Imp2 regulates GBM progression by activating IGF2/PI3K/Akt pathway

Abstract

Glioblastomas multiforme (GBM) are the most frequently occurring malignant brain cancers. Treatment for GBM consists of surgical resection and subsequent adjuvant radiation therapy and chemotherapy. Despite this, GBM patient survival is limited to 12–15 months, and researchers are continually trying to develop improved therapy options. Insulin-like growth factor 2 mRNA-binding protein 2 (Imp2) is known to be upregulated in many cancers and is known to regulate the signaling activity of insulin-like growth factor 2 (IGF2). However, relatively little is known about its role in malignant development of GBM. In this study, we first found Imp2 is upregulated in GBM tissues by using clinical samples and public database search. Studies with loss and gain of Imp2 expression in in vitro GBM cell culture system demonstrated the role of Imp2 in promoting GBM cell proliferation, migration, invasion and epithelial-to-mesenchymal transition (EMT). Additionally, our results show that Imp2 regulates the activity of IGF2, which further activates PI3K/Akt signaling, thereby to promote GBM malignancy. Inhibition of Imp2 was also found to sensitize GBM to temozolomide treatment. These observations add to the current knowledge of GBM biology, and may prove useful in development of more effective GBM therapy.

Abbreviations

EMT

epithelial-mesenchymal transition

GBM

glioblastomas multiforme

IGF2

insulin-like growth factor 2

K_D

knockdown

Imp2

insulin-like growth factor 2 mRNA-binding protein 2

OE

overexpressing

PI3K

phosphatidyl inositol 3-kinase

TMZ

temozolomide

GBM-P

semi-established GBM primary cells

GBM-RE

semi-established recurrent TMZ-resistant GBM primary cells

Introduction

Glioblastomas (GBM) constitute approximately 70% of malignant primary brain tumors.¹ The current standard for GBM treatment is surgical resection, subsequent adjuvant radiation therapy and chemotherapy.² Despite intense therapy, the median survival of patients with high grade glioma (GBM) is only 12–15 months.³ Thus, the aetiologies that underlie glioblastoma tumorigenesis require further elucidation in order to develop more effective and less invasive therapies.

The PI3K/Akt pathway is a major pathway that regulates tumor cellular growth, survival and mobility. Components of PI3K/Akt pathway are frequently abnormal in a variety of tumors, making them attractive targets for anti-cancer therapy.⁴ Activation of PI3K/Akt signaling is initiated by the binding of growth factors to their receptor tyrosine kinase (platelet-derived growth factor receptor, epidermal growth factor receptor, and insulin-like growth factor receptor), which subsequently leads to membrane recruitment of PI3K and phosphorylation of Akt via phosphoinositol-3,4,5-triphosphate. It has been reported that PI3K/Akt signaling is elevated in ~88% of GBM clinical samples.⁵ This pathway is central to GBM tumorigenesis and plays a very important role in GBM drug resistance. Several PI3K/Akt pathway inhibitors have been investigated either in preclinical models or undergoing clinical trials.^6,7

Imp2 (insulin-like growth factor 2 mRNA-binding protein 2, IGF2BP2) has been reported to be dysregulated in glioblastoma and is one of the top 100 primary glioblastoma-associated genes overexpressed compared to normal tissue.⁸ It has been implicated in type 2 diabetes and in regulation of smooth muscle cell adhesion and motility.^9-12 Imp2 has been ascribed functions of regulating subcellular mRNA localization, translation and stability and specifically regulates insulin-like growth factor 2 (IGF2) mRNA translation.^11,13 Although Imp2 is known to be dysregulated in several malignancies, relatively little is known about its role in cancers.^14,15 Recently, Imp2 was found to bind the mitochondrial respiratory complex IV mRNA and complex I proteins and to help in their assembly and function, thus exerting regulation of oxidative phosphorylation in GBM stem cells. Additionally, it was found that Imp2 is crucial for the survival and function of cells that are self-renewing and tumor-initiating.¹⁶ Here, provide evidence to suggest that 1) Imp2 is dysregulated in GBM clinical samples, and 2) Imp2 regulates IGF2-mediated activation of PI3K/Akt signaling, further inducing proliferation, migration and invasion ability and epithelial-mesenchymal transition (EMT) in GBM cells. Three) Inhibition of Imp2 can sensitize GBM to temozolomide (TMZ) treatment.

