Carcinogenesis vol.30 no.10 pp.1781–1788, 2009
doi:10.1093/carcin/bgp175
Advance Access publication July 14, 2009
P-cadherin induces an epithelial-like phenotype in oral squamous cell carcinoma by
GSK-3beta-mediated Snail phosphorylation
Karin Bauer, Albert Dowejko, A.-K.Bosserhoff1,
T.E.Reichert and Richard Josef Bauer
Department of Oral and Maxillofacial Surgery and 1Institute of Pathology,
University of Regensburg, Franz-Josef-Strauss-Allee 11, D-93053
Regensburg, Germany
To whom correspondence should be addressed. Tel: þ49 941 943 1627;
Fax: þ49 941 943 1631;
Email: [email protected]
Cadherins belong to a family of Ca21-dependent homophilic
cell–cell adhesion proteins that are important for correct cellular localization and tissue integrity. They play a major role in the
development and homeostasis of epithelial architecture. Recently, it has become more and more evident that P-cadherin
contributes to the oncogenesis of many tumors. To analyze the
role of P-cadherin in oral squamous cell carcinoma (OSCC), we
used a cell line that was deficient of the classical cadherins,
P-cadherin, E-cadherin and N-cadherin. This cell line was transfected with full-length P-cadherin (PCI52_PC). After overexpression of P-cadherin, PCI52_PC gained an epithelial-like
brickstone morphology in contrast to the mock-transfected cells
with a spindle-shaped mesenchymal morphology. Immunohistochemical analysis revealed a strong nuclear Snail staining in
mock-transfected cells compared with a significantly reduced
nuclear staining and translocation to the cytoplasm in
P-cadherin-overexpressing cells. Interestingly, the effects triggered by P-cadherin overexpression could be reversed by transfecting the cells with an antisense P-cadherin plasmid construct.
Additional investigations showed a reexpression of E-cadherin
in all P-cadherin-transfected cell clones in contrast to the mock
controls. Analyzing the signaling mechanism behind it, we found
glycogen-synthase-kinase-3beta (GSK-3beta) bound to Snail in
all cell clones. Furthermore, P-cadherin-overexpressing cell lines
showed activated GSK-3beta that phosphorylated Snail leading to
its cytoplasmic translocation. In summary, our results reveal Pcadherin as one major component in reconfiguring mesenchymal
cells with epithelial features by triggering GSK-3beta-mediated
inactivation and cytoplasmatic translocation of Snail in OSCC.
Introduction
Head and neck cancer is one of the 10 most frequent cancers worldwide and affects .500 000 people each year. Oral squamous cell
carcinoma (OSCC) represents .95% of all head and neck cancer
(1). Therefore, fighting against head and neck cancer primarily means
improving the diagnosis, biology and management of squamous cell
carcinoma. Although surgery, chemotherapy and radiation therapy
have improved over the last 30 years, the 5 year survival rate is still
only 50%. Frequent lymph node metastasis involving migration and
invasion of aberrant cells from the primary neoplasm to distant sites
results in poor prognosis (2). One of the critical steps in cancer metastasis is the transformation of cells from an epithelial to a mesenchymal phenotype [epithelial to a mesenchymal transition (EMT)].
The loss of epithelial features usually involves an increased cell motility and expression of mesenchymal genes like Snail, slug, vimentin
or twist (3–5). Under normal condition, this process occurs frequently
during embryonic development and allows cells to overcome physical
Abbreviations: EC, extracellular cadherin; EMT, epithelial to a mesenchymal
transition; GSK-3beta, glycogen-synthase-kinase-3beta; MMP, matrix metalloproteinase; OSCC, oral squamous cell carcinoma; PBS, phosphate-buffered
saline; PCR, polymerase chain reaction.
constraints like cellular adhesions. In an embryo, this event is highly
controlled in terms of endocrine signaling, autocrine signaling and
coordinated stromal events. Embryonic cells undergoing controlled
EMT are able to adopt a migratory phenotype orderly move to different embryonic sites and build up a proper tissue architecture. However, upon neoplastic transformation, cells that undergo EMT
disseminate in a dysregulated way throughout the body via lymphatic
and blood vessels. The first step in carcinoma EMT is similar to those
that occur during wound healing where basal keratinocytes downregulate their cytokeratin and cadherin settings and adopt a proliferative
and migratory phenotype; these cells have an elongated spindleshaped morphology with front–back asymmetry that facilitates
motility and locomotion. Interestingly, under normal conditions,
cells revert into basal keratinocytes when the wound defect is reepithelialized; in contrast, abnormally transformed keratinocytes do
not regain their epithelial features. In recent years, it has become
evident that cadherins play a major role in regulating EMT (6–9).
