Evaluation of serology, 13C-urea breath test, and polymerase chain reaction of stool samples to detect Helicobacter pylori in Bangladeshi children.

BACKGROUND: Serologic methods to detect Helicobacter pylori in infants, especially in developing countries, may be limited because of decreased immune response caused by malnutrition. The true prevalence may therefore be underestimated in this age group. Urea breath test is considered to be a good screening method in children but is expensive and therefore is not suitable for screening in developing countries. Simple, inexpensive, and accurate noninvasive methods to detect H. pylori in infants and young children are needed. METHODS: Enzyme immunoassay (EIA) and immunoblot (IB) serologic analyses, 13C-urea breath test (UBT), and immunomagnetic separation--polymerase chain reaction (IMS-PCR) were performed on stool specimens, to detect H. pylori in 68 children between 4 and 24 months of age (mean, 11.5 months) in an endemic area in Bangladesh and the results compared. RESULTS: The occurrence of H. pylori was 57% (n=39) using only UBT, 60% (n=41) using only IMS-PCR, and 78% (n=53) using UBT and IMS-PCR together. The concordance between UBT and IMS-PCR results was 62%. Immunoblot was positive in only 9% (n=6). Results in all 68 children were negative using EIA. DISCUSSION: The prevalence of H. pylori infection in this periurban community and age group was high. Only serologic methods seem to be unsatisfactory for screening of H. pylori infection in infants and may not reflect the true prevalence. Immunomagnetic separation-PCR is a simple and rapid method for detection of H. pylori in stool and is an attractive method for analysis of colonization in infants. However, it may reflect a different stage of disease than UBT. Further studies are needed to clarify this.

Severe gastritis in guinea-pigs infected with Helicobacter pylori.

An appropriate animal model is essential to study Helicobacter pylori infection. The aim of this study was to investigate if H. pylori can colonise the guinea-pig stomach and whether the infection causes gastritis and a serological response similar to that observed in man. Guinea-pigs were infected either with fresh H. pylori isolates from human gastric biopsies or with a guinea-pig passaged strain. When the animals were killed, 3 and 7 weeks after inoculation, samples were taken for culture, histopathology and serology. H. pylori was cultured from 22 of 29 challenged animals. All culture-positive animals exhibited a specific immune response against H. pylori antigens in Western blotting and gastritis in histopathological examination. Antibody titres in enzyme immunoassay were elevated among animals challenged with H. pylori. The inflammatory response was graded as severe in most animals and consisted of both polymorphonuclear leucocytes and lymphocytes. Erosion of the gastric epithelium was found in infected animals. These results suggest that the guinea-pig is suitable for studying H. pylori-associated diseases. Moreover, guinea-pigs are probably more similar to man than any other small laboratory animal as regards gastric anatomy and physiology.

Dietary factors influence the recovery rates of Helicobacter pylori in a BALB/cA mouse model.

The aim of this study was to assess the ability of different mouse diets to sustain an H. pylori infection in BALB/cA mice. Four commercially available mouse diets were compared. Experiment 1: Mice were fed the four diets for seven days before infection, infected three times at two-day intervals with 0.1 ml of 10(9) colony-forming units/ml H. pylori cells. H. pylori strains (n = 4) were cultured on GAB-Camp agar for 2 days, harvested and suspended in PBS. All animals were sacrificed at 2 and 4 weeks post inoculation. Experiment 2: Mice infected for 8 weeks were fed RM2, changed to the different diets for 10 days and sacrificed. Stomachs were collected, cultured on GAB-Camp agar to estimate H. pylori growth and stomach biopsies were analyzed by PCR. There were significant differences between diets in their ability to sustain growth of H. pylori. The range was from a few hundred colonies to no growth at all on the GAB-Camp agar. PCR signals showed good correlation with the culture results. All H. pylori-infected mice gave a significantly higher inflammation score compared to non-infected mice. The diet RM2, having the highest number of culturable H. pylori in the mouse stomach, also showed the highest inflammation. These results suggest that the dietary factors affect the amounts of H. pylori in an infection of BALB/cA mice.

Infection of BALB/c A mice by spiral and coccoid forms of Helicobacter pylori.

