Experimental Cell Research 312 (2006) 457 – 467 www.elsevier.com/locate/yexcr
Research Article
E-cadherin modulates Wnt-dependent transcription in colorectal cancer cells but does not alter Wnt-independent gene expression in fibroblasts Felix Kuphal, Ju¨rgen Behrens* Department of Experimental Medicine II, Nikolaus-Fiebiger-Center for Molecular Medicine, University of Erlangen, Glueckstr. 6, 91054 Erlangen, Germany Received 29 July 2005, revised version received 29 September 2005, accepted 8 November 2005
Abstract E-cadherin mediates homophilic adhesion of epithelial cells and is a major determinant of epithelial differentiation during embryonic development and tumor progression. At cell junctions, E-cadherin associates with h-catenin, which also functions as a transcriptional coactivator of the canonical Wnt signaling pathway by interacting with TCF transcription factors. Here, we have analyzed whether E-cadherin plays a role in the control of gene expression in Wnt-dependent and -independent cellular systems. In DLD-1 colorectal cancer cells, which show constitutive activation of Wnt signaling and exhibit E-cadherin-based cell contacts, the siRNA-mediated knock-down of E-cadherin led to the disturbance of cell junctions, translocation of h-catenin to the nucleus and an enhancement of h-catenin/TCF-dependent reporter activity. In L929 fibroblasts, which are deficient in Wnt signaling and E-cadherin-mediated cell adhesion, ectopic expression of E-cadherin induced the stabilization of h-catenin at the cell junctions and caused marked alterations in cellular morphology and phenotype. However, Ecadherin did not significantly change the transcriptional program of these cells as revealed by DNA microarray analysis. Our data indicate that E-cadherin may modulate Wnt-dependent gene expression by regulating the availability of h-catenin but has a surprisingly small impact on gene expression in the absence of Wnt signaling. D 2005 Elsevier Inc. All rights reserved. Keywords: E-cadherin; Wnt signaling; Gene expression; siRNA; Knock-down; DNA microarray
Introduction E-cadherin is a well-characterized cell – cell adhesion molecule which connects neighboring epithelial cells in a specialized structure termed the adherens junction. Classical cadherins are transmembrane glycoproteins that interact homophilically with cadherin molecules on opposing cell surfaces in a calcium-dependent manner. In order to provide physical rigidity to an epithelial tissue, interacting cadherin molecules are linked to the actin cytoskeleton via proteins of the catenin family [1]. The cytoplasmic portion of cadherins harbors binding sites for h-catenin or g-catenin/plakoglobin and p120 catenin. h-Catenin in turn binds to a-catenin, which mediates the anchorage of the E-cadherin/h-catenin
* Corresponding author. Fax: +49 9131 8529111. E-mail address:
[email protected] (J. Behrens). 0014-4827/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.yexcr.2005.11.007
complex to the underlying cytoskeleton, either by binding directly to F-actin or indirectly via interaction with several F-actin binding proteins [2]. The functional significance of E-cadherin expression during embryonic development is highlighted by the fact that E-cadherin knock-out mice die very early after the morula stage due to a decompaction of the preimplantation embryo [3,4]. In the adult, E-cadherin is an important determinant of the differentiated phenotype of epithelial cells. Mice with a conditional knock-out of E-cadherin display disruption of adherens junctions and severely disturbed differentiation of epidermal keratinocytes and hair follicles [5,6] or show impaired terminal differentiation of alveolar epithelial cells in the mammary gland [7]. Furthermore, a deregulation of the cadherin-based cell adhesion system is frequently observed in highly aggressive and invasive carcinomas [8]. Reintroduction of E-cadherin into invasive cancer cells restores cell – cell adhesion and
458
F. Kuphal, J. Behrens / Experimental Cell Research 312 (2006) 457 – 467
causes a reduction of invasion, which established Ecadherin as a suppressor of invasion [9,10]. In addition, a causal role for the loss of E-cadherin in the progression from adenoma to carcinoma has been demonstrated in vivo [11]. Formation and maintenance of cadherin-mediated cell – cell adhesion is a highly dynamic process, and a number of physiological or pathological conditions alter the stability of the cadherin/catenin complex. Prominent examples are epithelial-to-mesenchymal transitions during gastrulation, or the process of wound healing, in which cells have to loosen their intimate cell – cell contacts and acquire a more mobile phenotype [12]. The downregulation of E-cadherin can occur as a result of the action of transcriptional repressors, e.g., proteins of the Snail/Slug family, which bind to E-box elements within the promoter region of Ecadherin [13,14]. Also, tyrosine phosphorylation of Ecadherin or h-catenin can lead to the destabilization and internalization of adherens junction components [15,16]. In malignant tumors, E-cadherin is frequently downregulated by the aforementioned transcriptional repressors, hypermethylation of the promoter region or mutations in the Ecadherin gene [12,17,18]. Given its potent functional role in epithelial differentiation, the question arises whether E-cadherin exerts its effects in part by altering gene expression. In particular, E-cadherin might be involved in regulating its binding partner h-catenin, which holds a second function as a transcriptional coactivator of the canonical Wnt signaling pathway (for a review, see [19]). When Wnt signaling is inactive, cytosolic h-catenin is constantly degraded by a multiprotein complex. Activation of the pathway results in stabilization of h-catenin in the cytoplasm and its translocation to the nucleus, where it associates with T cell factor (TCF) proteins [20,21] and drives the transcription of Wnt target genes [22]. The action of the Wnt signaling pathway is fundamental during embryonic development and for maintaining the stem cell compartment in the adult, for instance in the crypts of the intestine [23]. Deregulation of the pathway is causally involved in the development of several types of cancer, the most prominent example of which is colorectal cancer [23]. Aberrant Wnt signaling arises in the majority of colorectal tumors from truncating mutations of the adenomatous polyposis coli (APC) tumor suppressor protein, which lead to the stabilization of cytosolic h-catenin, or in a smaller fraction of tumors from stabilizing mutations in the h-catenin gene itself [24]. In several experimental systems, forced expression of Ecadherin or the cadherin cytoplasmic tail is able to sequester h-catenin away from the nucleus and thereby attenuates hcatenin-dependent transcription [25 – 27]. Vice versa, Ecadherin knock-out mouse embryonal stem cells showed higher nuclear localization of h-catenin and higher activity of a h-catenin/TCF reporter system (TOPflash) compared to wild-type cells, which is an indication for increased Wnt signaling if E-cadherin expression is lost [27]. These results indicate that changes in the level of E-cadherin can modulate the activity of the Wnt signaling pathway by altering the
amounts of free h-catenin. In contrast to this, in several welldifferentiated (E-cadherin positive) and dedifferentiated (Ecadherin negative) breast cancer cell lines, E-cadherin deficiency does not correlate with increased nuclear signaling activity of h-catenin [28]. Thus, it is still not completely clear whether the loss of E-cadherin can directly activate h-catenin/ TCF signaling [29]. In this study, we addressed the question whether altering the levels of E-cadherin can lead to changes in gene transcription, either by influencing Wnt signaling or by changing global gene expression. We found that loss of Ecadherin by siRNA-mediated knock-down augments Wntdependent transcription in DLD-1 colorectal cancer cells in which the Wnt pathway is activated but has no effect in HaCaT keratinocytes that do not display Wnt signaling activity. Furthermore, forced expression of E-cadherin in L929 fibroblasts, which are devoid of Wnt-dependent transcription, strongly influences cell morphology and behavior but does not alter gene expression.
