We have recently reported that mice deficient in the myeloid Src-family tyrosine kinases Hck, Fgr, and Lyn (Src triple knockout [TKO]) had augmented innate lung clearance of Pneumocystis murina that correlated with a higher ability of alveolar macrophages (AMs) from these mice to kill P. murina. In this article, we show that despite possessing enhanced killing, AMs from naive Src TKO mice did not demonstrate enhanced inflammatory responses to P. murina. We subsequently discovered that both AMs and lungs from P. murina-infected Src TKO mice expressed significantly greater levels of the M2a markers RELM-α and Arg1, and the M2a-associated chemokines CCL17 and CCL22 than did wild-type mice. IL-4 and IL-13, the primary cytokines that promote M2a polarization, were not differentially produced in the lungs between wild-type and Src TKO mice. P. murina infection in Src TKO mice resulted in enhanced lung production of the novel IL-1 family cytokine IL-33. Immunohistochemical analysis of IL-33 in lung tissue revealed localization predominantly in the nucleus of alveolar epithelial cells. We further demonstrate that experimental polarization of naive AMs to M2a resulted in more efficient killing of P. murina compared with untreated AMs, which was further enhanced by the addition of IL-33. Administration of IL-33 to C57BL/6 mice increased lung RELM-α and CCL17 levels, and enhanced clearance of P. murina, despite having no effect on the cellular composition of the lungs. Collectively, these results indicate that M2a AMs are potent effector cells against P. murina. Furthermore, enhancing M2a polarization may be an adjunctive therapy for the treatment of Pneumocystis.
Pneumonia caused by Pneumocystis jiroveci continues to be a significant cause of morbidity and mortality in HIV+ (1). According to the 2009 World Heath Organization Global HIV/AIDS Epidemic Report, 1.4 million individuals are living with AIDS in North America, compared with 1.2 million in 2002 (2). Approximately 55,000 individuals (52,000 in 2002) were diagnosed in the year 2009, and 25,000 deaths (19,000 in 2002) were recorded as a result of HIV infection (2). Although the advent of highly active antiretroviral therapy (HAART) has decreased the overall incidence of Pneumocystis carinii pneumonia (3), the mortality rate of those requiring hospitalization remains high (3). A recent retrospective study in an academic medical center found that overall hospital mortality rate to P. jiroveci pneumonia was 11.6%, and in those patients requiring intensive care, 29.0% (4). An additional study summarized the experience with HIV-associated P. carinii pneumonia over a 21-y period in a single center and found that mortality was 10.1% for the period from 1985 through 1989, 16.9% for the period from 1990 through June 1996, and 9.7% for the period from July 1996 through 2006 (i.e., the HAART era) (5). A major concern voiced by this study is the similar mortality associated with P. jiroveci pneumonia pre- and post-HAART.
It is widely reported that phagocytosis by alveolar macrophages (AMs) is the predominant mechanism of Pneumocystis murina clearance from the lungs (6–8); however, the mechanisms by which macrophages kill P. murina are not completely understood. It is hypothesized that on phagocytosis, the oxidative burst by macrophages, primarily the generation of hydrogen peroxide and superoxide, has a direct cytotoxic effect on P. murina (9, 10). Indeed, studies have shown impaired hydrogen peroxide and superoxide production by AMs from HIV+ individuals with <200 CD4+ T cells/mm3 with and without P. jiroveci infection (11). However, mice with an inability to produce reactive oxygen or nitrogen species are not more susceptible to P. murina infection (12), leaving a role for oxidative elimination in doubt. In a surprising revelation, recent studies have identified significant AM apoptosis during P. murina infection (13) that appears to be mediated by immune cell polyamine and peroxide production (14). Mechanistic studies demonstrated that the observed AM apoptosis was a result of Pneumocystis organisms inducing antizyme inhibitor overexpression in AMs, leading to increased polyamine synthesis and uptake (15).
