Pneumocystis carinii Cell Wall β-Glucans Initiate Macrophage Inflammatory Responses through NF-κB Activation

Abstract

β-Glucans are major structural components of fungi. We have recently reported that the pathogenic fungus Pneumocystis carinii assembles a β-glucan-rich cell wall that potently activates alveolar macrophages to release pro-inflammatory cytokines and chemokines. Purified P. carinii β-glucans predictably induce both cytokine generation and associated neutrophilic lung inflammation. Herein, we demonstrate that P. carinii β-glucan-induced macrophage stimulation results from activation of NF-κB. Although analogous to macrophage activation induced by bacterial lipopolysaccharide (LPS), P. carinii β-glucan-induced macrophage NF-κB activation exhibits distinctly different kinetics, with slower induction and longer duration compared with LPS stimulation. Macrophage activation in response to P. carinii β-glucan was also substantially inhibited with the NF-κB antagonist pyrrolidine dithiocarbamate. In addition to different kinetics of NF-κB activation, P. carinii β-glucan and LPS also utilize different receptor systems to induce macrophage activation. Macrophages from Toll-like receptor 4-deficient and wild type mice produced equivalent amounts of tumor necrosis factor α when stimulated with P. carinii β-glucan. However, Toll-like receptor 4-deficient macrophages were refractory to stimulation with LPS. In contrast, MyD88-deficient macrophages exhibited a significant (though partial) blunted response to P. carinii β-glucan. These data demonstrate that P. carinii β-glucan acts as potent inducer of macrophage activation through NF-κB utilizing cellular receptors and signaling pathways distinct from LPS.

Pneumocystis carinii remains a common cause of pneumonia among individuals with compromised immune function, resulting in substantial morbidity and mortality in patients with AIDS or malignancies and following organ transplantation (1, 2). Similar to other related fungi, P. carinii assembles a β-glucan-rich cell wall comprised of glucose residues arranged in β-(1,3)-d-glucopyranosyl polymers with β-(1,6)-d-glucopyranosyl side chains of varying length and frequency distributions (3–5). We recently reported the isolation and characterization of a particulate carbohydrate cell wall fraction derived from P. carinii, which is composed largely of β-glucans (6, 7). In addition to providing structural support to the organism, P. carinii cell wall β-glucans potently induce inflammatory activation of lung cells, with release of pro-inflammatory cytokines and chemokines during this infection (6, 7).

Although essential for host responses and clearance of P. carinii, alveolar macrophage-induced lung inflammation contributes substantially to respiratory failure and death during P. carinii pneumonia (8). Indeed, exuberant host inflammatory responses to this organism stimulate alveolar edema and neutrophilic infiltration, resulting in acute lung injury and respiratory failure. Pulmonary inflammation exerts a greater effect on clinical outcome than direct actions of the organism (9). This concept is supported by the clinical observation that severe P. carinii pneumonia remains the only form of acute lung injury that has clearly been demonstrated to improve with agents that nonspecifically down-regulate host inflammatory responses, namely corticosteroid therapy (10). Thus, understanding the mechanisms of host inflammatory cell activation by P. carinii is of critical importance and may provide essential insights into means by which the inflammatory response may be selectively modulated to advantage the host.

Particulate β-glucans from fungal organisms interact with cognate receptors on macrophages stimulating the release of reactive oxidants, eicosanoid metabolites, cytokines, and chemokines (11). In an analogous fashion, whole P. carinii organisms and β-glucan isolates purified from P. carinii both strongly promote inflammatory activation of alveolar macrophages with subsequent release of tumor necrosis factor α (TNFα)¹ and macrophage inflammatory protein-2 (6, 12). P. carinii-initiated macrophage stimulation is mediated through ligation of glucan receptors and is abolished by digestion with β-glucanase, indicating that glucan components of the organism are chiefly responsible for macrophage inflammatory responses during this infection (6, 12). Although these investigations provide important insight into the initiation of P. carinii-induced lung inflammation, signal transduction mechanisms mediating macrophage activation during P. carinii challenge have not yet been defined.

NF-κB is a ubiquitous transcription factor that regulates inflammatory gene expression in many immune effector cells. NF-κB activation is induced by a variety of extracellular signals such as pro-inflammatory cytokines, bacterial lipopolysaccharide, and viral products (13–16). In quiescent unstimulated cells, NF-κB is composed of homodimers or heterodimers containing various Rel proteins, including NF-κB1 (p50), NF-κB2 (p52), Rel A (p65), Rel B, and c-Rel, complexed to an inhibitory anchor protein termed IκBα (17). Following stimulation with an appropriate agent, IκBα is phosphorylated and rapidly degraded through proteosomal mechanisms (17). Removal of IκBα permits translocation of active NF-κB components to the nucleus where they initiate cytokine gene transcription (17, 18).

