Original Article

Polarization of osteoclasts on dental implant materials is similar to thatobserved on bone

Takahiro Nakayama a,b, Gnanasagar J. Thirukonda b, Sakae Nagasawa b, Ichiro Kawahara b,c,Nobuyuki Udagawa d, Kimitoshi Yagami c, Makoto Kawatani e, Hiroyuki Osada e,f,Yutaka Doi g, Nobuo Yoshinari a, Naoyuki Takahashi b,n

a Department of Periodontology, School of Dentistry, Matsumoto Dental University, Nagano 399-0781, Japanb Institute for Oral Science, Matsumoto Dental University, Nagano 399-0781, Japanc Department of Social Dentistry, School of Dentistry, Matsumoto Dental University, Nagano 399-0781, Japand Department of Biochemistry, School of Dentistry, Matsumoto Dental University, Nagano 399-0781, Japane Antibiotics Laboratory, RIKEN, Saitama 351-0198, Japanf Chemical Biology Research Group, RIKEN CSRS, Saitama 351-0198, Japang Departments of Dental Materials Science, Division of Oral Functional Science and Rehabilitation, Asahi University School of Dentistry, Gifu 501-0296, Japan

a r t i c l e i n f o

Article history:Received 8 April 2014Received in revised form3 June 2014Accepted 4 June 2014Available online 13 August 2014

Keywords:TitaniumHydroxyapatiteActin ringTRAP-markReveromycin A

a b s t r a c t

Objective: Polarized osteoclasts form sealing zones, (also called clear zones) detectable as actin rings, andruffled borders to resorb bone. They secrete protons and catabolic enzymes, including tartrate-resistantacid phosphatase (TRAP), through the ruffled borders. We previously reported that polarized osteoclastsdevelop areas of TRAP activity (TRAP-marks) when cultured on dentin slices [11]. In this study, weexamined how osteoclasts recognize dental implant materials.Methods: Osteoclasts obtained from murine co-cultures were cultured on implant materials such astitanium (Ti), alumina, zirconia, and sintered hydroxyapatite (sHA), in addition to dentin. Osteoclastswere also treated with reveromycin A (RM-A), which specifically acts on polarized osteoclasts andinduces apoptosis. Polarization of osteoclasts cultured on implant materials was evaluated by measuringactin rings, TRAP-marks, and reveromycin A-induced apoptosis.Results: Osteoclasts formed actin rings on all substrates examined. The formation of actin rings on Ti byosteoclasts was inhibited by the GRGDS peptide, but not by the GRGES peptide, suggesting an integrin-mediated polarization of osteoclasts on Ti. Calcitonin, an inhibitory hormone of osteoclast function,disrupted the actin rings that were preformed on Ti and sHA. Osteoclasts put TRAP-marks on sHA anddentin and formed resorption pits on dentin, but failed to form resorption pits on sHA. RM-A inducedapoptosis in osteoclasts cultured on Ti and sHA; this was suppressed by calcitonin.Conclusions: These results demonstrate that osteoclasts are able to polarize on dental implant materialssimilar to the polarization observed on bone.

& 2014 Japanese Association for Oral Biology. Published by Elsevier B.V.

1. Introduction

Osteoclasts are bone-resorbing cells that have a role in boneremodeling [1,2]. Osteoclasts have been shown to create polarizedcytoplasmic organizations such as clear zones and ruffled bordersto resorb bone [3,4]. The recognition of extracellular matrix proteinscontaining the Arg-Gly-Asp-(RGD) sequence by integrins is theinitial step in osteoclast polarization. Osteoclasts express a largenumber of αvβ3 vitronectin receptors [5,6]. The clear zone serves forthe attachment of osteoclasts to bone surfaces and for the isolation

of resorption areas under ruffled borders from the surroundings.The clear zone consists of many distinct close contact points calledpodosomes and is observed as a ringed structure of F-actin dots(actin ring) [3,4]. The formation of actin rings in osteoclastsprecedes the start of bone resorption. A previous study demon-strated that the disruption of actin rings suppressed osteoclastfunction [7].

