The Gs-Linked Receptor GPR3 Inhibits the Proliferation of Cerebellar Granule Cells during ...

12 Pages·2009·3.46 MB·English

The Gs-Linked Receptor GPR3 Inhibits the Proliferation
of Cerebellar Granule Cells during Postnatal
Development
Shigeru Tanaka¤, Imran Mohammed Shaikh, E. Antonio Chiocca*, Yoshinaga Saeki
DardingerLaboratoryforNeuro-oncologyandNeurosciences,DepartmentofNeurologicalSurgery,TheOhioStateUniversity,Columbus,Ohio,UnitedStatesofAmerica
Abstract
Background:During postnatal murine and rodent cerebellar development, cerebellar granule precursors (CGP) gradually
stopproliferatingastheydifferentiateaftermigrationtotheinternalgranulelayer(IGL).Moleculareventsthatgovernthis
program remain to be fully elucidated. GPR3 belongs to a family of Gs-linked receptors that activate cyclic AMP and are
abundantly expressed inthe adult brain.
Methodology/PrincipalFindings:ToinvestigatetheroleofthisorphanreceptorinCGPdifferentiation,wedeterminedthat
exogenous GPR3 expression in rat cerebellar granule neurons partially antagonized the proliferative effect of Sonic
hedgehog (Shh), while endogenous GPR3 inhibition by siRNA stimulated Shh-induced CGP proliferation. In addition,
exogenousGPR3expressioninCGPscorrelatedwithincreasedp27/kipexpression,whileGPR3knock-downledtoadecrease
inp27/kipexpression.Inwild-typemice,GPR3expressionincreasedpostnatallyanditsexpressionwasconcentratedinthe
internalgranularlayer(IGL).InGPR32/2mice,theIGLwaswidenedwithincreasedproliferationofCGPs,asmeasuredby
bromodeoxyuridine incorporation. Cell cycle kinetics of GPR3-transfected medulloblastoma cells revealed a G0/G1 block,
consistentwith cellcycle exit.
Conclusions/Significance:These results thus indicate that GPR3 is a novel antiproliferative mediator of CGPs in the
postnatal developmentofmurine cerebellum.
Citation:TanakaS,ShaikhIM,ChioccaEA,SaekiY(2009)TheGs-LinkedReceptorGPR3InhibitstheProliferationofCerebellarGranuleCellsduringPostnatal
Development.PLoSONE4(6):e5922.doi:10.1371/journal.pone.0005922
Editor:LinMei,MedicalCollegeofGeorgia,UnitedStatesofAmerica
ReceivedFebruary24,2009;AcceptedMay3,2009;PublishedJune15,2009
Copyright: (cid:1)2009 Tanaka et al. Thisis an open-access article distributed under theterms of the Creative Commons Attribution License, which permits
unrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited.
Funding:TheauthorswishtoacknowledgethegenerosityoftheGerlachFoundationinfundingthisstudy.ThestudywasalsosupportedbytheDardinger
Neuro-oncologyFund.Thefundershadnoroleinstudydesign,datacollectionandanalysis,decisiontopublish,orpreparationofthemanuscript.
CompetingInterests:Theauthorshavedeclaredthatnocompetinginterestsexist.
*E-mail:[email protected]
¤ Current address: Department of Molecular and PharmacologicalNeuroscience, Graduate School of Biomedical Sciences, Hiroshima University, Minami-ku,
Hiroshima,Japan
Introduction the IGL over the 3 weeks of postnatal development are only
partially understood. The Hedgehog (HH) signaling pathway, an
In the adult, the cerebellum is organized into distinct layers, important mitogenic signal for GNPs[9], is the likely factor that
each containing specialized neuronal cell populations, positioned regulates granule progenitor cell (GPC) proliferation [10–12,13].
there during postnatal development with precise coordination of Indeed,mutationsthatactivatetheHedgehogpathwaycontribute
processes involving cell proliferation, differentiation, and migra- totheformationofmedulloblastoma[14–20,21].Sonichedgehog
tion. In the immediate postnatal period, the outermost layer, the (Shh) is the most widely studied member of this pathway [22–
external granule layer (EGL), contains dividing granule neuron 25,26],andoneofitsmostimportanttargetsistoactivatetheGli
progenitors(GNP).Thesegranulecellsmigrateinwards,guidedby family of transcription factors, including transcription of the Gli
Bergmann radial glial fibers through the molecular layer (ML), family of transcription factors [27,28]. Shh responses can be
pass through the Purkinje neuron layer (PL) containing the cell antagonized by increasing cAMP levels and PKA activity
bodies of Purkinje neurons and Bergmann glia and finally settle [29,30,31]. Regulators of Shh signaling include pituitary adeny-
intotheinternalgranularlayer(IGL)wheretheyexitthecellcycle late-cyclase activating peptide (PACAP) that raises PKA activity
andterminallydifferentiate[1–8].Thisprocessisrelativelyrapidly: and antagonizes Shh[31,32,33]. p27/Kip1 isanother factor that
in fact, by the end of the third postnatal week in the mouse, the has been reported to be predominantly expressed in the
EGLhasceasedtoexistandtheshapeofthecerebellumhasbeen developingcerebellumandtoplayacriticalroleintheregulation
transformedfromitsembryonicflat,ovalformtoamorecomplex of GCP proliferation [34] by promoting cell cycle exit of GCPs
structure withdeep fissuresandlarge folia. andantagonizing Shh action.
