SKA1 promotes malignant phenotype and progression of glioma via multiple signaling pathways

Data used in this study for bioinformatic analysis was obtained from public datasets, including GEO Datasets (https://​www.​ncbi.​nlm.​nih.​gov/​gds/​), TCGA (https://​cancergenome.​nih.​gov/​) and CGGA (http://​www.​cgga.​org.​cn/​). Gene set enrichment analysis (GSEA) was further performed to explore the potential roles of SKA1.

The human glioma cell lines U251, U87, LN229 and T98 were purchased from the Chinese Academy of Sciences (Shanghai, China). In the laboratory, all cell lines were grown in Dulbecco’s modified Eagle’s medium (DMEM) (Hyclone, Logan, UT) supplemented with 10% fetal bovine serum (FBS, Hyclone, Logan, UT) and incubated in a humidified atmosphere of 5% CO₂ at 37 °C.

Tumor tissues were collected from patients with pathologically and clinically confirmed glioma. All samples were confirmed by pathological diagnosis and classified according to the World Health Organization (WHO) criteria. Moreover, prior written informed consents were obtained from patients or their guardians, and approval from the Ethics Committees of Nanfang Hospital of Guangdong Province were also obtained.

The preparation of lentivirus expressing human SKA short hairpin RNA (sh#1, CCGGTGAAGAACCTGAACCCGTAAACTCGAGTTTACGGGTTCAGGTTCTTCATTTTTTG; sh#2, CCGGATAGAGTATAGAGGCTATTTCCTCGAGGAAATAGCCTCTATACTCTATTTTTTTG; sh#3, CCGGCCTGACACAAAGCTCCTAAATCTCGAGATTTAGGAGCTTTGTGTCAGGTTTTTG) were performed using the pLVTHM-GFP lentiviral RNAi expression system (Genechem, China). U87, U251, LN229 and T98 cells were then transfected with lentiviral particles containing specific or negative-control (shNC) vectors according to the manufacturer’s instructions, respectively.

The cells or tumor tissues were washed three times with PBS and lysed in RIPA Buffer (50 mM Tris–HCl pH 8.0, 1 mM EDTA pH 8.0, 5 mM DTT, 2% SDS) with protease inhibitor and phosphoric-acid protease inhibitor at 4 °C for 30 min. Then they were crushed with ultrasonic machine and centrifuged at 12000 rpm for 15 min at 4 °C. The protein concentration was measured using BCA assay (Beyotime Inc, China). 12.5 μl mixed solution including 30 mg protein and 2.5 μl SDS was resolved using a 10% SDS-PAGE gel and electro-transferred to polyvinylidene fluoride membranes (Invitrogen, Carlsbad, CA). Afterwards, the membranes were blocked with 5% BSA or nonfat milk in pH 7.0 TBST, and then were incubated with primary antibodies overnight at 4 °C. On the next day, membranes were washed three times with TBST and incubated with horseradish peroxidase conjugated secondary antibody for 1 h at room temperature. At last, membranes were washed three times with TBST again. Signals were detected using enhanced chemiluminescence reagents (Pierce, Rockford, IL, USA). All experiments were independently performed in triplicate.

The CCK8 assay was performed to examine cell viability. Glioma cells (1000/well) were seeded in 96-well plates with a volume of 200 μl medium. Subsequently, 10 μl of CCK8 reagent mixed with 100 μl DMEM medium without FBS were added to each well and incubated for 1 h. Then the absorbance value (OD) was measured at 450 nm. The observation duration lasted for one week at the same time every day.

The indicated cells were plated in 12-well plates (200 cells per well) and cultured for 2 weeks. The colonies were then fixed with methanol for 30 min and stained with 1% crystal violet for 1 min. All assays were independently performed in triplicate.

The proliferation of U87 and U251 cells were examined using the Cell-Light EdU Apollo488 In Vitro Imaging Kit (RiboBio, China) according to the manufacturer’s protocol. In brief, cells were incubated with 10 μM EdU for 2 h before fixation with 4% paraformaldehyde and permeabilization with 0.5% Triton X-100, and then stained with EdU kit. Cell nucleus were stained with 5 μg/ml DAPI (4′,6-diamidino-2 phenylindole) for 10 min. The number of EdU-positive cells was counted under a microscope in five random fields (200×). All assays were independently performed in triplicate.

