E. Gemmell, Immunopathology Laboratory, Oral Biology and Pathology, School of Dentistry, University of Queensland, Brisbane 4072, Australia. E-mail: firstname.lastname@example.org
T cell cytokine profiles and specific serum antibody levels in five groups of BALB/c mice immunized with saline alone, viable Fusobacterium nucleatum ATCC 25586, viable Porphyromonas gingivalis ATCC 33277, F. nucleatum followed by P. gingivalis and P. gingivalis followed by F. nucleatum were determined. Splenic CD4 and CD8 cells were examined for intracytoplasmic interleukin (IL)-4, interferon (IFN)-gamma and IL-10 by dual colour flow cytometry and the levels of serum anti-F. nucleatum and anti-P. gingivalis antibodies determined by an ELISA. Both Th1 and Th2 responses were demonstrated by all groups, and while there were slightly lower percentages of cytokine positive T cells in mice injected with F. nucleatum alone compared with the other groups immunized with bacteria, F. nucleatum had no effect on the T cell production of cytokines induced by P. gingivalis in the two groups immunized with both organisms. However, the percentages of cytokine positive CD8 cells were generally significantly higher than those of the CD4 cells. Mice immunized with F. nucleatum alone had high levels of serum anti-F. nucleatum antibodies with very low levels of P. gingivalis antibodies, whereas mice injected with P. gingivalis alone produced anti-P. gingivalis antibodies predominantly. Although the levels of anti-F. nucleatum antibodies in mice injected with F. nucleatum followed by P. gingivalis were the same as in mice immunized with F. nucleatum alone, antibody levels to P. gingivalis were very low. In contrast, mice injected with P. gingivalis followed by F. nucleatum produced equal levels of both anti-P. gingivalis and anti-F. nucleatum antibodies, although at lower levels than the other three groups immunized with bacteria, respectively. Anti-Actinobacillus actinomycetemcomitans, Bacteroides forsythus and Prevotella intermedia serum antibody levels were also determined and found to be negligible. In conclusion, F. nucleatum immunization does not affect the splenic T cell cytokine response to P. gingivalis. However, F. nucleatum immunization prior to that of P. gingivalis almost completely inhibited the production of anti-P. gingivalis antibodies while P. gingivalis injection before F. nucleatum demonstrated a partial inhibitory effect by P. gingivalis on antibody production to F. nucleatum. The significance of these results with respect to human periodontal disease is difficult to determine. However, they may explain in part differing responses to P. gingivalis in different individuals who may or may not have had prior exposure to F. nucleatum. Finally, the results suggested that P. gingivalis and F. nucleatum do not induce the production of cross-reactive antibodies to other oral microorganisms.
Periodontal disease results from the inflammatory response to bacteria in dental plaque (reviewed in ). Although there are well over 300 different bacterial species in the plaque, progression to periodontitis depends not on bacterial load, but on the presence of specific periodontopathic bacteria. The major pathogens identified at the 1996 World Workshop in Periodontics as causative agents are Porphyromonas gingivalis, Actinobacillus actinomycetemcomitans and Bacteroides forsythus. P. gingivalis has been found in 15% of subjects in an Australian population, the prevalence increasing with increasing pocket depth . In another study, Slots and Ting  found that 40–100% of adult periodontitis patients were positive for P. gingivalis. This high occurrence, together with the pathogenic potential of P. gingivalis, has made it a major pathogen of adult periodontitis .
Immunohistological studies have established that a T cell/macrophage lesion identical to a delayed hypersensitivity reaction  occurs within 4–8 days of plaque accumulation in an experimental gingivitis study . T cells with a Th1 cytokine profile may therefore be the major mediator in the early/stable lesion. On the other hand, B cells and plasma cells have been shown to predominate in the advanced/progressive lesion [8,9] suggesting a role for Th2 cells. A number of studies have attempted to delineate the Th1/Th2 profile in periodontal disease. However, different T cell subsets may predominate at different phases of disease and the inability to determine disease activity clinically in humans is a major limitation in these studies.
