Abstract

Cryptococcus neoformans produces a life-threatening meningitis in patients who are immunocompromised by AIDS. A striking feature of cryptococcosis in AIDS is high serum levels of the major capsular polysaccharide, glucuronoxylomannan (GXM). Soluble GXM has numerous biologic activities that may contribute to the pathogenesis of infection. The objective of the study was to further understand in vivo processing of GXM. Mice were injected intravenously with GXM, and the tissue distribution was determined. A macrophage suicide technique that used liposome-encapsulated dichloromethylene diphosphonate determined the role of macrophages. GXM was cleared from serum with a half-life of 24–48 h but was retained for an indefinite period in tissues rich in cells of the mononuclear phagocyte system. Ablation of macrophages decreased GXM in the liver and spleen and increased serum GXM. The results identify a key role for macrophages in the clearance of GXM from serum and identify macrophages as a long-term reservoir for storage

Cryptococcus neoformans is an encapsulated yeast that may produce a life-threatening meningitis in immunocompromised persons, particularly those with AIDS. The major component of the polysaccharide capsule is glucuronoxylomannan (GXM) [1]. GXM has an α-(1→3)–linked d-mannopyranosyl backbone. Side chains of single xylopyranosyl and glucuronopyranosyl residues are attached to the mannan backbone. The polysaccharide is also O-acetylated. The extent of xylose substitution and O-acetylation varies with the serotype

GXM is essential to the virulence of the yeast. Strains deficient in capsule production are avirulent [2–4]. Biologic properties of the capsule and of soluble GXM that may contribute to the pathogenesis of cryptococcosis include inhibition of phagocytosis [5, 6], induction of immune unresponsiveness [7, 8], inhibition of neutrophil influx into potential sites of inflammation [9], induction of shedding of l-selectin from neutrophils [10], blockade of neutrophil CD18 [11], contribution to cerebral edema and increased intracranial pressure [12–14], alterations in cytokine secretion by leukocytes [15–18], and enhanced infectivity of human immunodeficiency virus [19, 20]

Assay of serum and cerebrospinal fluid (CSF) for free GXM by latex agglutination [21] or ELISA [22] is an important tool in the diagnosis of cryptococcosis. Extraordinarily high serum antigen titers are a feature of cryptococcosis in patients with AIDS [23]. In an extensive study of cryptococcosis done before the advent of AIDS, high initial titers of cryptococcal antigen in serum and CSF were found to be negative prognostic indicators [24]. There is less agreement as to the prognostic value of serum or CSF antigen levels in AIDS patients with cryptococcosis. Chuck and Sande [23] reported that high initial titers of antigen in serum or CSF are not predictive of shorter survival in AIDS patients with cryptococcosis. In contrast, Zuger et al. [25] reported a significant correlation between initial serum and CSF antigen titers and mortality rate. Similarly, Saag et al. [26] found that high CSF antigen titers were predictors of death during therapy. Levels of serum antigen appear to be higher in AIDS patients with cryptococcosis than in non-AIDS patients with cryptococcosis [23, 24, 27, 28]. Eng et al. [28] suggested that patients with AIDS may have a defect in antigen elimination

Because of the importance of GXM as a virulence factor and the assay of body fluids for GXM in the diagnosis of cryptococcosis, there have been several studies of the clearance of GXM in animal models. A serum half-life of 14–48 h has been reported in studies of GXM clearance in mice, rats, and rabbits, with a half-life of ∼24 h being described most frequently [29–33]. A similar half-life was observed over challenge doses of 20 μg to 1 mg of GXM. Serum clearance appears to be similar in cryptococcosis; Eng et al. [32] reported that serum GXM in patients who respond to treatment with amphotericin B has a half-life of 48 h. In studies that examined the tissue distribution of GXM, clearance of GXM from serum was accompanied by the accumulation of GXM in tissues such as the liver and spleen, which are rich in cells of the mononuclear phagocyte system [29, 30, 33]. Immunohistochemistry showed that GXM is localized in the Kupffer cells of the liver and in the marginal zone macrophages of the spleen [30]

In the present study, we examined kinetic models for the elimination of GXM from serum, tissues, and the whole body and explored the role of tissue macrophages in the clearance of GXM. Our studies took advantage of a procedure described by Van Rooijen et al. [34, 35], in which liposome-encapsulated dichloromethylene diphosphonate (Cl2MDP) is a suicide vehicle that eliminates macrophages from the liver and spleen in mice but does not affect neutrophil levels

Materials and Methods

Capsular polysaccharideGXM from serotype A strain 24064 (American Type Culture Collection) was used throughout the study. GXM was isolated from culture supernatant fluids, as described elsewhere [36]

Monoclonal antibodies (MAbs)MAbs 439 and 3C2 were used. Both antibodies are IgG1 and are reactive with an epitope shared by C. neoformans serotypes A, B, C, and D. The characteristics of these antibodies and the manner by which they are produced and purified have been described elsewhere [37, 38]

MiceWe used 6–8-week-old female BALB/c mice (Charles River Laboratories) for all experiments

In vivo tissue distributionLyophilized GXM was hydrated in PBS (0.01 M phosphate and 127 mM saline [pH 7.3]) and was filter-sterilized before use. Mice were injected intravenously with various amounts of GXM in a 0.2-mL injection volume. At selected time points, mice were anesthetized with methoxyflurane (Pitman-Moore), and blood was collected via cardiac puncture. Blood samples were allowed to clot, and serum samples were frozen. Immediately after collection of blood, the mice were killed by cervical dislocation while under anesthesia. A sample of bile was obtained by gallbladder puncture, and fecal and urine samples were collected. The spleen, liver, kidneys, lungs, and brain were removed, were washed briefly with PBS, were weighed, were placed in ∼10 volumes (vol/wt) of PBS, and were stored at −20°C. Frozen tissue samples were thawed and homogenized in PBS with a 10-mm generator (Omni International) powered by an Omni 2000 variable speed motor. Tissues were homogenized for 12 s at 40% of maximum rotations per minute and were stored at −20°C until GXM content was determined by ELISA

