- Email: [email protected]
Osteoclasts but not osteoblasts are affected by a calcified surface treated with zoledronic acid in vitro
BBRC Biochemical and Biophysical Research Communications 338 (2005) 710–716 www.elsevier.com/locate/ybbrc
Osteoclasts but not osteoblasts are affected by a calcified surface treated with zoledronic acid in vitro Aaron Schindeler a,b,*, David G. Little a,b a
Department of Orthopaedic Research and Biotechnology, The ChildrenÕs Hospital at Westmead, Sydney, Australia b Discipline of Paediatrics and Child Health, Faculty of Medicine, University of Sydney, Sydney, Australia Received 26 September 2005 Available online 11 October 2005
Abstract Bisphosphonates are potent inhibitors of osteoclast-mediated bone resorption. Recent interest has centered on the effects of bisphosphonates on osteoblasts. Chronic dosing of osteoblasts with solubilized bisphosphonates has been reported to enhance osteogenesis and mineralization in vitro. However, this methodology poorly reflects the in vivo situation, where free bisphosphonate becomes rapidly bound to mineralized bone surfaces. To establish a more clinically relevant cell culture model, we cultured bone cells on calcium phosphate coated quartz discs pre-treated with the potent nitrogen-containing bisphosphonate, zoledronic acid (ZA). Binding studies utilizing [14C]-labeled ZA confirmed that the bisphosphonate bound in a concentration-dependent manner over the 1–50 lM dose range. When grown on ZA-treated discs, the viability of bone-marrow derived osteoclasts was greatly reduced, while the viability and mineralization of the osteoblastic MC3T3-E1 cell line were largely unaffected. This suggests that only bone resorbing cells are affected by bound bisphosphonate. However, this system does not account for transient exposure to unbound bisphosphonate in the hours following a clinical dosing. To model this event, we transiently treated osteoblasts with ZA in the absence of a calcified surface. Osteoblasts proved highly resistant to all transitory treatment regimes, even when utilizing ZA concentrations that prevented mineralization and/or induced cell death when dosed chronically. This study represents a pharmacologically more relevant approach to modeling bisphosphonate treatment on cultured bone cells and implies that bisphosphonate therapies may not directly affect osteoblasts at bone surfaces. 2005 Elsevier Inc. All rights reserved. Keywords: Bisphosphonates; Osteoblasts; Osteoclasts; Osteoporosis: Therapy; Bone resorption
Nitrogen-containing bisphosphonates (N-BPs) are a class of drugs used in the treatment of osteoporosis and diseases of high bone turnover [1]. After administration, a considerable proportion of an N-BP binds with high affinity to hydroxyapatite at the bone surface, while the remainder is rapidly excreted by the renal system [2–6]. Once released from hydroxyapatite and internalized by cells, N-BPs act as potent inhibitors of farnesyl diphosphate synthase (FPPS), an enzyme in the mevalonate pathway [7]. The presence of bound N-BP at the bone surface can effectively reduce osteoclast-mediated bone turnover. Inhibition of FPPS prevents the prenylation of small GTP-
*
Corresponding author. Fax: +61 2 98453078. E-mail address: [email protected] (A. Schindeler).
0006-291X/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2005.09.198
ases such as Ras, Rac, and Cdc42, thus resulting in decreased osteoclastogenesis, osteoclast survival, and/or osteoclast function [1,8]. In addition to affecting osteoclasts, it has been speculated that bound N-BPs may also influence the differentiation, function or survival of osteoblasts. Numerous studies have investigated the effects of various N-BPs, including zoledronic acid (ZA), alendronate, pamidronate, and risedronate, on cultured cells [9–17]. The majority of these studies have involved the chronic exposure of osteoblasts and osteoblast-like cell lines to solubilized N-BPs and then measuring cell viability, alkaline phosphatase (ALP) activity, mineralization, and/or osteogenic gene expression. Although the individual N-BPs, doses, and experimental outcomes vary considerably, it was frequently concluded that specific doses of N-BPs could have mildly positive
A. Schindeler, D.G. Little / Biochemical and Biophysical Research Communications 338 (2005) 710–716
effects on proliferation or differentiation at certain time points. These data conflict with animal and clinical studies that suggest long-term N-BP treatment can lead to reduced bone formation [18–23]. Recently, Coxon et al. [24] have investigated the release, uptake, and internalization of a fluorescently labeled N-BP by cultured cells. Resorbing osteoclasts were able to take up the labeled N-BP bound to a dentine substrate, while macrophages and non-resorbing osteoclasts could not. Thus, the acidic environment or enzymes produced specifically by osteoclasts may be required for the release of NBPs from a bone surface. Although osteoblasts have been reported to have some capacity to degrade organic components of the bone matrix [25], we speculated they lack the capacity to release and internalize bound N-BPs. To address this issue, we have employed a cell culture system that incorporates a calcified surface and a radiolabeled N-BP. Data obtained using this system suggest that osteoblasts are unable to release and take up pharmacologically significant amounts of bound bisphosphonate. Nevertheless, osteoblasts may be exposed to serum levels of unbound N-BP immediately following a clinical dosing. Certainly non-resorbing macrophages were able to take up solubilized fluorescent N-BP when cultured on plastic [24], consistent with other cell culture studies [8– 17]. The length of transient exposure may be dependent on the dosing regime. In the case of intravenous treatment, the method used for ZA dosing, bisphosphonate is rapidly cleared to the bones or kidneys within hours of injection [2]. In contrast, the bioavailability of oral bisphosphonates such as alendronate is less (leading to higher amounts of drug being administered), and absorption is prolonged compared with an intravenous dosing [3,4]. Thus, it is possible that these treatment regimes may have different effects on osteoblasts. To model the limited bisphosphonate exposure that follows a clinical dosing, MC3T3-E1 cells and primary osteoblasts were dosed with ZA for a 2 h pulse (analogous to intravenous dosing [2]), a 1 day pulse (analogous to oral weekly dosing [3,4]) or continuously (consistent with prior cell culture studies [9–17]). A range of ZA concentrations were used to reflect both the expected pharmacological exposure and the elevated doses used in vitro by other investigators. These studies attempt to generate more clinically relevant approaches to in vitro experimental design and may help explain some of the disparities between published in vitro and in vivo data. Materials and methods Calcium phosphate discs. The Osteologic disc system was used to model bisphosphonate binding and cell activity at the bone surface (BD Biosciences, NJ, USA). This product is a practical alternative to bone or dentine slices, exhibits greater consistency across the surface, is simple to visualize by phase contrast microscopy, and is feasible for both osteoblast and osteoclast culture [26,27]. The discs, supplied in 24-well plate format, were 12.7 mm in diameter and consisted of quartz layered with 0.65 lm of calcium phosphate. For bisphosphonate pre-treatment, discs were incu-
711
