It has become increasingly clear that there are notable parallels between normal development and tumorigenesis. Glioma is a classic model that links between tumorigenesis and development. We evaluated the expression of GRIM-19, a novel gene essential for normal development, in various grades of gliomas and several human glioma cell lines. We showed that GRIM-19 mRNA and protein expression were markedly lower in gliomas than in control brain tissues and negatively correlated with the malignancy of gliomas. Downregulation of GRIM-19 in glioma cells significantly enhanced cell proliferation and migration, whereas overexpression of GRIM-19 showed the opposite effects. We also showed that the activation of signal transducer and activator of transcription 3 (STAT3) and the expression of many STAT3-dependent genes were regulated by the expression of GRIM-19. In addition, GRIM-19 exerted its role probably through the non-STAT3 signaling pathway. Collectively, our data suggest that most gliomas expressed GRIM-19 at low levels, which may play a major role in tumorigenesis in the brain. (Cancer Sci 2011; 102: 1991–1999)
Gliomas represent the most frequent primary tumor of intracranial neoplasms and are graded as low-grade I and II and high-grade III and IV according to their histopathological and clinical features.(1,2) Unfortunately, low-grade gliomas will inevitably progress toward malignant high-grade gliomas.(3) Despite the maximal therapies, patients bearing high-grade gliomas still have a median life expectancy of <1 year.(4) Therefore, it is vital to develop new therapies against this devastating and invariably fatal disease.
Accumulating evidence has shown that there are notable similarities between normal development and tumorigenesis.(5,6) For example, the Sonic hedgehog–Patched signaling pathway plays a critical role in regulating growth in normal cerebellum development and is also a major target of mutation in the cerebellar tumor medulloblastoma.(7)NF1, a tumor suppressor gene, is important for the development of the central nervous system (CNS), and its inactivation has been involved in glioma formation.(8,9) Based on these findings, genes that are involved in normal developmental processes, if dysregulated, might contribute to tumorigenesis.
GRIM-19, a novel gene, was initially identified in β-interferon (IFN-β) and all-trans-retinoic acid (RA)-induced tumor cell death pathway using the antisense knockout technique.(10) It does not belong to any of the known cell death gene categories, such as the Bcl-2, caspase, and death receptor families.(11) GRIM-19 is mainly located in mitochondria, with only a small fraction in nucleus.(12) It has multiple functions in cell growth, embryonic development, and mitochondria.(12–14) We and others have shown that high levels of GRIM-19 induced apoptosis, but moderate expression caused growth retardation.(10,12) Furthermore, we found that homologous deletion of GRIM-19 led to embryonic lethality, suggesting that GRIM-19 is essential for normal embryonic development.(15)
Although dysregulated expression of GRIM-19 has been found in some tumors such as renal carcinoma, thyroid tumor, and cervical cancers, its underlying mechanisms in tumorigenesis are still poorly investigated.(16–18) Here, we examined mRNA and protein levels of GRIM-19 in human glioma and normal brain tissues. Additionally, we investigated the roles of GRIM-19 in human malignant glioma (MG) cells and the possible mechanisms of GRIM-19 in tumorigenesis.
Materials and Methods
Tumor samples. A total of 58 glioma specimens from patients with primary gliomas and five normal brain tissues from trauma patients were collected during surgical procedures at the Department of Neurosurgery, Qilu Hospital, Shandong University (Jinan, China). Collection was carried out in accordance with the National Regulation of Clinical Sampling in China.
Cell culture and transfection. Human glioma cell lines U251MG, U87MG, and A172 were maintained in DMEM with high glucose (Hyclone, Logan, UT, USA) and 10% FBS (Hyclone). The U251MG cells were exposed in the culture medium to either vehicle (culture medium without serum) or interleukin (IL)-6 (50 ng/mL; R&D Systems, Minneapolis, MN, USA), or a combination of IFN-β/RA (750 U/mL IFN-β, Millipore, Billerica, MA, USA; 1.5 μM RA, Sigma, St. Louis, MO, USA) followed by IL-6 (50 ng/mL). The GRIM-19 expression plasmids HA-PXJ40-GRIM-19, HA-pXJ40, Psuper.neo.GFP.KD-GRIM-19 (GRIM-19 siRNA), Psuper.neo.GFP (control scrambled siRNA), and Flag-PXJ40-STAT3 were kindly provided by Professor Xinmin Cao (Signal Transduction Laboratory, Institute of Molecular and Cell Biology, Singapore). Transfection of plasmids was carried out using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions.
