therapeutics renal cancer resistance to...

Prec

RenDel

RupalDavid

Abst

Intro

ThefrequLindaan E3inducIn thehypoxgrowt(PDGgenicaccou

AuthorBiologyCenterand 4DHarvard

Note: SCancer

R.S. Bh

Corresogy andMA 022bidmc.h

doi: 10

©2010

www.a

D

Published OnlineFirst August 10, 2010; DOI: 10.1158/1535-7163.MCT-10-0477
Published OnlineFirst on September 21, 2010 as 10.1158/1535-7163.MCT-10-0477
Molecular
Cancer
apeutics

linical Development

al Cancer Resistance to Antiangiogenic Therapy Is

Ther

ayed by Restoration of Angiostatic Signaling

S. Bhatt1, Xiaoen Wang2, Liang Zhang1, Michael P. Collins3, Sabina Signoretti3,4,
C. Alsop2, S. Nahum Goldberg2, Michael B. Atkins1, and James W. Mier1
ractTrea

growthowevleculagraftscontinday 3perfussunitinadmindownalthouin vivoperfus

factors,nts for t

s' Affiliatio, 2Departm, 3DepartmepartmentMedical S

upplementTherapeuti

att and X. W

ponding AuCancer Bi15. Phonearvard.edu

.1158/1535-

American A

acrjourna

ownload

tment of metastatic renal cell cancer (RCC) with antiangiogenic agents that block vascular endothelialh factor (VEGF) receptor 2 signaling produces tumor regression in a substantial fraction of patients;er, resistance typically develops within 6 to 12 months. The purpose of this study was to identify mo-r pathways involved in resistance. Treatment of mice bearing either 786-0 or A498 human RCC xeno-with sorafenib or sunitinib produced tumor growth stabilization followed by regrowth despiteued drug administration analogous to the clinical experience. Tumors and plasma were harvested atof therapy and at the time of resistance to assess pathways that may be involved in resistance. Serialion imaging, and plasma and tumor collections were obtained in mice treated with either placebo orib alone or in combination with intratumoral injections of the angiostatic chemokine CXCL9. Sunitinibistration led to an early downmodulation of IFNγ levels as well as reduction of IFNγ receptor andstream angiostatic chemokines CXCL9 to 11 within the tumor. Intratumoral injection of CXCL9,gh producing minimal effects by itself, when combined with sunitinib resulted in delayed resistanceaccompanied by a prolonged reduction of microvascular density and tumor perfusion as measured byion imaging relative to sunitinib alone. These results provide evidence that resistance to VEGF receptory is due at least in part to resumption of angiogenesis in association with reduction of IFNγ-related
therap
angiostatic chemokines, and that this resistance can be delayed by concomitant administration of CXCL9.Mol Cancer Ther; 9(10); OF1–10. ©2010 AACR.

treatm(TKI)and sagenttancesubstaThi

inhibmoderesistaproan

duction

pathogenesis of renal cell carcinoma (RCC)ently involves the inactivation of the von Hippelu (VHL) tumor suppressor gene that encodesubiquitin ligase complex that targets hypoxia-

ible factor for proteasome-mediated degradation.setting of VHL inactivation, a large repertoire ofia inducible genes including vascular endothelialh factor (VEGF) and platelet derived growth factorF) are overexpressed (1). This increase in proangio-

even in the absence of hypoxia, likelyhe nearly unique sensitivity of RCC to

basicanti-VmentMancangiogestedbetterresistaand Gcellsthe amay capproinhibiIFN

static

ns: 1Division of Hematology-Oncology and Cancerent of Radiology, Beth Israel Deaconess Medicalent of Pathology, Brigham and Women's Hospital,of Medical Oncology, Dana-Farber Cancer Institute,chool, Boston, Massachusetts

ary material for this article is available at Molecularcs Online (http://mct.aacrjournals.org/).

ang contributed equally to this work.

thor: Rupal S. Bhatt, Division of Hematology Oncol-ology, Beth Israel Deaconess Medical Center, Boston: 617-735-2062; Fax: 617-735-2060. E-mail: rbhatt@

7163.MCT-10-0477

ssociation for Cancer Research.

ls.org

on May 17, 2018. © 2010mct.aacrjournals.org ed from

ent with small molecule tyrosine kinase inhibitorsof the VEGF receptor (VEGFR) such as sunitiniborafenib. Treatment of patients with RCC with theses frequently leads to tumor regression, but resis-to treatment typically develops within one year,ntially limiting their benefit (2–6).s acquired “evasive” resistance to VEGF pathwayition has been observed in multiple preclinicalls and tumor types (7). In these settings proposednce mechanisms include an increase in alternativegiogenic factors such as interleukin-8 (IL-8) andfibroblast growth factor as seen in the setting ofEGFR2 antibody therapy (8, 9). The empty base-membrane sheaths and pericyte changes seen byuso et al. could also provide the scaffold for tumorgenesis in the resistant setting (10). Other sug-mechanisms include selection of cells that can

tolerate hypoxia (11) and, in the setting of intrinsicnce to VEGF inhibition, the recruitment of CD11br1-positive bone marrow–derived proangiogenic(12). Elucidation of the mechanisms underlyingcquired resistance to VEGFR blockade in RCContribute to the development of novel therapeuticaches that could enhance the efficacy of VEGFRtors in this patient population.
γ signaling leads to the production of three angio-chemokines, CXCL9 to 11 (mig, IP-10, and ITAC;
OF1

American Association for Cancer Research.

http://mct.aacrjournals.org/

ref. 13relativfavoraover,highlyits exsurvivof CXantituRCC,have spanieloss oConseof themightblockapotenstaticVEGFmode

Mate

ReagCX

ceptorbodyanti-vperoxanti-mSignawas fIFNγR

PlasmThe

purchcurve2 to 350 μLroomantiboroom(50 μL10 mition otion astep awereArraymanuare gi

ProteTum

(Cellfluori

fluorifor 10at 12,was dagent12% Steinsprobebodieand thsubstr

Cell cA49

lines (Collecfrozenment.cultursuppland 1turedof bowithchara

TumoFor

nude/The mcabinperimCareical CRC

turesEDTAtaininwashin PB>90%To

mor cof 6- tweighwereSorafe53.6 mweekto a dBardaon treResistby 2 mferencbe re

Bhatt et al.

