Aberrant ASPM expression mediated by transcriptional regulation of FoxM1 promotes the progression of gliomas

Abstract Gliomas are the most common form of malignant tumour in the central nervous system. However, the molecular mechanism of the tumorigenesis and progression of gliomas remains unclear. In this study, we used the GEO database to identify genes differentially expressed in gliomas and predict the prognosis of glioma. We observed that ASPM mRNA was increased obviously in glioma tissue, and higher ASPM mRNA expression predicted worse disease prognosis. ASPM was highly expressed in glioma cell lines U87‐MG and U251, and knockdown of ASPM expression in these cells significantly repressed the proliferation, migration and invasion ability and induced G0/G1 phase arrest. In addition, down‐regulation of ASPM suppressed the growth of glioma in nude mice. Five potential binding sites for transcription factor FoxM1 were predicted in the ASPM promoter. FoxM1 overexpression significantly increased the expression of ASPM and promoted the proliferation and migration of glioma cells, which was abolished by ASPM ablation. ChIP and dual‐luciferase reporter analysis confirmed that FoxM1 bound to the ASPM promoter at −236 to ‐230 bp and −1354 to ‐1348 bp and activated the transcription of ASPM directly. Collectively, our results demonstrated for the first time that aberrant ASPM expression mediated by transcriptional regulation of FoxM1 promotes the malignant properties of glioma cells.

in poor prognosis and high recurrence of glioma patients. Therefore, further clarification of the molecular mechanism of the pathogenesis and progression of gliomas is warranted for the development new therapeutic targets for the disease.
ASPM (abnormal spindle-like microcephaly) gene, also known as MCPH5, is the most common family microcephaly mutation gene involved in the regulation of neurogenesis and cerebral cortical size. 5,6 Intracellular ASPM is mainly distributed in centrosome and spindle microtubule intermediates, and ASPM deficiency leads to spindle assembly and mitotic process disruption. 7,8 In addition, ASPM is also implicated in tumorigenesis and tumour progression. [9][10][11] In gliomas, the expression of ASPM increased with the glioma grade, and ASPM expression is significantly higher in recurrent glioma than that in primary gliomas. 12 Depletion of ASPM suppresses the proliferation of glioma spheres and promotes cell death. 13 Yet, the underlying molecular mechanisms that mediate the up-regulation of ASPM expression in glioma tissues remain unknown.
FoxM1 is a member of the forkhead box (Fox) transcription factor family. 14 Previous evidence indicated that FoxM1 plays important roles in a variety of biological processes including mitotic G1/S and G2/M phase transition, mitotic spindle integrity, DNA damage repair, angiogenesis and tumour metastasis. 14,15,16 FoxM1 is highly expressed in a variety of solid tumours and contributes to malignant transformation and chemotherapy resistance, suggesting that FoxM1 may serve as a potential antitumour target. 17,18 FoxM1 is also highly expressed in glioblastoma (GBM), and higher FoxM1 expression is associated with worse overall survival in glioma patients. Mechanism studies have indicated that FoxM1 can promote glioma progression by enhancing MMP2 transcription 19 or temozolomide (TMZ) resistance by up-regulating RFC5 expression directly. 20 In an attempt to find potential transcript factors that may play a role in increased ASPM expression in gliomas, we predicted several potential FoxM1 binding sites in ASPM promoter using transcript factor binding site prediction databases. However, whether FoxM1 can directly regulate ASPM expression and whether the FoxM1-ASPM axis contributes to the pathogenesis and progression of gliomas remain largely unclear.
In this study, we aimed to identify genes related to the pathogen-

| Patient samples
All glioma and normal brain control tissue samples were obtained from patients undergoing neurosurgical resection in Xiangya Hospital of Central South University (Changsha, Hunan, China) between June 2015 and December 2017. All samples were confirmed by post-operative pathology and classified according to the WHO Classification of the Central Nervous System Tumors. All procedures were approved by the Ethics Committee of Xiangya Hospital of Central South University, and informed consent was obtained from patients.
gov/geo/. 100 glioma samples from GSE4271, 157 gliomas samples and 23 non-tumour samples from GSE4290 and 21 glioma samples from GSE45921 were used to screen genes differentially expressed in non-tumour brain tissue and glioma tissue. Genes with |log fold change| ≥ 1 and adjusted P < .05 were considered as differentially expressed genes (DEGs). 21

