- Email: [email protected]
Molecular mimicry of ferret gastric epithelial blood group antigen A by Helicobacter mustelae
GASTROENTEROLOGY 1998;114:690–696
Molecular Mimicry of Ferret Gastric Epithelial Blood Group Antigen A by Helicobacter mustelae ´ CRO ´ INI´N, MARGUERITE CLYNE, and BRENDAN DRUMM TADHG O Department of Paediatrics, University College Dublin, The Children’s Research Centre, Our Lady’s Hospital for Sick Children, Dublin, Ireland
Background & Aims: Molecular mimicry of Lewis blood group antigens by Helicobacter pylori may be involved in immune evasion by the bacteria and in the pathogenesis of chronic atrophic gastritis. Helicobacter mustelae infects ferrets naturally, causing gastritis, and may be involved in ulcerogenesis. The aim of this study was to determine if H. mustelae shows a similar form of molecular mimicry. Methods: Antibodies raised against H. mustelae were used to stain ferret gastric tissue by immunoblotting, immunohistochemistry, and flow cytometry. Epitopes recognized by cross-reactivity were characterized by proteinase K and sodium metaperiodate treatment. Results: H. mustelae antiserum reacted with H. mustelae and with ferret gastric tissue. Absorption of the antiserum with H. mustelae or ferret and rabbit gastric tissue removed the cross-reactive antibodies. Antibodies reacted with a blood group antigen A–like structure on ferret gastric epithelial cells and H. mustelae lipopolysaccharide. Conclusions: H. mustelae expresses a blood group–like antigen as part of its lipopolysaccharide that may be used as a method of immune evasion by mimicry of gastric epithelial cells. The cross-reactivity shown by H. mustelae–specific antibodies with gastric mucosa may suggest a role for autoantibodies in the pathogenesis of H. mustelae–induced gastritis in ferrets.
elicobacter pylori has been recognized as a major cause of gastritis and duodenal ulcer disease in humans. Although several virulence factors have been identified, the pathogenesis of H. pylori–induced gastritis is still poorly understood. Several animal models have been proposed to study H. pylori infection in vivo, including H. pylori infection of the gnotobiotic piglet1 or the recently described mouse model.2 However, H. pylori does not infect any of these animals naturally. Helicobacter mustelae is a gastric pathogen that infects ferrets naturally, colonizing the gastric mucosa. Infection is associated with gastric inflammation and is believed to contribute to ulcerogenesis.3,4 H. mustelae–induced gastritis in ferrets has been shown to closely resemble many aspects of H. pylori gastritis in humans.5 Several virulence factors of H. pylori have also been shown to be expressed by
H
H. mustelae, including a potent urease enzyme,6 sheathed flagellae,7 intimate adherence to gastric epithelial cells,5 and a CagA homologue.8 These factors indicate that H. mustelae infection of the ferret may be a useful natural animal model in which to study the pathogenesis of gastritis and duodenal ulcer disease caused by a Helicobacter infection. Despite a strong local and systemic inflammatory response induced by H. pylori,9 infection can persist throughout life and leads to the development of chronic gastritis. A recent study suggested that H. pylori may evade the immune response through a mechanism of molecular mimicry.10 O antigen regions of lipopolysaccharide (LPS) from some H. pylori strains have been shown to be structurally similar to the blood group antigens Lewis x and Lewis y that are expressed on the gastric mucosa of humans.11,12 It has been suggested that these Lewis blood group antigens could potentially mask H. pylori from the immune system.13 Greater than 85% of strains obtained from various parts of the world have been shown to express Lewis antigens.14 However, expression of the Lewis antigens has also been shown to be more common in strains isolated from patients with ulcer disease than from those without the disease.15 Many patients infected with H. pylori have autoantibodies that cross-react with antigens expressed on gastric epithelial cells.16 Recent reports suggested that the presence of autoantibodies may play a crucial role in the pathogenesis of chronic atrophic gastritis. 17 Both H. pylori–associated Lewis x and Lewis y have been shown to induce autoantibodies in humans, and the gastric H1, K1–adenosine triphosphatase found in the canaliculi of parietal cells has been identified as a possible target of the autoimmune response.10,17,18 The aim of this study was to investigate whether H. mustelae possessed a similar mechanism of molecular mimicry as that described for H. pylori. Abbreviation used in this paper: LPS, lipopolysaccharide. r 1998 by the American Gastroenterological Association 0016-5085/98/$3.00
April 1998
Materials and Methods Bacterial Strains H. pylori strain PU3 was isolated from antral biopsy material obtained from a child undergoing upper gastrointestinal endoscopy. H. mustelae strains NCTC 12198 and NCTC 12032 were obtained from the National Collection of Type Cultures (Public Health Laboratory Service, London, England). Both H. pylori and H. mustelae were cultured on Columbia blood agar plates (Oxoid, Columbia, MD) containing 7% defibrinated horse blood for 3 days at 37°C in an atmosphere of 5% O2 and 10% CO2.
Raising of Specific Anti–H. mustelae Antibodies H. mustelae strain NCTC 12198 was harvested from blood agar plates and fixed in phosphate-buffered saline (PBS) containing 0.5% (vol/vol) formaldehyde at a concentration of 1 g of bacteria (wet weight) per 100 mL of solution. For initial immunization, the bacterial suspension was suspended in Freund’s complete adjuvant and was injected intradermally into rabbits. Two subsequent boosters were suspended in Freund’s incomplete adjuvant and injected intramuscularly at 2-week intervals. Test bleeds were taken from the rabbits after the second booster. The serum reacted with a H. mustelae whole-cell lysate by Western immunoblotting at a 1:500 dilution. Rabbits were bled out, the blood was allowed to clot, and the serum was removed.
Antibody Titration H. mustelae antiserum and preimmune serum were tested for the presence of anti-A and anti-B antibodies by titration against blood group A, B, and O red blood cells. Serum was diluted serially in microtiter plates, and red blood cells of blood group A, B, or O were added. The suspension was mixed, spun, and resuspended, and agglutination patterns were recorded.
Isolation of Gastric, Colonic, and Duodenal Epithelial Cells Biopsy tissue was obtained from children undergoing endoscopy at Our Lady’s Hospital for Sick Children, Dublin, and gastric, duodenal, and colonic cells were isolated from biopsy specimens as described previously.19 Ethical approval for this study was obtained from the ethics committee at the hospital. Ferret and rabbit gastric duodenal and colonic tissues were obtained by killing 10–12-week-old ferrets or adult rabbits and removing the stomach, duodenum, and colon. All ferrets were negative for H. mustelae by culture and urease activity of biopsy specimens taken from the antrum, fundus, and duodenum (three samples from each site). The serosa was separated from the tissue by injecting RPMI 1640 tissue culture medium containing 10% (vol/vol) fetal calf serum underneath the serosa and peeling it away. The remaining tissue was minced using scalpels and incubated with Hank’s balanced salt solution containing 0.1 mmol/L ethylenediamine-
H. MUSTELAE AND MOLECULAR MIMICRY
691
tetraacetic acid plus 0.1 mmol/L dithiothreitol and RPMI 1640 medium containing 0.05% (wt/vol) collagenase, and cells were isolated. Cells were counted, and viability was determined by the trypan blue exclusion assay. Cells were frozen using a Planar Kryo 10 controlled rate freezer (Planar Products Ltd., Middlesex, England) and stored in liquid nitrogen.
