The Effects of Selective COX-2 Inhibitor/Celecoxib and Omega-3 Fatty Acid onMatrix Metalloproteinases, TIMP-1, andLaminin-5g2-Chain Immunolocalizationin Experimental PeriodontitisSaynur Vardar-Sengul,*† Eralp Buduneli,* Oya Turkoglu,* Nurcan Buduneli,* Gul Atilla,*Jaana Wahlgren,‡ Timo Sorsa,§ and Haluk Baylas*

Background: Matrix metalloproteinases (MMPs) play important roles in tissue-destruction mechanisms–associatedperiodontitis.MMP-8and-13are thepredominantcollagenases thatare important in theextracellularmatrix degradation in periodontal tissues. MMP-14 is a membrane-type MMP, whereas laminin-5 indicates basalmembrane modification and epithelial induction. The purpose of the present study was to evaluate the effects ofcelecoxib and omega-3 fatty acid administration on the gingival tissue expression of MMP-8, -13, and -14, tissueinhibitor of MMP (TIMP)-1, and laminin (Ln)-5g2-chain in rat experimental periodontitis induced by Escherichiacoli endotoxin (lipopolysaccharide [LPS]).

Methods: Experimental periodontitis was induced in rats by repeated LPS injection. Fifty-one adult maleSprague-Dawley rats were divided into six study groups: saline control, LPS, LPS + celecoxib, LPS + therapeuticomega-3 (TO3), prophylactic omega-3 + LPS + omega-3 (P+TO3), and LPS + celecoxib + omega-3 fatty acid.Celecoxibandomega-3 fattyacidweregivenasasingleagentorascombination therapy for14days.Onday15,all rats were sacrificed, and gingival tissues were analyzed immunohistochemically for the expression of MMP-8,-13, and -14, TIMP-1, and Ln-5g2-chain. Alveolar bone loss was evaluated morphometrically under a stereomi-croscope. Data were tested statistically by Kruskal-Wallis and Mann-Whitney tests and Spearman correlationanalysis.

Results: Alveolar bone loss was significantly higher in all study groups compared to the saline control group(all P <0.01). MMP-8 expression was significantly higher in the LPS group than in the saline group (P = 0.001).Very low expression of MMP-8 was found in the celecoxib, P+TO3, and combination groups. TO3 increasedTIMP-1 expression significantly compared to the LPS group (P <0.05). Individual celecoxib and P+TO3 admin-istration increased MMP-14 significantly compared to saline control and LPS groups (P <0.05). No significantdifferences were found among the study groups with regard to Ln-5g2-chain and MMP-13 expressions (P>0.05).

Conclusions: Selective cyclooxygenase-2 inhibitor, prophylactic omega-3 fatty acid, and a combination ofthese two agents can inhibit gingival tissue MMP-8 expression. Moreover, the individual administration of ther-apeutic omega-3 may increase gingival TIMP-1 expression in contrast to no effect on MMP-8, -13, and -14 ex-pressions in experimental periodontitis. These experimental findings in a rat model of LPS-induced periodontitisneed to be verified by clinical human studies. J Periodontol 2008;79:1934-1941.

KEY WORDS

Celecoxib; COX-2 inhibitor; laminin 5; matrix metalloproteinases; omega-3 fatty acid;periodontitis/drug therapy.

* Department of Periodontology, School of Dentistry, Ege University, _Izmir, Turkey.† Department of Pathology and Laboratory Medicine, School of Medicine, University of California, Irvine, CA.‡ Department of Central Pathology, Helsinki University Central Hospital, Helsinki, Finland.§ Department of Oral and Maxillofacial Diseases, Institute of Dentistry, Helsinki University Central Hospital.

doi: 10.1902/jop.2008.080001

Volume 79 • Number 10

1934

Periodontitis is a chronic inflammatory diseasecharacterized by loss of the tooth-supportingstructures, including connective tissue attach-

ment and alveolar bone. The inflammatory cascadestarts with the penetration of lipopolysaccharides(LPS) from the surface of Gram-negative bacteriainto periodontal tissues. LPS stimulates monocytes/macrophages to secrete inflammatory mediators, suchas prostaglandin E2 (PGE2), interleukin (IL)-1, -6, and-8, tumor necrosis factor-alpha (TNF-a), and matrixmetalloproteinases (MMPs), which, in turn, can triggervascular smooth muscle cells, fibroblasts, monocytes/macrophages, and osteoclasts to produce variousproteolytic enzymes and stimulate bone resorption.This leads to clinical inflammation and periodontalattachment loss. MMPs are also important in phys-iologic growth and tissue remodeling and can exertanti-inflammatory protective properties.1-3

