The transient receptor potential vanilloid-2 cation channel impairs glioblastoma stem-like cell...

The transient receptor potential vanilloid-2 cation channelimpairs glioblastoma stem-like cell proliferation andpromotes differentiation

Maria Beatrice Morelli1*, Massimo Nabissi1*, Consuelo Amantini1, Valerio Farfariello1,2, Lucia Ricci-Vitiani3,

Simona di Martino3, Roberto Pallini4, Luigi Maria Larocca5, Sara Caprodossi1, Matteo Santoni1,

Ruggero De Maria3 and Giorgio Santoni1

1 Section of Experimental Medicine, School of Pharmacy, University of Camerino, Camerino, Italy2 Department of Molecular Medicine, Sapienza University, Rome, Italy3 Department of Hematology, Oncology and Molecular Medicine, Istituto Superiore di Sanit�a, Rome, Italy4 Institute of Neurosurgery, School of Medicine, Catholic University, Rome, Italy5 Institute of Pathology, School of Medicine, Catholic University, Rome, Italy

Malignant transformation of cells resulting from enhanced proliferation and aberrant differentiation is often accompanied by

changes in transient receptor potential vanilloid (TRPV) channels expression. In gliomas, recent evidence indicates that TRPV

type 2 (TRPV2) negatively controls glioma cell survival and proliferation. In addition, cannabinoids, the ligands of both

cannabinoid and TRPV2 receptors, promote glioblastoma stem-like cells (GSCs) differentiation and inhibit gliomagenesis.

Herein, we provide evidence on the expression of TRPV2 in human GSCs and that GSCs differentiation reduces nestin and

progressively increases both the glial fibrillary acidic protein (GFAP) and TRPV2 expression. Therefore, we evaluated the role

of TRPV2 cation channel in GSC lines differentiation. Treatment of GSC lines with the TRPV antagonist Ruthenium Red, with

ethylene glycol-bis(2-aminoethylether)-N,N,N0,N0-tetraacetic acid or knockdown of TRPV2 gene during differentiation, decreases

GFAP and class III beta-tubulin (bIII-tubulin) expression; conversely, phorbol-12-myristate-13-acetate stimulates GSCs

proliferation, reduces TRPV2 expression and partially reverts astroglial differentiation. In addition, forced TRPV2 expression in

GSC lines by stable TRPV2 transfection increases GFAP and bIII-tubulin expression and parallelly reduces proliferation. Finally,

TRPV2 overexpression inhibits GSCs proliferation in a xenograft mouse model, as shown by reduced tumor diameter and

mitotic index, and promotes the differentiation of GSCs toward a more mature glial phenotype. Overall, our results

demonstrate that TRPV2 promotes in vitro and in vivo GSCs differentiation and inhibits their proliferation. Better

understanding of the molecular mechanisms that regulate the balance between proliferation and differentiation of GSCs would

lead to more specific and efficacious pharmacological approaches.

Malignant gliomas remain the most deadly human braintumors with a poor prognosis despite years of research inantitumoral therapeutic strategies. A hallmark characteristic ofgliomas is their molecular and cellular heterogeneity that isconsidered to be one of the reasons for their high malignancy

and recurrence.1,2 Neoplastic transformation of differentiatedglial cells was for many years the most accepted hypothesis toexplain the origin of gliomas1,2; however, recent findings sup-port the existence of a stem cell-derived origin for differenttypes of cancers including brain tumors.2,3 In particular,

Key words: glioblastoma, glioblastoma stem-like cells, TRP, TRPV2, differentiation

Abbreviations: bIII-tubulin: class III beta-tubulin; BrdU: 5-bromo-2-deoxyuridine; CBD: cannabidiol; dxRT-PCR: duplex quantitative real-

time polymerase chain reaction; EGTA: ethylene glycol-bis(2-aminoethylether)-N,N,N0,N0-tetraacetic acid; GAPDH: glyceraldehyde-3-

phosphate dehydrogenase; GFAP: glial fibrillary acidic protein; GSCs: glioma stem-like cells; PMA: phorbol-12-myristate-13-acetate; RR:

Ruthenium Red; TRPV2: transient receptor potential vanilloid type 2; TRPV2-GSCs: transient receptor potential vanilloid type 2-transfected

glioma stem-like cells; Vector-GSCs: vector-transfected glioma stem-like cells

Additional Supporting Information may be found in the online version of this article

*M.B.M. and M.N. contributed equally to this work

Grant sponsor: Associazione Italiana Ricerca sul Cancro (AIRC) Regional Grant 2009–2011; Grant number: 6353; Grant sponsor:

University of Camerino Grant 2009

DOI: 10.1002/ijc.27588

History: Received 14 Sep 2011; Accepted 29 Mar 2012; Online 11 Apr 2012

Correspondence to: Giorgio Santoni, Section of Experimental Medicine, School of Pharmacy, University of Camerino, Via Madonna delle

Carceri 9, 62032 Camerino, Italy, Tel.: þ39-0737-403312/403319, Fax: þ39-0737-403325, E-mail: [email protected]

