Protective Efficacy of Antigenic Fractions in Mouse Models of Cryptococcosis

Abstract

Infections due to the encapsulated fungus Cryptococcus neoformans are a significant cause of morbidity and mortality in patients with impaired T-cell function, particularly those with AIDS. Presumably then, T-cell responses to cryptococcal antigens are critical for protection against this ubiquitous fungus. To test the protective efficacy of these antigens as vaccine candidates, secreted cryptococcal antigens were separated by concanavalin A affinity chromatography into adherent (mannoprotein [MP]) and nonadherent (flowthrough [FT]) fractions, and the fractions were tested in murine models of disseminated cryptococcosis. Compared with adjuvant alone, C57BL/6 mice that received two inoculations of MP and FT exhibited prolonged survival and reduced brain and kidney fungal loads following intravenous challenge with C. neoformans strain B3501. MP-immunized animals had increased brain levels of tumor necrosis factor alpha, gamma interferon, and interleukin-2. Histopathologic examination revealed that compared with organs from mice that received only adjuvant, MP-immunized mice were able to recruit a stronger cellular infiltrate in brain, kidney, and liver in response to cryptococcal infection. Conjugated O-linked glycans were necessary for optimal MP-mediated protection, because chemical O deglycosylation reduced the protective efficacy of MP immunization. FT and MP immunization protected B-cell-deficient, but not T-cell-deficient mice, suggesting that protection was T-cell mediated. CBA/J mice also benefited from immunization with FT and MP, although the benefits were more modest than those seen with C57BL/6 mice. Thus, both MP and FT fractions of C. neoformans contain components that protect mice from disseminated cryptococcosis, and this protection appears to be T-cell mediated.

Persons with impaired CD4⁺ T-cell function, particularly those with AIDS and those receiving immunosuppression for solid organ transplants, are at high risk of developing clinically apparent infections due to the encapsulated fungus Cryptococcus neoformans (21, 39). Indeed, cryptococcosis has emerged as one of the most common causes of death worldwide in individuals afflicted with AIDS (10). Moreover, the recent epidemic of cryptococcosis on Vancouver Island, Canada (40), underscores the potential for this fungus to continue to emerge in unexpected geographic and clinical settings.

Cryptococcal capsular polysaccharide is a high-molecular-weight polysaccharide, of which glucuronoxylomannan is the major component. There is unequivocal evidence proving that capsule is a major virulence factor on C. neoformans with both shed and in situ glucuronoxylomannan contributing to virulence (3, 5). While capsule subverts virtually all aspects of host defenses, innate phagocytic (neutrophil and macrophage) and humoral (antibody and complement) defenses are particularly hard hit. The result of the relative ineffectiveness of phagocytic and humoral anticryptococcal defense mechanisms is that the host must rely heavily upon acquired T-cell defenses.

The requirement for T cells to effectively defend against cryptococcosis has led investigators to search for immunoreactive cryptococcal antigens that could serve as vaccine candidates. Murphy and colleagues isolated a crude culture supernatant, designated C. neoformans culture filtrate antigen (CneF), which stimulated delayed-type hypersensitivity (DTH) responses and cytokine production in immunized mice (30). Subcutaneous immunization of CBA/J mice with CneF in complete Freund's adjuvant resulted in protection against a challenge infection with C. neoformans (32, 33). Protection was associated with an increase in activated CD4⁺ T cells and macrophages, as well as production of gamma interferon (IFN-γ) and tumor necrosis factor alpha (TNF-α). In contrast, immunization with heat-killed C. neoformans in complete Freund's adjuvant did not confer protection against a challenge with viable fungi (32, 33).

In an effort to define the components of the CneF responsible for the T-cell responses, CneF has been separated on concanavalin A (ConA) affinity columns into adherent (mannoprotein [MP]) and nonadherent (flowthrough [FT]) fractions based upon the ability of the lectin ConA to bind terminal mannose and glucose groups. The MP fraction was found to be predominantly responsible for the DTH responses (31). It has also been shown that MP stimulates lymphoproliferative responses and cytokine production from patients recovered from cryptococcosis (12, 23). Moreover, preparations of MP induce TNF-α and IL-12 production by human monocytes and murine macrophages (4, 34, 36). These two cytokines are critical to host defenses in murine models of cryptococcosis (7, 15).

