Impact of Surfactant Protein D, Interleukin-5, and Eosinophilia on Cryptococcosis

Abstract

Cryptococcus neoformans is an opportunistic fungal pathogen that initiates infection following inhalation. As a result, the pulmonary immune response provides a first line of defense against C. neoformans. Surfactant protein D (SP-D) is an important regulator of pulmonary immune responses and is typically host protective against bacterial and viral respiratory infections. However, SP-D is not protective against C. neoformans. This is evidenced by previous work from our laboratory demonstrating that SP-D-deficient mice infected with C. neoformans have a lower fungal burden and live longer than wild-type (WT) control animals. We hypothesized that SP-D alters susceptibility to C. neoformans by dysregulating the innate pulmonary immune response following infection. Thus, inflammatory cells and cytokines were compared in the bronchoalveolar lavage fluid from WT and SP-D^−/− mice after C. neoformans infection. Postinfection, mice lacking SP-D have reduced eosinophil infiltration and interleukin-5 (IL-5) in lung lavage fluid. To further explore the interplay of SP-D, eosinophils, and IL-5, mice expressing altered levels of eosinophils and/or IL-5 were infected with C. neoformans to assess the role of these innate immune mediators. IL-5-overexpressing mice have increased pulmonary eosinophilia and are more susceptible to C. neoformans infection than WT mice. Furthermore, susceptibility of SP-D^−/− mice to C. neoformans infection could be restored to the level of WT mice by increasing IL-5 and eosinophils by crossing the IL-5-overexpressing mice with SP-D^−/− mice. Together, these studies support the conclusion that SP-D increases susceptibility to C. neoformans infection by promoting C. neoformans-driven pulmonary IL-5 and eosinophil infiltration.

INTRODUCTION

Cryptococcus neoformans is an opportunistic fungal pathogen of the respiratory tract. It is the leading cause of fungal meningoencephalitis, with 1 million infections and 600,000 attributable deaths occurring annually. C. neoformans is especially prevalent in sub-Saharan Africa, where it causes approximately 30% of the deaths of HIV/AIDS victims (1). Because C. neoformans is an opportunistic respiratory pathogen, an immunocompetent host is generally able to control the infection within the lung. Conversely, when a host becomes immunocompromised, the fungi disseminate out of the lung, through the blood, and into the central nervous system (CNS), where uncontrolled growth of cryptococcal organisms typically results in host morbidity and mortality.

The nature of the immune response is a critical determinant of host outcome during C. neoformans pathogenesis. For example, Th1-skewed immune responses are generally considered host protective, while Th2-biased immune responses, characterized by high levels of interleukin-4 (IL-4), IL-5, and eosinophil expression, are nonprotective in the context of C. neoformans infection (2–8). Furthermore, an overexuberant inflammatory response can lead to complications such as cryptococcal immune reconstitution inflammatory syndrome (IRIS) (9–11). Thus, fine-tuned regulation of the inflammatory response is necessary to ensure a favorable outcome for the host against C. neoformans infection.

Although it has a proclivity for the CNS, C. neoformans is a respiratory pathogen that is normally encountered as an aerosol in the environment. Thus, cells and proteins of the lung provide the first line of defense against this potentially fatal pathogen. One class of proteins that can regulate immune responses in the lung is surfactant proteins. Although classically known for mediating relief of surface tension in alveolar air spaces, two of the four defined surfactant proteins are now established in the literature as possessing immunomodulatory functions. Specifically, surfactant protein A (SP-A) and SP-D, both members of the collectin family of proteins, are able to interact with pathogens and regulate immune responses. During bacterial and viral infections, as well as allergic reactions, SP-A and SP-D have been extensively shown to play protective roles that benefit the host (reviewed in reference 12).

In contradiction to this host-protective paradigm, we have shown that SP-D^−/− mice are less susceptible than wild-type (WT) mice to C. neoformans, and thus SP-D is considered detrimental to the host during C. neoformans infection (13, 14). Our studies have further demonstrated that SP-D^−/− mice survive longer and have lower fungal burden than WT control animals. SP-D can opsonize C. neoformans, but increased phagocytosis has not been demonstrated to lead to increased fungal death. Instead, SP-D was shown to protect these fungi from macrophage-killing mechanisms, including reactive oxygen species (14). These observations led us to conclude that rather than playing a host-protective role, SP-D facilitates C. neoformans pathogenesis by protecting fungi from host immune responses and that this fungal protection was, in part, mediated via direct interaction with fungal cells. These findings prompted further in vivo investigations into the role of SP-D in mediating cryptococcal pathogenesis.

We hypothesized that SP-D exacerbates C. neoformans pathogenesis, at least in part, by dysregulating the early pulmonary immune response, causing heightened inflammation and detrimental cellular and cytokine responses. In the present study, using mice with genetically altered levels of IL-5, eosinophils, and SP-D, we examined the role of IL-5 and eosinophil infiltration during C. neoformans infection. We concluded that SP-D increases susceptibility to pulmonary C. neoformans infection by acting as a natural immune enhancer of infection-driven IL-5 production and eosinophil infiltration in the lung and that augmented levels of IL-5 and eosinophils contribute to increased host susceptibility to C. neoformans.

MATERIALS AND METHODS

Mice.