Can

cerCellBiology

Int. J. Cancer: 131, E1067–E1077 (2012) VC 2012 UICC

International Journal of Cancer

IJC

glioblastoma stem-like cells (GSCs) have been isolated fromboth human tumor tissues 4–8 and several glioma cell lines.6,9–11 Because GSCs retain the ability to respond to physiologicalsignals that induce neural stem cell differentiation into neuro-nal, astroglial and oligodendroglial cells, mechanism-induceddifferentiation could be a promising approach to eradicatethis tumor-driving cell population.12 Members of transient re-ceptor potential vanilloid (TRPV) channel family control cel-lular homeostasis by regulating the balance between cell pro-liferation and differentiation.13 Recently, an additionalpathophysiological role for TRP channels in cancer growthand progression has been demonstrated.13,14 Malignant trans-formation of cells resulting from enhanced proliferation andaberrant differentiation is often accompanied by changes inTRPV channel expression and consequently by abnormal pro-gression of the cellular responses.14–17 Among TRPV chan-nels, TRPV2, also called vanilloid receptor like-1, contains sixtransmembrane domains, a putative pore-loop region, a cyto-plasmic amino terminus with three ankyrin-repeat domainsand a cytoplasmic carboxy terminus.18 TRPV2 as cation chan-nel shows Ca2þ permeability, sensitivity to noxious heat withan activation threshold >52�C18 and changes in osmolarityand membrane stretch.19 In addition, TRPV2 can be triggeredby agonists, such as 2-aminoethoxydiphenyl borate,20 canna-bidiol (CBD) and D(9)-tetrahydrocannabinol.21 TRPV2 wasfound in neuronal and non-neuronal tissues,15,18,22,23 whereits activity can be regulated by signaling events triggered bygrowth factors including the insulin growth factor-I, the plate-let-derived growth factor and the neuropeptide head activatorthat can induce the translocation of TRPV2 channel from theintracellular compartments to the plasma membrane throughphosphatidylinositol-3 kinase (PI3-kinase)-dependent and -in-dependent pathways.24,25 Recent evidence indicates a specificrole of TRPV2 in cancer. 17,26–28 In this regard, we haverecently reported that TRPV2 negatively controls glioma cellsurvival and proliferation as well as resistance to Fas-inducedapoptosis in an extracellular signal-regulated kinases-depend-ent manner.17 Moreover, a role for TRPV2 in lysophospholi-pid-mediated migration, extracellular matrix invasion andprogression to androgen resistance of prostate cancer cells hasbeen reported.26,27 Finally, a clinicopathological significance ofTRPV2 expression as prognostic marker in human hepatocel-lular carcinoma has been shown.28

Herein, we provide the first evidence on the expression ofTRPV2 in GSC lines. In addition, by RNA interference andstable transfection to knockdown and enforce TRPV2 geneexpression, respectively, we demonstrate that TRPV2 channelinhibits proliferation and promotes astroglial GSC linesdifferentiation.

Material and MethodsCell cultures

GSC lines (#1, #30 and #83) previously characterized29 wereisolated from surgical samples of three adult patients, diag-nosed according to the WHO classification of tumors of the

central nervous system,30,31 who had undergone craniotomyat the Institute of Neurosurgery, Catholic University Schoolof Medicine, Rome.32 Informed consent was obtained beforesurgery according to the protocols approved at the CatholicUniversity. GSC lines were cultured in a serum-free mediumsupplemented with 20 ng/ml of epidermal growth factor and20 ng/ml of basic fibroblast growth factor. For differentiation,GSC lines were grown in medium supplemented with 5% fe-tal bovine serum as described.6,29

Reagents and antibodies

Phorbol 12-myristate 13-acetate (PMA), Ruthenium Red(RR) and ethylene glycol-bis(2-aminoethylether)-N,N,N0,N0-tetraacetic acid (EGTA) were obtained from Tocris Bio-science (Bristol, UK). 40-6-Diamino-2-phenylindole dihydro-chloride (DAPI), 5-bromo-2-deoxyuridine (BrdU) and hema-toxylin and eosin (H&E) were obtained from Sigma Aldrich(St. Louis, MO). The following mouse monoclonal antibodies(mAbs) were used: anti-glyceraldehyde-3-phosphate dehydro-genase (GAPDH)-peroxidase (Sigma Aldrich), anti-class IIIbeta-tubulin (bIII-tubulin), normal mouse IgG1 (Cell Signal-ing Technology, Danvers, MA), anti-nestin (Chemicon, Bill-erica, MA) and fluorescein isothiocyanate (FITC)-conjugatedanti-BrdU (Becton Dickinson Biosciences, San Jose, CA). Thefollowing polyclonal antibodies were used: goat anti-TRPV2(Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti-glialfibrillary acidic protein (GFAP), normal rabbit IgG (Dako,Milan, Italy) and rabbit anti-nestin (Sigma Aldrich). The fol-lowing secondary antibodies were used: phycoerythrin (PE)-conjugated goat anti-mouse IgG (GAM), PE-conjugated goatanti-rabbit IgG (GARB) and FITC-conjugated GARB (BectonDickinson Biosciences); biotin-conjugated rabbit anti-goatIgG (RAG), biotin-conjugated GARB and tricolor-conjugatedstreptavidin (Invitrogen, San Diego, CA); FITC-conjugateddonkey anti-goat (Santa Cruz Biotechnology); horseradishperoxidase (HRP)-conjugated donkey anti-goat IgG (SantaCruz Biotechnology); HRP-conjugated sheep anti-mouse IgGand HRP-conjugated donkey anti-rabbit IgG (GE Healthcare,Munich, Germany) and Alexa Fluor 488-conjugated GAMand RAG (Invitrogen) and Texas Red-conjugated GARB(Abcam, Cambridge, UK). Blocking peptide (BP) againstTRPV2 and normal goat IgG was obtained from Santa CruzBiotechnology.

Soft agar assay

Sphere-forming/self-renewal capability was tested by soft agarassay in a six-well plate. Briefly, a 500 ll underlayer of 0.6%soft agarose (sea plaque agarose diluted in medium plusgrowth factors) was prepared in each 35-mm well. A total of3 � 103 cells per well were added above the layer as a secondlayer composed of 2/3 soft agarose 0.4% and 1/3 cells plusmedium in a total volume of 500 ll. The plates were incu-bated for 10 min at 4�C and for 15 days at 37�C. Then, thedishes were stained with 70 ll of Crystal Violet (0.1%) andcolonies were counted.

Can

cerCellBiology

E1068 TRPV2 in glioblastoma stem-like cell differentiation


Confocal laser scanning microscopy analysis

For immunophenotyping, undifferentiated (U-GSC) and dif-ferentiated (D-GSC) lines were treated and fixed as previ-ously published.29 Then, GSC lines were stained with mouseanti-bIII-tubulin (1:50) mAb, rabbit anti-GFAP (1:100) andmouse anti-nestin (1:100) Abs followed by Alexa Fluor 488-GAM Ab and Texas Red-GARB. Nuclei were counterstainedwith DAPI (Vector Laboratories, Burlingame, CA). The slideswere analyzed with an Axioplan fluorescence microscope(Carl Zeiss MicroImaging GmbH, Munchen, Germany) withthe appropriate filter sets (Abbott Molecular, Abbott Park,IL). Images were captured using a high-resolution black andwhite CCD microscope camera AxioCam MRm REV 2 (CarlZeiss). The resulting images were reconstructed with green,orange and blue (DAPI) pseudocolor using AxioVision 4multichannel fluorescence basic workstation (Carl Zeiss)according to the manufacturer’s instructions.

