Zoledronic acid protects against osteosarcoma-induced bone destruction but lacks efficacy against pulmonary metastases in a syngeneic rat model

Zoledronic acid protects against osteosarcoma-induced bonedestruction but lacks efficacy against pulmonary metastasesin a syngeneic rat model

Agatha Labrinidis, Shelley Hay, Vasilios Liapis, David M. Findlay and Andreas Evdokiou

Discipline of Orthopaedics and Trauma, University of Adelaide, The Royal Adelaide Hospital and The Hanson Institute, Adelaide, South Australia, Australia

Osteosarcoma (OS) is the most common primary malignant tumor of bone in children and adolescents. In spite of successful

control of the primary tumor, death from lung metastasis occurs in more than a third of patients. To investigate the efficacy

of zoledronic acid (ZOL) on the development, progression and metastatic spread of OS, we used a rat model of OS, with

features of the disease similar to human patients, including spontaneous metastasis to lungs. Rat OS cells were inoculated

into the tibial marrow cavity of syngeneic rats. OS development was associated with osteolysis mixed with new bone

formation, adjacent to the periosteum and extended into the surrounding soft tissue. Metastatic foci in the lungs formed 3–4

weeks postcancer cell transplantation. Treatment with a clinically relevant dose of ZOL was initiated 1 week after tumors were

established and continued once weekly or as a single dose. ZOL preserved the integrity of both trabecular and cortical bone

and reduced tumor-induced bone formation. However, the overall tumor burden at the primary site was not reduced because

of the persistent growth of cancer cells in the extramedullary space, which was not affected by ZOL treatment. ZOL treatment

failed to prevent the metastatic spread of OS to the lungs. These findings suggest that ZOL as a single agent protects against

OS-induced bone destruction but lacks efficacy against pulmonary metastases in this rat model. ZOL may have potential value

as an adjuvant therapy in patients with established OS.

Osteosarcoma (OS) is defined as an osteoid-producing malig-nant sarcoma. It is the most frequent primary malignancy ofthe skeleton in children and adolescents, developing mainlybefore the age of 30.1,2 Current standard treatment consistsof surgery and various combinations of chemotherapy.3

Treatment of OS has undergone considerable changes overthe past 20 years, with more efficacious chemotherapy signifi-cantly improving long-term survival. However, response tochemotherapy depends on the type and combination of drugsused, the doses given and the sensitivity/resistance of the tu-mor cells. Further, despite the recent advances, the develop-ment of drug resistance to chemotherapy remains a problem

in OS therapy.4 Metastatic spread, preferentially to the lungscompared with other sites, is seen in more than a third ofpresenting patients, and 90% of recurrent patients, and iscorrelated with extremely poor survival statistics.5–9 There-fore, therapies that could inhibit metastatic OS disease wouldhave considerable potential to reduce mortality in OS.

Bone lesions caused by OS are characterized on the basisof their radiologic appearance and appear as osteolytic, osteo-blastic (osteosclerotic) or a combination of both.10 Osteolysisis a common manifestation of OS, even within predominantlyosteoblastic lesions, and is mediated primarily by osteoclastsand their bone resorbing activity.1,2 Factors released from thebone are believed to stimulate tumor growth and tumor cellsare in turn able to produce factors that stimulate osteoclastdifferentiation and activity, resulting in the establishment of amutually beneficial relationship, often termed ‘‘the viciouscycle’’ because of its progressively destructive nature.11 Con-versely, tumor cells associated with osteoblastic lesions maystimulate osteogenesis.12,13

Bisphosphonates (BPs) are commonly used for the pre-vention and treatment of various bone diseases characterizedby increased bone resorption, such as Paget’s disease andosteoporosis.14 The highly selective localization and retentionof BPs in bone is the basis for their use in skeletal disorders.Nitrogen-containing BPs, such as zoledronic acid (ZOL), in-hibit bone resorption by preventing prenylation of GTPases,such as Ras and Rho, which are required for many cellularprocesses and ultimately induce cell death in osteoclasts.15 In

Key words: osteosarcoma, zoledronic acid, bisphosphonate,

metastasis, osteolysis

Grant sponsors: National and Medical Research Council of

Australia (NHMRC), National Breast Cancer Foundation (NBCF),

Cancer Council of South Australia (CCSA), Australian Orthopaedics

Association (AOA)

DOI: 10.1002/ijc.25051

History: Received 17 Sep 2009; Accepted 12 Nov 2009; Online

18 Nov 2009

Correspondence to: Andreas Evdokiou, Discipline of Orthopaedics

and Trauma, University of Adelaide, Level 4, Bice Building, Royal

Adelaide Hospital, North Terrace, Adelaide 5000, South Australia,

Australia, Fax: 618-8232-3065,

E-mail: [email protected]

Can

cerCellBiology

Int. J. Cancer: 127, 345–354 (2010) VC 2009 UICC

International Journal of Cancer

IJC

addition to their antiresorptive activity, there is growing evi-dence supporting the direct effects of BPs on cancer cells, atleast in vitro. In this respect, ZOL exhibits the highest po-tency of its class. A wide variety of tumor cell types havebeen used to demonstrate the antitumor activity of BPs,including leukemia,16 breast cancer,17 prostate cancer18 andOS.19–23 These studies have shown that BPs can dosedependently inhibit proliferation and induce apoptosis intumor cells. A reduction in tumor cell adhesion, invasion andangiogenesis has also been reported, potentially makingBPs attractive agents in the treatment of metastaticmalignancies.24,25

