Portal de Periódicos da CAPES

N-cadherin-mediated adhesion and aberrant catenin expression in anaplastic thyroid-carcinoma cell lines

1999; Wiley; Volume: 83; Issue: 5 Linguagem: Inglês

10.1002/(sici)1097-0215(19991126)83

1097-0215

Johanna Husmark, Nils‐Erik Heldin, Mikael Nilsson,

Thyroid Cancer Diagnosis and Treatment

International Journal of CancerVolume 83, Issue 5 p. 692-699 Experimental CancerFree Access N-cadherin-mediated adhesion and aberrant catenin expression in anaplastic thyroid-carcinoma cell lines Johanna Husmark, Corresponding Author Johanna Husmark johanna.husmark@anatcell.gu.se Institute of Anatomy and Cell Biology, Göteborg University, Göteborg, SwedenInstitute of Anatomy and Cell Biology, Göteborg University, Box 420, Medicinaregatan 3, SE 405 30 Göteborg, Sweden. Fax: +4631-773 33 22.Search for more papers by this authorNils-Erik Heldin, Nils-Erik Heldin Department of Genetics and Pathology, Unit of Pathology, University Hospital, Uppsala, SwedenSearch for more papers by this authorMikael Nilsson, Mikael Nilsson Institute of Anatomy and Cell Biology, Göteborg University, Göteborg, SwedenSearch for more papers by this author Johanna Husmark, Corresponding Author Johanna Husmark johanna.husmark@anatcell.gu.se Institute of Anatomy and Cell Biology, Göteborg University, Göteborg, SwedenInstitute of Anatomy and Cell Biology, Göteborg University, Box 420, Medicinaregatan 3, SE 405 30 Göteborg, Sweden. Fax: +4631-773 33 22.Search for more papers by this authorNils-Erik Heldin, Nils-Erik Heldin Department of Genetics and Pathology, Unit of Pathology, University Hospital, Uppsala, SwedenSearch for more papers by this authorMikael Nilsson, Mikael Nilsson Institute of Anatomy and Cell Biology, Göteborg University, Göteborg, SwedenSearch for more papers by this author First published: 10 November 1999 https://doi.org/10.1002/(SICI)1097-0215(19991126)83:5 3.0.CO;2-1Citations: 39AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinked InRedditWechat Abstract The role of cadherins and catenins in the progression of thyroid carcinoma is unclear. We have investigated α-, β- and γ-catenins and p120ctn in relation to the expression of cadherins in human anaplastic thyroid-carcinoma cell lines (HTh7, HTh74, C643 and SW1736) with Western blotting and immunofluorescence. E-cadherin was lacking except in SW1736, which consisted of E-cadherin-positive (approx. 5%) and -negative cells. The α- and β-catenin levels were similar to those of primary cultured non-neoplastic (E-cadherin-positive) human thyrocytes. In contrast, the expression of γ-catenin was low and variable, correlating with the different levels of cytokeratin in the same cells (HTh74 > SW1736 > C643 > HTh7). p120ctn resolved as a doublet in Western blots; the approximately 100-kDa band also found in non-neoplastic epithelial cells was reduced whereas the approximately 115-kDa band, corresponding to the fibroblast-type isoform of p120ctn, was neo-expressed. A DNA-demethylating agent, 5-aza-2′-deoxycytidine, up-regulated E-cadherin in SW1736 and γ-catenin in SW1736 and C643, whereas the other cell lines were unresponsive; other catenins were not affected. The catenins were generally distributed along the cell borders. Immunostaining, cell-surface biotinylation and co-immunoprecipitation revealed that all cell lines expressed N-cadherin in connection with β-catenin at the plasma membrane. Incubation with an N-cadherin antibody disrupted cell-cell adhesion. We conclude that E-cadherin-negative anaplastic thyroid-carcinoma cell lines display functional N-cadherin/β-catenin complexes, partial or complete loss of γ-catenin, and isoform shift of p120ctn. The unequal expression of E-cadherin and γ-catenin and the variable response to DNA de-methylation suggest that anaplastic thyroid carcinoma is not a uniform entity. Int. J. Cancer 83:692–699, 1999. © 1999 Wiley-Liss, Inc. E-cadherin is a transmembrane glycoprotein which, in the presence of Ca2+, binds homotypically to the extracellular part of complementary molecules on adjacent epithelial cells (Overduin et al., 1995). E-cadherin is also connected to a number of cytoplasmic proteins, the catenins, which have both structural and regulatory roles in