Tel: +86(20)87332748; Fax: +86(20)87331209
Early Detection and Diagnosis
Sp1 is upregulated in human glioma, promotes MMP-2-mediated cell invasion and predicts poor clinical outcome
Article first published online: 9 JUN 2011
Copyright © 2011 UICC
International Journal of Cancer
Volume 130, Issue 3, pages 593–601, 1 February 2012
How to Cite
Guan, H., Cai, J., Zhang, N., Wu, J., Yuan, J., Li, J. and Li, M. (2012), Sp1 is upregulated in human glioma, promotes MMP-2-mediated cell invasion and predicts poor clinical outcome. Int. J. Cancer, 130: 593–601. doi: 10.1002/ijc.26049
- Issue published online: 24 NOV 2011
- Article first published online: 9 JUN 2011
- Accepted manuscript online: 5 APR 2011 07:58AM EST
- Manuscript Accepted: 14 FEB 2011
- Manuscript Received: 1 JUN 2010
- The Ministry of Science and Technology of China. Grant Number: (973)2005CB724605
- the Natural Science Foundation of China. Grant Numbers: U0632002, 40806059, 81001190
- Guangdong Provincial Natural Science Foundation. Grant Numbers: 06201946, 10451008901004231
- National High-Tech Research Program (863 Program). Grant Numbers: 2007AA09Z431, 2007AA09Z448
- National Science and Technique Major Project. Grant Numbers: 2009ZX09103-041, 201005022-2, 2009ZX10004-213
- the Key Science and Technique Research Project of Guangdong Province. Grant Numbers: 2007A03260001, 2009A020101006, 2009B020600001, 2010B030600003
Sp1, the first identified transcription factor, has been reported to be associated with the development and progression of various human cancer types. However, the clinical significance and biological role of Sp1 in glioma are less well understood. In this study, we found that the expression of Sp1 was markedly elevated in glioma cell lines and tissues. Immunohistochemistry analysis revealed that the vast majority of 222 paraffin-embedded archival glioma specimens tested displayed positive Sp1 expression, and 58.6% exhibited high-level Sp1 expression. Statistical analysis suggested that the high Sp1 expression was correlated strongly with the WHO grading (p < 0.001) and survival status (p < 0.001) of glioma patients. Patients with lower Sp1 expression had better overall survival than those with higher Sp1 expression. Multivariate analysis suggested that Sp1 expression might be an independent prognostic indicator of the survival of patients with glioma. Furthermore, overexpression of Sp1 in glioma cells was found to increase their invasiveness, and in contrast, silencing Sp1 by siRNA caused an inhibition of cell invasion. Moreover, we demonstrated that the up-regulation of Sp1 could increase activity and expression of MMP-2. Collectively, our data suggest that Sp1 might represent a valuable prognostic marker for glioma and is involved in modulation of tumor invasion.
Glioma is the most common central nervous system tumor in adults. Despite that great progress has been made in the surgical, radioactive and chemical antiglioma modalities, the prognosis of the disease remains poor, with the 5-year survival rate for glioblastoma multiforme being only 2%.1, 2 While the molecular pathology responsible for the aggressiveness of human glioma and the suboptimal response to conventional therapies is largely unknown, existing data suggest that a diversity of biological changes in glioma cells may account for the poor overall survival of patients with glioma.3 Understanding these molecular characteristics that are associated with the development, progression and response or resistance to therapies will bring new insights into gliomagenic mechanisms and new therapeutic strategies.
Specificity protein 1 (Sp1) is the first identified transcription factor that activates a broad spectrum of cellular and viral genes.4 Sp1 protein recognizes different but related G-rich motifs and interacts with DNA through 3 C2H2-type zinc fingers located at the C-terminus.5 A mass of evidence has shown that Sp1 plays a pivotal role in the regulation of multiple genes that have been recognized to be important for tumorigenesis.6 In a number of human cancer types, including breast cancer,7 thyroid tumor,8 gastric cancer,9 epidermal squamous cell carcinoma10 and pancreatic cancer,11 elevation of Sp1 expression has been reported, in some of which Sp1 overexpression is associated with poor clinical outcome.9, 11 To our knowledge, however, no study has examined the expression of Sp1 and its impact on the outcome of glioma patients.
