Toll-like receptors (TLR) 2 and 4 are cell surface receptors that in association with CD14 enable phagocytic inflammatory responses to a variety of microbial products. Activation via these receptors triggers signaling cascades, resulting in nuclear translocation of NF-κB and a proinflammatory response including TNF-α production. We investigated whether TLRs participate in the host response to Cryptococcus neoformans glucuronoxylomannan (GXM), the major capsular polysaccharide of this fungus. Chinese hamster ovary fibroblasts transfected with human TLR2, TLR4, and/or CD14 bound fluorescently labeled GXM. The transfected Chinese hamster ovary cells were challenged with GXM, and activation of an NF-κB-dependent reporter construct was evaluated. Activation was observed in cells transfected with both CD14 and TLR4. GXM also stimulated nuclear NF-κB translocation in PBMC and RAW 264.7 cells. However, stimulation of these cells with GXM resulted in neither TNF-α secretion nor activation of the extracellular signal-regulated kinase 1/2, p38, and stress-activated protein kinase/c-Jun N-terminal kinase mitogen-activated protein kinase pathways. These findings suggest that TLRs, in conjunction with CD14, function as pattern recognition receptors for GXM. Furthermore, whereas GXM stimulates cells to translocate NF-κB to the nucleus, it does not induce activation of mitogen-activated protein kinase pathways or release of TNF-α. Taken together, these observations suggest a novel scenario whereby GXM stimulates cells via CD14 and TLR4, resulting in an incomplete activation of pathways necessary for TNF-α production.
Cryptococcus neoformans, a ubiquitous yeast with worldwide distribution, predominantly infects individuals with defects in cellular immunity. Risk factors for cryptococcosis include infection with the human immune deficiency virus, therapy with cytotoxic agents, organ transplantation, hematological malignancies, and treatment with corticosteroids (1). Among patients with AIDS, cryptococcosis has emerged as one of the most common life-threatening mycosis. In the United States, cryptococcosis affects 5–10% of AIDS patients, whereas in parts of the developing world infection rates are even higher (2, 3). Despite advances in therapy, successful treatment of cryptococcosis, especially among severely immune-suppressed individuals, is difficult. Mortality can reach 25%, and many survivors are left with permanent neurological disability.
C. neoformans is the only human pathogenic fungus with a polysaccharide capsule. The capsule is composed primarily of a high molecular mass glucuronoxylomannan (GXM)3 (4). Capsule-deficient mutants of C. neoformans have greatly reduced virulence compared with their parent strains. In addition to coating the surface of C. neoformans, GXM is shed from the fungus. In humans and experimental animals infected with C. neoformans, shed GXM can circulate in the blood and CSF at concentrations well into the micrograms per milliliter range, and it is likely that in infected tissues, local concentrations of GXM in the milligram per milliliter range are achieved (5, 6, 7).
A myriad of mechanisms has been postulated by which capsular polysaccharide, including GXM, helps the fungus to elude host defenses (1, 8). Capsular polysaccharide inhibits phagocytosis by presenting a surface that is not well recognized by phagocytes. In the absence of opsonins, most phagocytic cells will not bind to encapsulated strains of C. neoformans (9). GXM inhibits neutrophil migration by triggering those cells to shed the adhesion molecule L-selectin (10). GXM may further reduce neutrophil adherence by binding to CD18, thereby impairing complex formation with endothelial ICAM-1 (11). Other immune-suppressive effects attributed to the cryptococcal capsule include induction of suppressor T cells, inhibition of lymphoproliferation, and impairment of Ag presentation (8).
Prevention of cryptococcosis requires an immune response composed of innate defenses and then, if unsuccessful, acquired immunity. The innate immune system identifies infectious agents by means of pattern recognition proteins. The host then attempts to dispose of the microbe and to activate effector defense mechanisms (12). Monocytes and macrophages are central to these processes. Whereas disposal is attempted via phagocytosis and the generation and/or release of antimicrobial substances, activation is accomplished by the production of proinflammatory cytokines such as TNF-α. Both processes utilize a series of events initiated by cell surface binding and mediated by intracellular signaling.
