Periodontiamedica.com

Hyperreactive Neutrophyls – A Mechanism of tissue destruction in Periodontitis

I– PREFACE
This thesis is basesd on the following articles, which will be referred to in the text by Roman numerals:

I– Protease activity in gingival crevicular fluid. Presence of three protease. Figueredo CMS & Gustafsson A. Journal of clinical Periodontology 1998:25; 306 –310.

II– Activity and inhibition of elastase in GCF.
Figueredo CMS & Gustafsson A. Journal of clinical Periodontology 1998:25; 531 –535.

III– Increased release of elastase from in vitro activated peripheral neutrophilis in patients with adult periodontitis. Figueredo CMS, & Gustafsson A, Asman B and Bergström K. Journal of clinical Periodontology 1999:26; 206 –211.

IV– Increased interleukin–1ß concentration in GCF as a characteristic tarit of patients with periodontitis. Figueredo CMS, Ribeiro MSM, Fischer RG, and Gustafsson A. Journal of Periodontology; resubmitted for publication in February 1999.

V– Increased amounts of laminin in GCF from untreated patients with periodontitis.
Figueredo CMS, & Gustafsson A. Journal of Clinical Periodontology; submitted for publication in March 1999.

II– Introduction 1

Periodontitis

A crevicular accumulation of microbes, follewed by inflammation and immune reaction, are the main features of gingivitis and periodontitis. A small but definite infiltrate of inflammatory cells can be detected in the coronal portion of the connective tissue, although the clinical findings are not typical of inflammation. Thus, a subclinical inflammatory reaction is generally present in clinically healthy gingiva (Page & Schroeder 1976).

Plaque accumulation in the crevice leads to an increase in the inflammatory reaction that can be seeen clinically and microscopically. Until a gengival lesion becomes “established” (Page & Schroeder 1976), this pattern persists, with inflammation following plaque accumulation. At a certain stage in this process, marked irreversible tissue destruction begins. Degradation of the connective tissue is followed by junctional epithelial migration and bone resorption. This stage is at the border line between gingivitis and periodontitis. It is not clear why some lesions remain localized on the marginal part of the gingival tissues, while others go on to loss of connective tissue attachment and supporting alveolar bone. Moreover the severity of periodontal tissue damage often varies from tooth to tooth and even from surface on the same tooth in the same subject.

Epidemiological studies show that the frequency of the severe form of adult periodontitis is very low in the population. The percentage of subjects with the disease tends to increase with age, reaching a peak at 50–60 years. Brown et al. (1989) reported that only 3.4% of the United States population between 19–44 years had 1 or more pockets deeper than 6 mm. This percentage increased to 16.5% in the population between 45–64 yeras, and decreased to 14 % in those 65 years or more.

Baelum et al. (1986) studied adult tanzanians aged 30–69 years, having large amounts of plaque and caculus, and found that only 19% showed pockets deeper than 3 mm and attachment loss > 6mm. Seventy–five percent of the tooth sites with attachment loss 7 mm were found in 31% of the subjects, showing that advanced periodontal disease does not correlate with supragingival plaque levels. The same investigators (Baelum et al.1988) observed, in a 15 to 65 year old population in Kenya where poor oral hygiene was reflected by plaque, calculus and gingivitis, that pockets 4 mm deep were found on less than 20% of the surfaces, and the proportion oof sites per subject with deep pockets and advancedloss of attachment showed a markedly skewed distribuition.

The current view of the role of micro–organisms as the principal etiologic factor in periodontal diseases can be summarized by saying that the periodontopathic bacterial flora is “necessary, but not sufficient to cause disease” or that periodontal disease are” specific mixed infections which cause periodontal destruction in an appropriately host” (Offenbacher 1996). This renewed emphasis on the host–response in pathoogenesis is based on three factors: (1) the inflammataryresponse varies greatly from one subject to another (Movig et al. 1988, Pociot et al. 1992, Shapira et al. 1994). (2) epidemiological population studies show that microbial parameters account for only some of the incidence and prevalence of the disease. Other factors, such as smoking, stress, systemic diseases, genetic and biochemical markers of inflammation, also contributesignificantly to multivariate models of disease expression, independently of the microbial infections (Beck at al. 1990; Beck et al. 1992; Christersson et al. 1992; Eheeler et al. 1994). (3)Epidemiologic studies in twins have suggested that, overall, about half the variability of periodontal disease expression is controlledby genetic, not microbial, factors (Michalowicz 1993; 1994).

Wolff et al. (1993) found that the frequency and levels of five gram–negative anaerobic bacteria, P. gingivalis, A. actinomycetemcomitans, P. intermidia, E. corrodens and F. nucleatum, correlated with the probing depth of 6905 sites sampled. This could be either an indicator of disease or simply the result of a more favourable environment. Deep pockets favour anaerobic bacteria, such as those mentioned above but they can also be found in shallow pockets, even in subjects without periodontitis (chen et al. 1989; Dahén et al. 1992). According to Wolff et al. (1993) the finding that periopathogens frequently inhabit sites not associated with advanced disease suggests that a susceptible host is necessary.

Although reports on the presence of still unidentified etiologic organisms or pathogenic clonal types cannot be disregarded, it is not correct to say that specific organisms are the cause of severe disease. According to Kornman & di Giovine (1998), the response to chronic inflammation may be modified by factors that do not directly cause the disease, but modify some aspects of it to aggravate the clinical conditions. The chronic reaction of an inflammation may be disproportionated to the pathogenic challenge, resulting in a response detrimental to the host. Smoking (Harber 1994), diabetes (Salvi et al. 1997c) and genetic influences (Kornman et al. 1997) may place some subjects at relatively high risk for an increase in the severity of periodontitis.

1.1– Risk factors

A risk factor is part of the causal chain of a particular disease or can lead to exposure of the host toa a disease (Beck 1994). The presence of a risk factor implies a direct increase in the probability of that a disease will occur, and if a risk factor is absent or removed, a reduction will probably ensue. This differs from markers, which generally result from a disease, vary over time and contribute to the natural history of disease progression (Salvi et al. 1997b)

1.1.1 Diabetes

Diabetes entails a risk of periodontitis, with an odds ratio of 2 to 3 for diabetics, as compared to non–diabetics (Salvi et al. 1997b). The increase in susceptibility seems to be associated more with an aberrant host–response (Salvi et al. 1997a;c) than with differences in the microbiota (Mandell et al. 1992). Diabetics may increase susceptibilityto periodontitis by impairing neutrophil chemotaxis and phagocytosis (McMullen et al. 1981). Moreover, monocytes from patients with diabetes secrete more prostaglandin E, TNF– and interleukin–1 after stimutalion with lipopolysaccharide (Salvi et al. 1997a; Salvi et al. 1998).

1.1.2– Smoking

Smoking is a well–accepted risk factor in periodontitis. In a meta–analysis, Papapanou (1996) confirmed that smoking is a risk in periodontitis, with an odds ratio of 2.8. The pathogenic mechanisms underlying this are not clear. regarding subjects with gingivitis, no differences were found between smokers and non–smokers in the composition of microbiota or the plaque accumulation rates (Bergström & Preber 1994). Preber et al. (1992) reported that patients with periodontitid who smoke do not differ from those who do not, with respect to possible periopathogens. However, Zambon et al. (1996) reported that smokers had ssignificantly higher levels of B. forsythus and were at significantly greater risk of infection with this periodontal pathogen than non–smokers. Smoking causes several changes in the response to inflammation – e.g., it can impair the chemotaxis and phagocytosis of normal peripheral neutrophils (Kraal & Kenney 1979) and alter their oxidative bursst (Ryder et al.1998). Locally, Numabe et al. (1998) showed that salivary neutrophilis intensify their phagocytic activity after smoking.

