Humoral IgG Antibodies to Oral Microbiota in a Population at Risk for Root-surface Caries R. KENT, D.J. SMITH1O3, K. JOSHIPURA, P. SOPARKAR2, and M.A. TAUBMAN1 Departments ofBiostatistics, 'Immunology, and 2Clinical Trials, Forsyth Dental Center, 140 The Fenway, Boston, Massachusetts 02115 Mutans streptococci have been strongly implicated in the initiation of dental caries on coronal surfaces. Their role in development of root-surface caries is less clear. The etiologic agents of both types of dental caries are likely to elicit systemic immune responses. The objective of the present study, therefore, was to study the association ofclinical variables ofdisease with humoral IgG antibodies to nine oral micro-organisms in 314 adult subjects, aged 45-65 years, who were at risk for root-surface caries. Antibody activity to Streptococcus mutans strain Ingbritt, S. mutans / S. sobrinus GTFs, S. faecalis strain 19433, Actinomyces viscosus strain WVU 626, Actinomyces naeslundii strain 12, Lactobacillus casei, Actinobacillus actinomycetemcomitans strain Y4, Porphyromonas gingivalis strain 381, Eikenella corrodens strain 1073, and Wolinella recta strain 371 was measured by ELISA. Pearson correlation coefficients among loglo antibody levels within subjects revealed marked positive correlations among subgingival bacteria, generally weak positive correlations among supragingival micro-organisms, and no correlations between elements of the supragingival battery with the subgingival battery. IgG antibody levels to mutans streptococcal antigens were significantly correlated with subject DMF scores (r = 0.23; p < 0.0001). No significant correlation was seen between DMF scores and antibody to any other supragingival micro-organism tested. Further relationships between levels of S. mutans antibody and individual clinical variables were analyzed by step-wise multiple linear regression, resulting in a model that was highly significant (p = 0.0001), with an r2 = 0.14. Numbers of missing teeth, coronal caries, root-surface caries, and root-surface restorations were each positively associated with antibody levels to mutans streptococci. The model indicated that the associations seen with S. mutans antibody reflected previous encounters with S. mutans antigens, since this model emphasized cumulative caries experience over measures reflecting a subject's current status. These observations strengthen the association between S. mutans and dental caries on coronal surfaces and suggest a further association with caries on root surfaces.
J Dent Res 71(7):1399-1407, July, 1992
Introduction. Many components of host defense are present in the oral cavity. Components that enter the oral cavity via the gingival crevicular fluid emanate both from the systemic circulation and from local gingival synthesis (Lehner, 1983). Glandular saliva contains both antibody that originates in interstitial plasma cells, and innate immune components (e.g., lysozyme, lactoferrin, sialoperoxidase, and mucins) that originate in acinar cells and ductal epithelium (reviewed by Brandtzaeg, 1984; Michalek and Childers, 1990). Although it is not yet clear how antibody-mediated responses to oral flora are specifically integrated into the Received for publication August 7, 1991 Accepted for publication January 30, 1992 3To whom correspondence and reprint requests should be addressed This investigation was supported by USPHS Research Grant DE-07009 from the National Institute of Dental Research, National Institutes of Health, Bethesda, MD 20892.
oral host-defense system, the antibody specificities that result from recognition of bacterial antigenic epitopes can provide a record of infection with micro-organisms that cause disease. Dental caries on coronal surfaces is an infectious disease that appears to be initiated on coronal surfaces by mutans streptococci (Loesche et al., 1975; Loesche and Straffon, 1979; Emilson and Krasse, 1985). To understand the role of the systemic immune response to the organisms causing disease, several investigators have attempted to correlate levels of serum antibody to plaque micro-organisms with indicators of oral microbial pathogenesis (e.g., DMF). Since these studies used children or young adults, nearly all disease occurred on coronal surfaces. Cross-sectional studies in these populations have generally shown that high levels of serum IgG antibody to mutans streptococcal cells or antigens were associated with low caries experience (Challacombe, 1980; Challacombe et al., 1984, 1986; Gregoryet al., 1986). In this regard, IgG, purified from the sera of caries-free young adults, significantly inhibited the activity of mutans streptococcal enzymes associated with plaque formation (Block et al., 1979). Exposed dental root surfaces may have a different susceptibility to caries than do coronal surfaces. Intrinsic differences in mineral composition and organic matrix may alter factors involved in the colonization potential of organisms within the oral milieu. This may lead to different cariogenic flora on coronal and root surfaces (Jordan and Hammond, 1972; Ellen et al., 1985; Nyvad and Kilian, 1990), although mutans streptococci have been implicated in dental caries on both surfaces (van Houte et al., 1990). The level and specificities of antibody to the microbiota colonizing root surfaces may provide evidence for the organisms responsible for root-surface caries. Therefore, this study evaluated potential associations between clinical variables and serum IgG antibodies to nine oral micro-organisms in a population at risk for root-surface caries.
Materials and methods. Subjects.-Three hundred and fourteen subjects between the ages of 45 and 65 years were included in this study. This ambulatory volunteer population was nearly equally divided between males (51.9%) and females (48.1%). Each subject received a comprehensive clinical examination at Forsyth Dental Center. Subjects had a minimum of ten teeth; 293 subjects had gingival recession of 1 mm or more on at least three teeth. The population statistics are given in Table 1. Subjects were later divided into three groups, based on their DMF scores, for description of any associations of antibody data with DMF levels. These three groups (low, middle, high) had the following characteristics (n) (mean ± standard deviation) (minimum-maximum): low third (98) (15.2 ± 2.8) (2-18); middle third (115) (20.4 ± 1.2) (19-22); upper third (101) (25.1 ± 2.0) (23-31). ELISA.-Blood was taken from the antecubital fossa. Serum was then obtained from the coagulated specimen after centrifugation and stored frozen at -20'C until use. Sera were tested for antibody activity by a previously described biotin-avidin, alkaline phosphatase enzyme-linked immunosorbent assay (ELISA) method (Smith et al., 1990). Polystyrene microtiter plates (Flow Laboratories, McLean, VA) were coated with approximately 108 bac-
Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
1399
1400
J Dent Res July 1992
KENT et al.
phosphatase (Zymed) and p-nitrophenylphosphate. Antibody activity was expressed as ELISA units (EU) after comparison of experimental optical densities (Artek Industries, Farmingdale, NY) that fell within the linear portion of a reference standard curve constructed from dilutions of pooled reference sera containing high levels of antibody activity to the respective antigen. Analyses of antibody to supragingival micro-organisms were performed on sera from all 314 subjects. Measurement of antibody to subgingival micro-organisms was performed on sera from 248 of the 314 subjects. The subgingival microbial antigen set was included in the analyses under the assumption that these micro-organisms would not be expected to be involved in the formation of dental caries. Antigen preparation. -All bacterial cell preparations were formalin-killed after growth in PPLO broth, supplemented with glucose for the supragingival micro-organisms, and supplemented with hemin and menadione for the subgingival micro-organisms. Bacterial cell antigens included Streptococcus mutans strain Ingbritt, S. faecalis strain 19433, Actinomyces viscosus strain WVU 626, Actinomyces naeslundii strain 12, Lactobacillus casei, Actinobacillus actinomycetemcomitans strainY4, Porphyromonas gingivalis strain 381,Eikenella corrodens strain 1073, and Wolinella recta strain 371. Glucosyltransferases, prepared from Streptococcus mutans (two strains) and S. sobrinus (one strain) by adsorption techniques previously described (Gahnberg et al., 1985), were pooled and used to coat plates. Clinical variables.-Each subject received an examination that included the following measurements: number and locations of missing teeth, coronal caries and restorations by tooth surface, root lesions and restorations by root surface, debris and calculus index (Greene and Vermillion, 1960) applied to every tooth, buccal and lingual recession (the maximum distance from the cemento-enamel junction to the gingival margin on the buccal and lingual surfaces of every tooth), pocket depth (taken as the distance in mm from the gingival margin to the epithelial attachment at the mid-line as well as the mesio- and disto-buccal, and mesio- and disto-lingual papilla of each tooth), and gingival status (Loe and Silness, 1963). Incipient, arrested, and secondary caries were combined with frank cavitation in the estimation of root-surface and coronal caries.
