INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Haemophilus ducreyi is a gram-negative, obligate human pathogen that causes chancroid, a sexually transmitted genital ulcer disease. Chancroid lesions involve cells of the epidermis and dermis and contain an immune cell infiltrate consisting of polymorphonuclear leukocytes (PMNs), T cells, and macrophages (2, 25, 26, 42, 43). Despite this immune cell infiltrate, viable H. ducreyi can be isolated from these ulcers weeks or months after initial infection (31). Inguinal lymphadenopathy occurs in up to 50% of untreated chancroid cases, and viable H. ducreyi can be isolated from infected lymph nodes (31). The mechanism used by H. ducreyi to persist and cause disease at these sites despite the presence of an apparently vigorous immune response is not understood.

Several potential virulence factors have been identified in H. ducreyi, including the cell-associated hemolysin (33, 48, 50), the secreted cytolethal distending toxin (8, 36), lipooligosaccharide (4), a hemoglobin uptake protein (11), a serum resistance protein (3, 12) and PAL, a peptidoglycan-associated lipoprotein (14). The target cell ranges of the two known H. ducreyi toxins, hemolysin and cytolethal distending toxin, have been studied in some detail. The cell-associated hemolysin lyses human foreskin epithelial cells (keratinocytes), fibroblasts, and immune cells including macrophages, T cells, and B cells and may thus contribute to tissue destruction and immune evasion (34, 50). The secreted cytolethal distending toxin is thought to contribute to immune evasion by inhibiting the proliferation of T cells and inducing apoptosis (17). Epithelial cells treated with cytolethal distending toxin become enlarged, arrest in the G₂ phase, and die several days following exposure (8, 9, 36). Both cytolethal distending toxin and hemolysin may thus contribute to the formation of ulcers, evasion of the immune response, and the delayed healing common to chancroid (8, 9, 36).

Because H. ducreyi survives in the presence of macrophages, both in chancroid lesions and in regional lymph nodes, we studied the interactions of H. ducreyi with the human macrophage-like cell line, U-937 (45). We found that H. ducreyi inhibits phagocytosis of both itself and a secondary phagocytosis target, opsonized sheep red blood cells (SRBC). The results of this study suggest a mechanism by which H. ducreyi avoids immune clearance in the human host.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. The bacterial strains and plasmids used in this study are described in Table 1. Plasmid pJL300, H. influenzae Rd, H. ducreyi strain CIP542, and H. ducreyi strain HMC56 were gifts from Eric Hansen (University of Texas Southwestern Medical Center), Marilyn Roberts (University of Washington), Leslie Slaney (University of Manitoba), and M. R. de Quinones (Instituto Dermatologico, Santo Domingo, Dominican Republic), respectively. H. ducreyi was cultured on chocolate agar plates or in Hd broth containing 10% fetal bovine serum (49); Escherichia coli and Salmonella enterica serovar Typhimurium were cultured on Luria-Bertani plates or broth (41). H. influenzae was grown on chocolate agar plates or in Hd broth lacking fetal bovine serum.

Bacterial adherence, uptake, and intracellular survival assays. The human macrophage-like cell line U-937 (45) was obtained from the American Type Culture Collection (Manassas, Va.) and cultured routinely in RPMI medium containing 10% heat-inactivated fetal bovine serum and 100 µg each of ampicillin and streptomycin per ml. Adherence to and phagocytosis by U-937 cells was measured as follows. U-937 cells were plated in 24-well tissue culture plates at a density of 5 × 10⁵ cells per well in medium containing 100 ng of phorbal myristate acetate (PMA; Sigma) per ml of to induce differentiation and adherence (23). After 48 h, the medium containing nonadherent cells was aspirated and the adherent U-937 cells were cultured for an additional 24 h in medium lacking PMA. The U-937 cells were washed three times in PBS, and then 1 ml of medium (without ampicillin and streptomycin) containing ~5 × 10⁶ CFU was added. The bacteria were centrifuged onto the U-937 cells at 150 × g for 10 min and then incubated at 33°C in humidified 5% CO₂-95% air for 30 min to allow phagocytosis or invasion. The nonadherent bacteria were then removed by washing three times with PBS. The U-937 cells in one set of three wells were lysed with trypsin-EDTA, and the total number of cell-associated (adherent and intracellular) bacteria was determined by dilution plate counts. The remaining wells were treated with gentamicin (30 µg/ml of RPMI) for 30 min to kill extracellular bacteria. The 0-h wells were then washed, the U-937 cells were lysed, and bacterial plate counts were done. The 6- and 24-h wells were incubated for an additional 5 or 23 h, respectively, in medium containing 3 µg of gentamicin per ml and then washed, the U-937 cells were lysed, and the bacteria were plated as described above. We have previously shown that H. ducreyi 35000, E. coli HB101, and S. enterica SL 1344 do not differ in their susceptibility to gentamicin (50).