The primary role of macrophages is to respond to pathogens, but they also manage the adaptive immune response by Ag processing and presentation functions. In addition, macrophages can play a central role in the generation and resolution of inflammation, as well as tissue repair (16). This diverse set of macrophage functions is governed by what type of receptor is activated on the macrophage and, more importantly, the cytokine milieu in which the macrophage exists (17). The original description of macrophage activation has been subsequently termed “classical activation” and involves an activation pathway induced by the cytokines IFN-γ and TNF-α (18). Classically activated macrophages, also called M1 macrophages, are considered highly proinflammatory, produce reactive oxygen species and reactive nitrogen species, and confer defense against multiple pathogens (19). A second type of macrophage is termed the “alternatively activated macrophage,” also called an M2 macrophage, in which cytokines such as IL-4 and IL-13 provide the activation signal. These macrophages are considered to be more associated with a tissue repair/extracellular matrix response and are often called “wound-healing macrophages” (20, 21). The host defense properties of alternatively activated macrophages are somewhat enigmatic, but these cells are clearly required for antihelminthic responses (21). M2 macrophages have now been further divided into individual subsets, M2a, M2b, and M2c, all of which have distinct gene expression profiles (22). M2a macrophages are activated by IL-4/IL-13, M2b macrophages are activated by immune complexes and TLR/IL-1R signaling, and M2c macrophages are activated by IL-10. Tumor-associated macrophages are an additional subset of macrophages that are often considered alternatively activated, although the major characteristic of this macrophage population is the production of IL-10, TGF-β, and multiple growth factors (23). Another distinction between M1 and M2a macrophages is their chemokine repertoire, with M1 macrophages producing high levels of the CXCR3 ligands CXCL9/Mig and CXCL10/IP-10, targeting the recruitment of Th1 and NK cells, and M2a macrophages producing high levels of the CCR3/CCR4 ligands CCL17/TARC and CCL22/MDC, targeting the recruitment of Th2 cells, eosinophils, and basophils (22). In terms of M2 macrophages and lung fungal infections, studies decidedly point to a detrimental role of the development of M2 responses on lung defense. IL-13 overexpressing mice have AMs that display M2 characteristics and fail to clear Cryptococcus neoformans (24). In turn, IFN-γ–deficient mice develop an allergic response to C. neoformans characterized by M2 development (25). In lung infections with Histoplasma capsulatum, absence of CCR2 results in M2 development and progressive infection (26).
We have recently reported that deficiency in the Src-family tyrosine kinases (SFKs) Hck, Fgr, and Lyn (Src triple knockout [TKO] mice) resulted in paradoxically augmented lung clearance of P. murina (27). Mechanisms correlating with clearance were enhanced cytokine and chemokine levels, and increased inflammatory cell recruitment to the lungs, as well as the ability of AMs from Src TKO mice to kill P. murina more efficiently. In this report, we show that Src TKO mice display an enhanced alternative macrophage activation lung environment during early P. murina infection, suggesting that M2a AM polarization enhances innate clearance of P. murina.
Materials and Methods
Male C57BL/6 mice, 6–8 wk of age, were purchased from the National Cancer Institute, National Institutes of Health (Bethesda, MD). Hck/Fgr/Lyn−/− mice (Src TKO mice) originally developed by Dr. Clifford Lowell, University of California at San Francisco (28), were provided by Dr. Shaoguang Li, University of Massachusetts. All animals were housed in a specific pathogen-free, Association for Assessment and Accreditation of Laboratory Animal Care-certified facility and handled according to Public Health Service Office of Laboratory Animal Welfare policies after review by the University of Alabama Institutional Animal Care and Use Committee.
P. murina isolation and inoculation
A preparation of P. murina was prepared as previously described (29). In brief, C.B-17 SCID mice previously inoculated with P. murina were injected with a lethal dose of pentobarbital, and the lungs were aseptically removed and frozen at −80°C in 1 ml PBS. Frozen lungs were homogenized through a 70-μm filter and pelleted at 300 × g for 10 min at 4°C. The pellet was resuspended in 1 ml PBS, and a 1/10 dilution was stained with modified Giemsa stain (Diff-Quik). The number of P. murina cysts was quantified microscopically, and the concentration was adjusted to 2 × 106 cysts/ml. For in vivo challenge, mice were anesthetized with isoflurane and administered 2 × 105 cysts in a volume of 0.1 ml via intratracheal inoculation. Some preparations were also adjusted to 2 × 106 cysts/ml, and 50 μl aliquots were placed into tubes containing 200 μl of 90% FBS supplemented with 10% DMSO and stored at −80°C. Using this storage method, stable P. murina viability, as determined by quantitative real-time PCR, can be maintained for >1 y (30).
AM isolation and stimulation
Naive mice were anesthetized with i.p. ketamine/xylazine and sacrificed by exsanguination. Thereafter, lungs were lavaged through an intratracheal catheter with prewarmed (37°C) calcium- and magnesium-free PBS supplemented with 0.6 mM EDTA. A total of 10 ml was used in each mouse in 0.5-ml increments with a 30-s dwell time. The lavage fluids were pooled and centrifuged at 600 × g for 10 min, and the cells were collected for the coculture assay. A total of 25,000 cells was cytospun onto slides and stained with H&E to ensure that each cell preparation was enriched for macrophages. Cell preparations were generally >98% enriched for AMs. For in vitro stimulation, macrophages (2 × 105 in 100 μl) were cocultured with P. murina (2 × 105 cysts in 100 μl) for 24 h at 37°C, 5% CO2 as described previously (30). Controls include macrophages incubated with medium alone. Supernatants were collected and clarified by centrifugation, followed by analysis for protein levels of 23 cytokines and chemokines using Bio-Plex multiplex suspension cytokine array (Bio-Rad Laboratories), according to the manufacturer’s instructions. The data were analyzed using Bio-Plex Manager software (Bio-Rad Laboratories).