TABLE 1 POPULATION STATISTICS
n
314
MALES
51.9%
FEMALES
48.1%
Mean
SD
Range
56.53
5.47
25.00
Gingival index
0.76
0.49
1.90
Debris
0.92
0.32
1.91
Coronal caries
1.75
4.08
43.00
Coronal fillings
32.81
17.26
91.00
% Pockets > 3 mm
7.58
11.02
62.50
Missing teeth
7.67
4.53
22.00
12.34
4.99
26.00
% Teeth recessed > 2 mm 11.49
14.05
72.73
Age
Filled teeth Root-surface caries
0.82
1.57
12.00
Root fillings
1.39
2.39
19.00
20.30
4.45
31.00
DMF
terial cells, or with 0.5 jig of glucosyltransferase (GTF) purified from three different mutans streptococcal strains by techniques previously described (Gahnberg et al., 1985). Antibody activity was then measured by incubation with 1:100-1:400 dilutions of sera. Plates were then developedfor IgG antibody with biotinylated affinity-purified anti-gamma-chain reagents (Zymed, South San Francisco, CA), followed in sequence by streptavidin-alkaline
TABLE 2
CORRELATIONS AMONG SERUM IgG ANTIBODY LEVELS Bacterial Strain
Pearson Correlation Coefficient*
S. mutans
1
GTF
0.04
1
A. naeslundii A. viscosus
0.39
0.00
1
0.11
0.12
0.31
1
L. casei
0.15
0.18
0.12
0.26
1
S. faecalis
0.14
0.09
0.06
0.4
0.09
1
W. recta
0.06
0.07
-0.02
0.12
-0.01
-0.05
1
A. actino. Y4
0.01
0.07
-0.09
0.08
0.09
-0.02
0.41
1
P. gingivalis
0.07
0.01
-0.05
0.02
-0.01
-0.07
0.32
0.49
1
E. corrodens
-0.02
0.03
-0.03
-0.01
0.08
-0.02
0.29
0.58
0.39
Supragingival
Subgingival
1
GTF Smut Anaes Avis Lcas Wrec AaY4 Bging Ecor Sfaec *Single underline indicates a p < 0.05 level of significance. Double underline indicates a p < 0.0001 level of significance. Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
HUMORAL ANTIBODY TO ORAL MICROBIOTA
Vol. 71 No. 7
Statistical methods.-
Summary descriptive statis-
1401
100
A. actinomycetmcomlans
go tics for clinical variables observed on 314 subjects are given in Table 1. Frequency 40. distributions are given for serum IgG antibody levels to 20 the battery of ten antigens 0 ° (Fig. 1). For reduction of _ skewness in the distributions, E W. so ELISAunits wereloglo-trans*-A. neeslundli formed for all analyses. All 40_ but the A. naeslundii distri* _ bution were unimodal, al_ _ though nearly all the distri10 butions still exhibited some o degree of skewness to the 32 3.6 OA 0.6 12 1.6 2.0 24 right. Product-moment correlations were calculated for Log10 Antibody Level (EU) evaluation ofall pair-wise asloo sociations among the antibody levels (Table 2). In adGTF dition, correlations were calW rea culated between log1o antiso body levels for the six su4* pragingival members of the battery and "caries-related" j variables (including DMF, coronal caries, root-surface C_ 701 lesions, missing teeth, coro_ . nal restorations, root-surface E so E r restorations, and debris) and Z so between log10 antibody lev*0 els for the four subgingival 30_ members of the battery and "periodontitis-related" clini10 cal measures (including recession, pocket depth, and 0 o 0 O3 A 0 12 1.5 1A 2.1 2.4 2.7 3.0 OA 0.9 12 1.5 1.6 2.1 2. 2.7 3.0 3.3 3.6 missing teeth). Multiple linear regression Log10 Antibody Leve (EU) Log1* Antibody Level (EU) was used for examination of so multivariate associations between levels ofS. mutans IgG A. viecosus S. mutans GO antibody, as the response variable in the statistical model, 40 and a variety of clinical variables, as explanatory vari_ ables in the model. An initial model related log antibody level to age, gender, and DMF score. The contribuP. glngivalls S. fecalls tions to this association of W the three components of the composite DMF score were *evaluated further by forward step-wise regression (Kleinbaum et al., 1988), in a which, instead ofDMF, separate variables-representing 0.4 GA 12 IA 2. 24 U 32 3A 4.0 4. 0.4 OA CA 1.0 11 IA IA 2.0 2.2 2.5 number of missing teeth, % Logo AntIbody LUr (EU) Log1o Antibody Level (EU) teeth (present) with enamel distributions of IgG antibody to supra- and subgingival micro-organisms. The ELISA coating Fig. for1-Frequency each formalin-killed bacterial restorations, and % teeth antigen species is given in the upper right corner of each plot. Antibody is reported (present) with enamel car- as log1o EU. ies-were utilized as explanatory variables. Age, gender, debris, and measures of root-surface variable to the resulting regression models (Tables 3 and 4) are lesions and root-surface restorations were also candidate ex- summarized by the associated regression coefficients which deplanatory variables in these analyses. The contributions of each fine the regression model and by mean squares and probability Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
KENT et al.
1402 2.6 2.4 2.2 LJ 2.0 0 0' 1.8 0 1.6 0 .4 1.4 1.2 2.4 2.2 2.0 1.8 1.6 1.4 1.2
-
J Dent Res July 1992
:~ ~ ~ ~
2.6
a) S. mutans
2.4
d) GTF
2.2 Li
(Table 2). Micro-organismswere grouped according to their supra-
2.0
or subgingival location. Correlation coefficients above 0.25 in Table 2 were significant at the p < 0.0001 level. With the given sample sizes, correlations larger than ( r >) 0. 1 were statisti-
1.8 0
0
1.6 1.4
1.2
- b) A. viscosus
2.4
transformed levels of IgG antibodies reactive with the ten micro-organisms in the battery
e) S. foecalis
cally significant (p
3 mm) was lowest for subjects with generally lower antibody levels to subgingival species and highest among subjects with generally higher antibody levels to-subgingival species (by the rank-sums). Means for gingival recession and, to some extent, missing teeth were lower among subjects having higher ranksums. Regression of the antibody rank-sum on age, pocket depth (% sites > 3 mm) and recession (% sites > 2 mm) was statistically significant (p = 0.003), but explained a minimal amount of variation (r2 = 0.06). Coefficients were positive for pocket depth (p = 0.04), and negative for recession (p = 0.02).
