Cabozantinib Inhibits Prostate Cancer Growth and Prevents ... · based on retrospective review. Among the patients with stable disease at week 12 that were randomly assigned to cabozantinib

Cancer Therapy: PreclinicalSee related article by Lee and Smith, p. 525

Cabozantinib Inhibits Prostate Cancer Growth and PreventsTumor-Induced Bone Lesions

Jinlu Dai1, Honglai Zhang1, Andreas Karatsinides1, Jill M. Keller1, Kenneth M. Kozloff2, Dana T. Aftab3,Frauke Schimmoller3, and Evan T. Keller1

AbstractPurpose: Cabozantinib, an orally available multityrosine kinase inhibitor with activity against mesen-

chymal epithelial transition factor (MET) and VEGF receptor 2 (VEGFR2), induces resolution of bone scan

lesions in men with castration-resistant prostate cancer bone metastases. The purpose of this study was to

determine whether cabozantinib elicited a direct antitumor effect, an indirect effect through modulating

bone, or both.

ExperimentalDesign:Usinghumanprostate cancer xenograft studies inmice,wedetermined the impact

of cabozantinib on tumor growth in soft tissue and bone. In vitro studies with cabozantinib were performed

using (i) prostate cancer cell lines to evaluate its impact on cell growth, invasive ability, and MET and (ii)

osteoblast cell lines to evaluate its impact on viability and differentiation and VEGFR2.

Results: Cabozantinib inhibited progression of multiple prostate cancer cell lines (Ace-1, C4-2B, and

LuCaP 35) in bone metastatic and soft tissue murine models of prostate cancer, except for PC-3 prostate

cancer cells inwhich it inhibited only subcutaneous growth.Cabozantinib directly inhibited prostate cancer

cell viability and induced apoptosis in vitro and in vivo and inhibited cell invasion in vitro. Cabozantinib had

a dose-dependent biphasic effect onosteoblast activity and inhibitory effect on osteoclast production in vitro

that was reflected in vivo. It blocked MET and VEGFR2 phosphorylation in prostate cancer cells and

osteoblast-like cells, respectively.

Conclusion: These data indicate that cabozantinib has direct antitumor activity, and that its ability to

modulate osteoblast activity may contribute to its antitumor efficacy. Clin Cancer Res; 20(3); 617–30.

�2013 AACR.

IntroductionMore than 80% of men with advanced prostate cancer

develop bone metastases (1). The appearance of bonemetastasis in men with advanced prostate cancer is associ-ated with compromised quality of life (QOL) and is aharbinger of death. Skeletal metastases result in skeletal-related events (SRE; fracture, spine compression and insta-bility, decreased mobility, pain, and hypercalcemia),immunosuppression, and anemia. Skeletal metastatic painis a problem in almost all patients and greatly impacts theQOL of a patient (2, 3). Both bisphosphonates and deno-sumab [via receptor activator of NF-kB ligand (RANKL)inhibition] have been demonstrated to decrease SRE andimproveQOL inpatientswith bonemetastases but havenot

shown a significant survival benefit (4). Thus, it is criticalthat we continue to define mechanisms that promote bonemetastasis to identify key targets to not only further enhanceQOL but also improve survival.

Because their importance in cell signaling and cancerprogression protein kinases have been explored for theirroles as anticancer targets. MET is a receptor tyrosine kinase,expressed in epithelial and endothelial cells (reviewed inref. 5). Under normal circumstances, MET is activated byhepatocyte growth factor (HGF) that is produced by stromalcells, such as fibroblasts, thus generating a paracrine acti-vation loop. MET has been found to be highly expressed inprostate cancer compared with benign prostate hyperplasiaand significantly correlated with higher tumor histologygrade (6). Another kinase that contributes to cancer pro-gression is VEGF receptor (VEGFR), which plays an impor-tant role in the progression of metastasis through its abilityto promote angiogenesis upon activation by VEGF. Inaddition, the VEGF pathway has been shown to promotethe development of osteoblastic bone lesions in prostatecancer (7, 8). Intriguingly, there is crosstalk between VEGFandMET pathways. VEGF induces phosphorylation of METand thus activates the HGF/MET pathway in prostate cancer(9). Taken together, these data demonstrate both the impor-tance of MET and VEGF pathways in prostate cancer and

Authors' Affiliations: Departments of 1Urology and 2OrthopaedicSurgery,University of Michigan, Ann Arbor, Michigan; and 3Exelixis Inc., South SanFrancisco, California

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

Corresponding Author: Evan T. Keller, Room no. 5308 CCGC, Universityof Michigan, 1500 E. Medical Center Drive, Ann Arbor, MI 48109-8940.Phone: 734-615-0280; Fax: 734-764-3013; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-13-0839

�2013 American Association for Cancer Research.

ClinicalCancer

Research

www.aacrjournals.org 617

on February 23, 2021. © 2014 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 4, 2013; DOI: 10.1158/1078-0432.CCR-13-0839

http://clincancerres.aacrjournals.org/

their ability to crosstalk indicating that targeting both ofthese pathways may have a greater benefit than targetingeach pathway individually.

Cabozantinib (XL184) is an orally bioavailable tyrosinekinase inhibitor with activity primarily against MET andVEGFR2 as well as other tyrosine kinases (10). Cabozanti-nib was tested in a phase II randomized discontinuationtrial in men with metastatic castration–resistant prostatecancer (CRPC; ref. 11). Cabozantinib treatment resulted inthe regression of soft tissue lesions in 72% of patientsevaluable for change in measurable disease. In addition,68% of evaluable patients had improvement on bone scan,including 12% with complete resolution. In addition, painimproved in 67%of evaluable patients with pain at baselinebased on retrospective review. Among the patients withstable disease at week 12 that were randomly assigned tocabozantinib or placebo, median PFS was 23.9 weeks withcabozantinib and 5.9 weeks with placebo. Although theseresults are promising, it remains to be better understoodwhether cabozantinib affects only the cancer or the bonecells, or both. This knowledgewill help to better understandthe mechanism of the activity of cabozantinib in prostatecancer skeletal metastases and to determine whether tumorinhibition occurs in conjunction with the response obs-erved in bone scan lesions. Accordingly, the purpose of thisstudy was to determine the efficacy of cabozantinib inmodels of prostate cancer bone metastasis and determinewhether therewas anantitumor effect, a boneeffect, or both.

Materials and MethodsAnimals

Six-week-old male severe combined immunodeficient(SCID) mice (Charles River Laboratories) were housed

under pathogen-free conditions in accordancewith theNIHguidelines using an animal protocol approved by the Uni-versity of Michigan Animal Care and Use Committee (AnnArbor, MI).

Cell cultureHuman prostate cancer cell lines LNCaP and PC3 were

obtained from the American Type Culture Collection(ATCC) and cultured in RPMI-1640 (Invitrogen Co.). C4-2B cells (UroCor), which are LNCaP sublines selected togrow in bone (12), were grown in T medium. Humanprostate cancer cell line PC-3, a spontaneously immortal-ized cell line derived fromahumanvertebral prostate cancermetastasis, was obtained from the ATCC and cultured inRPMI-1640 (Invitrogen Co.). The canine Ace-1luc prostatecancer cell line (kindly provided by Dr. Tom Rosol, TheOhio State University, Columbus, OH) was derived from aspontaneous dog prostate carcinoma (13) and was main-tained at 37�C in Dulbecco’s Modified Eagle Medium/Ham’s nutrient mixture F12 (DMEM/F12). The LuCaP35human prostate cancer androgen–responsive xenograft(ref. 14; kindly provided by Dr. Robert Vessella, Universityof Washington, Seattle, WA) was maintained as a xenograftin SCIDmice. TheMC3T3-E1 (cloneMC-4; kindly providedby Dr. Renny Franceschi, University of Michigan, AnnArbor, MI) consists of preosteoblasts derived from murinecalvariae that, when treated with ascorbate, express osteo-blast-specific markers and are capable of producing a min-eralized matrix (15) was routinely maintained in a-MEM.ST-2 cells, mouse bone marrow stromal cell line, wereobtained from RIKENCell Bank (Ibaraki, Japan) andmain-tained in a-MEM. C4-2B and PC3 cells infected with thepLazarus retroviral construct expressing luciferase wereselected for stable transfectants in G418.

Cell viabilityCell viability was measured using WST-1 assays (Roche

Applied Science) as directed by the manufacturer.

