Itraconazole, a Commonly Used Antifungal that Inhibits Hedgehog Pathway Activity and Cancer ...

12 Pages·2010·1.62 MB·English

Cancer Cell
Article
Itraconazole, a Commonly Used
Antifungal that Inhibits Hedgehog
Pathway Activity and Cancer Growth
JamesKim,1JeanY.Tang,2,9RuoyuGong,4JynhoKim,1,4JohnJ.Lee,1KarlV.Clemons,5,6CurtisR.Chong,7,8
KrisS.Chang,9MarkFereshteh,10DaleGardner,11TannishthaReya,10JunO.Liu,7,8ErvinH.Epstein,9DavidA.Stevens,5,6
andPhilipA.Beachy1,3,4,*
1DepartmentofDevelopmentalBiology
2DepartmentofDermatology
3InstituteforStemCellandRegenerativeMedicine
4HowardHughesMedicalInstitute
5DivisionofInfectiousDiseasesandGeographicMedicine
StanfordUniversity,Stanford,CA94305,USA
6DepartmentofMedicine,SantaClaraValleyMedicalCenterandCaliforniaInstituteforMedicalResearch,SanJose,CA95128,USA
7DepartmentofPharmacologyandMolecularSciences
8TheJohnsHopkinsClinicalCompoundScreeningInitiative
TheJohnsHopkinsUniversitySchoolofMedicine,Baltimore,MD21205,USA
9Children’sHospitalOaklandResearchInstitute,Oakland,CA94609,USA
10DepartmentofPharmacologyandCancerBiology,DukeUniversityMedicalCenter,Durham,NC27710,USA
11USDAARS,PoisonousPlantResearchLaboratory,Logan,UT84341,USA
*Correspondence:[email protected]
DOI10.1016/j.ccr.2010.02.027
SUMMARY
Inascreenofdrugspreviouslytestedinhumansweidentifieditraconazole,asystemicantifungal,asapotent
antagonistoftheHedgehog(Hh)signalingpathwaythatactsbyamechanismdistinctfromitsinhibitoryeffect
on fungal sterol biosynthesis. Systemically administered itraconazole, like other Hh pathway antagonists,
can suppress Hh pathway activity and the growth of medulloblastoma in a mouse allograft model and
does so at serum levels comparable to those in patients undergoing antifungal therapy. Mechanistically,
itraconazole appears to act on the essential Hh pathway component Smoothened (SMO) by a mechanism
distinctfromthatofcyclopamineandotherknownSMOantagonists,andpreventstheciliaryaccumulation
ofSMOnormallycausedbyHhstimulation.
INTRODUCTION Response to the Hh protein signal is governed by Patched
(PTCH), a 12-pass transmembrane protein that restrains the
The Hh signaling pathway is critical for embryonic growth and activity of Smoothened (SMO), a member of the 7-transmem-
patterninginorganismsrangingfrominsectstomammals(Varjo- brane family of serpentine receptors (Figure 1A). Hh protein,
salo and Taipale, 2008). Hh signaling also functions postem- when present (Taipale and Beachy, 2001), binds PTCH and
bryonically in tissue homeostasis through effects on stem or blocks its inhibition of SMO, thus permitting accumulation of
progenitor cells (Beachy et al., 2004). Inappropriate activity of SMO in the primary cilium (Corbit et al., 2005; Rohatgi et al.,
the Hh signaling pathway also has been linked totumor types 2007), and causing activation of the Gli family of transcription
thatarisesporadicallyoringeneticallypredisposedindividuals factors. Pathway activation via SMO thus can occur either by
(VarjosaloandTaipale,2008;Yauchetal.,2008). Hh protein stimulation or through loss of PTCH activity, as
Significance
TheinappropriateactivityoftheHhsignalingpathwaynotedinavarietyoftumortypeshaspromptedeffortstodevelop
small-molecule antagonists of Hh pathway signaling. Such drug development efforts, however, are costly and time-
consuming.Withtheidentificationofitraconazole,acommonlyusedsystemicantifungal,asapotentinhibitoroftheHh
pathwayinvitroandinvivo,itmaybepossibletoavoidsomeoftheexpenseanddelayassociatedwithdenovodevelop-
mentofHhpathwayantagonists.SupportingtherapidentryofitraconazoleintoclinicaltrialsinthetreatmentofHh-depen-
dentcancers,serumlevelsthatinhibitgrowthoftumorsinmicearecomparabletothosereportedinhumans.
388 CancerCell17,388–399,April13,2010ª2010ElsevierInc.
Cancer Cell
AntifungalItraconazoleAntagonizesHhSignaling
seeninsporadiccancersorthosethatariseinthefamilialcancer IC morethan10-foldlowerthanthatofketoconazole,itsnear-
50
predisposition syndrome, BCNS (basal cell nevus syndrome, estrival(IC z9mM;seeFigure1DandTableS1).Hydroxy-itra-
50
associated with heterozygous mutation of the human PTCH conazole, the major metabolite of itraconazole in humans, is
gene). equally potent in its antifungal activity (Harvey et al., 1980;
Cyclopamine and other small molecules that antagonize Hh Hostetler et al., 1993), and also inhibited Hh pathway activity
pathwayactivity(Chenetal.,2002a;Chenetal.,2002b;Cooper (IC z1.2mM;seeFigure1E).Interestingly,fluconazole,ananti-
50
et al., 1998; Frank-Kamenetsky et al., 2002; Incardona et al., fungal with similar potency for 14LDM as itraconazole (Lamb
1998;Taipaleetal.,2000)havebeenfoundtoactpredominantly, etal.,1999;Troskenetal.,2006),showsnodetectableinhibition
although not exclusively, by binding the essential pathway ofHhpathwayactivity,evenatalevel50-foldhigherthantheIC
50
component SMO. These small molecules have been effective ofitraconazole.Wethusconcludethatazoleantifungaldrugsas
inblockingHhpathway-dependentgrowthoftransformedcells, ageneralclassarenotinhibitorsofHhpathwayactivity.
bothinvitro(Taipaleetal.,2000)andinvivo(Bermanetal.,2002; The inability of fluconazole to inhibit Hh pathway activity,
Dierksetal.,2008;Romeretal.,2004;Yauchetal.,2008;Zhao despiteasimilarpotencyforinhibitionof14LDMasitraconazole,
et al., 2009), thus stimulating major efforts to develop small- suggeststhattheeffectofitraconazoleisnotduetoaneffecton
moleculeantagonistsoftheHhpathwayascancertherapeutics. cellularcholesterolbiosynthesis.Curiously,however,wenoted
However, drug development is time-consuming and costly thatincreasingserumlevelsinourassaysblockedpathwayinhi-
(DiMasietal.,2003;Frank,2003),andwesoughttocircumvent bition by itraconazole (Figure 2A). This blocking effect was
this delay and expense by identifying Hh pathway antagonists absent in lipid-depleted serum (Figure 2B), suggesting that a
among drugs that have been tested for toxicity in humans factorremovedbylipiddepletionofserumcaninterferewithor
or even approved for human use by the US Food and Drug reversetheactionofitraconazole.Asimilarinterferingeffectof
Administration(FDA). serum but not lipid-depleted serum was noted with pathway
activation by the small-molecule agonist SAG (Chen et al.,
RESULTSANDDISCUSSION 2002b)(FiguresS2AandS2B).