All mice were maintained in specific-pathogen-free housing at Duke University. SP-D^−/− (15), IL-5TgxSP-D^−/− (IL-5 transgenic [IL-5Tg] mice crossed with SP-D^−/− mice), and corresponding wild-type (WT) control mice (C57BL/6J) were bred in-house. IL-5Tg (NJ.1638) (16) mice (kind gift of James Lee) were also bred in-house, and WT (C57BL/6J) littermates were used as controls. As needed, additional C57BL/6J mice were purchased from Jackson Laboratories (Bar Harbor, ME). For survival studies mice were infected between 6 and 8 weeks of age. For all other studies, mice were infected between 8 and 12 weeks of age. Both males and females were used. The animal protocol was approved by the Institutional Animal Care and Use Committee of Duke University (protocol number A021-12-01). All surgery was performed under anesthesia with ketamine and xylazine, and every effort was made to minimize pain and suffering.

C. neoformans strains.

Mice were infected with H99 Stud, a highly virulent strain of C. neoformans serotype A that was created by passaging a mixed stock of C. neoformans, which included H99#1 (the original human isolate generated by John Perfect) through a rabbit (17). H99 Stud was used because of its effective and reproducible pathogenicity. H99 Stud was stored in a glycerol stock at −80°C. H99 Stud was maintained on yeast extract-peptone-dextrose (YPD) agar for up to 1 month. Liquid cultures were grown for 8 to 18 h at 30°C with shaking (∼250 rpm).

C. neoformans infection.

For aerosol exposures, H99 Stud was grown at 30°C overnight in YPD liquid medium, then subcultured (1:100) for a second night, and inoculated onto V8 agar plates (made with 5% V8 Original Tomato Juice, pH 5.0) in 10-μl spots (about 50 spots/plate) for 2 to 3 weeks. On the day of exposure, fungal colonies were gently removed with a cell scraper, dispersed into sterile phosphate-buffered saline (PBS), centrifuged, and resuspended to 1 × 10⁸ cells/ml in sterile deionized water. C. neoformans was aerosolized using a six-jet Collision Nebulizer (CN25; BGI Incorporated). Mice were exposed to H99 Stud aerosol in a whole-body Madison chamber for 60 min at 95% relative humidity, in a biosafety level 3 (BSL-3) biocontainment facility (NIAID Regional Biocontainment Laboratory at Duke University). After exposure, mice were handled in class III and class II biosafety cabinets connected to the Madison chamber. They were then kept in hermetically sealed cages. The aerosol concentration of C. neoformans was measured via a BioSampler (SKC Incorporated) attached to the Madison chamber as previously described (18).

For intranasal instillation, H99 Stud was prepared as previously described (14). Briefly, C. neoformans cells were washed in sterile saline three times, counted, and resuspended to the desired cell density (between 5 × 10⁴ and 5 × 10⁵ yeast cells per mouse in 25 to 50 μl of saline as specified below). Mice were sedated with a ketamine (Ketaset; Fort Dodge) (150 mg/kg) and xylazine (Anased; Lloyd Incorporated) (10 mg/kg) mixture and hung gently by their incisors on a taut thread. The cell suspension was delivered by careful pipetting into one nostril. Sham-infected mice were given an equivalent volume of sterile saline. The mice were allowed to hang for 5 min after C. neoformans instillation.

For fungal burden studies, histological analysis, survival studies, and flow cytometric analysis, mice were instilled with 5 × 10⁴ C. neoformans cells in 25 to 50 μl of saline. For survival studies, mice were monitored for weight loss and general health per an approved IACUC protocol. When mice lost more than 15% of initial body weight or became moribund, they were humanely euthanized according to an approved IACUC protocol. For bronchoalveolar lavage (BAL) fluid cellularity and cytokine analysis, mice were instilled with a higher dose of 5 × 10⁵ H99 Stud cells as a higher dose elicited greater differences in immune responses of WT and SP-D^−/− mice.

BAL fluid cellularity and cytokine analysis.

On the day of organ harvest, mice were euthanized with pentobarbital (Nembutol; Oak Pharmaceuticals) or with a solution of pentobarbital sodium and phenytoin sodium (Euthasol; Virbac Corporation) and exsanguinated via the renal artery. Lungs were perfused with a sufficient volume of sterile PBS to flush red blood cells (RBCs) (5 to 10 ml) and then lavaged with 0.1 mM EDTA in PBS. A total of 6 ml of lavage fluid was collected in 1-ml aliquots. BAL fluid cells were centrifuged at 350 × g. Supernatant from the first 1 ml was stored at −80°C for cytokine analysis. BAL fluid cells for each sample were pooled, counted, and adhered to slides via centrifuging 200 μl of BAL fluid in a Cytospin 2 (Shandon Incorporated) at 400 rpm for 4 min. A total of 50 to 100,000 cells were used for each cytospin. Cytospin products were air dried, fixed with methanol, and stained with hematoxylin and eosin (H&E) for differential counting.

Cytokine concentrations were measured in cell-free supernatants acquired from the first 1 ml of collected BAL fluid by multiplex assay using a 20-plex mouse inflammatory cytokine and chemokine panel (Invitrogen Life Sciences). For some experiments, IL-5 levels were assessed via enzyme-linked immunosorbent assay (ELISA; BD Biosciences/Pharmingen). Control samples showed comparable results for IL-5 concentration in both the bead-based multiplex assays and the ELISAs.

Fungal burden analysis.

On the day of harvest, mice were euthanized and exsanguinated as described above. For fungal burden analysis, lung tissue was excised and homogenized, and samples were diluted and plated onto YPD agar plates and incubated at room temperature or at 30°C until colonies became large enough to count (usually 48 to 72 h).

Tissue histology.

