STMN2 overexpression promotes cell proliferation and EMT in pancreatic cancer mediated by WNT/β-catenin signaling

STMN2, as a key regulator in microtubule disassembly and dynamics, has recently been shown to participate in cancer development. However, the corresponding role in pancreatic ductal adenocarcinoma (PC), to our knowledge, has not been reported yet. In the current study, we systematically investigate the potential role of STMN2 in the progression of PC in vitro and vivo. Overexpression of STMN2 was prevalently observed in 81 human cases of PC tissues compared with that in the paired adjacent pancreas (54.3% vs 18.5%, P < 0.01), which was positively associated with multiple advanced clinical stages of PC patients (tumor size, T stage, lymph-node metastasis and the poor prognosis). Meanwhile, a close correlation between high STMN2 and cytoplasmic/nuclear β-catenin expression (P = 0.007) was observed in PC tissues and cell lines. STMN2 overexpression induced EMT and cell proliferation in vitro via stimulating EMT-like cellular morphology, cell motility and proliferation, and the change of EMT (Snail1, E-cadherin and Vimentin) and Cyclin D1 signaling. However, XAV939 inhibited STMN2 overexpression-enhanced EMT and proliferation. Conversely, KY19382 reversed STMN2 silencing- inhibited EMT and cell proliferation in vitro. Furthermore, activated STMN2 and β-catenin were co-localized in cytoplasm/nuclear in vitro. β-catenin/TCF-mediated the transcription of STMN2 by the potential binding sites (TTCAAAG). Finally, STMN2 promoted subcutaneous tumor growth following the activation of EMT and Cyclin D1 signaling. STMN2 overexpression promotes the aggressive clinical stage of PC patients and promotes EMT and cell proliferation in vitro and vivo. β-catenin/TCF-mediated the transcription of STMN2.


INTRODUCTION
Pancreatic ductal adenocarcinoma (PC) is the most prevalent neoplastic disease of the pancreas accounting for more than 90% of all pancreatic malignancies. It is one of the most fatal digestive cancers, with a 5-year survival rate of <10% [1]. It would overtake breast cancer as the third leading cause of cancer death by 2025 in Europe [2] and would become the 2nd most cause of cancer-related death in the US by 2030 [3]. The malignant biological behavior of PC, such as the intense invasion and rapid metastase, contributes to the unfavorable outcomes of PC patients. One of a critical driving factor is epithelial-to-mesenchymal transition (EMT). EMT provides cancer cells with a dramatic cytoskeleton rearrangement and metastatic phenotype characterized by the loss of the epithelial phenotype (Ecadherin) and the gain of mesenchymal properties (N-cadherin and Vimentin) [4]. Thus, it is urgent to define the molecular mechanism target EMT during tumor development.
STMN2, as a neuronal growth-associated protein of Stathmin family [5], plays a significant role in neuronal growth, microtubule dynamics, cell motility and signaling regulation [6][7][8][9]. Decreased STMN2 have been associated with Down's syndrome and Alzheimer's diseases [10], whereas increased STMN2 participates in the progression of hepatocellular [11], neuroblastoma [12] and ovarian cancer [13]. However, its potential role in the development of PC, to our knowledge, has not been reported yet.
The WNT/β-catenin signaling pathway, as a classic and conserved signal pathway, participates in multiple physiological processes, including cell proliferation, differentiation, apoptosis, polarity, mobility and homeostasis [14]. Dysregulation of the WNT/β-catenin pathway is implicated in various cancers. Meanwhile, the WNT/β-catenin signaling is an indispensable component to drive EMT [15]. Previous study showed that STMN2 was a novel target of β-catenin/TCF-mediated transcription in human hepatoma cells [16,17]. Taken together, we systematically investigate the potential role of STMN2 in regulating malignant behavior of PC in combination with WNT/β-catenin pathway in vitro and vivo, which supplies a promising gene-targeted therapy for PC.

Clinical human samples and PC cell lines
This study was approved by the academic committee at the First hospital of China Medical University with the agreement of specimen consent signed by each patient. The study methodology has been admitted by the ethics committee from the same institution. Due to the different tumor pathology, patients with endocrine carcinoma, acinar cell carcinoma and invasive intraductal papillary mucinous carcinoma were excluded from this study. Finally, 81 cases of PC and paired adjacent pancreas were picked up from postoperative patients from 2010 to 2020 which were pathologically diagnosed as pancreatic ductal adenocarcinoma. PANC-1, BxPC-3, and SW1990 cells were purchased from the cell culture collection in Chinese Academy of Sciences. Capan-2 cells were purchased from the cell bank of the American Type Culture Collection.