Helicobacter pylori exists in two different morphological forms, spiral and coccoid. This study demonstrated that both forms can infect BALB/c A mice. The animals were inoculated orally three times at 2-day intervals with 10(8) cfu of both spiral and coccoid forms of strain CCUG 17874 (NCTC 11637), strain 25 and strain 553/93. Infection was followed over a 30-week period by histological scoring of the grade of inflammation in gastric biopsies. At each time point sera were collected for analysis in ELISA and immunoblot analysis. Both spiral and coccoid forms of all H. pylori strains gave significantly higher inflammation scores than a control group of animals 1 week after inoculation. The histological evidence persisted throughout the entire 30 weeks. The inflammation was most severe in the pylorus and duodenum. Infection with strain 553/93 displayed the most severe gastritis. The spiral form of strain CCUG 17874 gave an immune response after only 4 weeks, whereas its coccoid form as well as strains 25 and 553/93 (spiral and coccoid forms) gave a significant increase in antibody response in ELISA and immunoblot after 16 weeks. It is concluded that both spiral and coccoid forms of H. pylori can cause acute gastritis in BALB/c A mice.

Sulfatides inhibit binding of Helicobacter pylori to the gastric cancer Kato III cell line.

Helicobacter pylori adhere to Kato III and Hela S3 cells in monolayer cultures. To explore whether cell surface glycoconjugates on these two cell lines mediate binding of H. pylori, various carbohydrates, glycoproteins, and glycolipids were tested to inhibit H.pylori cell adhesion. The adhesion was measured (i) with a urease-based assay and (ii) by cells stained with fluorescein. Sodium periodate and sialidase treatment (but not alpha- or beta-galactosidase, heparitinase,lysozyme, or trypsin) inhibited H. pylori binding to both cell lines. Sulfatides and sulfated glycoconjugates (50 microg/ml) but not heparin or a number of simple carbohydrates inhibited binding (1 mg/ml). The two H.pylori strains studied (CCUG 17874 and strain 25) showed high binding of soluble 125I-labeled heparin and other sulfated carbohydrate compounds.

Immunoblot assay for serodiagnosis of Helicobacter pylori infections.

An immunoblot assay for the serological diagnosis of Helicobacter pylori infection was evaluated. Serum samples from patients whose gastric biopsy specimens were known to be positive or negative for H. pylori on culture were used to establish interpretive criteria for the immunoblot assay. A panel of sera from patients with diseases other than H. pylori infection and sera from healthy blood donors were included to validate these criteria. All sera were initially assessed in an enzyme immunoassay (Ge-EIA), based on acid glycine-extracted cell surface proteins of H. pylori NCTC 11637. The same antigen extract was used in the immunoblot assay. In addition, the Ge-EIA and the immunoblot assay were compared with a commercially available EIA (Seradyn, Color Vue Pylori). Bands of 110/120 kDa and/or two of five low-molecular-mass proteins (26, 29, 30, 31, and 33 kDa, in any combination) showed a strong correlation with the H. pylori culture-positive patients (97.5%) compared to the correlation obtained with the EIA results (Ge-EIA, 87.5%; Seradyn EIA, 92.5%), and the antibody responses to these proteins were considered specific reactions. In 37 of 40 serum samples from culture-negative patients and also in sera from patients with other disorders, a moderate antibody reactivity to the medium-size proteins (43 to 66 kDa) was observed, and these were considered not valuable for a specific immunoblot assay. Among sera from culture-positive patients, 39 of 40 serum samples were defined to be immunoblot positive, and from among sera from culture-negative patients, 3 of 40 serum samples were defined to be immunoblot positive. The use of sera from patients with negative cultures for H. pylori as negative controls may decrease the sensitivity due to sampling error and false-negative culture results. Immunoblot assay-positive results were detected among 10% of sera from patients with other diseases, whereas they were detected among 42.5% of sera by the Ge-EIA and 47.5% of sera by the Seradyn-EIA. The higher number of EIA-positive sera in this group reflects a possible cross-reactivity (false-positive EIA result). Of the blood donors, representing asymptomatic but possibly colonized subjects, 24% were immunoblot positive. In conclusion, our data indicate that immunoblotting is more sensitive as well as more specific than EIA. Moreover, it permits detection of antibody responses to specific antigens, e.g., the cytotoxin-associated CagA protein, which may have pathological implications.

Binding of human plasminogen and lactoferrin by Helicobacter pylori coccoid forms.

