Characterization of biomimetic calcium phosphate labeled with fluorescent dextran for

quantification of osteoclastic activity.

Salwa M. Maria1,2, Christiane Prukner3, Zeeshan Sheikh1, Frank Mueller3, Svetlana V.

Komarova1,2, Jake. E. Barralet1,4

1Faculty of Dentistry, McGill University, Canada, 2Shriners Hospital for Children-Canada,

Montreal, Canada, 3Institute of Materials Science and Technology, Friedrich Schiller

University, Jena, Germany, 4Orthopaedics Division, Department of Surgery, Faculty of

Medicine, Montreal General Hospital, McGill University

Corresponding author: Jake. E. Barralet, McGill University, Montreal, Quebec, Canada H3G IA6, Telephone: 514-398-3908

Abstract:

Bone resorbing osteoclasts represent an important therapeutic target for diseases associated

with bone and joint destruction, such as rheumatoid arthritis, periodontitis, and osteoporosis.

The quantification of osteoclast resorptive activity in vitro is widely used for screening new

anti-resorptive medications. The aim of this paper was to develop a simplified semi-

automated method for the quantification of osteoclastic resorption using fluorescently labeled

biomimetic mineral layers with which can replace time intensive, often subjective and clearly

non-sustainable use of translucent slices of tusks from vulnerable or endangered species such

as the elephant.. Osteoclasts were formed from RAW 264.7 mouse monocyte cell line using

the pro-resorptive cytokine receptor activator of nuclear factor kappa-B ligand (RANKL). We

confirmed that fluorescent labeling did not interfere with the biomimetic features of

hydroxyapatite, and developed an automated method for quantifying osteoclastic resorption.

Correlation between our assay and traditional manual measurement techniques was found to

be very strong (R2 = 99). In addition, we modified the technique to provide depth and volume

data of the resorption pits by confocal imaging at defined depths. Thus, our method allows

automatic quantification of total osteoclastic resorption as well as additional data not

obtainable by the current tusk slice technique offering a better alternative for high throughput

screening of potential antiresorptives.

Introduction:

Osteoclasts are bone cells responsible for the resorption of the mineralized tissues in the normal

bone remodeling process. The abnormal resorptive activity of osteoclasts occurring in serious

disorders, such as osteoporosis [1], rheumatoid arthritis [2]; [3]; [4]; [5], periodontitis [6] and

cancer metastasis to bone [7]. Therefore, targeting osteoclastic resorption is an important

therapeutic strategy. While a number of therapies including bisphosphonates [8] and anti-

RANKL denosumab [9] are now successfully used to limit osteoclastic activity, intolerant and

resistant cases necessitate the development of novel treatments. For this, the ability to rapidly

screen the candidate chemical compounds is important.

We have previously validated a method of precipitating a thin layer of hydroxyapatite on

tissue culture plates and glass coverslips [10] to study osteoclast resorptive activity [11]. Calcium

phosphate substrates, including hydroxyapatite, can be fluorescently labeled, for example with

calcein, which was used to quantify osteoblast mineralization [12], or fluoresceinamine labeled

chondroitin polysulfate or calcein to study osteoclastic resorption [13].

Dextran is a nontoxic, water-soluble polysaccharide of a complex branched glucan,

composed of many glucose molecules with variable chain lengths [14]. Dextran can be

conjugated to a variety of fluorophores which expand the possibility of using these substrates for

the fluorescence studies. Dextran has the ability to chemically bind to the hydroxyapatite [15];

[16]. Dextran can also be internalized by a number of cells capable of endocytosis, including

osteoclasts [17].

In this study we used dextran (10,000 Da), conjugated to Alexa Fluor488 or Rhodamine

B to fluorescently label hydroxyapatite, and validated that the new fluorescent substrate is

supportive of osteoclast formation and resorption and amenable for automated quantification.

Materials and Methods:

Reagents: Chemicals used for mineral coating were purchased from Sigma and Fisher Scientific.