Gene expression analysis

Total RNA from cell samples was extracted with the RNeasyMini Kit (Qiagen, Milan, Italy), and cDNA was synthesizedusing the High-Capacity cDNA Archive Kit (Applied Biosys-tems, Foster City, PA) according to the manufacturer’s instruc-tions. Duplex quantitative real-time polymerase chain reaction(dxRT-PCR) for TRPV2 was performed using the iQ5 Multi-color Real-Time PCR Detection System (Bio-Rad, Hercules,CA) as previously described.17 All samples were assayed intriplicates in the same plate. Measurement of b-actin levels wasused to normalize mRNA contents, and TRPV2 levels were cal-culated by the 2�DDCt method and expressed as relative foldcompared to the corresponding control.

GSC lentiviral infection

Human TRPV2 cDNA was subcloned under the citomegalo-virus promoter of a lentiviral vector that carries the greenfluorescence protein (GFP) reporter gene under the phospho-glycerate kinase promoter. Recombinant lentivirus was pro-duced as previously described.33 Empty lentivirus (Vector)was used as control. GSC lines were plated in the presence ofviral supernatant, and 4 lg/ml of polybrene was added to theviral supernatant to improve the infection efficiency.33 Afterthe infection cycles, GSC lines were washed twice andreplated in fresh medium. Transfected U-GSC (TRPV2-GSCand Vector-GSC) lines were sorted based on their fluores-cence intensity (FACS Aria, Becton Dickinson Biosciences).

TRPV2 gene silencing

siGENOME SMARTpools for TRPV2 (siTRPV2), siCON-TROL nontargeting siRNA (siGLO) used as negative controland siGLO RISC-Free siRNA (siFLUO, a siRNA nontargetCy3-labeled control siRNA cotransfected with functionalsiRNA) were purchased from Thermo Scientific-Dharmacon(Lafayette, CO). For gene silencing experiments, GSC lineswere plated at the density of 3 � 104/cm2 and 160 pmol of

siTRPV2 or siGLO or siFLUO was added to the wells, follow-ing the METAFECTENE SI PRO transfection protocol (Bion-tex Laboratories, San Diego, CA).

Western blot

Total proteins from GSC lines were obtained as previouslydescribed.17 GSC lines plasma membrane and cytosol frac-tions were isolated using the Subcellular Protein Fractiona-tion kit (Thermo Scientific, Rockford, IL), according to themanufacturer’s directions. Proteins were separated on 5–9%SDS polyacrylamide gel, transferred onto Hybond-C extramembranes (GE Healthcare) and blotted with the specificAbs as previously published.17 The Abs used were anti-GFAP(1:300), anti-nestin (1:2,000), anti-bIII-tubulin (1:1,000), anti-GAPDH (1:10,000) and anti-TRPV2 (1:200) followed by rela-tively HRP-conjugated Abs. Anti-TRPV2 Ab was preincu-bated in 5% milk, followed by dilution with blocking buffer(5 lg/ml). The preadsorption of anti-TRPV2 Ab with its BPwas performed to validate Ab specificity. Normal goat IgG(1:200) was used as isotype control Ab to evaluate nonspe-cific binding of TRPV2 primary Ab. Furthermore, the lysateswere incubated with secondary Ab. No reactivity wasobserved blotting with the secondary Ab alone. Immuno-staining was revealed by enhanced ECL Western Blottinganalysis system (GE Healthcare). Densitometric analysis wasperformed by ChemiDoc using the Quantity One software(Bio-Rad).

Fluorescence-activated cell sorting analysis

GSC lines were fixed with 4% paraformaldehyde (PFA) andlabeled with anti-GFAP (1:100) and anti-bIII-tubulin (1:50)Abs, followed by FITC-conjugated GARB and PE-conjugatedGAM polyclonal Abs. TRPV2 was evaluated by tricolor im-munofluorescence labeling cells with anti-TRPV2 (1:25) Ab,followed by biotin-conjugated RAG polyclonal Ab and tri-color-conjugated streptavidin. For proliferation analysis,BrdU incorporation was determined by seeding 3 � 104/cm2

cells per well and labeled by adding 20 lM BrdU. Cells werefixed and stained with anti-BrdU FITC-conjugated Ab as pre-viously described.17 For tricolor cytofluorimetric analysis,electronic compensation was used between green and orangeand between orange and red fluorescence to remove spectraloverlap. Samples were analyzed by a FACScan cytofluorime-ter using the CellQuest software (Becton Dickinson). Thespecificity of staining was confirmed using isotype-specificAbs as negative control. Fluorescence intensity was expressedin arbitrary units on logarithmic scale.

WST assay

For cell proliferation assays, 1 � 104 cells per square centi-meter were plated in quadruplicate and grown for 3 days. Atthe indicated time, cells were incubated for additional 3 hr inthe presence of 10 ll per well WST-1 (2-(2-methoxy-4-nitro-phenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium,monosodium salt) labeling mixture (Roche Molecular

Can

cerCellBiology

Morelli et al. E1069


Biochemicals, Basel, Switzerland). The absorbance of the sam-ples against a background control (medium alone) was meas-ured at 450 nm using an ELISA reader microliter plate (BioTekInstruments, Winooski, VT).

Calcium mobilization assay

For calcium influx analysis, GSC lines were resuspended inmedium supplemented with 1% FCS, 7 lmol/l FLUO 3-AM

(Invitrogen) and 1 lg/ml Pluronic F-127 (Invitrogen) andincubated in the dark for 30 min at 37�C and 5% CO2.FLUO 3-AM fluorescence was measured by FACS. [Ca2þ]iwas determined before and after the addition of various testcompounds. The following equation was used to determine[Ca2þ] free: [Ca2þ] free ¼ Kd[F � Fmin]/[Fmax � F], whereFmin and Fmax are the fluorescence intensities of FLUO 3without or with maximal [Ca2þ], and F is the fluorescence

Figure 1. TRPV2 expression increases during GSC differentiation. (a and b) The relative TRPV2 mRNA expression in U-GSC #1, #30 and #83

and in D-GSC lines (day 14 of differentiation) was evaluated by dxRT-PCR. TRPV2 mRNA levels were normalized for b-actin expression.