1.1.3– Genetic

the clearest evidence of genetic factors in adult periodontitis ccomes from periodontal findings in adult twins. More than 50% of the variance in several clinical and radiographic measures appears to be explained by genetic factors (Michaliwicz et al. 1991a; Michaliwicz et al. 1991b; Corey et al. 1993).

Kornman et al. (1997) showed that a composite genotype including 2 polymorphisms in the cluster of IL–1 genes (allele 2 of IL–1B +3953 plus allele 2 of IL–1A –889) was significantly overrepresented in the severe periodontitis group. Gore et al. (1998) found that the frequency of IL–1 genotypes, including the IL–1 + 3953 allele 2 (IL– 1 + 3953 1/2 and IL–1 + 3953 2/2), previously correlated with increased IL–1 production, was significantly higher in patients with advanced than in those with early–to–moderate adult periodontitis.

2. Neutrophils

Neutrophils are produced in the same bone mmarrow, where the myeloid precursor cells mature to segmented neutrophils in about 9 days, and are then released into the circulation. The cells remain in the bloodstream for a relatively short time (half–life = 6–7h) and in the tissue for 1–4 days . They can be rapidly mobilized within 30 mim to the site of an bacteria. In man, neutrophils are the predoominant leukocyte in blood, comprising between 50% and 70% of the circulation pool of leukocytes.

Neutrophils exist in three states: quiescent, primed and actiivated. Neutrophil activation seems to be mediated by a rise in cytosolic free Ca 2+, which correlates with activation of the oxidase system. “Primed” means that the cells are “ready to go”, but await a further stimulus before the oxidase response is elicited . A resting cell can receive a priming or an activation stimulus. Only the activated cells show oxidase activity but, if the primed cell also receives an activating stimulus, the ensuing oxidase activity will be higher than that in unprimed, activated cells (Hallet & Lloyds 1995). Greater chemotaxis and degranulation have also been found when neutrophils are primed with cytokines, such as TNF– (bajaj et al. 1992) and interleukin–1 (Brandolini et al. 1997).

Cytokines are ssolube protein mediators of immunity. They produce their effects on target cells via specific membrane receptors. Cytokines are generally rather small polypeptides (17 – 20 kDa), which are divided into four major groups: interferons, TNF, interleukins (IL) and growth factors. Most of them are not stored in secretory granules, but are synthesized when needed (Leffell 1997). Although T–cells, natural killer cells and monocytes/macrophages produce large amounts of cytokines, nearly all cells infiltrating an inflamed tissue are neutrophils, which may therefore be a large source of cytokines. At least two interleukins, IL–1 and IL–8, have a strong association with neutrophils and periodontal diseases.

A relation between interleukin–1 (IL–1) and perioddontal disease has been reported in many studies. IL–1 is a potent proinflammatory cytokine that can influence the host–response in the periodontal lession. It is produced in and forms, wich are different gene productos, but utilize the same cellular receptor. Both are produced mainly by macrophages, but alsso partly by other cells, including neutrophils (Tokoro et al. 1996; Galbraith et al. 1997). IL–1 is the major inflammatory cytokine in gingival tissue associated with periodontitis (Tokoro et al. 1996). It is formed and released in response to several immunostimlatory agents – e.g., lipopolysaccharide (LPS) and tissue degradation products (Hanazawa et al. 1985) – and can activate endothelial cells to upregulate ICAM–1 and E–selectin expression (Bevilacqua et al. 1985; Carlos & Harlan 1994), wich may increase the diapedesis of leukocytes. IL–1 has also been found to activate osteoclasts (Dewhirst et al. 1985). All these IL–1effects increase the inflammatory response and can subsequently cause degradation of the periodontal ligament and alveolar bone.

Interleukin–8 (IL–8) is a small cytokine that exhibits chemotactic activy against neutrophils, but not against nmonocytes 8Harada et al. 1996). It can be produced in vitro by a wide variety of cell types, including monocytes, lymphocytes, neutrophils and keratinocytes (mukaida et al. 1995). Il–8 is not constitutively produced, but occurs in response to an inflammatory stimulus and has a strong effect on the migration and activation of neutrophils – e.g., it can induce the degranulation of neutrophils (Brandolini et al. 1997).

2.1 Antimicrobial functions

The neutrophil recognizes and combats foreing substances by their receptors for several surface structures – for instance, the fc–portion of attached antibodies, complement factors, capsule reactive proteins and the LPS–complex. Antibodies(especially IgG) and complement proteins like C3b can opsonize and are therefore referred to as “opsonins”. Opsonization is the process whereby particles such as micro–organisms become coated with molecules, which allow them to bind to receptors on phagocytes.IgG antibodies bind to the antigens in the Fab region,leaving the Fc region sticking out. phagocytes have Fc gamma receptors(FcR) and they can therefore bind to the coated molecules and internalize them.

Three classes of FcR are currently distinguished: (1) FcRI (CD64) – a 72 kDa high–affinity receptor for IgG, constitutively expressed on monocytes, macrophages,myeloid progrenitor cells and dendritic cells; (2) FcRII (CD 32) –a broadly distributed 40 kDa molecule, found in most types of blood leukocytes, “Langerhans” cells, various populations of endothelial cells, dendritic cells, macrophages and platelets; and (3) FcRIII (CD16) – a glycoprotein having a molecular weight ranging from 50 – 80 kDa with two gene expressions. The FcRIIIA gene product is a transmembrane receptor named FcRIIIa, found on natural killer cells, macrophages, subpopulations of T–cells and fresh isolated blood monocytes, immature thymocytes and placental trophoblasts. The FcRIIIB gene product is FcRIIIb, which is celectively expressed on neutrophils (for a review, see Rascu et al. 1997).

Activated neutrophils csn destroy foreign substances by at least three mechanisms: 1) respiratory burst, 2) cytolysis and 3) phagocytosis and degranulation.

2.1.1 Respiratory burst

The respiratory burst realeases highly reactive oxygen metabolites extracellularly. Oxidative metabolism mainly generates hypochlorous acid (HOCI), because of the relatively high concentration of CL in plasma (Weiss 1989), which has both tissue–destructive and antimicrobial effects.

2.1.2 Cytolysis

This is a mechanism by which antibacterial substances reach the bacteria. Considering the high concentration of neutrophils in inflamed tissue, the release of antibacterial agents from dying cells may be an important host–defense mechanism(McNamara et al. 1988).

2.1.3 Phagosytosis and degranulation

Phagocytosis is a process in which the neutrophil isolates the organism intracellulary by creating a phagosome . Fusion between the phagosome and cytoplasmic granules deliveres high concentrations of antibacterial substances into the vacuole.

Degranulation leads to the release of anti–bacterial substances extracellularly(Wright 1988) from the primary (azurophil) granule(defensins, lysozyme, myeloperoxidase, cathepsin G and elastase) and the secondary 8specific) granule 8lactoferrin, collagenase and lysozyme). A tertiary granule containing gelatinase has also been described (Dewald et al. 1982). The degranulation of the primary and secondary granules is of special interest in this study. Bentwood & Henson(1980) reported that human neutrophils respondednto soluble stimuli withh sequential release of, first, the secondary(specific) granule constituent, followed by the constituents of the primary granule.