Dulbecco’s modification of eagle’s medium 1x, (DMEM), fetal bovine serum (FBS, 08050),

penicillin-streptomycin (450-202-EL), sodium pyruvate (600110-EL) and 0.25% trypsin with

0.1% EDTA, (325-043-EL) were from Wisent Inc, Quebec, Canada. Fluorescein isothiocyanate-

labeled-phalloidin (Phalloidin conjugate FITC) P5282, alendronate (A4978) were from Sigma-

Aldrich, Ontario, Canada. Recombinant glutathione S-transferase-soluble RANKL was purified

from the clones kindly provided by Dr. MF Manolson (University of Toronto, Canada) and

freshly reconstituted before each experiment. Dextran Alexa Fluor® 488 (D-22910) and

Rhodamine B (D-1824) of 10000 Da were from Invitrogen, New York, USA. Lactate

dehydrogenase activity assay kit plus (LDH) was from Roche Applied Science, UK.

Hydroxyapatite coating: The following solutions were prepared freshly before each coating as

described in detail in [10], Briefly, 2.5-fold simulated body fluid (SBF) was prepared by

mixing 50% Tris buffer (50 mM Tris base, pH=7.4 with 1M HC1), 25% calcium stock solution

(25 mM CaCl2·H2O, 1.37 M NaCl, 15 mM MgCl2·6H2O in Tris buffer, pH 7.4), and 25%

phosphate stock solution (11.1 mM Na2HPO4·H2O, 42 mM NaHCO3 in Tris buffer, pH=7.4).

Calcium phosphate solution (CPS) was prepared by first adding 41 ml HCl (1M) to 800 ml

MilliQ water, then dissolving 2.25 mM Na2HPO4·H2O, 4 mM CaCl2·2H2O, 0.14 M NaCl and

50 mM Tris, then bringing pH to 7.4 and the volume to 1 litre. The solutions were sterilized by

filtration with a 0.22 µm MillexGV, Millipore. Glass coverslips or tissue culture plates were

incubated with SBF (0.5 ml/well) for 3 days at room temperature. SBF was aspirated and CPS

(0.5 ml/well) was added for 1 day at room temperature, aspirated and 70% ethanol was added

and evaporated to sterilize the surface. The hydroxyapatite coated coverslips and plates washed

twice with distilled water and dried overnight at 37oC, were either used immediately or kept dry

at room temperature for up to 1 month. Prior to cell plating, coated plates were incubated with

FBS for 1 hour at 37oC.

Labeling hydroxyapatite with fluorescent dextran: The 25 nM solution of Alexa Fluor 488- or

Rhodamine B-conjugated dextran in distilled water was prepared and filtered. Sterile

hydroxyapatite-coated glass coverslips and tissue culture plates were incubated with 200 µl of

the dextran solution at 37oC for 48 h, washed twice with deionized distilled water (Milli-Q

water) and kept sterile and protected from light.

Characterization of fluorescent-hydroxyapatite layers: Low angle X-ray diffraction (XRD) data

was collected with Ni filtered CuKα radiation (λ = 1.54A) with 2 dimensional VANTEC area

detector at 40 kV and 40 mA using Bruker Discover D8 diffractometer. Step size of 0.02° was

used to measure from 10 to 50° over 2 frames with a count time of 400 s per frame. Scanning

electron microscopy (SEM) was performed using Hitachi S-4700. Scanning Electron Microscope

(Tokyo, Japan) with field emission gun (FEG) at an accelerating voltage of 10 kV and current of

10 amps. The sample stage was tilted to an angle of 35 degrees. Images were acquired using

secondary electron detector. For SEM the samples were coated Gold/Palladium using Hummer

VI sputter system with Argon gas under 80 millitorr vacuum, 10 milliamps: plasma discharge

current and voltage of 2.5 V as described previously [11].

Osteoclast cultures, transfer and identification: To generate osteoclasts, murine monocytic

cells RAW 264.7 (ATCC) were plated at 5,000 cells per cm2 in DMEM supplemented with

10% FBS, 1% penicillin-streptomycin, 1% sodium pyruvate and 50 ng/ml GST-RANKL as

described previously [18]; [19] on plastic tissue culture plates, on hydroxyapatite, or on

fluorescent hydroxyapatite. The medium was replaced every 2 days, until formation of

multinucleated osteoclasts (on day 5–7 day) was observed. To transfer mature osteoclasts onto

unlabeled or fluorescent hydroxyapatite, we first generated osteoclast-like cells from RAW

264.7 cells in 100 mm tissue culture plates. Then the osteoclast-like cells were purified, re-

suspended, and replated on the substrates at 20-200 cells/cm2 using the transfer protocol

described previously [11]. Tissue culture plates were returned to the incubator carefully and

maintained for 24-48 h. For fluorescence imaging, samples were fixed in 3.7%

paraformaldehyde stained with FITC-conjugated phalloidin for actin, and examined under

inverted fluorescence microscope (Nikon ECLIPSE TE 2000-U, USA).

