Porphyromonas gingivalis–dendritic cell interactions: consequences for coronary artery disease

Abstract

An estimated 80 million US adults have one or more types of cardiovascular diseases. Atherosclerosis is the single most important contributor to cardiovascular diseases; however, only 50% of atherosclerosis patients have currently identified risk factors. Chronic periodontitis, a common inflammatory disease, is linked to an increased cardiovascular risk. Dendritic cells (DCs) are potent antigen presenting cells that infiltrate arterial walls and may destabilize atherosclerotic plaques in cardiovascular disease. While the source of these DCs in atherosclerotic plaques is presently unclear, we propose that dermal DCs from peripheral inflamed sites such as CP tissues are a potential source. This review will examine the role of the opportunistic oral pathogen Porphyromonas gingivalis in invading DCs and stimulating their mobilization and misdirection through the bloodstream. Based on our published observations, combined with some new data, as well as a focused review of the literature we will propose a model for how P. gingivalis may exploit DCs to gain access to systemic circulation and contribute to coronary artery disease. Our published evidence supports a significant role for P. gingivalis in subverting normal DC function, promoting a semimature, highly migratory, and immunosuppressive DC phenotype that contributes to the inflammatory development of atherosclerosis and, eventually, plaque rupture.

• dendritic cells
• periodontitis
• atherosclerosis
• Porphyromonas gingivalis
• DC-SIGN

Introduction

The pathological manifestations of chronic periodontitis (CP), namely, destruction of the soft and hard tissues that support the dentition culminating in tooth loss, have been previously described, as has the high prevalence of the CP in the general adult population 1 2. Also well documented is the specific role of the anaerobic Gram-negative species, Porphyromonas gingivalis, in infection of the tissues around the dentition in CP and in initiation of CP 3. P. gingivalis, along with two other species, Tannerella forsythia and Treponema denticola, comprise the so-called red complex of pathogens 4 that function cooperatively within the subgingival plaque of CP 1 2 5 6 7 . P. gingivalis has been associated with several important systemic diseases such as cardiovascular disease, rheumatoid arthritis, preterm birth weight, and diabetes mellitus 8 9. Our group is particularly interested in the influx and efflux of various dendritic cells (DCs) in response to P. gingivalis and their role in local and systemic inflammatory processes. DCs are very active antigen-capture cells when immature. When DCs mature, they become potent antigen-presenting cells and are very efficient at stimulating T-naïve cells to differentiate into T-cell effectors [reviewed in Steinman 10, Cutler 11]. DCs have been implicated in a number of allergic and inflammatory diseases in the periphery [reviewed in Velazquez and Teran 12] including atherosclerosis 13 14. Being central to the development of immunologic memory and tolerance 15, the normal immunologic function of DCs is to patrol the periphery, capture infecting microbes and then migrate out to the secondary lymphoid organs where they can initiate and regulate the adaptive immune response. DC functions in this regard are tightly regulated and depend on the activation signals that DCs receive in the periphery [reviewed in Cutler and Jotwani 16]. These signals include inflammatory cytokines, chemokines, as well as pathogen-associated molecular patterns (PAMPS) of bacteria. The development of gingivitis and periodontitis involves the influx and efflux of different DC subsets at distinct stages of disease (Figure 1). Dendritic Langerhans cells (LC) infiltrate the gingival epithelium in gingivitis and then efflux into the lamina propria in CP, where they begin to undergo maturation 17. Dermal dendritic cells (DC-SIGN + ) increase in the lamina propria in CP and become localized toward the lymphatics and vasculature 18 19 20 . LCs have been implicated in both the initiation and regulation of contact-hypersensitivity responses in mice 21 [reviewed in Igyarto and Kaplan 22], while dermal DCs have been implicated in many other inflammatory diseases including rheumatoid arthritis and inflammatory bowel disease 23. This review will focus on P. gingivalis and DCs and their respective contributions to the development (and instability) of atherosclerotic plaques.