Materials and methods Plasmids and cloning Cloning of the E-cadherin cDNA into the doxycyclineregulatable expression vector pUHC10-3 [30] was performed as follows. A 650-bp fragment comprising the 5Vpart of the mouse E-cadherin cDNA was PCR-amplified from the full-size E-cadherin expression vector pBATEM2 [31] using the oligonucleotides 5V-AAAATCTAGAGCCGCCATGGGAGCCCGGTG-3V and 5V-ATGATGAAACGCCAACGGG-3V, thereby adding an XbaI site upstream of the translation start site. The PCR fragment was digested with XbaI and EcoRI (internal restriction site) and cloned into pBSK II as cloning vector (BD Biosciences, Heidelberg, Germany). To obtain the 3V part of the E-cadherin cDNA, pBATEM2 was restricted with EcoRI and BamHI and the resulting fragment ligated in frame to the 5V-fragment in pBSK II. Subsequently, the full-size E-cadherin cDNA was released by XbaI restriction and transferred into XbaI restricted pUHD10-3, generating pUHD E-cad. The doxycycline-inducible transactivator (rtTA2S-M2) was expressed from pWHE146 which carries the neomycin resistance gene on the same transcript linked by an internal ribosome entry site [32]. pUHC13-3 carries a doxycycline-dependent luciferase gene and pUHD16-1 a constitutively expressed h-galactosidase gene [30]. pTK Hyg was used for selection of hygromycin-resistant clones (BD Biosciences). pTOPflash or pFOPflash contains three optimal TCF binding sites (TOP) or mutated sites (FOP), respectively, upstream of the luciferase gene which is under the control of the minimal cFos promoter [33]. pcDNA-h-catenin S33A was used to express degradation-resistant h-catenin S33A [34]. The target sequences of the siRNA oligonucleotides are siE-cad #1: 5V-GAUUGCACCGGUCGACAAAdTdT-3V (Dharma-
F. Kuphal, J. Behrens / Experimental Cell Research 312 (2006) 457 – 467
con, Chicago, IL, USA), siGFP: 5V-GCUACCUGUUCCAUGGCCAdTdT-3V (Eurogentec, Seraing, Belgium), siE-cad #2 was purchased from Qiagen (Hilden, Germany; catalogue number: SI02654029). Cell culture and transfection All cell lines were cultured in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal calf serum (FCS) and penicillin/streptomycin (PAA Laboratories, Pasching, Austria). To generate L929 mouse fibroblasts constitutively expressing the doxycycline-inducible transactivator, cells were stably transfected with pWHE146 using Lipofectamine (Invitrogen, Karlsruhe, Germany) and selected in 800 Ag/ml Geneticin (G418, Invitrogen). Resistant clones were tested for inducibility by transient transfection of the luciferase reporter plasmid pUHC13-3 and induction with doxycycline (Sigma, Schnelldorf, Germany; for all experiments, doxycycline was used at a concentration of 1 Ag/ml). A well-regulatable clone (L20) was selected for co-transfection of pUHD E-cad and pTK Hyg. After selection with Hygromycin B (200 Ag/ml, Sigma), resistant clones were screened for E-cadherin expression by immunofluorescence staining. One clone which expressed E-cadherin in a large percentage of cells in an inducible manner was termed L20-22. For TOPflash assays, cells were transfected in 12-well tissue culture plates with either pTOPflash or pFOPflash (200 ng) and pUHD16-1 (100 ng) to normalize for transfection efficiency. Luciferase activity was assayed 48 h after transfection and normalized to h-galactosidase activity as described [35]. Transient transfections of L-cells for TOPflash assays were performed using Perfectin (Gene Therapy Systems, San Diego, CA, USA), and transient transfection of siRNA for immunofluorescence stainings or TOPflash assays were performed using TransIT TKO (Mirus, Madison, WI, USA), as recommended by the manufacturers. Western blotting Cells were cultured for indicated times, and protein was extracted with L-CAM assay buffer [36] [NaCl 8 g/l, KCl 0.35 g/l, MgSO4 0.16 g/l, CaCl2 0.18 g/l, HEPES (pH 7.4) 2.4 g/l]. Equal amounts of extracted protein were separated by SDS-PAGE and transferred to HybondN nitrocellulose membranes (Amersham, Freiburg, Germany). Antibodies DECMA-1 (Sigma) against mouse E-cadherin, mouse antih-catenin (BD Transduction Laboratories, Heidelberg, Germany) and mouse anti-h-actin (Sigma), as well as peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch, Cambridgeshire, UK) were used according to the manufacturer’s instructions. Signals were visualized using Enhanced Chemoluminescence Reagent (Perkin Elmer, Rodgau, Germany) and detected by exposure to Xray films (Kodak) or with a luminoimager (LAS-3000, Fuji).
