@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Dentistry, Faculty of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Coil, Albert J. M."@en ; dcterms:issued "2008-09-10T21:45:54Z"@en, "1992"@en ; vivo:relatedDegree "Doctor of Philosophy - PhD"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """Active periodontal disease is characterized by marked destruction of collagen in periodontally-involved tissues. Biochemical characterization of the disease is complicated as it is an episodic disorder that is believed to cycle through periods of high activity and quiescence. The first part of the study focused on the measurements and analysis of volatile sulphur compounds (VSC) collected at gingival sulci. Their concentrations in mouth air correlate with the severity of the disease. The second part was devoted to the evaluation of breakdown products of collagen metabolism found in the gingival crevicular fluid (GCF) in relation to periodontal disease. A novel method and device was developed for collection and analysis of volatile compounds from specific gingival crevice sites. It demonstrated that volatile sulphur compound profiles of crevicular air differ from that of mouth air. It showed that total sulphur content of either inflamed or deep (PD z4 mm) periodontal sites was significantly higher than the non inflamed or shallow (PD<4 mm) sites (p<.05). The ratio of CH3SH to H2S was significantly higher in inflamed than noninflamed sites (p<.05) and in deep versus shallow sites (p<.1). This is the first known study to quantitate VSC directly from individual crevicular sites. Furthermore, in the course of the study, methodology was developed for high performance liquid chromatography (HPLC) analysis of hydroxyproline (Hyp) from GCF using precolumn derivatization with phenylisothiocynate. The method was applied to a study of crevicular fluid collected from 30 periodontally-involved subjects that participated in a clinical study which evaluated spiramycin as an adjunct in treatment to scaling and root planing. The study compared Hyp values to several recorded clinical periodontal parameters. Analysis of hydrolyzed and unhydrolyzed samples indicated that the dominant source of Hyp was present in a peptide or bound form. Hyp did not correlate with pocket depth or crevicular fluid volume. Hyp levels measured at inflamed and noninflamed sites fluctuated at given examination points (0, 2, 8, 12, 24 weeks) and between these time points during the study. The most complete data were obtained for time points0 and 12 weeks which were used to make the best longitudinal comparison between these two time points. Accordingly, the Hyp levels in both treated groups at week twelve in periodontal sites that were inflamed at week zero and remained inflamed at week 12 were higher than in sites that remained noninflamed. This relationship was statistically significant in the spiramycin treated group (p<.05). In both treatment groups Hyp levels at week twelve were higher in healing sites that experienced a mm gain of attachment between weeks 0 and 12, than sites that remained unresolved. Again in the spiramycin treated group this difference was found to be statistically significant (p<.05). Hence the study indicated an increase in Hyp levels in GCF during episodes of increased collagen metabolism. This metabolism occurred during both healing and active phases of periodontal disease. A cross sectional study of GCF from inflamed and noninflamed periodontal sites was performed to confirm that Hyp content was higher in inflamed sites and to investigate the Hyp contribution from type I collagen and Clq sources. Again Hyp content was found to be higher ininflamed than noninflamed sites (p<.001). Using ELISA, type I collagen was found to contribute approximately 6-fold more Hyp than Clq to total Hyp content of fluid from inflamed and noninflamed sites. SDS/PAGE gels of GCF from inflamed sites exhibited more intense protein banding pattern than GCF from noninflamed sites. Western blot analyses showed markedly more intense staining for type I collagen peptides than for Clq reactive peptides. This finding, together with the ELISA assay results which indicated that Clq contributed less than 10% of total Hyp to both inflamed or noninflamed GCF, strongly implies that collagen is the dominant source of the total Hyp content in the processsed GCF samples. The presence of higher VSC content and CH3SH to H2S ratio in crevicular air and higher Hyp content in GCF, can distinguish inflamed from noninflamed periodontal sites. These indicators can distinguish the presence of disease at specific sites at the time of examination. Since Hyp levels were found to be increased at periodontal sites that exhibited healing during a 3 month period, Hyp can also ascertain the effectiveness of periodontal therapy."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/1829?expand=metadata"@en ; dcterms:extent "11833499 bytes"@en ; dc:format "application/pdf"@en ; skos:note "BIOCHEMICAL AND IMMUNOLOGICAL STUDIES OF PERIODONTALDISEASE IN HUMANS WITH EMPHASIS ON THE ANALYSES OFBREAKDOWN PRODUCTS EMANATING FROM THE GINGIVAL CREVICEbyALBERT JEFFREY MARTIN COILB. Sc.D.M.D.A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDoctor of PhilosophyinTHE FACULTY OF GRADUATE STUDIESDepartment of Oral Biology, School of DentistryWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIASeptember, 1992.©A.J.M. Coil, 1992.In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of ^Oral BiologyThe University of British ColumbiaVancouver, CanadaDate^CeptPmhe r 24, 1(1 92DE-6 (2/88)AbstractActive periodontal disease is characterized by marked destruction of collagen in periodontally-involved tissues. Biochemical characterization of the disease is complicated as it is an episodicdisorder that is believed to cycle through periods of high activity and quiescence. The first part ofthe study focused on the measurements and analysis of volatile sulphur compounds (VSC)collected at gingival sulci. Their concentrations in mouth air correlate with the severity of thedisease. The second part was devoted to the evaluation of breakdown products of collagenmetabolism found in the gingival crevicular fluid (GCF) in relation to periodontal disease.A novel method and device was developed for collection and analysis of volatile compoundsfrom specific gingival crevice sites. It demonstrated that volatile sulphur compound profiles ofcrevicular air differ from that of mouth air. It showed that total sulphur content of either inflamedor deep (PD z4 mm) periodontal sites was significantly higher than the noninflamed or shallow(PD<4 mm) sites (p<.05). The ratio of CH3SH to H2S was significantly higher in inflamed thannoninflamed sites (p<.05) and in deep versus shallow sites (p<.1). This is the first known studyto quantitate VSC directly from individual crevicular sites.Furthermore, in the course of the study, methodology was developed for high performanceliquid chromatography (HPLC) analysis of hydroxyproline (Hyp) from GCF using precolumnderivatization with phenylisothiocynate. The method was applied to a study of crevicular fluidcollected from 30 periodontally-involved subjects that participated in a clinical study whichevaluated spiramycin as an adjunct in treatment to scaling and root planing. The study comparedHyp values to several recorded clinical periodontal parameters. Analysis of hydrolyzed andunhydrolyzed samples indicated that the dominant source of Hyp was present in a peptide or boundform. Hyp did not correlate with pocket depth or crevicular fluid volume. Hyp levels measured atinflamed and noninflamed sites fluctuated at given examination points (0, 2, 8, 12, 24 weeks) andbetween these time points during the study. The most complete data were obtained for time points0 and 12 weeks which were used to make the best longitudinal comparison between these two timepoints. Accordingly, the Hyp levels in both treated groups at week twelve in periodontal sites thatwere inflamed at week zero and remained inflamed at week 12 were higher than in sites thatremained noninflamed. This relationship was statistically significant in the spiramycin treatedgroup (p<.05). In both treatment groups Hyp levels at week twelve were higher in healing sitesthat experienced a mm gain of attachment between weeks 0 and 12, than sites that remainedunresolved. Again in the spiramycin treated group this difference was found to be statisticallysignificant (p<.05). Hence the study indicated an increase in Hyp levels in GCF during episodesof increased collagen metabolism. This metabolism occurred during both healing and active phasesof periodontal disease.A cross sectional study of GCF from inflamed and noninflamed periodontal sites wasperformed to confirm that Hyp content was higher in inflamed sites and to investigate the Hypcontribution from type I collagen and Clq sources. Again Hyp content was found to be higher ininflamed than noninflamed sites (p<.001). Using ELISA, type I collagen was found to contributeapproximately 6-fold more Hyp than Clq to total Hyp content of fluid from inflamed andnoninflamed sites. SDS/PAGE gels of GCF from inflamed sites exhibited more intense proteinbanding pattern than GCF from noninflamed sites. Western blot analyses showed markedly moreintense staining for type I collagen peptides than for Clq reactive peptides. This finding, togetherwith the ELISA assay results which indicated that Clq contributed less than 10% of total Hyp toboth inflamed or noninflamed GCF, strongly implies that collagen is the dominant source of thetotal Hyp content in the processsed GCF samples.The presence of higher VSC content and CH3SH to H2S ratio in crevicular air and higher Hypcontent in GCF, can distinguish inflamed from noninflamed periodontal sites. These indicatorscan distinguish the presence of disease at specific sites at the time of examination. Since Hyplevels were found to be increased at periodontal sites that exhibited healing during a 3 monthperiod, Hyp can also ascertain the effectiveness of periodontal therapy.iiiTable of ContentsAbstract ^iiList of Tables ^viiList of Figures^ viiiAcknowledgement 1. INTRODUCTION1.1 ORAL CAVITY1.1.1 Structure 11.1.2 Environment ^ 11.2 GINGIVAL CREVICE1.2.1 Structure and function 21.2.2 Gingival vasculature and permeability 41.3 GINGIVAL CREVICULAR FLUID1.3.1 Production ^ 51.3.2 Composition 71.3.3 Methods of collection ^ 81.3.4 Volume ^ 111.3.5 Clinical significance 121.4 CONNECTIVE TISSUE IN THE PERIODONTIUM1.4.1 Collagen structure ^ 151.4.2 Collagen types 161.4.3 Intracellular and extracellular metabolism ^ 211.4.4 Non-collagenous components ^ 321.5 VOLATILE SULPHUR COMPOUNDS1.5.1 Substrates ^ 381.5.2 Bacteria 391.5.3 Analyses of volatile compounds produced in the oral cavity ^ 391.5.4 Implications in periodontal disease ^ 411.6 PERIODONTAL DISEASE1.6.1 Health versus disease state 431.6.2 Disease progression ^ 481.6.3 Clinical evaluations 511.6.4 Host interactions 54iv1.6.5 Treatment ^581.7 INDICATORS OF PERIODONTAL DISEASE1.7.1 Clinical indicators ^601.7.2 Bacterial indicators ^611.7.3 Enzyme indicators ^651.7.4 Immunological and inflammatory indicators ^ 701.7.5 By-products of metabolism ^741.8 PROPOSED STUDY1.8.1 Crevicular volatile sulphur production ^771.8.2 Hydroxyproline content of crevicular fluid as an indicator of disease activity 771.9 SIGNIFICANCE ^ 782. MATERIALS AND METHODS2.1 MATERIALS 792.2 VOLATILE SULPHUR COMPOUND STUDY2.2.1 Development of a collection device for sampling crevicular air ^802.2.2 Experimental design ^822.2.3 Collection of crevicular air ^832.2.4 Gas chromatographic analysis of crevicular air ^842.2.5 Identification and quantitation of volatile sulphides ^842.2.6 Data analysis ^862.3 DEVELOPMENT OF HPLC HYDROXYPROLINE ANALYSIS^ 872.4 SPIRAMYCIN STUDY2.4.1 Experimental design ^922.4.2 Sample collection and measurement ^932.4.3 HPLC analyses for hydroxyproline ^ 932.4.4 Data analysis ^962.5 INFLAMED AND NONINFLAMED PERIODONTAL SI'Z'ES STUDY2.5.1 Experimental design ^982.5.2 Clq antibodies^ 1002.5.3 Type I collagen antibodies^ 1032.5.4 Crevicular fluid model 1062.5.5 Sample collection and measurement ^ 1082.5.6 HPLC analyses for Hyp ^ 1092.5.7 ELISA analyses 1102.5.8 SDS/PAGE gel electrophoresis ^1112.5.9 Western blotting ^ 1132.5.10 Data analysis ^ 1153. RESULTS3.1 VOLATILE SULPHUR COMPOUNDS IN HUMAN GINGIVAL CREVICE3.1.1 Volatile sulphur composition ^ 1163.1.2 Deep versus shallow sites and inflamed versus noninflamed sites . . . ^ 1213.2 HYDROXYPROLINE LEVELS IN GINGIVAL CREVICULAR FLUID ANDSPIRAMYCIN EFFECT ON PERIODONTAL SITES3.2.1 Hydroxyproline content of hydrolyzed and unhydrolyzed samples^. 1273.2.2 Hydroxyproline levels in inflamed and non-inflamed sites ^ 1283.2.3 Hydroxyproline levels versus attachment levels ^ 1423.2.4 Spiramycin effect on periodontal measurements 1493.3 BIOCHEMICAL ANALYSES OF CREVICULAR FLUID FROM INFLAMED ANDNONINFLAMED PERIODONTAL SITES3.3.1 Hydroxyproline levels in inflamed and noninflamed sites ^ 1503.3.2 Clq content in crevicular fluid model ^ 1503.3.3 Clq content in crevicular fluid 1523.3.4 Type I collagen in crevicular fluid model 1543.3.5 Type I collagen in crevicular fluid^ 1553.4 IMMUNOLOGICAL FINDINGS IN CREVICULAR FLUID FROM INFLAMED ANDNONINFLAMED PERIODONTAL SITES3.4.1 SDS/PAGE gels ^ 1613.4.2 Western blotting 1624. DISCUSSION4.1 VOLATILE COMPOUNDS IN THE ORAL CAVITY ^ 1684.2 HYDROXYPROLINE AS A MEASURE OF PERIODONTAL TISSUE^INTEGRITY 1734.3 CONTRIBUTION BY Clq AND COLLAGEN TO THE HYDROXYPROLINECONTENT OF GINGIVAL CREVICULAR FLUID^ 1825. CONCLUSIONS 1876. RECOMMENDATIONS FOR FUTURE STUDY 1887. BIBLIOGRAPHY ^ 1898. APPENDIX^Abbreviations used in this thesis ^ 216viList of Tables1.1 Genetically distinct collagens ^ 171.2 Distribution of collagen types in periodontal structures ^ 202.1 HPLC gradient program for separation of PITC-derivatized amino acids ^ 962.2 Sequence of events for developing ascites fluid in Balb/c mice ^ 1022.3 Percent specificities of monoclonal and polyclonal Clq antibodies 1032.4 Percent specificities to Chemicon's polyclonal rabbit anti-human type I collagen 1042.5 Sources of antibody to type I collagen showing percentage specificity with relatedproteins ^1053.1a Levels of sulphides found in 10 ml of mouth air^ 1193.1b Levels of sulphides found in 10 ml of crevicular air 1203.2 Frequency of VSC in mouth and crevicular air from control (C) and test (1) sites . . 1233.3 Levels of VSC per 10 ml of mouth or crevicular air at control (C) and test (T) sites . 1233.4 Total Hyp levels in both the spiramycin (Drug = 1+) and non-drug (Drug ='-')treatment groups ^1333.5 Weighted Hyp levels in both the spiramycin (Drug ='+') and non-drug (Drug ='-')treatment group ^ 1333.6 Hydroxyproline values at week zero time point for inflamed and noninflamed sitesfor both spiramycin and non-drug treatment groups combined ^ 1383.7 Hydroxyproline values at various exam points for both spiramycin (Drug='+') andnon-drug (Drug='-') treatment groups ^ 1393.8 Hydroxyproline values at week zero time point for both shallow and deep pocketssubdivided into inflamed or noninflamed, for spiramycin (Drug='+'),non drug (Drug='-'), and all (Drug='+ and -') treatment groups^ 1433.9 Hydroxyproline values and bleeding on probing measurements at weeks 0 and 12 inspiramycin treatment group ^ 1473.10 Hydroxyproline values and bleeding on probing measurements at weeks 0 and 12 innon-drug treatment group 1473.11 Periodontal measurements at weeks 0 and 12 for both drug and non-drug groups 149viiList of Figures1.1 Attached epithelial cuff and gingival sulcus at an early stage of tooth eruption^31.2 Intracrevicular sampling ^101.3 Extracrevicular sampling ^111.4 Schematic diagram showing the major steps in collagen biosynthesis ^ 222.1 Standard plot for H2S of integrated peak area versus concentration ^852.2 Standard plot for CH3SH of integrated peak area versus concentration ^862.3 HPLC profile of PITC-derivatized amino acid mixture ^ 902.4 HPLC profile of PITC-derivatized crevicular fluid ^912.5 Standard plot of Periotron readings versus fluid volume usingfetal calf serum-soaked Periotron filter paper strips^ 942.6 Standard plot for high concentrations of hydroxyproline ^972.7 Standard plot for low concentrations of hydroxyproline ^982.8 Flow diagram of Clq precipitation experiment ^ 1073.1 Comparative profiles of tenax-trapped mouth and crevicular air from the same subject 1173.2 Comparative profiles of tenax-trapped crevicular air from two different subjects . . . 1183.3 Ratios of CH3SH to H2S and total methyl sulphides (CH3-S's)to H2S from control (C) and test (1) sites of crevicular air ^ 1253.4 Ratios of CH3SH to H2S of control (C) and test (T) and noninflamed (NI)versus inflamed (1) crevicular sites ^ 1263.5 Total sulphur content of {H2S, CH3SH, (CH3)2S, (CH3S)2} in crevicular air at control(C) verus test (T); and noninflamed (NI) versus inflamed (I) sites ^ 1273.6 HPLC profile of unhydrolyzed and PITC-derivatized crevicular fluid sample^ 1283.7 HPLC profile of hydrolyzed and PITC-derivatized crevicular fluid sample 1293.8 Total Hyp levels at periodontal sites from the spiramycin treatment groupat all examination points of the study^ 1343.9 Total Hyp levels at periodontal sites from the non-drug treatment groupat all examination points of the study 1353.10 Weighted Hyp levels at periodontal sites from the spiramycin treatmentgroup at all examination points of the study ^ 1363.11 Weighted Hyp levels at periodontal sites from the non-drug treatmentgroup at all examination points of the study 1373.12 Hydroxyproline levels in inflamed and noninflamed periodontal sites at week zerotime point^ 1383.13 Total Hyp levels at all examination points for both the spiramycinand non-drug treatment groups ^ 140viii3.14 Weighted Hyp levels at all examination points for both the spiramycinand non-drug treatment groups ^ 1413.15 Hydroxyproline levels for combined drug treatment groups at week zero forshallow and deep pockets 1443.16 Total Hyp levels for both drug treatment groups at week zero forboth bleeding on probing categories and shallow and deep pocket depths ^ 1453.17 Weighted Hyp levels for both drug treatment groups at week zero forboth bleeding on probing categories and shallow and deep pocket depths ^ 1463.18 Hydroxyproline levels versus changes in attachment level between weeks zeroand twelve for drug and non-drug groups ^ 1483.19 Weighted Hyp levels in inflamed and noninflamed periodontal sites asmeasured by HPLC ^ 1513.20 Type I collagen and Clq in a crevicular fluid model ^ 1523.21 ELISA results using a polyclonal antibody to Clq 1543.22 Detection of degraded collagen using ELISA, after one and two hour incubations withbacterial collagenase (A) and P. gingivalis blebs (B) ^ 1563.23 ELISA results using a polyclonal antibody to type I collagen 1583.24 Moles {Exp -12} of type I collagen and Clq, per p1 of crevicular fluid, in bothinflamed and noninflamed periodontal sites ^ 1593.25 Moles of Hyp {Exp -12} contributed by type I collagen and Clq, per p1 of crevicularfluid, from inflamed and noninflamed, periodontal sites ^ 1603.26 Crevicular fluid Hyp contributions from type I collagen and Clq frominflamed and noninflamed periodontal sites as measured by ELISA ^ 1613.27 SDS/PAGE gel of type I collagen and Clq standards ^ 1633.28 SDS/PAGE gel of native and denatured patient matched pooled GCF from inflamedand noninflamed periodontal sites ^ 1643.29 Western blot of type I collagen and Clq standards ^ 1653.30 Western blot of anti-human type I collagen reacted with denatured patient matchedpooled GCF from inflamed and noninflamed sites 1663.31 Western blot of anti-human Clq reacted with denatured patient matched pooled GCFfrom inflamed and noninflamed sites ^ 1674.1 Type I collagen and Clq components in whole, precipitate and supernatantfractions conceptualized to occur in acetate treated crevicular fluid ^ 186ixAcknowledgementsThroughout my study and especially during the preparation of this thesis, I wish to express mysincere gratitude to my supervisor, Dr. Joseph Tonzetich, for his advice, guidance and patience. Iam indebted to him for the many ways he has contributed to my education. He has been atremendous source of knowledge and wisdom and I truely appreciated his encouragement.I would like to thank the members of my committee for their advice and direction of the project.I especially would like to thank Dean Paul Robertson for his valuable assistance and guidance indata analyses of the spiramycin study. Thanks also to Dr. George Beagrie for is encouragementand support.Thanks are extended to the members of the laboratory for creating a helpful and pleasantenvironment in which to work. Special thanks are extended to Mr. Tony Ng who has shared manyof the frustrating experiences involving HPLC malfunctions.I would like to thank Sigrid, my wife, and my parents for their love, support andencouragement throughout the study, especially during difficult times. Thanks also go to myfriends at Granville Chapel who have prayed for me during my study.Finally, I would like to acknowledge the financial support which I received from the MedicalResearch Council of Canada and The B.C. Health Research Foundation.11. INTRODUCTION1.1 ORAL CAVITY1.1.1 StructureThe oral cavity is a unique mixture of specialized hard and soft tissues. Teeth are comprised ofenamel, dentin and cementum, and house a specialized connective tissue, the dental pulp. Salivaryglands, either major or minor, are found in various parts of the mouth. The oral cavity has amucosal covering which is clinically and biologically classified as masticatory, lining andspecialized types. They vary in structure depending on the function of the region concerned.Mucosal covery of the oral cavity has two basic components: epithelium and the connectivetissue lamina propria. Separating the two components is the basal lamina, an epithelial productformed by the basal epithelial cells. Papillae of connective tissue project toward the epithelium,which in turn has ridges that protrude into the lamina propria. The basal lamina forms an interfacebetween the connective tissue papillae and epithelial ridges (Bhaskar 1980).Depending upon the region, the stratified squamous epithelial lining of the oral cavity can bekeratinized, non-keratinized or parakeratinized (Cleaton-Jones and Fleisch 1973). The surfaces ofmore functionally stressed areas of the mouth such as the hard palate and gingiva are protected witha keratinized epithelium. However, in many individuals the gingival epithelium is parakeratinized(Bhaskar 1980). Gingival crevicular epithelium as well as delicate locations like alveolar, buccaland vestibular mucosae are composed of non-keratinized epithelium.1.1.2 EnvironmentThe environment of the oral cavity is complex having numerous anaerobic and aerobic sites.The emergence of teeth into the oral cavity results in the formation of a gingival crevice and an evenmore complex microenvironment. The gingival crevice is bounded by soft tissue on one side and2mineralized tissue on the other. From physical, metabolic and nourishment perspectives, thegingival crevice offers an unique environment where a wide variety of microorganisms flourishes.The superior aspect of the gingival crevice offers an aerobic environment, while deeper portionsallow for anaerobic conditions . Microorganisms can attach to either the surface of the sulcularepithelium or to the enamel or cemental surfaces of the teeth. In addition, the gingival sulcus isbathed in a fluid that emanates from the underlying connective tissue adjacent to the junctionalepithelium.1.2 GINGIVAL CREVICE1.2.1 Structure and functionThe gingival epithelium is divided into several subclasses based upon anatomical location andhistological characteristics. Included in a typical cross section of marginal gingiva are crevicular,junctional (attachment), as well as lining epithelia of free and attached gingivae and of alveolarmucosa (Brill and Krasse 1958). In particular, the crevicular epithelium extends from the gingivalmargin to the junctional epithelium, and forms a physical and biological boundary of the gingivalcrevice surrounding each tooth.As the tooth erupts into the oral cavity, the reduced enamel epithelium of the emerging crownfuses with the oral epithelium. The reduced enamel epithelium (REE) that remains attached to theenamel of the tooth which has not yet erupted is termed the primary attachment epithelium. As thetooth erupts the REE becomes shorter thus causing a development of a shallow groove, thegingival crevice, between the gingiva and the surface of the tooth (Figure 1.1). This gingivalsulcus extends from the apical attachment epithelium to the coronal gingival margin (Bhaskar1980).3EnamelGingival sulcusFree gingivaOral Epithelium^Attachment Epithelium(from Reduced enamelepithelium)^ CementoenameljunctionPulpFigure 1.1:^Attached epithelial cuff and gingival sulcus at an early stage of tooth eruption(Bhaskar 1980).The non-keratinizing crevicular epithelium consists of three layers of cells, namely the stratumbasale, stratum intermedium and stratum superficiale (basal, intermediate and superficial). Basalcells of both keratinizing and non-keratinizing epithelia are the same morphologically; however, theintermediate cells of non-keratinized epithelium are larger than those of the stratum spinosum foundin keratinized epithelium, and contain no granular or cornified cells (Bhaskar 1980).4A number of authors have microscopically studied the cytostructure of the crevicular epitheliumand have found it to resemble non-keratinizing oral mucosa (Listgarten 1964; Listgarten 1966;Schroeder and Theilade 1966; Weinstein et al. 1967; Freedman et al. 1968; Gavin 1968;Geisenheimer and Han 1971; Kaplan et al. 1977; Muller-Glauser and Schroeder 1982). The basalcells have a central ovoid nuclei bounded by a porous nuclear membrane, numerous basally placedmitochondria, Golgi apparartus in close proximity to the nucleus, and abundant amounts ofpolyribosomes. These features are suggestive of active protein synthesis. Intermediate cell layersshow a decrease in the size of the nucleus and number of mitochondria, and also lack keratohyalingranules. These flattened cells display an increase in the number of tonofilaments, but they have notendency to form prominent bundles as in keratinizing gingival epithelium. In the superficial layerof cells, cytoplasmic organelles are scarce and appear in a degenerative state. The cells lack a welldefined nucleus, contain electron dense bodies, varible sized vesicles and intact desmosomes.Crevicular epithelium is a specialized lining epithelium which is formed at the time of activeeruption of the dentition and is lost when the teeth are exfoliated. It forms the soft tissue wall ofthe gingival crevice together with the junctional epithelium, and opposes the hard structure of thetooth. While this specialized structure has an important function of maintaining a barrier againstmasticatory forces and bacterial plaque toxins, it lacks a protective keratin layer.1.2.2 Gingival vasculature and permeabilityThe lamina propria contains the vascular supply, innervation and lymphatics of the gingivaltissues. Ultrastructurally, individual capillary loops are present in each of the connective tissuepapillae. The gingival blood supply is derived from the supraperiosteal arterioles along the buccaland lingual surfaces of the alveolar bone, the arterioles that emerge from the crest of the interdentalalveolar bone and to some extent from the arterioles within the periodontal ligament (Cohen 1954).The permeability of crevicular epithelium to tissue fluid has been widely studied (Brill andKrasse 1958; Brill and Bjorn 1959; Brill and Krasse 1959; Thilander 1964). Using a fluoresceintracer, Brill and Bjorn found that tissue fluid permeates healthy gingival pocket tissues in dogs.5With the same tracer compound, Brill and Krasse observed that except for crevicular epithelium,fluorescein did not penetrate other regions of healthy oral mucosa.In all three layers of crevicular epithelium, the intercellular spaces are wide and generallyoccupied by leukocytes, with neutrophils being the most common type. Some of these cellsrupture as they migrate towards the surface of the crevicular epithelium. Thilander reported ageneral widening of the intercellular spaces after application to the pocket epithelium of leukocytehomogenate, containing proteases, amylases, peroxidases and lipases (Thilander 1964). This wasascribed to an effect on the intercellular substance and the mutual adhesion mechanism of the cells.Studies of the intercellular matrix of the oral mucosa have suggested that it is comprised of mainlypolysaccharides and proteins (Wislocki et al. 1951; Thonard and Scherp 1959; Cancellaro et al.1961; Klingsberg et al. 1961). It is believed that an enzymatic degradation of such constituentsresults in an increase in the uptake of water to maintain constant tonicity. This weakening of theadhesion effect may be responsible for widening of intercellular spaces (Thilander 1964).In general, the intercellular compartment tends to be narrowest in the intermediate layer,slightly wider in the basal layer and widest in the most superficial layer (Geisenheimer and Han1971). These continuous intercellular spaces provide an explanation for the passage of tissue fluidthrough the crevicular epithelium and not through other regions of healthy mucosa.1.3 GINGIVAL CREVICULAR FLUID1.3.1 ProductionThere are differences in the ratios of components between crevicular fluid and serum fromwhich it is derived (Weinstein et al. 1967). This indicates that the fluid is not a transudate, butrather an inflammatory exudate (Oliver et al. 1969; Orban and Stallard 1969). This apparentselectivity can be related to two morphological features of crevicular epithelium (Gavin 1968).First, the basal lamina appears to form the only continuous barrier between the connective tissue6and the gingival pocket. Secondly, the presence of prominent microvilli suggests a modification inconstituents of the intercellular spaces.In 1971, Geisenheimer and Han, using electron microscopy, (E.M.), observed the presence oftypical desmosomes in all cell layers of the crevicular epithelium (Geisenheimer and Han 1971).These intercellular bridges were most numerous in the intermediate layer, followed by thesuperficial layer, and the least in the stratum basale. They also found hemidesmosomes irregularlyspaced along the basal surface of the basal cells, at the interface between crevicular epithelium andlamina propria.An electron microscopic study by the same investigators showed that in marginal gingiva ofclinically healthy tissue the manifestations of inflammation, such as dense cellular infiltrate, dilatedvessels, ruptured cells and large amounts of edematous exudate are not observed (Geisenheimerand Han 1971). Their observations also indicated that clinically normal marginal gingiva willappear free of manifestations of inflammation when rigid standards of oral hygiene and tissueselection are applied.A stratified squamous epithelium is considered to be an effective barrier to certain substances.It is most permeable to small lipid soluble non-polar materials (Thilander 1964). However, pocketepithelium does not display the same properties, as large polar serum proteins can pass through thecrevicular epithelium (Brill and Krasse 1958; Brill and Bjorn 1959). Furthermore, pocketepithelium is thinner than the gingival epithelium and lacks a protective keratin layer.Holmberg and Killander (Holmberg and Killander 1971) and others (Mann 1963; Egelberg1964; Bjorn et al. 1965; Loe and Holm-Pedersen 1965; Hara and Liie 1969; Oliver et al. 1969),found gingival fluid flow higher in subjects with severe periodontitis. Extracellular passage ofserum protein through crevicular epithelium is the most important route of flow in an inflamedgingiva, which exhibits widening of the intercellular spaces (Thilander 1963; Thilander 1968).Since similar concentrations of immunoglobulins were found in gingival fluid from inflamedgingiva as in serum, these investigators concluded that serum was the origin of various crevicularproteins such as IgA, IgG and IgM.7Crevicular epithelium has been described in relation to the periodontal pocket and underlyingconnective tissue. Its unique ability to allow passage of tissue fluid through its layers has beenextensively studied. It is of immense interest how such a permeable mucosa can simultaneouslyact as an effective barrier to exogenous substances.1.3.2 CompositionWhether the gingiva is inflamed or healthy is of crucial importance when evaluating the sulcularregion. A potential problem in studying the gingival sulcus is an inadequate definition of the healthstatus of the epithelium. The degree of inflammation has an effect on the turnover of the epithelialcells (Marwah et al. 1960). Conflicting results amongst researchers may in part be due to thestudies being performed on gingival tissue specimens which have been classified normal andchronically inflamed .Egelberg reported that the number of inflammatory cells was increased in inflamed gingivalpockets (Egelberg 1963). However, the proportion of inflammatory cells appeared to beindependent of the degree of inflammation. Differential counts of inflammatory cells showed 95-97% neutrophils, 1-2% lymphocytes, and 2-3% monocytes (AttstrOm 1970). Although theproportion of leukocytes was found to be independent of the severity of inflammation, a positiverelation was observed between the number of cells and the presence of inflammation.Chemical analysis and observed ratios of various components of crevicular fluid differ fromthose of the serum samples. The Na+/K+ ratio of 2/1, which is much lower than in serum,support the earlier results of Krasse and Egelberg (Krasse and Egelberg 1962). The ratios ofprotein/ chloride and of protein/inorganic phosphorous were also reported much lower in serum.This indicated a decreased presence of higher anionic protein than in serum. Furthermore, sincethe proteins in the gingival fluid did not react with an anti-saliva sera, this implied that they werenot derived from saliva.Hara and LOe (Hara and Lit 1969), reported on the glucose, hexosamine and hexuronic acidcontent in crevicular fluid. They found glucose concentration in crevicular fluid three to six times8greater than in serum. The presence of hexosamine and hexuronic acid may represent breakdownproducts of bacterial action on host tissues. Similar substances are liberated during theinflammatory process by enzymatic degradation of the gingival connective tissue.Components in gingival fluid were found to react with specific antisera to alpha G-, alpha A-and alpha M-globulins, albumin and fibrinogen (Brandtzaeg 1965). This suggested that plasmawas the source of these proteins. Also, these components migrated upon electrophoresis, just asplasma proteins did, and were present in concentrations characteristic of plasma. Other reportshave also indicated that the concentrations of IgG, IgA and IgM in crevicular fluid were verysimilar to serum (Holmberg and Killander 1971).A significant amount of research has been devoted to enzyme characterization of crevicularfluid. Acid phosphatase, beta-glucuronidase, lysozyme, cathepsin D, proteases, alkalinephosphatase, and lactic dehydrogenase have all been found in gingival crevicular fluid (Brandtzaegand Mann 1964; Sueda and Cimasoni 1968; Ishikawa and Cimasoni 1970). These enzymes,however, could have originated from either the host tissue and/or the plaque microflora.1.3.3 Methods of collectionSampling of gingival fluid has been performed by a variety of methods. These have includedcollection of subgingival washings, and collection using either capillary tubes or filter paper strips.The gingival washing technique proved to be quite cumbersome and time-consuming. Thiscollection procedure was improved by using microcapillary tubes, which, in turn, more recentlyhas been superseded by the use of filter paper strips. These advances simplified the collection andhandling of the samples.Sampling with capillary tubes provided the potential for quantitative analysis of a knownvolume of crevicular fluid. Initially the use of such micropipettes was merely to collect the fluidwithout evaluating the volume or the exact procedure for sampling (Krasse and Egelberg 1962;Mann 1963; Brandtzaeg and Mann 1964). Egelberg (Egelberg 1963), described his capillary tubesampling procedure as simply bringing the tube into the pocket and moving it slightly in mesial and9distal directions, two to four times. Kaslick and workers (Kaslick et al. 1968), described atechnique for exact measurement of gingival fluid involving collection of fluid with capillary tubesof known inner diameter, followed by centrifugation and removal of sediment from the sample.This technique for collection and measurement of fluid volume has been subsequently adopted byother investigators (Sueda et al. 1969).More recently, two filter paper strip methods have been employed to sample crevicular fluid.The first, referred to as the intracrevicular technique, involves the gentle insertion of the filter paperinto the gingival sulcus. In the second, known as extracrevicular sampling, the strips arepositioned on the vestibular surfaces of the gingivae and teeth, without penetrating the sulcus. Thelatter method is preferred for repetitive sampling from the same site as it does not disturb thesulcular environment.The method of intracrevicular sampling was first described by Brill and co-workers (Brill andKrasse 1958; Brill 1959; Brill and Bjorn 1959). Filter paper strips used to absorb the fluid wereinserted into the gingival crevice until minimal resistance was felt (Figure 1.2). This method wascommonly used among researchers during the 1960's and early 1970's (Bjorn et al. 1965; Lindheand AttstrOm 1967; Weinstein et al. 1967; Hara and Ltie 1969; Golub et al. 1971; Sandalli andWade 1971; Golub et al. 1976; Ciancio et al. 1980). The strips were inserted in the crevices for asuitable sampling time before being removed and analyzed.A slight variation of this method was introduced by Mann (Mann 1963). By inserting threefilter paper strips into each crevice he not only limited the sampling to a specific area but alsoeliminated the chance of contamination by saliva. These three strips were used to dry the creviceand remained side by side in the crevice for three minutes. The central drying strip was removedand immediately replaced with a collecting strip. After five minutes the strip was removed forevaluation. This variation insured that any fluid collected after the drying sequence must originatefrom within the gingival sulcus.10Figure 1.2:^Intracrevicular sampling. 'A': For the Brill technique, the strip is passed into thecrevice until resistance is felt. 'B': In the 125e and Holm-Pedersen technique, the end of the strip isplaced at the entrance to the crevice. (Cimasoni 1983)Yet another modification of the intracrevicular technique was to sample only at the sulcularorifice. In this technique, filter paper strips are placed at the entrance of the crevice, avoidingphysical irritation of the crevicular epithelium (Liie and Holm-Pedersen 1965). Egelberg andAttstibm (Egelberg and AttstrOm 1973), used this technique in preference to the initialintracrevicular method (Figure 1.2). They found that while the intracrevicular samples yieldedgreater volumes of fluid than orifice samples, this was not an advantage, since the orifice methodshowed less variation between the samples. This technique has since been adopted by otherinvestigators (Egelberg 1966a-e; Waite 1976; Jameson 1979).11In addition to intracrevicular sampling, Brill and Krasse (Brill and Krasse 1958) described theextracrevicular method. Accordingly, strips of filter paper were placed against the vestibularsurfaces of the teeth and gingivae. This method has been used by other researchers (Liie andHolm-Pedersen 1965), and is illustrated in Figure 1.3.Figure 1.3:^Extracrevicular sampling. The strip fits closely to the tooth surface, gingivalmargin and attached gingiva, thus bridging the entrance of the gingival sulcus. (Cimasoni 1983)1.3.4 VolumeThe amount of crevicular fluid collected using the available techniques is quite small and onoccasion difficult to detect. In order to distinguish the presence of crevicular fluid on filter paperstrips, Brill and Bjorn (Brill and Bjorn 1959) and others (Mann 1963; Weinstein et al. 1967), have12utilized fluorescein which passes from the systemic circulation to the oral cavity via the gingivalsulcus, and is not found in the saliva. In these studies, subjects were administered sodiumfluorescein prior to crevicular fluid collection. After fluid collection, fluorescence of the filterpaper strips was viewed under ultraviolet light. This technique can be employed to detect GCFwhen small amounts of fluid are present.Another method of evaluating crevicular fluid volume was effected by staining the filter paperstrips. This technique, used to accentuate small volumes of gingival crevicular fluid (GCF) onfilter paper strips, utilized ninhydrin staining (Orban and Stallard 1969). The resulting blue orpurple colour, indicative of reaction with alpha-amino groups of amino acids, can be quantitatedmicroscopically using a suitable grid pattern.More recently, fluid soaked filter paper strips are analyzed electronically. In Jameson's studies(Jameson 1979), a gingival crevice fluid meter was employed. This instrument discriminatesminute fluid volumes by the reduction in capacitance between two sensors when in contact with astandardized filter strip containing fluid under investigation. Use of this instrument had thepotential of yielding a quick and suitable method for repeated fluid volume determinations.Furthermore, it obviated a need for injection of subjects with fluorescein.The most widely used instrument today for analyzing crevicular fluid volume on filter paperstrips is the Periotron 6000TM. This device measures conductance of a fluid absorbed filter paperstrip when placed between two electrodes. A digital readout between 0 to 200 is expressed, whichis then related to a standard curve of known volume and corresponding Periotron values. In thepast decade, the majority of crevicular fluid studies have performed crevicular fluid collection usingfilter paper strips, and volume measurements using a Periotron instrument.1.3.5 Clinical significanceIt was determined that crevicular fluid flow followed a circadian periodicity with the highestflow rate occuring in the late evening (Bissada et al. 1967). Using the Brill technique (Brill andKrasse 1958), Bissada and co-workers observed that after oral administration of sodium13fluorescein the maximum flow followed the evening crest in body temperature by about fourhours. Moreover, a considerable variation was found in rate of flow between different individualsand also between different crevices of the same individual.The gingival fluid flow during the menstrual cycle was studied by Lindhe and AttstrOm (Lindheand AttstrOm 1967). They observed a small gradual increase of gingival exudation during theproliferative phase and a peak on the day of ovulation. Following ovulation the secretory phaseshowed a gradual decrease in gingival fluid flow. These findings suggested that perhaps anincrease in female sex hormones results in a relatively higher gingival exudation, whereas a lowergingival fluid flow occurs during menstruation when the availability of female sex hormones is atthe lowest level.Numerous investigators examined the influence of diabetes upon the state of health of theperiodontium. In comparing diabetic patients to healthy controls it was suggested that the severityand duration of diabetes appeared to have little effect upon periodontal diseases (Hove and Stallard1970). No significant differences in pocket depth or gingival inflammation were found in a groupof diabetic and control patients with similar plaque indices. While no difference was found in theseverity of the inflammatory infiltrate in the two groups of patients, histological sections displayeda significantly higher incidence of vascular modifications in gingiva of diabetic patients (Cimasoni1983). In contrast, Ringelberg and co-workers (Ringelberg et al. 1977) reported a higher flow rateof GCF in a group of diabetic children than in a control group of children.There are conflicting reports as to the possible variation attributable to diabetes in theconcentration of glucose in the GCF. In an early study which compared serum and GCF levels ofglucose, it was determined that GCF glucose was up to six-fold higher than that found in serum(Hara and Lizie 1969). Subsequently, Ficara and co-workers showed similar concentrations ofglucose in GCF and serum of healthy and diabetic persons, especially in subjects exhibiting highblood glucose levels (Ficara et al. 1975). However, Weinberg and co-workers reported thatglucose levels in GCF from clinically inflamed sites were only approximately 10% of that of serum(Weinberg et al. 1986).14A positive correlation has been reported between the degree of gingival inflammation and theamount of gingival fluid flow (Brill 1960; Mann 1963; Bjorn et al. 1965; LOe and Holm-Pedersen1965; Orban and Stallard 1969; Rtidin et al. 1970; Wilson and McHugh 1971). This correlationhas been confirmed by the majority of histological findings.Perhaps of even greater significance is the fact that changes in gingival fluid flow rate may be asign of subclinical inflammation. A longitudinal study of experimental gingivitis demonstrated agradual increase in GCF flow even before clinical signs of gingivitis were evident (L6e and Holm-Pedersen 1965). This pre-gingivitis increase in the GCF flow prior to clinical evidence of diseaseduring experimental development of inflammation has also been observed by several otherinvestigators (LOe et al. 1967; Son et al. 1971; Egelberg and Attstr6m 1973; Greenstein 1984). Itis believed that these increases in GCF flow are a result of early changes in microvascularpermeability in vessels adjacent to the junctional epithelium (SOderholm and AttstrOm 1977).Both cross-sectional and longitudinal studies have compared GCF volume with measuredperiodontal parameters. Both approaches have found strong correlations between the amount ofgingival fluid collected and the Gingival Index (GI) scores (Bjorn et al. 1965; LOe and Holm-Pedersen 1965; Oliver et al. 1969; Wilson and McHugh 1971; Holm-Pedersen et al. 1975;Hancock et al. 1979; Kowashi et al. 1980; Solis-Gaffar et al. 1980). Moreover, gingival fluidflow was found to be dependent upon the degree of gingival inflammation but independent ofpocket depth (Mann 1963; Suppipat et al. 1977).The recognition that bacteria are a factor in the etiology of periodontal diseases created aninterest in the use of antimicrobial drugs in periodontal therapy. Early work by Bader andGoldhaber (Bader and Goldhaber 1966), demonstrated that systemically administered tetracyclinein dogs quickly emerged in the gingival fluid. The fmding led to the use of antibiotic drugs inperiodontal therapy. A variety of antibiotic drugs in conjunction with surgical and non-surgicalperiodontal therapy were evaluated (Mills et al. 1979; Ciancio et al. 1980; Walker et al. 1981).More recently, the fmding that tetracycline and its derivatives inhibit collagenase activity hasinfluenced their use in periodontal treatment (Golub et al. 1984; Golub et al. 1987).151.4 CONNECTIVE TISSUE IN THE PERIODONTIUM1.4.1 Collagen structureThe extracellular matrix is composed of two main classes of macromolecules; the collagens andthe polysaccharide glycosaminoglycans (GAG's). Collagen is the major structural protein of theextracellular matrix and is the most abundant protein found in mammals. GAG's are usuallybound to a core protein to form proteoglycans. Although collagens are the major proteins found intissue they should not be considered as the only molecules of importance, since the function of theextracellular matrix is dependent on the structure of all its various components.Collagen has a characteristic triple stranded helical structure. This rod-like conformationconfers onto collagen its strength and rigidity, especially when individual collagen molecules arearranged in fibers and bundles. This helical structure, which accounts for 95% of the total collagenmolecule, is composed of three separate alpha chains, each of which contains approximately 1000amino acid residues, twisted together to form a left handed triple helix. Each collagen molecule isstabilized by interchain hydrogen bonds.Formation of each component alpha chain is effected intracellulary. Each alpha chain follows arepeating amino acid pattern. Every third amino acid residue is a glycine residue (-333 residues /chain). The repeating sequence of amino acids is Gly-X-Y, where X and Y can be any amino acid.Generally the X position is occupied by a proline residue, whereas the Y position is frequentlyoccupied by hydroxyproline. Except for proline (Pro) and hydroxyproline (Hyp), no other aminoacid was found to exist in collagen in any recognizable pattern (Piez 1963; Hulmes et al. 1973).These amino acids each occur at about 100-120 sites per collagen molecule. Proline comprisesapproximately 12% of the amino acid residues in collagen, whereas Hyp is present in slightlylesser amounts, comprising 10 - 12% of the residues. These cyclic imino acids give fidigity to thewhole collagen molecule and influence the left-handed helical structure of the individual alphachains. Hydroxyproline residues not only stabilize the triple helical structure of collagen, but via16hydrogen bonds, cross link adjacent collagen molecules and contribute to stabilization of thestructure (Ramachandran et al. 1975). Thus, this repeating amino acid configuration is requiredfor stability and proper helical conformation.In addition to the helical portion of the collagen molecule, there are short non-triple-helicalregions found as extensions at the NH2- and COOH- terminal ends of each alpha chain (Traub andPiez 1971). These so-called \"extension\" domains have functional roles. They are the areas wherehydroxylysine is found which contributes to intra- and intermolecular cross-links in native fibrils.Such cross links stabilize the molecular arrangement of collagen and greatly decrease the solubilityof these molecules.1.4.2 Collagen typesThere are at least fifteen genetically distinct types of collagen which have been thus faridentified in the human body. The most useful way to classify collagen type has been to use asystem based upon the function of the collagen form. It is practical to group the collagens into twomain groups. The two classifications of collagens are those that form periodically banded fibersand those that do not (Burgeson and Morris 1987). A list of the known genetically distinctcollagens is given in Table 1.1.lal(1\\)12a2M0{al(V) }2a2(V){al(V)}3{a3(V)}3al(V)a2(V)a3(V){ al (V1)}2a2(V1){al(V11)}3{al(V1:11)}3al (IX)a2(IX)a3(IX)Type IV(Basement membrane collagen)Type VType VIType VII(long chain collagen)Type VIII(endothelial cell collagen)Type IX(HMW-LMW)Type X(G collagen)Major component of all basal laminaeMinor components of most tissuesexcept cartilage; fiber forms unknownFirst identified in aortic intima, but nowthought to have a broad, but yetunidentified, tissue distributionIdentified in amniotic membrane andskin; believed to be associated with allstratified epithelia; may be anchoringfibril proteinIdentified as product of a variety of celltypesCartilaginous tissuesProduct of cartilage hypertrophic cells17A. The interstitial collagens - Collagens that form broad-banded fibersFound in most tissues except cartilage;major component of bone, tendon, skin,and dentinSome fetal tissues; product of certainmalignant and normal cell linesCartilaginous tissuesCartilaginous tissuesFound together with type I; in relativelyhigh concentrations in extensible tissuessuch as blood vessels, skin and gutType I^{ al(1)}2a2(1)Type I 'trimer'^{al(I)}3Type IIM { al(IIM)}3Type IIm^{ al (11m)}3Type III {cciallthB. The minor collagens - Less abundant; do not form broad-banded fibersC. Minor cartilage collagensla2a3awith unknown subunit compositionsCompositionally similar to type V;structures, fiber forms, and functionsunknownCompositionally similar to type II.Table 1.1:^Genetically distinct collagens (Burgeson and Morris 1987)18The types of collagens that form large fiber groups are 300 nm long having periodic crossstriations of 67 nm which are evident when studied by electron microscopy. Newly secretedcollagen molecules aggregate to form collagen fibrils, which will combine to form a largercollagen fiber. Collagen fibers are the most common fiber found in connective tissue and areresponsible for the rigidity and strength of the connective tissue. The major types of these bandedfibers are collagen types I, II, and III, and these collagen types are refered to as the interstitialcollagens.Type I collagen is the most abundant and widely distributed form of collagen found throughoutthe connective tissues of the body. It is located in bone, tendon, skin, tooth structures, periodontalligament and gingiva (Bornstein and Sage 1980). Early work on collagen was performed onneutral-salt soluble and acid-soluble collagen derived from tissues composed entirely of type Icollagen (rat tail or tendon), or tissues in which this form is predominate. There is also an atypicalform of type I collagen referred to as type I trimer. Its designation is { al (I)}3. Type I trimer wasobserved in a polyoma virus-induced mouse tumor (Moro and Smith 1977), cirrhotic liver(Rojkind et al. 1979), human skin (Uitto 1979) and diseased human gingiva (Narayanan et al.1980; Narayanan et al. 1985). Whether this represents a separate gene product or is therecombination of three al (I) chains is still unknown.Type II collagen is a homotrimer that has been identified to exist in two forms. Types IIM andIIm are used to designate major and minor forms that have been separated based upon amino acidsequencing (Butler et al. 1977). The chain composition is designated foci (1))3. Type II collagenis the main constituent of cartilage, vitreous humour and intervertebral discs.Type Ill collagen is the second most abundant collagen type after type I. Except for tendon andbone, it is found in the same connective tissues as type I . It, like type II, is a homotrimerdesignated as al (Ill) }3. While the function of type DI collagen has not been demonstrated, thereappears to be a correlation between the proportion of type III to type I collagen and the extensibilityof the connective tissue (Burgeson and Morris 1987). Relatively high concentrations of type III19have been isolated from skin, blood vessels and placental tissues (Miller 1976), as well as gingivaand periodontal ligament.Collagens that do not form broad banded fiber systems are types N through X. Although theirmain distinguishing feature is their inability to form broad-banded collagen fibers, they like allcollagens contain both helical and non-helical structural domains.Type IV collagen is referred to as the basement membrane collagen. It contains two geneticallydistinct a chains, and is designated as fa 1 (IV) 120c2(IV). This form of collagen is highlyglycosylated and lacks the characteristic 67nm fibrillar structure of interstitial collagens (Kefalideset al. 1979). Studies using monospecific antibodies revealed that type N collagen molecules werelocalized to the lamina densa of the basal laminae. It is believed that this collagen in the basallamina is involved in membrane ultrafiltration, and in the anchoring of neighbouring cells duringthe development and maintenance of tissues (Martinez-Hernandez and Amenta 1983).Although type N collagen contains both helical and non-helical domains, the size andorientation differ from that observed in the interstitial collagens. The main triple-helical domain intype N collagen is longer than that found in the interstitial collagens (Kuhn et al. 1981). Inaddition to containing cysteine residues which form disulfide bonds within its triple helix, type Ncollagen contains several regions with a disrupted Gly-X-Y sequence, creating susceptible regionsto enzyme attack (Bornstein and Sage 1980). The amino and carboxy terminal domains of type IVcollagen are also different. The majority of the amino terminal end is triple helical, and this minortriple helical region is separated from the major central triple helix by a small globular domaincalled NC-2 (Timpl et al. 1979a). The designation NC indicates that this globular portion is notsusceptible to bacterial collagenase attack. In addition, this minor triple helical portion is the sitewhere type IV collagen molecules combine and later become part of the basement membranecomplex (Timpl et al. 1981a). The carboxy terminal globular domain is similar in size to thatfound for the interstitial collagens.Studies pertaining to the synthesis of basement membrane reveal that type IV collagen isincorporated directly into the matrix as a procollagen entity (Heathcote et al. 1978). Structural20studies of polymeric type IV collagen using electron microscopy suggest that four molecules forma tetramer joined by disulfide and other covalent linkages (Kuhn et at 1981).Type V collagen is composed of a group of molecules having genetically distinct chains andsimilar structure (Sage and Bornstein 1979). Presently three a chains with different chainorganization have been identified. Studies employing monoclonal antibodies suggest that type Vcollagen is present throughout connective tissue stroma (Linsenmayer et al. 1983), although itsfunction is not well understood. Other studies indicate that type V collagen may be a basementmembrane component (Stenn et al. 1979), and since it has been detected on cell surfaces it has alsobeen considered as a cytoskeleton collagen (Gay et al. 1981).Collagens are found throughout the hard and soft tissue components of the periodontium andcontribute significantly to its function. For example, in the gingiva, collagens comprise themajority of the protein, with more than 60% of the total tissue protein being collagen. For variouscomponents of the periodontium, the distribution of the collagen types are shown in Table 1.2STRUCTURE^COLLAGEN TYPES^PERCENT OF TOTALAlveolar bone I^ 100Periodontal ligament^I, in, v*^84, 16-18, 1Cementum^I, III 95, <5Healthy gingiva I, III, IV, V^91, 9, <1, <1Edentulous ridge mucosa^I, III, IV 85, 14, <1Table 1.2:^Distribution of collagen types in periodontal structures (Schroeder and Page 1990).'