TBD

APmaaipdoacwpmeOtrtcaiocrLRbmsLiLt

KCt

BH

oD0

Ed

Basic Research—Biology

4

he Role of Cyclooxygenase-2 (COX-2) in Inflammatoryone Resorptionavid Coon, DDS, Ajay Gulati, BDS, MSc, Cameron Cowan, BS, and Jianing He, DMD, PhD

Iarss1r

pi(ioait

hPclmbsa

mmawm

Aspat

irvorRtRo

bstractrostaglandin E2 (PGE2) is an important inflammatoryediator that plays an essential role in the development

nd progression of periradicular diseases. Cyclooxygen-se-2 (COX-2) is the inducible enzyme responsible for

ncreased PGE2 levels during inflammation and otherathologic processes. The purpose of this study was toetermine the role of COX-2-mediated PGE2 synthesis insteoclast formation in response to endodontic pathogensnd materials. Primary osteoblast cultures and osteoclastultures were prepared from COX-2 knockout (K/O) andild-type (WT) littermates. These cultured cells were ex-osed to lipopolysaccharide (LPS) or root canal obturationaterials including gutta-percha (GP), Resilon (RS), min-

ral trioxide aggregates (MTAs), and AH Plus (AH�).steoclast formation was evaluated using tartrate-resis-

ant acid phosphatase (TRAP) staining. The expression ofeceptor activator of NF-�B ligand (RANKL) and osteopro-egerin (OPG) was determined by real-time polymerasehain reaction (PCR) analysis. It was found that in both WTnd K/O cultures, treatment with LPS led to a marked

ncrease in osteoclast formation. The number of oste-clasts formed was significantly lower in K/O culturesompared to WT cultures. Exposure to endodontic mate-ials did not lead to any significant osteoclast formation.PS and endodontic materials caused a decrease in bothANKL and OPG expression in WT cells. In K/O cells, theaseline levels of RANKL and OPG expression were dra-atically decreased compared to the WT cells. In conclu-

ion, COX-2-mediated PGE2 expression is required forPS-induced inflammatory bone resorption and maintain-ng the baseline level of RANKL and OPG expression.PS-induced osteoclast formation may be independent ofhe RANKL pathway. (J Endod 2007;33:432–436)

ey Wordsyclooxygenase-2, inflammation, osteoclast, osteopro-

egerin, receptor activator of NF-�B ligand

From the Department of Endodontics, Biomedical Sciences,aylor College of Dentistry, The Texas A&M University Systemealth Science Center, Dallas, Texas.

Address requests for reprints to Dr. Jianing He, Departmentf Endodontics, Baylor College of Dentistry, 3302 Gaston Ave.,allas, TX 75246. E-mail address: jhe@bcd.tamhsc.edu.099-2399/$0 - see front matter

Copyright © 2007 by the American Association ofndodontists.oi:10.1016/j.joen.2006.12.001

b

32 Coon et al.

nflammatory bone resorption is a pathologic process associated with chronic perira-dicular periodontitis. Bacterial infection is the primary cause of endodontic disease

nd subsequent inflammatory bone resorption in the periradicular region (1). Bacte-ial cells and their structural components such as lipopolysaccharides (LPS) initiate aeries of inflammatory and immune reactions leading to the development and progres-ion of periradicular bone resorption (2). Inflammatory cytokines such as interleukin(IL-1), tumor necrosis factor � (TNF-�), and prostaglandins (PGs) play an important

ole in the formation and activation of osteoclasts that mediate bone resorption (2).Prostaglandins (PGs) are a family of cytokines with a wide range of biologic and

athologic functions, such as the maintenance of tissue homeostasis, the mediation ofnflammatory responses, and the development of neoplasms (3). CyclooxygenaseCOX) is an enzyme responsible for converting arachidonic acid to PGE2. At least twosoforms of COX have been elucidated: COX-1, which is responsible for the productionf PGs with homeostatic functions in tissues such as the stomach, kidney, and platelets;nd COX-2, the inducible enzyme, which is responsible for the production of PGsnvolved in inflammation. Inflammatory mediators including IL-1, TNF-�, growth fac-ors, LPS, and tumor cells are known stimulators of COX-2 expression (4, 5).

