Phagocytosis by Human Neutrophils Is Stimulated by a Unique Fungal Cell Wall Component

To determine the opsonic response of neutrophils to the yeast form of Candida albicans, we performed microarray experiments using whole genome Affymetrix arrays on RNA extracted from neutrophils that had ingested the fungus. Among the genes induced in neutrophils was a family of heat shock proteins (HSPs) (The 10 kDa HSPE1, 40 kDa DNAJB1 and DNAJB9, 70 kDa HSPA1A and HSPA9B, 90 kDa HSPCA and HSPCB, and 105/110 kDa HSPH1) (see Table S1 available in the Supplemental Data available with this article online). These inductions were corroborated by quantitative real-time PCR (Figure 1A).

The induction of HSPs was greater if Candida was first heated to unmask the underlying β-glucan (Figure 1B), suggesting that expression of HSPs is induced by the exposed β-glucan.

To assay which of the glucan components was responsible for induction of HSPs, we presented neutrophils with various glucan polymers. As unprimed neutrophils respond best to insoluble material (Fossati et al., 2002), we conjugated several sources of β-glucan to 6 μm polystyrene beads, which are similar in size to Candida albicans yeast form cells (5 μm). We utilized β-1,3-glucan and β-1,6-glucan purified from Candida cell walls, as well as nonfungal sources of β-1,3-glucan (laminarin) and β-1,6-glucan (pustulan), both used as standards for these polymers in previous studies (Brown and Gordon, 2001; Palma et al., 2006). The extent of coating was quantified as described in Experimental Procedures.

Beads coated with Candida-derived β-1,6-glucan were more efficient than beads coated with Candida-derived β-1,3-glucan at elicitation of HSPs in neutrophils (Figure 1C). Similar results were observed with the standards: pustulan (β-1,6-glucan) being more powerful stimulant than laminarin (β-1,3-glucan). Beads coated with pustulan elicited high levels of HSPs in neutrophils (Figure 1D), whereas beads coated with laminarin elicited only low levels of HSPs expression (Figure 1D) despite efficient coating of the beads. Glucans in the soluble form elicited only low levels of HSPs in neutrophils (Figure 1E), suggesting that neutrophils respond when the polysaccharides are presented on particles. β-glucan from barley, which is composed of β-1,3-glucan (30%) and β-1,4-glucan (70%), also did not elicit HSPs expression (Figure 1E). Beads coated with α-1,6-glucan (dextran) did not elicit any response (Figure 1E), suggesting that the neutrophil response is specific to the β configuration. Elicitation of HSPs required heat-labile serum components, as pustulan-coated beads did not elicit HSPs when opsonized with heat-inactivated (HI) serum (Figure 1D). We conclude that neutrophils express HSPs following ingestion of opsonized beads coated with β-1,6-glucan.

Digestion of pustulan with an endoglucanase specific for the β-1,6-glucan linkage (Lora et al., 1995) confirmed that it consisted of β-1,6-glucan. Pustulan digested with the enzyme resulted in as much as 80% reduction in elicitation of HSPs (Figure 2A). Analysis of the digestion products of pustulan by column chromatography (see Experimental Procedures) revealed the expected peak (Figure 2 compare B to C), composed of di-and trisaccharides, as determined by TLC (G2-G3, compare lanes 2 and 3, Figure 2D). These short polymers in the G2-G3 fraction did not coat the beads efficiently (see Experimental Procedures).

In addition to the expected digestion products there was a small amount of polysaccharide (∼5% of the starting material) that was resistant to the enzyme (Figure 2C, designated as Vo) and elicited expression of the HSPs in the neutrophils. Since pustulan is reported to have a small percentage of O-acetylated β-1,6-glucan (Nishikawa et al., 1970), the small peak resistant to enzymatic digestion could be the acetylated polymer. Consistent with this interpretation, deacetylation of pustulan prior to digestion and fractionation (see Experimental Procedures) virtually abolished this minor component (Figure 2E, insert) and rendered the digested material unable to elicit HSPs (Figure 2F). The residual activity of pustulan digested with the endo-β-1,6-glucanase is therefore attributable to O-acetylated β-1,6-glucan that was resistant to the enzyme and coated the beads efficiently. Chemically acetylated laminarin (β-1,3-glucan) did not elicit any expression of HSPs (Figure S1), suggesting that the stimulation of neutrophils by β-1,6-glucan is not due to acetylation per se. Based on these results we conclude that the activity of pustulan is due to the β-1,6-glucan.

