Regulation of glioblastoma stem cells by retinoic acid: role for Notch

ORIGINAL ARTICLE

Regulation of glioblastoma stem cells by retinoic acid: role for Notch

pathway inhibition

M Ying1,2,6, S Wang1,6, Y Sang1, P Sun1,2, B Lal1,2, CR Goodwin1, H Guerrero-Cazares3,4,A Quinones-Hinojosa3,4, J Laterra1,2,3,5 and S Xia1,2

1Department of Neuro-Oncology, Hugo W. Moser Research Institute at Kennedy Krieger, Baltimore, MD, USA; 2Department ofNeurology, Johns Hopkins School of Medicine, Baltimore, MD, USA; 3Department of Oncology, Johns Hopkins School of Medicine,Baltimore, MD, USA; 4Department of Neurosurgery, Johns Hopkins School of Medicine, Baltimore, MD, USA and 5Department ofNeuroscience, Johns Hopkins School of Medicine, Baltimore, MD, USA

It is necessary to understand mechanisms by whichdifferentiating agents influence tumor-initiating cancerstem cells. Toward this end, we investigated the cellularand molecular responses of glioblastoma stem-like cells(GBM-SCs) to all-trans retinoic acid (RA). GBM-SCswere grown as non-adherent neurospheres in growth factorsupplemented serum-free medium. RA treatment rapidlyinduced morphology changes, induced growth arrest atG1/G0 to S transition, decreased cyclin D1 expressionand increased p27 expression. Immunofluorescence andwestern blot analysis indicated that RA induced theexpression of lineage-specific differentiation markers Tuj1and GFAP and reduced the expression of neural stem cellmarkers such as CD133, Msi-1, nestin and Sox-2. RAtreatment dramatically decreased neurosphere-formingcapacity, inhibited the ability of neurospheres to formcolonies in soft agar and inhibited their capacity topropagate subcutaneous and intracranial xenografts.Expression microarray analysis identified B350 genesthat were altered within 48 h of RA treatment. Affectedpathways included retinoid signaling and metabolism, cell-cycle regulation, lineage determination, cell adhesion,cell–matrix interaction and cytoskeleton remodeling.Notch signaling was the most prominent of these RA-responsive pathways. Notch pathway downregulation wasconfirmed based on the downregulation of HES and HEYfamily members. Constitutive activation of Notch signal-ing with the Notch intracellular domain rescued GBMneurospheres from the RA-induced differentiation andstem cell depletion. Our findings identify mechanisms bywhich RA targets GBM-derived stem-like tumor-initiatingcells and novel targets applicable to differentiationtherapies for glioblastoma.Oncogene (2011) 30, 3454–3467; doi:10.1038/onc.2011.58;published online 7 March 2011

Keywords: glioblastoma; cancer stem cell; retinoic acid;differentiation; Notch

Introduction

Glioblastoma multiforme (GBM) is the most aggressiveprimary brain tumor in adults with a 2-year survival rateof 28% following surgical resection, chemotherapy andradiotherapy (Stupp et al., 2009). Recurrence is nearlycertain after initial treatment and there is currently notherapy proven to prolong survival after tumor recur-rence. The dismal prognosis associated with GBM hasfostered aggressive investigations into alternative thera-peutic paradigms. Radical improvements in clinicaloutcomes will require a better understanding of themolecular and cell biological bases of glioblastomapropagation and therapeutic resistance (Kumar et al.,2008).

Small subpopulations of neoplastic cells with stem-like properties have been identified in leukemia and solidtumors including glioblastoma (Galli et al., 2004; Singhet al., 2004). These stem-like cells display the character-istic cardinal features of unlimited growth potential,self-renewal and multilineage differentiation. GBMstem-like cells (GBM-SCs) grow in vitro as non-adherentclonal multicellular spheroids (variably referred to asneurospheres or oncospheres) and efficiently initiatetumor xenografts that recapitulate the genetic andhistopathological features of the original neoplasm fromwhich they were derived (Lee et al., 2006). Thus, GBM-SCs form highly infiltrative orthotropic xenografts thatare excellent models of the human disease. Conventionalradiotherapy and chemotherapy appear to mainly targetthe most proliferative cancer cells and spare the lessproliferative neoplastic stem-like cells that appear tobe relatively resistant to current cytotoxic therapeuticsdue to upregulated anti-apoptotic proteins, multi-drug transporters and DNA repair enzymes (Baoet al., 2006; Chalmers, 2007; Johannessen et al., 2008).It is currently hypothesized that targeting GBM-SCs ortheir tumor-initiating capacity will be more effectivethan current treatment regimens. Thus, understanding

Received 1 September 2010; revised and accepted 3 February 2011;published online 7 March 2011

Correspondence: Dr S Xia, Hugo W. Moser Research Institute atKennedy Krieger/The Johns Hopkins School of Medicine, 707 N.Broadway, Room 400K, Baltimore, MD 21205, USA.E-mail: [email protected] authors contributed equally to this work.

Oncogene (2011) 30, 3454–3467& 2011 Macmillan Publishers Limited All rights reserved 0950-9232/11

www.nature.com/onc

http://dx.doi.org/10.1038/onc.2011.58

http://www.nature.com/onc

the mechanisms that regulate neoplastic stem-like cellgrowth and differentiation could lead to new andparticularly effective anti-cancer strategies (Al-Hajjet al., 2004; Massard et al., 2006; Piccirillo and Vescovi,2007).

Strategies to induce cancer cell differentiation havebeen applied to select malignancies (Sell, 2004, 2006).Retinoic acid (RA) is the most common differentiatingagent in clinical practice and all-trans RA has beensuccessfully used to treat acute promyelocytic leukemia,a stem cell malignancy (Jurcic et al., 2007). The anti-tumor activity of RA is due to the activation of an as yetpartially defined genetic program that modulates cellproliferation, differentiation and death (Dragnev et al.,2003; Mongan and Gudas, 2007). RA activates nuclearretinoid receptors, and genes most proximately modu-lated by RA contain retinoid responsive elements. Inembryonic carcinoma (EC) cells and embryonic stem(ES) cells, two peaks of gene expression change occur inresponse to RA treatment (Niles, 2004; Soprano et al.,2007). One set of affected genes codes for regulatoryproteins that function at early stages of lineagecommitment. These include Notch, Wnt, Hedgehogand TGF-b, most of which have prominent functions asdevelopment regulators. A second wave of RA-inducedgene regulation occurs at 24–48 h. Many of the secondwave gene products, such as matrix metalloproteinasesand integrins, function in extracellular matrix remodel-ing, cell adhesion and cell–cell/cell–matrix interactions.Thus, RA influences multiple signaling pathwaysinvolved in stem cell maintenance, lineage-specific celldifferentiation and organogenesis. There have beenmixed results regarding RAs efficacy in most other solidmalignancies including glioblastoma (Yung et al., 1996).Understanding the molecular RA response in these lessresponsive cancers could lead to more innovative andeffective applications of RA.

We examined the molecular and cellular responsesof GBM-SCs grown as neurospheres to RA. We showthat RA potently inhibits GBM neurosphere proli-feration and induces neurosphere cell differentiation.RA was found to reduce the pool of tumor-initiatingcells as evidenced by the inhibition of clonogenic andtumorigenic potentials. The global gene expressionprofile of GBM neurosphere cells in response to RAwas examined by microarray analysis and severalRA-responsive molecular pathways are implicated tomediate the GBM stem cell responses.

Results

RA inhibits GBM neurosphere growth and clonogenicityWe studied the effect of all-trans RA on the growth offour neurosphere cell lines established from humanglioblastomas. GBM-derived neurosphere cells becameadherent to the tissue culture substratum and generatedprocesses within 48 h of exposure to RA (10 mmol/l,Figure 1a). RA was also found to inhibit GBMneurosphere cell growth as determined by cell counting

(Figure 1b) and to induce cell growth arrest at the G1-Sphase transition as determined by flow cytometry cell-cycle analysis (Figure 1c). As one example, exposure ofHSR-GBM1B neurosphere cells to RA for 48 h in-creased the G1/G0 fraction from 58 to 74% (Po0.05)and decreased the S and G2/M fractions from 21 to14% and 20 to 11%, respectively (Figure 1c). Similarresponses were seen in all of four additional neurospherelines examined, GBM-KK190156 (KK), HSR-GBM1Aand GBM-DM140207 (DM) neurospheres (Supplemen-tary Figure 1). Thus, neurospheres that are enriched inGBM-SCs appear to be uniformly sensitive to RA.