Fig. 1.  Schematic representation of how P. gingivalis-laden dendritic cells may promote atheroma formation and maturation. (A) Representation of the oral lymphoid foci, its organized inflammatory infiltrate. The left panel depicts the healthy oral biofilm in the gingival crevice, comprised predominantly of Gram-positive bacteria (green dots). Healthy gingival tissue is infiltrated with numerous Langerhans cells in the epithelium, with sparse dermal dendritic cells in the lamina propria. As disease progresses, the oral biofilm changes to a predominantly Gram-negative subgingival flora (red dots). In response, a dramatic loss (efflux) of Langerhans cells occurs from the epithelium toward the lamina propria. Also observed is an influx of myeloid-derived DC-SIGN+ dermal dendritic cells (DCs) into the lamina propria. Present are neutrophils, macrophages (mφ), B-cells, and CD4+ and CD8+ T-cells. The DCs form immune conjugates with CD4+ T-cells and also mobilize toward the vasculature. (B) Hypothetical model in periodontitis, showing mobilization of P. gingivalis-laden DC-SIGN+ myeloid DCs (mDCs) in the gingival lamina propria. These mDCs undergo reverse transmigration through the vascular endothelium. Once in the circulation, mDCs carrying P. gingivalis attach to endothelial integrins via DC-SIGN and, after rolling adhesion, undergo diapedesis between endothelial cells. As the atheroma continues to mature and DCs contribute to the foam cells and release MMP-9, the atheroma becomes highly unstable and, eventually, thrombus formation occurs.

Atherosclerosis and microbes

Atherosclerosis (ATH) is a progressive disease characterized by the accumulation of lipids, fibrous elements, and inflammatory cells in the large arteries. ATH constitutes the single most important contributor to the growing worldwide burden of cardiovascular disease. Only about 50% of patients with ATH have currently identified risk factors 24. This suggests how little we know about ATH risk. Inflammation in the arterial vessel wall is particularly important in the development of ATH. Four mechanisms have been proposed for how bacterial pathogens may induce or accelerate ATH [reviewed in Gibson et al. 25]. These include: (i) direct invasion of the vascular endothelium by pathogens in the blood, (ii) immunological sounding, (iii) molecular mimicry, and (iv) pathogen trafficking of microbes within leukocytes in peripheral blood.

Of particular relevance to atherogenesis are bacterial species that infect and survive within endothelial cells and within migrating leukocytes [reviewed in Rackley 26]. A prevailing hypothesis is that, regardless of their viability status, bacteria release PAMPs that serve as agonists for TLRs, thus activating inflammatory leukocytes and endothelial cells, and can contribute to the development of ATH 27. Large population studies support the role of bacterial species in ATH 28 29. However, the results of clinical trials using antibiotics to treat cardiovascular disease have been disappointing 30 31 32 33 . Many atherogenic bacteria, including P. gingivalis, are intracellular pathogens 34 35. An apparent consequence of this is that these pathogens are less susceptible to antibiotics when sequestered inside host cells. P. gingivalis is 100-fold more resistant to moxofloxocin, 10-fold more resistant to clindamycin and metronidazole when inside host epithelial cells 36. Clindamycin- and azithromycin-resistant P. gingivalis isolates have been identified in human subjects with CP 37.

P. gingivalis has also been identified in human ATH plaques 38, as have other atherogenic bacteria such as Chlamydophila pneumoniae and Helicobacter pylori 27 39 40. Experimental infection with P. gingivalis accelerates ATH in animal models 35 41. The FINRISK 1992 cohort study of 6,051 individuals implicates exposure to P. gingivalis or endotoxin in increased risk for cardiovascular diseases 42. A recent meta-analysis indicated that the level of systemic bacterial exposure in CP mediates ATH risk 43 44. Another meta-analysis studied human cohort studies, case-control studies, and cross-sectional studies and concluded that CP is a significant risk factor for developing coronary artery disease (CAD) 44. The degree of increased risk of CAD conferred by CP appears comparable to smoking 45 and elevated serum triglycerides 46. Viable and invasive P. gingivalis, though in a dormant state, have been cultured from human ATH plaques 38. This has not been shown with other atherogenic bacteria such as C. pneumoniae and H. pylori, which have only been identified by DNA-based methods 27 39 40. Overall, several infectious agents have been shown to be disseminated by pathogen trafficking leukocytes, they include Streptococcus pyogenes 47, C. pneumoniae 48, and Listeria monocytogenes 49.

