|Home | About | Journals | Submit | Contact Us | Français|
Astrocyte elevated gene 1 (AEG-1) is an oncoprotein that strongly promotes the development and progression of cancers. However, the detailed underlying mechanisms through which AEG-1 enhances tumor development and progression remain to be determined. In this study, we identified c-Jun and p300 to be novel interacting partners of AEG-1 in gliomas. AEG-1 promoted c-Jun transcriptional activity by interacting with the c-Jun/p300 complex and inducing c-Jun acetylation. Furthermore, the AEG-1/c-Jun/p300 complex was found to bind the promoter of c-Jun downstream targeted genes, consequently establishing an acetylated chromatin state that favors transcriptional activation. Importantly, AEG-1/p300-mediated c-Jun acetylation resulted in the development of a more aggressive malignant phenotype in gliomas through a drastic increase in glioma cell proliferation and angiogenesis in vitro and in vivo. Consistently, the AEG-1 expression levels in clinical glioma specimens correlated with the status of c-Jun activation. Taken together, our results suggest that AEG-1 mediates a novel epigenetic mechanism that enhances c-Jun transcriptional activity to induce glioma progression and that AEG-1 might be a novel, potential target for the treatment of gliomas.
Glioma is the most common and aggressive type of central nervous system tumor (1). Despite intensive research and clinical efforts, the prognosis for patients with this tumor type remains poor, largely attributable to its highly invasive and fast proliferating phenotype. The median life expectancy of patients with a grade IV glioma, known as a glioblastoma multiforme (GBM), is less than 1 year (2). Therefore, the definition of appropriate targets against which effective strategies to treat glioma may be developed represents a major goal in glioma research. A better comprehension of the molecular mechanisms mediating glioma progression is crucial to developing an efficacious therapeutic strategy that prevents the infiltration, invasion, and proliferation of glioma cells.
The product of the gene astrocyte elevated gene 1 (AEG-1), also known as the metadherin (MTDH) or LYRIC gene, was initially identified to be a novel protein whose expression is induced by human immunodeficiency virus type 1 (HIV-1) or by tumor necrosis factor alpha (TNF-α) in primary human fetal astrocytes (3,–6). AEG-1 is a multifunctional protein that interacts with diverse partners in different types of cancers and promotes the development of essentially all hallmarks of cancer (7,–12). Previous studies have found that through its protein-protein interactions, the product of the AEG-1 gene is a key pathological factor that regulates a variety of diseases associated with several signaling pathways, including the nuclear factor κB (NF-κB), phosphatidylinositol 3-kinase/AKT, mitogen-activated protein kinase (MAPK), and Wnt pathways (8, 13,–17). Recently, Li et al. demonstrated that AEG-1 promotes gastric cancer progression through a positive-feedback Toll-like receptor 4/NF-κB signaling-related mechanism (14). Moreover, Robertson and colleagues found that AEG-1-deficient mice display resistance to N-nitrosodiethylamine-induced hepatocellular carcinoma and lung metastasis (18).
Although numerous studies have established a clinicopathological correlation between the levels of AEG-1 expression and various aspects of malignant features of cancer, the underlying molecular mechanism through which AEG-1 exerts its oncogenic activities requires further clarification. We previously demonstrated that AEG-1 upregulates matrix metallopeptidase 9 (MMP-9) expression, possibly through both NF-κB- and c-Jun-mediated mechanisms (19). In agreement with our results, Emdad et al. reported that AEG-1 interacts with p65 and promotes the nuclear translocation of NF-κB (7). However, the molecular mechanism through which AEG-1 activates the c-Jun pathway remains largely unknown.
In the current study, we identified AEG-1 to be a novel interacting partner of c-Jun and p300. Furthermore, we demonstrated that these interactions promote c-Jun acetylation and chromatin remodeling, which led to upregulated expression of the c-Jun target genes and, consequently, promoted angiogenesis and tumor cell survival. Our findings reveal a new molecular mechanism through which AEG-1 promotes cancer progression, which might further suggest a clinical value of AEG-1 as a novel therapeutic target.
