Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Human cytomegalovirus (CMV) infection is a major source of morbidity and mortality in congenitally and perinatally infected infants as well as in immunosuppressed patients, e.g., transplant recipients and human immunodeficiency virus (HIV)-infected individuals (1). Active CMV infections may present with myriad clinical manifestations including life-threatening interstitial pneumonia, which is very common in perinatally infected infants and transplant recipients. CMV has the ability to infect and to replicate in various cell types. Although embryonal lung fibroblasts are a main target for virus replication in vitro, CMV in vivo mainly infects monocyte/macrophages and endothelial, epithelial, and microglial cells (2). In the lung, in addition to fibroblasts, alveolar macrophages and, most importantly, type II pneumocytes have been found to be infected (3).

Surfactant protein (SP)-A is the most abundant protein of pulmonary surfactant (for review, see References 4 and 5) and is produced and secreted by type II pneumocytes. SP-A has been shown to play a key role in the innate immune defense against various pathogens. Very recently it was found in SP-A-deficient mice that whereas growth, reproduction, and lung function are normal, infection with group B streptococcus is generally more severe compared with wild-type mice (6, 7). In vitro studies demonstrated that SP-A may bind specifically to some viruses, bacteria, fungi, and parasites, such as herpes simplex virus, influenza A virus, serum-opsonized Staphylococcus aureus, gram-negative Escherichia coli, Klebsiella, Aspergillus fumoigatus, and Pneumocystis carinii. It can increase phagocytosis by opsonizing these agents (5, 8). Moreover, SP-A facilitates attachment of Mycobacterium tuberculosis to alveolar macrophages and stimulates infection (12). Also, SP-A has been shown to have antiviral activity by reducing the infectivity of influenza A (not of influenza B virus), mumps virus, and semliki forest virus (15). On the other hand, entry of HIV-1 into alveolar macrophages and HIV-1 replication in these cells were not affected (16). Thus, depending on the pathogen, SP-A may facilitate infection or pathogen elimination in the lung.

The aim of this study was to investigate whether SP-A plays a role in human CMV infection of the lung. The interaction between SP-A and human CMV proteins was studied in in vitro experiments. Using flow-cytometry analysis and standard plaque assay we addressed the questions of whether SP-A binds to viral proteins exposed on the surface of infected human embryonal lung fibroblasts and whether it affects human CMV replication in embryonal lung fibroblasts.

To evaluate the influence of SP-A on virus uptake into pneumocytes and alveolar tissue macrophages (ATM), we used rat CMV and rat lung cells because primary human lung cells were not available. We chose the rat CMV model because pathogenesis of human and rat CMV infection of the lung are very similar (17). We used flow cytometry and confocal laser fluorescence microscopy to examine the influence of SP-A on binding and internalization of rat CMV by rat type II pneumocytes and ATM.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Preparation of Pulmonary SP-A

Sheep and bovine SP-A were purified as described by Hawgood and colleagues (18) from lavages of lungs obtained from the local slaughterhouse. The purity of each SP-A batch was analyzed by sodium dodecyl sulfate (SDS)/polyacrylamide gel electrophoresis and Western blot analysis using polyclonal anti-SP-A antibodies as described by Wissel and associates (19). Functionality of the SP-A preparations was tested by liposome aggregation, surfactant secretion assay, and liposome uptake assays, as described recently (19). As tested with antihuman or antisheep immunoglobulin (Ig) G the amount of contaminating IgG was << 1%. Contamination of SP-A preparations by endotoxin was investigated using the quantitative chromogenic limulus Amebocyte Lysate Kit (COATEST plasma endotoxin) from Chromogenix AB (Mölndal, Sweden). Endotoxin concentration was determined as described by the manufacturer. In different SP-A preparations, up to a maximum of 0.4 pg endotoxin/µg SP-A was detected. Purified human alveolar proteinosis SP-A and fluorescein isothiocyanate (FITC)-labeled human SP-A were a gift from Dr. J. F. F. van Iwaarden (Free University Amsterdam, The Netherlands). Preparation and characterization of SP-A have been described by van Iwaarden and coworkers (20). SP-A was biotinylated using sulphosuccinimidobiotin (Sulfo-NHS-Biotin; Sigma, Steinheim, Germany) at a 5:1 molar ratio of biotin:SP-A for 3 h at room temperature. Thereafter it was extensively dialyzed against phosphate-buffered saline (PBS) at 8°C.