Results

Imp2 is upregulated in GBM

To study the role of Imp2 during GBM development, we first analyzed Imp2 mRNA level in 49 clinical GBM patient samples. We found Imp2 expression was upregulated in 43 out of 49 GBM patient tissues (median log₂ ratio = 1.424) compared with 6 normal brain tissues (median log₂ ratio = −0.13) (Fig. 1A). Immunohistochemical staining results demonstrated significantly increased cytoplasmic level of Imp2 in GBM tissues (32/49; 65.3%) compared with normal brain samples (Fig. 1B). Our results was further confirmed using the public database NCBI/GDS1813, which showed that Imp2 levels were significantly increased in 30 GBM samples compared to 4 normal brain tissues (p = 0.001, Fig. 1C).¹⁷

Figure 1.

Imp2 is upregulated in GBM. (A) Relative mRNA level of Imp2 in 49 clinical GBM tissues compared to 6 normal brain tissues analyzed by RT-PCR. (B) Immunohistochemical staining of Imp2 in GBM and normal brain tissues. Scale bar: 50 μm. (C) Imp2 is increased in GBM according to NCBI/GDS1813 database search. n = 30 GBM tissues; n = 4 normal brain tissues. p = 0.001.

Imp2 upregulates IGF2 levels and activates PI3K/Akt pathway in GBM cells

Since Imp2 has been shown to promote IGF2 mRNA translation in other cell types, we next sought to see if Imp2 regulates IGF2 mRNA or protein levels in GBMs.¹³ We developed Imp2 overexpressing (OE) and knockdown (K_D) U87 and U251 stable cell lines, then assessed the IGF2 levels by RT-PCR and western blot. As expected, changes of Imp2 levels did not change IGF2 mRNA levels (Fig. 2A). However, overexpression of Imp2 significantly increased and silencing of Imp2 significantly decreased IGF2 protein levels in both U87 and U251 cell lines (Fig. 2B). Previous studies revealed that IGF2 is involved in the development of highly proliferative GBMs via PI3K/Akt pathway.¹⁸ Herein, we harvested Imp2 OE and K_D U87 and U251 cells and performed protein gel blot for Akt and p-Akt expression. As shown in Figure 2C, levels of p-Akt were decreased by knockdown of Imp2 and increased by overexpression of Imp2 compared to control cells. These data indicate that implicates Imp2 may activate IGF2/PI3K/Akt signaling axis in human GBMs.

Figure 2.

Imp2 affects IGF2 level and Akt activation. (A) RT-PCR analysis of IFG2 mRNA level in Imp2 overexpression (OE) or knockdown (K_D) U87 and U251 cells compared to vector plasmid (Vector) or scramble shRNA (Scramble) transfected control cells. The assays were performed in triplicate. (B) Western blot for Imp2 and IGF-2 in Imp2 OE or K_D U87 and U251 cells. (C) Western blot for Akt and p-Akt level in Imp2 OE or K_D U87 and U251 cells.

Imp2 promotes cell proliferation via activating IGF2/PI3K/Akt pathway

PI3K/Akt pathway plays central role in GBM biology and inhibition of PI3K/Akt can lead to reduced GBM proliferation.^19,20 To explore the role of PI3K/Akt activator Imp2 in GBM progression, we performed cell proliferation assay with Imp2 OE and K_D cell lines and primary GBM (GBM-P) cells. As shown in Figure 3A, Imp2 OE cells grow significantly faster than vector control cells. However, cells with Imp2 knockdown show delayed growth compared with the scramble shRNA transfected control cells (Fig. 3B). We have further extended these studies and monitored the effect of IGF2 and PI3K inhibition on cell growth in Imp2 OE cells using IGF2 neutralization antibody (IGF2 Ab) and PI3K inhibitor LY294002. We found that inhibition of IGF2 or PI3K partially or completely abrogated the proliferation-promoting effect of Imp2 overexpression (Figure 3C). These results suggest that Imp2 promotes GBM cell proliferation which is mediated by IGF2/PI3K/Akt pathway.

Figure 3.

