Periodontitis: from microbial immune subversion to systemic inflammation

Abstract

Periodontitis is a dysbiotic inflammatory disease with an adverse impact on systemic health. Recent studies have provided insights into the emergence and persistence of dysbiotic oral microbial communities, which can mediate inflammatory pathology at local as well as distant sites. This Review discusses mechanisms of microbial immune subversion that tip the balance from homeostasis to disease in oral or extraoral sites.

Periodontitis is a chronic inflammatory disease that compromises the integrity of the tooth-supporting tissues, that is gingiva, periodontal ligament and alveolar bone, collectively known as the periodontium¹(BOX 1). Known since antiquity, periodontitis became prevalent after the domestication of plants and animals in Neolithic societies (≈10,000 years ago) when the oral microbiota underwent a distinct compositional shift — with increased frequency of Porphyromonas gingivalis and other periodontitis-associated species — compared with earlier hunter-gatherer societies². In its severe form, which afflicts 8.5% of U.S. adults³, periodontitis may not only cause tooth loss, but can also affect systemic health by increasing the patients’ risk for atherosclerosis, adverse pregnancy outcomes, rheumatoid arthritis, aspiration pneumonia and cancer^4-9.

Box 1. Periodontitis and susceptibility factors.

Periodontitis has a complex etiology acting at multiple levels: at the microbial level, based on the presence of dysbiotic microbial communities with potential for destructive inflammation; at the host level, based on genetic factors that may predispose to or protect from disease; and at the level of environmental factors and systemic health status that modify the host response in either protective or destructive direction¹⁵¹. Accordingly, dysbiosis by itself may not necessarily precipitate periodontitis, but it could initiate disease in the context of other risk factors associated with host genotype, stress, diet or risk-related behaviour such as smoking^92,152-156. For instance, there might be individuals who can tolerate dysbiosis by virtue of their intrinsic immuno-inflammatory status; hyporesponsive or lack-of-function polymorphisms in immune response genes could attenuate inflammation and prevent development of overt disease²⁰. Bacterial dysbiosis will only lead to disease in susceptible hosts as there are individuals who remain periodontally healthy despite massive tooth-associated biofilm formation, whereas others with less biofilm accumulation are extremely susceptible to periodontitis¹⁵⁴. Although a genetic basis for periodontitis is supported by twin studies and familial aggregation of severe forms of the disease, the implication of identified specific genes — such as IL1, IL6, TNF, FCGR2A, C5, CD14, WNT5A — is debatable^153-155. This uncertainty is probably attributed to the fact that chronic periodontitis is a polygenic disease, in which multiple genes contribute cumulatively to the overall disease risk (or protection) by influencing the host immune response and the microbiota composition and structure¹⁵⁵. This notion stands in stark contrast to monogenic forms of the disease, such as aggressive periodontitis in young patients with leukocyte adhesion deficiency, where a single gene (ITGB2, encoding integrin β-2) invariably precipitates periodontal disease⁵⁷.

A triadic group of oral anaerobic bacteria that comprises P. gingivalis, Treponema denticola and Tannerella forsythia have traditionally been considered as causative agents of periodontitis, based on their virulence properties and strong association with diseased sites¹⁰. However, recent advances from metagenomic, metatranscriptomic and mechanistic studies^11-16 are consistent with a new model of periodontal disease pathogenesis, which suggests that a more diverse periodontitis-associated microbiota than previously thought is involved in disease. In this model, disease results not from individual pathogens but rather from polymicrobial synergy and dysbiosis, which perturbs the ecologically balanced biofilm associated with periodontal tissue homeostasis^17-19(FIG.1). The dysbiosis of the periodontal microbiota signifies an imbalance in the relative abundance or influence of microbial species, which mediate distinct roles that synergize to shape a pathogenic entity that can cause disease in oral or extraoral tissues of susceptible individuals^6,8,11,20. In this new context of pathogenesis, the roles of individual bacteria and their interactions with host need to be re-evaluated.

Figure 1. Polymicrobial synergy and dysbiosis in periodontitis.

Periodontitis is induced in susceptible hosts by a polymicrobial community, in which different members fulfil distinct roles that converge synergistically to cause destructive inflammation. Keystone pathogens, the colonization of which is facilitated by accessory pathogens, initially subvert the host response leading to a dysbiotic microbiota, in which pathobionts over-activate the inflammatory response and cause periodontal tissue destruction, including resorption of the supporting alveolar bone. Inflammation and dysbiosis positively reinforce each other because inflammatory tissue breakdown products are used as nutrients by the dysbiotic microbiota. The lower panel depicts the progression from periodontal health (swallow gingival crevice; ≤2 mm) to gingivitis (periodontal inflammation without bone loss; gingival crevice ≤3 mm) to periodontitis (formation of periodontal pockets ≥4 mm and inflammatory bone loss). Inflammation-induced collagenolytic enzymes can contribute to loss of tissue attachment to the teeth and the deepening and ulceration of the pockets (up to 10-12 mm covering a surface area of 8-20 cm²), which serve as a niche that can harbour 10⁸ to 10¹⁰ bacteria feeding on the inflammatory spoils (for example collagen peptides, haem-containing compounds) carried with the gingival crevicular fluid (GCF) that bathes the pocket.

Central to the new model of pathogenesis, and constituting the main theme of this Review, is the active bacterial subversion of the host immune response in ways that enables pathogen persistence in the local inflammatory environment of periodontitis and induction of pathology or complications at systemic sites. We discuss evidence and mechanisms whereby periodontal organisms disseminate from their oral habitat to distant sites, including to atherosclerotic plaques, the lungs and placenta where they can disrupt immune surveillance and homeostasis to promote or accelerate pathogenic processes^9,21-29. Moreover, P. gingivalis in particular is examined as a potential cause for the generation of autoantibodies in rheumatoid arthritis^5,30-33. Understanding how oral pathogens misdirect the host immune response can provide novel mechanistic insights into the pathogenesis of periodontitis and associated systemic conditions as well as reveal new therapeutic targets.

