Periodontitis decreases the antiatherogenic potency of high density lipoprotein

Periodontitis, a consequence of persistent bacterial infection and chronic inflammation, has been suggested to predict coronary heart disease (CHD). The aim of this study was to investigate the impact of periodontitis on HDL structure and antiatherogenic function in cholesterol efflux in vitro. HDL was isolated from 30 patients (age 43.6 6.1 years, mean SD) with periodontitis before and after (3.2 1.4 months) periodontal treatment. The capacity of HDL for cholesterol efflux from macrophages (RAW 264.7), HDL composition, and key proteins of HDL metabolism were determined. After periodontal treatment, phospholipid transfer protein (PLTP) activity was 6.2% ( P 0.05) lower, and serum HDL cholesterol concentration, PLTP mass, and cholesteryl ester transfer protein activity were 10.7% ( P 0.001), 7.1% ( P 0.078), and 19.4% ( P 0.001) higher, respectively. The mean HDL 2 /HDL 3 ratio increased from 2.16 0.87 to 3.56 0.48 ( P 0.05). HDL total phospholipid mass and sphingomyelin-phosphatidylcholine ratio were 7.4% ( P 0.05) and 36.8% ( P 0.001) higher, respectively. The HDL-mediated cholesterol efflux tended to be higher after periodontal treatment; interestingly, this increase was significant ( P 0.05) among patients whose C-reactive protein decreased (53.7% reduction, P 0.015) and who were positive by PCR for Actinobacillus actinomycetemcomitans . These results suggest that periodontitis causes similar, but milder, changes in HDL metabolism than those that occur during the acute-phase response and that periodontitis may diminish the antiatherogenic potency of HDL, thus increasing the risk for CHD. —Pussinen, P. J., M. Jauhiainen, T. Vilkuna-Rautiainen, J. Sundvall, M. Vesanen, K. Mattila, T. Palosuo, G. Alfthan, and S. Asikainen. Periodontitis decreases the antiatherogenic potency of high density lipoprotein. J. Lipid Res. 2004. 45: 139–147. Supplementary key words HDL metabolism • atherosclerosis • reverse cholesterol transport • cholesterol efflux • inflammation One of the central cells mediating atheroma development in atherosclerosis is a circulating monocyte-derived activated macrophage. Macrophages in arterial intima accumulate excess cholesterol, which is esterified and stored in the cytoplasm. The cholesteryl ester-laden macrophages, or foam cells, that accumulate in the arterial wall are the hallmark of early fatty streak lesions in atherogenesis. Among the bacterial components that trigger macrophage activation, the most widely studied is lipopolysaccharide (LPS), a potent virulence factor of gram-negative bacteria. Increasing evidence suggests that the pathway consisting of LPS-macrophage-inflammatory mediators plays a critical role in infection-associated atherogenesis and thromboembolism by inducing the release of cytokines, by enhancing platelet aggregation and endothelial monocyte adhesion, and by promoting the formation of lipid-laden foam cells (1). LPS also interferes with macrophage cholesterol metabolism by downregulating both scavenger receptor B1 and ATP binding cassette transporter A1 (ABCA1) expression (2). The functions of these two transmembrane proteins have been associated with cholesterol efflux. HDL is considered an antiatherogenic lipoprotein because of its direct role in neutralizing LPS in circulation (3) and protecting LDL against oxidation (4) as well as its role in reverse cholesterol transport (5). During the first phase of reverse cholesterol transport, HDL functions as an acceptor for cholesterol from cell membranes in a process of cholesterol efflux. This process is facilitated pasAbbreviations: apoA-I, apolipoprotein A-I; CETP, cholesteryl ester transfer protein; CHD, coronary heart disease; CRP, C-reactive protein; HL, hepatic lipase; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, phosphatidylinositol; PLTP, phospholipid transfer protein; PS, phosphatidylserine; SAA, serum amyloid A; SM, sphingomyelin; TG, triglyceride. 1 To whom correspondence should be addressed. e-mail: pirkko.pussinen@helsinki.fi Manuscript received 11 June 2003 and in revised form 27 August 2003. Published, JLR Papers in Press, September 16, 2003. DOI 10.1194/jlr.M300250-JLR200 by gest, on O cber 9, 2017 w w w .j.org D ow nladed fom

One of the central cells mediating atheroma development in atherosclerosis is a circulating monocyte-derived activated macrophage. Macrophages in arterial intima accumulate excess cholesterol, which is esterified and stored in the cytoplasm. The cholesteryl ester-laden macrophages, or foam cells, that accumulate in the arterial wall are the hallmark of early fatty streak lesions in atherogenesis. Among the bacterial components that trigger macrophage activation, the most widely studied is lipopolysaccharide (LPS), a potent virulence factor of gram-negative bacteria. Increasing evidence suggests that the pathway consisting of LPS-macrophage-inflammatory mediators plays a critical role in infection-associated atherogenesis and thromboembolism by inducing the release of cytokines, by enhancing platelet aggregation and endothelial monocyte adhesion, and by promoting the formation of lipid-laden foam cells (1). LPS also interferes with macrophage cholesterol metabolism by downregulating both scavenger receptor B1 and ATP binding cassette transporter A1 (ABCA1) expression (2). The functions of these two transmembrane proteins have been associated with cholesterol efflux.
