The Interaction of Pneumocystis with the C-Type Lectin Receptor Mincle Exerts a Significant Role in Host Defense against Infection

Pneumocystis jirovecii is a highly successful opportunistic infection causing life-threatening pneumonia in immunocompromised individuals and representing one of the top 10 severe fungal infections worldwide (1–3). Pneumocystis species are yeast-like fungi belonging to the ascomycetes and closely related to the genus Taphrina, whose members are pathogens of certain flowering plants (4, 5). Recent analyses of Pneumocystis demonstrate that the genome is much more compact compared with related fungi, with reduction and/or absences of certain metabolic pathways, such as de novo amino acid synthase and transporter genes, suggesting that the fungus must obtain these components from the host and, therefore, resides in the host lung during its proliferative life cycle (6). The genomes of Pneumocystis spp. encode for an abundance of major surface glycoproteins (MSGs), or MSG/glycoprotein A (gpA) proteins comprising ∼3–6% of the total genome. MSG/gpA proteins are expressed on the Pneumocystis cell surface and are thought to be important for organism binding to lung epithelial cells and potentially for avoidance of host elimination (1, 2, 7, 8). These abundant proteins, along with β-1,3- and β-1,6-glucans that provide cell wall rigidity, are the major constituents of the Pneumocystis cell wall (9, 10). Additionally, these Pneumocystis surface components are recognized by host cells and participate in the initiation and modulation of host inflammatory responses during infection (9, 11–13). Such host inflammatory responses, although clearly involved in control of infection, are also associated with lung injury and impairment of respiratory function (14). Hence, a more complete understanding of Pneumocystis-initiated lung inflammation is clearly required to enhance control of infection and to minimize lung damage during Pneumocystis pneumonia (PCP).

During fungal infection, innate immune responses are initiated by binding fungal cell wall surfaces displaying pathogen-associated molecular patterns (PAMPS) to their cognate pattern recognition receptors (PRRs) on host cells (15). Included within this group of PRRs is the macrophage-inducible C-type lectin (Mincle; also called Clec4e). First described as the receptor for the Mycobacterium tuberculosis mycobacterial glycolipid trehalose-6,6′-dimycolate (TDM; also known as cord factor), this type II transmembrane receptor is an activating receptor, through the FcRγ, leading to early innate signaling events such as Syk activation (16, 17). Overall interest in Mincle is quickly growing as the receptor–ligand associations are being expanded, thereby increasing the receptor’s significance in response to various bacterial and fungal host–pathogens interactions (18–20). Furthermore, recent evidence suggests that Mincle is coregulated with another C-type lectin receptor (CLR), namely MCL, leading to enhanced innate immune responses (21, 22).

Given the importance of Mincle in other host–microbial interactions and disease, we investigated the potential role of this CLR in the recognition of Pneumocystis spp. We noted significant binding of Mincle to Pneumocystis carinii life forms, as well as the cell wall component MSG/gpA on the organism. Enhanced expression of Mincle in the rodent PCP model was also noted. Furthermore, we identified that Mincle functions as an important signaling receptor upon P. carinii association. Significantly, we observed that Mincle^−/− mice exhibit significantly higher Pneumocystis murina burdens during infection compared with wild-type mice, and that these Mincle^−/− mice also demonstrate elevated levels of TNF-α, IL-6, and IL-1Ra during infection. Interestingly, the CD4-deficient Mincle^−/− mice with PCP did not demonstrate worsened survival, despite the elevated organism burdens. This is perhaps related to upregulation of anti-inflammatory pathways such as IL-1Ra. Furthermore, during infection the Mincle^−/− mice showed enhanced expression of other CLRs, including Dectin-1, Dectin-2, and MCL, in the face of this enhanced infection burden. Taken together, our observations support a significant role for Mincle in host defense against Pneumocystis infection.

Mice with a targeted mutation in Clec4e (Mincle^−/−) on the C57BL/6 background were generated previously (23). Wild-type C57BL/6 mice were from Charles River Laboratories (Wilmington, MA). Approximately equal numbers of both male and female mice aged 6–10 wk were used in all experiments. All animal protocols were approved and conducted in accordance with the procedures from the Mayo Institutional Animal Care and Use Committee.

RAW 264.7 macrophage cells (TB-71; American Type Culture Collection, Manassas, VA) were maintained in DMEM 4.5 g/l glucose, 2 mM l-glutamine supplemented with 10% heat-inactivated FBS (HyClone; GE Healthcare, Logan, UT), and Life Technologies antibiotic/antimycotic (Life Technologies, Grand Island, NY). Bone marrow–derived macrophages (BMDMs) were isolated from the respective wild-type and Mincle^−/− mice by standard procedures and used within 5–7 d after initial isolation and culture (24).