Toinvestigatethisserumeffectfurther,wedirectlytestedthe
Wescreenedalibraryof(cid:2)2400FDA-approvedorpost-phaseI effectsofcholesterolanditsbiosyntheticprecursorslathosterol
drugs (Chong et al., 2006a; Chong et al., 2006b) (now part of and desmosterol (see Figure 1C). In cultured cell signaling
the Johns Hopkins Clinical Compound Library) for activity in assays, our laboratory has shown that statins can inhibit Hh
inhibition of Hh signaling. This screen made use of a reporter signalresponse,butonlyiftransientcyclodextrinpretreatment
cell line, Shh-Light2 (Taipale et al., 2000) (see Experimental isfirstusedtodepletecellularsterols(Cooperetal.,2003);itra-
Procedures), which contains a stably integrated Gli-luciferase conazoleincontrastdoesnotrequirecyclodextrinpretreatment
reporterthatrespondstostimulationbySHHN,theactiveform toinhibittheHhpathway.Furthermore,althoughtheadditionof
oftheSonichedgehog(SHH)signalingprotein. cholesteroltocyclodextrinpretreatedculturesinlipid-depleted
Although several dozen hits were identified from preliminary serumreversedtheinhibitoryeffectsofstatinsonHhresponse
screening,onlyafewwereactiveinpathwayinhibitionatcon- (Figure2C)andcellproliferation(FiguresS2C–S2E),theinhibi-
centrationsachievedinhumans.Amongthesethemicrotubule toryeffectofitraconazolewasnotreversedbysupplementation
inhibitors,includingvincaalkaloidsandtheirderivatives,consis- ofcultureswithupto25mMlathosterolordesmosterol(Figures
tentlyactedasHhpathwayantagonists(seeFigureS1available 2D and 2E) or 250 mM cholesterol as an ethanolic solution or
online).Theircytotoxicnature,however,precludesanextended in water-soluble form (Figures S2F and S2G). These results
and continuous course of treatment, as may be required further reinforce the conclusion that itraconazole inhibition of
for therapy of cancers by antagonism of Hh pathway activity Hhresponseisnotmediatedthroughinhibitoryeffectsonsterol
(Berman et al., 2002; Romer et al., 2004; Yauch et al., 2008). biosynthesis.
A more promising candidate, itraconazole, was identified as We next considered whether specific lipoprotein particles in
apotentinhibitorofHhpathwayactivitywithanIC ofapprox- serummightaccountforthereversalofitraconazoleactionon
50
imately800nM(Figure1B).Thisantifungaliscommonlyusedin the Hh pathway, and found that low-density lipoprotein (LDL)
clinical practice as an orally delivered systemic treatment that butnothigh-densitylipoprotein orvery low-densitylipoprotein
canbesustainedformonths(Stevens,2001). particlesdramaticallyinterferedwithitraconazoleactiononHh
Itraconazoleinhibits14-a-lanosteroldemethylase(14LDM),by pathway activity (Figures 3A–3C). The megalin member of the
amechanisminwhichthetriazolegroupofitraconazolecoordi- LDL receptor family (Fisher and Howie, 2006; Hussain et al.,
nates the heme Fe2+ in the cytochrome P-450 active site of 1999)hasbeenshowntobindandendocytoseSHH(McCarthy
14LDM(GeorgopapadakouandWalsh,1996).14LDMiscritical et al., 2002), suggesting the possibility that LDL may interfere
for the biosynthesis of ergosterol in fungi and cholesterol in withitraconazolebyaffectingreceptionoftheHhsignal.Thisis
mammals(LepeshevaandWaterman,2007)(Figure1C).Itraco- unlikely,however,becauseitraconazoleactsbelowthelevelof
nazoleissignificantlymorepotentininhibitingthefungalenzyme the Hh receptor PTCH (see below); LDL furthermore interferes
(Lambetal.,1999;Troskenetal.,2006)buthasbeenshownto withitraconazoleactionevenincellslackingPTCH(Figure3D).
decrease human cholesterol levels at high doses (Schneider Wethusconcludethattheinterferingeffectofserumisnotdue
etal.,2007).Totestwhetherazoleantifungalsingeneralactto to reversal of a sterol biosynthetic defect, nor through a LDL-
inhibit Hh pathway activity, we examined five other imidazole mediated effect on Hh reception, but more likely through
and triazole antifungal drugs. We found that itraconazole was sequestrationofitraconazolebyLDLlipoproteinparticles,which
byfarthemostpotentinhibitorofHhpathwayactivity,withan have been shown to sequester and interfere with the action
CancerCell17,388–399,April13,2010ª2010ElsevierInc. 389
Cancer Cell
AntifungalItraconazoleAntagonizesHhSignaling
Figure1. ItraconazoleInhibitsHhSignaling
(A)AschematicviewoftheHedgehog(Hh)signalingpathway.IntheabsenceofHh,Patched(PTCH)suppressesSmoothened(SMO)function.Hh,whenpresent,
bindstoandinhibitsPTCH,permittingSMOaccumulationintheprimarycilium(notshown)andcausingactivationofthepathwayviatheGlifamilyoftranscription
factors.PtchandGli1arethemselvestranscriptionaltargetsofthepathway.Oxysterols(dashedgreenbracket)actbetweenPTCHandSMO,aspathway
activators,whereasstatins(dashedredbracket)actdownstreamofPTCHandatorupstreamofSMO,aspathwayinhibitors.SAGandcyclopamineactivate
andinhibitthepathway,respectively,bybindingtothetransmembranedomainofSMO.Activatorsandinhibitorsofthepathwayarelabeledingreenand
red,respectively.
(B)Hhsignalingassays.Luciferasereporteractivityunderthecontrolofan8x-GlibindingsiteintheShh-Light2reportercelllinewasmeasureduponstimulation
withSHHN-containingmedium.ItraconazoleblockedHhpathwayactivity(IC50z800nM).
