JOURNAL OF PHYSIOLOGY AND PHARMACOLOGY 2017, 68, 2, 209-214 www.jpp.krakow.pl

E. MAJEWSKA1, R. ROLA2, M. BARCZEWSKA3, J. MARQUEZ4, J. ALBRECHT1, M. SZELIGA1

TRANSCRIPTION FACTOR GATA3 EXPRESSION IS INDUCED BY GLS2 OVEREXPRESSION IN A GLIOBLASTOMA CELL LINE BUT IS GLS2-INDEPENDENT IN PATIENT-DERIVED GLIOBLASTOMA

1Department of Neurotoxicology, Mossakowski Medical Research Centre, Polish Academy of Sciences, Warsaw, Poland; 2Department of Neurosurgery and Paediatric Neurosurgery, Medical University of Lublin, Lublin, Poland; 3Department of Neurosurgery, Faculty of Medical Sciences, University of Warmia and Mazury, Olsztyn, Poland; 4Department of Molecular Biology and Biochemistry, Faculty of Sciences, Campus de Teatinos, University of Malaga, Malaga, Spain

Phosphate-activated (GA), a ubiquitous glutamine-metabolizing enzyme, is encoded by two , GLS and GLS2. In mammalian cancers, GLS isoforms are perceived as molecules promoting cell proliferation and invasion, whereas the role of GLS2 isoforms seems to be more complex and cell type-specific. Previous studies have shown abundance of GLS and lack of GLS2 transcripts in T98G human glioblastoma (GBM) cell line and patient-derived GBM. Transfection with GAB sequence, the whole GLS2 cDNA transcript, suppressed malignant phenotype of T98G cells. Microarray analysis revealed upregulation of GATA3, the product of which has been implicated in suppressing growth of some peripheral cancers. In this study we confirmed a significant upregulation of GATA3 expression in the transfected cells both at mRNA and level. Considerable expression of GATA3 was also observed in GBM tissues (previously shown as not expressing GLS2), while only traces or no GATA3 was detected in (GLS2-expressing) non-tumorigenic brain samples. In conclusion, while mechanistic relation between GAB and GATA3 expression is evident following in vitro manipulation of GBM cell line, it does not appear to be an intrinsic property of GBM nor non-tumorigenic brain tissue.

Key words: glutaminase isoforms, GLS2 , GATA3, transcription factor, human glioblastoma cell line, glioblastoma tissue, non-tumorigenic brain tissue