The molecular events responsible for this ‘‘switch’’ from The G-protein coupled receptor GPR3 (GPCR21) is a
proliferation in the EGL to cell cycle exit and differentiation in constitutive activator of intracellular cAMP [35–37]. GPR3
PLoSONE | www.plosone.org 1 June2009 | Volume 4 | Issue 6 | e5922
GPR3InhibitsGPProliferation
belongs to a family of G protein coupled receptors (GPCR), plated onto 24-well culture plates (BD Falcon, BD Biosciences,
predominantly expressed in mammalian brain. Up-regulation of Franklin Lakes, New Jersey) or chamber glass slides (Nalge Nunc
these receptors in cultured cells stimulates adenylate cyclases to International,Rochester,NY)coatedwith100mg/mLofpoly-D-
levels similar to those obtained with other G-coupled receptors lysine (PDL; Sigma-Aldrich, St. Louis, MO) and 2mg/mL of
s
that have been fully stimulated by their ligands [36,37]. GPR3 is laminin (BD Biosciences, San Jose, CA). Four to 6hours after
expressed in oocytes and maintains meiotic arrest with a Gs- electroporation, culture medium was replaced with neurobasal-A
dependent mechanism[38,39]. medium (Invitrogen, St. Louis, MO) with 200mM L-glutamine,
We hypothesized that GPR3 activation may represent one of 2% B27 supplement (Invitrogen, Carlsbad, CA), penicillin
thesignalsthatmediatesthepostnatalcellcycleexitandterminal (100 U/mL) (Invitrogen, Carlsbad, CA), and streptomycin
differentiationofGPCs.Thishypothesiswasbasedonthefinding (100 ug/mL) (Invitrogen, Carlsbad, CA). All cells were cultured
thatexpressionofGPR3up-regulatesintraneuronalcAMP[40].In at 37uC in an atmosphere containing 5% carbon dioxide.
addition, GPR3 mRNA expression was strongly increased in the Transfected GFP-positive neurons were imaged using a Zeiss
IGLduringcerebellumdevelopment[40].Inthisreport,weshow LSM510 META confocal microscope (Carl Zeiss Microimaging,
that in vitro GPR3 expression antagonizes Shh and is associated Inc., Thornwood, NY). In order to detect endogenous and
with p27/kip expression. In vivo, absence of GPR3 leads to a transfected levels of GPR3 , P7 rat cerebellar granule neurons
hyperproliferativeresponseofCGPsinIGL,suggestingthatitmay (56106 granule neurons) were electroporated with 3 mg of
be the granule cell receptor that mediates cAMP’s action on cell pHGCX (Empty) or pHGC-GPR3-RHA (GPR3-RHA). The
cycle exitof CGPsduring postnatal development. GPR3-RHA construct, which expresses mRFP RNA fused with
therecombinantGPR3RNA,allowstheuseofthemRFP-specific
Materials and Methods primers for quantitation of recombinant vs. endogenous GPR3
expression. Neurons were plated onto PDL/Laminin-coated
Ethics statement/Animals chamber slides. Twenty-four hours after transfection, total RNA
All animal experiments were performed according to the wasextractedfromtheneuronsandsubjectedtoquantitativereal-
guidelines of the Subcommittee on Research Animal Care of timeRT-PCRusingprimersandprobesspecificforratGPR3and
The Ohio State University, a IACUUC approved veterinary mRFP. Untransfected neurons (PBS) were also used as a control.
facility. C57BL/6 mice were obtained from Charles River ThesequencesofPCRprimersandVIC-labeledTaqManprobes
Laboratories (Wilmington, MA) and SD rats were from Harlan specific for mRFP were59-ACTACGACGCCGAGGTCAAG-39,
Laboratories (Indianapolis, IN). The GPR32/2 knockout mice 59-TGGGAGGTGATGTCCAGCTT-39, and 59-VIC-ACATG-
were obtained fromDeltagen(Redwood City,CA). GCCAAGAAGCCCGTGCA-TAMRA-39. Ready-made FAM-
labeled TaqMan probe and primers specific for rat b-actin
Chemicals, proteins, and siRNAs (Rn00667869_m1) were used as an internal control to normalize
thedata.
