Cryptococcus neoformans Growth and Protection from Innate Immunity Are Dependent on Expression of a Virulence-Associated DEAD-Box Protein, Vad1

Abstract

The fungus Cryptococcus neoformans has emerged as a major cause of meningoencephalitis worldwide. Host response to the fungus involves both innate and adaptive immunity, but fungal genes that modulate these processes are poorly understood. Previous studies demonstrated attenuated virulence of a mutant of a virulence-associated DEAD-box protein (VAD1) in mice, despite normal growth at host temperatures, suggesting modulation of the immune response. In the present study, the Δvad1 mutant demonstrated progressive clearance from lung and was unable to induce pathological lesions or to cause extrapulmonary disease, despite retaining its ability to grow in mouse serum and a J774.16 macrophage cell line. Pulmonary clearance occurred with a minimal cellular infiltrate, marked by reduced CD4 cells, CD11b⁺ Ly6C^high monocytes, and F4/80⁺ macrophages, but the mutant strain retained recruitment of CD8 cells, compared to infections with wild-type fungi. Adaptive cytokine responses were reduced, including Th1, Th2, and Th17 cytokines; however, early gamma interferon (IFN-γ) and tumor necrosis factor alpha (TNF-α) responses were retained while nonprotective interleukin 4 (IL-4) and IL-5 were diminished. Furthermore, the Δvad1 mutant was controlled in lungs despite CD4/CD8 cell depletion. These data, along with improved phagocytosis by macrophages and increases in early/innate IL-1α, IFN-γ, and chemokines elicited in the lungs within 3 days of infection with the Δvad1 mutant, indicate that VAD1 expression reduces innate recognition of C. neoformans, rendering the yeast resistant to elimination by the innate mechanisms of host defense. Thus, our studies define a novel role of the cryptococcal Vad1 protein as a central regulator of cryptococcal virulence and illustrate that Vad1 promotes microbe resistance to innate host defenses.

INTRODUCTION

Cryptococcus neoformans is a major pathogen in immunocompetent and immunocompromised patients, including those with AIDS in both the developed and the developing world. As cases decline in the West with the advent of AIDS-directed anti-retroviral therapy, there has been recognition of large numbers of cases in the developing world, accounting for approximately 600,000 deaths annually (1). In more-developed countries, the fungus continues to cause disease, predominantly in immunosuppressed patients, notably those receiving cancer chemotherapy and transplant-related immune conditioning (2).

Acquisition of the disease is thought to occur through inhalation of small desiccated yeast forms into the lung (3). The alveolar macrophage is thought to represent the first cellular line of defense against this facultative intracellular pathogen (4), and antigen-presenting cells such as dendritic cells (DC) and macrophages are key players in control of cryptococcal lung infections during both innate and adaptive phases of the immune response (5–10). Successful lung fungal clearance and prevention of systemic dissemination also depend on effector function of pulmonary CD4⁺ and CD8⁺ T cells and protective Th1 immune polarization, while the development of Th2 polarization is nonprotective (11–15). The role of Th17 responses is less clear, but studies suggest that interleukin 17 (IL-17) could contribute to protection against C. neoformans (16–19). Indeed, most recent studies demonstrate that the overall balance between multiple cytokine responses in C. neoformans-infected lungs translates into outcomes that range from successful fungal clearance to the development of a chronic “steady-state” infection or a progressive, lethal infection (11, 13, 20–23). In summary, these studies document that the innate mechanisms of the immune response on their own are insufficient to contain the microbe following infection and that adaptive Th immunity needs to be properly polarized to provide effective control of the pathogen.

Apart from the effects of the host immune status, quantitative differences in the expression of multiple virulence factors modulate the ability of C. neoformans to persist in the infected host and to cause central nervous system (CNS) dissemination (24, 25). Some of these factors have been shown to promote crucial steps in the pathogenesis of the yeast such as the ability to grow in and disseminate from the lungs into other organs and tissues and/or survive within the CNS (8, 26–30). A prominent virulence factor for the fungus is the production of an anti-phagocytic polysaccharide capsule (31). In addition, laccase is a copper-dependent cell wall-associated virulence factor that plays a role in survival within the CNS (27) and urease is a secreted enzyme that plays a role in the transmigration of the fungus through capillaries into the CNS (28, 30). Both laccase and urease also have the potential to modulate adaptive immune responses (8, 32). However, our understanding of the relationship between fungal factors and the immune response remains incomplete because many virulence-associated genes identified in C. neoformans have not been evaluated with respect to their role or mechanism in the pathogenesis of cryptococcosis (8, 25, 28, 33–35).

One of the novel factors identified by insertional mutagenesis is a virulence-associated DEAD-box protein (Vad1). Previous studies established that Vad1 serves as an important regulator of the virulence factor laccase and that a VAD1-deletant mutant of C. neoformans (Δvad1) exhibited markedly attenuated virulence in mouse models despite normal growth at host temperatures, suggesting that altered fungal fitness was not exclusively responsible for the attenuated virulence (36). Thus, to understand the relative contributions of fungal fitness and host response in the overall virulence composite due to VAD1 expression, we undertook further studies to assess anticryptococcal host defenses leading to accelerated clearance of the Δvad1 mutant and whether these effects could be chiefly or partially attributed to changes in laccase production. Our objective was to determine whether VAD1 expression affects the course of pulmonary infection and CNS dissemination in infected mice. Our results identified a crucial role for cryptococcal VAD1 expression in suppressing innate responses and regulating cryptococcal pathogenicity that are largely laccase independent. Understanding these novel aspects of fungal and host immune system interactions may provide opportunities for improving the control and treatment of cryptococcosis.

MATERIALS AND METHODS

Fungal strains and cell lines.

