Cadherin 6 Is a New RUNX2 Target in TGF-β Signalling
`Pathway
`Valentina Sancisi1, Greta Gandolfi1, Moira Ragazzi2, Davide Nicoli1, Ione Tamagnini2, Simonetta Piana2,
`Alessia Ciarrocchi1*
`
`1 Laboratory of Molecular Biology, Department of Oncology and Advanced Technologies, Azienda Ospedaliera Arcispedale S. Maria Nuova-IRCCS, Reggio
`Emilia, Italy, 2 Pathology Unit, Department of Oncology and Advanced Technologies, Azienda Ospedaliera Arcispedale S. Maria Nuova-IRCCS, Reggio Emilia,
`Italy
`
`Abstract
`
`Modifications in adhesion molecules profile may change the way tumor cells interact with the surrounding
`microenvironment. The Cadherin family is a large group of transmembrane proteins that dictate the specificity of the
`cellular interactions. The Cadherin switch that takes place during epithelial-mesenchymal transition (EMT) contributes
`to loosening the rigid organization of epithelial tissues and to enhancing motility and invasiveness of tumor cells.
`Recently, we found Cadherin-6 (CDH6, also known as K-CAD) highly expressed in thyroid tumor cells that display
`mesenchymal features and aggressive phenotype, following the overexpression of the transcriptional regulator Id1. In
`this work, we explored the possibility that CDH6 is part of the EMT program in thyroid tumors. We demonstrate that
`CDH6 is a new transforming growth factor-β (TGF-β) target and that its expression is modulated similarly to other
`EMT mesenchymal markers, both in vitro and in thyroid tumor patients. We show for the first time that CDH6 is
`expressed in human thyroid carcinomas and that its expression is enhanced at the invasive front of the tumor.
`Finally, we show that CDH6 is under the control of the transcription factor RUNX2, which we previously described as
`a crucial mediator of the Id1 pro-invasive function in thyroid tumor cells. Overall, these observations provide novel
`information on the mechanism of the EMT program in tumor progression and indicate CDH6 as a potential regulator
`of invasiveness in thyroid tumors.
`
`Citation: Sancisi V, Gandolfi G, Ragazzi M, Nicoli D, Tamagnini I, et al. (2013) Cadherin 6 Is a New RUNX2 Target in TGF-β Signalling Pathway. PLoS
`ONE 8(9): e75489. doi:10.1371/journal.pone.0075489
`Editor: Jun Li, Sun Yat-sen University Medical School, China
`Received April 12, 2013; Accepted August 15, 2013; Published September 12, 2013
`Copyright: © 2013 Sancisi et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
`unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
`Funding: This work was supported by grants from the Italian Association for Cancer Research (AIRC, MFAG:10745)(http://www.airc.it/) and the Guido
`Berlucchi Foundation (http://www.fondazioneberlucchi.com/). The funders had no role in study design, data collection and analysis, decision to publish, or
`preparation of the manuscript.
`Competing interests: The authors have declared that no competing interests exist.
`* E-mail: Alessia.Ciarrocchi@asmn.re.it
`
`Introduction
`
`The transdifferentiation of epithelial tumor cells towards a
`mesenchymal condition is a complex process that allows tumor
`cells to leave their original site and to invade adjacent tissues.
`During this transition (also known as epithelial-mesenchymal
`transition - EMT), the epithelial cells shed their differentiated
`characteristics, including cell-cell adhesion, polarity, and lack of
`motility, and acquire instead mesenchymal features, including
`motility,
`invasiveness, and resistance
`to apoptosis. The
`importance of the EMT program in mediating epithelial cancer
`progression is supported by a massive amount of evidence that
`has been published on this topic over the last 15 years [1-5].
`While the relevance of this process to the biology of tumors is
`well established and fully accepted, the complexity of the
`molecular events and regulatory pathways at the basis of EMT
`is far from being fully understood. Furthermore, some of the
`
`molecular players involved in this process remain unknown.
`The first functional consequence of EMT program activation is
`the alteration of the epithelial tumor cell interactions with the
`surrounding microenvironment. The epithelial adhesion
`molecules, in particular the E-Cadherin, are displaced by the
`multiprotein complexes at the adherent junctions and are
`substituted by mesenchymal Cadherins (such as N-Cadherin).