*' denotes the the proportion is not known.211.4.3 Intracellular and extracellular metabolismCollagen is synthesized by fibroblasts which secrete a precursor form, procollagen, into theextracellular space. Once procollagen is in the extracellular compartment, it is converted intocollagen following cleavage of the N- and C-terminal peptide domains. Each collagen molecule isapproximately 300,000 daltons, comprised of 3 unbranched polypeptide chains of approximately1000 amino acids in length. Each polypeptide a chain is coiled into a left handed helix, and allthree a chains are assembled in the collagen molecule to form a right handed helix.Collagen biosynthesis can be separated into several stages. First the DNA is transcribed intocorresponding mRNA's, which are translated into various alpha chains. Intracellular processingvia post-translational modifications are necessary for the formation of triple helical procollagenmolecules. Extracellularly, secreted procollagens are processed into molecules which, forexample, are crosslinked into fibers in types I and III collagen, or form molecular networks as intype IV collagen. The sequence of collagen synthesis is outlined in Figure 1.4.Transcription of DNA into mRNA for the pro-alpha chains is effected in the usual manner asfor any other protein. First, DNA is transcribed into pre-mRNA which contains exons andintrons. These are further modified by excision and splicing to form cytoplasmic mRNA. Thesesteps give rise to several types of mRNA's, which code for different forms of collagen.Furthermore, different sizes of mRNA have been identified for the pro-a2(I) chain (Myers et al.1981), and the pro-al (I) chains (Chu et al. 1982), for human type I procollagen.Post-translational modifications of collagen can be separated into intracellular and extracellularevents. Intracellularly, modifications such as hydroxylation and glycosylation of the polypeptidechains result in the formation of procollagen molecules. Once procollagen is secreted to theextracellular space, further modifications include cleavage of extension peptides and cross linkingof collagen molecules.Secretion to extracellular spaceV Cleavage of extension peptidesCollagen^V^I^Mature Collagen IE2nITE@CAOUllaffFigure 1.4:^Schematic diagram showing the major steps in collagen biosynthesis.'pro'= proline; 'lys'= lysine; 'hylys'= hydroxylysine.I ^IAssembly into microfibrilsand fibrilsAggregation of fibrils to formcollagen fibers22OrtH[rm©@Nhollmo .DNA Jr TranscriptionmRNATranslationHydroxylation of pro and lysGlycosylation of hylysTriple helix formationDisulphide linkingProcollagenSynthesis of pro-alpha chains23The synthesis of the polypeptide chains occurs on membrane-bound ribosomes, and they passthrough the membrane to the cistemae of the rough endoplasmic reticulum (Prockop et al. 1976).The pro-a chains are synthesized in a precursor form having an additional hydrophobic N-terminalpre-peptide sequence similar to sequences found for other secretory proteins. However, the pre-protein sequences of procollagen may be significantly longer than for other proteins. The pre-peptide in the newly synthesized pre-pro-al (I) chain contains more than 100 amino acid residues(Palmiter et al. 1979; Sandell and Veis 1980). It is believed that the function of the pre-peptide isto bind the rough endoplasmic reticulum during polypeptide synthesis and perhaps play a roletransporting the protein across the membrane (Davis and Tai 1980).The pre-peptides of collagen are cleaved within the rough endoplasmic reticulum membraneduring translation. One enzyme has been isolated and characterized from dog pancreas membranes(Strauss et al. 1979). This enzyme is inhibited by high concentrations of chymostatin and by someserine proteases. The specificity of this enzyme is uncertain as the amino acid sequence at theprotein cleavage site can vary. However, incorporation of leucine and threonine analoguesspecifically inhibits the cleavage of the pre-protein segment (Horan and Boime 1981). This wouldindicate that there is some level of control of the protease involved in the propeptide processing.The next major intracellular modification is the hydroxylation of both proline and lysineresidues. The hydroxylation of prolyl and lysyl residues are catalyzed by three distinct enzymes:prolyl 4-hydroxylase, prolyl 3-hydroxylase and lysyl hydroxylase. All three enzymes specificallyrequire that the amino acid residue must be present in a peptide linkage (Kivirikko and Myllyla1980), and that Fell-, 2-ketoglutarate, 02 and ascorbate be present as co-factors. The resulting 4-Hyp, 3-Hyp and hydroxylysine (hylys) are found almost entirely in collagen with smaller amountsfound in a few other proteins with collagen-like amino acid sequences (Adams and Frank 1980).Further studies on the requirement for peptide substrates for enzymatic hydroxylationdemonstrates a spatial requirement for hydroxylation of Pro and lysine residues. Studies on anumber of peptides that the minimum sequence requirement for interaction with both prolyl 4-hydroxylase was attained with a X-Pro-Gly triplet, and for interaction, lysyl hydroxylase a X-Lys24-Gly triplet (Kivirikko and Myllyla 1980). In agreement with these sequence requirements both 4-Hyp and Hylys have been identified in collagen and collagen amino acid sequences. The onlyexceptions are for the sequence of X-Hyp-Ala, in the first subcomponent of the first complementprotein (Clq), and one sequence of X-Hyl-Ser or X-Hyl-Ala in some short non-helical endportions of collagen a chains (Gallop and Paz 1975).The 3-Hyp has been identified in collagen only in the sequence Gly-3Hyp-4Hyp-Gly (Gryderet al. 1975). The reaction with peptidyl proline 3-hydroxylase appears to require a Pro-4Hyp-Glysequence rather that Pro-Pro-Gly sequence which is probably not hydroxylated (Tryggvason et al.1977). The prolyl 4-hydroxylase and lysyl hydroxylase is enhanced by the amino acid in the Xposition of the X-Y-Gly triplet being hydroxylated. Additionally, interaction with all threeenzymes is influenced by amino acids in other parts of the peptide, peptide conformation and chainlength (Prockop et al. 1976; Risteli and Kivirikko 1976; Risteli et al. 1976; Kivirikko and Myllyla1980). The maximal velocity of the hydroxylation reaction is influenced by the structure of theamino acid in the 'X' position of the same triplet. The effect of increasing chain length is todecrease the reaction constant, Km, of the peptide (Prockop et al. 1976; Kivirikko and Myllyla1980). And finally, triple helical peptides do not allow for hydroxylation reactions to occur, whichindicates that hydroxylation occurs before triple helical formation of pro a chains. However,within the non triple-helical domains of collagen pro a chains, the lysine before glycine order doesnot hold. Furthermore, Hylys that form in these regions are important since they are laterconverted to aldehydes which subsequently participate in cross linking collagen extracellularly(Olsen 1981).Another intracellular event in the biosynthesis of collagen is the glycosylation of Hylys andasparagine residues in the propeptides. Two specific enzymes, galactosylhydroxylysyl transferaseand glycosylgalactosylhydroxylysyl transferase, catalyze the 0-glycosylation of hylys in thecollagen domain of procollagen (Kivirikko and Myllyla 1979). The first enzyme, peptidylhydroxylysine galactosyltransferase, couples a galactose sugar to certain Hylys's, while the25second, peptidyl galactosylhydroxylysine glucosyltransferase, transfers glucose to certaingalactosylhydroxylysines.Glycosylation takes place mainly while the pro a chains are still being assembled on theribosomes, as there is no significant time lag between the hydroxylation of lysine and theglycosylation of Hylys (Kivirikko and Myllyla 1984). As in hydroxylation, glycosylationcontinues until the formation of procollagen triple helix prevents further reaction. Thus, the degreeof glycosylation will depend upon the rate of triple helix formation. The extent of glycosylation isalso different amongst the genetically distinct collagen types. This can be explained by regulationof the amounts of the single types of transferase enzymes.Glycosylation is also influenced by several factors including peptide amino acid sequence,peptide chain length, and peptide conformation (Kivirikko and Myllyla 1979). The number of -X-Hyl-Gly- triplets in a polypeptide chain is important in the overall glycosylation, whereas theamino acid sequence in the region around Hyl residues may be unimportant (Anttinen and Hulkko1980). The effect of peptide length is that the longer the chain the better it acts as a substrate.Triple helix formation prevents glycosylation, most likely due to steric hindrance with bothtransferases.The carbohydrate donor for both peptidyl hydroxylysine glycosyltransferase enzymes is thecorresponding uridine diphosphate (UDP) glycoside. Either UDP-galactose or UDP-glucosesupplies the sugar moiety for the peptidyl Hylys residue. The reaction requires a bivalent cation,best served by manganese (Kivirikko and Myllyla 1979), although in vitro Fe 2+ and Ca2+ haveserved as the cation cofactors (Myllyla et al. 1979).The function of 0-glycosylations in collagen is unknown. There have been suggestions thatthey play a role in the packing of collagen into fibrils. Intermolecular cross links are formed fromglycosylated Hylys residues, and these units may have a role in the attachment of some cells totype IV collagen (Berman et al. 1980). However, glycosylation was found not to be a requirementfor secretion of procollagen from cells, as secretion was not affected in patients with type VI26Ehlers-Danlos syndrome characterized by a deficiency of glycosylated and non-glycosylated Hylysresidues (Quinn and Krane 1976).Glycosylation of asparagine residues in collagen is affected by transfer of a mannose-richoligosaccharide side chain from a dolichol phosphate intermediate. The acceptor site for theoligosaccharide side chain contains the triplet Asn-X-Thr-, in agreement with structuralrequirements for N-glycosyltransferases (Pesciotta et al. 1981). In type I procollagen most or allof the carbohydrate moieties reside in the C-terminal propeptide, whereas for type II procollagen,these carbohydrates are present in both the N-terminal and C-terminal propeptides. It has beensuggested by some investigators that these aspargine-linked oligosaccharides are necessary tomaintain a normal rate of procollagen secretion from the cell (Housley et al. 1980), but this view isnot universally accepted (Kivirikko and Myllyla 1982).The formation of intramolecular and intermolecular propeptide disulfide bonds contributes tothe rate of formation and stability of triple helical procollagen. For the N-terminal peptides, intypes I and II procollagens, only intrachain disulfide bonds are formed (Prockop et al. 1979).However, interchain disulfide bonds are found in the C-terminal propeptides for both these typesof collagen. Type III procollagen has both N-terminal and C-terminal interchain disulfide bonds,and 2 additional interchain disulfide bonds at the C-terminal end of the triple helical domain. Bothtypes N and V procollagens consist of two or more polypeptide chains (Fessler et al. 198 la,b;Kumamoto 1981), which contain interchain disulfide bonds. The exception is for three pro al (V)chains which do not form disulfide linkages (Fessler et al. 1981a,b; Kumamoto 1981).As interchain disulfide bonds are found in the C-terminal region for types I, II and IIIprocollagens, the C-terminal propeptides are considered to contain the information for interchainrecognition and initial chain association (Rosenbloom et al. 1976). Supporting this concept is thatinterchain bonds between C-terminal pro-a chains are found when triple helix formation isprevented, whereas bonds between N-terminal pro-a chains in type BI procollagen are not formedunder the same situation. This indicates that disulfide bonds are formed earlier between C-terminal27propeptides than for N-terminal propeptides. Thus, triple helix formation from the three pro-achains is likely to begin at the C-terminal end of these propeptides (Fessler et al. 1981c).The hydroxyl groups of 4-Hyp are critical in that they stabilize the triple helix underphysiological conditions (Prockop et al. 1976). With no Hyp residues present, the triple helix ofcollagen melts at approximately 25 °C, whereas with about 100 Hyp residues per a chain, themelting point rises to 40°C (Berg and Prockop 1973; Rosenbloom et al. 1973). Thus the formationof stable triple helices of collagen requires that approximately 50% of the proline residuesincorporated into each pro-a chain become hydroxylated. Furthermore, it is known that non-helical pro-collagen is secreted by cells at one tenth the rate that normal pro-collagen is secreted(Kao et al. 1977; Kao et al. 1979).Extracellular processing of collagens involves cleavage of extension peptides and formation ofintermolecular cross links. The removal of the N- and C-terminal extension propeptides occurs viaspecific N- and C-terminal proteases. These are neutral endopeptidases and they require thepresence of a divalent cation such as Calf.. In type I collagen for example, removal of extensionpeptides occurs via specific proteases after the secretion of the entire propeptide (Bornstein andSage 1980). For type I procollagen, the cleavage site for the N-terminal protease in the pro-al(I)chain is through a proline-glutamine bond (HOrlein et al. 1979). Furthermore, it is known thatwithout removal of the propeptides, collagen fibril formation is inhibited (Prockop et al. 1979).Fibril formation occurs spontaneously for interstitial collagen molecules once propeptides havebeen removed. Although the rate to which this can occur in vivo is difficult to assess,macromolecules like fibronectin and proteoglycans are thought to affect the rate of fibril formation(Kleinman et al. 1981).Major collagen cross linking occurs via aldol condensation following oxidation of lysine andhylys residues by lysyl oxidase (Kuboki et al. 1981). The oxidative deamination of c-aminogroups in certain lysine and hylys residues yields reactive aldehyde groups termed allysine andhydroxyallysine respectively. These aldehydes can then condense with other lysine or hylysresidues or with allysines to form covalent intermolecular cross linkages.28There are two main mechanisms for intracellular and extracellular degradation of collagen.Intracellular degradation of fragments of collagen fibrils is believed to occur via lysosomalmetabolism. Extracellular degradation is also affected by enzymes, but is probably modulated bythe tissue, cell type, and the disease state.Newly synthesized collagen and collagen fibrils initially present extracellularly both canundergo intracellular collagen degradation. Degradation of newly synthesized collagen appears toserve both as a basal turnover rate phenomenon, as well as a mechanism for removing defectivecollagen. Intracellular degradation of extracellular collagen is believed to occur after disruption ofthe collagen fiber, with subsequent phagocytosis and digestion by the lysosomal enzymes.It has been shown that 10 to 20% of newly synthesized collagen is degraded intracellularly,while the remainder is secreted to the extracellular environment. This degradation based ondeterminations for radiolabelled Hyp has been demonstrated for cells synthesizing types I, Ill, andIV collagen (Bienkowski et al. 1978a,b; Palotie 1983). The observation that this intracellulardegradation is relatively consistent for a variety of collagen types, suggests that it may be auniversal phenomenon.In addition to the basal level of intracellular collagen degradation, the conformationallyabnormal newly synthesized collagen is degraded at an enhanced rate. The role of degradationappears to eliminate defective pro-a chains that are unable to form a stable helix. Diploid humanfibroblasts, which form pro-a chains that are deficient in 4-Hyp and do not form triple helicalmolecules, degrade approximately one third of their newly synthesized collagen (Berg et al. 1980).The same investigators also found that cis-4-Hyp, a proline analogue, prevents triple helixformation and significantly enhances the degradation of newly synthesized pro-a chains. Otherstudies demonstrated that the use of amino acid analogs not only rendered collagen moresusceptible to proteolysis (Uitto and Prockop 1974), but that the proteolysis occurred within ashort time after synthesis (Berg et al. 1984).Enhanced intracellular degradation was also found to occur in collagen that was thermallydenatured or altered by incorporation of amino acid analogues. Studies of incubated human skin29fibroblasts at 30°C and 42°C showed an increase in newly synthesized collagen degradation from12% to 49% due to increased temperature (Steinmann et al. 1981). Increased intracellular collagendegradation was also observed in systems which block the enzyme prolyl hydroxylase (Kao et al.1979), thus interfering with the conversion of Pro to Hyp. These studies support the concept thatcollagen conformation is important in the molecules' stability.In addition to proteolysis of non-helical or defective collagen, another control mechanism hasbeen demonstrated. Under-hydroxylated collagen is nonhelical at physiological temperature andpH, and it has been shown to be secreted at one tenth the rate of fully hydroxylated collagen (Kaoet al. 1979).Extracellular collagen degradation occurs via attack by enzymes in the matrix and to someextent by the actions of by-products of metabolism. The major enzyme responsible for collagendegradation is mammalian collagenase. Other enzymes, such as elastases, gelatinases, proteases,and bacterial collagenases also breakdown collagen. By-products of protein metabolism such asvolatile thiol compounds have also been shown to disrupt interstitial collagens (Johnson et al.1985, 1992).Collagen fibrils are stable at physiologic temperature and pH, and are resistant to cleavage byproteases within their helical domains. If denatured by heat which disrupts the triple helix forexample, the a chains can be cleaved by proteolytic enzymes. However, mammalian collagenases,which are metalloproteinases dependent upon Ca 2+, are able to cleave collagen at physiologicaltemperatures into 3/4 and 1/4 fragments. Once these fragments are formed they spontaneouslydenature and can then be further degraded by proteases. In contrast to mammalian collagenase 3/4and 1/4 cleavage products, the bacterial collagenase digests collagen to numerous small peptidesthat are readily dialyzable.Mammalian collagenase cleaves interstitial collagen in a specific helical region. In type Icollagen this cleavage occurs through the Gly-Ile bond in the al (I) chain (residues 775-776), andthrough the Gly-Leu bond in the a2(I) chain. In the a chains of type III collagen, the cleavage30also occurs through the Gly-Leu bond. This forms the N-terminal 3/4 and C-terminal 1/4fragments which spontaneously denature at 37°C.Type HI collagen differs from type I in that it can be degraded in its native form by enzymesother than mammalian collagenase. Type BEE collagen is susceptible to attack by other proteasessuch as elastase, trypsin, and thermolysin (Miller et al. 1976; Mainardi et al. 1980b), in the sameregion where collagenase acts. In addition, elastase is thought to further contribute to collagendegradation by cleaving the telopeptide region thus solubilizing the entire molecule (Starkey et al.1977).Mammalian collagenase is secreted in a latent form. Agents that convert the latent to an activeform include proteases such as trypsin and plasmin, and organomercurial compounds. Metalchelating agents such as EDTA produce total inhibition, while thiol agents like cysteine anddithiothreitol produce partial inhibition of collagenase activity. Two additional collagenaseinhibitors found in plasma are oc2-macroglobulin and Pi-anticollagenase (Woolley 1984).Type IV collagen is an integral part of basement membranes and its degradation may have aprofound effect on the integrity of this matrix. In studies of collagen degradation it was found thattype IV collagen was resistant to mammalian collagenase (Welgus et al. 1981). Severalinvestigators have identified and characterized a type IV specific degrading enzyme. It is also ametalloproteinase that is activated by trypsin or mercuric chloride (Salo et al. 1983). The cleavagesite was determined to be near the N-terminal portion of the molecule (Fessler et al. 1984).Other proteinases are also capable of degrading type IV collagen. Pepsin extracted type IVcollagen has been shown to be susceptible to a mast cell proteinase (Sage et al. 1979) and to ahuman neutrophil metallogelatinase (Murphy et al. 1980). Human leukocyte proteases were alsofound to degrade native type IV collagen (Mainardi et al. 1980a; Uitto et al. 1980). These resultssuggest that this lattice form of collagen is more labile than the fibrillar collagens of the interstitialgroup.Metabolism of collagen in the periodontium varies with each collagen type and the rate differsfrom other tissues of the body. Studies based on conversion of radiolabelled Pro into labelled Hyp31demonstrated a faster rate of synthesis in oral tissues than in skin. It was also determined that thelabel was lost at a faster rate in gingival tissues than other connective tissues of skin, tendon andbone, indicating a more rapid turnover in the gingiva.Additional studies on collagen metabolism in the periodontium have been performed usinganimal models (Sodek 1976; Sodek 1977; Sodek et al. 1977; Sodek 1978; Sodek and Limeback1979). Sodek has demonstrated that although collagen turnover is higher in gingiva than in boneor skin, it is lower than in periodontal ligament (Sodek 1977; Sodek 1978). These differentturnover rates are reflected in different half lives for collagen. For example, the half life for mid-root peridontal fibers is 7.5 days while for dentogingival fibers is 25 days. These studies haveshown that collagen metabolism in the oral cavity is complex.Destruction of periodontal connective tissue results from increased activity of tissue proteolyticenzymes, including matrix metalloprotease and plasminogen/plasmin cascades (Birkedal-Hansen1988). Since mammalian collagenases are the only tissue proteases capable of degrading nativecollagen, they can initiate degradation of collagen.Studies of collagenase activity in relation to severity of gingival inflammation havedemonstrated an increased enzyme activity in inflamed human gingiva and crevicular fluid ascompared to healthy tissue and fluid (Uitto and Raeste 1978; Overall et al. 1987). In addition,moderately inflamed tissue was found to release more latent collagenase than did severely inflamedgingiva. This latent form of the enzyme can be activated by trypsin or bacterial plaque (Uitto andRaeste 1978).The sources of the collagenases and mechanisms of their activation were subsequentlyelucidated. The role of gingival, GCF and salivary interstitial collagenases in human periodontaldiseases was investigated by Sorsa and co-workers (Sorsa et al. 1990). It was found that all threecollagenases had Mr of 70K and existed predominantly in a latent form that could be activated byaminophenylmercuric acetate, gold thioglucose and hypochlorous acid. While several serineproteases also activated gingival and salivary collagenases, plasmin and plasma kallikrein wereineffective. The collagenases from all three sources degraded types I and II collagens at32approximately equal rates which were considerably faster than degradation of type III collagen.These findings indicated that the collagenases found in inflamed gingiva, GCF and saliva resemblethe enzyme derived from neutrophils.In contrast, an investigation of the collagenase activity in crevicular fluid of patients withjuvenile periodontitis (JP) indicated that these samples contain an enzyme from a different source(Suomalainen et al. 1990). Gingival crevicular fluid from untreated JP pockets containedmammalian collagenase which cleaved type II at a markedly slower rate than types I and IIIcollagens. This substrate specificity is characteristic of collagenases produced by fibroblasts,epithelial cells and macrophages. Furthermore, they found that A. actinomycetemcomitansisolated from subgingival plaque samples of these patients was able to release collagenase fromPMN's in vitro. They postulated that due to a lack of normally functioning PMN's in theperiodontium of JP patients, the bacteria activate resident periodontal cells to produce increasedamounts of collagenase.Regardless of the source of mammalian collagenases, they are important initiators of collagendegradation. These are the only known enzymes that degrade the native interstitial collagens viacleavage within their helical domain. Once collagen is cleaved into 3/4 tissue collagenae A (TCA)and 1/4 tissue collagenase B (TCB) fragments, it is subject to denaturation at body temperature andfurther degradation by gelatinases and proteases. Proteinases such as telopeptidase, elastase andcathepsin G can cleave the cross link regions and eliminate TCA triple helical complexes, thusexposing the chains to further proteolytic degradation. Since bacterial collagenase can completelydegrade collagen to small peptides, the demonstration of TCA collagen cleavage fragments inmoderate to severely inflamed gingival tissues (Overall et al. 1987), suggests a dominant role formammalian collagenases in collagen destruction in inflamed gingiva.1.4.4 Non-collagenous componentsIn addition to the major connective tissue protein collagen, other macromolecules, namelyglycoproteins, glycosaminoglycans, proteoglycans and elastin account for most of the remaining33extracellular matrix (ECM) content in periodontal connective tissue. While extensive research hasfocused on collagen of gingiva and periodontal ligament, comparatively much less attention hasbeen devoted to these other proteins.Most extracellular proteins are glycosylated, and therefore, by definition are glycoproteins.Four types of carbohydrate linkages to protein have been distinguished. They are theproteoglycans which contain GAG's linked 0-glycosidically via xylose to either serine orthreonine, the collagens that have glucose and galactose 0-glycosidically linked to Hylys, themucins which contain oligosaccharide 0-glycosidic ally linked to serine or threonine and a furthergroup of glycoproteins that have oligosaccharide with a mannose core N-glycosidically linked toasparagine. Possible functions of these carbohydrate linkages include modification of physicalproperties, alterations to the susceptability of enzymatic degradation, regulation of metabolicbehavior of the protein and organization of proteins within cells, in membranes and in theextracellular compartment (Hakomori et al. 1984).Fibronectin (FN) is a major glycoprotein which is found in the extracellular spaces, cellsurface, basement membranes and body fluids. It is synthesized by fibroblasts, endothelial cellsand monocytes. It is a dimer consisting of polypeptide chains of approximately 2000 amino acidresidues, which are linked together via disulfide bonds to form a molecule of approximately450,000 daltons. FN has been implicated in a number of functions which have been observed invitro. Various observed functions have been related to plasma, extracellular matrix organization,homeostasis and wound healing.Fibronectin promotes cell attachment and spreading which are mediated by interaction of FNwith cell surface receptors. A specific region of FN, of about 150,000 daltons, has been identifiedas the specific domain required for these activities (Sekiguchi and Hakomori 1980).Fibronectin is also involved in organization of extracellular matrix and basement membranes.Its role in matrix organization is not well understood but is conjectured to be via binding of FN tocollagen (Kleinman et al. 1976; Pearistein 1976) and to proteoglycans (Yamada et al. 1980;Oldberg and Ruoslahti 1982). Covalent cross linkage of FN to collagen, fibrin and other34macromolecules occurs via the enzyme transglutaminase, which is thought to react with aglutamine residue on FN's N-terminal region (Mosher et al. 1980).Fibronectin also plays a role in homeostasis and wound healing by participating in severaldifferent processes. These include homeostasis at the site of injury, clearance of cell debris andblood clots, formation of a temporary scaffolding and recruitment of granulation tissue.FN is involved in homeostasis via interactions with several components. In early studies themechanism was thought to occur via the interaction of FN with fibrin (Ruoslahti and Vaheri 1975),and crosslinkage of FN by transglutaminase. Further studies showed that FN is covalently crosslinked to fibrin via factor VIlla (Mosher 1975), which in turn mediates cell adhesion to fibrin(Grinnell et al. 1980). Activated platelets have been shown to express surface FN, whereasnonactivated platelets do not (Ginsberg et al. 1980). In addition, FN enhances the spreading ofplatelets on collagen surfaces (Grinnell et al. 1979). Furthermore, due to the above interactions, ithas been suggested that FN plays a role in platelet adhesion to the subendothelium of injured bloodvessels (Bensusan et al. 1978).Clearance of cellular debris and blood clots from wound areas by macrophages is believed tobe mediated by FN acting as a nonspecific opsonin (Zinkevich et al. 1982). Both monocytes andpolymorphonuclear leukocytes have FN receptors on their cells surface (Bevilacqua et al. 1981;Pommier et al. 1984). Thus, via interaction with these receptors, FN may play a role in debrisclearance by mononuclear phagocytosis.The ability of FN to interact with a component of the first complement protein, Clq (Pearlsteinet al. 1982), may be important in promoting wound healing and homeostasis. Since Clqstructurally possesses a helical domain, aggregation of this molecule with FN may function as ascaffolding matrix. It has been recently shown that aggregated forms of Clq bind to the fibroblastClq receptor (Bordin et al. 1990), and thus may play a role in fibroblast cell adhesion. Potentiallythese interactions of FN, Clq and fibroblasts may be important in directing the wound healingmatrix. In addition, because biologically Clq has immunological activity, aggregation of thiscomplement protein can participate in immune-mediated opsonization.35Studies on experimental granulation tissue have implicated FN for additional roles in woundhealing. Not only was the concentration of FN increased after wounding, but also the factor VIIIa-mediated cross linked FN to fibrin served as a better substrate for fibroblast attachment andspreading than did fibrin alone (Grinnell et al. 1980). In addition, FN was found to act as achemoattractant for both endothelial cells (Bowersox and Sorgente 1982) and fibroblasts(Postlethwaite et al. 1981). Furthermore, FN in granulation tissue was found to be primarilyassociated with fine fibers of type III collagen, while later on in wound healing when there is apredominance of thicker, mature type I collagen fibers, FN is no longer present (Kurkinen et al.1980). This correlated with the observation that FN has the strongest association with type IIIcollagen, when FN is experimentally exposed to all collagen types in their native state (Engvall etal. 1978).Laminin is a glycoprotein which consists of three subunits of 200,000 - 220,000 daltons andone large subunit of 400,000 - 440,000 daltons, and is supported by interchain disulfide bonds(Timpl et al. 1979b). The small and large subunits are referred to as a and b subunits, and they arebiologically and immunologically distinct (Rao et al. 1982). This large glycoprotein is animportant component found within all basement membranes, where it serves as a major attachmentglycoprotein for epithelial cells (Terranova et al. 1980). In addition to having domains forepithelial cells, laminin is able to bind both type I collagen and proteoglycans on the ends of its aand b subunits (Yamada 1983). Furthermore, laminin was found to have the capacity for adhesionto fibroblasts, with a similar affinity to that of FN (Couchman et al. 1983). Through theinteractions with epithelial cells, proteoglycans, fibroblasts and collagen, laminin acts as anextracellular glue important for maintaining the integrity of basement membranes.Glycosaminoglycans (GAG's), which are long unbranched polysaccharide chains composed ofrepeating disaccharide units, are another important class of macromolecules found in theperiodontium. Formerly called mucopolysaccharides, the GAG's characteristically have one oftwo sugars in the repeating disaccharide always an amino sugar, either N-acetyl glucosamine or N-36galactosamine. Due to the presence of sulfate or carboxy groups on both or many of the sugarresidues, GAG's are very negatively charged.Glycosaminoglycan chains are synthesized by sequential addition of sugars from sugarnucleotide precursors. For example, UDP-glucuronic acid and UDP-N-acetylgalactosamine areprecursors for chondroitin sulfate (Silbert 1964). For GAG's acting as side chains inproteoglycans, the protein core is synthesized first and then sites for monosaccharide transferasesinvolved in GAG synthesis are presented (Roden 1980).Proteoglycans comprise a large group of macromolecules that are ubiquitous in extracellularconnective tissue (Lindahl and HO& 1978). Each is composed of a central protein core to whichmultiple GAG chains are attached via oligosaccharide linkages. The oligosaccharide linkage formost GAG's to this protein core is a trisaccharide consisting of galactosyl-galactosyl-xylose(Lindahl and Roden 1972). The xylose is glycosidically linked to the hydroxyl moiety of serine inthe protein chain. Each proteoglycan molecule is composed of 2 or more different types ofGAG's. The total number of GAG's can vary from one or two to more than a hundred chains perproteoglycan molecule (Heinegard and Paulson 1984). This forms the basic proteoglycan unit.These units can then be covalently bound via a link protein to a central core of hyaluronic acid(Hardingham and Muir 1972), a linear polymer of alternating glucuronic acid and N-acetylglycosamine units.The function of proteoglycans and GAG's has mainly been derived from studies on cartilage(Comper and Laurent 1978). Proteoglycans are able to retain water and thus can produce anincrease in tissue pressure. In cartilage, an increase tissue pressure is manifested in compressivestiffness, which has been shown experimentally to be related to the presence of the GAGcomponents of proteoglycans (Kempson 1980). Studies by Comper and Laurent have indicatedthat proteoglycans limit diffusion of macromolecules through connective tissue (Comper andLaurent 1978). It is thought that due to their anionic and hydrated form, proteoglycans areresponsible for maintaining tissue and cell functions via interaction with matrix and cell surfacecomponents.37The breakdown of matrix structure of GAG's and proteoglycans appear to be initiated byglycosidases and proteases (Fluharty 1982). It is believed that after enzymatic cleavage anddisaggregation of matrix components, cells take up the fragments via pinocytosis and furtherdegradation occurs via lysosomes .Proteoglycans present in the periodontium are primarily secreted by the resident fibroblasts.The major proteoglycans located in the gingiva are chondroitin sulfate and dermatan sulfate(Shibutani et al. 1989). Dermatan sulfate, the most prominent GAG in gingival connective tissueis found uniformly distributed in the periodontium. Although chondroitin 4-sulfate also has a widedistribution, it is present in relatively much smaller amounts. Furthermore, chondroitin 6-sulfate isfound only perivascularly.Chondroitin sulfate is the predominant GAG in alveolar bone. Chondroitin 4-sulfate accountsfor more than 90% of the total GAG's extracted, while smaller amounts of hyaluronic acid,dermatan sulfate, and heparin sulfate were also isolated (Waddington et al. 1989). In addition,proteoglycans in bone were found to be smaller in size than those present in the gingiva, butsimilar in size to those present in hard tissues of cementum and dentin (Bartold 1990).Elastin, another major fibrous extracellular protein besides collagen, is a prominent componentin connective tissues where resiliency is important. High concentrations are found in the aorta,large arteries, lung and skin. Elastin is composed of two distinct components: an amorphouselastin component which comprises 90% of the mature fiber, and a microfibrillar component ofsmall fibrils 10-12 tun in diameter.Tropoelastin, the precursor of elastin, is secreted by the cell as a 72,000 dalton polypeptide.These molecules are cross linked via the aldehyde groups of lysine formed by enzyme peptidyllysine oxidase (Siegel 1980), which is the same enzyme involved in collagen cross linking Theresulting insoluble polymer which is highly cross linked serves as the elastomer. Although thesame enzyme is involved, cross linking in elastin differs from that in collagen. Since there is nohylys and little histidine in elastin, they are not involved in the process. The final result of the38condensation reactions of four lysyl residues, is the formation of desmosine and isodesmosine,two cross linkage components unique to elastin.1.5 VOLATILE SULPHUR COMPOUNDS1.5.1 SubstratesVolatile suphur compounds (VSC) are a group of substances that are associated with andcontribute significantly to halitosis. These compounds are emitted in breath and originate from oraland non-oral sources, such as nasopharyngeal, respiratory and gastro-intestinal systems (Sulser etal. 1939).Most investigations on the mechanism of production of oral malodour were performed onputrescent saliva systems. Whole saliva contains a variety of substances including bacteria,desquamated epithelial cells, lysed leukocytes, food debris and blood. From this complex mixturean objectionable odour is generated within one hour of incubation (Massler et al. 1951). Thereduction in intensity of putrefactive odour following thorough brushing of teeth and associatedoral tissues demonstrates that the oral cavity is the principle source of breath malodour emitted viathe mouth (Sulser et al. 1939).Oral malodour is intensified in individuals whose saliva exhibits increased putrefactive activity(Tonzetich and Richter 1964). Investigations by Tonzetich and co-workers on head space analysisof putrescent systems, provided the initial insights on VSC substrates (Tonzetich and Kestenbaum1960; Tonzetich and Richter 1964). Their studies showed that cellular elements are the mainsources of substrate for the production of VSC. Chemical and organoleptic evaluations attributedthe putrescence of putrefied saliva principally to acidic compounds. The VSC are produced viaputrefaction by microorganisms of proteinaceous substrates, such as desquamated epithelial cells,blood cells and food debris (McNamara et al. 1972). In the process, the cells are disintegrated, thedisulphide substrate is liberated and reduced to thiols; these serve as immediate precursors of VSC(Tonzetich and Johnson 1977). It is important to note that the addition of sulphur-containing39amino acids intensified the putrescence of incubated saliva. While substrates such as cysteine thatpossess free thiol groups, were the most immediate sources of malodour, disulphide substrates likecystine produced odour of similar intensity but at a slower rate.Oral malodour has also been correlated with the severity of periodontitis, and was shown todecrease following periodontal therapy (Tonzetich and Spouge 1979). The initial elevatedmalodour in periodontitis subjects is attributed to several factors. In periodontitis, increasednumbers of desquamated epithelial cells and adherent bacteria are found in saliva (Dreizen et al.1956). Furthermore, haemorrhage and increased crevicular fluid flow from affected periodontalsites provide increased sources of putrescible substrates (Cimasoni 1983).1.5.2 BacteriaSulphur containing volatile compounds are formed through degradation of sulphur-containingsubstrates by oral microflora (Tonzetich and Kestenbaum 1960; Tonzetich and Carpenter 1971).At least five genre of anaerobic gram-negative oral microorganisms commonly found in the deepestregion of periodontal pockets are capable of forming H2S. Spirochaetes, fusiforms, vibrios,veillonellae and bacteroides species produce H2S (Schwabacke and Lucas 1947; Breed et al. 1957;Omata and Braunberg 1960; Sawyer et al. 1962; Rogosa and Bishop 1964). Porphyromonasasaccharolyticus which produces profuse amounts of VSC (Tonzetich and McBride 1981),especially CH3SH and (CH3S)2, is found in high numbers in the gingival crevice (Syed et al.1981). This organism has been implicated as a primary pathogen in the etiology of periodontaldisease.1.5.3 Analysis of volatile compounds produced in the oral cavityIndividual malodorous components of breath and saliva have been identified using gaschromatographic (GC) and/or mass spectrometric analyses (Tonzetich and Richter 1964; Larsson1965; Kostelc et al. 1980; Kostelc et al. 1981). Tonzetich introduced the GC method employing ahighly sensitive and selective flame photometric detector (FPD) coupled with a GC unit (Brody and40Chaney 1966) for direct measurement of subnanogram quantities of VSC present in mouth airsamples (Tonzetich 1971).The direct GC analysis identified three volatile sulphur compounds in mouth air of allindividuals (Tonzetich 1971). Hydrogen sulphide (H2S) and methyl mercaptan (CH3SH) wereestablished as the primary volatile components which accounted for approximately 90 per cent ofthe sulphur content of mouth air. The remaining minor component was identified as dimethylsulphide, (CH3)2S. Concentrated air samples also demonstrated the presence of dimethyldisulphide, (CH3S)2, with increased prevalence and concentration in periodontal cases (MacKayand Hussein 1978). Kaizu and co-workers provide convincing evidence that except forobjectionable breath emitted by periodontally diseased individuals, the malodorous concentration ofCH3SH emanating from the oral cavity can be controlled by tongue scraping (Kaizu et al. 1978).Other investigators have also shown that VSC produced on the tongue surface in periodontalpatients, contributes sixty percent of the total sulphide content of mouth air (Yaegaki and Sanada1991). Reduction of objectionable breath from these patients was further accomplished byperiodontal treatment which removed a significant portion of the VSC content from their mouthair.Rizzo reported a positive correlation between the amount of H2S produced in gingival creviceand the depth of corresponding periodontal pockets (Rizzo 1967). Similarly, mouth air studies inour laboratory have indicated that levels of VSC are considerably increased in the mouths ofpersons with periodontal disease. In a study of twenty periodontal subjects, the concentration ofH2S and CH3SH correlated with the incidence and depth of periodontal pockets greater than 3 mm.Furthermore, the levels of both compounds were markedly reduced following curettage andcorrective periodontal surgery (Tonzetich and Spouge 1979). Similar studies of salivary volatilesindicate that other volatile products of salivary putrefaction such as pyridine and its 2-, 3-, and 4-methyl analogues are present in elevated amounts in saliva collected from individuals withmoderate to severe periodontitis (Kostelc et al. 1980).411.5.4 Implications in periodontal diseaseIt has been previously reported that volatile thiol compounds, H2S and CH3SH producedthrough putrefactive activity in the gingival sulcus, have the capacity to increase the permeability ofnon-keratinized oral mucosa. Using a two-chamber diffusion apparatus, both ions, (35S)-SO4,and small molecules, (3H)-PGE2, were found to penetrate more readily across mercaptan exposedthan untreated mucosa. This penetration followed a gradual diffusion kinetics until a steady statewas reached (Ng and Tonzetich 1984). Using the same diffusion apparatus, it was found that thediffusion of 14C- labelled LPS from E. coli followed a different pattern. Initially a component ofthe (14C)- LPS was found to diffuse, which decreased with time, through both control and CH3SHtreated specimens. A marked increase in diffusion of ( 14C)-LPS was obtained with tissues withprior exposure to CH3SH (Tonzetich et al. 1987). These results indicated that CH3SH increasesthe permeability of mucosa to ions, small molecules, and LPS. Thus one mechanism wherebythiol agents may contribute to the etiology of periodontal disease is by disrupting the integrity ofthe mucosal barrier.Once volatile sulphur compounds penetrate the epithelial barrier, they can then react with theunderlying connective tissue and accompanying cells. Human gingival fibroblasts were shown tobe adversely affected when exposed to H2S and CH3SH. Exposure of cells to lOng H2S orCH3SH/ml air/CO2 during 26 hours of activation and 4 hrs of pulsing resulted in a 44-47%reduction in total DNA synthesis (Johnson et al. 1992).Direct effect of VSC on collagen has been demonstrated in our laboratory. Exposure of type Irat-tail tendon collagen to H2S and CH3SH resulted in the conversion of some mature fibrillar(Tonzetich and Lo 1978) and acid-soluble forms (Johnson and Tonzetich 1985; Johnson et al.1985) to a neutral salt-soluble product. CM cellulose and SDS/PAGE/fluorography analysis ofboth neutral salt-soluble and acid-soluble products generated from acid-soluble collagen reactedwith radiolabelled H2S, yielded radioactivity associated with al, NJ, 01,2, and a2 chainscharacteristic of type I collagen. Furthermore, it was determined that thiol groups (-SH) reactedwith 2 or more active sites on the molecule (Tonzetich and Johnson 1986).42The demonstration of decreased DNA synthesis suggested that VSC inhibits a certainpopulation of cells during their cell cycle or perhaps that there is a reduced availability ofsubstrates. By investigating Pro transport of fibroblast cells, a 35% and 36% reduction in Protransport was reported in the presence of lOng of H2S or CH3SH/m1 air/CO2, respectively(Tonzetich et al. 1985). Subsequent