Toll-like receptors (TLR) are a recently described family of cell surface receptors. Although no direct binding studies have been reported describing TLR ligands, molecular genetic studies suggest that the extracellular domain of TLRs recognizes conserved molecular patterns associated with a range of microbial pathogens. The cytoplasmic portion of all TLRs is homologous to the cytoplasmic domain of the IL-1R, and is responsible for signal transduction (13). The downstream pathways leading to the production and release of TNF-α and other proinflammatory cytokines are starting to be defined. TLR engagement leads to binding of a cytoplasmic adapter protein (MyD88), and activation of IL-1R-associated kinase, followed by phosphorylation and activation of TNFR-associated factor 6 (TRAF6). Activation of TRAF6 subsequently results in the activation of NF-κB-inducing kinase, followed by I-κB (inhibitory protein that dissociates from NF-κB) kinase, phosphorylation of I-κB, and its dissociation from NF-κB (14). These events ultimately result in the translocation of NF-κB to the nucleus and the initiation of gene transcription. In addition to pathways leading to NF-κB activation, three mitogen-activated protein kinase (MAPK) cascades, the extracellular signal-regulated kinase (ERK), p38 MAPK, and stress-activated protein kinase (SAPK)/c-JUN N-terminal kinase (JNK), appear necessary for optimal TNF-α cytokine production (15).
The diversity of microbial products to which TLR2 and 4, often in association with CD14, mediate intracellular signaling has been recently reviewed (16). Molecules derived from a range of pathogens, including Gram-negative and Gram-positive bacteria, mycobacteria, mycoplasmas, spirochetes, and the fungal product zymosan, have been reported to activate cells via TLR2 (17, 18). To date, TLR4 has only been reported to mediate intracellular signaling in response to LPS (17) and Mycobacterium tuberculosis (19). In the present study, we tested the hypothesis that TLRs participate in mediating signaling responses to C. neoformans. We found that whereas GXM, the major cryptococcal polysaccharide capsule component, binds to and activates cells to induce nuclear translocation of NF-κB via an interaction with TLR4 and CD14, it neither activates MAPK pathways nor induces TNF-α production. These findings suggest that the interaction between GXM and TLR may represent a mechanism of immune dysregulation whereby the organism incompletely activates cascades leading to TNF-α production.
Materials and Methods
PBS and trypsin/EDTA were obtained from BioWhittaker (Walkersville, MD). Hams F12, RPMI 1640, and G418 were obtained from Life Technologies (Gaithersburg, MD). FBS was from Life Technologies (Grand Island, NY). Ciprofloxacin was obtained from Bayer Pharmaceuticals (West Haven, CT). Hygromycin B was obtained from Calbiochem (La Jolla, CA). Puromycin, streptomycin, penicillin, Ficoll-Hypaque, and LPS (from Escherichia coli O111:B4) were obtained from Sigma (St. Louis, MO). Where indicated, an LPS preparation devoid of TLR2-stimulating activity (a kind gift of Dr. Stephanie Vogel, Uniformed Services University of the Health Sciences, Bethesda, MD) (20) was utilized in lieu of Sigma LPS. Anti-CD25 mAb conjugated with FITC was obtained from Caltag (Burlingame, CA). The 4-(4,6-dichlorotriazinyl)aminofluorescein (DTAF) was obtained from Molecular Probes (Eugene, OR). rIL-1β was obtained from Genzyme (Cambridge, MA).
All incubations were at 37°C in humidified air supplemented with 5% CO2, except where otherwise noted. All experiments were performed under conditions designed to minimize endotoxin contamination, as in our previous studies (21, 22, 23). Complete media is defined as RPMI 1640 with 10% FBS supplemented with penicillin and streptomycin. Media, FBS, and PBS contained less than 0.03 endotoxin U/ml, as certified by the manufacturer. Pooled human serum (PHS) was obtained by combining serum from >10 healthy donors under conditions designed to preserve complement activity. PHS was stored in aliquots at −70°C until use.