The resorption area under the ruffled border is acidic, whichfavors dissolution of bone mineral. Vacuolar type Hþ-ATPase(V-ATPase), which is specifically localized on the ruffled bordermembrane, has been shown to play a role in the secretion of protonsinto resorption lacunae [8]. Polarized osteoclasts also secrete cata-bolic enzymes including tartrate-resistant acid phosphatase (TRAP)through the ruffled borders in order to degrade the organic matrix

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/job

Journal of Oral Biosciences

http://dx.doi.org/10.1016/j.job.2014.06.0051349-0079/& 2014 Japanese Association for Oral Biology. Published by Elsevier B.V.

n Corresponding author. Tel.: þ81 263 51 2135; fax: þ81 263 51 2223.E-mail address: takahashinao@po.mdu.ac.jp (N. Takahashi).

Journal of Oral Biosciences 56 (2014) 136–142

Open access under CC BY-NC-ND license.

Open access under CC BY-NC-ND license.

of bone [9,10]. We previously investigated the relationship betweenpositions of actin rings and resorption pits in osteoclasts, usingdentin slices and osteoclasts obtained form murine co-cultures [11].We found that pit-forming osteoclasts put TRAP activity (namedTRAP-marks) on dentin slices. The position of the actin ringscorresponded to that of TRAP-marks. Thus, the detection of TRAP-marks on dentin slices is a simple method to confirm the location ofpolarized osteoclasts.

Calcitonin (CT) is a peptide hormone that inhibits the function ofosteoclasts, which generally express a large number of CT receptors[12,13]. CT has been shown to suppress bone-resorbing activity ofosteoclasts by disrupting the formation of actin rings [14,15].Reveromycin A (RM-A), which is produced by the genus Strepto-myces, is an osteoclast-targeting antibiotic [16,17]. Woo et al. [17]demonstrated that RM-A specifically induced apoptosis in function-ing osteoclasts that secreted protons, but not in non-functionalosteoclasts. RM-A can be incorporated into osteoclasts under acidicconditions. RM-A-induced apoptosis can be inhibited by concanamy-cin A, an inhibitor of V-ATPase, or through the disruption of actinrings. Therefore, CT is able to suppress RM-A-induced apoptosis inosteoclasts by disrupting the actin rings. Osteoprotegerin is a solubledecoy receptor for receptor activator of nuclear factor kB (RANKL), acritical inducer of osteoclast differentiation and function [18]. Previousstudies have reported that bone resorptionwas markedly upregulatedin osteoprotegerin-deficient mice [19,20]. Furthermore, in vivoadministration of RM-A to osteoprotegerin-deficient mice suppressedtooth movement induced by orthodontic forces [21].

Advances in biomaterial science and cell biology are intimatelyintertwined. Elucidating the recognition mechanism for biomaterialsby host cells is an important issue for the clinical success of ortho-pedics and dental implants. Pertinent questions include how osteo-clasts recognize implant materials, whether osteoclasts try to resorbimplant materials, and whether osteoclasts can resorb hydroxyapatite(HA) coated on titanium (Ti) implants. These questions are of impor-tance, as the success of dental implants largely depends on such vitalreactions in patients. To answer these questions, we examined howosteoclasts recognized implant materials by evaluating actin rings,TRAP-marks, and RM-A-induced apoptosis in osteoclasts. The resultsobtained demonstrated that osteoclasts have the ability to polarize onimplant materials in a manner similar to that observed on bone.

2. Materials and methods

2.1. Implant materials

Grade 2 pure Ti was purchased from Daido Steel (Nagoya, Japan).Alumina sintered compact (alumina) and yttrium-stabilized zirconiasintered compact (zirconia) were obtained from Taimei Chemicals(Shimoina, Japan) and Kikusui Chemical Industries (Nagoya, Japan),respectively. These material plates (1-mm thick) were polished usingpastes that were 1 μm in diameter and cut into small pieces(5 mm�5 mm). HA was synthesized as described previously [22].Precipitates of HA were dry-pressed and further compacted isostati-cally at 200 MPa. sHA was produced by heating at 1200 1C for 2 h.Approximately 1-mm thick disks were dry cut from sHA and cut intosmall pieces (3-mm squares). The surface of the sHA slices waspolished with pastes (1 μm in diameter). Dentin slices (4 mm�4 mm,0.2-mm thick) were made from an ivory block [23].