Recombinant mouse Shh C-terminal peptide was purchased
fromR&DSystem(Minneapolis,MN).db-cAMPwerepurchased
from Calbiochem (La Jolla, CA). All other chemicals were from Immunofluorescence analysis
Sigma (St.Louis, MO), unless indicated otherwise. G SilencerH Cerebellar cells were resuspended in Neurobasal A medium
siRNA against rat GPR3 (59-CCUACUACUCAGAGACAACtt/ supplementedwithB27,penicillin-streptomycin,andL-glutamine
59GUUGUCUCUGAGUAGUAGGtg)andnegativecontrol#1 (2 mM) and plated on PDL-precoated eight-well chamber slides
siRNAwere purchased fromAmbion,Inc. (Austin, TX). (BD) at the density of 5.06105cells/well. Cells were cultured for
Isolation of rat cerebellar granule neurons and DNA 48hr and then fixed with 4% paraformaldehyde. Permeabiliza-
electroporation. Cerebellar granule neurons were isolated tionwasperformedwith0.1%Triton-X,followedby10%normal
from Sprague Dawley rats (Harlan, Indianapolis, IN) according donkeyblocking serum.
to a published procedure [6,41] with modifications. Briefly, pups
were sacrificed at postnatal day 7 (P7) and the whole cerebellum Cell proliferation assay
wasremoved.Themeningeswerecarefullystrippedoff,andwhole Cell proliferation was evaluated using the bromodeoxyuridine
tissue was then washed in calcium- and magnesium-free (BrdU) cell proliferation ELISA assay (Roche). Rat cerebellar
phosphate-buffered saline (PBS: 10mM sodium phosphate, neuronsfromP5orP7neonateswereisolatedasdescribedabove
0.9% sodium chloride pH=7.3) and dissociated into single cells and transferred to PDL-coated 24-well plates (BD) or eight-well
using the Worthington Papain Dissociation System (Worthington Lab-Tek chamber slides (Nunc, Naperville, IL), at the density of
Biochemical Corporation, Lakewood, NJ). The dissociated 8.06105cells/cm2.Six hours afterplating, medium waschanged
cerebellar cells were applied immediately onto a two-step to Neurobasal A medium supplemented with B27, penicillin-
gradient of Percoll (35%/60% in PBS; Sigma-Aldrich). After streptomycin, andL-glutamine (2mM)(Invitrogen). Stimuli were
centrifugation at 2,0006g for 10min at 4uC, a fraction enriched added at the time of medium replacement, and the cells were
with granule neurons was collected between the 35% and 60% culturedfortheindicatedamountoftime(48 hr,unlessotherwise
Percoll layers. The isolated neurons were washed once with PBS noted). Cells were pulsed with 10uM of BrdU (Roche), 12hr
andsubjectedtoelectroporationusingtheNucleofectorTMsystem beforetheendofthecultureperiod.BrdU-detectingELISAassay
(Amaxa Inc., Gaithersburg, MD). Briefly, 56106 dissociated wasperformedaccordingtomanufacture’srecommendation.The
neurons were spun down at 1386g for 5min at 4uC, colorimetric signal of each sample was detected and analyzed
resuspended in 100mL of Rat Neuron NucleofectorTM solution usingaFLUOstarOPTIMAmicroplatereader(BMGLABTECH
keptatroomtemperature,combinedwith3 mgofplasmidDNA, Inc., Durham, NC). For anti-BrdU immunostaining, cells were
transferredintoacuvetteandelectroporatedusingprogramG-13 fixed with 4% paraformaldehyde-PBS for 10min at room
of the Amaxa system. The electroporated neurons were temperature followed by 50% formaldehyde in sodium-citrate
immediately resuspended in 600mL of DMEM (Dulbecco’s buffer at 65uC for 2 hours. Sections were then rinsed in 0.1%
modified essential medium) supplemented with 10% fetal bovine Triton X-100 in PBS for 20min and incubated with Rat anti-
serum,penicillin(100U/mL),andstreptomycin(100 ug/mL)and BrdU antibody (Oxford Biolab, UK) at 4uC overnight. After
PLoSONE | www.plosone.org 2 June2009 | Volume 4 | Issue 6 | e5922
GPR3InhibitsGPProliferation
washing with PBS, the sections were incubated with Rhodamine postnatal cerebellar development. We electroporated cultured
labeled secondary antibody (1:400) (Jackson Immunolab, West cerebellar granule precursor cells (GCPs), isolated from P7 rat,
Grove,PA)for1 houratroomtemperature.Co-localizationwith with a GPR3/EGFP-expression vector and then exposed trans-
transducted GFP-positive neurons were evaluated using a Zeiss fectedcellstoShh.Althoughtheinitialtransfectionefficiencywas
LSM510 META confocal microscope (Carl Zeiss Microimaging, likelytobesimilarforalldishes,GPR3andcyclicAMPprovidea
Inc.). survival advantage to neurons (data not shown and [40]),
accounting for the observed increase in GFP-positive cells in
Real time PCR analysis those dishes. The proliferation of untransfected control cells was
significantly enhanced by exposure to Shh as measured by
Total RNA was isolated from transfected cells using TRIzol
incorporation of BrdU (fig. 1a). However, exposure to Shh did
reagent(Invitrogen).RNAsamplesweretreatedwiththeTURBO
DNA-freeTM kit(Ambion)andthenreversetranscribedintofirst- notstimulateasmuchproliferationofGPR3transfectedcells,both
strand cDNA by Superscript IITM reverse transcriptase (Invitro- visually(Fig.1a)anduponquantitativeenumeration(Fig.1b).As
expected, exposure to high doses of dibutyryl-cyclic AMP
gen) using oligo-dT primers according to the manufacturer’s
abrogated the proliferating action of Shh. This result was also
protocols. The sequences of PCR primers and VIC-labeled
confirmedbyusinganELISAassaytodetectBrdU-positivecells.
