The epithelial–mesenchymal transition (EMT) is a key mechanism in both embryonic development and cancer metastasis. The EMT introduces stem-like properties to cancer cells. However, during somatic cell reprogramming, mesenchymal–epithelial transition (MET), the reverse process of EMT, is a crucial step toward pluripotency. Connective tissue growth factor (CTGF) is a multifunctional secreted protein that acts as either an oncoprotein or a tumor suppressor among different cancers. Here, we show that in head and neck squamous cell carcinoma (HNSCC), CTGF promotes the MET and reduces invasiveness. Moreover, we found that CTGF enhances the stem-like properties of HNSCC cells and increases the expression of multiple pluripotency genes. Mechanistic studies showed that CTGF induces c-Jun expression through αvβ3 integrin and that c-Jun directly activates the transcription of the pluripotency genes NANOG, SOX2, and POU5F1. Knockdown of CTGF in TW2.6 cells was shown to reduce tumor formation and attenuate E-cadherin expression in xenotransplanted tumors. In HNSCC patient samples, CTGF expression was positively correlated with the levels of CDH1, NANOG, SOX2, and POU5F1. Coexpression of CTGF and the pluripotency genes was found to be associated with a worse prognosis. These findings are valuable in elucidating the interplay between epithelial plasticity and stem-like properties during cancer progression and provide useful information for developing a novel classification system and therapeutic strategies for HNSCC. Cancer Res; 73(13); 4147–57. ©2013 AACR.

Epithelial–mesenchymal transition (EMT), which is hallmarked by suppression of the adherent protein E-cadherin, is a fundamental process in embryonic development, organ fibrosis, and cancer metastasis (1, 2). A recent breakthrough in EMT research has shown that EMT is capable of introducing stem-like properties to epithelial cells (3, 4). In human cancers, the acquisition of stem-like properties in cancer cells is critically involved in disease progression and treatment resistance (5). Different mechanisms have been shown to be responsible for the EMT-induced stem-like properties of cancer cells. For instance, the EMT inducer Zeb1 inhibits the expression of the microRNA 200 family, resulting in the upregulation of the polycomb protein Bmi1 and the induction of stemness in pancreatic cancer (6). In head and neck cancers, we previously showed that the other EMT inducer Twist1 directly activates BMI1 transcription and that Twist and Bmi1 act cooperatively to promote the EMT and stemness (7). However, the role of pluripotency genes in EMT-induced stemness remains unclear.

Recently, the most fascinating progress in stem cell research is the direct reprogramming of somatic cells into an embryonic stem cell (ESC)-like state by defined factors, as in the generation of induced pluripotent stem cells (iPSC; refs. 8–10). A further conceptual advance in iPSC research is that mesenchymal–epithelial transition (MET), the reverse process of EMT, is essential for the generation of iPSCs (11, 12). Moreover, E-cadherin, an important adhesion molecule that is downregulated during the EMT, has been shown to be highly expressed in mouse ESCs and reprogrammed cells, and E-cadherin itself induces pluripotency (13). All of these results from independent study groups show the critical role of the MET and E-cadherin in the induction of pluripotency in somatic cells. However, these findings seem to conflict with the concept of EMT-induced stemness in cancer cells (14). Clarification of the relationship between the EMT/MET and stem cell properties in cancer cells is necessary.

Connective tissue growth factor (CTGF, also known as CCN2) is a secreted protein that acts as a multifunctional signaling modulator in various biologic or pathologic processes (15). In human cancers, the pleiotropic functions of CTGF have been investigated among different types of cancers (16). CTGF acts as an oncoprotein in glioma and breast cancer (17, 18). However, we previously showed the tumor-suppressive effect of CTGF in lung cancer and colon cancer (19, 20). Head and neck squamous cell carcinoma (HNSCC), including cancers originating in the oral cavity, oropharynx, hypopharynx, and larynx, is one of the leading causes of cancer-related death worldwide. Local progression and lymph node involvement are the major causes of HNSCC-related mortality and the incidence of distal organ metastasis is relatively rare compared with other cancers (21). However, the mechanism responsible for the local progression of HNSCC is unclear. We recently showed that CTGF attenuates the invasiveness of oral squamous cell carcinoma cells through miR504 and FOXP1 (22). In this report, we further show the unique function of CTGF in HNSCC in the induction of the MET and stem-like properties, which results in hindering the dissemination but promoting the local progression of this cancer.

Cells and plasmids

HEK-293T and the human hypopharyngeal cancer cell line FaDu were obtained from the Bioresource Collection and Research Center of Taiwan. The human oral cancer cell line OECM-1 was originally provided by Dr. Ching-Liang Meng of the National Defense Medical College in Taiwan (23). The human oral cancer cell lines SAS and HSC3 were provided by Dr. Cheng-Chi Chang of National Taiwan University. The TW2.6 cells were provided by Dr. Mark Y.B. Kuo of National Taiwan University (24). The characteristics of the other HNSCC cell lines are described in our previous study (22). All of the cell lines were cultured in Dulbecco's Modified Eagle Medium (DMEM) with 10% FBS, except for FaDu and OECM-1, which were cultured in Roswell Park Memorial Institute (RPMI)-1640 medium with 10% FBS.

The pcDNA3-CTGF was generated by inserting the open reading frame of CTGF into the pcDNA3.1 vector, and pCDH-JUN was generated by inserting the open reading frame of JUN into the pCDH-MCV-MCS-EF1-puro vector. The short-interfering RNA (siRNA) vectors pSUPER-si-CTGF, pSUPER-si-JUN, and pSUPER-si-ESR1 were generated by inserting a short-hairpin sequence against the target genes into the pSUPER.puro vector, and a control vector for the siRNA experiments was constructed by inserting a scrambled sequence. Supplementary Table S1 lists the sequences used for the production of the siRNA constructs. Stable clones were generated by transfection of the expression vectors and/or siRNA plasmids and selected using the appropriate antibiotics.

Spheroid formation assay

Cells (1 × 104) were suspended in serum-free DMEM/F-12 (Gibco-BRL) containing N2 supplement, 10 ng/mL human recombinant bFGF, and 10 ng/mL EGF (all from R&D Systems Inc.). After cultivation for 14 days, primary spheroids were harvested by centrifugation and then dissociated and resuspended in this medium. The number of secondary spheroids larger than 100 μm was counted after 14 days.

cDNA microarray

The Affymetrix HG-U133 plus 2.0 whole-genome array was used for cDNA microarray. The microarray data were deposited in the NCBI GEO database with the accession number GSE30423.

Prediction of putative transcription factors

TESS (25) and PROMO 3.0 (26) were used to predict the putative transcriptional factors that regulate the expression of pluripotency genes.

