An Opsonizing Monoclonal Antibody That Recognizes a Noncapsular Epitope Expressed on Cryptococcus neoformans

Abstract

A mouse hybridoma secreting a monoclonal antibody (MAb) that bound a noncapsular epitope expressed on C. neoformans was developed by immunizing BALB/c mice with formalin-killed serotype A yeasts. The hybridoma, designated CSFi, secreted an immunoglobulin G2b MAb that reacted with all C. neoformans serotypes tested, including the acapsular mutant ATCC 52817 (Cap67). Postsectioned immune electron microscopy revealed extensive binding of the MAb to the cell walls of both encapsulated and acapsular yeasts. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blot analysis of secreted antigens recovered from concentrated culture supernatants from both encapsulated and acapsular strains was conducted. The results showed that this MAb bound predominantly to antigens with molecular masses of approximately 75 and 100 kDa. A competitive enzyme-linked immunosorbent assay was used to demonstrate that the MAb was not cross-reactive with purified glucuronoxylomannan derived from either serotypes A or D. Experiments conducted with mouse peritoneal phagocytes and the mouse phagocyte-like cell line, J774A.1, demonstrated that the CSFi MAb opsonized the yeasts and increased their adherence to both types of phagocytic cells. We conclude, therefore, that antibodies directed at noncapsular epitopes can serve as opsonins and may have a role in modulating cryptococcal infection.

The pathogenic yeast, Cryptococcus neoformans, is an important cause of life-threatening infections in individuals with AIDS and other immunocompromising diseases. Virulence factors associated with this yeast include the capsule, melanin production, mating type, mannitol production, and the ability to grow at 37°C (1, 2, 4, 11, 27). Of these virulence factors, the major capsular polysaccharide, glucuronoxylomannan (GXM) has been studied the most extensively with regard to its effects on the host’s immune system and the host’s specific immunologic response to it. While cell-mediated immunity and the innate immune responses are recognized as pivotal in the defense against disseminated cryptococcal infection, there is growing evidence that the humoral immune response may be at least as equally important. Studies on the role of antibody-mediated immune reactions have focused on characterizing antibodies directed at epitopes associated with the GXM. Monoclonal antibodies (MAbs) to GXM were first reported by Dromer et al. (5) and Eckert and Kozel (7). In the period since these publications, anti-GXM antibodies have been shown to opsonize the yeast and enhance phagocytosis (23, 24, 28, 35, 36), affect cytokine synthesis and T-cell activity (33), and activate complement (17). Additionally, the protective and nonprotective effects of various anti-GXM MAbs in experimental infections have been extensively studied (32; reviewed in reference 2).

The humoral responses to noncapsular antigens have received much less attention. Hamilton et al. (8) have reported on the production of MAbs that recognized epitopes on cytoplasmic and secreted proteins with molecular masses of between 36 and 110 kDa. Additionally, Hamilton and coworkers reported that secreted antigens from acapsular (9) and encapsulated (10) strains were recognized by antibodies in patients’ sera. Chen et al. (3) described a number of proteins that were secreted from acapsular and encapsulated cryptococcal strains and reported that antibodies to some of these proteins were produced during experimental infections in mice and rats. Kakeya et al. (14) also recently reported on the occurrence of antibodies in infected mice to a heat shock protein secreted by C. neoformans. While these reports do demonstrate a humoral response to noncapsular antigens, a potential role for these antibodies in the immune defense against C. neoformans was not investigated.

In this study we report on a novel MAb and present evidence that it recognizes a cell-associated and secreted antigen unrelated to the major capsular polysaccharides. We additionally provide the first in vitro evidence of a possible immunologic role for such noncapsular antibodies, namely, opsonization and enhancement of yeast interactions with phagocytes.

(This work was presented in part at the General Meeting of the American Society for Microbiology, Atlanta, Ga., 1998.)

MATERIALS AND METHODS

Yeast strains and culture conditions.

