glycogen synthase kinases-3β controls differentiation of malignant glioma cells

Glycogen synthase kinases-3b controls differentiation ofmalignant glioma cells

Yan Li1,2, Huimin Lu1, Yijun Huang1, Ru Xiao1, Xiaofeng Cai1, Songmin He1 and Guangmei Yan1

1 Department of Pharmacology, Zhongshan School of Medicine, Sun Yat-Sen University, Guangzhou, People’s Republic of China2 Department of Infectious Diseases, Third Affiliated Hospital of Sun Yat-Sen University, Guangzhou, People’s Republic of China

Malignant gliomas persist as a major disease of morbidity and mortality in adult. Differentiation therapy has emerged as a

promising candidate modality. However, the mechanism related is unknown. Here, we show that glycogen synthase kinase-3b

(GSK-3b) is highly expressed and activated during the cholera toxin-induced differentiation in sensitive C6 and U87-MG

malignant glioma cells, whereas the GSK-3a activity remains stable. GSK-3b inhibitors or small interfering RNA suppress the

induced-differentiation in sensitive C6 cells. Conversely, overexpression of a constitutively active form of human GSK-3b

(pcDNA3-GSK-3b-S9A) mutant in resistant U251 glioma cells restores their differentiation capabilities. In addition, GSK-3b

triggers cyclin D1 nuclear export and subsequent degradation, which is necessary for differentiation in C6 and U251 glioma

cells. Analysis of human glioma tissues further revealed overexpression of active GSK-3b. These findings suggest that GSK-3b

is a differentiation fate determinant, and shed new lights on the mechanism by which GSK-3b regulates cyclin D1 degradation

and cellular differentiation in gliomas.

Gliomas are the most common primary tumors of the adultcentral nervous system (CNS). The majority of gliomas inadults are highly malignant with a poor prognosis, in particu-lar with high-grade tumors such as glioblastoma multiforme.1

Current therapies such as irradiation and chemotherapy areoften ineffective and fail to improve its prognosis.2,3 Differen-tiation-inducing therapy, which modifies cancer cell differen-tiation and is the most successful for the treatment of acutemyelocytic leukemia,4 has been proposed to be a novel poten-tial approach to treat malignant tumors.5 The biotoxin chol-era toxin is reported to be capable of inducing differentiationof malignant gliomas.6 However, little progress has beenmade about the critical mechanism underlying the function

of this differentiation. And more importantly, the key factorsgoverning glioma differentiation, and therefore tumor malig-nancy, are still unknown.

Glycogen synthase kinase-3 (GSK-3) is an evolutionaryconserved, ubiquitous serine/threonine kinase that is highlyenriched in the brain and consists of 2 distinct isoforms, aand b, in mammals.7 One of the most notable qualities ofGSK-3 is the vast number of signaling pathways that con-verge on this enzyme and subsequently an even greater num-ber of biological targets.8,9 Numerous studies have indicatedthat GSK-3 is involved in key functions of the brain and isassociated with dysfunction in multiple neurological dis-eases.10 More recent studies indicate a role for GSK-3 in thecontrol of neoplastic transformation and tumor development,suggesting that GSK-3 is a potential therapeutic target inhuman cancers. However, most attention has focused on theb-isoform of GSK-3 and the exact role of GSK-3b in malig-nancies remains highly controversial due to the conflictingresults from different tumor models. Although some studiesfound that GSK-3b is a part of a tumor suppressor complexthat phosphorylates the oncoprotein b-catenin and that GSK-3b inactivation could possibly lead to tumor promotion,11,12

other studies have shown that inhibition of GSK-3b sup-presses cancer cell proliferation and induces apoptosis byabrogating nuclear factor NF-jB mediated gene transcrip-tion.13,14 In addition, the vast majority of research on GSK-3has been focused on the aspects of proliferation and apopto-sis, and little is known about its possible role involved in theprocess of cancer cell differentiation.

Regulation of cell proliferation and terminal differentiationis critical for normal development and homeostasis, but isfrequently disturbed during tumorigenesis. Positive regulationof cell cycle proteins (cyclins, cyclin-dependent kinases and

Key words: GSK-3b, differentiation, malignant gliomas, cholera toxin

Additional Supporting Information may be found in the online

version of this article

Yan Li, Huimin Lu and Yijun Huang contributed equally to this work

Grant sponsor: National Natural Science Foundation of China;

Grant numbers: 30830111, 30801408; Grant sponsor: National

Natural Science Foundation of Guangdong Province;

Grant number: 8451008901000297; Grant sponsor: Chinese

Postdoctoral Science Foundation; Grant numbers: 20080430801,

200801266; Grant sponsor: Guangzhou Scientific and

Technological Programs; Grant number: 2008Z1-E561

DOI: 10.1002/ijc.25020

History: Received 8 Jun 2009; Revised 30 Sep 2009; Accepted 27 Oct

2009; Online 30 Oct 2009

Correspondence to: Guangmei Yan, Department of Pharmacology,

Zhongshan School of Medicine, Sun Yat-Sen University, Guangzhou,

People’s Republic of China, Tel.: (86) 20-87333732,

Fax: 86-20-87330578, E-mail: [email protected]

Can

cerCellBiology

Int. J. Cancer: 127, 1271–1282 (2010) VC 2009 UICC

International Journal of Cancer

IJC

their inhibitors) promotes proliferation, whereas inhibition ofthese cell cycle proteins results in differentiation.15 D-typecyclins, among which cyclin D1 remains to be the mostextensively studied, play important roles in cell cycle controlat the G1/S boundary and overexpress in various human can-cers, including gliomas.16,17 Sustained nuclear localization ofcyclin D in NIH3T3 cells renders them tumorigenic.18 Theseresults have led to the proposal that cyclin D1 subcellularlocalization and degradation may participate in cancer celldifferentiation.

Here, we show intensive GSK-3b expression and activa-tion during differentiation in C6 and U87-MG malignant gli-oma cells. Inhibition of GSK-3b activity or knockdown of itsexpression by small interfering RNA (siRNA) suppresses thisdifferentiation, whereas constitutively active S9A-GSK-3b ini-tiates robust differentiation in differentiation-resistant U251glioma cells. Ubiquitin/proteasome is identified the majorpathway responsible for cyclin D1 reduction and sequentialdifferentiation. We also provide evidence that GSK-3b regu-lates differentiation by triggering cyclin D1 translocation anddegradation.

Material and MethodsCell culture and drug treatment

Cells were maintained in Dulbecco’s modified Eagle’s Me-dium (DMEM, Invitrogen, Grand Island, NY) supplementedwith 10% FBS (Invitrogen) in a humidified atmosphere of 5%CO2 at 37�C. Human malignant glioma tissue samples wereacquired from 10 astrocytic neoplasms patients undergoingsurgery with approval by the Ethical Committee of Sun Yat-Sen University. Tumors were classified according to theWHO classification system as anaplastic astrocytoma (WHOGrade III; 7 tumors) and glioblastoma multiforme (WHOGrade IV; 3 tumors). Primary cultures of human glioma cellswere prepared as previously described.6 Cell differentiationwas induced with cholera toxin (Sigma, St. Louis, MO) inDMEM containing 1% FBS. Control was treated with anequivalent volume of DMEM containing 1% FBS. For pri-mary human glioma cells, cholera toxin was added to the10% FBS-medium.