Cell viability assay. Cells (1000/well) were plated in 96-well plates. Each treatment group had eight replicates. The growth was monitored at different time points and cell viability was measured by MTT assay as previously described.(19)
Colony formation assay. Cells were seeded in six-well culture plates at 1000 cells/well. After 10 days of incubation, the cells were stained with Giemsa. The colonies with more than 50 cells were counted. The colony formation rate was acquired through colony numbers/total seeding cells.
Immunocytochemistry and immunohistochemistry. The cells were fixed with 4% paraformaldehyde and blocked with 10% goat or rabbit serum. They were then probed with primary anti-HA polyclonal antibody (1:1600; Cell Signaling Technology, Beverly, MA, USA), and anti-signal transducer and activator of transcription (STAT3) mAb (1:400; Cell Signaling Technology) followed by secondary antibody (FITC-conjugated goat anti-rabbit IgG or TRITC-conjugated goat anti-mouse IgG). The cell nuclei were counterstained with DAPI (Vector Laboratories, Burlingame, CA, USA).
Paraffin-embedded sections were deparaffinized with xylene and rehydrated through a graded series of alcohol. GRIM-19 antibody (1:200; Abnova, Walnut, CA, USA) or PTyr(705)-STAT3 antibody (1:100; Cell Signaling Technology) was applied to sections overnight at 4°C, followed by ABC kit (Santa Cruz Biotechnology, Santa Cruz, CA, USA). ABC kits include reagents for the avidinbiotin complex (ABC) technique, a highly sensitive method for immunohistochemical detection with biotinylated secondary antibodies. The average number of GRIM-19-positive cells was quantified in 20 fields randomly from each sample section at ×40 objective.
Semiquantitative RT-PCR and Western blot analysis. Total RNA was extracted using a Qiagen RNeasy Plus Mini Kit (Qiagen, Valencia, CA, USA). Total RNA (2 μg) was reverse-transcribed using ReverAid First Strand cDNA synthesis kit (Fermentas, Burlington, Ontario, Canada). Polymerase chain reaction was carried out with ExTaq (Takara, Kyoto, Japan). The primer sequence and the size of amplified product are shown in Table S1. The PCR products were separated on 2% agarose gels by electrophoresis. The intensity of bands was determined using Image-Pro Plus 6.0 software (http://www.mediacy.com/index.aspx?page=Home). Proteins were separated by SDS-PAGE and transferred to PVDF membrane probed with the respective antibodies.
Flow cytometric assay. At 48 h after transfection, a total of 5 × 105 cells were transferred to a tube in which 5 μL annexin V and 5 μL propidium iodide were added. The cells were then allowed to incubate at room temperature for 15 min and analyzed using flow cytometry.
Wound healing experiment. A scratch was made in monolayer of cells with a 200 μL pipette tip. Wound healing was monitored at 36 h. The relative migration distance was acquired by the following formula: the relative migration distance (%) = 100 (A − B)/A, where A represents the width of cell wounds before incubation, and B indicates the width of cell wounds after incubation.
Statistical analyses. The correlation between GRIM-19 or PTyr(705)-STAT3 expression and tumor grade was assessed by Spearman’s rank test. Multiple comparisons of means were determined by one-way anova followed by the Newman–Keuls test. The average number of GRIM-19-positive cells was compared using the unpaired Student’s t-test. The difference in the colony formation rate between each group was assessed by the chi square-test. The level of significance was set at P < 0.05.
Expression of GRIM-19 in glioma samples. Both mRNA and protein for GRIM-19 expression in glioma and normal brain tissues were examined using RT-PCR and Western blot analysis, respectively. Representative results are shown in Figure 1(a,b). Densitometric evaluation of the relative expression showed that GRIM-19 mRNA and protein levels were lower in high-grade than in low-grade gliomas or normal brain tissue and were inversely correlated with the malignancy of gliomas (*P < 0.05; **P < 0.01; Fig. 1c,d). Consistent with this observation, normal brain tissue, unlike glioma tissue, contained many diffusely distributed GRIM-19-positive cells using immunohistochemical assay (Fig. 2A). In 7 of 16 (43.8%) grade II glioma samples, 10 of 18 (55.6%) grade III, and 15 of 24 (62.5%) grade IV glioma samples, the number of GRIM-19-positive cells was markedly fewer than in normal brain tissues (Table S2). The average number of GRIM-19-positive cells was significantly lower in glioma than normal brain tissues and inversely correlated with grades of glioma (Fig. 2B).
