B R A I N R E S E A R C H 1 3 0 4 ( 2 0 0 9 ) 1 4 9 – 1 5 4

ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r . com/ loca te /b ra i n res

Research Report

The expression and significance of HIF-1α and GLUT-3 in glioma

Yang Liua,d,1, Yun-ming Lib,1, Rui-feng Tianc,1, Wei-ping Liua,⁎, Zhou Feia, Qian-fa Longa,Xiao-an Wanga, Xiang Zhanga

aDepartment of Neurosurgery, Xijing Hospital, Fourth Military Medical University,17 Changle West Road, Xi'an,710032, ChinabDepartment of Health Statistics, Fourth Military Medical University, Xi'an, 710032, ChinacDistrict Retired Cadres Entertainment Center, Fourth Military Medical University, Xi'an,710032, ChinadDepartment of Neurosurgery, The Third Hospital of Mianyang, Sichuan, 621000, China

A R T I C L E I N F O

⁎ Corresponding author. Fax: +86 29 84775567E-mail address: liuwpfmmu@163.com (WAbbreviations: HIF-1α, hypoxia inducible f

1 These authors contributed equally to this

0006-8993/$ – see front matter © 2009 Elsevidoi:10.1016/j.brainres.2009.09.083

A B S T R A C T

Article history:Accepted 19 September 2009Available online 25 September 2009

HIF-1α plays an indispensable role in tumor formation and histogenesis. Target genesinvolved in glucose transport are acutely transactivated by HIF-1α. GLUT-3 protein is therate-limiting factor related to glucose transport, which is classified as brain-type glucosetransporter. This study was the initial one aiming to probe into the co-expression andclinical significance of HIF-1α and GLUT-3 in glioma. One hundred and twenty cases ofglioma tissues and ten human normal cerebral tissues decompressed in glioma excisionwere examined using immunohistochemistry and Western blot. The expression of HIF-1αand GLUT-3 increased gradually with the increase of pathological grade of glioma,respectively. There was significant difference in the expression of HIF-1α and GLUT-3 inevery two groups, respectively. There was a positive correlation between HIF-1α and GLUT-3. In conclusion, the expression of HIF-1α and GLUT-3 in glioma was correlated significantlywith tumors' pathological grade, which can be taken as a pair of useful markers forpredicting the biological behavior of glioma.

© 2009 Elsevier B.V. All rights reserved.

Keywords:HIF-1αGLUT-3GliomaImmunohistochemistry

1. Introduction

It is well known that malignant tumors are regarded to be in amore or less hypoxic state, with highly malignant tumorsbeing specially hypoxia-resistant (Zhong et al., 1999). Hypoxia-inducible factor-1 (HIF-1), a heterodimer composed of α and βsubunit, plays an indispensable role in tumor formation andhistogenesis (Semenza, 2002). The expression and transcrip-tional activity of HIF-1 itself increase when oxygen availabilitybecomes a limiting factor, and several dozen target genesincluding those involved in glucose transport, glycolysis, andtissue vascularization are acutely transactivated by HIF-1(Mobasheri et al., 2005). In this process, HIF-1 transcription

.. Liu).actor-1α; GLUT-3, glucosework.

er B.V. All rights reserved

activity is regulated by the expression level of the HIF-1αsubunit (Semenza, 2002), and the loss of the von Hippel-Lindau tumor suppressor gene results in constitutive over-expression of HIF-1α (Huang et al., 1998). In all mammaliancells, one or more of the glucose transporter (GLUT) geneswere found. GLUTs, which are integral membrane glycopro-teins, play an essential role in facilitating glucose transport. Inprevious study, Otto Warburg observed that the rate ofanaerobic-like glycolysis was increased obviously in malig-nant tumor cells compared to non-malignant tumor cells ornormal tissue cells (Warburg, 1956). GLUT-3 is a representativemember of the family and is primarily expressed in neurons.The prominent localization of GLUT-3 to mature neuronal

transporter-3

.