Opsonization of SRBCs. SRBCs (150 µl of defibrinated sheep blood; PML Microbiologicals, Wilsonville, Oreg.) were washed in 150 mM NaCl, resuspended in 3 ml of 150 mM NaCl containing 6 µl of anti-SRBC antiserum (Rockland Immunochemicals), and then incubated at 37°C with shaking for 30 min. Opsonized SRBCs were then centrifuged for 5 min at 300 × g, resuspended in 3 ml of 150 mM NaCl, and added to the phagocytosis assay mixture as described below.

Phagocytosis assays. Phagocytosis assays were performed using U-937 cells treated with PMA as described above, except that the U-937 cells were allowed to adhere to glass coverslips placed in the tissue culture wells. Bacterial cultures were microcentrifuged for 1 min and then resuspended to an optical density at 540 nm of 0.7 to 1.0 in RPMI containing 10% fetal bovine serum. A 100-µl volume of bacterial suspension was added to each well, and the wells were incubated at 33°C in humidified 5% CO₂-95% air for 60 min. Note that the bacteria were not centrifuged onto the U-937 cells except in the dose response experiment (see Fig. 6), in which centrifugation for 10 min at 150 × g was used to synchronize contact with the U-937 cells. Opsonized or nonopsonized SRBCs (100 µl) were added, and the mixture was centrifuged for 2 min at 150 × g. The plate was then incubated at 37°C in humidified 5% CO₂-95% air for 60 min to allow phagocytosis. Extracellular SRBCs were lysed with distilled water as previously described (23). The coverslips were then fixed in methanol, mounted on glass slides, and stained with the Diff-Quik stain set (Dade Diagnostics, Aguada, Puerto Rico). The slides were viewed (blinded) by light microscopy, and the percentage of 100 to 200 cells containing intracellular SRBCs was determined (percentage of U-937 cells with intracellular SRBC).

Assay for cytolethal distending toxin activity. Supernatants of overnight cultures of H. ducreyi and E. coli strain DH5(pJL300) were assayed for cytolethal distending toxin activity against HeLa cells as previously described (8).

Fractionation of H. ducreyi. H. ducreyi was sonicated by the method of Carlone et al. (5). Briefly, overnight cultures were diluted 1:5 in Hd broth with 10% fetal bovine serum and incubated at 35°C with shaking for 4 to 5 h. A 20-µl sample of this logarithmic-phase culture was centrifuged at 6,000 × g for 10 min and then resuspended in 2 ml of 10 mM HEPES (pH 7.4). A 250-µl volume was set aside ("whole cells"; approximately 5 × 10⁸ CFU/ml), and the remaining suspension was sonicated (3 to 4 15-s bursts on ice) to disrupt the cells. The cell lysate was microcentrifuged for 2 min to remove unlysed cells and debris, and a 250-µl aliquot was set aside ("total-cell lysate"). The remaining lysate was microcentrifuged for 30 min at 4°C. The supernatant (1.5 ml; "soluble fraction") was collected, while the pellet ("membrane fraction") was suspended in 500 µl of 10 mM HEPES (pH 7.4). A 100-µl aliquot of each fraction was added to the phagocytosis assay mixtures as described above. The membrane fraction is thus concentrated threefold relative to the other fractions.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

