J Periodontol • October 2004

Effectiveness of Adjunctive Low-DoseDoxycycline Therapy on ClinicalParameters and Gingival Crevicular FluidLaminin-5 γγ2 Chain Levels in ChronicPeriodontitisGülnur Emingil,* Gül Atilla,* Timo Sorsa,† Pirjo Savolainen,† and Haluk Baylas*

1387

Background: Laminin-5 (Ln-5) is involved in the apical migration of epithelial cells during the develop-ment of periodontal pockets. Low-dose doxycycline (LDD) can therapeutically modulate the host responsewith its non-antimicrobial properties. In the present randomized, double-blind, placebo-controlled, parallelarm study, the effectiveness of LDD in combination with non-surgical periodontal therapy on gingival crevic-ular fluid (GCF) Ln-5 γ2 chain fragment levels and clinical parameters in patients with chronic periodontitiswas examined over a 12-month period.

Methods: GCF samples were collected and clinical parameters including probing depth (PD), clinicalattachment level, gingival index (GI), and plaque index were recorded. Thirty chronic periodontitis patientswere randomized either to low-dose doxcycline or placebo groups. LDD group received doxycycline (20 mg,b.i.d.) for 3 months plus scaling and root planing (SRP), while placebo group was given placebo capsulesb.i.d. for 3 months plus SRP. The patients were evaluated every 3 months during the 12-month study period.All clinical parameters and GCF sampling were repeated at each visit. GCF Ln-5 γ2 chain fragment levelswere determined by Western immunoblotting using specific antibody and quantitated by computerized imageanalysis. Friedman test was used for intragroup comparisons followed by Wilcoxon signed rank test to ana-lyze significance of changes over time. The Mann-Whitney test was used to determine differences betweenboth LDD and placebo groups.

Results: Both groups revealed significant improvements in all clinical parameters over the 12-monthperiod (P <0.0125). LDD group showed a significantly greater reduction in the mean PD scores at 9 and12 months and in the mean GI scores at all time points than the placebo group (P <0.05). In the LDDgroup, GCF Ln-5 γ2 chain fragment levels were significantly reduced at 3 months (P <0.0125) and thenslightly increased during the rest of the study period. In the placebo group, GCF 45 and 70 kDa Ln-5 γ2chain fragments tended to decrease at 3 months compared to baseline, but did not reach significance; theselevels continued to increase throughout the remainder of the study period. GCF Ln-5 γ2 chain fragment lev-els in LDD group were significantly lower than those of the placebo group during the study period (P <0.05).

Conclusions: The present data indicate that LDD therapy in combination with SRP therapy can reduceGCF Ln-5 γ2 chain fragment levels and improve clinical periodontal parameters in patients with chronic peri-odontitis. Since matrix metalloproteinases (MMP)-mediated fragmentation of laminin-5 can contribute topocket formation by stimulating epithelial cell migration, the reduction of Ln-5 γ2 chain fragment levels couldprovide a new mechanism by which LDD, adjunctive to SRP, inhibits periodontal disease more effectivelythan SRP alone. Thus, these results provide extended and additional information about the effectiveness ofthe LDD therapy as an adjunct to non-surgical periodontal therapy in the long-term management of peri-odontal disease. J Periodontol 2004;75:1387-1396.

KEY WORDSAdjunctive therapy; cells, epithelial; clinical trials, randomized; comparison studies; double-blind method;doxycycline/therapeutic use; gingival crevicular fluid/analysis; laminin; metalloproteinases, matrix; periodontaldiseases/therapy; planing; scaling.

* Ege University, School of Dentistry, Department of Periodontology, Izmir, Turkey.† Department of Oral and Maxillofacial Diseases, Helsinki University Central Hospital (HUCH), Institute of Dentistry, University of Helsinki, Helsinki, Finland.

30218.qxd 10/14/04 8:52 AM Page 1387

1388

Low-Dose Doxycycline Therapy in Chronic Periodontitis Volume 75 • Number 10

Connective tissue breakdown in periodontitis resultsfrom the interaction of periopathogens and thedestructive host response.1 During host response,

inflammatory and immune mediators released from res-ident and inflammatory cells and from bacteria con-tribute to periodontal tissue destruction by degradingextracellular matrix (ECM) and basement membrane(BM) components.2 Thus, they lead to the conversionof the junctional epithelium (JE) into the periodontalpocket, which is involved in the progression of peri-odontal disease. Destructive mechanisms leading to theconversion of the gingival sulcus into an infected peri-odontal pocket play a crucial role in the progression ofperiodontal disease, but the precise mechanisms in thisdegeneration process have not been established yet.3

Laminins represent large molecular glycoproteins andmajor components of the basement membrane and exertmultiple structural and functional roles including cell dif-ferentiation, proliferation, adhesion, and migration.4,5 Cellmigration is an essential part of several physiological andpathogenic processes including chronic inflammatoryconditions, tumor invasion, and wound healing.5,6 It hasbeen suggested that laminins are involved in the apicalmigration of epithelial cells during the development ofperiodontal pockets.3 Laminin-5 is a well-known epithe-lial cell-derived adhesive protein localized to the anchor-ing filaments within the lamina lucida space of the basalmembrane zone of the junctional and gingival epithe-lium.7,8 It is secreted by epithelial cells and composedof three polypeptide chains α3, β3, and γ2 of which theγ chain is specific for laminin-5.4 While intact laminin-5serves to anchor epithelial cells on the basement mem-brane, it stimulates epithelial cell migration after havingbeen cleaved by proteolytic enzymes.9-11 Certain matrixmetalloproteinases (MMPs) can process Ln-5 γ2 chain,producing 45 to 70 kDa fragments.12-15 The cleavedLn-5 γ2 chain can direct inflammatory reactions by reg-ulating cell adhesion, migration and proliferation offibroblasts and epithelial cells, and also act as chemoat-tractant for leukocytes.6,11 Pirilä et al.14 have shown thatLn-5 γ2 chain was co-localized with MMPs within thebasement membrane area at the sites of periodontalinflammation. It was recently demonstrated that the pres-ence of the elevated Ln-5 γ2 chain fragments in diseasedperi-implant sulcular fluid (PISF) could reflect basementmembrane disruption caused by active inflammation.16