Microbial synergy and dysbiosis

The transition from periodontal health to disease is associated with a dramatic shift from a symbiotic microbial community — composed mostly of facultative bacterial genera such as Actinomyces and Streptococci — to a dysbiotic microbial community structure composed mainly of anaerobic genera from the phyla Firmicutes, Proteobacteria, Spirochaetes, Bacteroidetes and Synergistetes. The dysbiotic oral microbiota is enriched in virulence factors and adapted to thrive in an inflammatory environment^11-14,34. The predominant habitat of periodontitis-associated bacteria is the subgingival crevice, where the bacteria are found in distinct microenvironments; the tooth-associated biofilm, the gingival crevicular fluid and the epithelium lining the crevice¹⁹ (FIG. 1). The subgingival environment is rich in immune and inflammatory mediators and provides unique challenges and opportunities for the bacteria^35-37. Periodontal health requires a controlled immuno-inflammatory state that can maintain host–microbe homeostasis in the periodontium³⁸. However, in periodontitis, the host immune response is dysregulated — either because it is subverted by the microbial community or because of host immunoregulatory defects — and is therefore ineffective to restrain bacterial outgrowth and overt pathogenicity¹⁹. A poorly controlled host immune response, in turn, can generate a self-perpetuating pathogenic cycle where dysbiosis and inflammation reinforce each other by forming a postive feedback loop (FIG. 1).

Most studies investigating microbial subversion of the periodontal host immune responses have primarily used P. gingivalis as a model pathogen, which is necessarily reflected in this Review. Decades of research have identified a plethora of documented or putative virulence factors of P. gingivalis that may contribute to its persistence in the subgingival region (for review see REFs.37,39). However, only now have we begun to understand how this Gram-negative asaccharolytic bacterium integrates its virulence attributes to enhance the pathogenicity of a polymicrobial community. Although P. gingivalis was thought to be capable of directly causing periodontitis in animal models⁴⁰, it is now established that P. gingivalis-induced periodontitis requires the presence of the commensal microbiota as P. gingivalis is unable to cause periodontitis in germ-free mice despite colinizing this host¹⁵. In this context, P. gingivalis — while present at a low frequency — is pathogenic owing to its ability to induce dysbiotic microbial communities, and thereby act as a keystone pathogen^16,41. Manipulation of the host immune response is fundamental to the capacity of P. gingivalis to instigate quantitative and qualitative alterations in the oral microbiota, which can thereby trigger inflammatory periodontal bone loss largely mediated by pathobionts^42-44. Although non-pathogenic by themselves in the oral environment, certain commensals such as Streptococcus gordonii promote P. gingivalis colonization and, as such, are implicated as accessory pathogens^19,45.

It should be noted that the presence of P. gingivalis does not necessarily precipitate disease. Indeed, P. gingivalis is detectable, albeit with decreased frequency, in periodontally healthy individuals^11,46. This might be explained by the considerable strain diversity within the population structure of P. gingivalis. Moreover, key virulence factors of this pathogen (such as gingipains and lipid A phosphatases) are regulated by local environmental conditions, which may differ among individuals⁴¹. In a related context, there might be individuals who can resist the conversion of a symbiotic microbiota into a dysbiotic one by virtue of their intrinsic immune status, for example they might have alterations in signalling pathways required for immune subversion by P. gingivalis or other keystone-like pathogens.

The concept that P. gingivalis co-operates with other periodontal organisms is supported by findings in animal models of periodontitis where combined inoculation of P. gingivalis with accessory pathogens or other keystone-like pathogens such as S. gordonii or T. forsythia, respectively, leads to enhanced alveolar bone loss compared with P. gingivalis alone^47-50. This synergism may not be restricted to manipulation of the host response, but may also entail co-operative interspecies communication that further promotes bacterial fitness and hence dysbiosis. At least in vitro, communication among periodontal bacteria causes reciprocal transcriptomic and proteomic responses that regulate nutrient acquisition, metabolic processes and production of virulence factors^51-53.

In summary, the emerging polymicrobial synergy and dysbiosis model suggests that the host immune response is initially subverted by keystone pathogens aided by accessory pathogens and is subsequently over-activated by pathobionts, thereby linking homeostasis breakdown and destructive inflammation in susceptible individuals (FIG. 1). Specific molecular mechanisms by which periodontal bacteria manipulate the host response to cause dysbiotic inflammation are discussed below.

Subversion of host immune responses

The dysbiotic periodontal community is faced with a survival conundrum: on the one hand, these bacteria need to evade immune-mediated killing; on the other hand, they require inflammation to procure nutrients from tissue breakdown such as degraded collagen peptides and haem-containing compounds²⁰. Hence, immunosuppression, though a common evasion strategy of many pathogens⁵⁴, is not a viable option for inflammophilic bacteria⁵⁵. Periodontal bacteria can manipulate the interaction with host immune responses — such as neutrophils and complement — to enhance bacterial fitness.

The role of neutrophils in periodontitis

Neutrophils are the most common leukocyte recruited to the subgingival crevice or periodontal pockets^35,36. Individuals with congenital deficiencies in neutrophil numbers or recruitment develop severe periodontitis, suggesting that neutrophils are required for periodontal tissue homeostasis^17,56,57. However, hyperactive, supernumerary or dysregulated neutrophils can cause collateral tissue damage through the release of inflammatory and toxic substances or tissue-degrading enzymes^35,58-60. Indeed, ample clinical evidence show that neutrophils mediate a significant portion of periodontal tissue destruction^61,62 and their local numbers positively correlate with the severity of chronic periodontitis⁶³. The chronic recruitment of excessive numbers of neutrophils to diseased periodontal pockets possibly arises from their inability to control the microbial challenge, despite being viable and capable of eliciting immune responses^20,58. This suggests that the microorganisms can evade neutrophil-mediated killing while promoting inflammation, thereby contributing to dysbiosis. Indeed, as discussed below, periodontal bacteria can subvert complement function in ways that interfere with neutrophil-mediated killing in a persisting inflammatory environment⁴².

Complement subversion

Complement is a cascade system of a network of proteins and receptors that are centrally involved in immunity and inflammation⁶⁴. Complement activation is initiated by distinct mechanisms, which converge at the third complement component (C3) and lead to the generation of effector molecules. These effector molecules mediate microbial opsonization and phagocytosis, for example the opsonin C3b interacts with complement receptor-1; recruitment and activation of inflammatory cells, such as anaphylatoxin C5a, which interacts with the C5a receptor (C5aR); and direct lysis of targeted microbes, mediated by the C5b-C9 membrane attack complex⁶⁴. P. gingivalis, T. forsythia and Prevotella intermedia can protect themselves and bystander bacteria from human complement-mediated opsonophagocytosis and killing by blocking complement activation through degradation of C3 or key upstream components such as the mannose-binding lectin^65-67. Importantly, the bacterial proteases involved — namely, gingipains, karilysin and interpain A, expressed by P. gingivalis, T. forsythia and P. intermedia, respectively — act synergistically to inhibit complement, which may lead to enhanced protection of complement-susceptible bystander species^65-67. P. intermedia and T. denticola can also evade human complement-mediated killing by capturing complement factor H, which is a physiological inhibitor of complement^68,69.

