Pneumocystis carinii Major Surface Glycoprotein Dampens Macrophage Inflammatory Responses to Fungal β-Glucan

Abstract

Background

Pneumocystis major surface glycoprotein (Msg) is a 120-kD surface protein complex on the organism with importance in adhesion and immune recognition. In this study, we show that Msg significantly impairs tumor necrosis factor (TNF)-α secretion by macrophages induced by Saccharomyces cerevisiae and Pneumocystis carinii (Pc) β-glucans.

Methods

Major surface glycoprotein was shown to greatly reduce β-glucan-induced Dectin-1 immunoreceptor tyrosine-based activating motif (ITAM) phosphorylation. Major surface glycoprotein also down regulated Dectin-1 receptor messenger ribonucleic acid (mRNA) expression in the macrophages. It is interesting that Msg incubation with macrophages resulted in significant mRNA upregulation of both C-type lectin receptors (CLR) Mincle and MCL in Msg protein presence alone but to even greater amounts in the presence of Pc β-glucan.

Results

The silencing of MCL and Mincle resulted in TNF-α secretions similar to that of macrophages treated with Pneumocystis β-glucan alone, which is suggestive of an inhibitory role for these 2 CLRs in Msg-suppressive effects on host cell immune response.

Conclusions

Taken together, these data indicate that the Pneumocystis Msg surface protein complex can act to suppress host macrophage inflammatory responses to the proinflammatory β -glucan components of the organisms.

Pneumocystis jirovecii (Pj) is an opportunistic fungus that causes Pneumocystis pneumonia (PCP) worldwide. Statistics suggests that Pj may cause more than 400 000 life-threatening infections per year, with mortality rates between 20% and 80% [1]. Despite the advent of highly active antiretroviral therapy, which has significantly reduced PCP in the human immunodeficiency virus (HIV)-infected population [2], PCP in the developing world continues to represent a major threat [1]. Furthermore, recent evidence suggests that even in developed countries, the incidence of non-HIV PCP is on the rise, and, in this cohort, these patients have worse outcomes and significant in-hospital mortality [3]. For these reasons, PCP continues to be a major health concern.

Host responses to Pneumocystis organisms are varied depending on the life cycle forms that interact with host cells. Both an ascus and asexual trophic form are thought to be essential components of the life cycle. The ascus forms, which contain copious amounts of β-glucans, have been shown to be recognized by the glucan receptor Dectin-1, whereas studies suggest that the smaller trophic forms containing an abundant surface protein complex, termed major surface glycoprotein ([Msg] also known as gpA), interact with C-type lectin receptors (CLRs), Dectin-2, Mincle, MCL, mannose receptor (MR), and dendritic cell-specific intercellular adhesion molecule-3-grabbing nonintegrin (DC-SIGN), respectively [4–9]. In a recent study, researchers showed that trophic forms that harbor large amounts of Msg on their surface can act to suppress β-glucan-induced proinflammatory markers in vitro [10]. Furthermore, these researchers and others have shown that in dendritic cells, either trophic forms or its major cell wall mannoprotein Msg inhibits CD4⁺ T-cell responses as well as dendritic cell activation failure [8, 11].

Based on these studies, we sought to further determine the effects of Msg on innate immune proinflammatory response by macrophages stimulated with fungal β-glucans. Next, we further examined the effects of Msg on Dectin-1 immunoreceptor tyrosine-based activating motif (ITAM) phosphorylation. Furthermore, the expression profiles of known fungal interacting CLRs that modulate innate immunity to Pneumocystis were evaluated.

MATERIALS AND METHODS

Cell Culture

The macrophage cell line RAW 264.7 (TB-71; American Type Culture Collection [ATCC]) was purchased and maintained according to the specification. RAW 264.7 cells expressing V5-epitope tagged wild-type Dectin-1 were obtained from Dr. David Underhill (Cedars Sinai, Los Angeles, CA).

