Glioblastoma multiforme is a common malignant brain tumor that portends extremely poor patient survival. Recent studies reveal that glioma stem-like cells (GSC) are responsible for glioblastoma multiforme escape from chemo-radiotherapy and mediators of tumor relapse. Previous studies suggest that AEG-1 (MTDH), an oncogene upregulated in most types of cancers, including glioblastoma multiforme, plays a focal role linking multiple signaling pathways in tumorigenesis. We now report a crucial role of AEG-1 in glioma stem cell biology. Primary glioblastoma multiforme cells were isolated from tumor specimens and cultured as neurospheres. Using the surface marker CD133, negative and positive cells were separated as nonstem and stem populations by cell sorting. Tissue samples and low passage cells were characterized and compared with normal controls. Functional biological assays were performed to measure stemness, self-renewal, differentiation, adhesion, protein–protein interactions, and cell signaling. AEG-1 was upregulated in all glioblastoma multiforme neurospheres compared with normal neural stem cells. Expression of AEG-1 was strongly associated with stem cell markers CD133 and SOX2. AEG-1 facilitated β-catenin translocation into the nucleus by forming a complex with LEF1 and β-catenin, subsequently activating Wnt signaling downstream genes. Through an AEG-1/Akt/GSK3β signaling axis, AEG-1 controlled phosphorylation levels of β-catenin that stabilized the protein.

Implications: This study discovers a previously unrecognized role of AEG-1 in GSC biology and supports the significance of this gene as a potential therapeutic target for glioblastoma multiforme. Mol Cancer Res; 15(2); 225–33. ©2016 AACR.

This article is featured in Highlights of This Issue, p. 115

Glioblastoma multiforme is the most prevalent and aggressive cancer of the central nervous system. Moreover, glioblastoma multiforme is highly proliferative, invasive, and vascular and is considered one of the most lethal human cancers. Despite the combination of surgery, chemotherapy, and radiotherapy, improvements in patient outcome have only improved modestly in the past decades, which highlight the dire need for an effective treatment of these aggressive tumors (1–3).

Cancer stem-like (CSC) cells represent a population of rare tumor cells characterized by their ability to self-renew and to induce tumorigenesis (4–6). Glioma-like stem cells (GSC) have been isolated from both human brain tumors and several glioma cell lines, and they play a seminal role in therapy resistance, immune evasion, and in mediating an increased invasive phenotype (7, 8). After surgical excision, the GSCs remaining in the brain can repopulate and reinitiate the pathogenic process. GSCs share a number of characteristics with normal neural stem cells, including the expression of neural stem cell markers, the capacity of self-renewal and long-term proliferation, the formation of neurospheres, and the ability to differentiate into multiple nervous system lineages. Recent studies have characterized and categorized GSCs into proneural and mesenchymal subtypes, which reflects the complexity and heterogeneous nature of glioblastoma multiforme (9, 10). Despite clinical significance, effective/selective targeting strategies for GSCs do not currently exist, and this may be due to the lack of comprehensive understanding of GSCs and the absence of new clinically effective targets (11, 12).

Of the various signaling pathways that have been implicated in tumor development and progression, the Wnt, Hedgehog, and Notch pathways have garnered considerable interest (13, 14). Wnt/β-catenin signaling was shown to play an important role in the maintenance of cancer stem cells (15). In addition to being associated with glioma growth or invasion, these pathways are necessary for the normal development of the nervous system and the maintenance of stem cells, which makes them potentially critical regulators of glioma initiation, self-renewal, and progression through different stages of malignancy (16, 17).

Astrocyte elevated gene (AEG)-1 (18), also known as metadherin (MTDH; ref. 19) and lysine-rich CEACAM1 co-isolated (LYRIC; ref. 20), was first cloned in our laboratory as an upregulated transcript in HIV-1–infected or gp-120- or TNFα-treated human astrocytes in 2002 (21) and has been found to be one of the key oncogenes in almost every cancer type (22–24). AEG-1 plays an essential role in glioblastoma multiforme pathogenesis (25, 26), and it is expressed at very high levels in high-grade gliomas and promotes brain tumor growth and invasion. In one recent study, we discovered a novel specific interaction between AEG-1 and Akt2 in glioma (27). More recently, AEG-1 has also been found to be upregulated in CD133+ tumor cells (28). AEG-1 has been shown to interact with β-catenin in colorectal carcinoma (29), and a positive correlation between AEG-1 and β-catenin nuclear expression in colorectal carcinoma was also reported. However, the detailed biological function of the AEG-1–β-catenin interaction is still ambiguous.

Given the unique expression and protumorigenic role of AEG-1 in gliomas, we investigated the function and molecular mechanisms of action of this protein, in GSC biology. Here, we report that AEG-1 acts through Akt-GSK3β as a key regulator of β-catenin/Wnt signaling to control the stemness, differentiation, resistance to apoptosis, as well as adhesion of GSCs. Our results are the first demonstration that the AEG-1 oncogene has the ability to activate the Wnt pathway in glioma stem cells to promote stemness and tumor progression.

Primary glioblastoma multiforme cell cultures

Glioblastoma samples were obtained from subjects who underwent surgical removal of their brain tumors. Informed consent was obtained according to the research proposals approved by the Institutional Review Board at the VCU Tissue & Data Acquisition & Analysis Core (TDAAC). Tissues were dissociated with trypsin (Invitrogen), hyaluronidase (Sigma), collagenase (Sigma), and DNase I (Sigma) mixture. Enzyme reaction was stopped by trypsin inhibitor (Sigma) and washed several times with 1× PBS. Digested samples were filtered with 70-μm nylon cell strainer (BD Biosciences) and resuspended in stem cell medium composed of DMEM/F-12 50:50 containing K27 supplements, glutamine 2 μmol/L (Invitrogen), and basic fibroblast and epidermal growth factors (PeproTech, 20 ng/mL each) for continuous culturing (27). Floating neurospheres were amplified and stored for further experiments. All primary cells were cultured as suspended spheres in uncoated T25 or T75 culture dishes (Corning) in Essential 8 medium (Invitrogen) and analyzed prior to five passages. All primary tumor cells were authenticated using the “CellCheck” service provided by the Research Animal Diagnostic Laboratory IDEXX Bioresearch.