`This alteration signals within
`the cells
`triggering
`the
`complicated cytoskeleton rearrangement that it is necessary to
`support cell motility [6]. The E-Cadherin and N-Cadherin are
`the prototypes and by far the most studied members of the
`large Cadherin family, which includes over 50 proteins in
`vertebrates and non-vertebrate organisms. Some evidence
`suggests that different Cadherins play non-redundant roles in
`cells and it is commonly believed that such large variability
`originates from the need of complex organisms to specifically
`differentiate intercellular interactions [7]. Despite this, the
`
`PLOS ONE | www.plosone.org
`
`1
`
`September 2013 | Volume 8 | Issue 9 | e75489
`
`1 of 16
`
`OnCusp
`Ex. 1027
`
`

`

`potential role of Cadherins other than the E- and N-Cadherin in
`cancer development and progression has being
`rarely
`investigated. Recently, we have shown that the aggressive
`phenotype induced by the transcription regulator Id1 in thyroid
`tumor cells is accompanied by acquisition of mesenchymal
`features and by deregulation of over 400 genes, most of which
`are known to be deregulated or partake in the EMT program.
`Among the Id1 target genes, the Cadherin-6 (CDH6, also
`known as K-Cadherin) was strongly induced by Id1 in thyroid
`tumor cells [8]. CDH6 is a class II Cadherin, which is expressed
`mainly in kidney and central nervous system [9-11]. CDH6 is
`highly expressed in renal tissue during embryogenesis in which
`it drives the mesenchymal-epithelial differentiation that is
`necessary for kidney morphogenesis [9,12,13]. In spite of its
`role
`in
`promoting
`the
`epithelial
`phenotype
`during
`embryogenesis, CDH6 has been described as strongly
`expressed in ovarian cancer and renal carcinoma [14,15]. In
`the latter, CDH6 expression has been shown to strongly
`correlate with aggressive tumor behavior and poorer patient
`outcome [16,17]. To date, nothing is known regarding the
`involvement of CDH6 in the EMT program during epithelial
`tumor development and progression.
`In
`this work, we
`investigated the ability of normal and tumor thyroid cells to
`activate the EMT program in response to transforming growth
`factor-β (TGF-β) and we explored the possibility that CDH6 is a
`TGF-β target during EMT in thyroid tumors. Intriguingly, we
`found that thyroid tumor cells constitutively display markers of
`an active EMT process as compared to normal thyrocytes, both
`in vitro and in human patients. We also showed that CDH6 is
`strongly induced by TGF-β treatment both in normal and tumor
`thyroid cells, and that its expression accompanies invasiveness
`in human thyroid tumor patients.
`
`Materials and Methods
`
`Cell cultures and treatments
`B-CPAP, TPC1, and WRO human cell lines were obtained
`from Dr. Massimo Santoro, University of Naples (Naples, Italy)
`[18,19,20,21]. Nthy.ori 3.1 and FTC133 cell
`lines were
`purchased from Sigma-Aldrich (Milan, Italy). All cell lines were
`grown at 37°C/5% CO2. B-CPAP, TPC1, and WRO were grown
`in DMEM supplemented with 10% fetal bovine serum. Nthy.ori
`3.1 were grown in RPMI supplemented with 10% fetal bovine
`serum. FTC133 were grown in DMEM:Ham’s F12 (1:1)
`supplemented with 10% fetal bovine serum. Id1- and RUNX2-
`overexpressing B-CPAP clones and control clones were grown
`in presence of 400 µg/ml geneticin (Life Technologies, Monza,
`Italy).
`Cell lines were starved with culture medium containing 1%
`fetal bovine serum 16-18 hours before TGF-β treatment. TGF-β
`(Peprotech, Rocky Hill, NJ) was added in starvation medium at
`concentrations of 5 ng/ml and 100 ng/ml for the indicated
`periods of time.
`Italy) and
`(Sigma-Aldrich, Milan,
`For Actinomycin D
`Cycloheximide (Sigma-Aldrich, Milan, Italy) treatments, cells
`were starved with 1% fetal bovine serum, were treated after
`16-18 hours with Actinomycin D 5 mg/ml or cycloheximide 50
`
`CDH6 in TGF-β Mediated EMT Program
`
`mg/ml for 4 hours, then TGF-β was added at 100 ng/ml for 24
`hours.
`For TGF-β inhibitor experiments, BCPAP and TPC1 cell lines
`were starved over night and treated with 10mM of SB-431542
`(Abcam, Cambridge, UK) for 24h or 48h. For cells treated for
`48h, fresh inhibitor was added after 24h by replacing the
`medium.
