SKA1 regulates actin cytoskeleton remodelling via activating Cdc42 and influences the migration of pancreatic ductal adenocarcinoma cells

Abstract Objectives Spindle and kinetochore–associated protein 1(SKA1), originally identified as a protein essential for proper chromosome segregation, has been recently linked to multiple malignancies. This study aimed to explore the biological, clinical role and molecular mechanism of SKA1 in pancreatic carcinogenesis. Materials and Methods SKA1 expression was detected in 145 pancreatic ductal adenocarcinoma (PDAC) specimens by immunohistochemistry. Biological behaviour assays were used to determine the role of SKA1 in PDAC progression in vitro and in vivo. Using isobaric tags for relative and absolute quantitation (iTRAQ), SKA1’s downstream proteins were examined. Moreover, cytochalasin B and ZCL278 were used to explore the changes of SKA1‐induced signalling and cell morphology, with further confirmation by immunoblotting and immunofluorescence assays. Results Increased SKA1 expression was significantly correlated with tumour size and cellular differentiation degree in PDAC tissues. Furthermore, elevated levels of SKA1 reflected shorter overall survival (P = .019). As for biological behaviour, SKA1 acted as a tumour promotor in PDAC, overexpression of SKA1 facilitates cell proliferation, migration and invasion in vitro and in vivo. Mechanistically, we demonstrated that SKA1 enhanced pancreatic cancer aggressiveness by inhibiting G2/M arrest and regulating actin cytoskeleton organization via activating Cdc42. Conclusions This study revealed novel roles for SKA1 as an important regulator of actin cytoskeleton organization and an oncogene in PDAC cells, which may provide insights into developing novel therapeutics.

is characterized by high malignancy and early metastatic potential, but the specific mechanisms involved are still elusive.
Consequently, early detection and novel therapeutic strategies are urgently needed.
Epithelial-mesenchymal transition (EMT) is the earliest event in tumour metastasis, and involves morphological changes of epithelial cells, accompanied by loss of cell polarity, disassembled tight junctions, looser cell-cell adhesion and increased motility. All these events contribute to key molecular changes and are relevant to challenges of PDAC treatment. 2 Understanding how a tumour cell experience EMT and establish itself at distant locations is indispensable, especially for cancers with early metastasis. One area of focus in EMT-related cellular events has been the role of the dra-  [3][4][5][6] Microtubules, as the largest cytoskeletal components, govern intracellular trafficking, organelle positioning and molecular signalling, also affect the interactions with other cytoskeletal elements. 7,8 Hence, many anti-microtubule agents (eg Taxanes) were derived, which increased survival in pancreatic cancer. 9 However, their simultaneous deleterious activity towards normal cells is common because of low specificity, serious side effects are inevitable. Meanwhile, whether factors associated with microtubules as well as their associated proteins play a role in EMT of PDAC cells remains largely unknown. Therefore, we take this as the starting point to seek therapeutic targets with higher specificity and less toxicity.
Based on two independent microarray profiling datasets (TGF-β stimulation-induced EMT PANC-1 cell line and control group) from the Gene Expression Omnibus and The Cancer Genome Atlas (TCGA) database. Spindle and kinetochore-associated protein 1 (SKA1) was focused on, because of its essential effects of promoting proper chromosome segregation and regulating the stability of both mitotic spindle and cytoskeletal microtubules in PDAC. [10][11][12][13] Recently, SKA1 was found to overexpressed in several malignancies and its oncogenic role has been demonstrated, including cell cycle distribution, chromosomal instability, cell proliferation and metastasis. [14][15][16][17] Additionally, SKA1 overexpression results in centrosome amplification in human prostate epithelial cells, which facilitates non-tumorigenic epithelial cells switch to tumorigenic in nude mice, 18 indicating that SKA1 may represent a master regulator of carcinogenesis. However, currently, the expression and molecular function of SKA1 in human pancreatic cancer remain undefined.
In this study, we determined the high expression and possible tumorigenic role of SKA1 in pancreatic cancer cells in vitro and in vivo, and the underlying mechanisms and signalling pathways were explored by iTRAQ assay. We firstly found that SKA1 overexpression induces Cdc42 activation, which in turn promotes actin cytoskeleton remodelling, providing a potential therapeutic strategy to overcome PDAC.

