CHN1 Promotes EMT via the Akt / GSK-3β / Snail Pathway in Cervical Carcinoma

Background: Metastasis and invasion are key points to lead the mortality of cervical squamous cell carcinoma (CSCC). Epithelial–mesenchymal transition (EMT) is now a universal explanation for the mechanisms of tumor metastasis. Α-chimeric protein (α-chimaerin, CHN1) plays an important role in the regulation of signal transduction and development. However, the molecular regulatory relationship between CHN1 and CSCC progression involved EMT has not been identied. Methods: The expression of CHN1 in CSCC tissues, adjacent tissues, and lymph node metastases of CSCC patients was detected by immunohistochemistry assay. Upregulation and knockdown of CHN1 was achieved by transfecting the plasmid into CSCC cells. The effect of CHN1 on proliferation was determined by CCK-8 and plate clone formation assay. Changes in migration and invasion capabilities were evaluated by scratch migration assay and trans-well invasion assay. The effect of CHN1 overexpression and interference on xenograft tumor growth was determined by measuring tumor weight and pathological analysis. The expression of EMT-related mRNA was measured by qRT-PCR assay in transfected CSCC cells. EMT-related proteins and Akt / GSK-3β / Snail signaling pathway-related proteins were also evaluated by using western blotting assay. Results: CHN1 was overexpression in CSCC tissues and was associated with lymph node metastasis and low survival of CSCC patients. The overexpression CHN1 promoted cell proliferation, migration and invasion bility in CSCC cells. On the contrary, silencing CHN1 inhibited these phenomena. In vivo experiments, the overexpression of CHN1 promoted tumors formed in xenograft tumor mouse model, with increased volume and weight of xenograft tumor. In addition, CHN1 induced the expression of EMT related transcription factors, accompanied by the decreased expression of epithelial markers and the increased expression of mesenchymal markers. The Akt / GSK-3β / Snail signaling pathway was activated by the overexpression of CHN1 in vitro, and the The and on the and Three independent experiments performed and GAPDH and


Background
Cervical carcinoma is one of the main malignancy threatening the health and life of women worldwide [1]. With the development of early diagnosis and treatment in cervical carcinoma, the incidence of cervical carcinoma in developed countries has declined signi cantly [2], but the incidence and mortality of cervical carcinoma among young women in developing countries remain high [1,3].
The metastatic spread of cancer cells from primary site to distant area, such as lymph nodes around the vessels of pelvic wall, is the major cause of unsuccessful treatment and ultimate death in cervical cancer [4,5]. Accumulated evidence demonstrated that the epithelial-mesenchymal transition (EMT) is responsible for the invasion and metastasis of various carcinomas and had been related to elevated resistance of chemotherapy and immunotherapy [6][7][8]. Despite enormous efforts veri ed that EMT reprogramming of tumor epithelial cells depends on the systematical occurrence at multiple regulation levels [6,9], the exact underlying mechanisms connect EMT, metastasis and cervical carcinoma remain to dig for a better understanding of cancer progression and exploring the related therapeutic methods.
Chimaerin (CHN1) is one kind of ras-related Rho GTPase-activating protein (RhoGAPs) involves in cytoskeletal regulation [10][11][12]. Two chimaerin isoforms, α and β, exist in mammalian genomes and each of them consists of at least two splice variants: a full-length type 1 transcript and a truncated type 2 transcript. Although a cysteine-rich domain followed by the GAP domain is shared by the two transcripts, a SH2-domain encompassed N-terminal is characterized in type 2 isoforms [13][14][15]. CHN1 is widely investigated in neurobiology, and it is pivotal in neuronal signal-transduction, brain development, synaptogenesis and cognitive ability [12,[16][17][18][19]. Besides, it also involved in the regulation of T cell adhesion and chemotaxis [20], transmission of signals in tumor progression [21], maintenance of epithelial morphology [22], development of oculomotor projection [23], and regulation of Duane retraction syndrome [19,24,25]. However, the role of CHN1 in tumorigenesis has not been elucidated clearly, especially in cervical carcinoma.
The aim of this study is to investigate the relationship between differential expressed CHN1 and the progression and metastasis of cervical carcinoma, and meanwhile reveal the potential mechanism of how CHN1 contribute to the related effect. In the present study, we prove that CHN1 is highly expressed in cervical carcinoma with a correlation of long-term prognosis. CHN1 can accelerate the occurrence of tumor biological behaviors, and it may stimulate the activation of EMT through Akt / GSK-3β / Snail pathway to promote the metastasis of cervical carcinoma.