EMT construction
In order to enhance EMT phenotype, STMN2-GFP and GFP transfected PANC-1 and Capan-2 cells were pre-cultured with medium containing 1% FBS for 24 h. Then cells were pretreated with XAV939 (20uM, Selleckchem, USA) for 12 h. Similarly, STMN2 silencing BxPC-3 and SW1990 cells were pretreated with KY19382 (1 μM, MedChemExpress, USA) for 24 h. One percent DMSO was used as the vehicle. We evaluated EMT model from three characters: EMT-like cellular morphology, cell motility and the activation of EMT signaling.

Transwell assays
Based on our previous study [19], STMN2-GFP and GFP transfected PANC-1 and Capan-2 cells were pretreated with XAV939 (20uM, Selleckchem, USA) for 12 h, while STMN2 silencing BxPC-3 and SW1990 cells were pretreated with KY19382 (1uM, MedChemExpress, USA) for 24 h. Cells were implanted into membrane inserts (BD Biosciences) covered with 10%matrigel with free serum medium. Medium containing 10%FBS was covered at the bottom. The crossed cells were calculated in at least 5 random fields/ well (×200). The migration assay was conducted in the similar way without matrigel. Transwell was repeated in triplicates.

XTT assay
The XTT assay provides an easy to use tool for cell viability and has more sensitivity and specificity than CCK-8 and MTT assays. Thus, XTT (Abcam, ab232856) was used to investigate the effect of STMN2 in regulating cell proliferation with different time points combining with XAV939 (20 μM for 12 h repeated three times) or KY19382 (1 μM for 24 h repeated twice) treatments. STMN2-GFP, GFP, si2-STMN2 and siCtrl transfected PC cells (the density of 4000 viable cells per well) were seeded into 96-well plates and incubated for 1-5 days. 50 µl XTT labeling mixture were finally added to each well and incubate for 4 h at 37°C at the end of time point. 96-well plates were finally measured at a wavelength of 450 nm in an ELISA 96-well microtiter plate reader (BIORAD680, USA). Mean OD from GFP or siCtrl group was used as the control following the time point-to-point comparison manner. The formula used for the cell viability assay was expressed as: % cell growth = [Mean OD (sample from 2 to 4 days) − Mean OD blank ] / [Mean OD (sample in 1 day) − Mean OD blank ] × 100.

In vivo xenograft model
Animals were kept according to the Animal Care Committee of China Medical University. Ten cases of 8-week-old nude mice (BALB/c, female, Beijing Vital River Laboratory Animal Technology Co., Ltd. China) were acclimatized for a week. The investigator was blinded to the group allocation and all animals were randomly assigned in two groups (n = 5 in each group). STMN2-GFP and GFP-transfected Capan-2 (5 × 10 6 ) cells were subcutaneously transplanted into the subcutaneous axilla, respectively. A cotton swab was used to avoid leakage from the injection site. 3 weeks later, all mice (n = 10) were treated with carbon dioxide for euthanasia. The following formula was used to calculate tumor size: length × width × height × 0.52 in millimeters. The final samples were extracted for late hematoxylin and eosin (HE) and IHC staining.

Statistical analysis
Based on our previous study [19], non-parametric paired, chi-squared and spearman testes were used to analyze the clinical and statistical data of IHC assays. The Kaplan-Meier curve in univariate analysis and Cox regression tests in multivariate analysis were used to analyze the prognostic data. The difference of the results involving WB, qRT-PCR, transwell and tumor size were represented as means ± standard deviation and were compared via independent t-test. P-value is regarded statistically significant as: *P < 0.05; **P < 0.01.