Histological analyses were performed similarly to those previously described (19). Lungs were gravity inflated with 10% neutral buffered formalin (NBF) and paraffin embedded. Sections of 4 to 5 μm were deparaffinized and stained with hematoxylin and eosin. Blinded samples were scored by a veterinary pathologist for necrosis, hemorrhage, edema, Cryptococcus, and inflammation. Scores ranged from 0 to 8, with the score roughly representing the percentage of tissue involved (where 0 is 0%, 1 is 10%, 2 is 20%, etc.).

Lung tissue digests and immunophenotyping by flow cytometry.

Lung digests for immunophenotyping were performed similarly to previously described methods (20). After perfusion and collection of the BAL fluid, the whole lung was excised and finely chopped with razor blades in a petri dish. Lung tissue was diluted in 4.5 ml of Hanks balanced salt solution (HBSS) with Ca²⁺ and Mg²⁺ and digested with DNase I and Collagenase XI (Roche). Single-cell suspensions were attained by passing the digested material through 40-μm-pore-size strainers, and cells of hematopoietic lineage were enriched by centrifuging the samples in a 4%/14.5% iodixanol gradient (Optiprep; Axis-Shield). Cells were stained with the following antibodies: fluorescein isothiocyanate (FITC)-Gr-1 (clone RB6-8C5; BD Biosciences), phycoerythrin (PE)-Cy5.5 CD11c (clone N418; eBiosciences), allophycocyanin (APC)-CD11b (clone M1/70; eBiosciences), biotin major histocompatibility complex class II (MHC-II) (clone 2G9; BD Biosciences), and PE-Cy7 CD45 (clone 30-F11; BD Biosciences). PE-Texas red-streptavidin (BD Biosciences) was used as the conjugate for biotin-labeled antibodies. Finally, cells were fixed with 10% NBF.

Immunophenotype list mode data were acquired on a BD-LSRII (BD Biosciences). Using Flow software (Treestar), sample data (FCS software, version 3.0) were first analyzed for FSC-H (where FSC is forward scatter and H is height) versus FSC-A (where A is area) to gate on single cells. CD45⁺ hematopoietic-lineage cells were then analyzed for Gr-1, CD11c, and MHC class II expression using bivariate dot plots. Polymorphonuclear leukocytes (PMNs) were defined as expressing high levels of Gr-1 and negative for or expressing only low levels of CD11c (Gr-1^hi CD11c^neg-lo). We also assessed CD11c⁺ MHC-II^int (expressing intermediate levels of MHC-II) cells (which should mainly consist of macrophages) and CD11c⁺ MHC-II^hi cells (which should mainly consist of dendritic cells). This gating strategy is based on the work of Lin et al. (21). Cells that were Gr-1^neg-lo CD11c^neg-lo MHC-II^neg-lo with high side scatter (SSC) were considered to be eosinophils (22, 23). The phenotype of these cells was confirmed by fluorescence-activated cell sorting (FACS) and cytospin analysis with H&E staining. The remaining CD45⁺ cells were designated “lymphocytes/other.” Isotype-matched antibodies were used for control staining.

RNA isolation, reverse transcription (RT), and real-time PCR analysis.

RNA was isolated from lung tissue with TRIzol reagent (Invitrogen) per the manufacturer's protocol, followed by chloroform extraction and isopropanol precipitation. Equal amounts of RNA were reverse transcribed with Moloney murine leukemia virus (MMLV) and random primers. Real-time PCR was performed with the double-stranded DNA probe SYBR green in a two-step reaction on a Bio-Rad CFX96 machine. Results were normalized to β-actin RNA as an internal control and analyzed by the 2^−ΔΔCT (where C_T is threshold cycle) method. The oligonucleotide primers for β-actin were 5′-GATTACTGCTCTGGCTCCTAG-3′ (forward) and 5′-GACTCATCGTACTCCTGCTTGC-3′ (reverse). The oligonucleotide primers for IL-5 were 5′-AGCACAGTGGTGAAAGAGACCTT-3′ (forward) and 5′-TCCAATGCATAGCTGGTGATTT-3′ (reverse).

Statistical analysis.

Statistical analyses were performed with GraphPad Prism software (version 5.0) or SAS Statview software. To compare two groups, data were analyzed by a two-tailed Student's t test. For multiple comparisons, data were analyzed by one-way or two-way analysis of variance (ANOVA) followed by Bonferroni's or Tukey's posttest correction. Survival studies were assessed via Kaplan-Meier survival curve analysis with the Mantel-Cox test. A P value of <0.05 was considered significant.

RESULTS

SP-D^−/− mice are less susceptible than WT mice to C. neoformans in an aerosol exposure model.

SP-D is a constitutively expressed pulmonary protein that is immediately available to interact with C. neoformans after the pathogen enters the lung. Our previous studies used intranasal instillation of inoculum by directly pipetting a suspension of cryptococcal cells into the nostrils of lightly sedated mice. Naturally occurring C. neoformans infections, however, generally arise after inhalation of aerosolized C. neoformans in the environment. Thus, we asked whether a more natural route of exposure to C. neoformans would yield comparable results to intranasal instillation, that is, whether SP-D would facilitate C. neoformans pathogenesis after exposure to aerosolized fungal cells.

