Carboxyl-Terminal Modulator Protein Positively Acts as an Oncogenic Driver in Head and Neck Squamous Cell Carcinoma via Regulating Akt phosphorylation

The exact regulatory mechanisms of carboxyl-terminal modulator protein (CTMP) and its downstream pathways in cancer have been controversial and are not completely understood. Here, we report a new mechanism of regulation of Akt serine/threonine kinase, one of the most important dysregulated signals in head and neck squamous cell carcinoma (HNSCC) by the CTMP pathway and its clinical implications. We find that HNSCC tumor tissues and cell lines had relatively high levels of CTMP expression. Clinical data indicate that CTMP expression was significantly associated with positive lymph node metastasis (OR = 3.8, P = 0.033) and correlated with poor prognosis in patients with HNSCC. CTMP was also positively correlated with Akt/GSK-3β phosphorylation, Snail up-regulation and E-cadherin down-regulation, which lead to increased proliferation and epithelial-to-mesenchymal transition, suggesting that CTMP expression results in enhanced tumorigenic and metastatic properties of HNSCC cells. Moreover, CTMP suppression restores sensitivity to cisplatin chemotherapy. Intriguingly, all the molecular responses to CTMP regulation are identical regardless of p53 status in HNSCC cells. We conclude that CTMP promotes Akt phosphorylation and functions as an oncogenic driver and prognostic marker in HNSCC irrespective of p53.

Head and neck squamous cell carcinomas (HNSCCs) are the sixth most common cancer worldwide in men and occur as a heterogeneous tumor with an aggressive phenotype 1 . Despite the advances in biology and medicine over the past several decades, HNSCC remains a major cause of morbidity and mortality due to its relatively poor prognosis. Even with current treatment strategies, more than 50% of patients die from HNSCC or related conditions within 5 years 2 . This is most likely due to a lack of understanding about the molecular basis of HNSCC, and a lack of biomarkers that predict HNSCC progression or therapeutic resistance 3 . However, the development of HNSCC is characterized by multistep carcinogenic processes in which the activation of oncogenes and inactivation of tumor suppressor genes, including p53, epidermal growth factor receptor, Ras, MYC, survivin, cyclin D1, and cyclin-dependent kinase inhibitor, occurs as a result of genetic and epigenetic alterations. These alterations result in the proliferation and aggressiveness of tumor cells 4 .
Epithelial-to-mesenchymal transition (EMT) is a complex cellular process that is intimately linked to aggressiveness of cancer cells such as metastasis or resistance to chemotherapy 5 . Therefore, understanding EMT biology is essential to improve patient outcome. Previously, it is reported that both invasion and metastasis may be critically dependent on the acquisition by the incipient cancer cell of EMT features 6 . More recently, primary HNSCC

Materials and Methods
HNSCC patients. We retrospectively reviewed the medical charts of 119 HNSCC patients who had undergone curative surgery (primary resection and appropriate cervical lymph node (LN) dissection according to disease stage) at the Department of Otolaryngology-Head and Neck Surgery of Chungnam National University Hospital from April 1999 to December 2011. This study was approved by the Institutional Review Board of Chungnam National University College of Medicine (Jung-gu Daejeon, Korea), and the informed consent requirement was waived. All experiments relating human tissue were performed in accordance with our institutional guidelines. Clinicopathological patient characteristics are summarized in Table 1. Of the patients, 40 (33.6%) had oral cavity cancer, 20 (16.8%) had oropharyngeal cancer, 11 (9.2%) had hypopharynx cancer, and 48 (40.4%) had larynx cancer. Tumor size and stage were classified according to the TNM system published by the American Joint Committee on Cancer (AJCC), and tumor differentiation was classified according to the World Health Organization (WHO) classification system. The mean follow-up duration was 40.6 months (range: 2-144 months).
Cell lines and reagents. RNA isolation and RT-PCR. Total cellular RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA, USA), reverse transcribed, and amplified using specific primers for CTMP and glyceraldehyde 3-phosphate dehydrogenase (GAPDH), as described previously 16  Cell proliferation assay. After transfection with control or CTMP cDNA or siRNA, cells were plated at a density of 5 × 10 3 /well in serum-free culture medium. After 48 h incubation for SNU1041 cells and 72 h for SCC15 cells, viability was measured using the Cell Proliferation Reagent WST-1 (Roche Diagnostics Corporation, Indianapolis, IN, USA) as described previously 7,17 . Results are presented as percentages relative to control cells. Western blot analysis. Cells were lysed in lysis buffer containing 150 mM NaCl, 1.0% nonidet-P 40