(b) TRPV2 mRNA levels were expressed as relative fold with respect to the corresponding GSC lines used as calibrator. *p < 0.01 (student’s

t-test). (c) Representative analysis of TRPV2 protein expression in U- and D-GSC lines evaluated by Western blot analysis. Proteins from

plasma membrane (M) and cytosolic (C) fractions were immunoblotted with anti-TRPV2 Ab. Preadsorption of anti-TRPV2 to its BP was

performed to validate antibody selectivity. The incubation with normal goat IgG was performed to estimate the nonspecific binding of TRPV2

Ab. GAPDH protein levels were used as loading control. (d) M and C fractions from U-GSC and D-GSC lines at different times of differentiation

were immunoblotted with anti-TRPV2 Ab, and protein levels were expressed as relative fold with respect to U-GSCs. Total lysates from U-GSC

and D-GSC lines at different times of differentiation were also immunoblotted with anti-GFAP, anti-bIII-tubulin and anti-nestin Abs, and the

relative protein levels were determined using GAPDH protein levels as loading control. Data shown are representative of one of three

separate experiments. Data shown are relative to U- and D-GSC #83 lines and are representative of the three GSC lines analyzed.

Can

cerCellBiology



intensity with an intermediate. Unstimulated cells were ana-lyzed to establish baseline fluorescence levels.

Grafting of GSCs into immunodeficient mice

Studies involving animals were approved by the EthicalCommittee of the Catholic University School of Medicine,Rome. Tumors were induced by subcutaneous flank inocu-lation of GSC lines into 4- to 5-week-old athymic nudemice of both sexes (Harlan, Udine, Italy). A total of 1 �105 GSCs in 100 ll of PBS supplemented with 0.1% glu-cose from wild-type, TRPV2- and Vector-transfected GSClines were grafted in nude athymic mice. Mice were sacri-ficed at 12 weeks; tumors were removed and measuredwith an external calliper. Tumor diameter was calculatedaccording to the equation D ¼ a þ b/2, where a is thelongest diameter and b is the shortest diameter.34 Tumors

were fixed in 4% PFA, paraffin embedded, cut in 5-mm-thick sections and stained with H&E. Mitotic index (MI)was determined on H&E-stained sections as the percentageof mitoses relative to the total number of cells. In each tu-mor, about 2,500 tumor cells were counted. Assessment ofthe MI was performed by a pathologist.

Immunohistochemistry

Immunohistochemistry analysis was performed on deparaffi-nized sections of wild-type, TRPV2- and Vector-transfectedtumor xenograft using the avidin–biotin–peroxidase complexmethod as previously published.32 The anti-human GFAP Abwas used to assess astrocytic differentiation of the tumor xen-ografts. The sections were counterstained with H&E andmounted, and GFAP expression was determined randomlyby using the BX51 microscope.

Figure 2. TRPV2-induced Ca2þ influx is implicated in U-GSC lines differentiation. (a) Intracellular free calcium levels were measured by

monitoring Fluo 3-AM in U- and D-GSC (differentiation days 3, 7 and 14) lines, *p < 0.01 vs. U-GSC lines (student’s t-test). Intracellular free

calcium levels in Vector- and TRPV2-transfected U-GSC lines (b) and siGLO and siTRPV2-transfected D-GSC lines (day 14) (c). Statistical

analysis was performed by comparing TRPV2- vs. Vector-U-GSC and siTRPV2- vs. siGLO-transfected D-GSC lines, *p < 0.01. U-GSC lines (d)

and D-GSC lines (e) were treated with 25 lM CBD alone or in combination with 25 lM RR. Time course of [Ca2þ]i rise in Fluo 3-loaded U-

and D-GSC lines was evaluated by FACS analysis. Statistical analysis was performed by comparing CBD vs. vehicle (*), CBD þ RR vs. CBD

(*) and RR vs. vehicle (#). Data shown are relative to U- and D-GSC #83 lines and are representative of the three U- and D-GSC lines

analyzed.

Can

cerCellBiology



Statistical analysis

Data represent the mean 6 standard deviation (SD) of atleast three separate experiments. The statistical significancewas determined by ANOVA or student’s t-test, *p < 0.01.

ResultsGSC lines differentiation results in increased

TRPV2 expression

U-GSC lines formed spheres and were immunoreactive fornestin in proliferation medium, whereas D-GSC lines acquiredadherent shape and expression of both glial (GFAP) and neu-ronal (bIII-tubulin) markers in differentiation medium (S 1A).

TRPV2 mRNA expression was evaluated in U-GSC andD-GSC lines by dxRT-PCR. We found that the U-GSC linesexamined express comparable TRPV2 mRNA levels (Fig. 1a),whereas a marked TRPV2 mRNA increase was detected inD-GSC lines with respect to the corresponding U-GSC line

(Fig. 1b). Because TRPV2 is mainly localized in plasma mem-brane (M) and in cytoplasm (C), we isolated M and C frac-tions from U- and D-GSC lines. By Western blot analysis, wefound that in U-GSC lines low TRPV2 protein levels weremainly localized in M fraction, whereas a TRPV2 levelincrease was detected both in M than in C fraction of D-GSC lines, with a prominent TRPV2 localization in M fraction(Fig. 1c). In addition, a time course analysis of differentiationmarkers during U-GSC lines differentiation was determined byWestern blot analysis. Data evidenced a rapid increase ofGFAP, bIII-tubulin and TRPV2 (both M and C fractions),accompanied by a rapid decrease of nestin expression (Fig. 1d).Moreover, analysis of TRPV2 expression in U- and D-GSClines was evaluated by FACS. Data confirmed an increase ofGFAP, bIII-tubulin and TRPV2 expression, whereas a concom-itant decrease of nestin level was observed in all D-GSC lineswith respect to the U-GSCs counterpart (S1). In addition, theanalysis of bIII-tubulin and GFAP expression on TRPV2þ-cells