Our previous studies that indicated the essential involvement of AP-1 activation in the effects of AEG-1 on gliomagenesis (19) prompted us to further study the role of AEG-1 in c-Jun signaling. To this end, we first performed gene expression microarray analysis and found that the c-Jun target gene sets were downregulated in AEG-1-specific small interfering RNA (siRNA)-transfected cells compared with their regulation in scrambled siRNA-transfected control cells (Fig. 1A; see also Table S1 in the supplemental material). Next, we comparatively assessed the DNA-binding affinity of the c-Jun complex in AEG-1-overexpressing and AEG-1-knockdown glioma cell models (Fig. 1B) using an electrophoretic mobility shift assay (EMSA). As anticipated, the formation of the c-Jun/DNA complex was enhanced by ectopic expression of AEG-1 but was decreased by silencing of AEG-1 in glioma cells (Fig. 1C). Furthermore, the activity level of a luciferase reporter driven by the binding of the c-Jun-binding motif (TRE) was significantly increased in AEG-1-tranduced cells and was decreased in AEG-1-silenced cells (Fig. 1D). Taken together, these data suggest that AEG-1 promotes c-Jun transactivation and induces the expression of c-Jun-dependent downstream genes in glioma cells.
Next, we explored the mechanism underlying AEG-1-mediated c-Jun transactivation. Because the increase in c-Jun activity is partially due to c-Jun phosphorylation, we began with examining the c-Jun phosphorylation levels in glioma cells subjected to AEG-1 overexpression or AEG-1 knockdown. Our results showed no significant change in the level of c-Jun phosphorylation upon AEG-1 overexpression or knockdown (Fig. 1E). As the transactivational activity of c-Jun is also modulated by c-Jun acetylation (20,–22), we further sought to determine whether AEG-1 regulates c-Jun acetylation. As shown in Fig. 1F, the acetylation level of c-Jun was significantly increased in cells overexpressing AEG-1 and was significantly decreased in cells in which AEG-1 was knocked down compared with the AEG-1 levels in the control cells. Notably, the expression of a c-Jun mutant (K3R-c-Jun), which blocks acetylation-dependent AP-1 transactivation (20, 22), drastically diminished the stimulatory effect of AEG-1 on the transactivational activity of c-Jun (Fig. 1G), suggesting that the effect of AEG-1 on c-Jun transactivation is associated with the increased acetylation of c-Jun.
Since the transcriptional efficiency of c-Jun could be modulated by the p300 acetylation of c-Jun (20,–22), we further investigated whether deregulated AEG-1 affects the acetyltransferase activity of endogenous p300 acetyltransferase. As shown in Fig. 2A, the level of p300 acetyltransferase activity was significantly higher in AEG-1-overexpressing cells than in control cells. In contrast, the level of endogenous p300 acetyltransferase activity was dramatically deceased in AEG-1-knockdown cells (Fig. 2A). Moreover, the results of a coimmunoprecipitation (co-IP) assay performed using an antibody directed against AEG-1 demonstrated that AEG-1 interacted with both p300 and c-Jun (Fig. 2B). Reciprocal co-IP assays using antibodies directed against p300 or c-Jun showed that AEG-1 coimmunoprecipitated with p300 and with c-Jun (Fig. 2B). Consistent with the observation in glioma cell lines, AEG-1 indeed interacts with both p300 and c-Jun in patient-derived glioma cells (Fig. S1A). These results was then further confirmed using cells ectopically expressing the three tagged proteins using agarose beads conjugated to antibodies directed against the tag epitopes (Fig. 2C). Additionally, amino-terminal (N) segments (N1 to N7), carboxyl-terminal (C) segments (C1 to C3), and middle segments (M4) of AEG-1 could be coimmunoprecipitated with hemagglutinin (HA)-tagged p300, whereas the middle portions of AEG-1 (M1, M2, and M3) did not interact with HA-tagged p300 (Fig. 2D). These data suggest that both the N terminus (amino acids [aa] 71 to 205) and the C terminus (aa 404 to 513) of AEG-1 contribute to its interaction with p300. Moreover, although the amino-terminal (N) segments (N5 to N7) and middle segment (M1) fragments of AEG-1 could not be coimmunoprecipitated with Myc-tagged c-Jun, the amino-terminal (N) segments (N1 to N4), carboxyl-terminal (C) segments (C1 to C3), and middle segments (M2 to M4) of AEG-1 could be coimmunoprecipitated with Myc-tagged c-Jun (Fig. 2D), suggesting that the middle region (aa 232 to 262) of AEG-1 contributes to its interaction with c-Jun.