Purification of GXM
GXM was purified from culture supernatants of two serotype A strains of C. neoformans, J11a (a gift of Arturo Casadevall, Albert Einstein College of Medicine, New York, NY) and 6 (ATCC 62066), using a modification of the method of Cherniak (24). Serotype A is the most common serotype seen in clinical infections, and the structure of the GXM from both strains has been well characterized (24, 25). C. neoformans was grown in yeast nitrogen base media, supernatants collected, and polysaccharide precipitated with sodium acetate (taking care to keep the pH at 7 in order not to destroy acetyl groups), followed by 2.5 vol of ethanol. The precipitate (consisting of GXM and galactoxylomannan) was collected, quantitated by the phenol-sulfuric acid method (26), and solubilized in 0.2 M NaCl. GXM was separated from galactoxylomannan by combining 0.3% cetyltrimethylammonium bromide (CTAB) with the polysaccharide mixture, at a ratio of 3:1 by mass, to yield a precipitate containing CTAB bound to GXM. The CTAB-GXM precipitate was solubilized in 0.2 M NaCl, and 2 vol ethanol added to precipitate selectively the GXM. The precipitate was dissolved in 2 M NaCl and extensively dialyzed with a 10 kDa molecular mass cutoff against 1 M NaCl and then H2O. Quantitation was again by the phenol-sulfuric acid method.
Size and purity of the GXM preparations were evaluated by SDS-PAGE, followed by Periodic Acid-Schiff staining. Both 6 and J11a GXM migrated approximately the same distance as a dextran standard of average molecular mass 473,000. Low molecular mass bands were not seen, even using overloaded gels, suggesting that both GXM solutions were free of significant contamination with galactoxylomannan. The GXM was endotoxin free to less than 0.03 endotoxin U/ml, as determined by the Limulus amebocyte lysate assay (Associates of Cape Cod, Woods Hole, MA). The negative Limulus amebocyte lysate assay also suggests that the GXM was free of contamination with β-glucan to less than 1 ng (27). Finally, the GXM preparation used did not contain detectable amounts of protein, as determined by the bicinchoninic acid protein assay (Pierce, Rockford, IL). The bicinchoninic acid assay was sensitive to 5 μg/ml.
Fluorescein-labeled GXM was prepared by covalent binding of the polysaccharide to DTAF per the method of Prigent (28), with modifications for GXM. In brief; DTAF and GXM were dissolved in a 0.05 M borate buffer solution at pH 9. The reaction was allowed to proceed for 24 h at 37οC, at which point 0.1 M glycine was added for 72 h to react with any unincorporated DTAF. DTAF-GXM was then purified by centrifuge filtration using a mini centrifuge filter (Nalgene, Rochester, NY) with a 10 kDa molecular mass cutoff.
The stably transfected cell lines used in the studies are listed in Table I. The cDNAs for TLR2 and TLR4 were cloned into the vector pFLAG, as described (29). In addition to expressing the indicated human TLR and/or CD14, all cell lines except Chinese hamster ovary (CHO)/TLR2 were transfected with the NF-κB reporter plasmid, ELAM.tac. This construct contains an NF-κB-dependent portion of the endothelial cell-leukocyte adhesion molecule (ELAM-1) promoter driving cell surface expression of human CD25 (tac) (30, 31, 32). CHO cells were grown as adherent monolayers in tissue culture flasks in F12 medium and passed at least twice weekly to maintain logarithmic growth.
|Cell Line .||Transfected Receptor(s) .||Refs. .|
|CHO/ELAM.tac||Control cell line||31|
|CHO/CD14/TLR2/ELAM.tac||CD14, TLR2||17, 32|
|CHO/CD14/TLR4/ELAM.tac||CD14, TLR4||31, 32|
|Cell Line .||Transfected Receptor(s) .||Refs. .|
|CHO/ELAM.tac||Control cell line||31|
|CHO/CD14/TLR2/ELAM.tac||CD14, TLR2||17, 32|
|CHO/CD14/TLR4/ELAM.tac||CD14, TLR4||31, 32|
CHO cells were transfected with the indicated human receptors. All cells except CHO/TLR2 also were transfected with ELAM.tac, a construct that contains an NF-κB dependent promoter driving cell surface expression of human CD25.