We examined the effects of RA on principal cell-cycleregulatory proteins including cyclins, cyclin-dependentkinase and the cyclin-dependent kinase inhibitors p21and p27 in GBM neurosphere cultures. Western blotanalysis revealed a 50% decrease in cyclin D1 andtwofold increase in p21 protein levels in response to RA(Figures 1d and e, Po0.001). RA had no effect on theexpression of cyclin A, B1, E and p27 (Figures 1d and e).Thus, RA-induced G1/G0 arrest is likely secondary tocyclin D1 inhibition and p21 induction.

Neurosphere size and number were used to assess theeffects of RA on cells expressing the stem-like pheno-type, as most stem-like neoplastic cells generate largeneurospheres (4100 mm, Galli et al., 2004; Bar et al.,2007; Sun et al., 2009) in contrast to small neuro-spheres generated by progenitor cells. RA treatmentfor 5 days dramatically reduced neurosphere formation.RA completely inhibited the formation of neurospheres4100mm (70 neurospheres to none) (Figure 1f, Po0.001).RA was also found to significantly inhibit neurosphereclonogenicity in soft agar (Figure 1g, Po0.001).

We also examined the effects of RA on neurospherecell viability and caspase activation. There was nosignificant increase in Annexin V-positive cells up to 48 hof RA exposure. There was a slight increase in caspase 3cleavage after 4 days of RA treatment (SupplementaryFigure 2).

RA induces GBM neurosphere cell differentiation andinhibits stem cell marker expressionRA induced a dramatic change in neurosphere cellmorphology. In response to RA, 490% of the sus-pended neurosphere cells became adherent to the tissueculture substratum and generated processes within 48 hof exposure to RA (Figure 1a). The effects of RA onneurosphere cell expression of the progenitor cellmarker nestin, the neuronal marker Tuj1 and theastrocytic marker GFAP were examined by usingmultiple methodologies. Immunofluorescence revealedthat RA generated cells with prominent Tuj1 and GFAPexpression and decreased nestin expression, consistentwith differentiation along neuronal and astroglial line-ages (Figure 2a). The oligodendroglial marker Gal-cwas not detected in either control or RA-treated cells(data not shown). Established GBM neurosphere linesand low passage primary GBM-derived neurospheres(JHH551, passage o5) responded similarly (Figure 2b).Western blot analysis showed that RA increased the

Differentiation of glioblastoma stem cellsM Ying et al

3455

Oncogene

expression of GFAP and Tuj1 after normalized to totalprotein and actin expression (Figure 2c). Flow cyto-metry analysis demonstrated an increase in numberof GFAPþ cells (from 31 to 44% in HSR-GBM1Bcells and from 9 to 32% in HSR-GBM1A cells) and anincrease in number of Tuj1þ cells (from 1 to 7% inHSR-GBM1B cells and from 9 to 25% in HSR-GBM1A

cells) (Figures 2d and e, Po0.001). In contrast to theseneurosphere cell responses, RA had no effect ondifferentiation marker expression in U87 glioblastomacells (Figure 2c).

Quantitative RT–PCR was used to examine theeffects of RA on neurosphere cell expression of msi-1,nestin and CD-133, markers associated with the stem

Figure 1 All-trans RA induces growth arrest of GBM neurospheres. (a) Phase contrast photomicrographs of neurospheres before andafter RA treatment for 48 h. Cells change morphology and attach to culture surface and form processes in response to RA.Bar¼ 100mm. (b) Cell growth curve after RA treatment. Cells were treated with RA continuously for 96 h. Compared with control, RAinhibited neurosphere cell growth (n¼ 6, ***Po0.001). (c) Cell-cycle analysis shows that RA (10mmol/L, 48 h) induced G1-S arrest inGBM neurospheres (n¼ 4, *Po0.05). (d, e) Immunoblot analysis shows that cyclin D1 was significantly downregulated (B50%, n¼ 4,***Po0.001), and p21 was significantly upregulated by RA (twofold, n¼ 4, ***Po0.001). The expression of cyclin A, B1, E and p27was not changed by RA. (f) RA reduces neurosphere number and size. Neurospheres were treated with RA for 5 days. RA dramaticallyinhibited the number of neurospheres between 50 and 100mm by 490%; and RA completely inhibited the formation of neurospheresbigger than 100mm (n¼ 6, ***Po0.001). (g) RA inhibits GBM neurosphere cell colony formation in soft agar. RA (10mmol/l) wasadded in both culture medium and soft agar and incubated for 7–10 days. Colonies 4100mm were counted (n¼ 6, ***Po0.001).


3456

Oncogene

and progenitor cell pools. RA treatment for 48 h drama-tically reduced CD-133, msi-1 and nestin expressionby 65, 50 and 75%, respectively (Figure 2f, Po0.001).Flow cytometry with propidium iodide-conjugatedanti-CD133 antibody confirmed that the fraction ofCD133þ cells decreased from B43 to B10% inresponse to RA (Figure 2g, Po0.001).

It has been reported that RA can induce a ‘differen-tiation-like’ state in tumor cells, but the effect is oftenreversible (Tang and Gudas, 2011). To determine if theeffect of RA on GBM-SC growth and colony formationis readily reversible, we pretreated GBM neurosphereswith RA for 96 h and then cultured the cells in normalmedium without RA. RA pretreatment for 96 h drama-tically inhibited cell growth (Supplementary Figure 3A).When RA-pretreated cells were plated in soft agar in theabsence of RA, the number of large colonies (4100 mmdiametre) decreased by nearly 50% (SupplementaryFigure 3B, Po0.001). In GBM-DM and KK cultures,RA and RA pretreatment both dramatically decreasedtheir colony formation in soft agar (SupplementaryFigures 3C and D, Po0.001). Flow cytometry analysisalso indicated that the number of CD133þ cells inRA-pretreated GBM-SCs were significantly decreasedcompared with control, indicating that RA depletes the

stem cell pool in the neurosphere cultures (Supplemen-tary Figure 3E).

RA inhibits GBM neurosphere tumor propagationWe examined the effect of RA on the capacity of GBM-derived neurospheres to support and initiate xenograftgrowth. Multiple complementary in vivo experimentswere performed. First, viable GBM neurosphere cellswere implanted subcutaneously to immunodeficiency(SCID) mice. After several days, when tumors reacheda measurable size, RA administration was initiated at1.5 mg/kg/day i.p. RA significantly inhibited the growthof these preestablished neurosphere-derived xenografts(Figure 3a). By day 12 of treatment, RA-treated tumorswere B75% smaller than the controls (n¼ 5, Po0.01)(Figure 3b). Western blot analysis of tumor extractsrevealed that in response to RA, Tuj1 and GFAPexpression significantly increased. Although there issome degree of variability in GFAP/Tuj1 expressionamong different xenografts, which is not unusualfor in vivo studies, we found that in response to RA,the expression of GFAP and Tuj1 increased B5.4- and3.2-fold, respectively (Figure 3c, Po0.05, n¼ 4). Thedecrease in tumor growth and increase in differentiation

Figure 2 RA induces GBM neurosphere differentiation. (a, b) GBM neurospheres were treated with RA for 96 h and stained withanti-nestin, anti-GFAP and anti-Tuj1. Nuclei were stained with DAPI. Nestin was downregulated while GFAP and Tuj1 wereupregulated in response to RA. Bars¼ 25 mm. (c) Immunoblot analysis showed increased GFAP and Tuj1 expression in neurospherecultures after RA treatment (10mmol/l, 96 h). In contrast, in U87 human glioblastoma cells, RA (10mmol/l, 96 h) did not increase theexpression of GFAP and Tuj1. (d, e) Flow cytometry analysis shows that RA increased the number of GFAPþ cells and Tuj1þ cells inneurosphere cultures (n¼ 3, ***Po0.001). (f) RT–PCR reveals that RA significantly decreased the expression of stem cell markersincluding CD133, nestin and Msi-1 (vs 18S, n¼ 4, ***Po0.001). (g) Flow cytometric analysis of CD133 expression shows a significantdecrease in cells expressing CD133 after RA treatment (from 44 to 10%, n¼ 4, ***Po0.001).