Imp2 regulates GBM cell proliferation. (A) MTT assay in Imp2 OE U87, U251 or GBM-P cells compared to Vector control cells. (B) MTT assay in Imp2 K_D U87, U251 or GBM-P cells compared to Scramble control cells. (C) MTT assay in Imp2 OE U87, U251 or GBM-P cells treated with IGF2 neutralization antibody (IGF2 Ab, 0.6 μg/ml) or PI3K inhibitor LY294002 (20 μM) for 4 d The assays were performed in triplicate. Data was shown as mean ± SD. *P < 0.05, **P < 0.01 compared to Vector control cells. #P < 0.05, ##P < 0.01 compared to OE control cells.

Imp2 promotes cell migration, invasion and epithelial-mesenchymal transition (EMT)

PI3K/Akt also facilitates the invasive phenotype of GBM in terms of cell motility.²¹ We further performed wound healing and cell invasion assays. Microscope examination of U87, U251 and GBM-P cultures post-wounding, revealed a significant delay in the wound closure rate of Imp2 K_D cells and increased wound closure rate in Imp2 OE cells compared to scramble shRNA or vector control cells, respectively (Fig. 4A–C). Additionally, the number of invaded cells were significantly decreased in Imp2 K_D groups, and significantly increased in Imp2 overexpressing cells compared to control cells (Fig. 5A–C).

Figure 4.

Imp2 regulates GBM cell migration. Wound healing assay in Imp2 K_D or OE U87 (A), U251 (B) and GBM-P (C) cells at 48 hours after cells were wounded. Magnification: 100 ×. The assays were performed in triplicate. Data was shown as mean ± SD. *P < 0.05, **P < 0.01 compared to Scramble or Vector control cells.

Figure 5.

Imp2 affects GBM cell invasion. Transwell assay in Imp2 K_D or OE U87 (A), U251 (B) and GBM-P (C) cells at 24 hours after cells were seeded. Magnification: 100×. The assays were performed in triplicate. Data was shown as mean ± SD. *P < 0.05, **P < 0.01 compared to Scramble or Vector control cells.

To study the role of IGF2/PI3K/Akt pathway in Imp2-induced GBM cell migration and invasion, we treated U87 and U251 cells with IGF2 neutralization antibody and found that blocking IGF2 significantly reduced cell migration and invasion ability, even in Imp2 OE cells (Fig. 6A). We also treated control or Imp2 OE cells with LY294002. As showed in Figure 6B, inhibition of PI3K significantly reduced U87 cell migration and invasion. However, Imp2 overexpression failed to rescue cell migration and invasion.

Figure 6.

Imp2 regulates GBM cell migration and invasion via IGF2/PI3K/Akt pathway. (A) Wound healing and transwell assay in Imp2 OE U87 cells treated with PBS or IGF2 Ab (0.6 μg/ml) for 48 hours after cells were wounded. *P < 0.05, **P < 0.01 compared to PBS treated Vector cells. #P < 0.05, compared to PBS treated OE cells. (B) Wound healing and transwell assay in Imp2 OE U87 cells treated with DMSO or LY294002 (20 μM) for 24 hours after cells were seeded. *P < 0.05, **P < 0.01 compared with DMSO treated Vector cells. ##P < 0.01 compared to DMSO treated OE cells. The assays were performed in triplicate. Data was shown as mean ± SD.

EMT is an important process during tumor progress and malignant transformation, and activation of PI3K/Akt axis plays central roles during EMT.²² From the results above, Imp2 affects GBM cell migration and invasion. We then evaluated the effect of Imp2 on GBM cell EMT, by analyzing expression of 3 EMT markers (E-cadherin, Vimentin and N-cadherin) in Imp2 OE and K_D U87 cells. RT-PCR results showed that Imp2 overexpression resulted in reduced E-cadherin and increased Vimentin and N-cadherin expression, while Imp2 silencing increased E-cadherin and reduced Vimentin and N-cadherin expression (Fig. 7A). These results were further confirmed by western blot (Fig. 7B). Taken together, these results indicate that Imp2 regulates GBM cell migration, invasion and EMT ability.

Figure 7.

Imp2 regulates expression of EMT markers. (A) Relative mRNA expression of E-cadherin, Vimentin and N-cadherin in Imp2 K_D or OE U87 cells. *P < 0.05, **P < 0.01 compared to control cells. (B) Western blot of E-cadherin, Vimentin and N-cadherin levels in Imp2 K_D and Imp2 OE U87 cells.