The mouse macrophage-like RAW 264.7 cell line (American Type Culture Collection, Manassas, VA) was grown in RPMI with 10% FBS and maintained by serial passage in culture flasks. Human PBMC were purified as in our previous studies (21, 22, 23). Briefly, peripheral blood obtained from healthy volunteers by venipuncture was anticoagulated with 10 U/ml of pyrogen-free heparin (Elkins-Sinn, Cherry Hill, NJ) and centrifuged at 500 × g for 15 min, and the leukocyte-rich buffy coat was harvested. PBMC were collected following centrifugation of the buffy coat over a gradient of Ficoll-Hypaque.
Flow cytometric analysis of CHO cells
CHO cells were plated in 12-well tissue culture dishes at a concentration of 1 × 105 cells/ml, incubated for 24 h, and then stimulated for 18 h with GXM or LPS. Cells were detached from the surface using trypsin/EDTA and assessed by flow cytometry for the presence of surface CD25, as in our previous studies (32). Note that CD14, TLR2, TLR4, and CD4 are all resistant to standard trypsin treatment (data not shown). Cells analyzed for binding of GXM were likewise grown as adherent monolayers and detached from the surface using trypsin/EDTA. The cells were washed once with PBS, resuspended in PBS/10% FBS, incubated with fluorescein-labeled J11a GXM at 37°C for 30 min, washed twice with PBS, and evaluated for fluorescence by flow cytometry. Cell surface expression of heterologous receptors was verified by flow cytometry using anti-FLAG and anti-CD14 Abs to measure TLR and CD14 expression, respectively (19, 33).
Analysis of NF-κB translocation
PBMC and RAW 264.7 cells were stimulated with 250 μg/ml of J11a GXM and 100 ng/ml of LPS for 1 h. Cells were washed with PBS, and nuclear extracts were prepared and analyzed using the EMSA, as described (21, 23, 31). Briefly, nuclear extracts from stimulated cells were prepared in the presence of protease inhibitors, and protein concentration was determined using a commercial kit (Bio-Rad Laboratories, Hercules, CA). An oligonuleotide containing the NF-κB-binding sequence (Promega) was end labeled with [α-32P]dATP and [α-32P]dCTP using Klenow DNA polymerase (Promega, Madison, WI). Unincorporated nucleotides were removed with a G-25 spin column (Pharmacia Biotech, Piscataway, NJ). Labeled probe (0.2 ng) and nuclear extracts (4 μg) were incubated at room temperature for 30 min in a 1× band shift buffer (10 mM Tris-HCl, pH 7.8, 1 mM EDTA, 40 mM KCl, 1 mM DTT) containing 50 mg/ml of poly(dI-dC) and 5% glycerol. Reactions were then size fractionated by electrophoresis in 4% native polyacrylamide gels, transferred to 3 MM filter paper (Whatman Laboratory Products, Clifton, NJ), dried, and visualized by autoradiography.
Measurement of TNF-α production and gene expression
In 96-well flat-bottom plates, 105 PBMC and RAW 264.7 cells were stimulated with 250 μg/ml of GXM or 100 ng/ml of LPS for 18 h in complete medium. PBMC were also stimulated in RPMI with 10% PHS. Human and mouse TNF-α were measured in culture supernatants by sandwich ELISA using commercial kits according to the manufacturer’s instructions (Duo Set ELISA Development System, human and mouse TNF-α; R&D Systems, Minneapolis, MN). The ELISA were sensitive over a range of 10 to 3000 pg/ml.
