Two Distinct Receptors Mediate Nonopsonic Phagocytosis of Different Strains of Pseudomonas aeruginosa

Abstract

Complement receptor 3 (CR3) mediates both opsonic and nonopsonic phagocytosis of bacteria. Leukocyte adhesion deficiency (LAD) allows for the study of CR3-dependent phagocyte-bacterial ingestion, since LAD phagocytes do not express this receptor. Phagocytes from an infant with LAD were unable to ingest 50% of the Pseudomonas aeruginosa strains studied, which indicates a requirement for CR3. However, the remaining strains were phagocytosed in the absence of CR3, and ingestion was blocked by monoclonal antibodies directed at CD14. This CR3/CD14 receptor bias was further confirmed by using thioglycollate-elicited murine peritoneal macrophages, which have nonfunctional CR3 before activation. Results indicate that either CR3 or CD14 is involved independently in nonopsonic phagocytosis of different P. aeruginosa strains. Clearance of P. aeruginosa from the endobronchial space may be facilitated by nonopsonic phagocytosis, since low levels of opsonins are present. The impact of lung infection with P. aeruginosa may be determined, in part, by the phagocytic receptor that mediates ingestion

We previously described a newborn infant who was admitted to British Columbia's Children's Hospital in 1999 with omphalitis, perianal abscess, and neutrophilia [1] and a clinical diagnosis of leukocyte adhesion deficiency (LAD) syndrome type 1 [2]. Flow cytometry revealed the absence of CD18, the common subunit of β₂-integrins on the surface of leukocytes, which confirmed the diagnosis [1, 3]. In this disease, the CD18-containing integrins are absent from leukocytes, including leukocyte function-associated antigen 1 (CD11a/CD18), complement receptor 3 (CR3; CD11b/CD18), and complement receptor 4 (CR4; CD11c/CD18). The functional defect results from the failure of leukocytes to adhere firmly to the vascular endothelium, which prevents migration to extravascular sites of infection [4, 5]. CR3-mediated phagocytosis also is attenuated

We have found that CR3 has a central role in nonopsonic phagocytosis of Pseudomonas aeruginosa by murine alveolar macrophages (authors' unpublished data). This organism is an opportunistic pathogen in the lungs of persons with cystic fibrosis (CF) and in persons undergoing mechanical ventilation. Since there are low levels of opsonins in the lung [6–8], the initial inflammatory response after bacterial challenge is probably through nonopsonic phagocytosis by alveolar macrophages and epithelial cells before the influx of opsonins and neutrophil recruitment, which is a result of inflammation. Although the dynamics of this process have not been established, resident alveolar phagocytes may be critical for general housekeeping in the lung in response to continual bacterial exposure. Neutrophils are probably the primary effector in the clearance of P. aeruginosa after recruitment through inflammation

Other mechanisms also may play a role in defense of the lung against bacterial infection, including surfactant proteins that promote phagocytosis of P. aeruginosa in the lung [9, 10] and have a role in resistance to infection in animals [11]. There also may be a functional classical complement pathway in the lung [12], but a major problem for complement-mediated phagocytosis in the lung is the lack of complement receptors on alveolar macrophages [13]. In addition, it appears that alveolar macrophages poorly bind P. aeruginosa even in the presence of serum [14]. The relative contribution of these various mechanisms requires further clarification. However, nonopsonic phagocytosis is a very efficient method of microbial ingestion and seems to be well preserved between human and murine phagocytes, which suggests that there has been some evolutionary pressure for its preservation. Others also have suggested that nonopsonic phagocytosis is important in microbial defense of the lung [15]. When there is exposure to small numbers of infecting microbes (which happens continually [16]) or before widespread inflammation occurs in response to bacterial invasion, it seems likely that nonopsonic phagocytosis may play a central role in lung defense