Reagents and Strains

All reagents were from Sigma-Aldrich unless specified otherwise [12]. All animal experiments were conducted in accordance with the guidelines of the Mayo Institutional Animal Care and Use Committee. Pneumocystis carinii (Pc) pneumonia was induced in rats immunosuppressed with dexamethasone as previously described [4]. Pneumocystis carinii organisms were derived from ATCC. Pneumocystis carinii was propagated for 10 weeks in immunosuppressed, corticosteroid-treated rats, as previously reported [13, 14]. Whole populations of Pc containing both cyst (ascus) and trophic forms were purified from infected rat lungs by homogenization and filtration through 10-μm filters. To exclude the presence of other infectious organisms in the Pc isolates, the preparations were stained (Diff-Quik modified Wright-Giemsa stain; Dade Diagnostics) and examined to exclude concurrent infection with bacteria or fungi. Isolates contaminated with contaminating microorganisms were discarded [15].

Major Surface Glycoprotein Isolation

Purified native Msg was generated from Pc based on its binding affinity to Concanavalin A [5, 8]. Organisms were suspended in 0.9% sterile saline with protease inhibitor tablets (Roche) and homogenized using a 7-mm steel ball in a TissueLyser LT (QIAGEN), followed by determining protein concentration. Concanavalin A agarose (Vector Laboratories) was prepared by washing several times with Dulbecco’s phosphate-buffered saline (DPBS) containing magnesium and calcium (pH 7.1) [16] and mixed with ∼50 mg of Pc homogenate and DPBS with a final protein concentration of Pc homogenate of ∼4 mg/mL. The slurry was incubated for 18 hours at 4°C end over end and then applied onto a glass chromatography column (Bio-Rad Laboratories, Hercules, CA). Column washing was conducted 3 times with 2 column volumes of DPBS, and associated Msg was eluted with 2 column volumes of 0.5 M α-methyl-d-mannoside. Collected fractions were dialyzed against 1× PBS in 12-kDa molecular mass cutoff Spectra/Por dialysis tubing (Repligen, Inc.). Western blotting was implemented to analyze final preparations utilizing a specific monoclonal antibody recognizing Msg (Figure 1A) [17]. Next, Msg purity was verified by silver stain and similar levels of purity verified as previously published [5, 18]. Finally, to verify that the Msg preparations contained no contaminating lipopolysaccharide (LPS), 10 ng/mL LPS alone or with 100 µg/mL Msg or bovine serum albumin (BSA) was incubated with RAW cells for 24 hours overnight at 37°C/5% CO₂. Tumor necrosis factor (TNF)-α secretion was analyzed by enzyme-linked immunosorbent assay (ELISA) (Life Technologies) (Figure 1B). Furthermore, to verify whether any possible residual Concanavalin A (~25 pg/well per test condition as detected by Concanavalin A ELISA in typical Msg preparations) was not interfering with the biological effects of Msg, this amount of Concanavalin A was also added to RAW cells as described below to determine whether TNF-α secretion was altered. As shown (Supplementary Figure 1), the presence of this amount of Concanavalin A had no effect on baseline or Pc β-glucan-stimulated TNF-α release.

In Vitro Stimulation of Macrophages

RAW 264.7 cells were plated at 2 × 10⁵/well in a 96-well plate for 4 hours at 37°C/5% CO₂. It has been reported that the bulk of β-glucan present on Pneumocystis is actually masked by Msg and other surface proteins, and, as organisms die, the glucans become more readily available to initiate inflammation [19]. By preincubating the macrophages with Msg before addition of the β -glucans, this preincubation mimics the situation that occurs during infection. Accordingly, after 4 hours of incubation, 100 µg/mL Msg was added to the respective wells and incubated for 1 additional hour. Next, 150 µg/mL Saccharomyces cerevisiae (Sc) (Sigma-Aldrich) or Pc β-glucans (isolated and characterized as we previously described) [15] were added, plates were incubated for 24 hours at 37°C/5% CO₂, and supernatants were collected as described above for TNF-α ELISA. Likewise, alveolar macrophages were isolated as previously described [20], Pc β-glucans were added, and TNF-α secretion was detected by ELISA as noted above.