The interactions between Helicobacter pylori spiral and coccoid forms, extracellular matrix (ECM) and plasma proteins were studied in an 125I-labelled protein assay. The range of binding of collagen V, plasminogen, human lactoferrin (HLf) and vitronectin to coccoid forms of H. pylori NCTC 11637 was 26-48%. In contrast, binding of radiolabelled fibronectin and collagen types I and III was low (3-8%). The coccoid forms of 14 strains of H. pylori showed significant HLf binding (median 26%). With plasminogen, no significant difference was found between binding to the coccoid (median = 13%) and spiral (median = 12%) forms, of 13 of the 14 strains of H. pylori tested; the exception was strain NCTC 11637. 125I-plasminogen showed a dose-dependent binding to both the coccoid and spiral forms. Plasminogen binding to both forms was specific; the binding was inhibited by non-labelled plasminogen, plasmin, lysine, EACA (epsilon-aminocaproic acid) but not by fetuin or various carbohydrates. Similarly, HLf binding was found to be specific and was inhibited by non-labelled HLf and BLf. The coccoid forms showed either similar or enhanced ECM binding capabilities compared with the spiral forms. As the binding of ECM proteins may be an important mechanism of tissue adhesion for various pathogenic bacteria, the coccoid differentiated form of H. pylori can be considered as an infective form in the pathogenesis of helicobacter infection and type B gastritis.

A collagen binding protein from Lactobacillus reuteri is part of an ABC transporter system?

The gene coding for a collagen binding protein from Lactobacillus reuteri NCIB 11951 was cloned and sequenced. A genomic lambda library was constructed and recombinant plaques were screened using antisera raised against purified collagen binding proteins from the same L. reuteri strain. The positive plaques were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blot analysis, which revealed the expression of a 29 kDa protein, which reacted with the antisera and bound 125I-labelled type I collagen. The sequence of the corresponding gene, cnb showed that the collagen binding protein has sequence similarities to the solute binding component of bacterial ABC transporters.

Gastrointestinal colonisation of BALB/cA mice by Helicobacter pylori monitored by heparin magnetic separation.

Immunocompetent and immunodeficient BALB/cA mice were fed orally with 10(8) colony forming units of 2-day-old spiral or coccoid (12 days old) Helicobacter pylori strain NCTC 11637. Immunocompetent BALB/cA mice were also fed orally with decreasing numbers of spiral or coccoid forms of H. pylori. The gastrointestinal colonisation process was monitored for 34 days post-infection by heparin magnetic separation and subsequent enzyme immunoassay (EIA) for the detection of the H. pylori cells. Both mice types were colonised with H. pylori. The coccoid form of H. pylori gave higher EIA absorbance values and more efficient colonisation in the mice than the spiral form. Immunocompetent BALB/cA mice fed with the coccoid form of H. pylori exhibited an acute inflammation process in histopathological samples from the stomachs. In conclusion, H. pylori can infect both immunocompetent as well as immunodeficient BALB/cA mice and coccoids (viable but non-culturable) obtained after 12 days of culturing can infect BALB/cA mice.

Purification of collagen-binding proteins of Lactobacillus reuteri NCIB 11951.

Collagen type-I-binding proteins of Lactobacillus reuteri NCIB 11951 were purified. The cell surface proteins were affinity purified on collagen Sepharose and eluted with an NaCl gradient. Two protein bands were eluted from the column (29 kDa and 31 kDa), and both bound radio-labeled collagen type I. Rabbit antisera raised against the 29 kDa and 31 kDa protein reacted with the affinity-purified proteins in a Western blot with whole-cell extract used as antigen. The N-terminal sequence of the 29-kDa and 31-kDa proteins demonstrated the closest homologies with internal sequences from an Escherichia coli trigger factor protein (TIG.ECOLI). Out of nine other lactobacilli, the antisera reacted only with the L. reuteri and not with the other species tested.

Cell surface hydrophobicity and adherence to extra-cellular matrix proteins in two collections of methicillin-resistant Staphylococcus aureus.

Non-specific and specific mechanisms of adherence have been examined in two collections of methicillin-resistant Staphylococcus aureus (MRSA). Determination of hydrophobicity by salt aggregation, hydrophobicity indices and of adherence to the extra-cellular matrix proteins fibronectin, vitronectin, laminin and collagen type 1 have failed to reveal any correlation with phage-type, plasmid profile or antibiogram. Further, the strain collections, made over a period of years in two countries, differ markedly in their adherence characteristics; MRSA are heterogeneous in this respect. Such heterogeneity may explain the polarization of views on the epidemicity or 'virulence' of MRSA. With the exception of adherence to collagen a small group of methicillin sensitive S. aureus had characteristics intermediate between the two groups of MRSA.

Adherence of haemagglutinating Helicobacter pylori to five cell lines.

Helicobacter pylori causes gastritis and is an important factor for the development of peptic ulcer disease in man. We used two different methods to examine the adhesion of nine H. pylori strains, with different haemagglutinating properties, to five cell lines, HeLa S3, HFI, Vero, SW1222 and WEHI cells. The adhesion studies were performed a) as bacterial adhesion to a monolayer of tissue culture cells, visualizing bacteria with fluorescein-isothiocyanate-labelled antibodies, b) as cell agglutination with bacteria and eucaryotic cells mixed in a suspension. The H. pylori strains were divided into three groups according to their cell adhesion properties. In general, H. pylori strains which showed the best adhesion to the five cell lines were strains which showed the best capability of agglutinating erythrocytes of several animal species. It is likely that the same adhesins are involved in cell adhesion and in haemagglutination. The two methods gave similar results.

