Role of Complement in Protection against Cryptococcus gattii Infection

Abstract

Previous studies have shown that the alternative pathway of complement activation plays an important role in protection against infection with Cryptococcus neoformans. Cryptococcus gattii does not activate the alternative pathway as well as C. neoformans in vitro. The role of complement in C. gattii infection in vivo has not been reported. In this study, we used mice deficient in complement components to investigate the role of complement in protection against a C. gattii isolate from an ongoing outbreak in northwestern North America. While factor B-deficient mice showed an enhanced rate of death, complement component C3-deficient mice died even more rapidly, indicating that the alternative pathway was not the only complement pathway contributing to protection against disease. Both C3- and factor B-deficient mice had increased fungal burdens in comparison to wild-type mice. Histopathology revealed an overwhelming fungal burden in the lungs of these complement-deficient mice, which undoubtedly prevented efficient gas exchange, causing death. Following the fate of radiolabeled organisms showed that both factor B- and C3-deficient mice were less effective than wild-type mice in clearing organisms. However, opsonization of C. gattii with complement components was not sufficient to prolong life in mice deficient in complement. Killing of C. gattii by macrophages in vitro was decreased in the presence of serum from factor B- and C3-deficient versus wild-type mice. In conclusion, we have demonstrated that complement activation is crucial for survival in C. gattii infection. Additionally, we have shown that the alternative pathway of complement activation is not the only complement pathway contributing to protection.

The complement system consists of a cascade of serum proteins that are involved in opsonization, membrane lysis, and chemotaxis. There are three pathways through which complement can be activated: classical, alternative, and lectin. Complement components C3 to C9 participate in all three complement cascade pathways. C1q is used only in the classical pathway, and factor B is used only in the alternative pathway. While C4 is used in both the classical and lectin pathways, recently it was reported that the lectin pathway can function in the absence of C2 and/or C4 if the alternative pathway is intact (23). Mannose binding lectin (MBL) (24), which is used only in the lectin pathway, exists in two forms in rodents (MBL-A and MBL-C) while only one form is found in humans (8). MBL-A and MBL-C have different levels in serum and differ in their affinity for d-glucose and α-methyl-d-glucose but have redundant function (24).

Cryptococcus spp. are fungal pathogens that possess a polysaccharide capsule composed mainly of glucuronoxylomannan (GXM). The capsule is antiphagocytic and anti-inflammatory but is known to activate the alternative complement pathway (11). Previously, Cryptococcus spp. were identified serologically and were all considered as one species, Cryptococcus neoformans. More recently, two species have been designated, C. neoformans (serotypes A and D) and Cryptococcus gattii (serotypes B and C). A key difference between the two species is that C. neoformans tends to infect immunocompromised individuals while C. gattii generally infects apparently immunocompetent people (15). Cryptococcus spp. most commonly cause pulmonary and central nervous system infections in humans (15, 25) and mouse models (3).

In 1999, an outbreak of C. gattii began on Vancouver Island, British Columbia, infecting people, companion animals, and porpoises. Strain A1MR265 is the major clinical reference isolate from the Vancouver Island outbreak. Recently, a case of cryptococcosis caused by the strain predominant in Vancouver Island was identified in Puget Sound, WA (27). A total of eight human cases have been reported in Washington State in the past 2 years; four of these individuals had not traveled out of state (29). Nine cases have been reported in Oregon (L. Hoang, presented at the 108th General Meeting of the American Society of Microbiology, Boston, MA, 1 to 5 June 2008). The potential spreading of a strain of Cryptococcus capable of infecting immunocompetent people is cause for concern, and C. gattii infection is now a reportable disease in Washington State (29).

The Kozel laboratory and others have shown that Cryptococcus spp. strongly activate the alternative pathway of the complement cascade (5, 14) while the polysaccharide capsule blocks the activation of the classical pathway that occurs at the cell wall in nonencapsulated strains (13). The capsule serves as a site for activation and deposition of C3 fragments, mainly iC3b, which promote phagocytosis of the yeast (12). C. gattii does not appear to activate the alternative pathway as potently as C. neoformans (28, 31). One study found that C. gattii binds fewer C3 molecules than C. neoformans (28). A later report indicated that while the maximum amount of bound C3 did not differ significantly between species, there was more rapid accumulation of C3 on C. neoformans before a steady state was achieved (31). In the absence of C5 in mouse strains B10.D2/oSn, DBA/2, and A/J, C. neoformans infection proceeds to a fatal pneumonia with higher fungal burdens in the blood, brain, lungs, and liver than in complement-sufficient animals (6, 21, 22). Together, these studies indicate a significant role for the alternative pathway of complement in protection against C. neoformans. While C. gattii has been reported to not activate the alternative pathway as vigorously as C. neoformans in vitro, the role of complement in in vivo C. gattii infection has not been determined.

In this study, we investigated the role of complement pathways in protecting against infection with C. gattii. Mice deficient in complement components C1q, C4, C3, and factor B have been generated on the C57BL/6J background (2, 16, 30). Using these mice and mice treated with cobra venom factor (CVF), which depletes C3 and prevents cascade progression from C3 to C9, we have now shown that complement activation plays an essential role in delaying disease progression in mice infected with C. gattii.