Cryptococcal MPs are heterogeneous, although at least some share structural features, including signal sequences, Ser/Thr-rich C-terminal regions (which likely serve as sites of extensive O glycosylation), and glycosylphosphatidylinositol anchor motifs (13, 22). Four cryptococcal proteins, including two MPs, which stimulate T-cell responses, have been purified, sequenced, and cloned (1, 13, 22, 26). The aim of the present study was to test the protective efficacy of MP and FT fractions in murine models of cryptococcosis. We found that both the MP and FT fractions afforded partial protection via a mechanism that appeared to be dependent upon T cells, but not B cells.

MATERIALS AND METHODS

Materials.

All chemical reagents were obtained from Sigma Chemical Company (St. Louis, Mo.), and all plasticware was obtained from Fisher Scientific (Pittsburgh, Pa.), unless otherwise specified.

Mice.

Specific-pathogen-free mice were purchased from The Jackson Laboratory, (Bar Harbor, Maine) and housed in microisolator cages at The Boston University Medical Center Laboratory Animal Sciences Center. The mice were adapted to their environment for at least 3 days prior to experimentation. To prevent unnecessary pain and suffering, infected mice were sacrificed when moribund, using objective criteria (periorbital edema, posturing, ataxia, and inability to feed) approved by The Boston University Medical Center Institutional Animal Care and Use Committee.

The B-cell-deficient (MuMT B6.129S2-Igh-6^tm1Cgn) mice carry a stop codon in the 5′ end of the first transmembrane exon of the μ chain. This prevents μ_m expression on pro-B cells, and the cells die by apoptosis (17, 25). Thus, MuMT mice lack mature B cells in the periphery and have near absent levels of circulating serum immunoglobulins. The T-cell-deficient (B6.129P2-Tcrb^tm1Mom Tcrd^tm1Mom) mice have deletions of the β and δ loci (18, 19). α/β and γ/δ T-cell receptors do not form, thus rendering these mice deficient in CD4⁺ T cells, CD8⁺ T cells, γ/δ T cells, and NK T cells.

Isolation of soluble C. neoformans CR, MP, and FT supernatant fractions.

Soluble fractions were isolated from culture supernatants of C. neoformans acapsular strain Cap 67 (ATCC 52817) as in previous studies (27). Briefly, supernatants from yeast cultures were filter sterilized and then concentrated with a tangential filtration system equipped with a 10-kDa cutoff regenerated cellulose membrane. The concentrated material was dialyzed against phosphate-buffered saline (PBS) to form the crude (CR) fraction. The CR fraction was then subjected to ConA affinity chromatography. The nonbinding FT fraction was collected. The ConA-binding MP fraction was eluted with a 0.2 M methyl-α-d-mannopyranoside dissolved in PBS. Fractions were dialyzed against distilled water by using 7-kDa-cutoff dialysis tubing, lyophilized, and stored at −80°C until use. The protein concentration was assessed with the bicinchoninic acid assay (Pierce Scientific, Rockford, Ill.), while total carbohydrate was measured with the phenol-sulfuric acid assay (9). As calculated by mass, the MP and FT fractions had carbohydrate/protein ratios of 5.4:1 and 6.5:1, respectively. β-Elimination of MP was performed as in previous studies (27) by treating the MP with 0.1 M NaOH at 37°C for 24 h, followed by neutralization with acetic acid. The MP fraction, but not the FT fraction, contained trace amounts of ConA, as determined by Western blotting (data not shown). Where indicated, the MP fraction was boiled for 10 min to destroy the biological activity of ConA.

Immunization of mice and mouse model of cryptococcosis.