Time-lapse microscopy of neutrophils ingesting beads revealed that beads coated with pustulan (β-1,6-glucan) were more efficiently internalized than beads coated with laminarin (β-1,3-glucan) (Figure 3A, compare Ad to Ab, and compare Movie S2 to Movie S1). Neutrophils readily ingested beads when coated with pustulan. In contrast, most neutrophils ignored the laminarin-coated beads, and only a few of them ingested beads (Figure 3Ab and Movie S1, a neutrophil ingesting one bead on the upper left corner, and one neutrophil ingesting two beads on the upper right corner).

Quantification of a population of neutrophils ingesting beads by Fluorescence Activated Cell Sorting (FACS) support the idea that pustulan mediates more efficient internalization of beads than laminarin, β-1,3-glucan from Candida, or β-glucan from barley, although laminarin and β-1,3-glucan from Candida promoted internalization somewhat better than uncoated beads (Figure 3B, compare panel Be to panels Bb–Bd, respectively).

Since killing of pathogens by neutrophils depends on a burst of ROS (Babior et al., 1973), we tested whether ROS were induced by either of the two glucans. ROS could not be detected in neutrophils alone or in neutrophils presented with beads that had not been treated (Figure 3C, red and green, respectively). Low levels were detected with beads coated with laminarin, β-glucan from Candida, or β-1,3-glucan from barley (Figure 3C, blue, brown, and purple, respectively). However, beads coated with pustulan stimulated the generation of significant amounts of ROS (Figure 3C, light blue). No ROS was detected by adding soluble pustulan (Figure 3C, pink). These data show that β-1,6-glucan evokes a massive production of ROS, whereas the response to β-1,3-glucan is much lower by comparison.

As serum is required for phagocytosis and the induction of HSPs by beads coated with β-1,6-glucan (Figure 1D), we determined whether there was a serum component that was differentially deposited on beads coated with either β-1,6-glucan or β-1,3-glucan. Beads were coated with laminarin or pustulan and opsonized as described in Experimental Procedures. Bound proteins were removed from beads and separated by SDS gel electrophoresis (see Experimental Procedures). Although a number of serum proteins bound both β-1,3-glucan or β-1,6-glucan, there were two prominent proteins that adhered more avidly to β-1,6-glucan (Figure 4A). These proteins were extracted from the gel and subjected to analysis by mass spectrometry. The peptides from these bands gave masses that identified both proteins as C3, suggesting that proteolytic fragments of C3 are deposited more avidly on β-1,6-glucan than β-1,3-glucan.

Western analysis revealed that, indeed, β-1,6-glucan was deposited with more C3 (Figure 4B). Antibodies specific for the alpha chain detected high molecular weight bands, including bands the size of the complete C3 or C3b (105 and 115 kDa, respectively), as well as a lower molecular doublet the size of C3d (31 and 33 kDa, Figure 4Ba). Beads coated with laminarin had low levels of these C3 fragments (Figure 4Ba). Antibodies specific for the C3 beta chain revealed only the high molecular weight C3/C3b (Figure 4Bb). The 75 kDa chain of C3b or iC3b were not detected (Figure 4Bb).

To assess further the differences in C3 deposition on β-1,6-glucan as compared with β-1,3-glucan, serum was preincubated with soluble pustulan or laminarin before using it to opsonize beads coated with pustulan. Preincubation of the serum with soluble pustulan eliminated phagocytosis (Figure 5A, compare Ac and Aa) and ROS production (Figure 5B, compare red and blue) by neutrophils, suggesting that serum C3 was titrated out by the soluble β-1,6-glucan. Serum that was preincubated with an equivalent amount of soluble laminarin still mediated phagocytosis (Figure 5A, compare Ab and Aa) and ROS production (Figure 5B, compare red and green), suggesting that β-1,3-glucan did not efficiently block the interaction of C3 with beads coated with pustulan.

Western analysis revealed that preincubation with soluble pustulan eliminated most of the C3/C3b and C3d deposition on the pustulan-coated beads (Figure 5C, compare 1 to 3). Preincubation with laminarin eliminated the low molecular weight C3d doublet, but retained the high molecular weight C3/C3b (Figure 5C, compare 2 to 3), suggesting that the remaining C3/C3b was mediating phagocytosis and induction of ROS (Figures 5A and 5B). Preincubation of serum with soluble pustulan but not laminarin reduced the killing of Candida albicans by 6-fold (Figure 5D). This inhibition by soluble β-1,6-glucan is consistent with the idea that complement deposition on β-1,6-glucan is required for efficient killing of the fungus.