TaqMan probes specific for mouse GPR3 genes were selected
Asexpected,thenumberofBrdU-positiveGCPswasincreasedby
using PrimerExpressHsoftware(Applied Biosystems, Foster City,
Shh treatment in control, electroporated GCPs (Figure 1c).
CA): 59-CTGACCGCGTGGCTCTAGA-39, 59-CACTTGGG-
However, the number of BrdU-positive GCPs was significantly
CTGTGAGACATTTC-39, and 59-VIC-TGTTCCAGATGGT-
decreased inGPR3-transfected GCPs.
CAGGGTCCCACTC-TAMRA-39 for mouse GPR3; 59-CGCC-
We then asked whether exogenous GPR3 gene expression had
AACTCTCCTCCTCTCTAC-39,59-CGGGTTGATCATGGA-
an effect on the antiproliferative marker, p27/Kip1. GCPs were
GTTGTAA-39, and 59-VIC-CCTACCTTACCCTGCTCCCT-
electroporatedwiththeGPR3/EGFPexpressionorcontrolvector.
GC-TAMRA-39 for rat GPR3. Ready-made FAM-labeled
Figure1dshowsthat,48hoursaftertransfectionwiththeGPR3/
TaqMan probes and primers specific for mouse b-actin
EGFP vector, there was an increase in p27/Kip-expressing cells
(Mm00607939_s1, Applied Biosystems) or rat b-actin
compared to control that was also confirmed by quantitative
(Rn00667869_m1) were used as internal controls. Real-time
evaluation(figure1e).ThemRNAlevelsoftheendogenousGPR3
PCR analyses were carried out in 96-well optical reaction plates
and of the transfected recombinant mouse GPR3 were measured
using an ABI PRISMH 7900 HT sequence detection system
using the rat GPR3-specific primers and primers for mouse red
accordingtothemanufacturer’sprotocol.After2-minincubation
fluorescent protein (mRFP), respectively. In this experiment, we
at 50uC, the samples were denatured by a 10-min incubation at
usedtheGPR3-RHAconstruct,whichexpressesmRFPRNAfused
95uC and subjected to 40 cycles of amplification (95uC for 15s,
with the recombinant GPR3 RNA, allowing us to use the mRFP-
60uC for 1 min). The fluorescence signal from each well was
specific primers for quantitation of recombinant vs. endogenous
normalized using an internal passive reference. The cycle
GPR3 expression. The data shows that recombinant GPR3 is
threshold (C ) values obtained with the GPR3-specific probes
T expressedatlevels10-foldhigherthanthoseofendogenousGPR3
and primers were compared with those of b-actin-specific probe
(supplemental figure S1). These findings thus showed that
andprimersusingthecomparativeC methodasdescribedinthe
T exogenous GPR3 expression into cultured CGPs from postnatal
user manual (User Bulletin #2 forABIPrismH7700).
cerebellum partially counteracted the proliferative action of Shh
andwasassociatedwithexpressionoftheantiproliferativemarker
Cell cycle analysis by flow cytometry
(p27/kip).
DNA-indices for cell cycle analysis were assayed by multi-
parametricflowcytometryusingstandardmethods.Analyseswere GPR3 mRNA ‘‘knock-down’’ is associated with increased
performed using a Becton Dickinson FACScan flow cytometer
proliferation of CGPs in response to Shh and with
(BD Biosciences, San Jose, CA) for the detection of cells stained
reduced expression of p27/Kip
with propidium iodide (PI) and a 488nm laser with filter
We then asked if inhibition of endogenous GPR3 expression
combination for GFP. Single cell suspensions were isolated from
stimulatedShh-inducedproliferationofCGPs.Wehadpreviously
culture, fixed in methanol, and stained with PI (100 mg/mL in
shown that GPR3 was up-regulated postnatally in cerebellar
PBS). Each histogram represents 10,000–100,000 cells for
granule neuron precursors [40]. To determine if endogenous
measuring DNA-index and cell cycle. Histogram analysis was
GPR3 expression contributed to Shh-induced GCP proliferation,
performed with the CellQuest software (BD Biosciences) for
GPR3expressionwasknockeddownbysiRNAbyavaluethatwas
multiparametric calculationsand analyses.
,60%thatofcontrol(seebarinfigure2d).AfteradditionofShh,
proliferation of GCPs with ‘‘knocked-down’’ GPR3 was signifi-
Statistical analyses
cantly elevated compared to that of control GCPs, both visually
Statistical analysis was performed by one-way ANOVA
(figure2a)andquantitatively(figure2b).Asexpected,theaddition
followedbyFisher’sPLSDtest.P,0.05wasconsideredstatistically
of dibutyril cyclic AMP (dbcAMP) in the medium abolished the
significant.
proliferative action of Shh in GPR3-siRNA transfected GCPs.
BrdUimmunohistochemistryalsoconfirmedtheseresults(Fig.2c).