P. murina viability assay
Macrophages (1 × 105 in 100 μl) were left alone or treated with 10 ng/ml IL-4, IL-13, or IFN-γ (all from eBioscience) for 24 h, followed by coculture with P. murina (1 × 103 cysts in 100 μl) for 6 h at 37°C, 5% CO2. In specific experiments, IL-33 (10 ng/ml) was added for an additional 6 h after IL-13; then P. murina was added for 6 h. Controls include P. murina incubated with medium alone or in the presence of each cytokine. The contents of each well were collected and pelleted at 800 × g for 5 min. The supernatants were discarded and total RNA was isolated from the cell pellets using TRIzol reagent (Invitrogen, Carlsbad, CA). Viability of P. murina was analyzed through real-time PCR measurement of rRNA copy number and quantified by using a standard curve of known copy number of P. murina rRNA as previously described (31, 32). This methodology detects viable P. murina organisms as evidenced by the absence of detectable P. murina rRNA in samples subjected to heat inactivation or exposure to trimethoprim/sulfamethoxazole. Percentage killing was defined as previously described (30).
Real-time PCR analysis for alternative activation marker expression in AMs and lung tissue
Mice were infected with P. murina for 3 d followed by either lung lavage or lung tissue collection. Lung lavage cells were adhered to plastic for 1 h at 37°C in 5% CO2 (to enrich for AMs), followed by the removal of nonadherent cells. Total RNA was isolated from enriched AMs or the right lung of infected mice by a single-step method using TRIzol reagent as per the manufacturer’s instructions. Thereafter, RNA was transcribed to cDNA (iScript cDNA synthesis kit; Bio-Rad), and real-time PCR for Retnla (Mm00445109_m1; Applied Biosystems), Arg1 (Mm00475988_m1; Applied Biosystems), and Ym1 (Mm00657889_mH; Applied Biosystems) was performed (iQ Supermix; Bio-Rad). mRNA levels were normalized to GAPDH mRNA levels (primers/probe from Applied Biosystems) using the 2−(ΔΔCt) method.
Cytokine/chemokine, RELM-α, CCL17, CCL22, and IL-33 protein analysis in lung tissue
The left lung was homogenized in PBS supplemented with Complete Mini protease inhibitor tablets (Roche), clarified by centrifugation, and stored at −80°C. Supernatants from lung homogenates were analyzed for protein levels of 23 cytokines and chemokines using Bio-Plex multiplex suspension cytokine array (Bio-Rad Laboratories) according to the manufacturer’s instructions. The data were analyzed using Bio-Plex Manager software (Bio-Rad Laboratories). RELM-α protein levels were quantified in lung homogenates by ELISA (Mouse RELM-α “Super-X” Pre-Coated ELISA Kit; Antigenix America, Huntington Station, NY). CCL17, CCL22, and IL-33 protein levels were quantified in lung homogenates by ELISA (all kits from R&D Systems). In specific experiments, lungs were inflated with optimal cutting temperature medium (Sakura Finetek, Torrance, CA), submerged in optimal cutting temperature medium, and flash frozen on liquid nitrogen, and 7-μm-long sections were cut using a cryostat. Sections were stained with purified goat anti-murine IL-33 (R&D Systems) followed by FITC-conjugated rabbit anti-goat IgG (Invitrogen). Control sections were stained with FITC-conjugated rabbit anti-goat IgG alone.
Treatment of mice with IL-33
Naive C57BL/6 mice were administered 2 μg recombinant murine IL-33 (PeproTech) i.p. in a volume of 200 μl four times, beginning 1 d before P. murina infection, the day of inoculation, and again on days 1 and 2 postchallenge. Control mice received an identical treatment regimen with diluent. Mice were sacrificed on day 3 postchallenge for P. murina lung burden, cytokine and chemokine, and M2a marker analysis. In specific experiments, mice were sacrificed, the lungs lavaged, and cells were stained with Abs against Siglec F, CD11b, and Ly-6G to identify myeloid populations (33). In addition, lavage cells were cytospun onto glass slides and stained with Wright-Giemsa.
Data were analyzed using GraphPad Prism statistical software (GraphPad Software, San Diego, CA). Comparisons between groups when data were normally distributed were made with the two-tailed unpaired Student t test and the two-tailed Mann–Whitney U test when not normally distributed. In real-time PCR analysis comparing wild-type (WT) versus Src TKO samples and in the AM killing assays, the two-tailed paired Student t test was used. Significance was accepted at p < 0.05.