Evaluation of osteoclastogenesis and osteoclastactivity

RAW 264.7 mouse macrophage/monocytes (ATCC) wereseeded in 96-well plates (103 cells/well) and allowed toattach to bovine bone slices in 96-well tissue culture plates.The culture medium was supplemented with 100 ng/mLrecombinantmurineRANKL. The cellswere incubated7daysat which time indicated levels of cabozantinib were added.The supernatants were collected at day 10 were stored at�80�C until analysis of tartrate-resistant acid phosphatase(TRACP)5b and carboxy-terminal collagen crosslinks (CTX).

TRACP 5b and CTX measurementsSecreted TRACP 5b was determined from the culture

medium using MouseTRAP assay (TRACP 5b) ELISA (IDSInc.). Secreted TRACP 5b accurately reflects the numberof osteoclasts formed in each well during the differen-tiation period. CTX, a measure of bone collagen degrada-tion products, was measured using Beta-Crosslaps (bCTx)

Translational SignificanceCabozantinib is a multikinase inhibitor that has the

greatest impact onMET, VEGFR2, and rearranged duringtransfection oncogene (RET) compared with otherkinases. In clinical trials, it has shown unprecedentedresolution of bone metastatic lesions, based on bonescans, inmenwith prostate cancer.However, it is unclearin these studies whether cabozantinib directly impactsthe cancer or indirectly impacts the cancer throughmodulating the bone. In this study, a bedside-to-benchevaluation of cabozantinib’s mechanism of action wasperformed. A combination of in vitro studies and in vivomurinemodels revealed that in addition to having directimpact on prostate cancer in both soft tissue and bonemetastases, cabozantinib had a biphasic impact onbone. Cabozantinib targeted both MET and VEGFR2 inprostate cancer and osteoblast cells, respectively. Theseresults reveal that targeting both the tumor and the bonemicroenvironment can have an important therapeuticimpact on prostate cancer bone metastases.

Dai et al.

Clin Cancer Res; 20(3) February 1, 2014 Clinical Cancer Research618




ELISA (Uscn Life Science Inc.). A resorption index dem-onstratingmean osteoclast activity was calculated by divid-ing the obtained resorption volume (CTX value) with thenumber of osteoclasts (TRACP 5b value), as describedpreviously (16).

Caspase-3/7 assayActivity of caspase-3/7 was determined using Apo-ONE

HomogeneousCaspase-3/7AssayKits (Promega) followingthe manufacturer’s instructions.

Matrigel invasion assayThe Matrigel invasion assay was perfomed using BD-

Biocoat InvasionChambers (BDBiosciences), as previouslydescribed (17).

Prostate-specific antigen measurementTotal prostate-specific antigen (PSA) levels in serum or

culture supernatants were determined using the HumanPSA ELISA Kit (Abazyme LLC) as described by the manu-facturer. The sensitivity of this assay is 1 ng/mL of PSA. Forculture supernatants, values were normalized to cell num-bers as determined by the modified WST-1 assay.

Bone remodeling assaysAlkalinephosphatase (ALP) activity in cellswasmeasured

using SensoLyte pNPP Alkaline Phosphatase Assay Kit(AnaSpec Inc.), as directed by the manufacturer. Osteocal-cin was measured using mouse-specific ELISA, as recom-mended by the manufacturer (Biomedical TechnologiesInc.). ALP activity and osteocalcin were normalized to totalprotein content determinedwithbicinchoninic acidproteinassay reagent (Thermo Scientific). The deposition of calci-um in cells was quantitated by the Calcium Assay Kit(Cayman), as directed by the manufacturer. Serum TRACP5b was measured using Mouse TRACP 5b assay (IDS Inc.)and serumprocollagen IN-terminal propeptide (PINP)wasmeasured using the Mouse PINP ELISA Kit (IDS Inc.) asdirected by the manufacturer.

Immunoblot analysisFor identification of MET, VEGFR2, AKT, and ERK1/2

phosphorylation, whole cell lysates were prepared by incu-bating cells in ice-cold RIPA lysis buffer (Millipore). Lysateswere precleared and the protein concentration was deter-mined by the bicinchoninic acid assay (Pierce Biochem-icals). For electrophoresis, lysates were supplemented withSDS loading buffer and separated on SDS-8% polyacryl-amide gels. Proteinswere transferred tonitrocellulosemem-branes. The blots were incubated in TBS containing 0.1%Tween 20 and 5% bovine serum albumin (BSA) during theblocking and the antibody incubation steps, followedby Western blot analysis with rabbit anti-human p-METmonoclonal antibody (mAb; 1:1,000; Invitrogen), rabbitanti-human MET polyclonal antibody (1:1000; SantaCruz Biotechnology), rabbit anti-human p-AKT mAb, rab-bit anti-human p-ERK1/2 mAb (1:1,000; Cell SignalingTechnology), rabbit anti-human AKT mAb, rabbit anti-

human ERK1/2 mAb (1:1000; Cell Signaling Technology),rabbit anti-mouse p-VEGFR2 polyclonal Ab (1:1,000; CellApplication Inc.), rabbit anti-mouse VEGFR2 polyclonal Ab(1:1,000, Santa Cruz Biotechnology), mouse anti-humanglyceraldehydes-3-phosphate dehydrogenase (GAPDH)mAb (1:5,000; Millipore). The antibody binding was rev-ealed using an HRP-conjugated anti-rabbit IgG (1:3,000,Cell Signaling Technology), or anti-mouse IgG (1:3,000;Amersham Pharmacia Biotech). Antibody complexes weredetected by SuperSignal West Pico Chemiluminescent Sub-strate, or SuperSignal West Dura Extended Duration Sub-strate (Thermo Scientific) and exposure to X-Omat film(Kodak). Densitometric analyses for protein quantificationwere done using ImageJ 1.38� software.

ImmunohistochemistryTibiae and subcutaneous tumors were fixed in 10%

neutral buffed formalin for 24 hours. The tibiae werethen decalcified for 48 hours in 10% EDTA and then bothtibiae and subcutaneous tumors were processed for par-affin embedding. Of note, 5-mm sections were used forhematoxylin and eosin (H&E) and IHC. Nonstained sec-tions were deparaffinized and rehydrated then stainedfor the indicated antigens including Ki 67, caspase-3/7,CD31, CD45, p-MET (pY1349; Novus Biologicals), MET(sc-161; Santa Cruz Biotechnology), VEGFR2 (Flk1, sc-504; Santa Cruz Biotechnology), and p-VEGFR2 (Abcam).The percentage of cells staining positive in each samplewas determined in sections by counting the positive cellsin 100 cells in 3 separate random sections of each slideat �40.

Bone histomorphometric analysisThe analysis was performed as we previously described

(18). Briefly, 3-mm sections of the mouse tibiae that wereprocessed for immunhistochemistry were stained withH&E or TRACP and counterstained with hematoxylin.Osteoblast number [Ob.No./mm] and osteoclast number[Oc.No./mm] were determined in trabecular bone, 0.25mm from the growth plate. Nine discontinuous randomregions of interest were examined within each bone torepresent the bone fragment. Two sections were analyzedper bone, usingBIOQUANT system(R&MBiometrics, Inc.).

Animal studiesAll experimental animal procedureswere approvedby the

University of Michigan Committee for the Use and Care ofAnimals (Ann Arbor, MI). For PC-3 tumor subcutaneousstudies, single-cell suspensions (1 � 106 cells) of PC-3luc

cells in RPMI media were injected in the flank at 100 mL persite using a 27-G 3/8-inch needle. Mice were randomizedto receive either cabozantinib (n ¼ 12; 60 mg/kg bodyweight/d, oral administration) or vehicle (n ¼ 12, distilledwater) for 15 days once tumors were established at 35 daysafter tumor inoculation. The dose of 60 mg/kg was chosenbecause it was shown to be the maximum tolerated dosefor longer-term dosing (10). The cabozantinib gavage solu-tion was made fresh daily. Subcutaneous tumor burden

Cabozantinib Inhibits Prostate Cancer Growth in Bone

www.aacrjournals.org Clin Cancer Res; 20(3) February 1, 2014 619




was determined using bioluminescence imaging (BLI), asdescribed below, every 5 days. All animals were sacrificed atthe end of day 15 of treatment. The tumor weights weremeasured. The subcutaneous tumors were harvested. Halfof each tumor was kept for histologic analysis and the otherhalf flash frozen for molecular analysis.