The management of periodontal disease includesmany treatment modalities such as conventional ther-apies involving non-surgical and surgical methods.Pharmacological agents have been advocated as adjunctsin the non-surgical treatment of periodontal infection,because of the infective nature of periodontitis.17 Tetra-cycline (TC) and TC analogs (CMTs) were shown tohave non-antimicrobial properties that therapeuticallymodulate the host response by inhibiting the mam-malian collagenase activity in gingival tissues and

gingival crevicular fluid (GCF) of chronic periodontitispatients.18-21 Through the inhibition of MMPs by TCs,both the connective tissue destruction and bone resorp-tion, which are key events in the pathogenesis of peri-odontal disease, decreased.20,22 The effectiveness ofLDD therapy in combination with non-surgical peri-odontal therapy was evaluated in several double-masked, placebo-controlled clinical studies in whichLDD was prescribed from 2 weeks to 12 months.23-30

These studies reported that a combined regimen ofLDD and non-surgical periodontal therapy resulted insignificant improvement in attachment levels and prob-ing depth in combination with a significant decreasein collagenase activity when compared to non-surgicalperiodontal therapy alone.24,25,28 It has also been re-ported that long-term usage of LDD did not induce thedevelopment of TC resistant microorganisms or exertany detrimental effect on the periodontal microflora.31-33

The aim of the present study was to examine theeffect of LDD in combination with non-surgical peri-odontal therapy on the GCF Ln-5 γ2 chain fragmentlevels as well as on clinical parameters in patients withchronic adult periodontitis over a 12-month period.

MATERIALS AND METHODSStudy PopulationThirty patients with chronic periodontitis (20 male and10 female; 37 to 61 years old) were enrolled from theDepartment of Periodontology, School of Dentistry, EgeUniversity, Izmir. The purpose and procedures werefully explained to all patients prior to participation, andpatients were entered into the study only after verbalconsent was obtained from each subject. The study wascarried out in accordance with the Helsinki Declaration.Complete medical and dental histories were taken fromall subjects. Patients were diagnosed with chronic peri-odontitis according to the clinical and radiographic find-ings and periodontal history.34 Study patients had atleast 14 natural teeth, at least eight pockets with ≥5 mmprobing depth (PD) and ≥4 mm clinical attachmentlevel (CAL), and radiographic evidence of moderate toadvanced chronic periodontitis. Subjects who had ahistory of systemic disease or a known hypersensitiv-ity to any type of tetracyclines were excluded, as werepatients who had received antibiotics or other medi-cines or periodontal treatment within the past 4 months,and pregnant or lactating women. None of the subjectswere heavy users of alcohol or tobacco. They weresmoking less than five cigarettes per day.

Study DesignThis clinical study was a randomized, double-blind,placebo-controlled, parallel-group study of 12 monthsduration. This study was organized into three phases:screening, treatment, and evaluation.

At the screening phase, the periodontal status of

30218.qxd 10/14/04 8:52 AM Page 1388

1389

J Periodontol • October 2004 Emingil, Atilla, Sorsa, Savolainen, Baylas

each patient was assessed by one examiner (GE). Thefull-mouth PD and CAL using the cemento-enamel junc-tion as the anatomical reference point were measuredwith a standard manual probe at six sites around eachtooth. The full-mouth gingival index (GI)35 and plaqueindex (PI)36 for each tooth were also recorded. In addi-tion, full-mouth, intraoral periapical radiographs weretaken. All clinical measurements were collected over thecourse of the study by the same examiner who wasunaware of the treatment provided. After being selectedfor the study, the subjects were randomized by the flipof a coin to either LDD or placebo groups.

GCF samples were collected at baseline examination.The treatment phase included non-surgical periodontaltreatment consisting of scaling and root planing (SRP),which was carried out in four to six sessions by the sameexaminer. Tooth and root surfaces were instrumentedunder local anesthesia up to 1 hour allowed per quadrantuntil they were free of all deposits as determined byvisual or tactile examination. Patients received oralhygiene instruction at each session, including tooth-brushing and the use of interdental flossing or interden-tal brushing as appropriate. The LDD group was givenlow-dose doxycycline (20 mg per capsule) b.i.d. for3 months plus SRP, while the placebo group receivedplacebo capsules (containing inactive filler; i.e., starch)b.i.d. for 3 months plus SRP. Patients were instructed totake study medication once in the morning and once inthe evening, 1 hour before eating. Study medicationswere identical in appearance. Compliance was encour-aged by distributing the capsules in labeled bottlesbiweekly. Adverse event information was collectedthroughout the study.

The evaluation phase consisted of four recall visits:recall 1, the last day of LDD or placebo therapy (i.e.,3 months into the study), and recalls 2, 3, and 4 at 3,6, and 9 months later. All clinical parameters (PD,CAL, PI, and GI and GCF sampling) were repeated ateach recall visit. Patients were also encouraged to con-tinue good oral hygiene and given regular maintenancetherapy (i.e., removal of any supragingival plaque andcalculus) at every recall appointment.