H. ducreyi adheres efficiently to U-937 cells, but few are phagocytosed. We first performed experiments to investigate the adherence, uptake, and intracellular survival of H. ducreyi in the human macrophage-like cell line, U-937. Overnight (nonhemolytic) cultures of H. ducreyi strain 35000 were added to differentiated U-937 macrophages at a multiplicity of infection (MOI) of ~10 bacteria per macrophage. After 30 min the wells were washed to remove nonadherent bacteria and the total number of cell-associated (adherent and intracellular) bacteria was determined by plate counts. In a parallel experiment, extracellular bacteria were killed with gentamicin and the intracellular bacteria were enumerated. Avirulent E. coli HB101 and virulent S. enterica serovar Typhimurium SL1344 served as controls in this experiment. HB101 was expected to adhere to macrophages and be phagocytosed but not survive intracellularly, while SL1344 should adhere, invade, and multiply intracellularly.

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Interaction of H. ducreyi, E. coli, and S. enterica serovar Typhimurium with U-937 macrophages. Adherence to (A) or phagocytosis (B) by U-937 macrophages of E. coli HB101, S. enterica SL1344, and H. ducreyi 35000 is shown. Bacteria at an approximate MOI of 10 were centrifuged onto macrophages, and the number of cell-associated or intracellular bacteria was determined as described in Materials and Methods. Cell-associated bacteria are shown as a percentage of the total inoculum, while intracellular bacteria are shown as a percentage of the cell-associated bacteria. The mean and standard error for a typical experiment performed in triplicate are shown; the experiment was performed three times with similar results.

Intracellular H. ducreyi are rapidly killed by U-937 cells. To investigate the fate of the few H. ducreyi that were phagocytosed by U-937 cells, we performed a time course experiment to monitor the survival of intracellular bacteria over a 24-h period. As shown in Fig. 2A, most (~70%) of the intracellular E. coli HB101 bacteria were killed during the 24-h incubation period. As expected, S. enterica SL1344 multiplied in U-937 cells, consistent with the known virulence capabilities of this organism (Fig. 2B) (16). In comparison, intracellular H. ducreyi strain 35000 were rapidly killed by the U-937 cells (Fig. 2C); after 6 h the number of intracellular bacteria was reduced by 99.8% and no intracellular H. ducreyi bacteria were viable after 24 h. Thus, H. ducreyi, like E. coli, did not replicate or survive in U-937 cells but instead were killed.

View larger version (15K): [in this window] [in a new window]   FIG. 2.   Intracellular survival of H. ducreyi in U-937 macrophages. The intracellular survival of E. coli HB101 (A), S. enterica SL1344 (B), or H. ducreyi 35000 (C) was monitored over time as described in Materials and Methods. The number of intracellular bacteria as a percentage of the total cell-associated bacteria is shown. The mean and standard error for a typical experiment performed in triplicate are shown; the experiment was performed three times with similar results.

H. ducreyi inhibits phagocytosis by U-937 macrophages. Our observation that few H. ducreyi bacteria are internalized by U-937 cells suggested that H. ducreyi is able to avoid or inhibit phagocytosis. To test the latter possibility, we developed an assay to measure the phagocytosis of a secondary phagocytosis target (opsonized SRBCs) after exposure of U-937 cells to H. ducreyi. We used this assay because (i) extracellular (nonphagocytosed) SRBCs can be lysed with distilled water, thus simplifying the identification of U-937 containing intracellular SRBCs (23), and (ii) inhibition of phagocytosis can be measured separately from bacterial killing by the U-937 cells. To avoid lysis of the target SRBCs by H. ducreyi hemolysin, a nonhemolytic mutant (35000APC) derived from 35000 (50) was used in these experiments. Strain 35000APC did not differ from wild-type strain 35000 in the ability to inhibit phagocytosis (see below). PMA-treated U-937 cells were exposed to cultures of H. ducreyi strain 35000APC or controls of cytochalasin B (an inhibitor of actin polymerization) or dimethyl sulfoxide (DMSO, solvent for cytochalasin B) and then allowed to phagocytose SRBCs. Extracellular SRBCs were lysed, and the percentage of U-937 cells containing intracellular SRBCs was determined by light microscopy. As shown in Fig. 3, approximately 30% of untreated (medium -only) U-937 cells phagocytosed opsonized SRBCs while only ~2% of U-937 cells phagocytosed nonopsonized SRBCs (data not shown), indicating that opsonization is required for phagocytosis of SRBCs by U-937 cells. The solvent control, DMSO, had no effect, while cytochalasin B treatment inhibited the phagocytosis of opsonized SRBCs. Treatment of U-937 cells with H. ducreyi strain 35000APC at a multiplicity of infection (MOI) of approximately 70 CFU per macrophage inhibited the phagocytosis of SRBCs to a level similar to that caused by cytochalasin B. This effect was observed with log-phase and stationary-phase cultures grown in broth and, to a lesser extent, with H. ducreyi grown on plates. Treatment of U-937 cells with avirulent E. coli HB101 or avirulent H. influenzae strain Rd had no effect on the phagocytosis of SRBCs (Fig. 3) suggesting that inhibition of phagocytosis was specific to H. ducreyi.