3457

Oncogene

marker expression are consistent with the depletion ofxenografts of their tumor-propagating stem-like cells.To test this hypothesis, primary neurospheres wereestablished from control and RA-treated xenografts.Tumors from RA-treated group generated 70% fewerneurospheres than controls (Figure 3d, n¼ 4, Po0.001).

GBM neurosphere cells were implanted intracraniallyand animals were treated with RA (1.5 mg/kg/day i.p.)beginning on postimplantation day 14. After 2 months,animals were sacrificed and tumor sizes quantifiedin brain sections. Tumors were significantly smaller inRA-treated animals compared with dimethyl sulfoxide(DMSO)-treated controls (5.7 vs 18.1mm3, n¼ 6,Po0.001) (Figures 4a–c). Ki-67 staining showed thatthere are about 34.3% Ki-67þ cells in controlxenografts vs 10.2% Ki-67þ cell in RA-treated tumors(Figure 4b, Po0.001, n¼ 6), consistent with thedecreased tumor size after RA treatment. In addition,RA treatment increased median survival by B50%,from 33.1±3.5 to 47.5±3.3 days (Figure 4d, Po0.05).Immunofluorescence was performed to determine if RAinduced differentiation of orthotropic xenografts. Fig-ures 4e and f show increased staining for GFAP in RA-treated brain slices. Co-labeling with anti-humannuclear antigen localized GFAP expression to thehuman tumor cells.

Neurosphere cells were pretreated with RA for 96 hbefore intracranial implantation to determine if RAreduced the capacity of neurospheres to propagatexenografts. Equal numbers of viable control or RA-pretreated cells were implanted to brain. After 2 months,animals were sacrificed and brain sections were analyzedfor tumor formation. Nine of 10 animals (90%) injectedwith control cells developed tumors. RA-pretreatedcells formed detectable tumors in only 3 of 10 animals(30%) (Figure 4g, Po0.01). Tumors propagated fromRA-pretreated cells were also significantly smaller thancontrols (Figure 4h, Po0.05).

Gene expression response to RAThe response of GBM-derived neurospheres to RAprovides an excellent model for identifying molecularmechanisms that regulate the proliferation, differentia-tion and tumor-propagating capacity of GBM-SCs. Thetranscriptional responses to RA at 8 h (early response)and 48 h (late response) after RA treatment wereexamined using Affymetrix gene expression arrays.A comprehensive list of gene expression changes canbe found in Supplementary Table 1 and a short genelist is in Table 1a. The expression of B350 genes wassignificantly altered by RA after 48 h treatment, withB20% downregulated and 80% upregulated. Promi-nent were changes in the expression of genes codingfor proteins involved in RA signaling and metabolism,cell-cycle regulation, neuronal differentiation, cell–celland cell–matrix interaction and cytoskeleton remodeling(Table 1a). RA-induced expression of both RA receptora and b (RARa and RARb) indicating positive feedbackfor RA signaling. RA also altered the expression ofbasic helix–loop–helix DNA binding proteins, includingATOH8 (atonal homolog 8, induced B3-fold), oligo-dendrocyte lineage transcription factor 1 and 2 (OLIG1,OLIG2, inhibited B50%), and basic helix–loop–helixfamily member e40 (BHLHE40, induced B2-fold). Thisis consistent with the preferential differentiation res-ponse along neural and astroglial lineage (not oligo-dendrocytic lineage) observed by immunofluorescent

Figure 3 RA inhibits subcutaneous tumor xenograft growth.Subcutaneous tumors derived from GBM-KK neurosphere cellswere treated daily with RA (1.5mg/kg, i.p.). Control animalsreceived DMSO. (a) Tumors retrieved from animals after 12 daysof RA treatment are shown. Bar¼ 5mm. (b) Tumor was measureddaily and volumes calculated as described in Materials andmethods. By treatment day 12, RA significantly inhibited thegrowth of tumors by 75% (n¼ 5, **Po0.01). (c) Proteins fromtumor samples were subjected to immunoblot analysis. There was asignificant increase in GFAP and Tuj1 expression in RA-treatedsubcutaneous tumors. Each lane represents individual xenograft.(d) Cells were dissociated from subcutaneous tumors and grown inneurosphere culture medium. RA-treated tumors generated fewerneurospheres after reculturing (n¼ 4, ***Po0.001).


3458

Oncogene

analysis. RA was found to inhibit cyclin D1 expressionby B40%, independently validating the decrease incyclin D1 protein levels (Figures 1d and e). We selected10 genes for further validation using quantitative

RT–PCR. These include those coding for regulatorsof RA-mediated gene transcription (RARs and co-activators: Sox6, RARa, RARb and nuclear receptorco-activator 3 (NCOA3)), neuronal differentiation

Figure 4 RA inhibits intracranial tumor xenograft growth. Intracranial tumors derived from HSR-GBM1B neurosphere cells(a, b, d–g) or GBM-KK (c) were treated daily with RA (1.5mg/kg, i.p.). Control animals received DMSO. (a) H&E stained brainsections from animals after 30 days of treatment. Bar¼ 1mm. Insets in (a) show a higher magnification of tumor staining. Bar¼ 20mm.(b) Ki-67 staining of DMSO and RA-treated xenografts. There are about 34.3% Ki-67þ cells in control xenografts vs 10.2% Ki-67þcell in RA-treated tumors (Po0.001, n¼ 6). Bar¼ 20 mm. (c). Tumor cross-sectional areas were measured and tumor volumes werecalculated as described in Materials and methods. RA reduced tumor size by 75% (n¼ 6, ***Po0.001). (d) RA treatment prolongedmedian survival (33.1 vs 47.5 days, Po0.05, n¼ 6). Data obtained from GBM-KK neurosphere cells are shown. (e, f)

Immunofluorescent staining shows that GFAP expression by human glioma cells in orthotropic xenografts was increased after RAtreatment. Bar¼ 20 mm (d) and 5 mm (e). (g, h). Neurosphere cells were pretreated with RA for 96 h before intracranial implantation.Equal numbers of viable control or RA-pretreated cells were implanted to brain. Nine of 10 animals (90%) injected with control cellsdeveloped tumors. RA-pretreated cells formed detectable tumors in 3 of 10 animals (30%, n¼ 3, **Po0.01). Tumors fromRA-pretreated cells were also significantly smaller than controls (*Po0.05).


3459

Oncogene

(basic helix–loop–helix member: ATOH8) and celladhesion (tenascin, cadherin, integrin b3 (ITGB3),intercellular adhesion molecule 1 (ICAM1) and fibro-nectin leucine rich transmembrane protein 1 (FLRT1)).All showed the same pattern of upregulation ordownregulation in HSR-GBM1A and 1B. We furthervalidated the expression change of 10 genes in twoother GBM neurosphere lines as well as in low passageGBM-derived neurospheres. The results are summarizedin Table 1b.

A web-based pathway analysis (http://www.ingenuity.com) identified several molecular pathways significantlyaffected during RA-induced neurosphere differentiation(Tables 2a and b). Among them, Notch signaling wasone of the most prominently altered pathways (P-value

3.006e-5). Six of the 34 known members of this pathwaywere altered by RA. Of these, Delta-like 1 (DLL1),Notch 1, NCID1, NEXT and Notch 1 precursor weredownregulated by RA after 48 h. Only Delta-like 4(DLL4) was upregulated by RA.

