Feasibility and therapeutic strategies of vaccines against Porphyromonas gingivalis

Abstract

Periodontitis is a chronic infectious disease that is highly prevalent worldwide and is characterized by inflammation of the gums, and loss of connective tissue and bone support. The Gram-negative anerobic bacterium Porphyromonas gingivalis is generally accepted as the main etiological agent for chronic periodontitis. The objective of this paper is to elucidate the feasibility of achieving protection against periodontitis though immunization against P. gingivalis. Until now, animal studies have showed no complete protection against P. gingivalis. However, current knowledge about P. gingivalis structures could be applicable for further research to develop a successful licensed vaccine and alternative therapeutic strategies. This review reveals that a multicomponent vaccine against P. gingivalis, which includes structures shared among P. gingivalis serotypes, will be feasible to induce broad and complete protection.

Keywords::

• capsule
• fimbriae
• gingipain
• immunization
• periodontitis
• Porphyromonas gingivalis
• vaccine

Figure 1. Schematic composition of the cell envelope of Porphyromonas gingivalis.

A P. gingivalis vaccine might be based on surface antigens: capsule, LPS, fimbriae, OMPs, gingipains and hemagglutinin, of which LPS and gingipains could also be localized within the outer membrane vesicles [21].

LPS: Lipopolysaccharide; OMP: Outer membrane protein.

Figure 2. Recognition of Porphyromonas gingivalis-specific surface antigens, LPS, fimbriae and hemagglutinin by TLRs.

LPS: Lipopolysaccharide; MyD88: Myeloid differentiation primary response gene 88; TLR: Toll-like receptor; TRIF: TIR-domain-containing adapter-inducing IFN-β.

From [41,45,46,61].

Porphyromonas gingivalis & periodontitis

Periodontitis is characterized by inflammation of the gums, and loss of connective tissue and bone support. The infection affects approximately 13–57% of people in different populations, depending on oral hygiene levels and socioeconomic status [1]. The prevalence, disease progression and severity of periodontitis vary worldwide, but it is well known that the disease is mostly observed in adults. In addition, association with adverse pregnancy outcomes, cardiovascular disease, stroke, pulmonary disease and diabetes stress the importance of knowledge about periodontitis [2].

An estimate of the species diversity in the subgingival plaque is proposed at 500 species [3]. Much evidence points to Porphyromonas gingivalis as the major etiological agent for chronic periodontitis in adults, whereas for example, Aggregatibacter actinomycetemcomitans is strongly associated with juvenile periodontitis [4–7]. A common finding in patients with periodontitis is the presence of P. gingivalis-specific antibodies in serum and gingival crevicular fluid [8–10]. Although there are contradictory reports on the correlation between serum IgG level, avidity or subclasses with disease severity in patients with rapidly progressive periodontitis, most studies show the association between P. gingivalis-specific antibodies and disease severity [11]. Whereas the production of antibodies generally indicates the activation of our major host defense mechanism, these antibodies are insufficient to clear P. gingivalis infection. Since it is desirable to prevent P. gingivalis-induced periodontitis, there is a wide interest in developing a vaccine against P. gingivalis.

Whole-cell immunization

Immunization with live-attenuated cells

One of the first experiments using whole-cell immunization against P. gingivalis showed no potential benefit [12]. Per-oral immunization of hamsters with formalin-killed (F-K) whole cells of P. gingivalis strain 381 showed no significant reduction of colonization after oral challenge, measured by the number of cultivable cells in ligature threads tied coronally to the gingival margins of mandibular first molars. This could be explained by the use of a noninvasive P. gingivalis strain, since only cells from invasive P. gingivalis strains are able to alter the course of subsequent infection [13]. P. gingivalis strains were denoted as invasive when migration from the injection site occurred after challenge by subcutaneous injection, and as noninvasive when these bacteria remained localized around the injection site. Challenging mice with subcutaneous injection with the invasive P. gingivalis strain ATCC 53977 (A7A1–28), after intraperitoneal immunization with invasive P. gingivalis ATCC 53977 (A7A1–28) or noninvasive P. gingivalis 381 strain subsequently leads to both localized and widespread infections. By contrast, another study has shown that immunization with invasive as well as noninvasive P. gingivalis strains could protect mice from secondary lesion formation and death after challenge with invasive P. gingivalis in a subcutaneous lesion chamber model [14]. The route of immunization will cause such differences found in achieved protection. Oral or subcutaneous immunization led to mucosal or systemic immune responses, respectively. In general, invasive strains are most appropriate as a whole-cell vaccine.

Inactivated whole-cell vaccines

Immune response induction in a BALB/c mouse model differs when using either F-K or heat-killed (H-K) whole cells from P. gingivalis W50 [15]. Both types of nonviable whole cells elevated P. gingivalis-specific serum IgG and IgM. However, H-K antigen reacted with more antibodies than the F-K antigen. When subcutaneous immunization with F-K and H-K whole cells are compared with subcutaneous immunization with lipopolysaccharide (LPS) or an outer membrane fraction (OMF), both LPS and OMF also elicited P. gingivalis-specific serum IgG and IgM, but in a lower content than whole cells. Even though LPS, OMF, F-K and H-K whole-cell immunization elicits an antibody production, animals were still partly protected from subcutaneously induced lesion formation like purulent phlegmonous abscesses or localized subcutaneous abscesses, as observed by reductions in lesion size compared with homologous P. gingivalis strains [15]. Salivary antibodies are possibly also required to induce protection, which were not achieved by subcutaneous and/or intramuscular immunization.

The elevation of P. gingivalis-specific serum antibodies by intramuscular or subcutaneous immunization with F-K whole cells from P. gingivalis has also been demonstrated in Macaca fascicularis, a nonhuman primate [16–19]. Although induction of serum IgG and inhibition of P. gingivalis colonization after immunization and ligature-induced periodontitis was shown, these results indicate that nonviable cells could elicit P. gingivalis-specific antibody production in serum. Nevertheless, it was insufficient to induce complete protection [20].

Immunization with P. gingivalis-specific antigens

Since many research groups have studied P. gingivalis over recent years, several structures of P. gingivalis are well defined and could be useful as a vaccine target or as an adjuvant to boost the immune response. As visualized in Figure 1, structures of P. gingivalis, like the capsule, LPS, fimbriae, outer membrane proteins (OMPs), gingipains and hemagglutinin, are localized at the cell surface. Most studies on vaccines against P. gingivalis have focused on these P. gingivalis-specific surface antigens (Table 1).