B cells and plasma cells produce and secrete immunoglobulins which protect the host by various methods including prevention of bacterial adherence, inactivation of bacterial toxins and by acting as opsonins for phagocytosis by neutrophils. The inability of specific antibodies to eliminate the causative organisms of periodontal disease could be due to a number of factors, including poor antigenicity of the virulence determinants and elicitation of antibodies with poor antibacterial properties . Morinushi et al. have also demonstrated that P. gingivalis may inhibit the production of specific antibodies. As these responses are regulated by immunoregulatory genes, it may be that antibody responses are protective in one individual but not another .
Although a periodontopathic organism is essential for periodontal disease progression to occur, interactions between the many species of oral bacteria must also be considered to be important factors in the development of periodontal disease. While animal studies have demonstrated the pathogenicity of mono-infections of periodontopathogens such as P. gingivalis[13–17], in recent years, reports have been published on the effects of mixed microbial infections including P. gingivalis and A. actinomycetemcomitans, P. gingivalis and F. nucleatum and P. gingivalis and B. forsythus which resulted generally in increased levels of pathogenicity with a synergistic effect observed in the humoral and cellular host responses. In the human, F. nucleatum colonizes the plaque prior to P. gingivalis and high levels of F. nucleatum have been demonstrated in association with P. gingivalis as well as other bacteria associated with periodontal disease, such as B. forsythus, Prevotella intermedia and Eikenella corrodens. F. nucleatum subspecies nucleatum is isolated most frequently from patients with adult periodontal disease and is the one most associated with P. gingivalis. The aim of this study therefore was to examine the effect of this subspecies of F. nucleatum on the splenic T cell cytokine and serum specific antibody responses to P. gingivalis.
MATERIALS AND METHODS
P. gingivalis ATCC 33277 and F. nucleatum ATCC 25586 were used in this study. The organisms were cultured anaerobically, as described by Bird and Seymour . Briefly, the organisms were grown on Wilkins Chalgrens agar (WCA) prepared from Wilkins Chalgrens broth (33 g/l) (BioMérieux Vitek, Hazelwood, MD, USA) to which was added agar (10 g/l) and 5% defibrinated horse blood. The plates were incubated for 2–4 days at 37°C in an atmosphere of 80% N2, 10% CO2 and 10% H2 in an anaerobic cabinet (Coy Laboratory Products Inc, Grass Lake, Michigan, USA). The purity of the cultures was monitored by colonial morphology and identification confirmed using API-ZYM . All manipulations were carried out in the anaerobic cabinet. Bacteria were harvested from the WCA plates using swabs moistened in reduced normal saline and then suspended in saline. Bacterial numbers for injection were determined by counting in a Helber bacterial counter chamber. The bacteria were suspended in saline in the anaerobic cabinet and then transported in an anaerobic state in tubes with injection caps to the animals to be injected.
This project was approved by the institutional animal ethics review committee. BALB/c female mice (6–8 weeks old) were obtained from the University of Queensland Central Animal Breeding House. The immunization protocol has been described previously by Bird et al.. Twenty-nine mice were divided into five groups (six per group, with five mice in the control group). The sham-immunized group of mice was injected with saline alone once a week for 4 weeks (days 0, 7, 14 and 21). The second group of mice received intraperitoneal injections of 1 × 108 viable F.nucleatum in saline once a week for 4 weeks as described for the control group. Group 3 received 1 × 108 viable P. gingivalis organisms in saline, as for Group 2. Group 4 were injected with 1 × 108 viable F. nucleatum in saline for the first 2 weeks followed by 1 × 108 viable P. gingivalis organisms in saline in weeks 3 and 4. Group 5 received the reverse of group 4 with injections of P. gingivalis for the first 2 weeks followed by F. nucleatum in weeks 3 and 4. All mice were injected at the same time using the individual protocols for each group.