C. neoformans wild-type (wt) strain H99 (ATCC 208821) was a kind gift of J. Perfect. Mutant Δlac1 and Δvad1 strains and the complemented Δvad1::VAD1 strains were derived from H99 and were described previously (36). The J774.16 cell line (here referred to as J774) was obtained from the American Type Culture Collection (ATCC, Manassas, VA).

C. neoformans labeling.

Calcofluor white dye (fluorescent brightener 28; Sigma-Aldrich) was used to label cell walls of C. neoformans where indicated. Yeast cells were collected from agar plates containing 2% glucose, 2% Bacto peptone, and 1% yeast extract (YPD), washed twice, and suspended in phosphate-buffered saline (PBS) (Invitrogen) at 5 × 10⁶ to 2 × 10⁷/ml. The cells were incubated with Calcofluor at 10 μg/ml in PBS for 30 min in the dark at room temperature and then washed twice in PBS as indicated.

Phagocytosis assay.

To assess the extent of Cryptococcus phagocytosis, J774 cell suspensions (10⁵ in fresh medium per well of a 8-well culture slide) were incubated at 37°C in 5% CO₂ overnight after activation with IFN-γ (100 U/ml) and lipopolysaccharide (0.6 μg/ml). Calcofluor-labeled yeast cells were opsonized using nonimmune mouse serum or monoclonal antibody (MAb) 18B7 (1 μg/ml) as indicated and incubated for 1 h at 37°C with 5% CO₂. Thereafter, Calcofluor-stained C. neoformans suspensions were added to a J774 cell monolayer and incubated at 37°C and 5% CO₂ for 1 h (phagocytosis assay; C. neoformans/J774 ratio, 5:1). To differentiate internalized yeast cells from noningested C. neoformans cells, Fun-1 staining (100 μl of PBS containing 2.5% fetal calf serum [FCS] and 1 μl Fun-1) was used as described previously (36). After coincubation at 37°C and 5% CO₂ for 30 min, nonadherent extracellular yeast cells were then removed by PBS washings, and incubation was stopped to assess phagocytosis. The phagocytic index (PI) was determined by the following equation: PI = (number of attached and ingested cryptococci)/(number of macrophages) (37). Three individual experiments for each condition were performed.

Murine pulmonary infections.

All experimental procedures were conducted under protocols approved by the Institutional Animal Care Committees (IACUC) of the Intramural NIH/NIAID and the University of Illinois at Chicago. CBA/J mice were inoculated intranasally with 1 × 10⁶ CFU of C. neoformans wt strain H99 or Δvad1, VAD1-complemented, or Δlac1 strains in 20 μl of sterile PBS. A second model to more closely equilibrate fungal loads between wt and Δvad1 infections uses an intranasal inoculation with 1 × 10⁴ CFU of the indicated strains, followed by an additional inoculation at 2 weeks of 1 × 10⁷ CFU of the Δvad1 strain followed by euthanasia and tissue preparation at week 3 as described below. The mice were fed ad libitum and were monitored by inspection twice daily. Mice were euthanized on days 1, 2, 3, 7, 14, and 21 postinoculation, and brain, spleen, and lung tissue was excised using an aseptic technique. Tissue was homogenized in 1 ml of sterile PBS, followed by culture of serial dilutions of each tissue on YPD agar supplemented with chloramphenicol (Mediatech, Inc., Herndon, VA). CFU were enumerated following incubation at 30°C for 48 h. In a separate experiment, mice were inoculated with 1 × 10⁶ CFU of the Δvad1 strain and then injected intraperitoneally (i.p.) with monoclonal antibody (MAb) (300 μg rat anti-mouse CD4, clone GK 1.5, and 300 μg rat anti-mouse CD8, clone 2.43) on day 21 and then 100 mg of each MAb on day 24 and weekly until mice were euthanized at 8 weeks, followed by analysis of tissue fungal burden.

Pulmonary leukocyte isolation.

Mice were euthanized by CO₂ asphyxiation. To exsanguinate the animals, the abdominal vena cava was severed to remove circulating blood cells from lungs prior to excision. Lungs were excised on days 1, 2, 3, 7, 14, and 21 postinoculation and digested enzymatically at 37°C in 15 ml of digestion buffer (RPMI 1640 containing 5% fetal calf serum, penicillin, streptomycin, 0.25 ml/mouse of DNase, and a 1-mg/ml concentration of collagenase type IV [Sigma]) with intermittent stomacher homogenizations every 15 min. The resultant cell suspensions were centrifuged (250 × g) for 1 min to remove tissue debris. Supernatants were then successively filtered through sterile nylon filters to enrich for leukocytes. Erythrocytes were lysed by incubation in NH₄Cl buffer (0.859% NH₄Cl, 0.1% KHCO₃, 0.0372% Na₂EDTA, pH 7.4) for 3 min on ice followed by addition of a 10-fold excess of Hanks' balanced salt solution (HBSS). The resulting leukocyte population was collected by centrifugation (800 × g) for 10 min, washed twice with sterile HBSS, and enumerated in a hemocytometer using trypan blue dye exclusion. Flow cytometric analysis was used to determine the absolute number of total leukocytes (CD45⁺) within the lung cell suspension for correction of hemocytometer counts.

Antibody staining and flow cytometry analysis.