Proinflammatory cytokine gene expression was measured by the RNase protection assay (RPA), as in our previous studies (23). Briefly, PBMC (5 × 106) were stimulated with 100 ng/ml LPS or 250 μg/ml GXM in six-well polystyrene plates in the presence of RPMI 1640 containing 10% FBS. Total cellular RNA was extracted from PBMC using TRIzol reagent (Life Technologies, Grand Island, NY) and analyzed for human TNF-α, IL-1β, and IL-6 mRNA by RPA using a custom kit (PharMingen, San Diego, CA). Integrity of the RNA and equal loading of lanes were verified using the housekeeping genes L32 and GAPDH.
Analysis of activation of MAPK pathway
Phosphorylation of MAPK was evaluated in human PBMC and RAW 264.7 cells according to the manufacturer’s (New England Biolabs, Beverly, MA) protocol. Briefly, following stimulation with LPS or J11a GXM, cell lysates were prepared, separated by SDS-PAGE, and analyzed by Western immunoblotting using HRP-labeled Abs directed against the phosphorylated (activated) forms of p44/42 MAPK (ERK1/2), p38 MAPK, and SAPK/JNK. These Abs do not react with unphosphorylated forms of the protein kinases. Identity of the bands was confirmed using molecular mass markers and positive controls supplied by the manufacturer.
Means and SE were compared by the two-tailed two-sample t test using a statistical software program (SigmaStat; Jandel Scientific Software, San Rafael, CA). Adjustments for significance of multiple comparisons were made using the Bonferroni correction.
TLR2, TLR4, and CD14 mediate GXM binding
To determine the contribution of TLR2, TLR4, and CD14 to binding GXM, fluorescein-labeled J11a GXM was incubated with the six stably transfected CHO cell lines listed in Table I: CHO/ELAM.tac, CHO/CD14/ELAM.tac, CHO/TLR2, CHO/TLR4/ELAM.tac, CHO/CD14/TLR2/ELAM.tac, and CHO/CD14/TLR4/ELAM.tac. Compared with the parent strain, an increase in fluorescence was observed in CHO cells expressing CD14, TLR, or both (Fig. 1). Preincubation of the cells with 250 μg/ml of unlabeled J11a GXM nearly completely competed out the increase in fluorescence observed following incubation with GXM-DTAF (Fig. 2). These data suggest CD14, TLR2, and TLR4 participate in binding to GXM.
TLR4, but not TLR2, confers GXM responsiveness on CHO/CD14 cells
We next sought to determine whether binding of GXM to transfected CHO cells led to cellular activation as measured by nuclear translocation of NF-κB. To accomplish this, we used CHO reporter cells stably transfected with an inducible promoter-driving surface expression of CD25 in response to NF-κB activation. Activation of cells was measured by flow cytometry using FITC-labeled Ab to CD25.
We exposed CHO/CD14/ELAM.tac, CHO/CD14/TLR2/ELAM.tac, and CHO/CD14/TLR4.tac cells to J11a GXM. As a positive control, cells were incubated with LPS, a known stimulant of CHO/CD14/ELAM.tac, CHO/CD14/TLR2/ELAM.tac, and CHO/CD14/TLR4/ELAM.tac cells (29, 34). Both GXM and LPS stimulated CHO/CD14/TLR4/ELAM.tac cells to translocate NF-κB to the nucleus, as indicated by increased surface expression of CD25 (Fig. 3). Surface expression of CD25 increased in a concentration-dependent manner in response to escalating concentrations of GXM from 62.5 to 250 μg/ml. LPS, but not GXM, increased surface expression of CD25 among CHO cells expressing either CD14 alone or CD14 and TLR2. Incubation of CHO cells with 25 μg/ml of polymixin B inhibited responses to LPS, but not GXM (data not shown). This result in combination with our finding that our GXM preparations were free of LPS, as determined by Limulus amebocyte lysates, made it unlikely that the results with GXM stimulation were secondary to LPS contamination. Similar results were obtained comparing GXM preparations from strains J11a and 6 (data not shown).