View larger version (24K): [in this window] [in a new window]   FIG. 3.   Inhibition of phagocytosis by H. ducreyi. Differentiated U-937 macrophages were treated with DMSO, cytochalasin B (cyto B), logarithmic (log), overnight (ON), or plate-grown (plate) cultures of H. ducreyi, or logarithmic cultures of E. coli HB101 (HB101) or H. influenzae strain Rd (Rd) as indicated for 1 h and then exposed to opsonized SRBCs for 1 h. Extracellular SRBCs were lysed, and phagocytosis was determined by light microscopy. The percentage of U-937 cells with intracellular SRBCs is shown. The mean and standard error for a typical experiment performed in triplicate are shown; the experiment was performed at least three times with similar results.

Characterization of the antiphagocytic activity of H. ducreyi. To determine whether the antiphagocytic activity of H. ducreyi was secreted or cell associated, U-937 cells were pretreated with whole cells or supernatants from overnight cultures of H. ducreyi 35000APC. As shown in Fig. 4, the antiphagocytic activity was associated only with whole cells and not with culture supernatants. Supernatants from E. coli DH5(pJL300) expressing plasmid-encoded cytolethal distending toxin (44) were also unable to inhibit phagocytosis by U-937 cells (Fig. 4). Both H. ducreyi 35000APC and E. coli DH5(pJL300) culture supernatants contained active cytolethal distending toxin as determined by the HeLa cell assay (reference 8 and data not shown). These results suggest that the antiphagocytic activity requires cell contact and that the secreted cytolethal distending toxin, known to act on immune cells (17), is not sufficient to inhibit phagocytosis.

View larger version (21K): [in this window] [in a new window]   FIG. 4.   Inhibition of phagocytosis by H. ducreyi cells and supernatants. U-937 macrophages were treated with the samples indicated for 1 h and then allowed to phagocytose opsonized SRBCs for an additional 1 h. cells, washed H. ducreyi cells; spnts, culture supernatants. The mean and standard error for a typical experiment performed in duplicate are shown; the experiment was performed three times with similar results.

View larger version (23K): [in this window] [in a new window]   FIG. 5.   Time course of inhibition of phagocytosis. U-937 macrophages were treated with H. ducreyi 35000APC at an MOI of 50 to 100 CFU per macrophage for the times indicated. Ampicillin and streptomycin were added to kill H. ducreyi, and the macrophages were allowed to phagocytose opsonzed SRBC for 1 h. The mean and standard error for a typical experiment performed in duplicate are shown; the experiment was performed three times with similar results.

View larger version (27K): [in this window] [in a new window]   FIG. 6.   Dose response of inhibition of phagocytosis. U-937 macrophages were treated with the dose of H. ducreyi 35000APC shown for 1 h then allowed to phagocytose opsonized SRBCs for 1 h. The mean and standard error for a typical experiment performed in duplicate are shown; the experiment was performed three times with similar results.

Antiphagocytic activity is a common characteristic of H. ducreyi isolates. Eight H. ducreyi isolates were tested for their ability to inhibit phagocytosis by U-937 cells. These strains (Table 1) were isolated at different times from various geographical locations and also differed in their antimicrobial susceptibilities, ribotyping patterns, and plasmid content (reference 20 and data not shown). As shown in Fig. 7, all of the strains tested inhibited phagocytosis, suggesting that this potential virulence trait is conserved among H. ducreyi strains. Interestingly, H. ducreyi strain CIP A75 inhibited phagocytosis to a lesser degree than the other strains did.