473
LGG patients and 539 patients GBM with complete clinical information were included in the TCGA-LGG data set and TCGA-GBM data set, respectively.

| Immunohistochemical staining
Paraffin-embedded glioma tissue sections were dewaxed with xylene and then hydrated with gradient ethanol. After antigen repair and blocking, the tissue sections were incubated with ASPM (Santa Cruz, CA, USA)-and FoxM1 (GeneTex, CA, USA)-specific primary antibodies in a wet box at 4°C overnight. Negative control sections were in parallel incubated with non-specific IgG (Sigma, MO, USA).
Then, tissue sections were incubated with a secondary antibody for 2 hours at room temperature. After staining with DAB and haematoxylin, tissue sections were observed and photographed under a microscope.

| Western blot analysis
The treated cells were collected and lysed with RIPA lysis buffer TBST, the membranes were incubated with appropriate secondary antibodies. The membranes were then exposed to enhanced Chemiluminescence-Plus reagents (Beyotime, Shanghai, China).
ChemiDoc™ XRS + chemiluminescence imaging system was used to capture the images. GAPDH protein was used as the internal reference to calculate the relative expression level of the protein by Image Lab 3.0 software.

| Cell proliferation detection
Cell viability and proliferation were measured by CCK8 assay according to the manufacturer's instructions. In brief, cells were seeded in 96-well plates at a density of 1 × 10 3 cells/well. At the indicated time-point, 90 μL culture medium and 10 μL CCK8 were added into each well. After gently mixing, cells were incubated at 37°C for 1 hour. Subsequently, the absorbance at 450 nm was detected by a microplate reader (Bio-Rad, CA, USA), and the proliferation rate of glioma cells was calculated. Clonal formation assay was used to determine the proliferation ability of glioma cells, and the procedures were carried out according to a previous study. 22

| Wound healing assay
Wound healing assay was performed according to the previous study. 22 In brief, cells were seeded in 6-well plates at a density of 5 × 10 5 cells/wells and cultured for 24 hours. After transfection for 24 hours, wounds were gently made using a micropipette tip, and the cells were washed with sterile PBS solution to remove floating cells.
Then, the cells were cultured with serum-free medium for 24 hours.
Cells were photographed at 0 hour and 24 hours after scratch under an inverted microscope.

| Cell cycle analysis
The procedure was according to a previous study, 23,24 and the distribution of cell cycle was determined by flow cytometry (Beckman Coulter, CA, USA). Subsequently, the firefly luciferase activity and Renilla luciferase activity were determined with the dual-luciferase assay system (Promega, WI, USA) according to the manufacturer's directions.

| Chromatin immunoprecipitation (ChIP) PCR assay
ChIP experiments were performed with the commercially available Pierce™ Agarose ChIP Kit (No. 26156) (Thermo Fisher, CA, USA) according to the manufacturer's instructions. Real-time PCR primers for the ASPM promoter region are shown in Table S3.

| Subcutaneous tumorigenesis experiment of nude mice
Female BALB/c nude mice aged 6 weeks and weighing about 18 g were selected for this experiment. Sixteen mice were randomly divided into sh-con group and sh-ASPM group. ASPM stable knockout U87-MG cells (U87-MG-sh-ASPM cells) and knockdown control U87-MG cells (U87-MG-sh-con cells) were collected and resuspended in pre-cooled DMEM containing 10% Matrigel and 2%

| Statistical analysis
All of the statistical analyses were carried out using SPSS 19.0 software (SPSS Inc, IL, USA) and GraphPad Prism 5.0 (GraphPad Inc, CA, USA). The data are presented as mean ± SD. Student's t test was used to analyse the significance of the differences between groups, and one-way ANOVA was used to analyse significance of the differences among groups. Kaplan-Meier survival analysis was used to determine the survival profiles, and log-rank test was carried out to assess the statistical significance of differences. P < .05 was considered statistically significant.