`
`RNA extraction from formalin-fixed paraffin-embedded
`tissues
`Total RNA was collected from 5 slices, 5 µm thick, of
`formalin-fixed and paraffin-embedded PTCs using the High
`Pure RNA paraffin kit (Roche, Milan, Italy). All samples
`(normal, primary
`tumor, and metastasis) were manually
`dissected under microscopic guidance by two pathologists (MR
`and SP). In order to minimize the biases that a different tissue
`processing may introduce in RNA quality, we performed gene
`expression analysis comparing normal tissue, primary tumor,
`and metastasis from the same patient. Primary tumor and
`normal tissues were collected from the same slide. Quantity
`and purity of the total RNA were checked using the NanoDrop
`2000 spectrophotometer (Thermo Scientific, Waltham, MA).
`Primers for qRT-PCR were designed in order to obtain
`amplicons less than 100 bp in length.
`
`RNA extraction and quantitative real time-PCR (qRT-
`PCR)
`Total RNA purification from cells was performed with
`RNAeasy Mini kit (Qiagen, Milan, Italy). 500 ng of total RNA
`was retrotranscribed using the iScript cDNA kit (Biorad,
`Segrate, Italy). qRT-PCR was conducted using Sso Fast
`EvaGreen Super Mix (BioRad, Segrate, Italy) in the CFX96
`Real Time PCR Detection System (BioRad, Segrate, Italy).
`Relative expression of target genes was calculated using the
`ΔΔCt method by normalizing to the geometric mean of three
`reference genes expression: Glyceraldehyde 3-phosphate
`dehydrogenase (GAPDH), Cyclophilin A (CYPA) and Beta-D-
`Glucuronidase (GUSB) unless otherwise specified. Sequences
`of primers are listed in Table S1.
`
`Western blot
`For Western blot analysis, cells were lysed in RIPA buffer.
`Equal amounts of protein extracts were analyzed by SDS-
`PAGE using
`the Bio-Rad
`(Segrate,
`Italy) Mini-Protean
`apparatus. Staining was performed with the ECL Western blot
`detection reagent (GE Healthcare, Milan,
`Italy). Primary
`antibodies were mouse anti-E-CAD
`(BD Biosciences,
`Buccinasco,
`Italy), mouse anti-N-CAD
`(BD Biosciences,
`Buccinasco,
`Italy), rabbit anti-FN 1 (H-300, Santa Cruz
`Biotechnology, Santa Cruz, CA, USA), mouse anti-pERK (E-4,
`Santa Cruz Biotechnology, Santa Cruz, CA, USA), rabbit anti-
`pAKT (Santa Cruz Biotechnology, Santa Cruz, CA, USA), and
`mouse anti-Actin (AC-15, Sigma-Aldrich, Milan, Italy). For
`CDH6 we tried several antibodies from several dealers (Abcam
`ab64917 and ab71434, Santa Cruz sc-59974, Epitomics
`T0233). None of these antibodies gave reliable results in either
`western blot or immunohistochemistry. Secondary antibodies
`
`PLOS ONE | www.plosone.org
`
`2
`
`September 2013 | Volume 8 | Issue 9 | e75489
`
`2 of 16
`
`OnCusp
`Ex. 1027
`
`

`

`CDH6 in TGF-β Mediated EMT Program
`
`were HRP-conjugated anti-rabbit (GE Healthcare, Milan, Italy)
`and anti-mouse (GE Healthcare, Milan, Italy).
`
`Additional materials and methods are provided as supporting
`information.
`
`Patient samples and immunohistochemistry
`All PTC samples (N=15) and matched LNMs (N=7) were
`retrieved from the archive of the Pathology Unit of Arcispedale
`S. Maria Nuova. Description of the clinicopathological features
`of the analyzed PTC patients is provided in Table S2. Tissue
`specimens were fixed in 10% formalin, paraffin embedded, and
`cut in 4 µm thick sections. For the immunohistochemistry
`analysis, we used the mouse monoclonal anti-CDH6 antibody
`(HPA007047) from Sigma-Aldrich (Milan, Italy), which resulted
`in a strong membrane staining. As control of the specificity of
`the staining, we tested the antibody on renal carcinomas. No
`specific staining was observed in 3 mammary carcinoma
`samples. As indicated in the datasheet, this antibody does not
`work in western blot. Images were captured using a Nikon
`Eclipse E80 microscope (Nikon Instruments, Calenzano, Italy).
`
`Ethics statement
`This project was approved by the “Comitato etico provinciale
`di Reggio Emilia” (local ethics committee of Reggio Emilia).
`Written informed consent was obtained by all the patients
`involved in this project. The local ethics committee approved
`the entire procedure of informed consent collection and
`patients data managing. This study has been conducted
`according to the Declaration of Helsinki.