(C) Schematic view of mammalian cholesterol biosynthesis from 3-hydroxy-3-methyl-glutaryl-CoA (HMG-CoA). Statins inhibit HMG-CoA reductase
whereasazoleantifungaldrugsinhibit14a-lanosteroldemethylase(14LDM),asindicated.Lathosterolanddesmosterolarecholesterolprecursorsdownstream
of14LDM.
390 CancerCell17,388–399,April13,2010ª2010ElsevierInc.
Cancer Cell
AntifungalItraconazoleAntagonizesHhSignaling
Figure 2. Inhibition of Hh Signaling by
ItraconazoleIsNotMediatedbyEffectson
CholesterolBiosynthesis
(A) Increasing serum concentrations attenuated
the inhibitory activity of itraconazole but not
KAAD-cyclopamine.
(B)Theattenuatingeffectofserumwasabolished
bylipiddepletion(LD).
(C) Shh-Light2 cells were pretreated with
methyl-b-cyclodextrin (MbCD) 8 mM in DMEM
for45mintoremovesterolsfromthecellsurface
(Cooper et al., 2003). Lovastatin, an HMG-CoA
reductaseinhibitor,inhibitedHhpathwayactivity
induced by SHHN. Cholesterol, in ethanol
solution, reversed the pathway inhibitiondueto
lovastatin.
(DandE)Lathosterol,desmosterol,andcholes-
terol failed to rescue Hh pathway inhibition by
itraconazole under low (D) or lipid-depleted (E)
serumconditions.
All signaling assays were performed with Shh-
Light2cellsanddataaremeanoftriplicates±SD.
SeealsoFigureS2.
anddownstreamofPTCH.Incontrastto
Ptch(cid:3)/(cid:3)cells,theinhibitoryactionofitra-
conazole was bypassed by expression
ofSmoA1(Taipaleetal.,2000;Xieetal.,
1998), a constitutively active oncogenic
variant ofmurineSMO(Figure 4B),sug-
gesting that itraconazole does not act
downstream of SMO. To define further
the target of itraconazole action, we
combined itraconazole treatment with
pathwayactivationbyoverexpressionof
SMOorbytreatmentwiththeoxysterols
20(S)- and 22(S)-hydroxycholesterol
(CorcoranandScott,2006;Dwyeretal.,
2007). We found that itraconazole
blocked pathway activity produced by
SMOoverexpression(Figure4C).Itraco-
nazolealsocompletelyblockedoxysterol
action (IC z270 nM; Figure 4D), sug-
50
of other hydrophobic compounds such as benzo(a)pyrene gesting that oxysterols act on a target distinct from and
(Remsen and Shireman, 1981; Shu and Nichols, 1979) and upstreamofthetargetofitraconazole.
p-dimethylaminoazobenzene(Chenetal.,1979). The ability to suppress pathway activity from SMO overex-
WeinvestigatedthetargetofitraconazoleactionwithintheHh pression but not from SMOA1 suggests that itraconazole may
pathwaybyexaminingcellswiththePtchgenedisruptedbyin act by binding to SMO as an inverse agonist (Bond et al.,
frame fusion of lacZ (Goodrich et al., 1997; Taipale et al., 1995;CostaandHerz,1989);bindingtoSMOalsocouldaccount
2000).LossofPTCHfunctioninthesecellscausesconstitutive foritraconazoleeffectsinPtch(cid:3)/(cid:3)cells,orincellstreatedwith
pathwayactivityandconsequentexpressionofb-galactosidase oxysterols.Therefore,wealsoexaminedtheeffectofitracona-
throughactivationofthePtchpromoter.Itraconazoletreatment zole treatment on localization of SMO, which has been shown
of these cells inhibited b-galactosidase expression at levels toaccumulateintheprimaryciliumuponHhstimulation(Corbit
similar to those required for inhibition of pathway activity in et al., 2005; Rohatgi et al., 2007; Figure 4E). We found that
SHHN-stimulated Shh-Light2 cells (IC z900 nM;Figure 4A). itraconazole treatment dramatically reduced Shh-induced
50
These results indicate that itraconazole acts independently of accumulation of SMO in the primary cilium (Figure 4E), similar
(D)Amongtheazoleantifungals,itraconazolewasthemostpotentinhibitorofHhpathwayactivity.
(E)Hydroxy-itraconazole,themajormetaboliteofitraconazoleinmammals,alsoinhibitedtheHhpathway(IC z1.2mM).Allsignalingassayswereperformed
50
withShh-Light2cellsin0.5%serummediaanddataareshownasthemeanoftriplicatesandstandarddeviation(±SD).SeealsoFigureS1andTableS1.
CancerCell17,388–399,April13,2010ª2010ElsevierInc. 391
Cancer Cell
AntifungalItraconazoleAntagonizesHhSignaling
Figure 3. Low-Density Lipoprotein Modu-
lates the Hh Pathway Inhibitory Effects of
Itraconazole
(A and B) High-density lipoprotein (HDL, A)
100 mg/ml and very low-density lipoprotein
(VLDL,B)6.5mg/mlcannotreverseHhpathway
inhibitionbyitraconazole.
(CandD)Titrationoflow-densitylipoprotein(LDL)
up to 120 mg/ml rescues either induced Hh
pathway activity by SHHN ligand in Shh-Light2
cells (C) or constitutively active pathway in
Ptch(cid:3)/(cid:3) cells (Goodrich et al., 1997; Taipale
etal.,2000)(D)fromtheinhibitoryeffectsofitraco-
nazole.Allsignalingassayswereperformedinlipid
depleted 10% calf serum media for Shh-Light2
cells (A–C) and 0.5% fetal bovine serum media
forPtch(cid:3)/(cid:3)cells.
Datarepresentmeanoftriplicates±SD.