The poor clinical outcome of glioma is largely due to their infiltrating nature and recurrence at adjacent or distant regions in the brain.12 Degradation of extracellular matrix (ECM) and basement membrane, catalyzed by proteases, is a crucial step for a tumor to invade, and matrix metalloproteinases (MMPs) are major proteolytic enzymes involved in degradation of these barriers.13 Human glioma cells express a variety of MMPs, among which MMP-2 and MMP-9 are believed to most effectively degrade most of ECM components.14 Increased expression of MMP-2 and MMP-9 have been strongly implicated in the invasion of various types of human cancers.15, 16 Our recent study showed that MMP-9 gene transcription can be activated by a mechanism involving the astrocyte elevated gene-1 (AEG-1) protein and promotes AEG-1 induced glioma invasion.17 MMP-2 has been found to be overexpressed and correlate with tumor invasion in glioma.13, 14 The mechanisms that activates MMP-2 gene transcription in human glioma cells, however, are still to be elucidated. Notably, the 5′ regulatory region of human MMP-2 gene contains a number of cis-acting elements, including Sp1, Sp3, and Ap-2. Moreover, Pan and coworkers18 have shown that Sp1 plays a key role in controlling MMP-2 expression in lung cancer. These studies suggest a possible involvement of Sp1 in the regulation of MMP-2 gene transcription and consequently the molecular mechanism mediating the invasive and angiogenic phenotype of glioma cells.
In this study, we examined the expression of Sp1 in human glioma and investigated the biological role of Sp1 in promoting the progression of the disease. Our results showed that Sp1 was upregulated in both glioma cell lines and glioma tissues and correlated inversely with the clinical outcome. Furthermore, we found that Sp1 induced invasion of glioma cells through transactivating the expression of MMP-2. Taken together, the data derived from this study provide a basis for further clarification of the molecular mechanism underlying the invasion ability of glioma cells and for future development of new therapeutic strategies.
Material and Methods
Cell lines and reagents
Primary normal human astrocytes (NHA) were purchased from the Sciencell Research Laboratories (Carlsbad, CA) and cultured according to the instruction of the manufacturer. Glioma cell lines LN-18, LN-382, U87MG, U251MG, LN-444, SNB19, U118MG and LN-428 were a kind gift from Dr. Shi-Yuan Cheng (Department of Pathology, University of Pittsburg) and maintained in DMEM supplemented with 10% FBS. MMP-2 inhibitor I (Cis-9-Octadeconyl-N- hydroxylamide, OA-Hy) was purchased from Calbiochem (San Diego, CA).
Patients and tissue specimens
This study was conducted on a total of 222 paraffin-embedded glioma specimens, which were histopathologically diagnosed at the First Affiliated Hospital of Sun Yat-Sen University from 2000 to 2005. Five biopsy specimens of human gliomas (Sample 1, WHO Grade III; Sample 2, WHO Grade IV; Sample 3, WHO Grade II; Sample 4, WHO Grade II; Sample 5, WHO Grade III) and the matched adjacent non-cancerous brain tissues were collected, frozen and stored in liquid nitrogen until assayed. Informed consent from patients and ethics approval from the Institutional Research Ethics Committee were achieved. Clinical information of the samples is presented in detail in Table 1. Normal brain tissues were taken from individual donations died in traffic accidents and confirmed to be free of any detectable pathological conditions.
Western blotting (WB) was performed according to a standard method as described previously.19 The following primary antibodies were purchased: anti-Sp1 (Upstate, Lake Placid, NY) and anti-α-tubulin (Sigma-Aldrich, Saint Louis, MO).
