Preclinical evaluation of zoledronate using an in vitro mimetic cellular model for breast cancer metastatic bone disease

Biochimica et Biophysica Acta 1830 (2013) 3625–3634

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbagen

Preclinical evaluation of zoledronate using an in vitro mimetic cellular model for breast cancer metastatic bone disease P.G. Dedes a, I. Kanakis a, Ch. Gialeli a, d, A.D. Theocharis a, T. Tsegenidis a, D. Kletsas b, G.N. Tzanakakis c, N.K. Karamanos a, d,⁎ a

Laboratory of Biochemistry, Department of Chemistry, University of Patras, 26110 Patras, Greece Laboratory of Cell Proliferation and Ageing, Institute of Biosciences and Applications, National Center of Scientific Research “Demokritos”, Athens, Greece Department of Histology, Medical School, University of Crete, University of Crete, Heraklion, Greece d Foundation for Research and Technology/Institute of Chemical Engineering Sciences (FORTH/ICE-HT), 26500 Patras, Greece b c

a r t i c l e

i n f o

Article history: Received 7 December 2012 Received in revised form 23 January 2013 Accepted 28 January 2013 Available online 5 February 2013 Keywords: Breast cancer Bone metastasis Cellular model Metalloproteinase Matrix macromolecule

a b s t r a c t Background: The interactions between metastatic breast cancer cells and host cells of osteoclastic lineage in bone microenvironment are essential for osteolysis. In vitro studies to evaluate pharmacological agents are mainly limited to their direct effects on cell lines. To mimic the communication between breast cancer cells and human osteoclasts, a simple and reproducible cellular model was established to evaluate the effects of zoledronate (zoledronic acid, ZOL), a bisphosphonate which exerts antiresorptive properties. Methods: Human precursor osteoclasts were cultured on bone-like surfaces in the presence of stimuli (sRANKL, M-CSF) to ensure their activation. Furthermore, immature as well as activated osteoclasts were co-cultured with MDA-MB-231 breast cancer cells. TRAP5b and type I collagen N-terminal telopeptide (NTx) were used as markers. Osteoclasts’ adhesion to bone surface and subsequent bone breakdown were evaluated by studying the expression of cell surface receptors and certain functional matrix macromolecules in the presence of ZOL. Results: ZOL significantly suppresses the precursor osteoclast maturation, even when the activation stimuli (sRANKL and M-SCF) are present. Moreover, it significantly decreases bone osteolysis and activity of MMPs as well as precursor osteoclast maturation by breast cancer cells. Additionally, ZOL inhibits the osteolytic activity of mature osteoclasts and the expression of integrin β3, matrix metalloproteinases and cathepsin K, all implicated in adhesion and bone resorption. Conclusions: ZOL exhibits a beneficial inhibitory effect by restricting activation of osteoclasts, bone particle decomposition and the MMP-related breast cancer osteolysis. General significance: The proposed cellular model can be reliably used for enhancing preclinical evaluation of pharmacological agents in metastatic bone disease. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Bone microenvironment provides a fertile ground for metastatic breast cancer cells to migrate, proliferate and finally colonize on bone surface, establishing skeletal metastases. The consequences are severe skeletal disorders due to stimulation of bone cells, which in turn lead to imbalanced bone remodeling that may be osteolytic, osteoblastic or a combination of both, causing mixed lesions [1]. The processing steps in the development of bone metastasis involve Abbreviations: BPs, bisphosphonates; ZOL, zoledronate; RANKL, receptor activator of nuclear factor-κB ligand; NTx, N-terminal telopeptide; CATK, cathepsin K; OPG, osteoprotegerin; PTHrP, parathyroid hormone-related peptide; M-CSF, macrophage colony-stimulating factor; ECM, extracellular matrix; MMP, matrix metalloproteinase; MT-MMP, membrane type-MMP; TIMP, tissue inhibitors of MMPs ⁎ Corresponding author at: Laboratory of Biochemistry, Department of Chemistry, University of Patras, 26110 Patras, Greece. Fax: +30 2610 997153. E-mail address: [email protected] (N.K. Karamanos). 0304-4165/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bbagen.2013.01.020

primary tumor cells’ invasion by modification and breakdown of the normal surrounding extracellular matrix (ECM), favored by proteolytic enzymes such as matrix metalloproteinases (MMPs) [2]. Tumor cells traverse the walls of both small normal and tumor-induced blood vessels to enter circulation and travel to distant target organs including skeletal tissue, where they establish respective metastases [3–7]. Each of these consecutive steps involves important molecular interactions between the tumor cells and the host cells, and is considered a potential target for the development of drugs that are designed to abrogate the metastatic process [8]. The most crucial process in breast cancer metastasis to bone is the molecular communication between tumor and bone cells [9]. Interactions between tumor cells and osteoclasts in bone microenvironment cause, not only osteoclast activation and subsequent bone resorption, but also aggressive growth and behavior of the tumor cells. Bisphosphonates (BPs), compounds based on a P–C–P spine similar to endogenous pyrophosphate, are able to inhibit osteoclast-induced