Leukocyte-enriched lung cells (1 × 10⁶) were incubated with Fc Block (BD Pharmingen), anti-murine CD16/CD32, and rat immunoglobulin G2b (IgG2b) in 100 μl of PBS supplemented with 2% heat-inactivated fetal bovine serum (PBS-FBS) for 5 min to block nonspecific binding of antibodies to cellular Fc receptors. All antibodies were purchased from BD Pharmingen. Data were analyzed using an LSRFortessa flow cytometer with DIVA software and using FlowJo software (Tree Star Inc., San Carlos, CA). CD45⁺ leukocytes, CD11b⁺/Ly6C^high cells, and CD11c⁺ cells were detected in enzymatically dispersed lungs over the first 3-week period postinfection. The gating strategy involved initial exclusion of debris, red blood cells (RBC), and cell clusters using light scatter, followed by a forward scatter (FSC)-versus-CD45 plot to select for CD45⁺ leukocytes. An FSC-versus-CD11c scatter plot was used to identify large CD11c^neg cells (see Fig. S1A, gate R1, in the supplemental material). Using this R1 population, a second plot of CD11c versus CD11b was used to differentiate monocytes (CD11c^neg/CD11b⁺Ly6C^high) from lung dendritic cells (DC) (CD11c/CD11b). To maintain consistency, the cytometer parameters and gate position were held constant during analysis of all samples. Macrophages were identified as cells expressing high levels of F4/80⁺ that were also CD11b⁺ and CD11c^neg. The percentage of AM and DC obtained from flow cytometry was used to calculate the total number of these cells from each lung by multiplying the frequency of each population by the total number of leukocytes identified within the sample.

Serum incubation CFU assay.

H99 wt, Δvad1, VAD1-complemented, and Δlac1 strains were combined with freshly isolated mouse serum and incubated at 37°C in triplicate, and CFU counts were determined on Sabouraud dextrose agar after incubation at room temperature for 48 h. Colony counts were performed and expressed as log CFU/well.

Cytokine protein measurements.

Cytokine levels in lung tissues were analyzed using the Bio-Plex protein array system (Luminex-based technology; Bio-Rad Laboratories, Hercules, CA) according to the manufacturer's directions. Briefly, lung tissue was excised and homogenized in ice-cold, sterile PBS (1 ml) and kept on ice. An aliquot (10 μl) was taken to quantify the pulmonary fungal burden, and an antiprotease buffer solution (1 ml) containing PBS, protease inhibitors (inhibiting cysteine, serine, and other metalloproteinases), and 0.05% Triton X-100 was added to the homogenate, which was then clarified by centrifugation (800 × g) for 5 min.

Histological analysis.

Lungs were fixed by inflation with 1 ml of 10% neutral buffered formalin, excised en bloc, and immersed in neutral buffered formalin. After paraffin embedding, 5-μm-thick sections of organs were cut and stained with hematoxylin and eosin. Sections were analyzed using light microscopy.

Statistics.

Data (means ± standard errors of the means [SEM]) for each experimental group were derived from at least two independent infections. Statistical significance was calculated using Student's t test for individual paired comparisons or the t test with Bonferroni adjustment when multiple groups were compared. A P value of <0.05 after controlling for multiple comparisons was considered to be statistically significant. Analyses were performed using Prism 5.0 (GraphPad Software Inc., La Jolla, CA).

RESULTS

A Δvad1 deletant strain of C. neoformans shows accelerated clearance from lungs of mice and absence of dissemination to spleen and brain.

Pulmonary growth and clearance of a wt C. neoformans strain were compared to those of a Δvad1 mutant strain using a CBA/J mouse intranasal infection model. A congenic Δlac1 strain was used as an additional control strain, as the Δvad1 strain exhibits reduced expression of laccase (36). Mice were inoculated intranasally with 10⁶ CFU, and the pulmonary fungal burden kinetics was evaluated. As shown in Fig. 1A, the wt strain showed progressive growth in lungs, resulting in a 2-orders-of-magnitude expansion of cryptococcal burden by day 21 (week 1, 2.1 × 10⁵; week 3, 5.7 × 10⁶). This was in sharp contrast with fungal burdens of the Δvad1 mutant strain, which showed a gradual decrease over time (Fig. 1A; week 1, 3.8 × 10²; week 3, 3.6 × 10²). Progressive growth of yeast was also observed in lungs infected with two additional control strains, a Δvad1::VAD1-complemented strain and a Δlac1 mutant strain, indicating that the effect of VAD1 deletion on microbial growth was specific and independent of that due to reductions of laccase expression during the afferent phase of the immune response to C. neoformans. An evaluation at an additional time point at week 8 demonstrated that clearance of the Δvad1 strain was progressive. As expected, no animals infected with wt, Δlac1, or Δvad1::VAD1 strains survived up to this point. In contrast, mice infected with the Δvad1 strain not only survived until the conclusion but showed only a minimal level of infection (i.e., either below or slightly above the 10-CFU/lung limit of detection), indicating that the Δvad1 strain was progressively cleared to near-sterility over 2 months (Fig. 1A).

Fig 1.

A Δvad1 deletant strain displays accelerated clearance from lungs of mice and poor rates of dissemination to spleen and brain. (A) Mice were inoculated intranasally with 1 × 10⁶ CFU of the indicated strains and pulmonary fungal burdens determined at 1-week intervals and at week 8 (n = 8). (B and C) Mice were infected as for panel A and CFU counts of spleens (B) (n = 8) and brains (C) (n = 8) determined at the indicated times. Results are expressed as mean CFU per organ ± SEM.

We further evaluated the effect of the cryptococcal VAD1 deletion on extrapulmonary dissemination of organisms. To illustrate this multistep process, which includes escape from the lung and infection of secondary lymphoid organs followed by CNS dissemination (29), fungal burdens were measured in spleens (Fig. 1B) and brains (Fig. 1C) of infected mice. While the wt and VAD1-complemented strains were able to successfully disseminate to spleens and brain, the Δvad1 deletion strain lost the ability to disseminate (Fig. 1C). This could only be partially accounted for by a defect in laccase activity, as parallel infections with the Δlac1 strain retained an ability to disseminate to spleens but not to brains of the infected mice, as previously described (27). Thus, VAD1 expression is required for progressive pulmonary growth and extrapulmonary dissemination of C. neoformans, in a largely laccase-independent manner.