RESULTS
Overexpression of STMN2 was closely associated with the clinicopathological characters of PC patients In IHC assays, STMN2 was mainly localized in cytoplasm and nuclear in PC and adjacent pancreas (Fig. 1A). STMN2 was overexpressed in human PC specimens compared with that in adjacent paired pancreas (44/81,54.3% vs 15/81, 18.5%, P < 0.01) (Fig. 1A). Low and high expression of STMN2 was defined in classic PC sample #8 and #15, respectively (Fig. 1A). Interestingly, PC patients with STMN2 overexpression was accompanied with cytoplasmic and nuclear expression of β-catenin (Fig. 1B, C). Normally, β-catenin showed membrane expression in normal pancreas (#3) and few cases of PC samples (#7), while most PC specimens exhibited cytoplasmic and nuclear expression of β-catenin (#25) (Fig. 1B). According to previous study [20], membrane/negative and cytoplasmic/nuclear expression of β-catenin was regarded as normal and abnormal expression, respectively. PC samples with STMN2 overexpression was associated with β-catenin abnormal expression (#25) in most serial sample slices, and vice versa (#7) (Fig. 1B, C), which is further verified by the Spearman's rank correlation coefficient (Table 1).
The mRNA level of STMN2 was also much higher in PC specimens in contrast with paired adjacent pancreas (P < 0.01) ( Fig. 2A). In 4 PC cell lines, both STMN2 and β-catenin protein (Fig.  2B) and mRNA (Fig. 2C) level were significantly higher in BxPC-3 and SW1990 cells than that of the two other PC cell lines. It is well known that Nuclear β-catenin is a key inducer of EMT [21]. The tight relationship between STMN2 and β-catenin in human PC tissues and cell lines drive us to focus on their potential function in regulating EMT in vitro and vivo.
WNT/β-catenin signaling mediated STMN2-promoted cell proliferation in vitro We next investigated the potential role of STMN2 in cell proliferation in vitro by XTT assays. In contrast with the GFP (scramble) group corresponding to the same time point, STMN2 overexpression promoted cell proliferation in PANC-1 cells in timedependent manner, especially in 4-5 culture of days (Fig. 5A). However, XAV939 reversed STMN2 overexpression-promoted cell proliferation in vitro at the same time point (Fig. 5A) compared with that in STMN2-GFP group. The similar experiment was also repeated in STMN2 overexpressing Capan-2 cells (Fig. 5B). Conversely, in contrast with the siCtrl (scramble) group, STMN2 silencing inhibited cell proliferation in BxPC-3 cells at the same time point (Fig. 5C). However, KY19382 reversed STMN2 silencing-inhibited cell proliferation in vitro compared with that in si2-STMN2 group (Fig. 5C). The similar experiment was also observed in STMN2 silencing SW1990 cells (Fig. 5D). Taken together, STMN2 promoted cell proliferation in vitro mediated by WNT/β-catenin signaling.
STMN2 regulating EMT and Cyclin D1 signaling mediated by WNT/β-catenin signaling We next investigated the potential mechanism of STMN2 in regulating EMT and cell proliferation in vitro. WB showed that STMN2 overexpression upregulated Vimentin, Snail1, and Cyclin D1, but downregulated E-cad expression in PANC-1 (Fig. 6A) and Capan-2 ( Fig. 6B) cells in contrast with corresponding GFP groups. β-catenin and N-cadherin expression was unchanged. XAV939 not only specially inhibited β-catenin and STMN2 expression, but also reversed STMN2 overexpression-induced the change of EMT and Cyclin D1 expression in STMN2-GFP plus XAV939 groups compared with that in STMN2-GFP groups (Fig. 6A, B). Conversely, STMN2 silencing downregulated Vimentin, Snail1, and Cyclin D1, but upregulated E-cad expression in BxPC-3 (Fig. 6C) and SW1990 cells (Fig. 6D) in contrast with corresponding siCtrl groups. KY19382 not only specially activated β-catenin and STMN2 expression, but also reversed STMN2 silencing-inhibited the change of EMT and Cyclin D1 expression in si2-STMN2 plus KY19382 groups compared with that in si2-STMN2 groups (Fig. 6C,  D). Meanwhile, upon KY19382, activated β-catenin and STMN2 were co-localized in the cytoplasm and nuclear in BxPC-3 cells by IF (Fig. 7A). To observe whether β-catenin directly interacts with STMN2 promoter, ChIP assays were conducted using an antibody against β-catenin in BxPC-3 cells pretreated with KY19382. Three potential TCF binding sites upon the promoter of STMN2 were shown in Fig. 7B. The immunoprecipitated DNA was detected by the specific primer pairs (F1-R1, F2-R2 and F3-R3) corresponding to each TCF binding site (Fig. 7B, C). Upon immunoprecipitation with anti-β-catenin, the DNA fragment containing the "F1-R1" TCF site was amplified at a significantly higher level from the chromatin of KY19382-activated BxPC-3 cells. However, the other two primer pairs (F2-R2 and F3-R3) did not show any increased amplification upon β-catenin activation (Fig. 7C). Therefore, the TCF binding site at −1816 to −1822 is crucial and specific for β-catenin/TCFmediated the regulation of STMN2.
STMN2 promoted subcutaneous tumor size in vitro STMN2-GFP and GFP transfected Capan-2 stable cells were used for subcutaneous tumor model. The subcutaneous tumor size and volume in STMN2-GFP group (n = 5) was significantly increased in contrast with GFP group (n = 5) (P < 0.05) (Fig. 8A,  C). HE staining confirmed the tumor pathology (Fig. 8B). IHC further showed that STMN2, Vimentin, Cyclin D1 and Ki67 expression were remarkably upregulated in STMN2-GFP group in contrast with the scramble GFP group (Fig. 8D-F). β-catenin showed abnormal (cytoplasm and nuclear) and normal (membrane) expression in STMN2-GFP and GFP group, respectively (Fig. 8E, F). Taken together, a tight relationship of STMN2 with EMT and Cyclin D1 signaling were prevalently existed in clinical PC samples and in vitro and vivo.