INTRODUCTION GBM cells (13, 14). Stable transfection of human glioblastoma cell line T98G with a sequence encoding GAB leads to Glioblastoma multiforme (GBM, WHO grade IV) belongs to suppression of malignant phenotype of these cells by decreasing the most malignant primary brain tumors with very poor prognosis their proliferation, viability and ability to migrate (15). (1). Glutamine (Gln) is a versatile amino acid promoting Moreover, the transfected cells, herein termed TGAB, are more hallmarks of malignancy (2). Phosphate-activated glutaminase sensitive to the alkylating agents used in glioma therapy (16). (GA, EC 3.5.1.2) metabolizes Gln to glutamate (Glu) and The exact mechanism of action of GAB remains unknown. Its ammonia and thus plays a significant role in tumor biology. nuclear localization in defined populations of neurons (17) and Human GA is encoded by two genes: GLS encodes kidney-type astrocytes (18) led to hypothesis that this molecule can modulate isoforms and GLS2 encodes liver type isoforms (3). So far, two gene expression. Consistent with this hypothesis, microarray transcripts arising from each of GA genes have been identified: analysis revealed that transfection of T98G cells with GAB KGA and GAC for GLS and GAB and LGA for GLS2 (3). sequence deregulated expression of 85 genes (15). Different studies revealed that glutaminase isoforms play GATA3 is one of the genes whose expression is upregulated opposite roles in tumorigenesis. GLS isoforms are associated in TGAB cells as compared to the controls (15). This gene codes with high rate of cell proliferation, whereas GLS2 are for a multifunctional transcription factor which plays an related to resting or quiescent cell states (4). Suppression of the important role in the development of several tissues including malignant phenotype can be obtained by inhibition of GLS (5-7) brain (19). or overexpression of GLS2 (8, 9). Of note, the GLS gene has Available data on the role of GATA3 in the progression and been shown to be regulated by the mediators of oncogenesis, development of tumors of different tissue origin are ambiguous. MYC (10), Rho GTPases (6) and Notch (7), whereas the GLS2 The loss of GATA3 expression in breast cancer is linked to gene was identified as a target of p53 tumor suppressor (8). aggressive tumor development, metastatic capability and poor In GBM tissues GLS isoforms are highly expressed, while prognosis (20). The low level of GATA3 is related to the expression of GLS2 is silenced (11) via DNA hypermethylation promotion of cell migration and invasion of bladder cancer cells (12). Recent findings indicate that isoforms arising from GLS (21). Knockdown of GATA3 downregulated suppressors play opposing roles to isoforms encoded by GLS2 in growth of (UGT1A, PTEN, p53, p21) and upregulated oncogenes (i.e. c- 210 myc, cyclin D1, cyclin D3, cyclin E, FGFR3) in SVHUC tumorigenic brain tissues: two GBM adjacent tissues retrieved immortalized normal urothelial cells (22). In renal cell from the margin of the tumor and two other obtained by carcinoma GATA3 is silenced via hypermethylation (23). Low intracranial decompression for severe traumatic brain injury. The frequency of GATA3 expression has been detected in control tissues were identified by a neuropathologist as non- adenocarcinomas of lung, colon, endometrium, ovary and neoplastic. prostate (24). All the above would indicate a tumor suppressing role of GATA3. On the other hand, common GATA3 presence in Protein isolation and western blot analysis skin adnexal tumors, mesothelioma, salivary gland and pancreatic ductal carcinomas may not favor its tumor Cells or tissues were lysed in ice-cold RIPA buffer with suppressing role (24). Furthermore, overexpression of GATA3 cocktails of sodium fluoride, protease and phosphatase inhibitors correlates with a poor outcome of neuroblastoma patients (25). (Sigma-Aldrich). Lysates were centrifuged at 12,000 g for 10 Strong expression of GATA3 has also been documented in min at 4°C, and the supernatants were collected. Proteins Hodgkin lymphoma (HL) and down-regulation of GATA3 in HL concentrations were determined by bicinchoninic acid Protein cell lines demonstrated its role in the regulation of IL-5, IL-13, Assay Kit (Pierce, Rockford, IL, USA). Fifty µg of cells or tissue STAT4 (26). proteins were separated by 10% SDS-PAGE and then Previously performed microarray analysis revealed electrotransferred to nitrocellulose membranes. Membranes increased expression of GATA3 in TGAB cells as compared to were blocked with 5% skim milk for 1 hour at room temperature, the controls (15). This prompted the question whether the high and incubated with a human-specific anti-GATA3 antibody GATA3 content in GAB-enriched cells is a coincidence or (Abcam, Cambridge, MA) overnight at 4°C. After washing, reflects the regulation of expression of GATA3 by GAB. We first membranes were incubated with anti-rabbit antibody (Sigma- attempted to verify this hypothesis by comparing the mRNA and Aldrich) conjugated to horseradish peroxidase for 1 h at room protein level of GATA3 in T98G and TGAB cell lines. temperature. Blots were visualized on X-ray film using a Subsequently, we asked the question whether and in what degree SuperSignal West Pico Chemiluminiscence Substrate (Pierce). the finding obtained in a cell line is reflected in clinical For loading control, membranes were stripped for 15 min two glioblastomas which are composed of a heterogeneous times in Stripping Buffer and reused with an antibody against population of cells. To this end we measured the levels of glyceraldehyde 3-phosphate dehydrogenase (ProteinTech, GATA3 mRNA and protein in GLS2-negative GBMs tissues and Chicago, IL, USA). Densitometric analysis was performed with in non-tumorigenic brain tissues previously shown to be GLS2- G:Box system and GeneTools software (Syngene, Frederick, positive (12). MD, USA).

