Active Efflux of Dasatinib from the Brain Limits Effi against … · Preclinical Development Active Efflux of Dasatinib from the Brain Limits Efficacy against Murine Glioblastoma:

Preclinical Development

Active Efflux of Dasatinib from the Brain Limits Efficacyagainst Murine Glioblastoma: Broad Implications for theClinical Use of Molecularly Targeted Agents

Sagar Agarwal1,6, Rajendar K. Mittapalli1,6, David M. Zellmer2, Jose L. Gallardo2, Randy Donelson2,Charlie Seiler2, Stacy A. Decker3, Karen S. SantaCruz4, Jenny L. Pokorny7, Jann N. Sarkaria7,William F. Elmquist1,6, and John R. Ohlfest2,5,6

AbstractThe importance of theblood–brain barrier inpreventing effectivepharmacotherapyof glioblastomahas been

controversial. The controversy stems from the fact that vascular endothelial cell tight junctions are disrupted in

the tumor, allowing some systemic drug delivery. P-glycoprotein (Pgp) and breast cancer resistance protein

(BCRP) efflux drugs from brain capillary endothelial cells into the blood. We tested the hypothesis that

although the tight junctions are "leaky" in the core of glioblastomas, active efflux limits drug delivery to tumor-

infiltrated normal brain and consequently, treatment efficacy. Malignant gliomas were induced by oncogene

transfer into wild-type (WT) mice or mice deficient for Pgp and BCRP (knockout, KO). Glioma-bearing mice

were orally dosedwith dasatinib, a kinase inhibitor anddual BCRP/PgP substrate that is being currently tested

in clinical trials. KOmice treated with dasatinib survived for twice as long asWTmice. Microdissection of the

tumor core, invasive rim, andnormal brain revealed 2- to 3-fold enhancement in dasatinib brain concentrations

in KO mice relative to WT. Analysis of signaling showed that poor drug delivery correlated with the lack of

inhibition of a dasatinib target, especially in normal brain. A majority of human glioma xenograft lines tested

expressed BCRP or PgP and were sensitized to dasatinib by a dual BCRP/Pgp inhibitor, illustrating a second

barrier to drug delivery intrinsic to the tumor itself. These data show that active efflux is a relevant obstacle to

treating glioblastoma and provide a plausible mechanistic basis for the clinical failure of numerous drugs that

are BCRP/Pgp substrates. Mol Cancer Ther; 11(10); 2183–92. �2012 AACR.

IntroductionGlioblastoma multiforme (GBM) is a devastating pri-

mary brain tumor that claims more than 12,000 lives eachyear in theUnited States (1). Despite aggressive treatment,nearly all malignant gliomas recur (2), eventually leadingto patient death. The median overall survival of patientswith GBM is 16 to 19 months (3); survival after tumorrecurrence is a dismal 5 to 7 months (4). Small moleculemolecularly targeted agents that target signaling path-ways critical for tumor cell growth have been extensivelyevaluated for therapy in GBM (5). Unfortunately, none of

these have resulted in any significant clinical benefitdespite success in numerous preclinical models (6).

A critical challenge in treatingmost neurologic diseasesis delivery of drugs into the central nervous system (7).The blood–brain barrier (BBB) separates the brain fromperipheral circulation and thus protects it from harmfultoxins and chemicals (8). This physical barrier is largelyattributable to the tight junctions between brain capillaryendothelial cells that limit paracellular diffusion. Drugtransporter proteins, such as the ATP-binding cassettetransporters P-glycoprotein (Pgp, encoded by ABCB1)and the breast cancer resistance protein (BCRP, encodedby ABCG2), constitute a major component of this barrierand restrict drug penetration into the brain by effluxingthem back into the blood (9). These transporters signifi-cantly limit the distribution of a number of drugs to thebrain, includingmultiple anticancer agents currentlyusedin clinic (10–17). Previous studies have shown that severalmolecularly targeted tyrosine kinase inhibitors (TKI) aredual substrates for Pgp and BCRP, and as a result, havepoor brainpenetration (10–12). Simultaneous inhibition ofboth transporters showed dramatic increases in brainpenetration of these drugs, suggesting that it is necessaryto inhibit both proteins to significantly improve the brainpenetration of dual substrates (10–12).

Authors' Affiliations: Departments of 1Pharmaceutics, 2Pediatrics, 3Neu-roscience, 4Laboratory Medicine and Pathology, and 5Neurosurgery;6Brain Barriers Research Center, University of Minnesota, Minneapolis;and 7Department of Radiation Oncology, Mayo Clinic, Rochester,Minneapolis

Note: Supplementary data for this article are available at Molecular CancerTherapeutics Online (http://mct.aacrjournals.org/).

Corresponding Author: John Ohlfest, Department of Pediatrics, Univer-sity ofMinnesota, 420DelawareSt. SE,MMC366,Minneapolis,MN55455.Phone: 612-626-2491; Fax: 612-626-2490; E-mail: [email protected]

doi: 10.1158/1535-7163.MCT-12-0552

�2012 American Association for Cancer Research.

MolecularCancer

Therapeutics

www.aacrjournals.org 2183

on July 14, 2020. © 2012 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst August 13, 2012; DOI: 10.1158/1535-7163.MCT-12-0552

http://mct.aacrjournals.org/

The relevance of the BBB in treating brain tumors hasbeen controversial. Imaging studies have long beenknown to reveal brain malignancies, because the "leaky"tight junctions in the bulk tumor mass allow the imagingcontrast agent to selectively accumulate in the tumor core.This clinical observation along with recent studies thatshow robust TKI delivery into the core tumormass (18, 19)have suggested that treatment failure with TKIs is notbecause of inadequate drug delivery. According to thishypothesis, the BBB is compromised by residual damagefromradiotherapy and/orby the tumor itself such that thefunctional integrity of the BBB is lost and drug delivery isnot the basis for treatment failure (18, 19).