ELISA for measurement of serum and tissue GXMWells of Immulon 1 microtiter plates (Dynatech Laboratories) were coated overnight at room temperature with 0.2 μg of anti-GXM MAb 439 in 100 μL of coating buffer (0.05 M sodium phosphate [pH 7.4] containing 10 mM EDTA). After coating, the wells were washed 3 times with 0.05 M sodium phosphate (pH 7.4) and were blocked by incubation for 90 min at room temperature with PBS containing 0.05% Tween 20 (PBS-Tween). After blocking, the plates were washed with PBS-Tween and were incubated for 90 min with serial 2-fold dilutions of homogenized tissue samples in PBS-Tween. Each 96-well microtiter plate included dilutions of purified GXM as a standard. By using the standard GXM preparations, 10 ng/mL GXM could be measured reliably. Standard curves prepared by using GXM added to serum or homogenized tissue (liver) were identical to curves prepared using GXM in PBS. Taking into account the initial dilution of the homogenized tissue with PBS, the limit of detection for GXM was ∼100 ng/g tissue. After incubation with the tissue samples or standards, the wells were washed 3 times with PBS-Tween and were incubated for 30 min with horseradish peroxidase (HRPO)–labeled indicator antibodies (anti-GXM MAbs 3C2 or 439). The choice of indicator antibody was determined by the availability of labeled antibody at the time that each experiment was done. MAbs 3C2 and 439 produced identical results when used in the indicator phase of the assay

The MAbs were conjugated to HRPO with a peroxidase labeling kit (EZ-Link Activated Peroxidase Kit; Pierce), according to the manufacturer’s directions. After incubation with the indicator antibody, the wells were washed 3 times with PBS-Tween and were incubated for 30 min with tetramethyl benzidine peroxidase substrate solution (Kirkegaard & Perry Laboratories). The reaction was stopped by the addition of 1 M H3PO4, and the absorbance was read within 30 min on a Ceres 900 EIA workstation (Bio-Tek Instruments)

ELISA data were analyzed by Kineticalc EIA application software (version 2.12; Bio-Tek Instruments), using a 4-parameter algorithm. GXM content for each sample was determined by comparison with standard curves generated with purified GXM

The GXM content in each organ was calculated by subtracting the GXM present in the plasma of each organ from the total GXM measured in each organ. With the exception of brain tissue, the plasma content of each organ was based on a published report of the plasma content in various murine organs [39]. Correction for plasma content in brain was done by using values reported for rat tissue [40]. The total body GXM was calculated as the sum of GXM in the serum, liver, spleen, kidney, lung, and brain. Calculation of serum volume assumed a blood volume of 5.77 mL/100 g of body weight [41]. In the macrophage ablation experiments, the lung and brain were not included in the calculation of total body GXM because, in initial studies of GXM clearance, GXM in the lungs and brains of GXM-injected mice did not contribute appreciably to the estimate of total body GXM

We used 4 mathematical models to analyze clearance of GXM from serum, total body clearance, and the accumulation and clearance of GXM in individual tissues. Clearance of GXM from serum was modeled by use of a simple 2-parameter equation for exponential decay:  

formula
where a is the Y intercept and b is the rate constant for clearance (exponential decay). Total body clearance of GXM was assessed by use of 3- and 4-parameter equations for exponential decay. In the 3-parameter exponential decay equation,  
formula
where y0 is the lower limit for total body GXM, a is the Y intercept, and b is the rate constant for clearance. In the 4-parameter exponential decay equation,  
formula
where a is the Y intercept for the initial rapid clearance step, b is the rate constant for early clearance, c is the Y intercept for the slower clearance step, and d is the rate constant for the slower clearance step. Finally, we derived an expression to model both accumulation (exponential rise to maximum) and clearance (exponential decay) in individual tissues:  
formula
where a is the Y intercept, b is the rate constant for accumulation, and c is the rate constant for clearance

Curve fits of serum, tissue, and total body GXM were calculated by use of SigmaPlot (SPSS), and the adjusted r2 (r2adj), a measure of how well the regression model describes the data, was calculated from the curve fit reports. The half-lives for GXM in serum and total body also were calculated from the curve fit reports. Statistical analysis of data where a pairwise comparison was made between a control group and a single treatment group (macrophage ablation experiment) was done by Student’s t test calculated with SigmaStat (version 2.03; SPSS). Analyses in which several treatment groups were compared with a single control (the effects of various doses of anti-GXM MAbs on tissue clearance) were determined by analysis of variance with pairwise comparisons done by the Tukey procedure. In a limited number of instances, the data failed a test for equality of variance. In such cases, the data were analyzed by the Kruskal-Wallis test, and pairwise multiple comparisons were done by the Dunn method

Macrophage ablationInjection of mice with liposome-encapsulated Cl2MDP efficiently depletes splenic and hepatic macrophage populations [35] but does not affect neutrophil function [34]. Mice were injected intravenously with 0.2 mL of liposome-encapsulated 12–15 mM Cl2MDP or PBS as a control, as described elsewhere [34, 42]. Then, 72 h later, each mouse was injected intravenously with 250 μg of GXM. Eight days after administration of GXM, the mice were killed, serum and other tissues were collected, and the amount of GXM was determined as described above