Analysis of Dectin-1 ITAM Phosphorylation

RAW 264.7 cells expressing V5-epitope tagged wild-type Dectin-1 (2 × 10⁶/well) [21] in 48-well plates were plated and incubated for 4 hours. Next, macrophages were preincubated with 100 µg/mL Msg or BSA for 1 hour. Cells were then stimulated with 150 µ g/mL Sc β-glucans for 1 hour. Cells were lysed in 300 µ L lysis buffer I (1% sodium dodecyl sulfate [SDS], 10 mM Tris, 0.2 mM sodium orthovanadate, pH 7.4, with protease inhibitors) and insoluble material was pelleted. Protein quantity was determined, and 50 µ g of each sample was added to buffer II (1% Triton X-100, 50 mM NaCl, 10 mM Tris, 0.2 mM sodium orthovanadate, pH 7.4, with protease inhibitors) for a final volume of 750 µ L. Next, 40 µL of a slurry of V5 agarose were added, and the mixture was incubated overnight at 4°C. Agarose beads were then washed 3 times with buffer II, and protein was separated by SDS-polyacrylamide gel electrophoresis. Determination of Dectin-1 ITAM phosphorylation was conducted by Western blotting with anti-Tyr(P) monoclonal antibody (clone 4G10; Cell Signaling Technology). Dectin-1 ITAM phosphorylation was quantified using Image Studio Lite (version 5.2.5) normalized against total Dectin-1 levels.

Dectin-1 Carbohydrate Recognition Domain Solid-State Binding Assay

Major surface glycoprotein or Pc β-glucans (2.0 µg per well) were plated onto 96-well microtiter plates and incubated at 4°C overnight. Next, plates were washed 3 times with 100 µL PBS-Tween (PBS-T), and wells were blocked with PBS/10% fetal bovine serum/2.5% milk at 4°C for 2 hours. After 3 washes with PBS-T, hFc-fusion control, or hFc-Dectin-1 carbohydrate recognition domain (CRD) (kindly provided by Chad Steele, Tulane University) proteins (200 ng) [6, 22] were added in lectin binding buffer (50 mM HEPES, 5 mM MgCl₂, and 5 mM CaCl₂) for 2 hours at 4°C. Next, 1:1000 dilution factor of horseradish peroxidase goat antihuman antibody (SouthernBiotech) in blocking buffer was added for 1 hour at 4°C followed by washing and development. Plates were read in a VERSAmax microplate reader (Molecular Devices) at 450 nm.

Quantitative Polymerase Chain Reaction Analysis of Macrophage C-Type Lectin Receptor Expression

After treatment of RAW 264.7 macrophages with Msg and Pc β-glucans as described above, the cells were lysed using a Qiashredder. Total ribonucleic acid (RNA) was purified with the RNeasy Mini Kit (QIAGEN). iScript Select cDNA synthesis kit (Bio-Rad) was used for reverse-transcription 200 ng of RNA using oligo(dT) primers and random hexamer primer mix. A SYBR green polymerase chain reaction (PCR) kit (Bio-Rad) was used for quantitative real-time PCR performed on a CFX96 Touch system (Bio-Rad). The sequences of the primer pairs are listed in Table 1.

Small Interfering Ribonucleic Acid Study

Small interfering RNA (siRNA) transfections of RAW 264.7 cells were performed as follows: double-stranded chemically synthesized siRNAs were purchased from integrated DNA technologies (IDT). Two sequences corresponding to mouse Mincle were used as targets for siRNAs: 5’-CGAGCACUUCUAUUAGACUUUCUCA-3’ and 3’-AUGCUCGUGAAGAUAAUCUGAAAGAGU-5’ (mm.Ri.Clec4e13.1) and 5’-ACGACAACCUUGACCUGGUCAUCAA-3’ and 3’-GAUGCUGUUGGAACUGGACCAGUAGUU-5’ (mm.Ri.Clec4e13.2). For MCL, target siRNAs sequences were 5’-CAUUGUUUCCUUGAUAGUAUUUUAA-3’ and 3’-UAGUAACAAAGGAACUAUCAUAAAAUU-5’ (mm.Ri.Clec4d.13.1). An unrelated sequence with no significant homology to any mouse gene was implemented as a negative control (scrambled-siRNA) and also obtained from IDT. Small interferring RNAs were transfected into macrophages using Lipofectamine LTX reagent according to the manufacturer’s protocol and described by others [23]. At 24 hours posttransfection, media was removed and Msg plus Pc β-glucans stimulation assays described above for TNF-α analysis were conducted. Before these experiments, siRNA knockdown efficiency was detected at 24 hours by quantitative PCR (qPCR).

