|Home | About | Journals | Submit | Contact Us | Français|
Cryptococcus neoformans is a fungal pathogen causing pulmonary infection and a life-threatening meningoencephalitis in human hosts. The fungus infects the host through inhalation, and thus, the host response in the lung environment is crucial for containment or dissemination of C. neoformans to other organs. In the lung, alveolar macrophages (AMs) are key players in the host lung immune response, and upon phagocytosis, they can kill C. neoformans by evoking an effective immune response through a variety of signaling molecules. On the other hand, under conditions not yet fully defined, the fungus is able to survive and proliferate within macrophages. Since the host sphingosine kinase 1 (SK1) regulates many signaling functions of immune cells, particularly in macrophages, in this study we determined the role of SK1 in the host response to C. neoformans infection. Using wild-type (SK1/2+/+) and SK1-deficient (SK1−/−) mice, we found that SK1 is dispensable during infection with a facultative intracellular wild-type C. neoformans strain. However, SK1 is required to form a host lung granuloma and to prevent brain infection by a C. neoformans mutant strain lacking the cell wall-associated glycosphingolipid glucosylceramide (Δgcs1), previously characterized as a mutant able to replicate only intracellularly. Specifically, in contrast to those from SK1/2+/+ mice, lungs from SK1−/− mice have no collagen deposition upon infection with C. neoformans Δgcs1, and AMs from these mice contain significantly more C. neoformans cells than AMs from SK1/2+/+ mice, suggesting that under conditions in which C. neoformans is more internalized by AMs, SK1 may become important to control C. neoformans infection. Indeed, when we induced immunosuppression, a host condition in which wild-type C. neoformans cells are increasingly found intracellularly, SK1−/− survived significantly less than SK1/2+/+ mice infected with a facultative intracellular wild-type strain, suggesting that SK1 has an important role in controlling C. neoformans infection under conditions in which the fungus is predominantly found intracellularly.
Cryptococcus neoformans is the etiological agent of the most common form of fungal meningoencephalitis worldwide in immunocompromised individuals. Upon environmental exposure, desiccated yeasts or basidiospores are inhaled into the alveolar spaces of the host lung. In the lung, C. neoformans can survive and replicate in the extracellular environment of alveolar spaces and/or, following phagocytosis, intracellularly within the phagolysosome of the alveolar macrophages (AMs) (13). Following internalization of C. neoformans, AMs initiate an immune response resulting in granuloma formation, which contains the infection in the lung, thereby preventing fungal dissemination to other organs or tissues. Hence, the high incidence of cryptococcal meningitis in patients with impaired cell-mediated immunity clearly links dissemination from the lungs to the lack of or impaired formation of a granuloma. The host cellular pathway(s) controlling the formation of a granuloma response has not been elucidated, mainly because a granulomatous response is very weakly evoked in mouse models of cryptococcosis. Recent studies, however, using a mutant strain of C. neoformans lacking the glucosylceramide synthase 1 (GCS1) gene have shown that immunocompetent but not immunodeficient mice are able to produce a granuloma that successfully contains this mutant strain within the lung tissue, thus preventing its dissemination to the brain and the development of meningoencephalitis (24, 36).
Sphingosine kinases 1 and 2 (SK1 and SK2) are two enzymes of the sphingolipid pathway that phosphorylate sphingosine to produce sphingosine-1-phosphate (S1P), a bioactive signaling molecule. In mammalian cells, S1P modulates numerous cellular processes, including immune cell development, differentiation, activation, and migration (14, 25). The predominant isoform expressed in the lung is SK1, and the regulation of SK1 activity is thought to have a central role in the lung immune response (1, 4, 31). For instance, S1P levels increase in the bronchoalveolar lavage (BAL) fluid upon antigen challenge (2); S1P modulates pulmonary epithelial and endothelial cell function, including the expression of adhesion molecules required for immune cell migration (35); and binding of S1P to their respective cell surface receptors (S1PRs) stimulates SK1 activity and the proinflammatory response in macrophages (20, 30, 32, 50). Importantly, SK1 also has a role in the antimicrobial activities of AMs against internalized pathogens, such as the Mycobacterium species, where its activity is required for phagosome maturation of the pathogen-containing vesicle into the microbicidal phagolysosome (17, 29, 46). Exogenous S1P has also been found to induce a Th1-associated phenotype, intracellular killing, and antigen presentation by human monocytes and macrophages containing intracellular Mycobacterium in vitro (16, 17, 40) and ex vivo (16). Furthermore, intravenous administration of S1P decreases bacterial burden and improves histopathology in the lungs of Mycobacterium-challenged mice during acute infection (16, 17, 39). Whereas these data clearly show that SK1 regulates facets of host inflammation in vitro and that its product, S1P, evokes antimycobacterial actions in vitro and in vivo, the role of SK1 during pulmonary infection in vivo and its effect on the antimicrobial actions of AMs against other facultative intracellular pathogens, such as C. neoformans, are unknown.
In this research, the effect of SK1 on the host immune response during pulmonary cryptococcosis and on the virulence of C. neoformans was examined. Using mice deficient in SK1, we obtained data demonstrating that SK1 modulates the host immune response involved in the formation of granulomas. We also show that SK1 is vital for the containment of C. neoformans during pulmonary infection but only when the fungus predominantly replicates intracellularly.