Preclinical animal models of metastatic cancer have dem-onstrated a reduction in tumor-induced osteolysis with ZOLtreatment.25,26 Reports of animal models of prostate cancerhave all shown reduced osteolysis with ZOL treatment, withconflicting results in osteoblastic lesions.27 In the clinic, BPtreatment is the current standard practice for palliative treat-ment of bone metastases.14 ZOL has proven to be effective inlarge nonrandomized clinical trials and is the first BP toshow significant clinical benefit in patients with bone metas-tases from various primary tumors.25 Prolonged treatmentwith ZOL seems to be safe and well tolerated and this combi-nation of potency and safety makes it a useful adjuvant ther-apy in bone metastasis,28 although prolonged treatment withnitrogen-containing BPs in cancer patients appears to beassociated with the risk of developing osteonecrosis of thejaws.29

The efficacy of ZOL against mouse and rat OS has beenevaluated previously by 2 separate studies using animal mod-els, in which tumor cells were injected either intravenously orwere in contact with the bone surface.30,31 OS originates inthe bone so that animal models with ectopic tumor implanta-tion, such as intravenous injections or subcutaneous implan-tation, lack relevance for patients with OS. We have recentlydemonstrated the ability of ZOL to inhibit the developmentand progression not only of the osteolytic but also the osteo-blastic component of primary OS lesions in an immunocom-promised murine model. However, in the mouse model, lungmetastases were not reduced and may even have been pro-moted with ZOL treatment, highlighting the need for furtherinvestigation before clinical application of ZOL is consideredfor OS therapy.23 An important limitation of the immuno-compromised moue model was that it did not take intoaccount the role of the immune system in the establishmentand metastasis of OS. Therefore, in our study, we establisheda syngeneic animal model, in which rat OS cells were inocu-lated directly into the tibial marrow cavity of immune-com-petent rats. We report here that treatment with clinically rel-evant doses of ZOL following tumor establishment had asignificant protective effect on OS-induced bone destructionbut lacked efficacy in reducing the overall tumor burden atthe primary site or in preventing pulmonary metastases inthis rat model that closely mimics the clinical features ofpatients with OS.

Material and MethodsCells and reagents

The rat OS cell line MSK-8G32 was kindly provided by Dr.Paul Reynolds (Hanson Institute, Adelaide, Australia), andthe cells were cultured in Dulbecco’s modified Eagle’s me-dium, supplemented with glutamine (2 mM), penicillin (100IU/ml), streptomycin (100 lg/ml), gentamicin (160 lg/ml)and 10% fetal bovine serum (Biosciences, Sydney, Australia),in a humidified atmosphere containing 5% CO2.

ZOL was generously provided by Novartis Pharma AG(Basel, Switzerland).

Measurement of cell viability

For determination of ZOL effects on cell growth, 1 � 104

cells per well were seeded into 96-well microtiter plates andallowed to adhere overnight. Cells were then incubated withfresh media containing increasing concentrations of ZOL(1–100 lM) for 72 hr. Cell viability was determined by stain-ing with crystal violet and measuring optical density at 570nm wavelength. The experiments were performed in quadru-plicate and were repeated at least twice. The results of repre-sentative experiments are given as the mean 6 SD.

Apoptosis analysis

DAPI staining of nuclei. Cells were seeded on plastic chamberslides and treated with 40 lM ZOL for 48 hr. After 2 washeswith PBS, cells were fixed in methanol for 5 min, washedagain with PBS and incubated with 0.8 mg/ml of 40,6-diami-dine-20-phenylindole dihydrochloride (DAPI, Roche Diagnos-tics, Castle Hill, NSW, Australia) in PBS for 15 min at 37�C.After several washes in PBS, the coverslips were mounted onPBS/glycerol. DAPI staining was visualized by fluorescencemicroscopy.

Measurement of DEVD-caspase activity. DEVD-caspase wasassayed by cleavage of zDEVD-AFC (z-asp-glu-val-asp-7-amino-4-trifluoro-methyl-coumarin), a fluorogenic substratebased on the peptide sequence at the caspase-3 cleavage siteof poly(ADP-ribose) polymerase. Cells (5 � 105) grown in24-well plates were treated as indicated, washed once withHBSS and resuspended in 200 ll of NP-40 lysis buffer con-taining 5 mM Tris-HCl, 5 mM EDTA and 0.5% NP40, pH7.5. After 15 min in lysis buffer at 4�C, insoluble materialwas pelleted at 15,000g and an aliquot of the lysate was testedfor protease activity. To each assay tube containing 8 lM ofsubstrate in 1 ml of protease buffer (50 mM Hepes, 10% su-crose, 10 mM DTT and 0.1% CHAPS, pH 7.4) was added 20ll of cell lysate. After 4 hr at room temperature, fluorescencewas quantified (Exc 400, Emis 505) in a Perkin Elmer LS50spectrofluorimeter. Optimal amounts of added lysate and du-ration of assay were taken from linear portions of curves asdetermined in preliminary experiments. One unit of caspaseactivity was taken as 1 fluorescence unit (at slit widths of10 nm) per 4 hr incubation with substrate.

