Expression of Surfactant Protein D in the Human Gastric Mucosa and during Helicobacter pylori Infection

Abstract

Helicobacter pylori establishes persistent infection of gastric mucosa with diverse clinical outcomes. The innate immune molecule surfactant protein D (SP-D) binds selectively to microorganisms, inducing aggregation and phagocytosis. In this study, we demonstrated the expression of SP-D in gastric mucosa by reverse transcription-PCR and immuohistochemical analysis. SP-D is present at the luminal surface and within the gastric pits, with maximal expression at the surface. Levels of expression are significantly increased in H. pylori-associated gastritis compared to those in the normal mucosa. Immunofluorescence microscopy was used to demonstrate binding and agglutination of H. pylori by SP-D in a lectin-specific manner. These activities resulted in a 50% reduction in the motility of H. pylori, as judged on the basis of curvilinear velocity measured by using a Hobson BacTracker. Lipopolysaccharides extracted from three H. pylori strains were shown to bind SP-D in a concentration-dependent manner, and there was marked variation in the avidity of binding among the strains. SP-D may therefore play a significant role in the innate immune response to H. pylori infection.

Helicobacter pylori is a gram-negative bacterium which colonizes gastric mucosa. It is one of the most common pathogens, with a prevalence of up to 90% in developing countries (17). Once infection is established, gastric mucosal inflammation develops, and although infection persists for life, only 30% of those infected become symptomatic. The outcome of infection is diverse and includes duodenal ulcer, gastric ulcer, and gastric malignancy—both carcinoma and lymphoma. Such heterogeneous consequences are dependent on the time span and topography of mucosal inflammation (28). The route of infection is not proven, but the most likely event is direct ingestion of gastric contents. Gastric mucosa is markedly adverse to bacterial colonization, as the physical and chemical barriers encountered (mucus, enzymes, and acid) inhibit colonization by common bacteria.

H. pylori-related gastritis is characterized by both lymphocytic and neutrophil infiltrates. In addition, there is a strong humoral immune response with specific antibodies of both immunoglobulin G (IgG) and IgA classes. Despite evidence of an immune response, H. pylori infection is frequently persistent, suggesting that the organism may evade conventional innate and adaptive immune responses. A number of factors appear to contribute to the virulence of H. pylori, including the possession of flagella, which confer motility (6).

Surfactant protein D (SP-D) is a collagenous glycoprotein which contains trimeric arrays of C-type (calcium-dependent) lectin domains and which belongs to a family of proteins implicated in innate immunity, termed the collectins (5, 25, 27). The protein and molecular structures for SP-D are well characterized, and the gene has been localized to human chromosome 10q22.2-23.1 (8, 13, 18). SP-D has been shown to recognize and bind selectively to the surfaces of viruses, bacteria, protozoa, and fungi and is currently being considered for use as a therapeutic agent for the treatment of both bacterial infections and allergies in humans (4, 19, 23, 24, 26). SP-D is believed to bind directly to the lipopolysaccharide (LPS) on the surface of gram-negative bacteria via the carbohydrate recognition domain (16). This process may result in the aggregation of microorganisms followed by enhanced phagocytosis by neutrophils and macrophages (11). Although phagocytosis of microorganisms appears to be enhanced by interaction with SP-D or SP-A, it is unclear whether such interactions promote further immunological or inflammatory responses (2, 9, 11, 30).

In humans, SP-D has been shown to be synthesized in the lungs, specifically alveolar type II and Clara cells, and to be present in the fluid phase of the airways. SP-D has also been detected at other sites, including tears, amniotic fluid, and fetal membranes. In the rat, SP-D has been found in mucus-secreting cells of the gastric mucosa but not in the duodenum or remaining intestine, and a speculative role in mucus barrier assembly and possibly host defense has been proposed (7).

In order to establish whether SP-D may play a role in defense against H. pylori infection or in the pathogenesis of established H. pylori infection, the aim of this study was to determine whether SP-D is present in the human gastric mucosa and whether SP-D interacts with H. pylori. Further studies were performed to determine if the motility of H. pylori is impaired by SP-D. In addition, SP-D binding studies were performed to establish whether SP-D binds to LPS from bacterial cell walls.

