|Home | About | Journals | Submit | Contact Us | Français|
Heat shock protein 60 derived from Chlamydia pneumoniae (cHSP60) activates Toll-like receptor 4 (TLR4) signaling through the MyD88 pathway in vitro, but it is not known how cHSP60 contributes to C. pneumoniae-induced lung inflammation. We treated wild-type (WT), TLR2−/−, TLR4−/−, or MyD88−/− mice intratracheally (i.t.) with recombinant cHSP60 (50 μg), UV-killed C. pneumoniae (UVCP; 5 × 106 inclusion-forming units/mouse), lipopolysaccharide (2 μg), or phosphate-buffered saline (PBS) and sacrificed mice 24 h later. Bronchoalveolar lavage (BAL) was obtained to measure cell counts and cytokine levels, lungs were analyzed for histopathology, and lung homogenate chemokine concentrations were determined. Bone marrow-derived dendritic cells (BMDDCs) were generated and stimulated with live C. pneumoniae (multiplicity of infection [MOI], 5), UVCP (MOI, 5), or cHSP60 for 24 h, and the expression of costimulatory molecules (CD80 and CD86) was measured by fluorescence-activated cell sorting. cHSP60 induced acute lung inflammation with the same intensity as that of UVCP-induced inflammation in WT mice but not in TLR4−/− or MyD88−/− mice. cHSP60- and UVCP-induced lung inflammation was associated with increased numbers of cells in BAL, increased neutrophil recruitment, and elevated BAL interleukin-6 (IL-6) levels. Both cHSP60 and UVCP induced IL-6 release and CD80 and CD86 expression in WT cells but not in MyD88−/− BMDDCs. cHSP60 stimulated DC activation in a TLR4- and MyD88-dependent manner with an intensity similar to that induced by UVCP. These data suggest that cHSP60 promotes lung inflammation and DC activation via TLR4 and MyD88 and therefore may play a significant role in the pathogenesis of C. pneumoniae-induced chronic inflammatory lung diseases.
Chlamydia pneumoniae is an obligate intracellular gram-negative bacterium that causes upper and lower respiratory tract infections throughout the world; it is responsible for 10% of community-acquired pneumonia (17). The estimated number of cases of C. pneumoniae-induced pneumonia is 300,000 cases per year. Approximately 50% of young adults and 75% of the elderly population have serological evidence of previous infection (8, 17). In addition to acute infection, there is increasing evidence that implicates C. pneumoniae and heat shock protein 60 derived from C. pneumoniae (cHSP60) in the pathogenesis of atherosclerosis and chronic inflammatory lung diseases such as asthma, chronic obstructive pulmonary disease, and bronchitis (7, 9, 10, 29).
The three major effector antigens of Chlamydia are lipopolysaccharide (LPS), the major outer membrane protein (MOMP), and cHSP60 (16). Chlamydia pneumoniae has a unique biphasic life cycle. The metabolically inactive infectious elementary bodies (EBs) attach and enter the host cell, where they differentiate into the metabolically active reticulate bodies. The reticulate bodies replicate within the expanding endosome, resulting in the development of characteristic cytoplasmic inclusions. These persistent intracellular inclusions contain increased quantities of cHSP60, a highly immunogenic protein that has been implicated in the stimulation of the innate immune system and the pathogenesis of chronic inflammatory lung diseases (9). Chlamydia can achieve a state of chronic intracellular infection in which they remain viable but quiescent and do not replicate. During such persistent infections, cHSP60 is abundantly produced and might stimulate the innate immune and inflammatory responses, thus contributing to chronic inflammatory lung diseases (9, 14, 16).