Unless specified otherwise, all reagents were from Sigma-Aldrich. P. carinii and P. murina organisms were derived from the American Type Culture Collection (25). P. carinii was propagated for 10 wk in immunosuppressed, corticosteroid-treated rats, as we reported (26, 27). Whole populations of P. carinii 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 P. carinii and P. murina 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 other microorganisms were discarded (28).

Native P. carinii MSG/gpA was isolated from purified P. carinii based on its binding affinity to Con A (29, 30). P. carinii 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 instrument (Qiagen), followed by protein concentration determination. Con A agarose (Vector Laboratories) was prepared by washing several times with Dulbecco’s PBS (DPBS) containing magnesium and calcium (pH 7.1) (31, 32) and mixed with ∼50 mg of P. carinii homogenate and DPBS with a final protein concentration of P. carinii homogenate of ∼4 mg/ml. The slurry was incubated for 18 h at 4°C end over end and then applied onto a glass chromatography column (Bio-Rad Laboratories). Column washing was conducted three times with two column volumes of DPBS and associated MSG/gpA was eluted with 2 column vol of 0.5 M α-methyl-d-mannoside. Fractions were collected and dialyzed against 1× PBS in 12,000-kDa molecular mass cut-off Spectra/Por dialysis tubing (Spectrum Laboratories). The purified preparations were analyzed by Western blot using a specific mAb recognizing P. carinii MSG/gpA (33). Furthermore, MSG/gpA purity was verified by silver stain as conducted previously (34) (Supplemental Fig. 1). Additionally, total unfractionated P. carinii homogenates were also generated from isolated P. carinii organisms. Purified organisms were suspended in 0.9% saline with protease inhibitor tablets (Roche) and homogenized as noted above. Final homogenates were centrifuged at 13,000 × g at 4°C to collect the insoluble cellular components, resuspended in saline, and aliquots were stored at −20°C.

Total P. carinii organisms were resuspended in 4% formaldehyde for 1 h at room temperature (9, 35). P. carinii organisms were then washed three times with PBS and either the Fc portion of human IgG1 fragment alone or human Fc fragment fused at the N terminus to the extracellular domain of mouse Mincle (aa 53–214) (36) was added to 1 × 10⁶ P. carinii (10 μg/ml) in 200 μl of binding buffer (20 mM Tris-HCl, 150 mM NaCl, 10 mM CaCl₂, 0.05% Tween-20 [pH 7.4]) and incubated overnight at 4°C. Organisms were then washed three times with binding buffer and incubated for 2 h at 4°C with 1:100 diluted FITC-conjugated goat anti-human Fc Abs (Life Technologies) in binding buffer. The P. carinii life forms were then washed three times with binding buffer and applied to poly-l-lysine–coated slides and analyzed for potential Mincle/P. carinii binding using an Olympus IX70 microscope with ×10 ocular and ×100 objective lenses under oil immersion with phase contrast or FITC filters (×1000 magnification total).

Lungs were harvested after 10 wk of P. carinii infection in rats or P. murina infection in mice and tissue samples (30 mg) were homogenized using a TissueLyser LT (Qiagen) at 50 oscillations/s for 5 min. Total RNA was isolated, and an aliquot (200 ng) was used to generate cDNA. Steady-state mRNA expression of Mincle in these samples was determined by quantitative PCR (qPCR) analysis conducted with the respective primer sets (Supplemental Table I) and was expressed as normalized to GAPDH to verify equal cDNA contents.

To assess binding of Pneumocystis surface components to Mincle, we used a solid-state binding assay. P. carinii homogenate or native P. carinii MSG/gpA (5.0 μg per well) was plated individually into wells of a 96-well microtiter plate and incubated at 4°C overnight to coat the well. The next day, the plates were washed with 3× 100 μl of 1× TBS with Tween 20 and then blocked with 1× PBS/10% FBS/2.5% milk at 4°C for 2 h. After three washes with 1× TBS with Tween 20, the Mincle-Fc fusion proteins or Fc fragments alone were added at 10 ng/ml for 2 h at 4°C. Next, a 1:1000 dilution of goat anti-human IgG Fc HRP in blocking buffer was added for 1 h at 4°C. Finally, after washing, the 3,3′,5,5′-tetramethylbenzidine substrate was added for 20 min at room temperature and 2.0 N H₂SO₄ was added to stop the reaction. The absorbance for each well was determined in a spectrophotometer at 450 nm. In a similar fashion, to further demonstrate specificity of Mincle-Fc fusion binding to specific P. carinii cell wall fungal components, crab shell chitin (5.0 μg per well) was also plated as described above and the binding percentage was compared with native P. carinii MSG/gpA. Lastly, to further demonstrate specificity and to determine whether the Mincle-Fc fusion carbohydrate recognition domain (CRD) binding site for P. carinii was similar to the binding region of the CLR to trehalose-6,6′-dibehenate (TDB), a synthetic analog of M. tuberculosis TDM was used. Briefly, increasing doses of TDB were incubated in the presence of the Mincle-Fc fusion or Fc fusion control (10 ng/ml) for 1 h at room temperature. This mixture was then applied to microtiter plates coated with P. carinii MSG/gpA as noted above and binding ability was determined as above.