The C. neoformans encapsulated clinical isolates, designated CSF-1 and BLD-1 (both serotype A), have been described previously (19–22). The acapsular strain, ATCC 52817, was purchased from the American Type Culture Collection (Manassas, Va.). This strain was originally described as Cap67 by Jacobson et al. (13). Yeasts were routinely grown at 25°C in yeast nitrogen base (YNB) (Difco Labs, Detroit, Mich.) with 0.5% (NH₄)₂SO₄, and 1.0% glucose. When radiometric adherence experiments were conducted, 2 μCi of l-[4,5-³H]leucine (140 Ci/mmol) (Amersham Pharmacia Biotech, Piscataway, N.J.) were added per ml. All media were sterilized by filtration. Yeast cell numbers were determined microscopically with a hemacytometer. Protein concentrations were determined by use of the BCA Protein Assay Reagent as described by the manufacturer (Pierce, Rockford, Ill.).

Hybridoma development, maintenance, and antibody preparation.

The procedures used were modified from those described previously (21). BALB/c mice were administered 5 weekly intraperitoneal injections of formalin-killed yeasts (strain CSF-1; 150 μg of whole-cell protein per injection). For the first injection, yeast suspensions were mixed with an equal volume of Freund complete adjuvant. For all subsequent injections, the yeast suspensions were mixed with equal volumes of Freund incomplete adjuvant. Hybridomas were produced by standard procedures modified from Kohler (16) and those previously described (21). Culture supernatants were screened for the presence of antibody recognizing cell surface epitopes by an immunofluorescence (IF) assay described previously (21). Cultures yielding positive results in this initial screening were cloned at 1 cell per well. Cultures that grew out were screened as described above, positive cultures were expanded, and culture supernatants were retained for antibody collection. The hybridomas were maintained in Dulbecco modified Eagle medium supplemented with 0.37% NaHCO₃, 200 U of penicillin per ml, and 200 μg of streptomycin per ml (hereafter referred to as DMEM) and also containing 4.5 mg of glucose per ml and 10% heat-treated fetal bovine serum (FBS) and then routinely subcultured every 3 to 4 days. The isotype and subclass of the MAb described here (designated CSFi MAb) was determined to be immunoglobulin G2b (IgG2b) by a mouse antibody typing kit (The Binding Site, San Diego, Calif.). Concentrations of the IgG MAb were measured with a mouse RID kit (The Binding Site).

The CSFi MAb failed to adhere to protein A resins and was therefore partially purified by first precipitating it from pooled hybridoma culture supernatants with 35% ammonium sulfate. Antibody was allowed to precipitate overnight at 5°C; the precipitate was then collected by centrifugation, dissolved in 25 mM HEPES–50 mM NaCl (HEPES-NaCl), and dialyzed against the same. The dialysate was applied to a Sephacryl S-300 (Amersham Pharmacia Biotech) column (2.6 by 98 cm) and then eluted with HEPES-NaCl at a flow rate of 15 ml/h. Fractions of 5-ml volumes were collected while the absorbance at 280 nm was monitored. Localization of the MAb in the eluted fractions was achieved by the IF assay described above. The fractions with the greatest activity were pooled and concentrated in Amicon Minicon concentrators (15,000 molecular weight cutoff; Amicon, Inc., Beverly, Mass.). The final solutions were filter sterilized, and the antibody concentration was determined by RID as described above.

Immune electron microscopy (IEM).

Yeasts were grown 48 h in YNB, washed with phosphate-buffered saline (PBS), and then fixed with 1% glutaraldehyde, followed by 1% osmium tetroxide, and processed as described previously (19). Sections were blocked with 1% bovine serum albumin, reacted with the CSFi MAb for 45 min, washed, and then reacted for 45 min with anti-mouse polyclonal antibody conjugated to 20-nm colloidal gold (E Y Laboratories, San Mateo, Calif.). After being washed, the sections were stained with uranyl acetate and lead citrate (Ladd Industries, Burlington, Vt.) and examined with an RCA electron microscope.

Phagocytic cells.

Mouse peritoneal exudate cells from nonelicited mice and from mice elicited with an intraperitoneal injection of 50 μg of concanavalin A (ConA) 24 h prior to harvest were prepared by peritoneal lavage. Mice were sacrificed, and 10 ml of cold DMEM, without FBS and with 10 U of heparin per ml, was injected into the peritoneal cavity; the flanks were massaged, and then the DMEM recovered by aspiration. The peritoneal cells were collected by centrifugation and suspended in DMEM containing 10% FBS, and then a viable cell count with trypan blue was conducted. The cells were seeded into 24-well plates (Costar, Cambridge, Mass.) containing 0.5 ml of DMEM with 10% FBS at a density of 10⁶ viable cells per well. In some experiments the wells contained Thermanox coverslips (Nalge Nunc, Rochester, N.Y.) to facilitate the microscopic counts (described below). The seeded plates were incubated at 37°C for 90 min, the media with nonadherent cells were removed, the monolayers were washed three times with DMEM without FBS, and DMEM with 10% FBS was added back. The cultures were then incubated overnight prior to conducting the adherence experiments.