In addition, receptor mechanisms participating in innate macrophage immune responses to β-glucans are not yet characterized. The Toll family of immune receptors represents a conserved family of pattern recognition receptors that respond to microbial products and other agonists to trigger host inflammatory responses (19). For instance, Toll-like receptor 4 (TLR-4) is a well defined mechanism conferring host cell response to bacterial LPS (20). Furthermore, many, although not all, Toll family responses also require activity of the MyD88 adaptor protein to initiate cell signaling (21, 22). Cellular activation through such innate immune receptors has not been extensively studied during challenge of cells with fungal cell wall β-glucans.

The branching structures and biological activities of fungal β-glucans vary considerably with the sources from which they are derived. Zymosan, a commonly studied β-glucan source prepared from the nonpathogenic fungus Saccharomyces cerevisiae and other soluble nonpathogenic glucans activate nuclear factor NF-κB in cell lines (23–26). Whether NF-κB participates in signaling macrophage inflammatory activation to release TNFα following stimulation with cell wall β-glucans derived from P. carinii and whether such signaling can be modified pharmacologically are currently unknown. The present investigation was therefore performed to determine the role of NF-κB in inflammatory activation of macrophages challenged P. carinii β-glucans. We further tested the roles of TLR-4 and MyD88 in signaling macrophage responses to P. carinii β-glucans and whether macrophage activation could be modified with pyrrolidine dithiocarbamate (PDTC), a potent inhibitor of NF-κB.

MATERIALS AND METHODS

Reagents—Endotoxin-free buffers and reagents were scrupulously used in all experiments. LPS from Escherichia coli 026:B6, PDTC, and other general reagents were obtained from Sigma, unless otherwise specified. P. carinii was derived originally from the American Type Culture Collection stock (Manassas, VA) and has been passaged in our immunosuppressed rat colony (8). Murine RAW 264.7 macrophages were also purchased from American Type Culture Collection and routinely cultured in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum and 2 mm l-glutamine, penicillin 10,000 units/liter, and streptomycin 1 mg/liter.

Generation of a P. carinii β-Glucan-rich Cell Wall Isolate—All animal experimentation was reviewed and approved by the Mayo Institutional Animal Care and Usage Committee. A glucan-rich cell wall fraction from P. carinii was prepared as recently described (6, 7). P. carinii pneumonia was induced in dexamethasone-treated immunosuppressed Harlan Sprague-Dawley rats (Harlan, Inc., Indianapolis, IN) (8). P. carinii organisms were isolated from lungs of heavily infected animals by homogenization and filtration through 10-μm filters. The organisms were autoclaved (120 °C, 20 min) and disrupted by ultrasonication (200 W for 3 min, six times), and the glucans were isolated by NaOH digestion and lipid extraction as previously detailed (6, 7). The final product contained predominantly carbohydrate (95.7%) and released 82% of its content as d-glucose following hydrolysis (6). Extensive measures were employed to ensure that the fraction was free of endotoxin. Prior to use in culture, the P. carinii cell wall fractions were washed with 0.1% SDS and then vigorously washed with distilled physiological saline to remove the detergent. The final preparation was assayed for endotoxin with the Limulus amebocyte lysate assay method and found to consistently contain <0.125 units of endotoxin (6).

Preparation of Nuclear and Cytosolic Lysate Fractions—To evaluate IκBα degradation and NF-κB activation, nuclear and cytosolic fractions were prepared from macrophages (RAW 264.7; 2 × 10⁶ cells/well) incubated in the presence or absence of particulate P. carinii β-glucans (2.5 × 10⁶ particles/ml) or LPS (0.1 μg/ml) in parallel cultures. At specified times, the cells were scraped and suspended in 1 ml of buffer A (10 mm HEPES, 1.5 mm MgCl₂, 10 mm KCl, pH 7.9) freshly supplemented with 0.5 mm dithiothreitol, 10 μg/ml leupeptin, 2 μg/ml aprotinin, 2 μg/ml pepstatin, and 1 μm phenylmethylsulfonyl fluoride. Cytosolic preparations were made by the addition of lysis buffer (buffer A containing 0.1% Nonidet P-40) for 5 min, followed by centrifugation (6,500 rpm, 3 min, 4 °C). The supernatants were collected as the cytosolic fractions. The remaining pellets were resuspended in 15 μl of buffer B (20 mm HEPES, pH 7.9, 25% (v/v) glycerol, 0.42 m NaCl, 1.5 mm MgCl₂, 0.2 mm EDTA) supplemented with protease inhibitors as above. Nuclear suspensions were incubated for 30 min at 4 °C with intermittent agitation and then centrifuged (12,000 × g, 20 min, 4 °C) to remove nuclear debris. The supernatants were collected and diluted with 30 μl of buffer D (20 mm HEPES, pH 7.9, 20% (v/v) glycerol, 0.05 m KCl, and 0.2 mm EDTA) containing the inhibitors described above and designated as the nuclear fractions. Cytosolic and nuclear fractions were stored at –70 °C until assay. Protein concentrations in respective extracts were determined by Coomassie protein assay (Pierce) referenced against a bovine serum albumin standard.