Morphological evaluation

The morphologies of cells were studied using an Olympus(Melville, NY) IX71 inverted microscope along with anOlympus DP Controller software.

Cell cycle analysis

A flow cytometry analysis of DNA content of cells was per-formed to assess the cell cycle phase distributions asdescribed.19 Cells were analyzed by a FACSCalibur flow cy-tometer (BD, Heidelberg, Germany) using a peak fluores-cence gate to discriminate aggregates.

Western blot analysis

After lysis of cells and measurement of protein concentration,the cells were dissolved in SDS sample buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 50 mM DTT and 0.1%bromophenol blue). Equal amount of proteins were analyzedby SDS-PAGE on 12% polyacrylamide gels. Proteins wereelectroblotted on a nitrocellulose membrane. Membraneswere incubated in 5% nonfat dry milk in TBST (Tris-bufferedsaline, 0.05% Tween-20) and then overnight at 4�C with anti-bodies against glial fibrillary acidic protein (GFAP), cyclinD1, p-GSK-3a, GSK-3a, p-GSK-3b, GSK-3b, GAPDH(1:1,000, Cell Signaling Technology, Beverly, MA), PCNA(1:10,000, Cell Signaling Technology) and b-actin (1:2,000,New England Biolabs) respectively. After incubation withhorseradish peroxidase-labeled secondary antibody (1:1,000,Cell Signaling Technology), visualization was achieved withenhanced chemiluminescence (Amersham Pharmacia Bio-tech) using a GeneGnome chemiluminescence imaging andanalysis system (Syngene Bio Imaging, Cambridge, UK).

Reverse transcription-polymerase chain reaction (RT-PCR)

Total RNA was isolated with TRIzol reagent (Gibco-BRL, USA)according to the manufacturer’s instructions. First strand cDNAwas synthesized using the SuperscriptTM II reverse transcriptasekit (Gibco-BRL) with oligo-dT as the primer following the man-ufacturer’s instructions. Sequences of the PCR primers are asfollows: Cyclin D1: 50-CCC TCGGTGTCCTACTTCAAA-30

(sense primer), 50-CAC CTCCTCCTCCTCCTCTTC-30 (anti-sense primer). GAPDH: 50-TCACCATCTTCCAGGAGC-GAGA-30 (sense primer), 50-ATGAGCCCTTCCACGATGC-30

(antisense primer). The PCR conditions contained an initialcDNA synthesis reaction at 48�C for 1 h, followed by a denatu-ration step for 2 min at 94�C and 32 cycles: 1 min at 94�C, 1min at 57�C and 1 min at 72�C. After the last cycle, the finalextension was performed at 72�C for 7 min. PCR product wasanalyzed by agarose gel electrophoresis and visualized with ethi-dium bromide staining.

RNA silencing of GSK-3bRat GSK-3b siRNA and negative siRNA (control) were pur-chased from Invitrogen. The coding strand sequence of theGSK-3b siRNA is 50-GCAGCAAGGUAACCACAGU-30. C6cells with 30–50% confluence were transfected using Lipofect-amine 2000 reagent (Invitrogen) following the manufacturer’sprotocol. Inhibition of GSK-3b protein expression wasassessed by immunoblot analysis 48 h after transfection.Thirty-six hours post-transfection, cells were incubated withcholera toxin for additional 48 h for further analysis.

Transient transfection

pcDNA3-GSK-3b S9A mutant is a kind gift of Prof. Z. Mao,Emory University. Asynchronously growing cells were platedin 6-well dishes 24 h before transfection. Transient transfec-tions were performed with 4 lg plasmid DNA per well using12 ll FuGENE HD transfection reagent (Roche DiagnosticsCorp, Mannheim, Germany). To monitor transfection

Can

cerCellBiology

1272 GSK-3b controls differentiation of glioma cells


efficiency, the constructs were cotransfected with pEGFP vec-tor and fluorescent cells were counted 48 h post-transfection.Such analyses indicated that 50–60% of the cell populationexpressed the foreign protein.

GSK-3b kinase assay

Cell extracts were prepared in cell lysis buffer (20 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1 mM EGTA,1% Triton X-100, 2.5 mM sodium pyrophosphate, 5 mM b-glycerolphosphate, 1 mM Na3VO4, 1 lg of leupeptin permilliliter, 1 mM phenylmethylsulfonyl fluoride and 1 lMmicrocystin) for 15 min at 4�C. After brief sonication, thelysates were clarified by centrifugation at 15,000g for 10 minat 4�C, and GSK-3b from 200 lg of cell extract was immu-noprecipitated with 1.0 lg of GSK-3b antibody for 2 h at4�C with rotation. Protein A-agarose (20 ll of a 50% suspen-sion) was then added, and the incubation was continued for1 h at 4�C with rotation. Immune complexes were recoveredby centrifugation at 4�C and were washed 3 times withextraction buffer and twice with kinase buffer. Kinase activityof the immunoprecipitated GSK-3b was assayed in a totalvolume of 40 ll containing 25 mM sodium glycerophosphate,20 mM Tris-HCl [pH 7.4], 10 mM MgCl2, 5 mM DTT,20 lM phosphoglycogen synthase peptide 2 (Upstate Bio-technology) and 50 lM [c-32P]ATP (1 lCi, Amersham Phar-macia Biotechnology). After 10 min of incubation at 30�C,reaction mixtures were centrifuged for 1 min, and 20 ll ofthe supernatant was spotted onto P81 phosphocellulose paper(Whatman, Piscataway, NJ), which was washed 4 times (30min each) in 0.6% phosphoric acid and dried, and boundradioactivity was quantified by scintillation counting. A sec-ond sample was assayed in the presence of 10 mM LiCl andspecific activity was calculated by subtraction of this sample(background) from the original.

Immunofluorescence, immunohistochemistry and

confocal microscopy

Immunofluorescent staining of primary cells and immunohis-tochemistry on 4 lm sections of paraffin-embedded sampleswere performed as described.20 To evaluate the GSK-3a,GSK-3b and p-GSK-3bY216 levels, immunostained slides werescored with the Image Pro Plus 6.0 software. The signal wasscored based on the percentage of cells with GSK-3a/GSK-3b/p-GSK-3bY216 staining on the following scale: þ, positivestaining for <10% of all tumor cells; þþ, positive stainingfor 10–60%; þþþ, positive staining for >60%.

For confocal analysis, immunostaining was observedunder a Zeiss LSM 510 Confocal Microscope (Carl ZeissMicroimaging, Thornwood, NY).

