Oral Infection With a Periodontal Pathogen Accelerates Early Atherosclerosis in Apolipoprotein E–Null Mice
Objective— Because recent epidemiologic evidence suggests that periodontal infections may increase the risk of atherosclerosis and related events in humans, we assessed the impact of oral inoculation with the periodontal pathogen Porphyromonas gingivalis on atherogenesis in hypercholesterolemic apolipoprotein E–null mice.
Methods and Results— In the absence of alterations in distinct risk factors, P gingivalis infection exacerbated the early stages of atherogenesis in this model. Infected animals displayed evidence of local periodontal infection, as the severity of alveolar bone loss, the hallmark of periodontitis, was increased. Generalized activation of host inflammatory responses was evident in infected mice, as demonstrated by serum IgG response to P gingivalis and elevated levels of interleukin-6. P gingivalis DNA was localized in the aortic tissue from a limited number of infected mice but not in any noninfected controls. Infected mice displayed enhanced vascular activation, as suggested by increased aortic expression of vascular cell adhesion molecule-1 and tissue factor.
Conclusions— Oral infection with P gingivalis accelerates early atherosclerosis. Thus, uncovering the underlying mechanisms is critical for the design of preventive and therapeutic strategies targeting atherosclerotic vascular disease and its sequelae.
Periodontal disease is a chronic infection affecting the tissues surrounding and supporting the teeth, caused by the bacteria of dental plaque. It begins as gingivitis, an inflammation of the soft tissues only, and can progress to periodontitis, where destruction of connective tissue attachment and alveolar bone can eventually lead to tooth loss. The global prevalence of periodontal diseases is high, and the severe forms of chronic periodontitis affect approximately 10% to 15% of individuals in most populations.1 Evidence from recent epidemiologic studies suggests a link between periodontal infections and increased risk of atherosclerosis and related cardiovascular and cerebrovascular events in human subjects.2–5⇓⇓⇓ In some studies, associations were observed between the degree of coronary atherosclerosis and alveolar bone loss or the number of missing teeth, suggesting that the degree of vascular activation correlates with periodontal disease severity.3,4⇓ Hypotheses to account for these observations are multiple, including systemic inflammation primed by periodontal infection and release of lipopolysaccharide into the periphery; activation of circulating inflammatory cells and endothelial cells; direct invasion of the vessel wall by oral bacteria consequent to bouts of transient bacteremia; exaggerated, hyperinflammatory responses to periodontal bacteria or cross-reactivity between microbial and human antigens; and underlying predisposing mechanisms, such as genetic predilection.
See page 1309 and cover
These concepts are of particular relevance, because mechanisms involved in the development and progression of atherosclerosis are complex. It is well established that distinct events, such as hyperglycemia, hyperhomocystinemia, and hypertension, for example, may magnify vascular inflammation and augment processes involved in innate lipid-driven atherogenesis.6–8⇓⇓ In addition to these amplifying factors, multiple studies have suggested a role for infection with a range of organisms, such as Chlamydia pneumoniae, cytomegalovirus, and others.9–11⇓⇓ As the precise role of these infections in the development and progression of atherosclerosis is not fully understood, it is logical to propose that chronic periodontal infections may also participate in this multiple hit model, accelerating mechanisms involved in atherogenesis.
Consistent with this premise, a recent report has suggested that large inocula of the major periodontal pathogen Porphyromonas gingivalis delivered via weekly intravenous injections into the systemic circulation of heterozygous apolipoprotein E (apoE) +/− mice accelerated atherosclerosis.12 In these experiments, P gingivalis and its key components had highly facilitated access to the systemic circulation. As this mode of delivery of the microorganism did not involve oral inoculation, the means by which P gingivalis ordinarily gains access to the tissues in human subjects, the goal of the present studies was to test whether introduction of P gingivalis into the oral cavity would modulate early atherogenesis in mice that spontaneously develop atherosclerosis. To address these concepts, we subjected apoE-null mice,13 in which spontaneous hypercholesterolemia leads to formation of atherosclerotic lesions similar in morphology and distribution to humans, to topical oral inoculation with P gingivalis. Our murine model of experimental periodontal infection is distinct from models that do not introduce P gingivalis into the oral cavity but rather via parenteral routes. Therefore, our studies directly address the role of oral periodontal infections in modulation of atherogenesis.