MMPs form a group of proteases able to degrade al-most all extracellular matrix (ECM) and bone matrix(BM) components. MMPs are divided according to theirsubstrate specificities and structures into interstitialcollagenases, gelatinases, membrane-type MMPs,stromelysins, matrilysins, and other MMPs.3 MMP ex-pression is regulated by proinflammatory cytokinesand growth factors as well as ECM components. Colla-genases (MMP-1, -8, and -13) can break collagens aswell as laminins, and they are considered to be the keyMMPs responsible forECM and BMdestruction inmanytissue-destructive inflammatory and malignant patho-logic conditions.4 MMP-14 (membrane-type I MMP)can activate proMMP-2, -8, and -13 and is responsiblefor collagenous as well as non-collagenous ECM de-struction. Subepithelial fibroblasts in inflamed gingivaare the main cells capable of expressing MMP-14.MMP-14 has been found in cell membranes, in solubleor shedded forms, in periodontitis-affected gingival tis-sue, and gingival crevicular fluid.5

Tissue inhibitors of MMPs (TIMPs) are major endog-enous downregulators of MMPs.3 An imbalance be-tween MMPs and TIMPs results in and reflects theperiodontal tissue destruction.6,7 TIMP-1 is more effec-tive on interstitial collagenases.8,9 Laminin (Ln)-5 is animportant molecule associated with epithelial cell ad-hesion and migration.10,11 Ln-5 is localized in gingivawithin the basement membranes between the junc-tional epithelium and the tooth/or connective tissue.12

Ln-5 was shown to be colocalized with MMP-2 and -13within the basement membrane at the sites of peri-odontal inflammation,11,13 and MMP-8, -13, and -14can directly cleave Ln-5.14,15

Conventional periodontal therapy (scaling and rootplaning) is effective and sufficient to control diseaseprogression in chronic periodontitis. However, in somecases, adjunctive chemical therapy may be requiredwhen patients do not respond to mechanical periodon-

tal treatment. In this regard, host response modulationhas arisen as a novel treatment approach.16 Non-ste-roidal anti-inflammatory drugs, one of the host-modu-lation therapies, have been used to reduce increasedprostanoids in periodontitis sites.17-19 Recently, selec-tive cyclooxygenase (COX)-2 inhibitors were reportedto be capable of modifying the progression of peri-odontitisbyreducingalveolarbone loss.20,21 Inourpre-vious studies,22,23 we demonstrated that the combinedadministration of celecoxib and omega-3 fatty acidhad a synergistic effect, resulting in significant reduc-tions in the gingival tissue levelsofPGE2, prostaglandinF2a (PGF2a), leukotriene (LT)-B4, and platelet activat-ing factor in LPS-induced experimental periodontitis inrats. Prostaglandins were reported to regulate the MMPproduction in synoviocytes24 and prostate epithelialtumor cells.25 Noguchi et al.26 also reported that PGF2a

can regulate MMP-1 production by human gingival fi-broblasts. Considering the potential regulatory effectsof arachidonic acid metabolites on MMPs, we investi-gated possible individual and combined effects ofCOX-2 inhibitor/celecoxib and omega-3 fatty acid onthe collagenolytic MMP-8, -13, and -14 expressions ina rat model of experimental periodontitis.

Host modulation therapy is required to enable theclinician to decide when and how to use adjunctivemedications to modulate host response in the treat-ment of periodontal diseases. Therefore, the presentstudy was planned to evaluate the individual and com-bined effects of selective COX-2 inhibitor/celecoxiband/or omega-3 fatty acid on gingival tissue MMP-8,-13, and -14, TIMP-1, and Ln-5g2-chain expressionsin LPS-induced periodontitis in rats and to test whetherprophylactic omega-3 fatty acid administration hadan additional effect on these proteins levels.

MATERIALS AND METHODS

Study DesignThe experimental protocol of the present study was ap-proved by the Ethics Committee on Animal Experi-mentation at Ege University. Fifty-one adult maleSprague-Dawley rats (205 – 29.3 g) were purchasedfrom the University’s facilities and housed in tempera-ture-controlled rooms and received water and food adlibitum.

Induction of experimental periodontitis was pursuedby repeated injection of endotoxin (LPS) according tothe model described by Ramamurthy et al.27 The ratswere injected with 10 ml saline or Escherichia coliendotoxini (serotype 055:B5, L2637; 1 mg/ml) intothe palatal gingiva between the first and second maxil-lary molars and into the labial and oral gingiva betweenthe central incisors in the mandible and maxilla everyother day for 3 days. Twenty-four hours before initiating

i Sigma Chemical, St. Louis, MO.

J Periodontol • October 2008 Vardar-Sengul, Buduneli, Turkoglu, et al.

1935

drug treatment, the rats were injected with endotoxin/saline using an insulin syringe to which a 0.30 · 13-mmneedle was attached. The rats were distributed intosix experimental groups:

1) Saline controls (n = 9) = gingiva was injected withsaline, and each rat was orally gavaged daily withsaline solution for 14 days.

2)Endotoxin(LPS)(n=9)=gingivawasinjectedwithLPS, and each rat was orally gavaged daily with salinesolution.

3) Celecoxib (n = 8) = LPS + celecoxib: gingiva wasinjectedwithLPS,andeachratwasorallygavagedwithselective COX-2 inhibitor/celecoxib (4 mg/kg) daily.

4) Therapeutic omega-3 fatty acid (TO3) (n = 10) =LPS + omega-3 fatty acid: gingiva was injected withLPS,andeachratwasorallygavagedwithomega-3 fattyacid (40 mg/kg; 60% eicosapentaenoic acid [EPA] and40% docohexaenoic acid [DHA]) daily for 14 days.