Furthermore, we found that the overexpression of AEG-1 promoted the interaction between p300 and c-Jun, whereas silencing of AEG-1 expression inhibited this interaction (Fig. 2E and andFF and S1B and C). However, the extent of the interaction between AEG-1 and p300 was not affected by c-Jun knockdown (Fig. 2G).
Because the acetylation of promoter or enhancer elements can play an important role in the transcriptional activation of the corresponding genes, we sought to investigate whether enhancing the level of AEG-1-meditated p300 acetyltransferase activity could increase the level of activation of chromatin modifications. We chose to analyze vascular endothelial growth factor (VEGF), a c-Jun target gene (23,–25), the expression of which was significantly downregulated in AEG-1 siRNA-transfected cells (Fig. 1A). As shown in Fig. 3A, the level of VEGF expression was markedly increased in cells overexpressing AEG-1 and was decreased in cells subjected to AEG-1 silencing. Meanwhile, the expression of VEGF could be partially but not completely rescued by expressing c-Jun in cells in which AEG-1 was interrupted by RNA interference (RNAi) (AEG-1-RNAi cells) (Fig. S2A). Binding site prediction analysis performed using the TRANSFAC (version 4.0) program revealed four putative c-Jun-binding sites within the VEGF promoter (Fig. 3B). The results of chromatin immunoprecipitation (ChIP) assays revealed that AEG-1 bound the second and third potential c-Jun-binding sites within the VEGF promoter (Fig. 3C). The results of a dual-luciferase reporter assay showed that the luciferase activities driven by the VEGF promoter, covering segments of nucleotides −1140 to +690, −1700 to +690, and −2200 to +690, were increased in AEG-1-overexpressing cells compared with those in the control cells (Fig. 3D and andE).E). Conversely, the luciferase activities driven by these segments of the VEGF promoter decreased in AEG-1-silenced cells (Fig. 3D and andE).E). Furthermore, we evaluated whether the c-Jun-binding site in the VEGF promoter is essential for AEG-1-induced luciferase activation. As shown in Fig. 3F, the luciferase activities in AEG-1-overexpressing cells decreased when the second or third potential c-Jun-binding site within the VEGF promoter was mutated, suggesting that AEG-1 promotes VEGF promoter transactivation through second and third potential c-Jun-binding sites.
As shown in Fig. 3G, the enrichment of AEG-1 and p300 binding within the VEGF promoter was significantly reduced when c-Jun was knocked down. Moreover, AEG-1 silencing reduced the DNA-binding affinities of both p300 and c-Jun to AEG-1-bound c-Jun target sites (Fig. 3H). Furthermore, our results showed a significant decrease in the levels of acetylated histones H3 and H4 in parallel with increased levels of histone deacetylase 1/3 (HDAC1/3) recruitment at the putative c-Jun-binding sites upon AEG-1 knockdown (Fig. 3I). Moreover, expression of c-Jun in AEG-1-RNAi cells could partially rescue the levels of acetylated histones H3 and H4 at the putative c-Jun-binding sites in the VEGF promoter (Fig. S2B). These results provide additional support for the hypothesis that AEG-1 plays an essential role in establishing an acetylated chromatin state that favors transcriptional activation.
We next determined whether AEG-1-mediated c-Jun acetylation contributes to the development of malignant properties in gliomas. As shown in Fig. 4A, human umbilical vein endothelial cells (HUVECs) cultured in conditioned medium (CM) transferred from AEG-1-overexpressing cells formed significantly larger numbers of capillary-like structures than the vector controls, whereas HUVECs cultured in CM derived from cultured AEG-1-knockdown cells formed fewer capillary-like structures, suggesting that AEG-1 provoked the angiogenesis-inducing ability of glioma cells. Moreover, we found that the proangiogenic effects of AEG-1 on glioma cells could be abrogated by overexpression of the dominant negative K3R-c-Jun mutant (Fig. 4B). Colony formation and 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assays showed that overexpression of AEG-1 promoted the propagation of glioma cells, whereas AEG-1 depletion inhibited the propagation of glioma cells (Fig. 4C and andE).E). Furthermore, the expression of the K3R-c-Jun mutant markedly inhibited the AEG-1-induced proliferation of glioma cells (Fig. 4D and andF).F). Taken together, these data obtained using glioma cells as a model reveal the effects of AEG-1-mediated c-Jun acetylation on tumor neovascularization and proliferation.