WT and SP-D^−/− mice were simultaneously exposed to aerosolized C. neoformans in a Madison whole-body exposure chamber (see Materials and Methods). The average aerosol concentration of three independent exposures was 8 ± 2 CFU/ml. The fungal load was quantified at 1 h, 24 h, 10 days, 14 days, and 21 days postinfection. At early stages of infection (1 h and 24 h), WT and SP-D^−/− mice had similar fungal loads, but as the infection progressed, fungal loads were higher in WT mice than in SP-D^−/− mice, with a significant difference between strains apparent at 14 and 21 days postinfection (Fig. 1A). Histological analysis at day 21 postinfection confirmed increased pathology in WT mice compared to SP-D^−/− mice (Fig. 1B and C and Table 1). These data corroborate that SP-D facilitates C. neoformans pathogenesis across different models of infection and support the continued use of the more accessible intranasal instillation method as a model infection route for C. neoformans in mice.

FIG 1.

SP-D^−/− mice are more resistant than WT mice to aerosol exposure of C. neoformans. (A) Mice were exposed to aerosolized H99 Stud in full-body Madison chambers, and fungal burden was assessed in the lung up to 21 days postinfection. Each time point was repeated at least twice for a total of 7 to 8 mice/group. Representative images from histological analyses of paraffin-embedded and H&E-stained lung tissue from WT mice (B) and SP-D^−/− mice (C) after aerosol exposure are shown. *, P < 0.05; **, P < 0.01.

TABLE 1.

Histological analysis of H&E-stained lung sections 21 days after aerosol challenge

Genotype	Mean histological scorea
	Necrosis	Hemorrhage	Edema	Cryptococcus	Inflammation	Composite
WT	2.7	1.3	1	7.7	6.7	19.3
SP-D−/−	3	1.7	0.3	3.7	4	12.7

^aData are for 2 to 5 mice/group. Sham-infected WT and SP-D^−/− mice received a score of 0 for each category.

SP-D^−/− mice have a decreased pulmonary cellular immune response to C. neoformans compared to WT mice.

To further characterize the effect of SP-D during C. neoformans infection, the pulmonary cellular responses in WT and SP-D^−/− mice were assessed. Using immunophenotyping and polychromatic flow cytometry, myeloid cell infiltration in both BAL fluid and interstitial lung tissue was quantified at 3 and 7 days post-C. neoformans infection. In lung interstitium and BAL fluid, the absolute number of infiltrating CD45⁺ cells (hematopoietic-lineage cells) in both WT and SP-D^−/− mice increased with infection (Fig. 2A and B). This CD45⁺ cell lung infiltration was statistically lower in infected SP-D^−/− mice than in infected WT control mice. A similar trend was observed in BAL fluid. Detailed analysis of the frequency of myeloid cell populations within the lung CD45⁺ compartment revealed that the difference between infected WT and SP-D^−/− mice could not be accounted for by differences in the proportion of polymorphonuclear cells (PMNs) (Fig. 2C and D), CD11c⁺ MHC-II^int cells (likely macrophages) (Fig. 2E and F), or CD11c⁺ MHC-II^hi cells (likely dendritic cells) (Fig. 2G and H). However, a large population of MHC-II^neg-low/CD11c^neg-low/SSC^hi cells was observed that was lower in infected SP-D^−/− mice than in infected WT mice (data not shown). Since this population of cells was undetermined by our flow cytometry panels, additional cytospin analysis with H&E staining was conducted and revealed that these cells were eosinophils (Fig. 3E to H). Thus, eosinophils account for the majority of the CD45⁺ pulmonary cell influx post-C. neoformans infection, and SP-D^−/− mice had reduced frequencies of eosinophil infiltration compared to WT mice. SP-D^−/− mice also had lower absolute numbers of eosinophils after 7 days of C. neoformans infection than WT mice in the interstitium (24 × 10⁵ ± 5 × 10⁵ cells in WT mice and 5 × 10⁵ ± 1 × 10⁵ cells in SP-D^−/− mice) and BAL fluid (9 × 10⁵ ± 3 × 10⁵ cells in WT mice and 7 × 10⁵ ± 3 × 10⁵ cells in SP-D^−/− mice), which further endorsed the substantial differences between WT and SP-D^−/− mice. WT and SP-D^−/− mice did not have significantly different frequencies of the remainder of CD45⁺ cells, which were primarily lymphocytes.

FIG 2.

Total CD45⁺ cells are decreased in SP-D^−/− mice compared to WT mice after C. neoformans infection. Flow cytometric analysis of total CD45⁺ cells in WT and SP-D^−/− mice after C. neoformans infection in lung interstitium (day 3, n = 5 to 9 mice/group; day 7, n = 14 to 21 mice/group) (A) and lung BAL fluid (day 3, n = 1 to 2 samples/group; day 7, n = 5 to 8 samples/group; samples were pooled and normalized to per mouse cell counts) (B). Shown in panels A and B are absolute numbers of cells per mouse calculated from total cellularity obtained and the percentage of CD45⁺ cells. Frequency of PMNs in lung interstitium (C) and BAL fluid (D), of CD11c⁺ MHC-II^int cells in lung interstitium (E) and BAL fluid (F), of CD11c⁺ MHC-II^hi cells in lung interstitium (G) and BAL fluid (H), and of lymphocytes/other CD45⁺ cells in lung interstitium (I) and BAL fluid (J) are shown. In panels C to J, data are shown as percentages of CD45⁺ cells. *, P < 0.05 for WT infected versus SP-D^−/− infected mice.

FIG 3.