For LuCaP 35 studies cells weremaintained in SCIDmiceas xenografts, tumors were harvested and made into single-cell suspensions as previously described. SCID mice werethen injected subcutaneously with LuCaP-35 cells (2� 106

in 100mLRPMI-1640) and allowed to establish tumors overa period of 42 days. After establishment of tumor, treatmentwith either cabozantinib (60mg/kg/d, oral administration;n¼ 12) or distilled water vehicle (n¼ 12) was initiated andcontinued for 10 weeks. Subcutaneous tumor growth wasmonitored weekly using calipers to measure two perpen-dicular axes. Tumor volume was calculated using the for-mula ðvolume ¼ length � width2=2Þ. At 10 weeks, mice

were euthanized and subcutaneous tumors were collected,weighed, and saved in formalin for additional studies.

For intratibial studies, Ace1luc cells, C4-2Bluc and PC-3luc

cells were inoculated intratibially to measure the effect ofcabozantinib on tumor growth. For intratibial injection,micewere anesthetizedwith 2.5% isofluorane/air, and bothlegs were cleaned with betadine and 70% ethanol. The kneewas flexed, and a 27-G3/8-inch needle was inserted into theproximal end of right tibia followed by injection of 20 mLsingle-cell suspensions of Ace1luc and PC-3luc cells (1� 105

cells) andC4-2Bluc cells (3�105 cells). BLI and radiographywere used to check tumor burden in bone as a primaryoutcome. Tumors were allowed to become established for 7days for PC-3, 14 days for Ace1, and 30 days for C4-2B.Micewere randomized to receive either cabozantinib (n¼ 12; 60mg/kg body weight/d, oral administration) or vehicle (n ¼12, distilled water) for 3 weeks (PC-3luc), or 5 weeks(Ace1luc), or 28 days (C4-2Bluc) once tumors had beenestablished. The cabozantinib gavage solution was madefresh daily. The blood samples were taken by cardiac aspi-ration under anesthesia before treatment for mice with C4-2Bluc cells inoculation for determination of serum PSAlevels. Tumor development in bone was checked by BLIandmicroradiography, weekly or every 10 days. All animalswere sacrificed at the end of week 3, or week 5, or week 7.Before sacrifice, blood samples and magnified flat radio-graphs were taken under anesthesia, and serum and tibiaewere collected.

Bioluminescence imagingBLI was performed as previously described (19).

Bone imagingBone mineral content (BMC) of the excised tibiae were

measured using dual-energy X-ray absorptiometry (DEXA)on an Eclipse peripheral DEXA Scanner using pDEXA Sabresoftware version 3.9.4 in research mode (Norland MedicalSystems) as previously described (18). Magnified flat radio-graphs of hind limbs were taken on a microradiographyapparatus (Faxitron X-ray Corp.). The tibiae were scanned

on a mCT system (GE Healthcare Systems) and recon-structed with a voxel size of 25 mm for mCT analysis.

Statistical analysisStatistical analysis was done using Statview Software

(Abacus Concepts). For comparison among two groups, aStudent t test was used. For multiple comparisons, ANOVAwas used for initial analyses followed by Fisher protectedleast significant difference for post hoc analyses. Differenceswith P < 0.05 were determined as statistically significant.

ResultsOur initial studies were based on the observation that

cabozantinib induced marked resolution of Tc99 bone scanlesions inmenwith prostate cancer skeletalmetastases (11).Accordingly, we first sought to determine whether cabo-zantinib inhibits prostate cancer cell growth in vivo in bonein amurinemodel.We injectedAce1luc cells, which induce astrong osteoblastic reaction, into the tibiae of mice. Tumorswere allowed to become established and then cabozantinib(60 mg/kg/os daily based on this being previously demon-strated as the maximum tolerated dose for long-termadministration; ref. 10) or vehicle administration was ini-tiated and continued for 5 weeks. Cabozantinib inhibitedtumor growth based on BLI (Fig. 1A and B). As anticipated,the Ace1luc cells created osteoblastic lesions based on radi-ography and microCT, which was associated with anincrease of BMC (Fig. 1C and D). Cabozantinib decreasedthe Ace1luc-induced osteoblastic lesions based on bothradiographs and microCT and a decrease of BMC towardthe normal baseline (Fig. 1C and D). Although cabozanti-nib administration altered Ace1luc-induced tumor boneremodeling, it did not alter serum markers of bone remo-deling although serum PINP showed a trend (P ¼ 0.08)toward increasing (Fig. 1E).

To determine the effect of cabozantinib on other prostatecancer cell lines, we further evaluatedC4-2Bluc, which createmixed osteoblastic/osteolytic lesions. In this model, cabo-zantinib administrationwas initiated after tumorhaddevel-oped after intratibial injection and was continued for aperiod of 7 weeks as C4-2B tumors grow slowly. Similar tothe results with Ace1luc, cabozantinib inhibited tumorgrowth in bone based on BLI (Supplementary Fig. S1A andS1B). Further evidence of an effect on tumor burden wasprovided by the observation that cabozantinib administra-tion also decreased serum PSA levels in the C4-2B–bearingmice (this cell line produces PSA, as opposed to PC-3 andAce1; Supplementary Fig. S1C). As expected, the C4-2B cellscreated mixed osteoblastic and osteolytic lesions based onradiography, microCT, and decline of BMC (Supplemen-tary Fig. S1D and S1E). Cabozantinib administrationreversed the tumor-induced reduction in BMC to levels ofnon–tumor-bearing bone (Supplementary Fig. S1E). Thiseffect was opposite to that in the Ace1luc cell model, inwhich tumor growth was associated with excessive BMCthat cabozantinib treatment reduced (Fig. 1D). No impacton systemic bone remodeling markers was observed (Sup-plementary Fig. S1F).

Dai et al.





Todeterminewhether the effects onBMC inAce1 andC4-2B tumors could be due to direct effects on bone, weevaluated the non–tumor-bearing tibiae in themice. Underthese treatment conditions, cabozantinib had no effect onBMC of the non–tumor-bearing bone (Fig. 1D and Sup-plementary Figs. S1E and S2D; compare the BMC in the notumor groups with and without cabozantinib). However,cabozantinib induced an increase in the osteoblast perim-eter and a decrease of the osteoclast perimeter in the non–tumor-bearing bone (Supplementary Fig. S3).It is now recognized that some prostate cancer lesions are

highly heterogeneous and in some cases may have a strong

osteolytic component. The cell line PC-3 is frequently usedto model prostate cancer both in vitro and in vivo and ishighly osteolytic. To determine whether the effects of cabo-zantinib extended to prostate cancer of an osteolytic nature,we assessed the impact of cabozantinib on PC-3luc. Afterintratibial injection, tumors were allowed to become estab-lished and then cabozantinib or vehicle administration wasinitiated and continued for 3 weeks. In contrast with theACE1luc and C4-2Bluc, cabozantinib had no impact on PC-3luc tumor growth in bone (Supplementary Fig. S2A andS2B). In addition, cabozantinib did not impact PC-3–induced bone loss (Supplementary Fig. S2C and S2D).

Weeks (treatment)

To

tal

flu

x (

p/s

, ¥1

010)

60

40

20

0

40

30

20

10

0

8

6

4

2

0

Vehicle

PIN

P (

ng

/mL

)O

ste

oc

alc

in (

ng

/mL

)T

RA

CP

5b

(U

/L)

Cabozantinib

Vehicle

Cabozantinib

Vehicle

Cabozantinib

Ve

hic

leC

ab

oza

nti

nib

Ve

hic

leC

ab

oza

nti

nib

Vehicle Cabozantinib


Ace1luc

RadiographA

D E

B C m CT

6

4

2

00 1 2 3 4 5

0.015

0.010

0.005

0.000

Control Tumor

BM

C (

g)

Figure 1. Cabozantinib inhibits the progression of Ace-1luc prostate cancer cells in bone in vivo. SCIDmicewere injected intratibially with Ace-1luc cells (1� 105

in 20 mL DMEM/F12) and allowed to become established over a period of 14 days. After establishment of tumor growth in bone, treatment with eithercabozantinib (60mg/kg/oncedaily,per os;n¼12) or distilledwater vehicle (n¼12)was initiated andcontinued for 5weeks.Micewere subjected toweeklyBLI.At 5 weeks after initiation of cabozantinib, mice were euthanized, bones subjected to Faxitron X-ray analysis, microCT, and DEXA, and blood wascollected and separated into serum thatwassubjected to enzyme-linked immunoassay for bonemarkers (PINP, osteocalcin andTRACP5b). A, representativeBLI imaging. Note the decreased signals in the tibiae of the cabozantinib-treated mice compared with vehicle-treated mice. B, tumor burden asmeasured using BLI. Results are reported as relative light units (RLU). �, P < 0.05 versus vehicle-treated mice. C, representative radiographic and microCTimaging. Note the decreased osteoblastic activity in the tibiae of the cabozantinib-treated mice compared with vehicle-treated mice. D, BMC measuredusing DEXA. #, P < 0.05 versus no tumor mice for each respective treatment group. �, P < 0.05 versus tumor-bearing vehicle-treated animals. E,serum PINP, osteocalcin, and TRACP 5b levels.