Antigen was admixed with Ribi adjuvant system according to the manufacturer's directions. Ribi adjuvant system is a formulation of squalene, Tween 80, monophosphoryl lipid A (from Salmonella enterica serovar Minnesota), trehalose dicorynomycolate (an analog of cord factor of the tubercle bacillus), and cell wall skeleton (deproteinized and delipidated cell wall from Mycobacterium). Mice (6 to 8 weeks old) received 50 μg of antigen (based upon protein content) intraperitoneally (i.p.) in 0.2 ml of Ribi adjuvant system. Three weeks later, a second i.p. injection of 50 μg of antigen was delivered. Mice were challenged with live C. neoformans 1 week following this second injection.

The encapsulated serotype D strain B3501 (ATCC 34873), which is the isogenic parent strain of Cap 67, was utilized for the in vivo challenge studies. B3501 has been extensively characterized and is virulent in mouse models of cryptococcosis (37). Fresh cultures of strain B3501 were started from frozen −80°C stock for each animal experiment. Yeast cells were harvested from 2-day-old cultures grown on Sabouraud dextrose agar at 30°C, washed, and suspended at 10⁷ per ml in ice-cold PBS. Intravenous (i.v.) injections were performed by cannulating the lateral tail vein and then administering 100 μl of the yeast suspension. The inoculum size was confirmed by assessing CFU following dilutions and spread plates.

Assessment of organ CFU.

Mice were euthanized by CO₂ asphyxiation, and the livers, lungs, brains, spleens, and kidneys were harvested, weighed, and placed in 12-ml sterile polypropylene tubes containing 1 ml of cold PBS supplemented with 40 U of penicillin per ml and 40 μg of streptomycin per ml. The tissue was homogenized at a setting of 1 on a PowerGen Model 700 tissue homogenizer (Fisher Scientific) for 5 to 10 s. The homogenates were diluted and spread on Sabouraud dextrose agar to determine the number of CFU per gram of tissue.

Assessment of tissue cytokine concentrations.

Organs were obtained as described above and homogenized in 1 ml of ice-cold PBS containing 0.5 μM phenylmethylsulfonyl fluoride, 0.5 μM dithiothreitol, and 8 μg of leupeptin per ml. Homogenates were clarified from tissue debris by centrifuging the organ samples at 19,000 × g for 5 min. Supernatants were collected and stored at −80°C. Analysis for TNF-α, IFN-γ, IL-4, IL-5, and IL-2 content was performed by flow cytometry with the mouse Th1/Th2 cytokine bead array (BD PharMingen, San Diego, Calif.) according to the manufacturer's directions.

Tissue histopathologic examination.

Following sacrifice, organs were removed and fixed in 10% buffered formalin. Tissue was embedded into paraffin wax in an automated tissue processor at the pathology facilities of the Skin Pathology Laboratories (Boston, Mass.). Sections (5 to 6 μm thick) then were cut on a microtome and stained with hematoxylin and eosin by using an automated slide processor. For each organ examined, three to five sections were performed, and five fields from each of those sections were read in a blinded fashion. The sections were scored for the degree of fungal infiltration and inflammation severity by using predefined criteria.

Statistical analysis.

Kaplan-Meier survival curves were compared by using the log rank test (NCSS Statistical Software, Kaysville, Utah). All other statistical comparisons utilized the Student's t test. Values of P < 0.05 were considered statistically significant.

RESULTS

Effect of immunization with cryptococcal fractions on survival following challenge with C. neoformans.

C57BL/6J mice were immunized twice, 3 weeks apart, with CR, MP, and FT fractions of C. neoformans admixed with the Ribi adjuvant system. Control mice received adjuvant alone or PBS. Mice were then i.v. challenged with 10⁶ C. neoformans cells and monitored daily for survival. Each of the three cryptococcal fractions afforded significant protection compared with the groups that received adjuvant alone (Fig. 1 and Table 1). In addition, mice that received adjuvant alone had significantly prolonged survival compared with mice that received PBS.

FIG. 1.

