Report

Chlamydia Infections and Heart Disease Linked Through Antigenic Mimicry

See allHide authors and affiliations

Science  26 Feb 1999:
Vol. 283, Issue 5406, pp. 1335-1339
DOI: 10.1126/science.283.5406.1335

Abstract

Chlamydia infections are epidemiologically linked to human heart disease. A peptide from the murine heart muscle–specific α myosin heavy chain that has sequence homology to the 60-kilodalton cysteine-rich outer membrane proteins of Chlamydia pneumoniae, C. psittaci, and C. trachomatiswas shown to induce autoimmune inflammatory heart disease in mice. Injection of the homologous Chlamydia peptides into mice also induced perivascular inflammation, fibrotic changes, and blood vessel occlusion in the heart, as well as triggering T and B cell reactivity to the homologous endogenous heart muscle–specific peptide.Chlamydia DNA functioned as an adjuvant in the triggering of peptide-induced inflammatory heart disease. Infection with C. trachomatis led to the production of autoantibodies to heart muscle–specific epitopes. Thus, Chlamydia-mediated heart disease is induced by antigenic mimicry of a heart muscle–specific protein.

Cardiovascular diseases are the major cause of death in Western societies. Various risk factors have been associated with the pathogenesis of heart diseases, including increased cholesterol levels, smoking, stress, high blood pressure, obesity, and hyperglycemia (1). Bacterial infections may be a causative event in the development of heart diseases (2, 3). Chlamydia infections cause pneumonia, conjunctivitis in children, and are a primary cause of sexually transmitted diseases and female infertility (4). The mechanism by which Chlamydia causes cardiovascular disease is unknown (5).

Inflammatory heart diseases and dilated cardiomyopathy in humans can be reproduced in mice by immunization with heart muscle myosin (6). Cardiac myosin–induced autoimmune myocarditis is dependent on CD4+ T cells that recognize a heart muscle–specific peptide in association with self major histocompatibilty complex (MHC) class II molecules (7). Various peptides of the α myosin heavy chain protein have been identified that can induce autoimmune myocarditis in mice (8, 9).

Immunization with a 30–amino acid peptide (amino acids 614 to 643) of the cardiac-specific α myosin heavy chain molecule [αmhc(614–643)] induces severe inflammatory heart disease in BALB/c mice (8). The first 16 amino acids [αmhc(614–629), SLKLMATLFSTYASAD] constituted a dominant autoaggressive epitope that was designated M7Aα (Table 1 and Fig. 1A) (10). In contrast, the homologous region of the β myosin heavy chain isoform, designated M7Aβ, did not induce disease (Table 1 and Fig. 1B). The introduction of single amino acid substitutions into M7Aα further revealed that the residues xxxMAxxxSTxxx (where x is any amino acid) were important for the pathogenicity of M7Aα in vivo (11). These immunogenic amino acids are conserved between murine and human α myosin heavy chains, and injection of the human M7Aα homolog into BALB/c mice also induced inflammatory heart disease (11).

Figure 1

Inflammatory heart disease in BALB/c mice that were immunized with (A) the endogenous mouse M7Aα peptide from the α myosin heavy chain, (B) the control endogenous M7Aβ peptide from the homologous region of the β myosin heavy chain, (C) the 60-kD CRP-derived peptide from C. trachomatis (ChTR1), or (D) the 60-kD CRP-derived peptide from C. pneumoniae (ChPN) (10). (E) Adoptive transfer of ChTR1 peptide–induced inflammatory heart disease into nonimmunized recipient mice (21). (F) Induction of inflammatory autoimmune heart disease in BALB/c mice with C. trachomatis DNA–derived CpG containing ODN as adjuvant (25). Perivascular inflammation is apparent in (A), (C), (D), and (F). (B) shows normal heart muscle morphology. Hearts were analyzed 21 days after the initial immunization. Staining was with hematoxylin and eosin (H&E). Magnification: ×320

Table 1

Sequence alignment ofChlamydia peptides, the immunogenic mouse M7Aα motif, and the nonimmunogenic mouse M7Aβ motif. Prevalence and severity of inflammatory heart disease as determined with these peptides are indicated. Six-week-old BALB/c mice were immunized twice at a 7-day interval with the indicated peptides (50 μg per mouse] in FCA and analyzed 21 days after the initial immunization for the presence and severity of myocarditis. Histological grading of severity was as follows: 0, no infiltration in heart muscle; 1, up to 5% of histological cross section is infiltrated; 2, 6 to 10%; 3, 11 to 20%; 4, >20%. Mean values of disease severity ± SD are indicated (6, 10).