2.2 Neutrophils and tissue destruction

Neutrophils are associated with tissue destruction in chronic inflammation – e.g., adult respiratory ditress syndrome (Lee et al. 1981) and emphysema(Janoff 1985). During periodontal inflammation, increased numbers of neutrophils have been found in sevarel morphologic studies of the inflamed gingiva. Attström & Egelberg (1970) examined leukocytes in clinically healthy and chronically inflamed sites in humans and dogs. The leukocytes were present in both healthy and inflamed sites. Differential counts showed 95–97% neutrophils, 1.2% lymphocytes and 1–3% monocytes.The proportions remained the same, although the number of leukocytes increased with inflammation. In another study by the same author (Attström 1971), leukopenia was induced with nitrogen mustard and with heterologous anti–neutrophil serum in dogs having chronically inflamed gingiva. The leukopeina rediuced the number of crevicular leukocytes, the enzymatic activity and the volume of gengival crevicular fluid. Kowashi et al.(1979), who counted neutrophils in gingival washings during experimental gingivitis in man, found a doubling of the number of cells from day 0 to day 20. Although the local number of neutrophils are associated with the degree of inflammation, it is not conclusively shown that a marginal chronic gingivitis evolves into a perodontitis lesion. Gustafsson et al. (1994) showed that, in chronic gingivitis and in periodontitis sites, the numbers of neutrophils seem to be very similar, supporting the concept that hyperactivity per cell, rather tham differences in the local cell numbers, is important for initiation of tissue destruction.

The term “hyperactive” was chosen to describe higher activity of neutrophils in an inflamed site, while “hyperreactive” was chosen to describe higher reactivity of peripheral after in vitro activation. Studies of peripheral neutrophils from patients with juvenile periodontitis, uisng the production of oxygen radicals as a marker for neutrophil activity, have shown hyperreactivity after Fc–receptor stimulation (Asman 1988; whyte et al.1989; Shapiran et al.1991). The same phenomenon has recently een seen in adults with periodontitis(Gustafsson & Asman 1996; Fredriksson et al. 1998). Chollet–Martin et al.(1992) called “hyperresponsive” a subpopulation of neuttrophils, isolated from patients withacute respiratory distress syndrome(ARDS), which showed an increased capacity to generate hydrogen peroxide (H2O2). Although the hyperreactivity of neutrophils from periodontitis patients might be related to the generation of oxygen species other than HO (Fredriksson et al.1998), the accumulating neutrophils may produce extensive tissue damage by generating reactive oxygen species in both diseases.

Besides reactive oxugen species, there are strong indications that neutrophil proteases can injure the lung and periodontal tissues. Neutrophil proteases can degrade lung and arterial–wall elastin(Janoff & Zeligs 1968), cause emphysema in animal(Weinbaum et al. 1974) and collagen degradation(Janoff 1983). Increased amounts of neutrophil–derived substances, such as (beta)–glucuronidase(Lamster et al. 19949, neutrophil collagenase(Overall et al. 1987) and elastase(Gustaffsson et al. 1992), have been found in GCF in deen pockets in periodontitis patients.

The release of elastase from in vitro–activated peripheral neutrophil has been shown to differ significantly in patients with juvenile periodomtitis, as compared to healthy controls(Asman 1988), but it has not been shown in patients with adult periodontitis.

2.2.1 Neutrophil elastase

One of major end–products of neutrophil activity is the extracellular release of active and inactive proteases. Elastase is a neutral serine protease(33kDa), synthesized primarily in promylocytes and stored in the cytoplasmatic azurophil granules of maturing neutrophils in amounts ranging up to 3 picogramas per cell(Janoff 1985). Although neutrophils contain large amounts of elastase, they have no elastase mRNA transcripts, indicating that matureneutrophils cannot produce this enzyme. The expression of the elastase gene is limited to a very short period in leukocytes differentiation(Fouret et al. 1989).

Elastase can degrade many important proteins the extracellular matrix, such as elastin (Janoff & Zeligs 1968), laminim (Heck et al. 1990), fibronectin and collagen (Janoff 1985, Owen & Campbell 1995). The destructive capacity of released elastase is usually offset by the protease inhibitors –1–antitrypsin(A1AT) and –2–macroglobulin (A2NG). A1AT (52 kDa) seems to be the main regulator of elastase. This is a glycoprotein synthesized in the liver, and it can also be produced locally by macrophages and neutrophils(Bois et al. 1991; Pääkkö et al.1996). Normally, A1AT is present in such abundance that all active proteases are inhibited within milliseconds. However, elastase can avoid inhibition by three mechanisms: 1) when the enzyme is released in closed compartments in concentrations sufficient to overwhelm the available inhibitor; 2) release of enzyme in close proximity to its substrates, which then successffully compete with A1AT ti bind elastase; 3) enzyme released in sites where local A1AT molecules have been inactivated by oxidation.

The respiratory burts generates reactive oxygen radicals that are released extracellularly. Oxigen radicals are tissue–destructive per se (Weiss 1989), but they can also act in concert with simultaneously released proteases. Extracellularly–released oxigen radicals can oxidatively inhibit A1AT and thus allow the proteases to degrade matrix proteins close to the neutrophils (Weiss 1989). The facilitation of neutrophil elastase activity by oxidative inactivation of an inflammatory reaction. It can degrade tissue inhibitors of metalloproteinases (Weiss 1989) and, at the same time, may be involved in the extracellular activation of latent metalloproteinase (Ferry et al. 1997). Elastase is able to activate pro–geletinase B, which is one of the majo factors in neutrophil migration across the basement membrane (Delclaux et al. 1996) while causing extensive damage to the basement membrane by degrading laminin (Heck et al. 1990).

Several studies have shown increased elastase activity against specific low molecular weight substrates in periodontitis (e.g., Gustafsson et al. 1994; Ingman et al.1994) and increased levels of elastase activity have also been suggested as a perdictor of periodontal disease progression8Palcnis et al. 1992; Armitage et al. 1994). Such substrates are hydrolyzed by free elastase and bound to A2NG (Travis & Salvesen 1983). This means that measurements of elastase activity with these substrates cannot distinguish between free elastase and the elastase–A2MG complex. Moreover, to our knowledge, the presence of free elastase activity in GCF from patients with adult periodontitis has not been convincingly shown.

In summary, elastase is very important for normal tissue turnover and for combating infection, but an excessive release in the active form together with a higher production of oxigen radicals, might cause partial inactivatiion of the major inhibitor, A1AT, and thus lead to extensive tissue destruction.

III – AIMS

The main goals in the thesis were: (1) to detect protease activity that might explain differences in site–specific tissue destruction; (2) to find in vitro evidence of peripheral neutrophil hyperreactivity related to the release of elastase; (3) to determine whether increased levels of interleukin–1(beta) could explain the local neutrophil hyperreactivity; (4) to show that inflamed sites from periodontits patients have more activated neutrophils tham inflamed sites from healthy controls, by measuring basement membrane degradation; (5) and to determine whether there is a relation between levels of IL–8 and site–specific tissue destruction.

1. Specific aims

Paper I – To determine whether free protease activity is present in GCF.

Paper II – To determine whether free elastase is present in GCF.

Paper III – To determine whether there free differences in the release of first and secundary granule between peripheral neutrophils from periodontitis patients and healthy controls.

Paper IV – To test the hyputhesis high levels of IL–1(beta) in GCF are characteristicof patients with periodontitis.

Paper V – To test the hypothesis that the presence of hyperactive neutrophils generates more basement membrane degradation in the inflamed sites of periodontitis patients.

IV MATERIAL AND METHODS

1– Clinical parameters

The presence of supraginginval plaque (PI) and gingival inflammation(GI) were recorded using the criteria of Silness & Löe (1964) (Papers I, II, IV and V). Only sites with a GI value of 1 or 2 were included. In sites without inflammation (value 0), the volume of gingival crevicular fluid (GCF) was insufficient to permit measurement. It was usually impossible to examine very infalmed sites (GI=3), because of difficulty in avoiding contamination with bllod. Pocket depht was measured with calibrated periodontal probe.