Discussion. Investigations of the etiology of dental caries consistently implicate mutans streptococci in the initiation of this disease on coronal surfaces (Loesche et al., 1975; Loesche and Straffon, 1979; Emilson and Krasse, 1985). Although the association of mutans streptococci with root-surface caries is less compelling, several observations suggest a similar etiology. Mutans streptococci have been reported to be more prevalent in the saliva (Ravald et al., 1986; Emilson et al., 1988; van Houte et al., 1990), on sound exposed root surfaces, and in incipient lesions on exposed root surfaces (Brown et al., 1986; van Houte et al., 1990) of root-cariesactive compared with root-caries-free subjects. The population described in the present investigation has had significant coronal caries experience and is also in an age range in which the degree
Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
HUMORAL ANTIBODY TO ORAL MICROBIOTA
Vol. 71 No. 7
1405
3.0 of gingival recession puts many 12 2.7 at risk for root-surface caries. d) GINGIVITIS TH 2.4 IgG antibody to S. mutans was POCKET a) DEPI l0 x Li 2.1 the only serological response 0 A; z 1.8 ofthe six supragingival plaque A -j 1.5 antigens examined that showed V) 1.2 significant associations with Li 0 z 0.9 clinical variables relating to U) 4 0.6 dental caries. For example, 0.3 only serum IgG antibody to S. 0 mutans showed a significant, positive correlation with sube) GENDER b) RECESSION 0.8 ject DMF scores (Fig. 2; Table Li. z 3). Furthermore, multiple lin- CNJ 0 0.6 ear-regression analysis of S. A : 0 mutans antibody on components a)L'i CL 0.4 I 0 ofthe DMF score also resulted V) CL in a model with very high sta0.2 tistical significance (Table 4). A degree of correspondence ( existed between serum IgG anarf) AGE 14 - c) MISSING TEETH tibody levels to oral microbiota 70K 12 and their oral locations (Table 60 H 2). For example, serum-anti- z 10 XXX 50U)J body levels to six supragingi- I 8 )OOC )O00C 4 >00( 20tial number of significant, posiXXX 2 >00( tive correlations among mem10 H 0 - . - I bers within each battery. 0 L H M L M Among the six supragingival L H organisms, serum-antibody ANTIBODY RANK SUM levels exhibited a pattern of low, positive correlations, which included several highly signifiFig. 4-Association of subgingival microbiota rank-sums with clinical variables. The EU values for antibody to cant correlations. Clearly, the P. gingivalis, A. actinomycetemcomitans, E. corrodens, and W. recta were rank-ordered, and a sum ofranks for the strong correlation between an- four antibody specificities was calculated for each subject. Means (± SE) are given for clinical variables for subjects tibody to A. viscosus and A. grouped, on the basis of this periodontal antibody rank-sum, into low (L), middle (M), and high (H) antibody levels. naeslundii cells is likely to be at least partly related to cross-reacting antigens (Cisar et al., suggest a measure of specificity, at least in the local immune 1987). It is unclear, however, why the correlation between S. response (Ebersole et al., 1985; Smith et al., 1985), since a mutans antibody andA. viscosus antibody should not be of at least collective elevation has not been found in gingival crevicular fluid the same order of magnitude as that between S. mutans and A. antibody levels to these organisms. It should be noted that these naeslundii, since both micro-organisms cohabit both non-carious associations were quite modest, possibly reflecting the fact that sites and caries lesions (Fure et al., 1987; Emilson et al., 1988; this population was not selected for those with periodontal disease. van Houte et al., 1990). No significant correlations were found DMF was positively associated with serum-antibody level for between antibody levels to supragingival organisms and antibodS. mutans (r = 0.23, p = 0.0001; Fig. 2; Table 3) but showed no ies to subgingival organisms. The most remarkable array of correlations existed among association with antibody levels to other supragingival organisms. When the associations among clinical variables and S. serum antibodies to A. actinomycetemcomitans, P. gingivalis, E. corrodens, and W. recta (r = 0.29 0.49). These four micro- mutans antibody were explored further, regression analysis of S. organisms are frequently cultured from subgingival locations mutans antibody level (log10) on age, gender, and DMF was highly (Moore et al., 1982; Tanner et al., 1987), and at least two of them significant (p = 0.0001), but explained only a small proportion of (A. actinomycetemcomitans and P. gingivalis) have been associ- the variation (r2 = 0.08). Step-wise regression analysis of S. ated with various forms of periodontal diseases. These correla- mutans antibody level (loglo) on a set of clinical parameters tions are unlikely to result from responses to common antigens, resulted in a model (r2 = 0.14) that included measures of cumusince studies with Actinobacillus (Zambon et al., 1983) and lative caries experience (missing teeth, enamel and root-surface Porphyromonas (Ebersole et al., 1986) strains and human or restorations) and a present/absent index variable for enamel animal sera indicated little cross-reactivity between these gen- lesions. The model supported the conclusion that the associations seen with S. mutans antibody reflected previous encounters with era. When IgG antibodies to these four micro-organisms were rank-ordered and summed for each subject, an apparent positive S. mutans antigens, since this model emphasized cumulative correlation existed between increased antibody level and pocket caries experience over measures reflecting a subject's current depth (Fig. 4a). It is conceivable that an immune (perhaps status. The associations seen with S. mutans were not seen with GTF, polyclonal) responses) occurred to several microbial species that had presented infectious challenges during the extension of the an antigen derived from mutans streptococci. This lack of corpocket. However, IgG antibody levels in gingival crevicular fluids relation may have occurred because the immunodominance of S. and in gingival tissues of subjects with periodontal disease would mutans components-such as antigen I/II, GTF, lipoteichoic acid, 8k
,
14
-
-
T-
I-
-
Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
1406
KENT et al.