Results In addition, downstream effects of Shh signaling, such as Gli1
mRNA expression, were significantly reduced upon siRNA
Exogenous GPR3 expression inhibits Shh-induced
inhibitionofGPR3(datanotshown).Toprovidefurtherevidence
proliferation of cerebellar granule cell precursors (GCP)
for modulation of antiproliferative effects by GPR3, we also
and correlates with increased p27/kip evaluatedeffectsonp27/kipgeneexpression.Figure2dshowsthat
We sought to determine whether GPR3 expression inhibited expressionofp27/Kipwasreducedinthesecultures,withonlyan
proliferativesignals(Sonichedgehog-Shh)and/orcorrelatedwith approximate 60% ‘‘knock-down’’ of GRP3 mRNA. Taken in
antiproliferative markers (p27/kip), known to be important in conjunction, this evidence thus suggested a significant role for
PLoSONE | www.plosone.org 3 June2009 | Volume 4 | Issue 6 | e5922
GPR3InhibitsGPProliferation
Figure1.ExogenousGPR3partiallyinhibitsShh-inducedproliferationofrodentcerebellargranuleneuronsandisassociatedwith
increasedp27/Kipexpression.GCPsisolatedfromratcerebellum(P7)wereelectroporatedwithaGPR3/EGFPorcontrolvectorusingtheAMAXA
nucleofectorsystem.Shh(1ug/ml)wasthenadded6hourslater.Thirtyhourslater,BrdU(finalconcentration=10mM)wasadded12hourspriorto
fixation.CellproliferationwasdeterminedusingBrdUincorporationassessedbyimmunohistochemistrywithaBrdU-specificmonoclonalantibody
(panelsa,b).dbcAMP(100mM)wasusedasachemicalcAMPactivatorandasapositivecontrol.Inpanela,representativefieldsfromdishesare
visualizedfromtheShhandcontrolvectorgroup(toprow),ShhandGPR3/EGFP-electroporatedgroup(middlerow)andtheShh,controlvector
groupanddbcAMP(bottomrow).VisualizationbyBrdUimmunohistochemistry(reddye),EGFPfluorescence(greendye),andthecombinedmerged
imageareshown.Fiverandomfieldswereselectedpereachdish(n=3)fromtheShh,Shh+GPR3andShh+GPR3+cAMPgroups.ThenumberofGFP-
positivecells(representingpositivelytransfectedneurons)wascounted(atleastonehundredGFP-positiveneuronspereachfield).Doublylabeled
cells (yellowcolor)werealsoenumeratedper eachdish.Valueswerethenexpressedas thepercentageofdoubly-labeledcellsoutoftotalGFP
positive cells in the Shh+GPR3 group or Shh+GPR3+dbcAMP groups compared to Shh group, i.e. % Shh-treated cells (panel b). In parallel, cell
proliferationwasalsodeterminedusinganELISAassaybasedonthemeasurementofBrdUincorporationduringDNAsynthesis(PanelC)Variations
betweenpercentageofcellsontheplatesinpanelbandcarelikelyduetodifferencesinassays(visualcountingvs.countingbycolorimetry).For
p27/kip1expression(paneld,e),GCPsisolatedfromratcerebellum(P4)wereelectroporatedwiththeGPR3/EGFPexpression(bottomrow,paneld)or
controlvector(toprow,paneld).p27/kip1expressionwasdetectedbyimmunohistochemistry,48hourslater.EGFPandp27/kip1doublypositive
cellswerecountedinGPR3andcontrolgroups(panele).*p,0.01;**p,0.001,#p,0.0001,scalebar=200mm.
doi:10.1371/journal.pone.0005922.g001
PLoSONE | www.plosone.org 4 June2009 | Volume 4 | Issue 6 | e5922
GPR3InhibitsGPProliferation
Figure 2. Endogenous GPR3 siRNA-mediated inhibition enhances Shh-induced rodent GCPs proliferation and inhibits p27/kip
expression.GCPswereisolatedfromratcerebelli(P7)(n=12–14toobtain108neurons).Cells(56106)wereelectroporatedwithGPR3(middlerow,
panela)orcontrolsiRNA(toprow,panela).Tovisualizetransfectedcells,anEGFPexpressionvectorwasalsoco-transfectedwitheachsiRNA.Shh
(1mg/ml)wasadministrated6hoursafter.BrdUwasadded12hourspriortofixation.dbcAMP(100mM)wasusedasachemicalcAMPactivatorand
asapositivecontrol(bottomrow,panela).Inpanela,representativefieldsfromdishesarevisualizedfromthecontrolsiRNA+Shhgroup(toprow),
GPR3siRNA+Shhgroup(middlerow)andGPR3siRNA+Shh+dbcAMP(bottomrow).VisualizationbyBrdUimmunohistochemistry(reddye),andEGFP
fluorescence (green dye) and the combined merged image are shown. Five random fields were selected per each dish (n=3) from the control
siRNA+Shh,Shh+GPR3siRNAandShh+GPR3siRNA+cAMPgroups.ThenumberofGFP-positivecells(representingpositivelytransfectedneurons)was
counted(atleastonehundredGFP-positiveneuronspereachfield).Doublylabeledcells(yellowcolor)werealsoenumeratedpereachdish.Values
werethenexpressedasthepercentageofdoubly-labeledcellsoutoftotalGFPpositivecellsintheShh+GPR3grouporShh+GPR3+dbcAMPgroups