Immunohistochemistry (IHC) analysis was performed to study altered protein expression in human glioma and normal tissues. The degree of immunostaining of Sp1 protein was evaluated and scored independently by 2 observers as previous described, considering both the proportion of positive staining tumor cells and the staining intensity.19 Scores representing the proportion of positive stained tumor cells was graded as: 0 (no positive tumor cells), 1 (<10%), 2 (10–50%), and 3 (>50%). The intensity of staining was determined as: 0 (no staining); 1 (weak staining), 2 (moderate staining) and 3 (strong staining). The staining index (SI) was calculated as the product of staining intensity and percentage of positive tumor cells, resulting in scores as 0, 1, 2, 3, 4, 6 and 9. Cutoff values for Sp1 were chosen based on a measurement of heterogeneity using the log-rank test with respect to overall survival. We identified the optimal cutoff as: the SI score of ≥4 was considered as high Sp1 expression, and ≤3 as low expression of Sp1.
IHC staining for protein expression in tumor and normal tissues was quantitative analyzed with the AxioVision Rel.4.6 computerized image analysis system assisted with the automatic measurement program (Carl Zeiss, Oberkochen, Germany). Briefly, to assess the Mean Optical Density (MOD), which represents the strength of staining signals by measuring per positive pixels, we evaluated the stained sections at 200× magnification and ten representative staining fields of each section were analyzed.
Plasmid and generation of stable cell lines
Ectopic expression of Sp1 in glioma cells was achieved using a retroviral gene delivery method. Briefly, human Sp1 cDNA was cloned into retroviral transfer plasmid pMSCV to generate Sp1 expression vector, co-transfecting with the pIK packaging plasmid in 293FT cells by using standard calcium phosphate transfection method as previously described.20 Totally, 36 hr after the cotransfection, supernatants were collected and incubated with LN-18 and LN-382 cells for 24 hr in the presence of polybrene (2.5 μg/mL). Following transduction, puromycin (1.5 μg/mL) was used as a selection antibiotic to select the infected cells for 10 days. Expression of the Sp1 was tested by WB analysis.
Silencing endogenous Sp1 with siRNA
Sp1 specific siRNA oligonulciotides were purchased from Ribobio (Guangzhou, China). siRNA sequences of sense strands are as follows: Sp1-siRNA1, 5′-GGCUGGUGGUGAUGGAAUAdTdT-3′; and Sp1-siRNA2, 5′-GGCUGGUGGUGAUGGAAUAdTdT-3′. siRNA was transfected into cells cultured in 6-well plates using Lipofectamine 2000 following the manufacturer's instructions (Invitrogen, Carlsbad, CA). Following transfection with siRNA, cells were cultured for 2 days before use.
Transwell invasion assay
Cell invasion assays were performed using an in vitro Matrigel assay as described previously.17 Briefly, Transwell inserts for 24-well plates (Corning Costar Corp., Cambridge, MA) were coated with prediluted Matrigel (BD Biosciences, Bedford, MA) and allowed to gel at 37°C for 30 min. Cells were seeded at a density of 3 × 105 per insert and the lower chamber of the Transwell was filled with 500 μL DMEM supplemented with 10% FBS. After 24 hr of incubation, cells remaining on the upper surface of the Transwell membrane were removed by a cotton swab. Cells that had invaded through the Matrigel to the bottom of the insert were fixed, stained, photographed and quantified by counting them in 6 random high-powered fields.
MMP-2 expression in conditioned medium of Sp1 overexpressed or knocked down glioma cells was analyzed by gelatin zymography as previously described.17
RNA extraction and real-time RT-PCR
RNA extraction, RT and real-time PCR were performed as described previously.19 Sequences of the primers and probes are as follows: MMP-2 forward primer 5′-TGAGCTATGG ACCTTGGGAGAA-3′, MMP-2 reverse primer 5′-CCAT CGGCGTTCCCATAC-3′, and MMP-2 probe 5′-FAM-CCAA GTGGTCCGTGTGA-TAMRA-3′; GAPDH forward primer 5′-GACTCATGACCACAGTCCATGC-3′, GAPDH reverse primer 5′-AGAGGCAGGGATGATGTTCTG-3′, and GAPDH probe 5′-FAM-CATCACTGCCACCCAGAAGACTGTG-TA MRA-3′. Expression data were normalized to the geometric mean of housekeeping gene GAPDH to control the variability in expression levels and calculated as 2-[(CTofSp1/MMP-2)−(CTofGAPDH)], where CT represents the threshold cycle for each transcript.