3626

P.G. Dedes et al. / Biochimica et Biophysica Acta 1830 (2013) 3625–3634

bone destruction. Zoledronic acid (ZOL), a third generation nitrogencontaining BP (N-BP) found to be the most effective, is approved by the FDA for breast cancer patients with osteolytic lesions to reduce the skeletal complications of malignancy. N-BPs inhibit farnesyl bisphosphate (FPP) synthase, a key enzyme in the mevalonate pathway [10–12]. The result is the loss of FPP and geranylgeranyl diphosphate (GGPP), which are required for the post-translational lipid modification (prenylation) of small GTPases such as Ras, Rho and Rac. Failure to activate small GTPases appears to disrupt downstream signaling pathways, resulting in apoptosis of osteoclasts and tumor cells via a caspasedependent pathway [13]. Preclinical and preliminary clinical data suggest that ZOL alone exerts direct or indirect antitumoral effects on a variety of cancers including breast cancer or when used in combination with known neoadjuvant chemotherapy, common in clinical settings for breast cancer improves its therapeutic effects [14–16]. Recently, we have shown [17] that ZOL inhibits the functional properties of breast cancer cells, as a result of its ability to modulate the expression of key matrix molecules implicated in breast cancer progression [16]. In addition, ZOL prevents the invasion of malignant cells and cell adhesion of osteoclasts on bone tissue, thus reducing bone metastatic status [18]. Furthermore, osteoprotegerin (OPG) production of human osteoblasts is stimulated by ZOL [19]. The osteoclast is a tissue-specific macrophage polykaryon created by the differentiation of monocyte/macrophage precursor cells at or near the bone surface. It is well established that the two hematopoietic factors necessary and sufficient for osteoclastogenesis, are the TNF-related cytokine RANKL and the macrophage colony-stimulating factor (M-CSF) [20–23]. Together, M-CSF and RANKL are required to induce expression of genes that typify the osteoclast lineage, including those encoding tartrate-resistant acid phosphatase (TRAP), cathepsin K (CATK), β3-integrin and MMPs, functional macromolecules of mature osteoclasts. TRAP5b is specifically released by active mature osteoclasts and activated macrophages [24]. In blood serum samples, TRAP can be found in two isoforms alternatively glycosylated, known as TRAP5a and 5b. TRAP5b is of osteoclast origin and 5a of active macrophage origin [25]. An important step in bone resorption is the adhesion of activated osteoclasts on bone surface, a procedure mainly mediated by integrins such as αvβ3 and β3 [26,27]. Collagen type I N-terminal peptides (NTx) are secreted during bone lysis as stable products of catabolism and are considered to be among the most trusted osteolysis markers, due to their unique amino acid sequence and the specific a2 (I) N-telopeptide orientation. Therefore, NTx is one of the byproducts of proteolytic enzyme catabolic activity, produced during bone collagen breakdown by active osteoclasts. The most accurate quantitative detection method of NTx is, to date, competitive ELISA, approved by the FDA, since it uses target-specific monoclonic antibodies against NTx. The tumor-induced activation of osteoclasts results in secretion of well-known proteolytic enzymes that may drive bone resorption e.g. MMP-1, -2, -9, MT1-MMP and CATK. A lysosomal cysteine protease, CATK, degrades the triple helix of collagen I at multiple sites and thus increases its susceptibility to collagenolytic enzymes such as MMP-1 [28]. Furthermore, membrane type 1-MMP (MT1-MMP or MMP-14) expressed in tumor cells activates proMMP-2 produced by stromal cells in the primary focus and at the metastatic site [29]. MMP-9 and MT1-MMP seem to play an important role in the process of osteoclasts’ migration toward bone breakdown sites [30,31]. Moreover, MMP-9 was found to localize to bone resorbing osteoclasts in human breast to bone metastases [32]. Several studies, in in vitro and murine models, have demonstrated that breast cancer cells influence osteoclastic maturation and activity either directly or by stimulating osteoblast-derived osteoclastogenic factors [33,34]. In a murine osteoblast–spleen cell co-culture system, MDA-MB-231, MDA-MB-435 and MCF-7 cells were found to induce osteoclast formation by promoting host IL-11 production and downregulating M-CSF [35]. In addition, osteoclast precursors are sensitive to MDA-MB-231-released factors, leading to increased expression of

key osteoclastogenic transcription factor NFATc1 after a short treatment with RANKL, and final mature osteoclast formation with subsequent release of CATK and MMP-9 [36]. Lau et al. [37] also showed that osteoclast formation and lacunar resorption took place, by a RANKLindependent mechanism, when the conditioned medium from MDAMB-231 and MCF-7 cells was added (with M-CSF) to monocyte cultures. It was, therefore, motivating to evaluate the effect of zoledronate on the function and expression of important ECM macromolecules derived from osteoclasts cultured on bone-like surfaces to facilitate a mimetic local bone microenvironment. More specifically, the present study involves the impact of ZOL on three subsequent levels: 1) on the maturation/activation of precursor (immature) osteoclasts, cultured in the presence of bone microenvironment growth factors, 2) on the maturation/activation of immature osteoclasts co-cultured with the highly invasive breast cancer cells MDA-MB-231, in the presence or absence of growth factors, and 3) on the activity of mature osteoclasts on bone destruction and the expression of certain matrix effective macromolecules. 2. Materials and methods 2.1. Chemicals and reagents FBS (Fetal Bovine Serum), EMEM, sodium pyruvate, sodium bicarbonate, L-glutamine, nonessential amino acids, penicillin, streptomycin, amphotericin B and gentamycin were all obtained from Biochrom (Berlin, Germany). Osteoclast Precursor Basal Medium (OPBM) and OsteoAssay ™ Human Bone Plate were obtained from Lonza (Walkersville, USA). Bovine insulin and p-aminophenylmercuric acetate (APMA) were obtained from Sigma (Steinhelm, Germany). MMP-2 and MMP-9 were obtained from Chemicon (Harrow, UK). Zoledronate was supplied by Novartis Pharmaceuticals (Basel, Switzerland). All other chemicals used were of the best commercially available grade. 2.2. Cell cultures MDA-MB-231 [ATCC HTB 26; human breast adenocarcinoma of high metastatic potential, estrogen receptor α (ERα)-negative] cell line was obtained from the American Tissue Culture Collection (Rockville, MD). MDA-MB-231 cells were cultured in EMEM supplemented with 10% (v/v) FBS, 2 mM L-glutamine, 1.0 mM sodium pyruvate, 1.5 g/L sodium bicarbonate, 0.1 mM nonessential amino acids, 0.01 mg/mL insulin and a cocktail of antimicrobial agents (100 IU/mL penicillin, 100 μg/mL streptomycin, 10 μg/mL gentamicin sulfate and 2.5 μg/mL amphotericin B). Cells were routinely grown at 37 °C in a humidified atmosphere of 5% (v/v) CO2. Culture medium was changed every 48–72 h and the cultures were not left to become confluent. Cells were harvested by trypsinization with 0.05% (w/v) trypsin in PBS containing 0.02% (w/v) Na2EDTA. Poietics™ Osteoclast Precursor Cell System was obtained from Lonza (Walkersville, USA). Precursor osteoclasts were cultured in OPBM supplemented with 10% (v/v) FBS in the presence of 2 mM L-glutamine. It is important to mention that pre- and mature osteoclasts are not able to be re-cultured. In order to achieve differentiation to mature and active osteoclasts, M-CSF and soluble RANKL were added in final concentrations of 33 ng/mL and 66 ng/mL, respectively for 7 days. 2.3. Pre-osteoclasts' culture and co-culture with MDA-MB-231 cells on bone substrate OsteoAssay™ Human Bone Plate remained at room temperature for 1 h. This plate provides a thin layer of adherent human bone particles for the culture of primary human or non-human osteoclasts [38], osteoclast precursors and osteoclast cell lines. Osteoclast precursors (1 × 10 4 cells) were placed in a 48-well plate and were cultured,

P.G. Dedes et al. / Biochimica et Biophysica Acta 1830 (2013) 3625–3634

3627

2.6. Gelatin zymography

in conditions as described above, for 7 days. To examine the effects of breast cancer cells on osteoclast activation and maturation, MDA-MB-231 cells (1 × 10 3 cells) were added to the co-culture system containing 5 × 103 immature osteoclast cells in the presence or absence of sRANKL and M-CSF. Every 3 days 0.3 ml of the old medium was replaced with fresh medium. The co-culture media was collected after 7 days and stored at −80 °C until assayed. Immature osteoclasts were treated in the absence and presence of ZOL within the 7 days of culture for the differentiation of the cells. For experiments with mature osteoclasts the aforementioned procedure was followed as described. On day 7 nutrition medium was removed and replaced with fresh medium to set measurements threshold. Mature osteoclasts were treated for 72 h incubation period and after the media were collected and stored at −80 °C until assayed.