459
Immunofluorescence microscopy Cells were grown on glass coverslips, fixed with 3% paraformaldehyde and permeabilized with 0.5% Triton X100. After blocking with 10% FCS in DMEM, cells were incubated with the following antibodies according to the manufacturer’s instructions: DECMA-1 (Sigma), rabbit antih-catenin (Santa Cruz, Santa Cruz, CA, USA) or HECD1 anti-human E-cadherin (TaKaRa, Genevilliers, France). Appropriate Cy2- or Cy3-labeled secondary antibodies (Jackson ImmunoResearch) were used to visualize primary antibodies. F-actin (filamentous actin) was stained using tetramethylrhodamine isothiocyanate labeled-phalloidin (Sigma). Cell micrographs were obtained with a CCD camera (Visitron, Munich, Germany) on a Zeiss Axioplan 2 imaging microscope (Zeiss, Oberkochen, Germany). Wounding assay Monolayers of confluent L20 – 22 cells, which had been precultured for 2 days in the presence or absence of doxycycline to allow for differential E-cadherin expression, were scraped with a plastic pipette tip. After removing dislodged cells by washing, cells were further cultured in fresh medium with or without 1 Ag/ml doxycycline and photographed on a Zeiss Axiovert 25 microscope. Analysis of differential gene expression Total RNA was prepared from L20 – 22 cells using TRIzol reagent (Invitrogen), digested with DNaseI and purified with RNeasy (Qiagen) according to the manufacturer’s instructions. RT-PCR and Northern blot were performed according to standard protocols [37]. Microarray analysis on Affymetrix MG-U74Av2 GeneChips (Affymetrix, Santa Clara, CA, USA) was carried out as described previously [38]. Data were processed using Microarray Suite 5.0 (Affymetrix; statistical algorithm), Data Mining Tool 3.0 (Affymetrix) and MS Excel and Access software. Genes were considered to be differentially expressed if the following criteria were met: (i) transcript got a Fpresent call_ under at least one condition (+ or doxycycline), which indicates that the gene is expressed according to the Affymetrix analysis; (ii) gene was increased (I) or decreased (D) by at least 2-fold in the Affymetrix Fdifference call_; (iii) change P value was 0.05 (difference call reliability); (iv) Affymetrix Fsignal intensity_ was higher than 100 in at least one sample (arbitrary threshold).
Results Impact of E-cadherin on canonical Wnt signaling To determine the impact of endogenous E-cadherin on Wnt signaling, its expression was knocked down using
460
F. Kuphal, J. Behrens / Experimental Cell Research 312 (2006) 457 – 467
small interfering RNA (siRNA) in two epithelial cell lines, DLD-1 colorectal cancer cells and HaCaT keratinocytes. We used two synthetic siRNA oligonucleotides which target different sequences within the coding region of the human E-cadherin gene (denoted siE-cad #1 and #2). Both siRNA oligonucleotides yielded essentially identical results, as documented below. An siRNA oligonucleotide against GFP (siGFP) served as a control. In siGFP transfected DLD-1 and HaCaT cells, as well as in untreated cells (not shown), E-cadherin and h-catenin colocalized at the plasma membrane (Figs. 1A, C; F, H). In both cell lines, transfection of both siE-cadherin oligonucleotides markedly reduced the staining for E-cadherin, in particular, towards the edges of epithelial cell clusters (Figs. 1B, BV, G, GV). In these regions, cells started loosening their cell – cell contacts and were not as tightly packed as in the siGFP control. Interestingly, h-catenin showed differences in its subcellular localization between the two cell lines when E-cadherin was depleted: in DLD-1 cells, h-catenin was lost at the cell junctions and markedly accumulated in the nucleus (Figs. 1D, DV). In HaCaT cells, the level of hcatenin at the cell membrane also dropped concomitantly with the reduction of E-cadherin. However, nuclear localization of h-catenin was not detectable (Figs. 1I, IV), indicating that h-catenin was efficiently degraded in these cells. Some faint residual staining of E-cadherin or hcatenin, which remained at the cell junctions of transfected cells, is likely due to incomplete knock-down of E-cadherin. The relocalization of h-catenin from the plasma membrane to the nucleus in DLD-1 cells prompted us to investigate Wnt-dependent transcription by using the TOPflash reporter system (see Materials and methods). Upon knock-down of E-cadherin in DLD-1 cells, the TOP/FOP ratio more than doubled when compared to the siGFP control (Fig. 1E), indicating that loss of E-cadherin stimulates Wnt-dependent transcription, in line with the stronger nuclear staining of h-catenin observed in Figs. 1D and DV. Co-transfection of the mouse E-cadherin expression vector pBATEM2, which is not targeted by siE-cad #1, reversed this effect, demonstrating that the increase in the TOP/FOP ratio was due to specific knock-down of Ecadherin. pBATEM2 transfection also reduced the basal TOP/FOP ratio in DLD-1 cells and reversed the increase in the TOP/FOP ratio in 293T cells, in which the Wnt pathway was activated by transient transfection of h-catenin S33A or Disheveled 2 (data not shown). In contrast to DLD-1, the siE-cadherin treatment did not lead to a detectable change in the TOP/FOP ratio in HaCaT cells (Fig. 1K). In principle, hcatenin/TCF-dependent transcription can be activated in HaCaT cells, as transient transfection of degradationresistant h-catenin S33A resulted in a 10-fold increased TOP/FOP ratio (Fig. 1K). Taken together, the knock-down of E-cadherin leads to a stimulation of Wnt-dependent transcription in DLD-1 but not in HaCaT cells, indicating a differential capacity of E-cadherin to control Wnt signaling in these different cell types.