Effect of immunization with cryptococcal antigenic fractions on survival of C57BL/6J mice challenged with C. neoformans. C57BL/6J mice received two i.p. injections 3 weeks apart of PBS, Ribi adjuvant system without antigen (ADJ), or Ribi adjuvant system admixed with 100 μg of crude MP or FT antigen. One week following the last injection, mice received an i.v. challenge of 10⁶ live C. neoformans cells. Mice were monitored for signs of disease and sacrificed when signs of disease were severe. P < 0.001 for comparison of survival of the MP, CR, and FT groups versus the ADJ control. P = 0.005 for comparison of the ADJ and PBS groups. The figure is representative of three to four independent experiments (summarized in Table 1), each of which had similar results.

TABLE 1.

Effect of immunization on survival of mice challenged with C. neoformans^a

Exp.	Median no. of survival days
	PBS	Adjuvant	CR	MP	FT
1	7	10	26	18	22
2	10	14	26	24	NDb
3	11	14	23	24	64
4	ND	7	ND	18	21
Combined	9	11	25	21	36

^aGroups of 8 to 10 C57BL/6J mice received two i.p. injections 3 weeks apart of PBS, Ribi adjuvant system without antigen (Adjuvant) or Ribi adjuvant system admixed with 100 μg of CR, MP, or FT antigen. One week following the last injection, mice received an i.v. challenge of 10⁶ live C. neoformans cells. Mice were monitored for signs of disease and sacrificed when signs of disease were severe.

^bND, not determined.

The MP used in the experiment described above was native protein obtained following ConA affinity chromatography. A potential pitfall to this approach is that bioactive ConA can leach off the column. To determine if bioactive ConA found as part of the MP fraction had a nonspecific effect on the immune response, a boiled (and therefore ConA inactivated) antigen preparation was compared to native MP with 50 μg per immunization (Fig. 2). No difference in protection was observed, suggesting that the observed effects were due to the MP. Moreover, administration of a 10-fold-smaller amount of MP (5 μg) resulted in similar levels of protection.

FIG. 2.

Influence of dose and boiling on MP-dependent protection. Groups of 10 C57BL/6J mice received two i.p. injections 3 weeks apart of PBS, Ribi adjuvant system without antigen (ADJ), or Ribi adjuvant system admixed with 50 μg of native MP (MP 50), 5 μg of native MP (MP 5), or 50 μg of boiled MP (MP-B). One week following the last injection, mice received an i.v. challenge of 10⁶ live C. neoformans cells. Mice were monitored for signs of disease and sacrificed when signs of disease were severe. P < 0.05 for comparison of the ADJ with the MP 50, MP 5, and MP-B groups.

Effect of immunization with cryptococcal fractions on organ fungal load.

Having determined that cryptococcal fractions prime a partially protective immune response, we sought to determine if the immune response generated resulted in a reduced organ fungal load. Mice were immunized with CR, MP, and FT; challenged with live C. neoformans cells; and then sacrificed at various time points postinfection for enumeration of organ CFU. At the 1-h time point following infection, all vaccinated groups had similar fungal organ loads, ensuring even inoculation (data not shown). While nonsignificant decreases in fungal organ loads were measured at day 7 postinfection (data not shown), at day 10, CR-, FT-, and MP-immunized mice showed a significant 1-log decrease in cryptococcal CFU formation in both brain and kidney (Fig. 3). Significant reductions in organ fungal burdens were not seen in the liver, lungs, or spleen (data not shown).

FIG. 3.

Effect of immunization with cryptococcal antigenic fractions on organ fungal burden. C57BL/6J mice received two i.p. injections 3 weeks apart of PBS, Ribi adjuvant system without antigen (ADJ), or the Ribi adjuvant system admixed with 100 μg of crude or MP antigen. One week following the last injection, mice received an i.v. challenge of 10⁶ live C. neoformans cells. Ten days postinfection, mice were sacrificed, and the number of CFU in the brain and kidneys was determined. Data are expressed as CFU per gram of tissue. P < 0.05 for comparison of adjuvant with CR or MP for brain and kidneys.

Histopathological analysis of tissue from C. neoformans-infected mice.

