Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Pneumocystis carinii pneumonia remains a major cause of morbidity and mortality in individuals with compromised immune function (1). Previously thought to be a protozoan parasite, recent studies indicate that P. carinii is a fungal pathogen based on ribosomal RNA sequence, enzyme biochemistry, and the demonstration that P. carinii possesses a -glucan-enriched cell wall (4, 5). -glucans are carbohydrate polymers consisting of (1,3)-linked -D-glucopyranosyl residues with (1,6)-linked -D-glucopyranosyl side chains of varying length and distribution frequency (6). In addition to providing structural cell wall integrity, -glucans also strongly induce the release of inflammatory mediators from alveolar macrophages, thereby promoting lung inflammation and injury (9, 10). We recently described isolation of a P. carinii cell wall fraction that is composed of complex carbohydrates, mainly -glucans, and largely depleted of protein (11). This -glucan cell wall fraction strongly induces lung inflammation in vivo and is a potent stimulant of tumor necrosis factor (TNF)- and macrophage inflammatory protein (MIP)-2 production by alveolar macrophages (10, 11).

A growing body of evidence implicates nonimmune factors as being important determinants of host response to P.  carinii (12, 13). P. carinii interacts with a variety of host proteins in the alveolar spaces, including immunoglobulins (Ig), fibrinogen, and the serum glycoproteins vitronectin (VN) and fibronectin (FN) (12, 13). VN and FN are biochemically distinct proteins, which promote cellular adhesion, inflammation, and wound repair (14). Neese and colleagues (12) demonstrated that both VN and FN bind to P. carinii and are present in significantly increased quantities in the lower respiratory tract of patients with P.  carinii pneumonia. Increased quantities of VN and FN in the lung during P. carinii pneumonia may represent capillary leakage of serum proteins into the alveolar space or increased local synthesis of VN and FN in the lung. VN and FN are synthesized in increased quantities during acute and chronic inflammatory conditions (18). Hepatic production of VN and FN during systemic inflammatory states is predominantly controlled by interleukin (IL)-6 (20). Although elevation in local and systemic production of IL-6 has been demonstrated in severe combined immunodeficiency (SCID) mice with P. carinii infection, the exact role of this cytokine in the host response to P. carinii remains poorly understood (23).

FN has been previously shown to interact, at least in part, through the gpA surface glycoprotein of P. carinii, enhancing adherence of the organisms to lung epithelial cells and macrophages (24). VN also binds P. carinii and augments interaction of the organism to lung cells (25). VN binding to P. carinii occurs through the glycosaminoglycan-binding domain of VN. The specific surface components of P. carinii interacting with VN are currently not known, although VN has been reported to bind to -glucans present on the phylogenetically related fungus Saccharomyces cerevisiae (27). In addition to enhancing adherence of P. carinii to lung epithelial cells, incubation of P. carinii with VN or FN results in augmented alveolar macrophage activation and TNF- release in response to the organism (12).

Emerging evidence suggests that invertebrate organisms possess specific proteins that bind to fungal -glucans. These proteins have been termed -glucan binding proteins and are believed to act as pattern recognition molecules potently enabling innate immune recognition of invading fungal pathogens (28, 29). Although to date -glucan binding proteins have not been identified in mammals, we have observed that in vitro responses to whole P. carinii and -glucan are substantially enhanced in the presence of serum, suggesting that circulating factors such as VN and FN in mammalian serum may function as -glucan binding proteins.

The current investigation was therefore undertaken to address the following hypotheses. First, we postulated that VN and FN interact with glucan carbohydrate components of the P. carinii cell wall and thereby enhance macrophage responses to the organism, acting in an analogous fashion to -glucan binding proteins present in invertebrates. Second, we hypothesized that a potential mechanism by which IL-6 responses modulate host inflammation in P. carinii pneumonia is by increasing hepatic and local synthesis of both VN and FN during P. carinii pneumonia. Finally, we investigated whether P. carinii glucan induces IL-6 secretion from alveolar macrophages, providing a potent stimulus for host generation of VN and FN.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials

Endotoxin-free reagents were used in all experiments. General reagents were from Sigma Chemical Co. (St. Louis, MO) unless otherwise specified. P. carinii f. sp. (formae specialis) carinii was originally obtained through the American Type Culture Collection (Manassas, VA). L929 cells were purchased from American Type Culture Collection and fetal bovine serum was from GIBCO BRL (Rockville, MD). Cell culture grade rat plasma VN and FN were purchased from Sigma Chemical Co. and Calbiochem Inc., respectively. Murine recombinant TNF- was obtained through Genzyme Corporation (Cambridge, MA). Polyclonal rabbit antihuman VN and FN antibodies were procured from Chemicon International (Temecula, CA). S. cerevisiae -glucan was from Sigma.

