Dr Kazuhisa Yamazaki, Department of Periodontology, Faculty of Dentistry, Niigata University, 5274, Gakkocho-Dori 2-ban-cho, Niigata 951-8514, Japan.E-mail: firstname.lastname@example.org
The presence of antibodies to the 60-kD human and Porphyromonas gingivalis GroEL hsp60 in the sera and inflamed gingival tissues of periodontitis patients was examined. In order to obtain the antigens, recombinant plasmids carrying human hsp60 and P. gingivalis GroEL genes were constructed and expressed as histidine-tagged recombinant proteins. Immunoreactivities of these proteins were confirmed by MoAbs specific to mammalian hsp60 and cross-reactive with both mammalian and bacterial hsp60. Western blot analysis clearly demonstrated that the number of periodontitis patients showing a positive response to P. gingivalis GroEL was higher than the number of periodontally healthy subjects. Furthermore, anti-P. gingivalis GroEL antibody was detected in all samples of gingival tissue extracts. For human hsp60, a higher frequency of seropositivity was found in the periodontitis patients than in the healthy subjects. In addition, the periodontitis patients demonstrated stronger reactivity compared with the healthy subjects. Quantitative analysis of serum antibodies by ELISA also demonstrated that the levels of antibodies in the sera of patients were significantly higher than those of control subjects. In the gingival tissue extracts, seven out of 10 patients demonstrated a positive response to human hsp60 and tso of these demonstrated strong positivity. Affinity-purified serum antibodies to human hsp60 and P. gingivalis GroEL from selected patients reacted with P. gingivalis GroEL and human hsp60, respectively, suggesting cross-reactivity of antibodies. These results suggest that molecular mimicry between GroEL of the periodontopathic bacterium P. gingivalis and autologous human hsp60 may play some role in immune mechanisms in periodontitis.
Hsp60 belong to a family of related proteins which have been conserved during evolution. Despite being highly homologous between prokaryotic and eukaryotic cells, hsp60 are strongly immunogenic and immune responses to microbial hsp60 are speculated to initiate chronic inflammatory diseases in which autoimmune responses to human hsp60 may be central to pathogenesis . T cell clones with self hsp60 reactivity and high titres of antibody responses to self hsp60 have been identified in patients with several chronic inflammatory diseases [2–4].
Chronic inflammatory periodontal disease is characterized by connective tissue destruction and alveolar bone resorption. Although periodontopathic bacteria are the primary aetiological agents, the ultimate determinant of disease progression and clinical outcome is the host’s immune response . It has been reported that GroEL-like protein belonging to the hsp60 family can be expressed by periodontopathic bacteria such as Porphyromonas gingivalis[6,7] and Actinobacillus actinomycetemcomitans. Further, antibodies against P. gingivalis GroEL are present in serum from periodontitis patients . Ando et al. demonstrated that hsp60 is expressed in periodontitis tissues using an anti-human hsp60 antibody which cross-reacted with bacterial GroEL . Therefore, it can be hypothesized that the immune system could be triggered by bacterial antigens, GroEL for example, which share a high degree of homology with self hsp60 proteins, resulting in an aberrant immune response and chronicity of inflammation. In addition, recent epidemiological reports suggested that periodontal diseases are associated with increased risk factors for coronary heart disease [10,11]. The underlying mechanisms which associate with these diseases have not yet been elucidated. However, as the immune response to hsp60 of Clamydia pneumoniae has been implicated in the pathogenesis of atherosclerosis , it is possible to assume that the antibodies against hsp60 derived from periodontopathic bacteria have similar effects on the process of vascular endothelial injury.
To test the hypothesis that molecular mimicry between human hsp60 and bacterial GroEL might be associated with periodontal disease, we first constructed recombinant plasmids in order to produce histidine-tagged human hsp60 and P. gingivalis GroEL. Then, by using these recombinant proteins as antigens we analysed the presence of antibodies against human and bacterial hsp60 in sera and tissue extracts of periodontitis patients. Finally, cross-reactivity of these antibodies was determined.
