Materials and methods

PATIENT SAMPLES

Serum samples from patients with various chronic liver diseases (n=144, age 31–88 years) were analysed. They represented a two year prospective series of patients investigated at the Department of Medicine, University Hospital of Malmö for suspected chronic liver disease (CLD) owing to increased concentrations of liver enzymes and/or clinical symptoms of more than six months duration.8Thirty patients with primary sclerosing cholangitis (PSC)9were also included. The diagnosis was verified in all cases by endoscopic retrograde cholecystography (ERC). Thirty per cent of the patients had alcoholic liver disease, 20% autoimmune liver disease (primary biliary cirrhosis, autoimmune hepatitis, and primary sclerosing cholangitis) and 10% cryptogenic liver disease. Other diagnostic groups were chronic viral hepatitis, metabolic liver disease, systemic disease, drug reactions, malignant disease, and heart failure.

Blood donor serum samples (n=48) randomly collected at the Blood Transfusion Unit at the University Hospital of Malmö, served as controls. All serum samples were kept frozen at −20°C until analysed.

BACTERIAL STRAINS AND CULTURE CONDITIONS

H hepaticus, a murine liver isolate (strain CCUG 33637) was cultured on Fastidious agar (Amersham Pharmacia Biotech, Rainham, UK), and grown for 7–10 days in a microaerophilic atmosphere (3% H₂, 10% CO₂, 5% O₂, 82% N₂) at 37°C. Agar plates were examined for colonies with a green metallic transparent surface, and tested positive for urease, catalase, and oxidase enzyme activities.

PREPARATION OF CELL SURFACE ANTIGENS FOR ENZYME IMMUNOASSAY AND IMMUNOBLOT

Ten agar plates (Φ5 cm) with confluent bacterial growth were harvested and washed once in 10 mM phosphate buffered saline (PBS) pH 7.2, with a total cell yield of 0.6 g wet weight. Cell surface antigens were extracted using a standard method with 0.2 M acid glycine buffer pH 2.2 as described previously.10 The protein concentration was 290 μg per ml as determined by the BioRad protein assay (BioRad, Richmond, UK). The protein profile ofH hepaticus was compared with acid glycine extracted cell surface proteins of H pylori(CCUG 17874) by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS- PAGE)11 followed by silver staining.

H HEPATICUS ENZYME IMMUNOASSAY (IgG-EIA)

The H hepaticus enzyme immunoassay (EIA) was performed as described for H pylori EIA.10 12 In brief, Maxisorp immunoplates (NUNC, Roskilde, Denmark) were coated with 100 μl per well with a protein concentration of 5 μg per ml; serum diluted 1/800 was added and the plates were incubated for 90 minutes at 37°C. An alkaline phosphatase goat antihuman antibody (Sigma, St Louis, Missouri, USA) was used as the second antibody. After repeated washings, 100 μl per well of a substrate solution containingp-nitrophenyl phosphate (Sigma), diethanolamine, and MgCl was added. The EIA test results are expressed as corrected mean absorbance values (A 405 nm) in per cent of a reference standard (human gammaglobulin, Pharmacia & Upjohn, Stockholm, Sweden). A relative antibody activity (RAA value) greater than 35 and less than 25 units was defined as positive and negative, respectively; values in between were regarded as low positive values (borderlines). The cut off values are based on EIA and immunoblot analyses of serum samples from patients with positive and negative gastric culture forH pylori, healthy blood donors, and children.10 12 Similar IgG EIA analyses were performed with acid glycine extracted cell surface proteins ofH pylori.

IMMUNOBLOT (IgG)

In the immunoblot assay, acid glycine extracted cell surface proteins of H hepaticus were separated by SDS-PAGE with an 8–18% gradient (ExelGel, Pharmacia Biotech AB, Uppsala, Sweden) according to the Pharmacia manual. Proteins were electroblotted onto a PVDF membrane (Millipore, Intertech, Bedford, Massachusetts, USA), and subsequent blocking and probing with patient and control serum samples were performed as described previously.13 Similar immunoblot analyses were performed with acid glycine extracted cell surface proteins ofH pylori.