Co-immunoprecipitation.
To identify the interaction between Akt and CTMP, co-immunoprecipitation was performed using a mouse monoclonal antibody to Akt (Cell Signaling Technology). For immunoprecipitation, 200 μ g of cell extract was precleared overnight at 4 °C with 10 μ L of protein G-Sepharose beads pre-blocked with phosphate-buffered saline containing 20 mg/mL bovine serum albumin. One microgram of antibody was then added and samples were incubated at 4 °C for 2 h, followed by a 2 h incubation with 10 μ L of pre-blocked protein G-Sepharose beads on a rotator. Beads were washed five times in lysis buffer. Samples were resolved by SDS-PAGE and detected by ECL.
Statistical analysis. All statistical analyses were performed using SPSS for Windows statistical software (ver. 20.0; SPSS Inc., Chicago, IL, USA). Pearson's chi-squared or Fisher's exact tests were used to determine the relationship between CTMP expression and clinicopathologic parameters. Factors deemed of potential importance by univariate analyses were included in a multinominal logistic regression analysis to adjust for other factors. Survival curves were constructed using the Kaplan-Meier method, and a log-rank test was used to determine statistical significance.
All in vitro experiments were repeated three times, and statistical significance was analyzed using two-sided Student's t-test or one-way analysis of variance (ANOVA) followed by Tukey's post hoc test. Data are presented as means ± standard deviation (SD), and a P value < 0.05 was considered statistically significant (* P < 0.05; * * P < 0.01; * * * P < 0.001). In addition, we assessed whether CTMP expression varied between normal and tumor tissue derived from the same patients at mRNA and protein levels. RT-PCR and Western blot analysis revealed that tumor tissues had higher CTMP expressions at both mRNA and protein levels compared to normal tissues (Fig. 1B,C). Therefore, we can assume that CTMP can be related with carcinogenesis in HNSCC.

CTMP expression in
CTMP expression independently correlates with lymph node metastasis in HNSCC patients. We next evaluated the relationship between CTMP expression and clinicopathological characteristics. Of the various parameters described in Table 2, T classification, lymph node metastasis, and AJCC stage were significantly associated with CTMP expression. The three characteristics mentioned above were then modeled together using multinomial logistic regression to adjust for other factors. As shown in Table 3, tumors with high CTMP expression significantly increased the odds of lymph node metastasis by more than three-fold (odds ratio [OR] = 3.788, 95% confidence interval [CI] = 1.113-12.896, P = 0.033).

CTMP expression is associated with poor prognosis in HNSCC.
To examine the potential correlation between high CTMP expression and HNSCC survival rate, disease-free survival (DFS) and overall survival (OS) curves were calculated using the Kaplan-Meier method and compared using a log-rank test. As shown in Fig. 2A,B, patients with high CTMP expression had significantly lower DFS and OS rates than patients with low CTMP expression (P = 0.004 and P = 0.000, respectively). The 5-year DFS rate was 34.0% with high CTMP expression, but was 67% with low CTMP expression. The 5-year OS rate was 34.0% with high CTMP expression and 75.0% with low CTMP expression. Collectively, these data suggest that CTMP is a potential molecular biomarker for HNSCC aggressiveness and prognosis.

CTMP expression in HNSCC cell lines.
To confirm our observations from clinical data, we examined CTMP protein levels in a panel of HNSCC cell lines (FaDu, SNU1041, SNU1076, and SCC15) using normal fibroblasts (hFB) for comparison (Fig. 3). All HNSCC cells demonstrated relatively high CTMP expression compared to hFB, to a greater or lesser extent. Since CTMP was upregulated in tumors compared to normal tissues, we hypothesized that CTMP may be involved in HNSCC tumorigenesis. The SNU1041 and SCC15 cell lines were used for subsequent experiments because they derive from different head and neck areas (pharynx and oral cavity, respectively) and have different p53 genetic background, the most commonly mutated gene both in HNSCC tumors and cell lines 18 (p53-wild-type 19 or -mutated 20 , respectively).