Mol COF2

D


). These chemokines are highly expressed in RCCe to normal kidney and have been associated withble prognosis in patients with RCC (14, 15). More-CXCR3, the receptor for CXCL9 to 11, is much moreexpressed in RCC than in normal kidney (16), and

pression is associated with improved disease-freeal following nephrectomy (17). Functional studiesCL9 and 10 also show that both chemokines exhibitmor activity in mouse models of lung cancer andrespectively (18, 19). Prior studies from our grouphown that resistance to VEGFR blockade is accom-d by restoration of angiogenesis. We also noted af IFNγ-regulated chemokines at day 3 of treatment.quently, because of the known angiostatic functionse chemokines, we hypothesized that this losscontribute to the acquired resistance to VEGFRde. We investigated this possibility by testing thetial value of restoration and maintenance of angio-chemokines in delaying the acquired resistance toR blockade in murine human tumor xenograftls.

rials and Methods

entsCL9 antibody was obtained from R&D. IFNγ re-1 (IFNγR1), CXCL10, CXCL11, and GAPDH anti-for Western Blot were obtained from Abcam. Theinculin antibody was from Sigma. Horseradishidase (HRP)-conjugated secondary antibodies (goatouse, goat anti-rabbit) were obtained from Cell

ling Technology. For immunohistochemistry, CD34rom Abcam, rabbit anti-goat from Dako, andfrom Santa Cruz.

a protein analysisMultiplex bead kit for cytokine measurement wasased from BioRad Laboratories, Inc. Standards for each mediator were generated, ranging from2,000 pg/mL. Plasma samples were incubated withof antibody-coupled microsphere sets for 1 hour attemperature. Freshly diluted secondary detectiondy (25 μL; 1 μg/mL) were added and incubated attemperature for 1 hour. Streptavidin-phycoerythrin; ×1) was added, followed by an incubation for

nutes at room temperature. After each step incuba-f samples with microsphere sets, secondary detec-ntibody, and streptavidin-phycoerythrin, a filteringnd three washing steps using a vacuum manifolddone. Each well was analyzed on a Bioplex ProteinSystem (BioRad Laboratories, Inc.) according to thefacturer's instructions. All protein concentrationsven in picograms (pg) per milliliter.

in extraction and Western blot analysisor tissues were homogenized in lysis solution

Signaling Technology) supplemented with sodiumde (10 μmol/L; Fisher) and phenylmethylsulfonyl

analogrowt

ancer Ther; 9(10) October 2010

on May 17, 2018. © 2010mct.aacrjournals.org ownloaded from

de (100 μg/mL; Sigma-Aldrich). After sonicationseconds, cell debris was removed by centrifugation000 g for 20 minutes at 4°C. Protein concentrationetermined by bicinchoninic acid protein assay re-(Pierce). Lysates were fractionated in either 8 orDS-polyacrylamide gels, and the separated pro-were transferred to nitrocellulose. The blots wered for the proteins of interest with specific anti-s followed by a second antibody-HRP conjugateen incubated with SuperSignal chemiluminescenceate (Pierce).

ulture8 and 786-O, two VHL-deficient human RCC cell20), were obtained from the American Type Culturetion and cultured for <1 month, and aliquots were. Fresh frozen aliquots were used for each experi-A498 was grown in Eagle's MEM. 786-O cells wereed in RPMI 1640 medium (Cellgro). All media wereemented with 2 mmol/L L-glutamine, 10% FCS,% streptomycin (50 μg/mL), and cells were cul-at 37°C with 5% CO2. Growth and morphologyth lines was observed and noted to be consistentprior descriptions of the lines; no further geneticcterization was done.

r xenograft inductions.c. xenograft tumor models, female athymicbeige mice (Charles River Laboratories) were used.ice were housed and maintained in laminar flow

ets under specific pathogen-free conditions. All ex-ents were approved by the Institutional Animaland Use Committee at Beth Israel Deaconess Med-enter.C cell lines were harvested from subconfluent cul-by a brief exposure to 0.25% trypsin and 0.02%. Trypsinization was stopped with medium con-g 10% fetal bovine serum, and the cells wereed once in serum-free medium and resuspendedS. Only suspensions consisting of single cells withviability were used for the injections.establish RCC tumor xenografts, 786-O or A498 tu-ells were injected s.c. (1 × 107 cells) into the flankso 8-week-old mice that were of 20 g average bodyt. Tumors developed in >80% of the mice andusually visible within a few days of implantation.nib (80 mg/kg; Bayer) or sunitinib (additive-free,g/kg; Pfizer) was administered 6 of 7 days perby gavage beginning when the tumors had growniameter of 12 mm as per Sabir et al. (21) and Schor-ch et al. (22). Tumors were measured daily whileatment, and the day of resistance was recorded.ance was defined as an increase in tumor diameterm from its pretreatment size of 12 mm. This dif-