IgA IMMUNOBLOT

As control for non-specific bacterial binding between proteins ofH hepaticus/H pylori and IgA, an immunoblot was performed. Strips, containing separated cell surface proteins of the two helicobacter species were overlaid with human IgA antibodies (myeloma produced) and then incubated with an enzyme conjugated anti-IgA antibody followed by substrate staining. Another two strips were probed with serum samples from H hepaticus/H pylori EIA positive patients. Enzyme conjugated antihuman IgA antibodies (P216, Dako A/S, Glostrup, Denmark) were added and after incubation the strips were developed.13

IMMUNOSTAINING OF HEAT SHOCK PROTEINS (Hsp60)

Immunoblot strips containing separated H hepaticus and H pylori cell surface proteins respectively were probed with a polyclonal antiserum to theH pylori Hsp60 protein. The Hsp60 protein was purified from H pylori strain CCUG 17874 by gel filtration (Superose 12, Pharmacia Biotech, Uppsala, Sweden) and ion exchange chromatography (MonoQ, Pharmacia Biotech), and then used for immunisation.14 Corresponding strips were incubated with an antihuman Hsp60 serum (Santa Cruz, Biotechnology Inc., USA). Immunoblot analyses were performed as described above.

ABSORPTION EXPERIMENTS

For absorption of potential cross reacting antibodies, sonicated whole cells of H pylori (CCUG 17874) were used. To 1 ml of sonicated cells (A 540 of OD 1.0) 10 μl serum was added and incubated for one hour at 22°C and for 16 hours at 6°C during constant shaking. Cells were removed by centrifugation and supernatants collected for Helicobacterserology.

Results

Protein profiles of acid glycine extracted cell surface proteins from H hepaticus were compared with a corresponding antigen preparation of H pylori after silver staining of the SDS-PAGE gel. Differences of the band pattern were observed with proteins migrating in the region of 43–94 kDa in molecular mass (M _r), whereas proteins with lower M _r showed a similar pattern (fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Protein profiles of the acid glycine extracted cell surface proteins of Helicobacter pylori (lanes 1 and 2) and Helicobacter hepaticus (lanes 3 and 4) separated by SDS-PAGE (8–18%) and stained with silver.

IgG-EIA

Antibody reactivity to extracted H hepaticus cell surface proteins in serum samples from patients with chronic liver disease was initially analysed by EIA. Of 144 serum samples, 37 (26%) showed increased antibody concentrations (RAA greater than 35 units), 19 (13%) showed moderate values (RAA 25–35 units), and 88 (61%) were negative (table 1). The distribution of antibody reactivity in the diagnostic subgroups amongH hepaticus EIA positive patients, including primary biliary cirrhosis and other autoimmune liver diseases and cryptogenic liver disease, did not differ from the whole patient population. In the H pylori EIA, 64 of 144 (44%) serum samples were positive and 15 serum samples gave borderline values (table 1). Of the 30 serum samples from PSC patients, six (20%) showed increased antibody concentrations, and 24 (77%) were negative in the H hepaticus EIA. In theH pylori EIA four serum samples were positive (table 1).

Serum samples from 48 healthy blood donors were assayed in the two EIAs and the seroprevalence was found to be 46% for H hepaticus and 48% for H pylori(table 1).

ABSORPTION EXPERIMENTS

Cross reactivity between cell surface proteins of the two helicobacter species was analysed with 37 serum samples from patients with chronic liver diseases showing high antibody concentrations in theH hepaticus EIA (14 of these wereH pylori EIA negative) and with 24 blood donor serum samples found positive by H hepaticus EIA (three of these were H pylori EIA negative). In 25 patients the H hepaticus EIA became negative but 12 serum samples remained positive. These serum samples were from patients who suffered from alcoholic related liver disease. All blood donor serum samples became negative after the absorption step except three that showed low values close to the lower cut off level (table 2).

IMMUNOBLOT (IgG)

Patient serum samples absorbed with H pylori whole cells and with a remaining positive value in theH hepaticus EIA were further analysed by immunoblot, and the antibody reactivity was compared with results with the same serum samples that were not absorbed. In theH hepaticus immunoblot the unabsorbed serum samples intensively stained proteins of 55–65 kDa inM _r (fig 2A). This was not observed in the immunoblot based on H pyloriantigens (fig 2B). When these serum samples, absorbed with sonicatedH pylori cells, were analysed in theH hepaticus immunoblot, reactivity to the 55–65 kDa proteins remained (fig 2C). However, when the serum samples were probed on strips containing separated H pylori proteins no reactivity was observed (figure2D).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Antibody response to cell surface proteins of Helicobacter hepaticus (A and C) and Helicobacter pylori (B and D) was analysed by immunoblotting with serum samples from patients with chronic liver disease (lanes 1 to 11). Lane 12 represents the reference standard. In (C) and (D) serum samples were absorbed with sonicated helicobacter cells.

