INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Cryptococcus neoformans is a fungus which is a frequent cause of life-threatening meningoencephalitis in patients with impaired immunity (22, 25). Cryptococcosis has been reported to occur in 6 to 8% of patients with AIDS (7). In immunocompromised individuals, C. neoformans infections are often incurable with conventional antifungal agents, and these patients frequently require lifelong therapy (45). The difficulties involved in the management of cryptococcosis in immunocompromised individuals have led to a reexamination of the potential of antibody-mediated immunity for prevention and therapy of cryptococcal infections. A polysaccharide-tetanus toxoid (TT) conjugate vaccine which is highly immunogenic and can elicit protective antibodies in mice has been made (3, 8, 9). In addition, several monoclonal antibodies (MAbs) have been shown to modify the course of infection in mice, and these may be useful in therapy of human infection (12, 14, 28, 42, 43).

Cell-mediated immunity is generally acknowledged to provide important host defense against C. neoformans infection (4, 20, 26, 31, 42). In contrast, the role of antibody-mediated immunity in host resistance is less certain (2), but there is considerable evidence that administration of some MAbs can modify the course of infection in mice (8, 12, 14, 16, 28, 33). C. neoformans is unusual among fungal pathogens in that it has a polysaccharide capsule composed primarily of glucuronoxylomannan (GXM) (6), which is important for virulence (5). The capsular polysaccharide has been shown to produce a variety of deleterious effects including inhibition of phagocytosis (21), interference with antigen presentation (39), shedding of adhesion molecules (11), inhibition of leukocyte migration (10), and alterations in cytokine production by host effector cells (24, 40, 41). Antibodies to the C. neoformans capsular polysaccharide may contribute to host defense through multiple effects including enhanced opsonization (13, 18, 23, 30, 44), clearance of polysaccharide antigen (15), promotion of granuloma formation (14), and release of oxygen- and nitrogen-derived oxidants (27, 38).

In previous studies, we demonstrated that immunoglobulin G3 (IgG3) MAbs are not protective in various mouse models of cryptococcal infection (32, 42). When one of these nonprotective IgG3 MAbs was switched to IgG1, the IgG1 significantly prolonged animal survival (32, 42). In the present study, we analyzed two families of IgG switch variants generated in vitro from two nonprotective IgG3 MAbs with different epitope specificities. We found that MAbs with different isotypes have different protective efficacies and that switching of nonprotective IgG3 MAbs to IgG1, IgG2b, and IgG2a significantly increased antibody protective efficacy. These studies demonstrate a complex relationship among efficacy of antibody protection, epitope specificity, and isotype.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

C. neoformans. Strain 24067 (serotype D) was obtained from the American Type Culture Collection (Rockville, Md.) and maintained on Sabouraud dextrose agar (Difco, Detroit, Mich.) at 4°C. For murine infection, C. neoformans was grown at 37°C in Sabouraud dextrose broth (Difco) for 24 h. Yeast cells were washed three times with phosphate-buffered saline (PBS), and the inoculum was determined by counting in a hemocytometer and was confirmed by counting the number of colonies after plating on Sabouraud dextrose agar.

MAbs. The 3E5 IgG3 MAb was generated in response to immunization with the GXM-TT vaccine (3), and it binds all four serotypes of C. neoformans (3). The 4H3 IgG3 MAb was generated in response to infection with the C. neoformans strain GH (3). The IgG1, IgG2a, and IgG2b switch variants of MAb 3E5 IgG3 MAb and the IgG1 and IgG2b variants of MAb 4H3 IgG3 were generated by in vitro isotype switching as described elsewhere (35, 42). Ascites fluid containing hybridoma protein was obtained by paracentesis of BALB/c mice injected with 10⁷ hybridoma cells into the peritoneal cavity. Prior to hybridoma injection, the mice were primed with Pristane (Sigma Chemical Co., St. Louis, Mo.). For in vitro studies, antibodies were purified on protein G-Sepharose columns (Pierce, Rockford, Ill.) and their concentrations were determined by enzyme-linked immunosorbent assaying (ELISA) relative to isotype-matched standards of known concentrations. Previous studies had established that the MAb constituted approximately 95% of the immunoglobulin of the relevant isotype in the ascites (28). In previous work, we have shown that ascites induced by the NSO myeloma cell line and ascites containing irrelevant isotype-matched MAbs do not modify infection with C. neoformans (28). In addition, in these experiments the ascites containing the different isotypes served as internal controls for each other, since the switch variants were products of the same hybridoma.