Gingival Crevicular Fluid SamplingGCF sampling was done before recording the clinicalparameters at each visit. GCF samples were collectedfrom mesio-buccal aspects of two single rooted teethexhibiting probing depth of 5 to 8 mm in differentquadrants in each subject. Prior to GCF sampling,supragingival plaque was removed from the inter-proximal surfaces with a sterile curet; these surfaceswere dried gently by an air syringe and were isolatedby cotton rolls. GCF was sampled with filter paper.‡

Paper strips were carefully inserted into the creviceuntil mild resistance was felt and left for 30 seconds.37

Care was taken to avoid mechanical injury. Strips con-

taminated with blood were discarded.38 The absorbedGCF volume of each strip was determined by an elec-tronic device§ and placed into a sterile polypropylenetube and kept at −40°C until analyzed. The readingsfrom the device were converted to an actual volume(µl) by reference to the standard curve.

Analysis of Ln-5 γγ2 Chain Fragments by WesternImmunoblottingWestern immunoblotting was used to detect Ln-5 frag-ments in GCF utilizing specific anti-Ln-5 γ2 chain anti-bodies.15,16 Ten µl samples of GCF strips extracted (for1 hour at 22°C) in 50 µl 50 mM Tris-HCL, 0.2 M NaCl,1.0 mM CaCl2, pH 7.8 were mixed with 5 µl of 2%β-mercaptoethanol and Laemmli’s buffer and heatedfor 5 minutes at 100°C, followed by protein separationwith 8% SDS-polyacrylamide gels. The protocol fol-lowed the manufacturer’s� instructions. A molecularanalysis scanning program¶ was used to quantitate theimmunoreactivities and molecular forms of Ln-5 γ2chain from the Western immunoblots. Intact and cleavedlaminin-5 γ2 chain were used as controls and in the cali-bration curve.15

Statistical AnalysisThe mean values of each subject’s whole-mouth clini-cal parameters were calculated. Mean values for PD,CAL, GI, and PI and GCF volume for two study sites perpatient were also averaged. The mean of Ln-5 γ2 chainfragment levels obtained from both study sites was alsocalculated and the patient was regarded as the unit ofanalysis. Intragroup comparisons of the GCF Ln-5 γ2chain fragment levels and the clinical parameters of thestudy sites, together with the mean values of the subjectwhole-mouth clinical parameters between baseline and3, 6, 9, and 12 months, were tested by the Friedmantest followed by Bonferroni corrected Wilcoxon signed-rank test to analyze significance of changes over time.The means of the changes from baseline for whole-mouth CAL scores in both study groups at each timepoint were also analyzed by Friedman test followed byBonferroni-corrected Wilcoxon signed-rank test, andP values <0.0125 were considered statistically significant.The comparison of both LDD and placebo groups wasperformed on per patient variables obtained at base-line and at 3, 6, 9, and 12 months. The Mann-Whitneytest was used to determine significant differences be-tween the LDD and placebo groups.

RESULTSOut of the 30 patients who entered the study, 20 com-pleted all examinations throughout the 12-month studyprotocol. Ten patients were exited due to their inability

‡ Periostrips, ProFlow, Inc., Amityville, NY.§ Periotron, ProFlow, Inc. � ECL System, Amersham Pharmacia Biotech, Buckinghamshire, U.K.¶ GS-700 Imaging Densitometer, Version 1.4, Biorad Laboratories, Hercules, CA.

30218.qxd 10/14/04 8:52 AM Page 1389

1390

Low-Dose Doxycycline Therapy in Chronic Periodontitis Volume 75 • Number 10

to attend the recall appointments. Random assignmentresulted in 10 patients in the LDD group and 10 patientsin the placebo group. Treatment with LDD was welltolerated and none of the patients reported any adverseeffects from the use of either LDD or placebo capsules.

Demographics for patients who completed the studyare shown in Table 1. The gender and mean ages weresimilar in both groups.

Clinical ResultsTable 2 shows the mean values of the subject whole-mouth clinical parameters for LDD and placebo groups.At baseline, there were no statistically significant dif-ferences between the study groups (P >0.05), except forCAL (P <0.05). Except for CAL, periodontal conditionsof both LDD and placebo groups markedly improvedbetween baseline and examinations at 3, 6, 9, and 12months (P <0.0125). The decreased PD and a sub-stantial reduction in GI and PI scores support this obser-vation. Both groups showed a similar reduction in PDat 3 months and 6 months. However, the LDD groupdemonstrated a significantly greater reduction in PD at9 months and 12 months (P <0.05). Improvement inwhole-mouth CAL scores was not significant in eithergroup at any time point compared to baseline (P >0.05),although the LDD group had more improvement thanthe placebo group at all time points. The improvementsin the whole-mouth CAL scores were similar for bothstudy groups for all time points (P >0.05). The LDDgroup showed statistically significant improvement inGI scores compared to placebo group at all time points(P <0.05). There was a significant reduction in the PIscores in both groups, with no significant differencesbetween them over the study period (P >0.05).

Figure 1 shows the mean PD scores of the study sitesfor both groups over the 12-month study period. Studysites in both LDD and placebo groups demonstratedsignificant improvement in PD at 3 months and thisimprovement was maintained over the course of thestudy (P <0.0125). The reduction in PD was similar forboth groups at 3 months and 6 months (P >0.05). How-ever, the LDD group showed asignificantly greater reduction inPD than the placebo group at 9months and 12 months (P <0.05).

The mean CAL scores of thestudy sites for LDD and placebogroups during the 12-monthstudy period are shown in Fig-ure 2. An improvement in CALwas observed in both groupswhen compared with baselinevalues over the course of thestudy (P <0.0125). LDD grouphad greater improvement in CALthan the placebo group, but the

differences were not significant for any time points(P >0.05). After 12 months, the mean gain in CAL was3 mm for the LDD group and 2.2 mm for the placebogroup, with no significant differences between both studygroups (P >0.05).