The above data established that GXM stimulated CHO/CD14/TLR4/ELAM.tac, but not CHO/CD14/ELAM.tac or CHO/CD14/TLR2/ELAM.tac, to translocate NF-κB to the nucleus. Next, we sought to determine whether TLR4 alone, in the absence of CD14, could activate NF-κB. We exposed CHO/TLR4/ELAM.tac to 250 μg/ml of J11a GXM, 1 μg/ml of LPS, and 3 ng/ml of IL-1β. CHO/TLR4/ELAM.tac cells were observed to increase surface expression of CD25 only minimally in response to GXM or LPS (Fig. 3). Consistent with previous observations, only IL-1β, but not LPS, activated the cells to translocate NF-κB in the absence of membrane-bound CD14 (35, 36). Thus, GXM stimulated a vigorous NF-κB response only in the presence of coexpression of both human TLR4 and CD14.
GXM induces nuclear translocation of NF-κB among PBMC and RAW 264.7 cells
The next set of experiments assessed whether GXM stimulates NF-κB nuclear translocation in professional phagocytic cells, as it does in the CHO/CD14/TLR4/ELAM.tac cell line. Human PBMC and murine RAW 264.7 cells were stimulated with 250 μg/ml of J11a GXM or 100 ng/ml of LPS for 1 h. Both cell types were noted to increase translocation of NF-κB to the nucleus in response to J11a GXM and LPS when compared with controls (Fig. 4). Unopsonized C. neoformans strain 145 (21), which presents a surface coated with serotype A GXM, also stimulated NF-κB translocation.
GXM fails to stimulate TNF-α gene expression and release
Having found that GXM stimulates NF-κB nuclear translocation, we next sought to determine whether TNF-α release was likewise stimulated. Mononuclear phagocytes from two sources were studied: human PBMC and RAW 264.7 cells. In both cell types, levels of TNF-α following an 18-h stimulation with 250 μg/ml of GXM were not significantly different from levels seen in unstimulated cells (Fig. 5 A). As expected, an 18-h incubation with 100 ng/ml of LPS stimulated large amounts of TNF-α. Substitution of FBS with PHS in the media did not alter TNF-α responses (data not shown). Moreover, in experiments using three different donors each tested in triplicate, preincubation of PBMC with 250 μg/ml of GXM did not significantly alter TNF-α release induced by 10 ng/ml of an LPS preparation that is devoid of TLR2-stimulating activity (mean ± SEM TNF release 1295 ± 214 and 1273 ± 216 ng/ml in the absence and presence of GXM, respectively).
Failure of GXM to stimulate TNF-α release was not secondary to posttranscriptional mechanisms as GXM also failed to stimulate TNF-α mRNA in PBMC (Fig. 5 B). In addition, cellular levels of IL-1β and IL-6 mRNA did not change following GXM stimulation. In contrast, LPS stimulated gene expression of all three proinflammatory cytokines.
GXM fails to activate MAPK pathways among PBMC and RAW 264.7 cells
The MAPK are a group of signaling kinases that, by modulating the phosphorylation status of transcription factors, link transmembrane signaling with gene induction. Three major groups of MAPKs have been identified in mammalian cells: ERK1/2, p38 MAPK, and SAPK/JNK. All three MAPKs are thought to be necessary for optimal TNF-α induction in LPS-stimulated macrophages (37, 38). To assess whether the above MAPK pathways were activated by GXM, PBMC (data not shown) and RAW 264.7 cells were exposed to 250 μg/ml of strain 6 GXM for 10, 20, 40, and 60 min. As a positive control, the cells were likewise stimulated with 100 ng/ml of LPS. Activation of the MAPK pathways was measured by Western blot using Abs reactive to phosphorylated ERK1/2, p38, and SAPK/JNK. LPS activated all three MAPK pathways, peaking by 20 min and falling off at 60 min (Fig. 6). In contrast, GXM failed to stimulate phosphorylation of any of the MAPKs. Simultaneous costimulation of PBMC and RAW 264.7 cells with 250 μg/ml of strain 6 GXM and 100 ng/ml of LPS did not abrogate LPS-induced MAPK pathway activation, suggesting that GXM does not deactivate stimulated pathways. These data suggest failure to stimulate MAPK pathways as a possible mechanism for the lack of TNF-α release by GXM.