HDL is considered an antiatherogenic lipoprotein because of its direct role in neutralizing LPS in circulation (3) and protecting LDL against oxidation (4) as well as its role in reverse cholesterol transport (5). During the first phase of reverse cholesterol transport, HDL functions as an acceptor for cholesterol from cell membranes in a process of cholesterol efflux. This process is facilitated pas-sively by the diffusion of cholesterol to HDL and actively by the interaction of lipid-poor apolipoprotein A-I (apoA-I) or pre ␤ -HDL and ABCA1 (5,6). ABCA1-facilitated cholesterol removal is the rate-limiting step in the production of nascent HDL. HDL cholesterol is esterified in circulation and transported directly, or via lipoproteins of lower density, to the liver for excretion. HDL-promoted cholesterol efflux has been studied extensively using different HDL subpopulations as acceptors of cholesterol and has been shown to be indirectly associated with arterial wall thickness (7). Less is known about the action of inflammatory HDL in the cholesterol removal process.
During both acute and chronic infections, changes occur in lipoprotein distribution and subclass composition (8)(9)(10). Increased catabolism of HDL may be the cause of low HDL cholesterol concentrations during infection (11,12). The main protein of HDL, apoA-I, is displaced by serum amyloid A (SAA), whose synthesis is largely regulated by inflammation-associated cytokines (13). The displaced lipid-poor apoA-I displays pre ␤ mobility upon electrophoresis (14). On the other hand, HDL triglyceride (TG) concentration increases; the formed TG-rich HDL is a good substrate for hepatic lipase (HL). The function of HL again leads to the formation of lipid-poor apoA-I and its accelerated catabolism via the kidneys (15). In fact, infection and inflammation may cause such dramatic changes in HDL composition that its substrate properties for transforming proteins and receptors change (16,17). Concentrations and activities of major HDL remodeling factors, which affect reverse cholesterol transport, are also modified during infection. Such factors include HL, cholesteryl ester transfer protein (CETP), lecithin-cholesterol acyltransferase (LCAT), and phospholipid transfer protein (PLTP) (8,10,18,19).
Infection-induced proatherogenic changes in lipoprotein profile may be one of the mechanisms underlying the increased risk of atherosclerosis in patients with chronic infections. Several pathogens that cause chronic infections may induce alterations in lipoprotein metabolism. These pathogens include Chlamydia pneumoniae (9), Helicobacter pylori (20), and periodontal pathogens (21). Contrary to the single-bacterium infections caused by the two former pathogens, the bacterial flora associated with periodontitis constitute a complex overgrowth of normal oral microbial species. Species with increased periodontopathogenic potential include mainly gram-negative bacteria, particularly Actinobacillus actinomycetemcomitans and Porphyromonas gingivalis (22).
Although periodontitis is associated with mild hypercholesterolemia or hypertriglyceridemia (21,23,24), its effects on the composition and remodeling factors of HDL are unknown. Therefore, the aim of this study was to investigate in detail the impact of periodontitis on HDL metabolism. Specifically, we investigated the changes in the composition and subclass distribution of HDL, the activities of lipid transfer proteins and enzymes involved in reverse cholesterol transport, and the potential of HDL in the cholesterol efflux in vitro among 30 patients with periodontitis before and after periodontal treatment.