staining of connective tissues with fluorescein diacetate andcounter staining with ethidium bromide demonstrated damage to cells exposed to CH3SH. Sincethe uptake of amino acids is a membrane-associated phenomenon, disruption of the integrity of cellmembranes by H2S and CH3SH would affect protein synthesis.Investigations of the effect of H2S- and CH3SH-treated human gingival fibroblast cultures onprotein and collagen metabolism have also been reported (Johnson 1983). Exposure to H2Sdecreased total protein synthesis by 18 to 20%, but did not alter the ratio of type I to type IIIprocollagen, 8.5/1. CH3SH-exposed fibroblast cultures showed a 35% decrease in total proteincontent. There was a reduction in type III collagen and almost a complete absence of type IIIprocollagen. Furthermore, type I collagen content was reduced, whereas type I procollagen and/ortype I protrimer increased approximately two-fold.Investigations on the effect of CH3SH on collagenolytic proteases and cAMP production wereperformed using human gingival fibroblast cultures. Following exposure to CH3SH, cathepsin Bactivity increased by 20% , while cAMP content of CH3S-exposed cultures was increased by 34%and by 58% in the same systems supplemented with IL-1 (Tonzetich et al. 1990).The effect of CH3SH on IL-1 secretion by leukocytes was studied using human tonsillarmononuclear cells. Exposure of these cells to 10 ng/ml of CH3SH was as effective as 5x10 -5 M 2-mercaptoethanol in augmenting a LPS augmentation of IL-1 production (Waterfield and Tonzetich1989). This implies that CH3SH may play a role in activation of the immune system.Recent investigations by Ratkay and Tonzetich, have shown that exposure of T lymphocytes toCH3SH resulted in a 30% increase in IL-1 production (Ratkay and Tonzetich 1992). In addition,exposure of fibroblasts to CH3SH increased collagenase production. Thus, it is possible that43CH3SH potentiates collagenase production either by direct interaction with fibroblasts, and/or viaan IL-1 mediated pathway.Recently Ouyang (Ouyang 1991) demonstrated that Zn ion can reverse the adverse metaboliceffects of CH3SH. It was found that 10 -4 to 0.5x10 -5 M Zn completely nullified the CH3SHadverse effects on protein content and proline transport of human fibroblast cell cultures.The described results suggest that VSC affects periodontal tissue directly as well as indirectlyby disrupting the tissue and normal fibroblast cell function. Thus, once H2S or CH3SH gainsentry into the tissue it can be involved in a number of vital biochemical reactions contributing topotentiation of periodontal disease.1.6 PERIODONTAL DISEASE1.6.1 Health versus disease stateThe periodontium consists of four tissues surrounding each tooth. The cementum, alveolarbone, periodontal ligament and gingival tissues function in several ways to maintain the integrity ofthe dentition. They secure the teeth within bone, resist forces of mastication, maintain a batherbetween the external and internal environments of the body, provide a source of cells to supportgrowth and repair, and defend against noxious agents.Periodontal disease is a process which destroys the tooth's supporting structures, theperiodontium. Although periodontal disease is clinically complex and heterogeneous, it iscategorized as an inflammatory process which, depending upon an individual's host response,results in loss of alveolar bone and loss of ligamentous attachment (Page and Schroeder 1981).The different classifications of inflammatory periodontal disease are broadly divided into twocategories. Gingivitis and periodontitis have been characterized as separate entities based uponhistopathologic, bacteriologic and immunologic evidence, which have provided the basis forunderstanding the disease process. Initially a reversible inflammation that is induced by dentalplaque which causes irritation in the gingival crevice, is considered gingivitis. Depending on44whether this source of irritation is removed, gingivitis can be present in either an acute or chronicform. Persistance of the inflammation can lead to deeper penetration into the underlying connectivetissues. Extension of the inflammatory process causes proliferation and apical migration of thejunctional epithelium, resulting in periodontal pocket formation. Loss of subjacent connectivetissue and underlying alveolar bone are indicative of the state of periodontitis.A number of different periodontitis lesions have been categorized. Children and adolescentscan be affected by generalized or localized juvenile periodontitis. The localized form ischaracterized by rapid loss of attachment and distinct bone loss around the mandibular incisors andaround the first molars. The generalized form also involves other teeth with the same distinct rapidbone loss. Adults can be affected by rapidly progressive or by adult periodontitis forms. Rapidlyprogressive periodontitis is seen most commonly in young adults in their twenties and earlythirties. As its name implies, it is a rapid form which is characterized by markedly inflamedgingiva and significant bone loss (Page 1983). The disease may have periods of remission inwhich the gingiva appears clinically normal, however the patients are now known to exhibit defectsin either neutrophil or monocyte chemotaxis. Adult periodontitis is the most common form ofperiodontitis. The age of onset is usually after 30 to 35 years of age. Unlike the other forms ofperiodontitis, there are no localized patterns of bone loss, no evidence of rapid progression and nosystemic abnormalities.Microorganisms have been implicated as the primary causative agents in the pathogenesis ofperiodontal disease. Both direct and indirect mechanisms of tissue breakdown have beendescribed. Direct tissue destruction is a result of endotoxins, exotoxins, proteolytic enzymes andcytotoxic substances, which have been shown to disrupt tissue and cellular integrity (Ng andTonzetich 1984; Kryshtalskyj and Sodek 1987; Overall et al. 1987; Uitto 1987). Indirectmechanisms involve activation of host cells in response to bacterial products such as antigens ormitogens. This host-mediated tissue injury is a result of inflammation, mediated via complementactivation and liberated lymphokines from sensitized lymphocytes (Genco and Slots 1984).45Pathogenesis of inflammatory periodontal disease is characterized into four distinct stages(Page and Schroeder 1976). These are described by Page and Schroeder as the initial, early,established and advanced stages. Ultrastructural and histopathological features have beenpreviously used to examine the stages of experimental gingivitis (Listgarten and Ellegaard 1973;Payne et al. 1975). The same techniques were applied to defining the nature of periodontalinflammatory lesions, with insight as to the progression of the disease. The following represents asummary of these stages as first described by Page and Schroeder (Page and Schroeder 1976).One of the major difficulties in understanding the pathogenesis of periodontal disease is theinability to discriminate between normal and pathological tissue. In experimental animals, in whichperiodontal tissues are maintained plaque-free, leukocytes have been observed within the junctionalepithelia (JE). It has generally been accepted that these features reflect enhanced levels of hostdefense mechanisms.The response of the gingival tissues within 2 to 4 days to an accumulation of microbial plaqueis localized to the region of the gingival sulcus. The initial lesion is characterized by acuteexudative vasculitis, exudation of GCF, and increased migration of leukocytes into the J.E. andgingival sulcus. Other features include the presence of serum proteins, alteration of coronal J.E.,and loss of perivascular collagen. The gingival tissues begin to show clinical signs and symptomsof gingivitis.The early lesion develops within 4 to 7 days after the beginning of plaque accumulation. It ischaracterized by a dense infiltrate of lymphocytes, pathologic alterations in local fibroblasts andcontinuing loss of the collagen fibers supporting the marginal gingiva. Additional features includeproliferation of the basal cells of the JE and an increase in the volume of GCF.Two to three weeks after the early lesion the development of the established lesion occurs. It isdistinguished by a predominance of plasma cells within the attached connective tissue, prior to anysignificant bone loss. Although the lesion is confined to the bottom of the sulcus, as in earlierstages, the plasma cells are not confined to the reaction site and are present along blood vesselsdeep within connective tissue between collagen bundles. Other features of the established lesion46include proliferation, apical migration and lateral extension of the JE, with or without pocketformation, and continued loss of connective tissue.The established lesion may remain stable for years or it may develop into a progressive,destructive lesion. Host response to this stage of periodontal disease is vascular proliferation at thesite of infiltration, while fibrosis and scarring may occur at sites distant from zones of continuingcollagen loss. Factors causing the conversion to an advanced lesion are not totally understood.The advanced lesion is characterized by extension of the lesion into alveolar bone andperiodontal ligament with accompanying bone loss. Collagen loss continues subjacent to thepocket epithelium with fibrosis at more peripheral sites. This lesion continues to be dominated bythe presence of plasma cells as in the established lesion. The result of these inflammatory changesis the formation of periodontal pockets. In addition, at this stage of disease, periods of acuteexacerbation and periods of quiescence will occur.The first three lesions, initial, early and established, are equivalent to the stages seen clinicallyin gingivitis. They account for the major portion of inflammatory gingivitis and periodontaldisease. The advanced lesion is manifest as periodontitis and this is seen as alveolar bone loss,increased tooth mobility, pocket formation and periodontal abscess formation (Page and Schroeder1976).Evidence supporting the bacteriological etiology of periodontal disease has led to the evaluationof microbial species associated with different types of the disease. At clinically healthy sites, thereare mainly gram-positive facultative anaerobic cocci and rods. Streptococci species predominate innormal flora as facultative anaerobic cocci. In addition, Actinomyces organisms, particularly A.viscosus and A. naeslundi are present as gram-negative rods (Slots 1977; Slots 1979).In gingivitis, there is a shift in microflora in subgingival plaque toward a higher content ofgram-negative forms. Established gingivitis is characterized by a shift from a Streptococcus to anActinomyces dominated plaque. The organisms found in established gingivitis include A.viscosus, A. israelii, Fusobacterium nucleatum, Campylobacter sputorum, black-pigmentedBacteroides species and Veillonella species (Loesche and Syed 1978). Moore and co-workers, in47extensive analyses of microorganisms associated with human experimental gingivitis, found thatthe subgingival plaque in gingivitis also contains A. naeslundi, A. odontolyticus, Streptococcusanginosus, Veillonella parvula and Treponema species (Moore et al. 1982).The composition of the subgingival flora associated with early onset forms of periodontaldisease, especially juvenile periodontitis (JP), has been extensively studied. The bacteria reportedin localized JP include Actinobacillus actinomycetemcomitans (Aa), Capnocytophaga, Eikenellacorrodens and Bacteroides species, excluding Porphyromonas gingivalis (P. gingivalis) (Slots1976; Slots 1979; Zambon et al. 1981; Mandel 1984).The flora of adult forms of periodontitis is predominantly composed of gram-negative,anaerobic and motile organisms (Moore 1987). The most consistent bacteria recovered fromperiodontally-involved sites includes both P. gingivalis and Aa (Tanner et al. 1984; Dzink et al.1988). With a recent emphasis on periodontally active sites, other micro-organisms found includeWolinella recta, other Bacteroides species and Treponema denticola (Moore 1987).Certain bacteria appear to have important roles in the initiation and maintenance of the majorforms of periodontal disease. For example, Aa and P. gingivalis are the significant organismsimplicated in the pathogenesis of juvenile periodontitis (Zambon 1985) and adult periodontitis(Genco and Slots 1984), respectively. Although the various mechanisms by which thesemicroorganisms participate in periodontal disease are well documented, there is a lack of evidenceto establish a correlation between their presence and disease activity (Listgarten 1987). This ispartly due to the fact that many periodontal pathogens are studied in vitro, which does not take intoaccount the influence of in vivo conditions and that of the host response. It is also due to the lackof an acceptable standard to measure discrete periodontal disease activity.The concept that periodontal disease is a gradual continuous process has been modified to anepisodic disorder that cycles through active and quiescent phases of infection. Monitoring ofindividual periodontally diseased sites indicates that active destruction occurs in short bursts oftime, followed by prolonged periods of tissue stability (Goodson et al. 1981; Goodson et al. 1982;Haffajee et al. 1983; Socransky et al. 1984). Paired measurements of attachment level at two48month intervals for one year revealed a significant increase in probable attachment loss in 2.8percent of 3414 sites in 22 subjects. Attachment level changes that were deemed significant rangedbetween 2 and 5 mm. This observation and fluctuations found when monitoring rates ofdestruction during two month intervals are not consistent with a continuous disease process. Otherauthors have also found that attachment level loss in periodontal disease can occur in bursts ofactivity within a two month period (Kennedy and Poison 1973; Schroeder and Lindhe 1975;Lindhe and Ericsson 1979; Schroeder and Lindhe 1980). If periodontal tissue destruction occursover short bursts of time at individual sites, then it would be beneficial to be able to detect suchperiods of high activity and possibly identify the active sites. This would aid in prevention andtreatment monitoring of the affected sites.1.6.2 Disease progressionThe progression from gingivitis to periodontal disease is not totally understood. It has beenassumed by clinicians that destructive periodontal disease is a result of progressive gingivitis.Although some investigators have found no firm evidence that this progression occurs (Schectmanet al. 1972; Page et al. 1975), other studies indicate that gingivitis always precedes periodontitis(Saxe et al. 1967; Lindhe et al. 1973). It has been shown in the dog model that with plaqueaccumulation, gingivitis does progress to periodontitis (Lindhe et al. 1973).The next consideration is whether periodontitis can occur without passing through a phase ofgingivitis. The idea that periodontal inflammation and destruction are discontinuous is an observedcharacteristic in the advanced lesion (Page and Schroeder 1976). Evidence that gingivalinflammation is not related to attachment loss comes from clinical studies in humans. Haffajee andco-workers found sites of progressive attachment loss through episodic bursts of activity,generally in the absence of clinical signs of gingivitis (Haffajee et al. 1983). They monitored 3414individual periodontal sites and found no significant correlations of gingival redness, plaque,bleeding on probing or suppuration with periodontal destruction determined by probing attachmentloss. In addition, of all sites that showed no significant attachment loss, the majority showed49positive signs for gingival redness, plaque, bleeding on probing and suppuration. These resultsindicated that some periodontal lesions occur with gingivitis as a precursor. However, in somecases gingivitis may not presage destruction of periodontal attachment .The concept on the pattern of progression of the periodontal lesion has changed during the pasttwo decades. Three models representing the progression of chronic destructive periodontal diseasehave been reviewed by Socransky and co-workers (Socransky et a. 1984). It was traditionallythought that periodontal disease was continuous in nature, following a linear model of progressivedestruction. Once periodontal disease was initiated at a particular site, there was progressiveattachment loss with time. This led to the belief that unless an affected tooth was treated theperiodontal condition would worsen.There are several reasons why the model for continuous or linear disease progression isincorrect. Firstly, studies have shown that attachment loss can occur very rapidly which isinconsistent with this model. Studies by Goodson and co-workers on progression patterns ofadvanced destructive periodontal disease have demonstrated significant increases in probeableattachment loss, between 2-5mm over two month intervals (Goodson et al. 1982; Haffajee et al.1983). Secondly, the fact that there are a few sites that exhibit no changes in periodontal activity isinconsistent with a slow progressive destructive disease model. Two longitudinal studies by thesame authors, each monitoring periodontal attachment for one year, have shown that the number ofsites exhibiting significant attachment loss were 2.8% and 5.7%. And finally, data from animalstudies indicate that periodontal disease does not progress in all lesions (Lindhe et al. 1975).The second concept of chronic periodontal disease progression is the random burst hypothesis.In this model the cumulative extent of the destruction varies depending upon the site. Some sitesshow several bursts of activity, while other sites show no activity and may be free of periodontaldisease throughout an individual's life. Sites that show activity may never demonstrate activityagain or could be subject to further activity in the future. The term random is related to time andprevious loss of attachment. There is good evidence to support the hypothesis that periodontaldisease is not random with regard to site of involvement (Lee et al. 1978).50The third concept of disease progression was the asynchronous multiple burst model.According to this model, the majority of periodontal activity occurs within a few years of anindividual's life. Sites can show repeated bursts of activity followed by prolonged periods ofinactivity. The major difference from the random burst model is that multiple sites can break downwithin a relatively short period of time, which is then followed by prolonged periods ofquiescence.The complement system plays an important role in the pathogenesis of and protection inperiodontal disease. Activation of complement by bacteria intiates a number of events includingopsonization, chemotaxis and release of inflammatory mediators, and bacteriolysis via themembrane attack complex. Complement cleavage products have been isolated in GCF andobserved in diseased gingival tissues. In a recent article, Schenkein (Schenkein 1991) reviewedthe role of complement in periodontal diseases. Although a number of bacteria from individualswith periodontal diseases can activate complement, some of these organisms evade opsonizationdue to their proteolytic activity. Schenkein concluded that there is insufficient concrete evidence toindicate that complement activation occurs in human periodontal disease and that it is important inits pathogenesis or in protection against bacteria.However, other investigators concluded that the presence of complement proteins in thecrevicular fluid provides evidence for their activation in active disease. In comparing healthy andperiodontally involved patients, the presence of C5 and decreased levels of C3 and C4 were foundin diseased sites (Attstriim et al. 1975). More recent studies have also shown that complementcleavage in crevicular fluid can be used to distinguish pre- from post-treatment sites, and adultfrom juvenile periodontitis (Niekrash and Patters 1985; Niekrash and Patters 1986). Suchobservations may be useful in elucidating pathogenic mechanisms of the host.Increased degradation of collagen is a characteristic feature of the onset and progression ofperiodontal diseases. Mammalian collagenases have a specific activity, whereby cleavage of thecollagen molecule generates 3/4 (TCA) and 1/4 (TCB) fragments. The demonstration of tissuecollagenase activity in vivo, elaborating TCA and TCB fragments, has been shown to correlate with51inflammation severity in human gingiva (Overall et al. 1987). The generation of 3/4 and 1/4fragments is a critical step in collagen degradation since the fragments are unstable at physiologicaltemperatures and are subject to further degradation by specific and non-specific gelatinases (Weiss1976).Since collagen loss occurs in both gingivitis and periodontitis, it is not surprising thatcollagenolytic activity in GCF has been correlated with the severity of periodontal involvement(Golub et al. 1976; Uitto and Raeste 1978; Kowashi et al. 1979; Kryshtalskyj and Sodek 1987;Villela et at 1987). The studies by Kryshtalskyj and Sodek and by Villela and co-workersdescribe an active collagenase which degraded collagen into 3/4-a and 1/4-a fragments. Thiscleavage pattern is characteristic of mammalian, and not microbial, collagenase.1.6.3 Clinical evaluationsClassical clinical parameters used to measure periodontal conditions include plaque index(P1.1), gingival index (GI), pocket depth (PD), and bleeding on probing (BoP). It has been widelyaccepted that bacterial plaque is not just associated with inflammatory disease, but is the cause ofthis destructive process. While NJ is used as an indirect measure of potential inflammation of thegingival tissues, both GI and BoP provide direct evidence that inflammation is present at the timeof examination. Measurements of increasing PD however, provide information that priorperiodontal destruction has occurred. While all these indices indicate the presence of periodontaldisease, they provide no information on disease activity at the time of examination.The periodontal probe has been and continues to be one of the most useful and fundamentaltools used in periodontal diagnosis. Assessment of periodontal defects such as pocket depths,width of keratinized tissue and gingival recession are measured with this instrument. However,statements as to the presence of active disease cannot be made when measuring such periodontalsites.Retrospective disease activity can be ascertained by comparing changes in probingmeasurements obtained at different time points. Goodson and co-workers (Goodson et al. 1982)52have demonstrated that changes in probing periodontal attachment levels can be discriminated witha high degree of accuracy.There is controversy as to the probing depth measurement and how it relates to JE attachment.Although the intent is to probe to the level of the cementoenamel junction (CEJ) (for a healthytooth), there are reports that show that this does not always occur (Listgarten 1980). The degree ofprobing force and the diameter of the probe both contribute to the extent of the penetration of theinstrument. In addition, the degree of inflammation in the periodontal tissues is also an importantfactor in how deep the probe penetrates. The greater the inflammation present, the deeper theprobe penetrates the JE beyond its coronal extent, towards the connective tissue attachment(Hancock and Wirthln 1981).The accuracy of probing measurements is further compromised in at least two ways. Themarkings on the probe can vary from 1 mm to 3 mm gradations, and therefore, accurate at best to1mm. Another source of error is the probing force used by the operator. To improve upon theselimitations, an electronic probe was developed which operates at a constant probing force and isaccurate to 0.1 mm.One of the first electronic probes that became available was the Florida probe (Florida ProbeCorp., Gainesville, FL.). Due to its features of a constant probing force and a 0.1 mm resolution,this instrument compared to the conventional probe reduces the inter-operator variability inmeasuring pocket depth (Walsh and Saxby 1989).Radiographs are frequently used to detect and assess periodontal disease. They offerinformation on the level of alveolar bone present and alterations in bony contour. However, singleradiographs only yield information as to the present state of the periodontium. They only representa one-frame look at the condition, and thus, only by comparing radiographs taken at different timescan one retrospectively demonstrate disease activity.There are a number of limitations in the use of radiographic interpretations. One of the mainlimitations of a radiograph is that it detects changes in bone density only after 30-50% of themineral content of the bone has been lost (Bender and Seltzer 1961a,b). In addition, the53reproducibility of angulation and quality of film can obscure the subtle bone changes on theradiograph.A method for comparing radiographs from different time points is possible through digitialsubtraction radiography. This is an imaging modality commonly used in medical radiology,whereby each image is scanned by a computer and structures that have not changed densitybetween examination points are subtracted to produce the image. The result is a generation of animage where mineralized changes have occurred. The benefit of this technology is that it not onlyeliminates operator subjectivity in examination, but that it can detect as little as a 5% change in themineral content (Ortman et al. 1982).A study of the relationship between attachment level loss and alveolar bone loss demonstratedthat attachment loss precedes radiographic evidence of crestal bone loss during periods of diseaseactivity (Goodson et al. 1984). In view of this result, comparison by subtraction radiography offilms taken at different times can demonstrate changes in periodontal mineral content, but againwould demonstrate disease activity retrospectively.Parameters such as the gingival index and bleeding on probing are sensitive indicators ofgingival inflammation. Several studies have shown correlations between the clinical signs ofinflammation and histological evidence of inflammatory gingival lesions (Applegren et al. 1979;Greenstein et al. 1981; Abrams et al. 1984). However, these clinical signs pertain to the status ofthe gingiva alone and are only useful in diagnosing gingivitis and are of little value in identifyingperiodontitis, or being able to predict the onset of periodontitis.Recognition of active destructive periodontal disease is difficult. Studies have shown thatperiodontal disease is cyclical and site specific, and that disease progresses for a relatively shorttime, soon replaced by longer or shorter periods of quiescence (Goodson et al. 1982).541. 6 . 4 Host interactionsMacrophages play an important role in the host response to periodontal disease. Directstimulation of these cells by bacterial products, such as lipopolysaccharide (LPS), or by indirectstimulation by lymphokines leads to the production of collagenase and other proteolytic enzymes(Wahl and Mergenhagen 1988). Macrophage interaction and processing of cell surface boundantigens, gives rise to the production of monokines which amplify the immune response. One ofthe better characterized monokines is interleukin-1 (IL-1), a potent stimulator of T cells. IL-1affects a number of target cells including thymocytes, T and B lymphocytes, fibroblasts andhepatocytes. Richards and Rutherford have recently studied IL-1 effects on the collagenolyticactivity and pro staglandin-E (PGE) secretion by human periodontal ligament fibroblasts (PLFB1)and gingival fibroblasts (GFB1) (Richards and Rutherford 1988). They reported that while PLFB1and GFB1 both responded to IL-1 by secreting PGE, only GFB1 produced increased levels ofcollagenolytic activity.Human GCF has been found to contain thymocyte activating factor (Mergenhagen 1984). Inaddition to macrophages, epithelial cells are capable of secreting an Epidermal Cell ThymocyteActivating Factor (ETAF), a protein which is similar if not identical to IL-1 (Luger et al. 1982).Since IL-1 and ETAF were found to be physiochemically and biologically similar, it is not possibleto differentiate whether the source of activity is from epithelial cells or macrophages.Analysis of gingival fluids collected from inflamed and noninflamed sites indicated thatthymocyte growth promoting activity was significantly increased in the inflamed sites (Charon etal. 1982). They collected specimens of fluid by a gingival washing technique (Skapski and Lehner1976), which is a crude method for collecting gingival samples. It is now preferred to collect GCFusing either filter paper strips or calibrated microcapillary tubes, which allow for site specificsampling and quantitation of fluid volume.For a considerable time, immunological mechanisms have been implicated in the pathogenesisof chronic periodontal disease. Cell mediated (Ivanyi et al. 1972) and humoral responses(Brandtzaeg 1966) have both been linked to progression of the disease. Initial opinions of55lymphocyte function within inflamed tissues were primarily based upon studies performed onperipheral blood lymphocytes (PBL) from patients with periodontal involvement (Campana 1981).As periodontal disease is localized to the supporting tissues of the tooth, activity of PBL may notreflect the activity of lymphocytes present within the lesion. In order to better understand thepathogenesis of the disease it is necessary to characterize cells present in the tissue. The study offunctional characteristics of lymphocytes in chronically inflamed tissue will lead to a betterunderstanding of the role of the immune system in periodontal disease.As stated previously the progression from gingivitis to periodontitis proceeds from the initiallesion, through early and then established lesions, and finally to the advanced lesion.Histochemical and immunological studies of lymphoid cell subpopulations in experimentalgingivitis in humans have characterized gingivitis as T-cell dominated (Seymour et al. 1983). Thedistinction between gingivitis and periodontitis is that in the established lesion of gingivitis there isno bone destruction, while the advanced lesion of periodontitis involves progressive bone loss.The switch from a stable lesion to a progressive lesion has been shown to be associated with aswitch in lymphocyte populations, from T cells to B cells Wackier et al. 1977). Phenotypiccharacterization of lymphocyte populations in established human periodontal disease suggests thatit be considered as a B cell lesion (Seymour and Greenspan 1979).The T lymphocyte role in the regulation of the immune system can range from enhancement tosuppression. The involvement of T helper (TH) and T suppressor (Ts) cells in periodontal lesionscan be either protective or destructive, depending upon the type of antigen stimulation. If anantigen is T cell dependent, TH cells would act protectively by stimulating B cells to become IgG,IgA or IgE secreting plasma cells. If however, an antibody response is T cell independent, as forLPS antigen, a polyclonal B cell activator, TH cells would give rise to overstimulation of B cells.This would lead to an overproduction of antibodies and Type DI hypersensitivity reactions.Alternatively, Ts cells in this instance would function in a protective role.Cell-mediated immunity (CM1), has been shown to relate to the periodontal condition in studiesof the lymphoproliferative response during human experimental gingivitis (Patters et al. 1979).56The sometimes protective, sometimes destructive consequences of CMI leaves one to questionwhether the host response is beneficial or harmful in cases of periodontal disease. Evidence ispresent to support both these contradictory roles of the immune response.In studies performed on animals, variations in T cell populations resulted in differing effects onperiodontal bone loss. Thymectomized (Tx) hamsters showed a greater bone loss then did sham-Tx controls (Barefoot and Silverman 1977). In contrast, studies on rats with an imbalance of Tcell regulation, either congenitally athymic or neonatally Tx, demonstrated elevated bone lossrelative to controls. Experiments involving T cell suppression offered similar contradictory results.T lymphocyte reduction from cyclosporin A treatment resulted in decreased bone loss(Guggenheim et al. 1981), while cyclophosphamide suppression resulted in dramatically increasedbone loss after ligature placement (Sallay et al. 1982). Therefore, the question whether the hostresponse is harmful or beneficial remains unresolved.Studies of host response in experimental periodontal disease in rats indicate that changes in Tcells, depending on the degree of sensitization and subset involved, result in differing effects onperiodontal bone loss (Taubman et al. 1984b). When sensitized T lymphocytes are adoptivelytransferred, bone loss is increased, presumably through the mechanism of host synthesis ofosteoclast activating factor (OAF) and other lymphokines. However, when athymic rats, havingan imbalance of TH and Ts cells, are reconstituted with a balanced population of thymus cells,bone loss is decreased (Yoshie et al. 1983). Although these results suggest that T cells play a rolein the regulation of periodontal bone loss, further studies are necessary to determine the T cellsubsets present and their influences at various stages of periodontal disease.The application of immunocytochemical methods to histological sections has led to thequantitation of lymphocyte phenotypic variations in experimental gingivitis and in periodontaldiseased tissues. Characterization of lymphocyte subpopulations in established periodontal diseasewas determined using indirect immunofluorescence to detect human thymocyte antigen, humanmyeloid antigen and IgG and IgM-bearing cells in situ in the lesion (Seymour and Greenspan1979). The results indicated that the majority of lymphoid cells were IgM-positive/thymus antigen-57negative B cells, while relatively few thymus antigen-positive T cells were found. In a similarstudy of cryostat sections from patients with chronic marginal periodontitis, a three layerimmunofluorescent technique was used to differentiate T helper cells (TH) from Tsuppressor/cytotoxic cells (Tsd (Johannessen et al. 1986). The ratio of TH to TSC was 1.13.These results are in agreement with Taubman and co-workers (Taubman et al. 1984a), who founda TH/Tsc ratio of 1.1, by measuring immunofluoresence of extracted cells separated by Ficoll-Hypaque gradient centrifugation.In studies of experimental gingivitis, Seymour and co-workers (Seymour et al. 1983), usedtwo and three layer immunofluorescent techniques to characterize lymphocyte differentiationantigens, and conjugated fluorescein isothiocyanate (tin C) rabbit anti human-Ig for demonstrationof plasma cells. Although these authors found a predominance of T cells throughout their 21 daystudy, further work was necessary to characterize T-cell subsets. These results became available ina subsequent immunohistological analysis of gingivitis in 1988, as Seymour and co-workers wereable to identify T-cell subsets and the pattern of Class II major histocompatibility complex (MHC)antigens. Using a panel of monoclonal antibodies in an avidin biotin immunoperoxidasetechnique, the ratio of TH cells (CD4+) to Tsccells (CD8+) varied only slightly from 2.18 to 2.48over the course of the lesion. A double staining immunofluorescence technique using 1-41 C-conjugated OKT8 antibody and phycoerythrin-conjugated anti-Leu-15, was used to differentiate Tsuppressor from T cytotoxic subpopulations. Examination with an epifluorescence microscoperevealed that all T8+ cells co-expressed Leu-15, indicating a suppressor rather than a cytotoxicphenotype, suggesting a regulatory role for these cells. In addition, the majority of T cellsexpressed the Class II MHC antigens HLA-DR and HLA-DQ; this implied that they were activated.Although these are the first studies to classify T-cell subpopulations and determine theiractivation, the clinical data are insufficient to differentiate the periodontal status with respect todisease activity. The chronology, stage of development and current destructive activity inindividual periodontal lesions, in the studies of experimental gingivitis and chronic marginalperiodontitis, were not recorded. The limit of controlled experimental periodontal disease in58humans is to the gingivitis stage, as it would be unethical in humans to allow gingivitis to progressto a course of untreated experimental periodontitis. For this reason an indicator of active disease,such as hydroxyproline (Coil et al. 1987), is essential in ascertaining the activities of individualsites and predicting clinical changes in the periodontium. Markers for active disease would notonly allow clinicians to monitor patient periodontal status but also enable scientists to correlate withclinical parameters with biological phenomena.1.6.5 TreatmentThe goal of periodontal therapy is to restore the periodontal tissues to a healthly state andfunction. The treatment is directed at reducing or eliminating bacterial plaque and resolvingperiodontal inflammation. This is accomplished by removal of local factors by conventional non-surgical methods of scaling and root planing. Following treatment, prevention of further diseasedepends mainly on the proper management of the bacterial plaque by mechanical and/or chemicalmeans.Scaling and root planing in the initial phase of treatment has been supplemented withantimicrobial drugs. The use of these antimicrobial agents can be divided into those for use inplaque control and those used in the elimination of subgingival microflora. A common therapeuticagent used for plaque control is chlorhexidine mouthrinse. When used on a daily basis it has beenshown to reduce supragingival plaque. The use of this agent in subgingival irrigation may alsoprove to be effective in reducing the subgingival microflora.Systemic antibiotics have also been used alone and as adjuncts with scaling and root planing intreatment of advanced forms of periodontal disease. A variety of antibiotics have been usedincluding tetracycline (Listgarten et al. 1978; Hellden et al. 1979; Slots et al. 1979), penicillin(Helovuo and Paunio 1989), metronidazole (Jenkins et al. 1989; Soder et al. 1990; Loesche et al1991, 1992), and spiramycin (Mills et al. 1979; Al-Joburi et al. 1989). Tetracycline when givensystemically in conjunction with conventional therapy, showed only minor differences in microbialand clinical parameters when compared to scaling and root planing alone. Overall, the studies have59shown conflicting results regarding the benefit of systemically administered antibiotics inperiodontal therapy.There has been an increased interest on the methods and effects of local delivery of antibioticsto specific periodontal sites (Goodson et al. 1979; Goodson et al. 1983; Goodson et al. 1985;Goodson et al. 1991; Heijl et al. 1991). Local delivery has advantages of directing site specifictreatment and maintaining high concentrations of antibiotic locally. Matrix systems for controlleddrug release are employed to maximize and sustain the local concentration of therapeutic agent.Investigations in this area continue to evaluate matrix support systems and different antimicrobialagents.Recent advances in the use of tetracycline in periodontal therapy have demonstrated a non-antimicrobial property of therapeutic significance. Golub and co-workers have shown thattetracycline directly inhibits the extracellular activity of a metalloproteinase collagenase (Golub etal. 1983; Golub et al. 1984; Golub et al. 1985). This agent may suppress connective tissuebreakdown during active periods of periodontal disease. This important finding has expandedexperiments into localizing this benefical property of tetracycline, and has led to the development ofchemically modified tetracyclines (CM1) (McNamara et al. 1986; Golub et al. 1987). The latteragents have lost their antimicrobial properties but retain their anticollagenase activity (Golub et al.1991). Although CMT's have not been approved for human use, they possess great potential infuture treatment of periodontal disease.After intial therapy, surgery may be required in more severe forms of the disease to furthereliminate inflammation and/or improve the architecture of the periodontal tissues. Althoughsurgical therapy is directed at the treatment of remaining periodontal pockets, creation of a moreaccessible contour which is conducive to more effective home care maintenance is also a goal. Themain approaches to surgery involve resective, inductive and reattachment procedures. Thesesurgical methods should be regarded as an extension of the basic principle of mechanicaldebridement performed during initial therapy.601.7 INDICATORS OF PERIODONTAL DISEASEAn ideal indicator of periodontal disease should be able to do the following: 1) diagnose thepresence of disease at the time of examination, 2) evaluate disease activity and identify active sitesat the time of examination, 3) predict and determine the degree of future periodontal attachment lossthereby serving as a predictor of subsequent breakdown of periodontal tissues that would occur inabsence of treatment, and 4) quantitatively monitor the response to therapy.1.7.1 Clinical indicatorsAn accurate assessment of a patient's periodontal health is a key to the diagnosis anddetermination when periodontal therapy is required. As previously indicated, once a patient has ahistory of periodontal disease, the active phases of the disease are characterized by short bursts ofactivity, generally followed by longer periods of quiescence. For this reason it is difficult to assessprecisely the progression of the disease.At present the only methods available to clinically measure periodontal disease activity are tomonitor retrospective changes in alveolar bone levels and changes in attachment level over time.Clinical parameters such as gingival index, bleeding on probing, crevicular fluid flow and plaqueindex, provide information on the state of gingival inflammation at a given examination point. Arecent 2 year study on the relationship of gingival bleeding and supragingival plaque to attachmentloss _. 2mm, confirmed that neither of these signs were prognosticators of attachment loss duringthis period (Kaldahl et al. 1990). Pocket depth, attachment and alveolar bone level, are indicatorsof patient history of periodontal involvement. Measurements of recent reductions in alveolar boneheight and/or attachment level loss indicate that active disease has occurred. However, monitoringof these clinical parameters yields only a retrospective diagnosis of active disease.The gingival crevice is an important milieu for the periodontal disease process. This is a logicallocation to seek for diagnostic aids for evaluating the destructive disease process. Crevicular fluidemanating from the crevice contains a multitude of substances. Although its composition is similar61to that of saliva, it differs from it with respect to bacteria, enzymes, immunological andinflammatory mediators, and by-products of metabolism. One means of investigating the complexnature of periodontal disease is by the analyses of this unique transudate.1. 7 . 2 Bacterial indicatorsStudies of microorganisms in periodontal pockets of periodontitis subjects demonstrateincreased numbers of gram-negative anaerobic rods and spirochetes as compared to unaffectedgingival sulci. Evidence that bacteria contribute to the etiology of periodontal disease has promptedinvestigators to determine which microorganisms are associated with different forms of thedisease. Although an overall increase in mircoorganisms is associated with periodontal conditions,their presence does not necessarily connote disease activity (Haffajee et al. 1983).The progressive destruction of the periodontium is the result of an active state of the disease.At a given time point, a previously diseased periodontium can clinically be in a healing, stable, or acontinuing active disease state. Although it is not presently known what triggers thetransformation from a stable to an active disease state, it has been suggested that a significantnumber of pathogenic bacteria may cause or initiate the destructive process. Furthermore, thepresence or increased numbers of certain bacteria may be indicative of the active disease state.A number of microbiological assays have been used to monitor periodontal diseases. Culturingmethods were first performed to detect and characterize bacteria associated with various forms ofthe disease. In localized juvenile periodontitis (LW), Slots reported that Capnocytophaga, Aa, andWolinella recta were the dominant isolated types (Slots 1976). In adult periodontitis, White andMayrand found that in comparison to less inflamed sites the inflamed sites had a greater portion ofgram-negative anaerobic rods, 32% of which were P. gingivalis (White and Mayrand 1981).The culture results were often compared to clinical parameters to seek information as to theseverity of the disease or the level of disease activity. By development of bacterial profiles it wasintended to develop a diagnostic test for various forms of periodontal disease. Slots and co-workers determined the presence of Aa, P. gingivalis and B. intermedius in destructive62periodontitis in adults (Slots et al. 1986). Their definition of progressive disease was based onradiographic evidence of alveolar bone loss over a minimum period of two years. They found thatat least one of these organisms was present in 99% of the investigated progressive lesions.However, their definition of progressive disease was actually a retrospective evaluation thatperiodontal disease had occurred.Using immunological techniques, investigations of antibodies to specific oral microorganismshas enabled better understanding of certain periodontal conditions. It has been demonstrated thatelevated serum antibody titers to Aa are present in individuals with localized juvenile periodonditis(LW) (Ebersole et al. 1982; Ranney et al. 1982). Similarly serum IgG titers to P. gingivalis havebeen shown to be elevated in adult and rapidly progressive periodontal diseases (Mouton et al.1981; Tolo and Schenck 1985). A recent study compared antibodies from both serum and GCFsources to P. gingivalis in patients with treated and untreated periodontal disease (Murray et al.1989). The results showed that untreated periodontitis patients had significantly higher serum andGCF levels of IgG antibody to P. gingivalis than did either patients with treated periodontitis orgingivitis controls.In a longitudinal study of serum antibodies to P. gingivalis in periodontitis patients, Moutonand co-workers (Mouton et al. 1987) observed an unexpected consistent serological dichotomy inchronic periodontitis patients. Throughout the one year study, half of the periodontitis group didnot exhibit measureable levels of IgA, while the other half remained seropostive.Correspondingly, in comparison to healthy subjects the chronic periodontitis patients eitherexibited normal antibody titers or significantly increased levels of antibody reactive to P.gingivalis. Although it could be conjectured that low levels of antibody reflect low colonization,this may not be true since it has been reported that large numbers of specific microorganisms canbe found in the absence of their corresponding specific antibody (Williams et al. 1985).Microbial measurements can be useful in selecting and evaluating treatment, and forestablishing recall schedules. Robertson and co-workers assessed subgingival microflora andclinical parameters at baseline and after 7 months in a group of severe peridontitis patients that63received scaling and root planing (Robertson et al. 1987). They found that after 7 months, thepercentage of sites with P. gingivalis was significantly lower in patients with resolved than inpatients with unresolved periodontal sites. However, they concluded that none of the clinical andmircobial measures used in the study could predict the response to repeated scaling and rootplaning.A recent 30 month investigation by Listgarten and co-workers, on treated patients who were ona 3 month recall, compared disease recurrence with analyses of specific culturable bacteria(Listgarten et al. 1991). They found that the presence or absence of Aa, Prevotella intermedia, orP. gingivalis, was not a reliable predictor of future episodes of recurrent disease.An indirect method for monitoring bacteria is by assessing microbial enzyme activity. Onemethod is based on hydrolysis of a specific peptide, benzoyl-arginine naphthylamide (BANA), bya trypsin-like enzyme found in a limited number of microorganisms. Periodontal microfloracapable of this cleavage are Treponema denticola, P. gingivalis, and Capnocytophaga (Schmidt etal. 1988). Clinical studies have shown that BANA cleavage is strongly associated with thepresence of T. denticola and P. gingivalis (Bretz et al. 1990). The presence of these specificbacteria may be an important prognostic marker. A recent study that examined subgingival plaquefor these two organisms before and after treatment indicated that T. denticola levels decreased athealing sites and either increased or remained the same at nonresponding sites (Simonson et al.1992).New detection methods for bacteria have enabled a more specific and expedient evaluation ofspecies present at periodontal sites. The use of DNA probes offers improved sensitivity inmicrobial detection. A recent study compared culture methods and DNA probe analyses for thedetection of Aa, P. gingivalis, and B. intermedius, in subgingival plaque samples (Savitt et al.1988). The results indicated that probe assays frequently identified these pathogens in samples thatwere culture negative. DNA probes also revealed a better correlation on an individual patient basisbetween the presence of a pathogen and clinical evidence of disease.64There are a number of limitations to the currently used assays. The premise for their use is thatspecific microorganisms were present in healthy states, whereas others are associated with specificperiodontal diseases. However, according to Listgarten (Listgarten 1988), an association betweencertain bacteria and clinical conditions does not make them causative agents, nor does it imply theirpresence is a useful diagnostic tool. It is possible that bacteria may be present in disease conditionsbecause the environment is conducive to their growth and proliferation. In this case the presenceof bacteria would be the result rather than the cause of the disease.Significant variations in bacterial composition were found at healthy and diseased periodontalsites (Listgarten and He'Men 1978; Listgarten 1984). Furthermore it has been shown that most ofthe variability is due to differences among subjects rather than differences amongst diseased siteswithin subjects (Evian et al. 1982). This has obvious implications on pooling of bacterial samplesand comparisons of healthy and diseased sites in the same individual.Also of concern is the reliability of diagnostic tests, as the host response plays a role indetermining whether the disease remains stable or progresses. The interpretations ofmicrobiological tests are limited as they do not measure host-microbiologic interactions. Hencethey cannot detect active states or predict the future progression of disease.Diagnostic tests for bacterial species such as Aa and P. gingivalis may serve to identify sitesand/or patients at risk for periodontitis. As there are no known longitudinal studies that compareand monitor specific subgingival microflora and disease progression, the ability of such a riskindicator is unknown. However, as discussed above, since the host response is important indetermining the onset and progression of disease, the presence or number of pathogenic bacteriamay be superfluous. The presence of the same pathogenic bacteria in one host may result indisease progression, whereas the same proportions of bacteria in another host may result in a stablelesion.The use of microbial indicators to detect the presence of microorganisms is most useful inmonitoring the effectiveness of therapy. They