| Patients and the clinical cohort
A total of 145 formalin-fixed and paraffin-embedded sections of human PDAC specimens were retrospectively collected from patients who underwent pancreatectomy without preoperative radiation or chemotherapy in Ruijin Hospital (Shanghai, China) before 2019. Histological diagnosis was performed by two experienced pathologists independently. This study was approved by the ethics committee of Ruijin Hospital.

| Lentivirus-mediated stable transfection
SKA1 interference and overexpression lentiviruses were constructed with GV208 by GeneChem Company. PDAC cells were seeded into 6-well plates, and lentiviral transfection with experimental constructs or respective controls (1 × 10 9 TU/mL) was carried out at 60% confluency, according to the manufacturer's protocol. After 48-72 hours, infection efficiency was detected and SKA1 expression levels were verified. To establish stably transfected cells (sh-SKA1, sh-ctr, SKA1 and vector), cells were further cultured in 10 μg/mL puromycin containing medium.

| Cell proliferation, migration and invasion assays
For the cell proliferation MTT assay, detailed procedures are provided as we have previously described. 19 Transwell migration assay was performed using transwell chambers (24-well, 8 μm pore size; Corning), and cell invasion assay was conducted with BD Falcon Cell culture inserts coated with BD matrigel matrix (BD Bioscience). About 2 × 10 5 cells were seeded in the upper chambers with DMEM, and DMEM supplemented with 10% FBS was added to the lower chambers. The plates were incubated at 37°C for 24 hours. Cells that did not migrate or invade through the pores were removed with cotton swabs. Cells on the lower side of the filter were fixed with 10% formalin and stained with crystal violet, finally counted to evaluate the migratory and invasive abilities.

| Wound-healing assay
Wound-healing assays were performed in 6-well plates with confluent cells, with scratches generated using 200 μL pipette tips. The wells were then washed three times with the culture medium and cultured for an additional 24 or 48 hours, followed by the assessment of relative wound closure areas.

| Immunoblotting and quantitative real-time PCR (qRT-PCR)
Immunoblotting procedures were described in our previous study. 20 Cytoskeletal fractions were obtained using a ProteoExtract Subcellular Proteome Extraction Kit (Merck Millipore) according to the manufacturer's instruction. Primary antibodies are shown in Table S1.
Immunoreactivity was detected by the Enhanced Chemiluminescence kit (GE Healthcare), and visualization was performed with the G BoxChemic XL system (Syngene). GAPDH was used as an internal reference control for the relative protein expression levels using the Quantity One software (Bio-Rad). For qRT-PCR, total RNA was iso-

| Cdc42-GTP pull-down assay
Cdc42 activation was examined using the Cdc42 Activation Assay Kit (ab211163, Abcam) following the manufacturer's instructions. Cells were harvested with cell lysis buffer, 293 cell lysate loaded with GDP or GTPγS and incubated with PAK1 PBD Agarose beads were as negative or positive controls, respectively. One mg of protein lysate in a 1 mL total volume at 4°C was immediately precipitated with 40 μL of PA1K-PBD beads for 60 minutes with rotation. After washing, beads were resuspended and processed for immunoblotting.

| Histology, immunocytochemistry and immunofluorescence
Immunohistochemical experiment was operated using standard techniques according to our previous study. 21 To evaluate SKA1 expression levels, a semi-quantitative immunoreactivity score analysis method scoring both the percentage of positive cells and staining intensity was used. The percentage of positive cells was scored from 0 to 3 (0, 0%; 1, 1%-49%; 2, 50%-70%; and 3, >70%); staining intensity was scored from 0 to 3 (0, negative; 1, weak staining; 2, intermediate staining; and 3, strong staining). The two sub-scores were added to determine the final score. We defined the samples having a final score ≤ 3 as low expression and samples, with a score of >3 as high expression. Each slide was evaluated double-blindedly by two experienced pathologists independently. For Immunofluorescence staining, procedures were described in our previous study, 22 and primary antibodies are shown in Table S1.
All slides were examined by confocal microscopy (Zeiss710), and photometric values were analysed by ImageJ.