The mean GI, PI, and GCF scores of the study sitesare presented in Table 3. No statistically significant dif-ferences were found in the mean GI and PI scoresbetween LDD and placebo groups at baseline (P >0.05).Friedman test followed by Wilcoxon signed-rank testrevealed a significant improvement in GI and PI scoresas a result of both therapies from baseline to the 12-month examination (P <0.0125). LDD group showedstatistically significant improvement in GI scores com-pared to placebo group at 9 months and 12 months(P <0.05), but not at other time periods (P >0.05).Although there was a significant reduction in the PIscores in both groups, there were no significant differ-ences between them over the study period (P >0.05).

Ln-5 γγ2 Chain FragmentsLn-5 γ2 chain was predominantly present in GCF as45 and 70 kDa fragmented species (Fig. 3); theseimmunoreactivities below 80 kDa evidently represent

Table 1.

Demographic Characteristics of PatientsCompleting the Study

LDD Group Placebo Group

Males (N) 5 6

Females (N) 5 4

Mean age (years) 44.90 ± 5.06 47.80 ± 8.40

Age range 37-53 39-61

Current smokers (N) 3 5

Non-smokers (N) 7 5

Table 2.

Whole-Mouth Mean (±±SE) Clinical Parameters for Study Groups

LDD Group

Parameter Baseline 3 Months 6 Months 9 Months 12 Months

PD (mm) 3.62 ± 0.20 2.38 ± 0.09* 2.26 ± 0.10* 2.09 ± 0.12*† 2.03 ± 0.13*†

CAL (mm) 3.91 ± 0.21 3.58 ± 0.29 3.74 ± 0.32 3.62 ± 0.33 3.70 ± 0.32

GI 1.74 ± 0.15 0.54 ± 0.09*† 0.68 ± 0.08*† 0.49 ± 0.07*† 0.46 ± 0.10*†

PI 3.68 ± 0.31 1.97 ± 0.27* 2.42 ± 0.22* 2.17 ± 0.32* 1.64 ± 0.26*

* Significant difference from baseline (P <0.0125).† Significant difference from placebo group (P <0.05).

30218.qxd 10/14/04 8:52 AM Page 1390

than baseline (P = 0.0069, P = 0.0093, respectively)(Fig. 4).

The mean changes in the total amount of the GCFLn-5 γ2 fragments at 3, 9, and 12 months in the SRPplus LDD group were greater than those obtained bythe SRP plus placebo therapy (P = 0.0211, P = 0.0102,P = 0.0233, respectively), although at 6 months themean changes were similar between the two groups (P >0.05) (Fig. 4).

When the GCF Ln-5 γ2 chain fragments data wereexpressed as concentrations, a different pattern ofchange was seen than in the data expressed as totalamount (ng/site) (Fig. 5). No significant decrease wasfound in the GCF Ln-5 γ2 chain fragment concentrationin either group at any time point compared to baseline (P >0.0125). In fact, the concentration increased in bothgroups at all time points compared to baseline values.In the placebo group, this increase was found to be sig-nificant at 9 and 12 months (P = 0.0051, P = 0.0051,respectively), while in the LDD group it was significantonly at 12 months (P = 0.0069). The mean changes in theconcentration of GCF Ln-5 γ2 chain fragments obtainedby the SRP plus LDD therapy were greater than thoseobtained by the SRP plus placebo therapy at 9 months(P <0.05).

DISCUSSIONIn the present double-blind, placebo-controlled, parallelarm study, the effectiveness of low-dose doxycyclinetherapy (LDD) as an adjunct to scaling and root plan-ing was evaluated in chronic adult periodontitis patients.The effect of LDD on both clinical parameters and GCFLn-5 γ2 chain fragments levels compared to that obtainedby non-surgical periodontal therapy and placebo wasinvestigated.

The severity of the chronic periodontal disease wassimilar in both LDD and placebo groups at the begin-ning of the study. Significant improvements in clinicalparameters were observed in both groups. The effectof both therapies was evident at 3 months and thisimprovement, as indicated by PD and GI scores, was

maintained throughout the study period.This was not unexpected since consider-able time was spent in SRP in both groupsby an experienced clinician. As can beexpected, SRP led to the resolution of theinflammatory response and cessation ofthe periodontal disease progression, therebyresulting in a relative gain of clinical attach-ment and reduction of probing depth.39 SRPalso facilitates the elimination of a favorableenvironment for the colonization of periodon-topathogens.40 In the present study, afteractive periodontal therapy was completed,patients were evaluated every 3 months.41

Our results showed that improvements in

1391

J Periodontol • October 2004 Emingil, Atilla, Sorsa, Savolainen, Baylas

Figure 1.Effect of LDD plus SRP therapy or placebo plus SRP therapy on PD(mean ± SE) in study sites. *Significant difference from baseline afterLDD plus SRP therapy or SRP plus placebo therapy (P <0.0125).†Significant difference from placebo group (P <0.05).