Imp2 regulates temozolomide (TMZ) anti-drug resistance in GBM cells

EMT has previously been found to be involved in TMZ resistance in human glioma cells.²³ Meanwhile, inhibition of PI3K/Akt pathways has been shown to sensitize GBM cells to TMZ treatment.²⁴ In order to investigate the possible role of Imp2 in GBM cell sensitivity to TMZ, we first develop TMZ-resistant subclones of U87 cells by exposing cells to TMZ for 6 months. U87 TMZ-resistant cells (U87-RE) was confirmed by increase of inhibitory concentration (IC₅₀) of approximate 4 folds (U87: 261.8 μM vs U87-RE: 1076 μM; Fig. 8A). Then we investigated whether Imp2 overexpression or downregulation could affect GBM cell response to TMZ. Chemotoxicity assay showed that Imp2 overexpression was associated with increased cell viability against different TMZ doses (IC₅₀: 1892 μM). However, downregulation of Imp2 resulted in increased sensitivity of GBM cells to TMZ (IC₅₀: 535 μM). Furthermore, we performed colony formation assay and confirmed Imp2 enhances TMZ sensitivity in U87-RE cells (Fig 8B). Lastly, these results were recapitulated in recurrent TMZ-resistant primary GBM (GBM-RE) cells. As shown in Figure 8C, overexpression of Imp2 led to increased IC₅₀ of TMZ in GBM-RE cells (770 μM in GBM-RE cells vs 1076 μM in GBM-RE + Imp2 OE cells). In contrast, knockdown of Imp2 reduced IC₅₀ of TMZ in GBM-RE cells to 436 μM. Consistently, the colony numbers were significantly lower in GBM-RE+ Imp2 K_D cells and higher in GBM-RE + Imp2 OE cells compared to GBM-RE cells (Fig. 8D) when treated with TMZ. Together, these data demonstrate that Imp2 mediates GBM cell TMZ-resistance and inhibition of Imp2 can sensitize GBM cells to TMZ.

Figure 8.

Imp2 expression is related to TMZ-resistance. (A) Cell viability assay in U87, U87 TMZ resistant cells (U87-RE) or U87-RE cells with overexpression or knockdown Imp2. Cells were treated with TMZ at indicated concentration for 2 d (B) Colony formation assay in U87-RE, or U87-RE cells with overexpression or knockdown Imp2. Cell were treated with TMZ at indicated concentration for 14 d (C) Cell viability assay in TMZ-resistant primary GBM cells (GBM-RE) or GBM-RE cells with overexpression or knockdown Imp2. Cells were treated with TMZ at indicated concentration for 2 d (D) Colony formation assay in GBM-RE cells or GBM-RE cells with overexpression or knockdown Imp2. Cells were treated with TMZ at indicated concentration for 14 d *P < 0.05 compared to U87-RE or GBM-RE group.

Discussion

In this study, we first analyzed Imp2 expression in 49 clinical samples and found that Imp2 expression was increased in GBM tissues compared with normal brain tissues. Further, a database search showed significantly upregulated Imp2 expression in GBM samples compared to normal brain tissue. Cell proliferation assay demonstrated that Imp2 stimulates cell growth. A wound healing assay was then performed and showed that Imp2 promoted wound healing. Invasion assay showed that Imp2 had significant positive effect on invasion ability of GBM cells in vitro. Further, we found evidence to support the notion that Imp2 regulates EMT in GBM cells by reducing E-cadherin expression and increasing Vimentin and N-cadherin expression. These three factors are intimately involved in cell adhesion, EMT and migration, and are commonly used markers of EMT. A previous study showed that EMT is involved in TMZ resistance in human glioma cells.²³ Our study adds to this knowledge by showing that, at least in vitro, one mechanism by which GBM cells can increase TMZ resistance, is by increasing expression of Imp2.