Conditional expression of E-cadherin in fibroblasts To investigate whether E-cadherin-mediated cell – cell adhesion can affect gene expression independently of a Wnt signal, we established a conditional expression system in L929 fibroblasts. L929 cells are devoid of endogenous cadherin expression [31] and do not display intrinsic Wntdependent reporter activity (not shown and cf. Fig. 2D). First, L929 cells were stably transfected with the doxycycline-inducible transactivator, generating the cell line L20, which was subsequently transfected with the E-cadherin cDNA under the control of a doxycycline-responsive promoter. This resulted in the cell line L20 – 22, which expresses E-cadherin in more than 90% of the cells when treated with doxycycline (Fig. 3B). In a time course experiment, E-cadherin protein was detectable about 3 h after the administration of doxycycline, and levels continuously increased until 48 h after induction (Fig. 2A). The upregulation of E-cadherin expression was also seen on the mRNA level by Northern blot and RT-PCR analysis (Figs. 2B, C). A faint signal for E-cadherin in L20 – 22 cells without doxycycline in Figs. 2A and C represents the very low basal rate of transcription in the absence of induction. Concomitant with the induction of E-cadherin protein, levels of endogenous h-catenin increased (Fig. 2A), which is likely due to its stabilization upon binding to the cadherin cytoplasmic tail [39]. When doxycycline was withdrawn, E-cadherin as well as h-catenin levels decreased progressively, indicating that the induction of E-cadherin is reversible (not shown). To determine whether E-cadherin expression influences Wntdependent transcription in this system, a TOPflash analysis was performed. As shown in Fig. 2D, neither the parental cell clone L20 nor L20 – 22 cells exhibited differential TOP/ FOP ratios with or without E-cadherin expression. Even when stabilized h-catenin S33A was transfected, the TOP/ FOP ratio did not increase noticeably. Thus, our conditional E-cadherin expression system is well suited to analyze changes in global gene expression which are independent of Wnt-dependent transcription. Expression of E-cadherin resulted in a remarkable change in the morphology of L20 – 22 cells. Upon induction of E-cadherin expression, L20 –22 cells with doxycycline formed epithelium-like clusters (Fig. 3A) and showed colocalization of E-cadherin and h-catenin at cell – cell contacts (Figs. 3B, C, arrowheads), whereas without doxycycline, L20 –22 have a fibroblastic appearance (Fig. 3D) similar to parental L929 cells. Treatment of E-cadherin expressing L20 – 22 cells with the function blocking antibody DECMA-1 caused a complete dissociation of cells, indicating that cell –cell contacts formed were indeed E-cadherin-dependent (not shown). In some cells, we also observed perinuclear localization of E-cadherin, which is probably due to incomplete transport of newly synthesized E-cadherin from the Golgi network to the plasma membrane (Fig. 3B, arrow). Neither E-cadherin nor h-catenin was detected in uninduced cells (Figs. 3E, F). E-cadherin
F. Kuphal, J. Behrens / Experimental Cell Research 312 (2006) 457 – 467
461
Fig. 1. Effect of E-cadherin knock-down using two different siRNA oligonucleotides in DLD-1 (A – E) and HaCaT (F – K) cells. (A – D, BV, DV, F – I, GV, IV) Doubleimmunofluorescence staining of E-cadherin (red) and h-catenin (green) 48 h after transfection of siGFP (A,C, F, H), siE-cad #1 (B, D, G, I) or siE-cad #2 (BV, DV, GV, IV). Arrows, cells showing loss of E-cadherin; arrowheads, DLD-1 cells showing nuclear accumulation of h-catenin; asterisks, sites of loosened cell – cell contacts. (E) TOPflash assays in DLD-1 cells performed 48 h after transfection. siE-cad #1 or siE-cad #2 increase the TOP/FOP ratio compared to siGFP control, which can be reversed by co-transfection of mouse E-cadherin (siE-cad #1+pBATEM2). (K) TOPflash assays in HaCaT cells. siE-cadherin does not increase the TOP/FOP ratio, which can be readily increased by transfection of h-catenin. The values are representatives of at least three independent experiments.
expressing L20 – 22 cells showed circumferential actin cables with emanating radial fibers towards the cell junctions (Fig. 4B). In contrast, in uninduced control cells, actin was largely organized in stress fibers which is typical
for motile fibroblastic cells (Fig. 4D). The reorganization of the actin cytoskeleton indicates that the expression of Ecadherin leads to the productive formation of adherens junctions. To investigate the impact of E-cadherin on cell
462
F. Kuphal, J. Behrens / Experimental Cell Research 312 (2006) 457 – 467
Fig. 4. Impact of E-cadherin expression on the organization of the actin cytoskeleton. L20 – 22 cells were cultured in the presence of doxycycline (A, B, +dox) for 48 h to induce E-cadherin expression or left untreated (C, D, dox) and double-stained for E-cadherin and F-actin. Cells expressing E-cadherin at the sites of cell – cell contacts (A) exhibit circumferential actin cables (B, arrows) with emanating fibers towards the cell junctions (arrowheads). In E-cadherin negative cells (C), the actin cytoskeleton is mainly organized in stress fibers (D).
Fig. 2. Conditional expression of E-cadherin in L20 – 22 fibroblasts. (A) Protein levels of E-cadherin and h-catenin were determined by Western blot analysis after administration of doxycycline (dox) to L20 – 22 or parental L20 cells for different time periods (h, hours). h-Actin was probed as a loading control. (B, C) E-cadherin mRNA levels were determined by Northern blot (B) and RT-PCR (C) analysis in L20 – 22 or L20 cells cultured in the presence or absence of doxycycline for 48 h. GAPDH (glyceraldehyde-3-phosphate dehydrogenase) served as control. (D) TOPflash assays were performed in L20 – 22 and L20 cells with or without doxycycline treatment for 48 h and after transfection of h-catenin.
motility, confluent cultures of L20 –22 with or without Ecadherin expression were wounded through scratching with a sterile plastic pipette tip, and cell movement into the wounded area was monitored for 24 h. Both the parental L20, and L20 –22 cell line without E-cadherin induction,
were able to infiltrate the wound area as single cells (Figs. 5D, F). E-cadherin expressing L20 –22 cells did also close the gap in the cell monolayer (Fig. 5B), however, as judged by videomicroscopy, this appeared to result from a more spread phenotype and epithelial cell sheet movement rather than from migration of single cells. Taken together, we conclude that the ectopically expressed E-cadherin in L20 – 22 cells is fully functional in the formation of adherens junctions and leads to changes in cell morphology, as well as to alterations in the pattern of cell migration. Global gene expression analyses in E-cadherin expressing versus nonexpressing L20– 22 cells Next, we addressed whether establishment of E-cadherindependent cell adhesion in L20 –22 cells leads to alterations
Fig. 3. Morphology of L20 – 22 cells with or without E-cadherin expression. L20 – 22 cells were cultured in the presence (A – C, +dox) or absence (D – F, dox) of doxycycline for 48 h. (A, D) Phase contrast microscopic images (B, C and E, F) double immunofluorescence stainings of E-cadherin and h-catenin. Immunostained E-cadherin localizes to cell – cell contacts and co-localizes with h-catenin (B, C, arrowheads). Some cells also display intracellular, perinuclear staining of E-cadherin (arrow in panel B). Uninduced cells show only background staining for E-cadherin or h-catenin (E, F).
