Anti-inflammatory doc effects of Lactobacillus brevis CD2 on periodontal disease

ORIGINAL ARTICLE

Anti-inflammatory effects of Lactobacillus brevis (CD2) onperiodontal disease

DN Della Riccia1, F Bizzini

1, MG Perilli

2, A Polimeni

3, V Trinchieri

4, G Amicosante

2and MG Cifone

1

1Department of Experimental Medicine, University of L’Aquila, L’Aquila, Italy; 2Department of Biomedical Sciences andTechnologies, University of L’Aquila, L’Aquila, Italy; 3Department of Odontostomatologic Sciences, University �La Sapienza’, Rome,Italy; 4Department of Infectious and Tropical Diseases, University �La Sapienza’, Rome, Italy

OBJECTIVES: To analyze the anti-inflammatory effects

of Lactobacillus brevis extracts on periodontitis patients

and to investigate the involved mechanisms in vitro on

activated macrophages.

METHODS: Eight healthy subjects and 21 patients with

chronic periodontitis were enrolled to analyze the effect

of L. brevis-containing lozenges on periodontitis-associ-

ated symptoms and signs. Before and after the treat-

ment, the patients received a complete periodontal

examination. Saliva samples, collected before and after

treatment, were analyzed for metalloproteinase and ni-

tric oxide synthase (NOS) activity, immunoglobulin-A

(IgA), prostaglandin E2 (PGE2) and c-interferon (IFN-c)

levels. Arginine deiminase (AD) and NOS activities were

determined through a radiometric assay. Metallopro-

teinases were assayed by zymogram and Western blot-

ting, whereas IgA, PGE2 and IFN-c were assayed by

enzyme-linked imunosorbent assay tests.

RESULTS: The treatment led to the total disappearance

or amelioration of all analyzed clinical parameters in all

patients. This was paralleled to a significant decrease of

nitrite/nitrate, PGE2, matrix metalloproteinase, and IFN-

c levels in saliva samples.

CONCLUSION: Our results suggest that the anti-

inflammatory effects of L. brevis could be attributed to

the presence of AD which prevented nitric oxide gen-

eration. Our findings give further insights into the

knowledge of the molecular basis of periodontitis and

have a potential clinical significance, giving the experi-

mental ground for a new innovative, simple and effica-

cious therapeutical approach of periodontal disease.

Oral Diseases (2007) 13, 376–385

Keywords: L. brevis; nitric oxide; metalloproteinases; prostaglan-

din E2; c-interferon

Introduction

Lactic acid bacteria (LAB) belong to a variety of generaused for milk fermentation. The primary metabolic endproduct from carbohydrate metabolism is lactic acid,which in turn preserves milk by providing the aciditynecessary for a tart flavor and for changes in thestructure of casein to achieve syneresis and desiredfunctional characteristics. The extensive knowledge onLAB has opened new possibilities for their application(for reviews see Konings et al, 2000; Caglar et al, 2005).Reports of many investigators have confirmed that LABand their products have beneficial effects on the healthof animals and humans, i.e., protection against entericinfections, use as an oral adjuvant, immunopotentiationin malnutrition, and prevention of chemically inducedtumors (Bodana and Rao, 1990; Roberfroid, 1998;Dugas et al, 1999; Gorbach, 2000). Almost all genera ofLAB are able to produce bacteriocins that are primarilylethal to other strains and species of bacteria form short-lived pores in biologic membranes by interactions withdifferent compounds including nisin, lipid II andphospholipids, thereby killing the target bacteria (Kon-ings et al, 2000). In the attempt to better characterize themetabolic features of some LAB strains, our group hasrecently shown the presence of high levels of argininedeiminase (AD) in Lactobacillus brevis (Di Marzio et al,2001). As bacterial AD may compete with nitric oxidesynthase (NOS) by using the same substrate, arginine,we decided to investigate the ability of L. brevis extractsto inhibit NOS activity as well other inflammatoryparameters [c-interferon (IFN-c), prostaglandin E2

(PGE2) and metalloproteinases], known to be associatedwith NOS induction (Cifone et al, 1999; Goodwin et al,1999; Di Marzio et al, 2001; Ishii et al, 2003). Inparticular, the chronic periodontal disease was chosenas the known inflammatory model to test the potentialin vivo anti-inflammatory effects of L. brevis.

Periodontal disease is a major cause of tooth loss inhumans and is one of the most prevalent diseasesassociated with bone loss. Following bacterial coloniza-tion, the gingiva becomes inflamed leading, in some cases,

Corresponding Author: Dr MG Cifone, Department of ExperimentalMedicine, University of L’Aquila, Via Vetoio, 10, Coppito 2, 67100L’Aquila, Italy. Tel: 0862 433503, Fax: 0862 433523, E-mail: [email protected] 19 November 2006; revised 9 January 2006; accepted 30January 2006

Oral Diseases (2007) 13, 376–385. doi:10.1111/j.1601-0825.2006.01291.x� 2007 The Authors. Journal compilation � 2007 Blackwell Munksgaard

All rights reserved

http://www.blackwellmunksgaard.com

to the destruction of the alveolar bone. Periodontitis hastwo distinct but interconnected etiologic components:periodontopathic bacteria adjacent to the periodontaltissues, and host-mediated connective tissue-destructiveresponses to the causative bacteria and their metabolicproducts (Golub et al, 1998). Treatment strategiesagainst periodontal diseases have evolved with the aimof eliminating specific pathogens or suppressing destruc-tive host response. Although the pathogenesis of thevarious forms of this disease is not completely under-stood, several factors are believed to play an importantrole, including inflammatory cytokines (McGee et al,1998; Graves, 1999), PGE2 (Offenbacher, 1996; Tsai et al,1998; Kornman, 1999), nitric oxide (Matejka et al, 1999;Takeichi et al, 2000) and salivary immunoglobulin-A(IgA) antibodies (Schenk et al, 1993).Moreover, many ofthe characteristics of periodontal diseases such as inflam-mation and attachment loss are associated with proteo-lytic events. Indeed, oral damage can result from therelease of an array of proteolytic enzymes by colonizingbacteria as well as host-derived proteases (Henskens et al,1996; Baron et al, 1999; Korostoff et al, 2000; Travis andPotemba, 2000). Selection of an appropriate deliverysystem is an important factor and, periodontitis being a�localized’ disease condition, local drug treatment shouldbe preferred.

The aims of this study were firstly to analyze theeffects of AD-expressing L. brevis-containing lozenges,both on the clinical signs of chronic periodontitispatients and on the NOS activity levels, matrix metal-loproteinase (MMP) levels, eicosanoid generation, IgAlevels, IFN-c levels of the saliva. The mechanismsunderlying the anti-inflammatory effect of L. brevisextract were investigated in vitro on rat peritonealmacrophages activated by lipopolysaccharide (LPS).