Statistical analysis

Data are presented as mean 6 standard deviation (SD) of 3separate experiments. Statistical significance was determinedby Student’s t-test. A result with a p-value of less than 0.05 isconsidered statistically significant.

ResultsCholera toxin induces differentiation of C6 and

U87-MG glioma cells

We have previously showed that cholera toxin induces differ-entiation in rat C6 and primary cultured human gliomacells.6 Here in established human malignant glioma cell lines,we further examined the induction of differentiation on ex-posure to cholera toxin. Different glioma cell lines exhibiteddifferent sensitivities to treatment with cholera toxin: Induc-tion of differentiation, characterized by morphologic changes,increase in GFAP expression and growth arrest, was observedin U87-MG cells after exposure to cholera toxin (data notshown). However, U251 and T98G malignant glioma cellswere resistant to the differentiation reagent cholera toxin(Fig. 4 and data not shown). This indicates that cholera toxininduces differentiation of certain malignant glioma cells intothe maturation process of astrocytic lineage and provides areliable model of differentiation.

Elevated expression and activation of GSK-3b in

differentiation-sensitive glioma cells

Numerous cellular stimuli cause the inactivation of GSK-3via the phosphorylation of an N-terminal serine residue, Ser-21 in GSK-3a and Ser-9 in GSK-3b. GSK-3a phosphorylationat Ser-21 and GSK-3b phosphorylation at Ser-9, which repre-sent the inactive form of GSK-3 kinase, were then deter-mined. As shown in Figure 1a, both phosphorylated and totalGSK-3a protein levels were essentially unchanged after 24 hincubation with cholera toxin in C6 cells. Meanwhile, by 6 hafter cholera toxin stimulation, there was a decrease in theintensity of the phosphorylated form of GSK-3b and minimalexpression of the protein occurred 12 h after treatment, whiletotal GSK-3b remains stable throughout. The observedchange in the phosphorylation of GSK-3b was further con-firmed by in vitro kinase assays, which were carried out fol-lowing immunoprecipitation of GSK-3b using phosphoglyco-gen synthase peptide 2 as the substrate. Figure 1b showedthat cholera toxin elevated GSK-3b activity by about 2.5-foldafter 12 h incubation. Similar results are also observed in thesensitive U87-MG cells (data not shown). These results indi-cate that GSK-3b, but not GSK-3a, is involved in choleratoxin-induced differentiation of malignant glioma cells.

To examine whether the activation of GSK-3b after expo-sure of cells to cholera toxin was a regulator of differentia-tion, we further examined GSK-3b expression in differentia-tion-sensitive (C6 and U87-MG) and resistant (U251 andT98G) glioma cells. We observed low levels of GSK-3bexpression in the resistant U251 and T98G glioma cells, com-pared with the robustly increased GSK-3b levels in the sensi-tive C6 and U87-MG cells (American Type Culture Collection,Manassas, VA) (p < 0.01, p < 0.05 compared with C6 andU87-MG cells, respectively, Fig. 1c). Our results suggest thatGSK-3b overexpression and activation possibly be a contribu-tory event in the process of cellular differentiation in malig-nant gliomas.

Can

cerCellBiology

Li et al. 1273


GSK-3 inhibitors suppress differentiation of C6 glioma

cells in a PI3K/PKB dependent pathway

To further confirm the role of GSK-3b in glioma cell differen-tiation, we examined the effects of GSK-3 inhibitors, lithiumchloride (LiCl, Sigma) and SB216732 (Sigma) on cholera toxin

induced-differentiation of C6 cells. As shown in Figure 2a, themorphological transformation induced by cholera toxin wasattenuated by inhibition of GSK-3b activity by LiCl andSB216732, while exposure alone to LiCl and SB216732 did notcause any alterations. Moreover, upregulation of GFAP anddownregulation of PCNA induced by cholera toxin werereverted by LiCl and SB216732 (Fig. 2b). Meanwhile, the cellcycle alterations (G1 cell cycle arrest and accelerated S-phaseentry) were markedly counteracted by LiCl and SB216732 (Fig.2c), demonstrating the crucial role of active GSK-3b in the dif-ferentiation of malignant glioma cells.

GSK-3b can be phosphorylated by multiple kinases, princi-pally from the phosphatidylinositol 3-kinase (PI3K)/protein ki-nase B (PKB) pathway.21 To elucidate the involvement ofGSK-3b upstream molecules in glioma cell differentiation, weassessed the expression of major proteins of the PI3K/PKB sig-naling pathway. Cholera toxin significantly inhibited PKBphosphorylation at Ser 473 and GSK-3b phosphorylation atSer 9 (data not shown). To assess a possible contribution ofPI3K/PKB inhibition to differentiation in C6 cells, we usedLY294002, a PI3K inhibitor. This inhibitory of PI3K/PKB acti-vation caused pronounced elongation of cellular processes andG1 cell cycle arrest similar to that previously observed on chol-era toxin-induced differentiation (data not shown). Moreover,the expression of GFAP in C6 cells was markedly upregulatedand PCNA downregulated in the presence of LY294002 (Fig.2b). These data indicate that PI3K inhibition probably partici-pated in the GSK-3b mediated differentiation of C6 cells.

GSK-3b gene knockdown abrogates differentiation ability

of C6 glioma cells

It has been shown that LiCl and SB216732 have multiplenoncompetitive targets in addition to GSK-3.22 To selectivelydownregulate expression of GSK-3a and GSK-3b proteins,we next evaluated the effect of isoform-specific GSK-3a andGSK-3b gene silencing mediated by siRNAs. As shown inFigure 3a, cells transfected with #2 GSK-3b siRNA (Invitro-gen) resulted in the most dramatic knockdown of the proteinlevels (18.4%, p < 0.01), whereas transfection with mock ornegative siRNA had no effect. Moreover, #2 GSK-3b siRNArestrained the GSK-3b protein levels in a dose-dependentmanner, with a maximum effect at concentration of 10 nM(Fig. 3b). In addition, morphological transformation and ele-vation of GFAP during C6 cell differentiation were blockedby 10 nM GSK-3b specific siRNA (Figs. 4c and 4d). At thesame time, silencing of GSK-3a did not significantly affectthe morphological transformation and elevation of GFAPduring C6 cell differentiation (Supporting Information Figure1). These results further indicate the requisite role of GSK3bin the differentiation of glioma cells and suggest that GSK-3bpositively regulates the differentiation.