may become an important diagnostic aid to ensurethat potentially pathogenic bacteria have been successfully eliminated. Although culture techniques65are able to identify a multitude of species and subspecies of microorganisms, they are not practicalfor routine use in clinical practice. DNA probes, immunofluoresence microscopy and enzymemarkers would provide faster, more sensitive results, provided specialized laboratories or chairsidetests are available to perform such analyses.In light of these limitations, a more practical use of microbiological assays appears to be formonitoring treatment. Quantitation of specific bacteria known to be present in higher amounts atdiseased sites can enable specific monitoring of treatment and targeting of those sites that requirefurther elimination of bacteria.1.7.3 Enzyme indicatorsA multitude of enzymes are involved in the metabolism and pathogenesis of periodontaltissues. They can be derived from microbial pathogens and host sources. The identification andcharacterization of such enzymes may provide rationale for their use as potential markers of thedisease process.As previously described, the process of collagen degradation by host enzymes involvesmammalian collagenases, which are produced by PMN's, fibroblasts, epithelial cells andmacrophages. As this enzyme is secreted as a proenzyme, activation requires cleavage by otherenzymes. This can be effected by such enzymes as neutrophil elastase, mast cell tryptase, andCathepsin G.Gingival crevicular fluid from periodontitis sites has been shown to possess activecollagenolytic activity, whereas control and gingivitis sites exhibit latent collagenase andcollagenase inhibitor complex (Kryshtalskyj et al. 1986). Subsequently it has been shown that theincreased collagenolytic activity associated with periodontitis sites is accompanied by otherenyzmes which further degrade collagen beyond the 3/4 and 1/4 fragments (Kryshtalskyj andSodek 1987).Other investigators have demonstrated higher collagenase activity in inflamed gingiva andcrevicular fluid than in healthy tissue and fluid (Uitto and Raeste 1978; Overall et al. 1987).66Furthermore, it was found that relatively less inflamed tissue released more latent collagenase thanthe more severely inflamed gingiva. In a study of collagenase in different forms of periodontaldisease, it was determined that amongst patients collagenase activity increased with the severity ofthe disease in the order of healthy < gingivitis < periodontitis (Villela et al. 1987). Amongst sites,a significant correlation was found between GCF collagenase activity and pocket depth in bothchronic adult periodontitis and LJP. Furthermore, it has been observed that interstitial collagenaseactivity decreases after periodontal treatment (Hakkarainen et al. 1988).The preferred methods of quantifying mammalian collagenase are quite elaborate and involvethe use of radiolabelled collagen substrates for SDS/PAGE analysis of resultant breakdownproducts. This methodology is not adaptable for a simple, quick, or practical clinical test.Immunological methods would be useful in detecting collagenase except that they also detect thelatent form of the enzyme. Thus, it appears that collagenase assays offer no additional diagnosticbenefit over the conventional clinical methods in identifying the disease state.A variety of other proteolytic enzymes are believed to play a role in the pathogenesis ofperiodontitis. It is likely that the destructive process is associated with increased levels of theseenzymes because they are capable of degrading tissue components. Using synthetic substratescoupled to fluorogenic agents, serine proteinase activity was found in crevicular fluid. A tryptase-like activity was found in GCF collected from gingivitis and periodontitis patients (Cox and Eley1989b). An expanded investigation employing synthetic peptides linked to fluorogenic agents wasperformed to detect cathepsin B and L, elastase, typtase and trypsin-like enzyme activitiesconsistent with a host cell origin. Also, dipeptidyl peptidase-like activity (DPP IV), which canresult from a mixture of host and bacterial sources, was also investigated. Cox and Eley (Cox andEley 1989a) found that all 5 classes of enzyme activities were detected, and that it is possible toanalyze all 5 activities using only 0.1 ill of GCF. Althougth enzyme concentrations varied widelyat individual patient sites, elastase-like activity was generally the highest.The same fluorogenic methods were also used to determine cathepsins B,H and L activities inGCF from chronic adult periodontitis and experimental gingivitis patients. Kunimatsu and co-67workers found higher levels of these cathepsins at sites with more severe signs of the disease(Kunimatsu et al. 1990). The total activity of each enzyme (per unit time) was positively correlatedwith the GCF volume, in contrast to a negative correlation found between specific activity of eachenzyme in GCF (activity units per mg of protein) and GCF volume. From this the investigatorsconcluded that these proteases are selectively released into the crevice. However, in the employedGCF collection procedure the first filter paper used to collect fluid during the first 30 seconds wasdiscarded. This has critical consequences as it has been recently demonstrated that theconcentration of host or bacterial substances is likely to be highest in the first collected sample ofGCF (Lamster et al. 1989).In an attempt to evaluate GCF proteases for their diagnostic potential, Eley and Cox employedthe same 5 classes of fluorogenic substrates to compare protease activity in GCF with clinicalparameters which included probing depth (PD), clinical attachment loss (CAL), gingival index(GI), bleeding index (BI) and plaque index (P1.1) (Eley and Cox 1992). In this cross-sectionalstudy they found that total enzyme activities had good diagnostic specificity and sensitivity aspredictors of clinical parameters. They obtained the following order of correlations with differentenzymes and parameters; cathepsin B/L- > elastase- > DPP IV- > trypsin- > tryptase-like activityand PD > CAL > GI > BI > PH. A relatively high level of correlation with pocket depth infersthat the total enzyme activities are, in part, dependent upon the size of the GCF reservoir in thepocket.Lactate dehydrogenase (LDH), 0-glucuronidase (BG) and arylsulfatase (AS) activities havealso been evaluated in GCF (Lamster et al. 1985a-c; Lamster et al. 1988). LDH is a cytoplasmicenzyme and its occurence in extracellular fluid indicates cellular necrosis. Both BG and AS areground substance degrading enzymes involved in degradation of GAG'S, and are constituents oflysosomal granules.Lamster and co-workers investigated these enzymes in a six month study to determine if theiractivities parallel clinical attachment loss seen in patients with chronic adult periodontitis (Lamsteret al. 1988). In a subgroup of patients that was identified as displaying a localized form of disease68activity, clinical attachment loss could be substantiated by individual GCF samples thatdemonstrated BG activity at least 4 times the population mean baseline value. In contrast, LDHand AS did not provide a statistically significant measure of localized attachment loss in this groupof patients. Both the sensitivity (89%) and specificity (89%) for the relationship of BG activity inGCF to detection and prediction of clinical attachment loss, indicated that the analysis haddiagnostic value. However, more studies are required to further examine this relationship, sincethis subgroup of patients displaying localized disease activity consisted of only 4 patients.Elastase relationship to periodontal health has been studied in both GCF and salivary systems.Zafiropoulos and co-workers (Zafiropoulos et al. 1991) performed a detailed cross sectional studythat compared GCF elastase inhibitor complex (GCF ELPai-PI) with clinical indices andsubgingival flora. They found that correlations between GCF ELPai-PI and P. gingivalis(r=0.642) or B. intermedius (r=0.774) were the highest for alveolar bone loss 20% and pocketdepth 3mm, respectively. These results indicate the measurement of GCF ELPai-PIconcentrations may be useful for evaluating sites with little or no tissue destruction.In a recent investigation Palcanis and co-workers used a prototype elastase diagnostic kit toassess the potential use of elastase as a marker of periodontal disease activity {Dentsply, YorkPA }(Palcanis et al. 1992). Total elastase was significantly higher in sites demonstratingprogressive attachment loss than in inactive sites and sites demonstrating bone loss. When thejoint presence of bone loss and attachment loss were considered together, the sensitivity andspecificity of the assay was 82% and 66% respectively. Since GCF elastase levels are significantlyhigher in sites demonstrating progressive attachment and bone loss assessed 6 months later,elastase may serve as a predictor for these changes.The relationship of salivary elastase and collagenase to periodontal health has been assessed byUitto and co-workers who studied the possibility of using salivary elastase detection from an oralrinse for the development of a simple periodontal disease screening test (Uitto et al. 1992). Theyfound that elastase activity was markedly elevated in periodontitis and the activity decreased closeto the normal levels following periodontal therapy. As a screening method for untreated patients, a69strong correlation was found between periodontitis and positive elastase test results. Similarly,salivary collagenase was significantly higher in periodontitis patients before than after treatment(Uitto et al. 1990).It is anticipated that the enzymes present in the cytoplasm of dying cells would serve asmarkers of tissue destruction. Aspartate aminotransferase (AST) is such an intracellular enzymethat is released from impaired or dying cells. It is seen in increased amounts in the serumfollowing myocardial infarction and during active hepatitis. Studies on AST levels in GCF werefirst performed in beagle dogs (Chambers et al. 1984). Recently this area of research has receivedconsiderable attention (Page et al. 1975; Persson et al. 1990a,b; Persson and Page 1990; Chamberset al. 1991; Cohen et al. 1991; Persson and Page 1992).AST levels have been evaluated in gingivitis and periodontitis models. In human experimentalgingivitis a statistically significant association was demonstrated between AST levels and gingivalindex scores for both developing lesions and resolving gingivitis (Persson et al. 1990b). In a 26week study of AST in a ligature-induced periodontitis in beagle dogs, a correlation was foundbetween elevated concentration of the enzyme and microscopic evidence of disease activity (Cohenet al. 1991).Persson and co-workers assessed the relationship between GCF AST levels and active tissuedestruction in treated chronic periodontitis patients (Persson et al. 1990a). They examined 25patients at 3 month intervals over a 2 year period, and determined that the maximum enzyme levelwas significantly elevated at sites with confirmed disease activity and attachment loss. AST levelshad median values approximately 600-800 IU at disease-active sites, compared to 400-500 IU atdisease-inactive sites. These results supported the claim that an objective diagnostic test basedupon AST levels could distinguish between disease-active and disease-inactive sites.In a recent longitudinal study by Chambers and co-workers, AST levels were monitored in 31patients with untreated adult periodontitis (Chambers et a. 1991). They found that only 2.6% (40of the 1536 periodontal sites) lost 2 mm or more of attachment over the two year study. Although70strong correlations were found between AST levels and degenerated sites, associations were alsoobserved with sites that maintained the same attachment level and sites that exhibited gingivitis.Further evaluations were performed by Persson and co-workers to ascertain the diagnosticcharacteristics of crevicular fluid AST levels associated with periodontal disease activity (Perssonand Page 1992). The ability of crevicular fluid AST activities at 600, 800, 1000, and 1200 1111Jlevels to recognize active disease was investigated. Eight of the 25 subjects investigateddemonstrated 1 or 2 sites that lost 2mm of attachment during a 2 year period. The AST 800AIUdemonstrated a sensitivity of 0.93 and specificity 0.68 for attachment loss 2mm. AST 8001.LIUwas the most suitable cutoff point to distinguish sites at risk for future attachment loss from thosethat were unlikely to progress further. These results appear to be the most significant finding increvicular fluid research in the past few years.1. 7 . 4 Immunological and inflammatory indicatorsImmunological products and inflammatory mediators are produced by affected periodontaltissues. Since these substances play important roles affecting the pathogenesis of the disease, thisled investigators to identify and relate them to the progression of periodontal disease.Interleukin- la and 13 are significant mediators of the destructive process of periodontal disease.They are produced by a number of cell types such as mononuclear phagocytes, PMN leukocytesand endothelial cells. Their effects include enhanced bone resorption and the production ofprostaglandin E2 and collagenase.Investigations of IL-1f3 in gingival tissue in relation to periodontal disease have recently beenperformed. In one study comparing chronic adult periodontitis patients to healthy control subjects,IL-113 was consistently recovered from GCF of disease affected subjects whereas no IL-1f3 couldbe found in normal gingival tissue (Honig et al. 1989). Although measureable attachment loss wasassociated with the presence of IL-113, it was not statistically significant. In another study usingimmunofluorescent staining techniques, there were almost 3-fold more IL-113 staining cells inperiodontally diseased tissue than in normal tissue (Jandinski et al. 1991).71IL-1 levels have also been measured in crevicular fluid. In a cross sectional study of GCFfrom adults with previous destructive periodontitis, measureable amounts of IL-i(3 were found at58% of the examined sites which displayed a wide range of values (Wilton et al. 1992). Since nostatistically significant correlations were found with measured clinical parameters, IL-10 could notbe related to previous evidence of destructive periodontal disease. This lack of correlation withmeasurements like pocket depth is not surprising since pocket depth reflects the cummulativehistory of periodontal involvement which does not necessarily reflect the current disease activity.Another study which examined IL- la and IL-1[3 in disease active sites of untreatedperiodontitis patients found that 90% of GCF samples from tested sites contained measureablelevels of IL-1, with more frequent occurrence of IL-113 (Masada et al. 1990). In addition, markedreductions in total IL-1 levels were observed after treatment. Furthermore, in a subset of thesepatients, 1L-1 messenger RNA was detected in all sampled gingival tissue, demonstrating that IL-1is produced and released locally. Although GCF IL-1 was recovered in a greater number ofcrevicular sites than in previous study, its presence was related only to a cross-sectional judgementof sites manifesting characteristic features of disease presence. A longitudinal study is required tocompare levels of IL-1 in GCF with accurate measures of disease activity such as changes inattachment level or reductions in alveolar bone height. Only then would it be possible to assess theutility of IL-1 to serve as an indicator of active disease.Immunoglobulins have been evaluated in GCF for class and subclass distributions and assayedas biologically active proteins against specific substrates. A recent investigaion by Reinhardt andco-workers examined IgG subclasses in GCF from active versus stable periodontal sites(Reinhardt et al. 1989). Significantly higher mean IgG1 and IgG4 concentrations were found inGCF at sites that exhibited 2mm of attachment loss than sites that were stable. Theserelationships remained significant when IgG subclass concentrations were adjusted per mgalbumin. In addition, mean adjusted IgG subclass concentrations in GCF were generally higherthan in serum, especially for IgG4 which displayed an active site GCF : serum ratio of 24.2: 1.72Thus IgG4 concentrations may be useful as a disease indicator, as shifts towards higher IgG4levels indicate immunopathological changes related to periodontitis.Autoimmunity to collagen has been evaluated in patients with periodontal disease. An earlystudy by Mammo and co-workers demonstrated an enhanced cellular immune response to nativeand denatured homologous type I collagen in patients with periodontal disease (Mammo et al.1982). Subsequently this same laboratory demonstrated that serum levels of antibody to type Icollagen was significantly higher in periodontally affected subjects than in the controls (Ftis et al.1986). This was followed by a study on 20 patients of antibody to type I collagen in GCF frominflamed sites before and 6 weeks after treatment (Refaie et al. 1990). It was found that IgGantibody levels to collagen in GCF were significantly higher than in control sera, but that theselevels were not significantly different from those in autologous sera. Furthermore, the levels ofIgG antibody in GCF and autologous sera did not change significantly after periodontal treatment.This result is not surprising due to the 10 to 15 minute GCF sampling time employed in this study.Prolonged GCF sampling time greater than 10 minutes has been shown to dilute the GCF withserum proteins (Curtis et al. 1988). Thus in the above study, GCF and autologous sera initiallymeant to comprise two distinct sources, would be expected to possess similiar compositions.In a recent study of autoimmunity to collagen in adult periodontal disease, immunoglobulinclasses in sera and gingival tissue extracts were examined in patients with chronic adultperiodontitis (Anusaksathien et al. 1992). While IgG and IgA antibodies to type I collagen werepresent in higher concentration in tissue extracts than in autologous serum, no significantdifferences were found for IgM antibodies. These results support the concept that theperiodontium is a major site for production of collagen antibodies. In addition, their findingssuggest a class switch of IgM to IgG in inflamed tissue, which may be the result of prolongedantigenic stimulation.Although antibodies to collagen were found in periodontally affected individuals, they wereonly associated with the presence of periodontal disease. In regards to their use as an indicator of73periodontal disease, analysis for antibodies to collagen is another example of elaborate testing todetermine what is already known by prior clinical evaluation.In addition to IL-1f3, another cytokine, tumour necrosis factor alpha (TNF-a), which isproduced by activated monocytes and leukocytes and found in GCF, was examined as a possibleindicator of periodontal disease (Rossomando et al. 1990). Although TNF-a levels were foundconsistently well above serum levels, the amounts recovered from individual sites varied widelyand when compared to clinical recordings of periodontal disease status, no significant differenceswere found. In this cross-sectional study TNF-a could not predict periodontal health or disease.However, its presence in some pockets that were 3mm and absence in other pockets that were4mm requires further investigation. It is possible that in a longitudinal study employing clinicalattachment loss measurements to identify disease activity, TNF-a may be an indicator for earlystages of disease activity.Prostaglandin E2 (PGE2), a potent mediator of inflammation produced via the cyclooxygenasepathway from arachidonic acid, has been detected in GCF of experimentally induced periodontitisin the dog (Tubb et al. 1990). Studies have been made on the relationship of PGE2 and 6KPGFlato cAMP, IgG, IgM and a-2macroglobulin in GCF in adult periodontitis subjects (Sengupta et al.1990). Twenty one days after scaling and root planing, the levels of all factors significantlydecreased except for 6KPGFi a and cAMP which were essentially unchanged. There was asignificant correlation between PGE2, 6KPGFia, and cAMP before but not after treatment. Theauthors concluded that this correlation pattern indicates the involvement of PGE2, 6KPGFi a, andcAMP in inflammation of the periodontium.The most significant studies on PGE2 in GCF were performed by Offenbacher and co-workers(Offenbacher et al. 1986; Offenbacher et al. 1989). In cases of untreated periodontitis patients,GCF PGE2 was found to substantially increase during active phases of periodontal destruction. Asubsequent investigation in experimental ligature induced periodontitis in monkeys demonstratedthat PGE2 level increases at three months correlated with attachment loss and bone loss. In thismodel, PG levels peaked at 6 months and returned to baseline by 12 months. This work confirms74the earlier findings in the initial investigation that PGE2 has value in diagnosing the disease activestate.1.7.5 By-products of metabolismByproducts of metabolism of the periodontium include cellular and extracellular tissueelements. One of the main features of active periodontitis is the destruction of the extracellularmatrix. It is plausible to propose that crevicular fluid would contain byproducts of tissuedestruction, and hence evidence of breakdown.By monitoring crevicular fluid for breakdown products of connective tissue, it is possible tofollow destruction of supporting tissue during active periodontal disease. In particular, Hyp, aunique amino acid which constitutes 10-12% of the collagen molecule and is not found inappreciable amounts in other tissue fluids, has been isolated in human crevicular fluid (Miller et al.1982). It is anticipated that during episodes of active periodontal disease, it should be possible tofollow collagen breakdown by measuring Hyp levels in GCF.In our laboratory a highly sensitive and specific method has been developed for determinationsof Pro and Hyp in biological systems using high performance liquid chromotography (Yaegaki etal. 1986). In particular, the method has been modified for evaluation of Hyp in crevicular fluid.In our earlier clinical study on 30 patients, we observed that Hyp levels in GCF were increased inperiodontally involved sites, and that Hyp existed mainly in a peptide form (Coil and Tonzetich,1986). Subsequently, it was determined that Hyp levels did not correlate with pocket depth orcrevicular fluid flow. Emerging results at the time indicated that Hyp levels were associated withthe level of epithelial attachment (Coil et al. 1987). It appeared that changes in Hyp levels precededchanges in attachment level by 2 to 4 weeks. Thus, Hyp could be promising as an indicator ofactive disease state.Most of the Hyp in GCF may be derived either from fragments of tissue collagen or serumClq. Clq, a subcomponent of the first complement protein, has a molecular weight ofapproximately 410,000 and contains 4.3% Hyp by weight. Svanberg used 0.02 M sodium acetate75to remove Clq from GCF samples (Svanberg 1987a,b). We have similarly removed Clqsuccessfully by precipitating it with 0.5M acetate solutions, followed by centrifugation, whicheffectively removed 5 times the serum concentration of Clq added to crevicular fluid samples (Coiland Tonzetich 1988). Thus, it appears from these results that the main source of Hyp followingthis procedure is degraded collagen. However, it is still possible that Clq could be degraded bymicroorganisms to fragments that are not eliminated by the employed precipitation procedure.Analysis of Hyp was performed in GCF and serum in three beagle dogs during a 5 weekcourse in experimental gingivitis (Svanberg 1987a). Hyp concentration from individual sitesshowed an irregular pattern of high and low values during the 5 week study indicating that collagenmetabolism was not a linearly continuous process. In a further study Svanberg examined Hypcontent of GCF using a 9 day ligature-induced periodontitis beagle dog model (Svanberg 1987b).Collagen derived Hyp (total GCF minus serum value) was maximal 4 days after removal ofligature. No indication was given as to the extent of tissue destruction at this examination point.Although it appeared that Hyp was temporally related to inflammation, further investigations arerequired to determine the accuracy of this amino acid as an indicator of disease.A number of GAG's have been detected from periodontal sites exhibiting previous destruction(Embery et al. 1982). Last and co-workers further investigated GAG's in GCF (Last et al. 1985).At sites of chronic gingivitis, hyaluronic acid was the only major component detected.Chondroitin-4-sulfate was also detected at periodontal sites that had 7mm pocket depth andradiographic evidence of bone loss, moderate pockets with some detectable alveolar bone crestloss, and at sites of LW involvement. Since chondroitin-4-sulfate is a prominent component ofbone, these results indicate that its detection in GCF may be a sensitive indicator of bonedestruction.Other potential markers of bone destruction have also been examined. In a preliminaryinvestigation, Bowers and co-workers (Bowers et al. 1989) used a dot blot assay to examine GCFfor the presence of osteonectin, bone phosphoprotein, and bone sialoproteins I and II. Whileneither of the bone sialoproteins were detected, osteonectin and bone phosphoprotein were found76in quantities that roughly correlated with the degree of periodontal involvement. This is the firstknown published work to examine GCF for this group of connective tissue-associated proteins.Evaluations of protein composition of GCF may also prove useful in identifying a marker ormarkers of periodontitis. Baseline data of a longitudinal study on the protein composition of GCFfrom subjects without periodontitis indicated that four non-plasma derived proteins are routinelydetected in the fluid (Curtis et al. 1990). It is believed that these represent products of normalturnover in the periodontium. It is anticipated the major metabolic changes that accompanydestruction of periodontal tissues will be detectable in the GCF protein profile.Elastin is a fibrous protein component of connective tissues, especially abundant in lungs,ligaments and arterial walls. Its degradation has been associated with degenerative disorders suchas emphysema and arteriosclerosis. Desmosine (DES) and isodesmosine (IDE) are specific cross-linking amino acids found in elastin, and their presence in the tissue is indicative of elastincatabolism.Although prominent elastin fibers are present in the periodontal ligaments of rabbit, dog, sheepand swine, they are found in the gingiva but not the periodontal ligament of man (Berkovitz et al.1982). Oxytalan fibers are found in the periodontal ligament of man, having a similar distributionto elastin fibers in the above named animals. Although it is impossible to distinguish between themhistochemically, these two fibers types are thought to be phylogenetically related.HPLC methods for the detection of the specific cross linking amino acids, DES and IDE, havebeen performed on cell cultures (Muramoto et al. 1984) and tissue hydrosylates (Yamaguchi et al.1987). Application of HPLC methods for analysis of crevicular fluid, will make it possible todetect DES and IDE at the nanogram level if they are present. This has not been attempted to date.771.8 PROPOSED STUDY1.8.1 Crevicular volatile sulphur productionIn view of the apparent importance of VSC production in periodontal pockets (Coil andTonzetich 1984; Coil and Tonzetich 1985), one of the purposes of this study was to develop adevice for collection of volatiles from the gingival crevice, and to analyze and compare thecomposition of VSC in gingival sulci with those of mouth air. The study is intended to quantifythe level of VSC production in periodontal crevice sites and relate it to clinical measurements ofperiodontal involvement. It is plausible that measurement of volatiles emanating from periodontalsites may reflect the level of disease activity occurring at the time of sampling.1.8.2 Crevicular fluid hydroxyproline content as an indicator of disease activityThe only known study to quantitate hydroxyproline levels in GCF was work done by Millerand co-workers (Miller et al. 1982). Using pre-column derivatization followed by HPLCseparation, Hyp was identified in GCF samples of three subjects. Since it has been demonstratedthat active stages of periodontitis are accompanied by significant loss of collagen (Page andSchroeder 1976), it is conjectured that Hyp levels in GCF may reflect the level of collagenmetabolism in the periodontium. By monitoring gingival fluid levels of Hyp, it may be possible tomonitor the level of collagen breakdown occurring at specific periodontal sites. By comparingHyp levels to clinically accepted periodontal assessments at specific periodontal sites, it would bepossible to evaluate Hyp as an indicator for presence of periodontal disease. Attachment levelmeasurements in a longitudinal study could also assess the ability of Hyp analyses to indicatedisease activity.781.9 SIGNIFICANCEIt is no longer appropriate to attempt to identify active periodontal disease using traditionalclinical measurements because they only give retrospective evidence of periodontal disease.Pathology recorded by gingival indexes, bleeding on probing, suppuration and pocket depth, areinadequate measurements for predicting future changes in the level of periodontal attachment(Haffajee et al. 1983; Kaldahl et al. 1990). There is a strong need for a non-invasive, sensitive andspecific test for periodontal disease activity.In considering such a test, one must first accept the concept that periodontitis in man is anepisodic disorder, occurring in short bursts followed by long periods of quiescence. Without thisperception of the disease activity, the traditional clinical measures would be sufficient to designatethe presence of disease if one accepts a linear model for progression of disease. Secondly, inrecent years it is generally acknowledged that periodontitis, specifically periodontal diseaseactivity, may be a site specific phenomena (Goodson et al. 1982; Haffajee et al. 1983).The assessment of periodontal diseases remains controversial because of a lack of reliablediagnostic procedures which can differentiate previous periodontal disease from currentlyprogressing or active periodontal destruction. Evaluations of disease progression have focusedupon loss of epithelial attachment and loss of alveolar bone, which act only as retrospectiveanalyses of disease activity. It is anticipated that correlation of Hyp levels in crevicular fluid withattachment level change will establish Hyp as a reliable predictor of disease activity.792 . MATERIALS AND METHODS2.1 MATERIALSUnless stated otherwise all reagents used in this study were of analytical grade and purchasedfrom either Fisher Scientific Limited (Fairlawn, NJ, USA) or Sigma Chemical Company (St.Louis, MO, USA). All materials used for the High Performance Liquid Chromatography (HPLC)were purchased from Waters (Millipore Corp., Milford MA, USA). Acetonitrile, ethanol andmethanol were HPLC chemical grade and obtained from BDH (BDH Chemicals, Vancouver,B .C. ,C anada).Materials and chemicals used for the electrophoresis and western blotting were purchased fromBio-Rad (Bio-Rad Laboratories, Richmond, CA, USA) and ELISA plates were from Falcon(Becton Dickinson and Co., Lincoln Park, N.J.,USA)Antibodies were obtained from a number of sources. Anti-type I collagen was bought fromChemicon (Temecula, CA, USA) and Southern Biotechnology Associates Inc. (Birmingham, AL,USA), and donated by Dr. Sampath Narayanan (Seattle, WA, USA). Hybridoma cell linesproducing monclonal antibodies to Clq were graciously donated by Dr. Linda Curtiss (ScrippsClinic, La Jolla, CA, USA). Polyclonal antibody to Clq was purchased from Calbiochem(Calbiochem Corp. La Jolla, CA, USA). Secondary antibodies, which included alkalinephosphatase conjugated goat anti-rabbit IgG and alkaline phosphatase conjugated goat anti-mouseIgG, were purchased from Calbiochem and BRL (Bethesda Research Laboratories, Gaithersburg,MD, USA), respectively. All other antibody products were purchased from Sigma.802.2 VOLATILE SULPHUR COMPOUND STUDY2.2.1 Development of a collection device for samplingcrevicular airThe concepts surrounding the development of a suitable gingival crevice air collection devicefocused upon several ideal parameters. An ideal collection device should be easily fabricated,custom fit for each patient's crevicular site, provide a tight seal, be reusable, and be non-invasiveto the tissues.The initial approach was to fabricate a collection device using a common silicone based dentalimpression material. This was accomplished by the following procedures. First, using alginateimpression material (JeltrateTM, L.D. Caulk Co., Miford, DE, USA), an impression was obtainedof the patient's dental arch using a standard metal tray. Subsequently a dental cast, a replica of thepatients' teeth, was made by pouring this impression in a die stone, (DensiteTM, Georgia-Pacific,Portlan, OR, USA). From this cast a segmental custom tray from acrylic was made thatoverlapped the crowns and gingival margins on several teeth adjacent to and including the test tooth(FAS TRAYTm, Harry J. Bosworth Co., Chicago, IL, USA). A section of micro polyethylenetubing (INTRAMEDICTM, Becton, Dickinson and Company, Parsippany, N.J., U.S.A.),measuring 0.76 mm I.D. x 1.22 mm 0.D., was held against the buccal surface of the tooth at thelevel of the free gingival margin using dental floss interproximally. A medium body silcone basedimpression material (ReprosilTm, L.D. Caulk Co, Miford, DE, USA) was mixed and placed in thesegmental tray, which in turn was placed in the patient's mouth and pressed over the tooth to besampled. To ensure that the tubing was completely surrounded by the impression material, it waspositioned and held so that it contacted the edge of the segmental tray while the impression materialwas setting. The setting took five minutes during which time this newly forming sampling devicewas held stable in the patient's mouth.After the impression had set, the sampling device was removed from the patient's mouth andthe dental floss was removed from around the tubing. The portion of the tubing which protruded81from the inside of the impression material and rested against the free gingival margin was cut anddiscarded. This created a space above the free gingival margin from which crevicular air could besampled, leaving two pieces of tubing protruding from the impression which served as inlet andoutlet sampling ports. A peristaltic pump was attached to the micro polyvinyl tubing exiting on thesame proximal surface of the gingival crevice to be sampled. The pump set at a flow rate of 0.6ml/min. directed a flow of air across the gingival site and out through the micro tube at the oppositeproximal surface.Because it is impossible to obtain sufficiently large crevicular air samples for direct gaschromatographic analyses, it was necessary to pre-trap and concentrate the volatiles emanatingfrom the gingival sites. The same methodology for collection and concentration of volatile organiccompounds from human mouth air was adapted for use in collecting volatiles from crevicular air(Tonzetich et al. 1991). A Tenax-GC trap, which was used to collect volatile compounds, wasplaced between a vacuum pump and the subject's oral cavity. Tenax traps were constructed from 7mm OD x 75 mm lengths of aluminum tubing, coated on the inside with teflon, packed with 0.7 gof Tenax-GC, and stoppered with glass wool end plugs. A total of six traps were used in thestudy.Each Tenax trap was tested individually for absorption and desorptive capabilities. Theretention times of Tenax-trapped sulphide components from mouth air were matched with thoseobtained by direct mouth air sampling. Room air was also analyzed in a similar manner to ensurethat there was no sulphur contribution from the ambient air. Prior to use, a Tenax trap that wasstored at -20°C in a stoppered test tube was removed and immediately desorbed on a modifiedBendix Flasher at 120°C. It remained at room temperature for no longer than one hour before itwas cooled and maintained in dry ice at -55°C while being utilized as a collection device.Pilot runs designed to test the recovery of volatiles from crevicular air were performed toascertain the validity of the sampling technique. After collecting the equivalent of 10 ml ofcrevicular air from several subjects, each Tenax trap was desorbed on the modified Bendix Flasherunit coupled to the gas chromatographic/ flame photometric detector (GC/FPD). As each Tenax82trap had been tested for its ability to retain and desorb volatile sulphur compounds prior to use andas none of the spectrums in these pilot runs revealed the presence of volatile sulphur compounds,this implied that there was a defect in construction or operation of the sampling device. Inaddition, since each patient exhibited VSC in their mouth air at the same appointment as the pilotruns for crevicular air sampling were performed, it was anticipated that there was a goodpossibility of detecting VSC in crevicular air samples.After contemplating the possible causes for failure using this technique, it was decided tochange the position of the peristaltic pump. Instead of pumping air across the gingival crevice andforcing it through tubing at the opposite proximal site, it was decided to aspire air from above thecrevice. When the Tenax trap was placed between the gingival site and the peristaltic pump,measureable amounts of VSC were collected as evidenced by the recordings on the GCchromatograms.The sampling device was further modified to allow for easier fabrication. Instead of producinga dental cast to construct a special custom segmental tray to take an impression, a soft acrylicdenture reline material, Coe-softTM (Coe Laboratories Inc., Chicago, IL, 60658) was used to takean impression without using a tray by placing it directly over the tested crevicular site. Thismodification yielded the same volatile compound collection capabilities and allowed for easierchairside fabrication.2.2.2 Experimental designTwenty gingival crevicular sites, thirteen test and seven control, were selected from seventeenpatients ( 12 male and 5 female, aged 24-67 yrs). Patients were selected on the basis of not havingreceived previous periodontal treatment or antibiotic medication within the past six months andhaving no known systemic disease. Pocket depths were measured using a Marquis probe andrecorded to the nearest millimeter by one competent examiner. Periodontal sites were separatedinto control and test groups based upon pocket depth measurements of 53 mm and z4 mm,respectively. Three patients that exhibited both shallow and deep crevices, were further classified83on the basis of bleeding on probing as having an inflamed or noninflamed site. Five patientscontributed three and two sites to the inflamed and noninflamed groups, respectively. Clinicalsigns of inflammation based upon bleeding on probing evaluations were performed by the sameexaminer.Patients were tested at least one week after periodontal probing and their periodontal status wasconfirmed immediately after VSC collection. The GC/FPD method developed by Tonzetich wasemployed for sampling mouth air (Tonzetich 1971). Patients were instructed to close their lips andbreathe through their nose. After one minute, a 6 cm length of a 15 cm long piece of teflon tubing(0.3 mm ID.) connected to the GC sampling line assembly was inserted between the lips into theoral cavity. The subject then voluntarily stopped breathing while 15 ml of mouth air was aspiredby means of a 25 ml Hamilton syringe, thus filling the 10 ml sample loop of the GC. The carriergas was then directed through the sample loop and the 10 ml sample transferred onto the GCcolumn for separation.2.2.3 Collection of crevicular airCollection of volatiles from gingival sulci was effected by sampling air from individual isolatedperiodontal sites. A small section of micro polyethylene tubing was held, as previously described,against the buccal surface of the tooth at the level of the gingival margin using dental flossinterproximally. Coe-Soft was mixed and molded into a cube before it was fitted over the toothand its adjacent tubing. After the material hardened, the impression was removed from the mouthand epoxy resin was used to secure the tubing as it exited the impression to the external portion ofthe acrylic. This created a closed system. Finally, the portion of tubing at the gingival margin wasremoved thereby creating inlet and outlet ports for sampling. Each custom sampling device had a60 cm outlet port of tubing extend from the crevicular site to a Tenax trap, while a 15 cm inlet portof tubing extended from the collection device to ambient air.These special devices for collection of volatiles were constructed for each site of eachindividual. They were prepared after the subjects provided mouth air samples for GC analysis.84The collection device was placed back in the mouth and the exit port tubing was connected viatygon tubing to a Tenax trap cooled to -55°C in dry ice. The other end of the trap was connected toa peristaltic pump and crevicular air was removed at a rate of 0.60 ml/min. for sixteen minutes.Only one device was required per site which allowed for repetitive sampling. Immediately after thecollection of volatiles, the tenax trap was removed from the dry ice chamber and quickly detachedfrom the tubing and prepared for GC analysis.2.2.4 Gas chromatographic analysis of crevicular airImmediately after collection, the compounds concentrated on Tenax traps were desorbed at120°C using a modified Bendix Flasher coupled to the GC unit and analyzed by GC. Theseparations of volatile sulphur compounds were performed on a Tracor 550 Gas Chromatographusing a 7.315 x 3.2 mm fluorinated ethylene propylene (FEP) Teflon column packed with 5 percent polyphenyl ether and 0.05 per cent phosphoric acid on 30-40 mesh Teflon (Micro TekInstruments Corp., Austin, Texas). The temperature of the column and detector were 70°C and150°C respectively. The pressures and flow rates of the employed high purity gases (UnionCarbide Canada Ltd.) were : air carrier gas (55 psig)- 10 cc/min; H2 (40 psig) - 80 cc/min; air (55psig) - 60 cc/min. Sulphide profiles were recorded using a Shimadzu, C-R3A Chromatopacapparatus (Kyoto, Japan). Permeation tube standards were used to internally program calculationsof sulphide content from recorded peak areas.2.2.5 Identification and quantitation of volatile sulphidesThe identification of sulphide components was based upon their retention times. Thechromatographic pattern exhibited sulphide peaks and retention times (Rt) as follows: hydrogensulphide, H2S (Rt:115-130 sec.), methyl mercaptan, CH3SH (Rt: 190-205 sec.), dimethylsulphide, (CH3)2S (Rt: 275-310 sec.), dimethyl disulfide, (CH3S)2 (Rt: 390-425 sec.). Standardplots of integrated peak area versus concentration for H2S and CH3SH are illustrated in figures2.1 and 2.2 respectively. For sulphide peaks that were below the threshold quantitation level85setting of the recorder (100 mV/sec), extrapolation calculations were performed using peak weightsof known peak areas and corresponding ng concentrations. For H2S the relationship of ng to peakarea was (y= .1266 + 1.556 exp-04 x - 1.214 exp-08 x 2 + 5.167 exp-13x3 ; r2=1.000), while forCH3SH the ng to peak area was (y= .2223 + 1.189 exp-04 x - 3.634 exp-09 x 2 + 5.43 exp-14 x3 ;r2=1.000). The peak weight to ng relationships for (CH3)2S and (CH3S)2 were based upon asimiliar response factor and the % weight of sulphur of each compound compared to CH3SH(52/67 and 68/67 respectively). Depending upon the amount of sampling time for each crevicularsite, the total volume of trapped crevicular air was calculated and the levels of recorded sulphideswere reported as ng/10 ml of crevicular air.Sulfide concentration [ng/10m1]Figure 2.1:^Standard plot for H2S of integrated peak area versus concentration8610 6 .^••^•^•1 0Sulfide concentration [ng/ml]Figure 2.2:^Standard plot for CH3SH of integrated peak area versus concentration2.2.6 Data analysisThe levels of crevicular VSC were converted to ng levels of sulphide per 10 ml of air, basedupon flow rate and time of sample collection. This enabled comparisons of crevicular and mouthair sulphides to be performed on equivalent volumes of air. The various sulphide levels, the ratiosof CH3SH to H2S and all methyl sulphides to H2S, and total sulphur content in crevicular air werecategorized according to pocket depth (shallow and deep) and inflammation (noninflamed and10 387inflamed). Comparisons based on student's t-test were made between mean values of sulphideratios and total sulphur levels in crevicular air.2.3 DEVELOPMENT OF HPLC HYDROXYPROLINE ANALYSISThe objective was to develop and apply a sensitive, specific, and reproducible chromatographicprocedure that would detect Hyp in crevicular fluid. In previous investigations Hyp content ofgingival exudate was determined by colorimetric methods (Hara and Takahashi 1975). Recentlythe analysis of Hyp in GCF by an HPLC method was described by Miller and co-workers (Milleret al. 1982). A logical place to begin the current study was to evaluate the applicability of the latterHPLC method.Miller and co-workers analyzed Hyp using precolumn derivatization with dansyl chloride (5-dimethylaminoaphthahalene-1-sulfonyl chloride) (Miller et al. 1982). Briefly, the procedureinvolved acid hydrolysis followed by incubation of the sample with dansyl chloride reagent for onehour at 55 °C. After drying, reconstitution in methanol, and filtration, the samples were injectedonto a g-bondpack C18 column. The components were eluted at a flow rate of 1.5 ml/min. using0.1 M sodium acetate/acetonitrile buffer system, programmed with acetate/acetonitrile gradients as95/5 constant for 10 minutes, increased to 80/20 during 15 minutes, followed by another gradientof 06/40 during 75 minutes. Fluorescence detection was effected using an excitation wavelengthof 405 nm with a 485 nm emission cut off filter.Our HPLC unit (Waters Assoc., Millipore Corp., Mississauga, ONT, Canada) consisted of aModel 721 programmable system controller, two Model 510 HPLC pumps, a Model 710 WISP, AModel 730 data module and a Model 481 LC UV variable-wavelength spectrophotometer. Thesystem was coupled to a 10 cm x 8 mm (5 gm) reversed-phase Nova-Pak C18 Radial-Pak column(Waters Assoc.) used with a Z-module radial compression separation system (Waters Assoc.) andwhen required a Gilson Model 201 fraction collector (Gilson, France).88In using the dansyl chloride derivatization method, it was found that the standard Hyp peakwas small, gave low responses, and variably appeared at retention times beyond 30 minutes ofelution. In addition, location of the Hyp peak was at times uncertain, especially in test systems ofcrevicular fluid. Due to the above limitations, spiking of the sample with relatively large amountsof Hyp was required to identify the Hyp peak. Other investigators have also demonstrated thatusing this method Hyp is difficult to detect, as it has a low responsiveness and is very close to aglutamine peak (Wiedmeier et al. 1982).Concurrent with the investigation for an appropriate Hyp analytic procedure, an efficienttechnique for analysis of Hyp and Proline (Pro) was developed in our laboratory by Yaegaki andco-workers (Yaegaki et al. 1986). A number of techniques were evaluated for determining Hypand Pro in biological systems including derivatization with dabsyl and dansyl chlorides, and NBD-Cl (4-chloro-7-nitrobenzolfurazan). These methods were found complicated and limited inapplicability due to poor separation, low recovery, low sensitivity, and complexity of analyses.For example, using radiolabelled Hyp and Pro, it was determined that derivatizing with dansylchloride resulted in less than 50% recovery of the amino acid. As Hyp is found in extremely lowconcentrations in biological materials, its chromatographic peak in complex samples wasovershadowed in previous Hyp derivatization analyses by large numbers of other constituents.The method developed by Yaegaki and co-workers was based on the removal first of primaryamino acids from the sample followed by analysis for secondary amino acids Hyp and Pro.Primary amino acids are removed by pre-column derivatization with o-phthalaldehyde (OPA) andseparation by reversed phase HPLC. Subsequent pre-column derivatization of the void volumewith phenylisothiocyanante (PITC) and reversed phase HPLC separates derivatized Hyp and Prowhich are quantified using a UV detection system. Using this new method Pro and Hyp werereadily analyzed without interference by other constituents.This method was successfully used in evaluating collagen metabolism by human gingivalfibroblasts cultured under different experimental conditions. The prior removal of primary amino89acids from the sample allows a clean chromatographic separation of Hyp and Pro, and hence, moreaccurate results. The procedure yields 93% recovery for both Hyp and Pro.In crevicular fluid systems the volume of collected fluid is in the submicroliter range. Due tothis volume limitation, it was necessary to reduce the manipulation of the sample steps to theminimum. For this reason the utilization of a single derivatizating step employing PITC, whichreacts with both primary and secondary amino acids, was evaluated.Using the above described improved modified HPLC method, a Pierce amino acid standard(Sigma), supplemented with 5 x 10 -3 M 4-Hyp, was processed for the PITC derivatizationaccording to the method outlined by Yaegaki and co-workers (Yaegaki et al. 1986). From thisstandard amino acid mixture the Hyp peak was cleanly separated from neighbouring Glu and Serpeaks. This was confirmed by spiking the mixture with radiolabelled Hyp (3H-Hyp) with specificactivity of 5.9 Ci/mmol (DuPont, Boston, MA, USA), and determination of radioactivity incollected peak fractions. Figure 2.3 displays the HPLC profile of the derivatized amino acidstandard exhibiting a peak corresponding to Hyp with a retention time of 3 min 20 seconds. Inpilot tests of crevicular fluid samples, the Hyp consistently remained separated from the Glu andSer peaks. This is demonstrated by a crevicular fluid profile shown in figure 2.4. Hence twoderivatization methods have been developed in our laboratory for the analysis of hydroxyproline(Hyp) and proline (Pro) in biological materials (Yaegaki et al. 1986). One of these, the singlephenylisothiocyanate (PITC) derivatization method was most suitable as it yielded cleanseparations of Hyp and other amino acids in crevicular fluid, without the need for initialderivatization with o- phthalaldehyde to remove primary amino acids. However, if one requiresaccurate analysis of both Pro and Hyp then Yaegaki's two derivatization technique is a method ofchoice.0.1 2^5^78910 13^169012114i14i63Hyp^_iI^ I10^120 5Elution time (min)Figure 2.3: HPLC profile of PITC derivatized amino acid mixture. The black outline within thepeak corresponding to Hyp indicates the location of the radioactivity. '1 1= Asp; '2'= Glu; '3'=Hyp; '4'=Ser; '5'=Gly; '6'=His; '7'= Arg; '8'=Thr; '9'= Ala; 1 10'=Pro; '11'=Tyr; '12'=Val;'13'=Met; '14'= Cys; '15'=Ile; '16'=Leu.0^A AI. v,.1J1, J J V I,T I10Hyp11 ‘ J10.111005^ 12Elution time (min)Figure 2.4: HPLC profile of PITC derivatized crevicular fluid. The black outline within thepeak corresponding to Hyp indicates the position of the radioactivity.91922.4 SPIRAMYCIN STUDY2.4.1 Experiment designThis study was performed on thirty patients selected from the Graduate Periodontal Clinic atthe University of British Columbia. The objective of the study was to test the effect ofantimicrobial treatment with spiramycin as an adjunct to scaling and root planing. Exclusioncriteria for patient selection included the following: 1) the presence of juvenile periodontitis, 2) ahistory of periodontal therapy within the past six months, 3) the use of antibiotics in the last sixmonths, 4) a history of diabetes, and 5) pregnancy or lactation. Each patient had a minimum oftwo teeth that were periodontally involved, each having at least one interproximal site of ?