| In vivo subcutaneous xenograft and metastasis assay
All mouse care and handling procedures were conducted in accordance with the recommendations of the Institutional Animal Care and Use Committee of Shanghai Experimental Animals Centre, Chinese Academy of Sciences. Briefly, PANC-1 cells as well as Capan-1 cells stably transfected with sh-SKA1, sh-ctr, SKA1 or empty vector were washed in 0.15 mL serum-free DMEM. In xenografts' assay, 3 × 10 6 cells were subcutaneously transplanted into the right or left flank of 4-to 5-week-old female BALB/c nude mice (n = 5/group), and tumour volume was monitored until the mice were sacrificed 30 days later. For metastasis analysis, 1 × 10 6 cells were intravenously injected into the tail vein to study lung metastasis (n = 10/ group); 12 weeks later, the mice were euthanized, and lung samples stained with haematoxylin and eosin were evaluated under a microscope.

| iTRAQ and bioinformatic analysis
We performed Isobaric Tags for Relative and Absolute Quantitation (iTRAQ) with stably transfected sh-SKA1 and sh-ctr in PANC-1 cells.
After protein extraction and trypsin digestion, the peptides were desalted with a Strata X C18 SPE column (Phenomenex) and vacuum dried. Peptides were reconstituted in 0.5 mol/L TEAB and processed according to the manufacturer's protocol for TMT/iTRAQ kit. Briefly, one unit of TMT/iTRAQ reagent was thawed and reconstituted in acetonitrile. The peptide mixtures were then incubated for 2 hours at room temperature and pooled, desalted and dried by vacuum centrifugation. Subsequently, bioinformatic analyses (GO annotation, KEGG pathway annotation, protein-protein interaction network, etc) were performed.

| Statistical analysis
Statistical analyses were performed with the SPSS 20.0 software (SPSS Inc). Excel and GraphPad Prism version 6 (GraphPad Software)

| Increased SKA1 is significantly correlated with poor survival outcomes in PDAC
SKA1 (identified as EMT-related dysregulated protein in PDAC cells, see Figure S1) is expressed in various types of tumours; however, whether SKA1 expression is involved in pancreatic cancer  (Table 1). Furthermore, we found that SKA1 expression was significantly correlated with PDAC cellular differentiation degree ( Figure 1A).    Figure 2B). In addition, SKA1 promotes PDAC cells proliferation was also evidenced by colony formation and cell apoptosis assays ( Figure S2).

| SKA1 enhances PDAC proliferation in vitro and in vivo by inhibiting G2/M arrest
Next, we examined cell cycle distribution by flow cytometry; significantly, increased amounts of PANC-1-sh-SKA1 cells were found in the G2/M phase (P < .001), indicating that SKA1 depletion was potentially associated with G2/M arrest ( Figure 2C). To elucidate its molecular basis, G2/M arrest-associated proteins were investigated.
Results showed that knock-down of SKA1 lead to G2/M arrest by phosphorylating cdc25C (Ser216) and regulating the p21, cyclinB1 in PANC-1 cells, and vice versa in SW1990 cells ( Figure 2D). These findings suggested that SKA1 increases proliferation by promoting G2/M cell cycle progression.
Finally, to evaluate the in vivo effect of SKA1, we performed subcutaneous xenograft assays in nude mice, and SKA1 overexpression significantly increased tumour growth, along with a marginally increased expression of Ki67 ( Figure 2E,F). Likewise, similar results were obtained in PANC-1 cells ( Figure S2).