Proteinase K and Sodium Metaperiodate Treatment Ferret gastric cells, H. mustelae cells, and H. pylori cells were incubated with proteinase K (50 µg/200 µg of protein in the cell suspension) for 1 hour at 60°C or with 50 mmol/L sodium acetate (pH 4.5) alone or containing 10 mmol/L sodium metaperiodate for 1 hour at room temperature in the dark. Cells were then lysed and analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and Western immunoblotting.
SDS-PAGE and Western Immunoblotting Gastric cells, H. mustelae cells, and H. pylori cells were lysed by boiling in sample buffer (Tris/mercapthoethanol) and separated by SDS-PAGE (10% polyacrylamide), and proteins were transferred to nitrocellulose. Membranes were probed with H. mustelae and H. pylori rabbit polyclonal antiserum and monoclonal antibodies raised against blood group antigens A, B, H (Dako, Carpinteria, CA), Lewis x (Dako), or Lewis y (Signet, Dedham, MD). Antigen antibody complexes were detected using either anti-mouse (Sigma Chemical Co., St. Louis, MO) or anti-rabbit (Sigma) antibodies. Blots were developed using enhanced chemiluminescence (Amersham Corp., Arlington Heights, IL).
Absorption of Serum Serum was diluted 1:100 in PBS and H. mustelae cells (1012/mL), or a mixture of ferret and rabbit gastric epithelial cells (107/mL) were added. The suspensions were mixed well and left at 4°C overnight with shaking. The cells were removed by centrifugation at 10,000g for 5 minutes, and the supernatant was aliquoted and stored at 220°C.
Flow Cytometry Measurement of cross-reactivity of H. mustelae–specific antibodies with ferret gastric cells was made with a FACScan flow cytometer (Becton Dickinson, San Jose, CA). Ferret gastric cells were washed once in PBS and pelleted by centrifugation at 200g for 5 minutes. Cells were then stained with a 1:100 dilution of H. mustelae–specific antiserum raised in a rabbit. The expression of blood group antigens on gastric cells was assessed similarly using monoclonal antibodies against blood group antigens A, B, and H. Preimmune rabbit serum and normal mouse immunoglobulins (Dako) were used as negative controls. Cells were washed twice and stained with a 1:80 dilution of fluorescein isothiocyanate–conjugated goat anti-rabbit or goat anti-mouse immunoglobulin G (Sigma Chemical Co.). Cells were then washed twice in PBS and analyzed by flow cytometry. A total of 10,000 events was
692
´ CRO ´ INI´N ET AL. O
GASTROENTEROLOGY Vol. 114, No. 4
collected, and analysis of the data was performed using the Lysis II software program from Becton Dickinson. This program produces histograms of each cell sample and calculates the mean channel fluorescence of the cell population, which relates directly to the amount of antibody bound to each cell. Mean fluorescence values of cells stained with the various antibodies were then compared.
Immunohistochemistry Antral gastric biopsy specimens were taken from the stomachs of four 10–12-week-old ferrets by endoscopy using a pediatric bronchoscope and biopsy forceps,20 fixed in formalin, and paraffin embedded. Sections were then cut using a microtome and mounted on glass slides. Slides were deparaffinated in xylene and rehydrated in graded ethanol solutions. Internal peroxidase was inactivated by incubation in 0.3% hydrogen peroxidase in methanol. Normal goat serum was used to block nonspecific binding of the secondary antibody. Slides were stained with H. mustelae–specific antiserum, preimmune serum, monoclonal antibodies raised against blood group antigen A (Dako), or normal mouse immunoglobulins (Dako). The secondary antibody used was either anti-mouse (Sigma) or anti-rabbit (Sigma) conjugated to peroxidase. All antibodies were used at a 1:100 dilution. Slides were developed for 5 minutes with 3,38-diaminobenzidine tetrahydrochloride medium and counterstained with Weigerts hematoxylin (Sigma). After dehydration, slides were mounted and analyzed using a light microscope.
Results Cross-reactivity Between H. mustelae–Specific Antibodies and Ferret Gastric Cells As determined by Western immunoblotting, H. mustelae–specific antibodies reacted with total cell lysates of ferret gastric, duodenal, and colonic epithelial cells, giving diffuse bands ranging from 60 kilodaltons upwards. These bands were also observed on rabbit gastric epithelial cells but not on human gastric epithelial cells. Preimmune serum showed no reaction with either H. mustelae or epithelial cell lysates (Figure 1). Flow cytometry confirmed the reactivity of H. mustelae– specific antibodies with ferret gastric epithelial cells (Figure 2). H. mustelae–specific serum (Figure 2A) gave a much greater fluorescence (mean channel number, 369.86) than cells stained with preimmune serum from the same rabbit (Figure 2B) (mean channel number, 13.62). Removal of Cross-reactive Antibodies by Absorption With Epithelial Cells and H. mustelae Whole Cells H. mustelae–specific antiserum was absorbed with either a mixture of ferret and rabbit gastroduodenal epithelial cells or H. mustelae whole cells. The absorbed
Figure 1. Analysis of gastric epithelial cells and H. mustelae strain 12198 using whole-cell antiserum raised in rabbits against H. mustelae strain 12198 and preimmune serum. Lanes 1 and 7, H. mustelae strain 12198; lanes 2 and 8, ferret gastric epithelial cells; lane 3, ferret duodenal epithelial cells; lane 4, ferret colonic epithelial cells; lane 5, rabbit gastric epithelial cells; and lane 6, human gastric epithelial cells. Lanes 1–6 were incubated with H. mustelae antiserum, and lanes 7 and 8 were incubated with preimmune serum. Molecular masses are indicated (in kilodaltons) on the left.
serum was then analyzed using flow cytometry (Figure 2) and Western immunoblotting (Figure 3) to investigate whether absorption would remove all cross-reactivity with ferret gastric epithelial cells. Absorption of the H. mustelae antiserum with ferret and rabbit gastroduodenal epithelial cells removed all crossreactivity with ferret gastric cells as shown by flow cytometry (Figure 2). Western immunoblotting (Figure 3) confirmed this result and also showed the removal of a 14–20-kilodalton band reacting on H. mustelae. Absorption of the serum with H. mustelae whole cells also removed all cross-reactivity with ferret gastric epithelial cells as shown by Western immunoblotting. Identification of Blood Group Antigens on H. mustelae, H. pylori, and Ferret Gastric Epithelial Cells H. pylori has been shown previously to express Lewis x and Lewis y blood group antigens that are also found on human gastric epithelial cells. Therefore, we investigated whether H. mustelae expressed any blood group antigens by Western immunoblotting using monoclonal antibodies (Figure 4). Both Lewis x and Lewis y antibodies reacted with the H. pylori cell lysate with diffuse bands ranging from 30 to 60 kilodaltons. No bands were detected in lanes with H. mustelae or ferret gastric epithelial cell lysates, indicating that neither expressed Lewis x or Lewis y. However, both H. mustelae and ferret gastric epithelial cell lysates reacted with a monoclonal antibody raised against blood group antigen A. The pattern of staining on the ferret gastric epithelial cells was identical to that observed using the H. mustelae
April 1998
H. MUSTELAE AND MOLECULAR MIMICRY
693
Figure 2. Flow-cytometric analysis of ferret gastric cells stained with (A ) H. mustelae antiserum, (B ) preimmune serum, and (C ) H. mustelae antiserum absorbed with ferret and rabbit epithelial cells. The mean fluorescence intensity is indicated in the top right of each histogram.