Generation of a P. carinii Glucan-Rich Cell Wall Isolate

We recently reported the generation and characterization of a glucan-rich cell wall fraction from P. carinii (11). To isolate this component, P. carinii pneumonia was induced in Harlan Sprague- Dawley (HSD) rats (Indianapolis, IN) by immunosuppression with dexamethasone, as we previously described (9). Rats moribund with P. carinii pneumonia were killed, the lungs homogenized for 10 min, and P. carinii organisms were obtained by filtration through 10-µm filters. P. carinii organisms were autoclaved (120°C, 20 min), disrupted by ultrasonication (200 W for 3 min, six times), and washed in chloroform/methanol (2:1) for 2 h. The isolated product was additionally washed with 1 N sodium hydroxide, rinsed with deionized water until neutral, and finally treated with ethanol, acetone, and diethyl ether extractions to dehydrate and fully remove lipids from the products. Approximately 1 mg (equivalent to 10⁸ cell wall particles/ml) of P. carinii cell wall isolates were obtained from roughly 30 rats. The final cell wall fraction was found to contain 95.7% carbohydrate and 4.3% protein by weight using the orcinol-sulfuric acid method and bicinchoninic acid protein determinations, respectively (11). A sample of the P. carinii cell wall isolate was partially hydrolyzed with 2 M HCl, thereby releasing 82% of its content as D-glucose measured by the glucose oxidase method (11). Thus, the majority of this material represents P. carinii-derived glucose polymer, namely glucans. This product was previously shown to be a potent stimulant of macrophage release of TNF- and MIP-2, and was inactivated by digestion with -1,3-glucanase. Before use, the -glucan cell wall fractions were washed in 0.1% sodium dodecyl sulfate to ensure stripping of any residual protein contamination on the cell wall and then vigorously washed three times in distilled physiologic saline (10 ml, pH 7.0). The final washes were free of endotoxin as assessed by Limulus amebocyte lysate assay. Before use, these preparations were sonicated (three × 30 s) to yield a disperse suspension of particles and enumerated using a hemacytometer.

Isolation of Serum -Glucan Binding Proteins

The institutional review board approved all studies involving the use of animal or human subjects. Human serum was prepared from healthy volunteers. Rat serum was derived from uninfected female HSD rats. Rat and human sera were used within 7 d of preparation. To identify major proteins that bind fungal -glucans, we precipitated sera with glucan-rich cell wall preparations from P. carinii and S. cerevisiae. Human, rat, and fetal bovine sera (1 ml each) were incubated with P. carinii -glucan (approximately 1 × 10⁷ particles) or S. cerevisiae -glucan (2.5 mg/ml serum) for 16 h at 4^oC with shaking. At the end of the incubation, the sera were centrifuged for 10 min (2,200 × g), and the pellet was washed three times in 50 mM Tris, 150 mM NaCl, pH 8.0. The proteins precipitated by the fungal glucan preparations were eluted by boiling in Laemmli buffer and subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions on 10 to 12% gradient porosity gels. Protein bands were identified on the gels by silver staining using a commercially available system (Bio-Rad, Hercules, CA).

Immunodetection of VN and FN as Serum -Glucan Binding Proteins

We postulated that VN and FN were major constituents among serum proteins precipitated by fungal -glucans. To demonstrate this, serum proteins precipitated by S. cerevisiae and P. carinii cell walls were fractionated on 12% SDS-PAGE gels and transferred to nitrocellulose membranes (100 V × 1 h). The membranes were subsequently washed in Tris-buffered saline (TBS) containing 0.05% Tween, and nonspecific binding sites were blocked by overnight incubation with TBS containing 5% dried milk at room temperature. Immunoblotting was performed using specific polyclonal rabbit antibodies to VN and FN (Chemicon, Inc.). The membranes were incubated with the primary antibodies (1:250 dilution) in TBS for 2 h and then washed extensively. Nonimmune rabbit IgG was used as a control at identical dilutions. Bound antibody was detected using a commercial immunodetection system (Bio Rad).

