Differential Macrophage Polarization from Pneumocystis in Immunocompetent and Immunosuppressed Hosts: Potential Adjunctive Therapy during Pneumonia

ABSTRACT

We explored differential polarization of macrophages during infection using a rat model of Pneumocystis pneumonia. We observed enhanced pulmonary M1 macrophage polarization in immunosuppressed (IS) hosts, but an M2 predominant response in immunocompetent (IC) hosts following Pneumocystis carinii challenge. Increased inflammation and inducible nitric oxide synthase (iNOS) levels characterized the M1 response. However, macrophage ability to produce nitric oxide was defective. In contrast, the lungs of IC animals revealed a prominent M2 gene signature, and these macrophages effectively elicited an oxidative burst associated with clearance of Pneumocystis. In addition, during P. carinii infection the expression of Dectin-1, a critical receptor for recognition and clearance of P. carinii, was upregulated in macrophages of IC animals but suppressed in IS animals. In the absence of an appropriate cytokine milieu for M2 differentiation, Pneumocystis induced an M1 response both in vitro and in vivo. The M1 response induced by P. carinii was plastic in nature and reversible with appropriate cytokine stimuli. Finally, we tested whether macrophage polarization can be modulated in vivo and used to help manage the pathogenesis of Pneumocystis pneumonia by adoptive transfer. Treatment with both M1 and M2 cells significantly improved survival of P. carinii-infected IS hosts. However, M2 treatment provided the best outcomes with efficient clearance of P. carinii and reduced inflammation.

INTRODUCTION

Pneumocystis jirovecii is an opportunistic fungal pathogen and a leading cause of morbidity and mortality in immunosuppressed individuals. Although the advent of highly active antiretroviral therapy has decreased the overall incidence of Pneumocystis pneumonia (PcP), the mortality rate of those requiring hospitalization remains quite high (1). Clinical outcome of Pneumocystis infection is highly dependent on adaptive host immune responses. Although host recognition and inflammatory responses are necessary for the clearance of Pneumocystis, they also strongly contribute to lung injury. Indeed, respiratory failure and death from Pneumocystis pneumonia are more strongly linked to lung inflammation than to direct toxic effects of Pneumocystis organisms themselves (2).

Although control of Pneumocystis depends on development of adaptive CD4 cell-driven immunity, alveolar macrophages are major effector cells that mediate clearance of the microorganism from the lung (3). Macrophages recognize hazardous signals through pattern recognition receptors and, consequently, release cytokines and chemokines to mediate their effector functions (4). It is necessary to keep such responses in balance so that the proinflammatory responses to innocuous antigens are adequately suppressed, while effective immune responses to pathogenic microorganisms are not compromised (5). Macrophages accomplish this by exhibiting plasticity, which has led to their broad classification as either classically activated M1 macrophages or alternatively activated M2 macrophages (5–7). It has been widely accepted that M1-polarized macrophages are effective fungicidal cells (8). In contrast, alternatively activated M2 macrophages exhibit anti-inflammatory function. The inflammation and antimicrobial properties of M1 macrophages are associated with high expression of the enzyme inducible nitric oxide synthase (iNOS) and with the secretion of proinflammatory cytokines such as tumor necrosis factor alpha (TNF-α), interleukin-6 (IL-6), and IL-1β (9). Alternatively, M2 macrophages display increased expression of the enzyme arginase and other proteins such as FIZZ1, MRC1, PPARγ, and the immunosuppressive cytokine IL-10, with little or no secretion of proinflammatory cytokines (10, 11). STAT signaling serves as a central pathway that controls macrophage M1-M2 polarization. Macrophages can be polarized to M2 phenotype through IL-4/IL-13/IL-4Rα and via activating STAT6 signaling or through IL-10/IL-10R and by activating STAT3 signaling. Similarly, macrophages can be polarized to M1 phenotype through gamma interferon (IFN-γ)-IFN-γR and by activating STAT1 signaling (12). Upon activation, STAT molecules are phosphorylated, form homodimers, and translocate to the nucleus where they promote transcription of STAT-targeted genes. Relatively little is known about the roles of STAT molecules during antifungal immunity. However, STAT1 signaling has been shown to be essential for protection against Cryptococcus infections and during chronic mucocutaneous candidiasis, coccidioidomycosis, and histoplasmosis (12–14).

Several in vitro and in vivo studies have described the regulatory role of macrophages in suppressing inflammation in various disease models, such as multiple sclerosis, kidney disease, and spinal cord injury (15–18). More specifically, during infections it is well accepted that M2 macrophages act to ameliorate disease severity and protect the host from detrimental effects of an excessive inflammatory response (19–22). Pneumocystis pneumonia is best treated using antibiotics such as trimethoprim-sulfamethoxazole (23). The disease is severe in AIDS and other immunosuppressed patients and corticosteroids are often used as adjunctive therapy to suppress deleterious lung inflammation during severe lung injury. The outcome of treatments depends on the severity of illness and PcP continues to have significant morbidity and mortality. Hence, new therapeutic strategies to improve the outcome of patients are clearly needed. In that light, immune modulation has been viewed as a potential adjunctive therapy to treat PcP (24). The potential cellular targets for immune modulation to treat PcP include CD4 T cells, CD8 T cells, regulatory T cells, neutrophils, macrophages, and epithelial cells (24, 25). Macrophages offer a novel approach to cell therapy in the treatment of infection, especially in the immunosuppressed hosts, but require additional investigation (26–31). There are relatively few reports concerning macrophage polarization and the resulting antifungal and anti-inflammatory activity as a potential treatment approach for fungal infections (12, 32, 33).

In the present study, we examined the lung immune polarization kinetics of macrophages using a rat model of chronic Pneumocystis pneumonia. We further present a potential therapeutic model using adoptive transfer of differentially polarized macrophages and demonstrate ways to alter their polarization states to improve the immunopathogenesis associated with Pneumocystis infection.

RESULTS

Pneumocystis carinii infection in immunosuppressed host enhances M1 pulmonary macrophage polarization in vivo.

We first evaluated the kinetics of macrophage polarization in vivo during P. carinii infection. To accomplish this, we used a rat model of Pneumocystis pneumonia (PcP). After 8 weeks of immune suppression, Pneumocystis pneumonia was well established, and bronchoalveolar lavage fluid (BALF) samples were serially collected and alveolar macrophages isolated for analysis of the polarization markers iNOS (M1) and Arg-1 (M2). Over the course of infection, we observed a steady increase in the iNOS mRNA expression after more than 4 weeks of infection with further progression as the P. carinii infection became more established (Fig. 1A). The levels of M2 macrophage polarization marker, Arg-1, were slightly increased, although not as much as iNOS compared to uninfected controls (Fig. 1B). Concurrently, the P. carinii burden was quantified using real-time PCR targeting the P. carinii Prt1 gene in the lung tissues (Fig. 1D). The increase in P. carinii burden after 4 weeks correlated with the increase in M1 response. Notably, the ratio of iNOS to Arg-1 levels steadily increased following P. carinii infection, supporting pulmonary M1 macrophage polarization during P. carinii infection in vivo (Fig. 1C). We further observed that progressively increasing P. carinii burdens and M1 responses were associated with severe lung inflammation, based on the presence of perivascular and alveolar aggregates and exudate (Fig. 1F and G). We further evaluated the IFN-γ and IL-4/IL-13 kinetics to determine the expression of the M1 and M2 polarizing cytokines in the lung. Correlating with the increase in M1 response, the release of IFN-γ, an M1 polarizing cytokine, followed the same trend as the iNOS expression in the BALF of P. carinii-infected animals (Fig. 1E). In addition, flow cytometry analyses of control and P. carinii-infected BALF recovered from animals after 8 weeks of infection revealed a significant decrease in the alveolar macrophage numbers. However, there were significant increases in total BALF cell numbers and cell numbers of other leukocytes reflecting enhanced inflammatory cell influx during PcP (Fig. 1I and see Fig. S1 in the supplemental material). Together, these data demonstrate that during P. carinii infection in the immunosuppressed host, >4 weeks after P. carinii infection, the lung exhibits predominantly a classically M1 proinflammatory phenotype rather than an alternatively activated M2 macrophage response.