Oral Infection in Apolipoprotein E–Deficient Mice
All animal studies were performed in accordance with the policies of the Institutional Animal Care and Use Committee at Columbia University. Fifty male homozygous apoE-null mice (backcrossed for more than 10 generations into the C57BL/6 background) were obtained from Jackson Laboratories and housed in a pathogen-free environment. At age 5 weeks, mice received 4 daily ampicillin/kanamycin administrations by gavage (2 mg per drug per day) in an attempt to facilitate the subsequent bacterial inoculation and enhance colonization. Mice were transferred to a non-pathogen–free environment, and 25 of them were inoculated with the human periodontal pathogen P gingivalis, strain 381. Approximately 0.2 mL of 1.5×1012 bacterial cells/mL in 2% carboxymethylcellulose with PBS was administered via oral topical application and anal topical application/gavage over a 3-week period for a total of 15 inoculations. Mice are coprophagic, and the rationale behind the latter mode of inoculation is that bacteria will be in the feces and will then return to the mouth, thereby establishing a cycle of oral reinfection. Twenty-five control apoE-null mice received antibiotics and vehicle (sterile PBS) on the same schedule. Body weight was recorded for animals in both groups throughout the experimental period (at 5, 9, 13, 15, and 17 weeks of age). At age 17 weeks, all mice were euthanized.
Quantification of Atherosclerotic Lesion Area
Mice were deeply anesthetized, and blood was collected into heparinized syringes from the inferior vena cava. For each mouse, the heart was gently perfused through the pulmonary vein with 0.5 mL of citrate buffer (0.05 mol/L; pH 7.4). On euthanasia, the aorta was retrieved in its entirety to the iliac bifurcation and was snap frozen for analyses. The heart was stored in buffered formalin (10%), and cryostat sections were prepared.14 Atherosclerotic lesion area was quantified using a Zeiss microscope interfaced with a Sony videocamera and an image analysis system (Media Cybernetics), and mean lesion area per mouse was calculated by an investigator blinded to the experimental protocol. Atherosclerotic lesion complexity (fatty streaks versus complex lesions, the latter defined by presence of fibrous caps, cholesterol clefts, or necrosis) was also assessed.6
Quantification of Alveolar Bone Loss
Evaluation of the severity of periodontal destruction by measurement of alveolar bone loss on the lingual surfaces of posterior mandibles was determined as previously described.15 Alveolar bone loss was defined as the total area in pixel units between the cementoenamel junction and the alveolar bone crest.
Lipoprotein and Serum Analysis
Levels of total cholesterol and triglycerides were determined before euthanasia in fasted mice (no caloric intake from 8 to 12 am) using chromogenic assays (Sigma Chemical Co and Boehringer Mannheim, respectively). Fast Performance Liquid Chromatography (Amersham Pharmacia) was performed using 2 Superose 6 columns run in series onto which pooled sterile filtered mouse plasma was applied (5 mice per condition per assay). Serum samples from blood collected at euthanasia were also assayed for interleukin (IL)-6 by ELISA (R&D Systems), glucose by a blood glucose meter (Accu-Chek Advantage, Roche Diagnostics Corporation), insulin by ELISA (Crystal Chem Inc), and creatinine (Anylitics Inc).
Assessment of serum IgG antibody response to P gingivalis was performed by means of the checkerboard immunoassay.16 Protein-A standards (2.5 μg/mL) and bacterial suspensions of P gingivalis 381 (ATCC 33277) and, as a control, a common oral streptococcal strain not associated with periodontitis (Streptococcus oralis, ATCC 35037) were prepared (109 cells/mL) and immobilized on nitrocellulose membranes. Mouse serum samples (diluted 1:300), as well as murine IgG standards (2000 and 1000 ng/mL) were loaded perpendicularly and were allowed to interact. The membrane was incubated with horseradish peroxidase–conjugated anti-mouse IgG. The ECL detection system (Amersham Pharmacia) was used, and the chemiluminescent signals were assessed in a LumiImager Workstation (Boehringer Mannheim).