There is also evidence to support the counter hypoth-esis: that the BBB is not uniformly leaky, and drug deliv-ery is a critical obstacle to treating GBM. Levin andcolleagues experimentally have shown that BBB disrup-tion in brain tumors is not uniform within the tumor orsurrounding brain (20, 21). Glioma cells have the ability toinvade normal brain tissue several centimeters away fromthe tumor, escaping surgical resection, and thus remainprotected by a relatively intact BBB (22, 23). Although theBBB might be disrupted at or near the tumor core, itremains intact in areas away fromthe core and consequent-ly restricts delivery of anticancer drugs in these regions(24). This suggests that limited drug delivery to the inva-sive tumor cellsmay be a significant contributor to therapyresistance in many GBM patients. This kind of drug deliv-ery failure is often not detected in preclinical models thatare used to justify clinical trials. Some preclinical modelsused are not established in the brain (e.g., flank model), orassessments of antitumor efficacy involves mice bearingwell-circumscribed brain tumors derived from implantedcell lines that exhibit a leaky BBB amid no appreciableinfiltration. The latter well-circumscribed phenotype is thegrowth pattern of the majority of standard implantedmodels that are typically used for preclinical validation inthe process of drug development (25). If inadequate drugdelivery is a plausible mechanism for the failure of TKIs inthe clinic, then it is not surprising that clinical trials derivedfrom these inadequate preclinicalmodels have culminatedin a series of disappointing failures.

On the basis of this premise, we hypothesize thatrestricted delivery of molecularly targeted agents to thebrain can significantly limit efficacy. We have used dasa-tinib as a model agent because we have shown that itsbrain penetration is limited by Pgp- and BCRP-mediatedactive efflux at the BBB (12) and because it is currentlybeing testing in clinical trials for GBM. Because activeefflux limits the brain penetration of many drugs used totreat glioma, dasatinib is considered a reasonable surro-gate to address the overarching question of how efflux atthe BBB limits treatment efficacy. Using a highly infiltra-tive model of glioma, we show that increased delivery ofdasatinib to the brain of mice deficient in Pgp and BCRPsignificantly enhances its efficacy. Our data suggest thatinadequate drugdelivery is a perfectly plausible, yet oftendiscounted mechanism to explain the clinical failure of

TKIs and warrants clinical investigation into using dualBCRP/Pgp pharmacologic inhibitors to improve drugdelivery to GBM.

Materials and MethodsChemicals

Dasatinib (Fig. 1) was purchased from LC Laboratories.Texas Red Dextran 3000 MW (TRD) was purchased fromMolecular Probes (Invitrogen). Elacridar (GF120918) wasobtained from Toronto Research Chemicals.

Animal careAll procedures were carried out in accordance with the

guidelines set by thePrinciples of LaboratoryAnimalCare(NIH, Bethesda, MD) and were approved by The Institu-tional Animal Care and Use Committee of the Universityof Minnesota (Minneapolis, MN). FVB wild-type (WT)and Mdr1a/b(�/�)Bcrp1(�/�) mice, hereafter referredto as knockout (KO) mice, were from Taconic Farms, Inc.

Spontaneous glioma modelWe used an oncogene induced de novomodel of malig-

nant glioma previously described by Wiesner and collea-gues (26). Plasmid vectors coding for mouse plateletderived growth factor beta (mPDGFb), enhanced greenfluorescent protein (eGFP) and a short hairpin RNAagainst p53 (p53shRNA), and flanked by transposonIR/DRs, were mixed with a transfection agent, in vivo-jetPEI (PEI; Polyplustransfection Inc.), and a sleepingbeauty (SB) transposase vector containing the luciferasereporter, in the ratio 2:2:1 (PDGF:p53:SB, respectively).Neonatal mice were secured in a cooled, "neonatal rat"

Figure 1. Chemical structures. The structure of dasatinib and GF120918are shown.

Agarwal et al.

Mol Cancer Ther; 11(10) October 2012 Molecular Cancer Therapeutics2184




stereotaxic frame (Stoelting) maintained at 4�C to 8�C andDNA–PEI complexes were injected into the right lateralventricle (1 mg total DNA in 2 mL) using a 10 mL Hamiltonsyringe fittedwith a 30-gauge hypodermic needle (Hamil-ton Company). Mice were allowed to recover on a heatedpad before moving them back to their cages. Tumorsspontaneously arose after 3 to 8 weeks as detected bybioluminescent imaging (Fig. 2; ref. 26). Animals wereimaged regularly and were in enrolled in experimentswhenameasurable tumorwasdocumented, as definedbya minimum signal of 2 � 105 photons/sec/cm2/Sr.

Regional breakdown of the BBBTumor-bearingWTmice were anesthetized using keta-

mine (100 mg/kg) and xylazine (10 mg/kg). The animalswere injected with TRD (1.5 mg/animal body weight) viathe tail vein. TRDwas allowed to circulate for 10minutes,after which the animal was perfused by a brief cardiacwashout for 30 seconds. After perfusion, the brain tissuewas removed and flash frozen in isopentane (�60�C).Care was taken to remove brain rapidly in less than 60seconds to minimize post mortem diffusion of the com-pound. Brains were then sectioned (20 mm) using a cryo-

stat (Leica Cryotome) at �20�C. Brain sections weremounted on glass slides and air-dried. After tissue prep-aration, fluorescent images were obtained using Leicainverted fluorescent microscopy (model DMI6000B). Thefluorescent image analysis was carried out using Image J.To convert the fluorescent intensity to concentration, astandard reference curve was generated in brain homog-enate as previously described (27).

Plasma and brain distribution of dasatinib after oraldosing in FVB WT and KO mice

WT and KO mice (n ¼ 28 per group) were adminis-tered 15 mg/kg dasatinib via oral gavage. Animals weresacrificed at predetermined time points post dose (n ¼ 4at each time point). Blood was collected by cardiacpuncture and plasma was harvested. Whole brain wasimmediately removed from the skull, rinsed with coldsaline, and flash frozen in liquid nitrogen. Plasma andbrain specimens were stored at �80�C until analysis byLC-MS/MS for determination of dasatinib concentra-tions. Pharmacokinetic parameters from the concentra-tion–time data in plasma and brain were obtained bynoncompartmental analysis carried out using PhoenixWinNonlin 6.1 (Mountain View). The area under theconcentration–time profiles for plasma (AUCplasma) andbrain (AUCbrain) was calculated using the linear trape-zoidal method. The sparse sampling module in Win-Nonlin was used to estimate the standard error aroundthe mean of the AUCs.