Determination of Cytosolic IκBα—To first determine the effect of P. carinii cell wall β-glucan stimulation on macrophage levels of IκBα, immunoblot analysis was performed on cytosolic lysates prepared from RAW 264.7 macrophages stimulated with particulate P. carinii cell wall β-glucan. Cytosolic lysates from P. carinii β-glucan-stimulated cells were compared with lysates prepared from LPS-treated (0.1 μg/ml) macrophages. The lysates were separated on 10% polyacrylamide gels and transferred electrophoretically to nitrocellulose membranes. The membranes were blocked with 5% milk in Tris-buffered saline and incubated with polyclonal antibody recognizing the carboxyl terminus of IκBα (200 μg/ml; Santa Cruz Biotechnology, Santa Cruz, CA) for 1–2 h at room temperature. Subsequently, the membranes were washed and incubated with horseradish peroxidase-conjugated secondary antibody and detected by a chemiluminescence detection system ECL (Amersham Biosciences).

Evaluation of NF-κB Activation—To determine activation and nuclear translocation of NF-κB, electromobility shift assays (EMSA) were performed on nuclear lysates prepared from RAW 264.7 macrophages stimulated with either P. carinii β-glucan or LPS. A double-stranded DNA probe that binds NF-κB (5′ AGT TGA GGG GAC TTT CCC AGC 3′; Santa Cruz Biotechnology, Santa Cruz, CA) was radiolabeled with [γ-³²P]ATP, and EMSA performed on 5 μg of nuclear extract as described previously (27). To confirm specificity of DNA binding, in some experiments a 50-fold excess of unlabeled NF-κB probe was added to the reaction for 15 min prior to addition of the radiolabeled NF-κB probe. In further studies, supershift analyses were performed using antibodies to each of the specified NF-κB subunits (p50, 200 μg/0.1 ml; p65, 200 μg/0.1 ml; or c-Rel, 200 μg/0.1 ml; Santa Cruz). These antibodies were incubated with nuclear extracts for 15 min prior to the addition of radiolabeled DNA probe recognizing NF-κB. All of the EMSA reactions were subsequently incubated for 30 min at room temperature, and DNA-protein complexes were separated on 6% polyacrylamide gels and visualized by autoradiography.

Fluorescence Visualization of NF-κB Nuclear Translocation—To further define the time course of NF-κB activation following macrophage stimulation with P. carinii β-glucans, immunofluorescence was employed to detect translocation of the p65 component of NF-κB to the nucleus. RAW 264.7 cells were seeded onto sterile 22 × 22-mm glass coverslips in 6-well tissue culture plates and cultured with P. carinii β-glucans or LPS for the indicated times prior to fixation with 2% paraformaldehyde for 10 min at 37 °C. In parallel experiments, the cells were treated for 30 min with the NF-κB inhibitor PDTC (10 μm) prior to the addition of P. carinii β-glucan. After macrophage stimulation and fixation, the cells were subsequently permeabilized with methanol for 2 min at 22 °C and rinsed with phosphate-buffered saline. Blocking of nonspecific binding sites was performed with 5% goat serum in phosphate-buffered saline for 30 min at room temperature. Goat anti-mouse p65 (1 μg/ml; Santa Cruz) was added and incubated for 60 min at room temperature. The coverslips were washed extensively and incubated for 60 min with a 1:50 dilution of Texas Red dye-conjugated donkey anti-goat polyclonal antibody (Jackson ImmunoResearch Laboratories, New Grove, PA) or 1:1000 dilution of fluorescein isothiocyanate-conjugated rabbit anti-goat polyclonal antibody (ICN-Cappel, Costa Mesa, CA). Nuclear staining was performed with 4′,6-diamidino-2-phenylindole (0.1 μg/ml, Sigma). Following extensive washing, the coverslips were mounted on slides and analyzed by fluorescence microscopy.

Conventional fluorescence microscopy was performed using an IX70 Olympus microscope equipped with filter packs. Fluorescein isothiocyanate and Texas Red labeling was observed under the fluorescence microscope using optics appropriate for these fluorophores (excitation = 470/40 nm, emission = 540/40 nm for fluorescein isothiocyanate; excitation = 540/25 nm, emission = 620/60 nm for Texas Red). Nuclear staining with 4′,6-diamidino-2-phenylindole was observed at excitation = 360/40 nm and emission = 460/50 nm. In any given experiment, all of the photomicrographs were exposed and printed identically. Quantitative image analysis was performed using the “Metamorph” image processing program (Universal Imaging Corp.) as previously described (28, 29).