Cabozantinib administration was associated with anincrease in serum levels of the bone resorption markerTRACP 5b, whereas it had no impact on bone productionmarkers (Supplementary Fig. S2E).

As cabozantinib inhibited both Ace-1 and C4-2B growthbut not PC-3 growth in bone, we next determine whethercabozantinib impacted soft tissue PC-3 lesions. We injectedPC-3luc in soft tissue on the flank of mice and cabozantinibadministration was initiated after tumors were establishedover 5 weeks and continued for 15 days. Cabozantinibinhibited the development of PC-3luc tumors in soft tissue(Fig. 2A–C), demonstrating that the ability of cabozantinibto inhibit tumor growth was not specific to tumors growingwithin bone. To ensure these results were not specific to PC-3luc cells, we also evaluated the effect of cabozantinib on theandrogen-dependent human prostate cancer xenograftLuCaP 35. To model the clinically relevant transition to

CRPC we implanted LuCaP 35 xenografts subcutaneously,then after tumors became established, mice were subjectedto orchiectomy at which time cabozantinib was initiated.We were able to measure serum PSA levels in this model asLuCaP 35 produces PSA. Cabozantinib prevented progres-sion ofCRPC tumor based on tumor burden (Fig. 2D andE)and serum PSA levels (Fig. 2F). Taken together, these resultsdemonstrate that cabozantinib can inhibit prostate cancergrowth independent of the bone microenvironment.

As cabozantinib had a discordant impact on PC-3 growthin bone versus soft tissue, we examined whether a differ-ential impact on tumor-associated angiogenesis or infiltra-tion of tumor by myeloid cell could account for this dif-ference, as these activities are known to impact tumorgrowth (20). Cabozantinib decreased tumor-associatedvasculature in subcutaneous but not intratibial PC-3 tumors(Supplementary Fig. S4A). In contrast, cabozantinib had no

Figure 2. Cabozantinib inhibits progression of PC-3luc and LuCaP-35 prostate cancer cells in soft tissue in vivo. A andC, PC-3luc cells: SCIDmicewere injectedsubcutaneously with PC-3luc cells (1� 106 in 100 mL RPMI-1640) and allowed to grow for 35 days to become established. After the establishment of tumor,treatmentwith either cabozantinib (60mg/kg/d,per os;n¼12) or distilledwater vehicle (n¼12)was initiatedandcontinued for 15days.Micewere subjected toBLI weekly. At 15 days after initiation of cabozantinib, mice were euthanized, subcutaneous tumors were collected, weighed, and saved in formalin foradditional studies. A, representative BLI of PC-3luc tumors. Note the decreased signals in the tumors of the cabozantinib-treatedmice comparedwith vehicle-treated mouse. B, tumor burden of PC-3luc tumors as measured using BLI. Results are reported as RLU. �, P < 0.05 versus vehicle-treated mice. C,tumor weight of PC-3luc tumors. �, P < 0.05 versus vehicle-treated mice. D and F, LuCaP-35 cells: LuCaP tumors maintained in SCID mice were made intosingle-cell suspensions. SCIDmicewere then injected subcutaneouslywith LuCaP-35 cells (2�106 in 100mLRPMI-1640) and allowed to develop into tumorsover a period of 42 days. After the establishment of tumor, treatment with either cabozantinib (60 mg/kg/d, oral administration; n ¼ 12) or distilledwater vehicle (n ¼ 12) was initiated and continued for 10 weeks. The tumors were measured by caliper weekly. At 10 weeks, mice were euthanized,subcutaneous tumors were collected, weighed, and saved in formalin for additional studies. D, tumor volume of LuCaP-35 tumors. �, P < 0.05 versus control;��, P < 0.05 versus vehicle plus castration. E, tumor weight of LuCaP-35 tumors. �, P < 0.05 versus control; ��, P < 0.05 versus vehicle plus castration.F, PSA levels from mice with LuCaP-35 tumors. �, P < 0.05 versus control; ��, P < 0.05 versus vehicle plus castration.

Dai et al.





impact on CD45þ myeloid cell numbers found withineither subcutaneous or intratibial PC-3 tumors (Supple-mentary Fig. S4B). These results suggest that a diminishedimpact of cabozantinib-mediated inhibition of angiogene-sis in bone versus soft tissue could contribute to thedecreased antitumor response observed in PC-3 intratibialtumors.On the basis of the overall inhibition of prostate cancer

growth, we next determined whether cabozantinib had animpact on cell proliferation and/or apoptosis in the cancercells. We therefore stained the tumor tissue from the cabo-zantinib-treated animals for either Ki67 to evaluate forproliferation or activated caspase-3 to evaluate for apopto-sis. Ki67 expression was undetectable in the Ace1luc tumors,which is due to lack of cross reactivity of the antibody withthe canine epitope in these cells (Fig. 3A and B). However,Ki67 was decreased in the cabozantinib-treated C4-2Bintratibial tumors and the PC-3luc subcutaneous tumors,respectively, compared with vehicle-treated animals (Fig.3A and B). In contrast, Ki67 expression was not altered bycabozantinib treatment in the PC-3luc intratibial tumors(Fig. 3A and B). Caspase-3 expression was increased in theAce1luc and C4-2Bluc intratibial tumors and the PC-3luc

subcutaneous tumors but not the PC-3luc intratibial tumorsfrom the cabozantinib-treated mice compared with vehi-cle-treated mice (Fig. 3C and D). Taken together, theseresults suggest that cabozantinib inhibits overall tumorgrowth through inhibition of proliferation and promo-tion of apoptosis.To determine whether cabozantinib directly modulated

the viability of prostate cancer cells, we examined its effecton three representative cell lines in vitro: LNCaP, C4-2B, andPC-3. LNCaP cells were used to represent an androgen-responsive cell line, but they do not growwell in bone, thuswere not used for the in vivo studies. Cabozantinib had noimpact on cell viability at 24 hours (not shown), whereas at72 hours it inhibited cell viability of these cell lines in adose-dependent fashion (Fig. 4A). We next sought to deter-mine whether cabozantinib achieved this effect, in part,through induction of apoptosis by measuring caspase-3/7activities. Cabozantinib induced caspase in all three celllines at 72 hours (Fig. 4B). To determine whether the abilityof cabozantinib to impact prostate cancer cells extendedinto diminishing theirmetastatic phenotype, we assessed itsimpact on the invasive ability of the prostate cancer cells.Cabozantinib inhibited the invasive ability of all three celllines (Fig. 4C). PSA is used to measure prostate tumorresponse to therapies. To determine whether cabozanti-nib-impacted PSA expression, we measured intracellularPSA mRNA and PSA protein from cell culture supernatantin LNCaP and C4-2B cells exposed to cabozantinib. Wefound that cabozantinib induced a biphasic effect, first anincrease thendecrease in PSAmRNAandprotein expressionin the androgen-dependent LNCaP cells, but had no impacton PSA expression in the androgen-nonresponsive C4-2Bcells (Fig. 4D). In contrast with this in vitro result in the C4-2B model, in vivo a decline in PSA was observed, which wasassociated with a reduction in tumor burden (Supplemen-

tary Fig. S1C). Taken together, these results indicate thatcabozantinib can directly diminish prostate cancer progres-sion and that measurement of PSA may not be the optimalassessment of tumor response.

