IgG Antibody Levels to Porphyromonas gingivalis and Clinical ...

2330.qxd

2/20/04

11:23 AM

Page 221

J Periodontol • February 2004

IgG Antibody Levels to Porphyromonas gingivalis and Clinical Measures in Children Cara L. Donley,*† Rachel Badovinac,† Shabtai Sapir,‡ Lior Shapira,‡ Yael Houri,‡ Alpdogan Kantarci,§ Martha L. Warbington,§ Serge Dibart,§ Thomas E. Van Dyke,§ Howard L. Needleman,*† Nadeem Karimbux,† and Enrique Bimstein

Background: Periodontopathic clinical markers are poorly understood in the pediatric population. Several studies have proposed Porphyromonas gingivalis (P. gingivalis) and an antibody response to the microorganism as factors in periodontal tissue destruction in children. The objective of this study was to examine the prevalence of P. gingivalis in dental plaque and of serum immunoglobulin G (IgG) antibody levels to P. gingivalis, and their relationship to periodontal clinical measures in children. Methods: Thirty-one subjects, aged 20 to 163 months, participated in this study. Clinical measures examined included gingivitis, plaque, alveolar bone height, age, gender, ethnicity, medical status, caries, and IgG antibody levels to P. gingivalis. Five ml of blood was collected for serum analysis, and IgG antibody levels to P. gingivalis were determined by using enzyme-linked immunosorbent assay. Plaque samples were examined for the presence of P. gingivalis by DNA-DNA checkerboard. Data were analyzed on a person-level basis for relationships to serum IgG antibody levels to P. gingivalis and on a site-specific level for relationships to the presence of P. gingivalis in plaque. Results: A majority (77%) of the subjects were systemically healthy, non-white (74%), and did not have detectable P. gingivalis in their plaque. Fifty-two percent of the subjects had positive serum IgG antibody levels to P. gingivalis. Based on univariate linear regression, factors related to IgG antibody levels to P. gingivalis (P 2 mm. When all clinical measures were considered together, only age remained statistically significantly related to serum IgG antibody levels to P. gingivalis. Conclusions: Age is one of the most important factors in the development of the immune response to putative microorganisms such as P. gingivalis in children. The role of IgG as a timesensitive measure of periodontal health in children needs to be investigated further. J Periodontol 2004;75:221-228. KEY WORDS Age factors; antibody response; children; IgG; immune response; Porphyromonas gingivalis. * The Children’s Hospital, Department of Dentistry, Boston, MA. † Harvard School of Dental Medicine, Boston, MA. ‡ Hadassah School of Dental Medicine, Hebrew University in Jerusalem, Jerusalem, Israel. § Boston University, Goldman School of Dental Medicine, Boston, MA. Hadassah School of Dental Medicine, Hebrew University in Jerusalem; currently, University of Florida, Gainesville, FL.

P

eriodontal disease is the result of an inflammatory and immunological response to pathogens such as Porphyromonas gingivalis, one of the principal players in the initiation and progression of various types of periodontitis. In longitudinal trials, the presence of P. gingivalis has been shown to be among the preeminent risk indicators of periodontitis.1-3 The pathogen has been associated consistently with increased probing depths, attachment loss, and alveolar bone loss, all indicators of periodontal disease.1,4 While some early reports based on cultivation seldom detected P. gingivalis in healthy children, more current and sensitive DNA-based detection methods report a prevalence greater than 30% in healthy children and have shown it to be equally common in children of all ages.5-8 In addition to these direct evaluation methods, the immune response against specific microorganisms also can be used as evidence for the presence and impact of the microbial etiology on the pathogenesis of periodontal disease.9 Immunoglobulin G (IgG) has been shown to be the major antibody class that mediates the host’s immune response against P. gingivalis.10 In adults, IgG antibody levels to P. gingivalis are higher in patients with periodontitis compared to age-matched controls, and the levels are a reflection of subgingival colonization of numerous additional periodontal pathogens.11-13 Differences in serum IgG antibody profiles to subgingival species have also been reported in minority populations.14 Serum IgG titer to 221