Five- to 7-week-old wild-type C57BL/6J mice (The Jackson Laboratory, Bar Harbor, ME), SK1-deficient mice (SK1−/−), and SK2-deficient mice (SK2−/−) were used for this research. SK1−/− and SK2−/− mice were previously generated and colonies were maintained as previously described (1). SK1−/− and SK2−/− mice were available to us through the MUSC COBRE Animal Core Facility, directed by T. Kawamori, who provided breeding pairs. Travis J. McQuiston performed all breeding, weaning, and genotyping (data not shown). For all experiments, SK1−/− and SK2−/− mice were age and sex matched with SK1/2+/+ wild-type mice (C57BL/6).
AMs were isolated from the lungs of mice using 1× sterile phosphate-buffered saline (PBS) at pH 7.0 by bronchoalveolar lavage (BAL). BAL fluid was subjected to centrifugation at 500 × g for 5 min. Cell pellets were resuspended in serum-free RPMI medium supplemented with 0.1% penicillin-streptomycin, and cell number was determined using a hematocytometer. For all coincubation assays, 1 × 105 cells were plated on the glass portion of a poly-d-lysine-coated glass-bottom confocal cell dish (MatTek Corporation, Ashland, MA). AMs were allowed to adhere for 30 min before the cell dishes were washed three times and fresh medium was added.
C. neoformans var. grubii serotype A strain H99 (wild type [WT]) and a mutant C. neoformans strain lacking GCS1 (Δgcs1) (36, 42) were used in this study. Both strains were grown in yeast extract-peptone-dextrose (YPD) medium for 16 to 18 h at 30°C in a shaking cell culture incubator.
mRNA was isolated from AMs using the RNeasy minikit from Qiagen. cDNA was generated from 0.5 μg RNA by using random hexamer primers and the SuperScript III First Strand cDNA synthesis system from Invitrogen (Carlsbad, CA). Real-time RT-PCR was conducted using a Bio-Rad iCycler to quantify mRNA levels of SK1 and SK2 or of the S1PRs, respectively. The standard real-time RT-PCR volume was 25 μl, which comprised 12.5 μl SYBR green PCR reagents (Bio-Rad, Hercules, CA), 5 μl cDNA template, 1 μl forward primer (4 μM), 1 μl reverse primer (4 μM), and 5.5 μl water. The sequences of primer pairs for SK isoforms, along with the RT-PCR steps for amplification, were described previously (3, 44). All reactions were performed in triplicate. Q-Gene software was used to analyze data, which were then expressed as the change (fold) between the mean normalized expression and the control value. The mean normalized expression is directly proportional to the amount of mRNA of the target gene relative to the amount of mRNA of the reference gene, the β-actin gene. Melting curves were also examined to ensure that the data corresponded to production of the single desired RT-PCR fragment for each target gene. Data are the averages from three separate experiments.
AMs were plated as described above. C. neoformans cultures were subjected to centrifugation at 500 × g for 10 min. YPD medium was removed, and the cell pellet was washed three times with sterile water. After washing, C. neoformans cells were resuspended in the desired cell medium, and cell number was calculated using a hematocytometer. Next, 1 × 106 C. neoformans cells were opsonized in 1 ml (final volume) RPMI medium containing either 10% fresh mouse serum, 10 μg/ml of the anti-glucuronoxylomannan (GXM) monoclonal IgG1 antibody 18B7 (kindly provided by Arturo Casadevall, Albert Einstein College of Medicine, Bronx, NY), or both. Serum was obtained from C57BL/6J mice immediately before the assay. These opsonization solutions were then vortexed vigorously and incubated for 20 min at 37°C. After opsonization of C. neoformans, the medium from the confocal dishes containing the AMs was removed and replaced with 100 μl opsonized C. neoformans solution containing 1 × 105 C. neoformans cells, thereby making the multiplicity of infection (MOI) 1:1. After 2 h of coincubation, the medium was removed, the plates were washed three times with PBS, fixed in ice-cold methanol, and stained with Giemsa for analysis by light microscopy using a 100× objective under oil immersion. For each confocal dish, a minimum of 500 macrophages were examined for C. neoformans internalization. As previously described, the phagocytic index is the percentage of macrophages with internalized fungal cells multiplied by the average number of internalized fungal cells (45).
The ability of internalized C. neoformans cells to replicate in a coculture within AMs was examined after 4 h of coincubation. C. neoformans opsonized cells with 10% fresh mouse serum and the anti-GXM IgG1 antibody 18B7 were coincubated with AMs, as described above for the in vitro phagocytosis assay. To determine the intracellular growth of C. neoformans, the medium from coincubations was removed after 2 h and plates were washed three times to remove any extracellular C. neoformans. Fresh serum-free RPMI medium was added for an additional 2 h. After a total of 4 h, plates were processed for light microscopy to allow visualization of daughter cells, also known as buds. A minimum of 100 internalized C. neoformans cells per plate were inspected for budding, and intracellular growth was calculated as the percentage of total internalized C. neoformans cells that exhibited buds.