The cell migration assays were carried out with Transwell assays. About 5 × 10⁴ cells in 100 μl DMEM medium without FBS were seeded on a fibronectin-coated polycarbonate membrane inserted in a Transwell apparatus (Costar, MA). In the lower chamber, 500 μl DMEM with 10% FBS was added as a chemoattractant. After the cells were incubated for an appropriate time according to specific cell lines in a 5% CO₂ atmosphere at 37 °C, the insert was washed with PBS, and cells on the top surface of the insert were removed with a cotton swab. Cells adhering to the lower surface were fixed with methanol for 30 min, stained with 1% crystal violet solution for 1 min and counted under a microscope in three random fields. The cell invasion assays were carried out with Boyden assays, and the procedure was similar to the cell migration assay, except for that polycarbonate membranes were precoated with 24 mg/ml Matrigel (R&D Systems, USA). All assays were independently performed in triplicate.

U87 and U251 cells were stably transfected with shSKA1 or empty vectors respectively, and then cultured in 6-well plates. After the cells grew to 90% confluence, three paralleled scratch wounds across each well were made with a P-10 pipette tip. Fresh medium supplemented with reduced FBS (5%) was added, and the wound-closing procedure was observed for 48 h. Photographs were taken at 0, 12 h, 24 h and 48 h, respectively. All assays were independently performed in triplicate.

About 1 × 10⁶ active U87 cells that transfected with shRNA or empty vectors and suspended in 0.1 ml DMEM medium were subcutaneously injected into a group of ten nude mice, respectively. These mice were grown in a barrier facility on HEPA-filtered racks. The animals were fed with an autoclaved laboratory rodent diet and water. All animal studies were conducted in accordance with the principles and procedures outlined in the National Institutes of Health Guide for the Care and Use of Animals under assurance number A3873-1. On day 30, the mice were sacrificed, and tumor tissues were excised and weighed, respectively.

Paraffin sections at 4 μm were deparaffinized in 100% xylene and rehydrated in descending ethanol series according to standard protocol. Heat-induced antigen retrieval was performed in citrate buffer for 15 min in microwave oven. Endogenous peroxidase activity and non-specific antigens were blocked with peroxidase blocking reagent containing 3% hydrogen peroxide and serum. The sections were then incubated with primary antibodies, including N-cadherin (1:100; CST, USA), E-cadherin (1:100; Proteintech, USA), MMP9 (1:100; Proteintech, USA), and PCNA (1:100; Abcam, USA), at 4 °C overnight. The next day, the sections were washed three times with PBS. The sections were followed by incubation with biotin-labeled secondary antibody for 1 h at room temperature and washed three times with PBS. Subsequently, Sections were visualized with DAB and counterstained with hematoxylin, mounted in neutral gum, and analyzed with a bright field microscope equipped with a digital camera (Nikon, Japan). The results were scored as previously mentioned: 0, no staining; 1, weak staining in < 50% cells; 2, weak staining in ≥ 50% cells; 3, strong staining in < 50% cells; and 4, strong staining in ≥ 50% cells [18]. All experiments were independently performed in triplicate.

All quantified data is presented as an average of triplicate technical replicates. SPSS 13.0 and Graph Pad Prism 4.0 software were used for statistical analysis. Data are represented as mean ± SD. One-way ANOVA or two-tailed Student’s t-test was used for comparisons between groups. Chi square test or Fischer’s were used to identify differences between categorical variables. Survival analysis was performed using Kaplan–Meier method. Multivariate Cox proportional hazards method was used for analyzing the relationship between the variables and patient’s survival status. Differences were considered statistically significant when P < 0.05.

First, SKA1 expression was determined in public data deposited in TCGA database and GEO dataset (GSE 4290). We found that SKA1 expression was markedly increased in high grade glioma (Grade IV and III), but there was no significant difference between low grade glioma (Grade II) and non-tumor brain tissues (Fig. 1a). Moreover, analysis of TCGA dataset showed significant increase of SKA1 expression in grade IV glioma compared with those in grade III glioma and grade II glioma (Fig. 1b). In brief, analysis of public dataset revealed that SKA1 expression was positively correlated with glioma grade in mRNA level. To further confirm these results, we examined the expression of SKA1 at protein level with tissue samples collected from Nanfang hospital, including glioma tissues (grade II, n = 5; grade III, n = 8; grade IV, n = 34) and non-tumor brain tissues (n = 5) with Western blot (Fig. 1c, d). Consistent with the results described above, significantly higher expression level of SKA1 was detected in Grade IV glioma.