cadherin-mediated cell-cell adhesion (reviewed by Aberle et al., 1996). β-catenin and γ-catenin (plakoglobin) are closely related members of the armadillo family of proteins which interacts directly with the cytoplasmic domain of E-cadherin in a mutually exclusive manner. α-catenin is an actin-binding protein which also binds to β-catenin or γ-catenin, thereby connecting the cadherin-catenin complex to the cortical cytoskeleton. A fourth catenin, p120ctn (formerly p120cas), has been shown to bind to E-cadherin independently of the others (Daniel and Reynolds, 1995). The precise function of p120ctn is not clarified, but a role in the regulation of cadherin binding has been suggested (Reynolds et al., 1996; Yap et al., 1998). Loss of expression or function of E-cadherin is often encountered in carcinomas, and correlates with invasiveness, metastatic spread and poor prognosis (for an overview, see Mareel et al., 1995). Conversely, transfected E-cadherin is able to convert E-cadherin-negative cancer cell lines to an epithelial and non-invasive phenotype (Frixen et al., 1991). Altogether, E-cadherin displays the characteristics of a typical tumor invasion suppressor. The mechanism by which E-cadherin is down-regulated is unclear, although hypermethylation of the promoter region of the E-cadherin gene might be involved (Graff et al., 1995). However, not only normal levels of E-cadherin, but also correct localization and proper cadherin-catenin interactions, are necessary for the adherens junction to be functional (Navarro et al., 1993; Bullions et al., 1997). Indeed, a number of studies indicate that defective cell-cell adhesion in tumor cells may also be related to alterations among the catenins (reviewed by Ben-Ze'ev and Geiger, 1998). Loss of E-cadherin expression in sub-sets of thyroid carcinomas is well documented (Brabant et al., 1993; Serini et al., 1996; von Wasielewski et al., 1997). Some alterations of catenin expression have also been demonstrated (Serini et al., 1996; Huang et al., 1998; Cerrato et al., 1998). However, a concise picture of the catenins involved in cadherin binding and adhesiveness in thyroid-tumor cells is lacking. Anaplastic thyroid carcinoma (ATC), the most malignant form of cancer originating from the follicular epithelium, is characterized by a marked reduction or complete loss of thyroid-specific proteins involved in the biogenesis of thyroid hormone. Unlike more differentiated forms (follicular and papillary) of thyroid carcinomas, ATC also shows a complete loss of the normal thyroid histotype, suggesting that cell-cell adhesion in such tumors is severely deranged. Our aim was to investigate possible aberrations in the expression and interactions of cadherins and catenins in human ATC cell lines. MATERIAL AND METHODS Cells, culture conditions and experimental procedures The 4 established ATC cell lines, SW1736 (obtained from Dr. A. Leibowitz, Scott and White Memorial Hospital, Temple, TX), HTh7, C643 and HTh74, used in the present study, were routinely cultured in minimum essential medium (MEM) supplemented with 10% FCS in humidified atmosphere (5% CO2) at 37°C. Primary non-neoplastic human thyroid cells were isolated from paradenomatous or Graves′ thyroid tissue and cultured in Coon's modified Ham medium as described (Nilsson et al., 1996). Additionally, MDCK and MRC-5 (ECACC, No. 84101801) cell lines cultured in MEM supplemented with 5% FCS and primary cultures of pig thyrocytes (Nilsson et al., 1991) were used as controls. All culture reagents were from GIBCO (Paisley, UK). The cells were grown on Petri dishes or Lab-Tech chamber slides (Nunc, Naperville, IL) as well as in Transwell chambers (Corning Costar, Cambridge, MA) pre-coated with collagen type I (Roche, Bromma, Sweden). For de-methylation experiments, cells were cultured in the presence of 1 μM 5-aza-2′-deoxycytidine (AzaC; Sigma, St. Louis, MO) for 5 days. The multiplication rate of the cell lines was estimated by EZ4U assay (Biomedica, Vienna, Austria), detecting the metabolic conversion of tetrazolium salt to formazan. For this purpose, cells were seeded at low density (10 000 cells/well) in a microtiter plate and grown for 4 days. The absorption value was measured