SUBJECTS and METHODS
Sera and gingival tissue extracts
Twenty-three subjects with moderate to advanced periodontitis (mean age 43·9 years, male:female 10:13), referred to the Periodontal Clinic of Niigata University Dental Hospital, took part in this study. Sera were obtained at the initial examination whereas gingival tissue samples were obtained from 10 patients at the time of periodontal surgery. Approximately 100 mg of gingival tissue were homogenized in 1 ml of PBS and tissue debris was removed by centrifugation at 200 g for 15 min. The resultant soluble extract was used for the determination of IgG concentration as described previously . All patients who had no systemic disorders identified by the usual medical history were classified as adult periodontitis or early onset periodontitis. The clinical characteristics of the sampling sites are as shown in Table 1. As a control, periodontally healthy subjects (mean age 32·5 years, male:female 15:3) were also included in this study. Written informed consent was obtained from all patients and control subjects, in line with the Helsinki declaration, before inclusion in the study.
Table 1. Clinical profile of sampling sites
Bone loss (%)
Data shown were from the most severe site of each tooth determined at the time of sampling.
GI, Gingival index ; PD, pocket depth; AL, attachment level; BOP, bleeding on probing.
pTrc99A-HSP60 containing a DNA fragment coding human hsp60 was kindly provided by I. Hirai (Sapporo Medical College, Sapporo, Japan). Details of this recombinant plasmid have been described elsewhere . From this plasmid, the cDNA insert was released with NcoI and HindIII and subcloned into the prokaryotic expression vector pRSET B (Invitrogen Co., San Diego, CA) digested with the same enzymes.
For P. gingivalis GroEL, an AvaI-HindIII fragment of the DNA originally described by Hotokezaka et al.  was modified so as to provide a translational start codon with a NcoI site in order to be cloned into multiple cloning sites in pRSET B. This modification was carried out by the addition of an oligonucleotide with NcoI and AvaI restriction sites which is a polymerase chain reaction (PCR)-amplified fragment using primers with these restriction sites and P. gingivalis GroEL gene (1–120) as a template. Nucleotide sequences of constructed plasmids were confirmed by automated DNA sequencing (Pharmacia Biotech, Uppsala Sweden).
Expression and purification of recombinant hsp60 and P. gingivalis GroEL
The recombinant plasmids were transformed into Escherichia coli BL21(DE3) pLysS and expressed as polyhistidine-tagged proteins with 1 m m isopropyl-βd-thiogalactopyranoside (IPTG; G ibco BRL, Life Technologies, Rockville, MD) induction for 7 h. The cell pellet prepared from 1 l of bacterial culture was resuspended in PBS containing 60 m m imidazole and 8 m urea and lysed. The cell lysates were applied to the Hi-Trap chelating column (Pharmacia) which had been precharged with nickel ions. After washing with 20 m m sodium phosphate buffer pH 7·4 containing 60 m m imidazole, samples were eluted with 20 m m sodium phosphate buffer pH 7·4 containing 500 m m imidazole. Aliquots of 1 ml fraction were collected and analysed by 10% SDS–PAGE and subsequent coomassie brilliant blue staining.
Confirmation of specificities of recombinant proteins
In order to confirm specificities of recombinant proteins, purified fractions were subjected to 10% SDS–PAGE and transferred onto nitrocellulose membrane using Trans-Blot SD (BioRad Labs, Hercules, CA) for Western blotting. The membranes were reacted with monoclonal anti-human hsp60 (LK-2; Stressgen Biotechnologies Co., Victoria, Canada) cross-reactive with bacterial GroEL, monoclonal anti-human hsp60 (LK-1; Stressgen) only reactive with mammalian hsp60  and monoclonal anti-pentahistidine antibodies (Qiagen GmbH, Hilden, Germany) followed by biotinylated horse anti-mouse IgG (Vector, Burlingame, CA) and finally with avidin-biotin-immunoperoxidase (Vector). The peroxidase was developed using 0·005% 3,3′-diaminobenzidine in Tris–HCl buffer pH 7·2 containing 0·01% hydrogen peroxide.