FIG 1.

P. carinii infection enhances M1 alveolar macrophage polarization in vivo in immunosuppressed host. Female CD IGS rats were immunosuppressed and inoculated with P. carinii organisms (10⁶ inoculum/rat) intratracheally. Alveolar macrophage mRNA was harvested from BALF at the indicated times and analyzed by qPCR for mRNA expression of iNOS (A) and Arg-1 (B), marker genes for M1 and M2 polarization, respectively. (C) Kinetics of M1 (iNOS) versus M2 (Arg-1) response. The mRNA expression was normalized to GAPDH. (D) The P. carinii burden was assessed in the lung tissues at the indicated time points by real-time PCR. (E) IFN-γ, IL-4 (undetectable), and IL-13 (undetectable) secretion was measured by ELISA in the first 5 ml of BALF. The data shown in panels A to D are expressed as means ± the SD of n ≥ 4 animals per group that were age and sex matched and are representative of three independent experiments. In panels A to E, the asterisk (*) indicates P < 0.05. At the indicated time points, P. carinii-infected and control lungs were formalin-fixed, paraffin-embedded and stained with H&E. (F) Histology results for P. carinii-infected lung tissues are shown as representative images of H&E-stained lung sections from n ≥ 4 animals per group and are representative of three independent experiments. (G) Inflammation was scored blindly from n ≥ 4 animals per group and analyzed using ImageJ software as described in Materials and Methods. The scores were representative of three independent experiments, with the bars representing medians. Inflammatory cell influx was analyzed by identifying the leukocytes in the BALF by flow cytometry in the control versus P. carinii-infected rats 8 weeks postinfection from n ≥ 5 rats per group. (H and I) Average of total cell numbers (H) and determinations of alveolar macrophages, neutrophils, eosinophils and lymphocytes (I) in the BALF of control versus P. carinii-infected rats at 8 weeks after infection. The data in panels H and I are means ± the SD of five or more rats per group and are representative of at least three independent experiments. In panels H and I, the asterisk (*) indicates P < 0.05.

P. carinii infection in immunocompetent hosts results in enhanced M2 alveolar macrophage polarization in vivo.

We next assessed macrophage polarization in vivo during P. carinii infection in immunocompetent animals to better understand the pathways used during natural clearance of the pathogen. We infected immunocompetent animals and monitored them for a period of 30 days. The P. carinii burden, measured by the copy number of P. carinii PRT1 mRNA, peaked at about 7 days and thereafter decreased until the organisms were almost completely eliminated by 30 days (Fig. 2A). We further determined the expression of iNOS (M1 marker) and Arg-1 (M2 marker) in the alveolar macrophage isolated from BALF, as well as the level of expression of IFN-γ and IL-4/IL-13, the M1- and M2 driving cytokines in the lung. Both iNOS and Arg-1 expression were significantly increased in the P. carinii-infected macrophages compared to the uninfected controls at 7 days and were decreased thereafter to almost basal levels by 30 days. This correlated well with the decrease in P. carinii burden observed after 7 days of infection (Fig. 2A to C). However, the level of expression of Arg-1 was much greater than that of iNOS in the alveolar macrophages of the P. carinii-infected animals compared to uninfected controls (Fig. 2B and C). Furthermore, although both IFN-γ and IL-4/IL-13 followed similar trends observed for iNOS and Arg-1 in the BALF, we observed nearly 2- to 4-fold-greater M2 driving cytokine (IL-4/IL-13) than IFN-γ (Fig. 2B to F). The IFN-γ secretion began to decrease after 7 days, whereas the IL-4 and IL-13 production persisted a while longer in the M1 and M2 lung macrophage environment (Fig. 2D and F). These data demonstrate that, in contrast to that observed in immunosuppressed animals, in the immunocompetent animals, there was a predominance of M2 greater than M1 responses during P. carinii infection.

FIG 2.

P. carinii infection in immunocompetent hosts results in enhanced M2 alveolar macrophage polarization in vivo. Female CD IGS rats were inoculated with P. carinii organisms (10⁶ inoculum/rat) intratracheally. (A) The P. carinii burden was assessed in the lung tissues at the indicated time points by real-time PCR. In addition, alveolar macrophage mRNA was harvested from BALF at the indicated time points and analyzed by qPCR for the mRNA expression of iNOS (B) and Arg-1 (C), marker genes for M1 and M2 polarization, respectively. The mRNA expression was normalized to GAPDH. (D to F) IFN-γ, IL-4, and IL-13 secretion was measured by ELISA in the first 5 ml of BALF. The data shown in panels A to F are expressed as means ± the SD of n = 4 animals per group that were age and sex matched and are representative of three independent experiments. In panels A to E, the asterisk (*) indicates P < 0.05.

Functional comparison of macrophage polarization in immunocompetent and immunosuppressed host during P. carinii infection.

We further evaluated various aspects of M1 and M2 macrophage activity associated with complete elimination of P. carinii organisms and minimal symptoms in immunocompetent hosts but the failure to clear P. carinii in immunosuppressed hosts despite enhanced M1 responses during P. carinii infection. To accomplish this, we compared the 7-day time point in the immunocompetent rats and the 8-week time point in the immunosuppressed rats. These represented time points correlating with the highest P. carinii burdens and evaluated macrophage responses at these times of greatest P. carinii burdens. Interestingly, through Western blot analysis of alveolar macrophages isolated from BALF, we observed that the expression of iNOS protein was increased in both the immunocompetent and the immunosuppressed host infected with P. carinii compared to their respective controls (Fig. 3). In contrast, the expression of ARG-1 protein was increased in the immunocompetent P. carinii-infected animals but decreased in the immunosuppressed P. carinii-infected animals compared to their controls (Fig. 3A). We analyzed the expression of several other markers using real-time PCR of alveolar macrophage mRNA at these time points to confirm whether similar trends were observed with other M1 and M2 genes. This was indeed the case, with the levels of M1 cytokines such as TNF-α and IL-6 being significantly elevated in both the immunocompetent and immunosuppressed animals, whereas the IL-1β level was increased only in the immunosuppressed hosts (Fig. 3B and see Fig. S3 in the supplemental material). Similarly, the expression of M2 genes, with the exception of Fizz1, followed the same trend as that of ARG-1, with the mRNA expression of M2 cytokine IL-10 and the M2 marker genes Pparγ and Mrc1 being significantly increased in the immunocompetent host infected with P. carinii but not in the immunosuppressed host (Fig. 3B and C and see Fig. S3 in the supplemental material). These data suggest that M2 responses are defective in the immunosuppressed host, despite having intact M1 responses.

FIG 3.