Increased production of PGE2 has been demonstrated in periradicular lesions; thisas been suggested to account for a great portion of the bone resorptive activity (6, 7).GE2 levels in periradicular exudates from root canals are closely correlated to thelinical symptoms of endodontic disease, especially the occurrence of pain (8, 9). Highevels of PGE2 are often linked to acute, painful periradicular periodontitis. Further-

ore, inhibition of PGE2 synthesis results in suppressed inflammatory changes andone resorption in experimental animals (10). The use of selective COX-2 inhibitors hashown promising results in pain relief following endodontic treatment without thedverse effects associated with the nonselective COX inhibitors (11, 12).

Periradicular periodontitis is commonly managed by nonsurgical root canal treat-ent. This treatment involves cleaning, shaping, and obturating the root canal withaterials that seal the canal throughout its length to eliminate bacteria, prevent leakage,

nd promote healing. Endodontic obturation materials may have prolonged contactith periradicular tissues and ideally should have good biocompatibility. However,ost obturation materials currently in use have a certain degree of cytotoxicity (13).

Endodontic obturation materials have been shown to affect cytokine expression.ctivation of COX-2 expression with subsequent PGE2 production has been hypothe-ized as one of the causes of pathogenesis associated with root canal sealer–inducederiapical inflammation (5). Activation of other cytokines such as IL-6 and IL-8 maylso be involved (14). The role of cytokines in obturation material–induced inflamma-ion is not fully understood.

Inflammatory bone resorption is a complex and tightly regulated process thatnvolves many cytokines and cross-talk between different cell types. The main cell typeesponsible for carrying out bone resorption is the osteoclast. Osteoclastogenesis in-olves cell– cell interactions between osteoblasts or bone marrow stromal cells andsteoclast progenitors. These interactions are mediated by key molecules includingeceptor activator of NF-�B ligand (RANKL), RANK, and osteoprotegerin (OPG) (15, 16).ANKL is expressed by osteoblastic cells as a membrane-associated factor. RANKL binds

o RANK, which is a receptor of RANKL expressed by osteoclast progenitors. RANK–ANKL interaction stimulates the differentiation of osteoclast progenitors into maturesteoclasts in the presence of macrophage-colony stimulating actor (M-CSF). Osteo-

lastic cells also produce a soluble factor OPG, which acts as a decoy receptor for

JOE — Volume 33, Number 4, April 2007

RbIOpeup

tmdvtbr

Pset

A

pCC(HCgo

DfcCTpCK

6uT

P

STMpbtp3ewmv

cc(t

P

dwwI0Tssct�(Eae

R

maaa

(at

T

T

Basic Research—Biology

J

ANKL and inhibits osteoclast formation by interrupting the interactionetween RANK and RANKL. Inflammatory cytokines including PGE2,L-1, and TNF-� have been shown to affect the expression of RANKL andPG (17–20). The exact role of COX-2-mediated PGE2 induction ineriradicular inflammatory bone resorption and in RANKL and OPGxpression remains to be determined. The efficacy and the rationale forsing selective COX-2 inhibitors in managing endodontic pain anderiradicular bone resorption need to be further explored.

COX-2 knockout (K/O) mice have no detectable levels of PGE2 andherefore can be used for investigating the role of PGE2 in the inflam-

atory process (21, 22). Mice lacking COX-2 expression display re-uced bone resorption in response to parathyroid hormone (PTH) oritamin D3 (23). Primary osteoblasts and osteoclasts prepared fromhese COX-2 deficient animals provide an ex vivo environment that cane used to study the modulation of inflammatory response and boneesorption in the absence of PGE2.

The aim of this study was to determine the role of COX-2-mediatedGE2 synthesis in inflammatory bone resorption. The effect of LPS andeveral common endodontic obturation materials on osteoclastogen-sis and the expression of bone resorption modulating factors in wildype (WT) and K/O cultures were determined.

Materials and Methodsnimals

All animal studies were conducted in accordance with the princi-les and procedures approved by the Texas A&M System Health Scienceenter Baylor College of Dentistry Institutional Animal Care and Useommittee. Male and female heterozygous COX-2 knockout micePtgs2 �/�, HET) were purchased from the Jackson Laboratory (Bararbor, ME). Mice were bred within the animal resource unit at Baylorollege of Dentistry. These mice were crossed to obtain the followingenotypes: wild type (Ptgs2 �/�, WT), HET (Ptgs2 �/�), and knock-ut (Ptgs2 �/�, K/O).