150 B R A I N R E S E A R C H 1 3 0 4 ( 2 0 0 9 ) 1 4 9 – 1 5 4

processes suggests an essential role for this transporter inregulating fuel requirements for dendritic and axonal traffic,therebymediating neurotransmission (Mantych et al., 1992). Avariety of researches have shown a close relationship betweenGLUT-3 and carcinogenesis, tumor development or theunfavorable prognosis of various malignancies (Airley andMobasheri, 2007; Riedl et al., 2007). However, to our knowledge,none of research about co-expression of HIF-1α and GLUT-3 inhuman glioma has been found. In this study, we focusedprimarily on the co-expression of HIF-1α and GLUT-3 on a totalof 120 cases of human gliomas and 10 cases of brain tissues by

Fig. 1 – Immunohistochemical expression of HIF-1α and GLUT-3proteins immunoreactivity was shown as brown staining in thebrown staining in the cell membrane and cytoplasm (ABC, originglioma, (1B) HIF-1α in grade II glioma, (1C) HIF-1α in grade III glglioma, (1F) GLUT-3 in grade II glioma, (1G) GLUT-3 in grade III g

the immunohistochemical method and Western blot withspecial antibody.

2. Results

2.1. HIF-1α expression was recognized through a nuclearstaining of glioma cells mainly

There was no expression of HIF-1α in normal brain tissues.There was statistical significance in the expression of HIF-1α

in human gliomas of different pathological grades. HIF-1αnucleus, GLUT-3 proteins immunoreactivity was shown asal magnifications 400× for 1A–1H). (1A) HIF-1α in grade Iioma, (1D) HIF-1α in grade IV glioma, (1E) GLUT-3 in grade Ilioma, and (1H) GLUT-3 in grade IV glioma.

Table 1 – HIF-1α-IRS and GLUT-3-IRS in glioma of different sex, age, and grade groups.

n HIF-1α-IRS GLUT-3-IRS

– + ++ +++ P – + ++ +++ P

Total 120 58 18 19 25 54 17 21 28Gender 0.910 0.116

Male 61 30 9 9 13 30 12 8 11Female 59 28 9 10 12 24 5 13 17

Age, years 0.293 0.316<30 39 22 4 4 9 22 4 5 8≥30<40 43 17 6 9 11 16 5 12 10≥40 38 19 8 6 5 16 8 4 10

Pathologic grade <0.001a <0.001b

Grade I 30 26 2 2 0 25 3 1 1Grade II 30 18 6 3 3 16 7 4 3Grade III 30 10 5 8 7 9 4 10 7Grade IV 30 4 5 6 15 4 3 6 17

a Difference of HIF-1α-IRS in different groups was significant, statistical significance was defined as P-value <0.05.b Difference of GLUT-3-IRS in different groups was significant, statistical significance was defined as P-value <0.05.

151B R A I N R E S E A R C H 1 3 0 4 ( 2 0 0 9 ) 1 4 9 – 1 5 4

between glioma and normal brain tissues (Z=2.948, P=0.003).There was no statistical significance between the expressionof HIF-1α and patient' sex and age (P>0.05). The expression ofHIF-1α increased gradually with the increase of pathologicalgrade of glioma (χ2=41.401, P<0.001), and the difference in theexpression of HIF-1α in every two groups was significant.Representative images of HIF-1α immunostaining and therelated results were shown in Fig. 1A–D and Tables 1, 2,respectively. The expression of GLUT-3was detected in gliomacell membrane and cytoplasm. There was no expression ofGLUT-3 in normal brain tissues. There was statistical signif-icance in the expression of GLUT-3 between glioma andnormal brain tissues (Z=3.101, P=0.002). There was nostatistical significance between the expression of GLUT-3and patient' sex and age. (P>0.05). The expression of GLUT-3increased gradually with the increase of pathological grade ofglioma (χ2=41.124, P<0.001), and the difference in the expres-sion of GLUT-3 between every two groups was significant.Representative images of GLUT-3 immunostaining and therelated results were shown in Fig. 1E–H and Tables 1, 2,respectively. There was a positive correlation between HIF-1α-IRS and GLUT-3-IRS (r=0.545, P<0.001).