Tissue samples and patients
Two para n-embedded tissue microarrays (TMA), one consisted of 62 paired SCC/adjacent noncarcinoma samples and another contained 84 CSCC and 8 lymph nodes from metastatic cervical carcinoma samples, were obtained from Shanghai Outdo Biotech (Shanghai, China) and Alena Bio (Xi'an, China), respectively. The other para n embedded CSCC samples were histopathologically diagnosed at PLA 251 Hospital. These patients did not undergo radiation and chemotherapy before the surgery and samples were stored at -80 ℃.

Cell lines, cell culture and gene transfection
The CSCC cell line SiHa was preserved by our lab. The cells were cultured in DMEM/F12 medium (Thermo Fisher Scienti c, USA) supplied with 10% fetal bovine serum (FBS) (Thermo Fisher Scienti c, USA) and maintained at 37 ℃ in a 5% CO 2 contained humidi ed atmosphere. The recombinant fulllength CHN1 gene expression plasmid and CHN1 shRNA interfered plasmid were synthesized and obtained from GenePharma Company (GenePharma, Shanghai, China). The cells were transfected with empty vector plasmid as negativ controls. Transfection was performed in the 24-well plate (Corning Life Science, USA) by Lipofectamine 3000 (Invitrogen, USA) following the instruction, and the e cacy of transfection was assessed through western blotting. Stable transfection clones were selected by 400ug/ml G418(Sigma, USA) for 14 days.

Immunohistochemistry (IHC) and hematoxylin & eosin (H&E) staining
The para n-embedded sections were de-waxed, hydrated, and antigen retrieved in 0.01M citrate solution at 95 ℃ for 20 min. Then they were treated with 3% hydrogen peroxide for 10 min to block the endogenous peroxidase. The mouse anti-human CHN1 monoclonal antibody (1:150, Abcam, UK) and rabbit anti-human Snail polyclonal antibody (1:500, Proteintech Group, USA) were incubated respectively followed by the serum blocking. A PV-6002 kit (ZSGB-Bio, Beijing, China) was used for the following steps. The specimens were observed and captured under an inverted microscope (Leica, Wetzlar, Germany). The tumors harvested from the xenograft nude mice were xed and embedded in para n and their pathological features were analyzed through H&E staining (Solarbio Life Science, Beijing, China).

RNA extraction, reverse transcription and quantitative real-time PCR (qRT-PCR)
Total RNA was isolated using TRIzol reagent (Invitrogen, USA) from cultured cells and prepared according to the instruction manual. The complementary DNA was reverse transcribed according to the PrimeScript RT reagent Kit (TaKaRa, China) and qPCR was performed with the SYBR Premix Ex Taq II (Tli RNaseH Plus) Kit (TaKaRa, China) on an ABI 7500 Real-time PCR system (Applied Biosystems, USA). All genes were conducted in triplicate and relative expression was normalized to the housekeep gene β-actin. The primers used here are listed in Supplementary Table1.

Western blotting
The extracted equal amounts of protein were separated through polyacrylamide SDS gels (SDS-PAGE) and transferred to polyvinylidene uoride (PVDF) membranes (Thermo Fisher Scienti c, USA). Then they were probed with different primary antibodies including CHN1 (Abcam, UK), Fibronectin (Invitrogen, USA), β-catenin, Snail (two from CST, USA), vimentin and E-cadherin (two from Proteintech Group, USA) and the blot was detected with peroxidase-conjugated secondary antibodies (ZSGB-Bio, China). The signals were acquired by chemiluminescence kit (Merckmillipore, Germany). β-tublin and GAPDH (two from Transgen Biotech, China) were used as the internal control.