LDH cytotoxicity assay: Cytotoxicity was assessed by the leakage of LDH into the culture

medium using [20] the LDH activity kit from Roche Applied Science (LDL50). Cells were

incubated on plastic surface or on the fluorescent hydroxyapatite for 24 h in 96-well plates

(Fisher), or on plastic surface with lysis solution, and the absorbance was measured using the

micro-plate readers (Infinite F200 TECAN).

Characterization of osteoclastic resorption: In the end of the culture period, the osteoclast-like

cells were washed twice with PBS and removed by incubation with 0.2% Triton-X-100 in

sodium chloride (1M) for 1-2 min at room temperature, then washed twice with distilled water

[10]. Unlabeled hydroxyapatite was examined using bright field microscopy. The fluorescent

hydroxyapatite was examined using the inverted fluorescence microscope (Nikon, ECLIPSE

TE 2000-U, USA). Three-five images per condition were collected and analyzed using ImageJ.

For manual quantification, each resorption pit was individually outlined and the areas were

quantified. For automatic quantitation the color images were first adjusted to obtain the best

brightness/contrast for the visual inspection of the resorption areas, and converted to binary

images, which were visually confirmed to correspond to the original images. The dark areas

were quantified. Using the training set of images, we determined that areas smaller than 100

pixel could not be reliably associated with resorption activity and therefore the particles below

100 pixels were removed by filtering from all the images. The areas of the remaining particles

were analyzed. The amount of the fluorescent dextran (Alexa Fluor 488 'green' and Rhodamine

B 'red') in the media was measured using the microplate reader (Infinite F200, TECAN). The

resorption volume was assessed using the confocal microscopy (Zeiss LSM510-META and

Olympus Fluoview, FV10-ASW1.7). A z-stack was obtained, and the volume was calculated as

sum of resorption areas in each level multiplied by the distance between the levels. To estimate

the resorption depth using fluorescence microscopy without confocal capabilities, images were

obtained at three levels for the same area of the substrates - superficial layer, at the glass level

and at the middle (3-4 µm depth).The volume of the resorption was calculated as sum of

resorption areas in each level multiplied by the estimated distance between the levels.

Statistical analysis: Data are expressed as mean ± standard error of the mean or standard

deviation, with n indicating the number of independent experiments or replicates recpectively.

Statistical difference was evaluated by Student t-test, or ANOVA followed by Tukey post-test

where appropriate and were considered to be significant at p < 0.05. Coefficient and significance

of correlation was assessed using Vassar Statistical Recourse page.

Results:

Characterization of the fluorescent hydroxyapatite: We used dextran conjugated to fluorescent

label Alexa Fluor 488 or Rhodamine B to improve the visualization of the thin film of

hydroxyapatite precipitated on the surface of glass coverslips or tissue culture plates as described

previously [11]. We have analyzed the degree of labeling of hydroxyapatite with fluorescent

dextran in the concentration range of 12.5-500 nM and found that sufficient fluorescence was

achieved even at low dextran levels (Fig. 1A). The labeled substrates were then incubated in

phenol red-free media for 48 h, and media fluorescence was measured. No significant

fluorescence was observed in media after incubation with hydroxyapatite labeled with 12.5-500

nM dextran, indicating stability of labeling. Low angle XRD patterns indicated the layer was

phase pure hydroxyapatite (Fig.1B). SEM demonstrated homogeneous and complete coating

prior to cell culture (Fig.1C) and after incubation with media without cells (Fig.1D). After

culture with RAW 264.7 cells differentiated with 50ng/ml into osteoclast-like cells for 5 days

characteristic resorption lacunae were evident (Fig.1E).

