Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Chlamydia pneumoniae has emerged as a common cause of community-acquired pneumonia, bronchitis, and pharyngitis. The organism has only recently been established as a third species of the obligate intracellular chlamydiae, but is characterized by its extraordinarily high seroprevalence: independent of geographic region, seroepidemiologic results indicate that virtually all persons are infected with the organism at least once during life, and that reinfections are common (1). The majority of C. pneumoniae infections are subclinical, but serious respiratory infection and pronounced lymphocytic alveolitis are observed (2). Because chlamydiae are notorious for causing persistent or recurrent disease with severe immunopathologic effects, concern about sequelae of chronic C. pneumoniae infection is justified. Meanwhile, C. pneumoniae infection has been associated with chronic pulmonary conditions, such as bronchial asthma (3), but the pathogen is apparently also related to extrapulmonary systemic disorders in the form of coronary heart disease (4) and rheumatic disease (5). Although respiratory epithelium has been identified as the primary target of C. pneumoniae infection, a wide range of host cells, such as vascular endothelial cells, smooth-muscle cells, and monocytes/macrophages, can be infected by the organism in vitro (6). However, mechanisms for a systemic dissemination of C. pneumoniae have not yet been described.

The importance of C. pneumoniae in respiratory infections has been acknowledged, but little information is available about its mode of transmission and growth in alveolar cells, and the inflammatory response it induces within the bronchoalveolar compartment. We therefore investigated the interaction between human alveolar macrophages (AM) and C. pneumoniae in vitro. AM play a central role in the nonspecific pulmonary immune system because of their phagocytic and microbicidal activity. However, certain chlamydiae, as well as other intracellular pathogens, can survive within human phagocytes. Mechanisms that may enable chlamydiae to evade the host defense include impairment of the release of bactericidal mediators. In this regard, controversial data exist on the oxidative response to chlamydiae (7) and the chlamydicidal efficacy of reactive oxygen species (ROS) (10). Although interferon- (IFN-) and tumor necrosis factor- (TNF-) inhibit chlamydial growth in vitro (11), they may be responsible for the establishment of chronic chlamydial infection (12). In this study we investigated the infection rate by C. pneumoniae of AM ex vivo and in vitro. Simultaneously, we measured the release of oxygen radicals and proinflammatory cytokines, and the expression of adhesion molecules on the surface of AM.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Culture and Purification of C. pneumoniae

Reference strain C. pneumoniae CWL-029 (ATCC VR-1310) was grown on HEp-2 monolayers (13). Briefly, HEp-2 cells were grown in tissue culture plates, with one well usually containing a cover slip for microscopic control. The growth medium was Eagle's minimum essential medium (Sigma, St. Louis, MO) with 10% fetal calf serum (FCS; Biochrom KG, Berlin, Germany), L-glutamine (2 mM; GIBCO/BRL GmbH, Eggenstein, Germany), nonessential amino acids (GIBCO), gentamicin (10 mg/liter, Sigma), vancomycin (50 mg/liter), and amphotericin B (2 mg/liter). Confluent monolayers were infected with C. pneumoniae by centrifugation (2,000 × g, 35°C, 45 min) of infectious inocula onto the host cells. Supernatants were replaced by infection medium, consisting of serum-free growth medium with cycloheximide (1 mg/liter; Sigma). Infected cells were incubated at 35°C in 5% CO₂, and chlamydial growth was assessed with a fluorescein isothiocyanate (FITC)-conjugated monoclonal antibody (DAKO, Hamburg, Germany). For purification, chlamydiae were released from host cells by mechanical disruption. Host-cell debris was removed by centrifugation at 170 × g for 10 min, and chlamydiae were concentrated by centrifugation of the supernatant (30,000 × g, 1 h). Elementary bodies (EB) were further purified in a urografin sucrose gradient (14), washed twice (30,000 × g, 1 h) in phosphate-buffered saline (PBS; 0.1 mol/liter, pH 7.2), resuspended in PBS, and stored at 70°C until used. Uninfected HEp-2 cells were treated in the same manner for control purposes.

