Journal of Neuro-Oncology

, Volume 111, Issue 3, pp 273–283

MicroRNA-183 upregulates HIF-1α by targeting isocitrate dehydrogenase 2 (IDH2) in glioma cells

Authors

  • Hirotomo Tanaka
    • Department of NeurosurgeryKobe University Graduate School of Medicine
    • Department of NeurosurgeryKobe University Graduate School of Medicine
  • Kazuhiro Tanaka
    • Department of NeurosurgeryKobe University Graduate School of Medicine
  • Satoshi Nakamizo
    • Department of NeurosurgeryKobe University Graduate School of Medicine
  • Masamitsu Nishihara
    • Department of NeurosurgeryNishi-Kobe Medical Center
  • Katsu Mizukawa
    • Department of NeurosurgeryKobe University Graduate School of Medicine
  • Masaaki Kohta
    • Department of NeurosurgeryKobe University Graduate School of Medicine
  • Junji Koyama
    • Department of NeurosurgeryKobe University Graduate School of Medicine
  • Shigeru Miyake
    • Department of NeurosurgeryKobe University Graduate School of Medicine
  • Masaaki Taniguchi
    • Department of NeurosurgeryKobe University Graduate School of Medicine
  • Kohkichi Hosoda
    • Department of NeurosurgeryKobe University Graduate School of Medicine
  • Eiji Kohmura
    • Department of NeurosurgeryKobe University Graduate School of Medicine
Laboratory Investigation

DOI: 10.1007/s11060-012-1027-9

Cite this article as:
Tanaka, H., Sasayama, T., Tanaka, K. et al. J Neurooncol (2013) 111: 273. doi:10.1007/s11060-012-1027-9

Abstract

MicroRNAs (miRs) are small, non-coding RNAs that regulate gene expression and contribute to cell proliferation, differentiation and metabolism. Our previous study revealed the extensive modulation of a set of miRs in malignant glioma. In that study, miR microarray analysis demonstrated the upregulation of microRNA-183 (miR-183) in glioblastomas. Therefore, we examined the expression levels of miR-183 in various types of gliomas and the association of miR-183 with isocitrate dehydrogenase 2 (IDH2), which has complementary sequences to miR-183 in its 3′-untranslated region (3′UTR). In present study, we used real-time PCR analysis to demonstrate that miR-183 is upregulated in the majority of high-grade gliomas and glioma cell lines compared with peripheral, non-tumorous brain tissue. The mRNA and protein expression levels of IDH2 are downregulated via the overexpression of miR-183 mimic RNA in glioma cells. Additionally, IDH2 mRNA expression is upregulated in glioma cells expressing anti-miR-183. We verified that miR-183 directly affects IDH2 mRNA levels in glioma cells using luciferase assays. In malignant glioma specimens, the expression levels of IDH2 were lower in tumors than in the peripheral, non-tumorous brain tissues. HIF-1α levels were upregulated in glioma cells following transfection with miR-183 mimic RNA or IDH2 siRNA. Moreover, vascular endothelial growth factor and glucose transporter 1, which are downstream molecules of HIF-1α, were upregulated in cells transfected with miR-183 mimic RNA. These results suggest that miR-183 upregulation in malignant gliomas induces HIF-1α expression by targeting IDH2 and may play a role in glioma biology.

Keywords

miR-183IDH2HIF-1αGlioblastoma

Introduction

MicroRNAs (miRs) are small, non-coding RNAs transcribed from DNA that are 18–24 nucleotides in length and subsequently regulate gene expression and contribute to cell proliferation, apoptosis, differentiation and metabolism [1, 2]. miRs are processed from primary transcripts, known as primary miRs, into short stem–loop structures called premature miRs and then finally processed into mature miRs [3]. A single miR has the capacity to regulate a large number of target messenger RNAs (mRNAs), and the main function of miRs is to downregulate gene expression [4, 5]. It has been proposed that depending on the role of their mRNA targets, miRs can function either as tumor suppressors or as oncogenes [6]. In glioblastoma, which is one of the most malignant brain tumors, a number of miRs are reported to display aberrant expression patterns [711]. Using miR microarray analysis, we previously described that miR-10b, miR-21, miR-183, miR-92b, and miR-106b are highly expressed in glioblastomas compared with normal brain tissue [12]. Several other reports have also identified these miRs as being upregulated in glioblastomas [1316]. Recently, the expression of miR-183 was identified as being upregulated in colorectal cancer and hepatocellular carcinoma and was also reported to be involved in tumor invasion or migration [1720].