compared to Shh group, i.e. % Shh-treated cells (panel b). In parallel, cell proliferation was also determined using an ELISA assay based on the
measurementofBrdUincorporationduringDNAsynthesis(PanelC).Forp27/kipexpression,GPR3siRNAorcontrolsiRNAwereelectroporatedinP4
PLoSONE | www.plosone.org 5 June2009 | Volume 4 | Issue 6 | e5922
GPR3InhibitsGPProliferation
GCPs(paneld).ThemRNAofGPR3andp27/kip1wasanalyzedbyrealtimePCR,48hoursaftertransfectionandeachvaluewasnormalizedtobeta-
actinexpression.TheexpressionofGPR3wasreducedtoapproximate40%ofthecontrolsiRNAvalueandthisinhibitedp27/kip1mRNAexpression
by20%.#p,0.0001;*p,0.001,scalebar=200mm
doi:10.1371/journal.pone.0005922.g002
GPR3expressionindown-regulatingtheproliferativeeffectofShh Anti-proliferative effect of GPR3 during postnatal
and up-regulatingp27/Kip gene expressionin CGPs. cerebellar development
To determine if GPR3 gene expression was associated with an
GPR3 is expressed during postnatal cerebellar anti-proliferativeeffectinthepostnatalcerebellum,weutilizedthe
development 5-bromo-29 deoxyuridine (BrdU) pulse chase labeling technique.
The distribution and expression of mouse GPR3 in the central When BrdU was injected 4hours prior to the sacrifice of P14
nervous system has been previously reported by us[40]. To mice,thenumberofBrdU-positivecellswasincreasedintheIGL
determine whether GPR3 also plays a role in the developing of GPR32/2 versus wild-type type (Figure 4a). Quantitatively,
cerebelluminvivo,wefirstsoughttodetermineifGPR3mRNAwas thisincreasewassignificantinallIGLareasofcerebellumexcept
expressed postnatally (P1, P7, and P21) in the adult cerebellum. for 10Cb (Table 1). Interestingly, there was no significant
Quantitative RT-PCR showed up-regulation of GPR3 mRNA in difference in the number of BrdU-positive cells in the molecular
rodent cerebellum from birth until adulthood (Fig. 3A). To layer(ML)orwhitematter(WM)ofthecerebellumbetweenwild
determinethepatternanddistributionofGPR3promoteractivity typeandGPR32/2mice.Similarresultswerealsoobtainedusing
in the developing cerebellum, a genetic model using GPR32/2; P12 GPR32/2 orwild typemice (datanot shown).
LacZ+/+ mice was employed. In these mice, the GPR3 gene is To provide further confirmation of increased proliferation in
genetically substituted by a LacZ gene under control of the GPR32/2mice,weemployedimmunohistochemicalstainingfor
endogenous GPR3 promoter. Figure 3B reveals that LacZ gene the proliferative markers, Ki67 and phospho-Histone H3.
expression and, thus, GPR3 gene promoter activity was gradually Figures 4B and Table 2 show that the number of Ki67-positive
up-regulated post-natally almost exclusively in the internal cellswasincreasedintheIGLofGPR3knockoutmicecompared
granular layer (IGL). A very small number of cells displayed with wild type mice at P14. A similar finding was also observed
GPR3promoteractivityinthemolecularlayer(ML).Theseresults with phospho-Histone H3 immunohistochemistry (data not
thus indicated that GPR3 gene expression was up-regulated shown). To determine the lineage of Ki67- positive cells, double
postnatally and that GPR3 promoter activity was observed to immunohistochemical staining was performed using a neuronal
increase primarily intheIGLafter birth. (btubulin) marker.Ki67-positivecellsco-localized withbtubulin -
Figure3.GPR3expressionduringpostnatalratandmousecerebellardevelopment.Panela:TotalmRNAwasextractedfromratcerebelli
(N=12–15pertimepoint)at4differentdevelopmentalstages(P1,P7,P21)andadult(7to8weeks).RNAsamplesweresubjectedtoquantitativeRT-
PCRanalysisusingprobesandprimersspecifictoratGPR3.Datawereadjustedusingb-actinmRNAascontrol.Panelb:Todeterminethedistribution
ofGPR3inthedevelopingpostnatalcerebellum,aGPR32/2;LacZ+/+mousewasemployed,wheretheE.coliLacZgenewassubstitutedintothe
GPR3locusandwasthusundertranscriptionalcontroloftheendogenousGPR3promoter.Stainingforb-galactosidaseexpressionrevealedincreased
activation of transcriptional activity of the GPR3 promoter in the internal granular layer of cerebellum from P5 to P12 stages of postnatal
development.Scalebar=100mm.