Lack of a heightened inflammatory response to P. murina by Src TKO AMs
In our previous work, we found that mice deficient in the SFKs Hck, Fgr, and Lyn (Src TKO mice) had augmented innate lung clearance of P. murina (27). At 3 and 7 d postchallenge, Src TKO mice displayed increased levels of IL-1β, IL-6, G-CSF, CXCL1/KC, and CCL2/MCP-1 (27). We further observed that AMs from naive Src TKO mice possess a greater capacity to kill P. murina in vitro (27). Because AMs from naive Src TKO mice were more efficient at killing P. murina, we questioned whether their inflammatory reactivity to P. murina was similarly enhanced, and thus a probable cellular source for the observed increase in inflammatory mediators. Surprisingly, naive Src TKO AMs stimulated 24 h with P. murina did not demonstrate enhanced inflammatory cytokine or chemokine production (Fig. 1). In fact, with most mediators, Src TKO AMs appeared to produce lower levels than WT AMs. Thus, although naive Src TKO AMs are more efficient at killing P. murina, these AMs do not respond in a more hyperinflammatory manner.
Enhanced expression of M2a markers in P. murina-infected Src TKO mice
Because AMs from Src TKO mice were able to more efficiently kill P. murina (27) in the absence of increased inflammatory reactivity (Fig. 1), we questioned whether Src TKO AMs were alternatively activated. For this, we collected lavage cells from WT and Src TKO mice challenged with P. murina-for 3 d and adhered them to plastic for 1 h at 37°C to enrich for macrophages. In a separate experimental design, the left and right lungs from WT and Src TKO mice challenged with P. murina for 3 d were collected. Quantitative real-time PCR was performed for a panel of markers associated with M2a development, including Retnla, Arg1, and Ym1 (16). Results in Fig. 2 show that both AMs and the lungs of P. murina-infected Src TKO mice had higher expression of Retnla and Arg1 compared with WT mice. Although the M2a marker Ym1 was not increased in AMs of P. murina-infected Src TKO mice (Fig. 2A), Ym1 was induced at greater levels in the lungs of P. murina-infected Src TKO mice at day 3 (Fig. 2B). We further measured RELM-α protein concentrations by ELISA in lung homogenates collected from WT and Src TKO mice challenged with P. murina for 3 and 7 d, and found >3-fold greater RELM-α concentrations in Src TKO mice at both time points (Fig. 2C). Although we observed RELM-α to be increased in the lungs of naive Src TKO, these levels were significantly enhanced by P. murina infection. Thus, augmented lung clearance of P. murina in Src TKO mice correlates with increased lung expression of M2a markers.
M2a polarization is associated with the production of specific chemokines that target the CCR3 and CCR4 receptors and recruit such cells as eosinophils, basophils, NK cells, and activated T cells (16, 21). We therefore assessed the levels of CCL17 and CCL22 in Src TKO mice to determine whether these M2a-associated chemokines were enhanced in response to P. murina lung challenge. CCL17 (Fig. 2D) was significantly increased in the lungs of Src TKO mice before infection and continued to increase after exposure to P. murina for 3 and 7 d. CCL22 (Fig. 2E) was also significantly increased in the lungs of Src TKO mice before infection, but only marginally increased in response to P. murina infection. The observed increase in CCL17 and CCL22 levels in Src TKO mice was not due to a general increase in CCR3/CCR4 ligands, as CCL11, a well-recognized CCR3 ligand (34), was not produced at increased levels in P. murina-challenged Src TKO mice (Fig. 2F). Thus, myeloid SFK deficiency also results in augmented M2a-associated chemokine production in response to P. murina lung challenge.
Lung IL-4 and IL-13 levels are not enhanced in P. murina-infected Src TKO mice
IL-4 and IL-13 are the dominant cytokines that drive M2a polarization (21) and are conventionally produced by CD4+ Th2 cells (35). However, innate cell sources of IL-4 and IL-13, such as basophils, have been reported (36). Because of the importance of IL-4/IL-13 signaling in alternative macrophage activation, we speculated that enhanced M2a development in P. murina-exposed Src TKO mice was a result of enhanced IL-4 and/or IL-13 expression in the lungs. Surprisingly, IL-4 was nearly undetectable in the lungs of mice challenged with P. murina, but was slightly increased in Src TKO mice at 7 d postchallenge, albeit produced at extremely low levels (Fig. 3A; actual values for day 7 are 1.7 ± 0.35 versus 3.5 ± 0.52 pg/ml for WT and Src TKO, respectively; p = 0.007). In contrast, IL-13 was induced by P. murina at much greater levels than IL-4, although there was no significant difference in IL-13 levels between WT and Src TKO mice at either 3 or 7 d postchallenge (Fig. 3B). Thus, SFK deficiency does not result in a significant increase in whole-lung levels of IL-4 and IL-13 after P. murina lung challenge that would explain the observed enhancement in M2a polarization in Src TKO mice.