Cadherins belong to a family of glycosylated Ca2þ-dependent adhesion molecules. They are single-pass transmembrane proteins that form
homophilic cellular interactions by means of several tandemly repeated
extracellular cadherin (EC) domains (10). There are .80 members of
the cadherin superfamily. Cadherin subfamilies are divided in classical
type I (contain an histidine, alanine, valine amino acid sequence in
EC1) and type II cadherins. Type I and type II cadherins are characterized by five EC repeats EC1–EC5 (11). Intracellularly, cadherins are
bridged to the actin cytoskeleton binding directly and indirectly to
various cytoplasmatic proteins. For cells, it is important to constantly
adapt their binding strength to environmental demands. There are controversial statements concerning cadherin interaction with cytoskeletal
proteins. The classic model proposes that cadherins are linked directly
to the actin cytoskeleton via binding to beta- and alpha-catenin, the
latter linking to actin filaments. The group of Nelson could not confirm
simultaneous interaction between E-cadherin/beta-/alpha-catenin and
actin. They state that binding plasticity can be controlled by monomerization and dimerization of alpha-catenin that regulates binding to betacatenin and actin filaments (12,13). Beta-catenin is known to be
involved in the canonical and non-canonical Wnt signaling and binds
and activates members of the Lef-1/Tcf family of DNA-binding proteins. These are responsible for a diverse array of signaling events
associated with cell cycle activities, proliferation, migration, invasion
and apoptosis. During embryonic development, cadherins control many
morphogenetic processes determining tissue boundaries and separate or
fuse different tissue layers, respectively. In pathological processes, they
play a prominent role in tumor metastasis and cell migration (14). Many
studies have shown that E-cadherin acts as tumor suppressor and is
a key molecule in establishing epithelial cell contacts and maintaining
epithelial tissue integrity. Moreover, it is known that Snail is a potent
repressor of E-cadherin expression. Supporting epithelial integrity via
E-cadherin expression, glycogen-synthase-kinase-3beta (GSK-3beta)
binds to Snail and phosphorylates it at two distinct sites that control
degradation and subcellular localization of Snail (15–19). GSK-3beta is
a serine/threonine kinase that phosphorylates many proteins, including
structural proteins, transcription factors and signaling proteins (20).
Under resting conditions, GSK-3beta is an active kinase taking Snail
and beta-catenin as substrates marking them for degradation. For regulation, GSK-3beta is inactivated by phosphorylation of an N-terminal
serine residue (Ser 9) that occupies the active site of the enzyme preventing substrate processing. GSK-3-inactivating kinases are cyclic
adenosine 3#,5#-monophosphate-dependent protein kinase A, protein
kinase B (Akt), protein kinase C and p90 ribosomal S6 kinase (21).
P-cadherin is another important epithelial molecule being exclusively expressed in the basal layer of epithelia. Its role in tissue integrity and homeostasis is still elusive and our main goal is to
Ó The Author 2009. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
1781
K.Bauer et al.
determine the role of P-cadherin in malignant oral epithelial cells. In
a recent study, we have shown that the generation of truncated P-cadherin is involved in early steps of OSCC. We have observed
that highly dedifferentiated OSCC cells tend to loose full-length
P-cadherin expression and display a mesenchymal fibroblastoid
phenotype. In this study, we demonstrate that forced expression of
full-length P-cadherin activates GSK-3beta and inactivates and translocates phosphorylated Snail to the cytoplasm. This mechanism triggers reexpression of E-cadherin and reverts the mesenchymal
phenotype of OSCC cell lines into an epithelial phenotype.
Material and methods
Cloning
The full-length complementary DNA clone of P-cadherin was obtained from
Imagene, Berlin, Germany (http://www.imagenes-bio.de, IRATp970F11102D
verified sequence, accession number BC040486) in pBluescriptR vector and
subcloned without 5# and 3# untranslated regions in pcDNA 3.1D vector
(pcDNA 3.1 Directional TOPO expression kit, Invitrogen, Karlsruhe,
Germany) cut at restriction sites 5# BamHI and 3# NdeI.
Stable transfections/transient transfections
For transfection experiments, full-length complementary DNA of human
P-cadherin was stably transfected into the OSCC cell line PCI 52 with FuGene
HD Transfection Reagent (Roche, Basel, Switzerland) and transfected PCI 52
cells were selected by picking single cells under microscopic control. Geneticin was used for selection (1.4 mg/ml, Gibco, Karlsruhe, Germany). Stable
mock control cell lines were transfected with pcDNA 3.1D vector without fulllength P-cadherin coding sequence.
For transient transfections, PCI52_PC cells were seeded in six-well plates
and transfected with 2 lg antisense P-cadherin or 2 lg pCMX-PL1-vector with
coding sequence for Snail with FuGene HD Transfection Reagent (Roche). For
control, PCI52_mock cells were transfected with 2 lg pCMX-PL1-vector.
Fig. 1. Immunohistochemical staining of the invasion front of squamous cell carcinoma. Paraffin sections of patients suffering from OSCC show a decrease of
P-cadherin expression at the invasion front (a–c) and at the same time, a strong nuclear expression of the EMT marker Snail (d–f). Magnification: a–b/d–e, 100;
c and f, 400.
Fig. 2. Western blot analysis of P-cadherin overexpression in the OSCC cell line PCI52_PC. The figure demonstrates three stably full-length P-cadherinoverexpressing cell clones (a; PCI52_PC1, PCI52_PC2 and PCI52_PC3) after P-cadherin transfection. Immunohistochemical stainings of PCI52_PC cell lines
show P-cadherin overexpression in the cell membrane (b and c; exemplified by PCI52_PC1 and PCI52_PC2) in contrast to the non-transfected mock cell lines (d;
exemplified by PCI52_mock1). Magnification 200.
1782
P-cadherin induces an epithelial-like phenotype in OSCC
Fig. 3. Bright field images of the OSCC cell line PCI52 after P-cadherin transfection show a conversion of morphology. Cells overexpressing P-cadherin change
from a spindle-shaped mesenchymal (mock control 1 in a and mock control 2 in b) to an epithelial-like brickstone phenotype. (c–e) depict P-cadherinoverexpressing clones PCI52_PC1–PC3. Magnification 200. Immunohistochemical staining of the EMT marker Snail in PCI52_PC cell clones. P-cadherinoverexpressing PCI52 cells indicate a strong decrease of nuclear Snail staining (f and g), whereas mock cells reveal a relatively high Snail expression in the cell
nuclei (h). Magnification 200 and 400.
