Background
The sex-determining region Y (SRY) box (SOX) gene family, characterized by the highly conserved HMG-domain responsible for sequence specific DNA binding, encodes transcription factors that are essential for embryonic development, cell fate determination, differentiation, and proliferation [
1]. So far, twenty pairs of SOX genes have been identified in the human and mouse genomes [
2].
SOX4 has been found to be over-expressed in adenoid cystic carcinoma (ACC), hepatocellular carcinoma, bladder tumors, acute myeloblastic leukemia, prostate cancer, endometrial cancer and glioblastoma [
3]-[
8]. SOX4 was further identified as a common transcription factors for neoplastic transformation and progression in a large-scale meta-analysis of cancer microarray data [
9]. However, SOX4’s mode of action in cancer is complicated as SOX4 can act either as an oncogene [
4],[
10],[
11] or a tumor suppressor [
4],[
12]. As an oncogene, SOX4 overexpression predicts poor outcome of colorectal cancer [
13]. Its overexpression in prostate cancer correlated strongly with Gleason score [
6]. Knock down of SOX4 induced apoptosis in prostate cancer cells [
6] and adenoid cystic carcinoma ACC3 cells [
3]. SOX4’s role in bladder is perplexing: SOX4 is over expressed in bladder cancer tissues compared to normal tissues, but strong SOX4 expression was found to be correlated with increased patient survival (P <0.05) of bladder cancer [
4], and when introduced to bladder cancer cell line HU609, it reduced cell viability by promoting apoptosis and necrosis [
4]. As a tumor suppressor, introduction of SOX4 into hepatocarcinoma Hep3B and HepG2 cells induced apoptosis via the caspase cascade with caspase-1 activation [
14]. In HeLa cells, SOX4 was shown to also induce apoptosis via the caspase dependent pathway [
15].
Glioblastoma multiforme is the most common and aggressive type of malignant gliomas (WHO grade IV) with an annual incidence of 2 to 3 per 100,000 population [
16]. Currently, the standard therapy for gliomas consists of maximal surgical resection, followed by chemotherapy [
16]. However, because of its malignant features manifested by fast growth and chemo- or radio-resistance, most of patients die from the recurrence with malignant gliomas within one year [
17]. Others and we have showed that SOX4 is a target of TGF-beta signaling and is involved in maintaining stemness of glioma-initiating cells [
4],[
8],[
18],[
19]. To further understand the molecular mechanism of SOX4 in GBM, in this study, we systematically studied the function of SOX4 in GBM cells using the
Cre/lox P system to generate gain or loss of SOX4 in GBM cells. We showed that SOX4 inhibited the growth of GBM cells. A gene expression profiling analysis showed that SOX4 influenced several key pathways including the Wnt/ beta-catenin and TGF-beta signaling pathways. Finally, we showed the activation of p53-p21 signaling and down-regulation of phosphorylated AKT1 by SOX4. These data provide new molecular insights into how SOX4 exerts its functions in glioma cells.
Methods
Survival analysis
Z-Scores of mRNA of SOX4 from all three platforms (U133 microarray, Agilent and RNA Seq V2 RSEM) for Glioblastoma Multiforme (TCGA, Provisional) dataset were downloaded using cBioPortal [
20],[
21]. Univariable survival analysis was performed by the Kaplan-Meier method and log-rank test with ‘survival’ R package version 2.37-7 [
22].
Cell lines and cell culture
Human glioma cell lines LN229, T98G, U87MG, U251MG, A172, M059J and M059K were obtained from the American Type Culture Collection. All cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin in a humidified incubator with 37°C and 5% CO2.
Plasmid construction, retrovirus infection and transfection
The SOX4 gene from a vector containing SOX4-eYFP (a gift from Carlos S. Moreno, Emory University) was cloned by PCR and inserted into a retrovirus plasmid pBrit-HA/Flag (Plasmid 17519, Addgene, Cambridge MA) between the two lox P sites using restriction enzyme BamHI and EcoRI. A 5’ primer containing a Flag tag was used and the final construct was designed to express the Flag-SOX4-HA fusion protein, forward primer (5’-cgcGGATCCgcgATG GATTACAAGGATGACGACGATAAGATGgtgcagcaaaccaaca-3’), reverse primer (5’-ccggaattcGTAGGTGAAAACCAG-3’).
Retrovirus was packaged by co-transfect pBrit-Flag-SOX4-HA/pBrit-Flag-HA together and pcl-Ampho plasmids into 293T cells with the protocol previously described [
23], SOX4 overexpression cells (named as LN229_pSFH or A172_pSFH or U87_pSFH respectively for the cell line used) and control cells (named as LN229_con or A172_con or U87_con respectively for the cell line used) were cultured by infecting the respective GBM cell lines with the virus. After 96hours, cells were selected in culture media contains 800ng/ml puromycin (Sigma) for 3 weeks. Survival cells were harvested for functional study and analysis.