We recently observed that a P. aeruginosa isolate from a patient with CF was not ingested by neutrophils isolated from the patient with LAD, which supports the view that CR3 plays a central role in phagocytosis of P. aeruginosa in the absence of opsonin [1]. We have proposed that the interaction between CR3 and P. aeruginosa may be important in determining the outcome in patients with CF; however, the mechanisms of P. aeruginosa phagocytosis are complex. We found that a strain isolated from an abscess on the infant with LAD was actively phagocytosed by LAD phagocytes [1], which suggests that different P. aeruginosa strains may be ingested by different phagocyte surface receptors. To further explore this finding and to identify alternative receptors that might be involved in phagocytosis of this organism, we studied phagocyte-bacterial interactions in both normal and CR3-deficient neutrophils and macrophages and multiple P. aeruginosa strains

Materials and Methods

Bacterial strains.P. aeruginosa P1 is a motile, nonmucoid revertant of a mucoid isolate from a patient with CF [17]. Strain 808 was cultured from a dermal abscess of an infant with LAD [1]. All other strains studied were from a collection of clinical, environmental, and laboratory P. aeruginosa strains stored at −70°C (table 1). Bacteria were grown overnight in Luria-Bertani broth (10 g of bacto-tryptone, 5 g of yeast extract, 10 g of NaCl, and 1000 mL of dH₂O [pH 7.0]) at 37°C in a shaking incubator. The bacteria were adjusted to an absorbance of 1.0 at 600 nm in Hanks' balanced salt solution (HBSS) and then were pelleted and resuspended in the same volume of HBSS

Table 1

Origin of Pseudomonas aeruginosa strains used in phagocytosis studies, phagocyte receptor required for ingestion, and phagocytosis enhancement by glucose of thioglycollate-elicited macrophages

Isolation of human neutrophils and monocytesHeparinized venous blood was obtained from the patient with LAD, 2 adult control subjects, and umbilical cords of 2 healthy neonates. Leukocytes and erythrocytes were separated by ficoll-hypaque density gradient centrifugation. The neutrophil and erythrocyte pellet was resuspended in 5–10 mL of lysing buffer (0.87% ammonium chloride) for 10 min at 37°C, was washed, and was resuspended in HBSS with 0.1% gelatin (gHBSS). Monocytes were washed and resuspended in RPMI (Sigma) supplemented with penicillin (100 U/mL; ICN Pharmaceuticals), streptomycin (100 μg/mL; ICN Pharmaceuticals), and 10% autologous serum, before incubation in Teflon beakers for 96 h at 37°C in a 5% CO₂ environment

Murine macrophagesResident peritoneal macrophages were harvested from 6-week-old female BALB/c mice as follows. We used a 5-cc hypodermic syringe and a 25-gauge needle to inject 5 mL of Dulbecco's MEM (Gibco-BRL) into the peritoneal cavity and then gently massaged the abdomen for 1 min. The fluid within the peritoneum was subsequently removed with a hypodermic syringe and an 18-gauge needle. Peritoneal exudate cells (PECs) were quantified by hemocytometer and, after pelleting by centrifugation at 400 g were resuspended to give a volume of 6×10⁶ PECs/mL. Thioglycollate-elicited murine peritoneal macrophages were obtained in a similar manner, except that 1 mL of 4% Brewer's complete thioglycollate broth was injected intraperitoneally, and macrophages were harvested after 3 days [18]. Adherent cells were used for phagocytosis assays

Flow cytometryHuman neutrophils or macrophages were suspended in wash buffer (RPMI, 5% fetal calf serum; Gibco-BRL) at a density of 10⁵ cells/mL. Cells were harvested by centrifugation and were resuspended in 100 μL of wash buffer containing 1 μg of primary antibody (fluorescein isothiocyanate or phycoerythrin labeled) directed against human CD antigens (PharMingen; table 2). After incubation for 45 min on ice, the volume was adjusted to 1 mL with wash buffer, and leukocytes were pelleted by centrifugation (400 g). The wash was repeated twice, once with wash buffer and then with ice-cold PBS. The harvested leukocytes were resuspended gently in 250 μL of PBS and 250 μL of 3% paraformaldahyde in PBS. Labeled leukocytes were stored in the dark at 4°C until sorted (FACSCalibur; Becton Dickinson), according to the manufacturer's protocols