Mouse tail genomic DNA was extracted using the Wizard genomicNA extraction kit (Promega, Madison, WI). Genotyping was per-

ormed using primer pairs specific for the WT cox-2 gene and the neoassette in the K/O gene. cox-2 forward primer (5=-CACCATAGAATC-AGTCCGG-3=) and cox-2 reverse primer (5=-ACCTCTGCGATGCTC-TCC-3=) generate an 875-bp fragment for the WT cox-2 gene; neo1rimer (5=-GCCCTGAATGAACTGCAGGACG-3=) and neo2 primer (5=-ACGGGTAGCCAACGCTATGTC-3=) generate a 500-bp fragment for the/O gene.

Polymerase chain reaction (PCR) cycles were: 94°C, 30 seconds;6°C, 1 minute; 72°C, 1 minute. The cycle number was 35. PCR prod-cts were fractionated by electrophoresis on a 1% agarose gel in 1�BE buffer and visualized by ethidium bromide staining.

reparation of Endodontic MaterialsGutta-percha (GP, Obtura Spartan, Fenton, MO), RealSeal (RS,

ybronEndo, Orange, CA), white ProRoot™ MTA (MTA, Dentsply,ulsa, OK), and AH Plus (AH�, Dentsply) were selected for testing.aterials were prepared according to manufacturer’s directions and

acked onto the bottom of 24-well dishes (Corning-Costar Corp., Cam-ridge, MA) to create a uniform surface area (1.9 cm2). For GP and RS,

hermoplasticized pellets were used to obtain the samples. The freshlyrepared samples were allowed to set for 24 hours in a cell incubator at7°C and 100% humidity. After 24 hours, 1 ml of Dulbecco’s modifiedagle medium (DMEM, Invitrogen, Grand Island, NY) supplementedith 10% fetal bovine serum (FBS, VWR International, Irving, TX) and aixture of 100 U/ml penicillin and 50 �g/ml streptomycin (P/S, In-

itrogen) was added to each well. Materials were incubated with the

OE — Volume 33, Number 4, April 2007

ulture medium at 37°C and 100% humidity. The material elutes wereollected after 24 hours and passed through a sterile 0.2-�m filterFisher Scientific, Pittsburgh, PA), aliquotted into 1.5-ml Eppendorfubes (VWR International), and stored at �20°C until use.

rimary Osteoblast CultureCalvariae halves from 6- to 8-week-old WT and K/O mice were

issected free of suture. These calvariae halves of the same genotypeere grouped and subjected to four sequential 15-minute digestionsith an enzyme mixture of 1.5 U/ml collagenase P (Roche Diagnostics,

ndianapolis, IN) in phosphate-buffered saline (PBS, Invitrogen) and.05% trypsin/1 mM EDTA (Invitrogen) at 37°C on a rocking platform.he first digest containing fibroblast contamination was discarded. Theecond to the fourth digest were pooled. After an initial in vitro expan-ion, cells were seeded in 6-well dishes (Corning-Costar) at 5,000 cells/m2 in DMEM supplemented with 10% FBS and P/S. Cells were allowedo grow at 37°C for 48 hours. Various treatments including LPS (10

g/ml) (E coli, Sigma, St. Louis, MO) and endodontic material elutes1:10 dilution) were then added. Treatment groups are listed in Table 1.ach group included 3 wells. Experiments were repeated three timesnd results were pooled. Cells from passages 2 to 4 were used for thexperiment.

eal-Time PCRRNA from primary osteoblasts was collected 24 hours after treat-

ent. RNA was isolated using 0.1 ml/cm2 of TRIzol reagent (Invitrogen)nd phenol/chloroform (Sigma). RNA was quantitated by measuring thebsorbance at 260/280 nm on a spectrophotometer. cDNAs were cre-ted by reverse transcription (RETROscript, Ambion, Austin, TX).

Primers specific for mouse receptor activator of NF-�B ligandRANKL) and osteoprotegerin (OPG) were utilized for real-time PCRmplification. An additional primer, GAPDH, served as an internal con-rol. Primer sequences are listed in Table 2. Real-time PCR was per-

ABLE 1. Treatment groups for osteoblast and osteoclast cultures

GroupTreatment

Osteoblast Culture Osteoclast Culture

1 (Negativecontrol)

Culture medium Culture medium

2 Lipopolysaccharide(LPS)(10 �g/ml)

Lipopolysaccharide(LPS)(1 �g/ml), first 3 days only

3 Gutta-percha elute(GP)

Gutta-percha elute(GP)