IL-5 production and eosinophil infiltration are reduced in infected SP-D^−/− mice compared to levels in WT mice. WT and SP-D^−/− mice were infected intranasally with 5 × 10⁵ H99 Stud cells. (A) Eosinophils in BAL fluid were assessed at 7 days postinfection via cytospin and H&E staining (n = 3 to 11 mice/group). (B) IL-5 protein levels were measured at 3 days postinfection in BAL fluid by ELISA (n = 8 to 10 mice/group). (C) IL-5 mRNA steady-state level was measured on days 1 to 3 postinfection in postlavage lung tissue (n = 4 to 10 mice/group). (D) Mice were administered IL-5 neutralizing antibodies at the time of infection. Eosinophils were assessed in BAL fluid at 7 days postinfection via cytospin and H&E staining. For the IL-5 blockade, data are shown as frequency of the respective isotype-treated control (n = 12 to 14 mice/group). *, P < 0.05; **, P < 0.01; ***, P < 0.001. Photomicrographs show representative cytospins used for analysis of saline-treated WT mice (E), saline-treated SP-D^−/− mice (F), C. neoformans-infected WT mice (G), and C. neoformans-infected SP-D^−/− mice (H). Arrowheads indicate macrophages, and arrows indicate eosinophils.

To confirm the findings of the flow cytometric analysis, eosinophils were quantified by classical cytospin/histologic analysis of BAL fluid cells. Eosinophil counts were significantly lower in infected SP-D^−/− mice than in infected WT mice at day 7 (Fig. 3A). The resultant eosinophilia values were lower by cytospin analysis than they were by FACS analysis; this is because the design of the flow cytometry panel tends to underestimate the number of eosinophils. Notably, the fold differences were similar for the two methods (i.e., WT mice had more eosinophils than SP-D^−/− mice, postinfection). Thus, we concluded that eosinophils dominate the lung environment after C. neoformans infection and that SP-D enhances eosinophilia during C. neoformans infection in both the interstitial and alveolar compartments of the lung.

SP-D^−/− mice have altered pulmonary cytokine/chemokine responses to C. neoformans compared to WT mice.

To determine the effect of SP-D on the lung inflammatory cytokine response to C. neoformans, we quantified levels of 20 cytokines in BAL fluid from WT and SP-D^−/− mice at 7 days postinfection (Table 2). In WT mice, C. neoformans infection stimulated higher production of gamma interferon (IFN-γ), tumor necrosis factor alpha (TNF-α), IL-4, IL-5, IL-6, IL-12, IL-13, IP-10, and monocyte chemotactic protein 1 (MCP-1) than in sham-infected WT mice. Of these infection-induced cytokines/chemokines, IFN-γ, IL-4, IL-6, and IL-13 were produced at comparable levels in SP-D^−/− mice, suggesting that SP-D does not alter the production of these immune mediators 7 days post-C. neoformans infection.

TABLE 2.

Levels of BAL specimen cytokines and chemokines (pg/ml) 7 days postinfection^a

Cytokinef	Mean concn for the group (± SE [pg/ml])a				Fold change in:
	WT control	SP-D−/− control	WT infected	SP-D−/− infected	WT mice	SP-D−/− mice
IFN-γ	9.2 (0.1)	12 (2)	50 (10)b	49 (8)b	6.0e	4.2
TNF-α	10.1 (0.5)	10.1 (0.5)	20 (3)b	16 (2)b,c	2.0	1.6
IL-4	16.2 (0.3)	16.1 (0.3)	110 (20)b	160 (20)b	7.0	9.8
IL-5	9.3 (0.07)	9.3 (0.07)	310 (40)b	550 (70)b,c	33.8e	59.2
IL-6	14 (1)	14 (1)	130 (20)b	200 (30)b	9.4	13.9
IL-12	9 (2)	40 (8)d	60 (8)b	81 (7)b	6.7e	2.0
IL-13	17 (2)	17 (2)	800 (80)b	800 (200)b	47.1e	49.1
IL-17	5 (1)	5 (1)	7.4 (0.6)	9 (1)b	1.4	1.6
IP-10	8 (2)	7 (2)	57 (6)b	140 (10)b,c	7.3e	18.8
MCP-1	19 (2)	24 (5)	300 (30)b	530 (50)b,c	16.0	22.2
MIG	5 (2)	7 (3)	29 (6)	42 (9)b	5.5	6.4
MIP-1α	27 (8)	29 (7)	120 (20)	190 (40)b	4.6	6.7
VEGF	90 (10)	160 (10)d	30 (8)b	50 (20)b	0.4	0.3
FGF basic	40 (9)	90 (20)	100 (20)	170 (30)	2.5	1.9
KC	180 (60)	230 (50)	94 (5)	130 (40)	0.6	0.6

^aData are from 10 to 15 mice/group and from three to four independent experiments.

^bValues for infected mice are significantly different than for the respective saline-treated control mice.

^cValues for infected SP-D^−/− mice are significantly different than for infected WT mice.

^dBaseline values for SP-D^−/− mice are significantly different than for WT mice.

^eFold change is significantly different in WT versus SP-D^−/− mice.

^fVEGF, vascular endothelial growth factor; FGF, fibroblast growth factor.