F. Kuphal, J. Behrens / Experimental Cell Research 312 (2006) 457 – 467
463
Fig. 5. Impact of E-cadherin expression on the migration pattern of L20 – 22 fibroblasts. L20 – 22 (A – D) or parental L20 (E, F) cells were precultured for 2 days with or without doxycycline (+, dox), wounded with a plastic pipette tip and photographed immediately after the removal of dislodged cells (0 h; A, C, E), or 24 h after wounding (B, D, F). E-cadherin negative L20 – 22 (D) and L20 (F) cells migrated into the wound area as single cells. L20 – 22 cells which expressed E-cadherin (B) did not infiltrate the wound area as single migrating cells.
in gene expression. We extracted total RNA from L20 –22 cells which had been cultured in the presence or absence of doxycycline and analyzed transcriptional changes using Affymetrix MG-U74Av2 GeneChip microarrays, by which more than 12,000 genes and EST sequences were interrogated. To assess early as well as long-term changes, analyses were performed after induction for 12 h, 24 h, 4 days or 3 weeks. Under all conditions, the vast majority of genes showed no detectable changes in mRNA levels between E-cadherin expressing and nonexpressing cells, with only a few genes displaying a difference of more than 2-fold (Fig. 6C, Supplementary Table 1). A representative scatter plot analysis of L20 – 22 cells after 24 h of + versus E-cadherin expression is depicted in Fig. 6A. The expression differences of selected genes could not be verified by RT-PCR (not shown), probably due to the low expression levels according to the Affymetrix signal intensities (below 500) and the rather small differences in transcription (mostly 2- to 3-fold). This indicates that the expression of E-cadherin did not substantially alter gene transcription. Although exogenous E-cadherin was found differentially expressed by RT-PCR (Fig. 2C), its mRNA levels did not show up as regulated in the microarray experiments, as the cDNA used for the generation of L20 – 22 cells did not contain the 3V untranslated region of the Ecadherin transcript probed by Affymetrix. To demonstrate our ability to detect transcriptional changes in L20 –22 cells using microarrays, we treated L20 – 22 cells for 4 h with 20 AM Forskolin, which is a strong inducer of the cAMP pathway [40]. This data analysis revealed more than 30 genes to be differentially
expressed upon Forskolin stimulation (Figs. 6B, C), and we were readily able to reproduce the transcriptional increase of the cAMP target ICER via RT-PCR (not shown). Furthermore, many of the identified genes have already been described as targets of cAMP signaling [41]. Taken together, although the inducible expression of E-cadherin in L20 – 22 leads to the establishment of functional adherens junctions and changes in cell morphology, the analysis of global gene transcription did not identify major transcriptional changes associated with this process.
Discussion As E-cadherin is a key determinant of epithelial differentiation and malignancy, it has been proposed to act as a signaling molecule besides its function as a cell adhesion receptor [42]. As one read-out for signaling activity, we have here analyzed whether E-cadherin affects gene expression. Since h-catenin is important for both Ecadherin-mediated cell – cell adhesion and Wnt-dependent signal transduction, it represents a putative link between these signaling pathways. In several systems, ectopic expression of E-cadherin negatively affects Wnt/h-catenindependent transcription, which can be explained by competition between E-cadherin and TCFs for binding to hcatenin [26,43,44]. Rather than overexpressing E-cadherin, we depleted endogenous E-cadherin by RNAi and monitored changes in h-catenin-dependent reporter activity (Fig. 1). DLD-1 colorectal cancer cells respond to the knockdown of E-cadherin with nuclear localization of h-catenin
464
F. Kuphal, J. Behrens / Experimental Cell Research 312 (2006) 457 – 467
Fig. 6. Analysis of global gene expression in L20 – 22 cells by Affymetrix DNA microarrays. (A) Representative scatter plot analysis of E-cadherin expressing (+dox) versus nonexpressing ( dox) cells 24 h after addition of doxycycline. The Affymetrix detection signal intensities of the two respective conditions are plotted against each other on a logarithmic scale, whereby each dot represents one probed gene. Diagonal lines labeled 2, 3 or 5 correspond to 2-fold, 3fold or 5-fold expression differences, respectively. (B) Scatter plot of Forskolin treated (20 AM, 4 h) versus DMSO (dimethylsulfoxide) solvent control treated cells. (C) Table giving the total numbers of genes being differentially expressed by a factor of at least 2-fold after E-cadherin induction (E-cad) or Forskolin treatment (FSK) for the indicated time points.