Capsule

The presence of a capsule on various P. gingivalis strains has been confirmed by electron microscopy [21]. The capsule is represented by capsular polysaccharide (CPS)/K antigens [22,23]. Many studies show evidence or indications that antibodies against CPS will protect against infections caused by a wide range of pathogens [24]. However, the extent of the immune response elicited by CPS as part of a vaccine depends on the structure and molecular weight of these components. Since only very high-molecular-weight polysaccharides are immunogenic, glycoconjugates based on low-molecular-weight oligosaccharides of CPS can overcome insufficient immunogenicity. Based on this knowledge, carrier proteins have been implemented in P. gingivalis vaccine development, for example, vaccination of Balb/c mice with an oral vaccine containing fimbriae in combination with cholera toxin (CT) [25]. Vaccination with P. gingivalis CPS and carrier proteins is still a potential field of study. Most pathogenic bacteria display differences between strains with respect to the expression of CPS, resulting in different serotypes or serogroups [26].

Capsule-based serotypes

Seven capsular serotypes, K1 to K7, and a nonencapsulated strain K^- have been described for P. gingivalis(Table 2)[22,23,27]. These K antigens are acidic, negatively charged, thermostable nonsomatic polysaccharides [22,23]. Studies on the prevalence and distribution of serotypes K1 to K6 showed that K-antigen serotypes were present in 45.4% of the tested isolates from patients with P. gingivalis-associated periodontitis. Serotypes K5 and K6 predominate in these isolates and were more prevalent among young people than adults. Microscopically visible encapsulation of P. gingivalis strains, not categorized as K1 to K6, support the presence of additional K-antigen serotypes [28].

Several studies have demonstrated substantial strain heterogeneity among P. gingivalis isolates, while intra-individual heterogeneity is limited [29]. For example, the CPS locus among K1 strains has been reported as being highly conserved, although slight variations do occur [30].

Immunization with CPS

Research on immunization with CPS showed that subcutaneous immunization of mice with either P. gingivalis A7436 H-K whole cells or CPS resulted in IgM and IgG titers, of which the whole-organism-specific IgG was significantly higher after immunization with H-K whole cells. Strikingly, both types of immunization resulted in similar protection against subsequent P. gingivalis-elicited alveolar bone loss by oral challenge [31].

The advantage of CPS in a subunit vaccine has been shown by intraperitoneal immunization of mice with a conjugate vaccine, polysaccharide of P. gingivalis ATCC 53977 and bovine serum albumin. This resulted in P. gingivalis-specific serum antibodies and reduced the severity of infection, as measured by lesion size, after subcutaneous challenge with P. gingivalis ATCC 53977 [32]. A polysaccharide–fimbriae protein conjugate vaccine resulted in a higher serum IgG response and the highest protection against P. gingivalis 53977-induced infection by subcutaneous injection after intraperitoneal immunization of the vaccine in hu-PBL-SCID mice when compared with CPS or fimbriae alone [33]. While all mice immunized with the conjugate vaccine survived a high-dose P. gingivalis 53977 challenge, mice immunized with CPS or fimbriae, and the control groups showed a death rate of one third. In this conjugate vaccine, CPS was derived from P. gingivalis 53977 and was conjugated to fimbriae from P. gingivalis 381 [33]. Although, these conjugate vaccines could not prevent infection, a significant reduction in disease severity was observed. Application of CPS should be further tested in oral or nasal vaccines to test its ability to induce mucosal immune responses.

New developments in the field of vaccination techniques have led to promising possibilities by application of DNA vaccination. Although human DNA vaccines are still experimental, a DNA vaccine against West Nile virus (WNV) has been licensed for use in horses and has achieved positive results in a Phase I clinical trial [34]. DNA vaccines are made up of genetically engineered plasmids that produce one or more proteins specific for the pathogen. The capsule encoding loci PG1135–PG1142 is, based on comparison with other bacteria, highly conserved among serotypes and could therefore be applied in future DNA vaccine against P. gingivalis, inducing a broad protection [35].

Lipopolysaccharide

The LPS antigens of P. gingivalis represent three serotypes, O1 to O3, which are described for P. gingivalis strains 381, HG1691 and W50 respectively (Table 2)[36]. Clinical studies show that serotypes K1/O3 and K6/O2 are most likely to participate in periodontitis. As a consequence of the serological variation seen for K- and LPS antigens among strains, IgG responses in patients infected with a particular serotype are specific for that particular antigen serotype [37].

Recognition of LPS by the immune system

Several studies on the recognition of Toll-like receptors (TLRs) by P. gingivalis LPS or lipid A show inconsistent results, either demonstrating the stimulation of TLR2 and TLR4, or TLR2 alone [38–41]. These contradictory results can be explained by the high risk of endotoxin contamination during purification by TLR-4-stimulating species, for example, by Gram-negative enterobacteria [40,41]. The difference in recognition between P. gingivalis and enterobacteria might be explained by the P. gingivalis lipid A, which differs substantially from enterobacteria and other bacterial lipid A [42]. Therefore, recognition of P. gingivalis by pattern-recognition receptors (PRRs) of the host immune system occurs predominantly via TLR2 (Figure 2)[41]. This recognition of P. gingivalis will activate phagocytes, such as macrophages, to respond to the pathogens by secreting chemokines and cytokines, and express the costimulatory molecules needed to activate adaptive immunity [43,44]. In addition to LPS or lipid A, fimbriae and hemagglutinin have also been shown to stimulate TLRs (Figure 2)[45,46].

Immunization with LPS

Examination of LPS as a vaccine showed no induction of serum IgG, antigen-reactive lymphocytes and protection against P. gingivalis-induced infection by subcutaneous injection after intraperitoneal immunization of BALB/c mice [47]. These results were in contradiction to immunization with whole cells or a protein–LPS extract from P. gingivalis cells using lithium diiodosalicylate extract. This lithium diiodosalicylate extract contained proteins and LPS; however, the LPS content was less than in a LPS preparation. Based on earlier studies, LPS will probably not induce a protective immune response but only an antigen-specific unresponsiveness [48]. Dose–response studies indicated that unresponsiveness is antigen specific and could be induced by a single intraperitoneal injection of LPS in BALB/c mice [48]. LPS-based vaccines are therefore hardly feasible to protect against P. gingivalis infection.