MATERIALS AND METHODS

Mice.

C57BL/6 mice were purchased from Charles River Laboratories or The Jackson Laboratories. C1q^−/− mice were made by Mark Walport (The Wellcome Trust, London, United Kingdom) (2) and provided by Andrea Tenner (University of California, Irvine, CA). C3^−/− and C4^−/− mice were a gift from Michael Carroll (Harvard Medical School, Boston, MA) (30). Factor B-deficient mice were provided by Rick Wetsel (University of Texas, Houston, TX) (16). All animals were maintained in accordance with the Chancellor's Animal Research Committee at the University of California, Los Angeles.

Murine infection.

Three milliliters of Sabouraud's dextrose broth (Becton Dickinson, Franklin Lakes, NJ) was inoculated with a clinical isolate of C. gattii (strain A1MR265; provided by James Kronstad, University of British Columbia, Canada) and incubated at 37°C with shaking for 48 h. The culture was diluted into 25 ml of Sabouraud's dextrose broth and incubated at 37°C with shaking for 18 h. C. gattii cells were pelleted by centrifugation (1,942 × g for 10 min at 4°C) and washed twice in 50 ml of cold phosphate-buffered saline (PBS). Cryptococci were resuspended in approximately 10 ml of cold PBS, counted using a hemacytometer, and diluted to the desired concentration in cold PBS. Approximately 2 × 10⁶ C. gattii cells in 0.2 ml were injected via lateral tail vein, with the actual inoculum determined by plating on Sabouraud's dextrose agar plates (Becton Dickinson). Animals were monitored daily for survival and euthanized if they exhibited hydrocephalus or paralysis. Temperatures were recorded via a rectal digital thermometer probe (Physitemp Thermalert TH-5; Clifton, NJ).

CVF.

One day prior to infection with Cryptococcus, CVF (Quidel, San Diego, CA) was injected intraperitoneally at 5 units per mouse two times at 4 h apart. Mice continued to be injected with 5 units of CVF every 4 days for 1 month. C3 deficiency following initial treatment was confirmed by Western blotting (data not shown).

Fungal burden.

Mice were dissected, and brains, lungs, livers, spleens, and kidneys were removed. Half of each organ was placed in cold PBS, and the remaining half was placed in 10% formalin for histopathology. Organs in PBS were weighed, homogenized in 5 ml of cold PBS using a Stomacher 180 (Seward, West Sussex, United Kingdom), and plated on Sabouraud's dextrose agar plates at various dilutions. Plates were incubated at room temperature for 3 days, and colonies were counted. Homogenized organs were stored in 10% glycerol at −80°C. The major organ fungal burden for each mouse was calculated by first multiplying organ weight by fungal burden per gram of tissue for each organ (brain, lungs, liver, and spleen) and then totaling these results.

Histopathology.

Organs in 10% formalin were submitted to the Department of Laboratory and Animal Medicine of the University of California, Los Angeles, where they were sectioned and stained with hematoxylin and eosin or mucicarmine. Slides were viewed using an Olympus BX41 microscope (Center Valley, PA), and representative images were taken using an Olympus digital camera.

TNF-α ELISA.

Serum was collected by cardiac puncture immediately after death and transferred to tubes with a gel barrier and clot activator (Terumo Medical Corporation, Somerset, NJ); samples were centrifuged, aliquoted, and stored at −80°C. Tumor necrosis factor alpha (TNF-α) levels were measured using a BD Biosciences OptEIA Mouse TNF (Mono/Poly) enzyme-linked immunosorbent assay (ELISA) set following the manufacturer's instructions.

Half-life of radiolabeled C. gattii.

Mice were given 0.1 mg/ml KI in their drinking water starting 1 week prior to injection with radiolabeled C. gattii. C. gattii was grown as above and heat killed at 60°C for 1 h. A total of 6 × 10⁷ cells were iodinated with 2 mCi of ¹²⁵I using IODO-Beads (Pierce, Rockford, IL). Immediately prior to injection, iodinated organisms were centrifuged at 800× g for 3 min and resuspended in 1.4 ml of PBS, and 200 μl was injected into the tail veins of mice. Assuming a 50% loss during the manipulations, each mouse was injected with approximately 4 × 10⁶ cells. Radioactivity was counted using a Ludlum 2200 Scaler Ratemeter and an NaI crystal with a well large enough to accommodate a mouse. Counts were plotted, and the line of best fit was calculated using the Microsoft Excel 2002 data analysis regression tool.

Murine infection with preopsonized Cryptococcus.

A total of 5.8 × 10⁵ C. gattii cells in PBS or 300 μl of serum from C57BL/6J, C3-deficient, or RAG-deficient mice were incubated for 30 min at 37°C and then pelleted by centrifugation (1,942 × g for 10 min at 4°C). Cryptococci were resuspended in 2 ml of PBS, and 200 μl was injected intravenously (i.v.) into lateral tail veins of C57BL/6J mice pretreated with CVF the day before, as described above. Mice were monitored for survival.