Brain concentrations of dasatinibTumor-bearingWT and KOmice were administered 15

mg/kg dasatinib via oral gavage (n ¼ 5 - 10 per group).Animals were sacrificed 90 minutes post dose and thewhole brain was harvested. With the aid of GFP visual-ization goggles (Biological Laboratory Equipment), thebrain was immediately dissected into core, rim (brainaround tumor), and normal brain. Because the tumorexpresses GFP, the region of concentrated green fluores-cence was deemed as the tumor core, the tissue surround-ing the tumor was called as tumor rim and normal braintissue was collected from the left (contralateral) hemi-sphere. Tissue specimens were flash frozen in liquidnitrogen and stored at � 80�C until further analysis.

Dasatinib concentrations in the tissue specimens weredetermined by a rapid and sensitive LC-MS/MSmethod.Frozen samples were thawed at ambient temperature.Brain samples were homogenized using 3 volumes of5% bovine serum albumin in PBS and a 100-mL aliquotof brain homogenate or plasma was used for analysis.Samples were spiked with 10 ng of gefitinib as internalstandard and alkalinized by addition of 200 mL of a pH 11buffer (1mmol/L sodiumhydroxide, 0.5mmol/L sodiumbicarbonate). The samples were then extracted by liquid–liquid extraction using ice-cold ethyl acetate. A 5-mLvolume of sample was injected in the HPLC system.Chromatographic analysis was carried out using an Agi-lentModel 1200 separation system and anAgilent Eclipse

Figure 2. Oncogene-induced glioblastoma model. A, schematicillustrating transposon vectors delivered into neonatal mice to initiateglioblastoma. Tumors were tractable by bioluminescent imaging asillustrated by a representative animal imaged longitudinally. B,representative histologic features of the tumor from a moribund animalby H&E staining. Black box in low-power sagittal image showscorresponding region highlighted at higher power. Top panel, documentinfiltrative growth pattern. Arrows in lower left panel highlight areas ofnecrosis. Lower right panels are examples of microvascular proliferation.

Brain Efflux Limits Efficacy of a Molecularly Targeted Agent

www.aacrjournals.org Mol Cancer Ther; 11(10) October 2012 2185




XDB-C18 RRHT threaded column (Santa Clara). Themobile phase was composed of acetonitrile: 20 mmol/Lammonium formate (containing 0.1% formic acid, pH 4;32:68 v/v). The column effluent was monitored using aThermo Finnigan TSQQuantum 1.5 detector (San Jose) toallow the [MH]þ ion of dasatinib at m/z 488 and that ofinternal standard atm/z 446.9 to pass into the collision cell.The product ions for dasatinib (m/z 401) and the internalstandard gefitinib (m/z 128.1)weremonitored through thethird quadrupole. The lower limit of detection of the assaywas at least 2.5 ng/mL with a corresponding CV ofapproximately 10%.

Regional efficacy of dasatinib by target inhibitionusing Western blot

A separate group ofWT and KOmice (n¼ 4 per group)were treated with vehicle or dasatinib (15 mg/kg) for 3days. Mice were sacrificed and the whole brain wasdissected into core, rim (brain around tumor), and normalbrain using the GFP visualization goggles as described.Western blotting was conducted in the tissue specimensfor determination of expression levels of AKT, phospho-AKT, SRC, phospho-SRC, and b-actin as a control forprotein loading. Tissue specimens were lysed in radio-immunoprecipitation assay buffer containing proteaseand phosphatase inhibitors. Protein concentration wasdetermined using the bicinchoninic acid assay and 40 mgwere loaded per lane on a 4% to 12% SDS-PAGE gel, andrun at 170 V for 1 hour. Gels were transferred to anitrocellulose membrane and blots were then incubatedwith the primary antibodies to pSRC, SRC, pAKT, AKT,

and actin (1:1000; Cell Signaling) overnight at 4�C.Membranes were washed and incubated with horserad-ish peroxidase (HRP)-conjugated secondary antibody(1:1000; Cell Signaling) for 1 hour. Proteins were detectedusing Amersham ECL Advance Western Blotting Detec-tion Kit (GE Life Sciences).

Efficacy of dasatinib in the spontaneous gliomamodel

Oncogene-injected mice were imaged regularly todetect tumor formation and were enrolled in a treatmentarm within a day of reaching the tumor signal threshold(minimum of 2 � 105 photons/sec/cm2/Sr). Mice wererandomly assigned to 1 of 4 groups (n¼ 5–8 per group): (i)WT mice treated with 15 mg/kg dasatinib, (ii) WT micetreated with vehicle, (iii) KO mice treated with dasatinib,and (iv) KO mice treated with vehicle. All mice receivedeither dasatinib or vehicle by oral gavage every 12 hoursfor 7 days and survival was monitored posttreatmentinitiation with the experimental endpoint being death ormoribund status of the mice.

Efficacy of dasatinib and transporter expression inhuman GBM cell lines

The short-term explant cell lines used in this studywerederived from primary human GBM specimens and havebeen maintained by serial xenotransplantation in mice aspreviously described in detail (28). The xenograft lineswere established and exclusively maintained in the Sar-karia laboratory. Therefore, no authentication was carriedout. For drug sensitivity assays, 5,000 cells per well were

Bright field image

15

10

5

0

6

4

2

0

***

A

B***

Brain BrainTumor

1.0 ± 0.4

3.6 ± 1.5

Tumor

Concentr

ation o

f

TR

D 3

kD

a (

µg/g

)

Fold

incre

ase in T

RD

3 k

Da

perm

eabili

ty

Green – eGFP tumor Red – TX red distribution

Figure 3. BBB permeability in thetumor visualized by Texas reddextran. A, 4 WT tumor-bearingmice were systemically injectedwith TRD. Representativequalitative images show higherTRD accumulation in the tumorrelative to the adjacent normalbrain. B, the concentration of TRDis approximately 3.6-fold higher inthe tumor region compared withnormal brain. Each symbolrepresents a measurement; 5 to6 measurements were made perbrain in 4 total animals(��, P < 0.0001).