Althoughitraconazolefailstocompete
withBODIPY-cyclopamine,itcouldnev-
ertheless bind SMO at a distinct site
from cyclopamine. If so, itraconazole
conceivably could bind simultaneously
with cyclopamine or with other small
molecule agonists and antagonists that
bindatthecyclopaminesite.Itraconazole
thusmaysynergizewithcyclopamineor
otherSMOantagonistsinpathwayinhibi-
tionanditmayactnoncompetitivelywith
SMO agonists that bind at the cyclop-
amine site. Indeed, we found that the
IC of KAAD-cyclopamine decreased
50
from20nMto2nMwithincreasingdoses
ofitraconazole(Figure5BandTableS2),
and that the IC of itraconazole de-
50
creases from 800 nM to 130 nM with
tootherSMOantagonists,SANT-1andSANT-2,butincontrast increasing doses of KAAD-cyclopamine (Figure 5C and Table
to cyclopamine, which induced accumulation of SMO in the S3). In addition, the maximal degree of pathway activation by
primarycilium(Rohatgietal.,2009;Wangetal.,2009).Theability SAG, whose SMO binding site overlaps that of cyclopamine,
ofitraconazoletosuppressSHH-inducedaccumulationofSMO was reduced by increasing concentrations of itraconazole
intheprimaryciliumisconsistentwithdirectactionofitracona- without altering the SAG EC (Figure 5D and Table S4), indi-
50
zoleonSMOforpathwayinactivation. cating that itraconazole acts as a non-competitive inhibitor
Like itraconazole, cyclopamine inhibits pathway activity in (Taylor and Insel, 1990). Itraconazole, in summary, behaves
Ptch(cid:3)/(cid:3)cells(Taipaleetal.,2000),incellstreatedwithoxysterols like an inverse agonist, suppresses the accumulation of SMO
(Dwyer et al., 2007), and in cells overexpressing SMO (Taipale intheprimarycilium,mutuallyenhancesthepotencyofpathway
et al., 2000), but not in SMOA1-expressing cells (Taipale et al., inhibition in combination with cyclopamine, and acts as
2000). As proposed above for itraconazole, cyclopamine and anoncompetitiveinhibitorofpathwayactivationbySAG.These
many other antagonists act by binding to SMO (Chen et al., characteristicsofitraconazoleareconsistentwiththebindingof
2002a;Chenetal.,2002b;Frank-Kamenetsky etal.,2002).We itraconazoletoSMOatadistinctsitefromthatofcyclopamine.
thereforetesteditraconazoleforSMObindingusingBODIPY-cy- Giventhatwehavenotdemonstrateddirectbindingofitracona-
clopamine,a fluorescentcyclopaminederivative,andacellline zoletoSMO,thepossibilityremainsthatitraconazolemayexert
thatexpressesSMOuponinductionbytreatmentwithtetracycline itsinhibitoryactiononSMOthroughotherindirectmechanisms.
(Dwyeretal.,2007).WefoundthatBODIPY-cyclopaminebinding HavingcharacterizedtheHhpathwayinhibitoryactivityofitra-
to SMO was not affected significantly by itraconazole, even at conazole in vitro, we examined its effect in vivo by systemic
10mM,aconcentration12.5-foldhigherthanitsIC (Figure5A). treatment of animals subcutaneously engrafted with a primary
50
This contrasts sharply with the potent cyclopamine derivative, medulloblastoma froma Ptch+/(cid:3);p53(cid:3)/(cid:3)mouse. Medulloblas-
KAAD-cyclopamine(IC =20nM),whichreducedboundBOD- tomas from these mice depend on constitutive Hh pathway
50
IPY-cyclopaminetonear-backgroundlevelsat200nM(Figure5A). activity for their growth (Berman et al., 2002; Goodrich et al.,
392 CancerCell17,388–399,April13,2010ª2010ElsevierInc.
Cancer Cell
AntifungalItraconazoleAntagonizesHhSignaling
Figure 4. Mapping of Itraconazole Action
withintheHhPathway
(A)ItraconazoleinhibitedconstitutiveHhpathway
activityinPtch(cid:3)/(cid:3)cells,measuredasexpression
of b-galactosidase activity from the Ptch-lacZ
locus(Goodrichetal.,1997;Taipaleetal.,2000).
The levels of b-galactosidase activity from cells
treatedwithKAAD-cyclopamineat40and200nM
(2x and 10x IC50) are shown as purple and red
lines,respectively.
(B) Shh-Light SmoA1 cells, stably expressing a
constitutivelyactiveSMOmutant,wereresistant
toHhpathwayinhibitionbyitraconazole.Cyclop-
amine and forskolin (orange and blue lines,
respectively)wereusedasnegativeandpositive
controlsfortheinhibitionofSMOA1.
(C)ItraconazoleandKAAD-cyclopaminebothin-
hibitedconstitutiveHhpathwayactivationinduced
bytransient overexpressionofSMOinNIH 3T3
cells.
(D) Itraconazole inhibited Hh pathway activity
inducedbythecombinedoxysterols,20(S)-and
22(S)-hydroxycholesterol (5 mM each). All sig-
naling assays were performed with Shh-Light2
cellsin0.5%serummediaanddataareshown
asthemeanoftriplicates±SD.
(E)Representativeimmunofluorescentimagesof
NIHT-3T3cells.SMOaccumulatedinprimarycilia
upon stimulation with SHHN (arrowheads). This
accumulationwasblockedbytreatmentwithitra-
conazole.Insetsshowenlargedviewsofprimary
ciliainthedashedbox,withSMOandacetylated
tubulinchannelsoffset.Arrowsshowprimarycilia
withnoSMOaccumulation.Asseeninthehisto-
gram, the accumulation of SMO in (cid:2)90% of
primary cilia of cells stimulated with SHHN, is
reduced to (cid:2)5% in the additional presence of
itraconazole. The number of primary cilia con-
tainingSMOoverthetotalnumberofprimarycilia
areshownaboveeachbar.
Datarepresentmeanoftenfields±SD.Thescale
barrepresents5mm.
1997; Romer et al., 2004) and this growth can be inhibited by itraconazole and hydroxy-itraconazole (Hostetler et al., 1993).
pathwayantagonists(Bermanetal.,2002;Romeretal.,2004). Itraconazole concentrations reported here therefore represent
Treatment with itraconazole indeed inhibited growth of these acombinedmeasurementofitraconazoleandhydroxy-itracona-
subcutaneousallografts,withsimilarinhibitionoftumorgrowth zole effects. Serum itraconazole concentrations were similar
notedforanimalstreatedwithitraconazoleat100mg/kgtwice betweenthe100mg/kgand75mg/kggroups(Figure6B).Itraco-
per day (b.i.d.) or 75 mg/kg b.i.d. (Figure 6A). Consistent with nazolelevelsachievedintumors(Figure6C)weresimilartodrug
the synergistic effect of itraconazole and cyclopamine in vitro, levels in serum, indicating that itraconazole has good tumor
combined treatment with itraconazole and cyclopamine pro- tissue penetration. The area-under-the curve (AUC) values for
ducedanevengreaterinhibitionoftumorgrowth.Micebearing serum concentrations of itraconazole reached during a 24 hr
control tumors were euthanized after 18 days as the tumors pharmacokinetic study (Figure 6D) were also similar for
exceeded the size limits set forth by our Institutional Animal itraconazolegivenat100mg/kgand75mg/kg(220mg$hr/ml
CareandUseCommittee. and238mg$hr/ml,respectively).Thesefindingsareconsistent
At the end of the experiment, 4 hr after the last dose, we with previously published pharmacokinetic results (Hostetler
obtainedserumandtumorsfromthreemiceineachtreatment etal.,1992).