Materials and methods

Materials

Nitrate reductase, L-lactic dehydrogenase (LDH), pyru-vic acid, b-nicotinamide adenine dinucleotide phosphate(NADPH), flavin adenine dinucleotide (FAD), flavinmononucleotide (FMN), (6R)-5,6,7,8-tetrahydrobiop-terin (H4B), L-valine, phenylmethanesulfonyl fluoride(PMSF), pepstatin A, leupeptin, aprotinin, ethyleneglycol-bis (2-aminoethlether)-N,N,N,‘N’-tetraaceticacid (EGTA), ethylenediamine tetraacetic acid (EDTA),4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid(HEPES), Tris base, gelatin, Triton X-100, Coomassiebrilliant blue and other chemicals were obtained fromSigma Aldrich srl (Milan, Italy). L-ornithine hydrochlo-ride and TPA were purchased from Alexis (San Diego,CA, USA). L-lisin monohydrochloride was obtainedfrom Merck (Merck Bioscience GmbH, Schwalback,Germany). EN3HANCE spray was obtained fromPerkin-Elmer (Perkin-Elmer Life and AnalyticalSciences, Boston, MA, USA). Methanol, acetic acid,chloroform and ammonium hydroxide were purchasedfrom J.T. Baker (Milan, Italy), Microconcentrators

(Centricon 10) from Amicon (Amicon, Millipore Co.,Bedford, MA, USA). IFN-c (Endogen, Inc., Walburn,MA, USA) and PGE2 Elisa kits (Assay Designs Inc.,Ann Arbor, MI, USA) were purchased from TEMAricerca srl (Bologna, Italy). IgAs Elisa kits wereobtained from IPR (Catania, Italy).

Preparation of bacterial extractsLactobacillus brevis, obtained from VSL Pharmaceuti-cals (Gaithersburg, MD, USA) in a pure lyophilizedform (108 colony-forming units per gram), was routinelygrown in MRS broth medium under aerobic conditions.For the determination of AD activity and for in vitroexperiments on macrophages, L. brevis cultures werewashed and resuspended in 10 ml phosphate-bufferedsolution (PBS), sonicated (30 min, alternating 10 ssonication and 10 s pause) with a Vibracell sonicator(Sonic and Materials Inc., Danbury, CT, USA), centri-fuged at 7500 g for 20 min at 4�C and the supernatantsstored at )20�C until used (bacterial extract). Forthe in vitro experiments, the bacterial extract wasadded to cell cultures at 0.5 mg/105cells/1 ml)1 (finalconcentration).

For gene expression studies, genomic DNA wasisolated from an overnight culture of L. brevis usinga WIZARD genomic DNA purification kit (Promega,Milan, Italy), according to the manufacturer’sinstructions.

Arginine deiminase activityArginine deiminase activity was determined in L. brevisextracts bymeasuring the conversion of [3H]-L-arginine to[3H]-L-citrulline. Samples were homogenized by sonica-tion in homogenization ice-cold buffer (25 mMTris-HCl,pH 7.4, 1 mMEDTA, 1 mMEGTA) containing proteaseinhibitors (0.2 mM PMSF, 10 lg ml)1 aprotinin,10 lg ml)1 pepstatin A and 10 l g ml)1 leupeptin). Thehomogenates were centrifuged at 4�C for 20 min at11 000 g. The supernatants were passed through anAG50WX-8 Dowex resin (Bio-Rad Laboratories, Mel-ville, NY, USA) to remove endogenous arginine. Theprotein content was determined by the bicinchoninic acidprocedure (BCA; Pierce, Rockford, IL, USA) using BSAas standard. Enzymatic reactions were conducted at 37�Cfor 30 min in 50 mMTris-HCl, pH 7.4, containing 1 mMCaCl2, 1 mM MgCl2, 1 mM L-citrulline (to inhibit thecatabolism of [3H]-L-citrulline), 60 mM L-valine, 60 mML-ornithine, 60 mM L-lysine (to inhibit nonspecific argi-nase activity) and 10 lCi ml)1 of [3H]-L-arginine (Amer-sham, Bukinghamshire, UK). The reaction was stoppedby the addition of 0.4 ml of stop buffer (50 mMHEPES,pH 5.5, 5 mM EDTA). The samples were spotted ontoTLC plates and developed in chloroform/methanol/ammonium hydroxide/water (1:4:2:1, by vol) as solventsystem. After drying, the radiolabeled spots were visual-ized by autoradiography, after spraying EN3HANCER(Perkin-Elmer Life and Analytical Sciences). [3H]-argin-ine and [3H]-citrulline, identified by co-chromatographywith unlabeled standards which were revealed withninhydrin spray, was scraped, and quantified by liquidscintillation. Where indicated, bacterial extracts were

Anti-inflammatory effects of L. brevisDN Della Riccia et al

377

Oral Diseases

preincubated for 60 min with 20 mM formamidinehydrochloride (Fluka Seelze, Germany), an inhibitor ofAD (Weikmann et al, 1978a,b).

Arginine deiminase gene PCR amplificationThe oligonucleotides were designed on the bases ofnucleotide sequences of AD genes conserved at EMBLand GeneBank databases (accession number AJ001330).For polymerase chain reaction (PCR) experiments thefollowing primers were used: primer arcA1 frame5¢-ATTTGCACAAACATTACGAGAC and primerarcA2 reverse 5¢-GTGAAGACTGTATCTAAATG-CAT. Polymerase chain reaction (PCR) to amplify theAD-encoding gene, was performed in a 100 ll volume.The amplification reaction mixture contained 2.5 U ofAmpliTaq Gold DNA polymerase (Perkin-Elmer Lifeand Analytical Sciences), the reaction buffer provided bythe Taq manufacturer (containing 1.5 mM MgCl2),200 lM each deoxyribonucleotide triphosphate(dNTPs), 0.5 lM each primer, and 1 lg of genomicDNA as template. Amplification reactions were carriedout as follows: 12 min at 95� C for activation ofAmpliTaq Gold DNA polymerase; denaturation at 94�C for 1 min, annealing at 55� C for 1 min, and extensionat 72�C for 1 min, repeated for 30 cycles. A final DNAextension step at 72�C for 7 min was allowed at the endof the termal cycling. Direct sequencing of PCR ampl-icons was performed on both strands by dideoxy-chaintermination method using a dRhodamina TerminatorCycle Sequencing Ready reaction Kit and the ABIPRISM 377 DNA Sequencer (Perkin-Elmer Life andAnalytical Sciences). Sequencing was performed on threePCR products derived from three independent reactions.

L. brevis-containing lozengesEach lozenge (VSL Pharmaceuticals) was composed ofthe following: lyophilized bacteria (200 mg), fructose(400 mg), mannitole (1020 mg), aspartame (40 mg), talc(60 mg), aerosil 200 (10 mg), Peg 6000 (50 mg), mag-nesium stearate (20 mg), tartaric acid (100 mg), andorange flavor (100 mg).