Gelatinase (MMP-2 and -9) activity in the conditioned medium of the co-culture of pre-osteoclasts and/or MDA-MB-231 cells was analyzed by gelatin zymography [40]. More specifically, cell culture supernatant was removed, cell monolayer was washed with PBS and following the treatment with ZOL cells were cultured for another 24 h in serum free media (conditioned media). Samples were analyzed with SDS-PAGE in 8% gel with 1% gelatin, followed by 5 washes with Tris–HCl, Triton X-100 buffer, 15 min each at room temperature followed by 18 h incubation at 37 °C in developing buffer, according to standard procedures. Gels were stained with 0.25% Coomassie Brilliant Blue G-250 (Pierce, Rockford, IL) and de-stained in methanol– acetic acid–water (4:1:4 v/v) at room temperature.

2.4. TRAP5b and NTx determination

2.7. Statistical analysis

TRAP5b was measured by a direct enzyme-linked immunosorbent assay (ELISA) in culture medium samples using the Bone TRAP® Assay (Immunodiagnostic Systems Inc., Boldon, UK) kit, according to the manufacturer's instructions. Quantitative determination of NTx in culture medium samples was also performed by ELISA using the Osteomark® NTx Serum test (Ostex International, Inc., Seattle, USA), following standard kit conditions. TRAP5b and NTx levels were calculated after estimating the total protein content with the Bradford method [39] and were expressed as Units/g (U/g) and nanomolar bone collagen equivalents per g (nM BCE/g), respectively.

In all experiments, the mean values±standard deviations (SD) were calculated for the determinations performed in triplicate. Statistically significant differences were evaluated using the ANOVA test. Differences were considered statistically significant at the level of at least p b 0.05.

2.5. RNA isolation and RT-PCR Total RNA was extracted from osteoclasts and MDA-MB-231 cells using the NucleoSpin RNA II Isolation System (Macherey-Nagal, Germany). The amount of isolated RNA was quantified by measuring its absorbance at 260 nm. All total RNA preparations were free of DNA contamination as assessed by RT-PCR analysis. Total RNA was reverse transcribed using the PrimeScript 1st strand cDNA synthesis kit (TAKARA) and DyNAzyme II DNA Polymerase kit (Finnzymes). Semi-quantitative analysis of cDNA sequences was carried out based on simultaneous amplification of a “house-keeping gene”, glyceraldehyde-3-phosphate dehydrogenase (GAPDH). All amplification products were separated by electrophoresis in a 2% agarose gel, containing Gel Star® stain (BioWhittaker, Rockland, ME, USA). Bands were visualized on a UV lamp and gels were photographed with a CCD camera. The sequences of primers as well as their prime characteristics for the genes of interest are provided in Table 1. For semi-quantitative analysis, gene expression was determined as relative fluorescence obtained for each molecule as compared to the reference gene (GAPDH). Image analysis was performed using the program UNIDocMv version 99.03 for Windows (UVI Tech, Cambridge, UK).

3. Results In order to evaluate the effects of zoledronate on activation of osteoclasts and bone resorption, human precursor osteoclasts were initially cultured in the presence of M-CSF and sRANKL, on bone-like surfaces. Moreover, to facilitate a mimetic local bone microenvironment, immature as well as activated osteoclasts were co-cultured with MDA-MB-231 breast cancer cells. TRAP5b gene expression and protein levels were determined to monitor osteoclast activation, whereas NTx levels were used for screening their actual boneresorbing activity. Furthermore, in order to unravel the importance of MMPs in metastatic bone disease, MMP-1, MMP-9 and MT1-MMP expression levels were determined; MMP-2 and -9 levels were assayed as well by zymography in conditioned media. In addition, cathepsin K and β3-integrin mRNA levels were used to explore osteoclastic resorption and adhesion on bone surface. Notably, as has been reported in a previous study of our labs [17], the use of EDTA doesn't affect the anti-proliferative effects of zoledronate on cancer cells. Thus, it is plausible to suggest that the chelation of Ca +2 is not responsible for the observed effects. 3.1. Zoledronate suppresses precursor osteoclast maturation, bone osteolysis and activity of MMPs To evaluate the effect of ZOL on the activation of precursor osteoclasts, they were cultured in wells coated with bone particles in the presence of normally available in bone microenvironment, sRANKL

Table 1 PCR primers used to amplify genes under investigation. Gene

Up-stream/down-stream

Base pairs of PCR product (bp)

Annealing temperature of primers Tan. (°C)

Cycles

MMP-1

ATTGGAGCAGCAAGAGGCTGGGA/ TTCCAGGTATTTCTGGACTAAGT CACACCACAACATCACCTATTG/ CAGGGTTTCCCATCAGCATT CGCTACGCCATCCAGGGTCTCAAA/ CGGTCATCATCGGGCAGCACAAAA ACCGGGGTATTGACTCTGAA/GAGGTCAGGCTTGCATCAAT TCGAGTTCCCAGTGAGTGAG/ GACAGGTCCATCAAGTAGTAG GATCCTGGGTGCAGACTTCA/ GCGCTTGGAGATCTTAGAGT ACATCATCCCTGCCTCTACTGG/ AGTGGGTGTCGCTGTTGAAGTC

183

60

35

515

57

35

497

62

35

190 202

57 60

30 30

210

57

30

MMP-9 MT1-MMP Cathepsin K Integrin β3 TRAP GAPDH

3628

P.G. Dedes et al. / Biochimica et Biophysica Acta 1830 (2013) 3625–3634

and M-CSF. The utilized model is schematically presented in Fig. 1A. The concentrations of ZOL used were 3 μM, as lower concentration limit, and 15 μM, which was previously defined by cell proliferation measurements as the IC50 for the MDA-MB-231 cells [17]. Moreover,

it has been reported in the literature [41] that even higher concentrations of ZOL (30 and 50 μM) had no effect on the viability of immature osteoclasts. Interestingly, the high level of TRAP5b determined (298.30 ± 18.97 U/g) in the supernatants of osteoclast cultures,

Fig. 1. Effects of ZOL on activation of osteoclasts, bone osteolysis and activity of MMPs. (A) Immature osteoclasts were incubated on bone particles in the presence of stimuli (sRANKL, M-CSF) and in the presence (3 and 15 μΜ) or absence of ZOL for 7 days of culture for the differentiation of the cells. (B) Effect of ZOL on the activation of immature osteoclasts determined by the level of TRAP5b. (C) Effects of ZOL on bone osteolysis determined by the level of NTx levels. (D) Effects of ZOL on the activity of MMPs as assessed by gelatin zymography. Symbols mark the statistical significant levels as follows: (**) and (***) indicate p b 0.01 and p b 0.001, respectively, as compared to control, (†) indicates p b 0.05 as compared to ZOL (3 μM)-treated cells and (N.D.) stands for not detected.