Preparation of GXM.

C. gattii was grown in Sabouraud's dextrose broth to obtain at least 1 × 10⁹ CFU. After centrifugation (1,942 × g for 10 min at 4°C) and two washes in 50 ml of cold PBS, the supernatant was filtered with a 0.45-μm-pore-size filter. The pH was adjusted to 4 to 5 using acetic acid. Sodium acetate was added at 10% (wt/vol). After sodium acetate dissolved, 2.5 volumes of 95% ethanol were added, and the culture was incubated at room temperature for 1 to 3 days until the solution was clear and a glaze covered the flask bottom. The supernatant was decanted, and then the pellet was air dried and resuspended in 2 to 3 ml of deionized water plus a 2- to 3-ml water rinse. To calculate the amount of C. gattii polysaccharide (CGPS), a series of dilutions of glucose in water were made (10, 20, 30, 40, 50, 60, 70, and 80 μg/ml). Fifty microliters of phenol was added to the glucose standard as well as to 10 μl of CGPS in 2 ml of water. Five milliliters of undiluted sulfuric acid was added, and the absorbance was read at 485 nm. The solution was adjusted to 0.2 M NaCl. A 0.3% solution of cetyltrimethyl ammonium bromide (CTAB; Sigma Aldrich, St. Louis, MO) in water was prepared, and 3× (wt/wt) CTAB was added to CTAB-CGPS with stirring to precipitate the GXM. The solution was centrifuged at 863 × g for 15 min at room temperature, and the supernatant was discarded. The pellet was washed with 20 ml of 10% ethanol in water, the supernatant was discarded, and the CTAB-GXM precipitate was dissolved in 5 ml of 1 M sodium chloride with rocking overnight at room temperature. Ethanol (95%) was added dropwise to CTAB-GXM with stirring to aid in GXM precipitation, and the solution was again centrifuged at 863 × g for 15 min at room temperature. Supernatant was discarded, and the CTAB-GXM precipitate was dissolved in 5 ml of 2 M sodium chloride and rocked overnight at room temperature. The solution was then dialyzed against 1 M sodium chloride using a 10,000-molecular-weight cutoff filter (Amicon Ultra centrifugal filter device; Millipore, Billerica, MA) until the solution was clear. The solution was then dialyzed against deionized water for 2 days and lyophilized, weighed, and resuspended in sterile water.

Serum antibody levels.

ELISA plates (Immulon 1B; Fisher Scientific, Pittsburgh, PA) were coated with C. gattii GXM at 10 μg/ml in PBS for 2 h at 37°C. Plates were blocked with 2% bovine serum albumin in PBS overnight at 4°C. Mouse serum was added in serial dilutions with blocking buffer, starting at 1:50, and incubated 1 h at 37°C. Plates were washed, and mouse antibodies were detected using goat anti-murine κ and goat anti-murine λ light chains (Southern Biotech, Birmingham, AL), both conjugated to alkaline phosphatase and added at 1:500. ELISA plates were developed with 1 mg/ml p-nitrophenyl phosphate (Sigma-Aldrich) in 1 M diethanolamine-0.25 mM MgCl, pH 9.8, and the absorbance was measured at 410 nm.

Killing assay.

Peritoneal macrophages were obtained by peritoneal lavage of C57BL/6J mice 5 days after intraperitoneal stimulation with 1.5 ml of 4% thioglycolate, as described previously (1). Mononuclear cells were plated at 4 × 10⁴/well on a 96-well tissue culture plate in 100 μl Iscove's modified Dulbecco's medium supplemented with 5% heat-inactivated fetal calf serum. Nonadherent cells were washed away with PBS after 2 h of incubation at 37°C. Adherent cells were stimulated overnight with 50 U/well of murine recombinant gamma interferon (R&D Systems, Minneapolis, MN). Mouse serum was added at a 30% final concentration, followed by the addition of C. gattii at 2,000 organisms/well (effector-target ratio of 5:1). Heat-inactivated mouse serum was prepared by incubation at 55°C for 30 min. After a 20-h incubation at 37°C, the supernatant was removed, and cells were lysed with sterile H₂O. The supernatant and lysate were combined, diluted, and plated on Sabouraud's dextrose agar in duplicate. Plates were incubated at 37°C for 2 days, and the number of CFU was quantitated. To determine whether serum alone had an effect on cryptococcal growth, conditions lacking macrophages were also examined.

Statistical analysis.

Survival of groups of mice was analyzed using a log rank test and StatView software (JMP, Cary, NC). Fungal burdens were analyzed using a Mann-Whitney U test and Prism, version 4.0 (GraphPad Software, Inc., San Diego, CA). A P value of ≤0.05 was considered significant.

RESULTS

C3 and, to a lesser extent, factor B play essential roles in protection against C. gattii infection.