View this table:

After identification of the crucial pathogenic amino acids within the M7Aα peptide, we screened public databases for viral and bacterial sequences containing the MAxxxST motif (12). Peptide sequences from the 60-kD cysteine-rich outer membrane protein (CRP) from different serovars of C. trachomatis matched the M7Aα motif and were designated ChTR1 (serovar E), ChTR2 (serovar C), and ChTR3 (serovars L1, L2, and L3) (13). The homologous peptides from the 60-kD CRPs of C. pneumoniae, designated ChPN, and C. psittaci, designated ChPS, shared sequence identities with the M7Aα motif, although to a lesser extent (Table 1) (14). Apart from identity at the MAxxxST motif, there were no other conserved regions in the primary sequences of the murine M7Aα peptide and all three Chlamydia 60-kD CRP peptides. A peptide from the p11 protein of C. trachomatis, designated ChTRp11, also shared sequence homology with the M7Aα motif (Table 1) (15).

We tested the possibility of antigenic mimicry betweenChlamydia peptides and the M7Aα motif in our murine model of antigen-induced inflammatory heart disease. We immunized BALB/c mice with murine M7Aα or the homologous 60-kD CRP or p11-derived peptides in Freund's complete adjuvant (FCA) (10). All of theChlamydia-derived peptides induced inflammatory heart disease at a similar frequency, although at significantly lower severity, as compared with M7Aα-immunized mice (Table 1). Like the inflammation that follows immunization with the endogenous autoantigen M7Aα, the disease induced by all the Chlamydia-derived peptides was characterized by perivascular and pericardial infiltration of mononuclear cells and fibrotic changes (Fig. 1, A, C, and D). Immunohistochemical characterization revealed that the inflammatory infiltrate in ChTR1 peptide–induced heart disease was similar to cardiac myosin– and cardiac myosin–derived peptide–induced myocarditis and consisted of about 11% CD4+ and 12% CD8+ T cells, 16% B220+ B cells, and 61% CD11b+ macrophages (16, 17). Inflammation was restricted to the heart and was not observed in skeletal muscle, lung, liver, pancreas, kidney, intestine, or uterus of peptide-immunized mice. Injection of mice with human immunodeficiency virus–2 [gp160 (371–383), INFIGPGKGSDPE]– or parainfluenza virus 1 [HT83b hemagglutinin-neuraminidase (291–309), DLVFDILDLKGKTKSPRYK]–derived peptides that shared homology with other immunogenic regions of the mouse αmhc molecule [αmhc (735–747), GQFIDSGKGAEKL, and αmhc (314–332), DSAFDVLSFTAEEKAGVYK] did not cause inflammatory heart disease (8, 11). Thus, antigenic mimicry between Chlamydia peptides and a heart muscle–specific myosin peptide can lead to the development of inflammatory heart disease.

The development of murine autoimmune myocarditis depends on the activation of CD4+ T cells (7). To directly address the hypothesis of antigenic mimicry between an endogenous cardiac specific peptide and Chlamydia-derived peptides, we immunized BALB/c mice with M7Aα, ChTR1, or another cardiac-specific αmhc-derived peptide, designated kkα (10). This kkα peptide is restricted to I-Ak MHC class II molecules, and kkα immunization induces myocarditis in A/J (I-Ak) mice (9) but not in BALB/c (I-Ad) mice (11). Immunization with M7Aα or ChTR1, but not with kkα or FCA alone, led to splenomegaly and large expansion of TCRαβ+ CD4+ T cells, TCRαβ+CD8+ T cells, B220+ B cells, and CD11b+ macrophages, beginning 8 days after the initial immunization (18). Most (>50%) of CD4+ and CD8+ T cells expressed CD69 and CD25, indicating that these cells had been activated in vivo (19). Splenic T cells from mice immunized with the endogenous peptide M7Aα proliferated when incubated with splenocytes pulsed with the M7Aα peptide (Fig. 2A) (20). Splenic T cells from these mice also showed a strong proliferative response to the C. trachomatis–derived peptide ChTR1 (Fig. 2A). T cells from M7Aα- or ChTR1-immunized mice did not proliferate above control when incubated with γ-irradiated splenocytes pulsed with the nonpathogenic kkα peptide. Splenic T cells from mice immunized with ChTR1 proliferated to ChTR1 and to the endogenous M7Aα peptide. Splenic T cells from control mice immunized with FCA only did not proliferate when activated with M7Aα. ChTR1, or kkα. Thus, ChTR1 peptide immunizations can cross-prime for T cell reactivity against the endogenous M7Aα.