Aortic tissue was homogenized in lysis buffer (50 mmol/L Tris, pH 7.4, NaCl 0.15 mol/L, phenylmethylsulfonyl fluoride, PMSF, 1 mmol/L), and approximately 20 to 25 μg of total protein per sample was analyzed by SDS-PAGE and immunoblotted with goat anti-human vascular cell adhesion molecule-1 (VCAM-1) IgG (Santa Cruz Biotechnology) and goat anti-rat tissue factor IgG (kindly provided by Dr Vijay Rao, University of Texas Health Sciences at Tyler). Corresponding secondary horseradish peroxidase–conjugated antibodies (Sigma) and the ECL detection system (Amersham-Pharmacia) were used for visualization by autoradiography. Densitometry was performed using the ImageQuant software (Molecular Dynamics).
The following primary antibodies were used on paraffin-embedded heart sections (5 μm thick): goat anti-rat tissue factor IgG, rat anti-mouse Mac-3 IgG (Pharmingen/BD Biosciences), and mouse anti-human smooth muscle actin IgG (Dako).
Polymerase Chain Reaction
Polymerase chain reaction (PCR) was used on DNA from aortic tissue (DNEasy tissue kit, Qiagen) using published primers and conditions.17 Standard housekeeping genes for PCR (G3PDH and β-actin) were included to ensure that adequate quantities and quality of isolated DNA were used. Amplified products were detected by electrophoresis. Genomic DNA extracted from P gingivalis 381 served as positive control, and PCR without template DNA served as negative control.
Unless otherwise noted, when 2 groups were compared, a Student’s t test was used. Statistical analysis was performed using the SAS analysis package (SAS Institute). Parametric correlation analyses were performed, and Pearson’s correlation coefficients were calculated. P<0.05 was considered statistically significant.
Oral Infection Enhances Atherosclerosis and Periodontal Destruction in apoE-Null Mice
Dissection of the hearts and proximal aortae and macroscopic examination at euthanasia (age 17 weeks) revealed increased atherosclerotic plaques in P gingivalis–infected versus noninfected apoE-null mice (Figures 1A through 3⇓⇓D). Morphometric analysis revealed that there was a statistically significant 40% increase in mean atherosclerotic lesion area at the aortic sinus in P gingivalis–infected animals compared with PBS-treated controls (Figure 1E; 12 753±1128 versus 9080±951 μm2, respectively; n=25/group, P=0.008). At age 17 weeks, indices of lesion complexity at the aortic sinus failed to reveal significant differences between the 2 groups, because most lesions were typical fatty streaks, with only a few larger fibrous plaques with cap formation.
To determine the degree of local periodontal inflammation/destruction, we dissected the mandibles at euthanasia and measured alveolar bone loss. Infected animals displayed significantly increased alveolar bone loss compared with controls (12 038±1047 versus 10 978±1330 pixels; n=21 mice/group, P=0.007). Suggestive of a potential link between the degree of vascular inflammation and the severity of alveolar bone loss, a trend for a positive correlation (r=0.27, P=0.08; n=42) was observed between alveolar bone loss and atherosclerotic lesion area.
Oral Infection and the Systemic Host Response
Activation of host immune responses occurred in infected animals, as demonstrated by increased levels of serum IgG to P gingivalis. Checkerboard immunoblot analysis performed on serum before inoculation revealed no evidence of IgG response to P gingivalis in any of the mice. However, intermediate serum analysis (age 13 weeks, 4 weeks after the last inoculation) revealed a strong IgG response to P gingivalis selectively in infected animals. This response decreased by approximately 10% at euthanasia (age 17 weeks, 8 weeks after the last inoculation). In contrast, at no time point did any of the control animals display an IgG response to P gingivalis, and none of the animals demonstrated a response to S oralis.
In concert with increased serum IgG to P gingivalis, levels of plasma IL-6 by ELISA were significantly increased in infected apoE-null mice. Because values did not follow a normal distribution, nonparametric analysis was performed, and median value, range and probability value are reported (Figure 2A; infected, 10 [5.3 to 19.1] pg/mL; controls, 7.3 [5.3 to 22.4] pg/mL; n=25 per group, P=0.04). A statistically significant positive correlation between levels of plasma IL-6 and extent of atherosclerotic lesion area was observed (Figure 2B; r=0.37, P=0.007; n=50). When only infected mice were included in the correlation analysis (n=25), the correlation coefficient amounted to 0.39 (P=0.05) despite the low number of observations.