Real-time PCR MATERIALS AND METHODS Total RNA from cell lines and tissues were extracted using a Cell lines and culture conditions commercially available TRI-Reagent (Sigma-Aldrich). Two µg of RNA were reverse-transcribed using High Capacity cDNA T98G human glioblastoma cell line (American Type Culture Reverse Transcription Kit (Applied Biosystems, Warrington, Collection), TGAB and TpcDNA cell lines were maintained in UK) according to the manufacturer’s protocol. The GATA3 Minimum Essential Medium Eagle (Sigma-Aldrich, St. Louis, primers were purchased from Applied Biosystems (assay ID: MO, USA) supplemented with 10% fetal bovine serum (Gibco, Hs00231122_m1), β-actin primers were purchased from Applied Thermo Fisher Scientific, Grand Island, NY, USA), 1% non- Biosystems (in case of cell line experiments, assay ID: essential amino acids (Gibco), and 1% antibiotics (penicillin and 4326315E) and Blirt (Gdansk, PL) (in case of biopsy material streptomycin) (Gibco). Cultures were maintained at 37°C in a experiments, cat no HK-DD-hu). Each reaction (total volume of humidified atmosphere with 5% CO2. TGAB and TpcDNA cell reaction 10 µl) contained 5 µl TaqMan Universal PCR Master lines were obtained by stable transfection of T98G cells with a Mix (Applied Biosystems), 1 µl cDNA, 0.5 µl primers and full cDNA sequence coding for human GAB or empty pcDNA3 RNAse/DNAse free water. The real time PCR reactions were vector respectively (15). The culture medium for TpcDNA and performed at 95°C for 10 min, followed by 40 cycles of 30 s at TGAB cells was additionally supplemented with 0.5 mg/mL 95°C and 1 min at 60°C. The results of the analysis were G418 (BioShop, Lab Empire, Rzeszow, PL). The expression of calculated by normalizing to the β-actin product. Results were the GLS2 gene in all cell lines was monitored by RT-PCR as described previously (15). Experiments were performed on cells harvested 24 hours after seeding.

Biopsy material Table 1. Data of patients from whom GBM samples were obtained. The clinical material wasobteined from the Department of Neurosurgery and Pediatric Neurosurgery of the Medical Glioblastoma tissue Sex Age University Lublin, Poland and from the Department of G1 Female 58 Neurology and Neurosurgery, Faculty of Medical Sciences, G2 Male 64 University of Warmia and Mazury, Olsztyn, Poland. Written G3 Female 65 informed consent was obtained from the patients and all G4 Female 29 procedures were carried out according to the local Ethics G5 Male 67 Committee approval. Nine cases of primary glioblastoma (WHO grade IV) were obtained at the time of surgery and immediately G6 Female 73 frozen. Patients’ age and gender are presented in Table 1. G7 Female 77 Histological examination of resected tissues was performed by a G8 Female 76 neuropathologist according to the World Health Organization G9 Female 68 criteria (27). The control group consisted of four non- 211 expressed by an equation (2–∆∆Ct) that gives the amount of target, higher level of GATA3 mRNA in TGAB cells compared to the normalized to an endogenous reference and relative to a controls (Fig. 1). Then we examined the level of GATA3 protein calibrator. in all three cell lines. Densitometric analysis of band intensity revealed a 400% increase in the expression of GATA3 protein in Statistical analysis TGAB cells as compared to the controls (Fig. 2).