ImmunohistochemistryLivers, spleens, and kidneys from Cl2MDP-treated and control mice were collected, were washed briefly in PBS, and were fixed for 4 h in 10% buffered formalin (Fisher Scientific). The organs were dehydrated in ethanol and xylene, were embedded in paraffin, and were sectioned. Immediately before immunostaining for GXM, tissues were deparaffinized in a xylene and graded alcohol series and were hydrated in water. Nuclei were stained with Gill 1 hematoxylin stain (Fisher Scientific). Stained sections were washed with water and were placed in 0.1% NaHCO3 blue-up solution for 2 min at room temperature

GXM in tissues was localized with Texas Red (TR)–labeled anti-GXM MAb 3C2. MAb was conjugated to TR (Molecular Probes), according to the manufacturer’s directions. Each tissue section was incubated with TR-labeled MAb (5 μg/mL PBS) for 90 min at room temperature. After staining, the tissues were rinsed with PBS and were covered with VECTASHIELD (Vector Laboratories) and a coverslip. The pattern of GXM deposition in the tissues was determined by epifluorescence microscopy (Nikon Eclipse E800 microscope). Both brightfield and fluorescence images were captured with an integrating charge-coupled device camera (Photonic Science) controlled by Image Pro Plus analysis software (Media Cybernetics). The fluorescence image was deconvolved with Micro-Tome (version 4.0; Vay-Tek), and the brightfield and fluorescence images were overlaid with Image Pro Plus software

Results

In vivo distribution and long-term clearance of GXMAn initial experiment evaluated the tissue distribution of GXM over a 32-day experimental period. Mice were injected intravenously with 62, 250, or 1000 μg of GXM. These doses were chosen to approximate the wide range of serum GXM levels that represent significant antigenemia in human cryptococcosis [23, 28]. After administration of GXM, mice were killed at 1, 2, 4, 8, 16, and 32 days, for the collection of serum, bile, urine, feces, and liver, spleen, kidney, lung, and brain tissue. Results are shown in figure 1

Figure 1

Kinetics for the clearance of glucuronoxylomannan (GXM) from serum, liver, spleen, and kidney after intravenous injection of 1000, 250, or 62 μg of GXM. Data are mean±SD for 6 mice at each time point. t1/2, Half-life

Figure 1

Kinetics for the clearance of glucuronoxylomannan (GXM) from serum, liver, spleen, and kidney after intravenous injection of 1000, 250, or 62 μg of GXM. Data are mean±SD for 6 mice at each time point. t1/2, Half-life

Analysis of serum clearance by the 2-parameter equation (equation 1) for exponential decay showed an excellent fit for all 3 doses (r2adj>.98), with a half-life (t1/2) that did not differ appreciably with GXM dose. Extrapolation of total GXM in serum to t=0 days showed an amount of GXM that approximated the amount injected intravenously when the injection dose was 250 or 1000 μg. Although the extrapolated amount of GXM in serum differed by 35% from the amount injected in the 62-μg GXM treatment dose, the similarity between the amount injected and the calculated amount based on the clearance curve suggests that exponential decay is an appropriate model for GXM clearance and further suggests that the use of antigen capture ELISA accurately measures serum GXM levels

In agreement with previous studies, we found appreciable accumulation of GXM in the liver, spleen, and kidney. Accurate assessment of GXM tissue levels required correction for GXM found in tissue serum. In most cases, the fit of the data to the growth and decay curve (equation 4) produced r2adj>.85. The fit of the data to the model was less satisfactory for the spleens of mice given 1000 μg (r2adj=.49) and 250 μg (r2adj=.80) of GXM. These lower r2adj values likely reflect a poor data point (16 days) in the clearance curve for mice given 1000 μg of GXM and the limited number of data points during the exponential rise portion of the curve

GXM was found only in trace amounts at the limit of assay sensitivity in lung and brain tissue after the results were corrected to reflect serum GXM (data not shown). GXM was found sporadically and in variable amounts in feces, urine, and bile; the amounts detected were most often at the lower limit of detection by antigen capture ELISA

Plots of total body GXM were generated by calculating the sums of serum, liver, spleen, kidney, lung, and brain GXM levels for each mouse (figure 2). According to the 3-parameter model (equation 2), GXM was cleared exponentially until the level reached a stable limit below which there was little or no additional elimination. The fit was very good for GXM doses of 1000 μg (r2adj=.88) and 250 μg (r2adj=.96). No appreciable improvement was accomplished by the use of a 4-parameter equation (equation 3) that included an additional parameter representing a second slow elimination rate. These results indicate a relatively rapid clearance of GXM from the body (t1/2, 1.1–2.7 days), followed by persistence of a readily measurable percentage of the total injected GXM for many days. Once past the exponential clearance, there was little or no further clearance of GXM. Extrapolation of the curve to the Y-axis showed an amount of total body GXM that closely approximated the amount of injected GXM. Again, the similarity between the actual amount of injected GXM and the amount calculated from the clearance model suggests that the model is correct and that the procedures used to measure total body GXM are appropriate

Figure 2

Kinetics for the clearance of total body glucuronoxylomannan (GXM) after intravenous injection of 1000, 250, or 62 μg of GXM. Total body GXM was calculated as the sum of GXM in serum and in liver, spleen, kidney, lung, and brain tissue. Data are mean±SD for 6 mice at each time interval. t1/2, Half-life

Figure 2

Kinetics for the clearance of total body glucuronoxylomannan (GXM) after intravenous injection of 1000, 250, or 62 μg of GXM. Total body GXM was calculated as the sum of GXM in serum and in liver, spleen, kidney, lung, and brain tissue. Data are mean±SD for 6 mice at each time interval. t1/2, Half-life