daily in triplicates on a microplate reader (Bio-Rad, Upplands-Väsby, Sweden) at 450 nm wavelength with a 620-nm reference. The cell number was calculated by comparing the absorption values with standard curves for each cell line within the linear range. Antibodies Monoclonal antibodies against E-cadherin, α-, β- and γ-catenins, and p120ctn were purchased from Transduction Laboratories (Lexington, KY). Poly- and monoclonal (clone CH-19) pan-cadherin antibodies (raised against a synthetic peptide corresponding to the C-terminal sequence of chicken N-cadherin), a neutralizing N-cadherin antibody (clone GC4) and cytokeratin antibodies (peptide 8, clone M20) were obtained from Sigma. A neutralizing E-cadherin antibody (SHE78-7) was from Zymed (San Francisco, CA). Antibodies against vimentin and biotin- and horseradish peroxidase-conjugated anti-mouse and anti-rabbit antibodies were from Dako (Glostrup, Denmark). Western blotting Confluent cells cultured in Petri dishes were solubilized in Laemmli sample buffer containing 5% mercaptoethanol and boiled for 4 min. Equal amounts of protein from whole-cell extracts, determined according to Karlsson et al. (1994), were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE; 4–15% polyacrylamide) and transferred to nitrocellulose filters (0.45 μm) in a mini-trans-blot cell (Bio-Rad). Blotted filters were incubated in blocking buffer, consisting of 5% dried milk in TBS-Tween (20 mM Tris, 137 mM NaCl, 0.1% Tween-20, pH 7.6), primary antibodies and secondary horseradish-peroxidase-conjugated antibodies for 1 hr each. Antibody incubations were followed by washing 6 × 5 min in TBS-Tween. Immunolabelled proteins were detected by enhanced chemiluminescence (ECL, Amersham Pharmacia Biotech, Uppsala, Sweden) according to the manufacturer's instructions. Molecular weights were estimated by comparison with pre-stained SDS-PAGE molecular-weight standards (Bio-Rad). Immunoprecipitation Cells were grown to confluence in 100-mm Petri dishes and solubilized in lysis buffer (0.5% Triton X-100, 0.01 M maleinic acid, 0.13 M NaCl, 0.03 M KCl, 5 mM glucose, 0.5 mM MgCl2, 1.8 mM CaCl2 and 0.01 M Tris, pH 6.8) for 20 min in the presence of protease inhibitors: 5 μg/ml leupeptin and aprotinin (Sigma) and 0.4 mM 4-(2-aminoethyl)-benzolsulfonylfluorid (Pefabloc; Roche). Lysed cells were scraped off the Petri dishes and cleared from insoluble material in a microfuge for 15 minutes at 14,000 g. The supernatants were diluted to equal protein concentrations and aliquots were incubated with MAbs for 1.5 hr at 4°C on a rocking platform. Protein A-Sepharose CL4-B (Amersham Pharmacia Biotech) was added and the mixture was further incubated for 1 hr. Immunoprecipitated and co-precipitated proteins were washed 3 times in lysis buffer, solubilized by boiling for 4 min in sample buffer, and analyzed by Western blotting as outlined above. Cell-surface biotinylation Petri-cultured cells were washed once in MEM and twice in ice-cold PBS pH 7.5, then incubated with 1 mg/ml NHS-SS-biotin (Pierce, Rockford, IL) diluted in PBS, pH 8.0, for 30 min on ice. Biotinylation was stopped by 3 washes for 10 min with 100 mM glycine in PBS, pH 7.5, and the cells were solubilized in a lysis buffer (50 mM Tris, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, pH 7.5) and cleared from insoluble material as described above. Biotinylated proteins were precipitated overnight at 4°C with streptavidin-agarose (Pierce), and washed sequentially 3 times in lysis buffer, twice in high-salt buffer (lysis buffer with 500 mM NaCl) and once in 10 mM Tris, pH 7.5. Precipitated proteins were solubilized in sample buffer and analyzed by SDS-PAGE and Western blotting. Immunofluorescence and cadherin blocking experiments Cells cultured on chamber slides were fixed in ice-cold ethanol for 15 min, washed with PBS, pH 7.4, and pre-incubated at room temperature with avidin-biotin blocking reagents (Vector, Burlingame, CA) for 2 × 10 min and with blocking buffer consisting of 5% fat-free milk, 0.1% gelatin and 7.5% sucrose in PBS for 10 min. The cells were then incubated in sequence with primary antibodies for 1 hr biotin-conjugated secondary antibodies for 30 min and fluorescein-isothiocyanate-conjugated