Can

cerCellBiology

346 ZOL protects against OS-induced bone destruction


Detection of unprenylated and total Rap1A

To determine the effect of ZOL on the prenylation of smallGTPases in the OS cells, lysates from ZOL-treated cells wereanalyzed by Western blotting for the presence of the unpre-nylated form of Rap 1A. Cells were seeded into 25 cm2 flasksand were incubated for 72 hr with media containing 40 lMZOL. Cells were lysed in buffer containing 10 mM Tris HCl,pH 7.6, 150 mM NaCl, 1% Triton X-100, 0.1% sodium do-decyl sulfate (SDS), 2 mM sodium vanadate and a cocktail ofprotease inhibitors (Roche Diagnostics, NSW AUS) andstored at �70�C until ready to use. Cell extracts were mixedwith an equal volume of sample buffer containing 12 mMTris HCl, pH 6.8, 6% SDS, 10% b-mercaptoethanol, 20%glycerol and 0.03% bromophenol blue. Protein samples wereboiled for 5 min and electrophoresed under reducing condi-tions in 4–20% polyacrylamide gels. Separated proteins wereelectrophoretically transferred to PVDF transfer membrane(Novex, San Diego, CA) and blocked in PBS containing 5%blocking reagent (Amersham, Castle Hill, NSW, Australia)for 1 hr at room temperature. Immunodetection was per-formed overnight at 4�C in PBS/blocking reagent containing0.1% Tween 20, using anti-Rap 1A antibody, which is specificfor the unprenylated form of Rap 1A, or an anti-Rap 1Aantibody, which detects total Rap 1A (both antibodies fromSanta Cruz Technology, CA), diluted according to the manu-facturer’s instructions. Filters were rinsed several times withPBS containing 0.1% Tween 20 and incubated with 1:5,000dilution of anti-goat alkaline phosphatase-conjugate (Amer-sham Biosciences, Castle Hill, NSW, Australia) for 1 hr.Bound proteins were detected and quantitated using the Vis-tra ECF substrate reagent kit (Amersham) using a FluorIm-ager (Molecular Dynamics, Sunnyvale, CA).

Rat model of osteosarcoma

Male Fischer 344 (F344) rats, 5–6 weeks old, were housed inaccordance with the guidelines approved by the Institute ofMedical and Veterinary Science animal ethics research com-mittee. Animals were acclimatized for 1 week before com-mencement of procedures. The rats were anesthetized by i.p.injection with 80 mg ketamine/kg body weight and 10 mgxylazine/kg body weight. The left tibia was shaved free ofhair, wiped with 70% ethanol and a 23-gauge needle wasinserted through the tibial plateau with the knee flexed, and1 � 105 MSK-8G cells resuspended in 50 ll of PBS wereinjected into the marrow space. All animals were injectedwith PBS in the contralateral tibia, as a control. After tibialinjection, rats were randomly assigned to 3 groups of 6 ani-mals each. Rats were weighed regularly and radiographs weretaken every 2 weeks to determine the extent of osteolysis. Atsacrifice, all the major organs and both hind limbs were col-lected for micro-CT and histological analysis.

Treatment with zoledronic acid

ZOL was dissolved in sterile water and prepared in sterile 1�PBS. ZOL at 100 lg/kg body weight was administered by s.c.

injection at weekly intervals for 5 weeks, starting 1 week aftercancer cell implantation. On a mg/kg basis, this dose of ZOLis approximately equivalent to the approved human dose of 4mg i.v., but the weekly dosing frequency is higher than theonce monthly regimen used clinically in oncology patientswith bone metastases. Therefore, in a second cohort of ani-mals, ZOL was administered as a single dose only of 100 lg/kg, equivalent to the clinical dose of 4 mg given to patientsonce monthly.33

Radiography

Animals were anesthetized by i.p. injection with 80 mg keta-mine/kg body weight and 10 mg xylazine/kg body weightand were laid onto X-ray film with hind limbs spread to ena-ble clear imaging of the tibiae. Radiographs were taken usingthe HP cabinet X-Ray System-Faxitron Series. Exposure timeand intensity was optimized with final settings of 18 sec at60 kVp. Kodak min R-2000 film was used (Kodak, Austral-asia Pty, Melbourne, Australia).

Microcomputer tomography analysis

For micro-CT imaging, hind limbs were dissected and placedin 100% ethanol. Both the right and left tibiae of each animalwere mounted in the CT specimen tube and placed securelyinto a SkyScan-1072 X-ray micro-CT Scanner (Aartselaar,Belgium). The program was commenced with magnificationset to give scan slices of 18 lm. Three-dimensional (3D)images were generated using Cone-Beam reconstruction and3D visualization (Skyscan). Using the 2D images obtainedfrom the micro-CT scan, the growth plate was identified and500 sections, starting from the growth plate/tibial interfaceand moving down the tibia, were selected. Histograms repre-senting trabecular bone volume (TBV, mm3) were generatedand compared to the control tibia.

Histology

Tibiae were fixed in 10% buffered formalin, followed byEDTA decalcification in 10% EDTA solution and 7% nitricacid at room temperature. Decalcification was confirmed byradiography before sectioning. Samples were paraffin embed-ded, sectioned longitudinally at 6 lM and stained with hema-toxylin and eosin (H&E). Analysis was performed on aNikon Eclipse TE300 inverted microscope (Nikon Corpora-tion, Tokyo, Japan). Tumor area was calculated in a blindedfashion from digitized images of histological slides obtainedfrom a standard 5 megapixel resolution camera coupled to amicroscope and using Scion Imaging software (Scion Corpo-ration, MA). Tumor size, defined as the tumor area, was cal-culated from the section of the tibia that best represents thecenter of the tumor mass and expressed as an average tumorarea per group in absolute units (mm2). For pulmonary me-tastases, sections were stained with H&E, and the numberand size of each lung metastasis were measured with Scionimage analysis software. Representative lung samples wereresin embedded, sectioned longitudinally at 6 lM and stained

Can

cerCellBiology

Labrinidis et al. 347


with Von Kossa or Alizarin Red. Slides were fixed in 10%formalin and washed twice with dH2O. One percent silver ni-trate was added for 30 min in direct sunlight for Von Kossastaining, or 2% Alizarin Red for 5 min at room temp, thenwashed twice with dH2O. Von Kossa stains were finallyrinsed with 2.5% sodium thiosulfate for 5 sec and allowed todry. Tartrate-resistant acid phosphatase (TRAP) staining ofosteoclasts was performed using the leukocyte acid phospha-tase (TRAP) kit, as per the manufacturer’s instructions(Sigma-Aldrich, St. Louis, MO).