2– Gingival crevicular fluid samples

In Papers I and II, GCF samples ewrw collected with repeated intracrevicular washings, as described by Salonen & Paunio (1991), to avoid retention of free elastase in the more commonly used paper strips (Gustafsson 1996). Each pocket was washed five times with 5l of PBS during continuous aspiration . All samples from sites in the same category in each person were pooled and diluted with PBS up to 1 ml. Since one can not measure the volume of GCF with the intracrevicular washing method, we related protease activity to the concentration of transferrin in the wash–fluid, to obtain a semiquantitative estimate (Asma et al.1981). The amount of transferrin in GCF increases with fluid volume, probably because of leakage from plasma(Adonogianaki et al. 1994) and can therefore, be viewed as a marker of GCF volume.

In Papers IV and V, the sites were sampled with paper strips. The strip was inserted into the pocket until mild resistance was felt and kept there for 30sec. The volume of GCF was measured with a Periotron 8000 ®GCF meter: Before each study, the unstrument was calibrated, using saline delivered with a Hamilton syringe.

3– Protease activity

Free preotease activity was measured with FITC–conjugated casein. The mixture of casein and sample and incubated for 20h at +37°C, during agitation. After incubation, trichloroacetic acid 0.6M/l was added to stop the reaction and percipitate the intact casein molecules. After 30 minutes, the samples were centrifuged at 3000g for 10 minutes. The amountesof FITC–conjugated cassein fragments in the aupernant corresponds to the amount of protease activity in the samples. Fluorescence was analysed in a fluorescence spectrophotometer, at an excitation wavelength of 488 nm and an emission wavelength of 520 nm.

4– Elastase activity

The levels of elastase activity were measured with the low molecular weight substrate(445.5 Da) L–pyroglutamy–L–propyl–L–valine–p–nitroanilide. The substrate is highly specific for granulocytes elastase(Tanaka et al. 1990) but it is also hydrolyzed by the elastase––2–macroglobulin complex (Wewers 1988). To distinguish between activity derived from free elastase and elastase bound to A2NG, an excess amount of A1At was added to the samples. A1AT is a very effective inhibitor of free elastase activity, but it can innhibit activity from the elastase bound to A2MG (Travis & Salvesen 1983). The activity inhibited by A1AT could thus be seen as derived from free elastase and the remaining activity from the elastase–A2MG complex.

5– ELISA assays

5.1 –Elastase – A1AT complex

A polyclonal antibody against A1AT was coated on a 96–well microtitre plate overnight at + 40°C. The wells were washed five times with PBS + 0.05% Tween ®20. Samples or standards were added to each well and incubated for one hour at +37°C. The wells were washed again, as above. and an alkanline phosphate–conjugated polyclonal sheep antibody aginst elastase was added and incubated for one hour at + 37°C. After a final washing, the substrate p–nitrophenol–phosphate was added to each well and the absorbance was read after 10 min at 405 nm in a spectrophotometer.

5.2– Antigenic elastase

A monoclonal antibody against elastase was coated on a 96–well microtitre plate overnightnat +4¼C. The wells were washed four times with PBS + 0.05% Tween®. Diluted samples or standards were added to each well and incubated for 1h at +37¼C. The wells were washed again, as above, and a polyclonal sheep antibody against elastase was added to them and incubated for one hour at + 37¼C. After another wash, the third antibody, an alkaline phosphatase–conjugated rabbit ant–goat IgG, was added and incubated, as above. After a final washing, the substrate p–nitrophenol–phosphate was added to each well. The absordance at 405nm was read in a spectrophotometer after 10 min. The antigenic elastase assay does not detect elastase bound to A2MG.

5.3– Lactoferrin

A microtitre plate was coated with a monoclonal mouse antibody against lactoferrin and incubated overnight at + 4¼C. The wells were washed five times with PBS + 0.05% Tween¨. Samples and standards were added to the wells and incubated for one hour, washed once again and the second antibody, a polyclonal rabbit anti–lactoferrin, was added. The plates were incubated for one hour at +37¼C and washed five times before adding of the third antibody, an alkaline phosphatase–cconjugated polyclonal goat antibody against rabbit IgG, and incubating for one hour. The phosphatase activity was measured after the additing of the substrate p–nitrophenol– phosphate and the absorbance at 405 nm was read in a spectrophotometer after 10 minutes.

5.4– –1–antitrypsin (A1AT) and –2macroglobulin (A2MG)

Antigenic A1AT and A2MG were measured with a one –layer ELISA. The microtitre plate was coated with samples and standard, diluted in carbonatee buffer, pH 9.6, and icubated overnight at +4¼C. After incubation, the plate was washed four times with PBS + 0.05% Tween¨. The antigen was detected with a polyclonal rabbit antibody against A1AT or A2MG during incubation at +37¼C for one hour. After another washing, as described above, the final antibody, a goat anti–rabbit IgG conjugated with alkaline phosphatase, was added. The plate once again incubated at +37¼C for one hour and washed, as above. The phosphatase activity was measured after the addition of the substrate, p– spectrophotometer after 10 minutes.

5.5– Interleukin–1 andInterleukin–8

A monoclonal antibody to IL–1 or IL–8, diluted in carbonate buffer, was coated onto the microtitre plates overnight at +4¼C. After the coating, the plates were washed four times with PBS+0.05% Tween, blocked with 1% HSA, and incubated for 1h at room temperature. After washing, as above, a standard curve and samples were coated onto the plates. The plates were incubated at +37¼C and shaken for 45 min and washed as above. The detection antibody, a biotinylatedpolyclonal goat anti– IL–1 or IL–8, was incubated at +37¼C and shaken for 45min. After washing, as above, the horse radish peroxidase (HRP)–conjugated streptavidin was added to the plates and incubated in the same way as the detection antibody. The plates were washed once again, as above, before the HRP–substrate, 3,3«, 5,5«–tetramethyl–benzidine, was added. The reaction was stopped with 1M H2SO4 after approximately 10min and the absorbance at 450 nm was read in a spectrophotometer.

5.6– Laminin

the standard curve and diluted samples were added to the microtitre plates and incubated overnight at +4¼C. After washing the plates, a polyclonal rabbit antibody against laminin, diluted in PBS + 0.1%HSA, was used as a primary antibody. The plates were then incubated during shaking 45 min at 37¼C. After another washing, the second antibody, a biotinylated polyclonal goat anti–rabbit IgG diluted in PBS+0.1% HSA, was added and incubated during shaking at +37¼C fro 45 min. After a third wash, as above, HRP–conjugated streptavidin was added to the plates and incubated in the same way as the detection antibody. The plates were once again washed, as above, before the HRP–substrate, 3,3« , 5,5«–tetramethyl–benzidine, was added. The reaction was stopped with 1M H2SO4

5.6– Laminin

the standard curve and diluted samples were added to the microtitre plates and incubated overnight at +4¼C. After washing the plates, a polyclonal rabbit antibody against laminin, diluted in PBS + 0.1%HSA, was used as a primary antibody. The plates were then incubated during shaking 45 min at 37¼C. After another washing, the second antibody, a biotinylated polyclonal goat anti–rabbit IgG diluted in PBS+0.1% HSA, was added and incubated during shaking at +37¼C for 45 min. After a third wash, as above, HRP–conjugated streptavidin was added to the plates and incubated in the same way as the detection antibody. The plates were once again washed, as above, before the HRP–substrate, 3,3« , 5,5«–tetramethyl–benzidine, was added. The reaction was stopped with 1M H2SO4 after approximately 40 min, and the absorbance at 450nm was read in a spectrophotometer.

5.7– Transferrin

Samples and standards were coated on a 96 micro–well plate overnight at +4¼C. After incubation, the plate was washed four times with PBS +0.05% Tween. One hundred l of a polyclonal rabbit antibody against trasferrin was added to the plate and incubated for onw hour +37¼C. The plate was washed again and the second antibody, an alkaline phosphatase–conjugated swine anti–rabbit immunoglobbulin, was added. After one more hour if incubation and still another wash, 100 l of the substrate p–nitrophenol–phosphate was added. The absorbency at 450nm was read after 30 min in a spectrophotometer.