Oral Infection and Bacterial Invasion in the Vessel Wall
To determine if direct bacterial invasion of the vessel wall may account, at least in part, for the effects of P gingivalis on atherosclerosis, we performed PCR on DNA prepared from aortic tissue collected at euthanasia, 8 weeks after the last inoculation, using published primers specific for P gingivalis. Although none of 9 noninfected mice tested displayed presence of P gingivalis DNA, 2 of 9 infected mice tested demonstrated presence of DNA for this organism. These findings do not unequivocally indicate that bacteria directly invaded the vessel wall but are supportive of this concept.
Oral Infection and Vascular Activation/Inflammation
We sought evidence for altered levels of proinflammatory and prothrombotic mediators in the vascular tissue of infected animals. By immunoblotting, levels of VCAM-1 antigen at euthanasia were significantly increased in aortae retrieved from infected mice versus noninfected controls (Figure 3A; n=8 per group, P=0.03). Levels of tissue factor antigen in the aorta were increased 4.4-fold in infected versus noninfected mice (Figure 3B, n=4 per group; P=0.03). Immunohistochemistry of adjacent sections at the aortic sinus using rat anti-mouse Mac-3 IgG suggested that an important source of tissue factor in the aorta was the macrophage, not smooth muscle cells (the latter as determined using anti-smooth muscle actin IgG) (Figure 4).
Oral Infection and the Effect on Risk Factors for Atherosclerosis
Importantly, it was necessary to determine if infection with P gingivalis modulated established risk factors for atherogenesis in this model. No differences in total plasma cholesterol (infected, 444±22 mg/dL; noninfected, 468±15 mg/dL) or triglyceride (infected, 112±6.8; noninfected, 126±7.3) were observed between the 2 groups (Table; n=25 per group, P=0.4 and 0.2, respectively). However, because it is possible that lack of differences in total lipid levels might belie an altered profile, for instance, a shift toward proatherogenic particles, we separated plasma lipoproteins by size. Fast performance liquid chromatography using sterile-filtered pooled plasma (n=5 mice/pool) was performed and revealed no differences between infected versus noninfected mice (Figures 5A and 5B). Body weight was also monitored throughout the experimental period and was similar in infected and noninfected mice at each of 5 assessments. Similarly, at euthanasia, no differences in glucose, insulin, or serum creatinine levels were observed between infected and noninfected animals (Table; n=5 per group).
In periodontitis, destruction of the attachment apparatus of the teeth is initiated by bacterial colonization of root surfaces and release of bacterial substances onto the epithelial lining of the gingival crevice.18,19⇓ Once bacteria gain access to these sites, their byproducts incite periods of exacerbation and remission of inflammation, leading to increased production of proinflammatory cytokines and matrix metalloproteinases, often by augmenting mechanisms linked to evasion of host defenses.20,21⇓ A consequence of this heavy burden of cytokines and tissue-destructive mediators is the development of hyperpermeability and loss of epithelial integrity, which creates an opportunity for invading bacteria, and their byproducts, to gain access to the systemic circulation. In addition, it has been demonstrated that events such as toothbrushing, scaling/root planing, extractions, and periodontal surgery may promote hematogenous spread of bacteria.22
Definitive support for the concept that oral infection may engage the systemic host response is evident from the observation that specific antibody responses to oral bacteria emerge on infection.23,24⇓ Furthermore, such antibodies could, potentially, exert endothelial cytotoxic effects, as reported in the case of antibodies to Escherichia coli and C pneumoniae.25 Importantly, bacteria and release of lipopolysaccharide into the periphery may act as systemic triggers, activating a range of cytokines and tissue-destructive mediators in circulating inflammatory effector cells.