Overexpression of active GSK-3b restores differentiation

capacity of resistant U251 glioma cells

We next examined whether forced expression of GSK-3b issufficient to induce differentiation in resistant U251 glioma

Figure 1. GSK-3 is highly expressed in C6 glioma cells and

activated by cholera toxin. (a) Immunoblot of p-GSK-3bser9,

GSK-3b, p-GSK-3aser21 and GSK-3a levels in C6 cells treated with

10 ng/ml cholera toxin (CT) for the time indicated. (b) After

incubation with 10 ng/ml CT for 12 h, C6 cells were lysed and

GSK-3b was immunoprecipitated. GSK-3b activity was determined

as described in ‘‘Material and Methods’’ section. The results are

expressed as fold activity of control cells and are mean values

6 SD from 3 experiments. **Statistical significance according to

Student’s t-test (p < 0.01 compared with 0 h). (c) Immunoblot of

the GSK-3b levels in the indicated cell lines. The graph represents

the mean 6 SD (bars) of 3 independent experiments. **p < 0.01

compared with C6, #p < 0.05 compared with U87-MG.

Can

cerCellBiology



cells (Cell Bank, Chinese Academy of Science, Shanghai,China) (Figs. 5a and 5b), which expresses lower levels of en-dogenous GSK-3b than sensitive C6 cells (Fig. 1b). Cells weretransfected with a constitutively active pcDNA3-GSK-3b S9Amutant (GSK-3b S9A) in which the N-terminal serine-9 resi-due was substituted with an alanine residue, and thereforecannot undergo inhibitory phosphorylation. As seen in Figure4c, transfection of GSK-3b S9A initiates morphological trans-formation from ovate or polygon appearance in the emptyvector group to smaller cell bodies with much longer, fine,tapering processes, similar to that of mature astrocytes. West-ern-blotting analysis further confirmed significant upregula-tion of GFAP protein expression in GSK-3b S9A gene trans-ferred cells compared with those in the empty vector controls(Fig. 4d). Meanwhile, the level of PCNA in GSK-3b S9Atransfected cells was markedly reduced (Fig. 4d), i.e., U251cells transfected with GSK-3b S9A exhibited hallmarks of dif-ferentiation. These data indicate that overexpression of activeGSK-3b is necessary and sufficient to promote differentiationin resistant U251 glioma cells.

Cyclin D1 degradation is required for glioma cell

differentiation

Since the cholera toxin-induced differentiation of C6 cellsleads to G1 arrest of a majority of cancerous cells,6 cyclinD1, a specific early G1 phase cell cycle regulatory protein isinvestigated. Figure 5a depicts the remarkable reduction ofthe cellular cyclin D1 protein levels during the time course of6 h to 48 h after cholera toxin stimulation. However, themRNA levels of cyclin D1 remained unaltered after 48 h ex-posure as detected by RT-PCR (Fig. 5b), suggesting that apost-transcriptional mechanism might underlie the reducedlevel of cyclin D1 protein during glioma cell differentiation.One such system, cytoplasmic ubiquitin-proteasome degrada-tion in the control of cyclin D1 has been recently reported inmouse embryo fibroblasts.23

To assess whether blockage of this pathway in glioma cellscould interfere with cyclin D1 turnover and the regulatoryeffects of cell differentiation, we used MG132 (benzyloxycar-bonyl-leucyl-leucyl-leucinal, Sigma), the most specific protea-some inhibitor available currently24 in our study. The

Figure 2. Pharmacological inhibition of GSK-3 represses differentiation of C6 glioma cells in a PI3K/PKB dependent pathway. C6 cells were

pretreated with 10 mM LiCl or 10 lM SB216732 for 2 h and then treated with CT for further 48 h. (a) Morphological transformation,

(b) GFAP and PCNA expressions and (c) cell cycle distributions.

Can

cerCellBiology

Li et al. 1275


morphological transformation, cell cycle alterations (G1 arrestand accelerated S-phase entry), upregulation of GFAP anddownregulation of PCNA induced by cholera toxin were atte-nuated by MG132 while exposure alone to MG132 had noeffect on these parameters (Figs. 5c–5e). These observationsdemonstrate that cholera toxin leads to decline of cyclin D1through the ubiquitin-proteasome pathway, which determinesthe degradation of this protein and may counteracted cellulardifferentiation in C6 glioma cells.

Active GSK-3b translocation triggers cyclin D1 nuclear

export and cytoplasmic degradation

To further explore the mechanism underlies cytoplasmic deg-radation of cyclinD1 in glioma cell differentiation, the subcel-lular distribution of GSK-3b and cyclin D1 is investigated.GSK-3b has been shown to remain in the cytoplasm in aninactivated form. However, once activated, it is translocatedinto the nucleus to phosphorylate cyclin D1 at threonine 286,and then the phosphorylated cyclin D1 is in turn excludedfrom the nucleus to the cytoplasm, where the phosphorylatedcyclin D1 is degraded through a proteasome-mediated path-way.25,26 We assayed the subcellular distribution of GSK-3band cyclin D1 proteins by immunofluorescent labeling of C6cells treated with cholera toxin. As shown in Figure 6a-a0,

GSK-3b was most abundant in the cytoplasm of unstimulatedcells, but was then found in the nucleus after choleratoxin stimulation. In contrast, cyclin D1 could hardly bedetected in cholera toxin-exposed C6 cells presumablybecause of its reduced stability (Fig. 6a-d0). Blockade of theproteosome-mediated degradation pathway with MG132revealed that cholera toxin-induced cyclin D1 translocationfrom the nucleus to the cytoplasm in C6 cells (Figs. 6a-c0

and 6a-e0).To further confirm that the nuclear exit of cyclin D1 dur-

ing differentiation induced by cholera toxin in glioma cells isdue to the activation of GSK-3b, we evaluated cyclin D1localization in GSK-3b gene-deleted C6 cells. As seen in Fig-ures 6b-a0 and 6b-f 0, GSK-3b was almost completely knockeddown by siGSK-3b in contrast to the negative control siRNA.GSK-3b knockdown cells showed that cyclin D1 localize pre-dominantly within the nuclear fraction (Fig. 6b-h0), similar tothat of control cells (Fig. 6b-c0). Echoing with suppressed dif-ferentiation after GSK-3b knockdown (Fig. 3), cyclin D1cytoplasmic migration was almost completely abolished andmainly restricted to nucleus in GSK-3b-silencing cells (Fig.6b-i0). While in the absence of MG132, cyclin D1 was foundto be localized within both the nucleus and cytoplasm (Fig.6b-j0).

Figure 3. Silencing GSK-3b blocks differentiation of C6 cells. (a, b) Immunoblot of the GSK-3b protein levels after transfection with negative

control (Neg), different number (a) or indicated concentration (b) of siGSK-3b or no-treat (No) for 36 h. Mock control cells received

transfection reagent only. (n ¼ 3, t-test, *p < 0.05, vs. the Neg). (c, d) Morphology (c) and immunoblot of the GFAP levels (d) in GSK-3b

knockdown cells subsequently stimulated with CT for 48 h.

Can

cerCellBiology



In contrast, recovery differentiation capacity of resistantU251 cells by mutant active GSK-3b gene transfection (Fig.4) results in increased nuclear GSK-3b accumulation com-pared with the pcDNA3 vector control (Fig. 6c-a0 and 6c-b0).This accumulation of GSK-3b is accompanied by activecyclin D1 translocation from nucleus to cytoplasm, asdetected in the presence of MG132 (Figs. 6c-c0–6c-e0).