ELISA. Antibody binding to GXM was studied by ELISAs that have been described elsewhere (3). Briefly, the wells in the ELISA plates were coated overnight with 50 µl of a solution of 1-µg/ml GXM from C. neoformans 24067. The plates were incubated for 1 h at 37°C with serial dilutions of purified MAbs, washed, and incubated with alkaline phosphatase-labeled goat anti-mouse IgG MAbs (Fisher Scientific, Orangeburg, N.Y.). The plates were then washed and developed by adding substrate and determining absorbance at 405 nm. Competition assays between the 3E5 IgG and 4H3 IgG switch variants were performed by adding a constant amount of one MAb (1 to 5 µg/ml) and varying the concentration of the other isotype (0 to 10 µg/ml).

Animal experiments. Female C5 complement-deficient A/JCr mice aged 6 to 8 weeks were purchased from the Jackson Laboratory (Bar Harbor, Maine). Protection experiments were performed as described elsewhere (42). Briefly, 10 mice per group were given 1 mg of IgG1, IgG2a, IgG2b, or IgG3 MAb or PBS. MAbs were administered via intraperitoneal injection (i.p.) 24 h prior to intravenous (i.v.) challenge with 2 × 10⁶ C. neoformans 24067 cells, and mouse deaths were recorded daily.

Macrophage experiments. Bronchoalveolar macrophages were obtained from A/JCr mice by lung lavage as described previously (14). For the phagocytosis assay, 10⁵ cells per well were plated in 96-well tissue culture plates (Falcon; Becton Dickinson, Mountain View, N.J.) and cultured overnight at 37°C in Dulbecco's modified Eagle's medium with 10% fetal calf serum (FCS) with or without 500 U of gamma interferon (IFN-) (Genzyme, Cambridge, Mass.) per ml. Virtually all of the cells attach, spread out, and extend pseudopods. C. neoformans was incubated with purified MAb for 1 h at 37°C and then added to the macrophages. After the addition of C. neoformans, the plates were incubated at 37°C for 4 h and washed three times with sterile PBS to remove nonadherent yeast cells and the occasional cell from the lung lavage that had not attached, and the attached cells were fixed with cold absolute methanol and stained with a 1:20 solution of Giemsa (Sigma). Phagocytic indices were determined with a microscope at a magnification of ×600 (Diaphot; Nikon Inc., Melville, N.Y.). Internalized organisms were distinguished from ones that were attached by their presence in vacuoles. In prior studies, we used Uvitex B dye (CIBA, Summit, N.J.) to confirm that we could distinguish attached organisms from internalized ones (14, 29, 30). The phagocytic index was defined as the number of macrophages that had internalized two or more yeast cells over the number of total macrophages per field. Eight fields were counted for each experiment.

Statistics. Data were analyzed with statistical software for MacIntosh (Instat version 2.01; GraphPDA Software for Science, San Diego, Calif.) by the unpaired Student t test for CFUs. The alternative Weltch t test was used for animal survival studies. All results were also analyzed by the unpaired Wilcoxon test, which gave similar results (32).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Binding of isotype switch variants to GXM. Generation of the IgG1, IgG2a, and IgG2b switch variants of MAb 3E5 and the IgG1 and IgG2b switch variants of MAb 4H3 has been described elsewhere (36, 42). Several ELISAs were used to test whether the switch variants had similar levels of antigen binding to the parental IgG3 MAbs. At low antibody concentrations, the IgG3 MAb 3E5 and its IgG1, IgG2b, and IgG2a switch variants bind to GXM in a similar fashion (Fig. 1A). However, at higher antibody concentrations, the IgG3 MAb produces higher absorbancy, in agreement with the ability of IgG3 subclass antibodies to bind more strongly as a result of polymerization after binding antigen (17). The 4H3 MAbs bound GXM less strongly than the 3E5 MAbs, and the 4H3 switch variants had similar antigen binding curves (Fig. 1B). Each of the switch variants of 3E5 competed with each other for binding, confirming that switching was not associated with a change in specificity (Fig. 1C). Hence, the levels of binding of each member of a family of switch variants to GXM were similar, as expected for antibodies which have the same antigen binding site but differ in isotype.

View larger version (28K): [in this window] [in a new window]   FIG. 1.   (A and B) ELISAs of 3E5 IgG (IgG1, IgG2a, IgG2b, and IgG3) and of 4H3 IgG (IgG1, IgG2b, and IgG3) MAb binding to GXM antigen, respectively; (C and D) competition assays. The details of each ELISA are described in Results and are diagrammed in the insert of each panel.