Healthy volunteers, patients, treatment, and samplecollectionThe study was conducted as a double-blind paired-comparison study, with treatment assignments random-ized and balanced to a group of eight healthy volunteers(people of our laboratories; male and female, age 24–47 years, mean age 33.6 ± 7.96), who were clinically freeof periodontal disease. Twenty-one patients with chronicperiodontitis, 16 men and five women, 30–51 years old(meanage 41.9 ± 7.23)were enrolled to analyze the effectof L. brevis-containing lozenges (4 lozenges/day for4 days) on periodontitis-associated symptoms and signs,from the clinical point of view. Patients in this studydisplayed a range of periodontal disease frommoderate tosevere periodontitis. Pocket depth, measured to thenearest millimeter marking, was >4 mm. Mean loss ofattachment, evaluated bymeasuring the distance from thecemento-enamel junction to the bottom of the probingpocket, was 4.7 mm.No systemic disease was observed in

any patient, and antibiotics were not taken during theprevious 4 months. No patient was subjected to oralhygienics. Before and after the treatment with lozenges,the patients received a periodontal examination includingplaque index (Silness and Loe, 1964), gingival index,bleeding on probing, calculus and temperature sensitivityscores. All periodontal disease measurements were per-formed in four quadrants at six sited per tooth for allteeth, with the exception of plaque index for which foursites were examined. The results were expressed as ameanvalue accompanied by standard error of the mean. Thesaliva samples, collected from controls and patients byspitting into an ice-cooled vessel before and after thetreatment, were analyzed forMMPandNOS activity (thelatter determined as nitrite/nitrate levels in the fluid salivafraction and as citrulline levels in the cellular salivafraction), and IgA, PGE2, IFN-c. Five patients, four menand one women, 29–48 years old (mean age 40 ± 9.6)manifesting periodontitis were also enrolled in the trialand treated for 4 days with conventional oral antibiotictherapy (spiromycin, 1 · 300 mg day)1) with the aim ofcomparing the effects of this conventional treatment withthose exerted by the bacterial extracts of L. brevis. Allparticipants were carefully informed of the aim of theinvestigation and were free to withdraw from the study atany time. Written consent was obtained from all patientsprior to the collection of samples. The studywas approvedby the Institute ofOdontoiatric Clinics of the La SapienzaUniversity, Rome, and was conducted in according withthe principles of the Declaration of Helsinki.

AnimalsWistar rats (5–6 weeks old) were purchased fromCharles River Breeding Laboratories (Calco, Como,Italy).

Rat peritoneal macrophagesRat macrophages were obtained by peritoneal washeswith PBS, centrifuged and resuspended in Roswell ParkMemorial Institute (RPMI) 1640 medium containing2 mM glutamine, 10% heat-inactivated fetal calf serum(FCS) and 100 IU ml)1 penicillin and 100 lg ml)1

streptomycin. The cells were counted and their viability,as assessed by Trypan blue dye exclusion, was routinelygreater than 98%. Two milliliters of cell suspension(4 · 106cell ml)1) was then plated on a 12-well tissueculture plate (Falcon; Becton Dickinson, Buccinasco,Milan). The following day, cell cultures were washedthrice with medium without FCS and then cultured inthe same medium for 18 h with or without LPS(100 ng ml)1) in the presence or absence of bacterialextracts (0.5 mg ml)1). At the end of the treatments, cellpellets were used for NOS activity determination andsupernatants, concentrated 10-fold by Centricon C10(Amicon, Millipore Co.), were used in zymogramexperiments to detect metalloproteinase activity.

Nitrite/nitrate levelsIn aqueous solution, NO reacts rapidly with O2 andaccumulates in the culture medium as nitrite and nitrateions. Twenty microliters of saliva samples or superna-


378

Oral Diseases

tants from macrophage cultures was used for nitritemeasurement by a colorimetric assay based on the Griessreaction. Briefly, 5 ll of HEPES (50 mM, pH 7.4), 5 llof FAD (5 lM), 10 ll NADPH (0.1 mM), 58 ll of H2Oand 2 ll of nitrate reductase (0.2 IU ml)1) were added to20 ll of samples, which were then incubated for 30 minat 37� C. At the end of the incubation time 10 ll ofpyruvic acid (100 mM) and 1 ll of LDH (1500 IU ml)1)were added to the samples. After 10 min of incubation at37� C, samples were mixed with an equal volume ofGriess reagent in a microtiter plate and incubated atroom temperature in the dark for 10 min. The OD wasmeasured at 550 nm using a micro-enzyme-linked imu-nosorbent assay (ELISA) reader (Easy Reader EAR 400;Kontron Analytic, London, UK). KNO3 dissolved inH2O was used as standard and H2O as blank.

Nitric oxide synthase activityNitric oxide synthase activity was determined in bothmacrophage extracts and cell fraction of salivary sam-ples by measuring the conversion of [3H]-L-arginine to[3H]-L-citrulline as above described for AD assay. Theassay, performed with the addition to the reaction bufferof tetrahydrobiopterin (BH4), 2 lM FAD, 2 lM flavinadenine mononucleotide, CaCl2 (0.6 mM) and calmod-ulin (CAM) (0.1 lM), measures both the calcium-dependent (constitutive) and calcium-independent(inducible) isoenzyme activities. Any activity inhibitedby AG or detected in the absence of Ca/CAM and in thepresence of the CAM antagonist W13 (100 lM) (toinhibit endogenous CAM), represented iNOS activity.Where indicated, 20 mM formamidine or 10 lM ami-noguanidine (AG) or 500 lM L-NMMA were added tothe assay system to measure AD and/or NOS activity.

Determination of saliva IFN-c, PGE2, and IgA levelsThe IFN-c (Endogen Inc., Woburn, MA, USA), PGE2

(Assay Designs) and IgA (IPR S.p.A., Catania, Italy)levels were determined in the fluid saliva fraction usingELISA kits, according to the manufacturer’s instruc-tions. Where indicated, PGE2 levels were also assessedin macrophage culture medium.

Substrate gel electrophoresis (zymogram)SDS-PAGE zymograms containing 0.1% gelatin wereperformed as described previously (Thorgeirsson andMacKay, 1992) using approximately 2 lg of sampleprotein from saliva samples or cultured macrophagesupernatants to reveal the gelanolytic activity. Gelano-lytic activity secreted by cultured macrophages in basalor stimulated conditions was analyzed using equal10-fold concentrated sample volume (10–20 ll). Follow-ing electrophoresis, gels were rinsed in 50 mM Tris-HCl(pH 7.4) containing 2% Triton X-100 followed by50 mM Tris-HCl (pH 7.4) and incubated overnight forgelatinases, in a buffer containing 50 mM Tris-HCl (pH7.4), 0.2 M NaCl, 5 mM CaCl2, and 1% Triton X-100 at37�C. Enzyme activity was detected following stainingwith 0.1% Coomassie brilliant blue in a mixture ofacetic acid:methanol:water (1:3:6), destained in the samemixture without dye and dried. Sample collagenase

activity was compared with that of partially purifiedgelatinases (MMP-2 and MMP-9) obtained from thesupernatants of HT1080 human fibrosarcoma cell line(ICLCHTL98016, InterlabCellLineCollection,Genova,Italy) stimulated with TPA (10)7M) for 48 h. Allexperiments were conducted with or without EDTA,serine or cystein proteinase inhibitors. Bands relative toenzyme activity were quantified by scanning densito-metry using Molecular Analyst PCTM software for theBio-Rad model 670 scanning densitometer.