Isocitrate dehydrogenases (IDHs) are a group of enzymes involved in monocarbon metabolism, which catalyze the conversion of isocitrate to α-ketoglutarate by oxidative decarboxylation [2125]. IDH1 is a cytosolic and peroxisomal enzyme, whereas IDH2 and IDH3 are mitochondrial enzymes [25, 26]. Only IDH1 and IDH2 are known to be mutational targets in human cancers. Recently, mutations in IDH1 and IDH2 were reported in association with gliomas, acute myeloid leukemia and other cancers [2732].

In this study, we determined that miR-183 is upregulated in a majority of malignant (high-grade) gliomas compared with peripheral, non-tumorous brain tissue. Because miR-183 has complementary sequences in the 3′-untranslated region (3′UTR) of IDH2 mRNA, we hypothesized that miR-183 regulates IDH2 expression. We demonstrated that miR-183 potently inhibits IDH2 expression at both the mRNA and protein levels in glioma cell lines. We next demonstrated that the expression of IDH2 mRNA and protein was downregulated in most malignant gliomas compared with peripheral, non-tumorous brain tissue. Additionally, we demonstrated that the overexpression of miR-183 or knockdown of IDH2 by siRNA in glioma cells elevated the protein levels of HIF-1α. These results imply that overexpressing miR-183 might repress the mitochondrial enzyme IDH2 and upregulate HIF-1α in glioma cells.

Materials and methods

Cell culture

Five GBM cell lines (U251, U87MG, T98G, A172 and SF126) were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Nacalai tesque, Kyoto, Japan) containing glutamine, 10 % fetal bovine serum, and penicillin/streptomycin. Immortalized human astrocyte cell line (SVGp12) were maintained in Eagle’s minimum essential medium (E-MEM, Wako, Osaka, Japan) containing glutamine, 10 % fetal bovine serum, and penicillin/streptomycin. Cells were grown in a 37 °C incubator with 5 % CO2.

Clinical specimens

Glioma tissues were obtained from therapeutic procedures performed as part of routine clinical management at the Department of Neurosurgery at the University of Kobe. Tissue samples were resected during surgery and immediately frozen in liquid nitrogen for subsequent investigation. A total of 88 glioma specimens (53 glioblastomas, 14 anaplastic astrocytomas, 3 anaplastic oligodendrogliomas, 10 diffuse astrocytomas, 5 oligodendrogliomas, and 3 pilocytic astrocytomas) and 6 peripheral brain tissue samples were included in this study. This study was approved by the ethics committee at the Kobe University Graduate School of Medicine.

Transfection protocol

The miR-183 mimic RNA (MISSION microRNA Mimic-has-miR-183) and the negative control RNA (MISSION microRNA Mimic, Human, Negative Control 1) were purchased from Sigma-Aldrich (St. Louis, MO, USA). The anti-miR-183 (mirVana™ Inhibitors) and negative control (mirVana™ Inhibitors Negative Control-1) were purchased from Applied Biosystems/Ambion (Foster City, CA, USA). The IDH2 siRNA and negative control siRNA were purchased from Sigma-Aldrich (St. Louis, MO, USA). HiPerFect Transfection Reagent (QIAGEN, Hilden, Germany) was used to transfect the miR-183 mimic/inhibitor or IDH2 siRNA into cells, and the cells were then subjected to real-time PCR and western blotting, according to the manufacturer’s instructions. The miR-183 mimic/inhibitor or IDH2 siRNA were transfected at a final concentration of 20 nM. A total of 1.0 × 105 glioma cells were plated into a 12-well plate in 1 ml of DMEM containing serum. The miR-183 mimic/inhibitor or siRNA was diluted in 100 μl OptiMEM (Invitrogen, Carlsbad, CA, USA) without serum. Then, 6 μl of HiPerFect Transfection Reagent was added to the diluted miR-183 mimic/inhibitor or siRNA solution and mixed, and the complexes were added drop-wise onto the cells. The cells were incubated with the transfection complexes under their normal growth conditions (37 and 5 % CO2) for 48–96 h (mRNA assay: 48–72 h, protein assay: 72–96 h) before the assays were performed. A Cy3 dye-labeled Anti-miR Negative Control (Applied Biosystems/Ambion) was used to monitor the transfection efficiency of the transfection experiment. Transfection efficiency was greater than 80 % in all glioma cell lines (U251, U87, A172, T98, and SF126) (Supplementary Fig. 1B).