streptavidin (Dako) for 30 min. The slides were mounted with Vectashield (Vector) and examined in a Nikon Microphot FXA epifluorescence microscope. Some of the cultures subjected to immunostaining were exposed to neutralizing antibodies against N-cadherin (GC4; 80 μg/ml) or E-cadherin (SHE78-7; 1 μg/ml) present in the culture medium for 24 hr prior to fixation. Electron microscopy Cultures were fixed for 1 hr in 2.5% glutaraldehyde in 0.05 M sodium cacodylate, pH 7.4, followed by post-fixation for 1 hr in 1% OsO4, dehydration in ethanol series, and embedding in epoxy resin. Ultrathin sections were contrasted with uranyl acetate and lead citrate and examined in a Philips 400 T electron microscope. RESULTS General characteristics of cell lines Four established ATC cell lines (HTh7, HTh74, SW1736 and C643) were examined. Their duplication rates were fast, ranging between 19 and 29 hr (data not shown). All cells had a spindle-shaped morphology, with ruffled edges during outgrowth (Fig. 1a). At confluence, the cells acquired a more epithelial-like shape (Fig. 1b) but formed multilayers (Fig. 2a). Electron microscopy revealed that most cells were closely connected to each other (Fig. 2b), but typical epithelial junctions (tight junctions, adherens junctions and desmosomes) were not observed. The expression of cytokeratin in Western blots was high in SW1736 and Hth74, but markedly reduced in HTh7 and C643 (Fig. 3a). However, immunofluorescence revealed that most cells were weakly positive for cytokeratin (not shown). For comparison, the vimentin content of the tumor cells was overall equal to that found in primary non-neoplastic human thyrocytes (Fig. 3b). Figure 1Open in figure viewerPowerPoint Light microscopy of sub-confluent (a) and confluent (b) cultures of the SW1736 anaplastic thyroid-carcinoma cell line. There is a change from scattered to epithelial-like morphology, also noted for the other cell lines (not shown), when confluence is reached. Scale bar = 50 μm. Figure 2Open in figure viewerPowerPoint Electron micrographs of filter-cultured SW1736 cells. (a) The confluent cells form multilayers. The cell-cell contact shown in (b) at a higher magnification is indicated (arrow). Scale bar = 10 μm. (b) The plasma membranes of adjacent cells are closely apposed along a considerable length of the cell-cell contact (arrow), but ultrastructurally distinguishable junctions are not present. Scale bar = 2 μm. Figure 3Open in figure viewerPowerPoint Densitometric evaluation of cytokeratin (a) and vimentin (b) expression in Western blots of whole-cell lysates obtained from 4 ATC cell lines (HTh7, C643, SW1736 and HTh74) and primary cultured non-neoplastic human thyrocytes (control). Expression and localization of E-cadherin and catenins The cell lines were examined for the presence of E-cadherin and catenins by Western blotting. None of the cell lines showed any immunoreactivity to E-cadherin in Western blots after normal exposure times (Fig. 4a), but a faint 120-kDa band corresponding to E-cadherin was detected in SW1736 after prolonged exposure (not shown). Only in the SW1736 cell line were small colonies of E-cadherin-positive cells, amounting to less than 5 percent of the total cell number, distinguished by immunofluorescence (Fig. 5a). The levels of α-catenin were similar in the cell lines but slightly lower than that of non-neoplastic thyrocytes (Fig. 4b). The expression of β-catenin was equal in all cell lines and comparable to the level found in control cultures (Fig. 4c). In contrast, more pronounced differences were found for γ-catenin, which was undetectable in HTh7 cells and expressed to a variable extent, although much less than in controls, in the other cell lines (Fig. 4d). Figure 4Open in figure viewerPowerPoint Expression of E-cadherin (a), α-catenin (b), β-catenin (c), γ-catenin (d) and p120ctn (e) proteins in ATC cell lines (HTh7, C643, SW1736, HTh74) and primary control cultured non-neoplastic human thyrocytes (control). Equal amounts of protein from whole-cell extracts were applied in SDS-PAGE. Figure 5Open in figure viewerPowerPoint Immunofluorescence of cadherins and β-catenin in ATC cell lines. (a) E-cadherin in SW1736 cells. Sparse colonies of