2.2. In Western blot, the expression levels of HIF-1α orGLUT-3 were represented by the gray value of HIF-1α orGLUT-3/gray value to their β-actin

The cases of Western blot were the same to those ofimmunohistochemistry, and the coincident conclusion wasacquired. The results showed that protein expression levels ofHIF-1α or GLUT-3 increased with the increase of pathologicalgrades and the difference of HIF-1α or GLUT-3 in every twogroupswas significant. (P<0.05, Figs. 2, 3). Therewas a positive

Table 2 – Comparison of the expression of HIF-1α and GLUT-3

Variables Grade I vs. II Grade I vs. III Grade I vs. IV

HIF-1α 0.019 <0.001 <0.001GLUT-3 0.013 <0.001 <0.001

correlation between HIF-1α/β-actin and GLUT-3/β-actin(r=0.550, P<0.001).

3. Discussion

Glioma is the most frequent malignant tumor in the brain.Consistent proliferation of tumor cells results in a sharpincrease of cell oxygen consumption, which causes anoxicmicroenvironment. It is known that HIF-1α plays an essentialrole in adapting the cellular environment to a hypoxic statusby inducing the expression of various hypoxia responsemolecules, including members of GLUT family. Scientistshave reported that increased HIF-1α is overexpressed as aresult of intratumoral hypoxia, which leads to treatmentfailure and a poor prognosis for the malignant tumors(Aebersold et al., 2001; Semenza, 2002). Hypoxia underlies anumber of biologic processes including cellular migration andinvasion. For example, oxygen deficiency inducesmigration ofendothelial cells (Meininger et al., 1988) and stimulatestransendothelial migration of monocytes (Kalra et al., 1996).Oxygen deficiency has also been shown to increase theinvasive capacity of murine sarcoma, carcinoma, and mela-noma cells (Cuvier et al., 1997). Our studies have identifiedthat HIF-1α expression increased with the increase of patho-logic grade in glioma (Fig. 1A–1D). Our results can also beinterpreted as follows: oxygen deficiency may be associatedwith invasion of gliomas, which needs further investigation.The correlation between HIF-1α expression and pathologicalgrade of glioma suggests that HIF-1α expression also should beup-regulated in response to oncogene activation and/or tumorsuppressor gene inactivation during glioma cells progression.For instance, loss of p53 activity has been testified to promote

between different pathological grades (P-value).

Grade II vs. III Grade II vs. IV Grade III vs. IV

0.019 <0.001 0.0270.020 <0.001 0.013

Fig. 3 – HIF-1α/β-actin and GLUT-3/β-actin in glioma ofdifferent grades groups in Western blot. (3A) HIF-1α/β-actin,(3B) GLUT-3/β-actin.

Fig. 2 – Expression of HIF-1α and GLUT-3 in glioma ofdifferent grades in Western blot.

152 B R A I N R E S E A R C H 1 3 0 4 ( 2 0 0 9 ) 1 4 9 – 1 5 4

HIF-1α expression, leading to increased VEGF-mediated tumorangiogenesis (Ravi et al., 2000). We found that HIF-1αimmunohistostaining was heterogeneous with signal concen-trated primarily within the nucleus. The patterns of immu-nohistochemical staining in gliomas of different pathologicalgrades suggest that HIF-1α overexpression may result fromboth physiological and nonphysiological mechanisms. It isobviously clear from previous studies that a lot of humantumors have regions of significant hypoxia (Brizel et al., 1996;Hockel et al., 1996). In gliobalstoma multiforme in whichHIF-1αwas detected in energetic tumor cells that were closestto areas of necrosis and farthest from a blood vessel, thispattern was most evident, as previously demonstrated by theexpression of VEGF mRNA in some tumors (Chan et al., 1998).In addition, under normoxic conditions, HIF-1α undergoesdegradation quickly due to VHL-mediated ubiquitin-protea-some system; therefore, it is hard to be detected (Semenza,2002). In our study, there was no expression of HIF-1α innormal brain tissues, which was consistent with what wasreported in the previous research (Zhong et al., 1999).