doi:10.1371/journal.pone.0005922.g003
PLoSONE | www.plosone.org 6 June2009 | Volume 4 | Issue 6 | e5922
GPR3InhibitsGPProliferation
Figure4.IncreasedproliferationofgranuleprecursorcellsintheIGLofaGPR32/2mouse.Inpanela,BrdUwasadministeredtoP14
GPR32/2(rightpanel)orwildtypemice(leftpanel)4hourspriortosacrifice.Sections(20mm)wereincubatedwitharatanti-BrdUmonoclonal
antibodyandFITC-labelledanti-ratsecondaryantibody.FluorescentcellsareindicatedbyarrowsintheIGLofGPR32/2micebutnotinwild-type
mice.Asexpected,meningealcellsoncerebellarsurfacealsoshowedBrdUincorporation.Inpanelb,parallelsectionswerestainedwithananti-ki67
antibodytodetectproliferatingcellsinIGL.PositiveKi67cellswerevisualizedinIGLofGPR32/2vs.wild-typemice.Inpanelc,mergedimagesof
Ki67positivecellsandbtubulinpositiveareshownintherightsubpanel.Whitearrowheadspointtodoublypositivecells.Scalebar=100mm(panel
a,b)and20mm(panelc).
doi:10.1371/journal.pone.0005922.g004
positive cells (Fig. 4c). These findings thus confirmed that GPR3 In order to determine if proliferating neurons in the IGL of
gene expression was associated with a neuronal antiproliferative GPR32/2 mice led to an anatomic difference, BrdU was
effect intheIGLof thepostnatal mouse cerebellum. administered to P5 mice that were then sacrificed 13 days later.
PLoSONE | www.plosone.org 7 June2009 | Volume 4 | Issue 6 | e5922
GPR3InhibitsGPProliferation
Table1. BrdUpositivecells percerebellar lobe.
2Cb 3Cb 4&5Cb 6&7Cb 8Cb 9Cb 10Cb Sum
Wildtype 4.460.91 6.561.0 13.861.4 15.461.9 5.260.72 8.160.79 3.760.59 57.063.4
GPR32/2 8.261.1 10.761.8* 19.362.5* 25.862.0* 11.661.4* 16.761.8** 3.260.56 95.468.0*
*p,0.05.
**p,0.001(Mean6SE).
doi:10.1371/journal.pone.0005922.t001
At P18, the number of BrdU positive cells in the IGL layer of G1 phase increased by approximately 10% and that in S phase
GPR32/2micewasgreaterthanthatinwildtypemice(Fig.5a). decreased by an equal amount in GPR3-transfected DAOY cells
This difference was significant upon quantitation (Fig. 5b). The (Table 3). These findings thus showed that GPR3 functioned to
continuedneuronalproliferationdidresultinanatomicalthicken- promote G1 arrest during thecellcycle.
ing of the IGL layer, both visually (Fig. 5c) and
quantitatively.(Fig. 5d). These results thus confirmed that the Discussion
continued neuronal proliferation inthe cerebellum of GPR32/2
GPR3 is a member of a family of G-protein couple receptors
mice did result inanatomical wideningof theIGLlayer.
whose activation of PKA and subsequent increase of cyclic AMP
level promotes meiotic arrest in the oocyte[38]. Mice deficient in
In vivo expression of p27/Kip1 is inhibited in GPR32/2
GPR3[38] display premature ovarian aging and loss of fertili-
mice during postnatal development
ty[42]. This report also reports another abnormality related to
To provide additional confirmatory evidence for a regulatory
increased proliferation of CGPs in IGL of cerebellum. Our
relationshipbetweenGPR3andp27/Kip,wesoughttodetermine
previously published finding that GPR3 was highly expressed in
the expression of p27/Kip1 in wild type or GPR32/2 mice
theinternalgranulelayer(IGL)oftherodentpostnatalcerebellum
during postnatal cerebellar development (P5, P8 and P14). The
ledustohypothesizethatGPR3wasoneofthefactorsresponsible
expression of p27/Kip1 in wild-type P5 mice was scarce and for inhibition of granule cell precursors’ proliferation and
observedprimarilyinIGLinP5cerebellum(figure6A).However, induction of their terminal differentiation during develop-
thisexpression visiblyincreasedat theP8andP14timepointsin ment[40]. In this report, we show for the first time that GPR3
IGL,butalsointheinnersideofEGLandinmigratingGCPsin expression: 1) decreases Shh-induced proliferation and increases
ML. In contrast, in GPR32/2 mice theexpression of p27/Kip1 p27/Kip expression of GPCs when added exogenously, while it
was visibly reduced in the EGL and IGL of P5, P8 and P14 promotestheoppositeeffectwhen‘‘knocked-down’’bysiRNA;2)
cerebella compared to that of wild-type mice at the same time gradually increases postnatally in IGL and, when absent, there is
points (figure 6a). These data thus indicate that GPR3 gene increased proliferation of cells in IGL with a visible neuroana-
expression was involved in the regulation of p27/Kip gene tomicalabnormality;and3)islinkedinvivotoinductionofp27/
expression inthedeveloping postnatal mousecerebellum. Kip,previouslylinkedtoCGPcellcycleexit,andtoanincreasein
Since a downstream effect of elevated cAMP in neuronal cells phosphorylated CREB, suggesting that its action is mediated by
consists of phosporylation of the CREB transcription factor, we cyclic AMP. Therefore, these results indicate that GPR3
also determined whether levels of phosphorylated CREB were represents one of the molecules responsible for the regulation of
decreasedintheGPR32/2mice.Figure6bshowsthatphospho- the antiproliferative and differentiation program in the postnatal
CREB expression could be detected in the IGL and ML of P12 cerebellum.