P. gingivalis virulence and targeting of dendritic cells (DCs)

P. gingivalis is a an amino acid fermentor with an absolute requirement for hemin 2. The bacteria utilizes its many virulence factors to fulfill its complex nutritional requirements, while still enabling it to evade and even modulate the host immune system 1 2 5. Several virulence factors including the polysaccharide capsule, fimbriae, proteases for opsonins C3, proteases for IgG, gingipains, bacterial lipopolysaccharides (LPS), toxins, and hemagglutinins enable P. gingivalis to persist in the oral mucosa and help facilitate some of the physiopathology of CP 19 50 51 52 53 54 55 56 .

There are two fimbriae that are essential adhesins for the invasion and colonization of the oral mucosa by P. gingivalis. These adhesins, termed the major and minor fimbriae, are distinct antigenically by amino acid composition and by size 57 58. The major fimbriae form long projections from the bacteria and have been shown in most reports to facilitate their adhesion to and invasion of the host cells. The major fimbriae is a 41 kDa protein, encoded by the fimA gene 59, and many of its cellular receptors have been identified as either the β-1 integrins (CD29) 60 61 or the β-2 integrins (CD18) 62 63 64 . The minor fimbriae (though much shorter) are comprised of a 67 kDa protein encoded by the mfa1 gene 57. We have recently shown that the minor fimbriae targets DC specific ICAM-3 grabbing non-integrin (DC-SIGN or CD209) on monocyte-derived DCs for entry 5. DC-SIGN is a type II membrane protein in which the extracellular domain consists of a stalk that promotes tetramerization 65. It contains a C-terminal carbohydrate recognizing domain (CRD) that belongs to the C-type lectin superfamily 65. Early studies by Feinberg et al. showed that the DC-SIGN CRD preferentially binds to the high-mannose N-linked oligosaccharides GlcNAc (N-acetylglucosamine) and Manα1-3[Manα1-6] Man (mannose) 65. Furthermore, Appelmelk et al. showed that DC-SIGN also binds to fucose-containing Lewis blood antigens 66. Guo et al. utilized an extensive glycan array and showed that DC-SIGN will bind high mannose-containing glycans or glycans that contain terminal fucose residues 67. Previous studies showed that DC-SIGN is used by microorganisms such as Neisseria gonorrhoeae 68, Mycobacterium tuberculosis 66 69 70, Mycobacterium leprae, HIV 71, H. pylori 66, and P. gingivalis 5 to target DCs for entry and immune suppression.

Pathological consequences of DC-SIGN targeting

The ability of minor fimbriae to specifically target DC-SIGN on DCs has significant pathological and immunological repercussions. Periodontitis lesions contain an intense infiltrate of DC-SIGN+ DCs 17 72. DCs in the lesions become activated and appear to mobilize toward the capillary-rich lamina propria 17. We propose that infection of DCs by P. gingivalis triggers reverse transmigration of infected gingival DCs into circulation and these may contribute to the pathogenesis of ATH (Figure 1). Evidence exists for the presence of activated DC-SIGN+ myeloid DCs in rupture prone unstable plaques 13 73. The sources of these DC-SIGN+ plaque DCs are not clear, but presumably may include DCs from inflamed peripheral tissues. Other sources of DCs may include CD14+ CD16− monocytes, CD14low CD16+ monocytes 74 that differentiate into DCs in situ 5. DCs that infiltrate rupture-prone atherosclerotic plaques express atherogenic markers including C1q (a classical complement pathway component involved in apoptotic cell clearance); HSP60 and HSP70 (chaperone proteins involved in autoimmune responses); chemokine receptors CCR2, CCR5, CX3CR1; and chemokines CXCL16, CCL19, and CCL21 (involved in DC transmigration and leukocyte homing) 74 75 76 . Also expressed are DC maturation markers CD40, CD80, and CD86 73 77. Matrix-metalloproteinase-9 (MMP-9), which are highly expressed in vulnerable regions of the atherosclerotic plaques and suggested to be causally involved in plaque rupture 78, is produced by leukocytes including DCs. The P. gingivalis LPS is a particularly potent inducer of MMP-9 but not TIMP-1 by DCs, suggesting that P. gingivalis induces an MMP-9/TIMP-1 imbalance in DCs 79. Our published data indicated that fimbriated strains of P. gingivalis infect DCs and induce atherogenic biomarkers in vitro, but the mechanisms are not presently clear 5.