P.G. Dedes et al. / Biochimica et Biophysica Acta 1830 (2013) 3625–3634

following a 7-days culture period to ensure activation of osteoclasts, was significantly decreased by the addition of increasing concentrations of ZOL (ca 3 and 6-times using 3 and 15 μΜ ZOL, respectively) as compared to control osteoclasts (Fig. 1B). A similar inhibitory pattern was seen when NTx was measured to evaluate the osteolytic activity of activated osteoclasts on bone particles. Therefore, the high NTx level (392.53 ±60.81 nmol BCE/g) found in the absence of ZOL, suggesting a high degree of osteolysis, was significantly decreased (ca 7 to 10-times using 3 and 15 μΜ ZOL, respectively) by ZOL (Fig. 1C). Taking into account the importance of specific MMPs on collagen degradation, we further evaluated the bone-resorbing activity of osteoclasts using zymographic analysis to visualize MMP-2 and -9 levels. As shown in Fig. 1D the conditioned media of activated osteoclasts is rich in pro-MMP-9; lower levels of pro-MMP-2 and activated MMP-9 and MMP-2 were also found. Notably, total MMP-9 activity (determined as total band density of MMP-9) was reduced by 53 and 77% and active MMP-9 levels were significantly decreased by 80% and 94%, using 3 and 15 μΜ of ZOL, respectively. Pro-MMP-2, which was detected only under control conditions, was not identified after treatment with ZOL (Fig. 1D). These results clearly indicate that the activity of MMP-2 and -9 that targets bone collagen and facilitates bone breakdown is significantly inhibited by ZOL even at the low concentration of 3 μM. 3.2. Effect of ZOL on breast cancer cells-induced osteoclast activation It is well established that breast cancer cells represent a principal player of the bone metastasis “vicious cycle”. For this purpose, we conducted co-cultures of immature osteoclasts and MDA-MB-231 breast cancer cells on bone substrate using two experimental approaches (Figs. 2 & 3). In the first experimental set-up, we examined the effects of ZOL on osteoclasts’ maturation/activation in the presence of breast cancer cells and upon specific stimulation (Fig. 2A). Thus, the co-culture of precursor osteoclasts with MDA-MB-231 cells was conducted in the presence of bone microenvironment growth factors (sRANKL, M-CSF). As expected in the absence of ZOL, osteoclasts were activated as assessed by the level of TRAP5b, whereas the incubation with ZOL resulted in the inhibition of precursor osteoclasts’ activation (Fig. 2B). More specifically, the level of TRAP5b (210.65±20.87 U/g) was significantly decreased in the presence of ZOL (ca 2- and 4-times using 3 and 15 μM, respectively). Notably, MDA-MB-231 cells, cultured under the same conditions without osteoclasts, didn't express any TRAP5b, a fact validating that the estimated TRAP5b levels were produced only by the active osteoclasts (Fig. 2B). On the other hand, ZOL also significantly downregulated the level of NTx at both concentrations used by 58 and 87%, respectively. However, it was found that the presence of MDA-MB-231 cells in the same culture condition results in the release of NTx from the bone substrate. Therefore, NTx cannot be used as an osteolysis marker in these co-cultures (Fig. 2B). The previous statement is in accordance with the detection of cathepsin K gene in MDA-MB-231 cells (Fig. 2C). When osteoclasts were undergoing maturation in the co-culture system with the MDA-MB-231 cells, and in the presence of M-CSF and sRANKL (±ZOL) their MMP-2 and -9 levels were also monitored by gelatin zymography. As demonstrated in Fig. 2D, in the presence of M-CSF and sRANKL, ZOL reduced by 27% and 46% of total MMP-9. Percentage reductions of 79% and 88% for active MMP-9, and 53% and 72% for pro-MMP-2 using 3 and 15 μM ZOL, respectively, were also observed. In order to elucidate the direct effect of breast cancer cells in the activation of osteoclasts as well as the effect of ZOL on this activation, a second experimental approach was conducted in the absence of bone microenvironment growth factors (sRANKL, M-CSF) from the co-culture system (Fig. 3A). The incubation of co-cultures with ZOL resulted in the inhibition of immature osteoclasts' activation, translated in reduced TRAP5b levels for both concentrations of ZOL by 26 and 72%, respectively. No statistically significant changes were observed in the