Mice received intraperitoneal injections of 125 mg/kg of cortisone acetate (CA; Sigma Chemical Co., St. Louis, MO) 24 h prior to, immediately prior to, and at days 1, 4, and 7 after intranasal challenge with 5 × 105 C. neoformans cells. The CA regimen was carried out as described in previous studies examining pulmonary fungal infections and virulence (5, 8). Control animals received CA injections but were not challenged with C. neoformans.
Blood samples were collected from CA-treated mice on the same days immediately prior to CA treatments to assess the effect of the corticosteroid on the white blood cell (WBC) population using C. neoformans infection. Blood was obtained by lancing the saphenous vein with a 23 1/2-gauge needle, and approximately 100 μl of blood was collected from each mouse. The blood from three mice was pooled, totaling 300 μl in an EDTA-coated Microtainer collection tube (catalog no. 365973; Becton Dickinson, Franklin Lakes, NJ). A complete blood count (CBC) was conducted on each pooled sample to determine the total WBC count.
Mice were anesthetized with an intraperitoneal injection of 60 μl of a xylazine-ketamine mixture containing 5 mg xylazine and 95 mg ketamine per kg body weight. All strains of C. neoformans were grown in YPD medium for 16 to 18 h at 30°C. C. neoformans cells were washed and resuspended in PBS. Mice were challenged intranasally with 20 μl of the inoculum solution containing 5 × 105 C. neoformans cells. Mice were fed ad libitum and monitored twice a day for signs of appearing moribund or in pain and for any clinical signs suggesting meningoencephalitis. Mice showing any of these signs were immediately sacrificed using CO2 inhalation followed by cervical dislocation.
Organs were harvested and fixed overnight in 37% formaldehyde, paraffin embedded, sectioned, and stained with hematoxylin and eosin (H&E), Movat, mucicarmine, Verhoeff-Van Gieson (VVG), or Giemsa for identification and visualization of host components and C. neoformans cells using light microscopy. The Histology Core Facility at the Medical University of South Carolina performed all staining and tissue processing. Russell Harley and Masha Bilic performed histological analyses on separate occasions.
Data from each experimental group were subjected to an analysis of normality and variance. The statistical significance of the difference between the means of two experimental data sets composed of normally distributed values was analyzed using Student's two-tailed t test. Nonparametric evaluation of independent data sets was performed with the Mann-Whitney rank sum test (also known as the Wilcoxon rank sum test). For both types of analyses, P values less than 0.05 were considered significant.
Previous studies have suggested that SK1 activity is essential to the killing of internalized Mycobacterium cells by human monocytes and macrophages in vitro (29, 46). Thus, we wondered whether SK1 has a role during C. neoformans infection, as C. neoformans is a facultative intracellular pathogen.
To determine if SK1 affects host susceptibility to cryptococcosis, SK1/2+/+ and SK1−/− mice were infected intranasally with 5 × 105 C. neoformans cells and monitored for morbidity. This well-established mouse model of cryptococcal meningitis was chosen because inhalation is the most common route of infection following environmental exposure and, therefore, provides the best infection model to examine the pathogenicity of C. neoformans as it relates to the clinical setting. Figure Figure11 shows that there was no significant difference in the survival of SK1/2+/+ and SK1−/− mice infected with WT C. neoformans strain H99 cells (19.6 ± 1.5 days versus 20.6 ± 1.2 days). Since mice are innately susceptible to cryptococcosis, it is possible that the virulence of this WT strain of C. neoformans may mask any effect that SK1 has on the development of cryptococcosis. Thus, to determine if SK1 affects the dissemination of WT C. neoformans and disease progression, the fungal burdens in the organs from H99-infected SK1/2+/+ and SK1−/− mice were determined at days 5, 10, and 15 after C. neoformans challenge. Tissues from three mice were isolated and processed to determine CFU at each time point. There were no significant differences in the fungal burdens in the lungs (Fig. (Fig.2A)2A) and brains (Fig. (Fig.2B)2B) of H99-infected SK1/2+/+ and SK1−/− mice at any time point inspected. No differences were found in the fungal burdens of the liver, spleen, and kidneys (data not shown). These results are not surprising, as there was no difference in the symptoms between SK1/2+/+ and SK1−/− mice challenged with WT C. neoformans. Since no differences in organ fungal burden or symptoms were found, histopathology was not performed on the organs from H99-infected SK1/2+/+ and SK1−/− mice. Together, Fig. Fig.11 and and22 suggest that SK1 does not play a role in the susceptibility of an immunocompetent mouse to a WT (e.g., facultatively intracellular) C. neoformans strain.