Additionally, in vivo animal experiments were performed to examine the effect of AEG-1-dependent c-Jun acetylation on tumor progression. Using a flank xenograft tumor model, we found that tumors formed by cells overexpressing AEG-1 grew faster than those formed by control cells, whereas cells with stable knockdown of AEG-1 formed relatively smaller tumors (Fig. 5A). Furthermore, the tumor-promoting activity of AEG-1 could be blocked by K3R-c-Jun expression (Fig. 5A). As shown in Fig. 5B, the tumors formed by AEG-1-overexpressing glioma cells in the brains of nude mice were far larger and displayed markedly increased levels of CD31 and higher levels of Ki67 signals compared with the ones formed by vector control cells. Moreover, the tumors formed by AEG-1-overexpressing glioma cells transfected with the K3R-c-Jun mutant exhibited a smaller size and reduced CD31 and Ki67 staining compared with those formed by AEG-1-overexpressing cells or AEG-1-overexpressing cells transfected with wild-type c-Jun. These data suggest an important role of c-Jun acetylation in mediating these AEG-1-induced biological effects.
Finally, we examined whether the experimentally observed AEG-1-mediated c-Jun activation is clinically relevant for human gliomas. Immunohistochemistry (IHC) analysis of 149 glioma clinical specimens revealed that the level of AEG-1 was significantly correlated with the expression of markers of glioma aggressiveness, such as those of the endothelial marker CD31 (P < 0.001) and the proliferative marker Ki67 (P < 0.001), as well as with the levels of expression of c-Jun downstream gene transcripts, including those of VEGF (P < 0.001), c-MET (P < 0.001), cyclin D1 (P < 0.001), and MMP-2 (P < 0.001) (Fig. 6A and andB).B). In addition, EMSA results showed that the level of AEG-1 expression was strongly associated with the level of c-Jun transcriptional activity (r = 0.762, P = 0.028) (Fig. 6C and andD).D). Taken together, these data demonstrate the importance of AEG-1 overexpression for the induction of the expression of c-Jun downstream genes and for the promotion of glioma aggressiveness.
To identify a c-Jun signaling pathway that was strongly associated with AEG-1 expression, gene set enrichment analysis (GSEA) was conducted. As shown in Fig. 7A, the GSEA results obtained using tumor samples from glioblastoma patients that were deposited in The Cancer Genome Atlas (TCGA) data set (n = 605) showed a marked enrichment in the levels of AP-1-activated gene expression in high-grade glioblastoma tumors expressing AEG-1 (enrichment score [ES] = 0.438; normalized enrichment score [NES] = 1.5; P = 0.019). The levels of c-Jun-activated gene signatures were markedly enriched in the breast cancer group expressing high levels of AEG-1, as analyzed using GSEA of the relevant data sets (Gene Expression Omnibus database accession number GSE20685; P = 0.014), and in the prostate cancer data sets (Gene Expression Omnibus database accession number GSE17951; P = 0.043) (Fig. 7B and andC).C). Together, these results further support the hypothesis that AEG-1 binds p300 and c-Jun and enhances the acetylation of c-Jun, leading to the activation of the AP-1 signaling pathway and the enhancement of transcription of AP-1 target genes (Fig. 7D).
As the transcriptional hub of multiple upstream oncogenic signals, c-Jun transcription factors have been found to be constitutively activated in many types of tumors (26,–28) and to participate in various processes that promote tumor development and progression, such as angiogenesis and cell survival (29,–32). The essential role of c-Jun in the progression of gliomas has been demonstrated in previous studies (33). Therefore, delineating the regulatory roles of c-Jun in cancer cells is of great importance. Posttranslational modifications of c-Jun, such as its phosphorylation and acetylation, are important regulatory factors involved in the transactivation of c-Jun (20,–22). Driven by the activation of oncoproteins such as epidermal growth factor receptor (34) and Jun N-terminal protein kinase (29, 31, 35) or by the inactivation of the tumor suppressor PTEN (36), the phosphorylation of c-Jun (mainly at Ser63/Ser73) increases in a transient, pulse-like fashion as the cell cycle progresses or in response to treatment with a mitogen or 12-O-tetradecanoylphorbol-13-acetate (37). In contrast, mutation of the K271 acetylation site of c-Jun and the flanking lysine residue results in the blockage of its transactivation activity even in the presence of mitogens (20), suggesting that the activation of c-Jun may require not only phosphorylation but also acetylation. In the current study, we found that overexpression of AEG-1 significantly increased the level of c-Jun acetylation and drove its transactivation activity. A c-Jun acetylation site mutant (K3R-c-Jun), which blocks acetylation-dependent AP-1 transactivation (20, 22), drastically diminished the stimulatory effect of AEG-1 on the transactivational activity of AP-1. These results provide new insights into the mechanisms underlying the oncogenic modulatory activity of c-Jun in cancer cells.