CTMP promotes proliferation of HNSCC cells.
The finding that T stage III and IV had a higher proportion of CTMP high than those of earlier stages led us to hypothesize that CTMP may drive cellular proliferation. To examine this hypothesis further, we performed a proliferation assay after CTMP up-/down-regulation in SNU1041 and SCC15 cell. As shown in Fig. 4, CTMP-overexpressing cells showed significantly increased proliferation (26.4%, P < 0.01; and 13.4%, P < 0.05, respectively, Fig. 4A,C) whereas CTMP-suppressed cells showed markedly decreased proliferation in both cell lines (20.5%, P < 0.01; and 11.7%, P < 0.001, respectively, Fig. 4B,D) suggesting contribution of CTMP signaling to cancer cell proliferation.
CTMP positively regulates HNSCC cell invasiveness. Since CTMP-high tumors almost always presented with lymph node metastases (Table 1), we speculated that CTMP may promote invasive properties of HNSCC cells, which helps tumor cells to travel to a lymph nodes or distant site and to settle there. To test this in vitro, we performed a Transwell invasion assay using Matrigel, Invasion of cancer cells requires degradation of the basement membrane and extracellular matrix, cytoplasmic extension, and cell migration. Transwell invasion assay imitate this environment for tumor invasion. The attached cells in the lower section passed through the filter of the chamber, indicating invasive cells. As shown in Fig. 5, the number of cells that moved to the bottom chamber was significantly increased in CTMP-overexpressing SNU1041 and SCC15 cells (193.3%, P < 0.001; and 23.0%, P < 0.01, respectively, Fig. 5A,C), but markedly decreased in CTMP-suppressed cells compared to control cells (48.7%, P < 0.01 and 31.0%, P < 0.01, respectively, Fig. 5B,D) suggesting contribution of CTMP signaling to cancer cell migration and invasion.

CTMP promotes epithelial-to-mesenchymal transition (EMT), associated with an increase in
Snail. Within a primary tumor, a subpopulation of cells can undergo EMT, which confers novel migratory, invasive, and stem-like properties that promote metastasis. To examine whether CTMP promotes EMT, we examined both epithelial and mesenchymal markers and the well-known EMT-inducing transcription factors, Snail and Slug, by immunoblot analysis. As shown in Fig. 6A,C, CTMP-overexpressing cells exhibited a significant decrease in E-cadherin, while the mesenchymal markers vimentin and Snail were increased in both SNU1041 and SCC15 cells. In addition, CTMP-silenced cells showed the exact opposite tendency; up-regulation of E-cadherin and down-regulation of vimentin and Snail (Fig. 6B,D). However, there were no significant changes in the expression of Slug (Fig. 6). These data mean that CTMP promotes Snail-associated EMT.

CTMP binds Akt in HNSCC cells and upregulates Akt signaling.
Akt signaling is one of the most important oncogenic driver in a wide range of human cancers including HNSCC 9,21 , thus understanding the regulatory mechanism of Akt is important for developing therapeutic tactics against cancers. Although, CTMP   Table 3. Multinomial logistic regression for the association of CTMP expression with T classification, lymph node metastasis, and AJCC stage. Exp (β) indicates odd ratio; CI, confidence interval; LN, lymph node; AJCC, American Joint Committee on Cancer. * P < 0.05 between the two categories for a given variable. has been emerged as one of Akt regulating protein, there are disparate views regarding the direction of regulation and no report about the association between CTMP and Akt regulation 10,11 . To evaluate the association between CTMP expression and Akt activation, we evaluated Akt phosphorylation at Ser 473 after CTMP up-/down-regulation. Although, Akt is activated via two sequencial phosphorylation (Thr 308 is followed by Ser 473), we examined only Ser 473 phosphorylation, because additional phosphorylation at Ser 473 is essential to accomplish Akt full activation in spite of Thr 308 phosphorylation 9 . As shown in Fig. 7A,C, CTMP up-regulation markedly enhanced Akt phosphorylation without changing total Akt amount. In contrast, CTMP down-regulation significantly attenuated Akt phosphorylation with stationary expression of total Akt (Fig. 7B,D). Moreover, the phosphorylation of GSK-3β, the downstream protein kinase of Akt, at Ser 9 was increased or decreased after CTMP-overexpression/-suppression, respectively, indicating that CTMP regulates Akt activity (Fig. 7A-D). Furthermore, to determine the direct interaction between CTMP and Akt, co-immunoprecipitation was performed. The cell lysate of SNU1041 and SCC15 cell lines was precipitated using an anti-Akt antibody. CTMP was detected in the anti-Akt-IP products but CTMP was not detected in normal Ig G immunoprecipitates, indicating that there is a direct association between CTMP and Akt proteins (Fig. 7E,F). Taken together, CTMP binds Akt and upregulates Akt pathway in SNU1041 and SCC14 cells.

