Article Figures & Data
Figures
- Figure 1.
Stable overexpression of PAX6 in GBM cells suppresses cell invasiveness. A, depiction of PAX6 mutation constructs. Western blotting using polyclonal PAX6 antibody in whole-cell lysate or nuclear extract of U251HF (B) and whole-cell lysate of LN229 (E) determined positive stable transfectants of wild-type and mutants of PAX6. p, parental wild-type GBM cell line; Vec, vector transfectants. PAX6 and 344 in vitro synthesized proteins were made from plasmid constructs in pRC-CMV vector using TnT T7 Quick Transcription/Translation System (Promega, Madison, WI) and used as a control for PAX6 antibody. Actin antibody detected actin in the same blot after stripping as control of protein loading. C, representative pictures of cells above Matrigel (left) and cells that invaded through Matrigel (right) after 16 hours of culture. D and F, comparison of cell invasiveness of GBM cell lines U251HF and LN229, respectively, after stable transfection with wild-type or mutant PAX6. The position of the bar is the mean of the normalized cell counts from duplicate or triplicate invasion assays of three to five independent experiments. Bonferroni-adjusted P < 0.0001 from Wilcoxon rank-sum tests was used for statistically significant difference between groups.
- Figure 2.
Stable overexpression of PAX6 suppresses the expression of MMP2 in U251HF cells grown in vitro. Attached subconfluent cells were cultured in serum-free medium for 48 hours before RNA extraction from attached cells and precipitation of secreted proteins, as described in Materials and Methods. A, real-time quantitative RT-PCR quantified MMP2 expression ratios relative to enolase-α and GAPDH, expressed as percent of parental cells (p). B, a representative gelatin zymography gel. C, densitometry of gelatin degradation by MMP2. Data for statistical analysis were from three to six independent experiments. Expression ratios were log transformed to stabilize variance for mixed model ANOVA with Construct as fixed effect and Experiment as random effect. Columns, mean as calculated by back-transformation from the least-square means and ANOVA; bars, SD analyzed by mixed model ANOVA. Horizontal lines above the columns indicate Construct Groupings, with Bonferroni-adjusted P values shown for the indicated contrasts among Construct Groupings.
- Figure 3.
Stable overexpression of PAX6 suppresses the expression of MMP2 grown in vivo. A, representative MMP2 immunoreactivity in intracranial xenografts of U251HF (a, ×400; b, ×100), PAX6-2.3 (c, ×400), and PAX6-2.2 (d, ×400). Small and large arrows, infiltrating cells and infiltrated tumor with positive MMP2 staining, respectively. Positive MMP2 staining (brown) was shown in the cytoplasm of tumor cells whereas the nucleus was stained blue with hematoxylin. B, MMP2 expression ratios times 1,000 relative to enolase-α, β-actin, or GAPDH in intracranial xenografts of U251HF, PAX6-2.2, and PAX6-2.3, each quantified in real-time quantitative RT-PCR by the human gene-specific primers and the gene-specific standards. For survival days of the animal, refer to Zhou et al. ( 6).
- Figure 4.
Adenoviral-mediated overexpression of PAX6 suppresses MMP2 expression in malignant GBM cell lines U251HF, LN229, and U87 in vitro. Equal amounts of cells were infected with Ad-PAX6 (lane 2), Ad-344 (lane 3), and Ad-GFP (lane 4; 50 viral particles per cell) and incubated in serum-free medium for 48 hours before extraction of RNA and precipitation of secreted protein. Mock infection (lane 1) was used as a control. A to C, comparison of MMP2 expression quantified by real-time quantitative RT-PCR and normalized to GAPDH (black column) or enolase-α (white column). D to F, representative gelatin zymography of U251HF (4 μg protein per lane), LN229 (0.6 μg protein per lane), and U87 (1.5 μg protein per lane), respectively. Right, densitometry of the MMP2 band. Columns, group mean from three to five independent experiments as calculated by back-transformation from the least-square means and ANOVA; bars, SD analyzed by mixed model ANOVA. Post hoc comparisons were of infection (noncontrol) to mock (control) of each cell line. Symbols above noncontrol columns indicate Bonferroni-adjusted P values for difference from control.
- Figure 5.
PAX6 binds directly to MMP2 promoter and suppresses MMP2 promoter activity in glioma cells. A, depiction of the MMP2 promoter constructs in the luciferase reporter vector pGL3B and probe location using EMSA. MMP2 promoter activity in U251HF and its stable transfectants of PAX6 and PAX6-344 (B) and in LN229 and its stable transfectants of PAX6 (C) after transient transfection with three MMP2 promoter luciferase constructs: MMP2P1/pGL3B (black column), MMP2P5/pGL3B (white column), or MMP2P6/pGL3B (gray column). Columns, group mean from three to five independent experiments at three repeats per experiment; bars, SD. D, EMSA narrowed PAX6 binding to a 245-bp region in MMP2 promoter, MMP2P5-1, using in vitro synthesized PAX6 (lanes 1-12) or PAX6-344 (344; lanes 12-22), or nuclear extract from U251HF cells (lanes 23 and 24) with the [α-32P]-labeled MMP2P5-1. PAX6 binding specificity was determined by addition of PAX6- or paired domain–binding competitors Tsp11Bs or CD19, respectively, mutant oligomers of Tsp11Bs (Tsp11Bm1), and rabbit antibodies for PAX6 (PAX6 Ab) and C-Myb (C-Myb Ab). Minus, no addition of competitor/antibody; arrowheads, MMP2P5-1/PAX6 complex or MMP2P5-1/PAX6-344 complex; arrow, supershifted complex of PAX6 antibody/PAX6/MMP2P5-1. Data from each promoter and experiment were converted to percentages of the mean value for its control (U251HF with same promoter), square root transformed to stabilize variance, then analyzed by mixed model ANOVA. Post hoc comparisons were of noncontrols to same-promoter controls. Symbols above noncontrol columns indicate Bonferroni-adjusted P values for difference from U251HF.
- Figure 6.
MMP2 is one of the PAX6 target genes mediating PAX6 suppression of glioma cell invasion and reverse correlation between PAX6 and MMP2 expression in GBM. A, gelatin zymography assay showing knocked down MMP2 protein secreted from a 344 stably transfected U251HF (344-1) after stable expression of inhibitor RNA of MMP2 from retroviral construct V2HS-48431, as described in Materials and Methods. B, comparison of cell invasiveness by Matrigel invasion assay of three 344-1(shMMP2) clones with different levels of MMP2. Invasion data for the wild-type cells (U251HF) and a PAX6-transfected line (PAX6-2.2) were included for comparison. The invaded cell counts were normalized to U251HF. The position of the bar is the mean of the normalized cell counts from duplicate or triplicate invasion assays of three to five independent experiments. Bonferroni-adjusted P < 0.0001 from Wilcoxon rank-sum tests was used for statistically significant difference between groups. C, depictions of correlations between the PAX6 and MMP2 expression ratios relative to β-actin in 43 astrocytomas and 7 adjacent normal tissues (top) and 41 GBMs and 7 adjacent normal tissues (bottom). Spearman correlation coefficients (R value) and the significance (P value) are shown.
Volume 66, Issue 20