Western blotSaliva collected from healthy volunteers and patientswere centrifuged at 600 g for 15 min and the superna-tants were assayed for protein using the BCA method(Pierce). Twenty-eight microliters of each sample (40 lgof protein) was mixed with 7 ll of nonreducing PAGEbuffer 5X [0.5 mol/l Tris-HCl (pH 6.8), 28% glycerol,4.4% SDS, 0.55% bromophenol blue] and separated onan 8.5% polyacrylamide gel. After electrophoresis, SDSwas removed by washing the gel in 50 mM Tris HCl pH7.4, 2% Triton X-100 for 15 min at 37�C and then withthe same buffer lacking Triton X-100 for 15 min at 37�C.Afterwards, the gel was equilibrated in electrotransferbuffer (25 mM Tris-HCl, 192 mM glycine, 10% meth-anol) for 20 min and transferred to polyvinyl difluoride(PVDF) membrane (Millipore) for 1 h at 4�C at 100Vusing a Mini Trans-Blot Cell apparatus (Bio-RadLaboratories, Hercules, CA, USA). Nonspecific bindingsites were blocked with 10% nonfat dry milk in TBS-T(TBS and 0.05% Tween 20) for 16 h at 4�C. Themembrane was then incubated for 1 h at 25�C with amouse monoclonal antihuman MMP-9 (Calbiochem-Novabiochem, Darmstadt, Germany) diluted 1:400 in5% nonfat dry milk in TBS-T. After incubation themembrane was incubated for 1 h at 25�C with goatantimouse-horseradish peroxidase-conjugated IgG (San-ta Cruz Biotechnology, Santa Cruz, CA, USA) diluted1:3000 in 2.5% non-fat dry milk in TBS-T. Immunore-activity was assessed by chemiluminescence reactionusing ECL Western blotting detection system (Pierce).

Statistical analysisThe differences in the analyzed parameters between thecontrol and patient groups were tested by using theStudent’s t-test. The statistical significance of the differ-ences in clinical and laboratory parameters between T0and T1 was analyzed using the paired sample t-test.The STATPAC program was used to perform statisticalanalysis and statistical difference set at P < 0.05.

Results

Arginine deiminase activity in L. brevis extractsArginine deiminase activity was analyzed in bacterialextracts as described in the Materials and methodssection. The enzymatic assay, based on the conversionof L-arginine to L-citrulline in appropriate conditionsallowing to exclude NOS and arginase activity, showedthat AD activity was present at high levels in L. brevis(Figure 1a). The generation of citrulline by L. brevis


379

Oral Diseases

extract could be attributed to AD activity, as it wasobserved in the absence of NOS cofactors (NADPH,FAD, FMN, and BH4) and it was not affected either byL-valine, an arginase inhibitor, or by L-NMMA, a NOSinhibitor, and aminoguanidine, a relatively specificinhibitor of the inducible form of NOS. Moreover, theabrogation of citrulline generation observed in thepresence of formamidine, a specific inhibitor of AD,definitively established that the conversion of arginine tocitrulline was due to the presence of high levels of thisenzyme in L. brevis extracts.

The presence of AD in L. brevis was further con-firmed through the structural analysis of the geneencoding the enzyme. A couple of primers designed onthe basis of consensus sequences known to be aninternal region of the AD encoding gene was used toattempt PCR amplification. The PCR, performed ongenomic DNA of L. brevis, yielded an amplimer of665 bp (Figure 1b) which was subjected to directsequencing. The sequencing of the PCR fragment,named arcA1-2, resulted in a completely identified478 bp sequence (Figure 1c) encoding a polypeptide of

124 amino acid residues which showed several conservedstructural elements typical of AD genes. Comparingamino acids residues of the AD of L. brevis and AD ofL. sakei (Figure 1d), a stringent homology (52%) wasfound. Although the AD gene is scarcely conserved,including the internal regions, our results clearlydemonstrated the presence of this gene in L. brevis.

Clinical evaluationThe scores for gingival inflammation, plaque, tartar,temperature sensitivity, and bleeding on probing beforeand after the treatment with L. brevis lozenges arereported in Table 1. All clinical parameters that wereanalyzed reached statistically significant differencesbetween patients before treatment (T0) and the sameafter treatment with L. brevis containing lozenges (T1).On the mean, gingival inflammation was rated asmoderate/diffuse by 20 patients at enrollment, and mildby one subject. After treatment, almost all patientsshowed no gingival inflammation; two individualscomplained of a mild inflammation, while just onepatient had a still moderate gingival inflammation. The

05

1015202530354045

(a)

(c)

(d)

(b)C

itrul

line

(pm

ol m

g–1 p

rot m

in–1

)

L. b

revi

s +

Form

amid

ine

+ L-

NA

ME

+ A

min

ogua

nidi

ne

665 bp

Mar

kers

5’ aaagaggttt ctagcgatgt cttcgatggc gtccgccgtcgtccgaggga aaccccaatt gctaagacgt gtcggacaa gacgagctcgtcgccccctt cgatgtggga ttcgtggttc cggtcgcgcc atacctcaacttgccccgca aaccgtgggt ggtggttgat gatgtattcc gtgatcagggattcacgttg acgggccggg aaagtcatct tgttaatgga gatcccgtcaccaatggagg cctgcggatc cgcgtgaag taagcgtttg gcattgggtccattaagaac ggccagtccg tattttcgga aacggattgc agatccggttgggcgatgtc aatgtcagtc ccccgcaccc cttcgatgat cttgtcaaccatttcctgcg gggtcatttg ggtagaggta gtcatataag gcgtcgtgggtggtacccgc aacgtaacca gattcggcga tcatccgg 3’

L. sakei L D M V E K I M A G V R T N E I D I K S KL. brevis M E T V D K I I E G V R G T D I D I A Q P

L. sakei A L I D V S A - - - - - - - - - - - - - -L. brevis D L Q S V S E N T D W P F L M E T D P M E

L. sakei - - N L Y F T R D P A A S M G D G L T I NL. brevis T P N A Y F T R D P Q A S I G D G I S I N

L. sakei K M - - T F E A R Q R E S M F M E V I M QL. brevis K M E T T F P A R Q R E S L I T E Y I I N

L. sakei H H P R F A N Q G A Q V W R D R D H I D RL. brevis H H P R F A G O – V E V W R D R N H E S H

L. sakei M E G G D E L I L S D K V L A I G I SL. brevis I E G G D E L V L S D H V L A I G V S

Figure 1 Arginine deiminase expression inLactobacillus brevis extracts. (a) Argininedeiminase activity in bacterial extracts in thepresence or absence of formamidine, a specificinhibitor of arginine deiminase, L-NMMA, anitric oxide synthase (NOS) inhibitor, andaminoguanidine, a relatively specific inhibitorof the inducible form of NOS. The resultsrepresent the mean values of duplicate deter-minations and are representative of one ofthree experiments. SD values were ever lowerthan 10% of mean value. *P < 0.001 whencompared with L. brevis extracts alone. (b)Analysis of arginine deiminase gene, afterPCR amplification performed on genomicDNA, through 1.2% agarose gel electro-phoresis. (c) Analysis of the direct sequencingof the PCR-obtained fragment amplimer(arcA1-2). (d) The sequence of a 124 aminoacid residue encoded by the PCR-obtainedfragment, corresponding to arginine deimi-nase protein. The stringent homology (52%)with the arginine deiminase of L. sakei isshown


380

Oral Diseases

tolerability was considered as very good by all patients.Overall, the clinical evaluation of both the dentist andthe patient overlapped, and the results confirmedrecovery in 18 patients and improvement in threeindividuals. The antibiotic treatment for 4 days didnot significantly modify signs and symptoms (data notshown), as expected.

