Cell migration into the clean region was recorded using Axiovert 40 CFL (Zeiss International) at 0, 4, 8 and 12 h, and the percentage of wound healing was determined via dividing the migrated distance by the scratched distance. For transwell invasion assay, BCL6-overexpressed J82 (8  10 5 ) and BCL6-knockdown BFTC905 (1  10 6 ) and BFTC909 cells (8  10 5 ) were starved in media (250 L) containing 0.5% FBS at 37C overnight, seeded in ECMatrix-coated inserts in a 24-well plate. Literally 500 L of media containing 10% (J82 and BFTC905 cells) or 15% (BFTC909 cells) FBS were added into the lower chambers and cells were cultured for another 48 h. The inserts were removed, placed into new lower chambers and the penetrated cells were detached with Cell Detachment Solution and lysed with Lysis Buffer/Dye Solution. The lysed mixtures were transferred to a 96-well plate for fluorescence measurement at wavelengths 480/520 nm using a Beckman Coulter PARADIGM™ Detection Platform. For transwell migration assay, ECMatrix-coated inserts were replaced by BD Falcon Cell Culture Inserts (#353097).

Gelation zymography and MMP2/MMP9 assay
Gelatin zymography was performed based on an early study (Fredriksson et al. 2006; Am J Physiol Lung Cell Mol Physiol 290:L326-33), with some modifications. Cells (5  10 5 in 6-cm dishes) were cultured in serum-free medium (2 mL) for 24 h at 37C with 5% CO 2 before gelatin zymography analysis. Culture media were centrifuged at 4C, 300  g for 10 min and the supernatants were collected. Literally 10 L of the supernatant and 10 L of 2X sample buffer (2% SDS, 0.1% bromophenol blue and 40% glycerol in stacking gel buffer which contains 84 mM ammediol/HCl and 0.02% NaN 3 ) were mixed. Samples were separated by electrophoresis on 6% SDS-polyacrylamide gels containing 0.1% gelatin without reducing agent at 4C. After electrophoresis, gels were washed with 2.5% Triton X-100 twice on a shaker for 30 min at room temperature to remove the SDS. Gels were next incubated at 37C for 48 h in 20 mL of development buffer (200 mM Figure S5. Gelatin zymography on stable BCL6-overexpressed J82 and -knockdown BFTC905 and BFTC909 cells showed MMP2 and/or MMP9 activities were increased and decreased, respectively. Experiments were performed in triplicate and representative images are shown. Figure S6. Stable transfection of the pCMV-BCL6 plasmid into J82 and T24 cells downregulated TP53, CDKN1A and CDKN1B, while knockdown of the BCL6 gene with shBCL6#1 and shBCL6#2 in BFTC905 and/or BFTC909 cells upregulated TP53, CDKN1A, CDKN1C and CDKN2D mRNA levels. All experiments were performed in triplicate and results are expressed as the mean  SEM. Statistical significance: *P < 0.05; **P < 0.01; ***P < 0.001. Figure S7. The correlation coefficients between BCL6 and FOXO3 mRNA levels with two FOXO3 probes were -0.454 and -0.353, respectively, in GSE13507 (Gene Expression Omnibus database, NCBI).  . Stable transfection of the pCMV-BCL6 plasmid into J82 and T24 cells or knockdown of the BCL6 gene in BFTC905 and BFTC909 cells were not able to consistently downregulate or upregulate PTEN mRNA levels. Figure S10. Amplification of the BCL6 gene accounted for 43.5% and 43.8% in all urothelial and muscle-invasive urothelial carcinoma, respectively, from data deposited in the TCGA database.