Cell sorting and FACS

Neurospheres were disassociated with Accutase (Invitrogen) and labeled with CD133 (AC133)-PE conjugated antibody (Miltenyi Biotec). Stained cells were sorted through a BD Aria II sorting station. Antibody-negative and -positive cell populations were counted and collected for further culturing. GSC CD133+ cells were validated by cell self-renewal, differentiation, and tumorigenicity assays. FACS of CD44 and CD133 double staining in glioblastoma multiforme cells were performed on BD Caliber flow cytometry.

Monolayer primary glioblastoma multiforme cell growth

Plastic dishes or glass slides were precoated with CELLstart (Invitrogen) in PBS for one hour in 37°C humidified incubator. Neurospheres were disassociated with Accutase (Invitrogen) and counted with cellometer Auto T4 (Nexcelom). Seeded cells were grown as monolayer in glioma stem cell medium.

DNA constructs, siRNAs, adenovirus, and lentivirus

A clone containing the full-length coding sequence of AEG-1 driven by CMV promoter was cloned in pcDNA3.1 vector. Adenovirus containing AEG-1 FL and lentivirus shRNA against AEG-1 were described in previous studies (27). siRNA oligonucleotides against AEG-1 (MTDH) were purchased from Qiagen and validated at mRNA and protein levels.

IHC and immunofluorescence

Human tumor tissue sections (5 μm) were processed for IHC with antibodies against AEG-1 (α-chicken) and β-catenin (Cell Signaling Technology). Images (20×) were taken with a Nikon microscope. Correlations between AEG-1 and β-catenin staining were analyzed with Prism 5 software. GSC spheres were disassociated with Accutase and attached to CELLstart (Invitrogen) precoated glass slides. These cells were fixed with 4% PFA in PBS, immunostained with primary and Alexa-conjugated secondary antibodies, followed by imaging by laser confocal microscopy (Leica).

TCGA data analysis

AEG-1 (MTDH) and β-catenin (CTNNB1) RNA sequencing (RNA-seq) data were analyzed using an online tool from NCBO BioPortal.

Cell fractionation

Primary glioblastoma multiforme cells were fractioned with cell fractionation kit (Qiagen). β-Tubulin and HDAC3 were used as biomarkers for cytoplasm and nuclear fractions, respectively.

Biochemical assays

Cultured cells were lysed and processed for Western blotting using standard protocols (antibodies are listed in Supplementary Data; refs. 26, 27). For qRT-PCR, cells or tissue samples were processed using RNeasy Kit (Qiagen) and resuspended in RNase-Free Water. RNA concentration was measured with NanoDrop (Thermo). Real-time PCR was performed with Applied Biosystems SYBR Green PCR Master Mix Kit. Primers for AEG-1 and Wnt target genes synthesized at VCU oligo nucleotide facility core, AEG-1 (MTDH): forward: 5′-CCAGGCTCCTTCATCAACTT-3′, reverse: 5′-AAAGCAGCCACCAGAGATTG-3′; CD133 forward: 5′-ACCAGGTAAGAACCCGGATC-3′, reverse: 5′-CAAGAATTCCGCCTCCTAGC-3′; SOX2 forward: 5′-GCTTAGCCTCGTCGATGAAC-3′, reverse: 5′-AACCCCAAGATGCACAACT-3′; GFAP forward: 5′-CACCACGATGTTCCTCTTGA-3, reverse: 5′-GTGCAGACCTTCTCCAACCT-3′; AXIN2 forward: 5′-CTGGTGCAAAGACATAGCCA-3′, reverse: 5′-AGTGTGAGGTCCACGGAAAC-3′; CTNNB1 forward: 5′-CGCTGGATTTTCAAAACAGT-3′, reverse: 5′-CTGAGGAGCAGCTTCAGTCC-3′.

Co-Immunoprecipitation assay

Nuclear extracts were prepared from VG2 and VG9 GSCs using Cell Fractionation Kit (Qiagen). Pierce Protein G conjugated Agarose (Thermo Scientific) was used in co-immunoprecipitation (Co-IP) experiments. Corresponding mouse or rabbit IgG (Jackson ImmunoResearch) were used as controls. Primary antibodies used were AEG-1 rabbit polyclonal antibody (26, 27), β-catenin antibody #9562, and LEF1 (C12A5) rabbit mAb #2230 (Cell Signaling Technology). Horseradish peroxidase–conjugated secondary antibody with light chain IgG (mouse/rabbit) was used to eliminate potential overlapping heavy chain background in all IP experiments.

Adhesion assays

Cell adhesion assays were performed as described previously (27). Briefly, dissociated stained GSCs were counted and seeded on cultured human brain microvascular endothelial cell (HBMEC) monolayer. Floating cells were washed off with PBS. Fluorescence microscopy was used to capture attached cells. Images were quantified with ImageJ software.

Statistical analysis

GraphPad Prism Version 5.00 for Windows (GraphPad Software Inc.) was used to perform a one-way ANOVA with Newman–Keuls posttest or a paired two-way Student t test as described previously (27). P values of <0.05 were considered significant. Results were analyzed by one-factor or multifactorial ANOVA, followed by the Bonferroni post hoc test.