Statistical analysis was performed using GraphPad Prism 5 Expression of GATA3 in glioblastoma clinical samples (GraphPad Software, La Jolla, CA, USA). Statistical significance was determined by one-way analysis of variance The observation that GATA3 expression is evoked by (one-way ANOVA) followed by Tukey’s test or by t-Test overexpression of GLS2 in glioblastoma cells prompted us to followed by Mann Whitney’s test. P < 0.05 was considered examine GATA3 level in the clinical material. We used 5 GBM statistically significant. samples in which we have not detected GLS2 transcript and 2 non-tumorigenic brain tissues previously identified as expressing considerable amounts of GLS2 (12). More recently RESULTS we extended our analysis to 4 additional GBM tissues in which we observed lack or extremely low amounts of GLS2 transcript Expression of GATA3 in glioblastoma cell lines and two non-tumorigenic brain tissues identified as GLS2- positive (data not shown). Real-time PCR analysis revealed a Previously performed microarray analysis revealed GATA3 relatively similar level of GATA3 transcript in all 4 control as one of the 50 genes upregulated in TGAB cells as compared samples examined, with 2–∆∆Ct values ranging from 0.6 to 1.5 to both control T98G and TpcDNA cells (15). Here we validated (Fig. 3A). By contrast, a large variability in the expression level this finding by real-time PCR analysis which showed 12 times of this transcript was observed between GBM samples: 2–∆∆Ct values were ranging from 0.0 to 169.0 (Fig. 3A). The difference in GATA3 mRNA level between GBM and control tissues was statistically significant (P < 0.05) (Fig. 3B). We further examined the level of GATA3 protein in the same clinical samples. We found GATA3 band in all GBM samples examined, although its intensity was heterogeneous (Fig. 4A). A faint signal was observed in one sample, a moderate in six samples and very sharp in two. A lack or trace of GATA3 band was observed in two control tissues, and a weak signal was found in another two (Fig. 4A). Altogether, statistical analysis showed significantly higher levels of GATA3 protein in GBM as compared to the controls (Fig. 4B).

DISCUSSION

A growing body of evidence indicates that particular GA isoforms play opposing roles in tumorigenesis. GLS proteins are generally perceived as molecules driving cancer cell metabolism by degrading Gln. The idea that the role of GLS2 proteins in Fig. 1. Expression of mRNAs coding for GATA3 in T98G, tumorigenesis may be more complex was derived from the TpcDNA and TGAB cell lines assessed by real-time PCR. Results surprising finding that they are located not only in mitochondria, are mean ± S.D. (n = 3 – 4). *P < 0.05 versus T98G and TpcDNA tested with the one-way ANOVA followed by Tukey’s test.

Fig. 2. Expression of GATA3 protein in T98G, TpcDNA and TGAB cell lines. (A) Representative immunoblot. (B) Quantification of GATA3 bands intensity normalized to GAPDH. Results are mean ± S.D. (n = 5). *P < 0.05 versus T98G and TpcDNA tested with the one-way ANOVA followed by Tukey’s test. 212

Fig. 3. Expression of mRNAs coding for GATA3 in non-tumor and tumor tissue assessed by real-time PCR. 2–∆∆Ct values of 4 controls and 9 GBM tissues in a table (A) and presented as a scatter graph (B). Results are mean ± S.D. (n = 4 – 9). *P < 0.05 tested with the t-Test followed by Mann Whitney’s test.

Fig. 4. Expression of GATA3 protein in non-tumor and tumor tissue. (A) Representative immunoblot. (B) Quantification of GATA3 bands intensity normalized to GAPDH. Results are mean ± S.D., (n = 4 – 9). *P < 0.05 tested with the t-Test followed by Mann Whitney’s test.