MATERIALS AND METHODS

Preparation of SP-D and antisera to SP-D.

SP-D was purified from lung lavage fluid from patients with alveolar proteinosis by the method of Strong et al. (29). Briefly, lavage fluid was centrifuged at 10,000 × g for 40 min. The supernatant was applied to a maltosyl-agarose column, and bound SP-D was specifically eluted by using MnCl₂. SP-D was further purified by gel filtration on Superose-6.

A recombinant preparation of human SP-D head and neck regions expressed in Escherichia coli was used to raise polyclonal antisera in rabbits. The IgG fraction was purified from rabbit serum by sodium sulfate precipitation followed by protein A chromatography as previously described (20). The specificity of the antisera for native human SP-D was confirmed by an enzyme-linked immunosorbent assay (ELISA) and Western blotting.

Detection of SP-D mRNA.

mRNA for SP-D was detected by reverse transcription (RT)-PCR. Twenty-eight patients with nonulcer dyspepsia were selected at the time of endoscopy and gave informed consent. Thirteen patients had normal histological findings, and 7 had H. pylori-associated antral gastritis. RNA was extracted from three gastric antral biopsy specimens taken from each patient. The biopsy specimens were snap-frozen in liquid nitrogen and stored at −70°C until processed. When required, biopsy specimens were homogenized in saline and RNA was isolated by using a total RNA isolation kit according to the manufacturer's protocol (Ambion, AMS Biotechnology, Europe Ltd.). cDNA synthesis and PCR amplification were carried out as a single step by using a one-step RT-PCR kit according the manufacturer's protocol (ABgene, Surrey, United Kingdom). Primers from exons 3 and 4 of the SP-D gene were designed to give product sizes of 475 bp from genomic DNA and 155 bp from cDNA. The primer sequences were as follows: exon 3, 5′-GAACATAGGACCTCAGGGCA-3′, and exon 4, 5′-TGTGTTTCCAGGGACTCCAC-3′. Parallel RT-PCRs were performed with primers for the human β-actin gene to serve as a reference.

First-strand synthesis was performed with a single cycle at 47°C for 30 min. The reverse transcriptase was then inactivated by denaturation at 94°C for 2 min. PCR amplification was performed with 43 cycles of annealing at 63°C for 30 s, extension at 72°C for 1 min, and denaturation at 94°C for 30 s. There was a final extension step at 72°C for 5 min. Negative and positive control amplifications were performed in every experiment. RT-PCR products for SP-D and β-actin were run on a 2% (wt/vol) agarose gel, stained with ethidium bromide, and visualized under UV light. Digital images were obtained, and densitometry was performed by using a Syngene gel documentation system (Synoptics Ltd., Cambridge, United Kingdom) and Genetools v3.0 software (Synoptics). Semiquantitative measurement of SP-D mRNA was achieved by comparing the area under the peak of the SP-D band with the area under the peak of the β-actin band and expressing the result as a simple ratio.

Immunohistochemical analysis.

Immunohistochemical analysis was performed with the anti-SP-D polyclonal antisera described above and with monoclonal antibody 245.2 (a gift from U. Holmskov, Odense, Denmark). The specificity of this antibody has been confirmed by Western blotting and ELISA analyses (21).

Formalin-fixed, paraffin-embedded gastric biopsy specimens were randomly selected from the pathology archives for immunocytochemical staining, and 24 specimens were classified according to histological appearances. Of the 24 cases selected, 11 were H. pylori-associated antral gastritis, 4 were H. pylori-negative chemical antral gastritis, and 9 were H. pylori-negative normal antral mucosa. Two 5-mm sections were cut from each biopsy specimen; one of these was stained with the rabbit anti-SP-D polyclonal antisera, and the other was stained with the SP-D-specific monoclonal antibody. Staining was demonstrated by using a routine avidin-biotin-horseradish peroxidase system. Negative controls were stained with the monoclonal antibody which had been preincubated with purified SP-D, with an irrelevant antibody, and with preimmune sera from the same rabbits.