A recent study showed that Imp3 (an activator of IGF2 mRNA translation, similar to Imp2) regulates the IGF2/Akt axis in GBM.²⁵ Although the authors did not mention the following in the main content of their article, their supplementary data shows that Imp2 expression is significantly different in GBM compared to normal control (p = 0.0446). The authors probably did not discuss this because they found other Imps far more significant in their study, leading them to investigate those Imps. Our results however, suggest that the differential expression of Imp2 in GBM is much more significant than the previous report suggested (p = 0.001 obtained by database analysis of 30 GBM samples compared to 4 normal brain tissue samples, Fig. 1C). This discrepancy may be explained by differences in study design. Suvasini et al.²⁵ normalized their RT-PCR data with 6 different internal reference genes (AGPAT, GARS, ATP5G1, GAPDH, ACTB and RPL35A) and used 121 GBM samples, whereas the present report only utilized β-actin for normalization of 49 GBM samples.

Similarly to the above Imp3 data by Suvasisni et al.²⁵, our experiment found that Imp2 also positively regulates IGF2/PI3K/Akt. Akt is phosphorylated to become its active form p-Akt, a process which can occur in both PI 3-kinase dependent or independent manners.^26-28 Our observation of altered p-Akt, but not Akt, is consistent with the previous report on Imp3,²⁵ and we also found that Imp2 overexpression could not rescue cells treated with the PI3K inhibitor LY294002. Thus, it appears that Imp2 regulates Akt signaling by affecting PI3K-dependent (and not independent) phosphorylation of Akt, rather than affecting the expression of the Akt itself.

IGF2 is known to interact with IGFR1R, which causes downstream activation of PI3K/Akt signaling and finally protein translation and proliferation.²⁵ Further, Imp2 was previously shown to bind IGF2 mRNA and upregulate its translation, increasing the bioavailabilility of IGF2 in glioma cells and breast cancer.^29–32 Consistently with this, our data shows that the bioavailability of IGF2 to bind IGF1R is influenced by Imp2 expression. Furthermore, we found that blocking IGF2 with antibodies reduces proliferation, migration and invasion ability, even in Imp2 overexpressing GBM cells, which provides further evidence for the notion that Imp2 may regulate GBM cell tumorigenesis by regulating IGF2-mediated activation of the PI3K/Akt cascade (Fig. 9). Thus we found evidence to suggest that Imp2 regulates the binding of IGF2 to IGFR1R, which further activates the PI3K/Akt signaling pathway.

Figure 9.

Schematic model demonstrating Imp2 function in GBM and potential mechanisms.

E-cadherin is a type-1 transmembrane protein that plays important roles in cell adhesion by forming adherens junctions, which bind cells within tissues together, and is a known tumor suppressor.^33,34 In many cancers, E-cadherin expression loss is often accompanied by de novo expression of mesenchymal cadherins, such as N-cadherin³⁵ and several studies have reported that E-cadherin is rare in high grade gliomas (anaplastic astrocytoma and GBM), while N-cadherin is highly expressed,^36-40 which is consistent with our own observations. Recently, E-cadherin was reported to be an inhibitor of the PI3K/Akt signaling pathway in ovarian cancer, and that loss of E-cadherin resulted in aberrant PI3K/Akt signaling.³⁴ Similarly, we also found that E-cadherin levels inversely correlated with PI3K/Akt signaling activity in GBM, and we also found that E-cadherin expression was negatively regulated by Imp2-induced Akt signaling. It should be noted that E-cadherin itself can activate PI3K/Akt, at least in ovarian cancer, although we did not investigate this in GBM.⁴¹ N-Cadherin is a cell-cell adhesion glycoprotein, frequently found in cancer cells where it facilitates trans-endothelial migration by upregulating src kinase signaling. Although the mechanism still needs elucidation, increased expression of N-cadherin is frequently associated with loss of E-cadherin expression,²² a notion that our data is in agreement with. Vimentin is a type III intermediate filament (IF) protein expressed in mesenchymal cells and is their predominant cytoskeletal component and responsible for cell integrity. We found that Vimentin expression was upregulated in GBM cells overexpressing Imp2, which along with elevated N-cadherin and reduced E-cadherin, suggests that Imp2 induces EMT in GBM.

In summary, we present confounding evidence that aberrantly overexpressed Imp2 in GBM may stimulate GBM progression by upregulating translation of IGF2 mRNA to protein, making IGF2 protein available to bind IGF1R. Imp2 was also found to induce phosphorylation and thereby activation of Akt. These two events appear to induce the PI3K/Akt signaling pathway and cause the promotion of cell proliferation, migration, invasion and EMT. Inhibition of Imp2 can sensitize GBM to TMZ treatment. Our discoveries add to the current knowledge of GBM biology and may prove useful in future GBM therapy development.

