Antibody Response to Cryptococcus neoformans Proteins in Rodents and Humans

Abstract

The prevalence and specificity of serum antibodies to Cryptococcus neoformans proteins was studied in mice and rats with experimental infection, in individuals with or without a history of potential laboratory exposure to C. neoformans, human immunodeficiency virus (HIV)-positive individuals who developed cryptococcosis, in matched samples from HIV-positive individuals who did not develop cryptococcosis, and in HIV-negative individuals. Rodents had little or no serum antibody reactive with C. neoformans proteins prior to infection. The intensity and specificity of the rodent antibody response were dependent on the species, the mouse strain, and the viability of the inoculum. All humans had serum antibodies reactive with C. neoformans proteins regardless of the potential exposure, the HIV infection status, or the subsequent development of cryptococcosis. Our results indicate (i) a high prevalence of antibodies reactive with C. neoformans proteins in the sera of rodents after cryptococcal infection and in humans with or without HIV infection; (ii) qualitative and quantitative differences in the antibody profiles of HIV-positive individuals; and (iii) similarities and differences between humans, mice, and rats with respect to the specificity of the antibodies reactive with C. neoformans proteins. The results are consistent with the view that C. neoformans infections are common in human populations, and the results have implications for the development of vaccination strategies against cryptococcosis.

Infection with Cryptococcus neoformans is associated with an impaired cell-mediated immune response (reviewed in reference 31). Individuals with AIDS, renal transplants, and lymphoproliferative diseases and individuals receiving immunosuppressive therapy are at significantly greater risk for cryptococcosis than are immunocompetent individuals. Histopathological studies of experimental rodent and rabbit cryptococcosis indicate that granulomatous inflammation is essential for successful host immunity (16, 36). Thus, cellular immunity makes a critical contribution to host defense against C. neoformans (34).

In the past decade, several laboratories have shown that humoral immunity can also be important for host defense against C. neoformans (for reviews, see references 4, 5, and 37). Most studies of the antibody response to C. neoformans have focused on capsular polysaccharide and cell wall antigens (9, 12, 24, 41). In contrast, few studies have investigated the antibody response to protein antigens. Hamilton and colleagues have generated murine monoclonal antibodies to glycoprotein antigens of 36 to 38 kDa and of 30 kDa and studied the human and rodent response to these antigens (19, 21, 39). These authors also analyzed the antibody response to cryptococcal proteins in human immunodeficiency virus (HIV)-infected patients with cryptococcosis by isoelectric focusing and concluded that there may be several immunodominant antigens (20). Kakeya et al. reported that a 77-kDa protein belonging to the Hsp70 family was the immunodominant protein antigen in murine cryptococcal infection (23). Characterization of the antibody response to C. neoformans proteins in both humans and experimental animals is important because it may provide clues to the pathogenesis of infection and help to identify antigens recognized by the immune system. This study reports the serum antibody responses to cryptococcal proteins in HIV-positive and -negative humans and in rodent models of experimental cryptococcosis.

MATERIALS AND METHODS

Strains and growth conditions.

Strain 24067 (serotype D) was obtained from the American Type Culture Collection (Rockville, Md.). Strain SB4 (serotype A) is a clinical isolate obtained from E. Spitzer (Stony Brook, N.Y.), and strain J32 is a recent clinical isolate from New York City (40). Candida albicans SC5314 and Saccharomyces cerevisiae 1H1701 were obtained from M. Ghannoum (Cleveland, Ohio) and L. Marsh (Bronx, N.Y.), respectively. All fungi were grown in Sabouraud dextrose broth (Difco Laboratories, Detroit, Mich.) and stored in 50% glycerol at −80°C.

Fungal protein extracts.