Luciferase assay

We used three sets of luciferase reporter systems (pLightSwitch-3′UTR-empty-Luc vector and pLightSwitch-wild-type-IDH2-3′UTR-Luc vector, pLightSwitch-mutant-IDH2-3′UTR-Luc vector) (Switchgear genomics, Menlo Park, CA, USA) (Fig. 3a). The pLightSwitch-mutant-IDH2-3′UTR-Luc vector was generated from pLightSwitch-wild-type-IDH2-3′UTR-Luc vector by replace the binding site of miR-183 using QuikChange Site-Directed Mutagenesis Kit (Agilent technologies, Santa Clara, CA, USA). The empty vector contains neither the 3′UTR nor a control insert. DharmaFECT Transfection Reagent (Promega, Fitchburg, Wisconsin, USA) was used to co-transfect the miR and plasmid into the cells according to the manufacturer’s protocol. After transfection, the cells were incubated in 96-well plates and harvested for the luciferase assays approximately 48 h later. The Dual-Luciferase Reporter Assay System (Promega) was used to measure the luciferase activity in the cells transfected with either the pLightSwitch-3′UTR-control or the pLightSwitch-3′UTR-IDH2 vector and pGL4-renilla luciferase-control vector using a GloMax® 96 Microplate Luminometer (Promega). Expression was calculated as the relative firefly luciferase activity normalized to the activity of the co-transfected control renilla luciferase.

Western blotting and antibodies

Cells were lysed on ice for 30 min in lysis buffer containing 1.0 % NP-40, 0.1 % SDS, 100 mM Tris–HCl (pH 7.4), 150 mM NaCl, 5 mM EDTA, 1 mM ZnCl2, 50 mM NaMoO4, 10 mM NaF, 0.1 mM sodium vanadate, and a protease inhibitor cocktail. Electrophoresis was performed for approximately 90 min at 20 mA in a 1.0 mm polyacrylamide gel. The proteins were transferred electrophoretically at 120 mA for 60 min and assessed by immunoblotting using primary antibodies against IDH2 (Abcam, Cambridge, MA, USA), HIF-1α (H1alpha67) (Novus Biologicals, Littleton, CO, USA) and ß-actin (Ambion, Inc., Austin, TX, USA). Immunoblot development was performed as recommended by the protocol provided in the ECL Advance Western Blotting detection kit (GE Healthcare Life Sciences, Little Chalfont, UK). For quantitative analysis by normalizing to β-actin, Image J software (National Institutes of Health (NIH), Bethesda, MD, USA) was used.

Real-time RT-PCR for miRs

Reverse transcription was performed using the TaqMan® microRNA Reverse Transcription Kit, according to the manufacturer’s instructions. TaqMan® microRNA assays (Applied Biosystems) were used to quantify mature miR expression, as previously described [12]. RNU6 (Applied Biosystems) was used as the endogenous control for the miR expression studies. Real-time PCR reactions for miRs were performed in triplicate in 20 μl volumes. Quantitative miR expression data were acquired and analyzed using an Applied Biosystems 7500 real-time PCR system (Applied Biosystems) and the ΔΔ-Ct method.

Real-time RT-PCR for IDH2, HIF-1α, VEGF, and GLUT1

Reverse transcription was performed using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems), according to the manufacturer’s instructions. Expression levels of IDH2 and HIF-1α mRNA were quantitatively assayed by real-time RT-PCR using TaqMan® Gene Expression Assays (Applied Biosystems). The 18S ribosomal RNA transcript was used as the endogenous control. Quantitative mRNA expression data were acquired and analyzed by the ΔΔ-Ct method using an Applied Biosystems 7500 real-time PCR system (Applied Biosystems).

Immunocytochemistry of IDH2

Cells were fixed with 4 % paraformaldehyde in sodium phosphate (NaPi) buffer (pH 7.4) for 15 min and then incubated in 0.3 % triton-X/PBS for 15 min. Both endogenous peroxidase and biotin were blocked, and the cells were incubated with antibodies against IDH2 (Abcam) at 4 °C overnight. The cells were allowed to react with peroxidase-conjugated anti-mouse IgG monoclonal antibody (Histofine Simple Stain MAX-PO; Nichirei, Tokyo, Japan) for 60 min and visualized by immersing the sections in 0.03 % diaminobenzidine solution containing 2 mM hydrogen peroxide for 5 min. The nuclei were lightly stained with Mayer’s hematoxylin.