Effect of Serum, VN, and FN on Fungal -Glucan-Induced TNF- Release from Alveolar Macrophages

Prior investigations indicated that TNF- production by alveolar macrophages in response to P. carinii and other fungi is enhanced by incubation with either nonimmune or immune serum. Therefore, we next determined whether the presence of serum, or the specific serum proteins FN and VN, augments alveolar macrophage inflammatory responses to fungal cell wall -glucans. Normal alveolar macrophages were obtained by whole lung lavage of HSD rats, using 50 ml of sterile physiologic saline. After centrifugation (400 × g for 10 min), the cells were suspended in mixed media (1:1 of RPMI and medium 199 containing 2 mM L-glutamine, 10,000 U penicillin/liter, 1 mg streptomycin/liter, and 25 µg/ml amphotericin B/liter) and counted using a hemacytometer. Smears consistently showed more than 95% macrophages in the lavage. A total of 2 × 10⁵ macrophages was plated per well in 96-well tissue culture plates, allowed to adhere for 60 min, and gently washed to remove any unattached cells. Subsequently, the macrophages were incubated with the P. carinii glucan cell wall fraction (1 × 10⁷ particles/ml of medium) either in the presence or absence of fetal bovine serum (final concentration, 10%). Parallel experiments were conducted with macrophages stimulated with 100 µg/ml of S. cerevisiae glucan in the presence or absence of fetal calf serum (10% serum). After 8 to 12 h of incubation, the media were assayed for TNF-.

To test the effect of incubation of particulate -glucan with VN and FN, additional experiments were performed in which the P. carinii cell wall fraction (1 × 10⁷ particles/ml) or S. cerevisiae glucan (100 µg/ml) was incubated with either VN or FN at a final concentration of 100 µg/ml or with saline alone (control) for 2 to 4 h at 37°C. After coating with VN or FN, the P. carinii glucan-rich cell wall fraction was washed with 10 ml of saline to remove unbound protein and incubated with 2 × 10⁵ alveolar macrophages to assess macrophage TNF- release.

Quantification of TNF-

Immune reactive TNF- was quantified using a commercial enzyme-linked immunosorbent assay (ELISA) (Genzyme Corp.). Additionally, the bioactivity of TNF- was further analyzed by a sensitive L929 cytotoxicity assay. Murine L929 monolayers were plated on 96-well plates (4 × 10⁴ per well) and incubated overnight to achieve confluence. The following day, media were replaced by Earle's minimum essential medium containing 10% fetal calf serum and 1 µg/ml actinomycin D. Recombinant murine TNF- (Genzyme Corp.) was used as a relative standard of bioactivity. Standards or samples were plated onto the L929 cells, and the cells were incubated for 18 h at 37°C. Subsequently, media were removed, cells were stained with 0.5% crystal violet, and viability was measured by determining the optical density at 490 nm. A standard curve was constructed from the standards and used to estimate TNF- in the samples. There were no discernible differences in values of TNF- using either the bioassay or ELISA.

Analysis of Host Tissue VN and FN Steady-State Messenger RNA Content

Prior investigations demonstrated that VN and FN accumulate in the lower respiratory tract during P. carinii pneumonia. Accordingly, we sought to determine if increased local transcription of VN and FN messenger RNA (mRNA) occurred in lung during P. carinii pneumonia. In addition, because the liver is a predominant source of these circulating serum proteins, we further evaluated whether there was an increase in hepatic expression of VN and FN mRNA during P. carinii pneumonia. Total RNA was extracted from the lung and liver of normal female rats and P. carinii-infected rats using TriZol (GIBCO BRL), and quantified spectrophotometrically at 260 nm. A total of 10 µg of total RNA from both liver and lung samples was separated on 1% agarose gels and transferred to nitrocellulose. The relative quantities of steady mRNA for each gene of interest was assessed by Northern blot analysis. Rat VN and FN mRNA levels were detected with specific complementary DNA (cDNA) probes generated from rat genomic DNA using primers for rat VN and FN. For rat VN, the primer sequences used were 5'-CTG GCT GTT TTG ACA ACG GG-3' and 5'-GCT CCA TGT GTC TCC AAT TCT GTA G-3'. Rat FN primers used were 5'-AAG AGG CAG GCT CAG CAA ATC GTG-3' and 5'-ATT GGC TTG CAG GTC CAT TCC C-3'. Rat VN (708 bp) and FN (600 bp) cDNA probes were synthesized using 30 cycles of polymerase chain reaction (PCR) with the following reaction mixtures: 5 µl 10 × PCR buffer (GIBCO BRL), 25 mM MgCl₂ (1.5 µM final concentration), 36.5 ml water, 1 µl deoxynucleotide triphosphate mix (0.2 µM final concentration), 1 µl of both forward and backward primers (0.2 µM final concentration), 2 µl of rat cDNA, and 1 µl (2.5 U) of Taq polymerase enzyme (GIBCO BRL). PCR (for both VN and FN cDNA synthesis) was performed with denaturing at 94°C for 30 s, annealing at 55°C for 45 s, and extension at 72°C for 1 min for 30 cycles. ³²P end-labeling of the probes was performed with RadPrime labeling kit (GIBCO BRL) according to the manufacturer's instructions.