Cloning of the proximal promoter regions of POU5F1, SOX2, NANOG, and JUN, generation of the promoter reporter constructs, and luciferase reporter assay

The genomic regions flanking the promoter region of human POU5F1 (−1464 ∼ +53 bp to ATG), NANOG (−1544 ∼ +160 bp to ATG), SOX2 (−1580 ∼ +300 bp to ATG), and JUN (−1140 ∼ +80 bp to ATG) were amplified by PCR and inserted into the SacI/BglII sites of the pGL4.2 vector to generate the corresponding reporter constructs (Fig. 3E and 5A). Promoter constructs containing mutated c-Jun–binding sites were generated by site-directed mutagenesis (Fig. 5A). The luciferase reporter assay was conducted by transfecting the reporter construct with or without the siRNA vector into the indicated cell lines. A plasmid expressing the bacterial β-galactosidase gene (pCMV-βgal) or the renilla luciferase gene (pRL-TK) was cotransfected in each experiment as an internal control for transfection efficiency. Cells were harvested after 24 hours of transfection, and the luciferase activities were assayed as previously described (7). All values are expressed as the fold change in luciferase activity after normalization to the β-galactosidase activity.

Statistical analysis

An independent Student t test was used to compare the continuous variables between 2 groups, and a χ2 test was applied for the comparison of dichotomous variables. A Kaplan–Meier estimation and a log-rank test were used to compare the difference in the survival period between patient groups. The level of statistical significance was set at 0.05 for all tests.

Please see Supplementary Methods for the other methods used in this study.

CTGF induces mesenchymal–epithelial transition in head and neck cancer cells

Because CTGF has been shown to play different roles, either to promote or inhibit metastasis in different types of human cancers (16), we herein investigated the effect of CTGF on the migration of HNSCC cells. First, we screened the expression level of CTGF in 4 HNSCC cell lines, including SAS, HSC3, FaDu, and TW2.6. CTGF was found to be significantly higher in TW2.6 than in the other 3 cell lines (Supplementary Fig. S1A). We therefore selected SAS and FaDu as parental cell lines to generate stable CTGF overexpression lines and TW2.6 to establish stable CTGF knockdown lines. The ectopic expression of CTGF inhibited the migration of the SAS and FaDu cells (Fig. 1A and B), and knockdown of endogenous CTGF in the TW2.6 cells augmented their migratory ability (Fig. 1C). However, CTGF did not have significant impact on proliferation of the SAS and FaDu cells (Supplementary Fig. S1B), suggesting that CTGF reduces cellular migration without affecting proliferative ability in HNSCC cells.

Because the EMT is a crucial process in promoting cancer cell migration and metastasis (2), we investigated whether CTGF could induce the MET, the reverse process of EMT, to inhibit migration in HNSCC. To this end, we examined the EMT phenotype in our established stable cell lines. A switch from N-cadherin to E-cadherin was shown in the SAS and FaDu CTGF transfectants (Fig. 1D and E), which indicates the occurrence of the MET. Consistently, the suppression of CTGF in TW2.6 caused a shift from E-cadherin to N-cadherin (Fig. 1F). In the SAS and FaDu cells, overexpression of CTGF induced an epithelial morphology and expression of membranous E-cadherin (Fig. 1G and H). In contrast, knockdown of CTGF dissociated membranous E-cadherin in the TW2.6 cells (Fig. 1I). Collectively, these results suggest that CTGF inhibits migration and promotes the MET in HNSCC cells.

Figure 1.

CTGF induces MET in HNSCC. A–C, wound-healing migration assay of SAS cells stably transfected with CTGF (SAS-C1 and SAS-C2) versus a control vector (SAS-Neo; A), FaDu cells stably transfected with CTGF (FaDu-C1 and FaDu-C2) versus a control vector (FaDu-Neo; B) and TW2.6 cells receiving siRNA against CTGF (TW2.6-siC1 and TW2.6-siC2) versus a scrambled sequence (TW2.6-scr; C). Left, photos of wound-healing assay. Right, relative migratory ability at different time points. D–F, Western blot analysis of CTGF, E-cadherin, and N-cadherin in SAS (D), FaDu (E), and TW2.6 stable cells (F). G–I, phase contrast image (top) and immunofluorescence staining of E-cadherin (bottom) in SAS (G), FaDu (H), and TW2.6 stable cells (I). Scale bar, 200 μm for phase contrast image; 20 μm for immunofluorescence image.

Figure 1.

CTGF induces MET in HNSCC. A–C, wound-healing migration assay of SAS cells stably transfected with CTGF (SAS-C1 and SAS-C2) versus a control vector (SAS-Neo; A), FaDu cells stably transfected with CTGF (FaDu-C1 and FaDu-C2) versus a control vector (FaDu-Neo; B) and TW2.6 cells receiving siRNA against CTGF (TW2.6-siC1 and TW2.6-siC2) versus a scrambled sequence (TW2.6-scr; C). Left, photos of wound-healing assay. Right, relative migratory ability at different time points. D–F, Western blot analysis of CTGF, E-cadherin, and N-cadherin in SAS (D), FaDu (E), and TW2.6 stable cells (F). G–I, phase contrast image (top) and immunofluorescence staining of E-cadherin (bottom) in SAS (G), FaDu (H), and TW2.6 stable cells (I). Scale bar, 200 μm for phase contrast image; 20 μm for immunofluorescence image.

Close modal

CTGF promotes stem-like properties and upregulates pluripotency genes

Because the EMT generates cells with stem-like properties (3, 4), we reasoned that CTGF may reduce cellular stemness in HNSCC. To examine this notion, we conducted experiments to evaluate the stem-like properties of HNSCC cells using our stable cell lines, including the analysis of the putative HNSCC stem cell marker CD44+ (27), aldehyde dehydrogenase 1 (ALDH1) activity, the proportion of side-population cells, anchorage-independent growth, and spheroid-forming ability. To our surprise, in the SAS cells, the ectopic expression of CTGF increased the proportion of CD44+ (Fig. 2A), side-population (Supplementary Fig. S2A), and ALDH1+ cells (Supplementary Fig. S2B). Ectopic CTGF also enhanced the anchorage-independent growth and spheroid-forming ability (Fig. 2B and C). This result was also shown in another HNSCC cell line FaDu: an increased proportion of CD44+ cells and enhanced colony-forming and spheroid-forming abilities were shown in FaDu–CTGF transfectants (Supplementary Fig. S3A–C). Knockdown of CTGF in TW2.6 cells, which express a high level of endogenous CTGF, reduced the CD44+ population, spheroid-forming ability, and anchorage-independent growth (Fig. 2A–C). Because CTGF is a secreted protein, we used the recombinant CTGF protein (rCTGF) to treat SAS and FaDu cells, which produce a low level of CTGF, and then observed the impact of exogenous CTGF on the stem-like properties of these cells. rCTGF increased the proportion of CD44+ cells and enhanced the colony-forming and spheroid-forming abilities in both the SAS and FaDu cells (Supplementary Fig. S4).