2330.qxd

2/20/04

11:23 AM

Page 222

IgG Antibody to P. gingivalis

P. gingivalis has been proposed to serve as an indicator for the susceptibility to periodontal disease.12,15 Studies in children have shown that early contact with P. gingivalis in childhood could lead to development of a humoral immune response.5 It also has been noted that serum antibodies to P. gingivalis are highly prevalent in children.11 Furthermore, in children with aggressive periodontitis, elevated levels of antibodies to several periodontal pathogens, including P. gingivalis, have been reported.16 IgG levels to periodontopathic organisms in children increase with age and with increasing disease severity.11,17 High levels of antibodies in the presence of disease indicate protection is insufficient and the antibodies may be ineffective, even contributing to the pathogenesis of early-onset periodontal disease.13,18 Further research has shown that the clinical diagnosis and severity of gingivitis could be associated with the presence of P. gingivalis and the onset of periodontitis in adolescence.19 The prevalence and severity of gingivitis have been shown to progressively intensify with age and with maturation from the primary to the permanent dentition.20-22 The quality and quantity of dental plaque influence the development and severity of gingivitis in children.20,23 In addition to gingivitis and IgG antibody levels, age plays a role in numerous clinical parameters of periodontal diagnosis.24 The alveolar bone crest to cementoenamel junction (ABC-CEJ) distance changes with age, a phenomenon related to growth and development, attrition, and periodontal disease.25-27 Historically, an ABC-CEJ distance of 2 mm has been utilized to differentiate between normal alveolar bone level and marginal bone loss in children.28 The depth of the gingival sulcus increases with age, favoring a more anaerobic environment for the colonization and stabilization of pathogens such as P. gingivalis.29,30 Few studies have examined the presence of and antibody response to P. gingivalis in children with respect to clinical evaluation of the periodontal status. This information might provide insight for understanding the natural course of periodontal disease and could be utilized to design novel strategies for prevention. Thus, the purpose of this study was to examine the prevalence of P. gingivalis in dental plaque and serum IgG antibody levels to P. gingivalis in children, and their relationship to periodontal clinical measures. MATERIALS AND METHODS Patient Selection and Clinical Evaluation Thirty-one subjects, aged 20 to 163 months, undergoing restorative dental treatment with general anesthesia at The Children’s Hospital Boston participated in this study after obtaining informed parental consent. The study was approved by the Committee on Clini-

222

Volume 75 • Number 2

cal Investigation, Children’s Hospital Boston. Subjects were excluded from participation if antibiotics were taken within 14 days of the operative date; if they were immunosuppressed, status post-transplant, oncology patients, known HIV patients; or taking inhaled steroids. Children were not recruited from a periodontal disease cohort, nor were they screened for gingival inflammation/periodontal disease prior to complete dental treatment under general anesthesia. Each child required at least six permanent or deciduous teeth present for clinical measurements and plaque sampling. Clinical measures examined included gingival inflammation,31 plaque level assessment,32 probing depth, alveolar bone height (ABC-CEJ), age, gender, ethnicity, systemic condition using the American Society of Anesthesiologists’ Physical Class System (ASA),33 and caries evaluation (dmfs + DMFS). Under general anesthesia, six specified teeth and their surfaces (three buccal surfaces of maxillary teeth [#3/A, #8/E, #12/I] and three lingual surfaces of mandibular teeth [#19/K, #24/0, #28/S]) were assigned a gingival score31 and plaque score32 in addition to the routine dental exam, charting, and diagnostic bitewing radiographs. ABC-CEJ distances were measured on bitewing radiographs. In addition to assigning a gingival and plaque score to each of the six teeth for each patient, an average gingival index score (GI average) and average plaque index score (PI average) were determined by adding individual tooth scores and then dividing by the total number of sites scored. Plaque Sample Collection and DNA-DNA Checkerboard Analysis Subgingival plaque samples were obtained with a sterile curet from the same surface of each of the six teeth and placed into six separate Eppendorf tubes containing 150 µl TE buffer (10 mM Tris HCl, 1 mM EDTA, pH 7.6). Within 30 minutes, 100 µl of 0.5 N fresh NaOH was added. Plaque analysis was performed using DNADNA hybridization by checkerboard analysis.34 Briefly, DNA was transferred to a nylon membrane and crosslinked. Membranes were incubated with prehybridization and hybridization solutions. The hybridization solution contains digoxigenin-labeled whole chromosomal DNA probes to Actinomyces naeslundii (43146), Bacteroides forsythus (338), Campylobacter concisus (484), Campylobacter curva (9584), Campylobacter rectus (371), Capnocytophaga sputigena (33562), Eikenella corrodens (23834), Fusonucleatum ss. vincentii (364), Porphyromonas gingivalis (381), Prevotella intermedia (25611), Prevotella nigrescens (33563), Streptococcus oralis (SSII), Treponema denticola (B1), and Actinobacillus actinomycetemcomitans (Y4). After stringent washes to eliminate unbound probes, an antidigoxigenin antibody conjugated to alkaline phosphatase