MAbs 3E5 and 4H3 have different epitope specificities. In previous studies, we have shown that MAbs 3E5 and 4H3 have different reactivities with C. neoformans strains of the various serotypes, a finding that suggests differences in epitope specificities (28). To more rigorously investigate the epitope specificities of MAbs 3E5 and 4H3, we carried out competition experiments by ELISA. No competition was observed between MAbs 3E5 and 4H3 for binding GXM (Fig. 1D). The inabilities of MAbs 3E5 and 4H3 to inhibit each other in binding GXM confirm that these MAbs bind to different epitopes. Additional evidence for a difference in epitope specificity comes from peptide-binding studies, which show that peptides (37) that bind to the 3E5-binding site do not bind to 4H3. Hence, 3E5 and 4H3 bind to different epitopes on C. neoformans GXM.

Efficacy of protection by IgG3 MAbs 3E5 and 4H3 and their switch variants. Administration of MAb 3E5 IgG3 to mice prior to i.v. infection did not prolong survival, in agreement with prior findings that this MAb was not protective against C. neoformans (42). Although there appeared to be a reduction in survival compared to controls (Fig. 2A), the difference was not significant (P = 0.05). In contrast, groups of mice given 3E5 IgG1, IgG2b, and IgG2a switch variants each lived significantly longer than the control group given PBS. Mice given 3E5 IgG2a lived longer than those given 3E5 IgG1, but this difference was not significant (P > 0.08). These results confirm prior experiments which show that isotype switching of 3E5 IgG3 to IgG1 converts a nonprotective antibody to a protective antibody (42) and demonstrate that the IgG2a and IgG2b variants of 3E5 can also mediate protection.

View larger version (22K): [in this window] [in a new window]   FIG. 2.   Survival of A/JCr mice infected with C. neoformans after the administration of IgG3 isotype switch variant MAbs. A dose of 1.0 mg of each antibody was given i.p. 24 h prior to i.v. challenge with 2 × 106 C. neoformans cells. (A) Average survival rates ± standard deviations for the 3E5 IgG3, IgG1, IgG2b, and IgG2a and PBS groups were 11.6 ± 0.5, 14.4 ± 2.6, 14.0 ± 1.7, 15.8 ± 2.6, and 12.1 ± 1.0 days, respectively. IgG1, IgG2b, and IgG2a MAbs prolonged animal survival significantly (P < 0.003), whereas the IgG3 MAbs showed a trend toward decreased survival that was not statistically significant (P > 0.18). (B) Average survival rates ± standard deviations for the 4H3 IgG3, IgG1, and IgG2b and PBS groups were 5.8 ± 3.8, 16.3 ± 3.2, 17.9 ± 2.8, and 14.1 ± 3.1 days, respectively. Administration of 4H3 IgG3 MAb reduced survival significantly (P < 0.0001). However, the 4H3 IgG1 switch variant showed no effects on survival (P > 0.14), and 4H3 IgG2b prolonged animal survival (P < 0.01).

View larger version (15K): [in this window] [in a new window]   FIG. 3.   Survival of A/JCr mice infected with C. neoformans after two administrations of the 3E5 IgG3 MAb. A dose of 1 mg of the 3E5 MAb was given i.p. 24 h prior to and 3 days after i.v. challenge with 2 × 106 C. neoformans cells. Average survival rates ± standard deviations for the groups given IgG3 once and twice and the PBS group were 11.5 ± 1.1, 9.0 ± 3.3, and 14.1 ± 1.2 days, respectively. Administration of two doses of IgG3 3E5 reduced survival relative to one dose (P < 0.002).

View larger version (18K): [in this window] [in a new window]   FIG. 4.   Survival of A/JCr mice infected with C. neoformans after administration of IgG3 MAbs 4H3 and 3E5. A dose of 0.5 mg of each antibody was given i.p. 24 h prior to i.v. challenge with 2 × 106 C. neoformans. Average survival rates ± standard deviations for the IgG3 (3E5), IgG3 (4H3), IgG3 (3E5 plus 4H3), and PBS groups were 8.4 ± 2.3, 6.6 ± 2.1, 4.1 ± 0.7, and 11.3 ± 2.6 days, respectively. Administration of IgG3 MAbs 3E5 and 4H3 significantly reduced survival (P < 0.004) compared to IgG3 3E5 or IgG3 4H3 treatment alone.