Three types of C. neoformans protein extracts were used in this study: whole-cell, cytosolic, and membrane extracts. For each of these, C. neoformans 24067 was grown for 1 day at 30°C in Sabouraud dextrose broth. Culture volumes were usually 50 ml, and the starting cell concentration was approximately 10⁴/ml. The cells were collected by centrifugation (12,000 × g, 10 to 15 min, 4°C), and the pellet was washed twice with sterile cold distilled water and suspended 0.5 ml of cold lysis buffer containing 1 μM pepstatin A, 2 mM phenylmethylsulfonyl fluoride (PMSF), 10 mM Tris-HCl (pH 7.5), and 5 mM EDTA. The sample was divided into two microcentrifuge tubes each containing 1.2 ml, 0.6 g of glass beads (0.15 to 0.2 mm in diameter) was added to each tube, and the cells were disrupted by vortexing three times for 15 s at 4°C. For cell disruption the total volume of the bead-cell suspension was 2 to 2.5 ml. In other experiments the culture was scaled up to 300 ml, the cells were collected, washed, suspended in 10 ml of lysis buffer, and disrupted by vortexing a 10-ml suspension of cells with 5 to 7 ml of glass beads. For some experiments, the cell suspension was sonicated three times in an Ultrasonic Processor XL (Misonix, Farmingdale, N.Y.) at a setting of 3.5 for 30 s to enhance lysis. Cell debris was removed by centrifugation (3,000 × g, 2 min, 4°C), and the supernatants, representing the total cell extracts, were placed in tubes. The total protein yield for the scaled-up preparations was ∼15 mg and is referred to as whole-cell extract. For the separation of cytosol and membrane fractions, these extracts were centrifuged in a Beckman Optima Ultracentrifuge with the TLA100.3 rotor or a Beckman LS-50 Ultracentrifuge (Beckman Instruments, Inc., Palo Alto, Calif.) by using a Ti 60 rotor (100,000 × g, 1 h, 4°C). The membranes were washed with buffer containing protease inhibitors and collected by centrifugation (100,000 × g, 30 min, 4°C), and the wash was pooled with the previous supernatants as part of the cytosolic fraction. Membrane proteins were extracted by adding 3 ml of 50 mM Tris-HCl (pH 7.5)–1 mM EDTA–1 mM PMSF–1% sodium dodecyl sulfate (SDS) buffer to the pellets, heating the mixture to 65°C for 20 min, and removing the cell debris by centrifugation (12,000 × g, 15 min, 4°C). Protein concentrations were determined by using the Bio-Rad Protein Assay (Bio-Rad, Hercules, Calif.). Protein samples with concentrations of <1 mg/ml were lyophilized, dissolved in buffer containing protease inhibitors, and stored at −80°C. Protein extracts for C. albicans and S. cerevisiae cells were prepared as described above for C. neoformans cells except that the protein yields were 10 to 30 times greater than for cryptococcal cultures of comparable volume.

Animal experiments.

A/JCr and BALB/c mice and male Fischer rats were purchased from the National Cancer Institute (Bethesda, Md.). CBA/J mice were purchased from Jackson Laboratories (Bar Harbor, Maine), and Swiss Webster [Crl:CFW(SW)BR] and CF1 (Crl:CF-1BR) mice were purchased from Charles River Laboratories (Wilmington, Mass.). The numbers of mice used in each experiment are given in the tables. Mice were infected intratracheally (i.t.) with 10⁵ C. neoformans cells in one of the following combinations: strain 24067 alone; strains 24067 and SB4 (1:1); or strains 24067, SB4, and J32 (1:1:1). For the experiment with the live or dead inoculation, log-phase C. neoformans cells were divided into two batches, one of which was killed by treatment with either 0.5 M sodium azide for 3 h or heat at 65°C for 2 h. Killing was confirmed by plating. Killed cells were washed and suspended in sterile phosphate-buffered saline (PBS) prior to use in animal experiments. Mice were injected with either live or dead cryptococci intraperitoneally, and the serum was analyzed at day 35. This time was selected for analysis because it allowed sufficient time for the development of an immunoglobulin G (IgG) response, yet it was not so prolonged that the animals became sick and died.

Rats were infected i.t. with 10⁷ C. neoformans 24067 cells and killed at various times (1, 5, 6, 12, and 18 months) after infection, and blood was removed from the inferior vena cava. For some rats, dexamethasone phosphate (Sigma Chemical Co., St. Louis, Mo.) was added to the drinking water (1.5 mg/liter) at either 1 week or 11 months after infection, and this was continued for 6 weeks, at which time the animals were killed. Plating of the lung homogenates to determine the CFUs as described previously (16) revealed that all rats were chronically infected with C. neoformans. Sera were stored at −20°C.

Human serum.