The involvement of the cHSP60 of C. trachomatis in the immunopathogenesis of trachoma, pelvic inflammatory disease, and tubal infertility is well established, but the role of cHSP60 in the pathogenesis of C. pneumoniae-induced infections and chronic lung disease has not been discerned (25, 26, 32). Evidence that cHSP60 indeed plays a role comes from studies showing the presence of anti-cHSP60 antibodies (Abs) in adult patients with asthma who developed symptoms after an acute respiratory illness (11). However, the cHSP60 responses might simply represent a marker for exposure to chlamydial infection. Conversely, Huittinen et al. showed an independent association between cHSP60 immunoglobulin A (IgA) Abs and the severity of allergic asthma, even after controlling for the effects of migration inhibition factor IgA Abs to C. pneumoniae (12). Importantly, these authors demonstrated that only Abs against cHSP60, and not those against C. pneumoniae EBs, were associated with chronic airway disease (31). These results clearly suggest the possibility that the persistent presence of cHSP60 after C. pneumoniae infection in the lungs participates in the immunopathology of chronic airway disease. Furthermore, several studies have linked cHSP60 with asthma and decreased pulmonary functions (11, 12, 34).
Previous reports indicate that C. pneumoniae infection and cHSP60 alone can activate the innate immune system through Toll-like receptors (TLRs), one of the sensors of innate immunity (5, 24, 28, 30, 33). Live Chlamydia pneumoniae infection can activate the innate immune response by both TLR2 and TLR4, and we have demonstrated a critical role of MyD88 in host defense against C. pneumoniae (22, 28). Some studies with dendritic cells (DCs) indicate that the immune recognition of live Chlamydia pneumoniae was dependent largely on TLR2 and only to a lesser extent on TLR4 (24, 27). However, we reported that cHSP60 is a potent inducer of vascular endothelial cells (EC) and macrophage inflammatory responses, and these inflammatory effects are mediated through the innate immune receptor complex TLR4-MD2 (4). These responses proceed via the MyD88-dependent signaling pathway.
In this study, we show that the i.t. administration of cHSP60 induces acute lung inflammation in a TLR4- and MyD88-dependent manner with severity comparable to that seen after the inoculation of UVCP in mice. We also demonstrate that cHSP60 induces lung inflammation and stimulates DC activation and maturation in a MyD88-dependent manner and with intensity similar to that induced by UVCP. These data suggest that cHSP60 is a key inflammatory component of Chlamydia pneumoniae that induces lung inflammation and DC activation, and therefore it may play a significant role in the pathogenesis of C. pneumoniae-induced acute and chronic inflammatory lung diseases.
Specific-pathogen-free C57BL/6 mice 8 to 12 weeks of age were used throughout the study. C57BL/6 wild-type (WT) mice and TLR2−/− mice were purchased from Jackson Laboratories and bred at our facility. MyD88−/− and TLR4−/− mice were kindly provided by Shizuo Akira (Osaka University, Japan), and a homogenous population was established by backcrossing these mice to C57BL/6 mice for at least 8 generations. All experiments received prior approval from the Cedars-Sinai Medical Center Institutional Animal Care and Use Committee.
C. pneumoniae CM-1 (ATCC, Manassas, VA) was propagated in HEp-2 cells as previously described (22). C. trachomatis was obtained from Kathleen Kelly, UCLA. C. pneumoniae stocks were determined to be free of Mycoplasma contamination by PCR (22). UV-treated C. pneumonia (UVCP; 5 × 106 inclusion-forming units [IFU]/mouse) and UV-treated C. trachomatis (UVCT; 5 × 106 IFU/mouse) were prepared by placing cultures 1 cm below a UV light for 15 min (34). UV inactivation was confirmed by the subculturing of treated bacteria in HEp-2 cells. Recombinant cHSP60 protein was isolated and purified as described earlier (4). The concentration of LPS measured in the undiluted recombinant protein was <0.06 ng/ml, which corresponds to <0.0012 ng/ml LPS in 10 μl of protein, a concentration well below what is needed for cellular activation (data not shown). To further exclude the possibility of endotoxin contamination, cHSP60 and Escherichia coli LPS preparations were heat treated (100°C, 20 min) before incubation. Heat treatment abolished the ability of cHSP60 to induce NF-κB activation but had no effect on the activation of NF-κB by LPS, as we described earlier (4). In addition, the preincubation of cells with a specific monoclonal Ab (MAb) raised against cHSP60 (A57-B9; 50 μg/ml) significantly blocked cHSP60-induced NF-κB activation, again establishing that the effects are specific to cHSP60 and not to any contaminating LPS, as we published earlier (4). E. coli K235 LPS was obtained from S. Vogel (Uniformed Services University, Bethesda, MD). The purity of this LPS preparation has been demonstrated previously, and this preparation of LPS is active on TLR4-transfected HEK 293 cells and not on TLR2 transfectants (S. N. Vogel, unpublished observation) (4).