Mice were immunized with MP or adjuvant alone, challenged with C. neoformans, and then sacrificed 4 and 10 days postinfection. Specimens of brain, liver, and kidney were subjected to histopathologic analysis. All animals had fungal colonies present in the cerebrum, cerebellum, brainstem, and occasionally the meninges. Interestingly, brain sections from mice immunized with MP had numerous fungal colonies (average of 41 per section), but some of the colonies were surrounded by and/or infiltrated with mononuclear cells (Fig. 4A). Although fungal colonies were present in the brains of control mice immunized with adjuvant (average of 15 per section; Fig. 4B), only rarely were inflammatory infiltrates observed. Mice immunized with MP had multifocal, small hepatic pyogranulomas (Fig. 4C), which frequently contained aggregates of eosinophils and, rarely, one or two fungi. Similar to mice immunized with MP, control mice had numerous hepatic pyogranulomas, many of which were quite large (Fig. 4D), containing focal hepatic necrosis. In addition, control mice had numerous, large fungal colonies randomly scattered throughout the liver, sometimes associated with the inflammatory infiltrates. The kidneys of mice immunized with MP contained focal to focally extensive areas of interstitial nephritis with tubular degeneration and regeneration (Fig. 4E). Small fungal colonies (one to three organisms) were present in the renal interstitium and were often associated with the foci of inflammation. In contrast, kidney specimens from control animals contained an increased number of fungal colonies with more organisms per colony, but only occasionally were the colonies associated with the interstitial inflammation (Fig. 4F).

FIG. 4.

Histopathological examination of tissues from immunized mice. C57BL/6J mice received two i.p. injections 3 weeks apart of Ribi adjuvant system without antigen or Ribi adjuvant system admixed with 100 μg of MP antigen. One week following the last injection, mice received an i.v. challenge of 10⁶ live C. neoformans cells. Ten days postinfection, mice were sacrificed, and their brains, livers, and kidneys were analyzed. (A) Brain specimens from mice immunized with MP plus adjuvant contained individual fungal organisms (arrow) or colonies, some of which were infiltrated with mononuclear cells (arrowhead). (B) Brain specimens from control mice immunized with adjuvant also contained fungal colonies, some of which were rather large, but no inflammation. (C) Liver specimens from mice immunized with MP plus adjuvant contained multifocal pyogranulomas (arrowheads), some of which contained prominent aggregates of eosinophils. The few organisms observed in these sections were associated with inflammatory foci. (D) Control mouse liver specimens contained numerous fungal organisms, which were randomly distributed throughout the liver, present individually or in colonies (arrows). Compared to mice immunized with MP plus adjuvant, control mouse liver specimens also contained an increased number of pyogranulomas, only some of which were associated with the fungal colonies. (E) Examination of kidney specimens from mice immunized with MP plus adjuvant revealed mild, multifocal interstitial nephritis, dominated by mononuclear cells and fewer neutrophils (arrowhead). Occasionally, these foci contained small numbers of fungal organisms (arrow). (F) In contrast, control mouse kidney specimens contained numerous fungal organisms, which were present individually or in small colonies (arrows). These kidney specimens contained fewer foci of interstitial nephritis, and only occasionally were the inflammatory foci associated with organisms. The results are representative of two to three mice per group.

Dependence on conjugated O-linked glycans for MP-mediated protection.

Our previous in vitro studies demonstrated that optimal T-cell responses to MP require the presence of O-linked mannose residues on MP (27). These O-glycans bind to mannose receptors on antigen-presenting cells, thus allowing for efficient entry, degradation, and presentation to T cells. To examine the role of O-linked carbohydrates in vivo, MP was chemically deglycosylated by β-elimination and used to vaccinate C57BL/6J mice (100 μg per mouse). Positive controls received native MP, while negative controls received PBS or adjuvant alone. Mice were then challenged with live C. neoformans, and survival was monitored (Fig. 5A). Mice immunized with deglycosylated MP died sooner than the mice immunized with native MP (median survival times of 13 and 18 days, respectively), although the differences between the groups did not reach statistical significance. However, the group that received the deglycosylated MP did fare better than the groups receiving PBS or adjuvant alone. Similar trends were seen when examining brain and kidney CFU at 10 days after infection, with mice immunized with β-eliminated MP having higher organ CFU than mice receiving native MP (Fig. 5B). Thus, in vivo, the presence of O-linked glycans on MP appears to augment host defenses against C. neoformans.