VAD1 deletion significantly limited pathological lesions in pulmonary C. neoformans infection.

Since the lung is the primary organ responsible for containment of C. neoformans and prevention from dissemination, we examined lung appearance and histopathology after infection, to generate hypotheses regarding host factors governing effective control of the Δvad1 mutant strain. Consistent with the quantitative CFU analysis, the Δvad1 strain showed an attenuated inflammatory response evidenced by reduced lung enlargement (swelling/cryptococcomas) in mice infected with the mutant strain (Fig. 2A). At the same time point, lungs of mice infected with wt C. neoformans contained large numbers of fungi, which formed large nodular clusters and masses (cryptococcomas). These C. neoformans clusters were surrounded by marginal zones of mixed leukocyte infiltrates forming diffuse granulomas containing cryptococci. The areas of lung consolidation were significantly less extensive in the lungs of mice infected with the Δvad1 strain in comparison with lungs infected with the wt strain (Fig. 2B). Under high-power magnification, we detected a decreased presence of both cryptococcal organisms and leukocytes in the lungs infected with the Δvad1 strain, as opposed to the lungs infected with control strains (wt, complemented, and Δlac1 strains), which contained a mixed leukocyte infiltration with occasional giant cells (Fig. 2C). Furthermore, histological features of the repair process, including groups of fibroblasts surrounded by modest mononuclear cell infiltration, were noted in the mutant but not the wt strain. Thus, infection with the Δvad1 mutant resulted in significant protection from pathological changes, while infections with wt, Δlac1, and VAD1-complemented strains exhibited focal destruction of lung tissue due to a chronic, nonprotective inflammatory response and progressive expansion of the yeast burden in the lungs. Collectively, these data suggest that C. neoformans requires the VAD1 gene for progressive growth and pathogenicity, largely independent of laccase expression.

Fig 2.

Infection with a vad1Δ deletant strain results in reduced inflammatory burden in lungs of mice. (A) Gross pathology of resected lungs of infected mice. Mice were inoculated as described in the Fig. 1 legend and harvested at 3 weeks from mice infected with the wt and Δvad1 strains. (B and C) Histology overview of lungs from panel A: low-power (2×) (B) and higher-power (200×) (C) views of lungs. Red arrows, C. neoformans cells.

VAD1 deletion results in a reduced inflammatory response in C. neoformans-infected lungs.

To further characterize the effect of cryptococcal VAD1 expression on the inflammatory response in terms of quantitative measures of overall inflammation, we compared total leukocyte (CD45⁺), CD4⁺ T-cell, and CD8⁺ T-cell numbers from the lungs of infected mice. Pulmonary cell counts were determined from enzymatic digests of whole lungs as described in Materials and Methods. Lungs were excised at weeks 1, 2, and 3 postinfection and leukocytes isolated from whole lungs by enzymatic digestion and mechanical dispersion. As shown in Fig. 3A, the numbers of total leukocytes and CD4⁺ T cells in the lungs infected with the Δvad1 strain were significantly lower than cell numbers in mice infected with the other strains. However, the number of CD8⁺ T cells in the lungs of mice infected with the Δvad1 strain showed only minor reductions at week 2 that were not significantly different from levels in the wt strain at week 1 or week 3 (Fig. 3C). To determine possible contributions of the reduced CD4 accumulation in retention of the low but persistent fungal burden, mice were similarly infected with Δvad1 mutant strains at the same inoculum and, beginning at 3 weeks, were given a combination of CD4 and CD8 monoclonal antibodies in a regimen used previously to deplete these cell reservoirs (38). Interestingly, CD4/CD8 depletion resulted in no increase in Δvad1 pulmonary burden (2 ± 1.5 versus 11 ± 5; P = 0.15, n = 5), indicating that the progressive lung clearance over the first month postinfection was not related to changes in host CD4/CD8 cell populations and that the Δvad1 strain can be contained by the host immune system even in the absence of CD4 and CD8 T cells.

Fig 3.

Lung leukocyte and CD4⁺ T cell recruitment is reduced in Δvad1 pulmonary infections. The indicated strains were inoculated as in the Fig. 1 legend, and at the indicated times, pulmonary total leukocytes (CD45⁺ [A], CD4⁺ [B], and CD8⁺ [C] T cells) were determined from enzymatic digests of whole lungs as described in Materials and Methods. Recruited leukocytes were expressed as total numbers from infected mice minus the mean number of leukocytes in uninfected mice and expressed as the mean number of leukocytes per mouse lung ± SEM; n = 6 per time point pooled from two independent experiments. (D) Fungal cells were inoculated as in the Fig. 1 legend, and, at the indicated times, pulmonary inflammatory monocytes were identified as CD45⁺ CD11c⁻ CD11b⁺ Ly-6c^high by flow cytometry as detailed in Materials and Methods. (E) Cells were harvested from lungs at 2 weeks postinfection, and recruitment of F4/80⁺ macrophages was quantified by the presence of CD45⁺ CD11c⁻ F4/80⁺ CD11b⁺ cells by flow cytometry.

Recruited dendritic cell and macrophage populations are present but mildly reduced in mice infected with a cryptococcal Δvad1 deletant strain.