Human studies were done in accordance with a protocol approved by the Committee on Clinical Investigations at the Albert Einstein College of Medicine. Sera were obtained from laboratory personnel who work with C. neoformans and from those who do not. Serum samples from HIV-positive individuals were obtained from the Multi-Center AIDS Cohort Study (MACS). The samples were received coded for a blinded study and included the following: 13 samples from HIV-positive patients who developed cryptococcosis subsequent to but within 6 months of the time the serum sample was obtained (HIV⁺/CN⁺); 13 samples from HIV-positive patients who did not develop cryptococcosis (HIV⁺/CN⁻); and 26 samples from HIV-negative patients (HIV⁻/CN⁻). Each individual who subsequently developed cryptococcosis was matched for CD4⁺ T-cell count, geographic location, and ethnic background to one HIV-positive individual who did not develop cryptococcosis and two HIV-negative individuals. The mean CD4⁺ T-cell count of the HIV-positive group was <100 cells/μl. All HIV-positive individuals were included only if they had no history of prior opportunistic infection and had not taken antifungal therapy. Serum samples were stored at −80°C prior to use. Upon thawing, the samples were heat inactivated at 56°C for 30 min and then stored at 4°C.

Immunoblotting.

Electrophoresis was done in the Bio-Rad Mini-Protean II system at 150 V in 25 mM Tris-HCl–192 mM glycine–0.1% SDS (pH 8.3). Transfer to nitrocellulose membranes (Schleicher and Schuell, Keene, N.H.) was done by the buffer tank method, with a buffer composed of 25 mM Tris-HCl, 192 mM glycine, and 0.1% SDS (pH 8.3) with 20% methanol at 250 mA for 0.7, 1.0, or 1.4 h for 7.5, 10, or 12% gels, respectively. Protein transfer was ascertained by staining the membranes with 0.1% Ponceau S in 5% acetic acid. The dye was washed off, and the membrane was blocked with 5% milk in Tris-buffered saline (TBS; 10 mM Tris, 150 mM NaCl [pH 7.2]) for 1 h at room temperature. Individual channels on a blotting frame (Idea Scientific, Minneapolis, Minn.) were incubated with sera diluted 1:50 for mouse sera; 1:100 or 1:250 for human sera; and 1:100, 1:250, or 1:750 for rat sera in blocking buffer. After primary antibody incubation the channels were washed with 0.1% Tween 20 (U.S. Biochemical Corp., Cleveland, Ohio) in TBS and then incubated with a horseradish peroxidase-conjugated goat antibody to human IgG (Southern Biotechnology Associates, Birmingham, Ala.). All antibody incubations were done for 2 to 4 h at room temperature or overnight at 4°C. The blots were then developed with chemiluminescent substrate (SuperSignal; Pierce, Rockford, Ill.) and exposed to film. Blots were aligned on the film, and bands were assigned to individual lanes. For the detection of rat antibodies, alkaline phosphatase-labeled goat antibody to rat IgG was used at a dilution of 1:1,000 and color developed with 5-bromo-4-chloro-3-indolyl phosphate (Sigma). For some experiments, sera were absorbed with protein extracts from C. neoformans, C. albicans, or S. cerevisiae prior to incubation with the blots. Absorption was done by mixing 5 μl of serum with 150 μg of either C. albicans or S. cerevisiae whole-cell extract (75 μg in the case of C. neoformans extract) in a total volume of 25 μl, followed by incubation for 1.5 h at room temperature.

Data analysis.

Immunoblot bands were traced onto transparency film. Protein bands were counted by aligning the gels and counting the number of sera from each group that showed reactivity to the protein. Data from the laboratory workers was analyzed by the Fisher exact test. Data from the MACS samples were analyzed by a 2 × 3 χ² test, and the significance was set at P ≤ 0.05 as described earlier (25).

RESULTS

Rodent studies. (i) Mouse antibody response to cryptococcal proteins after inoculation with live or dead cryptococci.

The serum antibody response in A/JCr mice was analyzed after inoculation of live or dead C. neoformans 24067 cells to determine whether infection was necessary to elicit an antibody response. Prior to infection, the mice had little or no antibody reactive with C. neoformans proteins (data not shown). Mice given live cells produced antibody responses to two to three proteins in the 61.5- to 83.6-kDa range and to several low-molecular-mass proteins (Fig. 1). Of five mice given dead cells, three (60%) produced antibodies to proteins of ca. 55 to 60 kDa, and all of them (100%) had antibodies to proteins of <37.6 kDa. The intensity of bands from the sera of mice given dead cells was weaker than that of mice given live organisms.