Fluorescent hydroxyapatite is not toxic and supports osteoclastogenesis: To examine if the

fluorescent hydroxyapatite have toxic effects on cultured cells, LDH cytotoxicity assay was

performed. RAW 264.7 cells were incubated on the fluorescent hydroxyapatite or the tissue

culture plates (control) for 24 h, and LDH release to the media was assessed. As a positive

control, cells were incubated with lysis solution. No cell toxicity of the fluorescent

hydroxyapatite was found (Fig.2A). To examine if fluorescent hydroxyapatite supports

osteoclastogenesis, were plated and cultured either untreated or treated with RANKL (50 ng/ml)

for 5 days to induce osteoclastogenesis. The fluorescent hydroxyapatite was stable after the

incubation with culture media alone (Fig.2B) or RAW 264.7 cells (Fig.2C). Multinucleated

osteoclast-like cells successfully formed on the fluorescent hydroxyapatite and were surrounded

by resorption zone free from coating (Fig.2D, white arrowhead). Both undifferentiated

RAW264.7 and osteoclasts internalized fluorescent dextran (Fig.2C,D yellow arrows). The

osteoclast-like cells formed on fluorescent hydroxyapatite exhibited actin ring normally

associated with resorption [21] (Fig.2E). We directly examined osteoclast resorptive activity on

the parallel samples of hydroxyapatite either unlabeled or labeled with fluorescent dextran, and

observed a significant correlation between the average resorption areas resorbed by osteoclasts

incubated on labeled and unlabeled hydroxyapatite, R2= 0.99, p  <0.0001 (Fig.2F).

Automated quantification of osteoclastic resorption: Fluorescent labeling of the resorbable

substrates facilitated visualizing the substrates-free-areas resulted from osteoclastic resorption

(Fig.3A). Mature osteoclasts were transferred onto fluorescent hydroxyapatite as described

previously [11] and incubated with RANKL (0, 12.5, 25, 50 ng/ml), or with alendronate (100

µM) for 24-48 h. The cells were removed the images were obtained using the fluorescence

microscopy and quantified with ImageJ either manually by individually outlining each resorption

pit, or automatically by first adjusting the color image to obtain the best contrast, then converting

to binary image, then filtering the particles below 100 pixels and finally analyzing the size of the

remaining particles. We have found that the total resorption area increased significantly with the

increase in RANKL concentration (Fig.3B, left), and decreased significantly with osteoclast

inhibitor alendronate (Fig.3B, right). The correlation between the values obtained using manual

and automated quantification was significant at R2 = 0.82 (Fig.3C).

To examine if the amount of the fluorescent dextran in the media can be used as an

indicator of resorption, we measured the fluorescence intensity of the conditioned media

collected after 24 h of incubation of fluorescent hydroxyapatite with osteoclasts treated with

RANKL (0, 12.5, 25, 50 ng/ml) or alendronate (100 µM), or with untreated RAW 264.7 cells, or

with media only. No correlation between the resorptive activity in the sample and the amount of

dextran in the media was observed (Fig.4), likely indicating that internalization of dextran

interferes with its release into the incubation media.

Fluorescent hydroxyapatite facilitates measuring resorption volume: To obtain more detailed

3-dimentional information of the resorptive activity, we examined the osteoclastic resorption pits

formed on the fluorescent hydroxyapatite using the confocal microscopy. The z-stack was

obtained at 1 µm interval (Fig.5A). To facilitate visualization of the resorption depth, a color-

coded height stack with the color scale indicating the actual height was constructed (Fig. 5B).

The resorption pits of different depths were evident: while in some areas the fluorescent

hydroxyapatite was removed to the glass level (Fig. 5B, white outlined black arrow), in other

areas the resorption pit was more shallow (Fig.5B, white arrow). We have estimated the

resorption volume, and have found that it although there was a trend of increased volume with

increasing RANKL concentration it was not significant (P<0.05, ANOVA). (Fig. 5C).

We next developed a simplified method to determine the osteoclast-like cells’ resorption

volume using the fluorescent hydroxyapatite. Using fluorescence microscope without confocal

capability, the images were obtained focusing as the surface of the substrate, at the surface of the

glass and at the middle position (Fig.6A-C). Similar to confocal images, we were able to

distinguish between the resorption pits of different depth (Fig. 6D, white arrow and white

outlined black arrow). We have found significant correlation between the resorption volume of

the same samples estimated using fluorescence and confocal microscopy, R² = 0.93 (Fig.6E).