6. Flow cytometric analysis

To confirm the presence of intrcellular stores of A1At, 250 000 leukocytes were fixed and permeabilized. After permeabilization, 2l of a rabbit polyclonal antibody against A1AT was added to the cells and incubated for 20 min during aditation, followed by the addition of 2l of a FITC–conjugated polyclonal swine antibody against rabbit IgG. A rabbit IgG fraction was used as the negative control. Flow cytometric meassurements were mada after gating the subpopulations on a histogram with foward– and side– scatter. The mean fluorescence intensity and the staining in percentage were recorded.

7– Cell Preparation

Venous blood was collected into vacuum tubes containing EDTA and allowed to rest for 1 hour at room temperature. A leukocyte–rich preparation from 7ml of venous was made by lysing the red blood cell in whole blood at 4¼C for 10 min. with 0.83% NH4Cl solution (NH4 Cl 8.3g, KHCO3 1.0g, Na2 EDTA 2H2O 0.04 g and distilled water to 100 ml) suplemented with 0.25% HSA, ratio of blood to reagent = 1:7. After centrifugation (350 g, 5 min), the harvested leukocyte pellet was washed twice with PBS/HSA and stored at + – 0¼C in the same buffers for one hour.

8– Degranulation

The leukocytes (1.0 x 106 netrophils) stored in PBS + 0.25% HSA, were mixed with IgG–opsonized Staphylococcus aureus (75 bacteria per neutrophil) (Bergström & Asman 1993). Tubes without bacterial stimulation were used as controls for each subject. The mixture was further diluted with Hankks«Balanced Salt Solution up to a final reaction volume of 1 ml, and incubated for one hour during horizontal agitation. The reaction was stopped by centrifuging at 1000 g for 5 minutes. The supernatant was removed and stored at –70¼C, pending analysis. the pellet containing the leukocytes was diluted up to 1ml with distilled watter and the cells were homogenized by repeated freezing and thawing (four times), centrifuged again (1000g for 5 min.) to remove the cell fragments and stored at –70¼C, pending analysis.

9– Statistical analysis

The significances of differences between patients and controls were calculated with the Mann–Whitney U–test and of differences between the categories of sites in periodontitis patients with the Wilcoxon signed–rank test. The correlations were calculated with the Spearman rank correlation coefficient or the Kendall rank correlation coefficient.

V– RESULTS

In Paper I, we measured non–specific protease activity in GCF from inflamed sites with or without tissue ddistruction. The clinical findings are shown in Table 11. The free protease activity was higher in samples from inflamed sites with tissue destruction in subjects having periodontitis (PP), than in samples form inflamed sites without such destruction in the same subjects (GP) and samples from subjects with gingivitis alone (GG). The difference between PP and GP was significant (p=0.0019). There was a large variation in protease activity, but the highest activities were found mostly in samples from the periodontitis patients. PP showed a mean protease activity, corresponding to 230 ng trypsin (SD± 481.6), 52 ng in GP (SD±174.7) and 25 ng in GG (SD± 33.6). Median values are presented in Table 2. The residual protease activity, i.e. the activity measured after the addition of excess A1AT, was usually low in the GP samples than in GG and PP samples. The addition of A1AT inhibited a larger proportion of the protease activity in the GP sample than in the GG and PP samples (Table 2). The transferrin concentration was significantly lower in the GP samples than in the other two categories . We found no difference between GG and PP samples. The ratio of protease activity to trasnferrin was higher in PP than in GP. A significant difference was observed between GP and PP (p=0.0001) (Table 2).

Table 1. Comparison of mean values of clinical findings (± standard deviation) between GG (gingivitis in gingivitis patients), GP (gingivitis in periodontitiss patients) and PP (periodontitis in periodontitis patients). GI– gingival index; PI plaque index; PPD–probing pocket depth.
GG (n12)     GP (n19)     PP (n:19)
GI     1.0 (±0.4)     1.1 (±0.3)     1.3 (±0.4)
PI     0.9 (±0.5)     1.3 (±0.6)     1.5 (±0.5)
PPD (mm)     1.7 (±0.7)     2.4 (±0.6)     6.3 (±0.9)

Table 2. Median of protease activity (after inhibition with A1AT) expressed in ng, median of the amounts of transferrin expressed in ng, the ratio between protease activity and transferrin and mean of protease inhibition expressed as a %.
Test / Site     GG     P1     GP     P2     PP     P3
Protease activity     15.7     NS     5.0     p<0.002     56.9     NS
Residual activity     2.1     NS     0.3     p=0.02

2.2     NS
Transferrin     355.1     p<0.04     194.6     p=0.003     345     NS
protease/transf     0.038     NS     0.023     p<0.001     0.140     NS
Mean inhibition     86     NS     94     p=0.041     88     NS

P1: Differences between GG and GP, calculated with Mann – Whitney U–Test. P2: Differences between GP and PP, calculated with Wilcoxon signed–rank test. P3:Differences between GG and PP, calculated with Mann–Whitney U–teste.

In paper II, elastase activity was measured with a low molecular weight substrate and elastase bound to A1AT was measured with with an ELISA. To distinguish between free elastase and elastase bound to A2MG, we added an excess of A1AT to the samples. The activity inhibited by A1AT was considered as free elastase and the uninhibited activity as derived from the elastase–A2MG complex (c.f. materials and methods, page 19). The amounts of free elastase, total elastase and elastase bound to A2MG, were significantly higher in PP than in GP and GG (Table 3; Fig. 1). There were no significant differences in the amounts of ealstase bound to A1AT betwenn the groups. Mean values are shown in Table 3.

The amounts of transferrin in the wash fluid were similar in the PP and GG sites while they low in the GP sites. Since the distribution was skewed, the median concentration of transferrin is shown in Table 3 .

The ratio of total elastase to tranferrin was significantly higher in PP samples than in GP and GG samples. The ratio between free elastase and transferrin was also significantly higher in PP sites than in Gp and GG sites. It was almost twice as high in GP sites as in clinically similar GG sites(table 3). The was a strong negative correlation between the percentage of free elastase and the percentage of A1AT complex (R= – 0.93) calculated in all samples, n= 50 (Fig.2). We found a positive correlation between free elastase and the non–specific protease activity analyzed previously(Paper I), R= 0.41 in GG and 0.71 in PP (date not shown).

Table 3. Mean values of total elastase, free elastase, elastase inhibited by –1–antitrypsin (A1AT) and by –2–macroglobulin (A2MG), exprressed as cell equivalents x 10(3). The ratios of total elastase to tranferrin and of free elastase to transferrin are expressed as mean the amounts of transferrin are expressed as median.
Test / Site     GG (n12)     P1     GP (n 19)     P2     PP (n19)     P3
Total elastase     67.5     NS     64.3     <0.001     162.9     0.005
Free elastase     17.7     NS     28.8     <0.001     79.9

0.007
Elastase – A1AT     42.8     NS     27.6     NS     61.8     NS
Elastase A2MG     8.4     NS     9.6

0.001     21.1

0.014
Free elastase/transf     59.7     NS     104.5     0.025     339.0     0.012
Total elastase/transf     148.6     NS     263.4     0.05     572.5

<0.001
Transferrin – ng     355     0.036     195     0.003     345     NS

p1: Differences between GG and GP, calculated with the mann–Whitney u ––test. p2: Differences between GP and PP, calculated with the Wilcoxon signend–rank test. p3: Significance of differences between GG and PP, calculated with the Mann–whitney u –test. NS: not significant.

Fig.1. Mean values of total elastase, free elastase, and of elastase bound to A2MG and to A1AT, expressed in cell equivalents x 10(3).