Immunological consequence of targeting DC-SIGN on dendritic cells (DCs)

DC-SIGN targeting by minor fimbriae results in a dampening of the maturation status and the inflammatory cytokine profile of DCs. Conversely, removal of the minor fimbriae results in robust DC maturation coupled with a strong proinflammatory cytokine response 5. This regulation of DC immunogenic functions based on minor fimbriae expression extends to the T-cell effector response elicited by DCs. In DC-CD4+ T-cell coculture experiments, P. gingivalis strains expressing solely the minor fimbriae induced DCs to prime T-cells into a Th₂ effector phenotype, whereas, strains expressing solely the major fimbriae induced DCs to prime T-cells into a Th₁ effector phenotype 5. Intriguingly, the wild-type strain was able to stimulate a mixed or anergic T-cell effector phenotype 5. Furthermore, in vitro studies in the presence of DC-SIGN targeting agonists (e.g. mannan from S. cerevisiae or glycosylated HIV gp120) resulted in a diminished association of minor fimbriated strains with cells 5. When the wild-type strain of P. gingivalis was cocultured with glycosylated HIV gp120, we observed a dramatic uncoupling of DC maturation from DC inflammatory cytokine secretion 5. The presence of HIV or even Candida albicans (etiological agent of oral thrush) might act synergistically with P. gingivalis to further exasperate this uncoupling and act to greatly diminish DC maturation.

Mechanisms of DC mobilization, access to peripheral blood

Experimental studies demonstrate the important role that chemokines and chemokine receptors play in trafficking of leukocytes to and invasion of the arterial wall in ATH 80 81 82 83 84 . Immature DCs express inflammatory chemokine receptors (Table 1) that direct their migration into infected tissues. In response to capture of antigens or to TLR-mediated recognition of microbes, DCs undergo a process called functional maturation in which they downregulate inflammatory chemokine receptors and upregulate homeostatic chemokine receptors (Table 1). This directs DC migration out of the tissues toward lymph nodes. When the DC maturation process is disrupted, as occurs in DC-SIGN ligation 5, activated DCs ostensibly undergo reverse transmigration into the blood (Figure 1), which can lead to systemic inflammation. This process contributes to the initiation, progression, and instability of arterial plaque in patients with CAD. Recent studies have investigated the presence of blood DCs in patients with CAD, but the results are controversial 85 86. Shi et al. found that the total peripheral blood CD11c + DCs were significantly higher in patients with CAD compared to healthy controls 85. Conversely, Yilmaz et al. found a decrease in circulating DCs in patients with CAD 86. However, the patient population and markers used for DC isolation were different. None of the studies reported the presence of CP on these patients, which might have affected these conflicting findings. An attempt has been made to correlate the presence of blood DCs as a risk factor for CAD 87. Further, much speculation has been made about the source of these DCs that infiltrate the atheromas.