AEG-1 has been detected in various intracellular compartments and has been shown to possess pleiotropic properties. One important function of AEG-1 is to serve as a scaffold protein to mediate the formation of multiprotein complexes, thereby contributing to the regulation of diverse signaling pathways. In the cytoplasm, AEG-1 interacts with SND1, a component of the RNA-induced silencing complex (RISC), thus contributing to the onco-microRNA-mediated degradation of tumor suppressor mRNAs (10,–12). In the nucleus, AEG-1 functions as a transcriptional coactivator by directly interacting with the p65 subunit of NF-κB and by functioning as a transcriptional repressor through its interaction with the promyelocytic leukemia zinc finger, thereby increasing the level of c-Myc transcription (7,–9). These findings, together with the results obtained in the present study, suggest that AEG-1 is a key contributor to cancer development and progression by regulating the activity of several signaling cascades in different cellular compartments.
The results of this study demonstrate the essential role of AEG-1 in promoting c-Jun activity by interacting with p300 and c-Jun, which results in the subsequent acetylation of c-Jun. Notably, AEG-1 is a lysine-rich highly basic protein (15). Therefore, further investigation into whether its interaction with p300 results in the acetylation of AEG-1, which could lead to the occurrence of posttranslational modifications that modulate the biological effects of AEG-1, is of great interest.
In our current study, we identified AEG-1 to be a novel interacting partner of c-Jun and demonstrated that these interactions promote c-Jun activation. In a similar context, the activation of specific c-Jun complexes has been shown to be involved in the regulation of cellular death/survival in neurodegenerative diseases, possibly mediated through various signal transduction pathways, depending on different c-Jun-binding partners (38). Thus, the AEG-1/c-Jun interaction observed in glioma may also exist in neurodegenerative processes, implicating that AEG-1 may affect neurological integrity.
In summary, although previous publications have illustrated the multifunctional oncogenic characteristics of AEG-1, our results provide details of the mechanism involved in driving dysregulated c-Jun activity in human gliomas. We found that the interaction among AEG-1, p300, and c-Jun could play an important role in regulating c-Jun transcriptional activity, which could result in the development of a more aggressive phenotype of human glioma cells in vitro and in vivo. Our results suggest that a novel and promising therapeutic anticancer strategy could involve the targeting of AEG-1 to inhibit the aberrant function of acetyltransferases.
The U87MG and U251MG glioma cell lines were kindly provided by Shi-Yuan Cheng (University of Pittsburgh, Pittsburgh, PA) and were grown in Dulbecco's modified Eagle medium (DMEM; Gibco, Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum (HyClone, Logan, UT) at 37°C with 5% CO2. The authenticity of the cell lines was verified by short tandem repeat fingerprinting by the Medicine Laboratory of the Forensic Medicine Department of Sun Yat-sen University (Guangzhou, China).
A total of 149 paraffin-embedded, archived specimens in which glioma was clinically and histopathologically diagnosed at the First Affiliated Hospital of Sun Yat-sen University from 2000 to 2005 were used. The clinicopathological characteristics of the samples are summarized in Table 1. Fresh brain tumor tissue samples obtained from the First Affiliated Hospital of Sun Yat-sen University were collected and processed within 30 min after resection. Prior donor consent and study approval by the Institutional Research Ethics Committee of Zhongshan School of Medicine, Sun Yat-sen University, were obtained for our study.