SP-D expression was evaluated by a quantitative method with a standard grid (10). Positive immunostaining was assessed by site in the gastric mucosa (lumen, foveola, and pit) by counting cells expressing SP-D in a given area delineated by the grid.

Binding and agglutination.

H. pylori strain J178 (cagA positive) was cultured in brain heart infusion broth at 37°C under microaerophilic conditions to a concentration of 10⁸ to 10⁹ bacteria ml⁻¹. The organisms were centrifuged at 3,000 × g and resuspended in Tris-buffered saline (pH 7.4) (TBS).

Binding of SP-D to H. pylori was demonstrated by immunofluorescence. H. pylori was incubated with SP-D (10 μg/ml) in TBS containing 10 mM CaCl₂ or 10 mM EDTA to demonstrate calcium-dependent binding. Controls were incubated in calcium buffer alone. To assess lectin specificity, 100 mM maltose was added to bacteria treated with SP-D and 10 mM CaCl₂. Finally, SP-D binding to bacteria was detected by probing with biotinylated anti-human SP-D antisera (1:200 dilution of a 1-mg/ml IgG stock solution) followed by streptavidin-fluorescein isothiocyanate (1:200; Sigma Chemical Co., Dorset, United Kingdom). Agglutination was demonstrated by direct observation of the bacteria with phase-contrast microscopy. H. pylori organisms resuspended in TBS containing 10 mM CaCl₂ at 10⁸ organisms/ml were incubated for 30 min with buffer alone, with 10 μg of SP-D/ml, with 10 μg of SP-D/ml and 10 mM EDTA, or with 10 μg of SP-D/ml and 100 mM maltose. Phase-contrast microscopy and fluorescence microscopy were performed by using a Zeiss Axioskop transmitted-light microscope. H. pylori was photographed with a Zeiss MC 100 camera (final magnification, ×400) by using TX 400-ASA black and white film for phase contrast and Provia 1600 color slide film for fluorescence.

Quantitative estimates of SP-D-mediated agglutination were obtained for clinical isolates of H. pylori. Bacteria were placed in optical cuvettes suspended in TBS plus 10 mM CaCl₂. Two aliquots of each isolate were adjusted to an optical density at 700 nm (OD₇₀₀) of 0.7, and SP-D at a final concentration of 2.5 μg/ml was added to one aliquot. OD₇₀₀ values were measured at 15-min intervals up to 60 min and again at 90 min. Agglutination was expressed as the difference in OD₇₀₀ units between control and SP-D-containing aliquots at 60 min.

Inhibition of motility.

Assessment of H. pylori motility was made by using a Hobson BacTracker as previously described (15). The technique gives a quantitative measure of motility by using an image-processing computer combined with a phase-contrast microscope and video equipment. Motility was assessed in the presence and absence of 5 μg of SP-D/ml in calcium buffer with and without the addition of 100 mM maltose. Motility is expressed as curvilinear velocity, which is the length of a track (total path length) divided by the time taken to travel it in micrometers per second.

LPS binding studies.

Binding of SP-D to LPS was demonstrated in a competitive inhibition ELISA. LPS was extracted from three strains of H. pylori: clinical strain J178, cagA-positive reference strain 007, and mouse-adapted strain SS1. After pretreatment of bacterial biomass with pronase (Calbiochem, Los Angeles, Calif.), LPS was extracted by the hot phenol-water technique. The resulting crude LPS preparations, recovered from the water phase of extracts, were purified by treatment with RNase, DNase II, and proteinase K (Sigma) and by ultracentrifugation as described previously (22).