Table 1. Summary of results of immunohistochemistry (IHC), cDNA microarray, and qRT-PCR analyses of MoDCs and gingival tissues

Chemokine receptors (their ligands)	Class	P. gingivalis pulsed MoDC (fold change)	Chronic periodontitis gingival tissue (fold change)
CCR2 (CCL2/MCP-1, CCL7/MCP3, CCL8/MCP2, CCL13/MCP4)	Inflammatory	5.7× ↑^e 10×↑^b	2,000×↑^e
CXCR6 (CXCL16 or SR-PSOX)	Inflammatory	2×↑^e	300×↑^e
CCR5 (CCL3 or MIP-1α, CCL4 or MIP-1β, CCL5 or RANTES, CCL11, CXCL8 or IL-8)	Inflammatory	12×↑^b	20× ↑^c
CXCR3 (CXCL10)	Inflammatory	NT^a	2.2×↑^d
CX3CR1 (CX3CL1)	Inflammatory	NT^a	NT^a
CCR4 (CCL22)	Homeostatic	NT^a	2.3×↑^d
CCR6 (MIP-3α)	Homeostatic/Inflammatory	NT^a	10×↑^e
CCR7 (CCL19 or MIP-3β, CCL20 or LARC)	Homeostatic	3×↑^b	25×↑^e

^aNT, not tested.

^bcDNA microarray (GEArray, Superarray) of P. gingivalis pulsed MoDC (3 h) or gingival tissues from diseased vs. control patients, normalized vs. β Actin.

^c# cells/per field (IHC), 11 healthy vs. 11 control.

^dAffymetrix chip results, normalized against internal standard.

^eqRT-PCR results normalized against GAPDH.

Fully matured DCs lose their ability to uptake and process antigens, stop migrating, and express CD83 and other molecules involved in antigen presentation to T-cells 88. Full maturation of DCs also results in loss of expression of many endocytic receptors, including DC-SIGN; thus, expression of DC-SIGN is one indicator that DC are not fully mature 89. There is evidence that a particular subset of blood DCs expressed DC-SIGN and that this receptor may be involved in the uptake and dissemination of HIV 90. Engering et al. showed that DC-SIGN+ blood DCs are able to stimulate proliferation of allogeneic T-cells, as well as infect these T-cells in trans 90. Potentially, DC-SIGN+ blood DCs can disseminate pathogens, increase systemic inflammation, and contribute to plaque instability.

Environmental regulation of fimbriae: possible role in systemic immunosuppression, dissemination of P. gingivalis

Wu et al. discovered that the major and minor fimbriae are regulated by a two component regulatory system termed FimS/FimR 91. It was also determined that while FimR binds directly to mfa1, it will only bind to the first gene of the fimA gene cluster, pg2130 91 92. Moreover, this two component regulatory system responds to environmental cues like heme and temperature 91. However, it is still not clear how these adhesive fimbriae are regulated in vivo. The possibility of major and minor fimbriae being differentially regulated in response to different environmental cues or stimuli may allow this organism to modulate the immune system to expand its ecological niche. Moreover, our observations of intact bacteria inside DC-SIGN rich vesicles might explain the dissemination of this organism to atherosclerotic plaques 5. We propose here that P. gingivalis interacts with DC-SIGN on dermal DCs from the gingiva mucosa (unpublished observation). This interaction facilitates uptake of P. gingivalis and immunomodulation of normal DC functions. Directed migration is disrupted and DCs then migrate through the endothelium instead of the lymphatic system (Figure 1). Once in the endothelium, the DC undergoes partial maturation resulting in adherence to the endothelium and recruitment of other leukocytes. Soilleux et al. previously described the presence of immature (lacking CD83 but expressing HLA-DR and LAMP) DC-SIGN⁺ DCs on atherosclerotic plaques 73. Strikingly, this maturation profile is very similar to our results with minor fimbriated strains (Pg381 and DPG-3) and DCs 5. Thus, P. gingivalis infected DCs may exit into the bloodstream and migrate to the site of developing atheroma, where they adhere to and invade the arterial endothelium. Recent reports suggest that P. gingivalis is able to spread from infected epithelial, endothelial, and smooth muscle cells to new host cells where it multiplies 93. These mechanisms might explain how an oral opportunistic pathogen is able to disseminate throughout its host and potentially facilitate the formation of atherosclerotic plaques.

Conflict of interest and funding

This study was supported by US Public Health Service grants from the NIH/NIDCR R01 DE014328, R21 DE020916 to Christopher W. Cutler, F31 DE020014 to Amir E. Zeituni, and K23DE018187-01 to Julio Carrion.