e represents the smallest increase in size that couldproducibly measured by calipers and is roughly
gous to the clinical criteria for resistance (20%h by Response Evaluation Criteria for Solid Tumors)
Molecular Cancer Therapeutics



usedouslyin thissureddetermtumorrificedrespodescri

CXCLCXC

into tweeklceivedwas d12 mmsunitidaily,20 mdurinous tu

ImmuFor

frommensheatedin citroomdrogeperoxproteiwas afollow30 mDakorabbitdine cscopeusingTechnFor

tionsair-dride foanti-Iappliewas dHRP-lowed

TumoTum

magnouslyand trconstr

Interadifferimagesion aTo

drawnmor bthat wbloodest waA 16-incremflow vgreenperfu

Resu

Moduinhib786

descrinitini12 mmstabiltinuedcollecand ascreenwereseverastrominclud(P = 0treatetumortumortance.immucyte mand gouslySignifnecromentaTo

sion oin thesunitiof IFNof IFNsistan(Fig. 1can amunoregula

CXCL9 Delays Resistance to VEGFR Inhibition

www.a

D


in patients. Furthermore, this difference was previ-shown to be associated with restored angiogenesismodel (22). Tumor long and short axes were mea-

, and long axis and tumor volume were followed toine growth curves. Treatment was continued untils grew to 20 mm, at which point the mice were sac-. Tumor tissue was obtained pretreatment, duringnse, and at time of resistance for various analysesbed below.

9 administrationL9 (1 ug in 200 uL; R&D Systems) was injected

he central portion of tumor xenografts three timesy as described previously (18). Control mice re-injections of PBS, the vehicle in which the CXCL9issolved. Injections began when tumors reachedand were given either alone or concomitant with

nib as described above. Tumors were measuredand time to resistance (growth by 2 mm) and

m were ascertained. Tumor tissue was obtainedg response and at time of resistance for the vari-mor tissue analyses described below.

nohistochemistryCD34 analysis, 4-um-thick sections were preparedformalin-fixed, paraffin-embedded tumor speci-. Sections were deparaffinized, rehydrated, andwith a pressure cooker to 125°C for 30 seconds

rate buffer for antigen retrieval. After cooling totemperature, sections were incubated in 3% hy-n peroxide for 5 minutes to quench endogenousidase, and then for 20 minutes in Dako serum-freen block (Dako). The anti-CD34 antibody (Abcam)pplied at a 1:100 dilution to sections for 1 hour,ed by rabbit anti-goat secondary antibody for

inutes. Detection was done by incubating withEnVision+ System HRP-labeled polymer anti-for 30 minutes, followed by 3,3 ′-diaminobenzi-hromogen. Slides were scanned using the Scan-XT (Aperio Technologies Inc.) and analyzeda modified microvessel analysis algorithm (Aperioologies Inc.).IFNγR staining, frozen sections were used. Sec-were fixed in -20°C acetone for 5 minutes and thenied. Sections were incubated in 3% hydrogen perox-r 5 minutes to quench endogenous peroxidase. TheFNγR antibody (Santa Cruz Biotechnology) wasd at a 1:50 dilution to sections for 1 hour. Detectionone by incubating with Dako EnVision + Systemlabeled polymer anti-mouse for 30 minutes, fol-by 3,3 ′-diaminobenzidine chromogen.

r perfusion imagingor perfusion imaging [arterial spin labeled (ASL)etic resonance imaging (MRI)] was done as previ-described (22). ASL sequence raw data were saved
ansferred to the analysis workstation for image re-uction by using custom software written within the
micepared

acrjournals.org


ctive Data Language (Research Systems). The ASLence image, between average label and controls, was then converted to quantitative tumor perfu-s previously described (23).determine tumor perfusion, a region of interest wasfreehand around the peripheral margin of the tu-

y using an electronic cursor on the reference imageas then copied to the perfusion image. The meanflow for the tumor tissue within the region of inter-s derived, and image window and level were fixed.color table was applied in 10 mL/100 g/minuteents ranging from 0 to 160 mL/100 g/minute, withalues represented as varying shades of black, blue,, yellow, red, and purple, in order of increasingsion.

lts

lation of IFNγ signaling with VEGFRitor therapy-O–derived tumors were implanted into mice asbed in Materials and Methods. Treatment with su-b or sorafenib was initiated at a tumor size ofin long axis. Tumors exhibited a period of growth

ization followed by growth resumption despite con-therapy as previously described (22). Plasma was

ted from untreated mice (n = 12), on day 3 (n = 16),t resistance (n = 13). Twenty-seven cytokines wereed using human-specific panels and 23 cytokinesscreened using murine-specific panels. Changes inl cytokines were noted in both the tumor and thea at the time of resistance (Supplementary Table S1),ing a significant decrease in the production of IFNγ.018 for day 3 versus resistant, and 0.0120 for un-d versus resistant; Fig. 1A). Human cytokines arederived in this system; these findings suggest that-derived IFNγ may be downmodulated with resis-Other significant changes were observed in severalne cytokines, including decreased human granulo-acrophage colony-stimulating factor (GM-CSF)

ranulocyte colony-stimulating factor, and, as previ-reported, increased human IL-8 at resistance (24).icant changes in murine IL-3, IL-4, IL-13, tumorsis factor α, and GM-CSF were also noted (Supple-ry Fig. S1).further interrogate this cytokine pathway, expres-f IFNγR was measured in the tumors from miceabsence of therapy or with treatment with eithernib or sorafenib. There was abundant expressionγR in untreated tumors and a loss of expressionγR both at day 3 of therapy and at the time of re-ce in mice receiving either sorafenib or sunitinibB). These data suggest that a loss of IFNγ signalingccompany resistance to therapy. Furthermore, im-histochemical analysis showed dramatic down-tion of IFNγR expression in tumors derived from
that were treated with 3 days of sunitinib as com-with untreated tumors (Fig. 1C).
Mol Cancer Ther; 9(10) October 2010 OF3



To faccompressiwas atherapthree c(Fig. 2TKI tr

Chemof resThe

led towithwas i

comitcauseactivigrowtplusCXCLalonegrowttion otreate1.3 dathe aCXCL

Figure(218.5decreas3 of thefrom twtumor a ing is p

Bhatt et al.