IL-6 Release from Alveolar Macrophages Stimulated with the P. carinii Glucan-Rich Cell Wall Isolate

Because IL-6 is a major regulator of VN and FN synthesis, we next determined whether the isolated glucan cell wall component of P. carinii would directly activate macrophage IL-6 release. To accomplish this, alveolar macrophages were collected by whole lung lavage from pathogen-free HSD rats. After centrifugation (400 × g for 10 min), recovered cells were suspended in mixed medium. A total of 2 × 10⁵ macrophages was plated per well in 96-well tissue culture plates, allowed to adhere for 60 min, and gently washed to remove any unattached cells. Subsequently, varying concentrations of the P. carinii cell wall isolate (1 to 5 × 10⁶ particles/ml) were incubated with the alveolar macrophages for 12 h (37°C, 5% CO₂). After incubation, the medium was removed, centrifuged (10,000 × g for 5 min) to remove any particulate material or cells, and assayed for the presence of soluble IL-6 using a commercially available ELISA (Biosource International, Camarillo, CA).

Statistical Analysis

Data are expressed as mean ± standard error of the mean (SEM). Differences between experimental and control data groups were determined using two-tailed Student's t tests for normally distributed variables. Nonparametric parameters were assessed using the Mann-Whitney U test. Statistical testing was performed on the Statview II statistical package (Abacus Concepts, Inc., Berkeley, CA). Statistical differences between groups were considered significant if P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Serum Enhances Alveolar Macrophage Generation of TNF- in Response to Fungal -Glucan

The -glucan-rich cell wall preparations from S. cerevisiae and P. carinii induced TNF- release from cultured alveolar macrophages, which was significantly augmented by the presence of serum (Figure 1). Alveolar macrophages stimulated with -glucan from S. cerevisiae (100 µg/ml) in the absence of fetal calf serum released 3,823.1 ± 275.9 pg/ ml of TNF-, whereas in the presence of serum, macrophages released 5,753.2 ± 212.9 pg/ml of TNF- (P = 0.004). Similarly, macrophages stimulated with -glucan- rich P. carinii cell wall preparations (1 × 10⁷ particles) generated 6,314.8 ± 37.7 pg/ml of TNF- in the absence of serum compared with 8,993.6 ± 294.3 pg/ml of TNF- in the presence of serum (P = 0.006).

View larger version (11K): [in this window] [in a new window]   View larger version (9K): [in this window] [in a new window]   Figure 1.   Alveolar macrophage release of TNF- in response to -glucan is augmented by serum. (A) Macrophages (2 × 105) were stimulated with -glucans derived from S. cerevisiae (100 µg/ml) in the absence or presence of fetal calf serum (10%). After 12 h of incubation, TNF- release was measured using the L929 bioassay (*P < 0.005 compared with macrophages stimulated without serum). (B) Similarly, alveolar macrophages (2 × 105) were stimulated with the glucan-rich P. carinii cell wall isolate (1 × 107 particles) in the absence or presence of fetal calf serum (10%) (*P < 0.01 compared with macrophages stimulated without serum). In both cases, the presence of serum significantly enhanced macrophage TNF- generation in response to glucan cell wall components. Bars represent the mean ± SEM of four determinations.

It was highly unlikely that the observed TNF- release from macrophages was due to contaminating bacterial endotoxin in the -glucan preparations interacting with lipopolysaccharide binding protein in serum. Extensive measures were undertaken to eliminate any potential contamination of the fungal cell wall preparations with endotoxin (11). All cell wall isolates were washed multiple times in sterile physiologic saline, washed once with a solution of the endotoxin chelating agent polymyxin B (10 µg/ml), and assayed for the presence of endotoxin using a Limulus amebocyte lysate assay. After polymyxin B and multiple large volume saline washes, the detectable levels of endotoxin in the cell wall preparations were consistently < 10 pg/ml, a level of endotoxin incapable of stimulating this degree of macrophage activation. Thus, the observed serum-associated augmentation of TNF- release by macrophages challenged with fungal -glucan is likely due to the presence of serum factors or proteins that enhance alveolar macrophage activation in response to -glucan.

Serum VN and FN Bind Fungal -Glucans

These observations prompted us to search for serum proteins that bind to -glucan and participate in host responses to these fungal components. To address this, we precipitated glucan-binding proteins by incubating freshly obtained sera with the glucan-rich P. carinii cell wall isolate or with S. cerevisiae -glucan. Two predominant protein bands were isolated in both precipitations (Figure 2). These consisted of a high molecular weight protein of approximately 220 kD in size and another major band migrating at roughly 70 kD. Similar bands were isolated from precipitations using human, rat, and bovine sera.