Placebo Group

Baseline 3 Months 6 Months 9 Months 12 Months

3.97 ± 0.16 2.62 ± 0.14* 2.57 ± 0.14* 2.64 ± 0.18* 2.65 ± 0.17*

4.72 ± 0.25 4.50 ± 0.29 4.68 ± 0.23 4.69 ± 0.21 4.67 ± 0.23

1.91 ± 0.10 0.98 ± 0.13* 0.99 ± 0.12* 0.97 ± 0.14* 0.86 ± 0.14*

4.20 ± 0.22 2.27 ± 0.26* 2.29 ± 0.24* 2.17 ± 0.27* 2.25 ± 0.21*

fragmented molecular species of the Ln-5 γ2 chain byMMPs.11,15 After LDD plus SRP, the total amount ofGCF Ln-5 γ2 chain fragments significantly decreasedat 3 months compared to baseline (Fig. 4) (P =0.0125). The 6-month levels were similar to those at3 months. The Ln-5 γ2 chain fragment levels slightlyincreased at 9 and 12 months, with no significant dif-ferences compared to baseline (P >0.0125). Afterplacebo plus SRP therapy, total amount of GCF Ln-5γ2 chain fragments tended to decrease at 3 monthscompared to baseline, but did not reach significance(P >0.0125). At 6 months, GCF Ln-5 γ2 chain fragmentlevels were similar to the baseline values (P >0.0125),but were increased compared to 3 months and thistrend continued to the end of the study period (i.e.,levels at 9 and 12 months were significantly higher

30218.qxd 10/14/04 8:52 AM Page 1391

1392

Low-Dose Doxycycline Therapy in Chronic Periodontitis Volume 75 • Number 10

Figure 3.A representative Western immunoblot of Ln-5 γ2 chain immunoreactivities in low-dose doxycycline (LDD) and placebo (PL) GCF samples. 45 and70 kDa Ln-5 γ2 chain fragments and changes in their GCF levels during LDD plus SRP therapy and SRP plus placebo therapy are shown. Intact Ln-5γ2 chain is shown as a control. Mobilities of the molecular weight markers are shown to the left.

clinical parameters obtained in both groups during3 months of active periodontal therapy were maintaineduntil the end of the 12-month study. It should be notedthat the periodontal maintenance therapy, which includedsupragingival scaling and oral hygiene instruction, givenat each recall visit might result in markedly improvedoral hygiene conditions in both groups.

In the present study, the use of LDD in combinationwith SRP in chronic periodontitis patients provided moreclinical improvement than that obtained by SRP ther-apy alone. In particular, the significant decrease in GIscores at 3 months and PD reduction at 9 monthsobserved in the LDD group compared to placebo groupwas maintained until the end of the 12-month studyperiod.

In previous studies involving chronic periodontitispatients, LDD plus scaling and prophylaxis or SRPwas shown to result in significant reduction of probingdepths and improvement in clinical attachment levelabove those seen with non-surgical periodontal proceduresalone.26,27 Our findings thatLDD plus SRP has more clinicalbenefit than SRP alone in termsof PD reduction and clinicalattachment gain are in agree-ment with previous reports.24-27

In contrast, Walker et al.32 couldnot find any significant differ-ence between LDD and placebogroups in terms of clinicalattachment gain and probing

Figure 2.Effect of LDD plus SRP therapy or placebo plus SRP therapy on CAL(mean ± SE) of study sites. *Significant difference from baseline afterLDD plus SRP therapy or SRP plus placebo therapy (P <0.0125).

Table 3.

Clinical Parameters of Study Sites in Both Groups (mean ±± SE)

LDD Group

Parameter Baseline 3 Months 6 Months 9 Months 12 Months

GI 2.15 ± 0.16 0.55 ± 0.22* 0.85 ± 0.25* 0.65 ± 0.25*† 0.70 ± 0.24*†

PI 3.20 ± 0.49 1.85 ± 0.21* 2.00 ± 0.31* 1.90 ± 0.35* 1.60 ± 0.27*

GCF (µl) 0.58 ± 0.04 0.23 ± 0.06* 0.13 ± 0.02*† 0.19 ± 0.01* 0.15 ± 0.04*

* Significant difference from baseline (P <0.0125).† Significant difference from placebo group (P <0.05).

30218.qxd 10/14/04 8:52 AM Page 1392

Among BM components involved in cellular inter-actions, Ln-5 is likely to play an essential role duringcell migration.6 The presence of cleaved Ln-5 γ2 chainfragments in tissues was suggested to be important inactive, physiological, and pathological tissue remod-eling and in epithelial induction.9,10,16 In the presentstudy, the presence of 45 and 70 kDa Ln-5 γ2 chainfragments in periodontitis patients’ GCF strongly sug-gests the proteolytic processing of the intact Ln-5 byMMPs in periodontal disease,11,14,15 and correspondingcleaved fragments have recently been reported in peri-implantitis sulcular fluid.16 The cleaved Ln-5 γ2 chainin GCF can originate from the degraded basementmembrane between JE and connective tissue becausethe JE cells can express Ln-5 mRNA, while in otherparts of gingival epithelium the Ln-5 expression is lessprominent.7,8 Ln-5 is regarded as the main cell-matrixattachment factor for human keratinocytes and theenhanced fragmentation of Ln-5 γ2 chain by MMPs11,14

stimulate epithelial cell migration; i.e., pocketformation.3,10 Figueredo and Gustafsson47

examined laminins in GCF of patients withperiodontal disease and gingivitis and foundelevated total laminin levels in GCF from peri-odontitis patients. However, in this studylaminin isoforms were not identified.47 Theauthors suggested that activated neutrophilsseemed to be generating extensive destruc-tion of the basement membrane.47 It has alsobeen shown that elevated MMP-8 and Ln-5 γ2chain 45 and 70 kDa fragments in peri-implant

1393

J Periodontol • October 2004 Emingil, Atilla, Sorsa, Savolainen, Baylas

Figure 4.GCF Ln-5 γ2 chain fragment total amount (ng/sample) of both LDDand placebo groups from baseline to 12 months. Each valuerepresents mean values of both study sites in each patient. *Significantdifference from baseline after LDD plus SRP therapy or SRP plusplacebo therapy (P <0.0125). †Significant difference from placebogroup (P <0.05).