Modeling of total body GXM in mice given 62 μg of GXM produced a relatively poor fit with the 3-parameter equation (r2adj=0.38) and a Y intercept that was less than half the amount of injected GXM. Such a discrepancy suggests a deficiency in either the model or the experimental design. The most likely deficiency is the need for additional data points at times <24 h that would allow for a more accurate extrapolation to t0

Role of tissue macrophagesPrevious studies of tissue localization of GXM found significant amounts of GXM in cells of the mononuclear phagocyte system [29, 30, 33]. One means to study the role of macrophages in biologic processes in vivo is the use of the liposome-mediated macrophage suicide technique. In one example of this procedure, experimental animals are injected intravenously with liposome-encapsulated Cl2MDP [34, 35, 42]. Macrophages ingest the liposomes, causing the release of the Cl2MDP and destruction of the cell. In an initial experiment, we evaluated the effect of treating mice intravenously with liposomes that contain either PBS or Cl2MDP. The mice were injected intravenously with 250 μg of GXM 72 h after treatment with the liposomes. Eight days later, tissues were collected, were fixed, and were processed for staining

Mice treated with liposome-encapsulated PBS (control mice) showed GXM localized to the hepatic sinusoids (figure 3). This pattern of staining is consistent with the presence of GXM in Kupffer cells, the hepatic component of the mononuclear phagocyte system. In contrast, livers of mice treated with liposome-encapsulated Cl2MDP showed markedly reduced staining around the sinusoids, with increased staining in the central vein, which suggests the presence of serum GXM. The control spleens had prominent staining of marginal zone macrophages of both white and red pulp. This localization is consistent with resident populations of macrophages in the spleen. The spleens from macrophage-ablated mice showed diminished GXM staining, particularly from marginal zone macrophages. Finally, staining for GXM in the kidneys of both control and macrophage-ablated mice was consistent with the presence of GXM in the capillary networks in the renal medulla. The absence of differences in the localization of GXM in the kidneys from control and macrophage-ablated mice suggests that localization to kidneys occurs via a mechanism that is independent of macrophages

Figure 3

Effect of ablation of macrophages on the localization of glucuronoxylomannan (GXM) in the liver, spleen, and kidney. Mice were injected with liposomes containing either PBS (control) or liposome-encapsulated dichloromethylene diphosphonate (L-Cl2MDP), followed 72 h later by intravenous injection of 250 μg of GXM. Mice were killed, and tissues were collected 8 days after injection of GXM. Tissues were lightly stained with Gill 1 hematoxylin, and GXM was localized with a Texas Red–labeled monoclonal antibody specific for GXM. Original magnification, ×400

Figure 3

Effect of ablation of macrophages on the localization of glucuronoxylomannan (GXM) in the liver, spleen, and kidney. Mice were injected with liposomes containing either PBS (control) or liposome-encapsulated dichloromethylene diphosphonate (L-Cl2MDP), followed 72 h later by intravenous injection of 250 μg of GXM. Mice were killed, and tissues were collected 8 days after injection of GXM. Tissues were lightly stained with Gill 1 hematoxylin, and GXM was localized with a Texas Red–labeled monoclonal antibody specific for GXM. Original magnification, ×400

The effect of macrophage ablation on the distribution of GXM between tissues was determined. Mice were treated with PBS- or Cl2MDP-encapsulated liposomes and were injected with 250 μg of GXM, as described above. Eight days later, mice were killed; serum, livers, spleens, and kidneys were collected; and the amount of GXM in each tissue was determined by antigen-capture ELISA. The GXM content of lungs and brains was not assessed, because previous studies (figure 2) found that these latter tissues did not contribute appreciably to total body GXM

Results of an analysis of GXM in tissues from control and macrophage-ablated mice (figure 4) showed that GXM levels dropped to almost 0 in the livers of macrophage-ablated mice (P⩽.001, control vs. macrophage-ablated mice). The GXM levels in the spleens of macrophage-ablated mice were less than half those of control mice (P<.001). In contrast, serum GXM increased in the macrophage-ablated mice 5-fold over levels in control mice (P=.002). There were no differences in the levels of GXM in the kidneys of control and macrophage-ablated mice, which confirms the suggestion (figure 3) that GXM accumulation in kidney is macrophage independent. The increase in serum GXM in macrophage-ablated mice identifies an essential role for macrophages in the clearance of GXM from serum. However, calculation of whole body GXM levels produced the surprising result that total GXM levels were 5 times less in macrophage-ablated mice than in control mice (P<.001). These results indicate that localization of GXM in macrophages prevents eventual clearance of GXM from the body

Figure 4

Effect of the ablation of macrophages in the lung and spleen on glucuronoxylomannan (GXM) levels in serum, liver, spleen, kidney, and the total body. Total body GXM was calculated as the sum of GXM found in the serum, liver, spleen, and kidney. Mice were injected with liposomes containing either PBS or dichloromethylene diphosphonate (Cl2MDP) and were injected intravenously 72 h later with 250 μg of GXM. GXM levels were assessed 8 days after GXM injection. Data are mean±SD for 8 mice

Figure 4

Effect of the ablation of macrophages in the lung and spleen on glucuronoxylomannan (GXM) levels in serum, liver, spleen, kidney, and the total body. Total body GXM was calculated as the sum of GXM found in the serum, liver, spleen, and kidney. Mice were injected with liposomes containing either PBS or dichloromethylene diphosphonate (Cl2MDP) and were injected intravenously 72 h later with 250 μg of GXM. GXM levels were assessed 8 days after GXM injection. Data are mean±SD for 8 mice