Figure 2.

CTGF promotes stem-like properties of HNSCC cells. A, flow cytometry for analyzing CD44+ expression in SAS cells transfected with the CTGF-expressing vector (SAS-C1 and SAS-C2) or with an empty vector as a control (SAS-Neo; left) and TW2.6 cells receiving siRNA against CTGF (TW2.6-siC1) versus a scrambled sequence (TW2.6-scr; right). The percentage of CD44+ cells is shown in the right bottom quadrant. B, spheroid formation assay. Top, representative pictures. Bottom, quantification (n = 3). C, soft agar colony formation assay (n = 3). Colonies larger than 0.1 mm were counted. D, a heat-map summarizing the results of relative expression of 14 stemness genes in SAS-C1 versus SAS-Neo or TW2.6-siC1 versus TW2.6-scr. E, Western blot analysis of CTGF, Oct4, Nanog, and Sox2 in SAS and TW2.6 stable cell lines. In B and C, data represent means ± SEM. *, P < 0.05 by Student t test.

Figure 2.

CTGF promotes stem-like properties of HNSCC cells. A, flow cytometry for analyzing CD44+ expression in SAS cells transfected with the CTGF-expressing vector (SAS-C1 and SAS-C2) or with an empty vector as a control (SAS-Neo; left) and TW2.6 cells receiving siRNA against CTGF (TW2.6-siC1) versus a scrambled sequence (TW2.6-scr; right). The percentage of CD44+ cells is shown in the right bottom quadrant. B, spheroid formation assay. Top, representative pictures. Bottom, quantification (n = 3). C, soft agar colony formation assay (n = 3). Colonies larger than 0.1 mm were counted. D, a heat-map summarizing the results of relative expression of 14 stemness genes in SAS-C1 versus SAS-Neo or TW2.6-siC1 versus TW2.6-scr. E, Western blot analysis of CTGF, Oct4, Nanog, and Sox2 in SAS and TW2.6 stable cell lines. In B and C, data represent means ± SEM. *, P < 0.05 by Student t test.

Close modal

Next, we investigated the correlation between the expression of CTGF and the pluripotency genes in HNSCC cells. We first tested the expression of 14 human ESC-enriched genes (9) in the stable CTGF-manipulated SAS and TW2.6 cell lines. Most of the ESC-enriched genes were upregulated in the CTGF-overexpression cells and downregulated in the CTGF-knockdown cells (Fig. 2D). Among these genes, we investigated the impact of CTGF on the expression of POU5F1 (POU class 5 homeobox 1, which encodes the Oct4 protein), NANOG, and SOX2 due to their importance in both stem cells and cancers (28, 29). First, we compared the expression level of the pluripotency genes among different HNSCC cell lines and found that the cells with low endogenous CTGF (e.g., SAS and HSC3) tended to have a lower level of POU5F1/NANOG/SOX2. In contrast, in cells with high endogenous CTGF (e.g., TW2.6), the level of POU5F1/NANOG/SOX2 was relatively higher (Supplementary Fig. S5A). Stable expression of CTGF in the SAS and FaDu cells increased the level of Oct4, Sox2, and Nanog (Fig. 2E left; Fig. S5B and S5C). Consistently, knockdown of CTGF in TW2.6 cells reduced Oct4, Sox2, and Nanog (Fig. 2E right). Exogenous CTGF upregulated the expression of the pluripotency genes in a dose- and time-dependent manner, and the effect of rCTGF on inducing POU5F1/NANOG/SOX2 was more prominent than its effect on the other 2 pluripotency genes KLF4 and MYC (Supplementary Fig. S6). Taken together, these results suggest that CTGF promotes stem-like properties and induces the expression of pluripotency genes, especially POU5F1, NANOG, and SOX2, in HNSCC.

c-Jun is a major player in the CTGF-induced MET and stem-like properties

Next, we investigated how CTGF induces the MET and stem-like properties in HNSCC. We hypothesized that CTGF regulates the pluripotency genes POU5F1, NANOG, and SOX2 simultaneously, resulting in the enhancement of stem-like properties and the induction of the MET. Because the pluripotency genes were upregulated by CTGF at the mRNA level (Fig. 2D), we first aimed to identify the putative transcription factor(s) mediating CTGF-induced pluripotency gene expression. To this end, cDNA microarray analysis was conducted in FaDu cells treated with rCTGF versus vehicle control, and the transcripts that increased more than 2-fold were considered to be upregulated upon rCTGF treatment. In addition, we used 2 software systems, TESS and PROMO 3.0, to predict the transcription factors that may simultaneously regulate the transcription of NANG, SOX2, and POU5F1. Putative transcription factors that are responsible for CTGF-mediated pluripotency gene induction were identified by overlapping these 3 datasets (Fig. 3A).

Figure 3.

c-Jun is a major factor responsible for CTGF-induced pluripotency genes expression, and CTGF activates JUN transcription through αvβ3 integrin. A, schematic representation of the strategy for mining the candidate transcriptional factors mediating CTGF-induced pluripotency genes expression. B, quantitative RT-PCR analysis of the mRNA levels of 6 candidate transcriptional factors in CTGF-expressing SAS stable clones (SAS-C1 and SAS-C2) and a control clone (SAS-Neo; n = 3). C, relative mRNA levels of LEF1, ESR1, and JUN in TW2.6 cells receiving siRNA against CTGF (TW2.6-siC1 and TW2.6-siC2) or a scrambled sequence (TW2.6-scr). D, Western blot analysis of total c-Jun and phosphorylated c-Jun (p-c-Jun) in SAS (top) and TW2.6 (bottom) stable cells. E, top, schematic representation of the reporter construct of JUN promoter (JUN-Luc). TSS, transcription start site. Bottom left, relative luciferase activity of FaDu cells transfected with JUN-Luc and treated with rCTGF (0.2 μg/mL) or a vehicle control (n = 3). Bottom right, relative luciferase activity of SAS stable cells transfected with JUN-Luc (n = 3). F, relative luciferase activity in FaDu cells transfected with JUN-Luc and treated with rCTGF for 12 hours, with or without an αvβ3 neutralizing antibody. IgG was a control of the antibody neutralization experiment (n = 3). In B, C, E, and F, data represent means ± SEM. *, P < 0.05 by Student t test.