group. Itraconazole concentrations were measured using a We also tested the ability of itraconazole to inhibit pathway
bioassay method based on antifungal activity (Hostetler et al., activity in tumor tissue in a separate experiment that made
1992; Rex et al., 1991) that does not distinguish between useofmicebearingsecondarysubcutaneousmedulloblastoma
CancerCell17,388–399,April13,2010ª2010ElsevierInc. 393
Cancer Cell
AntifungalItraconazoleAntagonizesHhSignaling
Figure5. ItraconazoleActsataSiteDistinct
fromtheCyclopamine-BindingSiteofSMO
(A)BODIPY-cyclopamine(B-C)boundtotetracy-
cline-induced SMO in 293S cells as monitored
byFACS.Asnegativecontrols,thegraytracing
showsfluorescenceofcellsnottreatedwithB-C
andthebluetracing(bluearrow)showsuninduced
cells treated with B-C; the black tracing (black
arrow)represents maximalfluorescence ofcells
induced for SMO expression and treated with
B-C. Itraconazole treatment (no arrows) at con-
centrationsupto12.53IC50(10mM)didnoteffec-
tivelycompeteforB-Cbinding,whereasKAAD-
cyclopamine at 23 IC50 and 103 IC50 (40 nM
and200nM;tanandredarrowheads,respectively)
causedsignificantdecreasesinfluorescence.
(B and C) Synergistic action of itraconazole
andKAAD-cyclopamine.Insignalingassaysper-
formedwithShh-Light2cells,thepresenceofitra-
conazolecausedadownwardshiftintheIC of
50
KAAD-cyclopamine(from21nMto1.8nMinthe
presenceof900nMitraconazole;[B]andTable
S2);KAAD-cyclopaminecausedadownwardshift
intheIC ofitraconazole(from805to135nMin
50
the presence of 30 nM KAAD-cyclopamine; [C]
andTableS3).
(D)KAAD-cyclopamineisacompetitiveinhibitorof
SAG,aknownHhpathwayagonist,asshownby
shiftsoftheSAGEC50by(cid:2)5and(cid:2)13foldat40nMand100nMconcentrationsofKAAD-cyclopamine.Incontrast,itraconazoledidnotaltertheSAGEC50
butdecreasedthemaximalactivityofSAG,thusactingasanoncompetitiveinhibitor(seeTableS4).
AllsignalingassayswereperformedwithShh-Light2cellsin0.5%serummediaanddatarepresentmeanoftriplicates±SD.SeealsoTablesS2,S3,andS4.
allograftsfromthesameoriginaltumorasthatusedforstudies been reported in patients treated for severe fungal infections.
of tumor growth (see above). These mice were treated either This high-dose itraconazole therapy has ranged from 600 to
with 100mg/kg itraconazole b.i.d. or37.5 mg/kg cyclopamine 900mg/day(Sanchezetal.,1995;Sharkeyetal.,1991;Takagi
b.i.d.ThetreatedtumorsshoweddecreasedGli1mRNAlevels, etal.,2001)overtreatmentperiodsrangingfrom3to16months,
indicative of reduced Hh pathway activity as compared with withoccasionaltoxicitiesthatweremanageable(Sharkeyetal.,
cyclodextrinvehicle-treatedcontrols(Figure6E). 1991).Itisofinterestthatketoconazole,anotherazoleantifungal,
Tofurtherassesstheactivityofitraconazoleonsuppressionof has been used as second-line therapy for androgen-indepen-
Hh-dependent tumors, we used K14-CreER; Ptch+/(cid:3); p53 fl/fl dent metastatic prostate cancer (Scholz et al., 2005; Trump
mice to generate endogenous basal cell carcinomas (BCC) of et al., 1989) at doses 3- to 6-fold higher than its normal dose
theskin(seeExperimentalProcedures).BCCtumorsgenerated andwithtoxicitiesgreaterthanthoseforhigh-doseitraconazole
from these mice have increased levels of Hh pathway activity (Sugar etal.,1987). Theuse of ketoconazole in thiscontext is
andtheirgrowthissuppressedbyHhpathwayinhibitors(Xiao believedtoactbyinhibitingandrogenproduction,butwecannot
et al., 2009). Itraconazole treatment suppressed the growth of ruleoutthepossibilityofaneffectonHhpathwayactivity.
the endogenous BCC tumors when compared with the cyclo- Hhpathwayactivityhasbeencharacterizedascontributingto
dextrin vehicle control (Figure 7A). When itraconazole was growthofmanytumortypes(VarjosaloandTaipale,2008;Yauch
stoppedafter18daysoftreatment(Figure7A),thetumorsbegan et al., 2008). Recent studies in the mouse have shown that
togrowagainataratesimilartothatofvehicle-treatedcontrol anothercancer,chronicmyelogenousleukemia(CML),isdepen-
tumors. At the end of the study, tumors from itraconazole- dentonHhsignaling.SMOactivityisnecessaryformaintenance
treated mice appeared to have greater levels of necrosis as of CMLstem cellsand mouse and human CMLthat becomes
compared with tumors from vehicle-treated mice, based on resistant to tyrosine kinase inhibitor therapy remains sensitive
destruction of basophilic tumor cell nests and theeosinophilic tomanipulationofSMOactivity(Dierksetal.,2008;Zhaoetal.,
stroma surrounding these nests (Figure 7B). Tabulation of 2009).
necrosisscores,withahigherscoreindicatingworsenecrosis, The importance of the Hh pathway in tumor growth has led
confirmedthathistologicallyassessednecrosiswasgreaterfor to an active interest in the development of small-molecule Hh
itraconazole-treated BCC as compared with vehicle-treated pathwayantagonists(Tremblayetal.,2009).Mostofthesemole-
tumors(Figure7C). cules are either cyclopamine derivatives or they antagonize
Theserumconcentrationsofitraconazoleassociatedwithan pathway by binding SMO at the same site as cyclopamine.
antitumor effect in our in vivo studies are higher than those Recent clinical tests of one of these cyclopamine mimics,
reachedinhumansatregulartherapeuticdoses(Baroneetal., GDC-0449(Rudinetal.,2009;VonHoffetal.,2009),produced
1993) (i.e., 400 mg/day orally for 15 days). Similar high serum promising results, suggesting that Hh pathway inhibition may
levels of itraconazole (Sanchez et al., 1995), however, have represent an effective approach to cancer therapy. Further
394 CancerCell17,388–399,April13,2010ª2010ElsevierInc.
Cancer Cell
AntifungalItraconazoleAntagonizesHhSignaling
Figure 6. Itraconazole Inhibits the Growth
ofHh-DependentMedulloblastomaInVivo
(A) Itraconazole, either alone or in combination
withcyclopamine,suppressedthegrowthofallog-
rafted medulloblastomas from a Ptch+/(cid:3)p53(cid:3)/(cid:3)
mouse,ascomparedwithvehicle-treatedtumors.