`
`Immunofluorescence
`Cells were grown in 4 well Lab-Tek Chamber slides
`(NunThermo Scientific, Waltham, MA, USA). After appropriate
`treatments, cells were fixed in 4% PFA in PBS 1X for 15
`minutes at room temperature, permeabilized with 0.1% Triton
`in PBS 1X for 2 minutes, blocked with 20% FBS and 2% BSA
`in PBS 1X for 1 hour, then incubated with a mouse anti-
`Smad2/3 antibody (C-8, Santa Cruz Biotechnology, Santa
`Cruz, CA, USA) in a humidified chamber for 1 hour. Antibody
`binding was revealed with a secondary anti-mouse Alexa 488
`conjugated antibody (Life Technologies, Monza, Italy). Cell
`nuclei were stained with DAPI (Life Technologies, Monza,
`Italy). Slides were mounted using the SlowFade mounting
`medium (Life Technologies, Monza, Italy) and observed using
`an Axiophot fluorescent microscope (Zeiss, Arese, Italy).
`
`Small interfering RNA (siRNA) transfection
`Stealth RNA interference oligos against RUNX2 and control
`oligos were purchased from Life Technologies (Monza, Italy),
`and 30 nM of either anti-RUNX2 or control siRNA was
`transfected using the RNAiMax Lipofectamine reagent (Life
`Technologies, Monza, Italy) using the reverse transfection
`protocol.
`
`Statistical analysis
`Statistical analysis was performed using GraphPad Prism
`Software (GraphPad, San Diego, California, USA). When
`statistical analyses were performed, the specific tests used are
`indicated in the figure legend.
`
`Results
`
`CDH6 is a new TGF-β target gene
`In order to investigate the effect of the TGF-β in thyroid
`tumor progression and the possibility that CDH6 is one of the
`TGF-β targets, five different thyroid-derived cell lines were
`selected and used as a model. The origin and specific features
`of the cell lines used are summarized in Table 1. Two different
`doses of TGF-β (5 ng/ml and 100 ng/ml) were given to cells for
`24h, after which changes
`in
`the expression
`levels of
`established EMT markers and CDH6 were measured by means
`of qRT-PCR (Figure 1). Two epithelial markers - E-CAD and
`Cadherin-16 (CDH16) - and four mesenchymal markers - N-
`CAD, Tenascin C (TNC), Vimentin (VIM) and Fibronectin 1
`(FN1) - were analyzed to monitor the activation of the EMT
`program [8,22-24]. All cell lines responded to TGF-β to some
`extent (Figure 1 A-E). However, the strength of the response
`was different between the Nthy.ori 3.1 thyrocytes and the tumor
`cell lines. In the Nthy. ori3.1 cells, E-CAD was significantly
`repressed, by the TGF-β treatment, while N-CAD, TNC, VIM,
`and FN 1 expression were considerably increased (Figure 1E).
`In contrast, TGF-β treatment induced only a modest effect on
`the four tumor-derived cell lines analyzed (Figure 1A-D). No E-
`CAD expression could be detected in any condition in B-CPAP,
`TPC1, and WRO cells, whereas CDH16 expression was
`detected in all cell lines and was significantly repressed only in
`TPC1 cells. Western Blot analysis of some of these markers in
`TGF-β treated Nthy.ori 3.1, B-CPAP and TPC1 cells confirmed
`the validity of the qRT-PCR results (Figure 1F). TGF-β
`treatment on the metastatic FTC133 cells resulted in increased
`E-CAD expression. This may be consistent with the need of
`metastatic cells to reverse the EMT phenotype necessary for
`metastatic site colonization [1]. Noticeably, CDH6 was strongly
`induced in all cell lines analyzed, with the exception of the
`WRO cells. While in tumor-derived cells CDH6 expression
`increased two-threefold after TGF-β treatment, in Nthy.ori 3.1
`CDH6 induction was much more pronounced (eight-fifteen
`fold), similar to the one observed for TNC and superior to the
`induction observed for N-CAD and FN 1. These results
`demonstrate that CDH6 is a TGF-β target in thyroid cells and
`that its expression is modulated similarly to other mesenchymal
`markers. With the intent to understand whether this could be a
`thyroid specific mechanism, we have extended this analysis to
`other types of tumor. To this purpose, A549 lung cancer, A375
`melanoma and MDA-MB-231 breast cancer cell lines were
`treated with TGF-β and expression of CDH6 was monitored by
`qRT-PCR. We were not able to detect CDH6 expression in
`either A375 or MDA-MB-231 cell lines (data not shown). In
`A549 lung cancer cell line CDH6 was expressed and induced
`upon TGF-β treatment at levels comparable to the thyroid
`tumor cell lines (Figure S1). These results suggest that CDH6
`expression is tissue specific and the ability of TGF-β to target
`this gene is common to other types of tumor.