FIG. 5.

Effect of deglycosylation of MP on survival and organ fungal burden following cryptococcal challenge. C57BL/6J mice received two i.p. injections 3 weeks apart of PBS, Ribi adjuvant system without antigen (ADJ), or Ribi adjuvant system admixed with 50 μg of MP or β-eliminated MP (MP-BE). One week following the last injection, mice received an i.v. challenge of 10⁶ live C. neoformans cells. (A) Mice (10 per group) were monitored for signs of disease and sacrificed when signs of disease were severe. P < 0.001 for comparison of survival of the ADJ group with that of the PBS, MP, and BE-MP groups. (B) Ten days postinfection, mice (three to four per group) were sacrificed, and the numbers of CFU per gram of tissue in the brain and kidneys were determined.

Contribution of T and B cells to cryptococcal fraction-dependent protection.

With the observation that cryptococcal fractions extend survival and decrease fungal organ loads in an i.v. model of murine cryptococcosis, we next assessed which branch of the adaptive immune responses is most critical in the contribution to protection. To investigate this, B-cell-deficient (MuMT B6.129S2-Igh-6^tm1Cgn) and T-cell-deficient (B6.129P2-Tcrb^tm1Mom Tcrd^tm1Mom) β/δ^−/− mice were immunized with adjuvant, FT, or MP fractions. The wild-type, B-cell-deficient, and T-cell-deficient mice were infected i.v. with live C. neoformans cells, and survival was measured. For the wild-type and B-cell-deficient mice, the groups immunized with MP and FT fractions had significant prolongation of survival compared with mice immunized with adjuvant alone (Fig. 6). In contrast, immunization failed to afford significant protection for the T-cell-deficient mice. Thus, the partial protection elicited by MP and FT appears to be dependent on the T-cell component of the adaptive immune system.

FIG. 6.

Contribution of T and B cells to cryptococcal fraction-dependent protection. C57BL/6J wild-type, B-cell-deficient (knockout [KO]) (MuMT B6.129S2-Igh-6^tm1Cgn), or T cell-deficient (B6.129P2-Tcrb^tm1Mom Tcrd^tm1Mom) mice received two i.p. injections 3 weeks apart of Ribi adjuvant system without antigen (ADJ) or Ribi adjuvant system admixed with 100 μg of MP or FT antigen. One week following the last injection, mice received an i.v. challenge of 10⁶ live C. neoformans cells. Mice were monitored for signs of disease and sacrificed when signs of disease were severe. P < 0.001 for comparison of survival of the FT group versus ADJ for the wild-type and B-cell-deficient mice. P = 0.015 and P < 0.001 for comparison of survival of the MP group versus ADJ for the wild-type and B-cell-deficient mice, respectively.

Effect of MP immunization on tissue cytokine levels.

Having identified T cells as critical to the protective mechanism elicited by the MP fraction, we next sought to determine whether protection correlated with in vivo tissue cytokine levels. C57BL/6 mice received either adjuvant alone or MP immunization and then were i.v. challenged with C. neoformans. Mice were sacrificed 10 days postchallenge, and TNF-α, IFN-γ, IL-5, IL-4, and IL-2 levels in the brain, liver, spleen, lungs, and kidneys were determined. Uninfected, unimmunized mice served as additional controls. Compared to mice treated with adjuvant alone, MP-immunized mice had increased levels of TNF-α, IFN-γ, and IL-2 in the brain (Fig. 7). The kidneys and livers of MP-immunized mice also had increased levels of TNF-α, and the kidneys also had decreased levels of IL-5. Cytokine levels in other organs did not vary significantly. Surprisingly, few differences were noted when comparing cytokine levels in organs of uninfected mice with those in infected mice treated with adjuvant alone.