E-cadherin-positive cells are present within the otherwise unlabelled cell layer. (b) β-catenin in C643 cells. Immunoreactivity is distributed mainly along the cell borders. (c) N-cadherin in SW1736 cells. The culture was treated with AzaC for 5 days before immunolabelling (see last section of Results for further comments). In contrast to the limited expression of E-cadherin (a), all cells exhibit N-cadherin at intercellular contacts. Scale bars (a,c) = 100 μm; (b) 50 μm. The expression of p120ctn was also altered. In primary non-neoplastic thyrocytes, p120ctn appeared as a single band of approximately 100 kDa in size (Fig. 4e). However, in the ATC cell lines, this band was markedly decreased and an additional 115 kDa protein was recognized (Fig. 4e). The possibility that the p120ctn doublet corresponded to phenotype-specific isoforms, as originally described by Reynolds et al. (1996), was examined by comparing p120ctn expression in the ATC cells with that of MDCK epithelial and MRC-5 fibroblast cell lines. As shown in Figure 6, the p120ctn pattern in HTh74 (and in the other cell lines; not shown) was similar to that found in MRC-5 cells. Figure 6Open in figure viewerPowerPoint Comparison of p120ctn expression between HTh74 anaplastic, MRC-5 fibroblastic and MDCK epithelial cell lines. The 2 bands detected in Western blots on whole-cell lysates probably correspond to different p120ctn isoforms. Immunostaining of the ATC cell lines revealed that the catenins were located mainly along the cell borders (exemplified by β-catenin in C643 cells in Fig. 5b). However, unlike the continuous peripheral localization in normal thyrocytes (not shown), catenin distribution in the tumor cells was interrupted and, in addition, numerous punctuate assemblies of immunoreactivity were observed. Catenin binding to aberrant cadherins The membrane localization of catenins indicated the presence of a catenin-binding protein, presumably a cadherin, in the E-cadherin-negative cell lines. Accordingly, a pan-cadherin MAb stained the cell-cell contacts (not shown) and detected a 135-kDa protein in lysates of all cell lines (Fig. 7). Small amounts of this cadherin were also present in primary cultures of non-neoplastic human thyrocytes (Fig. 7). Notably, the pan-cadherin MAb did not stain the 120-kDa E-cadherin band in Western blots (not shown). To further characterize the 135-kDa cadherin, the cell lines were subjected to cell-surface biotinylation, after which labeled proteins were precipitated and analyzed with Western blotting. This revealed that the 135-kDa protein was indeed expressed at the cell surface (Fig. 8a,b). Furthermore, β-catenin was co-precipitated with the biotinylated proteins (Fig. 8c). Since N-cadherin has a molecular weight of about 135 kDa (Volk and Geiger, 1986), we investigated whether N-cadherin was present in the cell lines. As shown for SW1736 in Figure 5c, weak but clear-cut staining of N-cadherin was found at the cell borders in all ATC cells. The available N-cadherin antibody did not directly detect N-cadherin when used in Western blots, but polyclonal pan-cadherin antibodies stained a 135-kDa band in N-cadherin immunoprecipitates (Fig. 9a). Moreover, the N-cadherin antibody co-precipitated β-catenin (Fig. 9b). Figure 7Open in figure viewerPowerPoint Cadherin expression in ATC cell lines. Equal amounts of protein from whole-cell lysates were subjected to Western blotting with antibodies against E-cadherin (a) and pan-cadherin (b), the latter raised against the conserved cytoplasmic domain of N-cadherin. Note that the pan-cadherin MAb, which stains a 135 kDa band strongly in the ATC cell lines (HTh7, C643, SW1736, HTh74) and weakly in non-neoplastic thyrocytes (control), fails to detect a protein at a molecular mass (approx. 