bated for 2 h with 1–50 lM ZA dissolved in 1 ml of culture medium. Discs were washed with 3 · 1 ml of culture media prior to seeding with cells. New bone formation was measured on calcium phosphate discs by treating cells for 6 days with 4 lg/ml calcein and evaluating fluorescence on a DC500 microscope (Leica Microsystems, NSW, Australia). This system for measuring mineralization on discs in vitro is based on previously published methods [26]. [14C]ZA binding assays. [14C]-radiolabeled zoledronic acid ([14C]ZA) was generously supplied by Novartis Pharma AG, Basel, Switzerland. The central carbon atom between the two phosphonic acid groups (carbon 1) of ZA was labeled with [14C]; since N-BPs are metabolically stable we expect no re-arrangement of the [14C] within the molecule nor its release under physiological conditions. Bound [14C]ZA was rapidly eluted from discs by decalcifying with 0.5 ml of 5 M HCl. Unbound and bound [14C]ZA were quantified using OptiPhase SuperMix (Perkin-Elmer, MA, USA) in a Tri-Carb scintillation counter (Packard, MN, USA). Values in counts per minute (cpm) were calibrated using a control vial containing a known quantity of [14C]ZA. MC3T3-E1 cell culture. MC3T3-E1 cells were grown in a-MEM (Gibco Laboratories, CA, USA) supplemented with 10% FBS, 2 mM Lglutamine, and antibiotics (100 U/ml penicillin and 0.1 mg/ml streptomycin). Cultures were incubated at 37 C in the presence of 5% CO2. Osteogenic differentiation was initiated on confluent MC3T3-E1 cells using medium further supplemented with 50 mg/L ascorbic acid and 10 mM b-glycerophosphate. Treatment with Zoledronic Acid (ZA, kindly supplied by Novartis Pharmaceuticals, NSW, Australia) was conducted in normal medium supplemented with the drug. Medium changes were made every 3–4 days, including the addition of fresh ZA for continuous dosing regimes. Preparatin and culture of bone marrow-derived murine osteoclasts. Femoral bone marrow from 3- to 4-month- old male mice was extracted using a 21 gauge needle and the mononuclear lymphocytes were subsequently purified using Histopaque-1077 (Sigma) as per the manufacturerÕs instructions. Mononuclear cells were plated at 2.5 · 105 cells/well in 24well plates, cultured in media consisting of a-MEM containing 10% FBS, 2 mM L-glutamine, and antibiotics, and supplemented with 30 ng/ml RANK-L (R&D Systems), and 5 ng/ml M-CSF (R&D Systems). Media changes were performed every 3 days. Preparation and culture of primary calvarial murine osteoblasts. Neonatal mice (days 2–3) were killed by decapitation. Calvaria were then dissected, thoroughly minced, and the fragments were extensively washed with sterile PBS. Bone pieces were digested with 0.05% Trypsin (CSL, Victoria, Australia), 0.3 mM EDTA, and 2.5 mg/ml Collagenase A (Roche Molecular Biochemicals, NSW, Australia) in PBS. Cells were passed through 19G and 21G needles and then filtered through a sterile mesh. Cells were counted using a hemocytometer and plated at a concentration of 5 · 104 cells/ml in 6, 12 or 24-well plates. Primary calvarial murine osteoblasts were grown in a-MEM supplemented with 10% FBS, 2 mM L-glutamine, antibiotics (100 U/ml penicillin and 0.1 mg/ml streptomycin), 50 mg/L ascorbic acid, 10 mM b-glycerophosphate, 20 mM Na2CO3, 34 mM Hepes, and 34 mg/L gentamycin. Drug interventions commenced after osteoblasts had reached confluence (days 5–7 after preparation). Cell viability and alkaline phosphatase assays. Osteoblast viability was determined at day 3 using the CellTiter 96 AQueous One Solution Cell Proliferation Assay system (Promega, NSW, Australia), as per the manufacturerÕs instructions. Alkaline phosphatase (ALP) activity assays were performed at day 4. Osteoblasts were harvested in 50 ll of a lysis buffer containing 20 mM Tris–HCl (pH 8.2), 2 mM MgCl2, and 0.05% Triton X100. Subsequently, cell lysate (normalized for total protein) was added to 300 ll ALP assay buffer containing 0.1 M NaHCO3, 1 mM MgCl2, and 10 mM p-nitrophenyl phosphate, and incubated for 15 min at 37 C. Absorbance at 405 nm was read on a microplate reader (Labsystems, MA, USA). Tartrate-resistant acid phosphatase staining. Osteoclasts were identified using a stain for tartrate-resistant acid phosphatase (TRAP). Cells were then incubated for 5 min at 37 C with freshly made TRAP staining solution utilized in tissue histology [28].
712
A. Schindeler, D.G. Little / Biochemical and Biophysical Research Communications 338 (2005) 710–716
Alizarin Red S staining. In brief, cells were rinsed twice with PBS, fixed with 4% paraformaldehyde in PBS, washed twice with deionized water, and then stained with freshly made 40 mM Alizarin Red S (pH 4.2) for 15 min. The Alizarin Red S solution was then removed by aspiration and cells were rinsed five times with deionized water. Mineralization was quantified using NIH Image software. Statistical methods. Statistical significance was calculated using unpaired two-tailed studentÕs t test to compare treated samples with untreated controls. Error bars on graphs represent ± the standard error. Experiments were performed in duplicate or triplicate wells, and independent trials were repeated a minimum of two times, unless otherwise noted.
affected by ZA released from the discs acting in a paracrine manner. Osteoblast viability is unaffected on a ZA-treated calcified surface
Binding experiments utilizing [14C]ZA revealed that the bisphosphonate bound and concentrated at the disc surface (Table 1). A range of ZA concentrations were used (1– 50 lM) with the lowest dose approximating the serum concentration of ZA immediately following a clinical dosing [2] and the highest dose reported to be lethal when chronically dosed in vitro [9–14]. After 2 h of incubation, 5–10% of ZA in the liquid phase bound to the 0.65 lm calcium phosphate substrate (i.e., 0.2–1.83 nmol ZA). Based on a matrix volume of 8.23 · 107 lm3 for a 12.7 mm disc, this translated to local matrix ZA concentrations of 1.0–9.0 mM, which was up to 180-fold higher than the highest soluble dose.
To test whether osteoblasts behaved in the same manner as osteoclasts, mouse MC3T3-E1 osteoprogenitor cells were plated on calcium phosphate discs and compared to culture on plastic. Cell proliferation assays at day 3 showed a 1.25-fold increase in the number of osteoblasts on the discs as compared to plastic (Fig. 1B). When cultured on calcified discs treated with 1–50 lM ZA, no significant difference in cell viability was seen (Fig. 1B), despite ZA being concentrated at the disc surface (Table 1). In a separate experiment, mineralized matrix production on calcium phosphate discs was measured by calcein incorporation, with ZA-coating having no significant effect on mineralization (data not shown). In contrast, osteoblasts treated with 1–50 lM ZA in the media and grown on tissue culture plastic showed decreased viability in a dose-dependent manner. This dose–response curve is consistent with several studies that have chronically treated cells with ZA in vitro [10,11,15,16]. Significant cell death was seen at the maximal drug dose (50 lM ZA), which was not evident on discs pre-treated or post-treated (after cell plating) for 2 h with 50 lM ZA (Fig. 1C).
Osteoclast viability is reduced on a ZA-treated calcified surface
Osteoblasts inefficiently take up ZA bound to calcium phosphate discs
To confirm the validity of the in vitro model, bone marrow derived lymphocytes were cultured in the presence of the pro-osteoclastic agents RANK-L and M-CSF on untreated discs and discs pre-treated with 10 lM ZA. Following 12 days of culture, cells were stained for TRAP expression and revealed high numbers of multinucleate TRAP+ cells on untreated discs. In comparison TRAP+ cells were vastly reduced on ZA-pretreated discs (Fig. 1A). These data suggest that the bisphosphonate bound to CaPO4 discs can act specifically on resorbing osteoclasts in vitro. In addition, normal numbers of viable multinucleate osteoclasts were found on plastic surrounding both ZA-treated and untreated discs (data not shown). This indicates that cells on ZA-treated discs were not being
To directly measure bisphosphonate absorption by osteoblasts, MC3T3-E1 cells were treated with [14C]ZA. Following detergent lysis, osteoblasts treated with unbound 10 lM [14C]ZA (i.e., grown on plastic) were found to take up 80 pmol of drug over 3 days. Despite a relatively high local concentration in the CaPO4 matrix (Table 1), osteoblasts grown on [14C]ZA coated discs absorbed a dose of only 50 pmol over the same interval. These results further suggest that bound ZA cannot be efficiently released and internalized by osteoblasts.
Results ZA binds to calcium phosphate discs
Table 1 [14C]ZA binding to calcium phosphate discsa ZA treatment (nmol)
Bound ZA (nmol)
Matrix ZA conc. (mM)
0 1 10 50
0 0.20 0.74 1.83
0.0 1.0 3.6 9.0
a ZA treatments were performed in 1 ml of osteoblast growth media, so that 1 nmol ZA gave a final concentration of 1 lM. The concentration of bound ZA within the 0.65 lm CaPO4 layer (Matrix ZA conc.) reached far higher levels than in the soluble ZA treatments.