Cells and Viruses

Type II cells and ATM were isolated from the lungs of adult male Wistar rats (body weight 120 to 140 g) according to Dobbs and colleagues (21). Rat alveolar macrophages were separated from bronchial lavage fluid by ficoll/paque (Lympholyte-M; Cedarlane Laboratories Ltd., Hornby, ON, Canada) gradient centrifugation and repeated washes with PBS.

Human embryonic lung fibroblasts (HELF) were maintained in mimimum essential medium/Eagle medium (Biochrom, Berlin, Germany) supplemented with 7.5% fetal calf serum (Biochrom). Rat embryonal fibroblasts were prepared from 17-d old rat embryos using the methods described by Bruggeman and associates (22). For human and rat CMV viruses we used the strains AD169 and RA67 ("Maastricht strain"; kindly provided by C. Bruggeman, Maastricht, The Netherlands), respectively. For virus isolation, HELF or rat embryonal fibroblasts were infected with the appropriate CMV strain at a multiplicity of infection (MOI) of 0.001 and incubated at 37°C until a complete cytopathic effect was observed. The virus was pelleted from culture fluid by ultracentrifugation (90 min at 40,000 × g and 4°C). Virus purification was carried out by density-gradient centrifugation using a continuous 20 to 60% (wt/vol) sucrose gradient (60 min at 90,000 × g and 4°C). The visible virus-containing fraction was collected, washed once with STE (0.1 M NaCl, 0.01 M Tris/HCl [pH 8.0], 0.001 M EDTA) buffer, aliquotted, and stored in liquid nitrogen.

Rat CMV was labeled with FITC (isomer I; Sigma) according to the method described by Benne and coworkers (8). To remove free FITC, the virus was dialyzed against PBS overnight at 8°C, aliquotted, and stored at 70°C.

Ligand Blots

Purified virus, virus-infected, or uninfected HELF were denaturated for 3 min at 100°C in 5 mM Tris/HCl, pH 6.8, containing 200 mM dithiothreitol, 5% SDS (wt/vol), 20% glycerol (vol/vol), and 0.1% bromphenol blue (wt/vol), and size-fractionated on a 10% SDS-polyacrylamide gel. The proteins were transferred to nitrocellulose membrane (Schleicher & Schüll, Dassel, Germany) by overnight electrotransfer (0.8 V/cm²) (Mini Tank Elektroblotter; OWLScientific). To minimize unspecific binding of SP-A, the membrane was blocked with 3% bovine serum albumin (BSA) (wt/vol) in 10 mM Tris/HCl, pH 7.4, containing 150 mM NaCl, 0.05% Tween 20, and 5 mM CaCl₂ for 30 min at room temperature. After washing for 10 min twice with the same buffer, the membrane was incubated with 3 to 5 µg of biotinylated SP-A (sheep) for 12 h at 8°C with slight agitation. Bound SP-A was visualized by avidin-peroxidase and o-phenylenedianine (OPD) peroxidase substrate (Sigma). To prove specificity of SP-A binding, biotinylated SP-A was preincubated with the monoclonal SP-A antibody 17A12 or a polyclonal anti-SP-A serum (19). The 17A12 antibody was produced in our lab and was shown to recognize native SP-A only. It does not interfere with binding of SP-A by type II cells. Alternatively, the blot was preincubated with anti-CMV hyperimmunsera (Cytotect; Biotest, Dreieich, Germany). The influence of mannan on SP-A binding to CMV proteins was tested by preincubation of biotinylated SP-A with 50 mg/ml mannan (Sigma).