FIG. 1.

Immunoblots showing the reactivity of sera from A/JCr mice inoculated with either live or dead cells of C. neoformans 24067 with whole protein extracts from cryptococcal cells. Proteins were separated electrophoretically by SDS–7.5% polyacrylamide gel electrophoresis (PAGE). The treatment of the cells is indicated above the lanes, and the molecular mass standards are shown on the left in kilodaltons. Each lane contains sera from a different mouse. The dye front runs ahead of the 25.4-kDa marker.

(ii) Mouse antibody response to cryptococcal proteins of a mixed infection with genetically different C. neoformans strains.

To establish whether mixed infection would result in different antibody responses, A/JCr mice (n = 5 per group) were infected i.t. with strain 24067 alone; strains 24067 and SB4; or strains 24067, SB4, and J32. Strain SB4 has been previously shown to elicit antibody responses in rats after pulmonary infection (17). A control group received an i.t. administration of PBS alone. Western blot analysis revealed relatively few differences in antibody responses between mice receiving single and mixed infections at day 30 of infection (Fig. 2). One-half of all mice (9 of 18) had antibody responses to several proteins in the 61.5- to 83.6-kDa range. Antibody to proteins of <14 kDa was observed in 50% of the mice regardless of the infection group.

FIG. 2.

Immunoblots showing the reactivity of sera from A/JCr mice with whole protein extracts from cryptococcal cells after infection with a single or with multiple C. neoformans strains. Proteins were electrophoresed by SDS–10% PAGE. Labels above the lanes indicate the strain combination used to infect the mice. The PBS group represents mice given PBS i.t. Each lane contains sera from a different mouse.

(iii) Mouse antibody response to cryptococcal proteins in genetically different mouse strains.

Three inbred (A/JCr, BALB/c, and CBA/J) and two outbred (Swiss Webster and CF1) mouse strains were infected i.t. with strain 24067. Western blot analysis revealed significant mouse strain-to-strain variation in the specificity of the serum antibody response to cryptococcal proteins at day 30 of infection (Fig. 3 and Table 1). Serum from CF1, CBA/J, and A/JCr mice reacted with two to three cryptococcal proteins in the 61.5- to 83.6-kDa mass range (12 of 14 mice [79%]). Three of nine (33%) BALB/c and SW mice developed antibodies to these proteins. There was significant individual mouse-to-mouse variation in the intensity and specificity of the antibody response to cryptococcal proteins for all mouse strains studied.

FIG. 3.

Immunoblots showing the reactivity of sera from five mouse strains with whole protein extracts from cryptococcal cells after infection with strain 24067 at day 35 of infection. Each panel shows three lanes corresponding to three individual mice. Protein extracts were electrophoresed by SDS–10% PAGE and transferred to nitrocellulose.

TABLE 1.

Percentage of mice in five strains producing an antibody response to C. neoformans protein antigens in response to infection

Mousea	MHCb	nc	% Mice recognizing protein of approx mass (kDa)d:
			<14	20	30	37.6	40	61.5	61.6–70	83.6	110
A/JCr	a	5	20	0	0	20	0	40	40	20	100
BALB/c	d	5	80	0	0	20	0	20	20	0	40
CBA/J	k	5	100	80	40	0	20	80	80	0	60
SW	N/A	4	50	25	25	50	50	50	25	25	25
CF1	N/A	4	100	0	0	50	25	100	75	50	100

^aEach mouse strain was infected with a single i.t. dose of C. neoformans 24067. 

^bMajor histocompatibility complex (MHC) H-2 haplotype. SW and CF1 are outbred mouse strains. N/A, not applicable. Note that the haplotype may not be the only difference contributing to different antibody responses. 

^cn = number of mice in each group. 

^dMolecular mass was calculated relative to protein standards in 10% acrylamide gel. The numbers represent the percentage of mice with sera that recognized the specific protein band in immunoblots containing whole protein extracts from cryptococcal cells. 

(iv) Rat antibody response to cryptococcal proteins after infection.