To further characterize cellular interactions of Mincle with Pneumocystis, a murine macrophage cell line overexpression the Mincle CLR was constructed. A full-length mouse Mincle cDNA containing the C-terminal V5 epitope was synthesized by GenScript Biotech containing BamHI (5′ end) and NotI (3′ end) restriction sites for subsequent cloning. This cDNA fragment was then subcloned into the lentiviral vector plasmid, pHR-SIN-CSGWdlNotI. COS-1 cell gene delivery and subsequent infection of RAW 264.7 mouse macrophage cells were as described previously (37). Western blotting with a V5-HRP Ab (Life Technologies) as well as a specific mouse Mincle Ab (InvivoGen) was used to verify protein expression in the transfected RAW cells. FACS analysis using the anti-V5 Ab was further used to confirm surface expression.

Total P. carinii organisms were isolated and labeled with ⁵¹Cr as described previously (38). After labeling, 50 P. carinii organisms were added per RAW macrophage cell in six-well tissue culture plates for 30 min at 37°C. Next, nonadherent P. carinii organisms were removed by gentle washing. The surface-bound macrophages containing associated P. carinii organisms were solubilized with 1 N NaOH, and total bound ⁵¹Cr-labeled P. carinii organisms were quantified by scintillation counting. Similarly, ⁵¹Cr P. carinii organisms were applied at the same organism-to-cell ratio to either wild-type or Mincle^−/− BMDMs and incubated and washed as described above for the RAW cells. Lastly, in a separate sets of experiments, 50 Saccharomyces cerevisiae yeast cells labeled with ⁵¹Cr were added per RAW macrophage cell and measured for the ability to bind the representative RAW cell lines in a head-to-head comparison versus P. carinii organisms.

To initially assess cell signaling activation induced by Pneumocystis, we assessed total tyrosine phosphorylation in RAW macrophage cells following stimulation with whole P. carinii organisms. RAW cells (2.5 × 10⁶) were starved for 1 h in serum-free DMEM at 37°C, and subsequently ∼10 P. carinii per macrophage were added and incubated for the times indicated. Similarly, wild-type or Mincle^−/− BMDMs were stimulated with the same P. carinii organism-to-cell ratio. Total tyrosine-phosphorylated proteins were analyzed by Western blot as described previously (37).

To determine Mincle-based cellular signaling following interaction with Pneumocystis, the T cell hybridoma line B3Z containing the lacZ β-galactosidase reporter and control constructs was employed as previously described (39). Briefly, 1 × 10⁵ B3Z/FcRγ or B3Z/FcRγ-expressing Mincle cells per well were cultured in a 96-well culture plate and incubated for 18 h with P. carinii organisms at the concentrations indicated. This cell-based reporter construct mediates generation of lacZ activity following ligation of Mincle with relevant ligands and subsequent FcRγ activation. The lacZ activity was measured in total cell lysates using chlorophenol red–β-d-galactopyranoside (Roche) as a substrate with the OD at 560–620 nm being measured using a microtiter plate reader (Molecular Devices).