Measurement of FITC-CMV Binding to and Internalization by Rat Type II Cells and ATM, as well as Binding of FITC-SP-A to Uninfected or CMV-Infected HELF

To study rat CMV binding and uptake by type II cells and ATM, 50 µl of FITC-labeled RA67 virus was preincubated with SP-A for 10 min at 37°C; added to 2.5 × 10⁶ cells resuspended in 500 µl Dulbecco's modified Eagle's medium (DMEM) containing 1% glutamine, 1% BSA, 0.1% NaN₃, and 5 mM CaCl₂; and incubated at 37°C for 1 h. Thereafter, the cells were centrifuged (10 min at 22 × g and 8°C), washed twice with 1 ml ice-cold PBS, and fixed in 500 µl PBS containing 1% paraformaldehyde (PFA). To differentiate between bound and internalized virus, half of the cells were treated as described while the other half was mixed with 100 µl of quenching solution (Orpegen Pharma, Heidelberg, Germany) and 3 ml of ice-cold PBS. Then the cells were centrifuged, washed, and fixed as described.

To investigate binding of FITC-labeled SP-A to virus-encoded proteins exposed on the surface of infected lung fibroblasts, HELF were mock-infected or infected with AD169 at a MOI of 1. After 6 d the cells were harvested by trypsinization and washed three times with buffer A (10 mM N-2-hydroxyethylpiperazine- N'-ethane sulfonic acid [Hepes], 140 mM NaCl, 5 mM KCl, 25 mM Na₂HPO₄, 2 mM MgSO₄, 6 mM glucose, 2 mM CaCl₂, and 0.1% [wt/vol] gelatine, pH 7.4) for 10 min at 500 × g and at room temperature. Binding of FITC-labeled SP-A was measured essentially as described by van Iwaarden and coworkers (23). Briefly, in a final volume of 400 µl of buffer A containing 5 mM CaCl₂, 2.5 × 10⁵ cells were incubated with various amounts of FITC-conjugated human SP-A for 30 min at 37°C in a shaking water bath. The reaction was terminated by adding 2 ml of ice-cold PBS. The cells were washed twice with PBS for 10 min at 500 × g (4°C) and fixed in 1 ml PBS/1% PFA. To prove the influence of mannan on the interaction of the SP-A/CMV complex with rat lung cells, 50 mg/ml mannan were added to the reaction mixture.

Cell suspensions were analysed by FACScan and the Lysis II software (Becton Dickinson, Heidelberg, Germany). Type II cells, ATM, and lymphocytes were gated in a forward/side scatter. To verify this gate, an additional tube was provided for staining with CD45-FITC (OX-1) and phycoerythrin-labeled antimacrophage subset monoclonal antibody (mAb) (HIS36), both from Becton Dickinson/PharMingen (Heidelberg, Germany).

Confocal Laser Fluorescence Microscopy

To visualize intracellular FITC-labeled rat CMV, 5 × 10⁶ type II cells and ATM were resuspended in 1 ml DMEM/0.1% BSA and incubated for 1 h at 37°C with FITC-labeled rat CMV that was preincubated for 10 min with or without SP-A. The cells were washed three times with cold PBS/0.1% BSA and once in cold PBS and then fixed in 500 mM Hepes (pH 7.4)/1% PFA for 30 min at 37°C. Thereafter, the cells were washed again twice with cold PBS. The cells were placed on glass slides, covered with Prolong Antifade Kit under a glass coverslip, and examined using an epifluorescence microscope interfaced with a laser scanning confocal microscope (Leica CLSM; Leica Lasertechnik GmbH, Heidelberg, Germany). Images of cells were created using standard objectives and photomultiplier tubes dedicated to fluorescent excitation and emission spectra for FITC (excitation 490 nm, emission 520 nm) and nonconfocal transmitted light for visualization of cell borderline. Using the dual filter system of the confocal microscope, dual-emission (535/590 nm) images were recorded simultaneously with a scanning speed of 1 s/frame (512 lines). Serial sectioning of cells at 1 µm depth was performed to distinguish material adhering to the cell membrane from internalized material and to assess intracellular CMV. Quantitative and statistical functions were performed using the Quantify Tool-Window, Leica TCS NT Version 1.5.451 (Leica Lasertechnik GmbH). Using the stacks function xz, FITC-fluorescence intensity profiles of marked areas of 10 representative cells were estimated.