Our data, using heterologously transfected CHO cells, demonstrate that CD14, TLR2, and TLR4, independently and in concert, facilitate binding of cryptococcal GXM. These interactions point toward a role for CD14 and TLR as pattern recognition receptors for C. neoformans and add to the already impressive list of microbial products that interact with these receptors (16, 17, 18, 34). In support of our findings with CD14, others have demonstrated that Abs directed against CD14 block uptake of unopsonized C. neoformans by swine microglia (39). A report showing that TLR2 enhances zymosan-induced NF-κB activation provides precedence for fungal stimulation of cells via TLR (18). However, our studies are the first to demonstrate a role for mammalian TLR in interactions with products derived from a fungus of unquestionable pathogenicity. Interestingly, in the fruit fly Drosophila melanogaster, the TLR homologue Toll mediates antifungal responses by inducing genes that encode for the antifungal peptide drosomycin. This interaction occurs by way of Rel-Cactus complexes, which are functionally and structurally equivalent to vertebrate NF-κB/I-κB complexes (40).
The conjugation of fluorescein to GXM was performed at an alkaline pH. It is possible that these conditions resulted in loss of O-acetyl groups on the polysaccharide. However, it is unlikely that this substantially altered the binding characteristics of the labeled GXM as unconjugated native GXM was able to displace its binding to transfected CHO cells.
Nuclear translocation of NF-κB, as measured by reporter gene expression, was stimulated by GXM only in CHO cells heterologously expressing both CD14 and TLR4. NF-κB activation in response to GXM did not occur in CHO cells expressing CD14 and TLR2 despite our finding that those cells effectively bound GXM. Thus, receptor occupancy by GXM does not necessarily lead to a signaling response. In addition to stimulating nuclear translocation of NF-κB in CHO/CD14/TLR4/ELAM.tac, GXM activated NF-κB in human PBMC and the murine macrophage cell line RAW 264.7. This is consistent with the finding that mononuclear phagocytes express both CD14 and TLR4 (16). The failure of CHO/CD14 cells, which express endogenous TLR4, to react to stimulation with GXM may be due to the inability of hamster TLR4 to respond to this ligand. Recently, it has been reported that human and hamster TLR4 mediate disparate, species-specific responses to identical ligands (34). Furthermore, it is possible that GXM stimulation of transfected CHO cells, but not mononuclear phagocytes, is dependent upon overexpression of TLR4.
Our data demonstrating the capacity of GXM to stimulate nuclear translocation of NF-κB provide a putative mechanism for the observation that capsular polysaccharide enhances HIV-1 infection (41, 42). NF-κB binding sites are present in the HIV long terminal repeat promoter, and as a consequence, stimuli that induce nuclear translocation of NF-κB generally also induce HIV replication (43, 44).
Cytokine dysregulation may contribute to the capacity of GXM to act as a virulence factor. GXM has been reported to stimulate IL-6 and IL-10 production by monocytes and to inhibit LPS-induced secretion of cytokines (45, 46). Vecchiarelli et al. reported that GXM induces IL-8 secretion by human PMN and that the effect was due to complement activation with generation of C3a and C5a (47). Whereas opsonized C. neoformans stimulates cells to produce TNF-α (22, 48), this response was not seen with either GXM or unopsonized organisms. In contrast to our finding, two laboratories have reported that GXM induced small amounts of TNF-α release from human PBMC, but only in the presence of complement-sufficient PHS (49, 50). The reasons for this disparity are unclear, although the possibility remains that some of the GXM preparations used might have been degraded and/or contaminated by galactoxylomannan, mannoprotein, β-glucans, and/or bacterial endotoxin. Using sensitive techniques, we demonstrated that the GXM preparations utilized in our studies were free of significant amounts of these contaminants and migrated as a single high molecular mass band.