View larger version (26K): [in this window] [in a new window]   FIG. 7.   Inhibition of phagocytosis by various strains of H. ducreyi. U-937 macrophages were untreated or treated with the strains indicated (MOI ~ 100 bacteria per macrophage) for 1 h and then allowed to phagocytose opsonized SRBCs. The mean and standard error for a typical experiment performed in duplicate are shown; the experiment was performed twice with similar results.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In vivo, H. ducreyi is found in association with macrophages in naturally acquired chancroid lesions and in the human experimental infection model (2, 25, 26, 31, 42, 43). To understand the mechanisms by which H. ducreyi persists and survives in the presence of these phagocytes, we explored the in vitro interactions of H. ducreyi with the human macrophage-like cell line, U-937. In this study, we found that H. ducreyi strain 35000 adhered to U-937 cells efficiently but resisted phagocytosis. The few H. ducreyi bacteria that were phagocytosed in our experiments were killed by the U-937 cells within 24 h, suggesting that H. ducreyi lacks the ability to survive intracellularly under these conditions. Using a secondary target, opsonized SRBCs, we were able to show that H. ducreyi inhibits Fc-mediated phagocytosis by U-937 macrophages.

We characterized the antiphagocytic activity of H. ducreyi and found that it is associated with whole cells and not with culture supernatants, whole-cell lysates, soluble or membrane fractions, or heat-killed bacteria. Expression of antiphagocytosis was not growth phase dependent and was found in bacteria grown either on plates or in broth. The antiphagocytic activity acts quickly since a 15-min exposure of U-937 cells to H. ducreyi was sufficient to inhibit phagocytosis. Low, physiologically relevant MOIs (~1 bacterium per macrophage) were sufficient to inhibit phagocytosis, while higher MOIs were more inhibitory. Based on these results, we hypothesize that the antiphagocytic activity of H. ducreyi is active in vivo and enhances the ability of H. ducreyi to avoid clearance by the macrophages found in chancroid lesions. Our finding that H. ducreyi inhibits the phagocytosis of secondary targets (opsonized SRBCs) suggests that antiphagocytic activity may also contribute to the establishment of secondary infections often found in chancroid lesions, because the ability of macrophages to phagocytose other pathogens may be impaired. Antiphagocytosis might also be a mechanism for survival of H. ducreyi in inguinal lymph nodes.

Other organisms can avoid the killing mechanisms of phagocytes by inhibiting attachment to phagocytes or by inhibiting phagocytosis itself. Production of a capsule or slime layer prevents the attachment and uptake of a number of organisms including Streptococcus pyogenes, S. pneumoniae, Cryptococcus neoformans, and Klebsiella pneumoniae (30). S. pyogenes M protein, considered the major virulence factor of group A streptococci, prevents phagocytosis by reducing the opsonization of streptococci by complement (22). Similarly, YadA of Yersinia enterocolitica prevents opsonization by complement and thereby avoids uptake and killing by PMNs (7). These mechanisms are different from the antiphagocytic activity expressed by H. ducreyi, which adhered to U-937 cells and prevented the phagocytosis of a secondary target, opsonized SRBCs. While we cannot exclude the possibility that H. ducreyi avoids opsonization (possibly mediated through DsrA [12]), the antiphagocytic activity described here is distinct since complement was omitted in our experiments.

Other bacteria use toxins to inhibit phagocytosis. For example, the E. coli alpha-hemolysin inhibits leukocyte function, including phagocytosis, at sublytic doses (6), while the YopE and ExoS cytotoxins of Yersinia species and Pseudomonas aeruginosa, respectively, disrupt actin microfilaments, thereby avoiding uptake by macrophages (15, 40). We were unable to detect overt toxicity in our assays since treated macrophages were viable (by trypan blue exclusion) in our assays. The H. ducreyi cytolethal distending toxin is known to act on immune cells, but supernatants containing active secreted cytolethal distending toxin were unable to inhibit phagocytosis in our experiments. Furthermore, whole-cell sonicates, previously shown to contain active cytolethal distending toxin (36), did not inhibit phagocytosis, suggesting that cell-associated cytolethal distending toxin is also not sufficient for antiphagocytosis. However, definitive proof awaits experiments comparing the antiphagocytic activity of wild-type H. ducreyi whole cells to cells of a mutant deficient in cytolethal distending toxin expression. Our experiments also suggest that the cell-associated hemolysin does not play a predominant role in antiphagocytosis since a hemolysin mutant, 35000APC, retains antiphagocytic activity.