120 kDa) corresponding to E-cadherin. Molecular-weight markers are indicated to the right. Figure 8Open in figure viewerPowerPoint Cadherin-catenin complex at the surface of C643 cells. Cell-surface proteins were biotinylated and precipitated with streptavidin-agarose. The resulting precipitate was separated by SDS-PAGE and analyzed with Western blotting using pan-cadherin MAb (a) and polyclonal pan-cadherin (b) and β-catenin (c) antibodies. A 135-kDa cadherin reactive to both poly- and monoclonal pan-cadherin antibodies is present on the cell surface. β-catenin is co-precipitated by cell-surface biotinylated proteins. Molecular-weight markers are indicated to the left. Figure 9Open in figure viewerPowerPoint Immunoprecipitation of N-cadherin. Cell lysates of SW1736 and C643 were incubated with a N-cadherin MAb and the immunoprecipitated and co-precipitated proteins were analyzed by Western blotting with polyclonal pan-cadherin (a) and β-catenin (b) antibodies. N-cadherin is precipitated and detected by pan-cadherin (a); the reason for using anti-pan-cadherin for detection was that the N-cadherin antibody was not reactive in Western-blot analysis. β-catenin is co-precipitated with N-cadherin in the ATC cell lines (b). A negative control, in which the N-cadherin antibody was omitted in the precipitation step, is included in the first lane (−). Incubation of the cultures with the N-cadherin antibody, which is known to disrupt pre-existing homotypic binding (Volk and Geiger, 1986), induced changes in the distribution of the cadherin stained by the pan-cadherin antibody (Fig. 10a,b), as well as of β-catenin (Fig. 10c,d). Similar treatment with an E-cadherin-neutralizing antibody (SHE78-7) had no effect (not shown). Figure 10Open in figure viewerPowerPoint Effect of neutralizing N-cadherin antibody on cell-cell adhesion in C643 cells. Confluent cultures were incubated with the antibody (80 μg/ml) for 24 hr. The distribution of cadherin (a, b), detected by polyclonal anti-pan-cadherin, and β-catenin (c, d) present in untreated (a, c) and antibody-treated (b, d) cultures are compared. Both proteins are changed from an almost linear to a punctate distribution at the cell periphery. In addition, small aggregates appear in the cytoplasm of the antibody-treated cells. Scale bar = 100 μm. Effect of AzaC on E-cadherin and catenins Treatment with the de-methylating agent AzaC (1 μM) for 5 days led to a large increase in the expression of E-cadherin in SW1736 but not in the other cell lines (Fig. 11a). This effect of AzaC corresponded to an increased number of E-cadherin-positive cells in SW1736 (Fig. 12). Another finding was that γ-catenin was increased by AzaC both in SW1736 and in C643, whereas HTh7 and HTh74, which had the lowest and highest amounts of γ-catenin expression before treatment, were unresponsive (Fig. 11c). β-catenin (Fig. 11b) and p120ctn (not shown) were not influenced by AzaC treatment in any of the cell lines. Figure 11Open in figure viewerPowerPoint Effect of 5-aza-2′-deoxycytidine (AzaC) on protein expression of E-cadherin and catenins in ATC cell lines. The cells were incubated with (+) or without (−) AzaC (1 μM) for 5 days. Cell lysates with equal amounts of protein were subjected to SDS-PAGE and Western blotting for E-cadherin (a), β-catenin (b) and γ-catenin (c). Figure 12Open in figure viewerPowerPoint Effect of 5-aza-2′-deoxycytidine (AzaC) on E-cadherin expression in confluent SW1736 cultures. The cells were treated as outlined in Figure 11. After immunofluorescent labelling, 10 areas of 0.16 mm2 were counted for E-cadherin-positive cells in each culture (mean ± SD; n = 3). DISCUSSION Loss of cell-cell adhesion is considered to be a main determinant of the invasive properties and metastasizing capacity of tumors. Most studies concerning altered adhesion in carcinomas have been concentrated on the transmembrane protein E-cadherin. However, there is evidence that the catenins, which normally connect E-cadherin to the actin-based cytoskeleton and regulates its function, might also be involved in tumor progression. In the present study, we investigated the expression and localization of cadherins and all hitherto known catenins in 4 cell lines established from anaplastic thyroid carcinomas. We found that E-cadherin was essentially lacking, but that adhesion instead was mediated by N-cadherin. In addition, the ATC cell lines had altered expression of the catenins, most notably for γ-catenin and p120ctn. The presence of N-cadherin was confirmed by immunofluorescence with specific N-cadherin antibodies. Moreover, β-catenin was shown to co-precipitate with N-cadherin, and N-cadherin antibodies