High-level AEG-1 expression associates with high-level expression of stem cell makers in CD133+ tumor cell populations

In collaboration with VCU TDAAC, we successfully obtained and maintained multiple primary glioma tumor cells with corresponding pathologic reports providing tumor grade information. Higher levels of AEG-1 were evident in both primary glioblastoma multiforme cell cultures VG2, VG4, VG6, VG9, and their original tissue samples as compared with normal neural stem cell H9NSC and normal brain tissue lysates (Fig. 1A). AEG-1 level was also highly upregulated in many primary glioblastoma multiforme neurospheres obtained from external collaborators (Drs. Chiocca and Nakano, Department of Neurosurgery, Ohio State University, Columbus, OH) as compared with normal neural stem cells (H9NSC; Supplementary Fig. S1). After cell sorting with the CD133 cell surface marker, the majority of AEG-1 protein was associated with the CD133+ cell population in primary glioblastoma multiforme neurospheres VG2 and VG9 (Fig. 1B; Supplementary Fig. S2). To maintain the stemness of cells, cell cultures were used within five passages of being established as suggested in the literature. Most strikingly, overexpression of AEG-1 using adenovirus transduction in CD133 glioma cells enhanced the expression of glioma stem cell markers, including CD133, SOX2, and tumor mesenchymal marker CD44, which were detected at both protein and mRNA levels (Fig. 1C and D). CD133+ cells exhibited multilineage differentiation potential when cultured under astrocyte and neuronal medium conditions (Supplementary Fig. S3).

Figure 1.

AEG-1 is upregulated in neurospheres and associated with a high level of stem cell makers in CD133+ tumor cell populations. A, AEG-1 is upregulated in VG neurospheres derived from glioma tumor samples compared with normal neural stem cell H9NSC. B, CD133+ populations of primary glioblastoma multiforme VG cells have higher levels of AEG-1 expression than CD133 negative cells. C, Overexpression of AEG-1 in CD133 VG2 and VG9 cells increases the expression of stem cell markers, CD133 and SOX2, as well as the glioblastoma multiforme mesenchymal stem marker, CD44. D, Upregulation of Wnt/β-catenin target genes measured by qRT-PCR in CD133 populations of VG2 and VG9 cells overexpressing AEG-1. OE, overexpression. E, AEG-1 shRNA KD reduces VG2 and VG9 neurosphere self-renewal ability. After two passages, the reduction of self-renewal becomes more dramatic. *,P < 0.05, and ***, P < 0.001 with mean ± SD are shown in the figures.

Figure 1.

AEG-1 is upregulated in neurospheres and associated with a high level of stem cell makers in CD133+ tumor cell populations. A, AEG-1 is upregulated in VG neurospheres derived from glioma tumor samples compared with normal neural stem cell H9NSC. B, CD133+ populations of primary glioblastoma multiforme VG cells have higher levels of AEG-1 expression than CD133 negative cells. C, Overexpression of AEG-1 in CD133 VG2 and VG9 cells increases the expression of stem cell markers, CD133 and SOX2, as well as the glioblastoma multiforme mesenchymal stem marker, CD44. D, Upregulation of Wnt/β-catenin target genes measured by qRT-PCR in CD133 populations of VG2 and VG9 cells overexpressing AEG-1. OE, overexpression. E, AEG-1 shRNA KD reduces VG2 and VG9 neurosphere self-renewal ability. After two passages, the reduction of self-renewal becomes more dramatic. *,P < 0.05, and ***, P < 0.001 with mean ± SD are shown in the figures.

Close modal

AEG-1 knockdown inhibits primary neurosphere self-renewal

Expression of AEG-1 was knocked down (KD) under stem cell culture conditions in both VG2 and VG9 primary glioblastoma multiforme neurospheres using adenoviral shRNA targeting AEG-1. AEG-1 KD neurospheres showed significant loss of self-renewal ability compared with control shRNA group during two passages (Fig. 1E). These observations support the hypothesis that high-level expression of AEG-1 may contribute to GSC self-renewal.

AEG-1 expression correlates with β-catenin and activates its downstream signaling pathways

Analysis of TCGA RNA-seq data using NCBO BioPortal online analytic tool indicated a strong positive correlation between AEG-1 (MTDH) and β-catenin (CTNNB1) with Spearman correlation coefficient r = 0.67 (Fig. 2A). The positive correlation of AEG-1 and β-catenin was established at a protein level by analysis of tumor samples obtained at VCU Cancer Center (Richmond, VA; Fig. 2B; Supplementary Table S1). These results suggest that AEG-1 may associate with β-catenin/Wnt signaling in glioblastoma multiforme. Further gain-of-function and loss-of-function experiments using overexpression or KD of AEG-1 in CD133+ GSCs from VG2 and VG9 confirmed that phosphorylation of β-catenin (nonfunctional form) is negatively regulated by AEG-1 through the GSK3β signal cascade (Fig. 2C). Moreover, CD133+ VG2 and VG9 GSCs show higher levels of β-catenin than CD133 cells with a modest decrease in p-β-catenin levels (Fig. 2D). This suggests that higher levels of AEG-1 in CD133+ GSCs play a protective role for β-catenin. A similar trend was observed in the regulation of target genes of the Wnt/β-catenin pathway, such as AXIN2, CD44, and Myc, when AEG-1 expression levels were modified in GSCs (Fig. 2E).

Figure 2.

AEG-1 expression correlates with β-catenin expression and activates its downstream signaling pathways. A, Analysis of TCGA RNA-seq data (cBioportal) confirms a strong correlation between expression of AEG-1 (MTDH) and β-catenin (CTNNB1), Spearman correlation r = 0.67. B, Protein expression profiles of clinical human brain tumor samples obtained from the Brain Tumor Tissue Bank of VCU Medical Center. C, AEG-1 regulates the level of p-β-catenin through GSK3β signaling in VG2 and VG9 glioblastoma multiforme neurospheres. AOE, AEG-1 overexpression; Asi, AEG-1 siRNA. D, CD133+ VG2 and VG9 cells express lower levels of p-β-catenin than CD133 cell populations. E, mRNA expression of Wnt/β-catenin target genes, including AXIN2, CD44, and MYC, along with stem cell markers CD133 and SOX2 following overexpression or KD of AEG-1. OE, overexpression. ***,P < 0.001 with mean ± SD are shown in the figures.

Figure 2.