The levels of cyclin D1 mRNA and protein were thenexamined. siGSK-3b subdued the cholera toxin-induced deg-radation of cyclin D1 protein (Fig. 6d) in C6 cells. However,there is a lack of regulation of cyclin D1 mRNA levels bysiGSK-3b (Fig. 6d). In addition, transfection of a constitu-tively active S9A GSK-3b mutant initiates a decline of cyclinD1 protein without affecting its mRNA levels in U251 cells(Fig. 6d). These results indicate that the degradation of cyclinD1 is mediated through a GSK-3b dependent pathway.

Altogether, these findings provide biochemical and cellularevidence showing that cyclin D1 nuclear-to-cytoplasmicrelocalization and subsequent degradation during differentia-tion is indeed regulated by GSK-3b.

Expression and activity of GSK-3b in human

malignant gliomas

To extend our findings to clinical malignant gliomas, we fur-ther prepared a series of 10 primary cell cultures grown from7 Grade III and 3 Grade IV malignant astrocytoma explants.Differentiation was examined in the first passage in vitro ofthe primary cells. Exposure to the differentiation agent chol-era toxin also resulted in differentiated characteristics with astellar shape with filamentous processes and increased GFAPexpression in all the primary cultures examined (Fig. 7a).Furthermore, the phosphorylated form of GSK-3b shows thesame alteration panel as in the rat C6 cell line (Fig. 7b) andGSK-3b inhibitors LiCl and SB216732 could block theincreased GFAP and decreased PCNA levels induced by dif-ferentiation agent cholera toxin (Fig. 7c). These results con-firm our findings in C6 cells and, moreover, suggest a generalcorrelation of GSK-3b activity with differentiation in malig-nant glioma cells.

To further confirm the possibility that GSK-3b expressionis associated with sensitivity to differentiation therapy, weexamined GSK-3b expression in human glioma tissue sam-ples (WHO Grade III and IV) by immunohistochemicalstaining. Tissues were scored on the basis of the percentageof GSK-3b-positive tumor cells (see ‘‘Material and Methods’’section). Using immunohistochemistry, we found elevatedexpression of GSK-3b in the tumor samples compared withnormal brain tissues. In addition, overexpression of GSK-3bwas observed frequently in the cytoplasm of tumor cells (rep-resentative immunostaining image from Grade IV malignantglioma tissues shown in Fig. 7d), whereas weak expressiononly was observed in the cytoplasm of cells from normalbrain. p-GSK-3bY216 that represents the active kinase form ofGSK-3b was also detected to assess the activation state of theprotein. Figure 7d showed a higher expression of p-GSK-3bY216 in malignant gliomas compared with normal brain tis-sues (representative image from Grade IV malignant gliomatissues). The immunohistochemical results of GSK-3b expres-sion and its Y216 phosphorylation in patients were summar-ized in Supporting Information Table 1. Meanwhile, primarycultured human malignant glioma cells were facile to beinduced-differentiation (Fig. 7a). Together, it is likely thatoverexpression of active GSK-3b is a pathological characteris-tic of clinical malignant gliomas, which are sensitive to theinduced-differentiation.

DiscussionThis work had led to the identification of GSK-3b as an im-portant regulator of differentiation in malignant glioma cells.Importantly, overexpression of GSK-3b in otherwise resistantglioma cells enables them to acquire differentiation ability.

Figure 4. Overexpression of active GSK-3b promotes differentiation

of resistant U251 glioma cells. (a, b) Morphology (a) and

immunoblot of the GFAP and PCNA levels (b) of cells treated with

10 ng/ml CT for 48 h. (c) Morphology of cells transfected with

GSK-3b S9A mutants (S9A) or empty vector (vector). (d) Cells

transfected with S9A for 48 h followed by western-blotting for

GSK-3b and GFAP expression.

Can

cerCellBiology

Li et al. 1277


Conversely, GSK-3b suppression via its inhibitors or siRNA-triggered gene silencing inhibits cholera toxin-induced differ-entiation. We also characterized GSK-3b-triggered cyclin D1nuclear export and degradation underlying this regulation ofglioma cell differentiation.

Despite that the GSK-3a and GSK-3b isoforms havehighly homologous kinase domains, their C- and N-terminalsequences are dissimilar. Furthermore, the 2 GSK-3 isoformsshow different tissue expression patterns and their expressionis differentially regulated, suggesting that they are function-ally distinct.27 Recent study has shown that only GSK-3b canmediate the function of the transcription factor NF-jB andGSK-3a was not able to compensate for the loss of GSK-3b.28 Keratinocyte migration was found to be activated byGSK-3a, while inhibited by GSK-3b.29 Furthermore, GSK-3aand GSK-3b were shown to regulate the production of Alz-heimer’s disease amyloid-b peptides in opposite ways.30 Ourdata show that both GSK-3a and GSK-3b are expressed inmalignant gliomas (Figs. 1a and 7; Supporting InformationFigure 1). However, they seemed to act differentially in theregulation of cell differentiation. GSK-3a silencing did notsignificantly affect the differentiation during C6 cell (Support-ing Information Figure 1), whereas knockdown of GSK-3breduced differentiation in glioma cells and overexpression ofcatalytically active GSK-3b showed a tendency to initiate it,

suggesting that GSK-3a and GSK-3b have distinct biologicalroles and the activation of GSK-3b isoform contributes to theoutcome of glioma cell differentiation.

Although initially viewed as a specific regulator of glycogenmetabolism, GSK-3b has recently been shown to be a crucial en-zymatic regulator of diverse number of cellular function includ-ing cell structure, survival, proliferation and apoptosis.9,31

Increasing evidence has also shown a global role for GSK-3b inhuman malignancies.32–35 Nevertheless, to our knowledge, thereare no previous reports that assess the role of GSK-3b and itseffect on malignant gliomas in the field of cellular differentiation.Microtubule-associated protein 1B (MAP1B) phosphorylation bya novel isoform of GSK-3 is induced during PC12 cell differen-tiation.36 GSK-3b is nuclear accumulated in poorly differentiatedpancreatic adenocarcinoma and inhibition of GSK-3b arrestspancreatic tumor growth.13 These findings imply a possible roleof GSK-3b in cellular differentiation of malignancies. Here, wefind that GSK-3b is overexpressed in both rat C6 and humanprimary cultured malignant glioma cells, both of which are dif-ferentiation-sensitive. Additionally, we show that GSK-3b phar-macological inhibition or silencing by RNA interference sup-presses cellular differentiation in gliomas, whereas forcedexpression of active GSK-3b promotes differentiation. Theseindicate that GSK-3b is a central point of regulation for the cel-lular differentiation of malignant glioma cells to normal

Figure 5. Cyclin D1 degradation is required for differentiation in C6 glioma cells. (a, b) Cholera toxin regulates cyclin D1 at the post-

translation level. Cyclin D1 protein (a) and mRNA (b) levels treated with cholera toxin as demonstrated by western blot and RT-PCR

analysis, respectively. (c, e) Morphology (c), immunoblot of GFAP and PCNA levels (d) and cell cycle distributions (e) in C6 cells on

stimulation with CT in the absence or presence of 10 lM MG132.Can

cerCellBiology



astrocytes. Noticeably, we also found that the inhibitor or geneticdownregulation of GSK-3b induces apoptosis and attenuates sur-vival (data not shown). These data are consistent with the recentdemonstrations.37,38 Together, our results suggested that GSK-3binhibition has dual actions in malignant glioma cells by stimulat-ing apoptosis and promoting cell differentiation.