Forty-eight hours after transfection, the PCI52_PC and PCI52_mock cells were
stained or cell lysates were prepared. Antisense P-cadherin construct was used
as described previously (22,23).
MA) for 30 min at 30°C. The reaction was stopped by the addition of 2
sample buffer and incubation for 5 min at 95°C.
Cell lines and culture conditions
PCI52_PC and PCI52_mock cells were maintained in Dulbecco’s modified
Eagle’s medium (Pan-Biotech, Aidenbach, Germany) with 10% fetal calf serum (Gibco), 5% penicillin/streptomycin (Gibco) and 5% L-glutamine (Gibco).
All cells were grown under selection with Geneticin (1.4 mg/ml).
Western blot analysis
Lysates were prepared from PCI52_PC and PCI52_mock cells using radio
immuno precipitation assay buffer (Sigma). An equivalent amount of total
protein was separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis using a 12.75% resolving gel and transferred onto polyvinylidene
difluoride membrane. Blots were blocked in 3% bovine serum albumin (Sigma)
in phosphate-buffered saline (PBS) (Sigma) containing 0.1% Tween 20
(Sigma) and probed. Primary antibodies were obtained as follows: E-cadherin
(1:5000), N-cadherin (1:5000) and P-cadherin (1:1000) from BD Transduction
Laboratories (Franklin Lakes, NJ) and Snail (1:1000, Abcam), p-GSK-3beta
and GSK-3beta (1:1000) from Cell Signaling Technology, Danvers, MA. All
appropriate secondary antibodies conjugated with horseradish peroxidase were
purchased from Pierce, Rockford, IL. For protein visualization, the Roti Lumin
kit (Roth, Karlsruhe, Germany) was used. Equal loading was verified with
monoclonal antibody to beta-actin (1:10 000, Abcam). Densitometric analysis
was carried out using Bio1-D software (Peqlab, Erlangen, Germany). All experiments were repeated at least three times.
Immunoprecipitation
For immunoprecipitation experiments, 100 lg of cell lysate was precleared with
25 ll sepharose protein A beads (Sigma, St Louis, MO) for 2 h at 4°C with
rotation and then beads were centrifuged. Afterward the precleared protein lysates were immunoprecipitated with Snail antibody 1:100 (Abcam, Cambridge,
UK) for 4 h at 4°C with rotation. The immunoprecipitation mixture was incubated with 50 ll sepharose protein A beads for 90 min at 4°C with rotation. After
centrifugation, the complex was washed three times with radio immuno precipitation assay buffer, dissociated from the beads by addition of 2 sample
buffer and heated for 5 min at 95°C. For protein analysis, samples were loaded
on sodium dodecyl sulfate–polyacrylamide gel electrophoresis gel.
Inhibition of GSK-3beta
PCI52_PC and PCI52_mock cells were incubated with 20 mM lithium chloride
(Merck, Darmstadt, Germany) for 30 h. Then the cells were lysed with radio
immuno precipitation assay buffer and prepared for sodium dodecyl sulfate–
polyacrylamide gel electrophoresis.
Dephosphorylation assay
In total, 50 lg cell lysates were incubated with 400 U of lambda protein
phosphatase in the presence of 2 mM MnCl2 (New England BioLabs, Ipswitch,
Immunohistochemistry
Paraffin-embedded tissue sections were deparaffined in xylene, rehydrated
through a series of graded ethanol and washed in aqua dest. Before endogenous
peroxidase blocking, the slides were heated in 10 mM sodium citrate buffer for
20 min. Anti-P-cadherin antibody (BD Transduction Laboratories) and antiSnail antibody (Abcam) were used as primary antibodies at a dilution of 1:100
for 1 h at 37°C. Afterward the paraffin sections were incubated with EnVision/
horse radish peroxidase solution (DakoCytomation, Glostrup, Denmark) for
30 min and diaminobenzidene solution (DakoCytomation) for 4 min. The cell
1783
K.Bauer et al.
Fig. 4. Bright field images and immunohistochemical stainings of PCI52_mock and PCI52_PC cells transfected transiently with antisense P-cadherin (b–f) and
antisense Snail (h and i). PCI52_PC cells transfected with antisense P-cadherin reveal a clear reversion of morphology back to a mesenchymal phenotype (b and c)
comparable with the mock control (a). Transiently transfected PCI52_PC cell lines partly show nuclear Snail (e and f) comparable with the non-transfected mock
control (d). After transient transfection with antisense Snail, PCI52_mock cells show a reversion of cell morphology to an epithelial-like phenotype (h and i) in
contrast to the non-transfected mock cells with a spindle-shaped phenotype (g). Magnification 200.
nuclei were counterstained with hematoxylin. All images were taken using the
microscope Olympus BX 61.
Immunocytochemistry
Cells were plated on glass coverslips and fixed with 4% paraformaldehyde in
PBS for 10 min. After blocking with peroxidase blocking solution (DAKO) for
10 min and 5% normal goat serum (DakoCytomation) in PBS for 1 h at 37°C,
the following primary antibodies were used for 1 h at 37°C: anti-Snail (1:100,
Abcam) and anti-P-cadherin (1:100, BD Transduction Laboratories). Afterward the cells were incubated with EnVision/horse radish peroxidase solution
(DakoCytomation) for 30 min and diaminobenzidene solution (DakoCytomation) for 4 min. The cell nuclei were counterstained with hematoxylin.