Macrophage experiments. Primary lung macrophages from A/JCr mice were used to study the opsonic efficacies of 3E5 and 4H3 IgG3 MAbs and their switch variants. In the absence of antibody, there was little or no phagocytosis even if the macrophages were stimulated with IFN- (Fig. 5). All MAbs promoted phagocytosis, and phagocytosis was significantly greater in cells treated with IFN-. For both 3E5 and 4H3, the relative efficacy of the various subclasses in promoting phagocytosis was IgG2b > IgG1 > IgG3. To assess the relative efficacy of the 3E5 IgG3 MAb in enhancing macrophage antifungal activity against C. neoformans, CFUs were determined after 2 and 24 h of macrophage-yeast coculture in the presence or absence of 3E5 MAbs. Table 1 shows that IgG1-, IgG2a-, and IgG2b-mediated phagocytosis led to a decrease in CFU at 2 and 24 h compared to controls. In contrast, in the presence of IgG3 MAb, the number of C. neoformans cells increased after 24 h in culture compared to controls that lacked antibody (P < 0.006).

View larger version (23K): [in this window] [in a new window]   FIG. 5.   Effects of 3E5 and 4H3 MAb switch variants on phagocytosis by A/JCr bronchoalveolar macrophages in the presence or absence of IFN-. All MAbs promoted phagocytosis, and phagocytosis was significantly greater in cells treated with IFN- (P < 0.01).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The specificities of antibody isotype and epitope have previously been shown to be important determinants of the efficacy of antibodies against C. neoformans (28). The importance of the isotype was first suggested by Kozel and collaborators, who showed that IgG2a and IgG2b isotype switch variants of MAb 471 differed in their abilities to reduce organ fungal burden in murine experimental infections (33). The IgG isotype variants also differed in their abilities to promote phagocytosis, with the relative efficacy being IgG2b > IgG2a > IgG1 (33). Mukherjee et al. subsequently demonstrated differences among IgM, IgG1, IgG3, and IgA MAbs (28). Most striking was the fact that IgG3 MAbs did not protect against C. neoformans infection despite the ability of this class to protect against other encapsulated pathogens such as Streptococcus pneumoniae (1). Isotype switching of 3E5 IgG3 to IgG1 converted a nonprotective antibody into a protective antibody (42), providing unequivocal evidence for the importance of isotype in antibody-mediated protection against C. neoformans. IgG3 MAbs were subsequently shown to function as blocking antibodies in cryptococcal infections, with the potential to either enhance infection or reduce the efficacy of protective MAbs, depending on the experimental conditions (32). Evidence for specificity as a critical determinant of efficacy against C. neoformans was obtained by comparison of IgM MAbs which differ in their locations of capsular binding. For another fungus, Candida albicans, the efficacy of antibody protection has been shown to depend on epitope specificity (19). Hence, both isotype and specificity are known to contribute to the efficacy of protection, but their relative contributions and mechanisms of action remain obscure.

In this study, we analyzed two families of isotype switch variants derived from the 3E5 and 4H3 nonprotective IgG3 MAbs. MAbs 3E5 and 4H3 differ in heavy- and light-chain variable gene usage. MAb 3E5 is a member of group II anticryptococcal MAbs, which include many protective MAbs, and is assembled from V_H441, J_H3, V lamda 1, and J lambda 1 gene elements (3). MAbs 3E5 and 4H3 did not compete for binding to capsular polysaccharide and hence must bind to different epitopes. The switch variants reacted with capsular polysaccharide in a manner similar to that of each parental IgG3 MAb, in agreement with the identical epitope specificities expected for isotype switch variants.

In all experiments with immunocompetent mice, the two IgG3 MAbs were either not protective or enhanced infection. Additional evidence for the deleterious effects of this isotype in cryptococcal infection was obtained by showing that two doses of MAb 3E5 shortened survival more than a single dose. Furthermore, the combination of the two IgG3 MAbs 3E5 and 4H3 shortened survival more than either antibody alone. The inability of IgG3 MAbs to mediate protection is intriguing when one considers that IgG3 class antibodies tend to be elicited by type 2 T-cell-independent antigens such as cryptococcal capsular polysaccharide. Hence, it appears that the IgG subclass most likely to be elicited by the capsular polysaccharide antigen is also the least protective. However, it is interesting that a GXM-TT vaccine primarily elicits IgG1 antibodies to GXM (3) and that antibody response is protective (8). Isotype switch variants of both 3E5 and 4H3 IgG3 MAbs were significantly more effective at prolonging survival than the parental IgG3 MAbs. For MAb 3E5, the IgG1 switch variant has previously been shown to be protective (42). Here, the IgG2a and IgG2b switch variants have also been shown to be protective. For MAb 4H3, the IgG1 switch variant was previously shown to be nonprotective (32). Here, the 4H3 IgG2b switch variant is shown to prolong survival, demonstrating that the efficacy of protection of an IgG1 MAb can be increased by subsequent isotype switching to downstream isotypes. This result shows that conclusions regarding the protective potential of antibodies against a given epitope cannot be made from studies with one or even two isotypes.