Fig.2. Correlation between the percentage of a free elastase and the percentage of elastase –– 1–atitrypsin complex (A1AT) in all samples (n = 50).

In Paper III, we found that the release of elastase from peripheral neutrophils after stimulation with opsonized bacteria was significantly higher in patients with periodontitis than in healthy controls (p= 0.01) (Table 4, Fig.3). There was also a significant difference between the groups in the unstimulated sample s(p= 0.024)(Table 4). The total amount of elastase, i.e. the sum of the elastase activity released and the remaining activity in the cells, were similar in the unstimulated samples from the two groups. During bacterial stimulation, however, the total amount of elastase increased in samples from patients (p= 0.013). In contrast, no corresponding increase could be seen in the control group(Fig. 4). The total amount of A1AT and the amount of elastase –A1AT complex in the supernatant from the stimulated cells were higher than from the unstimulated cells in both patients and controls, but the differences between the groups were not significant (Table 4). The flow cytometric analysis showed that the granulocytes contained intracellular stores of A1AT and that only negligible amounts of the inhibitor were present in monocytes and lymphocytes. The release of lactoferrin, used as a marker of the release of secondary granula, was the same in both groups.

Table 4. Significances of differences in elastase released, total elastase–A1AT complex (E–A1AT), total A1AT and lactoferrin. The degranulation was analysed in four groups: CO – Control subjects without bacterial stimulation; CB – Control subjects after bacterial stimulation; PO – Patients without bacterial stimulation and PB – Patients after bacterial stimulation.
CO – CB*     PO – PB*     CO – PO**     CB – PB**     pairs
Elastase released     0.002     0.001     0.024     0.010     15
Total elastase     NS     0.013     NS     0.044     15
Total A1AT     0.041     0.002     NS     NS     13
E–A1AT complex     0.007     0.002     NS     NS     13
Lactoferrin     0.012     0.008     NS     NS     10

* Significance of difference calculated with the Wilcoxon signed–rank test. ** Significance of difference calculated with the Mann–Whitney U – test. NS not significant.

Fig.3. Elastase activity, expressed as cell equivalent x 10(3), released from peripheral neutrophyls during Fc R–mediated simulation for 60 min in 15 healhy controls and 15 patients with adult periodontitis. Horizontal bars indicate mean.

Fig.4. Mean values of elastase activity released extracellularly and elastase extracted from the pellet (expressed as cell equivalents x 10(3). Samples from 15 control subjects incubated in Hank’s buffer with opsonized bacteria (CB) or in Hank’s buffer alone (C0) and from 15 patients with (PB) or without simulation by bacteria (P0).

In Paper IV, we measured the concentration of IL–1 and the elastase activity in GCF samples from untreated subjects with periodontitis and gingivitis. The plaque index (PI), gingival index (GI) and the volume of GCF collectec were similar in the periodontitis sites from patients and in the gingivitis sites fromhealthy controls, but the values were lower in inflamed shallow pockets from periodontitis patients (Table 5). No significant correlations, were found between the clinical data at various sites and the IL–1 concentrations, with the exception of GCF volume, were a significant positive correlation was found (p<0.001) (data not shown).

Both the concentration and the total amount of IL–1 were significantly higher in shallow (GP) and deep pockets (PP) in patients with periodontitis than in gingivitis sites(GG) in healthy sobjects. No signgficant difference was noted between GP and PP. PP also showed higher elastase activity, althogh the difference was not signgficant (Tables 6a and b). A weak positive correlation was present between elastase activity and IL–1 (p= 0.06). This correlation was most pronounced in the GG samples (p= 0.03).

The amount of elastase–A1AT complex was significantly higher in PP tahn in GG and GP(Table 6b). The concentrations of E–A1AT, A1AT and A2MG were similar in the three types of sites (Table 6a).

Table 5. Comparison of mean values at various sites in 13 subjects with gingivitis and 18 patients with periodontitis. GG– gingivitis sites in gingivitis in patients.GP– gingivitis sites in periodontitis patients,PP– periodontitis sites in periodontitis patients. PI– plaque index; GI– gingival index; PPD– probing pocket depth expressed in mm and GCF– gingival crevicular fluid expressed in l.
Site     GG     p1     GP     p2     PP     p3
PI     1.7     NS     1.2     0.001     1.7     NS
GI     1.7     0.05     1.4     0.003     1.9     0.021
PPD     1.9     NS     2.1     0.001     6.1     0.001
GCF     1.8     0.013     1.3     0.006     1.9     NS

p 1 : Differences between GG and GP, calculated with the Mann–Whitney U –test. p 2: Differences betwenn GP and PP, calculated with the Wilcoxon signed–rank test. p3: Significance of differnces between GG and PP, calculated with the Mann–Whitney U–test. NS: nit significant.

Tables 6a and b. Mean concentrations (6a) and total amounts (6b) of elastase activity (E activity), elastase––1–antitrypsin complex(EA1AT), –1–antitrypsin(A1AT), –2–macroglobulin (A2MG) and interleukin–1 (IL–1) in GG (gingivitis sites in gingivitis patients), GP (gingivitis sites in periodontitis patients) and PP(periodontitis sites in periodontitis patients), n= 13 control subjects with gingivitis and 18 periodontitis patients.

Table 6a
Site     GG     p1     GP     p2     PP     p3
E activity ( cell equiv/l )     26600     NS     22400     NS     30000     NS
EA1AT (nmol/l)     0.26     NS     0.23     NS     0.44     NS
A1AT (nmol/l)     0.24     NS     0.17     NS     0.24     NS
A2MG (ng/l)     458     NS     738     NS     621     NS
IL–1 (pg/l)     2.0     0.009     5.0     NS     5.4     0.0036

Table 6b
Site     GG     p1     GP     p2     PP     p3
E activity ( cell equiv)     46500     0.04     27700     0.03     68500     NS
EA1AT (nmol)     0.42     NS     0.29     0.001     0.75     0.04
A1AT (nmol)     0.38     0.04     0.20     NS     0.40     NS
A2MG (ng)     656     NS     823     0.002     1164     NS
IL–1 (pg)     2.9     0.02     5.5     0.002     9.0     0.001

p1: Differences between GG and GP, calculated with the Mann–Whitney U –test. p2: Differences between Gp and PP, calculated with the Wilcoxon signied–rank test. p3: Significance of differences between GG and PP, calculated with the Mann–Whitney U –test. NS: not significant.

In paper V, we studied laminin and IL–8 in GCF samples from untreated subjects with periodontitis and gingivitis. The clinical findings are shown in table 7. The amounts of laminin were significantly higher in PP, as compared to GG, but no significant differences was noted between shallow and deep pockets in the patients group (table 8). The distribution of the amounts of laminin in the samples is shown in Figure 5. The concentrations of laminin tended to be higher in the GP samples (table 8). The amounts of IL–8 and lactoferrin were very similar in the three groups (table 8). Strong negative correlations were found when the GCF volume was correlated with the concentrations of laminin (Fig. 6) lactoferrin (Fig. 7), while no correlation with the IL–8 concentration was observed.

A strong positive correlation was observed between the concentrations of laminin and lactoferrin (R= 0.72,p= 0.001). This correlation was more pronouced in patients with preiodontitis (R= 0.81, p=0.003) than in control subjects (R= 0.69, p= 0.02).

Table 7. Mean values for PI– plaque indexm GI–gingival index, PPD– probing pocket depth expressed in mm and GCF– gingival crevicular fluid expressed in ml, in 12 subjects with gingivitis and 13 with periodontitis. GG– gingivitis sites in subjecs; GO– gingivitis sites in periodontitis patients; PP– periodontitis sites in periodontitis patients.
Site     GG     p1     GP     p2     PP     p3
PI     1.4     NS     1.6     NS     1.6     NS
GI     1.3     NS     1.7     NS     1.9     0.01
PPD     2.4     NS     2.6     0.0001     5.5     0.0001
GCF     1.8     NS     1.6     0.02     2.4     0.03

p1: Differences between GG and Gp, calculated with the Mann–Whitney U–test. p2: Differences between GP and PP, calculated with the Wilcoxon signed–rank test. p3: Significance of differences between GG and OO, calculated with the Mann–Whitney U –test. NS: not significant.