First, RAW 264.7 macrophage cells were transfected with a plasmid containing Mincle small interfering RNA (siRNA) to reduce Mincle expression. The siRNA-containing plasmids for Mincle and GFP control plasmid were obtained from InvivoGen. RAW 264.7 cells (2 × 10⁶) were transfected with 2 μg of each plasmid using Amaxa Cell Line Nucleofector kit V with the Amaxa Nucleofector 2b instrument (Lonza). Transfected cells were then transferred to six-well plates in 1.5 ml of media and incubated for 48 h at 37°C. Cells were then transferred to T-75 cm flasks in media supplemented with selection antibiotic (50 μg/ml Zeocin) per the manufacturer’s recommendations using optimal concentrations determined in preliminary experiments. Silencing of mRNA was confirmed with quantitative real-time PCR using the primers listed in Supplemental Table I. Selected Mincle and control GFP siRNA RAW cells were seeded with the appropriate selection media on six-well plates at 5 × 10⁵ cells per well and incubated at 37°C/5% CO₂ for 4 h prior to treatment. Likewise, either wild-type or Mincle^−/− BMDMs were also isolated as described and used in a similar fashion. P. carinii was added to the cells at a 10:1 cell ratio and incubated for 30 min at 37°C/5% CO₂ with the siRNA RAW cells, or for the times listed for the BMDMs. The cells were then rinsed with ice-cold PBS, lysed, and extracted proteins were analyzed for total and phospho-Syk levels by SDS-PAGE and Western blotting. The phospho-Syk Ab (1:1000 dilution; Cell Signaling Technology, catalog no. 2710) was used to measure Syk phosphorylation levels. Subsequently, the blots were stripped and reprobed for total Syk (1:2000 dilution; Cell Signaling Technology, catalog no. 2712) to confirm equal protein loading.

For all animal studies, the mice were age and sex matched. Both wild-type and Mincle^−/− mice were immune suppressed by CD4 lymphocyte depletion using the rat anti-mouse CD4 mAb GK1.5 as previously reported (40). Spleen cells of both Mincle^−/− and wild-type animals were depleted of Th lymphocytes following two i.p. injections of 0.3 mg GK1.5 during the first week. The mice were anesthetized with i.p. ketamine (100 mg/kg) and xylazine (10 mg/kg) and then inoculated intratracheally with P. murina organisms (5 × 10⁵ in 50 μl of saline). After the first week of infection, the mice received a second intranasal infection with P. murina organisms. Control animals were given an equal volume of saline intratracheally and were treated during 8- to 10-wk study periods with 250 mg of sulfamethoxazole and 50 mg trimethoprim (Teva Pharmaceutical Industries) per kilogram body weight in the drinking water to prevent any secondary infections during immune suppression. During the entire study period, mice continued to receive weekly injections of GK1.5 and were additionally immune suppressed throughout the study by injecting 1 mg of hydrocortisone i.p. two times per week. For the cytokine and organisms burden analysis experiment, the mice had fulminant P. murina infections after 10 wk, and they were euthanized and organisms and cytokine burdens quantified as we have previously reported (40). Lung samples were fixed in 10% phosphate-buffered formalin, paraffin embedded, and 5-μm sections were stained with Gomori methenamine silver and H&E using conventional methods. Slides were analyzed for the presence and severity of inflammation and for P. murina organisms. Additional lung tissue samples were collected for RNA isolation and determination of P. murina organism burden by quantifying 16S mitochondrial rRNA (41) using the PCR primers listed in Supplemental Table I. Survival analysis was also performed with identical groups of mice in two separate experiments containing 11–12 mice in each group. No significant survival differences were noted for up to 10 wk of PCP. As the mice were excessively moribund, further time points were deemed unethical. Finally, mouse lung cytokine levels were determined on lung tissue lysates by ELISA.

Statistical differences between various experimental conditions were first assessed using ANOVA and then by Student t tests as indicated. Nonparametric statistics were used when data were distributed in a non-Gaussian manner. Statistical testing was performed using GraphPad Prism version 5.0b software, and statistical differences were considered significant when p < 0.05.

To first define the potential role of Mincle in innate immune interactions with Pneumocystis, we assessed the ability of a Mincle-Fc fusion protein to recognize and bind to P. carinii life forms. We observed that Mincle-Fc fusion protein indeed bound to P. carinii very strongly, with homogeneous staining of the cyst wall (Fig. 1B). In contrast, there was no appreciable staining observed when using the Fc fusion fragment alone without the Mincle receptor component (Fig. 1D). Thus, the carbohydrate recognition site of Mincle recognizes and binds to P. carinii cysts forms.

Previous studies have indicated that Mincle mRNA is elevated in macrophages infected with bacteria in vitro as well as in bacterial and fungal pneumonia models (20, 42–44). Therefore, in a parallel fashion, we evaluated whether overall Mincle mRNA expression was increased in a rat PCP infection model. Notably, Mincle mRNA was enhanced to a significant degree in P. carinii–infected lung tissue compared with normal uninfected lung RNA, documenting increased levels of Mincle during PCP (ratio of Mincle RNA to GAPDH RNA in uninfected rat lungs = 0.36 ± 0.29 versus PCP rat lungs = 2.12 ± 0.43, p < 0.05).