Statistical Analysis

For multigroup data, initial analysis was first performed with analysis of variance (ANOVA) to determine overall different differences. If ANOVA indicated overall differences, subsequent group analysis was then performed by 2-sample unpaired Student t test for normally distributed variables. Statistical analysis was performed using Prism version 5.0b software (GraphPad Inc.), and differences were considered to be statistically significant when P < .05.

RESULTS

Macrophage Inflammatory Responses Are Dampened by Pneumocystis carinii Major Surface Glycoprotein

Utilizing the macrophage cell line RAW 264.7, we investigated the activation of the proinflammatory marker TNF-α in response to fungal β-glucans in the presence or absence of Msg. Macrophages were preincubated for 1 hour with Msg and then stimulated with either Pc or Sc β-glucans for 24 hours. As we and others have previously shown, fungal β-glucans induce significant increases in TNF-α production by macrophages [24, 25]. Similar to the studies of Sassi et al, incubation with Msg alone causes little if any TNF-α production [8]. It is interesting to note that preincubating Msg before the addition of either Pc or Sc fungal β-glucans resulted in significant reduction in the amount of TNF-α secreted by the macrophages (Figure 2A and B). To verify that these Msg inflammatory dampening effects were not just observed in the immortal RAW 264.7 cell line, the same experiment was also conducted in freshly isolated primary alveolar macrophages. Similar to what was noted in the RAW 264.7 cell line, Pc Msg also resulted in significant reduction in TNF-α secretion when preincubated with the macrophages before Pc β-glucan addition (Supplementary Figure 2).

Pneumocystis carinii Major Surface Glycoprotein Reduces Macrophage Dectin-1 ITAM Phosphorylation in Response to β-Glucans

Due to the significant reduction of TNF-α production after incubation of macrophages with fungal β-glucan stimulation in the presence of Msg, we further investigated the activation of the principal myeloid β-glucan receptor, Dectin-1, under these conditions. Due to the large amounts of total reagents required for these immunoprecipitation experiments, we used Sc β-glucans for these studies, as the source for these carbohydrates, because it is available in large quantities. Furthermore, as we demonstrated above, Sc β-glucans behave similarly to Pc β-glucans in the magnitude of TNF-α released from RAW264.7 macrophages in response to the fungal cell wall component. As previously demonstrated [21, 26], when macrophages were treated with fungal β-glucan stimulation, we noted that these particles trigger robust phosphorylation of Dectin-1 in RAW cells expressing Dectin-1. It is worth noting that when macrophages were incubated with β-glucan in the presence of Msg, we observed significant reductions in the degree of Dectin-1 phosphorylation. Incubation of macrophages with fungal β -glucans in the presence of an irrelevant protein such as BSA resulted in no significant reduction in Dectin-1 phosphorylation (Figure 3A and B). To verify that the observed reduction of fungal β-glucan-induced Dectin-1 phosphorylation was not the result of Dectin-1 CRD interacting with Msg, we implemented an ELISA approach [4, 5, 22]. As shown in Figure 3C, the hFc-Dectin-1 CRD bound to β-glucan to a highly significant degree, whereas the binding of Pc Msg to hFc-Dectin-1 CRD was minimal and below the level of hFc control. Thus, Pc Msg results in reduced activation (phosphorylation) of Dectin-1, a principal β -glucan receptor, and this does not appear to occur through a direct binding of Msg to the host receptor.

Major Surface Glycoprotein Alters the Expression of C-Type Lectin Receptors Known to Interact With Pneumocystis