View larger version (40K): [in this window] [in a new window]   View larger version (25K): [in this window] [in a new window]   Figure 2.   Fungal -glucans precipitate binding proteins from mammalian sera. (A) Human, bovine, and rat sera were incubated with the P. carinii -glucan cell wall preparation, washed extensively, eluted in Laemmli buffer to remove attached proteins, and subjected to SDS-PAGE. Silver staining was performed to identify relevant protein bands. Two major protein bands were recovered from all serum samples tested with relative molecular weights of 220 and 65 to 75 kD. (B) Similar glucan binding proteins were precipitated from serum by incubation with S. cerevisiae -glucan.

The respective molecular sizes of these proteins precipitated by -glucan, and our prior studies of adhesive protein interaction with whole P. carinii (12) strongly suggested that these serum glucan binding proteins might represent FN and VN. To address this, immunoblotting was performed using specific polyclonal antibodies for these serum proteins (Figure 3). Immune analysis revealed the higher molecular weight protein (220 kD) to react with antibodies to FN, whereas the lower weight band (65 to 75 kD) reacted with antibodies to VN. Nonimmune antibodies gave no specific staining with the precipitated proteins. Abundant immunoreactive VN and FN were precipitated from sera using either the P. carinii -glucan cell wall isolate or -glucan derived from S. cerevisiae.

View larger version (27K): [in this window] [in a new window]   View larger version (27K): [in this window] [in a new window]   Figure 3.   VN and FN bind to fungal -glucans. (A) To determine if VN and FN bind to the P. carinii glucan, proteins precipitated by incubating bovine serum with the P. carinii cell wall preparations were separated on 12% polyacrylamide gels, transferred to nitrocellulose membranes, and analyzed by immunoblotting. Immunoblotting with polyclonal anti-VN and anti-FN antibodies, respectively, reveals the presence of immunoreactive VN and FN among P. carinii cell wall-precipitated proteins (arrows). In contrast, immunoblotting with nonimmune isotype-matched antibody does not reveal any specific reactivity. (B) Similarly, VN and FN bind to S. cerevisiae -glucan. Immunoblotting with polyclonal anti-VN and anti-FN antibodies, respectively, reveals the presence of immunoreactive VN and FN among S. cerevisiae -glucan-precipitated proteins (arrows). Nonimmune antibody does not reveal any specific reactivity.

The major bands precipitated by both the P. carinii and S. cerevisiae -glucan preparations represent VN and FN as demonstrated by the Western blots in Figures 3A and 3B. A small amount of additional material was precipitated by the P. carinii glucan-rich cell wall isolate. These appear as additional bands of ~ 100 kD in the anti-VN blot and as ~ 120 kD bands in the anti-FN blot. These materials reacted specifically with the anti-VN and anti-FN antibodies. In the VN blot, these minor bands potentially represent glycosylated or other alternatively modified VN. The additional FN-related bands of ~ 120 kD likely represent minor proteolytic fragments of intact FN. In either event, the major reactive proteins precipitated by both fungal -glucans appear to be VN and FN.

Binding of VN and FN to Fungal -Glucans Significantly Augments Macrophage TNF- Response

We previously reported that VN and FN bind to intact P. carinii, enhancing the ability of the organism to stimulate macrophage TNF- generation (12). Therefore, we next sought to determine whether VN or FN binding to isolated fungal -glucans would similarly augment macrophage TNF- release in response to these cell wall components. We observed that coating of S. cerevisiae -glucans (100 µg/ml) with either VN or FN (100 µg/ml each) significantly augmented TNF- release from alveolar macrophages (Figure 4). Pretreatment of S. cerevisiae -glucan with VN resulted in 1,764 ± 132 pg/ml of TNF- release compared with 1,258 ± 152 pg/ml in response to uncoated -glucan particles (P = 0.005). Binding of FN to S. cerevisiae -glucan particles similarly resulted in enhanced macrophage TNF- generation (1,481 ± 289 pg/ml TNF- compared with 1,024 ± 160 pg/ml with uncoated -glucan; P = 0.008).