4 Real Seal elute(RS)

Real Seal elute(RS)

5 ProRoot MTA elute(MTA)

ProRoot MTA elute(MTA)

6 AH plus elute(AH�)

AH plus elute(AH�)

7 — 1, 25-dihydroxyvitamin D3(1, 25-D, 10 nM)

ABLE 2. Primer sequences used in real-time PCR

Gene Primer Sequence

RANKL 5=-GGGAATTACAAAGTGCACCAG-3=5=-GGTCGGGCAATTCTGAATT-3= (34)

OPG 5=-TCCTGGCACCTACCTAAAACAGCA-3=5=-CTACACTCTCGGCATTCACTTTGG-3= (35)

GAPDH 5=-GGTCGGTGTGAACGGATTTGG-3=

5=-ATGTAGGCCATGAGGTCCACC-3= (35)

The Role of Cyclooxygenase-2 (COX-2) in Inflammatory Bone Resorption 433

fp

vtcM0cct5apfc

as(oi

B

dcAbcgtpmttea1�oi1

3mragw

O

ts(5Nstat

fiaTcTp

S

ap

O

goaosfctdFf

FD0(

Basic Research—Biology

4

ormed using the Brilliant SYBR Green system with the Mx4000 Multi-lex Quantitative PCR system (Stratagene, Cedar Creek, TX).

Optimization was performed by establishing a standard curve witharious concentrations of cDNA template and a primer matrix to obtainhe optimal primer concentration. Reactions were 20 �l in total volumeonsisting of a final concentration of 1� Brilliant SYBR Green QPCRaster Mix (Stratagene, Cedar Creek, TX), 0.15 �M RANKL primers,

.3 �M OPG primers, or 0.3 �M GAPGH primer solution, and 0.4 �MDNA template. Amplification conditions consisted of an initial prein-ubation at 95°C for 10 minutes (FastStart Taq DNA polymerase activa-ion), followed by amplification for 50 cycles (95°C for 30 seconds,5°C for 60 seconds, and 72°C for 30 seconds). Dissociation curvenalysis was performed immediately after amplification at a linear tem-erature transition rate of 0.1°C/s from 65°C to 95°C with continuousluorescence acquisition. All sample amplifications were run in dupli-ate.

The model proposed by Pfaffl (24) was used to calculate the rel-tive expression level of the target gene (RANKL and OPG) in compari-on to GAPDH. This model is based on the corresponding PCR efficiencyE) of one cycle in the exponential phase and the amplification thresh-ld cycle value (Ct). Results are presented as means � SD. Three

ndependent experiments were performed and results were pooled.

one Marrow CulturesFemurs and tibias from 6- to 8-week-old WT and K/O mice were

issected free of adherent soft tissue. Epiphyses were removed asepti-ally and marrow was flushed out with Minimum Essential Mediumlpha (�-MEM, Invitrogen) supplemented with 10% FBS and P/S anti-iotic mixture. Single cell suspension was obtained by pipetting cells inulture medium and cell numbers were counted. A microscope coverlass slip with a diameter of 12 mm (Fisher Scientific) was inserted athe base of each well of the 24-well plates (Corning-Costar). Cells werelated onto the cover slips at a density of 2.0 � 106 cells/well. Thisethod allowed the transfer and mounting of the cover slips with at-

ached cells onto a microscope slide for better morphologic observa-ion at the time of analysis. The LPS (1 �g/ml) (E coli, Sigma) andndodontic material elutes were added to the wells at the time of plating,s shown in Table 1. In addition, one group was treated with 10 nM,25-dihydroxyvitamin D3 (1,25-D, Sigma) as a positive control. 1g/ml LPS was used in this experiment because of the high cytotoxicity

f the 10 �g/ml LPS in bone marrow cultures observed in a trial exper-ment (data not shown). The endodontic material elutes were used in:10 dilution.