Mice were anesthetized with isoflurane, followed by the intratracheal (i.t.) installation of 100 μl of sucrose-phosphate-glutamate-phosphate-buffered saline buffer (SPG-PBS) containing either 5 × 106 IFU UVCP, 5 × 106 IFU UVCT, 50 μg cHSP60, 2 μg LPS, or SPG-PBS alone. SPG-PBS alone does not induce inflammation after i.t. application (data not shown). The dose of 50 μg cHSP60 was used as in the case of in vitro experiments, as this dose of cHSP60 and 5 × 106 IFU UVCP gave equivalent tumor necrosis factor alpha release from primary mouse bone marrow-derived macrophages (data not shown). In preliminary experiments of 24 and 40 h after cHSP60 injection, we saw a robust inflammation at 24 h that started to clear at 40 h. Therefore, we chose 24 h as the time point for analysis. At the time of sacrifice, bronchoalveolar lavage (BAL) was collected with 0.5 ml of PBS. The lavage fluid was centrifuged, and the supernatant was used to determine interleukin-6 (IL-6) levels using an enzyme-linked immunosorbent assay (ELISA) (BD Bioscience). The pellet was resuspended in PBS, and the number of cells was counted. Half of the lung was homogenized in 1 ml of sucrose-phosphate-glutamate medium and stored at −70°C. The remainder of the lung was fixed in 10% formalin and paraffin embedded, and hematoxylin and eosin (H&E)-stained sections were evaluated for inflammation.
Lung lymphocytes were isolated by digesting the lung tissue at 37°C for 1 h with HANKS' solution containing 100 μg/ml Blenzyme 3 (Roche) and 50 U/ml DNase I (Roche), and then it was filtered through a 70-μm cell strainer (BD Falcon). Erythrocytes were depleted by lysis buffer before staining. Isolated single cells were stained with specific MAbs against Gr1 (clone RB6-8C5) and CD11b (clone M1/70), both purchased from eBioscience, as direct conjugates to phycoerythrin (PE) or PECy5. Flow cytometric analysis was performed by a FACScan flow cytometer (BD Biosciences), and the data were analyzed by Summit (Dako, Denmark). The percentages of gated positive cells are indicated.
Cytokine and chemokine concentrations in BAL and lung homogenate were analyzed using mouse IL-6 (BD Biosciences, Franklin Lakes, NJ) and mouse keratinocyte-derived chemokine (KC) (R&D Systems, Minneapolis, MN) according to the manufacturers' instructions.
Bone marrow-derived DCs (BMDDCs) were generated by incubating bone marrow cells with 10 ng/ml recombinant mouse granulocyte-macrophage colony-stimulating factor (Biosource) for 6 days, with medium changes at days 3 and 5. DCs were harvested at day 6 and were purified with CD11c microbeads (Miltenyi Biotec) according to the manufacturer's suggested protocol. Purity was checked by flow cytometry using anti-CD11c Abs and was routinely above 90%. BMDDCs were generated from WT, TLR2−/−, TLR4−/−, and MyD88−/− mice stimulated in vitro with either live C. pneumoniae (multiplicity of infection [MOI], 5:1), UVCP (MOI, 5:1), LPS (100 ng/ml), Pam3CSK4 (2 μg/ml), or cHSP60 for 24 h. IL-6 was measured by ELISA in the supernatants as described above. The expression of costimulatory molecules was measured by fluorescence-activated cell sorting using fluorescein isothiocyanate-conjugated CD80 and CD86 MAbs (eBiosciences).