The GLUT-3 protein is considerably conserved, ubiquitousmolecules involved in a variety of biologic events, such as cellcycle control and apoptosis. Glucose transport is the rate-limiting factor. The expression of GLUT-3 was found inprimary human malignant tumors. In addition, glioma cellgrowth is related to glucose transport, which is closelyassociated with glucose metabolism (Nagamatsu et al., 1996).GLUT-3, which is classified as brain-type glucose transporter,has been shown to be localized particularly in cortical neuronand hippocampus (Nagamatsu et al., 1993; Vannucci et al.,1997). However, there was no expression of GLUT-3 in normalbrain tissues in our study. It might be related with that of thecontrol normal brain tissues, which were almost composed ofnormal white matter adjacent to glioma in decompression,and contained little special functional location for GLUT-3.Boado et al. (1994) concluded that GLUT-3 status is anindependent prognostic factor for the response to chemothe-rapy in advanced glioma, especially high-grade glioma. Theirstudy proved that in malignant glial cells, the expression levelof GLUT-3 correlated with the biologic aggressiveness of thetumor. Our data, which closely paralleled these observations,testified that GLUT-3 was more detectable in the high-gradegliomas than low-grade gliomas (Fig. 1E-1H), which indicatedthat GLUT-3 may be the predominant glucose transporter inGLUT family of high malignant human glioma.

Furthermore, hypoxia stimulates glucose transport byenhancing the function of the glucose transporters GLUT-1,

GLUT-3, and GLUT-4 (Zhang et al., 1999). Our immunohisto-chemical results showed that GLUT-3 expression wasconsiderably related with HIF-1α expression, as most ofGLUT-3-positive areas were contained within HIF-1α-positiveareas. We performed Western blot analysis to confirm theaccuracy and significance of the immunohistochemicalresults. Protein expressions of both HIF-1α and GLUT-3detected in the Western blot were testified to be consistentwith the immunohistochemical results (Figs. 2, 3A, B). In ourstudy, there was a positive correlation between HIF-1α andGLUT-3 to some extent. These facts demonstrated theexpression of both two markers closely corresponded topathological grade of glioma, and there are some mecha-nisms about tumor formation and progression betweenHIF-1α and GLUT-3 in glioma. Therefore, we believe thatHIF-1α and GLUT-3 in glioma are a pair of reliable hallmarkindicating glioma progression.

Some experts found that HIF-1α overexpression indicates adiminished prognosis in oligodendrogliomas (Vordermark,2002). Other regarded hypoxia score as an independentprognostic factor for all of the primary tumors (Birner et al.,

153B R A I N R E S E A R C H 1 3 0 4 ( 2 0 0 9 ) 1 4 9 – 1 5 4

2004). Currently, reports about relation between GLUT-3 inglioma and survival rate are scarcely found. We are workingon analyzing the association between the prognosis and theoverexpression of both two markers by post-operative follow-up of glioma patients.

HIF-1α suppression has been achieved through the admin-istration of a certain agent in vitro (Hudson et al., 2002).Increased HIF-1α activity promotes tumor progression, where-as inhibition of HIF-1α could represent a novel approach tocancer therapy (Dai et al., 2003). The inhibition of GLUT-3 isalso expected to suppress glioma growth. Sanchez-Alvarezet al. (2005) reported a unique strategy, which involvedthe increase in gap junctional communication caused bytolbutamide and dbcAMP, resulting in a rapid inhibition in thefunction of both GLUT-1 and GLUT-3. This inhibitiondecreased energy supply to glioma cells by a change in thelocalization of both type I and type II hexokinase. We candeduce the result from the literatures abovementioned thatsuppressing the expression of GLUT-3 and HIF-1α may be animportant biological therapy for glioma.

In conclusion, our study suggests that HIF-1α and GLUT-3expression tend to be correlated, but with significant differ-ences between them. There are certain unknownmechanismssuch as the regulative relationship between HIF-1α andGLUT-3, which will be the focus of further study. HIF-1α andGLUT-3 can be taken as a pair of useful markers for predictionof the biological behavior and future treatment of glioma.