and increased TOPflash activity. This indicates that loss of E-cadherin in these cells releases h-catenin from the cell junctions, which then translocates to the nucleus and activates gene transcription. In contrast, in HaCaT keratinocytes, the knock-down of E-cadherin affected cell junctional integrity similarly to DLD-1 cells but had no effect on h-catenin-dependent transcription or nuclear localization. The difference between both cellular systems is likely due to their differential capacity for h-catenin degradation. DLD-1 cells express a truncated version of the tumor suppressor APC which renders them incapable of degrading h-catenin in the proteasome [45], whereas HaCaT cells represent immortalized keratinocytes that presumably contain an intact h-catenin destruction machinery. hCatenin released from cell junctions might thus be efficiently degraded in HaCaT but not in DLD-1 cells. E-cadherin is frequently downregulated in tumors which correlates with increased invasiveness and metastasis formation. Our data indicate that the loss of E-cadherin will boost Wnt pathway activity in tumors which are deficient for h-catenin degradation, such as colorectal carcinomas. In these tumors, the downregulation of E-
cadherin might not only passively promote metastasis by loss of cell – cell adhesion but, in addition, lead to a stimulation of oncogenic Wnt/h-catenin-dependent transcription. Indeed, colorectal tumor cells at the invasive front exhibit reduced expression of E-cadherin and, concomitantly, pronounced nuclear localization of h-catenin, which correlates with expression of invasion-related genes [46]. The failure of HaCaT cells to activate the TOPflash reporter upon E-cadherin knock-down shows that loss of Ecadherin per se does not activate Wnt signaling unless hcatenin degradation is compromised. In line with this, breast cancer cell lines deficient for E-cadherin expression do not show constitutive activation of the Wnt pathway probably because they retain the capacity for degrading excess hcatenin [28]. One can also speculate that physiological Wnt signaling makes use of the reservoir of h-catenin at the cell junctions during processes in which E-cadherin is downregulated, e.g., during epithelial-to-mesenchymal transitions in embryonic development. It was shown recently that activation of the Wnt pathway leads to the induction of the transcription factor Snail, which represses E-cadherin [47]. Via such a
F. Kuphal, J. Behrens / Experimental Cell Research 312 (2006) 457 – 467
mechanism, Wnt signaling might transiently boost its own activity through increasing the levels of free h-catenin by preventing its degradation and, in addition, promoting its release from E-cadherin. To determine whether E-cadherin-mediated cell adhesion elicits changes in gene transcription which are independent of Wnt signaling, we analyzed global gene expression in L20 – 22 fibroblasts that display E-cadherin-mediated cell – cell contact formation in an inducible manner and do not show h-catenin/TCF-dependent reporter gene activity. We identified only very few genes that appeared to be differentially expressed in the Affymetrix microarray analysis after E-cadherin induction, although the criteria for identifying such genes were deliberately set at a low stringency. Furthermore, none of these genes showed up as being regulated in more than one analysis. In addition, as the expression differences could not be verified by conventional mRNA analysis techniques, we decided not to further analyze these genes. Lack of expression differences in the microarray analyses is not due to technical reasons, as internal control spike-RNAs added to the samples in different amounts were found correctly (not shown). Moreover, the control experiment using Forskolin treatment identified several transcriptional target genes of the cAMP pathway as being regulated, proving the validity of our microarray approach. Although we found no transcriptional changes which manifested stably after various time periods of E-cadherin mediated cell – cell adhesion, it is still possible that E-cadherin triggers early and temporary changes in gene expression. For instance, it was shown previously that the reformation of E-cadherin mediated cell –cell contacts after a calcium switch experiment in MDCK cells leads to the activation of PKB/Akt via PI 3-kinase stimulation, which could participate in the regulation of gene expression. This effect though was transient, and original levels were restored within approximately 1 h [48]. Such changes were not detectable in our analysis, which started 12 h after induction of E-cadherin expression. We did not attempt to analyze earlier time points, e.g., at 6 h, because most of the E-cadherin protein was still located in vesicle-like structures in the cytoplasm and only punctually stained at cell– cell contacts, as judged by immunofluorescence staining (data not shown). The induction of E-cadherin expression in L20 – 22 cells led to marked phenotypic differences, including adherens junction formation, reorganization of the actin cytoskeleton and changes in the pattern of cell migration. Although these changes are reminiscent of mesenchymal-to-epithelial transitions, the expression of epithelial or mesenchymal marker genes, such as vimentin, was unaltered (not shown). This is in line with previous studies demonstrating that E-cadherin expression per se is not sufficient to induce a complete mesenchymal-to-epithelial transition [49,50]. It can be envisaged that fully established epithelial polarity, which we probably did not achieve in our fibroblasts, will strongly affect gene expression, in particular by altering various
465
aspects of cellular physiology [51]. In this context, it might be of interest to perform gene expression analysis after knock-down of E-cadherin or other epithelial determinants in epithelial cells. However, it can be assumed that the loss of E-cadherin in a differentiated epithelial cell line will result in a multitude of secondary effects following the disruption of adherens junctions, which would have to be dissected carefully. Yet, it is astonishing that L20 – 22 cells do not respond to the drastic phenotypic alterations, in particular to the newly established contacts to neighboring cells, by changing their transcriptional program. Therefore, we have to assume that post-transcriptional alterations play a role in mediating the effects of E-cadherin in L20 – 22 cells. A well-known example for such a modification is the stabilization of hcatenin upon E-cadherin expression, which was also seen to accumulate at the cell junctions in the E-cadherin expressing L20 – 22 cells. h-Catenin provides a link to the actin cytoskeleton via its binding to a-catenin [2] and might suffice to induce structural rearrangements within the cell that underlie some of the phenotypic changes observed here. In addition, the rearrangements of the actin cytoskeleton could be accomplished by members of the Rho family of GTPases, which are potent regulators of the cytoskeleton. Several studies have implicated initial E-cadherin-dependent cell – cell contact formation in the transient recruitment and activation of Rho GTPases, i.e., Rac1 and Cdc42 [52,53]. While we focussed on transcriptional alterations in this study, our inducible system will also allow us to analyze post-transcriptional changes in a controlled fashion, e.g., by applying proteomics.
Acknowledgments We thank Dr. L. Klein-Hitpass for the Affymetrix analyses, Drs. C. Berens and W. Hillen for providing plasmids and Dr. B. Mayr for assistance with the Forskolin experiments. This study was supported by grant KR-S05T01 of the ‘‘Nationales Genomforschungsnetz (NGFN)’’.
Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version at doi:10.1016/j.yexcr.2005. 11.007.