Independent experiments were conducted at least in triplicate, except as otherwise noted. Results were summarized as means ± standard deviations and were compared using one-way analysis of variance (ANOVA) followed by a Bonferroni post test. A P value of less than 0.05 was required to reject the null hypothesis.
To investigate whether cHSP60 induces lung inflammation, WT, MyD88−/−, TLR2−/−, and TLR4−/− mice were injected i.t. with either PBS, cHSP60, UVCP, or LPS and then sacrificed 24 h later. Both cHSP60- and UVCP-injected lungs revealed patchy interstitial and diffuse pneumonia with massive infiltration of neutrophils into the lung tissue (Fig. (Fig.1).1). Lung histopathology revealed increased interstitial infiltration 24 h after cHSP60, UVCP, and LPS injections in WT and TLR2−/− mice (Fig. (Fig.1).1). However, these changes were not observed in MyD88−/− and TLR4−/− mice that were similarly treated with i.t. cHSP60 injection. Although there was increased interstitial infiltration in UVCP-treated WT mice, the magnitude of infiltration was less in TLR4−/− mice and TLR2−/− mice, and infiltration in MyD88−/− mice was no different from that of PBS-injected control mice (Fig. (Fig.1).1). We also examined the BAL cell count as another measure of inflammation. WT, MyD88−/−, TLR4−/−, and TLR2−/− mice were treated as described above and sacrificed 24 h later. The installation of cHSP60 into WT and TLR2−/− mice resulted in significantly increased BAL cell counts (Fig. (Fig.2),2), but there were no significant increases in BAL cell counts in MyD88−/− and TLR4−/− mice. Unlike cHSP60, significant BAL cell counts were found in WT mice only after UVCP installation. LPS injection also resulted in significantly increased BAL cell counts in WT and TLR2−/− mice, but as expected, it did not have any effect on MyD88−/− and TLR4−/− mice (Fig. (Fig.22).
Since it is known that Chlamydia pneumoniae strongly recruits neutrophils into the lung early in infection (23), we investigated the effect of cHSP60 on neutrophil recruitment. WT mice were injected i.t. with either PBS, cHSP60, UVCP, or LPS and sacrificed after 24 h. The recruitment of neutrophils into the lung was measured by flow cytometry. cHSP60 induced a nearly twofold increase in neutrophils in the lung 24 h after installation (Fig. (Fig.3A).3A). UVCP and LPS, as expected, also induced an increase in neutrophils (Fig. (Fig.3A).3A). Since neutrophils were indeed present in greater numbers following cHSP60 administration, we measured the levels of the neutrophil attractant chemokine KC using ELISA. KC levels in lung homogenates also were significantly increased 24 h after cHSP60, UVCP, and LPS injection compared to those in lung homogenates after PBS injection (Fig. (Fig.3B3B).
To investigate the mechanism by which cHSP60 induces acute lung inflammation, we measured the levels of IL-6 in the BAL of WT, TLR4−/−, MyD88−/−, and TLR2−/− mice treated with either PBS, cHSP60, UVCP, or LPS as described above. We observed significantly increased IL-6 production in WT and TLR2−/− mice after cHSP60 installation (Fig. (Fig.4A).4A). There was no increase in IL-6 production in TLR4−/− and MyD88−/− mice, as expected. LPS and UVCP also induced significant IL-6 production in WT mice but not in MyD88−/− mice. Compared to that of WT mice, UVCP-induced IL-6 production was reduced in TLR4−/− and TLR2−/− mice (Fig. (Fig.4A).4A). However, LPS-induced IL-6 in TLR2−/− mice was similar to that of WT mice, but as expected, it did not occur in TLR4−/− mice.
Similarly, cHSP60 induced IL-6 into the supernatant of BMMDCs but did not induce IL-6 in TLR4−/− and MyD88−/− BMMDCs (Fig. (Fig.4B).4B). The IL-6 production of TLR2−/− BMDDCs exposed to cHSP60 was normal compared to that of WT cells. As expected, LPS-induced IL-6 production was significantly reduced in TLR4−/− and MyD88−/− BMDDCs, while normal levels were induced in WT and TLR2−/− BMDDCs. Finally, Pam3CSK4 exposure induced IL-6 only in WT and TLR4−/− BMDDCs and was inhibited in MyD88−/− and TLR2−/− BMDDCs.