Functional comparison of macrophage polarization between the immunocompetent and immunosuppressed host during P. carinii infection. Female CD IGS rats that were either immunosuppressed or left immunocompetent were inoculated with P. carinii (10⁶ inoculum/rat) intratracheally. On day 7 for immunocompetent animals and after 8 weeks for the immunosuppressed animals, when there was the greatest respective P. carinii accumulation, the two groups were compared. Uninfected animals were used as controls. (A) Lysates were prepared from pooled alveolar macrophages harvested from BALF of four rats and analyzed by Western blot for protein expression of iNOS and ARG-1. β-Actin was used as a loading control, and the the data are representative of two independent experiments. (B and C) Alveolar macrophage mRNA was harvested from BALF and analyzed by qPCR for mRNA expression of M1 genes (IL-1β, IL-6, and TNF-α) (B) or M2 genes (Fizz1, Pparγ, Mrc1, and IL-10) (C). The mRNA expression levels were normalized to GAPDH. (D) Alveolar macrophages harvested from BALF were restimulated with P. carinii in vitro for 4 h after a 90-min adherence and the removal of nonadherent cells. Nitrite concentrations as a measure of NO (nitrosative burst) and H₂O₂ (oxidative burst) release into the cell culture supernatant were measured using calorimetric assays according to the manufacturer's instructions. Data shown in panels A to F are expressed as means ± the SD of n = 4 animals per group that were age and sex matched and are representative of three independent experiments. In panels A to D, the asterisk (*) indicates P < 0.05.

We further sought to determine whether the increase in iNOS levels in the M1 cells translated functionally into NO production that might mediate the killing of P. carinii organisms by forming nitrogen intermediates through a nitrosative burst. Of note, the M1 polarized macrophages from immunosuppressed animals were still defective in eliminating P. carinii despite their induction of iNOS. To examine this, we assessed the release of hydrogen peroxide and nitrite as measure of reactive oxidants and NO⁻ release, respectively, from P. carinii-infected and uninfected alveolar macrophages derived from immunocompetent and immunosuppressed animals that were restimulated with P. carinii organisms in vitro for 4 h. Interestingly, the immunocompetent cells that exhibited enhanced M2 response predominantly exhibited an oxidative burst rather than a nitrosative burst, whereas immunosuppressed cells were defective in the production of both oxidative and nitrosative products (Fig. 3D). These data demonstrate that the enhanced M2 response in immunocompetent animals correlates with P. carinii clearing activity likely by forming reactive oxygen intermediates (oxidative burst), while the immunosuppressed cells were defective in production of both nitrogen and oxygen intermediates despite intact iNOS expression.

Macrophage polarization during Pneumocystis infection is plastic in nature.

We next sought to determine whether macrophage polarity could be manipulated during P. carinii infection reflecting dynamic plasticity of macrophages. This was first addressed in vitro by challenging macrophages with P. carinii organisms to assess whether macrophage polarization could be achieved by P. carinii stimulation independent of the lung microenvironment. For these studies, freshly isolated macrophages were directly infected with P. carinii over 24 h. Interestingly, P. carinii significantly induced increased mRNA expression for both the M1 and the M2 marker genes iNOS and Arg-1 in the P. carinii-challenged macrophages compared to controls. However, the M1 responses were greater than those observed for the M2 markers, in contrast to what had been observed in the immunocompetent hosts in vivo (Fig. 4A). In a parallel fashion, bone marrow-derived macrophages (BMDM) were also challenged with P. carinii. Interestingly both alveolar and bone marrow-derived macrophages responded similarly after P. carinii infection (Fig. 4B). These observations indicate an overall M1 predominant polarization of macrophages resulting from P. carinii infection in the absence of the native cytokine milieu found in the lung microenvironment.

FIG 4.

Macrophage polarity during Pneumocystis infection is plastic. (A) Alveolar macrophages were stimulated with 5 or 10 P. carinii organisms per macrophage or left unstimulated. After 24 h, macrophage mRNA was harvested and analyzed for the M1 (iNOS) or M2 (Arg-1) marker genes by real-time PCR. BMDM were stimulated with IL-4 or IFN-γ at concentrations of 20 and 100 ng/ml, respectively, or with P. carinii at a ratio of 10 organisms per macrophage for 24 h. In addition, BMDM were stimulated with P. carinii (10 P. carinii per alveolar macrophage) for 24 h (primary stimulation), followed by a second stimulation with IL-4 or IFN-γ for an additional 24 h (secondary stimulation). (B, D, and E) After stimulation, macrophage mRNA samples were harvested and analyzed by qPCR for mRNA expression of iNOS and ARG-1 (B) and other M1 genes (IL-1β, IL-6, and TNF-α) (D) or analyzed for M2 genes (Fizz1, Pparγ, Mrc1, and IL-10) (E). The mRNA expression levels from A, B, D, and E were normalized to GAPDH, and the values are presented as the relative expression compared to the mRNA levels of unstimulated controls, whose values were set to 1. The data shown are integrated means ± the SD from three independent experiments with an asterisk (*) indicating P < 0.05. (C) Macrophage lysates from BMDM were prepared after stimulation and analyzed by Western blot for protein expression of iNOS and ARG-1. β-Actin was used as a loading control, and the data are representative of three independent experiments.

We further tested the ability of cytokines to modulate P. carinii-mediated polarization in vitro. BMDM were either stimulated for 24 h with a primary stimulus alone and/or for an additional 24 h with a secondary stimulus following the primary stimulus. When provided as the primary stimulus, IL-4 induces an M2 polarized state with high Arg-1 expression and low iNOS expression. In contrast, IFN-γ polarized the macrophages to an M1 state with high iNOS expression and low Arg-1 expression compared to unstimulated controls (Fig. 4B). We further observed that when provided as a secondary stimulus, IL-4 reversed the primary P. carinii-induced M1 polarization. In comparison, when provided as a secondary stimulus, IFN-γ augmented the primary M1 response compared to samples that were induced with P. carinii alone. The increase in M2 markers that was observed following P. carinii infection was completely suppressed when secondarily stimulated with IFN-γ for 24 h and was significantly boosted following secondary IL-4 stimulation after an initial overnight challenge with P. carinii (Fig. 4B). We further confirmed that the protein levels of M1 and M2 marker genes, iNOS and Arg-1, were altered in a similar fashion to their respective mRNA expression (Fig. 4C and see Fig. S2 in the supplemental material). Taken together, these data indicate that P. carinii-induced macrophage polarization in BMDM can be successfully reversed in vitro with appropriate cytokine stimulation.

We further evaluated the expression of a panel of M1 and M2 marker genes to understand whether the reversal of P. carinii-induced macrophage polarity was mirrored by changes in the overall macrophage gene expression profiles. Specifically, we analyzed the expression of M1-related genes (TNF-α, IL-1β, and IL-6), as well as the M2-related genes (IL-10, Fizz1, Pparγ, and Mrc1), using real-time PCR and enzyme-linked immunosorbent assay (ELISA). The M1-related genes were induced following stimulation of the macrophages with either P. carinii or IFN-γ consistent with the iNOS expression. In contrast, the M2 genes either were inhibited or remained stable compared to the levels of unstimulated controls following P. carinii infection (Fig. 4D and E and Fig. S4 in the supplemental material). IL-4 stimulation triggered the opposite gene expression in macrophages, with the M1 genes being suppressed, while the M2 genes were enhanced compared to controls. The macrophages that were secondarily stimulated with IL-4 for 24 h following initial overnight stimulation with P. carinii demonstrated inhibition of the P. carinii-induced M1 gene expression profile, while the M2 gene signature was significantly elevated compared to macrophages stimulated with P. carinii alone (Fig. 4D and E and Fig. S4 in the supplemental material). As anticipated, we observed the opposite pattern with increased M1 gene signatures and suppressed M2 gene expression profiles when the macrophages were secondarily treated with IFN-γ for 24 h following an initial overnight stimulation with P. carinii. Overall, Pneumocystis stimulation was largely shown to be an M1 stimulus to the macrophages, similar to IFN-γ. The exceptions to this overall trend were IL-10 and Arg-1, which are M2 genes. Their levels were significantly elevated after stimulation with P. carinii alone. However, the expression of IL-10 and Arg-1 was significantly inhibited when the macrophages were secondarily treated with IFN-γ and further elevated when treated secondarily with IL-4 similar to other M2 genes (Fig. 4D and E and see Fig. S4 in the supplemental material). Together, these data indicate that the P. carinii-induced M1 macrophage polarization could be reversed toward M2 phenotypes using bone marrow-derived macrophages in vitro.