α-Ketoglutarate (α-KG) Assay

α-KG Assay was performed using an α-KG Assay Kit (Abcam), according to the manufacturer’s instructions. Cells were rapidly homogenized with 100 μl of ice cold buffer containing 1 % NP-40, and 1–50 μl of supernatant was transferred into triplicate wells of a 96-well plate. For each well, a total 50 μl Reaction Mix containing α-KG Assay Buffer, α-KG Converting Enzyme, α-KG Enzyme Mix, and α-KG Probe were prepared and 50 μl of the Reaction Mix were added to each well. Samples were incubated for 30 min at 37 °C, and then measured OD at 570 nm. The background control was subtracted from the α-KG reading.

Statistical analyses

The Mann–Whitney U-test or Student’s t test was used to analyze the differences between groups, and a p value of less than 0.05 was considered to be significant. Statistical analyses were performed using the SPSS 12.0 software package (SPSS Inc., Chicago, IL, USA).

Results

miR-183 is upregulated in gliomas

First, we examined the expression level of miR-183 in glioblastoma in several different locations, including in the peripheral, non-tumorous brain tissues. Real-time RT-PCR analysis revealed that the expression level of miR-183 in glioblastoma was upregulated in the range of 40- to 320-fold compared to its expression level in peripheral non-tumorous brain tissue (Fig. 1a). Next, to compare miR-183 expression levels between tumor and non-tumorous brain tissue, we examined the levels of miR-183 using six glioblastoma samples and peripheral, non-tumorous brain tissues from the same patients. We observed increased expression of miR-183 in all glioblastoma samples compared to the levels in the peripheral brain tissues (Fig. 1b). We also analyzed the miR-183 expression levels in 88 glioma samples. Real-time RT-PCR analysis revealed that miR-183 expression was higher in the glioma samples compared to the peripheral brain tissue (Fig. 1c). Furthermore, miR-183 expression in high-grade gliomas (grade III–IV gliomas) was significantly higher than in low-grade gliomas (grade I–II gliomas) (Fig. 1d). We then examined the expression level of miR-183 in five glioma cell lines (U251, U87MG, A172, T98G, and SF126) and immortalized human astrocyte cell line (SVGp12), and detected increased expression of miR-183 in A172, SF126, T98G, and SVGp12 cells compared to the peripheral, non-tumorous brain tissue (Supplementary Fig. 1A).
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-012-1027-9/MediaObjects/11060_2012_1027_Fig1_HTML.gif
Fig. 1

miR-183 expression in gliomas. a A comparison of the expression levels of miR-183 from various regions in glioblastoma, including in peripheral, non-tumorous brain tissue. The left panels display images of hematoxylin and eosin staining at each region. The upper right panel depicts the gadolinium-enhanced MRI. Red circles indicate each region analyzed. The lower right panel displays the relative expression levels of miR-183 in each region. The graph indicates the mean ± SD values. b A comparison of the expression levels of miR-183 in six glioblastomas to miR-183 expression in peripheral, non-tumorous brain tissues from the same patients. The graph indicates the mean ± SD values of each sample. Pt, patient. c Relative miR-183 expression in various grades of gliomas. The expression level of miR-183 in the peripheral brain tissue from Fig. 1b served as the control. d The comparison of relative miR-183 expression between low-grade gliomas (grade I and II) and high-grade gliomas (grade III and IV) (**p < 0.01)

miR-183 downregulates the expression of IDH2

We searched for the target of miR-183 using the “miRanda” (http://www.microrna.org/microrna/home.do), “TargetScan” (http://www.targetscan.org/), “PicTar” (http://pictar.mdc-berlin.de/) website, and found that miR-183 contains complementary sequences to the 3′UTR of IDH2 mRNA (Fig. 2a). IDH2 is a mitochondrial enzyme and IDH mutations were recently reported in several types of gliomas. Therefore, we focused on the association between miR-183 and IDH2. To test the IDH2 mRNA levels in each cell line, we performed real-time RT-PCR and compared the results to normal brain tissues. We identified that the expression levels of IDH2 in A172, SF126, T98G, U87, and U251 were 0.27, 0.45, 1.07, 1.60 and 1.83-fold greater, respectively, compared to the expression level in peripheral brain tissues (Supplementary Fig. 1D). This result was somewhat expected based on the expression of miR-183 in these cell lines. Next, we transfected glioma cell lines with miR-183 mimic RNA and analyzed the expression of IDH2 using real-time RT-PCR, western blotting and immunocytochemistry. The expression of IDH2 mRNA was significantly reduced by the overexpression of the miR-183 mimic RNA in four glioma cell lines (U251, U87, T98G, and SF126) and an immortalized human astrocyte cell line (SVGp12) (Fig. 2b). Western blot analysis also showed that the miR-183 mimic RNA downregulated the protein levels of IDH2 relative to the control (Fig. 2c). In addition, immunocytochemical examination also revealed the downregulation of IDH2 protein in T98 cells overexpressing the miR-183 mimic RNA (Fig. 2d). Moreover, transfection with an anti-miR-183 inhibitor resulted in increased expression of IDH2 mRNA compared with the negative control (Fig. 2e).
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-012-1027-9/MediaObjects/11060_2012_1027_Fig2_HTML.gif
Fig. 2