Adherence experiments were also conducted with the mouse macrophage-like cell line, J774A.1 (ATCC TIB-67). These cells were grown in 24-well plates, with or without coverslips as described above, containing (per well) 0.5 ml of DMEM with 10% FBS.

Adherence assays.

The effects of the CSFi MAb on the interaction of the C. neoformans-encapsulated isolates with murine phagocytes was studied by use of two techniques. In one technique, phagocytes (both the peritoneal cells and J774A.1) that were grown on coverslips in well cultures were washed with Earle balanced salts (EBS) (Sigma Chemical Co., St. Louis, Mo.) containing 0.22% NaHCO₃, and fresh DMEM with 10% FBS was added back to the wells. Yeasts, grown in YNB at 25°C for 48 h, were washed with PBS and resuspended to the appropriate concentration in DMEM with 10% FBS. Each well culture was inoculated with 8.25 × 10⁵ yeasts, and the cultures were incubated for the indicated times at 37°C. After the incubation period, the cultures were washed with EBS and fixed with 10% formalin in PBS, and then the coverslips were removed, placed on a slide, and covered with buffer and another coverslip. The percentage of phagocytes with bound or ingested yeasts was then determined microscopically by counting the total number of phagocytes and the number of phagocytes with bound yeasts in five microscopic fields that contained 100 or more phagocytic cells. These values were averaged, and the percentage of phagocytes with bound yeast was calculated. The phagocytic index is defined here as the average total number of bound yeasts divided by the average total number of macrophages counted per microscopic field. This calculation for the phagocytic index was similar to that described by Mukherjee et al. (23, 24), except that we did not differentiate between yeasts that were bound and those that were actually internalized.

The second technique used to measure yeast adherence to the phagocytes was a radiometric assay (21). This technique allowed us to determine the percentage of yeasts used to inoculate the phagocyte cultures that had bound to the phagocytic cells and allowed us to further substantiate the activity of the CSFi MAb. Phagocytes, cultured in wells without coverslips, were processed as described above. Yeasts, grown at 25°C for 48 h in YNB containing [³H]leucine, were washed with PBS and suspended in DMEM with 10% FBS. Each well culture was inoculated with 8.25 × 10⁵, [³H]leucine-labeled yeasts, and the cultures were incubated for the indicated times at 37°C. After the incubation period, the cultures were washed to remove unbound yeasts, and the phagocytes were detached and lysed with 0.05% Triton X-100 in H₂O. This concentration of Triton had no effect on the integrity of the yeasts but completely lysed the phagocytes and liberated the yeasts that were both bound to the surface and internalized. The resulting suspensions were processed, and the counts per minute (cpm) were determined as described previously (21). The percentages of the yeast inoculum that had bound to the phagocyte cultures were calculated based on the cpm of the individual inocula, and these adherence values are presented as the average percent bound per well culture.

Secreted antigens.

Yeasts were grown 48 h in YNB, the yeasts were removed by centrifugation, and the culture medium was filtered (0.45 μm [pore size]), frozen, and lyophilized. The material remaining after lyophilization was suspended in distilled water and dialyzed (12,000 to 14,000 molecular weight cutoff) extensively against the same. The dialysate was frozen and lyophilized again, and the material remaining suspended in a minimal volume of PBS. This constituted the crude secreted antigen (SA). The 52817 SA preparation was filter sterilized and then subjected to gel filtration on a Sepharose CL-6B (Amersham Pharmacia Biotech) column (2.6 by 98 cm) at a flow rate of 24 ml/h with HEPES-NaCl as the mobile phase. Fractions of 5-ml volumes were collected, and the appropriate fractions were pooled, dialyzed against distilled water, and then frozen and lyophilized. The material remaining after lyophilization was suspended in a minimal amount of PBS and was designated as the CL-6B-A, -B, or -C fractions. Protein concentrations were determined by use of the BCA Protein Assay Reagent as described above. Carbohydrate concentrations were determined by the phenol-sulfuric acid assay (6) with glucose as the standard.