Agarwal et al.





cultured in the presence of 0, 0.3, 1, and 10 mmol/L ofdasatinib with or without 5 mmol/L of GF120918, a dualBCRP/PgP inhibitor. Seventy-two hours later, cell viabil-ity was measured using an MTS assay as previouslydescribed (29). The cell-survival data in the GBM cell lineswere analyzed using an inhibitory sigmoid Emax model.The IC50valueswere estimatedandcomparedbetween thecontrol and GF120918-treated groups. For Western blots,primary antibodies to b-actin (1:1,000; Cell Signalling),ABCB1 (p-glycoprotein [c219], 1:100), and ABCG2 (BXP-

21, 1:500; Enzo Life Sciences) were used. HRP-conjugatedantimouse (1:20,000; Pierce), and anti-rabbit (1:5000; CellSignaling) secondary antibodies were used for detection.

Statistical analysesAnimal survival was analyzed with Prism 4 software

(Graph Pad Software, Inc.) using a log-rank test. For allother experiments, statistical analyses were made withSigmaStat, version 3.1 (Systat Software, Inc.) using a 2-tailed t test.

ResultsRegional breakdown of the vascular endothelial celltight junctions

We utilized an oncogene-induced de novo model ofmurine GBM (26) with 1 modification, the replacementof theNRAS oncogene withmPDGFb. The firefly luciferase,eGFP, and p53shRNA genes were delivered on separatevectors as before (Fig. 2A). Tumors arose within 4 to 8weeks of birth and progressed rapidly, killing animalswithin 2 weeks or less from the time of initial detectionusing bioluminescent imaging. Representative histologicanalysis of a tumor from a moribund animal is shownin Figure 2B. Hallmark features of GBM were apparent,including diffuse infiltration of tumor cells into thesurrounding normal brain tissue (e.g., absence of cleartumor:brain boundary), necrosis, and microvascular pro-liferation. Importantly, similar to human patients, thesetumors formed at different locations within the brain ofeach animal, introducing a source of spatial heterogeneitynot modeled in transplanted models. To determine if thismodel had defects in the capillary endothelial cell tightjunctions similar to human GBM, a 3 kDa TRD probe wassystemically administered and allowed to circulate for10minutes. TRD brain penetrationwas analyzed inmicro-dissected tumor (based on GFP) and normal brain. Micro-scopic analysis revealed large tumorswith aheterogeneousdistribution of TRD within the center of the mass and lessTRD accumulation in adjacent normal brain (Fig. 3A).Quantitation of TRD showed over a 3-fold increase in TRDin the GFPbright tissue harboring the tumor core relative tothe adjacent normal brain (Fig. 3B). This animal modeltherefore resembles a spatially heterogeneous breakdownof the tight junctions similar to human GBM.

Plasma and brain distribution of dasatinib in miceWe compared the plasma and brain distribution of

dasatinib between WT and KO mice. Plasma concentra-tions of dasatinib after oral dosing in the KOmicewere notsignificantly different than that in WT mice (P > 0.05; Fig.4A). The AUCplasma was 1.31 mg�hour/mL in the WTmicecompared with 1.78 mg�hour/mL in the KOmice (P > 0.05;Supplementary Table S1). This indicates that Pgp andBCRP have minimal influence on the oral absorption orsystemic clearance of dasatinib at the 15 mg/kg dose.Dasatinibbrainconcentrationsweresignificantlyenhancedin KO mice compared with those in the WT mice (P <0.05; Fig. 4B). The AUCbrain was 0.160 mg�hour/mL in the

Figure 4. Comparison of dasatinib brain distribution in WT and KO mice.A, and B, the plasma and brain concentrations of dasatinib, respectively,after 15 mg/kg oral dose. The brain-to-plasma ratio for dasatinib in WTandKOmice is shown inC.Data representmean�SD;n¼3–4 for all datapoints (�, P < 0.05, ��, P < 0.01, ��, P � 0.001).






WT mice and increased by approximately 10-fold to1.63 mg�hour/mL in the KO mice (P ¼ 0.001). The ratio ofAUCbrain to AUCplasma in WT mice was approx. 0.12,indicating the restriction to brain penetration of dasatinib.This ratio increased by 7-fold to 0.93 in the KO mice(SupplementaryTable S1). These results conclusively showthat dasatinib brain distribution is dramatically enhancedwhen P-gp and BCRP are absent at the BBB. The data alsoshow that systemic exposure of dasatinib (the driving forcefor brain distribution) remained relatively unchanged inthe KO mice.

Differential distribution and efficacy of dasatinib inthe brain

We examined the ability of dasatinib to traverse theBBB in the different brain regions of tumor-bearingmice. GFP-aided tumor dissection allowed the brain to

be divided into tumor core, tumor rim, and normalbrain, as illustrated in Figure 5A. In WT mice, meandasatinib concentrations in the tumor core were approx-imately double of that measured in tumor rim andnormal brain (Fig. 5B). In the KO mice, brain concentra-tions increased by over 2-fold in the normal brain, andby up to 3-fold in the tumor rim and core relative to WTmice (P < 0.05).