Western blot analysis of sera and gingival tissue extracts
After SDS–PAGE of 1 μg of recombinant human hsp60 and P. gingivalis GroEL, the gels were transferred onto nitrocellulose membrane. After incubation with blocking solution (5% skim milk in PBS), the membranes were incubated with either diluted serum (1:100) or tissue homogenates (equivalent to 50 μg/ml IgG) followed by horseradish peroxidase (HRP)-conjugated goat anti-human IgG (Sigma Chemical Co., St Louis, MO). The peroxidase was developed using 0·005% 3,3′-diaminobenzidine in Tris–HCl buffer pH 7·2 containing 0·01% hydrogen peroxide. Two examiners (K.T. and K.Y.) judged the positivity and there was no discrepancy between the examiners.
ELISA for anti-human hsp60 and anti-P. gingivalis GroEL antibody levels in sera and gingival tissue extracts
For ELISA, microtitre plates were coated with hsp60 and P. gingivalis GroEL (100 μl/well; 5 μg/ml in 0·05 m carbonate buffer, pH 9·6) overnight at 4°C. After washing three times with PBS containing 0·05% Tween-20 (PBS–T; pH 7·4), non-specific binding sites were blocked with PBS–T containing 1% bovine serum albumin (BSA) for 1·5 h at room temperature. The plates were washed with PBS–T and 100 μl of test samples diluted with blocking reagent (1:100) were added and incubated for 1 h at room temperature. After washing three times with PBS–T, HRP-conjugated goat anti-human IgG (Sigma) was added and incubated further for 1 h at room temperature. The peroxidase was developed using 0·005% 3,3′-diaminobenzidine in Tris–HCl buffer pH 7·2 containing 0·01% hydrogen peroxide. Colour development was stopped by the addition of 4 m sulphuric acid and absorbance was read at a wavelength of 490 nm using an automated ELISA reader (Labsystems Oy, Helsinki, Finland) and analysed using GENESIS LITE software (Labsystems).
Determination of cross-reactivity of serum antibody to hsp60
Cross-reactivity of the serum antibody to hsp60 and P. gingivalis GroEL was determined by the technique of Hirata et al. . After 36 μg of hsp60 and P. gingivalis GroEL were blotted onto nitrocellulose membrane using a BioRad Biodot SF apparatus (BioRad), the membranes were incubated with test sera for 1 h at room temperature. After nitrocellulose membranes were washed extensively with PBS, bound antibodies were eluted in 0·6 ml of glycine–HCl pH 2·5 supplemented with 0·5% Tween 20 and 100 μg/ml of BSA. Affinity-purified antibodies thus obtained were transferred to a second tube, restored to physiologic pH immediately, and dialysed against PBS for 18 h.
After adjusting the volume of absorbed sera and gingival tissue extracts to be equivalent to affinity-purified and eluted antibodies, the samples were then subjected to Western blot analysis.
Serum antibody levels between two groups were evaluated using the Mann–Whitney U-test. The χ2 test was used to compare the frequency of antibody-positive samples in sera between periodontitis patients and healthy control subjects (2 × 2 contingency table).
Production and determination of the specificities of recombinant proteins
Protein profiles, on SDS–PAGE, of recombinant human hsp60 and P. gingivalis GroEL are shown in Fig. 1. Although a few very faint bands were found after affinity purification by a nickel chelate column, both human hsp60 and P. gingivalis GroEL were purified almost homogeneously. The molecular weight of the recombinant proteins coincided with the expected sizes from the nucleotide sequences of the genes. Western blot analysis demonstrated that human hsp60 reacted with both LK-1 and LK-2 whereas P. gingivalis GroEL reacted only with LK-2. Furthermore, the anti-pentahistidine antibody reacted with both the recombinant proteins ( Fig. 2). Similarity of the amino acid sequence between human hsp 60 and P. gingivalis GroEL or A. actinomycetemcomitans GroEL, and epitopes recognized by LK-1 and LK-2 are shown in Table 2.