FIG. 7.

Effect of MP immunization on tissue cytokine levels following infection. C57BL/6J mice received two i.p. injections 3 weeks apart of Ribi adjuvant system without antigen (Adjuvant-infected) orRibi adjuvant system admixed with 100 μg of MP antigen (MP-infected). One week following the last injection, mice received an i.v. challenge of 10⁶ live C. neoformans cells. Ten days postinfection, mice were sacrificed, and cytokine levels in homogenized organs were assayed. Organ cytokine levels in uninfected, unimmunized mice (Non-infected) were also determined to establish baseline levels. Data are expressed as picograms of cytokine per gram of tissue. n = 3 mice per group. ND, not detectable. Asterisks denote P < 0.05 for comparison of adjuvant-infected with MP-infected groups.

Effect of murine strain on cryptococcal fraction-dependent protection.

C57BL/6 mice have been shown to be relatively susceptible to C. neoformans infection compared to BALB/C (moderately resistant) and CBA/J (highly resistant) mice when studied in an intratracheal model (14). Many of the studies examining immune responses to C. neoformans and cryptococcal components have utilized CBA/J mice (32, 33). Therefore, in the final set of experiments, we tested our cryptococcal fractions in CBA/J mice by using the identical immunization and inoculation protocol utilized for the experiments with the C57BL/6 mice. Mice immunized with CR or MP fractions survived longer than controls given adjuvant alone (Fig. 8). However, the increased survival was more modest than that seen in the experiments performed with C57BL/6 mice described above.

FIG. 8.

Effect of immunization with cryptococcal antigenic fractions on survival of CBA/J mice challenged with C. neoformans. The conditions are identical to those described in the legend to Fig. 1, except CBA/J mice were utilized. P < 0.001 for comparison of survival of the PBS and CR groups versus the ADJ control. P = 0.03 for comparison of the ADJ and MP groups.

DISCUSSION

The data reported herein define some of the factors responsible for partially protective responses to secreted antigens from an acapsular C. neoformans strain. We found the MP and FT fractions offer partial T-cell-mediated protection when used to vaccinate mice against experimental cryptococcosis. While previous studies have demonstrated that both of these fractions contain T-cell stimulatory antigens (1, 13, 22, 26), the MP fraction has been shown to be predominantly responsible for DTH responses in mice and lymphoproliferative responses in humans (23, 31). An advantage of using an antigenic mixture is it allows the host to develop an immune response against multiple antigens. However, a pitfall of this approach is that not all antigens are likely to be protective, and some antigens may even be deleterious. In this regard, while most immunological responses to MP are proinflammatory, a recent report demonstrates that an MP fraction, designated MP-4, can desensitize neutrophils to chemotactic challenge (6). Ultimately, individual antigens will need to be tested to determine which have protective efficacy.

The exact mechanisms by which immunization with MP and FT affords partial protection remains to be determined. The studies with T- and B-cell-deficient mice demonstrated that T cells, but not B cells, were necessary for the protective effect. In addition to stimulating an antigen-specific T-cell response, we cannot eliminate the possibility that at least some of the protective effect was due to antigen-nonspecific stimulation of the immune system. In support of this concept, Pietrella et al. demonstrated that C. neoformans MP was able to induce a Candida albicans-directed protective Th1 response (35).

Paradoxically, the B-cell-deficient mice had increased survival compared to wild-type mice. While the mechanisms remain speculative, in studies examining the protective efficacy of anticryptococcal monoclonal antibodies, it has been demonstrated that individual antibodies can be protective, deleterious, or neutral (2). Thus, it is tempting to hypothesize that the antibody response in our model was deleterious. However, it has been observed that MuMT mice have reduced production of IL-4, IL-10, and transforming growth factor beta (TGF-β) following oral administration of antigen (11). This raises the possibility that the increased survival of the B-cell-deficient mice challenged with C. neoformans is due to a more vigorous Th1 response. Similar to our data, using models of invasive pulmonary aspergillosis and primary candidiasis, Montagnoli et al. found that B-cell-deficient MuMT mice had prolonged survival compared with wild-type mice and that this was associated with the induction of antifungal Th1 immune responses (28). Moreover, the number of dendritic cells producing IL-12 was higher in the MuMT mice, while the numbers of IL-4- and IL-10-producing dendritic cells were higher in the wild-type mice (28). Finally, it should be noted that while MuMT mice have a pro-B-cell developmental block, immunoglobulin A is selectively expressed, and thus they are not totally antibody deficient (25).