References [1] M. Perez-Moreno, C. Jamora, E. Fuchs, Sticky business: orchestrating cellular signals at adherens junctions, Cell 112 (2003) 535 – 548. [2] A. Kobielak, E. Fuchs, Alpha-catenin: at the junction of intercellular adhesion and actin dynamics, Nat. Rev., Mol. Cell Biol. 5 (2004) 614 – 625. [3] L. Larue, M. Ohsugi, J. Hirchenhain, R. Kemler, E-cadherin null
466
[4]
[5]
[6]
[7]
[8] [9]
[10]
[11]
[12] [13]
[14]
[15]
[16]
[17]
[18]
[19] [20]
[21]
[22] [23]
F. Kuphal, J. Behrens / Experimental Cell Research 312 (2006) 457 – 467 mutant embryos fail to form a trophectoderm epithelium, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 8263 – 8267. D. Riethmacher, V. Brinkmann, C. Birchmeier, A targeted mutation in the mouse E-cadherin gene results in defective preimplantation development, Proc. Natl. Acad. Sci. U. S. A. 92 (1995) 855 – 859. C.L. Tinkle, T. Lechler, H.A. Pasolli, E. Fuchs, Conditional targeting of E-cadherin in skin: insights into hyperproliferative and degenerative responses, Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 552 – 557. P. Young, O. Boussadia, H. Halfter, R. Grose, P. Berger, D.P. Leone, H. Robenek, P. Charnay, R. Kemler, U. Suter, E-cadherin controls adherens junctions in the epidermis and the renewal of hair follicles, EMBO J. 22 (2003) 5723 – 5733. O. Boussadia, S. Kutsch, A. Hierholzer, V. Delmas, R. Kemler, Ecadherin is a survival factor for the lactating mouse mammary gland, Mech. Dev. 115 (2002) 53 – 62. J. Behrens, Cadherins and catenins: role in signal transduction and tumor progression, Cancer Metastasis Rev. 18 (1999) 15 – 30. U.H. Frixen, J. Behrens, M. Sachs, G. Eberle, B. Voss, A. Warda, D. Lochner, W. Birchmeier, E-cadherin-mediated cell – cell adhesion prevents invasiveness of human carcinoma cells, J. Cell Biol. 113 (1991) 173 – 185. K. Vleminckx, L. Vakaet Jr., M. Mareel, W. Fiers, F. van Roy, Genetic manipulation of E-cadherin expression by epithelial tumor cells reveals an invasion suppressor role, Cell 66 (1991) 107 – 119. A.K. Perl, P. Wilgenbus, U. Dahl, H. Semb, G. Christofori, A causal role for E-cadherin in the transition from adenoma to carcinoma, Nature 392 (1998) 190 – 193. J.P. Thiery, Epithelial – mesenchymal transitions in tumour progression, Nat. Rev., Cancer 2 (2002) 442 – 454. E. Batlle, E. Sancho, C. Franci, D. Dominguez, M. Monfar, J. Baulida, A. Garcia De Herreros, The transcription factor snail is a repressor of E-cadherin gene expression in epithelial tumour cells, Nat. Cell Biol. 2 (2000) 84 – 89. A. Cano, M.A. Perez-Moreno, I. Rodrigo, A. Locascio, M.J. Blanco, M.G. del Barrio, F. Portillo, M.A. Nieto, The transcription factor snail controls epithelial-mesenchymal transitions by repressing E-cadherin expression, Nat. Cell Biol. 2 (2000) 76 – 83. J. Behrens, L. Vakaet, R. Friis, E. Winterhager, F. Van Roy, M.M. Mareel, W. Birchmeier, Loss of epithelial differentiation and gain of invasiveness correlates with tyrosine phosphorylation of the Ecadherin/beta-catenin complex in cells transformed with a temperature-sensitive v-SRC gene, J. Cell Biol. 120 (1993) 757 – 766. Y. Fujita, G. Krause, M. Scheffner, D. Zechner, H.E. Leddy, J. Behrens, T. Sommer, W. Birchmeier, Hakai, a c-Cbl-like protein, ubiquitinates and induces endocytosis of the E-cadherin complex, Nat. Cell Biol. 4 (2002) 222 – 231. G. Berx, A.M. Cleton-Jansen, F. Nollet, W.J. de Leeuw, M. van de Vijver, C. Cornelisse, F. van Roy, E-cadherin is a tumour/invasion suppressor gene mutated in human lobular breast cancers, EMBO J. 14 (1995) 6107 – 6115. K. Yoshiura, Y. Kanai, A. Ochiai, Y. Shimoyama, T. Sugimura, S. Hirohashi, Silencing of the E-cadherin invasion-suppressor gene by CpG methylation in human carcinomas, Proc. Natl. Acad. Sci. U. S. A. 92 (1995) 7416 – 7419. M.J. Seidensticker, J. Behrens, Biochemical interactions in the Wnt pathway, Biochim. Biophys. Acta 1495 (2000) 168 – 182. J. Behrens, J.P. von Kries, M. Kuhl, L. Bruhn, D. Wedlich, R. Grosschedl, W. Birchmeier, Functional interaction of beta-catenin with the transcription factor LEF-1, Nature 382 (1996) 638 – 642. M. Molenaar, M. van de Wetering, M. Oosterwegel, J. PetersonMaduro, S. Godsave, V. Korinek, J. Roose, O. Destree, H. Clevers, XTcf-3 transcription factor mediates beta-catenin-induced axis formation in Xenopus embryos, Cell 86 (1996) 391 – 399. M. Bienz, H. Clevers, Linking colorectal cancer to Wnt signaling, Cell 103 (2000) 311 – 320. T. Reya, H. Clevers, Wnt signalling in stem cells and cancer, Nature 434 (2005) 843 – 850.