Fimbriae

Fimbriae of P. gingivalis have been studied extensively as a potential vaccine target, not only owing to their outward protrusion from the cell surface but also owing to their potency to attach to host tissues or other microorganisms. The beneficial contribution of fimbriae in a vaccine could be demonstrated by the current application of fimbriae as part of a five-component acellular vaccine against Bordetella pertussis[49].

Genes encoding fimbriae

Research on P. gingivalis fimbriae has revealed two distinct fimbrillins: long major and short minor fimbriae called FimA and Mfa1, respectively [50,51]. A fimA-inactivated mutant strain of P. gingivalis ATCC 33277 was unable to produce the 41-kDa FimA protein and could not adhere to in vitro cultured cells but did produce a 67-kDa minor fimbrial protein [51]. The loci encoding FimA and Mfa1 are separated from each other, while their gene arrangements are similar. They contain a structural gene (fimA or mfa1), an unknown gene (fimB or mfa2) and three genes encoding minor components (fimC–fimE of 50, 80, 60 kDa or pg0180–pg0182) [52–54]. The latter has been suggested to be involved in autoaggregation and adhesion of P. gingivalis to glyceraldehyde-3-phosphate dehydrogenase of Streptococcus oralis, fibronectin and type I collagen [53]. Inactivation of one of the minor components leads to the absence of all minor components, referred to as devoid of all accessory proteins fimbriae (‘DAP fimbriae’) [53].

Fimbriae-based serotypes

Major fimbriae can be classified based on the presence of various fimA genes resulting in six fimbriae types: I, Ib, II, III, IV and V [55–57]. The fimbriae variations and their corresponding P. gingivalis strains are outlined in Table 2. Of these fimbriae-based serotypes, types II and IV predominate in periodontitis patients, while healthy individuals carry predominantly type I and V [58]. The involvement of type II fimbriae in periodontal disease may be due to their greater adhesive abilities than type I fimbriae [59].

Recognition of fimbriae by the immune system

Proinflammatory cytokines, such as IL-1β, IL-6, IL-8 and TNF-α, have shown to be induced in human peripheral blood monocytes and macrophage cell cultures upon simulation with P. gingivalis fimbriae or synthetic peptides [60]. These reactions do not depend on a single PRR. For example, fimbriae have shown to induce TLR2-dependent cell activation in a monocyte/macrophage model, using TLR1 or TLR6 as a signaling partner (Figure 2)[45]. DAP fimbriae showed only TLR2-dependent cell activation in association with TLR1 [45]. Besides the TLR1 and TLR6 signaling partners, other PRRs such as CD14 and the β₂-integrin CD11b/CD18 seem to be involved as coreceptors of TLR2 (Figure 2)[61]. CD14 is able to bind P. gingivalis fimbriae directly and induces binding by CD11b/CD18 in cooperation with TLR2 and phosphatidylinositol-3-kinase [61]. CD14 also interacts with other epitopes of P. gingivalis fimbriae than CD11b/CD18 does, as shown by synthetic peptides covering the entire FimA [62].

Recognition of fimbriae by a combination of PRRs, macrophage activation and further induction of the immune response by cytokines will eventually mediate P. gingivalis internalization by macrophages as shown by mice deficient in various PRRs [63]. Internalization of P. gingivalis is reduced if only DAP fimbriae are present [64]. Furthermore, fimbriae appear to downregulate the production of IL-12 p70 by macrophages, which classically activates natural killer cells and induces the differentiation of CD4 T cells into Th1 cells to induce bacterial clearance [62].

Immunization with fimbriae

Immunization of germ-free rats by subcutaneous injection with purified fimbriae protein showed protection against periodontal destruction when infection with P. gingivalis by gavage was accompanied by high levels of fimbriae-specific antibodies [65]. These results were based on fimbriae isolation from noninvasive P. gingivalis strains 381 and 2561 (ATCC 33277). Polyclonal antibodies induced by subcutaneous and intramuscular immunization of rabbits with purified fimbriae of a single P. gingivalis strain only bind closely related strains of the same fimbrial serotype, although not all strains of the same type show opsonization [66]. This indicates that a vaccine against P. gingivalis based on a single specific P. gingivalis fimbria variation strain might have negative consequences for the immunization outcome.

As mentioned previously, bacterial toxoids are often used in vaccines as immunomodulating adjuvants to enhance the uptake of antigens and therefore improve the induced immune response. In addition to serum IgM, IgG and IgA antibodies, salivary IgA antibodies specific for fimbriae were significantly increased after oral vaccination of mice with fimbriae in combination with CT when compared with fimbriae only [25]. CT is an exotoxin produced by Vibrio cholerae consisting of a single A subunit and five B subunits. Oral immunization requires a higher amount of fimbriae and CT for an antigen-specific immune response than a nasal vaccine does [67]. Coadministration of fimbriae and a recombinant B subunit of the CT by intranasal immunization of BALB/c mice stimulated both systemic and mucosal immune responses by inducing serum and secretory IgA. Furthermore, P. gingivalis-induced alveolar bone loss was reduced [68]. CPS–fimbriae conjugate vaccine against P. gingivalis further induced serum IgG and protection against P. gingivalis-induced infection when compared with CPS or fimbriae alone [33]. These results indicate that P. gingivalis fimbriae could not only be used as single antigen target but might be more beneficial as an immunomodulating adjuvant in an acellular vaccine. In addition, intranasal immunization has been shown to induce mucosal immune responses, which could be crucial for a safe and protective vaccine against P. gingivalis.

Fimbriae have been studied as part of a DNA vaccine. Mice immunized intranasally or intramuscularly with a plasmid pIRES-fimA or coexpression plasmid pIRES-fimA:IL-15, in which IL-15 is used as an adjuvant, showed induction of FimA-specific IgG in serum [69]. In addition to intransal and intramuscular immunization, targeted salivary gland immunization via injection of a fimbriae/fimA DNA vaccine into the salivary gland, using plasmid pcDNA3/fimA, showed enhanced fimbriae-specific IgA and IgG in saliva, and IgG in serum [70]. On the other hand, oral delivery of P. gingivalis FimA epitopes in germ-free rats via Streptococcus gordonii vectors results in FimA-specific serum IgG and IgA, salivary IgA, and subsequent protection against oral P. gingivalis-induced bone loss [71]. To further increase the immunogenicity and efficacy of these vaccines, P. gingivalis antigens could be fused to cytotoxic T lymphocyte-associated antigen 4 (CTLA4), thereby targeting these antigens to antigen-presenting cells [72]. As previously mentioned, these results also show that immunization via an oral or intranasal route is able to induce systemic as well as mucosal immune responses.