Susceptibility to cryptococcosis varies greatly among mouse strains. In pulmonary models of cryptococcosis, C57BL/6 mice are relatively susceptible, whereas BALB/C and CBA/J mice are resistant. Susceptibility correlates with development of pulmonary eosinophilia and a Th2-type response (14, 41, 42). However, in our studies, we found that CBA/J mice were more susceptible to i.v. challenge with C. neoformans than were C57BL/6 mice. Both mouse strains derived partial protection from immunization with cryptococcal fractions, although the protection was more dramatic in the C57BL/6 mice. In agreement with our data, in an i.v. model, Decken et al. found C57BL/6 mice to be more resistant to cryptococcosis than BALB/C mice (7). Taken together, these results suggest that the immune responses to C. neoformans differ in the pulmonary and systemic compartments. Pulmonary models of infection mimic the likely natural route of exposure to C. neoformans, whereas the i.v. route of infection mimics the clinical situation when the fungus hematogenously disseminates (21). Future studies will examine the protective efficacy of the cryptococcal fractions in pulmonary models of infection.

Our studies utilized serotype D C. neoformans strain B3501. This strain was chosen because it is the parent strain of Cap 67 (used to generate the immunogens), has been nearly completely sequenced, and is virulent in mouse models of cryptococcosis (37). One must be careful though not to generalize data obtained with this strain. Differences in virulence have been noted within and between serotypes of clinical strains of C. neoformans, even using the same animal model (3, 16). Clearly, ideal vaccine candidates should protect against the vast majority of isolates likely to be encountered clinically.

The antigen preparations used in our studies were administered with the Ribi adjuvant system. Few studies have systematically compared the elicited immune responses induced by different adjuvants (20, 24). Responses skewed towards either Th1 or Th2 have been noted with the Ribi adjuvant system (29, 38). In vivo cytokine analysis of brains from C. neoformans-infected mice that were immunized with MP in Ribi adjuvant system revealed a Th1 bias. Although the effects were modest, significantly increased levels of IFN-γ, IL-2, and TNF-α were found in the brains of the animals immunized with MP in adjuvant compared with mice that just received adjuvant alone. Undetectable levels of the Th2 cytokines IL-4 and IL-5 were found in the brains of infected mice. Consistent with these results, examination of brain histopathology revealed an inflammatory response only in the mice immunized with MP. Interestingly, the adjuvant utilized in our studies, Ribi adjuvant system, afforded the mice a modest degree of protection against cryptococcosis even in the absence of specific antigens. In other cryptococcal model systems, salutary nonspecific effects also have been noted when using complete Freund's adjuvant (32, 33).

The development of an effective vaccine against C. neoformans will be a formidable challenge. Most humans who develop cryptococcosis have qualitative or quantitative disorders of CD4⁺ T-cell function. A vaccine that protects by stimulating antigen-specific CD4⁺ T-cell responses might have reduced efficacy in situations in which T-cell function is compromised. However, having some memory T-cell response, even if diminished, might be enough to afford some degree of protection. Moreover, antigen-specific CD8⁺ T-cell-mediated immunity may be able to compensate in situations in which CD4⁺ T cells are depleted, as was recently demonstrated in murine models of blastomycosis and histoplasmosis (43). Alternative approaches to cryptococcal vaccine development have focused on eliciting an antibody response to capsular polysaccharide (8). Ultimately, to elicit maximal protection, a vaccine that elicits both T-cell and antibody responses may be required. The data we present here provide experimental support that both MP and FT fractions elicit partially protective T-cell responses and thus could serve as components of a C. neoformans vaccine.