Effects of GRIM-19 on glioma cell growth and motility. The U251MG, A172, and U87MG cells expressed GRIM-19 at relatively high, moderate, and weak levels, respectively (data not shown). At 48 h after transfection, GRIM-19 expression was suppressed in U251MG cells with GRIM-19 siRNA or GRIM-19 siRNA + STAT3siRNA and was overexpressed in U87MG cells with pXJ40-GRIM-19 (Fig. S1). Downregulation of GRIM-19 significantly promoted the cell growth of U251MG cells, whereas STAT3 knockdown attenuate the effect of GRIM-19 downregulation on the promotion in cell growth (Fig. 3a). However, upregulation of GRIM-19 evidently suppressed the cell growth of U87MG cells (Fig. 3b). In addition, colony formation rates were significantly increased in U251MG cells with GRIM-19 siRNA (45.3 ± 4.6%) when compared with that of control cells (27.9 ± 3.2%, P < 0.05; Fig. 3c), whereas STAT3 knockdown worked against this effect of GRIM-19 siRNA on promotion in cell colony formation (34.8 ± 3.9%, P < 0.05; Fig. 3c). Consistently, the colony formation rate was drastically decreased in U87MG cells with forced expression of GRIM-19 (16.8 ± 4.4%) when compared with that of control cells (30.8 ± 4.0%, P < 0.05; Fig. 3c). Furthermore, GRIM-19 overexpression significantly induced apoptosis of U87MG cells transfected with pXJ40-GRIM-19 compared with control cells (P < 0.05; Fig. 4).
In the injury-induced migration assays, a few U251MG control cells migrated into the “denuded” area (Fig. 5a). However, more U251MG cells with GRIM-19 siRNA migrated under the same conditions (Fig. 5a). Very few U87MG cells with pXJ40-GRIM-19 migrated into the “injured” area, whereas many more control cells migrated into the “injured” area (Fig. 5a). The wound healing capacity was significantly enhanced in U251MG cells with GRIM-19 siRNA (200 ± 13.9%) compared with control cells (100 ± 14.9%; Fig. 5b). However, the wound healing capacity was drastically suppressed in U87MG cells with pXJ40-GRIM-19 vector (51.9 ± 11.8%) compared with control cells (100 ± 8.6%; Fig. 5b).
Level of PTyr(705)-STAT3 inversely correlated to GRIM-19. The level of PTyr(705)-STAT3 was significantly increased in glioma tissues compared with the normal control (Fig. 6a) and positively correlated with the malignancy of gliomas (P < 0.05; Fig. 6b). However, the expression of total STAT3 remained relatively unchanged in various glioma tissues (Fig. 6a). Interestingly, the level of GRIM-19 protein expression gradually decreased, whereas PTyr(705) STAT3 protein expression gradually increased as the grade of glioma stepped up (Fig. 6a,b). Consistently, PTyr(705)-STAT3 expression was inversely correlated to GRIM-19 expression in glioma tissues using immnuohistochemistry analysis (Fig. 6c,d).
GRIM-19 regulated phosphorylation of STAT3 in glioblastoma cells. STAT3 phosphorylation in the GRIM-19 siRNA group was significantly enhanced compared with the control group, although the total protein level of STAT3 remained relatively unchanged in U251 MG cells (Fig. 7a). Forced expression of GRIM-19 suppressed STAT3 phosphorylation but had no effect on total STAT3 level in U87MG cells (Fig. 7a). We found that in U87MG cells transfected with Flag-STAT3 alone, STAT3 protein was distributed in the cytoplasm and the nucleus. However, in cells cotransfected with GRIM-19, the majority of STAT3 was located at the nuclear periphery, which colocalized with GRIM-19 (Fig. S2), suggesting that GRIM-19 prevents the nuclear translocation of STAT3 in glioma cells. Furthermore, a small amount of STAT3 was detected in the nucleus of U87MG cells, which was enhanced on treatment with IL-6. However, the expression of nuclear STAT3 was downregulated when cotransfected with GRIM-19, whereas expression of the nuclear protein poly(ADP-ribose) polymerase remained unchanged (Fig. 7b).