SK1 is required for the antimicrobial activity of macrophages against internalized Mycobacterium (16, 17, 29, 39, 46). To investigate the effect of SK1 on the pathogenicity of intracellular C. neoformans, we used a C. neoformans strain in which the gene encoding glucosylceramide synthase 1 (GCS1) is deleted (Δgcs1) (24, 36). C. neoformans Δgcs1 cells do not replicate in the neutral or alkaline pH and physiological concentrations of CO2 characteristic of extracellular environments, such as in alveolar spaces and the bloodstream. However, this mutant strain lacking the GCS1 gene has no growth defect under acidic conditions, such as within the phagolysosome (36). It is important to note that the Δgcs1 strain does not disseminate and is avirulent in immunocompetent mice (24, 36). Therefore, we refer to C. neoformans Δgcs1 as obligately intracellular for replication in vivo. Thus, SK1/2+/+ and SK1−/− mice were infected intranasally with C. neoformans Δgcs1, and mouse survival and tissue burden cultures were examined. All SK1/2+/+ and SK1−/− mice infected with Δgcs1 survived 100 days postchallenge (Fig. (Fig.3A).3A). At this time point, the lungs and brains of these mice were extracted and processed either to determine fungal burden using CFU or for histopathological analysis using light microscopy. SK1−/− mice had a significantly greater number of Δgcs1 cells in the lungs than SK1/2+/+ mice (P < 0.05) (Fig. (Fig.3B).3B). Tissue burden culture of brains showed that none of the seven brains from SK1/2+/+ mice contained Δgcs1 cells, whereas all seven brains from SK1−/− mice contained C. neoformans cells, representing a statistically significant difference (P < 0.005) (Fig. (Fig.3B).3B). Organ fungal burdens in the livers, spleens, and kidneys were also examined using CFU. C. neoformans cells were not found within the livers, spleens, and kidneys of Δgcs1 mutant-infected SK1/2+/+ mice, while the Δgcs1 strain was present in the spleens of two SK1−/− mice but absent in all other organs examined (data not shown). Together, these data suggest that SK1 protects the host from pulmonary C. neoformans infection and prevents dissemination to the central nervous system only when dealing with a C. neoformans strain that can replicate only intracellularly.
Histopathology of the lungs of Δgcs1-infected SK1/2+/+ mice compared to Δgcs1-infected SK1−/− mice corroborated the hypothesis that SK1 was instrumental in the host's ability to prevent dissemination of C. neoformans Δgcs1 from the lung to the brain. The lungs of three SK1/2+/+ and SK1−/− mice were excised, sectioned, mounted onto microscope slides, and subjected to various stainings for inspection of the host immune response at 100 days after intranasal challenge with C. neoformans Δgcs1. We found that SK1/2+/+ mice challenged intranasally with Δgcs1 cells showed a granuloma in the lungs that, in light of the data presented in Fig. Fig.3B,3B, successfully contained and prevented dissemination. These results, obtained with C57BL/6J immunocompetent mice, are similar to previous findings with CBA/J immunocompetent mice (36). Figure Figure4A4A shows that C. neoformans cells, stained with Alcian blue by Movat staining, are contained within the center region of a well-organized granuloma, where the host cellular components are stained shades of red. Activated fibroblasts deposited extracellular matrix, including significant amounts of collagen, which is stained pink by VVG staining, to form a thick fibrotic tissue encapsulating the Δgcs1 strain-containing granuloma, thereby acting as a physical barrier preventing dissemination from the lungs (Fig. (Fig.4B).4B). The encapsulation of the fibrotic tissue is further demonstrated with Movat staining, which showed that fibrotic tissues (stained light red) and the leukocyte infiltrate (stained deep red) prevent the egress of Δgcs1 cells from the lungs (Fig. (Fig.4C).4C). Mixed populations of leukocytes were observed to reside between the fibrotic tissue and the region containing C. neoformans cells (Fig. (Fig.4B,4B, D, and E). When H&E and Giemsa stains were used to examine the cell types comprising the leukocyte infiltration during pulmonary cryptococcosis (9, 37), neutrophils were identified as the dominant granulocyte cell type, whereas plasma cells were also present (Fig. 4D and E). Although eosinophils have been shown to be a common component of the pulmonary inflammatory response of C57BL/J6 mice to wild-type C. neoformans challenge through 28 days (9, 11, 21), eosinophils were rarely observed in the granulomas within the lungs of Δgcs1 strain-infected SK1/2+/+ mice and do not represent a significant cell type in the inflammatory response at 100 days after C. neoformans Δgcs1 challenge. However, crystal-like structures were observed both extracellularly and intracellularly in host phagocytes with and without internalized C. neoformans throughout the granulomas in the lungs from Δgcs1 strain-challenged SK1/2+/+ mice (Fig. 4D and F). These unique structures are mostly likely derived from the eosinophil chemotactic factors Ym1 and/or Ym2, whose formation commonly occurs in C. neoformans-challenged C57BL/J6 mice (11, 21). The presence of these crystals indirectly suggests eosinophil involvement at some junction during the host inflammatory response.
C. neoformans cells were localized to the center of the granulomas in the lungs from Δgcs1 strain-challenged SK1/2+/+ mice (Fig. 4A and F). In developing granulomas, C. neoformans cells were found to reside mostly intracellularly within host phagocytes. The center regions of the more developed granulomas were necrotic and thus contained cellular debris (Fig. (Fig.4F).4F). Numerous extracellular C. neoformans cells were also found within the necrotic cores of these advanced granulomas (Fig. (Fig.4F).4F). Since the vast majority of C. neoformans Δgcs1 cells are observed intracellularly in developing granulomas, while the more-developed granulomas contain cellular debris and extracellular Δgcs1 cells, it is likely that the extracellular C. neoformans cells in the necrotic regions are Δgcs1 cells that had been previously internalized by macrophages and then released following lysis of the host phagocytes.