5) Prophylacticplus therapeuticomega-3 fattyacid(P+TO3)(n =7) =omega-3+LPS +omega-3 fattyacid.Daily oral gavage with omega-3 fatty acid (40 mg/kg)was started 14 days before baseline, LPS injection intogingivawasstartedatbaseline,andomega-3fattyacidgavage was continued for 14 days after baseline.

6) Combination (n = 8) = LPS + celecoxib + omega-3fatty acid. Gingiva was injected with LPS and theneach rat was treated with a combination of celecoxibplus omega-3 fatty acid for 14 days.

All rats were anesthetized and sacrificed on day 15.The palatal gingival tissues from the maxillary molarsand labial and oral gingival tissue from the mandibularand maxillary central incisors were removed and fixedin phosphate buffered saline–buffered formalin. Theparaffin blocks were processed for immunohisto-chemical analysis. The whole heads were removedand frozen for subsequent morphometric evaluations.

Measurement of Alveolar Bone LossTheheadswereboiled for10minutes, soft tissueswereremoved manually, and the upper and lower jaws weredefleshed andcleaned. Theskullswere soaked in 0.2NNaOH at room temperature for 5 minutes to removethe remaining soft tissue debris. Then the jaws werewashed,airdried,andstainedwithaqueousmethyleneblue (1%) to identify the cemento-enamel junction(CEJ). The alveolar bone height was measured at 22sites in the left and right maxillary molar regions undera stereomicroscope (·40 magnification) by recordingthe distance from the CEJ to the alveolar bone crest.The histomorphometric measurement of alveolar boneloss was performed by a single examiner who wasmasked to the study groups.

Immunohistochemical AnalysisImmunolocalization of MMP-8, -13, and -14, TIMP-1,and Ln-5g2-chain were evaluated in formalin-fixed/paraffin-embedded gingival tissue sections. For im-

munostaining, APC peroxidase kits¶ were used asdescribed by Hanemaaijer et al.28 Six-micrometer for-malin-fixed paraffin sections were deparaffinized, hy-drated, and treated with 0.4% pepsin for 45 minutesandwith0.6%H2O2 inmethanol for30minutes.There-after, the sections were incubated with normal goatserum (1:50) for 3 hours, with polyclonal anti–MMP-8(1:250),28,29 anti–MMP-13 (1:100),4,30 anti–MMP-14(1:100),5 and anti–Ln-5g2-chain (1:50)31 overnightat +4�C, and then with biotinylated anti-rabbit immu-noglobulin G (1:250) for 1 hour and with avidin-biotincomplex (1:125) for 30 minutes. The sections werestained with 3-amino-9-ethycarbazole and counter-stained with Mayer’s hematoxylin. Then, the sectionswere stained with a chromogenic substrate kit for per-oxidase.# For double immunohistochemical staining,sections were incubated with polyclonal antibody forLn-5g2-chain. Ln-5g2-chain immunoreactivity wasvisualized with a chromogenic substrate kit for perox-idase,** which yields a dark blue-gray color. For neg-ative controls, the antibodies were replaced with 2 mg/ml rabbit immunoglobulin G as well as with 4 mg/mlmouse immunoglobulin G.

Immunoreactivity was semiquantified: 0 = no ex-pression; 1 = low expression (less than eight cells);2 = moderate expression (nine to 14 cells); 3 = intenseexpression(15 to20cells); and4 = very intenseexpres-sion (>20 cells).32

Statistical AnalysisStatistical analysis was performed using non-para-metric techniques. Comparisons among the six studygroups were performed using the Kruskal-Wallis test.In case of significant differences, post hoc, two-groupcomparisons were done with the Mann-Whitney Utest. The Spearman correlation analysis was used todetect correlations between immunohistochemicalparameters and alveolar bone loss. P values <0.05were considered statistically significant.

RESULTS

Alveolar Bone LossMeasurement of alveolar bone loss in the maxillarymolar teeth revealed significantly higher values inall study groups compared to the saline control group(P <0.05). Alveolar bone loss in treatment groups wasless than in the LPS group, but the differences did notreach statistical significance (P >0.05) (Fig. 1).

Immunohistochemical AnalysisExpression scores of the investigated immunohis-tochemical parameters are outlined in Table 1 andFigure 2. MMP-8 expression in the LPS and TO3

¶ Vectastain, ABC Elite Kits, Vector Laboratories, Burlingame, CA.# SG color, Vector.** SG color, Vector.

MMPs in Rats Administered Celecoxib and Omega-3 Volume 79 • Number 10

1936

groups was significantly increased compared to thatin the saline control group (P <0.01). TO3 group ex-hibited lower MMP-8 expression than the LPS group,but the difference was not statistically significant (P>0.05), whereas in the celecoxib, P+TO3, and com-bined treatment groups, the expression of MMP-8was very low and significantly lower (P <0.001) thanin the LPS group. MMP-13 expression was similar inall studied groups (P >0.05). The saline control,LPS, TO3, and combined drug treatment groups ex-hibited similar MMP-14 expressions (P >0.05).MMP-14 expression in the celecoxib and P+TO3groups was significantly higher than in the saline con-trol (both P = 0.004) and LPS (both P = 0.001) groups.