Outer membrane proteins

Outer membrane proteins have been extensively studied in Escherichia coli and other enterobacteria. Approximately 50% of the outer membrane is composed of proteins, of which porins and OmpA-like proteins are the most abundant [54]. Purification of the outer membrane [73] and the availability of the genomic sequence of P. gingivalis W83 [74] have led to the identification of abundant OMPs in P. gingivalis. RagA, RagB, Arg-gingipain, Lys-gingipain and OmpA-like proteins were thereafter further studied.

OmpA-like proteins

OmpA-like proteins identified in P. gingivalis: Pgm6/7 (40 and 120 kDa), PG32/33 (43 and 42 kDa) or OmpA40–41 (40 and 41 kDa), are encoded by open-reading frames pg0694 and pg0695[54]. Of these, PG32/33 has shown to be highly conserved among P. gingivalis strains and are homologous to OprF from Pseudomonas aeruginosa[75]. Inactivation of one or both genes led to significant growth impairment of P. gingivalis, underlining their importance in the bacterial metabolism. Transmembrane proteins, such as PG32/33, form a complex hydrophobic β-barrel structure that is difficult to mimic in soluble form. Correct conformation of the recombinant protein is required to obtain protection against P. gingivalis. In order to enhance the solubility and correct folding of the proteins, truncated versions of PG32 and PG33 were generated [75].

RagA & RagB proteins

RagA (115 kDa) is homologous to TonB-dependent outer membrane receptors, which are involved in outer membrane transport. RagB (55 kDa) has been predicted to be a lipoprotein. The genes encoding RagA and RagB are located adjacent to each other and are potentially functionally linked [54]. As shown in a mouse model, ragA and ragB mutants are significantly less virulent than wild-type strains [76]. Subcutaneous inoculation of bacteria, either an insertion mutant of ragA in strain W50 or its wild type, resulted in a higher survival rate with ragA mutant-inoculated mice. Since the ragA mutant was phenotypically ragB negative, results were ascribed to the inactivation of the rag locus. There are four variants of the rag locus, rag1 to rag4[77]. W50 and W83, which both belong to the invasive serotype K1, contain the rag1 gene.

Other OMPs

Other TonB-dependent OMPs are demonstrated or suggested to be involved in hemin aquisition, such as hemin-uptake receptors HmuR and Tlr /Tla, outer membrane receptor HemR or hemin-binding protein IhtB (Pga30) [54]. Although hemin acquisition is essential for growth and hemins are therefore a potential target for vaccination, there are no reports linking these proteins to immune responses. The reports on the cloning and sequencing of the tlr/tla and ihtB genes might include data that could be applied to a vaccine [78,79].

Immunization with OMPs

Of the OMPs proteins described above, only the OmpA-like protein PG32/33 and the 40-kDa OMP have been studied with respect to immunization possibilities against P. gingivalis. Subcutaneous immunization with recombinant PG32 and PG33, cloned for expression in E. coli, significantly reduced the lesion size and therefore induced protection against P. gingivalis oral challenge in a murine lesion model [80]. For oral immunization with the 40-kDa OMP, it is described that mouse-derived human monoclonal antibody against recombinant 40-kDa OMP (anti-r40-kDa OMP human monoclonal antibody) of P. gingivalis protects against oral P. gingivalis-induced alveolar bone loss in rats [81]. Second, transcutaneous or nasal immunization of mice with 40-kDa OMP and CT has shown to induce serum IgG, IgA and IgG in saliva [82,83]. Presently, adjuvant-free transcutaneous immunization of rats with 40-kDa OMP induced a similar immune response and protection against subcutaneous challenge with P. gingivalis 381, a nonvirulent strain [84]. To circumvent the adverse effects of the CT, a nontoxic chimeric variant of CT could be used. Mice nasally immunized with 40-kDa OMP together with mCTA/LTB, a nontoxic chimeric adjuvant that combines the A subunit of mutant cholera toxin E112K with the pentameric B subunit of heat-labile enterotoxin from enterotoxigenic E. coli, induced high levels of 40-kDa specific serum IgG and IgA, and IgA in saliva, and a concomitant reduction of oral P. gingivalis-induced alveolar bone loss [85]. Comparable with immunization with 40-kDa and CT, sublingual immunization of mice with 40-kDa and a cDNA vector plasmid encoding Flt3 ligand (pFL) induced 40 kDa OMP-specific IgG and IgA responses in serum, and IgA in saliva, and resulted in a reduction of alveolar bone loss caused by oral infection with P. gingivalis ATCC 53977 [86].

Gingipains

Proteases are assumed to play an important role in destruction of gingival tissue and alveolar bone caused by P. gingivalis-induced periodontitis. Cysteine proteases have been isolated from P. gingivalis and most of the trypsin-like activities are due to the presence of Arg- and Lys-specific cysteine proteases and are collectively nominated as gingipains [87,88]. Although, there is no vaccine based on gingipains so far, the application of gingipains in a vaccine has been studied extensively.

Two highly related genes, rgpA and rgpB, encode products that account for the major arginine-specific cysteine protease activity in P. gingivalis, which is conserved among P. gingivalis strains. rgpA also encodes polypeptides that are involved in hemagglutination and adherence. The lysine protease activity of P. gingivalis is derived from a single gene product encoded by kgp.