The immunopathogenesis caused by P. carinii infection can be modulated in vivo through adoptive transfer of M1 and M2 macrophages.

The cumulative in vitro data indicated that the P. carinii-induced M1 and M2 macrophage response could be altered using BMDM with appropriate cytokine stimulation. Accordingly, we next investigated whether this could be achieved in vivo by adoptive transfer of polarized M1 and M2 macrophages intratracheally using our rat model of Pneumocystis pneumonia. Animals were immunosuppressed for 1 week prior to infection and were subsequently inoculated with P. carinii. The animals were immunosuppressed continuously throughout the infection period. After 8 weeks of infection, in vitro-derived and polarized M0, M1, or M2 macrophages were transferred into the P. carinii-infected rats intratracheally. One week later, BALF and lung tissues were collected for analysis (see Fig. S5 in the supplemental material).

We first assessed the net effect of these transferred cells on the intensity of infection in vivo by evaluating the P. carinii burdens using real-time PCR. Interestingly, all of the animals that received macrophages, whether MO, M1, or M2 polarized ones, cleared P. carinii significantly better than the control animals receiving phosphate-buffered saline (PBS) alone (Fig. 5A). Strikingly, there was greater P. carinii clearance in the animals that received M1 macrophages in comparison to controls that received PBS alone and also compared to the animals that received M0 macrophages (Fig. 5A).

FIG 5.

The immunopathogenesis caused by P. carinii infection can be modulated in vivo through adoptive transfer of M1 and M2 macrophages. Animals were immunosuppressed with Depo-Medrol at 20 mg/kg 1 week prior to infection and inoculated with P. carinii (10⁶ P. carinii/rat). The animals were immunosuppressed every week throughout the subsequent infection period. At 8 weeks postinfection, polarized M0, M1, or M2 BMDM (5 × 10⁶ cells/rat) were transferred into the infected rats intratracheally. One week later, samples were collected for analysis. (A) The P. carinii burden was assessed in the lung tissues by real-time PCR. The data shown are means ± the SD of n ≥ 8 animals per group that were age and sex matched and are representative of two independent experiments. (B and C) Pathology of P. carinii-infected lung tissue treated for GMS staining (B) and H&E staining (C) showing Pneumocystis cysts obtained from formalin-fixed, paraffin-embedded lungs treated with M0, M1, and M2 macrophages or PBS. Representative images of stained lung sections from n ≥ 8 animals per group and representative of two independent experiments are shown. (D) The data shown are survival curves of animals monitored for 2 weeks posttransfer of macrophages and PBS-injected animals, with median survivals of the PBS-injected group at 12 days and the M0-transferred group at 13 days (log rank [Mantel-Cox] statistic of PBS versus M1, P = 0.0445; log rank statistic of PBS versus M2, P = 0.0450). The data shown are from n = 12 animals per group at initiation. (E) Inflammation was scored blindly from n ≥ 8 animals per group and analyzed using ImageJ software. The scores are representative of two independent experiments. The bars represent medians.

We further stained the lung sections of these animals with hematoxylin and eosin (H&E) and Gomori methenamine silver (GMS). Consistent with our PCR data, we observed a reduction in P. carinii burden in the lungs of animals transferred with macrophages compared to PBS-treated animals. Furthermore, the P. carinii reduction was more striking in the lung sections from M1-transferred animals compared to the M0-treated mice (Fig. 5B). Interestingly, the P. carinii burden in the lungs of M2-transferred animals, though significantly lower compared to PBS-treated controls, was comparable to that of M0-transferred animals rather than the M1-treated animals (Fig. 5C). These data demonstrate that although transferred macrophages in general were effective in reducing organism burdens, the M1 macrophages exerted the greatest effect in clearing P. carinii organisms in vivo.

Furthermore, the H&E staining suggested that overall lung inflammation was greater in the lungs of M1-transferred animals compared to the lungs of M0-transferred animals and similar compared to the PBS-injected controls. In contrast, inflammation appeared to be reduced in the lungs of M2-transferred animals compared to control animals that received PBS and to the animals that received M0 macrophages (Fig. 5C and E). These data indicate that more severe lung inflammation occurs as a result of M1 polarization in the lung during the course of P. carinii infection despite greater clearance of organisms under these conditions.

In addition, when we monitored the animals for about 2 weeks posttreatment with macrophages, we observed a better survival with both the groups of animals that received M1 and M2 macrophages in comparison to the ones that received M0 macrophages or just PBS (Fig. 5D). We terminated the experiment after 2 weeks posttransfer because the PBS-treated controls were moribund, and ethical considerations prevented longer observation.

M1 and M2 gene signatures are altered in lung after treatment with polarized macrophages in P. carinii-infected animals.

We further questioned whether the gene signatures of macrophages in the lungs of macrophage-treated animals were altered in concert with the improvement in organism clearance, inflammation, and survival. Very interestingly, iNOS, an M1 marker gene that aids in mediating the antimicrobial activity of macrophages, was significantly elevated in the BALF of all animals that received polarized macrophages in comparison to PBS-treated control animals (Fig. 6A). We further evaluated the expression of a panel of M1 and M2 genes both at the mRNA and protein levels using real-time PCR in macrophages isolated from the BALF and ELISA from the BALF of these animals 1 week after the adoptive transfer of macrophages. We observed elevated levels of TNF-α mRNA in the lungs of all animals that received activated macrophages, including the ones that received M2 cells, compared to PBS-treated control animals, whereas the IL-6 and IL-1β mRNA levels were elevated only in M1-treated animals compared to M0-treated or PBS-treated control animals (Fig. 6A and see Fig. S6 in the supplemental material). In contrast, Arg-1 and other M2-related genes were elevated only in the M2-transferred animals compared to controls, with IL-10 being an exception since its expression was also elevated in the M0-treated groups as well (Fig. 6B and see Fig. S6 in the supplemental material). In summary, these data indicate that there was generally a predominance of M1 gene signature in P. carinii-infected animals, whether M0 or M1 cells were transferred. However, the presence of M2 cells was able to drive the expression of M2 markers during P. carinii infection.

FIG 6.

M1 and M2 gene signatures are altered in the lungs after treatment with polarized macrophages in P. carinii-infected animals. Macrophage mRNA was prepared from BALF of P. carinii-infected rats 1 week after adoptive transfer of M0, M1, and M2 macrophages or of those that were injected with PBS intratracheally and analyzed by qPCR for mRNA expression of M1 genes (iNOS, IL-1β, IL-6, and TNF-α) (A) and M2 genes (Arg-1, Fizz1, Pparγ, Mrc1, and IL-10) (B). The mRNA expression levels were normalized to GAPDH. The data show the means ± the SD of n ≥ 8 animals per group and are representative of two independent experiments. An asterisk (*) indicates P < 0.05.

Macrophage-mediated clearance activity and molecular mechanisms controlling M1 and M2 functions after treatment with polarized macrophages in P. carinii-infected animals.