Figure 2

(A) Splenic T cell proliferation and (B) serum IgG antibody production. (A) Proliferative responses to M7Aα, ChTR1, or kkα peptides. Splenic T cells from mice immunized with the indicated peptides were cultured with γ-irradiated syngeneic splenocytes pulsed with the indicated peptides (20). [3H]thymidine uptake (counts per minute) in triplicate cultures is shown (mean ± SD). One representative result of three different experiments is shown. (B) Serum IgG antibodies reactive to cardiac-specific epitopes and ChTR1. Specific antibody production was determined by ELISA (23). For each immunization, representative results of three individual mice are shown.

Cardiac myosin–induced autoimmune myocarditis can be transferred adoptively into nonimmunized recipient mice (7). To establish the autoimmune basis of Chlamydia peptide–induced heart disease, we injected splenic T cells from ChTR1-immunized mice, restimulated in vitro with ChTR1 peptide and murine recombinant interleukin-2 (mrIL-2), into syngeneic BALB/c recipient mice (four mice per group) (21). All animals developed inflammatory heart disease similar (severity 1.0 ± 0.0) to that seen after direct immunization with ChTR1 peptide (Fig. 1E). Splenic T cells from FCA-immunized donors, stimulated in vitro with ChTR1 pepide and mrIL-2, did not induce myocarditis. Thus, ChTR1 peptide–induced myocardits can be transferred adoptively into nonimmunized recipient mice.

Murine autoimmune myocarditis is accompanied by the T cell–dependent production of autoantibodies to cardiac epitopes (6,22). Immunization with endogenous M7Aα peptide led to the production of serum antibodies to the M7Aα peptide used for the induction of the disease and to the ChTR1 peptide (Fig. 2B) (23). Likewise, immunization with the C. trachomatis–derived peptide ChTR1 induced the production of serum antibodies to ChTR1 and to the endogenous M7Aα peptide. Mice immunized with M7Aα or ChTR1 also produced antibodies to the kkα peptide (Fig. 2B), suggesting that M7Aα- and ChTR1-induced heart disease leads to epitope spreading at the B cell level.

How can Chlamydia infections in the lung or reproductive organs lead to the development of myocarditis? In our experimental model of inflammatory heart disease we used FCA as a potent immunoactivator. Bacterial DNA, but not mammalian DNA, has direct immunostimulatory effects in vitro and in vivo (24). We tested whether bacterial DNA–-derived synthetic oligodeoxynucleotides (ODNs) containing unmethylated CpG islands could act as adjuvant for peptide-mediated autoimmunity. Various synthetic CpG motif–containing ODNs could trigger inflammatory autoimmune heart disease in M7Aα peptide–immunized BALB/c mice (Table 2 and Fig. 1F) (25). Immunization of BALB/c mice with a CpG ODN derived from the C. trachomatis CRP gene plus the M7Aα autoantigen induced inflammatory heart disease in the absence of FCA (Table 2 and Fig. 1F) (25). Immunizations in which a control non-CpG ODN was used plus peptide did not induce disease (Table 2). Thus, CpG motif–containing bacterial DNA, including ChlamydiaDNA, can function as potent immunoactivator for autoimmunity.

Table 2

Prevalence and severity of M7Aα peptide–induced myocarditis as determined with synthetic ODNs or FCA as adjuvant. CpG motive–containing ODNs were derived either from C. trachomatis DNA (CpG 1) or from previously reported bacterial DNA sequences (CpG 2 and 3) (25). The CpG motif or the reversed non-CpG motif (non-CpG) is underlined. For severity of myocarditis, seeTable 1. One result representative of three independent experiments is shown.