LPS injection resulted in a significant decrease inTIMP-1 expression compared to the saline controlgroup (P = 0.000). TO3 fatty acid administration re-sulted in similar TIMP-1 expression to that in the salinecontrol group (P = 0.719) and higher expression thanin the LPS, celecoxib, P+TO3, and combinationgroups (P = 0.001, P = 0.001, P = 0.002, and P =0.004, respectively). The combined drug treatment,celecoxib, and P+TO3 groups exhibited lower TIMP-1

expression than the saline control group (P = 0.002;P = 0.001; and P = 0.001, respectively) and similar ex-pression to the LPS group (P >0.05). No significant dif-ferences in Ln-5g2-chain expression (P >0.05) werefound among the study groups.

According to the Spearman correlation analysis,the amount of alveolar bone loss had a significant pos-itive correlation with MMP-8 expression (r = 0.559; P =0.000) and a negative correlation with TIMP-1 (r =-0.425; P = 0.004). There were significant correla-tions between TIMP-1 and MMP-14 (r = -0.385; P =0.01) as well as between Ln-5g2-chain and MMP-14expressions (r = -0.367; P = 0.014).

DISCUSSION

E. coli LPS injection into the gingiva in the presentstudy resulted in a significant amount of alveolar boneloss, further supporting and extending the validity ofthe experimental periodontitis model in rats describedby Ramamurthy et al.27 The present significant posi-tive correlations between MMP-8 and the amount ofalveolar bone loss and negative correlations betweenTIMP-1 and alveolar bone loss provide further supportfor the prominent roles of MMP-8 and TIMP-1 in peri-odontal tissue destruction.3 This is in agreement withthe findings demonstrating that bone cells can ex-press MMP-8,33 which can contribute to gingival tis-sue MMP-8 levels in the rat model of experimentalperiodontitis.

In vivo studies4-6,34-36 reported increased or un-changed expressions of MMPs and distinct findingswith respect to their localization in periodontitis-affected sites. In the present study, increased MMP-8expression was found in the LPS group compared tothe saline control, whereas MMP-13 expression wasnot affected significantly. Our findings are in agree-ment with and extend those of Llavaneras et al.,35

who used a corresponding experimental rat peri-odontitis model and reported increased MMP-8 and

Figure 1.Alveolar bone loss (mean – SD) in study groups. *Significant differencefrom all other groups (P <0.05).

Table 1.

Expression Scores of Studied Proteins (mean – SD)

Saline Control (n = 9) LPS (n = 9) Celecoxib (n = 8) TO3 (n = 10) P+TO3 (n = 7) Combination (n = 8)

MMP-8 0.22 – 0.44 1.56 – 0.73* 0.00 – 0.00† 1.10 – 0.56* 0.00 – 0.00† 0.00 – 0.00†

MMP-13 1.33 – 0.70 1.44 – 0.73 1.88 – 0.64 1.80 – 0.63 1.71 – 0.48 1.63 – 0.74

MMP-14 1.11 – 0.93 1.33 – 0.50 2.63 – 0.52*† 1.60 – 0.52 2.71 – 0.48*† 1.75 – 1.03

TIMP-1 2.22 – 0.67 0.56 – 0.53* 0.50 – 0.53*‡ 2.10 – 0.74† 0.86 – 0.37*‡ 0.75 – 0.71*‡

Ln-5g2-chain 1.78 – 0.67 1.67 – 0.50 1.25 – 0.46 1.50 – 0.70 1.14 – 0.38 1.25 – 0.45

Expressions of the proteins were semiquantified as follows: 0 = none; 1 = very mild; 2 = mild; 3 = moderate; 4 = abundant.* Significantly different from the saline control group (P <0.05).† Significantly different from the LPS group (P <0.05).‡ Significantly different from the TO3 group (P <0.05).

J Periodontol • October 2008 Vardar-Sengul, Buduneli, Turkoglu, et al.

1937

unchanged MMP-13 immunoreactivities in the LPSgroup as assessed by Western blot techniques. Veryfew investigations exist on MMP-14 expression in peri-odontitis. In this regard, we found similar MMP-13 and-14 immunoreactivity expression patterns in the LPSand saline control groups. Our findings are in agree-ment with the previous report34 that found similarMMP-14 expressions in inflamed human gingival tissueand healthy control tissues. Eventually, the degree ofMMP-13 and -14 activation may play a more promi-nent role than simply their expression levels in theinflamed gingival tissue. The lack of change in Ln-5g2-chain expression in the LPS group can be ex-plained, at least in part, by the unaffected MMP-14expression, because MMP-14 is one of the key enzymescapable of degrading or processing Ln-5g2-chain.14,15