MC3T3-E1 cells survive transitory dosing with ZA Although osteoblasts may not efficiently take up bound ZA, we postulated that these cells would still be exposed to transitory doses of unbound ZA during a clinical dosing. To model this event in culture, we treated MC3T3-E1 cells grown on plastic with transient doses of 1–50 lM ZA for 2 h and 1 day. The 2 h dose was used to reflect the pharmacokinetics of an intravenous bisphosphonate [3] while the 1 day dose was chosen to reflect the prolonged presence of an oral bisphosphonate [4]. Both oral and intravenous bisphosphonate are typically given to patients on a weekly or monthly basis, so cells were transiently dosed a single time. Cell viability was measured at day 3 of treatment and compared with that of untreated cells. As previously
A. Schindeler, D.G. Little / Biochemical and Biophysical Research Communications 338 (2005) 710–716
713
Fig. 1. Osteoclasts but not osteoblasts are affected on calcium phosphate discs pre-treated with ZA. (A) TRAP+ multinucleated cells (MNCs) generated from murine bone marrow cells were counted on calcium phosphate discs in five random fields. Significantly more TRAP+ MNCs were present at day 12 on untreated discs compared with discs pre-treated with 10 lM ZA *P < 0.05. Data reflect the results of two independent experiments. (B) Cultured MC3T3-E1 cells were exposed to either soluble ZA (on plastic) or to bound ZA (on discs pre-treated with ZA solution). All treatment groups were statistically compared to osteoblasts grown on plastic. Osteoblasts dosed with soluble ZA (on plastic) were sensitive to treatment with 10 lM ZA (*P < 0.07) and 50 lM ZA (**P < 0.05); however, no significant reduction in cell viability was seen on discs pre-coated with similar doses of ZA. A significant 1.3-fold increase in viable osteoblasts was observed on untreated calcium phosphate discs and discs pre-treated with 1 lM ZA (***P < 0.05). These data summarize the results from three independent experiments. (C) Phase contrast images of differentiating osteoblasts on plastic and calcium phosphate discs after 4 days. When exposed to 50 lM soluble ZA, significant cell death was seen in osteoblast cultures cultured on plastic. In contrast, osteoblasts grown on calcium phosphate discs either pre-treated or post-treated for 2 h with 50 lM ZA showed no phenotypic effects.
reported by others, cell viability was reduced in chronic 10– 50 lM ZA doses (Figs. 1C and 2A). At 50 lM ZA, significant cell death was observed (Fig. 1C), which previous reports have demonstrated to be apoptosis [13,16]. In contrast, more limited treatments (2 h or 1 day) did not significantly affect cell viability (Fig. 2A), even at the highest ZA dose used. Transitory ZA treatment slightly enhances osteogenic differentiation and mineralization Next, alkaline phosphatase (ALP) activity was measured as a marker of osteogenic differentiation. Assays were normalized for total protein, thus accounting for differences in cell viability. Chronic ZA dosing did not have any significant benefit on osteoblast differentiation, even causing a
slight decrease in differentiation at higher doses (Fig. 2B). Transient dosing with ZA achieved a slight increase in ALP activity, with the most significant being a 2-fold increase in the 1 lM ZA for 1 day treatment sample (Fig. 2B). A more functionally relevant endpoint is to measure osteoblast mineralization, which was achieved by staining extracellular matrix calcium deposits with Alizarin Red S stain. MC3T3-E1 cells treated as previously described were fixed and stained at day 8. All wells displayed an even staining pattern, reflective of the homogeneous nature of the cell line. Staining intensity was measured using imaging software and the findings paralleled those obtained from earlier ALP assays (Fig. 3A). These small increases in mineralization were consistently maintained when cultures were taken out to longer time points of 14 and 21 days (data not shown).
714
A. Schindeler, D.G. Little / Biochemical and Biophysical Research Communications 338 (2005) 710–716
Fig. 2. Transitory ZA dosing does not negatively affect viability and ALP activity in MC3T3-E1 cells. (A) Chronic dosing with ZA caused a decrease in cell viability compared with untreated cells that corresponded with ZA dose (*P < 0.05). No negative effects on cell viability were noted with transient doses of ZA, even at the highest concentrations of ZA (50 lM). (B) ALP activity at day 4 was negatively affected by chronic ZA and positively affected by brief exposure to low doses of ZA (*P < 0.05 compared to untreated cells). These data summarize the results from three independent experiments.
To confirm that these effects were not specific to the MC3T3-E1 cell line, primary osteoblasts were isolated and cultured from neonatal mouse calvaria and treated with identical ZA dosing regimes. Mineralization was measured at day 14 after reaching confluence, by which time cells had formed numerous mineralized nodules (Fig. 3B). Quantification using image analysis software revealed that mineralization nodule formation by primary osteoblasts mirrored the results from the MC3T3-E1 cells (data not shown). However, the overall effect of ZA on primary osteoblast mineralization was reduced compared to the MC3T3-E1 mineralization (e.g., only a 1.4-fold increase was seen in the 1 lM ZA for 1 day sample). Discussion Clinical success using bisphosphonates has led to speculation that these drugs may generate a pro-anabolic re-
Fig. 3. Transitory ZA does not negatively affect the mineralization of MC3T3-E1 cells or primary calvarial osteoblasts. (A) Monolayers of differentiated MC3T3-E1 cells were stained for mineralization using Alizarin Red S and quantified on digital images. Decreased mineralization was observed in continuously treated samples, while increased mineralization was seen in 1 lM/1 day dose, 10 lM/1 day dose, and 1 lM/2 h dose (*P < 0.05 compared with untreated MC3T3-E1 cells). (B) Primary osteoblasts were similarly stained at day 14. Mineralized nodules (red patches) were apparent in all samples, not including cells treated continuously with 50 lM ZA. Image analysis indicated results were compatible those observed from the MC3T3-E1 cell line.
sponse in addition to their anti-catabolic activity. While evidence from in vivo studies generally suggests that bone formation is reduced by bisphosphonate action [18–23], experiments with cultured osteoblasts have often implied
A. Schindeler, D.G. Little / Biochemical and Biophysical Research Communications 338 (2005) 710–716
a potential anabolic mechanism [9–17]. In vitro tests have revealed increases in proliferation, ALP activity, mineralization, and osteogenic markers with specific bisphosphonate treatment regimes [9–14,17]. We hypothesized that this apparent conflict in the literature may be attributed to the prolonged soluble delivery of bisphosphonate to osteoblasts in vitro, which differs from circumstances in vivo. To test this premise, we have compared standard in vitro techniques to what we propose to be more pharmacologically relevant in vitro techniques, including the provision of a calcified surface and transitory soluble dosing. We have demonstrated that the effects of exposing osteoblasts to bisphosphonate bound to a calcified surface can differ from treatment with soluble bisphosphonate. Experiments using [14C]ZA confirmed that ZA binds to calcified discs, which concentrates at the disc surface. While osteoclasts were negatively affected when cultured on a ZA-coated calcified surface, osteoblasts survived, proliferated, and produced mineralized matrix under the same conditions. This has significant implications for the future of cell culture modeling of bisphosphonate treatment. Our results are also consistent with observations made by Coxon et al. [24], which indicate that fluorescently labeled alendronate bound to dentine could be taken up by resorbing osteoclasts, but not macrophages or non-resorbing osteoclasts. To further extend our in vitro model, we simulated the transient exposure of cells to soluble bisphosphonate prior to its binding to mineralized surfaces. This is analogous to the situation that immediately follows a clinical N-BP treatment [2–4]. Although cell culture media may not precisely represent the cation and plasma protein mix normally seen in human serum, the 1 lM ZA dose corresponds to the plasma concentration after an intravenous dose [2]. Osteoblasts were grown on plastic and treated with soluble ZA. Transient dosing (for 2 h or 1 day) was compared with chronic dosing, which is already well reported in the literature [9–17]. In our transitory treatment model, we demonstrated that ZA can have a subtle (never greater than 2-fold) but positive effect on ALP activity and mineralization. It is not clear whether this small increase could be translated into a clinically meaningful anabolic effect. Certainly ALP activity is increased far more substantially when osteoblasts are treated in vitro with established proanabolic agents. For example, treatment with 100 ng/ml BMP-2 was found to increase ALP activity 13-fold in MC3T3-E1 cells [29]. Thus our results, like other in vitro studies [9–14,17], suggest the potential for small but positive anabolic benefits on bone formation via direct effects on osteoblasts. These results diverge from animal studies showing that bisphosphonates have negligible or negative effects on mineral apposition rate or bone formation rate [18–21]. Bone formation has also been reported to be reduced in patients undergoing long-term bisphosphonate therapies [22,23]. However, there is no evidence that these examples of re-