In contrast, the inflammatory response to the Δgcs1 strain in lungs of SK1−/− mice did not result in the formation of well-developed granulomas, as observed in Δgcs1 strain-challenged SK1/2+/+ mice (Fig. (Fig.4G).4G). The nodule-like structures in the lungs of Δgcs1 strain-challenged SK1−/− mice lacked necrotic centers, defined rings of recruited immune cells, and encapsulating fibrotic tissue (Fig. (Fig.4G,4G, H, and I). Although collagen deposition was visible in some areas, it was present in relatively low abundance and did not surround the outer perimeter of the nodule-like structure (Fig. (Fig.4H).4H). In comparison to the leukocyte infiltration of the lungs from SK1/2+/+ mice, significantly more foamy histiocytes, giant macrophages, and lymphocytes were observed in lungs from SK1−/− mice (Fig. 4I and J). In addition, Δgcs1 cells were almost exclusively localized intracellularly within macrophages (Fig. (Fig.4L).4L). These phagocytic cells also contained numerous C. neoformans cells and/or a copious amount of capsule fragments (Fig. (Fig.4L).4L). Interestingly, these Δgcs1 mutant-containing host cells were abundantly distributed throughout the tissue and were frequently found near blood vessels (Fig. 4I and L), an observation possibly foreshadowing the escape from the host defense perimeter, leading to dissemination. Neutrophils were less prevalent in the inflammatory response of SK1−/− mice to the Δgcs1 mutant, whereas the lymphocyte infiltration occurred to a much greater extent than in SK1/2+/+ mice (Fig. 4J and H). As with Δgcs1 strain-challenged SK1/2+/+ mice, eosinophils were rarely observed in the inflammatory response of Δgcs1 strain-challenged SK1−/− mice. However, the crystal formation commonly found in the lung granulomas of SK1/2+/+mice was not present in the lung granulomas of SK1−/−mice. These observations suggest that SK1 affects host factors influencing granulocyte recruitment and lymphocyte infiltration. As a result, the inflammatory response in SK1−/− mice to the Δgcs1 strain does not promote the formation of a proper granuloma, leading to increased lung fungal burden and ultimately resulting in C. neoformans Δgcs1 dissemination to the brain (Fig. (Fig.2B2B).
Phagocytosis is an essential component of the effector activities of AMs against C. neoformans. Histopathology of the lungs from SK1/2+/+ and SK1−/− mice challenged with the Δgcs1 mutant revealed a much greater number of intracellular C. neoformans cells residing within AMs from SK1−/− mice than within AMs from SK1/2+/+ mice, suggesting that SK1 may affect the internalization and/or intracellular growth of C. neoformans. Analysis of the mRNA expression of AMs from SK1/2+/+ mice using real-time RT-PCR revealed that these phagocytes express both SK1 and SK2 genes, although SK1 was the predominant isoform, as its expression was approximately 2.7-fold times higher than that of SK2 (data not shown). To determine the specific role of SK1 in phagocytosis of C. neoformans, AMs from SK1/2+/+, SK1−/−, and SK2−/− mice were isolated, coincubated with C. neoformans at a MOI of 1:1 for 2 h, and processed for light microscopy to determine the number of internalized C. neoformans cells. Prior to coincubation with AMs, C. neoformans cells were opsonized with either complement from fresh mouse serum, the anti-GXM monoclonal IgG1 antibody 18B7, or both. Figure Figure5A5A shows the phagocytic index of complement-opsonized C. neoformans Δgcs1 cells by AMs isolated from SK1/2+/+, SK1−/−, and SK2−/− mice. All three groups of AMs had low phagocytic indices, and no differences were observed. However, AMs from SK1−/− mice had significantly greater internalization of antibody-opsonized Δgcs1 cells than AMs from SK1/2+/+ mice (65.5 ± 9.9 versus 27.5 ± 8.5, P < 0.05) (Fig. (Fig.5B).5B). The phagocytic index of the Δgcs1 strain by SK2−/− AMs was approximately the same as the phagocytic index of AMs possessing both SK isoforms (34.5 ± 14.5 versus 27.5 ± 8.5) (Fig. (Fig.5B).5B). Similar results were found when Δgcs1 cells were opsonized with both complement and antibody (Fig. (Fig.5C).5C). These data reveal that the deficiency in SK1 increases phagocytosis of C. neoformans Δgcs1 when the fungal cells have been opsonized with an IgG1 antibody. The phagocytic indices of AMs isolated from SK1/2+/+, SK1−/−, and SK2−/− mice were also analyzed for C. neoformans wild-type strain H99. As was observed with the phagocytosis of the Δgcs1 mutant, complement-opsonized H99 cells were poorly internalized, with no differences between groups of AMs. We found that the phagocytic indices of antibody-opsonized wild-type C. neoformans cells significantly increased, similarly to those of C. neoformans Δgcs1, upon deletion of SK1 (Fig. (Fig.5E5E and andF),F), suggesting that this increase is due solely to the lack of SK1 and not to the lack of GlcCer. Additionally, AMs from SK2−/− mice had phagocytic indices that were comparable to those of AMs from SK1/2+/+ mice, demonstrating that the increased internalization of C. neoformans cells by AMs from SK1−/− mice is specific to the deficiency of SK1.