antiserum (lane 3, Figure 3), with bands ranging from 60 kilodaltons upwards. The band recognized on the H. mustelae cell lysate was the 14–20-kilodalton band removed by absorption of the H. mustelae antiserum with gastroduodenal epithelial cells (lane 2, Figure 3). This indicated that the antigen recognized by the H. mustelae antiserum had a blood group antigen A–like structure. This structure was present on both H. mustelae strains NCTC 12198 and 12032. No staining was observed on the H. pylori cell lysate using anti–blood group antigen A antibodies. Monoclonal antibodies raised against blood groups B and H were also tested, and no reaction was observed on either H. mustelae or H. pylori (data not shown). Blood group antigen A expression on ferret gastric
Figure 3. Detection of cross-reacting bands using whole cell antiserum raised against H. mustelae strain 12198 adsorbed with H. mustelae whole cells or rabbit and ferret epithelial cells. Lanes 1 and 2, H. mustelae strain 12198; and lanes 3–5, ferret gastric epithelial cells. Lanes 1 and 3 were incubated with H. mustelae antiserum, lanes 2 and 4 were incubated with H. mustelae antiserum adsorbed with rabbit and ferret epithelial cells, and lane 5 was incubated with H. mustelae antiserum adsorbed with H. mustelae strain 12198. Molecular masses are indicated (in kilodaltons) on the left.
epithelial cells was also shown using flow cytometry (Figure 5). Characterization of Epitopes Recognized by Cross-reactive Antibodies Lewis x and Lewis y have been shown to be expressed as part of the O antigen regions of LPS of some H. pylori strains. We therefore investigated whether the blood group antigen A–like epitope expressed by H. mustelae was part of the LPS on the surface of the bacteria. We used proteinase K to digest proteins and sodium metaperiodate to oxidize sugar residues (Figure 6). In the case of ferret gastric epithelial cells, the epitope was removed by digestion with proteinase K or by oxidation of sugars using sodium metaperiodate. However, in the case of H. mustelae whole cells, treatment with sodium metaperiodate removed the 14–20-kilodalton band, whereas proteinase K treatment removed all other bands apart from the 14–20-kilodalton band. This
Figure 4. Identification of blood group antigens on H. mustelae, H. pylori, and ferret gastric cells. Lanes 1, 4, and 7, ferret gastric cells; lanes 2, 5, and 8, H. mustelae strain 12198; and lanes 3, 6, and 9, H. pylori strain PU3. Lanes 1–3 were incubated with monoclonal antibodies against blood group antigen A, lanes 4–6 were incubated with monoclonal antibodies against Lewis x, and lanes 7–9 were incubated with monoclonal antibodies against Lewis y. The molecular masses (in kilodaltons) are indicated on the left.
694
´ CRO ´ INI´N ET AL. O
GASTROENTEROLOGY Vol. 114, No. 4
suggests that the blood group antigen A–like epitope on ferret gastric cells was part of a glycoprotein(s), whereas the epitope on H. mustelae appeared to be part of the LPS. Immunohistochemical Staining of Ferret Gastric Tissue Immunohistochemistry was used to compare the staining of sections of ferret gastric mucosa by H. mustelae–specific antibodies and monoclonal antibodies to blood group antigen A and also to investigate whether the staining of each was specific for epithelial cells (Figure 7). Both H. mustelae–specific antiserum and anti–blood group A antibodies showed a similar pattern of staining of epithelial cells. Epithelial cells from the superficial, neck, and glandular epithelium were all stained by both antibodies, whereas no staining was observed using preimmune rabbit serum or normal mouse immunoglobulins. Identification of Anti–Blood Group A Antibodies in Cross-reactive Antiserum
Figure 5. Flow-cytometric analysis of ferret gastric cells stained with (A ) monoclonal antibodies raised against blood group antigen A and (B ) normal mouse immunoglobulins. The mean fluorescence intensity is indicated in the top right of each histogram.
A
B
Figure 6. (A ) Treatment of ferret gastric cells with proteinase K and sodium metaperiodate. Lane 1, ferret gastric cells; lane 2, ferret gastric cells incubated with sodium metaperiodate; and lane 3, ferret gastric cells incubated with proteinase K. (B ) Treatment of H. mustelae with proteinase K and sodium metaperiodate. Lane 1, H. mustelae strain 12198; lane 2, H. mustelae 12198 treated with sodium metaperiodate; and lane 3, H. mustelae 12198 treated with proteinase K. All lanes were incubated with anti–H. mustelae serum. Molecular masses (in kilodaltons) are indicated on the left.
H. mustelae antiserum and preimmune serum were tested for the presence of anti-A and -B antibodies by antibody titration against red blood cells. Preimmune serum was shown to have antibody titers of 2 and 1 for blood group A and B antibodies, respectively. However, the postimmune serum showed antibody titers of 16 and 2, respectively, showing a large increase in anti-A antibodies after immunization with H. mustelae.
Discussion In this study, we have shown cross-reactivity between H. mustelae–specific antibodies and ferret gastric epithelial cells, with cross-reactive antigens reacting to a blood group antigen A–like structure found on both cell surfaces. This suggests that the ferret may be an excellent animal model for studying the relevance of molecular mimicry between Helicobacter species and the gastric mucosa, which may be of significance in relation to possible immune evasion that has been suggested as an explanation for the persistence of H. pylori in the gastric mucosa.10 This cross-reactivity is extremely important in relation to H. pylori and the human gastric mucosa where such autoantibodies have been proposed as being of significance in relation to the development of chronic atrophic gastritis.17 Ferrets infected with H. mustelae have also been shown to develop multifocal atrophic gastritis,5 suggesting that a similar mechanism of autoantibody production may be used. However, the cross-reactivity was not restricted to gastric tissue because the antibodies did cross-react with ferret duodenal and colonic epithelial cells. This is in contrast with a previous study in which
April 1998
H. MUSTELAE AND MOLECULAR MIMICRY
695
Figure 7. Cross-reactivity of various antibodies with ferret gastric epithelial cells. (A ) Ferret gastric tissue stained with preimmune serum from a rabbit. (B ) Ferret gastric tissue stained with H. mustelae antiserum raised in a rabbit. (C ) Ferret gastric tissue stained with monoclonal antibodies against human blood group A. (D ) Ferret gastric epithelia stained with normal mouse immunoglobulin G.