There are varying data regarding the gingival tissueexpression of TIMP-1 in periodontitis compared tonormal gingiva.6,37-39 Our present findings demon-strating decreased TIMP-1 expression in the LPSgroup are in agreement with the studies by Soellet al.38 and Choi et al.39 Taken together, the decreasein gingival tissue TIMP-1 levels in concert with the sig-nificant increase in MMP-8 immunoreactivity in theLPS group in the present study supports the conjunc-ture that an imbalance between MMPs and their in-

hibitors results in connective tissue degradation inperiodontitis.3

Prostaglandins can regulate MMP production bysynoviocytes24 and prostate epithelial tumor cells.25

Sakaki et al.40 reported that the expression of MMP-1is regulated, in part, by COX-2 and its product PGE2.They showed that PGE2 induced the upregulationof MMP-1 (collagenase-1) mRNA levels, and COX-2inhibitor/indomethacin suppressed the IL-1b–inducedMMP-1 expression in gingival fibroblasts.40 Likewise,COX-2 inhibitor/NS-398 could also reduce IL-1b–and TNF-a–upregulated MMP-1 and -3 production,41

indicating that prostaglandins may be involved in theregulation of MMPs by gingival fibroblasts. Nishikawaet al.42 reported that in periodontal ligament fibro-blasts, NS-398 increased TNF-a–induced MMP-13mRNA levels whereas PGE2 decreased them. Thismay explain, in part, why COX-2 inhibitor/celecoxibincreased gingival tissue MMP-13 levels in the pres-ent study but decreased MMP-8 levels, indicating dis-tinct interactions between prostaglandin and MMPpathways further suggesting an MMP-type specificeffect.

Noguchi et al.26 reported that PGF2a could regulateMMP-1 production by human gingival fibroblasts. Wepreviously showed that PGE2, PGF2a, and LTB4 were

Figure 2.Immunohistochemical analyses of MMP-8, -13, and -14, TIMP-1, and Ln-5g2-chain expressions in the gingival tissue sections from the study groups.Representative figures are shown. (Original magnification, ·20.)

MMPs in Rats Administered Celecoxib and Omega-3 Volume 79 • Number 10

1938

reduced by omega-3 fatty acid administration aloneand in combination with celecoxib, and PGE2 was de-creased by celecoxib administration alone.22,23 In thepresent study, we found very low MMP-8 expression inthe celecoxib, prophylactic omega-3, and combina-tion groups, reflecting the complete inhibitory effectof these agents on MMP-8. Taken together, MMP-8 ex-pression might be regulated by COX-2 and/or lipox-ygenase (LOX) pathways in periodontitis. However,the administration of celecoxib and omega-3 fattyacid, alone or in combination, did not reduce the ex-pression of MMP-13 or -14. Therefore, it might be sug-gested that the efficacy of celecoxib and/or omega-3fatty acids in periodontal treatment may be related totheir inhibitory effects on MMP-8 and eicosanoidlevels associated with the increasing effect on TIMP-1,which thereby decreases the MMP-8/TIMP-1 ratio ininflamed gingival tissue.

In the present study, we did not detect an inhibitionof MMP-13 expression in gingival tissue with a 2-weekadministration of omega-3 fatty acid, but we did notea significant increase in TIMP-1 protein levels. A2-week administration of omega-3 fatty acid maybe too short to affect MMP-13 levels, but this coursewas sufficient to result in a TIMP-1 increase. The useof omega-3 fatty acid as a dietary supplement hasbeen increasing for various reasons; therefore, an eval-uation of clinical periodontal parameters, togetherwith gingival crevicular fluid levels of inflammatorymediators and MMP/TIMP ratio, in these long-termomega-3 fatty acid users may provide further in-sights into its adjunctive effects on the treatment ofperiodontitis.

Omega-3 fatty acids can play a key role in the pro-gression or prevention of human diseases by affectingcellular membrane lipid composition, metabolism,or signal transduction pathways or by direct controlof gene expression in a number of tissues and celltypes.43,44 The exact mechanisms by which omega-3 fatty acid mediate the inhibition of inflammationremains unclear. Alterations in arachidonic acid me-tabolism and eicosanoid pathway intermediates andthe induction of apoptosis and/or proinflammatorycytokine production are among the suggested mech-anisms.45 The present findings showed that omega-3fatty acid, as an adjunctive medication in periodontaltreatment, might be beneficial because of its inhibi-tory effect on MMP-8 and increasing effect on TIMP-1in concert with direct regulation of the eicosanoid path-way, which we demonstrated in previous studies.22,23

The prophylactic administration of omega-3 fattyacid for 2 weeks provides a maximum increase inthe levels of EPA and DHA in the cell membrane46

and eventually maintains the membrane stabilityand fluidity in the physiologic state. A cell membranewith these properties is more resistant to the attacks of

bacteria and their virulence factors.46 Therefore, inthe present study, we evaluated the possible effectsof prophylactic omega-3 fatty acid administered for2 weeks before the induction of experimental peri-odontitis and compared these effects to those of thera-peutic omega-3 fatty acid. The prophylactic use ofomega-3 acid exerted an inhibitory effect on MMP-8levels, but not on MMP-13 or -14.