Treatments were given twice per day by oral
gavage.Thenumberoftumorsusedinthisexper-
imentwere:Cyclodextrincontroln=8,itracona-
zole 100 mg/kg n = 12, itraconazole 75 mg/kg
n=10,cyclopamine37.5mg/kgn=10,itracona-
zole/cyclopaminen=10.Dataareexpressedas
mean percentage change in volume ±standard
errorofthemean(SEM).
(B and C) Mean serum and tumor itraconazole
levelsobtainedattheendoftheallograftstudyin
(A)areshowninpanels(B)and(C),respectively.
Thedashedlinesrepresenttheminimumdetect-
ablelevelsoftheassay;vehiclecontrolswereat
orbelowtheselevels.
(D)Pharmacokineticsofitraconazolelevelsinserum
ofathymicnudemiceafteroneoraldoseofitracona-
zole100mg/kgor75mg/kg.Allitraconazolelevels
wereobtainedusingthebioassaymethod(Clemons
etal.,2002;Harveyetal.,1980;Hostetleretal.,
1992;Rexetal.,1991).Dataforitraconazolelevels
in(B)–(D)representthemeanofthreesamples±SD.
(E) Hh pathway is decreased in allografted
medulloblastomatumorswhentreatedwitheither
itraconazole or cyclopamine as compared to
cyclodextrincontroltreatment.Hhpathwaywas
monitoredbyGli1mRNAlevels.Gli1mRNAlevels
ofdrugtreatmentswerenormalizedtothemean
ofGli1mRNAlevelsofcyclodextrincontrol.Treat-
mentsweregiventwiceperdayfor4daysbyoral
gavage.**p=0.0197;***p=0.0386.
EXPERIMENTALPROCEDURES
CellCultureAssaysforHhPathway
Activation
Shh-Light2cellline(Taipaleetal.,2000)isaclonal
NIH3T3celllinestablytransfectedwith83-Gli1-
bindingsitedependentfireflyluciferaseandRenilla
luciferasereporters.SmoA1Lightcelllineisaclonal
sublineofShh-Lightcellsstablytransfectedwith
oncogenicSmoA1(Taipaleetal.,2000).Shh-Light2
clinical testing will be required to establish efficacy and long-
cellsweremaintainedinDulbecco’smodifiedEagle’smedium(DMEM)contain-
termsafetyofsuchagents(DlugoszandTalpaz,2009).
ing10%(v/v)calfserum,penicillin,streptomycin,L-glutaminesupplemented
Itraconazole has been studied for nearly 25 years, and we withGeneticin(Invitrogen)400mg/ml,andZeocin(Invitrogen)100mg/ml.SmoA1-
thereforehaveagoodunderstandingofitssafetyandpotential LightcellsweremaintainedinDMEMcontaining10%(v/v)calfserum,penicillin,
sideeffects.Furthermore,itraconazoleisalreadyusedinhuman streptomycin, L-glutamine supplemented with Geneticin 400 mg/ml and
patientsatdosesthatproduceserumconcentrationslikethose Hygromycin (CalBiochem) 100 mg/ml. Ptch(cid:3)/(cid:3) MEF cell line (Taipale et al.,
2000)wasmaintainedinDMEMcontaining10%fetalbovineserum,penicillin,
foundtobeeffectiveinourmurinemedulloblastomaandBCC
streptomycin,L-glutaminesupplementedwithGeneticin400mg/ml.
models. It is worth noting that itraconazole acts at a distinct
SHHN conditioned medium (CM) was prepared as previously described
site from cyclopamine; itraconazole and related compounds (Maity et al., 2005). HEK293 CM was used for control. HEK293 CM was
may therefore represent a helpful adjunct or alternative to the preparedinthesamemannerasSHHNCM.StockSHHNCMandHEK293
cyclopaminemimicscurrentlyunderwidespreaddevelopment. CM(forcontrol)werediluted10-foldwiththeappropriatemediaforvarious
For these reasons, particularly its well-known safety profile, it Hhpathwaysignalingassays.
ForSMOoverexpressionassayinNIH3T3cells,Smo,GFP,Gli-dependent
shouldbepossibletoquicklyevaluateitraconazoleeitheralone
fireflyluciferasereporter,andTK-Renillaluciferasereportervectorshavebeen
or in combination with other targeted therapies or cytotoxic
describedpreviously(Taipaleetal.,2000).Fugene6(RocheAppliedScience)
agents in patients with tumors of known Hedgehog pathway was used for transient transfection of the vectors at a ratio of 4 ml:1.5 mg
dependence. (Fugene6:DNA). Each well of a 24 well multiwell plate was transfected with
CancerCell17,388–399,April13,2010ª2010ElsevierInc. 395
Cancer Cell
AntifungalItraconazoleAntagonizesHhSignaling
Figure 7. Itraconazole Inhibits the Growth
of Hh-Dependent Basal Cell Carcinoma
InVivo
(A) Compared with cyclodextrin vehicle-treated
tumors, itraconazole suppressed the growth of
endogenous basal cell carcinomas from a
K-14CreER;Ptch+/(cid:3);p53 fl/fl mouse. Treatments
weregiventwiceperdaybyoralgavage.Grayarrow
indicatesthedaythatitraconazoletreatmentwas
stopped.Thenumberoftumorsusedinthisexper-
imentwereasfollows:Cyclodextrincontrol,n=10;
itraconazole100mg/kgn=20untilday19when
itraconazole treatment was discontinued then
n=6aftersomemicewereeuthanizedandtumors
dissectedforfurtherexamination(seeExperimental
Procedures).Forstatisticalanalysis,pairwisecom-
parisonsoftreatmentarmsweremadeforcorre-
spondingdayoftreatmentwiththeexceptionof
cyclodextrincontrolday17beingcomparedwith
itraconazoleday18.Dataareexpressedasmean
percentagechangeintumorlength±SEM.