J Dent Res July 1992
or serotype carbohydrate-is likely to vary among subjects. Regardless of which S. mutans immunodominant antigen is recognized by a subject, all subjects would still react with the intact S. mutans antigen, since it contains all components. However, those subjects for whom an antigen other than GTF was immunodominant would be positively associated with the intact S. mutans, but not associated with GTF. Alternatively, the specificity seen with the intact S. mutans species may have been blurred by the use of pooled GTF from S. mutans and S. sobrinus species. Many previous studies have examined the association between serum IgG antibodies to supragingival plaque micro-organisms (reviewed in Michalek and Childers, 1990). Antigens in most studies were limited to oral streptococci, and studies were performed with sera from 18-25-year-old subjects. Both positive (Huis in't Veld et al., 1979) and negative associations (Block et al., 1979; Challacombe, 1980; Levine et al., 1984; Gregory et al., 1986) were found between serum IgG antibody to mutans streptococci and caries experience. The present study, where a positive association was observed, differed importantly from essentially all of the foregoing studies in that the cohort was much older (mean age, 56.7 years). This resulted in a population whose mean levels of caries experience (DMF) were generally equivalentto those ofthe so-called "high caries" groups ofthe earlier studies. Furthermore, the cohort in the present study was decades removed from the "caries-prone" years of adolescence. Thus, if mutans streptococci frequently initiate dental caries, and if mutans streptococcal antigens elicit serum IgG antibody during this process (Challacombe, 1980; Challacombe et al., 1984), which takes place relatively early in life, and if a memory lymphocyte cell population is subsequently maintained as a result of the underlying mutans streptococcal infection, then itis not surprisingthat serum IgG antibody levels to S. mutans were most closely associated with cumulative caries experience. With respect to root-surface caries, it is of interest to note that both root-surface lesions and root-surface restorations were positively associated with serum IgG levels to S. mutans (see, for example, Fig. 3), and that, in the final model, root-surface fillings maintained this significant positive association, even after adjustment for the other model components. The manner in which immune responses to S. mutans antigens participate in oral host defenses cannot be deduced from this study. We are likely to be observing a cumulative host response to a life-long association with S. mutans. Immune responses to these micro-organisms may play several roles over one's lifetime. Initial immune responses to S. mutans antigens may limit or delay initial infection. Antibody may also serve to limit accumulation of mutans streptococci at various stages in the infection. Disease may progress in spite of these protective antibody activities. Disease would increase the antigenic load and the consequent immune response. Teeth lost because of caries would be major contributors to this stimulatory process. Thus, the crosssectional picture of antibody levels in later life may bear little relationship to potentially protective interactions of the host immune network and initial infections with S. mutans.
Acknowledgments. The skillful technical assistance of Mary Ritchie is greatly appreciated. The authors also acknowledge the cheerful cooperation of Peggy Resker in the clinical aspects of the study. REFERENCES Block MS, Smith DJ, Ebersole JL, Taubman MA (1979). Effects of saliva and sera from caries-free and caries-prone subjects on glucosyltransferase (GTF) (abstract). J Dent Res 58:145. Brandtzaeg P (1984). The oral secretary immune system with special
emphasis on its relation to dental caries. Proc Finn Dent Soc 79:7 184. Brown LR, Billings RJ, Kaster AG (1986). Quantitative comparisons of potentially cariogenic micro-organisms cultured from noncarious and carious root and coronal tooth surfaces. Infect Immun 51:765770. Challacombe SJ (1980). Serum and salivary antibodies to Streptococcus mutans in relation to the development and treatment of human dental caries. Arch Oral Biol 25:495-502. Challacombe SJ, Bergmeier LA, Czerkinsky C, Rees AS (1984). Natural antibodies in man to Streptococcus mutans: specificity and quantitation. Immunol 52:143-150. Challacombe SJ, Stephenson PA, Giel HM, Wilton JMA (1986). Specific antibodies and opsonic activity in human crevicular fluid. Quantitation and relationships with disease. In: Lehner T, Cimasoni G, editors. The borderland between caries and periodontal disease. Vol. III. Geneva (Switzerland): Medecine et Hygiene, 87-103. Cisar JO, Sandberg AL, Mergenhagen SE (1987). The function and distribution of different fimbriae on strains of Actinomyces viscosus and Actinomyces naeslundii. J Dent Res 63:393-396. Ebersole JL, Taubman MA, Smith DJ (1985). Gingival crevicular fluid antibody to oral microorganisms. II. Distribution and specificity of local antibody responses. J Periodont Res 20:349-356. Ebersole JL, Taubman MA, Smith DJ, Frey DE (1986). Human immune responses to oral microorganisms: patterns of systemic antibody levels to Bacteroides species. Infect Immun 51:507-513. Ellen RP, Banting DW, Fillery ED (1985). Streptococcus mutans and Lactobacillus detection in the assessment of dental root-surface caries risk. J Dent Res 64:1245-1249. Emilson CG, Klock B, Sanford CB (1988). Microbial flora associated with the presence of root-surface caries in periodontally treated patients. Scand J Dent Res 96:40-49. Emilson CG, Krasse B (1985). Support for an implication ofthe specific plaque hypothesis. Scand J Dent Res 93:96-104. Freund R, Littell R, Spector P (1986). SAS system for linear models. Cary (NC): SAS Institute Inc., chapter 1. Fure S, Romaniec M, Emilson CG, Krasse B (1987). Proportions of Streptococcus mutans, Lactobacilli and Actinomyces spp. in rootsurface plaque. Scand J Dent Res 95:119-123. Gahnberg L, Smith DJ, Taubman MA, Ebersole JL (1985). Salivary IgA antibody to glucosyltransferase of oral microbial origin in children. Arch Oral Biol 30:551-556. Greene JC, Vermillion JR (1960). The oral hygiene index: a method for classifying oral hygiene status. J Am Dent Assoc 61:172-179. Gregory RL, Filler SJ, Michalek SM, McGhee JR (1986). Salivary immunoglobulin A and serum antibodies to Streptococcus mutans ribosomal preparations in dental caries-free and caries-susceptible human subjects. Infect Immun 51:348-351. Huis in't Veld JHJ, van Palenstein-Helderman WH, Sampaio Carmago P, Backer Dirks 0 (1979). Antibodies against Streptococcus mutans and glucosyltransferases in caries-free and caries-active military recruits. Adv Exp Med Biol 107:369-381. Jordan HV, Hammond BF (1972). Filamentous bacteria isolated from human root-surface caries. Arch Oral Biol 17:1333-1342. Kleinbaum D, Kupper L, Muller K (1988). Applied regression analysis and other multivariate methods. 2nd ed. Boston (MA): PWS-KENT Publishing Co., chapter 16. Lehner T (1983). Scientific basis for vaccination against dental caries. Proc Finn Dent Soc 79:62-70. Levine M, Parker DW, Strober JA (1984). Human serum precipitins to oral bacteria related to dental caries. Arch Oral Biol 29:191-194. Loe H, Silness J (1963). Periodontal disease in pregnancy. I. Prevalence and severity. Acta Odontol Scand 21:532-551. Loesche WJ, Rowan J, Straffon LH, Loos PJ (1975). Association of Streptococcus mutans with human dental decay. Infect Immun 11: 1252-1260. Loesche WJ, Straffon LH (1979). Longitudinal investigation of the role
Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
Vol. 71 No. 7
HUMORAL ANTIBODY TO ORAL MICROBIOTA
of Streptococcus mutans in human fissure decay. Infect Immun 26:498-507. Michalek SM, Childers NK (1990). Development and outlook for a caries vaccine. Crit Rev Oral Biol Med 1:37-54. Moore WEC, Holdeman LV, Smibert RM, Hash DE, Burmeister JA, Ranney RR (1982). Bacteriology of severe periodontitis in young adult humans. Infect Immun 38:1137-1148. Morrison D (1967). Multivariate statistical methods. New York (NY): McGraw-Hill, chapter 6. Nyvad B, Kilian M (1990). Microflora associated with experimental rootsurface caries in humans. Infect Immun 58:1628-1633. Ravald N, Hamp S-E, Birkhed D (1986). Long-term evaluation of root surface caries in periodontally treated patients. J Clin Periodontol 13:758-767. Smith DJ, Gadalla LM, Ebersole JL, Taubman MA (1985). Gingival
1407
crevicular fluid antibody to oral micro-organisms. III. Association of antibody levels in gingival homogenates with gingival crevicular fluid. J Periodont Res 20:357-364. Smith DJ, King WF, Taubman MA (1990). Salivary IgA antibody to oral streptococcal antigens in predentate infants. OralMicrobiolImmunol 5:57-62. Tanner ACR, Dzink JL, Ebersole JL, Socransky SS (1987). Wolinella recta, Campylobacter concisus, Bacteroides gracilis, and Eikenella corrodens from periodontal lesions. J Periodont Res 22:327-330. Van Houte J, Jordan HV, Laraway R, Kent R, Soparkar PM, DePaola PF (1990). Association of the microbial flora of dental plaque and saliva with human root-surface caries. J Dent Res 69:1463-1468. Zambon JJ, Slots J, Genco Rd (1983). Serology of oral Actinobacillus actinomycetemcomitans and serotype distribution in human periodontal disease. Infect Immun 41:19-27.
Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
J Dent Res 71(7):1399-1407, July, 1992
Introduction. Many components of host defense are present in the oral cavity. Components that enter the oral cavity via the gingival crevicular fluid emanate both from the systemic circulation and from local gingival synthesis (Lehner, 1983). Glandular saliva contains both antibody that originates in interstitial plasma cells, and innate immune components (e.g., lysozyme, lactoferrin, sialoperoxidase, and mucins) that originate in acinar cells and ductal epithelium (reviewed by Brandtzaeg, 1984; Michalek and Childers, 1990). Although it is not yet clear how antibody-mediated responses to oral flora are specifically integrated into the Received for publication August 7, 1991 Accepted for publication January 30, 1992 3To whom correspondence and reprint requests should be addressed This investigation was supported by USPHS Research Grant DE-07009 from the National Institute of Dental Research, National Institutes of Health, Bethesda, MD 20892.
oral host-defense system, the antibody specificities that result from recognition of bacterial antigenic epitopes can provide a record of infection with micro-organisms that cause disease. Dental caries on coronal surfaces is an infectious disease that appears to be initiated on coronal surfaces by mutans streptococci (Loesche et al., 1975; Loesche and Straffon, 1979; Emilson and Krasse, 1985). To understand the role of the systemic immune response to the organisms causing disease, several investigators have attempted to correlate levels of serum antibody to plaque micro-organisms with indicators of oral microbial pathogenesis (e.g., DMF). Since these studies used children or young adults, nearly all disease occurred on coronal surfaces. Cross-sectional studies in these populations have generally shown that high levels of serum IgG antibody to mutans streptococcal cells or antigens were associated with low caries experience (Challacombe, 1980; Challacombe et al., 1984, 1986; Gregoryet al., 1986). In this regard, IgG, purified from the sera of caries-free young adults, significantly inhibited the activity of mutans streptococcal enzymes associated with plaque formation (Block et al., 1979). Exposed dental root surfaces may have a different susceptibility to caries than do coronal surfaces. Intrinsic differences in mineral composition and organic matrix may alter factors involved in the colonization potential of organisms within the oral milieu. This may lead to different cariogenic flora on coronal and root surfaces (Jordan and Hammond, 1972; Ellen et al., 1985; Nyvad and Kilian, 1990), although mutans streptococci have been implicated in dental caries on both surfaces (van Houte et al., 1990). The level and specificities of antibody to the microbiota colonizing root surfaces may provide evidence for the organisms responsible for root-surface caries. Therefore, this study evaluated potential associations between clinical variables and serum IgG antibodies to nine oral micro-organisms in a population at risk for root-surface caries.
Materials and methods. Subjects.-Three hundred and fourteen subjects between the ages of 45 and 65 years were included in this study. This ambulatory volunteer population was nearly equally divided between males (51.9%) and females (48.1%). Each subject received a comprehensive clinical examination at Forsyth Dental Center. Subjects had a minimum of ten teeth; 293 subjects had gingival recession of 1 mm or more on at least three teeth. The population statistics are given in Table 1. Subjects were later divided into three groups, based on their DMF scores, for description of any associations of antibody data with DMF levels. These three groups (low, middle, high) had the following characteristics (n) (mean ± standard deviation) (minimum-maximum): low third (98) (15.2 ± 2.8) (2-18); middle third (115) (20.4 ± 1.2) (19-22); upper third (101) (25.1 ± 2.0) (23-31). ELISA.-Blood was taken from the antecubital fossa. Serum was then obtained from the coagulated specimen after centrifugation and stored frozen at -20'C until use. Sera were tested for antibody activity by a previously described biotin-avidin, alkaline phosphatase enzyme-linked immunosorbent assay (ELISA) method (Smith et al., 1990). Polystyrene microtiter plates (Flow Laboratories, McLean, VA) were coated with approximately 108 bac-
Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
1399
1400
J Dent Res July 1992
KENT et al.
phosphatase (Zymed) and p-nitrophenylphosphate. Antibody activity was expressed as ELISA units (EU) after comparison of experimental optical densities (Artek Industries, Farmingdale, NY) that fell within the linear portion of a reference standard curve constructed from dilutions of pooled reference sera containing high levels of antibody activity to the respective antigen. Analyses of antibody to supragingival micro-organisms were performed on sera from all 314 subjects. Measurement of antibody to subgingival micro-organisms was performed on sera from 248 of the 314 subjects. The subgingival microbial antigen set was included in the analyses under the assumption that these micro-organisms would not be expected to be involved in the formation of dental caries. Antigen preparation. -All bacterial cell preparations were formalin-killed after growth in PPLO broth, supplemented with glucose for the supragingival micro-organisms, and supplemented with hemin and menadione for the subgingival micro-organisms. Bacterial cell antigens included Streptococcus mutans strain Ingbritt, S. faecalis strain 19433, Actinomyces viscosus strain WVU 626, Actinomyces naeslundii strain 12, Lactobacillus casei, Actinobacillus actinomycetemcomitans strainY4, Porphyromonas gingivalis strain 381,Eikenella corrodens strain 1073, and Wolinella recta strain 371. Glucosyltransferases, prepared from Streptococcus mutans (two strains) and S. sobrinus (one strain) by adsorption techniques previously described (Gahnberg et al., 1985), were pooled and used to coat plates. Clinical variables.-Each subject received an examination that included the following measurements: number and locations of missing teeth, coronal caries and restorations by tooth surface, root lesions and restorations by root surface, debris and calculus index (Greene and Vermillion, 1960) applied to every tooth, buccal and lingual recession (the maximum distance from the cemento-enamel junction to the gingival margin on the buccal and lingual surfaces of every tooth), pocket depth (taken as the distance in mm from the gingival margin to the epithelial attachment at the mid-line as well as the mesio- and disto-buccal, and mesio- and disto-lingual papilla of each tooth), and gingival status (Loe and Silness, 1963). Incipient, arrested, and secondary caries were combined with frank cavitation in the estimation of root-surface and coronal caries.
