Interactions of Penicillium marneffei with Human Leukocytes In Vitro

Abstract

Penicillium marneffei, a dimorphic fungus endemic in parts of Asia, causes disease in those with impaired cell-mediated immunity, especially persons with AIDS. The histopathology of penicilliosis marneffei features the intracellular infection of macrophages. We studied the interactions between human leukocytes and heat-killed yeast-phase P. marneffei. Monocyte-derived macrophages bound and internalized P. marneffei in the presence of complement-sufficient pooled human serum (PHS). Binding and phagocytosis were still seen if PHS was heat inactivated or omitted altogether. The binding of unopsonized P. marneffei to monocyte-derived macrophages occurred in the absence of divalent cations and was not affected by inhibitors of mannose and β-glucan receptors or monoclonal antibodies directed against CD14 and CD11/CD18. Binding was profoundly inhibited by wheat germ agglutinin. A vigorous respiratory burst was seen in peripheral blood mononuclear cells (PBMC) stimulated with P. marneffei, regardless of whether the fungi were opsonized. However, tumor necrosis factor alpha (TNF-α) release from PBMC stimulated with P. marneffei occurred only if serum was present. These data demonstrate that (i) monocyte-derived macrophages bind and phagocytose P. marneffei even in the absence of opsonization, (ii) binding is divalent cation independent but is inhibited by wheat germ agglutinin, suggesting that the major receptor(s) recognizing P. marneffei is a glycoprotein with exposed N-acetyl-β-d-glucosaminyl groups, (iii) P. marneffei stimulates the respiratory burst regardless of whether opsonins are present, and (iv) serum factors are required for P. marneffei to stimulate TNF-α release. The ability of unopsonized P. marneffei to parasitize mononuclear phagocytes without stimulating the production of TNF-α may be critical for the virulence of this intracellular parasite.

Penicillium marneffei is a dimorphic fungal pathogen endemic in southeast Asia, south China, and Hong Kong. Prior to the AIDS epidemic, cases of penicilliosis marneffei were limited to scattered case reports. However, with the spread of human immunodeficiency virus (HIV) into regions where P. marneffei is endemic, the number of cases of penicilliosis marneffei has increased greatly. For example, over a 7-year period, 1,115 cases of penicilliosis marneffei were seen in human immunodeficiency virus-infected patients at the Chiang Mai University Hospital in northern Thailand (44). There, penicilliosis marneffei is the third most common opportunistic infection after tuberculosis and cryptococcosis (45). Fever, anemia, weight loss, lymphadenopathy, hepatomegaly, and skin lesions are relatively common clinical manifestations, and mortality is substantial if the infection is not treated in a timely fashion (45, 46). As with many other AIDS-related opportunistic infections, even if initial therapy is successful, lifetime prophylaxis of penicilliosis marneffei is required to prevent relapse (46).

Infection with P. marneffei is presumed to originate in the lung following the inhalation of airborne conidia. The recognition of laminin by P. marneffei conidia may facilitate attachment to the bronchoalveolar epithelium (12). In susceptible hosts, the fungus undergoes phase transition and reproduces as yeast cells. The histopathology of penicilliosis marneffei is characterized by the intracellular infection of macrophages (9). Yeast cells of P. marneffei measure approximately 2 to 3 by 2 to 7 μm and reproduce inside the macrophage by schizogony (9). In AIDS patients, the fungal burden usually is high, and intracellular P. marneffei organisms are readily seen in infected tissue specimens. In murine models of penicilliosis marneffei, T cells are critical for protection. Athymic mice are hypersusceptible to infection, and partial protection can be adoptively transferred via nylon wool nonadherent splenocytes (21). In immunocompetent mice inoculated intranasally with P. marneffei, an exuberant CD4⁺ T-cell infiltration into the lungs is observed (22). Presumably then, analogous to many other intracellular infections (23), the activation of macrophages by T cell-derived cytokines is necessary to control penicilliosis marneffei. In the present study, the interactions of human peripheral blood mononuclear cells (PBMC) and monocyte-derived macrophages (MDM) with P. marneffei were examined.