SFK deficiency results in greater lung IL-33 levels after P. murina challenge
Although the major determinants of M2 macrophage development are IL-4 and IL-13, other mediators, such as IL-3 (37, 38), IL-21 (39), and galectin-3 (40), can act as synergistic factors to drive macrophage polarization toward M2a. Because IL-4 and IL-13 were not differentially expressed between P. murina-exposed WT and Src TKO mice, we hypothesized that such a synergism factor may be expressed at greater levels in Src TKO mice. Recently, the IL-1 family member cytokine IL-1F11, also known as IL-33, has been shown to synergize with both IL-4 and IL-13 in the lungs to promote M2a polarization of AMs (41). Because the initial report of IL-33 described strong expression in epithelial cells of the lung (42), we questioned whether P. murina induced IL-33 expression and whether it was modulated in Src TKO mice. Results in Fig. 4A show that P. murina induces IL-33 production in the lungs at 3 and 7 d postchallenge, which was significantly enhanced in the absence of SFKs. IL-33 is considered both a cytokine and an intracellular NF with transcriptional regulatory properties (43). Immunohistochemical staining of lung tissue from WT and Src TKO mice 7 d after P. murina challenge demonstrated that IL-33 was highly localized in the nucleus of alveolar epithelial cells (Fig. 4B), with higher staining observed in Src TKO mice (Fig. 4B, middle image). Control slides demonstrated no nuclear IL-33 staining (Fig. 4B, right image). In addition, myeloid cells observed in the lung sections did not demonstrate any nuclear localization of IL-33 (Fig. 4B, middle image, arrows). Thus, although IL-4 and IL-13 were not increased in P. murina-exposed Src TKO mice, increased levels of the pro-M2a cytokine IL-33 were observed, suggesting a role for IL-33 in promoting M2a polarization and augmented AM effector function against P. murina.
Experimental polarization of AMs to M2a results in higher P. murina killing
As we have previously reported that AMs from Src TKO mice kill P. murina more efficiently than WT AMs (27), and report here that Src TKO AMs are polarized toward an M2a phenotype (Fig. 2), we sought additional evidence that M2a AMs were better effector cells against P. murina. Although IL-4 and IL-13 were not expressed in the lungs at greater levels in P. murina-exposed Src TKO mice, these cytokines are the best described factors driving M2a polarization (21) and, in the case of IL-13, produced at significant levels in the lungs in response to P. murina (Fig. 3). To this end, we determined the effects of IL-13 on AM-mediated killing of P. murina. We chose IL-13 because it was produced in the lungs of both P. murina-infected WT and Src TKO mice at much greater levels than IL-4 (Fig. 3B). Results in Fig. 5 show that pretreating AMs with IL-13 for 24 h significantly increased killing of P. murina by 2-fold compared with untreated AMs. In contrast, IFN-γ treatment did not modulate baseline AM killing of P. murina (data not shown). Because IL-33 was identified as a putative M2a synergism factor in Src TKO lungs, we next determined whether IL-33 modulated the increased killing of P. murina by AMs pretreated with IL-13. Indeed, addition of IL-33 to IL-13–treated AMs resulted in an additional increase in M2a AM killing of P. murina (Fig. 5). IL-33 treatment of naive AMs had no effect on killing (data not shown). Thus, AMs experimentally polarized toward M2a with IL-13 kill P. murina more efficiently than those polarized toward M1, and IL-33 can augment M2a activity against P. murina.
IL-33 augments innate lung clearance of P. murina
To further determine the extent by which IL-33 could modulate innate immunity to P. murina, we administered IL-33 i.p. to mice and analyzed P. murina lung burden, M2a marker expression, and cytokine and chemokine levels. Results in Fig. 6A show that i.p. administration of IL-33 to mice led to a significant reduction in P. murina lung burden (Fig. 6A). IL-33 administration also significantly increased the levels of RELM-α (Fig. 6B) and CCL17 (Fig. 6C) in the lungs. IL-33 administration did not increase the levels of IL-4 and IL-13; however, mice that received IL-33 did have a significant increase in IL-5 (Fig. 6D). Levels of proinflammatory mediators such as IL-1β, IL-6, and CCL2 were not affected by the administration of IL-33 (data not shown). We further did not observe an effect of IL-33 administration on lung cell populations after P. murina exposure as Siglec F+, CD11b−, and Ly-6G− macrophages (33) were the dominant cell types in lung lavage fluid from both infected untreated and infected IL-33–treated mice (Fig. 6E). Wright-Giemsa staining of lung lavage cells from both infected untreated and infected IL-33–treated mice indicated nuclear and morphological characteristics of macrophages (Fig. 6F). Thus, IL-33 can augment innate clearance of P. murina and is an important contributor to optimal innate lung immune responses to P. murina.