C57BL/6J, C1q^−/−, C4^−/−, factor B^−/−, or C57BL/6J mice treated with CVF were infected i.v. with 1.3 × 10⁶ C. gattii cells. CVF-treated mice were all dead at day 2 postinfection, which was significantly faster (P = 0.0001) than untreated wild-type mice (Fig. 1A). Factor B^−/− mice died significantly faster than wild-type mice (P = 0.01) but lived significantly longer than CVF-treated mice (P = 0.0001), indicating that although the alternative pathway plays a role in protection against C. gattii, other aspects of complement activation are also important. For factor B^−/− mice, approximately half the mice died on days 5 to 6, and the remainder died 15 to 31 days postinfection. This biphasic pattern of death was consistently observed in three other experiments (Fig. 1B). Wild-type mice and mice deficient in C1q or C4 died at approximately the same rate, with deaths beginning at 11 days and all mice dead 60 days postinfection. Thus, complement activation contributes to survival, but the classical pathway does not play a role in protection.

FIG. 1.

C3-deficient, factor B-deficient mice, and CVF-treated mice died more rapidly than wild-type mice after infection. (A) Groups of eight mice (C57BL/6J, CVF-treated, C1q^−/−, C4^−/−, and factor B^−/−) were infected i.v. with 1.3 × 10⁶ C. gattii cells and monitored for survival. Compared to C57BL/6J mice, CVF-treated and factor B^−/− mice died earlier (P = 0.0001 and P = 0.01, respectively). (B) Groups of eight mice (C57BL/6J, C3^−/−, and factor B^−/−) were infected i.v. with 7.7 × 10⁵ C. gattii cells and monitored for survival. Compared to C57BL/6J mice, C3^−/− and factor B^−/− mice died early (P < 0.0001 and P = 0.0002, respectively).

In a separate experiment, C57BL/6J, factor B^−/− and C3^−/− mice were infected with 7.7 × 10⁵ C. gattii cells (Fig. 1B). Similar to mice treated with CVF, C3^−/− mice died within 3 days of infection (P < 0.0001). Therefore, the rapid death of CVF-treated mice was not the result of complement cascade hyperactivation elicited by CVF treatment but could be explained solely by the depletion of C3. All CVF-treated or C3-deficient mice were dead by the time the first deaths were seen in factor B^−/− mice.

C3- and factor B-deficient mice infected with C. gattii have higher fungal burdens and different fungal organ distributions than wild-type mice.

C3-deficient mice are completely deficient in complement activation, whereas factor B-deficient mice are deficient in only the alternative pathway. To determine how these differences in complement activation impacted organism localization, distribution, and clearance, C57BL/6J, CVF-treated, and factor B^−/− mice were infected with 6.3 × 10⁵ C. gattii cells and euthanized 21 h after infection. Fungal burdens in brains, lungs, livers, and spleens were determined from organ homogenates (Fig. 2A). Compared to wild-type mice, CVF-treated mice had increased numbers of organisms in the lungs but fewer in the liver (P = 0.03 for both organs); no significant differences were seen in the fungal burdens of spleens or brains. In contrast, factor B^−/− mice had significantly more organisms than wild-type mice in all organs examined (P = 0.03 for all organs). Factor B^−/− mice also had more organisms than CVF-treated mice in all organs except lungs (brain and liver, P = 0.03; spleen, P = 0.05). Although CVF-treated mice had more organisms in the lungs than factor B-deficient mice, this difference did not quite reach statistical significance (P = 0.06).

FIG. 2.

CVF-treated and factor B^−/− mice had higher fungal burdens than wild-type mice 21 h after infection. (A) Groups of four C57BL/6J, CVF-treated, and factor B^−/− mice were infected i.v. with 6.3 × 10⁵ C. gattii cells and euthanized 21 h after infection. Homogenized organs were plated on Sabouraud's dextrose agar plates. Colonies were counted, and fungal burden was calculated per gram for each organ. Error bars represent range. *, P < 0.05 compared to C57BL/6J mice using a Mann-Whitney test. (B) Fungal burdens per organ (brain, lungs, liver, and spleen) were added to give an estimate of the major organ fungal burden. Bars represent the median. *, P < 0.05 in comparison to C57BL/6J mice using a Mann-Whitney test. (C) Ratios of fungal burden per organ were calculated. Horizontal bars indicate the median. *, P < 0.05 in comparison to C57BL/6J mice; ❖, P < 0.05 in comparison to CVF-treated mice using the Mann-Whitney test.

The major organ fungal burden for each mouse was estimated from the organ weights and fungal burdens per gram of tissue in brain, lungs, liver, and spleen (Fig. 2B). CVF-treated mice had the largest major organ fungal burden with significantly more organisms than wild-type C57BL/6J mice (mean CFU/mouse, 24,000 versus 3,500; P = 0.03). Factor B-deficient mice also had significantly more organisms than wild-type mice (mean CFU/mouse, 8,200; P = 0.04). Although as a group factor B-deficient mice had fewer organisms than CVF-treated mice, there was one outlier in the CVF-treated group, so this difference did not quite reach statistical significance (P = 0.06).