To further demonstrate whether the Mincle CRD-containing fusion protein could bind P. carinii components, total P. carinii homogenates, as well as the isolated P. carinii cell wall components MSG/gpA, were tested using a plate-binding assay where the P. carinii components were bound to microtiter plates and the Fc Mincle fusion protein or Fc fragment alone was allowed to bind (Fig. 2). Immobilization of either the P. carinii homogenate or purified P. carinii cell wall MSG/gpA yielded significant binding of Mincle (Fig. 2). Furthermore, incubation of Mincle-Fc fusion with another known fungal component, chitin, showed relatively little binding, indicating that Mincle interacts with specific P. carinii cell wall components rather than widely with all fungal constituents (Supplemental Fig. 2A). Lastly, competition experiments with a known Mincle agonist, the mycobacterial synthetic cord factor analog TDB, showed significant inhibition of Mincle-Fc fusion binding to native P. carinii MSG/gpA in a dose-dependent manner, suggesting that Mincle engages the same binding site when interacting with this P. carinii glycoprotein (Supplemental Fig. 2B). Taken together, these findings demonstrate, to our knowledge for the first time, that Mincle recognizes not only another important fungal pathogen, but also MSG/gpA, a specific P. carinii cell wall ligand.

RAW 264.7 macrophages expressing full-length Mincle cDNA in an expression vector with a C-terminal V5 epitope tag were generated. Western blotting with either Mincle-specific or anti-V5 Abs confirmed the expression of the protein by the RAW cells (Fig. 3A). The surface expression of Mincle was further confirmed by FACS analysis using the anti-V5 Ab (Fig. 3B). Next, ⁵¹Cr-labeled P. carinii organisms were incubated with the transfected RAW cells, washed to remove nonadherent P. carinii life forms, and bound P. carinii life forms were enumerated by scintillation counting. RAW cells overexpressing Mincle exhibited significantly greater binding of P. carinii compared with control RAW cells expressing basal levels of Mincle. ⁵¹Cr-labeled P. carinii organisms alone had very little binding to the microtiter plate (Fig. 3C). Furthermore, and consistent with previous reports of Mincle binding to S. cerevisiae (23), we also noted no increase in binding of radiolabeled S. cerevisiae using the Mincle overexpressing RAW cells compared with the parental RAW cell line alone (Supplemental Fig. 3A). Lastly, application of ⁵¹Cr-labeled P. carinii organisms on either wild-type or Mincle^−/− BMDMs demonstrated significantly less binding of the organisms to the Mincle^−/− BMDM cells (Supplemental Fig. 3B). These findings indicate that Mincle CLR in native configuration expressed on the surface of macrophages can significantly bind P. carinii organisms, further supporting a role for these CLR interactions during Pneumocystis infection.

It has previously been demonstrated with C. albicans that Dectin-2, another CLR family relative of Mincle, leads to the induction of pan-specific tyrosine phosphorylation signaling through association of Dectin-2 with FcRγ and subsequent cell activation (37). Because Mincle also activates downstream signaling via FcRγ interactions, we examined whether challenging macrophages that overexpress Mincle with P. carinii organisms would similarly induce increased tyrosine phosphorylation of macrophage proteins. After applying P. carinii organisms to this cell line or to the parental cell line alone, a pan-specific antiphosphotyrosine Ab was used for Western blotting of the macrophage cellular lysates. Although the parental cell line demonstrated some increases in protein tyrosine phosphorylation following addition of P. carinii, the macrophage cells overexpressing Mincle displayed substantially greater overall tyrosine phosphorylation following stimulation with P. carinii (Fig. 4A). Furthermore, similar experiments involving the addition of P. carinii organisms to either wild-type or Mincle^−/− BMDMs demonstrated that Mincle^−/− BMDMs have considerably less tyrosine phosphorylation compared with their wild-type counterparts (Fig. 4B). Thus, taken together, these observations further suggest that interactions of P. carinii with Mincle CLRs stimulated tyrosine-based cell signaling.

To test this, increasing concentrations of P. carinii organisms were added to T cell hybridoma cells expressing both Mincle and FcRγ or FcRγ alone. In these reporter cells, the binding of Mincle to an appropriate ligand results in ligation of Mincle to the FcRγ and subsequent signaling as demonstrated by activation of lacZ (39). Accordingly, the addition of P. carinii to the reporter cells resulted in a significant dose-dependent increase in lacZ reporter activity in the reporter cells expressing both Mincle and FcRγ compared with the T cell line expressing FcRγ alone (Fig. 5). These results indicate that not only can Mincle bind P. carinii, but that these binding events lead to FcRγ-mediated intracellular signaling.