Interestingly, elevated plasma levels of acute-phase reactants such as C-reactive protein, serum amyloid A, fibrinogen, and IL-6 have been reported in periodontitis in human subjects.26–29⇓⇓⇓ These observations suggest that recruitment of the systemic inflammatory response machinery may accompany localized diseases of the periodontium. Alternatively, or in addition to this premise, it is possible that common underlying predisposing factors, such as genetic predilection of a hyperinflammatory phenotype, emerge as distinct phenotypes in the oral cavity and in the vasculature.30,31⇓
In addition to these indirect means by which periodontitis may prompt vascular perturbation, evidence is emerging that periodontal bacteria, once released into the circulation, may directly invade the blood vessel wall. Indeed, 2 studies in human subjects32,33⇓ have suggested that P gingivalis may be localized to atherosclerotic plaques by PCR and immunostaining. Evidence for definitive molecular mechanisms underlying these observations are being uncovered, as studies in vitro suggest that P gingivalis can invade primary cultures of human coronary artery endothelial cells.34,35⇓ These observations suggest that once these bacteria adhere to and invade endothelial cells, they may directly activate signal transduction pathways linked to the proinflammatory response.
Our murine model of experimental periodontal infection has advantages over other models that do not introduce P gingivalis into the oral cavity but rather via intravenous or subcutaneous routes.12 Of course, the limitation of a mouse model is that because of the very small size of the periodontal tissues, direct application of bacteria in the sulci of the animals is impossible. However, there is evidence that periodontal infection occurs in our model, as alveolar bone loss (the hallmark of periodontal disease and the direct result of infection/inflammation locally in the periodontium) was significantly increased in inoculated versus noninoculated mice. In this study, we believe that the microorganism first gained access to the oral epithelium, followed by engagement of systemic host response mechanisms. The only difference between test and control groups of mice is the bacterial inoculation; it is thus reasonable to conclude that the observed effects here are attributable to this bacterial infection. We propose that basal vascular hyperpermeability and dysfunction, primed by elevated levels of circulating lipoproteins, likely facilitated direct invasion of the blood vessel wall in this murine model. In addition, primed vasculature displays enhanced expression of a wide array of adhesion molecules and chemokines, which may additionally trigger and sustain access of P gingivalis to the vessel wall.
Certainly, periodontitis does not ensue from infection by a single microorganism, as the oral environment is very complex with respect to microbial diversity. However, P gingivalis is a most prominent pathogen in chronic periodontitis and has been extensively studied and characterized to date. Therefore, in these first studies, we used oral infection with P gingivalis to address the role of periodontal infections in modulation of atherogenesis. In this context, it is important to note that although the inoculum of bacteria per mouse was high, it is only a small fraction that likely gained access to the epithelia, because the efficiency of local application and gavage is countered naturally by epithelial integrity. Additional studies are needed in this model system to test lower inocula of P gingivalis and different schedules of administration. Such experimentation will be necessary for the study of the precise mechanisms by which this periodontal pathogen seems able to engage the vasculature and, thereby, mediate direct effects on atherogenesis or lesion progression. It is important to emphasize again, however, that evidence linking P gingivalis specifically to vascular/inflammatory cell perturbation is already emerging.32–36⇓⇓⇓⇓ Pilot studies from our group have shown that P gingivalis infection of human monocytes incubated with LDL significantly increased foam cell formation compared with LDL incubation alone. DPG3, the fimbriae-deficient mutant of P gingivalis that lacks the ability to adhere to and invade host cells, an important determinant of its pathogenicity in human periodontitis, did not have such an effect.37 Furthermore, it is possible that the effects we observed are not specific to P gingivalis and that they might also be induced by other microorganisms. This aspect, as well, is a central focus for future experimentation.
The next step in this research is to use this newly developed animal model to address the mechanistic impact of oral infection at early time points after bacterial inoculation. In this study, our focus was to develop and validate the model; it is highly possible that despite the elevated vascular/plasma indices of inflammation at euthanasia, even additional augmentation of these mediators may have occurred at times more proximal to that of infection.
Lastly, it is well recognized that multiple factors may disrupt vascular homeostasis. In human subjects, elevated levels of lipoproteins and their oxidized forms in particular pose a specific threat. On superimposed perturbation, such as hyperglycemia and its products, increased homocysteine levels, or hypertension, augmentation of vascular dysfunction ensues, leading to acceleration of atherosclerosis and lesion progression, ultimately causing increased incidence of heart attacks and strokes. We propose that the results of the present studies add oral infection with P gingivalis to this array of factors linked to exacerbation of atherosclerosis. Our findings present the first experimental evidence that entry of this periodontal pathogen via the oral route into the host may magnify vascular inflammation and early atherosclerotic lesion formation. These data support the contention put forth by epidemiologic data that unifying mechanisms, at least in part, link the 2 seemingly distinct disorders, periodontal disease and atherosclerosis.