(B) Cyclodextrin control treated tumors showed
nests of well-organized tumor cells (arrowhead),
whereas itraconazole-treated tumors showed
greaterevidenceofnecrosisasseenbydestruction
oforganizednestsofbasophilictumorcellsand
pyknosis (arrow) and destruction of eosinophilic
stroma lining the BCC nests (black arrowhead).
ImagesareH&Esections.Thescalebarrepresents
100mm.
(C)Fromthetumorgrowthinhibitionstudy,BCC
tumors from itraconazole-treated mice (including
those that had itraconazole discontinued) show
moreevidenceofnecrosisthancyclodextrincontrol
treatedtumorsasmeasuredbynecrosisscore(see
Experimental Procedures) with higher numbers
indicatingworseningnecrosis.
FluorescentBOPDIPY-Cyclopamine
CompetitionAssays
Maintenance and induction of SMO in the
HEK293Scelllinestablyexpressingatetracycline
inducible Smo gene has been described previ-
375ngtotalDNA.SignalingassayforPtch(cid:3)/(cid:3)cellswasperformedsimilarlyexcept ously(Dwyeretal.,2007).After2daysofSMOinduction,cellswereincubated
forFugene:DNAratioof3ml:10mgwith250ngtotalDNAperwellofa24-wellplate. withBODIPY-cyclopamine10nM±inhibitorsat37(cid:4)Cfor4-6hr.Cellswere
AllassaysforHhpathwayactivationwereperformedin24welltissueculture trypsinized,washedwithphosphate-bufferedsaline(PBS)aftercentrifugation,
plates.Cellsweregrownin10%(v/v)serummediauntilcompletelyconfluent. pelleted by centrifugation and resuspended in 0.5% calf serum/phenol
Drugs,293CM,andSHHNCMweredilutedinDMEMcontaining0.5%(v/v) red-freeDMEM.BODIPYfluorescencewasmeasuredonFACSAria(Becton
serum unlessotherwise notedinmaintext andfigure legends.Cellswere Dickinson). Fluorescence-activated cell sorting data were analyzed with
incubatedfor40-46hrwithdrugtreatments,exceptforthecomparisonof FlowJosoftware.
variousazole(30hr)andSMOA1(30hr)tominimizetoxicityfromthehigh
concentrationsofazoleantifungalsandforskolin,respectively.Afterinternal Immunofluorescence
normalizationoffireflyluciferasewithtk-Renillaluciferaseforeachwell,rela- NIH3T3cellsweregrowntoconfluenceontissueculturetreatedglasscover-
tiveluciferaseunitswereobtainedbynormalizationofHhpathwaystimulated slips(Fisher)andtreatedwithvariousconditionsfor6hrasdescribedprevi-
cellswithunstimulatedcells(allotherconditionsbeingequal)grownonasepa- ouslyforHhpathwaysignalingassays.Attheendoftheincubationperiod,
rateplateinparallel.AllluminescenceassayswereperformedonaFluostar samples were fixed in methanol at (cid:3)20(cid:4)C and subsequently stained with
Optima (BMG Labtech). Luminescence assays were performed according appropriateantibodies.Theantibodiesusedwereasfollows:rabbitpolyclonal
tomanufacturer’sinstructionsusingDualLuciferaseAssayReporterSystem anti-SMO 1:750dilution, mousemonoclonal anti-acetylated tubulin 1:1000
(Promega). b-Galactosidase activity of Ptch(cid:3)/(cid:3) cells were obtained with dilution (Sigma), AlexaFluor 488goat anti-rabbit 1:667dilution (Invitrogen),
Galacto-LightPlusReporterSystem(AppliedBiosystems)permanufacturer’s and AlexaFluor 594 goat anti-mouse 1:667 dilution (Invitrogen). Details of
instructions.TotalproteinwasobtainedusingBCAProteinAssaykit(Thermo anti-SMOantibodyhavebeendescribedelsewhere(Kimetal.,2009).
Scientific)accordingtothemanufacturer’sinstructions.
Dose-responsecurvesandIC valuesweredeterminedbyfittingthedata Microscopy
50
withfourparameterlogarithmicnonlinearregressionanalysiswithGraphPad ImagesforciliaryaccumulationofSMOwereobtainedonaZeissLSM510
Prismsoftware. MetaConfocalmicroscopeattheStanfordInstituteforNeuro-Innovation&
396 CancerCell17,388–399,April13,2010ª2010ElsevierInc.
Cancer Cell
AntifungalItraconazoleAntagonizesHhSignaling
TranslationalNeuroscience.MaximumintensityoverlayimagesofZ-stacks tworemainingwerefolloweduntil22days.Twomicewith10BCCtumors
wereobtainedwithZeissZensoftware.Ciliafromtheseimageswerequanti- were treated with 40% w/v HPCD (Sigma) for an average of 22 days at
fiedmanually. which point the mice were euthanized because of their large tumor size
Forpresentationofimages,representative3003300pixelportionsofeach that exceeded our animal care guidelines. We measured the change in
metafile were cut and maximum intensity overlay images of Z-stack were the long axis of BCC tumor size with calipers every 3 days (So et al.,
produced using Zeiss Zen software. The cut images were enlarged using 2004).Pairwisecomparisonsbetweencyclodextrincontrolanditraconazole
AdobePhotoshopsoftware.Forinsetimages,portionswerecutandenlarged treatmentarmsweredoneusingatwo-sidedunpairedttestwithGraphPad
usingAdobePhotoshop.Subsequently,redandgreenchannelswereoffset Prismsoftware.
using ImageJ software. All images were then recombined and brightened
equallyusingAdobePhotoshop.
NecrosisScoringofBCCTumors
InVivoDrugTreatmentStudies Hematoxylin and eosin stained (H&E) sections of paraffin-embedded BCC
All mouse studies were approved by and conformed to the policies and tumorsafterthegrowthinhibitionstudywereexaminedfornecrosis.Necrosis
regulationsoftherespectiveInstitutionalAnimalCareandUseCommittees wasscoredasfollows:0=noevidenceofnecrosiswithlargebasaloidcells;
atStanfordUniversityandtheChildren’sHospitalOaklandResearchInstitute. 1=mildnecrosiswithnuclearpyknosis;2=moderatenecrosiswith<25%
necrosis;3=severenecrosiswith>25%necrosis.Aftertabulationofnecrosis
StudyofMedulloblastomaTumorGrowthInhibition scores,pairwisecomparisonbetweencyclodextrincontrolanditraconazole
PrimarymedulloblastomafromPtch+/(cid:3);p53(cid:3)/(cid:3)mousewasdissected,minced treatmenttumorsweremadeusingatwo-sidedunpairedttestwithGraphPad
andtrituratedthrougha20Gneedleundersterileconditions.Tumorwasmixed Prismsoftware.