715
duced bone formation in vivo are due to the direct effects of bisphosphonates on osteoblasts. Rather, we speculate that decreases may represent indirect, compensatory reductions in bone formation over time. In conclusion, MC3T3-E1 cells and primary osteoblasts were resistant to the negative effects of high ZA doses when grown on calcium phosphate discs or when treated transiently. These experiments may better reflect the prevailing in vivo conditions than those that employ tissue culture plastic and persistent dosing. From a clinical perspective it is reassuring that, under conditions that we consider more representative of osteoblast exposure following ZA dosing, cultured osteoblasts were not adversely affected. Nevertheless, osteoblasts in vivo may also be exposed to bisphosphonate released by resorbing osteoclasts and we plan to examine this phenomenon by co-culturing osteoblasts and osteoclasts. Future experiments will be facilitated by our development of an in vitro system that utilizes a calcium phosphate substrate and/or transitory dosing. This system will provide a key technical framework for us to further investigate the cellular targets of bisphosphonate action. Acknowledgments This work was funded by grants from the Australian Orthopaedic Association and the NF1 Society, as well as charitable donations. We also thank Dr. Andreas Evdokiou and Dr. Jonathan Green for assistance in preparation of the manuscript. References [1] M.J. Rogers, S. Gordon, H.L. Benford, F.P. Coxon, S.P. Luckman, J. Monkkonen, J.C. Frith, Cellular and molecular mechanisms of action of bisphosphonates, Cancer Suppl. 88 (2000) 2961–2978. [2] T. Chen, J. Berenson, R. Vescio, R. Swift, A. Gilchick, S. Goodin, P. LoRusso, P. Ma, C. Ravera, F. Deckert, H. Schran, J. Seaman, A. Skerjanec, Pharmacokinetics and pharmacodynamics of zoledronic acid in cancer, J. Clin. Pharmacol. 42 (2002) 1228–1236. [3] A. Hoffman, D. Stepensky, A. Ezra, J.M. Van Gelder, G. Golomb, Mode of administration-dependent pharmacokinetics of bisphosphonates and bioavailability determination, Int. J. Pharm. 220 (2001) 1– 11. [4] J.H. Lin, G. Russell, B. Gertz, Pharmacokinetics of alendronate: an overview, Int. J. Clin. Pract. Suppl. 101 (1999) 18–26. [5] Y. Ogura, A. Gonsho, J.C. Cyong, H. Orimo, Clinical trial of risedronate in Japanese volunteers: a study on the effects of timing of dosing on absorption, J. Bone Miner. Metab. 22 (2004) 120–126. [6] Y. Ogura, A. Gonsho, J.C. Cyong, H. Orimo, Clinical trial of risedronate in Japanese volunteers: single and multiple oral dose studies, J. Bone Miner. Metab. 22 (2004) 111–119. [7] J.E. Dunford, K. Thompson, F.P. Coxon, S.P. Luckman, F.M. Hahn, C.D. Poulter, F.H. Ebetino, M.J. Rogers, Structure–activity relationships for inhibition of farnesyl diphosphate syntase in vitro and inhibition of bone resorption in vivo by nitrogen-containing bisphosphonates, J. Pharmacol. Exp. Ther. 296 (2001) 235–242. [8] 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. [9] N. Giuliani, M. Pedrazzoni, G. Negri, G. Passeri, M. Impicciatore, G. Girasole, Bisphosphonates stimulate formation of osteoblast precursors and mineralized nodules in murine and human bone marrow
716
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
A. Schindeler, D.G. Little / Biochemical and Biophysical Research Communications 338 (2005) 710–716 cultures in vitro and promote early osteoblastogenesis in young and aged mice in vivo, Bone 22 (1998) 445–461. G.G. Reinholz, B. Getz, L. Pederson, E.S. Sanders, M. Subramaniam, J.N. Ingle, T.C. Spelsberg, Bisphosphonates directly regulated cell proliferation, differentiation, and gene expression in human osteoblasts, Cancer Res. 60 (2000) 6001–6007. G.G. Reinholz, B. Getz, E.S. Sanders, M.Y. Karpeisky, N.S. Padyukova, S.N. Mikhailov, J.N. Ingle, T.C. Spelsberg, Distinct mechanisms of bisphosphonate action between osteoblasts and breast cancer cells: identity of a potent new bisphosphonate analogue, Breast Cancer Res. Treat. 71 (2002) 257–268. O. Fromigue, J.J. Body, Bisphosphonates influence the proliferation and the maturation of normal human osteoblasts, J. Endocrinol. Invest. 25 (2002) 539–546. B. Pan, L.B. To, A.N. Farrugia, D.M. Findlay, J. Green, S. Gronthos, A. Evdokiou, K. Lynch, G.J. Atkins, A.C. Zannettino, The nitrogencontaining bisphosphonate, zoledronic acid, increases mineralisation of human bone-derived cells in vitro, Bone 34 (2004) 112–123. G.-I. Im, S.A. Qureshi, J. Kenney, H.E. Rubash, A.S. Shanbhag, Osteoblast proliferation and maturation by bisphosphonates, Biomaterials 25 (2004) 4105–4115. B. Peter, P.Y. Zambelli, J. Guicheux, D.P. Pioletti, The effect of bisphosphonates and titanium particles on osteoblasts: an in vitro study, J. Bone Joint Surg. Br. 87 (2005) 1157–1163. J. Yu, S.S. Chang, S. Suratwala, W.S. Chung, P. Abdelmessieh, H.J. Lee, J. Yang, F.Y. Lee, Zoledronate induces apoptosis in cells from fibro-cellular membrane of unicameral bone cyst (UBC), J. Orthop. Res. 23 (2005) 1004–1012. F. von Knoch, C. Jaquiery, M. Kowalsky, S. Schaeren, C. Alabre, I. Martin, H.E. Rubash, A.S. Shanbhag, Effects of bisphosphonates on proliferation and osteoblast differentiation of human bone marrow stromal cells, Biomaterials 26 (2005) 6941–6949. M. Glatt, A. Pataki, G.P. Evans, S.B. Hornby, J.R. Green, Loss of vertebral bone and mechanical strength in estrogen-deficient rats is prevented by long-term administration of zoledronic acid, Osteoporos. Int. 15 (2004) 707–715. D.G. Little, R.A. Peat, A. McEvoy, P.R. Williams, E.J. Smith, P.A. Baldock, Zoledronic acid treatment results in retention of femoral head structure after traumatic osteonecrosis in young Wistar rats, J. Bone Miner. Res. 18 (2003) 2016–2022.
[20] R.G. Erben, L. Mosekilde, J.S. Thomsen, K. Weber, K. Stahr, A. Leyshon, S.Y. Smith, R. Phipps, Prevention of bone loss in ovariectomized rats by combined treatment with risedronate and 1alpha,25-dihydroxyvitamin D3, J. Bone Miner. Res. 17 (2002) 1498– 1511. [21] P.M. Chavassieux, M.E. Arlot, J.P. Roux, N. Portero, A. Daifotis, A.J. Yates, N.A. Hamdy, M.P. Malice, D. Freedholm, P.J. Meunier, Effects of alendronate on bone quality and remodeling in glucocorticoid-induced osteoporosis: a histomorphometric analysis of transiliac biopsies, J. Bone Miner. Res. 15 (2000) 754–762. [22] F. Rauch, R. Travers, H. Plotkin, F.H. Glorieux, The effects of intravenous pamidronate on the bone tissue of children and adolescents with osteogenesis imperfecta, J. Clin. Invest. 110 (2002) 1293– 1299. [23] C.V. Odvina, J.E. Zerwekh, D.S. Rao, N. Maalouf, F.A. Gottschalk, C.Y. Pak, Severely suppressed bone turnover: a potential complication of alendronate therapy, J. Clin. Endocrinol. Metab. 90 (2005) 1294–1301. [24] F.P. Coxon, J. Langton, M.J. Rogers, Visualisation of the uptake of a novel fluorescent bisphosphonate by resorbing osteoclasts and J774 cells, Bone 34 (Supp. 1) (2004) S51. [25] M.T. Mulari, Q. Qu, P.L. Harkonen, H.K. Vaananen, Osteoblastlike cells complete osteoclastic bone resorption and form new mineralized bone matrix in vitro, Calcif. Tissue Int. 75 (2004) 253–261. [26] D.R. Sindrey, D. Kusljie, P.C. Kwong, Human parathyroid hormone (1–84) stimulates bone formation in rate bone marrow cultures during spaceflight, ASBMR 21st Annual Meeting (1999) SA424. [27] N. Kurihara, J. Tatsumi, F. Arai, A. Iwama, T. Suda, Macrophage-stimulating protein (MSP) and its receptor, RON, stimulate human osteoclast activity but not proliferation: effect of MSP distinct from that of hepatocyte growth factor, Exp. Hematol. 26 (1998) 1080–1085. [28] S.R. MacNeill, C.M. Cobb, J.W. Rapley, A.G. Glaros, P. Spencer, In vivo comparison of synthetic osseous graft materials. A preliminary study, J. Clin. Periodontol. 26 (1999) 239–245. [29] G. Rawadi, B. Vayssiere, F. Dunn, R. Baron, S. Roman-Roman, BMP-2 controls alkaline phosphatase expression and osteoblast mineralization by a Wnt autocrine loop, J. Bone Miner. Res. 18 (2003) 1842–1853.