To determine if SK1 modulates the ability of AMs to control the initial growth of internalized C. neoformans wild-type and Δgcs1 mutant cells, the presence of daughter cells, also referred to as budding, was analyzed in C. neoformans cells internalized within SK1/2+/+, SK1−/−, and SK2−/− AMs. To enable phagocytosis, C. neoformans cells were opsonized with both complement from fresh mouse serum and the anti-GXM IgG1 antibody 18B7, which was used in the in vivo phagocytic assays. After 4 h of coincubation, there was no difference in intracellular growth of the C. neoformans Δgcs1 or wild-type strain among different types of AMs (Fig. 6A and B), suggesting that SK1 and SK2 do not have roles in controlling C. neoformans intracellular growth.
Several lines of evidence suggest that AMs can promote and exacerbate dissemination, particularly when they are not activated by Th1-specific cytokines (6, 12, 24, 28, 36). Very interestingly, downregulation of SK1, but not SK2, by small interfering RNA (siRNA) in mice leads to a suppression of collagen-specific proinflammatory/Th1 cytokines (e.g., IL-6, TNF-α, and IFN-γ) (26), suggesting that SK1 plays a role in regulating the immune response to the granuloma formation. Since we found that lack of SK1 does not affect the infection of C. neoformans WT H99 under conditions of immunocompetence, we wondered whether we could exacerbate C. neoformans infection in SK1−/− compared to SK1/2+/+ mice by inducing immunodeficiency. Thus, SK1/2+/+ and SK1−/− mice were treated with a regimen of the corticosteroid cortisone acetate (CA) and challenged intranasally with a C. neoformans WT strain. Corticosteroids suppress inflammation by inhibiting the generation of Th1 response-associated cytokines, which results in reduced T-cell proliferation and numbers. To make sure of the effectiveness of the CA regimen (decrease in the number and proliferation of circulating lymphocytes), we performed pilot experiments to measure the effect of the immunosuppression on white blood cell (WBC) count in our mouse model of pulmonary cryptococcosis. We found that treatment with CA in mice profoundly reduced the WBC count (Fig. (Fig.7A),7A), specifically that of lymphocytes (Fig. (Fig.7B).7B). Circulating lymphocytes account for 60 to 90% (depending on mouse strains) of all WBCs (34, 41), and thus, the observed decrease in the WBC counts upon CA regimen results in a specific depletion of lymphocytes as the number of the other cell types examined (e.g., neutrophils, monocytes, eosinophils, and basophils) was not significantly affected by CA treatment (data not shown). Interestingly, untreated SK1/2+/+ mice had a significantly greater number of WBCs and lymphocytes than untreated SK1−/− mice (P < 0.05) (Fig. (Fig.7),7), which appears to be in contrast with previous studies in which no abnormalities in the number of lymphocytes between SK1/2+/+ and SK1−/− mice were reported (1). However, it is possible that this discrepancy may be due to the different sites of blood collection (e.g., venous versus heart), as WBC count varies according to which blood vessel is utilized for collection (34, 41). Regardless of this, CA treatment clearly results in a decrease in WBCs, and more specifically in lymphocytes, in both SK1/2+/+ and SK1−/− mice.
As expected, we found that CA-treated mice were significantly more susceptible than untreated mice to the infection by C. neoformans WT strain (Fig. (Fig.88 versus Fig. Fig.1).1). Interestingly, Fig. Fig.88 reveals that CA-treated SK1−/− mice showed a significantly greater susceptibility than CA-treated SK1/2+/+ mice (14.2 ± 3.4 days versus 17.5 ± 2.3 days, P < 0.05). These results suggest that, under conditions of immunosuppression, SK1 may have a role in controlling infection by a wild-type C. neoformans strain.
In this study, we show that SK1 is involved in the regulation of the host immune response to a fungal human pathogen. More specifically, we found that SK1 plays an essential role in the granuloma formation required to prevent dissemination of the C. neoformans Δgcs1 strain. In fact, in lungs of SK1−/− mice infected with the Δgcs1 mutant, we found a significantly increased fungal burden and lack of granuloma. Thus, the C. neoformans Δgcs1 strain was able to disseminate to the brains of SK1−/− mice. In addition, in vitro experiments revealed that the deficiency in SK1 increases phagocytosis of C. neoformans WT and Δgcs1 by AMs but does not promote intracellular fungal growth. Interestingly, this increased internalization of C. neoformans was dependent on antibody- and not complement-mediated opsonization. Finally, we found that the effect of SK1 on host susceptibility to cryptococcosis is more pronounced when the host is immunocompromised. To our knowledge, this is the first report implicating SK1 in the regulation of phagocytosis and virulence of a fungal pathogen.
Evidence supporting a role of SK1 during infection is provided by previous studies showing that the SK1/S1P pathway regulates immune cell function in vitro and by the fact that this pathway is of particular importance in the lung, as S1P level is modulated in lung diseases (16, 22, 38). Additional support is also provided by investigations examining the effects of intravenous administration of S1P on the histopathology of Mycobacterium (16, 17, 39). However, to date, the role of SK1, which is the predominant SK isoform expressed in lung tissue, in the modulation of the host immune response during pulmonary infection has not been directly examined.