_7 mmpocket depth and interproximal contact with a natural tooth. If several teeth were found suitable,then the most posterior tooth in each arch was chosen, as they are the most difficult to maintain.For each tooth, four sites (distal buccal, mesial buccal, distal lingual and mesial lingual) wereclinically evaluated using the following parameters: 1) plaque index, 2) crevicular fluid flow, 3)probing depth, 4) attachment level and 5) bleeding on probing.For those patients that met the above criteria at the initial examination, alginate impressionswere taken of patients' dental arches and poured in dental stone, and an acrylic stent wasconstructed over the resulting plaster casts. The acrylic was extended to within 3 mm of thegingival margins interproximally and where possible it extended two teeth beyond the selected testsites. The acrylic opposing each examinable interproximal site was marked with a vertical line, sothat interproximal positioning of the periodontal probe would be reproducible at later examinationperiods.The patients began the study within one week after the initial examination, which involvedmeasurement and recording of the above clinical parameters at times 0, 2, 8, 12 and 24 weeksafter the onset of the study. All clinical examinations were performed by the same experiencedexaminer. After plaque index was recorded for each site, supragingival plaque was removed with93a sterile curette. Following collection of crevicular fluid on filter paper strips and measurement ofvolume using a Periotron 6000 (Harco Electronics, Irvine, CA, USA), each paper strip wassubsequently assayed for Hyp (Coil and Tonzetich 1986). This was followed by pocket depth andattachment levels measurements using a Marquis probe. It was imperative to adhere to thissequence of measurements so as not to interfere with the crevicular site prior to fluid collection.2.4.2 Sample collection and measurementEach site was isolated using cotton rolls and gently dried with air. A standardized filter paperstrip (PeriopaperTM, Harco, Tustin, CA, U.S.A.), was positioned at each interproximal site andplaced at the entrance of the crevice to collect fluid for five seconds. Crevicular fluid volume wasimmediately measured using a Periotron 6000 instrument and recorded. Each filter paper strip wasplaced in a 13 mm pyrex test tube which was sealed using parafilm wax. The relationship betweenthe Periotron 6000 reading and fluid volume was determined using 0.0 - 1.2 ptl of fetal calf serumand is depicted in figure 2.5.2.4.3 HPLC analysis for hydroxyprolineCrevicular fluid was eluted from filter paper strips by soaking them for 24 hours in 100 pi of 5mM Na2HPO4. The strips were then washed with 50 of 0.5 N acetic acid and left to soak for anadditional one hour. After the two eluants were combined, they were placed in 1.5 ml Eppendorftubes and centrifuged at 15,600 g for 30 minutes at 4° C. The supernatants were collected in 6 x 50mm pyrex micro culture tubes (Corning glass works, Corning, NY). These sample tubes werethen placed in a reaction vial, which could accomodate up to twelve samples and one Hypstandard. Further processing of the samples in the reaction vial was carrried out using the Pico-Tag Workstation (Millipore, Milford, MA).94I^I^I1200y = 9.9929 + 148.96x RA2 = 0.965a0g100 -0.0^0.2^0.4 0.6 0.8 1.0 1.2ao 0^Fluid volume [pi]Figure 2.5: A standard plot of Periotron readings versus fluid volume using fetal calf serum-soaked Periotron filter paper stripsSamples were vacuum dried to 65 mtorr before being hydrolyzed in 350 p.1 of 6 N HC1containing 0.1% phenol, which was added to the bottom of each reaction vial. Oxygen wasremoved from reaction vials by three successive vacuum evacuations interposed with nitrogenflushing. After the third evacuation the reaction vial was sealed under vacuum and placed in theworkstation oven, where the samples underwent hydrolysis at 150 °C for one hour. After coolingto room temperature the reaction vial was cleaned and dried, and each sample tube was wiped drybefore being placed back into the reaction vial and dried to 65 mtorr.95Twenty microliters of freshly prepared redrying agent (ethanol: water: triethylamine 2: 2: 1)was added to each sample and again the reaction mixture was brought to dryness. Then 50 IA offreshly made PITC derivatizing reagent (ethanol: water: phenylisothiocyanate: triethylamine 7: 1:1: 1), was added to each sample vial, vortexed for 15 seconds, placed back and sealed into thereaction vial, and left at room temperature for 20 minutes. Excess reagent was purged undervacuum at 65 mtorr, to remove all traces of PITC, which otherwise would interfere withsubsequent amino acid separation.Sample tubes were then removed from the reaction vial and processed for HPLC analysis.Samples were reconstituted in 200 ill of sample diluent (5 mM Na2HPO4 solution adjusted to pH7.4 with 10% orthophosphoric acid combined with acetonitrile, 95:5, v/v) and vortexed for 30seconds. The resulting mixtures were filtered through 0.45 12M HV SJHVOO4NS filters(Millipore, Milford, MA) using 1 cc tuberculin syringes, into individual limited volume insertssuspended within HPLC sample vials.All analyses were performed on the same HPLC unit as described in section 2.4.3. In additionto the Nova-Pak C18 column, two Pico-Tag C18 columns (Waters Assoc.), quality controlledspecifically for amino acid analyses, were also used during the study. Also, a temperature controlmonitored column heater (Waters Assoc.) replaced a water bath for maintaining a constanttemperature of 38°C.HPLC solvent buffers A and B were comprised of the following solutions. Non-organic waterwas prepared by passage of 18 MS2 pure water through a Norganic cartridge (Waters Assoc.,Mississauga, Ont.,Canada). Buffer A of the mobile phase was composed of 60 ml of acetonitrileand 940 ml of 138 mM sodium acetate buffer containing 0.05% triethylamine (TEA), adjusted topH 6.4 with glacial acetic acid Buffer B consisted of 60% acetonitrile in 40% non-organic waterby volume. The flow program for the solvent gradient of buffers A and B is given below in Table2.1. A 100 Al volume of sample was automatically injected onto the C18 stainless steel columnmaintained at 38 °C, for separation and determination of the phenylthiocarbamyl derivative of Hyp.96Each sample was separated over a period of 12 minutes followed by a gradient washing and re-equilibration for 13 minutes.Time (min) Flow (ml/min) % A % B Curve0.0 1.0 100 5 0510 1.0 54 46 0510.5 1.0 0.0 100 0511.5 1.5 0.0 100 0612.0 1.5 0.0 100 0512.5 1.5 100 0.0 0520 1.5 100 0.0 0520.5 1.0 100 0.0 0550 1.0 100 0.0 05Table 2.1: Gradient program for separation of PITC-derivatized amino acid residues.As described earlier, the retention time for Hyp was established using a radioactively labelledstandard 3H-Hyp. A standard amino acid mixture (Pierce) containing Hyp was included andprocessed with each set of 12 samples. This acted as an internal standard for each set of analyses,and monitored any fluctuations in detector response between sets of sample runs and/or separationdifferences amongst different columns employed. Standard plots for high and low levels of Hypare displayed in figures 2.6 and 2.7, respectively.2.4.4 Data analysisComparisons were made between Hyp levels and clinical measurements amongst the variousdata collection time points. Due to a variety of circumstances, such as the crevicular fluid samplesnot being available for processing, too low a collected fluid volume, chromatographic fluctuationsthat were uninterpretable, and significant HPLC instrument malfunctions over the course of the9750^100 ^150I r^ I200150y = 0.87931 + 0.62190x RA2 = 0.971100 —a)e.)eaea50 —Moles of Hyp [Exp-12]0Figure 2.6:^Standard plot for high concentrations of hydroxyproline.study, Hyp data points for all examination periods were incomplete. After an in depth analysis ofexisting data points that were common to the same patient, it was determined that the mostcomplete data was available for time points 0 and 12 weeks. Thus, Hyp analyses were comparedto clinical measurements between weeks 0 and 12 of the study. This offered a comparison ofparameters before and after periodontal treatment, a range of time that has commonly been used byother investigators (Lamster et al. 1988; Sengupta et al. 1988; Reinhardt et al. 1989; Kaldahl et al.1990; Persson et al. 1990a; Chambers et al. 1991; Deas et al. 1991; Persson and Page 1992).0.0 I^0 10^20Moles of Hyp [Exp-12]I^ I30 4098Figure 2.7:^Standard plot for low concentrations of hydroxyproline.2.5 INFLAMED AND NONINFLAMED PERIODONTAL SITES STUDY2.5.1 Experimental designThe intent of this study was to confirm that more Hyp is found in inflamed periodontal sites, aresult of the spiramycin study, and to determine the source of Hyp. During the spiramycin study aquestion arose regarding the contribution of Hyp from Clq, a subunit of the first complementcomponent, to the total HPLC measured Hyp content of crevicular fluid. It is known thatapproximately 50% of the Clq molecule has a helical collagen-like structure containing 4.3% by99weight Hyp, and that Clq is normally present in the serum. Therefore, due to the leakage ofplasma proteins through the junctional epithelium, it is likely that this protein can gain entry to thegingival crevicular fluid.The processing of the crevicular fluid in the spiramycin study included acetate precipitationfollowed by subsequent centrifugation to reduce the contribution derived from Clq. Acetateprecipitation has been reported to remove intact Clq molecules from crevicular fluid (Svanberg1987a). Thus, a study was performed to ascertain the efficiency of Clq removal from crevicularfluid by precipitating the Clq with acetate buffer. A further study was also planned to investigatethe efficiency of Clq removal by acetate in a crevicular fluid model containing type I collagen andClq, using ELISA with antibodies to collagen and Clq. Therefore, it was necessary to employ acrevicular fluid model specifically to determine the effects on type I collagen and Clq.Potential subjects were chosen for the study on the basis of criteria cited under section 2.4.1.In addition, subjects taking systemic medications that would influence gingival health such as anti-inflammatory drugs, including aspirin, within 4 weeks of the study were excluded.After clinically examining each patient without probing, gingival crevicular fluid was collectedwith Periopaper from several periodontal sites that clinically appeared inflamed or noninflamed.Each site was isolated with cotton rolls, gently air dried and supragingival plaque was removed.Periopaper was placed at the orifice of the gingival crevice for 30 seconds. GCF volume wasassessed using a Periotron 6000, and GCF samples were placed in 1.5 ml Eppendorf tubescontaining 50 121 of 0.05 M sodium acetate and immediately placed in an ice bath, before beingtransferred to storage at -70°C until analysis. Then pocket depth and bleeding on probingmeasurements of all sites were performed using a Marquis periodontal probe.As required the GCF samples that were stored at -70°C were thawed at room temperaturebefore being centrifuged at 15,600 g for 30 min. to remove the precipitated Clq (Svanberg 1987a).The resulting supernatants were allocated for analyses as follows. A 10 .t.1 aliquot was used forHPLC analysis for total Hyp present in each sample; while 25 1.11 of the original sample wasutilized for ELISA experiments to determine the levels of type I collagen and Clq in the GCF. The100remaining GCF sample was divided equally and employed for SDS/PAGE gels and western blotanalyses to ascertain collagen and Clq content at inflamed and noninflamed sites.2.5.2 Clq antibodiesAnalyses for Clq were performed using antibodies to it. Two hybridoma cell lines thatproduce monoclonal antibodies directed against Clq were graciously donated by Dr. Linda Curtiss(Research Instiute, Scripps Clinic, La Jolla, CA). These two hybridoma cell lines, 1H11 and2A10, were shown to have the greatest binding specificity to the head and stalk regions of Clqrespectively (Kilchher et al. 1985). The head and the stalk regions represented the globular andcollagen-like helical tail regions of the Clq molecule.Both the 1H11 and 2A10 hybridoma vials supplied 5 x 10-6 cells which were processed in thefollowing manner. The frozen cells were quickly thawed by immersing the vials in warm water.The cells were transferred to a 10 ml sterile polyethylene tube containing 5 ml of DMEM, andcentrifuged for 4 min. at 1500 g at room temperature to remove the medium. Each pellet was thenreconstituted in 5 ml of HT medium and transferred to a 25 cm 2 culture flask and incubated at37°C. The whole procedure was completed within 12 minutes.The HT medium was made from the following ingredients: two ml. of HT stock solutionwhich when reconstituted in 10 ml yielded 5 x 10 -3 M hypoxanthine and 8 x 10-4 M thymidine, tenml. of fetal calf serum, 0.5 ml. of Hepes stock solution, one ml. of 100 mM cis oxaloacetate stocksolution, five ml. of mixed supplement (containing 1 ml Pen/Strep; 1 ml Na pyruvate (0.1M); 1 mlNEAA solution (0.01M); 0.1 ml vitamin C (5 mg/ml) and 2.0 ml L-glutamine (0.2M)), and 81.5ml of the DMEM made up the total volume to 100 ml.The cells were observed after two days and each day thereafter for changes in colour of themedia, cell number and vitality. After one week the 1H11 cells reached a count of 1 x 106cells/mi.. These cells were split equally into two culture flasks and supplied with an equal amountof HT medium. As the 2A10 cells showed no evidence of proliferation, they were reincubated at10137°C. After two additional days the 1H11 cells were diluted to a ratio of 1:3 and 1:2 of cells andHT medium, respectively. However, there was still no evidence of proliferation of the 2A10 cells.Over the next few weeks the 1H11 cell line continued to grow in culture and the cells wereexpanded and stored in liquid nitrogen. The 2A10 cells did not proliferate and according to trypanblue vital staining (0.2% w/v) examination under the light microscope they were non-viable.Two additional vials (Q58 and 2A) of 2A10 hybridoma cells were donated by Dr. Curtiss'slaboratory and again processed in the same manner as described above. One day after incubation at37 °C, the cells from each culture flask were examined. The Q58 cells appeared to be slightly morenumerous and healthier than the 2A cells. However, after 5 days, the viable cell counts for theQ58 and 2A cells were 6 x 10 and zero, respectively. Continued careful monitoring and cellcounting every second day indicated that the Q58 cell cultures also became non-viable after anadditional two weeks of incubation.Since the 1H11 cells were predictably proliferating and several vials had been stored in liquidnitrogen, it was decided not to further pursue antibody production from the 2A10 cell line. 1H11antibody was known to be directed against the globular head of the Clq molecule. It was expectedthat this antibody would have less cross reactivity than the 2A10 antibody to collagen, because thelatter binds to the collagen-like helical domain of Clq.In order to increase the amount of Clq antibody (MAB-Clq) produced from 1H11 cells, micewere used to produce ascites fluid containing MAB-Clq. Five 9 week old Balb/c mice weresubjected for 5 seconds to a CO2 environment to make them drowsy before being primed with 0.5ml of pristane i.p. One week later the mice received, in the same manner, a booster of anadditional 0.5 ml of pristane. Three days later each mouse was injected with 0.5 ml i.p. of (5 x106) 1H11 cells in HT medium. After 11 days two mice developed ascites tumors. Ascites fluidwas harvested by puncturing the abdomen of each mouse with a 20 gauge needle and the fluid wascollected drop-wise in a 10 ml polyethylene tube. Table 2.2 depicts the sequence of events.102Vari-ablePrstneInj - 1PrstneInj-21H11cell injAsctsCol.AsctsCol.AsctsCol.AsctsCol.AsctsCol.AsctsCol.AsctsCol.Day -11 -3 0 11 12 13 14 16 18 21Vol-1 0.5 0.5 0.5 .75 5.0 2.5 2.5 7.0 3.0 2.5*Vol-2 0.5 0.5 0.5 2.75 0.5**Table 2.2:^Sequence of events for developing ascites fluid in Balb/c mice. 'Prstne' = pristane;'Ascts' = ascites fluid; 'Inj' = injection; '*'= mouse was sacrificed; '**' = mouse expiredThe specificity of monoclonal antibody to Clq (MAB-Ciq) was evaluated using commontissue macromolecules and molecules of similar structure. These included Clq, collagen types I,and IV, fibronectin (FN), laminin, and albumin. This MAB-Clq was found highly sensitiveand specific for Clq, and exhibited virtually no cross reactivity to the above mentionedcomponents. This was evaluated using ELISA by matching optical densities of serially dilutedClq with other serially diluted substrates, and then determining their relative concentrations for agiven level of Clq.A polyclonal antibody, rabbit anti-human Clq, was obtained from Calbiochem. Specificitytesting of this product was performed in the same manner and to the same components as theMAB-Clq. The results are displayed in Table 2.3. Even though both of these antibodies exhibitedexcellent specificity and high responsiveness, the monoclonal antibody was chosen for furtheranalyses since it provided the best specificity to Clq.103Specificity^Monoclonal AB^Polyclonal ABmouse anti-human Clq^rabbit anti-human ClqClqhuman collagen type Ihuman collagen type BIhuman collagen type IVhuman fibronectinmouse lamininhuman albumin100< 0.5< 0.5< 0.5<0.5<0.5< 0.5100<0.5<0.5<0.5< 1<0.5<0.5Table 2.3:^Percent specificities of monoclonal and polyclonal Clq antibodies. Evaluated bymatching optical densities of Clq with other substrates, and then determining their relativeconcentrations for a given level of Clq2.5.3 Type I collagen antibodiesIn order to perform type I collagen analyses, it was necessary to obtain appropriate antibodiesto it. A considerable effort was spent in locating, obtaining, and testing suitable antibodies to typeI collagen. The few commercial sources that offered the product had backorders, which actuallytranslated into a reality that they were having problems with their hybridoma cells lines. Limitedquantities were obtained from Seattle (Nayaranan, S., Seattle, WA), and preparations ofunacceptable purity were available from another laboratory (Sodek 1990).Finally a source of polyclonal rabbit anti-human type I antibody was obtained from ChemiconInternational. This antibody held great promise as suggested by the provided specificity sheet andthe list of recent publications that had used this product. This antibody was tested using ELISA forspecificity and cross reactivity with common tissue macromolecules and molecules of similarstructure; included were types I, III, and IV collagen, FN, laminin and Clq. Comparisons ofChemicon's quoted specificities and specificities determined by our testing are displayed in Table2.4.104Results of using this antibody to human type I collagen were not encouraging. Its sensitivitywas poor since the alkaline phosphatase reaction took 24 hours to develop. In addition, thespecificities to type I collagen were not acceptable as indicated by the obtained extensive andintense cross reactions. The reactivities demonstrated that type III collagen, type IV collagen, andClq had much better reactivity to this antibody than the intended type I collagen, indicating aheteroclitic specificity. These shockingly unfavourable specificities were confirmed by repeatedanalyses. The results obtained in these inital experiments were subsequently confirmed by in-housetesting of Chemicon's rabbit anti-human type I collagen antibody.Substrate^Chemicon's % specificity^ELISA measured % specificityhuman collagen type I^100^ 100human collagen type II^0.4 NThuman collagen type 0.8 120human collagen type N^0.4^ 195human collagen type V^0.5 <5human fibronection 0.4 <5mouse laminin^0.4^ NTbovine collagen type I^0.7 NTrat collagen type I 0.4 NTchicken collagen type I^' 0.4^ NTClq^ NT 400Table 2.4:^Percent specificities to Chemicon's polyclonal rabbit anti-human type I collagen.Evaluated by matching optical densities of type I collagen with other substrates, and thendetermining their relative concentrations for a given level of type I collagen. 'NT'= not tested.Selective adsorption experiments using mixed human sera were performed to try and reduce thecross reactivities. After sera adsorption of the type I collagen antibody ELISA evaluations wereperformed using Clq and types I and N collagens. Unfortunately, it resulted in a marked decreasein the overall sensitivity of the antibody with relatively no change in relative specificity to thesesubstrates.Substrate ChemiconpolyclonalChemiconmonoclonalNarayananpolyclonal Southern Biotech.polyclonalCollagen type ICollagen type InCollagen type NFibronectinLamininClq• wo^100^, 100^100120 180 28 38195 65 13 15<5 NT NT NMR<5 NT NT NMR400 190 NMR NMR105Subsequently, another antibody to human type I collagen became available from Chemicon.This preparation was classified as a monoclonal antibody to human type I collagen raised inrabbits. Our ELISA evaluations demonstrated that this product also had significant crossreactivities, particularly to Clq, and types III and IV collagens.Two additonal sources of antibody to type I collagen were evaluated. The first was a limitedamount of rat anti-human type I collagen antibody graciously provided by Dr. S. Narayanan(University of Washington, Seattle, WA). The second was an affinity purified goat anti-humantype I collagen antibody obtained from Southern Biotechnologies Associates Inc. (Birmingham,AL). Specificity testing of both of these antibodies demonstrated the appropriate specificity tocollagen with no measurable cross reactivity to Clq. Finally, for reasons of specificity andavailability, the affinity purified goat anti-human type I collagen antibody from SouthernBiotechnologies was chosen for use in ELISA and western blotting experiments. Table 2.5compares the specificities of all of the above mentioned type I collagen antibodies that wereevaluated.Table 2.5: Sources of antibody to type I collagen showing percentage specificity with relatedproteins. Evaluated by matching optical densities of type I collagen with other substrates, and thendetermining their relative concentrations for a given level of type I collagen. 'NMR'= nomeasurable response; 'NT'= not tested.1062.5.4 Crevicular fluid modelBecause it was desirable to ascertain the relative contributions of type I collagen and Clq inGCF, crevicular fluid models were used to evaluate how these components are affected by ouranalyses and how they can be affected in the periodontal environment. To test the effectiveness ofClq precipitation, two experiments were performed. The first evaluated Clq influence on totalHyp content in GCF, after processing it in different buffer systems and analysis by HPLC. Thesecond evaluated an acetate mixture of type I collagen and Clq after centrifugation, by ELISAusing antibodies to these components. Also, the effect of proteolytic enzymes from pathogenicbacteria on the recovery of type I collagen was evaluated using ELISA.The crevicular fluid collected on filter paper strips in the Spiramycin study was eluted in 5 mMNa2HPO4 and 0.5 N acetic acid which solubilizes neutral salt-soluble and acid-solublecompartments of collagen. The design of the experiment was to ascertain the efficiency of removalof Clq from the system by precipitating the Clq with acetate buffer.Sample preparation and procedures for testing Clq removal in sample buffers are outlined inFigure 2.8. Crevicular fluid samples were collected on filter paper strips and the volume wasdetermined using a Periotron 6000. Fluid was extracted twice from the filter paper strips usingtwo 100 ill aliquots of 0.005 M phosphate buffer. The two eluents were combined and dividedinto phosphate buffer and acetate buffer groups. In each of these groups the eluents were furthersub-divided: to one half human Clq was added at either one or five times the serum concentration(70 or 350 µg/ml); and to the other half no Clq was added. To each portion 100 11.1 of theappropriate buffer (either 0.005 M phosphate buffer or 0.5 N acetic acid) was added. The acetatebuffer systems were centrifuged at 15,600 g for 30 minutes at 4°C to remove precipitated Clq.Then the eluents from both the phosphate and acetate systems were processed for HPLC Hypanalyses (Coil and Tonzetich 1988).•Add Acetate Buffer• Centrifuge•Add ClqCentrifugeCrevicular Fluid onFilter Paper StripExtract twice with 100p10.005 M phosphate buffer107Combine Eluants•Add Phosphate BufferLI Non-centrifuged Add ClqNon-centrifuged• •^•^•Processing of Samples for HPLC Hydroxyproline analysisFigure 2.8: Flow diagram of Clq precipitation experimentClq removal in a sodium acetate buffer system from a crevicular fluid model was verified byELISA. In the ELISA system, elimination of Clq from crevicular fluid was assayed usingmonoclonal antibodies to Clq (MAB-Cq). For this test one p,1 of 200 pg/m1 human Clq wasadded to 49 p.1 of 0.05 M sodium acetate in a 1 ml Eppendorf tube. After centrifugation at 15,600gfor 30 min at 4°C, 25 IA of the supernatent was transferred to an ELISA plate well. The pellet wasresuspended in 25 pl of PBS and dispensed into an ELISA plate well. A 25 ill volume of108uncentrifuged sample was taken from a solution consisting of 1 pl of 200 pg Clq/ml in 49 pl PBSand added to an ELISA plate well. Sample volumes in each well were brought with PBS to a finalvolume of 50 pl. Incorporation of human Clq standards and ELISA analyses are described later insection 2.5.7.Identical evaluations for type I collagen in sodium acetate buffer were also performed on 1 glvolumes of 1 mg/ml of human type I collagen standard.To simulate an enzymatic activity in a periodontal environment 2 proteolytic enzyme sourceswere each incubated with type I collagen, and percent recovery of type I collagen was determinedusing ELISA. Bacterial collagenase (BC) and P. gingivalis bleb preparations (blebs) wereindividually combined with type I collagen . For this experiment two pi of BC, 2 pi of lmg/m1type I collagen, and 2 pl PBS containing 50 mM calcium were added to an ELISA plate well.Similary 5 pl of blebs, 2 pl of type I collagen, and 3 gl of PBS containg 50 mM calcium wereadded to an ELISA plate well. After 1 and 2 hours of incubation at 37°C analyses for type Icollagen were performed according to the ELISA analyses described in section 2.5.7.2.5.5 Sample collection and measurementPatients who satisfied the inclusion criteria were examined as potential patients for the study.Each patient received a thorough clinical examination of soft tissues, with particular attentionfocused on the gingiva. Periodontal sites that visually appeared to be inflamed, as indicated bylack of gingival stippling, papillary edema, erythematous papillary or marginal gingival werecatalogued along with several periodontal sites that visually appeared clinically healthy.All recorded sites were isolated and dried prior to sampling of crevicular fluid. Then aperiotron filter paper strip was placed at the entrance to the crevice and fluid was collected for 30seconds. After GCF volume was determined using a Periotron 6000, samples were placed inindividual 1.5 ml Eppendorf tubes containing 50 pl of 0.05 M sodium acetate and maintained in anice bath while being transferred to storage at -70°C.109All recorded sites were then probed using a Marquis periodontal probe to determine the pocketdepth and evaluate bleeding on probing response. Pocket depths were measured to the nearestmillimeter in the usual accepted manner. A positive bleeding on probing response was recordedfor those sites that bled within 10 seconds of probing. Sites which bled immediately upon probingwere also noted. A negative bleeding on probing response was assigned to those sites that did notshow evidence of bleeding within 30 seconds after probing.Patients were selected for the study if they exhibited at least one inflamed and one noninflamedperiodontal site. In addition, sites were chosen that were not on the same tooth or that shared acommon interproximal space. A total of fifty five subjects screened from a pool of patients whomet all the above criteria were chosen for the study.2.5.6 HPLC analysis for HypHPLC was used to determine total Hyp content of inflamed and noninflamed GCF samples.Individual crevicular fluid samples were removed from -70°C storage, thawed and centrifuged at15,600 g for 30 minutes. A 10 .tl aliquot of the supernatant was pipetted from each Eppendorftube and transferred to an individual 6 x 50 mm pyrex culture tube. Processing of samples forHPLC analyses was carried out using the Pico-tag Workstation in the same manner previouslydescribed in section 2.4.3.Important points of emphasis for the handling samples include the following. The sampleswere hydrolyzed at 150 °C for one hour. After derivatization the samples were dried thoroughly to65 millitorr, to remove all traces of PITC, which usually took 4 to 5 hours. This is an importantstep as incomplete drying would adversely affect the recovery of derivatized products.All analyses were performed on the same HPLC assembly described in section 2.4.3, with thefollowing changes in apparatus; an Automated Gradient Controller and a Model 740 data module(Waters Assoc.) were incorporated into the assembly unit and separation of the derivatized aminoacids was effected using a Pico-tag C18 column.110Crevicular fluid samples from individual inflamed and noninflamed sites in 54 patients wereanalyzed. These samples were grouped together so that the inflamed and noninflamed samplescollected from the same patient were included together in the same reaction vial. Amino acidstandards included in each HPLC reaction vial yielded the same profiles, indicating excellentbetween-run precision for the analyses.2.5.7 ELISA analysesELISA was used to test for the presence of type I collagen and Clq in GCF and crevicular fluidmodel. The primary antibodies employed in the study were affinity purified goat anti-human type Icollagen antibody from Southern Biotechnology, and mouse monoclonal human Clq antibodyfrom Dr. Curtiss. It was also necessary to use rabbit anti-human Clq (Chemicon). Secondaryantibodies employed were alkaline phosphatase conjugated goat anti-mouse IgG (BRL) andalkaline phosphatase conjugated rabbit anti-goat IgG (Sigma).For the type I collagen analyses, 55 paired samples of inflamed and noninflamed GCF wereanalyzed in the following manner. GCF samples stored at -70°C were thawed and centrifuged at15,600 g for 30 min. Duplicate four ill aliquots of each sample were taken and individually addedto 46 of phosphate-buffered-saline (PBS) in an ELISA plate well. Inflamed and noninflamedsamples from the same patient were included on the same plate. Triplicate standards of human typeI collagen were serially diluted from an initial concentration of 0.114/50 A total of four ELISAplates (Falcon) were used for these analyses.Ninety-six-well ELISA plates were coated overnight at 4°C with individual 4 gl aliquots of theGCF samples in 46 ill of PBS and with 50 ill of serially diluted type I collagen as a standard,leaving lane 1 as a blank. After washing twice with PBS the nonsaturated binding sites of theplates were blocked with 2% casein/PBS during 2 hours of incubation at 37°C followed by twosequential washes with PBS. Fifty microliters of 1% casein/0.05% Tween 20/PBS buffercontaining goat anti-human type I collagen antibody (1:5000 dilution) was added to all the wellsand incubated at 37°C for one hour. After two washes with PBS/0.1% Tween 20, the plates were111incubated with 50 p.1 of alkaline phosphatase-conjugated rabbit anti-goat-IgG antibody(Calbiochem) in 1% casein/0.05% Tween/PBS (1:1000 dilution) for one hour at 37°C. The plateswere then washed four times with PBS/0.1% Tween 20 and developed with 50pl/well (0.5mg/ml) of p-nitrophenyl phosphate (Sigma) in diethanolamine buffer. The optical densities weremeasured after 6 hours at room temperature, at 405 tun using a Titertek ELISA recorder.Analyses for Clq were performed in a similar manner on 55 paired samples of inflamed andnoninflamed GCF. Three p.1 aliquots of each sample were taken in triplicate and individually addedto 47 p.1 of PBS in an ELISA plate well. Inflamed and noninflamed samples from the same patientwere included on the same plate. Triplicate standards of human Clq were serially diluted from aninitial concentration of 0.1pW50 pl. A total of six ELISA plates were used for these analyses.Ninety-six-well ELISA plates were coated for 1.5 hours at 37°C with individual 3 gl aliquotsof the GCF samples in 47 p.1 of PBS and with 50 p.1 of serially diluted human Clq as a standard,leaving lane 1 as a blank. The procedure was the same as described above for type I collagenanalysis, except the secondary antibody used was alkaline phosphatase conjugated rabbit anti-goatIgG antibody (Sigma) in 50 of 1% casein/0.05% Tween/PBS (1:1000 dilution). The opticaldensities were measured after standing for 1 hour at room temperature.In the same manner, 24 paired samples of inflamed and noninflamed. GCF, already analyzedfor type I collagen content, were again analyzed for Clq content using the polyclonal antibody,rabbit anti-human Clq. Duplicate three pl aliquots of each GCF sample were subjected to ELISAanalysis as outlined above.2.5.8 SDS/PAGE gel electrophoresisSodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS/PAGE) was performed undernon-reducing conditions to separate proteins of each crevicular fluid sample. For analysis, a 6 p.1volume of each diluted GCF sample was combined with 4 p.1 of SDS/PAGE sample buffercontaining distilled water (4 ml), 0.5 M Tris-HC1, pH 6.8 (1.0 ml), glycerol (0.8 ml), 10% (w/v)112SDS (1.6 ml), and 0.05% (w/v) bromophenol blue (0.2 ml). Samples were denatrued byplacement in boiling water for 5 minutes.Polyacrylamide gel electrophoresis was performed using mini-gels with stacking gel andrunning gel acrylamide concentrations of 4% and 7.5%, respectively. The gel dimensions were 8cm (L) x 4 cm (H) x 0.75 mm (D) with 15 sample wells each of 3 mm width. The GCF sampleswere grouped together according to patient, having an inflamed site adjacent to a noninflamed site.A solution of prestained high range protein molecular weight standards (BRL, Burlington, Ontario,Canada), was added to a separate lane. Ten ill volumes of samples and standards were loaded intothe wells using a micro syringe and electrophoresed at a constant voltage of 120V. Once thebromophenol blue marker reached the bottom of the gel, the electrophoresis was terminated and thegel was removed and placed in 40% methanol solution prior to silver staining.Silver staining was performed using modified procedures developed by Oakley, Switzer andco-workers (Switzer et al. 1979; Oakley et al. 1980). Each gel was placed in 50 ml of the firstfixative (57.5 g TCA; 17.25 g 5-sulfosalicylic acid, dihydrate; in 500 ml dH2O) and gently shakenfor 30 minutes. Following a brief transfer to dH2O, the gel was exposed to the second fixative (50ml of 2.5% glutaraldehyde) and shaken for 45 minutes. Following three 10 min. rinses withdeionized water (dH2O), the gel was left overnight in 10% ethanol. A solution of diammine wasprepared by titrating 2 ml of 19.4% AgNO3 with base solution (10 ml 0.35% NaOH + 0.67 mlconc. NH3OH) until the red precipitate disappeared. To this diammine solution 1.5 ml of additionalbase solution, 5 ml ethanol and 32 ml dH2O were added. The gel was placed in a dish containingthis solution and orbitally shaken for 15 minutes. The gel was then thoroughly rinsed with threealiquots of fresh dH2O shaken for 20 min each, then left in 200 ml of dH2O for two hours.Reducer, consisting of 0.5 ml 1% citric acid, 0.05 ml formaldehyde, and 10 ml ethanol brought toa final volume of 100 ml with dH2O, was used to develop the stained gels. Once the desired stainintensity was reached, which usually took between 5 and 10 minutes, the gel was soaked in dH2Ofor at least another day before drying. Over-stained gels or background was lightened by soakingthe gel for 15 minutes in freshly prepared 5% sodium thiosulfate (w/v) in dH2O.1132.5.9 Western blottingFollowing SDS/PAGE, the gel was equilibrated in 50 ml of the blotting transfer buffer (25 mMTris, 192 mM glycine, 20% v/v methanol). After equilibration was achieved by gently shaking thegel for 10 minutes in the transfer buffer, the solution was discarded and replaced with a freshbuffer. This process removed electrophoresis salts and detergents which would otherwise increasethe conductivity of the transfer buffer and generate heat during the electrophoretic transfer. Inaddition, equilibration also allowed the gel to adjust to its final size, as the gel shrinks in thepresence of methanol.Powderless gloves were worn during gel equilibration and preparation of mini trans-blotmaterials in order to avoid contamination of the membrane. Four pieces of filter paper were cut tothe size of the gel during electrophoresis. The same size of an Immobulin-P transfer membrane(Millipore Corp., Bedford, MA, U.S.A.), was cut and labelled in the top and bottom corners toassist with future orientation of the membrane. The membrane was wetted in an organic solvent byslowly sliding it at a 45° angle into methanol and allowing it to soak for 30 seconds. Themembrane was then transferred to a dish containing transfer buffer and allowed to soak for at least15 minutes prior to blotting. The filter papers along with the fiber pads were slid into transferbuffer at a 45° angle to avoid entrapment of air bubbles, and left to soak for a minimun of fiveminutes before the assembly of the mini trans-blot apparatus.Prior to assembly of the transfer sandwich, the Bio-Ice cooling unit and 500 ml of transferbuffer precooled at 4° C, were placed in the buffer chamber containing a magnetic stirrer. The gelholder cassette was assembled sequentially to form a sandwich consisting of a fiber pad, two filterpapers, gel, membrane, two filter papers and the remaining fiber pad. The assembly wascontinually immersed in transfer buffer during the assembly of all components. At the stage whenthe membrane was placed over the gel, a small glass test tube was rolled over the membrane toensure an intimate contact between the gel and the membrane and also to remove any air bubblestrapped between the gel and the membrane. Following this step the remaining saturated two filterpapers and fiber pad were placed over the membrane and the gel holder cassette was closed.114The assembly was placed in the buffer chamber so that that gel faced the cathode electrode andadditional transfer buffer was added to within 1 cm of the upper edge of the chamber. Transferswere effected in 6 hours using an initial power setting of 35 V and 80 mA, which reached a fmalcurrent level of 120-140 mA.Immediately after the transfer, the membrane was ready for probing analysis for either Clq ortype I collagen. Unless stated otherwise, all procedures were performed at ambient temperature.Using gloves and forceps, the membrane was placed in 50 pi of 2% casein/PBS as a blockingreagent, and shaken for a minimum of 2 hours or left overnight at 4° C. The membrane wastransferred to another dish containing the first antibody solution. Either rabbit anti-human Clq(Calbiochem) or goat anti-human type I collagen (Southern Biotechnologies) were used in 50milliliters of 1% casein/0.05 % Tween 20/PBS (1:1000 dilution), and gently shaken for 3 to 4hours. The membrane received a brief rinse and two 50 ml. washes of 15 minutes each usingPBS/0.1% Tween 20, employing an orbital shaker.The membrane was then immersed in the second antibody solution which contained thecorresponding species specific anti IgG - alkaline phosphatase conjugate. Anti rabbit IgG - APconjugate or anti goat IgG - AP conjugate were used separately in 1:1000 or 1:2000 dilutionsrespectively, in fifty milliliters of 1%casein/0.05% Tween 20/PBS, which was agitated for 3 or 4hours using an orbital shaker. Each membrane received a brief rinse and three 100 ml washes for10 minutes each using PBS/0.1% Tween 20 under gentle shaking conditions.The blot was developed using the bromochloroindolyl phosphate/ nitro blue tetrazolium(BCIP/NBT) substrate from BRL (Bethesda Research Laboraties, MD, U.S.A.), which generatesan intense dark purple precipitate at the site of enzyme binding. During the fmal rinses of themembrane a fresh BCIP/NBT developing solution was prepared. One hundred microliters of NBTstock (0.1 g NBT in 2 ml of 70% dimethylformamide) was mixed in 25 ml of alkaline phosphatasebuffer (100 mM NaCI, 5 mM MgC12 ,100 mM Tris, pH 9.5). Fifty microliters of BCIP stock(0.1 g BCIP in 2 ml of 100% dimethylformamide) were added to complete the developing reagent.The washed immunoblot was placed in a 14 cm petri dish containing the BCIP/NBT reagent. The115blot was gently shaken until the bands appeared. Some bands appeared within a few minutes,while the weaker bands required more than 15 minutes to appear. After the desired intensity of thebands were reached and before the background colour increased, the membrane was immersed in20 mM EDTA which chelated Mg2+ ions and thereby stopped the precipitation reaction.2.5.10 Data analysisComparisons of total Hyp levels generated by HPLC analysis were performed on inflamed andnoninflamed sites in the same patient. They were evaluated using paired t-tests. ELISAdeterminations of type I collagen and Clq content were made using corresponding levels ofstandards plotted against the measured optical densities. Standard curves were generated for eachELISA plate analyzed. ELISA results for levels of Clq and type I collagen found in GCF aliquots,were normalized to the fluid volume of GCF initially collected. The units are reported in moles/illGCF. Then based upon the amount of Hyp known to be present in Clq and type I collagen, thesevalues were converted into moles Hyp/ill GCF, for each of these components. These in turn wererelated to the total moles of Hyp/p1 GCF, which was determined by HPLC analyses. The valueswere expressed in % (Clq or type I collagen) Hyp contribution of total Hyp. These values werecompared using the student's t-test. All statistical tests were performed at the 95% and 99% levelof significance.SDS/PAGE gels and western blots were qualitatively analyzed for banding pattern andpresence and position of antibody reactions. Qualitative assessment of any differences betweeninflamed and noninflamed sites was performed where appropriate.1163. RESULTS3.1 VOLATILE SULPHUR COMPOUNDS IN HUMAN GINGIVAL CREVICE3.1.1 Volatile sulphur compositionQualitative and quantitative differences were found in VSC compositions of mouth andcrevicular air samples. Figure 3.1 compares profiles for Tenax-trapped mouth air and Tenaxtrapped crevicular air from the same patient. These are reproducible profiles exhibiting peakscorresponding to H2S (Rt: 115-130 sec.), CH3SH (Rt: 190-205 sec.), (CH3)2S, (Rt: 270-310sec.), and (CH3S)2, (Rt: 390-425 sec.). Whereas both chromatograms exhibit profiles for H2S,CH3SH and (CH3)2S, the crevicular air sample shows significantly more CH3SH and (CH3)2S.These profiles demonstrate that the composition of crevicular air can differ from that of mouth air.The results further indicate that it is possible to collect sufficient amounts of VSC from specificgingival crevice sites for gas chromatographic analyses. Figure 3.2 compares profiles of Tenax-trapped crevicular air collected from two subjects. Whereas both chromatograms exhibit profilesof H2S, (CH3)2S and (CH3S)2, one chromatogram shows a prominent peak for CH3SH, itsconcentration exceeding that of H2S.Of the 17 patients studied, three contributed to both test (4mm) and control (53mm) crevicegroups, while the remaining subjects were divided into either group. Results of the gaschromatographic identification of VSC from control and test groups indicate that there arecomparative differences in composition of mouth and crevicular air. Tables 3.1a and 3.1b depictthe levels of H2S, CH3SH, (CH3)2S, and (CH3S)2, and the ratios for CH3SH to H2S and for allmethyl sulphides (CH3-S's) to H2S, found in mouth and crevicular air samples respectively. Totalsulphur content is also shown for crevicular air.Tenax-trapped Crevicular Air198 8118 8Tenax-trapped Mouth Air120 8130 8197 89033 5855^7! t10 8Time (seconds)Figure 3.1:^Comparative profiles of 10 ml of tenax-trapped mouth and crevicular air from thesame subject.117100^300^500^100^300^500Time (seconds)118Figure 3.2: Comparative profiles of tenax-trapped crevicular air from two different subjects.Rati Ratio18 .31 .27^.00^.00^.86^.861111111MI .29 .00 11011.4814 MEM .431.19048 .66 .00^036•.41 .00 .18 .78.00 .00 .00.00 .00.00 .00 .40.64 .001111.00 .00 .00 .00.09 .00 .24 .44.19 .00 .66 .90.00 1.051111111111111111111111114^.00^.00^.00^.00^.00.44.611.68.61.48 .09 .001.281.03.28.58.61.00.00.00.00111111101111 .00.68.24.451.16.38.47119Mouth AirTable 3.1a: Levels of sulphides found in 10 ml of mouth air. Ratios for CH3SH to H2S and forall methyl sulphides to H2S are presented.120Crevicular AirTotal^RatioPnt PD* IREGINZEIMMIERBEIBIBMOLIERMEHIMIIIMBIRatiogigs 24MB .42 MI= .00^1.03^.65^1.48mmi 46MB .49^.05^.05^.08^.68^.11^.384 1111 46MB .44 .07^.00^.00^mimmEminag 36MB .56^.07^.14^.14^.90^IMI .616^6^16MB .64 IMMI .10^.19^1.63^MilinallNMI 26MB .74 OM .64^.00^1.65^MI 1.248^4^46DB .68^.46^.00^.00^1.14^.68^.6810 MI 15MB In .65^MI .00^OM 1.97^INNIIMI 6^26MB .50^.00^.24^.89^11=1 .794^4^36DB .63^.28^.00^Mil .44^.87rIm 36DB Iffil .28 1^1.41^.55^1.74 ni 8^26DB .62 neal .36^1.20NEN 26mBri .19^.20^ristgamm. 1.49111 4^36DB .48^.47 .86^1.98^.98^3.1014 gm 26MB .00^.00^.00^.00^.00^.00^.00el 4^12DB .34 .00^.00^.91^1.67^1.6716^4^13MB .38^.00^.00^.91^1.41^1.414^25MB .74^.59^.00^.00 .80^.8018^4^12DB .63^.00^.00^.00^.63^.00^.00Table 3.1b: Levels of sulphides found in 10 ml of crevicular air. Ratios for CH3SH to H2S andfor all methyl sulphides to H2S, as well as the total sulphur content are presented. 'MB'=mesiobuccal; 'DB'= distobuccal.121H2S was the dominant VSC found in both the mouth and crevicular air samples. CH3SH waspresent in mouth air of all 5 control subjects and absent in 3 of the 10 test subjects, while increvicular air it was detected in variable amounts in all test and control subjects. In controls,(CH3)2S was detected in 4 out of 5 samples of both mouth and crevicular air. In test subjects,(CH3)2S was detected in 6 of 10 mouth air and in 8 of 10 crevicular air samples. It is noteworthythat 3 of the 8 test subjects exhibiting (CH3)2S in crevicular air did not exhibit detectable levels inmouth air. The fact that in these subjects the VSC composition of mouth air differed fromcrevicular air is additional support that the customized collection device created a closed system andthat crevicular air is distinct from mouth air in the same patient.3.1.2 Deep versus shallow and inflamed versus noninflamed sitesThe frequency of appearance of VSC components in mouth and crevicular air of both controland test subjects is displayed in Table 3.2. The subjects in control and test groups were furthersubdivided into depth and inflammation categories. Of the 17 patients examined, three contributedto all four categories, having one periodontal site that was both shallow and noninflamed, whileanother site was both deep and inflamed. Five and ten subjects contributed VSC values for thecontrol and test depth groups respectively (C: shallow and T: deep). Three of these patientscontributed separate sites, while five additional subjects contributed two and three sites, for thecontrol and test inflammation groups respectively (C: noninflamed and T: inflamed). Thefrequencies of mouth air VSC in the control and test inflammation groups, therefore, contain threecontributions of the same value. The frequency of VSC in these four groups is tabulated as anumber and percentage of times that each of the four examined sulphides were present.In both mouth and crevicular air H2S was the most frequently observed VSC in all four sitecategories. The occurrence of both CH3SH and (CH3)2S was similar, or slightly greater increvicular air over mouth air. There appears also to be no great differences in frequency betweencontrol and test groups for both of these sulphides in mouth and crevicular air. However,(CH3S)2 does appear in greater frequency in crevicular air than mouth air. As was the case with122the other methyl sulphides, (CH3S)2 did not appear to predominate in either the control or testgroups.Table 3.3 portrays the levels of VSC per 10 ml of mouth or crevicular air from control and testsites. In both instances, H2S was the VSC collected in higher amounts from all subjects regardlessof site category. Furthermore, there was no evidence of dominance of H2S in either of the controlor test groups. Although CH3SH was found to the same proportion in control and test groups foreach of the mouth air samples, the levels of this sulphide were significantly different in control andtest groups of crevicular air. In both the depth and inflammation categories, crevicular air levels ofCH3SH were higher in the test over the control groups at the 95% level of significance. The sametrend held true for (CH3S)2, although in the crevicular air samples, the differences in the sulphidelevels were not statistically significant. The levels of (CH3)2S were similar in mouth air for bothsubcategories in control and test groups. However, in crevicular air, the levels of this sulphidewere statistically greater at the 90% level of significance in the inflamed over the noninflamedgroup. No significant difference was found in the level of this sulphide between deep and shallowgingival crevices.123outh^Air^Crevicula r AirCrevice^H2S^CH SH (CH )2S (CH S)2 H2S^CH SH (CH )2S (CH S)2T:inflamed 6{100} 5 {83}^4 {67}^0 {0}^6 {100} 6 {100} 3{50}^2 {33} Table 3.2: Frequency of VSC in mouth and crevicular air from Control (C) and Test (I) sites.Mouth^Air^Crevicula^r^AirCrevice H2S CH3SH (CH3)2S (CH3S)2 H2S CH3SH (CH3)2S (CH3S)2Sites ng {SD} ng {SD} ng {SD} ng {SD} ng {SD} ng {SD} ng {SD} ng {SD}C: shallow . .72{.19} .45{.42} .22{.21} 0{0} .45{.09} .14{.11} .19{18}^.04{.06}C: noninfl ; .53{.22} .25{.20} ! .14{.18} 0{0} .42{.25} .04{.04} .04{.06}^.04{.06}T: deep • .85{.68} .39{.44} .17{.23} .02{.06} .56{.12} .37{.19} .20{.18}^.24(.28}T:inflamed . 