H. pylori antibodies raised in mice were shown to cross-react only with gastric mucosa.16 Cross-reactivity was not observed with the sample of human gastric mucosa. However, these samples were also shown to have no reactivity with monoclonal antibodies against blood group antigen A (data not shown), suggesting that the individual was blood group A negative. The crossreactivity was shown to be specific for H. mustelae antibodies because absorption of the serum with H. mustelae whole cells removed cross-reactivity and because no reactivity was observed using preimmune serum. Although other antigens may be involved, cross-reactive antibodies were shown to be directed to a blood group A–like antigen structure, with the epitope on H. mustelae being characterized as part of the LPS and with the epitopes on the ferret gastric tissue being a range of high-molecular-weight glycoproteins. The titer of antibodies directed against blood group antigen A was also shown to increase in the postimmune serum compared with the preimmune serum. It has been shown previously that A, B, and H blood group structures are present on certain gram-negative bacteria and that the presence of such antigens is thought
to be connected with the origins of human anti-A and anti-B isoantibodies.21 However, the expression of a blood group antigen A–like structure as part of its LPS is of particular significance in the case of H. mustelae because it may represent a similar method of molecular mimicry as that shown by H. pylori.13 Infection with H. pylori is thought to occur early in life, and despite a strong local and systemic immune response,9 the infection is lifelong in most individuals. The organism is highly adapted to the environment in the stomach22 and may have developed mechanisms to avoid clearance by the immune system. Previous studies have shown that ferrets also become infected with H. mustelae at an early age and that almost 100% of adult ferrets are infected,4 suggesting that H. mustelae must also have a mechanism of evading the immune response. The fucosylation of LPS on the surface of H. pylori is thought to be a form of molecular mimicry used to evade the immune system. Various strains of H. pylori have been shown to express Lewis x, Lewis y, Lewis a, Lewis b, and sialyl Lewis x.15 It is speculated that the expression by H. pylori of antigens commonly found on human gastric epithelial cells may mask the bacterium from the immune system13
696
´ CRO ´ INI´N ET AL. O
and may ensure its survival in the gastric mucosa. Thus, the fucosylation of the LPS of H. mustelae may also play a role in masking the bacteria from the immune system. The expression of Lewis x by H. pylori has also been suggested to be connected to the process of ulcerogenesis, and Lewis x expression has been shown to be more common among isolates from patients with ulcer disease than from those without the disease.15 H. mustelae infection is thought to be associated with ulcerogenesis in the ferret and, thus, may also prove to be a good model to investigate the relationship between strains with fucosylated LPS and ulcerogenesis. In our study, two of two strains were shown to express the blood group antigen A–like structure as part of their LPS. It would be interesting to see if other H. mustelae strains express any other blood group antigens or whether more virulent strains express blood group antigen A. The expression of Lewis antigens by H. pylori is also thought to be involved in the production of autoantibodies10 that may play an important role in the pathogenesis of gastritis and gastric atrophy. Balb/c mice inoculated with hybridoma cells secreting H. pylori–specific antibodies have been shown to develop histopathologic features in the gastric mucosa.16 However, an animal model involving natural infection leading to molecular mimicry or secretion of autoantibodies would be very important in furthering our understanding of the part played by Lewis antigens in the pathogenesis of H. pylori infection. Experiments are ongoing to investigate whether autoantibodies are present in the serum of H. mustelae–positive ferrets and to investigate the nature of any antigens recognized by such autoantibodies.
References 1. Krakowka S, Morgan DR, Kraft WG, Leunk RD. Establishment of gastric Campylobacter pylori infection in the neonatal gnotobiotic piglet. Infect Immun 1987;55:2789–2796. 2. Marchetti M, Arico B, Burroni D, Figura N, Rappuoli R, Ghiara P. Development of a mouse model of Helicobacter pylori infection that mimics human disease [see comments]. Science 1995;267: 1655–1658. 3. Fox JG, Edrise BM, Cabot EB, Beaucage C, Murphy JC, Prostak KS. Campylobacter-like organisms isolated from gastric mucosa of ferrets. Am J Vet Res 1986;47:236–239. 4. Fox JG, Cabot EB, Taylor NS, Laraway R. Gastric colonization by Campylobacter pylori subsp. mustelae in ferrets. Infect Immun 1988;56:2994–2996. 5. Fox JG, Correa P, Taylor NS, Lee A, Otto G, Murphy JC, Rose R. Helicobacter mustelae–associated gastritis in ferrets. An animal model of Helicobacter pylori gastritis in humans. Gastroenterology 1990;99:352–361. 6. Dunn BE, Sung CC, Taylor NS, Fox JG. Purification and characterization of Helicobacter mustelae urease. Infect Immun 1991;59: 3343–3345. 7. Suerbaum S, Josenhans C, Labigne A. Cloning and genetic characterization of the Helicobacter pylori and Helicobacter mustelae flaB flagellin genes and construction of H. pylori flaA- and flaB-negative mutants by electroporation-mediated allelic exchange. J Bacteriol 1993;175:3278–3288.
GASTROENTEROLOGY Vol. 114, No. 4
8. Andruitas K, Fox JG, Schauer DB. Identification of a cag A gene in Helicobacter mustelae (abstr). Gut 1995;37(Suppl 1):A30. 9. Rathbone BJ, Wyatt JI, Worsley BW, Shires SE, Trejdosiewicz LK, Losowsky MS. Systemic and local antibody responses to gastric Campylobacter pyloridis in non-ulcer dyspepsia. Gut 1986;27:642– 647. 10. Appelmelk BJ, Simoons-Smit I, Negrini R, Moran AP, Aspinall GO, Forte JG, De Vries T, Quan H, Verboom T, Maaskant JJ, Ghiara P, Kuipers EJ, Bloemena E, Tadema TM, Townsend RR, Tyagarajan K, Crothers JM, Montiero MA, Savio A, De Graaff J. Potential role of molecular mimicry between Helicobacter pylori lipopolysaccharide and host Lewis blood group antigens in autoimmunity. Infect Immun 1996;64:2031–2040. 11. Sherburne R, Taylor DE. Helicobacter pylori expresses a complex surface carbohydrate, Lewis X. Infect Immun 1995;63:4564–4568. 12. Moran AP. The role of lipopolysaccharide in Helicobacter pylori pathogenesis. Aliment Pharmacol Ther 1996;10(Suppl 1):39–50. 13. Moran AP, Prendergast MM, Appelmelk BJ. Molecular mimicry of host structures by bacterial lipopolysaccharides and its contribution to disease. FEMS Immunol Med Microbiol 1997;16:105– 115. 14. Simoons-Smit IM, Appelmelk BJ, Verboom T, et al. Typing of Helicobacter pylori with monoclonal antibodies against Lewis antigens in lipopolysaccharide. J Clin Microbiol 1996;34:2196– 2200. 15. Wirth HP, Yang M, Karita M, Blaser MJ. Expression of the human cell surface glycoconjugates Lewis x and Lewis y by Helicobacter pylori isolates is related to cagA status. Infect Immun 1996;64: 4598–4605. 16. Negrini R, Lisato L, Zanella I, Gavazzini L, Gullini S, Villanacci V, Poiesi C, Abertini A, Ghielmi S. Helicobacter pylori infection induces antibodies cross-reacting with human gastric mucosa. Gastroenterology 1991;101:437–445. 17. Negrini R, Savio A, Poiesi C, Appelmelk BJ, Buffoli F, Paterlini A, Cesari P, Graffeo M, Vaira D, Franzin G. Antigenic mimicry between Helicobacter pylori and gastric mucosa in the pathogenesis of body atrophic gastritis. Gastroenterology 1996;111:655– 665. 18. Faller G, Steininger H, Eck M, Hensen J, Hann EG, Kirchner T. Antigastric autoantibodies in Helicobacter pylori gastritis: prevalence, in-situ binding sites and clues for clinical relevance. Virchows Arch 1996;427:483–486. 19. Clyne M, Drumm B. Adherence of Helicobacter pylori to primary human gastrointestinal cells. Infect Immun 1993;61:4051– 4057. 20. Batchelder M, Fox JG, Hayward A, Yan L, Shames B, Murphy JC, Palley L. Natural and experimental Helicobacter mustelae reinfection following successful antimicrobial eradication in ferrets. Helicobacter 1996;1:34–42. 21. Springer GF, Williamson P, Brandes WC. Blood group activity of gram-negative bacteria. J Exp Med 1961;113:1077–1093. 22. Clyne M, Labigne A, Drumm B. Helicobacter pylori requires an acidic environment to survive in the presence of urea. Infect Immun 1995;63:1669–1673.
Received July 22, 1997. Accepted December 9, 1997. Address requests for reprints to: Marguerite Clyne, Ph.D., Department of Paediatrics, University College Dublin, The Children’s Research Centre, Our Lady’s Hospital for Sick Children, Crumlin, Dublin 12, Ireland. Fax: (353) 1-4555307. Supported by grants from the Health Research Board, Ireland, and The Children’s Research Centre, Our Lady’s Hospital for Sick Children, Dublin, Ireland. The authors thank Annette Power for performing the blood group titrations, Francis Owens for paraffin embedding and cutting the tissue, and Dr. Anthony Moran for helpful discussions.