Infection of AM with C. pneumoniae

AM were collected from 12 healthy male nonsmoking volunteers (mean age: 25.8 yr) without clinical or laboratory signs of airway inflammation (mean FEV₁ = 102%, VC = 99%, C-reactive protein [CRP] < 6 mg/liter, and immunoglobulin E [IgE] = 31 IU/ml). Bronchoalveolar lavage (BAL) was performed under standard conditions in the right middle lobe, and the mean BAL differential cell count disclosed 93% AM, 5.8% lymphocytes, 0.5% neutrophils, and 0.2% eosinophils. The cells were separated by centrifugation at 170 × g, washed twice in PBS, and counted on a hemocytometer. Wright-Giemsa-stained cytospin slides were prepared (Cytospin II; Shandon, Sewickley, PA), and the cell viability was determined by trypan blue dye exclusion. AM were suspended at a concentration of 10⁶/ml in M199 (GIBCO) containing 5% FCS, L-glutamine, and penicillin/streptomycin (1 g/liter; GIBCO), and were allowed to adhere in polystyrene tubes for chemiluminescence measurement, on 13-mm cover slips for microscopic evaluation, or in culture plates at 37°C in a totally humidified atmosphere for 3 h. AM monolayers were then washed twice in M199, inoculated with the EB preparation at concentrations of 0 EB per AM, 4 EB per AM, and 40 EB per AM, and incubated for up to 120 h. Cycloheximide treatment and the centrifugation amplification step were omitted to provide more physiologic conditions. At 4, 8, 16, 48, 72, 96, and 120 h, the infected monolayers were fixed with methanol and stained for C. pneumoniae with the FITC-conjugated antibody according to the manufacturer's instructions. Slides of bronchoalveolar lavage fluid (BALF) from six patients with clinical C. pneumoniae pneumonia, as determined by polymerase chain reaction (PCR) assay on lavage fluid specimens (2), were stained in the same manner.

Production of Reactive Oxygen Species

The release of oxygen radicals was determined through luminol- and lucigenin-amplified chemiluminescence (15). Lucigenin preferentially reacts with superoxide anion, whereas luminol was included for control of nonspecific or neutrophil-derived light emission, since it reacts with intermediates in the myeloperoxidase (MPO)-hypochlorous acid pathway of reactive oxygen species (ROS) metabolism that are not generated by AM. For the measurement of ROS, 100 µl of the resuspended cells were incubated in 4-ml polystyrene tubes and infected with 0, 4, or 40 EB/ AM. After 0.5 h and 3 h incubation, respectively, the spontaneous and phorbol myristate acetate (PMA)-stimulated (1.4 µg/ml) production of ROS was measured in duplicate in a semiautomatic luminometer (LB 953; Berthold, Wildbad, Germany) for a period of 30 min. The peak value was taken for calculation and expressed in counts per minute (cpm).