Surface properties, connective tissue protein binding and Shiga-like toxin production of Escherichia coli isolated from patients with ulcerative colitis.

Escherichia coli strains isolated from intestinal biopsies of patients with ulcerative colitis (n 146), Crohn's disease and colonic polyposis (n 41) were analysed for binding of collagen I, collagen IV, fibronectin and laminin. Strains expressed varying degrees of binding of one or more of the four connective tissue proteins. Only 32 strains did not express binding of any of the proteins. The strains expressed low or moderate cell surface hydrophobicity. There was no correlation between protein binding and expression of cell surface hydrophobicity. E. coli isolated from inflamed rectal mucosa were slightly less negatively charged than strains isolated from healthy intestinal mucosa. Shiga-like toxins I and II were detected in 32 strains from 28 patients. Of these, 5 strains had been isolated from normal or healed tissue. In patients with inflammatory bowel disease, connective tissue proteins are exposed in intestinal ulcerations. Strains expressing binding of one or several of these proteins may have a selective advantage to colonize these lesions.

Type I and IV collagen and fibrinogen binding to Aeromonas species isolated from various infections.

Collagen binding is a common property of strains of Aeromonas species. However, agglutination of latex beads coated with types I and IV collagen and fibrinogen with Aeromonas cells varied among strains of Aeromonas species and their source of isolation. Culture media and growth conditions greatly influenced expression of Aeromonas cell surface receptors to bind collagen (types I and IV) and fibrinogen immobilized on the latex particles as suggested by the particle agglutination assay (PAA). Aeromonas cells aggregated with the differentially coated latex beads in a specified manner. Furthermore, the PAA method was found to be rapid, easy to perform and sensitive for routine screening of a large number of strains for serum and connective-tissue protein cell surface receptors.

Particle agglutination assays to identify fibronectin and collagen cell surface receptors and lectins in Aeromonas and Vibrio species.

A rapid particle agglutination assay (PAA) utilizing latex beads coated with connective tissue and serum proteins was evaluated for its ability to identify fibronectin, collagen (types I and IV), fibrinogen, and transferrin cell surface receptors on Vibrio and Aeromonas strains isolated from diseased fish, human infections, and the environment. Similar tests were performed to screen for cell surface lectins. Vibrio as well as Aeromonas strains were found to bind connective tissue proteins (collagen types I, II, and IV and fibronectin), serum proteins (i.e., fibrinogen), and glycoproteins (bovine submaxillary mucin, hog gastric mucin, orosomucoid, and fetuin) immobilized on the latex particles. The specificity of the agglutination reaction was studied by particle agglutination inhibition assays performed by preincubating bacterial suspensions in solutions containing either gelatin (for the various connective tissue protein PAA reagents) or sialic acid-rich glycoproteins (for the various glycoprotein PAA reagents). Expression of cell surface receptors for connective tissue proteins was found to depend on culture methods.

Crystal violet binding, cell surface properties and extracellular enzyme profiles of Staphylococcus aureus producing toxic shock syndrome toxin-1.

Staphylococcus aureus isolated from clinically diagnosed cases of toxic shock syndrome (TSS) showed susceptibility to phage types belonging to both I and III groups (90.5%). Phage typing patterns showed a wide diversity among 87 toxic shock syndrome toxin-1 (TSST-1) positive strains isolated from different non TSS clinical sources. Toxin producing strains isolated from both TSS and non TSS showed a remarkable ability to bind to crystal violet (pattern C/D, 97.2%) incorporated into brain heart infusion agar media at subinhibitory concentrations and these isolates were traced to biotype var. hominis. The cellular fatty acid compositions of TSS and non-TSS strains belonging to the three biotypes S. aureus var. hominis, S. aureus var. bovis and S. aureus var. canis did not differ. TSST-1 producing strains demonstrated a high salt aggregation test value (above 1.5) indicating a low cell surface hydrophobicity. Both TSS and non TSS strains demonstrated a high lipolytic activity. TSST-1 positive strains in general, showed significantly higher lipase activity than strains isolated from septicemia (p less than 0.0001) and superficial (p less than 0.0001) infections. The proteolytic activity is higher among TSS (median value 0.075 U/ml) than to non TSS (median value 0.045 U/ml) strains. There was no correlation with the quantity of toxin production in vitro and to the properties described.