Figure 3. TRPV2 promotes GSCs differentiation. (a and b) In U- and D-GSC lines (differentiation days 3, 7 and 14), treated with RR (25 lM)

or EGTA (1 mM), GFAP (a) and bIII-tubulin (b) levels were quantified by FACS analysis. Data shown are expressed as mean fluorescence

intensity, *p < 0.01 vs. untreated. (c and d) The expression of GFAP (c) and bIII-tubulin (d) in siTRPV2-siGLO-transfected and untransfected

U- and D-GSC lines (differentiation days 3, 7 and 14) was evaluated by FACS analysis. Data are expressed as mean fluorescence

intensity,*p < 0.01 vs. siGLO-transfected D-GSC lines. Data shown are relative to U- and D-GSC #83 lines and are representative of the

three U- and D-GSC lines analyzed.

Can

cerCellBiology



showed the presence of an aberrant differentiated TRPV2þ-GSC population coexpressing both bIII-tubulin and GFAP.Moreover, a time-dependent increase of GFAP expression thatpeaks at day 14 and a transient increase of bIII-tubulin expres-sion until day 7 that is reverted at later time were observed(S2). Overall, U-GSC lines express low TRPV2 levels, and theirdifferentiation is characterized by an increase of TRPV2 anddifferentiation markers.

TRPV2-induced Ca21 influx is implicated in

GSC differentiation

Because TRPV2 is a cation channel,18 we examined [Ca2þ]ibasal level in U-GSC and D-GSC lines during differentia-

tion. As shown in Figure 2a, [Ca2þ]i basal level increasedduring differentiation in a time-dependent manner (aboutfourfold increase at day 14). To demonstrate that TRPV2affects [Ca2þ]i levels, we stable transfected TRPV2 in U-GSC lines or we silenced TRPV2 gene during U-GSC linesdifferentiation (Supporting Information Fig. S3). Calciummobilization assay indicated that overexpression or knock-down of TRPV2 strongly influenced [Ca2þ]i level (Figs. 2band 2c). To evaluate if TRPV2 was functional, we stimu-lated TRPV2 channels with 25 lM CBD21 to induce[Ca2þ]i rise in U- and D-GSC lines. CBD treatmentsresulted in a higher increase of [Ca2þ]i levels in D-GSClines than in U-GSC lines (Figs. 2d and 2e). Moreover, the

Figure 4. PMA inhibits GSC lines differentiation. (a) Cell viability on untreated or 3-day PMA (10 nM)-treated D-GSC lines at day 11 of

differentiation was determined by WST assay. *p < 0.01 vs. untreated cells (student’s t-test). (b) The effects of PMA treatment above

described on differentiation markers were evaluated by FACS analysis. Untreated or PMA-treated D-GSC lines were analyzed for GFAP or bIII-

tubulin (gray area) and secondary Abs (white area). Data shown are expressed as percentage of positive cells when compared to respective

secondary Ab and are representative of one of three separate experiments. (c) TRPV2 expression (gray area) on untreated or PMA-treated D-

GSC lines, detected by FACS analysis, and secondary Abs (white area). (d) M and C fractions from untreated or PMA-treated D-GSC lines

were immunoblotted with anti-TRPV2 Ab, and protein levels were expressed as relative fold with respect to D-GSCs untreated. GAPDH

protein levels were used as loading control, *p < 0.01 vs. untreated D-GSC lines. Data are representative of one of three separate

experiments. Data shown are relative to D-GSC #83 line and are representative of the three D-GSC lines analyzed.

Can

cerCellBiology



TRP antagonist RR35 at 25 lM rapidly reverted the CBD-induced [Ca2þ]i influx and decreased basal [Ca2þ]i levelsboth in U- and D-GSC lines (Figs. 2d and 2e). Further-more, we examined the effect of TRPV2-mediated Ca2þ

influx during differentiation by growing U-GSC lines inthe presence of 1 mM of EGTA (as Ca2þ chelator) or 25lM of RR. We found that EGTA as RR treatmentsstrongly reduced the differentiation markers GFAP andbIII-tubulin (Figs. 3a and 3b). To ascribe a direct role ofTRPV2-mediated Ca2þ influx, we knocked down TRPV2during differentiation of U-GSC lines (Supporting Informa-

tion Fig. S3), and GFAP and bIII-tubulin levels weredetected up to 14 days of differentiation. The resultsshowed a significant reduction of GFAP and bIII-tubulinexpression in siTRPV2- when compared to siGLO-trans-fected D-GSC lines, suggesting that loss of TRPV2 delayedGSC differentiation (Figs. 3c and 3d). No significative dif-ferences were found in untransfected and siGLO-transfectedD-GSC lines (Figs. 3c and 3d). On the basis of these evi-dences, we suggest that differentiation is faster inGSC lines, where major TRPV2 levels are expressed, com-pared to U-GSC lines (Supporting Information Fig. S3).

Figure 5. TRPV2 expression increases GSC lines differentiation. (a) The expression of GFAP and bIII-tubulin in Vector- and TRPV2-U-GSC

lines was evaluated by FACS analysis, after 3 incubation days. Transfected U-GSC lines were stained with anti-GFAP and anti-bIII-tubulin Abs

(red lines). Black lines indicate secondary Abs. (b) The percentage of BrdUþ cells from Vector- and TRPV2-transfected U-GSC lines was

determined by FACS analysis. (c) The percentage of BrdUþ GFAPþ or BrdUþ bIII-tubulinþ from Vector- and TRPV2-transfected U-GSC lines

was evaluated by staining transfected U-GSC and pulsed with BrdU, anti-GFAP or anti-bIII-tubulin Abs. Data are expressed as percent of

positive cells 6 SD, *p < 0.01 vs. Vector-U-GSC lines (Student’s t-test). All data shown, expressed as percent of positive cells, are

representative of one of three separate experiments. (d) Representative phase-contrast photomicrographs depicting colony formation in soft

agar assay for Vector- and TRPV2-transfected GSCs. Total soft agar colony counts were done visualizing individual colonies in ten random

fields. Data shown are representative of one of two separate experiments, *p < 0.01 vs. Vector-GSCs. Bar: 500 lm. All data shown are

relative to transfected U-GSC #83 line and are representative of the three U-GSC lines analyzed.