MATERIALS AND METHODS

Materials.

Unless otherwise noted, reagents were purchased from Sigma Chemical Co. (St. Louis, Mo.). Monoclonal antibodies OKM10 and IB4 were gifts from Patricia Rao (Ortho Pharmaceutical Corp., Raritan, N.J.) and Samuel Wright (Merck Pharmaceutical, Rahway, N.J.), respectively. Monoclonal antibody MY4 was purchased from Coulter Immunology (Hialeah, Fla.). OKM10 reacts with the complement binding site of the α chain of CD11b/CD18, IB4 reacts with the common β chain of the CD11/CD18 complex, and MY4 reacts with the lipopolysaccharide (LPS) receptor CD14 (50, 51). RPMI 1640 and phosphate-buffered saline (PBS) were obtained from BioWhittaker, Inc. (Walkersville, Md.). Pooled human serum (PHS) was obtained by combining sera from more than 10 healthy donors under ice-cold conditions and storing it in aliquots at −70°C to preserve complement activity. Supplemented PBS was PBS containing 1 g of glucose per liter, 100 mg of CaCl₂ per liter, and 100 mg of MgCl₂ per liter. All experiments were performed under conditions carefully designed to minimize endotoxin contamination as described (28). Except where otherwise indicated, incubations were performed at 37°C in humidified air supplemented with 5% CO₂.

Fungi.

P. marneffei 30 was obtained from William G. Merz (Johns Hopkins Medical Institute) and cultured in brain heart infusion broth for 7 to 21 days at 37°C to obtain yeast-phase organisms. A previously described strain of Candida albicans (29) was grown in the yeast phase for 48 h on Sabouraud dextrose agar. Fungi were heat killed (60°C for 30 min) and washed six times in PBS. P. marneffei organisms were surface labeled with rhodamine isothiocyanate (RITC) by incubation in a 1-mg/ml solution of RITC dissolved in 50 mM borate buffer, pH 8.0, for 18 h at 4°C (27). Fungi were washed extensively in PBS prior to addition to phagocytes.

Isolation of PBMC and MDM.

Peripheral blood was obtained by venipuncture from healthy adult volunteers. For each set of experiments, the same donor was not used more than once. Blood was anticoagulated with pyrogen-free heparin (Elkins-Sinn Inc., Cherry Hill, N.J.) and centrifuged at 500 × g for 15 min. The buffy coat was collected, and PBMC were purified by centrifugation on a Ficoll-Hypaque density gradient as in previous studies (25, 26). MDM were obtained by incubating PBMC (2 × 10⁵/well) in 96-well flat-bottom polystyrene tissue culture plates with RPMI 1640 containing 10% male AB serum. Plates were incubated at 37°C with 5% CO₂ in a humidified chamber, washed after 24 h to remove nonadherent cells, and cultured for an additional 5 to 8 days.

Binding and internalization assays.

Wells containing MDM were washed three times with RPMI 1640 warmed to 37°C. P. marneffei yeast cells (5 × 10⁴/well) were added and left in the wells for 60 min. To facilitate the microscopic identification of fungi, RITC-labeled P. marneffei organisms were utilized (27). In preliminary experiments, labeling did not affect binding. At the end of incubation, unattached fungi were washed free and the cells were fixed with 1% formaldehyde buffered in PBS. Wells were read with an inverted microscope under epifluorescence. The binding index was calculated as the number of cell-associated yeasts per 100 MDM. When putative inhibitors and monoclonal antibodies were utilized, they were added for 30 min prior to the addition of P. marneffei organisms. The percent inhibition of binding was calculated by the following equation: [1 − (binding index of studied group/binding index of control)] × 100.

The assays described above measured all cell-associated fungi, regardless of whether they were surface attached or internalized (phagocytosed). To measure internalization, the Uvitex assay was performed as previously described (14, 27). MDM were incubated with RITC-labeled P. marneffei organisms for 60 min, washed free of unattached fungi, and then fixed with 1% formaldehyde buffered in PBS. Cells were then stained with 0.1% Uvitex (Fungiqual A; Specialty Chemicals for Medical Diagnostics, Kandern, Germany) for 1 min followed by washing in PBS. Cell-associated fungi that are stained with Uvitex under these conditions are bound but not internalized. Internalized fungi are stained with RITC but not Uvitex. Data are presented as percentages of the cell-associated fungi that are internalized.