The activity of GSK-3b is increased by the dephosphoryl-ation of Ser 9 or the phosphorylation of Tyr 216 by upstreamkinases.8,39,40 PI3K/PKB is often constitutively active inhuman gliomas41 and phosphorylates GSK-3b on Ser 9 toinactivate its kinase activity. A selective blockade of PKB acti-vation caused glioma cell morphologic alteration and raisedGFAP expression.42 In our report, cholera toxin inhibitedPI3K and downstream PKB, which in turn dephosphorylatedand activated GSK-3b. Furthermore, the PI3K inhibitormimic the cholera toxin-induced cell differentiation, suggest-ing that the positively regulated-differentiation by GSK-3bwas mediated by PI3K/PKB pathway.

Cholera toxin catalyzes ADP-ribosylation of Gs protein,thereby inducing permanent adenylate cyclase activation andresulting in an increase of intracellular cyclic adenosinemonophosphate (cAMP).43 Inhibition of PI3K/PKB pathwaysby cAMP has been reported in Swiss 3T3, HEK293, COS andRat2 cells, although the mechanism by which this occurs islargely uncharacterized.44 In C6 glioma cells, however, it waswell documented that cAMP inhibits PI3K/PKB pathways byinhibiting the activity of Rap1, which activates PI3K by directinteraction.45,46 Thus, it is reasonable to conclude that the in-hibitory role of cholera toxin for PI3K/PKB pathway in gli-oma cells is the result of suppressed Rap1 activity. However,further studies are needed to elucidate this mechanism fully.

Aberrations in the regulatory circuits that govern transitthrough the G1 phase of the cell cycle occur frequently inhuman cancer.47 Our data show that induction of differentia-tion triggered by increasing of intracellular cAMP was accom-panied by retardation of G1 transition associated with pro-found downregulated G1 control protein cyclin D1 levels.Cyclin D1 is expressed in the early G1 phase and plays a keyrole in the initiation and progression of this phase. Amplifica-tion or overexpression of cyclin D1 is one of the most com-monly observed alterations in various types of human can-cers.47,48 Inhibition of cyclin D1 expression either by antisensemethodology or antibody microinjection lengthens the durationof the G1 phase and suppresses cell proliferation.49,50 The levelsof cyclins have been shown to be regulated by both transcrip-tional and post-transcriptional processes.51–54 Targeted proteol-ysis of cyclin D1 allows for rapid removal and prevents it frominterfering with downstream cell cycle events and cell cycleprogression. However, cyclin D1 mRNA was not modified dur-ing differentiation in our study, and a mechanism that specifi-cally suppresses translation for cyclin D1 is eradicated. The evi-dence that a proteasome inhibitor MG132 restored the levels ofcyclinD1 and counteracted the differentiation strongly suggeststhat cAMP activate proteasome-mediated proteolysis, finallyleading to G1 arrest and thus differentiation.

Figure 6. GSK-3b translocation triggers cyclin D1 nuclear export

and degradation. (a–c) Subcellular location of GSK-3b and cyclin

D1 in C6 or U251 glioma cells. (a) C6 cells were treated with

10 ng/ml CT in the absence or presence of 10 lM MG132 for

24 h. (b) C6 cells were transfected with siGSK-3b ( f 0–j0) ornegative control siRNA (Neg) (a0–e0) for 36 h followed by treatment

with CT in the absence or presence of MG132. (c) U251 cells were

transfected with GSK-3b S9A mutants (S9A) or empty vector

(vector). Cells were fixed and stained with anti-GSK-3b (green) and

anti-cyclin D1 (red) antibodies. Nuclei were stained with Hoechst

33258 (blue). Optical sections were captured using a confocal

microscope. (d–g) Cyclin D1 protein (d and f) and mRNA (e and g)

levels in GSK-3b knockdown C6 cells subsequently stimulated with

CT or in S9A transfected U251 cells.

Can

cerCellBiology

Li et al. 1279


Following its discovery, cyclin D1 is localized to the nucleusand its rapid ubiquitin-dependent degradation requires phos-phorylation at Thr286 by GSK-3b in fibroblasts.26 Additionalstudies led to the recognition that GSK-3b migrates into thenucleus where it phosphorylates cyclin D1,55 resulting in ubiq-uitylation, nuclear export and degradation of the cyclin in thecytoplasm.26 Cyclin D1 nuclear export is dependent on thechromosome region maintenance 1 (CRM1) complex andrequires prior phosphorylation of cyclin D1 by GSK-3b. Inhibi-tion of CRM1 with leptomycin B, GSK3b inhibition or murinefibroblasts with T286A mutation inhibits ubiquitin-dependentcyclin D1 degradation.25,26,55 Here, we clearly show that in theprocess of differentiation in glioma cells, activated GSK-3btranslocated to the nucleus and triggers cyclin D1 nuclearexport and in turn degradation by phosphorylation in the cyto-plasm. We further show that depletion of endogenous GSK-3binterrupts cyclin D1 translocation while overexpression ofactive GSK-3b promote cyclin D1 nuclear export, suggestingthe indispensable role of GSK-3b in cyclin D1 translocation,degradation and thus differentiation.

In this article, we demonstrate that GSK-3b initiates re-sistant glioma cells susceptible to differentiation and that lossof GSK-3b activity interrupts cyclin D1 proteolysis necessaryfor the astrocytic differentiation of malignant glioma cells.Our data provide strong evidence that GSK-3b acts as a tu-mor suppressor to induce cellular differentiation, thus sup-pressing tumor development. As a pathological characteristicof differentiation-sensitive glioma cells, GSK-3b may be anovel therapeutic target and a determinant of cellular differ-entiation in malignant tumors. Our data also show that GSK-3a and GSK-3b have distinct biological roles, as the formeris not involved in the differentiation-inducing mechanism ofcholera toxin in malignant gliomas.

AcknowledgementsThe authors thank Prof. Zixu Mao for the generous gift of pcDNA3-GSK-3bS9A, Prof. Xueyun Zhong for kindly providing the U251 glioma cells, Prof.Ying Guo, Prof. Chunkui Shao and Prof. Jianyong Shao for providinghuman primary gliomas.