Effects of L. brevis-containing lozenges on saliva sampleNOS activity, PGE2, IFN-c, and IgA levelsThe levels of nitrites and nitrates and those of PGE2,IFN-c, and IgA, determined in the fluid saliva fractionof healthy controls and periodontitis patients, asdescribed in Materials and methods in an unblindedfashion, are reported in Table 2. Periodontitis patients’saliva had a significantly higher level of nitrites/nitrateswhen compared with normal subjects (P < 0.01). In thesame Table are also reported the results obtained whenperiodontitis patients’ saliva was analyzed for nitrite/nitrate levels before (T0) and after (T1) treatment withL. brevis-containing lozenges. In all patients, the treat-ment was associated with a significant reduction innitrite and nitrate levels. Similar results were obtainedwhen the conversion of radiolabeled arginine to citrul-line was analyzed in the cell saliva fraction (data notshown). In addition the PGE2 and IFN-c levels weresignificantly higher in patients with periodontitis than incontrols (P < 0.01). Interestingly, treatment withL. brevis-containing lozenges was associated with astrong decrease in the levels of these inflammation-associated molecules (P < 0.01). In contrast, IgA levels,which were significantly reduced in patients comparedwith controls (P < 0.05), were not significantly mod-ified by the treatment.

Effects of L. brevis-containing lozenges on saliva sampleMMP activity and levelsThe measurements of MMP activity in the solublefraction of saliva samples disclosed that the salivarysamples of periodontitis patients contained higher

amounts of MMPs than the control samples, thusconfirming previous reports. Figure 2 shows gelatinaseactivity present in five different control salivary samples(panel a) as well as in five different patients’ salivarysamples (panel b), representative of all examined cases.Enzymatic activity was compared with that of gelatin-ases obtained from the supernatants of the HT1080 cellline stimulated with tissue plasminogen activator (TPA)(10)7 M) for 48 h (the 72 kDa type IV collagenase,gelatinase A, MMP-2, and the 92 kDa type IV collag-enase, gelatinase B, MMP-9). The zymogram of healthycontrol salivas (Figure 2a) shows bands with relativemolecular weights of 92 kDa, of about 120 kDa andhigher than 200 kDa. Aside from a band of about37 kDa, little or no enzymatic activity was visible at

Table 1 Effect of the treatment with Lactobacillus brevis-containinglozenges on gingival inflammation, plaque and calculus, temperaturesensitivity, and bleeding in periodontitis patients

T0 T1 P-value

Gingival inflammation 2.24 ± 0.12 0.14 ± 0.10 <0.0001Plaque 2.10 ± 0.14 1.10 ± 0.12 <0.0001Calculus 1.62 ± 0.21 1.29 ± 0.16 <0.0001Temperature sensitivity 2.29 ± 0.14 0.43 ± 0.11 <0.0001Bleeding 1.38 ± 0.19 0.05 ± 0.05 <0.0001

Table 2 Effect of the treatment with Lactobacillus brevis-containing lozenges on laboratory parameters in periodontitis patients

Samples Nitrite/nitrate (lM) PGE2 (ng ml)1) IFN-c (pg ml)1) IgA (lg ml)1)

HC 4.33 ± 1.31 P < 0.0001 0.32 ± 0.08 10.06 ± 0.84 885.62 ± 118.28T0 38.86 ± 6.07* P < 0.0001 1.64 ± 0.1* 28.23 ± 2.11* 309.76 ± 143.06*P < 0.05T1 9.39 ± 1.47** 0.46 ± 0.07** 8.82 ± 4.74**1.04 466.51 ± 237.1

HC, healthy controls; T0, before treatment; T1, after treatment. *P ¼ T0 vs HC; **P ¼ T1 vs T0.

37

100150 250

100 150 250

50

50

75

37

75

HT

1080

Healthy controls

HT

1080

Patients

92

92

72

72

Figure 2 Salivary matrix metalloproteinase (MMP) activity in healthysubjects and periodontitis patients. Gelatinase activity in the solublesalivary fractions of representative healthy subjects and periodontitispatients is shown. The electrophoretic mobility of MMP-9 and MMP-2, obtained from supernatants of HT1080 cell, is indicated on the left-hand side. Molecular weights of the standard are indicated on the right


381

Oral Diseases

molecular weights lower than 92 kDa. The profile ofmultiple forms of gelatinases found in patients’ saliva isshown in Figure 2b. In addition to the same bands ofcontrols, a strong gelatinolytic activity at molecularweights lower than 92 kDa could be seen in all patients’salivas. In some cases (i.e., lanes 2 and 3) the highmolecular weight activities of gelatinase were almostabsent, being prevailing below 92 kDa. These latterspecies mainly migrated with a connecting smear, beingevident as separate bands only at the 37–50 kDa region.The gelatinolytic activity detected by zimogramappeared to be due to the presence of MMPs, asprotease activity was completely abrogated by theaddition of EDTA, but not by the addition of serineand cysteine proteinase inhibitors to the incubationbuffer (data not shown). The treatment with lozenges ledto the near disappearance of these MMP bands in allpatients’ samples. Figure 3 shows the gelatinolytic activ-ity of five representative periodontitis patients before(T0) and after (T1) the treatment with experimentallozenges. The profile of multiple forms of gelatinaseswas similar in all patients. The treatment led to the neardisappearance of the low molecular weight bands andrestored the control profile of gelatinolytic activity.

Western blot experiments demonstrated that the bandsrevealed by zymograms were duemainly to the enzymaticactivity ofMMP-9. Figure 4 shows aWestern blot of onerepresentative control and three representative patientsbefore and after treatment with experimental lozenges. Inthe saliva of patients antibody against MMP-9 revealedsome high molecular weight forms (>200 kDa) togetherwith bands at the 180 kDa region and at 120 kDa. Othermolecular weights ranging from 92 to 64 kDa andapproximately at 54 and 44 kDa were seen in the salivaof patients with chronic periodontitis. After treatment theenzymatic pattern resembled that of healthy controls.Indeed, a relevant reduction in high molecular weightbands was observed and little or noMMP-9 species below

92 kDa were detected. Western blotting with specificantibodies against human MMP-2, MMP-3, MMP-1,and MMP-8 showed no immunoreactivity, suggestingthat theseMMPswere not detectable in the saliva of eithercontrols or patients (data not shown).