GRIM-19 regulated STAT3 target genes in malignant glioma cells. Knockdown of GRIM-19 in U251MG cells enhanced expression of a series of STAT3 target genes including Survivin, CyclinD1, BCl-XL, Bcl-2, Hif1-α, and MMP-9 at mRNA level. However, all of these genes decreased in U251MG cells with GRIM-19 knockdown followed by STAT3 siRNA, suggesting that additional depletion of STAT3 attenuates the effect of GRIM-19 depletion on the induction of STAT3 targeted genes (*P < 0.05, **P < 0.01; Figs 8,S3). Previous studies have shown that a combination of IFN-β and RA is a potent inducer for GRIM-19 expression in various cells.(10,20) It has also been reported that IL-6 promotes STAT3 activation in glioblastoma multiforme cells.(21) Indeed, many STAT3-dependent genes were significantly enhanced in cells treated with IL-6 compared with control cells. However, all of the genes tested were significantly reduced in glioma cells treated with IL-6 + IFN-β/RA compared with IL-6 alone (*P < 0.05, **P < 0.01; Figs 9,S4a). However, STAT3 remained unchanged in glioma cells regardless of IL-6 + IFN-β/RA or IL-6 treatment (Fig. 9a). Furthermore, STAT3 translocated into the nucleus after IL-6 stimulation in U251MG cells. However, the amount of nuclear STAT3 was reduced after treatment with IFN-β/RA (Fig. S4b, top panel, compare lanes 2 and 4), which was associated with a concomitant upregulation of endogenous GRIM-19 expression (Fig. S4b, bottom panel), suggesting a physiological role for GRIM-19 in the regulation of STAT3 nuclear localization. But the amount of nuclear STAT3 increased after treatment with IFN-β/RA + IL-6 followed by GRIM-19 siRNA (Fig. S4b, top panel, compare lanes 4 and 5), suggesting GRIM-19 depletion antagonizes the effect of IFN-β/RA.
GRIM-19 regulates STAT3-independent target genes in malignant glioma cells. GADD153 expression was significantly downregulated and COX-2 expression was evidently upregulated in U251MG cells with GRIM-19 siRNA (*P < 0.05; Figs 10a a,S5). Consistently, COX-2 expression was significantly downregulated and GADD153 expression was evidently upregulated in cells treated with IL-6 + IFN-β/RA compared with IL-6 alone (Figs 10b,S5b).
Recent discoveries have shown that the causes of CNS tumors are closely correlated with genes that control cell growth, differentiation, and death during normal development.(22,23) More and more genes, such as the epidermal growth factor receptor and the phosphatase and tensin PTEN, essential for normal growth and development, have been identified as being involved in glioblastoma.(24,25) Here, we showed that GRIM-19, a novel gene essential for development, was downregulated in glioma tissues and its expression was negatively correlated with malignancy grade. More importantly, interference with GRIM-19 expression in glioma cells promoted cell growth and migration through STAT3-dependent and STAT3-independent pathways.
In contrast to normal brain tissue, we observed that GRIM-19 expression was reduced in the majority of glioma samples. Furthermore, its expression was inversely correlated with the malignancy grade of gliomas. This is the first published research to show that GRIM-19 is downregulated in various glioma tissues. In light of this, it is suggested that GRIM-19 might be involved in the pathogenesis of glioma. GRIM-19 is mapped to human chromosome 19p.(26) It has been reported that deletion of 19p was closely associated with astrocytoma histologic phenotype.(27) It is speculated that this reduction of GRIM-19 expression may be due to deletion of 19p. Methylation of GRIM-19 promoter could be another explanation.(28) However, the exact mechanisms that cause GRIM-19 inactivation in gliomas remain to be clarified.
GRIM-19 is involved in numerous cellular functions, including apoptosis and cell cycle progression.(10,12,29) Indeed, in the present study, upregulation of GRIM-19 resulted in a significant decrease in the cell growth and colony formation rate of glioma cells. The inhibition of cell growth may be attributed partly to apoptosis. Apoptosis has been reported to play an important role during malignant transformation of a normal cell.(30,31) These findings, along with ours, suggest that dysregulation of GRIM-19 may disrupt the balance between proliferation and apoptosis and, as such, may represent an important pathogenic step in the development of glioma.