Of interest, IL-5, IP-10, and MCP-1 levels were significantly higher in SP-D^−/− mice than in WT mice after 7 days of infection, revealing a possible mechanism by which the SP-D protein may cause suppression of some cytokines/chemokines during C. neoformans infection. Additionally, C. neoformans infection stimulated production of monokine induced by IFN-γ (MIG) and macrophage inflammatory protein 1α (MIP-1α) specifically in SP-D^−/− mice. Since neither was detected in WT mice, this finding suggests that SP-D suppresses certain host cytokine/chemokine responses in fungal infection. Conversely, C. neoformans infection stimulated TNF-α production in WT mice significantly more than in SP-D^−/− mice, showing that SP-D might enhance TNF-α expression. IL-17, keratinocyte-derived chemokine (KC), granulocyte-macrophage colony-stimulating factor (GM-CSF), IL-1α, IL-1β, IL-2, and IL-10 were not induced by C. neoformans infection (data not shown for GM-CSF, IL-1α, IL-1β, IL-2, and IL-10). Overall, cytokine/chemokine analysis did not reveal a clear bias toward Th1 or Th2 responses; rather, SP-D has a dysregulating, sometimes stimulating and sometimes suppressing, effect on certain chemokines and Th1 cytokines at 7 days postinfection.

Interestingly, higher IL-5 levels in SP-D^−/− mice than in WT mice at 7 days postinfection was incongruent with our cellular analyses showing higher eosinophilia in WT mice. We postulated that there were earlier increases in IL-5 that could account for the elevated eosinophilia observed at 7 days postinfection. Thus, IL-5 levels were quantified at 3 days postinfection in both strains of mice. As predicted, BAL fluid IL-5 was highly induced in WT mice at 3 days postinfection compared to SP-D^−/− mice (Fig. 3B). This early rise in lung IL-5 protein levels in WT mice corresponded with increased IL-5 steady-state mRNA levels in lung tissue of WT mice during day 2 postinfection (Fig. 3C). This early induction of IL-5 expression levels was not detected in SP-D^−/− mice. Additionally, it should be noted that multiplex analysis was performed at day 3 for the same cytokine panel as day 7, and we did not observe any significant differences that seemed as relevant to eosinophilia as the results with IL-5 (data not shown).

To determine whether IL-5 was necessary for the observed eosinophil recruitment in this model, IL-5 neutralizing antibodies were administered by intraperitoneal (i.p.) injection on the day of infection, and eosinophilia was measured at 7 days postinfection (Fig. 3D). Eosinophilia induced by C. neoformans infection was drastically reduced in both WT and SP-D^−/− mice with antibody administration, thus affirming the critical importance of IL-5 for eosinophil recruitment during acute C. neoformans infection. Together these data demonstrate that delayed IL-5 induction accounts for decreased eosinophilia in SP-D^−/− mice compared to WT mice.

Mice overexpressing IL-5 and eosinophils are more susceptible to C. neoformans than WT mice.

Having established that IL-5 and eosinophil levels are augmented more in WT mice than in SP-D^−/− mice during the early phase of infection, it was next necessary to define the role of IL-5 and eosinophils during cryptococcal pathogenesis in relationship to SP-D. Thus, IL-5 transgenic (IL-5Tg) mice that overproduce IL-5 under a T cell-specific promoter and therefore have a constitutive overabundance of eosinophils in the peripheral blood were infected with 5 × 10⁴ H99 Stud cells for histological studies, pulmonary fungal burden comparisons, and survival analyses (16). As predicted, after C. neoformans infection, IL-5Tg mice displayed high overexpression of both IL-5 and eosinophils in BAL fluid (see Fig. S1 in the supplemental material). At 7 days postinfection, fungal burden was significantly higher in IL-5Tg mice (Fig. 4A). Histological analysis confirmed the fungal burden analysis and revealed that IL-5Tg mice had increased necrosis and inflammation compared to WT mice (Table 3; see also Fig. 6). Finally, IL-5Tg mice died more rapidly in survival studies than WT control mice (Fig. 4B). These data indicated that augmentation of IL-5 production and resulting eosinophilia contributed significantly to C. neoformans pathogenesis.

FIG 4.

IL-5-overexpressing mice are more susceptible to C. neoformans than WT mice. (A) Pulmonary fungal burden at 7 days post-C. neoformans infection in WT and IL-5Tg mice. ***, P < 0.001. (B) Survival analysis of WT and IL-5Tg mice (two independent experiments for a total of at least 18 mice/group; P < 0.0001).

TABLE 3.

Histological analysis of H&E-stained lung sections at 7 days postinfection

Genotype	Mean histological scorea
	Necrosis	Hemorrhage	Edema	Cryptococcus	Inflammation	Composite
WT	2.3	1.3	0.8	4.3	4.0	12.8
SP-D−/−	1.4	1.0	0.0	3.6	2.6	8.6
IL-5Tg	3.3	2.3	1.6	6.5	5.8	19.4
IL-5TgxSP-D−/−	1.7	1.0	0.7	3.0	3.0	9.3

^aData are for 2 to 5 mice/group. All sham-infected mice received a score of 0 for each category except IL-5Tg mice, which received a mean score of 0.3 for inflammation, and IL-5TgxSP-D^−/− mice, which received a mean score of 0.5 for hemorrhage and 1 for inflammation.

FIG 6.

Representative photomicrographs from H&E-stained lung sections 7 days after C. neoformans infection: WT sham-infected mice (A), WT C. neoformans-infected mice (B), SP-D^−/− sham-infected mice (C), SP-D^−/− C. neoformans-infected mice (D), IL-5Tg sham-infected mice (E), IL-5Tg C. neoformans-infected mice (F), IL-5TgxSP-D^−/− sham-infected mice (G), and IL-5TgxSP-D^−/− C. neoformans-infected mice (H). Arrows indicate C. neoformans cells; the cell wall of C. neoformans is stained by hematoxylin.

IL-5 overexpression reverses the SP-D^−/− response to C. neoformans.