Effect of L. brevis extracts on macrophage NOS activityRat peritoneal macrophages incubated in vitro in thepresence of LPS for 18 h were induced to express iNOSactivity (Figure 5a). NOS activity was inhibited by theNOS inhibitors AG and L-NMMA. In the presence ofbacterial extracts, NOS activity disappeared and argin-ine was converted in citrulline through bacterial AD,as demonstrated either by the unsensitivity of thisconversion to NOS inhibitor treatment (Figure 5a) orby the disappearance of enzymatic activity in thepresence of the AD inhibitor, formamidine (data notshown). Similar results were obtained with L. fermentumextracts, but not with Streptococcus thermophilus, aLAB in which the AD activity was not detectable (notshown).

These results suggest that bacterial AD was able tototally utilize the arginine present in the medium thuspreventing its uptake by macrophages and consequentlyNOS activity.

Effect of L. brevis extracts on macrophage PGE2 releaseWhen rat peritoneal macrophages were cultured for upto 18 h with LPS, a gradual accumulation of PGE2 inthe supernatants occurred (fivefold increase over that incontrol cells) (Figure 5b). This effect was almost totallyabrogated in the presence of L. brevis extracts asoccurred when the cells were previously treated withL-NMMA or aminoguanidine.

Effect of L. brevis extracts on MMP production bymacrophages in the presence of LPSTo investigate the L. brevis effect on MMP production,activated rat peritoneal macrophages were treated with

HT

1080

T0 T1 T0 T1 T0 T1 T0 T1 T0 T1

4 5 7 3 6Patient n.

92

72

37

100150250

50

75

Figure 3 Effects of Lactobacillus brevis-containing lozenges on salivasample matrix metalloproteinase (MMP) activity. Zymography com-paring gelatinolitic activities in saliva of five periodontitis patientsbefore treatment (T0) with gelatinolitic activities in saliva of the samepatients after treatment with L. brevis-containing lozenges (T1). Theelectrophoretic mobility of MMP-9 and MMP-2, obtained fromsupernatants of HT1080 cell line, is indicated on the left-hand side.Molecular weights of the standard are indicated on the right-hand side

Patient n.

150

100

75

50

250

C T0 T1

4 5 6

T0 T1 T1 T0

Figure 4 Effects of Lactobacillus brevis-containing lozenges on salivasample matrix metalloproteinase levels. Western blot analysis ofgelatinase expression in one representative healthy control and threeperiodontitis patients’ saliva samples before and after treatment withlozenges containing L. brevis. C, healthy control; T0, before treatment;T1, after treatment. Molecular mass markers are indicated on the left-hand side


382

Oral Diseases

bacterial extracts in the presence or absence of LPS andcell culture supernatants were analyzed by zymographyfor MMP-9 and MMP-2 levels. Compared with macr-ophages in culture medium alone, LPS increased MMP-9 activity, whereas MMP-2 was not detectable. Thepresence of L. brevis extract in the culture medium led toa relevant decrease in MMP-9 activity, either whenadded alone or in the presence of stimuli. Similarly, atotal inhibition of MMP activity was observed in thepresence of iNOS inhibitors. A representative zymogra-phy is illustrated in Figure 5c. Enzymatic activity wascompared with that of partially purified gelatinases(MMP-2 and MMP-9) obtained from the supernatantsof HT1080 cell line stimulated with TPA (10)7M) for48 h. The bands revealed by zymograms could beattributed to MMPs as they disappear with EDTA butnot by the addition of other proteinase inhibitors to theincubation buffer (data not shown).

Discussion

Periodontitis is characterized by extensive destruction ofthe gingival tissues and associated supporting structuresof the teeth. In the present work, we first analyzed thepotential anti-inflammatory action in vivo of L. brevis-containing lozenges in 21 adult patients with periodontaldisease. Previously, our group reported the presence ofarginine-deiminase in L. brevis, a strain belonging toLAB (Di Marzio et al, 2001). The presence of high levelsof this enzyme, which metabolize arginine to citrullineand ammonia, aids L. brevis extracts to inhibit NOgeneration, by competing with NOS for the samesubstrate, arginine. To confirm the presence of AD inL. brevis, previously shown through the assay of

enzymatic activity (Di Marzio et al, 2001), we checkedthe presence of the AD gene. The analysis of a fragmentobtained by PCR unambiguously revealed the presenceof this gene, thus confirming our previous findings.

Nitric oxide is known to be an important inflamma-tory mediator, and is involved in the pathophysiology ofseveral inflammatory disorders. An increase in NOproduction has been also demonstrated in periodontitisas its role in the inflammatory response of periodontaltissues has been suggested (Matejka et al, 1998, 1999;Lappin et al, 2000; Hirose et al, 2001; Kendall et al,2001; Lohinai et al, 2001; Daghigh et al, 2002). Thesefindings are consistent with the hypothesis suggestingthat treatment with agents able to block the productionof NO or its effects might be therapeutically valuable forperiapical lesions (Lohinai et al, 1998; Paquette andWilliams, 2000; Daghigh et al, 2002). Considering therole of NO in inflammatory events in general and inperiodontitis in particular, we analyzed the potentialbeneficial effects of L. brevis-based treatment in patientswith periodontal disease. Of note, the proposed experi-mental treatment led to the total disappearance oramelioration of all analyzed clinical parameters (gingi-val inflammation, plaque, calculus, temperature sensi-tivity, and bleeding on probing) in almost all 21 patients(recovery in 18 patients and improvement in threeindividuals). The beneficial effects of the treatment onclinical signs and symptoms were paralleled by a strongand statistically significant decrease in oral NOS activ-ity. Besides its direct role as an inflammatory modulator,NO represents a multifunctional messenger moleculeable to modulate the production of inflammatorycytokines (Cifone et al, 1999), PGE2 (Goodwin et al,1999), and MMPs (Maeda et al, 1998; Tronc et al, 2000;

0

0.5

1

1.5

2

2.5

None L-NAME Aminoguanidine Formamidine

Citr

ullin

e (p

mol

mg–1

pro

t min

–1)

Control

LPSL. brevis

LPS + L. brevis

3

MMP9

MMP2

HT

108

0

Con

trol

L. b

revi

s

LPS

LPS

+ L

. bre

vis

PGE

2 (n

g/10

6 cel

ls)

0

10

20

30

40

50

60 ControlL. brevisLPSLPS+ L.brevis LPS+ L-NMMALPS+ AG

(a)

(b) (c)

Figure 5 Effect of Lactobacillus brevisextracts on LPS-activated macrophages.Effect of L. brevis extracts on iNOS activity(a), PGE2 levels (b) and MMP activity (c) inrat peritoneal macrophages incubated in vitroin the presence or absence of LPS for 18 h.Where indicated, the effect of L-NMMA,aminoguanidine (AG), and formamidine isshown