Rats chronically infected with C. neoformans had serum antibodies reactive with multiple cryptococcal proteins (Table 2). Reactivity with proteins of 52, 55, 60.5, 64, 68, 72, 87, and 92 kDa was evident in sera from infected rats at all of the times examined, whereas reactivity for proteins of 38, 39, 105, and 110 kDa was detected only after 1.5 months of infection. Dexamethasone treatment resulted in a reduction in the number of proteins recognized by rat sera relative to the sera from nontreated rats at the various times after infection (Table 2). Rat sera reacted with several proteins of the same apparent molecular mass as those recognized by sera from infected mice (Table 3). No reactivity was observed with serum from noninfected rats with for the majority of these antigens (Table 2).

TABLE 2.

Percentage of rats that had serum antibodies to specific C. neoformans proteins at various times postinfection

Time (mos postinfection)a	nb	% Mice recognizing protein of approx mass (kDa) ofc:
		38	39	52	55	60.5	64	68	72	87	92	105	110	115
1.5	3	0	0	100	33	100	100	SM	SM	SM	100	0	0	0
1.5 D	3	0	0	0	0	0	33	SM	SM	SM	33	0	0	0
6	5	20	80	100	0	100	100	80	80	100	20	80	80	100
12	4	50	75	100	75	100	100	100	100	100	50	75	75	100
12 D	3	0	33	66	33	66	66	33	33	66	33	0	0	33
18	3	100	66	100	100	100	100	100	100	100	66	100	100	100
Control	2	0	0	0	0	1	0	0	0	0	0	0	0	0

^aD, rats treated with dexamethasone. Control refers to uninfected rats. 

^bn = number of rats studied at the indicated time. 

^cMolecular mass was calculated relative to protein standards in 7.5% acrylamide gel. The numbers represent the percentage of rats with sera that recognized the specific protein band in immunoblots containing whole protein extracts from cryptococcal cells. SM denotes a smeared band where precise identification of the molecular mass was not possible. 

TABLE 3.

Percentage of mice, rats, and humans with reactivity to specific protein antigens

Groupa	nb	% Individuals with antibody to protein antigen with approx mass (kDa)c of:
		31	38	39	52	55	60.5	64	68	72	87	92	105	110
Mice	22	14	27	18	0	0	50	Sm	Sm	Sm	18	0	0	68
Rats	15	NA	40	73	83	47	83	83	93	93	100	53	67	67
Rats (dexamethasone)d	6	NA	0	17	33	17	33	50	67	50	83	33	0	0
Exposed lab worker	5	40	60	80	40	40	40	0	0	0	Sm	Sm	Sm	0
Nonexposed lab worker	15	60	47	60	40	40	93	0	0	0	Sm	Sm	Sm	0

^aGroups include the cumulative prevalence of reactivity in all experiments. For rats the table includes the reactivity at the various time intervals. For mice the table includes the reactivity of the various mouse strains. To generate this table we considered common antigens between 31 and 111 kDa having those apparent molecular masses within 3.5-kDa range. This range was selected after taking into account interexperimental variation when comparing bands in different immunoblots. 

^bn is the number of individuals in the group. 

^cNumbers in the table are percentages of individuals with antibody to that protein antigen; values were calculated on the basis of the number of individuals in each group who had reactivity with the indicated protein in immunoblots containing whole protein extracts. Sm means “smear” and indicates an area in the blot where defined bands were not resolvable. NA means not available. For the rat studies reactivity below 38 kDa was not studied or analyzed. 

^dThis group includes rats given dexamethasone. 

Human studies. (i) Laboratory workers.

Sera from laboratory workers with or without potential exposure to C. neoformans contained antibodies reactive with cryptococcal proteins (Fig. 4 and Table 3). The intensity of the bands reactive with specific proteins varied between individuals. Individual sera detected as few as 2 to as many as 20 proteins. Figure 4 shows representative samples from seven individuals. The most prevalent antibody response was to proteins of ∼61.5, 40, and <25.4 kDa (Table 3). However, sera from some individuals exhibited greater reactivity to other proteins. Sera of individuals with potential laboratory exposure to C. neoformans were less likely to have reactivity to a protein of ∼61.5 kDa (P = 0.032; Fisher exact test).

FIG. 4.

Immunoblots showing the reactivity of sera from seven human subjects with cytosolic proteins from strain 24067 by separated by SDS–7.5% PAGE. Numbers above the lanes indicate the serum sample. All serum samples except sample 1 were used at a dilution of 1:250. Sample 1 is shown here at a dilution of 1:100. Molecular masses are indicated on the left in kilodaltons. The 25.4-kDa marker is just above the dye front.