The plates were then cultured for 8 days at 37°C in 5% CO2. At day, half of the culture medium was changed and replaced with freshedium. At day 7, a full culture medium change was done. LPS was

emoved at the first medium change. Fresh endodontic material elutesnd 1,25-D were supplemented at each medium change. Osteoclasto-enesis was assessed after 8 days. Each treatment group included 6ells.

steoclast FormationOsteoclastogenesis was assessed by cytochemical staining for tar-

rate-resistant acid phosphatase (TRAP), a marker of osteoclasts. TRAPtaining was performed using a leukocyte acid phosphatase staining kitSigma). Cells were rinsed, fixed with 2.5% glutaraldehyde (Sigma) forminutes, and incubated with the TRAP staining solution, containingaphthol AS-BI phosphate solution, Diazotized Fast Garnet GBC baseolution, acetate solution, and tartrate solution. After 2 hours incuba-ion at 37°C, the wells were rinsed with deionized water. A drop of goldntifade reagent (Slowfade) was placed on each microscope slide, and

he 12-mm microscope cover glass slip with attached cells was removed c

34 Coon et al.

rom the bottom of the culture well, inverted, and placed on the mount-ng reagent. Slides were analyzed using Zeiss Axioplan microscope withSpot Camera system (Diagnostic Instruments, Sterling Heights, MI).RAP-positive cells appeared dark purple. The TRAP-positive, multinu-leated (3 or more nuclei) osteoclast-like cells per well were counted.hree independent experiments were performed and results wereooled.

tatistical AnalysisThe results are presented as means � SD. Statistical differences

mong groups were determined by one-way ANOVA followed by Tukey’sost hoc test. Values of p � 0.05 were considered significant.

Resultssteoclast Formation

No TRAP positive osteoclast was seen in any of the negative controlroups. The positive control group treated with 1,25-D showed markedsteoclast formation in WT cultures (Fig. 1). LPS treatment also causedn increase in osteoclast numbers in WT cultures. No osteoclast wasbserved in the GP-treated groups. Occasional TRAP-positive cells wereeen in RS-treated and MTA-treated groups, but at a much lowerrequency compared to the positive control and LPS groups. In K/Oulture, 1,25-D and LPS also stimulated osteoclast formation. However,he number of osteoclasts formed was lower compared to WT. Thisifference was statistically significant (p � 0.05) in LPS-treated groups.urthermore, endodontic materials did not stimulate any osteoclastormation in K/O cultures.

igure 1. (A) TRAP-positive osteoclast number in control and treatment groups.ata are shown as means � SD. a, significantly different from WT control (p �.05); b, significantly different from WT with the same treatment (p � 0.05).B) A representative image of a TRAP-positive multinucleated osteoclast in WT

ulture treated with LPS (1 �g/ml). Original magnification �40.

JOE — Volume 33, Number 4, April 2007

ji

R

optqw

enRNm

RdT(

mc

mEdrdeC

dtPrtmiots

amvtadTwr

Wsttstratt

rihaRcw(ptleLcf

FRpsf

Basic Research—Biology

J

AH� showed marked cytotoxicity leading to cell death in the ma-ority of the cells in the culture. Therefore, this group was not includedn the analysis.

ANKL and OPG ExpressionThe effects of the LPS and obturation materials on the expression

f RANKL and OPG were examined by real-time PCR. Results are ex-ressed as relative expression in comparison to GAPDH. Cells exposedo AH� elute showed significant cell death. Consequently, a sufficientuantity of RNA could not be obtained. Therefore, this group (AH�)as not included in the analysis.

In WT cells, LPS treatment led to a marked decrease in RANKLxpression, as did the obturation materials, when compared to theegative control (p � 0.05) (Fig. 2A). In K/O cells, the baseline level ofANKL was dramatically diminished in the negative control (p � 0.05).o significant change was seen with the treatment of LPS and endodonticaterials.

OPG gene expression pattern followed a similar trend to that of theANKL results (Fig. 2B). LPS and the endodontic materials caused aecrease in OPG expression levels compared to control in WT cultures.his decrease was slightly more pronounced in the GP and RS groupsp � 0.05). In K/O cells, baseline levels of OPG also decreased dra-

igure 2. RANKL and OPG mRNA expression determined by real-time PCR. (A)ANKL expression. (B) OPG expression. Results are expressed as relative ex-ression ratio in comparison to GAPDH. Data are shown as means � SD. a,ignificantly different from WT control (p � 0.05); b, significantly different

arom WT with the same treatment (p � 0.05).

OE — Volume 33, Number 4, April 2007

atically in the control group. LPS and endodontic materials did notause any significant change when compared to the negative control.

DiscussionLPS is an important etiologic factor contributing to the develop-

ent and progression of periradicular inflammatory bone resorption.ndodontic materials may be in prolonged direct contact with perira-icular tissue and therefore can contribute to the disease process in thisegion. This study was designed to investigate the mechanisms of perira-icular inflammatory bone resorption in response to LPS and selectedndodontic obturation materials, with an emphasis on the role ofOX-2.

