E-Cadherin mediates MMP down-regulation in highly invasive bronchial tumor cells

Abstract

The disorganization of E-cadherin/catenin complexes and the overexpression of matrix metalloproteinases (MMPs) are frequently involved in the capacity of epithelial cells to acquire an invasive phenotype. The functional link between E-cadherin and MMPs was studied by transfecting invasive bronchial BZR tumor cells with human E-cadherin cDNA. Using different in vitro (cell dispersion, modified Boyden chamber) and in vivo assays (human airway epithelial xenograft), we showed that E-cadherin-positive clones displayed a decrease of invasive abilities. As shown by immunoprecipitation, the re-expressed E-cadherin was able to sequestrate one part of free cytoplasmic beta-catenin in BZR cells. The decrease of beta-catenin transcriptional activity in E-cadherin-transfected clones was demonstrated using the TOP-FLASH reporter construct. Finally, we observed a decrease of MMP-1, MMP-3, MMP-9, and MT1-MMP, both at the mRNA and at the protein levels, in E-cadherin-positive clones whereas no changes in MMP-2, TIMP-1, or TIMP-2 were observed when compared with control clones. Moreover, zymography analysis revealed a loss of MMP-2 activation ability in E-cadherin-positive clones treated with the concanavalin A lectin. These data demonstrate a direct role of E-cadherin/catenin complex organization in the regulation of MMPs and suggest an implication of this regulation in the expression of an invasive phenotype by bronchial tumor cells.

Figure 1.

Expression of E-cadherin in transfected BZR cells. A: Screening by Western blot analysis of E-cadherin (E-cad) expression in E-cadherin-transfected clones (BZR-PJ3ECAD) and vector-transfected clones (BZR-PJ3). Anti-actin staining indicates equivalent protein loading. B: Phase-contrast photographies showing the phenotype of E-cadherin-transfected clone 24 (a) and vector-transfected clone 23 (b). Clones 24 and 23 are shown as representative of E-cadherin or control transfectants, respectively. Bars, 100 μm.

Figure 2.

Effect of E-cadherin expression on cell dispersion. A: Comparison of partitions obtained by the three geometrical models, namely Voronoi’s partition (Voronoi), Delaunay’s graph (Delaunay), and minimum spanning tree (MST) (a, b, c, respectively), of the E-cadherin-transfected clone 24 and the vector-transfected clone 23 (d, e, f, respectively) after 24 hours of culture. B: Comparison of parameters calculated from each of the three graphs (a, b, c) for the E-cadherin-transfected clone 24 and the vector-transfected clone 23. The three geometrical models showing that the RFH (roundness factor homogeneity) value decreases while the AD (area disorder) value increases and that the m (average length) value decreases while the σ (SD) value increases, indicate that the E-cadherin-transfected clone 24 displayed a more cohesive cluster spatial distribution than vector-transfected clone 23 (P < 0.05).

Figure 3.

Effect of E-cadherin expression on in vitro cell invasion. Analysis of the invasive properties of E-cadherin-transfected clones and vector-transfected clones in a modified Boyden chamber invasion assay. The specific synthetic MMP inhibitor BB-94 (BB94) was used to demonstrate that invasiveness of vector-transfected clones was MMP dependent. Results are expressed as the mean of three different experiments ± SD. Thirty fields (×400) were counted on each filter (**, P < 0.01; *, P < 0.05).

Figure 4.

Effect of E-cadherin expression on in vivo cell invasion. A human airway epithelial xenograft model was used to mimic bronchial tumor invasion. HPS staining of paraffin sections of tracheae removed 3 days (a and b), 1 week (c and d), 3 weeks (e and f), 4 weeks (g and h), and 8 weeks (i and j) after the graft was performed. The E-cadherin-transfected clone 24 and the vector-transfected clone 17 are shown as example of E-cadherin and of control transfectants, respectively. After cells had repopulated the denuded tracheae (arrowheads) (a and b), they proliferated (c and d) and formed a tumor mass (T) well delimited by the basement membrane (arrowheads) (e and f). Only control vector-transfected cells gave rise to an aggressive tumor that colonized the submucosal space (SU) and infiltrated beneath the cartilage (CA) to form tumor islands (TI) into the surrounding host tissue (g and h) up to the mouse skin (SK) (i and j). Bars, 84 μm.

Figure 5.

Effect of E-cadherin expression on β-catenin transcriptional activity. A: Western blot analysis of overall expression of β-catenin (β-cat) in E-cadherin-transfected clones and vector-transfected clones. Anti-actin staining indicates equivalent protein loading. B: Analysis of interaction between E-cadherin and β-catenin in E-cadherin-transfected clones by E-cadherin immunoprecipitation coupled to β-catenin Western blot. C: Measure of β-catenin/TCF activity in E-cadherin-transfected clones and vector-transfected clones by transient transfection of the TOP-FLASH firefly luciferase reporter construct. Data are expressed as RLUs (relative light units) normalized to the co-transfected Renilla luciferase-encoding phRG-TK plasmid (**, P < 0.01).

Figure 6.

Effect of E-cadherin expression on in vitro MMP production. A: RT-PCR analysis of levels of MMP and TIMP mRNA in E-cadherin-transfected clones and vector-transfected clones. The levels were normalized to the control 28S and results are expressed as the mean of three different experiments ± SD (*, P < 0.05; **, P < 0.01). B: Western blot analysis of conditioned media or cellular extracts of E-cadherin-transfected clones and vector-transfected clones using monoclonal antibodies directed against MMP-1, MMP-3, and MT1-MMP. HT1080 cells were used as a control showing the main forms (pro-form and active form, respectively) of MMP-1 (52 and 42 kd), MMP-3 (57 and 45 kd), and MT1-MMP (63, 60, and 43 kd). For MMP-1 and MMP-3 detection, conditioned medium from equal cell numbers was used. For MT1-MMP detection, anti-actin staining indicates equivalent protein loading. C: Analysis by zymography of gelatinolytic activities (92 kd, pro-MMP-9; 72 kd, pro-MMP-2; 62/59 kd, active MMP-2) in conditioned media of E-cadherin-transfected clones and vector-transfected clones treated (+) or not (−) with Con A. An MT1-MMP RT-PCR was performed on corresponding cells to demonstrate its involvement in MMP-2 activation. MT1-MMP levels were normalized to the control 28S.

Figure 7.

Effect of E-cadherin expression on in vivo MMP production. Immunohistochemistry with monoclonal antibodies directed against human MMP-1 (a and b, respectively), MMP-2 (c and d, respectively), MMP-3 (e and f, respectively), MMP-9 (g and h, respectively), and MT1-MMP (i and j, respectively) was performed to compare the MMP expression profile of E-cadherin- and vector-transfected clones in the in vivo human airway epithelial xenograft model. Bars, 21 μm.