AEG-1 expression correlates with β-catenin expression and activates its downstream signaling pathways. A, Analysis of TCGA RNA-seq data (cBioportal) confirms a strong correlation between expression of AEG-1 (MTDH) and β-catenin (CTNNB1), Spearman correlation r = 0.67. B, Protein expression profiles of clinical human brain tumor samples obtained from the Brain Tumor Tissue Bank of VCU Medical Center. C, AEG-1 regulates the level of p-β-catenin through GSK3β signaling in VG2 and VG9 glioblastoma multiforme neurospheres. AOE, AEG-1 overexpression; Asi, AEG-1 siRNA. D, CD133+ VG2 and VG9 cells express lower levels of p-β-catenin than CD133 cell populations. E, mRNA expression of Wnt/β-catenin target genes, including AXIN2, CD44, and MYC, along with stem cell markers CD133 and SOX2 following overexpression or KD of AEG-1. OE, overexpression. ***,P < 0.001 with mean ± SD are shown in the figures.

Close modal

AEG-1 activates Wnt signaling by β-catenin translocation into the nucleus, and AEG-1 levels affect differentiation, self-renewal, and adhesion of GSCs

When Wnt signaling is activated by upstream signals, β-catenin proteins translocate from cytoplasm to the nucleus and form a complex via binding with other transcriptional proteins to activate specific target genes. Overexpression of AEG-1 resulted in translocation of β-catenin into the nucleus of GSCs, and when AEG-1 was knocked down, the majority of β-catenin remained in the cytoplasm as evidenced by confocal laser microscopy imaging (Fig. 3A). These results suggest that AEG-1 may control β-catenin localization. Similar results were obtained by Western blot analysis after subcellular fractionation. More β-catenin accumulates in the nucleus when AEG-1 is overexpressed, and more β-catenin is localized in the cytoplasm when AEG-1 is knocked down (Fig. 3B). In VG2 and VG9 tumor samples, a strong overlapping localization between AEG-1 and β-catenin was observed in tumor regions (Fig. 3C).

Figure 3.

AEG-1 regulates β-catenin translocation into the nucleus. A, Immunofluorescence microscopy shows translocalization of β-catenin (green) when overexpressed or after KD of AEG-1 (red) in VG2 GSCs. B, Cell fractionation of VG2 GSCs demonstrates β-catenin localization after AEG-1 overexpression or KD. β-Tubulin and HDAC3 were used as cytoplasm and nuclear markers, respectively. OE, overexpression. C, Confocal images of immunofluorescence staining of AEG-1 (red) and β-catenin (green) show that both molecules are highly expressed in tumor regions of VG2 and VG9 tumor sections (10-μm thickness).

Figure 3.

AEG-1 regulates β-catenin translocation into the nucleus. A, Immunofluorescence microscopy shows translocalization of β-catenin (green) when overexpressed or after KD of AEG-1 (red) in VG2 GSCs. B, Cell fractionation of VG2 GSCs demonstrates β-catenin localization after AEG-1 overexpression or KD. β-Tubulin and HDAC3 were used as cytoplasm and nuclear markers, respectively. OE, overexpression. C, Confocal images of immunofluorescence staining of AEG-1 (red) and β-catenin (green) show that both molecules are highly expressed in tumor regions of VG2 and VG9 tumor sections (10-μm thickness).

Close modal

KD of AEG-1 with specific lentiviral shRNAs only reduced self-renewal in CD133+ VG2 cells and not in CD133 cells (Fig. 4A). This confirmed that AEG-1 contributes to neurosphere self-renewal only in CD133+ GSCs. Previous studies by other groups demonstrated that the microvascular endothelial cell surface serves as an anchor for the GSC niche (30–32). We tested the ability of GSCs to adhere to HBMEC monolayers. KD of AEG-1 in VG2 CD133+ cells greatly reduced adhesion to HBMEC, whereas AEG-1 KD slightly increased adherence to HBMEC in CD133 cells (Fig. 4B).

Figure 4.

AEG-1 regulates GSC self-renewal, adhesion, and lineage differentiation. A, AEG-1 KD only affects self-renewal of CD133+ VG2 cells. B, AEG-1 KD significantly inhibits VG2 CD133+ cell adherence to HBMEC monolayer cultures. C, AEG-1 (red) regulates the astrocyte marker GFAP expression in GSCs, which is demonstrated by immunofluorescence staining of GFAP (green) in VG2 GSCs. D, Neural marker MAP2 (green) is regulated by AEG-1 (red) overexpression or KD in GSCs. Neural dendritic branches are detected by MAP2 immunofluorescence staining. Top, control; middle, AEG-1 overexpression; bottom, AEG-1 KD. **,P < 0.01 with mean ± SD are shown in the figures.

Figure 4.

AEG-1 regulates GSC self-renewal, adhesion, and lineage differentiation. A, AEG-1 KD only affects self-renewal of CD133+ VG2 cells. B, AEG-1 KD significantly inhibits VG2 CD133+ cell adherence to HBMEC monolayer cultures. C, AEG-1 (red) regulates the astrocyte marker GFAP expression in GSCs, which is demonstrated by immunofluorescence staining of GFAP (green) in VG2 GSCs. D, Neural marker MAP2 (green) is regulated by AEG-1 (red) overexpression or KD in GSCs. Neural dendritic branches are detected by MAP2 immunofluorescence staining. Top, control; middle, AEG-1 overexpression; bottom, AEG-1 KD. **,P < 0.01 with mean ± SD are shown in the figures.

Close modal

To investigate AEG-1′s role in GSC differentiation, we cultured the VG2 and VG9 cells in monolayer culture under different inducing conditions for astrocytes or neurons. The differentiation ability of these cells was analyzed by an immunofluorescence assay with antibodies against AEG-1, GFAP, and MAP2. As shown in Fig. 4C and D, overexpression of AEG-1 decreased the intensities of the astrocyte maker GFAP and neuronal maker MAP2; on the other hand, AEG-1 KD increased the intensities of staining when compared with control cells. Under neural culture condition, AEG-1 KD increased the number of dendritic branches and their lengths in comparison with control and AEG-1 overexpression (Fig. 4D).