Discussion:

In this study, we developed an assay for an automated analysis of the osteoclastic resorption.

Fluorescent dextran labeling of the resorbable substrates facilitated visualizing the substrates-free

areas resulted from osteoclast resorption. We have shown that the data obtained using an

automated method to estimate the osteoclastic resorption significantly correlated with the manual

method of the resorption estimation. In addition, we demonstrated that fluorescently labeled

hydroxyapatite allows obtaining additional information regarding the resorption volume and

developed a simplified method to analyze resorption volume.

We have established that fluorescent labeling with dextran did not affect the basic

properties of the thin hydroxyapatite layers characterized for their stability and lack of non-

specific dissolution in the previous studies [10, 11]. In a similar approach, Miyazaki and

colleagues examined the fluorescent labeling of a carbonated calcium phosphate with three

different fluorescent polyanions: fluoresceinamine-labeled chondroitin polysulfate, Hoechst

33258-labeled deoxyribonucleic acid and calcein [13]. While demonstrating very promising

capabilities for quick assessment of the resorption index with fluoresceinamine-labeled

chondroitin polysulfate or Hoechst 33258-labeled deoxyribonucleic acid calcium phosphate, this

study had certain drawbacks, first in the evidence of a non-specific dissolution due to higher

solubility of the carbonated calcium phosphate, and second in a nonlinear correlation of

fluorophore release with the resorption index, suggesting certain lack of stability in fluorophore-

calcium phosphate binding. In contrast, the protocol used in this study allows for the

development of fluorescently labeled hydroxyapatite coatings with superior properties.

We have found that labeling the resorbable substrates with the fluorescent dextran was

not toxic to the cells and did not affect osteoclast differentiation or resorptive activity. Osteoclast

exhibited normal morphology and formed the actin ring structure that is associated with

resorption. Although osteoclasts and their precursors were to internalize dextran, their resorptive

activity was not affected by dextran internalization, as evident by a significant correlation and

similar range of the resorption areas resorbed by osteoclasts on labeled and unlabeled

hydroxyapatite, R2= 0.99, p <0.0001. These data confirm that this layer is suitable for analyzing

osteoclast resorptive activity.

We have used labeling of hydroxyapatite with the fluorescent dextran, which has the

capacity to bind effectively to hydroxyapatite [15]; [16], to facilitated visualizing and automate

the analysis of the substrates-free areas resulted from osteoclast resorption. In general, the

automated evaluation of the substrate-free areas on calcium phosphate substrates is complicated

by the crystallinity of the substrate that results in uneven illumination of the surface and poses

difficulties for automated thresholding. The automated quantification is commonly performed

without the demonstration of substrate appearance or validation of the quantification protocol

[13], and the when the substrates are shown it is often clear that automated quantification

represents only a rough estimate of osteoclastic activity [22]. The stability of hydroxyapatite

substrate and of dextran binding allowed for a high correlation between the manual and

automated quantification demonstrated in this study. Another approach to estimate a rapid

resorption index in live cultures, is to measure the release of fluorescent label into the media

[13]. However, we have found that level of the fluorescent dextran in the incubation media did

not correlate with the osteoclast activity likely due to it endocytosis by monocytes and

osteoclasts [23]; [17]. Similarly, it was previously demonstrated that when calcium phosphate

was labeled with calcein no correlation between the amount of fluorescence in the supernatant

and resorptive activity was observed [13]. Thus, labeling with different fluorescent indicators can

fine-tune the presented methodology to the specific experimental needs.

Measurement of the resorption volume represents a true measure of osteoclast resorptive

activity, however it is difficult to achieve in practice. Numerous studies proposed the use of

scanning electron microscopy [24], [25]; [26], confocal microscopy [27] and vertical scanning

profilometry [28], however most techniques are either time consuming or require specialized

instrumentation. In this study, we demonstrated that the measurements of the resorption volume

per pit can be simplified on the fluorescently labeled hydroxyapatite substrates, and developed a

convenient protocol to estimate the resorption volume using the fluorescence microscopy

without the confocal capability.

Thus, the method developed in this study allows simplifying and automating the analysis

of osteoclast resorptive activity, and provides additional measures of resorption volume that can

facilitate the development of novel osteoclast-targeting therapies.

Acknowledgment:

We thank Dr. MF Manolson (University of Toronto) for providing reagents used in this study.