Selective COX-2 inhibitors and omega-3 fattyacids are potent prostaglandin inhibitors, and theymay also have an indirect inhibitory effect on MMP-8.Accordingly, the present study revealed the mostsignificant MMP-8 inhibition in the celecoxib, prophy-lactic omega-3, and combined treatment groups.Therapeutic omega-3 had no effect on MMP-8 expres-sion, but it resulted in a significant increase in TIMP-1immunoreactivity. However, this was not the casefor MMP-13 and -14. It can be speculated that thismodel, which primarily evaluated the acute response,did not enable us to fully address the possible long-term effects of celecoxib and/or omega-3 fatty acidon the expression of MMP-13 and -14 in gingival tis-sue. Moreover, our findings indicated no significant ef-fect of drug treatment on Ln-5g2-chain expression. Inthis regard, we could not assess whether the investi-gated enzymes degraded the Ln-5g2-chain, becausethe immunohistochemical technique used could notspecifically determine the degraded fragment ofLn-5g2-chain in tissue. This may have resulted in un-changed gingival tissue expressions of Ln-5g2-chain,preventing the detection of any possible effects of thetested medications.

The reduction of MMP-8 expression in LPS-inducedinflamed gingiva by the studied therapeutic strategieswas effective, but not complete. Complete inhibitionor reduction of MMP-8 in periodontitis-affected gin-giva may not be desirable because MMP-8 exertsphysiologic and protective functions by processinganti-inflammatory cytokines and chemokines re-quired for sustained host defense.1-3,47

The lack of a statistically significant effect on alve-olar bone loss, despite significant changes in MMP-8and TIMP-1 expression, may be explained by the pre-ceding character of biochemical changes to the clin-ically detectable effects. It may be considered that 14days of celecoxib and/or omega-3 fatty acid adminis-tration, at this dosage, is not long enough to observetheir effects on alveolar bone loss. Longer periods ofdrug administration might be required for celecoxiband omega-3 fatty acid therapy to affect bone levels.

CONCLUSIONS

Selective COX-2 inhibitor/celecoxib, prophylacticomega-3 fatty acid, and the combination of thesetwo agents can inhibit pathologically excessive gingivaltissue MMP-8 expression. Moreover, the administration

J Periodontol • October 2008 Vardar-Sengul, Buduneli, Turkoglu, et al.

1939

of therapeutic omega-3 fatty acid alone resulted in asignificant increase in gingival TIMP-1 expression, butit had no inhibitory effect on MMP-8, -13, and -14 ex-pression in experimental periodontitis. The studiedadjunctive medications can beneficially reduce theMMP-8/TIMP-1 ratio in inflamed gingiva. These ex-perimental findings in a rat model of periodontitisneed to be verified in clinical human studies.

ACKNOWLEDGMENTS

This study was supported by grants from the Ege Uni-versity Research Foundation (2002 Disx 015), theAcademy of Finland, Helsinki, Finland, and the Hel-sinki University Central Hospital Research Founda-tion. Dr. Sorsa is listed as an inventor on four oralfluid biomarker/diagnostic patents. Drs. Vardar-Sengul,E. Buduneli, Turkoglu, N. Buduneli, Atilla, Wahlgren,and Baylas report no conflicts of interest related to thisstudy.

REFERENCES1. Gueders MM, Balbin M, Rocks N, et al. Matrix metallo-

proteinase-8 deficiency promotes granulocytic allergen-induced airway inflammation. J Immunol 2005;175:2589-2597.

2. Owen CA, Hu Z, Lopez-Otin C, Shapiro SD. Mem-brane-bound matrix metalloproteinase-8 on activatedpolymorphonuclear cells is a potent, tissue inhibitor ofmetalloproteinase-resistant collagenase and serpinase.J Immunol 2004;172:7791-7803.

3. Sorsa T, Tjaderhane L, Konttinen YT, et al. Matrixmetalloproteinases: Contribution to pathogenesis, di-agnosis and treatment of periodontal inflammation.Ann Med 2006;38:306-321.

4. Kiili M, Cox SW, Chen HY, et al. Collagenase-2 (MMP-8) and collagenase-3 (MMP-13) in adult periodontitis:Molecular forms and levels in gingival crevicular fluidand immunolocalisation in gingival tissue. J Clin Peri-odontol 2002;29:224-232.

5. Tervahartiala T, Pirila E, Ceponis A, et al. The in vivoexpression of the collagenolytic matrix metalloprotei-nases (MMP-2, -8, -13, and -14) and matrilysin (MMP-7)in adult and localized juvenile periodontitis. J Dent Res2000;79:1969-1977.

6. Aiba T, Akeno N, Kawane T, Okamoto H, Horiuchi N.Matrix metalloproteinases-1 and -8 and TIMP-1 mRNAlevels in normal and diseased human gingivae. Eur JOral Sci 1996;104:562-569.

7. Pozo P, Valenzuela MA, Melej C, et al. Longitudinalanalysis of metalloproteinases, tissue inhibitors ofmetalloproteinases and clinical parameters in gingivalcrevicular fluid from periodontitis-affected patients.J Periodontal Res 2005;40:199-207.

8. Birkedal-Hansen H. Role of matrix metalloproteinasesin human periodontal diseases. J Periodontol 1993;64:474-484.

9. Howard EW, Bullen EC, Banda MJ. Preferential inhi-bition of 72- and 92-kDa gelatinases by tissue inhibitorof metalloproteinases-2. J Biol Chem 1991;266:13070-13075.