Mol COF4

D


urther define the changes in the IFNγ pathway thatpany resistance to sorafenib and sunitinib, the ex-on of the angiostatic chemokines regulated by IFNγnalyzed. Western analysis of the tumors at day 3 ofy and at the time of resistance showed a loss of allhemokines (CXCL9–11) relative to untreated tumors). This is consistent with the hypothesis that VEGFReatment leads to a decrease in IFNγ signaling.

okine administration delays the developmentistance to sunitinibtreatment-induced loss of angiostatic chemokinesthe hypothesis that resistance could be delayed

s compared with a tumor that was treated with sunitinib for 3 days. Stain

chemokine replacement. To study this, CXCL9 (Fig.



ant with sunitinib gavage. CXCL9 was selected be-of the prior data that this chemokine exhibits

ty in murine RCC models (18). Figure 3 shows theh of tumors treated with vehicle gavage, sunitinibintratumoral PBS, sunitinib plus intratumoral9, or intratumoral CXCL9 alone. Although CXCL9and treatment with sunitinib alone slowed tumorh, this effect was further enhanced by the combina-f sunitinib and injections of CXCL9 (Fig. 3). Tumorsd with sunitinib plus PBS increased by 2 mm in 11.2 ±ys and this was extended to 18.9 ± 3.2 days withddition of CXCL9 (P = 0.001). Treatment with9 alone exhibited a similar effect as sunitinib alone

resent in the untreated tumor and absent in the treated tumor.

3). Thus, the early administration of CXCL9 withce to sunitinib.

ureCL9ayitinipaultsorsthe described conditions are shown.

njected into the 786-O tumors either alone or con- sunitinib delays resistan

FigCXat dsuncomRestumwith

1. A, plasma analysis of mice harboring tumors that were untreated (389.8 ± 118.6 pg/mL; n = 12 mice), at day 3 after initiation of sorafenib± 45.5 pg/mL; n = 16 mice), and that were resistant (90.76 ± 20.0 pg/mL; n = 13 mice). In the setting of resistance to sorafenib, plasma IFNγes (P = 0.018 for day 3 versus resistant and 0.012 for untreated versus resistant). B, Western analysis showing that IFNγR is downmodulated at dayrapy with sorafenib or sunitinib (D3) and at the time of resistance (R) as compared with untreated tumors (Un). Results from tumors harvestedo to three different mice treated with the described conditions are shown. C, a representative immunohistochemical stain for IFNγR in an untreated

Mo

American Associatio

2. Western analysis showing that, 10, and 11 are downmodulated3 of therapy (D3) with sorafenib orb and at the time of resistance (R) asred with untreated tumors (Un).from two to three representativeharvested from different mice treated

lecular Cancer Therapeutics

n for Cancer Research.


CXCLsunitTo

the eflyzedfromCXCLthe m20 mmshowCD34treatenitinibCXCLage Mwas 4PBS (PTo f

tumordone.perfusnib plplus Pof perof pertinibgreate3 (dayP = 0.0treate

reducweek(P =perfuperfuneverobserchemoverCXCL

CXCLxenogTo

therderivplus786-Osunitiity astreateshowpared(P =growtcontamicePBS,

Figure(n = 5),of tumo in the*, vehic , P = 0


www.a

D


9 treatment prolongs antiangiogenic effect ofinibassess the mechanism by which CXCL9 prolongsfect of sunitinib, 786-O–derived tumors were ana-at the time of sacrifice. Tumors were harvestedboth the sunitinib-plus-PBS and the sunitinib-plus-9 groups at the average time when the tumors inice treated with sunitinib plus PBS reachedin size (day 43 ± 2.4 posttreatment). Figure 4

s representative immunohistochemical sections ofassessment of microvessel density (MVD) of miced with vehicle, CXCL9, sunitinib plus PBS, and su-plus CXCL9 (Fig. 4A–D, respectively). Although

9 itself had no significant effect on MVD, the aver-VD of tumors treated with sunitinib plus CXCL90% lower than tumors treated with sunitinib plus= 0.014).

urther assess the angiogenic capability of the treateds, serial tumor perfusion imaging by ASL MRI wasFigure 5A shows a time course of tumor (786-O)ion in a mouse treated with vehicle, CXCL9, suniti-us PBS, or sunitinib plus CXCL9. Whereas sunitinibBS and sunitinib plus CXCL9 produced initial lossfusion compared with control followed by recoveryfusion at the time of resistance, treatment with suni-plus CXCL9 suppressed perfusion longer and to ar extent. Resumption of perfusionwas seen onweeks 22–26) and was significant by week 6 (days 42–46;

rs treated with sunitinib + CXCL9 exhibit a longer time to resistance shownle versus CXCL9, P = 0.016; §, sunitinib + PBS versus sunitinib + CXCL9

3) in sunitinib-plus-PBS-treatedmice, whereasmiced with sunitinib plus CXCL9 exhibited a greater

suniti0.001)

acrjournals.org


tion in tumor perfusion that was maintained at both3 (days 22–26) and week 6 (days 42–46) of therapy0.03 and 0.04, respectively). In contrast to tumorsion in mice treated with sunitinib plus PBS, tumorsion at time of resistance to sunitinib plus CXCL9attained the pretreatment perfusion levels. These

vations are consistent with the CD34 immunohisto-ical data. Tumor perfusion exhibited little changetime in mice treated either with PBS alone or9 alone (Fig. 5A and B).