| ASPM is related to glioma risk and prognosis
Three gene expression profiles (GSE4290, GSE4271 and GSE45921) were used to identify genes related to the pathogenesis and progression of gliomas. As shown in Figure 1A,B and Table 1, a total of 59 differentially expressed genes (DEGs) were found between normal brain tissues and glioma tissues. GO analysis and KEGG pathway enrichment analysis were performed to explore potential biological functions of DEGs (Table S1). GO analysis showed that DEGs were significantly enriched in cell components such as centrosome, nucleocentric centromere and spindle tubulin and were involved in the biological processes such as cerebral cortex development, mitosis and DNA repair. KEGG pathway enrichment analysis showed that DEGs were mainly enriched in cell cycle, P53 signalling pathway, progesterone-mediated oocyte maturation and ECM-receptor interaction. Then, we selected twelve genes (ASPM, CCNB2, CENPA, CENPF, COLA2, DLGAP5, GINS1, HSPG2, KLHDC8A, KIF14, LAMC1 and RRM2) that were not well studied at present, and verified the expression of these candidate genes by real-time PCR in 30 clinical samples. It was found that ASPM, CCNB2, HSPG2, KLHDC8A and RRM2 mRNA were differentially expressed in glioma tissues of different grades.
And among them, ASPM changes most significantly ( Figure S1).
More importantly, further validation in the CGGA and TCGA-LGG data sets and clinical samples also demonstrated that the expression of ASPM mRNA was increased with the increase in glioma grades, and the high expression of ASPM indicated a worse prognosis ( Figure 1C,D). These results suggested that ASPM may act as a molecular marker for the diagnosis and prognosis prediction of glioma patients.

| ASPM inhibition affects proliferation, migration and invasion of glioma cells
In order to investigate the role of ASPM in glioma, we primarily

| ASPM inhibition induces cell cycle arrest in the mitosis process of glioma cells
Studies have shown that ASPM participates in the regulation of symmetrical mitosis and the maintenance of neuronal granulosa cells by affecting the cell cycle. 6,26 However, whether ASPM is in-

| Altering FoxM1 expression affects ASPM expression and proliferation, migration and invasion of glioma cells
Next, we further determined whether FoxM1 has a regulatory effect on ASPM expression. We examined the mRNA and protein expression levels of FoxM1 in different glioma cell lines and found that FoxM1 was highly expressed in U87-MG cells and less expressed in Hs683 cells ( Figure 5A). As shown in Figure 5B

| FoxM1 transcriptionally regulates ASPM expression by directly binding to the promoter region
To ascertain whether FoxM1 can directly bind to ASPM promoter region and thus regulate ASPM expression, we first used bioinformatics to predict the potential transcription factor binding sites of ASPM gene promoter and found that ASPM promoter region contains 5 potential FoxM1 binding sites ( Figure 6A), and the binding motif of FoxM1 is TGCAAA ( Figure 6B). Then, ChIP experiment was used to detect the direct binding of FoxM1 to ASPM promoter. As shown in

| ASPM is essential for the growth of subcutaneous tumours in nude mice
A xenograft nude mouse model was established to evaluate the effect of ASPM on the growth of gliomas in vivo. Firstly, we used lentiviral infection to establish ASPM stable knockdown U87-MG cells. Figure 7A, the fluorescence-positive rate of cells was >90% after lentiviral infection. In agreement, compared with U87-MG-sh-nc cells, ASPM expression was significantly lower in U87sh-ASPM cells ( Figure 7B,C). In the subcutaneous tumorigenesis