Figure 3.

c-Jun is a major factor responsible for CTGF-induced pluripotency genes expression, and CTGF activates JUN transcription through αvβ3 integrin. A, schematic representation of the strategy for mining the candidate transcriptional factors mediating CTGF-induced pluripotency genes expression. B, quantitative RT-PCR analysis of the mRNA levels of 6 candidate transcriptional factors in CTGF-expressing SAS stable clones (SAS-C1 and SAS-C2) and a control clone (SAS-Neo; n = 3). C, relative mRNA levels of LEF1, ESR1, and JUN in TW2.6 cells receiving siRNA against CTGF (TW2.6-siC1 and TW2.6-siC2) or a scrambled sequence (TW2.6-scr). D, Western blot analysis of total c-Jun and phosphorylated c-Jun (p-c-Jun) in SAS (top) and TW2.6 (bottom) stable cells. E, top, schematic representation of the reporter construct of JUN promoter (JUN-Luc). TSS, transcription start site. Bottom left, relative luciferase activity of FaDu cells transfected with JUN-Luc and treated with rCTGF (0.2 μg/mL) or a vehicle control (n = 3). Bottom right, relative luciferase activity of SAS stable cells transfected with JUN-Luc (n = 3). F, relative luciferase activity in FaDu cells transfected with JUN-Luc and treated with rCTGF for 12 hours, with or without an αvβ3 neutralizing antibody. IgG was a control of the antibody neutralization experiment (n = 3). In B, C, E, and F, data represent means ± SEM. *, P < 0.05 by Student t test.

Close modal

Using this strategy, we identified 6 factors (TCFL2, ESR1, YY1, MYB, LEF1, and JUN) as candidates that coordinate the expression of NANOG, SOX2 and POU5F1 simultaneously. To validate this result, quantitative RT-PCR analysis was conducted in the CTGF overexpression and knockdown systems. In the SAS cells, 3 of the 6 transcription factors (ESR1, LEF1, and JUN) were upregulated in the CTGF transfectants compared with the control cells (Fig. 3B). In the TW2.6 cells, however, only ESR1 and JUN were significantly decreased when CTGF was knocked down (Fig. 3C). We therefore considered that c-Jun (the protein encoded by JUN) and ESR1 (estrogen receptor 1) are major factors contributing to the CTGF-induced expression of POU5F1/NANOG/SOX2. To confirm this, we repressed the expression of JUN and ESR1, respectively, using siRNA in the SAS–CTGF transfectants. Repression of either JUN or ESR1 reduced the CTGF-induced expression of POU5F1/NANOG/SOX2 in the SAS cells (Supplementary Fig. S7), suggesting that both c-Jun and ESR1 contribute to CTGF-mediated pluripotency genes expression.

Next, we focused on the regulation of c-Jun by CTGF because c-Jun is an important factor in cancer progression (30), and head and neck cancer is a male-predominant disease (31, 32). In SAS cells, overexpression of CTGF enhanced the expression of total and phosphorylated c-Jun; consistently, knockdown of CTGF in TW2.6 reduced the level of total and phosphorylated c-Jun (Fig. 3D). Because JUN was found to be upregulated by CTGF at the mRNA level, we assumed that CTGF activates the transcription of JUN. To address this notion, we generated a reporter construct containing the proximal promoter of JUN and tested whether CTGF could activate it. Both exogenous CTGF treatment and ectopic CTGF expression activated the JUN promoter (Fig. 3E). Because CTGF has been shown to induce target gene expression through interaction with αvβ3 integrin (33), we investigated whether CTGF induces JUN transactivation through αvβ3 integrin. The data showed that in FaDu cells, the activation of the JUN promoter by exogenous CTGF was abrogated in the presence of an anti-αvβ3 antibody (Fig. 3F), suggesting that αvβ3 is an essential receptor in CTGF-induced JUN transactivation.

We next investigated the role of c-Jun and individual pluripotency factors in the CTGF-induced MET and stem-like characteristics. In stable SAS–CTGF cells, knockdown of JUN reversed the MET, i.e., upregulation of N-cadherin and downregulation of E-cadherin, appearance of a mesenchymal-like morphology, and dissociation of E-cadherin at intercellular junctions (Fig. 4A). Interestingly, the ectopic expression of c-Jun promoted the MET in SAS cells (Fig. 4B). Repression of c-Jun reduced the CD44+ population in the SAS–CTGF transfectants (Fig. 4C and D). However, knockdown of individual pluripotency genes, including POU5F1, NANOG, and SOX2, only partially affected the expression of EMT markers in the SAS–CTGF cells. Repression of POU5F1 enhanced N-cadherin expression (Fig. 4E), and silencing NANOG or SOX2 attenuated the levels of both E-cadherin and N-cadherin (Fig. 4F and 4G). Taken together, these results indicate that CTGF induces the transcriptional activation of JUN through αvβ3 integrin and c-Jun is a major factor involved in the CTGF-mediated MET and stemness in HNSCC.

Figure 4.

c-Jun is critical in CTGF-induced MET and stemness, and suppression of single pluripotency factor only partially affects the EMT markers. A, left, Western blot analysis of c-Jun, E-cadherin, and N-cadherin in SAS-C1 cells receiving siRNA against c-Jun (siJUN) or a scrambled control (scr). Right, phase contrast image (top) and immunofluorescence staining of E-cadherin (bottom). Scale bar, 200 μm for phase contrast image and 20 μm for immunofluorescence. B, Western blot of CTGF, c-Jun, epithelial markers (E-cadherin and γ-catenin), and mesenchymal markers (N-cadherin and vimentin) in SAS cells transfected with a c-Jun expression vector (SAS-c-Jun) or a control vector (SAS-CDH). C, representative results of flow cytometry for analyzing CD44+ expression in SAS-C1-siJUN versus SAS-C1-scr. The percentage of CD44+ cells was shown in the right top quadrant. D, quantification of the CD44+ flow cytometry results (n = 3). Data represent means ± SEM. *, P < 0.05 by Student t test. E–G, Western blot analysis of the epithelial marker E-cadherin, mesenchymal marker N-cadherin and vimentin, and pluripotency factors Oct4/Nanog/Sox in SAS-C1 cells receiving siRNA against a scrambled control (scr) or Oct4 (E), Sox2 (F), or Nanog (G).