Osteoclasts but not osteoblasts are affected by a calcified surface treated with zoledronic acid in vitro Aaron Schindeler a,b,*, David G. Little a,b a
Department of Orthopaedic Research and Biotechnology, The ChildrenÕs Hospital at Westmead, Sydney, Australia b Discipline of Paediatrics and Child Health, Faculty of Medicine, University of Sydney, Sydney, Australia Received 26 September 2005 Available online 11 October 2005
Abstract Bisphosphonates are potent inhibitors of osteoclast-mediated bone resorption. Recent interest has centered on the effects of bisphosphonates on osteoblasts. Chronic dosing of osteoblasts with solubilized bisphosphonates has been reported to enhance osteogenesis and mineralization in vitro. However, this methodology poorly reflects the in vivo situation, where free bisphosphonate becomes rapidly bound to mineralized bone surfaces. To establish a more clinically relevant cell culture model, we cultured bone cells on calcium phosphate coated quartz discs pre-treated with the potent nitrogen-containing bisphosphonate, zoledronic acid (ZA). Binding studies utilizing [14C]-labeled ZA confirmed that the bisphosphonate bound in a concentration-dependent manner over the 1–50 lM dose range. When grown on ZA-treated discs, the viability of bone-marrow derived osteoclasts was greatly reduced, while the viability and mineralization of the osteoblastic MC3T3-E1 cell line were largely unaffected. This suggests that only bone resorbing cells are affected by bound bisphosphonate. However, this system does not account for transient exposure to unbound bisphosphonate in the hours following a clinical dosing. To model this event, we transiently treated osteoblasts with ZA in the absence of a calcified surface. Osteoblasts proved highly resistant to all transitory treatment regimes, even when utilizing ZA concentrations that prevented mineralization and/or induced cell death when dosed chronically. This study represents a pharmacologically more relevant approach to modeling bisphosphonate treatment on cultured bone cells and implies that bisphosphonate therapies may not directly affect osteoblasts at bone surfaces. 2005 Elsevier Inc. All rights reserved. Keywords: Bisphosphonates; Osteoblasts; Osteoclasts; Osteoporosis: Therapy; Bone resorption
Nitrogen-containing bisphosphonates (N-BPs) are a class of drugs used in the treatment of osteoporosis and diseases of high bone turnover [1]. After administration, a considerable proportion of an N-BP binds with high affinity to hydroxyapatite at the bone surface, while the remainder is rapidly excreted by the renal system [2–6]. Once released from hydroxyapatite and internalized by cells, N-BPs act as potent inhibitors of farnesyl diphosphate synthase (FPPS), an enzyme in the mevalonate pathway [7]. The presence of bound N-BP at the bone surface can effectively reduce osteoclast-mediated bone turnover. Inhibition of FPPS prevents the prenylation of small GTP-
*
Corresponding author. Fax: +61 2 98453078. E-mail address: [email protected] (A. Schindeler).
0006-291X/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2005.09.198
ases such as Ras, Rac, and Cdc42, thus resulting in decreased osteoclastogenesis, osteoclast survival, and/or osteoclast function [1,8]. In addition to affecting osteoclasts, it has been speculated that bound N-BPs may also influence the differentiation, function or survival of osteoblasts. Numerous studies have investigated the effects of various N-BPs, including zoledronic acid (ZA), alendronate, pamidronate, and risedronate, on cultured cells [9–17]. The majority of these studies have involved the chronic exposure of osteoblasts and osteoblast-like cell lines to solubilized N-BPs and then measuring cell viability, alkaline phosphatase (ALP) activity, mineralization, and/or osteogenic gene expression. Although the individual N-BPs, doses, and experimental outcomes vary considerably, it was frequently concluded that specific doses of N-BPs could have mildly positive
A. Schindeler, D.G. Little / Biochemical and Biophysical Research Communications 338 (2005) 710–716
effects on proliferation or differentiation at certain time points. These data conflict with animal and clinical studies that suggest long-term N-BP treatment can lead to reduced bone formation [18–23]. Recently, Coxon et al. [24] have investigated the release, uptake, and internalization of a fluorescently labeled N-BP by cultured cells. Resorbing osteoclasts were able to take up the labeled N-BP bound to a dentine substrate, while macrophages and non-resorbing osteoclasts could not. Thus, the acidic environment or enzymes produced specifically by osteoclasts may be required for the release of NBPs from a bone surface. Although osteoblasts have been reported to have some capacity to degrade organic components of the bone matrix [25], we speculated they lack the capacity to release and internalize bound N-BPs. To address this issue, we have employed a cell culture system that incorporates a calcified surface and a radiolabeled N-BP. Data obtained using this system suggest that osteoblasts are unable to release and take up pharmacologically significant amounts of bound bisphosphonate. Nevertheless, osteoblasts may be exposed to serum levels of unbound N-BP immediately following a clinical dosing. Certainly non-resorbing macrophages were able to take up solubilized fluorescent N-BP when cultured on plastic [24], consistent with other cell culture studies [8– 17]. The length of transient exposure may be dependent on the dosing regime. In the case of intravenous treatment, the method used for ZA dosing, bisphosphonate is rapidly cleared to the bones or kidneys within hours of injection [2]. In contrast, the bioavailability of oral bisphosphonates such as alendronate is less (leading to higher amounts of drug being administered), and absorption is prolonged compared with an intravenous dosing [3,4]. Thus, it is possible that these treatment regimes may have different effects on osteoblasts. To model the limited bisphosphonate exposure that follows a clinical dosing, MC3T3-E1 cells and primary osteoblasts were dosed with ZA for a 2 h pulse (analogous to intravenous dosing [2]), a 1 day pulse (analogous to oral weekly dosing [3,4]) or continuously (consistent with prior cell culture studies [9–17]). A range of ZA concentrations were used to reflect both the expected pharmacological exposure and the elevated doses used in vitro by other investigators. These studies attempt to generate more clinically relevant approaches to in vitro experimental design and may help explain some of the disparities between published in vitro and in vivo data. Materials and methods Calcium phosphate discs. The Osteologic disc system was used to model bisphosphonate binding and cell activity at the bone surface (BD Biosciences, NJ, USA). This product is a practical alternative to bone or dentine slices, exhibits greater consistency across the surface, is simple to visualize by phase contrast microscopy, and is feasible for both osteoblast and osteoclast culture [26,27]. The discs, supplied in 24-well plate format, were 12.7 mm in diameter and consisted of quartz layered with 0.65 lm of calcium phosphate. For bisphosphonate pre-treatment, discs were incu-
711