10. Hormia M, Sahlberg C, Thesleff I, Airenne T. Theepithelium-tooth interface – A basal lamina rich in

laminin-5 and lacking other known laminin isoforms.J Dent Res 1998;77:1479-1485.

11. Pirila E, Maisi P, Salo T, Koivunen E, Sorsa T. In vivolocalization of gelatinases (MMP-2 and -9) by in situzymography with a selective gelatinase inhibitor. Bio-chem Biophys Res Commun 2001;287:766-774.

12. Atsuta I, Yamaza T, Yoshinari M, et al. Ultrastructurallocalization of laminin-5 (gamma2 chain) in the ratperi-implant oral mucosa around a titanium-dentalimplant by immuno-electron microscopy. Biomate-rials 2005;26:6280-6287.

13. Uitto VJ, Airola K, Vaalamo M, et al. Collagenase-3(matrix metalloproteinase-13) expression is inducedin oral mucosal epithelium during chronic inflamma-tion. Am J Pathol 1998;152:1489-1499.

14. Koshikawa N, Giannelli G, Cirulli V, Miyazaki K,Quaranta V. Role of cell surface metalloproteaseMT1-MMP in epithelial cell migration over laminin-5.J Cell Biol 2000;148:615-624.

15. Pirila E, Sharabi A, Salo T, et al. Matrix metalloprotei-nases process the laminin-5 gamma 2-chain andregulate epithelial cell migration. Biochem BiophysRes Commun 2003;303:1012-1017.

16. Kornman KS. Host modulation as a therapeutic strat-egy in the treatment of periodontal disease. Clin InfectDis 1999;28:520-526.

17. Abramson MM, Wolff LF, Offenbacher S, Aeppli DM,Hardie ND, Friedman HM. Flurbiprofen effect on gin-gival crevicular fluid prostaglandin and thromboxanelevels in humans. J Periodontal Res 1992;27:539-543.

18. Offenbacher S, Williams RC, Jeffcoat MK, et al. Effectsof NSAIDs on beagle crevicular cyclooxygenase me-tabolites and periodontal bone loss. J Periodontal Res1992;27:207-213.

19. Jeffcoat MK, Reddy MS, Haigh S, et al. A comparisonof topical ketorolac, systemic flurbiprofen, and pla-cebo for the inhibition of bone loss in adult periodon-titis. J Periodontol 1995;66:329-338.

20. Bezerra MM, de Lima V, Alencar VB, et al. Selectivecyclooxygenase-2 inhibition prevents alveolar boneloss in experimental periodontitis in rats. J Periodontol2000;71:1009-1014.

21. Holzhausen M, Rossa Junior C, Marcantonio Junior E,Nassar PO, Spolidorio DM, Spolidorio LC. Effect ofselective cyclooxygenase-2 inhibition on the develop-ment of ligature-induced periodontitis in rats. J Peri-odontol 2002;73:1030-1036.

22. Vardar S, Buduneli E, Turkoglu O, et al. Therapeuticversus prophylactic plus therapeutic administration ofomega-3 fatty acid on endotoxin-induced periodontitisin rats. J Periodontol 2004;75:1640-1646.

23. Vardar S, Buduneli E, Baylas H, Huseyinov Berdeli A,Buduneli N, Atilla G. Individual and combined effectsof selective cyclooxygenase-2 inhibitor and omega-3fatty acid on endotoxin-induced periodontitis in rats.J Periodontol 2005;76:99-106.

24. DiBattista JA, Martel-Pelletier J, Fujimoto N, Obata K,Zafarullah M, Pelletier JP. Prostaglandins E2 andE1 inhibit cytokine-induced metalloprotease expres-sion in human synovial fibroblasts. Mediation bycyclic-AMP signalling pathway. Lab Invest 1994;71:270-278.

25. Attiga FA, Fernandez PM, Weeraratna AT, ManyakMJ, Patierno SR. Inhibitors of prostaglandin synthesisinhibit human prostate tumor cell invasiveness andreduce the release of matrix metalloproteinases. Can-cer Res 2000;60:4629-4637.

MMPs in Rats Administered Celecoxib and Omega-3 Volume 79 • Number 10

1940

26. Noguchi K, Tominaga Y, Matsushita K, et al. Upregu-lation of matrix metalloproteinase-1 production byprostaglandin F2alpha in human gingival fibroblasts.J Periodontal Res 2001;36:334-339.

27. RamamurthyNS,GolubLM,GwinnettAJ,SaloT,DingY,Sorsa T. In vivo and in vitro inhibition of metallo-proteinases including MMP-13 by several chemicallymodified tetracyclines (CMTs). In: Davidovitch D, Mah J,eds. Biologic Mechanisms of Tooth Eruption, Resorptionand Replacement by Implants. Boston: Harvard Societyfor the Advancement of Orthodontics; 1998:271-277.

28. Hanemaaijer R, Sorsa T, Konttinen YT, et al. Matrixmetalloproteinase-8 is expressed in rheumatoid syno-vial fibroblast and endothelial cells. Regulation bytumor necrosis factor-alpha and doxycycline. J BiolChem 1997;272:31504-31509.