Figure 5.GCF Ln-5 γ2 chain fragment concentration (ng/µl) of LDD andplacebo groups from baseline to 12 months. Each value representsmean values of both study sites in each patient. *Significant differencefrom baseline after LDD plus SRP therapy or SRP plus placebo therapy(P <0.0125). †Significant difference from placebo group (P <0.05).

Placebo Group

Baseline 3 Months 6 Months 9 Months 12 Months

2.05 ± 0.05 1.30 ± 0.26* 1.30 ± 0.20* 1.25 ± 0.13* 1.20 ± 0.18*

4.15 ± 0.27 2.20 ± 0.40* 1.90 ± 0.35* 2.20 ± 0.48* 2.10 ± 0.52*

0.50 ± 0.05 0.27 ± 0.03* 0.29 ± 0.02* 0.15 ± 0.02* 0.29 ± 0.06*

depth reduction. Caton et al.27 found a statisticallysignificant difference of around 0.5 mm in both prob-ing depth reduction and gain of clinical attachmentbetween LDD and placebo groups at tooth sites withsevere disease, which they observed was a relativelysmall mean change. The clinical importance of a0.5 mm change has to be taken into considerationin evaluating the clinical significance of the data.42

However, it has also been stated that even small gainsin clinical attachment and reductions in probing depthcould be important in deciding if surgical interven-tion was needed.43 Furthermore, adjunctive therapycould have an advantage of reducing the expenseand discomfort of surgery to the patient.17 We sug-gest that such clinical improvement in the LDD groupmight not only be attributed to the effectiveness ofnon-surgical periodontal therapy, but also be due tothe long-term beneficial host modulatory effect ofdoxycycline.44-46

30218.qxd 10/14/04 8:52 AM Page 1393

1394

Low-Dose Doxycycline Therapy in Chronic Periodontitis Volume 75 • Number 10

sulcular fluid could reflect the active phase of the inflam-matory peri-implant disease.16

GCF Ln-5 γ2 chain fragment levels showed markedreductions at 3 months where active periodontal ther-apy was performed in combination with LDD. Thereafter,no additional decrease was observed, but these levelsremained constant until the end of the study period.GCF Ln-5 γ2 chain fragment levels showed only a smalldecrease after SRP plus placebo therapy, and these lev-els gradually increased until the end of the study period.Our data have shown that 3-month usage of LDD as anadjunct to SRP could result in lower GCF Ln-5 γ2 chainfragment levels compared to placebo group. Low GCFLn-5 γ2 chain fragment levels in the LDD group duringactive periodontal therapy might suggest the effective-ness of non-surgical periodontal therapy plus LDDtherapy in reducing not only the bacterial but also inflam-matory load in the periodontal environment. Based onthese findings, it can be suggested that decreasedGCF Ln-5 γ2 chain fragment levels in LDD group couldbe dependent on the host modulatory properties ofLDD.44-46 LDD therapy might also provide beneficialeffect on Ln-5 γ2 chain fragments of the BM. More-over, the properties of doxycycline to bind to the cal-cified tooth surfaces and its gradual release in activeform might help to maintain low GCF Ln-5 γ2 chainfragments levels compared to placebo group.48,49 Ourstudy is novel in showing the effects of LDD therapyin combination with non-surgical periodontal therapyon GCF Ln-5 γ2 chain fragments levels. We were notable to compare our results with other studies becausewe were not able to find any other report investigat-ing the effect of LDD therapy on GCF Ln-5 γ2 chainfragments levels in periodontal disease. However, arecent study investigating the efficacy of estrogen andCMT-8 on wound healing in ovariectomized rats foundthat Ln-5 γ2 chain expression decreased after CMT-8therapy.50 This finding suggests that CMT-8 treatmentcan promote wound healing by recovering the expres-sion of key molecules in wound healing, including theLn-5 γ2 chain.

In the present study, decreased GCF Ln-5 γ2 chainfragments levels after LDD plus SRP therapy could bedue to the promotion of the wound healing by LDD.In previous studies, LDD therapy in combination withnon-surgical periodontal therapy was shown to inhibitthe activity or down-regulate the expression of hostMMPs by a mechanism unrelated to the antimicrobialefficacy of the drug.20,24,25,28 During the host inflam-matory response, many rather than single MMPs cancleave Ln-5.11,12,14 These MMPs have the ability tocontribute to pocket formation as well as epithelial dis-integration through degradation of the internal basallamina, which acts as a barrier against infectiveagents.3 Therefore, LDD therapy might decrease GCF

Ln-5 γ2 chain fragment levels by inhibiting the destruc-tive GCF MMP levels to produce Ln-5 γ2 chain fragmentsin periodontal tissues. In this regard, we have previouslyshown that LDD therapy in combination with SRP ther-apy reduced GCF MMP-8 levels and improved clinicalperiodontal parameters of patients with chronic peri-odontitis.51 Thus, our present findings reporting on thereduction in the levels of Ln-5 γ2 chain breakdownproducts might represent an additional new mechan-ism by which LDD therapy, adjunctive to SRP, inhibitsthe progression of periodontal disease more effectivelythan SRP alone.

In conclusion, the present data indicate that adjunc-tive LDD in combination with SRP therapy reducesGCF Ln-5 γ2 chain fragment levels and improves clin-ical periodontal parameters in patients with chronicadult periodontitis. These results might provide addi-tional information about the usefulness of the LDDtherapy as an adjunct to non-surgical periodontal ther-apy in the long-term management of chronic adultperiodontal disease.

ACKNOWLEDGMENTSThis work was supported by grants from The FinnishTechnology Development Center (TEKES), Finnish Den-tal Society Apollonia, Helsinki University Central Hos-pital Research (grants HUCH-EVO-TI 020 Y 0002 andTYH 4113), Willhelm and Else Stockman Foundation(Helsinki), and Helsinki University Research Funds.