Statistical analysis

The continuous outcome bone volume was analyzed usingmixed model ANOVA to allow for clustering of rats (i.e.,more than 1 observation per rat). Post hoc pair-wise compari-sons were undertaken with no adjustment made for multiplecomparisons. In all cases, p < 0.05 was considered statisti-cally significant.

ResultsEffect of ZOL on viability of MSK-8G cells in vitro

The effect of ZOL on the viability of the MSK-8G rat OScells was tested in vitro. Treatment with ZOL for 72 hrresulted in a dose-dependent inhibition of cellular prolifera-tion of MSK-8G cells in monolayer cultures (Fig. 1a). Expo-sure to greater than 10 lM ZOL resulted in cells detachingfrom the substratum within 24–48 hr, similar to effects wereported previously for human OS cells.20 Loss of attachmentwith ZOL treatment was concomitant with a dose-related

increase in caspase-3-like activity (Fig. 1a). We have previ-ously published data showing that caspase-3 activation is sec-ondary to the apoptotic activity of ZOL in OS.20 Morphologi-cal evidence characteristic of apoptosis, including chromatincondensation, nuclear fragmentation and the formation ofdense rounded apoptotic bodies, was seen at doses greaterthan the half maximal effective dose of 40 lM ZOL, asassessed by DAPI staining of nuclei (Fig. 1b). Lysates fromuntreated cells and cells treated with 40 lM ZOL were col-lected after 48 hr and analyzed by Western blotting. Con-comitant with apoptosis induction was cleavage of the apo-ptosis target protein PARP with ZOL treatment (Fig. 1c).Also, consistent with the importance of the mevalonate path-way as an intracellular target for the effects of ZOL, theuntreated cells expressed only the prenylated form of Rap1A,whereas in the ZOL-treated cells both prenylated and unpre-nylated RAP 1A were present (Fig. 1c).

Rat model of OS development, progression and

metastatic spread

The rat model chosen for these studies recapitulates the fea-tures of development of OS and the spontaneous metastasisto lungs seen in the human disease, in the context of anintact immune system. MSK-8G rat OS cells were injecteddirectly into the tibial marrow cavity of F344 rats, using themethod we have previously described.23 All animals injectedwith OS cells developed tumors, which were visible by 3weeks after injection. Visible and palpable tumors confirmedsuccessful implantation and indicated that tumors had

Figure 1. Activity of Zol against MSK-8G OS cells in vitro. (a) The rat MSK-8G OS cells line were seeded in 96-well plates at 1 � 104 cells

per well and treated with increasing doses of ZOL, as indicated. Cell viability (n ) was assessed by crystal violet staining 72 hr after

treatment. Data are presented as the mean 6 SD of quadruplicate wells and are expressed as a percentage of the number of control cells.

Cell lysates were collected and used to determine caspase-3-like activity (~), using the caspase-3-specific fluorogenic substrate, zDEVD-

AFC, as described in the Material and Methods. Data points show means of quadruplicate results from a representative experiment,

repeated at least twice. (b) Cells were treated with 40 lM ZOL and left for 48 hr before cells were fixed with methanol and incubated with

DAPI, before washing in PBS and mounting on PBS/glycerin. DAPI staining was visualized by fluorescence microscopy. (c) Western blot

analysis revealed the presence of unprenylated Rap1A protein and cleavage of PARP in cell lysates treated with ZOL but not in the

untreated control cells. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Can

cerCellBiology



penetrated the tibial cortex and extended into the surround-ing soft tissue. X-ray images of tumor-bearing tibiae takenjust before sacrifice confirmed the presence of osteolysis inall animals within the vehicle (PBS)-treated group (Fig. 2a).A characteristic radiodense ‘‘sunburst’’ configuration, consist-ent with new bone formation, was also evident, extendingperpendicular to the cortex and into the surrounding soft tis-sue mass (Fig. 2a). Micro-CT images of the tibiae werereconstructed to produce 3D images. These revealed extensivebone remodeling, characterized by areas of osteolysis mixedwith areas of osteosclerosis in the tumor-bearing tibiae (Fig.2b). Cross sections of the micro-CT images of the normalnontumor-bearing tibiae demonstrated significant corticalbone destruction and loss of trabecular bone in all tumor-bearing tibiae, representing the osteolytic component of OSlesions. In addition, marked spicular new bone formation,extending from the cortex, was clearly evident in all tumor-bearing tibiae, confirming the plain radiographic data (Figs.2b–2d). Histological analysis demonstrated that the spicularnew bone was mineralized woven bone, and although it wasmainly present as an extension from the cortex, it was alsopresent within the extramedullary tumor mass highlightingthe bone-forming ability of these tumors (Fig. 2e).