TABLE 1 POPULATION STATISTICS
n
314
MALES
51.9%
FEMALES
48.1%
Mean
SD
Range
56.53
5.47
25.00
Gingival index
0.76
0.49
1.90
Debris
0.92
0.32
1.91
Coronal caries
1.75
4.08
43.00
Coronal fillings
32.81
17.26
91.00
% Pockets > 3 mm
7.58
11.02
62.50
Missing teeth
7.67
4.53
22.00
12.34
4.99
26.00
% Teeth recessed > 2 mm 11.49
14.05
72.73
Age
Filled teeth Root-surface caries
0.82
1.57
12.00
Root fillings
1.39
2.39
19.00
20.30
4.45
31.00
DMF
terial cells, or with 0.5 jig of glucosyltransferase (GTF) purified from three different mutans streptococcal strains by techniques previously described (Gahnberg et al., 1985). Antibody activity was then measured by incubation with 1:100-1:400 dilutions of sera. Plates were then developedfor IgG antibody with biotinylated affinity-purified anti-gamma-chain reagents (Zymed, South San Francisco, CA), followed in sequence by streptavidin-alkaline
TABLE 2
CORRELATIONS AMONG SERUM IgG ANTIBODY LEVELS Bacterial Strain
Pearson Correlation Coefficient*
S. mutans
1
GTF
0.04
1
A. naeslundii A. viscosus
0.39
0.00
1
0.11
0.12
0.31
1
L. casei
0.15
0.18
0.12
0.26
1
S. faecalis
0.14
0.09
0.06
0.4
0.09
1
W. recta
0.06
0.07
-0.02
0.12
-0.01
-0.05
1
A. actino. Y4
0.01
0.07
-0.09
0.08
0.09
-0.02
0.41
1
P. gingivalis
0.07
0.01
-0.05
0.02
-0.01
-0.07
0.32
0.49
1
E. corrodens
-0.02
0.03
-0.03
-0.01
0.08
-0.02
0.29
0.58
0.39
Supragingival
Subgingival
1
GTF Smut Anaes Avis Lcas Wrec AaY4 Bging Ecor Sfaec *Single underline indicates a p < 0.05 level of significance. Double underline indicates a p < 0.0001 level of significance. Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
HUMORAL ANTIBODY TO ORAL MICROBIOTA
Vol. 71 No. 7
Statistical methods.-
Summary descriptive statis-
1401
100
A. actinomycetmcomlans
go tics for clinical variables observed on 314 subjects are given in Table 1. Frequency 40. distributions are given for serum IgG antibody levels to 20 the battery of ten antigens 0 ° (Fig. 1). For reduction of _ skewness in the distributions, E W. so ELISAunits wereloglo-trans*-A. neeslundli formed for all analyses. All 40_ but the A. naeslundii distri* _ bution were unimodal, al_ _ though nearly all the distri10 butions still exhibited some o degree of skewness to the 32 3.6 OA 0.6 12 1.6 2.0 24 right. Product-moment correlations were calculated for Log10 Antibody Level (EU) evaluation ofall pair-wise asloo sociations among the antibody levels (Table 2). In adGTF dition, correlations were calW rea culated between log1o antiso body levels for the six su4* pragingival members of the battery and "caries-related" j variables (including DMF, coronal caries, root-surface C_ 701 lesions, missing teeth, coro_ . nal restorations, root-surface E so E r restorations, and debris) and Z so between log10 antibody lev*0 els for the four subgingival 30_ members of the battery and "periodontitis-related" clini10 cal measures (including recession, pocket depth, and 0 o 0 O3 A 0 12 1.5 1A 2.1 2.4 2.7 3.0 OA 0.9 12 1.5 1.6 2.1 2. 2.7 3.0 3.3 3.6 missing teeth). Multiple linear regression Log10 Antibody Leve (EU) Log1* Antibody Level (EU) was used for examination of so multivariate associations between levels ofS. mutans IgG A. viecosus S. mutans GO antibody, as the response variable in the statistical model, 40 and a variety of clinical variables, as explanatory vari_ ables in the model. An initial model related log antibody level to age, gender, and DMF score. The contribuP. glngivalls S. fecalls tions to this association of W the three components of the composite DMF score were *evaluated further by forward step-wise regression (Kleinbaum et al., 1988), in a which, instead ofDMF, separate variables-representing 0.4 GA 12 IA 2. 24 U 32 3A 4.0 4. 0.4 OA CA 1.0 11 IA IA 2.0 2.2 2.5 number of missing teeth, % Logo AntIbody LUr (EU) Log1o Antibody Level (EU) teeth (present) with enamel distributions of IgG antibody to supra- and subgingival micro-organisms. The ELISA coating Fig. for1-Frequency each formalin-killed bacterial restorations, and % teeth antigen species is given in the upper right corner of each plot. Antibody is reported (present) with enamel car- as log1o EU. ies-were utilized as explanatory variables. Age, gender, debris, and measures of root-surface variable to the resulting regression models (Tables 3 and 4) are lesions and root-surface restorations were also candidate ex- summarized by the associated regression coefficients which deplanatory variables in these analyses. The contributions of each fine the regression model and by mean squares and probability Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
KENT et al.
1402 2.6 2.4 2.2 LJ 2.0 0 0' 1.8 0 1.6 0 .4 1.4 1.2 2.4 2.2 2.0 1.8 1.6 1.4 1.2
-
J Dent Res July 1992
:~ ~ ~ ~
2.6
a) S. mutans
2.4
d) GTF
2.2 Li
(Table 2). Micro-organismswere grouped according to their supra-
2.0
or subgingival location. Correlation coefficients above 0.25 in Table 2 were significant at the p < 0.0001 level. With the given sample sizes, correlations larger than ( r >) 0. 1 were statisti-
1.8 0
0
1.6 1.4
1.2
- b) A. viscosus
2.4
transformed levels of IgG antibodies reactive with the ten micro-organisms in the battery
e) S. foecalis
cally significant (p
3 mm) was lowest for subjects with generally lower antibody levels to subgingival species and highest among subjects with generally higher antibody levels to-subgingival species (by the rank-sums). Means for gingival recession and, to some extent, missing teeth were lower among subjects having higher ranksums. Regression of the antibody rank-sum on age, pocket depth (% sites > 3 mm) and recession (% sites > 2 mm) was statistically significant (p = 0.003), but explained a minimal amount of variation (r2 = 0.06). Coefficients were positive for pocket depth (p = 0.04), and negative for recession (p = 0.02).