Cytokines

The activity of TNF- was measured with a L929 lytic bioassay (16). Infected and noninfected AM were grown in 96-well tissue-culture plates and the supernatants were removed after 16 h, pooled from triplicate wells, and immediately frozen at 70°C. Threefold serial dilutions of the collected samples and a human recombinant TNF- standard (Bender, Vienna, Austria) were distributed in the wells of a 96-well tissue culture plate and incubated for 20 h in the presence of L929 fibroblast cells (60,000 cells/ well) and actinomycin D. Remaining cells were stained with crystal violet, absorbance values were measured at 570 nm, and TNF- concentrations were obtained by interpolation from a standard curve. Interleukin-1 (IL-1) was measured in the supernatants after 16 h, using an enzyme immunoassay (Quantikine; R&D Systems Inc., Abingdon, UK) that is specific for IL-1 and does not show cross-reactivity with TNF- or IL-8. IL-8 was measured in the supernatants after 16 h with a sandwich enzyme-linked immunosorbent assay (ELISA) (17). In this assay a polyclonal goat anti-IL-8 antibody was bound to the wells of 96-well microtiter plates (4°C for 20 h). Nonspecific binding sites were blocked with 0.1% BSA. Supernatants (diluted 1:50) and recombinant human IL-8 standard (Biermann, Bad Nauheim, Germany) were added and incubated for 90 min, followed by incubation for 20 h with monoclonal mouse anti-IL-8 antibody. Peroxidase-conjugated goat antimouse antibody/(2,2'-azino-bis-[3-ethylbenzothiazoline-6-sulfonic acid]) (ABTS) was used as a developing system, and plates were read at 405 nm. Results were obtained by interpolation from a standard curve. IL-8 was also measured in the lavage fluids from the 12 volunteers and nine patients with C. pneumoniae pneumonia. BALF was concentrated by centrifugation (5,000 × g) in microconcentrators (Centricon 10; Amicon, Witten, Germany) (17) and used undiluted in the IL-8 assay.

Expression of Intercellular Adhesion Molecule-1 and Human Leukocyte Antigen-DR

Intercellular adhesion molecule-1 (ICAM-1) and human leukocyte antigen subclass DR (HLA-DR) were measured on the AM surface with a cell-associated ELISA (18). After 16 h, supernatants were removed and cells were fixed with paraformaldehyde (10 g/liter; 21°C for 15 min). Free binding sites were blocked with 2% BSA (30 min at 37°C). A mouse monoclonal anti-ICAM-1 antibody (1:1,000, 0.1% BSA) was added at 100 µl per well for 1 h at 37°C, or an anti-HLA-DR antibody (1:1,000, 0.1% BSA) was added at 100 µl per well for 12 h at 37°C. Peroxidase-conjugated goat antimouse antibody/ABTS served as a developing system, and plates were read at 405 nm after 30 min for ICAM-1 and 2.5 h for HLA-DR. Results were calculated from quadruplicate wells and expressed in absorbance values. An even cell distribution was quantified by staining with crystal violet.

Modification of the Interaction between Chlamydiae and AM

IFN- (100 to 200 U/ml) was added for 16 h to infected AM from a subgroup of five volunteers as a positive control for cytokine release and surface-marker expression. To inhibit attachment and lipopolysaccharide (LPS)-induced stimulation of target cells, chlamydiae were preincubated at 21°C for 30 min with heparin (5 U/ml), which interferes with binding of chlamydiae on the host cell, or with polymyxin B (100 µg/ml), which as a polycation binds directly to the anionic lipid A portion of LPS.

Statistics

For statistical analysis, nonparametric tests were used throughout the study. Wilcoxon's signed rank test was used to test the significance of chlamydia-stimulated effects in comparison with noninfected controls. Correlations were calculated with Spearman's rank correlation test. Values of P < 0.05 were considered statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Infection and Phagocytosis

EB were readily ingested by AM, with an infection rate at 16 h of 47 ± 22% (mean ± SD) for the concentration of 4 EB/AM and 76 ± 10% for the concentration of 40 EB/ AM. The number of infected cells decreased to 37 ± 12% (4 EB/AM) and 60 ± 12% (40 EB/AM) after 96 h. Distinct inclusions were observed after 72 h. The number of inclusions decreased by 120 h. Figure 1, top panel and middle panel, shows infected AM in vitro after 16 h and 96 h, respectively; morphologically identical inclusions were demonstrated in AM from patients with chlamydial pneumonia (Figure 1, bottom panel). Under ex vivo conditions, 32 ± 11% of AM were shown to be infected, and inclusions were observed in about 4.3 ± 2% of AM.