2330.qxd

2/20/04

11:23 AM

Page 223

Donley, Badovinac, Sapir, et al.


was added, and the signals were detected by incubating the membranes with a chemiluminescent solution and exposing it to a radiographic film. Standards of known DNA quantity for the various species were included on the membrane to allow quantitation. Enzyme-Linked Immunosorbent Assay (ELISA) for Serum IgG Antibody Levels Five milliliters of peripheral blood was collected from each child and placed into a sterile test tube without heparin or anticoagulants. Tubes were placed in a refrigerator at 4°C for 24 hours then centrifuged for 20 minutes (200 to 400 g) at 1,200 rpm. Serum was collected, divided into two subsets, and stored at −70°C until analysis. The serum was analyzed using the ELISA method described by Kojima et al.13 to measure IgG antibody levels to P. gingivalis, strain ATCC 33277, with the serum of a patient with aggressive periodontitis used as a control. Briefly, 96-well plates were coated with 10 µg protein/ml heat-killed P. gingivalis and incubated overnight at −4°C. After washing with 0.05% Tween in phosphate buffered saline and blocking with 2% bovine serum albumin for 1 hour at 37°C, serial dilutions of sera were added and incubated overnight at 4°C. Biotin-conjugated goat anti-human IgG¶ was added and incubated for 1 hour at 37°C, followed by streptavidin-horseradish peroxidase conjugate. O-Phenylendiamine# was used as the substrate. The reaction was stopped by the addition of 4N sulfuric acid, and the optical density was read with a microplate reader** at 490 to 650 nm. The results were expressed as antibody titers by reference to serial dilutions of a serum pool prepared from a positive patient with high levels of the specific antibody. Statistical Methods Person-level descriptive statistics for the study population were calculated. Frequencies and percentages were calculated for the categorical variables. Means and standard deviations, as well as medians and interquartile ranges, were determined for the continuous variables. The following variables were investigated in connection with serum IgG antibody levels to P. gingivalis: age in months, gender, white race, ASA systemic condition,33 P. gingivalis, average plaque index, average gingival index, average probing depth, and number of teeth with ABC-CEJ distance greater than 2 mm. All IgG analyses were conducted on person-level data. First, univariate linear regressions were performed. Those variables that were statistically significant in univariate regression were considered for inclusion in a multivariate model. Pearson correlation coefficients between gingival health indices (gingival index, probing depth, and number of teeth with ABC-CEJ distance greater than 2 mm) found to be significantly

Table 1.

Descriptive Statistics of Study Population (n = 31) N

%

Gender Male Female

17 14

54.8 45.2

Race Non-white African-American Hispanic Asian Other White

23 8 7 2 6 8

74.2 25.8 22.6 6.5 19.4 25.8

ASA 2 mm

1.26 (2.10)

0 (0-2)

IQR = interquartile range.

related to IgG in the univariate analysis were calculated to assess for potential collinearity in the multivariate model. Because significant correlation was detected, the gingival health index that maximized the R2 of the multivariate model was selected for entry into the final model. In connection with the presence of site-specific P. gingivalis, the following variables were considered: age in months, gender, white race, ASA systemic condition,33 site-specific gingival index, site-specific plaque index, average probing depth, number of teeth with ABC-CEJ distance greater than 2 mm, and serum IgG. Descriptive statistics were calculated for those sites with ¶ Jackson Immunoresearch, West Grove, PA. # Zymed Laboratories, Inc., South San Francisco, CA. ** Molecular Devices, Sunnyvale, CA.

223

2330.qxd

2/20/04

11:24 AM

Page 224


and without detected P. gingivalis. Univariate analyses were performed using a generalized linear model with a logit link and an exchangeable correlation structure (Stata command xtgee) to account for the nonindependence of the site-specific data. All statistical analyses were performed using a software program.†† In all cases, the significance level was set at 0.05. RESULTS Description of Population A total of 31 children were included in the study. The median age of the population was 4.6 years, and the majority (77%) of the population was < ASA II (normal, healthy patient, without organic, physiologic, or psychiatric disturbance), was of a non-white race, and was free of detected P. gingivalis. Table 1 shows the complete descriptive statistics calculated for this population. Association Between Serum IgG and Demographic and Clinical Variables The results of the univariate and multivariate linear regressions are shown in Table 2. Fiftytwo percent (n = 16) of the subjects had positive IgG antibody levels to P. gingivalis. On a univariate basis, age in months, average gingival index, average probing depth, and number of teeth with ABC-CEJ >2 mm were significantly associated with serum IgG antibody levels (Table 2). A significant weak


Table 2.