AEG-1–Akt signaling involves regulation of phosphorylated β-catenin

To further confirm our finding that AEG-1 regulates p-β-catenin in GSCs, we overexpressed AEG-1 in primary astrocytes isolated from AEG-1 KO pups. The result from immunoblotting showed that overexpression of AEG-1 increased the levels of p-GSK3β and decreased phosphorylation of β-catenin (Fig. 5A). In addition, the downstream genes of Wnt/β-catenin were upregulated following overexpression of AEG-1 (Fig. 5B). These results confirm the direct regulation of AEG-1 on GSK3β–β-catenin signaling and its downstream effector molecules. It would be of interest to comapre the expression levels of GSK3β–β-catenin signaling genes in the astrocytes/neuronal cells derived from AEG-1−/− mice and their wild-type (WT) littermates, which would permit analysis of the effects of the absence of AEG-1 on specific signaling cascades.

Figure 5.

AEG-1 regulates β-catenin through GSK-3β and also forms a protein complex in nuclei. A, Immunoblots show that overexpression of AEG-1 upgreulates the levels of p-GSK3β and downregulates the levels of p-β-catenin in astrocytes isolated from AEG-1 KO mice. B, AEG-1 OE upregulates mRNA levels of Wnt/β-catenin signaling target genes in AEG-1 KO astrocytes. KO, knock-out; OE, overexpression. C, Co-IP indicates that AEG-1, β-catenin, and LEF-1 form a protein complex in GSCs. D, AEG-1 increases the proportion of CD133 and CD44 double-positive cell populations in both CD133-negative and -positive cells, as determined by FACS analysis. AOE, AEG-1 overexpression. *,P < 0.05, and **,P < 0.01 with mean ± SD are shown in the figures.

Figure 5.

AEG-1 regulates β-catenin through GSK-3β and also forms a protein complex in nuclei. A, Immunoblots show that overexpression of AEG-1 upgreulates the levels of p-GSK3β and downregulates the levels of p-β-catenin in astrocytes isolated from AEG-1 KO mice. B, AEG-1 OE upregulates mRNA levels of Wnt/β-catenin signaling target genes in AEG-1 KO astrocytes. KO, knock-out; OE, overexpression. C, Co-IP indicates that AEG-1, β-catenin, and LEF-1 form a protein complex in GSCs. D, AEG-1 increases the proportion of CD133 and CD44 double-positive cell populations in both CD133-negative and -positive cells, as determined by FACS analysis. AOE, AEG-1 overexpression. *,P < 0.05, and **,P < 0.01 with mean ± SD are shown in the figures.

Close modal

AEG-1, LEF-1, and β-catenin form a protein complex in the nucleus

An interaction between AEG-1 and β-catenin was previously reported in colon cancer (29). To confirm this interaction in the context of GSCs, we performed Co-IP analysis and found that AEG-1, β-catenin, and LEF-1 form a protein complex in VG2 and VG9 GSC cell nuclei (Fig. 5C). This physical interaction between AEG-1 and β-catenin may prevent its degradation by proteasomes and facilitate β-catenin accumulation in the nuclei of GSCs.

High levels of AEG-1 in CD133+ GSCs associate with mesenchymal marker CD44

Recent studies showed that CD133+ GSCs consist of several subtypes, including proneural and mesenchymal (9). AEG-1 was found to be upregulated in the mesenchymal subtype of glioblastoma multiforme (unpublished data). CD44 is established as a mesenchymal cell marker as well as a stem cell marker in several cancers (33). Using FACS, we measured CD44 and CD133 double-positive cell populations in primary glioblastoma multiforme neurospheres after overexpression or KD of AEG-1. As shown in Fig. 5D, overexpression of AEG-1 increased CD133+ populations in CD133 cells. Cells transfected with pcDNA-AEG-1 show approximately 5- to 30-fold (depending on the cell type) enhanced expression of AEG-1 as compared with the parental cells (see Figs. 1, 2, and 5). The level of expression usually remains comparable over the assay period. More importantly, CD133-CD44 double-positive populations increased in CD133+ GSCs, and KD of AEG-1 noticeably decreased these double-positive populations (Fig. 5D). This supports the hypothesis that AEG-1 regulates stemness and aggressiveness of GSCs.

The concept of cancer stem cells as major contributors to cancer aggression, resistance to conventional therapies, and disease recurrence has gained acceptance and is an area of intense investigation (6, 7, 11, 17, 34). GSCs are considered essential components in glioma initiation and progression and display inherent resistance to conventional therapies and thus facilitate tumor recurrence. GSCs share many similarities with neural stem cells in biomarker expression, self-renewal, and multiple-lineage differentiation potential, but also distinguish themselves from other stem cells by their highly invasive nature, immune evasion, and chemo-radiotherapy resistance (5, 17, 35). We successfully isolated and cultured primary brain tumor cells and GSCs from patient tissue samples using established methodologies (27, 36, 37). CD133 is a well-recognized and functional GSC marker, although its biological relevance is still poorly defined and its role as a sole biomarker of GSC is controversial (7, 38). We detected a small population of CD133+ cells in the total tumor cell population, and this ratio was different from sample to sample as reported in previous publications (36, 37). Tumor cells positive for CD133 showed high levels of neural stem cell markers, including SOX2, Oct4, and Nestin (39). CD133+ cells were able to maintain self-renewal, differentiate into astrocytes and neurons, and generate tumors when injected into the brains of nude mice (Supplementary Figs. S2 and S3; ref. 7), supporting the suggestion that these CD133+ cells isolated from glioblastoma multiforme samples are GSCs. AEG-1 expression positively correlated with stem cell makers in CD133+ GSC populations. Gain of AEG-1 expression in CD133 glioma cells enhanced the expression of glioma stem cell markers (CD133 and SOX2) and the tumor mesenchymal marker, CD44. Interestingly, KD of AEG-1 in VG2 and VG9 neurospheres (mixed populations of CD133 and CD133+; Fig. 1E) or in CD133+ GSCs (Fig. 4A) resulted in decreased GSC self-renewal. CD133 is one of the markers, but not necessarily the sole marker, associated with self-renewal. It would be interesting to determine the impact of AEG-1 overexpression in CD133 cells on self-renewal ability, which is something worth pursuing in future studies. One recent study suggests that AEG-1 is continuously required for the full functionality of tumor-initiating cells in mammary tumors (40). In another recent study, the authors found that AEG-1 drives breast cancer stem cell expansion by promoting the expression of TWIST1, a transcription factor critical for cancer cell stemness and metastasis (41). In line with these observations, we now demonstrate a novel and important role of AEG-1 as a regulator of stemness and self-renewal of glioma stem cells.