Comparison of the ratios of total fungal burden in the lungs versus livers and lungs versus spleens showed that CVF-treated mice had preferential distribution to the lungs and decreased dissemination to the liver compared to wild-type mice (Fig. 2C). In factor B-deficient mice, organisms also preferentially localized to the lungs but not to the same extent seen in CVF-treated mice. In wild-type mice there were more total organisms in the liver than in the spleen although there were equivalent numbers of organisms per gram of tissue in these two organs.

Rectal temperatures of factor B-deficient mice (median, 34.3°C; range, 33.0 to 35.2°C) were significantly decreased compared to C57BL/6J mice (median, 36.9°C; range, 35.7 to 37.5°C) 21 h after infection with 6.3 × 10⁵ C. gattii cells (P = 0.03). Though CVF-treated mice also had lower rectal temperatures (median, 35.1°C; range, 34.5 to 36.8°C), the difference was not statistically significant from C57BL/6J mice (P = 0.1). A decrease in core body temperature is suggestive of septic shock, which can be mediated by certain cytokines, in particular, TNF-α. However, no TNF-α was detectable by ELISA in serum from C57BL/6J, factor B-deficient, or CVF-treated mice euthanized at 21 h postinfection (data not shown).

To explore further the influence of complement activation on disease progression and fungal burden in C. gattii infection, C3-deficient mice were examined at the time of death. C57BL/6J and C3^−/− mice were infected with 2 × 10⁶ C. gattii cells, but in this case, the C3^−/− mice were dissected at the time of death, and one C57BL/6J mouse was euthanized and dissected whenever a C3-deficient mouse died. All C3^−/− mice were dead by 49 h. As observed at 21 h, the major organ fungal burden was significantly elevated in C3^−/− mice (Fig. 3) (P = 0.03), with significantly more organisms in the lungs and kidneys of C3^−/− mice than in organs of C57BL/6J mice (P = 0.03 for both). In contrast, C57BL/6J mice had significantly more organisms in the liver (P = 0.03). There were no significant differences in the fungal burdens in the brains and spleens.

FIG. 3.

At the time of death, C3^−/− mice had higher fungal burdens than wild-type mice. Groups of four C57BL/6J and C3^−/− mice were infected with 2 × 10⁶ C. gattii cells i.v. When a C3^−/− mouse died, a C57BL/6J mouse was euthanized. Homogenized organs were plated on Sabouraud's dextrose agar. Fungal burden per gram was multiplied by the weight of each organ to determine the total fungal burden of each organ. Organ fungal burdens were totaled for each mouse. The horizontal bar represents the median. *, P < 0.05 using the Mann-Whitney test.

During the course of infection, rectal temperatures remained normal at 14 h postinfection. However, by 33 h (near the time of death), the CVF-treated mice had markedly decreased body temperatures (median, 26.3°C; range, 24.9 to 30.3°C) compared to C57BL/6J mice (median, 35.3°C; range, 34.5 to 35.9°C; P = 0.006). We could not determine whether hypothermia observed in these complement-deficient mice infected with C. gattii was the cause of death.

Lungs, heart, thymus, kidneys, liver, and spleen were also sectioned and examined microscopically. In the lungs, C57BL/6J mice had a range of rare individual to small clusters of mucicarmine-positive organisms, consistent with C. gattii, that expanded and filled the lumens of alveolar capillaries (Fig. 4A). In contrast, C3^−/− mice had large numbers of mucicarmine-positive organisms scattered throughout the alveoli that expanded and filled capillaries (Fig. 4B). The organisms tended to cluster multifocally throughout all lung fields. Both types of mice had rare individual organisms in the heart that expanded and filled the capillaries of the myocardium, and rare individual organisms were randomly located in the medulla of the thymus in C57BL/6J mice (data not shown). In C3^−/− but not C57BL/6J mice, random glomeruli and interstitial capillaries of the kidneys contained organisms that expanded and filled capillaries primarily at the glomerular hilus (Fig. 4C).

FIG. 4.

Organ histology showing that C3^−/− mice had more organisms in the lungs and kidneys than wild-type mice. Groups of four C57BL/6J and C3^−/− mice were infected with 2 × 10⁶ C. gattii cells i.v. When a C3^−/− mouse died, a C57BL/6J mouse was euthanized. Organs of two mice per group were sectioned and stained. Shown are mucicarmine-stained sections. Arrows indicate clusters of mucicarmine-positive cells. (A) C57BL/6J lung at a magnification of ×200. (B) C3^−/− lung at a magnification of ×100. (C) C3^−/− kidney at a magnification of ×400.

Clearance of C. gattii is reduced in factor B-deficient and CVF-treated mice.

To investigate the effect of complement activation on fungal clearance, we examined clearance of radiolabeled C. gattii in CVF-treated, factor B^−/−, and C57BL/6J mice. Approximately 4 × 10⁶ heat-killed and iodinated C. gattii organisms were injected into the tail veins of mice, and the clearance of radiolabeled organisms from animals was followed by whole-body counting of residual radioactivity over time (Fig. 5). By 60 h, less than 10% of the injected radioactivity remained in C57BL/6J mice, while 20% was present in factor B-deficient mice and 50% remained in CVF-treated mice. For C57BL/6J and factor B-deficient mice, there were two clearance phases. During the first 11 h, rapid clearance was seen, with cryptococci clearing more rapidly in the wild-type mice (half-life, 8 to 10 h) than in the factor B-deficient mice (half-life, 14 to 16 h). In the second phase, organisms cleared with a half-life of approximately 50 h. For CVF-treated mice, all organisms cleared with a half-life of approximately 50 h.