Our prior studies indicate that the expression of the various CLRs that interact with the organism are significantly altered during Pneumocystis infection [5]. Therefore, we further examined whether notable host CLR expression patterns were altered in the presence of Msg alone and when coincubated with Pc fungal β-glucan. It is interesting to note that Msg alone resulted in significantly reduction of Dectin-1 mRNA expression in macrophages incubated overnight with the Pc cell surface protein. After 24 hours, we noted appreciable lower Dectin-1 protein levels on repeated experiments in the presence of Pc Msg added to the β -glucan-stimulated macrophage cultures. These levels demonstrated a trend toward significance (P = .0517) but were slightly less than statistical significance (Supplementary Figure 3). Furthermore, incubation of the macrophages with Pc β-glucan in the presence of Msg also reduced Dectin-1 mRNA levels, but those levels did not achieve significant reduction (Figure 4). In contrast, we observed that Pc Msg alone could significantly elevate the expression levels of both Mincle and MCL. Furthermore, incubation of β-glucans in the presence of Msg resulted in even more profound increases in both of the respective transcript levels for Mincle and MCL (Figure 4). Taken together, these data indicate that Pc Msg can significantly alter the expression of various macrophage cell CLRs that are known to interact with the organism.

Inhibition of Macrophage C-Type Lectin and Mincle Expression Restores Macrophage Tumor Necrosis Factor-α Responses to Pneumocystis carinii β-Glucan

From the above observations demonstrating that MCL and Mincle were significantly induced by Msg, we next investigated the effects that silencing these 2 CLRs individually or in tandem might have on β -glucan-induced inflammatory activation of the macrophages. To this end, siRNAs were generated to both MCL and Mincle and used to knockdown their expression. Dual transfection of Pc Msg and MCL and Mincle siRNA together with the Pc β-glucan-challenged macrophages resulted in significant restoration of TNF- α secretion similar to the Pc β-glucan treatment group alone (Figure 5). Unlike the cotransfection of tandem transfection of both MCL and Mincle together, individual siRNA knockdowns of either MCL or Mincle alone by their respective siRNAs resulted in some restoration of TNF- α response by the β-glucan-stimulated macrophages, but not to a significant extent, compared with the macrophages stimulated with Pc β-glucan alone (Supplementary Figure 4). These findings suggest that Mincle and MCL may act together to modulate inflammatory activation of macrophages during β-glucan stimulation.

DISCUSSION

Interesting work by Evans et al [10] has demonstrated that life cycle forms of Pneumocystis murina can modulate host immune responses, with trophic forms reducing the inflammatory potential of macrophages in the presence of the water-insoluble β-glucan, curdlan. Furthermore, these researchers show convincingly that the cyst form, which is high in β-glucans [27], mediates bone marrow-derived dendritic cells and early T-cell proinflammatory responses, effects that are suppressed by trophic forms of the organism [10]. These data, along with others, further corroborates that the Pneumocystis cyst forms and their rich β-glucan cell wall [28] are the immune-stimulating life form [13, 29, 30] and that the diminutive, but numerous, tropic forms containing Msg are less immune sensitizing and may actually help the organism evade host immune responses. Indeed, Msg has been shown not to activate dendritic cells [8]. These studies provide interesting insights but do not directly address the role of Pc cell surface Msg in mediating these effects, nor do they address the potential roles of host CLR in these interactions.

Accordingly, we sought to directly determine whether Msg, the most abundant surface mannoprotein on the Pneumocystis trophic cell surface, may contribute to reducing inflammatory responses by macrophages in response to Pc β-glucan. Herein, we demonstrate that Msg itself [31, 32] has little inflammatory stimulatory effect in macrophages as measured by proinflammatory TNF-α release, in a manner similar to dendritic cells [8]. Strikingly, however, ours results further show that treating macrophages with Msg in addition to β-glucan challenge significantly reduces the inflammatory response by macrophages. Sassi et al has previously shown that even though the cyst forms contain Msg that may mask β-1,3 glucans, the trophic forms possess much greater levels of Msg [8]. Furthermore, intact mixtures of both cyst and life forms, versus just trophic forms alone, still elicit a robust inflammatory response [10], suggesting that β-glucan are still exposed to some degree on the surface of Pneumocystis cyst forms. Similarly, others have shown that early during Candida albicans infection, the organism is able to “cloak” its β-glucan content under mannan, whereas later during the infection, the β-glucan becomes exposed allowing Dectin-1 recognition and immune signaling [33]. The preponderance of trophic forms to cyst forms, as high as 10:1 [34] during PCP, may suggest that trophic forms containing little or no appreciable levels of β-glucans [27] but abundant levels of Msg, may act to mask cystic form surface carbohydrate, reducing host inflammation amd thereby providing an ecological niche for the organism to avoid immune surveillance, at least early during infection.