Previous work in our laboratory has demonstrated that recognition of P. murina by AMs leads to the activation of the SFKs Hck and Lyn (27), providing us an initial hypothesis that SFKs were essential for AM function and host defense against P. murina. Unexpectedly, however, mice triple deficient in the SFKs Hck, Fgr, and Lyn (Src TKO mice) had augmented innate lung clearance of P. murina that correlated with both an increased lung inflammatory response and a greater ability of Src TKO AMs to kill P. murina in vitro (27). Because AMs are critical effector cells required for P. murina lung clearance (7), and because Src TKO AMs killed P. murina more efficiently (27), we hypothesized that Src TKO AMs would be hyperresponsive to P. murina stimulation resulting in increased inflammatory mediator production. Surprisingly, the opposite was true, as Src TKO AMs produced lower amounts of multiple mediators known to be produced by AMs in response to P. murina, including G-CSF, IL-1β, and IL-6. These data provided the initial evidence that SFK deficiency in AMs resulted in a different type of macrophage activation.
Classically activated (M1) macrophages often act in a more proinflammatory manner, producing such cytokines as TNF-α, IL-1, and IL-12, chemokines such as CXCL9 and CXCL10, and reactive oxygen species (16). Because Src TKO AMs were not hyperinflammatory, yet better effector cells against P. murina, we speculated that they may be alternatively activated. Indeed, analysis of AMs and lungs from P. murina-infected Src TKO mice demonstrated increased mRNA expression of the M2a markers Retnla, Arg1, and to a lesser extent, Ym1. Moreover, signature chemokines associated with M2a polarization, CCL17 and CCL22, were also found to be increased in P. murina-infected Src TKO mice. Because RELM-α, CCL17, and CCL22 are secreted factors, we subsequently showed that the protein levels of each were also significantly increased in the lungs of P. murina-infected Src TKO mice at 3 and 7 d postchallenge, time points at which P. murina lung burden is significantly lower in these mice compared with WT control animals (27). In addition, we observed that baseline levels of RELM-α, CCL17, and CCL22 were increased in the lungs of naive Src TKO mice. As we have previously reported that AMs from naive Src TKO kill P. murina more efficiently than WT AMs (27), these results suggest that AMs from naive Src TKO maintain an M2a phenotype.
Recent effort has focused on identifying specific functional roles for various M2a markers. RELM-α is a cysteine-rich secreted protein originally identified in the lung (44) and subsequently observed to be expressed by macrophages, eosinophils, epithelial cells, and endothelial cells (44, 45). RELM-α stimulates production of MCP-1 (46) and IL-6 (46), both of which we have previously reported to be increased in the lungs of P. murina-infected Src TKO mice (27). RELM-α can also bind to F4/80+ cells (45), which we also reported to be recruited in greater numbers to the lungs of P. murina-infected Src TKO mice (27). Although RELM-α is considered a product of Th2 activation, models of Nippostrongylus brasiliensis GI infection (47) and Schistosoma mansoni lung infection (45) in RELM-α–deficient mice define a paradoxical role for RELM-α in suppressing Th2 responses. Finally, in a murine model of Pneumocystis-associated pulmonary hypertension, although a Th2 bias was not observed, increases in RELM-α correlated with an increase in perivascular fibrosis (48). Arginase-1 (Arg-1) is a cytosolic enzyme expressed in macrophages on M2 polarization that hydrolyzes arginine to urea and ornithine (49). Similar to RELM-α, mice with Arg-1 deficiency in macrophages have exacerbated Th2 responses during S. mansoni lung infection (50). Interestingly, P. carinii infection in steroid-treated rats results in increased ornithine decarboxylase expression leading to increased polyamine production (14, 15). These results suggest that increased ornithine decarboxylase levels during P. carinii infection are a result of increased expression of Arg-1 and subsequent increased production of ornithine. CCL17 and CCL22 are recognized as Th2 chemokines and are often produced in diseases with strong Th2 polarization, including atopic dermatitis, asthma, and pulmonary fibrosis (51, 52). The receptors for CCL17 and CCL22 are CCR3 and CCR4, which are abundantly expressed on eosinophils, basophils, and activated Th2 cells (53). Th2 and M2a development in parasitic infections promote the recruitment and activation of eosinophils (21, 36). Intriguingly, it has been shown that eosinophils are the predominant leukocyte population found in the lungs of immunocompetent mice infected with P. murina for 2 wk (54). Moreover, a recent study reported that microarray analysis of AMs from P. murina-infected IFN-αR–deficient mice demonstrated reduced CCL22 expression, which correlated with impaired eosinophil recruitment in these mice compared with P. murina-infected control mice (55).