3629

NTx levels under ZOL treatment (Fig. 3B). No statistically significant changes were observed as regards the NTx level under ZOL treatment (Fig. 3B). It is worth noting that the activation of osteoclasts is low, depicted by the low levels of TRAP5b (control 19.06± 2.08 U/g) as compared with the other experimental procedures, indicating that the observed level of NTx is attributed only to the presence of breast cancer cells. When the cells were cultured in the absence of maturation factors, the results concerning the levels of MMP-2 and MMP-9 were similar, since 25% and 48% were obtained for total MMP-9 at 3 and 15 μΜ ZOL, respectively. The active form of MMP-9 was decreased by 25% and 90% and the respective values for pro-MMP-2 reached 41% and 86%, as compared to control (Fig. 3C). These data suggest that MDA-MB-231 cells stimulate osteoclastic MMP-resorbing activity through a RANKL-independent manner. 3.3. Effect of ZOL on the osteolytic activity of active osteoclasts and the expression of functional macromolecules As has been established [33–36], when breast cancer cells migrate to bone microenvironment, they exploit bone-remodeling mechanisms in order to create a “friendly” environment for their survival. Therefore, we evaluated the effect of ZOL on the activity of mature/active osteoclasts. For this purpose, precursor osteoclasts were first cultured in the presence of bone microenvironment growth factors (sRANKL and M-CSF) for 7 days as to ensure their activation. Then, the culture media was renewed and supplemented with or without 3 and 15 μM ZOL for 72 h. It is worth noting that the concentrations of ZOL used in the present study were in all cases lower than those reported (20 μM to 2 mM) to cause low to moderate apoptosis of mature osteoclasts [42]. Following another 72 h culture, the supernatants were collected and assayed for NTx (Fig. 4A). The results showed that the NTx level (453.91 ±40.87 nmol BCE/g) was significantly decreased in the presence of both concentrations (ca 7.5- and 13-times for 3 and 15 μM of ZOL, respectively) (Fig. 4B). It has been well established [28–31] that several macromolecules, such as TRAP5b, integrin β3, cathepsin K, MMP-1, MMP-9 and MT1-MMP play crucial roles in osteoclastic bone resorbing procedure. Therefore, we evaluated the gene encoding for these macromolecules at transcriptional level following incubation of mature osteoclasts with 3 and 15 μΜ of ZOL for 72 h. The incubation of mature osteoclasts with ZOL caused statistically significant downregulation of TRAP5b expression by 23% at the concentration of 15 μΜ, but no noticeable change was observed at 3 μΜ. Integrin β3 levels were also reduced by 21% and 45% at 3 and 15 μΜ of ZOL, respectively as compared to control. Concerning cathepsin K mRNA levels, they dropped by 40% at the higher concentration of 15 μΜ. For MMP-1 and MT1-MMP, the expression levels were significantly reduced by 73% and 87%, and 51% and 57% using 3 and 15 μΜ ZOL, respectively. Finally, MMP-9 expression was reduced by 45% at the highest concentration of 15 μΜ, as compared to control (Fig. 4C). Additionally, in order to validate gelatinases levels of mature osteoclast, gelatin zymography was performed. Although, total MMP-9 activity levels did not seem to be affected significantly by ZOL, active MMP-9 showed a statistically significant decrease of 60% at 15 μΜ, which was also the case for MMP-2 at 15 μΜ (Fig. 4D). 4. Discussion Bone metastasis occurs with a high incidence in a broad range of tumors, including breast, lung, prostate cancer and other solid tumors as well as hematological malignancies, such as multiple myeloma. It is important to note that physiologically bone tissue undergoes continuous metabolic remodeling. This metabolic activity is mediated through osteoblasts, which are responsible for the formation of new bone, and osteoclasts, which mediate bone resorption [43]. Advanced

3630

P.G. Dedes et al. / Biochimica et Biophysica Acta 1830 (2013) 3625–3634

Fig. 2. Evaluation of ZOL effects on the osteoclast precursors maturation in co-culture with breast cancer cells and in the presence of sRANKL and M-CSF. (A) Description of the experimental set up where precursor osteoclasts were co-cultured with MDA-MB-231 breast cancer cells in wells coated with bone particles and in the presence of bone microenvironment growth factors (sRANKL and M-CSF). Immature osteoclasts were treated with ZOL within the 7 days of culture for the differentiation of the cells. Following a 7 days culture period, the conditioned media was collected and assayed accordingly. Effect of ZOL (3 and 15 μΜ) on the activation of osteoclasts was assessed by measuring (B) TRAP5b and NTx levels of precursor osteoclasts, and (D) MMPs activity assessed by gelatin zymography. (C) The effect of ZOL on the gene expression of cathepsin K by breast cancer cells was assayed by RT-PCR. Symbols mark the statistical significant levels as follows: (*), (**) and (***) indicate pb 0.05, p b 0.001 and p b 0.001, respectively, as compared to control, while (†) indicate p b 0.05 as compared to ZOL (3 μM)-treated cells.

P.G. Dedes et al. / Biochimica et Biophysica Acta 1830 (2013) 3625–3634

3631

Fig. 3. Evaluation of ZOL effects on the breast cancer cells-induced osteoclast activation in the absence of sRANKL and M-CSF. (A) Description of the experimental set up where precursor osteoclasts were co-cultured with MDA-MB-231 breast cancer cells in wells coated with bone particles in the absence of bone microenvironment growth factors. Immature osteoclasts were treated with ZOL within the 7 days of culture for the differentiation of the cells. Following a 7 days culture period, the conditioned media and RNA were collected and assayed accordingly. Effect of ZOL (3 and 15 μΜ) on the activation of osteoclasts was assessed by measuring (Β) TRAP5b and NTx levels of precursor osteoclasts, and (C) MMPs activity assessed by gelatin zymography. Symbols mark the statistical significant levels as follows: (*), (**) and (***) indicate pb 0.05, p b 0.001 and p b 0.001, respectively, as compared to control, while (†), (††) and (†††) indicate p b 0.05, p b 0.01 and p b 0.001, respectively, as compared to ZOL (3 μM)-treated cells.

3632

P.G. Dedes et al. / Biochimica et Biophysica Acta 1830 (2013) 3625–3634

Fig. 4. Effects of ZOL on the osteolytic activity of mature osteoclasts and the expression of matrix macromolecules by breast cancer cells. (A) Description of the experimental set-up where precursor osteoclasts were first co-cultured for 7 days to ensure their activation. Then, mature osteoclasts were co-cultured with MDA-MB-231 breast cancer cells in wells coated with bone particles in the presence of bone microenvironment growth factors (sRANKL and M-CSF). After a 72 hour culture in the presence or absence of ZOL the conditioned media and RNA were collected and assayed accordingly. Effect of ZOL (3 and 15 μΜ) on (B) the level of NTx of mature osteoclasts, (C) the gene expression of functional molecules, and (D) MMPs activity assessed by gelatin zymography. Symbols mark the statistical significant levels as follows: (**) and (***) indicate p b 0.01 and p b 0.001, respectively, as compared to control, while (†) and (††) indicate p b 0.05 and p b 0.01, respectively, as compared to ZOL (3 μM)-treated cells.

solid tumors, such as breast cancer, have a predilection to intersperse to bone resulting in the development of osteolytic skeletal-related events. When breast cancer cells approach the bone microenvironment, they induce the recruitment and activation of osteoclasts, a procedure that in return releases molecules that stimulate cancer cells, establishing a perpetual positive feedback [44,45]. This molecular communication between malignant and bone cells is mutually sustained, but not yet fully understood.