Effect of ZOL on OS-induced bone destruction

Eighteen animals were inoculated with cancer cells on Day 0and animals were randomly assigned into 3 groups of 6 ratseach group: (i) the control vehicle group, (ii) the ZOL single

dose only group and (iii) the ZOL once weekly treated group.ZOL treatment commenced 7 days after cancer cell trans-plantation, to allow establishment of tumor cells within thebone environment, and was administered at 100 lg/kg subcu-taneously (s.c.)

Radiographic images of bones from both the ZOL-treatedgroup of animals were highly radiodense when comparedwith the vehicle-treated group, reflecting increased bone den-sity due to inhibition of bone resorption. This effect of ZOLwas more pronounced in areas of increased bone turnover,such as the distal femurs and proximal tibiae, and was notrestricted to the tumor site, as the contralateral nontumor-bearing tibiae also showed this effect (Figs. 3a and 3b). Qual-itative micro-CT assessment of bone architecture showed noevidence of osteolysis in both the ZOL treatment groupswhen compared with vehicle treatment. In addition, theamount of spicular new bone formation extending from thecortex was also significantly reduced with ZOL treatment(Fig. 3a). Longitudinal and cross-sectional micro-CT imagesconfirmed the increase in trabecular bone density in both tu-mor- and nontumor-bearing tibiae that were treated withZOL (Figs. 3a and 3b). To quantify the TBV, we comparedthe left tumor-bearing tibiae of the ZOL-treated animals withthat of the vehicle-treated animals at a selected region begin-ning at the growth plate and extending downward 500 lm �18 lm slices, which encompassed all of the OS lesions. TheTBV of both the ZOL-treated groups was significantly higherthan the vehicle alone-treated group (Fig. 4a), thus

Figure 2. OS-induced bone destruction in an immune-competent rat model. (a) Radiographs showing areas of osteolysis and areas of new

bone formation extending from the cortex. (b) Ex vivo representative 3D reconstructed micro-CT images. (c, d) Serial cross section m-CT

views with progressive loss of trabecular and cortical bone. Also shown is an extensive network of mineralized new bone extending

perpendicular from the cortex. (e) Histological sections at low and high magnification showing that new bone formation is mineralized

woven bone predominantly extending from the periosteum but also present within the extramedullary tumor mass (arrow).

Can

cerCellBiology



Figure 3. Radiographic and ex vivo micro-CT images. (a) Tumor-bearing and (b) contralateral nontumor-bearing tibiae from vehicle-treated

and ZOL-treated rats. Transverse and cross section views of representative tibiae show loss of cortical and trabecular bone in tumor-

bearing tibia with spicular new bone extending from the cortex in the vehicle group. ZOL treatment given once only or weekly inhibits OS-

induced bone destruction, reduced spiculation and increased trabecular density in both tumor- and nontumor-bearing tibiae. ZOL treatment

increased radiodensity in areas of higher bone turnover, including proximal tibia and distal femur in both tumor- and nontumor-bearing

tibiae. Depicted are representative images that best illustrate the effect of treatment. White arrows indicate increased radiodensity.

Can

cerCellBiology



confirming the qualitative assessment which showed anincrease in trabecular bone density with ZOL treatment. Theprotective effect of ZOL on OS-induced bone destruction wasdue to a decrease in osteoclastic bone resorption as evidentby the significant decrease in the number of TRAP-positiveosteoclasts lining the bone surface (Fig. 4b). This effect wasmost pronounced in the ZOL weekly treatment group.

Effect of ZOL on primary tumor growth and on

pulmonary metastases

The intra- and extraosseous tumor burden in the tibiae follow-ing ZOL treatment was calculated from histological sections

and expressed as an average tumor area per group (Fig. 5a).Treatment with ZOL either as a single dose or as a weeklydose did not decrease the extramedullary tumor growth in thetibiae as assessed by histomorphologic analysis at the end ofthe experiment. In contrast, from histological inspection, theintraosseous tumor burden was significantly decreased withboth of the ZOL treatment regimens and cancer cells werebarely detectable within the bone marrow space, which hadnow been significantly reduced and spatially constrainedbecause of the increase in trabecular bone density (Fig. 5b).

Pulmonary metastases were present in all animals, irre-spective of treatment. The metastatic tumor nodules in the

Figure 4. Effect of ZOL on bone volume. Trabecular bone volume was measured from 500 micro-CT slices of the proximal tibia starting from the growth

plate, using the program CTan. (a) Graph represents the trabecular bone volume in mm3 in the tumor-bearing tibiae of each group, **represents p¼0.001. Data shown in each case are the average bone volume from all animals in that group: points are means6 SEM. (b) Histological sections (n¼ 4

per group ) showing TRAP staining of osteoclasts in the tibia of untreated and ZOL single and weekly treated rats, respectively.

Figure 5. Quantification of the effect of ZOL on primary tumor growth. (a) Total tumor burden was measured using the histology images and the

area expressed as an average per group is shown in the graph as mm2 (data are expressed as mean6 SEM). (b) Histological sections from

representative tibiae stained with H&E, showing intraosseous (IO) tumor burden in the vehicle animal only and extramedullary tumor burden (EM).

Can

cerCellBiology



lungs were highly mineralized and stained positive for VonKossa and Alizarin Red, confirming the presence of phos-phate deposits within the tumor mass (Fig. 6a). Tumor bur-den was quantified ex vivo using histology and calculated asa percent of total lung area. The tumor burden in the lungswas similar between the 2 ZOL treated groups and was notsignificantly different from that of the vehicle-treated animals(Fig. 6b). Similarly, the number of metastatic nodules was nodifferent between the groups (Fig. 6c). Histological analysis ofkidney, liver, heart and spleen tissue sections obtained at au-topsy from animals transplanted with OS cells showed noevidence of metastatic spread in these organs (data notshown).