Study subjects and samples
The study population comprised 30 generally healthy patients [age 49.8 Ϯ 7.4 years (mean Ϯ SD), 16 males and 14 females] with periodontitis. The inclusion criteria of the patients and periodontal examination were described in our previous article (25). Each patient had clinical and radiographic periodontal attachment loss at more than six teeth. In each patient, periodontal probing depths, gingival bleeding on probing, and suppuration were recorded for all teeth at six sites per tooth. Periodontal treatment consisted of conventional mechanical therapy and metronidazole 500 mg twice a day for 7 days when indicated (n ϭ 7). Serum and plasma samples were obtained before periodontal treatment and after approximately 3 months (108 Ϯ 32 days). The patients were not advised to change their diet or smoking habits during the study. The study protocol was approved by the Ethics Committee of the Institute of Dentistry, University of Helsinki, and informed consent was obtained from all participants. The samples were stored at Ϫ 70 Њ C until use. The results of the PCR detection of A. actinomycetemcomitans and P. gingivalis from subgingival bacterial samples were available from our earlier study (26).

Determination of serum parameters
Serum total cholesterol, HDL cholesterol, and triglyceride concentrations were determined by fully enzymatic methods. Sensitive C-reactive protein (CRP) and SAA concentrations were determined by UC CRP ELISA (Eucardio Laboratory, San Diego, CA) and the Cytoscreen Immunoassay Kit (BioSource International, Camarillo, CA). PLTP mass and PLTP, CETP, and LCAT activities were assayed as reported (27)(28)(29)(30). Specific activity of PLTP was calculated as the ratio of PLTP phospholipid transfer activity and mass. Serum antibodies to A. actinomycetemcomitans and P. gingivalis were analyzed by multiserotype ELISA using four dilutions of the samples in duplicate (26).

HDL-facilitated cholesterol efflux from macrophages
Permanent murine macrophages (RAW 264.7) were cultivated in RPMI 1640 medium supplemented with 10% fetal calf serum, 100 U/ml penicillin, and 100 mg/ml streptomycin. Cells were cultured as monolayers in a humidified atmosphere of 95% air and 5% CO 2 at 37 Њ C and transferred or harvested at 70% confluence. The cells were revived exactly 2 weeks before the efflux experiments and passaged four times. On day 1 before the experiment, the cells were harvested using a rubber policeman, counted under a microscope, and divided at a density of 7 ϫ 10 4 /ml on 12-well plates. The experiment medium was Macrophage-SFM medium (GIBCO, Gaithersburg, MD) supplemented with 100 U/ml penicillin, 100 mg/ml streptomycin, and 10 ng/ ml Leucomax (1.67 million IU Novartis; stock 5 g/ml in sterile water). On day 2, the medium was changed and the cells were labeled with 50 nCi/ml [1 ␣ ,2 ␣ (n)-3 H]cholesterol in ethanol (49.0 Ci/mmol; Pharmacia Amersham, Uppsala, Sweden). The final concentration of ethanol in the medium was 0.1% (v/v). On day 3, the medium with the radioactive label was removed and the cells were allowed to equilibrate overnight in fresh medium. On day 4, the medium was removed, the cells were washed twice with PBS, and 1 ml of fresh medium containing 100 g (as total protein) of HDL isolated from the patients was added in duplicate wells. The spontaneous cholesterol diffusion (mean 15.6 Ϯ 1.9% from the control HDL, n ϭ 10) was monitored in wells in which no HDL was present. HDL stored at Ϫ 70 Њ C was used in duplicate wells to control the weekly interexperiment variation (coefficient of variation, 2.9%).
At time points of 1, 3, and 6 h, the medium was removed and centrifuged at 6,000 g for 2 min. Of the supernatant, 250 l was transferred to scintillation vials, stored at 4 Њ C overnight with the scintillation cocktail, and counted for radioactivity. In the preliminary experiments (n ϭ 6), the cells were lysed with 0.3 M NaOH after washing twice with PBS containing 2% BSA and twice with PBS without BSA. The mean recovery, 94.5%, of the total choles-terol-associated 3 H radioactivity was counted from the cell lysates and the culture supernatants.