Gel electrophoresis and immunoblotting.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was conducted with precast, 4% stacking, 12% separating gels (Bio-Rad, Hercules, Calif.) by using modified Laemmli (18) buffers and conditions.

Components in the SDS-PAGE gels were electrophoretically transferred onto polyvinylidene difluoride membranes as previously described (30). The sheets were cut appropriately, and sections were stained for protein or carbohydrate as described below. After a soaking for 1 h in blocking buffer, the membranes were washed and then reacted for 1 h with the CSFi MAb at a concentration of 0.1 mg/ml in blocking buffer. This was followed by incubating the membranes for 1 h with peroxidase-conjugated anti-mouse polyclonal antibody (Sigma) diluted 1:2,000 in blocking buffer. After being washed, the blots were developed with the substrate-chromagen consisting of 0.48 mM 4-chloro-1-naphthol (Sigma), 17% methanol, and 0.01% H₂O₂. Membranes were stained for protein with 0.1% Coomassie brilliant blue R and for carbohydrate by the periodic acid-Schiff reagent (PAS; Sigma). Carbohydrate epitopes were removed from the membranes by the technique of Woodward et al. (34).

ELISA.

Wells of enzyme-linked immunosorbent assay (ELISA) plates (Costar, Cambridge, Mass.) were coated overnight at 5°C with SA in TBS (20 mM Tris, 500 mM NaCl [pH 7.5]) at a concentration of 1 μg/well. Wells were washed with TBS containing 0.05% Tween 20 and then blocked with TBS containing 1% BSA for 45 to 60 min at room temperature. After a washing, the CSFi MAb (final dilution, 1:500) was added with or without the various inhibitors described for the individual experiments. After a 45- to 60-min incubation at room temperature, the wells were washed and peroxidase-conjugated polyclonal anti-mouse antibody (diluted 1:2,000; Sigma) was added to the wells and incubated for 45 min at room temperature. The wells were washed again, and the substrate and chromagen was added (0.5 mg of O-phenylenediamine in 0.05 M sodium citrate buffer [pH 5.0], and H₂O₂ was added to 0.005%). After 30 min of incubation at room temperature, the reaction was stopped by the addition of 1N H₂SO₄ and the absorbance read at 490 nm.

Statistical analysis and reproducibility.

The Student’s t test (unpaired) was performed to assess statistical significance, and P values of <0.05 were considered to be significantly different. All experiments were repeated on separate days and yielded results similar to those reported here.

RESULTS

Binding characteristics of the CSFi MAb.

Initial immunofluorescent detection methods demonstrated that the MAb bound to antigenic determinants expressed on the surface of the immunizing antigen (CSF-1, Fig. 1a and b) and on the surface of the acapsular mutant 52817 (Fig. 1c and d). The binding pattern was patchy, and the MAb seemed to bind to some cells more than others. A number of different C. neoformans strains were tested, and similar results were obtained. The strains tested were ATCC 732 (serotype A), ATCC 6352 (serotype A), ATCC 28958 (serotype D), ATCC 32608 (serotype C), ATCC 56991 (serotype BC), and ATCC 64062 (serotype B) (results not shown).

FIG. 1.

Binding of the CSFi MAb to the surface of whole, fixed yeasts as demonstrated by IF microscopy. Phase-contrast (a and c) and IF (b and d) photomicrographs of the CSF-1 and 52817 strains, respectively, are shown.

Postsectioned IEM of the CSF-1 (Fig. 2a) and 52817 (Fig. 2b) strains demonstrated that the antigenic determinants recognized by the CSFi MAb appeared to be concentrated in the cell wall of both strains. Dense aggregates of the antigen were observed in the cell wall, at the cell wall-plasma membrane interface, and also within the cytoplasm (arrows).

FIG. 2.

IEM of sections of the CSF-1 (A) and 52817 (B) strains reacted with the CSFi MAb, followed by anti-mouse antibody conjugated with 20-nm colloidal gold. Magnification, ca. ×54,000 (A) and ca. ×28,000.