`All together, our observations suggested that tumor-derived
`cells are less responsive to the TGF-β than normal thyrocytes.
`
`PLOS ONE | www.plosone.org
`
`3
`
`September 2013 | Volume 8 | Issue 9 | e75489
`
`3 of 16
`
`OnCusp
`Ex. 1027
`
`

`

`Table 1. Origin and mutational status of thyroid-derived cell
`lines.
`
`FTC133
`B-CPAP TPC1 WRO
`Nthy.ori 3.1
`Cell line
`FTC-LNM
`PTC
`PTC FTC
`Normal thyroid
`Origin
`wt
`V600E
`wt
`wt/V600E*
`nd
`BRAF
`wt
`wt
`wt
`wt
`nd
`NRAS
`wt
`wt
`wt
`wt
`nd
`HRAS
`wt
`wt
`wt
`wt
`nd
`KRAS
`wt
`wt
`wt
`wt
`nd
`PI3KCA
`R130stop
`wt
`wt
`wt
`nd
`PTEN
`R273H
`D259Y
`wt
`wt
`nd
`TP53
`-
`-
`+
`-
`nd
`RET/PTC1
`-
`-
`-
`-
`nd
`RET/PTC3
`-
`-
`-
`-
`nd
`PAX8/PPARg
`PTC, Papillary Thyroid Carcinoma; FTC, Follicular Thyroid Carcinoma; FTC-LNM,
`Lymph Node Metastasis from Follicular Thyroid Carcinoma; nd, not determined; wt,
`wild-type; * BRAF V600E mutation in WRO cell line is reported only by some
`authors; + presence of rearrangement; - absence of rearrangement (modified from
`Saiselet et al. 2012) [21].
`doi: 10.1371/journal.pone.0075489.t001
`
`A lower expression of the TGF-β receptors in thyroid tumor
`cells could result in a minor capacity to transduce the TGF-β
`signal. To test this hypothesis, we compared the expression
`levels of the TGF-β receptors 1 and 2 in the five cell lines by
`qRT-PCR analysis. As shown in Figure 1G and H, all cell lines
`displayed comparable levels of the TGFR1, while a more
`heterogeneous expression of the TGFR2 was observed.
`However, in contrast with the extent of the TGF-β response, all
`tumor-derived cells, with the exception of the FTC133 cells,
`had higher levels of the TGFR2 than the Nthy.ori 3.1 cells.
`
`The EMT program is constitutively active in tumor-
`derived cells.
`We investigated the possibility that tumor-derived cells have
`already undergone an EMT transformation by comparing the
`expression
`levels of
`the above-mentioned EMT markers
`between normal Nthy.ori 3.1 thyrocytes and tumor-derived
`cells. Noticeably, the expression of all mesenchymal markers
`analyzed (N-CAD, TNC, VIM and FN 1) was significantly higher
`in the tumor cells than in the Nthy.ori 3.1 thyrocytes. By
`contrast, E-CAD expression could be detected only in the
`Nthy.ori 3.1 and FTC133 cells, while no expression of this
`epithelial marker was detected in the other tumor-derived cells
`(Figure 2A). CDH16 expression was higher in TPC1 and
`FTC133 than in the rest of the cell lines. Noticeably, CDH6
`expression was greatly enhanced in all thyroid tumor-derived
`cells as compared to the normal thyrocytes (Figure 2B).
`Western blot analysis was in accordance with the gene
`expression data (Figure 2C). Then we tested whether the EMT-
`like phenotype displayed by
`thyroid
`tumor cells was
`accompanied by a constitutive activation of
`the TGF-β
`downstream pathways. TGF-β signals within the cells through
`three major pathways: nuclear translocation of the SMAD2/3
`proteins, the activation of the MAPK, and of PI3K cascade [25].