Intriguingly, the arginine-specific gingipains of P. gingivalis can cleave C5 to generate high local concentrations of C5a independently of canonical complement activation^65,70. Hence, in neutrophils, which recognize P. gingivalis through Toll-like receptor 2 (TLR2)⁷¹, this pathogen can co-activate C5aR and TLR2, which in turn leads to signalling crosstalk⁴². In both human and mouse neutrophils, this C5aR–TLR2 crosstalk leads to ubiquitination and proteasomal degradation of the TLR2 signalling adaptor MYD88, and thereby suppress its antimicrobial effects⁴²(FIG. 2). Moreover, the C5aR–TLR2 crosstalk activates an alternative pathway in which the TLR2 MYD88 adaptor-like protein (Mal) induces phosphoinositide 3-kinase (PI3K) signalling, which in turn inhibits GTPase RhoA-dependent actin polymerization and hence the phagocytosis of P. gingivalis and bystander bacteria⁴². The PI3K pathway also stimulates a robust inflammatory response⁴²(FIG. 2). Importantly, the local inhibition of C5aR, TLR2 or PI3K in the periodontium of P. gingivalis-colonized mice leads to elimination of P. gingivalis, reverses the increase in total microbiota numbers induced by P. gingivalis colonization and blocks periodontal inflammation⁴². In summary, P. gingivalis manipulates neutrophils through distinct mechanisms that together ensure the survival of the microbial community and the perpetuation of inflammation.

Figure 2. Porphyromonas gingivalis subversion of neutrophils leads to dysbiotic inflammation.

Porphyromonas gingivalis expresses ligands that activate the Toll-like receptor 2 (TLR1)–TLR2 complex and enzymes (HRgpA and RgpB gingipains) with C5 convertase-like activity that generate high local concentrations of C5a ligand. The organism can co-activate C5aR and TLR2 in neutrophils and the resulting crosstalk leads to ubiquitination and proteasomal degradation of the TLR2 adaptor MYD88, thereby inhibiting a host-protective antimicrobial response. This proteolytic event requires C5aR–TLR2-dependent release of transforming growth factor-β (TGF-β1), which mediates MYD88 ubiquitination via the E3 ubiquitin ligase Smurf1 (enlarged inset). Moreover, the C5aR–TLR2 crosstalk activates phosphoinositide 3-kinase (PI3K), which prevents phagocytosis through inhibition of RhoA GTPase and actin polymerization, while stimulating the production of inflammatory cytokines. In contrast to MyD88, another TLR2 adaptor, Mal, contributes to immune subversion by acting upstream of PI3K. These functionally integrated pathways, as manipulated by P. gingivalis, provide ‘bystander’ protection to otherwise susceptible bacterial species and promote polymicrobial dysbiotic inflammation in vivo. C5aR, complement C5a receptor; HRgpA, high molecular mass arginine-specific gingipain A; Mal, MyD88 adaptor-like; MyD88, myeloid differentiation primary response protein 88; RgpB, arginine-specific gingipain B; Smurf1, Smad ubiquitin regulatory factor 1. Reproduced with permission from REF.⁴².

Interestingly, the karilysin of T. forsythia was recently shown to also cleave C5 to release C5a⁶⁶. Since T. forsythia activates TLR2⁷², it would be important to determine whether T. forsythia can instigate a C5aR–TLR2 subversive crosstalk similar to that of P. gingivalis or whether the two organisms can synergize in that regard. Intriguingly, both gingipains and karilysin readily degrade the C5b component of C5, and thereby prevent formation of the membrane attack complex^65,66.

Additional immune subversive mechanisms

Major immune-subversive organisms probably use additional strategies to protect bystander bacteria and elevate the virulence of the entire microbial community, although most of these putative mechanisms have not been confirmed in vivo. For instance, the capacity of P. gingivalis to degrade and inactivate antimicrobial peptides might confer in vivo protection to bystander bacteria^36,67. In a related context, T. denticola blocks the production of human β-defensins by gingival epithelial cells in response to Fusobacterium nucleatum, which promotes the colonization of periodontitis-associated bacteria⁷³. Mechanistically, T. denticola blocks the fusion of internalized F. nucleatum with lysosomes and suppresses induction of intracellular reactive oxygen species, and thereby inhibits TLR responses that control human β-defensin expression⁷³. Macrophages can engulf periodontal bacteria that invade the gingival connective tissue. However, P. gingivalis can induce cAMP-dependent activation of protein kinase A, which inhibits the expression of inducible nitric oxide synthase, thereby suppressing nitric oxide-dependent intracellular killing in macrophages⁷⁴. Moreover, P. gingivalis suppresses human and mouse macrophage endocytosis of F. nucleatum, an event required for NLRP3 inflammasome activation in response to this bacterium⁷⁵. This mechanism may promote the fitness of the periodontal community since inflammasome activation induces pyroptosis, a proinflammatory mode of lytic cell death protecting the host against infection⁷⁶. P. gingivalis is also thought to manipulate adaptive immune responses by favouring the differentiation and recruitment of CD4⁺ T helper 17 (Th17) cells at the expense of the Th1 lineage^77-79. Although it has been argued that Th17 cells contribute to destructive periodontal inflammation whereas Th1 cells are involved in protective cell-mediated immunity^80,81, more research is warranted to elucidate the role of various T cell subsets in periodontal disease pathogenesis.

When the periodontium is chronically exposed to a dysbiotic microbial community — which apparently evolved to evade the host immune response while promoting its inflammatory aspects — it probably has an adverse impact on systemic health. Below, we examine established and emerging mechanisms whereby periodontal bacteria can subvert host-signalling pathways to instigate chronic inflammation in extraoral sites.

Periodontitis and cardiovascular disease

Numerous cross-sectional, case-control and cohort epidemiological studies suggest that periodontitis is associated with atherosclerotic cardiovascular disease, independently of confounding factors such as smoking and obesity^6,82,83. Moreover, clinical interventional studies indicate that treatment of periodontitis reduces systemic inflammation and has favourable effects on subclinical markers of atherosclerosis, including improved endothelial function as determined by flow-mediated dilatation^83-85. Moreover, a recent study has shown that longitudinal improvement in periodontal health is related to a decreased progression of carotid atherosclerosis in humans⁸⁶.