Immunofluorescence
For beta-catenin immunofluorescence, PCI52_PC and PCI52_mock cells were
seeded on glass coverslips. After growing to confluence, cells were washed
with PBS and fixed for 10 min in 4% paraformaldehyde in PBS. Cells were
then permeabilized with 0.1% Triton X-100 for 5 min. After washing with
PBS, cells were blocked with 3% bovine serum albumin (Sigma) in PBS for 1 h
and incubated with beta-catenin antibody (1:500, BD Transduction Laboratories) for 1 h at 37°C, following incubation with the corresponding secondary
antibody Alexa 488 (1:500, Invitrogen) for 1 h at 37°C. Afterward the cells
were washed with PBS and mounted with mounting medium (Vectashield,
Vectorlabs, Peterborough, UK).
Real-time polymerase chain reaction analyses
For real-time polymerase chain reaction (PCR), total RNA was isolated using
RNeasy Mini Kits (Qiagen, Hilden, Germany) according to the manufacturer’s
instructions. Reverse transcription of RNA to complementary DNA was performed using Superscript II Kit (Invitrogen). Quantitative gene expression was
done using the SybrGreen system from Peqlab.
Migration assay
The migration assays were performed using 24-well dishes and ThinCert cell
culture inserts (Greiner Bio-One, Frickenhausen, Germany). As chemoattractant, the lower compartment was filled with Dulbecco’s modified Eagle’s medium containing fibroblast conditioned medium. The upper compartment was
filled with 1 105 cells in Dulbecco’s modified Eagle’s medium and 0.2%
bovine serum albumin. After incubation for 16 h at 37°C, the migrated cells
were harvested by trypsinization, stained and counted.
1784
Genomic DNA isolation, bisulfite modification and analysis of the E-cadherin
promoter by methylation-specific PCR
Genomic DNA was isolated using the Wizard genomic DNA Purifications Kit
(Promega, Mannheim, Germany) according to the manufacturer’s instructions.
Methylation status of the CpG sites in the E-cadherin promoter region was
analyzed based on the principle that bisulfite modification of the genomic DNA
would convert unmethylated cytosine residues to uracil, whereas methylated
cytosine is resistant to the treatment. For bisulfite conversion and postmodification cleanup of DNA samples, the Imprint DNA Modification Kit (Sigma)
was used according to the manufacturer’s instructions. Modified DNA was
amplified using primers specific for the methylated sequence, 5#-TTAGGTTAGAGGGTTATCGCGT-3# and 5#-TAACTAAAAATTCACCTACCGAC-3#,
and for the unmethylated sequence, 5#-TAATTTTAGGTTAGAGGGTTATTGT-3# and 5#-CACAACCAATCAACAACACA-3#. The conditions of
PCR amplification used were 35 cycles composed of 30 s at 94°C, 30 s at
56°C and 30 s at 72°C. PCR products (116 bp for methylated and 97 bp for
unmethylated) were run on 7.5% Tris–borate–ethylenediaminetetraacetic
acid–polyacrylamide gels. CpG-methylated HeLa genomic DNA and HeLa
genomic DNA (New England BioLabs) were utilized as control for the methylated and unmethylated E-cadherin gene.
Results
In our previous study, we have shown that dedifferentiating OSCC cells
gradually loose their membrane-bound full-length P-cadherin that
renders them capable of migration by means of a truncated 50 kDa
form of P-cadherin (24). This raised the question what role fulllength P-cadherin plays in the process of dedifferentiation and the progression of OSCC. To address this issue, in vivo immunohistochemical
stainings of P-cadherin were performed on paraffin sections of patients
suffering from OSCC. Figure 1a–c shows a decrease and loss of
membrane-bound P-cadherin at the invasion front of OSCC. At the
same time, we observed a strong nuclear expression of the EMT marker
Snail (Figure 1d–f). These results and the results of our previous paper
hinted at P-cadherin playing a role in the differentiation process by
preventing EMT of malignant oral epithelial cells. Therefore, by clonal
selection we generated cell lines derived from the original E-, N- and
P-cadherin induces an epithelial-like phenotype in OSCC
Fig. 5. (a and b) Western blot analysis of E- and N-cadherin expression in PCI52_PC cell lines. The figure demonstrates reexpression of E-cadherin in all
P-cadherin-transfected PCI52 cell lines compared with the mock controls (a). Transfection of P-cadherin does not increase expression of N-cadherin in the
PCI52_PC cell clones. PCI 13 shows a well-differentiated OSCC cell line (b). (c) Analysis of E-cadherin promoter methylation in PCI52_PC and PCI52_mock cell
lines. Displayed from left to right are HeLa (control cell line for no promoter methylation), CpG methylated HeLa (control cell line for promoter methylation),
PCI52_mock1-2, PCI52_PC1-3 and negative control. Methylated (M) and non-methylated (NM) templates show amplification fragments of 116 bp and 97 bp,
respectively. Both in PCI52_PC and in PCI52_mock cell lines, no promoter methylation was detectable. The lowest DNA band is caused by primer dimerization.
(d) Quantitative PCR of the mesenchymal marker vitronectin. The figure shows a significant decrease vitronectin expression the P-cadherin-transfected cell lines
in contrast to the mock control cells. (e–g) Immunofluorescent stainings show the distribution of beta-catenin in PCI52_PC versus PCI52_mock cell lines.
PCI52_mock cells display a diffuse nuclear and cytoplasmatic beta-catenin staining, whereas in PCI52_PC cells, beta-catenin is translocated to the cell membrane. (h)
Cell migration assays show a significant slower migration of P-cadherin-transfected cell clones in contrast to non-transfected mock cell lines ( P , 0.05, P , 0.001).