(ii) Serum antibody in HIV-positive and -negative groups.

Serum from all HIV-positive and HIV-negative individuals had antibodies reactive with multiple cryptococcal proteins. Hence, we surveyed various conditions to identify those that would reveal differences between the groups, including reactivity with cytosolic and membrane fractions and various gel conditions. We noted differences in reactivity to several smaller-molecular-weight proteins and focused our efforts on the protein antigens of <60 kDa separated in the 12.5% gels (Table 4). With these conditions, we noted some differences in the pattern of cryptococcal proteins recognized by each group. Three major complexes were identified based on groupings of protein antigens recognized by the human samples (Table 4). There were considerable individual differences in the prevalence of antibodies to lower-molecular-mass proteins. Analysis of the reactivity of the MACS samples to the group of 24067 cytosolic proteins of 15, 17 to 19, and 21 kDa revealed a significant difference between the HIV⁺/CN⁺, HIV⁺/CN⁻, and HIV-negative groups (P = 0.039; χ² analysis). There is also a significant difference between the HIV-positive and -negative groups in their reactivity to these proteins (P = 0.026; χ² analysis). Sera from HIV-positive individuals were less likely to react with the proteins of 15, 17 to 19, and 21 kDa than were the HIV-negative individuals. There was no difference between the HIV⁺/CN⁺ and HIV⁺/CN⁻ groups.

TABLE 4.

Percentage of HIV-positive and HIV-negative individuals with serum antibody reactive with specific C. neoformans proteins

Groupa	nb	% Individuals with sera that recognized the protein band complex of approxc:
		15, 17–19, and 21 kDa		26, 27–28, 30, and 35 kDa		37–39, 42, and 48 kDa
		Cytosol	Membrane	Cytosol	Membrane	Cytosol	Membrane
HIV+/CN+	13	23	23	77	77	69	38
HIV+/CN−	13	31	38	85	85	77	77
HIV− negative	26	62	54	85	73	88	46
Pd		0.039	0.177	0.817	0.723	0.328	0.103

^aHIV⁺/CN⁺ are HIV-positive individuals who subsequently developed cryptococcosis; HIV⁺/CN⁻ are HIV-positive individuals who did not develop cryptococcosis. 

^bn denotes the number of individuals in each group. 

^cComplexes are sets of proteins recognized by individual serum samples. Molecular was mass calculated relative to protein standards in a 12% acrylamide gel. For this study we used an acrylamide gel of higher percentage because preliminary studies suggested differences in proteins with smaller molecular masses. The numbers represent the percentage of individuals with sera that recognized the specific protein band by Western blot analysis. “Cytosol” and “membrane” refer to the type of protein preparation of strain 24067 used in the Western blot analysis. Sera were evaluated at a dilution of 1:100. 

^dP values were calculated from a 2 × 3 χ² analysis. P values of <0.05 indicate significant differences between the three groups. 

(iii) Cross-reactivity with other fungal proteins.

Cross-reactivity with other fungal proteins was determined by immunoblotting before and after absorption of serum with protein extracts from C. albicans and S. cerevisiae. Adsorption of human serum with C. albicans protein extract abolished reactivity for C. albicans but did not abolish reactivity to C. neoformans proteins (data not shown). Adsorption of serum with S. cerevisiae protein extracts did not abolish reactivity to cryptococcal proteins (data not shown). Absorption of human serum with C. neoformans protein extracts reduced, but did not abolish, the intensity of some bands, a fact that may reflect a higher antibody affinity for immobilized antigens.

DISCUSSION

Serologic studies provide the foundation for understanding the epidemiology and pathogenesis of many infectious diseases. Many studies have analyzed the serum antibodies of mice and humans that are reactive with GXM (3, 6, 9, 10, 13–15, 22). Two major concepts have emerged from these studies: (i) rodents have little preexisting antibody reactive with GXM, but mount antibody responses after infection and vaccination (6, 10, 15, 16); and (ii) humans have preexisting antibodies reactive with GXM and these antibodies are found in both HIV-positive and HIV-negative individuals, i.e., in a population that is susceptible to cryptococcosis and in one that is relatively resistant, respectively (9, 22, 24, 41). The latter finding supports the concept that humans may have ongoing exposure to C. neoformans. However, definite conclusions from serologic studies of GXM-binding antibodies can be difficult because GXM has cross-reactive epitopes with other fungal polysaccharides (11) and GXM is a T-cell-independent antigen that classically elicits weak antibody responses. Thus, there is a need to develop more-specific serologic tools to investigate problems in the epidemiology and pathogenesis of C. neoformans infections. Analysis of antibody responses to protein antigens can provide insight into immunodominant antigens. For example, studies of the antibody response to Histoplasma capsulatum protein antigens led to the discovery of protein antigens that also elicit protective cell-mediated responses (reviewed in reference 8).