bated for 2 h with 1–50 lM ZA dissolved in 1 ml of culture medium. Discs were washed with 3 · 1 ml of culture media prior to seeding with cells. New bone formation was measured on calcium phosphate discs by treating cells for 6 days with 4 lg/ml calcein and evaluating fluorescence on a DC500 microscope (Leica Microsystems, NSW, Australia). This system for measuring mineralization on discs in vitro is based on previously published methods [26]. [14C]ZA binding assays. [14C]-radiolabeled zoledronic acid ([14C]ZA) was generously supplied by Novartis Pharma AG, Basel, Switzerland. The central carbon atom between the two phosphonic acid groups (carbon 1) of ZA was labeled with [14C]; since N-BPs are metabolically stable we expect no re-arrangement of the [14C] within the molecule nor its release under physiological conditions. Bound [14C]ZA was rapidly eluted from discs by decalcifying with 0.5 ml of 5 M HCl. Unbound and bound [14C]ZA were quantified using OptiPhase SuperMix (Perkin-Elmer, MA, USA) in a Tri-Carb scintillation counter (Packard, MN, USA). Values in counts per minute (cpm) were calibrated using a control vial containing a known quantity of [14C]ZA. MC3T3-E1 cell culture. MC3T3-E1 cells were grown in a-MEM (Gibco Laboratories, CA, USA) supplemented with 10% FBS, 2 mM Lglutamine, and antibiotics (100 U/ml penicillin and 0.1 mg/ml streptomycin). Cultures were incubated at 37 C in the presence of 5% CO2. Osteogenic differentiation was initiated on confluent MC3T3-E1 cells using medium further supplemented with 50 mg/L ascorbic acid and 10 mM b-glycerophosphate. Treatment with Zoledronic Acid (ZA, kindly supplied by Novartis Pharmaceuticals, NSW, Australia) was conducted in normal medium supplemented with the drug. Medium changes were made every 3–4 days, including the addition of fresh ZA for continuous dosing regimes. Preparatin and culture of bone marrow-derived murine osteoclasts. Femoral bone marrow from 3- to 4-month- old male mice was extracted using a 21 gauge needle and the mononuclear lymphocytes were subsequently purified using Histopaque-1077 (Sigma) as per the manufacturerÕs instructions. Mononuclear cells were plated at 2.5 · 105 cells/well in 24well plates, cultured in media consisting of a-MEM containing 10% FBS, 2 mM L-glutamine, and antibiotics, and supplemented with 30 ng/ml RANK-L (R&D Systems), and 5 ng/ml M-CSF (R&D Systems). Media changes were performed every 3 days. Preparation and culture of primary calvarial murine osteoblasts. Neonatal mice (days 2–3) were killed by decapitation. Calvaria were then dissected, thoroughly minced, and the fragments were extensively washed with sterile PBS. Bone pieces were digested with 0.05% Trypsin (CSL, Victoria, Australia), 0.3 mM EDTA, and 2.5 mg/ml Collagenase A (Roche Molecular Biochemicals, NSW, Australia) in PBS. Cells were passed through 19G and 21G needles and then filtered through a sterile mesh. Cells were counted using a hemocytometer and plated at a concentration of 5 · 104 cells/ml in 6, 12 or 24-well plates. Primary calvarial murine osteoblasts were grown in a-MEM supplemented with 10% FBS, 2 mM L-glutamine, antibiotics (100 U/ml penicillin and 0.1 mg/ml streptomycin), 50 mg/L ascorbic acid, 10 mM b-glycerophosphate, 20 mM Na2CO3, 34 mM Hepes, and 34 mg/L gentamycin. Drug interventions commenced after osteoblasts had reached confluence (days 5–7 after preparation). Cell viability and alkaline phosphatase assays. Osteoblast viability was determined at day 3 using the CellTiter 96 AQueous One Solution Cell Proliferation Assay system (Promega, NSW, Australia), as per the manufacturerÕs instructions. Alkaline phosphatase (ALP) activity assays were performed at day 4. Osteoblasts were harvested in 50 ll of a lysis buffer containing 20 mM Tris–HCl (pH 8.2), 2 mM MgCl2, and 0.05% Triton X100. Subsequently, cell lysate (normalized for total protein) was added to 300 ll ALP assay buffer containing 0.1 M NaHCO3, 1 mM MgCl2, and 10 mM p-nitrophenyl phosphate, and incubated for 15 min at 37 C. Absorbance at 405 nm was read on a microplate reader (Labsystems, MA, USA). Tartrate-resistant acid phosphatase staining. Osteoclasts were identified using a stain for tartrate-resistant acid phosphatase (TRAP). Cells were then incubated for 5 min at 37 C with freshly made TRAP staining solution utilized in tissue histology [28].
712
A. Schindeler, D.G. Little / Biochemical and Biophysical Research Communications 338 (2005) 710–716
Alizarin Red S staining. In brief, cells were rinsed twice with PBS, fixed with 4% paraformaldehyde in PBS, washed twice with deionized water, and then stained with freshly made 40 mM Alizarin Red S (pH 4.2) for 15 min. The Alizarin Red S solution was then removed by aspiration and cells were rinsed five times with deionized water. Mineralization was quantified using NIH Image software. Statistical methods. Statistical significance was calculated using unpaired two-tailed studentÕs t test to compare treated samples with untreated controls. Error bars on graphs represent ± the standard error. Experiments were performed in duplicate or triplicate wells, and independent trials were repeated a minimum of two times, unless otherwise noted.
affected by ZA released from the discs acting in a paracrine manner. Osteoblast viability is unaffected on a ZA-treated calcified surface
Binding experiments utilizing [14C]ZA revealed that the bisphosphonate bound and concentrated at the disc surface (Table 1). A range of ZA concentrations were used (1– 50 lM) with the lowest dose approximating the serum concentration of ZA immediately following a clinical dosing [2] and the highest dose reported to be lethal when chronically dosed in vitro [9–14]. After 2 h of incubation, 5–10% of ZA in the liquid phase bound to the 0.65 lm calcium phosphate substrate (i.e., 0.2–1.83 nmol ZA). Based on a matrix volume of 8.23 · 107 lm3 for a 12.7 mm disc, this translated to local matrix ZA concentrations of 1.0–9.0 mM, which was up to 180-fold higher than the highest soluble dose.
To test whether osteoblasts behaved in the same manner as osteoclasts, mouse MC3T3-E1 osteoprogenitor cells were plated on calcium phosphate discs and compared to culture on plastic. Cell proliferation assays at day 3 showed a 1.25-fold increase in the number of osteoblasts on the discs as compared to plastic (Fig. 1B). When cultured on calcified discs treated with 1–50 lM ZA, no significant difference in cell viability was seen (Fig. 1B), despite ZA being concentrated at the disc surface (Table 1). In a separate experiment, mineralized matrix production on calcium phosphate discs was measured by calcein incorporation, with ZA-coating having no significant effect on mineralization (data not shown). In contrast, osteoblasts treated with 1–50 lM ZA in the media and grown on tissue culture plastic showed decreased viability in a dose-dependent manner. This dose–response curve is consistent with several studies that have chronically treated cells with ZA in vitro [10,11,15,16]. Significant cell death was seen at the maximal drug dose (50 lM ZA), which was not evident on discs pre-treated or post-treated (after cell plating) for 2 h with 50 lM ZA (Fig. 1C).
Osteoclast viability is reduced on a ZA-treated calcified surface
Osteoblasts inefficiently take up ZA bound to calcium phosphate discs
To confirm the validity of the in vitro model, bone marrow derived lymphocytes were cultured in the presence of the pro-osteoclastic agents RANK-L and M-CSF on untreated discs and discs pre-treated with 10 lM ZA. Following 12 days of culture, cells were stained for TRAP expression and revealed high numbers of multinucleate TRAP+ cells on untreated discs. In comparison TRAP+ cells were vastly reduced on ZA-pretreated discs (Fig. 1A). These data suggest that the bisphosphonate bound to CaPO4 discs can act specifically on resorbing osteoclasts in vitro. In addition, normal numbers of viable multinucleate osteoclasts were found on plastic surrounding both ZA-treated and untreated discs (data not shown). This indicates that cells on ZA-treated discs were not being
To directly measure bisphosphonate absorption by osteoblasts, MC3T3-E1 cells were treated with [14C]ZA. Following detergent lysis, osteoblasts treated with unbound 10 lM [14C]ZA (i.e., grown on plastic) were found to take up 80 pmol of drug over 3 days. Despite a relatively high local concentration in the CaPO4 matrix (Table 1), osteoblasts grown on [14C]ZA coated discs absorbed a dose of only 50 pmol over the same interval. These results further suggest that bound ZA cannot be efficiently released and internalized by osteoblasts.
Results ZA binds to calcium phosphate discs
Table 1 [14C]ZA binding to calcium phosphate discsa ZA treatment (nmol)
Bound ZA (nmol)
Matrix ZA conc. (mM)
0 1 10 50
0 0.20 0.74 1.83
0.0 1.0 3.6 9.0
a ZA treatments were performed in 1 ml of osteoblast growth media, so that 1 nmol ZA gave a final concentration of 1 lM. The concentration of bound ZA within the 0.65 lm CaPO4 layer (Matrix ZA conc.) reached far higher levels than in the soluble ZA treatments.