Table 8. Mean values of the total amounts and concentrations of laminin, interleukin–8(IL–8), expressed in picogramas(pg) and picogramas per microlitre(pg/ml), and lactoferrin expressed in arbritary units(a.u) and a.u/ml, in GG(gingivitis sites in gingivitis subjects) GP(gingivitis sites in periodontitid patients) and PP(periodontitis sites in periodontitis patients). N= 12 gingivitis and 13 periodontitis patients.
Site     GG     GP     PP
Laminin (pg)     287     397     385*
IL–8 (pg)     7.4     7.1     9.2
Lactoferrin (a.u.)     526.6     423.6     546.8
Laminin concentration (pg/l)     182     397     179
IL–8 concentration (pg/l)     4.4     8.9     3.8
Lactoferrin concentration (a.u./l)     352.8     378     237.1

* Difference between GG and PP, calculated with the Mann–Whitney U –test.

Fig 5. Boxplot showing the total amounts of laminin in 12 GG (gingivitis sites in control subjects), 13 GP (gingivitis sites in patients with periodontitis) and 13 PP (periodontitis sites in patients with periodontitis). Median, 10th , 25th , 75th and 90th percentiles in boxplot.

Fig 6. Correlation between laminin concentration and GCF volume in all samples (n= 38).

Fig 7. Correlation between lactoferrin and GCF volume in all samples (n= 38)

VI – DISCUSSION

The studies in this thesis show that neutrophil hyperreactivity is probaby one of the mechanisms causing tissue destruction in patients with adult periodontitis. This view is supported by the following findings: 1) sites with tissue destruction showed a higher free elastase activity than sites without; 2) neutrophils from periodontitis patients reacted more strongly against Fcy–receptor–mediated stimulation in vitro; and 3) basement membrane degradation tenteded to be greater in these patients. Moreover, such patients also have increased concentrations of interleukin–1b in gingival crevicular fluid(GCF), regardless of the degree of tissue destruction.

For he past few years, several authors have studied protease activity in GCF from patients with periodontitis(Ohlsson et al. 1973; Kowashi et al. 1979). Here we assessed protease activity by evaluating as the degradation of FITC–conjugated casein. Casein is degraded by most protease but, in contrast to low molecular weight substrates, not by protease complexed to A2MG. We showed that free protease activity can be detected in GCF in inflamed sites with or without tissue destruction, and that it is higher in sites with tissue destruction. The addition of A1AT to the GCF samples inhibited almost all activity(approx. 90%), suggesting that the protease activity was due to an imbalance between protease and antiprotease rather than to activity of proteases not sensitive to inhibition by A1At.

Elastase activity in GCF was measured with a low molecular weight substrate, highly specific for granulocyte elastase(Kramps et al. 1983). Such a substrate is hydrolyzed by free elastase and by elastase bound to A2MG (Trevis & Salvesen 1983). This means that measurements of elastase activity with this substrate cannot distinguish between free elastase and the elastase–A2MG complex. To distinguish between activity derived from free elastase and elastase bound to A2MG, an excess amount of A1AT was to the samples. A1AT is a very effective inhibitor of free elastase activity, but it cannot inhibit activity from the elastase – A2MG complex(Teavis & Salvesen 1983). The activity inhibited by A1AT can thus be seen as derived from free elastase and the remaining activity from the elastase–A2MG complex.

The total amount of neutrophil elastase was significantly higher in sites with tissue destruction and a large proportion of this elastase was free – i.e., still active against important proteins such as collagen and elastin . The free neutrophil elastase activity was stronbly correlated with the free protase activity, and showed the highest R–value in sites with tissue destructio (R= 0.71). Incontrast to our results, Giannopoulou et al.(1992) found no free elastase in GCF in patients with periodontitis. The reason for this discrepancy is not clear, but one explanation might be the sampling technique. Giannopoulou collected samples with microcapillaries, a method that may disturb gingival vessels ans cause a leakage of plasma. A large influx of plasmacould innhibit all elastase activity in GCF.

Higher elastase activity in sites with tissue destruction may be due to both a higher relase of elastase per all and/or an increased number of neutrophils. Gustafsson et al.(1994) showed that a higher elastase activity in GCF from periodontitis patients was due to increased release from each cell rather than to differences in the number of local cells, which mght be due to local hyperactivity of neutrophils. We studied the association between peripheral neutrophil reactivity and destructive periodontal disease by means of Fcy–receptor–stimulated elastase release. Significantly more elastase was released by in vitro–activated peripheral neutrophils from adult patients having periodontitis than by those from healthy controls. This was also true of juvenile periodontitis, using tha same type of stimulation(Asman 1988), but it has not been shown before in adult periodontitis. On the other hand , a similar study showed no differences in the release of elastase between subjects with various degrees of periodontal disease when peripheral neutrophils were activated with formyl–methionyl–leucyl–phenylalanime(fMLP) (Giannopoulou et al.1994). However, the increased degranulation of neutrophils in such patients may depend on the kind of stimuli used. Corresponding studies of oxygen radical prodiction, using Fcy–receptor–mediated stimulation, have consistently shown a difference between periodontitis patients and healthy controls(Asman 1988; Whyte et al. 1989; Fredriksson et al.1998), while studies using other methods of activation have shown conflicting results(Henry et al. 1984; Mouynet et al. 1994).

The higher release of elastase by peripheral neutrophils from patients with adult periodontitis might be due to a priming/activation of neutrophils already in the circulation, which could influence the intracellular activity of elastase. Elastase is present in both an active(Gardim et al. 1991) and a proenzyme form inside the azurophilic granules(Cavarra et al. 1997), but the mechanism controlling the transition from one to another is not well understood. According to Tamura et al. (1998), primed neutrophils activated with fMLP release almost twice as much elastase as fMLP activation does alone. Priming is a process that enhances the cellÕs ability to respond to a second stimulus. An increased priming response may reflect an increased susceptibility to priminig or increased amounts substances with a priming effect. Such substances may be derived from the host, e.g., proinflammatory cytokines and degradation products(Wachtfogel et al. 1988; Steinbeck & Roth 1989), or from bacteria, e.g., formyl–methionyl–leucyl–phenylalanime(Allred & hill 1978) and LPS(Forehand at al. 1991) . We recently found that unstimulated neutrophils from periodontitis patients expressed higher intracellular elastase activity than in those from healthy controls. Unlike elastase activity, the total amounts of antigenic elastase were similar in patients and controls(Figueredo et al. Unpublished data). This might indicated greater intracellular activation of latent elastasein perioheral neutrophils from patients with adult periodontitis.

Although elastase may have a salutary effect on the normal turnover of tissues and on combating infection, excessive release, especially under conditions that compromise the function of regulatory innhibitors – for instance, a–1–antitrypsin (A1AT) and a–2–macroglobulin(A2MG) – can damage tissue(for a review, see Janoff 1985). A1AT inactivates virtually all mammalian serine proteases. However, the rate of inhibition shows that elastase activity is inactivated by A1AT at a rate tenfold higher tahn of other proteases, indicating that the primary function of A1AT is to contorl neutrophil elastase activity(Travis & Savesen 1983). Most A1AT is produced in the liver, but cells like macrophages and neutrophils can contribute to local production of this inhibitor. In neutrophils, A1AT is stored in the same granula as elastase, but it does not seem to bind elastase intracellularly. Elastase tends to be present in the peripheral of the granule(Gramer et al. 1989), while A1At is usually found in the granule matrix(Mason et al. 1991). One reason why elastase activity was higher in the supernatant from the patientsÕ cells could be release of A1AT. However, we found no differences in the release of A1AT that explained the increase in elastase activity. Another possibility could be differences in inhibitory capacity of the A1AT released, but this possibility was ruled out by determining the levels of the elastase–A1AT complex in the supernatantm which were also similar in patients and controls.