Effects of the MAb on yeast interactions with phagocytes.

Experiments were conducted with two different types of phagocytic cells to determine whether the CSFi MAb was an opsonin and could enhance yeast adherence. The mouse peritoneal exudate cells were judged to be predominantly phagocytes based upon their adherence to plastic and their morphology. Phagocytes prepared from ConA-elicited and nonelicited mice yielded similar results (data not shown). However, since more cells could be recovered from elicited mice, ConA-elicited phagocytes were used for all of the remaining experiments described here.

The CSFi MAb enhanced yeast adherence to the mouse phagocyte-like cell line, J774A.1 (Fig. 3). ³H-labeled yeasts were used to inoculate confluent well cultures of phagocytes containing the indicated amounts of the CSFi MAb. These cultures were then incubated for 2 h and washed, and then phagocytes and yeasts removed from the wells with Triton X-100. The cpm values of the well contents and that of the yeast inocula were determined and used to calculate the percentage of the yeast inoculum that bound to the phagocytes. For these experiments the yeast/phagocyte ratio was approximately 10 to 1. As shown, there was a dose-dependent relationship between the MAb concentration and the percentage of yeasts that bound. Both clinical isolates appeared to be equally opsonized and yielded nearly identical adherence values.

FIG. 3.

Concentration-dependent MAb-induced adherence of the CSF-1 and BLD-1 strains to J774A.1 monolayers. ³H-labeled yeasts were incubated with the phagocytes for 2 h before the percentage of the yeast inocula that had bound per well culture was determined. The results presented are averages of triplicate determinations, with bars indicating the standard deviations.

Table 1 summarizes experiments demonstrating the effects of the MAb on C. neoformans interactions with both types of phagocytic cells. As shown, 100 μg of the MAb significantly increased both measures of adherence and resulted in a 10-fold increase in the phagocytic index. Microscopic observations indicated that MAb opsonization also resulted in increased internalization; however, quantitation of these values will be the subject of a later study. It should be noted that a mouse IgA MAb that was partially purified by the same technique used to prepare the CSFi MAb had no effect on yeast adherence to either type of phagocytic cells.

TABLE 1.

The effects of the CSFi MAb on C. neoformans interactions with phagocytes^a

Cell type	% Phagocytes with bound/ingested yeasts (P)	Phagocytic index (P)	% Yeast inoculum bound (P)
J774A.1 macrophage-like cell line
Control	5.6 ± 1.8	0.1 ± 0.03	7.6 ± 0.6
100 μg of MAb	55.9 ± 2.8 (<0.0001)	1.6 ± 0.2 (<0.0001)	32.6 ± 0.8 (<0.0007)
Mouse peritoneal cells
Control	8.4 ± 2.7	0.16 ± 0.06	3.3 ± 1.2
100 μg of MAb	59.7 ± 2.8 (<0.0001)	1.6 ± 0.08 (<0.0001)	11.6 ± 0.2 (0.01)

^aYeasts (CSF-1 strain) were incubated with the phagocytic cells for 2 h. The results are presented as the averages of at least triplicate determinations ± the standard deviation. The P values are derived from Student’s t test comparing values in the absence of the CSFi MAb to those obtained in the presence of the MAb at a concentration of 100 μg/well. 

CL-6B fractionation and SDS-PAGE–Western analysis of the SAs.

Supernatants from CSF-1 and 52817 cultures were collected, dialyzed, and concentrated, yielding the CSF-1 SA and the 52817 SA. Because of the high viscosity of the CSF-1 SA, it was not subjected to further purification on the CL-6B column. The 52817 SA was applied to the column, and both the absorbance at 280 nm and the carbohydrate concentrations were monitored on the eluted fractions. Typical 52817 SA preparations had protein concentrations of approximately 0.8 mg/ml and carbohydrate concentrations of approximately 4 mg/ml. Fractionation of the 52817 SA resulted in three major peaks (data not shown). Based on the elution pattern of standards, the SA components ranged in size from approximately 150 to <12 kDa. Fractions corresponding to the three peaks were pooled (designated CL-6B-A, -B, and -C) and concentrated.