Results of Univariate and Multivariate Linear Regressions (final model) with Serum IgG as the Dependent Variable Univariate

P Value

95% CI

Age (months)

0.11

0.034*

0.0091, 0.21

dmfs

0.022

0.731

−0.11, 0.15

0.0041

0.15

Female

−0.26

0.93

−6.61, 6.09

0.0003

White race

−4.78

0.17

−11.77, 2.22

0.063

4.77

0.19

−4.72

0.14

−11.09, 1.63

0.074

Average plaque index

4.11

0.16

−1.69, 9.91

0.068

Average gingival index

6.15

0.02*

1.02, 11.28

0.17

Average probing depth

4.05

0.013*

0.92, 7.19

0.19

Number of teeth with ABC-CEJ >2 mm

1.74

0.016*

0.36, 3.12

0.19

Coefficient

P Value

95% CI

Age (months)

0.12

0.015†

0.024, 0.21


1.80

0.007†

0.53, 3.06

ASA II+ P. gingivalis

Multivariate

−2.57, 12.11

0.057

* Statistically significant association with serum IgG antibody levels to P. gingivalis (P 2 mm. To avoid the ill effects of collinearity, only one of these variables was entered into the multivariate model with age in months. The number of teeth with ABC-CEJ >2 mm was selected because the model containing age in months and this variable had the highest R2 (0.344). The respective R2 of the models containing average gingival index and average probing depth were 0.337 and 0.342. Table 2 contains the results of the final multivariate analysis. Association Between Site-Specific P. gingivalis and Demographic and Clinical Variables Descriptive statistics for those sites with and without P. gingivalis are shown in Table 3. Eleven of the 31 patients had detectable P. gingivalis in their plaque. Considering all the sampled sites (n = 186), the paucity of P. gingivalis-positive sites in this population is note†† Intercooled Stata 7.0, Stata Corporation, College Station, TX.

2330.qxd

2/20/04

11:24 AM

Page 225

Donley, Badovinac, Sapir, et al.


Table 3.

Table 4.

Descriptive Statistics for Plaque Sample Sites with P. gingivalis (n = 15) and without P. gingivalis (n = 169)*

Results of Univariate Generalized Linear Models (final model) with Site-Level Presence of P. gingivalis as the Outcome*

P. gingivalis Present N (%)

P. gingivalis Absent N (%)

Gender Male Female

10 (66.67) 5 (33.33)

91 (53.85) 76 (46.15)

Race Non-white White

11 (73.33) 4 (26.67)

126 (74.56) 43 (25.44)

ASA 2 mm

0.73 (1.39) 0 (0-1)

1.36 (2.11) 0 (0-2)

Serum IgG

5.35 (6.85) 0.00 (0.00-12.08)

8.30 (8.43) 9.84 (0.00-15.58)

Age (months) dmfs

* Table demonstrates little difference between sites with P. gingivalis and sites without P. gingivalis. IQR = interquartile range.

worthy: 15 sites were detected to be P. gingivalis positive, 169 sites were determined to be P. gingivalis negative, and two sites were P. gingivalis indeterminate. Those two sites were treated as missing data. In the univariate analyses, none of the variables considered was statistically significantly associated with the presence of P. gingivalis. The results of these analyses are provided in Table 4. DISCUSSION In this study, 35% of the patients had detectable P. gingivalis in their plaque, but on a site-specific basis,

Odds Ratio

P Value

95% CI

Age (months)

1.00

0.614

0.99, 1.02

dmfs

0.99

0.562

0.97, 1.02

Female

0.58

0.393

0.17, 2.01

White race

1.06

0.931

0.27, 4.11

ASA II+

0.84

0.812

0.19, 3.65

IgG

0.95

0.249

0.88, 1.03

Average plaque index

1.27

0.542

0.59, 2.72

Average gingival index

1.02

0.969

0.46, 2.26

Average probing depth

0.45

0.120

0.16, 1.23


0.83

0.314

0.57, 1.20

* Each of the variables taken alone was not related to P. gingivalis in the plaque.