Proteolytic activity

RgpA, RgpB and Kgp are the major contributors in the proteolytic activity of P. gingivalis, as has been demonstrated by gelatin degradation assays using wild-type strain ATCC 33277 and five gingipain-deficient mutants which are deficient in rgpA, rgpB, rgpA and rgpB, kgp, or all three gingipains: rgpA, rgpB and kgp, respectively [89]. These gingipains promote growth of P. gingivalis by producing assimilate peptides and are required for bacterial resistance against the bactericidal activity of human serum. Moreover, these mutants show differences in hemagglutination and hemolysis. Arg-gingipain gene inactivation results in decreased hemagglutination. Lys-gingipain gene inactivation results in increased hemolytic activity, while the gingipain-deficient mutant showed similar hemolytic activity to the wild type. This suggests that there are also other factors involved in the lysis of erythrocytes. In addition, Arg-gingipains are involved in cell-damaging effects on human gingival fibroblasts, as shown by morphological changes and cell death [89]. P. gingivalis use hemoglobin efficiently as an important iron source required for bacterial growth. Hemagglutination and hemolytic activities may therefore be indicative for utilization of this iron source by P. gingivalis. Lys-gingipain isolated from P. gingivalis 381 has been shown to bind to human hemoglobin, which is suggested to be mediated by active domains distinct from those for protease activity [90]. These results demonstrate that gingipains are important factors in P. gingivalis growth and its proteolytic activity towards the host. Gingipains, therefore, seem to be useful in a vaccine against P. gingivalis.

Immunization with gingipains

Subcutaneous immunization of Macaca fascicularis with purified cysteine protease from P. gingivalis W12 resulted in a high induction of serum IgG titers, reduction of P. gingivalis in subgingival plaque and inhibition of the onset and progression of alveolar bone loss [91]. In addition, the extent of the reduction in alveolar bone loss was of a greater and more uniform extent than that achieved by the whole-cell vaccine, even though the same protocol was used [17,91].

The protective effect of anti-Kgp and anti-Rgp IgG seems to vary among P. gingivalis serotypes, as shown by subcutaneously and intramuscularly immunized rabbits with a vaccine containing purified HRgpA or Kgp from P. gingivalis HG66/W83 [92]. Anti-Kgp and anti-HRgpA antibodies have similar characteristics in binding to whole P. gingivalis cells for all of the serotypes tested, although the serotype classification based on capsular antigens (K^-, K1 to K7) is not completely equal to the classification based on antibody responses to outer membrane antigens (A–D): ATCC33277 (serotype A, noninvasive), A7A-28 (serotype B, K3), W50 (serotype C, K1) and 381 (serotype D, K^-). While differences were seen for the enhancement of polymorphonuclear leukocyte- mediated opsonization, phagocytosis and bacterial killing immunization with gingipains derived from one P. gingivalis serotype might therefore show no protection against other serotypes [92].

Subcutaneous immunization with RgpA can stimulate the production of P. gingivalis- and RgpA-specific serum IgG in a murine chamber and oral challenge model [93]. These antibodies are directed to the hemagglutinin domain of RgpA and coincide with the protection against alveolar bone loss elicited by P. gingivalis A7A1–28. On the other hand, immunization with H-K P. gingivalis or purified RgpB alone resulted in elevated levels of P. gingivalis-specific serum IgG but gave no protection against bone loss. It was concluded that the enhanced serum IgG and protective effect were due to the hemagglutinin domain of RgpA, which is absent in RgpB [93]. In contradiction to these results, the catalytic domain of RgpA has also shown to be involved in inducing protective immune responses (serum IgG) against P. gingivalis A7436 in a mouse chamber model [94]. P. gingivalis-specific serum IgG against functionally defined peptide fragments derived from the catalytic and hemagglutitin/adhesion domains of RgpA were induced after intraperitoneal immunization of mice with RgpA, RgpB or multiple antigenic peptide-conjugates of Rgp-derived peptides [94]. Although these results illustrate that specific domains of a specific gingipain enhance the immune response against that specific gingipain, this alone might not induce sufficient protection against P. gingivalis. Since all three gingipains are involved in the proteolytic activity, only immune responses to a common gingipain domain will induce sufficient protection.

Immunization with RgpA–Kgp protease–adhesin complex

By applying the RgpA–Kgp protease–adhesin complex in a vaccine, immunization against two important gingipains could be gained at once. First of all, serum IgG, IgM and IgG subclass responses to the RgpA–Kgp protease–adhesin complex of P. gingivalis have been studied in adult periodontitis patients [95]. RgpA–Kgp-specific serum IgG was significantly elevated in the diseased group and was associated with mean probing depths and the percentage of sites positive for P. gingivalis, which is indicative of disease severity. On the other hand, serum IgG subclass distribution was similar for diseased and healthy subjects, and the concentrations of thes IgG subclasses decrease in the following order: IgG₄ > IgG₂ > IgG₃ = IgG₁.

Subcutaneous immunization of mice with the purified RgpA–Kgp protease–adhesin complex, derived from P. gingivalis W50, protects against subsequent subcutaneous challenge with invasive W50 and noninvasive ATCC 33277 strains of P. gingivalis, as measured by the lesion size [96]. This type of immunization restricted the colonization by P. gingivalis and alveolar bone loss in rats [97]. Furthermore, peptides derived from this complex, active-site and adhesin-binding motif peptides also gave protection against subcutaneously induced infection with P. gingivalis when BALB/c mice were subcutaneously immunized with these peptides conjugated to diphtheria toxoid. These vaccines even protects against orally P. gingivalis-induced alveolar bone loss in a murine periodontitis model where BALB/c mice were subcutaneously immunized [98]. Recombinant A1 adhesin domains were found to significantly attenuate P. gingivalis infection after subcutaneous immunization, while immunization with rRgpA_cat had no effect in a murine lesion model [99]. By contrast, investigation of T-cell proliferative and cytokine responses to RgpA–Kgp complexes in BALB/c mice has shown that their reactions were mainly directed to the proteolytic domain [100]. Recombinant proteins based on domains or regions of the RgpA–Kgp proteinase–adhesin complex might be applied in a vaccine [99]. When domains of RgpA–Kgp proteinase–adhesin complex could be combined with, for example, the catalytic domain of RgpB, this could lead to immunization against the main contributors of the proteolytic activity.

