MicroRNA-153 regulates glutamine metabolism in glioblastoma through targeting glutaminase

Abstract

Glioblastoma is the most aggressive manifestation of malignant gliomas and considered to be among the deadliest forms of human cancers. MicroRNAs are found to tightly regulate diverse biological processes and considered to play important roles in cancer etiology. In this study, we found that microRNA-153 was significantly downregulated in glioblastoma tissues compared to matched non-tumor tissues and in glioblastoma cell lines. To investigate the potential function of microRNA-153 in glioblastoma, we transfected glioblastoma cell line U87MG as well as U373MG with synthetic microRNA-153 oligos and observed decreased cell proliferation and increased apoptosis. We further found that microRNA-153 restrained glutamine utilization and glutamate generation. Bioinformatics analysis revealed that glutaminase, which catalyzed the formation of glutamate from glutamine, is the potential target of microRNA-153. Indeed, microRNA-153 cannot further reduce glutamine utilization when glutaminase was knocked down. Overexpression of glutaminase abrogates the effect of microRNA-153 on glutamine utilization. Furthermore, the relative expression of microRNA-153 and glutaminase in glioblastoma versus matched non-tumor tissues showed a reverse correlation, further indicating that microRNA-153 may negatively regulate glutaminase in vivo. These results demonstrate an unexpected role of microRNA-153 in regulating glutamine metabolism and strengthen the role of microRNA-153 as a therapeutic target in glioblastoma.

Introduction

MicroRNAs (miRNAs) are an abundant class of non-coding RNA molecules of about 21–24 nt in length that are critical in maintaining cellular homeostasis post-transcriptionally. miRNAs have been highly conserved during evolution in both plants and animals.^1–3 Once in their mature form, miRNAs imperfectly pair with complementary sites within the 3′-untranslated regions (3′-UTRs) of messenger RNAs (mRNAs) and either degrade or inhibit translation of multiple transcripts.^4–6 Since their discovery in the 1990s,^7,8 thousands of miRNAs have been shown to play roles in regulation of various important biological processes including cancers, triggering great interest in studying miRNA dysregulations in human diseases.^3,9,10

Glioblastoma (GBM) is the most frequent primary brain tumor in adults. Although the incidence rate of GBM is relatively low compared with other cancers, patients with malignant GBM have a very poor prognosis and are with a high mortality rate. ¹¹ Much effort has been placed into understanding the molecular basis of GBM and generating promising pharmacological targets. Over a decade, the interplay between the protein-coding and non-protein-coding genome, especially the miRNAome (microRNA genome), has been the most exciting yet unexpected discovery in oncology.

It is reported that miRNA dysregulation could play important roles in GBM development and progression. In 2005, the first miRNA, miR-21-5p, was identified in GBM. ¹² miR-21-5p was overexpressed in GBM tissues when compared with non-neoplastic human control tissues. Targeting miR-21-5p led to decreased cell number and increased apoptosis.^12,13 However, the complete patterns of miRNAs regulated and involved in GBM remain to be fully elucidated.

MicroRNA-153 (miR-153) has been identified as a tumor suppressor in some cancers such as hepatocellular carcinoma, ¹⁴ osteosarcoma, ¹⁵ and non–small cell lung cancer. ¹⁶ It is reported that miR-153 is highly expressed in brain while its expression is significantly downregulated in GBM compared with normal brains, which suggests that miR-153 might play an important part in the development of GBM. Indeed, some studies showed that miR-153 could restrain cell proliferation and promotes apoptosis by suppressing the expression of B-cell lymphoma 2 (Bcl-2), myeloid cell leukemia sequence 1 (Mcl-1), and insulin receptor substrate-2 (Irs-2) in GBM cell lines.^17,18 However, the role of miR-153 in the metabolism control of GBM cells remains to be fully defined.