added to the cultures caused a redistribution both of cadherin (recognized by anti-pan-cadherin) and of β-catenin. Thus, anaplastic thyroid-tumor cells display functional N-cadherin/β-catenin complexes at the cell surface. Since N-cadherin is not detected in isolated normal thyroid follicles (data not shown), and we found that only small amounts are present in primary cultured normal human thyrocytes, E-cadherin may be replaced by N-cadherin in poorly differentiated thyroid-carcinoma cells. In support of this possibility is the finding of Islam et al. (1996) that N-cadherin is expressed in squamous-cell carcinomas but not in the normal tissue surrounding the tumor. The presence of N-cadherin appears to correlate with loss of epithelial differentiation in breast-carcinoma cell lines (Hazan et al., 1997). In addition, N-cadherin transfected to squamous epithelial cells results in a scattered phenotype (Islam et al., 1996). N-cadherin, therefore, may be one of several aberrantly expressed molecules determining the malignant properties of epithelial tumors. One proposed mechanism (Hazan et al., 1997) is that N-cadherin present at the surface of tumor cells mediates adhesion to stromal cells, thereby facilitating the local and metastatic spread of the tumor. Based on the present data, it will be important to find out whether N-cadherin is a marker of de-differentiation in thyroid carcinomas in vivo. Reports on catenins in thyroid carcinomas are few. α-catenin has been studied in only a limited number of tissue samples, mainly from differentiated carcinomas, which show normal protein levels (Serini et al., 1996), and in a follicular-carcinoma cell line, in which the distribution of α-catenin appeared to be cytoplasmic and not present at the plasma membrane (Huang et al., 1998). The tumor-suppressing activity of α-catenin is well documented (Shimoyama et al., 1992; Bullions et al., 1997). Whether α-catenin is aberrant in anaplastic thyroid carcinoma is presently unknown. The findings in this study that the ATC cell lines exhibited fairly high levels of α-catenin and also a membranous localization of the protein suggest that the anaplastic phenotype is probably not characterized by loss of α-catenin. We also found that the amount of β-catenin in the ATC cell lines was not different from that in primary cultured non-neoplastic thyrocytes. This is somewhat surprising in view of the findings of Cerrato et al., (1998), indicating a lack of β-catenin in anaplastic tumors, as revealed by immunohistochemistry on de-paraffinized tissue sections. This discrepancy might be explained by differences in the methods employed. Another possibility is that β-catenin might be depressed by some unknown factors in vivo and that the expression is revoked when cells are grown in culture. In support of such a mechanism is the finding that wild-type E-cadherin expressed in tumor cell lines is down-regulated during in vivo growth in nude mice (Mareel et al., 1991). Although β- and γ-catenins display strong structural homologies, their functions are distinct and in some respects opposite to each other (Butz et al., 1992). As shown for endothelial cells, γ-catenin becomes associated to the cell-cell contacts first when the cells are tightly confluent, suggesting that it is important for stabilizing the mature junction (Lampugnani et al., 1995). Unlike β-catenin, γ-catenin is also a plaque protein in desmosomes (hence the synonymous term plakoglobin), in which it provides the connection between desmosomal cadherin and intermediate filaments. Several lines of evidence suggest that γ-catenin/plakoglobin may counteract malignant transformation. Plakoglobin-deficient carcinoma cells retain a spindle-shaped and scattered morphology, despite the fact that functional complexes between E-cadherin and α-/β-catenin are established (Navarro et al., 1993). Transfected γ-catenin has been shown to suppress tumorigenicity independently of cadherin expression (Simcha et al., 1996). Moreover, loss of γ-catenin might be a negative factor in determining the outcome of malignancies (Pantel et al., 1998). In the ATC cell lines examined in the present study, we found that the expression of γ-catenin was reduced ov