| D ISCUSS I ON
In this study, GSE4290, GSE4271 and GSE45921 were used to screen out 59 differentially expressed genes in normal brain tissues and gliomas of different grades. Through verification of the CGGA data set, TCGA-LGG data set and clinical samples, we found that the expression of ASPM mRNA increased with the increase of glioma grade, and the high expression of ASPM predicted worse prognosis, suggesting that ASPM may function as a molecular marker for the diagnosis and prognosis prediction of glioma patients.
ASPM plays an important role in controlling the neurogenesis and cerebral cortical size, and its homozygous mutation can lead to apoptosis of neural progenitor cells and microcephaly. 5,27 Previous studies have reported that ASPM is mainly distributed in the centrosomes, spindle microtubule intermediates and midbody 28,29 and plays a vital role in cell division and proliferation in foetal tissues and human cancer cells. 30,31 In the development of mouse brain, ASPM maintains the symmetrical division and proliferation of embryonic neuroepithelial cells by controlling mitotic spindle orientation. 6 In gliomas, ASPM is elevated in glioma tissues, and ASPM knockdown significantly represses the proliferation of glioma spheres. 12,13 Consistent with previous studies, our results also indicated that ASPM is significantly positively correlated with the pathological grade of glioma and poor prognosis of patients, and knockdown of ASPM could inhibit the proliferation of glioma