Cancer cells call for abnormal glutamine metabolism to benefit their growth, proliferation, and survival. ¹⁹ Glutaminase (GLS) catalyzes the formation of glutamate from glutamine, which is called glutaminolysis. Some cancer cell lines display increased glutamine uptake and catabolism to fuel the tricyclic acid (TCA) cycle that renders cells addicted to glutamine.

Here, we demonstrated the critical role of miR-153 in GBM metabolism especially glutamine metabolism. We found that miR-153 was downregulated in GBM tissues compared with matched non-tumor tissues and GBM cell lines.

miR-153 overexpression inhibited cell growth and promoted apoptosis of GBM cells. We further found that miR-153 restrained glutamine utilization and glutamate generation through directly targeting GLS expression. Overexpression of GLS completely abrogated the effects of miR-153 on GBM cells. Moreover, a reverse correlation was observed between relative GLS and miR-153 expression in human GBM tissues compared with matched non-tumor tissues. These findings bring us new perspectives in the regulation of GBM glutamine metabolism by miR-153 and strengthen the role of miR-153 as a therapeutic target in GBM.

Material and methods

Clinical specimens

GBM specimens, including adjacent non-tumor tissues, were obtained from the Affiliated Bayi Brain Hospital. The experiments were undertaken with the understanding and written consent of each subject. The study methodologies conformed to the standards set by the Declaration of Helsinki. The Institutional Ethics Committee approved the study protocol and use of clinical specimen. The specimens were obtained after surgical resection, immediately frozen, and stored in liquid nitrogen.

Cell lines

All cell lines were obtained from the American Type Culture Collection and cultured according to standard protocols and sterile techniques. Human GBM cell lines U87MG and U373MG were maintained in MEM media (Invitrogen, Carlsbad, USA) and supplemented with 10% (v/v) fetal bovine serum, 100 U/mL penicillin, and 100 mg/mL streptomycin. Cell culture was conducted at 37°C in a humidified 5% CO₂ incubator. MiR-153 mimics was purchased from Sigma–Aldrich (St. Louis, USA). Small interfering RNA (siRNA) targeting GLS was synthesized from Thermo Fisher Scientific (Dreieich, Germany).

Quantitative reverse transcription polymerase chain reaction

Total RNA was isolated from tissues and cell lines using the TRIzol reagent (Invitrogen). Expression of miR-153 was detected by quantitative real-time polymerase chain reaction (RT-PCR)-based TaqMan® MicroRNA Assay (Applied Biosystems, Foster City, USA) by following the manufacturer’s instructions. U6 small nuclear RNA (snRNA) was used as an endogenous control. For GLS measurement, RNA was reverse-transcribed into complementary DNA (cDNA) using Kit from Thermo Fisher Scientific. Forward primer: 5′-AGGGTCTGTTACCTAGCTTGG-3′ and reverse primer: 5′-ACGTTCGCAATCCTGTAGATTT-3′ were used for GLS measurement. Inner-control β-Actin was measured by forward primer: 5′-AGAGCTACGAGCT GCCTGAC-3′ and reverse primer: 5′-AGCACTGTGTTGGCGTACAG-3′. SYBR green mix was from Thermo Fisher Scientific. The results were determined using the 7500 Fast System SDS software (Applied Biosystems). The delta-delta-Ct (ddCt) algorithm was used to calculate the relative quantification.

In situ hybridization

In situ hybridization was performed using the digoxigenin-labeled locked nucleic acid (LNA) probes of miR-153 (Exiqon, Vedbaek, Denmark). The probes were visualized using a horseradish peroxidase (HRP)-conjugated anti-digoxigenin antibody (Abcam, Cambridge, USA) followed by enzymatically reaction with 3,3′-diaminobenzidine (DAB) substrate. The nuclei were further stained by hematoxylin. The signal was evaluated by assessing staining intensity using a BX51 microscope (Olympus, Tokyo, Japan).

Immunohistochemistry

Immunohistochemistry was performed according to standard protocols. The tissue sections were incubated with anti-GLS antibody (Sigma–Aldrich; 1:200 dilution) overnight at 4°C followed by incubation with the HRP-conjugated secondary antibody (1:100 dilution). The signal was visualized using DAB substrate. The nuclei were further stained by hematoxylin. The signal was evaluated by assessing staining intensity using a BX51 microscope (Olympus).