Human AEG-1 was amplified by PCR and was cloned into the pMSCV-puro retroviral vector. Knockdown of endogenous AEG-1 was performed by cloning two short hairpin RNAs (shRNAs) using the following oligonucleotides: oligonucleotide 1 (AACAGAAGAAGAAGAACCGGA) and oligonucleotide 2 (GAAATCAAAGTCAGATGCTA) (synthesized at Invitrogen) into the pSuper-retro-puro vector to generate pSuper-retro-AEG-1-RNAis (AEG-1-RNAi#1 and AEG-1-RNAi#2, respectively). Fragments of the human AEG-1-coding sequence, which were generated by PCR amplification, were cloned into the pMSCV-puro retroviral vector. Different segments of the human VEGF promoter, including fragments covering nucleotides −200 to +690, −1140 to +690, −1700 to +690, and −2200 to +690 (numbered in relation to the transcription initiation site), were generated by PCR amplification using U87MG cell DNA as the template. Each of the products was cloned into the pGL3-basic luciferase reporter plasmid (Promega, Madison, WI). Mutations in the promoter segments were created using primers and a Stratagene mutagenesis kit according to the protocol recommended by the manufacturer. The sequences of all primers used for plasmids construction and site-specific mutagenesis are listed in the Tables 2 and and3,3, respectively. Small interfering RNA (siRNA) duplexes were synthesized and purified by Ribobio Inc. (Guangzhou, Guangdong, China). The sequences of the siRNAs used were as follows: for AEG-1 siRNA, 5′-AACAGAAGAAGAAGAACCGGA-3′; for c-Jun siRNA#1, 5′-CGCAGCAGTTGCAAACATT-3′; and for c-Jun siRNA#2, 5′-GACCTTATGGCTACAGTAA-3′. Transfection of the siRNAs was performed using the Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA) according to the instructions recommended by the manufacturer.
Western blotting (WB) was performed as previously described (39) using an anti-AEG-1 antibody (1:1,000; catalog number 40-6500; Thermo Fisher Scientific), an anti-p300 antibody (1:500; catalog number 05-257; Merck Millipore, Billerica, MA), an anti-c-Jun antibody (1:1,000; catalog number 9165; Cell Signaling, Danvers, MA), an anti-phospho-c-Jun antibody (Ser 63) (1:1,000; catalog number 9261; Cell Signaling), an anti-phospho-c-Jun (Ser73) antibody (1:1,000; catalog number 3270; Cell Signaling), an anti-acetyl lysine antibody (1:500; catalog number ab21623; Abcam, Cambridge, MA), an anti-HA tag antibody (1:5,000; catalog number H9658; Sigma, St. Louis, MO), an anti-Flag tag antibody (1:5,000; catalog number F3165; Sigma), and an anti-Myc tag antibody (1:5,000; catalog number ab9132; Abcam). After the initial Western blotting assay was performed, the membranes were stripped and were reprobed with an anti-α-tubulin antibody (1:1,000; catalog number T9026; Sigma) to determine the levels of the loading control.
Three thousand cells were seeded in triplicate in 48-well plates and were incubated for 24 h. One hundred nanograms of the luciferase reporter plasmid or the control luciferase plasmid plus 1 ng of the pRL-TK Renilla plasmid (Promega) were transfected into glioma cells using the Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA) according to the protocol recommended by the manufacturer. The luciferase and Renilla signals were measured 24 h after transfection using a dual-luciferase reporter assay kit (Promega) according to the manufacturer's instruction. Three independent experiments were performed. The data are presented as the mean values ± standard deviations (SDs).
Human umbilical vein endothelial cells (HUVECs) were purchased from ATCC (Manassas, VA) and were cultured according to the protocol recommended by the manufacturer. Matrigel (BD Biosciences, Bedford, MA) was pipetted into each well of a 24-well plate and was allowed to polymerize for 30 min at 37°C. HUVECs (2 × 104) in 200 μl of conditioned medium were added to each well, and the plate was incubated at 37°C with 5% CO2 for 20 h. Photographs were taken under a bright-field microscope (magnification, ×100), and the total lengths of the completed capillary tubular structures were measured. The effect of each experimental condition was assessed at least in triplicate.
The concentrations of VEGF in the cell-conditioned medium were determined using a commercially available VEGF-specific enzyme-linked immunosorbent assay (ELISA) kit (Keygen Co. Ltd., Nanjing, China). The ELISA was performed according to the protocol recommended by the manufacturer. Briefly, the conditioned medium was added to a well coated with a VEGF polyclonal antibody, and then a biotinylated monoclonal anti-human VEGF antibody was added and the plate was incubated at room temperature for 2 h. The color development catalyzed by horseradish peroxidase was terminated using 2.5 M sulfuric acid, and the absorption at 450 nm was measured. The protein concentration was determined by comparing the absorbance levels of the samples with those of the standards.