Role of TLR-4 and MyD88 in Macrophage TNFα Release following P. carinii β-Glucan Stimulation—We have previously shown that isolated P. carinii cell wall β-glucans directly activate alveolar macrophages to release TNFα, a centrally important cytokine promoting lung inflammation during P. carinii pneumonia (6, 12). To begin to evaluate the potential roles of Toll-like receptors and the MyD88 toll adaptor protein in mediating macrophage activation, alveolar macrophages were collected from TLR-4 –/– mice (30), MyD88 –/– mice (31, 32), and strain-matched wild type controls (C57Bl/6). Macrophages were recovered by BAL as previously described (6). Two × 10⁵ alveolar macrophages were plated per well in 96-well tissue culture plates, allowed to adhere for 60 min, and gently washed to remove any unattached cells. Subsequently, varying P. carinii β-glucan concentrations (2.5 to 5 × 10⁶ particles/ml) were incubated with the alveolar macrophages for 12 h (37 °C, 5% CO₂). Following incubation, the medium was removed, centrifuged (10,000 × g for 5 min) to remove any particulate material, and assayed for the presence of TNFα by ELISA (BIOSOURCE International, Camarillo, CA).

RESULTS

P. carinii β-Glucans Induce Degradation of Cytosolic IκBα in Macrophages—Phosphorylation, ubiquitination, and degradation of IκBα represent key events in the activation and translocation of NF-κB to the nucleus (14, 17). We therefore sought to evaluate whether P. carinii β-glucans stimulate degradation of IκBα during the course of macrophage activation with P. carinii β-glucans. RAW 264.7 macrophages were stimulated with P. carinii β-glucans and IκBα levels evaluated by immunoblot analysis of cytosolic fractions (Fig. 1). Stimulation of macrophages with P. carinii β-glucans induced degradation of IκBα, which was evident following 2 h of stimulation and appeared to be virtually absent in cytosolic extracts following 4 h of incubation. To contrast these effects of P. carinii β-glucan to another potent stimulus of IκBα degradation, we further compared the kinetics to those resulting from LPS stimulation. In comparison, LPS induced more rapid degradation of IκBα, evident after as little as 15 min of macrophage stimulation. This rapid degradation of cytosolic IκBα following LPS stimulation is consistent with prior published observations (33) and illustrates that although P. carinii β-glucans and LPS both induce IκBα degradation, they exhibit markedly different time courses, strongly suggesting the use of different receptor-ligand interactions.

**Fig. 1** **Stimulation of RAW 264.7 macrophages with P. carinii β-glucan induces cytosolic degradation of IκBα**. RAW macrophage cells were treated with *P. carinii* β-glucan (2.5 × 10⁶ particles/ml) for the indicated times, and degradation of IκBα levels in cytoplasmic lysates was determined by immunoblot analysis. In parallel, the cells were stimulated with LPS (0.1 μg/ml) and IκBα levels in cytoplasmic lysates were also determined. Shown are representative blots of three experimental runs.

P. carinii β-Glucans Stimulate Nuclear Translocation of p65 NF-κB in Macrophages—Degradation of cytosolic IκBα enables NF-κB subunits to migrate to the nuclear compartment, where they promote gene transcription. We next investigated whether P. carinii β-glucans induce translocation of NF-κB components to nuclear compartments of stimulated macrophages. Because the p65 subunit of NF-κB has been demonstrated to exert critical activity in the transcription of many inflammatory genes (14, 17), we performed immunostaining to detect the nuclear translocation of the p65 subunit of NF-κB (Fig. 2). Although p65 NF-κB was localized predominantly in the cytoplasm of unstimulated RAW macrophages, stimulation with P. carinii β-glucans causes p65 NF-κB translocation to the nucleus in a time-dependent manner. This response was detectable as early as 1 h after stimulation and peaked between 2 and 4 h of stimulation (Fig. 2A). Nuclear translocation of p65 NF-κB began to decline after 6 h of continuous stimulation with P. carinii β-glucan.

**Fig. 2** **P. carinii β-glucans stimulate nuclear translocation of p65 NF-κB in RAW macrophages.**A, cells were treated with *P. carinii* β-glucan (*PCBG*; 2.5 × 10⁶ particles/ml) for the indicated times, and nuclear translocation of p65 NF-κB was visualized by immuno-staining with a specific antibody recognizing p65. Immunostaining was performed with fluorescein isothiocyanate-labeled secondary antibody yielding fluorescence in the *green* range. For reference the nuclei are stained with 4′,6-diamidino-2-phenylindole (*blue*). B, parallel experiments were performed on cells treated with 0.1 μg/ml of LPS. Nuclear translocation of p65 NF-κB was visualized by immunostaining with a specific antibody recognizing p65 and a Texas Red conjugated secondary antibody yielding fluorescence in the red range. C, quantitative differences in fluorescence intensities between nuclei and cytoplasm were calculated and plotted from three experiments.