miR-183 inhibits IDH2 expression. a The predicted seed matching between miR-183 and the IDH2 3′UTR sequence. b miR-183 downregulates the expression of IDH2 mRNA. Glioma cells (U251, U87, A172, T98, and SF126) and immortalized human astrocyte cells (SVGp12) were transfected with miR-183 mimic RNA for 48 h, and IDH2 expression levels were analyzed using real-time RT-PCR by the ΔΔ-CT method. The data represent the mean ± SD from three separate experiments performed in triplicate (**p < 0.01). c IDH2 protein expression is downregulated by miR-183. Glioma cells (U87, T98, and A172) were transfected with miR-183 mimic RNA for 72 h and expression levels were analyzed by western blotting. The relative expression levels of IDH2 were obtained using densitometric evaluation of the immunoblots normalized to β-actin. Relative values are displayed below the β-actin panel. d Immunostaining of IDH2 in T98G cells. Cells were transfected with miR-183 mimic RNA for 72 h, and IDH2 expression levels were analyzed by immunocytochemistry. e The expression levels of IDH2 mRNA increased after transfecting with the miR-183 inhibitor RNA. Four glioma cell lines (U251, T98, A172, and SF126) were transfected with an anti-miR-183 inhibitor for 48 h and expression levels were analyzed using real-time RT-PCR by the ΔΔ-Ct method. The data represent the mean ± SD (**p < 0.01)

miR-183 directly targets IDH2

To further evaluate the effect of miR-183’s ability to regulate IDH2, we generated a luciferase reporter plasmid containing the wild-type IDH2 3′UTR, which contains the target region predicted for miR-183 and a luciferase reporter plasmid linked to mutant IDH2 3′UTR (Fig. 3a). As shown in Fig. 3b, miR-183 mimic RNA inhibited the luciferase activity in the luciferase reporter plasmid linked to the wild-type IDH2 3′UTR compared with the luciferase reporter plasmid linked to mutant IDH2 3′UTR. Also, miR-183 mimic RNA inhibited the luciferase activity in the luciferase reporter plasmid linked to the wild-type IDH2 3′UTR compared with the luciferase reporter plasmid without IDH2 3′UTR (Supplementary Fig. 2A). On the other hand, miR-183 inhibitor increased the luciferase activity in the luciferase reporter plasmid linked to the wild-type IDH2 3′UTR compared with the luciferase reporter plasmid linked to mutant IDH2 3′UTR (Fig. 3c). In addition, miR-183 inhibitor increased the luciferase activity compared with negative control (Supplementary Fig. 2B).
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-012-1027-9/MediaObjects/11060_2012_1027_Fig3_HTML.gif
Fig. 3

The 3′UTR of IDH2 mRNA is a direct target of miR-183. a Diagram of luciferase reporter plasmids. Luciferase IDH2 3′UTR vector contains the wild-type IDH2 3′UTR or mutant IDH2 3′UTR. b miR-183 downregulates IDH2 mRNA via the IDH2 3′UTR. Five glioma cell lines (U251, U87, A172, T98, and SF126) were co-transfected with miR-183 mimic RNA and luciferase reporter plasmids containing the IDH2 3′UTR. Luciferase activity was measured 48 h post-transfection. The data represent the mean ± SD from three separate experiments performed in triplicate (*p < 0.05). c Five glioma cell lines were co-transfected with anti-miR-183 inhibitor and luciferase reporter plasmids. The data represent the mean ± SD from three separate experiments performed in triplicate (*p < 0.05). WT wild-type, Mut. mutant, IDH2 IDH2 3′UTR-linked luciferase vector, NC negative control, 183m miR-183 mimic RNA

mRNA and protein expression levels of IDH2 are downregulated in malignant gliomas