For transient transfection, SOX4-HA was cloned to the pCMV-Tag2C vector to generate the Flag-SOX4-HA plasmid.
Cre recombinase cDNA from the pACN vector [
24](a gift from Dr. Mario Capecchi, University of Utah) was cloned into the pEGFP-N1 to construct the
Cre -GFP plasmid. Flag-HA-GFP plasmid was used as the control. 48 hours after transfection, cells were harvested for further analysis and cell functional test.
Cell proliferation assay
Cell growth curves were obtained by measuring cell viability with the Cell Counting Kit-8 (CCK-8, Dojindo, Japan) according to the manufacturer’s instruction. 103 cells/well were seeded into 96-well plate on day zero and then cultured for five days. Cell viability was assessed at 6 h (hours), 24 h, 72 h, and 120 h, and normalized to the cell viability measurement at 6 h for each group. Each group was measured in triplicate.
Cell cycle analysis
For cell cycle analysis, 5 × 105 cells from each sample were trypsinized to make single cell suspension. The cells were then washed with PBS and fixed in 95% ethanol. The cells were treated with RNAse A and stained with propidium iodide (PI) and subjected FACS analysis.
Cells were plated into 6-well plates at 400 cells/well in DMEM culture medium and further incubated for 2 weeks when colonies were sufficiently large for visualization. Colonies were then fixed in methanol and stained with 0.5% crystal violet (Sigma). The stained cells were quantified. Each group was measured in triplicate.
The soft agar assay was performed in 12-well plates containing two layers of Sea Plague Agar (Invitrogen). The bottom layer consisted of 0.8% agar in 1 ml of DMEM with 10% FBS and the top layer consisted 103 cells of stable overexpression cells or control cells. Cells were seeded in each well and cultured for three weeks. Colonies were photographed under a microscope and counted.
Reverse transcription and real-time PCR
RNAs from cell lines were extracted using TRIZOL (Invitrogen) and 1μg of total RNAs was reverse transcribed into cDNAs using ReverTra Ace qPCR RT Kit (Toyobo, Japan). For quantitative real-time PCR, 20ng of cDNAs was used as template using the SYBR® Premix Ex Taq™ II (Taraka, Japan). GAPDH was used as control. Primers for target genes were listed in (Additional file
1: Table S1). Relative expression was calculated as described [
25].
Immunoblotting and analysis
Cells were washed twice with PBS and lysed in RIPA buffer (25mM Tris HCl, pH 7.6, 150mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS, 1mM PMSF [P7626, Sigma] with the Protease Inhibitor Cocktail [Roche]). Nuclear and cytosolic fractions were extracted using a nuclear protein extraction kit (Beyotime, Jiangsu, China) according to the manufacturer's instructions. Protein concentration was determined using the BCA protein kit (23225, Thermo, USA). Total cell lysate (30 μg) and nuclear or cytosolic fractions (15 μg) were loaded on SDS-PAGE and electrophoretically transferred to polyvinylidene difluoride membrane (Milipore, Bedford, MA). The membranes were blocked with 5% milk protein in TBST (0.6% Tween, 10 mM Tris pH7.6 and 150 mM NaCl in H2O) and incubated with primary antibodies at 4°C overnight followed by incubation with secondary antibody for 1h at room temperature before exposure. Primary antibodies used were: anti-HA antibody (1:10000, ab9110, Abcam), anti-beta-catenin antibody(1:2000, ab32572, Abcam), anti-p53 antibody (1:1000, ab32389, Abcam), anti-AKT1 antibody(1:500, ab81283, Abcam), anti-AKT1p antibody(1:500, ab81283, Abcam), anti-p21 antibody(1:500, AM2385, Abzoom),anti-Ki67 antibody(1:250, ab92742, Abcam), anti-GAPDH antibody (1:2500, ab9485, Abcam), anti-Histone H3 antibody (1:2500, 2134–1,Epitomics), Anti-beta Catenin (phospho S33 + S37) antibody (1:1000, ab75777, Abcam).
Expression profiling and Gene Ontology analysis
In order to identify genes subjected to SOX4’s regulation in GBM cells, total RNAs from SOX4 overexpression stable cell lines LN229_pSFH and LN229_con were used for expression profiling using the PrimeView Human Gene Expression Array (Affymetrix) using a standard Affymetrix protocol. The array data were submitted to the GEO database (GSE51301). Gene Ontology analysis was done using the DAVID program (
http://david.abcc.ncifcrf.gov/) [
26],[
27].
Statistical analysis
Statistical analysis was performed using the Student’s t-test. P <0.05 was considered to be statistically significant. The results are reported as mean ± SD.