Seeing that cHSP60 did promote IL-6 production, we next measured the ability of cHSP60 to induce the maturation and activation of BMDDCs by measuring CD80 and CD86 activation from WT and MyD88−/− mice. BMDDCs were incubated with cHSP60 for 24 h, and then CD80 and CD86 expression levels were measured by flow cytometry. Indeed, the ability of cHSP60 alone to induce DC maturation was similar to that of live C. pneumoniae and UVCP (Fig. (Fig.5).5). This maturation was MyD88 dependent, and CD86 induction was especially contingent upon MyD88 activity (Fig. (Fig.55).
Having shown that UVCP and recombinant cHSP60 can induce lung inflammation, we wished to determine if UVCT also induces pulmonary inflammation. The i.t. injection of UVCT also resulted in a significant increase in BAL cell counts, IL-6 secretion in BAL fluid, and KC production in the lung homogenates in WT mice (Fig. (Fig.6).6). Lung histopathology examination revealed increased interstitial infiltration 24 h after UVCT injections in WT mice (Fig. (Fig.66).
We investigated the role of cHSP60 and innate immune signaling pathways in C. pneumoniae-induced acute lung inflammation. We showed that the i.t. administration of cHSP60 was able to induce an acute lung inflammation in 24 h. WT and TLR2−/− mice were able to mount a successful host defense against cHSP60, but TLR4−/− and MyD88-deficient mice were unable to upregulate proinflammatory cytokines. TLR4−/− and MyD88−/− mice also were unable to recruit polymorphonuclear leukocytes and showed no histopathological evidence of lung inflammation. These findings are consistent with our previous in vitro studies, which showed that cHSP60 is a potent inducer of vascular EC and macrophage inflammatory responses, and that these inflammatory effects are mediated through the TLR4-MD2 complex and depend upon MyD88 (4). While MyD88 also plays a role in other potential chlamydial signaling pathways (TLR9, IL-1, and IL-18), our data clearly indicate that the TLR4/MyD88 pathway is responsible for cHSP60 signaling. In the present study, UVCP injection also induced acute inflammation in WT mice; however, the cytokine response and BAL cell recruitment were diminished in both TLR2−/− and TLR4−/− mice and was absent from MyD88−/− mice. Since UVCP contains all three major ligands of Chlamydia (LPS, MOMP, and cHSP60), we expected to see some inflammatory activity in both TLR2−/− and TLR4−/− mice; however, this response was significantly diminished in MyD88−/− mice.
Parallel to these findings, both cHSP60 and UVCP induced costimulatory molecules such as CD80 and CD86 in WT BMDDC, but there was a notable reduction of costimulatory molecules in MyD88-deficient mice. Also, cHSP60, UVCP, and LPS induced neutrophil recruitment into lungs in 24 h. KC is a potent chemoattractant for murine neutrophils (3), and our results demonstrated that lung homogenate KC production increased 24 h after cHSP60, UVCP, or LPS injection in WT mice. Our studies are consistent with recent studies that indicate that cHSP60 induces T-cell polarization and the maturation and functional activation of monocyte-derived DCs (1).