View larger version (10K): [in this window] [in a new window]   View larger version (9K): [in this window] [in a new window]   Figure 4.   Coating of particulate -glucans with VN or FN significantly enhances alveolar macrophage TNF- release. (A) S. cerevisiae -glucan (1 mg/ml) was incubated with either VN or FN (100 µg/ml each) or buffer only. After 2 h, the -glucan preparations were washed with phosphate-buffered saline, sonicated to render particulate, and incubated with macrophages. After 12 h, the culture media were assayed for TNF-. Coating of S. cerevisiae -glucan with either VN or FN resulted in significant enhancement of TNF- release (* P = 0.01 compared with macrophages stimulated with uncoated -glucan). (B) Parallel experiments were conducted using glucan-rich P. carinii cell wall preparations (1 × 107 particles/ml). Similarly, coating of P. carinii glucan particles with either VN or FN was associated with an increase in observed TNF- release from the cultured macrophages (*P < 0.05 and **P < 0.01 compared with macrophages stimulated with uncoated P. carinii cell wall glucan preparations). Bars represent the mean ± SEM from six determinations.

In a parallel manner, coating of P. carinii glucan cell wall isolate with VN was also associated with an increased TNF- liberation from macrophages. Alveolar macrophages stimulated with VN-coated P. carinii glucan isolate released 16.34 ± 0.5 ng/ml of TNF- compared with 9.0 ± 1.2 ng/ml using the uncoated P. carinii glucan isolate (P = 0.007). In addition, coating of P. carinii cell wall glucan isolate with FN enhanced macrophage TNF- release to 12.6 ± 0.3 ng/ml compared with 9.0 ± 1.2 ng/ml after stimulation with uncoated P. carinii glucan preparations (P = 0.026). Thus, interaction of VN and FN with -glucan components of fungal cell walls enhances macrophage inflammatory activation.

Expression of VN in Liver and Lung during P. carinii Pneumonia

Because the liver is the predominant source of plasma VN, we analyzed steady-state hepatic mRNA in rats with P. carinii pneumonia in comparison to uninfected control rats (Figure 5). During P. carinii pneumonia, we observed a marked upregulation of hepatic VN mRNA abundance. In contrast, there was minimal increase in lung VN mRNA in rats with P. carinii pneumonia. Although slight increases in lung VN mRNA levels were observed after prolonged exposure of the autoradiographs, the difference in expression between normal and infected lungs was very slight. These data indicate that the predominant source of VN in the lower respiratory tract during P. carinii pneumonia originates predominantly from circulating VN of hepatic origin. This is consistent with prior observations that VN found in other tissues is most likely to represent plasma-derived protein rather than locally synthesized VN (20, 30).

View larger version (36K): [in this window] [in a new window]   Figure 5.   Vitronectin steady-state mRNA expression is enhanced in the liver during P. carinii pneumonia. Total RNA was extracted from lung and liver of rats with overt P. carinii pneumonia as well as uninfected normal rats. VN mRNA levels were assessed by Northern blot analysis using a specific radiolabeled cDNA probe for rat VN. Equal RNA loading was verified by probing for the -actin housekeeping gene. The liver of P. carinii-infected rats demonstrated significantly increased levels of VN mRNA transcripts when compared with the liver of control rats. Minimal differences in lung VN mRNA levels were detected between infected and control rats. Shown is a blot representative of three separate experiments.

Expression of FN in Liver and Lung during P. carinii pneumonia.

Although plasma FN is mainly of hepatic origin, several lines of evidence indicate that local synthesis of FN can occur in the lung during inflammatory states (31). Northern blot analysis of total RNA obtained from P. carinii-infected rat lung and uninfected controls demonstrated that while minimal synthesis of FN occurs in the uninfected lung, there is substantial increase in FN steady-state mRNA abundance during P. carinii pneumonia (Figure 6). Similarly, analysis of total hepatic RNA revealed increased hepatic FN mRNA during P. carinii pneumonia compared with control liver (Figure 6). These data, therefore, suggest that the increased amounts of FN detected in the lower respiratory tract during P. carinii pneumonia are likely related both to increased hepatic synthesis as well as to increased local elaboration of FN.

View larger version (39K): [in this window] [in a new window]   Figure 6.   FN steady-state mRNA increases in both liver and lung during P. carinii pneumonia. Total RNA was extracted from lung and liver of rats with overt P. carinii pneumonia as well as normal animals. FN mRNA levels were measured by Northern blot analysis using a specific radiolabeled cDNA probe for rat FN. Equal RNA loading was confirmed by analysis of 28S and 18S ribosomal RNA (rRNA) subunits. During P. carinii pneumonia, FN levels of mRNA increase significantly in both liver and lung. Shown is a blot representative of three separate experiments.