1.1{1.2} .27{.22}091.281 0{ 0} .47{.11} .41{.17} .131.121, .25{.38}Table 3.3: Levels of VSC per 10 ml of mouth or crevicular air at Control (C) and Test (1) sites.124Figures 3.3 and 3.4 illustrate sulphide ratios found in crevicular air. Figure 3.3 demonstratesthe sulphide ratios of CH3SH/H2S and total CH3-S's/H2S in control and test groups. A similartrend was found for both the ratios of CH3SH to H2S and all methyl sulphides to H2S. Deepercrevicular sites exhibited higher sulphide to H2S levels than the shallow sites. In particular, theCH3SH to H2S ratio of the test group was significantly greater than the control (p<.10). Figure3.4 shows the sulphide ratio of CH3SH to H2S for the control versus test groups and compares thedepth and inflammation categories. Comparisons in each category revealed that the ratio ofCH3SH to H2S was greater in the test than the control groups. In both the deeper and inflamedsites the ratios were significantly larger (p<0.1 and p<.05) than the shallow and noninflamed sites,respectively.Quantitation of the total VSC content in crevicular air of control and test subjects is displayed inFigure 3.5. It shows the same trend for the CH3SH to H2S ratio; the deeper and inflamed sitesshow higher total sulphur content than the corresponding shallow and non-inflamed sites. Thetotal sulphur content in deeper sites was significantly different from shallow sites (p<.05).Similarly, the total sulphur content at inflamed sites was significantly different than fornoninflamed sites (p<.05).125C*^T*^C^TCH3SH / H2S CH3-S's / H2SFigure 3.3: Ratios of CH3SH to H2S and total methyl sulphides (CH3-S's) to H2S fromcontrol (C) and test (T) sites of crevicular air. T* is not significantly greater thanC* (p>.05). Error bars represent the standard error of the mean.126C*^T*^NI**^I* *Crevice sitesFigure 3.4: Ratios of CH3SH to H2S of control (C) and test (T) and noninflamed (NI)versus inflamed (1) crevicular sites. T* is significantly larger than C* (p<.10). I**is significantly larger than that NI** (p<.05). Error bars represent the standarderror of the mean.127C*^T*^NI**^I**Crevice sitesFigure 3.5: Total sulphur content of H2S, CH3SH, (CH3)2S, (CH3S)2 in crevicular air atcontrol (C) versus test (1); and noninflamed (NI) versus inflamed (1) sites. T* issignificantly different than C* (p<.05). I** is significantly different than NI**(p<.05). Error bars represent the standard error of the mean.3.2 HYDROXYPROLINE LEVELS IN GINGIVAL CREVICULAR FLUID ANDSPIRAMYCIN EFFECT ON PERIODONTAL SITES3.2.1 Hydroxyproline content in hydrolyzed and unhydrolyzed samplesThe results of HPLC analyses indicate that it is possible to demonstrate and quantitate Hyplevels using small volumes of crevicular fluid in the range of 0.2 to 1.0 p1 collected on strips of128filter paper. HPLC analyses using PITC precolumn derivatization yielded distinctly differentresults for hydrolyzed and unhydrolyzed crevicular fluid. The analyses showed that acidhydrolysis yielded significantly more Hyp and other amino acid residues then did the unhydrolyzedsamples. Most of the unhydrolyzed crevicular fluid samples exhibited no peak for Hyp and barelydetectable presence of other amino acids. The corresponding hydrolyzed samples exhibited Hyppeaks and strong responses for remaining amino acids. Representative HPLC chromatograms ofPITC-derivatized unhydrolyzed and hydrolyzed crevicular fluid are shown in Figures 3.6 and 3.7,respectively.These results indicate that the preponderance of amino acid content of crevicular fluid is presentin a peptide form, and a disproportionately lesser amount in a free unbound form. Consequently,the Hyp analyses of crevicular fluid collected during the spiramycin study were performed on thehydrolyzed samples.3.2.2 Hydroxyproline levels in inflamed and noninflamed sitesAll crevicular fluid samples were analyzed by reversed phase HPLC. The amount of Hyp wasdetermined and presented as either moles of Hyp for the entire collected GCF sample or as molesof Hyp per ill of GCF. Hydroxyproline levels in the total GCF sample from inflamed and non-inflamed sites of both the spiramycin and non-drug treatment groups, for all examination periodsof the study, are shown in Table 3.4. Parallel values for Hyp levels in moles of Hyp per pi ofGCF are shown in Table 3.5. These Hyp levels represent mean values in inflamed or noninflamedcategories during each examination point and for each drug treatment group.\\...0.111^ I110^1 20 5Elution time (min)Figure 3.6: HPLC profile of unhydrolyzed PITC-derivatized crevicular fluid sample.1291V_ Az.5O 110 120.1I IiElution time (min)Figure 3.7: HPLC profile of hydrolyzed P1TC-derivatized crevicular fluid sample.130011I1 1HypI^L,.../ tv,A.\\___k____ v ,,A,131Tables 3.4 and 3.5 are graphically presented in Figures 3.8 through 3.11. In Figures 3.8 and3.9, in which the data is plotted as the total moles of Hyp per sample, the levels of Hyp are higherin the inflamed than in the noninflamed periodontal sites. This trend holds true for all examinationpoints, regardless of the drug treatment group, with one exception. In Figure 3.8 at week two inthe spiramycin treatment group, the Hyp levels for the noninflamed group are greater than those forthe inflamed group, however this difference is not statistically significant at the 95% level ofsignificance. But, in Figure 3.9 at week 8 in the non-drug group where the Hyp level is muchhigher for the inflamed group compared to the noninflamed group, this difference is statisticallysignificant at the 95% level of confidence. All other comparisons of Hyp content in inflamedversus non-inflamed sites were not statistically significant (p>.05).Figures 3.10 and 3.11 display Hyp levels per of GCF during all examination points for bothinflamed or noninflamed sites, for either the spiramycin or non-drug treatment groups,respectively. Unlike the previous two figures, Hyp levels in Figures 3.10 and 3.11 follow nodistinct trend at various examination points. In Figure 3.10, Hyp levels in inflamed sites are lowerthan found in noninflamed sites from weeks 0 to 12, but this trend is reversed at week 24. For thenon-drug treatment group in Figure 3.11, the Hyp level in inflamed sites is less during the initaland final exam points, but slightly greater during the remaining exam points, when compared toHyp levels in noninflamed sites. All comparisons of Hyp content in inflamed versus non-inflamedsites were not statistically significant (p>.05).The average Hyp levels at zero time point for combined drug and non-drug groups aredisplayed in Table 3.6 and graphically represented in Figure 3.12. Although the inflamedperiodontal sites show more total Hyp in the collected GCF sample (MHyp) than in noninflamedsites, an opposite relationship is seen for the Hyp levels calculated on the basis of Hyp per ill ofGCF. For these weighted Hyp (WHyp) values, the Hyp level is slightly higher for thenoninflamed versus the inflamed site. The difference in Hyp levels between inflamed andnoninflamed sites, for either total Hyp or weighted Hyp, is not statistically significant (p>.05).132Hydroxyproline values for all examination points for both drug groups are listed in Table 3.7.Mean values of total Hyp for each time point are displayed for drug and non-drug groups in Figure3.13. The drug group shows a trend of decreasing Hyp values through to week 12, which at week24 reverses back to slightly above the week zero baseline level. The non-drug group exhibits aslight drop in Hyp value at week two, which is followed by a rise in week 8 and a further riseslightly above baseline in week 12, and ends with a decrease at week 24. In comparingspiramycin-treated versus non-drug treated groups the only significant differences in the totalmoles of Hyp in a given examination period are found in weeks 12 and 24. At week 12 the totalmoles of Hyp are significantly higher in the non-drug treated group than in the drug treated group(p<.05). Conversely, at week 24 the total moles of Hyp are significantly larger in the spiramycintreated group compared to the non-drug treated group (p<.05). In addition, there is a significantdifference at the 95% level between total Hyp levels at week zero versus week 12 in the drugtreatment group.Mean values of weighted Hyp for each time point are displayed for drug and non-drugtreatment groups in Figure 3.14. Although there is a difference in the levels of Hyp in spiramycinversus non-drug treatment groups at week zero, this difference is not statistically different (p>.05).For the spiramycin treated group, there is a trend for an increase in Hyp content throughout theexamination periods. This is first observed as a significant rise in Hyp at week two compared toweek zero (p<.05). Subsequent examination points exhibit rising Hyp levels above that of weekzero. No other significant differences were found in the spiramycin treated group.The general trend for the non-drug treatment group is for a decrease in Hyp levels after weekzero. At week 12 there is a slight increase in Hyp, but the level is similiar to that found at week 2,and still below the level found at week zero. Although there is a tendency for Hyp levels to bebelow those found at week zero, no statistically significant differences in Hyp levels were foundamongst the examination periods (p>.05).133Drug Stat WeekBP=10BP=OWeekBP=12BP=OWeekBP=18BP=0WeekBP=112BP=0WeekBP=124BP=O+ X 2.19 1.57 1.76 2.20 1.94 1.59 1.59 1.27 2.32 2.11+ SD 1.91 1.77 1.37 2.28 1.99 1.30 3.21 1.21 1.73 1.52+ n 61 15 40 59 26 31 25 38 7 23+ SE .24 .46 .22 .30 .39 .23 .64 .20 .65 .32- X 2.52 1.67 2.22 1.48 2.86* 1.51* 2.59 1.97 1.46 1.30- SD 4.93 1.08 3.40 0.84 2.63 1.32 2.66 1.79 1.92 1.19- n 50 27 44 40 27 39 30 27 17 45- SE .70 .21 .51 .13 .51 .21 .49 .34 .47 .18Table 3.4: Total Hyp levels in both the spiramycin (Drug= +) and non-drug (Drug= -)treatment groups. Units for Hyp are in moles {Exp -111. 'BP=1 1= bleeding on probing positive;'BP=O' = bleeding on probing negative; X= mean; SD= standard deviation; n= number ofperiodontal sites; SE= standard error of the mean. '*' indicates a statistically significant differenceat the 95% level of significance.Drug Stat WeekBP=10BP=0WeekBP=12BP=0WeekBP=18BP=0WeekBP=112BP=0WeekBP=124BP=0+ X 4.15 4.38 4.61 8.17 5.31 5.78 4.70 6.87 10.52 6.00+ SD 5.73 5.33 4.05 12.13 6.64 8.09 4.93 12.97 14.60 8.79+ n 16 15 40 59 26 31 25 38 7 23+ SE 0.73 1.38 0.64 1.58 1.30 1.45 0.99 2.10 5.52 1.83- X 6.20 6.46 5.42 5.01 4.50 3.99 5.32 4.89 2.90 4.63- SD 8.65 8.72 8.44 8.56 5.90 3.70 6.48 5.31 3.47 7.97- n 50 27 44 40 27 39 30 27 17 45- SE 1.22 1.68 1.27 1.35 1.14 0.59 1.18 1.02 0.84 1.19Table 3.5: Weighted Hyp levels in both the spiramycin (Drug= +) and non-drug (Drug= -)treatment groups. Units for Hyp are in moles {Exp -11} per gl of GCF. 'BP=1'= bleeding onprobing positive; 'BP=O' = bleeding on probing negative; X= mean; SD= standard deviation; n=number of periodontal sites; SE= standard error of the mean.Week 0^Week 2^Week 8^Week 12^Week 24Examination points for spiramycin treatment goupFigure 3.8:^Total Hyp levels at periodontal sites from the spiramycin treatment group at allexamination points of the study. Hydroxyproline units are in moles Hyp {Exp-11}. Error barsrepresent the standard error of the mean.134Week 0^Week 2^Week 8^Week 12^Week 24Examination points for non -drug treatment groupFigure 3.9:^Total Hyp levels at periodontal sites from the non-drug treatment group at allexamination points of the study. Hydroxyproline units are in moles Hyp {Exp-11}. Error barsrepresent the standard error of the mean.135Week 0^Week 2^Week 8^Week 12^Week 24Examination points for spiramycin treatment groupFigure 3.10: Weighted Hyp levels at periodontal sites from the spiramycin treatment group at allexamination points of the study. Hydroxyproline units are in moles Hyp {Exp-11) per Ill GCF.Error bars represent the standard error of the mean.13614.0 -aa) 12.0 -0.\"14 to.°8.0 -a)S6.0-+0-a4.0 -2.0 -OWeek 0^Week 2^Week 8^Week 12^Week 24Examination points for non-drug treatment groupFigure 3.11: Weighted Hyp levels at periodontal sites from the non-drug treatment group at allexamination points of the study. Hydroxyproline units are in moles Hyp {Exp-11} per ill GCF.Error bars represent the standard error of the mean.13700138Statistic BoP = 1WHypBoP = 1MHypBoP = 0WHypBoP = 0MHypX 5.07 2.34 5.72 1.64SD 7.231 3.58 7.68 1.34n 111 111 24 42SE 0.69 0.34 1.18 0.21Table 3.6: Hydroxyproline values at week zero time point for inflamed and noninflamed sitesfor both spiramycin and non-drug treatment groups combined. WHyp= Moles Hyp {Exp -111 / .tlGCF; MHyp= Moles Hyp {Exp -11}; X= mean; SD= Standard deviation; n= Number ofperiodontal sites; SE= Standard error of the mean.1^ 0Bleeding on probing indexFigure 3.12: Hydroxyproline levels in inflamed and noninflamed periodontal sites at week zerotime point. 'Mhyp'= Moles Hyp {Exp-11} for total GCF sample; 'Whyp'= Moles Hyp {Exp-11}per ill GCF. Error bars represent the standard error of the mean.139Drug Stat Week 0WHyp MHypWeek 2WHyp MHypWeek^8WHyp MHypWeek^12WHyp MHypWeek 24WHyp MHyp+ X 4.12 2.06 6.73 2.02 5.57 1.75 6.01 1.40 7.05 2.16+ SD 5.62 1.89 9.83 1.96 7.41 1.64 10.31 1.20 10.32 1.55+ n 76 76 99 99 57 57 63 63 30 30+ SE .64 .22 .99 .20 .98 .22 1.30 .15 1.88 .28- X 6.29 2.22 5.23 1.87 4.12 2.06 5.12 2.30 4.16 1.34- SD 8.62 4.03 8.45 2.54 4.69 2.05 5.91 2.29 7.04 1.41- n 77 77 84 84 66 66 57 57 62 62- SE .98 .46 .92 .28 .58 .25 .78 .30 .89 .18Table 3.7: Hydroxyproline values at various exam points for both spiramycin (Drug= +) andnon-drug (Drug= -) treatment groups. Stat= Statistic; WHyp= Moles Hyp {Exp-11 } / ill GCF;MHyp= Moles Hyp {Exp-11 }; X= Mean; SD= Standard deviation; n= Number of periodontalsites; SE= Standard error of the mean.140Examination pointFigure 3.13: Total Hyp levels at all examination points for both the spiramycin and non-drugtreatment groups. 'Drug = +' = spiramycin treated group; 'Drug = -' = non-drug group. Errorbars represent the standard error of the mean.141Week 0^Week 2^Week 8^Week 12^Week 24Examination pointFigure 3.14: Weighted Hyp levels at all examination points for both the spiramycin and non-drugtreatment groups. Drug = +' = spiramycin treated group; 'Drug = -' = non-drug group. Errorbars represent the standard error of the mean.142Hydroxyproline values at week zero time point for shallow and deep periodontal sites that areeither inflamed or noninflamed are given in Table 3.8. The mean, standard deviation, number ofsubjects and standard error are listed for all drug treatment groups. These data are graphicallyrepresented in Figures 3.15 through 3.17. For the combined drug treatment groups, the Hyplevels for shallow pockets show a different trend than for deeper sites. In Figure 3.15, the totalHyp in the collected sample is greater in inflamed versus noninflamed sites. This same trend istrue for the weighted Hyp levels in shallow sites, showing that inflamed sites exhibit more Hypthan do the noninflamed sites. In pocket depths greater than or equal to 4 mm, Hyp levels forinflamed and noninflamed crevicular fluid are approximately the same. No statistically significantdifferences were found amongst these baseline Hyp measurements (p>.05).Hydroxyproline levels were compared to bleeding on probing measurements in 127 periodontalsites of 17 patients that had complete clinical and Hyp analysis data for weeks zero and twelve ofthe spiramycin study. Bleeding on probing scores were compared amongst periodontal sitesbetween these time points. The results indicated that inflamed periodontal sites at week zero whichremained inflamed at week twelve had higher Hyp levels at week twelve than sites that remainednoninflamed. In the drug group a significant difference was found between Hyp levels at weektwelve in inflamed and noninflamed sites (p<.05). Tables 3.9 and 3.10 display mean Hyp valuesand bleeding on probing recordings at weeks 0 and 12, for spiramycin and non-drug treatmentgroups respectively.3.2.3 Hydroxyproline levels versus attachment levelsIn comparing Hyp levels to changes in attachment levels in 127 periodontal sites in the abovespiramycin study, changes in probable attachment level were deemed significant if changes inmeasurements were greater than or equal to 2 mm. In both treatment groups Hyp levels werehigher in sites that experienced a gain of attachment of mm between weeks 0 and 12, than sitesthat remained unresolved. In the drug group at week twelve a significant difference was found in143Hyp levels between healing and nonresolved sites (p<.05). Changes in attachment level betweenweeks 0 and 12 for both treatment groups are displayed in Figure 3.18.Drug Statistic Pocket^Depth^< 4rnm_ Pocket Depth^...4mmBoP=1^BoP=1WHyp MHypBoP=0WHypBoP=0MHypBoP=1WHypBoP=1MHypBoP=0WHypBoP=0MHyp+ and - X 8.55 3.52 6.72 1.58 4.26 2.06 . 4.39 1.71+ and - SD 12.49 7.39 9.06 1.44 5.10 1.78 5.29 1.24+ and - n 21 21^24 24 90 90 18 18+ and - SE 2.73 1.61 ' 1.85 0.29 0.54 0.19 1.25 0.29+ X 2.86 1.74 5.76 2.02 4.44 2.28 , 2.32 0.89+ SD 1.70 0.89 6.46 2.19 6.25 2.06 . 2.03 0.35+ n 11 11^9 9 50 50 .6 6+ SE 0.51 0.27 I 2.15 0.73 0.88 0.29 ; 0.83 0.14- X 14.81 5.48 7.30 1.31 4.04 1.78 5.42 2.13- SD 16.14 10.59 10.49 0.68 3.19 1.34 6.15 1.32- n 10 10 15 15 40 40 12 12- SE 5.10 3.35 2.71 0.17 0.50 0.21 1.78 0.38Table 3.8: Hydroxyproline values at week zero time point for both shallow and deep pocketssubdivided into inflamed or noninflamed, for spiramycin (Drug= +), non-drug (Drug= -), and all(Drug= + and -) treatment groups. 'BoP=1'= Bleeding on probing positive; 'BoP=O'= Bleedingon probing negative; WHyp= Moles Hyp {Exp-11} / GCF; MHyp= Moles Hyp {Exp-11}; X=mean; SD= Standard deviation; n= Number of periodontal sites; SE= Standard error of the mean.W-Hyp, BoP=1W-Hyp, BoP=0^ M-Hyp, BoP=1• M-Hyp, BoP=010 -1110c.0• 412Pocket Depth < 4mm^Pocket Depth 4mmPocket depth category at week zeroFigure 3.15: Hydroxyproline levels for combined drug treatment groups at week zero forshallow and deep pockets. W-Hyp= moles Hyp {Exp-11} / GCF; M-Hyp= moles Hyp {Exp-11 }; 'BoP=1'= Bleeding on probing positive; 'BoP=O'= Bleeding on probing negative. Error barsrepresent the standard error of the mean.144• Drug= +, BoP=10 Drug= -, BoP=1^ Drug= +, BoP=0^ Drug= -, BoP=O145Pocket Depth < 4mm^ Pocket Depth ?_ 4mmPocket depth category at week zeroFigure 3.16: Total Hyp levels for both drug treatment groups at week zero for both bleeding onprobing categories and shallow and deep pocket depths. 'Drug= +'= Spiramycin treatment group;'Drug= -'= Non-drug treatment group; 'BoP= +'= Bleeding on probing positive; 'BoP= -'=Bleeding on probing negative Error bars represent the standard error of the mean.• Drug= +, BoP=10 Drug= -, BoP=1^ Drug= +, BoP =0E2 Drug= -,BoP =0•146Pocket Depth < 4mm^Pocket Depth 4mmPocket depth category at week zeroFigure 3.17: Weighted Hyp levels for both drug treatment groups at week zero for both bleedingon probing categories and shallow and deep pocket depths. 'Drug= +'= Spiramycin treatmentgroup; Drug= -'= Non drug treatment group; 'BoP= +'= Bleeding on probing positive; 'BoP= -'= Bleeding on probing negative. Error bars represent the standard error of the mean.147N= 15Week 0,Week (0,12) BoP (+/+)BoP +^Week 12, BoP +N= 26 Week (0,12) BoP (+/-)Week 0, BoP +^Week 12, BoP -MHyp{SE}WHyp{SE}[MHyp{SE}WHyp{SE}MHyp{SE}WHyp{SE}MHyp{SE}WHyp{SE)1.43{0.40}2.81{1.01}0.72{0.25}1.58*{0.59}1.67{0.52}2.74{0.97}0.23{0.12}0.45(0.131N=2 Week (0,12) BoP (-/+)Week 0, BoP -^Week 12, BoP +N=11^Week (0,12) BoP (-/-)Week 0, BoP -^Week 12, BoP -MHyp{SE}WHyp{SE}MHyp{SE}WHyp{SE}MHyp{SE}WHyp{SE}MHyp{SE}WHyp{SE}4.02{3.22}3.97{2.81}0.55{0.25}1.22{0.22}0.41{0.15}0.80{0.30}0.41{0.35}0.23*{0.12}Table 3.9: Hydroxyproline values and bleeding on probing measurements at weeks 0 and 12in spiramycin treatment group. 'N'= number of subjects, 'Mhyp'=Moles Hyp {Exp-11},'Whyp'=Moles Hyp {Exp-11}/41 GCF, 'SE'=Standard error. ' 4' 1= Hyp level is significantlylower (p<.05).N= 18Week 0,Week (0,12) BoP (+/+)BoP +^Week 12, BoP +N= 28Week 0,Week (0,12) BoP (+/-)BoP +^Week 12, BoP -MHyp{SE}WHyp^MHyp{SE}^{SE}WHyp{SE}MHyp{SE}WHyp{SE}MHyp, {SE}WHyp{SE}2.80{1.91}3.83^0.71{2.52}^{0.26}1.47{0.58}1.23{0.27}2.75{0.62}0.36{0.10}0.75{0.28}N=9 Week (0,12) BoP (-/+)Week 0, BoP -^Week 12, BoP +N=18 Week (0,12) BoP (-/-)Week 0, BoP -^Week 12, BoP -MHyp{SE}WHyp{SE}MHyp{SE}WHyp{SE}MHyp{SE}WHyp{SE}MHyp{SE}WHyp{SE}1.11{0.34}2.29{1.02}0.05{0.05}0.18{0.17}0.85{0.28}2.59{1.00}0.53{0.28}0.81{0.33}Table 3.10: Hydroxyproline values and bleeding on probing measurements at weeks 0 and 12in non-drug treatment group. 'N'= number of subjects, 'Mhyp'=Moles Hyp {Exp-11 },'Whyp'=Moles Hyp {Exp-11}/121 GCF, 'SE'=Standard error.148^AAL 2^-2 .05). A comparison of the Clq derived Hyp contribution from inflamed andnoninflamed sites to total Hyp levels is presented in Figure 3.21. Furthermore, the ELISA assayusing this polyclonal antibody resulted in a narrow range of Clq levels in GCF. The range of thedata obtained with this antibody preparation was deemed a more accurate representation of the Clqcontent in GCF since none of the Clq values contributed more than 100% to the total Hyp levelsdetermined by HPLC.154Cl q^Grey. fluid^Clq^Crev. fluidInflamed sites VS^Noninflamed sitesFigure 3.21: ELISA results using a polyclonal antibody to Clq. Clq derived Hyp contributionsto crevicular fluid from inflamed and noninflamed sites compared to total Hyp content of crevicularfluid. Error bars represent the standard error of the mean.3.3.4 Type I collagen in crevicular fluid modelThe effects of sodium acetate buffer and proteolytic enzymes on the recovery of type I collagenwere assessed in crevicular fluid models. After centrifugation of a 1 Al volume of 1 mg/m1 type Icollagen added to 49 ill of 0.05 sodium acetate buffer, collagen content was determined usingpolyclonal antibodies in an ELISA assay. Analysis of the supernatant indicated in Figure 3.20 thatessentially all of the collagen was recovered following centrifugation. Thus, whereas whole155collagen is soluble in acetate and remains in the supernatant, the intact Clq is precipitated andremoved by this procedure.The reliability of an ELISA assay to quantitate degraded collagen was assayed in systems inwhich collagen was incubated with proteolytic enzymes. In this study type I collagen wasincubated with either bacterial collagenase (BC) or bleb preparation from P. gingivalis (blebs), forone and two hour periods at 37°C. After one hour of incubation, 82.1% and 90.3% of the addedcollagen was detected in the BC and blebs' enzyme systems, respectively. After 2 hours ofincubation the collagen content was further reduced to 78.8% and 77.7%. The results depicted inFigure 3.22 show that the employed enzymatic preparations caused a reduction of anti-collagen Ideterminants due to degradation of collagen which increased with incubation time.3.3.5 Type I collagen in crevicular fluidThe content of type I collagen in crevicular fluid was determined from both inflamed andnoninflamed sites in each of the 54 original subjects. Employing a polyclonal antibody to type Icollagen in an ELISA assay, collagen content of the crevicular sites was determined on duplicatesamples.Each ELISA reading was converted to a molar value of collagen, which was then normalized toan amount of collagen per ill of crevicular fluid. Finally, a percentage of Hyp contributed by acollagen was calculated and compared to the total Hyp level as determined by HPLC analyses. Theresults indicated that in inflamed sites type I collagen contributed 27.68% (+/-SE 6.12) of the totalHyp content of GCF. In noninflamed sites type I collagen contributed 51.65% (+/-SE 9.56) to thetotal Hyp content. Analysis of ELISA results using the student's t-test indicated that noninflamedsites showed significantly greater amounts of collagen-derived Hyp than the inflamed sites(p<.05). Figure 3.23 exhibits the percent Hyp content derived from collagen measured frominflamed and noninflamed sites and total Hyp content of crevicular fluid.1 0080 -bAOu 60 —I-1a)tn.:212 40 —as04JVati20 —C1 2hr^A 1hr^B 1hr^A 2hr^B 2hrEnzyme fraction and time pointFigure 3.22: Detection of degraded collagen using ELISA, after one and two hour incubationswith bacterial collagenase (A) and P. gingivalis blebs (B).156157Comparisons of the amount of type I collagen and Clq in inflamed and noninflamed crevicularfluid were made on the 24 subjects whose GCF Clq levels were analyzed using the polyclonalClq antibody. Figure 3.24 displays the mean ELISA values of moles of type I collagen and Clqper ill of GCF, from inflamed and noninflamed periodontal sites. In inflamed sites the mean type Icollagen content was 0.65 (+/-SE .18) picomoles/g1GCF, while the content of Clq was 0.25 (+/-SE .04) picomoles/p1 GCF. In noninflamed sites the mean type I collagen content was 0.82 (+/-SE .19) picomoles/1.11 GCF, while the content of Clq was 0.19 (+/-SE .04) picomoles/111 GCF.The results indicated that the moles of type I collagen were significantly higher then moles of Clqin GCF, for both inflamed and noninflamed groups (p<.025 and p<.0025 respectively).Figure 3.25 depicts the mean ELISA values of moles of Hyp contributed by type I collagen andClq per ptl of GCF, from inflamed and noninflamed periodontal sites. In inflamed sites the meanHyp contribution from type I collagen was 160 (+/-SE 43.9) picomoles/111 GCF, while the Hypcontribution from Clq was 33.85 (+/-SE 5.06) picomoles/g1 GCF. In noninflamed sites the meanHyp contribution from type I collagen was 202 (+/-SE 45.9) picomoles/p1 GCF, while the Hypcontribution from Clq was 26.05 (+/-SE 5.02) picomoles4t1 GCF. The results indicated themoles of Hyp contributed from type I collagen were significantly larger than moles of Hypcontributed by Clq per ill of GCF, for both inflamed and noninflamed groups (p<.01 and p<.001respectively).Figure 3.26 displays the combined type I collagen and Clq Hyp percent contributions to totalHyp content for inflamed and noninflamed periodontal sites from 24 subjects. The percentcontributed from type I collagen derived Hyp is significantly greater than by Clq-derived Hyp, inboth inflamed and noninflamed groups (p<.01 and p<.0025 respectively). Both the type I collagenand Clq-derived Hyp levels are lower in the inflamed sites than the corresponding Hyp levels inthe noninflamed group.Total HPLC^Inflamed^NoninflamedSource of collagen Hyp contributionFigure 3.23: ELISA results using a polyclonal antibody to type I collagen. Type I collagenderived Hyp contributions to crevicular fluid from inflamed and noninflamed sites, compared tototal Hyp content of crevicular fluid. Error bars represent the standard error of the mean.158Inflamed^ NoninflamedPeriodontal site categoryFigure 3.24: Moles {Exp -12 } of type I collagen and Clq, per .t1 of crevicular fluid, in bothinflamed and noninflamed periodontal sites. Error bars represent the standard error of the mean.159300U Type I collagenO C1q200 -100 -Inflamed^ NoninflamedPeriodontal site categoryFigure 3.25: Moles of Hyp {Exp -12} contributed by type I collagen and Clq, per ill ofcrevicular fluid, from inflamed and noninflamed periodontal sites. Error bars represent thestandard error of the mean.160161Inflamed^NoninflamedPeriodontal site categoryFigure 3.26: Crevicular fluid Hyp contributions from type I collagen and Clq from inflamed andnoninflamed periodontal sites as measured by ELISA. The 100 percent level corresponds to thelevel of Hyp determined by HPLC.3 . 4 IMMUMOLOGICAL FINDINGS IN CREVICULAR FLUID FROMINFLAMED AND NONINFLAMED PERIODONTAL SITES3. 4. 1 SDS/PAGE gelsNative and denatured type I collagen and Clq preparations were separated by SDS/PAGE gelsto determine banding profiles for these components. Figure 3.27 shows silver stained gel profilesof native and denatured forms of these components. Alternating denatured and native samples areshown for 0.5 gg of rat tail type I collagen (lanes 2, 3, 9, 10), 0.5 1.tg human placental type I162collagen (lanes 4, 5, 11, 12), and 0.2 lag human Clq (lanes 6, 7, 13, 14). Typical profiles fornative type I collagen exhibiting (31,1, (31,2, al and a2 bands are seen in lanes 3, 5, 10, and 12.Denatured samples demonstrate a significant loss in these bands, particularly (31,1 and (31,2, andan emergence of numerous bands with molecular weight below 100K. A typical strong band fornative AB chain of Clq at 66K is present in lanes 7 and 14, while a prominent band typical fordenatured AB chain of Clq at 39K is present in lanes 6 and 13.A SDSIPAGE profile of native and denatured pooled GCF from inflamed and noninflamedsites is shown is Figure 3.28. Pooled GCF from two groups of patient matched inflamed andnoninflamed sites are represented. From a group F designated patient pool (lanes 7-10), native anddenatured GCF samples exhibited more intense and diverse banding for inflamed than noniflamedsites. This same trend was also found in a group E designated patient pool (Lanes 11-14).3.4.2 Western BlottingA western blot of anti-human type I collagen and anti-human Clq reacted to duplicate type Icollagen and Clq standards are shown in Figure 3.29. When lanes 1 through 7 were developedusing anti-human type I collagen antibody, distinct reactions to type I collagen (lanes 4-7) but noneto Clq (lanes 2, 3) were seen for both native and denatured preparations. Similarly, when lanes 9through 15 were developed using anti-human Clq antibody, distinct and intense reactions wereobtained to Clq (lanes 9, 10) with barely detectable reaction to native type I collagen (lanes 11,13).Western blot analyses were performed on denatured pooled GCF from four groups of patientmatched inflamed and noninflamed sites. These substrates were developed using either anti-humantype I collagen (Figure 3.30) or anti-human Clq (Figure 3.31). In anti-collagen developed blots,all four patient matched groups exhibited intense reactions for type I collagen with more intensereaction for small peptides in inflamed than noninflamed sites. In anti-Clq developed blots, allfour patient matched groups exhibited extremely weak reactions for Clq in comparison to the Clqstandard (lanes 7, 14).3 4 5 6 7 8 9 10 11 12 13 14 15OMMON.111.4E=•••Nme^ 11110.1111• Ore^•^4111.1111-163.111111. ONO-41•■••11111.Figure 3.27 SDS/PAGE gel of type I collagen and Clq standards. Arrows indicate molecularweight markers, from top to bottom, respectively: 200,000 D (myosin H chain), 97,400 D(phosphorylase b), 68,000 D (bovine serum albumin), 43,000 D (ovalbumin), 29,000 D (carbonicanhydrase).Lane 1:^High molecular weight standardLane 2/9: Denatured (D) rat tail type I collagenLane 3/10: Native (N) rat tail type I collagenLane 4/11: D - human placental type I collagenLane 5/12: N - human placental type I collagenLane 6/13: D - human ClqLane 7:114 N - human ClqLane 8:115 High molecular weight standard1 2 3 4 5 6 7 8 9 10 11 12 13 14 15III^N^Poi loot• 4•■•■••■••IMMO164$•I•-SDI'!--*to\"a\" 4.11110 41111111P^yip tab Imp .4„Figure 3.28 SDS/PAGE gel of native and denatured patient matched pooled GCF from inflamedand noninflamed periodontal sites. Arrows indicate molecular weight markers, from top tobottom, respectively: 200,000 D (myosin H chain), 97,400 D (phosphorylase b), 68,000 D(bovine serum albumin), 43,000 D (ovalbumin), 29,000 D (carbonic anhydrase).Lane 1: High molecular weight standardLane 2: Denatured (D) human ClqLane 3: Native (N) human ClqLane 4: N - human placental type I collagenLane 5: D - human placental type I collagenLane 6 High molecular weight standardLane 7 Patient group F: D - noninflamed GCFLane 8: Patient group F: N - noninflamed GCFLane 9: Patient group F: D - inflamed GCFLane 10: Patient group F: N - inflamed GCFLane 11: Patient group E: D - noninflamed GCFLane 12: Patient group E: N - noninflamed GCFLane 13: Patient group E: D - inflamed GCFLane 14: Patient group E: N - inflamed GCFLane 15: High molecular weight standard1651^2^3^4^5^6^7^9 10 11 12 13 14 15ONOFigure 3.29 Western blot of type I collagen and Clq standards. Arrows indicate molecular weightmarkers, from top to bottom, respectively: 200,000 D (myosin H chain), 97,400 D(phosphorylase b), 68,000 D (bovine serum albumin), 43,000 D (ovalbumin), 29,000 D (carbonicanhydrase).Lane 1: High molecular weight standardLane2: anti-human type I collagen reacted with Native (N) human ClqLane 3: anti-human type I collagen reacted with Denatured (D) human ClqLane 4: anti-human type I collagen reacted with N - human placental type I collagenLane 5: anti-human type I collagen reacted with D - human placental type I collagenLane 6: anti-human type I collagen reacted with N - type I rat tail collagenLane 7: anti-human type I collagen reacted with D - type I rat tail collagenLane 8: High molecular weight standardLane 9: anti-human Clq reacted with N - human ClqLane 10: anti-human Clq reacted with D - human ClqLane 11: anti-human Clq reacted with N - human placental type I collagenLane 12: anti-human Clq reacted with D - human placental type I collagenLane 13: anti-human Clq reacted with N- type I rat tail collagenLane 14: anti-human Clq reacted with D - type I rat tail collagenLane 15: High molecular weight standard1661^2^3^4^5^6^7^8^9 10 11 12 13 14 15!!!^ riirfFigure 3.30 Western blot of anti-human type I collagen reacted with denatured patient matchedpooled GCF from inflamed and noninflamed sites. Arrows indicate molecular weight markers,from top to bottom, respectively: 200,000 D (myosin H chain), 97,400 D (phosphorylase b),68,000 D (bovine serum albumin), 43,000 D (ovalbumin), 29,000 D (carbonic anhydrase).Lane 1: High molecular weight standardLane 2: Patient group A: Inflamed GCFLane 3: Patient group A: Noninflamed GCFLane 4: Patient group B: Inflamed GCFLane 5: Patient group B: Noninflamed GCFLane 6: Human placental type I collagenLane 7: Human ClqLane 8: High molecular weight standardLane 9: Patient group E: Inflamed GCFLane 10: Patient group E: Noninflamed GCFLane 11: Patient group F: Inflamed GCFLane 12: Patient group F: Noninflamed GCFLane 13: Human placental type I collagenLane 14: Human ClqLane 15: High molecular weight standard1671^2^3^4^5^6 7^8^9 10 11 12 13 14 15Figure 3.31 Western blot of anti-human Clq reacted with denatured patient matched pooled GCFfrom inflamed and noninflamed sites. Arrows indicate molecular weight markers, from top tobottom, respectively: 200,000 D (myosin H chain), 97,400 D (phosphorylase b), 68,000 D(bovine serum albumin), 43,000 D (ovalbumin), 29,000 D (carbonic anhydrase).Lane 1: High molecular weight standardLane 2: Patient group A: Inflamed GCFLane 3: Patient group A: Noninflamed GCFLane 4: Patient group B: Inflamed GCFLane 5: Patient group B: Noninflamed GCFLane 6: Human placental type I collagenLane 7: Human ClqLane 8: High molecular weight standardLane 9: Patient group E: Inflamed GCFLane 10: Patient group E: Noninflamed GCFLane 11: Patient group F: Inflamed GCFLane 12: Patient group F: Noninflamed GCFLane 13: Human placental type I collagenLane 14: Human ClqLane 15: High molecular weight standard1684 . DISCUSSION4.1 VOLATILE COMPOUNDS IN THE ORAL CAVITYThis study demonstrates that it is possible to collect and analyze volatile sulphur compoundsfrom specific gingival crevicular sites, and to compare their content to that of mouth air. This wasaccomplished through the development of an apparatus for collection and total retention of volatilecompounds found in crevicular air. The study provides clear evidence that both the compositionand content of sulphides from these two sources are distinct.In contrast to sulphide levels of mouth air that were determined by other investigators underrestricted conditions, the differences in crevicular air sulphides for the various parameters in thisstudy were achieved on relatively low levels of VSC that were collected without any restrictions.Hence, their levels of VSC of mouth air are significantly greater than the concentrations found inthe patients used in the present study. Our results indicate that it is possible to obtain statisticallysignificant results regardless of the levels of VSC found in any patient. The fact that thecomparisons were performed on the ratio of low amounts of sulphides found in the gingivalcrevice, underscores the sensitivity and significance of this comparison.Furthermore, the studies of mouth air in periodontal patients also utilized an optimum samplingtime for patient breath. Yaegaki (Yaegaki 1990; Yaegaki 1991), in his mouth air studies ofperiodontal subjects, assayed mouth air in the early morning under restricted conditions, which isknown to yield the highest levels of sulphides (Tonzetich 1978). In the present study, the patientswere sampled throughout the day, and were not restricted in eating, drinking or exercising oralhygiene procedures prior to testing, as is generally the case in mouth air studies. Hence, the levelsof mouth air volatiles in this study are considerably lower than seen in controlled early morningmouth air studies. It is interesting that the ratio for CH3SH to H2S in crevicular air of controlsubjects (.37 +1- .18SE) is similiar to the CH3SH to H2S ratio in mouth air of the control group(.5) observed in another investigation (Yaegaki 1990). However, this sulphide ratio in test groupsfor the mouth air study of periodontally involved subjects (4.6) is quite different from that found in169crevicular air of either inflammed (.95 +/- .21SE) or deep (.71 +/- .16SE) gingival sites. The factthat the tongue is the principal source of mouth air sulphides in periodontal patients and that mouthair analyses were performed on early morning samples provides a feasible explanation for thedisparity of the test group ratio. Yaegaki and Sanada found that the VSC production on tongues ofpatients with periodontal disease was significantly higher than those of controls, and that theCH3SH / H2S ratio was significantly reduced by removal of the tongue coatings (Yaegaki andSanada 1991). From this they concluded that salivary putrefaction did not substantially contributeto the elevated ratio of mercaptan in mouth air and that in addition to periodontal involvement,tongue coatings play an important role in VSC production leading to an elevated concentration ofmethyl mercaptan in periodontally involved individuals.It is noteworthy that the VSC content of an individual periodontal site is dependent upon thedegree of gingival inflammation. Other investigators have shown that H2S and CH3SH content inmouth air correlated with the incidence and depth of periodontal pockets in excess of 3 mm. Inaddition, implementation of periodontal therapy consisting of gingival curettage and periodontalsurgery was found to reduce levels of both H2S and CH3SH of mouth air (Tonzetich and Spouge1979). Furthermore, Kaizu and coworkers, found that in a group of periodontal subjects withhalitosis, that the CH3SH content in mouth air was observed to have a higher correlation withgingival inflammation but showed no correlation with pocket depth or bone loss (Kaizu et al.1978). The results of the current study are the first known to date that compare gingivalinflammation with the volatile sulphur content of a specific periodontal site.With regard to sulphide ratios, it is noteworthy that CH3SH to H2S was the most significantratio investigated. A recent study by Yaegaki on the composition of H2S and CH3SH in mouth airof periodontally involved subjects, found that not only were the levels of CH3SH greater thanthose of H2S, but also that the ratio of CH3SH to H2S was significantly higher (p <.05) than for acontrol group (Yaegaki and Sanada 1991). Compared to the total methyl sulphide to H2S ratio, thepresent study found that the CH3SH to H2S ratio was the only sulphide ratio in crevicular air thatgave a statistically significant result (p <.01). Contributing factors to this result were frequencies170and amounts of dimethyl sulphide and dimethyldisulfide whose levels were below that of themethyl mercaptan. In addition, the levels of (CH3S)2 in deep and inflamed sites showed morevariation in frequency and amount than did the levels of CH3SH. The resulting large standarddeviations of the means gave rise to levels of (CH3S)2 in control and test systems that were notsignificantly different (p>.05). The fluctuation and weak appearance of this sulphide are furtherevidence that methyl mercaptan is a preferred component to use for sulphide comparisons.Although CH3SH occurred in all but one of the examined crevicular sites, the levels in controlsubjects were well below that for the corresponding test group. As the levels of both H2S andCH3SH are dependent upon a number of factors such as microbial activity and sampling time, itseems appropriate to compare the ratios of these sulphides as an indicator of the composition ofVSC in the gingival crevice.A related investigation which measured the levels of H2S production in GCF concluded that thelevels of H2S correlated with both the gingival index and the crevicular fluid volume (Solis-Gaffaret al. 1980). This conclusion is in agreement with the present study, which shows that the levelsof total sulphides is greater in inflamed over noninflamed sites. These investigators measured H2Safter each filter paper strip containing crevicular fluid was incubated with L-cysteine for 3 days.They reported that the generation of H2S in the GCF could be due to either the presence or gram-negative organisms or cysteine desulfhydrase activity. In the present study, the levels of varioussulphides were measured directly on individual crevices, as they existed in vivo.The generation of higher levels of sulphides in deeper and inflamed gingival crevices overcontrol systems suggests that such an environment is conducive for establishment of gram-negativeanaerobic pathogens. Because H2S and CH3SH are known to be both directly and indirectlydetrimental to mucosal tissues (Ng and Tonzetich 1984), the measurement and characterization ofthese volatiles achieved in this investigation could be a useful method for monitoring the initiationand activity of periodontal disease.Not only can VSC serve as indicators for distinguishing inflamed from noninflamed sites, butthey have been shown to have significant effects on periodontal tissue. There are several171mechanisms whereby the VSC can act on periodonal connective tissues. They can increase thepermeability of the mucosa, change cell metabolism, and directly and indirectly alter the structureof the tissue components.Ng and Tonzetich have shown that the permeability from epithelial surface to underlyingconnective tissue is potentiated by exposure of the epithelial surface to thiols (Ng and Tonzetich1984). They observed that after exposure to mercaptan, both small ions and molecules such asPGE2 penetrated more readily across the mucosa, thus providing indirect evidence for disruptionof basement membrane components. Subsequently it has been shown that thiols degrade FN,laminin, and type IV collagen, all which are known to possess disulfide linkages, as evidened byanalyses using SDS/PAGE gels (MacKay et al. 1989). Specifically it was demonstrated thatmercaptan directly disrupted these known basement membrane components.Once volatile sulphur compounds penetrate the epithelial barrier, they can then react with theunderlying connective tissue and accompanying cells. Direct effects of VSC on collagen have beendemonstrated. Exposure to either H2S or CH3SH of type I collagen resulted in the conversion ofsome of the mature fibrillar (Tonzetich and Lo 1978) and acid-soluble preparations (Johnson andTonzetich 1985; Johnson et al. 1985) to a neutral salt-soluble product. Analyses of these productsrevealed that radiolabelled H25 was associated with al, 013, P1,2, and a2 chains characteristic oftype I collagen. Furthermore, it was determined that thiol groups (-SH) reacted with 2 or moreactive sites on the molecule (Tonzetich and Johnson 1986). This provided evidence that thiolsdirectly disrupt extracellular collagen, which may make it more susceptible to enzymaticdegradation and contribute to increased collagen destruction observed in periodontal disease.Methyl mercaptan was also found to have a more adverse effect on fibroblast cell metabolism.In the presence of 10 ng of H2S or CH3SH/ml air/CO2, a 35% and 36% reduction in prolinetransport, respectively, was found in fibroblast cell cultures (Tonzetich et al. 1985). Furtheranalyses using vital cell staining indicated damage to cells exposed to CH3SH. Since the uptake ofamino acids is a membrane-associated phenomenon, disruption of cell membranes would affectprotein synthesis.172Other detrimental cellular effects of VSC have recently been described (Johnson et al. 1992).Exposure of fibroblast cell cultures to CH3SH resulted in a 25% decrease in total protein content,and a 44% supression of DNA synthesis. There was also a reduction in type III collagen andalmost a complete absence of type III procollagen. In addition, type I collagen content wasreduced, whereas type I procollagen and/or type I protrimer was increased two-fold. These resultsdemonstrated the adverse effects of CH3SH on fibroblast cells.Methyl mercaptan has also been shown to potentiate collagenolytic proteases, IL-1 and cAMPproduction by human gingival fibroblast cultures. Cathepsin B activity is increased by 20% whilecAMP content of CH3SH exposed cultures is increased by 34% (Tonzetich et al. 1990).CH3SH was also shown to activate the immune system. Recent investigations by Ratkay andTonzetich, have shown that exposure of T lymphocytes to CH3SH results in a 30% increase in IL-1 production (Ratkay and Tonzetich 1992). In addition, exposure of fibroblasts to CH3SHincreased collagenase production. Thus, it is possible that CH3SH potentiates collagenaseproduction either by interacting directly with the fibroblast cell, and/or via an IL-1 mediatedpathway. It has been shown that IL-1 produces more collagenase and plasminogen activator,which under the influence of CH3SH creates plasmin which converts procollagenase tocollagenase. Richards and Rutherford have shown that recombinant IL-1 added to fibroblasts fromthe periodontal ligament and gingival connective tissue will induce the production of prostaglandinE2 (PGE2) and collagenase (Richards and Rutherford 1988). Since IL-1, PGE2 and collagenaseeach play a role in tissue degradation, CH3SH is seen here to potentiate the pathogenesis ofperiodontal disease.Although thiols are considered inhibitory to collagenase we observed here that they also causean increase in collagenase production. It is known that collagenase can be activated byorganomercurials which may change the molecular conformation of the enzyme, indicating that thethiol-reacted collagenase can be reactivated. Whereas thiol groups activate bacterial collagenase,they inhibit mammalian collagenase. For example, cysteine normally is used to activatecollagenase in bacterial systems.173The same situation is analogous to production of a2 macroglobulin (a2-MG) in periodontalpockets. Although one would expect increased inactivation of collagenase by increased levels ofa2-MG in periodontal pockets, in reality, higher levels of active collagenase in periodontal pocketswas observed (Uitto and Raeste 1978). It was postulated that there are other mechanisms, such asaction of proteolytic enzymes, which cleave the a2-MG-collagenase complex to free collagenase(Uitto and Raeste 1978)The present study is the first known successful attempt to collect and analzye VSC fromindividual periodontal sites using a non-invasive sampling technique. Thiols have been shown todisrupt basement membrane components and interstitial collagen molecules and stimulate theimmune system. Exposure of fibroblast cell cultures to thiol compounds increases both IL-1 andcAMP production. These in turn increase the production of proteases, such as collagenase,elastase and plasminogen activator. These enzymes would lead to the destruction of collagen. Thepresence of VSC in periodontal pockets parallels the pattern of establishment of gram-negativemicororganisms (Tonzetich and McBride 1981). These collagenase-producing organisms arestrong producers of VSC which probably play a role in activation of their proteolytic enzymes.4.2 HYDROXYPROLINE AS A MEASURE OF PERIODONTAL TISSUEACTIVITYThe initial question was: can collagen breakdown products be detected at periodontal sites thatare undergoing disease alteration? Since collagen makes up most of the protein in theperiodontium, it was reasonable to study its metabolism. Collagen is known to have a uniqueamino acid composition, specifically the presence of Hyp and Hylys. A further question arises:can Hyp be quantified in periodontal sites?; furthermore, can Hyp levels be correlated with activeperiodontal breakdown? The hypothesis of this study was that increased levels of Hyp occur atperiodontal sites undergoing active breakdown. It was conjectured that increased Hyp levels inGCF would reflect the amount of collagen metabolism in periodontal connective tissue.174Analyses of Hyp in body fluids and tissue have been used as measures of collagen metabolismin pathological states. Urinary Hyp excretion was found to be increased in a group of 11 subjectsduring phases of mild vitamin C deficiency (Hevia et al. 1990). In a study of 33 patients withchronic liver disease, both the urinary excretion of Hyp and the hepatic content of Hyp wereincreased in relation to the severity of the liver disease (Yamada et al. 1989).At the time of our investigation the most recent study of Hyp in GCF was reported by Millerand co-workers (Miller et al. 1982). They used HPLC employing pre-column derivatization withdansyl chloride to measure Hyp content in gingival exudate. They found that the Hyp contentranged from 300 to 1480 ng/mg exudate weight.In testing this dansyl chloride pre-column derivatization method, we observed that the standardHyp peak small and that it gave low responsiveness and variably appeared at retention times greaterthan 30 minutes. The same limitations using dansyl chloride were also found by otherinvestigators who demonstrated that the Hyp peak was difficult to detect, had a low responsivenessand was very close to the glutamine peak (Wiedmeier et al. 1982).A recently developed HPLC method in our laboratory utilized derivatizations with PITC aloneor in combination with OPA to detect both Hyp and Pro in biological materials (Yaegaki et al.1986). In the crevicular fluid analysis, it was found that a single derivatization step using PITCwas sufficient and reliable to separate Hyp from other amino acids. This one step techniqueyielded Hyp as a cleanly separate peak. However, the proline peak in the chromatogram remainedunresolved and could only be separated from alanine using a combined OPA/PITC derivatizationtechnique.The results of Hyp analysis of GCF showed that there was more Hyp in hydrolyzed than inunhydrolyzed crevicular fluid. There are also more amino acids in general as evidenced bycomparing profiles of hydrolyzed versus unhydrolyzed GCF. This observation implied that inGCF there was less Hyp in a free amino acid form than in bound or peptide form. This finding isin agreement with Hyp analyses of urine using HPLC. Hughes and co-workers demonstrated thatfree urinary Hyp levels are substantially lower (2-29 gmoles/24 hrs) than total Hyp urinary levels175(122-374 pmoles/24 hrs) in healthly individuals (Hughes et al. 1986). Therefore, all subsequentHyp analyses were performed on hydrolyzed samples of GCF.In the spiramycin data, Hyp levels were presented as either total moles of Hyp per GCF sampleor in moles Hyp/µl of GCF. These units were used to compare the values of total Hyp in a givenGCF sample versus Hyp level per unit volume. Comparisons with periodontal measurementswere performed using both types of Hyp values.The total Hyp content was generally higher for inflamed than for noninflamed sites at each timepoint in the study. This trend was seen in spiramycin and non-drug treatment groups, but was notobserved for weighted Hyp values. At week zero, total moles of Hyp in each sample were foundto be greater in inflamed (23.4 +/- 3.4 picomoles) then in noninflamed (16.4 +/-2.1 picomoles)sites. However, when weighted, Hyp values at inflamed sites (50.7 +/-6.9 picomoles) wereslightly less than those at noninflamed (57.2 +/-11.8 picomoles) sites. None of these differenceswere large enough to be statistically significant (p>.05).When Hyp values were examined in both treatment groups throughout the study regardless ofclinical parameters, a distinct trend was evident. For weighted Hyp values, in comparison to thebaseline, Hyp content in the spiramycin treated group was consistently higher than the values inthe non-drug group. This was an important finding as it was later found that overall increases inweighted Hyp levels were present at week 12 for healing periodontal sites. However, total Hyplevels for all examination points for both drug and non-drug treatment groups showed no particulartrend.At week zero both weighted and total Hyp values were higher in inflamed than in noninflamedsites of pockets < 4 mm, whereas these Hyp measures were both similar in pockets 4 mm.When only pocket depths were considered both weighted and total moles were larger for pocketsdepths < 4 mm. However, none of these differences were statistically significant (p>.05).Hyp content at inflamed and noninflamed periodontal sites of patients in the 24 weekspiramycin study exhibited fluctuations at given time points (0, 2, 8, 12 and 24 weeks) andbetween time points throughout the study. These fluctuations were seen with both types of Hyp176measurements. Similar fluctuations were also seen in weekly measurements of Hyp by Svanbergin a 5 week experimental gingivitis study using a beagle dog model (Svanberg 1987a). Thisindicated to him that collagen metabolism was not a linear process. In a further beagle dog studyusing a 9 day ligature-induced periodontitis model, Svanberg found that collagen derived Hypcontent was maximal four days after removal of the ligature. In our 24 week spiramycin study thevariations of Hyp levels did not follow any distinct pattern. It was conjectured that since the Hypvalues were derived from different sites at different time points, they were not being compared onan equal basis. Since there were no significant trends in Hyp values seen in either the spiramycinor non-drug treated group, this suggested that absolute values of Hyp content were not significantand useful for determining either the present disease state or the disease active state. Therefore,one may suggest that the data which would represent matched sites for different time points in thestudy may give better comparisions for Hyp levels.Analyses of the data showed that the most complete matched data was available for 17 patientsat examination points 0 and 12 weeks. Analyses of the 127 periodontal sites studied indicated thatHyp values correlated to bleeding on probing and attachment level changes. Hyp levels at weektwelve in periodontal sites that were inflamed at week zero and remained inflamed at week 12 werehigher than in sites that remained noninflamed. This relationship was stronger for weighted Hypvalues than for total Hyp content of the GCF sample. In the spiramycin treated group a significantdifference was found between Hyp levels at week 12 in these inflamed and noninflamed sites(p<.05). These results revealed it was necessary to examine the same site at different time pointsin order to detect and relate changes in Hyp content. Previous analysis of the entire data, notmatched for site and patient, demonstrated non-significant fluctuations in Hyp content.In this same group of 17 patients it was also determined that Hyp content was higher in sitesthat exhibited healing than sites that remained unresolved. This relationship was found to be truefor weighted Hyp measures and not for total Hyp content found in the GCF sample. Healing wasdefined by a gain in attachment by ?