We already have established in an earlier study that the in vitro biological effects of the recombinant cHSP60 used in this study are specific to HSP60 and are not due to any contaminating LPS as listed in Materials and Methods (4). The concentration of LPS measured in the undiluted recombinant protein was <0.06 ng/ml, which corresponds to <0.0012 ng/ml LPS in 10 μl of protein, a concentration well below what is needed for cellular activation. Indeed, the i.t. administration of E. coli LPS at 0.01 ng/100 μl (concentrations of LPS were greater than those in our HSP60 preparation) per mouse did not induce any lung inflammation in 24 h compared to results with PBS. There was no significant difference in lung histopathology, BAL cell count, BAL and lung homogenate IL-6 levels, and KC levels in these two groups (data not shown). Moreover, another recent study with ultrapure recombinant cHSP60 reported that highly purified recombinant cHSP60 stimulated human peripheral monocyte cytokine synthesis and vascular endothelial cell adhesion protein expression, and that this activity was blocked by the antibody neutralization of TLR4 (19). Therefore, all of these observations, combined with our current data, now suggest that recombinant cHSP60 has direct immune and inflammatory effects in vivo as well as in vitro, and that these biological effects are specific and not due to LPS contamination. Despite the biochemical and functional evidence that we have provided to support the notion that the activity of cHSP60 is not due to contaminating LPS or TLR2 ligands, it remains possible that cHSP60, as a chaperon protein, tightly associates with other, yet-to-be defined microbial pathogen-associated molecular patterns and leads to TLR4-mediated signaling.
Several studies demonstrated that cHSP60 is capable of eliciting the profound inflammatory characteristics of delayed-type hypersensitivity (18, 21). Morrison et al. showed that purified cHSP60 triggered an ocular immunopathologic response in sensitized animals that was histologically and clinically comparable to that of chronic chlamydial disease. Chlamydia LPS and MOMP did not elicit such responses (21). When Chlamydia remains viable but dormant and does not replicate, the production of cHSP60 is unaltered but that of MOMP decreases markedly (2, 20). In addition, patients with advanced upper genital tract disease possess Abs to cHSP60 that cross-react with human HSP60 (18). Since cHSPs share significant homology with the mammalian homologues, this raises the possibility that chronic infection with Chlamydia triggers autoimmune responses (6, 21).
Recent work suggests that cHSP60 is involved in the pathogenesis of chronic inflammatory diseases such as atherosclerosis, asthma, and arthritis (11, 12, 14, 16, 20, 34). Both cHSP60 and human HSP60 can be found in atherosclerotic plaques and activate proinflammatory cytokines in human vascular endothelium, smooth-muscle cells, and macrophages (14, 16). cHSP60 may function in two ways to promote atherosclerosis: first, by direct antigenic stimulation, and second, as a signal transducer that triggers the activation of cells within atheromatous lesions (4, 14, 16). Several studies have linked cHSP60 with asthma and decreased pulmonary functions (11, 12, 34). The presence of Abs against cHSP60 has been demonstrated in patients with asthma who developed the symptoms after an acute respiratory tract infection (11). In addition, Huittinen et al. demonstrated that IgA Abs to cHSP60 were associated with decreased pulmonary function in asthma patients (12). Only Abs against cHSP60, and not those against C. pneumoniae EBs, were associated with asthma (31). The continuous presence of these highly conserved proteins could result in direct tissue damage, or they may cause the induction of cross-reactivity with host heat shock protein homologues and induce autoimmunity (18).
Recent evidence supports the conclusion that cHSP60 activates inflammatory cytokines (13, 15) and innate immunity through TLRs. Recombinant purified cHSP60, devoid of LPS contamination, potently induces vascular endothelial cells and macrophage inflammatory responses through TLR4-MD2 and MyD88 (4). cHSP60 stimulates vascular smooth-muscle cells, monocytes, and epithelial cells via TLR4 (19, 30), but others have reported that cHSP60 can signal via TLR4 and TLR2 (5, 30, 33). However, our in vitro (4) and current in vivo studies strongly suggest that recombinant cHSP60 is recognized predominantly by TLR4. We conclude that cHSP60 may be a key inflammatory component of persistent Chlamydia pneumoniae infections that promotes lung inflammation and DC activation via MyD88-dependent signaling pathways, and it may play a significant role in the pathogenesis of C. pneumoniae-induced acute and chronic inflammatory lung diseases.
This work was supported by NIH grant 8AI062938-02 to Y.B. and 5R01AI058128, 5R01AI067995, RO1HL66436 to M.A.
We thank Richard P. Morrison (Department of Microbiology and Immunology, University of Arkansas, Little Rock) for providing us with recombinant cHSP60.
Editor: S. R. Blanke
Published ahead of print on 27 April 2009.