P-cadherin-deficient cell line PCI52. We stably overexpressed full-length
P-cadherin in these cell lines and chose clones that did not produce
any truncated versions of these cadherins (Figure 2a). We denominated
these cell lines PCI52_PC. Immunohistochemical stainings of PCI52_
PC cell lines revealed that P-cadherin was located in the cell membrane
(Figure 2b and c), whereas in mock controls there was no visible Pcadherin expression (Figure 2d). A further conspicuous feature of all
P-cadherin-overexpressing cell lines was a clear change in morphology
compared with the mock controls (Figure 3a–e). Figure 3a and b depicts
the mock-transfected PCI52 cells (PCI52_mock), which display thin,
elongated and spindle-shaped cell morphology. Moreover, the mocktransfected cell lines continued to produce several cell layers even
when grown confluently as opposed to PCI52_PC cell lines that displayed a monolayer after growth arrest under confluent conditions.
After P-cadherin transfection, the cell morphology gained the typical
form of normal squamous epithelial cells, i.e. the cells displayed
brickstone morphology (Figure 3c–e). To analyze Snail expression
in the cell lines, immunohistochemical stainings were performed.
They revealed a strong decrease of nuclear Snail in PCI52_PC cell
lines and a cytoplasmatic translocation (Figure 3f and g). In contrast,
the mock controls yielded a relatively high level of nuclear Snail
expression and no cytoplasmatic staining was observed (Figure 3h).
To underline that these effects resulted from P-cadherin induction,
we transfected PCI52_PC cell lines with antisense P-cadherin. Figure
4b and c shows that antisense P-cadherin-transfected cell clones re-
gained the spindle-shaped morphology (Figure 4b and c) and started
to reexpress Snail in the nucleus (Figure 4e and f) like the mock
controls did (Figure 4a and d). To test the effect of Snail on wild-type
PCI52 cells, PCI52_mock cell lines were transfected with antisense
Snail. These transiently transfected mock cells partly recovered brickstone morphology (Figure 4h and i) in contrast to the control cell lines
(Figure 4g).
To analyze whether cytoplasmic Snail translocation had any effect
on E-cadherin expression, western blot analysis was performed with
cell lysates of P-cadherin-transfected clones and the mock controls.
Our results exhibit a strong reexpression of E-cadherin in all
P-cadherin-induced cell clones (Figure 5a). A change in N-cadherin
expression, however, was not observed (Figure 5b).
To determine the effect of promoter methylation on E-cadherin
expression, the methylation status of PCI52_PC and PCI52_mock
cells was analyzed. Figure 5c shows that in none of the cell lines
promoter methylation was detectable. Apart from E-cadherin reexpression, we analyzed further features revealing that P-cadherin induced an epithelial phenotype in oral keratinocytes. Figure 5d shows
a significant reduction of chemotactic migration in P-cadherintransfected cell lines compared with the controls. Moreover, we have
observed a marked translocation of beta-catenin to the cell membrane
(Figure 5e–g) PCI52_PC cells and quantitative reverse transcription–
PCR revealed a reduced expression of the mesenchymal marker
vitronectin in cells expressing P-cadherin (Figure 5h).
1785
K.Bauer et al.
Fig. 6. (a) Co-immunoprecipitation of GSK-3beta with Snail shows a complex of the proteins in all cell clones. (b) Western blot analysis of serine-9
phosphorylation of GSK-3beta in total cell lysates of PCI52_PC, mock controls and human oral keratinocytes (HOKs) reveals decreased phosphorylation of GSK3beta in PCI52_PC cell lines and human oral keratinocytes meaning increased activity of the kinase in P-cadherin-overexpressing cell lines and primary normal
keratinocytes. (c) Western blot analysis of Snail shows phosphorylated Snail in all P-cadherin-overexpressing cell clones compared with mock control cell lines
meaning Snail is marked for cytoplasmic shuttling and degradation. (d) Western blot analysis of Snail demonstrates a markedly reduced Snail phosphorylation in
PCI52_PC cells after inhibition of GSK-3beta activity with 20 mM LiCl compared with untreated PCI52_PC cells. Densitometric analysis was performed as
described in Materials and Methods. LP 5 lambda phosphatase. (e) Western blot analysis of E-cadherin in Snail-overexpressing PCI52_PC cell lines. The figure
shows less E-cadherin expression in all Snail-overexpressing cell lines. Interestingly, bands of E-cadherin shedding (at 65, 55, 40 and 35 kDa) appear only in Snailoverexpressing cell lines (arrows). Densitometric evaluation was done for the full-length version of E-cadherin. Snail OE 5 Snail overexpression.