Hydrogen peroxide generation and TNF-α release.

Levels of H₂O₂ production by stimulated PBMC were measured by the H₂O₂-dependent, horseradish peroxidase-mediated oxidation of homovanillic acid to the fluorescent dimer 2,2′-dihydroxy-3,3′-dimethoxydiphenyl-5,5′-diacetic acid (14, 41). PBMC (2 × 10⁶/well) were added to 24-well flat-bottom plates and left unstimulated or challenged with P. marneffei organisms and zymosan (10⁷ particles/well) for 2 h in supplemented PBS containing homovanillic acid and horseradish peroxidase. The fluorescence of the supernatants was compared with standards containing known amounts of H₂O₂. Tumor necrosis factor alpha (TNF-α) release was measured by enzyme-linked immunosorbent assay as previously described (28, 29). For both the H₂O₂ and TNF-α assays, PBMC were not washed free of nonadherent cells prior to stimulation.

Statistics.

Means and standard errors of the mean (SEM) were compared by using the two-tailed unpaired t test (SigmaStat Statistical Software; Jandel Corporation, San Rafael, Calif.). The Bonferroni correction was used to adjust for multiple comparisons. Statistical significance was considered to be achieved only if the P value multiplied by the number of comparisons was less than 0.05.

RESULTS

Binding and internalization.

MDM were challenged with P. marneffei organisms in the presence of PHS, heat-inactivated PHS (HI-PHS), or no serum, and binding levels were assayed (Fig. 1). Regardless of opsonic conditions, MDM bound P. marneffei. However, optimal binding required complement-sufficient serum. Thus, in the presence of PHS approximately two and three times as many organisms bound to MDM as with HI-PHS and no serum, respectively. However, once bound, approximately 80% of the fungi subsequently were internalized, regardless of opsonic conditions.

FIG. 1.

Opsonic requirements for binding and internalization of P. marneffei to MDM. Binding and internalization assays were performed as described in Materials and Methods in the presence of 10% PHS, 10% HI-PHS, and no serum (comparing binding indices with PHS versus no PHS [P = 0.004]). Error bars indicate SEM.

Receptors involved in MDM binding of unopsonized P. marneffei organisms.

Having established that MDM bound unopsonized P. marneffei organisms, we next sought to determine the responsible receptor(s). Agents serving as competitive ligands of β-glucan, mannose, and β-integrin receptors had no effect on binding (Table 1). In addition, binding was not affected significantly by the presence of EDTA, indicating that divalent cations were not necessary. Dextran sulfate, which binds to macrophage scavenger receptors (38), significantly inhibited binding by 32.9%. However, two other agents that serve as ligands for scavenger receptors, heparin and fucoidin, did not significantly inhibit binding once the Bonferroni correction was made.

TABLE 1.

Effects of receptor-blocking agents on binding of P. marneffei to MDM

Agent	Concn (μg/ml)	Receptor(s)	% Inhibition of bindinga	Pb
Laminarin	1,000	β-Glucan	−8.2 ± 45.2	NS
Mannan	1,000	Mannose	−15.4 ± 45.1	NS
Ovalbumin	1,000	Mannose	−23.6	NS
EDTA	1,000	Divalent cation dependent	17.1 ± 9.1	NS
Fibronectin	100	β-Integrin	7.4 ± 12.3	NS
Dextran	10	Scavenger	32.9 ± 5.4	0.001
Heparin	10	Scavenger	9.2 ± 6.3	NS
Fucoidin	10	Scavenger	19.8 ± 4.5	0.02
OKM10	10	CD11b/CD18	25.4 ± 27.0	NS
IB4	10	CD18 complex	−5.2 ± 25.8	NS
MY4	10	CD14	19.8 ± 19.7	NS

^aMDM were incubated with P. marneffei organisms in the absence of serum, and the percent inhibition of binding was calculated as described in Materials and Methods. OKM10, IB4, and MY4 are monoclonal antibodies directed at the indicated receptors. Data are means ± SEM for three to five experiments, each performed in triplicate. 