Molecular Mimicry of Ferret Gastric Epithelial Blood Group Antigen A by Helicobacter mustelae ´ CRO ´ INI´N, MARGUERITE CLYNE, and BRENDAN DRUMM TADHG O Department of Paediatrics, University College Dublin, The Children’s Research Centre, Our Lady’s Hospital for Sick Children, Dublin, Ireland
Background & Aims: Molecular mimicry of Lewis blood group antigens by Helicobacter pylori may be involved in immune evasion by the bacteria and in the pathogenesis of chronic atrophic gastritis. Helicobacter mustelae infects ferrets naturally, causing gastritis, and may be involved in ulcerogenesis. The aim of this study was to determine if H. mustelae shows a similar form of molecular mimicry. Methods: Antibodies raised against H. mustelae were used to stain ferret gastric tissue by immunoblotting, immunohistochemistry, and flow cytometry. Epitopes recognized by cross-reactivity were characterized by proteinase K and sodium metaperiodate treatment. Results: H. mustelae antiserum reacted with H. mustelae and with ferret gastric tissue. Absorption of the antiserum with H. mustelae or ferret and rabbit gastric tissue removed the cross-reactive antibodies. Antibodies reacted with a blood group antigen A–like structure on ferret gastric epithelial cells and H. mustelae lipopolysaccharide. Conclusions: H. mustelae expresses a blood group–like antigen as part of its lipopolysaccharide that may be used as a method of immune evasion by mimicry of gastric epithelial cells. The cross-reactivity shown by H. mustelae–specific antibodies with gastric mucosa may suggest a role for autoantibodies in the pathogenesis of H. mustelae–induced gastritis in ferrets.
elicobacter pylori has been recognized as a major cause of gastritis and duodenal ulcer disease in humans. Although several virulence factors have been identified, the pathogenesis of H. pylori–induced gastritis is still poorly understood. Several animal models have been proposed to study H. pylori infection in vivo, including H. pylori infection of the gnotobiotic piglet1 or the recently described mouse model.2 However, H. pylori does not infect any of these animals naturally. Helicobacter mustelae is a gastric pathogen that infects ferrets naturally, colonizing the gastric mucosa. Infection is associated with gastric inflammation and is believed to contribute to ulcerogenesis.3,4 H. mustelae–induced gastritis in ferrets has been shown to closely resemble many aspects of H. pylori gastritis in humans.5 Several virulence factors of H. pylori have also been shown to be expressed by
H
H. mustelae, including a potent urease enzyme,6 sheathed flagellae,7 intimate adherence to gastric epithelial cells,5 and a CagA homologue.8 These factors indicate that H. mustelae infection of the ferret may be a useful natural animal model in which to study the pathogenesis of gastritis and duodenal ulcer disease caused by a Helicobacter infection. Despite a strong local and systemic inflammatory response induced by H. pylori,9 infection can persist throughout life and leads to the development of chronic gastritis. A recent study suggested that H. pylori may evade the immune response through a mechanism of molecular mimicry.10 O antigen regions of lipopolysaccharide (LPS) from some H. pylori strains have been shown to be structurally similar to the blood group antigens Lewis x and Lewis y that are expressed on the gastric mucosa of humans.11,12 It has been suggested that these Lewis blood group antigens could potentially mask H. pylori from the immune system.13 Greater than 85% of strains obtained from various parts of the world have been shown to express Lewis antigens.14 However, expression of the Lewis antigens has also been shown to be more common in strains isolated from patients with ulcer disease than from those without the disease.15 Many patients infected with H. pylori have autoantibodies that cross-react with antigens expressed on gastric epithelial cells.16 Recent reports suggested that the presence of autoantibodies may play a crucial role in the pathogenesis of chronic atrophic gastritis. 17 Both H. pylori–associated Lewis x and Lewis y have been shown to induce autoantibodies in humans, and the gastric H1, K1–adenosine triphosphatase found in the canaliculi of parietal cells has been identified as a possible target of the autoimmune response.10,17,18 The aim of this study was to investigate whether H. mustelae possessed a similar mechanism of molecular mimicry as that described for H. pylori. Abbreviation used in this paper: LPS, lipopolysaccharide. r 1998 by the American Gastroenterological Association 0016-5085/98/$3.00
April 1998
Materials and Methods Bacterial Strains H. pylori strain PU3 was isolated from antral biopsy material obtained from a child undergoing upper gastrointestinal endoscopy. H. mustelae strains NCTC 12198 and NCTC 12032 were obtained from the National Collection of Type Cultures (Public Health Laboratory Service, London, England). Both H. pylori and H. mustelae were cultured on Columbia blood agar plates (Oxoid, Columbia, MD) containing 7% defibrinated horse blood for 3 days at 37°C in an atmosphere of 5% O2 and 10% CO2.
Raising of Specific Anti–H. mustelae Antibodies H. mustelae strain NCTC 12198 was harvested from blood agar plates and fixed in phosphate-buffered saline (PBS) containing 0.5% (vol/vol) formaldehyde at a concentration of 1 g of bacteria (wet weight) per 100 mL of solution. For initial immunization, the bacterial suspension was suspended in Freund’s complete adjuvant and was injected intradermally into rabbits. Two subsequent boosters were suspended in Freund’s incomplete adjuvant and injected intramuscularly at 2-week intervals. Test bleeds were taken from the rabbits after the second booster. The serum reacted with a H. mustelae whole-cell lysate by Western immunoblotting at a 1:500 dilution. Rabbits were bled out, the blood was allowed to clot, and the serum was removed.
Antibody Titration H. mustelae antiserum and preimmune serum were tested for the presence of anti-A and anti-B antibodies by titration against blood group A, B, and O red blood cells. Serum was diluted serially in microtiter plates, and red blood cells of blood group A, B, or O were added. The suspension was mixed, spun, and resuspended, and agglutination patterns were recorded.
Isolation of Gastric, Colonic, and Duodenal Epithelial Cells Biopsy tissue was obtained from children undergoing endoscopy at Our Lady’s Hospital for Sick Children, Dublin, and gastric, duodenal, and colonic cells were isolated from biopsy specimens as described previously.19 Ethical approval for this study was obtained from the ethics committee at the hospital. Ferret and rabbit gastric duodenal and colonic tissues were obtained by killing 10–12-week-old ferrets or adult rabbits and removing the stomach, duodenum, and colon. All ferrets were negative for H. mustelae by culture and urease activity of biopsy specimens taken from the antrum, fundus, and duodenum (three samples from each site). The serosa was separated from the tissue by injecting RPMI 1640 tissue culture medium containing 10% (vol/vol) fetal calf serum underneath the serosa and peeling it away. The remaining tissue was minced using scalpels and incubated with Hank’s balanced salt solution containing 0.1 mmol/L ethylenediamine-
H. MUSTELAE AND MOLECULAR MIMICRY
691
tetraacetic acid plus 0.1 mmol/L dithiothreitol and RPMI 1640 medium containing 0.05% (wt/vol) collagenase, and cells were isolated. Cells were counted, and viability was determined by the trypan blue exclusion assay. Cells were frozen using a Planar Kryo 10 controlled rate freezer (Planar Products Ltd., Middlesex, England) and stored in liquid nitrogen.
Proteinase K and Sodium Metaperiodate Treatment Ferret gastric cells, H. mustelae cells, and H. pylori cells were incubated with proteinase K (50 µg/200 µg of protein in the cell suspension) for 1 hour at 60°C or with 50 mmol/L sodium acetate (pH 4.5) alone or containing 10 mmol/L sodium metaperiodate for 1 hour at room temperature in the dark. Cells were then lysed and analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and Western immunoblotting.