Because complement receptor 3 (CR3) is known to mediate phagocytosis of opsonized yeast and the yeast cell wall preparation zymosan, as well as ROS production (Cain et al., 1987; Ross et al., 1985), we surmised that CR3 could mediate phagocytosis of beads coated with β-1,6-glucan. Anti-human CR3 blocking antibodies reduced the extent of phagocytosis (Figure 6A, compare Ab to Aa) and ROS production (Figure 6B, compare red to green), suggesting that CR3 recognized C3b proteolytic fragments (C3d) that are deposited on particulate β-1,6-glucan.

β-1,6-glucan also plays an important role in the neutrophil’s recognition of whole fungal cells, which contain other polysaccharides such as chitin and β-1,3-glucan as well as many cell surface proteins. Candida albicans cells were heat killed to expose their masked β-glucan, and one sample was digested with an endo β-1,6-glucanase, and the other was incubated without the enzyme. Both samples were opsonized (see Experimental Procedures) and tested for their reactivity. The cells exposed to the enzyme (as compared with those that were not) exhibited an ∼50% reduction in phagocytosis (Figure 7A) and ROS production (Figure 7B), as well as expression of HSPs (Figure 7C). As EM93, a feral progenitor of many Saccharomyces cerevisiae laboratory strains, has exposed β-glucan (I.R.-B and G.F., unpublished data), it was possible to test the effect of endo-β-1,6-glucanase treatment on recognition without heat killing. Live Saccharomyces treated with enzyme was also less effective than the untreated strain (Figure S2), suggesting that, in fungi, β-1,6-glucan is an important component of fungal recognition.

The Candida albicans strain was the commonly used laboratory strain CAF2-1. Cells were grown on standard media (YPD), as described (Sherman, 1991).

We used overnight cultures in all experiments (about 3 × 108 cells/ml), because we found the Candida population to be more homogenous (contain >99% yeast form cells) than in midlogarithmic phase.

In order to test how neutrophils recognize Candida and to avoid any alteration of the fungus by media or neutrophils or manipulation of neutrophils by the fungus, we inactivated the Candida cells by UV, which kills the cells but does not alter the fungal cell wall structure (Wheeler and Fink, 2006). When indicated, Candida was heat killed as described (Wheeler and Fink, 2006).

A pool of fresh human serum was generated from ten healthy volunteers and was used in all experiments. Candida cells or beads were preopsonized in 50% serum in Phosphate-Buffered Saline without calcium chloride and without magnesium chloride (PBS) (GIBCO) for 15 min at 37°C on a mixer. Cells or beads were then incubated on ice for 5 min, washed twice with 0.04 mg/ml of the protease inhibitor AEBSF (Sigma) in PBS, and then washed twice with PBS without AEBSF. Cells or beads were recounted after this treatment.

Neutrophils were mixed with Candida or beads at ratio of 1:5 neutrophil: target. Neutrophils were cultured with opsonized Candida or beads or alone in RPMI1640 at 37°C for 2 hr. For RNA extraction, the neutrophils were frozen in TRI reagent (MRC) at −80°C.

Total RNA was prepared following the TRI reagent protocol, except that for RNA precipitation it was incubated with isopropanol overnight at 4°C. First and second strand synthesis, in vitro transcription, hybridization, and scanning were done as described before (Rubin-Bejerano et al., 2003). The microarrays were GeneChip Human Genome U133A 2.0 array (Affymetrix).

Data sets were normalized and floored to 20. Ratios of expression from neutrophils cultured with Candida divided by that from the neutrophils alone control were calculated. Induced and repressed genes were defined as those with an expression ratio greater than two standard deviations from the mean for a given experiment. Only genes that were consistently induced or repressed in two experimental duplicates were considered as induced or repressed.

Total RNA was prepared following the TRI reagent protocol. cDNA was created using High Capacity cDNA Archive Kit (Applied Biosystems). Quantitative RT-PCR was done using TaqMan Gene Expression Assays (Applied Biosystems) and 7500 Real Time PCR system (Applied Biosystems), following the manufacturer protocol. The following TaqMan Gene Expression Assays were used: ACTB (Hs99999903_m1), DNAJB1 (Hs00428680_m1), HSPCB (Hs00607336_gH), and HSPH1 (Hs00198379_m1). Fold induction was calculated as the ratio of expression from neutrophils cultured with Candida over a neutrophil alone control, or neutrophils cultured with carbohydrate-coated beads over neutrophils cultured with untreated beads.

β-1,6-glucan was extracted from Candida albicans as described (Roemer et al., 1994). β-1,3-glucan was extracted from Candida albicans by digesting β-glucan (Roemer et al., 1994) with Chitinase (Sigma) and endo β-1,6-glucanase (Biomarin Pharmaceutical, Inc).