Materials and Methods

Patient samples

The study protocol was approved by the Medical Ethics and Human Clinical Trial Committee of First Hospital of Jilin University. All of the patients gave their written consent. None of the patients had received adjuvant chemotherapy or radiotherapy before surgery. Diagnosis of GBM was histopathologically confirmed by 2 pathologists according to WHO carcinoma. The clinicopathological characteristics of patients were shown in Table 1.

Table 1.

Clinical characteristics in archival GBM patients

Factor	Cases
Gender
Male	38
Female	11
Age (Years)
< 50	26
≥ 50	23
KPS
< 75	23
≥ 75	26
Extent of resection
GTR	34
STR	15
Tumor size (cm)
< 6	25
≥ 6	24

KPS: Karnofsky Performance Score.

GTR: Gross-total Resection.

STR: Subtotal Resection.

^*P < 0.05.

Cell lines and reagents

The U87 and U251 cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and grown in Dulbecco's Modified Eagle's Media (DMEM) (Gibco, Grand Island, NY, USA) containing 10% (v/v) fetal bovine serum (FBS, Gibco), 2 mM l-glutamine (Sigma-Aldrich, St. Louis, MO, USA), 100U/ml penicillin and 100 μg/ml streptomycin (Gibco). The primary and recurrent TMZ-resistant GBM cells were established from primary and recurrent GBM surgical cases. The method of isolation and culture was described by Hasselbach, L. A. et al.⁴² Cells were cultured at 37°C in a humidified atmosphere containing 5% CO₂. LY294002, TMZ and IGF2 antibody were all purchased from Sigma-Aldrich.

Immunohistochemical staining

Immunohistochemical studies were performed on 5-µm sections of formalin fixed, paraffin-embedded tissue. Antigen retrieval was carried out with 0.01 mol/l citrate buffer at pH 6.0 in an 800-W microwave oven for 15 min before immunostaining. The sections were blocked for endogenous protein binding and peroxidase activity with an application of Dual Endogenous Block (Dako, Santa Clara, CA, USA) for 10 min. The sections were then incubated with a monoclonal antibody specific for Imp2 (Abcam, Cambridge, MA) at a 1:25 dilution overnight at 4°C. The next day, sections were incubated with biotinylated secondary antibody (Santa Cruz Biotechnology, Dallas, TX, USA) for 1 h at room temperature. VECTSTAIN Elite ABC complex (Vector lab, Burlingame, CA, USA) and AEC+ Substrate Chromogen (Dako) were used for primary antibody detection. Negative control sections were incubated in PBS instead of primary antibody.

Real-time PCR (RT-PCR)

To quantify gene expression, total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad CA, USA) according to the manufacturer's instructions. Reverse transcription involved the Superscript III First Strand Synthesis kit (Invitrogen). Expression of genes was detected by use of SYBR® Green Real-Time PCR Master (Invitrogen) and normalization involved the 2−^ΔΔCt method relative to β-actin. The primers used for E-cadherin: sense, 5′-AATCCAAAGCCTCAGGTCATAAACA-3′ and antisense, 5′-TTGGGTCGTTGTACTGAATGGTC-3′; Vimentin: sense, 5′-GTTTCCCCTAAACCGCTAGG-3′ and antisense, 5′-AGCGAGAGTGGCAGAGGA-3′; N-cadherin: sense, 5′-TGGATGGGCTGCCTCCAGGTGAC-3′ antisense, 5′-ACCAGCCCACCCCTCGAGCCC-3′; β-actin: sense, 5′-GTGGGGCGCCCCAGGCACCA-3′ and antisense, 5′-CTCCTTAATGTCACGCACGATTT-3′.