Role for Notch pathway inhibition in GBM neurosphereresponse to RABased on the gene expression results, we hypothesizedthat RA promotes GBM growth inhibition and neuro-sphere differentiation, at least in part, by inhibitingNotch signaling. RA was found to downregulate neuro-sphere cell expression of the Notch pathway targetsHes2, Hey1 and Hey2 (Figure 5a). As the Notch

Table 1a Gene expression change after retinoic acid treatment for 48 h

Gene symbol Gene title Fold change P-value

Lineage determination-related genesASCL1 Achaete–scute complex homolog 1 0.26 4.57E-07OLIG1 Oligodendrocyte transcription factor 1 0.45 4.89E-06OLIG2 Oligodendrocyte transcription factor 2 0.49 1.57E-05SOX6 SRY-box 6 0.78 5.48E-06SOX9 SRY-box9 0.86 6.23E-06DCX Doublecortin 0.38 9.04E-06ATOH8 Atonal homolog 8 2.79 9.59E-08KCNE4 Potassium voltage-gated channel 2.26 1.3E-05KCNJ5 Potassium inwardly-rectifying channel 2.55 7.82E-06

Retinoids metabolism-related genesDHRS3 Dehydrogenase/reductase (SDR family) member 3 4.25 8.21E-10RBP1 Retinol binding protein 1 2.32 1.8E-07RLBP1 Retinaldehyde binding protein 1 0.44 1.01E-06

Retinoids signaling-related transcription factors/co-activatorsZNF436 Zinc finger protein 436 5.13 6.03E-09NCOA3 Nuclear receptor co-activator 3 2.37 6.73E-06NRIP1 Nuclear receptor interacting protein 1 2.56 1.13E-05RARB Retinoic acid receptor, b 2.11 1.13E-05BHLHE40 Basic helix–loop–helix family member 40 4.19E-05

Cell adhesion-related genesFLRT1 Fibronectin leucine rich transmembrane protein 1 3.18 6.73E-08ICAM1 Intercellular adhesion molecule 1 6.92 6.44E-10CDH6 Cadherin 6, type 2, K-cadherin 3.45 4.04E-08CHL1 Cell adhesion molecular, homolog of L1 2.49 1.28E-05

Table 1b Microarray analysis confirmed by RT–PCR

Gene symbol(NCBI ref sequence)

Fold changein microarray

Fold change inHSR-GBM1A

Fold change inHSR-GBM1B

Fold changein GBM-KK

Fold changein GBM-DM

Fold changein JHH-551

SOX6 (NM_19829) 0.58 0.2 ND ND ND NDATOH8 (NM_1267389) 2.79 39.3 20.63 ND ND 39.47ICAM1 (NM_1267389) 6.92 71.4 23.08 3.44 1.46 15.9FLRT1 (NM_1267389 3.18 4.26 ND ND ND 28.6ITGB3 (NM_1267389) 4.89 13.4 6.09 0.79 6.55 27.26TNC (NM_1267389) 0.47 0.18 0.63 ND ND 0.45RARA (NM_1267389) NS 2.64 18.86 ND ND NDRARB (NM_1267389) 2.11 ND 13.6 5.59 2.67 NDNCOA3 (NM_1267389) 2.38 2.07 1.06 1.23 0.99 5.33CDH6 (NM_1267389) 3.46 4.53 2.25 3.88 1.28 1.64

Abbreviations: ND, not determined; NS, not significant.Fold change is shown here.


3460

Oncogene

receptor intracellular domain (NICD1) is a proximalactivator of notch signaling, constitutive expression ofNICD1 would be expected to rescue GBM neurospheresfrom RA responses mediated by the downregulationof Notch receptor expression or activation by RA.The forced expression of NICD1 in GBM neurospherecells partially abrogated the inhibition of Hes2, Hey1and Hey2 expression by RA (Figure 5a, Po0.01). Weobserve no obvious morphology change in cells trans-fected with NICD1. However, forced NICD expressionpartially reversed the RA-induced decrease in theexpression of several stem cell markers (Figure 5b,Po0.05). Flow cytometric analysis of CD133 expressionrevealed that RA in conjunction with NICD1 over-expression increased CD133þ cells to 28.5%, comparedwith 16% in the presence of RA (Figure 5c, Po0.01).In cells treated with RA, NICD1 over-expressionincreased cell growth as evidenced by a decrease in cellsat G1/G0 phase (from 70.4 to 65.5%) and an increase atS phase (from 10.1 to 15%, Figure 5d). Compared withRA alone, NICD1 increased cell number by 40%(Figure 5e, Po0.05) and colonies 4100 mm in soft agarby 1.5-fold (Figure 5f, Po0.01). As RA has a strongerinfluence on glial differentiation, we examined the effectof NICD1 on GFAP expression. NICD1 expression

decreased upregulation of GFAP in response to RA(B40% less vs RA, Figure 5g, Po0.01). Similar resultswere found in other GBM neurosphere lines (Supple-mentary Figure 4). These data indicate that Notchpathway downregulation mediates RA effects on GBM-SCs including cell growth arrest, differentiation andstem cell pool loss.

Discussion

In this study, we examined the biological effects of RAon glioblastoma-derived stem-like cells grown as neuro-spheres and identified transcriptional responses to RAthat mediate its biological functions. We found that RAinduces cell growth arrest and differentiation of GBMneurosphere cells in vitro and in vivo. More importantly,we found that RA changes the expression of severalgroups of genes in GBM-SCs. These RA-responsivegenes and pathways can guide further studies of GBM-SC differentiation. Moreover, we also found that Notchpathway downregulation by RA mediates the GBM-SCresponse. Developing these results further could lead tonovel drug combinations for targeting GBM-SCsthrough differentiating mechanisms.

Table 2a Pathways affected by RA in GBM neurospheres (48 h vs control), analyzed with the ingenuity systems

Abbreviations: GBM, glioblastoma multiforme; RA, retinoic acid.

Table 2b Members in Notch pathway affected by retinoic acid (RA)

Pathway name P-value Network objects altered by RA Genes Expression change

8 h vs control 48 h vs control

Development Notch 3.006e-5 Six out of 34 members in DLL1 Down Downsignaling pathway this pathway DLL4 Up Up

Notch1 Down DownNCID Down DownNEXT Down Down

Notch1 precursor Down Down


3461

Oncogene

Several aspects of the cellular response of GBM-SCsto RA need to be commented. We found that RAinduced differentiation of GBM-SCs as evidenced by theexpression of neuronal markers and decrease of stem cellmarkers. The marker expression change is consistent invarious cell lines we have studied, including one lowpassage primary neurosphere isolates directly frompatient tumor. It has been reported that RA mainly

induces glial differentiation in mouse neural stem cellsby upregulating GFAP expression (Ray and Gage,2006). We also found that RA has a stronger influenceon GFAP expression vs Tuj1 expression, the marker ofneurons, suggesting that RA mainly induces glialdifferentiation in human GBM-SCs as well. As markerexpression is only one indication of the differentiationeffect of RA, we further performed functional analysis

Figure 5 NICD rescues RA-induced neurosphere differentiation. (a) HSR-GBM1B neurosphere cells were treated with lentiviruscontaining empty control (con) or NICD1 for 48 h before adding RA. RT–PCR shows that RA downregulated the Notch pathwaytargets Hes2, Hey1 and Hey2, which were rescued by NICD1 over-expression. (b) NICD1 partially restored the expression of msi-1,nestin and Sox-2, which was downregulated by RA, as revealed by RT–PCR (n¼ 9, *Po0.05; **Po0.01; ***Po0.001). (c) Flowcytometric analysis of CD133 expression indicated that NICD1 over-expression increased CD133þ cells to 28.5%, compared with16% with RA alone (n¼ 4, **Po0.01). (d) Cell-cycle analysis of GBM neurosphere cells treated with RA±NICD1 over-expression.Compared with RA alone, NICD1 over-expression decreased cells at G1/G0 phase from 70.4 to 65.5% and increased cells at S phasefrom 10.1 to 15% (n¼ 4). (e, f) In the presence of RA, NICD1 increased cell number in medium by 40% (e, n¼ 4, *Po0.05) and colonynumber in soft agar by 1.5-fold (f, n¼ 6, **Po0.01). (g) Western blot analysis shows that RA induced differentiation marker GFAPexpression, which was partially reversed by NICD1 over-expression (n¼ 3, **Po0.01).