The mouse and rat studies reported herein revealed no significant serum antibody levels to C. neoformans proteins before infection. These animals were maintained in cages without taking extraordinary precautions to avoid ambient pathogens. Since these rodents are colonized with endogenous microbial flora, the lack of antibody in the absence of infection suggests that naturally occurring antibodies in rodents are not cross-reactive with C. neoformans proteins. Both mice and rats generated antibodies reactive with C. neoformans proteins after infection. For both rats and mice, there was individual variation in the antibody response. The genetic background of the mice influenced the magnitude of the antibody response with respect to the intensity and number of protein bands. The specificity and intensity of the antibody response in mice inoculated with live or dead C. neoformans was different, suggesting that different protein antigens are recognized during an infection. Live inoculation may also produce a larger antigenic burden as C. neoformans reproduce in tissues. Mouse infections with one or more C. neoformans strains elicited similar antibody responses, suggesting that the protein antigens recognized in mixed C. neoformans infection are the same. Hence, mice and rats make specific antibody responses to C. neoformans proteins after experimental infection, and their responses depend on the viability of the inoculum and the genetic background of the host.

In contrast to mice and rats, human sera contained antibodies reactive with C. neoformans proteins even when there was no history of cryptococcal infection. Comparison of the reactivity of human, rat, and mouse sera for C. neoformans antigens revealed several noteworthy similarities and differences (Table 3). Overall, the serum antibodies in the three species react with proteins of similar molecular mass, a fact that supports the specificity of the reactivity of the human sera with C. neoformans antigens. Some antigens (i.e., those of 64, 68, 72, and 111 kDa) were recognized by sera of rats and mice, whereas others (i.e., those of 52, 55, 92, and 107 kDa) were recognized by sera of human and rats. Since humans and rats are highly resistant to C. neoformans infection, whereas mice are very susceptible, these serological differences could provide insight into the antigens that elicit protective immune responses.

The most provocative result was the finding that all human sera tested had antibodies reactive with cryptococcal proteins. Analysis of the human sera revealed the presence of antibodies reactive with cryptococcal proteins in the sera of both HIV-positive and HIV-negative individuals. These antibodies may have been elicited by cryptococcal proteins or by cross-reactive antigens. Although our data cannot unambiguously distinguish between these two possibilities, three lines of evidence suggest that the antibodies in human sera are specific for C. neoformans: (i) C. neoformans protein antigens of similar mass (see below) are recognized by both human and rodent antibodies; (ii) the reactivity of human sera with C. neoformans proteins was not abolished by absorption with C. albicans or S. cerevisiae extracts; and (iii) infection with C. neoformans was required to elicit antibodies reactive with cryptococcal infection in both mice and rats. We interpret the ubiquity of antibodies to C. neoformans proteins in human sera to be indicative of and consistent with past exposure to subclinical infection with or latent asymptomatic infection with C. neoformans. This interpretation is consistent with the following: (i) a high likelihood for frequent exposure to C. neoformans because it is common in urban environments (7, 26), (ii) delayed-hypersensitivity skin reactions in a some individuals with no history of clinical cryptococcosis (1, 18, 32, 33, 38), (iii) the presence of healed pulmonary cryptococcal lesions in human lungs consistent with primary infection (2), and (iv) the proposal that asymptomatic C. neoformans infection is common in human populations (27, 28) and the suggestion that cryptococcosis represents reactivation of a latent infection (35).