Figure 7. Expression and phosphorylation of GSK-3b in primary human malignant gliomas. (a) Morphology (a0, b0) and GFAP

immunofluorescence (c0, d 0) in human malignant glioma cells treated by 10 ng/ml CT for 48 h. (b) Immunoblot of p-GSK-3bser9 and GSK-3b

levels in CT-treated human malignant glioma cells. (c) Immunoblot of GFAP and PCNA levels in human malignant glioma cells pretreated

with 10 mM LiCl or 10 lM SB216732 for 2 h and then treated with CT for further 48 h. (d) GSK-3b and p-GSK-3bY216 immunocytochemistry

of stained sections from representative tissues of primary gliomas and normal brains. (e) Percentage of patients with malignant gliomas in

their GSK-3b and p-GSK-3bY216 immunostaining scored þ (positive staining for <10%), þþ (positive staining for 10–60%), or þþþ(positive staining for >60%).

Can

cerCellBiology



References

1. Trog D, Fountoulakis M, Friedlein A,Golubnitschaja O. Is current therapy ofmalignant gliomas beneficial for patients?Proteomics evidence of shifts in gliomacells expression patterns under clinicallyrelevant treatment conditions. Proteomics2006;6:2924–30.

2. Bao S, Wu Q, McLendon RE, Hao Y, ShiQ, Hjelmeland AB, Dewhirst MW, BignerDD, Rich JN. Glioma stem cells promoteradioresistance by preferential activation ofthe DNA damage response. Nature 2006;444:756–60.

3. Curran WJ, Jr, Scott CB, Horton J, NelsonJS, Weinstein AS, Fischbach AJ, ChangCH, Rotman M, Asbell SO, Krisch RE,et al. Recursive partitioning analysis ofprognostic factors in three RadiationTherapy Oncology Group malignantglioma trials. J Natl Cancer Inst 1993;85:704–10.

4. Huang ME, Ye YC, Chen SR, Chai JR, LuJX, Zhoa L, Gu LJ, Wang ZY. Use of all-trans retinoic acid in the treatment ofacute promyelocytic leukemia. Blood 1988;72:567–72.

5. Leszczyniecka M, Roberts T, Dent P, GrantS, Fisher PB. Differentiation therapy ofhuman cancer: basic science and clinicalapplications. Pharmacol Ther 2001;90:105–56.

6. Li Y, Yin W, Wang X, Zhu W, Huang Y,Yan G. Cholera toxin induces malignantglioma cell differentiation via the PKA/CREB pathway. Proc Natl Acad Sci USA2007;104:13438–43.

7. Woodgett JR. cDNA cloning andproperties of glycogen synthase kinase-3.Methods Enzymol 1991;200:564–77.

8. Cohen P, Frame S. The renaissance ofGSK3. Nat Rev Mol Cell Biol 2001;2:769–76.

9. Jope RS, Johnson GV. The glamour andgloom of glycogen synthase kinase-3.Trends Biochem Sci 2004;29:95–102.

10. Grimes CA, Jope RS. The multifacetedroles of glycogen synthase kinase 3beta incellular signaling. Prog Neurobiol 2001;65:391–426.

11. Hinoi T, Yamamoto H, Kishida M, TakadaS, Kishida S, Kikuchi A. Complexformation of adenomatous polyposis coligene product and axin facilitates glycogensynthase kinase-3 beta-dependentphosphorylation of beta-catenin and down-regulates beta-catenin. J Biol Chem 2000;275:34399–406.

12. Rask K, Nilsson A, Brannstrom M,Carlsson P, Hellberg P, Janson PO, HedinL, Sundfeldt K. Wnt-signalling pathway inovarian epithelial tumours: increasedexpression of beta-catenin and GSK3beta.Br J Cancer 2003;89:1298–304.

13. Ougolkov AV, Fernandez-Zapico ME,Savoy DN, Urrutia RA, Billadeau DD.Glycogen synthase kinase-3beta participatesin nuclear factor kappaB-mediated genetranscription and cell survival inpancreatic cancer cells. Cancer Res2005;65:2076–81.

14. Ougolkov AV, Bone ND, Fernandez-Zapico ME, Kay NE, Billadeau DD.Inhibition of glycogen synthase kinase-3activity leads to epigenetic silencing ofnuclear factor kappaB target genes andinduction of apoptosis in chroniclymphocytic leukemia B cells.Blood 2007;110:735–42.

15. Nurse P. A long twentieth century of thecell cycle and beyond. Cell 2000;100:71–8.

16. Sherr CJ. Mammalian G1 cyclins.Cell 1993;73:1059–65.

17. Sherr CJ, Roberts JM. CDK inhibitors:positive and negative regulators of G1-phase progression. Genes Dev 1999;13:1501–12.

18. Liu JJ, Chao JR, Jiang MC, Ng SY, Yen JJ,Yang-Yen HF. Ras transformation resultsin an elevated level of cyclin D1 andacceleration of G1 progression in NIH 3T3cells. Mol Cell Biol 1995;15:3654–63.

19. Roz L, Gramegna M, Ishii H, Croce CM,Sozzi G. Restoration of fragile histidinetriad (FHIT) expression induces apoptosisand suppresses tumorigenicity in lung andcervical cancer cell lines. Proc Natl AcadSci USA 2002;99:3615–20.

20. Matilla A, Koshy BT, Cummings CJ, IsobeT, Orr HT, Zoghbi HY. The cerebellarleucine-rich acidic nuclear proteininteracts with ataxin-1. Nature 1997;389:974–8.

21. Cross DA, Alessi DR, Cohen P,Andjelkovich M, Hemmings BA. Inhibitionof glycogen synthase kinase-3 by insulinmediated by protein kinase B. Nature 1995;378:785–9.

22. Coghlan MP, Culbert AA, Cross DA,Corcoran SL, Yates JW, Pearce NJ, RauschOL, Murphy GJ, Carter PS, Roxbee Cox L,Mills D, Brown MJ, et al. Selective smallmolecule inhibitors of glycogen synthasekinase-3 modulate glycogen metabolismand gene transcription. Chem Biol 2000;7:793–803.

23. Diehl JA, Zindy F, Sherr CJ. Inhibition ofcyclin D1 phosphorylation on threonine-286 prevents its rapid degradation via theubiquitin-proteasome pathway. Genes Dev1997;11:957–72.

24. Lee DH, Goldberg AL. Proteasomeinhibitors: valuable new tools for cellbiologists. Trends Cell Biol 1998;8:397–403.

25. Alt JR, Cleveland JL, Hannink M, DiehlJA. Phosphorylation-dependent regulationof cyclin D1 nuclear export and cyclin

D1-dependent cellular transformation.Genes Dev 2000;14:3102–14.