View this table:

Chlamydia pneumoniae has been linked to atherosclerosis and the clogging of blood vessels (3, 26 ). Experimental C. pneumoniae infections in rabbits and mice accelerate atherosclerosis and lead to focal periarteritis (27) and C. trachomatis infections lead directly to myocarditis (28). Mice immunized withChlamydia peptides developed perivascular fibrosis (Fig. 3, A and B), fibrinous occlusions of cardiac blood vessels (Fig. 3, C and D), and thickening of the arterial walls (29, 30). Fibrinous occlusion originating from blood vessel endothelium (Fig. 3C), a minimum of one per individual heart, occurred in 19 out of 32 (60%) hearts analyzed from mice immunized with Chlamydia-derived peptides. Similarly, fibrinous occlusion originating from blood vessel endothelium occurred in 14 out of 21 (67%) hearts analyzed from mice immunized with M7Aα. No fibrinous occlusions were detected in hearts from mice immunized with FCA only.

Figure 3

Blood vessels in mice immunized with C. trachomatis 60-kD CRP-derived peptide (10,30). (A) Thickening of the arterial wall and perivascular fibrotic changes in mice immunized with ChTR1. The perivascular mononuclear inflammatory cells are apparent. (B) Normal morphology of the cardiac artery in mice immunized with FCA alone. (C) Occlusion of cardiac blood vessels in mice immunized with ChTR1. (D) No occlusions in cardiac blood vessels were seen in control mice immunized with FCA alone. (A and B) Elastica staining for collagen (red) to detect fibrotic changes. (C and D) H&E staining. Magnification, ×320.

Because activation of autoaggressive T and B cells occurred in the absence of an overt bacterial infection, we then determined whether actual Chlamydia infections would lead to the activation of autoaggressive lymphocytes reactive to heart-specific antigens. BALB/c mice were infected with C. trachomatis through the respiratory tract and the reproductive organs (31). Inflammation of both the respiratory tract or the reproductive organs led to the production of immunoglobulin G (IgG) antibodies to heart-specific epitopes in BALB/c mice (Fig. 4). Because in the mouse model of autoimmune myocarditis, the production of IgG antibodies to heart-specific epitopes is dependent on the activation of autoaggressive T and B cells (8), these data show that infection by C. trachomatis can activate autoaggressive lymphocytes in BALB/c mice.

Figure 4

Serum IgG antibody production inC. trachomatis–infected mice. Eight-week-old female BALB/c mice were inoculated either intranasally or intravaginally with the indicated doses of C. trachomatis MoPn IFUs (31). Thirty-six days (intranasal infection) or 42 days (intravaginal infection) after the inoculation, serum was collected and specific IgG antibody production was determined by ELISA (Fig. 2B). Representative data from individual mice are shown.

Our results lead us to propose that Chlamydia infection of an organ can lead to a local immune response followed by systemic activation of autoreactive T and B lymphocytes. BecauseChlamydia peptides can mimic the effects of heart muscle α myosin heavy chain–derived immunogenic epitopes, T cells activated by Chlamydia-derived peptides may trigger organ-specific inflammation within the heart. Dendritic cells, which are resident within the heart and localize in the vicinity of blood vessels, can present cardiac myosin peptides even in healthy animals (7). This observation could account for the invasion of autoaggressive T cells that were activated in other organs. In light of the above data, it is conceivable that, during the course of a bacterial infection, the bacterial DNA acts as a potent adjuvant facilitating the activation of autoaggressive T cells (24).

In mice, the development of peptide-triggered inflammatory heart disease is related to genetic differences among inbred mouse strains (6). Similarly, genetic and environmental risk factors may determine susceptibility to Chlamydia-related heart diseases in humans. Chlamydia infections are common, and most people can expect to experience a Chlamydiainfection at least once during their lifetime (32). Our data suggest that antigenic mimicry of autoaggressive myosin epitopes by peptides present not only in C. pneumoniae but also inC. trachomatis and C. psittaci may be linked to inflammatory heart disease. Molecular mimicry between bacterial and viral proteins and endogenous molecules has been implicated in various autoimmune diseases, including insulin-dependent diabetes, multiple sclerosis, and autoimmune herpes stromal keratitis (33). After initiation of the disease, epitope spreading leads to the maintenance and progression of inflammation. Other mechanisms that could also contribute to the pathogenesis of cardiovascular diseases after Chlamydia infection include the production of inflammatory cytokines, bystander activation of lymphocytes, or both (34). Our results provide experimental in vivo and in vitro evidence of molecular mimicry between bacterial antigens and heart-specific proteins and indicate that bacterial peptides can trigger tissue-specific inflammation of the heart. In particular, this study establishes a causal link between Chlamydia infection and heart disease.

  • * To whom correspondence should be addressed. E-mail: jpenning{at}amgen.com

REFERENCES AND NOTES

View Abstract

Navigate This Article