Cell growth assay

MTT assay was used to measure the cell growth. Cells were seeded in a 96-well plate at a density of 1.5 × 10³ cells per well. MTT solution (20 µL; Promega, Madison, USA) was incubated with cells at 37°C for 2 h. The absorbance values were determined in a microplate reader (Calculator, Pharmacia Biotech, Uppsala, Sweden) at a wavelength of 490 nm. Experiments were performed three times. The growth curves were generated by SigmaPlot graphing software (Systat Software Inc., San Jose, USA).

Glutamine and glutamate content assay

Glutamine was detected using a kit from Biovision (K556-100, California, USA). Glutamate was detected using a kit from Sigma–Aldrich (MAK004). The assays were done according to the manufacturer’s manual. The results were measured using a Multiskan™ FC Microplate Photometer (Thermo Fisher Scientific).

Luciferase assay

Dual-luciferase reporter assay was performed in 293T and U87MG cells by following the manufacturer’s instructions (Promega). Cells were co-transfected with pmirGLO vector inserted with wild-type (WT-3′-UTR) or mutant (Mut-3′-UTR) 3′-UTR, along with miR-153 mimics or control miRNA. Forward primer: 5′-GCTCTAGATGGTCTCAAATCCCAAGATT-3′ and reverse primer: 5′-GCGTCGACGATACATATATTTATTATGC-3′ were used to clone the 3′-UTR of GLS into vector. 5′-AGCTGGCATAGTTTCCTGTTTTTATGTTTTCCATAG-3′ and 5′-CAGGAAACTATGCCAGCTCTCACTT-3′ were further used to clone mutant 3′-UTR of GLS. Dual-luciferase assay was performed in cells 48 h after transfection. Luciferase activity was measured using a Victor Luminometer (PerkinElmer, Waltham, USA). The firefly luciferase activity was normalized using co-transfected Renilla luciferase for transfection efficiency.

Immunoblotting

To prepare total protein, cells were collected and lysed and then subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS/PAGE) and transferred onto polyvinylidene difluoride (PVDF) membrane. The membranes were incubated with a primary antibody (anti-GLS, Sigma–Aldrich; anti-β-actin, Santa Cruz Biotechnology, Dallas, Texas). After incubation with appropriate HRP-conjugated secondary antibody (Santa Cruz Biotechnology), the membranes were treated with an enhanced chemiluminescence reagent (Thermo Fisher Scientific) and exposed to X-ray film (Kodak, Rochester, USA).

Statistical analysis

Data were analyzed using the SPSS software (SPSS Inc., Chicago, USA). Quantitative data were presented as mean ± standard deviation. Differences between miR-153 expressions in tumor tissues and adjacent non-tumor tissues were analyzed by Wilcoxon’s matched pairs test. The correlation was identified by Pearson’s test. Statistical differences between groups were determined by Student’s t-test. Differences were considered significant when p < 0.05.

Results

MiR-153 is downregulated in GBM

It is reported that a number of miRNAs are involved in GBM development. To examine the potential role of miR-153 in GBM development, we first analyzed mRNA expression levels of miR-153 in 30 pairs of GBM tissues and matched non-tumor tissues by quantitative polymerase chain reaction (qPCR) analysis. Results showed that the mRNA levels of miR-153 were significantly downregulated in GBM tissues (Figure 1(a)). MiR-153 protein expression in the 30 paired tissues was also analyzed by in situ hybridization. Accordingly, the collective data indicated that the protein levels of miR-153 were dramatically decreased in non-tumor tissues (Figure 1(b)). The representative staining was shown in Figure 1(c). qPCR analysis was further carried out to test the level of the miR-153 in five GBM cell lines such as U87MG, T98G, U373MG, LN299, and U138MG. Normal human astrocyte (NHA) cells were used as the control cell line. The expression of miR-153 was lower in all five GBM cell lines compared to NHA cell line (Figure 1(d)). These results suggested that miR-153 expression was downregulated in GBM tissues and cell lines, suggesting miR-153 might be involved in GBM tumor progression.