Mannan (10 μg/ml) in carbonate buffer (pH 9.6) was used to coat a Maxisorb Immunoplate (Life Technologies, Paisley, United Kingdom) overnight at 4°C. The plate was washed three times in TBS and then blocked with TBS plus 3% (wt/vol) bovine serum albumin for 1 h at 37°C. Between each step, the plate was washed three times in TBS plus 0.05% Tween 20 (TBST). LPS (200 to 1 μg/ml) from H. pylori (J178, 007, or SS1) was incubated with a fixed concentration of SP-D (1 μg/ml) for 1 h at room temperature before addition to the mannan-coated, blocked plate. The LPS-SP-D mixtures and SP-D alone (1,000 to 0 ng/ml) were incubated on the plate for 3 h at 37°C. An aliquot (100 μl) of biotinylated rabbit anti-SP-D polyclonal antisera (1:1,000 dilution of a 1-mg/ml stock) in TBST was added to each well and incubated for 2 h at 37°C. After washing, 100 μl of ExtrAvidin-peroxidase conjugate (Sigma) diluted 1:10,000 in TBST was added and incubated for 30 min at 37°C. The plate was developed with tetramethylbenzidine substrate (Bio-Rad, Hemel Hempstead, United Kingdom), the reaction was stopped with 100 μl of 1 N H₂SO₄, and readings were carried out at 450 nm (Titertek Multiscan PLUS MKII). The resulting graphs (see Fig. 5) were generated by using Prism software (GraphPad Software Inc., San Diego, Calif.), and the data were used to calculate the concentration of LPS required for 50% inhibition of SP-D binding.

FIG. 5.

Quantitative assessment of SP-D-mediated agglutination of H. pylori clinical isolates. Agglutination was measured as the difference in OD₇₀₀ between isolates incubated with or without SP-D at 2.5 μg/ml.

RESULTS

Detection of SP-D mRNA.

SP-D expression at the mRNA level was assessed in gastric mucosa by use of RT-PCR. mRNA for SP-D could be detected in 12 (100%) of 12 gastric tissue samples where H. pylori infection was present and in 10 (62.5%) of 16 samples from individuals who did not have the infection (Fig. 1A). The levels of gastric SP-D mRNA expression in individuals with H. pylori infection were significantly higher than those in individuals without the infection. The median SP-D/β-actin ratios were 0.34 (range, 0.01 to 0.79) in individuals with H. pylori infection and 0.08 (range, 0 to 0.45) in individuals without the infection (Fig. 1B) (P = 0.004) (Mann-Whitney U test).

FIG. 1.

SP-D mRNA is expressed in gastric mucosa. (A) RT-PCR results for RNA extracted from gastric mucosal tissue samples taken from patients with dyspepsia. Bands specific for actin and SP-D are seen at 325 and 155 bp, respectively. “Helicobacter” indicates the presence (+) or absence (−) of H. pylori infection, as determined by histological examination; s, standards. (B) Level of SP-D mRNA expression, assessed by densitometry of the RT-PCR bands and presented as plots (mean and standard error of the mean) of the SP-D/β-actin ratio. H. pylori-positive and H. pylori-negative gastric samples were compared.

Immunohistochemical analysis.

In order to assess, at the cellular level, the site of SP-D expression, immunocytochemical analysis was performed on endoscopic gastric mucosal biopsy specimens taken routinely from 24 patients with dyspepsia. Immunohistochemical analysis was performed with rabbit polyclonal antisera and the SP-D-specific monoclonal antibody. Negative controls, performed with the monoclonal antibody preabsorbed with SP-D, an irrelevant antibody, and preimmune sera from the same rabbits, revealed no immunohistochemical staining. Staining with the anti-SP-D monoclonal antibody gave results similar to those seen with the polyclonal antisera.

The staining demonstrated SP-D protein expression in epithelial cells, and SP-D coating of H. pylori was observed in situ (Fig. 2). The level of SP-D expression was observed to vary at different epithelial levels in the mucosa, with maximal intensity in the luminal mucosa compared to the foveolar region and the gastric pits. No staining was observed in sections stained with preimmune sera.

FIG. 2.

SP-D expression demonstrated by immunohistochemical analysis. Immunohistochemical analysis was performed with an anti-SP-D monoclonal antibody and antral gastric mucosa by using a standard avidin-biotin technique as described in Materials and Methods. (A) Negative control after the antibody was preabsorbed with native SP-D. (B) Low-level expression of SP-D (as judged by the brown staining) in the basal regions of epithelial cells on the luminal surface of the mucosa. (C) Luminal epithelial cells of H. pylori-positive antral gastric mucosa, with strong SP-D expression on the surface and in the basal regions of the cells. (D) Coating of H. pylori with SP-D (arrowheads). Original magnifications, ×1,000 (A, B, and C) and ×3,000 (D).