COX-2 is an inducible enzyme responsible for elevated PGE2 pro-uction during inflammation. Selective COX-2 inhibitors have been used

o manage endodontic pain with promising results (11, 12). BecauseGE2 has been shown to play an important role in mediating boneesorption during inflammation, inhibiting PGE2 production throughhe use of COX-2 inhibitors has the potential of influencing the inflam-

atory bone resorption process in addition to reducing pain. Althoughncreased cardiovascular adverse events associated with prolonged usef COX-2 inhibitors has raised safety concerns regarding these drugs,

he short-term use in endodontic pain management does not warrantimilar concerns (25).

COX-2 K/O mice have no detectable levels of PGE2 production andre otherwise relatively normal (21, 22). They provide an excellentodel to study the role of COX-2-mediated PGE2 in the modulation of

arious inflammatory processes. In this study, primary bone cell cul-ures were prepared from COX-2 WT and K/O mice. Inflammatory re-ctions were induced by LPS treatment and exposure to various en-odontic obturation materials. Osteoclast formation was evaluated byRAP staining. The mechanisms for the changes in osteoclastogenesisere further explored by analyzing the expression of RANKL/OPG via

eal-time PCR.LPS treatment caused a marked increase in osteoclast formation in

T cells. This result is in agreement with previous studies and is con-istent with the known effect of LPS as a potent inducer of bone resorp-ion (26, 27). In the absence of PGE2, osteoclast formation in responseo LPS was significantly decreased, as observed in the K/O cultures. Thisuggests COX-2-mediated PGE2 is necessary in LPS-induced osteoclas-ogenesis. LPS has been shown to stimulate PGE2 production, which isesponsible for increased osteoclast formation (17). These findingslso support the hypothesis that decreasing PGE2 production throughhe use of COX-2 inhibitors may have an inhibitory effect on inflamma-ory bone resorption caused by bacterial components (28).

The RANK/RANKL/OPG pathway is critically involved in the matu-ation and activation of osteoclasts. This pathway is regulated duringnflammation and other pathologic processes (15, 16). Previous studiesave found that the regulation of RANKL and OPG expression by LPSppears to be complex and biphasic. Suda et al. (17) showed increasedANKL and suppressed OPG expression in response to LPS treatment inocultures of mouse osteoblasts and bone marrow cells. Mice injectedith LPS had decreased serum RANKL levels and increased OPG levels19). In periodontal ligament fibroblasts, LPS led to upregulated ex-ression in both factors (29). When Saos-2 cells were exposed to LPS,

here was an initial increase in OPG levels and a decrease with pro-onged treatment (30). This discrepancy can be attributed to the differ-nce in the type of cell culture, culture conditions, the concentration ofPS, and the sensitivity in detection methods. It also reveals the intrinsicomplexity in the regulatory mechanism. In the present study, LPS wasound to reduce both RANKL and OPG expression in WT cells. OPG has

n inhibitory effect on bone resorption. Decreased OPG levels correlate

The Role of Cyclooxygenase-2 (COX-2) in Inflammatory Bone Resorption 435

weoo(taRfff

RdTOeRt

sdwsnabcdot

Lnuamcc

F

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

Basic Research—Biology

4

ith elevated osteoclast formation. On the other hand, decreased RANKLxpression is inconsistent with the stimulatory role of RANKL in oste-clast formation. A few studies have suggested that LPS-induced oste-clastogenesis may be independent of RANKL and osteoblastic cells31–33). It is hypothesized that during bacteria-mediated inflamma-ion, LPS binds to its cell membrane receptors, Toll-like receptors, toctivate NF-�B and phosphatidylinositol-3 kinase independent of theANKL-RANK pathway. The lack of RANKL stimulation in response to LPS

ound in our study further supports this hypothesis, where osteoclastormation during bacterial infection utilizes an alternative mechanismor activation.