Can

cerCellBiology



Taken together, our data indicate that TRPV2 is involved inpromoting Ca2þ-dependent serum-induced U-GSCsdifferentiation.

PMA reverts the differentiation of U-GSC lines

PMA has been found to stimulate the proliferation of glio-blastoma cells.36 Thus, we evaluated the effect of PMA onTRPV2 and differentiation markers expression in D-GSC.Three-day treatment of D-GSC lines with PMA, at differen-tiation day 11, resulted in stimulation of cell proliferation(Fig. 4a) with concomitant reduction of GFAP, bIII-tubulinand TRPV2 expression, as evaluated by cytofluorimetric anal-ysis (Figs. 4b and 4c). In addition, by Western blot analysis,we demonstrated that TRPV2 levels decreased in both M andC fractions of PMA-treated D-GSC lines with respect tountreated ones (Fig. 4d). Overall, these results indicate thatPMA-induced D-GSC proliferation was correlated withdecreased expression of TRPV2 and differentiation markers,further supporting an interplay between TRPV2 and U-GSClines differentiation.

TRPV2 overexpression increases the differentiation

of U-GSC lines

To enforce the suggested TRPV2 involvement in influenc-ing differentiation of U-GSC lines, we analyzed prolifera-tion and differentiation markers in TRPV2-transfected U-GSC lines. The results showed a strong increase of GFAPand bIII-tubulin expression in TRPV2-GSC compared toVector-GSC lines (Fig. 5a). In addition, by BrdU incorpo-ration assays, we found that TRPV2-GSC lines showed areduced cell proliferation with respect to Vector-GSC lines(Fig. 5b). Moreover, by FACS analysis, we showed thatTRPV2 overexpression mainly affects BrdUþ-bIII-tubulin

þ-U-GSC populations compared to Vector-GSCs (Fig. 5c).Finally, by soft agar assays, we confirmed that TRPV2 lev-els influenced U-GSC lines proliferation, because a signifi-cant reduction of neurosphere formation was observed inTRPV2-transfected GSC lines when compared to Vector-GSCs by conducting serial sphere formation (up to ter-tially) (Fig. 5d). No significative differences were found inuntransfected and Vector-transfected GSC lines (results not

Figure 6. Tumors derived from transplanted TRPV2-transfected U-GSC lines show an impaired MI and increased astroglial phenotype. (a

and b) U-GSC, Vector- and TRPV2-transfected U-GSC lines were subcutaneously injected into nude mice (n ¼ 5). Graphs show the tumor

diameter (a) and MI (b) of implanted tumors at 12 weeks. Statistical analysis was performed by ANOVA comparing TRPV2- vs. Vector- and

wild-type-derived tumors, *p < 0.01. (c) Tumor xenograft sections were analyzed for histological features by staining with H&E dyes and for

GFAP expression. Bar: 50 lm. Arrow indicates MNGCs. Data shown are relative to transfected U-GSC #83 line and are representative of the

three U-GSC lines analyzed.

Can

cerCellBiology



shown). Therefore, these data indicate that by modulatingTRPV2 expression and activation, the proliferative andstemness U-GSCs phenotype may be influenced. Altogether,these findings suggest TRPV2 overexpression results in areduced proliferation as well as an enhanced U-GSC linesdifferentiation.

TRPV2 overexpression reduces the proliferation and

promotes the differentiation of GSC-derived tumors

On the basis of our finding that TRPV2 regulates glioma cellproliferation/differentiation in vitro, we sought to determinethe impact of TRPV2 on the ability of U-GSC lines to stimu-late the gliomagenesis in vivo. Thus, we injected 1 � 105

TRPV2- or vector-transfected U-GSC lines in nude mice toinduce tumor formation. Tumors generated from TRPV2-U-GSC-injected mice showed a significant reduction of tumordiameter (Fig. 6a) and MI (Fig. 6b) when compared to thosearisen from empty vector U-GSC lines. Moreover, to evaluatethe involvement of TRPV2 in the promotion of U-GSC dif-ferentiation in vivo, we analyzed the expression of GFAP inGSC-derived tumors by immunohistochemistry. Tumorsderived from wild-type or vector-transfected U-GSCs werecharacterized by an epitheloid appearance with strands ofanaplastic cells showing round morphology and perinuclearexpression of GFAP (Fig. 6c). Conversely, the tumor xeno-graft developed upon injection of TRPV2-U-GSC lines wasmainly composed by GFAPþ cells with a multipolar mor-phology extending their cytoplasmic processes into the ma-trix to form a fibrillary network. GFAP selectively labeled thetumor cells as demonstrated by the absence of immunostain-ing in the vascular and fibrous structures of the tumor xeno-graft (data not shown). Multinucleated giant glioblastomacells (MNGCs) were also found in the TRPV2-U-GSC lines-derived xenograft (Fig. 6c).

DiscussionAs GSCs may have a pivotal role in tumor development,improving knowledges on their biology may have importantimplications both for the development of new therapeuticstrategies and for evaluation of potential pitfalls and benefitsof currently available treatments.37 The malignancy of humanbrain tumors inversely correlates with their degree of differ-entiation as shown by increased nestin expression,38 whereastheir mitotic activity is inversely associated with increasedexpression of mature glial and neuronal markers.39 The aimof our work was to first demonstrate the expression ofTRPV2 channel in GSC lines at mRNA and protein levels.Herein, we demonstrated that TRPV2 is expressed in GSClines and that their differentiation was associated to a highincrease of TRPV2 levels. In addition, we also found thatGSC lines differentiation, evaluated by the induction of anadherent phenotype and glial morphology, reduced nestinand increased GFAP and bIII-tubulin expression with manycells coexpressing both markers and was associated with amarked enhancement of TRPV2. Moreover, by Western blot