In addition to alterations in the gene signature of the lung, we investigated the macrophage-mediated clearance activity in the lungs treated with polarized macrophages. To address this, we measured both the oxidative and nitrosative bursts from macrophages isolated from BALF of animals treated with M0, M1, and M2 cells or PBS and restimulated these cells with P. carinii organisms in vitro for 4 h. There was an effective production of NO by all the macrophages isolated from M0-, M1-, or M2-treated lungs, with macrophages isolated from M1-treated lungs being a potent generator of nitrogen intermediates compared to those isolated from PBS-treated lungs (Fig. 7A). These data are consistent with induced iNOS expression observed in all groups transferred with activated macrophages in comparison to the PBS-treated controls. This is in contrast to what was observed in the endogenous immunosuppressed alveolar macrophages that were not able to activate their iNOS expression into functional production of NO (Fig. 6A and D). In addition, there was an effective release of H₂O₂ by macrophages isolated from M2-treated lungs compared to PBS-treated controls (Fig. 7A). These data, together with the significant generation of nitrosative burst in the macrophages isolated from lungs of animals that received IFN-γ activated M1 macrophages, might explain the improved survival observed in these adoptively treated animals (Fig. 5B). Although there was some nitrogen intermediate production by the macrophages isolated from M0-treated lungs, it was not as high as macrophages isolated from lungs of animals that received IFN-γ-activated M1 macrophages and was probably not sufficient to eliminate P. carinii organisms effectively (Fig. 7A and B).

FIG 7.

Molecular mechanisms controlling M1 and M2 macrophage activity after treatment with polarized macrophages in P. carinii-infected animals. One week after adoptive transfer of P. carinii-infected rats with M0, M1, and M2 macrophages or PBS intratracheally, the animals were sacrificed. (A) Macrophages harvested from BALF were restimulated with P. carinii in vitro for 4 h after a 90-min adherence and removal of nonadherent cells, after which nitrite concentrations (nitrosative burst) and H₂O₂ (oxidative burst) release into the cell culture supernatant were measured using calorimetric assays. (B) IFN-γ, IL-4, and IL-13 secretion were measured by ELISA in the first 5 ml of BALF. (C) Lysates were prepared from pooled macrophages harvested from BALF of 5 rats and analyzed by Western blot for protein expression of phospho STAT6 (Y641), phospho STAT1 (Y701), and total STAT1 and STAT6. Data shown in panels A and B are expressed as means ± the SD of n = 5 animals per group that were age and sex matched and are representative of three independent experiments. An asterisk (*) indicates P < 0.05. The data shown in panel C are representative of two independent experiments.

To examine molecular mechanisms that may underlie the phenotypic alterations observed in the lung microenvironment after adoptive transfer of differentially polarized macrophages, we investigated STAT signaling, a central pathway in the control of macrophage M1 and M2 polarization (12). The levels of phosphorylated STAT6 were increased in the macrophages isolated from BALF when M2 cells were transferred, in contrast to increased levels of phosphorylated STAT1 observed when M1 cells were transferred (Fig. 7C). Furthermore, STAT1 was also increased when M0 cells were transferred, but the levels of phosphorylated STAT1 were relatively low compared to macrophages derived from control animals that received PBS alone (Fig. 7C). We further evaluated the upstream cytokines driving polarization in the lung microenvironment and observed that the levels of proinflammatory IFN-γ and anti-inflammatory IL-4/IL-13 cytokines were increased in the lung microenvironment of infected animals following adoptive transfer of M1 and M2 macrophages, respectively (Fig. 7B). These data clearly demonstrate the predominance of STAT1 and STAT6 activation by the IFN-γ and IL-4/IL-13 increases that result from treatment of the P. carinii-infected mice with M1- and M2-polarized cells, respectively.

Dectin-1 expression is elevated in the lungs of M2-treated and immunocompetent animals during P. carinii infection.

We next examined potential mechanisms through which M2 cells may further promote clearance of P. carinii during infection. Traditional concepts suggest that M1 cells are the prototypic cells mediating fungal killing during infection. However, our data indicate that M2 polarization is the dominant pathway adopted in the immunocompetent host during native clearance of Pneumocystis, and we further observed substantial reduction in P. carinii burden and increased survival when P. carinii-infected animals were treated with M2 cells. Dectin-1 has been described to be a major innate immune receptor mediating the clearance of Pneumocystis during infection (34). In addition, Dectin-1 has been shown to be elevated in M2 macrophages (35). Accordingly, we further evaluated Dectin-1 expression in our disease and treatment models of Pneumocystis infection. Interestingly, Dectin-1 mRNA and protein were significantly elevated in the M2-treated macrophages isolated from BALF compared to that from other groups. In addition, M1 treatment did not greatly alter Dectin-1 expression (Fig. 8A and B). Furthermore, Dectin-1 expression was significantly upregulated in the P. carinii-infected immunocompetent alveolar macrophages but was significantly reduced in the P. carinii-infected immunosuppressed alveolar macrophages compared to uninfected controls (Fig. 8C). These data indicate that Dectin-1 induction in M2 cells appears to be an inherent phenomenon during P. carinii infection and is defective in the immunosuppressed hosts that succumb to this infection.

FIG 8.

Dectin-1 expression is elevated in the lungs of M2-treated and immunocompetent animals during P. carinii infection. One week after adoptive transfer of P. carinii-infected rats with M0, M1, and M2 macrophages or after the injection of only PBS (control) intratracheally, the animals were sacrificed. (A) Macrophage mRNA was prepared and analyzed by qPCR for the mRNA expression of Dectin-1. The mRNA expression levels were normalized to GAPDH. Data shown in panels A and B are means ± the SD of n = 4 animals per group and are representative of three independent experiments, with an asterisk (*) indicating P < 0.05. (B) Lysates were prepared from pooled macrophages harvested from BALF of four rats and analyzed by Western blotting for protein expression of Dectin-1. (C) Lysates were prepared from pooled alveolar macrophages harvested from BALF of four immunocompetent or immunosuppressed animals that were infected with P. carinii intratracheally for 7 days or 8 weeks, respectively, and analyzed by Western blotting for protein expression of Dectin-1. Uninfected animals were used as controls. In panels B and C, β-actin was used as a loading control, and the data are representative of two independent experiments.

Dectin-1 induction acts as one of the potential M2-mediated Pneumocystis clearance mechanisms.

We next sought to determine whether Dectin-1 was enhanced on M2 cells and thus may represent one mechanism by which these cells may assist in the clearance of P. carinii. We tested this in vitro using BMDM and observed that Dectin-1 expression was increased in M2-polarized cells compared to both M1 and M0 cells both in the presence and in the absence of P. carinii stimulation (Fig. 9A and B). However, when the M2 cells were inhibited by neutralizing the IL-4 and IL-13 cytokines using anti-IL-4 antibody and stimulated with P. carinii organisms, Dectin-1 expression was suppressed. Furthermore, the addition of IL-4/IL-13 back into the macrophage-P. carinii cocultures rescued the Dectin-1 expression (Fig. 9C and D). M1 inhibition using anti-IFN-γ antibody or its rescue with additional IFN-γ did not affect Dectin-1 expression in the P. carinii-infected macrophages. We simultaneously evaluated the ability of these macrophages to clear P. carinii using real-time PCR. We observed that inhibiting M2 cells using anti-IL-4 antibody significantly reduced P. carinii elimination and that P. carinii clearance was restored when IL-4 was added back, in a fashion parallel to the reduction and restoration of Dectin-1 expression under these same conditions (Fig. 9E). These data demonstrate that Dectin-1 is an IL-4-regulated M2 gene and is correlated with the M2 cells mediating organism clearance during Pneumocystis infection.