IMMUNOBLOT (IgA)

Patients with chronic alcoholic liver disease often show high serum IgA levels compared with healthy individuals.15 The immunoblots performed for control of a non-specific binding of bacterial cell material from H hepaticus andH pylori showed no visible IgA reactivity (data not shown).

IMMUNOSTAINING OF Hsp60

By immunoblot the polyclonal anti-H pylori Hsp60 serum stained two major bands with an approximateM _r of 62–65 kDa with the two helicobacter species. The antihuman Hsp60 serum weakly stained a 65 kDa protein of H pylori, and withH hepaticus no bands were observed (data not shown).

Discussion

Antibodies against H hepaticus andH pylori were detected in serum samples from patients with various liver diseases. The frequency of antibody reactivity was, however, not increased in any specific diagnostic groups and serum positivity was not increased in patients with chronic liver diseases compared with the healthy individuals. Some of these antibodies may represent cross reactivity between the two species, while a specific antibody response to cell surface proteins ofH hepaticus was detected in serum samples from patients with alcoholic related liver disease.

Several new species of the genus helicobacter have recently been identified in various mammals, with some species inducing gastrointestinal infection while others seem to colonise the gastrointestinal tract without causing infection and tissue inflammation.16-18 Recent reports on multifocal hepatitis in dogs19 induced by bile tolerant helicobacter, suggested that helicobacter organisms with liver tropism may also cause hepatitis in humans and other primates. However, up to now no human liver bile duct isolates of helicobacter have been reported although bile samples were shown to contain H pyloriDNA.5 In susceptible strains of mice,H hepaticus causes an acute focal hepatitis, which progresses to chronic hepatitis with the presence of relatively few infectious organisms.20 Franklin and colleagues21 reported on a novel helicobacter species,H cholecystus, isolated from the bile tract of Syrian hamsters with chronic gall bladder inflammation and pancreatitis. Interestingly, Siringo and colleagues22reported a high prevalence of H pyloriinfection in cirrhotic patients, which was not confirmed by Tsai.23 Thus, it is possible that not yet isolated urease positive as well as urease negative helicobacter species may be able to colonise the human intestinal tract, as recently described in dogs and cats.24 Fox and colleagues6 concluded that helicobacter might induce a chronic gall bladder inflammation in humans with late sequelae such as bile tract and gall bladder malignancies, and a potential role of H hepaticus, H bilis, and other bile tolerant helicobacter species in chronic cholecystitis and liver disease. The importance of further studies on the possible role of these new helicobacter species in human liver disease was emphasised as was the need to develop non-invasive serological assays to determine the seroprevalence of hepatic helicobacter infections.

It is assumed that bile tolerant helicobacter species are antigenically and genetically similar. Our findings in this study, showing 12 patient serum samples with high antibody reactivity to cell surface proteins ofH hepaticus even after absorption withH pylori, indicate that some helicobacter-like organism may be involved in human liver diseases. By immunoblotting, antibodies mainly reacted to the 55–65 kDa proteins and at least two of these proteins were recognised by an anti-H pylori Hsp60 antibody, suggesting that patient serum samples intensively staining the 60–65 kDa proteins of H hepaticus contained increased concentrations of antibodies to helicobacter Hsp60 proteins (fig 2A). Hsps are highly conserved, immunogenic molecules induced by heat, inflammatory mediators, and other forms of physiological stress25; ethanol might be such a factor. The pathogenic mechanism of this finding is unknown, but it is a fact that ethanol abuse induces a local immune defect and activation of a latent helicobacter infection. Ward and colleagues26 found that antibodies to helicobacter (Hsp70) immunoreacted with hepatocytes from infected mice livers and thus could be responsible for autoimmune mediated hepatic lesions of bile duct epithelium infected withH hepaticus.

In conclusion, antibodies to H hepaticus, often cross reacting with H pylori, were detected in serum samples from patients with alcoholic related liver disease. The antibody response to H hepaticus proteins withM _r of 60–65 kDa remained even after absorption experiments. Attempts to culture strains from human liver biopsy specimens and bile samples will be an attractive approach, together with improved serological assays, to explore further the possible role of bile tolerant helicobacters in various biliary tract and liver diseases.