but also in the cell nuclei and as such may contribute to doubt on their interaction in determining the tumor phenotype. transcription modulation (17, 18). Out of the nine GBM samples, all of which were GLS2- Our previous study revealed abundance of GLS and lack of deficient (12), six showed clear GATA3 upregulation as GLS2 expression in GBM tissues (11). Transfection of T98G compared to the non-tumorigenic samples, three of them glioblastoma cells with a sequence encoding GAB, the longer presented relatively lower GATA3 level. Non-tumorigenic transcript arising from GLS2, decreased cell proliferation and samples, all of which presented abundant GLS2 expression, had their ability to migrate (15). The mechanisms underlying the none to low GATA3 levels. Altogether, GATA3 expression in mode of action of GAB in glioblastoma cells remains to be GBM samples exceeded that in non-tumorigenic samples over 6 elucidated. T98G cells transfected with GAB sequence, termed times, a value comparable to the GLS2-driven enrichment of TGAB, presented deregulated expression of many genes, of GATA3 in a GBM cell line. These results suggest that which some could be potentially related to oncogenesis (15). mechanistic relation, if any, between GAB and GATA3 One of the downregulated genes codes for a DNA-repair protein, expression appears to be a result of in vitro manipulation carried O6-methylguanine DNA methyltransferase (MGMT) that is out in this particular GBM cell line, but is not an intrinsic associated with resistance to therapies using alkylating agents. reproducible property of GBM tissue. The above suggestion is Our later study showed that TGAB cells are more sensitive to consistent with an increasing body of evidence indicating that these compounds as compared to the controls (16). biological features of glioblastoma cell lines do not genuinely According to the microarray analysis, GATA3 was another reflect those of tumors grown in vivo (28) and that the metabolic gene deregulated in TGAB cells (15). As outlined in the incompatibility of cell lines with cells from which they directly Introduction, GATA3 is assumed to play important, albeit originate goes down to the DNA profile (29). variable roles in progression of tumors in different tissues (20- Going back to the GLS2/GATA3 interrelations, the analysis 26). Herein, we confirmed a marked increase in both GATA3 of data generated by the TCGA (The Cancer Genome Atlas) mRNA and protein in TGAB cells as compared to the controls. Research Network (http://cancergenome.nih.gov) did not reveal However, the results obtained with human GBM samples and a survival benefit related to high GLS2 or GATA3 expression in non-tumorigenic brain tissue do not support positive correlation native tumors. Moreover, examination of the PathCards database between the expression of the two proteins and as such cast (https://pathcards.genecards.org) appears to indicate that GLS2 213 and GATA3 pathways are not very likely to be interdependent 6. Wang JB, Erickson JW, Fuji R, et al. Targeting (30). Clearly, the above considerations do not exclude some mitochondrial glutaminase activity inhibits oncogenic indirect mutual control of their expression: lack of success in transformation. Cancer Cell 2010; 18: 207-219. finding the interdependence may in part reflect differences in the 7. Kalhert UD, Cheng M, Koch K, et al. Alterations in cellular cellular composition of the examined material: cellular metabolome after pharmacological inhibition of Notch in homogeneity of the cell line versus molecular and cellular glioblastoma cells. Int J Cancer 2016; 138: 1246-1255. heterogeneity of GBM tissues, which is their hallmark (1). In the 8. Hu W, Zhang C, Wu R, Sun Y, Levine A, Feng Z. future experiments the GATA3/GLS2 relationship will have to Glutaminase 2, a novel p53 target gene regulating energy be analyzed by measuring the effects of in situ manipulations metabolism and antioxidant function. Proc Natl Acad Sci with GATA3 expression in cell lines and GBM-derived native USA 2010; 107: 7455-7460. cell cultures. Such experiments are under way in this laboratory. 9. Liu J, Zhang C, Lin M, et al. Glutaminase 2 negatively On the other hand, the variability in GATA3 protein level among regulates the PI3K/AKT signaling and shows tumor GBMs observed in this study can result from the activation of suppression activity in human hepatocellular carcinoma. different responses in GBM tissues with different genomic Oncotarget 2014; 5: 2635-2647. profiles, per se reflecting their heterogeneity. Of note, a great 10. Gao P, Tchernyshyov I, Chang TC, et al. c-Myc suppression divergence in GATA3 protein level was documented in GBM of miR-23a/b enhances mitochondrial glutaminase tissues carrying a wild-type PTEN, while a decreased amount of expression and glutamine metabolism. Nature 2009; 458: GATA3 protein was found in GBM samples having PTEN 762-765. nonsense mutations (31). 