The level of p300-histone acetyltransferase (p300-HAT) activity was determined using an immunoprecipitation-HAT assay kit (catalog number 17-284; Merck Millipore). Briefly, endogenous p300 was precipitated using 5 μg of an anti-p300 antibody (Merck Millipore), and the immunoprecipitate was subsequently incubated with histone H4 peptide and 3H-labeled acetyl coenzyme A. The level of HAT activity was determined on the basis of the substrate radioactivity, which was measured using a liquid scintillation counter.
Cells (2 × 106) plated in a 100-mm culture dish were treated with 1% formaldehyde to cross-link the proteins to the DNA. The cell lysates were sonicated to shear the DNA into fragments of 300 bp to 1,000 bp. Equal aliquots of the chromatin supernatants were incubated with 1 μg of an anti-AEG-1 antibody (catalog number 40-6500; Thermo Fisher Scientific), an anti-c-Jun antibody (catalog number 9165; Cell Signaling), an anti-p300 antibody (catalog number 05-257; Merck Millipore), an anti-acetyl-histone H3 antibody (catalog number 17-615; Merck Millipore), an anti-acetyl-histone H4 antibody (catalog number 17-630; Merck Millipore), an anti-HDAC1 antibody (catalog number ab7028; Abcam), or an anti-HDAC3 antibody (catalog number ab7030; Abcam) overnight at 4°C with rocking. As a negative control, an anti-IgG antibody (catalog number ab172730; Abcam) was used to ensure antibody specificity. After reverse cross-linking of the protein/DNA complexes to free the DNA, PCR was performed using the following specific primers: for site 1, forward primer 5′-GTTCATCAGCCTAGAGCATG-3′ and reverse primer 5′-ATCATTCGTGCACTAGTCCT-3′; for site 2, forward primer 5′-CGAAACCCCCATTTCTATTCAG-3′ and reverse primer 5′-ACCCGCCAGCACTAAGGAA-3′; for site 3, forward primer 5′-ATGGAGCGAGCAGCGTCTT-3′ and reverse primer 5′-CCTTCTCCCCGCTCCAA-3′; and for site 4, forward primer 5′-AAGAGGTAGCAAGAGCTCCAGAGA-3′ and reverse primer 5′-GGCGGTCACCCCCAAA-3′.
Cells grown in 100-mm culture dishes were lysed using 500 μl of lysis buffer (25 mM HEPES [pH 7.4], 150 mM NaCl, 1% NP-40, 1 mM EDTA, 2% glycerol, 1 mM phenylmethylsulfonyl fluoride [PMSF]). After being maintained on ice for 30 min, the lysates were clarified by microcentrifugation at 12,000 rpm for 10 min. To preclear the supernatants, the lysates were incubated with 20 μl of agarose beads (Calbiochem, Cambridge, MA) for 1 h with rotation at 4°C. After centrifugation at 2,000 rpm for 1 min, the supernatants were incubated with 20 μl of antibody-cross-linked protein G-agarose beads overnight at 4°C. Then, the agarose beads were washed six times with wash buffer (25 mM HEPES [pH 7.4], 150 mM NaCl, 0.5% NP-40, 1 mM EDTA, 2% glycerol, 1 mM PMSF). After removing all the liquid, the pelleted beads were resuspended in 30 μl of 1 M glycine (pH 3), after which 10 μl of 4× sampling buffer was added, the samples were denatured, and the sample components were electrophoretically separated on SDS-polyacrylamide gels.
Electrophoretic mobility shift assay (EMSA) analyses were performed using a LightShift chemiluminescent EMSA kit obtained from Pierce Biotechnology (Rockford, IL). The following DNA probes containing specific binding sites were used: for c-Jun, sense probe 5′-CGCTTGATGACTCAGCCGG-3′ and antisense probe 5′-CCGGCTGAGTCATCAAGCG-3′, and for OCT-1, sense probe 5′-TGTCGAATGCAAATCACTAGAA-3′ and antisense probe 5′-TTCTAGTGATTTGCATTCGACA-3′.
Female BALB/c nude mice (age, 6 to 7 weeks; weight, 23 to 25 g) were purchased from the Experimental Animal Center of the Guangzhou University of Chinese Medicine (Guangzhou, China) and were randomly divided into four groups (n = 5 mice per group). The indicated cells (5 × 106) were suspended in 200 μl of sterile phosphate-buffered saline and were subcutaneously implanted into the flanks of the nude mice. Tumor size was measured with calipers every 6 days, and the volume was calculated as (width)2 × length × 0.52. At day 32 after implantation, all mice were sacrificed, and their tumors were excised. The usage of mice in this work was approved by the Animal Care Committee of Zhongshan School of Medicine, Sun Yat-sen University.