Dendritic cell (DC) and macrophage accumulation is crucial for the control of cryptococcal lung infections during both innate and adaptive phases of the immune response (5, 6, 8–10). DC accumulation in response to fungal infections in the lung results from the differentiation of Ly-6C^high inflammatory monocytes (3). We used a plot of CD11c-versus-forward scatter to identify CD11c-negative cells, and inflammatory monocytes were identified as Ly-6C^high CD11b⁺ cells (see Fig. S2 in the supplemental material). These data showed pulmonary Ly-6C^high monocyte accumulation at both weeks 1 and 2 that was reduced from that of the VAD1-complemented strain and showed a similar trend in comparison with the wild-type and Δlac1 strains (Fig. 3D). These trends were verified using CD11c⁻/F4/80⁺ CD11b⁺ gating (F4/80⁺ monocytes) at week 2. We observed significantly reduced accumulation of F4/80⁺ monocytes in lungs during infection with the Δvad1 deletant strain compared to infections with all VAD1-expressing groups at 2 weeks postinfection (Fig. 3E). Furthermore, a mild but not as pronounced decrease in pulmonary F4/80⁺ cell population occurred in response to cryptococcal laccase deletion alone. Thus, deletion of VAD1 results in improved control of pulmonary C. neoformans growth that occurs without an increase in DC or macrophage accumulation in infected lungs.

VAD1 deletion alters the balance between Th-polarizing, pro- and anti-inflammatory cytokines.

Our prior studies showed that pulmonary clearance in infected lungs correlates well with the type of the adaptive immune response (11, 13, 18–20). Thus, we analyzed cytokine profiles in lung homogenates at weeks 1, 2, and 3 postinfection. Focusing on the earliest time point (1 week) when clearance of the Δvad1 strain was maximal and fungal burdens were most equivalent, both gamma interferon (IFN-γ) (Fig. 4A) and tumor necrosis factor alpha (TNF-α) (Fig. 4C) were expressed at equivalent levels in mice infected with the Δvad1 strain and in those infected with the wt strain during the innate phase of response to C. neoformans. In contrast, IL-4 induction and IL-5 induction at the same early time point, associated with nonprotective responses, were both dramatically diminished. This, however, occurred with concurrent decreases in IL-12 production and good induction of IL-10 and IL-13, which in turn could shift the balance toward a nonprotective response and may have prevented more rapid clearance of the Δvad1 strain. At the later time point, infection with the Δvad1 strain showed overall lower levels for most of the cytokines associated with Th1, Th2, and Th17 production that became more prominent as the fungal burden of that strain decreased. Specifically, diminished production of Th1-type cytokines (IFN-γ, IL-12, TNF-α [Fig. 4A to C]), Th2-type cytokines (IL-4 and IL-5 [Fig. 4D and E]), and IL-17 (Fig. 5A) was observed at various points postinfection. Interestingly, the regulatory cytokines IL-10 and IL-13 (associated with chronic Th2 response [Fig. 5A and 4F]) were expressed at levels equivalent to that after infection with the wt strain. Furthermore, a decrease in proinflammatory factor induction, associated with VAD1 deletion, was observed for a large group of chemokines, including CCL2, CCL3, CCL4, CCL5, CXCL1, and CCL26 (see Fig. S2 in the supplemental material), and general proinflammatory cytokines IL-1α, IL-1β, and granulocyte colony-stimulating factor (G-CSF), consistent with reduced inflammation and reduced cytokine production across the board, likely due to reductions in fungal burden. Laccase deletion appeared to have little effect on cytokine levels compared to that in wt cells, although when an effect was present, it was most apparent in the latter stages of infection, as was the case with IFN-γ and IL-17 (Fig. 4A and 5A); this is consistent with our reports that cryptococcal laccase can interfere with polarization of the immune response in infected lungs (32).

Fig 4.

Pulmonary infection with a Δvad1 strain is associated with selective retention of key Th cytokines. Mice were inoculated as in the Fig. 1 legend, and at the indicated times, lung were harvested and assayed by enzyme-linked immunosorbent assay (ELISA) from supernatants of lung homogenates. (A) IFN-γ; (B) IL-12; (C) TNF- α; (D) IL-4; (E) IL-5; (F) IL-13. Bars represent mean cytokine concentrations ± SEM (pM/ml; n = 6). *, P < 0.05; **, P < 0.01; ***, P < 0.001 versus that of the Δvad1 strain. Dotted lines indicated levels of cytokines in uninfected mice ± SEM.

Fig 5.

Pulmonary infection with a Δvad1 strain is associated with reductions in Th17 cytokines and alterations in key cytokine ratios. Indicated strains were inoculated as in the Fig. 1 legend and pulmonary cytokines analyzed 1, 2, and 3 weeks postinfection. (A) IL-17; (B) IL-13; (C) IL-10; (D to F) ratios of IL-4/IFN-γ (D), IL-10/TNF-α (E), and IL-10/IL-12 (F). Bars represent mean cytokine concentrations ± SEM (pM/ml; n = 6). *, P < 0.05; **. P < 0.01; ***, P < 0.001 versus that of the vad1Δ strain. Dotted lines indicated levels of cytokines in uninfected mice ± SEM.