CTMP suppression increases cisplatin cytotoxicity in HNSCC cells. Cells that undergo EMT often
acquire properties associated with cancer stem cells, which are related to chemoresistance 22 . Therefore, given that CTMP is associated with EMT in HNSCC cells, we hypothesized that CTMP affects chemoresistance and aimed to determine the role of CTMP silencing on cisplatin efficacy in SNU1041 and SCC15 cells. As shown in Fig. 8, even 50 μ M of cisplatin did not significantly attenuate cell viability in SNU1041 or SCC15 cell lines, both of which have high endogenous CTMP expression. However, although CTMP knockdown alone did not have significant To sum up our in vitro data, CTMP acts as an oncogenic driver in HNSCC via promoting Snail-associated EMT and Akt phosphorylation. Moreover, CTMP suppression increased chemosensitivity.

Discussion
In general, the transformation of normal epithelium to squamous cell carcinoma (SCC) occurs in multiple steps, involving the sequential activation of oncogenes and the inactivation of tumor suppressor genes 23 . Although remarkable progress has been made in the identification of altered tumor onco-/suppressor-genes and their related protein products in HNSCC 23 , the proliferative pathways driving uncontrolled cell growth are still poorly defined, which limits our ability to identify more comprehensive mechanism-based therapeutic approaches. In addition, given the minimal success of EGFR targeting strategy and the commonly occurring resistance 24 , the need for additional treatment options that improve outcome of existing therapeutic tactics for HNSCC is pressing.
Akt (also called protein kinase B) is a major downstream target of receptor tyrosine kinases that signal via phosphatidylinositol 3-kinase. Recently, dysregulation of Akt activity has been found in a wide range of human cancers including HNSCC, and the Akt signaling pathway has emerged as a key regulator of cell growth and decision of cell fate 25,26 . Of the genetic or epigenetic alterations known in HNSCC, one of the most frequently altered pathways is EGFR/Ras/PI3K, which, along with other pathways (p53/DNp63, pRb/CycD1, TGF-β /Smad and NF-κ B), can lead to Akt dysregulation which has indeed been found in 20~60% of tumor samples and in the majority of HNSCC-derived cell lines 25,[27][28][29] . Moreover, Akt overexpression is predictive of poor clinical outcome, and is associated with advanced disease, local recurrence, and decreased survival [30][31][32] . Given the importance of Akt in HNSCC, understanding the regulatory mechanisms of Akt is important to develop new therapeutic strategies against HNSCC.
Akt activation is managed by various interacting proteins and upstream regulators. Since Maira et al. published the original study describing 27-kDA CTMP as an Akt-interacting protein in 2001 11 , accumulating evidence suggests that CTMP binds to the Akt carboxy terminus and negatively regulates its activity in lung cancer 33 , glioblastoma 34,35 , pancreatic adenocarcinoma 36 , and cervical cancer cell lines 37 , suggesting CTMP is a tumor suppressor. However, opposite results were found by Ono et al. 38 . In addition, Liu et al. found that CTMP is a positive regulator of Akt phosphorylation and is an oncogenic driver in breast cancer 9 . These conflicting results could be due to different biological responses between different cell types. Therefore, determination of the role of CTMP in HNSCC is needed, and no previous studies have examined this. Our results show that CTMP functions as a positive regulator of Akt, and has oncogenic activity in HNSCC.