Effect of anti-GXM MAbsPrevious studies showed that passive immunization with anti-GXM MAbs reduces the level of GXM in serum and increases the levels in the liver and spleen [29]. Given the apparent role of macrophages in sequestering GXM and preventing clearance, we examined the effect of passive immunization with an anti-GXM MAb on both tissue distribution of GXM and on whole body clearance. Mice were injected intravenously with 250 μg of GXM. After 48 h, the mice were injected intraperitoneally with 31, 125, or 500 μg of MAb 439. Mice were killed, and tissues were collected 48 or 96 h after treatment with the MAb. Although there was a trend toward a dose-response effect, significant (P<.05) results were observed only with the highest antibody dose (figure 5). There was a significant reduction in serum GXM at both 48 and 96 h in mice treated with antibody. A significant reduction in liver GXM was observed at 96 h but not at 48 h. A significant increase in spleen GXM was observed at both sample times. No effect was observed in kidneys. Finally, there was only a slight reduction in total body GXM that was found 96 h after treatment with antibody

Figure 5

Effect of passive immunization with an anti-glucuronoxylomannan (GXM) monoclonal antibody (MAb) on tissue distribution of GXM. Mice were injected intravenously with 250 μg of GXM. After 48 h, mice were injected intraperitoneally with PBS or with 31, 125, or 500 μg of MAb 439. Mice were killed, tissues were collected 48 or 96 h after treatment with the MAb, and GXM levels were determined. Ten mice were used for the PBS controls, and 5 mice per group were used for each MAb treatment group. Data are mean±SD. *P<.05, treatment groups vs. controls

Figure 5

Effect of passive immunization with an anti-glucuronoxylomannan (GXM) monoclonal antibody (MAb) on tissue distribution of GXM. Mice were injected intravenously with 250 μg of GXM. After 48 h, mice were injected intraperitoneally with PBS or with 31, 125, or 500 μg of MAb 439. Mice were killed, tissues were collected 48 or 96 h after treatment with the MAb, and GXM levels were determined. Ten mice were used for the PBS controls, and 5 mice per group were used for each MAb treatment group. Data are mean±SD. *P<.05, treatment groups vs. controls

Discussion

In vivo clearance of GXM has been studied previously [7, 29–33] in mouse, rat, and rabbit models. Our results confirm several previous observations: that serum GXM has a half-life of ∼1–2 days [29–33], that it accumulates in tissues that are rich in cells of the mononuclear phagocyte system and is retained in such tissues for an extended time [29, 30, 33], and that passive immunization with anti–GXM MAbs accelerates clearance of GXM from serum and enhances sequestration in spleen [29, 30]. Our studies contribute to the further understanding of the in vivo fate of GXM by demonstrating that, once past a phase of exponential elimination, there is little or no further clearance of residual GXM, macrophages facilitate clearance of GXM from serum but serve as a reservoir for retention of GXM in the body, and that, despite the transfer of GXM from one body compartment to another, passive immunization does not greatly enhance overall clearance from the body. One caveat should be noted: our studies examined the clearance of GXM in otherwise normal mice. It is possible that infection of the liver and spleen by C. neoformans could alter the clearance function

A previous report from our laboratory noted accumulation of GXM in the tubular epithelium of the kidney [7]. Although GXM was found in the kidney in the present study, accumulation in the tubular epithelium was not a prominent feature of our results. There was an important difference in the preparations of GXM used in the 2 studies. Our earlier study used polysaccharide prepared by precipitation with ethanol and deproteinization by the method of Sevag, as modified by Heidelberger et al. [43]. This latter step involved mixing polysaccharide with chloroform and butanol at high speed in a blender. As a consequence, the preparation would have been contaminated with cryptococcal galactoxylomannan (GalXM) [44], and the polysaccharide would have been sheared by the deproteinization procedure. Polysaccharide used in the present study was prepared by differential precipitation with ethanol and Cetavlon [44]. Thus, GXM in the present study would be free of GalXM and would have a higher molecular weight than the polysaccharide used previously. Since the earlier report relied on immunoassay by using polyclonal antibodies raised against polysaccharide that likely contained both GXM and GalXM, the antigen found in tubular epithelium may have been GalXM rather than GXM. Alternatively, shearing of GXM to a smaller size may have increased filtration in the kidney. Additional variables that might account for differences between results in the 2 studies are differences in the mouse strain or possible strain-specific differences in GXM

Little is known about the mechanism by which GXM might bind to macrophages or the fate of GXM that is bound. GXM might bind via specific interaction with macrophage surface receptors or by nonspecific interactions. Lendvai et al. [29] found that nonlabeled capsular polysaccharide competed with 125I-labeled polysaccharide for organ sequestration, which suggests the involvement of specific cellular receptors. One candidate receptor is CD18. Dong and Murphy [11] found that MAbs specific for CD18 blocked the binding of 14C-labeled CneF, a concentrated culture filtrate of C. neoformans to neutrophils. More recently, Shoham et al. [45] reported that GXM binds to CHO cells transfected with human toll-like receptors 2 and 4 and/or CD14, which suggests that there are multiple sites on macrophages that might bind GXM. The macrophage mannose receptor is an unlikely candidate for binding GXM because little, if any, of the mannose in GXM is likely to be displayed as the terminal mannose needed for recognition by the mannose receptor. The absence of terminal mannose residues is suggested by the failure of concanavalin A to agglutinate encapsulated cryptococci [46]. Once bound to macrophages, GXM could remain at the cell surface or be internalized via receptor-mediated endocytosis, absorptive pinocytosis, phagocytosis, or diffusion. If internalized, neither the site of localization nor the signaling associated with internalization are known. Some signaling appears to occur because GXM induces secretion of interleukin-6 and -10 by human monocytes [15, 18]