Since our readouts were complicated by significantly lower fungal burden/inflammation in the Δvad1 group, we next compared ratios between the opposing cytokine groups as previously described (18). This approach was used to highlight Vad1-specific effects on cytokine balance, since differences in fungal burden were likely to affect all cytokine groups. The IL-4/IFN-γ ratio (Fig. 5C) demonstrates that the IL-4/IFN-γ cytokine balance for the Δvad1 group at each analyzed time point shows the most pronounced shift toward a protective Th1 polarization. This observation matches well with the improved clearance of the Δvad1 strain and demonstrates that cryptococcal VAD1 expression contributes to a nonprotective Th2 response, rather than to the protective Th1. However, IL-10/TNF-α and IL-10/IL-12 ratios in lungs infected with the Δvad1 strain were both significantly skewed toward regulatory/anti-inflammatory responses at weeks 2 and 3 postinfection, consistent with the decreased inflammatory response and ongoing repair process observed in histological samples of lungs infected with the Δvad1 strain. Thus, cryptococcal VAD1 expression contributes to the generation of a cytokine environment that supports proinflammatory and immunopathological aspects of the anticryptococcal host response. In an attempt to improve equilibration of fungal burdens, we inoculated mice with the same inoculum of the vad1 strain (1 × 10⁶ CFU) but a low inoculum of the wt strain (1 × 10⁴ CFU), followed by an additional inoculation of the Δvad1 strain (1 × 10⁷ CFU) at week 2 and analysis of cytokines at week 3. As shown in Fig. S4 in the supplemental material, while dissemination to brain still did not occur with the higher inoculum, the lung fungal burdens of the two strains at week three more closely approximated each other. This resulted in higher levels of TNF-α but continued reductions in IFN-γ and IL-17 at these later time points, suggesting a continued role of TNF-α in innate clearance of the fungus under these conditions.

VAD1 deletion promotes early/innate clearance of C. neoformans in the lungs and enhances fungistatic macrophage and serum activities in vitro.

Thus far, our data demonstrate that VAD1 deletion resulted in fungal clearance that begins during the first week of infection and progresses with a mild inflammatory response and minimal damage to host tissues. This could be a result of an inability of the microbe to adapt to the host tissue environment or decreased resistance of the microbe to the innate mechanisms of host defenses. To explore these scenarios and to determine the onset of VAD1-associated effects on pulmonary control of C. neoformans, fungal burdens of the wt and Δvad1 strains were compared over a wider range of early time points (Fig. 6A). The fungal burdens of the wt strain remained stable during the first 2 to 3 days of infection, after which exponential growth of organism was observed, indicating that, following a brief adaptation to the host tissue environment, wt C. neoformans can grow in an unopposed fashion in the infected mice. In contrast, clearance of the Δvad1 strain began between days 1 and 2, resulting in a log drop by day 3, compared to the initial lung burden at 1 h.

Fig 6.

Early pulmonary retention of the Δvad1 strain is associated with retained but weakened ability to grow in mouse serum and macrophages. (A) Indicated strains were inoculated as in the Fig. 1 legend and pulmonary fungal burdens determined at indicated times postinfection. (B) Fungal suspensions of the indicated strains were combined with freshly isolated mouse serum and aliquots removed at the indicated times and assayed for CFU (n = 3). (C) Phagocytic index calculated after opsonization of the indicated fungal strain with either fresh mouse serum (+serum) or capsular monoclonal antibody 18B7 (+18B7); n = 6; mean ± SEM; ***, P < 0.001. (D) Indicated strains were opsonized with anti-capsular monoclonal 18B7 and incubated with J774.16 cells, and internalized/adherent cells were assayed for CFU at indicated time points after macrophage lysis (results are the averages of 4 independent experiments ± SEM).

Given that neither enhanced CD4⁺/CD8⁺ T cell nor DC/macrophage recruitment could explain the improved pulmonary control of fungal cells, we next examined the effect of other innate mechanisms known to be important for fungal control. Previously, we had shown that the Δvad1 mutant grew normally on 2% glucose, 2% Bacto peptone, 1% yeast extract (YPD) agar at 37°C (36), indicating the fairly robust growth potential of the strain. However, many humoral factors found in nonimmune serum, such as complement, may contribute to innate defenses and growth under nutrient-limiting conditions may further stress the fungus. Thus, we examined the effect of VAD1 deletion on C. neoformans growth in nonimmune fresh mouse serum. As shown in Fig. 6B, the Δvad1 mutant was capable of growth in mouse serum, although at a somewhat lower rate than wild-type cells, suggesting that the Δvad1 strain can successfully adapt and grow under conditions of host serum.

We next examined the efficiency of microbial uptake by macrophages and the survival of macrophage-associated fungus. As shown in Fig. 6C, phagocytosis of the Δvad1 deletant was significantly improved compared to that of the wt strain, in an assay utilizing J774 cells and anti-capsular antibody 18B7 as opsonin. In contrast, previous studies have shown no change in phagocytic index after LAC1 deletion (39). Despite the improved phagocytosis, the Δvad1 mutant strain retained the ability to grow in J774 cells, although at a reduced rate compared to the wt strain (Fig. 6D). Additional experiments under the same conditions determined that the Δlac1 mutant displayed mildly reduced growth rates in J774 cells intermediate between those of wt and Δvad1 fungal cells (48 h: wt, 5.1 ± 0.2; Δvad1, 4.3 ± 0.2; Δlac1, 4.7 ± 0.2; log₁₀ CFU ± standard error of the mean [SEM]; n = 5), similar to that reported previously (39). Collectively, these data indicate that VAD1 expression contributes to increased resistance of C. neoformans to a number of innate host defense mechanisms, such as the fungistatic effects of nonimmune serum, phagocytic uptake by macrophages, and macrophage-induced fungistasis, which together could partially explain the improved innate control of the Δvad1 strain in infected lungs.

Cryptococcal VAD1 expression interferes with early innate cytokine and chemokine production in C. neoformans-infected lungs.