One week after the final immunizations, the mice from each group were anaesthetized lightly with halothane/O2 and blood samples collected immediately by heart puncture, after which the mice were killed by cervical dislocation. Serum was separated for the determination of specific antibody levels. The spleens were removed, worked through cell strainers (Falcon, Becton Dickinson and Company, Franklin Lakes, NJ, USA) and the resulting suspensions washed and centrifuged on Ficoll-Paque gradients to obtain mononuclear cell suspensions. The T cells were then stained for intracytoplasmic cytokines and analysed by two-colour flow cytometry as described below.
Weight loss and gain and general health parameters such as subdued behaviour and ruffled hair were monitored throughout the study.
Flow cytometric analysis
The percentage of CD4 and CD8 cells extracted from the spleens staining positive for intracytoplasmic IL-4, IFN-gamma and IL-10 were determined as described previously [26–28]. Briefly, surface membrane staining of CD4 and CD8 cells was achieved using phycoerythrin (PE) conjugated rat antimouse CD4 or CD8 (PharMingen, San Diego, CA, USA), followed by fixation of these cells in paraformaldehyde, permeabilization using proteinase K and then incubation with fluorescein isothiocyanate (FITC) conjugated rat antimouse IL-4, IFN-gamma or IL-10 (PharMingen). For the assessment of non-specific binding of the rat antibodies to the mouse cell surface antigens, PE-and FITC-conjugated specific rat Ig isotypes (PharMingen) were used in place of the CD4 or CD8 antibodies and the anticytokine antibodies. Ten thousand stained cells from each sample were analysed using dual colour flow cytometry on a FACSCalibur (Becton Dickinson, Mountain View, CA, USA) and the percentages of CD4 and CD8 cells which were positive for IL-4, IFN-gamma and IL-10 determined.
Detection of anti-F. nucleatum and P. gingivalis antibodies
Serum samples were assayed for the presence of anti-F. nucleatum and P. gingivalis antibodies using an ELISA technique described by Bird et al.[25,27,28]. Briefly, the protein concentrations of both organisms were determined (BCA Protein Assay Kit, Pierce Rockford, IL, USA) and the optimum coating concentration of F. nucleatum and P. gingivalis ascertained to be 5·0 μg/ml protein. Whole cells were coated onto 96-well high-binding plates (Maxisorb Immunoplates, Nunc, Roskilde, Denmark) using the optimum coating concentration ensuring that equivalent amounts of protein were coated onto each plate to allow comparisons between plates. After blocking nonspecific sites with 1% bovine serum albumin (Commonwealth Serum Laboratories, Melbourne, Australia) in PBS-Tween 20 (0·05%), diluted serum samples were added followed by peroxidase-conjugated sheep antimouse IgG (detects all 4 IgG subclasses) (The Binding Site Limited, Birmingham, UK) at a dilution of 1/2000. The substrate containing 0·0075% H2O2 and 2·5 mM O’Tolidine (Eastman Kodak, Rochester, NY, USA) was then added and the blue colour reaction stopped after 10 min with 3 M HCl. The optical density of the wells was read at an absorbance of 450 and 655 nm on a BIO-RAD Microplate reader Model 3550. Negative controls consisted of substituting PBS in place of the serum samples. Levels of anti-F. nucleatum and P. gingivalis IgG antibodies in the serum samples were determined from a standard curve of dilutions of a known concentration of normal mouse IgG (Caltag Laboratories, Burlingame, CA, USA) which were coated onto each plate. Dilutions of IgG (0·4–50 ng/ml) were coated on to each plate at the same time as the bacterial coating. The ELISA procedure was then followed exactly as described above with the exception of the addition of PBS in place of the mouse serum samples. The levels of specific antibodies were expressed as being equivalent to μg/ml IgG.