Because it is known from the literature that the observed nuclear/
cytoplasmic Snail shuttling is facilitated by GSK-3beta-mediated
phosphorylation, we analyzed whether Snail binds GSK-3beta in
P-cadherin-overexpressing OSCC cell lines. The coimmunoprecipitation experiment in Figure 6a demonstrates that Snail binds to
GSK-3beta in all PCI52_PC and mock cell clones. We speculated if
PCI52_PC showed a different activation status compared with the
mock control cell lines. Interestingly, western blot analysis with an
antibody directed against GSK-3beta serine-9 phosphorylation revealed a significantly decreased phosphorylation of GSK-3beta in cell
lysates of all P-cadherin-overexpressing cell lines and in the primary
human oral keratinocytes in contrast to the mock controls. In sum-
1786
mary, these results show more activated GSK-3beta in PCI52_PC
cells and human oral keratinocytes than in the mock controls (Figure
6b). As a consequence of GSK-3beta activation, we speculated if
Snail was phosphorylated in PCI52_PC cell lines. As expected, western blot analysis revealed a band that migrated slower than the Snail
protein in all P-cadherin-overexpressing cell lines as opposed to the
mock controls (Figure 6c). To confirm that the slow migrating band
was phosphorylated, Snail cell lysates of PCI52_PC were treated with
lambda protein phosphatase. Western blot analysis verified Snail
phosphorylation because the slow migrating band disappeared (Figure
6d, last lane). To show that GSK-3beta was responsible for Snail
phosphorylation in the P-cadherin-overexpressing cell lines, cells
P-cadherin induces an epithelial-like phenotype in OSCC
were treated with 20 mM lithium chloride to inhibit GSK-3beta activity. Western blot and densitometric evaluation in Figure 6d show
a significant decrease in Snail phosphorylation after blocking GSK3beta activity in PCI52_PC cell lines. This means that Snail is marked
for cytoplasmatic shuttling and possible proteasomal degradation in
P-cadherin-overexpressing cell lines by phosphorylation via activated
GSK-3beta. To confirm that Snail expression has an influence on
E-cadherin expression, we overexpressed Snail in PCI52_PC cell
lines. The western blot analysis and densitometric analysis in Figure 6e
show that Snail overexpression led to a decrease of E-cadherin expression on protein level. Interestingly, all Snail-expressing cell lines
showed E-cadherin shedding (see arrows Figure 6e).
In summary, our results show reduced migration, translocation of
beta-catenin to the cell membrane and downregulation of vitronectin
expression in cell lines transfected with full-length P-cadherin. Furthermore, our data reveal that forced expression of P-cadherin leads to
a nuclear decrease of the E-cadherin repressor Snail via binding to
GSK-3beta and subsequent cytoplasmatic translocation. Additionally,
E-cadherin expression is induced and OSCC can revert from a mesenchymal to an epithelial-like phenotype.
Discussion
In our study, we demonstrate that forced expression of P-cadherin
leads to a cytoplasmatic translocation of the mesenchymal marker
Snail that triggers the reexpression of E-cadherin in malignant oral
keratinocytes and reverts them from a mesenchymal to an epitheliallike phenotype. E-cadherin is uniquely expressed in cells throughout
the body mainly in epithelial keratinocytes where it controls tissue
homeostasis. Usually, low E-cadherin or a switch from E- to Ncadherin expression is associated with adverse morphological and
clinicopathological features, for example, in malignant melanoma
and prostate cancer (25–28). N-cadherin expression does not seem
to be involved in EMT of OSCC because only trace amounts of
N-cadherin expression were detected in P-cadherin-transfected cell
clones and mock control cell lines.
We have found that P-cadherin expression is reduced at the invasive
front of OSCC in vivo, which confirms findings of us and other groups
showing that decreased P-cadherin expression or truncation is associated with low differentiation and poorer overall and disease-free
survival rate (23,29–31). Our studies on OSCC lines deficient of
classical E-, N- and P-cadherin showed strong nuclear Snail staining.
Snail is a labile protein with a short half-life (16). It plays a major role
in normal EMT processes like gastrulation and neural crest formation.
Moreover, it is known to be one of the molecules involved in dedifferentiation and EMT during tumor progression (32). Lyons et al.
(33) showed that Snail contributes to malignancy and negatively affects differentiation in oral keratinocytes. We found that our control
mock cell lines revealed a significant inactivation of GSK-3beta visible by phospho-serine-9-immunodetection. In contrast, primary human oral keratinocytes showed a marked attenuation of GSK-3beta
inactivation. Inducing P-cadherin expression in OSCC cell lines also
led to an attenuation of GSK-3beta inactivation. GSK-3beta inactivation by phosphorylating its serine-9 residue has been shown to be
involved in tumor progression. In laterally spreading colorectal tumors, phospho-GSK-3beta was significantly increased (34). In mouse
F9 teratocarcinoma cells, p38 mitogen-activated protein-kinase
regulates the Wnt-beta-catenin signaling pathway by inhibition
of GSK-3beta (35). Moreover, Mullholand et al. (36) showed that
phospho-serine-9-inactivation of GSK-3beta is a key regulatory element in the progression of prostate cancer. The upstream mediators of
GSK-3beta phosphorylation in OSCC are still elusive. Recently, Hong
et al. (37) demonstrated that Akt inhibition induced EMT and restored
E-cadherin expression in OSCC. Therefore, this kinase might be
a suitable candidate for further investigations upstream of GSK-3beta
linking to P-cadherin. Furthermore, overexpressing P-cadherin in
OSCC cell lines led to Snail phosphorylation and its cytoplasmatic
translocation. Zhou et al. (16) demonstrated that Snail is a GSK-3beta
substrate and can be dually phosphorylated leading to subcellular
translocation and proteasomal degradation controlling EMT. Yook
et al. (19) corroborated these findings showing that Snail harbors
a beta-catenin like consensus motif in the N-terminal domain that
supports its GSK-3beta-dependent phosphorylation, ubiquitination
and subsequent proteasomal degradation
Lithium chloride which is a direct inhibitor of GSK-3 affirms the
involvement of GSK-3beta in Snail phosphorylation because lithium
chloride treatment resulted in decrease of Snail phosphorylation.
