A Novel Splice Variant of GLI1 That Promotes Glioblastoma Cell Migration and Invasion

tGLI1 variant retains the ability to activate consensus GLI1-binding sites and undergo nuclear translocalization similar to GLI1. A, characterization of GBM cell lines stably expressing GLI1/tGLI1. Three GBM stable transfectants and a primary GBM specimen were analyzed for GLI1/tGLI1 expression by RT-PCR (left) and immunoblotting (right). In immunoblotting, anti-flag and anti-GLI1 (rabbit) antibodies were used to detect GLI1/tGLI1. B, tGLI1 activates a GLI1-binding site–containing promoter similar to the wild-type GLI1. Stable transfectants were cotransfected with a firefly luciferase reporter containing GLI1-binding sites, 8x3′Gli-BS Luc, and a Renilla reporter to control for transfection efficiency. Forty-eight hours following transfection, cells were treated with and without Shh (100 ng/mL) for 4 h and harvested and the level of transactivation (luciferase activity) was measured ( 11). Relative luciferase activity for the firefly luciferase activity was obtained after normalization against Renilla luciferase. Mean ± SD of three independent experiments. C, tGLI1 protein retains the ability to undergo nuclear translocalization. Top, U87MG-GLI1 and U87MG-tGLI1 cells were fractionated into nuclear and nonnuclear fractions. Lamin B and α-tubulin were used as nuclear and cytoplasmic markers, respectively. Bottom, immunofluorescence staining/confocal microscopy. The flag antibody was used to detect GLI1/tGLI1. Nuclei were visualized by staining with propidium iodide (red). Green, GLI1/tGLI1; yellow, nuclear GLI1/tGLI1. D, expression of GLI1 and tGLI1 did not alter growth rate of U87MG cells. U87MG stable transfectants were monitored for proliferation for 72 h.

tGLI1 variant has a higher propensity to promote the migratory and invasive phenotype in GBM cells than GLI1. A, U87MG-tGLI1 cells migrate at a significantly higher rate than U87MG-vector and U87MG-GLI1 cells. U87MG-vector, U87MG-tGLI1, and U87MG-GLI1 cells were subjected to the scratch wound assay. Columns, width of the initial scratch gap at the start of the experiment. A migratory index, Im, was defined as Im = (g₀ - g_t) / g₀, where g_t and g₀ are the gap widths at time t and time 0, respectively. B, U87MG-tGLI1 cells are more invasive than U87MG-vector and U87MG-GLI1 cells. Following 24-h incubation, the fluorescence in cells that had invaded the basement membrane was quantified. Cell proliferation was simultaneously determined before and after the time required for the invasion assay. Net invasiveness was determined using an invasion/proliferation ratio with U87MG-GLI1 cells as 1.0. P values were determined by the Student's t test. Bottom, invasive cells after staining with crystal violet. C and D, tGLI1-expressing T98G GBM cells are more migratory and invasive than those with GLI1. T98G cells were transiently transfected with the control, GLI1, and tGLI1 expression vectors and subjected to the scratch wound assay for migration (C, top) and RT-PCR and immunoblotting (C, bottom) for expression of GLI1/tGLI1 and invasion assay for invasiveness (D).

tGLI1 leads to a unique expression profile and activates CD24 expression. A, differential gene expression patterns of U87MG-tGLI1 cells compared with U87MG-vector and U87MG-GLI1 cells. GeneChip DNA microarray and subsequent data analyses were conducted to compare gene expression profiles of the U87MG transfectant lines. Cluster analysis included 75 genes that were expressed at significantly higher levels (>2-fold; P < 0.05) in U87MG-tGLI1 cells than U87MG-vector and U87MG-GLI1 cells. A color scale (bottom) shows the range of gene expression. Arrows, CD24 and MEST genes that were validated in subsequent studies. B to D, CD24 expression is up-regulated in tGLI1-expressing GBM cells. RT-PCR (B), quantitative RT-PCR (C), and immunoblotting (D) were done to determine CD24 expression level in U87MG stable transfectants. Compared with U87MG-vector cells, U87MG-GLI1 and U87MG-tGLI1 cells expressed higher levels of PTCH1, a well-known GLI1 target gene. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

tGLI1 protein binds to the CD24 promoter and activates CD24 gene expression. A and B, tGLI1, but not GLI1, binds to the promoter of the CD24 gene. Regular (A) and quantitative (B) chromatin immunoprecipitation assay was done to determine the binding affinity of GLI1 and tGLI1 to promoters of the CD24 gene. Mouse IgG was used as negative immunoprecipitation controls. Chromatin inputs were used as PCR templates to control for loading. CD24 promoter fragments were visualized in agarose gels showing the binding of tGLI1 to the CD24 promoter (A). B, mean ± SD of three independent quantitative chromatin immunoprecipitation experiments. P values were determined using the Student's t test. Right, representative amplification plot. C, tGLI1 activates the CD24 promoter via a 0.06 kb region. Three U87MG stable transfectants were transfected with five reporter constructs, the expression of which is driven by successively truncated CD24 promoter. The parental vector, pGL3-basic, was used as a baseline control. Mean ± SD of three independent experiments. D, tGLI1 is associated with increased CD24 expression in vivo. Using U87MG-GLI1 (n = 3) and U87MG-tGLI1 (n = 3) xenografts established in the flanks of nude mice, immunohistochemistry showed that CD24 expression was significantly higher in U87MG-tGLI1 tumors than U87MG-GLI1 xenografts (representative tumors shown in top). U87MG-tGLI1 tumors are more invasive than U87MG-GLI1 counterparts as indicated by H&E staining and by increased infiltration of U8MG-tGLI1 cells into the smooth muscles (sm). Bottom, RT-PCR results that confirmed expression of GLI1 and tGLI1 transcripts in U87MG-GLI1 and U87MG-tGLI1 xenografts, respectively.