REFERENCES1. Listgarten MA. Nature of periodontal diseases: Patho-

genic mechanisms. J Periodontal Res 1987;22:172-178.2. Birkedal-Hansen R. Role of cytokines and inflammatory

mediators in tissue destruction. J Periodontal Res 1993;28:500-510.

3. Pöllänen MT, Salonen JI, Uitto VJ. Structure and func-tion of the tooth-epithelial interface in health and disease.Periodontology 2000 2003;31:12-31.

4. Colognato H, Yurchenco PD. Form and function: The lamininfamily of heterotrimers. Dev Dyn 2000;218:213-234.

5. O’Toole EA. Extracellular matrix and keratinocyte migra-tion. Clin Exp Dermatol 2001;26:525-530.

6. Terranova VP, Lyall RM. Chemotaxis of human gingivalepithelial cells to laminin. A mechanism for epithelialcell apical migration. J Periodontol 1986;57:311-317.

7. Gürses N, Thorup AK, Reibel J, Carter GW, HolmstrupP. Expression of VLA-integrins and their related basementmembrane ligands in gingiva from patients of variousperiodontitis categories. J Clin Periodontol 1999;26:217-224.

8. Hormia M, Sahlberg C, Thesleff I, Airenne T. The epithe-lium-tooth interface—a basal lamina rich in laminin-5 andlacking other known laminin isoforms. J Dent Res 1998;77:1479-1485.

9. O’Toole EA, Marinkovich MP, Hoeffler WK, Furthmayr H,Woodley DT. Laminin-5 inhibits human keratinoytemigration. Exp Cell Res 1997;233:330-339.

10. Gagnoux-Palacios L, Allergra M, Spirito F, et al. The shortarm of the laminin γ2 chain plays a pivotal role in the incor-

30218.qxd 10/14/04 8:52 AM Page 1394

1395

J Periodontol • October 2004 Emingil, Atilla, Sorsa, Savolainen, Baylas

poration of laminin 5 into the extracellular matrix and in celladhesion. J Cell Biol 2001;153:835-849.

11. Giannelli G, Falk-Marzillier J, Schiraldi O, Stetler-StevensonWG, Quaranta V. Induction of cell migration by matrix met-alloproteinase-2 cleavage of laminin-5. Science 1997;277:225-228.

12. Heck LW, Blackburn WD, Irwin MH, Abrahamson DR.Degradation of basement membrane laminin by humanneutrophil elastase and cathepsin G. Am J Pathol 1990;136:1267-1274.

13. Mäkelä M, Larjava H, Pirilä E, et al. Matrix metallopro-teinase 2 (gelatinase A) is related to migration of ker-atinocytes. Exp Cell Res 1999;251:67-78.

14. Pirilä E, Maisi P, Salo T, Koivunen E, Sorsa T. In vivo local-ization of gelatinases (MMP-2 and -9) by in situ zymogra-phy with a selective gelatinase. Biochem Biophys ResCommun 2001;287:766-774.

15. Pirilä E, Sharabi A, Salo T, et al. Matrix metalloproteinasesprocess the laminin-5 γ2-chain and regulate epithelial cellmigration. Biochem Biophys Res Commun 2003;303:1012-1017.

16. Kivela-Rajamaki MJ, Teronen OP, Maisi P, et al. Laminin-5γ2-chain and collagenase-2 (MMP-8) in human peri-implant sulcular fluid. Clin Oral Implants Res 2003;14:158-165.

17. Loesche WJ. The antimicrobial treatment of periodontaldisease: Changing the treatment paradigm. Crit Rev OralBiol Med 1999;10:245-275.

18. Golub LM, Wolff M, Lee HM, et al. Further evidence thattetracyclines inhibit collagenase activity in human crevic-ular fluid and from other mammalian sources. J Periodon-tal Res 1985;20:12-23.

19. Golub LM, McNamara TF, D’Angelo G, Greenwald RA,Ramamurthy NS. A non-antibacterial chemically-modifiedtetracycline inhibits mammalian collagenase activity. JDent Res 1987;66:1310-1314.

20. Golub LM, Ciancio S, Ramamurthy NS, Leung M, McNamaraTF. Low-dose doxycycline therapy: Effect on gingival andcrevicular fluid collagenase activity in humans. J PeriodontalRes 1990;25:321-330.

21. Golub LM, Wolff M, Roberts S, Lee HM, Leung M, PayonkGS. Treating periodontal diseases by blocking tissue-destructive enzymes. J Am Dent Assoc 1994;125:163-171.

22. Golub LM, Evans RT, McNamara TF, Lee HM, RamamurthyNS. A non-antimicrobial tetracycline inhibits gingival matrixmetalloproteinases and bone loss in Porphyromonas gin-givalis–induced periodontitis in rats. Ann NY Acad Sci1994;732:96-111.

23. Caton JG. Evaluation of Periostat for patient management.Compend Contin Educ Dent 1999;20:451-463.

24. Crout RJ, Lee HM, Schroeder K, et al. The “cyclic” reg-imen of low-dose doxycycline for adult periodontitis: Apreliminary study. J Periodontol 1996;67:506-514.

25. Golub LM, Lee HM, Greenwald RA, et al. A matrix metal-loproteinase inhibitor reduces bone-type collagen degra-dation fragments and specific collagenases in gingivalcrevicular fluid during adult periodontitis. Inflamm Res 1997;46:310-319.

26. Ciancio S, Ashley R. Safety and efficacy of sub-antimi-crobial-dose doxycycline therapy in patients with adultperiodontitis. Adv Dent Res 1998;12:27-31.