Lithium chloride directly inhibits the kinase by inducing GSK-3
N-terminal phosphorylation that causes kinase inactivation (38). It is
well known from the literature that the transcription factor Snail is
a repressor of E-cadherin transcription (39). Batlle et al. demonstrated
that that Snail represses E-cadherin gene expression in epithelial tumors. They clearly revealed that Snail binds to three E-boxes present
in the human E-cadherin promoter (40). Therefore, it is unlikely that
E-cadherin reexpression is due to promoter demethylation after
cytoplasmic Snail translocation. This is confirmed by our showing no
E-cadherin promoter methylation in the cell lines we used. Forced
Snail expression in these cell lines leads to a decrease of E-cadherin
on protein level. Interestingly, all Snail-transfected PCI52_PC cell
lines showed markedly increased E-cadherin shedding in their cell
lysates. This result suggests that in OSCC Snail might not necessarily
be responsible for suppressing E-cadherin on the transcriptional level
but might also be able to initiate posttranslational truncation of
E-cadherin. It is known that E-cadherin shedding plays a role in tumor
progression (41). Our data reveal binding of GSK-3beta to Snail in
OSCC cells; therefore, this complex might be able to trigger
E-cadherin shedding via induction of matrix metalloproteinase (MMPs).
Miyoshi et al. and other groups showed that MMP-1, MMP-2, MMP-7
and MT1-MMP are strongly upregulated by Snail in hepatocellular
carcinoma cells (42,43). MMP-7 has been shown to be responsible for
E-cadherin shedding (44). Moreover, Noe et al. (45,46) demonstrated
that matrilysin-1 (MMP-7) and stromelysin-1 (MMP-3) can cleave
E-cadherin ectodomain. Additionally, a disintegrin and metalloproteinases have been shown to cleave E-cadherin at distinct sites.
ADAM15 is important in the progression of prostate and breast cancer
by shedding E-cadherin (47). Moreover, it has been revealed that
ADAM10 mediates the truncation of E-cadherin (48). Surely, further
studies will reveal the role of Snail, its impact on proteases in OSCC
and their relevance in cadherin shedding.
Funding
Deutsche Forschungsgemeinschaft (BA3696/1-1) to R.B.
Acknowledgements
Conflict of Interest Statement: None declared.
References
1. Chen,S.C. et al. (2008) Plasminogen activator inhibitor-1 inhibits prostate
tumor growth through endothelial apoptosis. Mol. Cancer Ther., 7, 1227–
1236.
2. Bray,F. et al. (2002) Estimates of cancer incidence and mortality in Europe
in 1995. Eur. J. Cancer, 38, 99–166.
3. Yanjia,H. et al. (2007) The role of epithelial-mesenchymal transition in oral
squamous cell carcinoma and oral submucous fibrosis. Clin. Chim. Acta,
383, 51–56.
4. Rosivatz,E. et al. (2002) Differential expression of the epithelialmesenchymal transition regulators snail, SIP1, and twist in gastric cancer.
Am. J. Pathol., 161, 1881–1891.
5. Savagner,P. (2001) Leaving the neighborhood: molecular mechanisms involved during epithelial-mesenchymal transition. Bioessays, 23, 912–923.
6. Baum,B. et al. (2008) Transitions between epithelial and mesenchymal
states in development and disease. Semin. Cell Dev. Biol., 19, 294–308.
7. Nakaya,Y. et al. (2008) Epithelial to mesenchymal transition during
gastrulation: an embryological view. Dev. Growth Differ., 50, 755–766.
8. Wu,Y. et al. (2008) New insights of epithelial-mesenchymal transition in
cancer metastasis. Acta Biochim. Biophys. Sin. (Shanghai), 40, 643–650.
1787
K.Bauer et al.
9. Moustakas,A. et al. (2007) Signaling networks guiding epithelialmesenchymal transitions during embryogenesis and cancer progression.
Cancer Sci., 98, 1512–1520.
10. Gumbiner,B.M. (2005) Regulation of cadherin-mediated adhesion in
morphogenesis. Nat. Rev. Mol. Cell Biol., 6, 622–634.
11. Hirano,S. et al. (1987) Calcium-dependent cell-cell adhesion molecules
(cadherins): subclass specificities and possible involvement of actin
bundles. J. Cell Biol., 105, 2501–2510.
12. Drees,F. et al. (2005) Alpha-catenin is a molecular switch that binds
E-cadherin-beta-catenin and regulates actin-filament assembly. Cell, 123,
903–915.
13. Hartsock,A. et al. (2008) Adherens and tight junctions: structure, function
and connections to the actin cytoskeleton. Biochim. Biophys. Acta, 1778,
660–669.
14. Wheelock,M.J. et al. (2003) Cadherins as modulators of cellular
phenotype. Annu. Rev. Cell Dev. Biol., 19, 207–235.
15. Doble,B.W. et al. (2007) Role of glycogen synthase kinase-3 in cell fate and
epithelial-mesenchymal transitions. Cells Tissues Organs, 185, 73–84.
16. Zhou,B.P. et al. (2004) Dual regulation of Snail by GSK-3beta-mediated
phosphorylation in control of epithelial-mesenchymal transition. Nat. Cell
Biol., 6, 931–940.
17. Schlessinger,K. et al. (2004) GSK-3beta sets Snail’s pace. Nat. Cell Biol., 6,
913–915.
18. Bachelder,R.E. et al. (2005) Glycogen synthase kinase-3 is an endogenous
inhibitor of Snail transcription: implications for the epithelial-mesenchymal
transition. J. Cell Biol., 168, 29–33.
19. Yook,J.I. et al. (2005) Wnt-dependent regulation of the E-cadherin repressor snail. J. Biol. Chem., 280, 11740–11748.
20. Woodgett,J.R. (2001) Judging a protein by more than its name: GSK-3. Sci.
STKE, 2001, RE12.
21. Doble,B.W. et al. (2003) GSK-3: tricks of the trade for a multi-tasking
kinase. J. Cell Sci., 116, 1175–1186.