only 8% of the total sites sampled had detectable P. gingivalis. The paucity of sites with the pathogen is noteworthy. Previously, it was reported that the prevalence of P. gingivalis varied between 10% to 60% in all age groups from birth to 18 years.6 The latter study utilized polymerase chain reaction (PCR) data to suggest that the pathogen is acquired early in life, immediately on exposure, and maintained throughout the life of the individual. In a recent study on the tongue flora and plaque samples utilizing the DNA probe checkerboard method, P. gingivalis was detected in 68.8% of the samples collected from children 18 to 48 months old. 7 On the other hand, another recent study using PCR technology failed to report the detection of P. gingivalis in the plaque of periodontally healthy children.35 While differences might be explained by the sensitivity of techniques, the sites sampled (e.g., tongue versus teeth), and differences in ages and disease status, questions still arise whether P. gingivalis is indigenous in the oral cavity of periodontally healthy children and whether its presence can be used as a risk indicator of periodontal disease in children. Our study population was obtained from available subjects having dental treatment under general anesthesia who were not screened prior for detectable signs of gingivitis/periodontal disease. Our findings indicate that P. gingivalis can be detected in 35% of the plaque samples of the subjects. Several studies have 225

2330.qxd

2/20/04

11:24 AM

Page 226


suggested that P. gingivalis found in the plaque of healthy children is temporary, suggestive of a transitory colonization by transmission from another heavily colonized individual.36-38 It is speculated that the anaerobic P. gingivalis cannot survive in the oral cavity of a young child without a permanent niche such as a deep anaerobic pocket. In contrast with this speculation, the periodontal status of the mother was not found to be a determinant for detection of periodontal pathogens in the flora samples from children.7 Thus, taken together with our findings, the available data suggest that P. gingivalis can be readily detected in the microflora of young healthy individuals and can present an important risk factor. On the other hand, our finding that none of the clinical parameters was related to P. gingivalis might be indicative that detection of the microorganism does not necessarily account for the disease, and periodontal disease is a multifactorial process. This is in contrast to previous pediatric studies that reported relationships between the pathogen and clinical measures. In one study, the presence of P. gingivalis was associated with the onset and severity of gingivitis,19 while another study reported that the prevalence of P. gingivalis was higher for sites with subgingival calculus, pockets exceeding 3 mm, and bleeding on probing in adolescents.39 In addition, a positive relationship between age and the persistent colonization of P. gingivalis was shown.38 This obvious discrepancy between the existing knowledge and our findings could be linked to the fact that different P. gingivalis clones might play different roles at different stages of periodontal disease. There are multiple clonal structures and ribotypes of P. gingivalis, and the pathogenicity of each clone or type has not been clearly defined.40,41 The lack of agreement between our work and the previous studies demonstrates that further detailed experiments are required to clarify the role of different types or clones of P. gingivalis in the development of periodontal disease. A majority of previous studies have also found that serum IgG antibody levels to P. gingivalis correlated with subgingival P. gingivalis colonization,11,13,15 while our study failed to show any such association. In one report,19 no direct relationship between colonization of P. gingivalis and antibody titers was observed, and this is in agreement with our findings. Earlier exposure to P. gingivalis in our sample group may have resulted in the formation of IgG antibodies, which have since effectively eliminated periodontopathogenic P. gingivalis colonization by opsonization and cytolysis. This could explain our low detection rate of the pathogen. Other explanations for finding IgG antibody levels to P. gingivalis in 52% of the patients, but small numbers of the pathogen, include insufficient

226


sampling, the idea that the formation of the antibodies is the result of stimulation from cross-reactive antigens unrelated to the periodontal status, or the idea that sensitization can be due to chronic low-level exposure, not a reflection of acute changes. This study supports a protective role of IgG antibodies to P. gingivalis, in contrast to a suggestion that the high levels of antibodies contribute to the pathogenesis of periodontal disease in children with aggressive periodontitis.18 However, the antibodies, as proposed,13 may be ineffective in diseased states versus nondiseased states. Our study population was not limited to subjects with periodontal disease or prior signs of gingival inflammation. Additionally, in terms of antibody titers, mere colonization or transient exposures to the pathogen may not induce production of high protective levels, whereas infection progressing to disease can. The presence of IgG antibodies to P. gingivalis was examined in this study, not the functional capabilities. While the IgG antibody levels to P. gingivalis were not related to the pathogen in this study, several clinical parameters were statistically significantly related to serum IgG antibody levels to P. gingivalis. Individually, the clinical measures of age (P