This study was supported by the Canadian Institutes for Health Research (SVK) and Natural

Sciences and Engineering Research Council of Canada (JEB). SM was supported by the Faculty

of Dentistry, McGill University. JEB held Canada Research Chair in Osteoinductive

Biomaterials and SVK holds Canada Research Chair in Osteoclast Biology.

References:

[1] Hughes DE, Dai A, Tiffee JC, Li HH, Mundy GR, Boyce BF. Estrogen promotes apoptosis of

murine osteoclasts mediated by TGF-beta. Nat Med 1996;2:1132-6.

[2] Ritchlin CT, Haas-Smith SA, Li P, Hicks DG, Schwarz EM. Mechanisms of TNF-alpha- and

RANKL-mediated osteoclastogenesis and bone resorption in psoriatic arthritis. J Clin

Invest 2003;111:821-31.

[3] Amft N, Curnow SJ, Scheel-Toellner D, Devadas A, Oates J, Crocker J, et al. Ectopic

expression of the B cell-attracting chemokine BCA-1 (CXCL13) on endothelial cells and

within lymphoid follicles contributes to the establishment of germinal center-like

structures in Sjogren's syndrome. Arthritis and Rheumatism 2001;44:2633-41.

[4] Burman A, Haworth O, Hardie DL, Amft EN, Siewert C, Jackson DG, et al. A chemokine-

dependent stromal induction mechanism for aberrant lymphocyte accumulation and

compromised lymphatic return in rheumatoid arthritis. Journal of Immunology

2005;174:1693-700.

[5] Page G, Lebecque S, Miossec P. Anatomic localization of immature and mature dendritic

cells in an ectopic lymphoid organ: correlation with selective chemokine expression in

rheumatoid synovium. J Immunol 2002;168:5333-41.

[6] Bartold PM, Cantley MD, Haynes DR. Mechanisms and control of pathologic bone loss in

periodontitis. Periodontology 2000 2010;53:55-69.

[7] Halvorson KG, Sevcik MA, Ghilardi JR, Rosol TJ, Mantyh PW. Similarities and differences

in tumor growth, skeletal remodeling and pain in an osteolytic and osteoblastic model of

bone cancer. Clin J Pain 2006;22:587-600.

[8] Hussein O, Tiedemann K, Komarova SV. Breast cancer cells inhibit spontaneous and

bisphosphonate-induced osteoclast apoptosis. Bone 2011;48:202-11.

[9] Iqbal J, Sun L, Mechanick JI, Zaidi M. Anti-cancer actions of denosumab. Curr Osteoporos

Rep 2011;9:173-6.

[10] Patntirapong S, Habibovic P, Hauschka PV. Effects of soluble cobalt and cobalt

incorporated into calcium phosphate layers on osteoclast differentiation and activation.

Biomaterials 2009;30:548-55.

[11] Maria SM, Prukner C, Sheikh Z, Mueller F, Barralet JE, Komarova SV. Reproducible

quantification of osteoclastic activity: Characterization of a biomimetic calcium

phosphate assay. J Biomed Mater Res B Appl Biomater 2013.

[12] Hale LV, Ma YF, Santerre RF. Semi-quantitative fluorescence analysis of calcein binding as

a measurement of in vitro mineralization. Calcified Tissue International 2000;67:80-4.

[13] Miyazaki T, Miyauchi S, Anada T, Imaizumi H, Suzuki O. Evaluation of osteoclastic

resorption activity using calcium phosphate coating combined with labeled polyanion. J

Anal Biochem 2011;410:7-12.

[14] Hoorfar M, Kurz MA, Policova Z, Hair ML, Neumann AW. Do polysaccharides such as

dextran and their monomers really increase the surface tension of water? Langmuir : the

ACS journal of surfaces and colloids 2006;22:52-6.

[15] Fricain JC, Schlaubitz S, Le Visage C, Arnault I, Derkaoui SM, Siadous R, et al. A nano-

hydroxyapatite--pullulan/dextran polysaccharide composite macroporous material for

bone tissue engineering. Biomaterials 2013;34:2947-59.

[16] Niwa M, Li W, Sato T, Daisaku T, Aoki H. The adsorptive properties of hydroxyapatite to

albumin, dextran and lipids. Biomed Mater Eng 1999;9:163-9.