9 restoration delays resistance in a secondraft modelconfirm these findings and extend them to ano-tumor model, mice bearing xenograft tumorsed from A498 cells were treated with sunitinibPBS or sunitinib plus CXCL9. As noted in the–derived tumors, CXCL9 when initiated withnib extended the duration of relative tumor stabil-measured by time to increase by 2 mm. Tumorsd with the combination of CXCL9 and sunitinibed a period of stability of 39.8 ± 6.1 days com-with 22.2 ± 4.1 days for sunitinib plus PBS

0.0008; Fig. 6A). CXCL9 itself also slowed tumorh but not to as great an extent as the sunitinib-ining treatments. Serial ASL MRI was done ontreated with vehicle, CXCL9 alone, sunitinib plusand sunitinib plus CXCL9. Both administration of

accompanying table (#, vehicle versus sunitinib + PBS, P = 0.002;.001). Comparisons were done by Student's t-test.

3. Growth curves depicting average tumor volume of 786-O–derived tumors from mice treated with vehicle gavage (n = 3), sunitinib (Su) + CXCL9sunitinib+ PBS (n = 5), or intratumoral CXCL9 (n = 5), with SE. As compared with vehicle-treated controls or sunitinib-treated mice, the growth

nib plus PBS (P ≤ 0.001) and CXCL9 alone (P =significantly reduced perfusion relative to vehicle




(Suppthat tin a gtumorplus0.001)perfureturnIFNγRtumorto 786exhibsistan

of anfor ea

Discu

Treand sresistmodeops apathw

FigureD, suniCXCL9sacrific on of

Bhatt et al.

Mol COF6

D


lementary figure). Representative images showhe combination of sunitinib and CXCL9 resultedreater reduction in tumor blood flow and lowerperfusion at all time points relative to sunitinib

PBS (Fig. 6B and Supplementary figure, P =. As in 786-O tumors, ASL MRI–measured tumorsion in sunitinib-plus-CXCL9-treated mice nevered to the pretreatment perfusion level. Levels ofand CXCL9 were downmodulated in the A498

s with exposure to sunitinib (Fig. 6C). In contrast-O tumors, however, the majority of A498 tumors

ed at 20 mm. E, quantification of average MVD. P = 0.014 for the comparis

ited increases in these molecules at the time of re-ce to sunitinib. This early, but not sustained, loss

Our dtance



giostatic chemokines is consistent with the needrly supplementation of sunitinib with CXCL9.

ssion

atment of patients with the VEGFR TKIs sorafenibunitinib can lead to periods of tumor stability, butance to therapy is inevitable. We used a mousel to define mechanisms by which resistance devel-nd found that components of the IFNγ signalingay are lost with sunitinib or sorafenib therapy.

sunitinib + PBS versus sunitinib + CXCL9. Scale bar, 100 μmol/L.

4. Tumor CD34 immunohistochemistry at the time of sacrifice is shown in the following conditions: A, vehicle. B, CXCL9. C, sunitinib + PBS.tinib + CXCL9. Mice treated with sunitinib + PBS (n = 3) were sacrificed when they reached 20 mm in long axis, and mice treated with sunitinib +(n = 4) were sacrificed at the average day that the sunitinib treated mice were sacrificed. Mice treated with vehicle (n = 3) and CXCL9 alone (n = 3) were

ata also show that CXCL9 treatment delays resis-to sunitinib in 786-O- and A498-derived tumors




Figureby ASLshownThe mictreatedgreater$, P = 0


www.a

D


5. A, serial tumor perfusion in a representative tumor treated with vehicle as control, CXCL9 alone, sunitinib+ PBS, or sunitinib + CXCL9 as measuredMRI. Red lines, regions of tumor. The tumor size was measured with long and short axes (in mm) and the mean blood flow (in mL/100 g/minute) arebelow each image. Color scale, range of perfusion values from 0 to 160 mL/100 g/minute. B, average perfusion with SE (n ≥ 3 mice in all arms).e treated with sunitinib + PBS exhibited decreased perfusion that began to resume by week 3 (*, P = 0.04 for comparison of vehicle versus sunitinibmice; £, P = 0.03 for comparison of day 3 versus day 45 of therapy with sunitinib + PBS). In contrast, mice treated with sunitinib + CXCL9 exhibited a
reduction in tumor perfusion than with sunitinib + PBS that was maintained at week 3 (days 22–26) and week 6 (days 42–46; ¶, P = 0.03 and.04, respectively).
Mol Cancer Ther; 9(10) October 2010acrjournals.org OF7

on May 17, 2018. © 2010 American Association for Cancer Research. mct.aacrjournals.org ownloaded from


FigureSE (n =time tosunitiniCXCL9vehiclewas mevaluessunitini

Bhatt et al.