FIG 9.

Dectin-1 induction acts as an M2-mediated Pneumocystis clearance mechanism. Dectin-1 protein expression was analyzed by Western blotting from cell lysates prepared from BMDM. (A) BMDM were polarized to M0, M1, or M2 phenotypes for 7 days. (B) BMDM were polarized for 7 days and stimulated with P. carinii (10 P. carinii per BMDM) for another 18 h. (C) BMDM were polarized for 7 days and cultured in the presence of anti-IFN-γ (20 μg/ml) or anti-IL-4 (20 μg/ml) and anti-IL-13 (20 μg/ml) or anti-IgG for an additional day to neutralize their M1 and M2 signatures and then further stimulated with P. carinii organisms (10 P. carinii per BMDM) for an additional 18 h. (D) BMDM were polarized for 7 days, and their M1 and M2 signatures were neutralized by using anti-IFN-γ or anti-IL-4 and anti-IL-13 or anti-IgG antibodies as described above for an additional day (day 8), after which they were replated and rescued by the addition of IL-4 (20 ng/ml) and IL-13 (20 ng/ml) or IFN-γ (100 ng/ml) or PBS and stimulated with P. carinii at a ratio of 10 organisms per macrophage for another 18 h (day 9). (E) Total RNA was extracted from the contents of each well described above, and quantitative real-time PCR for P. carinii mRNA copies was performed. The data shown in panels A to E are representative of three independent experiments.

DISCUSSION

Although host recognition and inflammatory responses are required for clearance of P. carinii, such outcomes also result in severe lung injury and respiratory compromise. Indeed, respiratory failure and death during Pneumocystis pneumonia in humans have been more strongly linked to lung inflammation rather than to direct toxic effects of Pneumocystis (2). Antibiotics used to treat the infection may lead to increased exposure of proinflammatory components such as β-glucans present on dying P. carinii organisms, resulting in worsening gas exchange. Anti-inflammatory corticosteroids, given in addition to antimicrobial agents, significantly improve outcomes during severe PcP. However, steroids are immunosuppressive and further increase the risk of coinfection with other organisms (12). Our recent studies have focused on developing strategies to effectively promote clearance of Pneumocystis, while reducing lung inflammation during pneumonia.

Macrophages are the primary lung effector cells promoting clearance of Pneumocystis organisms during infection (36–38). However, the mechanisms by which macrophages clear Pneumocystis are not fully understood. Specifically, the functional differentiation of macrophages during Pneumocystis infection has not been completely defined. Figure 10 provides a potential model for macrophage differentiation and clearance during Pneumocystis infection based on the findings of our study. We demonstrate a strong induction of M1 responses following P. carinii infection both in vitro and in vivo in the immunosuppressed host. However, in the immunocompetent host, with appropriate lung cytokine microenvironment, macrophage polarization is strongly driven toward an M2 phenotype during P. carinii infection. The adoptive transfers of M1 and M2 cells demonstrated that STAT1 and STAT6 activation by IFN-γ and IL-4/IL-13 might serve as molecular mechanisms driving macrophage polarization during P. carinii infection. The immunocompetent host exhibited a greater M2 response with elevated Dectin-1 expression, although the M1 pathway was also active and functional nitrogen intermediates were being generated in these animals. In general, enhanced Dectin-1 expression has been associated with increased clearance of Pneumocystis, most probably through more robust recognition of the organism, with better uptake and eventually elimination of the organism. Altogether, this results in few significant symptoms during the native clearance of P. carinii in immunocompetent hosts. In contrast, in the absence of sufficient M2-driving cytokine, the cells differentiate toward an M1 state with intact inducible nitric oxide synthase (iNOS) expression but defective generation of active nitrogen intermediates, resulting in P. carinii accumulation and excessive inflammation. Treatment with both M1 and M2 cells improved the survival of the P. carinii-infected immunosuppressed hosts. However, M2 treatment provided the best overall outcomes with efficient P. carinii killing and reduced inflammation, a state reminiscent of the inherent M2 response in the immunocompetent host challenged with Pneumocystis.

FIG 10.

Hypothetical model for the differential role of classical (M1) and alternatively activated (M2) macrophage polarization in immunocompetent and immunosuppressed hosts during pulmonary Pneumocystis infection.

An earlier study of modulating macrophage polarization during PcP using sulfasalazine demonstrated that this drug diminishes pulmonary inflammation while preserving the host defense by alternatively activating M2 lung macrophages (39). An additional study had shown that alternatively activated M2 macrophages are potent effector cells against Pneumocystis murina and suggests that enhancing M2a polarization may be an adjunctive therapy for the treatment of Pneumocystis infections (33). Parallel to our observation, these investigators observed that alveolar macrophages from P. murina-infected Src TKO mice expressed significantly greater levels of the M2a markers and possessed enhanced organism clearance activity and yet did not demonstrate enhanced inflammatory responses to P. murina (33). In addition, we observed that the lung microenvironment in immunocompetent animals that rapidly clear P. carinii organisms favors M2 polarization with increasing amounts of M2-driving cytokines. Additional reports indicate that immunocompetent mice infected with P. murina had a greater proportion of CD4 T cells producing IL-4 rather than IFN-γ within the first 7 days through 3 weeks of infection (40). Furthermore, IL-5 and IL-13 levels in BALF were higher than IFN-γ secretion in immunocompetent mice infected with P. murina (41). Taken together with our observations, these findings support that an M2-polarized lung environment promotes the natural clearance of Pneumocystis in immunocompetent hosts.

In immunosuppressed hosts that lack a lung environment conducive for M2 differentiation, sufficient IFN-γ in the lungs drives M1 polarization during P. carinii infection. In fact, Pneumocystis was able to independently drive M1 polarization of macrophages in vitro, and in general there was a predominance of M1 phenotype in the lungs of immunosuppressed and P. carinii-infected animals that were treated with either M0 or M1 cells. However, the transfer of M2 cells was able to drive a greater M2-like response during P. carinii infection in the immunocompromised host. In addition, the macrophages from immunosuppressed and P. carinii-infected hosts were defective in the production of nitric oxide, which would aid in Pneumocystis clearance despite possessing increased iNOS levels. These data of ours were indeed supported by another finding reported by Lasbury et al. (42). These researchers detected low levels of NO in the BALFs from immunosuppressed, P. carinii-infected rats and mice compared to those from uninfected animals. Consistent with our observations, these researchers also reported that the iNOS mRNA and protein concentrations were higher in alveolar macrophages from P. carinii-infected and immunosuppressed rats and mice. These researchers further demonstrated that the lack of calmodulin in the alveolar macrophages from P. carinii-infected rats inhibited homodimerization of iNOS and the production of functional NO in these animals (42). However, we demonstrated that when macrophages are activated in vitro and then transferred into immunosuppressed animals they are able to generate nitrogen intermediates that are associated with clearance of P. carinii. This strong response induced by in vitro-primed and activated macrophages is consistent with data from other infectious models that document that activated macrophages exhibit enhanced antimicrobial activity (43–46).