3462

Oncogene

of neurosphere formation, clonogenic and tumor propa-gation potential. All assays showed that RA treatmentdepletes the stem cell pool in the neurosphere cultures.The neurosphere assay used to identify the stem cellpool is based on the fact that most stem-like neoplasticcells generate large neurospheres in contrast to smallneurospheres generated by progenitor cells (Galli et al.,2004; Bar et al., 2007; Sun et al., 2009). As RA alsosignificantly inhibits cell growth, we used a secondaryneurosphere/clonogenic assay, in which we pretreatedGBM-SCs with RA, and analyze their colony formationability and marker expression. There was persistentinhibition of clonogenic ability after RA withdrawalcompared with controls, indicating fewer stem-likecells in the cultures. This also suggests that the effectof RA on GBM-SCs is stable. The in vitro results arealso confirmed in the in vivo setting in which pretreatedcells were injected into mouse brain and both the tumorsize and tumor propagation rate are significantlydecrease. Therefore, the results of marker expression,neurosphere/clonogenic assay, tumor propagation assayagree in concert that RA depletes stem cell pool andinduces differentiation in GBM neurosphere cultures.

RA induces differentiation of EC cells and ES cellsin vitro (Soprano et al., 2007) and is used to treat acutepromyelocytic leukemia in clinical practice (Monganand Gudas, 2007; Nasr et al., 2009). Previous studiesexamined the gene expression changes concurrent withRA-induced differentiation of EC and ES cells. Avariety of genes were found to respond either directlyor indirectly to RA including transcription factors, RAmetabolism and transporter proteins, extracellularmatrix proteins, proto-oncogenes, growth factors andtheir receptors, cytoskeletal proteins, apoptosis-relatedproteins, cell-cycle control proteins and proteins thatmediate intracellular and extracellular signaling (Bainet al., 2000; Harris and Childs, 2002; Wei et al., 2002;Sangster-Guity et al., 2004; An et al., 2005). Our geneexpression data in GBM neurosphere cells partiallyoverlaps with the findings in EC and ES cells. Forexample, we found that RA upregulates genes related toretinol signaling and metabolism such as RAR a/b,NCOA3, retinol binding protein 1, nuclear receptorinteracting protein 1 (NRIP1), suggesting there is apositive feedback for RA signaling (Mark et al., 2006).We also found that RA downregulates some genesinvolved in lineage determination, including OLIG1/2,achaete–scute complex homolog 1 (ASCL1), SRY-box6/9 (Sox6/9) and so on. In the meantime, RA reducedneural progenitor marker expression (CD133, nestin,Sox-2, Msi-1 and doublecortin) and upregulated genesexpressed in differentiated cells, KCNE4 and KCNJ5,which encode potassium channels. We also confirmedthat RA regulates cell-cycle control proteins (cyclin Dand cyclin D1), and proteins that mediate intracellularand extracellular signaling, including ICAM1, cadherin6 (CDH6), ITGB3, FLRT1 and so on.

In our gene expression array data, we found thatseveral Notch pathway components were significantlydownregulated by RA. Notch signaling was indeedfound to be downregulated by determining the expres-

sion of its downstream target genes Hes2, Hey1, Hey2and Hey5. The fact that expression of the active form ofNotch 1 (NICD1) can partially reverse the GBM stemcell response to RA further confirmed that Notchpathway mediates RA-induced differentiation and stemcell depletion. The involvement of Notch is particularlyinteresting because of the high level of conservationof this signaling pathway in cell fate determinationand pattern formation across metazoans (Wilson andRadtke, 2006). It is well known that Notch signaling hasa role in cell fate determination during development.Notch signaling also has an important role in themaintenance of neural stem cells (Kageyama et al.,2009). Our findings also agree with the results publishedby others (Murata-Ohsawa et al., 2005; Fan et al., 2010;Wang et al., 2010). Murata-Ohsawa et al. (2005) showedthat delta-1 ligand activated Notch signaling altersRA-induced differentiation responses and reducesRA-induced apoptosis in myeloid leukemia cells. Fanet al. (2010) proposed that Notch pathway blockadeinhibits the tumor-propagating capacity of GBM-derivedneurospheres and depletes CD133þ stem-like cells.

Our gene expression profile also indicated that down-regulation of several genes by RA may be mediated byNotch pathway inhibition. For example, nestin is shownto be downregulated by RA and restored by NICD.Nestin is in fact a direct Notch/RBPJk signaling target(Shih and Holland, 2006). The extracellular matrixprotein tenascin C is significantly downregulated by RA.Studies from Sivasankaran et al. (2009) have suggesteda possible link between Notch signaling and tenascin Cin glioblastoma.

Wnt and Notch signaling pathways have importantroles in regulating stem cell proliferation and celldifferentiation. In general, Wnt activation tends topromote while Notch tends to inhibit neuronal differen-tiation (Shi et al., 2008). In our GBM neurospheres,some Wnt pathway members were upregulated by RA.The relationship between RA and the Notch and Wntpathways is variable, contextual and cell type specific.In pluripotent human EC cells (Walsh and Andrews,2003), RA treatment alters expression of Wnt receptorfrizzled family members, the Frizzled Related Proteinfamily, and receptors of the Notch pathway. Ginestieret al. (2009) found a role for retinoid signaling in theregulation of breast cancer stem cell self-renewal anddifferentiation and analyzed gene sets and pathwaysassociated with retinoid signaling. They found Wnt tobe upregulated and the Notch pathway unaffected inbreast cancer stem cells treated with RA. The interplaybetween the Wnt and Notch pathways and their rolein cancer stem cell maintenance and differentiationrequires more detailed work in the future.

RA has been used to target GBM in the preclinicaland clinical settings, but responses have been modest atbest. Ray and colleagues have shown that RA down-regulates telomerase activity and survival pathways andenhances chemotherapy-induced apoptosis in the U87GBM cells grown under traditional serum-containingconditions (Zhang et al., 2007; Das et al., 2008;Karmakar et al., 2008). Chearwae and Bright (2008)


3463

Oncogene

found that RA inhibits the growth of U87-derivedCD133þ cells grown as neurospheres and concludedthat this activity was related to the peroxisomeproliferator-activated receptor-g agonist activity ofRA. See et al. (2004) performed a retrospective analysisof patients treated with RA for recurrent glioblastoma.Phase II studies combining temozolomide, radiotherapyand cis-RA (cRA) in patients with GBMs havedemonstrated mixed results. For instance, Butowskiet al. (2005) demonstrated that cRA in combinationwith chemotherapy and radiotherapy had no effect onnewly diagnosed malignant gliomas, whereas both Yunget al. (1996) and Jaeckle et al. (2003) concluded thatcRA in combination with current standard of care wasbeneficial in recurrent gliomas. Our findings regardingthe biological and transcriptional responses to RA inGBM-SCs suggest new avenues for exploiting RA or itstargets for brain cancer therapeutics. Further studies areneeded to identify the mechanistic basis for these effectsin hopes of elucidating the best therapeutic strategy forinduced cell-differentiation therapy.

In conclusion, the GBM neurosphere response toRA provides a platform for understanding the pathwaysimportant for cancer stem cell maintenance, proli-feration, differentiation and tumorigenicity. Figure 6presents a summary of our findings. RA binds to itsnuclear receptors RARa or RARb. The subsequentupregulation of these receptors generates a positivefeedback loop for RA signaling. RARs regulate geneexpression including those related to lineage determina-tion and cell differentiation. This results in loss of stemcell markers and differentiation. RA also regulates cell-cycle control genes with cell-cycle arrest. In addition,RA alters genes coding for proteins that regulate cell–cell and cell–matrix interactions. These adhesion inter-actions have the capacity to further enhance cellulardifferentiation. All these responses to RA act together toreduce tumor propagation and inhibit tumor growth.More specifically, we identified the Notch pathway as a

downstream target of retinoids and the inhibition of Notchsignaling by RA mediates, in part, the biological effect ofRA in GBM neurospheres. A more detailed understandingof the molecular responses of neoplastic stem-like cellsto RA may ultimately contribute to novel therapeuticapproaches to stem cell-based cancer therapy.