FIG. 5.

Complement-deficient mice had delayed clearance of radiolabeled heat-killed C. gattii in comparison to wild-type mice. Heat-killed C. gattii cells were labeled with radioactive iodine, and approximately 4 × 10⁶ organisms were injected i.v. into pairs of C57BL/6J, factor B^−/−, or CVF-treated C57BL/6J mice. Radioactivity was measured as counts per minute (CPM) using a mouse whole-body counter, and values were plotted as percent initial radioactivity. Error bars represent the range.

Fungal burdens vary in factor B-deficient mice depending on the time of death.

Factor B-deficient mice consistently showed a biphasic pattern of death following infection with C. gattii, with some mice dying on days 5 to 6 and the remainder dying on days 15 to 31. To determine whether fungal burden and localization differed in factor B-deficient mice that died early in infection compared to mice that died later in infection, factor B-deficient mice were infected with 7.6 × 10⁵ C. gattii cells, and organ fungal burdens for the first (day 6) and last (day 19) mice to die were determined. In the mouse that died on day 19, more organisms were seen in the brain and kidneys, as well as the lungs, and were too numerous to count (Fig. 6). In contrast, the mouse that died on day 6 had more organisms in the liver and spleen. However, the mouse that died at the later time had a 10-fold increase in the major organ fungal burden in comparison to the mouse that died earlier, largely as a consequence of increased fungal burden in the lungs (data not shown).

FIG. 6.

Factor B-deficient mice succumbing late to infection with C. gattii had larger fungal burdens in the brain, kidneys, and lungs than mice that succumbed early after infection. Five factor B-deficient mice were infected i.v. with 7.6 × 10⁵ C. gattii cells and dissected at the time of death. Fungal burdens are shown for the first (day 6) and last (day 19) mice to die. #, minimum estimate of fungal burden. The measured fungal burden was greater than 10⁸ CFU/g.

The lungs, heart, thymus, liver, spleen, kidneys, and brain from these two mice were also sectioned for histopathology. In the lungs of the mouse that died on day 6, large numbers of individual organisms and clusters of organisms were randomly scattered throughout terminal bronchioles and alveoli and were also found expanding and filling the lumina of alveolar capillaries (Fig. 7A). Some pulmonary veins contained large numbers of neutrophils and fewer mononuclear cells that extended into the perivascular interstitium. In the mouse that died on day 19, massive numbers of organisms were scattered within alveolar lumens and peribronchiolar connective tissue of all lung fields (Fig. 7B). The organisms were clustered and had well-developed mucinous capsules that completely filled and expanded alveoli. Alveolar macrophages filled with lipid (surfactant) were clustered in scattered alveoli, especially at the periphery of the lung (endogenous lipid pneumonia), and scattered pulmonary veins contained intraluminal organisms in close association with large numbers of neutrophils. Organisms were found in kidneys of both mice as clusters or individual organisms that expanded and filled the lumina of glomerular capillaries (Fig. 7C and D). In the mouse that died on day 19, moderate to large numbers of randomly scattered organisms expanded and filled the lumens of cortical capillaries. Glomeruli appeared to be spared in the mouse that died on day 19. Both mice had moderate numbers of individual and clustered organisms expanding and filling the hepatic portal vasculature of the liver (Fig. 7E and F). In the mouse that died on day 6, the organisms were sometimes surrounded by moderate numbers of individual macrophages and giant cell macrophages with foamy cytoplasm (lipid). In the mouse that died on day 19, there were moderate numbers of luminal neutrophils and cellular debris surrounding the organisms. Both mice also had low numbers of organisms in the lumen of the myocardial capillaries (data not shown). There were no organisms seen in the tissue sections from the thymus, brain, or spleen examined.

FIG. 7.

Organ histology shows that factor B-deficient mice that died later had more organisms in the lungs and kidneys but a similar number of organisms in the liver. Five factor B-deficient mice were infected i.v. with 7.6 × 10⁵ C. gattii cells and dissected at the time of death. Organs were sectioned and stained. Shown are mucicarmine-stained sections. (A) Early-death (day 6) lung at a magnification of ×100. (B) Late-death (day 19) lung at a magnification of ×100. The arrow points to an organism with a large capsule in an alveolus completely filled with yeast particles and proteinaceous material. (C) Early-death kidney at a magnification of ×200. (D) Late-death kidney at a magnification of ×200. (E) Early-death liver at a magnification of ×200. (F) Late-death liver at a magnification of ×400.

Opsonization with complement-sufficient serum does not protect against C. gattii infection in CVF-treated mice.