| Loss of SKA1 suppresses migration and invasion and confers resistance to EMT
It is universally acknowledged that EMT is one of the most important factors associated with three major steps (invasion, dissemination and metastasis) in pancreatic cancer. 23 Due to the fact that poorly differentiated cancer cells are more prone to early metastasis, and poorly differentiated pancreatic cancer tissues/cells showed higher SKA1 expression levels than well-differentiated counterparts (see above), whether SKA1 facilitates migration and invasion in PDAC cells is an interesting question.
We For in vivo PDAC metastatic models, tail vein metastasis assay demonstrated that SKA1 significantly enhanced the number of metastasis of lung in nude mice at 12 weeks ( Figure 3D). And there were fewer metastatic foci for SKA1-knock-down group. Immunofluorescence also revealed that SKA1 overexpression led to increased vimentin and N-cadherin signals but decreased E-cadherin immune reactivity ( Figure 4A). To understand the mechanism by which SKA1 reduces E-cadherin expression, we checked the expression of EMT regulatory transcription factors. As shown in Figure 4B, SKA1 overexpression significantly increased the level of Snail and Twist, whereas these protein levels were decreased in SKA1-knock-down cells, suggesting that SKA1 may facilitate EMT via Snail and Twist.
Taken together, these data implied that depletion of SKA1 likely suppressed the migration-invasion cascade of PDAC.

| Comparative proteome-wide analysis of SKA1induced biological processes
Since SKA1 exerted obvious effects for PDAC, we conjectured that more detailed and concrete pathways as well as potential effectors must underlie this phenomenon. To further unveil this mechanism, PANC-1 cells stably transfected sh-SKA1 and sh-ctr were generated as in vitro models to perform iTRAQ analysis. A total of 5351 proteins were obtained, with 868 identified as differentially expressed between groups with a cut-off fold change of 1.5 and P < .05. Of these, 448 proteins were upregulated and 420 were downregulated ( Figure 5A). Subcellular distribution of the differentially expressed proteins was annotated as Figure 5B.
Next, in order to obtain an overview of the functions of differentially expressed proteins, GO functional classification was conducted based on biological processes, cellular component and molecular function. Within the category of molecular function, the most significant group of GO annotated proteins was involved in actin binding activities, mainly including but not limited to calcium ion binding, cytoskeletal protein binding and cell adhesion molecule binding.
According to biological processes, functions in epithelial cell differentiation were significantly enriched ( Figure 5C). Moreover, enrichment analysis (P < .01) using the KEGG was performed to investigate the possible roles that were preferred targets of downstream pathways by SKA1 regulation. As shown in Figure 5D, the largest group of differential proteins was involved in regulation of actin cytoskeleton, followed by proteoglycans in cancer, which as a major component of the ECM can interact with numerous regulators of cell behaviour through signalling to the actin cytoskeleton and cell adhesion. 24,25 Additionally, major GO terms including the related annotated proteins were also presented as a PPI network ( Figure 5E), which high-

| SKA1 remodels the actin cytoskeleton via activating Cdc42
What are the specific molecular mechanisms leading to the occurrence of the above phenomenon? By protein functional prediction and differential protein expression folds analysis, we hypothesized that the biological processes induced by SKA1 in PDAC cells could be mediated through regulation of Cdc42 (with a cut-off fold change of 5.4 and P = 7.8629E-09), which could facilitate filopodia formation, regulate cell migration and regulate bipolar attachment of spindle microtubules to kinetochores in metaphase. [27][28][29] We first did the correlation analysis between SKA1 and Cdc42, as shown in Moreover, we used ZCL278, a compound that directly binds to Cdc42 and displays most inhibitory effects in a morphological assay of Cdc42 function, to assess its activity at the biochemical level by incubation at concentrations of 1, 5 and 20 μmol/L for 10 minutes, respectively. As depicted in Figure 7D,E, application of ZCL278 resulted in a dose-dependent decrease in Cdc42 activity, total Cdc42 level as well as N-WASP and Arp2/3 levels.
Meanwhile, ZCL278-treated cells displayed an obvious inhibition of microspike formation in a dose-dependent manner, leading to cell morphology changes such as decreased development of long pseudopodia-like protrusions; when administered at 20 μmol/L, the microfilament structure was completely destroyed and Golgi organization disruption was also seemingly observed ( Figure 7F).
We also examined the other main Rho-GTPase, and only a total of Rac1/2/3 increase accompanied by SKA1 overexpression was observed ( Figure 7G).
After cytoskeletal remodelling, cancer cells become more invasive and develop altered affinity to facilitate migration and invasion through basement membrane and ECM, and metastases are triggered subsequently. 25 Focal adhesions are subcellular structures that mediate the regulatory effects (ie signalling events) of a cell in response to ECM adhesion. Focal adhesion proteins also help hold actin filaments together to form actin stress fibres, which latter are specifically localized to filopodia. 30,31 We extracted cytoskeletal Collectively, we drew a simplified schematic diagram illustrating the facilitating effects of SKA1 in actin cytoskeleton organization and related pathways in PDAC cells ( Figure 7I).