Standard Plaque Assay

HELF in calibrated culture flasks were infected with AD169 at an MOI of 0.002. The virus was allowed to adsorb for 60 min at 37°C. Thereafter, nonadsorbed virus was removed from the cells. The infected monolayer was overlayered with 5% methocel medium and cultivated at 37°C. After 10 d the number of plaque-forming units (PFU) was determined by light microscopy. Significance of SP-A effects on CMV replication was tested by Wilcoxon's test for paired samples.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Binding of SP-A to Immobilized Human CMV Proteins

Specific binding of biotinylated SP-A to denatured human CMV proteins of strain AD169 as well as CMV-encoded proteins expressed in infected embryonal lung fibroblasts was studied in ligand blots. In protein extracts from purified virus (Figure 1, lane 4) and from virus-infected HELF (Figure 2B, lane 3), there are several prominent bands reacting with biotinylated SP-A. For comparison, no specific binding of biotinylated SP-A to proteins from uninfected HELF was observed (Figure 2A, lane 3). As a control, the absence or presence of CMV late structural proteins in uninfected and infected HELF, respectively, were shown by incubation of the blot with antibodies corresponding to the gB (gp58) envelope protein and the p68 late protein (Figure 2, lanes 1 and 2).

View larger version (66K): [in this window] [in a new window]   Figure 1.   Binding of biotinylated sheep SP-A to immobilized AD169 proteins. Proteins of purified AD169 virus were separated by electrophoresis, transferred to nitrocellulose membrane, and incubated with: lane 1: biotinylated sheep SP-A (5 µg) that was preincubated for 12 h at 8°C with a polyclonal anti-SP-A antiserum; lane 2: biotinylated sheep SP-A (5 µg) in the presence of 5 mM EDTA; lane 3: biotinylated sheep SP-A (5 µg) in the presence of 50 mg/ml mannan; or lane 4: biotinylated sheep SP-A (5 µg).

View larger version (96K): [in this window] [in a new window]   Figure 2.   Binding of biotinylated sheep SP-A to immobilized proteins from noninfected and human CMV-infected HELF. Proteins from uninfected (A) and AD169-infected (B) HELF were separated by electrophoresis, transferred to nitrocellulose membrane, and incubated with: lane 1: mAb against human CMV late protein p68; lane 2: mAb against human CMV gB protein; lane 3: biotinylated sheep SP-A (5 µg); lane 4: protein stained with Ponceau reagent; lane 5: molecular weight standards.

Specificity of the binding was verified by addition of (1) a polyclonal anti-SP-A antibody, (2) a polyclonal antiserum against human CMV, or (3) preincubation of the membrane with an excess of unlabeled sheep SP-A (not shown). Preincubation of the biotinylated SP-A with a polyclonal anti-SP-A serum (Figure 1, lane 1) or the monoclonal anti- SP-A antibody 17A12 (not shown) completely blocked the interaction between SP-A and CMV proteins. Similarly, preincubation of immobilized human CMV proteins with two different polyclonal anti-hCMV sera or a 10-fold excess of unlabeled sheep SP-A prevented SP-A binding (data not shown).

The binding of SP-A to human CMV proteins was independent of the biologic source of the SP-A. Identical binding patterns as with sheep SP-A were observed for bovine SP-A (data not shown). Because biotinylation of the human SP-A from an alveolar proteinosis patient was ineffective, we could not study its binding in this test; we were, however, able to inhibit the binding of sheep SP-A by preincubation of the membrane with an excess of unbiotinylated human SP-A (data not shown).