Cell proliferation and migration assays
Cellular growth curves and colony-forming assay were used to evaluate the cell proliferation rate. Brie y, cells were seeded, cultured, and incubated with CCK-8 Kit (Dojindo, Japan) and the optical density ( and subcutaneously injected into the oxter of each mouse. 5 mice were used for each group. Mice were anesthetized, sacri ced after 8 weeks, and tumors were harvested, measured and stained. All the mice used here were bred at the SPF Animal Laboratory of Institute of Science and Technology, National Research Institute for Family Planning. Statistical analysis SPSS 18.0 statistical software (IBM Corporation, USA) was used for data analysis. Data were expressed as mean ± standard deviation (SD) after at least three independent experiments. Pearson chi-square test was assessed for clinical correlation analysis and Kaplan Meier method was performed to analyze the overall survival rate of cervical cancer patients. The difference between the two groups was tested by Student's t-test. As long as P-value < 0.05, the results were statistically signi cant.

Results
Overexpression of CHN1 in CSCC was related to metastasis and was negatively connected with the overall survival time of patients In order to investigate the expression level of CHN1, western blotting was used for testing its expression in both CSCC and paracancerous tissues. It proved that CHN1 was highly expressed in tumors compared to the corelated non-carcinoma tissues in 5 pairs of CSCC and matched adjacent tissues (Fig. 1a). The location of CHN1 was mainly concentrated in the cytoplasm by IF staining (Fig. S1a) and IHC assays (Fig. S1b). IHC was further operated to con rm its expression status in a 62-paired CSCC/ adjacent noncarcinoma TMA and three pairs were shown as representatives (Fig. S1c). Among those samples, a strong positive signal of CHN1 was obviously observed in high, middle and low differentiated CSCC tissues ( Fig. 1b) with an overall positive rate of 85.5% (53/62), which was signi cantly different from the 43.5% (27/62) proportion in non-carcinoma tissues (Table 1). However, CHN1 mainly expressed in the cervical epithelium base of adjacent non-carcinoma tissues due to the rapid cell division and active cell proliferation (Fig. 1b). To further study the patient survival time and CHN1 status, we collected another 65 samples of CSCC patients for the IHC detection, then 25 of them were participated in the follow-up study for the assessment of survival time. The overall survival rate was varied among the patients with differentially expressed CHN1. Compared to the negative expressed CHN1, both the strong positive and positive expression of CHN1 presented a shorter overall survival time, with the shortest time in the strong positive group (Fig. 1c). The median survival time of the strong positive patients was 48 months. These data demonstrated that the overexpressed CHN1 might negatively concerned with the overall survival time of patients. Subsequently, the clinicopathological signi cance of CHN1 expression was then analyzed by chi-square test in the 128 CSCC tissue assay. As shown, CHN1 level was connected with lymph node metastasis (P = 0.003), but not with the age, clinical stage, and histological differentiation of patients (Table 2). Since the pelvic lymph node metastasis is one of the common types during cervical cancer progression, we then detected the correlation between the metastatic lymph nodes and CHN1 expression by IHC. It was presented that CHN1 also highly expressed in the metastatic lymph nodes of cervical cancer tissue, which indicated that CHN1 might be connected with cervical cancer metastasis (Fig. 1d). Up-regulation of CHN1 promotes the proliferation of SiHa cell in vitro and enhances tumorigenicity in vivo As CHN1 was highly expressed in cervical carcinoma tissues, the in uence of its up regulation on cell proliferation was next tested by establishing a CHN1 stable transfected model in SiHa cell line. After con rmed by western blotting, the CHN1-overexpressed cells had been successfully acquired when compared to the vector-only transfected cells (Fig. 2a). Then these cells had been tested for their ability of proliferation both in vitro and in vivo. As presented, an increasing of CHN1 level, coupled with a signi cantly enhanced cell proliferation, was observed in the CHN1-overexpressed SiHa cells compared to the vector-transfected group by using the CCK-8 assay (Fig. 2b); the same as the above-mentioned results, the colony-forming assay further veri ed that the numbers of colonies in SiHa-CHN1 were much more than that in SiHa-vector (Fig. 2c). These data demonstrated that the overexpression of CHN1 may increase the proliferation capacity of SiHa cells in vitro. Besides, the tumorigenicity of SiHa-CHN1 and SiHa-vector was followed in nude mouse for testing their ability of proliferation in vivo. 8 weeks post xenograft, tumors could be observed in both of these two groups. However, the ones formed by CHN1 upregulated SiHa cells showed a much bigger weight than the vector group with a statistical signi cance (p = 0.001, Fig. 2d). The H&E staining certi ed that the tumors consisted by these two kinds of cells presented an extinction of cell polarity in the parenchyma, enlarged cell nuclei, pathological mitoses, and apparent cell heteromorphism, all of which are the distinctive features of tumor cells (Fig. 2e, up panel). In contrast to SiHa-vector, the further IHC data showed that CHN1 was positively expressed in SiHa-CHN1 formed tumors (Fig. 2e, down panel), which implied that the xenograft tumors might be caused by the upregulated CHN1. Taken together, those data demonstrated that the increased CHN1 protein level may promote cell proliferation both in vitro and in vivo.