Mol COF8

D


6. A, growth curves of average volume of A498-derived xenograft tumors treated with sunitinib + PBS, sunitinib + CXCL9, CXCL9, or vehicle, with5 in all arms). As compared with untreated controls or sunitinib-treated mice, the growth of tumors treated with sunitinib + CXCL9 exhibit a longergrow by 2 mm shown in the accompanying table and a prolonged time of overall tumor growth (days of tumor growth from 12 mm to 14 mm:b + PBS versus sunitinib + CXCL9; §, P = 0.0008). CXCL9 also slowed tumor growth but to a lesser extent than sunitinib + CXCL9 (*, P < 0.0001 forversus vehicle; #, P < 0.0001 for CXCL9 versus sunitinib + CXCL9). B, representative set of perfusion images from a set of mice treated withas control, CXCL9 alone, sunitinib + PBS, or sunitinib + CXCL9 (representative of 3 mice per arm). Red lines, regions of tumor. The tumor sizeasured with long and short axes (in mm) and the mean blood flow (in mL/100 g/minute) are shown below each image. Color scale, range of perfusion
from 0 to 160 mL/100 g/minute. C, Western analysis for IFNγR and CXCL9 in A498 tumors in representative untreated tumors (Un), day 3 ofb (D3), and at the time of resistance (R). Comparisons were done by Student's t-test.
ancer Ther; 9(10) October 2010 Molecular Cancer Therapeutics

on May 17, 2018. © 2010 American Association for Cancer Research. mct.aacrjournals.org ownloaded from


and tproloThesepressefor thIn t

go exand amorsthe inpose tand dsignalsuch mmoleclinkedThus,likelyWit

seemevent s(by dato ourminisCXCLchangthougantianof angmetastherasignalCXC

tumortumorRCCgrowtwith sing thin patexplosplicefor diCXCLtion vour epossibto thesee evtumorCXCLmunonotedtreatmquireding anTo m

pressi

out tuhere tsion,to asspy. Othat icompmals aanalytokineare adgiogecouldwhichhallmtumorsion aour grmentperfutreatmindepVEGFmor mto theThe

CXCLmentthesein unmorscatestumorly trupathwCXCLever, parrestof tumtreatethat tVEGFand ttissuecirculantagpressiprediand that restors cthe prsuchAlt

conceCXCR


www.a

D


hat one mechanism by which this occurs is byngation of the antiangiogenic effects of sunitinib.data suggest that angiostatic pathways are sup-d as a result of VEGFRTKI therapy and set the stagee subsequent development of resistance to therapy.he setting of VEGFR inhibition, RCC tumors under-tensive necrosis (22). In the setting of this necrosisccompanying hypoxia/nutrient deprivation, tu-may undergo compensatory changes, includingduction of salvage angiogenesis pathways. We pro-hat the environmental stress resultant from a rapidramatic reduction in tumor vasculature and VEGFRing is particularly conducive to the development ofolecular changes. We hypothesize that one of these

ular changes is the loss of IFNγR and that this loss isto the downmodulation of angiostatic chemokines.this revascularization of VEGF-deprived tumors isphysiologically distinct from de novo angiogenesis.hin just a few days of therapy, RCC xenograftsd particularly vulnerable to agents that could pre-alvage angiogenesis. In fact, we show an early lossy 3 of treatment) of IFNγR and CXCL9, which ledhypothesis that treatment with CXCL9 must be ad-tered early. Thus, we administered sunitinib and9 concurrently in an effort to overcome possiblees resulting from this early loss of angiostasis. Al-h other recently published studies have shown thatgiogenic therapy can lead to increased productioniogenic molecules and increased invasiveness andtatic potential (25, 26), we show that antiangiogenicpy also leads to downmodulation of angiostaticing.R3, the receptor for CXCL9, is expressed on bothcells and endothelium. CXCR3 signaling in othertypes has proinvasive properties (27). By contrast,

xenografts in our models did not show acceleratedh compared with untreated tumors when injectedingle-agent CXCL9. This is consistent with the find-at CXCR3 expression confers a favorable prognosisients with localized RCC (17). Future studies couldre CXCR3 expression in the tumors, including thed subtype expression as well as the potential rolerect effects of CXCL9 on tumor cells. Additionally,9 has been shown to function in leukocyte activa-ia its interaction with CXCR3 on Th1 cells. Althoughxperiments used immunocompromised mice, theility remains that CXCL9 also recruits immune cellstumor, but in preliminary experiments we did notidence of increased lymphocyte recruitment to thes (data not shown). Exclusion of an immune role for9 would likely require studies in more severely im-deficient mice such as RAG KO mice. Finke et al.an increase in type I IFNγ producing cells afterent with sunitinib (28). Further studies will be re-to understand the relationship between this find-d the loss of IFNγR signaling that we have seen.
easure tumor vasculature, we analyzed tumor ex-
on of CD34 by immunohistochemistry and carriedto clinlation

acrjournals.org


mor perfusion imaging using ASL MRI. We showhat CD34 staining correlates with ASL MRI perfu-giving us confidence to use this imaging modalityess response and resistance to antiangiogenic thera-ne noteworthy advantage of ASL MRI over CD34 ist enables serial imaging of tumors as opposed toaring distinct tumors removed from different ani-t different time points. Although plasma biomarkerses have shown that there are changes in plasma cy-s that may predict for antitumor activity (29), therevantages to having a visual representation of the an-nic status of a tumor in an individual patient. Thisprovide early clues to the timing and means byescape from therapy occurs. Our data show that aark of sunitinib treatment is dramatically reducedperfusion followed by a restoration of tumor perfu-ccompanied by tumor regrowth. Prior work fromoup has shown similar findingswith sorafenib treat-(22). We did note, however, that the restored tumorsion at the time of resistance never reaches the pre-ent levels. This finding suggests that angiogenesis-endent factors may also contribute to resistance toR blockade. For example, metabolic changes in a tu-ay occur that limit its oxygen requirements relativetreatment-naïve setting (30).biological significance of the downmodulation of9 and IFNγR noted shortly after beginning treat-with a VEGFR antagonist is unclear. The fact thatangiostatic molecules are expressed by tumor cellstreated mice and even overexpressed in some tu-that have developed resistance to sunitinib indi-that their presence is not an absolute deterrent togrowth and vascularization. This may be especial-