analysis, we found that as reported in TRPV2-transfectedcells,25 serum-induced differentiation resulted in enhancedTRPV2 localization both in the cytosolic and plasma mem-brane fractions of GSC lines. Because TRPV2 is a cationchannel showing Ca2þ permeability,18 we evaluated if incre-ments in TRPV2 expression, detected during GSC differentia-tion, were associated with changes in [Ca2þ]i levels. The dataindicated that TRPV2 influenced basal [Ca2þ]i levels duringU-GSC lines differentiation. In addition, transfection ofTRPV2 cDNA in GSC lines or knockdown of TRPV2 gene indifferentiated GSC lines markedly enhanced or reduced,respectively, the intracellular Ca2þ levels. Furthermore, CBDstrongly stimulated [Ca2þ]i levels in differentiated GSC lines,and this effect was completely reverted by the TRP antago-nist, RR. Moreover, RR decreased the basal [Ca2þ]i level ofU-GSC lines and markedly reduced [Ca2þ]i level in D-GSClines, suggesting that other calcium channels may be involvedin [Ca2þ]i rise during GSCs differentiation.40 To better clarifythe role of [Ca2þ]i influx in influencing GSC lines differentia-tion, we grew GSC lines up to day 14 of differentiation in thepresence of RR or EGTA. Data indicated that RR treatmentscompletely inhibited calcium rise and markedly impairedGSC differentiation as evaluated by the reduced GFAP andbIII-tubulin expression. Similar results were obtained byEGTA treatments. Finally, to determine a specific role forTRPV2 in GSC differentiation process, we knocked downTRPV2 gene, which resulted in a reduced percentage ofGFAPþ and bIII-tubulin

þ GSC subpopulations. In addition,reduced TRPV2 expression was observed upon treatment ofD-GSC lines with PMA, which stimulated GSCs growth andsignificantly impaired GSCs differentiation. On the contrary,TRPV2 overexpression in U-GSC-transfected cell linesresulted in a more mature glial phenotype as demonstratedby increased expression of GFAP and bIII-tubulin markers,associated with a reduced proliferation. Overall, these resultsare consistent with our recent evidence showing that TRPV2negatively controls glioma cell survival and proliferation.17

Data obtained in vitro have also been evaluated in vivo. Wefound that tumors derived from TRPV2-overexpressing GSClines showed a significant reduction in diameter and MI witha differentiated morphology and a GFAP-positive phenotype.Interestingly, TRPV2-expressing tumors evidenced the pres-ence of several GFAPþ MNGCs resembling to that found inGlioblastoma multiforme subtypes with low proliferative rateand better prognosis.41–43 Regarding TRP channels andMNGCs, loss of TRPC1, by causing incomplete cytokinesis,results in multinucleated glioma cell generation and slow gli-oma growth.44,45 Overall, our data in TRPV2-overexpressingGSCs-derived xenografts, showing the presence of GSCs witha mature astrocyte phenotype and an impaired proliferativepotential, suggest for a role of TRPV2 in GSC-derived tumorprogression. The existence of a brain tumor-initiating cellphenotype with stem cell features may lead in the future topotential therapeutic strategies based on enforced stem celldifferentiation.12 In this context, deeper investigation will be

Can

cerCellBiology



required, also by using GSC orthotopic xenograft, to com-pletely identify the role of TRPV2 channel in the control ofthe balance between proliferation and differentiation ofGSCs. In conclusion, our results demonstrating the ability of

the TRPV2 channel to stimulate in vitro and in vivo GSClines differentiation and in parallel to inhibit their prolifera-tion may open new opportunities for the use of TRPV2 ago-nist-based antitumor strategies.

References

1. Maher EA, Furnari FB, Bachoo RM, et al.Malignant glioma: genetics and biology of a gravematter. Genes Dev 2001;15:1311–33.

2. Sanai N, Alvarez-Buylla A, Berger MS. Neuralstem cells and the origin of glioma. N Engl J Med2005;353:811–22.

3. Vescovi AL, Galli R, et al. Brain tumor stem cells.Nat Rev Cancer 2006;6:425–36.

4. Ignatova T, Kukekov VG, Laywell ED, et al.Human cortical glial tumors contain neuralstem-like cells expressing astroglial andneuronal markers in vitro. Glia 2002;39:193–206.

5. Hemmati HD, Nakano I, Lazareff JA, et al.Cancerous stem cells can arise from pediatricbrain tumors. Proc Natl Acad Sci USA 2003;100:15178–83.

6. Galli R, Binda E, Orfanelli U, et al. Isolation andcharacterization of tumorigenic stem-like neuralprecursors from human glioblastoma. Cancer Res2004;64:7011–21.

7. Singh SK, Clarke ID, Terasaki M, et al.Identification of a cancer stem cell in humanbrain tumors. Cancer Res 2003;63:5821–8.

8. Singh SK, Hawkins C, Clarke ID, et al.Identification of human brain tumour initiatingcells. Nature 2004;432:396–401.

9. Kondo T, Setoguchi T, Taga T. Persistence of asmall subpopulation of cancer stem-like cells inthe C6 glioma cell line. Proc Natl Acad Sci USA2004;101:781–6.

10. Patrawala L, Calhoun T, Schneider-Broussard R,et al. Side population is enriched in tumorigenic,stem-like cancer cells, whereas ABCG2þ andABCG2� cancer cells are similarly tumorigenic.Cancer Res 2005;65:6207–19.

11. Dirks PB. Cancer: stem cells and brain tumors.Nature 2006;444:687–8.

12. Campos B, Wan F, Farhadi M, et al.Differentiation therapy exerts antitumor effectson stem-like glioma cells. Clin Cancer Res 2010;16, 2715–28.

13. Prevarskaya N, Zhang L, Barritt G. TRP channelsin cancer. Biochim Biophys Acta 2007;1772:937–46.

14. B€odding M. TRP proteins and cancer. Cell Signal2007;19:617–24.

15. Caprodossi S, Lucciarini R, Amantini C, et al.Transient receptor potential vanilloid type 2(TRPV2) expression in normal urothelium and inurothelial carcinoma of human bladder:correlation with the pathologic stage. Eur Urol2008;54:612–20.