All statistical analyses were carried out using the SPSS 11.0 statistical software package. The relationship between Sp1 expression and the clinicopathologic characteristics was assayed by the χ2 test. The Spearman rank correlation coefficients were calculated to determine the bivariate correlations between study variables. The Kaplan-Meier method was used to construct the survival curves and the survival differences were compared by the log-rank test. Using multivariate Cox regression analyses to verify the independent effect of each variable involved in this study. Continuous data were compared using Student's t test. In all cases, p < 0.05 was considered statistically significant.
Expression of Sp1 in glioma cell lines and in paired glioma and nontumorous tissues
To investigate the expression levels of Sp1 protein in glioma cell lines, WB analysis was done in NHA and glioma cell lines LN-18, LN-382, U87MG, U251MG, LN-444, SNB19, U118MG and LN-428. All 8 cell lines showed higher level expression of Sp1 protein than that in the NHA (Fig. 1a). To further determine whether Sp1 is also overexpressed in clinical samples of human glioma, WB analysis was conducted with 5 pairs of matched glioma tissue and nontumorous tissue adjacent to the malignant lesion, with each pair obtained from a same patient. As shown in Figure 1b, compared with their paired adjacent nontumorous tissues from the same patient, Sp1 was found to be overexpressed in all 5 examined human glioma samples. Thus, our data indicate that Sp1 was upregulated in glioma and might represent a significant molecular feature of glioma pathogenesis.
Expression of Sp1 in archival glioma tissues
Based on the above observation that Sp1 was upregulated both in 8 glioma cell lines and in 5 glioma tumor specimens, we further examined, using IHC, the expression pattern of Sp1 in 222 paraffin-embedded, archived glioma tissues (including 25 cases of Grade I, 61 cases of Grade II, 90 cases of Grade III, and 46 cases of Grade IV). As shown in Figure 1c, Sp1 protein was found to be upregulated in all grades of human glioma compared with that in normal brain tissue. Notably, the subcellular location of Sp1 was detected mainly in the nucleus. Furthermore, quantitative analysis verified that average staining of Sp1 in Grades I to IV tumors was significantly higher than that in normal brain tissue (p < 0.05; Fig. 1d). As summarized in Table 1, Sp1 protein was detectable in 210 of 222 (94.6%) examined glioma cases, with 130 cases (58.6%) displaying high level expression of Sp1.
Correlation between Sp1 protein expression and glioma clinicopathologic features
To further the understanding of the role of Sp1 in the development of glioma, we next analyzed the correlation between Sp1 expression level and clinicopathologic features of glioma patients. While no significant correlation between the Sp1 protein expression and the age (p = 0.905) or sex (p = 0.107) of glioma patients (Table 2) was found, the expression of Sp1 closely correlated with WHO grading (p < 0.001) or survival status of glioma patients (p < 0.001). Moreover, Spearman correlation analysis further revealed that the Sp1 protein expression was positively correlated with WHO grades of glioma (r = 0.263, p < 0.001). Taken as a whole, these data support the notion that the progression of glioma is associated with increased Sp1 expression.
Kaplan-Meier analysis using the log-rank test was performed to calculate the effect of Sp1 expression on glioma patients' survival (Fig. 1e). The result showed that high expression of Sp1 was markedly associated with a reduced overall survival (p < 0.001). The median survival time of patients with high Sp1 expression (30.87 months, 95% confidence interval: 26.05–35.68) was significantly shorter than that of patients who had low Sp1 expression level (51.90 months, 95% confidence interval: 43.16–60.63). Moreover, the cumulative 3-year survival rate was 56.4% in the low Sp1 expression group, whereas it was only 27.2% in the high Sp1 expression group. We next investigated whether Sp1 expression represented an independent prognostic factor for glioma. Univariate analysis of individual variables revealed strong relationships between overall survival and age (p = 0.006), WHO grade (p < 0.001), and Sp1 expression (p < 0.001). Furthermore, multivariate analysis confirmed Sp1 expression (HR, 2.218; 95% CI, 1.164–4.225; p = 0.015) and WHO grade (HR, 4.410; 95% CI, 2.225–8.740; p < 0.001) as significant prognostic factors for glioma (Table 3).