Interaction of SP-A with Immobilized CMV Proteins Is Ca²⁺-Dependent and Likely Involves the Carbohydrate Recognition Domain of the SP-A Molecule

To determine whether the binding of SP-A to human CMV is Ca²⁺-dependent, the appropriate incubation buffers were supplemented with 5 mM ethylenediaminetetraacetic acid (EDTA) to complex free Ca²⁺ ions. As shown in Figure 1, addition of EDTA to the ligand blot completely abrogated the binding of SP-A to human CMV proteins when compared with the control lane (Figure 1, lanes 2 and 4). Interaction between SP-A and human CMV proteins could also be competed for by the addition of mannan, a high-mannose glycan (Figure 1, lane 3), suggesting involvement of the carbohydrate recognition domain (CRD) of SP-A in binding to the human CMV proteins in this assay.

Binding of SP-A to Human CMV Proteins Exposed on the Surface of Infected HELF

To further characterize the CMV proteins interacting with SP-A we studied binding of FITC-labeled human SP-A to uninfected and infected HELF. Glycosylated CMV envelope proteins are exposed on the surface of infected cells permissive for virus replication. Human CMV AD169- infected HELF were harvested 6 d after infection, trypsinated, and incubated for 30 min with FITC-labeled human SP-A. Fluorescence intensity was analyzed by fluorescence-activated cell sorter (FACS). In all experiments a significantly increased binding activity of FITC-SP-A to virus-infected compared with uninfected control cells was obtained (Figure 3). Specificity of the binding was confirmed by preincubation of the SP-A with the polyclonal anti-SP-A serum, which reduced binding by about 50% (data not shown). Also, the interaction of SP-A and virus-infected as well as uninfected fibroblasts depended on the presence of Ca²⁺ ions. As shown in Figure 3, addition of EDTA suppressed the binding of SP-A to infected and uninfected cells.

View larger version (23K): [in this window] [in a new window]   Figure 3.   Binding of FITC-labeled human SP-A to human CMV proteins expressed on the surface of AD169-infected human embryonal lung fibroblasts. HELF were mock-infected or infected with human CMV AD169 strain at a MOI of 1, harvested at Day 5 after infection, and incubated for 30 min with different amounts of FITC-labeled human SP-A in the presence of 5 mM CaCl2 with or without 5 mM EDTA. Binding of SP-A was measured by FACS analysis. The graph represents mean value (± standard deviation [SD]) of three independent experiments.

Thus, CMV envelope proteins exposed on the surface of infected cells as well as on the viral envelope can be targets for SP-A binding.

SP-A Increases Binding and Uptake of FITC-Labeled Rat CMV by Type II Cells and ATM

A mixture of rat lung type II cells (60 to 70%), ATM (15 to 20%), and lymphocytes (10%) prepared from lungs of adult rats was incubated for 1 h at 37 °C with FITC-labeled rat CMV strain RA67 in the absence or presence of increasing amounts of sheep SP-A. After fixation, the main fluorescence of the cells as well as the number of fluorescent cells in the different cell populations were measured by FACS analysis.

In the presence of increasing amounts of SP-A the number of fluorescent cells as well as the mean fluorescence (not shown) were enhanced in a concentration-dependent manner for both type II cells and ATM (Figure 4A). The amount of 5 to 10 µg sheep SP-A per milliliter enhanced the cell-associated fluorescence by averages of 100 and 50% when compared with the untreated control for type II cells and ATM, respectively. In the presence of 27 µg SP-A/ ml the number of fluorescent type II cells increased up to 3-fold.

View larger version (14K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   Figure 4.   (A) SP-A-dependent increase of binding/entry of FITC-labeled rat CMV by rat type II cells and ATM. Relative number of fluorescent type II cells and ATM after 1 h incubation with FITC-labeled rat CMV RA67 that was preincubated with increasing amounts of sheep SP-A. The graphs represent mean values (± SD) of three independent experiments. (B) Influence of LPS on SP-A-mediated stimulation of rat CMV binding/uptake by rat type II cells. Number of fluorescent cells (%) after 1 h incubation with FITC-labeled rat CMV that was preincubated for 10 min at 37°C in the absence or presence of 20 µg sheep SP-A and/or 8 pg LPS and/or 300 ng of endotoxin inhibitor. The graph represents one representative experiment.