_2 mm. Again, in the spiramycin treated group, this result wassignificant (p<0.5).177This result was in contrast to our preliminary observations of the data comparing Hyp contentwith attachment level changes. Preliminary analysis of data in the spiramycin study revealed that again in attachment between zero and subsequent examination intervals was accompanied by adecrease in Hyp content (Coil et al. 1987). These changes in Hyp levels were considered of greatimportance as it was anticipated that they precede changes in attachment level. However, this wasonly preliminary data and mainly observed changes in attachment level between weeks zero andweek two of the study. This fall in Hyp content and improvement in attachment level wasconsidered healing.In Svanberg's ligature-induced periodontitis model in beagle dogs, it was also seen that 2weeks following removal of ligature the Hyp values were significantly less than when the ligaturewas in place. This decrease in Hyp content was observed to persist for the remaining 4 weeks ofthe study. The Hyp values in this post-ligature phase were considered to represent collagen-derived Hyp, since total GCF Hyp was subtracted from serum Hyp. Thus, the maximal Hyp levelat the time of ligature removal represents maximal collagen breakdown, whereas subsequent weeksindicate that collagen degradation is not as prominent.Although no clinical measurements of the periodontium were presented in Svanberg'sinvestigation, it is plausible that the four week period after ligature removal represents a stabilizingperiodontal phase. Hyp levels have decreased as there is expected to be a 70% loss in collagen inthe early destructive phase of periodontal disease (Page and Schroeder 1976). It is conjectured thatafter the stabilizing phase the periodontium would shift to a healing phase. This would be seen asimproved attachment level, increased collagen production and increased collagen turnover. Thelatter would be reflected in a rise in Hyp content at healing sites. In the fourth and last week of theSvanberg study, the Hyp content was found to be the same or higher than the value at week 3 at50% of the sites. This possibly represents a shift from a stabilizing phase to a healing phase.It has been demonstrated that as much as 2 mm differences in probing attachment can occur dueto inflammation of the junctional epithelium (Listgarten 1980). If early changes in attachment leveloccur, it is expected that they are due to improvements in the junctional epithelium, and do not178represent changes in functional attachment levels. Thus it is possible to observe early changes inattachment level measurements after therapy, which, in fact, only reflect changes in theinflammation of the JE. This phenomenon along with the decreased levels of Hyp found bySvanberg up to 4 weeks after ligature removal would explain why prelimary results during earlyexamination points of the spiramycin study detected improvements in attachment levels withcorresponding decreased in Hyp content.As mentioned previously, our results showed that Hyp was increased in periodontal sites thatexhibited healing as evidenced by a gain in probing attachment levels of 2mm. This finding wasunexpected, and differed from our preliminary observations, where it was hypothesized that higherHyp values would be found in clinically active sites. It was conjectured that sites that wereundergoing periodontal breakdown would exhibit higher Hyp levels than other sites due to loss ofconnective tissue collagen. In fact, these sites which were losing probing attachment of ?. 2mmover the twelve week period exhibited Hyp levels that were similar or less than sites thatdemonstrated no clinical change in attachment level. The finding that Hyp levels were higher inperiodontal sites that were healing than sites that were unresolved was statistically significant inpatients receiving spiramycin therapy.In the spiramycin study it was of interest that weighted Hyp levels were highest in healingsites, as compared to those sites that either remained the same or even lost periodontal attachmentover the twelve week time period. It was conjectured that the relative increase in Hyp levels wasreflected by a higher turnover of collagen in healing sites. At all periodontally affected sites it wasexpected that 70% of the collagen was already lost during inflammatory periodonal disease(Narayanan and Page 1983). Once this collagen was lost, there would be relatively less collagenavailable for breakdown, due to a low amount present. This may provide an explanation as to whyHyp levels were lower in non-healing sites since they may represent a turnover of less collagenthan that occurring in a healing site.This result is in agreement with a recent investigation examining collagen metabolites inpatients undergoing periodontal therapy. Talonpoika and co-workers examined procollagen179aminoterminal propeptide (PIIINP) levels in GCF from periodontal patients before and aftertreatment (Talonpoika and HamEldinen 1992). Using a radioimmune assay they found that therewere significantly elevated PIIINP levels after periodontal therapy, which peaked at approximately20 days post treatment and then gradually decreased back towards baseline levels by about day 40.These results indicated that the elevated PUMP levels in GCF after periodontal treatment reflectedan increased type DI collagen synthesis in the gingiva. This is not only a significant finding fromthe standpoint of showing an increased presence of a collagen fragment in healing sites, but that themetabolism of collagen was over and above that found in the preexisting inflamed periodontaltissue. This was the only other known investigation to demonstrate that a specific collagenfragment, as a result of extracellular processing, was present in the GCF in greater amounts inhealing periodontal sites than in pretreatment values.One important distinction in the spiramycin study was that the Hyp determinations and clinicalmeasurements were performed on a group of periodontitis-treated patients. The Hypmeasurements beyond week 0 are taken from periodontal sites that have been identified to have aprevious disease and have undergone treatment. Hyp values can be compared with clinicalparameters only in relation to the effect of treatment. This is in contrast to studies that evaluateindicators in groups of untreated periodontitis subjects. In such patients, when a loss ofattachment is observed, the disease process is unaltered by treatment therapies and is considered tooccur spontaneously. Here the true effect of the disease process can be compared to a particularindicator being evaluated.Some sites in the spiramycin study were observed to have lost mm of attachment within 3months of receiving therapy. They can be viewed differently than sites that had undergone clinicalsigns of attachment loss in untreated subjects, investigated in other studies. In treated sites theelements of the periodontium have been altered. Subsequent to treatment the host and microfloraadapted to these new changes. By removing etiologic factors in the periodontium, the host couldrespond. It has been shown that an inflamed periodontal site that undergoes therapy can exhibitimproved clinical probings values due to healing of the junctional epithelium (Listgarten 1980).180The next study involved the analyses of GCF from untreated patients who exhibited aninflamed and noninflamed site. The results indicated Hyp content was higher in inflamedperiodontal sites than in noninflamed sites compared within the same patient. The difference wasstatistically significant (p<0.001) and in agreement with the spiramycin study of weighted Hypmeasures at inflamed and noninflamed sites at weeks 0 and 12. Hence, both studies showed thatHyp can be used as an indicator of periodontal inflammation provided it is compared to healthy sitemeasurements within the same patient.Connective tissue degradation occurs as part of the active disease process. Monitoring forproducts of degradation would seem a logical means of following the course of the disease.Factors such as bacterial enzymes, and immunological factors are dependent upon the hostresponse. In order to shift the balance away from a stable host response to a disease state, achange is required. The effect on connective tissue would represent shifts in host-parasite balance.Changes in connective tissue metabolism would reflect the status of the disease state.A large number of indicators have been claimed to identify an active state of disease. Actuallyit is more correct to state that they represent a change that has occurred in the periodontium sincethe last examination point. It may be that the site has already undergone its destructive phase, andthat the marker is merely an indication that the destruction has taken place. If, however, furtherclinical destruction is seen at the next time interval then this indicator should be considered to haverepresented a disease active state at the time of the previous examination. Furthermore, such anindicator would be considered a predictor, since it preceded clinical changes, such as futureattachment level changes.When evaluating a potential indicator of periodontal disease, one must consider that itsusefulness will depend on whether it can determine or predict a particular state in a given patient.A good marker should differentiate a single positive test result for a condition in one individualfrom a test result in a control patient without the condition. For example, the range of values for aproposed indicator in both active and inactive disease states may overlap, and hence, for a givenmeasured value for an individual, it would not be possible to predict with certainty his/her clinical181outcome. A good diagnostic marker should be quantitative and with reasonable confidencediscriminate between one clincial outcome versus another.There has been some controversy as to the detection of active disease sites using cutoff pointsin the measurements of changes in attachment level. Cut off points of mm for periodontaldisease activity considerations are discriminations that only recognize that changes in attachmentlevel beyond that cutoff point are considered significant. This is thought to be an underestimationof actual sites undergoing disease activity. By using electronic probes with features of constantprobing force and 0.1 mm resolution, it is possible to demonstrate earlier and more subtle changesin attachment level. It is expected that longitudinal monitoring of attachment level would detectmore deteriorating sites that are actually undergoing changes in attachment level. It would bebeneficial to monitor and detect more subtle changes in attachment level, which are below thethreshold level. These sites may actually be undergoing destructive changes but go undetected andare not recognized until a considerable amount of attachment loss has occurred, which may beirreversible.Several notable advances have been made toward the development of a diagnostic test for activeperiodontal disease. Analysis of crevicular fluid for aspartate amino transferase (AST) as anindicator of cell death, has been found to be significantly increased in GCF of patients thatsubsequently undergo periodontal degeneration as indicated by loss of periodontal attachment.Two well controlled studies by Perrson and co-workers have shown that increases in GCF ASTcorrelates with loss of periodontal attachment and that GCF AST rises during experimentalgingivitis and returns to normal when gingival health is restored (Persson et al. 1990a,b; Perssonand Page 1992). These studies pose a question: Which cell death is being measured? Epithelialcells? Fibroblasts cells? Or is it simply host PMN's which are found in increased number ininflamed sites? Regardless of its source high levels of AST are a consequence of periodontalinflammation and destruction.The reason why AST appears to be a good indicator is that is monitors a by-product ofmetabolism. Although AST is an enzyme, it is an intracellular enzyme whose presence182extracellulary implies cellular disruption. Such a product of metabolism would be seen inincreased amounts during periods of high cell turnover. AST has been shown to act as a predictoras its levels were found to be increased before changes in attachment level were observedclinically. As it is considered a by-product of metabolism, concentrations of AST reflect the host-pathogen interaction balance, and indicate when a shift towards a pathogenic state is occurring.Recently Uitto and co-workers have correlated salivary elastase levels with periodontalinvolvement in patients screened at the UBC dental clinic. Their results indicate that elastaseactivity in salivary rinses of periodontally involved subjects is significantly higher than in rinsesfrom less severely involved periodontal patients and healthy controls. The sensitivity of thisenzyme test identified true positive periodontitis patients 87%, and a true negative periodontitispatients 82 % of the time. The specificity of the test produced a low false positive rate of 18% andfalse negatives of 13%. This appears to be a good screening device to detect the existence ofperiodontal disease in the oral cavity.In our studies, high levels of Hyp were found at specific periodontal sites during the diseasestate and at healing sites of treated patients. In terms of monitoring effectiveness of therapy, thisinformation would be useful in identifying those sites which do not respond to treatment. In orderto determine how Hyp levels respond in untreated patients, a further study is required that wouldmonitor Hyp levels versus changes in attachment level.4.3 CONTRIBUTION BY Clq AND COLLAGEN TO THE HYDROXYPROLINECONTENT OF GINGIVAL CREVICULAR FLUIDDuring the spiramycin study a question arose regarding the relative contributions by collagenand serum derived Clq, to the total HPLC measured Hyp content of GCF. Clq, a subcomponentof the first complement component, has a molecular weight of approximately 409,600 and contains4.3% Hyp by weight. It is found locally at the periodontal sites and binds strongly to fibrin, FNand laminin (Pear'stein et al. 1982; Entwhistle and Furcht 1988). Clq may play a role as part of a183scaffold during wound healing. Clq has been implicated to serve as an intercellular glue that actsas a matrix which is laid down at the periphery of tissue undergoing inflammation and perhapspathological processes. It has been proposed that such a Clq matrix could act like a fibrin clot,and that the formation of Clq matrix precedes the fibrin clot.In a coition to forming a matrix, Clq has also been shown to bind to cells. In vitro experimentsusing periodontally-derived fibroblasts have shown that Clq acts as a cementing substance througha cell surface receptor (Bordin et al. 1990). Studies are in progress to determine the structure ofthis receptor. Thus, Clq not only acts as a matrix, but also binds cells of the connective tissue toform a diffusion barrier.In order to determine the collagen-derived Hyp , it was necessary to remove Clq from GCF.To remove Clq from crevicular samples Svanberg (Svanberg 1987a,b) precipitated it with 0.02 Msodium acetate then removed it by centrifugation. We have also removed Clq by precipitating itwith 0.5 M acetate solutions, followed by centrifugation, which was effective in removing fromcrevicular fluid samples 5 times greater concentrations of Clq than in serum (Coil and Tonzetich1988). From these results it appears that the main source of Hyp following precipitation andcentrifugation would be from collagen and possibly degraded Clq.An additional experiment was performed to further investigate the contribution of Hyp fromboth Clq and collagen,and to confirm that higher levels of Hyp are present at inflamed periodontalsites. This study analyzed GCF collected from inflamed and noninflamed sites of 54 patients. TheHPLC results confirmed that Hyp content was significantly higher (p<.001) in inflamedperiodontal sites (684 +/- 63 picomoles/41 GCF) than in noninflamed sites (460 +/- 53picomoles/g1GCF).Aliquots of all GCF samples were examined for the presence of type I collagen using ELISAemploying polycolonal antibodies to it. The results indicated that type I collagen was present inhigher amounts in noninflamed sites than inflamed sites. A percentage of Hyp contributed by acollagen source was calculated and compared to the total Hyp level as determined from prior HPLCanalyses. It was determined that type I collagen contributed 27.7 +/-6.1% of the total Hyp content184in GCF from inflamed sites. However, noninflamed sites contributed 51.7 +1-9.6% of the totalHyp content. This difference was statistically significant (p<.05).Analyses of the same GCF for Clq was performed using ELISA employing monoclonalantibodies to Clq. These antibodies were harvested from hybridoma cells cultures, amplified inascites fluid collected from mice, and were shown to be specific for Clq. The results obtainedwith this antibody preparation were unexpected. It was found that levels of Clq varied widelyamongst the crevicular sites. When the Clq levels were determined and Hyp contribution wascalculated and compared to total levels of Hyp, it was found that in a number of cases Hypcontribution far exceeded the maximum 100% level. This was unanticipated since the monocloncalantibody was tested for specificity and cross reactivity to related molecules. This antibodypreparation was directed to the globular head region of Clq molecule and was expected to reactspecifically with Clq and not type I collagen (Kilchher et al. 1985). Therefore, these highlydiverse Clq values in GCF are believed to be due an undetermined cross reactivity. Since thisantibody was directed against the globular portion of Clq, it is possible that a cross reactivity mayhave occurred with another globular component.When the ELISA was repeated using a polyclonal antibody to Clq the results demonstrated thatthe Hyp contribution by C1q to the overall levels of Hyp was minimal. Clq accounted for 6.9 +/-1.1% of the total Hyp content in inflamed sites and 9.9 +1- 2.5% of the total Hyp content innoninflamed sites. The same trend as observed in collagen-derived Hyp also applied to Clq; ahigher contribution of Hyp in noninflamed than in inflamed sites.These ELISA results indicate that collagen is the major source of hydroxyproline in crevicularfluid with only very low amounts ascribed to Clq reactive peptides.These results pose two perplexing questions. If Hyp content was higher in inflamed sites,why were the levels determined by ELISA of type I collagen and Clq higher in noninflamed sites?And secondly, what was the nature of the Clq found in the ELISA experiment? To solve thesequestions, crevicular fluid models were used to evaluate how type I collagen and Clq wereaffected by the analyses and how they could be affected in the periodontal environment.185Crevicular fluid models were used to evaluate the various forms of type I collagen and Clq thatcould be represented in a crevicular fluid sample. Due to enzymatic activity in the periodontiumthese parent molecules were expected to be present in various stages of degradation. It wasconjectured that based upon the degree of cleavage, various sizes of fragments are present increvicular fluid. These fragments could be separated in the following ways: ones that could orcould not be precipitated in acetate (FrPPT+ or FrPPT-), and ones that could or could not react toantibodies directed against the parent molecule (FrAB+ or FrAB-). Conceptualization of whichcomponents would appear in crevicular fluid and acetate-precipitated GCF are shown in Figure4.1.The effectiveness of acetate precipitation was evaluated for both removal of Clq and theretention of type I collagen. It was found that 91.3% of Clq was removed in a 0.05 M sodiumacetate buffer followed by centrifugation at 15,600g for 30 minutes at 4°C, while the level of type Icollagen virtually was the same. This confirmed earlier precipitation reports which also removedClq using acetate buffer (Svanberg 1987a,b). Thus the amount of Clq detected by ELISA wouldrepresent only that portion that would avoid precipitation and remain reactive to antibodies. In thecase of collagen, the same type of fragments would also be present. However in the collagensystem, it was expected that the fragments would not be precipitated.The question remained: How could more collagen exist in noninflamed sites yet there be moreHyp present at inflamed sites? It was hypothesized that in inflamed sites molecules such ascollagen and Clq were degraded further to smaller peptides due to increased proteolytic activitythan in noninflamed sites, which resulted in reduced number of antigenic determinant sites. Thispossibility was investigated using type I collagen incubated with bacterial collagenase and P.gingivalis bleb preparation. The results indicated that less of the initial collagen was detected usingELISA with increasing incubation time. This demonstrated that there was a loss of antigenicdeterminants due to incubation of collagen with these proteolytic enzyme preparations.CollagenCollagen-Fr-AB(+)Collagen-Fr-AB(-)Collagen (100%)Collagen-Fr-AB(+)Collagen-Fr-AB(-)Clq (100%)^Clq (92%)Clq-FrPPT(+)AB(+)^Clq-FrPPT(+)AB(+)Clq-FrPPT(+)AB(-) Clq-FrPPT(+)AB(-)Clq-FrPPT(-)ABHClq-FrPPT(-)AB(-)Clq (8%)Clq-FrPPT(-)AB(+)Clq-FrPPT(-)AB(-)186Acetate precipitation removes Supematent after acetateWhole crevicular fluid^these components:^precipitation and centrifugationFigure 4.1: Type I collagen and Clq components in whole, precipitate and supernatant fractionsconceptualized to occur in acetate treated crevicular fluid. 'Fr'= fragment; 'PPT'= precipitation;'(AB)'= antibody; '(+)'= sensitive; '(-)'= insensitive.SDS/PAGE gels and western blot analyses provided further evidence on the composition ofinflamed and noninflamed GCF. In pooled GCF samples there were more smaller fragmentproteins present in inflamed GCF than noninflamed GCF. Western blots using antibodies to type Icollagen demonstrated that there were more smaller fragments reactive to the antibody in inflamedsites than in noninflamed sites. The same trend applied to Clq. Markedly more intense reactionswere obtained for type I collagen than for Clq western blots. These results provide a feasibleexplanation for ELISA results which indicated higher collagen and Clq content in noninflamedsites.As the results from the ELISA experiments indicate, more type I collagen than Clq exists in thesamples, hence more of the Hyp is attributed to a collagen source. The fractions of type I collagenand Clq have been considered with the aid of a crevicular fluid model. In this model it appearsthat the Clq component that was detected in GCF samples had to avoid precipitation and also be187antibody reactive. Earlier HPLC results showed that Clq at serum concentrations would give riseto barely detectable recordings for total HPLC analysis. This finding, together with the result thatClq contributed less than 10% to total Hyp in either inflamed or noninflamed GCF, indicated thatClq was not a prominent contributor to the total Hyp content of GCF.5. CONCLUSIONSThe experiments of this thesis were designed to evaluate volatile sulphur production fromgingival crevice, crevicular fluid for Hyp as an indicator of disease activity, and type I collagen andClq for their contributions to Hyp levels in GCF. A device was developed for the collection ofvolatiles from the gingival crevice, while GCF was collected from the gingival sulcus using filterpaper strips. Specific conclusions based on the analyses of these components are summarized inthe following statements:1. Volatile suphur compounds can be collected and analyzed from individual periodontal sites.2. The total suphur content in crevicular air is higher at inflamed and deep, than correspondingnoninflamed and shallow periodontal sites.3. The ratio of CH3SH to H2S in crevicular air is significantly higher at inflamed and deep, thancorresponding noninflamed and shallow periodontal sites.4. The preponderance of Hyp content of crevicular fluid is present in a peptide form, and adisproportionately lesser amount in a free unbound form.5. Since more type I collagen than Clq exists in GCF samples, the predominant source of GCFHyp is attributed to a collagen source.6. Hyp levels in GCF are reflective of the amount of collagen turnover at specific periodontalsites.7. Higher Hyp levels were found at inflamed than noninflamed periodontal sites.1888. Higher Hyp levels were present at healing periodontal sites, than those that remainedunresolved after spiramycin and scaling and root planing treatment.6. RECOMMENDATIONS FOR FUTURE STUDYIn order to determine the response of Hyp GCF levels in untreated patients, a longitudinalstudy is required to monitor its levels versus changes in periodontal attachment. Since collagenmetabolism is increased, as evidenced by elevated Hyp levels in periodontal sites that experiencedhealing after therapy, it is of interest to determine collagen metabolism at untreated periodontalsites. Measurements with high resolution, pressure sensitive periodontal probes can determinesubtle changes in attachment level. This will allow monitoring of attachment level changes thatmight otherwise be below a cutoff threshold level and go undetected. This will permit closercomparison of Hyp levels to attachment level measurements.A study requiring further investigation is the production and testing of CH3SH reacted collagenantibodies as potential markers for periodontal disease activity. In a pilot study it was observedthat polyclonal serum antibodies raised in mice by injection of CH3SH treated collagen, reactedwith GCF gave a clear differentiation between diseased and control sites (Ratkay et al. 1990).Intense ELISA reaction and western blot staining were observed for collagen in GCF fromdiseased sites and only a trace or absence of antigen response to that in GCF samples from controlsites.This result may be similar to that obtained with N-ethylmaleimide (NEM) by other investigators(Dawson et al. 1987). NEM-containing small protein fragments were shown to have strongantigenic properties. This may also be true for CH3SH bound collagen fragments. It is ofsignificance that antibodies formed from the above substances are different from the commerciallyavailable products. Antibodies from commercial sources generated by exposure of whole collagen189resulted in antibodies that are directed against the entire molecule. Hence, disruption andfragmentation of collagen would lead to loss of antigenic sites. This would explain why ELISAvalues for healthy sites showed more collagen than inflamed sites using commercially availableantibodies. As Hyp values in GCF were significantly higher in sites that showed higher levels ofmetabolic activity (inflamed and healing sites), the development of this CH3SH treated collagenantisera merits strong consideration.7. BIBLIOGRAPHYAbrams, K., J. Canton, and A. Polson (1984). \"Histologic comparisons of interproximal gingivaltissues related to the presence or absence of bleeding.\" J. Periodontol. 55: 629-632.Adams, E. and I. Frank (1980). \"Metabolism of proline and the hydroxyprolines.\" Ann. Rev. Biochem. 49: 1005-1061.Al-Joburi, W., T. Chin Quee, C. Lautar, I. Iugovaz, J. Bourgouin, F. Delorme, and E. C. S.Chan (1989). \"Effects of adjunctive treatment of periodontitis with tetracycline and spiramycin.\"LPeriodontol. 60: 533-539.Anttinen, H. and A. Hulkko (1980). \"Regulation of the glycosylations of collagen hydroxylysinein chick embryo tendon and cartilage cells.\" Biochimica et Biophysica Acta 632: 417-427.Anusaksathien, 0., G. Singh, N. Matthews, and A. E. Dolby (1992). \"Autoimmunity to collagenin adult periodontal disease: Immunoglobulin classes in sera and tissue.\" J. Perio. Res. 27: 55-61.Applegren, R., P. J. Robinson, and E. J. Kaminski (1979). \"Clinical and histologic correlation ofgingivitis.\" J. Periodontol. 50: 540-543.Attstrtim, R. (1970). \"Presence of leucocytes in crevices of healthy and chronically inflamedgingivae.\" J. Periodontol. 5: 42-47.Attstriim, R., A. B. Laurel, V. Lahsson, and A. SjOholm (1975). \"Complement factors in gingivalcrevice material from healthy and inflamed gingiva.\" J. Perio. Res. 10: 19-27.Bader, H. J. and P. Goldhaber (1966). \"The passage of intravenously administered tetracycline inthe gingival sulcus of dogs.\" J. Oral Ther. Pharmacol. 2: 324-329.Barefoot, D. H. and M. S. Silverman (1977). \"Alveolar bone loss in thymectomized goldenhamsters.\" J. Periodontol. 48: 699-704.Bartold, P. M. (1990). \"A biochemical and immunohistochemical study of the proteoglycans ofalveolar bone.\" J, Dent, Res. 69: 7-19.190Bender, I. B. and S. Seltzer (1961). \"Roentgenographic and direct observation of experimentallesions in bone: I.\" J. Am. Dent. Assoc. 62: 152-160.Bender, I. B. and S. Seltzer (1961). \"Roentgenographic and direct observation of experimentallesions in bone: IL\" J. Am. Dent. Assoc. 62: 708-716.Bensusan, H. B., T. L. Koh, K. G. Henry, B. A. Murray, and L. A. Culp (1978). \"Evidence thatfibronectin is the collagen receptor on platelet membranes.\" Proc. Natl. Acad. Sci. U.S.A. 75:5874-5868.Berg, R. A. and D. J. Prockop (1973). \"Affinity column purification of protocollagen prolinehydroxylase from chick embryos and further characterization of the enzyme.\" Liudfsl in 248:1175-1182.Berg, R. A., M. L. Schwartz, and R. G. Crystal (1980). \"Regulation of the production ofsecretory proteins: Intercellular degradation of newly synthesized 'defective' collagen.\" Proc. Natl. Acad. gLLLS,A„ 77: 4746-4750.Berg, R. A., M. L. Schwartz, L. Rome, and R. G. Crystal (1984). \"Lysosomal function in thedegradation of defective collagen in cultured lung fibroblasts.\" Biochemistry 23: 2134-2138.Berkovitz, B. K. B., B. J. Moxham, and H. N. Newman (1982). The periodontal ligament in health and disease. Ontario, Willowdale.Berman, M. D., J. G. Waggoner, J.-M. Foidart, and H. K. Kleinman (1980). \"Attachment tocollagen by isolated hepatocytes from rats with induced hepatic fibrosis.\" J. Lab Clin Med 95:660-671.Bevilacqua, M. P., D. Amrani, M. W. Mosesson, and C. Bianco (1981). \"Receptors for cold-insoluble globulin (plasma fibronectin) on human monocytes.\" J. Exp. Med. 153: 42-60.Bhaskar, S. N. (1980). Orban's Oral Histology and Embryology. C.V. Mosby, St. Louis.Bienkowski, R. S., B. J. Baum, and R. G. Crystal (1978). \"Fibroblasts degrade newlysynthesised collagen within the cell before secretion.\" Nature 276: 413-416.Bienkowski, R. S., M. J. Cowan, J. A. McDonald, and R. G. Crystal (1978). \"Degradation ofnewly synthesized collagen.\" J. Biol. Chem. 253: 4356-4363.Birkedal-Hansen, H. (1988). \"From tadpole to family of matrix metalloproteases.\" J. Oral Path. 17: 445-451.Bissada, N. F., E. M. Schaffer, and E. Haus (1967). \"Circadian periodicity of human crevicularfluid flow.\" J. Periodontol. 38: 36-40.Bjorn, A. L., G. Kock, and J. Lindhe (1965). \"Evaluation of gingival fluid measurements.\"Odont. Rev. 16: 300-307.Bordin, S., B. Ghebrehiwet, and R. C. Page (1990). \"Participation of Clq and its receptor inadherence of human diploid fibroblast.\" J. Inununol. 145(8): 2520-2526.Bornstein, P. and H. Sage (1980). \"Structurally distinct collagen types.\" Ann. Rev. Biochem. 49:957-1003.191Bowers, M.A., L.W. Fisher, J.D. Termine, and M.J. Somerman (1989). \"Connective tissue -associated proteins in crevicular fluid: potential markers for periodontal diseases.\" J. Periodontol. 60: 448-451.Bowersox, J. C. and N. Sorgente (1982). \"Chemotaxis of aortic endothelial cells in response tofibronectin.\" Cancer Res. 42: 2547-2551.Brandtzaeg, P. (1965). \"Immunochemical comparison of proteins in human gingival pocket fluid,serum and saliva.\" Arch. oral Biol. 10: 795-803.Brandtzaeg, P. (1966). \"Local factors of resistance in the gingival tissue.\" J. Perio. Res. 1: 19-42.Brandtzaeg, P. and W. V. Mann (1964). \"A comparative study of the lysozyme activity of humangingival pocket fluid, serum and saliva.\" Acta Odont. Scand. 22: 441-445.Breed, R. S., E. G. D. Murray, and N. R. Smith (1957). Burgey's manual of determinativebacteriology. Baltimore, Maryland, The Williams and Wilkins Co.Bretz, W. A., D. E. Lopatin, and W. J. Loesche (1990). \"Benzoyl-arginine naphthylamide(BANA) hydrolysis by Treponema denticola and/or Bacteroides gingivalis in periodontal plaques.\"Oral Microbiol. Immun. 5: 275-279.Brill., N. (1959). \"Influence of capillary permeability on flow of tissue fluid into gingival pockets.\"Acta Odont. Scand. 17: 23-33.Brill, N. (1960). \"Gingival conditions related to flow of tissue fluid into gingival pockets.\" ActaOdont. Scand. 18: 421-446.Brill, N. and H. Bjorn (1959). \"Passage of tissue fluid into human gingival pockets.\" ActaOdontol. Scand. 17: 11-21.Brill, N. and B. Krasse (1958). \"The passage of fluid into the clinically healthy gingival pocket.\"Acta Odontol. Scand. 16: 233-245.Brill, N. and B. Krasse (1959). \"Effect of mechanical stimulation on flow of tissue fluid throughgingival pocket epithelium.\" Acta Odontol. Scand. 17: 115-130.Brody, S. S. and J. E. Chaney (1966). \"Flame Photometric Detector: The application of a specificdetector for phosphorus and for sulfur compounds - sensitive to subnanogram quantities.\" J. GasChromatogr. 4: 42-46.Burgeson, R. E. and N. P. Morris (1987). The collagen family of proteins. Connective TissueDiseases. Molecular Pathology of Extracellular Matrix. New York, Marcel Dekker. 1st, ed. 3-28.Butler, W. T., J. E. Finch, and E. J. Miller (1977). \"The covalent structure of cartilage collagen.\"J. Biol. Chem. 252: 639-643.Campana, L. (1981). \"Lymphocyte transformation responses in gingivitis and periodontitis.\" LPeriodontol. 52: 367-373.Cancellaro, L. A., J. Kingsberg, and E. 0. Butcher (1961). \"Histochemistry of oral mucousmembrane: Total protein, sulfhydryls, disulfides, ribonucleic acid and deoxyribonucleic acid.\" LDent. Res. 40: 436-445.192Chambers, D. A., J. M. Crawford, S. Muklerjee, and R. L. Cohen (1984). \"Aspartateaminotransferase increases in crevicular fluid during experimental periodontitis in beagle dogs.\" LPeriodontol. 52: 526-530.Chambers, D. A., P. B. Imrey, R. L. Cohen, J. M. Crawford, M. E. A. F. Alves, and T. A.McSwiggin (1991). \"A longitudinal of aspartate aminotransferase in human gingival crevicularfluid.\" J. Petit). Res. 26: 65-74.Charon, J. A., T. A. Luger, S. E. Mergenhagen, and J. J. Oppenheim (1982). \"Increasedthymocyte-activating factor in human gingival fluid during gingival inflammation.\" Infect. Immun. 38: 1190-1195.Chu, M.-L., J. C. Myers, M. P. Bernard, J.-F. Ding, and F. Ramirez (1982). \"Cloning andcharacterization of five overlapping cDNA's specific for human pro al (I) collagen chain.\" NucleicAcids Res. 10: 5925-5934.Ciancio, S. G., M. L. Mather, and J. A. McMullen (1980). \"An evaluation of minocycline inpatients with periodontal disease.\" J. Periodontol. 51: 530-534.Cimasoni, G. (1983). Crevicular Fluid Updated. Basel, New York, Karger.Cleaton-Jones, P. and L. Fleisch (1973). \"A comparative study of the surface of keratinized andnon-keratinized oral epithelia.\" J. Perio. Res. 8: 366-377.Cohen, B. (1954). \"The blood supply of the gingival tissues.\" J. Periodontol. 25: 64.Cohen, R. L., M. E. A. F. Alves, J. M. Crawford, T. McSwiggin, and D. A. Chambers (1991).\"Association of gingival crevicular fluid aspartate aminotransferase levels with histopathologyduring ligature-induced periodontitis in the beagle dog.\" J. Dent. Res. 70(6): 984-987.Coil, J. and J. Tonzetich (1988). \"Hydroxyproline contribution from Clq in crevicular fluid atdiseased sites.\" J. Dent. Res. 67: 372 (Abstr # 2074).Coil, J. M. and J. Tonzetich (1984). \"Collection and analysis of volatile compounds produced inthe gingival crevice sites of humans.\" J. Dent. Res. 63: 249.Coil, J. M. and J. Tonzetich (1985). \"Volatile sulphur production at healthy and diseased gingivalcrevice sites.\" J. Dent. Res. 64: 306.Coil, J. M. and J. Tonzetich (1986). \"HPLC analysis of crevicular fluid for markers of periodontaldisease.\" J. Dent. Res. 65: 837 (Abstr # 1004).Coil, J. M., J. Tonzetich, and G. S. Beagrie (1987). \"Hydroxyproline content of crevicular fluidcollected from periodontally involved sites.\" J. Dent. Res. 66: 213 (Abstract# 1011).Comper, W. D. and T. C. Laurent (1978). \"Physiological function of connective tissuepolysaccharides.\" Physiol. Rev. 58: 255-315.Couchman, J. R., M. Hook, D. A. Rees, and R. Timpl (1983). \"Adhesion, growth and matrixproduction by fibroblasts on laminin substrates.\" J. Cell Biol. 96: 177-183.193Cox, S. W. and B. M. Eley (1989a). \"Detection of cathepsin B- and L-, elastase-, tryptase-,trypsin-, and dipeptidyl peptidase IV-like activities in crevicular fluid from gingivitis andperiodontitis patients with peptidyl derivatives of 7-amino-4-trifluoromethyl coumarin.\" 1,12gdg iRes. 24: 353-361.Cox, S. W. and B. M. Eley (1989b). \"Tryptase-like activity in crevicular fluid from gingivitis andperiodontitis patients.\" J. Perio. Res 24: 41-44.Curtis, M. A., G. S. Griffiths, S. J. Price, S. K. Coulthurst, and N. W. Johnson (1988). \"Thetotal protein concentration of gingival crevicular fluid. Variation with sampling time and gingivalinflammation.\" J. Clin. Perio. 15: 628-632.Curtis, M. A., J. A. C. Sterne, S. J. Price, G. S. Griffiths, S. K. Coulthurst, J. M. A. Wilton,and N. W. Johnson (1990). \"The protein composition of gingival crevicular fluid sample frommale adolescents with no destructive periodontitis: baseline data of a longitudinal study.\" J. Perio, Res. 25: 6-16.Davis, D. B. and P.-C. Tai (1980). \"The mechanism of protein secretion across membranes.\"Nature 283: 433-438.Dawson, K. M., C. H. J. Sear, and C. M. Moorhouse (1987). \"Production of monoclonalantibodies recognising N-ethylmaleimide during attempted generation of monoclonal antibodies tohuman type I collagen.\" Coll. Rel. Res. 7: 125-134.Deas, D. E., L. A. Pasquali, C. H. Yuan, and K. S. Korman (1991). \"The relationship betweenprobing attachment loss and computerized radiographic analysis in monitoring progression ofperiodontitis.\" J. Periodontol. 62: 135-141.Dreizen, S., E. J. Gilley, and T. D. Sipes (1956). \"A comparison of the prevailing cell types insaliva in persons with and without periodontal disease.\" Oral Surg. 9: 278-283.Dzink, J. C., S. S. Socransky, and A. D. Haffajee (1988). \"The predominate cultivable microbiotaof active and inactive lesions of destructive periodontal diseases.\" J. Clin. Perio. 15: 316-323.Ebersole, J. L., M. A. Taubman, and D. J. Smith (1982). \"Humoral immune responses anddiagnosis of periodontal disease.\" L 121do,20„ 17: 478-480.Egelberg, J. (1963). \"Cellular elements in gingival pocket fluid.\" Acta Odont. Scand. 21: 283-287.Egelberg, J. (1964). \"Gingival exudate measurements for evaluation of inflammatory changes ofthe gingivae.\" Odont, Rev. 15: 381-398.Egelberg, J. (1966a). \"The blood vessels of the dento-gingival junction.\" J. Perio. Res. 1: 163-179.Egelberg, J. (1966b). \"Permeability of the dento-gingival blood vessels. I. Application of thevascular labelling method and gingival fluid measurements.\" J. Perio. Res. 1: 180-191.Egelberg, J. (1966c). \"Permeability of the dento-gingival blood vessels. II. Clinically healthygingivae.\" J. Perio. Res. 1: 276-286.Egelberg, J. (1966d). \"Permeability of the dento-gingival blood vessels. III. Chronically inflamedgingivae.\" J. Perio, Res. 1: 287-296.194Egelberg, J. (1966e). \"Permeability of the dento-gingival blood vessels. IV. Effect of histamine onvessels in clinically healthy and chronically inflamed gingivae.\" J. Perio. Res. 1: 297-302.Egelberg, J. and R. AttstrOm (1973). \"Comparison between orifice and intracrevicular methods ofsampling gingival fluid.\" J. Perio. Res. 8: 384-388.Eley, B. M. and S. W. Cox (1992). \"Cathepsin B/L-, elastase-, tryptase-, trypsin- and dipeptidylpeptidase IV-like activities in gingival crevicular fluid: Correlation with clinical parameters inuntreated chronic periodontitis patients.\" J. Perio. Res. 27: 62-69.Embery, G., W. M. Oliver, J. B. Stanbury, and J. A. Purvis (1982). \"The electrophoreticdetection of acidic glycosamino-glycans in human gingival sulcus fluid.\" Archs oral Biol. 27: 177-179.Engvall, E., E. Ruoslahti, and E. J. Miller (1978). \"Affinity of fibronectin to collagens of differentgenetic types and to fibrinogen.\" J. Exp. Med. 147: 1584-1595.Entwhistle, R. A. and L. T. Furcht (1988). \"Clq component of complement binds to fibrinogenand fibrin.\" Biochem. 27: 507-512.Evian, C. I., E. S. Rosenberg, and M. A. Listgarten (1982). \"Bacterial variability within diseasedperiodontal sites.\" J. Periodontol. 53: 595-598.Fessler, L. I., K. Duncan, J. H. Fessler, T. Salo, and K. Tryggvason (1984). \"Characterizationof the procollagen IV cleavage products produced by a specific tumor collagenase.\" J. Biol. Chem. 259: 9783-9789.Fessler, L. I., C. Kumamoto, M. E. Meis, and J. H. Fessler (1981a). \"Assembly and processingof procollagen (AB) in chick blood vessels and other tissues.\" J. Biol Chem 256: 9640-9646.Fessler, L. I., W. J. Robinson, and J. H. Fessler (1981b). \"Biosynthesis of procollagen [(proalV)2 pro a2V] by chick tendon fibroblasts and procollagen (pro alV)3 by hamster lung cellcultures.\" J. Biol Chem 256: 9646-9651.Fessler, L. I., R. Timpl, and J. H. Fessler (1981c). \"Assembly and processing of procollagentype DI in chick embryo blood vessels.\" J. Biol Chem 256: 2531-2537.Ficara, A. J., M. P. Levin, M. F. Grower, and G. D. Kramer (1975). \"A comparison of theglucose and protein content of gingival fluid from diabetics and nondiabetics.\" J. Perio. Res. 10:171-175.Fluharty, A. L. (1982). \"The mucopolysaccahidoses: a synergism between clinical and basicinvestigations.\" J. Invest. Dermatol. 79: 38s-44s.Freedman, H. A., M. A. Listgarten, and N. S. Taichman (1968). \"Electron microscopic featuresof chronically inflamed human gingiva.\" J. Perio. Res. 3: 313-327.Ftis, A., G. Singh, and A. E. Dolby (1986). \"Antibody to collagen type I in periodontal disease.\"J. Periodontol. 57: 693-698.Gallop, P. M. and M. A. Paz (1975). \"Posttranslational protein modifications, with specialattention to collagen and elastin.\" Physiological Reviews 55: 418.195Gavin, J. B. (1968). \"The ultrastructure of the crevicular epithelium of cat gingiva.\" Am. J. Mat. 123: 283-295.Gay, S., R. K. Rhodes, R. E. Gay, and E. J. Miller (1981). \"Collagen molecules comprised ofal (V)-chains (B-chains): An apparent localization in the exocytoskeleton.\" Coll. Rel. Res. 1: 53-58.Geisenheimer, J. and S. S. Han (1971). \"A quantitative electron microscopic study ofdesmosomes and hemidesmosomes in human crevicular epithelium.\" J. Pefiodontol. 42: 396-405.Genco, R. J. and J. Slots (1984). \"Host responses in periodontal diseases.\" J. Dent. Res. 63:441-451.Ginsberg, M. H., R. G. Painter, J. Forsyth, C. Birdwell, and E. F. Plow (1980). \"Thrombinincreases expression of fibronectin antigen on the platelet surface.\" Proc. Natl. Acad. Sci. U.S.A.77: 1049-1053.Golub, L. M., S. M. Borden, and I. Kleinberg (1971). \"Urea content of gingival crevicular fluidand its relation to periodontal disease in humans.\" J. Perio. Res. 6: 243-251.Golub, L. M., H. M. Lee, G. Lehrer, A. Nemiroff, T. F. McNamara, R. Kaplan, and N. S.Ramamurthy (1983). \"Minocycline reduces gingival collagenolytic activity during diabetes:preliminary observations and a proposed new mechanism of action.\" J. Perio. Res. 18: 516-526.Golub, L. M., T. F. McNamara, G. D'Angelo, and R. A. Greenwald (1987). \"A non-antibacterialchemically modified tetracycline inhibits mammalian collagenase activity.\" J. Dent. Res. 66: 1310-1314.Golub, L. M., N. S. Ramamurthy, T. F. McNamara, B. Gomes, M. Wolff, S. Ciancio, and H.Perry (1984). \"Tetracyclines inhibit tissue collagenase activity: a new mechanism in the treatmentof periodontal disease.\" J. Perio. Res, 19: 651-655.Golub, L. M., N. S. Ramamurthy, T. F. NcNamara, R. A. Greenwald, and B. R. Rifkin (1991).\"Tetracyclines inhibit connective tissue breakdown: new therapeutic implications for an old familyof drugs.\" Crit. Rev. Oral Biol. Med. 2: 297-322.Golub, L. M., K. Stiegel, N. S. Ramamorthy, and I. D. Mandel (1976). \"Some characteristics ofcollagenase activity in gingival crevicular fluid and its relationship to gingival diseases in humans.\"J. Dent. Res. 55: 1049-1057.Golub, L. M., M. Wolff, H. M. Lee, T. F. McNamara, N. S. Ramamurthy, J. Zambon, and S.G. Ciancio (1985). \"Further evidence that tetracyclines inhibit collagenase in human crevicularfluid and from other mammalian sources.\" J. Perio. Res. 20: 12-23.Goodson, J. M., M. A. Cugini, R. L. Kent, G. C. Armitage, C. M. Cobb, D. Fine, M. E. Fritz,E. Green, M. J. Imoberdorf, W. J. Killoy, C. Mendieta, R. Niederman, S. Offenbacher, E. J.Taggart, and M. Tonetti (1991). \"Multi-center evaluation of tetracycline fiber therapy: I.Experimental design.\" ,J. Perio. Res. 26: 361-370.Goodson, J. M., A. D. Haffajee, and S. S. Socransky (1979). \"Periodontal therapy by localdelivery of tetracycline.\" J. Clin. Perio. 6: 83-92.196Goodson, J. M., A. D. Haffajee, and S. S. Socransky (1984). \"The relationship betweenattachment level loss and alveolar bone loss.\" J. Clin. Perio. 11(5): 348-359.Goodson, J. M., P. Hogan, and S. Dunham (1985). \"Clinical responses following periodontaltreatment by local drug delivery.\" J. Periodontol. 56: 81-87.Goodson, J. M., D. Holborow, R. L. Dunn, P. Hogan, and S. Dunham (1983). \"Monolithictetracycline-containing fibers for controlled delivery to periodontal pockets.\" J. Periodontol. 54:575-579.Goodson, J. M., A. C. R. Tanner, A. D. Haffajee, and S. S. Socransky (1981). \"Evidence forepisodic peridontal disease activity.\" J. Dent. Res. 60A: 387 (Abstr # 305).Goodson, J. M., A. C. R. Tanner, A. D. Haffajee, G. C. Sornberger, and S. S. Socransky(1982). \"Patterns of progression and regression of advanced destructive peridontal diesease.\" LClin. Periodont. 9: 472-481.Greenstein, G. (1984). \"The role of bleeding upon probing in the diagnosis of periodontaldisease.\" J. Periodontol. 55: 684-688.Greenstein, G., J. Caton, and A. M. Poison (1981). \"Histological characteristics associated withbleeding after probing and visual signs of inflammation.\" J. Periodontol. 52((8)): 420-425.Grinnell, F., M. Feld, and D. Minter (1980). \"Fibroblast adhesion to fibrinogen and fibrinsubstrata: requirement for cold-insoluble globulin (plasma fibronectin).\" Cell 19: 517-525.Grinnell, F., M. Feld, and W. Snell (1979). \"The influence of cold insoluble globulin on plateletmorphological response to substrata.\" Cell. Biol. hit Rep. 3: 585-592.Gryder, R. M., N. Lamon, and E. Adams (1975). \"Sequence position of 3-hydroxyproline inbasement membrane collagens. Isolation of glycyl-3-hydroxyprolyl-4-hydroxyproline from swinekidney.\" J. Biol Chem 250: 2470-2474.Guggenheim, B. G., R. Gaegauf-Zollinger, A. Hefti, and J. H. Burckhart (1981). \"The effect ofcyclosporin A on periodontal disease in rats monoassociated with actinomyces viscosus NY1.\" LPerio. Res. 16: 26-38.Haffajee, A. D., S. S. Socransky, and J. M. Goodson (1983). \"Clinical parameters as predictorsof destructive periodontal disease activity.\" J. Clin. Periodont. 10: 257-265.Hakkarainen, K., V.-J. Uitto, and J. Ainamo (1988). \"Collagenase activity and protein content ofsulcular fluid after scaling and occlusal treament of teeth with deep periodontal pockets.\" J. Perio. Res, 23: 203-210.Hakomori, S., M. Fukuda, K. Sekiguchi, and W. G. Carter (1984). Fibronectin, laminin, andother extracellular glycoproteins. Extracellular Matrix Biochemistry. New York, Elservier. 277-328.Hancock, E. B., R. J. Cray, and T. J. O'Leary (1979). \"The relationship between gingivalcrevicular fluid and gingival inflammation.\" J. Peridontol. 50: 13-19.Hancock, E. B. and M. R. Wirthln (1981). \"The location of the periodontal probe tip in health anddisease.\" J. Periodontol. 52(3): 124-129.197Hara, K. and H. LOe (1969). \"Carbohydrate components of the gingival exudate.\" J. Perio. Res. 4: 202-207.Hara, K. and T. Takahashi (1975). \"Hydroxyproline content in gingival exudate before and afterperiodontal surgery.\" J. Perio. Res. 10: 270-274.Hardingham, T. E. and H. Muir (1972). \"The specific interaction of hyaluronic acid with cartilageproteoglycans.\" Biochim. Biophys. Acta 279: 401-405.Heathcote, J. G., C. H. Sear, and M. E. Grant (1978). \"Studies on the assembly of rat lenscapsule. Biosynthesis and partial characterization of the collagenous components.\" Biochem. J. 176: 283-294.Heijl, L., G. Dahlen, Y. Sundin, A. Wenander, and J. M. Goodson (1991). \"A 4 quadrantcomparative study of periodontal treatment using tetracycline containing drug delivery fibers andscaling.\" J. Clin. Perio. 18: 111-116.Heinegard, D. and M. Paulson (1984). Structure and metabolism of proteoglycans. ExtracellularMatrix Biochemistry. New York, Elsevier. 277-328.HeHan, L. B., M. A. Listgarten, and J. Lindhe (1979). \"The effect of tetracycline and/or scalingon human periodontal disease.\" ISm2gdo,l .^6: 222-230.Helovuo, H. and K. Paunio (1989). \"Effects of penicillin and erythromycin on the clinicalparameters of the periodontium.\" J. Periodontol. 60(8): 467-472.Hevia, P., S. T. Omaye, and R. A. Jacob (1990). \"Urinary hydroxyproline excretion and vitaminC status in healthy young men.\" Am. J. Clin. Nutrit. 51(4): 644-648.Holm-Perersen, P., N. Agerbaek, and E. Theilade (1975). \"Experimental gingivitis in young andelderly individuals.\" J. Clin. Periodontol. 2: 14-24.Holmberg, K. and J. Killander (1971). \"Quantitative determination of immunoglobulin (IgG, IgAand IgM) and identification of IgA-type in the gingival fluid.\" J. Perio. Res. 6: 1-8.Honig, J., C. Rordorf-Adam, C. Siegmund, W. Wiedemann, and F. Erard (1989). \"Increasedinterleukin-1 beta (IL-1i3) concentration in gingival tissue from periodontitis patients.\" J. Perio. Res. 24: 362-367.HOrlein, D., P. P. Fietzek, E. Wachter, C. M. Lapiere, and K. Kuhn (1979). \"Amino acidsequence of the aminoterminal segment of dermatosparactic calf-skin procollagen type I.\" Eur. J. Biochem. 99: 31-38.Hortin, G. and I. Boime (1981). \"Transport of an uncleaved preprotein into the endoplasmicreticulum of rat pituitary cells.\" J. Biol Chem 256: 1491-1494.Housley, T. J., F. N. Rowland, P. W. Ledger, J. Laplan, and M. L. Tanzer (1980). \"Effects oftunicamycin on the biosynthesis of procollagen by human fibroblasts.\" J. Biol Chem 255: 121-125.Hove, K. A. and R. E. Stallard (1970). \"Diabetes and the periodontal patient\" J. Periodont. 41:713-718.198Hughes, H., L. Hagen, and R. A. L. Sutton (1986). \"Determination of urinary hydroxyproline byhigh performance liquid chromatography.\" unpublished : 1-8.Hulmes, D. J. S., A. Miller, D. A. D. Darry, K. A. Piez, and J. Woodhead-Galloway (1973).\"Analysis of the primary structure of collagen for the origins of molecular packing.\" J. Mol. Biol. 79: 137-148.Ishikawa, I. and G. Cimasoni (1970). \"Alkaline phosphatase in human gingival fluid and itsrelation to periodontitis.