Down-regulation of CHN1 inhibits the proliferation of SiHa cell in vitro and reduces tumorigenicity in vivo
Besides the up-regulation, CHN1 was then interfered by using two CHN1 targeted shRNAs, shCHN1-1 and shCHN1-2, a scrambled shRNA was involved as a negative control (shVector), and the ability of cell proliferation was again detected both in vitro and in vivo. Post western blotting detection, the protein level of CHN1 was strongly decreased both in shCHN1-1 and shCHN1-2 compared to shVector (Fig. 3a). Meanwhile, the proliferation of CHN1 down-regulated cells were signi cantly inhibited tested by the CCK-8 assay (Fig. 3b), as well as their ability of colony-forming (Fig. 3c). The in vivo tumorigenicity assay further con rmed that, even though the tumors developed in all the three groups, the ones observed in shCHN1-1 and shCHN1-2 were signi cantly smaller than that in shVector (Fig. 3d).

Interference of CHN1 inhibits invasion and migration ability of SiHa cells
The capacity of cell migration was tested simultaneously in SiHa cells with CHN1 knockdown. The transwell assay proved that CHN1 knockdown could signi cantly inhibit the cell invasion (Fig. 4a); consistently, this inhibition of migration was also observed in wound-healing assay (Fig. 4b).
Taken together, these results demonstrated that the invasion and migration of SiHa cells could be inhibited by the down regulation of CHN1 in vitro, as well as the tumorigenicity in vivo.

CHN1 induces EMT though AKT / GSK-3β / Snail pathway
Since EMT is one of the key elements for cancer metastasis, as mentioned before, the inhibition of cell migration was correlated with CHN1 knockdown, the expression of two crucial markers of EMT, epithelial marker E-cadherin and mesenchymal marker Vimentin, were then detected to reveal the connection between EMT and down-regulated CHN1. Immuno uorescent staining showed that, in contrast to the control, increased E-cadherin and decreased Vimentin were observed in CHN1 interfered cells (Fig. 5a). In addition, the level of CHN1 and Snail, another EMT relevant transcription factor, was tested by IHC in serial sliced 24 CSCC samples to further verify the relativity of their expression. It could be seen that both of them were brown stained with the similar localization in tumor samples (Fig. 5b), which implied that the level Snail might related to the expression of CHN1 during the cervical cancer development.
The qRT-PCR analysis showed that an increased CHN1 gene level could decrease the level of E-cadherin and β-catenin, but increased the level of Vimentin, Fibronectin, and Snail (Fig. 5c). Western blotting on the other hand, presented the same tendency in protein level that overexpressed CHN1 not only increased the expression of Snail, Fibronectin, Vimentin, but also decreased the expression of E-cadherin and β-catenin ( Fig. 5d). But when CHN1 was interfered, the expression of those EMT related markers showed a reversed pattern (Fig. 5e).
It was published that activated AKT is essential for the induction of EMT through the inhibition of GSK-3β, which leads to the stabilization and nuclear localization of Snail to trigger the migration of cells and EMT [26]. In that case, several key factors involved in the ATK / GSK-3β / Snail pathway were detected by western blotting and the results showed that the expression of phosphorylated AKT, phosphorylated GSK-3β, and Snail was enhanced when CHN1 was up-regulated (Fig. 5f). To further make sure the effect of CHN1 was via the activation of the ATK / GSK-3β / Snail pathway, the expression of those three proteins were tested again under the presence of the PI3K inhibitor LY294002. It demonstrated that the level of phosphorylated AKT, phosphorylated GSK-3β, and Snail were effectively reduced with the attendance of LY294002 (Fig. 5f).