Figure 4.

c-Jun is critical in CTGF-induced MET and stemness, and suppression of single pluripotency factor only partially affects the EMT markers. A, left, Western blot analysis of c-Jun, E-cadherin, and N-cadherin in SAS-C1 cells receiving siRNA against c-Jun (siJUN) or a scrambled control (scr). Right, phase contrast image (top) and immunofluorescence staining of E-cadherin (bottom). Scale bar, 200 μm for phase contrast image and 20 μm for immunofluorescence. B, Western blot of CTGF, c-Jun, epithelial markers (E-cadherin and γ-catenin), and mesenchymal markers (N-cadherin and vimentin) in SAS cells transfected with a c-Jun expression vector (SAS-c-Jun) or a control vector (SAS-CDH). C, representative results of flow cytometry for analyzing CD44+ expression in SAS-C1-siJUN versus SAS-C1-scr. The percentage of CD44+ cells was shown in the right top quadrant. D, quantification of the CD44+ flow cytometry results (n = 3). Data represent means ± SEM. *, P < 0.05 by Student t test. E–G, Western blot analysis of the epithelial marker E-cadherin, mesenchymal marker N-cadherin and vimentin, and pluripotency factors Oct4/Nanog/Sox in SAS-C1 cells receiving siRNA against a scrambled control (scr) or Oct4 (E), Sox2 (F), or Nanog (G).

Close modal

c-Jun directly regulates the pluripotency genes POU5F1/NANOG/SOX2

Next, we sought to determine whether CTGF transactivates POU5F1/NANOG/SOX2 through c-Jun or ESR1. Reporter constructs containing the proximal promoter of POU5F1, NANOG, or SOX2 were generated (Fig. 5A). In SAS cells, the ectopic CTGF or treatment with rCTGF activated the promoters of POU5F1, NANOG, and SOX2, and mutation of the c-Jun–binding sites abrogated this activation (Fig. 5B and C). Knockdown of JUN in SAS–CTGF cells reduced the CTGF-induced wild-type pluripotency gene promoter activation but had no effect on the c-Jun–binding site–mutated promoters (Fig. 5D–F). Silencing of ESR1 also abrogated the CTGF-induced promoters activation (Supplementary Fig. S8). Here, we also focused on investigating the role of c-Jun in the regulation of POU5F1/NANOG/SOX2 by CTGF. A quantitative chromatin immunoprecipitation (qChIP) assay confirmed the direct binding of c-Jun to the promoters of POU5F1/NANOG/SOX2 in 2 independent SAS–CTGF transfectants (Fig. 5G–I). These results indicate that in CTGF-overexpressing HNSCC cells, c-Jun activates the transcription of POU5F1, NANOG, and SOX2 through direct binding to their promoters.

Figure 5.

CTGF activates POU5F1/NANOG/SOX2 transcription through c-Jun. A, schematic representation of the wild-type and c-Jun–binding sites–mutated promoter constructs of NANOG (Nanog-Luc and Nanog-Luc-mut), POU5F1 (Oct4-Luc and Oct-Luc-mut), and SOX2 (Sox2-Luc and Sox2-Luc-mut). BS1, BS2, and BS3 indicate c-Jun–binding sites. B, luciferase reporter assay. The relative luciferase activity in SAS–CTGF stable cells transfected with the wild-type (top) or mutant (bottom) promoter reporter constructs (n = 3). C, luciferase reporter assay. The relative luciferase activity in SAS cells transfected with the wild-type (top) or mutant (bottom) promoter reporter constructs and treated with an rCTGF 0.2 μg/mL or a vehicle control (n = 3). D–F, relative luciferase activity in SAS-C1 cells cotransfected with the wild-type or mutant promoter reporter construct of NANOG (D), POU5F1 (E), or SOX2 (F), and the vector containing siRNA against c-Jun (si-JUN) or a scrambled sequence (scr; n = 3). G–I, qChIP assay. Top, organization of the NANOG (F), POU5F1 (G), and SOX2 (H) promoter and schematic representation of the primer design. TSS, transcription start site. P1, P2, and P3 indicate the amplified sequences. BS1, BS2, and BS3 indicate the predicted c-Jun–binding sites. IgG was a control of immunoprecipitation experiments. n = 3 for each experiment. In B–I, data represent means ± SEM. *, P < 0.05 by Student t test.

Figure 5.

CTGF activates POU5F1/NANOG/SOX2 transcription through c-Jun. A, schematic representation of the wild-type and c-Jun–binding sites–mutated promoter constructs of NANOG (Nanog-Luc and Nanog-Luc-mut), POU5F1 (Oct4-Luc and Oct-Luc-mut), and SOX2 (Sox2-Luc and Sox2-Luc-mut). BS1, BS2, and BS3 indicate c-Jun–binding sites. B, luciferase reporter assay. The relative luciferase activity in SAS–CTGF stable cells transfected with the wild-type (top) or mutant (bottom) promoter reporter constructs (n = 3). C, luciferase reporter assay. The relative luciferase activity in SAS cells transfected with the wild-type (top) or mutant (bottom) promoter reporter constructs and treated with an rCTGF 0.2 μg/mL or a vehicle control (n = 3). D–F, relative luciferase activity in SAS-C1 cells cotransfected with the wild-type or mutant promoter reporter construct of NANOG (D), POU5F1 (E), or SOX2 (F), and the vector containing siRNA against c-Jun (si-JUN) or a scrambled sequence (scr; n = 3). G–I, qChIP assay. Top, organization of the NANOG (F), POU5F1 (G), and SOX2 (H) promoter and schematic representation of the primer design. TSS, transcription start site. P1, P2, and P3 indicate the amplified sequences. BS1, BS2, and BS3 indicate the predicted c-Jun–binding sites. IgG was a control of immunoprecipitation experiments. n = 3 for each experiment. In B–I, data represent means ± SEM. *, P < 0.05 by Student t test.

Close modal

CTGF promotes tumor growth but inhibits invasiveness in vivo

Next, we evaluated the effect of CTGF on tumor growth and invasiveness in vivo using a xenotransplantion assay. Different doses of TW2.6-siCTGF or TW2.6-scr cells were subcutaneously injected into nude mice. Knockdown of CTGF not only reduced the incidence of tumors but also decreased the volume of the formed tumors (Fig. 6A and B). However, repression of endogenous CTGF in TW2.6 promoted invasion of the tumor cells into the adjacent tissues and reduced the E-cadherin expression in the implanted tumors (Fig. 6C). Finally, we confirmed the experimental results in HNSCC patient samples. Quantitative RT-PCR analysis was conducted on 78 pairs of HNSCC samples to detect the expression of CTGF, NANOG, POU5F1, SOX2, and CDH1. The mRNA level of CTGF positively correlated with the levels of POU5F1, NANOG, SOX2, and CDH1 (Fig. 6D), suggesting the induction of epithelial and pluripotency genes expression by CTGF in the HNSCC samples. CTGF expression was significantly associated with the primary tumor size (Fig. 6E), i.e., T-stage of the American Joint Committee on Cancer staging system (34). However, CTGF expression did not significantly differ between patients with or without metastasis (Fig. 6F). We then analyzed the prognostic impact of CTGF and pluripotency gene expression in the HNSCC cases (Fig. 6G). Patients with both high CTGF and high pluripotency genes expression (group 3) or with low CTGF expression (group 2) tended to have a worse prognosis than those with both high CTGF and low pluripotency genes expression (group 1).