In contrast, parallel experiments conducted with LPS revealed significant p65 NF-κB translocation to the nucleus as rapidly as 15–30 min of stimulation. Subsequent time points showed decreasing levels of p65 NF-κB inside the nuclear compartment (Fig. 2B). This redistribution of p65 NF-κB over time was evaluated by quantifying the nuclear/cytoplasmic ration of fluorescence intensity (Fig. 2C). P. carinii cell wall β-glucans transiently induced p65 NF-κB nuclear translocation peaking between 2 and 4 h (nuclear-to-cytoplasmic ratios of 0.848 ± 0.050 and 0.826 ± 0.026, respectively), whereas LPS-induced p65 NF-κB translocation had an abrupt increase between 15 and 30 min (0.903 ± 0.023 and 0.884 ± 0.020, respectively) and rapidly declined by 1 h (0.640 ± 0.030). Thus, consistent with our earlier observations on IκBα degradation, P. carinii cell wall β-glucan stimulates NF-κB translocation in macrophages in a delayed fashion, but NF-κB activation is sustained for a much longer period of time, compared with that following LPS challenge.

Kinetics of NF-κB Activation in RAW 264.7 Macrophages Stimulated with P. carinii β-Glucan—To further understand P. carinii β-glucan activation of macrophages, DNA binding activity of NF-κB was examined in nuclear extracts of RAW 264.7 cells stimulated with P. carinii cell wall β-glucan by EMSA. An increases in nuclear NF-κB content occurred in nuclear extracts of macrophages stimulated with P. carinii cell wall β-glucans (Fig. 3A). P. carinii β-glucan-induced NF-κB binding was evident 1 h after stimulation and appeared to decline by 6 h, once again in marked contrast with LPS-induced nuclear NF-κB binding activity. LPS-induced NF-κB binding was evident following 10 min of stimulation, peaked between 15 and 30 min, and was largely reduced at later time points (Fig. 3, B and C). In contrast, P. carinii β-glucan-induced NF-κB binding was slower in onset and was not detected at these earlier time points (Fig. 3C).

**Fig. 3** **Time course of DNA binding activity of transcription factor NF-κB in the nuclei of P. carinii β-glucan and LPS-stimulated RAW macrophages.**A, RAW cells were treated with *P. carinii* β-glucan (2.5 × 10⁶ particles/ml) for the indicated times. The nuclear fractions were isolated, and DNA binding activity of NF-κB was determined using EMSA. B, in parallel, the nuclear extracts were obtained from RAW macrophages stimulated with LPS (0.1 μg/ml) at the indicated times, and DNA binding activity of NF-κB was determined by EMSA. C, DNA binding activity of NF-κB from RAW cells treated with LPS (0.1 μg/ml) or *P. carinii* (PC) β-glucan (2.5 × 10⁶ particles/ml) for shorter periods of time was also analyzed by electromobility shift assay. *Lanes C*, control.

To next determine which components of NF-κB were present in the DNA-protein complexes, supershift and competition assays were performed (Fig. 4). Supershift analysis demonstrated an upper band formed by a heterodimer of p65 and p50 subunits and a lower band representing a homodimer of p50. Furthermore, oligonucleotide competition assays demonstrate that the DNA-protein complexes were formed specifically by interactions between the NF-κB probe and cognate nuclear proteins, because DNA binding of the radiolabeled NF-κB probe was fully inhibited by the addition of cold NF-κB probe but not by equimolar concentrations of unlabeled AP-2 probe. Taken together, these data strongly indicate that P. carinii β-glucans induce macrophage degradation of IκBα and nuclear translocation of p65 NF-κB. Furthermore, the kinetics of P. carinii β-glucan-induced NF-κB activation are distinctly different from those following LPS stimulation.

**Fig. 4** **Competitive inhibition and supershift analysis of NF-κB DNA binding in RAW macrophages challenged with P. carinii β-glucan.***Lane 1*, EMSA performed using nuclear fractions derived from unstimulated RAW cells (control). *Lanes 2–6*, EMSA performed with nuclear proteins derived from RAW macrophages stimulated with *P. carinii* β-glucan (2.5 × 10⁶ particles/ml) for 2 h. *Lane 2*, EMSA using radiolabeled NF-κB probe but no competitive probe nor added antibody. *Lane 3*, addition of excess unlabeled NF-κB probe inhibits binding of radiolabeled NF-κB probe, confirming specificity of the binding interactions. *Lanes 4* and 5, supershift assays performed with antibodies recognizing the p65 and p50 subunit of NF-κB protein, respectively, confirm the presence of these subunits in the DNA protein complexes. *Lane 6*, 50-fold excess unlabeled AP-1 probe did not alter the electromobility shift assay, further confirming the specificity of these interactions.