In models of parasitic infections, the immune response decidedly relies on the generation of Th2 responses and the production of RELM-α, Arg-1, and Ym-1 for regulatory and protective mechanisms. Although not universal, in many models of fungal infection, generation of Th2 responses leads to nonprotective defense mechanisms. Specifically, the generation of Th2 responses and M2a polarization in Cryptococcus neoformans (24) and Histoplasma capsulatum (26) renders mice more susceptible to infection. Mice that are biased toward Th2 responsiveness are more susceptible to disseminated (56) and mucosal infection (57) with Candida albicans. However, promotion of M2a polarization in gastrointestinal Candida albicans infection has been shown to limit fungal colonization (58). Finally, in lung infection with Aspergillus fumigatus, generation of Th1/regulatory T cell responses over Th2 responses favor resistance to infection (59), suggesting that M2a polarization would complicate clearance of this organism. In contrast, immunocompetent mice infected with P. murina have a 4:1 ratio of CD4 T cells producing IL-4 versus IFN-γ within the first 7 d of infection that remain at a 2:1 ratio through 3 wk of infection (60). Moreover, IL-5 and IL-13 are detected in lung lavage fluid earlier and in greater amounts than IFN-γ in immunocompetent mice infected with P. murina (54). These results provide additional support for P. murina infection inducing a lung environment conducive for M2a development.
Because the Th2 cytokines IL-4 and IL-13 are required for M2a development (21), we naturally expected one or both of these mediators to be significantly increased in P. murina-infected Src TKO mice. However, this was not the case, as IL-4 was barely detectable in the lungs of P. murina-infected WT and Src TKO. IL-13 was produced at much greater levels but was not different between WT and Src TKO mice. Preliminary analysis of CD4+ T cells from the lungs and spleens of WT and Src TKO mice infected with P. murina for 7 d indicated increases in IL-4, IL-5, IL-13, IL-10, and IL-17A, but not IFN-γ production after stimulation with ConA or anti-CD3 Abs (unpublished observations). Therefore, we cannot exclude the possibility that skewing of Th responses in Src TKO mice may be playing a role in enhanced M2a development. However, because IL-10 and IL-17A were also produced at greater levels, it is possible the SFK deficiency results in an overall heightened T cell response. Nevertheless, without observing an increase in whole-lung IL-4 and IL-13 levels in P. murina-infected Src TKO mice, it is difficult for us to attribute increased M2a development in these mice solely to a partially skewed Th2 response. Recent focus on M2a development has uncovered additional Th2-associated cytokines that may synergistically act with IL-4 and IL-13 to drive M2a development. The newest member of this group is IL-33, an IL-1 family cytokine predominantly produced by epithelial and endothelial cells, and highly expressed in such tissues as stomach, skin, and lung (42). In humans, IL-33 is constitutively expressed in lung smooth muscle cells and epithelial cells forming the bronchus and small airways (42). Because P. murina is a potent activator of lung epithelial cells (61, 62) and IL-33 is highly associated with lung tissue, we hypothesized that IL-33 may be a factor in driving M2a development during P. murina lung infection. In turn, we found that IL-33 was highly induced in the lungs by P. murina infection in WT mice and was further enhanced in Src TKO mice. In fact, IL-33 is produced at increasing levels in the lungs of normal mice throughout the duration of P. murina infection (unpublished observations). IL-33 expression in the lungs during P. murina infection was alveolar epithelial cell associated and highly localized in the nucleus of these cells. IL-33 has recently been shown to enhance IL-13–mediated M2a polarization in both bone marrow macrophages and AMs (41). Because IL-13 is produced in copious amounts (relative to IL-4) in response to P. murina, we postulate that increased IL-33 in the presence of IL-13 drives the heightened M2a responses observed in Src TKO mice.