[17] Stenbeck G, Horton MA. Endocytic trafficking in actively resorbing osteoclasts. J Cell Sci

2004;117:827-36.

[18] Akchurin T, Aissiou T, Kemeny N, Prosk E, Nigam N, Komarova SV. Complex Dynamics

of Osteoclast Formation and Death in Long-Term Cultures. PLoS ONE 2008;3.

[19] Le Nihouannen D, Hacking SA, Gbureck U, Komarova SV, Barralet JE. The use of

RANKL-coated brushite cement to stimulate bone remodelling. Biomaterials

2008;29:3253-9.

[20] Fotakis G, Timbrell JA. In vitro cytotoxicity assays: comparison of LDH, neutral red, MTT

and protein assay in hepatoma cell lines following exposure to cadmium chloride.

Toxicol Lett 2006;160:171-7.

[21] Lakkakorpi PT, Väänänen KH. Kinetics of the osteoclast cytoskeleton during the resorption

cycle in vitro. Journal of Bone and Mineral Research 1991;6:817-26.

[22] Kartner N, Yao Y, Li K, Crasto GJ, Datti A, Manolson MF. Inhibition of Osteoclast Bone

Resorption by Disrupting Vacuolar H+-ATPase a3-B2 Subunit Interaction. Journal of

Biological Chemistry 2010;285:37476-90.

[23] Seto H, Kawakita H, Ohto K, Harada H, Inoue K. Membrane porosity control with dextran

produced by immobilized dextransucrase. Journal of Chemical Technology and

Biotechnology 2007;82:248-52.

[24] Boyde A, Ali NN, Jones SJ. Resorption of dentine by isolated osteoclasts in vitro. British

dental journal 1984;156:216-20.

[25] Fuller K, Thong JT, Breton BC, Chambers TJ. Automated three-dimensional

characterization of osteoclastic resorption lacunae by stereoscopic scanning electron

microscopy. Journal of bone and mineral research : the official journal of the American

Society for Bone and Mineral Research 1994;9:17-23.

[26] Grimandi G, Soueidan A, Anjrini AA, Badran Z, Pilet P, Daculsi G, et al. Quantitative and

reliable in vitro method combining scanning electron microscopy and image analysis for

the screening of osteotropic modulators. Microscopy research and technique

2006;69:606-12.

[27] Yamada Y, Ito A, Sakane M, Miyakawa S, Uemura T. Laser microscopic measurement of

osteoclastic resorption pits on biomaterials. Materials Science and Engineering: C

2007;27:762-6.

[28] Pascaretti-Grizon F, Mabilleau G, Baslé M, Chappard D. Measurement by vertical scanning

profilometry of resorption volume and lacunae depth caused by osteoclasts on dentine

slices. Journal of microscopy 2011;241:147-52.

Figure legends:

Figure 1: Characterization of the fluorescent hydroxyapatite. A: Representative images of

hydroxyapatite surfaces labeled with 500, 125, 50, and 12.5 nM fluorescent dextran. Same

exposure was used for all the images. B: XRD patterns of the fluorescent dextran-labeled-

calcium phosphate layer, crosses indicate standard hydroxyapatite peaks. C-E: SEM images of

fluorescent hydroxyapatite before and after culture, top and bottom rows are different

magnifications of the same image. The fluorescent hydroxyapatite untreated; dry (C), the

incubated fluorescent hydroxyapatite for 24 h with media with no cells (D), and with osteoclasts

(E), then cells were removed, and resorption pits formed underneath osteoclasts were imaged.

Figure 2: Fluorescent dextran labeling does not affect osteoclast response to hydroxyapatite. A:

RAW 264.7 cells were incubated on tissue culture plastic or fluorescent hydroxyapatite for 24 h,

and LDH levels in the media were assessed. As a positive control, lysis solution was added to

RAW 264.7 incubated on plastic. Data are means ± SD, n = 3 replicates, no significant difference

between plastic and fluorescent hydroxyapatite. B-E: Fluorescent hydroxyapatite (dextran

Rhodamine B 'red') was incubated for 24 h with media with no cells (B), with RAW 264.7 cells