Discussion
The progression of cervical cancer is a multifactor and multistep involved process. Even though the techniques of cervical screening and diagnostic have made some progress, as well as the improvement of its vaccine, it is now still one of the gynecologic malignant tumors worldwide that the effective way to restrain its development and to inhibit its relapse and metastasis still needs to be gured out.
In this study, the carcinogenesis of CHN1 is investigated to gure out its role during the occurrence and development of cervical carcinogenesis. We found that the overexpression of CHN1 in cervical carcinogenesis was related to a higher metastasizing degree and pathological stage, as well as total survival rate. The further in vitro functional study demonstrated that the metastasis and tumorigenesis of SiHa cells could be signi cantly enhanced by the increased level of CHN1. In contract, an inhibited effect was observed when its expression was interrupted. The nude mice in vivo study also supported the same impacts that higher level of CHN1 promoted the tumorigenesis while lower level inhibited it. One of the main reasons to cause the recurrence and death of CSCC patients is due to the distant metastasis, for example, pelvic lymph node [27][28][29]. EMT plays a crucial role during the lymph node metastasis of cervical cancer [30]. It has been reported that an increased tumor progression, invasion, metastasis, and distortion of epithelial integrity could be observed if the primary cervical cancer coupled with the occurrence of EMT [31]. Then it is important to test whether EMT is involved in the tumorigenesis of CHN1. According to our data, SiHa cells, with an increased expression of CHN1, promoted the down regulation of epithelial markers, but induced the up regulation of mesenchymal markers in both RNA and protein levels. However, the expression of these EMT markers was signi cantly reserved when CHN1 level was disrupted in SiHa cells.
In addition, the location of CHN1 was found related to the position of Snail, a key EMT principal transcription factor, by the IHC staining of lymph nodes with metastases of CSCC. The expression of Snail is regulated by the level of CHN1, an increased expression of CHN1 correlated with a high level of Snail while a decreased expression of CHN1 was relevant to a low level of Snail.
It is published that the activation of Akt / GSK-3β / Snail pathway involves in the occurrence of EMT in cervical cancer [32,33]. Our further results showed that the phosphorylated level of Akt and GSK-3β, as well as the expression of Snail, could be raised when the protein level of CHN1 was strongly enhanced, and the same tendency happened when CHN1 was attenuated, which suggesting that CHN1 might be required for the activation of Akt / GSK-3β / Snail to promote the metastasis of cervical carcinoma.

Conclusions
Taken together, overexpression of CHN1 promoted the proliferation, migration and invasion of CSCC cells in vitro. In addition, CHN1 overexpression was signi cantly relevant to high metastatic degree, low survival rate and poor prognosis in CSCC patients. CHN1 could induce EMT through the Akt / GSK-3β / Snail pathway, as well as enhance metastasis and progression of cervical carcinoma. CHN1 might be considered as a novel marker of cervical carcinoma for the clinical diagnosis of metastasis and poor prognosis. It could be a potential target for a better cervical carcinoma management with a further digging of its characters.  CHN1 overexpression promotes tumorigenicity of SiHa cells. a CHN1 was overexpressed in SiHa cells after transfection. GAPDH was used as the internal control. b Cell proliferation was enhanced after the overexpression of CHN1 in SiHa cells (data represent means ± SD, n=3 **p<0.01). c Foci formation was effectively promoted after CHN1 overexpressed in SiHa cells (data represent means ± SD, n=3 **p<0.01). d Xenograft tumors were formed after CHN1 and empty vector transfection in nude mice. Weight assessment of xenograft tumors (right, n=5 for each group, **p<0.01). e Representative H&E and IHC staining of CHN1 expression in CHN1-and empty vector-transfected xenograft tumors (original magni cation, 400×).

Figure 3
CHN1 interference inhibits the tumorigenicity of SiHa cells. a CHN1 expression was reduced after shRNA knockdown. b Cell proliferation was inhibited after CHN1 knockdown in SiHa cell line (data represent means ± SD, n=3 **p<0.01). c Foci formation was inhibited after the interference of CHN1 expression in SiHa cells (data represent means ± SD, n=3 **p<0.01). d Smaller xenograft tumors were generated after the injection of SiHa-shCHN1 cells in the nude mice. Weight assessment of xenograft tumors (right, n=5 for each group, **p<0.01).  c The expression of EMT markers and EMT-related transcription factors were affected after up regulation