Table 2. Sequence of human hsp60 identified by LK-1 and LK-2 and its similarity to Porphyromonas gingivalis GroEL and Actinobacillus actinomycetemcomitans GroEL
LK-1 and LK-2 recognize epitopes located between amino acid positions 383 and 447, and 383 and 419, respectively. Identical amino acid residues between human and bacteria are indicated (*). Sequence of human hsp60 is obtained from database (GenBank; accession number M34664).
Western blot analysis of the antibodies to human hsp60 and P. gingivalis GroEL in periodontally healthy subjects
As shown in Fig. 3, serum of one subject (h1) demonstrated moderate reactivity to hsp60 and those of four subjects (h3, h9, h14 and h17) showed very weak but obvious reactivities to hsp60. However, sera of 13 out of 18 subjects did not contain the antibodies reactive with hsp60. The number of subjects whose sera had reactivity to P. gingivalis GroEL was slightly higher than that for hsp60 (8/18; h1, h3, h4, h8, h9, h11, h13 and h17). The reactivity was mostly very weak except the serum of subject h4. Those who had antibody to human hsp60 except h14 were positive for P. gingivalis GroEL.
Comparison of anti-hsp60 antibodies between sera and gingival tissue extract in periodontitis patients
The results of Western blot analysis of sera and gingival tissue extracts derived from periodontitis patients are shown in Fig. 4. The frequency of patients who had serum antibody to human hsp60 was 61%, which was significantly higher than that of control subjects with apparently stronger reactions seen in the blots. On the other hand, most of the patients (91%) except patients p7 and p9 were seropositive for P. gingivalis GroEL and the frequency of positive samples was also significantly higher than that of control subjects ( Table 3). Their reactivities were also stronger than those of healthy subjects. In the gingival tissue extracts seven out of 10 patients demonstrated a positive response to human hsp60 and two of those demonstrated strong positivity (patients p2 and p4). Three patients showed a positive response to human hsp60 only in tissue extracts (patients p1, p7 and p10). The anti-P. gingivalis GroEL antibody was detected in all samples of gingival tissue extracts. In general, higher reactivity of antibody to P. gingivalis GroEL was found in tissue extracts than in sera. Although similar patterns of reactivity to P. gingivalis GroEL and hsp60 were found between gingival tissue extracts and sera in each patient, it is notable that the reactivities of sera and gingival tissue extracts to the recombinant proteins did not necessarily correlate with each other in some patients. For example, whereas the reactivity to P. gingivalis GroEL was weaker in gingival tissue extract than in serum, that to hsp60 was stronger in gingival tissue extract in patient p1. This trend was also observed in patient p6.
Table 3. Distribution of individuals with anti-human hsp60 and anti-Porphyromonas gingivalis GroEL antibodies in sera between periodontitis patients and healthy control subjects
No. (%) of positive subjects
Subject group (n)
Anti-P. gingivalis GroEL
Significantly higher frequencies in periodontitis group than in control are noted (*P < 0·05; **P < 0·002).
Quantitative assessment of antibodies in sera
In order to evaluate the antibodies quantitatively, the test sera were analysed by ELISA. As shown in Fig. 5, sera from patients contained both antibodies to human hsp60 and P. gingivalis GroEL at levels significantly higher than those of sera from control subjects. The median values of antibody levels to hsp60 and P. gingivalis GroEL were 0·25 (range 0·16–0·35) and 0·20 (range 0·13–0·34), respectively, in control subjects, and 0·35 (range 0·19–0·91) and 0·26 (range 0·16–0·66), respectively, in periodontitis patients. Due to the limited amount of samples, gingival tissue extracts could not be analysed by ELISA.