`
`CDH6 in TGF-β Mediated EMT Program
`
`Since our interest was mainly on the mechanism controlling the
`biology of papillary thyroid tumors, we limited the analysis to
`papillary-derived tumor cells. Nthy.ori 3.1, B-CPAP, and TPC1
`cells were treated with TGF-β and changes in the activation of
`the three signaling pathways were monitored. In the Nthy.ori
`3.1 thyrocytes, the levels of both pERK and pAKT rose early
`after the addition of TGF-β. In B-CPAP cells, pERK levels did
`not change in response to TGF-β, while AKT phosphorylation
`was significantly induced. In TPC1 cells, phosphorylation levels
`of both ERK and AKT did not change in response to TGF-β
`(Figure 2D). Furthermore, in untreated Nthy. ori.3.1 cells
`SMAD2/3 were localized in the cytoplasm, to be strongly
`translocated to nuclei early after TGF-β exposure (Figure 2E
`upper panels). In untreated B-CPAP cells SMAD2/3 were
`localized mainly in the nuclei and their localization did not
`change significantly in the presence of TGF-β (Figure 2E
`middle panels). In untreated TPC1 cells, the SMAD2/3 staining
`was visible in both nucleus and cytoplasm. However, after
`TGF-β treatment, the cytoplasmic staining was no longer
`detectable, indicating that all the protein was translocated to
`the nuclei (Figure 2E lower panels). Overall, these data
`suggest
`that TGF-β-dependent signaling cascades are
`constitutively active in thyroid tumor cells. These results may
`be partially explained by the fact that tumor cells, including the
`B-CPAP and TPC1, often harbor somatic mutations
`in
`components of these signaling pathways (see Table 1).
`However, tumor cells retain a partial ability to respond to the
`TGF-β signal by further activating some of these pathways, like
`AKT in the B-CPAP and SMAD2/3 in the TPC1. The EMT
`program is controlled by a number of transcription factors,
`which are known targets of TGF-β [5,26]. Both Id1 and RUNX2
`have been shown to be TGF-β targets and have been
`proposed as crucial regulators of the EMT program [8,27-29].
`We
`investigated how
`the expression of EMT-related
`transcription factors changed in response to TGF-β in thyroid
`cell lines. SNAI1, SNAI2, Id1, and RUNX2 were induced by
`TGF-β in both Nthy.ori 3.1 and TPC1 cells. In accordance with
`what was observed for the EMT markers (Figure 1), the
`activation of SNAI1, SNAI2, and Id1 was lower in TPC1 than in
`Nthy.ori 3.1 (Figure 2F-H). In B-CPAP cells slight change in
`SNAI1, SNAI2, ZEB1 and TWIST expression was observed
`after 24h (Figure 2G). ZEB1 and TWIST levels were not
`affected in Nthy.ori 3.1 and TPC1 cells. It is worth pointing out
`that the induction of Id1 and RUNX2 preceded the activation of
`the other factors, suggesting that these factors are early targets
`of the TGF-β-dependent EMT program.
`
`CDH6 is overexpressed in human PTCs and mainly
`localized at the invasive front of the tumors.
`To confirm in vivo that thyroid tumor cells display features of
`a constitutively active EMT-program, we analyzed
`the
`expression of EMT markers
`in human papillary
`thyroid
`carcinoma
`(PTC) samples and matched
`lymph node
`metastases (LNMs). We collected total RNA from tumor and
`surrounding normal tissue for 15 PTC patients and we
`analyzed EMT markers by means of qRT-PCR (Figure 3A).
`Noticeably, the expression of mesenchymal markers (N-CAD,
`TNC, and FN 1) was significantly higher in the vast majority of
`
`PLOS ONE | www.plosone.org
`
`4
`
`September 2013 | Volume 8 | Issue 9 | e75489
`
`4 of 16
`
`OnCusp
`Ex. 1027
`
`

`

`CDH6 in TGF-β Mediated EMT Program
`
`Figure 1. Nthy.ori 3.1 thyrocytes are more responsive to the TGF-β-mediated EMT program than tumor-derived cells. A-E)
`qRT-PCR analysis of EMT markers (E-CAD; N-CAD; CDH16; TNC; VIM; FN 1) and CDH6 in non-treated (NT; black bars) or TGF-β
`treated (5 ng/ml grey bars; 100 ng/ml white bars) thyroid-derived cell lines. The bars represent the average fold change of indicated
`genes in TGF-β treated cells as compared to non-treated cells, normalized to the geometric mean of levels of three reference
`genes: GAPDH, CYPA, GUSB. F) Western Blot analysis of E-CAD, N-CAD, FN 1, and Actin in Nthy.ori 3.1 cells, B-CPAP and TPC1
`cells non-treated (NT) or treated with 100 ng/ml of TGF-β for 6h and 24h. G, H) qRT-PCR analysis of TGFR1 (G) and TGFR2 (H) in
`non-treated thyroid-derived cell lines. The bars represent the average fold change of TGFR1 and TGFR2 in tumor cells (B-CPAP,
`TPC1, WRO, and FTC-133) as compared to thyrocytes (Nthy.ori 3.1), normalized to the to the geometric mean of GAPDH, CYPA,
`GUSB levels. Error bars represent s.e.m. (n=3). p-value was calculated by two-tailed Student’s t-test. *** p≤ 0.001; ** p≤ 0.01; * p≤
`0.05. Each experiment has been replicated a minimum of two times with comparable results.