ZOL is able to inhibit osteoclastic bone destruction and exhibits antineoplastic properties [45–47]. The current theory on bisphosphonate action on osteoclastogenesis is based on their influence on the RANK– RANKL–OPG system. They tend to induce OPG expression from osteoblasts in a dose dependent manner and reduce RANKL availability by down-regulating gene expression [48] or promoting RANKL protein degradation through activation of a specific gene [49]. OPG binds to RANKL and blocks the interaction with RANK. On the other hand, it

P.G. Dedes et al. / Biochimica et Biophysica Acta 1830 (2013) 3625–3634

has been documented that the inhibitory action of ZOL on osteoblasts, the other counter-player of the “vicious cycle” is mediated through soluble factors, suggesting an indirect inhibition of osteoclast activity [50]. In our experimental system, we introduced the two principal players of this procedure, “immature” or “mature” osteoclasts and cancer cells. However, osteoblasts and other cellular players of the tumor microenvironment that may have key roles in the activation procedure and subsequent osteolysis were not involved in the present cellular mimetic model. Encountering such partners may overcome the limitations and resemble a more clinically relevant model in the future. In the experimental procedure followed, precursor osteoclasts, cultured on the bone particles, were used to study the effect of ZOL on the activation process when lacking the OPG factor. Using a different in vitro model van Beek et al. [51] documented that BPs suppress bone resorption by a direct effect on early osteoclast precursors and that the mechanism of N-BP action in precursor osteoclasts was the same as in mature osteoclasts. In the present study, ZOL inhibits in a dose-dependent manner the activation of osteoclasts as well as bone resorption as verified by utilizing TRAP5b and NTx levels as respective markers. The diminished levels of MMP-9, a marker of precursor osteoclasts [38], observed in conditioned media reinforce the action of ZOL in precursor osteoclasts. It is postulated that through this effect on MMP-9, ZOL mediates inhibition of bone resorption. Considering the above, it proposed that ZOL inhibits the activation of immature osteoclasts in a manner independent of the RANK–RANKL–OPG system. In order to mimic the conditions present in the “vicious cycle”, we conducted co-culture of osteoclasts on bone matrices with the highly invasive breast cancer cells, MDA-MB-231, in the presence of M-CSF and sRANKL which are highly specific for the bone microenvironment. The results highlighted the dose–response inhibition of ZOL of osteoclast activation and destruction of bone substrate, leading to the inhibition of the “vicious cycle” of osteolytic metastasis. Comparing the usefulness of markers used, TRAP5b was found to be a better marker of osteoclast activation, as it wasn't detected in culture media of breast cancer cells in contrast to NTx. These results come in accordance with the detection of cathepsin K gene transcription in this study and by other research groups, where the diminished levels of NTx by ZOL are attributed to the inhibition of cathepsin K expression by ZOL in breast cancer cells. It is well documented throughout the literature that cytokines, like IL-6, TNF-α, IL-1 and IL-8, secreted by cancer cells are able to induce osteoclastogenesis independently of RANKL [52]. Considering that the RANKL-independent pathway leading to activation of osteoclasts in correlation with osteolytic cancer remains to be elucidated, we propose that bone resorption, observed in co-cultures without bone growth factors, may be correlated with the inherent osteolytic activity of breast cancer cells and that in turn ZOL is a key suppressor of this process. The well known effects of ZOL on small GTPases and their related downstream signaling pathways may encountered for possible effects on genes encoded for matrix macromolecules. Indeed, Yoneda et al. [53] have shown that breast cancer bone metastasis could be inhibited by combined treatment with bisphosphonate, and, TIMP-2, a natural inhibitor of MMPs. Since previous studies reported that recombinant TIMP-1 and TIMP-2 inhibited bone resorption in vitro [53,54], TIMP-2 may suppress bone metastasis by inhibiting both tumor invasion and bone resorption [55]. The results of this study indicate that ZOL inhibits dramatically the MMP-2 and -9 activities, which is well correlated with decreased bone breakdown. As far as gene expression levels are concerned, ZOL is responsible, in a concentration dependent manner, for rapid and enhanced downregulation of integrin β3, MMP-1 and MT1-MMP. On the other hand, apart from their collagenolytic activity, MMP-1 and MT1-MMP, play an important role in the migration of osteoclasts as well as in establishing cancer cells to bone destruction sites and, so significant inhibition by ZOL even at low concentrations is of vital importance. Nevertheless, the

3633

low concentration of ZOL (3 μΜ) was not able to reduce CATK expression levels in the incubation time chosen, whereas the higher dose (15 μΜ) was proven to succeed. These data show that cathepsin K is probably not the main enzyme for collagen I catabolism in our model system, since NTx levels drop while cathepsin K expression levels remain unaltered at 3 μM. Therefore, MMPs seem to play a direct proteolytic role in bone resorption since they are mainly produced by host osteoclasts. Conclusively, the application of cellular models in cancer-induced skeletal complications is of great importance for the deeper understanding of this complicated procedure. Pre-clinical studies of pharmaceutical candidates, such as RANK inhibitors or blocking antibodies and overexpressed OPG, must be carried out in reliable systems, which allow direct and easy determination of the implicated molecular mechanisms. In this study, a mimetic model using bone particle-osteoclasts and cancer cells for this purpose was applied, by which we unraveled the crucial implication of MMPs as key players in breast cancer metastatic bone disease. Acknowledgments This study was supported by the research project (PENED 2003, code 03ΕΔ970) which is co-financed by the E.U. — European Social Fund (80%) and the Greek Ministry of Development-GSRT (20%). The authors are grateful to Novartis SA for providing zoledronate active substance. We also thank Prof. D. Gullberg for critical reading and valuable advice. References [1] H.K. Brown, J.H. Healey, Metastatic cancer to the bone, in: V.T. De Vita, S. Hellman, S.A. Rosenberg (Eds.), Principles and Practice of Oncology, Lippincott Williams and Wilkins, 2001, pp. 2713–2729. [2] R.V. Iozzo, N.K. Karamanos, Proteoglycans in health and disease: emerging concepts and future directions, FEBS J. 277 (19) (2010) 3863. [3] R.E. Coleman, A. Lipton, G.D. Roodman, T.A. Guise, B.F. Boyce, A.M. Brufsky, P. Clézardin, P.I. Croucher, J.R. Gralow, P. Hadji, I. Holen, G.R. Mundy, M.R. Smith, L.J. Suva, Cancer Treat. Rev. 36 (8) (2010) 615–620. [4] L.A. Liotta, E.C. Kohn, The microenvironment of the tumour–host interface, Nature 411 (2001) 375–379. [5] I.J. Fidler, Tumor heterogeneity and the biology of cancer invasion and metastasis, Cancer Res. 38 (1978) 2651–2660. [6] B.R. Zetter, The cellular basis of site-specific tumor metastasis, N. Engl. J. Med. 322 (1990) 605–612. [7] Ch. Gialeli, A.D. Theocharis, N.K. Karamanos, Roles of matrix metalloproteinases in cancer progression and their pharmacological targeting, FEBS J. 278 (1) (2011) 16–27. [8] G.R. Mundy, Bisphosphonates and tumor burden, J. Clin. Oncol. 20 (15) (2002) 3191–3192. [9] G.R. Mundy, Metastasis to bone: causes, consequences and therapeutic opportunities, Nature 2 (2002) 584–591. [10] V.T. Labropoulou, A.D. Theocharis, A. Symeonidis, S.S. Skandalis, N.K. Karamanos, H.P. Kalofonos, Pathophysiology and pharmacological targeting of tumor-induced bone disease: current status and emerging therapeutic interventions, Curr. Med. Chem. 18 (2011) 1584–1598. [11] J.J. Body, R. Bartl, P. Burckhardt, P.D. Delmas, I.J. Diel, H. Fleisch, J.A. Kanis, R.A. Kyle, G.R. Mundy, A.H. Paterson, R.D. Rubens, Current use of bisphosphonates in oncology, J. Clin. Oncol. 16 (1998) 3890–3899. [12] S.P. Luckman, D.E. Hughes, F.P. Coxon, R. Graham, G. Russell, M.J. Rogers, Nitrogen-containing bisphosphonates inhibit the mevalonate pathway and prevent post-translational prenylation of GTP-binding proteins, including Ras, J. Bone Miner. Res. 13 (1998) 581–589. [13] F.P. Coxon, M.J. Rogers, The role of prenylated small GTP-binding proteins in the regulation of osteoclast function, Calcif. Tissue Int. 72 (2003) 80–84. [14] G. Morgan, A. Lipton, Antitumor effects and anticancer applications of bisphosphonates, Semin. Oncol. 37 (2010) 30–40. [15] R.E. Coleman, M.C. Winter, D. Cameron, R. Bell, D. Dodwell, M.M. Keane, M. Gil, D. Ritchie, J.L. Passos-Coelho, D. Wheatley, R. Burkinshaw, S.J. Marshall, H. Thorpe, The effects of adding zoledronic acid to neoadjuvant chemotherapy on tumour response: exploratory evidence for direct anti-tumour activity in breast cancer, Br. J. Cancer 102 (2010) 1099–1105. [16] K. Zarogoulidis, E. Boutsikou, P. Zarogoulidis, E. Eleftheriadou, T. Kontakiotis, H. Lithoxopoulou, G. Tzanakakis, I. Kanakis, N.K. Karamanos, The impact of zoledronic acid therapy in survival of lung cancer patients with bone metastasis, Int. J. Cancer 125 (2009) 1705–1709. [17] P.G. Dedes, Ch. Gialeli, A.I. Tsonis, I. Kanakis, A.D. Theocharis, D. Kletsas, G.N. Tzanakakis, N.K. Karamanos, Expression of matrix macromolecules and functional