Statistical analyses
The statistical significance of the differences in the continuous variables between the groups was tested with the Mann-Whitney U-test or the Wilcoxon signed ranks test. A two-tailed Pearson correlation was used for correlation analyses. The associations of kinetic parameters of cholesterol efflux as dependent variables with other parameters determined were examined with linear regression analyses. A P value of Ͻ 0.05 was considered statistically significant. All statistical analyses were performed with SPSS 9.0 for Windows.

RESULTS
The mean periodontal and serum parameters before and after periodontal treatment are summarized in Table  1 . The proportion of teeth with deepened periodontal pockets and the number of bleeding or suppurating periodontal pockets were significantly ( P Ͻ 0.001) lower after treatment. Additionally, there was a significant ( P Ͻ 0.001) decrease in mean serum IgG-class antibody levels to P. gingivalis but not to A. actinomycetemcomitans . In the whole study group, the mean CRP and SAA concentrations did not change from the pretreatment value. Serum HDL cholesterol and concomitantly serum total cholesterol concentrations increased 10.7% and 6.7%, respectively, after treatment ( P Ͻ 0.001). PLTP activity decreased (6.2%, P Ͻ 0.05), whereas PLTP concentration increased (7.1%, not significant), leading to a decrease in PLTP specific activity (9.3%, P Ͻ 0.05). After periodontal treatment, there was only a nonsignificant increase in LCAT activity but a significant increase in CETP activity (19.4%, P Ͻ 0.001).
Total HDL was isolated from plasma samples obtained before and after treatment and analyzed for size and sub- class distribution by native gradient gel electrophoresis (Fig. 1). The mean proportion of HDL 2 particles increased by 11.3% of the pretreatment values (not significant), whereas the proportion of HDL 3 particles decreased by 33.2% (P Ͻ 0.05) (Fig. 1, inset). The HDL 2 / HDL 3 ratio increased from 2.16 Ϯ 0.87 to 3.56 Ϯ 0.48 (P Ͻ 0.05). After treatment, the relative amount of HDL particles in subclasses 2b1 and 2a increased by 13.3% (not significant) and 30.6% (P Ͻ 0.05), whereas the amount of particles in subclasses 2b2, 3a, and 3b decreased by 17.4% (not significant), 33.0% (not significant), and 33.5% (P Ͻ 0.05), respectively. Subclasses 1 and 3c were not detected in the electrophoresis.
To further characterize HDLs before and after periodontal treatment, their mass composition and phospholipid subclass distributions were analyzed (Fig. 2). In the mean mass composition, there were small, nonsignificant decreases in cholesteryl ester and triglyceride concentrations and increases in free cholesterol and protein concentrations. However, there was a significant increase in HDL total phospholipid concentration after periodontal treatment (7.4%, P Ͻ 0.05). Regarding HDL phospholipid subclass distribution, a statistically nonsignificant decrease in HDL-PC and increases in HDL-PE and HDL-PI concentrations were observed. On the other hand, HDL-PS and HDL-SM concentrations increased significantly (P Ͻ 0.05) by 15.2% and 16.0%, respectively. These changes caused a highly significant increase in the SM/PC ratio after the treatment (36.8%, P Ͻ 0.001).