To further characterize whether interactions of macrophage Mincle receptors with P. carinii result in signal transduction events, we used siRNA to specifically evaluate whether P. carinii interactions with macrophage Mincle receptors activate Syk signaling. Using this approach, we observed that P. carinii organisms indeed activate the Syk cascade as measured by phospho-Syk, and that phosphorylation of Syk is substantially decreased by reducing macrophage Mincle expression using siRNA-targeting plasmid (Fig. 6A, 6B). We did still observe some levels of phospho-Syk in the Mincle siRNA-treated macrophages following stimulation with P. carinii, which may be due to Dectin-1 and/or other CLR activation that has previously been implicated in innate responses to Pneumocystis (45, 46).

Likewise, addition of P. carinii organisms to BMDMs lacking Mincle displayed significant reduction in phospho-Syk activity compared with the wild-type BMDMs with intact Mincle receptors (Fig. 6C). As a whole, these observations indicate that interactions of Mincle CLRs with P. carinii substantially increase innate immune signaling through phospho-Syk.

With the abundance of multiple redundant CLRs and other receptors implicated in host defense, the most rigorous manner to determine whether a receptor such as Mincle participates in host responses during PCP is to assess the impact of the absence of that receptor during active infection. To address this, Mincle^−/− mice were depleted of CD4 cells and inoculated with P. murina organisms. After 10 wk of ongoing immune suppression, the total P. murina organism burdens were determined in the lung by qPCR (Fig. 7). Significantly greater P. murina organism burdens were observed in the Mincle^−/− mice compared with wild-type P. murina–infected mice by qPCR (Fig. 7A). We further sought to determine whether there was enhanced mortality in the CD4-depleted mice lacking the Mincle CLR compared with wild-type CD4-depleted mice with P. murina pneumonia (Fig. 7B). Interestingly, for as long as 10 wk of P. murina pneumonia, we did not observe excess mortality in the Mincle^−/− mice with PCP despite the markedly enhanced organism burdens. Interestingly, there was a trend toward better mortality in Mincle^−/− mice with PCP (p = 0.071 by Fisher exact test, n = 23 mice inoculated per group). However, this did not reach statistical significance. Later time points were not possible, as the mice were excessively moribund, and further time points were deemed unethical. Silver staining for P. murina organisms in lung tissue from the Mincle^−/− mice confirmed visually the presence of significant larger and more abundant clusters of cysts as compared with the wild-type P. murina–infected mice that were identically immune suppressed and inoculated (Fig. 8). Overall, these finding indicate that Mincle CLRs are needed for optimal clearance of Pneumocystis organisms during infection.

Owing to our observations that Mincle^−/− lungs exhibit greater Pneumocystis organism burdens compared with wild-type mice, we further assessed whether markers of lung inflammation were altered in the two infected mouse populations studied (Fig. 9). We observed significant increases in several key cytokines and regulators of inflammation in Mincle^−/− P. murina–infected mice lungs compared with wild-type P. murina–infected mice (Fig. 9). We observed significant increases in IL-1β. Interestingly, we also observed a very significant increase in IL-1Ra, with nearly a 3.2 ± 0.7-fold increase in expression of this inflammation-modulating cytokine in the P. murina–infected Mincle^−/− compared with P. murina–infected wild-type mice (p ≤ 0.01 comparing IL-1Ra in P. murina–infected Mincle^−/− to P. murina–infected wild-type mice). Additionally, we observed a trend toward increased IL-1α in the infected Mincle^−/− mice. Of interest, we observed significantly decreased levels of IL-17, a cytokine implicated in Pneumocystis host defense (45, 47), in the P. murina–infected Mincle^−/− compared with P. murina–infected wild-type mice. Finally, although there were some increases in the levels of TNF-α and IL-6 production in the P. murina–infected Mincle^−/− mice, these differences did not meet statistical significance. Taken together, these data indicate that in the absence of Mincle, the clearance of Pneumocystis is impaired and cytokine production and host defense against infection are abnormally regulated.

Although the reasons for these increases in these inflammatory cytokines in the absence of Mincle receptors are not fully known, the increased organism burdens may be driving these proinflammatory responses. Additionally, we further postulated that other CLRs implicated in Pneumocystis host defense (48, 49) or receptors reported to participate in host response to other fungi (20, 23, 27, 50–52) might be upregulated to compensate for the loss of Mincle in the infected Mincle^−/− animals during infection. Strikingly, our analysis revealed that the fungal immune receptors Dectin-1, Dectin-2, and MCL were significantly upregulated in the Mincle^−/− P. murina–infected mice compared with the control P. murina–infected mice (Fig. 10). Importantly, note that we observed markedly increased P. murina organism burdens in the absence of Mincle but in the presence of enhanced levels of Dectin-1, a receptor that has been previously implicated in control of Pneumocystis infection. Thus, we conclude that Mincle CLRs participate in control of PCP, and their absence leads to decreased Pneumocystis clearance, despite increased levels of other known fungal CLRs postulated to suppress PCP infection in the host. Interestingly, despite this increased burden of organisms, survival was not worsened in the Mincle^−/− mice with P. murina pneumonia. This perhaps occurs through alteration of the balance between proinflammatory and anti-inflammatory cytokines such as IL-1Ra during this infection.