4. Materials and methods

4.1. Case Selection

With institutional review board approval, 120 paraffin-embed-ded human glioma samples used in this study were obtainedfrom patients admitted to the Department of Neurosurgery,The ThirdHospital ofMianyang between 2005 and 2007. A totalof 120patients [61males, 59 females; ranged from12 to 70 yearsold (mean±standard deviation [SD]: 36.69±13.47)] wereincluded in the study. All the samples of glioma were got atthe time of primary tumor excision. All of the 120 gliomasamples (grade I, 30 of grade II, 30 of grade III and 30 of grade IV)were definitely diagnosed by pathological examination. Everypatient was at the first visit without radiation therapy andchemotherapy. The samples in control group were got fromnormal brain tissue adjacent to the neoplasm of patients [5males, 5 females; aged from 22 to 41 years ([SD]: 32.93±6.28)].The normal brain tissues directly adjacent to the neoplasmwere available for this study as control samples. The histo-pathological diagnosis and pathological grade of the 120tumorswere identified by two pathologists separately, accord-ing to the classification criteria of central nervous systemtumors established by WHO in 2007 (Louis et al., 2007).

4.2. Immunohistochemistry

Formalin-fixed and paraffin-embedded tissue blocks were cutinto 4-μm-thick sections. Paired sections from the sameparaffin-embedded sample were stained by anti-HIF-1α andanti-GLUT-3 antibody separately for comparison of expres-

sions of the two different proteins. The deparaffinizedsections were pretreated with 0.3% H2O2 in the methanol for30 min at room temperature to prevent endogenous peroxi-dase activity. After these sections were blocked with 0.1 Mphosphate-buffered saline (PBS) plus 3% skimmed milk andthe relevant 3% normal serum for 2 h at room temperature, theanti-HIF-1α (Santa Cruz, CA) and anti-GLUT-3 antibody(CHEMICON,CA) diluted in 0.1 M PBS (1:200) was applied ontothe sections, which were incubated for 48 h at 4 °C in ahumidified chamber. These samples were reacted with therelevant biotinylated secondary antibody (Sigma) diluted in0.1 M PBS (1:500) for 2 h at room temperature. Subsequently,these sections were incubated with an avidin–biotin peroxi-dase complex (Sigma) diluted in 0.1 M PBS (1:500) for 2 h atroom temperature. These sections were developed in astaining solution (DAB). All steps were manipulated accordingto operation manual. These sections were examined under alight microscope (Zeiss, Germany) for a satisfactory developedstaining, and hematoxylin was used for counterstaining tofacilitate visualization of the immunostained product. Inaddition, normal serum from the same animal speciesreplaced the primary antibody as the negative controls.

4.3. Immunohistochemical evaluation

Microscopic analysis was independently performed by twoneuropathologists. For each slide, 10 high power (400×) fields(about 1000 cells) were at random picked for quantification.HIF-1α-positive tumor cells (HIF-1α-PP) and HIF-1α stainingintensity (HIF-1α-SI) were estimated as follows: HIF-1α-PP: 0point, <1%; 1 point, 1–25%; 2 points, 26–50%; 3 points, 51–75%; 4points, >75%. HIF-1α-SI was rated on a scale of 0 to 3 with 0representing for no staining, 1 point representing for weakstaining, 2 points representing for moderate staining, 3 pointsrepresenting for strong staining. HIF-1α-IRS was calculated bymultiplication of HIF-1α-PP and HIF-1α-SI. HIF-1α-IRS of 0 wasscored as no staining (−), 1–3 points, weak intensity (+); 4–6 points, moderate intensity (++); 8–12 points, strong intensity(+++). The GLUT-3 immunoreactivity score (GLUT-3-IRS) wasobtained by the same method.

4.4. Western blot analysis

Frozen tissues were homogenized and lysed with RIPA lysisbuffer. Protein concentration was detected using the BioRadprotein assay kit (Hercules, CA) and measured using themicro-BCA protein assay (Pierce, Rockford, IL). Samples wereresolved by electrophoresis on a 12% SDS-PAGE gel. Theproteins were then transferred onto a nitrocellulose mem-brane, and nonspecific binding was blocked with 5% non-fatmilk in TBST buffer at room temperature for 1 h before beingfurther incubated with the specific anti- HIF-1α antibody(Santa Cruz, CA) or anti-GLUT-3 antibody (CHEMICON,CA ) for1.5 h at room temperature. These specific antibodies wereprobed with HRP-anti-rat and HRP-anti-rabbit IgG (Santa Cruz,CA). Signal was detected using ECLWestern blotting detection(Amersham Life Science, Piscataway, NJ) and the luminolexcitation was imaged on X-ray film. Positive protein bandswere visualized by enhanced chemiluminescence and nor-malized by the level of β-actin protein.