Also of note, we observed that the suppression of the inflammation response in the presence of MSG was complete for Pneumocystis β-glucan and less prominent for Saccharomyces β-glucan. Our prior studies indicate that fungal cell wall β-glucans vary in their potency and proinflammatory activities depending on the fungal source [26, 28, 35]. Indeed, as natural products, this is likely related to the relative amount of β-1,3 versus β-1,6 side chains, with the β-1,6 chains being largely responsible for the proinflammatory activity of the overall glucan preparation [28]. Our earlier work further demonstrates that overall Saccharomyces β-glucan preparations behave similarly to Pneumocystis β-glucan preparations, although the proinflammatory potency varies between the sources [26, 35]. To the best of our understanding, Saccharomyces β-glucan appears to activate inflammatory cell signaling through similar mechanisms compared with Pneumocystis β-glucan preparations [26, 28, 35].

We further analyzed whether the initial immune signaling event, phosphorylation of the ITAM motif of Dectin-1 upon β-glucan engagement, was altered by Msg. We did note a significant reduction in Dectin-1 phosphorylation in cells pretreated with Msg before carbohydrate addition. We further showed that Msg does not directly bind the CRD of Dectin-1, suggesting that this reduction in phosphorylation was not due to a direct ligand-receptor interaction. Furthermore, mRNA analysis by qPCR shows that Msg alone could significantly reduce the expression profile of Dectin-1, while concurrently significantly increasing MCL and Mincle transcripts. For both MCL and Mincle, incubation of macrophages with Msg in the presence of β-glucan resulted in even more significant expression levels of these 2 transcripts, respectively. Taken together, our data suggest that Msg may inhibit early Dectin-1 macrophage signal events induced by β-glucan, as well as longer term proinflammatory cytokine release. Furthermore, the silencing of both MCL and Mincle together, but not individually, can completely restore TNF-α secretion of macrophages responding to β-glucans, suggesting that these 2 receptors interact together or in conjunction with Dectin-1 in the presence of Msg to reduce host cell proinflammatory signaling.

CONCLUSIONS

Recent work on CLR biology has provided evidence that these receptors not only interact individually with their respective ligand(s), but they also interact together and provide signaling via “cross-talk.” For instance, multiple CLR binding of Schistoma mansoni eggs by DC-SIGN, Dectin-2, and Mincle are needed for effective clearance of the organism [36]. In C albicans, both Dectin-2 and MCL form a heterodimeric pattern-receptor, which summoned together provide a more potent inflammatory response than the respective homodimers alone [37]. A similar immune magnifying event was reported with MCL aiding the expression of Mincle in recognizing the Mycobacterium tuberculosis antigen trehalose-6,6’idimycolate (TDM) [38]. More recently, Del Fresno et al [39] have reviewed the flexible signaling events downstream of various CLRs. Several of these CLRs, depending on ligand, have the ability to provide not only activating immune signaling but also anti-inflammatory outcomes as well depending on ligand-receptor interactions. These CLR receptors include DC-SIGN, MR, Mincle, and dendritic cell immunoreceptor. Why Msg causes dampening of initial Dectin-1 dampening of macrophage inflammatory signaling followed by TNF-α cytokine release is not yet fully known. It has been suggested that Msg binding to various CLRs with different affinities may yield a dampening inflammatory result [8], such as that in a mixed population of trophic and cyst forms challenging the host, as occurs during infection. These early preliminary results along with previous evidence [8, 10, 11, 34] suggest that Msg might indeed be a unique virulence factor that the organism uses to create an anti-inflammatory environment beneficial to fungal proliferation and host immune avoidance. Additional studies will be required to dissect the multiple possible roles of multiple CLRs (such as Dectin-1, Dectin-2, MCL, and Mincle) in these observed altered host inflammatory responses.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. We thank David Underhill for providing the RAW 264.7 cells expressing V5-epitope tagged wild-type Dectin-1.

Financial support. This work was funded by the Mayo Foundation, the Walter and Leonore Annenberg Foundation, and the National Institutes of Health (Grant R01-HL62150; to A. H. L).

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.

References