Determination of cross-reactivity of antibodies in sera
In order to analyse cross-reactivity of the antibody against human hsp60 and P. gingivalis GroEL, the serum samples demonstrating strong positivity for both hsp60 and P. gingivalis GroEL were used to determine cross-reactivity. As shown in Fig. 6, affinity-purified antibodies to human hsp60 not only retained their reactivity to human hsp60 after purification but also reacted with P. gingivalis GroEL ( Fig. 6a). Similarly, affinity-purified antibodies to P. gingivalis GroEL demonstrated reactivity to human hsp60 ( Fig. 6b). This cross-reactivity of antibodies was found in all samples tested.
The mechanisms of tissue destruction in chronic inflammatory periodontal disease have not been clearly defined, though several inflammatory mediators such as proinflammatory cytokines and prostaglandins are considered to be involved . The immune responses to self-antigens such as collagen type I, a major component of the periodontium, have also been considered as one of the pathogenic pathways due to the fact that high titres of anti-collagen type I antibody are found in the sera  and that collagen type I-specific T cell clones can be identified in the inflamed gingival tissues of periodontitis patients . Although the mechanisms that induce immune response to self components are not fully elucidated, molecular mimicry has drawn much attention to explain the linkage between infection with bacteria and subsequent autoimmune mechanisms .
Heat shock proteins, particularly the hsp60 family of proteins, are thought to play important roles in the causal relationship between microbial infections and autoimmunity because of conservation of the amino acid sequence during evolution and their strong immunogenicity. To date, there have been a number of reports on the role of hsp and autoimmune diseases . Recently, representative periodontopathic bacteria such as P. gingivalis[6,7], Bacteroides forsythus and A. actinomycetemcomitans were shown to express hsp60 family proteins which are homologous with E. coli GroEL. Serum antibody to the periodontopathic bacteria-derived hsp60 was frequently detected in periodontitis patients [7,8]. Furthermore, it has been reported that, using the MoAb reactive with both bacterial and mammalian hsp60 (LK-2), it could be detected in host cells in the inflamed gingival tissues of periodontitis patients . However, whether serum antibody to P. gingivalis GroEL is also reactive to autologous hsp60 remained unclear. In the present study recombinant proteins were used as antigens in order to demonstrate the presence of cross-reactive antibody. These recombinant proteins were expressed with a polyhistidine tail using the same E. coli host. This minimal modification of six amino acids is known to be non-immunogenic and allows for a simple one-step purification of the recombinant proteins by metal affinity chromatography. These purified proteins reacted with the respective MoAbs; human hsp60 reacted with LK-1 while P. gingivalis GroEL did not react with LK-1 but with LK-2, indicating that B cell epitopes, at least to these MoAbs, were retained after the purification procedure.
Sera from most patients and a few control subjects demonstrated immunoreactivity to P. gingivalis GroEL in Western blots. The difference in frequency between patients and healthy subjects may be attributable to the difference in the degree of infection by P. gingivalis. Although P. gingivalis has strong pathogenic characteristics and has been reported to increase in periodontitis patients, it can also be found in periodontally healthy subjects .
Virtually all the gingival tissue extracts contained antibodies reactive with P. gingivalis GroEL and demonstrated higher binding reactivity than the corresponding serum in most patients. As it is impossible to exclude contamination of gingival tissue extracts by serum components, IgG levels in the gingival tissue extracts were determined and adjusted to those of serum prior to Western blotting. If the antibodies contained in gingival tissue extracts are all derived from sera, their reactivities to antigens should be the same. However, above findings indicate that the concentration of IgG antibodies to P. gingivalis GroEL in gingival tissue extracts was higher than that in the serum of each patient, hence it is possible that anti-P. gingivalis GroEL antibody was being produced locally.