Semiquantitative analysis of cellular expression at different levels in the mucosa was performed by comparing the intensities of expression in patients with H. pylori-associated gastritis, reactive (chemical) gastritis, and normal histological findings. SP-D was expressed at equal intensities in the basal region of epithelial cells at each of the anatomical sites examined in all three patient groups. However, in H. pylori-associated gastritis, there was a significant increase in cell surface expression at all sites (P < 0.05) (Mann-Whitney U test) (Fig. 3). Increased cell surface expression was not seen in reactive gastritis.

FIG. 3.

Site-specific expression of SP-D in gastric mucosa. (Left) Schematic representation of the sites used for immunocytochemical assessment. (Right) Semiquantitative comparison of SP-D expression in H. pylori-associated gastritis, reflux-type gastritis, and samples with normal histological findings at the epithelial surface (top), the foveolar region (middle), and the gastric pit region (bottom).

Binding and agglutination.

Binding of SP-D to heat-fixed H. pylori was clearly seen with immunofluorescence staining. The calcium dependence and lectin-specific nature of this binding were confirmed by the lack of binding in the absence of calcium and competitive inhibition by maltose (Fig. 4A to D). SP-D caused agglutination of live H. pylori into large clumps that was inhibited by both EDTA and maltose (Fig. 4E to H).

FIG. 4.

Binding of SP-D to H. pylori and agglutination. (A) Immunofluorescence detection of SP-D binding to H. pylori by biotin-labeled anti-human SP-D antibody (1:200 dilution of a 1-mg/ml stock). (B) Phase-contrast image of the same bacteria as in panel A. (C) Effect of 100 μM maltose on SP-D binding to H. pylori, as detected by immunofluorescence. (D) Phase-contrast image of the same bacteria as in panel C. (E to H) Agglutination of H. pylori after incubation with 10 μg of SP-D/ml and 10 mM CaCl₂ buffer (E), 10 mM EDTA (F), or 100 μM maltose (G) or without treatment (H). Immunofluorescence microscopy and phase-contrast microscopy were performed with a Zeiss Axioskop transmitted-light microscope with a Zeiss MC 100 camera. Final magnifications, ×400.

Agglutination of clinical isolates of H. pylori was measured as the difference in the OD₇₀₀ after 60 min of incubation with 2.5 μg of SP-D/ml (Fig. 5). There was a wide range of values, from 0.100 to 0.475; the mean was 0.218. After 60 min of incubation, the difference in the OD₇₀₀ reached a plateau in some isolates.

Motility.

The motility studies demonstrated that SP-D reduced the curvilinear velocity of H. pylori by approximately 50%. H. pylori in TBS plus 10 mM CaCl₂ had a curvilinear velocity of 17.1 μm/s; this value fell to 9.65 μm/s in the presence of 5 μg of SP-D/ml. H. pylori motility was not reduced in the presence of maltose or in the absence of calcium. Direct observation of binding showed that the bacteria formed encircling clusters. Free movement was impeded by agglutination but was not halted.

LPS binding.

SP-D bound to mannan in a concentration-dependent fashion with a linear relationship to the absorbance at 450 nm at between 50 and 1,000 ng/ml. LPS inhibited SP-D binding in a dose-dependent manner. There was clear variation in the ability of each LPS preparation to inhibit SP-D binding, indicating variability in the avidity of SP-D for LPS binding. The 50% inhibitory concentration for H. pylori J178 LPS was 5.3 μg/ml, that for SS1 LPS was 13.4 μg/ml, and that for 007 LPS was 91.5 μg/ml. The 50% inhibitory concentration obtained with E. coli LPS was 73.7 μg/ml (Fig. 6).

FIG. 6.