SDS-PAGE and Western Immunoblotting Gastric cells, H. mustelae cells, and H. pylori cells were lysed by boiling in sample buffer (Tris/mercapthoethanol) and separated by SDS-PAGE (10% polyacrylamide), and proteins were transferred to nitrocellulose. Membranes were probed with H. mustelae and H. pylori rabbit polyclonal antiserum and monoclonal antibodies raised against blood group antigens A, B, H (Dako, Carpinteria, CA), Lewis x (Dako), or Lewis y (Signet, Dedham, MD). Antigen antibody complexes were detected using either anti-mouse (Sigma Chemical Co., St. Louis, MO) or anti-rabbit (Sigma) antibodies. Blots were developed using enhanced chemiluminescence (Amersham Corp., Arlington Heights, IL).
Absorption of Serum Serum was diluted 1:100 in PBS and H. mustelae cells (1012/mL), or a mixture of ferret and rabbit gastric epithelial cells (107/mL) were added. The suspensions were mixed well and left at 4°C overnight with shaking. The cells were removed by centrifugation at 10,000g for 5 minutes, and the supernatant was aliquoted and stored at 220°C.
Flow Cytometry Measurement of cross-reactivity of H. mustelae–specific antibodies with ferret gastric cells was made with a FACScan flow cytometer (Becton Dickinson, San Jose, CA). Ferret gastric cells were washed once in PBS and pelleted by centrifugation at 200g for 5 minutes. Cells were then stained with a 1:100 dilution of H. mustelae–specific antiserum raised in a rabbit. The expression of blood group antigens on gastric cells was assessed similarly using monoclonal antibodies against blood group antigens A, B, and H. Preimmune rabbit serum and normal mouse immunoglobulins (Dako) were used as negative controls. Cells were washed twice and stained with a 1:80 dilution of fluorescein isothiocyanate–conjugated goat anti-rabbit or goat anti-mouse immunoglobulin G (Sigma Chemical Co.). Cells were then washed twice in PBS and analyzed by flow cytometry. A total of 10,000 events was
692
´ CRO ´ INI´N ET AL. O
GASTROENTEROLOGY Vol. 114, No. 4
collected, and analysis of the data was performed using the Lysis II software program from Becton Dickinson. This program produces histograms of each cell sample and calculates the mean channel fluorescence of the cell population, which relates directly to the amount of antibody bound to each cell. Mean fluorescence values of cells stained with the various antibodies were then compared.
Immunohistochemistry Antral gastric biopsy specimens were taken from the stomachs of four 10–12-week-old ferrets by endoscopy using a pediatric bronchoscope and biopsy forceps,20 fixed in formalin, and paraffin embedded. Sections were then cut using a microtome and mounted on glass slides. Slides were deparaffinated in xylene and rehydrated in graded ethanol solutions. Internal peroxidase was inactivated by incubation in 0.3% hydrogen peroxidase in methanol. Normal goat serum was used to block nonspecific binding of the secondary antibody. Slides were stained with H. mustelae–specific antiserum, preimmune serum, monoclonal antibodies raised against blood group antigen A (Dako), or normal mouse immunoglobulins (Dako). The secondary antibody used was either anti-mouse (Sigma) or anti-rabbit (Sigma) conjugated to peroxidase. All antibodies were used at a 1:100 dilution. Slides were developed for 5 minutes with 3,38-diaminobenzidine tetrahydrochloride medium and counterstained with Weigerts hematoxylin (Sigma). After dehydration, slides were mounted and analyzed using a light microscope.
Results Cross-reactivity Between H. mustelae–Specific Antibodies and Ferret Gastric Cells As determined by Western immunoblotting, H. mustelae–specific antibodies reacted with total cell lysates of ferret gastric, duodenal, and colonic epithelial cells, giving diffuse bands ranging from 60 kilodaltons upwards. These bands were also observed on rabbit gastric epithelial cells but not on human gastric epithelial cells. Preimmune serum showed no reaction with either H. mustelae or epithelial cell lysates (Figure 1). Flow cytometry confirmed the reactivity of H. mustelae– specific antibodies with ferret gastric epithelial cells (Figure 2). H. mustelae–specific serum (Figure 2A) gave a much greater fluorescence (mean channel number, 369.86) than cells stained with preimmune serum from the same rabbit (Figure 2B) (mean channel number, 13.62). Removal of Cross-reactive Antibodies by Absorption With Epithelial Cells and H. mustelae Whole Cells H. mustelae–specific antiserum was absorbed with either a mixture of ferret and rabbit gastroduodenal epithelial cells or H. mustelae whole cells. The absorbed
Figure 1. Analysis of gastric epithelial cells and H. mustelae strain 12198 using whole-cell antiserum raised in rabbits against H. mustelae strain 12198 and preimmune serum. Lanes 1 and 7, H. mustelae strain 12198; lanes 2 and 8, ferret gastric epithelial cells; lane 3, ferret duodenal epithelial cells; lane 4, ferret colonic epithelial cells; lane 5, rabbit gastric epithelial cells; and lane 6, human gastric epithelial cells. Lanes 1–6 were incubated with H. mustelae antiserum, and lanes 7 and 8 were incubated with preimmune serum. Molecular masses are indicated (in kilodaltons) on the left.
serum was then analyzed using flow cytometry (Figure 2) and Western immunoblotting (Figure 3) to investigate whether absorption would remove all cross-reactivity with ferret gastric epithelial cells. Absorption of the H. mustelae antiserum with ferret and rabbit gastroduodenal epithelial cells removed all crossreactivity with ferret gastric cells as shown by flow cytometry (Figure 2). Western immunoblotting (Figure 3) confirmed this result and also showed the removal of a 14–20-kilodalton band reacting on H. mustelae. Absorption of the serum with H. mustelae whole cells also removed all cross-reactivity with ferret gastric epithelial cells as shown by Western immunoblotting. Identification of Blood Group Antigens on H. mustelae, H. pylori, and Ferret Gastric Epithelial Cells H. pylori has been shown previously to express Lewis x and Lewis y blood group antigens that are also found on human gastric epithelial cells. Therefore, we investigated whether H. mustelae expressed any blood group antigens by Western immunoblotting using monoclonal antibodies (Figure 4). Both Lewis x and Lewis y antibodies reacted with the H. pylori cell lysate with diffuse bands ranging from 30 to 60 kilodaltons. No bands were detected in lanes with H. mustelae or ferret gastric epithelial cell lysates, indicating that neither expressed Lewis x or Lewis y. However, both H. mustelae and ferret gastric epithelial cell lysates reacted with a monoclonal antibody raised against blood group antigen A. The pattern of staining on the ferret gastric epithelial cells was identical to that observed using the H. mustelae
April 1998
H. MUSTELAE AND MOLECULAR MIMICRY
693
Figure 2. Flow-cytometric analysis of ferret gastric cells stained with (A ) H. mustelae antiserum, (B ) preimmune serum, and (C ) H. mustelae antiserum absorbed with ferret and rabbit epithelial cells. The mean fluorescence intensity is indicated in the top right of each histogram.