Our earlier observation that cabozantinib had no impacton non–tumor-bearing bone suggested that cabozantinibhas no considerable direct effect on bone cells. However,impacts on BMC in healthy bone may take longer than the7-week period of administration of cabozantinib and maybe dose-dependent. Furthermore, the observations thatcabozantinib induced a response on bone scan lesions inthe clinical trial and that it impacted tumor-induced boneremodeling in the animal models suggests that there mayalso be a direct effect of cabozantinib on bone cells. Wetherefore examined whether cabozantinib modulates oste-oblastic activity and determined its impact on the ability ofthe preosteoblast cells MC3T3-E1 and ST-2 to differentiateand function as osteoblasts by measuring cell viability,alkaline phosphatase (an indicator of early osteoblast dif-ferentiation), osteocalcin (a measure of late osteoblastdifferentiation), and calcium (an indicator of mineraliza-tion). Cabozantinib decreased cell viability of both cell linesin a dose-dependent fashion with ST-2 cells being moresensitive to this activity (Fig. 5A). However, cabozantinibmodulated alkalinephosphatase activity inboth cell lines ina biphasic fashion (Fig. 5B). In contrast, cabozantinibdiminished osteocalcin expression in both cell lines (Fig.5C). Finally, cabozantinib had no impact on the ability ofMC3T3-E1 cells to mineralize, except at the highest doseevaluated (Fig. 5D). Because ST-2 cells do not readilymineralize in vitro these cells were not evaluated. Theseresults show that cabozantinib induces early osteoblastdifferentiation at low doses, but at higher doses can inhibitosteoblast differentiation, whichmay be due, in part, to theoverall impact of cabozantinib on osteoblast viability. Inaddition to an impact on osteoblasts, we explored theability of cabozantinib tomodify osteoclast biology in vitro.Cabozantinib inhibited the differentiation of RAW preos-teoclast cells into mature osteoclasts in a dose-dependentfashion (Fig. 5E). Similarly, cabozantinib inhibited theoverall resorptive activity in these cultures (Fig. 5E). How-ever, when resorptive activity was normalized for osteoclastnumbers, no change in osteolytic activity was observed (Fig.5E). These results indicate that the effect of cabozantinib onresorption in vivo was primarily due to a reduction in thenumbers of osteoclasts, as opposed to inhibition of theactivity of individual mature osteoclasts.

Cabozantinib is amultikinase inhibitorwith high affinityfor MET and VEGFR2 relative to other kinases. To ensurethat cabozantinib was acting on these expected intracellulartargets, we first confirmed the ability of cabozantinib toblockHGF-mediated activation ofMET bymeasuring phos-pho-MET in PC-3 cells (Fig. 6A). We next determinedwhether cabozantinib targeted VEGFR2 in the preosteoblastcell lines MC3T3-E1 and ST-2 (VEGFR2 is not expressed inthe prostate cancer cell lines; data not shown). VEGFinduced phospho-VEGFR2 expression in the MC3T3-E1cells, but not in the ST-2 cells, which had high basal






phospho-VEGFR2 expression (Fig. 6B). Cabozantinibinhibited basal and VEGF-induced phospho-VEGFR2expression, respectively, in ST-2 and MC3T3 cell lines (Fig.6B). In addition, as the PI3K/Akt pathway is downstream ofVEGFR2, we evaluated if cabozantinib impacted VEGF-mediated activation of AKT. VEGF induced phospho-AKTexpression in the MC3T3-E1 cells, but not in the ST-2 cells,which had high basal phospho-AKT expression, similarto the high basal phospho-VEGFR2 expression observed

(Fig. 6B). Cabozantinib inhibited basal and VEGF-inducedphospho-AKT expression, respectively, in ST-2 and MC3T3cell lines (Fig. 6C).

Todeterminewhether these results extended to the abilityof cabozantinib to target these pathways in tumor cells invivo we evaluated for phospho-Met and phospho-VEGFR2expression in the tumors. To accomplish this, we subjectedthe intratibial and subcutaneous PC-3 tumors, the intrati-bial C4-2B tumors, and the subcutaneous LuCaP35 tumors

Ki 67A B

C D

Caspase 3


Vehicle

100

80

60

40

20

0

100

80

60

40

20

0

Cabozantinib

Vehicle

Cabozantinib


Ace

1(lT

)C

4-2B

(lT)

PC

3(lT

)P

C3(

SQ

)

Ace 1(lT) C4-2B (lT) PC-3 (lT) PC-3 (SQ) Ace 1(lT) C4-2B (lT) PC-3 (lT) PC-3 (SQ)

Cas

pase

-3 p

ositi

ve c

ells

(%

)

ki67

pos

itive

cel

ls (

%)

B

N

B

B

Figure 3. Impact of cabozantinib on cellular proliferation and apoptosis of intratibial and soft tissue tumors inmice. Tumors from themice as described in Fig. 1through Fig. 4 were subjected to IHC for (A) proliferation (using anti-Ki67) and (C) apoptosis (using antiactivated caspase-3/7). �B, bone; �N, necrotic tissue.B and D, the percentage of positively stained cells was measured in three random �40 fields for each section. Results are shown as mean � SD foreach section. ��, P < 0.05 versus untreated for each cell line.

Dai et al.





to IHC. Unfortunately, the Ace-1 tumors are a canine cellline and although we attempted IHC on them, we were notsuccessful. These results are shown in Supplementary Figs.S5 and S6. We found that cabozantinib inhibited phospho-MET in all tumors in both subcutaneous and intratibial

sites; however, inhibition of phospho-MET in the intratibialPC-3 cells was about 25% of untreated level; whereas,inhibition of phospho-MET in the subcutaneous tumorswas approximately 50%. Cabozantinib inhibited phospho-VEGFR2 in all tumors in both subcutaneous and intratibial

Figure 4. Cabozantinib inhibitsmultiple parameters of prostate cancer tumor progression in prostate cancer cells. A and B, LNCaP, C4-2B, and PC-3 prostatecancer cells were plated in 96-well plates (2 � 103 cells/well) in medium plus 10% FBS and incubated overnight, then media was replaced with 2% FBS-containing media and the indicated concentrations of cabozantinib. After 72 hours, (A) the cell viability was measured using WST-1 assays and (B)apoptosis activity was assayed bymeasuring caspase-3/7 activity based on cleavage of DEVD substrate using the Apo-ONEKit (Promega). Data are from fivereplicates and shown as mean � SD. �, P < 0.05 versus control (0 mmol/L cabozantinib). The experiments were repeated three times. C, LNCaP, C4-2B, andPC-3 cells (5�104 cells) were added to the inserts ofmodifiedBoyden chambers and treatedwith the indicated concentrations of cabozantinib (or saline). Theplates were incubated for 22 hours in a CO2 incubator at 37�C. The chamber inserts were then stained using the Diff-Quick staining Kit (Dade-Behring)according to the manufacturer's instructions. The invasion was determined as the percentage of cells that migrated through the membrane. Data are fromtriplicate samples and reported as mean � SD% of control. �, P < 0.05 versus control. The experiment was repeated three times. D, (i) PSA mRNAexpression: LNCaPandC4-2Bcellswereplated at 5�105 cells/mL in 6-well plates and then treatedwithDHT (1 nmol/L) aspositive control, dimethyl sulfoxide(DMSO) as negative control and the indicated levels of cabozantinib. After 24 hours, total RNA was collected. Total RNA was subjected to PCR forPSA mRNA expression. (ii) PSA protein expression: LNCaP and C4-2B cells were plated at 2� 103 cells/mL in 96-well plates and after 24 hours treated withDHT (1 nmol/L) as positive control, DMSO as negative control and the indicated levels of cabozantinib. After 48 hours, the supernatants were collectedand PSA level in the supernatants was measured by PSA ELISA and values were normalized to cell numbers as determined by the modified WST-1 assay.Data are from triplicate samples and reported as mean � SD. P < 0.05 versus DMSO control.






sites; however, similar to what was observed for phospho-Met inhibition, cabozantinib-mediated inhibition of phos-pho-VEGFR2 was greater in the subcutaneous PC-3 versusintratibial PC-3 tumors. These results indicate that cabo-zantinib successfully targeted MET and VEGR2; however,intratibial PC-3 cells were more resistant to cabozantinib.The diminished inhibition of both phospho-MET andphospho-VEGFR correspond to the decreased antitumorresponse and antiangiogenic response in the intratibialPC-3 tumors compared with subcutaneous PC-3 tumors.Taken together, these results suggest that cabozantinibcan effectively target MET and VEGFR2 signaling in tumorcells in both the soft tissue and the bone metastasis micro-environments; however, there may be instances in whichtumors may have innate resistance to cabozantinib-medi-ated effects.