26. Guo Y, Yang K, Harwalkar J, Nye JM,Mason DR, Garrett MD, Hitomi M, StaceyDW. Phosphorylation of cyclin D1 at Thr286 during S phase leads to its proteasomaldegradation and allows efficientDNA synthesis. Oncogene 2005;24:2599–612.

27. Lau KF, Miller CC, Anderton BH, ShawPC. Expression analysis of glycogensynthase kinase-3 in human tissues.J Pept Res 1999;54:85–91.

28. Hoeflich KP, Luo J, Rubie EA, Tsao MS,Jin O, Woodgett JR. Requirement forglycogen synthase kinase-3beta in cellsurvival and NF-kappaB activation.Nature 2000;406:86–90.

29. Koivisto L, Jiang G, Hakkinen L, Chan B,Larjava H. HaCaT keratinocyte migrationis dependent on epidermal growth factorreceptor signaling and glycogen synthasekinase-3alpha. Exp Cell Res 2006;312:2791–805.

30. Phiel CJ, Wilson CA, Lee VM, Klein PS.GSK-3alpha regulates production ofAlzheimer’s disease amyloid-beta peptides.Nature 2003;423:435–9.

31. Doble BW, Woodgett JR. GSK-3: tricks ofthe trade for a multi-tasking kinase. J CellSci 2003;116:1175–86.

32. Kang T, Wei Y, Honaker Y, Yamaguchi H,Appella E, Hung MC, Piwnica-Worms H.GSK-3 beta targets Cdc25A for ubiquitin-mediated proteolysis, and GSK-3 betainactivation correlates with Cdc25Aoverproduction in human cancers.Cancer Cell 2008;13:36–47.

33. Li Y, Wang Z, Kong D, Murthy S, DouQP, Sheng S, Reddy GP, Sarkar FH.Regulation of FOXO3a/beta-catenin/GSK-3beta signaling by 3,30-diindolylmethanecontributes to inhibition of cellproliferation and induction of apoptosis inprostate cancer cells. J Biol Chem 2007;282:21542–50.

34. Wang Y, Lam JB, Lam KS, Liu J, Lam MC,Hoo RL, Wu D, Cooper GJ, Xu A.Adiponectin modulates the glycogensynthase kinase-3beta/beta-cateninsignaling pathway and attenuatesmammary tumorigenesis of MDA-MB-231cells in nude mice. Cancer Res 2006;66:11462–70.

35. Ding Q, Xia W, Liu JC, Yang JY, Lee DF,Xia J, Bartholomeusz G, Li Y, Pan Y, Li Z,Bargou RC, Qin J, et al. Erk associates withand primes GSK-3beta for its inactivationresulting in upregulation of beta-catenin.Mol Cell 2005;19:159–70.

36. Goold RG, Gordon-Weeks PR.Microtubule-associated protein 1Bphosphorylation by glycogen synthase

Can

cerCellBiology

Li et al. 1281


kinase 3beta is induced during PC12 celldifferentiation. J Cell Sci 2001;114:4273–84.

37. Miyashita K, Kawakami K, Nakada M, MaiW, Shakoori A, Fujisawa H, Hayashi Y,Hamada J, Minamoto T. Potentialtherapeutic effect of glycogen synthasekinase 3beta inhibition against humanglioblastoma. Clin Cancer Res 2009;15:887–97.

38. Kotliarova S, Pastorino S, Kovell LC,Kotliarov Y, Song H, Zhang W, Bailey R,Maric D, Zenklusen JC, Lee J, Fine HA.Glycogen synthase kinase-3 inhibitioninduces glioma cell death through c-MYC,nuclear factor-kappaB, and glucoseregulation. Cancer Res 2008;68:6643–51.

39. Kim L, Kimmel AR. GSK3, a master switchregulating cell-fate specification andtumorigenesis. Curr Opin Genet Dev2000;10:508–14.

40. Harwood AJ. Regulation of GSK-3: acellular multiprocessor. Cell 2001;105:821–4.

41. Kapoor GS, O’Rourke DM. Mitogenicsignaling cascades in glial tumors.Neurosurgery 2003;52:1425–34;discussion 34–5.

42. Kerkhoff E, Rapp UR. Cell cycle targets ofRas/Raf signalling. Oncogene 1998;17:1457–62.

43. Moss J, Vaughan M. Activation ofadenylate cyclase by choleragen. Annu RevBiochem 1979;48:581–600.

44. Kim S, Jee K, Kim D, Koh H, Chung J.Cyclic AMP inhibits Akt activity byblocking the membrane localization ofPDK1. J Biol Chem 2001;276:12864–70.

45. Wang L, Liu F, Adamo ML. Cyclic AMPinhibits extracellular signal-regulated kinaseand phosphatidylinositol 3-kinase/Aktpathways by inhibiting Rap1. J Biol Chem2001;276:37242–9.

46. Downward J. Role of phosphoinositide-3-OH kinase in Ras signaling. Adv SecondMessenger Phosphoprotein Res 1997;31:1–10.

47. Diehl JA. Cycling to cancer with cyclin D1.Cancer Biol Ther 2002;1:226–31.

48. Fu M, Wang C, Li Z, Sakamaki T, PestellRG. Minireview: cyclin D1: normal andabnormal functions. Endocrinology 2004;145:5439–47.

49. Hinz M, Krappmann D, Eichten A, HederA, Scheidereit C, Strauss M. NF-kappaBfunction in growth control: regulation ofcyclin D1 expression and G0/G1-to-S-phase transition. Mol Cell Biol 1999;19:2690–8.

50. Filmus J, Robles AI, Shi W, Wong MJ,Colombo LL, Conti CJ. Induction of cyclin

D1 overexpression by activated ras.Oncogene 1994;9:3627–33.

51. Qin C, Burghardt R, Smith R, Wormke M,Stewart J, Safe S. Peroxisome proliferator-activated receptor gamma agonists induceproteasome-dependent degradation ofcyclin D1 and estrogen receptor alpha inMCF-7 breast cancer cells. Cancer Res2003;63:958–64.

52. Buschges R, Weber RG, Actor B, Lichter P,Collins VP, Reifenberger G. Amplificationand expression of cyclin D genes(CCND1, CCND2 and CCND3)in human malignant gliomas. Brain Pathol1999;9:435–42; discussion 2–3.

53. Tetsu O, McCormick F. Beta-cateninregulates expression of cyclinD1 in colon carcinoma cells. Nature1999;398:422–6.

54. Vallance SJ, Lee HM, Roussel MF, ShurtleffSA, Kato JY, Strom DK, Sherr CJ.Monoclonal antibodies to mammalianD-type G1 cyclins. Hybridoma 1994;13:37–44.

55. Diehl JA, Cheng M, Roussel MF, Sherr CJ.Glycogen synthase kinase-3beta regulatescyclin D1 proteolysis and subcellularlocalization. Genes Dev 1998;12:3499–511.

Can

cerCellBiology



glycogen synthase kinases-3β controls differentiation of malignant glioma cells

Documents