[24] P. Polakis, Wnt signaling and cancer, Genes Dev. 14 (2000) 1837 – 1851. [25] M. Conacci-Sorrell, I. Simcha, T. Ben-Yedidia, J. Blechman, P. Savagner, A. Ben-Ze’ev, Autoregulation of E-cadherin expression by cadherin – cadherin interactions: the roles of beta-catenin signaling, Slug, and MAPK, J. Cell Biol. 163 (2003) 847 – 857. [26] C.J. Gottardi, E. Wong, B.M. Gumbiner, E-cadherin suppresses cellular transformation by inhibiting beta-catenin signaling in an adhesion-independent manner, J. Cell Biol. 153 (2001) 1049 – 1060. [27] S. Orsulic, O. Huber, H. Aberle, S. Arnold, R. Kemler, E-cadherin binding prevents beta-catenin nuclear localization and beta-catenin/ LEF-1-mediated transactivation, J. Cell Sci. 112 (Pt. 8) (1999) 1237 – 1245. [28] M. van de Wetering, N. Barker, I.C. Harkes, M. van der Heyden, N.J. Dijk, A. Hollestelle, J.G. Klijn, H. Clevers, M. Schutte, Mutant Ecadherin breast cancer cells do not display constitutive Wnt signaling, Cancer Res. 61 (2001) 278 – 284. [29] W.J. Nelson, R. Nusse, Convergence of Wnt, beta-catenin, and cadherin pathways, Science 303 (2004) 1483 – 1487. [30] M. Gossen, H. Bujard, Tight control of gene expression in mammalian cells by tetracycline-responsive promoters, Proc. Natl. Acad. Sci. U. S. A. 89 (1992) 5547 – 5551. [31] A. Nose, A. Nagafuchi, M. Takeichi, Expressed recombinant cadherins mediate cell sorting in model systems, Cell 54 (1988) 993 – 1001. [32] A. Knott, K. Garke, S. Urlinger, J. Guthmann, Y. Muller, M. Thellmann, W. Hillen, Tetracycline-dependent gene regulation: combinations of transregulators yield a variety of expression windows. BioTechniques, 32 (2002) 796, 798, 800 [33] M. van de Wetering, R. Cavallo, D. Dooijes, M. van Beest, J. van Es, J. Loureiro, A. Ypma, D. Hursh, T. Jones, A. Bejsovec, M. Peifer, M. Mortin, H. Clevers, Armadillo coactivates transcription driven by the product of the Drosophila segment polarity gene dTCF, Cell 88 (1997) 789 – 799. [34] B. Martin, R. Schneider, S. Janetzky, Z. Waibler, P. Pandur, M. Kuhl, J. Behrens, K. von der Mark, A. Starzinski-Powitz, V. Wixler, The LIM-only protein FHL2 interacts with beta-catenin and promotes differentiation of mouse myoblasts, J. Cell Biol. 159 (2002) 113 – 122. [35] C. Krueger, C. Berens, A. Schmidt, D. Schnappinger, W. Hillen, Single-chain Tet transregulators, Nucleic Acids Res. 31 (2003) 3050 – 3056. [36] J. Behrens, M.M. Mareel, F.M. Van Roy, W. Birchmeier, Dissecting tumor cell invasion: epithelial cells acquire invasive properties after the loss of uvomorulin-mediated cell – cell adhesion, J. Cell Biol. 108 (1989) 2435 – 2447. [37] J. Sambrook, D.W. Russell, Molecular Cloning—A Laboratory Manual, 3rd edR, Cold Spring Harbor Laboratory Press, New York, 2001. [38] J. Durig, H. Nuckel, A. Huttmann, E. Kruse, T. Holter, K. Halfmeyer, A. Fuhrer, R. Rudolph, N. Kalhori, A. Nusch, S. Deaglio, F. Malavasi, T. Moroy, L. Klein-Hitpass, U. Duhrsen, Expression of ribosomal and translation-associated genes is correlated with a favorable clinical course in chronic lymphocytic leukemia, Blood 101 (2003) 2748 – 2755. [39] M. Ozawa, H. Baribault, R. Kemler, The cytoplasmic domain of the cell adhesion molecule uvomorulin associates with three independent proteins structurally related in different species, EMBO J. 8 (1989) 1711 – 1717. [40] B. Mayr, M. Montminy, Transcriptional regulation by the phosphorylation-dependent factor CREB, Nat. Rev., Mol. Cell Biol. 2 (2001) 599 – 609. [41] X. Zhang, D.T. Odom, S.H. Koo, M.D. Conkright, G. Canettieri, J. Best, H. Chen, R. Jenner, E. Herbolsheimer, E. Jacobsen, S. Kadam, J.R. Ecker, B. Emerson, J.B. Hogenesch, T. Unterman, R.A. Young, M. Montminy, Genome-wide analysis of cAMP-response element binding protein occupancy, phosphorylation, and target gene activation in human tissues, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 4459 – 4464.
F. Kuphal, J. Behrens / Experimental Cell Research 312 (2006) 457 – 467 [42] M.J. Wheelock, K.R. Johnson, Cadherin-mediated cellular signaling, Curr. Opin. Cell Biol. 15 (2003) 509 – 514. [43] F. Fagotto, N. Funayama, U. Gluck, B.M. Gumbiner, Binding to cadherins antagonizes the signaling activity of beta-catenin during axis formation in Xenopus, J. Cell Biol. 132 (1996) 1105 – 1114. [44] M. Shtutman, J. Zhurinsky, I. Simcha, C. Albanese, M. D’Amico, R. Pestell, A. Ben-Ze’ev, The cyclin D1 gene is a target of the betacatenin/LEF-1 pathway, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 5522 – 5527. [45] R. Rosin-Arbesfeld, A. Cliffe, T. Brabletz, M. Bienz, Nuclear export of the APC tumour suppressor controls beta-catenin function in transcription, EMBO J. 22 (2003) 1101 – 1113. [46] T. Brabletz, A. Jung, S. Reu, M. Porzner, F. Hlubek, L.A. KunzSchughart, R. Knuechel, T. Kirchner, Variable beta-catenin expression in colorectal cancers indicates tumor progression driven by the tumor environment, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 10356 – 10361. [47] J.I. Yook, X.Y. Li, I. Ota, E.R. Fearon, S.J. Weiss, Wnt-dependent regulation of the E-cadherin repressor snail, J. Biol. Chem. 280 (2005) 11740 – 11748. [48] S. Pece, M. Chiariello, C. Murga, J.S. Gutkind, Activation of the
[49]
[50]
[51] [52]
[53]
467
protein kinase Akt/PKB by the formation of E-cadherin-mediated cell – cell junctions. Evidence for the association of phosphatidylinositol 3-kinase with the E-cadherin adhesion complex, J. Biol. Chem. 274 (1999) 19347 – 19351. E. Lozano, A. Cano, Induction of mutual stabilization and retardation of tumor growth by coexpression of plakoglobin and E-cadherin in mouse skin spindle carcinoma cells, Mol. Carcinog. 21 (1998) 273 – 287. P. Navarro, E. Lozano, A. Cano, Expression of E- or P-cadherin is not sufficient to modify the morphology and the tumorigenic behavior of murine spindle carcinoma cells. Possible involvement of plakoglobin, J. Cell Sci. 105 (1993) 923 – 934. W.J. Nelson, Adaptation of core mechanisms to generate cell polarity, Nature 422 (2003) 766 – 774. M. Betson, E. Lozano, J. Zhang, V.M. Braga, Rac activation upon cell – cell contact formation is dependent on signaling from the epidermal growth factor receptor, J. Biol. Chem. 277 (2002) 36962 – 36969. J.S. Ehrlich, M.D. Hansen, W.J. Nelson, Spatio-temporal regulation of Rac1 localization and lamellipodia dynamics during epithelial cell – cell adhesion, Dev. Cell 3 (2002) 259 – 270.