To determine the impact of this difference in drugdelivery on signal transduction, individual WT and KOmice were dosed with vehicle or dasatinib, and Westernblotting was carried out on tissue samples from the tumorcore, rim, and normal brain to examine the efficacy ofdasatinib in inhibiting phosphorylation of SRC, a directdasatinib target, and AKT, an indirect target (30). In thevehicle-treated WT and KO mice, relatively uniformlevels of phosphorylated (p) SRC were detected in the

Figure 5. Regional differences indasatinib concentration and signaltransduction. A, representativeexample illustrating how the braintissue was dissected using GFPgoggles. The brain was sliced inhalf coronally to expose the tumorexpressing GFP. The center wasconsidered the core, the tissueimmediately adjacent to centerwasconsidered the rim, and theopposite hemisphere wasconsidered normal brain. B, brainconcentration of dasatinib in eachregion of brain tissue wasquantified. Each symbolrepresents ameasurement from anindividual animal (n ¼ 9 WT, n ¼ 5KO; �, P < 0.05, ��, P < 0.01). C,Western blot analysis of braintissue from WT and KO mice.Representative blots from 3 to 4WT and 3–4 KO total mice areshown. Each panel represents apair of individual mice (1 WT, 1 KO)that were analyzed side-by-side onthe same gel.

Agarwal et al.





tumor core, rim, and normal brain tissue (Fig. 5C). Incontrast, p-AKTwaselevated in the tumormass relative tonormal brain in all animals tested. Total SRC and AKTwere relatively uniform, making these comparisonsinterpretable. We also analyzed PDGFR expression andphosphorylation because it is a direct dasatinib targetand presumably a driver oncogene in this animal model.Unfortunately, total PDGFR expression was extremelyvariable within individual animals (regionally) andbetween animals, making any meaningful comparison ofp-PDGFR nearly impossible (data not shown). The moststriking finding we observed was the difference betweenp-SRC in WT relative to KO mice. Dasatinib administra-tion completely extinguished all detectable p-SRC in all 3brain regions in KO mice. In marked contrast, dasatinibdid not appreciably change p-SRC levels in the normalbrain,minimally affectedp-SRC in the rim, andnoticeably(yet incompletely) inhibited p-SRC in the tumor core ofWT mice. Analysis of AKT activation followed a similartrend, whereby dasatinib inhibited p-AKT levels morerobustly in all 3 brain regions in KO mice relative to WT(Fig. 5C). These data clearly establish a relationshipbetween the brain penetration of dasatinib, as effected byBCRP/Pgp-mediated efflux, and inhibition of kinase-mediated phosphorylation.

BCRP and Pgp loss of function increases survival ofglioma-bearing mice treated with dasatinibThe therapeutic efficacy of dasatinib was examined by

monitoring survival inWT andKOmice that were treatedwith a single 7-day cycle of dasatinib. The vehicle treatedcohort died rapidly after initiation of the study, with themedian survival time being 7 days in theWTand 9days inthe KO mice (Fig. 6). When treated with dasatinib, WTmice survived longer, with a median survival time of 14days. Importantly, the most significant effect was seen inthe KO mice treated with dasatinib, where survivalincreased dramatically to yield a median survival timeof 33 days (P ¼ 0.03; Fig. 6). These results conclusivelyshow that dasatinib is much more effective against thetumor when drug delivery to the brain, and hence thetargets, is enhanced in KO mice.

Efficacy of dasatinib in Pgp and BCRP expressinghuman GBM cell linesA subset of glioma cells from humans and murine

models express BCRP and/or Pgp (31, 32). Drug effluxmediated by these transporterswithin the tumor cell itselfcan effectively create a second barrier to drug delivery,and this has beenproposed as an additionalmechanismofchemoresistance in GBM (reviewed in ref. 33). We usedprimaryGBMxenograft lines thatwere establisheddirect-ly from human GBM resections and screened them forexpression of Pgp and BCRP. On the basis of expressionpatterns of these 2 transporters, 6 xenograft lines wereidentified that expressed Pgp and BCRP at different levels(Fig. 7A). The efficacy of dasatinib was determined inshort-term explant cultures from these xenograft lines in

presence and absence of the dual Pgp/BCRP inhibitor,GF120918. Generally, there was a trend for higher BCRP/Pgp expression correlatingwith a higher IC50 of dasatinib.Five of the 6 cell lines testedwere sensitized to dasatinib inthe presence of GF120918.Notably, GF120918 had a great-er effect in lines 12, 22, 26, and 46 that expressed both Pgpand BCRP, relative to lines 6 and 10,which expressed onlyPgp (Fig. 7A and B).

DiscussionMany promising molecularly targeted agents have

failed to show any clinical benefit in patients withglioma. Although studies investigating the failure ofthese agents have proposed various hypotheses, limiteddrug delivery is sometimes disregarded as a likelyexplanation. In fact, clinical reports suggest that highdrug concentrations measured in resected tumor tissueare indicative of a disrupted BBB (19), leading theseinvestigators to conclude that poor drug delivery is nota plausible mechanism for upfront treatment failure(18). The objective of this study was to highlight 2fundamental realities in treatment of an infiltrativebrain tumor such as GBM: (i) delivery of chemothera-peutic agents into the brain can be significantly restrict-ed by a BBB that is not disrupted throughout the entirebrain; (ii) successful therapy in GBM will requireadequate drug delivery throughout the entire brain,especially to tumor cells that may efflux drugs them-selves, nested in normal brain tissue with a relativelyintact BBB. A detailed discussion of these conceptsfocused on drug delivery can be found in our recentreview (33). It is equally important to understand that

Figure 6. Enhanced efficacy of dasatinib in BCRP/Pgp-deficient mice.Mice were enrolled into the indicated treatment arm as soon astumors became apparent by bioluminescent imaging. Each grouprepresents the cumulative survival of 5 to 8 individual mice that weretreated at different times. The time scale indicates elapsed time from theday of enrollment. Significant differences in survival were determined bythe log-rank test.






migratory GBM cells that invade normal brain areintrinsically more resistant to chemotherapy and radio-therapy comparedwith bulk tumor (reviewed in ref. 34).These fundamental differences in treatment sensitivityfurther underscore the importance of using infiltrativepreclinical models to assess therapeutics, and the needto deliver higher drug levels to invasive glioma cellsthat may be harder to kill than bulk tumor.