To evaluate the function of HDL in reverse cholesterol transport, we determined the cholesterol efflux capacity of the isolated HDL preparations that were used as acceptors of labeled cholesterol from cultivated macrophages. The mean kinetic parameters describing the efflux capacity are summarized in Table 2. K m , V max , initial velocity (V init ), and catalytic efficacy (V max /K m ) were all 1.3-3.4% higher after periodontal treatment, although none of these differences reached statistical significance. To further analyze the efflux capacity, the patients were divided into two groups according to the presence (n ϭ 10) or absence (n ϭ 20) of A. actinomycetemcomitans in subgingival bacterial samples by PCR detection and according to the presence (n ϭ 20) or absence (n ϭ 10) of P. gingivalis (Fig. 3). As determined by V max , the A. actinomycetemcomitans-positive patients had a 9.3% lower (P Ͻ 0.01) efflux Fig. 1. HDL particle size and subclass distribution. Total HDL was isolated from plasma samples obtained from 30 patients with periodontitis before and after treatment, and HDL particle size was analyzed by 4-30% native gradient PAGE. The mean distribution of HDL 2 and HDL 3 subclasses before (open circles) and after (closed circles) periodontal treatment is shown in detail. The mean distribution of HDL 2 and HDL 3 main subclasses before (open circles) and after (closed circles) periodontal treatment is shown in the inset. The asterisks indicate statistically significant differences (P Ͻ 0.05) before versus after treatment.  To clarify how the HDL cholesterol efflux parameters related to the other variables determined and to the pe-riodontal variables, correlation analysis was performed on the pretreatment variables ( Table 3). There was a positive correlation between the efflux parameters and serum HDL cholesterol concentration and CETP activity as well as with HDL free cholesterol, cholesteryl ester, and phospholipid concentrations. The parameters describing the efflux velocity (V max and V init ), HDL free cholesterol, and HDL phospholipid concentrations were negatively correlated with the proportion of teeth with periodontal pockets deeper than 5 mm. The proportion of teeth with deep periodontal pockets and the number of periodontal pockets with suppuration correlated negatively with serum HDL cholesterol concentration and PLTP mass. Periodontal pockets with suppuration were also positively correlated with SAA concentration in serum. To ascertain the results from the correlation analysis, two linear regression models with V init as a dependent variable were tested. In a model adjusted for age, V init was significantly associated with PCR positivity for A. actinomycetemcomitans (␤ ϭ 0.240, P ϭ 0.049), CRP (␤ ϭ Ϫ0.262, P ϭ 0.044), and HDL cholesterol concentration in serum (␤ ϭ 0.474, P Ͻ 0.001). Among HDL components, the strongest association with V init was found with free cholesterol (␤ ϭ 0.298, P ϭ 0.098) and SM/PC (␤ ϭ 0.558, P ϭ 0.104), but none of the components reached statistical significance.

DISCUSSION
Periodontitis is a persistent bacterial infection that causes chronic inflammation in periodontal tissues. The disease is characterized by the formation of deep periodontal pockets and the destruction of connective tissue attachment and alveolar bone. Untreated periodontitis may eventually lead to loss of teeth. The systemic response to periodontitis can be measured as increased serum antibody levels against periodontopathogenic bacteria (35) and CRP concentrations (36). Also in the present study, the levels of serum antibodies to P. gingivalis and the concentrations of CRP and SAA in A. actinomycetemcomitanspositive patients were decreased after periodontal treatment. In addition to data from early cross-sectional clinical studies reporting that dental infections are significantly more prevalent in coronary heart disease (CHD) patients than in their matched controls (37), a number of studies have confirmed and extended these findings (38,39). Because severe periodontitis is relatively common, affecting ‫%02ف‬ of middle-aged and elderly populations worldwide (40), its relevance to public health is extensive. An increased risk for CHD in individuals with periodontitis may be one of the long-term effects of the daily systemic spread of live subgingival/oral bacteria and their components, which have access to the bloodstream as a result of mechanical injury (e.g., during eating or tooth brushing) to the inflamed gingivae. However, it is still unclear what mechanisms are involved in periodontitis-promoted atherogenesis.
A strong negative association between untreated periodontitis and HDL cholesterol concentration was observed in the present study. This finding is of major importance, Fig. 3. Kinetic parameters of cholesterol efflux capacity of HDL, and C-reactive protein (CRP) concentrations in serum before and after periodontal treatment. Total HDL was isolated from plasma samples obtained from 30 patients with periodontitis before and after treatment. Cultivated mouse macrophages (RAW 264.7) were labeled with [ 3 H]cholesterol. After removing the label, equilibration, and washing the cells, 100 g (as total HDL protein) of HDL was added. Cholesterol efflux promoted by HDL was analyzed from the cell medium after 1, 3, and 6 h of incubation. The kinetic parameters K m (A), V max (B), and initial velocity (vinit; C) were calculated before (white columns) and after (gray columns) treatment of patients positive (ϩ) or negative (Ϫ) for A. actinomycetemcomitans (Aa) and P. gingivalis (Pg). Sensitive CRP was determined from serum samples. One