383

Oral Diseases

Hirai et al, 2001). On the other hand, inflammatorycytokines such as IFN-c, tumor necrosis factor-a andinterleukin-1-a have been shown to be significantlyhigher in periodontitis patients than in healthy controls(Takeichi et al, 2000; Hirose et al, 2001; Tervahartialaet al, 2001; Ukai et al, 2001). Moreover, the generationof PGE2, a potent eicosanoid lipid mediator derivedfrom phospholipase-released arachidonic acid that isinvolved in numerous homeostatic biologic functionsand inflammation, has been shown to be stronglyincremented in periodontal lesions (Cavanaugh et al,1998; Leibur et al, 1999; Paquette and Williams, 2000;Tsilingaridis et al, 2003). Furthermore, MMP levelshave been previously shown to play a critical role in theinflammatory process leading to alveolar bone andconnective tissue loss in periodontal disease (Makelaet al, 1994; Ding et al, 1995; Ejeil et al, 2003; Sorsa et al,2004). Indeed, one of the most salient features ofperiodontitis and gingivitis is the increase in the levels ofbacterial and host-derived proteolytic enzymes in oralinflammatory exudates. Bacterial pathogens involved inperiodontal diseases partly trigger and induce host cellsto elevate their secretion of MMP. Proteolytic enzymesreleased by the host cells and/or by pathogen bacteriaare associated with tissue destruction and MMP havethe primary role in this process as they can degrade mostof the extracellular matrix components (Ding et al,1995).

Altogether, these observations prompted us to ana-lyze the potential effects of L. brevis extracts containinglozenges on IFN-c, PGE2, and MMP salivary levels inperiodontitis patients. Our findings clearly showed thatall these inflammatory-associated factors were drastic-ally reduced in periodontal disease patients after theL. brevis-based treatment. However, no significanteffects were observed on salivary IgA levels after theproposed treatment.

In the attempt to define the mechanisms underlyingthe anti-inflammatory action of L-brevis-based treat-ment, a set of experiments were performed by using ratperitoneal macrophages activated in vitro with LPS inthe presence or absence of bacterial extracts. The resultsindicated that the LPS-induced iNOS activity, PGE2

generation, and MMP activation were almost totallyabrogated when macrophages were cultured in thepresence of bacterial extracts, thus strongly supportingthe in vivo observations. The finding that iNOS inhib-itors were able to inhibit either PGE2 release or MMPactivity in LPS-activated macrophages, prompted us tohypothesize that NO could be responsible for the latterevents, as previously reported in several systems (Maedaet al, 1998; Goodwin et al, 1999; Tronc et al, 2000;Hirai et al, 2001). The anti-inflammatory effects ofL. brevis extracts could thus be attributed to theirability to prevent iNOS activity and, consequentlyNO-dependent PGE2 release and MMP activation.

Considering recent reports showing that Lactobacillusand Bifidobacterium spp. or their components protectedthe host against certain oral diseases by inhibiting thegrowth of potential pathogens (Smith et al, 2001;Grudianov et al, 2002; Volozhin et al, 2004), we cannot

exclude the possibility that L. brevis extract-basedtreatment, besides the anti-inflammatory effects repor-ted here, is also able to antagonize the growth of specificperiodontopathic pathogens.

Taken together, our findings give further insights intothe knowledge of the molecular basis of periodontitisand have a potential clinical significance, representingthe experimental ground for a new innovative, simpleand efficacious therapeutical approach of periodontaldisease.

Acknowledgements

The authors thank Gasperina De Nuntiis (Department ofExperimental Medicine, University of L’Aquila, L’Aquila,Italy) for technical assistance and Miss Florence Pryen forcareful language revision of the manuscript. This work wasfinancially supported by grant �PROGETTO DI INTERESSEGENERALE DI ATENEO’ from the University of L’Aquila.

References

Baron AC, Gansky SA, Ryder MI et al (1999). Cysteineprotease inhibitory activity and levels of salivary cystatins inwhole saliva of periodontally diseased patients. J Periodon-tal Res 34: 437–444.

Bodana AR, Rao DR (1990). Antimutagenic activity of milkfermented by Streptococcus thermophilus and Lactobacillusbulgaricus. J Dairy Sci 73: 3379–3384.

Caglar E, Kargul B, Tanboga I (2005). Bacteriotherapy andprobiotics’ role on oral health. Oral Dis 11: 131–137.

Cavanaugh PFJR, Meredith MP, Buchanan W et al (1998).Coordinate production of PGE2 and IL-1 beta in thegingival crevicular fluid of adults with periodontitis: itsrelationship to alveolar bone loss and disruption by twicedaily treatment with ketorolac tromethamine oral rinse.J Periodontal Res 33: 75–82.

Cifone MG, D’alo S, Parroni R et al (1999). Interleukin-2-activated rat natural killer cells express inducible nitric oxidesynthase that contributes to cytotoxic function and inter-feron-gamma production. Blood 93: 3876–3884.

Daghigh F, Borghaei RC, Thornton RD et al (2002). Humangingival fibroblasts produce nitric oxide in response toproinflammatory cytokines. J Periodontol 73: 392–400.

Di Marzio L, Russo PF, Biordi L et al (2001). Apoptoticeffects of selected strains of lactic acid bacteria on a humanT leukaemia cell line are associated with bacterial argininedeiminase and/or sphingomyelinase activities. Nutr Cancer40: 185–196.

Ding Y, Uitto VJ, Firth J et al (1995). Modulation of hostmatrix metalloproteinases by bacterial virulence factorsrelevant in human periodontal diseases. Oral Dis 1: 279–286.

Dugas B, Mercenier A, Lenoir-Wijnkoop I et al (1999).Immunity and probiotics. Immunol Today 20: 387–390.

Ejeil AL, Igondjo-Tchen S, Ghomrasseni S et al (2003).Expression of matrix metalloproteinases (MMPs) and tissueinhibitors of metalloproteinases (TIMPs) in healthy anddiseased human gingiva. J Periodontol 2: 188–195.

Golub LM, Ryan ME, Williams RC (1998). Modulation of thehost response in the treatment of periodontitis. Dent Today17: 102–106.

Goodwin DC, Landino LM, Marnett LJ (1999). Effects ofnitric oxide and nitric oxide-derived species on prostaglan-din endoperoxide synthase and prostaglandin biosynthesis.FASEB J 13: 1121–1136.


384

Oral Diseases

Gorbach SL (2000). Probiotics and gastrointestinal health. AmJ Gastroenterol 95 (Suppl. 1): S2–S4.

Graves DT (1999). The potential role of chemokines andinflammatory cytokines in periodontal disease progression.Clin Infect Dis 28: 482–490.

Grudianov AI, Dmitrieva NA, Fomenko EV (2002). Use ofprobiotics Bifidumbacterin and Acilact in tablets in therapyof periodontal inflammation. Stomatologiia Mosk 81: 39–43.