154 B R A I N R E S E A R C H 1 3 0 4 ( 2 0 0 9 ) 1 4 9 – 1 5 4

4.5. Statistics

All statistic analyses were performed using SPSS® version 16.0software (SPSS Inc., Chicago, USA). The Kruskal-Wallis H testand the Mann-Whitney U test were used to compare thedifferences between four pathological grade gliomas and inpairs, and corrected P-values were calculated for multipletesting by the Bonferroni method. The Spearman rank testwas used to identify a correlation betweenHIF-1α and GLUT-3.A P-value <0.05 was considered statistically significant.

Acknowledgments

We thank Mrs. Ling Sun, Xiaoyan Chen and Juan Li for theirassistance in this research. Shaolan Yu and Jicheng Xu are alsoworth of great thanks for their professional help.

R E F E R E N C E S

Aebersold, D.M., Burri, P., Beer, K.T., Laissue, J., Djonov, V., Greiner,R.H., Semenza, G.L., 2001. Expression of hypoxia-induciblefactor-1alpha: a novel predictive and prognostic parameter inthe radiotherapy of oropharyngeal cancer. Cancer Res. 61,2911–2916.

Airley, R.E., Mobasheri, A., 2007. Hypoxic regulation of glucosetransport, anaerobic metabolism and angiogenesis in cancer:novel pathways and targets for anticancer therapeutics.Chemotherapy 53, 233–256.

Birner, P., Preusser, M., Gelpi, E., Berger, J., Gatterbauer, B., Ambros,I.M., Ambros, P.F., Acker, T., Plate, K.H., Harris, A.L., Hainfellner,J.A., 2004. Expression of hypoxia-related tissue factorscorrelates with diminished survival of adjuvantly treatedpatients with chromosome 1p aberrant oligodendroglialneoplasms and therapeutic implications. Clin. Cancer Res. 10,6567–6571.

Boado, R.J., Black, K.L., Pardridge, W.M., 1994. Gene expression ofGLUT3 and GLUT1 glucose transporters in human braintumors. Brain Res. Mol. Brain Res. 27, 51–57.

Brizel, D.M., Scully, S.P., Harrelson, J.M., Layfield, L.J., Bean, J.M.,Prosnitz, L.R., Dewhirst, M.W., 1996. Tumor oxygenationpredicts for the likelihood of distant metastases in human softtissue sarcoma. Cancer Res. 56, 941–943.

Chan, A.S., Leung, S.Y., Wong, M.P., Yuen, S.T., Cheung, N., Fan, Y.W., Chung, L.P., 1998. Expression of vascular endothelialgrowth factor and its receptors in the anaplastic progression ofastrocytoma, oligodendroglioma, and ependymoma. Am. J.Surg. Pathol. 22, 816–826.

Cuvier, C., Jang, A., Hill, R.P., 1997. Exposure to hypoxia, glucosestarvation and acidosis: effect on invasive capacity of murinetumor cells and correlation with cathepsin (L + B) secretion.Clin. Exp. Metastasis 15, 19–25.

Dai, S., Huang, M.L., Hsu, C.Y., Chao, K.S., 2003. Inhibition ofhypoxia inducible factor 1alpha causes oxygen-independentcytotoxicity and induces p53 independent apoptosis inglioblastoma cells. Int. J. Radiat. Oncol. Biol. Phys. 55,1027–1036.

Hockel, M., Schlenger, K., Aral, B., Mitze, M., Schaffer, U., Vaupel, P.,1996. Association between tumor hypoxia and malignantprogression in advanced cancer of the uterine cervix. CancerRes. 56, 4509–4515.

Huang, L.E., Gu, J., Schau, M., Bunn, H.F., 1998. Regulation ofhypoxia-inducible factor 1alpha is mediated by anO2-dependent degradation domain via theubiquitin-proteasome pathway. Proc. Natl Acad. Sci. USA 95,7987–7992.

Hudson, C.C., Liu, M., Chiang, G.G., Otterness, D.M., Loomis, D.C.,Kaper, F., Giaccia, A.J., Abraham, R.T., 2002. Regulation ofhypoxia-inducible factor 1alpha expression and function bythe mammalian target of rapamycin. Mol. Cell. Biol. 22,7004–7014.