antiserum (lane 3, Figure 3), with bands ranging from 60 kilodaltons upwards. The band recognized on the H. mustelae cell lysate was the 14–20-kilodalton band removed by absorption of the H. mustelae antiserum with gastroduodenal epithelial cells (lane 2, Figure 3). This indicated that the antigen recognized by the H. mustelae antiserum had a blood group antigen A–like structure. This structure was present on both H. mustelae strains NCTC 12198 and 12032. No staining was observed on the H. pylori cell lysate using anti–blood group antigen A antibodies. Monoclonal antibodies raised against blood groups B and H were also tested, and no reaction was observed on either H. mustelae or H. pylori (data not shown). Blood group antigen A expression on ferret gastric
Figure 3. Detection of cross-reacting bands using whole cell antiserum raised against H. mustelae strain 12198 adsorbed with H. mustelae whole cells or rabbit and ferret epithelial cells. Lanes 1 and 2, H. mustelae strain 12198; and lanes 3–5, ferret gastric epithelial cells. Lanes 1 and 3 were incubated with H. mustelae antiserum, lanes 2 and 4 were incubated with H. mustelae antiserum adsorbed with rabbit and ferret epithelial cells, and lane 5 was incubated with H. mustelae antiserum adsorbed with H. mustelae strain 12198. Molecular masses are indicated (in kilodaltons) on the left.
epithelial cells was also shown using flow cytometry (Figure 5). Characterization of Epitopes Recognized by Cross-reactive Antibodies Lewis x and Lewis y have been shown to be expressed as part of the O antigen regions of LPS of some H. pylori strains. We therefore investigated whether the blood group antigen A–like epitope expressed by H. mustelae was part of the LPS on the surface of the bacteria. We used proteinase K to digest proteins and sodium metaperiodate to oxidize sugar residues (Figure 6). In the case of ferret gastric epithelial cells, the epitope was removed by digestion with proteinase K or by oxidation of sugars using sodium metaperiodate. However, in the case of H. mustelae whole cells, treatment with sodium metaperiodate removed the 14–20-kilodalton band, whereas proteinase K treatment removed all other bands apart from the 14–20-kilodalton band. This
Figure 4. Identification of blood group antigens on H. mustelae, H. pylori, and ferret gastric cells. Lanes 1, 4, and 7, ferret gastric cells; lanes 2, 5, and 8, H. mustelae strain 12198; and lanes 3, 6, and 9, H. pylori strain PU3. Lanes 1–3 were incubated with monoclonal antibodies against blood group antigen A, lanes 4–6 were incubated with monoclonal antibodies against Lewis x, and lanes 7–9 were incubated with monoclonal antibodies against Lewis y. The molecular masses (in kilodaltons) are indicated on the left.
694
´ CRO ´ INI´N ET AL. O
GASTROENTEROLOGY Vol. 114, No. 4
suggests that the blood group antigen A–like epitope on ferret gastric cells was part of a glycoprotein(s), whereas the epitope on H. mustelae appeared to be part of the LPS. Immunohistochemical Staining of Ferret Gastric Tissue Immunohistochemistry was used to compare the staining of sections of ferret gastric mucosa by H. mustelae–specific antibodies and monoclonal antibodies to blood group antigen A and also to investigate whether the staining of each was specific for epithelial cells (Figure 7). Both H. mustelae–specific antiserum and anti–blood group A antibodies showed a similar pattern of staining of epithelial cells. Epithelial cells from the superficial, neck, and glandular epithelium were all stained by both antibodies, whereas no staining was observed using preimmune rabbit serum or normal mouse immunoglobulins. Identification of Anti–Blood Group A Antibodies in Cross-reactive Antiserum
Figure 5. Flow-cytometric analysis of ferret gastric cells stained with (A ) monoclonal antibodies raised against blood group antigen A and (B ) normal mouse immunoglobulins. The mean fluorescence intensity is indicated in the top right of each histogram.
A
B
Figure 6. (A ) Treatment of ferret gastric cells with proteinase K and sodium metaperiodate. Lane 1, ferret gastric cells; lane 2, ferret gastric cells incubated with sodium metaperiodate; and lane 3, ferret gastric cells incubated with proteinase K. (B ) Treatment of H. mustelae with proteinase K and sodium metaperiodate. Lane 1, H. mustelae strain 12198; lane 2, H. mustelae 12198 treated with sodium metaperiodate; and lane 3, H. mustelae 12198 treated with proteinase K. All lanes were incubated with anti–H. mustelae serum. Molecular masses (in kilodaltons) are indicated on the left.
H. mustelae antiserum and preimmune serum were tested for the presence of anti-A and -B antibodies by antibody titration against red blood cells. Preimmune serum was shown to have antibody titers of 2 and 1 for blood group A and B antibodies, respectively. However, the postimmune serum showed antibody titers of 16 and 2, respectively, showing a large increase in anti-A antibodies after immunization with H. mustelae.
Discussion In this study, we have shown cross-reactivity between H. mustelae–specific antibodies and ferret gastric epithelial cells, with cross-reactive antigens reacting to a blood group antigen A–like structure found on both cell surfaces. This suggests that the ferret may be an excellent animal model for studying the relevance of molecular mimicry between Helicobacter species and the gastric mucosa, which may be of significance in relation to possible immune evasion that has been suggested as an explanation for the persistence of H. pylori in the gastric mucosa.10 This cross-reactivity is extremely important in relation to H. pylori and the human gastric mucosa where such autoantibodies have been proposed as being of significance in relation to the development of chronic atrophic gastritis.17 Ferrets infected with H. mustelae have also been shown to develop multifocal atrophic gastritis,5 suggesting that a similar mechanism of autoantibody production may be used. However, the cross-reactivity was not restricted to gastric tissue because the antibodies did cross-react with ferret duodenal and colonic epithelial cells. This is in contrast with a previous study in which
April 1998
H. MUSTELAE AND MOLECULAR MIMICRY
695
Figure 7. Cross-reactivity of various antibodies with ferret gastric epithelial cells. (A ) Ferret gastric tissue stained with preimmune serum from a rabbit. (B ) Ferret gastric tissue stained with H. mustelae antiserum raised in a rabbit. (C ) Ferret gastric tissue stained with monoclonal antibodies against human blood group A. (D ) Ferret gastric epithelia stained with normal mouse immunoglobulin G.