cells.
Previous studies have found that high expression of ASPM is significantly positively correlated with vascular invasion and early metastasis of hepatocellular carcinoma, and its high expression indicates poor prognosis of hepatocellular carcinoma patients. 32 In metastatic melanoma tissues, ASPM gene is highly expressed, and ASPM overexpression improves the invasion ability of melanoma cells, 33 indicating that ASPM may mediate the invasion and metastasis of tumours. However, the effect of ASPM on the migration and invasion ability of glioma cells has not been reported. In the present study, we found that knockdown of ASPM significantly reduced the migration ability of glioma cells. EMT plays an important role in the invasion and migration of gliomas, 34 and MMPs can promote the migration, invasion and distant metastasis of tumour cells to surrounding tissues by degrading the extracellular matrix. 35 In this study, we found for the first time that knockdown In order to further investigate the mechanism that induced ASPM up-regulation in gliomas, we utilized the JASPAR and ENCODE database prediction combined with clinical samples and the CGGA and TCGA database validation, and first discovered FoxM1 may be a potential control ASPM expression transcription factors.
FoxM1, a proliferation-specific transcription factor, was first discovered in the cervical cancer cell line HeLa. 14 FoxM1 is highly expressed in embryonic mouse tissues with a high proliferation index, such as thymus, small intestine and testicles, and it is also elevated in a variety of tumour tissues and is associated with tumour malignancy and patient prognosis. 17,18,40,41 It has been reported that FoxM1 is overexpressed in GBM and suggests poor prognosis in glioma patients. 19,42 Similar to previous studies, we also found that FoxM1 is highly expressed in gliomas and its high expression predicts the poor prognosis of glioma patients. In addition, studies have demonstrated that FoxM1 is involved in maintaining tumorigenicity of GBM stem cells, and its mechanism may be related to FoxM1 promoting nuclear localization of beta-catenin. 43 However, whether FoxM1 plays a role in tumour development by transcriptional regulation of ASPM expression is still unknown.
In this study, we demonstrated for the first time that FoxM1 can directly regulate ASPM transcription in glioma cells, thereby affecting their expression and thereby regulating the proliferation, migration and invasion of glioma cells.
In summary, the present study is the first to reveal that ASPM promoted glioma cell proliferation, migration and invasion and tumour growth, which is mediated by the transcriptional activation of FoxM1. In the present study, using the TCGA and CGGA data set analysis, we found that ASPM was highly correlated with FoxM1 expression. Due to the presence of tumour heterogeneity, FoxM1 was unfortunately present only in the DEG list of GSE4290 and GSE45921, and therefore not in the list of 59 DEGs we finally found. This phenomenon suggested that using multiple data sets to screen for DEGs may yield more reliable results but may also lose valuable information. It has been confirmed that FoxM1 is involved in chemotherapy resistance in a variety of tumours. 20,44,45,46 Moreover, studies have shown that FoxM1 can F I G U R E 5 Altering FoxM1 expression affects ASPM expression, proliferation and migration of U251 cells. A, The expression of FoxM1 mRNA and protein in different glioma cell lines; *P < 0.05 vs HEB, **P < 0.01 vs HEB, ***P < 0.001 vs HEB. B, C, The U87-MG cells were transfected with a control siRNA or FoxM1 siRNA, and the mRNA and protein expression of FoxM1 and ASPM were measured by real-time PCR and Western blot; **P < 0.01 vs si-nc group, ***P < 0.001 vs si-nc group. The U251 cells were transfected with FoxM1 overexpression vector and a control siRNA or ASPM siRNA. D, Protein expression of FoxM1 and ASPM was measured by Western blot. E, F, CCK8 assay and clonal formation assay was used to measure cell proliferation; ***P < 0.001 vs con group, ### P < 0.001 vs FoxM1 + si-nc group. G, Wound healing and transwell migration assays were used to detect the migration ability of U251 cells. H, The protein expression of E-cadherin, N-cadherin, Vimentin, MMP2 and MMP9 in U251 cells was measured by Western blot F I G U R E 6 FoxM1 transcriptionally regulates ASPM expression by directly binding to the promoter region. A, Bioinformatics predicted the binding region between FoxM1 and ASPM promoter. B, The binding motif of FoxM1. C, ChIP assay was used to detect the binding of FoxM1 to ASPM promoter region in U87-MG cells; *P < 0.05 vs IgG group, **P < 0.01 vs IgG group, ***P < 0.001 vs IgG group. D, The firefly luciferase reporter gene vectors with different lengths of ASPM promoter. E, Hs683 cells were transfected with ASPM promoter luciferase reporter gene vectors and FoxM1 overexpression vector for 24 h, and then, the dual-luciferase activity was detected; **P < 0.01 vs con group, ***P < 0.001 vs con group. F, U87-MG cells were transfected with ASPM promoter luciferase reporter gene vectors and FoxM1 siRNA for 24 h, and then, the dual-luciferase activity was detected; *P < 0.05 vs si-nc group, **P < 0.01 vs si-nc group F I G U R E 7 ASPM is essential for the growth of subcutaneous tumours in nude mice. A, The ASPM stable knockout U87-MG cells (U87-MG-sh-ASPM cells) and the control U87-MG cells (U87-MG-sh-nc cells) were imaged using fluorescence microscopy. B, C, The mRNA and protein expression of ASPM were measured by real-time PCR and Western blot; ***P < 0.001 vs U87-MG-sh-nc cells. D, Subcutaneous tumour was stripped and photographed after transplantation for 18 days. E, Subcutaneous tumour growth curve of nude mice. F, Subcutaneous tumour weight in nude mice. G, The histopathological characteristics and the expression levels of ASPM and Ki-67 protein in subcutaneous tumours of nude mice were detected by HE staining and IHC, respectively directly promote the expression of RFC5 at the transcription level, thus leading to TMZ resistance independent of MGMT activity. 20 Knockdown of FoxM1 can inhibit the expression of Rad51 and increase the TMZ sensitivity of recurrent GBM cells. 46 These studies suggest that FoxM1 is involved in the regulation of TMZ resistance and drug resistance in glioma. In the future, more experiments are needed to explore the role of the FoxM1-ASPM axis in TMZ chemotherapy resistance.

ACK N OWLED G EM ENTS
This work was supported by the National Natural Science

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings described in the current study are available in the article.