In K/O cells with no PGE2 production, the baseline levels of bothANKL and OPG were dramatically diminished, suggesting COX-2-me-iated PGE2 is critical in maintaining baseline levels of RANKL and OPG.his is consistent with the finding that PGE2 stimulates both RANKL andPG expression (34). In addition, in the absence of PGE2, LPS andndodontic materials were unable to cause any further changes inANKL and OPG levels. It has been suggested that PGE2 is required for

he modulation of OPG expression during osteoclastogenesis (17).The endodontic obturation materials selected in this study did not

timulate any significant osteoclast formation, although they did cause aecrease in both RANKL and OPG expression to a similar level as seenith LPS. This finding again indicates that bacteria-induced bone re-

orption uses a pathway independent of RANKL-RANK, whereas exoge-ous materials fail to activate this unique pathway. Bone remodeling istightly coupled process in which resorption needs to occur for newone formation to take place. Decreased RANKL/OPG expressionaused by LPS and exposure to endodontic materials may lead to aecrease in overall bone turnover rate. This suggests that extrudedbturation materials and the persistent presence of LPS may contributeo delayed healing in periradicular lesion.

In conclusion, COX-2-mediated PGE2 production is required forPS-induced bone resorption during bacterial infection. This mecha-ism is independent of the RANK–RANKL interaction. The short-termse of COX-2 selective inhibitors in endodontic patients may provide thedditional benefit of inhibiting inflammatory bone resorption whileanaging pain. Commonly used endodontic obturation materials in-

luding GP, RS, and MTA do not directly cause bone resorption but mayontribute to delayed healing if extruded into the periradicular tissue.

AcknowledgmentsThis study was supported by a research grant from the AAE

oundation.

References1. Kakehashi S, Stanley HR, Fitzgerald RJ. The effects of surgical exposures of dental

pulps in germ-free and conventional laboratory rats. Oral Surg Oral Med Oral Pathol1965;20:340 –9.

2. Takahashi K. Microbiological, pathological, inflammatory, immunological and mo-lecular biological aspects of periradicular disease. Int Endod J 1998;31:311–25.

3. McCarthy JA. Prostaglandins: an overview. Adv Pediatr 1978;25:121– 49.4. Fung HB, Kirschenbaum HL. Selective cyclooxygenase-2 inhibitors for the treatment

of arthritis. Clin Ther 1999;21:1131–57.5. Huang FM, Chou MY, Chang YC. Induction of cyclooxygenase-2 mRNA and protein

expression by epoxy resin and zinc oxide-eugenol based root canal sealers in humanosteoblastic cells. Biomaterials 2003;24:1869 –75.

6. McNicholas S, Torabinejad M, Blankenship J, Bakland L. The concentration of pros-taglandin E2 in human periradicular lesions. J Endod 1991;17:97–100.

7. Wang CY, Tani-Ishii N, Stashenko P. Bone-resorptive cytokine gene expression inperiapical lesions in the rat. Oral Microbiol Immunol 1997;12:65–71.

8. Takayama S, Miki Y, Shimauchi H, Okada H. Relationship between prostaglandin E2concentrations in periapical exudates from root canals and clinical findings of peri-

apical periodontitis. J Endod 1996;22:677– 80.

36 Coon et al.

9. Shimauchi H, Takayama S, Miki Y, Okada H. The change of periapical exudateprostaglandin E2 levels during root canal treatment. J Endod 1997;23:755–58.

0. Oguntebi BR, Barker BF, Anderson DM, Sakumura J. The effect of indomethacin onexperimental dental periapical lesions in rats. J Endod 1989;15:117–21.

1. Barden J, Edwards JE, McQuay HJ, Moore RA. Single dose oral celecoxib for post-operative pain. Cochrane Database Syst Rev 2003:CD004233.

2. Gopikrishna V, Parameswaran A. Effectiveness of prophylactic use of rofecoxib incomparison with ibuprofen on postendodontic pain. J Endod 2003;29:62– 4.

3. Rappaport HM, Lilly GE, Kapsimalis P. Toxicity of endodontic filling materials. OralSurg Oral Med Oral Pathol 1964;18:785– 802.

4. Huang FM, Tsai CH, Yang SF, Chang YC. Induction of interleukin-6 and interleukin-8gene expression by root canal sealers in human osteoblastic cells. J Endod2005;31:679 – 83.

5. Anderson DM, Maraskovsky E, Billingsley WL, Dougall WC, Tometsko ME, Roux ER,et al. A homologue of the TNF receptor and its ligand enhance T-cell growth anddendritic-cell function. Nature 1997;390:175–9.

6. Lacey DL, Timms E, Tan HL, Kelley MJ, Dunstan CR, Burgess T, et al. Osteoprotegerinligand is a cytokine that regulates osteoclast differentiation and activation. Cell1998;93:165–76.