e in the resistant setting, in which new angiogenicays may dominate over the angiostatic forces.9 initiated concurrently with the sunitinib, how-rolonged the duration of sunitinib-induced growthand delayed the revascularization and resumptionor perfusion that otherwise rapidly ensues in miced with sunitinib alone. This observation indicateshe early phases of tissue remodeling induced byR blockade are affected by the presence of CXCL9hat the disappearance of this chemokine from thefacilitates the re-establishment of the tumor micro-ation and the development of resistance to VEGFRonists. Thus, the early loss of CXCL9 and IFNγR ex-on seen in the A498 tumors is likely the dominantctor of the utility of chemokine supplementatione observation that these molecules are re-expressedistance may indicate that other proangiogenic fac-an be sufficient to mediate tumor growth even inesence of enhanced expression of angiostatic factorsas CXCL9.hough intratumoral CXCL9 injection validates thept that sunitinib resistance can be delayed with3 ligands, this method of delivery is not amenable
ical practice. Future studies that would enable trans-to the clinical setting could involve the application



of a CsystempatienAlthobe abCXCLmaintkinesWe arThe

regresably tcologsuggemechablockepathwappro

sorafepossib

Discl

No p

Grant

DF/Grant,Medicacollabofrom th

Thepaymenadvertisthis fac

Refe1. Ka

oxy2. Mo

a mtorme

3. Mome

4. Moin m

5. EsSo20

6. Rincin

7. Beap

8. Caevapa

9. Mizsercel

10. Matum261

11. Yuon

12. Shtre200

13. StrCaCa

14. Kocara T

15. Koprecel

16. Suferattrec

Bhatt et al.

Mol COF10

D


XCR3 agonist. Additionally, the potential use ofic CXCL9 could be explored as could treatment ofts with IFNγ or IFNγ-inducing agents such as IL-12.ughwe find that tumor cells lose IFNγRandmaynotle to respond to IFNγ treatment by producing9-11, it is possible that other nontumor cell typesain the ability to upregulate the angiostatic chemo-in response to IFNγ despite antiangiogenic therapy.e currently exploring this issue.inability to sustain the initial tumor stabilization orsion induced by VEGF pathway blockers is argu-he most vexing problem now encountered by on-ists who care for patients with RCC. Our studiesst that relative loss of CXCL9 is one of the molecularnisms that curtail the initial effectiveness of VEGFRrs. Future studies with augmentation of angiostatic
ays might lead to the elucidation of therapeutic Rece
on γ (IFN-γ)-inducible chemokines IFN-inducible T-cell α chemo-ractant and monokine induced by IFN-γ and of their receptor CXCeptor 3 in human renal cell carcinoma. Cancer 2005;103:258–67.

17. KlaCXcle

18. Paicano

19. Areprosmtas

20. Iliothe

21. SaeaAJ

22. SclabceRa

23. AlsmaBlo

24. Huan20

25. Pàmatan

26. EbKea p

27. ZipinvCa

28. Finsuppa

29. Ebfacgrogro

30. Araofba



nib, or other VEGF pathway blockers in RCC andly other tumors.

osure of Potential Conflicts of Interest

otential conflicts of interest were disclosed.

Support

HCCKidney Cancer SPORE: 2 P50 CA101942 VHL Family AllianceClinical Investigator Training Program: Beth Israel Deaconessl Center - Harvard/MIT Health Sciences and Technology, inration with Pfizer Inc. and Merck & Co (R.S. Bhatt), and a grante National Comprehensive Cancer Network, Jenkintown PA.costs of publication of this article were defrayed in part by thet of page charges. This article must therefore be hereby markedement in accordance with 18 U.S.C. Section 1734 solely to indicatet.

ived 05/26/2010; revised 07/23/2010; accepted 07/29/2010;
aches that extend the effectiveness of sunitinib, published OnlineFirst 08/10/2010.
renceselin WG, Jr. The von Hippel-Lindau protein, HIF hydroxylation, andgen sensing. Biochem Biophys Res Commun 2005;338:627.tzer RJ, Michaelson MD, Redman BG, et al. Activity of SU11248,ultitargeted inhibitor of vascular endothelial growth factor recep-and platelet-derived growth factor receptor, in patients withtastatic renal cell carcinoma. J Clin Oncol 2006;24:16–24.tzer RJ, Rini BI, Bukowski RM, et al. Sunitinib in patients withtastatic renal cell carcinoma. JAMA 2006;295:2516–24.tzer RJ, Hutson TE, Tomczak P, et al. Sunitinib versus interferon αetastatic renal-cell carcinoma. N Engl J Med 2007;356:115–24.