MC3T3-E1 cells survive transitory dosing with ZA Although osteoblasts may not efficiently take up bound ZA, we postulated that these cells would still be exposed to transitory doses of unbound ZA during a clinical dosing. To model this event in culture, we treated MC3T3-E1 cells grown on plastic with transient doses of 1–50 lM ZA for 2 h and 1 day. The 2 h dose was used to reflect the pharmacokinetics of an intravenous bisphosphonate [3] while the 1 day dose was chosen to reflect the prolonged presence of an oral bisphosphonate [4]. Both oral and intravenous bisphosphonate are typically given to patients on a weekly or monthly basis, so cells were transiently dosed a single time. Cell viability was measured at day 3 of treatment and compared with that of untreated cells. As previously
A. Schindeler, D.G. Little / Biochemical and Biophysical Research Communications 338 (2005) 710–716
713
Fig. 1. Osteoclasts but not osteoblasts are affected on calcium phosphate discs pre-treated with ZA. (A) TRAP+ multinucleated cells (MNCs) generated from murine bone marrow cells were counted on calcium phosphate discs in five random fields. Significantly more TRAP+ MNCs were present at day 12 on untreated discs compared with discs pre-treated with 10 lM ZA *P < 0.05. Data reflect the results of two independent experiments. (B) Cultured MC3T3-E1 cells were exposed to either soluble ZA (on plastic) or to bound ZA (on discs pre-treated with ZA solution). All treatment groups were statistically compared to osteoblasts grown on plastic. Osteoblasts dosed with soluble ZA (on plastic) were sensitive to treatment with 10 lM ZA (*P < 0.07) and 50 lM ZA (**P < 0.05); however, no significant reduction in cell viability was seen on discs pre-coated with similar doses of ZA. A significant 1.3-fold increase in viable osteoblasts was observed on untreated calcium phosphate discs and discs pre-treated with 1 lM ZA (***P < 0.05). These data summarize the results from three independent experiments. (C) Phase contrast images of differentiating osteoblasts on plastic and calcium phosphate discs after 4 days. When exposed to 50 lM soluble ZA, significant cell death was seen in osteoblast cultures cultured on plastic. In contrast, osteoblasts grown on calcium phosphate discs either pre-treated or post-treated for 2 h with 50 lM ZA showed no phenotypic effects.
reported by others, cell viability was reduced in chronic 10– 50 lM ZA doses (Figs. 1C and 2A). At 50 lM ZA, significant cell death was observed (Fig. 1C), which previous reports have demonstrated to be apoptosis [13,16]. In contrast, more limited treatments (2 h or 1 day) did not significantly affect cell viability (Fig. 2A), even at the highest ZA dose used. Transitory ZA treatment slightly enhances osteogenic differentiation and mineralization Next, alkaline phosphatase (ALP) activity was measured as a marker of osteogenic differentiation. Assays were normalized for total protein, thus accounting for differences in cell viability. Chronic ZA dosing did not have any significant benefit on osteoblast differentiation, even causing a
slight decrease in differentiation at higher doses (Fig. 2B). Transient dosing with ZA achieved a slight increase in ALP activity, with the most significant being a 2-fold increase in the 1 lM ZA for 1 day treatment sample (Fig. 2B). A more functionally relevant endpoint is to measure osteoblast mineralization, which was achieved by staining extracellular matrix calcium deposits with Alizarin Red S stain. MC3T3-E1 cells treated as previously described were fixed and stained at day 8. All wells displayed an even staining pattern, reflective of the homogeneous nature of the cell line. Staining intensity was measured using imaging software and the findings paralleled those obtained from earlier ALP assays (Fig. 3A). These small increases in mineralization were consistently maintained when cultures were taken out to longer time points of 14 and 21 days (data not shown).
714
A. Schindeler, D.G. Little / Biochemical and Biophysical Research Communications 338 (2005) 710–716
Fig. 2. Transitory ZA dosing does not negatively affect viability and ALP activity in MC3T3-E1 cells. (A) Chronic dosing with ZA caused a decrease in cell viability compared with untreated cells that corresponded with ZA dose (*P < 0.05). No negative effects on cell viability were noted with transient doses of ZA, even at the highest concentrations of ZA (50 lM). (B) ALP activity at day 4 was negatively affected by chronic ZA and positively affected by brief exposure to low doses of ZA (*P < 0.05 compared to untreated cells). These data summarize the results from three independent experiments.
To confirm that these effects were not specific to the MC3T3-E1 cell line, primary osteoblasts were isolated and cultured from neonatal mouse calvaria and treated with identical ZA dosing regimes. Mineralization was measured at day 14 after reaching confluence, by which time cells had formed numerous mineralized nodules (Fig. 3B). Quantification using image analysis software revealed that mineralization nodule formation by primary osteoblasts mirrored the results from the MC3T3-E1 cells (data not shown). However, the overall effect of ZA on primary osteoblast mineralization was reduced compared to the MC3T3-E1 mineralization (e.g., only a 1.4-fold increase was seen in the 1 lM ZA for 1 day sample). Discussion Clinical success using bisphosphonates has led to speculation that these drugs may generate a pro-anabolic re-
Fig. 3. Transitory ZA does not negatively affect the mineralization of MC3T3-E1 cells or primary calvarial osteoblasts. (A) Monolayers of differentiated MC3T3-E1 cells were stained for mineralization using Alizarin Red S and quantified on digital images. Decreased mineralization was observed in continuously treated samples, while increased mineralization was seen in 1 lM/1 day dose, 10 lM/1 day dose, and 1 lM/2 h dose (*P < 0.05 compared with untreated MC3T3-E1 cells). (B) Primary osteoblasts were similarly stained at day 14. Mineralized nodules (red patches) were apparent in all samples, not including cells treated continuously with 50 lM ZA. Image analysis indicated results were compatible those observed from the MC3T3-E1 cell line.
sponse in addition to their anti-catabolic activity. While evidence from in vivo studies generally suggests that bone formation is reduced by bisphosphonate action [18–23], experiments with cultured osteoblasts have often implied
A. Schindeler, D.G. Little / Biochemical and Biophysical Research Communications 338 (2005) 710–716
a potential anabolic mechanism [9–17]. In vitro tests have revealed increases in proliferation, ALP activity, mineralization, and osteogenic markers with specific bisphosphonate treatment regimes [9–14,17]. We hypothesized that this apparent conflict in the literature may be attributed to the prolonged soluble delivery of bisphosphonate to osteoblasts in vitro, which differs from circumstances in vivo. To test this premise, we have compared standard in vitro techniques to what we propose to be more pharmacologically relevant in vitro techniques, including the provision of a calcified surface and transitory soluble dosing. We have demonstrated that the effects of exposing osteoblasts to bisphosphonate bound to a calcified surface can differ from treatment with soluble bisphosphonate. Experiments using [14C]ZA confirmed that ZA binds to calcified discs, which concentrates at the disc surface. While osteoclasts were negatively affected when cultured on a ZA-coated calcified surface, osteoblasts survived, proliferated, and produced mineralized matrix under the same conditions. This has significant implications for the future of cell culture modeling of bisphosphonate treatment. Our results are also consistent with observations made by Coxon et al. [24], which indicate that fluorescently labeled alendronate bound to dentine could be taken up by resorbing osteoclasts, but not macrophages or non-resorbing osteoclasts. To further extend our in vitro model, we simulated the transient exposure of cells to soluble bisphosphonate prior to its binding to mineralized surfaces. This is analogous to the situation that immediately follows a clinical N-BP treatment [2–4]. Although cell culture media may not precisely represent the cation and plasma protein mix normally seen in human serum, the 1 lM ZA dose corresponds to the plasma concentration after an intravenous dose [2]. Osteoblasts were grown on plastic and treated with soluble ZA. Transient dosing (for 2 h or 1 day) was compared with chronic dosing, which is already well reported in the literature [9–17]. In our transitory treatment model, we demonstrated that ZA can have a subtle (never greater than 2-fold) but positive effect on ALP activity and mineralization. It is not clear whether this small increase could be translated into a clinically meaningful anabolic effect. Certainly ALP activity is increased far more substantially when osteoblasts are treated in vitro with established proanabolic agents. For example, treatment with 100 ng/ml BMP-2 was found to increase ALP activity 13-fold in MC3T3-E1 cells [29]. Thus our results, like other in vitro studies [9–14,17], suggest the potential for small but positive anabolic benefits on bone formation via direct effects on osteoblasts. These results diverge from animal studies showing that bisphosphonates have negligible or negative effects on mineral apposition rate or bone formation rate [18–21]. Bone formation has also been reported to be reduced in patients undergoing long-term bisphosphonate therapies [22,23]. However, there is no evidence that these examples of re-