Figure 6.

CTGF induces MET, promotes tumor growth, and attenuates invasiveness in vivo. A, representative pictures of nude mice 6 weeks after injection of TW2.6-siC1 versus TW2.6-scr cells. The black arrows indicate the xenotransplanted tumor. Cell dose = 1 × 105 cells. B, top, a table showing the result of xenotransplantation study (n = 4 for each group). Bottom, comparison of the tumor volume of TW2.6-siC1 or TW2.6-scr formed xenotransplanted tumors (cell dose = 1 × 105; n = 4). C, top, H&E stain showing the implanted tumor. The black arrows indicate the border of implanted tumor, and the gray arrows indicate the muscle infiltration of tumor cells. Scale bar, 500 μm. Middle and bottom, immunohistochemistry of CTGF (middle) and E-cadherin (bottom) in implanted tumor. Scale bar, 200 μm. D, relative mRNA expression levels of POU5F1, NANOG, SOX2, and CDH1 in CTGF low (2−ΔΔCT <1) versus high (2−ΔΔCT ≥1) HNSCC samples (n = 78). The boxplots represent sample maximum (top end of whisker), top quartile (top of the box), median (band in the box), bottom quartile (bottom of the box), and sample minimum (bottom end of whisker). E, relative mRNA level of CTGF in HNSCC samples with different T stages. *, P < 0.05 by Student t test. F, relative mRNA level of CTGF in HNSCC samples with or without lymph node/distant metastasis. G, comparison of the disease-free survival of HNSCC patients with different expression patterns of CTGF and pluripotency genes. P values were estimated by a log-rank test.

Figure 6.

CTGF induces MET, promotes tumor growth, and attenuates invasiveness in vivo. A, representative pictures of nude mice 6 weeks after injection of TW2.6-siC1 versus TW2.6-scr cells. The black arrows indicate the xenotransplanted tumor. Cell dose = 1 × 105 cells. B, top, a table showing the result of xenotransplantation study (n = 4 for each group). Bottom, comparison of the tumor volume of TW2.6-siC1 or TW2.6-scr formed xenotransplanted tumors (cell dose = 1 × 105; n = 4). C, top, H&E stain showing the implanted tumor. The black arrows indicate the border of implanted tumor, and the gray arrows indicate the muscle infiltration of tumor cells. Scale bar, 500 μm. Middle and bottom, immunohistochemistry of CTGF (middle) and E-cadherin (bottom) in implanted tumor. Scale bar, 200 μm. D, relative mRNA expression levels of POU5F1, NANOG, SOX2, and CDH1 in CTGF low (2−ΔΔCT <1) versus high (2−ΔΔCT ≥1) HNSCC samples (n = 78). The boxplots represent sample maximum (top end of whisker), top quartile (top of the box), median (band in the box), bottom quartile (bottom of the box), and sample minimum (bottom end of whisker). E, relative mRNA level of CTGF in HNSCC samples with different T stages. *, P < 0.05 by Student t test. F, relative mRNA level of CTGF in HNSCC samples with or without lymph node/distant metastasis. G, comparison of the disease-free survival of HNSCC patients with different expression patterns of CTGF and pluripotency genes. P values were estimated by a log-rank test.

Close modal

We propose a model to summarize our findings in Fig. 7. In HNSCC cells, CTGF induces the MET and stemness via the following mechanism. CTGF induces the expression of c-Jun through αvβ3, and c-Jun directly activates the transcription of the pluripotency genes POU5F1, NANOG, and SOX2. Upregulation of multiple pluripotency genes results in enhanced stem-like properties and the MET in HNSCC cells. The induction of the MET and stemness by CTGF promotes the local progression but reduces the invasiveness of HNSCC.

Figure 7.

A model depicting the mechanism of CTGF-induced MET and stemness in HNSCC.

Figure 7.

A model depicting the mechanism of CTGF-induced MET and stemness in HNSCC.

Close modal

The phenotypic association between the EMT and cancer stem cells (CSC) has been firmly established. However, recent studies implicate that self-renewal is not bound to the mesenchymal phenotype and that epithelial-type cells are favorable for tumor colonization and proliferation. In prostate cancers, the epithelial gene expression program is enriched in tumor-initiating cells (35). In breast cancers, miR-200, a major miRNA that is downregulated during the EMT and is associated with the epithelial phenotype, facilitates metastatic colonization through the repression of Sec23a (36). The iPSC studies revealed that the MET, the reverse process of EMT, is prerequisite for somatic cell reprogramming (11, 12). In this study, we showed that CTGF induces stem-like properties and local progression in HNSCC through the MET, and this is the first study to show the impact of the secreted protein-induced MET on tumor progression. This finding provides insight for understanding the interplay between epithelial–mesenchymal plasticity and self-renewal during cancer progression and also uncouples the “inevitable” association between the EMT and CSCs. We suggest that CSCs are derived from the EMT and harbor migratory abilities or are derived from the MET with lower invasiveness but higher colonizing abilities.

The critical role of the pluripotency transcription factors (e.g., Oct4, Nanog, and Sox2) in somatic cell reprogramming has been extensively investigated. However, the association between these factors and the EMT in cancer cells is poorly understood. Sporadic reports link a pluripotency gene signature to the EMT. Among these reports, most of them suggest that pluripotency gene expression is associated with the mesenchymal phenotype and cancer invasiveness (37–39). However, 1 report showed the opposite result, i.e., the silencing of Oct4 promotes the EMT of cancer cells (40). In addition to the undefined role of pluripotency factors in the EMT, how these genes are regulated during cancer progression has also been elusive. Here, we identify a novel pathway in which CTGF coordinates the expression of the pluripotency genes through c-Jun to promote the MET. Knockdown of the individual pluripotency factors is not able to completely reverse the MET process, suggesting that the pluripotency factors act collaboratively rather than individually to regulate the epithelial plasticity of cancer cells.