29. Sorsa T, Mantyla P, Ronka H, et al. Scientific basis ofmatrix metalloproteinase-8 specific chair-side test formonitoring periodontal and peri-implant health anddisease. Ann N Y Acad Sci 1999;878:130-140.

30. Lindy O, Konttinen YT, Sorsa T, et al. Matrix metallo-proteinase 13 (collagenase 3) in human rheumatoidsynovium. Arthritis Rheum 1997;40:1391-1399.

31. Pyke C, Salo S, Ralfkiaer E, Romer J, Dano K,Tryggvason K. Laminin-5 is a marker of invadingcancer cells in some human carcinomas and is coex-pressed with the receptor for urokinase plasminogenactivator in budding cancer cells in colon adenocarci-nomas. Cancer Res 1995;55:4132-4139.

32. Prikk K, Maisi P, Pirila E, et al. In vivo collagenase-2(MMP-8) expression by human bronchial epithelialcells and monocytes/macrophages in bronchiectasis.J Pathol 2001;194:232-238.

33. Sasano Y, Zhu JX, Tsubota M, et al. Gene expressionof MMP8 and MMP13 during embryonic developmentof bone and cartilage in the rat mandible and hindlimb. J Histochem Cytochem 2002;50:325-332.

34. Dahan M, Nawrocki B, Elkaim R, et al. Expression of ma-trix metalloproteinases in healthy and diseased humangingiva. J Clin Periodontol 2001;28:128-136.

35. Llavaneras A, Ramamurthy NS, Heikkila P, et al. Acombination of a chemically modified doxycycline anda biphosphonate synergistically inhibits endotoxin-induced periodontal breakdown in rats. J Periodontol2001;72:1069-1077.

36. Achong R, Nishimura I, Ramachandran H, HowellTH, Fiorellini JP, Karimbux NY. Membrane type (MT)1-matrix metalloproteinase (MMP) and MMP-2 ex-pression in ligature-induced periodontitis in the rat.J Periodontol 2003;74:494-500.

37. Ejeil AL, Igondjo-Tchen S, Ghomrasseni S, Pellat B,Godeau G, Gogly B. Expression of matrix metallo-

proteinases (MMPs) and tissue inhibitors of metallo-proteinases (TIMPs) in healthy and diseased humangingiva. J Periodontol 2003;74:188-195.

38. Soell M, Elkaim R, Tenenbaum H. Cathepsin C, matrixmetalloproteinases, and their tissue inhibitors in gin-giva and gingival crevicular fluid from periodontitis-affected patients. J Dent Res 2002;81:174-178.

39. Choi DH, Moon IS, Choi BK, et al. Effects of sub-antimicrobial dose doxycycline therapy on crevicularfluid MMP-8, and gingival tissue MMP-9, TIMP-1 andIL-6 levels in chronic periodontitis. J Periodontal Res2004;39:20-26.

40. Sakaki H, Matsumiya T, Kusumi A, et al. Interleukin-1beta induces matrix metalloproteinase-1 expressionin cultured human gingival fibroblasts: Role of cyclo-oxygenase-2 and prostaglandin E2. Oral Dis 2004;10:87-93.

41. Domeij H, Yucel-Lindberg T, Modeer T. Signal path-ways involved in the production of MMP-1 and MMP-3in human gingival fibroblasts. Eur J Oral Sci 2002;110:302-306.

42. Nishikawa M, Yamaguchi Y, Yoshitake K, Saeki Y.Effects of TNFa and prostaglandin E2 on the expres-sion of MMPs in human periodontal ligament fibro-blasts. J Periodontal Res 2002;37:167-176.

43. Ntambi JM, Bene H. Polyunsaturated fatty acid regu-lation of gene expression. J Mol Neurosci 2001;16:273-278.

44. Jump DB, Clarke SD. Regulation of gene expressionby dietary fat. Annu Rev Nutr 1999;19:63-90.

45. Tisdale MJ. Inhibition of lipolysis and muscle proteindegradation by EPA in cancer cachexia. Nutrition 1996;12:S31-S33.

46. Croft KD, Beilin LJ, Loece FM, Vandongen R. Effectsof diet enriched eicopentaenoic and docosahexaenoicacid or prostanoid metabolism in the rat. Lipids 1987;22:647-650.

47. Sorsa T, Golub LM. Is the excessive inhibition ofmatrix metalloproteinases (MMPs) by potent syn-thetic MMP inhibitors (MMPIs) desirable in periodonti-tis and other inflammatory diseases? That is: ‘Leaky’MMPIs vs excessively efficient drugs. Oral Dis 2005;11:408-409.

Correspondence: Dr. Saynur Vardar-Sengul, Departmentof Periodontology, School of Dentistry, Ege University,Bornova-35100, _Izmir, Turkey. Fax: 90-232-388-0325;e-mail: saynurvardar@yahoo.com.

Submitted January 2, 2008; accepted for publication April7, 2008.

J Periodontol • October 2008 Vardar-Sengul, Buduneli, Turkoglu, et al.

1941