Adisruption in the BBB at the central core ofGBM is notsurprising because of the pathophysiologic characteristics

of the necrotic tumor that damages the surroundingcapillaries. The resulting high drug concentrations inthe tumor, however, are essentially irrelevant in themajority of operable patients because the tumor core isusually removed at surgery. Drug concentrations in thetissue adjacent to the tumor are themost clinically relevantbecause inmost cases recurrence happensnear themarginof surgical resection (2). An intact BBB in such areascan shield small nests of invasive tumor cells that ulti-mately grow to give rise to the recurrent tumor. There

Figure 7. Inhibition of BCRP/Pgpsensitizes human glioma cells todasatinib. A, the relativeexpression of BCRP and Pgp isshown by Western blot in 6 humanGBM xenograft lines. The IC50

values in the presence ofGF120918 is shown directly abovethe corresponding cell lines in theWestern blot. B, representative cellsurvival curves of 4 of the cell linesshown in A.

Agarwal et al.





are several studies that show that whereas drug con-centrations in the resected tumor can be high (evengreater than plasma), concentrations are dramaticallylower in brain tissue away from the tumor. This hasbeen elegantly shown in a study by Blakeley and col-leagues, where in a human GBM patient, concentrationsof methotrexate in areas centimeters away from thetumor were about an order of magnitude lower thanthat in the tumor (24). Pitz and colleagues recentlyreported that concentrations of several anticanceragents in the noncontrast enhancing regions of the brainwere significantly lower than that in the contrastenhancing tumor (35). Similarly, it has been shown thatconcentrations of paclitaxel (36) and temozolomide (37)in the tumor periphery were lower than that in thetumor core. However, in all these prior studies, therelative importance of tight junctions versus the expres-sion of BCRP and Pgp was not entirely clear. Whenusing drugs that are substrates for these transporters,distinguishing this difference is critical to rationallydevelop effective drug delivery solutions. The majorcontribution of this study was to unequivocally linkBCRP/Pgp expression to treatment failure at the levelof hitting the drug target (p-SRC in this case) and overallsurvival. Results from pharmacokinetic studies showedthat the plasma concentrations of dasatinib in WT andKO mice are not significantly different, so the increasedefficacy observed in KO mice is not because of changingthe bioavailability of dasatinib; but because of changingbrain penetration. Our results show that drug concentra-tions in the tumor core are not representative of wholebrain concentrations. Therefore, the use of drug concen-tration measurements in the resected tumor tissue as aguide for clinical assessment of drug delivery to theentire brain can be misleading.The results of this and previous studies suggest that

many drugs used to treat GBM have "failed for thewrong reason" and could be revisited assuming drugdelivery obstacles are overcome. The finding thatincreasing drug delivery to the brain translates toimproved efficacy can be valuable in choosing futurecombinatorial drug therapy for GBM. We have shownthat inhibition of Pgp and BCRP is a useful strategy toenhance brain penetration of substrate drugs (10, 11),including dasatinib (12). Elacridar (GF120918) is a well-known dual inhibitor of Pgp and BCRP that been testedextensively in mice to improve brain penetration ofnumerous drugs. This study shows that GF120918 canalso enhance the efficacy of dasatinib in human GBMcells. There are also reports suggesting the use of drugsthat are dual Pgp/BCRP substrates to competitivelyinhibit both transporters. These include several TKIsthat have been shown to be substrates for both Pgp andBCRP. In vitro studies show that erlotinib (38), gefitinib(39), lapatinib (40), and sunitinib (41) inhibit ABC trans-porters, and suggest the potential use of these agents ascombination therapy to improve drug delivery to brain.It has been shown that concurrent treatment with gefi-

tinib results in a significantly increased brain (42) andtumor exposure to topotecan in a mouse model ofglioma (43). However, such combination therapies areoften confounded by additive toxicities that hinder theability to effectively saturate Pgp and BCRP, clearlyindicating the need to develop and employ morepotent inhibitors. However, if both BCRP and Pgp areinhibited, it may be possible to give a lower systemicdose of a drug that is an efflux substrate, although stillachieving enhanced brain concentrations (presumablyreducing systemic toxicities). Combination therapywith a molecularly targeted agent and a modulator ofABC transporters could be an attractive approach tocounter one of the most significant challenges in gliomachemotherapy.

In summary, this study highlights the influence ofrestricted drug delivery across the BBB on the efficacyof dasatinib in glioma. The heterogeneous permeability ofthe BBB in the tumor suggests that clinical assessment ofdrug delivery by using the tumor core drug concentra-tions (in resected tissue) as a guide for the adequacy ofdrug delivery can be highly misleading. Inhibition of Pgpand BCRP can be explored in future clinical trials as apossible strategy to improve delivery and hence efficacyof molecularly targeted agents.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: S. Agarwal, R. K.Mittapalli, Jose L. Gallardo, J. N.Sarkaria, W. F. Elmquist, J. R. OhlfestDevelopment of methodology: S. Agarwal, R. K. Mittapalli, Jose L.Gallardo, R. Donelson, K. S. SantaCruz, W. F. Elmquist, J. R. OhlfestAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): S. Agarwal, D. M. Zellmer, Jose L. Gallardo, R.Donelson,K. S. SantaCruz, J. L. Pokorny, J.N. Sarkaria,W. F. Elmquist, J. R.OhlfestAnalysis and interpretation of data (e.g., statistical analysis, biostatis-tics, computational analysis): S. Agarwal, R. K. Mittapalli, D. M. Zellmer,Jose L. Gallardo, R. Donelson, K. S. SantaCruz, J. N. Sarkaria, W. F.Elmquist, J. R. OhlfestWriting, review, and/or revision of the manuscript: S. Agarwal, R. K.Mittapalli, D.M. Zellmer, K. S. SantaCruz, J. N. Sarkaria,W. F. Elmquist, J.R. OhlfestAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): D. M. Zellmer, Jose L. Gallardo, C.Seiler, S. A. Decker, J. R. OhlfestStudy supervision: Jose L. Gallardo, W. F. Elmquist, J. R. Ohlfest

AcknowledgmentsThe authors thank Lajos Mates, Zsuzsanna Izsvak, Zoltan Ivics for

developing and providing the SB100 transgene used in our studies.