To exclude the possibility that endotoxin contamination of SP-A contributed to the observed effects, we studied the influence of lipopolysaccharide (LPS) as well as an endotoxin inhibitor (soluble endotoxin binding protein) (kindly provided by H. D. Volk, Institute of Medical Immunology, Charite, Berlin, Germany) on CMV binding/ uptake by rat lung cells. In two independent experiments, 0.4 pg LPS/µg of SP-A did not show any measurable effect on virus uptake, as demonstrated in Figure 4B. Similarly, the endotoxin inhibitor had no effect on SP-A-mediated stimulation of CMV binding/uptake by rat type II cells (Figure 4B) and ATM (data not shown).

To discriminate between bound and internalized virus, the cells were treated with quenching solution to eliminate the signal from labeled virus attached to the outer cell membrane. We could show that SP-A stimulates not only binding but also internalization of CMV by these cells (Figure 5). Correspondingly, confocal laser scanning microscopy demonstrated enhanced internalization of FITC-labeled rat CMV into rat type II cells and ATM, as shown in Figure 6. Mean fluorescence intensity observed in type II cells incubated with FITC-labeled CMV in the presence of SP-A (10 µg/ml) was 1.4-fold higher than in the untreated control (P  0.0445).

View larger version (42K): [in this window] [in a new window]   Figure 5.   SP-A stimulates internalization of FITC-labeled rat CMV by type II cells (A) and ATM (B). Number of fluorescent rat type II cells and ATM (%) after incubation for 1 h at 37°C with FITC-labeled rat CMV RA67 in the absence or presence of SP-A and with or without quenching. Figure shows results of one representative experiment, which was repeated twice.

View larger version (30K): [in this window] [in a new window]   Figure 6.   Photographs of confocal laser fluorescence microscopy from rat type II cells (A and B) and ATM (C and D) incubated with FITC-labeled rat CMV RA67 in the absence (A and C) or presence (B and D) of 10 µg sheep SP-A. Representative sections through the center of a cell are shown. Green: FITC-labeled rat CMV; red: cell borderline.

SP-A-mediated binding and internalization of rat CMV into type II cells and ATM as determined by FACS analysis could be abrogated by the addition of 5 mM EDTA, indicating that the effect is Ca²⁺-dependent (not shown) as demonstrated earlier for binding of SP-A to immobilized CMV proteins and virus-infected fibroblasts. However, 50 mg/ml mannan, which was shown to inhibit the interaction of SP-A and immobilized CMV proteins (Figure 1, lane 3), did not influence SP-A-mediated increase of FITC-CMV binding/uptake by type II cells and ATM (data not shown).

To confirm the specificity of SP-A-mediated uptake, type II cells incubated with liposomes (for experimental details, see Reference 19) in the absence of SP-A internalized very little if any lipids (2,020 ± 106 disintegrations/ min [dpm]/10⁶ cells). In the presence of SP-A (2 µg) a considerably larger amount of lipids was internalized (3,640 ± 425 dpm/10⁶ cells), whereas in the presence of C1q (2 µg) no increase in lipid uptake was observed (1,915 ± 167 dpm/ 10⁶ cells). Very similar results were obtained when we compared the influence of apoprotein B and SP-A on the uptake of [³H]dipalmitoylphosphatidylcholine-labeled liposomes. APO B had no influence (1,220 ± 90 versus 1,380 ± 80 dpm/10⁶ cells), whereas SP-A increased the uptake by 100% compared with the control (1,220 ± 90 versus 2,500 ± 80 dpm/10⁶ cells).

Influence of SP-A on the Infectivity of Human CMV

We used a standard plaque assay to study the influence of SP-A on human CMV infectivity to HELF. Human CMV strain AD169 was preincubated with increasing amounts of sheep (Figure 7A) or human (Figure 7B) SP-A before allowing adsorption to fibroblasts. After adsorption, the infected cells were incubated for 10 d. As shown in Figure 7A, preincubation of human CMV AD169 with 5 or 12.5 µg/ml of sheep SP-A has a very marginal stimulatory effect on CMV replication. However, for other concentrations the effect was not consistently significant. Human SP-A also slightly stimulated CMV replication, but the effect was not significant compared with the untreated control.