DNA vaccine based on gingipains

Rgp has been investigated as a rgpA DNA vaccine in BALB/c mice [101]. This vaccine contains pVax1 harboring the whole rgpA open-reading frame from P. gingivalis ATCC 33722. Besides the induction of specific serum IgG against P. gingivalis, which they also reported in previous work [102], different immunization routes seem to be of great influence. Intranasal immunization of BALB/c mice with rgpA DNA vaccine resulted in higher serum IgG titers and an additional IgA in saliva, when compared with intradermal immunization. Furthermore, the reduction in alveolar bone loss after challenge with P. gingivalis ATCC 33277 was more pronounced in the intranasal immunization group [101]. Although it has been shown that intradermal immunization with the rgpA DNA vaccine protects against challenge with invasive P. gingivalis strain W50 in a mouse lesion model, it is still unclear whether this vaccine also inhibits oral colonization by P. gingivalis, since this bacterium is not resident in the oral cavity of mice [102]. On the other hand, sera from immunized mice inhibited hemagglutination and adherence to type I collagen by P. gingivalis, which supports the role of RgpA in P. gingivalis hemagglutination, which was already suggested by decreased activity in an rgpA rgpB double mutant [103]. Intradermal immunization of mice with the rgpA DNA vaccine further showed reduced lethality against infection by a lethal dose of P. gingivalis W50 [104]. These results show that the different immunization routes of a DNA vaccine induce diverse mechanisms of protection. For example, intranasal immunization induced an enhanced systemic immune response and additional mucosal immune response when compared with intradermal immunization. Although remarkable results were gained, immunization with only a rgpA DNA vaccine does not protect against proteolytic activity, which is caused by a combination of all three gingipains.

All three gingipains are not yet combined in one vaccine, although both catalytic domains from Kgp (KGP_cd) and Rgp (RGP_cd) of P. gingivalis ATCC 33277 were used to construct plasmids pSeq2A/kgp_cd and pSeq2B/rgp_cd. These plasmids were tested for their applicability as a DNA vaccine in BALB/c mice [105]. Intramuscularly immunized mice challenged with intraperitoneal inoculation of P. gingivalis W50 showed preventive inflammatory responses, serum IgG, and prolonged survival rate of 43 and 27% for mice immunized by pSeq2A/kgp_cd and pSeq2B/rgp_cd, respectively, compared with the nonimmunized mice, which died within 72 h. In addition, anti-KGP_cd IgG inhibited hemoglobin binding by P. gingivalis, a response enhanced by antifimbriae IgG, while anti-RGP_cd IgG had a negligible inhibitory effect on hemoglobin binding. On the other hand, both anti-KGP_cd and anti-RGP_cd IgG had an inhibitory effect on Lys- and Arg-specific proteolytic activities and growth of P. gingivalis under iron-restricted conditions [105]. If both plasmids, pSeq2A/kgp_cd and pSeq2B/rgp_cd, could be combined or administered at once, this would be able to induce protection against all gingipains.

Hemagglutinin

Hemagglutinins are capable of agglutinating erythrocytes, suggesting a mechanism of adhering to host tissues and their involvement in virulence [106]. hagA, a gene revealed in P. gingivalis 381, encodes the surface protein hemagglutinin A. In addition to HagA, HagB, C, D and E are reported, of which HagB has been tested for applicability in a vaccine against adult periodontitis [46].

In contradiction to both LPS and fimbriae from P. gingivalis, Hag B has shown to signal through TLR4 instead of TLR2 (Figure 2)[41,45,46]. Membrane-bound CD14, adaptor molecule myeloid differentiation primary response gene 88 (MyD88) and TIR-domain-containing adapter-inducing IFN-β (TRIF) are essential participants of the TLR4 signaling complex [46]. Furthermore, activation of dendritic cells via TLRs differs from macrophages with respect to cytokine production, requirement of signaling partners and the induced signaling pathways. This indicates distinct functions of these cell types in the subsequent immune response [46]. In view of the fact that P. gingivalis whole cells appear to induce an immune response predominantly through TLR2, in which LPS or fimbriae could be the main immunodominant antigens, HagB alone shows unique responses but is no immunodominant antigen [41,45,46].

Immunization with hemagglutinin

Subcutaneous immunization of rats with recombinant HagB derived from P. gingivalis 381 and subsequent oral infection with P. gingivalis invasive and noninvasive strains (ATCC 33277, 381, A7A1–28 and W50) showed no HagB-specific salivary IgA induction and only a slight serum IgM induction, but an enhanced serum IgG response [107]. The low level of serum IgM is suggested to be the result of cross-reactivity with components of the indigenous microorganisms in the rat. Although HagB is not a dominant surface component, its potency to act as a vaccine antigen is stated by its capability to mediate hemagglutination, reduce bone loss as measured by radiographic assessment and induce the production of several cytokines [107].

Expert commentary

Periodontal disease is one of the most prevalent diseases in the world, comprising 5–20% of any population with severe periodontitis, leading to extended tooth loss. Current treatment modalities are based on oral hygiene programs, maintenance of plaque control and antimicrobial therapy. The latter option is regarded as optional, mainly owing to a lack of data regarding the clinical efficacy in specific patient groups. This leads to the absence of consensus in treatment of periodontitis as a real infection.

However, this does not mean that distinct factors cannot be specified as crucial in the etiology of the disease. From a microbiological point of view, periodontitis is caused by a huge variety of species comprising a range of very low- to very high-virulence ones. Socransky and coworkers attempted to define clusters of species by virulence and simultaneous prevalence [108]. For example, P. gingivalis clusters with Tannerella forsythia and Treponema denticola. The two latter are almost unambiguous in periodontal diseased patients and healthy subjects. These two, however, are prone to eradication or reduction in number by standard periodontal therapy, in contrast to P. gingivalis and A. actinomycetemcomitans[109–111].

Since P. gingivalis is pointed out as one of the major etiological agents in chronic periodontitis [4,112], largely more than A. actinomycetemcomitans, a prevention program using a vaccine this pathogen would be of enormous gain for dental health. However, a vaccine should ideally protect against infection with other virulent periodontal pathogens too.

Immunization with several P. gingivalis-specific antigens has been shown to enhance the immune response against P. gingivalis, as demonstrated by the induction of specific antibodies and reduction of P. gingivalis-induced alveolar bone loss in animal models. Although complete protection through immunization has not yet been achieved, new knowledge about specific P. gingivalis antigens holds promising possibilities for the future.

Studies comprising serotype differences, either CPS, LPS or fimbriae variations, have shown that these variations among P. gingivalis strains have a huge influence on the extend of protection after immunization. Vaccines should therefore contain P. gingivalis-specific epitopes that are shared among these serotypes. New developments in immunization techniques, such as the DNA vaccine, might fill these requirements, although national and local regulations will hamper the introduction of such programs for use in healthy subjects. However, sequences known for their shared homology among serotypes, for example, K-antigen genes PG1135–PG1142, could be applied in vaccines to develop broad protection among P. gingivalis strains [35]. In order to use P. gingivalis-specific antigens, persuade broad strain protection and induce a sufficient immune response, a combination of several antigens in a vaccine is essential. Current licensed vaccines, for example, the five-component acellular vaccine against B. pertussis, are based on this concept [49].