Interference with innate defense mechanisms by pathogens is frequently related to insufficient evasion of recognition by cells/receptors of the innate immune system. Thus, our final goal was to determine if VAD1 has an important role in dampening innate recognition and subsequent cytokine responses in C. neoformans-infected lungs. We compared early cytokine and chemokine production within the first 3 days after infection, when the fungal burdens were also most equivalent, and observed that cryptococcal VAD1 deletion resulted in an increase in KC, MIP-1β, and IL-1α production as early as 24 h postinfection. These effects were not only independent of laccase but showed trends that were opposite those of infections with strains having a LAC1 deletion. Similarly, a less pronounced relationship was observed for early IL-12p70 production, in which VAD1 deletion produced only a trend compared to the infection with the wt strain; however, laccase deletion unmasked a significant difference. We further observed a significant increase in day 3 IFN-γ production in lungs infected with the Δvad1 strain compared to that in other groups, suggesting that VAD1 expression results in a reduction in early innate responses that controlled C. neoformans growth during the first 72 h postinfection. Thus, cryptococcal VAD1 expression independently dampens the induction of early/innate signals contributing to the evasion of the innate host defenses by the yeast and, most likely, to limited innate control of VAD1-expressing C. neoformans strains in the infected lungs.

DISCUSSION

In this study, we demonstrate that cryptococcal VAD1 expression is required for fungal virulence during pulmonary infections and contributes to the ability of the organism to induce lung pathology. This was based on progressive clearance of a Δvad1 deletant strain that was shown previously to retain important virulence attributes, including growth at 37°C and polysaccharide capsule production (36). In the present studies, the Δvad1 deletant strain was also found to exhibit a retained ability to grow in both serum and macrophages, though at reduced rates. The data further suggest that VAD1 expression prevents rapid innate recognition of C. neoformans and renders the yeast resistant to its effective elimination by mechanisms of the host's innate immunity. Interestingly, most of the VAD1 effects appear to be unrelated to cryptococcal laccase expression, of which VAD1 is a critical inducer and regulator. These conclusions are based on the following effects of cryptococcal VAD1—but not LAC1—deletion observed in the mouse model of cryptococcal infection: (i) improved innate clearance of C. neoformans starting from the early days of infection, (ii) successful control of C. neoformans growth in mice after combined CD4 and CD8 T-cell depletion, (iii) inability of mutant yeast to induce lung pathology and to disseminate from infected lungs, (iv) improved production of early proinflammatory cytokines and chemokines in the infected lungs, and (v) improved phagocytosis and diminished resistance of the fungus to macrophage and serum-induced fungistasis. Thus, in the absence of VAD1 expression, a highly virulent wt strain (H99) lost many of its major virulence attributes.

The present studies focused on the effects of an important regulator of cryptococcal pathogenicity and the course of the immune response to C. neoformans in a mouse model; therefore, we chose an inhalational model of infection that closely follows the natural history of infection, including (i) the primary site of infection (the lungs), (ii) subsequent extrapulmonary dissemination occurring with progression of disease, and (iii) gradual development of the immune response capturing both innate and adaptive phases of host response to the microbe (29). Our previous studies established that a VAD1 deletant mutant of C. neoformans exhibits attenuated virulence in mouse models and reduced production of laccase (36). Since laccase is an important factor modulating pulmonary host responses (9, 32), in addition to wt and VAD1-complemented controls, we used parallel infections with a congenic Δlac1 strain. These controls enabled us to dissect which effects of VAD1 deletion could be attributed to defective laccase production and which were laccase independent.

Among the attributes of high virulence displayed by the wt H99 strain, the most important include (i) rapid evasion of the innate responses, especially resistance to antimicrobial factors in plasma/serum, anti-phagocytic properties, and interference with macrophage activation, (ii) its ability to skew the adaptive immune response toward Th2, and (iii) a high potential to disseminate from infected lungs, contributing to its high CNS tropism (29). Blending of these important virulence attributes identifies the serotype A H99 strain as one of the most virulent cryptococcal clinical isolates used in experimental infections. In fact, strain H99 been reported to be lethal for all studied immunocompetent strains of mice at a wide range of tested inocula (13, 18, 19, 26, 28, 40). The loss of these major aspects of cryptococcal virulence associated with VAD1 gene deletion signifies that Vad1 plays a vital role in conferring cryptococcal virulence within mammalian hosts.

Interestingly, VAD1 expression appears to be linked with cryptococcal immune responses rather than exclusively with the pathogenic fitness of the microbe (41). Several features of the Δvad1 mutant support this view of a role of VAD1 in expressing cryptococcal virulence. The Δvad1 mutant displays (i) no apparent decrease in growth of the microbe in vitro at 37°C (36), (ii) a growth rate similar to that of the wt strain in the presence of macrophages during the initial 9 h of incubation, (iii) relatively good (although diminished) ability to grow in mouse serum, and (iv) survival in murine lungs within the initial 24-h postinfection period, comparable to that of a wt strain. These data indicate that the Δvad1 strain has not completely lost its ability to survive or grow under the conditions present in host tissues but, rather, there is a contribution from a diminished ability to evade immune recognition by the components of the innate immune system and/or to counteract at least some of its protective mechanisms. Thus, it is interesting to compare the present studies with that of a strain, derived from the same fungal background, defective in calcineurin (Δcna1), that has a severe defect in fungal fitness by virtue of an inability to grow at 37°C (42). In the previous studies, higher inocula of the Δcna1 strain failed to increase the fungal burden 1 week after inoculation whereas reinfection with a larger inoculum of the Δvad1 strain was able to significantly increase pulmonary fungal burdens (see Fig. S4 in the supplemental material). However, it is difficult to completely separate effects of fungal fitness versus innate control. Indeed, reductions in TUF1, which is involved in mitochondrial sufficiency and was described previously in the Δvad1 strain (36), could reduce fungal fitness during mouse infections and may help to explain the modest reductions in growth in macrophages and serum. Nevertheless, progressive and significant clearance in pulmonary fungal burden in intact mice versus the mutant's positive growth in serum and the J774 macrophage-like cell line suggests an important contribution of host mechanisms to fungal clearance of the Δvad1 mutant.