We and other research groups have shown that GRIM-19 interacts with STAT3 and inhibits its transcriptional activity.(12,32) The activation of STAT3 protein is rapid and transient in normal cells, but is persistently activated in numerous human tumors and cancer cell lines.(33,34) Indeed, in this study, the constitutive activation of STAT3 705 tyrosine phosphorylation was found in glioma tissues and positively correlated with malignancy grade of gliomas, suggesting that aberrant STAT3 plays an important role in glioma progression. Furthermore, PTyr(705)-STAT3 expression is inversely correlated to GRIM-19 expression in various grades of gliomas. More importantly, GRIM-19 negatively regulated the phosphorylation of STAT3 and prevented translocation of STAT3 from cytoplasm to nucleus in malignant glioma cells. These data suggested that the relationship between GRIM-19 and STAT3 may be involved in the development and progression of gliomas. This would be consistent with the fact that GRIM-19 knockdown evidently promoted the expression of a series of STAT3-dependent genes; elevated endogenous GRIM-19 levels through IFN-β/RA showed an opposite effect. Furthermore, depletion of STAT3 attenuates the effect of GRIM-19 depletion on the induction of STAT3-targeted genes. Thus, GRIM-19 probably controls glioma cell growth through the inactivation of STAT3. As a convergent point of several signal pathways, STAT3 regulates the expression of numerous tumor-related genes such as Survivin, BCL-XL, Bcl-2, cyclinD1, cyclinB1, Hif1-α, and vascular endothelial growth factor (VEGF).(35,36)Survivin, BCL-XL, Bcl-2, cyclinD1, and cyclinB1 play an important role in preventing cell death and promoting cell cycle progression in glioma cells.(37–40)Hif1-α and VEGF are closely associated with glioma angiogenesis.(41,42) Taken together, it is suggested that the role of GRIM-19 in glioma is at least partly mediated through the STAT3 pathway in glioma cells.
In addition to the STAT3-dependent pathway, we also explored the effect of GRIM-19 on two STAT3-independent genes, GADD153 and COX-2. Interestingly, GRIM-19 depletion evidently promoted the expression of COX-2 but suppressed the expression of GADD153. GADD153 has been shown to be associated with apoptosis induced by retinoic acid, anticancer agents, and nutrient deprivation.(43) COX-2 is an inducible enzyme responsible for prostaglandin production and carries out important functions in tumorigenesis.(44) Previous reports revealed that COX-2 protein is overexpressed in the majority of gliomas and involved in inhibition of apoptosis.(45) Taken together, GRIM-19-induced glioma cell apoptosis may be partly mediated through STAT3-independent mechanism such as GADD153 and COX-2.
GRIM-19 not only negatively regulated growth of glioma cells, but also suppressed glioma cell migration induced by injury signal. This is shown when confluent monolayer was scratched with a pipette tip to make a wound, and images were captured to monitor the cell movement into the wounded area. In addition to intense proliferation, glioma cells have a remarkable ability to degrade the surrounding ECM and destroy tissue boundaries for tumor invasion.(30) The MMP family of zinc-dependent proteinases have the ability to degrade ECM components.(46) Once released by tumor cells or surrounding stroma, MMPs promote tumor invasion and metastasis through the degradation of the ECM and basement membranes.(47) Accumulated evidence has suggested that there are strong correlations between high levels of MMPs and glioma cell invasiveness.(48) Among all the MMP members, MMP-2 and MMP-9 stand out because of their constitutive activation in many tumors, including glioma.(48,49) In our study, GRIM-19 knockdown increased MMP-9 expression. In contrast, there was no modulation of MMP-2 expression, regardless of upregulation or downregulation of GRIM-19. These findings suggest that GRIM-19 suppresses glioma cell migration probably through inhibition of MMP-9, but not MMP-2.
In conclusion, we have provided evidence for the first time that GRIM-19 downregulation occurs in glioma samples, and that GRIM-19 suppresses glioma cell growth and migration. Additionally, we showed that the function of GRIM-19 in glioma cells is exerted through both STAT3-dependent and STAT3-independent pathways. Considering the role of GRIM-19 in glioma pathogenesis, restoration of GRIM-19 may be proven to be a novel and potential strategy to treat gliomas.
This research was supported by funding from: the National Basic Research Program of China (973 Program, Grant Nos. 2007CB512001, 2011CB966201); the National Natural Science Foundation of China (Grant Nos. 30771142, 81071057); and the Natural Science Foundation of Shandong Province (Grant Nos. Z2007C11, JQ200823).