P. carinii β-Glucan-stimulated Nuclear Translocation of p65 NF-κB Is Inhibited by Pyrrolidine Dithiocarbamate—To begin to investigate whether pharmacological inhibitors could be potentially used to alter macrophage activation during challenge with P. carinii β-glucan, we studied whether PDTC, a specific and potent inhibitor of NF-κB (34, 35), would suppress activation of this transcription factor under these conditions. RAW 264.7 macrophages were preincubated with PDTC for 30 min, followed by 4 h of stimulation with the P. carinii cell wall β-glucan isolate. Treatment with PDTC dramatically reduced p65 translocation to the nucleus upon stimulation with the P. carinii cell wall isolate (Fig. 5A). RAW cell viability was not impaired by these concentrations of PDTC using the XTT viability assay.

**Fig. 5** **PDTC inhibited nuclear translocation of p65 NF-κB and also TNFα release in RAW macrophages stimulated with P. carinii β-glucan.**A, RAW cells were treated for 30 min with the NF-κB inhibitor PDTC (10 μm) prior to stimulation with *P. carinii* β-glucan (*PCBG*, 2.5 × 10⁶ particles/ml) for 4 h. Nuclear translocation of p65 NF-κB was visualized by immunostaining with a specific antibody recognizing p65 and a Texas Red-conjugated secondary antibody yielding fluorescence in the red range. The nuclei are stained *blue* with 4′,6-diamidino-2-phenylindole for reference. Immunofluorescence demonstrates suppressed nuclear translocation of p65 NF-κB in PDTC-treated cells challenged with *P. carinii* β-glucan. B, to further determine whether PDTC also inhibited inflammatory activation of macrophages, RAW cells were treated with PDTC at the indicated concentrations for 30 min prior to and throughout subsequent stimulation with *P. carinii* (PC) β-glucan (*PCBG*, 2.5 × 10⁶ particles/ml) over 8 h. Release of TNFα into the medium was determined by ELISA. TNFα release was significantly suppressed by the NF-κB inhibitor PDTC. Shown are the means ± S.E. *, p < 0.05 comparing *P. carinii* β-glucan stimulated RAWs in the presence and absence PDTC.

To further determine whether PDTC suppression of NF-κB nuclear translocation was also associated with inhibition of inflammatory activation of macrophages, RAW cells were stimulated with P. carinii β-glucan in the presence or absence of PDTC (Fig. 5B). Release of the pro-inflammatory cytokine TNFα was subsequently determined by ELISA. TNFα release following P. carinii β-glucan challenge was significantly suppressed by the NF-κB inhibitor in a concentration-dependent fashion (p < 0.05 comparing P. carinii β-glucan-stimulated RAW cells in the presence versus the absence of PDTC). RAW cell viability was not adversely affected by the agent. Taken together, these data indicate that P. carinii β-glucan-induced translocation of NF-κB is associated with macrophage activation to release inflammatory mediators such as TNFα.

P. carinii β-Glucan Activation of Alveolar Macrophages Is Independent of TLR-4 but Is Partially Mediated through MyD88-dependent Mechanisms—The striking similarities between LPS and fungal β-glucan-induced macrophage activation and cytokine release suggests parallel, although independent, mechanisms of inflammatory activation. The differential kinetics of IκBα degradation and nuclear translocation of NF-κB strongly support that macrophage activation induced by P. carinii β-glucan isolates is unrelated to contaminating LPS. In addition, we performed exhaustive washes of our P. carinii cell wall isolate and confirmed the absence of contaminating endotoxin in the final P. carinii β-glucan preparations.

Toll receptors that interact with fungal β-glucans have not yet been defined. To address the potential roles of TLR-4 in mediating β-glucan responses, we investigated the ability of P. carinii cell wall β-glucan and LPS to differentially activate TLR-4-deficient macrophages. Alveolar macrophages obtained from TLR-4 knockout mice were stimulated in parallel with either P. carinii β-glucan or LPS (Fig. 6). As anticipated, LPS-induced TNFα release in TLR-4-deficient alveolar macrophages was strikingly suppressed. For instance, LPS stimulation of TLR-4 –/– macrophages induced the release of only 5,829 ± 1,216 pg/ml of TNFα compared with 27,043 ± 1,506 pg/ml of TNFα released from wild type alveolar macrophages (p = 0.0001). In contrast, stimulation of TLR-4 –/– macrophages with P. carinii β-glucan resulted in 33,627 ± 3,952 pg/ml of TNFα, which was not significantly different compared with wild type macrophages. Similar data were obtained upon stimulation of alveolar macrophages from TLR-4-deficient mice with glucan preparations of the phylogenetically related fungus S. cerevisiae (data not shown). Taken together, these data convincingly demonstrate that the P. carinii cell wall β-glucans activate alveolar macrophages through receptors systems distinct from the classical LPS receptor TLR-4. In addition, these findings conclusively indicate that LPS contamination is not responsible for macrophage stimulation following challenge with P. carinii β-glucan.