Perhaps the best understood mechanism of antiphagocytosis in macrophages is that used by Yersinia species. The Yersinia YopH protein is a tyrosine phosphatase that is specifically translocated into target cells via a contact-dependent type III secretion system, where it interferes with host cell phosphotyrosine signaling (1, 35). Mutants lacking YopH are unable to inhibit phagocytosis and are avirulent (1, 13, 39). Similar to H. ducreyi, Yersinia species not only prevent their own uptake but also inhibit the uptake of other phagocytosis targets (13). Inhibition of phagocytosis by enteropathogenic E. coli is similar to that of Yersinia spp. in that a type III secretion system is required and dephosphorylation of host proteins occurs (18). However, no tyrosine phosphatase activity is associated with the enteropathogenic E. coli-secreted proteins, suggesting that E. coli uses a different mechanism from Yersinia to alter macrophage protein phosphorylation and thus inhibit phagocytosis (18). BLAST searches of the recently completed H. ducreyi genome failed to identify homologs of a type III secretion system and the known antiphagocytic proteins YopH, YopE, and M protein, suggesting these mechanisms may not exist in H. ducreyi (data not shown). Finally, Helicobacter pylori inhibits phagocytosis via a mechanism that involves a type IV secretion system, although the secreted factor responsible for antiphagocytosis is not yet known (37). Actinobacillus actinomycetemcomitans contains a gene cluster (tadA-G) involved in adherence, and the TadA protein is similar to a component of type IV secretion systems (24). The tad gene cluster is also present in H. ducreyi, where it plays a role in adherence (E. J. Hansen, personal communication), but the role of these gene products in the antiphagocytic activity of H. ducreyi has not yet been examined.

The work described here complements that of other groups who studied the interactions of H. ducreyi with phagocytes in vitro and in vivo (2, 28, 32). Odumeru et al. demonstrated that virulent H. ducreyi strains resist phagocytosis and killing by PMNs while avirulent strains do not (32). We tested two of their avirulent strains, CIP A75 and CIP A77, in our macrophage phagocytosis assay and found that both strains were antiphagocytic although CIP A75 was less inhibitory than the other strains. Lagergard et al. (28) also studied the interactions of PMNs with H. ducreyi and found that H. ducreyi resists killing by PMNs but that killing could be greatly increased by the addition of human serum containing either complement or anti-H. ducreyi antibodies. However, even in the presence of 70% fresh serum, 10% of the bacteria survived exposure to PMNs, suggesting the presence of a mechanism to avoid phagocytic killing (28). It will be interesting to perform similar experiments using human macrophages to study the effects of normal and immune serum on the phagocytosis of H. ducreyi. Finally, Bauer et al. (2) found that H. ducreyi associates with phagocytes (macrophages and PMNs) in experimental human lesions but remains extracellular. The authors suggest that H. ducreyi resists phagocytic killing in vivo, a hypothesis supported by our in vitro data.

Although determination of the in vivo role of antiphagocytosis awaits further experiments, the following observations support a role for antiphagocytosis in H. ducreyi virulence: (i) all strains tested were antiphagocytic; (ii) inhibition of phagocytosis occurs at low, physiologically relevant MOIs; and (iii) phagocytosis is inhibited shortly after exposure to H. ducreyi. Furthermore, H. ducreyi is found in association with macrophages in vivo, suggesting that a mechanism must exist to avoid killing by these cells. The goal of future studies will be the elucidation of the antiphagocytic mechanism including identification of the gene product(s) involved and the in vivo importance of this potential virulence trait in the animal models of chancroid.