The BALF from immunosuppressed and P. carinii-infected animals has also been shown to exhibit low levels of granulocyte-macrophage colony-stimulating factor, which suppress the levels of PU.1, a critical regulator of Dectin-1 expression (47, 48). Consistent with this, we observed severe reductions in Dectin-1 expression in macrophages isolated from P. carinii-infected immunosuppressed animals. In contrast, Dectin-1 expression was significantly upregulated in the immunocompetent animals. Dectin-1 expression is critical for host defense against Pneumocystis (34, 49). Our data and other studies indicate that Dectin-1 is upregulated in the cell surface of M2 cells, most probably through IL-4/STAT6-mediated regulation (50, 51). In our present study, we propose that indeed M2 cells mediate their effector function against Pneumocystis by enhancing Dectin-1 expression. In fact, Dectin-1 expression was substantially elevated in the immunocompetent animals, which displayed M2 dominance during P. carinii infection. Furthermore, when we inhibited the M2 phenotype by neutralizing the IL-4 stimulation, Dectin-1 levels and concurrently the clearance of P. carinii by these M2 cells were significantly suppressed. Steele et al. demonstrated that phagocytosis of P. carinii by alveolar macrophages is mediated by Dectin-1 receptor with subsequent generation of hydrogen peroxide but not the generation of nitrogen intermediates (34, 52). Our data are consistent with these observations since we observed an effective release of hydrogen peroxide in the immunocompetent alveolar macrophages that displayed enhanced M2 phenotype and Dectin-1 expression. We observed similar findings in the macrophages isolated from M2-transferred lungs. Together, these data clearly support the contribution of an alternatively activated M2 response with elevated Dectin-1 levels as effective macrophage host defense promoting the clearance of Pneumocystis. In addition, our study further supports that P. carinii-induced macrophage polarization can completely be reversed at both the levels of gene and protein expression levels if the macrophage is provided with appropriate cytokine stimuli, a finding consistent with other models (32).

We further investigated whether polarized macrophages would function in vivo to provide therapeutic benefit during Pneumocystis pneumonia. We found that the adoptive transfer of macrophages was effective in alleviating the inflammatory pathogenesis of the disease. Adoptive transfer of M2 cells significantly reduced the inflammation and transfer of any type of macrophages (M0, M1, or even the anti-inflammatory M2 cells) to the P. carinii-infected animals was associated with improved clearance of P. carinii organisms. Overall, we observed a predominance of M1 phenotype unless our exogenous M2 cells were provided to override the P. carinii-stimulatory M1 response. In fact, some level of M1 response was detected in the lung environment even with adoptive transfer of M2 cells in the P. carinii-infected animals. The reason for this is unknown, but P. carinii organisms were shown to independently induce a strong M1 response in vitro without any need for exogenous cytokine signals. Since Pneumocystis stimulation was very much comparable to IFN-γ, an M1 stimulus, we tested whether macrophages prestimulated with P. carinii organisms produced similar effects as we observed for M1 adoptive transfers in the P. carinii-infected immunosuppressed animals. This was not the case (see Fig. S7 in the supplemental material), since P. carinii infection, unless accompanied by cytokine activation (43–46), renders the macrophages defective in NO production. The exact factors contributing to IL-4 and IFN-γ increases in the lung microenvironment of the adoptively transferred animals that drive the persistence of M1 and M2 phenotype remain unknown but may be related to other cells and tissues in the lung microenvironment that contribute to secretion of cytokines and factors that maintain the M1 and M2 phenotypes (53–56).

Taken together, our study emphasizes the plastic nature of macrophage polarization during Pneumocystis pneumonia, as well as the potential benefit of M2 macrophages in mediating clearance of P. carinii and reducing lung inflammation during infection. In addition, we demonstrate a possible strategy for immunotherapy that could be tailored to control and balance host defense and inflammation, which proves lethal to patients with severe Pneumocystis pneumonia.

MATERIALS AND METHODS

Animals studied and rat Pneumocystis pneumonia model.

The Mayo Clinic Institutional Animal Care and Utilization Committee approved all animal studies. P. carinii organisms were derived from the American Type Culture Collection (Rockville, MD) and maintained in immunosuppressed rats as we described (57). Female Long Evans rats (Harlan, Inc., Indianapolis, IN) were provided with 20 mg of Depo-Medrol (American Regent, Shirley, NY)/kg subcutaneously once weekly to induce immunosuppression. Antibiotics were simultaneously administered to suppress bacterial superinfection, using ampicillin (500 mg/liter; DAVA Pharmaceuticals, Fort Lee, NJ) or cephalexin (500 mg/liter; Ranbaxy, India) added to the drinking water on alternate days (57). Control groups of the immunosuppressed animals that were uninfected also received equal dosage of Depo-Medrol and antibiotics parallel to the P. carinii-infected animals. After 5 days, rats were anesthetized and intratracheally inoculated with P. carinii (10⁶ P. carinii nuclei) prepared by homogenizing infected rat lung. The animals were sacrificed after an additional 8 weeks of immunosuppression, and samples were collected for analysis or used for adoptive-transfer experiments as described below. Analyses consisted of parallel subgroups from each treatment group for histology and lung tissue collection and a second subgroup for BALF. Immunocompetent animals were infected similarly and monitored for ∼30 days. Control groups of the immunocompetent animals were uninfected but were administered with equal dosage of antibiotics similar to P. carinii-infected animals.

Isolation of P. carinii.

After 8 weeks of infection, the lungs were harvested, minced, and homogenized in PBS (pH 7.3). Homogenates were strained through sterile gauze and examined on cytopreparation smears. Bacterial or fungal contaminated samples were discarded. Samples were centrifuged in a Beckman J2-MC centrifuge at 4,960 × g for 10 min at 4°C, and the supernatants were discarded. Red blood cells were lysed using distilled deionized water, and cells were resuspended in PBS (pH 7.3) and centrifuged as described above. The cellular pellets were resuspended, and homogenates were passed through a 10-μm-pore-size filter (Millipore, Inc., Billerica, MA) (57). Duplicate 10-μl aliquots of suspension were spotted onto slides and stained with modified Wright-Giemsa stain (Diff-Quik), and the P. carinii organisms were quantified (57).

BMDM isolation and culture.

Female Long Evans rats were euthanized as approved by the Mayo Clinic Institutional Animal Care and Utilization Committee. The tibia and femur bones were excised, and bone marrow cells were flushed and dispersed into culture medium. Bone marrow-derived macrophages (BMDM) were derived in culture using Dulbecco minimal essential medium containing 20% fetal bovine serum, l-glutamine/l-alanyl l-glutamine supplementation (GlutaMAX; Thermo Fisher Scientific), MEM-nonessential amino acids, sodium pyruvate, penicillin, and streptomycin (all purchased from Life Technologies, Inc., Gaithersburg, MD) and recombinant rat macrophage colony-stimulating factor (M-CSF; 50 ng/ml; Peprotech, Rocky Hill, NJ) for 6 days in a non-tissue culture dish. An additional volume of the above-mentioned medium was added once on the fourth day of culture. For subsequent experiments, BMDM cells were removed from differentiation dishes using cold sterile pyrogen-free PBS containing 5 mM EDTA (all purchased from Life Technologies) and subsequently cultured in RPMI 1640 medium (Life Technologies) containing 10% fetal bovine serum, GlutaMAX, MEM-nonessential amino acids, sodium pyruvate, penicillin, and streptomycin and recombinant rat M-CSF (5 ng/ml). Recombinant rat M-CSF (5 ng/ml) was included in all BMDM infection and stimulation experiments.

Bronchoalveolar lavage and macrophage isolation.

Alveolar macrophages were isolated by bronchoalveolar lavage as described previously (58). Rat lungs were lavaged with sterile, pyrogen-free PBS. A minimum of 30 ml of BALF was recovered from the lungs of each rat. The cells in the lavage fluid samples were pelleted by centrifugation at 300 × g for 5 min at 25°C and resuspended to 1.5 × 10⁶/ml in RPMI 1640 medium containing 10% fetal bovine serum, GlutaMAX, MEM-nonessential amino acids, sodium pyruvate, penicillin, and streptomycin. The bronchoalveolar lavage cells were counted using a hemacytometer, and macrophages were selected from BALF by 90-min adherence selection on tissue culture plastic dishes, followed by removal of nonadherent cells by washing.