We next examined the in vivo linkage between IL-5 and SP-D by asking whether augmenting IL-5 levels in SP-D-deficient mice could reverse the SP-D^−/− phenotype and result in increased susceptibility to C. neoformans infection. We created IL-5TgxSP-D^−/− mice and assessed fungal burden, histology, and survival in these mice. Cytokine and cytospin analysis confirmed high levels of eosinophils and IL-5 at 7 days postinfection in IL-5TgxSP-D^−/− mice (see Fig. S1 in the supplemental material). At 14 days postinfection, the fungal burden in IL-5TgxSP-D^−/− mice was higher than that SP-D^−/− mice, lower than that of IL-5Tg mice, and similar to that of WT mice (Fig. 5A). Survival analysis, likewise, revealed that IL-5TgxSP-D^−/− mice were significantly more susceptible to C. neoformans than SP-D^−/− mice, less susceptible than IL-5Tg mice, and not statistically different from WT mice (Fig. 5B). Histological analysis confirmed that IL-5TgxSP-D^−/− mice were less susceptible to C. neoformans than IL-5Tg mice as they had lower levels of hemorrhage, necrosis, edema, cryptococcal cells, and inflammation (Table 3 and Fig. 6). Together, these data demonstrated that overexpression of IL-5, and thus eosinophils, restores SP-D^−/− mice to the WT phenotype and supports the conclusion that SP-D protein augments C. neoformans disease in the murine host by promoting excessive IL-5 expression and eosinophil infiltration in response to this fungal infection.

FIG 5.

IL-5 overexpression increases the susceptibility of SP-D^−/− mice to C. neoformans. (A) Pulmonary fungal burden at 14 days post-C. neoformans infection in WT, IL-5Tg, SP-D^−/−, and IL-5TgxSP-D^−/− mice. *, P < 0.05; **, P < 0.01; ***, P < 0.001. (B) Survival analysis of WT, IL-5Tg, SP-D^−/−, and IL-5TgxSP-D^−/− mice (two independent experiments for a total of at least 3 mice/group; for the WT versus SP-D^−/−, P <0.001; for the WT versus IL-5TgxSP-D^−/−, not significant; for IL-5TgxSP-D^−/− versus SP-D^−/−, P < 0.05; and for IL-5TgxSP-D^−/− versus IL-5Tg, P < 0.001).

DISCUSSION

SP-D is well documented as a host-protective immunoregulatory protein, yet our previous studies have demonstrated that SP-D fails to protect mice against C. neoformans by promoting fungal growth and survival to the detriment of the host (13, 14). The uniqueness of the interaction between SP-D and C. neoformans prompted further investigation into the mechanisms by which SP-D affects host outcome in the face of C. neoformans infection. Because of the immunoregulatory role of SP-D, we hypothesized that one mechanism by which SP-D increases susceptibility to C. neoformans may be through dysregulation of host immune responses. Analysis of C. neoformans infection in WT and SP-D^−/− mice revealed that SP-D increases inflammatory cell infiltrate and dysregulates the balance of pulmonary cytokines in response to C. neoformans. We specifically identified a role for SP-D in augmenting eosinophils and IL-5, an eosinophil recruitment and survival cytokine, during active C. neoformans infection. Select mutant mouse models were used to further examine the roles of IL-5, eosinophils, and SP-D during C. neoformans pathogenesis. Together these studies demonstrate a dynamic interplay between SP-D, IL-5, and eosinophils in determining the outcome of C. neoformans disease during the critical early phase of pulmonary infection.

To mimic the natural route of infection, we examined the ability of SP-D to facilitate cryptococcal pathogenesis after exposure to aerosolized C. neoformans cells. WT mice were more susceptible to aerosol exposures of C. neoformans than SP-D^−/− mice (Fig. 1). This confirmed that the ability of SP-D to affect C. neoformans disease is free of artifact due to the intranasal instillation method, corroborating our previous survival and fungal burden analyses with intranasal infections and firmly establishing that SP-D increases susceptibility to C. neoformans infection (13). This conclusion is further supported by models using inducible SP-D mice and via rescue studies in which SP-D protein instilled in the nostrils of SP-D^−/− mice increased susceptibility of SP-D^−/− mice to the levels of WT mice (13). These findings clearly validate the intranasal instillation method as a biologically relevant and robust model for early pulmonary C. neoformans disease, and, thus, in the immunological studies that followed, we used the traditional and more readily accessible intranasal method.

Immunophenotyping revealed C. neoformans induced higher levels of inflammatory (CD45⁺) cells in WT mice than in SP-D^−/− mice (Fig. 2), indicating that SP-D contributes to C. neoformans pathogenesis by exacerbating overexuberant immune responses that might be detrimental to the host. This was in agreement with our hypothesis. Furthermore, the presence of SP-D in WT mice was associated with increased eosinophil recruitment (Fig. 3). This is in direct contrast with the mold Aspergillus fumigatus, a medically relevant fungal species that has been studied extensively in the context of SP-D, in which the main role of SP-D is to suppress excessive inflammation, including eosinophilia, thereby reducing damage to the host (24–28). Thus, we postulate that while SP-D protects against A. fumigatus by dampening and thus preventing overexuberant immune responses and excessive eosinophilia after A. fumigatus exposure, it conversely contributes to C. neoformans pathogenesis by augmenting eosinophilic responses to C. neoformans, leading to excessive inflammation and potentially damaging the surrounding host tissue. The histology from this study and from previous studies supports this hypothesis (13) as WT mice displayed more extensive damage within the pulmonary airspaces than SP-D^−/− mice (Fig. 1 and Table 1).