asterisk indicates statistical significance (P Ͻ 0.05) before versus after treatment, and two asterisks indicate statistical significance (P Ͻ 0.01) of the difference between Aa-negative and Aa-positive patients before treatment. Error bars indicate SD.  because low HDL cholesterol concentration is one of the established independent CHD risk factors confirmed in numerous epidemiological studies (41). In earlier clinical studies, patients with periodontitis had higher serum total cholesterol or triglyceride concentrations than did healthy controls (21,23,24), but the differences in HDL cholesterol concentrations have not been significant in any of the reports. On the other hand, in a random population sample (n ϭ 1,163) of middle-aged and elderly men, high combined serum antibody levels against A. actinomycetemcomitans and P. gingivalis were significantly associated with low HDL cholesterol concentration (42). In the present study, the higher HDL cholesterol concentration after periodontal treatment was attributed specifically to the increase of the HDL 2a subpopulation, which has been proposed to be more protective than total HDL or HDL 3 (43). Cell culture as well as genetic family and population studies have revealed that serum HDL cholesterol concentration does not necessarily reflect the efficacy and antiatherogenicity of reverse cholesterol transport (44). Instead, HDL subclass distribution, cholesterol efflux capacity, and the kinetics of HDL metabolism are putatively important parameters that regulate the role of HDL against atherosclerosis. As suggested here, defective cholesterol removal attributable to inflammation-induced changes in HDL may provide one of the mechanisms underlying the relationship between periodontitis and atherosclerosis as well as between other chronic infections and atherosclerosis (45). HDL composition, in addition to HDL cholesterol concentration and subclass distribution, is an important determinant of the cholesterol removal capacity (46). In the present study, the changes in HDL composition before versus after periodontal treatment were substantial, except for HDL phospholipid mass and subclass distribution. It has been shown that the efflux capacity of plasma has the strongest correlation with the HDL-SM concentration (47)(48)(49). Moreover, SM is the phospholipid most efficient in solubilizing free cholesterol. The significant increase in the HDL-SM/PC ratio after periodontal treatment thus supports our present observation of the increased cholesterol efflux from macrophages.
Also, the higher LCAT and CETP activities after periodontal treatment support the hypothesis of an enhanced reverse cholesterol transport during periodontitis, because these proteins are among the key factors in this process (5). PLTP provides a supply of primary cholesterol acceptors through HDL conversion (50)(51)(52) and is thought to increase the binding of HDL to human skin fibroblasts (53), thus promoting cholesterol efflux. Nevertheless, the exact role of PLTP, especially the balance between recently reported active and inactive forms of PLTP, in the efflux capacity of HDL is not yet resolved (54). In the present study, PLTP activity decreased but PLTP mass increased after periodontal treatment, in agreement with the earlier study of patients with an acute-phase response (10).
Although the statistical power of the present study was limited, the results indicate that there is an association be-tween HDL cholesterol efflux capacity and clinical and microbiological findings of periodontitis. First, we found a significant negative correlation between the proportion of teeth with deep periodontal pockets (indicating the extent of periodontitis in the dentition) and the kinetic parameters of efflux capacity (V max and V init ). Second, the differences in the efflux kinetics before and after periodontal treatment were significant only in a subgroup of patients who were PCR-positive for A. actinomycetemcomitans. This was not the case, however, for patients who were PCR-positive for P. gingivalis. This difference may be attributable to the dissimilar pathogenic characteristics of the two pathogens, for example, LPS structure or activity (55). If substantiated in a larger study population, this difference may also be related to the lower SAA concentration in A. actinomycetemcomitans-positive patients after periodontal treatment. Through its association with HDL, SAA results in better binding of SAA-enriched HDL to macrophages, increased cholesteryl ester uptake, impaired activity of LCAT, and reduced cellular cholesterol efflux (16,17). In addition, SAA effectively displaces apoA-I from HDL and thereby diminishes the formation of lipid-poor apoA-I, the major specific cholesterol acceptor from macrophages.
We found that periodontitis may decrease serum total HDL cholesterol concentration and HDL 2 /HDL 3 ratio. Periodontitis-induced changes in HDL composition may impair its efflux capacity; indeed, the efflux capacity of HDL was significantly lower among patients who were PCR-positive for A. actinomycetemcomitans. Our results indicate that the changes in HDL metabolism during periodontitis are similar to but milder than those found during an acute-phase response. The present data suggest that periodontitis diminishes the antiatherogenic potency of HDL and may thus increase the risk for CHD.
Tiina Karvonen is acknowledged for skillful and precise technical assistance. This study was sponsored by the Academy of Finland (grants 75953 and 77613 to P.J.P., grants 71413 and 72971 to S.A.), by the Paulo Foundation, and by the Finnish Dental Society.