The glioma cells (5 × 105) indicated above were stereotactically implanted into the brains of individual mice (n = 5 mice per group). The mice were monitored daily and euthanized when they were moribund. Whole brains were removed, paraffin embedded, sectioned onto 5-μm-thick slides, and hematoxylin and eosin (H&E) stained or immunostained with the following antibodies: anti-CD31 (1:100; catalog number M0823; Dako, Carpinteria, CA) and anti-Ki67 (1:100; catalog number M7240; Dako). Images were captured using an AxioVision (release 4.6) computerized image analysis system (Carl Zeiss, Oberkochen, Germany). The proliferation index was quantified by determining the proportion of Ki67-positive cells by counting. Microvascular density was quantified by counting the CD31-positive cells. The usage of mice in this work was approved by the Animal Care Committee of Zhongshan School of Medicine, Sun Yat-sen University.
Immunohistochemistry (IHC) analysis of 149 paraffin-embedded, archived samples of glioma specimens was performed. Following deparaffinization, the sections were immunostained using antibodies directed against CD31 (1:100; catalog number M0823; Dako), Ki67 (1:100; catalog number M7240; Dako), VEGF-A (1:200; catalog number ab46154; Abcam), c-MET (1:200; catalog number 3189; Cell Signaling), cyclin D1 (1:200; catalog number ab134175; Abcam), or MMP-2 (1:200; catalog number ab37150; Abcam). The intensity of immunostaining was scored separately by two independent pathologists. The staining index (SI) values were calculated as follows: staining intensity score × proportion of positively stained tumor cells. The staining intensity and proportion of tumor cells positively stained for the proteins indicated above were determined in 10 randomly selected microscopic fields, and we calculated the average SI for the 10 selected fields. Cutoff values for high and low levels of expression of the proteins of interest were chosen to reflect the heterogeneity of the immunostaining results. The chi-square test was used to analyze the relationship between the level of AEG-1 expression and the levels of expression of the other genes. Images of the immunostained sections were captured using an AxioVision (release 4.6) computerized image analysis system (Carl Zeiss).
Microarray hybridization, data generation, and normalization were performed at the Shanghai Biochip Corporation following standard Agilent protocols. Bioinformatics analysis and visualization of the microarray data were performed using MeV (version 4.6) software (http://www.tm4.org/#/welcome).
GSEA was performed using the GSEA (version 2.0.9) platform (http://www.broadinstitute.org/gsea/). The samples were divided into groups with high levels of AEG-1 and groups with low levels of AEG-1 on the basis of the median level of AEG-1 expression. The c-Jun-activated genes included in our analysis were BDKRB1, BTK, CCND1, CCR7, CDH5, CHST1, DYRK1A, EDN1, GFAP, GLI2, HS3ST3B1, HSPB7, IGSF3, IL-9, KCNK10, LAPTM5, LOR, MCF2L, MET, MMP2, MMP9, MYC, PDE4D, PLA2G2E, PRDM13, SERPINE1, SOST, TFB1M, and VEGFA.
All statistical analyses were conducted using the SPSS (version 10.0) statistical software package. The relationship between the level of AEG-1 expression and the levels of expression of CD31, Ki67, VEGF-A, c-MET, cyclin D1, and MMP-2 determined by immunohistochemistry analysis were analyzed using the chi-square test. Spearman's correlation coefficients were calculated to determine the strength of the relationship between the expression levels of AEG-1 and the transcriptional activity levels of c-Jun. Mean values ± SDs were calculated, and two-tailed Student's t tests were performed to compare paired samples using the data analysis tools provided in the software package. In all cases, a P value of <0.05 was considered to indicate a significant difference.
This work was supported by the Natural Science Foundation of China (no. 81272339, 81330058, 91529301, and 81325013), Guangdong Natural Science Funds for Distinguished Young Scholars (2014A030306023), the Science and Technology of Guangdong Province (no. 2014A030313008), and the National Science and Technique Major Project (201305017).
We declare no competing financial interests.
Supplemental material for this article may be found at https://doi.org/10.1128/MCB.00456-16.