H. pylori antibodies raised in mice were shown to cross-react only with gastric mucosa.16 Cross-reactivity was not observed with the sample of human gastric mucosa. However, these samples were also shown to have no reactivity with monoclonal antibodies against blood group antigen A (data not shown), suggesting that the individual was blood group A negative. The crossreactivity was shown to be specific for H. mustelae antibodies because absorption of the serum with H. mustelae whole cells removed cross-reactivity and because no reactivity was observed using preimmune serum. Although other antigens may be involved, cross-reactive antibodies were shown to be directed to a blood group A–like antigen structure, with the epitope on H. mustelae being characterized as part of the LPS and with the epitopes on the ferret gastric tissue being a range of high-molecular-weight glycoproteins. The titer of antibodies directed against blood group antigen A was also shown to increase in the postimmune serum compared with the preimmune serum. It has been shown previously that A, B, and H blood group structures are present on certain gram-negative bacteria and that the presence of such antigens is thought
to be connected with the origins of human anti-A and anti-B isoantibodies.21 However, the expression of a blood group antigen A–like structure as part of its LPS is of particular significance in the case of H. mustelae because it may represent a similar method of molecular mimicry as that shown by H. pylori.13 Infection with H. pylori is thought to occur early in life, and despite a strong local and systemic immune response,9 the infection is lifelong in most individuals. The organism is highly adapted to the environment in the stomach22 and may have developed mechanisms to avoid clearance by the immune system. Previous studies have shown that ferrets also become infected with H. mustelae at an early age and that almost 100% of adult ferrets are infected,4 suggesting that H. mustelae must also have a mechanism of evading the immune response. The fucosylation of LPS on the surface of H. pylori is thought to be a form of molecular mimicry used to evade the immune system. Various strains of H. pylori have been shown to express Lewis x, Lewis y, Lewis a, Lewis b, and sialyl Lewis x.15 It is speculated that the expression by H. pylori of antigens commonly found on human gastric epithelial cells may mask the bacterium from the immune system13
696
´ CRO ´ INI´N ET AL. O
and may ensure its survival in the gastric mucosa. Thus, the fucosylation of the LPS of H. mustelae may also play a role in masking the bacteria from the immune system. The expression of Lewis x by H. pylori has also been suggested to be connected to the process of ulcerogenesis, and Lewis x expression has been shown to be more common among isolates from patients with ulcer disease than from those without the disease.15 H. mustelae infection is thought to be associated with ulcerogenesis in the ferret and, thus, may also prove to be a good model to investigate the relationship between strains with fucosylated LPS and ulcerogenesis. In our study, two of two strains were shown to express the blood group antigen A–like structure as part of their LPS. It would be interesting to see if other H. mustelae strains express any other blood group antigens or whether more virulent strains express blood group antigen A. The expression of Lewis antigens by H. pylori is also thought to be involved in the production of autoantibodies10 that may play an important role in the pathogenesis of gastritis and gastric atrophy. Balb/c mice inoculated with hybridoma cells secreting H. pylori–specific antibodies have been shown to develop histopathologic features in the gastric mucosa.16 However, an animal model involving natural infection leading to molecular mimicry or secretion of autoantibodies would be very important in furthering our understanding of the part played by Lewis antigens in the pathogenesis of H. pylori infection. Experiments are ongoing to investigate whether autoantibodies are present in the serum of H. mustelae–positive ferrets and to investigate the nature of any antigens recognized by such autoantibodies.
References 1. Krakowka S, Morgan DR, Kraft WG, Leunk RD. Establishment of gastric Campylobacter pylori infection in the neonatal gnotobiotic piglet. Infect Immun 1987;55:2789–2796. 2. Marchetti M, Arico B, Burroni D, Figura N, Rappuoli R, Ghiara P. Development of a mouse model of Helicobacter pylori infection that mimics human disease [see comments]. Science 1995;267: 1655–1658. 3. Fox JG, Edrise BM, Cabot EB, Beaucage C, Murphy JC, Prostak KS. Campylobacter-like organisms isolated from gastric mucosa of ferrets. Am J Vet Res 1986;47:236–239. 4. Fox JG, Cabot EB, Taylor NS, Laraway R. Gastric colonization by Campylobacter pylori subsp. mustelae in ferrets. Infect Immun 1988;56:2994–2996. 5. Fox JG, Correa P, Taylor NS, Lee A, Otto G, Murphy JC, Rose R. Helicobacter mustelae–associated gastritis in ferrets. An animal model of Helicobacter pylori gastritis in humans. Gastroenterology 1990;99:352–361. 6. Dunn BE, Sung CC, Taylor NS, Fox JG. Purification and characterization of Helicobacter mustelae urease. Infect Immun 1991;59: 3343–3345. 7. Suerbaum S, Josenhans C, Labigne A. Cloning and genetic characterization of the Helicobacter pylori and Helicobacter mustelae flaB flagellin genes and construction of H. pylori flaA- and flaB-negative mutants by electroporation-mediated allelic exchange. J Bacteriol 1993;175:3278–3288.
GASTROENTEROLOGY Vol. 114, No. 4
8. Andruitas K, Fox JG, Schauer DB. Identification of a cag A gene in Helicobacter mustelae (abstr). Gut 1995;37(Suppl 1):A30. 9. Rathbone BJ, Wyatt JI, Worsley BW, Shires SE, Trejdosiewicz LK, Losowsky MS. Systemic and local antibody responses to gastric Campylobacter pyloridis in non-ulcer dyspepsia. Gut 1986;27:642– 647. 10. Appelmelk BJ, Simoons-Smit I, Negrini R, Moran AP, Aspinall GO, Forte JG, De Vries T, Quan H, Verboom T, Maaskant JJ, Ghiara P, Kuipers EJ, Bloemena E, Tadema TM, Townsend RR, Tyagarajan K, Crothers JM, Montiero MA, Savio A, De Graaff J. Potential role of molecular mimicry between Helicobacter pylori lipopolysaccharide and host Lewis blood group antigens in autoimmunity. Infect Immun 1996;64:2031–2040. 11. Sherburne R, Taylor DE. Helicobacter pylori expresses a complex surface carbohydrate, Lewis X. Infect Immun 1995;63:4564–4568. 12. Moran AP. The role of lipopolysaccharide in Helicobacter pylori pathogenesis. Aliment Pharmacol Ther 1996;10(Suppl 1):39–50. 13. Moran AP, Prendergast MM, Appelmelk BJ. Molecular mimicry of host structures by bacterial lipopolysaccharides and its contribution to disease. FEMS Immunol Med Microbiol 1997;16:105– 115. 14. Simoons-Smit IM, Appelmelk BJ, Verboom T, et al. Typing of Helicobacter pylori with monoclonal antibodies against Lewis antigens in lipopolysaccharide. J Clin Microbiol 1996;34:2196– 2200. 15. Wirth HP, Yang M, Karita M, Blaser MJ. Expression of the human cell surface glycoconjugates Lewis x and Lewis y by Helicobacter pylori isolates is related to cagA status. Infect Immun 1996;64: 4598–4605. 16. Negrini R, Lisato L, Zanella I, Gavazzini L, Gullini S, Villanacci V, Poiesi C, Abertini A, Ghielmi S. Helicobacter pylori infection induces antibodies cross-reacting with human gastric mucosa. Gastroenterology 1991;101:437–445. 17. Negrini R, Savio A, Poiesi C, Appelmelk BJ, Buffoli F, Paterlini A, Cesari P, Graffeo M, Vaira D, Franzin G. Antigenic mimicry between Helicobacter pylori and gastric mucosa in the pathogenesis of body atrophic gastritis. Gastroenterology 1996;111:655– 665. 18. Faller G, Steininger H, Eck M, Hensen J, Hann EG, Kirchner T. Antigastric autoantibodies in Helicobacter pylori gastritis: prevalence, in-situ binding sites and clues for clinical relevance. Virchows Arch 1996;427:483–486. 19. Clyne M, Drumm B. Adherence of Helicobacter pylori to primary human gastrointestinal cells. Infect Immun 1993;61:4051– 4057. 20. Batchelder M, Fox JG, Hayward A, Yan L, Shames B, Murphy JC, Palley L. Natural and experimental Helicobacter mustelae reinfection following successful antimicrobial eradication in ferrets. Helicobacter 1996;1:34–42. 21. Springer GF, Williamson P, Brandes WC. Blood group activity of gram-negative bacteria. J Exp Med 1961;113:1077–1093. 22. Clyne M, Labigne A, Drumm B. Helicobacter pylori requires an acidic environment to survive in the presence of urea. Infect Immun 1995;63:1669–1673.
Received July 22, 1997. Accepted December 9, 1997. Address requests for reprints to: Marguerite Clyne, Ph.D., Department of Paediatrics, University College Dublin, The Children’s Research Centre, Our Lady’s Hospital for Sick Children, Crumlin, Dublin 12, Ireland. Fax: (353) 1-4555307. Supported by grants from the Health Research Board, Ireland, and The Children’s Research Centre, Our Lady’s Hospital for Sick Children, Dublin, Ireland. The authors thank Annette Power for performing the blood group titrations, Francis Owens for paraffin embedding and cutting the tissue, and Dr. Anthony Moran for helpful discussions.