To compare the mRNA expression level of IDH2 in glioblastoma with that in peripheral, non-tumorous brain tissue, we performed real-time RT-PCR for IDH2 mRNA using six glioblastoma samples and peripheral brain tissues from the same patients. The number of patients in each group is displayed in Fig. 1b. We observed reduced IDH2 mRNA expression in all glioblastoma samples compared to the level observed in the peripheral brain tissues (Fig. 4a). Next, we performed real-time RT-PCR for IDH2 mRNA using 43 malignant glioma samples (30 cases of glioblastoma and 13 cases of grade III glioma) and five peripheral, non-tumorous brain tissues. The expression level of IDH2 mRNA was decreased in both glioblastomas and grade III gliomas compared to the peripheral, non-tumorous brain tissues (Fig. 4b). Western blotting revealed that the expression level of IDH2 protein was lower in the glioblastoma samples compared to the peripheral brain tissue (Fig. 4c). These results indicate that IDH2 expression is downregulated in malignant gliomas compared to normal brain tissue. To determine the relationship between miR-183 expression and survival, we analyzed miR-183 expression and overall survival time of malignant glioma patients (Grade III, n = 17: Grade IV, n = 53). Kaplan–Meier curves were used to estimate the overall survival between the higher miR-183 group and lower miR-183 group; however, there was no significant difference between these two groups (Supplementary Fig. 3A). Additionally, in Grade IV patients or Grade III patients alone, there was no significant difference between the higher and lower groups either. We tested the relationship between MGMT promotor methylation and miR-183 expression levels in malignant glioma patients. However, the analysis demonstrated that miR-183 expression levels were not significantly different between the methylated group and the non-methylated group (Supplementary Fig. 3B). We analyzed the relationship between IDH mutations and miR-183 expression levels in glioma patients, and the level of miR-183 was significantly higher in gliomas without IDH mutation than that with IDH mutation (Supplementary Fig. 3C).
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-012-1027-9/MediaObjects/11060_2012_1027_Fig4_HTML.gif
Fig. 4

IDH2 expression in malignant glioma samples. a Comparing the expression levels of IDH2 mRNA in glioblastoma to expression levels in peripheral, non-tumorous brain tissues from the same patients. The graph displays the mean ± SD values of each sample. The patient numbers are the same as those presented in Fig. 1b. Pt, patient. b A comparison of the expression levels of IDH2 mRNA in grade III glioma (n = 13), grade IV glioma (n = 30), and peripheral, non-tumorous brain tissues (n = 5). The graph indicates the mean ± SD values of each group (**p < 0.01). c Western blotting for IDH2 in four patients with glioblastoma. IDH2 protein levels are lower in glioblastoma tissues compared with peripheral brain tissue from the same patient. The number of patients is equal to that in Fig. 1b. P and T indicate peripheral brain tissue and tumor, respectively. Pt, patient

miR-183 overexpression elevates the expression of HIF-1α

Because IDH2 has the ability to convert isocitrate to α-ketoglutarate (α-KG) in cells, the downregulation of IDH2 by miR-183 may reduce the levels of α-KG. To determine whether miR-183 downregulates the levels of α-KG in glioma cells, the cells were transfected with miR-183 mimic RNA and analyzed for their cellular α-KG levels. In U87 and T98 cells, α-KG levels were significantly reduced in miR-183 transfected cells, and α-KG levels were moderately reduced in A172 cells transfected with miR-183 (Fig. 5a). α-KG is a substrate of prolyl hydroxylases (PHDs), which downregulate HIF-1α. In addition, recent studies revealed that a deficiency of IDH1 induces the expression of HIF-1α in glioma cells [33]. Therefore, we analyzed HIF-1α protein expression levels by performing western blots on protein from glioma cells transfected with miR-183 mimic RNA. HIF-1α protein levels were increased following the overexpression of miR-183 (Fig. 5b). Next, to examine whether the inhibition of IDH2 induces the upregulation of HIF-1α, we transfected glioma cells with siRNA targeting IDH2 and performed a western blot. The level of HIF-1α protein was upregulated in glioma cells transfected with IDH2 siRNA (Fig. 5c). These results indicate that the knockdown of IDH2 increases the expression of HIF-1α protein. To test whether miR-183 regulates HIF-1α mRNA levels, we transfected glioma cell lines with miR-183 mimic RNA and analyzed the expression of HIF-1α using real-time RT-PCR. We observed significantly decreased levels of IDH2 and increased levels of HIF-1α in all glioma cell lines transfected with miR-183 mimic RNA compared with those transfected with control RNA (Fig. 5d). The mRNA levels of HIF-1α were elevated in six glioblastoma samples compared to the peripheral brain tissues taken from the same patients (Fig. 5e). Additionally, real-time RT-PCR revealed that the mRNA transcription levels of vascular endothelial growth factor (VEGF) and glucose transporter 1 (GLUT1) were increased in glioma cells transfected with the miR-183 mimic RNA compared with those transfected with control RNA (Fig. 5f). These results indicate that the overexpression of miR-183 induces the upregulation of HIF-1α in glioma cells.
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-012-1027-9/MediaObjects/11060_2012_1027_Fig5_HTML.gif
Fig. 5