Our results from bioinformatics and biochemical characterization of tumor samples showed a positive correlation between AEG-1 and β-catenin expression in glioblastoma multiforme. Previous studies showed that canonical Wnt/β-catenin signaling was involved in regulation of cancer stem cell stemness and resistance to chemo-radiotherapy (42, 43). AEG-1 was able to regulate phosphorylation levels of β-catenin and its transportation into the nucleus. Phosphorylation of β-catenin directs it toward protein ubiquitination, followed by proteasome degradation. Recently, we reported that AEG-1 specifically interacts with Akt2 in glioblastoma multiforme and regulates GSK3β signaling (27). Higher levels of AEG-1 increased p-GSK3β though the Akt pathway. This process results in lower GSK3β kinase activity, and β-catenin is a direct substrate of GSK3β kinase (44). LEF-1 has been shown to interact with β-catenin and function as a transcriptional coactivator for Wnt target genes (45). To confirm previous reports in colorectal cancer that AEG-1 could directly interact with β-catenin (29), we performed Co-IP analyses and found that AEG-1 not only interacted with β-catenin, but it also formed a protein complex with LEF1 to possibly function as a transcription activator for Wnt signaling downstream genes. Further investigations are necessary to define the contribution of this interaction complex between these three proteins in GSCs and other cancer stem cells. Our results suggest that AEG-1 may promote β-catenin protein stability in GSCs through AEG-1/Akt2–GSK3β. Also, the physical interaction between AEG-1 and β-catenin may protect its exposure to the proteasome and facilitate β-catenin accumulation in the nuclei of GSCs. In total, these results imply that AGE-1 regulates β-catenin–mediated GSC stemness, differentiation, and self-renewal.

High levels of AEG-1 in CD133+ cells also resulted in highly activated Wnt/β-catenin signaling with upregulated target genes, including AXIN2, CD44, and Myc. CD44 is one of the mesenchymal stem cell markers and plays an important role in cell adhesion and invasion (10, 46–48). Our results also demonstrated that AEG-1 KD strongly reduced GSC adhesion to monolayer HBMEC cells. It is possible that CD44 is one of the facilitators that regulate AEG-1/β-catenin–mediated glioma stem cell adhesion (49). Recently, two subtypes of GSCs were identified and characterized: proneural and mesenchymal (10). CD133 regulated the proneural GSC subtype, whereas CD44 regulated the mesenchymal GSC subtype (7, 10). Our results indicated that AEG-1 was capable of regulating both CD133 and CD44 double-positive cell populations, thereby suggesting a potential role of AEG-1 in both GSC subtypes.

Overall, we report that higher levels of AEG-1 play a seminal role in GSC biology to maintain stemness, self-renewal, and mesenchymal characteristics. AEG-1 maintains lower levels of p-β-catenin to improve its protein stability though the Akt2–GSK3β signal cascade. Thus, AEG-1–Akt2–GSK3β pathway in general promotes glioblastoma multiforme tumor cell survival, invasion, and proliferation (27) and also contributes to GSC maintenance via Wnt/β-catenin signaling (Fig. 6). AEG-1 as one of the key driver oncogenes upregulated in almost every type of cancer and serves as a major focal point in connecting multiple critical cancer signaling pathways (22). Our novel results confirming that AEG-1 provides important functions in glioma stem cells provide new insights into the role of AEG-1 in tumor initiation and progression, reinforcing the relevance of AEG-1 as a viable therapeutic target for glioblastoma multiforme and potentially other cancers.

Figure 6.

Schematic representation of the role of AEG-1 in GSC biology.

Figure 6.

Schematic representation of the role of AEG-1 in GSC biology.

Close modal

No potential conflicts of interest were disclosed.

Conception and design: B. Hu, L. Emdad, P.B. Fisher

Development of methodology: B. Hu, T.P. Kegelman

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): B. Hu, T.P. Kegelman, P.B. Fisher

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): B. Hu, L. Emdad, S.K. Das, P.B. Fisher

Writing, review, and/or revision of the manuscript: B. Hu, L. Emdad, D. Sarkar, P.B. Fisher

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L. Emdad, X.-N. Shen, S.K. Das

Study supervision: L. Emdad, S.K. Das, P.B. Fisher

We thank Mohammad A. Alzubi for excellent technical assistance. Human glioma samples were also provided through the Department of Pathology and VCU Massey Cancer Center of VCU, School of Medicine.