Cell stimulation and infection experiments.

For experiments requiring M1 and M2 polarization in vitro, BMDM cells were plated into fresh RPMI 1640 medium containing 5 ng/ml M-CSF. At the end of 6 days of differentiation, the cells were stimulated with recombinant rat cytokine IFN-γ (100 ng/ml to drive M1 differentiation) or IL-4 and IL-13 (20 ng/ml to drive M2 differentiation), respectively, overnight. BMDM cells that were not stimulated with any cytokines were termed M0. For cell infection with P. carinii, BMDM or alveolar macrophages (replated after 3 h of adherence selection) were plated into fresh RPMI 1640 medium and stimulated with P. carinii at 5:1 or 10:1 (P. carinii/macrophage ratio), followed by incubation at 37°C overnight.

To assess macrophage plasticity, polarized cells were secondarily restimulated with the appropriate M1 and M2 driving cytokines. To accomplish this, the cultured cells were replated at the same density as the primary cultures and secondarily stimulated with IFN-γ (100 ng/ml) to promote M1 polarization or IL-4 and IL-13 (20 ng/ml each)-containing media to promote M2 polarization for an additional 18 h. All cytokines were purchased from Peprotech, Rocky Hill, NJ. After these stimulations, the cells were washed with cold sterile pyrogen-free PBS (without Ca/Mg) containing 5 mM EDTA and harvested for analysis. Determinations of markers of M1 and M2 polarization were accomplished using quantitative PCR for mRNA of M1 and M2 genes and ELISAs for M1 and M2 cytokines.

For neutralization and rescue of M1 and M2 differentiation, BMDM cells were polarized as described on day 7 after 6 days of differentiation of BMDM cells with M-CSF. On day 8, the cells were further cultured in the presence of anti-IFN-γ (20 μg/ml, catalog no. 507802; BioLegend, San Diego, CA) or anti-IL-4 (20 μg/ml; catalog no. 507802; BD Biosciences, San Jose, CA) and anti-IL-13 (20 μg/ml; catalog no. ab10783; Abcam, Cambridge, MA) neutralizing antibodies. The following day, the cells were replated and stimulated with only P. carinii organisms (10:1 P. carinii/macrophage ratio) for 18 h or with P. carinii organisms in the presence of IFN-γ or of IL-4 and IL-13 for 18 h. Subsequently, samples were collected for analysis. Equal concentrations of anti-IgG of the specific isotypes were used as control antibodies for neutralization, and volumes of PBS equal to that of cytokines were used as controls for cytokine stimulation.

Adoptive transfer of macrophages.

Adoptive transfer was performed after 8 weeks of immunosuppression and P. carinii infection. M0, M1, and M2 macrophages (as described above) or M-P. carinii (BMDM stimulated with a 10:1 ratio of P. carinii for 24 h) and PBS as a vehicle control were transferred into the P. carinii-infected rats intratracheally (31, 59, 60). One week after the adoptive transfer, the animals were sacrificed, and samples were collected for analysis. Parallel subgroups from each treatment group were used for histology (left lung) and lung tissue collection for P. carinii burden determination (right lung) and a second subgroup for BALF collection.

Nitrite and H₂O₂ assays.

Alveolar macrophages harvested from BALF were purified by adherence to plastic dishes (90 min) and removal of nonadherent cells, followed by restimulation with P. carinii in vitro for 4 h. Nitrite concentrations as a measure of NO (nitrosative burst) and H₂O₂ (oxidative burst) in the cell culture supernatant were measured using calorimetric assays. The nitrite measurement kit was purchased from Cayman Chemicals (catalog no. 780001; Ann Arbor, MI), and the H₂O₂ kit was purchased from Abcam (catalog no. ab102500; Cambridge, MA).

Quantitation of M1 and M2 gene expression using real-time PCR.

Alveolar macrophages isolated from BALF or lung tissues were lysed using a Qiashredder or a TissueLyser LT (Qiagen, Germantown, MD), respectively. Total RNA was purified with the RNeasy minikit (Qiagen). An iScript Select cDNA synthesis kit (Bio-Rad, Hercules, CA) was used for reverse transcription using oligo(dT) primers and random hexamer primer mix. A SYBR green PCR kit (Bio-Rad) was used for quantitative real-time PCR performed on an ICycler IQ (Bio-Rad). The sequences of the primer pairs are listed in Table S1 in the supplemental material.

Determination of P. carinii burden.

After sacrifice, the right lung was stored in RNAlater (Ambion, Inc.) for quantification of the P. carinii burden using real-time PCR. These samples were washed, and RNA was isolated as described. Quantitative PCR was performed to enumerate the P. carinii using the Bio-Rad iCycler with primers to the P. carinii prt1 multigene (spanning P. carinii prot1 and prot3) (61). Amplifications of unknown samples were compared to standard samples containing rat P. carinii prt1 RNA, and the resulting measure of RNA copies was proportioned to the net P. carinii burden (62). To quantify the P. carinii organisms in cell cultures, total RNA was isolated from the entire contents of the well, and similar procedures were followed.

ELISA determination of cytokine release.

Cytokines were analyzed from cell culture supernatants or from the first 5 ml of BALF extracted from rat lungs that was centrifuged at 200 × g for 10 min, and supernatants were collected for analysis. ELISA kits to measure rat TNF-α and IL-10 were purchased from R&D Systems, Minneapolis, MN.

Immunohistochemistry.

Paraffin embedding and staining were performed at the Mayo Clinic Histology Core, Scottsdale, AZ. The left lungs were inflation fixed with 10% buffered formalin overnight and embedded in paraffin. Sections (5 μm) were stained with H&E and GMS and graded blindly for the extent of P. carinii infection and pulmonary inflammation. The sections were scored as follows: 1+, mild perivascular aggregates; 2+, heavy perivascular aggregates; 3+, mild alveolar aggregates; 4+, alveolar exudate and heavy alveolar aggregates; and 0, normal. These scores were based upon grading of the entire lung surface area present on the slide section. Grading analysis was performed using ImageJ software.

Flow cytometry analysis of cell populations.

BALF cells from healthy and infected animals were first stained for 20 min at 4°C with CD16/CD32 Fc-blocking antibody, in flow cytometry buffer (PBS with 1% fetal bovine serum [FBS]), followed by incubation with anti-rat antibodies against CD3, B220, MHC-II (BioLegend), CD11C antigens or anti-mouse CCR3, or their corresponding isotype controls, conjugated with various fluorophores for an additional 20 min (63). Cells were then washed and resuspended in FACS buffer. All antibodies were purchased from eBioscience (San Diego, CA) unless otherwise stated. Data were collected on FACSCanto II (BD) and analyzed using FlowJo software (Tree Star, Inc.). The leukocytes in the BALF were gated out and cell numbers calculated using the method of van Rijt et al. (64).

Western blot analysis.

Lysates were prepared from alveolar macrophages isolated from BALF or from cell cultures using RIPA buffer (BD Biosciences, San Jose, CA), and samples were separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis. After transfer to nitrocellulose membranes, immunoblotting analyses were performed as previously described (65).

Statistical analysis.

Data are shown as means ± the standard deviations (SD) from multiple experimental determinations. Data were analyzed using one-way or two-way analysis of variance with Bonferroni adjustments using GraphPad Prism version 5.0b software, and statistical differences were considered significant if the P value was <0.05.