Mincle is one of the main CLR PRRs that interacts with various bacterial and fungal cell wall components (16, 23, 50, 53). In the present study, we demonstrate that Mincle is also an important receptor participating in host defense during Pneumocystis infections. Mincle expression is enhanced during pneumonia, and the Mincle CRD interacts specifically with cell surface components of the organism, thereby inducing signaling through ligations of FcRγ. Furthermore, Mincle^−/− mice were shown to exhibit significantly higher burdens of Pneumocystis organisms. Interestingly, despite the enhanced levels of organisms in Mincle^−/− mice with P. murina infection, these mice did not exhibit worse survival during pneumonia. In fact, there was a trend toward better survival in the Mincle^−/− mice with PCP, but this did not reach statistical significance. We further observed higher expression of certain proinflammatory cytokines, as well as the anti-inflammatory molecule IL-1Ra in the lung during P. murina infection in the absence of Mincle. These observed differences in cytokine expression could have been, in part, related to the substantially enhanced organism burdens in the CD4-depleted Mincle^−/− mice during P. murina infection. However, note that the enhanced burdens of P. murina organisms in the Mincle^−/− mice occurred despite enhanced expression of Dectin-1 in these mice, a receptor previously implicated in PCP defense (54). We further posit that the overall relatively preserved survival during PCP in mice lacking Mincle may reflect an alteration of the proinflammatory and anti-inflammatory milieu in the lungs of these animals.

There has been considerable recent interest in CLRs as mediators of innate immune signaling following their binding to various microbial ligands (55, 56). These CLRs include Dectin-1, Dectin-2, MCL, and Mincle, which are analyzed in this study. All of these CLRs have been shown to have importance in mediating adequate host response to various fungal infections, including C. albicans and C. glabrata, as well as Aspergillus fumigatus, Blastomyces dermatitidis, Fonsecaea pedrosoi, and Malassezia infections (39, 57–64). However, their role in Pneumocystis infection is lacking, with only a few reports linking one of these CLRs, Dectin-1, with Pneumocystis spp. innate signaling (45, 54). Because of this paucity of information, we attempted to determine whether the CLR Mincle participates in innate immune response and host defense during PCP.

As noted, Mincle mRNA expression was significantly upregulated during PCP. These findings are similar to other reports of increased mRNA expression of Mincle in C. albicans, Malassezia spp., and M. tuberculosis infections, suggesting the importance of this CLR in control of these organisms (20, 23, 44). We further demonstrated that macrophages overexpressing full-length Mincle protein bound P. carinii organisms to greater extents than did parental macrophages expressing basal levels of Mincle, further supporting a role for Mincle in recognizing Pneumocystis in the host lung. To further define the potential PAMPs on Pneumocystis that Mincle may bind, we incubated Mincle-Fc fusion proteins with P. carinii homogenates and purified native P. carinii MSG/gpA, the major surface Ag present on Pneumocystis. Our results demonstrated that Mincle-Fc fusion proteins bind whole organisms, as well as organism-native P. carinii MSG/gpA. We noted the strong avidity on a ligand concentration basis with P. carinii MSG/gpA, the major Ag present on the surface of Pneumocystis.

Recent genetic analysis of and biochemical evidence on Pneumocystis species suggest that although the organism contains genes encoding for the synthesis of N- and O-linked glycan core structure, the fungus lacks genes for synthesizing the longer and more elaborate outer mannose chains seen in other fungi, such as C. albicans (6). These authors suggested that the lack of N-linked mannan on the Pneumocystis cell surface might contribute to host defense avoidance mechanisms (6). Our results may indicate otherwise, because Mincle was shown to indeed bind to low mannose–containing native P. carinii and purified P. carinii MSG/gpA, in a fashion similar to Mincle binding to intact C. albicans cells, which possess abundant N-linked outer chain mannan (65). These findings suggest that Mincle may well recognize multiple-length fungal mannose chains. Our data indicate that Mincle can recognize this important surface component of the organism, supporting a role for these interacting in innate immunity against this fungus.