P. carinii Cell Wall -Glucan Is a Potent Stimulant of IL-6 Generation by Alveolar Macrophages

Finally, to determine potential mechanisms by which increased systemic production of VN and FN occurs during P. carinii pneumonia, we assessed whether the glucan-rich P. carinii cell wall isolate would induce IL-6 release from alveolar macrophages (Figure 7). Macrophages were obtained from pathogen-free rats and stimulated with varying concentrations of the P. carinii -glucan preparation (Figure 7). Although unstimulated macrophages released only 14.5 ± 19.2 pg/ml of IL-6 into the medium, alveolar macrophages incubated with > 1.0 × 10⁶ particles/ml of the P. carinii glucan cell wall isolate generated substantially increased amounts of IL-6. Maximally, macrophages incubated with 5.0 × 10⁶ particles/ml produced 368.2 ± 18.3 pg/ml of IL-6 (P = 0.0001 compared with control). Thus, P. carinii cell wall -glucan is a potent stimulus of IL-6 generation by alveolar macrophages. In addition, as observed by others, we documented enhanced release of circulating IL-6 during the course of P. carinii pneumonia. P. carinii-infected rats (n = 14) exhibited substantially higher circulating IL-6 levels of 92.2 ± 28.1 pg/ml compared with healthy uninfected rats (n = 8) that exhibited 5.0 ± 1.7 pg/ml of circulating IL-6 (mean ± SEM; P = 0.0004). Taken together, these data support a potential role for IL-6 mediating systemic inflammatory responses during P. carinii infection.

View larger version (10K): [in this window] [in a new window]   Figure 7.   P. carinii glucan cell wall extract induces IL-6 release from macrophages. Alveolar macrophages (2 × 105) were stimulated with increasing amounts of the -glucan-rich P. carinii cell wall isolates (1 to 5 × 106 particles/ml). After 12 h of incubation, IL-6 release in the medium was measured by ELISA. P. carinii glucan-rich cell wall isolate stimulated substantial release of IL-6 from the macrophages. (*P = 0.01, **P = 0.0001 compared with unstimulated control macrophages). Bars represent the mean ± standard deviation of four determinations.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Our investigation demonstrates that serum components significantly enhance the ability of alveolar macrophages to produce TNF- in response to -glucan derived from P. carinii and from the phylogenetically related fungus S. cerevisiae. The effects of serum on macrophage recognition of -glucan are related, in part, to interactions of circulating VN and FN with these cell wall components. In addition, isolates of P. carinii rich in -glucan stimulate alveolar macrophage generation of IL-6, a pleiotropic cytokine that promotes VN and FN expression during states of inflammation (27, 28). Whereas increased levels of VN observed in the lungs during P. carinii pneumonia occur predominantly through enhanced hepatic secretion of VN, accumulation of lung FN is related both to hepatic and pulmonary synthesis of FN. Through interaction with cell wall -glucan, VN and FN in turn enhance alveolar macrophage recognition and inflammatory response to this organism.

Previous studies demonstrate that FN binds to the gpA glycoprotein complex present on the surface of P. carinii through the Arg-Gly-Asp amino-acid sequence of FN (26). However, FN additionally contains a glycosaminoglycan (heparin) binding domain, capable of interacting with complex carbohydrate conjugates in a fashion similar to VN (26). Previous studies also indicate that the P. carinii component that binds VN is compatible with a complex carbohydrate consistent with fungal -glucan (25, 26, 32). It remains possible that VN and FN will further interact with other P. carinii epitopes, modifying additional facets of host-microbial interaction.

Among the numerous inflammatory cytokines induced during P. carinii infection, TNF- exerts particularly critical activities. The importance of TNF- in host defense against this infection is underscored by multiple studies demonstrating delayed clearance of P. carinii infection and increased mortality in animals lacking functional TNF- or its cognate receptors (33). The mechanisms by which TNF- facilitates clearance of P. carinii are not fully understood, though TNF- promotes expression of intercellular adhesion molecule (ICAM)-1 during P. carinii pneumonia (36). ICAM-1 strongly potentiates activation of natural killer cells, CD8⁺ T lymphocytes, and B lymphocytes, and is involved in the regulation of class I major histocompatibility complex expression (37). Additional studies suggest that TNF- may further exert a direct toxic effect on P. carinii (40).

In previous studies, the ability of P. carinii or isolated P. carinii cell wall to stimulate TNF- release from alveolar macrophages was reduced by digestion of -glucan components (9, 11). Furthermore, intact P. carinii initially stripped of VN and FN had reduced ability to trigger TNF- release from macrophages compared with organisms coated with these matrix proteins (12). Our current data suggest that this enhanced response of macrophages may in part be due to binding of VN and FN to cell surface -glucan on the organism. In addition, it is known that FN also binds to gpA on P. carinii (41). We propose that binding of VN and FN to epitopes on P. carinii enhances macrophage recognition and response to this organism. Alternate explanations remain possible. For instance, coating of VN and FN on the organism's surface might promote independent ligation of integrin matrix receptors and receptors for fungal -glucan also resulting in enhanced macrophage release of TNF-.