Table 2

Assessment of CD antigen expression on phagocytic cells from a patient with leukocyte adhesion deficiency (LAD) and control subjects

Assessment of phagocytosisPhagocytosis assays for human neutrophils or macrophages were done, as described elsewhere [1]. Serum was removed before the phagocytosis assay, which was done with cells in suspension rather than on microscope coverslips, since CR3-deficient human phagocytes do not adhere to glass. Controls and LAD phagocytes were handled in the same way. We added ∼5×10⁵ phagocytes and 1.5×10⁸ bacteria to HBSS in a final volume of 3 mL, which were rotated (30 rpm) at 37°C for 1 h. An identical tube was prepared and was incubated at 4°C for 1 h with frequent inversions. Phagocytes were pelleted by centrifugation (400 g for 10 min at 4°C) and were resuspended in 1 mL of ice-cold lysozyme solution (5 mg/mL lysozyme and 0.25 Tris [pH 8.0]), to disrupt extracellular bacteria. The phagocytes were washed twice with HBSS, were resuspended in 100 μL of gHBSS, and were mounted onto a glass microscope slide by cytocentrifugation (Cytospin 2; Shandon). The slides were air dried overnight and were stained with 2% Giemsa (BDH Laboratory Supplies). Phagocyte-associated bacteria were enumerated by light microscopy (mean bacteria/phagocyte; n=60). Purity of neutrophil populations was determined by morphology; macrophages were distinguished from lymphocytes by morphologic characteristics

Phagocytosis was determined by 2 observers who were unaware of the strain type by subtracting the mean number of bacteria per phagocyte associated at 4°C from the mean bacteria associated per phagocyte at 37°C. Because phagocytosis does not occur at temperatures <18°C, the mean number of associated bacteria at 4°C represented those that were adherent to the macrophage surface membrane [17]. We used lysozyme to disrupt clumps of bacteria. Osmotic shock with distilled water is used typically to lyse extracellular bacteria, as described elsewhere [19], but this step was omitted, because it disrupts human neutrophils and macrophages

Phagocytosis by murine macrophages was assessed by using glass-adherent macrophages, as described elsewhere [20]. In brief, 3×10⁵ macrophages were matured for 1 day on glass microscope coverslips in RPMI at 37°C in a 5% CO₂ environment. The phagocytosis assay then was performed by incubating bacteria that were prepared, as described above, with adherent macrophages on coverslips for 1 h in HBSS. At the end of the assay, the coverslips were fixed and stained, as described above. The assay was done in HBSS with and without glucose. Assays that used antibodies to block phagocytosis of P. aeruginosa strains were identical to those described above, except that the blocking antibody was incubated with the phagocytes at a concentration of 10 μg/mL for 10 min before the addition of the bacteria. CR3-mediated phagocytosis was assessed by using complement-coated sheep erythrocytes (EIgMC), as described elsewhere [21]

Results

Flow cytometryExpression of surface antigens shared by both human neutrophils and monocytes was assessed by means of flow cytometry (table 2). We are unaware of any previous published description of LAD surface receptor expression. Neutrophils and macrophages from the infant with LAD expressed negligible CD18 and little CD11b, which demonstrates deficient CR3 expression. CD11c, the α subunit of CR4 (CD11c/CD18), also was not detected on the surface of the LAD phagocytes. All other CD antigens were expressed at similar levels on the LAD phagocytes as on phagocytes from adults or from an age-matched control infant