View larger version (49K): [in this window] [in a new window]   Figure 7.   Influence of sheep and human SP-A on replication of human CMV AD169 in HELF. Human CMV AD169 was preincubated for 20 min at 8°C with different amounts of sheep (A) or human (B) SP-A before adsorption to HELF. For control, virus was preincubated in the absence of SP-A. PFU were counted 10 d after infection. Mean value (± SD) of three independent experiments, each in duplicate, are presented. *P < 0.05 versus control; Wilcoxon's test.

In additional experiments we could show that preincubation of HELF with SP-A before incubation with the virus did not influence the efficiency of subsequent virus adsorption and infection (data not shown).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

It has been proposed that SP-A functions as a preimmune opsonin for rapid recognition and elimination of a wide range of pathogenes including viruses, bacteria, and fungi (5, 8). On the other hand, on some pathogens, such as HIV-1 (16), it has no effect or may even be deleterious, as described for M. tuberculosis (12) and P. carinii (11).

CMV is a major cause of pulmonary infection in perinatally infected newborns and immunosuppressed patients, including transplant recipients and HIV-infected individuals. In this study we present data that SP-A facilitates CMV binding to and infection of various types of lung cells, especially type II cells and ATM. Further, SP-A does not inhibit CMV infection of and replication in embryonal lung fibroblasts, indicating that SP-A has no antiviral activity against CMV. Whether the observed enhancement of virus replication in fibroblasts is of pathophysiologic relevance is as yet unknown. Thus, by enhancing virus entry, SP-A may contribute to the pathogenesis of CMV infection in the lung.

We found that SP-A binds both to immobilized human CMV proteins as well as to human CMV proteins exposed on the surface of infected embryonal lung fibroblasts in a Ca²⁺-dependent manner. Further, binding of SP-A to CMV proteins in ligand blots can be inhibited by mannan, a high-mannose glycan. Therefore, at least in those in vitro experiments, interaction between SP-A and CMV proteins likely involves the carbohydrate recognition domain of SP-A and the envelope glycoproteins of the virus expressed on the viral surface as well as on the surface of the virus-infected cell. The target protein(s) for the binding of SP-A to human CMV remain speculative. Sequence analysis of the human CMV genome has predicted as many as 65 unique glycoproteins but only a few of them have been studied in regard to function and structure (reviewed in Reference 24). Four glycoprotein homologues of herpes simplex virus (HSV)-1 have been identified: gB, gH, gM, and gL. The gB (gp65) protein, being the most abundant envelope protein of human CMV, and the gH protein have been proposed to participate in both attachment and fusion of the virion with the host cell membrane (24). In immunoprecipitated SP-A-CMV complexes, no gB protein could be detected by Western blot analysis, indicating that the envelope protein gB is not a target for SP-A binding (data not shown). These data are supported by our observation that SP-A has no antiviral effect in HELF, as gB protein is important in receptor-specific attachment and fusion of CMV with HELF (24). The involvement of other known envelope proteins was not tested for lack of commercially available antibodies.

Carbohydrate-dependent binding of SP-A to virus proteins is not unique to CMV. The mechanisms of SP-A involved in the interaction with specific viruses, however, may be different between viruses. Binding of SP-A to HSV-1 proteins exposed on the surface of infected cells was mediated via the carbohydrate moiety of SP-A and not the CRD itself (25). This interaction was insensitive to mannan, whereas deglycosylation of SP-A abrogated the interaction. Binding of SP-A to influenza virus A involved the sialic acid residues on the SP-A molecule (15, 26). Deglycosylation of SP-A or enzymatic digestion to remove only the sialic acid residues inhibited the interaction of the collectin with influenza A virus, whereas mannan, which binds to the mannose-binding region, had no effect (15). The influence of SP-A deglycosylation on the interaction with CMV proteins was not tested, but in contrast to HSV-1 and influenza A virus, mannan was able to block binding of biotinylated SP-A to CMV proteins in ligand blots, indicating that the mannose-binding region of SP-A may be involved.