Ablation of tissue macrophages by the liposome suicide technique produced an increase in serum GXM but also reduced the overall level of total body GXM 8 days after the injection of GXM. These results suggest a dual function for macrophages. Macrophages facilitate clearance of GXM from serum but also serve as a reservoir for long-term retention of GXM. The fate of GXM that is not cleared by cells of the mononuclear phagocyte system is not known, and there are several possibilities. GXM may not be cleared at the rate indicated in the present and prior studies. For example, GXM may be degraded such that the polysaccharide is not detected by antigen capture ELISA. The long-term persistence of GXM in macrophages suggests this is not the case. In addition, Lendvai et al. [29] observed a similar serum half-life for GXM by use of antigen capture ELISA and clearance of radiolabeled GXM

One potential route for the excretion of GXM is via urine. Jones et al. [47] reported that >89% of injected pneumococcal type III polysaccharide is secreted in the urine during the first 24 h after injection. Our results showed low levels of GXM in urine at the time mice were killed, but our experimental design did not include collection of total urine over the course of the experiments, so we cannot estimate the efficiency of renal excretion of GXM. GXM has been found in urine in some experimental models of GXM clearance [33] but not in others [31, 32]. GXM also is found in the urine of patients with cryptococcosis; however, titers are usually low [32, 48]. Eng et al. [32] calculated that levels of GXM in the urine of patients with cryptococcosis are too low to account for renal elimination of cryptococcal antigen from serum

An alternative for clearance is transfer via the hepatobiliary route to the gastrointestinal system and degradation or excretion of GXM in feces. Consistent with this mechanism, we detected low amounts of GXM in the contents of the gallbladder and in feces. GXM was noted in bile fluid in one previous study of GXM clearance [33] but not in another [29]. Support for hepatobiliary transport is provided by Russell et al. [49], who found antigen in the bile fluid of mice after intravenous injection of pneumococcal type III capsular polysaccharide. Our experimental design did not include an assessment of the total GXM in feces, so we can provide no further data in this regard. However, it is possible that GXM is degraded by glycosidases produced by bacteria in the fecal flora. Gadebusch [50] reported production of GXM degrading enzymes by an Alcaligenes isolate. Consistent with this hypothesis, Muchmore et al. [33] found GXM in the cecal contents of mice injected with GXM but much lesser amounts in feces, which suggests possible degradation by fecal bacteria

Serum GXM levels in AIDS patients with cryptococcosis are considerably higher than those in non-AIDS patients with cryptococcosis [23, 24, 27, 28]. In addition, serum antigen levels do not decline as rapidly in response to antifungal therapy as serum antigen levels in non-AIDS patients with cryptococcosis [28]. In contrast, CSF levels in AIDS and non-AIDS patients are similar, and levels of GXM in the CSF of AIDS patients decline in a predictable manner in response to effective antifungal therapy [28]. It is possible that the amount of GXM produced during cryptococcosis in AIDS patients is much greater than the amount produced in non-AIDS patients and that this high level of antigenemia accounts for the abnormally high levels of serum GXM. However, Eng et al. [28] argued that the high levels of serum antigen in AIDS patients with cryptococcosis are due to a severe defect in antigen elimination. Given the importance of macrophages in the reduction of serum GXM levels (figure 4), our results raise the possibility that a deficiency in macrophage clearance of GXM might account for defective elimination of GXM in AIDS patients with cryptococcosis. If so, an alternative means to enhance GXM sequestration, such as passive immunization with anti–GXM MAbs [29, 30], might be an important supplement to standard antifungal therapy