To determine the ability of F. nucleatum and P. gingivalis to induce the production of cross-reactive antibodies, levels of anti-A. actinomycetemcomitans, B. forsythus and P. intermedia were also determined. A. actinomycetemcomitans (Y4) and P. intermedia ATCC 25611 were grown on WCA plates as described above for F. nucleatum and P. gingivalis. B. forsythus ATCC 43037 was grown on Trypticase Soy agar with a disk containing N-acetylmuramic acid and co-cultured with a streak of Staphylococcus aureus, as described by Bird et al.. The protein concentrations of the three bacterial suspensions were determined and the organisms coated onto plates as for P. gingivalis and F. nucleatum at 5·0 μg/ml protein so that again, comparisons of the levels of specific antibodies to all five bacteria could be made and the ELISAs carried out as described above.
Multivariate analysis of variance using the general linear model was used to test for differences in the expression of each cytokine by CD4 and CD8 cells within each group of mice and comparisons were also made between the five groups. The same procedure was carried out for the analysis of levels of specific antibodies. Selected pairs of groups were then tested for significance using Student’s t-test. A significance level of 0·03 was determined to reduce the probability of significant differences occurring by chance. The Minitab statistical package (Minitab Inc., State College, PA, USA) was used to perform the analyses.
Splenic T cell cytokine responses
The mean percentages of IL-4+ IFN-gamma+ and IL-10+ CD8 cells were increased significantly in comparison with the corresponding cytokine positive CD4 cells in Groups 3 (P. gingivalis) (P = 0·000, 0·000 and 0·002, respectively), and 4 (F. nucleatum followed by P. gingivalis) (P = 0·020, 0·006 and 0·002, respectively) (Figs 1–3). There was also a significant increase in Group 2 (F. nucleatum) (P = 0·004) in the case of IL-4+ CD8 cells and Group 5 (P. gingivalis followed by F. nucleatum) for IFN-gamma+ CD8 cells (P = 0·004) (Figs 1 and 2).
The percentages of IL-4+ CD4 and CD8 cells were increased significantly in Groups 4 (P = 0·027 and 0·000, respectively) and 5 (P = 0·008 and 0·004, respectively) in comparison with control Group 1. The percentage of IL-4+ CD8 cells was also increased significantly in Group 2 (P = 0·005) and Group 3 (P = 0·000) (Fig. 1).
The percentage of IFN-gamma+ CD8 cells was increased significantly in groups 3 (P = 0·000), 4 (P = 0·000) and 5 (P = 0·01) in comparison with control Group 1. The percentage of IFN-gamma+ CD8 cells was also significantly higher in Group 5 compared with Group 2 (P = 0·027) (Fig. 2).
The percentage of IL-10+ CD4 cells was increased significantly in Group 5 (P = 0·029) and IL-10+ CD8 cells in Groups 3 (P = 0·007) and 4 (P = 0·004) compared with Group 1 (Fig. 3).
Serum-specific antibody responses
Anti-F. nucleatum antibody levels were significantly increased in Groups 2, 4 and 5 compared with control Group 1 (P = 0·000). Levels in each of Groups 2 and 4 were increased in comparison with Group 5 (P = 0·000) (Fig. 4).
Anti-P.ggingivalis antibody levels were significantly increased in Groups 3, 4 and 5 compared with control Group 1 (P = 0·000) and in comparison with Group 2 (P = 0·000). Levels in Group 3 were increased in comparison with Groups 4 and 5 (P = 0·000) and levels in Group 5 were increased in comparison with Group 4 (P = 0·000) (Fig. 4).
The levels of anti A.actinomycetemcomitans, B. forsythus and P. intermedia antibodies in the serum samples of mice immunized with bacteria were similar to or slightly increased in comparison with the control group. Anti-F. nucleatum antibody levels in Groups 2 and 4 were significantly increased compared with anti-P. gingivalis, A. actinomycetemcomitans, B. forsythus and P. intermedia levels and in Group 5 compared with all except anti-P. gingivalis antibody levels (P = 0·000). Anti-P. gingivalis antibody levels were significantly increased in Group 3 compared with anti-F. nucleatum, A. actinomycetemcomitans, B. forsythus and P. intermedia levels and in Group 5 compared with all except anti-F. nucleatum antibody levels (P = 0·000) (Fig. 4).