At least two biologically plausible mechanisms may account for a causative link between periodontitis and atherosclerosis^4,83 (FIG. 3). First, gingival ulceration in periodontal pockets (FIG. 1) enables the translocation of bacteria into the systemic circulation, that is, causing bacteraemia that are well documented in patients with periodontitis and may provide an atherogenic stimulus^4,83. This notion is supported by findings that recurrent experimental bacteraemia (using P. gingivalis as a model pathogen) promotes coronary and aortic atherogenesis in pigs with or without hypercholesterolemia⁸⁷. Additionally, in a subset of patients with severe periodontitis and ulcerated gingival epithelium, locally produced proinflammatory cytokines, such as tumour necrosis factor (TNF), interleukin-1β (IL-1β) and IL-6, can enter the systemic circulation and induce an acute-phase response in the liver (including elevated C-reactive protein, fibrinogen and serum amyloid A), and thereby promote atherogenesis^4,88. In support of this mechanism, patients with severe periodontitis have increased systemic inflammation — determined by elevated cytokines and acute-phase markers such as IL-6 and C-reactive protein, respectively — compared with healthy controls, whereas treatment of periodontitis reduces systemic inflammation in patients with or without history of cardiovascular disease^83,85. An alternative mechanism was suggested by a recent study in mice, which showed that P. gingivalis can cause alterations to the gut microbiota, leading to indirect induction of systemic inflammation⁸⁹(FIG. 3). Specifically, mice orally infected with P. gingivalis showed increased proportion of Bacteroidetes and decreased proportion of Firmicutes relative to sham-infected controls, which correlated with decreased expression of tight-junction proteins in the ileum, endotoxemia and systemic inflammation. Although large quantities of oral bacteria are constantly swallowed via the saliva in humans and animals, P. gingivalis was not detected in the gut of infected mice; therefore, the mechanism by which it caused compositional changes to the gut microbiota remains uncertain. The mechanistic basis of the association between periodontitis and atherosclerosis is substantiated by studies in animal models based on oral infection with P. gingivalis, which has been detected in human atherosclerotic tissue, and by clinical or in vitro studies attesting to the atherogenic potential of periodontal bacteria^{6,25-29,90-94}.

Figure 3. Biologically plausible mechanisms linking periodontitis to systemic inflammation and disease.

In periodontitis, locally produced pro-inflammatory cytokines can enter the systemic circulation and induce an acute-phase response in the liver — which is characterized by increased levels of C reactive protein, fibrinogen and serum amyloid A — in turn contributing to atherosclerosis or exacerbating intra-uterine inflammation. Moreover, gingival ulceration in periodontal pockets enables the egress and systemic dissemination of periodontal bacteria. Certain bacteria including Porphyromonas gingivalis have been detected in circulating leukocytes and in atherosclerotic lesions, where they may act as pro-atherogenic stimuli. Other periodontal bacteria such as Fusobacterium nucleatum have been detected in the placenta where they can cause adverse pregnancy outcomes. Large quantities of oral bacteria are constantly swallowed on a daily basis via the saliva into the gut. In this context, an alternative, or additional, mechanism linking periodontitis to systemic inflammation was recently proposed: Swallowed P. gingivalis causes alterations to the gut microbiota, thereby leading to increased gut epithelial permeability and endotoxemia, which causes systemic inflammation. Although independent, the depicted events are not mutually exclusive but could in principle occur simultaneously. CRP, C-reactive protein; IL, interleukin; TNF, tumor necrosis factor.

Animal model-based evidence

Oral infection with P. gingivalis of rabbits fed a high-fat diet or of atherosclerosis-prone (hyperlipidemic) apolipoprotein E-deficient (ApoE^−/−) mice on a standard chow diet causes not only local bone loss but also systemic inflammation and atherosclerotic lesions^28,90,91. The ApoE^−/− mouse model of combined periodontitis and atherosclerosis has been extensively used and provided insights into the role of TLR signalling in the disease process. Whereas TLR2 has an important role in mediating P. gingivalis-induced inflammatory atherosclerosis⁹⁵, the disease phenotype requires P. gingivalis to evade TLR4-mediated detection²⁹.

In this regard, the organism can enzymatically modify the lipid A moiety of its lipopolysaccharide to either evade or antagonize TLR4 activation in macrophages¹⁷. The shifting of lipid A activity from TLR4-evasive to TLR4-antagonistic depends on endogenous lipid A phosphatase activity⁹⁶, which is regulated by growth phase or environmental factors including temperature and haemin availability^97,98. Differential activities of lipid A 1’- and 4′-phosphatases are associated with the synthesis of different lipid A structures, comprising non-phosphorylated tetra-acylated lipid A (which is inert for TLR4 activation), mono-phosphorylated penta-acylated lipid A (which is a weak TLR4 agonist) and mono-phosphorylated tetra-acylated lipid A (a TLR4 antagonist). Genetic ablation of 4′-phosphatase activity (or bacterial growth at ≥39°C) leads to the synthesis of TLR4 agonist lipid A, whereas ablation of 1-phosphatase activity — or growth in haemin-replete conditions expected to be present in an inflammatory environment — leads to TLR4 antagonist lipid A^96,99. Remarkably, in addition to preventing TLR4 activation, the production of inert or antagonistic lipid A also increases the resistance of P. gingivalis to cationic antimicrobial peptides, owing to changes in the outer surface charge of the bacteria that affect the binding of cationic peptides^29,96,97,99. Importantly, the TLR4-antagonist lipid A of P. gingivalis inhibits TLR4 activation in response to other bacteria that express TLR4-agonist lipid A species^100,101.

In ApoE^−/− mice orally infected with wild-type P. gingivalis or isogenic phosphatase mutants with a ‘locked’ lipid A profile, the expression of inert or antagonistic lipid A was associated with elevated vascular inflammation, macrophage infiltration and progression of atherosclerosis²⁹. In macrophages infected with the same set of wild-type or mutant strains, the expression of lipid A structures that prevent or block TLR4 activation was associated with evasion of non-canonical inflammasome activation and increased bacterial survival, as compared to the expression of TLR4 agonist lipid A²⁹. Taken together with an earlier study by the same group⁹⁵, these findings suggest that the capacity of P. gingivalis to prevent TLR4 signalling while activating TLR2 leads to atherogenic inflammation at sites distant from initial infection. Using a similar ApoE^−/− model, an independent study confirmed the causal link between periodontitis and atherosclerosis and, moreover, detected P. gingivalis in aortic tissues, including vascular endothelial cells, using fluorescent in situ hybridization²⁸.