DiscussionMore than 80% of men with advanced prostate cancer

develop bone metastases and most of those men will alsohave soft tissue metastases. Our results suggest that cabo-zantinib is effective against prostate cancer in both softtissue and bone sites. These results indicate that cabozanti-nib effectively induces a tumor response independent of thetumor microenvironment. Furthermore, our results suggestthat the marked cabozantinib induced responses observedon the bone scans of men with prostate cancer bonemetastases were due to an antitumor response and not justan impact on bone remodeling.

Prostate cancer bone metastases are characterized asprimarily osteoblastic; however, it is now recognized thatmetastatic lesions are heterogeneous and contain areas ofosteolytic activity (21). The models we evaluated included

Figure 5. Cabozantinib impacts preosteoblast viability and differentiation. A, MC3T3-E1 and ST2 cells were plated in 96-well plates (2 � 103 cells/well) ina-MEM medium plus 10% FBS and incubated overnight, then media was replaced with 2% FBS-containing media and the indicated concentrations ofcabozantinib. After 72 hours, the cell viability was measured using WST-1 assays. Data are from triplicates and shown as mean � SD. �, P < 0.05versus control (0 mmol/L cabozantinib). The experiments were repeated three times. B and C, MC3T3-E1 and ST2 cells were plated in 12-well plates (5� 104

perwell) andgrown ina-MEMcontaining 10%FBS.After the cellswereconfluent, themediumwas replacedwithosteoblast differentiatingmedia (a-MEMwith10% FBS, 50 mg/mL ascorbic acid, and 10 mmol/L b-glycerophosphate) the indicated concentrations of cabozantinib. Medium was refreshed every 3days, at days 9, supernatants and cells were collected. B, alkaline phosphatase in cell lysates wasmeasured using ALP assay kit and C, osteocalcin in mediawas measured using mouse osteocalcin EIA assay. Data are from triplicates and shown as mean � SD. �, P < 0.05 versus control (0 mmol/L cabozantinib).The experiments were repeated three times. D, MC3T3-E1 cells were treated as in (B), but allowed to continue growth for 21 days after confluence at whichtime cells were collected and calcium in cell lysates was measured using a calcium assay that measure the reaction between o-cresolphthalein andcalcium. Results were normalized to protein concentration in cells. Data are from three experiments and shown as mean � SD. �, P < 0.05 versus control(0 mmol/L cabozantinib). E, for measurement of osteoclast numbers and function RAW cells were plated in 96-well plates with bovine bone chips andRANKL to induce osteoclastogenesis. At day 7, cabozantinib was added to the indicated concentration, and at day 10 supernatant was collected formeasurement of TRACP 5b and CTX. The resorption index per osteoclast is calculated as CTX divided by TRACP 5b. Data are from three experiments andshown as mean � SD. �, P < 0.05 versus control (0 mmol/L cabozantinib).

Dai et al.





highly osteoblastic, mixed osteoblastic/osteolytic, and hig-hly osteolytic tumor types. Cabozantinib demonstratedeffective antitumor activity for the osteoblastic and mixedlesions, but did not have an impact on the highly osteolyticPC-3 cells when grown in the bone. As pure osteolyticlesions, similar to those induced by PC-3, are rarely obs-erved in prostate cancer, these findings suggest that cabo-zantinib would be most active against the type of bonemetastases that are present in the majority of men withprostate cancer.The observation that cabozantinib inhibits PC-3 tumor

growth in soft tissue but not in bone demonstrates that PC-3cells are sensitive to cabozantinib in vivo and also suggeststhat there may be a microenvironment effect that protectsPC-3 cells from cabozantinib in the bone. Furthermore, theobservation that cabozantinib mediated inhibition of bothphospho-MET and phospho-VEGFR2 was diminished inintratibial versus subcutaneous PC-3 cells supports thehypothesis that the context of the tumormicroenvironmentimpacts response to cabozantinib. However, this seemedspecific to PC-3 cells as intratibial C4-2B cells respondedwell to cabozantinib and this was associated with reduc-tions in phospho-MET and phospho-VEGFR2 in these cells.This in vivo observation is supported by the in vitro studiesthat demonstrated cabozantinib diminished viabilityand induced apoptosis in PC-3 cells. However, in vitro,

PC-3 cells were less sensitive to cabozantinib inhibitionthan cells with more pronounced osteoblastic features suchas C4-2B. Taken together, these results indicate that PC-3cells are less susceptible to cabozantinib than the otherprostate cancer cell lines evaluated, and that an osteolyticbonemicroenvironmentmay potentially protect PC-3 fromthe antitumor activity of cabozantinib. The mechanismsthroughwhich osteolysismight protect prostate cancer cellsfrom cabozantinib were not determined at this time; how-ever, it is well recognized that resorbing bone releasesa variety of growth factors that could diminish the inhibi-tion of proliferation or induction of apoptosis caused bycabozantinib.

PSA levels have generally been a good indicator of anti-tumor activity and thus are often used tomeasure treatmentresponse and to monitor for recurrence. However, in thecabozantinib clinical trial, which was previously performedin men with CRPC, PSA levels did not always correlate withthe antitumor effects in bone and soft tissue (11). Thissuggests that measurements other than PSA should be usedin monitoring tumor response and recurrence in menundergoing therapy with cabozantinib. In ourmurine stud-ies, PSA levels correlated with the C4-2B tumor responsebased on the decline in tumor volume.However, our in vitrostudies demonstrated that cabozantinib induced a dose-dependent biphasic response in LNCaP cells, whereas it had

p-META

C

B

p-AKT

MET

p-AKT

AKT

AKT

GAPDH

GAPDH

1 0.60 1.83 0.95

1 1.37 0.65 0.63

1 1.12 1.52 0.70

1 1.72 0.56 0.86

HGF (50 ng/mL)

Cabozantinib (1 µmol/L)VEGF (50 ng/mL)

Cabozantinib (1 µmol/L)

VEGF (50 ng/mL)

Cabozantinib (1 µmol/L)

MC3T3-E1 ST2

MC3T3-E1 ST2

p-VEGFR2

VEGFR2

p-ERK1/2

ERK1/2

GAPDH

1 5.72 0.28 0.31 1 0.98 0.22 0.29

1 2.23 12.45 11.87 1 0.78 1.27 1.32

1 0.81 0.53 0.44 1 0.84 0.71 0.72

1 0.72 1.16 1.17 1 1.24 1.22 1.27

1 3.84 0.79 0.70 1 0.98 0.16 0.25

1 0.98 0.89 1.01 1 0.97 0.99 0.94

Figure 6. Cabozantinib inhibits c-MET, VEGFR2, and AKT phosphorylation in prostate cancer cells and preosteoblasts. A, PC-3 cells were plated at 2 � 106

cells on 100mmplates. After 12 hours, the cells were pretreatedwith cabozatinib (1 mmol/L) for 3 hours, then treated with either DMSO (as a negative control),or HGF (50 ng/mL, as a positive control) for 20minutes. Then total protein was extracted from the cells and subjected to immunoblot analysis using anti-MET,anti-phosphorylated-MET (p-MET), anti-Akt, anti-phosphorylated-Akt (p-Akt), and anti-GAPDH primary antibodies and appropriate secondaryantibodies. GAPDHwas used as an internal control. B andC, ST2 andMC3T3-E1 cells were plated at 2� 106 cells on 100mmplates. After 12 hours, the cellswere pretreated with cabozantinib (1 mmol/L) for 3 hours, then treated with either DMSO (as a negative control), or VEGF (50 ng/mL) for 20 minutes.Protein was then extracted from the cells and subjected to immunoblot analysis using (B) anti-VEGFR2, anti-phosphorylated-VEGFR2 (p-VEGFR2), anti-ERK1/2, anti-phosphorylated-ERK1/2 (p-ERK1/2), or (C) anti-AKT, anti-phosphorylated-AKT (p-AKT), and anti-GAPDH primary antibodies and appropriatesecondary antibodies. GAPDH was used as an internal control. These results were obtained from at least three replicate experiments. Gel images werethen subjected to densitometry using ImageJ 1.38� software and densitometry values normalized to GAPDH band values for each lane are reportedbelow each band.






no impact on PSA response in C4-2B cells. In the context ofLNCaP cells, the induction of PSA levels occurred at dosesthat decreased cell viability, thus suggesting that cabozan-tinib-induced PSA expression (as opposed to just anincrease in the number of cells producing PSA). Further-more, the reduction of PSA expression in the LNCaP cellsoccurred at the dose of cabozantinib that had markeddecrease in viability and increase of apoptosis. Takentogether these data suggest that in the androgen-responsiveLNCaP model, cabozantinib induces PSA expression thatmay be counteracted by cell death at higher concentrationsresulting in overall reduction of PSA expression. This isconsistent with the observation that cabozantinib had noimpact on PSA expression in C4-2B cells in vitro, yet dimin-ished PSA expression in vivo, suggesting that the PSA declinein vivowas reflective of the decline in tumor burden. To drawdefinitive conclusions about the impact of cabozantinib onPSA levels in vivo, correlation of cabozantinib serum levelswith their impact on PSA expression must be performed. Inour models, we did not measure cabozantinib serum levels,thus our results about PSA expression should be consideredexploratory at this time.