cudier B, Eisen T, Stadler WM, et al. TARGET Study Group.rafenib in advanced clear-cell renal-cell carcinoma. N Engl J Med07;356:125–34.i BI, Atkins MB. Resistance to targeted therapy in renal-cell car-oma. Lancet Oncol 2009;10:992–1000.rgers G, Hanahan D. Modes of resistance to anti-angiogenic ther-y. Nat Rev Cancer 2008;8:592–603.sanovas O, Hicklin D, Bergers G, Hanahan D. Drug resistance bysion of antiangiogenic targeting of VEGF signaling in late-stagencreatic islet tumors. Cancer Cell 2005;8:4:299–309.ukami Y, Jo WS, Duerr EM, et al. Induction of interleukin-8 pre-ves the angiogenic response in HIF-1α-deficient colon cancerls. Nat Med 2005;11:992–7.ncuso MR, Davis R, Norberg SM, et al. Rapid vascular regrowth inors after reversal of VEGF inhibition. J Clin Invest 2006;116:0–21.JL, Rak JW, Coomber BL, Hicklin DJ, Kerbel RS. Effect of p53 statustumor response to antiangiogenic therapy.Science2002;295:1526–8.ojaei F, Wu X, Malik AK, et al. Tumor refractoriness to anti-VEGFatment is mediated by CD11b+Gr1+ myeloid cells. Nat Biotechnol7;25:911–20.ieter RM,BurdickMD,Mestas J,GompertsB,KeaneMP,Belperio JA.ncer CXC chemokine networks and tumour angiogenesis. Eur Jncer 2006;42:768–78.ndo T, Nakazawa H, Ito F, et al. Favorable prognosis of renal cellcinoma with increased expression of chemokines associated withh1-type immune response. Cancer Sci 2006;97:780–6.ndo T, Ito F, Nakazawa H, Horita S, Osaka Y, Toma H. High ex-ssion of chemokine gene as a favorable prognostic factor in renall carcinoma. J Urol 2004;171:2171–5.yama T, Furuya M, Nishiyama M, et al. Up-regulation of the inter-

tte T, Seligson DB, Leppert JT, et al. The chemokine receptorCR3 is an independent prognostic factor in patients with localizedar cell renal cell carcinoma. J Urol 2008;179:61–6.n J, Burdick MD, Belperio JA, et al. CXCR3/CXCR3 ligand biolog-l axis impairs RENCA tumor growth by a mechanism of immu-angiostasis. J Immunol 2006;176:1456–64.nberg DA, Kunkel SL, Polverini PJ, et al. Interferon-γ-inducibletein 10 (IP-10) is an angiostatic factor that inhibits human non-all cell lung cancer (NSCLC) tumorigenesis and spontaneous me-tases. J Exp Med 1996;184:981–92.poulos O, Kibel A, Gray S, Kaelin WG, Jr. Tumour suppression byhuman von Hippel-Lindau gene product. Nat Med 1995;1:822–6.bir A, Schor-Bardach R, Wilcox CJ, et al. Perfusion MDCT enablesrly detection of therapeutic response to antiangiogenic therapy.R Am J Roentgenol 2008;191:133–9.hor-Bardach R, Alsop DC, Pedrosa I, et al. Does arterial spin-eling MR imaging-measured tumor perfusion correlate with renalll cancer response to antiangiogenic therapy in a mouse model?diology 2009;251:731–42.op DC, Detre JA. Reduced transit-time sensitivity in noninvasivegnetic resonance imaging of human cerebral blood flow. J Cerebod Flow Metab 1996;16:1236–49.ang D, Ding Y, Zhou M, et al. Interleukin-8 mediates resistance totiangiogenic agent sunitinib in renal cell carcinoma. Cancer Res10;70:1063–71.ez-Ribes M, Allen E, Hudock J, et al. Antiangiogenic therapy elicitslignant progression of tumors to increased local invasion and dis-t metastasis. Cancer Cell 2009;15:220–31.os JM, Lee CR, Cruz-Munoz W, Bjarnason GA, Christensen JG,rbel RS. Accelerated metastasis after short-term treatment withotent inhibitor of tumor angiogenesis. Cancer Cell 2009;15:232–9.in-Roitman A, Meshel T, Sagi-Assif O, et al. CXCL10 promotesasion-related properties in human colorectal carcinoma cells.ncer Res 2007;67:3396–405.ke JH, Rini B, Ireland J, et al. Sunitinib reverses type-1 immunepression and decreases T-regulatory cells in renal cell carcinoma

tients. Clin Cancer Res 2008;14:6674–82.os JM, Lee CR, Bogdanovic E, et al. Vascular endothelial growthtor-mediated decrease in plasma soluble vascular endothelialwth factor receptor-2 levels as a surrogate biomarker for tumorwth. Cancer Res 2008;68:521–9.gonés J, Schneider M, Van Geyte K, et al. Deficiency or inhibitionoxygen sensor Phd1 induces hypoxia tolerance by reprogramming
sal metabolism. Nat Genet 2008;40:170–80.



Published OnlineFirst August 10, 2010.Mol Cancer Ther Rupal S. Bhatt, Xiaoen Wang, Liang Zhang, et al. by Restoration of Angiostatic SignalingRenal Cancer Resistance to Antiangiogenic Therapy Is Delayed

Updated version

10.1158/1535-7163.MCT-10-0477doi:

Access the most recent version of this article at:

Material

Supplementary

http://mct.aacrjournals.org/content/suppl/2010/08/11/1535-7163.MCT-10-0477.DC1

Access the most recent supplemental material at:

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

Subscriptions

Reprints and

[email protected] at

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

Permissions

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

.http://mct.aacrjournals.org/content/early/2010/09/17/1535-7163.MCT-10-0477To request permission to re-use all or part of this article, use this link

on May 17, 2018. © 2010 American Association for Cancer Research. mct.aacrjournals.org Downloaded from


http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-10-0477

http://mct.aacrjournals.org/content/suppl/2010/08/11/1535-7163.MCT-10-0477.DC1

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

mailto:[email protected]

http://mct.aacrjournals.org/content/early/2010/09/17/1535-7163.MCT-10-0477


therapeutics renal cancer resistance to...

Documents