| D ISCUSS I ON
Early metastasis is quite a challenge to be overcome and is therefore a therapeutic target for preventing the spread of many types of cancer like PDAC. 1,32 SKA1 was reportedly increased as a candidate oncogene in several common human cancers, contributing to various steps of oncogenesis and poorer prognosis. [14][15][16][17][18] Not only SKA1 genomic mutations F I G U R E 6 SKA1 modulates the formation of cytoskeletal actin filament and shared partial co-localization with F-actin in PDAC cells. A, Combined rhodamine-phalloidin staining for F-actin (red) and SKA1 (green) immunofluorescence, nuclei were counterstained with DAPI (blue). Upper two panels: stable knock-down of SKA1 reduced actin protrusions and stress fibres in PANC-1 cells. Lower two panels: stable overexpression of SKA1 altered the cellular morphology and lead to longer invadopodia formation compared with vector group in Capan-1 cells. Additionally, SKA1 and phalloidin-stained F-actin shared partial co-localization to the cytoplasm. Co-localization correlation of intensity distributions curves between two channels was drawn by ImageJ and measured by Pearson's correlation coefficient analysis (right and Capan-1 cells' infectants to verify the proteins that regulate actin cytoskeleton. C, Cdc42-GTP pull-down and Western blot analyses of the activation state of Cdc42 expression and total expression of Cdc42 in the indicated cells. D and E, The application of ZCL278 resulted in a dose-dependent decrease in Cdc42 activity, followed by inhibition of endogenous Arp2/3, N-WASP. F, Cells were fixed and stained with rhodamine-phalloidin to label filamentous actin following ZCL278 treatments. Results showed that ZCL278 inhibits microspike formation and stress fibres in PANC-1 and Capan-1 cells. White asterisks indicate the subcellular locations that normally show stress fibre distribution. White arrows point to the seemingly disruption Golgi organization. G, RhoA, RhoB, RhoC and Rac1/2/3 protein expression were detected, and only a total of Rac1/2/3 increase accompanied by SKA1 overexpression was observed. H, FAK, Talin and ɑ-actinin levels increased with SKA1 expression in Capan-1 cells, FAK, Paxillin and ɑ-actinin levels decreased with SKA1 knockout in PANC-1 cells. I, Proposed mechanistic scheme of SKA1 in promoting tumour progression in PDAC Cdc42, which binds to a variety of effector proteins to regulate actin cytoskeleton remodelling and epithelial cell polarization, was strongly involved in SKA1-induced PDAC progression, and inhibition of Cdc42 may represents a promising strategy for precise cancer therapy. [46][47][48] In this study, we used ZCL278 as a powerful tool to assess Cdc42 function. 49,50 We first provided evidence that treatment with ZCL278 of PDAC cells had similar but less powerful effects compared with cytochalasin B, largely by affecting the downstream Arp2/3 complex and N-WASP, which play critical roles in actin cytoskeleton organization and metastasis in cancer cells. 26,51,52 Moreover, although we preliminarily confirmed that cell adhesion-related proteins increased with SKA1 expression, it still remains to be investigated how SKA1 regulates focal adhesion during cell migration in-depth.
In summary, we comprehensively characterized SKA1, as a oncogene, could constitute a reasonable biomarker and prognostic factor of PDAC. Also, our findings substantiate that SKA1 plays significant roles in facilitating PDAC cells' proliferation and metastasis by inhibiting G2/M arrest and remodelling actin cytoskeleton via activating Cdc42, which endow the potential of related downstream inhibitors as interference therapeutic targets for PDAC.

ACK N OWLED G EM ENTS
This work was supported by National Natural Science Foundation of China (81870385, 81672719, 81702740, 81800491).

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are included in this published article and its supporting information files.