Kalra, V.K., Shen, Y., Sultana, C., Rattan, V., 1996. Hypoxia inducesPECAM-1 phosphorylation and transendothelial migration ofmonocytes. Am. J. Physiol. 271, H2025–2034.

Louis, D.N., Ohgaki, H., Wiestler, O.D., Cavenee, W.K., Burger, P.C.,Jouvet, A., Scheithauer, B.W., Kleihues, P., 2007. The 2007 WHOClassification of Tumours of the Central Nervous System. ActaNeuropathol. 114, 97–109.

Mantych, G.J., James, D.E., Chung, H.D., Devaskar, S.U., 1992. Cellularlocalization and characterization of Glut 3 glucose transporterisoform in human brain. Endocrinology 131, 1270–1278.

Meininger, C.J., Schelling, M.E., Granger, H.J., 1988. Adenosine andhypoxia stimulate proliferation and migration of endothelialcells. Am. J. Physiol. 255, H554–H562.

Mobasheri, A., Richardson, S., Mobasheri, R., Shakibaei, M.,Hoyland, J.A., 2005. Hypoxia inducible factor-1 and facilitativeglucose transporters GLUT1 and GLUT3: putative molecularcomponents of the oxygen and glucose sensing apparatus inarticular chondrocytes. Histol. Histopathol. 20, 1327–1338.

Nagamatsu, S., Nakamichi, Y., Inoue, N., Inoue, M., Nishino, H.,Sawa, H., 1996. Rat C6 glioma cell growth is related to glucosetransport and metabolism. Biochem. J. 319 (Pt 2), 477–482.

Nagamatsu, S., Sawa, H., Kamada, K., Nakamichi, Y., Yoshimoto,K., Hoshino, T., 1993. Neuron-specific glucose transporter(NSGT): CNS distribution of GLUT3 rat glucose transporter(RGT3) in rat central neurons. FEBS Lett. 334, 289–295.

Ravi, R., Mookerjee, B., Bhujwalla, Z.M., Sutter, C.H., Artemov, D.,Zeng, Q., Dillehay, L.E., Madan, A., Semenza, G.L., Bedi, A., 2000.Regulation of tumor angiogenesis by p53-induced degradationof hypoxia-inducible factor 1alpha. Genes Dev. 14, 34–44.

Riedl, C.C., Akhurst, T., Larson, S., Stanziale, S.F., Tuorto, S.,Bhargava, A., Hricak, H., Klimstra, D., Fong, Y., 2007. 18F-FDGPET scanning correlates with tissue markers of poor prognosisand predicts mortality for patients after liver resection forcolorectal metastases. J. Nucl. Med. 48, 771–775.

Sanchez-Alvarez, R., Tabernero, A., Medina, J.M., 2005. Theincrease in gap junctional communication decreases the rateof glucose uptake in C6 glioma cells by releasing hexokinasefrom mitochondria. Brain Res. 1039, 189–198.

Semenza, G.L., 2002. HIF-1 and tumor progression:pathophysiology and therapeutics. Trends Mol. Med. 8, S62–67.

Vannucci, S.J., Maher, F., Simpson, I.A., 1997. Glucose transporterproteins in brain: delivery of glucose to neurons and glia. Glia21, 2–21.

Vordermark, D., 2002. Expression of hypoxia-induciblefactor-1alpha in oligodendrogliomas: its impact on prognosisand on neoangiogenesis. Cancer 94, 2317–2318 author reply2318-2319.

Warburg, O., 1956. On respiratory impairment in cancer cells.Science 124, 269–270.

Zhang, J.Z., Behrooz, A., Ismail-Beigi, F., 1999. Regulation ofglucose transport by hypoxia. Am. J. Kidney Dis. 34, 189–202.

Zhong, H., De Marzo, A.M., Laughner, E., Lim, M., Hilton, D.A.,Zagzag, D., Buechler, P., Isaacs, W.B., Semenza, G.L., Simons, J.W., 1999. Overexpression of hypoxia-inducible factor 1alpha incommon human cancers and their metastases. Cancer Res. 59,5830–5835.