715
duced bone formation in vivo are due to the direct effects of bisphosphonates on osteoblasts. Rather, we speculate that decreases may represent indirect, compensatory reductions in bone formation over time. In conclusion, MC3T3-E1 cells and primary osteoblasts were resistant to the negative effects of high ZA doses when grown on calcium phosphate discs or when treated transiently. These experiments may better reflect the prevailing in vivo conditions than those that employ tissue culture plastic and persistent dosing. From a clinical perspective it is reassuring that, under conditions that we consider more representative of osteoblast exposure following ZA dosing, cultured osteoblasts were not adversely affected. Nevertheless, osteoblasts in vivo may also be exposed to bisphosphonate released by resorbing osteoclasts and we plan to examine this phenomenon by co-culturing osteoblasts and osteoclasts. Future experiments will be facilitated by our development of an in vitro system that utilizes a calcium phosphate substrate and/or transitory dosing. This system will provide a key technical framework for us to further investigate the cellular targets of bisphosphonate action. Acknowledgments This work was funded by grants from the Australian Orthopaedic Association and the NF1 Society, as well as charitable donations. We also thank Dr. Andreas Evdokiou and Dr. Jonathan Green for assistance in preparation of the manuscript. References [1] M.J. Rogers, S. Gordon, H.L. Benford, F.P. Coxon, S.P. Luckman, J. Monkkonen, J.C. Frith, Cellular and molecular mechanisms of action of bisphosphonates, Cancer Suppl. 88 (2000) 2961–2978. [2] T. Chen, J. Berenson, R. Vescio, R. Swift, A. Gilchick, S. Goodin, P. LoRusso, P. Ma, C. Ravera, F. Deckert, H. Schran, J. Seaman, A. Skerjanec, Pharmacokinetics and pharmacodynamics of zoledronic acid in cancer, J. Clin. Pharmacol. 42 (2002) 1228–1236. [3] A. Hoffman, D. Stepensky, A. Ezra, J.M. Van Gelder, G. Golomb, Mode of administration-dependent pharmacokinetics of bisphosphonates and bioavailability determination, Int. J. Pharm. 220 (2001) 1– 11. [4] J.H. Lin, G. Russell, B. Gertz, Pharmacokinetics of alendronate: an overview, Int. J. Clin. Pract. Suppl. 101 (1999) 18–26. [5] Y. Ogura, A. Gonsho, J.C. Cyong, H. Orimo, Clinical trial of risedronate in Japanese volunteers: a study on the effects of timing of dosing on absorption, J. Bone Miner. Metab. 22 (2004) 120–126. [6] Y. Ogura, A. Gonsho, J.C. Cyong, H. Orimo, Clinical trial of risedronate in Japanese volunteers: single and multiple oral dose studies, J. Bone Miner. Metab. 22 (2004) 111–119. [7] J.E. Dunford, K. Thompson, F.P. Coxon, S.P. Luckman, F.M. Hahn, C.D. Poulter, F.H. Ebetino, M.J. Rogers, Structure–activity relationships for inhibition of farnesyl diphosphate syntase in vitro and inhibition of bone resorption in vivo by nitrogen-containing bisphosphonates, J. Pharmacol. Exp. Ther. 296 (2001) 235–242. [8] 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. [9] N. Giuliani, M. Pedrazzoni, G. Negri, G. Passeri, M. Impicciatore, G. Girasole, Bisphosphonates stimulate formation of osteoblast precursors and mineralized nodules in murine and human bone marrow
716
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
A. Schindeler, D.G. Little / Biochemical and Biophysical Research Communications 338 (2005) 710–716 cultures in vitro and promote early osteoblastogenesis in young and aged mice in vivo, Bone 22 (1998) 445–461. G.G. Reinholz, B. Getz, L. Pederson, E.S. Sanders, M. Subramaniam, J.N. Ingle, T.C. Spelsberg, Bisphosphonates directly regulated cell proliferation, differentiation, and gene expression in human osteoblasts, Cancer Res. 60 (2000) 6001–6007. G.G. Reinholz, B. Getz, E.S. Sanders, M.Y. Karpeisky, N.S. Padyukova, S.N. Mikhailov, J.N. Ingle, T.C. Spelsberg, Distinct mechanisms of bisphosphonate action between osteoblasts and breast cancer cells: identity of a potent new bisphosphonate analogue, Breast Cancer Res. Treat. 71 (2002) 257–268. O. Fromigue, J.J. Body, Bisphosphonates influence the proliferation and the maturation of normal human osteoblasts, J. Endocrinol. Invest. 25 (2002) 539–546. B. Pan, L.B. To, A.N. Farrugia, D.M. Findlay, J. Green, S. Gronthos, A. Evdokiou, K. Lynch, G.J. Atkins, A.C. Zannettino, The nitrogencontaining bisphosphonate, zoledronic acid, increases mineralisation of human bone-derived cells in vitro, Bone 34 (2004) 112–123. G.-I. Im, S.A. Qureshi, J. Kenney, H.E. Rubash, A.S. Shanbhag, Osteoblast proliferation and maturation by bisphosphonates, Biomaterials 25 (2004) 4105–4115. B. Peter, P.Y. Zambelli, J. Guicheux, D.P. Pioletti, The effect of bisphosphonates and titanium particles on osteoblasts: an in vitro study, J. Bone Joint Surg. Br. 87 (2005) 1157–1163. J. Yu, S.S. Chang, S. Suratwala, W.S. Chung, P. Abdelmessieh, H.J. Lee, J. Yang, F.Y. Lee, Zoledronate induces apoptosis in cells from fibro-cellular membrane of unicameral bone cyst (UBC), J. Orthop. Res. 23 (2005) 1004–1012. F. von Knoch, C. Jaquiery, M. Kowalsky, S. Schaeren, C. Alabre, I. Martin, H.E. Rubash, A.S. Shanbhag, Effects of bisphosphonates on proliferation and osteoblast differentiation of human bone marrow stromal cells, Biomaterials 26 (2005) 6941–6949. M. Glatt, A. Pataki, G.P. Evans, S.B. Hornby, J.R. Green, Loss of vertebral bone and mechanical strength in estrogen-deficient rats is prevented by long-term administration of zoledronic acid, Osteoporos. Int. 15 (2004) 707–715. D.G. Little, R.A. Peat, A. McEvoy, P.R. Williams, E.J. Smith, P.A. Baldock, Zoledronic acid treatment results in retention of femoral head structure after traumatic osteonecrosis in young Wistar rats, J. Bone Miner. Res. 18 (2003) 2016–2022.
[20] R.G. Erben, L. Mosekilde, J.S. Thomsen, K. Weber, K. Stahr, A. Leyshon, S.Y. Smith, R. Phipps, Prevention of bone loss in ovariectomized rats by combined treatment with risedronate and 1alpha,25-dihydroxyvitamin D3, J. Bone Miner. Res. 17 (2002) 1498– 1511. [21] P.M. Chavassieux, M.E. Arlot, J.P. Roux, N. Portero, A. Daifotis, A.J. Yates, N.A. Hamdy, M.P. Malice, D. Freedholm, P.J. Meunier, Effects of alendronate on bone quality and remodeling in glucocorticoid-induced osteoporosis: a histomorphometric analysis of transiliac biopsies, J. Bone Miner. Res. 15 (2000) 754–762. [22] F. Rauch, R. Travers, H. Plotkin, F.H. Glorieux, The effects of intravenous pamidronate on the bone tissue of children and adolescents with osteogenesis imperfecta, J. Clin. Invest. 110 (2002) 1293– 1299. [23] C.V. Odvina, J.E. Zerwekh, D.S. Rao, N. Maalouf, F.A. Gottschalk, C.Y. Pak, Severely suppressed bone turnover: a potential complication of alendronate therapy, J. Clin. Endocrinol. Metab. 90 (2005) 1294–1301. [24] F.P. Coxon, J. Langton, M.J. Rogers, Visualisation of the uptake of a novel fluorescent bisphosphonate by resorbing osteoclasts and J774 cells, Bone 34 (Supp. 1) (2004) S51. [25] M.T. Mulari, Q. Qu, P.L. Harkonen, H.K. Vaananen, Osteoblastlike cells complete osteoclastic bone resorption and form new mineralized bone matrix in vitro, Calcif. Tissue Int. 75 (2004) 253–261. [26] D.R. Sindrey, D. Kusljie, P.C. Kwong, Human parathyroid hormone (1–84) stimulates bone formation in rate bone marrow cultures during spaceflight, ASBMR 21st Annual Meeting (1999) SA424. [27] N. Kurihara, J. Tatsumi, F. Arai, A. Iwama, T. Suda, Macrophage-stimulating protein (MSP) and its receptor, RON, stimulate human osteoclast activity but not proliferation: effect of MSP distinct from that of hepatocyte growth factor, Exp. Hematol. 26 (1998) 1080–1085. [28] S.R. MacNeill, C.M. Cobb, J.W. Rapley, A.G. Glaros, P. Spencer, In vivo comparison of synthetic osseous graft materials. A preliminary study, J. Clin. Periodontol. 26 (1999) 239–245. [29] G. Rawadi, B. Vayssiere, F. Dunn, R. Baron, S. Roman-Roman, BMP-2 controls alkaline phosphatase expression and osteoblast mineralization by a Wnt autocrine loop, J. Bone Miner. Res. 18 (2003) 1842–1853.