2.2. Materials for cell culture

Collagenase, calcitriol, and prostaglandin E2 were purchasedfrom Wako Pure Chemical (Osaka, Japan). A type I collagen gelmatrix was obtained from Nitta Gelatin (Osaka, Japan). Elcatonin,a synthetic analog of eel CT, was kindly provided by Asahi Kasei

(Tokyo, Japan). The Gly-Arg-Gly-Asp-Ser (GRGDS) peptide and Gly-Arg-Gly-Glu-Ser (GRGES) peptide were obtained from the PeptideInstitute (Osaka, Japan) and Peptide International (Louisville, KY,USA), respectively. Reveromycin A (RM-A) was isolated and pur-ified from the fermentation broth of Streptomyces reveromyceticusSN-593 strain as described previously [16]. Rhodamine-conjugatedphalloidin was obtained from Molecular Probes (Eugene, OR, USA).

2.3. Preparation of osteoclasts

Six-week-old male and newborn ddY mice were obtained fromJapan SLC (Shizuoka, Japan). Primary osteoblasts were obtained fromcalvariae of newborn mice, as described previously [23]. Bonemarrow cells were obtained from the tibiae of adult mice. Osteoclastswere prepared from co-cultures of osteoblasts and bonemarrow cells[23]. Primary osteoblasts (1�106 cells/dish) and bone marrow cells(1�107 cells/dish) were co-cultured in α-minimal essential medium(α-MEM) containing 10% fetal bovine serum (FBS, JRH Biosciences,Lenexa, KS, USA), 10�8 M calcitriol, and 10�6 M prostaglandin E2 in10-cm dishes pre-coated with 0.2% type I collagen gel matrix. Afterbeing cultured for 7 days, the cells were treated with 0.2% bacterialcollagenase. Cells recovered from the dish were suspended in 10 mLof α-Modified Eagle's Medium (α-MEM) containing 10% FBS and usedas osteoclast preparations [23].

2.4. Identification of actin rings

Osteoclast preparations were cultured for 1 h on slices of Ti,alumina, zirconia, sHA, and dentin in 48-well culture plates(0.4 ml/well). The slices were then transferred into new 48-wellculture plates and cultured in α-MEM containing 10% FBS. Forcertain experiments, 10�8 M of CT was added to some cultures.The GRGDS and GRGES peptides were suspended in phosphate-buffered saline (PBS). Osteoclasts were placed on Ti in thepresence of 10% FBS with or without 1 mM GRGDS and GRGESpeptides. After being cultured for 3 or 7 h, cells were processed foractin staining [14]. The number of osteoclasts having actin ringswas counted.

2.5. TRAP staining

Osteoclasts on dentin, Ti, and sHA were fixed and processed forTRAP staining [23]. Cells on the slices were thoroughly removedusing cotton swabs for TRAP-mark staining [11]. Slices were alsoincubated with a TRAP staining solution for 30 min to detectTRAP-marks.

2.6. Scanning electron microscopy

Cells were removed from the slices by using cotton swabs andthe fine structure of lacunae on the surface of sHA and dentin wereobserved. The slices were fixed in 2.5% glutaraldehyde in phos-phate buffer (pH 7.4), washed, air-dried, and coated with carbonby using a vacuum evaporator (JEOL JEE-420, Tokyo Japan).Surfaces of the samples were observed using SEM (JEOL JSM-6360LA, Tokyo) with accelerating voltages of 30 kV. The workingdistance was 10 mm and samples were imaged in the SE mode.

2.7. RM-A-induced apoptosis in osteoclasts

48-well culture plates were pre-coated with 150 μL of 0.2%type I collagen gel matrix. Osteoclasts were cultured for 12 h onplastic, collagen gel, Ti, and sHA in the presence or absence of10�6 M RM-A in α-MEM containing 10% FBS. Some plates werealso treated with 10�8 M CT. Osteoclasts were then stained for

T. Nakayama et al. / Journal of Oral Biosciences 56 (2014) 136–142 137

TRAP. TRAP-positive cells in which the longitudinal axis of the cellbody was more than 100 μm were counted as osteoclasts.

2.8. Statistical analysis

All experiments were performed at least three times and theresults are expressed as the meansþSD for three or four cultures.Student's t-test was used to assess statistical differences.