We found a strong negative correletion between the percentage of free elastase and the percentage of elastase–A1AT complex in GCF. Since the concentrations of A1AT are high in plasma and in GCF, local inactivation rather than a shortage of A1AT probably occurred, in combination with increased release of elastase. Systemic inactivation of A1AT has also been suggested in a recentstudy by our group, in whuch plasma levels of A1AT were analysed in patients with adult periodontitis and compared to those in healthy controls. We found higherconcentrations of antigenic A1AT in patients with periodontitis, but a similar A1AT protease inhibitor capacity. Moreover, when this capacity was related to the antigenic A1AT, the levels of functional A1AT had a tendency to be lower in patients with periodontitis(Fredriksson et al. Unpublished data).

A1AT is sensitive to oxidative and proteolytic inactivation. Thus, an excessive realease of active protease in combination with simultaneous release of oxigen radicals might cause an inactivation of A1AT and subsequent tissue destruction(Weiss 1989). A similar process has been described in patients with lung emphysema, where high levels of oxidants, mainly hydrogen peroxide, can produce a situation in which the amounts of A1AT are normal, but the functional periodontal are near zero(Travis 1988). The situation may be even worse in periodontal disease, due to the activity of microbial protease that can degrade and inactivate A1AT(Carlsson et al.1984; Travis et al. 1994).

Apart from A1AT; –2–macroglobulin(A2MG) can inhibit elastase and almost all other proteases in periodontitis sites. A2MG is a rapid and efficient clearing for free proteases in the circualtion (Travis & Savesen 1983). We found that a higher percentage of elastase was inhibited by A2MG in samples from sites with tissue destruction, than in those without, an observation that accords with studies of GCF samples collected with paper strips (Gustafsson et al. 1994; Meyer et al. 1997). This may because impaired inhibition by A1AT is patly compensated by increased inhibition by A2MG. However, the simultaneous realease of oxigen species and proteases can also inactivate A2MG (Abbink et al.1991), which could explain the presence of free elastase in our samples. Moreover, the large molecular weight (725 kDa) of A2MG makes it difficult to compensate an extravascular deficiency/inactivation of A1AT.

Having assessed local peripheral differnces in elastase activity, we also wished to find a pro–inflammatory substance that might be characteristic of a given patient. IL–1 has been reported to be much higher in periodontitis sites than in healthy ones (Stashenko et al. 1991; Tsai et al. 1995; Ishihara et al.1997). This could be the result of severer inflammation and/or constitutional differences in Il–1b production. Kornman et al. (1997) and Gore et al. (1998) showed that patients with severe periodontitis have a significantly higher frequency of an IL–1 genotype associed with increased production of IL–1 in vitro (Pociot et al. 1992). We compared the levels of interleukin–1b in GCF from inflamed shallow and deep pockets in sujects without attachment loss. We found that the concentration of IL –1b in GCF is higher in periodontitis patients, even when inflamed shallow pockets in patients are compared to those in subjects with gingivitis alone. There were no significant differences in IL–1b concentration between shallow and deep pockets in the same patient. Since the GCF volumes and concentrations of elastase–A1AT, A1AT and A2MG were similar, the local degree of inflammation could be assumed to be the same in the three categories of sites. This supports the hypothesis that IL–1 is more a characteristic of a given patients and less a result of tissue destrution in the sampled site. Wilton et al. (1992; 1993) studied IL–1 levels in GCF from patients with adult periodontitis and found no correlation with plaque index, bleeding or pocket depth. Reinhardt et al. (1993) likewise found no differnce in IL–1 levels between shallow and deep pockets in patients with periodontitis. Salvi et al. (1997c) reported significantly higher IL–1 levels in GCF from patients with moderated/severe periodontitis than in those with gingivitis/mild periodontitis. However, no comparisom was made between shallow and deep pockets within the same patient.

Our findings differ to some extent from those of others, who have reported increasing levels of IL–1 with increasing inflammation and pocket depht (Ishihara et al. 1997; Hou et al. 1995; Yavuzylmaz et al. 1995). This may be because our subjects had not been treated and had similar degrees of inflammation in shallow and deep pockets. It seems likely that even if the level of IL–1 is characteristic of a given patient, it is also influenced by the degrees of inflammation since IL–1 is not constitutively produced, but needs external stimuli.

In our last investigation (Paper V), we found higher amounts of laminin in GCF from patients with adult periodontitis , suggesting the presence of hyperactivated leukocytes during the process of transmigration through the basement membranes. This difference was clearest when comparing PP and GG sites (p=0.03), but there was also a tendency towards a difference between GP and GG. However, due to the wide range of values, it did not reach significance (p=0.16). This finding indicates that these patients have more activated leukocytes that cause greater destruction of the basal membrane during their migration through the endothelium /epithelium. The high prevalence of neutrophils in the inflammatory lesions the periodontitis (Attström & Egelberg 1970). Together with previous evidence of local and peripheral neutrophil hyperreactivity (e.g., Gustafsson et al. 1994; Fredriksson et al.1998), suggest that these cells act the primary mediator of local basement membrane degradation in patients with adult periodontitis . The similar amounts of lactofferrin indicated that the differences in basal membrane degradation are not due to differences in the number of neutrophils in the lesion.

Activated neutrophils are said to express membrane–bound neutrophil elastase om the cell membrane (Owen et al. 1995). It is an important bioactive form of enzyme, where elastase is catalytically active against extracellular matrix macromolecules but resistant to inhibition by naturally occurring protease inhibitors (Campbell & Campbell 1988). Owen et al. (1995) showed that N–formyl–methiony–leucyl–phenylalanine (fMLP), complement component 5a (C5a) and phorbol myristate acetate (PMA) stimulate a dose–dependent increase in membrane–bound elastase expression. They also showed that priming neutrophils with bacterial LPS and then incubating them with fMLP induces a threefold increase in the cell surface expression of elastase, as compared to cells incubated with fMLP alone. We speculated that membrane–bound elastase activity might be the machanism by which activated neutrophils cause basement membrane degradation.

Even if neutrophil hyperreactivy occurs in periodontitis patients, sites–specific differences in tissue destruction still remain to be explained. We hypothesized that local differences in IL–8 could be one of the factors involved in these differences, but we found no association between hogher Il–8 levels and pocket depht. Chung et al.(1997) reported a lower concentration of IL–8 in patients with periodontal disease prior to tratament, but similar total amounts, which accords with our findings. Our results suggest that the amounts of IL–8 in GCF are influenced by the severity gengival inflammation rather than by constitutional differences in the local production that could explain site–specific tissue destruction.

VII – CONCLUSION

The main conclusion of this thesis are: (1) Increased levels of free protease activity are associated with tissue destruction in periodontitis patients, and part of this activity is derived from neutrophil elastase. Free elastase activity is associated with a local impairement in the majo inhibitor of elastase, A1AT. (2) Peripheral neutrophil hyperreactivity in patients having adult periodontitis is also associated with increased release of elastase after Fcy–receptor stimulation. (3) Increase concentration of interleukin–1 is a characteristic trait of a given patiens. (4) The tendency towards higher basal membrane degradation in patients having periodontitis seems to confirm the presence of local hyperactivityof neutrophils. (5) No evidence was found of a correaltion between IL–8 and site–specific differences in tissue destruction.

VII – ACKNOWLEDGEMENTS

IX – REFERENCES