Using SK1/2+/+ and SK1−/− mice as models of pulmonary cryptococcosis, we sought to determine if SK1 was involved in regulating the host immune response to the human pathogenic fungus C. neoformans. We found that SK1 did not affect the virulence of the WT C. neoformans strain H99, a facultative intracellular pathogen. In previous studies, histopathology of pulmonary cryptococcosis in mice revealed that the changes in the morphology of C. neoformans cells (capsule structure and size, cell wall thickness, and fungal cell size) occur throughout the course of infection and correspond to changes in the localization of the fungal cells (extracellular versus intracellular) within the host (10, 12, 15). Hence, the ability of WT C. neoformans cells to reside both extracellularly and intracellularly could be a result of an adaptive advantage toward a normal host immune response and, as a consequence, it may neglect the role of SK1 in this particular host condition.
Since SK1 and its specific production of S1P regulate antimicrobial actions of macrophages against internalized Mycobacterium, we hypothesized that SK1 may affect the pathogenesis and virulence of C. neoformans under conditions that allow the fungus to replicate only intracellularly within host phagocytes. To investigate this hypothesis, we examined the effect of SK1 during pulmonary infection with the C. neoformans Δgcs1 strain, which cannot replicate at the pH (7.0) and carbon dioxide concentration (5%) that characterize the extracellular spaces of the lungs. Therefore, this mutant strain of C. neoformans must reside intracellularly to replicate in vivo (24, 36). Under normal host conditions (e.g., wild-type mice), internalization of Δgcs1 cells stimulates an anticryptococcal response that leads to containment and killing of the fungus by AMs. In this study, comparison of histopathology specimens of lungs from SK1/2+/+ and SK1−/− mice reveals that there are vast differences in the host immune response to the Δgcs1 strain after 100 days of infection. In SK1/2+/+ mice, but not in SK1−/− mice, the host inflammatory response contains Δgcs1 cells within granulomas, thereby limiting disease progression (e.g., organ fungal burden and dissemination).
An abundance of crystal-like structures was observed throughout the granulomas in the lungs of Δgcs1 strain-infected SK1/2+/+ mice, particularly in the extracellular space of the necrotic core of well-developed granulomas. These crystal-like structures derived from Ym1 and/or Ym2, eosinophil chemotactic factors possessing chitinase activity (19, 33). These crystal structures have been previously described to occur in the different murine lung diseases, including pulmonary cryptococcosis (11, 21). Here, we observed crystal formation in SK1/2+/+ mice following challenge with the Δgcs1 mutant upon a well-developed granulomatous response. Interestingly, these crystal-like structures were not observed in the lungs of SK1-deficient mice, which do not show granulomas. Additionally, we did not observe granuloma formation or the presence of these crystal structures in the lungs of SK1/2+/+ or SK1−/− mice infected with the wild-type strain C. neoformans H99 (data not shown). Together, these results suggest that lack of the GCS1 gene (microbial side) and presence of SK1 (host side) may have a role in contributing to the production of these crystals.
Since SK1 mediates proinflammatory stimuli and its activity is required for cytokine generation and secretion (20, 30, 32, 50), it is hypothesized that SK1 promotes the ability of AMs to recruit other effector immune cells and is therefore essential to the host immune response during pulmonary cryptococcosis. The decreased presence of neutrophils at the site of the infection observed in the lungs from SK1−/− mice compared to SK1/2+/+ mice supports this hypothesis. Intriguingly, neutrophils and not eosinophils were the most represented granulocyte population in the host inflammatory response to C. neoformans Δgcs1 in both SK1/2+/+ and SK1−/− mice (Fig. (Fig.4D,4D, E, J, and K). This is seemingly contradictory to other studies using C57BL/6J mice in models of pulmonary cryptococcosis, where eosinophils are the major granulocyte cell type represented during the host inflammatory response (9, 21). However, there are two major differences between previous reports and the data presented here. First, in previous studies, WT C. neoformans strains H99 (serotype A) and ATCC 24067 (serotype D) were used. Both of these strains are well established as virulent C. neoformans strains in mouse models of pulmonary cryptococcosis. The strain used in our mouse model of pulmonary cryptococcosis for this experiment, the C. neoformans Δgcs1 strain, is an avirulent mutant strain that evokes a granulomatous inflammatory response to a degree that has not been described for infections with WT C. neoformans strains. Importantly, WT C. neoformans strains are facultative intracellular pathogens, whereas the Δgcs1 strain is obligately intracellular for replication in vivo. Therefore, the Δgcs1 strain elicits an immune response that is different than that of the WT C. neoformans strain used in previous studies. Second, previous studies analyzed the host inflammatory response at earlier time points (mostly less than 30 days after C. neoformans challenge). Here, we examined the host inflammatory response at 100 days after C. neoformans challenge. Thus, although eosinophils may have been involved at an earlier junction of the host immune response, eosinophils do not represent a major component of the lung host immune response at the time point we examined. Furthermore, new data are emerging that suggest a concerted action of neutrophils and macrophages in the induction of adaptive immunity against intracellular pathogens, including Mycobacterium tuberculosis, Toxoplasma gondii, Listeria monocytogenes, and Salmonella species (43). It is interesting to speculate that, under conditions where C. neoformans cells are pressured to reside intracellularly, neutrophils may replace eosinophils as the main granulocyte recruited to the site of infection.