PhagocytosisOf the 10 P. aeruginosa strains studied, 5 (including P1) were not ingested by neutrophils or macrophages from the patient with LAD but were ingested by phagocytes from an adult or from cord blood (figures 1 and 2). For each of these strains and for EIgMC, there was a significant difference (P<.05) in levels of ingestion by CR3-deficient and normal phagocytes. The remaining 5 strains of P. aeruginosa including the patient's strain (808), were ingested by LAD phagocytes and by both adult and cord blood phagocytes with similar efficiency (P>.05)

Figure 1

Neutrophil phagocytosis of Pseudomonas aeruginosa. Ten P. aeruginosa strains were assessed for susceptibility to phagocytosis by leukocyte adhesion deficient (LAD) and normal neutrophils. Strains that were poorly phagocytosed by LAD neutrophils are indicated (⊗; P<.05 vs. normal neutrophils). Controls were neutrophils from 2 adults and 2 healthy neonates. The normal adult neutrophils and neonatal umbilical cord neutrophils did not display significantly different levels of phagocytosis and are reported together. Strain V119 is not shown, since it was cytotoxic to neutrophils. A control particle, EIgMC, was included to permit assessment of CR3-dependent phagocytosis. Error bars indicate SE from 4 experiments (mean phagocytosis of 60 neutrophils; n=4)

Figure 2

Macrophage phagocytosis of Pseudomonas aeruginosa. Ten strains of P. aeruginosa were assessed for susceptibility to phagocytosis by leukocyte adhesion deficient (LAD) and normal macrophages. The same strains that were not ingested by LAD neutrophils also were not ingested by LAD macrophages (⊗; P<.05 vs. controls). Controls were neutrophils from 2 adults and 2 healthy neonates. The normal adult neutrophils and neonatal umbilical cord neutrophils did not display significantly different levels of phagocytosis and are reported together. A control particle, EIgMC, was included to permit assessment of CR3-dependent phagocytosis. Error bars show SE of 4 experiments (mean phagocytosis of 60 macrophages; n=4)

A panel of antibodies to the phagocyte surface receptors (PharMingen) that are common to both human macrophages and neutrophils (table 2) was used to assess their capacity to block phagocytosis by LAD phagocytes. Only 1 of the 13 antibodies, anti-CD14, blocked phagocytosis of strain 808. Anti-CD14 antibody, incubated with LAD neutrophils, blocked phagocytosis by ∼50% (figure 3). Anti-CD14 incubated with normal adult neutrophils also blocked phagocytosis of strain 808 (figure 3). Anti-CD14 did not effect phagocytosis of strain P1 by adult neutrophils, but antibodies directed against CR3 blocked phagocytosis of this strain (figure 3). None of the other antibodies blocked phagocytosis of either bacterial strain by any phagocytic cell tested (data not shown)

Figure 3

Antibody blocking of Pseudomonas aeruginosa phagocytosis. Phagocytosis of P. aeruginosa strains P1 and 808 were assessed after leukocyte adhesion deficient (LAD) or normal adult neutrophils were preincubated with monoclonal antibodies to CR3 and/or CD14 (see Materials and Methods). Error bars show SE of 2 experiments (mean phagocytosis of 60 neutrophils; n=2)

We used murine macrophages to study phagocytosis of both P. aeruginosa and EIgMCs, because thioglycollate-elicited peritoneal macrophages express CR3 in an nonactivated state [21]. Resident peritoneal macrophages express CR3 that is activated [21]. Thus, we hypothesized that thioglycollate-elicited macrophages would provide data similar to studies with LAD phagocytes. In the presence of glucose, EIgMCs (ligated by CR3) were phagocytosed by murine resident peritoneal macrophages, but fewer were ingested by thioglycollate-elicited macrophages (table 3). By comparison, P. aeruginosa P1 and 808 were ingested at the same levels by both thioglycollate-elicited and resident peritoneal macrophages. Phagocytosis of EIgMCs by thioglycollate-elicited peritoneal macrophages was enhanced if the particles were pelleted by centrifugation and were resuspended in strain P1 supernatant from an overnight culture before being added to the phagocytosis assay. In such circumstances, EIgMC phagocytosis was not significantly different between thioglycollate-elicited and resident peritoneal macrophages. The supernatant from an overnight culture of strain 808 did not enhance phagocytosis of EIgMCs by thioglycollate-elicited peritoneal macrophages