^bNS, not significant. 

Monoclonal antibodies directed at the complement binding site of complement receptor 3 (CD11b/CD18), the β chain of the CD18 complex, and the LPS receptor CD14 failed to significantly inhibit binding (Table 1). These three receptors are involved in the binding of other unopsonized fungi (36, 37).

The lectin wheat germ agglutinin (WGA) profoundly inhibited P. marneffei binding to MDM in a dose-dependent fashion, with nearly 90% inhibition seen at the highest concentration used (Fig. 2). The inhibition of binding was reversed by the addition of N-acetylglucosamine (NAG), a preferred ligand for WGA. NAG by itself had no effect on binding (inhibition, 0.0% ± 13.4%). The lectin concanavalin A, used at a concentration of 100 μg/ml, also did not inhibit binding (inhibition, −35.8% ± 9.6%).

FIG. 2.

Effects of WGA on MDM binding to unopsonized P. marneffei. MDM were incubated with WGA, and binding levels were determined as described in Materials and Methods. WGA demonstrated a significant (P < 0.001) inhibition of binding at concentrations of 10 μg/ml and greater. The inhibitory effect of 100 μg of WGA per ml was competed by 100 μg of NAG per ml. Data are means ± SEM (error bars) for three experiments each performed in triplicate.

The above-mentioned experiments demonstrated that WGA inhibited P. marneffei binding to MDM but did not establish whether WGA was acting primarily on the MDM or the fungi. To make that distinction, MDM and P. marneffei organisms were preincubated for 30 min with 100 μg of WGA per ml, washed three times, and then tested in a binding assay in the absence of serum. Compared with no WGA, the preincubation of MDM with WGA resulted in a 74.1% ± 2.3% inhibition of binding (mean ± SEM for three experiments each performed in triplicate; P < 0.001). In contrast, when P. marneffei organisms were preincubated with WGA, the inhibition of binding was 29.8% ± 6.5% (P value not significant).

Stimulation of the respiratory burst.

The ability of P. marneffei to stimulate the respiratory burst of PBMC, measured by H₂O₂ generation, was examined next (Fig. 3). P. marneffei vigorously stimulated PBMC H₂O₂ production, with about equal quantities seen in the presence of PHS and HI-PHS. In the absence of serum, stimulated H₂O₂ production was reduced modestly (by approximately 25%) but significantly (P < 0.001). The amounts of H₂O₂ generated were similar for PBMC stimulated by P. marneffei and by zymosan particles, the latter being a potent stimulus of monocyte H₂O₂ release (41).

FIG. 3.

Opsonic requirements for respiratory burst of PBMC stimulated with P. marneffei organisms. P. marneffei organisms and zymosan were preopsonized with PHS, HI-PHS, or no serum and added to PBMC for 2 h. H₂O₂ was measured as described in Materials and Methods. In comparing unopsonized P. marneffei organisms and zymosan with particles preopsonized in PHS or HI-PHS, P < 0.001. Data are means ± SEM (error bars) for four triplicate experiments. Unstimulated PBMC released 0.28 ± 0.06 nmol of H₂O₂.

Stimulation of TNF-α release.

In the final set of experiments, the level of fungal stimulation of TNF-α production was determined (Fig. 4). P. marneffei stimulated TNF-α release in the presence of PHS and HI-PHS at all four ratios of P. marneffei organisms to PBMC tested. Levels were comparable to that seen with 10 ng of LPS per ml, although less than that seen with heat-killed C. albicans opsonized with PHS. When serum was omitted from the system, P. marneffei and LPS failed to stimulate significant amounts of TNF-α. In contrast, C. albicans stimulated similar amounts of TNF-α in the presence of HI-PHS compared with no serum.

FIG. 4.