The relative efficacy of protection of the 3E5 isotype switch family differed from that observed with isotype variants from the IgM MAb 2D10, which followed the order IgG2a > IgG1 > IgG2b (30). In contrast, the relative efficacy of protection of the 3E5 switch family closely paralleled that observed by Schlageter and Kozel (34) with isotype variants of MAb 471, which followed the order IgG2a > IgG2b > IgG1. Furthermore, for 4H3 the efficacy of IgG2b was greater than that of IgG1. MAbs 3E5, 2D10, and 471 are similar to each other but have amino acid differences in their variable regions which may contribute to differences in fine specificity. In this regard, epitope mapping with phage peptide libraries has shown differences between the 3E5 and 2D10 binding sites (37). As demonstrated here, MAb 4H3 clearly binds to a different epitope than does MAb 3E5. Hence, these results suggest that isotype efficacy varies, depending on the specificity of the antibody in question.

All of the IgG classes were opsonic for C. neoformans with bronchoalveolar macrophages. Treatment of bronchoalveolar macrophages with IFN- enhanced phagocytosis for all isotypes. This phenomenon may reflect the ability of IFN- to increase Fc receptor expression. For both MAbs 3E5 and 4H3, the relative opsonic efficacy for C. neoformans ingestion by mouse bronchoalveolar macrophages was greater for IgG2b and IgG1 than that for IgG3. The opsonic efficacies of the various isotype switch variants of MAbs 3E5 and 4H3 are different from those observed by Schlageter and Kozel (34), who found a relative efficacy of IgG2a > IgG2b > IgG1 for isotype switch variants of MAb 471. Furthermore, this result differs from that of Mukherjee et al. (29), who found the relative efficacy for isotype switch variants of the IgM antibody 2D10 to be IgG1 > IgG2a > Ig2b. Opsonic efficacy is likely to reflect antibody affinity for antigen, surface epitope density on the target cell, and Fc receptor density on the effector cell. These differences may reflect variation in epitope specificity or experimental differences, given the use of different macrophages and C. neoformans strains. For example, Schlageter and Kozel used a serotype A strain with peritoneal macrophages (34), Mukherjee et al. used a serotype D strain with J774.16 cells (30), and the present study used a serotype D strain with bronchoalveolar macrophages. In any case, all IgG subclasses are opsonic for C. neoformans, and it is unclear if the minor differences observed among switch variants of the different MAbs are biologically relevant, given that both protective and nonprotective antibodies were opsonic.

Previous studies have shown that IgG-type MAbs can enhance the antifungal efficacies of murine J774.16 cells (29, 30) and alveolar macrophages (14). However, none of the prior studies has demonstrated qualitative differences between protective and nonprotective MAbs in vitro. In this study, we found that IgG1, IgG2a, and IgG2b MAbs reduced CFUs when they were added to mixtures of IFN--stimulated alveolar cells and C. neoformans. In contrast, the 3E5 IgG3 MAb was not effective in promoting reduction in CFUs at either 2 or 24 h. This result is different from that observed with IFN-- and LPS-stimulated J774.16 cells, in which all of the IgG subclasses, including IgG3, were found to enhance the antifungal efficacy of the macrophage-like cells (30). Presumably, the differences in experimental results reflect differences in the use of primary macrophage cells versus an immortal macrophage-like cell line. This observation suggests that primary macrophage cells may provide a useful system for dissecting MAb isotype differences in vitro. In any case, the findings in this study imply that the interaction of IgG3 with C. neoformans and macrophages is different from that of the other isotypes and suggest a plausible explanation for the lack of efficacy of this IgG subclass.

In summary, the experience with MAb 4H3 indicates that one cannot conclude that an epitope elicits nonprotective antibodies solely on the basis of MAbs of one or two isotypes to that epitope. The results suggest that to make definitive conclusions about the potential of a given epitope to generate protective or nonprotective antibodies, one must evaluate several, if not all, isotype variants of a MAb with that epitope specificity. Protective antibodies can be made more protective by switching their constant regions to other isotypes. The contributions of isotype and epitope specificities to the efficacy of protection appear to be interdependent and provide yet another layer of complexity to the structure-function relationship of anticryptococcal antibodies.