(C) or with osteoclasts (D, E). Yellow arrows point at RAW 264.7 cells and an osteoclast

containing dextran intracellularly; white arrowhead points at a resorption area free from the

substrates next to an osteoclast. E: Osteoclast cultures on fluorescent hydroxyapatite (red, left)

were fixed and actin was labeled with FITC-conjugated phalloidin (green, middle). The overlay

(right) demonstrates an osteoclast with an actin ring that is surrounded by a resorption area free

of fluorescent substrates. F: Differentiated osteoclasts were transferred to the parallel samples of

unlabeled and fluorescent hydroxyapatite and incubated for 24-48 h with RANKL (0, 12.5, 25 or

50 ng/ml). Correlation of the resorption areas resorbed by osteoclasts on labeled and unlabeled

hydroxyapatite, R2= 0.99, p <0.0001.

Figure 3: Automated quantification of the osteoclast resorption on the fluorescent

hydroxyapatite. The fluorescent hydroxyapatite (dextran alexa Fluor 488) was incubated with

osteoclasts with media containing RANKL (0, 12.5, 25, 50 ng/ml), or media with alendronate

(100 µM) and no RANKL, then cells were removed and the substrates were imaged using

fluorescence microscopy. A: Representative images of fluorescent hydroxyapatite that was

incubated with osteoclasts without RANKL (left), with RANKL (middle), or with osteoclast

inhibitor, alendronate (100 µM) (right) for 24-48 h. B: Automated quantification of the

resorption areas per 111210 µm2. Data are mean ± SE, n = 4, *p < 0.0001 indicate statistical

significance assessed using ANOVA for the RANKL dependence and t-test for the effect of

alendronate. C: The resorption areas on fluorescent hydroxyapatite were quantified manually and

automatically and then correlation was assessed. The manual and automated quantification

methods of resorption exhibited significant correlation R2= 0.82, p <0.0001.

Figure 4: Media release of fluorescent dextran cannot be used as an indicator for the osteoclastic

resorption. Osteoclasts were incubated on fluorescent hydroxyapatite for 24 h with RANKL (0,

12.5, 25, 50 ng/ml, labeled as OC, R12.5, R25 and R50 respectively), or with alendronate

(Alend). As control, fluorescent hydroxyapatite was incubated with media only without cells (M)

or with RAW 264.7 cells in media without RANKL. Conditioned media was collected and

fluorescence was measured (black bars). In a selected experiment, cells were lysed, intracellular

fluorescence was measured and added to the measurements in conditioned medium (white bars).

Data are means ± SE, normalized to media only n = 4, no significant difference.

Figure 5: Assessment of the osteoclast resorption volume using the fluorescent hydroxyapatite.

Osteoclasts were incubated with RANKL (0, 12.5, 25 or 50 ng/ml) for 24-48 h on the fluorescent

hydroxyapatite (alexa-488, green), cells were removed and then images were taken using

confocal microscopy. A,1-9: The z-stack was obtained at 1 µm interval. The depth of the calcium

phosphate layers was 9 ± 3 µm. B: A color-coded height stack with the color scale indicating the

actual height above the bottom slice. White outlined black arrow indicates the resorption pit that

almost reached the glass level, white arrow indicates the resorption pit that did not reach the

glass level. C: The resorption volume increased with the increase in RANKL concentration. Box-

plots indicate the minimal values and 90th percentile (whiskers), the 25th and 75th percentiles

(the limits of the box), and the median values (the line within the box), n = 10-24.

Figure 6: A simplified method to determine the osteoclast resorption volume with the

fluorescent hydroxyapatite. Osteoclasts were incubated in media with RANKL (0, 12.5, 25, 50

ng/ml) or with alendronate (100 µM) for 24-48 h, the cells were removed and the substrates were

imaged at three levels: surface layer, 4 µm and 8 µm in depth and pseudo-colored red, green and

blue respectively. A-C: representative images of the substrate appearance following incubation

of osteoclasts with RANKL 50 ng/ml at 3 different levels: at the surface (A), 4 µm deep (B) and

8 µm deep (C). D: merge of A, B and C. White outlined black arrow indicates the resorption pit

that reached the glass level, white arrow indicates the resorption pit that did not reach the glass

level. E: The average resorption volume in the same samples was estimated using fluorescence

and confocal microscopy. Data are means ± SE, n = 4, the correlation R2= 0.93, *p <0.001.