3. Results

3.1. Actin ring formation by osteoclasts cultured on dental materials.

We first examined the formation of actin rings by osteoclastscultured on various materials such as Ti, alumina, zirconia, and sHA.Osteoclasts formed actin rings on all the materials examined. Theactin rings were large on Ti, alumina, zirconia, and sHA, and smallon dentin. This suggests that osteoclasts may recognize someproteins that are projected on the surface of dentin. We previouslyreported that the formation of actin rings was inhibited by RGDpeptides, but not by RGE peptides [6]. Therefore, we subsequentlyexamined actin ring formation by osteoclasts cultured on Ti (Fig. 1B).Osteoclasts suspended in α-MEM containing 10% FBS were cultured

for 3 or 7 h on Ti, in the presence and absence of the GRGDS peptideor GRGES peptide. Osteoclasts formed actin rings on Ti within 3 h.The formation of actin rings by osteoclasts after 3 h was completelyinhibited by the addition of the GRGDS peptide, but not GRGESpeptide. However, the inhibition of actin ring-formation by the RGDSpeptide slightly recovered within 7 h. These results suggested thatactin ring formation by osteoclasts on Ti may be mediated byintegrin-induced signaling and that binding of the GRGDS peptideto integrins expressed by osteoclasts can be gradually replaced withserum-derived RGD containing peptides on Ti.

3.2. Effects of CT on actin rings in osteoclasts cultured on Ti and sHA

We previously reported that CT disrupted actin rings formed byosteoclasts cultured on plastic as well as on dentin [14]. Therefore,we investigated the effect of CT on actin rings in osteoclastscultured on Ti and sHA (Fig. 2). Osteoclasts cultured on Ti for 8 hformed actin rings (Fig. 2A). Treating these cultures with CTdisrupted most of the preformed actin rings within 2 h. Osteo-clasts cultured on sHA for 8 h also formed actin rings, which weredisrupted by the addition of CT (Fig. 2B). Thus, the characteristicsof actin rings formed by osteoclasts on Ti and sHA were similar tothose formed by osteoclasts on dentin, in terms of their sensitivityto CT.

Control

3h

7h

GRGDS GRGES

DentinAlumina sHATi Zirconia

200 µm

µm

0

20

40

60

80

100

Control GRGDS GRGES

Ost

eocl

asts

hav

ing

actin

ring

s/un

it

: 7 h

: 3 h

**

*

100

Fig. 1. Actin ring formation by osteoclasts on various implant materials: (A) actin ring formation by osteoclasts. Osteoclast preparations were placed on titanium (Ti), alumina,zirconia, sintered hydroxyapatite (sHA), and dentin in the presence of 10% FBS. After being cultured for 12 h, cells were fixed and stained for F-actin with rhodamine-conjugatedphalloidin. (B) Effects of the GRGDS peptide on actin ring formation in osteoclasts on Ti plates. Osteoclast preparations were placed on Ti plates in the presence of 10% FBStogether with the synthetic GRGDS peptide or GRGES peptide at 1 mM concentration. Osteoclasts were cultured for 3 or 7 h and cells were fixed and stained for F-actin.Osteoclasts having actin rings with a diameter of more than 100 μm were counted. The results are expressed as the means7S.D. of three cultures. *: po0.01.

T. Nakayama et al. / Journal of Oral Biosciences 56 (2014) 136–142138

3.3. Detection of TRAP-marks on Ti, sHA, and dentin slices

We previously reported that polarized osteoclasts put TRAP-marks on dentin [11]. Therefore, we examined TRAP-marks on Ti,sHA, and dentin, after osteoclasts were cultured for 48 h (Fig. 3).TRAP-positive osteoclasts were observed on these materials (Fig. 3A upper panels). After the removal of cells by using cotton swabs,the materials were subjected to TRAP staining. TRAP-marks wereobserved on dentin and sHA, but not on Ti (Fig. 3A lower panels).The size of the TRAP-marks on sHA was larger than that of theTRAP-marks on dentin. However, the staining intensity of TRAP-marks was greater on dentin than on sHA. TRAP-marks on Tiappeared to have been removed by the cotton swab treatment.

3.4. Detection of resorption pits on sHA and dentin slices

Using scanning electron microscopy, we further investigatedwhether osteoclasts formed resorption pits on sHA as well as ondentin (Fig. 3B). Osteoclasts were cultured on dentin and sHAslices. After being cultured for 48 h, cells were removed from theslices. The slices were then processed for scanning electronmicroscopy (Fig. 3B upper panels). A large number of resorptionpits were detected on dentin. However, osteoclasts failed to makeresorption pits on sHA (Fig. 3B lower panels), consistent with thefindings of previous studies [24,25].