**Fig. 6** **Alveolar macrophage TNFα response to P. carinii β-glucan is independent of TLR-4.** Alveolar macrophages harvested from TLR-4 –/– and wild type BALB/c control mice were treated with *P. carinii* β-glucan (2.5 × 10⁶ particles/ml) or LPS (0.1 μg/ml) overnight, and TNFα levels were measured in the medium by ELISA. Although macrophage response to LPS is dependent on TLR-4, no significant differences were observed between *P. carinii* β-glucan stimulation in macrophages from wild type and TLR-4 –/– mice. *, p < 0.05 comparing identical stimulation conditions in wild type and TLR-4 –/– macrophages.

Many Toll-mediated responses require activity of the Toll adaptor protein MyD88 (22). To further assess the activity of this family of innate immune receptors in mediating β-glucan responses, alveolar macrophages were obtained from MyD88-deficient mice and challenged with P. carinii β-glucan (Fig. 7). MyD88 –/– alveolar macrophages exhibited significantly reduced, although not completely suppressed, release of TNFα compared with wild type controls (Fig. 7). Although wild type control macrophages stimulated with P. carinii β-glucan released 22,743 ± 669 pg/ml of TNFα, MyD88 macrophages released only 13,026 ± 487 pg/ml (p = 0.001). These findings indicate that macrophage activation in response to P. carinii β-glucan is in part mediated through MyD88-dependent mechanisms.

**Fig. 7** **Alveolar macrophage TNFα response to P. carinii β-glucan is partially mediated through the MyD88, Toll adaptor protein.** Alveolar macrophages harvested from MyD88 –/– and wild type C57BL/6 control mice were treated with *P. carinii* β-glucan overnight, and TNFα release into the medium was measured by ELISA. MyD88 –/– exhibited significant, although partial, suppression of TNFα release in response to *P. carinii* β-glucan. *, p < 0.05 comparing *P. carinii* β-glucan stimulation in wild type and MyD88 –/– macrophages.

DISCUSSION

Macrophages orchestrate innate immunity by phagocytosing invading pathogens and coordinating local inflammatory responses. Both particulate and soluble β-glucans in fungi initiate macrophage-driven immune responses. We recently established that the opportunistic fungal pathogen P. carinii assembles a β-glucan-rich cell wall, which in addition to providing structural support for the organism also strongly induces the release of TNFα and macrophage inflammatory protein-2 from alveolar macrophages (3, 6). P. carinii β-glucans activate innate host defense responses through ligation of cognate glucan receptors on leukocytes (6, 12). However, specific signaling pathways involved in such activation have not been previously characterized. Our current investigations reveal that β-glucans components of the P. carinii cell wall strongly induce macrophage inflammatory activation through activation and nuclear translocation of NF-κB. We further demonstrate that glucan-driven inflammatory activation can be significantly suppressed by pyrrolidine dithiocarbamate inhibition of NF-κB activity. Finally, our investigations reveal that P. carinii β-glucans stimulate macrophage activation through mechanisms distinct from those utilized by bacterial LPS. In particular, P. carinii β-glucan activation of macrophage occurs independently of the classical LPS receptor, TLR-4, but does appear to be partially mediated by the MyD88 toll receptor adaptor protein.

Unlike the adaptive immune system, which requires days to weeks for generation of an effective response to infection, the innate system provides immediate and crucial host defenses against invading pathogens (19). This innate surveillance system responds by recognizing a restricted repertoire of highly conserved surface molecules representing indispensable structural components found in large groups of microorganisms. These structures (LPS, mannan, and peptidoglycan) are referred to as pathogen-associated molecular patterns and are recognized by pattern recognition receptors present on many cell types. One family of receptors that exerts pivotal functions in innate immunity is the interleukin-1 receptor/TLR family. Stimulation via the interleukin-1 receptor/TLR family leads to initiation of signaling cascades that culminate in activation of NF-κB and mitogen-activated protein kinases (36). Prototypic of such responses is TLR-4, which serves as a critical receptor conferring host cell recognition of LPS found on Gram-negative bacteria. In addition, most, although not all, toll receptor-mediated cellular responses involve interaction of cognate toll receptors with the MyD88 adaptor protein to effectively initiate cell signaling resulting in inflammatory activation (21, 31, 37).