Discussion. Investigations of the etiology of dental caries consistently implicate mutans streptococci in the initiation of this disease on coronal surfaces (Loesche et al., 1975; Loesche and Straffon, 1979; Emilson and Krasse, 1985). Although the association of mutans streptococci with root-surface caries is less compelling, several observations suggest a similar etiology. Mutans streptococci have been reported to be more prevalent in the saliva (Ravald et al., 1986; Emilson et al., 1988; van Houte et al., 1990), on sound exposed root surfaces, and in incipient lesions on exposed root surfaces (Brown et al., 1986; van Houte et al., 1990) of root-cariesactive compared with root-caries-free subjects. The population described in the present investigation has had significant coronal caries experience and is also in an age range in which the degree
Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
HUMORAL ANTIBODY TO ORAL MICROBIOTA
Vol. 71 No. 7
1405
3.0 of gingival recession puts many 12 2.7 at risk for root-surface caries. d) GINGIVITIS TH 2.4 IgG antibody to S. mutans was POCKET a) DEPI l0 x Li 2.1 the only serological response 0 A; z 1.8 ofthe six supragingival plaque A -j 1.5 antigens examined that showed V) 1.2 significant associations with Li 0 z 0.9 clinical variables relating to U) 4 0.6 dental caries. For example, 0.3 only serum IgG antibody to S. 0 mutans showed a significant, positive correlation with sube) GENDER b) RECESSION 0.8 ject DMF scores (Fig. 2; Table Li. z 3). Furthermore, multiple lin- CNJ 0 0.6 ear-regression analysis of S. A : 0 mutans antibody on components a)L'i CL 0.4 I 0 ofthe DMF score also resulted V) CL in a model with very high sta0.2 tistical significance (Table 4). A degree of correspondence ( existed between serum IgG anarf) AGE 14 - c) MISSING TEETH tibody levels to oral microbiota 70K 12 and their oral locations (Table 60 H 2). For example, serum-anti- z 10 XXX 50U)J body levels to six supragingi- I 8 )OOC )O00C 4 >00( 20tial number of significant, posiXXX 2 >00( tive correlations among mem10 H 0 - . - I bers within each battery. 0 L H M L M Among the six supragingival L H organisms, serum-antibody ANTIBODY RANK SUM levels exhibited a pattern of low, positive correlations, which included several highly signifiFig. 4-Association of subgingival microbiota rank-sums with clinical variables. The EU values for antibody to cant correlations. Clearly, the P. gingivalis, A. actinomycetemcomitans, E. corrodens, and W. recta were rank-ordered, and a sum ofranks for the strong correlation between an- four antibody specificities was calculated for each subject. Means (± SE) are given for clinical variables for subjects tibody to A. viscosus and A. grouped, on the basis of this periodontal antibody rank-sum, into low (L), middle (M), and high (H) antibody levels. naeslundii cells is likely to be at least partly related to cross-reacting antigens (Cisar et al., suggest a measure of specificity, at least in the local immune 1987). It is unclear, however, why the correlation between S. response (Ebersole et al., 1985; Smith et al., 1985), since a mutans antibody andA. viscosus antibody should not be of at least collective elevation has not been found in gingival crevicular fluid the same order of magnitude as that between S. mutans and A. antibody levels to these organisms. It should be noted that these naeslundii, since both micro-organisms cohabit both non-carious associations were quite modest, possibly reflecting the fact that sites and caries lesions (Fure et al., 1987; Emilson et al., 1988; this population was not selected for those with periodontal disease. van Houte et al., 1990). No significant correlations were found DMF was positively associated with serum-antibody level for between antibody levels to supragingival organisms and antibodS. mutans (r = 0.23, p = 0.0001; Fig. 2; Table 3) but showed no ies to subgingival organisms. The most remarkable array of correlations existed among association with antibody levels to other supragingival organisms. When the associations among clinical variables and S. serum antibodies to A. actinomycetemcomitans, P. gingivalis, E. corrodens, and W. recta (r = 0.29 0.49). These four micro- mutans antibody were explored further, regression analysis of S. organisms are frequently cultured from subgingival locations mutans antibody level (log10) on age, gender, and DMF was highly (Moore et al., 1982; Tanner et al., 1987), and at least two of them significant (p = 0.0001), but explained only a small proportion of (A. actinomycetemcomitans and P. gingivalis) have been associ- the variation (r2 = 0.08). Step-wise regression analysis of S. ated with various forms of periodontal diseases. These correla- mutans antibody level (loglo) on a set of clinical parameters tions are unlikely to result from responses to common antigens, resulted in a model (r2 = 0.14) that included measures of cumusince studies with Actinobacillus (Zambon et al., 1983) and lative caries experience (missing teeth, enamel and root-surface Porphyromonas (Ebersole et al., 1986) strains and human or restorations) and a present/absent index variable for enamel animal sera indicated little cross-reactivity between these gen- lesions. The model supported the conclusion that the associations seen with S. mutans antibody reflected previous encounters with era. When IgG antibodies to these four micro-organisms were rank-ordered and summed for each subject, an apparent positive S. mutans antigens, since this model emphasized cumulative correlation existed between increased antibody level and pocket caries experience over measures reflecting a subject's current depth (Fig. 4a). It is conceivable that an immune (perhaps status. The associations seen with S. mutans were not seen with GTF, polyclonal) responses) occurred to several microbial species that had presented infectious challenges during the extension of the an antigen derived from mutans streptococci. This lack of corpocket. However, IgG antibody levels in gingival crevicular fluids relation may have occurred because the immunodominance of S. and in gingival tissues of subjects with periodontal disease would mutans components-such as antigen I/II, GTF, lipoteichoic acid, 8k
,
14
-
-
T-
I-
-
Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
1406
KENT et al.
J Dent Res July 1992
or serotype carbohydrate-is likely to vary among subjects. Regardless of which S. mutans immunodominant antigen is recognized by a subject, all subjects would still react with the intact S. mutans antigen, since it contains all components. However, those subjects for whom an antigen other than GTF was immunodominant would be positively associated with the intact S. mutans, but not associated with GTF. Alternatively, the specificity seen with the intact S. mutans species may have been blurred by the use of pooled GTF from S. mutans and S. sobrinus species. Many previous studies have examined the association between serum IgG antibodies to supragingival plaque micro-organisms (reviewed in Michalek and Childers, 1990). Antigens in most studies were limited to oral streptococci, and studies were performed with sera from 18-25-year-old subjects. Both positive (Huis in't Veld et al., 1979) and negative associations (Block et al., 1979; Challacombe, 1980; Levine et al., 1984; Gregory et al., 1986) were found between serum IgG antibody to mutans streptococci and caries experience. The present study, where a positive association was observed, differed importantly from essentially all of the foregoing studies in that the cohort was much older (mean age, 56.7 years). This resulted in a population whose mean levels of caries experience (DMF) were generally equivalentto those ofthe so-called "high caries" groups ofthe earlier studies. Furthermore, the cohort in the present study was decades removed from the "caries-prone" years of adolescence. Thus, if mutans streptococci frequently initiate dental caries, and if mutans streptococcal antigens elicit serum IgG antibody during this process (Challacombe, 1980; Challacombe et al., 1984), which takes place relatively early in life, and if a memory lymphocyte cell population is subsequently maintained as a result of the underlying mutans streptococcal infection, then itis not surprisingthat serum IgG antibody levels to S. mutans were most closely associated with cumulative caries experience. With respect to root-surface caries, it is of interest to note that both root-surface lesions and root-surface restorations were positively associated with serum IgG levels to S. mutans (see, for example, Fig. 3), and that, in the final model, root-surface fillings maintained this significant positive association, even after adjustment for the other model components. The manner in which immune responses to S. mutans antigens participate in oral host defenses cannot be deduced from this study. We are likely to be observing a cumulative host response to a life-long association with S. mutans. Immune responses to these micro-organisms may play several roles over one's lifetime. Initial immune responses to S. mutans antigens may limit or delay initial infection. Antibody may also serve to limit accumulation of mutans streptococci at various stages in the infection. Disease may progress in spite of these protective antibody activities. Disease would increase the antigenic load and the consequent immune response. Teeth lost because of caries would be major contributors to this stimulatory process. Thus, the crosssectional picture of antibody levels in later life may bear little relationship to potentially protective interactions of the host immune network and initial infections with S. mutans.