Considering that SKA1 was overexpressed in grade IV glioma, we used Chinese Glioma Genome Atlas (CGGA) dataset to determine whether SKA1 could be used as a biomarker to distinguish between GBM and non-GBM patients (Grade II and III). The area under the receiver operating characteristic (ROC) curve of SKA1 for differential diagnosis was 0.774 (95% CI 0.716–0.832), indicating that SKA1 could serve as an effective diagnosis marker to distinguish glioblastoma patients from non-GBM patients (Grade II and III) (Fig. 1e).

In TCGA database, we observed that higher SKA1 expression was associated with worse overall survival (OS) and progression free survival (PFS) (Fig. 1f, g). The median OS in patients with higher SKA1 expression was 32.90 months compared with 95.83 months in those with lower expression (P < 0.0001). The median PFS of glioma patients with higher and lower expression of SKA1 was 10.27 months and 38.47 months, respectively (P < 0.0001). Consistently, SKA1 overexpression was also confirmed to be associated with worse OS in CGGA database (Fig. 1h).

To assess the function of SKA1 in glioma, three different lentiviral shRNA targeting SKA1 were used to specifically and stably knock down the SKA1 expression in four glioma cell lines including U87, U251, LN229 and T98. Among these three lentiviral particles, the most efficient shRNA vector, sh-SKA1-3, was confirmed with Western blot analysis and selected for further experiments (Fig. 2a).

CCK8 assays were subsequently performed to evaluate the effect of SKA1 on cell viability. After knockdown of SKA1, both U87 and U251 showed a slower rate of proliferation compared with the control group (Fig. 2b). The EdU incorporation assay revealed that the percentage of cells in S phase decreased after SKA1 knockdown in U87 and U251 cells (Fig. 2c). The results of colony forming assay performed in U87, U251, LN229 and T98 glioma cells further confirmed that suppression of SKA1 expression attenuated cell viability and proliferation of glioma cells in vitro (Fig. 2d).

To validate this result in vivo, subcutaneous xenograft tumor model was established in nude mice, which were divided into NC group and shSKA1 group with 10 mice per group. Mice were sacrificed at 30 days after tumor inoculation, and the average tumor weight was 0.925 g and 0.360 g, respectively (Fig. 3a, P < 0.0001). Furthermore, immunochemistry staining for the proliferation marker, PCNA, indicated that suppression of SKA1 significantly inhibited glioma proliferation in vivo (Fig. 3b, c).

To examine the effect of SKA1 on glioma cell migration and invasion, Transwell assay, wound healing assay and Boyden assay were then performed. In Transwell assay, knockdown of SKA1 significantly decreased the percentage of migrated cells in the shSKA1 group (Fig. 4a), consistent with that found in wound healing assay (Fig. 4b). Additionally, Boyden assay revealed that SKA1 knockdown significantly suppressed the invasion of glioma cells (Fig. 4c). These results demonstrated that inhibition of SKA1 could significantly suppressed both migration and invasion of glioma cells in vitro. In xenograft models, we confirmed that knockdown of SKA1 could led to decreased N-cadherin and MMP9 and increased E-cadherin in xenograft tumor tissue (Fig. 3b, c).

To provide further insights into the mechanisms of SKA1 regulation, we used GSEA to investigate the possible biological functional of SKA1 in glioma with public dataset, including TCGA database and GEO dataset. GSEA results showed that SKA1 might be involved in cell cycle phase transition in glioma (Fig. 5a). Consistently, significantly decreased expressions of several cell cycle regulators and markers, such as FOXM1, CCNI, CCND1 and CDK1, were observed after SKA1 knockdown (Fig. 5b).

Considering that IHC staining of xenograft tumors showed decreased N-cadherin and MMP9 under SKA1 inhibition, we then examined the effect of SKA1 knockdown on the expression of epithelial–mesenchymal transition (EMT) markers in vitro. Results showed that expression level of E-cadherin and Claudin was increased, and that of N-cadherin, PLOD2, MMP2 and Snail was significantly decreased after SKA1 knockdown in glioma cells (Fig. 5c).

Additionally, GSEA results also indicated that SKA1 might be involved in Wnt/β-catenin signaling pathway. GSEA analysis of both RNA-seq and microarray dataset indicated that genes involved in Wnt/β-catenin signaling were enriched in glioma tissues with high expression of SKA1 (Fig. 5d). Consistently, with Western blot, significantly decreased expression level of β-catenin, and canonical Wnt/β-catenin signaling targets including CD44, c-Jun and c-Myc were observed under SKA1 inhibition (Fig. 5d).