16. Amantini C, Ballarini P, Caprodossi S, et al.Triggering of transient receptor potentialvanilloid type 1 (TRPV1) by capsaicin inducesFas/CD95-mediated apoptosis of urothelialcancer cells in an ATM-dependent manner.Carcinogenesis 2009;30:1320–9.

17. Nabissi M, Morelli MB, Amantini C, et al.TRPV2 channel negatively controls glioma cellproliferation and resistance to Fas-inducedapoptosis in ERK-dependent manner.Carcinogenesis 2010;31:794–803.

18. Caterina MJ, Rosen TA, Tominaga M, et al. Acapsaicin-receptor homologue with a highthreshold for noxious heat. Nature 1999;398:436–41.

19. Muraki K, Iwata Y, Katanosaka Y, et al. TRPV2is a component of osmotically sensitive cationchannels in murine aortic myocytes. Circ Res2003;93:829–38.

20. Zimmermann K, Leffler A, Fischer MM, et al.The TRPV1/2/3 activator 2-aminoethoxydiphenylborate sensitizes native nociceptive neurons toheat in wild type but not TRPV1 deficient mice.Neuroscience 2005;135:1277–84.

21. Qin N, Neeper MP, Liu Y, et al. TRPV2 isactivated by cannabidiol and mediates CGRPrelease in cultured rat dorsal root ganglionneurons. J Neurosci 2008;28:6231–8.

22. Kowase T, Nakazato Y, Yoko-O H, et al.Immunohistochemical localization of growthfactor-regulated channel (GRC) in human tissues.Endocr J 2002;49:349–55.

23. Saunders CI, Kunde DA, Crawford A, et al.Expression of transient receptor potentialvanilloid 1 (TRPV1) and 2 (TRPV2) in humanperipheral blood. Mol Immunol 2007;44:1429–35.

24. Kanzaki M, Zhang YQ, Mashima H, et al.Translocation of a calcium-permeable cationchannel induced by insulin-like growth factor-I.Nat Cell Biol 1999;1:165–70.

25. Penna A, Juvin V, Chemin J, et al. PI3-kinasepromotes TRPV2 activity independently ofchannel translocation to the plasma membrane.Cell Calcium 2006;39:495–507.

26. Monet M, Gkika D, Lehen’kyi V, et al.Lysophospholipids stimulate prostate cancer cellmigration via TRPV2 channel activation. BiochimBiophys Acta 2009;1793:528–39.

27. Monet M, Lehen’kyi V, Gackiere F, et al. Role ofcationic channel TRPV2 in promoting prostatecancer migration and progression to androgenresistance. Cancer Res 2010;70:1225–35.

28. Liu G, Xie C, Sun F, et al. Clinical significance oftransient receptor potential vanilloid 2 expressionin human hepatocellular carcinoma. Cancer GenetCytogenet 2010;197:54–9.

29. Ricci-Vitiani L, Pallini R, Biffoni M, et al.Tumour vascularization via endothelialdifferentiation of glioblastoma stem-like cells.Nature 2010;468:824–8.

30. Kleihues P, Cavenee WK, eds. Pathology andgenetics of tumours of the nervous system, WorldHealth Organization classification of tumours,2nd edn. Lyon: IARC Press, 2000. 10–21.

31. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK,Burger PC, Jouvet A, Scheithauer BW, KleihuesP. The 2007 WHO classification of tumours ofthe central nervous system. Acta Neuropathol2007;114:97–109.

32. Ricci-Vitiani L, Pallini R, Larocca LM, et al.Mesenchymal differentiation of glioblastoma stemcells. Cell Death Differ 2008;15:1491–8.

33. Ricci-Vitiani L, Pedini F, Mollinari C, et al.Absence of caspase 8 and high expression ofPED protect primitive neural cells from celldeath. J Exp Med 2004;200:1257–66.

34. Pallini R, Ricci-Vitiani L, Bez A, et al.Homologous transplantation of neural stem cellsto the injured spinal cord of mice. Neurosurgery2005;57:1014–25.

35. Szallasi A, Blumberg PM. Vanilloid (Capsaicin)receptors and mechanisms. Pharmacol Rev 1999;51:159–212.

36. Aeder SE, Martin PM, Soh JW, et al. PKC-etamediates glioblastoma cell proliferation throughthe Akt and mTOR signaling pathways. Oncogene2004;23:9062–9.

37. Ho A, Fusenig N. Cancer stem cells: a promisingconcept and therapeutic challenge. Int J Cancer2011;129:2309.

38. Jang T, Litofsky NS, Smith TW, et al. Aberrantnestin expression during ethylnitrosurea-(ENU)-induced neurocarcinogenesis. Neurobiol Dis 2004;15:544–52.

39. Englund C, Alvord EC, Folkerth RD, et al.NeuN expression correlates with reducedmitotic index of neoplastic cells in centralneurocytomas. Neuropathol Appl Neurobiol2005;31:429–38.

40. McFerrin MB, Sontheimer H. A role for ionchannels in glioma cell invasion. Neuron GliaBiol 2006;2:39–49.

41. McLendon RE, Friedman AH, Gray L.Glioblastoma, 7th edn. London: Hodder Arnold,2006.

42. Kawano H, Kubota T, Sato K, et al.Immunohistochemical study of giant cellin glioblastoma. Clin Neuropathol 1995;14:118–23.

43. Ohgaki H, Biernat W, Reis R, et al. Gliosarcoma.In: Kleihues P, Cavenee WK, eds. Pathology andgenetics of tumours of the nervous system, WorldHealth Organization classification of tumours.Lyon: IARC Press, 2000. 40–41.

44. Bomben VC, Sontheimer H. Inhibition oftransient receptor potential canonical channelsimpairs cytokinesis in human malignant gliomas.Cell Prolif 2008;41:98–121.

45. Bomben VC, Sontheimer H. Disruption oftransient receptor potential canonical channel 1causes incomplete cytokinesis and slows thegrowth of human malignant gliomas. Glia 2010;58:1145–56.

Can

cerCellBiology



The transient receptor potential vanilloid-2 cation channel impairs glioblastoma stem-like cell...

Documents

Transcript of The transient receptor potential vanilloid-2 cation channel impairs glioblastoma stem-like cell...