11. Szeliga M, Sidoryk M, Matyja E, Kowalczyk P, Albrecht J. In summary, we documented a transcriptional regulation of Lack of expression of the liver-type glutaminase (LGA) GATA3 in glioblastoma cells upon transfection with GAB mRNA in human malignant gliomas. Neurosci Lett 2005; sequence. On the other hand, we have also shown considerable, 374: 171-173. though variable GATA3 level in GBM clinical samples not 12. Szeliga M, Bogacinska-Karas M, Kuzmicz K, Rola R, expressing GLS2. The tumor-promoting effect of the absence of Albrecht J. Downregulation of GLS2 in glioblastoma cells is GLS2 noted in the GBM cell line is likely to hold for GBM related to DNA hypermethylation but not to the p53 status. tissue. However, the ubiquity of this relation remains to be Mol Carcinog 2016; 55: 1309-1316. proven, as reports regarding the role of GLS2 in peripheral 13. Szeliga M, Bogacinska-Karas M, Rozycka A, Hilgier W, tumors are conflicting; in hepatocellular carcinoma this isoform Marquez J, Albrecht J. Silencing of GLS and has been claimed to either promote (32) or inhibit (33) growth overexpression of GLS2 genes cooperate in decreasing the and invasion of cancer cells. Moreover, downstream proliferation and viability of glioblastoma cells. Tumour mechanisms disclosed in the cell line may not reflect that Biol 2014; 35: 1855-1862. operating in GBM tissue. It remains to be elucidated whether 14. Martin-Rufian M, Nascimento-Gomes R, Higuero A, et al. and how overexpression of GATA3 affects the progression of Both GLS silencing and GLS2 overexpression synergize with GBM in the clinical setting. oxidative stress against proliferation of glioma cells. J Mol Med (Berl) 2014; 92: 277-290. Acknowledgments: The authors are greatly indebted to the 15. Szeliga M, Obara-Michlewska M, Matyja E, et al. teams of the regional pathology units (the Laboratory of Transfection with liver-type glutaminase cDNA alters gene Neuropathology, Department of Neurosurgery and Paediatric expression and reduces survival, migration and proliferation Neurosurgery, Medical University of Lublin, Poland and the of T98G glioma cells. Glia 2009; 57: 1014-1023. Department of Pathomorphology, Faculty of Medical Sciences, 16. Szeliga M, Zgrzywa A, Obara-Michlewska M, Albrecht J. University of Warmia and Mazury, Olsztyn, Poland) for Transfection of a human glioblastoma cell line with liver- providing histopathological services. type glutaminase (LGA) down-regulates the expression of DNA-repair gene MGMT and sensitizes the cells to The study was supported by the Leading National Research alkylating agents. J Neurochem 2012; 123: 428-436. Centre (KNOW-MMRC) project (EM). 17. Olalla L, Gutierrez A, Campos JA, et al. Nuclear localization of L-type glutaminase in mammalian brain. J Biol Chem Conflict of interests: None declared. 2002; 277: 38939-38944. 18. Cardona C, Sanchez-Mejias E, Davila JC, et al. Expression of Gls and Gls2 glutaminase isoforms in astrocytes. Glia REFERENCES 2015; 63: 365-382. 19. Tsarovina K, Reiff T, Stubbusch J, et al. The Gata3 1. Aldape K, Zadeh G, Mansouri S, Reifenberger G, von transcription factor is required for the survival of embryonic Deimling A. Glioblastoma: pathology, molecular mechanisms and adult sympathetic neurons. J Neurosci 2010; 30: 10833- and markers. Acta Neuropathol 2015; 129: 829-848. 10843. 2. Hensley CT, Wasti AT, DeBerardinis RJ. Glutamine and 20. Yoon NK, Maresh EL, Shen D, et al. Higher levels of cancer: cell biology, physiology, and clinical opportunities. GATA3 predict better survival in women with breast cancer. J Clin Invest 2013; 123: 3678-3684. Hum Pathol 2010; 41: 1794-1801. 3. Campos-Sandoval JA, Martin-Rufian M, Cardona C, Lobo 21. Li Y, Ishiguro H, Kawahara T, Kashiwagi E, Izumi K, C, Penalver A, Marquez J. Glutaminases in brain: multiple Miyamoto H. Loss of GATA3 in bladder cancer promotes isoforms for many purposes. Neurochem Int 2015; 88: 1-5. cell migration and invasion. Cancer Biol Ther 2014; 15: 4. Perez-Gomez C, Campos-Sandoval JA, Alonso FJ, et al. Co- 428-435. expression of glutaminase K and L isoenzymes in human 22. Li Y, Ishiguro H, Kawahara T, Miyamoto Y, Izumi K, tumour cells. Biochem J 2005; 386: 535-542. Miyamoto H. GATA3 in the urinary bladder: suppression of 5. Lobo C, Ruiz-Bellido MA, Aledo JC, Marquez J, Nunez De neoplastic transformation and down-regulation by Castro I, Alonso FJ. Inhibition of glutaminase expression by androgens. Am J Cancer Res 2014; 4: 461-473. antisense mRNA decreases growth and tumourigenicity of 23. Cooper SJ, Zou H, Legrand SN, et al. Loss of type III tumour cells. Biochem J 2000; 348: 257-261. transforming growth factor-beta receptor expression is due 214