3.5. Effects of reveromycin A on apoptosis in osteoclasts cultured onvarious materials

We determined whether osteoclasts on sHA and Ti secretedprotons toward the surfaces using RM-A. We previously reportedthat V-ATPase in osteoclasts cultured on plastic was localized along

the apical membrane at higher densities than along the basolateralmembrane, whereas V-ATPase in osteoclasts cultured on collagengel was distributed in the cytoplasm without polarity [6]. We firstconfirmed whether RM-A could induce apoptosis in osteoclastscultured on plastic, but not in those cultured on collagen gel.Osteoclasts were cultured on plastic and collagen gel (Fig. 4A, B),and following treatment with RM-A, osteoclasts spread on plasticand shrank on collagen gel, indicating that RM-A treatment inducedapoptosis in osteoclasts cultured on plastic, but not in thosecultured on collagen gel (Fig. 4A, B). CT was then added to bothplastic and collagen gel cultures. CT prevented RM-A-inducedapoptosis in osteoclasts cultured on plastic, while CT had no effecton osteoclasts cultured on collagen gel. These results suggested thatosteoclasts cultured on plastic secreted protons toward the surface,while those cultured on collagen gel did not. Additionally, theseresults suggested that CT disrupted the polarization of osteoclastscultured on plastic. We then examined the effects of RM-A onapoptosis in osteoclasts cultured on Ti and sHA in the presence orabsence of CT (Fig. 4C, D). RM-A induced apoptosis in osteoclastscultured on Ti and sHA, which could be suppressed by CT.

4. Discussion

Two kinds of crystals, biologically degradable and non-degrad-able, consist of calcium and phosphate. sHA is a non-degradablecrystal. Okuda et al. [26] reported that when sHA was implanted inrabbit femurs, osteoclasts were observed on the surface of sHA,which remained almost unchanged until 72 weeks after implanta-tion. Resorption pits were also absent on the surface of sHA onwhichosteoclasts derived from newborn rabbits were cultured [24]. Twopossibilities have been suggested to explain the resistance of sHA to

sHATi

CTControlCTControl

50

0Control CT

150

100

50

0 Control CT

250

200100

150

100 µm 100 µm

Ost

eocl

asts

hav

ing

actin

ring

s/un

it

Ost

eocl

asts

hav

ing

actin

ring

s/un

it

* *

Fig. 2. Effects of Calcitonin (CT) on actin rings preformed by osteoclasts cultured on Ti and sHA. Osteoclast preparations were cultured for 8 h on Ti (A) and sHA (B) slices. CTwas added at a concentration of 10�8 M for 2 h where indicated. Cells were fixed and processed for actin ring formation. Osteoclasts having actin rings with a diameter ofmore than 100 μm were counted (lower panels). The results are expressed as the means7S.D. of three or four cultures. *: po0.01.

T. Nakayama et al. / Journal of Oral Biosciences 56 (2014) 136–142 139

biodegradability: osteoclasts cannot polarize on non-resorbablematerials, or osteoclasts can polarize on sHA but cannot resorb sHAbecause of its insolubility. The results of the present study revealedthe latter as a possibility. Osteoclasts formed actin rings on sHA, andthese were sensitive to CT. Osteoclasts also put TRAP-marks on thesurface of sHA. Although osteoclasts could not form resorption pitson sHA, RM-A induced apoptosis in osteoclasts cultured on sHA.These results suggested that osteoclasts secrete protons and catabolicenzymes toward the surface of sHA. Resorption pits were not

observed on sHA, which may be attributed to the acid resistance ofsHA [24].

A previous study demonstrated that osteoclasts could notpolarize on collagen gel [6]. Furthermore, osteoclasts cultured oncollagen gel failed to form actin rings [6]. We demonstratedthat osteoclasts cultured on collagen gel showed resistance toRM-A. In contrast, osteoclasts cultured on Ti formed actinrings, which were disrupted by CT and the GRGDS peptide. Theseresults indicated that the characteristics of actin rings formed on

Ti sHA

sHA

200 µm200 µm200 µm

TRA

PTR

AP

-mar

kD

entin

Dentin

Fig. 3. Tartrate-resistant acid phosphatase (TRAP)-mark and resorption pit formation by osteoclasts cultured on dentin, Ti, and sHA. (A) TRAP-mark formation by osteoclasts.Osteoclast preparations were cultured for 48 h on dentin, Ti, and sHA slices in the presence of 10% FBS. Cells were then fixed and stained for TRAP (upper panels). Afterosteoclasts were examined on each slice, the cells were thoroughly removed from dentin, Ti, and sHA slices with cotton swabs, and the slices were then further stained forTRAP (lower panels). TRAP-marks are indicated by arrows. For sHA slices, the area of the upper panel is the same as that of the lower panel, and the right panels are enlargedareas of the boxed areas of the left panels. (B) Formation of resorption pits by osteoclasts. Osteoclast preparations were cultured for 48 h on dentin and sHA. Cells werecompletely removed with cotton swabs, and slices were processed for scanning electron microscopy. The right panels show the boxed areas on the left panels. Note thatmany pits were observed on dentin slices, but not on sHA slices.