Inhibition of SP-D binding to mannan by H. pylori LPS. Dose-dependent competitive inhibition of SP-D-mannan binding by LPSs prepared from three H. pylori strains (J178, 007, and SS1) and one E. coli strain (O128:B8) is shown. Curve fitting was performed by using GraphPad Prism software.

DISCUSSION

This study has demonstrated the expression of SP-D in human gastric antral mucosa at both the mRNA and the protein levels. Expression at this site suggests that SP-D may play a role either in the structure of the gastric mucus layer or in host defense. Low-level SP-D expression in the stomach was recently reported by Madsen et al. (21). However, this study did not take into account the H. pylori status or the anatomical site of tissue collection. In contrast, our data suggest that the gastric epithelium is capable of high levels of SP-D expression in the context of H. pylori infection.

Recent studies with SP-D gene knockout mice have indicated that SP-D is not an essential component of the mucus layer in respiratory epithelium, as these mice develop without significant respiratory pathology (3). However, these experiments do not completely rule out a structural role for SP-D in the gastric mucus layer. SP-D is related to the collectin family of proteins, which are components of the innate immune system. Evidence suggests that the major role of SP-D is the recognition of foreign carbohydrate structures expressed on the cell walls of microorganisms (26). SP-D may therefore be one of the first lines of defense in the gastric mucosa.

H. pylori is a common human pathogen which is capable of establishing a chronic infection of gastric mucosa. The binding, agglutination, and motility studies presented here indicate that SP-D recognizes and interacts with H. pylori, causing agglutination of the organism and inhibition of bacterial motility. Agglutination is thought to facilitate phagocytosis by polymorphonuclear cells or macrophages (9). While some reports suggest that there may be a specific SP-D receptor on neutrophils and macrophages, this idea is controversial, and other studies suggest that SP-D does not act as an opsonin (1, 12).

Motility is thought to be an important virulence determinant in H. pylori, allowing the organism to migrate to suitable niches within the gastric mucosa; inhibition of motility prevents colonization (6, 31). In these experiments, motility was significantly inhibited even in the absence of agglutination.

An attempt was made to measure the levels of SP-D in gastric juice and to correlate these measurements with the histological and immunocytochemical findings. Although it was possible to detect low concentrations of SP-D in gastric juice (results not shown), SP-D appeared to be rapidly digested by proteases. It has been suggested that SP-D secreted from salivary glands may also interact with H. pylori. However, our observations suggest that SP-D from salivary secretions would be rapidly digested in the gastric lumen, whereas SP-D expressed by gastric epithelial cells would be effectively protected under the mucus-bicarbonate barrier.

SP-D is believed to bind to LPS, which is the major component of the cell wall in gram-negative bacteria. LPS was therefore considered to be the most likely target for SP-D binding to H. pylori, and the data from the competitive inhibition assays appear to confirm this notion. Our data do not rule out the possibility that other components of the H. pylori cell wall are also involved in SP-D binding. Significant interstrain variations in the terminal O-chain structures of H. pylori LPS and in the core oligosaccharide have been described (14). Wide variations in agglutination measurements in clinical isolates and variations in binding affinities between LPSs extracted from different H. pylori strains and SP-D were observed, potentially reflecting these structural differences. In the inhibition studies, LPS of H. pylori J178 was the most inhibitory, with H. pylori SS1 LPS being next and H. pylori 007 LPS being the least inhibitory. The structure of J178 LPS is presently under investigation, and serological analysis indicates that LPS of H. pylori SS1 expresses Le^y (A. P. Moran et al., unpublished results); however, the O chain of H. pylori 007 LPS consists of a polymeric Le^x chain, lacking terminal Le^y, thus indicating variations in the O-chain structures of these strains. However, these later studies also have indicated variations in the core oligosaccharide of these strains. Therefore, at present, no clear conclusions can be made as to the structures within H. pylori LPS involved in SP-D binding.

The expression of SP-D in human gastric mucosa and a functional interaction between SP-D and H. pylori have been demonstrated. It therefore remains to be explained how persistent infection with this organism is established and how the organism can evade this component of innate immunity.