Conditions facilitating intracellular growth of C. neoformans within AMs transform the host phagocytes into a niche that may actually promote fungal dissemination. Furthermore, the lack of fibrotic tissue surrounding the granuloma in lungs of SK1−/− mice may enable the dissemination of Δgcs1 cells from the lungs. Interestingly, SK1 activity and S1P induce human lung fibroblasts to differentiate into myofibroblasts producing extracellular matrix (ECM) (47), and in several tissue-specific fibroblasts, they increase collagen synthesis, which is required for proper granuloma formation (18, 49). In addition, studies have recently suggested that SK1 is involved in promoting the stimulation of collagen-specific proinflammatory Th1 cytokines, such as IL-6, TNF-α, and IFN-γ (26). Also, treatment with SK1 inhibitors suppresses the production of IFN-γ by T cells and IL-12 by dendritic cells (23), suggesting that SK1 is a regulator of the Th1 immune response. These studies could explain why the formation of fibrotic tissues and collagen deposition was almost absent in nodule-like structures found in lungs of SK1−/− mice, further exemplifying the importance of SK1 in granuloma formation. It is important to emphasize that this phenomenon (granuloma formation) cannot be studied using C. neoformans wild-type strain H99 because the strain is too virulent in mice and the animals succumb before a lung granuloma can be formed.
Unlike many intracellular pathogenic microorganisms, C. neoformans does not possess a mechanism to actively enter host phagocytes. Internalization of C. neoformans by host phagocytes, including AMs, occurs through receptor-mediated pathways, which mostly involve complement and antibody opsonization (27). We found that AMs from SK1−/− mice showed significantly increased phagocytosis of both the C. neoformans WT and Δgcs1 strains in comparison to AMs from SK1/2+/+ mice in vitro. In our assays, SK1 deficiency increased the internalization of C. neoformans cells only when opsonization included antibody. Importantly, we found no differences in phagocytosis between C. neoformans wild-type and Δgcs1 strains in SK1/2+/+ and SK1−/− AMs when fungal cells were opsonized only with complement (Fig. 5A and D). Thus, SK1 may have a role in the phagocytosis of C. neoformans not in the time period shortly following inhalation but rather only after the production of C. neoformans-specific antibodies. From our in vitro phagocytosis assay data, we hypothesize that SK1 may regulate the expression of cell surface receptors on AMs that mediate phagocytosis, specifically FcγRs that recognize IgG molecules. This hypothesis is supported by studies showing that inhibition of SK1, but not SK2, regulates the expression of surface molecules such as CD40, CD80, CD86, and MHC class II in certain immune cells (23), suggesting that SK1 may also be important for the expression of receptors in AMs involved in the phagocytosis of C. neoformans. Additionally, treatment with exogenous S1P has been shown to increase FcγRII expression in macrophages (7). On the other hand, a possibility exists that SK1 regulates phagocytosis of C. neoformans after host recognition. In macrophages, the interaction of IgG with FcγRI triggers SK1 activity, which then evokes a signaling cascade involving PI-PLC, cPKC, ERK1/2, and PI 3-kinase, which has been suggested to modulate phagocytosis of Mycobacterium species (48). Our results, along with these previous reports, suggest that SK1 and its product, S1P, may have a role in controlling the antibody-mediated phagocytosis of C. neoformans.
As an opportunistic pathogen, C. neoformans most commonly infects individuals with impaired cellular immunity caused by human immunodeficiency virus (HIV) infection, solid organ transplantation, or administration of potent immunosuppressive regimens. We found that loss of SK1 increased susceptibility to WT C. neoformans when a corticosteroid regimen drastically reduced the number of circulating lymphocytes, suggesting that SK1 may affect host susceptibility to WT C. neoformans (i.e., a facultative intracellular pathogen) during an immunocompromised state. When immunosuppression occurs, AMs are not activated and, therefore, are unable to kill internalized C. neoformans, which can then replicate intracellularly to exacerbate disease progression and eventually lead to decreased mouse survival. Thus, the difference in susceptibility to WT C. neoformans between CA-treated SK1/2+/+ mice and CA-treated SK1−/− mice (Fig. (Fig.8)8) may result from the higher number of C. neoformans cells found intracellularly in the lungs of SK1-deficient mice (Fig. (Fig.44).
In conclusion, in this research, we show that SK1 is dispensable in mice infected with a facultative intracellular C. neoformans wild-type strain, but it is required for total containment of an obligate C. neoformans intracellular pathogen for replication. We also show that SK1 may be important during the infection caused by a C. neoformans wild-type strain under conditions of immunodeficiency.
We thank all members of M. Del Poeta's and C. Luberto's laboratories for helpful and constructive discussion. We are particularly grateful to Russell Harley and Masha Bilic for helping with histology analysis.
This work was supported by grants AI56168 and AI72142 (to M.D.P.) and was conducted in a facility constructed with support from the National Institutes of Health, grant number C06 RR015455 from the Extramural Research Facilities Program of the National Center for Research Resources. T. McQuiston was supported by the Graduate Assistance in Areas of National Need (GAANN) training grant in Lipidology and New Technologies (to M.D.P.) from the United States Department of Education. Maurizio Del Poeta is a Burroughs Wellcome New Investigator in Pathogenesis of Infectious Diseases.
Editor: G. S. Deepe, Jr.
Published ahead of print on 1 March 2010.