`doi: 10.1371/journal.pone.0075489.g001
`
`PLOS ONE | www.plosone.org
`
`5
`
`September 2013 | Volume 8 | Issue 9 | e75489
`
`5 of 16
`
`OnCusp
`Ex. 1027
`
`

`

`CDH6 in TGF-β Mediated EMT Program
`
`Figure 2. Tumor-derived cell lines display a constitutive EMT-like phenotype. qRT-PCR analysis of EMT markers (A) (E-
`CAD; N-CAD; CDH16; TNC; VIM; FN 1) and CDH6 (B) in non-treated thyroid-derived cell lines. The bars represent the average fold
`change of the indicated genes in tumor cells (B-CPAP, TPC1, WRO, and FTC-133) as compared to thyrocytes (Nthy.ori 3.1),
`normalized to the geometric mean of levels of three reference genes: GAPDH, CYPA, GUSB. C) Western Blot analysis of E-CAD,
`N-CAD, FN1 and Actin in non-treated Nthy.ori 3.1; B-CPAP and TPC1 cells. D) Western Blot analysis of phosphorylated ERK,
`phosphorylated AKT, and Actin in Nthy.ori 3.1 (Left panels); B-CPAP (middle panels) and TPC1 (right panels) cells untreated (NT)
`or after TGF-β exposure for the indicated times. E) Immunofluorescence staining of SMAD2/3 proteins (green) in Nthy.ori 3.1 (upper
`panels); B-CPAP (middle panels) and TPC1 (lower panels), non-treated (NT) or after TGF-β exposure for the indicated times. DAPI
`(Blue) stains the nuclei. Magnification 200X. F-H) qRT-PCR analysis of transcription factors known to partake in the EMT program
`(SNAI1, SNAI2, ZEB1, TWIST, Id1, and RUNX2) in Nthy.ori 3.1 (E), B-CPAP (F) and TPC1 (G) cells, non-treated (NT) or treated
`with TGF-β for the indicated times. The bars represent the fold change of the indicated genes in TGF-β treated cells as compared to
`the non-treated cell levels, normalized to the geometric mean of GAPDH, CYPA, GUSB levels. p-value was calculated by two-tailed
`Student’s t-test. *** p≤ 0.001; ** p≤ 0.01; * p≤ 0.05. Error bars represent s.e.m. (n=3).
`doi: 10.1371/journal.pone.0075489.g002
`
`PLOS ONE | www.plosone.org
`
`6
`
`September 2013 | Volume 8 | Issue 9 | e75489
`
`6 of 16
`
`OnCusp
`Ex. 1027
`
`

`

`tumors (black circles) than in normal tissues (baseline). By
`contrast, CDH16, marker of the epithelial phenotype, was
`significantly less expressed in tumor cells than in normal
`thyrocytes. E-CAD expression was not significantly different in
`tumor sample as compared with the normal tissue. LNMs were
`available for 7 of the analyzed PTC samples. The expression
`pattern of EMT markers in LNM (black and white circles)
`reflected the trend observed in the primary lesions, confirming
`that active EMT features are a constitutive hallmark of tumor
`transformation in thyroid tumors. To confirm that CDH6 is a
`marker of the mesenchymal phenotype in thyroid tumors, we
`investigated the expression of CDH6 in the aforementioned
`PTC samples by qRT-PCR (Figure 3B). As for the other
`mesenchymal markers, CDH6 expression was higher in tumor
`tissue (black circles) and LNMs
`than
`in normal
`tissue
`(baseline). The CDH6 expression levels seemed to decrease in
`the transition from primary to metastatic site, suggesting that
`this factor may be required particularly in the early phases of
`the metastasization process. Next, we analyzed the CDH6
`protein expression by immunohistochemistry in the same PTC
`samples previously analyzed by qRT-PCR. As expected, CDH6
`expression was clearly higher in tumor cells than in normal
`tissue (Figure 3C and Figure S2A-F). However, the expression
`of CDH6 was not uniform within the tumor but was restricted to
`groups of cells. Intriguingly, CDH6-positive cells were localized
`preferentially at the invasion front of the tumor, strongly
`supporting the hypothesis that CDH6 is important in controlling
`cell motility and invasiveness in PTCs (Figure 3C and Figure
`S2A and B).