In contrast with M1 macrophages, M2a macrophages are not thought to possess efficient antimicrobial activity, but rather function more in resolution of inflammation and wound healing. Indeed, the size of most parasitic organisms hampers the ability of macrophages to internalize them. However, a component of M2a polarization is the upregulation of innate pattern recognition receptors, such as Dectin-1, mannose receptor, and scavenger receptors (21, 36), with known antimicrobial functions. In turn, interactions between Dectin-1 (30), MR (8), and SR-A (63) and P. murina have been identified, suggesting that M2a macrophages function in P. murina host defense. We now have additional evidence to support this hypothesis by showing that primary AMs pretreated with IL-13 have superior killing activity against P. murina when compared with untreated or IFN-γ–treated (M1) AMs. However, the phagocytic/killing capability of M2a/IL-4–treated macrophages is controversial. IL-4 has recently been shown to negatively impact macrophage phagocytosis of Gram-negative bacteria and particulate Ags such as zymosan, despite increasing the expression of specific pattern recognition receptors (64). In contrast, other studies have shown that IL-4–treated macrophages have increased killing activity against the parasitic organism Leishmania major (65). We have previously reported that killing of P. murina by untreated AMs is dependent on phagocytosis, yet independent of reactive oxygen species and reactive nitrogen species (30). A goal of future studies will be to determine which AM killing mechanisms against P. murina are enhanced by IL-4/IL-13.
Another finding in this study is that IL-33 further enhanced killing of P. murina by IL-13–treated (M2a) AMs. To our knowledge, this is the first demonstration of IL-33 augmenting macrophage antimicrobial activity. Because of the heightened effector function afforded to AMs by IL-13 + IL-33 treatment, we postulated that treating mice with IL-33 would enhance lung clearance of P. murina. Indeed, mice treated with IL-33 had a 2-fold decrease in P. murina lung burden by 3 d postinfection. A recent study has shown that naive mice administered IL-33 intranasally for 6 consecutive days (4 μg each day) had increased IL-5, IL-13, CCL11, and CCL17 levels (41). Although we did observe increases in the levels of CCL17 and IL-5 in our study, we did not observe IL-33–associated increases in IL-13 or CCL11. Differences in our study include administering a 2-fold lower concentration of IL-33, IL-33 administered i.p. rather than intranasally, administering IL-33 to mice in the presence of an active lung infection, and a shorter experimental time course. However, it was abundantly clear that IL-33 was an effective inducer of RELM-α, which was 3-fold greater in IL-33–treated mice. Here again, as observed in Src TKO mice, the presence of increased RELM-α and CCL17 correlated with enhanced clearance of P. murina from the lungs, suggesting that IL-33, RELM-α, and CCL17 are critical mediators of innate immune-mediated P. murina clearance. Intriguingly, IL-33 treatment did not affect the type or number of cells in the lungs during P. murina infection, as flow cytometric analysis of lung lavage cells from untreated versus IL-33–treated mice demonstrated macrophages as the predominant population. We hypothesize that lack of observing differences in cell recruitment or cell type between these mice may be a result of the relatively short (3 d) time course of treatment coupled with administering IL-33 to mice during an active P. murina infection. We further hypothesize that a major function of IL-33 administered during P. murina infection is most likely the activation of macrophages rather than cell recruitment. Evidence to support this includes the observation of increased RELM-α and CCL17 levels in the lungs of IL-33–treated mice (Fig. 6B, 6C) coupled with a predominant macrophage population in these mice (Fig. 6E, 6F). Moreover, in a pilot assessment, we observed that macrophages isolated from the lungs of IL-33–treated mice infected with P. murina for 3 d demonstrated increased mRNA expression of the scavenger receptor MARCO (unpublished observations). A better of understanding for the role of IL-33 in P. murina lung infection will be pursued in mice deficient in the IL-33R ST2 in future studies. In addition, we will actively pursue the roles of RELM-α and CCL17 in lung host defense against P. murina.
M2a polarization in several lung fungal infections results in a compromised ability to eliminate these organisms from the lungs (24, 26); therefore, the identification and correlation of M2a development with enhanced lung clearance of P. murina (in both Src TKO and IL-33–treated mice) is a unique observation for a fungal pathogen. Furthermore, our study is the first to show IL-33 production in vivo in response to a fungal pathogen, which we hypothesize is critical for M2a development during P. murina infection. The lung M2a response to P. murina involves three additional mediators, RELM-α, CCL17/TARC, and CCL22/MDC, and the role of each in P. murina lung defense are currently being elucidated. In summary, our results indicate that M2a AMs are potent effector cells against P. murina. Furthermore, enhancing M2a polarization, potentially via IL-33, may be an adjunctive therapy for the treatment of Pneumocystis infection in the setting of immunosuppression or immunodeficiency.
We thank Dr. Floyd Wormley (University of Texas–San Antonio) and Dr. Hubert Tse (University of Alabama at Birmingham) for helpful discussions and critical reading of the manuscript.
This work was supported by Public Health Service Grants HL080317, AI068917, and HL096702 (to C.S.).
Abbreviations used in this article:
alveolar macrophage, HAART, highly active antiretroviral therapy
Src-family tyrosine kinase
The authors have no financial conflicts of interest.