Table 3

Murine macrophage phagocytosis assessed in the presence and absence of glucose

In the absence of glucose, phagocytosis of P. aeruginosa P1 was reduced (∼55%) with thioglycollate-elicited, but not resident, peritoneal macrophages. Similar conditions had no effect on phagocytosis of strain 808. We characterized 99 strains for their ability to be phagocytosed in the absence of glucose, and 50% exhibited the same depressed ingestion observed with strain P1 (authors' unpublished data)

Discussion

The occurrence of leukocyte adhesion deficiency provides a rare opportunity to study the role of CR3 in nonopsonic phagocytosis. By using both neutrophils and macrophages from this human CR3 knockout, we have demonstrated that CR3 appears to be the exclusive receptor for nonopsonic phagocytosis for some strains of P. aeruginosa (CR3 strains). Conversely, other strains of the same species were ingested through interaction with CD14 (CD14 strains; table 1). CR3 is the major receptor for opsonic phagocytosis of many bacteria [22], and previous studies also have shown nonopsonic ingestion of microbial pathogens (e.g., Listeria monocytogenes [23], Cryptococcus neoformans [24], Mycobacterium tuberculosis [21, 25], and Leishmania [26]) through CR3-mediated phagocytosis. We proposed elsewhere that this may be a critical receptor for nonopsonic phagocytosis in the human lung, in which there are low levels of opsonins [1]

CD14 is a 55-kDa gpI-anchored glycoprotein on neutrophils, monocytes, and mature macrophages, which also is present in plasma as a soluble protein (sCD14) [27]. CD14 can bind lipopolysaccharide (LPS) molecules [28, 29], and binding is facilitated by LPS-binding protein (LBP) [30], which shares homology with lipid transfer proteins [31, 32]. By virtue of its ability to bind LPS, CD14 has been implicated in phagocytosis of bacteria, including Escherichia coli [33], M. tuberculosis [34], and Salmonella typhimurium [35]. CD14 also promotes cytokine release after stimulation by a variety of ligands: lipoarabinamannan from M. tuberculosis [34], manuronic acid polymers from P. aeruginosa [36], soluble peptidoglycan fragments from Staphylococcus aureus [37], rhamnose-glucose polymers from Streptococcus mutans [38], and insoluble cell walls from several gram-positive species [39]. CD14 is unable to transmit signals directly to the phagocyte cytoplasm [40], yet LPS and bacterial ligation by CD14 transmit a signal that leads to internalization and cytokine production [33]. A mutant toll-like receptor 4 (TLR4) gene, which negates LPS sensitivity, is the likely candidate for CD14 signal transduction after LPS ligation [41, 42]. TLR4 may have an important signaling role in the phagocytosis of these CD14-internalized P. aeruginosa strains

Of the 10 different P. aeruginosa isolates that we studied, 5 exhibited CR3-mediated phagocytosis, and 5 were ingested after interaction with CD14. Of note, there was no obvious phenotypic segregation of these 2 groups of strains (other than receptor bias) such that clinical, environmental, and CF isolates appeared in either CR3 or CD14 groups