References

1
Cherniak
R
Sundstrom
JB
Polysaccharide antigens of the capsule of Cryptococcus neoformans
Infect Immun
 , vol. 
62
 (pg. 
1507
-
12
)
2
Bulmer
GS
Sans
MD
Gunn
CM
Cryptococcus neoformans. I. Nonencapsulated mutants
J Bacteriol
 , vol. 
94
 (pg. 
1475
-
9
)
3
Fromtling
RA
Shadomy
HJ
Jacobson
ES
Decreased virulence in stable, acapsular mutants of Cryptococcus neoformans
Mycopathologia
 , vol. 
79
 (pg. 
23
-
9
)
4
Kwon-Chung
KJ
Rhodes
JC
Encapsulation and melanin formation as indicators of virulence in Cryptococcus neoformans
Infect Immun
 , vol. 
51
 (pg. 
218
-
23
)
5
Bulmer
GS
Sans
MD
Cryptococcus neoformans. III. Inhibition of phagocytosis
J Bacteriol
 , vol. 
95
 (pg. 
5
-
8
)
6
Kozel
TR
Mastroianni
RP
Inhibition of phagocytosis by cryptococcal polysaccharide: dissociation of the attachment and ingestion phases of phagocytosis
Infect Immun
 , vol. 
14
 (pg. 
62
-
7
)
7
Kozel
TR
Gulley
WF
Cazin
J
Jr
Immune response to Cryptococcus neoformans soluble polysaccharide: immunological unresponsiveness
Infect Immun
 , vol. 
18
 (pg. 
701
-
7
)
8
Murphy
JW
Cozad
GC
Immunological unresponsiveness induced by cryptococcal capsular polysaccharide assayed by the hemolytic plaque technique
Infect Immun
 , vol. 
5
 (pg. 
896
-
901
)
9
Dong
ZM
Murphy
JW
Intravascular cryptococcal culture filtrate (CneF) and its major component, glucuronoxylomannan, are potent inhibitors of leukocyte accumulation
Infect Immun
 , vol. 
63
 (pg. 
770
-
8
)
10
Dong
ZM
Murphy
JW
Cryptococcal polysaccharides induce l-selectin shedding and tumor necrosis factor receptor loss from the surface of human neutrophils
J Clin Invest
 , vol. 
97
 (pg. 
689
-
98
)
11
Dong
ZM
Murphy
JW
Cryptococcal polysaccharides bind to CD18 on human neutrophils
Infect Immun
 , vol. 
65
 (pg. 
557
-
63
)
12
Denning
DW
Armstrong
RW
Lewis
BH
Stevens
DA
Elevated cerebrospinal fluid pressures in patients with cryptococcal meningitis and acquired immunodeficiency syndrome
Am J Med
 , vol. 
91
 (pg. 
267
-
72
)
13
Hirano
A
Zimmerman
HM
Levine
S
The fine structure of cerebral fluid accumulation. III. Extracellular spread of cryptococcal polysaccharides in the acute stage
Am J Pathol
 , vol. 
45
 (pg. 
1
-
19
)
14
Hirano
A
Zimmerman
HM
Levine
S
The fine structure of cerebral fluid accumulation. IV. On the nature and origin of extracellular fluids following cryptococcal polysaccharide implantation
Am J Clin Pathol
 , vol. 
45
 (pg. 
195
-
207
)
15
Delfino
D
Cianci
L
Lupis
E
, et al.  . 
Interleukin-6 production of human monocytes stimulated with Cryptococcus neoformans components
Infect Immun
 , vol. 
65
 (pg. 
2454
-
6
)
16
Retini
C
Vecchiarelli
A
Monari
C
Tascini
C
Bistoni
F
Capsular polysaccharide of Cryptococcus neoformans induces proinflammatory cytokine release by human neutrophils
Infect Immun
 , vol. 
64
 (pg. 
2897
-
903
)
17
Vecchiarelli
A
Retini
C
Pietrella
D
, et al.  . 
Downregulation by cryptococcal polysaccharide of tumor necrosis factor α and interleukin-1β secretion from human monocytes
Infect Immun
 , vol. 
63
 (pg. 
2919
-
23
)
18
Vecchiarelli
A
Retini
C
Monari
C
Tascini
C
Bistoni
F
Purified capsular polysaccharide of Cryptococcus neoformans induces interleukin-10 secretion by human monocytes
Infect Immun
 , vol. 
64
 (pg. 
2846
-
9
)
19
Pettoello-Mantovani
M
Casadevall
A
Kollmann
TR
Rubinstein
A
Goldstein
H
Enhancement of HIV-1 infection by the capsular polysaccharide of Cryptococcus neoformans
Lancet
 , vol. 
339
 (pg. 
21
-
3
)
20
Pettoello-Mantovani
M
Casadevall
A
Smarnworawong
P
Goldstein
H
Enhancement of HIV type 1 Infectivity in vitro by capsular polysaccharide of Cryptococcus neoformans and Haemophilus influenzae
AIDS Res Hum Retroviruses
 , vol. 
10
 (pg. 
1079
-
87
)
21
Bloomfield
N
Gordon
MA
Elmendorf
DF
Jr
Detection of Cryptococcus neoformans antigen in body fluids by latex particle agglutination
Proc Soc Exp Biol Med
 , vol. 
114
 (pg. 
64
-
7
)
22
Gade
W
Hinnefeld
SW
Babcock
LS
, et al.  . 
Comparison of the Premier cryptococcal antigen enzyme immunoassay and the latex agglutination assay for detection of cryptococcal antigens
J Clin Microbiol
 , vol. 
29
 (pg. 
1616
-
9
)
23
Chuck
SL
Sande
MA
Infections with Cryptococcus neoformans in the acquired immunodeficiency syndrome
N Engl J Med
 , vol. 
321
 (pg. 
794
-
9
)
24
Diamond
RD
Bennett
JE
Prognostic factors in cryptococcal meningitis: a study of 111 cases
Ann Intern Med
 , vol. 
80
 (pg. 
176
-
81
)
25
Zuger
A
Louie
E
Holzman
RS
Simberkoff
MS
Rahal
JJ
Cryptococcal disease in patients with the acquired immunodeficiency syndrome: diagnostic features and outcome of treatment
Ann Intern Med
 , vol. 
104
 (pg. 
234
-
40
)
26
Saag
MS
Powderly
WG
Cloud
GA
, et al.  . 
Comparison of amphotericin B with fluconazole in the treatment of acute AIDS-associated cryptococcal meningitis. NIAID Mycoses Study Group and the AIDS Clinical Trials Group
N Engl J Med
 , vol. 
326
 (pg. 
83
-
9
)
27
White
M
Cirrincione
C
Blevins
A
Armstrong
D