Immune subversion and atherogenic potential

Although periodontal bacteria have been identified in atheromas^25-27,92-94, the mode of their relocation is uncertain and could involve mechanisms alternative or additional to the bacteraemic route. In principle, bacteria might exploit recirculating leukocytes — such as macrophages and/or dendritic cells (DCs) — as ‘Trojan horses’ for dissemination to systemic tissues. In this regard, the ability of P. gingivalis to survive intracellularly in macrophages^29,102 and DCs²⁷ is intriguing. To persist intracellularly in macrophages, P. gingivalis needs to enter the macrophage via complement receptor-3 in cholesterol-rich lipid rafts^102,103, which is consistent with observations that pathogens which invade through lipid rafts are not readily directed to late endosomes and lysosomes where they would be killed¹⁰⁴. To survive within DCs, P. gingivalis needs to enter via the C-type lectin DC-specific ICAM-3 grabbing non-integrin (DC-SIGN)^105,106(FIG. 4). Both processes are mediated by fimbrial proteins of P. gingivalis: FimA fimbriae directly interact with complement receptor-3 on macrophages, whereas Mfa1 fimbriae interact with DC-SIGN on DCs^102,106.

Figure 4. Microbial immune subversion in atherogenesis.

In addition to a bacteraemic route, periodontal bacteria may hijack leukocytes or erythrocytes (to which they attach via a C3b–CR1 interaction) to disseminate from the oral mucosa to aortic tissues. Bacteria not only invade but also activate endothelial cells (upregulation of cell adhesion molecules and chemokines) in ways that promote the transmigration of leukocytes that may harbour viable intracellular bacteria. The bacteria can spread to deeper tissues where they can induce smooth-muscle-cell proliferation in the intima. The uptake of low-density lipoprotein (LDL) by transmigrated macrophages is enhanced in the presence of bacteria leading to accelerated foam cell formation and atherogenesis. At later stages, atherosclerotic plaque rupture can be facilitated by bacterially induced production of matrix metalloproteinases (MMPs) by lymphocytes or myeloid cells. Bacteria-induced platelet aggregation (directly or through the induction of prothrombotic autoantibodies) may contribute to thrombotic vessel occlusion. Most of these studies supporting the above-discussed model utilized Porphyromonas gingivalis as model pathogen, the survival of which within leukocytes depends in part upon Toll-like receptor 4 (TLR4) evasion as well as on its capacity to exploit CR3 (in macrophages) or DC-SIGN (in dendritic cells) for safe intracellular entry. Abs, antibodies; CR1, complement receptor-1; CR3, complement receptor-3; DC, dendritic cell; DC-SIGN, DC-specific ICAM-3 grabbing nonintegrin; SMC, smooth muscle cell; TLR4, Toll-like receptor 4.

The exploitation of DCs as transport vehicles for P. gingivalis is supported by a recent clinical study that identified P. gingivalis (using 16S rDNA sequencing) within blood myeloid DC from chronic periodontitis patients²⁷. The rate and frequency of DC carriage of P. gingivalis was increased after bacteraemia elicited by debridement, which is a routine clinical procedure involving removal of dental plaque biofilm and calculus from the teeth and gingiva. Moreover, immunofluorescence analysis revealed co-localization of P. gingivalis with myeloid DCs that are infiltrating either the oral mucosa or atherosclerotic plaques of patients with chronic periodontitis²⁷.

Mechanistic underpinnings for these clinical observations were provided by in vitro studies showing that P. gingivalis subverts human DC function^27,105,106. Specifically, following binding to DC-SIGN, P. gingivalis not only enters and survives within myeloid DCs, but also promotes an atherogenic phenotype in the cells, as indicated by upregulation of matrix metalloproteinase-9 and complement C1q which are indicators of plaque rupture risk²⁷. Furthermore, P. gingivalis selectively upregulates the expression of chemokine receptor CXCR4, but not that of CC-chemokine receptor 7 (CCR7), and thereby disrupt DC migration toward CC-chemokine ligand 19 (CCL19) and promote migration toward CXC-chemokine ligand 12 (CXCL12)¹⁰⁷. Consistently, peripheral blood myeloid DCs from patients with chronic periodontitis are characterized by high CXCR4 and low CCR7 expression compared with DCs from healthy individuals¹⁰⁷. CCR7 mediates homing to secondary lymphoid organs, whereas CXCR4 mediates homing to sites of neovascularization such as atherosclerotic plaques^108,109; therefore, the authors suggested that P. gingivalis hijacks and directs DC migration to inflammatory vascular sites, where the pathogen can exacerbate inflammatory pathology¹⁰⁷.

An alternative transport vehicle for P. gingivalis in the blood might be erythrocytes, which bind C3b-opsonized P. gingivalis in a complement receptor-1-dependent manner¹¹⁰(FIG. 4). Although leukocytes fail to bind and internalize erythrocyte-attached P. gingivalis, this may not necessarily be an evasion strategy since erythrocyte-associated bacteria can eventually be cleared by liver macrophages (Kupffer cells). However, because the interaction of P. gingivalis with erythrocytes gradually declines over time — which is attributed to bacterial degradation of bound C3b — resulting in the release of viable bacteria in vitro, the authors speculated that erythrocyte-carried P. gingivalis can potentially infect endothelial cells at extraoral sites, and thereby contribute to vascular inflammation¹¹⁰(FIG. 4). This intriguing hypothesis remains to be tested in animal models.

P. gingivalis can indeed invade human aortic endothelial cells by means of its FimA fimbriae and subsequently suppresses the levels of key intracellular molecules involved in cell death and host defence (such as NLRP3 and RIPK1) leading to a permissive intracellular environment^6,94,111,112. FimA fimbriae also induce TLR2-dependent expression of endothelial adhesion molecules (ICAM-1, VCAM-1 and E-selectin) and chemokines (MCP-1 and CXCL8) involved in leukocyte recruitment, as well as TLR2 inside-out signalling that activates leukocyte integrins that mediate transendothelial migration^6,113,114. However, additional TLR2 ligands of P. gingivalis, including serine lipids and lipoproteins^115,116, probably also participate in these pro-atherogenic activities.

Other studies using P. gingivalis as a pro-atherogenic model bacterium showed that it can accelerate atherothrombosis via the recruitment and activation of neutrophils¹¹⁷ and can induce platelet aggregation — an activity facilitated by interactions of P. gingivalis hemagglutinins with the platelet glycoprotein IIb–IIIa (also known as αIIbβ3 integrin) on platelets via a fibrinogen bridge¹¹⁸. P. gingivalis may contribute to cardiovascular disease pathology also via molecular mimicry. Specifically, compared with healthy individuals, a substantial subset of patients with periodontitis have elevated gingival crevicular fluid and serum concentrations of cardiolipin-specific antibodies, that is, prothrombotic autoantibodies associated with atherosclerosis and also adverse pregnancy outcomes (see discussion below). Such autoantibodies can be induced in response to bacterial epitopes — such as those found in P. gingivalis arginine-specific gingipains and T. denticola phosphoglycerate kinase — that bear homology to the TLRVYK peptide of the phospholipid-binding serum protein β2-glycoprotein-I^119,120. Furthermore, P. gingivalis enhances low-density lipoprotein-uptake and foam cell formation by upregulating CD36 (a scavenger receptor mediating lipid uptake) and downregulating ATP-binding cassette transporter A1 (which mediates the efflux of cholesterol from macrophages)^121,122. In a similar context, the pathogen induces vascular smooth-muscle-cell proliferation leading to neo-intima formation, and stimulates production of matrix metalloproteinases that contribute to plaque rupture and thrombotic vessel occlusion⁶(FIG. 4).