View larger version (33K): [in this window] [in a new window]   View larger version (26K): [in this window] [in a new window]   View larger version (89K): [in this window] [in a new window]   Figure 1.   Human AM infected with Chlamydia pneumoniae in vitro (top, middle) and ex vivo (bottom). AM from nonsmoking volunteers were grown on 13-mm cover slips and depleted of nonadherent cells, infected in vitro with 40 EB/AM, and stained for C. pneumoniae with an FITC-conjugated monoclonal antibody after 16 h (top) and 96 h (middle). BALF slides of six patients with clinically diagnosed and PCR-confirmed C. pneumoniae pneumonia were stained in the same manner ex vivo (bottom).

Oxidative Response to Chlamydia pneumoniae Infection

The oxidative response of AM to Chlamydia pneumoniae was evaluated by luminol- and lucigenin-amplified chemiluminescence. As expected, there was no significant change in the luminol-dependent assay, which exclusively measures ROS of the MPO-hypochlorous acid pathway. In contrast, after amplification with lucigenin, infected AM showed a marked, dose-dependent increase in spontaneous release of ROS after 30 min as compared with controls (1.6 ± 0.9 × 10⁵ cpm at 4 EB/AM and 3.3 ± 2.5 × 10⁵ cpm at 40 EB/ AM, versus 1.1 ± 0.7 × 10⁵ cpm for controls; P < 0.05 and P < 0.01, respectively). The results of nine experiments, shown in terms of peak values, are summarized in Figure 2. The effect was almost completely inhibited after addition of superoxide dismutase (SOD; 2.0 µg/ml). The release of ROS decreased over the following 3 h, and after 6 h there was no difference in the release of ROS by stimulated cells and controls. The formation of ROS was further enhanced by preincubation with IFN- (100 U/ml) over a period of 3 h (data not shown).

View larger version (15K): [in this window] [in a new window]   Figure 2.   Release of ROS by C. pneumoniae-infected AM (4 EB/AM or 40 EB/AM) in vitro, in comparison with noninfected AM (control). AM (106) from healthy, nonsmoking volunteers (n = 9) were infected with 0 (control), 4 EB/AM, or 40 EB/AM, and incubated at 37°C in 5% CO2 for 30 min. The spontaneous, PMA-stimulated, and SOD-inhibited release of ROS was determined by lucigenin-amplified chemiluminescence in a semiautomatic luminometer over a period of 30 min. Mean peak values ± SD expressed in cpm are shown. * P < 0.05; ** P < 0.01 in comparison with control values.

Release of Cytokines

Figure 3 summarizes the measurement of TNF-, IL-1, and IL-8 from the 12 volunteers. Spontaneous cytokine secretion was low (TNF-: 27 ± 26 IU/ml; IL-1: 15 ± 5 pg/ ml; IL-8: 73 ± 23 ng/ml) in all cases. Infection with C. pneumoniae resulted in significant increases in the measured cytokines: TNF- release was enhanced 4-fold (4 EB/AM) or 12-fold (40 EB/AM) after stimulation; there was a 14-fold increase (4 EB/AM) and a nearly 60-fold increase (40 EB/AM) in IL-1 release. The secretion of IL-8 showed a 3-fold increase (4 EB/AM) and a 4-fold increase (40 EB/ AM) after infection with chlamydiae. The release of cytokines after additional preincubation with IFN- (100 U/ ml) was tested in five volunteers, and was further enhanced to 379 IU/ml versus 327 IU/ml for TNF- (release from AM stimulated with 40 EB/cell + IFN- versus chlamydiae alone) and to 317.5 pg/ml versus 213 pg/ml for IL-1, respectively, whereas the release of IL-8 was not increased by addition of IFN-.

View larger version (16K): [in this window] [in a new window]   Figure 3.   Cytokine release by C. pneumoniae-infected AM compared with noninfected controls. AM from healthy volunteers were depleted of nonadherent cells and incubated in 96-well microtiter plates with 0 (control), 4 EB/AM, or 40 EB/AM at 37°C in 5% CO2 for 16 h. Cytokines were quantitated in the supernatants with an L929 fibroblast lytic bioassay (TNF-) or by enzyme immunoassay (IL-1, IL-8), and the values are shown as mean value ± SD. * P < 0.05; ** P < 0.01 in comparison with control values.

TNF- release from infected macrophages correlated with that of IL-1 (r = 0.68, P = 0.021). In contrast, there was an inverse relation between the basal ROS secretion and the release of TNF- and IL-1 (r = 0.67, P < 0.05). The IL-8 level determined in the BALF of patients with C. pneumoniae pneumonia (316 ± 152 ng/ml) in comparison with healthy controls (73 ± 25 ng/ml) showed a highly significant (P < 0.0001) 4-fold increase.

Expression of Surface Molecules

ICAM-1 expression on the AM surface was not significantly changed by chlamydial infection (4 EB/cell: 1.06 ± 0.43 ELISA units (EU)/ml; 40 EB/cell: 0.8 ± 0.35 EU/ml versus 0.99 ± 0.53 EU/ml in controls). In contrast, there was a dose-dependent increase in HLA-DR expression (4 EB/ cell: 0.98 ± 0.36 EU/ml; 40 EB/cell: 1.07 ± 0.34 EU/ml versus 0.84 ± 0.34 EU/ml in controls, P < 0.05), which exceeded the stimulating effect of 100 U/ml IFN- (0.92 ± 0.4 EU/ml). There was no correlation between the expression of surface molecules and the release of cytokines or oxygen radicals.

Modification of Chlamydia/Macrophage Interaction

As shown in Table 1 (mean values, polymyxin B: n = 7; heparin: n = 3), preincubation of chlamydiae with polymyxin B resulted in a marked inhibition of the release of oxygen radicals (by 48%), TNF- (by 57%), infection rate (by 53%), and development of inclusion bodies (P < 0.05). The number of inclusions decreased from 4.4% to 1.3% at 96 h. Despite a reduced infection rate (by 49%), preincubation with heparin resulted in a less prominent inhibition of oxygen radical release (by 14%) or even in stimulation of TNF- release, which was not seen after addition of heparin alone (data not shown).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Chlamydia pneumoniae can establish and maintain a productive infection in AM in vitro. The concomitant dose-dependent release of inflammatory mediators (ROS, TNF-, IL-1, IL-8) found in our study did not prevent chlamydial replication for several days. These in vitro results are confirmed by the demonstration of similar infection patterns in AM from patients with severe C. pneumoniae pneumonia. AM apparently act as host cells for C. pneumoniae in vitro and in vivo. The decrease in the number of infected AM observed in our study after 96 h might suggest a limited capacity to inhibit chlamydial growth in vitro, which could be due to the microbiostatic effects of cytokines and oxidants produced in response to chlamydia (19, 20). Alternatively, the number of infected cells may decrease by apoptosis or cell death. However, in our study, the percentage of dead cells did not differ in experiments with infected and those with control AM, as estimated by trypan blye dye exclusion (data not shown). A similar decrease in inclusion-forming units in AM in comparison with infected HEp cells after 96 h was reported recently (6). However, we observed a considerable number of infected AM even after an incubation time of 120 h.

Possible mechanisms for chlamydial survival in phagocytes include inhibition of phagosome-lysosome fusion and ineffective oxidative and nonoxidative killing mechanisms. Previous investigators did not find C. trachomatis to induce a luminol-dependent chemiluminescence in PMN (21) or reactive oxygen species in AM as measured by ferricytochrome-c reduction (7). These contradicting results may be due to the chlamydial species, the cells, or the assays used. Further studies are needed to clarify the impact of ROS on the viability of chlamydiae. In addition, chlamydiae can be inhibited by oxygen-independent mechanisms mediated by lymphokines, as shown in oxidatively deficient macrophages (22). Our in vitro finding of an increased release of inflammatory cytokines after chlamydial infection was confirmed ex vivo by detection of increased IL-8 levels in the BALF of patients with clinical C. pneumoniae infection.

The production of TNF-, IL-1, and IL-6 by C. pneumoniae-infected macrophages can also be induced in the permanent monocytic MonoMac 6 cell line (23). The proinflammatory cytokine and chemokine network, of which only a minor part was evaluated in this study, is known to play a central role in the activation and recruitment of leukocytes to the pulmonary compartment. Interstitial infiltrates consisting of PMN and mononuclear cells have also been demonstrated in a mouse model of C. pneumoniae pneumonitis (24). Whereas neutrophilic infiltrations are found mainly in the early phase after infection, which was mimicked by our in vitro experiments, a mononuclear infiltrate predominates beyond 5 to 8 d after infection (24). In a clinical study, we observed a distinct lymphocytic alveolitis in the BALF of patients with C. pneumoniae pneumonia, most of whom presented with a subacute phase of disease (2). The continuous release of inflammatory mediators during persistent infection has the potential to exert local destructive effects on lung tissue.

The effect of IFN- and TNF- on chlamydial infection may be ambiguous. In C. trachomatis infection, the cytokine response is suspected of inducing persistent infection by inhibiting replication of the pathogen without its complete elimination. A concomitant upregulation of chlamydial heat-shock proteins then may lead to the severe immunopathology often observed (25). Growth of C. trachomatis in HEp cells can be inhibited by TNF- (26), possibly through the depletion of tryptophan, which chlamydiae cannot synthesize. Tryptophan depletion may also be involved in the IFN--mediated persistence of chlamydiae (25). Similar mechanisms may be responsible for pulmonary and extrapulmonary manifestations of C. pneumoniae infection. A microbiostatic effect of the cytokines produced in response to chlamydiae apparently becomes irreversible only after extended coincubation with activated macrophages (20). The continuing viability of chlamydiae within AM may lead to persistence of the pathogen and permit its transportation into other tissues.

Recent data show that C. pneumoniae can replicate within vascular cells (e.g., human aortic and pulmonary artery-derived endothelial cells, and aortic smooth muscle cells) (6). Moreover, viable pathogen has been recovered (27) from atheromatous tissues. The mode of transmission of C. pneumoniae from the respiratory tract to the circulation is not clear. Infected monocytes/macrophages are possible vectors for the pathogen. In an in vitro model, C. pneumoniae infection can be transmitted from MonoMac 6 cells to endothelial cells (unpublished data), which indicates a possible mechanism for the transportation of C. pneumoniae from inflamed tissues to the vascular system. Alternatively, chlamydiae may be disseminated directly via bacteremia.

The process of chlamydial attachment and ingestion by the host cell remains unclear. It is presumed that ligand- receptor binding leads to entry of the organism into the host cell. The binding is trypsin sensitive (28) and can be inhibited by heparin (29), which was confirmed for C. pneumoniae in our study. According to recent studies of C. trachomatis, the inflammatory immune response to the organism is endotoxin mediated (30). To neutralize LPS-induced effects, polymyxin B was added, which, as a polycation, binds directly to the anionic lipid A portion of the LPS (31). The inhibitory effect of polymyxin B on the release of oxygen radicals and cytokines, the infection rate, and the formation of inclusion bodies found in our study suggests that the inflammatory response to C. pneumoniae infection is at least partly an endotoxin-mediated effect. An alternative explanation could be that polymyxin has an antibiotic effect on C. pneumoniae that was not investigated in our study. However, this mechanism seems unlikely, since the drug acts almost exclusively on extracellular pathogens.

In conclusion, our findings both ex vivo and in vitro indicate that the AM plays an important role as a target cell in C. pneumoniae infection and elicits a marked inflammatory response to the microorganism. Because of the intracellular survival of the pathogen, AM may represent a reservoir for persistent C. pneumoniae infection. Further studies of the pathogenesis of systemic C. pneumoniae infection, and of the antimicrobial efficacy of the inflammatory mediators released in response to the organism are clearly warranted.