22. Bauer,R. et al. (2006) Functional implication of truncated P-cadherin
expression in malignant melanoma. Exp. Mol. Pathol., 81, 224–230.
23. Bauer,R. et al. (2006) Prognostic relevance of P-cadherin expression in
melanocytic skin tumours analysed by high-throughput tissue microarrays.
J. Clin. Pathol., 59, 699–705.
24. Bauer,R. et al. (2008) Truncated P-cadherin is produced in oral squamous
cell carcinoma. FEBS J., 275, 4198–4210.
25. Margineanu,E. et al. (2008) Correlation between E-cadherin abnormal
expressions in different types of cancer and the process of metastasis.
Rev. Med. Chir. Soc. Med. Nat. Iasi., 113, 432–436.
26. Kuphal,S. et al. (2006) Influence of the cytoplasmic domain of E-cadherin
on endogenous N-cadherin expression in malignant melanoma. Oncogene,
25, 248–259.
27. Haass,N.K. et al. (2004) The role of altered cell-cell communication in
melanoma progression. J. Mol. Histol., 35, 309–318.
28. Gravdal,K. et al. (2007) A switch from E-cadherin to N-cadherin expression indicates epithelial to mesenchymal transition and is of strong and
independent importance for the progress of prostate cancer. Clin. Cancer
Res., 13, 7003–7011.
29. Lo Muzio,L. et al. (2004) P-cadherin expression predicts clinical outcome
in oral squamous cell carcinomas. Histol. Histopathol., 19, 1089–1099.
30. Lo Muzio,L. et al. (2005) P-cadherin expression and survival rate in oral
squamous cell carcinoma: an immunohistochemical study. BMC Cancer, 5, 63.
1788
31. Munoz-Guerra,M.F. et al. (2005) P-cadherin expression reduced in squamous cell carcinoma of the oral cavity: an indicator of poor prognosis.
Cancer, 103, 960–969.
32. Nieto,M.A. (2002) The snail superfamily of zinc-finger transcription
factors. Nat. Rev. Mol. Cell Biol., 3, 155–166.
33. Lyons,J.G. et al. (2008) Snail up-regulates proinflammatory mediators
and inhibits differentiation in oral keratinocytes. Cancer Res., 68,
4525–4530.
34. Wang,J. et al. (2009) Increased expression of beta-catenin, phosphorylated
glycogen synthase kinase 3beta, cyclin D1, and c-myc in laterally spreading
colorectal tumors. J. Histochem. Cytochem., 57, 363–371.
35. Bikkavilli,R.K. et al. (2008) p38 mitogen-activated protein kinase regulates
canonical Wnt-beta-catenin signaling by inactivation of GSK3beta. J. Cell
Sci., 121, 3598–3607.
36. Mulholland,D.J. et al. (2006) PTEN and GSK3beta: key regulators of
progression to androgen-independent prostate cancer. Oncogene, 25,
329–337.
37. Hong,K.O. et al. (2009) Inhibition of Akt activity induces the mesenchymal-to-epithelial reverting transition with restoring E-cadherin expression
in KB and KOSCC-25B oral squamous cell carcinoma cells. J. Exp. Clin.
Cancer Res., 28, 28.
38. Zhang,F. et al. (2003) Inhibitory phosphorylation of glycogen synthase
kinase-3 (GSK-3) in response to lithium. Evidence for autoregulation of
GSK-3. J. Biol. Chem., 278, 33067–33077.
39. Thuault,S. et al. (2008) HMGA2 and Smads co-regulate SNAIL1 expression during induction of epithelial-to-mesenchymal transition. J. Biol.
Chem., 283, 33437–33446.
40. Batlle,E. et al. (2000) The transcription factor snail is a repressor of
E-cadherin gene expression in epithelial tumour cells. Nat. Cell Biol., 2,
84–89.
41. Masterson,J. et al. (2007) Posttranslational truncation of E-cadherin and
significance for tumour progression. Cells Tissues Organs, 185, 175–179.
42. Miyoshi,A. et al. (2004) Snail and SIP1 increase cancer invasion by
upregulating MMP family in hepatocellular carcinoma cells. Br. J. Cancer,
90, 1265–1273.
43. Yokoyama,K. et al. (2003) Increased invasion and matrix metalloproteinase2 expression by Snail-induced mesenchymal transition in squamous cell
carcinomas. Int. J. Oncol., 22, 891–898.
44. McGuire,J.K. et al. (2003) Matrilysin (matrix metalloproteinase-7)
mediates E-cadherin ectodomain shedding in injured lung epithelium.
Am. J. Pathol., 162, 1831–1843.
45. Noe,V. et al. (2001) Release of an invasion promoter E-cadherin fragment
by matrilysin and stromelysin-1. J. Cell Sci., 114, 111–118.
46. Lochter,A. et al. (1997) Matrix metalloproteinase stromelysin-1 triggers a cascade of molecular alterations that leads to stable epithelial-to-mesenchymal
conversion and a premalignant phenotype in mammary epithelial cells. J. Cell
Biol., 139, 1861–1872.
47. Najy,A.J. et al. (2008) The ectodomain shedding of E-cadherin by
ADAM15 supports ErbB receptor activation. J. Biol. Chem., 283, 18393–
18401.
48. Maretzky,T. et al. (2005) ADAM10 mediates E-cadherin shedding and
regulates epithelial cell-cell adhesion, migration, and beta-catenin
translocation. Proc. Natl Acad. Sci. USA, 102, 9182–9187.
Received April 10, 2009; revised June 7, 2009; accepted July 5, 2009