The frequency of positive samples containing anti-human hsp 60 antibodies in serum was lower than that of P. gingivalis GroEL in periodontitis patients. The frequency of positive samples of gingival tissue extracts for both antigens was much higher than autologous serum samples. Moreover, although relative reactivities to P. gingivalis GroEL and hsp60 in gingival tissue were similar to that in serum in each patient, there were obvious increases in reactivity to hsp60, despite being decreased to P. gingivalis GroEL in two patients. It can be concluded therefore that the anti-human hsp60 antibodies were being produced in the periodontitis lesion, rather than in other inflammatory lesions elsewhere in the body.
Although the precise mechanisms of the elevated anti-human hsp60 antibody in periodontitis patients remain to be elucidated, those who had high titres of antibody may have been sensitized with bacteria other than periodontopathic bacteria. In this context, Hasan et al. reported that the T cells from patients with recurrent oral ulceration (ROU) showed significantly greater proliferative responses to both mycobacterial and human hsp65–60 . Although we could not find the current disease or history of ROU in any of the subjects in the present study, we can not exclude the possibility that some subjects had been sensitized to hsp by previous infection.
To confirm the specificity of anti-hsp60 antibodies in patients’ sera, antibodies were affinity-purified with either P. gingivalis GroEL or human hsp60 and then tested for binding reactivity with two different species of hsp60. Antibodies from five patients examined showed equal binding reactivity with human hsp60 and P. gingivalis GroEL. These results support the concept of molecular mimicry between the human and bacterial antigens which may be a factor in the immune response seen in periodontal disease. In addition to molecular mimicry, there are a number of other factors which may stimulate autoreactive antibody in periodontitis lesions . Those are: (i) Gram-negative bacteria-derived lipopolysaccharides are polyclonal B cell activators (PBA) which result in the accumulation of large amounts of antigen-specific B cells including autoreactive B cells in chronically inflamed tissues; (ii) expression of costimulatory molecules is up-regulated in the lesions. Antigen-specific T cell-dependent B cell activation requires signal transduction via costimulatory molecules in addition to the ligation of T cell receptor with MHC–antigen peptide complex. We have recently demonstrated that large numbers of CD80- and CD86-expressing cells, most of which were B cells, were identified in periodontitis lesions together with an increased number of CD40 ligand-positive T cells . There is also evidence to suggest that inflammation results in the up-regulation of hsp [27,28] and a number of hsp60-expressing cells can be found in inflammatory cell infiltration and epithelium in inflamed gingival tissues (unpublished observation). It is hypothesized that as a result of the up-regulation of hsp60, T cells specific for self hsp60 might be activated as part of the normal inflammatory process.
The findings presented in the present study are the first to show the presence of cross-reactive antibodies to P. gingivalis GroEL (an exogenous antigen) and human hsp60 (a self antigen) in periodontitis lesions. Clearly further studies to elucidate the overall mechanisms by which periodontopathic bacteria stimulate autoimmune responses, particularly T cell responses, and the pathogenic significance of anti-hsp60 antibodies in periodontitis are necessary. However, considering the expression of hsp60 in the inflammatory lesion of periodontitis and the presence of antibodies to hsp60, anti-hsp60 antibodies may not only exert pathogenic effects at the site of inflammation and contribute to the progression of periodontal destruction through complement fixation and T cell cytotoxicity dependent on antibody, but also be involved in the increased risk factors for coronary heart disease in periodontitis patients.
The authors would like to thank I. Hirai (Department of Pathology, Sapporo Medical University, Japan) for kindly providing pTrc99A-HSP60. The authors are also grateful to G. J. Seymour (Oral Biology and Pathology, Department of Dentistry, The University of Queensland, Australia) for critical reading of this manuscript. This work was supported by grants from the Ministry of Education, Science Sports and Culture of Japan (10470458, 10357020, 10307054) and Special grant for Development of Advanced Medical Technology.