Besides the combination of P. gingivalis-specific antigens, antigen delivery and adjuvant systems contribute to a successful vaccine. Mucosal vaccines especially, which are of particular interest for periodontal disease, due to their action on the mucosal surface, require these systems [113]. Adjuvants and antigen delivery systems may have different modes of action. The main contributions of cholera toxoids as an adjuvant, for example, might be its potential to enhance antigen presentation, promoting isotype differentiation in B cells, stimulatory and/or inhibitory effects on T-cell proliferation and cytokine production [113]. The mode of action of IL-15 as an adjuvant could be its stimulating effects on natural killer cells, B-cell proliferation and differentiation, and maintenance and activation of CD8⁺ memory T cells [69]. CTLA4, for example, will act as coupled carrier, targeting specific antigens towards antigen-presenting cells by interaction with the B7 molecule [72].

A newly developed multivalent vaccine against P. gingivalis that shows induced protection in animal models, might not be valid for humans per se. P. gingivalis, which does not naturally occur in animal models, causes periodontitis in humans as a main contributor in a multispecies disease. The experimental model used should therefore be taken into account. Models using, for example, oral infection, are more comparable to the human P. gingivalis infection than subcutaneously induced infections. In addition, the route of immunization should be adjusted to the target. Mucosal, as well as systemic, immune responses are effective in inducing protection against P. gingivalis. Hence, oral or nasal immunizations, which in many studies have shown to induce mucosal as well as systemic immune responses, are preferred. If protection against P. gingivalis in humans is achieved, other bacteria might prevail, with unknown consequences. For that reason, a vaccine against P. gingivalis might not meet with the safety rules applying to vaccine licensure. Furthermore, since periodontitis is not a life-threatening disease, this acts as a curb on further development of such a vaccine, although there is increasing evidence of the role of periodontitis in the development of cardiovascular diseases. Even though, complete protection with a vaccine against P. gingivalis might not be achieved, current knowledge about immune responses against P. gingivalis antigens could lead to other treatment options. Prevention of recolonization after manual treatment or neutralizing proteolytic activity by P. gingivalis-specific antibodies could already substantially decrease the disease burden, resulting in a lifelong protection of P. gingivalis-induced periodontal infection. In other words, the pathogen becomes a commensal microorganism, similar to Haemophilus influenzae and Neisseria meningitidis in the pharynx. A challenging test for further research is to develop a multivalent and/or a multicomponent vaccine that induce broad protection against P. gingivalis strains.

Five-year view

With increasingly ageing populations in developed countries, chronic periodontitis will become a serious problem in a large part of the global population. The impact of periodontitis on general health becomes clearer. However, current surgical and nonsurgical treatment protocols are available for only a small part of the population due to insurance issues. In addition, the widespread use of antibiotics without bacteriological diagnostic guidance is very undesirable.

The next period in dentistry will focus on prevention. Inflammatory manifestations of the soft tissues in particular are suspected to be of a growing concern in dentistry and general medicine. Prevention of periodontitis will be one of the main topics. However, new insights and knowledge about virulence factors of P. gingivalis is needed to develop a promising and safe multicomponent vaccine. The combination of fimbrial antigens, which are well-known for their immunogenic boosting activity, and a panel of CPS glycoconjugates to broaden the spectrum in combination of lifelong protection is an ambitious but realistic approach. A crucial step is to develop programs to start with Phase I/II clinical trials. Results from such trials will be of significant value for testing vaccine safety, but also for other preventive modi. The mode of administration should be focused on a mucosal immune response via oral or nasal vaccination since IgA induction is highly preferable to IgG production by B cells in the tissue, which can trigger osteoclasts to prolong alveolar bone loss. Whether such a vaccine will be available in practice depends on the efforts of governments, academic institutes, general practitioners and industries.

Table 1. Immunization with Porphyromonas gingivalis surface antigens.