Opsonic requirements for TNF-α release from PBMC stimulated with P. marneffei (Pm). PBMC (2 × 10⁵/well) were incubated with P. marneffei at the indicated fungus-to-PBMC ratio in the presence of PHS, HI-PHS, and no serum at 37°C for 18 h. Supernatants were collected and assayed for TNF-α by enzyme-linked immunosorbent assay. LPS (10 ng/ml) and C. albicans (2:1 ratio of fungus:PBMC) served as positive controls. Data represent means ± SEM (error bars) for four triplicate experiments. Unstimulated PBMC generated a mean of 114 pg of TNF-α per ml, which was not significantly different from that seen with LPS and P. marneffei in the absence of serum.

DISCUSSION

Whether a microorganism is able to establish itself as a pathogen depends to a large extent upon which receptors on the host cells bind to it and upon the consequences of that binding (27). Only a paucity of immunological research has been performed on P. marneffei, and studies examining the mechanisms by which this medically important fungus gains access to macrophages have not previously been reported. The data presented here establish that P. marneffei binds to and is phagocytosed by human mononuclear phagocytes. While binding occurred in the absence of opsonins, it was enhanced approximately threefold by the presence of complement-sufficient PHS. In contrast, binding was not enhanced significantly if the PHS was heat inactivated under conditions which prevent activation of the complement system (19). Thus, P. marneffei can gain access to macrophages even in the absence of serum, but binding is enhanced by the presence of a serum factor(s), most likely complement.

P. marneffei reacts with a monoclonal antibody specific for galactomannan; however, binding occurs on the plasma membrane and not the outer surface of the fungal cell wall (39). In contrast, a mannoprotein, designated MP1, reportedly is exposed on the cell wall of P. marneffei (6). Nevertheless, mannose receptors do not appear to be responsible for the binding interaction of unopsonized P. marneffei organisms to MDM, as yeast mannans, the mannosylated protein ovalbumin, and the mannose-binding lectin concanavalin A all failed to inhibit binding. Similarly, blocking β-glucan receptors had no significant effect on P. marneffei binding to MDM. Mannose and β-glucan receptors have been implicated in the interactions of C. albicans and Aspergillus fumigatus with macrophages (7, 16, 17, 31).

Monoclonal antibodies directed at the CD11/CD18 β2-integrin complex and CD14 also had no effect on the binding of unopsonized P. marneffei to MDM. CD11/CD18 is responsible for macrophage binding of unopsonized Histoplasma capsulatum, whereas both CD11b/CD18 and CD14 have been implicated in macrophage binding of Blastomyces dermatitidis (36, 37). CD11b/CD18 also is a receptor for complement and β-glucan, whereas CD14 recognizes LPS and other microbial components (27, 40, 47). Lack of inhibition of binding by antireceptor monoclonal antibodies does not rule out a role for the receptor in binding, as the antibody could be recognizing an epitope distinct from the fungal binding site. Further evidence against a role for CD11/CD18 and other members of the β-integrin family is the lack of inhibition by fibronectin and EDTA (10, 11, 42). The β-integrin receptors are divalent cation dependent.

Macrophage scavenger receptors bind a broad array of polyanionic ligands including surface components of gram-negative and gram-positive bacteria as well as the intact bacteria themselves (13, 20, 38). The anionic polysaccharides dextran sulfate and fucoidin as well as heparin inhibit at least some of the scavenger receptors (38). We observed a modest, but significant, inhibition of P. marneffei binding to MDM by using dextran sulfate and a trend towards significant inhibition with fucoidin. Thus, scavenger receptors might participate in macrophage binding of P. marneffei; however, these receptors do not appear to be the predominant ones responsible for binding.

WGA is a lectin with affinity for N-acetyl-β-d-glucosaminyl residues and N-acetyl-β-d-glucosamine oligomers (34). Our finding that WGA treatment of MDM profoundly inhibited the binding of unopsonized P. marneffei organisms strongly suggests that the major receptor(s) recognizing this fungus is a glycoprotein with exposed N-acetyl-β-d-glucosaminyl groups. While many cell surface glycoproteins on macrophages bind WGA (8, 43), the functions of only a few are defined (2). Treatment of macrophages with WGA has also been reported to inhibit the binding of other particulate stimuli including iron beads (18) and Leishmania mexicana mexicana promastigotes (3). Like many other lectins, WGA stimulates macrophages (8). However, the inhibition of fungal binding observed with WGA is unlikely to be due to a nonspecific activation of the cells because another lectin that stimulates macrophages, concanavalin A, did not inhibit binding.