DiscussionOS has a variable bone-forming ability but is destructive byvirtue of its ability to expand in bone by inducing osteoclast-mediated bone resorption. The effect of BPs on OS cells, spe-cifically alendronate and pamidronate, has been reported ashaving proapoptotic effects on canine21 and human34,35 OScells in vitro. Three recent studies have reported on the effectof ZOL in animal models of mouse and rat OS. Ory et al.investigated a mouse model, in which mouse OS cells wereinjected intravenously and showed that ZOL treatment sup-pressed lung metastases and prolonged the overall survival ofOS-bearing mice.31 Heymann et al. also showed that ZOLcompletely prevented lung metastases and caused reductionin osteolytic lesions in a rat model, in which intact rat OStissue was implanted subcutaneously and in contact with thetibial bone surface.30 Using a subcutaneous implanted mouseOS model, Koto et al. showed that clinically relevant doses ofZOL administered i.p inhibited lung metastases but had noeffect on the growth of OS at the primary site.36 Our experi-mental model involves the direct transplantation of OS cellsinto the tibial marrow cavity, attempting to simulate the nor-mal development, progression and metastatic spread of OSand to enable the evaluation of ZOL treatment in the boneenvironment. We previously demonstrated that ZOL is ableto inhibit the development and progression not only of theosteolytic but also the osteoblastic component of primary OSlesions in a nude mouse model using human OS cells. How-ever, we showed in the same study that lung metastases werenot reduced and may even have been promoted, with ZOLtreatment, indicating that caution is required before any clin-ical application of ZOL is considered for OS therapy.23 Morerecently, using a similar animal model, Dass and Choong37

have shown that ZOL administration to SAOS-2 tumor-bear-ing mice resulted in primary tumor growth inhibition, reduc-tion in lung metastases and a dramatic decrease in osteolysis.The reasons for these different observations are not clear butmay relate to cell type, dose of ZOL and scheduling of treat-ment protocols. For example, the dosing schedule of the Dassstudy equates to 8 times the current clinical dose.

Our previous mouse study had 1 important limitation, inthat it did not take into account the role of the immune sys-

tem in tumor growth and metastasis. Therefore, in our study,we established a syngeneic animal model, in which rat OScells were inoculated directly into the tibial marrow cavity ofimmune-competent rats. We selected the rat MSK-8G OS cellline, a derivative of the MSK cell line.32 MSK-8G cells whenimplanted directly into the tibial marrow cavity developlocally, growing tumors of mixed osteolytic/osteoscleroticlesions when analyzed radiographically and histologically.One added advantage is that, as in the human disease, MSK-8G cells injected intratibially form pulmonary metastases thatare easily quantifiable upon gross and histological inspectionof the lungs 3–4 weeks after tumor cell inoculation.

Although ZOL treatment offered considerable protectionagainst OS-induced bone destruction, there was no significanteffect on the overall tumor burden (intraosseus and extrame-dullary) at the primary site. However, the intraosseus tumor

Figure 6. Effects of ZOL treatment on pulmonary metastases.

Histological sections of the lungs were photographed and used to

measure tumor area and total lung area. (a) Tumor within lung was

confirmed using histological slides stained with H&E and

compared to normal lung histology. The mineralization status of

the lung metastases was confirmed using histological slides

stained with Von Kossa and Alizarin Red. (b) Graph represents the

tumor area as a percent of total lung area and is expressed as an

average per group. (c) Graph represents the average number of

metastatic loci per group. Data shown in each case are an average

from a representative section of each animal: points are means 6

SEM.

Can

cerCellBiology



burden diminished considerably such that tumor cells werebarely detectable within the bone marrow, as shown in histo-logic sections of the tibiae. It should be noted that the trabec-ular bone density was considerably increased with ZOL treat-ment, an effect that was more pronounced with the weeklyZOL treatment regimen. This effect of ZOL translated to asignificant reduction in bone marrow volume, contributing tothe spatial constraints preventing tumor cell growth withinthe marrow space but not in the surrounding soft tissue. Theeffects of ZOL on trabecular bone density were due to theactions on osteoclast formation and activity as ZOL treat-ment decreased the number of TRAPþve osteoclasts liningthe bone surfaces. Consistent with our findings, Buijs et al.38

have recently shown that ZOL or Fc-OPG treatment follow-ing intratibial injection of highly osteolytic breast cancer cellsprotected against cancer-induced osteolysis and although theintraosseus tumor burden was diminished, the overall tumorburden was also not significantly affected because of the per-sistent growth in the extramedullary space.

In the context of the effects of ZOL on OS-induced bonedestruction and on primary tumor growth, the results pre-sented here are in complete agreement with those we haverecently published.23 However, although our previous studydemonstrated an increase in pulmonary metastases with ZOLtreatment in immune-compromised animals, this effect wasnot seen in this syngeneic rat model of OS, where theimmune system is intact. ZOL was shown previously todirectly stimulate gamma-delta T cells to secrete high levelsof proinflammatory cytokines leading tumor cell death

in vitro.39–42 A previous study in severe combined immuno-deficiency mice demonstrated that mice reconstituted withhuman peripheral blood lymphocytes or purified gamma-delta T cells were able to kill tumor cells in vivo.39 In a clini-cal trial, Wilhelm et al.43 reported that BP therapy led to astrong correlation between gamma-delta T cell expansion andpatient response, highlighting the potential contribution ofthe immune system in mediating the anticancer actions ofZOL in vivo. It should be noted, however, that in our study,ZOL treatment failed to restrain pulmonary metastases, indi-cating that the immune system is unlikely to provide signifi-cant therapeutic benefit in the treatment of OS.