In summary, periodontal pathogens evade the immune system to induce atherogenic inflammation, although their impact beyond the oral cavity is not restricted to the cardiovascular system but includes additional tissues (see below).

Periodontitis and adverse pregnancy outcomes

Epidemiological and clinical studies suggest that maternal periodontitis may be associated with increased risk of adverse pregnancy outcomes, such as low birthweight, pre-term birth, miscarriage and/or stillbirth^9,21,123. Similar to atherosclerosis, two major plausible biological mechanisms have been proposed: firstly, periodontal pathogens that disseminate systemically may cross the placenta into the fetal circulation and amniotic fluid, and secondly, inflammatory mediators produced locally in the periodontium could enter the systemic circulation and stimulate an acute-phase response and thereby adversely affect the placenta and fetus (FIG. 3). The notion that periodontal bacteria can cause pregnancy complications is supported by mechanistic studies in animal models.

F. nucleatum — a potential accessory pathogen that facilitates the colonization of periodontitis-associated bacteria — becomes an overt pathogen if it translocates to extraoral sites⁸. Indeed, the bacterium can be isolated from abscesses in several internal organs and is implicated in adverse pregnancy outcomes and colorectal cancer (BOX 2). Clinical studies have linked F. nucleatum to several pregnancy complications, including premature birth, stillbirth and neonatal sepsis⁸. Moreover, the identification of identical F. nucleatum clones in the subgingival biofilm of a mother with pregnancy-associated gingivitis and her stillborn infant suggest that the bacterium can disseminate from the mother’s gingiva to her uterus²¹. Experiments in pregnant mice have provided insights into how F. nucleatum can cause intra-uterine infection and inflammation. Specifically, intravenously administered F. nucleatum (to mimick bacteraemia) uses its E-cadherin-binding FadA adhesin to cross the endothelium and colonize the fetal–placental compartment, where it induces TLR4-dependent necroinflammatory responses¹²⁴. In contrast to F. nucleatum, E. coli fails to induce fetal loss in the same model¹²⁵. Whereas F. nucleatum can colonize the placenta of TLR4-deficient mice, these mice display substantially decreased fetal death rate compared with wild-type mice, which indicates that fetal loss is caused by inflammation rather than bacterial colonization per se¹²⁴.

Box 2. Oral bacteria and cancer.

Accumulating evidence indicates a temporal and spatial association between oral bacteria and cancer^7,157. Fusobacterium nucleatum is associated with colorectal cancer¹⁵⁸, which has been attributed to its ability to stimulate the growth of colorectal cancer cells¹⁵⁹. This activity depends upon interactions between E-cadherin and the F. nucleatum adhesin FadA, the expression of which (FadA) is elevated in the cancerous colon tissue and is correlated with the expression of oncogenic and inflammatory genes¹⁵⁹. Porphyromonas gingivalis is associated with oral squamous cell carcinoma (OSCC)^160,161, orodigestive cancer (independently of periodontitis)¹⁶² and pancreatic cancer¹⁶³. Certain immune subversive mechanisms of P. gingivalis are consistent with a role in cancer development (see figure). In OSCC cells, P. gingivalis induces the expression of pro-matrix metalloproteinase-9 (MMP-9) by triggering proteinase activated receptor-2 (PAR-2)-mediated NF-κB activation (extracellular mechanism involving gingipain secretion) or by activating ERK1/2 and p38 MAPK pathways (intracellular mechanism requiring β1-integrin-dependent invasion)¹⁶⁴. In addition, the gingipains additionally cleave the secreted proenzyme into mature MMP-9 (gelatinase), which promotes carcinoma cell migration¹⁶⁴. P. gingivalis invasion of epithelial cells suppresses apoptosis and stimulates cell proliferation by inhibiting the p53 tumor suppressor¹⁶⁵. The ability of P. gingivalis to activate Janus kinase 1 (JAK1) and signal transducer and activator of transcription 3 (STAT3) pathway as well as the phosphoinositide 3-kinase (PI3K)-AKT signalling causes inhibition of intrinsic mitochondrial apoptosis pathways^149,166. Specifically, STAT3, which upregulates the anti-apoptotic molecules survivin and BCL-2, and AKT, which inhibits the pro-apoptotic molecule BCL-2-associated agonist of cell death (BAD), lead to caspase-3 inhibition^148,149(inset). Moreover, extracellular release of nucleoside diphosphate kinase (NDK) by P. gingivalis cleaves ATP and prevents induction of apoptosis via the purinergic receptor P2X₇¹⁶⁷. Although suppressor of cytokine signalling 3 (SOCS3) can induce apoptosis by targeting STAT3¹⁶⁸, P. gingivalis upregulates miR-203 which directly inhibits SOCS3¹⁶⁹. P. gingivalis can additionally induce B7 family ligands (B7-H1, B7-DC) that interact with the programmed death-1 receptor on T cells¹⁷⁰ potentially leading to their immunosuppression¹⁷¹.

Certain other periodontal bacteria, including P. gingivalis, were also shown to colonize the placenta and fetal tissues of mice or rats and thereby causing inflammation and pregnancy complications (reviewed in REF.9). P. gingivalis may also induce adverse pregnancy outcomes via alternative mechanisms. Cardiolipin-specific antibodies are associated with certain disorders, including adverse pregnancy outcomes¹¹⁹, and a subset of periodontitis patients have increased concentrations of such autoantibodies¹²⁶, which could be induced in response to cross-reactive bacterial epitopes such as the arginine-specific gingipains of P. gingivalis¹¹⁹. The possible connection involving P. gingivalis, cardiolipin-specific antibodies and pregnancy complications was strengthened recently. A study showed that antibodies raised against P. gingivalis — but not against an isogenic gingipain-deficient mutant — cause fetal loss when passively administered to pregnant female mice¹²⁷. Importantly, this effect was substantially inhibited when cardiolipin-specific antibodies were removed from the antibody preparations¹²⁷.

In summary, pregnancy complications can be caused by periodontal pathogen-induced inflammatory responses or autoantibodies, which have also been implicated in the association of periodontitis with rheumatoid arthritis.