ina1:1solutionofPBS:growthfactorreducedMatrigel(BDBioscience)and
equalvolumeswereallograftedsubcutaneouslyintorightandleftflanksof
6to7wkoldfemaleoutbredathymicnudemice(CharlesRiverLaboratories, ItraconazoleLevels
Wilmington,MA).Tumorswereestablishedwithanaveragetumorvolumeof Fortheitraconazolepharmacokineticstudy,7-to8-week-oldfemaleoutbred
eacharm(cid:2)110mm3beforetreatment.Tumorsweremeasuredwithvernier athymicnudemiceweregivenonedoseofitraconazole100mg/kgor75mg/
calipersevery3-4daysandtumorvolume(Bermanetal.,2002)wascalculated kgorequivolumeof40%HPCDsolutionattime=0hrpoint.Ateachtimepoint,
as(width3length3([0.53(length+width)]).Drinkingwaterwasacidified. threemicefromeachtreatmentarmwereanesthetized,exsanguinated,blood
Animalsweretreatedtwiceperday,10-14hrapartbyoralgavagewithequal collected,andeuthanized.Bloodwasallowedtoclotandtheresultantserum
volumes of 40% (w/v) 2-hydroxypropyl-b-cyclodextrin (HPCD) solution as usedforbioassaystomeasureitraconazolelevels.Serumfromuntreatedathy-
controlanddrugs.Oralitraconazolesolution 10mg/ml(Ortho-Biotec) was micnudemicewasusedforobtainingthestandardcurveforthebioassays.
used for itraconazole 100 mg/kg arm. Oral itraconazole solution contains Areaunderthecurve(AUC)from0–24hrwerecalculatedwithGraphPadPrism
40%(w/v)HPCDtosolubilizeitraconazole(Martin,1999).Fortheitraconazole softwarewiththetrapezoidalrulemethod.
75mg/kgarm,oralitraconazolesolutionwasdilutedto7.5mg/mlwithsterile Foritraconazoleconcentrationsfromthemedulloblastomaallograftstudy,
water.Cyclopaminewasdissolvedin10%(w/v)HPCD/0.1Msodiumcitrate/ serumfromthreemiceofeachtreatmentandcontrolarmwereobtainedat
phosphate solution (pH 3) (Feldmann et al., 2007) at a concentration of theendofthestudyinasimilarmannertothepharmacokineticstudy.
3.75mg/ml.Foritraconazole100mg/kg-cyclopamine25mg/kg,cyclopamine For itraconazole concentrations of tumors from the medulloblastoma
powderwasdissolvedinoralitraconazole10mg/mlsolutiontoafinalconcen- allograft study, mice were exsanguinated under isoflurane anesthesia and
trationof2.5mg/ml.Aone-wayanalysisofvariancewasusedforcomparingall perfused with PBS via the left ventricle to remove any blood that might
treatmentarmswithGraphPadPrismsoftware. confoundtumoritraconazolelevels.Tumorswereexcisedandfrozenondry
ice,ground,andextractedwithsolventfordrugassayaspreviouslydescribed
StudyofTreatmentforHhPathwayInhibition (Clemonsetal.,2002).Serumandtissueantifungalitraconazoleconcentra-
Anuntreatedprimarymedulloblastomaallograftfromthetumorgrowthstudy tionsweredeterminedbybioassaymethod,withitraconazolestandards,as
was minced, digested into a single cell suspension with collagenase IV described elsewhere (Clemons et al., 2002; Harvey et al., 1980; Hostetler
(Worthington Biochemical), and stored in liquid nitrogen. Medulloblastoma etal.,1992;Rexetal.,1991).
cells were thawed and incubated overnight in defined serum-free media
(Kalanietal.,2008)containingB-27supplementwithoutvitaminA(Invitrogen),
mouseEGF20ng/ml(Peprotech),mousebFGF20ng/ml(Invitrogen),minimal SUPPLEMENTALINFORMATION
nonessentialaminoacids(Invitrogen),N-acetylcysteine60mg/ml,Glutamax
(Invitrogen)inNeurobasalAmedium(Invitrogen).Cellswerepelleted,washed SupplementalInformationincludestwofigures,fourtables,andSupplemental
withPBS,andmixedin1:1solutionofPBS:growth-factor-reducedMatrigel Experimental Procedures and can be found with this article online at doi:
asnotedabove.Approximately1.2millioncellswereinjectedintotheflank 10.1016/j.ccr.2010.02.027.
of nude mice as noted above. Tumors were established and mice were
treatedtwiceperdaybyoralgavagefor4days.Drugswerepreparedasnoted
above. ACKNOWLEDGMENTS
GenerationofEndogenousBasalCellCarcinoma WethankN.SeverforprovidingHEK293SinducibleSMOcellline,M.Martinez
DifferentmicecarryingtheK14-Cre-ERtransgene(Metzgeretal.,2005),the andA.J.Tongfortechnicalassistance,P.LovelaceforassistancewithFACS,
floxed p53 allele (Jonkers et al., 2001), and the Ptch1+/(cid:3) (Goodrich et al., A.Olsonforassistancewithconfocalmicroscopy,andM.Scottforhelpful
1997) alleles were bred together for generating Ptch1+/(cid:3); K14-Cre-ER; commentsanddiscussionofthemanuscript.J.K.isarecipientoftheAmer-
p53fl/flmice.Wetreatedthesemiceatage6weekswith100mg/daytamoxifen icanSocietyofClinicalOncologyYoungInvestigatorAward.Thisresearch
administeredintraperitoneallyforthreeconsecutivedays.At8weeksofage, wassupportedinpartbygrantsfromtheProstateCancerFoundation,from
Ptch1+/(cid:3)K14-Cre-ERp53fl/flmicewereexposedto4Gyofionizingradiation the Stanford University Center for Children’s Brain Tumors, and from the
(IR).MicedevelopvisibleBCCtumorsat6monthsofage. NationalInstitutesofHealth.P.A.BisaninvestigatoroftheHowardHughes
MedicalInstitute.
StudyofBCCTumorGrowthInhibition
Fourmicewithatotalof20visibleBCCtumorsweregavagedtwicedaily Received:May29,2009
withoralitraconazolesolution(Ortho-Biotec)100mg/kg,for18days.Itraco- Revised:December3,2009
nazolewasdiscontinuedonday19,twomicewereeuthanizedonday19 Accepted:February12,2010
and the tumors dissected for further examination and BCC tumors of the Published:April12,2010
CancerCell17,388–399,April13,2010ª2010ElsevierInc. 397