We next assessed the prognostic significance of expression level of Sp1 protein in different subgroups of glioma patients stratified according to the WHO grading. Notably, high Sp1 expression also significantly correlated with shorter overall survival time in different glioma subgroups. Overall survival of patients with high Sp1 expressing tumors was significantly decreased than those with tumors expressing low levels of Sp1 in either Grades I+II subgroup (n = 44; p = 0.019, log-rank; Fig. 1f) or Grades III+IV subgroup (n = 76; p = 0.016, log-rank; Fig. 1g). Moreover, the number of patients of low- and high-Sp1 expression in each WHO grade is as following: Grade I, low expression: 10, high expression: 4; Grade II, low expression: 18, high expression: 12; Grade III, low expression: 10, high expression: 44; and Grade IV, low expression: 2, high expression: 20. These data collectively suggest that Sp1 expression level strongly and significantly correlate with the prognosis of glioma and the disease outcome.
Effect of Sp1 on glioma cell invasion
To investigate whether and how Sp1 impacts on the aggressive nature of glioma cells, LN-18 and LN-382 glioma cells were engineered to overexpress Sp1 (Fig. 2a) and tested for their ability of invasion. As shown in Figures 2c and 2e, Sp1 overexpression efficiently promoted glioma cells invasion as demonstrated by using the Matrigel invasion assay. To further demonstrate the role of endogenous Sp1 in glioma cell invasion, the effect of Sp1 knockdown on glioma cell invasion was examined. Knockdown of Sp1 expression by specific siRNA1 (Fig. 2b) markedly inhibited the invasive capability of LN-18 and LN-382 glioma cells (Figs. 2d and 2f), suggesting that Sp1 plays a role in the development of the invasion ability of glioma cells.
Induction of MMP-2 expression by Sp1
To further analyze the molecular mechanism via which Sp1 enhances glioma cell invasion, we compared the activity of MMP-2 in control and Sp1-overexpressing cells by using zymography assays. As shown in Figure 3a, Sp1 overexpression significantly enhanced the enzyme activity of MMP-2, indicating a possible correlation between MMP-2 expression and the function of transcription factor Sp1. Indeed, real time RT-PCR showed that ectopic expression of Sp1 resulted in increased MMP-2 mRNA expression (Fig. 3c). In contrast, depletion of endogenous Sp1 caused markedly decreased activity and mRNA level of MMP-2 (Figs. 3b and 3d). Furthermore, it is of interest that when the MMP-2 activity was suppressed by a specific inhibitor, as shown in Figure 3e, the enhancing effect of Sp1 on glioma cell invasion was negated, suggesting that MMP-2 may be required for Sp1 induced glioma cell invasion.
The present study shows, for the first time, that Sp1 is overexpressed in human glioma, and that the Sp1 overexpression is significantly associated with the clinical and pathological features, as well as the prognosis, of the disease. Together with this finding, we also have demonstrated that Sp1 might promote the invasiveness of glioma cells, possibly through upregulating MMP-2 expression.