Cryptococcal meningitis: outcome in patients with AIDS and patients with neoplastic disease
J Infect Dis
 , vol. 
165
 (pg. 
960
-
3
)
28
Eng
RHK
Bishburg
E
Smith
SM
Cryptococcal infections in patients with acquired immune deficiency syndrome
Am J Med
 , vol. 
81
 (pg. 
19
-
23
)
29
Lendvai
N
Casadevall
A
Liang
Z
Goldman
DL
Mukherjee
J
Effect of immune mechanisms on the pharmacokinetics and organ distribution of cryptococcal polysaccharide
J Infect Dis
 , vol. 
177
 (pg. 
1647
-
59
)
30
Goldman
DL
Lee
SC
Casadevall
A
Tissue localization of Cryptococcus neoformans glucuronoxylomannan in the presence and absence of specific antibody
Infect Immun
 , vol. 
63
 (pg. 
3448
-
53
)
31
Kappe
R
Muller
J
Rapid clearance of Candida albicans mannan antigens by liver and spleen in contrast to prolonged circulation of Cryptococcus neoformans antigens
J Clin Microbiol
 , vol. 
29
 (pg. 
1665
-
9
)
32
Eng
R
Chmel
H
Corrado
M
Smith
SM
The course of cryptococcal capsular polysaccharide antigenemia/human cryptococcal polysaccharide elimination kinetics
Infection
 , vol. 
11
 (pg. 
132
-
6
)
33
Muchmore
HG
Scott
EN
Fromtling
RA
Cryptococcal capsular polysaccharide clearance in nonimmune mice
Mycopathologia
 , vol. 
78
 (pg. 
41
-
5
)
34
Qian
Q
Jutila
MA
van Rooijen
N
Cutler
JE
Elimination of mouse splenic macrophages correlates with increased susceptibility to experimental disseminated candidiasis
J Immunol
 , vol. 
152
 (pg. 
5000
-
8
)
35
van Rooijen
N
The liposome-mediated macrophage “suicide” technique
J Immunol Methods
 , vol. 
124
 (pg. 
1
-
6
)
36
Houpt
D
Pfrommer
GST
Young
B
Larson
T
Kozel
TR
Characteristics of antibodies in normal human serum that are reactive with the Cryptococcus neoformans glucuronoxylomannan
Infect Immun
 , vol. 
62
 (pg. 
2857
-
64
)
37
Eckert
TF
Kozel
TR
Production and characterization of monoclonal antibodies specific for Cryptococcus neoformans capsular polysaccharide
Infect Immun
 , vol. 
55
 (pg. 
1895
-
9
)
38
Savoy
AC
Lupan
DM
Manalo
PB
, et al.  . 
Acute lethal toxicity following passive immunization for treatment of murine cryptococcosis
Infect Immun
 , vol. 
65
 (pg. 
1800
-
7
)
39
Friedman
JJ
Organ plasma volume of normal unanesthetized mice
Proc Soc Exp Biol Med
 , vol. 
88
 (pg. 
323
-
5
)
40
Everett
NB
Simmons
B
Lasher
EP
Distribution of blood (Fe59) and plasma (I131) volumes of rats determined by liquid nitrogen freezing
Circ Res
 , vol. 
4
 (pg. 
419
-
24
)
41
Dreyer
G
Ray
W
The blood volume of mammals as determined by experiments upon rabbits, guinea-pigs and mice, and its relationship to the body weight and surface area expressed in a formula
Proc R Soc Lond B
 , vol. 
82
 (pg. 
545
-
6
)
42
Han
Y
van Rooijen
N
Cutler
JE
Binding of Candida albicans yeast cells to mouse popliteal lymph node tissue is mediated by macrophages
Infect Immun
 , vol. 
61
 (pg. 
3244
-
9
)
43
Heidelberger
M
MacLeod
CM
Markowitz
H
Roe
AS
Improved methods for the preparation of the specific polysaccharides of pneumococcus
J Exp Med
 , vol. 
91
 (pg. 
341
-
9
)
44
Cherniak
R
Reiss
E
Slodki
ME
Plattner
RD
Blumer
SO
Structure and antigenic activity of the capsular polysaccharide of Cryptococcus neoformans
Mol Immunol
 , vol. 
17
 (pg. 
1025
-
32
)
45
Shoham
S
Huang
C
Chen
JM
Golenbock
DT
Levitz
SM
Toll-like receptor 4 mediates intracellular signaling without TNF-α release in response to Cryptococcus neoformans polysaccharide capsule
J Immunol
 , vol. 
166
 (pg. 
4620
-
6
)
46
Mishra
S
Mishra
A
Gugnani
H
Reactivity of Cryptococcus neoformans and Candida species with concanavalin A and blood grouping sera
Indian J Med Res
 , vol. 
91
 (pg. 
382
-
5
)
47
Jones
JM
Amsbaugh
DF
Prescott
B
Kinetics of the antibody response to type III pneumococcal polysaccharide (SSS-II). I. Use of 125I-labeled SSS-II to study serum antibody levels, as well as the distribution and excretion of antigen after immunization
J Immunol
 , vol. 
116
 (pg. 
41
-
51
)
48
Chapin-Robertson
K
Bechtel
C
Waycott
S
Kontnick
C
Edberg
SC
Cryptococcal antigen detection from the urine of AIDS patients
Diagn Microbiol Infect Dis
 , vol. 
17
 (pg. 
197
-
201
)
49
Russell
MW
Brown
TA
Claflin
JL
Schroer
K
Mestecky
J
Immunoglobulin A–mediated hepatobiliary transport constitutes a natural pathway for disposing of bacterial antigens
Infect Immun
 , vol. 
42
 (pg. 
1041
-
8
)
50
Gadebusch
HH
Johnson
JD
Decomposition of the capsular polysaccharide of Cryptococcus neoformans 3723 by a soil microorganism
Naturwissenschaften
 , vol. 
47
 (pg. 
329
-
40
)
Presented in part: 97th general meeting, American Society for Microbiology, Miami Beach, Florida, May 1997 (abstract F-66).
Financial support: National Institutes of Health (AI-14209 to T.R.K.; AI-24912 to J.E.C.)
Portions of this work were done by M.G. to fulfill MD and PhD requirements at the University of Nevada School of Medicine and the Graduate School of the University of Nevada, Reno
Present affiliation: Department of Pediatrics, University of Virginia Health Sciences Center, Charlottesville