siRNA, shRNA, plasmid and transfections

Imp2 siRNA, scramble shRNA and Imp2 shRNA were purchased from Santa Cruz Biotechnology. Imp2 plasmid was obtained from Addgene (Cambridge, MA, USA).⁴³ To raise stable clones overexpressing Imp2, cells were transfected with the pcDNA3-GFP-IMP2–2 construct using X-tremeGENE HP DNA Transfection Reagent (Roche, Shanghai, China) according to the manufacturer's protocol. Twenty-four h after transfection, the transfectants were selected on neomycin sulfate (200–500 μg/ml) for 4 weeks after which individual clones were further amplified. Stable clones were then confirmed by real time qPCR and protein gel blotting before further characterization. Stable Imp2 knockdown cells were established as described previously.¹⁶ Plasmid or siRNA transient transfection was performed using X-tremeGENE HP DNA Transfection Reagent or X-tremeGENE siRNA Transfection Reagent respectively (Roche), according to the manufacturer's instructions. Subsequent experiments with these cells were performed 48 h after transfection.

Western blot analysis

U87 and U251 cells were harvested and lysed with lysis buffer (100 mM NaCl, 10 mM Tris-HCl pH 7.5, 2 mM EDTA, and 0.5% w/v deoxycholate) containing phosphatase and protease inhibitor cocktail (Santa Cruz) added to the buffer immediately before use. Cells were harvested using a cell scraper, transferred to microfuge tubes, and centrifuged at maximum speed for 10 min at 4°C. The supernatants were collected and the protein content was determined using Bradford protein assay (Bio-Rad, Irvine, CA, USA). Proteins were separated by electrophoresis on 10% sodium dodecyl sulfate (SDS)-polyacrylamide gels, and transferred onto nitrocellulose membranes (Bio-Rad) for western blotting. Membranes were blocked (with 3% BSA in TBS) and then incubated overnight with primary antibody at 4°C. After HRP-conjugated secondary antibody incubation for 1 h, the bound antibodies were detected using the Supersignal West Pico ECL chemiluminescence kit (Thermo Scientific, Rockford, IL, USA). The antibodies for Imp2, IGF2, E-cadherin, Vimentin, N-cadherin, p-AKT, AKT, β-actin and secondary antibodies were purchased from Santa Cruz Biotechnology.

Wound healing assay

Transfected U87 or U251 cells were seeded into each well of a 6-well plate and incubated overnight to reach near confluence. The cell monolayers were wounded with a 1 ml pipette tip and cultured with DMEM medium supplemented with 10% FBS at 37°C with 5% CO₂ for 48 h. The gaps in cell monolayers were photographed and quantified using Image J software. Each experiment was repeated in triplicates.

Transwell assay

Invasion chambers (Corning Inc.., Corning, NY, USA) in a 24-well plate coated with 50 ml matrigel in DMEM were incubated at 37°C for 2 h. Transfected U87 or U251 cells (1 × 10⁵) were suspended in 200 µl serum-free DMEM medium and seeded into the upper layer of chamber. Cell culture medium (700 µl) was added to each lower layer and then the chambers incubated for 24 h at 37 °C with 5% CO₂. The chambers were washed with PBS 3 times, fixed with methanol and finally stained with crystal violet. Percentage of invaded cell number was calculated. Each experiment was repeated in triplicates.

Cell viability assay

Cell viability was analyzed by MTT assay (Promega, Madison, WI, USA). Briefly, transfected glioblastoma cells (5 × 10³ per well) were seeded into 96-well plates and incubated overnight. Then, the cells were treated with various concentrations of TMZ (0, 100, 250, 500, 1000 and 2000 µM) and incubated further. After 48 h incubation, MTT solution was added to the treated cell cultures, incubated for 2 h, and then their absorption at 570 nm read using a microplate reader. Each experiment was repeated in triplicates.

Colony formation assay

GBM cells (500 per well) were seeded on 6-well plate. After incubated overnight, the cells were treated with different concentration of TMZ. Cells were incubated at 37°C, and the medium was replaced with fresh medium every 3 d Colonies were fixed and then stained with 1% crystal violet 14 d later and counted. The value was presented as percentage of the colony number under the treatment conditions relative to the control. The experiments were conducted in triplicate.

Statistics

Values were expressed as mean ± standard deviation of samples measured in triplicate. Statistical analysis was performed using SPSS V19.0. Quantitative data from 2 groups were analyzed using the student's t-test and 2-tailed distribution. Three or more groups were analyzed using analysis of variance (ANOVA). A p value < 0.05 was considered significant.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.