P. gingivalis and rheumatoid arthritis

Several studies indicate an epidemiological association between periodontitis and rheumatoid arthritis, even after adjusting for common risk factors such as smoking^128-130. In a distinct clinical phenotype of rheumatoid arthritis, anti-citrullinated protein antibodies (ACPA) serve as diagnostic markers as they are detected in the serum prior to the onset of the disease and their serum levels correlate strongly with disease severity¹³¹. A recent study showed that patients with rheumatoid arthritis — particularly those with ACPA-positive rheumatoid arthritis — have higher frequency of periodontitis than control patients with osteoarthritis. Furthermore, in these patients the detection of P. gingivalis in subgingival biofilm samples was associated with increased levels of ACPA irrespective of smoking status³². The molecular underpinnings of these associations are examined below.

Peptidyl-arginine deiminase

P. gingivalis is unique among other periodontal bacteria, and possibly among all prokaryotes, with regard to expression of a peptidyl-arginine deiminase, which is an enzyme that converts protein arginine residues to citrulline^5,33. Protein citrullination has been implicated in several physiological processes¹³²; however, because citrullination can dramatically alter protein structure and function, dysregulated host protein citrullination by microbial enzymes could interfere with normal host cell signalling and immune or other homeostatic functions. For instance, P. gingivalis peptidyl-arginine deiminase (PPAD) citrullinates the C-terminal arginine of epidermal growth factor (EGF), and thereby inhibits its biological activity as shown by impaired EGF-induced fibroblast proliferation and migration¹³³. Interestingly, citrullination of two internal arginine residues of EGF by human peptidyl-arginine deiminase enzymes does not abrogate EGF function¹³³. As EGF has an essential role in wound healing and tissue regeneration, its inactivation by P. gingivalis may interfere with periodontal tissue healing and thus delay the resolution of inflammation.

Link between P. gingivalis and rheumatoid arthritis

The unique ability of P. gingivalis to citrullinate proteins has attracted considerable interest in the field of rheumatoid arthritis given the importance of ACPA in its pathogenesis^5,131. In this context, P. gingivalis was shown to citrullinate human fibrinogen and α-enolase, which, in their citrullinated form, are two major rheumatoid arthritis autoantigens³¹. This activity requires concerted action between PPAD and arginine-specific gingipains, which co-localize with PPAD in the outer membrane of P. gingivalis³¹. Specifically, the cleavage of fibrinogen or α-enolase by gingipains exposes C-terminal arginine residues that are subsequently citrullinated by PPAD³¹. In principle, the unique mode of proteolytic processing and post-translational modification of host antigens by P. gingivalis could generate neoepitopes to which immunologic tolerance does not exist, leading to the generation of autoantibodies (FIG. 5). It should be noted that the breakdown of immune tolerance to citrullinated proteins requires susceptible individuals, such as carriers of HLA-DRB1 shared epitope alleles that bind selectively to citrullinated sequences and may influence antigen presentation in ways that lead to ACPA production¹³⁴. Intriguingly, rheumatoid arthritis-specific autoantibodies to citrullinated α-enolase peptide 1 (the immunodominant B cell epitope of human α-enolase) were shown to cross-react with citrullinated enolase from P. gingivalis, suggesting that molecular mimicry can contribute to autoantibody generation¹³⁵.

Figure 5. Porphyromonas gingivalis-mediated citrullination and induction of ACPA in rheumatoid arthritis.

P. gingivalis peptidylarginine deiminase (PPAD) citrullinates host-derived or bacterial proteins in the inflammatory environment of periodontitis. In susceptible individuals (carriers of HLA-DRB1 shared epitope (SE) alleles), distinct citrullinated peptides are presented in the context of HLA-DRB1 SE to activate T cells, which, in turn stimulate B-cell production of anti-citrullinated protein antibodies (ACPA). The induction of autoantibodies may be explained by mechanisms involving neoepitope formation or molecular mimicry. Citrullination of host proteins, such as α-enolase, fibrinogen and collagen type II, by human peptidylarginine deiminase (PAD) enzymes can occur in injured or inflamed joints. ACPA bind citrullinated proteins and form immune complexes that can mediate local synovial inflammation by activating complement or Fcγ receptors (FcγR).

A ‘two-hit’ model of rheumatoid arthritis pathogenesis was proposed involving initial breakdown of tolerance to citrullinated peptides generated by P. gingivalis in inflamed gingiva followed by epitope spreading to other host-citrullinated proteins in the inflamed joint^5,31(FIG. 5). In this regard, citrullinated proteins have been detected in the gingiva of periodontitis patients⁵. Moreover, the notion that immunity to citrullinated proteins is initially triggered in inflamed mucosal surfaces distant from the joints is consistent with the presence of ACPA prior to signs of inflammation in the joints¹³⁶. Mechanisms of epitope spreading or molecular mimicry could lead to cross-reactivity with citrullinated joint proteins and subsequent formation of immune complexes could exacerbate or perpetuate the inflammatory process in rheumatoid arthritis through several mechanisms, including activation of complement or Fcγ receptors^5,31(FIG. 5). Epitope spreading precedes the development of rheumatoid arthritis and leads to maintenance and progression of inflammation as it sustains the generation of high-affinity ACPA to host citrullinated proteins^5,137.

The ‘two-hit’ model is consistent with the results of two recent mechanistic studies by independent groups. Specifically, infection of mice with wild-type P. gingivalis exacerbates collagen- or collagen antibody-induced arthritis, as revealed by accelerated progression and enhanced severity of bone and cartilage destruction^30,138. The ability of wild-type P. gingivalis strains to aggravate arthritis is strictly dependent on PPAD expression, since isogenic mutants lacking this enzyme fail to influence the disease outcome^30,138. Furthermore, infection of mice with wild-type, but not with PPAD-deficient, P. gingivalis was associated with detection of citrullinated proteins at the site of infection and with production of antibodies to citrullinated proteins^30,138. These findings suggest that, by virtue of its PPAD, P. gingivalis may constitute a mechanistic link between periodontitis and rheumatoid arthritis.