wild-type mice. However, this expression was visibly reduced in Invitro,increasesofGPR3geneexpressionbyexogenousmeans
GPR32/2 mice. In addition, p27/Kip and phosphoCREB did not affect proliferation of postnatal GPCs suggesting that by
expression were observed to co-localize in cerebellar granule itselftheantiproliferativeactionofthegenewasalreadyatitspeak.
neurons in the IGL and ML , while there was little or no However, addition of more GPR3 did significantly inhibit the
expression of either in the IGL and ML of GPR32/2 mice proliferative action of Shh. Although this inhibition seems to be
(figure6c).Takeninconjunction,thisdatathusprovidedevidence onlypartial,electroporationexperimentsmaybeunder-estimating
for a cAMP-mediated regulatory effect of GPR3 expression on itstruemagnitude.Firstofall,theefficiencyofelectroporationisat
p27/kip expression inthedeveloping cerebellum. bestabout50%ofcellsontheplate,whiletheeffectofShh,orof
Finally to provide further evidence for effects of GPR3 on cell thepositivecontroldibutyrylcyclicAMP,islikelytobeonallcells
cyclekinetics,DAOYmedulloblastomacellsweretransfectedwith on the plate. Second, more detailed dose-response assays could
controlorGPR3expressingvector.ThepopulationofcellsinG0/ havebeenperformedtodeterminethestoichiometryofGPR3gene
Table2. Ki67positivecellsper cerebellar lobe.
2Cb 3Cb 4&5Cb 6&7Cb 8Cb 9Cb 10Cb Sum
Wildtype 7.060.55 11.461.9 12.061.2 5.860.58 5.860.58 13.662.8 3.260.97 58.865.4
GPR32/2 11.861.5* 21.862.5** 27.662.2** 17.860.97** 8.261.7 27.462.3** 7.861.5 122.468.9**
*p,0.05.
**p,0.001(Mean6SE).
doi:10.1371/journal.pone.0005922.t002
PLoSONE | www.plosone.org 8 June2009 | Volume 4 | Issue 6 | e5922
GPR3InhibitsGPProliferation
Figure 5. The IGL of GPR32/2 mice is hyperproliferative and enlarged. Wild-type and GPR32/2 mice were administered BrdU at P5.
Thirteendayslater,miceweresacrificedandcerebelliwerestainedwithaBrdUantibody.Inpanela,BrdUstainingisshownprimarilyinIGLofwild-
typeand,evenmoreso,ofGPR32/2mice.Inpanelb,BrdUpositivecellsintheIGLwereenumeratedin9Cblobeofwild-typevs.GPR32/2mice.In
panelc,IGLthicknesswasvisuallyevaluatedinwild-typevs.GPR32/2mice(P14)afterhematoxilynandeosinstain.ThicknessoftheIGLin9Cblobe
(P14)wasmeasuredinwild-typevs.GPR32/2mice(P14).*=p,0.05;**=p,0.01.Scalebar=100mm.
doi:10.1371/journal.pone.0005922.g005
expressioninhibitionofShh-inducedproliferation.Inspiteofthis, PACAP) and signaling pathways are also required to completely
theexperimentperformedstillshowedsignificantinhibitionofShh abolish GPR3 action. Finally, experiments in vitro may not
actionbyGPR3,thusansweringthepositedquestion.Third,itis faithfully recapitulate in vivo events, due to artifacts of tissue
likely that GPR3 is not the only mediator of inhibition of Shh- culture. Similar considerations apply to the siRNA experiments.
inducedactionandthatotherfactors(forinstance,otherGPRsor To really determine if GPR3 possesses a significant role in
PLoSONE | www.plosone.org 9 June2009 | Volume 4 | Issue 6 | e5922
GPR3InhibitsGPProliferation
Figure6. Expression of p27/kip andphosphoCREB expression duringpostnatalcerebellar development. In panel a, cerebelli from
GPR32/2orwildtypemicewereharvestedatP5,P8,andP14.p27/kip1immunohistochemistrywasdetectedinIGL(aswellasotherlayers)butwas
visuallymorereadilyapparentinwild-typevs.GPR32/2miceatallstages.Inpanelb,pCREBwasdetectedbyimmunohistochemistryinwild-type
andGPR32/2mice(P12).WhilepCREB,wasdetectedinthemolecularlayer(ML)andIGLofwidl-typemice,itwasnotasreadilyvisibleintheIGLof
theGPR32/2.Inpanelc,co-localizationofpCREBandp27/kipwasdeterminedbydoubleimmunofluorescence.Scalebar=50mm.
doi:10.1371/journal.pone.0005922.g006
PLoSONE | www.plosone.org 10 June2009 | Volume 4 | Issue 6 | e5922