Figure 1.

MiR-153 is downregulated in glioblastoma: (a) miR-153 is downregulated in glioblastoma tissues compared to non-tumor tissue. In total, 30 paired tissue samples were analyzed by qPCR. (b) MiR-153 expression in the 30 paired tissues was analyzed by in situ hybridization. The staining intensity of miR-153 in tumor tissue compared to the matched non-tumor tissue was shown. (c) The representative staining figures in tumor tissue and matched non-tumor tissue were shown. (d) MiR-153 is downregulated in glioblastoma cell lines compared to normal human astrocyte cells.

MiR-153 inhibits GBM cell growth

To find out the potential role of miR-153 in GBM tumor progression, U87MG as well as U373MG cell line was adopted. MiR-153 was successfully overexpressed in the above two cell lines by transfecting miR-153 mimics (Figure 2(a)). Methylthiazolyldiphenyl-tetrazolium bromide (MTT) assay was performed to test whether miR-153 would affect GBM cell proliferation in the following 6 days after miR-153 overexpression. The proliferation was markedly suppressed in both cell lines transfected with miR-153 (Figure 2(b)). The effect of miR-153 on apoptosis was also examined. Cleaved caspase-3 is a hallmark of apoptosis. MiR-153 overexpression significantly promoted cleaved caspase-3 expression (Figure 2(c)), suggesting miR-153 may also be involved in GBM cell apoptosis. We also found that the growth inhibitory effect of miR-153 on GBM cells was eliminated with glutamine but not glucose deprivation (Figure 2(d)). Accordingly, the upregulated cleaved caspase-3 expression was also abolished by glutamine deprivation (Figure 2(e)). These results suggested that glutamine metabolism may be involved in the function of miR-153 in GBM cell growth inhibition and apoptosis promotion.

Figure 2.

MiR-153 inhibits glioblastoma cell growth: (a) validation of MiR-153 increase induced by miR-153 mimics, (b) miR-153 mimics inhibits glioblastoma cell proliferation in U87MG and U373MG cells, (c) miR-153 mimics increases glioblastoma cell apoptosis, (d) glutamine deprivation eliminated the growth inhibitory effect of miR-153 on glioblastoma cells, and (e) caspase-3 activation induced by miR-153 was inhibited by glutamine deprivation.

MiR-153 functions through regulating glutamine metabolism

Glutamine metabolism is recognized as a distinctive feature in addition to Warburg effect in cancer. Cancer cells acquire energy and nutrients through glutamine metabolism to support cell growth and proliferation. Several oncogenes and anti-oncogenes in cancers have been reported to influence glutamine metabolism. We next evaluated whether miR-153 played a role in glutamine metabolism. U87MG as well as U373MG cell was transfected with miR-153. Glutamine and glutamate levels were determined. Results showed that miR-153 overexpression increased remaining glutamine while reduced glutamate generation in both cell lines (Figure 3(a) and (b)). Remaining glutamine and glutamate generation were controlled by glutamine uptake and GLS activity, respectively. Glutamine uptake mediated by glutamine transporters SLC1A5 could be blocked using SLC1A5 inhibitor γ-l-glutamyl-p-nitroanilide (GPNA). The small-molecule GLS inhibitor bis-2-(5-phenylacetamido-1,2,4-thiadiazol-2-yl)ethyl sulfide (BPTES) treatment could inhibit GLS activity, lowered glutamate and α-KG levels. Results showed that GPNA or BPTES treatment abrogated the effects of miR-153 on the remaining glutamine content and glutamate generation (Figure 3(c) and (d)). These results suggested that miR-153 functioned through regulating glutamine metabolism.

Figure 3.