We also found that there is a biological link between CTMP and cellular invasion of HNSCC cells, which is essential in tumor progression and metastasis. Since HNSCC is characterized by local invasion and lymphatic metastasis, understanding the molecular mechanisms that mediate tumor invasion and metastasis is critical for the identification of novel therapeutic targets. To verify the effect of CTMP-Akt regulation on invasive cellular phenotypes at the molecular level, we analyzed protein levels of Slug and Snail, transcription factors and master regulators of EMT, and the levels of E-cadherin and vimentin, cellular machinery proteins associated with the invasion and thus hallmarks of EMT 8,39 . In our data, CTMP upregulation decreased E-cadherin expression and increased the expression of vimentin and Snail without altering Slug signaling, suggesting Snail-dependent/ Slug-independent E-cadherin regulation and EMT induction in HNSCC cell lines. Although, further studies are needed to figure out potential association with other EMT activating transcription factors such as Twist and Zeb1/2, to the best of our knowledge, this is the first evidence that CTMP modulates EMT in cancer. In addition, consistent with the in vitro EMT results, we demonstrated that CTMP was upregulated in human HNSCC and was positively associated with LN metastasis and poor prognosis in HNSCC patients.
Recently, Akt has been shown to repress the transcription and protein expression of E-cadherin, resulting in EMT 40 . Grille et al. showed that SCC cells producing a constitutively active form of Akt produce a transcription factor, Snail, which represses the expression of the E-cadherin gene 40 . In addition, Akt induces the inactivation (phosphorylation) of GSK-3β , which suppresses the phosphorylation of Snail. This results in Snail nuclear localization and up-regulation (due to protein stabilization) and E-cadherin downregulation, which subsequently results in EMT 41,42 . The role of the Akt/GSK-3β/Snail pathway in EMT has been reported previously in hepatocellular carcinoma 43 and in a gefitinib-resistant HNSCC cell line 44 . Consistent with the previous reports, we found that Akt/GSK-3β and Snail were simultaneously up/down regulated by CTMP, while E-cadherin expression had the opposite reaction. Thus, as a downstream of CTMP, the Akt/GSK-3β/Snail/E-cadherin pathway is likely involved in EMT in HNSCC cell lines. Additionally, in consistent with previous study showed that downregulation of E-cadherin enhances proliferation of HNSCC through transcriptional regulation of EGFR 45 , CTMP up-regulation, which significantly decreased E-cadherin expression in both SNU1041 and SCC15 cell lines, promoted cell proliferation and vice versa.
Intriguingly, we also found that both cell lines, despite harboring different p53 genetic backgrounds, demonstrated the same molecular response to CTMP up-/down-regulation in all experiments. Given that gene expression differences are related to heterogeneous biological responses such as sensitivity to chemotherapy and malignancy 46 , and that p53 mutations are involved in resistance to platinum based 47 and EGFR-targeting therapies 48 , these findings suggest that CTMP targeting stratagem may possess general anticancer efficacy regardless of p53 status. Thus, we can highlight that further studies should be performed to determine the potential of CMTP targeting to overcome chemoresistance.
In conclusion, our data suggest that CTMP functions as a positive regulator of Akt and facilitates HNSCC invasion such as LN metastasis by regulating EMT in a Snail-dependent manner indicating the oncogenic activity of CTMP in HNSCC, and its expression is an independent predictor of clinical prognosis and response to platinum-based chemotherapy in HNSCC. Given the data in this study, examining CTMP further in patient tumor tissues could identify patients at risk for tumor progression and chemotherapy resistance. Moreover, if validated in further studies including in vivo and clinical trial, a CTMP-targeting agent partnered with an existing anticancer modality could be a novel tactic to overcome drug resistance and ultimately improve HNSCC outcomes.

Figure 8. CTMP suppression increases cisplatin cytotoxicity in HNSCC cells.
Following exposure of the indicated cells to various concentrations of cisplatin for 24 h, cell viability was measured using a WST-1 proliferation assay. Although the cell lines were resistant to cisplatin treatment, CTMP siRNA-transfected cells demonstrated marked cisplatin-induced cytotoxic effects compared with control siRNA-transfected cells. Data represent the means ± SD of three independent experiments. NS, not significant; * P < 0.05; * * P < 0.01; * * * P < 0.001.