In addition to analysis of the cellular immune response, we also examined cytokine/chemokine responses after C. neoformans infection. Multiple proinflammatory markers were induced by C. neoformans infection in both WT and SP-D^−/− mice, with several key chemokines and Th1 cytokines expressed to a lesser extent in WT mice (Table 2). These observations indicate that SP-D may contribute to the susceptibility of WT mice by suppressing key inflammatory cytokines and chemokines, such as IL-12, IP-10, MIP-1α, MIG, and MCP-1, which are normally protective against C. neoformans (4–6, 29–34).

Following the immune phenotyping observations of eosinophil differences in the lung interstitial and lavage compartments, specific examination of IL-5 levels showed delayed IL-5 induction in SP-D^−/− compared to WT mice (Fig. 3 and Table 2). These data revealed a dynamic role for SP-D in modulating IL-5 expression, and we postulated that a delayed IL-5 response in SP-D^−/− mice resulted in decreased eosinophilia, as detected 7 days after infection. Furthermore, IL-5 blockade studies presented herein confirm that IL-5 is required for eosinophilia at an early stage of acute infection. This is in accordance with previous work in which IL-5 was also necessary for eosinophilia in a chronic C. neoformans infection model (8). During chronic C. neoformans infection, CD4⁺ T cells were critical producers of IL-5, but we expect that other cell types are a more significant source of IL-5 during the initial stages of an acute response. Potential candidates include lung stromal cells, such as epithelial cells (35, 36), and innate immune cells, such as mast cells, monocytes, or a newly identified IL-5 producing non-T lymphoid cell type that can rapidly produce this cytokine (37–40).

The role of IL-5 and eosinophils during C. neoformans pathogenesis has remained elusive. Along with the widely published fact that eosinophilia has often been associated with ineffective immune responses against C. neoformans, it has been shown that eosinophils bind C. neoformans in the presence of opsonizing antibodies (41), that late-phase C. neoformans-induced pulmonary eosinophilia is IL-5 dependent (8), and that eosinophils have C. neoformans-derived antigen-presenting capabilities (42, 43). Recently eosinophils were found to be important producers of IL-4 and thus are contributors to nonprotective Th2 responses (44). It has remained unclear whether IL-5 and eosinophils are mere bystanders during an otherwise pathogenic Th2 process or whether they were active contributors to C. neoformans disease. The present studies help to elucidate the role of eosinophils and IL-5 in C. neoformans pathogenesis in the context of SP-D. Overexpression of IL-5 and the large, resultant eosinophil population significantly increased host susceptibility to C. neoformans (Fig. 4). This is in contrast with recent bacterial studies that demonstrated that IL-5Tg mice were protected against bacterial infection due to bactericidal capabilities of eosinophils (45). Furthermore, studies with IL-5TgxSP-D^−/− mice determined that augmenting expression of IL-5 and eosinophils in SP-D^−/− mice was sufficient to restore susceptibility to C. neoformans, providing strong evidence that decreased IL-5 and eosinophils are an important part of explaining the mechanistic role of SP-D in promoting C. neoformans pathogenesis (Fig. 5). However, we did examine whether components of these interactions had an impact on the phenotype. When fungal burden and survival were additionally assessed in PHIL (eosinophil-deficient) and IL-5^−/− mice after C. neoformans infection (data not shown), we found no impact and minimal impact, respectively, on outcomes. Taken together, these data suggest that IL-5, eosinophils, and SP-D work together to determine host fate in C. neoformans disease and highlight the importance of balance in host cytokine production, collectin presence, and host cell influx to the site of infection during specific microbial assaults.

In the present study, SP-D acts as a natural enhancer of C. neoformans-driven IL-5 production and eosinophil infiltration in the lung, leading to increased host susceptibility. This model is in contrast to A. fumigatus (24–27, 46), in which SP-D plays a host-protective role by reducing Th2 responses. The ability of SP-D to act as a suppressor of IL-5 and eosinophilia in A. fumigatus models while acting as an enhancer of IL-5 and eosinophilia during C. neoformans infection can be explained by the dual role of SP-D in inflammation. SP-D can act as an anti-inflammatory molecule when present as a dodecamer that interacts with signal regulatory protein alpha (SIRPα). Conversely, SP-D can act as a proinflammatory molecule leading to NF-κB activation and proinflammatory signaling cascades after S-nitrosylation, trimer formation, and interaction with CD91/calreticulin (47). Previous work showing that oligomeric SP-D inhibits eosinophil chemotaxis, while S-nitrosylated SP-D does not, supports this dual-role paradigm for SP-D in eosinophilia (48, 49). Future studies should focus on gaining a deeper understanding of the dual role of SP-D in infection and inflammation by examining the oligomeric structure and posttranslational modifications of SP-D during both C. neoformans and A. fumigatus infections. Although both are pathogenic fungi, they diverge widely in nucleic acid sequence, morphology, pathogenic features, and host susceptibility. Comparing the status of SP-D protein in both models should lend significant insight into fundamental differences about how SP-D differentially affects host outcome in the face of two divergent types of fungal infections.

In conclusion, the effects of SP-D, IL-5, and eosinophils are interwoven to establish the early innate immune response to this encapsulated yeast and promote fungal pathogenesis. Future studies focused on disrupting the interactions of this triad of factors could suggest potential therapeutic strategies leading to disease prevention.