3634

[18]

[19]

[20]

[21]

[22]

[23] [24] [25]

[26] [27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

P.G. Dedes et al. / Biochimica et Biophysica Acta 1830 (2013) 3625–3634 properties of breast cancer cells are modulated by the bisphosphonate zoledronic acid, Biochim. Biophys. Acta 1820 (2012) 1926–1939. S. Boissier, S. Magnetto, L. Frappart, B. Cuzin, F.H. Ebetino, P.D. Delmas, P. Clezardin, Bisphosphonates inhibit prostate and breast carcinoma cell adhesion to unmineralized and mineralized bone extracellular matrices, Cancer Res. 57 (1997) 3890–3894. V. Viereck, G. Emons, V. Lauck, K.H. Frosch, S. Blaschke, C. Grundker, L.C. Hofbauer, Bisphosphonates pamidronate and zoledronic acid stimulate osteoprotegerin production by primary human osteoblasts, Biochem. Biophys. Res. Commun. 291 (2002) 680–686. T. Suda, N. Takahashi, N. Udagawa, E. Jimi, M.T. Gillespie, T.J. Martin, Modulation of osteoclast differentiation and function by the new members of the tumor necrosis factor receptor and ligand families, Endocr. Rev. 20 (3) (1999) 345–357. H. Yasuda, N. Shima, N. Nakagawa, K. Yamaguchi, M. Kinosaki, S. Mochizuki, A. Tomoyasu, K. Yano, M. Goto, A. Murakami, E. Tsuda, T. Morinaga, K. Higashio, N. Udagawa, N. Takahashi, T. Suda, Osteoclast differentiation factor is a ligand for osteoprotegerin/osteoclastogenesis-inhibitory factor and is identical to TRANCE/ RANKL, Proc. Natl. Acad. Sci. 95 (1998) 3597–3602. T. Onishi, N. Hayashi, R.L. Theriault, G.N. Hortobagyi, N.T. Ueno, Future directions of bone-targeted therapy for metastatic breast cancer, Nat. Rev. Clin. Oncol. 7 (11) (2010) 641–651. R. Hanada, T. Hanada, J.M. Penninger, Physiology and pathophysiology of the RANKL/RANK system, Biol. Chem. 391 (12) (2010) 1365–1370. C. Minkin, Bone acid phosphatase: tartrate-resistant acid phosphatase as a marker of osteoclast function, Calcif. Tissue Int. 34 (1982) 285–290. J.M. Halleen, H. Ylipahkala, S.L. Alatalo, A.J. Janckila, J.E. Heikkinen, H. Suominen, S. Cheng, H.K. Väänänen, Serum tartrate-resistant acid phosphatase 5b, but not 5a, correlates with other markers of bone turnover and bone mineral density, Calcif. Tissue Int. 71 (1) (2002) 20–25. J.G. Schneider, S.R. Amend, K.N. Weilbaecher, Integrins and bone metastasis: integrating tumor cell and stromal cell interactions, Bone 48 (1) (2011) 9, (54–65). R. Faccio, D.V. Novack, A. Zallone, F.P. Ross, S.L. Teitelbaum, Dynamic changes in the osteoclast cytoskeleton in response to growth factors and cell attachment are controlled by beta3 integrin, J. Cell Biol. 162 (3) (2003) 499–509. P. Garnero, O. Borel, I. Byrjalsen, M. Ferreras, F.H. Drake, M.S. McQueney, N.T. Foged, P.D. Delmas, J.M. Delaissé, The collagenolytic activity of cathepsin K is unique among mammalian proteinases, J. Biol. Chem. 273 (48) (1998) 32347–32352. H. Sato, T. Takino, Y. Okada, J. Cao, A. Shinagawa, E. Yamamoto, M. Seiki, A matrix metalloproteinase expressed on the surface of invasive tumour cells, Nature 370 (6484) (1994) 61–65. M.T. Engsig, Q.J. Chen, T.H. Vu, A.C. Pedersen, B. Therkidsen, L.R. Lund, K. Henriksen, T. Lenhard, N.T. Foged, Z. Werb, J.M. Delaissé, Matrix metalloproteinase 9 and vascular endothelial growth factor are essential for osteoclast recruitment into developing long bones, J. Cell Biol. 151 (4) (2000) 879–889. P. Garnero, M. Ferreras, M.A. Karsdal, R. Nicamhlaoibh, J. Risteli, O. Borel, P. Qvist, P.D. Delmas, N.T. Foged, J.M. Delaissé, The type I collagen fragments ICTP and CTX reveal distinct enzymatic pathways of bone collagen degradation, J. Bone Miner. Res. 18 (5) (2003) 859–867. S. Thiolloy, J. Halpern, G.E. Holt, H.S. Schwartz, G.R. Mundy, L.M. Matrisian, C.C. Lynch, Osteoclast-derived matrix metalloproteinase-7, but not matrix metalloproteinase-9, contributes to tumor-induced osteolysis, Cancer Res. 2 (69) (2009) 6747–6755. R.J. Thomas, T.A. Guise, J.J. Yin, J. Elliott, N.J. Horwood, T.J. Martin, M.T. Gillespie, Breast cancer cells interact with osteoblasts to support osteoclast formation, Endocrinology 140 (1999) 4451–4458. A.T. Mancino, V.S. Klimberg, M. Yamamoto, S.C. Manolagas, E. Abe, Breast cancer increases osteoclastogenesis by secreting M-CSF and upregulating RANKL in stromal cells, J. Surg. Res. 100 (1) (2001) 18–24. H. Morgan, A. Tumber, P.A. Hill, Breast cancer cells induce osteoclast formation by stimulating host IL-11 production and downregulating granulocyte/macrophage colony-stimulating factor, Int. J. Cancer 109 (2004) 653–660.