`
`CDH6 splicing variants display a different profile in
`thyroid tumor-derived cells
`The CDH6 gene codes for two different splicing isoforms,
`starting from a common promoter. The long isoform comprises
`both an extracellular domain (containing 4 ectodomain modules
`–EC– and a membrane proximal extracellular domain–MPE)
`and a cytoplasmic domain (containing the membrane proximal
`conserved domain–MPC- and the conserved catenin binding
`site–CBS) [7] (Figure S2G). Instead, the short isoform displays
`only the extracellular domain but lacks the cytoplasmic domain.
`Due to the ability of transmembrane proteins to transduce the
`signals within the cell through their cytoplasmic domains, it is
`reasonable to suppose that the two variants may play different
`roles. To address this issue, we analyzed the expression of the
`two CDH6 isoforms by qRT-PCR using specific primers (Figure
`S2G). Both isoforms were expressed in all cell lines (Figure 4A)
`and more strongly expressed in tumor-derived cells than in
`Nthy.ori 3.1 (Figure 4B), as observed for the total CDH6
`(Figure 2B). Intriguingly, the expression change in tumor-
`derived cells (B-CPAP and TPC1) as compared to the Nthy.ori
`3.1 was more pronounced for the CDH6-long isoform (CDH6-L)
`than for the short isoform (CDH6-S). Further, we analyzed
`changes in the expression levels of CDH6-L and CDH6-S
`isoforms by TGF-β treatment. Both isoforms were similarly
`induced by TGF-β in all cell lines (Figure 4C, D, E). This
`suggests
`that TGF-β controls CDH6 expression at
`the
`transcriptional level. To confirm this hypothesis we investigated
`the effect of TGF-β treatment on CDH6 expression in the
`
`CDH6 in TGF-β Mediated EMT Program
`
`absence of active transcription (with Actinomycin D) or in the
`absence of active protein synthesis (with Cycloheximide). As
`shown in Figure 5A, the addition of Actinomycin D completely
`abolished the TGF-β-mediated CDH6 induction, confirming that
`the regulatory effect of TGF-β requires an active transcription.
`Furthermore, the addition of Cycloheximide also resulted in a
`complete suppression of TGF-β induction, demonstrating that
`the effect on the CDH6 expression is indirect and requires the
`previous synthesis of an activator factor.
`
`RUNX2 mediates the TGF-β effect in thyroid cells
`In a recent microarray analysis, we found CDH6 as a target
`of Id1 in aggressive thyroid tumor cells [8]. Further, we showed
`that the transcription factor RUNX2, also a target of Id1, is a
`crucial mediator of the Id1 pro-invasive function in thyroid
`tumor cells [30]. Both RUNX2 and Id1 have been proposed to
`control EMT in epithelial tumors. We reasoned that RUNX2
`could be responsible for the TGF-β-dependent CDH6 induction.
`To test this hypothesis, Nthy.ori 3.1 cells were transfected with
`siRNA against RUNX2 or control siRNA and treated with TGF-
`β for 24h. As shown in Figure 5B, the siRNA-mediated RUNX2
`depletion profoundly impaired the TGF-β effect on CDH6
`expression. These experiments confirm that CDH6 is a target
`of RUNX2 in the TGF-β pathway. However, siRNA mediated
`RUNX2 ablation does not completely abolish the TGF-β effect
`on CDH6 expression suggesting that other TGF-β–dependent
`factors beside RUNX2 may be involved in controlling this gene
`in thyroid tumor cells.
`RUNX2 has been shown to control the expression of SNAI
`factors [29]. After TGF-β treatment, RUNX2 induction occurs
`earlier that SNAI1 and SNAI2 activation. Thus, it is possible
`that the effect of RUNX2 on CDH6 expression could be indirect
`and mediated by SNAI factors. However, RUNX2 silencing
`does not affect TGF-β- dependent SNAI1 or SNAI2 induction
`suggesting that these transcription factors are not involved in
`this regulation (Figure S2H-I).
`We showed that TGF-β signaling pathway is constitutively
`activated in thyroid tumor cells and suggested that RUNX2 and
`CDH6 is part of this pathway. If this hypothesis is correct
`blocking the TGF-β signaling in thyroid cancer cells should
`affect RUNX2 and CDH6 expression levels. To address this
`issue we treated B-CPAP and TPC1 cells with SB-431542, a
`small molecule known to block kinase activity of the TGF-β
`receptors family [30]. Noticeably, exposure to SB-431542
`results in a strong repression of CDH6 and RUNX2 in both B-
`CPAP and TPC1 cells, which was clearly visible after 24h and
`increased after 48h from the beginning of the treatment (Figure
`5C,D). The inhibitory effect of th