Following the binding of P. marneffei to MDM, approximately 80% of the yeast cells were phagocytosed within 1 h. Phagocytosis proceeded regardless of the presence of opsonins. In this regard, the MDM response to P. marneffei is similar to that seen with H. capsulatum (36) but is different from the situation with encapsulated Cryptococcus neoformans, where most bound organisms are not readily phagocytosed even if opsonized with normal human serum (27).

A major mechanism by which phagocytes exhibit antimicrobial activity is oxidative killing, a process whereby molecular oxygen is reduced to form reactive oxygen products (33). Some intracellular pathogens have evolved mechanisms to escape killing by macrophages either by inhibiting production of oxygen metabolites or by neutralizing the metabolites that are produced (32, 33, 49). We found that P. marneffei potently stimulated H₂O₂ production by PBMC and that such stimulation was seen even in the absence of opsonization. Future studies are needed to determine whether the amounts of oxidants generated are sufficient to kill P. marneffei. Compared with bacteria, many fungi are relatively resistant to oxidants generated by phagocytes (24).

In addition to their role as effector cells capable of inhibiting and killing microorganisms, mononuclear phagocytes secrete a variety of bioactive substances that serve to modulate the inflammatory response (35). One such substance is TNF-α, a proinflammatory cytokine that has a broad spectrum of immunoregulatory, metabolic, and inflammatory activities (28, 48). TNF-α promotes inflammation, possibly through the induction of cell adhesion molecules, neutrophil and macrophage chemotactic factors, and acute phase proteins and the generation of other proinflammatory cytokines such as interleukin 1 and interleukin 6. We found that P. marneffei stimulated human PBMC to secrete TNF-α, in amounts comparable to that stimulated by 10 ng of LPS per ml, provided that serum was present. Thus, despite stimulating a vigorous respiratory burst, unopsonized P. marneffei organisms failed to stimulate TNF-α secretion. In animal models of fungal infection, TNF-α is critical for host defenses (1, 15, 30). Thus, it is tempting to speculate that the ability of unopsonized P. marneffei organisms to parasitize mononuclear phagocytes without stimulating the production of TNF-α may enhance the virulence of this intracellular parasite. Moreover, while incubation with HI-PHS did not significantly increase the binding of P. marneffei to MDM, it did dramatically increase TNF-α secretion in PBMC. This suggests that heat-stable serum factors (such as antibody or mannose-binding lectin) on the surface of P. marneffei are interacting with their corresponding leukocyte receptors to stimulate TNF-α production.

In many respects, macrophage interactions with P. marneffei are quite similar to that seen with H. capsulatum. Yeast cells from both species are bound and phagocytosed in the absence of opsonins and both stimulate a respiratory burst (5, 36). Similarly, the pathophysiology and clinical presentation of histoplasmosis and penicilliosis marneffei share overlapping features including the parasitism of macrophages (4, 9). However, clearly there are differences, most notably in the mechanisms by which the two fungi gain access to the macrophage. H. capsulatum entry into macrophages is divalent cation dependent and uses the CD18 complex (5, 36), whereas P. marneffei is divalent cation independent and gains access through a WGA-inhibitable process.

The data presented herein begin to define the events leading to the parasitism of macrophages by P. marneffei. However, many unanswered questions remain, including the definition of the specific receptor(s) responsible for binding and the conditions necessary to activate the macrophage to inhibit and kill P. marneffei. As the human immunodeficiency virus epidemic continues to spread in regions of Asia where P. marneffei is endemic, the incidence of penicilliosis marneffei is likely to increase. Insights into the mechanisms whereby the immune system responds to infection by P. marneffei could lead to novel approaches to prevention and therapy of this medically important fungus. Moreover, the study of the interactions between macrophages and P. marneffei could serve to increase our general understanding of the mechanisms of intracellular parasitism.