CTGF is known to play distinct roles among different cancers, acting either as an oncoprotein or tumor suppressor (16–20). Here, we show a unique role for CTGF in HNSCC in that CTGF promotes local progression but reduces invasiveness. Because both factors are important for HNSCC prognosis, the impact of CTGF on HNSCC cells is relatively complicated, and the overall impact of CTGF may rely on whether the CTGF–stemness axis is activated. In this study, we found that patients with high CTGF and low pluripotency factor expression had a better prognosis, whereas patients with both high CTGF and pluripotency factor expression and low CTGF expression had a worse prognosis. This result indicates that in patients with high expression of both CTGF and pluripotency genes, the CTGF–stemness axis is activated and the stemness of the tumor cells is increased. In low CTGF cases, the tumors are more mesenchymal-like and invasive. Both of them will have a worse prognosis than those with high CTGF and low pluripotency factor expression. This finding provides a reasonable explanation for the clinical observation that head and neck cancers severely destruct local tissues but rarely metastasize to distal organs (21). According to our results, we suggest that combining the CTGF and pluripotency gene expression profiles in HNSCC will be a better prognostic indicator than an individual marker.

In summary, our study is the first one to show that the induction of pluripotency gene expression by a secreted protein promotes the MET in cancer cells, which leads to local progression but reduces distal dissemination. This conceptual breakthrough provides insight into understanding the interplay between epithelial plasticity and stemness during cancer progression. Furthermore, this report also provides useful information for developing a novel classification system and therapeutic strategy for HNSCC according the expression patterns of CTGF and pluripotency genes.

No potential conflicts of interest were disclosed.

Conception and design: C.C. Chang, W.H. Hsu, C.C. Wang, M.Y.P. Kuo, M.H. Yang

Development of methodology: C.C. Chang, W.H. Hsu, C.C. Wang, C.H. Chou

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C.C. Chang, W.H. Hsu, C.C. Wang, C.H. Chou, M.Y.P. Kuo, S.K. Tai

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C.C. Chang, W.H. Hsu, C.C. Wang, C.H. Chou, M.Y.P. Kuo, B.R. Lin, M.L. Kuo, S.T. Chen, M.H. Yang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): B.R. Lin, S.T. Chen, S.K. Tai