The different rates of clearance in the CVF-treated mice suggested that a failure to opsonize and clear C. gattii in the absence of C3 may make an important contribution to disease progression. To determine whether preopsonization with complement components was sufficient to prevent rapid death in CVF-treated mice, C57BL/6J mice were pretreated with CVF as described above and infected with 5.8 × 10⁵ C. gattii cells preopsonized with serum from C57BL/6J, C3^−/−, or RAG-deficient mice or incubated in PBS. C3 opsonization of C. gattii was detected by flow cytometry in a separate set of experiments (unpublished data). RAG-deficient mice, which do not produce antibodies, were used to evaluate the role of classical pathway activation from naturally occurring antibodies. CVF-treated mice infected with preopsonized C. gattii, regardless of serum source, died faster than mice infected with C. gattii alone (Fig. 8), and similar results were seen when C3-deficient mice were used (data not shown).

FIG. 8.

Preopsonization of C. gattii with complement-sufficient serum did not delay disease progression in mice depleted of complement. CVF-treated C57BL/6J mice were infected with 5.8 × 10⁵ C. gattii cells incubated in PBS or preopsonized with serum from C57BL/6J, C3^−/−, or RAG-deficient mice, and survival was monitored.

Complement plays a role in macrophage killing of C. gattii.

There was greater killing of C. gattii by mouse peritoneal macrophages in the presence of C57BL/6J serum (69.8%) than serum from C3^−/− (26.4%) or factor B^−/− (37.7%) mice (Fig. 9). There was reduced killing with heat-treated C57BL/6J serum compared to C57BL/6J serum (33.0%) (data not shown). The difference between C3^−/− and factor B^−/− results was not significant. Serum alone did not have an effect on cryptococcal growth. EDTA, which can be used to block complement activation, inhibited cryptococcal growth and therefore was not used in this assay.

FIG. 9.

Complement assists macrophage killing of C. gattii. C. gattii cells were incubated with peritoneal macrophages for 20 h with the indicated serum or no serum. Macrophages were then lysed, and the supernatant was collected and plated on Sabouraud's dextrose agar. The percent killing indicates the difference of C. gattii growth in the presence of mouse peritoneal macrophages and the indicated mouse serum versus no mouse serum and was calculated using the following formula: 100 × [1 − (number of CFU with serum/number of CFU without serum)]. Shown are the results of four separate experiments. Error bars indicate the standard deviations. The result for C57BL/6J mice was significantly different from that for C3^−/− or factor B^−/− mice (P < 0.05, Student's t test).

Serum levels of natural antibodies to C. gattii GXM were similar for wild-type and complement-deficient mice.

Sera from uninfected C57BL/6J, C3^−/−, and factor B^−/− mice were tested for the presence of antibodies to C. gattii GXM. C57BL/6J mice had a median inverse titer of 150 (range, 50 to 800), and factor B^−/− mice had a median inverse titer of 200 (range, 200 to 800) while C3^−/− had a median inverse titer of 400 (range, 100 to 800). However, the differences were not significant (data not shown).

DISCUSSION

Previous research has indicated that the alternative pathway is the only complement pathway activated by Cryptococcus spp. (5, 14) and that animals deficient in C3 but not C4 have faster disease progression (4). In this report, we have now explored the role of complement in infection with C. gattii, an emerging pathogen in North America. Moreover, this is the first study to examine cryptococcal infection in mice specifically deficient in components of the classical or alternative pathways of complement activation. The classical pathway does not function in C1q-deficient mice, and factor B^−/− mice are deficient in the alternative pathway. It was previously thought that C4-deficient mice lacked both the classical and lectin pathways, but given the finding that the lectin pathway can function in the absence of C4 (23), these mice are completely deficient only in the classical pathway. C3-deficient mice are completely deficient in the late steps of complement activation regardless of the pathway used. CVF forms a C3 convertase that rapidly cleaves C3, resulting in hyperactivation of the complement cascade and depletion of C3 to C9. Thus, mice treated with CVF are also completely deficient in the late steps of complement activation.

In our study using C. gattii, factor B-deficient mice died significantly faster than wild-type mice and had a significantly higher fungal burden, indicating that the alternative pathway does play an important role in resisting C. gattii infection. The C1q-deficient and C4-deficient mice did not die significantly faster than wild-type mice, indicating that the classical pathway was not essential for innate protection. However, C3 deficiency was more detrimental than factor B deficiency, as demonstrated by much more rapid death. There are two key differences between factor B- and C3-deficient mice: (i) the former can still opsonize with C3 fragments, and (ii) the lectin pathway remains functional in the absence of factor B although the amount of C3 deposited is expected to be decreased (19). Therefore, one possible explanation for the difference in rates of death observed between factor B- and C3-deficient mice is that the lectin pathway is also playing a role during cryptococcal infection. Previous studies have shown that human MBL does not bind to encapsulated C. neoformans and that MBL deficiency does not appear to predispose humans to cryptococcal disease (7), but studies in our laboratory have shown that murine MBL can bind to C. gattii (unpublished data).