Elevated expression of HIF-1α after transfection with miR-183 mimic RNA or siRNA targeting IDH2. a miR-183 reduces α-ketoglutarate levels in glioma cells. Glioma cells (U87, T98, and A172) were transfected with miR-183 mimic RNA or control RNA for 48 h, and the expression level of α-ketoglutarate was analyzed. The data represent the mean ± SD from three separate experiments performed in triplicate (**p < 0.01). b miR-183 upregulates the protein levels of HIF-1α. Glioma cells were transfected with miR-183 mimic RNA or control RNA for 72 h and expression levels were analyzed by western blotting. The relative expression levels of IDH2 were obtained using densitometric evaluation of immunoblots normalized to β-actin. The relative values are displayed below each panel. The experiments were performed in triplicate. c Knockdown of IDH2 by siRNA leads to the elevated expression of HIF-1α protein. Glioma cells were transfected with IDH2 siRNA or control siRNA for 72 h and expression levels were analyzed by western blotting. The relative values are displayed below the panel. The experiments were performed in triplicate. d Glioma cells were transfected with miR-183 mimic RNA or control RNA for 48 h and expression levels were analyzed using real-time RT-PCR by the ΔΔ-Ct method. The data represent the mean ± SD from three separate experiments performed in triplicate (*p < 0.05, **p < 0.01). e mRNA levels of HIF-1α were increased in six glioblastoma tissues compared to the respective peripheral brain tissues in the same patients. The patient numbers are the same as in Figs. 1b and 4a. The data represent the mean ± SD from three separate experiments. f miR-183 upregulates the mRNA levels of VEGF and GLUT1. Glioma cells were transfected with miR-183 mimic RNA or control RNA for 48 h and expression levels were analyzed using real-time RT-PCR by the ΔΔ-CT method. The data represent the mean ± SD from three separate experiments performed in triplicate (**p < 0.01)

Discussion

As noted above, we found that miR-183 is upregulated in a majority of malignant gliomas. Additionally, the overexpression of miR-183 inhibits the expression of IDH2, which is one of the mitochondrial enzymes in glioma cells related to the TCA cycle. Additionally, we determined that IDH2 expression was downregulated in a majority of malignant gliomas. Moreover, we demonstrated that the overexpression of miR-183 increased both the protein and mRNA levels of HIF-1α in glioma cells. The aberrant expression of miR-183 in various cancers, such as lung cancer, colon cancer, hepatocarcinoma, and medulloblastoma have been reported, suggesting that there is a close correlation between miR-183 and cancer development [3436]. It is reported that miR-183 regulates the expression of ezrin, integrin β1, and kinesin 2α and affects tumor cell invasion and migration [19, 37]. Additionally, miR-183 inhibits TGF-β-1-induced apoptosis via the downregulation of PDCD4 expression in human hepatocellular carcinoma cells [18]. Our study demonstrates that miR-183 upregulated HIF-1α in glioma cells, which may in turn contribute to glioma formation or survival.

Interestingly, we identified a direct relationship between miR-183 and IDH2 in glioma cells. To our knowledge, this is the first report demonstrating the interaction between miR-183 and IDH2 in glioma cell lines. Although the expression of miR-183 in U87 and U251 is significantly lower than in A172 and SF126, miR-183 mimic RNA has very similar effects on IDH2 expression. This may be because the expression levels of IDH2 mRNA are different in each cell line. Moreover, because the miR-183 mimic RNA also reduced the level of IDH2 mRNA in human astrocyte cell line SVGp12, IDH2 inhibition by miR-183 is not a tumor cell specific phenomenon.