T. Nakayama et al. / Journal of Oral Biosciences 56 (2014) 136–142140

Ti were similar to those of actin rings formed on bone in nature.Apoptosis in osteoclasts cultured on Ti as well as on sHA wasinduced by RM-A. These results also suggested that osteoclastsmight recognize implant materials as calcified tissues and begin toresorb them.

We attempted to determine the requirements for the polariza-tion of osteoclasts. The results obtained in the previous studyindicated that (1) the polarization of osteoclasts was substrate-dependent, (2) certain proteins containing the RGD sequence werenecessary for the polarization of osteoclasts, and (3) certain

0

100

200

300

Control CT RM-A CT+

RM-A

0

40

80

120

160

Control CT RM-A CT+

RM-A

Cont

CT

RM-A

CT+RM-A

Plastic Collagen gel

0

100

150

200

250

50

Control RM-A CT+

RM-A

Ost

eocl

asts

/uni

t

020406080

100120140

Control RM-A CT+

RM-A

sHATi

RM-A CT+RM-ACont

Cont

CT

RM-A

CT+RM-A

200 µm200 µm

200 µmRM-A CT+RM-ACont 200 µm

Ost

eocl

asts

/uni

t

Ost

eocl

asts

/uni

t

Ost

eocl

asts

/uni

t

**

*

*

*

*

Fig. 4. Effects of reveromycin A (RM-A) on apoptosis in osteoclasts cultured on various materials. Osteoclast preparations were cultured for 8 h on plastic [A] and a collagengel [B]. Cells were then treated with or without 10�6 M RM-A. Some cultures were also treated with 10�8 M CT. After being cultured for 12 h, cells were fixed and stained forTRAP. TRAP-positive cells in which the longitudinal axis of the cell body was more than 100 μm were counted as living osteoclasts. The results were expressed as themeans7S.D. of four cultures. [C, D] Osteoclast preparations were cultured for 8 h on Ti [C] or sHA [B]. Cells were then treated with or without 1 mM RM-A together with orwithout CT (10�8 M). After being cultured for 12 h, cells were fixed and stained for TRAP. TRAP-positive cells in which the longitudinal axis of the cell body was more than100 μm were counted as living osteoclasts. The results are expressed as the means7S.D. of three or four cultures. *: po0.01.

T. Nakayama et al. / Journal of Oral Biosciences 56 (2014) 136–142 141

physical properties of the substrate, such as rigidity and hardness,were also essential [6,7]. These results raise the question of howosteoclasts recognize physical properties of the substrate. Wepreviously reported that the adopter protein p130Cas (Crk-asso-ciated substrate) was highly expressed in osteoclasts [27]. p130Cas

co-localized with actin rings, and the tyrosine residues of p130Cas

were highly phosphorylated in association with the formation ofactin rings. Sawada et al. [28] showed that p130Cas acted as aprimary mechanosensor in cells adhering to hard substrates. Thetensile force of cell adhesion induced a mechanical extension ofp130Cas, resulting in the tyrosine phosphorylation of p130Cas andsubsequent activation of downstream signaling to induce polar-ization of osteoclasts. Thus, osteoclasts were shown to recognizethe hardness of implant materials as bone through p130Cas [29].This mechanism must be activated in osteoclasts attaching tosurfaces of implant materials.

Osteoclastic bone resorption is tightly coupled with osteoblas-tic bone formation [30]. Bone formation is usually observed oncement lines, which are previously resorbed surfaces of bone.Therefore, cement lines have been speculated to contain so-called“coupling factors”, which promote differentiation and function ofosteoblasts. Not only bone matrix-derived factors, but alsoosteoclast-secreted factors are proposed to participate in such acoupling phenomenon [31]. Osseointegration of dental implants isa functional connection between load-bearing implants and livingbone [32] and involves osteoblastic bone formation. In ourexperiments, osteoclasts secreted protons and catabolic enzymestoward the surface of implant materials when adhering to them.These findings suggest that osteoclasts attached onto implantmaterials may secrete some factors that promote bone formationby osteoblasts. Further studies will elucidate the role of osteoclastsin osseointegration of dental implants.