Antigen type	Modification	Administration	Model	Immunization results	Ref.
Capsule	Whole CPS	sc.	Mouse	CPS-specific IgG and IgM titers elevated. Protection against P. gingivalis-elicited alveolar bone loss	[31]
	Polysaccharide–BSA conjugate	ip.	Mouse	Induction of serum antibodies and reduced severity of infection after challenge with P. gingivalis	[32]
	Polysaccharide–fimbriae protein conjugate	ip.	Mouse	Higher serum IgG and highest protection against P. gingivalis-induced infection compared with CPS or fimbriae alone	[33]
LPS	LPS	ip.	Mouse	No induction of serum IgG, antigen-reactive lymphocytes and protection against P. gingivalis-induced infection	[47]
	LPS/dose-response study	ip.	Mouse	Unresponsiveness is antigen-specific and could be induced by a single injection of LPS	[48]
Fimbriae	Fimbriae accompanied by fimbriae-specific antibodies	sc.	Rat	Protection against periodontal destruction	[65]
	Fimbriae of a single fimbria variation	sc., im.	Rabbit	Induced polyclonal antibody only binds closely to related strains of the same fimbrial biovar	[66]
	Fimbriae in combination with CT	Oral	Mouse	In addition to serum IgM, IgG and IgA antibodies, salivary IgA specific for fimbriae was significantly increased	[25]
	Fimbriae coadministration with rCTB	in.	Mouse	Stimulation of both systemic and mucosal immune responses and reduced P. gingivalis-induced alveolar bone loss	[68]
	DNA vaccine: plasmid pIRES-fimA, coexpression plasmid pIRES-fimA:IL-15	im, in.	Mouse	in. and im. administration both induced FimA-specific IgG in serum. Only in. administration was able to enhance FimA-specific IgA in saliva, which was further increased by including IL-15 as an adjuvant	[69]
	DNA vaccine: plasmid pcDNA3/fimA	Injection into salivary gland	Mouse	Enhanced fimbriae-specific IgA and IgG in saliva and IgG in serum	[70]
	DNA vaccine: Streptococcus gordonii vectors	Oral	Rat	Enhanced FimA-specific serum IgG, IgA and salivary IgA	[71]
OMPs	OmpA-like protein: PG32/33	sc.	Rat, mouse	Significantly reduced the lesion size and therefore induced protection against P. gingivalis	[80]
	Anti-r40-kDa OMP hMAb	Oral	Rat	Protects against P. gingivalis-induced alveolar bone loss	[81]
	40-kDa OMP and CT	tc., in.	Mouse	Induced serum IgG, IgA and IgG in saliva	[82,83]
	40-kDa OMP	tc.	Rat	Induced serum IgG, IgA and IgG in saliva. Protection against challenge with P. gingivalis 381	[84]
	40-kDa OMP with mCTA/LTB	in.	Mouse	Induced high levels of 40-kDa-specific serum IgG, IgA and IgA in saliva. Reduction of P. gingivalis-induced alveolar bone loss	[85]
	40-kDa OMP with pFL	sl.	Mouse	Induced high levels of 40-kDa-specific serum IgG, IgA and IgA in saliva. Reduction of P. gingivalis-induced alveolar bone loss	[86]
Gingipains	Purified cysteine protease	sc.	Macaque	Induction of serum IgG and reduction of P. gingivalis in subgingival plaque. Onset and progression of alveolar bone loss was inhibited	[91]
	RgpA, RgpB, H-K whole cells	sc.	Rat, mouse	All vaccines induced specific serum IgG, but only RgpA induces protection against P. gingivalis- induced alveolar bone loss	[93]
	RgpA, RgpB, multiple antigenic peptide conjugate Rgp	ip.	Mouse	P. gingivalis-specific serum IgG against functionally defined peptide fragments derived from the catalytic and hemagglutinin/adhesion domains of RgpA	[94]
	RgpA-Kgp protease–adhesin complex	sc.	Mouse	Protects against challenge with invasive and noninvasive P. gingivalis strains	[96]
	Active site and ABM peptides from RgpA-Kgp complex conjugated to diphtheria toxoid	sc.	Mouse	Protection against P. gingivalis-induced alveolar bone loss	[98]
	A1 adhesin domain, rRgpAcat	sc.	Rat, mouse	Adhesin domains significant attenuate P. gingivalis infection, while rRgpAcat does not	[99]
	rgpA DNA vaccine	in., id.	Mouse	in. administration results in higher serum IgG, additional IgA in saliva and reduction of alveolar bone loss, when compared with id. administration	[101]
	rgpA DNA vaccine	id.	Mouse	Reduced lethality against infection by a lethal dose of P. gingivalis	[104]
	DNA vaccine: plasmids pSeq2A/kgpcd and pSeq2B/rgpcd	im.	Mouse	Preventive inflammatory responses, serum IgG, and prolonged survival rate	[105]
Hemagglutinin	rHagB	sc.	Rat	No HagB-specific salivary IgA, slight serum IgM, while serum IgG was actively enhanced. Reduction of bone loss	[107]
	rHagB alone or with MPL	in.	Mouse	Significant increase of HagB-specific salivary IgA and serum IgG compared with rHagB alone	[114]
	Salmonella typhimurium strain χ4072 expressing hagB	ig.	Mouse	Induced HagB-specific Hag serum IgG and salivary IgA	[115]
	S. typhimurium strain expressing hagB fused to Lpp-OmpA	oral	Mouse	Higher serum anti-HagB IgG and IgA, than by the cytoplasmic expressing strain	[116]
	130-kD HMGD	id.	Mouse	hMAb significantly inhibits hemagglutination by P. gingivalis and its vesicles	[117]

ABM: Adhesin-binding motif peptides; BSA: Bovine serum albumin; CPS: Capsule polysaccharide; CT(B): Cholera toxin (B subunit); H-K: Heat-killed; (h)MAb: (Human) monoclonal antibody; HMGD: Hemagglutinin domain; id.: Intradermal; ig.: Intragastric; im.: intramuscular; in.: Intranasal; ip.: Intraperitoneal; Kgp: Lysine gingipain; LPS: Lipopolysaccharide; mCTA/LTB: Mutant A subunit cholera toxin/B subunit heat-labile toxin; MPL: Monophosporyl lipid A; OMP: Outer membrane protein; pFL: Plasmid containing the Flt3 ligand; pIRES: Plasmid containing an internal ribosome entry site; rCTB: Recombinant cholera toxin B; Rgp(A/B): Arginine gingipain (A/B); (r)Hag: (Recombinant) hemagglutinin; sc.: Subcutaneous; sl.: Sublingual; tc.: Transcutaneous.

Table 2. Antigenic variation among Porphyromonas gingivalis strains.

P. gingivalis strain	Capsular serotype	LPS serotype	Fimbrial variation	Invasive/virulent	Ref.
1112	ND	ND	I	ND	[29]
1432	ND	ND	I	ND	[29]
2561	ND	ND	I	No	[29]
34–4	K7	ND	ND	Yes	[27]
381	K^-	O1	I	No	[22,23,27,29,36]
6/26	ND	ND	III	ND	[56]
9–14K-1	ND	ND	IV	Yes	[29]
A7A1–28/ATCC 53977	K3	ND	II	Yes	[22,23,29,57]
A7A2–10	ND	ND	II	Yes	[29]
ATCC 33277	ND	ND	I	No	[29,56]
ATCC 49417/RB22D-1	K4	ND	II	Yes	[22,23,57]
HG 1690	K5	ND	II	Yes	[22,23,29,57]
HG 1691	K6	O2	Ib	Yes	[22,23,36,57]
HG 1694	ND	ND	ND	No	[29]
HG 184	K2	ND	II	Yes	[22,23,57]
HG 564	ND	ND	IV	ND	[57]
HG 565	ND	ND	I	ND	[58]
HNA-99	ND	ND	V	Yes	[29,56]
HW24D1	ND	ND	II	ND	[56]
W50	K1	O3	IV	Yes	[22,23,29]
W83 / HG 66	K1	ND	IV	Yes	[22,23,29]

LPS: Lipopolysaccharide; ND: Not documented.

Key issues

• • Whole-cell vaccines against Porphyromonas gingivalis are not appropriate owing to the large variety of serotypes.

• • A combination of fimbrial P. gingivalis-specific antigens in combination with a capsular polysaccharide glycoconjugate in a vaccine is essential to induce a sufficient immune response with lifelong protection.

• • To ensure protection against P. gingivalis without the induction of osteoclast activation by IgG-induced triggering of neutrophils, mucosal immunization via oral or nasal administration is preferable.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.