A major goal of this study was to explore whether thecabozantinib-induced bone scan effects were reflective ofantitumor efficacy or a direct effect on bone because bonescans only measure incorporation of radionuclide intothe bone. Bone scans are thought to be a measure ofosteoblastic activity, and thus theoretically any effect onosteoblastic activity could impact the bone scan indepen-dent of a direct effect on tumor. Our results provideevidence that the impact on bone remodeling inducedby cabozantinib in the context of tumor was, in part, dueto an antitumor response. The evidence for this conclu-sion include (i) cabozantinib decreased tumor burden inbone; (ii) in the osteoblastic tumor, the cabozantinib-induced tumor response was associated with a decrease ofBMC toward normal levels; whereas, in the mixed osteo-lytic/osteoblastic tumor, the cabozantinib-induced tumorresponse was associated with an increase of BMC towardnormal levels and in the highly osteolytic tumor, the lackof a cabozantinib-induced tumor response was associatedwith no change in BMC; and (iii) cabozantinib had noimpact on BMC in non–tumor-bearing bone. However,the observation that cabozantinib altered both osteoblastand osteoclast perimeters in non–tumor-bearing bonesuggests that there is a primary bone effect that couldimpact tumor growth. These results are consistent withthe in vitro observations that cabozantinib impacted oste-oblastic differentiation and osteoclast production. Clearlydistinguishing between direct versus indirect antitumoreffects that are mediated through altering the bone willbe challenging in these in vivo models. Several possibil-ities could account for the apparent contrast amongin vitro and in vivo observations in the context of thetumor-induced bone phenotype including (i) that tumor-induced bone remodeling overshadows any cabozantinibbone–remodeling effect and (ii) the time span thesestudies were carried out in is relatively short compared

with modest effects induced by cabozantinib on osteo-blast differentiation.

Several lines of evidence suggest that the VEGF axis isimportant in prostate cancer progression. Perhaps themost recognized impact of VEGF is on its ability to induceangiogenesis, which supports tumor growth. A role of theVEGF pathway in prostate cancer is supported by studiesthat show (i) increased plasma levels of VEGF correlatewith the presence of bone metastasis in prostate cancer(22) and (ii) that VEGFR overexpression is associatedwith metastasis and poor outcome, whereas it seems toregulate epithelial-to-mesenchymal transition of prostatecancer cells (23, 24). In addition to proangiogenic effects,the VEGF axis has been shown to promote osteoblasticactivity in prostate cancer (7, 8). Because of its impor-tance in tumor progression, targeting angiogenesis hasreceived much attention. However, in general, clinicaltrial results with pure antiangiogenesis inhibitors havebeen disappointing for multiple tumor types. One sug-gestion to account for this observation is that the hypoxiainduced by angiogenesis inhibition leads to activationof the HGF/MET pathway, which may promote tumorgrowth (25).

Multiple lines of evidence indicate that the HGF/METpathway plays a critical role in prostate cancer progres-sion and seems to be an appropriate candidate for tar-geted therapy. HGF/MET expression has been associatedwith prostate cancer aggressive behavior in tissue samples(26, 27). Overexpression of MET is an independent pre-dictor of invasion and metastasis in prostate cancer andits expression has been shown to be highly prevalent inbone metastases (28, 29). Moreover, increased serumlevels of HGF are an independent prognostic marker inpatients with advanced disease stage (30, 31). Upregula-tion of HGF and its receptor MET is associated with thetransition to androgen-independent growth of prostatecancer (32, 33). Pharmacologic inhibition of the METsignaling pathway by a variety of methods has beenshown to reduce both the development and progressionof prostate cancer metastases in vitro and in vivo in animalmodels (34–36) including suppression of prostate cancergrowth in a mouse model (37). The observation thatangiogenesis is important in tumor progression and thattargeting angiogenesis promotes HGF/MET activation,which in turn may promote tumor growth, led to theconcept that targeting both the VEGF and HGF/METpathways may have greater antitumor efficacy than target-ing either pathway alone.

Cabozantinib targets both VEGFR2 and MET and inhib-ited their activation in prostate cancer and bone stromalcells, respectively. The ability to target MET in the cancercells is consistent with the ability of cabozantinib toimpact tumor growth. The ability of cabozantinib to targetVEGFR2 in bone stromal cells is consistent with the obser-vation that cabozantinib is active against bone marrowmicroenvironment cells and could account for the in-hibition of stromal cell differentiation to osteoblastswe observed in vitro. These observations indicate that

Dai et al.





cabozantinib can target both the cancer and microenvi-ronment components of the bone metastases throughdifferent pathways.In conclusion, cabozantinib can effectively inhibit tumor

growth and tumor-induced bone remodeling in murinemodels of prostate cancer that have osteoblastic compo-nents. Cabozantinib-mediated inhibition of tumor-ind-uced bone changes seemed to be primarily due to its abilityto inhibit tumor growth as opposed to a direct effect onbone. The ability of cabozantinib to target both the tumorcells themselves in addition tomicroenvironment cellsmayresult in effective antitumor therapy.

Disclosure of Potential Conflicts of InterestD.T Aftab is a senior Vice President and has ownership interest (including

patents), F. Schimmoller has ownership interest (including patents) inExelixis Inc. E.T. Keller received commercial research grant from ExelixisInc. and Pfizer. No potential conflicts of interest were disclosed by the otherauthors.

Authors' ContributionsConception and design: J.M. Keller, D.T. Aftab, F. Schimmoller, E.T. KellerDevelopment of methodology: J. Dai, H. Zhang, J.M. Keller, E.T. KellerAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): J. Dai, H. Zhang, A. Karatsinides, K.M. KozloffAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): J. Dai, A. Karatsinides, J.M. Keller, K.M.Kozloff, E.T. KellerWriting, review, and/or revision of the manuscript: J. Dai, K.M. Kozloff,D.T. Aftab, F. Schimmoller, E.T. KellerStudy supervision: J. Dai, E.T. Keller

Grant SupportThis work was supported by NIH grant P01 CA093900 and Exelixis Inc.

(to E.T. Keller).The costs of publication of this article were defrayed in part by the

payment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

ReceivedMarch 25, 2013; revisedAugust 20, 2013; accepted September 10,2013; published OnlineFirst October 4, 2013.

References1. Shah RB, Mehra R, Chinnaiyan AM, Shen R, Ghosh D, Zhou M, et al.

Androgen-independent prostate cancer is a heterogeneous group ofdiseases: lessons from a rapid autopsy program. Cancer Res2004;64:9209–16.

2. Dy SM, Asch SM, Naeim A, Sanati H, Walling A, Lorenz KA. Evidence-based standards for cancer pain management. J Clin Oncol 2008;26:3879–85.

3. Mercadante S, Fulfaro F. Management of painful bone metastases.Curr Opin Oncol 2007;19:308–14.

4. Rajpar S, Fizazi K. Bone targeted therapies in metastatic castration-resistant prostate cancer. Cancer J 2013;19:66–70.

5. Stella GM, Benvenuti S, Comoglio PM. Targeting the MET onco-gene in cancer and metastases. Expert Opin Investig Drugs 2010;19:1381–94.

6. Pisters LL, Troncoso P, Zhau HE, Li W, von Eschenbach AC, ChungLW. c-met proto-oncogene expression in benign and malignanthuman prostate tissues. J Urol 1995;154:293–8.

7. Dai J,KitagawaY, Zhang J,YaoZ,MizokamiA,ChengS, et al. Vascularendothelial growth factor contributes to the prostate cancer-inducedosteoblast differentiation mediated by bone morphogenetic protein.Cancer Res 2004;64:994–9.

8. Kitagawa Y, Dai J, Zhang J, Keller JM, Nor J, Yao Z, et al. Vascularendothelial growth factor contributes to prostate cancer-mediatedosteoblastic activity. Cancer Res 2005;65:10921–9.