Ethical approval

All experiments using mice were approved by the MatsumotoDental University Experimental Animal Committee and conductedin accordance with their guidelines for studies with laboratoryanimals.

Conflict of interest

The authors have no conflicts of interest to declare with respectto the authorship and/or publication of this article.

Acknowledgments

The authors thank Dr. Yasuhiro Kobayashi (Matsumoto DentalUniversity, Nagano, Japan) for his critical reading of the manuscript.This work was supported in part by a Grants-in-Aid for ScientificResearch from the Ministry of Education, Culture, Sports, Scienceand Technology of Japan (21390498, 22592201, and 25221310).

References

[1] Takahashi N, Udagawa N, Suda T. A new member of tumor necrosis factorligand family, ODF/OPGL/TRANCE/RANKL, regulates osteoclast differentiationand function. Biochem Biophys Res Commun 1999;256:449–55.

[2] Teitelbaum SL. Bone resorption by osteoclasts. Science 2000;289:1504–8.[3] Vaananen HK, Horton M. The osteoclast clear zone is a specialized cell-

extracellular matrix adhesion structure. J Cell Sci 1995;108(Pt 8):2729–32.[4] Saltel F, Chabadel A, Bonnelye E, Jurdic P. Actin cytoskeletal organisation in

osteoclasts: a model to decipher transmigration and matrix degradation. Eur JCell Biol 2008;87:459–68.

[5] Horton MA. The alpha v beta 3 integrin vitronectin receptor. Int J Biochem CellBiol 1997;29:721–5.

[6] Nakamura I, Takahashi N, Sasaki T, Jimi E, Kurokawa T, Suda T. Chemical andphysical properties of the extracellular matrix are required for the actin ringformation in osteoclasts. J Bone Miner Res 1996;11:1873–9.

[7] Suda T, Nakamura I, Jimi E, Takahashi N. Regulation of osteoclast function. JBone Miner Res 1997;12:869–79.

[8] Blair HC, Teitelbaum SL, Ghiselli R, Gluck S. Osteoclastic bone resorption by apolarized vacuolar proton pump. Science 1989;245:855–7.

[9] Kirstein B, Chambers TJ, Fuller K. Secretion of tartrate-resistant acid phospha-tase by osteoclasts correlates with resorptive behavior. J Cell Biochem2006;98:1085–94.

[10] Fuller K, Ross JL, Szewczyk KA, Moss R, Chambers TJ. Bone is not essential forosteoclast activation. PLoS One 2010;5:e12837.

[11] Nakayama T, Mizoguchi T, Uehara S, Yamashita T, Kawahara I, Kobayashi Y,Moriyama Y, Kurihara S, Sahara N, Ozawa H, Udagawa N, Takahashi N.Polarized osteoclasts put marks of tartrate-resistant acid phosphatase ondentin slices – a simple method for identifying polarized osteoclasts. Bone2011;49:1331–9.

[12] Chambers TJ, Magnus CJ. Calcitonin alters behaviour of isolated osteoclasts. JPathol 1982;136:27–39.

[13] Lin HY, Harris TL, Flannery MS, Aruffo A, Kaji EH, Gorn A, Kolakowski Jr. LF,Lodish HF, Goldring SR. Expression cloning of an adenylate cyclase-coupledcalcitonin receptor. Science 1991;254:1022–4.

[14] Suzuki H, Nakamura I, Takahashi N, Ikuhara T, Matsuzaki K, Isogai Y, Hori M,Suda T. Calcitonin-induced changes in the cytoskeleton are mediated by asignal pathway associated with protein kinase A in osteoclasts. Endocrinology1996;137:4685–90.

[15] Takahashi N, Ejiri S, Yanagisawa S, Ozawa H. Regulation of osteoclastpolarization. Odontology 2007;95:1–9.

[16] Takahashi H, Osada H, Koshino H, Sasaki M, Onose R, Nakakoshi M, YoshihamaM, Isono K. Reveromycins, new inhibitors of eukaryotic cell growth. II.Biological activities. J Antibiot 1992;45:1414–9.

[17] Woo JT, Kawatani M, Kato M, Shinki T, Yonezawa T, Kanoh N, Nakagawa H,Takami M, Lee KH, Stern PH, Nagai K, Osada H. Reveromycin A, an agent forosteoporosis, inhibits bone resorption by inducing apoptosis specifically inosteoclasts. Proc Natl Acad Sci USA 2006;103:4729–34.