CR3 has ⩾3 distinct binding sites that can adopt various states of functionality: a C3bi site that binds complement-coated particles, a lectin-like site that binds glucan, and a site distinct from the C3bi site that binds LPS [43–45]. Thioglycollate-elicited peritoneal macrophages display CR3 on their surface but are not functional until a secondary signal triggers ingestion [21]. The secondary signal can be the phorbol ester, such as phorbol myristate acetate, or proinflammatory cytokines, such as tumor necrosis factor–α [21, 46]. Therefore, thioglycollate-elicited peritoneal macrophages ingest EIgMC poorly in the absence of a secondary signal (table 3). We found that supernatant from the CR3-ingested strain P1 could provide the secondary signal for phagocytosis of EIgMC and, presumably, for the bacterium itself (table 3). The factor responsible for this phenomenon apparently was not present in supernatant from the CD14-ingested strain, 808. Since the bacterial factor that induces CR3 activation in this model is found in culture supernatant from CR3-ingested strains, it is possible that this putative second signaling molecule is a bacterial component, possibly LPS. The murine resident peritoneal macrophage population displays activated CR3 and does not require a secondary signal for phagocytosis [21]; therefore, resident peritoneal macrophages did not display differential stimulation by the 2 bacterial supernatants. This further underlines the observed phenotypic differences that lead to internalization of P. aeruginosa by phagocytes; strains internalized by CR3 are capable of activating that portal of entry, whereas strains internalized by CD14 are not

We reported elsewhere that the absence of glucose in the phagocytosis buffer abrogates ingestion of some P. aeruginosa strains by thioglycollate-elicited peritoneal macrophages and murine alveolar macrophages [18, 47]. This “glucose effect” was seen in half of 99 different P. aeruginosa strains (authors' unpublished data). Of interest, the P. aeruginosa strains studied here that exhibited this glucose effect are those that are phagocytosed via CR3. The same effect was also observed in the phagocytosis of EIgMC suspended in P1 culture broth. These observations suggest that glucose is a requirement for CR3-mediated phagocytosis of P. aeruginosa perhaps as an energy source [48] or to regulate kinase pathways [49] or both [50]. The glucose effect attenuates over time. Freshly explanted thioglycollate-elicited peritoneal macrophages will not ingest any P. aeruginosa P1 strain in the absence of glucose [18]; however, if incubated overnight in tissue culture media (as done in the present experiments), modest levels of phagocytosis will be observed (see table 3). Correspondingly, the activation of CR3 on thioglycollate-elicited macrophages over time in culture media has been reported elsewhere [21]

In this study, we demonstrated that either CD14 or CR3 is involved in the phagocytosis of different strains of P. aeruginosa in an exclusive manner. These findings suggest that ⩾2 host phagocytic mechanisms have evolved to overcome phenotypic differences in gram-negative bacteria such as P. aeruginosa. Conversely, the bacteria may have developed phenotypic differences that allow for the adaptation to a changing host environment. The phagocyte-bacterial interactions that lead to ingestion of the organism are probably more complex in vivo. A landscape on the surface of phagocytes is emerging, in which distinct receptors are involved in close contact and regulation by common kinase pathways [51, 52]. There is evidence that physical association of CR3 with gpI-anchored receptors (e.g., CD14 and CD16 [FcγRII]) occurs [53]. LPS from gram-negative bacteria can colocalize CD14 with CR3, and this common ligand may serve to join both receptors, such that CD14 is able to transmit cellular signals via CR3 [54]. CD14 also transmits signals to the cell cytoplasm at ligation via TLR4 [41]. Our findings suggest that phenotypic heterogeneity in P. aeruginosa strains leads to different mechanisms of nonopsonic phagocytosis. For patients with CF, in whom pulmonary infection with P. aeruginosa is usual, this finding may have important implications for long-term pulmonary function

Acknowledgment

We thank Susan Ursuliak (Department of Pathology, University of British Columbia, Vancouver, Canada) for technical assistance

Informed consent was obtained from the patients. The study was approved by the University of British Columbia Ethical Review Board

Financial support: Medical Research Council of Canada (grant to D.P.S.); Canadian Bacterial Diseases Network (grant to D.P.S.; funding to J.-P.H.); Pediatric Infectious Disease Society via educational grant from Pfizer (fellowship to A.J.P.)