9. ZhangS,ZhauHE,OsunkoyaAO, Iqbal S,YangX, FanS, et al. Vascularendothelial growth factor regulatesmyeloid cell leukemia-1 expressionthrough neuropilin-1-dependent activation of c-MET signaling inhuman prostate cancer cells. Mol Cancer 2010;9:9.

10. YakesFM,Chen J, Tan J, YamaguchiK, Shi Y, YuP, et al. Cabozantinib(XL184), a novel MET and VEGFR2 inhibitor, simultaneously sup-presses metastasis, angiogenesis, and tumor growth. Mol CancerTher 2011;10:2298–308.

11. Smith DC, Smith MR, Sweeney C, Elfiky AA, Logothetis C, Corn PG,et al. Cabozantinib in patients with advanced prostate cancer: resultsof a phase II randomized discontinuation trial. J Clin Oncol 2013;31:412–9.

12. Wu TT, Sikes RA, Cui Q, Thalmann GN, Kao C, Murphy CF, et al.Establishing human prostate cancer cell xenografts in bone: inductionof osteoblastic reaction by prostate-specific antigen-producingtumors in athymic and SCID/bg mice using LNCaP and lineage-derived metastatic sublines. Int J Cancer 1998;77:887–94.

13. LeRoy BE, Thudi NK, NadellaMV, Toribio RE, Tannehill-Gregg SH, vanBokhoven A, et al. New bone formation and osteolysis by ametastatic,highly invasive canine prostate carcinoma xenograft. Prostate 2006;66:1213–22.

14. Ellis WJ, Vessella RL, Buhler KR, Bladou F, True LD, Bigler SA, et al.Characterization of a novel androgen-sensitive, prostate-specific anti-gen-producing prostatic carcinoma xenograft: LuCaP 23. Clin CancerRes 1996;2:1039–48.

15. Franceschi RT, Iyer BS, Cui Y. Effects of ascorbic acid on collagenmatrix formation and osteoblast differentiation in murine MC3T3-E1cells. J Bone Miner Res 1994;9:843–54.

16. Rissanen JP, Ylipahkala H, Fagerlund KM, Long C, Vaananen HK,Halleen JM. Improved methods for testing antiresorptive compoundsin human osteoclast cultures. J Bone Miner Metab 2009;27:105–9.

17. FuZ, Smith PC, Zhang L, RubinMA,DunnRL, YaoZ, et al. Effects of rafkinase inhibitor protein expression on suppression of prostate cancermetastasis. J Natl Cancer Inst 2003;95:878–89.

18. Dai J, Keller J, Zhang J, Lu Y, Yao Z, Keller ET. Bone morphogeneticprotein-6 promotes osteoblastic prostate cancer bone metastasesthrough a dual mechanism. Cancer Res 2005;65:8274–85.

19. Zhang J, Lu Y, Dai J, Yao Z, Kitazawa R, Kitazawa S, et al. In vivo real-time imaging of TGF-beta-induced transcriptional activation of theRANK ligand gene promoter in intraosseous prostate cancer. Prostate2004;59:360–9.

20. Coffelt SB, Hughes R, Lewis CE. Tumor-associated macrophages:effectors of angiogenesis and tumor progression. Biochim BiophysActa 2009;1796:11–8.

21. Keller ET. The role of osteoclastic activity in prostate cancer skeletalmetastases. Drugs Today 2002;38:91–102.

22. Chen J, De S, Brainard J, Byzova TV. Metastatic properties of prostatecancer cells are controlled by VEGF. Cell Commun Adhes 2004;11:1–11.

23. Ferrer FA,Miller LJ, AndrawisRI, KurtzmanSH,AlbertsenPC, LaudoneVP, et al. Angiogenesis and prostate cancer: in vivo and in vitroexpression of angiogenesis factors by prostate cancer cells. Urology1998;51:161–7.

24. Yang AD, Camp ER, Fan F, Shen L, Gray MJ, Liu W, et al. Vascularendothelial growth factor receptor-1 activation mediates epithelial tomesenchymal transition in human pancreatic carcinoma cells. CancerRes 2006;66:46–51.

25. Aftab DT, McDonald DM. MET and VEGF: synergistic targets incastration-resistant prostate cancer. Clin Transl Oncol 2011;13:703–9.

26. Russo AL, Jedlicka K, Wernick M, McNally D, Kirk M, Sproull M, et al.Urine analysis and protein networking identify met as a marker ofmetastatic prostate cancer. Clin Cancer Res 2009;15:4292–8.

27. Zhu X, Humphrey PA. Overexpression and regulation of expression ofscatter factor/hepatocyte growth factor in prostatic carcinoma. Urol-ogy 2000;56:1071–4.






28. Humphrey PA, Zhu X, Zarnegar R, Swanson PE, Ratliff TL, Vollmer RT,et al. Hepatocyte growth factor and its receptor (c-MET) in prostaticcarcinoma. Am J Pathol 1995;147:386–96.

29. Knudsen BS, Edlund M. Prostate cancer and the met hepatocytegrowth factor receptor. Adv Cancer Res 2004;91:31–67.

30. Gupta A, Karakiewicz PI, Roehrborn CG, Lotan Y, Zlotta AR, ShariatSF. Predictive value of plasma hepatocyte growth factor/scatter factorlevels in patients with clinically localized prostate cancer. Clin CancerRes 2008;14:7385–90.

31. Yasuda K, Nagakawa O, Akashi T, Fujiuchi Y, Koizumi K, Komiya A,et al. Serum active hepatocyte growth factor (AHGF) in benign pros-tatic disease and prostate cancer. Prostate 2009;69:346–51.

32. Maeda A, Nakashiro K, Hara S, Sasaki T, Miwa Y, Tanji N, et al.Inactivation of AR activates HGF/c-Met system in human prostaticcarcinoma cells. Biochem Biophys Res Commun 2006;347:1158–65.

33. Verras M, Lee J, Xue H, Li TH, Wang Y, Sun Z. The androgen receptornegatively regulates the expression of c-Met: implications for a novel

mechanism of prostate cancer progression. Cancer Res 2007;67:967–75.

34. Davies G, Watkins G, Mason MD, Jiang WG. Targeting the HGF/SFreceptor c-met using a hammerhead ribozyme transgene reduces invitro invasion and migration in prostate cancer cells. Prostate2004;60:317–24.

35. Kim SJ, Johnson M, Koterba K, Herynk MH, Uehara H, Gallick GE.Reduced c-Met expression by an adenovirus expressing a c-Metribozyme inhibits tumorigenic growth and lymph node metastases ofPC3-LN4prostate tumor cells in anorthotopic nudemousemodel. ClinCancer Res 2003;9:5161–70.

36. Dai Y, Siemann DW. BMS-777607, a small-molecule met kinaseinhibitor, suppresses hepatocyte growth factor-stimulated prostatecancer metastatic phenotype in vitro. Mol Cancer Ther 2010;9:1554–61.

37. Tu WH, Zhu C, Clark C, Christensen JG, Sun Z. Efficacy of c-Metinhibitor for advanced prostate cancer. BMC Cancer 2010;10:556.

Dai et al.





2014;20:617-630. Published OnlineFirst October 4, 2013.Clin Cancer Res Jinlu Dai, Honglai Zhang, Andreas Karatsinides, et al. Tumor-Induced Bone LesionsCabozantinib Inhibits Prostate Cancer Growth and Prevents

Updated version

10.1158/1078-0432.CCR-13-0839doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2014/02/04/1078-0432.CCR-13-0839.DC2 http://clincancerres.aacrjournals.org/content/suppl/2013/10/03/1078-0432.CCR-13-0839.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/20/3/617.full#ref-list-1

This article cites 37 articles, 14 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/20/3/617.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://clincancerres.aacrjournals.org/content/20/3/617To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-13-0839

http://clincancerres.aacrjournals.org/content/suppl/2013/10/03/1078-0432.CCR-13-0839.DC1

http://clincancerres.aacrjournals.org/content/suppl/2014/02/04/1078-0432.CCR-13-0839.DC2

http://clincancerres.aacrjournals.org/content/20/3/617.full#ref-list-1

http://clincancerres.aacrjournals.org/content/20/3/617.full#related-urls

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/20/3/617


Documents

Cabozantinib Inhibits Prostate Cancer Growth and Prevents ... · based on retrospective review. Among the patients with stable disease at week 12 that were randomly assigned to cabozantinib