[36] Y. Guo, K. Tiedemann, J.A. Khalil, C. Russo, P.M. Siegel, S.V. Komarova, Osteoclast precursors acquire sensitivity to breast cancer derived factors early in differentiation, Bone 43 (2) (2008) 386–393. [37] Y.S. Lau, L. Danks, S.G. Sun, S. Fox, A. Sabokbar, A. Harris, N.A. Athanasou, RANKL-dependent and RANKL-independent mechanisms of macrophage–osteoclast differentiation in breast cancer, Breast Cancer Res. Treat. 105 (1) (2007) 7–16. [38] G.D. Roodman, Cell biology of the osteoclast, Exp. Hematol. 27 (8) (1999) 1229–1241. [39] M.M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding, Anal. Biochem. 72 (1976) 248–254. [40] K. Kusano, C. Miyaura, M. Inada, T. Tamura, A. Ito, H. Nagase, K. Kamoi, T. Suda, Regulation of matrix metalloproteinases (MMP-2, -3, -9, and -13) by interleukin-1 and interleukin-6 in mouse calvaria: association of MMP induction with bone resorption, Endocrinology 139 (3) (1998) 1338–1345. [41] K. Kimachi, H. Kajiya, S. Nakayama, T. Ikebe, K. Okabe, Zoledronic acid inhibits RANK expression and migration of osteoclast precursors during osteoclastogenesis, Naunyn Schmiedebergs Arch. Pharmacol. 383 (3) (2011) 297–308. [42] M. Kellinsalmi, H. Mönkkönen, J. Mönkkönen, H.V. Leskelä, V. Parikka, M. Hämäläinen, P. Lehenkari, In vitro comparison of clodronate, pamidronate and zoledronic acid effects on rat osteoclasts and human stem cell-derived osteoblasts, Basic Clin. Pharmacol. Toxicol. 97 (6) (2005) 382–391. [43] R.H. Christenson, Biochemical markers of bone metabolism: an overview, Clin. Biochem. 30 (8) (1997) 573–593. [44] T.A. Guise, J.J. Yin, S.D. Taylor, Y. Kumagai, M. Dallas, B.F. Boyce, T. Yoneda, G.R. Mundy, Evidence for a causal role of parathyroid hormone-related protein in the pathogenesis of human breast cancer-mediated osteolysis, J. Clin. Invest. 98 (1996) 1544–1549. [45] S.M. Käkönen, G.R. Mundy, Mechanisms of osteolytic bone metastases in breast carcinoma, Cancer 97 (Suppl. 3) (2003) 834–839. [46] J.R. Green, Antitumor effects of bisphosphonates, Cancer 97 (2003) 840–847. [47] C.L. Edward, Zoledronic acid: a new parenteral bisphosphonate, Clin. Ther. 25 (2003) 2669–2708. [48] B. Pan, A.N. Farrugia, L.B. To, D.M. Findlay, J. Green, K. Lynch, A.C. Zannettino, The nitrogen-containing bisphosphonate, zoledronic acid, influences RANKL expression in human osteoblast-like cells by activating TNF-alpha converting enzyme (TACE), J. Bone Miner. Res. 19 (1) (2004) 147–154. [49] P.S. Mackie, J.L. Fisher, H. Zhou, P.F. Choong, Bisphosphonates regulate cell growth and gene expression in the UMR 106-01 clonal rat osteosarcoma cell line, Br. J. Cancer 84 (7) (2001) 951–958. [50] M. Nishikawa, T. Akatsu, Y. Katayama, Y. Yasutomo, S. Kado, N. Kugal, M. Yamamoto, N. Nagata, Bisphosphonates act on osteoblastic cells and inhibit osteoclast formation in mouse marrow cultures, Bone 18 (1) (1996) 9–14. [51] E.R. Van Beek, C.W. Löwik, S.E. Papapoulos, Bisphosphonates suppress bone resorption by a direct effect on early osteoclast precursors without affecting the osteoclastogenic capacity of osteogenic cells: the role of protein geranylgeranylation in the action of nitrogen-containing bisphosphonates on osteoclast precursors, Bone 30 (1) (2002) 64–70. [52] M.S. Bendre, A.G. Margulies, B. Walser, N.S. Akel, S. Bhattacharrya, R.A. Skinner, F. Swain, V. Ramani, K.S. Mohammad, L.L. Wessner, A. Martinez, T.A. Guise, J.M. Chirgwin, D. Gaddy, L.J. Suva, Tumor-derived interleukin-8 stimulates osteolysis independent of the receptor activator of nuclear factor-kappaB ligand pathway, Cancer Res. 65 (23) (2005) 11001–11009. [53] T. Yoneda, A. Sasaki, C. Dunstan, P.J. Williams, F. Bauss, Y.A. De Clerck, G.R. Mundy, Inhibition of osteolytic bone metastasis of breast cancer by combined treatment with the bisphosphonate ibandronate and tissue inhibitor of the matrix metalloproteinase-2, J. Clin. Invest. 99 (10) (1997) 2509–2517. [54] P.A. Hill, J.J. Reynolds, M.C. Meikle, Inhibition of stimulated bone resorption in vitro by TIMP-1 and TIMP-2, Biochim. Biophys. Acta 1177 (1) (1993) 71–74. [55] T. Ohshiba, C. Miyaura, M. Inada, A. Ito, Role of RANKL-induced osteoclast formation and MMP-dependent matrix degradation in bone destruction by breast cancer metastasis, Br. J. Cancer 88 (8) (2003) 1318–1326.