DISCUSSION
Previous studies pay much more attention on the function of STMN1 in several cancers, including hepatocellular, gastric, colon, pancreatic and lung cancers [23][24][25][26][27]. However, STMN2, as a novel discovered oncogene, is poorly understood, especially in PC. In current study, we first identified STMN2 as a novel target of β-catenin/TCF-mediated transcription in PC. Overexpression of STMN2 contributes to the aggressive clinical stage of PC patients in coordination with WNT/β-catenin signaling. Meanwhile, WNT/ β-catenin-targeted STMN2 promotes cell proliferation and EMT in PC via activating EMT and Cyclin D1 signaling, which has not been reported yet, to our knowledge. We first found that STMN2 was overexpressed in PC patients, which was positively associated with tumor size, T stage, lymph node metastasis and the poor survival of PC patients. STMN2 is also overexpressed in hepatocellular, neuroblastoma and ovarian cancer [11][12][13], which is associated with advanced clinical characters and worse prognosis in hepatocellular cancer [11]. Meanwhile, it is an independent unfavorable prognostic factor in ovarian cancer [13]. Thus, STMN2 trends to act as a potential oncogene in several cancers. It was noteworthy that the combination of high STMN2 and cytoplasmic/nuclear β-catenin expression contributed to the much worse survival of PC patients. Meanwhile, the parallel expression of STMN2 and  β-catenin were observed in both PC tissues and cell lines. WNT/ β-catenin signaling is well known as a classic tumor signal pathway involving with the characteristic of EMT and proliferation potency [15,28], which drive us to investigate the coordinate function of STMN2 and WNT/β-catenin signaling in the malignant biology of PC. In current study, STMN2 overexpression promoted EMT and cell proliferation in vitro. EMT-like cell morphology, cell mobility and  proliferation were significantly enhanced in STMN2 overexpressing PC cells, which was reversed by the WNT/β-catenin signaling inhibitor XAV939. Conversely, the WNT/β-catenin signaling activator KY19382 reversed STMN2 silencing-inhibited EMT and cell proliferation. Only one study reports that STMN2 promotes cell migration, invasion and metastasis in hepatocellular cancer by triggering EMT [11]. Taken together, STMN2 promotes EMT and proliferation in cancer development.
Further potential mechanism showed that STMN2 overexpression upregulated Snail1, Vimentin, Cyclin D1, but  downregulated E-cad in vitro. XAV939 not only inhibited STMN2 expression, but also reversed STMN2 overexpression-induced the change of EMT and Cyclin D1 signaling. Conversely, KY19382 reversed STMN2 silencing-induced EMT and Cyclin D1 signaling in vitro. Snail1, as a critical EMT stimulator, induces EMT by repressing E-cad and claudins with concomitant upregulation of Vimentin [29]. Thus, STMN2 induced EMT partially by regulating Snail1 signaling. Similarly, STMN2 mediates nuclear translocation of Smad2/3 and enhances TGFβ signaling by destabilizing microtubules to promote EMT in hepatocellular cancer [11]. It is well known that Cyclin D1 plays a crucial role in regulating proliferation to cell cycle progression [30]. High Cyclin D1 expression drives unchecked cellular proliferation promoting tumor growth [31]. Therefore, STMN2 promoted cell proliferation by activating Cyclin D1 signaling.
Previous study showed that STMN2 was a novel target of β-catenin/TCF-mediated transcription in human hepatoma cells [16,17]. Similarly, the oncogenic function of STMN2 in PC was mediated by WNT/β-catenin signaling in current study. Activated β-catenin and STMN2 were co-localized in the cytoplasm and nuclear in BxPC-3 cells. ChIP assays further showed that TCF binding site at −1816 to −1822 upon STMN2 promotor is the crucial transcriptional site by β-catenin/TCF. Taken together, overexpression of STMN2 promotes cell proliferation and EMT in PC mediated by WNT/β-catenin signaling.
Finally, STMN2 overexpression promoted subcutaneous tumors formation in vivo with the overexpression of EMT and Cyclin D1 signaling, which was consistent with the results in vitro.

CONCLUSION
In conclusion, we first identified STMN2 as a novel target of β-catenin/TCF-mediated transcription in PC cells. Overexpression of STMN2 contributes to the advanced clinical stage of PC patients in coordination with WNT/β-catenin signaling. Meanwhile, WNT/β-catenin targeted STMN2 promotes cell proliferation and EMT in PC via activating EMT and Cyclin D1 signaling. STMN2 would serve as a promising prognostic biomarker and potential therapeutic gene target for PC.