\" Arch. oral Biol. 15: 1401-1404.Ivanyi, L., J. M. A. Wilton, and T. Lehner (1972). \"Cell-mediated immunity in periodontaldisease. Cytotoxicity, migration, inhibition and lymphocyte transformation studies.\" Immunology22: 141-145.Jameson, L. M. (1979). \"Comparison of the volume of crevicular fluid from restored andnonrestored teeth.\" J. Pros. Dent. 41: 209-214.Jandinski, J. J., P. Stashenko, L. S. Feder, C. C. Leung, W. J. Peros, J. E. Rynar, and M. J.Deasy (1991). \"Localization of interleukin-113 in human periodontal tissue.\" J. Periodontol. 62:36-43.Jenkins, W. M. M., T. W. MacFarlane, W. H. Gilmour, I. Ramsay, and D. MacKenzie (1989).\"Systemic metronidazole in the treatment of periodontitis.\" J. Clin. Perio. 16: 433-450.Johannessen, A. C., R. Nilsen, G. E. Knudsen, and T. Kristoffersen (1986). \"In situcharacterization of mononuclear cells in human chronic marginal periodontitis using monoclonalantibodies.\" J. Perio. Res. 21: 113-127.Johnson, P. W. (1983). Protein synthesis by human gingival fibroblasts in the presence of thiolcompounds. University of British Columbia.Johnson, P. W., W. Ng, and J. Tonzetich (1992). \"Modulation of human gingival fibroblast cellmetabolism by methyl mercaptan.\" J. Perio. Res. (In press).Johnson, P. W. and J. Tonzetich (1985). \"Sulphur uptake by type I collagen from methylmercaptan/dimethyl sulfide air mixtures.\" J. Dent. Res. 64: 1361-1364.Johnson, P. W., J. Tonzetich, and R. H. Pearce (1985). \"Characterization of hydrogen sulphidereaction with rat-tail tendon type I collagen in vitro.\" J. Periodont. Res. 20: 403-410.Johnson, P.W., K. Yaegaki, and J. Tonzetich (1992). \"Effect of volatile thiol compounds onprotein metabolism by human gingival fibroblasts.\" J. Perio. Res. (In press)Kaizu, T., M. Tsunoda, H. Aoki, and K. Kimura (1978). \"Analysis of volatile sulphurcompounds in mouth air from clinically healthy subjects and patients with periodontal disease.\"Bull. Tokyo dent. Coll. 19: 43-52.Kaldahl, W. B., K. L. Kalkwarf, K. D. Patil, and M. P. Molvar (1990). \"Relationship of gingivalbleeding, gingival suppuration, and supragingival plaque to attachment loss.\" J. Periodontol. 61:347-351.Kao, W. W.-Y., R. A. Berg, and D. J. Prockop (1977). \"Kinetics for the secretion of procollagenby freshly isolated tendon cells.\" J. Biol Chem 252: 8391-8397.199Kao, W. W.-Y., D. J. Prockop, and R. A. Berg (1979). \"Kinetics for the secretion of non-helicalprocollagen by freshly isolated tendon cells.\"LBiolCisnt 254: 2234-2243.Kaplan, G. B., C. H. Pameijer, and M. P. Ruben (1977). \"Scanning electron microscopy ofsulcular and junctional epithelia correlated with histology (Part I).\" J. Periodontol 48: 446-451.Kaslick, R. S., A. J. Chasen, D. Weinstein, and R. Waldman (1968). \"Ultramicromethod for thecollection of gingival fluid and quantitative analysis of its sodium content.\" J. Dent. Res. 47:1192.Kefalides, N. A., R. Alper, and C. C. Clark (1979). \"Biochemistry and metabolism of basementmembranes.\" Int. Rev. Ctyology 61: 167-228.Kempson, G. E. (1980). Mechanical properties of articular cartilage. Adult Articular Cartilage.Turnbridge Wells, U.K, Pitman Medical. 2nd, ed. 333-414.Kennedy, J. E. and A. M. Polson (1973). \"Experimental marginal periodontitis in squirrelmonkeys.\" J. Periodont. 44: 140-144.Kiihn, K., H. Wiedermann, R. Timpl, J. Risteli, H. Dieringer, T. Voss, and R. W. Glanville(1981). \"Macromolecular structure of basement membrane collagens. Identification of 7S collagenas a cross-linking domain of type IV collagen.\" FEBS Letters 125: 123-128.Kilchher, E., V. N. Schumaker, and L. K. Curtiss (1985). \"Activation of Clq by monoclonalantibodies directed against Clq.\" Biochem. Biophys. Res. Commun. 126: 785-791.Kivirikko, K. I. and R. Mylly1N. (1979). \"Collagen glycosyltransferases.\" Int. Rev. Conn. )'iss.Res. 8: 23-72.Kivirikko, K. I. and R. Myllyla (1980). Hydroxylation of prolyl and lysyl residues.Enzymology of Post-Translational Modification of Proteins. London, Academic Press. 53.Kivirikko, K. I. and R. Myllyla (1982). \"Posttranslational enzymes in the biosynthesis ofcollagen: Intracellular enzymes.\" Methods Enzymology 82A: 245-304.Kivirikko, K. I. and R. Myllyla (1984). Biosynthesis of the Collagens. Extracellular MatrixBiochemistry. New York, Elsevier. 83-118.Kleinman, H. K., E. B. McGoodwin, and R. J. Klebe (1976). \"Localization of the cell attachmentregion in types I and II collagens.\" aosho1131Qphysaun„ 72: 426-432.Kleinman, H. K., C. M. Wilkes, and G. R. Martin (1981). \"Interaction of fibronectin withcollagen fibrils.\" Biochemistry 10: 2325-2330.Klingsberg, J., L. A. Cancellaro, and E. 0. Butcher (1961). \"The comparative distribution ofglycogen, succinic dehydrogenase, and acid mucopolysaccharides in different aged animals.\" LDent. Res. 40: 461-469.Kostelc, J. G., G. Preti, P. R. Zelson, N. H. Stoller, and J. Tonzetich (1980). \"Salivary volatilesas indicators of periodontitis.\" J. Perio. Res. 15: 185-192.Kostelc, J. G., G. Preti, P. R. Zelson, J. Tonzetich, and G. R. Huggins (1981). \"Volatiles ofexogenous origin from the human oral cavity.\" J. Chromatogr. 226: 315-323.200Kowashi, Y., F. Jaccard, and G. Cimasoni (1979). \"Increase of free collagenase and neutralprotease activities in the gingival crevice during experimental gingivitis in man.\" Arch. oral Biol. 24: 645-650.Kowashi, Y., F. Jaccard, and G. Cimasoni (1980). \"Sulcular polymorphonuclear leukocytes andgingival exudate during experimental gingivitis in man.\" J. Perio. Res. 15: 151-158.Krasse, B. and J. Egelberg (1962). \"The relative proportions of sodium, potassium and calcium ingingival pocket fluid.\" Acta Odont. Scand. 20: 143-152.Kryshtalskyj, E. and J. Sodek (1987). \"Nature of collagenolytic enzyme and inhibitor activities increvicular fluid from healthy and inflamed periodontal tissues of beagle dogs.\" J. Perio. Res. 22:264-269.Kryshtalskyj, E., J. Sodek, and J. M. Ferrier (1986). \"Correlation of collagenolytic enzymes andinhibitors in gingival crevicular fluid with clinical and microscopic changes in experimentalperiodontitis in the dog.\" Archs. Oral Biol. 31: 21-31.Kuboki, Y., T. Takagi, S. Sasaki, S. Saito, and G. L. Mechanic (1981). \"Comparative collagenbiochemistry bovine periodontium, gingiva and dental pulp.\" J. Dent. Res. 60: 159-163.Kumamoto, C. A. (1981). \"Propeptides of procollagen V (A,B) in chick embryo crop.\" J. Biol Chem 256: 7053-7058.Kunimatsu, K., K. Yamamoto, E. Ichimaru, Y. Kato, and I. Kato (1990). \"Cathepsins B,H and Lactivities in gingival crevicular fluid from chronic adult periodontitis patients and experimentalgingivitis subjects.\" J. Perio. Res. 25: 69-73.Kurkinen, M., T. Vartio, and A. Vaheri (1980). \"Polypeptides of human plasma fibronectin aresimilar but not identical.\" Biochim. Biophys. Acta 624: 490-492.Lamster, I. B., D. S. Harper, S. Goldstein, R. S. Celenti, and R. L. Oshrain (1989). \"The effectof sequential sampling on crevicular fluid volume and enzyme activity.\" J. Clin. Perio. 16: 252-258.Lamster, I. B., L. J. Hartley, and R. I. Vogel (1985a). \"Development of a biochemical profile forgingival crevicular fluid: Methodological considerations and evaluation of collagen-degrading andground substance-degrading enzyme activity during experimental gingivitis.\" J. Periodontol. 56(Suppl): 13-21.Lamster, I. B., R. D. Mandella, and J. M. Gordon (1985b). \"Lactate dehydrogenase activity ingingival crevicular fluid collected with filter paper strips: analysis in subjects with non-inflamedand mildly inflamed gingival.\" J. Clin. Perio. 12: 153-161.Lamster, I. B., R. L. Oshrain, D. S. Harper, R. S. Celenti, C. A. Hovliaras, and J. M. Gordon(1988). \"Enzyme activity in crevicular fluid for detection and prediction of clinical attachment lossin patients with chronic adult periodontitis.\" J. Periodontol. 59(8): 516-523.Lamster, I. B., R. I. Vogel, L. J. Hartley, C. A. DeGeorge, and J. M. Gordon (1985c). \"Lactatedehydrogenase,P-glucuronidase and arysulfatase activity in gingival crevicular fluid associatedwith experimental gingivitis in man.\" J. Periodontol. 56(3): 139-147.201Larsson, B. T. (1965). \"Gas chromatography of organic volatiles in human breath and saliva.\"Acta Chem. Scand. 19: 159-164.Last, K. S., J. B. Stanbury, and G. Embery (1985). \"Glycosaminoglycans in human gingivalcrevicular fluid as indicators of active periodontal disease.\" Archs oral Biol. 30: 275-281.Lindahl, U. and V. HOOk (1978). \"Glycosaminoglycans and their binding to biologicalmacromolecules.\" Ann. Rev. Biochem, 47: 385-417.Lindahl, U. and L. Roden (1972). Glycoproteins. Glycoproteins. Amsterdam, Elsevier. 491-517.Lindhe, J. and R. AttstrOm (1967). \"Gingival exudation during the menstrual cycle.\" J. Perio. Res. 2: 194-198.Lindhe, J. and I. Ericsson (1979). \"Effect of ligature placement and dental plaque on periodontaltissue breakdown in the dog.\" J. Periodont. 49: 343-350.Lindhe, J., S.-E. Hamp, and H. L6e (1975). \"Plaque induced periodonal disease in beagle dogs.A 4 year clinical, roentgenographical and histometrical study.\" J. Perio. Res. 10: 243-255.Lindhe, J., S. E. Hamp, and H. LOe (1973). \"Experimental periodontitis in the beagle dog.\" L.Perio. Res. 8: 63-.Linsenmayer, T. F., J. M. Fitch, T. M. Schmid, N. B. Zar, E. Gibney, R. D. Sanderson, and R.Mayne (1983). \"Monoclonal antibodies against chicken type V collagen: production, specificity,and use for immunocytochemical localization in embryonic cornea and other organs.\" J. Cell Biol. 96: 124-132.Listgarten, M. A. (1964). \"The ultrastructure of human gingival epithelium.\" Am. J. Anat. 114:49-70.Listgarten, M. A. (1966). \"Electron microscopic study of the dentogingival junction of man.\" AmJ. Anat. 119: 147-178.Listgarten, M. A. (1980). \"Periodontal probing: what does it mean?\" J. Clin. Perio. 7(3): 165-176.Listgarten, M. A. (1984). \"Subgingival microbiological differences between periodontally healthysites and diseased sites prior to and after treatment.\" Int. J. Periodont. Rest. Dent. 1: 27-33.Listgarten, M. A. (1987). \"Nature of periodontal diseases: Pathogenic mechanisms.\" J. Perio. Res. 22: 172-178.Listgarten, M. A. (1988). \"A rationale for monitoring periodontal microbiota after periodontaltreatment.\" J. Periodontol. 59(7): 439-444.Listgarten, M. A. and B. Ellegaard (1973). \"Experimental gingivitis in the monkey.\" J. Periodont. Res. 8: 199-214.Listgarten, M. A. and L. Heal& (1978). \"Relative distribution of bacteria at clinically healthy andperiodontally diseased sites in humans.\" J. Clin. Perio. 5: 115-132.202Listgarten, M. A., J. Lindhe, and L. B. Hellden (1978). \"Effect of tetracycline and/or scaling onhuman periodontal disease. Clinical microbiologic and histological observations.\" J. Clin. Perio. 5: 246-271.Listgarten, M. A., J. Slots, A. H. Nowotny, J. Oler, J. Rosenberg, B. Gregor, and P. Sullivan(1991). \"Incidence of periodontitis recurrence in treated patients with and without cultivableActinobacillus actinomycetemcomitans, Prevotella intermedia, and Poiphyromonas gingivalis: Aprospective study.\" J. Periodontol. 62: 377-386.Lbe, H., A. Anerud, H. Boysen, and M. Smith (1978). \"The natural history of periodontal diseasein man. The rate of periodontal destruction before 40 years of age.\" J. Periodontol. 49: 607-620.Liie, H. and P. Holm-Pedersen (1965). \"Abscence and presence of fluid from normal andinflamed gingivae.\" Periodontics 3: 171-177.Ltie, H., E. Theilade, S. B. Jensen, and C. R. Schiott (1967). \"Experimental gingivitis in man.The influence of antibiotics on gingival plaque development.\" J. Perio. Res. 2: 282-289.Loesche, W.J., J.R. Giordano, P. Hujoel, J. Schwarz, and B.A. Smith (1992). \"Metronidazole inperiodontitis: reduced need for surgery.\" .LClim12eLo Li 19: 103-112.Loesche, W.J., E. Schmidt, B.A. Smith, E.C. Morrison, R. Caffesse, and P.P. Hujoel (1991).\"Effects of metronidazole on periodontal treatment needs.\" J. Periodontol. 62: 247-257.Loesche, W. J. and S. A. Syed (1978). \"Bacteriology of human gingivitis: Effect on plaque andgingivitis scores.\" Infect. Immun. 21: 830-839.Luger, T. A., B. M. Stadler, B. M. Luger, B. J. Mathieson, M. Mage, J. H. A. Schmidt, and J.J. Oppenheim (1982). \"Murine epidermal cell-derived thymocyte-activating factor resemblesmurine interleukin 1.\" J. Immunol. 128: 2147-2152.MacKay, D. A. M. and M. M. Hussein (1978). Analysis of Food and Beverages. Academic PressInc.MacKay, R., D. Grenier, B. McBride, and J. Tonzetich (1989). \"Proteolytic effect of Bacteriodesgingivalis vesicles on oral mucosa.\" J. Dent. Res. 68: 1006 (Asbtr#1113).Madder, B. F., K. B. Frostad, P. B. Robertson, and B. M. Levy (1977). \"Immunoglobulinbearing lymphocytes and plasma cells in human periodontal disease.\" J. Perio. Res. 12: 37-45.Mainardi, C. L., S. N. Dixit, and A. H. Kang (1980a). \"Degradation of type IV (basementmembrane) collagen by a proteinase isolated from human polymorphonuclear leukocyte granules.\"J. Biol. Chem. 255: 5435-5441.Mainardi, C. L., D. L. Hasty, J. M. Seyer, and A. H. Kang (1980b). \"Specific cleavage of humantype III collagen by human polymorphonuclear leukocyte elastase.\" J. Biol. Chem. 255: 12006-12010.Mairuno, W., G. Singh, and A. E. Dolby (1982). \"Enhanced cellular immune response to type Icollagen in patients with periodontal disease.\" Int. Archs Allergy appl. Immun. 67: 149-154.Mandel, R. L. (1984). \"A longitudinal microbial investigation of Actinobacillusactinomycetemcomitans and Eikenella conodens in juvenile periodontitis.\" Infect. Immun. 45:778-779.203Mann, W. V. (1963). \"The correlation of gingivitis, pocket depth and exudate from the gingivalcrevice.\" J. Periodontol. 34: 379-387.Martinez-Hernandez, A. and D. S. Amenta (1983). \"The basement membrane in pathology.\" LabInvest. 48: 656-677.Marwah, A. S., J. P. Weinmann, and J. Meyer (1960). \"Effect of chronic inflammation of theepithelial turnover of the human gingiva.\" Arch. Path. 69: 147-153.Masada, M. P., R. Persson, J. S. Kenney, S. W. Lee, R. C. Page, and A. C. Allison (1990).\"Measurement of interleukin-la and -113 in gingival crevicular fluid: Implications for thepathogenesis of periodontal disease.\" J. Perio. Res. 25: 156-163.Massler, M., R. D. Emslie, and T. E. Bolden (1951). \"Fetor ex ore.\" Oral Surg. 4: 110-125.McNamara, T. F., J. F. Alexander, and M. Lee (1972). \"The role of microorganisms in theproduction of oral malodor.\" Oral Surg. 34: 41-48.McNamara, T. F., L. M. Golub, G. D'Angelo, and N. S. Ramamurthy (1986). \"The synthesisand characterization of a non-bacterial chemically modified tetracycline.\" J. Dent. Res. 65: (Abstr#515).Mergenhagen, S. E. (1984). \"Thymocyte Activation Factor(s) in Human Gingival Fluids.\" J.Dent. Res. 63: 461-464.Miller, E. J. (1976). \"Biochemical characteristics and biological significance of the geneticallydistinct collagens.\" Molecular and Cell Biochemistry 13: 165-192.Miller, E. J., J. E. J. Finch, E. Chung, W. T. Butler, and P. B. Robertson (1976). \"Specificcleavage of the native type HI collagen molecule with trypsin; Similarity of the cleavage product tocollagenase-produced fragments and primary structure at the cleavage site.\" Arch. Biochem. Biophys. 173: 631-637.Miller, R. A., J. R. Wynkoop, and K. Johannsen (1982). \"Analysis of hydroxyproline in gingivalexudate by high performance liquid chromatography.\" J. Dent. Res. 61: 275 (Abstract# 866).Mills, W. H., G. W. Thompson, and G. S. Beagrie (1979). \"Clinical evaluation of spiramycinand erythromycin in control of periodontal disease.\" J. Clin. Perio. 6: 308-316.Moore, W. E. C. (1987). \"Microbiology of periodontal disease.\" J. Perio. Res. 22: 335-341.Moore, W. E. C., L. U. Holdeman, R. M. Smibert, I. J. Good, J. A. Burmeister, K. J. Palcanis,and R. R. Ranney (1982). \"Bacteriology of experimental gingivitis in young adult tissues.\" Infect.Immun. 38: 651-657.Moro, J. and B. S. Smith (1977). \"Identification of collagen a 1 (I) trimer and normal type Icollagen in a polyoma virus - induced mouse tumour.\" Arch. Biochem. and Biophys. 182: 33-41.Mosher, D. F. (1975). \"Cross-linking of cold-insoluble globulin by fibrin-stabilizing factor.\" J.Biol Chem. 250: 6614-6621.204Mosher, D. F., P. E. Schad, and J. M. Vann (1980). \"Cross-linking of collagen and fibronectinby factor Villa.\" J. Biol. Chem. 255: 1181-1188.Mouton, C., M. Desclauriers, H. Allard, and M. Bouchard (1987). \"Serum antibodies toBacteroides gingivalis in periodontitis: A longitudinal study.\" J. Perio. Res. 22: 426-430.Mouton, C., P. G. Hammond, and J. Slots (1981). \"Serum antibodies to oral Bacteriodesasaccharolyticus (Bacteroides gingivalis): Relationship to age and periodontal disease.\" Infect. Immun. 31: 182-192.Muller-Glauser, W. and H. E. Schroeder (1982). \"The pocket epithelium: A light andelectronmicroscopic study.\" 3. Periodontol. 53: 133-144.Muramoto, K., J. Ramachandran, J. Hall, A. Hui, and R. Stem (1984). \"A rapid sensitive assayfor the quantitation of elastin.\" Conn. Tiss. Res. 12: 307-317.Murphy, G., U. Bretz, M. Baggiolini, and J. J. Reynolds (1980). \"The latent collagenase andgelatinase of human polymorphonuclear neutrophil leukocytes.\" Biochem. J. 199: 516-525.Murray, P. A., D. A. Burstein, and J. R. Winkler (1989). \"Antibodies to Bacteriodes gingivalis inpatients with treated and untreated periodontal disease.\" J. Periodontol. 60(2): 96-103.Myers, J. C., F.O. Sangiori, M.-L. Chu, G.Ding, D. J. Prockop, and F. Ramirez (1981).\"Isolation and characterization of part of the human pro a2 collagen gene.\" Fed. Proc. 40: 1650.Myllyla, R., H. Anttinen, and K. I. Kivirikko (1979). \"Metal activation of galactosylhydroxylysylglucosyltransferase, an intracellular enzyme of collagen biosynthesis.\" Eur. J. Biochem. 101:261-269.Narayanan, A. S., J. A. Clagett, and R. C. Page (1985). \"Effect of inflammation on thedistribution of collagen types I, III, IV, V, and type I timer in normal and inflamed humangingivae.\" J. Dent. Res. 64: 1111-1116.Narayanan, A. S. and R. C. Page (1983). \"Connective tissues of the periodontium: a summary ofcurrent work.\" Coll. Rel. Res. 39: 61-65.Narayanan, A. S., R. C. Page, and D. F. Meyers (1980). \"Characterization of collagens ofdiseased human gingiva.\" Biochemistry 19: 5037-5043.Ng, W. and J. Tonzetich (1984). \"Effect of hydrogen sulfide and methyl mercaptan on thepermeability of oral mucosa.\" J. Dent. Res. 63: 994-997.Niekrash, C. E. and M. R. Patters (1985). \"Simultaneous assessment of complement componentsC3, C4 A and B, and their cleavage products in human gingival fluid. II. Longitudinal changesduring periodontal therapy.\" J. Perio. Res. 20: 268-275.Niekrash, C. E. and M. R. Patters (1986). \"Assessment of complement cleavage in gingival fluidin humans with and without periodontal disease.\" J. Perio. Res. 21: 233-242.Oakley, B. R., D. R. Kirsch, and N. R. Morris (1980). \"A simplified ultrasensitive silver stain fordetecting proteins in polyacrylamide gels.\" Anal. Biochem. 105: 361-363.205Offenbacher, S., B. M. Odle, L. D. Braswell, H. G. Johnson, C. M. Hall, H. McGuire, J. L.Orkin, E. A. Strobert, and M. D. Green (1989). \"Changes in cyclooxygenase metabolites inexperimental periodontitis in Macaca mulatta.\" J. Perio. Res. 24: 63-74.Offenbacher, S., B. M. Odle, and T. E. Van Dyke (1986). \"The use of crevicular fluidprostaglandin E2 levels as a predictor of periodontal attachment loss.\" J. Perio. Res. 21: 101-112.Oldberg, A. and E. Ruoslahti (1982). \"Interactions between chondroitin sulfate proteoglycan,fibronectin, and collagen.\" J. Biol. Chem. 257: 4859-4863.Oliver, R. C., P. Holm-Pedersen, and H. Liie (1969). \"The correlation between clinical scoring,exudate measurements and microscopic evaluation of inflammation in the gingiva.\" J. Periodontol. 40: 231-235.Olsen, B. R. (1981). Collagen Biosynthesis. Cell Biology of Extracellular Matrix. New York,Plenum Press. 139-137.Omata, R. R. and R. C. Braunberg (1960). \"Oral fusobacteria.\" J. Bact. 80: 737-740.Orban, J. E. and R. E. Stallard (1969). \"Gingival crevicular fluid. A reliable predictor of gingivalhealth?\" J. Periodontol. 40: 321-325.Ortman, L. F., K. McHenry, and E. Hausmann (1982). \"Relationship between alveolar bonemeasured by R51 absorptiometry with analysis of standardized radiographs.\" J. Periodontol. 53(5): 311-314.Ouyang, Y. (1991). Effect of zinc on the metabolism of thiol-treated-human gingival fibroblasts.University of British Columbia.Overall, C. M., 0. W. Wiebkin, and J. C. Thonard (1987). \"Demonstration of tissue collagenaseactivity in vivo and its relationship to inflammation severity in human gingiva.\" J. Perio. Res. 22:81-88.Page, R. C. (1983). \"Rapidly progressive periodontitis: a distinct clinical condition.\" J.Periodontol, 54: 197-209.Page, R. C. and H. E. Schroeder (1976). \"Pathogenesis of inflammatory periodontal disease.\"Lab. Invest. 33: 235-249.Page, R. C. and H. E. Schroeder (1981). \"Current status of the host response in chronic marginalperiodontitis.\" J. Periodontol. 52: 477-491.Page, R. C., D. M. Simpson, and W. F. Ammons (1975). \"Host tissue response in chronicinflammatory periodontal disease. III. The periodontal and dental status of a group of aged greatapes.\" J. Periodontol 46: 144-155.Palcanis, K. G., I. K. Larjava, B. R. Wells, K. A. Suggs, J. R. Landis, D. E. Chadwick, and M.K. Jeffcoat (1992). \"Elastase as an indicator of periodontal disease progression.\" J. Periodontol. 63: 237-242.206Palmiter, R. D., J. M. Davidson, J. Gagnon, D. W. Rowe, and P. Bornstein (1979). \"NH 2 -terminal sequence of the chick-proal (I) chain synthesized in the reticulocyte lysate system.\" LBiol. Chem. 254: 1433-1436.Palotie, A. (1983). \"Effect of elevated temperature on the intracellular degradation of differentcollagen types.\" Cgll. Rel. Res, 3: 105-113.Patters, M. R., N. Sedransk, and R. J. Genco (1979). \"The lymphoproliferative response duringhuman experimental gingivitis.\" J. Perio. Res. 14: 269-278.Payne, W. A., R. C. Page, A. L. Ogilvie, and W. B. Hall (1975). \"Histopathologic features of theintial and early stages of experimental gingivitis in man.\" J. Perio. Res. 10: 51-64.Pearlstein, E. (1976). \"Plasma membrane glycoprotein which mediates adhesion of fibroblasts tocollagen.\" Nature 262: 497-500.Pearlstein, E., J. Sorvillo, and I. Giglii (1982). \"The interaction of human plasma fibronectin witha subset of the first component of complement, Clq.\" J. Immunol. 128: 2036-2039.Persson, G. R., T. A. DeRouen, and R. C. Page (1990a). \"Relationship between gingivalcrevicular fluid levels of aspartate aminotransferase and active tissue destruction in treated chronicperiodontitis patients.\" J. Perio. Res. 25: 81-87.Persson, G. R., T. A. DeRouen, and R. C. Page (1990b). \"Relationship between levels ofaspartate aminotransferase in gingival crevicular fluid and gingival inflammation.\" J. Perio. Res. 25: 17-24.Persson, G. R. and R. C. Page (1990). \"Effect of sampling time and repetition on gingivalcrevicular fluid and aspartate aminotransferase activity.\" J. Perio. Res. 25: 236-242.Persson, G. R. and R. C. Page (1992). \"Diagnostic characteristics of crevicular fluid aspartateaminotransferase (AST) levels associated with periodontal disease activity.\" J. Clin. Perio. 19: 43-48.Pesciotta, D. M., L. A. Dickson, A. M. Showalter, E. F. Eikenberry, B. DeCrombrugghe, P. P.Fietzek, and B. R. Olsen (1981). \"Primary structure of the carbohydrate-containing regions of thecarboxy propeptides of type I procollagen.\" FEBS Lett 125: 170-174.Piez, K. A. (1963). Primary structure. Biochemistry of Collagen. Plenum Press, New York. 1-44.Pommier, C. G., J. O'Shea, T. Chused, K. Yancey, M. M. Frank, T. Takahashi, and E. J.Brown (1984). \"Studies on the fibronectin receptors of human peripheral blood leukocytes;Morphologic and functional characterization.\" J. Exp. Med. 159: 137-157.Postlethwaite, A. E., J. Keski-Oja, G. Balian, and A. H. Kang (1981). \"Induction of fibroblastchemotaxis by fibronectin; Localization of the chemotactic region to a 140,000-molecular weightnon-gelatin-binding fragment.\" J. Exp. Med. 153: 494-498.Prockop, D. J., R. A. Berg, K. I. Kivirikko, and J. Uitto (1976). Intracellular steps in thebiosynthesis of collagen. Biochemistry of Collagen. New York, Plenum. 163-273.207Prockop, D. J., K. I. Kivirikko, L. Tuderman, and N. A. Guzman (1979). \"The biosynthesis ofcollagen and its disorders.\" New Engl. J. Med. 301: 13 & 77.Quinn, R. A. and S. M. Krane (1976). \"Abnormal properties of collagen lysyl hydroxylase fromskin fibroblasts of siblings with hydroxylysine-deficient collagen.\" J. Clin Invest 57: 83-93.Ramachandran, G. N., M. Bansal, and Ramakrishnan (1975). \"Hydroxyproline stabilises bothintra-fibrillar structure as well as inter-protofibrillar linkages in collagen.\" Current Science 44(1-3):Ranney, R. R., N. R. Yanni, J. A. Burmeister, and J. G. Tew (1982). \"Relationship betweenattachment loss and precipitating serum antibody to Actinobacillus actinomycetemcomitans inadolescents and young adults having severe periodontal destruction.\" J. Periodontol, 53: 1-7.Rao, C. N., I. M. K. Margulies, T. S. Tralka, U. P. Terranova, J. A. Madri, and L. A. Liotta(1982). \"Isolation of a subunit of larninin and its role in molecular structure and tumor cellattachment.\" J. Biol. Chem. 257: 9740-9744.Ratkay, L. and J. Tonzetich (1992). Personal communication: Studies of M-1 production by Tlymphocytes after exposure to CH3SH.Ratkay, L., J. Tonzetich, J. Coil, and J. D. Waterfield (1990). \"Immunological analysis of humangingival crevicular fluid.\" J. Dent. Res. 70(Special Issue): 326 (Abstr#1739).Refaie, A., 0. Anuksaksathiem, G. Singh, J. Moran, and A. E. Dolby (1990). \"Antibody tocollagen type I in gingival crevicular fluid.\" J. Periodontol. 60: 289-292.Reinhardt, R. A., T. L. McDonald, R. W. Bolton, L. M. DuBois, and W. B. Kaldahl (1989).\"IgG subclasses in gingival crevicular fluid from active versus stable periodontal sites.\" LPeriodontol. 60(1): 44-50.Richards, D. and R. B. Rutherford (1988). \"The effects of interleukin 1 on collagenolytic activityand prostaglandin-E secretion by human periodontal-ligament and gingival fibroblast.\" Archs. o Biol. 33: 237-243.Ringelberg, M. L., D. 0. Dixon, A. 0. Francis, and R. W. Plummer (1977). \"Comparison ofgingival health and gingival crevicular fluid flow in children with and without diabetes.\" J. Dent. Res. 56: 108-111.Risteli, J. and K. Kivirikko (1976). \"Intracellular enzymes of collagen biosynthesis in rat liver as afunction of age and in hepatic injury induced by dimethylnitrosamine. Changes in prolylhydroxylase, lysyl hydroxylse, collagen galactosyltransferase and collagen glucosyltransferaseactivities.\" Biochem. J. 158: 361-367.Risteli, J., L. Tuderman, and K. Kivirikko (1976). \"Intracellular enzymes of collagen biosynthesisin rat liver as a function of age and in hepatic injury induced by dimethylnitrosamine. Purificationof rat prolyl hydroxylase and comparison of changes in prolyl hydroxylse activit with changes inimmunoreactive prolyl hydroxylase.\" Biochem. J. 158: 369-376.Rizzo, A. A. (1967). \"The possible role of hydrogen sulfide in human periodontal disease. I.Hydrogen sulfide production in periodontal pockets.\" Periodontics 5: 233-236.Robertson, P. B., S. A. Buchanan, G. C. Armitage, E. Newbrun, E. J. Taggart, and C. I.Hoover (1987). \"Evaluation of clinical and microbiological measures to predict treatment responsein severe periodontitis.\" J. Perio. Res. 22: 230-232.208Roden, L. (1980). Structure and metabolism of connective tissue proteoglycans. The Biochemisny,of Glycoproteins and Proteoglycans. New York, Plennum. 267-371.Rogosa, M. and F. S. Bishop (1964). \"The genus veillona. III. Hydrogen sulfide production bygrowing cultures.\" J. Bact. 88: 37-41.Rojkind, M., M.-A. Giambrone, and L. Biempica (1979). \"Collagen types in normal and cirrhoticliver.\"igukggit ig:Diggy 76: 710-719.Rosenbloom, J., R. Endo, and M. Harsch (1976). \"Termination of procollagen chain synthesis bypuromycin.\" J. Biol Chem 251: 2070-2076.Rosenbloom, J., M. Harsch, and S. A. Jimenez (1973). \"Hydroxyproline content determines thedenaturation temperature of chick tendon collagen.\" Arch Biochem Biophys 158: 478-484.Rossomando, E. F., J. E. Kennedy, and J. Hadjimichael (1990). \"Tumour necrosis factor alpha ingingival crevicular fluid as a possible indicator of periodontal disease in humans.\" Archs oral Biol. 35(6): 431-434.Riidin, H. J., H. F. Overdiek, and K. H. Rateitschak (1970). \"Correlation between sulcus fluidrate and clinical and histological inflammation of the marginal gingiva.\" Hely. Wont. Acta 14: 21-26.Ruoslahti, E. and A. Vaheri (1975). \"Interaction of soluble fibroblast surface antigen withfibrinogen and fibrin. Identity with cold insoluble globulin of human plasma.\" J. Exp. Med. 141:497-501.Sage, H. and P. Bornstein (1979). \"Characterization of a novel collagen chain in human placentaand its relation to AB collagen.\" J3iochem. 18: 3815-3822.Sage, H., R. G. Woodbury, and P. Bornstein (1979). \"Structural studies on human type IVcollagen.\" J. Biol. Chem. 254: 9893-9900.Sallay, K., F. Sanavi, I. Ring, P. Pham, U. H. Behling, and A. Nowothy (1982). \"Alveolar bonedestruction in the immunosuppressed rat.\" J. Perio. Res, 17: 263-274.Salo, T., L. A. Liotta, and K. Tryggvason (1983). \"Purification and characterization of a murinebasement membrane collagen-degrading enzyme secreted by metastatic tumor cells.\" J. Biol, Chem. 258: 3058-3063.Sandalli, P. and A. B. Wade (1971). \"Crevicular fluid flow in young children.\" J. Peirodontol. 42: 713-716.Sandell, L. and A. Veis (1980). \"The molecular weight of the cell-free translation product of al (I)procollagen mRNA.\" Biochem Biophys Res Commun 92: 554-562.Savitt, E. D., M. N. Strzempko, K. K. Vaccaro, W. J. Peros, and C. K. French (1988).\"Comparison of cultural methods and DNA probe analyses for the detection of actinobacillusactinomycetemcomitans, Bacteroides gingivalis, and Bacteroides intermedius in subgingival plaquesamples.\" J. Periodontol, 59(7): 431-438.209Sawyer, S. J., J. B. Macdonald, and R. J. Gibbons (1962). \"Biochemical characteristics ofBacteroides melaninogenicus. A study of thirty one strains.\" Arch. Oral Biol. 7: 685-691.Saxe, S. R., J. C. Greene, H. M. Bohannan, and J. B. Vermillion (1967). \"Oral debris, calculus,and periodontal disease in the beagle dog.\" Periodontics 5: 217-225.Schectman, L. R., W. F. Ammons, D. M. Simpson, and R. C. Page (1972). \"Host tissueresponse in chronic periodontal disease. II. Histologic features of the normal periodontium, andhistopathologic and ultrastructural manifestations of disease in the marmoset.\" J. Perio. Res. 7:195-212.Schenkein, H. A. (1991). \"The role of complement in periodontal diseases.\" Crit. Rev. Oral Biol. Med. 2(1): 65-81.Schmidt, E. F., W. A. Bretz, R. A. Hutchinson, and W. J. Loesche (1988). \"Correlation of thehydrolysis of benzoyl-arginine naphthylamide (BANA) by plaque with clinical parameters andsubgingival levels of spirochetes in periodontal patients.\" J. Dent. Res. 67: 1505-1509.Schroeder, H. E. and J. Lindhe (1975). \"Conversion of stable established gingivitis in dog intodestructive periodontitis.\" Arch. Oral Biol. 20: 775-782.Schroeder, H. E. and J. Lindhe (1980). \"Conditions and pathological features of rapidlydestructive experimental periodontitis in dogs.\" J. Periodont. 51: 6-19.Schroeder, H. E. and R. C. Page (1990). The Normal Periodontium. Periodontal Diseases. Basicphenomena. clinical management. and occlusal and restorative interrelationships. Lea & Febiger.2nd, ed. 3-52.Schroeder, H. E. and J. Theilade (1966). \"Electron microscopy of normal human gingivalepithelium.\" J. Perio. Res. 1: 59-119.Schwabacke, H. and D. R. Lucas (1947). \"Bacterium melaninogenicus - a misnomer.\" J. Gen. Microbiol. 1: 109-120.Sekiguchi, K. and S. Hakomori (1980). \"Functional domain structure of fibronectin.\" Proc. Natl. Acad. Sci. U.S.A. 77: 2661-3665.Sengupta, S., J. Fine, C.-Y. Wu-Wang, J. Gordon, V. L. N. Murty, A. Slomiany, and B. L.Slomiany (1990). \"The relationship of prostaglandins to cAMP, IgG, Ig,M and a-2macroglobulinin gingival crevicular fluid in chronic adult periodontitis.\" Archs oral Biol. 35(8): 593-396.Sengupta, S., I. B. Lamster, A. Khocht, T. A. Duffy, and J. M. Gordon (1988). \"The effect oftreatment on IgG, IgA, IgM and a-2macroglobulin in gingival crevicular fluid from patients withchronic adult periodontitis.\" Archs oral Biol. 33(6): 425-431.Seymour, G. J. and J. S. Greenspan (1979). \"The phenotypic characterization of lymphocytesubpopulations in established human periodontal disease.\" J. Perio. Res. 14: 39-46.Seymour, G. J., R. N. Powell, K. L. Cole, J. F. Aitken, D. Brooks, I. Beckman, H. Zola, J.Bradley, and G. F. Burns (1983). \"Experimental gingivitis in humans. A histochemical andimmunological characterization of the lymphoid cell subpopulations.\" J. Perio. Res. 14: 39-46.210Shibutani, T., Y. Murahashi, and Y. Iwayama (1989). \"Immunohistochemical localization ofchondroitin sulfate and dermatan sulfate proteoglycan in human gingival connective tissue.\" LPerio. Res. 24: 310-313.Siegel, R. C. (1980). \"Lysyl oxidase.\" Int. Rev. Conn. Tissue Res. 8: 73-118.Silbert, J. E. (1964). \"Incorporation of 14C and 3H from labeled nucleotide sugars into apolysaccharide in the presence of a cell-free preparation from cartilage.\" J. Biol. Chem. 239:1310-1315.Simonson, L. G., P. J. Robinson, R. J. Pranger, M. E. Cohen, and H. E. Morton (1992).\"Treponema denticola and Prophyromonas gingivalis as prognostic markers following periodontaltreatment.\" J. Periodontol. 63: 270-273.Skapski, H. and T. Lehner (1976). \"A crevicular washing method for investigating immunecomponents of crevicular fluid in man.\" J. Period. Res. 11: 19-24.Slots, J. (1976). \"The predominate cultivable flora in juvenile periodontitis.\" Scand. J. Dent. Res. 84: 1-10.Slots, J. (1977). \"Microflora in the healthy gingival sulcus in man.\" Scand. J. Dent, 85: 247-254.Slots, J. (1979). \"Subgingival microflora and periodontal disease.\" J. Clin, Perio, 6: 351.Slots, J., L. Bragd, M. Wikstriim, and G. Dahl& (1986). \"The occurrence of Actinobacillusactinomycetemcomitans, Bacteriodes gingivalis and Bacteroides intermedius in destructiveperiodontal disease in adults.\" J. Clin. Perio. 13: 570-577.Slots, J., P. Mashimo, M. J. Levine, and R. J. Genco (1979). \"Periodontal therapy in humans I.Microbiological and clinical effects of a single course of periodontal scaling and root planing and ofadjunctive tetracycline therapy.\" J. Periodontol. 50: 495-509.Socransky, S. S., A. D. Haffajee, J. M. Goodson, and J. Lindhe (1984). \"New concepts ofdestructive periodontal disease.\" J. Clin. Periodont. 11: 21-32.Sodek, J. (1976). \"A new approach to assessing collagen turnover by using a micro-assay. Ahighly efficient and rapid turnover of collagen in rat periodontal tissues.\" Biochem. J. 160: 243-246.Sodek, J. (1977). \"A comparison of the rates of synthesis and turnover of collagen and non-collagen protein in adult rat periodontal tissue and skin using a microassay.\" Arch. Oral Biol. 22:655-665.Sodek, J. (1978). \"A comparison of collagen and non-collagenous protein metabolism in rat molarand incisor periodontal ligaments.\" Arch. Oral Biol. 23: 977-982.Sodek, J. (1990). Personal communication: Unacceptable purification of affinity purified type Icollagen antibody.Sodek, J., D. M. Brunette, J. Feng, J. N. M. Heersche, H. F. Limeback, A. H. Melcher, and B.Ng (1977). \"Collagen synthesis is a major component of protein synthesis in the periodontalligament of various species.\" Arch. Oral Biol. 22: 647-653.211Sodek, J. and H. F. Limeback (1979). \"Comparison of the rates of synthesis, conversion, andmaturation of type I and type III collagens in rat periodontal tissues.\" J. Biol. Chem. 254: 10496-10502.Soder, P., L. Frithiof, S. Wikner, F. Wouters, P. Engstrom, B. Rubin, U. Nedlich, and B. Soder(1990). \"The effect of systemic metronidazole after non-surgical treatment in moderate andadvanced periodontitis in young adults.\" J. Periodontol. 61: 282-288.SOderholm, G. and R. AttstrOm (1977). \"Vascular permeability during initial gingivitis is dogs.\" LPerk). Res. 12: 395-401.Solis-Gaffar, M. C., K. N. Rustogi, and A. Gaffar (1980). \"Hydrogen sulfide production fromgingival crevicular fluid.\" J. Periodontol, 51: 603-606.Son, S., P. Hotz, and H. Miihlemann (1971). \"The effect of marginal gingivitis on toothmobility.\" Hely. Odont. Acta 15: 103-106.Sorsa, T., K. Suomalainen, and V.-J. Uitto (1990). \"The role of gingival crevicular fluid andsalivary interstitial collagenases in human periodontal diseases.\" Archs Oral Biol. 35(Suppl.):193s-196s.Starkey, P. M., A. J. Barrett, and M. C. Burleigh (1977). \"The degradation of articular collagenby neutrophil proteinases.\" Biochim. Biophys. Acta 483: 389-397.Steinmann, B., V. H. Rao, and R. Gitzelrnann (1981). \"Intracellular degradation of newlysynthesized collagen is conformation-dependent; Studies in human skin fibroblasts.\" FEBS Lett. 133: 142-144.Stenn, K. S., J. A. Madri, and F. S. Roll (1979). \"Migrating epidermis produces AB2 collagenand requires continual collagen synthesis for movement.\" Nature 277: 229-232.Strauss, A. W., M. Zimmerman, I. Boime, B. Ashe, R. S. Mumford, and A. W. Alberts (1979).\"Characterization of an endopeptidase involved in pre-protein processing.\" Proc. Natl. Acad. Sci. U.S.A. 76: 4225-4229.Sueda, T., J. Bang, and G. Cimasoni (1969). \"Collection of gingival fluid for quantitativeanalysis.\" J. Dent. Res. 48: 159.Sueda, T. and G. Cimasoni (1968). \"The origins of acid phosphatase in human gingival fluid.\"Arch. oral Biol. 13: 553-558.Sulser, G. F., K. H. Brening, and L. S. Fosdick (1939). \"Some conditions that affect the odorconcentrations of the breath.\" J. Dent. Res. 18: 355.Suomalainen, K., T. Sorsa, L. Saxen, M. Vauhkonen, and V.-J. Uitto (1990). \"Collagenaseactivity in gingival crevicular fluid of patients with juvenile periodontitis.\" Oral Microlbiol. Immun. 5: 1-6.Suppipat, N., J. R. Johansen, and P. Gjermo (1977). \"Influence of 'time of day', pocket depthand scaling on gingival fluid flow.\" J. Clin. Periodontol. 4: 48-55.Svanberg, G. K. (1987a). \"Hydroxyproline determination in serum and gingival crevicular fluid.\"J. Perio. Res. (22): 133-138.212Svanberg, G. K. (1987b). \"Hydroxyproline titers in gingival crevicular fluid.\" J. Perio. Res. 22:212-214.Switzer, R. C. I., C. R. Merril, and N. R. Morris (1979). \"A highly sensitive silver stain fordetecting proteins and peptides in polyacrylamide gels.\" Anal. Biochem. 98: 231-237.Syed, S., W. Loesche, E. Morrison, B. Laughon, and S. Ramford (1981). \"Subgingival plaqueflora of untreated moderate to advanced periodontitis patients.\" J. Dent. Res. 60: Abstr #81.Talonpoika, J. and M. Hamalainen (1992). \"Collagen III aminotenninal propeptide in gingivalcrevicular fluid before aned after periodontal treatment.\" Scand. J. Dent. Res, (in press):Tanner, A. C. R., S. S. Socransky, and J. M. Goodson (1984). \"Microbiota of periodontalpockets losing alveolar bone.\" J. Perio. Res. 19: 279-291.Taubman, M. A., E. D. Stoufi, J. L. Ebersole, and D. J. Smith (1984a). \"Phenotypic studies ofcells from periodontal disease tissues.\" J. Perio. Res. 19: 587-590.Taubman, M. A., H. Yoshie, J. L. Ebersole, D. J. Smith, and D. L. Olson (1984bf). \"Hostresponse in experimental periodontal disease.\" 3. Dent. Res. 63(3): 455-460.Terranova, U. P., D. H. Rohrbach, and G. R. Martin (1980). \"Role of laminin in the attachmentof PAM 212 (epithelial) cells to basement membrane collagen.\" Cell 22: 719-729.Thilander, H. (1963). \"The effect of leukocytic enzyme activity on the structure of the gingivalpocket epithelium in man.\" Acta. Odont. Scand. 21: 431-451.Thilander, H. (1964). \"Permeability of the gingival pocket epithelium.\" Int. Dent. J. 14: 416-421.Thilander, H. (1968). \"Epithelial changes in gingivits - An electron microsopic study.\" J. Perio. Res. 3: 303-312.Thonard, J. C. and H. W. Scherp (1959). \"Characterization of human gingival intercellularcementing substance.\" J. Dent. Res. 38: 711.Timpl, R., J. Risteli, and H. P. Bachinger (1979a). \"Identification of a new basement membranecollagen by the aid of a large fragment resistant to bacterial collagenase.\" FEBS Lett. 101: 265-268 .Timpl, R., H. Rohde, P. G. Robey, S. I. Rennard, J.-M. Foidart, and G. R. Martin (1979b).\"Laminin - a glycoprotein from basement membranes.\" J. Biol. Chem. 254: 9933-9937.Timpl, R., H. Wideman, V. VanDelden, H. Furthmayr, and K. Kuhn (1981). \"A network modelfor the organization of type IV collagen molecules in basement membranes.\" Eur. J. Biochem. 120: 203-211.Tolo, K. and K. Schenck (1985). \"Activity of serum immunoglobulins G, A, and M to sixanaerobic oral bacteria in diagnosis of periodontitis.\"J.12gth i ats„ 20: 113-121.Tonzetich, J. (1971). \"Direct gas chromatographic analysis of sulphur compounds in mouth air inman.\" ABLOral Biol. 16: 587-597.213Tonzetich, J. (1978). \"Oral Malodour: an indicator of health status and oral cleanliness.\" Int. Dent. L 28: 309-319.Tonzetich, J. and P. A. W. Carpenter (1971). \"Production of volatile sulphur compounds fromcysteine, cystine and methionine by human dental plaque.\" Arch. Oral Biol. 16: 685-607.Tonzetich, J., G. Cimasoni, and J. Coil (1987). \"Effect of volatile thiols of penetration of oralmucosa by bacterial endotoxins.\" J. Dent. Res. 66: 326.Tonzetich, J., J. M. Coil, and W. Ng (1991). \"Gas chromatographic method for trapping anddetection of volatile organic compounds from human mouth air.\" J. Clin. Dent. 2: 79-82.Tonzetich, J. and P. W. Johnson (1977). \"Chemical analysis of thiol, disulphide and total sulphurcontent of mouth air.\" Arch. Oral Biol. 22: 125-131.Tonzetich, J. and P. W. Johnson (1986). \"Radiochromatographic analysis of H2S-collagencomplexes in rat-tail type I collagen.\" Archs. Oral Biol. 31: 297-300.Tonzetich, J. and R. C. Kestenbaum (1960). \"Odour production by human salivary fractions andplaque.\" Arch. Oral Biol. 14: 815-827.Tonzetich, J. and K. K. C. Lo (1978). \"Reaction of hydrogen sulphide with proteins associatedwith the human mouth.\" Arch. Oral Biol. 23: 875-880.Tonzetich, J. and B. C. McBride (1981). \"Characterization of volatile sulphur production bypathogenic and non-pathogenic strains of oral bacteroides.\" Arch. Oral Biol. 26: 963-969.Tonzetich, J., P. Powell, and J. D. Waterfield (1990). \"Effect of methylmercaptan oncollagenolytic proteinases and cAMP production.\" J. Dent. Res. 69: 274 (Abst#1327).Tonzetich, J. and V. J. Richter (1964). \"Evaluation of volatile odoriferous compounds of saliva.\"Arch. Oral Biol. 9: 39-45.Tonzetich, J. and J. D. Spouge (1979). \"Effect of periodontal therapy on volatile sulphur contentof mouth air.\" J. Dent. Res. 63(Special Issue A): 249.Tonzetich, J., K. Yaegaki, and W. Ng (1985). \"Inhibition of proline transport in fibroblastcultures by methyl mercaptan.\" J. Dent. Res. (64): 306 (Abstract# 1180).Traub, W. and K. A. Piez (1971). \"The chemistry and structure of collagen.\" Adv. Protein Chem.25: 243-352.Tryggvason, K., J. Risteli, and K. I. Kivirikko (1977). \"Separation of prolyl 3-hydroxylase and4-hydroxylase activities and the 4-hydroxyproline requirement for synthesis of 3-hydroxyproline.\"Biochem. Biophys. Res. Commun. 76: 275-281.Tubb, M., F. E. Williams, and L. W. Frazier (1990). \"Gingival crevicular fluid prostaglandinand gingival phospholipids in experimentally induced periodontitis in the dog.\" Archs oral Biol. 35(7): 529-533.Ditto, J. (1979). \"Collagen polymorphism. Isolation and partial characterisation of al (I) timermolecules in normal human skin.\" Arch. Biochem. and Biophys. 192: 371-379.214Uitto, J. and D. J. Prockop (1974). \"Incorporation of proline analogues into collagenpolypeptides; Effects on the production of extracellular procollagen and on the stability of thetriple-helical structure of the molecule.\" Biochim. Biophys. Acta 336: 234-251.Uitto, V.-J. (1987). \"Human gingival proteases. I: Extraction and preliminary characterization oftrypsin-like and elastase-like enzymes.\" J. Perio. Res. 22: 58-63.Uitto, V.-J., H. Larjava, J. Coil, and A. Hieminen (1992). \"Oral rinse elastase: A simpleperiodontal disease screening method.\" funpublished) :Uitto, V.-J. and A.-M. Raeste (1978). \"Activation of latent collagenase of human leukocytes andgingival fluid by bacterial plaque.\" LDent. Res. 57: 844-851.Uitto, V.-J., K. Suomalainen, and T. Sorsa (1990). \"Salivary collagenase. Origin, characteristicsand relationship to periodontal health.\" J. Perio. Res. 25: 135-142.Uitto, V.-J., D. Schwartz, A. Veis (1980). \"Degradation of basement-membrane collagen byneutral proteases from human leukocytes.\" Eur. J. Biochem. 105: 409-417.Villela, B., R. B. Cogen, A. A. Bartolucci, and H. Birkedal-Hansen (1987). \"Crevicular fluidcollagenase activity in healthy, gingivitis, chronic adult periodontitis and localized juvenileperiodontitis patients.\" J. Perk,. Res. 22: 209-211.Waddington, R. J., G. Embery, and K. S. Last (1989). \"Glycosaminoglycans of human alveolarbone.\" Arch. Oral Biol. 34: 587-589.Wahl, L. M. and S. E. Mergenhagen (1988). \"Regulation of monocyte/macrophage collagenase.\"J. Oral Path. 17(452-455):Waite, I. M. (1976). \"A comparison between conventional gingivectomy and a non-surgicalregime in the treatment of periodontitis.\" J. Clin. Perio. 3: 173-185.Walker, C. B., J. M. Gordon, H. A. Cornwall, J. C. Murphy, and S. S. Socransky (1981).\"Gingival crevicular fluid levels of clindamycin compared with its minimal inhibitoryconcentrations for periodontal bacteria.\" Antimicrob. Agents Chemother. 19: 867-871.Walsh, T. F. and M. S. Saxby (1989). \"Inter and intra-examiner variability using standard andconstant force probes.\" J. Clin. Perio. 16: 140-143.Waterfield, J. D. and J. Tonzetich (1989). \"The effect of methyl mercaptan of IL-1 secretion byleukocytes.\" J. Dent. Res. 68: 932 (Abst #522).Weinberg, M., I. Lamster, L. Fiorello, S. Klein, and R. Oshrain (1986). \"Sensitive assay forglucose in human gingival crevicular fluid collected with filter paper strips.\" J. Dent. Res. 65: 206(Abstr.#327).Weinstein, E., I. D. Mandel, A. Salkind, and H. Oshrain (1967). \"Studies on gingival fluid.\"Periodontics 5: 161-166.Weiss, J. B. (1976) Enzymatic degradation of collagen. New York, Academic Press.Welgus, H. G., J. J. Jeffrey, and A. Z. Eisen (1981). \"The collagen substrate specificity ofhuman skin fibroblast collagenase.\" J. Biol. Chem. 255: 9511-9515.215White, D. and D. Mayrand (1981). \"Association of oral bacteroides with gingivitis and adultperiodontitis.\" J. Perio. Res. 16: 259-265.Wiedmeier, V. T., S. P. Porterfield, and C. E. Hendrich (1982). \"Quantitation of Dns-amino acidsfrom body tissues and fluids using high-performance liquid chromatography.\" J. Chromatog. 231: 410-417.Williams, B. L., J. L. Ebersole, and M. D. Spektor (1985). \"Assessment of serum antibodypatterns and analysis of subgingival microflora of members of a family with a high prevalance ofearly onset periodontitis.\" Infect. Immun. 48: 742-750.Wilson, G. B. and W. D. McHugh (1971). \"Gingival exudate. An index of gingivitis?\" Dent.,Practnr. dent. Rec. 21: 261-266.Wilton, J. M. A., J. L. M. Bampton, G. S. Griffiths, M. A. Curtis, H. S. Life, N. W. Johnson,H. R. Powell, G. J. Harrap, and P. Critchley (1992). \"Interleukin-1 beta (IL-1m levels in gingivalcrevicular fluid from adults with previous evidence of destructive periodontitis; a cross sectionalstudy.\" J. Clin. Perio. 19: 53-57.Wislocki, G. B., D. W. Fawcett, and E. W. Dempsey (1951). \"Staining of stratified squamousepithelium of mucous membranes and skin of man and monkey by the periodic acid Schiffmethod.\" Anat. Rec. 110: 359-375.Woolley, D. E. (1984). Mammalian Collagenases. Extracellular Matrix Biochemistry. New York,Elsevier. 119-157.Yaegaki, K. (1990). \"Composition of oral malodorous components and periodontitis.\" J. Dent. Res. 69: 382 (Abstr #2187).Yaegaki, K. (1991). \"Biochemical and clinical factors influencin oral malodour in periodontalpatients.\" J. Dent. Res. 70: 337 (Symposium #574).Yaegaki, K. and K. Sanada (1991). \"Volatile sulfur compounds in mouth air from clinicallyhealthy subjects and patients with periodontal disease.\" J. Perio. Res. : (Submitted).Yaegaki, K., J. Tonzetich, and A. S. K. Ng (1986). \"Improved high-performance liquidchromatography method for quantitation of proline and hydroxyproline in biological methods.\" LChromoto gr. 356: 163-170.Yamada, K. M. (1983). Fibronectin and other structural proteins. Cell Biology of the ExtracellularMatrix. New York, Plenum Press. 95-110.Yamada, K. M., D. W. Kennedy, K. Kimata, and R. M. Pratt (1980). \"Characterization offibronectin interation with glycosaminoglycans and identificationof active proteolytic fragments.\"J. Biol. Chem. 255: 6055-6063.Yamada, S., Y. Aoto, T. Suou, and C. Hirayama (1989). \"Urinary hydroxyproline andhydroxylysine excretions in relations to hepatic hydroxyproline content in chronic liver disease.\"Clin. Biochem. 22(5): 389-393.Yamaguchi, Y., J. Haginaka, M. Kunitomo, H. Yasuda, and Y. Bando (1987). \"High-performance liquid chromatographic determination of desmosine and isodesmosine in tissues and216its application to studies of alteration of elastin induced by atherosclerosis.\" J. Chromatog. 422:53-59.Yoshie, H., M. A. Taubman, D. L. Olson, J. L. Ebersole, and D. J. Smith (1983). \"Periodontalbone loss and immune characteristics of congenitally athymic and thymus cell reconstituted athymicrats.\" J Dent. Res. 62: 273 (Abstract# 932).Zafiropoulos, G.-G. K., L. Flores-de-Jacoby, G. Todt, G. Kolb, K. Havemann, and D. N.Tatnicis (1991). \"Gingival crevicular fluid elastase-inhibitor complex: correlation with clinicalindices and subgingival flora.\" J. Perio. Res. 26: 24-32.Zambon, J. J. (1985). \"Actinobacillus actinomycetemcomitans in human periodontal disease.\" LClin. Perio. 12: 1-20.Zambon, J. J., H. S. Reynolds, and J. Slots (1981). \"Black-pigmented Bacteriodes ssp. in thehuman oral cavity.\" Infect. Immun. 32: 198-203.Zinkevich, 0. D., R. I. Litvinov, and M. S. Kurayskaya (1982). \"Fibronectin and its receptors onthe surface of polymorphs.\" Bull. Exp Biol. Med. 94: 86-88.8. APPENDIXAbbreviations used in this thesisAa^Actinobacillus actinomycetemcomitansAla^alanineAS^arylsulfataseAST^aspartate aminotransferaseBANA^benzoyl-arginine naphthylamideBC^bacterial collagenaseBG^(3-glucuronidasebelbs^P. gingivalis bleb preparationsBI^bleeding indexBoP^bleeding on probingBRL^Bethesda Research LaboratoriesCAL^clinical attachment lossCEJ^cementoenamel junction217CM^carboxy methylCMI^cell mediated immunityCMT^chemically modified tetracylinesDB^distobuccalDES^desmosinedH2O^deionized waterDL^distolingualDPP^dipeptidyl peptidaseEDTA^ethylenediaminetetraacetic acidELISA^enzyme linked immunosorbent assayEM^electron microscopeETAF^epidermal cell thymocyte activating factorFN^fibronecting^gravityGAG^glycosaminoglycansGC^gas chromatographyGCF^gingival crevicular fluidGI^gingival indexGly^glycineHPLC^high performance liquid chromatographyhylys^hydroxylysineHyp^hydroxyprolineIDE^isodesmosineIL-1^interleukin- 1He^isoleucineJE^junctional epitheliumJP/LJP^juvenile periodontitis / localized juvenile periodontitis218LDH^lactate dehydrogenaseLPS^lipopolysaccharidelys^lysineMAB^monoclonal antibodyMB^mesiobuccalMHC^major histocompatability complexMHyp^moles of hydroxyprolineML^mesiolingualMr^molecular weightOAF^osteoclast activating factorOPA^o - phthaldehydePBS^phosphate buffered salinePD^pocket depthPGE^prostaglandin EPITC^phenylisothiocyanatePH^plaque indexPMN^polymorphonuclear leukocytesPro^prolineREE^reduced enamel epitheliumRt^retention timeSDS/PAGE sodium dodecyl sulphate polyacrylamide gel electrophoresisSer^serineTCA^trichloroacetic acidTCA/TCB tissue collagenase A/BTEA^trieklamineTH (CD4+) T helper cellsTNF-a^tumor necrosis factor a219Ts^T suppressor cellsTSC (CD8+) T suppressor/cytotoxic cellsTx^thymectomizedUV^ultravioletVSC^volatile sulphur compoundsWHyp^weighted hydroxyproline"@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "1993-05"@en ; edm:isShownAt "10.14288/1.0086391"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Dental Science"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Biochemical and immunological studies of periodontal disease in humans with emphasis on the analyses of breakdown products emanating from the gingival crevice"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/1829"@en .