Discussion
We reported our finding that high SOX4 expression was significantly associated with good prognosis of primary GBM patients (Figure
1). However, overexpression of SOX4 in GBM cell lines inhibited cell proliferation and G0/G1 cell cycle arrest (Figures
2 and
3). These seemly contradictory observations between observation in primary clinical samples and in vitro cell line functional analysis is similar to what Aaboe et al. observed in bladder cancer [
4]. They found that SOX4 is over expressed in bladder cancer tissues compared to normal tissues, but strong SOX4 expression was found to be correlated with increased patient survival (P <0.05) of bladder cancer patients [
4], and when introduced to bladder cancer cell line HU609, it reduced cell viability by promoting apoptosis and necrosis [
4]. The exact mechanism of this seemly contradict phenomena need further investigation. Sox factors function as either tumor activators or repressors depending on the cellular context and associated interacting proteins [
38]. One of the differences between the cell lines we used and the GMB tumor tissues is the P53 mutation status. P53 are often mutated in GBM tumors, seen in 25–30% of primary GBM and 60–70% of secondary GBM [
39]. However, the GBM cell line LN229, A172 and U87 that we used in the study turned out to contain wide type p53 [
40]. One possible mechanism is that SOX4 acts differently on the wild-type P53 than the mutated P53. Further experimentation is needed to verify this speculation.
The p53 protein is an important tumor suppressor that is inactivated in most tumors. It controls growth arrest and apoptosis by its transcriptional activation and repression of target genes [
41]. p53 is thought to exert its function in G1 checkpoint control mainly through increased expression of the p21 gene [
42]. We found p53 were increased and were accumulated in the nucleus with SOX4 (Figure
6B), while p53 s downstream target p21 was up-regulated (Figure
6A), and leading to G0/G1 cell cycle arrest (Figure
4C). In addition, SOX4 reduced expression levels of AKT1 and phosphorylated AKT1 (Figure
5A), which could enhance Mdm2-mediated ubiquitination and degradation of p53 [
36],[
43]. Thus, we propose a model to explain SOX4’s action in GBM cell lines: SOX4 could down-regulate Akt, which might result in increased stability of p53, as shown by increased p53 protein expression in SOX4 overexpressed GBM cell lines and induced p21 signal which leading to G0/G1 cell cycle arrest.
SOX4 has been proposed either as an oncogenor or a tumor suppressor. Our finding that SOX4 acts as a tumor suppressor in GBM is consistent with previously reported roles of SOX4 as a tumor suppressor. For example, SOX4 induces cell apoptosis through caspase activation in hepatocarcinoma cells and HeLa cells [
14],[
15]. Nuclear overexpression of SOX4 in HCC samples is correlated with diminished risk of recurrence and improved overall survival time in HCC patients [
32]. Pan
et al. reported that SOX4 promotes cell cycle arrest and apoptosis, and inhibits tumorigenesis by blocking Mdm2-mediated p53 degradation and facilitating p300/CBP/p53complex formation [
34]. In lung cancer cell line H1299, SOX4 increases PUMA expression in response to trichostatin A (TSA) and induces apoptosis in a p53 independent manner [
44].
Our data proved that SOX4 inhibited cell growth and induced G0/G1 cell cycle arrest in GBM cells. We found that AKT1 and AKT1p were reduced while SOX4 overexpression in GBM and p53 protein level was increased and co-localized in nuclear with SOX4. However, SOX4 increased beta-catenin expression seems to limited to the cytoplasm and failed to initiate cytoplasm-to-nuclear translocation of beta-catenin, suggesting the action of SOX4 seems unlikely to involve directly transcriptional changes of beta-catenin directly regulated genes.
Microarray analysis showed that SOX4 regulated genes were enriched in Gene Ontology terms related to cell cycle progression (Table
1 and Additional file
5: Table S3), consistent with our functional assay showing SOX4’s role in cell cycles. We also identified many interesting genes regulated by SOX4, suggesting SOX4’s diverse roles in cancer cells. For example, we found SOX2 is upregulated by SOX4 (Additional file
5: Table S3), which is consistent with previous findings that the transforming growth factor-β (TGF-β)-Sox4-Sox2 pathway is essential for glioma-initiating cells to retain their stemness [
18],[
19]. We also identified several other genes involved in TGF-beta signaling include BMP2 and SOX10, whose expression were also changed after SOX4 overexpression (Additional file
5: Table S3).
Finally, we demonstrated that beta-catenin was upregulated by SOX4 (Figure
5B-D), suggesting a role of SOX4 in WNT signaling in GBM cells. SOX4 was previously shown to enhance beta-catenin/TCF activities in several cancer types [
29]-[
31]. p53 can reduce the proliferation-supporting effect of beta-catenin/TCF signaling by downregulating Tcf-4 expression even with transcriptional activation by beta-catenin in the nucleus [
45]. Considering SOX4 increased both p53 and beta-catenin in GBM cells, there might be a cross talk between these two signals.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
BL, and JZ designed this study. JZ, HJ, RM performed the experiments. JF performed the TCGA data analysis and survival analysis. JZ, JL performed data analysis. SZ, YM, XF, NZ help to design the study and wrote the paper. All authors read and approved the final manuscript.