Materials and methods

ReagentsAll reagents were purchased from Sigma Chemical Co.(St Louis, MO, USA) unless otherwise stated. All-trans RAwas prepared as stock solutions in DMSO and diluted in cellculture medium. In all the experiments, the final DMSOconcentration was p0.1% and DMSO had no demonstrableeffect on neurosphere cultures.

Cell cultureThe human glioblastoma neurosphere lines HSR-GBM1A(20913) and HSR-GBM 1B (10627) were originally derived byVescovi and colleagues and maintained in serum-free mediumsupplemented with epidermal growth factor and fibroblastgrowth factor (Vescovi et al., 1999; Galli et al., 2004; Bar et al.,2007; Sun et al., 2009). Cells were incubated in a humidifiedincubator containing 5% CO2/95% air at 37 1C, and passagedevery 4–5 days. The GBM-DM140207 and GBM-KK190156neurosphere lines were derived from glioblastomas at theUniversity of Freiburg and kindly provided by Dr JaroslawMaciaczy. The primary neurospheres JHH551 was derivedfrom a malignant glioma at Johns Hopkins using the samemethods and culture conditions as described in Galli et al.(2004). JHH551 neurospheres were used at passage p10. Allhuman materials were obtained and used in compliance withthe Johns Hopkins IRB.

Clonogenic assaysFor soft agar clonogenic assays, 1% agarose in Dulbecco’sModified Eagle’s medium was cast on the bottom of plastic6-well plates. Dissociated neurosphere cells (5� 103) weresuspended in neurosphere culture medium containing 0.5%

Figure 6 Schematic diagram of the molecular and cellular response of GBM-SCs to all-trans RA. RA binds to its nuclear receptors toactivate gene expression including those related to lineage determination and cell differentiation, cell cycle and cell–matrix interaction.These result in loss of stem cell markers, differentiation, cell-cycle arrest and morphology change. All these cellular responses to RA acttogether to reduce tumor propagation and inhibit tumor growth. In addition, Notch pathway is the downstream target of retinoids andthe inhibition of Notch signaling partially mediates the biological effect of RA in GBM neurospheres. RARE, RA response element.


3464

Oncogene

agarose and placed on top of the bottom layer. Cells wereincubated in neurosphere culture medium±RA for 7–14 daysand colonies were fixed and stained with Wright’s solution(1%). Colony formation was analyzed by computer-assistedmorphometry (MCID) by measuring the number of neuro-spheres 4100 mm in diameter in three random microscopicfields per well.

Flow cytometric analysisCell cycle was analyzed by flow cytometry on a FACSCalibur(Becton-Dickinson, Mountain View, CA, USA) (Reznik et al.,2008). Briefly, GBM neurosphere cells were collected anddissociated by vigorously pipetting. Cells were fixed with 75%ethanol at 4 1C for 30min. Cells were incubated with DNase-free RNase at 37 1C for 30min followed by propidium iodide(100 ng/ml) for 1 h at 37 1C. The percentage of cells at eachcell-cycle phase (G1/G0, S and G2/M) was analyzed usingCellQuest software (Becton-Dickinson).The number expression of CD133, Tuj1 and GFAP were

analyzed by flow cytometry. For the cell surface markerCD133, dissociated neurosphere cells (1� 106) were suspendedin 100ml assay buffer (phosphate buffered saline (PBS) pH 7.2,0.5% bovine serum albumin and 2mM EDTA) and incubatedwith 10ml of phycoerythrin-conjugated anti-CD133 antibody(Miltenyi Biotec, Auburn, CA, USA) for 10min in the dark at4 1C. For Tuj1 and GFAP expression, cells were first fixed with4% paraformaldehyde for 30min at 4 1C and permeabilizedwith PBS containing 0.5% Triton X-100 for 5min. The cellswere then incubated with primary antibodies (anti-Tuj1,1:1000, Millipore, Billerica, MA, USA; anti-GFAP, 1:5000,DAKO, Carpinteria, CA, USA) for 2 h and then incubatedwith appropriate corresponding secondary antibodies conju-gated with fluorescent dye (1:1000, The Jackson Laboratory,Bar Harbor, ME, USA) for 30min. Cells were rinsed andpelleted by centrifugation. The cell pellet was resuspended inPBS and analyzed by FACSCalibur.

Western blot analysisCells were lysed with radioimmunoprecipitation assay buffer(50mM Tris–HCl, pH 7.4, 150mM NaCl, 1% NP-40 and0.25% Na-deoxycholate) containing 1� protease and phos-phatase inhibitor cocktail (Calbiochem, San Diego, CA,USA). After sonication for 15 s, the suspensions werecentrifuged at 3000 g for 10min. Protein concentrations weredetermined using the Coomassie Protein Assay Reagent(Pierce, Rockford, IL, USA). SDS-PAGE was performed on30mg of cellular protein per lane using 4–20% gradient Tris-glycine gels according to the method of Towbin et al. (1979)with some modifications (Reznik et al., 2008). Proteins wereelectrophoretically transferred onto nitrocellulose membranes(GE Healthcare, San Francisco, CA, USA). Membranes wereincubated for 1 h in Odyssey Licor Blocking Buffer (LI-CORBiosciences, Lincoln, NE, USA) at room temperature and thenovernight with primary antibodies at 4 1C in Odyssey BlockingBuffer. After rinsing, membranes were incubated with IRDyesecondary antibodies (1:15 000–1:20 000, LI-COR Biosciences)and protein expression changes were quantified by dualwavelength immunofluorescence imaging (Odyssey InfraredImaging System, LI-COR Biosciences) scanning of themembranes as previously described (Goodwin et al., 2010).The primary antibodies were anti-GAPDH (1:7500, SantaCruz Biotechnologies, Santa Cruz, CA, USA), anti-b-actin(1:6000), anti-GFAP (1:500) and anti-Tuj1 (1:1000). Antibo-dies against cyclin, cyclin A, cyclin E, p21 and p27 werepurchased from Santa Cruz and were diluted according to themanufacturers’ recommendations.

Immunofluorescence and immunohistochemistryNeurosphere cells were grown on coverslips or collectedby cytospin onto glass slides. The cells were fixed with4% paraformaldehyde for 30min at 4 1C and permeabilizedwith PBS containing 0.5% Triton X-100 for 5min. The cellswere then incubated with primary antibodies for 2 h andthen incubated with appropriate corresponding secondaryantibodies conjugated with fluorescent dye (1:1000) for30min. The primary antibodies were anti-nestin (1:200,Sigma), GFAP (1:400) and Tuj1 (1:1000). Coverslips weremounted with Vectashield Antifade solution containing40-6-Diamidino-2-phenylindole (DAPI) (Vector Laboratories,Burlingame, CA, USA) and observed under fluorescentmicroscopy. Immunofluorescent images were taken andanalyzed using Axiovision software (Zeiss, Oberkochen,Germany).

Tumor xenograftsFemale 4- to 6-week-old Athymic Nude mice were injected s.c.in the flank with 5� 106 viable cells in 0.1ml of PBS. Whentumors reached B50mm3, the mice were randomly dividedinto two groups and treated with RA 1.5 mg/kg, i.p. daily orwith solvent DMSO as control. Animal were treated with RAduring the entire period of experiments because of rapidmetabolism of some retinoids and the duration of RAtreatment is based on previous published RA work onglioblastoma xenografts (Karmakar et al., 2008). Tumor sizeswere determined daily by measuring two dimensions (length(a) and width (b)) and volumes (V) were estimated using theformula V¼ ab2/2 (Lal et al., 2005). At the end of eachexperiment, tumors were minced and dissociated tumor cellswere cultured in neurosphere culture medium.For intracranial xenografts, SCID immunodeficient mice

received 5000 (GBM-KK) or 10 000 (HSR-GBM1B) viableneurosphere cells in 5ml of culture medium by stereotacticinjection to the right caudate/putamen. Cell viability wasdetermined by trypan blue dye exclusion. Groups of mice(n¼ 6) were sacrificed at the indicated times and the brainswere removed for histological studies. The tumor cross-sectional areas were measured based on H&E stained cryostatsections from perfusion-fixed brains using computer-assistedimage analysis as previously described (Lal et al., 2005).Tumor sizes were determined according to the followingformula: tumor volume¼ (square root of maximum cross-sectional area)3. All animal protocols used in this study wereapproved by the Johns Hopkins School of Medicine AnimalCare and Use Committee.