In conclusion, we demonstrate that ZOL as a single agentprotects against OS-induced bone destruction but lacks effi-cacy against pulmonary metastases in this syngeneicimmune-competent rat model. The important question thatneeds to be addressed is whether a combinatorial approachof ZOL with clinically relevant chemotherapeutic drugs maybe a useful therapeutic option for the treatment of OSpatients in the future. This issue will need to be addressed inproperly designed preclinical and clinical studies.

AcknowledgementsA. Evdokiou is a research fellow of the National Health and MedicalResearch Council of Australia (NHMRC) and A Labrinidis is a postdoctoralfellow supported by the National Breast Cancer Foundation (NBCF). Theauthors thank Dr. Jonathan Green (Novartis, Switzerland), Prof. PeterChoong (St Vincent’s Hospital, Vic, Australia), Dr. Paul Reynolds (IMVS,SA, Australia), Dr. Masakazo Kogawa (IMVS, SA, Australia) and Dr. KikuoTakahashi (Chiga University Hospital, Japan) for help and advice.

References

1. Campanacci M. Bone and soft tissuetumors, 2nd edn. Padova: Piccin NuovaLibraria, 1999.

2. Pringle JAS. Bone-forming neoplasmsarising within bone. In: Helliwell TR, ed.Pathology of bone and joint neoplasmsed.Philadelphia: Saunders, 1999; p. 168–92.

3. Picci P. Osteosarcoma (Osteogenicsarcoma). Orphanet J Rare Dis 2007;2:6.

4. Chan HS, Grogan TM, Haddad G, DeBoerG, Ling V. P-glycoprotein expression:critical determinant in the response toosteosarcoma chemotherapy. J Natl CancerInst 1997;89:1706–15.

5. Ferrari S, Palmerini E. Adjuvant andneoadjuvant combination chemotherapyfor osteogenic sarcoma. Curr Opin Oncol2007;19:341–6.

6. Khanna C. Novel targets with potentialtherapeutic applications in osteosarcoma.Curr Oncol Rep 2008;10:350–8.

7. Link MP, Goorin AM, Horowitz M, MeyerWH, Belasco J, Baker A, Ayala A, ShusterJ. Adjuvant chemotherapy of high-gradeosteosarcoma of the extremity. Updatedresults of the multi-institutional

osteosarcoma study. Clin Orthop Relat Res1991;270:8–14.

8. Longhi A, Errani C, De Paolis M, MercuriM, Bacci G. Primary bone osteosarcoma inthe pediatric age: state of the art. CancerTreat Rev 2006;32:423–36.

9. Saeter G, Alvegard TA, Elommaa I, WiebeT, Bjork O, Strander H, Solheim OP.Chemotherapy for osteosarcoma andEwings sarcoma. Acta Orthop Scand Suppl1997;273:120–5.

10. Mundy GR. Metastasis to bone: causes,consequences and therapeuticopportunities. Nat Rev Cancer 2002;2:584–93.

11. Chirgwin JM, Guise TA. Molecularmechanisms of tumor-bone interactions inosteolytic metastases. Crit Rev EukaryotGene Expr 2000;10:159–78.

12. Goltzman D. Mechanisms of thedevelopment of osteoblastic metastases.Cancer 1997;80:1581–7.

13. Goltzman D, Karaplis AC, Kremer R,Rabbani SA. Molecular basis of thespectrum of skeletal complications ofneoplasia. Cancer 2000;88:2903–8.

14. Body JJ. Current and future directions inmedical therapy: hypercalcemia. CancerCell 2000;88:3054–8.

15. Zhang FL, Casey PJ. Protein prenylation:molecular mechanisms and functionalconsequences. Annu Rev Biochem 1996;65:241–69.

16. Kimura S, Kuroda J, Segawa H, Sato K,Nogawa M, Yuasa T, Ottmann OG,Maekawa T. Antiproliferative efficacy ofthe third-generation bisphosphonate,zoledronic acid, combined with otheranticancer drugs in leukemic cell lines. IntJ Hematol 2004;79:37–43.

17. Senaratne SG, Colston KW. Direct effectsof bisphosphonates on breast cancer cells.Breast Cancer Res 2002;4:18–23.

18. Lee MV, Fong EM, Singer FR, GuenetteRS. Bisphosphonate treatment inhibits thegrowth of prostate cancer cells. Cancer Res2001;61:2602–8.

19. Benassi MS, Chiechi A, Ponticelli F,Pazzaglia L, Gamberi G, Zanella L, ManaraMC, Perego P, Ferrari S, Picci P. Growthinhibition and sensitization to cisplatin byzoledronic acid in osteosarcoma cells.Cancer Lett 2007;250:194–205.

Can

cerCellBiology



20. Evdokiou A, Labrinidis A, Bouralexis S,Hay S, Findlay DM. Induction of cell deathof human osteogenic sarcoma cells byzoledronic acid resembles anoikis. Bone2003;33:216–28.

21. Farese JP, Ashton J, Milner RAL, VanGilder J. The effect of the bisphosphonatealendronate on viability of canineosteosarcoma cells in vitro. In Vitro CellDev Biol Anim 2004;40:113–17.