Recent studies have reported a high frequency of IDH1 mutations in gliomas. However, compared to IDH1, IDH2 mutations are much less common in gliomas; however, IDH2 mutations are more common than IDH1 mutations in acute myeloid leukemia [38]. Although IDH2 mutations are rare in gliomas, our analysis demonstrated that reduced IDH2 expression is frequently observed in malignant glioma. Therefore, we considered that the increased expression of miR-183 reduces the level of IDH2 in glioma cells. Very recently, Vohwinkel et al. [39] reported that elevated CO2 concentrations (hypercapnia) lead to increased levels of miR-183, which downregulates IDH2 and impairs mitochondrial function and cell proliferation in fibroblasts and alveolar epithelial cells.

IDH has the ability to convert isocitrate to α-KG. The present study revealed that the downregulation of IDH2 by miR-183 reduces the cellular levels of α-KG in glioma cells (Fig. 5a). Prolyl hydroxylases (PHDs), which use α-KG as substrates, hydroxylate HIF-1α to promote its proteasomal degradation in the presence of oxygen [40, 41]. Zhao et al. [33] previously reported the association between cytoplasmic IDH1 and HIF-1α . The authors revealed that mutant IDH1 dominantly inhibited wild-type IDH1 resulting in a decrease in cellular levels of α-KG, which causes an elevation in HIF-1α protein levels. Based on this mechanism, we assumed that a miR-183-mediated reduction in IDH2 might elevate the expression of HIF-1α protein. In addition, our data revealed that miR-183 elevates not only HIF-1α protein but also HIF-1α mRNA levels; however, the mechanism of this is unclear. To elucidate this phenomenon, further study is necessary.

Although the mechanism for how miR-183 affects HIF-1α is not completely understood, we provide evidence that the overexpression of miR-183 increases HIF-1α expression. HIF-1α activates critical pro-angiogenic growth factors such as vascular endothelial growth factor (VEGF) and erythropoietin. Furthermore, HIF-1α stimulates glucose uptake by regulating the expression of the glucose transporters GLUT1 and GLUT3 and regulates a number of glycolysis-related genes [42]. Moreover, HIF-1α also plays a key role in suppressing mitochondrial activity by inducing pyruvate dehydrogenase-kinase1 (PDK1), which inactivates pyruvate dehydrogenase (PDH) [43]. PDH provides the link between glycolysis and the TCA cycle by catalyzing the conversion of pyruvate into acetyl-CoA. As a result, the inactivation of PDH leads to suppression of both the TCA cycle and mitochondrial respiration.

Tumor cells obtain energy by using aerobic glycolysis, and they exhibit a concomitant defect in mitochondrial respiration. In addition, tumor cells require ongoing lipid synthesis. Fatty acids are essential components of all biological lipid membranes and are critical substrates for energy metabolism. Under aerobic conditions in tumor cells, cell metabolism shifts from an oxidative to a glycolytic mode, which leads to an increase in the mitochondrial concentration of citrate [44]. Citrate is an essential component for fatty acid synthesis [45, 46]. Therefore, we speculate that the downregulation of IDH2 by miR-183 overexpression might induce mitochondrial dysfunction and de novo fatty acid synthesis.

In conclusion, the major findings presented in this study are that miR-183 directly targets the 3′UTR of IDH2 mRNA in glioma cells, which induces the downregulation of IDH2 and the upregulation of HIF-1α. Because HIF-1α plays a significant role in angiogenesis, metabolism and survival in tumor cells, the overexpression of miR-183 might contribute to gliomagenesis. Further studies are necessary to clarify the association between miR-183 and tumor formation, growth, and metabolism, and the mechanisms responsible for the upregulation of HIF-1α by miR-183.

Acknowledgments

We thank Mitsuharu Endo and Ryosuke Doi (Department of Physiology and Cell Biology, Kobe University, Japan) for helping with the luciferase reporter assays. We also thank Mariko Ueda for helping with the immunocytochemical analysis. This work was supported in part by a Grant-in-Aid for Scientific Research to Takashi Sasayama (22591610) and Katsu Mizukawa (22791344) from the Japanese Ministry of Education, Culture, Sports, Science and Technology.

Conflicts of interest

None declared.

Supplementary material

11060_2012_1027_MOESM1_ESM.tif (300 kb)
Supplementary material 1 (TIFF 299 kb)
11060_2012_1027_MOESM2_ESM.tif (116 kb)
Supplementary material 2 (TIFF 115 kb)
11060_2012_1027_MOESM3_ESM.tif (100 kb)
Supplementary material 3 (TIFF 100 kb)
11060_2012_1027_MOESM4_ESM.docx (13 kb)
Supplementary material 4 (DOCX 13 kb)

Copyright information

© Springer Science+Business Media New York 2012