This work was supported in part by the Surgical Research Fund of the College of Physicians and Surgeons, Columbia University, and grants from USPHS to A.M.S. A.M.S. is a recipient of a Burroughs Wellcome Fund Clinical Scientist Award in Translational Research.
- Received March 24, 2003.
- Accepted May 29, 2003.
- ↵Mattila KJ, Nieminen MS, Valtonen VV, Rasi VP, Kesaniemi YA, Syrjala SL, Jungell PS, Isoluoma M, Hietaniemi K, Jokinen MJ. Association between dental health and acute myocardial infarction. BMJ. 1989; 298: 779–781.
- ↵Paunio K, Impivaara O, Tiekso J, Maki J. Missing teeth and ischaemic heart disease in men aged 45–64 years. Eur Heart J. 1993; 14 (suppl K): 54–56.
- ↵Grau AJ, Buggle F, Ziegler C, Schwarz W, Meuser J, Tasman AJ, Buhler A, Benesch C, Becher H, Hacke W. Association between acute cerebrovascular ischemia and chronic and recurrent infection. Stroke. 1997; 28: 1724–1729.
- ↵Assmann G, Cullen P, Jossa F, Lewis B, Mancini M. Coronary heart disease: reducing the risk: the scientific background to primary and secondary prevention of coronary heart disease. A worldwide view. International Task Force for the Prevention of Coronary Heart Disease. Arterioscler Thromb Vasc Biol. 1999; 19: 1819–1824.
- ↵Elkind MS, Lin IF, Grayston JT, Sacco RL. Chlamydia pneumoniae and the risk of first ischemic stroke: the Northern Manhattan Stroke Study. Stroke. 2000; 31: 1521–1525.
- ↵Li L, Messas E, Batista EL Jr, Levine RA, Amar S. Porphyromonas gingivalis infection accelerates the progression of atherosclerosis in a heterozygous apolipoprotein E-deficient murine model. Circulation. 2002; 105: 861–867.
- ↵Zhang SH, Reddick RL, Piedrahita JA, Maeda N. Spontaneous hypercholesterolemia and arterial lesions in mice lacking apolipoprotein E. Science. 1992; 258: 468–471.
- ↵Madianos PN, Papapanou PN, Sandros J. Porphyromonas gingivalis infection of oral epithelium inhibits neutrophil transepithelial migration. Infect Immunol. 1997; 65: 3983–3990.
- ↵Ebersole JL, Holt SC. Immunological procedures for diagnosis and risk assessment in periodontal diseases. In: Johnson NW, ed. Risk Markers for Oral Diseases. Cambridge: Cambridge University Press; 1991.
- ↵Mayr M, Metzler B, Kiechl S, Willeit J, Schett G, Xu Q, Wick G. Endothelial cytotoxicity mediated by serum antibodies to heat shock proteins of Escherichia coli and Chlamydia pneumoniae: immune reactions to heat shock proteins as a possible link between infection and atherosclerosis. Circulation. 1999; 99: 1560–1566.
- ↵Wu T, Trevisan M, Genco RJ, Falkner KL, Dorn JP, Sempos CT. Examination of the relation between periodontal health status and cardiovascular risk factors: serum total and high density lipoprotein cholesterol, C-reactive protein, and plasma fibrinogen. Am J Epidemiol. 2000; 151: 273–282.
- ↵Dorn BR, Dunn WA Jr, Progulske-Fox A. Invasion of human coronary artery cells by periodontal pathogens. Infect Immunol. 1999; 67: 5792–5798.
- ↵Giacona MB, Papapanou PN, Lamster IB, Schmidt AM, Lalla E. Determinants of foam cell formation by human monocyte-derived macrophages infected with Porphyromonas gingivalis. J Dent Res. 2003; 82: B-27. Abstract.