The results of the present study have demonstrated that in general the percentages of cytokine-positive CD8 cells in the spleens of immunized mice were significantly higher than those of the CD4 cells, as has been reported previously [27,28]. Also, while the percentages of cytokine positive CD8 cells were increased for the most part compared with the control group, only the percentage of IL-4+ CD4 cells was increased in the two groups immunized with both F. nucleatum and P. gingivalis. In general there was a trend for lower cytokine positive CD8 cells in mice immunized with F. nucleatum alone than in the other groups, whose immunization protocol included injections of P. gingivalis. However, F. nucleatum did not have any effect on the T cell production of cytokines induced by P. gingivalis.
In a recent study, Choi et al. established P. gingivalis-specific T cell clones from the spleens of BALB/c mice immunized with viable P. gingivalis alone or with F. nucleatum followed by P. gingivalis. Interestingly, all the clones were CD4+ T cells and while those established from mice injected with P. gingivalis had a polarized Th1 profile, those immunized with both F. nucleatum and P. gingivalis, had a polarized Th2 cytokine pattern. Proliferation assays also demonstrated that there was considerable cross-reactivity against the two organisms. It was concluded that F. nucleatum provided an immunomodulatory role where T cells initially primed by F. nucleatum antigens may result in the development of Th2 cells. While the strains of bacteria were different from those used in the present study, this does not explain the diverging results. As well, five times the dose of organisms was used in the former protocol, although the immunizing dose of P. gingivalis has been shown to have no effect on the T cell cytokine profiles of immunized mice . T cell clones were not established in the present study, which may be partly responsible for the differences found in the two studies, although based on the splenic T cell profiles of the present study, both CD4 and CD8 specific T cells with a Th1/Th2 profile would possibly be expected.
Although F. nucleatum did not have an effect on the T cell cytokine profiles induced by P. gingivalis in the present study, the levels of anti-P. gingivalis antibodies were affected by F. nucleatum immunization. Mice immunized with F. nucleatum alone had high levels of serum anti-F. nucleatum antibodies with very low levels of P. gingivalis antibodies. The opposite effect was observed in mice injected with P. gingivalis alone with the production of predominantly anti-P. gingivalis antibodies. Although the levels of anti-F. nucleatum antibodies in mice injected with F. nucleatum followed by P. gingivalis were approximately the same as in mice immunized with F. nucleatum alone, levels of antibodies to P. gingivalis were very low. This suggested that F. nucleatum inhibited the production of anti-P. gingivalis antibodies. In contrast, mice injected with P. gingivalis first followed by F. nucleatum produced equal levels of both anti-P. gingivalis and anti-F. nucleatum antibodies, although the levels were significantly lower than in mice injected with P. gingivalis alone, F. nucleatum alone or F. nucleatum followed by P. gingivalis, respectively. The injection of P. gingivalis initially has allowed the production of some anti-P. gingivalis antibodies before the inhibitory effect of F. nucleatum, although P. gingivalis immunization has also had a partial inhibitory effect on anti-F. nucleatum production.
Co-infection with periodontopathic organisms including P. gingivalis and B. forsythus and P. gingivalis and Treponema denticola has been reported to induce a synergistic effect on the size of lesions in a murine model. Primary infection of mice with P. gingivalis and F. nucleatum also resulted in significantly greater lesion size compared with infection with P. gingivalis alone . However, when the ratio of F. nucleatum to P. gingivalis was 1:1 or greater, spreading lesion formation and progression were decreased. This suggested an inhibitory effect on the virulence of P. gingivalis, although infection with F. nucleatum prior to or 1 h after P. gingivalis infection enhanced the ability of P. gingivalis to form large phlegmonous lesions. Following on from this study, Ebersole et al. showed that active immunization of mice with P. gingivalis protected against challenge with P. gingivalis as well as P. gingivalis together with F. nucleatum and this protection in terms of lesion size, correlated with the levels of specific serum IgG antibody. This has also been demonstrated in another study, in which strains of mice demonstrating a faster healing pattern of lesions after challenge with P. gingivalis had higher serum anti-P. gingivalis antibody levels than those strains which had slower healing lesions .