Our study adds to the current understanding on the role of Sp1 in tumorigenesis and progression of malignant tumors. With its central regulatory role in many pathways, many of which are associated with oncogenesis, such as cell proliferation, resistance to apoptosis, cell invasion and angiogenesis, Sp1 may be a powerful predictor of the outcome of glioma patients. Sp1 overexpression has been reported in several cancer types. Moreover, Sp1 was indicated as an independent predictor of patients survival.9, 11 Its connection with human glioma, however, has remained unknown. Evidence in support of such a connection, provided by our current study, includes positive results of Sp1 detection in 8 glioma cell lines, in tumor tissues paired with adjacent nontumorous tissue controls, and in the vast majority of a cohort of 222 archived clinical glioma specimens. Further support for a possible role of Sp1 in glioma pathogenesis derived from the analysis that revealed a strong correlation of Sp1 expression with the histopathological staging and inversely, with the survival of the disease. These data not only suggest that Sp1 is likely to be biologically involved in the progression of glioma, but also might represent a valuable independent prognostic biomarker for the disease.
While this study has provided strong evidence for the upregulation of Sp1 in glioma, the molecular mechanism underlying the observed Sp1 upregulation is yet to be elucidated. Sp1 overactivation during tumor development and progression has been shown to be mediated by genetic and epigenetic pathways.21, 22 Nicolás et al. have previously demonstrated that Sp1 protein regulates its own promoter,23 therefore the Sp1 gene can be autoregulated. In addition, Sp1 is regulated by interaction with other proteins, among them prominent oncogenes and tumor suppressors. For example, binding of MDM2 (mouse double minute 2) to C-terminal domain of Sp1 suppresses the DNA-binding capability of Sp1.24 Furthermore, pRb can abolish this suppression by releasing Sp1 from the MDM2-Sp1 complex, thus restoring its DNA binding capability. Moreover, at the protein level, DNA-binding activity and transactivation potential of Sp1 are modulated by several post-translational modifications, including phosphorylation, acetylation, glycosylation, ubiquitylation, sumoylation and poly(ADP-ribosyl)ation.25 Therefore, either increased gene expression or posttranslational modification can cause Sp1 overactivation. Further studies that aim at delineating the mechanisms of Sp1 regulation in glioma will improve our understanding not only of the function of Sp1 but also of gliomagenesis and are underway in the laboratory.
There have been studies addressing how Sp1 influences the outcome of several other cancers. In a previous report, for instance, Yao and coworkers9 have demonstrated that deregulated Sp1 expression plays an important role in gastric cancer tumorigenesis by enhancing VEGF expression. In the scenario of glioma, a biologically and clinically relevant cause for the high mortality lies in their capability to invade extensively within the brain.12 Invasion of glioma cells is a multistep process involving cancer cell attachment to ECM, degradation of ECM components and subsequent infiltration of into adjacent brain tissue.26 The accomplishment of this process, as shown by several lines of research, is largely attributable to the activation of MMPs.27 On the other hand, compelling evidence has linked the up-regulation of Sp1 protein to enhanced invasion of cells of various types of solid tumors.28–30 Our study demonstrates the pathological role of Sp1 in enhancing MMP-2 expression and mediating the invasive phenotype of human glioma cells. However, as a transcription factor, Sp1 plays critical roles in the metastasis of many tumor types by regulating various MMP family members. For example, several lines of evidence suggest that the transcription factor Sp1 is required for constitutive MMP-9 gene expression in various cancer types.31, 32 Sroka et al.'s studies implicated a role for Sp1 in regulating constitutive levels of membrane type-1 matrix metalloproteinase (MT1-MMP, MMP-14) in prostate cancer cell lines.33 Thus, further work is necessary to evaluate the full mechanisms by which Sp1 exerts its proinvasion effects in glioma. The data derived from the mechanistic investigation of this study provide a possible explanation for the observed relevance of Sp1 overexpression to predicting the prognosis of human glioma. It would also be of great interest to investigate whether similar molecular cascade, namely, upregulation of MMP-2 by Sp1 overexpression, is also present in other caner types than gliomas.
In summary, our study indicates that Sp1 is an independent prognostic marker for human glioma and that its deregulated expression may contribute to glioma progression by enhancing MMP-2 expression and tumor invasion. It is hoped that these findings will provide new insights in the development of novel antiglioma strategies.
- 2Cell-mediated immunotherapy: a new approach to the treatment of malignant glioma. Cancer Control 2003; 10: 138–47., , .