Laminarin (Sigma) is the standard for β-1,3-glucan, and pustulan (Calbiochem) is the standard for β-1,6-glucan (Hellerqvist et al., 1968; Lindberg and McPherson, 1954). Dextran (Fluka) is an α-1,6-glucan. Pustulan was cleaned as described (Lindberg and McPherson, 1954). Glucan from barley (Sigma) is composed of β-1,3-glucan (30%), and β-1,4-glucan (40%). When indicated, pustulan was specifically digested using an endo β-1,6-glucanase (Lora et al., 1995), a kind gift from Dr. Nick Zecherle (Biomarin Pharmaceutical, Inc). The reaction products were analyzed by gel filtration and thin-layer chromatography.

Twenty milligrams of Pustulan was applied to a BioGel P6 column (1.5X120 cm, BioRad, 200–400 mesh). The column was equilibrated in 0.1 M acetic acid and run at a constant flow rate of 15 ml/hr, 1.5 ml fractions were collected and carbohydrate was measured by the phenol-sulfuric acid method (Duboius et al., 1956). Fractions containing the peaks were dried twice for complete elimination of acetic acid and suspended in water at 10–20 mg/ml. Cytochrome C was used as a marker of the exclusion volume and glucose as a marker of the inclusion volume.

Samples (5 μl) were ascended twice on 20 cm silica gel 60 plates (Merck, 0.25 mm). The solvent system was n-butanol/ethanol/water (5:3:2). The samples and standards were visualized by heating the plates at 80°C after spraying with phenol-sulfuric acid. Standard gentiooligosaccharides were prepared from partial acid hydrolysate of pustulan as described (Magnelli et al., 2002).

Pustulan (3 ml, 15 mg/ml) was adjusted to 0.1 M NaOH, incubated for 1 hr at 37°C and dialyzed against water.

Laminarin (50 mg) was dried, and resuspended in 1.5 ml of acetic anhydride (Mallindcrodt). A few crystals of 4-dimethylaminopyridine (Avocado Research Chemist, Ltd) were added as catalyst. The reaction was allowed to proceed at room temperature for 20 min and stopped with 2 volumes of water. The sample was dialyzed against water.

Polybead polystyrene 6 μm microspheres (Polysciences, Inc.) (beads) were coated with polysaccharides as described before (Schlesinger et al., 1994). Pustulan tends to solidify at room temperature. We solubilized it in boiling water, and let it cool down to room temperature before applying to beads. We detected coating of the beads by the phenol -sulfuric acid method, which measures carbohydrates (Duboius et al., 1956). Only beads with 6-15 μg glucose per 1 ml (2 × 108 beads) were included in the analysis. We note that short polysaccharides did not coat the beads efficiently by this method. Moreover, by this method, acetylated glucans were coating the beads better than unacetylated glucans. When not acetylated, more of the glucan was added to the beads to achieve similar levels of coating.

Beads coated with an equivalent amount of pustulan or laminarin were opsonized and cultured with neutrophils for 40 min at 37°C. Images were taken every 10 s using a Nikon Eclipse TE-300, with a 100x DIC objective.

ROS production was assayed using DHR123 (Molecular Probes), which becomes fluorescent (rhodamine123) when oxidized (Conrads et al., 1999). 5 × 106 neutrophils were cultured with indicated beads at ratio of 1:5 in volume of 1 ml for 15 min at 37°C. 1 μl of DHR123 (D-23806) was added to 200 μl of the culture. Following incubation at room temperature for 30 min the cells were assayed by Fluorescence Activated Cell Sorting (FACS). In Figure 3, 5, and 6, an overlay of the signal from highly phagocytic cells is presented.

Beads were coated with equivalent amounts of β-1,3-glucan and β-1,6-glucan, and opsonized. Beads were suspended in 2% SDS 1 M ammonium hydroxide buffer, and incubated at 37°C for 1 hr. The supernatant representing proteins extracted from beads was loaded on 4%–20% acrylamide SDS gel. The gel was stained with silver stain, and bands were cut for analysis by mass spec.

C3 deposition was assayed using monoclonal antibodies directed against the alpha or the beta chain of C3b (RDI Reasearch Diagnostics).

Viability of Candida albicans was determined using XTT as described before (Meshulam et al., 1995).

Serum was preincubated with equivalent amount of soluble laminarin or pustulan for 5 min at 37°C. The serum was then used to opsonize pustulan-coated beads as described above.

Neutrophils were preincubated with anti human Mac-1 or isotype control IgG (Bender Med Systems) for 30 min on ice before adding to opsonized pustulan-coated beads.