Writing, review, and/or revision of the manuscript: S.T. Chen, M.H. Yang

Study supervision: C.C. Chang, M.L. Kuo, M.H. Yang

This work was supported by the Taipei Veterans General Hospital, National Taiwan University Hospital joint grant (VN101-02 to M.H. Yang and B.R. Lin); National Science Council (101-2321-B-010-015 to M.H. Yang); Excellent Translational Medicine Research Projects of National Taiwan University, College of Medicine, and National Taiwan University Hospital (100-C101-014 to C.C. Chang); a grant from Ministry of Education, Aim for the Top University Plan (M.H. Yang), and a grant from Department of Health, Center of Excellence for Cancer Research (DOH101-TD-C-111-007 to M.H. Yang and DOH101-TD-PB-111-007 to C.C. Chang and M.L. Kuo).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Kalluri
R
,
Weinberg
RA
. 
The basics of epithelial–mesenchymal transition
.
J Clin Invest
2009
;
119
:
1420
8
.
2.
Thiery
JP
,
Acloque
H
,
Huang
RY
,
Nieto
MA
. 
Epithelial–mesenchymal transitions in development and disease
.
Cell
2009
;
139
:
871
90
.
3.
Mani
SA
,
Guo
W
,
Liao
MJ
,
Eaton
EN
,
Ayyanan
A
,
Zhou
AY
, et al
The epithelial–mesenchymal transition generates cells with properties of stem cells
.
Cell
2008
;
133
:
704
15
.
4.
Morel
AP
,
Lièvre
M
,
Thomas
C
,
Hinkal
G
,
Ansieau
S
,
Puisieux
A
. 
Generation of breast cancer stem cells through epithelial–mesenchymal transition
.
PLoS ONE
2008
;
3
:
e2888
.
5.
Nguyen
LV
,
Vanner
R
,
Dirks
P
,
Eaves
CJ
. 
Cancer stem cells: an evolving concept
.
Nat Rev Cancer
2012
;
12
:
133
43
.
6.
Wellner
U
,
Schubert
J
,
Burk
UC
,
Schmalhofer
O
,
Zhu
F
,
Sonntag
A
, et al
The EMT-activator ZEB1 promotes tumorigenicity by repressing stemness-inhibiting microRNAs
.
Nat Cell Biol
2009
;
11
:
1487
95
.
7.
Yang
MH
,
Hsu
DS
,
Wang
HW
,
Wang
HJ
,
Lan
HY
,
Yang
WH
, et al
Bmi1 is essential in Twist1-induced epithelial–mesenchymal transition
.
Nat Cell Biol
2010
;
12
:
982
92
.
8.
Takahashi
K
,
Yamanaka
S
. 
Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors
.
Cell
2006
;
126
:
663
76
.
9.
Yu
J
,
Vodyanik
MA
,
Smuga-Otto
K
,
Antosiewicz-Bourget
J
,
Frane
JL
,
Tian
S
, et al
Induced pluripotent stem cell lines derived from human somatic cells
.
Science
2007
;
318
:
1917
20
.
10.
Park
IH
,
Zhao
R
,
West
JA
,
Yabuuchi
A
,
Huo
H
,
Ince
TA
, et al
Reprogramming of human somatic cells to pluripotency with defined factors
.
Nature
2008
;
451
:
141
6
.
11.
Li
R
,
Liang
J
,
Ni
S
,
Zhou
T
,
Qing
X
,
Li
H
, et al
A mesenchymal-to-epithelial transition initiates and is required for the nuclear reprogramming of mouse fibroblasts
.
Cell Stem Cell
2010
;
7
:
51
63
.
12.
Samavarchi-Tehrani
P
,
Golipour
A
,
David
L
,
Sung
HK
,
Beyer
TA
,
Datti
A
, et al
Functional genomics reveals a BMP-driven mesenchymal-to-epithelial transition in the initiation of somatic cell reprogramming
.
Cell Stem Cell
2010
;
7
:
64
77
.
13.
Redmer
T
,
Diecke
S
,
Grigoryan
T
,
Quiroga-Negreira
A
,
Birchmeier
W
,
Besser
D
. 
E-cadherin is crucial for embryonic stem cell pluripotency and can replace OCT4 during somatic cell reprogramming
.
EMBO Rep
2011
;
12
:
720
6
.
14.
Ocaña
OH
,
Nieto
MA
. 
Epithelial plasticity, stemness and pluripotency
.
Cell Res
2010
;
20
:
1086
8
.
15.
Luft
FC
. 
CCN2, the connective tissue growth factor
.
J Mol Med
2008
;
86
:
1
3
.
16.
Chu
CY
,
Chang
CC
,
Prakash
E
,
Kuo
ML
. 
Connective tissue growth factor (CTGF) and cancer progression
.
J Biomed Sci
2008
;
15
:
675
85
.
17.
Pan
LH
,
Beppu
T
,
Kurose
A
,
Yamauchi
K
,
Sugawara
A
,
Suzuki
M
, et al
Neoplastic cells and proliferating endothelial cells express connective tissue growth factor (CTGF) in glioblastoma
.
Neurol Res
2002
;
24
:
677
83
.
18.
Chen
PS
,
Wang
MY
,
Wu
SN
,
Su
JL
,
Hong
CC
,
Chuang
SE
, et al
CTGF enhances the motility of breast cancer cells via an integrin-αvβ3–ERK1/2-dependent S100A4-upregulated pathway
.
J Cell Sci
2007
;
120
:
2053
65
.
19.
Chang
CC
,
Shih
JY
,
Jeng
YM
,
Su
JL
,
Lin
BZ
,
Chen
ST
, et al
Connective tissue growth factor and its role in lung adenocarcinoma invasion and metastasis
.
J Natl Cancer Inst
2004
;
96
:
364
75
.
20.
Lin
BR
,
Chang
CC
,
Che
TF
,
Chen
ST
,
Chen
RJ
,
Yang
CY
, et al
Connective tissue growth factor inhibits metastasis and acts as an independent prognostic marker in colorectal cancer
.
Gastroenterology
2005
;
128
:
9
23
.
21.
Garavello
W
,
Ciardo
A
,
Spreafico
R
,
Gaini
RM
. 
Risk factors for distant metastases in head and neck squamous cell carcinoma
.
Arch Otolaryngol Head Neck Surg
2006
;
132
:
762
6
.
22.
Yang
MH
,
Chang
CH
,
Lin
BR
,
Chen
ST
,
Lin
SK
,
Kuo
MY
, et al
Connective tissue growth factor modulates oral squamous cell carcinoma invasion by activating a miR-504/FOXP1 signalling
.
Oncogene
2012
;
31
:
2401
11
.
23.
Wong
YK
,
Chang
KW
,
Chen
CF
,
Chang
CS
. 
Characterization of two new cell lines derived from oral cavity human squamous cell carcinomas—OC1 and OC2
.
J Oral Maxillofac Surg
1990
;
48
:
385
90
.
24.
Kok
SH
,
Hong
CY
,
Lin
SK
,
Lee
JJ
,
Chiang
CP
,
Kuo
MY
. 
Establishment and characterization of a tumorigenic cell line from areca quid and tobacco smoke-associated buccal carcinoma
.
Oral Oncol
2007
;
43
:
639
47
.
27.
Prince
ME
,
Sivanandan
R
,
Kaczorowski
A
,
Wolf
GT
,
Kaplan
MJ
,
Dalerba
P
, et al
Identification of a subpopulation of cells with cancer stem cell. Properties in head and neck squamous cell carcinoma
.
Proc Natl Acad Sci USA
2007
;
104
:
973
8
.
28.
Glinsky
GV
. 
Stemness” genomics law governs clinical behavior of human cancer: implications for decision making in disease management
.
J Clin Oncol
2008
;
26
:
2846
53
.
29.
Chambers
I
,
Tomlinson
SR
. 
The transcriptional foundation of pluripotency
.
Development
2009
;
136
:
2311
22
.
30.
Shaulian
E
. 
AP-1-the jun proteins: oncogenes or tumor suppressors in disguise?
Cell Signal
2010
;
22
:
894
9
.
31.
Mork
J
,
Glattre
E
. 
Squamous cell carcinomas of the head and neck in Norway, 1953–92: an epidemiologic study of a low-risk population
.
Cancer Causes Control
1998
;
9
:
37
48
.
32.
Chiang
CT
,
Hwang
YH
,
Su
CC
,
Tsai
KY
,
Lian
IeB
,
Yuan
TH
, et al
Elucidating the underlying causes of oral cancer through spatial clustering in high-risk areas of Taiwan with a distinct gender ratio of incidence
.
Geospat Health
2010
;
4
:
230
242
.
33.
Nakerakanti
SS
,
Bujor
AM
,
Trojanowska
M
. 
CCN2 is required for the TGF-β induced activation of Smad1-Erk1/2 signaling network
.
PLoS ONE
2011
;
6
:
e21911
.
34.
Greene
FL
,
Page
DL
,
Fleming
ID
,
Fritz
A
,
Balch
CM
,
Haller
DG
, et al
Head and neck
. In: 
American Joint Committee on Cancer (AJCC) cancer staging manual. 6th ed.
,
New York, NY
:
Springer
; 
2002
. p.
17
68
.
35.
Celià-Terrassa
T
,
Meca-Cortés
O
,
Mateo
F
,
de Paz
AM
,
Rubio
N
,
Arnal-Estapé
A
, et al
Epithelial–mesenchymal transition can suppress major attributes of human epithelial tumor-initiating cells
.
J Clin Invest
2012
;
122
:
1849
68
.
36.
Korpal
M
,
Ell
BJ
,
Buffa
FM
,
Ibrahim
T
,
Blanco
MA
,
Celià-Terrassa
T
, et al
Direct targeting of Sec23a by miR-200s influences cancer cell secretome and promotes metastatic colonization
.
Nat Med
2011
;
17
:
1101
8
.
37.
Kong
D
,
Banerjee
S
,
Ahmad
A
,
Li
Y
,
Wang
Z
,
Sethi
S
, et al
Epithelial to mesenchymal transition is mechanistically linked with stem cell signatures in prostate cancer cells
.
PLoS ONE
2010
;
5
:
e12445
.
38.
Chiou
SH
,
Wang
ML
,
Chou
YT
,
Chen
CJ
,
Hong
CF
,
Hsieh
WJ
, et al
Coexpression of Oct4 and Nanog enhances malignancy in lung adenocarcinoma by inducing cancer stem cell-like properties and epithelial–mesenchymal transdifferentiation
.
Cancer Res
2010
;
70
:
10433
44
.
39.
Han
X
,
Fang
X
,
Lou
X
,
Hua
D
,
Ding
W
,
Foltz
G
, et al
Silencing SOX2 induced mesenchymal–epithelial transition and its expression predicts liver and lymph node metastasis of CRC patients
.
PLoS ONE
2012
;
7
:
e41335
.
40.
Hu
J
,
Qin
K
,
Zhang
Y
,
Gong
J
,
Li
N
,
Lv
D
, et al
Downregulation of transcription factor Oct4 induces an epithelial-to-mesenchymal transition via enhancement of Ca2+ influx in breast cancer cells
.
Biochem Biophys Res Commun
2011
;
411
:
786
91
.