A consistent finding was a biphasic death curve in factor B-deficient mice infected with C. gattii, with approximately half the mice dying by days 5 to 7 and the remainder of the mice dying around day 20. Adaptive immunity may be playing a role in the mice that survive longer as previous studies have demonstrated a role for complement in mounting a T-cell-mediated adaptive response (10, 18, 20). It is possible that some factor B^−/− mice that survive longer are able to mount an adaptive response to the infection, albeit not at a level sufficient to provide sustained protection.

Factor B^−/− mice that died later had greater fungal burdens, particularly in the lungs and surprisingly in the kidneys. Histopathology showed that the fungal burden in the lungs was so large as to likely prevent effective gas exchange and was the probable cause of death. Overwhelming pneumonia in the absence of complement activation is consistent with previous studies (21). In contrast, factor B-deficient mice that died earlier did not have sufficient fungal burden in the lungs to prevent efficient gas exchange, suggesting that pneumonia was not the direct cause of death. Similar to C3-deficient mice, the mice that died early did not exhibit any symptoms of illness but suddenly succumbed to infection. However, compared to survival of C3-deficient mice, death was delayed by 3 to 4 days. We were unable to determine the specific cause of death although factor B-deficient mice did exhibit hypothermia prior to death.

C3-deficient mice appeared to die from suffocation due to overwhelming fungal burden in the lungs. Although these mice were severely hypothermic, which was consistent with septic shock, TNF-α was not detectable in serum 21 h postinfection, suggesting that septic shock was not the cause of death. Since C3-deficient mice die more rapidly than factor B-deficient mice, there must be factors, such as lectin or unidentified pathways, in addition to the absence of alternative pathway activation that are playing a role in their rapid death.

Both factor B- and C3-deficient mice had significantly greater major organ fungal burdens than wild-type mice though this is probably an underestimate of the total fungal burden as only lungs, liver, brain, and spleen were examined. This may be at least partially explained by decreased killing of C. gattii by macrophages in the absence of complement. Factor B-deficient mice had more organisms in all organs examined. C3-deficient mice also had increased lung fungal burden but interestingly had a decreased number of organisms in the liver compared to wild-type mice. The localization of organisms to the liver did not correlate with the rate of disease progression as C3-deficient mice had fewer organisms in the liver than wild-type mice, while factor B-deficient mice had more; yet both died more rapidly than the wild type. The newly discovered CrIg receptor that is present on Kupffer cells in the liver could be a potential target of complement-opsonized C. gattii (9), and a failure to opsonize organisms with C3 cleavage products could explain the decrease in the number of organisms in the liver of C3-deficient mice. However, it is expected that there would also be a decrease in C3 deposition in factor B-deficient mice. Therefore, it is difficult to explain why there would be increased localization of organisms to the liver in these animals.

Organisms may localize to the lungs for several reasons. First, our model system uses tail vein injection, which introduces the organisms in such a way that the lung is the first major organ encountered. Second, at least one strain of C. neoformans is known to have a tropism for the lungs (26), so it is possible that C. gattii also has a lung tropism. Differences in lung fungal burdens in complement-sufficient and -deficient mice could be the result of targeting to the lungs or the absence of targeting elsewhere when complement is depleted. For example, a decrease in localization to the liver could result in more organisms trafficking to the lungs.

Although complement activation played an important role in clearance of organisms, preopsonization of C. gattii with C57BL/6J serum was not sufficient to delay disease progression in C3-deficient mice, suggesting that more than C3 opsonization is required. However, it should be noted that in vivo complement is continually available in wild-type mice, whereas in the preopsonization experiment it was available only prior to infection. Interestingly, all of the mice infected with preopsonized C. gattii died more rapidly than the mice injected with C. gattii incubated with PBS, including organisms preopsonized with serum from RAG^−/− mice. C3-deficient and RAG-deficient serum produced results similar to C57BL/6J serum, suggesting that a serum component besides C3 and naturally occurring antibodies may enhance cryptococcal infection, but we were unable to identify that component.

In the absence of complement activation, iodinated C. gattii had a half-life of approximately 50 h. In animals with an intact complement pathway, the majority of the organisms (approximately 90%) cleared rapidly, with a half-life of about 8 to 10 h. In factor B-deficient mice, there was also initial rapid clearance (half-life of approximately 15 h) of organisms although not as many organisms cleared rapidly as in wild-type animals. Interestingly, after the initial rapid clearance in wild-type and factor B-deficient animals, the remainder cleared at approximately the rate seen in C3-deficient mice, suggesting that these organisms represented a subpopulation that was not opsonized. Heterogeneity in capsule composition is known to occur in Cryptococcus (17) and may explain the second clearance phase.

In these studies, we have demonstrated a critical role for complement activation in delaying infection with C. gattii, an emerging pathogen in North America. Using mice genetically deficient in various complement components allowed us to dissect the contribution of the different complement activation pathways. While the alternative pathway plays a major role, we show for the first time that the alternative pathway is not the only pathway that contributes to protection in cryptococcal infection as factor B deficiency is not as detrimental as C3 deficiency. In a related study, we have shown that murine MBL is capable of activating in vitro C3 deposition on C. gattii (unpublished data). Thus, it is possible that the lectin pathway or as yet undescribed complement activation pathways may contribute to protection against infection.