Regarding the host response, the following findings suggest that diminished virulence displayed by the Δvad1 mutant resulted from failure to effectively counteract the innate mechanisms of host defense: (i) improved early cytokine/chemokine induction in the lung (Fig. 7A to C and E), (ii) the absence of suppression of the innate IFN-γ on day 3 (Fig. 7D), which marks the beginning of the progressive growth of strain H99 (Fig. 6A), and (iii) no effect of CD4/CD8 cell depletion on fungal burden in mice infected with the Δvad1 strain. These all point out that failure to effectively counteract the innate host defense was an important mechanism of the diminished virulence displayed by the Δvad1 strain. In the absence of VAD1 expression, mice appear to overcome the wt-induced IFN-γ suppression (which occurs at the initiation of progressive wt fungal growth), potentiating better control of the Δvad1 mutant. This, together with improved phagocytosis and better suppression of growth by macrophages, could account for the improved early innate control of the Δvad1 strain. However, subsequently, the wt strain triggers additional IFN-γ production in the lungs by day 7 but is accompanied by an increase in IL-4, while the Δvad1 strain does not induce IL-4, thus maintaining a low IL-4/IFN-γ ratio (Fig. 5C), known to be beneficial for cryptococcal clearance (19, 43). Increased production of inflammatory cytokines/chemokines such as KC and retained production of TNF-α likely contributed to effective innate clearance of the mutant strain as suggested by previous studies (44, 45). High TNF-α levels induced during the initial infective period followed by reductions could be due to relatively low levels of secreted capsular polysaccharide, as reported previously (46). Increased phagocytosis of the Δvad1 strain may also suggest that alveolar macrophages contribute to the retention of these inflammatory cytokines, although other cells such as resident NK cells could also be a source (47). The result of the altered cytokine balance, along with the direct effects of Δvad1 deletion on macrophage efficiency, is that we observe progressive clearance of the Δvad1 strain, in contrast with a progressive growth of the wt strain in lungs. Subsequent low cytokine levels elicited by the mutant are likely due to reduced antigen load, although the IL-4/IFN-γ ratio remains favorable, thereby promoting further control/clearance of the Δvad1 strain in the presence of reduced inflammation. A control experiment, reinfection with a supplemental fungal inoculum in the Δvad1-infected mice (see Fig. S4 in the supplemental material), followed by cytokine analysis 1 week later failed to show increases in IFN-γ and IL-17 levels relative to those in the wt strain, suggesting that late suppression was not exclusively due to low fungal burdens. However, TNF-α levels did approach that of the wild-type strain after reinfection, documenting that the presence of Δvad1 was a strong trigger of TNF-α production in the lungs throughout all stages. Such continuous induction of TNF-α would be important to maintain continued clearance of the Δvad1 strain over the ensuing weeks. Again, comparison with previous studies of mouse infections with the Δcna1 mutant identifies important differences between host responses to strains differing in fungal fitness. In the earlier studies, the Δcna1 mutant failed to induce a CD8 response whereas the Δvad1 strain did, suggesting that a degree of fungal persistence is required for CD8 recruitment. Furthermore, the Δcna1 mutant showed progressive IFN-γ and TNF-α induction during the later stages of infection, despite rapid clearance of the organism, whereas the Δvad1 mutant elicited a more vigorous early response, showing reductions in levels of these cytokines compared to those in the wt strain at 3 weeks. These findings suggest that an intact innate response is capable of the control of weakly virulent strains such as the Δvad1 mutant but is overwhelmed by the suppressive properties and vigorous growth of the parental wt strain.

Fig 7.

Deletion of VAD1 is associated with increased innate cytokine and chemokine production in C. neoformans-infected lungs. Indicated strains were inoculated as in the Fig. 1 legend and pulmonary cytokines analyzed 1, 2, and 3 days postinfection. (A) KC; (B) IL-12; (C) CCL4/MIP-1β; (D) IFN-γ; (E) IL-1α. Bars represent mean cytokine concentration ± SEM (pM/ml; n = 6). *, P < 0.05; **, P < 0.01; ***, P < 0.001 versus that of the vad1Δ strain. Dotted lines indicated levels of cytokines in uninfected mice ± SEM.

Although Vad1 is a major regulator of cryptococcal laccase expression, interestingly, the majority of the effects of Vad1 expression were laccase independent. Laccase deletion resulted in modest decreases in fungal burden in the lung and prevented CNS dissemination; however, it did not prevent progressive pulmonary growth of C. neoformans and eventual death of the animals infected with the Δlac1 strain (32). The effects of laccase have been attributed to moderate effects on cellular immune polarization, specifically inhibition of early IL-17 production in the lungs, the effects on diminished systemic Th2 polarization, and the organisms' inability to invade the CNS (32). These effects were unlikely to be responsible for improved clearance of the Δvad1 strain, which occurred largely independent of the adaptive host responses. These data support the notion that VAD1 plays a role in the regulation of additional factors that interfere with the innate host defense, in addition to laccase. Indeed, the first VAD1 studies implicated roles for PCK1, TUF1, and MPF3 as VAD1-regulated genes involved in gluconeogenesis, mitochondrial sufficiency, and cell wall integrity, respectively (36). Future studies will be needed to identify additional VAD1-dependent factors that alter innate recognition of the fungus.

In summary, our studies defined a novel role of the cryptococcal Vad1 protein as a central regulator of cryptococcal virulence. Our data suggest that an important group of Vad1 effects are related to its role in promoting microbe resistance to mechanisms of the innate host defenses, beyond what could be explained by fungal fitness alone. Such host response comparisons of selective fungal mutants can thus allow a “molecular dissection” of the relative contributions of innate and adaptive host responses alongside fungal pathogenic fitness that comprise the overall virulence composite.