7. Suda K, Udagawa N, Sato N, Takami M, Itoh K, Woo JT, et al. Suppression of osteo-protegerin expression by prostaglandin E2 is crucially involved in lipopolysaccha-ride-induced osteoclast formation. J Immunol 2004;172:2504 –10.

8. Okada Y, Lorenzo JA, Freeman AM, Tomita M, Morham SG, Raisz LG, et al. Prosta-glandin G/H synthase-2 is required for maximal formation of osteoclast-like cells inculture. J Clin Invest 2000;105:823–32.

9. Maruyama K, Takada Y, Ray N, Kishimoto Y, Penninger JM, Yasuda H, et al. Receptoractivator of NF-kappaB ligand and osteoprotegerin regulate proinflammatory cyto-kine production in mice. J Immunol 2006;177:3799 – 805.

0. Rossa C, Ehmann K, Liu M, Patil C, Kirkwood KL. MKK3/6-p38 MAPK signaling isrequired for IL-1beta and TNF-alpha-induced RANKL expression in bone marrowstromal cells. J Interferon Cytokine Res 2006;26:719 –29.

1. Dinchuk JE, Car BD, Focht RJ, Johnston JJ, Jaffee BD, Covington MB, et al. Renalabnormalities and an altered inflammatory response in mice lacking cyclooxygenaseII. Nature 1995;378:406 –9.

2. Lim H, Paria BC, Das SK, Dinchuk JE, Langenbach R, Trzaskos JM, et al. Multiplefemale reproductive failures in cyclooxygenase 2-deficient mice. Cell 1997;91:197–208.

3. Okada Y, Pilbeam C, Raisz L, Tanaka Y. Role of cyclooxygenase-2 in bone resorption.J Uoeh 2003;25:185–95.

4. Pfaffl MW. A new mathematical model for relative quantification in real-time RT-PCR.Nucleic Acids Res 2001;29:e45.

5. McGettigan P, Henry D, Brophy JM, Klasser GD, Epstein J. Cardiovascular risk andinhibition of cyclooxygenase: a systematic review of the observational studies ofselective and nonselective inhibitors of cyclooxygenase 2. JAMA2006;296:1633– 44.

6. Zou W, Bar-Shavit Z. Dual modulation of osteoclast differentiation by lipopolysac-charide. J Bone Miner Res 2002;17:1211– 8.

7. Hong CY, Lin SK, Kok SH, Cheng SJ, Lee MS, Wang TM, et al. The role of lipopolysac-charide in infectious bone resorption of periapical lesion. J Oral Pathol Med2004;33:162–9.

8. Igarashi K, Woo JT, Stern PH. Effects of a selective cyclooxygenase-2 inhibitor, cele-coxib, on bone resorption and osteoclastogenesis in vitro. Biochem Pharmacol2002;63:523–32.

9. Wada N, Maeda H, Yoshimine Y, Akamine A. Lipopolysaccharide stimulates expres-sion of osteoprotegerin and receptor activator of NF-kappa B ligand in periodontalligament fibroblasts through the induction of interleukin-1 beta and tumor necrosisfactor-alpha. Bone 2004;35:629 –35.

0. Tanaka H, Tanabe N, Shoji M, Suzuki N, Katono T, Sato S, et al. Nicotine and lipo-polysaccharide stimulate the formation of osteoclast-like cells by increasing macro-phage colony-stimulating factor and prostaglandin E2 production by osteoblasts. LifeSci 2006;78:1733– 40.

1. Suda K, Woo JT, Takami M, Sexton PM, Nagai K. Lipopolysaccharide supports survivaland fusion of preosteoclasts independent of TNF-alpha, IL-1, and RANKL. J CellPhysiol 2002;190:101– 8.

2. Jiang Y, Mehta CK, Hsu TY, Alsulaimani FF. Bacteria induce osteoclastogenesis via anosteoblast-independent pathway. Infect Immun 2002;70:3143– 8.

3. Jiang J, Li H, Fahid FS, Filbert E, Safavi KE, Spangberg LS, et al. Quantitative analysisof osteoclast-specific gene markers stimulated by lipopolysaccharide. J Endod2006;32:742– 6.

4. Katagiri T, Takahashi N. Regulatory mechanisms of osteoblast and osteoclast differ-

entiation. Oral Dis 2002;8:147–59.

JOE — Volume 33, Number 4, April 2007