Grant SupportThis work was supported in part by grants R01 CA138437 and the

Children’s Cancer Research Fund (W. F. Elmquist, J. R. Ohlfest), F31 NS67937-1 (S. A. Decker), and NS69753 and CA108961 (J. N. Sarkaria).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received May 31, 2012; revised July 20, 2012; accepted August 6, 2012;published OnlineFirst August 13, 2012.






References1. Davis FG, Kupelian V, Freels S, McCarthy B, Surawicz T. Prevalence

estimates for primary brain tumors in theUnitedStates bybehavior andmajor histology groups. Neuro Oncol 2001;3:152–8.

2. WenPY,Kesari S.Malignant gliomas in adults. NEngl JMed2008;359:492–507.

3. GrossmanSA,YeX,PiantadosiS,Desideri S,Nabors LB,RosenfeldM,et al. Survival of patients with newly diagnosed glioblastoma treatedwith radiation and temozolomide in research studies in the UnitedStates. Clin Cancer Res 2010;16:2443–9.

4. WenPY, Brandes AA. Treatment of recurrent high-grade gliomas. CurrOpin Neurol 2009;22:657–64.

5. Collins VP. Mechanisms of disease: genetic predictors of response totreatment in brain tumors. Nat Clin Pract Oncol 2007;4:362–74.

6. DeWitt Hamer PC. Small molecule kinase inhibitors in glioblastoma: asystematic review of clinical studies. Neuro Oncol 2010;12:304–16.

7. Deeken JF, Loscher W. The blood-brain barrier and cancer: transpor-ters, treatment, and Trojan horses. Clin Cancer Res 2007;13:1663–74.

8. Pardridge WM. The blood-brain barrier: bottleneck in brain drugdevelopment. NeuroRx 2005;2:3–14.

9. Schinkel AH, Jonker JW. Mammalian drug efflux transporters of theATP binding cassette (ABC) family: an overview. Adv Drug Deliv Rev2003;55:3–29.

10. Agarwal S, Sane R, Gallardo JL, Ohlfest JR, Elmquist WF. Distributionof gefitinib to the brain is limited by P-glycoprotein (ABCB1) and breastcancer resistance protein (ABCG2)-mediated active efflux. J Pharma-col Exp Ther 2010;334:147–55.

11. Agarwal S, Sane R, Ohlfest JR, Elmquist WF. Role of breast cancerresistance protein (ABCG2/BCRP) in thedistribution of sorafenib to thebrain. J Pharmacol Exp Ther 2011;336:223–33.

12. Chen Y, Agarwal S, Shaik NM, Chen C, Yang Z, Elmquist WF. P-glycoprotein and breast cancer resistance protein influence braindistribution of dasatinib. J Pharmacol Exp Ther 2009;330:956–63.

13. Dai H, Marbach P, Lemaire M, Hayes M, Elmquist WF. Distribution ofSTI-571 to the brain is limited by P-glycoprotein-mediated efflux.J Pharmacol Exp Ther 2003;304:1085–92.

14. deVriesNA,BuckleT,ZhaoJ,BeijnenJH,SchellensJH, vanTellingenO.Restricted brain penetration of the tyrosine kinase inhibitor erlotinib dueto the drug transporters P-gp and BCRP. Invest New Drugs 2010:1–7.

15. de Vries NA, Zhao J, Kroon E, Buckle T, Beijnen JH, van TellingenO. P-glycoprotein and breast cancer resistance protein: two dominanttransporters working together in limiting the brain penetration oftopotecan. Clin Cancer Res 2007;13:6440–9.

16. Polli JW, Olson KL, Chism JP, John-Williams LS, Yeager RL, WoodardSM, et al. An unexpected synergist role of P-glycoprotein and breastcancer resistance protein on the central nervous systempenetration ofthe tyrosine kinase inhibitor lapatinib (N-{3-chloro-4-[(3-fluorobenzyl)oxy]phenyl}-6-[5-({[2-(methylsulfonyl)ethy l]amino}methyl)-2-furyl]-4-quinazolinamine; GW572016). Drug Metab Dispos 2009;37:439–42.

17. Tang SC, Lagas JS, Lankheet NA, Poller B, Hillebrand MJ, Rosing H,et al. Brain accumulation of sunitinib is restricted by P-glycoprotein(ABCB1) and breast cancer resistance protein (ABCG2) and can beenhanced by oral elacridar and sunitinib coadministration. Intl J cancer2011;130:223–33.

18. Hofer S, Frei K. Gefitinib concentrations in human glioblastoma tissue.J Neurooncol 2007;82:175–6.

19. Hofer S, Frei K, Rutz HP. Gefitinib accumulation in glioblastoma tissue.Cancer Biol Ther 2006;5:483–4.

20. LevinVA,PatlakCS, LandahlHD.Heuristicmodeling of drugdelivery tomalignant brain tumors. J Pharmacokinet Biopharm 1980;8:257–96.

21. Levin VA, Freeman-Dove M, Landahl HD. Permeability characteristicsof brain adjacent to tumors in rats. Arch Neurol 1975;32:785–91.

22. Berens ME, Giese A. ". . .those left behind." Biology and oncology ofinvasive glioma cells. Neoplasia 1999;1:208–19.

23. Pitz MW, Desai A, Grossman SA, Blakeley JO. Tissue concentration ofsystemically administered antineoplastic agents in human braintumors. J Neurooncol 2011;104:629–38.

24. Blakeley JO, Olson J, Grossman SA, He X, Weingart J, Supko JG.Effect of blood brain barrier permeability in recurrent high grade

gliomas on the intratumoral pharmacokinetics of methotrexate: amicrodialysis study. J Neurooncol 2009;91:51–8.

25. Lee J, KotliarovaS, Kotliarov Y, Li A, SuQ,DoninNM, et al. Tumor stemcells derived from glioblastomas cultured in bFGF and EGF moreclosely mirror the phenotype and genotype of primary tumors thando serum-cultured cell lines. Cancer Cell 2006;9:391–403.

26. Wiesner SM, Decker SA, Larson JD, Ericson K, Forster C, Gallardo JL,et al.De novo induction of genetically engineered brain tumors in miceusing plasmid DNA. Cancer Res 2009;69:431–9.