NF-κB nuclear translocation is thought to be necessary, but not sufficient, for TNF-α production. Our finding that GXM stimulated NF-κB without release of TNF-α prompted us to examine whether GXM stimulated other pathways thought necessary for optimal TNF-α release, notably the MAPK cascades (15, 38, 51). None of three MAPK pathways were activated by GXM, suggesting a mechanism to explain the lack of TNF-α. At present, significant gaps remain in our understanding of how signaling via TLR results in activation of MAPK and NF-κB. The adapter protein MyD88 acts to transduce signals in response to activation of CD14/TLR complexes. The TLR signaling pathway that leads to SAP/JNK and NF-κB activation appears to diverge downstream of MyD88 and upstream of TRAF6 (52). Nevertheless, MyD88 knockout mice activate MAPK and NF-κB in response to LPS, although activation does not lead to TNF-α production (53). Our experiments indicate that GXM activates NF-κB, but not the MAPK cascades in mononuclear phagocytes. Moreover, we found that GXM does not inhibit MAPK pathways or TNF-α production induced by LPS. The mechanisms responsible for this are unclear at the present time. It could be postulated that GXM and LPS interact with various components of the TLR/CD14 complex in a fashion dissimilar enough to yield divergent downstream signals. Alternatively, internalization of GXM may influence the activity of proteins distal to the site at which NF-κB and MAPK pathways separate.
Our studies utilized concentrations of GXM up to 250 μg/ml. The molecular mass of GXM and LPS have been estimated to be approximately 1 × 106 and 1 × 104, respectively (4, 54). Due to the very high molecular mass of GXM, on a molar basis, LPS and GXM did not vary greatly in their ability to activate NF-κB. From a clinical standpoint, in patients with AIDS and cryptococcosis, the fungal burden tends to be very high, and GXM concentrations approaching 1 mg/ml may be attained in the serum and cerebrospinal fluid (5). More importantly though, in infected tissues such as the brain, even larger concentrations of GXM are presumably found (6). Consistent with our data, accumulation of GXM in vivo preferentially occurs in macrophages (5), cells known to express CD14 and TLR4 (55, 56). In addition to the effects of shed GXM on cell function, our data demonstrate that whole C. neoformans, which presents a surface predominately consisting of GXM, also stimulates NF-κB. Like purified GXM, unopsonized encapsulated C. neoformans does not stimulate PBMC to make TNF-α (22).
Taken together, our data suggest a novel scenario whereby GXM interacts with CD14 and TLR4 to stimulate NF-κB nuclear translocation without MAPK activation or TNF-α production. The clinical consequences of these findings remain speculative. As discussed above, GXM is present in high concentrations in patients with cryptococcosis and is the major virulence factor of C. neoformans. Thus, the interaction between GXM and CD14/TLR4 could exert an immunodysregulatory effect deleterious for the host.
We thank Dr. Arturo Casadevall for expert advice on purifying GXM, and Stephen M. Weber for help with the RPA.
This work was supported by Grants AI37532, AI25780, and GM54060 from the National Institutes of Health. S.M.L. is the recipient of a Burroughs Wellcome Fund Scholar Award in Pathogenic Mycology.
Abbreviations used in this paper: GXM, glucuronoxylomannan; CHO, Chinese hamster ovary; CTAB, cetyltrimethylammonium bromide; DTAF, 4-(4,6-dichlorotriazinyl)aminofluorescein; ELAM, endothelial cell-leukocyte adhesion molecule; ERK, extracellular signal-regulated kinase; I-κB, inhibitory protein that dissociates from NF-κB; JNK, c-Jun N-terminal kinase; MAPK, mitogen-activated protein kinase; PHS, pooled human serum; RPA, RNase protection assay; SAPK, stress-activated protein kinase; TLR, Toll-like receptor; TRAF6, TNFR-associated factor 6.