to methylation silencing of the transcription factor GATA3 in news. Sci Transl Med 2016; 8: 354re3. doi: renal cell carcinoma. Oncogene 2010; 29: 2905-2915. 10.1126/scitranslmed.aaf6853 24. Miettinen M, McCue PA, Sarlomo-Rikala M, et al. 30. Belinky F, Nativ N, Stelzer G, et al. PathCards: multi-source GATA3: a multispecific but potentially useful marker in consolidation of human biological pathways. Database surgical pathology: a systematic analysis of 2500 (Oxford) 2015; pii: bav006. doi: 10.1093/database/bav006 epithelial and nonepithelial tumors. Am J Surg Pathol 31. Xu J, Li Z, Wang J, Chen H, Fang JY. Combined PTEN 2014; 38: 13-22. mutation and protein expression associate with overall and 25. Peng H, Ke XX, Hu R, Yang L, Cui H, Wei Y. Essential role disease-free survival of glioblastoma patients. Transl Oncol of GATA3 in regulation of differentiation and cell 2014; 7: 196-205. proliferation in SK-N-SH neuroblastoma cells. Mol Med Rep 32. Lee YZ, Yang CW, Chang HY, et al. Discovery of selective 2015; 11: 881-886. inhibitors of glutaminase-2, which inhibit mTORC1, 26. Stanelle J, Doring C, Hansmann ML, Kuppers R. activate autophagy and inhibit proliferation in cancer cells. Mechanisms of aberrant GATA3 expression in classical Oncotarget 2014; 5: 6087-6101. Hodgkin lymphoma and its consequences for the cytokine 33. Kuo TC, Chen CK, Hua KT, et al. Glutaminase 2 stabilizes profile of Hodgkin and Reed/Sternberg cells. Blood 2010; Dicer to repress Snail and metastasis in hepatocellular 116: 4202-4211. carcinoma cells. Cancer Lett 2016; 383: 282-294. 27. Louis DN, Ohgaki H, Wiestler OD, et al. The 2007 WHO classification of tumours of the central nervous system. Acta Received: June 28, 2016 Neuropathol 2007; 114: 97-109. Accepted: April 15, 2017 28. Grabacka M, Waligorski P, Zapata A, et al. Fenofibrate subcellular distribution as a rationale for the intracranial Author’s address: Dr. Monika Szeliga, Department of delivery through biodegradable carrier. J Physiol Pharmacol Neurotoxicology, M. Mossakowski Medical Research Centre, 2015; 66: 233-247. Polish Academy of Sciences, 5 Pawinskiego Street, 02-106 29. Allen M, Bjerke M, Edlund H, Nelander S, Westermark B. Warsaw, Poland. Origin of the U87MG glioma cell line: good news and bad E-mail: [email protected]