The 52817 SA and the CL-6B pooled fractions were analyzed by SDS-PAGE–Western immunoblotting, and Fig. 4 shows typical results. The MAb bound two bands from the SA and the CL-6B-A fractions (lanes B” and C”, respectively) that had molecular masses of approximately 75 and 100 kDa. The 100-kDa band was also observed on the Coomassie blue-stained membrane (lanes B and C). Both the 75- and 100-kDa bands reacted strongly when stained for carbohydrate with PAS (lanes B’ and C’). The CSFi MAb also reacted weakly with a 28-kDa band in the CL-6B-B fraction; however, binding to this band was difficult to demonstrate. A 28-kDa band also stained for protein (lane D) and carbohydrate (lane D’). The MAb did not react with any components of the CL-6B-C fraction and was not included in this figure. Western blots conducted with the CSF-1 SA or whole-cell extracts of CSF-1, BLD-1, and 52817 (extracted with SDS sample buffer) yielded results similar to those obtained with the 52817 SA (22). Treating the membranes with sodium periodate and sodium borohydride to remove carbohydrate epitopes resulted in no binding of the MAb to the treated membranes (results not shown).

FIG. 4.

SDS-PAGE–Western analysis of the 52817 SA and pooled CL-6B fractions. Molecular mass standards (represented as kilodaltons) are shown in lanes A, A’, and A”; the separation of components in the 52817 SA are shown in lanes B, B’, and B”; and the separation of components in the CL-6B-A and CL-6B-B fractions are shown in lanes C, C’, and C” and D, D’, and D”, respectively. The membranes were stained with either Coomassie blue (lanes A to D) or PAS (lanes A’ to D’) or were reacted with the CSFi MAb (lanes A” to D”).

ELISA.

A competitive ELISA was used to test for cross-reactivity between the CSFi MAb and a variety of purified GXMs (kindly provided by R. Cherniak). When the CSF SA was bound to the ELISA plate wells and the homologous antigen was used as the inhibitor, 90% inhibition of CSFi binding was obtained at 10 μg of SA protein/well (Fig. 5). This particular preparation of CSF-SA had a protein/carbohydrate ratio of 1:9.5; therefore, there was 95 μg of carbohydrate per well at the highest concentration tested, and 50% inhibition occurred at 9.5 μg of carbohydrate per well. GXMs from A and D serotypes were tested at up to 50 μg (dry weight). As can be seen, there was no notable inhibition by any of the GXM preparations. Similar results were obtained in competitive ELISAs, where the 52817 SA was used as the coating antigen and the homologous antigen or GXMs were used as competitors (22). For example, binding of the CSFi MAb to wells coated with the 52817 SA was nearly completely inhibited by 10 μg of either the 52817 SA or the CSF-1 SA, but 50 μg of either A or D serotype GXMs resulted in only 15 to 20% inhibition.

FIG. 5.

Competitive ELISA with the CSF SA and purified GXMs. ELISA plate wells were coated with the CSF SA, and the CSFi MAb was added without inhibitor or together with the homologous antigen or the GXMs shown. The percent inhibition was calculated by use of the following formula: inhibition = 100 − (average A₄₉₀ with inhibitor/average A₄₉₀ without inhibitor × 100).

DISCUSSION

We have developed a MAb that recognizes an antigenic determinant found in the yeast cell wall, on its surface, and also secreted into the culture medium. This is the first description of an opsonizing antibody that binds an antigenic determinant unrelated to the major capsular polysaccharide, GXM. On the other hand, many MAbs to the capsular GXM have been described that function as opsonins (17, 23, 24, 28, 33, 35, 36). In contrast, however, Houpt et al. (12) characterized antibodies in normal sera that reacted with GXM but did not contribute to opsonization. Additionally, Keller et al. (15) identified an anti-glucan antibody in human sera that also did not contribute to opsonization and enhance phagocytosis.

The pattern of binding of our CSFi MAb demonstrated by IF assay showed that the antigen was unevenly distributed over the surface of the encapsulated strains. Additionally, in any given population of yeasts tested, there were some yeast cells that didn’t react with the MAb, implying that perhaps expression of the antigen was related to the yeasts’ growth phase. Similar results were demonstrated with the acapsular mutant; however, those cells that were stained seemed to be more reactive with the MAb than the encapsulated strains. This implied that the capsule might either restrict the expression of the antigen or inhibit the binding of the MAb to its epitope.