27. Caton JG, Ciancio SG, Blieden TM, et al. Treatment withsubantimicrobial dose doxycycline improves the efficacyof scaling and root planing in patients with adult peri-odontitis. J Periodontol 2000;71:521-532.

28. Golub LM, McNamara TF, Ryan ME, et al. Adjunctive treat-ment with subantimicrobial doses of doxcycline: Effects ongingival fluid collagenase activity and attachment loss inadult periodontitis. J Clin Periodontol 2001;28:146-156.

29. Caton JG, Ciancio SG, Blieden TM, et al. Subantimi-crobial dose doxycycline as an adjunct to scaling androot planing: Post-treatment effects. J Clin Periodontol2001;28:782-789.

30. Novak MJ, Johns LP, Miller RC, Bradshaw MH. Adjunctivebenefits of subantimicrobial dose doxycycline in the man-agement of severe, generalized, chronic periodontitis. J Peri-odontol 2002;73:762-769.

31. Thomas JG, Metheny RJ, Karakiozis JM, Wetzel JM, CroutRJ. Long-term sub-antimicrobial doxycycline (Periostat)as adjunctive management in adult periodontitis: Effectson subgingival bacterial population dynamics. Adv DentRes 1998;12:32-39.

32. Walker C, Thomas J, Nango S, Lennon J, Wetzel J, PowalaC. Long-term treatment with subantimicrobial dose doxy-cycline exerts no antibacterial effect on the subgingivalmicroflora associated with adult periodontitis. J Periodontol2000;71:1465-1471.

33. Thomas J, Walker C, Bradshaw M. Long-term use of sub-antimicrobial dose doxycycline does not lead to changesin antimicrobial susceptibility. J Periodontol 2000;71:1472-1483.

34. Armitage GC. Development of a classification system forperiodontal diseases and conditions. Ann Periodontol1999;4:1-7.

35. Löe H, Silness J. Periodontal disease in pregnancy. I.Prevalence and severity. Acta Odontol Scand 1963;21:532-551.

36. Turesky S, Gilmore ND, Glickman I. Reduced plaqueformation by the chloromethyl analogue of vitamin C. JPeriodontol 1970;41:41-43.

37. Lamster IB, Hartley LJ, Oshrain RL. Evaluation andmodification of spectrophotometric procedures foranalysis of lactate dehydrogenase, beta-glucuronidaseand arylsulphatase in human gingival crevicular fluidcollected with filter-paper strips. Arch Oral Biol 1985;30:235-242.

38. Cimasoni G. Method of collection. Crevicular FluidUpdated. Basel: S. Karger; 1983;29-36.

39. Greenstein G. Periodontal response to mechanical non-surgical therapy: A review. J Periodontol 1992;63:118-130.

40. Socransky SS, Haffajee AD. Effect of therapy onperiodontal infections. J Periodontol 1993;64:754-759.

41. The American Academy of Periodontology. Parameteron periodontal maintenance. J Periodontol 2000;71:849-850.

42. Greenstein G, Lamster IB. Efficacy of subantimicrobialdosing with doxycycline. Point/counterpoint. J Am DentAssoc 2001;132:457-465.

43. Killoy WJ. The use of locally-delivered chlorhexidine inthe treatment of periodontitis. Clinical results. J Clin Peri-odontol 1998;25:953-958.

44. Golub LM, Lee HM, Ryan ME, Giannobile WV, Payne J,Sorsa T. Tetracyclines inhibit connective tissue break-down by multiple non-antimicrobial mechanisms. AdvDent Res 1998;12:12-26.

45. Golub LM, Ramamurthy NS, McNamara TF. Tetracyclinesinhibit connective tissue breakdown: New therapeuticimplications for an old family of drugs. Crit Rev Oral BiolMed 1991;2:297-322.

30218.qxd 10/14/04 8:52 AM Page 1395

1396

Low-Dose Doxycycline Therapy in Chronic Periodontitis Volume 75 • Number 10

46. Golub LM, Suomalainen K, Sorsa T. Host modulation withtetracyclines and their chemically modified analogues.Curr Opin Dent 1992;2:80-90.

47. Figueredo CMS, Gustafsson A. Increased amounts oflaminin in GCF from untreated patients with periodon-titis. J Clin Periodontol 2000;27:313-318.

48. Demirel K, Baer PN, McNamara TF. Topical applicationof doxycycline on periodontally involved root surfacesin vitro: Comparative analysis of substantivity on cemen-tum and dentin. J Periodontol 1991;62:312-316.

49. Michalowicz BS, Pihlstrom BL, Drisko CL, et al. Evaluationof periodontal treatments using controlled-release tetracy-cline fibers: Maintenance response. J Periodontol 1995;66:708-715.

50. Pirilä E, Parikka M, Ramamurthy NS, et al. Chemicallymodified tetracycline (CMT-8) and estrogen promotewound healing in ovariectomized rats: Effects on matrixmetalloproteinase-2, membrane type 1 matrix metallo-proteinase, and laminin-5 γ2-chain. Wound Rep Reg 2002;10:38-51.

51. Emingil G, Atilla G, Sorsa T, Luoto H, Kirilmaz L, Baylas H.The effect of adjunctive low-dose doxyccline therapy onclinical parameters and gingival crevicular fluid matrixmetalloproteinase-8 levels in chronic adult periodontitis.J Periodontol 2004;75:106-115.

Correspondence: Dr. Gülnur Emingil, Ege Üniversitesi,Dishekimligi Fakültesi, Periodontoloji Anabilim Dali, Bornova-35100, Izmir, Turkey. Fax: 90-232-3880325; e-mail: gemingil@yahoo.com.

Accepted for publication February 12, 2004.

30218.qxd 10/14/04 8:52 AM Page 1396