Microarray analysisTotal cellular RNA was extracted using the RNeasy Mini kit(Qiagen, Inc., Chatsworth, CA, USA) and purified usingRNeasy columns according to the manufacturer’s instructions.The integrity of rRNA was checked using agarose gelelectrophoresis. Two GBM neurosphere lines, HSR-GBM1Aand HSR-GBM1B were used. For each cell line, threetime points (0, 8 and 48 h after RA treatment) were analyzed.The hybridizations were performed in the Johns HopkinsMicroarray core facilities using ExonArray (Affymetrix,Santa Clara, CA, USA). We combined the data set for eachcondition from the two cell lines (n¼ 4) and analyzed thedata with Significant Analysis of Microarrays (http://www-stat.stanford.edu/Btibs/SAM/) with the false discoveryrate set at 20%. The raw data have been submitted to theNCBI website (http://www.ncbi.nlm.nih.gov/geo/, accessionnumber GSE21336). Pathways affected by RA were analyzedusing the Ingenuity Systems (http://www.ingenuity.com).


3465

Oncogene

Quantitative real-time PCRQuantitative real-time PCR (RT–PCR) was performed accord-ing to Pfaffl (2001). The primers used are listed in Supple-mentary Table 2. Total RNA (1 mg) was reverse transcribedusing the oligo(dT)12–18 primer and Superscript II (LifeTechnologies, San Diego, CA, USA) according to themanufacturer’s instructions. RT–PCR was performed withan Applied Biosystems Prism 7900 HT Sequence DetectionSystem using SYBR Green PCR Master Mix (Life Technol-ogies). The thermal cycling conditions were as follows: 95 1Cfor 5min, followed by 40 cycles of 95 1C for 10 s, 55 1C for 10 sand ended with 72 1C for 30 s. Samples were amplified intriplicate and data were analyzed using the Applied BiosystemsPrism Sequencer Detection Software Version 2.3 (LifeTechnologies). Human 18S rRNA was amplified as endogen-ous control. Relative expression of each gene was normalizedto the 18S rRNA control.

Viral expressionViral expression vector FugW and FugW containing theNotch intracellular domain 1 were obtained from Dr LinZhaoChen (Johns Hopkins University) (Yu et al., 2008). Virus waspackaged using the viral power package system (fromInvitrogen) according to the manufacturer’s instructions. Viruswas collected by centrifuging at 3000 r.p.m. for 10min.Neurosphere cells were transfected with virus containingFugW or NICD 48h before RA was added.

Statistical analysisData were analyzed using parametric statistics with one-wayanalysis of variance. Post hoc tests included the Student’s

t-Test and the Tukey multiple comparison tests as appro-priate using Prizm (GraphPad, San Diego, CA, USA). Allexperiments reported here represent at least three independentreplications. All data are represented as mean value±s.e.significance was set at Po0.05.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgements

This work is supported by the Maryland Stem Cell ResearchFund (MSCRFE) 2009-0126-00 (SX); MSCRF (MY, HG);NIH NS43987 and the James McDonnell Foundation (JL); aswell as NIH KO8 and HHMI grants (AQ). Dr CR Goodwin isa UNCF-Merck Science postdoc fellow. We thank Dr JiangQian and Ms Yanqing Yang from Wilmer Eye Institute, JohnsHopkins University, for their assistance with microarrayanalysis.Author contributions: MY and SW: collection and assembly

of data, data analysis and interpretation, final approval; YS, PS,BL and CRG: collection and assembly of data, final approval;HG, AQ and AV: provision of study materials, final approval;JL: conception and design, financial support, administrativesupport, data analysis and interpretation, manuscript writing,final approval. SX: conception and design, financial support,administrative support, collection and assembly of data, dataanalysis and interpretation, manuscript writing, final approval.

References

Al-Hajj M, Becker MW, Wicha M, Weissman I, Clarke MF. (2004).Therapeutic implications of cancer stem cells. Curr Opin Genet Dev

14: 43–47.An J, Yuan Q, Wang C, Liu L, Tang K, Tian HY et al. (2005).

Differential display of proteins involved in the neural differentiationof mouse embryonic carcinoma P19 cells by comparative proteomicanalysis. Proteomics 5: 1656–1668.

Bain G, Mansergh FC, Wride MA, Hance JE, Isogawa A, RancourtSL et al. (2000). ES cell neural differentiation reveals a substantialnumber of novel ESTs. Funct Integr Genomics 1: 127–139.

Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB et al.(2006). Glioma stem cells promote radioresistance by preferentialactivation of the DNA damage response. Nature 444: 756–760.

Bar EE, Chaudhry A, Lin A, Fan X, Schreck K, Matsui W et al.(2007). Cyclopamine-mediated hedgehog pathway inhibitiondepletes stem-like cancer cells in glioblastoma. Stem Cells 25:2524–2533.

Butowski N, Prados MD, Lamborn KR, Larson DA, Sneed PK, WaraWM et al. (2005). A phase II study of concurrent temozolomide andcis-retinoic acid with radiation for adult patients with newlydiagnosed supratentorial glioblastoma. Int J Radiat Oncol Biol

Phys 61: 1454–1459.Chalmers AJ. (2007). Radioresistant glioma stem cells—therapeutic

obstacle or promising target? DNA Repair (Amst) 6: 1391–1394.Chearwae W, Bright JJ. (2008). PPARgamma agonists inhibit growth

and expansion of CD133+ brain tumour stem cells. Br J Cancer 99:2044–2053.

Das A, Banik NL, Ray SK. (2008). Retinoids induced astrocyticdifferentiation with down regulation of telomerase activity andenhanced sensitivity to taxol for apoptosis in human glioblastomaT98G and U87MG cells. J Neurooncol 87: 9–22.

Dragnev KH, Petty WJ, Dmitrovsky E. (2003). Retinoid targets incancer therapy and chemoprevention. Cancer Biol Ther 2: S150–S156.

Fan X, Khaki L, Zhu TS, Soules ME, Talsma CE, Gul N et al. (2010).NOTCH pathway blockade depletes CD133-positive glioblastomacells and inhibits growth of tumor neurospheres and xenografts.Stem Cells 28: 5–16.

Galli R, Binda E, Orfanelli U, Cipelletti B, Gritti A, De Vitis S et al.(2004). Isolation and characterization of tumorigenic, stem-like neuralprecursors from human glioblastoma. Cancer Res 64: 7011–7021.

Ginestier C, Wicinski J, Cervera N, Monville F, Finetti P, Bertucci Fet al. (2009). Retinoid signaling regulates breast cancer stem celldifferentiation. Cell Cycle 8: 3297–3302.

Goodwin CR, Lal B, Zhou X, Ho S, Xia S, Taeger A et al. (2010).Cyr61 mediates hepatocyte growth factor-dependent tumor cellgrowth, migration, and Akt activation. Cancer Res 70: 2932–2941.

Harris TM, Childs G. (2002). Global gene expression patterns duringdifferentiation of F9 embryonal carcinoma cells into parietalendoderm. Funct Integr Genomics 2: 105–119.

Jaeckle KA, Hess KR, Yung WK, Greenberg H, Fine H, Schiff D et al.(2003). Phase II evaluation of temozolomide and 13-cis-retinoic acidfor the treatment of recurrent and progressive malignant glioma: aNorth American Brain Tumor Consortium study. J Clin Oncol 21:2305–2311.