The current study provides further support that cell wall β-glucans represent pattern recognition molecules triggering innate immune responses to pathogenic fungi such as P. carinii. The current study indicates that β-glucans components of the P. carinii cell wall activate macrophage NF-κB to stimulate expression of TNFα, a central mediator of lung inflammation during Pneumocystis pneumonia (38–40). Our study further reveals that the P. carinii cell wall glucan activates macrophages via receptors distinct from the TLR-4 receptor mediating cell responses to LPS. However, we further demonstrate that the MyD88 toll adaptor protein is at least partially responsible for conferring macrophage response to P. carinii cell wall glucan. Recent studies further indicate that curdulan, a linear nonpathogenic cell wall carbohydrate, signals macrophage activation through a Toll-like receptor pathway because expression of a dominant negative mutant of MyD88 rendered macrophages refractory to curdulan stimulation (41). Taken together, these data strongly implicate Toll-mediated processes in glucan-initiated macrophage activation.

Alveolar macrophages exert essential functions in host defense against P. carinii, mediating uptake and degradation of the organisms (8). In addition to Pneumocystis clearance, macrophages also strongly initiate inflammatory responses in the lung following interactions with β-glucan components of the organism, including secretion of TNFα, interleukin-6, interleukin-8, macrophage inflammatory protein-2, eicosanoid metabolites, and reactive oxidant species (6, 12, 38, 39, 42–44). These exuberant host inflammatory responses, although necessary for clearance of infection, further promote lung injury resulting in respiratory failure and excessive morbidity. Therapeutic maneuvers that limit excessive inflammation are beneficial to patients with Pneumocystis pneumonia (45–48). These observations provide considerable initiative to fully understand the mechanisms and signaling pathways induced by P. carinii cell wall glucans, which trigger inflammatory activation of macrophages (8).

The delay in macrophage NF-κB activation following fungal β-glucan stimulation is rather unique compared with activation of other innate immune responses. This delay may be related to the necessity of glucan particle internalization, novel protein synthesis, and/or recruitment of other secondary receptors to efficiently trigger cell activation. A previous study has shown that differing phagocytic mechanisms occur during ingestion of β-glucan through distinct binding sites of the CR3 β-glucan receptor, again possibly contributing to a delay in NF-κB activation (49). In addition, it is likely that several different glucan receptors, including CR3, toll receptors, and dectin-1, are present on the surface of inflammatory cells, exhibiting different affinities for various glucan isolates (50–54). Thus, the composition and conformation of the β-glucans in the fungal cell wall isolate as well as the combination of glucan-binding sites and receptors present on macrophages and other host cells will modify host responses to these fungal cell wall structures.

Effective NF-κB activation also likely requires optimal receptor recruitment, co-receptor interactions, and/or engagement of other signaling intermediates. Many Toll-like receptors depend on homodimeric or heterodimeric interactions with other Toll-like family members to elicit optimal responses (55). For example, two members of the TLR family, TLR-2 and TLR-6, together coordinately mediate macrophage activation by Gram-positive bacteria and yeast cell wall (56). Additional studies suggest that glucan polymers may cross-link spatially separated membrane receptors to induce immunomodulatory effects. Thus, the length and structural complexity of any species β-glucan likely modulate both the type and the strength of the elicited host responses (57, 58).

In summary, our study demonstrates that the β-glucan rich P. carinii cell wall is a potent stimulant of macrophage inflammatory responses through activation of NF-κB. The mechanisms of macrophage NF-κB activation induced by P. carinii are distinctly different from those employed by LPS, with β-glucan exhibiting delayed but more persistent activation kinetics. Furthermore, P. carinii β-glucan-induced TNFα release occurs through a TLR-4-independent pathway but is significantly suppressed in the absence of the Toll adaptor protein MyD88, suggesting elements of Toll activity in these responses. These studies provide a basis for initial characterization of the specific immune signaling pathways involved in macrophage activation during the course of P. carinii pneumonia. Careful characterization of these pathways may suggest novel therapeutic strategies to attenuate the host inflammation and the related mortality that accompany this infection.

Acknowledgments

We thank Drs. Zvezdana Vuk Pavlovic and Richard Pagano for many helpful discussions. We further appreciate the kind gift of ciprofloxacin from Dr. Barbara Painter (Miles Pharmaceuticals, Inc.). We also thank Kathy Streich for assistance during the final preparation of the manuscript.

References

Phair, J. ∙ Munoz, A. ∙ Detels, R. …

N. Engl. J. Med. 1990; 322:161-165