In further studies we could demonstrate that preincubation of CMV with SP-A enhances binding/uptake of the virus by rat lung cells by more than 100% when compared with untreated CMV. This increase was Ca²⁺-dependent but could not be competed for by addition of mannan, indicating that the CMV-binding domain and the cell-binding domain of SP-A are different. The precise mechanism(s) by which SP-A stimulates the entry of CMV into different lung cells remains unknown. Generally, there are two possible mechanisms: direct and indirect. SP-A may opsonize CMV by binding to the virus and acting as a "bridge molecule" to enhance interaction of the virus with its target cell. For SP-A, specific and unspecific receptors on type II cells and alveolar macrophages have been described (27). Interaction of the SP-A with the receptors identified by Stevens and coworkers (unpublished observations) or Kresch and colleagues (29) may involve the C-terminal globular domain of SP-A inasmuch as it is Ca²⁺-dependent but not inhibited by mannan. Thus, binding to an SP-A binding protein by the globular domain may be one possible explanation for increased attachment of SP-A-opsonized CMV by lung cells. Similarly, for influenza A virus, interaction between the virus/SP-A complex and the target cell has been shown to be mediated via the collectin domain of SP-A (26).

CMV attachment to its target cell includes binding to its specific receptors (mediated by glycoprotein B and glycoprotein CII) and to cell-surface heparan sulfate as coreceptor (30). Binding of virus-associated SP-A to other glycans on type II cells or ATM may involve an alternative coreceptor way. Thus, binding of virus-associated SP-A to SP-A- specific and/or -unspecific receptors on the cell surface, thereby enhancing binding of the virus to its specific receptors, may be one possible mechanism for the increased binding/uptake of CMV by rat lung cells in the presence of SP-A.

Alternatively, SP-A may act indirectly by interaction with the target cell and subsequent stimulation of expression of specific or unspecific receptors involved in CMV attachment and penetration. Recently, such an indirect mechanism in addition to a potential opsonin effect was shown to mediate increased attachment of M. tuberculosis to alveolar macrophages (12). The authors demonstrated that preincubation of the target cells with SP-A results in an upregulation of mannose receptors, which are able to bind the bacilli. Our experimental setup does not allow us to exclude this possibility because the SP-A was not washed out after preincubation with the virus. In contrast to our observation, the SP-A-mediated increase of M. tuberculosis attachment was inhibited by mannan. Also, deglycosylation of SP-A abrogated the effect. The influence of SP-A deglycosylation on the enhanced CMV binding/uptake by rat lung cells was not tested, but addition of mannan did not interfere with uptake, indicating that upregulation of mannose-binding receptors may not be involved.

If SP-A is involved in the pathogenesis of CMV infection of the lung, patients with elevated levels of SP-A in the lung should have an increased incidence of CMV infection. HIV-infected individuals are a well characterized patient group with significantly increased SP-A concentrations (31). In agreement with this, CMV can be detected postmortem in the lungs of 44 to 65% of HIV-infected individuals, despite the absence of pneumonitis (32). Similarly, enhanced SP-A levels in bronchial lavages of HIV-infected patients correlate to an increased susceptibility to M. tuberculosis infection (12). Elevated SP-A levels were also detected in lavages from patients with sarcoidosis and hypersensitivity pneumonitis (33, 34). Interestingly, there are many publications reporting a high incidence of M. tuberculosis in the lungs of sarcoidosis patients (for review, see Reference 35). Unfortunately, there are no data concerning the incidence of CMV in the lungs of these patients.

In conclusion, our results demonstrate that SP-A enhances CMV binding and uptake by type II cells and ATM, the main target cells of CMV in pneumonitis. This may contribute to the pathogenesis of CMV lung infection.