MiR-153 functions through regulating glutamine metabolism: (a) miR-153 mimics increases remaining glutamine in glioblastoma cells, (b) miR-153 mimics reduces glutamate generation in glioblastoma cells, (c) blocking glutamine uptake abrogates the effect of miR-153 on glutamine metabolism, and (d) suppressing GLS activity abrogates the effect of miR-153 on glutamine metabolism.

MiR-153 targets GLS

We next try to identify the exact role of miR-153 in glutamine metabolism. Computational analysis results revealed that miR-153 targets sequences in 3′-UTR of GLS (Figure 4(a)). Then, GLS 3′-UTR sequence and mutant GLS-3′-UTR sequence that including several mutations in miR-153 binding site were both cloned downstream of the luciferase reporter gene (Figure 4(a)). The dual-luciferase reporter assay was performed in 293T cells. Results showed that miR-153 reduced the GLS-3′ UTR but not GLS-3′ UTR-mut luciferase levels (Figure 4(b)). These data suggest that GLS is the target of miR-153. We also found that both the mRNA and protein levels of GLS were suppressed by miR-153 (Figure 4(c) and (d)). In addition, blocking miR-153 with anti-miR-153 reversed the downregulated expression of GLS (Figure 4(c) and (d)). These results suggested that miR-153 modulated GLS expression by directly targeting the 3′-UTR of GLS mRNA.

Figure 4.

MiR-153 targets GLS: (a) schema of binding site of miR-153 in GLS 3′-UTR region, (b) GLS 3′-UTR luciferase activity regulated by miR-153, (c) miR-153 reduces endogenous GLS protein levels in glioblastoma cells, and (d) miR-153 reduced GLS mRNA levels in glioblastoma cells.

Overexpression of GLS abrogates the effects of miR-153 on GBM cells

To further validate GLS was the target of miR-153, we test the function of miR-153 on glutamine metabolism by knockdown or overexpression of GLS. We first used siRNA to knock down GLS expression in U87MG and U373MG cells. Results showed that siRNA was efficient in lowering GLS expression (Figure 5(a)). GBM cell growth on day 4 was obviously suppressed when knockdown of GLS, while the cell growth on day 1 was not affected (Figure 5(b)). MiR-153 significantly decreased glutamine utilization when co-transfected with control siRNA in both U87MG and U373MG cells (Figure 5(c)). However, miR-153 cannot further reduce glutamine utilization when co-transfected with GLS siRNA further strengthens the idea that GLS was the target of miR-153 (Figure 5(c)). Moreover, we successfully overexpressed GLS in U87MG and U373MG cells (Figure 5(d)). Accordingly, the decreased effects of miR-153 were abrogated in GLS-overexpression GBM cells (Figure 5(e)). All these data suggested that miR-153 regulates glutamine metabolism through targeting GLS.

Figure 5.

Overexpression of GLS abrogates miR-153 effect on glioblastoma cells: (a) validation of GLS inhibition induced by siGLS, (b) knockdown of GLS inhibits glioblastoma cell growth, (c) miR-153 cannot further reduce glutamine utilization in GLS knocked-down glioblastoma cells, (d) validation of GLS overexpression in glioblastoma cells, and (e) overexpression of GLS abrogates the effect of miR-153 on glutamine utilization.

Relative GLS and miR-153 expression shows a reverse correlation in GBM

To further validate the critical role of GLS in GBM, we then analyzed expression of GLS mRNA in 30 pairs of GBM tissues and matched non-tumor tissues by qPCR analysis. Results showed that the mRNA levels of GLS were significantly promoted in GBM tissues compared to matched non-tumor tissues (Figure 6(a)). The protein levels of GLS expression in GBM tissues were examined by immunohistochemistry staining. GLS was dramatically upregulated in GBM tissues (Figure 6(b)). The relative intensity of GLS protein in GBM tissues was calculated and shown in Figure 6(c). Notably, a reverse correlation was observed between relative GLS and miR-153 expressions (Figure 6(d)). These results suggested that miR-153 may negatively regulate GLS in vivo.

Figure 6.