This work was supported in part by NIH/NCI grant R01 CA134721, National Foundation for Cancer Research (NFCR) and the VCU Massey Cancer Center (to P.B. Fisher) and NIH/NCI grant R01 CA138540 and the James S. McDonnell Foundation (to D. Sarkar). The VCU TDAAC Facility provided human glioma samples, which were supported in part through funding from NIH/NCI grant P30 CA016059.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Reardon
DA
,
Rich
JN
,
Friedman
HS
,
Bigner
DD
. 
Recent advances in the treatment of malignant astrocytoma
.
J Clin Oncol
2006
;
24
:
1253
65
.
2.
Dolecek
TA
,
Propp
JM
,
Stroup
NE
,
Kruchko
C
. 
CBTRUS statistical report: primary brain and central nervous system tumors diagnosed in the United States in 2005–2009
.
Neuro Oncol
2012
;
14
:
v1
49
.
3.
Chaudhry
IH
,
O'Donovan
DG
,
Brenchley
PE
,
Reid
H
,
Roberts
IS
. 
Vascular endothelial growth factor expression correlates with tumour grade and vascularity in gliomas
.
Histopathology
2001
;
39
:
409
15
.
4.
Lapidot
T
,
Sirard
C
,
Vormoor
J
,
Murdoch
B
,
Hoang
T
,
Caceres-Cortes
J
, et al
A cell initiating human acute myeloid leukaemia after transplantation into SCID mice
.
Nature
1994
;
367
:
645
8
.
5.
Rapp
UR
,
Ceteci
F
,
Schreck
R
. 
Oncogene-induced plasticity and cancer stem cells
.
Cell Cycle
2008
;
7
:
45
51
.
6.
Visvader
JE
,
Lindeman
GJ
. 
Cancer stem cells in solid tumours: accumulating evidence and unresolved questions
.
Nat Rev Cancer
2008
;
8
:
755
68
.
7.
Singh
SK
,
Hawkins
C
,
Clarke
ID
,
Squire
JA
,
Bayani
J
,
Hide
T
, et al
Identification of human brain tumour initiating cells
.
Nature
2004
;
432
:
396
401
.
8.
Bao
S
,
Wu
Q
,
McLendon
RE
,
Hao
Y
,
Shi
Q
,
Hjelmeland
AB
, et al
Glioma stem cells promote radioresistance by preferential activation of the DNA damage response
.
Nature
2006
;
444
:
756
60
.
9.
Nakano
I
. 
Stem cell signature in glioblastoma: therapeutic development for a moving target
.
J Neurosurg
2015
;
122
:
324
30
.
10.
Mao
P
,
Joshi
K
,
Li
J
,
Kim
SH
,
Li
P
,
Santana-Santos
L
, et al
Mesenchymal glioma stem cells are maintained by activated glycolytic metabolism involving aldehyde dehydrogenase 1A3
.
Proc Natl Acad Sci U S A
2013
;
110
:
8644
9
.
11.
Dean
M
,
Fojo
T
,
Bates
S
. 
Tumour stem cells and drug resistance
.
Nat Rev Cancer
2005
;
5
:
275
84
.
12.
Chen
J
,
Li
Y
,
Yu
TS
,
McKay
RM
,
Burns
DK
,
Kernie
SG
, et al
A restricted cell population propagates glioblastoma growth after chemotherapy
.
Nature
2012
;
488
:
522
6
.
13.
Takebe
N
,
Harris
PJ
,
Warren
RQ
,
Ivy
SP
. 
Targeting cancer stem cells by inhibiting Wnt, Notch, and Hedgehog pathways
.
Nat Rev Clin Oncol
2011
;
8
:
97
106
.
14.
Sato
N
,
Meijer
L
,
Skaltsounis
L
,
Greengard
P
,
Brivanlou
AH
. 
Maintenance of pluripotency in human and mouse embryonic stem cells through activation of Wnt signaling by a pharmacological GSK-3-specific inhibitor
.
Nat Med
2004
;
10
:
55
63
.
15.
Reya
T
,
Clevers
H
. 
Wnt signalling in stem cells and cancer
.
Nature
2005
;
434
:
843
50
.
16.
Li
Z
,
Wang
H
,
Eyler
CE
,
Hjelmeland
AB
,
Rich
JN
. 
Turning cancer stem cells inside out: an exploration of glioma stem cell signaling pathways
.
J Biol Chem
2009
;
284
:
16705
9
.
17.
Rich
JN
. 
Cancer stem cells in radiation resistance
.
Cancer Res
2007
;
67
:
8980
4
.
18.
Kang
DC
,
Su
ZZ
,
Sarkar
D
,
Emdad
L
,
Volsky
DJ
,
Fisher
PB
. 
Cloning and characterization of HIV-1-inducible astrocyte elevated gene-1, AEG-1
.
Gene
2005
;
353
:
8
15
.
19.
Brown
DM
,
Ruoslahti
E
. 
Metadherin, a cell surface protein in breast tumors that mediates lung metastasis
.
Cancer Cell
2004
;
5
:
365
74
.
20.
Britt
DE
,
Yang
DF
,
Yang
DQ
,
Flanagan
D
,
Callanan
H
,
Lim
YP
, et al
Identification of a novel protein, LYRIC, localized to tight junctions of polarized epithelial cells
.
Exp Cell Res
2004
;
300
:
134
48
.
21.
Su
ZZ
,
Kang
DC
,
Chen
Y
,
Pekarskaya
O
,
Chao
W
,
Volsky
DJ
, et al
Identification and cloning of human astrocyte genes displaying elevated expression after infection with HIV-1 or exposure to HIV-1 envelope glycoprotein by rapid subtraction hybridization, RaSH
.
Oncogene
2002
;
21
:
3592
602
.
22.
Sarkar
D
,
Fisher
PB
. 
AEG-1/MTDH/LYRIC: clinical significance
.
Adv Cancer Res
2013
;
120
:
39
74
.
23.
Shi
X
,
Wang
X
. 
The role of MTDH/AEG-1 in the progression of cancer
.
Int J Clin Exp Med
2015
;
8
:
4795
807
.
24.
Yoo
BK
,
Emdad
L
,
Lee
SG
,
Su
ZZ
,
Santhekadur
P
,
Chen
D
, et al
Astrocyte elevated gene-1 (AEG-1): A multifunctional regulator of normal and abnormal physiology