Johannessen TC, Bjerkvig R, Tysnes BB. (2008). DNA repair andcancer stem-like cells—potential partners in glioma drug resistance?Cancer Treat Rev 34: 558–567.

Jurcic JG, Soignet SL, Maslak AP. (2007). Diagnosis and treatment ofacute promyelocytic leukemia. Curr Oncol Rep 9: 337–344.

Kageyama R, Ohtsuka T, Shimojo H, Imayoshi I. (2009). Dynamicregulation of Notch signaling in neural progenitor cells. Curr Opin

Cell Biol 21: 733–740.


3466

Oncogene

Karmakar S, Banik NL, Ray SK. (2008). Combination of all-transretinoic acid and paclitaxel-induced differentiation and apoptosis inhuman glioblastoma U87MG xenografts in nude mice. Cancer 112:596–607.

Kumar HR, Zhong X, Sandoval JA, Hickey RJ, Malkas LH. (2008).Applications of emerging molecular technologies in glioblastomamultiforme. Expert Rev Neurother 8: 1497–1506.

Lal B, Xia S, Abounader R, Laterra J. (2005). Targeting the c-Metpathway potentiates glioblastoma responses to gamma-radiation.Clin Cancer Res 11: 4479–4486.

Lee J, Kotliarova S, Kotliarov Y, Li A, Su Q, Donin NM et al. (2006).Tumor stem cells derived from glioblastomas cultured in bFGFand EGF more closely mirror the phenotype and genotype ofprimary tumors than do serum-cultured cell lines. Cancer Cell 9:391–403.

Mark M, Ghyselinck NB, Chambon P. (2006). Function of retinoidnuclear receptors: lessons from genetic and pharmacologicaldissections of the retinoic acid signaling pathway during mouseembryogenesis. Annu Rev Pharmacol Toxicol 46: 451–480.

Massard C, Deutsch E, Soria JC. (2006). Tumour stem cell-targetedtreatment: elimination or differentiation. Ann Oncol 17: 1620–1624.

Mongan NP, Gudas LJ. (2007). Diverse actions of retinoid receptors incancer prevention and treatment. Differentiation 75: 853–870.

Murata-Ohsawa M, Tohda S, Kogoshi H, Sakano S, Nara N. (2005).The Notch ligand, Delta-1, alters retinoic acid (RA)-inducedneutrophilic differentiation into monocytic and reduces RA-inducedapoptosis in NB4 cells. Leuk Res 29: 197–203.

Nasr R, Lallemand-Breitenbach V, Zhu J, Guillemin MC, de The H.(2009). Therapy-induced PML/RARA proteolysis and acutepromyelocytic leukemia cure. Clin Cancer Res 15: 6321–6326.

Niles RM. (2004). Signaling pathways in retinoid chemopreventionand treatment of cancer. Mutat Res 555: 81–96.

Pfaffl MW. (2001). A new mathematical model for relative quantifica-tion in real-time RT-PCR. Nucleic Acids Res 29: e45.

Piccirillo SG, Vescovi AL. (2007). Brain tumour stem cells: possibilitiesof new therapeutic strategies. Expert Opin Biol Ther 7: 1129–1135.

Ray J, Gage FH. (2006). Differential properties of adult rat and mousebrain-derived neural stem/progenitor cells. Mol Cell Neurosci 31:560–573.

Reznik TE, Sang Y, Ma Y, Abounader R, Rosen EM, Xia S et al.(2008). Transcription-dependent epidermal growth factorreceptor activation by hepatocyte growth factor. Mol Cancer Res

6: 139–150.Sangster-Guity N, Yu LM, McCormick P. (2004). Molecular profiling

of embryonal carcinoma cells following retinoic acid or histonedeacetylase inhibitor treatment. Cancer Biol Ther 3: 1109–1120.

See SJ, Levin VA, Yung WK, Hess KR, Groves MD. (2004). 13-cis-retinoic acid in the treatment of recurrent glioblastoma multiforme.Neuro Oncol 6: 253–258.

Sell S. (2004). Stem cell origin of cancer and differentiation therapy.Crit Rev Oncol Hematol 51: 1–28.

Sell S. (2006). Cancer stem cells and differentiation therapy. Tumour

Biol 27: 59–70.Shi Y, Sun G, Zhao C, Stewart R. (2008). Neural stem cell self-

renewal. Crit Rev Oncol Hematol 65: 43–53.

Shih AH, Holland EC. (2006). Notch signaling enhances nestinexpression in gliomas. Neoplasia 8: 1072–1082.

Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T et al.(2004). Identification of human brain tumour initiating cells. Nature

432: 396–401.Sivasankaran B, Degen M, Ghaffari A, Hegi ME, Hamou MF,

Ionescu MC et al. (2009). Tenascin-C is a novel RBPJkappa-induced target gene for Notch signaling in gliomas. Cancer Res 69:458–465.

Soprano DR, Teets BW, Soprano KJ. (2007). Role of retinoic acid inthe differentiation of embryonal carcinoma and embryonic stemcells. Vitam Horm 75: 69–95.

Stupp R, Hegi ME, Mason WP, van den Bent MJ, Taphoorn MJ,Janzer RC et al. (2009). Effects of radiotherapy with concomitantand adjuvant temozolomide versus radiotherapy alone on survivalin glioblastoma in a randomised phase III study: 5-year analysis ofthe EORTC-NCIC trial. Lancet Oncol 10: 459–466.

Sun P, Xia S, Lal B, Eberhart CG, Quinones-Hinojosa A, Maciaczyk Jet al. (2009). DNER, an epigenetically modulated gene, regulatesglioblastoma-derived neurosphere cell differentiation and tumorpropagation. Stem Cells 27: 1473–1486.

Tang XH, Gudas LJ. (2011). Retinoids, retinoic acid receptors, andcancer. Annu Rev Pathol 6: 345–364.

Towbin H, Staehelin T, Gordon J. (1979). Electrophoretic transfer ofproteins from polyacrylamide gels to nitrocellulose sheets: procedureand some applications. Proc Natl Acad Sci USA 76: 4350–4354.

Vescovi AL, Parati EA, Gritti A, Poulin P, Ferrario M, Wanke E et al.(1999). Isolation and cloning of multipotential stem cells fromthe embryonic human CNS and establishment of transplantablehuman neural stem cell lines by epigenetic stimulation. Exp Neurol

156: 71–83.Walsh J, Andrews PW. (2003). Expression of Wnt and Notch

pathway genes in a pluripotent human embryonal carcinomacell line and embryonic stem cell. APMIS 111: 197–210; discussion210–191.

Wang J, Wakeman TP, Lathia JD, Hjelmeland AB, Wang XF,White RR et al. (2010). Notch promotes radioresistance of gliomastem cells. Stem Cells 28: 17–28.

Wei Y, Harris T, Childs G. (2002). Global gene expression patternsduring neural differentiation of P19 embryonic carcinoma cells.Differentiation 70: 204–219.

Wilson A, Radtke F. (2006). Multiple functions of Notch signaling inself-renewing organs and cancer. FEBS Lett 580: 2860–2868.

Yu X, Zou J, Ye Z, Hammond H, Chen G, Tokunaga A et al. (2008).Notch signaling activation in human embryonic stem cells isrequired for embryonic, but not trophoblastic, lineage commitment.Cell Stem Cell 2: 461–471.

Yung WK, Kyritsis AP, Gleason MJ, Levin VA. (1996). Treatmentof recurrent malignant gliomas with high-dose 13-cis-retinoic acid.Clin Cancer Res 2: 1931–1935.

Zhang R, Banik NL, Ray SK. (2007). Combination of all-transretinoic acid and interferon-gamma suppressed PI3K/Akt survivalpathway in glioblastoma T98G cells whereas NF-kappaB survivalsignaling in glioblastoma U87MG cells for induction of apoptosis.Neurochem Res 32: 2194–2202.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)


3467

Oncogene

Documents

Regulation of glioblastoma stem cells by retinoic acid: role for Notch