Fungal -glucan-induced TNF- secretion by macrophages is substantially augmented when alveolar macrophages are stimulated in the presence of serum components. This observation is analogous to the effect of serum on lipopolysaccharide-induced mononuclear cell activation through the activities of lipopolysaccharide binding protein (LBP) and cell surface CD14 receptors (42, 43). Prior investigations further reveal that invertebrates such as the freshwater crayfish Pacifastacus leniusclus and the insects Blaberus craniifer and Bombyx mori possess other serum proteins that bind to fungal -glucans and enhance host defense responses, such as activation of the prophenoloxidase system (28, 29, 44). Both LBP and glucan binding proteins represent pattern-recognition proteins that identify repeating structures of carbohydrate (and lipid conjugates in the case of LBP) derivatives on pathogenic microbes as "non-self" and trigger immune activation. Our study suggests that the serum proteins VN and FN may act in an analogous fashion to these pattern-recognition proteins by binding cell surface -glucans and enhancing immune activation to this fungal cell wall component.

The increased macrophage activation due to VN and FN interactions with -glucan was somewhat less in magnitude than that associated with whole serum. Other serum factors including interferon-gamma, nonimmune antibodies, and possibly, other glucan binding proteins, also likely enhance macrophage activation under these conditions. Furthermore, the increased production of TNF- induced by VN or FN binding to -glucan particles is not related to contaminating endotoxin. Macrophages incubated with up to 100 µg/ml VN or FN exhibited no augmentation of basal TNF- release in our hands. Prior studies have also consistently demonstrated no induction of inflammatory cell cytokine release by VN or FN (45, 46).

Although TNF- substantially aids in the elimination of P. carinii, excessive quantities of this cytokine are also deleterious to respiratory function of the host (47). TNF- induces endothelial cell permeability, promotes edema formation, and enhances exuberant inflammatory cell recruitment in the lungs, each of which has been associated with further impairment of gas exchange (48). The net host effect of enhanced macrophage inflammatory activation as a consequence of VN and FN interactions with P. carinii is not well understood. In addition to enhancing inflammatory responses, the interaction between VN and FN with P. carinii has other important activities during the development of pneumonia. For instance, VN and FN are well known to enhance attachment of fungal organisms to the alveolar epithelium and have been postulated to promote life cycle completion of the organism (26, 32). In addition, FN and VN both increase attachment of P. carinii to alveolar macrophages, and VN (but not FN) has been documented to enhance phagocytosis of coated cells (41, 52).

Previous studies in P. carinii-infected SCID mice demonstrate increased levels of IL-6 both in the lungs as well as in the circulation (23, 53). The interaction of P. carinii with alveolar cells has been further shown to stimulate epithelial production of this cytokine (54). However, the overall role of IL-6 in host defense against P. carinii remains incompletely understood. Indeed, treatment of P. carinii- infected SCID mice with neutralizing antibodies to IL-6 did not significantly inhibit the ability of these animals to clear infection, but did strongly influence production of specific antibodies, as well as the relative numbers of neutrophils and lymphocytes infiltrating the lung (23). These data suggest that although IL-6 may not be essential for clearance of P. carinii organisms, this cytokine may possess important roles in the regulation of inflammatory responses to this pathogen.

IL-6 further represents an important mechanism controlling host expression of VN and FN (20, 21, 55). Increased hepatic synthesis of FN and VN accompanying acute phase responses during several infectious conditions are largely regulated by IL-6 activity (20, 57). Our study demonstrates that the glucan-rich P. carinii cell wall induces substantial IL-6 secretion by alveolar macrophages, suggesting that recognition of complex carbohydrate epitopes of the organism contributes to the induction of VN and FN during this infection as well. In addition, a recent study has also shown that IL-6 induces fibrinogen synthesis by lung epithelium during P. carinii pneumonia, which may further modulate lung inflammation (13).

The current investigation suggests that IL-6 drives accumulation of VN and FN in the lung by enhanced hepatic synthesis and subsequent leakage of these serum proteins into the damaged alveolar spaces. The data further support local synthesis of FN in the lung during P. carinii infection. Regardless of their sources, VN and FN bind to -glucan components of the fungal cell wall, enhancing macrophage recognition and inflammatory host responses during development of P. carinii pneumonia.