Acknowledgments. The skillful technical assistance of Mary Ritchie is greatly appreciated. The authors also acknowledge the cheerful cooperation of Peggy Resker in the clinical aspects of the study. REFERENCES Block MS, Smith DJ, Ebersole JL, Taubman MA (1979). Effects of saliva and sera from caries-free and caries-prone subjects on glucosyltransferase (GTF) (abstract). J Dent Res 58:145. Brandtzaeg P (1984). The oral secretary immune system with special
emphasis on its relation to dental caries. Proc Finn Dent Soc 79:7 184. Brown LR, Billings RJ, Kaster AG (1986). Quantitative comparisons of potentially cariogenic micro-organisms cultured from noncarious and carious root and coronal tooth surfaces. Infect Immun 51:765770. Challacombe SJ (1980). Serum and salivary antibodies to Streptococcus mutans in relation to the development and treatment of human dental caries. Arch Oral Biol 25:495-502. Challacombe SJ, Bergmeier LA, Czerkinsky C, Rees AS (1984). Natural antibodies in man to Streptococcus mutans: specificity and quantitation. Immunol 52:143-150. Challacombe SJ, Stephenson PA, Giel HM, Wilton JMA (1986). Specific antibodies and opsonic activity in human crevicular fluid. Quantitation and relationships with disease. In: Lehner T, Cimasoni G, editors. The borderland between caries and periodontal disease. Vol. III. Geneva (Switzerland): Medecine et Hygiene, 87-103. Cisar JO, Sandberg AL, Mergenhagen SE (1987). The function and distribution of different fimbriae on strains of Actinomyces viscosus and Actinomyces naeslundii. J Dent Res 63:393-396. Ebersole JL, Taubman MA, Smith DJ (1985). Gingival crevicular fluid antibody to oral microorganisms. II. Distribution and specificity of local antibody responses. J Periodont Res 20:349-356. Ebersole JL, Taubman MA, Smith DJ, Frey DE (1986). Human immune responses to oral microorganisms: patterns of systemic antibody levels to Bacteroides species. Infect Immun 51:507-513. Ellen RP, Banting DW, Fillery ED (1985). Streptococcus mutans and Lactobacillus detection in the assessment of dental root-surface caries risk. J Dent Res 64:1245-1249. Emilson CG, Klock B, Sanford CB (1988). Microbial flora associated with the presence of root-surface caries in periodontally treated patients. Scand J Dent Res 96:40-49. Emilson CG, Krasse B (1985). Support for an implication ofthe specific plaque hypothesis. Scand J Dent Res 93:96-104. Freund R, Littell R, Spector P (1986). SAS system for linear models. Cary (NC): SAS Institute Inc., chapter 1. Fure S, Romaniec M, Emilson CG, Krasse B (1987). Proportions of Streptococcus mutans, Lactobacilli and Actinomyces spp. in rootsurface plaque. Scand J Dent Res 95:119-123. Gahnberg L, Smith DJ, Taubman MA, Ebersole JL (1985). Salivary IgA antibody to glucosyltransferase of oral microbial origin in children. Arch Oral Biol 30:551-556. Greene JC, Vermillion JR (1960). The oral hygiene index: a method for classifying oral hygiene status. J Am Dent Assoc 61:172-179. Gregory RL, Filler SJ, Michalek SM, McGhee JR (1986). Salivary immunoglobulin A and serum antibodies to Streptococcus mutans ribosomal preparations in dental caries-free and caries-susceptible human subjects. Infect Immun 51:348-351. Huis in't Veld JHJ, van Palenstein-Helderman WH, Sampaio Carmago P, Backer Dirks 0 (1979). Antibodies against Streptococcus mutans and glucosyltransferases in caries-free and caries-active military recruits. Adv Exp Med Biol 107:369-381. Jordan HV, Hammond BF (1972). Filamentous bacteria isolated from human root-surface caries. Arch Oral Biol 17:1333-1342. Kleinbaum D, Kupper L, Muller K (1988). Applied regression analysis and other multivariate methods. 2nd ed. Boston (MA): PWS-KENT Publishing Co., chapter 16. Lehner T (1983). Scientific basis for vaccination against dental caries. Proc Finn Dent Soc 79:62-70. Levine M, Parker DW, Strober JA (1984). Human serum precipitins to oral bacteria related to dental caries. Arch Oral Biol 29:191-194. Loe H, Silness J (1963). Periodontal disease in pregnancy. I. Prevalence and severity. Acta Odontol Scand 21:532-551. Loesche WJ, Rowan J, Straffon LH, Loos PJ (1975). Association of Streptococcus mutans with human dental decay. Infect Immun 11: 1252-1260. Loesche WJ, Straffon LH (1979). Longitudinal investigation of the role
Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
Vol. 71 No. 7
HUMORAL ANTIBODY TO ORAL MICROBIOTA
of Streptococcus mutans in human fissure decay. Infect Immun 26:498-507. Michalek SM, Childers NK (1990). Development and outlook for a caries vaccine. Crit Rev Oral Biol Med 1:37-54. Moore WEC, Holdeman LV, Smibert RM, Hash DE, Burmeister JA, Ranney RR (1982). Bacteriology of severe periodontitis in young adult humans. Infect Immun 38:1137-1148. Morrison D (1967). Multivariate statistical methods. New York (NY): McGraw-Hill, chapter 6. Nyvad B, Kilian M (1990). Microflora associated with experimental rootsurface caries in humans. Infect Immun 58:1628-1633. Ravald N, Hamp S-E, Birkhed D (1986). Long-term evaluation of root surface caries in periodontally treated patients. J Clin Periodontol 13:758-767. Smith DJ, Gadalla LM, Ebersole JL, Taubman MA (1985). Gingival
1407
crevicular fluid antibody to oral micro-organisms. III. Association of antibody levels in gingival homogenates with gingival crevicular fluid. J Periodont Res 20:357-364. Smith DJ, King WF, Taubman MA (1990). Salivary IgA antibody to oral streptococcal antigens in predentate infants. OralMicrobiolImmunol 5:57-62. Tanner ACR, Dzink JL, Ebersole JL, Socransky SS (1987). Wolinella recta, Campylobacter concisus, Bacteroides gracilis, and Eikenella corrodens from periodontal lesions. J Periodont Res 22:327-330. Van Houte J, Jordan HV, Laraway R, Kent R, Soparkar PM, DePaola PF (1990). Association of the microbial flora of dental plaque and saliva with human root-surface caries. J Dent Res 69:1463-1468. Zambon JJ, Slots J, Genco Rd (1983). Serology of oral Actinobacillus actinomycetemcomitans and serotype distribution in human periodontal disease. Infect Immun 41:19-27.
Downloaded from jdr.sagepub.com at UNIVERSITE LAVAL on July 14, 2015 For personal use only. No other uses without permission.