Relative GLS and miR-153 expression shows a reverse correlation in glioblastoma: (a) GLS is upregulated in glioblastoma. The 30 paired glioblastoma tissue samples were analyzed by qPCR. (b) GLS protein levels are upregulated in glioblastoma. The representative staining figures in tumor tissue and matched non-tumor tissue were shown. (c) GLS protein levels are upregulated in glioblastoma (intensity). (d) A reverse correlation between miR-153 and GLS in glioblastoma.

Discussion

Cell growth relies on coordination of multiple fluxes involving glucose, glutamine, several nonessential amino acids, and the cellular one-carbon pool. Glutamine, which is highly transported into proliferating cells, ²⁰ is a major source of energy and nitrogen for biosynthesis, and a carbon substrate for anabolic processes in cancer cells. Here, we reported that miR-153 regulated glutamine metabolism in GBM through targeting GLS. Bioinformatics analysis revealed that miR-153 directly targeted 3′-UTR sequence of GLS. Luciferase assay results confirmed the results. MiR-153 specifically reduced the GLS-3′-UTR but not GLS-3′-UTR-mut luciferase levels. Functionally, knockdown of GLS cannot further reduce glutamine generation induced by miR-153. To the contrary, overexpression of GLS efficiently abolished the effects of miR-153, further confirming that miR-153 functioned through targeting GLS.

It has been investigated that oncogene c-Myc could promote tumor cell growth and cell-cycle entry, with the strict demand for glutamine in order to maintain viability. ²¹ c-Myc promoted glutamine uptake through increasing the expression of glutamine transporters SLC1A5 and SLC7A5/SLC3A2. ²² In addition, c-Myc could facilitate glutaminolysis by increasing the expression of GLS. c-Myc promoted GLS expression by suppressing GLS repressor miR-23a/b in human P-493 B-lymphoma cells and human leukemic Jurkat cells.^23,24 miRNA-153 displayed low levels in GBM tissues and cell lines. It is possible that c-Myc may promote GLS expression in GBM by negatively regulating the GLS-targeting miRNA-153 in GBM, which further enhances the utilization of glutamine.

It is anticipated that functional miRNAs could be available as therapeutic targets in clinical translation as an alternative approach. let-7a is aberrantly expressed in GBM and has been linked to the regulation of cell growth and glucose metabolism in GBM through directly targeting c-K-ras (KRAS) protein.^25–27 MiR-7-5p, ²⁸ miR-128-3p, ²⁹ miR-491-5p, ³⁰ and miR-218-5p, ³¹ negatively regulating the expression of epidermal growth factor receptor (EGFR) in human GBM, were shown to be downregulated in GBM tissues and/or cell lines to drive gliomagenesis. Three upregulated miRNAs, miR-23a-3p, ³² miR-26a-5p, ³³ and miR-17-5p, ³⁴ directly target phosphatase and tensin homolog (PTEN) and promote excessive proliferation in GBM. miRNA dysregulation could play important roles in GBM development and progression.

Although a few miRNA-based therapeutics are trialed for clinical significance, such as Miravirsen in hepatitis C virus infection (LNA-antimiR-122-5p, phase IIa, NCT01200420), no clinical trial on miRNA intervention is tailored to GBM. ³⁵ Considering the utilization of such molecularly targeted drugs only received modest clinical benefits, there is an urgent need to bring miRNA therapeutics to the frontline either alone or in combination.

Taken together, we identified miRNA-153 in GBM tissues with low levels. MiR-153 mimics inhibits GBM cell proliferation in U87MG and U373MG cells. Glutamine deprivation eliminated the growth inhibitory effect of miR-153 on GBM cells. Mechanically, miRNA-153 restrained the utilization of glutamine and glutamate generation by directly targeting GLS expression. Moreover, a reverse correlation was observed between relative GLS and miR-153 expressions in vivo. Our study provides a new perspective in the role of miR-153 in GBM glutamine metabolism and strengthens the role of miR-153 as a therapeutic target in GBM.