IEM results showed that the antigen recognized by our MAb seemed to be concentrated in the cell wall and at the membrane-wall interface. The epitope was also observed in the cytoplasm. The organization and location of the antigen was similar in sections of both the encapsulated and acapsular strains. Todaro-Luck et al. (29) studied anti-GXM MAb; however, none of them reacted with the acapsular mutant, Cap67 (ATCC 52817), as did our MAb. Additionally, our IEM results were very different than those of Vartivarian et al. (31), who characterized the location of mannoprotein in the cell wall of C. neoformans. Therefore, the nature and distribution of this epitope appears unique in comparison to these well studied antigens.

All of the experiments with the phagocytic cells were conducted without activating the phagocytes with gamma interferon or lipopolysaccharide. Nevertheless, our results with the peritoneal phagocytes and the J774A.1 cell line demonstrated that the CSFi MAb significantly increased yeast adherence as measured by both adherence techniques. Additionally, the CSFi MAb increased the phagocytic index 10-fold. It is noteworthy that the J774A.1 cell line has been used extensively by other investigators studying the interactions of C. neoformans with phagocytic cells (23–25, 35). While we did not distinguish between bound and internalized yeasts for our phagocytic index calculation, the results that we obtained were similar to those reported by Mukherjee et al. (23) with anti-GXM MAb. However, whether our MAb enhances the antifungal activity of phagocytes, as do some anti-GXM antibodies, remains to be investigated.

We hypothesized that the antigen recognized by the CSFi MAb was secreted into the medium. Indeed, a plausible explanation for why an antibody directed at a noncapsular antigen might serve as an opsonin would be that, as the antigen is being secreted, it may be exposed for enough time to allow antibody binding while it is still associated with the capsule or surface of the cell. Using both the encapsulated clinical isolate and the acapsular mutant, we were able to demonstrate that the antigen was secreted and that it eluted from the CL-6B column in a broad peak with a molecular mass of approximately 150 kDa. SDS-PAGE–Western analysis confirmed that the CSFi-reactive epitope was present on antigens with molecular masses of approximately 75 and 100 kDa. The CSFi-reactive epitope was probably a carbohydrate, since treating the membranes with periodate and borohydride abrogated MAb binding. Indeed, the 100-kDa antigen appeared to be a glycoprotein, with bands visible on both the Coomassie blue-stained membrane and the PAS-stained membrane. A band corresponding to the 75-kDa band was not apparent on the Coomassie blue-stained membrane; however, because of the low sensitivity of this strain, a protein component cannot be ruled out. How these antigens might be related to the 28-kDa antigen that reacted weakly with the MAb is unknown. It is additionally not possible to determine from our results if the 75- and 100-kDa antigens are actually subunits of a larger native glycoprotein. Antibodies directed at secreted proteins with similar molecular masses have been described in the literature. For example, Hamilton and coworkers (8–10) have reported on antibodies that recognized 115- and 65-kDa antigens. Chen et al. (3) have described numerous extracellular proteins that were immunogenic, and antibodies were identified to 75- and 30-kDa proteins secreted from one of the strains they studied. Additionally, Kakeya et al. (14) identified antibodies in infected mice to a 77-kDa heat shock protein. However, an immunologic function, such as opsonization, was not reported to be associated with any of these previously described antibodies.

Competitive ELISAs clearly demonstrated that, because the GXMs did not block binding of the CSFi MAb to either SAs derived from the encapsulated strain or the acapsular mutant, the epitope that the MAb was reacting with was not associated with the capsular polysaccharide. The B and C serotypes were not tested because the immunizing antigen was an A serotype, and the fact that the CSFi MAb bound to the acapsular mutant made it unlikely that GXM epitopes were involved. Additionally, since the molecular masses of GalXM and mannoprotein fractions are much less than that of the antigen bound by our MAb (26), it seemed unlikely that the CSFi-reactive epitope was associated with either of these antigens.

While the exact nature of the antigen bound by this MAb remains to be determined, we have provided compelling evidence that a secreted antigen can be bound by antibody while still associated with the yeast. This antibody then can serve to opsonize the yeasts and enhance their interaction with phagocytes. Casadevall and Perfect have stated that the function of antibodies to secreted antigens is unknown (2), but our results provide evidence that may help delineate an immunologic function for these antibodies.