.
Pharmacol Ther
2011
;
130
:
1
8
.
25.
Sarkar
D
,
Emdad
L
,
Lee
SG
,
Yoo
BK
,
Su
ZZ
,
Fisher
PB
. 
Astrocyte elevated gene-1: far more than just a gene regulated in astrocytes
.
Cancer Res
2009
;
69
:
8529
35
.
26.
Emdad
L
,
Sarkar
D
,
Lee
SG
,
Su
ZZ
,
Yoo
BK
,
Dash
R
,
Yacoub
A
, et al
Astrocyte elevated gene-1: a novel target for human glioma therapy
.
Mol Cancer Ther
2010
;
9
:
79
88
.
27.
Hu
B
,
Emdad
L
,
Bacolod
MD
,
Kegelman
TP
,
Shen
XN
,
Alzubi
MA
, et al
Astrocyte elevated gene-1 interacts with Akt isoform 2 to control glioma growth, survival, and pathogenesis
.
Cancer Res
2014
;
74
:
7321
32
.
28.
Guo
J
,
Chen
X
,
Xi
R
,
Chang
Y
,
Zhang
X
. 
AEG-1 expression correlates with CD133 and PPP6c levels in human glioma tissues
.
J Biomed Res
2014
;
28
:
388
95
.
29.
Zhang
F
,
Yang
Q
,
Meng
F
,
Shi
H
,
Li
H
,
Liang
Y
, et al
Astrocyte elevated gene-1 interacts with beta-catenin and increases migration and invasion of colorectal carcinoma
.
Mol Carcinog
2013
;
52
:
603
10
.
30.
Gilbertson
RJ
,
Rich
JN
. 
Making a tumour's bed: glioblastoma stem cells and the vascular niche
.
Nat Rev Cancer
2007
;
7
:
733
6
.
31.
Rao
S
,
Sengupta
R
,
Choe
EJ
,
Woerner
BM
,
Jackson
E
,
Sun
T
, et al
CXCL12 mediates trophic interactions between endothelial and tumor cells in glioblastoma
.
PLoS One
2012
;
7
:
e33005
.
32.
Calabrese
C
,
Poppleton
H
,
Kocak
M
,
Hogg
TL
,
Fuller
C
,
Hamner
B
, et al
A perivascular niche for brain tumor stem cells
.
Cancer Cell
2007
;
11
:
69
82
.
33.
Sahlberg
SH
,
Spiegelberg
D
,
Glimelius
B
,
Stenerlow
B
,
Nestor
M
. 
Evaluation of cancer stem cell markers CD133, CD44, CD24: association with AKT isoforms and radiation resistance in colon cancer cells
.
PLoS One
2014
;
9
:
e94621
.
34.
Schmidt
C
. 
Cancer stem cells in the crosshairs
.
Cancer Discov
2012
;
5
:
384
.
35.
The Cancer Genome Atlas Research Network
. 
Comprehensive genomic characterization defines human glioblastoma genes and core pathways
.
Nature
2008
;
455
:
1061
8
.
36.
Seidel
S
,
Garvalov
BK
,
Acker
T
. 
Isolation and culture of primary glioblastoma cells from human tumor specimens
.
Methods Mol Biol
2015
;
1235
:
263
75
.
37.
Reynolds
BA
,
Weiss
S
. 
Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system
.
Science
1992
;
255
:
1707
10
.
38.
Joo
KM
,
Kim
SY
,
Jin
X
,
Song
SY
,
Kong
DS
,
Lee
JI
, et al
Clinical and biological implications of CD133-positive and CD133-negative cells in glioblastomas
.
Lab Invest
2008
;
88
:
808
15
.
39.
Sanai
N
,
Alvarez-Buylla
A
,
Berger
MS
. 
Neural stem cells and the origin of gliomas
.
N Engl J Med
2005
;
353
:
811
22
.
40.
Wan
L
,
Lu
X
,
Yuan
S
,
Wei
Y
,
Guo
F
,
Shen
M
, et al
MTDH-SND1 interaction is crucial for expansion and activity of tumor-initiating cells in diverse oncogene- and carcinogen-induced mammary tumors
.
Cancer Cell
2014
;
26
:
92
105
.
41.
Liang
Y
,
Hu
J
,
Li
J
,
Liu
Y
,
Yu
J
,
Zhuang
X
, et al
Epigenetic Activation of TWIST1 by MTDH promotes cancer stem-like cell traits in breast cancer
.
Cancer Res
2015
;
75
:
3672
80
.
42.
Yamashita
T
,
Budhu
A
,
Forgues
M
,
Wang
XW
. 
Activation of hepatic stem cell marker EpCAM by Wnt-beta-catenin signaling in hepatocellular carcinoma
.
Cancer Res
2007
;
67
:
10831
9
.
43.
Woodward
WA
,
Chen
MS
,
Behbod
F
,
Alfaro
MP
,
Buchholz
TA
,
Rosen
JM
. 
WNT/beta-catenin mediates radiation resistance of mouse mammary progenitor cells
.
Proc Natl Acad Sci U S A
2007
;
104
:
618
23
.
44.
Morin
PJ
,
Sparks
AB
,
Korinek
V
,
Barker
N
,
Clevers
H
,
Vogelstein
B
, et al
Constitutive transcriptional activation by a beta-catenin-Tcf complex in APC-/- colon carcinoma
.
Science
1997
;
275
:
1784
7
.
45.
Huber
O
,
Korn
R
,
McLaughlin
J
,
Ohsugi
M
,
Herrmann
BG
,
Kemler
R
. 
Nuclear localization of beta-catenin by interaction with transcription factor LEF-1
.
Mech Dev
1996
;
59
:
3
10
.
46.
Williams
K
,
Motiani
K
,
Giridhar
PV
,
Kasper
S
. 
CD44 integrates signaling in normal stem cell, cancer stem cell and (pre)metastatic niches
.
Exp Biol Med
2013
;
238
:
324
38
.
47.
Hiraga
T
,
Ito
S
,
Nakamura
H
. 
Cancer stem-like cell marker CD44 promotes bone metastases by enhancing tumorigenicity, cell motility, and hyaluronan production
.
Cancer Res
2013
;
73
:
4112
22
.
48.
Goodison
S
,
Urquidi
V
,
Tarin
D
. 
CD44 cell adhesion molecules
.
Mol Pathol
1999
;
52
:
189
96
.
49.
Ponta
H
,
Sherman
L
,
Herrlich
PA
. 
CD44: from adhesion molecules to signalling regulators
.
Nat Rev Mol Cell Biol
2003
;
4
:
33
45
.