Fig 2: RNAi-mediated knockdown of BCL9L diminishes proliferation, migration and invasion of pancreatic cancer cellsRelative levels of BCL9L mRNA (A) and protein expression (B) were determined in Panc-1 cells stably transduced with BCL9L-shRNA and control constructs by qRT-PCR and western blotting, respectively. shRNA-knockdown using the shBCL9L_2 construct resulted in ~80% reduction of BCL9L mRNA expression and almost complete abrogation of BCL9L protein expression compared to control cells. BCL9L mRNA levels were normalized to GAPDH. Tubulin served as the loading control in western blotting experiments. Shown is a representative image of 3 experiments (C) Stable knockdown of BCL9L significantly reduced the proliferative capacity of Panc-1 cells as quantified by DNA BrdU incorporation assay. (D) Wound healing assays revealed significantly impaired migration of Panc-1 cells stably transduced with BCL9L shRNA compared to controls. (E) Analysis of cell invasion showed a significantly reduced ability of BCL9L-knockdown Panc-1 cells to invade through a matrigel-coated transwell membrane in comparison to control cells. Right panel shows representative phase-contrast images of crystal-violet-stained BCL9L-knockdown (right) and control cells (left) on the lower side of a transwell membrane. All data represent mean ± s.e.m. of at least three independent experiments. **p = 0.01, ***p = 0.001 (Student's t-test).

Fig 3: BCL9L knockdown does not affect transcriptional activity of ß-catenin in pancreatic cancer cellsLuciferase activity of reporter constructs containing wild-type (TOPFLASH; A) or mutated (FOPFLASH, negative control; B) TCF/LEF binding site repeats, respectively, in Panc-1 BCL9L-knockdown and control cells. Following transient transfection with luciferase constructs, cells were treated overnight with 10 µM CHIR inhibitor or left untreated and subjected to analysis of luciferase activity. Data represent mean ± s.e.m. of three independent experiments.

Fig 4: RNAi-mediated inhibition of BCL9L expression counteracts epithelial-mesenchymal transition in pancreatic cancer cells treated with TGF-ß(A) Panc-1 cells stably transduced with control vector or BCL9L shRNA (shBCL9L) were treated with 5 ng/ml TGF-ß or left untreated for 72 h and subsequently visualized using a phase-contrast microscope. Control cells responded to TGF-ß treatment by adopting a mesenchymal, spindle-like phenotype whereas BCL9L-knockdown cells largely retained the cobblestone-like epithelial morphology. (B) Western Blot analysis of E-cadherin and a-tubulin (loading control) protein levels in Panc-1 cells treated with 5 ng/ml TGF-ß for 72 h. shRNA- and siRNA-mediated knockdown of BCL9L (shBCL9L, siRNA1, siRNA2, siRNA pool) induced an upregulation of E-cadherin expression in comparison to controls (Vector ctrl, siCTR). Shown are representative images of 2 experiments. mRNA expression levels of epithelial (CDH1) and mesenchymal (SNAI2, VIM) genes were quantified in BCL9L-knockdown and control Panc-1 cells treated or not with TGF-ß for 24 h (C), 48 h (D), 72 h (E) and 96 h (F), respectively. Data shown represent mean and standard deviation of 3 independent experiments. *p = 0.05, **p = 0.01, ***p = 0.001 (Student's t-test).

Fig 5: Knockdown of BCL9L stimulates E-cadherin expression and plasma-membrane retention in complex with ß-cateninE-cadherin, a-catenin, BCL9L and a-tubulin (loading control) protein levels were determined in stably shRNA-transduced Panc-1 cells (A) as well as S2-007, S2-028 and Panc-1 cells transiently transfected with siRNA (B). Knockdown of BCL9L by both shRNA (shBCL9L) and siRNA (siRNA1, siRNA2, siRNA pool) induced an up-regulation of E-cadherin protein in all cell lines compared to controls (Vector control, siCTR). Shown are representative images of 3 western blot experiments. qRT-PCR analyses of BCL9L and E-cadherin mRNA expression levels in Panc-1 (C), S2-007 (D), and S2-028 (E) cells transfected with control siRNA or BCL9L-targeting siRNA showed a significant up-regulation of E-cadherin expression in response to knockdown of BCL9L. Data represent mean ± s.e.m. of three independent experiments. **p = 0.01, ***p = 0.001 (Student's t-test). (F) Untransduced (UNT), Vector control and BCL9L-shRNA transduced Panc-1 cells were treated or not with 10 µM CHIR and subjected to immunofluorescence co-staining of E-cadherin (left panel; visualized by Alexa 555-tagged secondary antibody) and ß-catenin (middle panel; visualized by Alexa 488-tagged secondary antibody), respectively. Irrespective of CHIR treatment BCL9L-knockdown cells show increased retention of E-cadherin and ß-catenin staining at the plasma membrane compared to controls (see arrows, bottom right panel). (G) Quantification of CHIR-induced nuclear translocation of ß-Catenin in samples from (F). Data shown represent mean and standard deviation of mean grey values from nuclear ROIs of each 20 cells quantified as described in Materials and Methods. (H) ß-catenin was immunoprecipitated from total cell lysates of Panc-1 BCL9L-knockdown and control cells treated or not with CHIR inhibitor. Western Blot analysis of ß-catenin, E-cadherin and GAPDH (control) protein levels in input (IP) and bound (B) fractions of the immunoprecipitation revealed increased binding of E-cadherin to ß-catenin protein in presence and absence of CHIR in BCL9L knockdown cells compared to controls. Shown is a representative image of 2 experiments.