27. Lockman PR, Mittapalli RK, Taskar KS, Rudraraju V, Gril B, Bohn KA,et al. Heterogeneous blood-tumor barrier permeability determinesdrug efficacy in experimental brain metastases of breast cancer. ClinCancer Res 2010;16:5664–78.

28. Carlson BL, Pokorny JL, Schroeder MA, Sarkaria JN. Establishment,maintenance and in vitro and in vivo applications of primary humanglioblastoma multiforme (GBM) xenograft models for translationalbiology studies and drug discovery. Curr Protoc Pharmacol 2011;Chapter 14:Unit 14 6.

29. Xiong Z, Gharagozlou S, Vengco I, Chen W, Ohlfest JR. Effective CpGimmunotherapy of breast carcinoma prevents but fails to eradicateestablished brain metastasis. Clin Cancer Res 2008;14:5484–93.

30. Montero JC, Seoane S, Ocana A, Pandiella A. Inhibition of SRC familykinases and receptor tyrosine kinases by dasatinib: possible combi-nations in solid tumors. Clin Cancer Res 2011;17:5546–52.

31. Bleau AM, Huse JT, Holland EC. The ABCG2 resistance network ofglioblastoma. Cell Cycle 2009;8:2936–44.

32. WuA,OhS,Wiesner SM, EricsonK, Chen L, HallWA, et al. Persistenceof CD133(þ) cells in human and mouse glioma cell lines: detailedcharacterization of GL261 glioma cells with cancer stem cell-likeproperties. Stem Cells Dev 2008;17:173–84.

33. Agarwal S, Sane R, Oberoi R, Ohlfest JR, Elmquist WF. Delivery ofmolecularly targeted therapy to malignant glioma, a disease of thewhole brain. Expert Rev Mol Med 2011;13:e17.

34. Lefranc F, Brotchi J, Kiss R. Possible future issues in the treatment ofglioblastomas: special emphasis on cell migration and the resistance ofmigratingglioblastomacells toapoptosis.JClinOncol2005;23:2411–22.

35. Pitz MW, Desai A, Grossman SA, Blakeley JO. Tissue concentration ofsystemically administered antineoplastic agents in human braintumors. J Neurooncol 2011;104:629–38.

36. Fine RL, Chen J, Balmaceda C, Bruce JN, Huang M, Desai M, et al.Randomized study of paclitaxel and tamoxifen deposition into humanbrain tumors: implications for the treatment ofmetastatic brain tumors.Clin Cancer Res 2006;12:5770–6.

37. RossoL,BrockCS,Gallo JM,SaleemA,PricePM,Turkheimer FE, et al.A new model for prediction of drug distribution in tumor and normaltissues: pharmacokinetics of temozolomide in glioma patients. CancerRes 2009;69:120–7.

38. Shi Z, Peng XX, Kim IW, Shukla S, Si QS, Robey RW, et al. Erlotinib(Tarceva, OSI-774) antagonizes ATP-binding cassette subfamily Bmember 1 andATP-binding cassette subfamily Gmember 2-mediateddrug resistance. Cancer Res 2007;67:11012–20.

39. LeggasM, Panetta JC, Zhuang Y, Schuetz JD, Johnston B, Bai F, et al.Gefitinib modulates the function of multiple ATP-binding cassettetransporters in vivo. Cancer Res 2006;66:4802–7.

40. Dai CL, Tiwari AK, Wu CP, Su XD, Wang SR, Liu DG, et al. Lapatinib(Tykerb, GW572016) reverses multidrug resistance in cancer cells byinhibiting the activity of ATP-binding cassette subfamily B member 1and G member 2. Cancer Res 2008;68:7905–14.

41. Dai CL, Liang YJ, Chen LM, Zhang X, Deng WJ, Su XD, et al. Sensi-tization of ABCB1overexpressing cells to chemotherapeutic agentsbyFG020326 via binding to ABCB1 and inhibiting its function. BiochemPharmacol 2009;78:355–64.

42. Zhuang Y, Fraga CH, Hubbard KE, Hagedorn N, Panetta JC, WatersCM, et al. Topotecancentral nervous systempenetration is altered byatyrosine kinase inhibitor. Cancer Res 2006;66:11305–13.

43. Carcaboso AM, Elmeliegy MA, Shen J, Juel SJ, Zhang ZM, CalabreseC, et al. Tyrosine kinase inhibitor gefitinib enhances topotecan pen-etration of gliomas. Cancer Res 2010;70:4499–508.


Agarwal et al.




2012;11:2183-2192. Published OnlineFirst August 13, 2012.Mol Cancer Ther Sagar Agarwal, Rajendar K. Mittapalli, David M. Zellmer, et al. Molecularly Targeted AgentsMurine Glioblastoma: Broad Implications for the Clinical Use of Active Efflux of Dasatinib from the Brain Limits Efficacy against

Updated version

10.1158/1535-7163.MCT-12-0552doi:

Access the most recent version of this article at:

Material

Supplementary

http://mct.aacrjournals.org/content/suppl/2012/08/10/1535-7163.MCT-12-0552.DC1

Access the most recent supplemental material at:

Cited articles

http://mct.aacrjournals.org/content/11/10/2183.full#ref-list-1

This article cites 41 articles, 21 of which you can access for free at:

Citing articles

http://mct.aacrjournals.org/content/11/10/2183.full#related-urls

This article has been cited by 8 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mct.aacrjournals.org/content/11/10/2183To request permission to re-use all or part of this article, use this link



http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-12-0552

http://mct.aacrjournals.org/content/suppl/2012/08/10/1535-7163.MCT-12-0552.DC1

http://mct.aacrjournals.org/content/11/10/2183.full#ref-list-1

http://mct.aacrjournals.org/content/11/10/2183.full#related-urls

http://mct.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mct.aacrjournals.org/content/11/10/2183


Documents

Active Efflux of Dasatinib from the Brain Limits Effi against … · Preclinical Development Active Efflux of Dasatinib from the Brain Limits Efficacy against Murine Glioblastoma: