INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Trachoma-causing strains of Chlamydia trachomatis are endemic in many developing nations and are a leading cause of preventable blindness worldwide (57). In the United States, infection with sexually transmitted strains of C. trachomatis is the most common infectious disease reported to the Centers for Disease Control and Prevention and state health departments, estimated at 4 million to 5 million new cases annually (11). Approximately 60% of adults worldwide are seropositive to Chlamydia pneumoniae (28), and infection is associated with atherosclerosis and is a risk factor for myocardial infarction (54). Chlamydial disease requires chronic infection or reinfection, implicating the immune system in pathogenesis (27).

Natural infections begin with chlamydiae binding to epithelial cells, although macrophages, endothelial, and smooth muscle cells can be infected (24). Dissemination of chlamydiae from the mucosal portal of entry to distal sites of infection, such as lymph nodes or arteries, is thought to be mediated by macrophages. Bacteria are internalized and replicate within cells in a specialized vacuole termed the inclusion. In vitro, Chlamydia cells induce their uptake (40) and elicit secretion of proinflammatory cytokines and chemokines by infected epithelial cells (48).

Phosphatidylserine (PS), a lipid sequestered to the cytofacial plasma membrane (PM) leaflet, is externalized to the outer leaflet when cells enter apoptotic states. Within hours, influenza (22) or Legionella pneumophila (23) infections increase host cell PS exposure, representing an early host response to intracellular parasites. Externalized PS can be detected using annexin V, a cytoplasmic protein that binds PS and other acidic phospholipids in the presence of millimolar concentrations of calcium (62). Two- to 100-fold enhancements in mean annexin V-dependent fluorescence are commonly obtained (2, 7, 8, 38). Externalized PS has roles in physiologic processes, including coagulation, phagocytosis, and complement activation (14, 17, 65). Externalized PS on platelets (70), endothelial cells (7, 10), and monocytes/macrophages (2) serves to accelerate clotting reactions. Procoagulant enzyme complexes form on PS-rich domains, accelerating thrombin activation and fibrin generation (14). PS signals for phagocytosis of apoptotic cells by macrophages (56) via an uncharacterized receptor, among which CD36 is a candidate (18).

Associated with the early stages of apoptosis, PS externalization can be accompanied by increased PM permeability and caspase activation. The relationship between infection with Chlamydia and apoptosis is apparently developmental stage specific, with induction of proapoptotic (25, 42) and antiapoptotic (19) phenotypes previously described.

We tested whether Chlamydia would generate a PS response by infecting epithelial, endothelial, granulocytic, and monocytic cells. C. trachomatis or C. pneumoniae infection led to every cell type tested rapidly and transiently exposing PS on their PMs without an increase in PM permeability. PS externalization was dose dependent and extracellular calcium dependent but was unaffected by pretreatment of cells with a broad-spectrum caspase inhibitor shown to inhibit induced PS exposure (61, 69). Externalized PS was functional as Chlamydia-infected cells accelerated plasma clot formation and were phagocytosed by macrophages in a PS-dependent manner.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents. Annexin V-biotin was purchased from Clontech (Palo Alto, Calif.), and recombinant annexin V, streptavidin-fluorescein isothiocyanate (SA-FITC), and SA-phycoerythrin (SA-PE) were purchased from Pharmingen (San Diego, Calif.). Recombinant annexin V-green fluorescent protein (V-GFP) (16) was a kind gift from Joel Ernst, San Francisco General Hospital and University of CaliforniaSan Francisco, San Francisco, Calif. Percoll was purchased from Pharmacia (Piscataway, N.J.), fatty acid-free bovine serum albumin (BSA), 7-aminoactinomycin D (7-AAD), and actinomycin D were purchased from Calbiochem (San Diego, Calif.), and Russell's viper venom (RVV), Sigmacote, cephalin, PS, and phosphatidylcholine (PC) dissolved in chloroform were purchased from Sigma (St. Louis, Mo.).

Bacteria and cells. C. trachomatis serovar L2 (434/Bu/L2) and C. pneumoniae strain CWL-029 were obtained from the American Type Culture Collection (Manassas, Va.). HeLa, THP-1, and L929 cells were purchased from the American Type Culture Collection. Human coronary artery endothelial (HCAE) cells were purchased from Clonetics (San Diego, Calif.). HCAE cells were maintained in endothelium growth medium-MV per manufacturer's instructions. Neutrophils (>95% pure) were obtained from citrated whole blood by dextran sedimentation, followed by fractionation on a discontinuous Percoll gradient as described previously (55). Monocytes were obtained from buffy coats by adhesion onto glass or plastic in Iscove's modified Dulbecco's medium (IMDM), and nonadherent cells were removed after 1 h by gentle washing. Citrated whole human blood was obtained from the Alameda County Chapter of the American Red Cross (Oakland, Calif.). Work involving human samples was approved by the University of CaliforniaBerkeley Committee for the Protection of Human Subjects.

Cell culture and infection. HeLa, THP-1, and L929 cells were grown in IMDM (Gibco, Gaithersburg, Md.) plus 10% fetal bovine serum (Hyclone, Logan, Utah) plus 50 µg of vancomycin HCl (Sigma)/ml. Plastic tissue cultureware was used to maintain cells for assays, and glass spinner flasks were used to grow HeLa cells for C. trachomatis serovar biovar LGV propagation. HEp2 cells in six-well plastic tissue culture plates (Costar, Corning, N.Y.) were used for C. pneumoniae propagation. Macrophages were maintained in X-Vivo-10 medium (Biowhittaker, Walkersville, Md.) supplemented with 10% human AB serum (Sigma) and 25 µg of vancomycin/ml. HeLa cells were aspirated and overlaid with 1 ml (per 25 cm²) of IMDM containing 2 inclusion forming units (IFU) of C. trachomatis or C. pneumoniae placed on a rotating shaker, and infection was allowed to proceed for 2 h at room temperature. Inocula were prepared by brief sonication of bacteria in ice-cold unsupplemented endothelial cell basal medium. THP-1 and HCAE cells were overlaid with the inocula and returned to the 37°C incubator. Inocula for HCAE cells were prepared in unsupplemented endothelial cell basal medium (Clonetics). Uninfected cells were mock infected by using unsupplemented medium lacking chlamydiae and subjecting the cells to the same handling procedures as infected cultures.

View larger version (81K): [in this window] [in a new window]   FIG. 1.   Fluorescence micrographs of HeLa cells binding the PS-specific probe annexin V after acute infection with C. trachomatis or treatment with staurosporine. Infected cells probed with annexin V (A) stained bright green compared to uninfected cells (B). HeLa cells cultured with 1 µM staurosporine (C) or C. trachomatis (D) exposed PS on their plasma membranes. C. trachomatis cells were pretreated with 50 µg of unlabeled recombinant annexin V/ml or BSA and were used to infect HeLa cells. After washing the infected cell monolayers and staining with annexin V-GFP, no significant change in PS externalization (as green fluorescence) is seen between cells infected with C. trachomatis pretreated with annexin V (E) or BSA (F). The concentration of unlabeled annexin V used (E) was adequate to block binding of annexin V-GFP (100 ng/ml), as staurosporine-treated cells subsequently incubated with unlabeled annexin V (G) followed by annexin V-GFP staining showed no green fluorescent staining compared to staurosporine-treated cells incubated with BSA and then stained with annexin V-GFP (H). Cells in panels A through D were probed with annexin V-biotin and SA-FITC; all other cells were probed with annexin V-GFP. Cells were counterstained with Evans blue (background red fluorescence). A 40× objective (A, B, E, F, G, and H) and a 100× objective (C and D) were used to record images.

Fluorescence microscopy and flow cytometry. For fluorescence microscopy, 100,000 HeLa cells were plated overnight in 24-well tissue culture plates (Costar) with 12-mm-diameter glass coverslips. The next day, inocula were prepared in unsupplemented IMDM as was described. Infections were allowed to proceed for 1 to 2 h, and then the inoculum was removed. Wells were washed once with IMDM, aspirated, and then overlaid with annexin V-GFP (0.35 ng/ml) or annexin V-biotin (50 ng/ml) diluted in annexin V binding buffer (AVB; 140 mM NaCl, 10 mM HEPES [pH 7.4], 2.5 mM CaCl₂). After 10 min, the wells were aspirated and washed twice with AVB and then were fixed with ice-cold 2% paraformaldehyde-0.2% glutaraldehyde in phosphate-buffered saline (PBS) for 30 min on ice. The wells were washed, covered, and incubated at room temperature for 45 min with a 0.0005% Evans blue (Sigma) solution in PBS. The coverslips were mounted using ProLong antifade (Molecular Probes, Eugene, Oreg.), and simultaneous green and red fluorescence images (original magnification, ×400) were taken using an Olympus camera mounted on an Axiophot (Zeiss) fluorescence microscope. For blocking experiments, inocula (0.8 IFU/ml) resuspended in ice-cold AVB were incubated with recombinant annexin V or fatty acid-free BSA (50 ng/ml) for 30 min on ice. After 10-fold dilution with ice-cold AVB, bacteria were pelleted by ultracentrifugation (18,000 × g) at 4°C. Bacteria were resuspended in AVB by sonication, and 0.25 ml was applied to each well. For the PS-positive control wells, cells were pretreated with 1 µM staurosporine for 1 h at room temperature and then were aspirated and washed twice with ice-cold AVB. Either recombinant annexin V or BSA (at 50 ng/ml) in ice-cold AVB was added to the wells and allowed to bind for 30 min on ice. Wells were aspirated and washed twice with AVB, and the monolayers were probed with 0.35 ng of annexin V-GFP/ml, counterstained, and fixed as described above.

Liposome preparation. PS was dissolved in chloroform to 10 mg/ml. The PS-to-PC ratios was 30:70. Lipids were dried under nitrogen, PBS was added, and then lipids were sonicated on ice for a stock concentration of 1.0 mM total lipid. Liposomes were stored on ice and used within 1 h of preparation.

Plasma clotting assay. The method used for the plasma clotting assay was adapted from that described by Casciola-Rosen et al. (10), but we used 1.25 ng of RVV per reaction mixture instead of 12.5 ng. Platelet-poor plasma was prepared by centrifugation of citrated blood for 10 min at 400 × g and drawing the upper two-thirds for coagulation studies. PPP was stored at 70°C. Clotting was initiated by adding calcium to the reaction mixture prewarmed in a 37°C water bath. Siliconized 13- by 100-mm borosilicon glass tubes were visually examined every 10 s for clot formation. The endpoint of the assay was reached when a firm fibrin clot formed.

Phagocytosis assays. The method used for the phagocytosis assay was described by Fadok et al. (18). Macrophages were plated as described onto glass coverslips in 24-well plates, and medium was changed every 3 day. On day 5 or 6, UV-sterilized glucan (Sigma) was added to the medium to a final concentration of 25 µg/ml. After two further days of culture, the cells were washed twice in unsupplemented medium before addition of inhibitors or target cells. Liposomes were added to a final concentration of 0.1 mM total lipid, and RGDS peptide was added to a final concentration of 1 mM. Neutrophils were added at the indicated ratios in X-Vivo-10 medium and cocultured at 37°C for 2 h. Wells were washed twice with medium, were washed once with PBS, were fixed, and then were esterase stained and hematoxylin counterstained (Sigma) for counting. Target cells were scored as phagocytosed if 50% or more of a blue target cell fell within a macrophage PM border. The phagocytic index was derived as follows: the percent phagocytosing macrophages were multiplied by the average number of neutrophils ingested (18). Neutrophils were scored as ingested if 50% or more were within a macrophage border. Between 400 and 500 macrophages were counted per sample.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Chlamydial infection induces host cell PS exposure. Annexin V was used to detect PS (62) in human cells after infection with C. trachomatis or C. pneumoniae. Infected HeLa cells (Fig. 1A) acquired a bright green punctate staining pattern localized to the PM, while uninfected cells did not stain (Fig. 1B). Cells treated with the kinase inhibitor staurosporine externalized phosphatidylserine (Fig. 1C). The morphology of the annexin V staining patterns of staurosporine-treated and Chlamydia-infected cells (Fig. 1D) appeared to be dissimilar. Staurosporine-treated cells tended to generate stained areas localized to one side of the cell, while infected cells displayed a more even, punctate staining over the entire cell surface.

View larger version (19K): [in this window] [in a new window]   FIG. 2.   Detection of PS during Chlamydia infection under conditions that influence cellular PS transport and metabolism. (A) Effect of extracellular calcium concentrations upon PS externalization in uninfected HeLa cells () and in HeLa cells infected with 1 IFU of Chlamydia/cell (+) and stained with annexin V-GFP. (B) PS response of HeLa cells pretreated with the PS synthesis inhibitor serinol upon infection with Chlamydia. Shown are flow cytometric histograms of HeLa cells cultured with the PS synthesis inhibitor serinol and infected with 0.5 IFU of C. trachomatis/cell (Serinol+C.t.), infected cells cultured with L-serine (L-Serine+C.t.), and uninfected cells cultured in serinol or L-serine. Cells were stained with annexin V-biotin and SA-PE. (C) Flow cytometric dot plots of data of forward light scatter versus 7-AAD uptake of cells (B) demonstrate no significant changes in cell size or viability induced by culturing HeLa cells for 4 days with 2 mM serinol or L-serine or by acute infection with 2 IFU of C. trachomatis per cell.

Chlamydial binding is required for induction of host cell PS externalization. Inhibiting chlamydial protein synthesis with chloramphenicol did not affect peak fluorescence values or kinetics of PS exposure over a 48-h assay period (Fig. 3A). A second peak of PS exposure was sometimes noted in other chloramphenicol-treated cocultures (data not shown). Mild heating of EB abolishes their ability to infect and reduces their ability to bind host cells (40). Heated EB provoked less HeLa cell PS externalization (Fig. 3B). Pellets and supernatants from the high-speed centrifugation of heated EB were tested to induce PS exposure, and all activity was restricted to the pellet fraction (data not shown). It has been shown previously that excess heparin blocks chlamydial binding and entry, with the analog chondroitin sulfate having no effect (68). If Chlamydia binding triggered host PS exposure, then added heparin should have blocked PS externalization. Heparin treatment abolished Chlamydia-induced PS exposure (Fig. 3C). This suggests that chlamydial binding, not a soluble component, is minimally required to elicit host cell PS exposure.

View larger version (10K): [in this window] [in a new window]   FIG. 3.   Kinetics of host PS exposure by treatments affecting chlamydial protein synthesis and adhesion. (A) Effect of chloramphenicol upon infected host cell PS exposure. HeLa cells treated with chloramphenicol (50 µg/ml) before, during, and after pulse inoculation with 2 IFU of C. trachomatis per cell (squares) were compared to infected cells without chloramphenicol (triangles) and uninfected cells with (circles) and without (crosses) chloramphenicol over a 48-h period. Data are representative of three experiments. Data are from 10,000 gated events gathered by flow cytometry per time point. (B) PS externalization in HeLa cells infected with heat-treated bacteria (60°C, 10 min; squares), untreated bacteria (triangles), or buffer alone (circles). Data from 10,000 events per time point were gathered by flow cytom- etry at 1, 0, 2, 4, 6, 8, 12, 24, and 48 h postinfection. (C) Effect of pretreating host cells with heparin upon Chlamydia-induced PS externalization. Flow cytometric histograms of C. trachomatis-infected HeLa cells pretreated with 0.5 mg of heparin/ml (Heparin+C.t.) or with 0.5 mg of chondroitin sulfate/ml (Chondroitin+C.t.) and uninfected treated cells are shown. Data represent 20,000 events gathered by flow cytometry. Cells were stained with annexin V-biotin and SA-FITC.

Dose response and kinetics of PS exposure by Chlamydia-infected HeLa cells. Both the amplitude and rate of PS externalization were influenced by the inoculum dose. In Fig. 4A, cells continuously infected with 2 IFU of Chlamydia/cell increasingly externalized more PS over the assay. Cells infected with 12.5 IFU/cell displayed a steady level of PS exposure throughout the infection period. Interestingly, cells treated with the highest dose, 50 IFU/cell, began to lose detectable PS 30 min after inoculum addition; after 2 h of continuous infection, PS was near or at baseline uninfected-cell levels (n = 3 experiments). The mechanism behind cellular PS loss is not known; this may be a consequence of the cell reinternalizing PS-rich PM as the assay was conducted at 23 to 25°C, which permits attachment and EB endocytosis (40). Alternatively, PS loss may occur via the shedding of cell membranes that are enriched in the supernatants of cells treated with chemical inducers of apoptosis (3).

View larger version (11K): [in this window] [in a new window]   FIG. 4.   Kinetic dose-response relationship between infectious bacterial titer and PS exposure in HeLa cells infected with C. trachomatis. (A) Annexin V-dependent fluorescence of HeLa cells undergoing continuous short-term infection with 2 (circles), 12.5 (squares), 50 (closed triangles), or zero (open triangle) IFU of C. trachomatis per cell over a 2-h period. Doses of 50 IFU or more per cell over the 2-h period led to a decline in annexin V binding in three of three assays. (B and C) Relationship between subsaturating doses of infectious C. trachomatis and PS exposure by bacterium-pulsed HeLa cells. The percentage of HeLa cells binding annexin V (B) and mean fluorescence data (C) are shown. Data representing doses of 2 (closed triangles), 0.67 (squares), and 0.22 (circles) IFU per cell and buffer control (open triangles) are shown. Cells were processed for flow cytometry and mean fluorescence data are shown. Cells were stained with annexin V-GFP (A) or with annexin V-biotin and SA-FITC (B and C).

Host cell PS exposure induced by Chlamydia infection is transient and reinducible. PS exposure depended on the continuous presence of Chlamydia since inoculum removal led to PS disappearance (Fig. 3A and 4B). The transient nature of Chlamydia-mediated PS exposure appeared to be a novel finding, and we tested whether staurosporine- and ethanol-induced PS exposure was also transient. Twenty-four hours after stimulus withdrawal, staurosporine- and ethanol-treated cells still externalized PS while Chlamydia-infected cells were negative (Fig. 5A). We next examined whether loss of externalized PS by infected cells was reversible. Infected cells treated with either C. trachomatis, ethanol, or staurosporine all reexternalized PS (Fig. 5B). These results indicate that host cell entry and/or binding by Chlamydia is necessary to induce PS exposure, and the host PS response to subsequent agonists is not blunted by a concurrent infection with Chlamydia.

View larger version (15K): [in this window] [in a new window]   FIG. 5.   The transient, reinducible nature of Chlamydia-induced PS externalization. (A) Flow cytometric histograms of C. trachomatis-infected cells (+C.t.), staurosporine-treated cells (+St), or ethanol-treated cells (+EtOH) stained for PS 24 h after agonist withdrawal. Data represent 3,500 events. (B) Reinduction of PS exposure in infected cells. Twenty-four hours postinfection, HeLa cells were pulsed with either 2 IFU of C. trachomatis/cell (+C.t.), 1 µM staurosporine (+St), 5% ethanol (+EtOH), or medium for 2 h and then PS exposure was measured. Cells were stained with annexin V-biotin and SA-FITC.

C. pneumoniae infection induces PS exposure in arterial endothelial cells and in epithelial cells. C. pneumoniae accesses vascular cells; therefore, the bacterium encounters epithelial cells at least once during infection. We tested whether C. pneumoniae elicits PS exposure in HeLa epithelial cells. Cells externalized PS in response to C. pneumoniae (Fig. 6A), and paralleling the above results with C. trachomatis, chloramphenicol left PS exposure unaffected, but mildly heating the inoculum lowered mean fluorescence (Fig. 6A).

View larger version (21K): [in this window] [in a new window]   FIG. 6.   Induction of PS exposure by infectious C. pneumoniae. (A) Mean flow cytometric data of HeLa epithelial cells infected with C. pneumoniae. HeLa cells were inoculated with the following: infectious C. pneumoniae at 1 IFU per cell (+Cpn), C. pneumoniae and 50 µg of chloramphenicol/ml (+Cpn+chloramp), C. pneumoniae cells that were heated at 60°C for 10 min (Heated), 50 µg of chloramphenicol/ml (Chloramp), or a volume of sucrose buffer equal to that used for the infectious inoculum in medium (SPG). (B) Flow cytometric histograms of HCAE cells infected with C. pneumoniae and stained for PS. HCAE cells were pulsed with bacteria and then stained with annexin V and second-step reagent (HCAE+Cpn) or with second-step reagent alone (HCAE + Cpn, SA-FITC only); uninfected cells were stained with both reagents (HCAE, no Cpn) or with second-step reagent alone ("HCAE, no Cpn SA-FITC only"). (C) Flow cytometric dot plots of forward light scatter versus side scatter (ss log) and annexin V-biotin and SA-FITC staining versus 7-AAD uptake of HCAE cells from panel B. Infected HCAE cells (HCAE+Cpn) did not demonstrate significant changes in cell size and granularity (Forward Scatter) or PM permeability (7-AAD fluorescence) compared to uninfected controls. Data represent 6,000 gated events.

Human monocytes and neutrophils externalize PS in response to inoculation with Chlamydia. Intense inflammation accompanies acute Chlamydia infections, with neutrophils and monocytes accumulating at the infection site (35). Since infected epithelial and endothelial cells externalize PS, we tested whether this would occur in macrophages and neutrophils, as both cell types come into contact with the bacteria. Infected THP-1 promonocytes (Fig. 7A) externalized PS, as did infected neutrophils (Fig. 8A). PS externalization kinetics of inoculated neutrophils (Fig. 8B) revealed a pattern similar to that of HeLa cells (Fig. 3A and 4C), but no second PS peak was seen.

View larger version (27K): [in this window] [in a new window]   FIG. 7.   PS exposure and viability of THP-1 promonocytes infected with C. trachomatis. (A) Flow cytometric histograms of THP-1 cells stained for PS externalization that were either infected with C. trachomatis (THP-1+C.t.) or uninfected (THP-1). Data represent 10,000 events. (B) Flow cytometric dot plots of forward scatter versus side scatter (ss log) and annexin V-biotin plus SA-FITC versus 7-AAD staining of THP-1 cells from panel A demonstrate that cell size and granularity (forward scatter) and PM permeability (7-AAD fluorescence) of C. trachomatis-infected cells (THP-1+C.t.) do not change compared to those of uninfected THP-1 cells.

View larger version (28K): [in this window] [in a new window]   FIG. 8.   (A) PS exposure in human neutrophils inoculated with C. trachomatis. Shown are flow cytometric histograms of uninfected neutrophils (no C.t.) and neutrophils inoculated with C. trachomatis (+C.t.). Data represent 5,000 gated events. (B) Kinetic assay of PS exposure by neutrophils that were either inoculated with C. trachomatis (triangles) or left uninfected (circles). PS was detected by flow cytometry of annexin-V biotin and SA-PE-stained neutrophils (A and B). (C) Macrophage-mediated phagocytosis of C. trachomatis-inoculated neutrophils can be inhibited by PS liposomes. Neutrophils (at a ratio of five neutrophils to one HMDM) that were inoculated (+C.t.) or mock infected (no C.t.) were added to wells containing RGDS peptide or 0.1 mM liposomes. RGDS peptides and PS liposomes were added 30 min prior to neutrophil addition and remained during the coculture period. As described in Materials and Methods, 400 to 500 macrophages per sample were counted. (D) Chlamydia-infected cells are procoagulant, and acquired procoagulant activity is dependent upon exposed PS. THP-1 cells inoculated with C. trachomatis (THP-1+C.t.) significantly accelerated clotting over uninfected THP-1 cells (THP-1). Addition of the PS-binding protein annexin V abrogated enhanced coagulant activity in infected cells (THP-1+C.t.+Annexin V). The molar ratio of annexin V to calcium was 1:4,000, making it unlikely that annexin V exerted its anticoagulant effect by sequestering calcium. Assay range response was indicated by reactions run with cephalin as a positive clotting control or with added buffer only. Data are mean clotting times ± standard deviation of triplicates. Data are representative of three experiments.

Externalized PS on Chlamydia-infected neutrophils enhances their phagocytosis by HMDMs. Since neutrophils are among the first leukocytes to encounter Chlamydia cells during acute infection (35) and since neutrophils externalize PS when inoculated (Fig. 7A), we tested whether PS on neutrophils would augment their uptake by human monocyte-derived macrophages (HMDMs). Infected neutrophils (Fig. 8C) were phagocytosed by glucan-stimulated HMDMs. PS-dependent phagocytosis is upregulated in HMDMs by including -glucan, a glucose polymer, in the culture medium (18). Glucan-stimulated HMDMs recognized Chlamydia-inoculated neutrophils through a PS-dependent mechanism, as PS liposomes (Fig. 8C) reduced phagocytosis.

PS on Chlamydia-infected cells accelerates plasma clotting. Activated monocytes/macrophages express tissue factor (TF) enzyme, which is involved in initiating coagulation (50), and we tested whether PS on C. trachomatis-inoculated THP-1 cells would accelerate clot formation in a plasma-clotting assay. Infected THP-1 cells accelerated clotting in platelet-poor plasma, and recombinant annexin V abolished enhancing activity (Fig. 8D). The clotting activity present in uninfected THP-1 cells was likely due to basal TF expression (41).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Eukaryotic cells externalize PS while undergoing death, differentiation (67), and infection by influenza virus (21) and L. pneumophila (23). Here, we show that infection of human cells with the obligate intracellular bacterium Chlamydia elicits PS exposure. While induction of maximal PS externalization requires extracellular calcium, Chlamydia-induced PS exposure differs significantly from those induced by chemical agents or antibodies to the Fas-TNF receptor family. First, Chlamydia cells immediately induced PS exposure (<5 min), whereas staurosporine or ethanol required 1 to 2 h to elicit PS. Second, Chlamydia-induced PS exposure was transient; other inducers initiated a sustained PS exposure after stimulus withdrawal. Third, Chlamydia-induced PS exposure was unaffected by a broad-spectrum caspase inhibitor which represses PS exposure mediated by Fas-TNF receptor ligation or DNA-damaging agents. These findings demonstrate the unique nature of PS exposure elicited by Chlamydia.

Our results differ from those observed in monocytes/macrophages infected with the intracellular bacterium L. pneumophila (23), as PM permeability increased at the same time of annexin V binding, a finding consistent with necrosis rather than apoptosis. PS access may have occurred via membrane injury due to the high L. pneumophila infectious titers employed (>20 IFU of L. pneumophila/cell versus 2 IFU of Chlamydia/cell) and may not represent a specific response to infection. Influenza-infected epithelial cells externalize PS 9 to 12 h postinfection; however, significant PM permeability was also observed (22). CD47 ligation on activated T cells results in rapid (~30 min) PS externalization (45); again, an increase in PM permeability was seen. Clearly, Chlamydia-mediated PS externalization differs from that induced by other agents; this may reflect activation of a different host signaling pathway(s).

Our results imply that chlamydial binding is required to induce host cell PS exposure. Chlamydia-mediated PS externalization is contact dependent (heparin inhibitable), is localized to the outer PM leaflet, and is induced in four human cell types as well as the mouse connective tissue-derived line L929 (data not shown).

The rapid onset and disappearance of PS after inoculum withdrawal (Fig. 4B and C) may reflect endocytosis-related changes in the lipid composition of the PM, as Chlamydia cells are ingested within 5 min after cell contact (43). Increased cyclic GMP and calcium influx increase the efficiency of chlamydial entry (40), and tyrosine phosphorylation of host proteins occurs 15 min after infection (6, 20). Rapid PS externalization upon contact suggests that Chlamydia cells initiate host cell signaling during invasion.

PS exposure is an early marker of apoptosis (38) and is associated with the proinflammatory triggering of the alternative pathway of complement activation (39, 65). We did not address whether these two processes occurred in Chlamydia-infected cells. Induction of pro- or antiapoptotic phenotypes in host cells by Chlamydia remains unclear. Chlamydia-induced apoptosis has been reported (25, 42); however, another study reported induction of an antiapoptotic phenotype (19). PS exposure is associated with apoptosis, but we did not find a link between caspase activation and Chlamydia-mediated PS exposure. As Chlamydia-triggered PS exposure occurs rapidly after inoculum addition, it may be that caspase activation is not required or is bypassed. Our findings agree with those of Ojcius et al. (42), in which the appearance of apoptosis markers in Chlamydia-infected cells were not attenuated by caspase inhibitors. Although we cannot rule out the involvement of caspase(s) in PS exposure elicited by Chlamydia, it appears that their contribution to this process is minor.

Neutrophils kill C. trachomatis cells (49, 67) and have a role in preventing their dissemination (4). Macrophages fed PS-expressing apoptotic neutrophils or opsonized zymosan release soluble Fas ligand, which is capable of killing neutrophils, monocytes, eosinophils, and lymphocytes (9). We demonstrated that Chlamydia-infected neutrophils are phagocytosed by macrophages in a PS-dependent manner. Our results agree with those of Fadok et al. (18), in which uptake of PS-positive apoptotic neutrophils by glucan-stimulated HMDMs was strongly inhibited by PS liposomes, but not RGDS peptide. That RGDS reduced uptake in our experiments may reflect the fact that fewer of our macrophages switched exclusively to PS-specific receptors. The rapid (<2 h) PS-mediated HMDM phagocytosis of neutrophils we have seen may represent an exploitation of host homeostatic mechanisms by Chlamydia, in which the bacterium rids itself of a noxious opponent(s), but whether phagocytosis of infected neutrophils benefits or harms the host during a natural infection is unknown.

Segregation of macrophages and dendritic cells into "classically" and "alternatively" activated phenotypes (26) is hypothesized to explain how these antigen-presenting cells (APCs) generate proinflammatory ("classical") or anti-inflammatory ("alternative") responses and modify immune outcomes. APCs acquire alternative activation phenotypes when they ingest PS-exposing apoptotic cells (17, 51, 64). Alternatively activated APCs secrete anti-inflammatory cytokines, enhance endocytic and antigen presentation capacities, up-regulate receptors for innate immunity, and reduce production of nitric oxide and reactive oxygens (26). Delayed proinflammatory cytokine release may affect chlamydial pathogenesis, as resident macrophages may ingest PS-expressing cells acutely infected with Chlamydia. As interleukin-1 (IL-1), IL-8, and TNF- are secreted by epithelial cells 24 h postinfection (48) and Chlamydia cells are rapidly internalized (5 min after contact; 43), there is little chance of proinflammatory macrophage activation by exogenous cytokines or free bacteria. After ingesting PS-expressing cells, which may be enhanced by PS-mediated C3bi deposition on target cells (39), macrophages express antiinflammatory cytokines (17, 64), release soluble Fas ligand (9), and are poor stimulators of T and B cells (64), all hallmarks of alternative activation status.

Procoagulant states are induced in monocytes and endothelial cells after stimulation with lipopolysaccharide, TNF-, or IL-1 (14, 50, 53, 70) and, in this report, with Chlamydia-infected THP-1 promonocytes. Thrombin formation, a central event in coagulation and inflammation, catalyzes fibrin generation and binds thrombin receptors, activating transcription in vascular cells, including monocytes (12) and endothelial cells (1, 60). Fibrin deposition, associated with wound repair, inflammation, and coagulation (13), attracts platelets, neutrophils, and leukocytes (34) and may initiate vascular plaques (5, 52). Thus, cellular PS exposure following chlamydial infection may modulate the inflammatory, tissue remodeling, and immune responses of the host.

Chlamydia-infected mesothelial and endothelial cells upregulated TF, PAI-1 (22, 63), and platelet adhesion (22) activities several hours postinfection. Our data, with which we show that infected epithelial and aortic endothelial cells immediately externalize PS, may provide a molecular understanding of the results reported by Fryer et al. (21) and van Westreenen et al. (63), in which C. trachomatis-infected cells acquired factor Xa-generating activity. We showed that bacterial metabolism was not a prerequisite, as chloramphenicol-treated Chlamydia cells induced PS exposure. Induction of TF, PAI-1, and PS suggests that Chlamydia-inoculated endothelium can initiate and sustain coagulation reactions. Platelets (32), monocytes (33), neutrophils (34), and T cells (46), elements commonly found in atherosclerotic plaques, adhere to thrombin-treated endothelial cells. PS exposure, leading to macrophage-mediated phagocytosis of and procoagulant activity by infected cells, may be important in the genesis of chlamydial dissemination and systemic diseases, including atherosclerosis caused by Chlamydia.