A particularly interesting finding of this study was that both the wild-type and Mincle^−/− mice demonstrated increases of both proinflammatory (TNF-α, IL-6, IL-1α, and IL-1β) as well as anti-inflammatory (IL-1Ra) cytokine responses during PCP. Of interest were the striking 3-fold increased IL-1Ra levels in the Mincle^−/− mice compared with wild-type P. murina–infected mice. To our knowledge, the association of an increased level of IL-1Ra expression during PCP, as well as significant enhancement of IL-1Ra in the absence of Mincle, has not been previously reported. We further observed relatively preserved survival of the Mincle^−/− mice with PCP, despite impaired clearance of the organism. We postulate that this preserved survival might be related to the balance between proinflammatory and anti-inflammatory factors in the lung during pneumonia. Also of note was the observation that despite the elevated expression of priming cytokines for IL-17, such as IL-6 and IL-1, the levels of IL-17 were significantly reduced during PCP in the absence of Mincle. This is important, because IL-17 has been linked to host response to other fungal infections (45, 47).

Others have shown Dectin-2–related increases in both proinflammatory IL-1α and IL-1β followed by dramatic increases in IL-1Ra anti-inflammatory response in macrophages challenged with C. albicans (37). These authors postulated that the observed differences in host cytokine proinflammatory and anti-inflammatory production might be related to responses to different life stages of the organism (37). In C. albicans infections, Gow and colleagues (66) have noted a feedback loop with an initial burst of proinflammatory cytokines through recognition of PAMPs (β-glucans) by PRRs, including Dectin-1, followed by lessening of inflammation after release of host chitinases and host binding to chitin polymers via the mannose receptor after fungal death and subsequent release of anti-inflammatory IL-10. Furthermore, it has recently been shown that Mincle not only regulates proinflammatory cytokine production upon binding M. tuberculosis TDM, but it also can activate the anti-inflammatory cytokine IL-10, further indicating an essential role for Mincle in modulating host innate immune responses (67).

We hypothesize that because Pneumocystis species are obligate host fungal parasites requiring the lung for proliferation and life cycle progression (6), the organism would better survive with the dampening of robust prolonged inflammatory conditions. Release of proinflammatory cytokines is a consequence of the interactions of host cells with P. carinii β-glucans. The counterregulatory generation of anti-inflammatory IL-1Ra release mediated by Mincle receptors and not previously reported during PCP, may, on the one hand, represent a means to limit lung tissue damage during infection, but, on the other hand, it may also represent an important mechanism through which Pneumocystis species persist in the lung environment by dampening the host responses to infection. Indeed, Garvy et al. (68) recently demonstrated that the trophic life cycle forms of P. murina suppress β-glucan–induced proinflammatory cytokine responses, associated with the β-glucan–rich P. murina cysts wall. It will be interesting in future studies to determine whether Pneumocystis trophic forms might actively modulate anti-inflammatory immune cytokines such as IL-1Ra as a mechanism to persist and propagate in the host.

Additionally, the Mincle^−/− mice also exhibited significant transcriptional increases in other CLRs induced in the absence of Mincle. This includes Dectin-1, another potentially important new finding in Pneumocystis–host interactions. Also note that MCL and Mincle have recently been reported to coregulate the expression of other CLRs during other infectious challenges (21, 69).

The present study defines a previously unrecognized and important role for host CLR Mincle in recognizing and responding to Pneumocystis infection. Mincle was found to bind whole Pneumocystis organisms and Pneumocystis cell wall MSG/gpA, and it is expressed at enhanced levels during infection. Furthermore, in the CD4-depleted mouse PCP model, the absence of Mincle led to significant increases in organism burden, but this was not associated with worsened survival during pneumonia. Additionally, the lack of Mincle was associated with the induction of other fungal CLR expression. Despite enhanced levels of Dectin-1 and other CLRs previously implicated in P. carinii host defense, Pneumocystis organism clearance was impaired in the Mincle^−/− mice. The mechanisms underlying the total lung expression of Dectin-1, Dectin-2, and MCL in the Mincle^−/− mice with PCP are not fully understood and may reflect increased recruitment of macrophages and other effector cells into the heavily infected lungs. Furthermore, we postulate that Mincle participates in the balance between proinflammatory and anti-inflammatory cytokines in the lung. Taken together, these data indicate important roles for Mincle in Pneumocystis recognition and clearance, as well as in the overall inflammatory response to the organism during pneumonia.

We greatly appreciate the assistance of Dr. Yasuhiro Ikeda at the Mayo Clinic in Rochester, MN, for generating the Mincle lentiviral constructs for the RAW cell experiments. Assistance from Stephanie Zimmermann in generation of the Mincle-Fc fusion protein is also gratefully acknowledged.

The authors have no financial conflicts of interest.