Periodontitis and respiratory diseases

Aspiration pneumonia

The tooth-associated bacterial biofilm is thought to be a reservoir for respiratory infections, and oral anaerobic bacteria are common isolates from aspiration pneumonia and lung abscesses^22,139,140. Oropharyngeal aspiration of bacteria is a major cause of pneumonia in old or immunocompromised individuals^139,140 and periodontitis is epidemiologically implicated as a mortality risk factor for aspiration pneumonia at least in the elderly¹⁴¹. Since patients probably aspirate fragments of biofilm composed of mixed bacterial species, the polymicrobial synergistic interactions seen in periodontitis might also occur in the lung tissue. Despite limited research in this area, mixed infection with P. gingivalis and T. denticola in a mouse model of aspiration pneumonia has been shown to cause considerably higher inflammatory responses, impaired bacterial clearance and more severe lung pathology compared with single infection with either bacterium¹⁴². Importantly, the control of the oral microbial burden substantially decreases the incidence of aspiration pneumonia in frail elderly people^143,144, which suggests a direct association between oral bacteria and lung pathology in susceptible individuals. These results warrant more basic studies to understand the mechanisms involved.

Chronic obstructive pulmonary disease

Periodontitis is also associated with chronic obstructive pulmonary disease (COPD)^23,145. Polymicrobial infections including the opportunistic pathogen Pseudomonas aeruginosa are associated with exacerbations of COPD increasing its morbidity and mortality^23,146. P. gingivalis is readily detected with P. aeruginosa in tracheal aspirates of patients with acute COPD exacerbations²⁴ and can enhance the pathogenicity of P. aeruginosa in the context of lower airway infection^147,148. Indeed, P. gingivalis promotes the ability of P. aeruginosa to invade respiratory epithelial cells and modulates its apoptosis-inducing capacity^147,148. Specifically, compared with P. aeruginosa invasion alone, co-invasion of respiratory epithelial cells with both bacteria leads to diminished apoptosis as a result of enhanced signal transducer and activator of transcription 3 (STAT3) signalling, which upregulates the expression of the anti-apoptotic molecules survivin and B cell lymphoma 2 (BCL-2), in turn leading to caspase 3 inhibition¹⁴⁸. Moreover, co-invasion is followed by downregulation of the pro-apoptotic molecule BCL-2-associated agonist of cell death (BAD)¹⁴⁸, possibly mediated by phosphoinositide 3-kinase (PI3K)–AKT signalling¹⁴⁹(see Figure in BOX 2). However, the inhibition of apoptosis is transient since the signalling pathways involved start to decline 8 hours post-invasion of both bacteria, leading to dramatically increased caspase-3 activity by 12 hours post-invasion¹⁴⁸. Rapid apoptosis of infected epithelial cells is thought to contribute to effective clearance of P. aeruginosa¹⁵⁰. This notion suggests that inhibition of epithelial cell apoptosis for a substantial time following P. aeruginosa and P. gingivalis co-invasion may provide the bacteria with a safe intracellular niche and the opportunity to proliferate and establish infection.

Conclusions and perspective

Dysbiotic microbial communities in the periodontium resist immune elimination and create permissive conditions for growth in a nutritionally favourable inflammatory environment^{11,13,17-19,41,42}(FIG. 1-2). The immune-subversive and pro-inflammatory strategies that promote the fitness of periodontal bacteria not only cause collateral damage to the periodontium but also have repercussions that link periodontitis to systemic afflictions (FIG. 3-5). The virulence of individual periodontal pathogens is maximized in the context of a polymicrobial infection^{17,19,47-51,53} and its impact on the host depends on genetic predispositions and environmental modifiers^92,151-155(BOX 1). Hence, to better understand mechanisms of pathogenesis of periodontitis and associated systemic conditions, data from epidemiological and animal model studies need to be meaningfully integrated with those from metatranscriptomic and metaproteomic approaches as well as whole-genome transcriptomic and proteomic analyses of host tissue in health and different disease stages. This integration can offer insights into the dynamic nature of host–microbe interactions in disease development, and, moreover, can facilitate the formulation of novel hypotheses for further knowledge discovery. More importantly, key findings from basic research need to be translated into clinical applications for host-modulation therapies to counteract the immune-subversive mechanisms of periodontal bacteria, and thereby contribute to the treatment for periodontitis and associated systemic inflammatory disorders. Such host-modulation strategies are more likely to succeed than direct antimicrobial approaches — especially when targeting keystone pathogens as they can act at low abundance and will probably not be completely eradicated, partly because they can hide within permissive host cells.

Glossary terms

Microbiota

A complex and diverse community of microorganisms living within a given anatomical niche, for example an environmentally exposed surface of a multicellular eukaryotic organism.

Dysbiosis

A condition characterized by an imbalance in the relative abundance or influence of species within a microbial community associated with disease, for instance periodontitis or inflammatory bowel disease.

Homeostasis

A condition of equilibrium or stability in a system maintained by adjusting physiological processes to counteract external changes, such as a balanced relationship between host tissues and the resident microbiota that prevents destructive inflammation or disease.

Subgingival crevice

Narrow space between the tooth surface and the free gingiva.

Gingival crevicular fluid

Serum exudate that originates in the gingival capillaries and flows into the gingival crevice carrying locally produced immune and inflammatory mediators such as complement, cytokines and antimicrobial peptides.

Keystone pathogen

A pathogen with a disproportionately large effect on its environment relative to its abundance, for example low-abundance P. gingivalis remodels a commensal microbial community into a dysbiotic and disease-provoking microbiota.

Pathobiont

Commensal with the potential to induce pathology under conditions of disrupted homeostasis.

Accessory pathogen

A commensal bacterium that is not pathogenic by itself in a given niche, but which can enhance the virulence of keystone pathogens by, for example, facilitating their colonization or providing metabolic support.

Gingipains

A family of trypsin-like cysteine proteinases which are secreted by P. gingivalis and contribute to its virulence and the pathogenesis of periodontitis. Members include the high molecular mass arginine-specific gingipain A (HRgpA), arginine-specific gingipain B (RgpB) and lysine-specific gingipain (Kgp).

Inflammophilic

Refers to bacteria that thrive on inflammation as they feed off inflammatory tissue breakdown products; literally meaning attracted to inflammation, from the combined meaning of inflammation and the Greek suffix philic denoting fondness.

Periodontal pocket

The pathologically deepened subgingival crevice in periodontitis, which is pathognomonic for the disease.

Inflammasome

A cytosolic, multiprotein complex which responds to infection or tissue injury by activating pro-inflammatory caspases, mainly caspase-1, leading to the cleavage and release of pro-inflammatory cytokines such as IL-1β and IL-18 and — under certain conditions (myeloid cells infected with pathogenic bacteria) — to pyroptosis, a form of necrotic cell death.

Non-canonical inflammasome

A caspase-11–dependent pathway of inflammasome activation which is critical for controlling infection by Gram-negative bacteria and can induce cell death (pyroptosis) independently of caspase-1.

Atheroma

Accumulated fatty deposits in the inner lining (intima) of an artery, leading to restriction of blood flow and a risk of thrombosis.