The concept that polyclonal B cell activation may be important in the pathogenesis of periodontal disease was first introduced in the early 1980s [33–35] and Tew et al. suggested that in the lesion, this would result in the expansion of B cell clones and their differentiation into immunoglobulin producing plasma cells. However, the results of the present study indicated that neither P. gingivalis nor F. nucleatum induced the production of cross-reactive antibodies to B. forsythus, P. intermedia or A. actinomycetemcomitans. Chen et al. also showed that a subcutaneous injection of viable P. gingivalis together with A. actinomycetemcomitans resulted in the production of anti-P. gingivalis antibodies in mice injected with P. gingivalis alone with no anti-A. actinomycetemcomitans activity. Similarly, mice injected with A. actinomycetemcomitans alone were negative for antibodies to P. gingivalis, while antibodies to both P. gingivalis and A. actinomycetemcomitans were demonstrated in mice injected with the two organisms. The use of Western blot analysis as well as ELISAs would validate whether or not cross-reactive antibodies were produced. In contrast to these studies, Takahashi et al. demonstrated that P. gingivalis outer membrane antigens could induce higher polyclonal B cell activity in mouse splenocytes as determined by a direct plaque forming assay against sheep red blood cells than T. denticola. It was suggested that regulation of B cell responses to polyclonal B cell activators may lead to the production of auto-antibodies such as antitype Ι and ΙΙΙ collagens leading to a variety of pathological processes in the periodontal lesion. Holt et al. also showed that immunization of the non-human primate Macaca fascicularis with P. gingivalis cell wall/envelope antigens resulted in slightly lower levels of antibodies to P. intermedia but higher levels to Campylobacter rectus compared with anti-P. gingivalis antibodies. A mixed preparation of F. nucleatum, C. rectus and A. viscosus induced similar IgG levels to P. gingivalis, P. intermedia, F. nucleatum and A. viscosus and again much higher levels to C. rectus. Microbiological sampling of the gingival crevicular demonstrated that while P. gingivalis immunization resulted in high specific antibody levels, there was a reduction in the numbers of recoverable P. gingivalis organisms. At the same time, levels of F. nucleatum species decreased although the levels of other bacteria including Prevotella increased and others such as A. actinomycetemcomitans and Capnocytophaga species remained at preligation levels postimmunization, indicating that P. gingivalis can interfere with the progression of specific members of the periodontal microbiota. This report concluded that while a selected group of organisms may be important in the disease process, other members of this microbiota may play a significant role in the events leading to disease progression . Another study that examined the effect of immunization of BALB/c mice initially with viable F. nucleatum followed by P. gingivalis showed that while mice injected with P. gingivalis alone had high levels of anti-P. gingivalis antibody, there was also a rise in the level of anti-F. nucleatum antibody although to only about 20–25% of the level of the former. Mice injected with both organisms had a similar serum IgG profile. The results suggested that there may be cross-reactive antigen stimulation at both the T and B cell level between these two organisms .
In conclusion, the results of this study have shown that F. nucleatum immunization does not affect the splenic T cell cytokine response to P. gingivalis. However, F. nucleatum immunization prior to that of P. gingivalis inhibited almost completely the production of anti-P. gingivalis antibodies while P. gingivalis injection before F. nucleatum demonstrated a partial inhibitory effect by P. gingivalis on the production of antibodies to F. nucleatum. The mechanism of this suppression is unclear. Whether this involves active suppression by suppressor T cells or some other mechanism of antigen-specific non-responsiveness at the level of the T cell or B cells remains to be determined. Finally, the results suggested that P. gingivalis and F. nucleatum do not induce the production of cross-reactive antibodies to other oral microorganisms.
We thank Daria Love for the kind gifts of the type strains of P. gingivalis, F. nucleatum and P. intermedia, Anne Tanner for A. actinomycetemcomitans and Aaron Weinberg for B. forsythus.