22. Kubista B, Trieb K, Sevelda F, Toma C,Arrich F, Heffeter P, Elbling L, SutterlutyH, Scotlandi K, Kotz R, Micksche M,Berger W. Anticancer effects of zoledronicacid against human osteosarcoma cells.J Orthop Res 2006;24:1145–52.

23. Labrinidis A, Hay S, Liapis V, PonomarevV, Findlay D, Evdokiou A. Zoledronic acidinhibits both the osteolytic and osteoblasticcomponents of osteosarcoma lesions in amouse model. Clin Cancer Res 2009;15:3451–61.

24. Coleman R. On the horizon: canbisphosphonates prevent bone metastases?Breast 2007;16:21–7.

25. Green JR. Bisphosphonates: preclinicalreview. Oncologist 2004;9:3–13.

26. Winter MC, Holen I, Coleman RE.Exploring the anti-tumour activity ofbisphosphonates in early breast cancer.Cancer Treat Rev 2008;34:453–75.

27. Corey E, Brown LG, Quinn JE, Poot M,Roudier MP, Higano CS, Vessella RL.Zoledronic acid exhibits inhibitory effectson osteoblastic and osteolytic metastases ofprostate cancer. Clin Cancer Res 2003;9:295–306.

28. Ali SM, Esteva FJ, Hortobagyi G, HarveyH, Seaman J, Knight R, Costa L, Lipton A.Safety and efficacy of bisphosphonates

beyond 24 months in cancer patients.J Clin Oncol 2001;19:3434–7.

29. Wutzl A, Eisenmenger G, Hoffmann M,Czerny C, Moser D, Pietschmann P, EwersR, Baumann A. Osteonecrosis of the jawsand bisphosphonate treatment in cancerpatients. Wien Klin Wochenschr 2006;118:473–8.

30. Heymann D, Ory B, Blanchard F,Heymann MF, Coipeau P, Charrier C,Couillaud S, Thiery JP, Gouin F, Redini F.Enhanced tumor regression and tissuerepair when zoledronic acid is combinedwith ifosfamide in rat osteosarcoma. Bone2005;37:74–86.

31. Ory B, Heymann MF, Kamijo A, Gouin F,Heymann D, Redini F. Zoledronic acidsuppresses lung metastases and prolongsoverall survival of osteosarcoma-bearingmice. Cancer 2005;104:2522–9.

32. Takahashi K, Sato K, Egami F, Miya T,Narukawa Y, Kanazawa H. Comparison ofosteoblastic markers of nine clonal celllines from rat osteosarcoma cells. J BoneMiner Metab 1991;9:241–7.

33. Daubine F, Le Gall C, Gasser J, Green JR,Clezardin P. Antitumor effects of clinicaldosing regimens of bisphosphonates inexperimental breast cancer bone metastasis.J Natl Cancer Inst 2007;99:322–30.

34. Cheng YY, Hueng L, Lee KM, Li K, KumtaSM. Alendronate regulates cell invasionand MMP-2 secretion in humanosteosarcoma cell lines. Pediatr BloodCancer 2004;42:410–15.

35. Sonnemann J, Eckervogt V, TruckenbrodB, Boos J, Winkelmann W, van Valen F.The bisphosphonate pamidronate is apotent inhibitor of human osteosarcomacell growth in vitro. Anticancer Drugs 2001;12:459–65.

36. Koto K, Horie N, Kimura S, Murata H,Sakabe T, Matsui T, Watanabe M, AdachiS, Maekawa T, Fushiki S, Kubo T.Clinically relevant dose of zoledronic acidinhibits spontaneous lung metastasis in amurine osteosarcoma model. Cancer Lett2009;274:271–8.

37. Dass C, Choong P. Zoledronic acid inhibitsosteosarcoma growth in an orthotopicmodel. Mol Cancer Ther 2007;6:3263–70.

38. Buijs J, Que I, Lowik C, Papapoulos S, vander Pluijm G. Inhibition of bone resorptionand growth of breast cancer in the bonemicroenvironment. Bone 2009;44:380–6.

39. Kabelitz D, Wesch D, Pitters E, Zoller M.Characterization of tumor reactivity ofhuman V gamma 9V delta 2 gamma deltaT cells in vitro and in SCID mice in vivo.J Immunol 2004;173:6767–76.

40. Kunzmann V, Bauer E, Feurle J,Weissinger F, Tony H, Wilhelm M.Stimulation of gammadelta T cells byaminobisphosphonates and induction ofantiplasma cell activity in multiplemyeloma. Blood 2000;96:384–92.

41. Lopez R. Human gammadelta-T cells inadoptive immunotherapy of malignant andinfectious diseases. Immunol Res 2002;26:207–21.

42. Sato K, Kimura S, Segawa H, Yokota A,Matsumoto S, Kuroda J, Nogawa M, YuasaT, Kiyono Y, Wada H, Maekawa T.Cytotoxic effects of gammadelta T cellsexpanded ex vivo by a third generationbisphosphonate for cancer immunotherapy.Int J Cancer 2005;116:94–9.

43. Wilhelm M, Kunzmann V, Eckstein S,Reimer P, Weissinger F, Ruediger T, TonyH. Gammadelta T cells for immunetherapy of patients with lymphoidmalignancies. Blood 2003;102:200–6.

Can

cerCellBiology



Documents

Zoledronic acid protects against osteosarcoma-induced bone destruction but lacks efficacy against pulmonary metastases in a syngeneic rat model