Henskens YM, Van Den Keijbus PA, Veeman EC et al (1996).Protein composition of whole and parotid saliva in healthyand periodontitis subjects. Determination of cystatins,albumin, amylase and IgA. J Periodontol Res 31: 57–65.

Hirai Y, Migita K, Honda S et al (2001). Effects of nitric oxideon matrix metalloproteinase-2 production by rheumatoidsynovial cells. Life Sci 68: 913–920.

Hirose M, Ishihara K, Saito A et al (2001). Expression ofcytokines and inducible nitric oxide synthase inflamedgingival tissue. J Periodontol 72: 590–597.

Ishii Y, Ogura T, Tatemichi M et al (2003). Induction ofmatrix metalloproteinase gene transcription by nitric oxideand mechanisms of MMP-1 gene induction in humanmelanoma cell lines. Int J Cancer 103: 161–168.

Kendall HK, Marshall RI, Bartold PM (2001). Nitric oxideand tissue destruction. Oral Dis 7: 2–10.

Konings WN, Kok J, Kuipers OP et al (2000). Lactic acidbacteria: the bugs of the new millennium. Curr OpinMicrobiol 3: 276–282.

Kornman KS (1999). Host modulation as a therapeuticstrategy in the treatment of periodontal disease. Clin InfectDis 28: 520–526.

Korostoff JM, Wang JF, Sarment DP et al (2000). Analysis ofin situ protease activity in chronic adult periodontitispatients: expression of activated MMP-2 and a 40 kDaserine protease. J Periodontol 71: 353–360

Lappin DF, Kjeldsen M, Sander L et al (2000). Inducible nitricoxide synthase expression in periodontitis. J Periodontal Res35: 369–373.

Leibur E, Tuhkanen A, Pintson U et al (1999). ProstaglandinE2 levels in blood plasma and in crevicular fluid of advancedperiodontitis patients before and after surgical therapy. OralDis 5: 223–228.

Lohinai Z, Benedek P, Feher E et al (1998). Protective effectsof mercaptoethylguanidine, a selective inhibitor of induciblenitric oxide synthase, in ligature-induced periodontitis in therat. Br J Pharmacol 123: 353–357.

Lohinai Z, Stachlewitz R, Virag L et al (2001). Evidence forreactive nitrogen species formation in the gingivomucosaltissue. J Dental Res 80: 470–475.

Maeda H, Okamoto T, Akaike T (1998). Human matrixmetalloproteinase activation by insults of bacterial infectioninvolving proteases and free radicals. Biol Chem 379:193–200.

Makela M, Salo T, Uitto VJ et al (1994). Matrix metallopro-teinases (MMP-2 and MMP-9) of the oral cavity: cellularorigin and relationship to periodontal status. J Dental Res73: 1397–1406.

Matejka M, Partyka L, Ulm C et al (1998). Nitric oxidesynthesis is increased in periodontal disease. J PeriodontalRes 33: 517–518.

Matejka M, Ulm C, Partyka L et al (1999). Nitric oxidesynthesis is increased in periodontitis. Adv Exp Med Biol469: 419–424.

McGee JM, Tucci MA, Edmundson TP et al (1998). Therelationship between concentrations of proinflammatorycytokines within gingival and the adjacent sulcural depth.J Periodontol 69: 865–871.

Offenbacher S (1996). Periodontal diseases: pathogenesis. AnnPeriodontol 1: 821–878.

Paquette DW, Williams RC (2000). Modulation of hostinflammatory mediators as a treatment strategy for perio-dontal diseases. Periodontol 2000 24: 239–252.

Roberfroid MB (1998). Prebiotics and synbiotics: concepts andnutritional properties. Br J Nutr 80: S197–S220.

Schenk K, Poppelsdorf D, Denis C et al (1993). Levels ofsalivary IgA antibodies reactive with bacteria from dentalplaque are associated with susceptibility to experimentalgingivitis. J Clin Periodontol 20: 411–417.

Silness J, Loe H (1964). Periodontal disease in pregnancy. II.Correlation between oral hygiene and periodontal condi-tion. Acta Odontol Scand 22: 121–135.

Smith SI, Aweh AJ, Coker AO et al (2001). Lactobacilli inhuman dental caries and saliva. Microbios 105: 77–85.

Sorsa T, Tjaderhane L, Salo T (2004). Matrix metalloprotein-ases (MMPs) in oral diseases. Oral Dis 6: 311–318 (Review).

Takeichi O, Haber J, Kawai T et al (2000). Cytokine profilesof T-lymphocytes from gingival tissues with pathologicalpocketing. J Dent Res 79: 1548–1555.

Tervahartiala T, Koski H, Xu JW et al (2001). Tumor necrosisfactor-alpha and its receptors, p55 and p75, in gingiva ofadult periodontitis. J Dent Res 80: 1684–1687.

Thorgeirsson UP, MacKay AR (1992). Characterization ofmetastatic tumor cell. In: Gallegner CT, Rees RC, ReynoldsCW, eds. Tumor immunology: a practical approach. Univer-sity Press: Oxford, pp. 82–90.

Travis J, Potemba J (2000). Bacterial proteinases as targets forthe development of second-generation antibiotics. BiochimBiophys Acta 1477: 35–50.

Tronc F, Mallat Z, Lehoux S et al (2000). Role of matrixmetalloproteinases in blood flow-induced arterial enlarge-ment: interaction with NO. Arterioscler Thromb Vasc Biol20: E120–E126.

Tsai CC, Hong CC, Chen CC et al (1998). Measurement ofprostaglandin E2 and leukotriene B4 in the gingival crevic-ular fluid. J Dent 26: 97–103.

Tsilingaridis G, Yucel-Lindberg T,Modeer T (2003). Enhancedlevels of prostaglandin E2, leukotriene B4, and matrixmetalloproteinase-9 in gingival crevicular fluid from patientswith Down syndrome. Acta Odontol Scand 3: 154–158.

Ukai T, Mori Y, Onoyama M et al (2001). Immunohistolog-ical study of interferon-gamma- and interleukin-4-bearingcells in human periodontitis gingival. Arch Oral Biol 46:

901–908.Volozhin AI, Il’in VK, Maksimovskii IuM et al (2004).Development and use of periodontal dressing of collagenand Lactobacillus casei 37 cell suspension in combinedtreatment of periodontal disease of inflammatory origin (amicrobiological study). Stomatologiia (Mosk) 83: 6–8

Weikmann JL, Himmel ME, Smith DW et al (1978a). Argin-ine deiminase: demonstration of two active sites and possiblehalf-of-the-sites reactivity. Biochem Biophys Res Commun83: 107–113.

Weikmann L, Himmel ME, Squire PG et al (1978b). Argininedeiminase from Mycoplasma arthritidis. Properties of theenzyme from log phase cultures. J Biol Chem 253: 6010–6015.


385

Oral Diseases

Health & Medicine

Anti-inflammatory doc effects of Lactobacillus brevis CD2 on periodontal disease