[18] Simonet WS, Lacey DL, Dunstan CR, Kelley M, Chang MS, Luthy R, Nguyen HQ,Wooden S, Bennett L, Boone T, Shimamoto G, DeRose M, Elliott R, ColomberoA, Tan HL, Trail G, Sullivan J, Davy E, Bucay N, Renshaw-Gegg L, Hughes TM,Hill D, Pattison W, Campbell P, Sander S, Van G, Tarpley J, Derby P, Lee R, BoyleWJ. Osteoprotegerin: a novel secreted protein involved in the regulation ofbone density. Cell 1997;89:309–19.

[19] Bucay N, Sarosi I, Dunstan CR, Morony S, Tarpley J, Capparelli C, Scully S, TanHL, Xu W, Lacey DL, Boyle WJ, Simonet WS. Osteoprotegerin-deficient micedevelop early onset osteoporosis and arterial calcification. Genes Dev1998;12:1260–8.

[20] Mizuno A, Amizuka N, Irie K, Murakami A, Fujise N, Kanno T, Sato Y, NakagawaN, Yasuda H, Mochizuki S, Gomibuchi T, Yano K, Shima N, Washida N, Tsuda E,Morinaga T, Higashio K, Ozawa H. Severe osteoporosis in mice lackingosteoclastogenesis inhibitory factor/osteoprotegerin. Biochem Biophys ResCommun 1998;247:610–5.

[21] Tanaka M, Miyazawa K, Tabuchi M, Yabumoto T, Kadota M, Yoshizako M,Yamane C, Kawatani M, Osada H, Maeda H, Goto S. Effect of reveromycin A onexperimental tooth movement in OPG-/- mice. J Dent Res 2012;91:771–6.

[22] Doi Y, Koda T, Wakamatsu N, Goto T, Kamemizu H, Moriwaki Y, Adachi M, SuwaY. Influence of carbonate on sintering of apatites. J Dent Res 1993;72:1279–84.

[23] Takahashi N, Udagawa N, Kobayashi Y, Suda T. Generation of osteoclastsin vitro, and assay of osteoclast activity. Methods Mol Med 2007;135:285–301.

[24] Doi Y, Shibutani T, Moriwaki Y, Kajimoto T, Iwayama Y. Sintered carbonateapatites as bioresorbable bone substitutes. J Biomed Mater Res 1998;39:603–10.

[25] Nakamura M, Hentunen T, Salonen J, Nagai A, Yamashita K. Characterization ofbone mineral-resembling biomaterials for optimizing human osteoclast dif-ferentiation and resorption. J Biomed Mater Res A 2013;101:3141–51.

[26] Okuda T, Ioku K, Yonezawa I, Minagi H, Gonda Y, Kawachi G, Kamitakahara M,Shibata Y, Murayama H, Kurosawa H, Ikeda T. The slow resorption withreplacement by bone of a hydrothermally synthesized pure calcium-deficienthydroxyapatite. Biomaterials 2008;29:2719–28.

[27] Nakamura I, Jimi E, Duong LT, Sasaki T, Takahashi N, Rodan GA, Suda T.Tyrosine phosphorylation of p130Cas is involved in actin organization inosteoclasts. J Biol Chem 1998;273:11144–9.

[28] Sawada Y, Tamada M, Dubin-Thaler BJ, Cherniavskaya O, Sakai R, Tanaka S,Sheetz MP. Force sensing by mechanical extension of the Src family kinasesubstrate p130Cas. Cell 2006;127:1015–26.

[29] Nakamura I, Takahashi N, Jimi E, Udagawa N, Suda T. Regulation of osteoclastfunction. Mod Rheumatol 2012;22:167–77.

[30] Parfit AM. The coupling of bone formation to bone resorption: a criticalanalysis of the concept and of its relevance to the pathogenesis of osteoporo-sis. Metab Bone Dis Relat Res 1982;4:1–6.

[31] Henriksen K, Karsdal MA, Martin TJ. Osteoclast-derived coupling factors inbone remodeling. Calcif Tissue Int 2014;94:88–97.

[32] Le Guéhennec L, Soueidan A, Layrolle P, Amouriq Y. Surface treatments oftitanium dental implants for rapid osseointegration. Dent Mater 2007;23:844–54.

T. Nakayama et al. / Journal of Oral Biosciences 56 (2014) 136–142142