Another striking observation was the heterogeneity in the specificity of antibodies reactive with cryptococcal proteins, as reflected by the fact that no two individuals had identical immunoblot patterns. Although the mechanisms responsible for this phenomenon in human responses are not understood, there are several potential explanations. First, individual differences in the timing and extent of exposure to cryptococcal antigens may have elicited different antibody responses. Second, differences in the type and timing of exposure relative to serum sampling may contribute to individual variation. In this regard, chronic exposure to endogenous proteins generated during a latent infection could contribute to the complex patterns of antibody reactivity observed with the human sera. Heterogeneous and variable antibody responses to the latent intracellular pathogen Mycobacterium tuberculosis have also been described in humans and cows (29, 30). The findings with M. tuberculosis are consistent with our observation that antibody responses diversify over time in chronically infected rats, possibly as new antigens are exposed. Third, the observation that the mouse antibody response is influenced by the genetic background of the host suggests that genetic factors may contribute to the diversity among human antibody responses.

We did not observe major differences in the intensity or specificity of the antibody responses of individuals with or without potential laboratory exposure to C. neoformans. Those individuals with potential laboratory exposure to C. neoformans had each worked with strain 24067. Individuals with potential laboratory exposure were less likely to have antibodies to a 61.5-kDa protein than were individuals without a history of laboratory exposure to strain 24067. This result must be interpreted with caution because of the small size of the laboratory worker group. However, the possibility of different antibody responses resulting from work-related exposure is consistent with reports that laboratory microbiologists are more likely to have skin reactivity to cryptococcal antigens (38). The absence of major serological differences between individuals with or without potential laboratory exposure is consistent with infection or exposure to C. neoformans before laboratory exposure with this organism or the absence of significant exposures in the laboratory environment.

We noted some differences in the prevalence of serum antibodies to specific proteins among individuals with or without HIV infection. The implications of this finding are uncertain. In the setting of a compromised immune system, human C. neoformans infections can disseminate, perhaps interrupting a “balance” between latent infection and the host response. In light of this possible scenario, there was a statistically significant trend toward less reactivity among HIV-positive samples and a trend toward even less reactivity among the patients who subsequently developed cryptococcosis. Several explanations could support this observation: (i) preexisting antibodies may be depleted by the binding to proteins produced by multiplication of organisms during infection; (ii) new antibodies to cryptococcal proteins may not be produced in the setting of T-cell deficits that impair the ability to generate new antibodies; and (iii) the epitopes of cryptococcal proteins may be different in the setting of disseminated infection. Although our data clearly demonstrate the existence of heterogeneous antibody responses to cryptococcal proteins in both HIV-positive and HIV-negative individuals, additional studies are needed to identify the specific proteins recognized.

Only one other study has investigated the human antibody response to C. neoformans protein antigens. Hamilton et al. recently reported the reactivity of sera from 20 HIV-positive patients with a history of infection with C. neoformans var. neoformans and from 15 control patients with no history of cryptococcosis with cryptococcal protein extracts (20). These investigators reported that patient sera frequently recognized protein antigens of 26, 52, 74, 110, and 114 kDa by Western blot analysis. Each of these proteins has a molecular mass comparable to one of the protein antigens recognized by the human and rodent sera in our study. Hamilton et al. also found patient-to-patient variation in the serological response (20). However, in contrast to our results, only 3 of 15 HIV-negative patients in their study had serum antibody reactive with cryptococcal proteins. This difference may reflect differences in the methodology used by the two studies, given that the earlier study used fractions purified by isoelectric focusing (20). Furthermore, there may be differences in the exposure between individuals in the United Kingdom and the United States. Nevertheless, heterogeneity and subject-to-subject variation in antibody response was observed in both studies.

In summary, C. neoformans infection elicits detectable antibody responses to cryptococcal proteins in rodents. In humans, antibodies reactive with many cryptococcal proteins are common despite the absence of clinical cryptococcosis or a history of previous infection. This observation strongly supports the notion that humans may be immunized by chronic exposure to environmental C. neoformans and/or that they harbor chronic, latent foci of infection. Our observations may have important implications for the prevention and therapy of C. neoformans. If humans harbor latent infections, then vaccine strategies for adults should aim to control reactivation, rather than prevention of the new infection. If cryptococcosis does represent reactivation, then vaccine development may require the identification of targets that may not necessarily be present in a primary infection or in laboratory cultures. If infection does occur in early childhood, vaccination strategies to prevent infection would need to be offered at that time. Thus, our results strongly support the need for longitudinal studies of antibody profiles to cryptococcal proteins in children and adults.