|Home | About | Journals | Submit | Contact Us | Français|
We recently described helicobacter-associated progressive, proliferative, and dysplastic typhlocolitis in aging (18- to 24-month-old) Syrian hamsters. Other pathogens associated with typhlocolitis in hamsters, Clostridium difficile, Lawsonia intracellularis, and Giardia spp., were not indentified. The presence of Helicobacter genus-specific DNA was noted by PCR in cecal and paraffin-embedded liver samples from aged hamsters by the use of Helicobacter-specific PCR primers. By 16S rRNA analysis, the Helicobacter sp. isolated from the liver tissue was identical to the cecal isolates from hamsters. The six hamster 16S rRNA sequences form a genotypic cluster most closely related to Helicobacter sp. Flexispira taxon 8, part of the Helicobacter bilis/H. cinaedi group. Livers from aged helicobacter-infected hamsters showed various stages of predominantly portocentric and, to a lesser extent, perivenular fibrosis. Within nodules, there was cellular atypia consistent with nodular dysplasia. The livers also exhibited a range of chronic active portal/interface and lobular inflammation, with significant portal hepatitis being present. The inflammation was composed of a mixture of lymphocytes, neutrophils, and macrophages, indicative of its chronic-active nature in these aged hamsters infected with Helicobacter spp. The isolation of novel Helicobacter spp., their identification by PCR from the diseased livers of aged hamsters, and their taxonomic classification as belonging to the Helicobacter bilis cluster strengthen the argument that H. bilis and closely related Helicobacter spp. play an etiological role in hepatobiliary disease in both animals and humans.
We recently described progressive, proliferative, and dysplastic typhlocolitis in aging (18- to 24-month-old) Syrian hamsters (31). The lesions in these hamsters were more severe in the older animals aged 7 to 24 months than in the younger, 1- to 6-month-old hamsters. The presence of Helicobacter spp. in the large bowel of all 24 hamsters included in the study was confirmed by culture and Helicobacter-specific PCR (31). Other pathogens associated with typhlocolitis in hamsters, Clostridium difficile, Lawsonia intracellularis, and Giardia spp., were not indentified in this study (31). Enterohepatic Helicobacter spp. are associated with the development of inflammatory bowel disease in mice and, as determined more recently, humans (2, 18, 27, 51). These helicobacters can also induce hepatitis and hepatocellular carcinoma in susceptible strains of mice (1, 14, 19). We undertook a study to examine in detail the histological profile of the livers of hamsters aged 18 to 24 months and to determine whether Helicobacter sp. DNA was present in their livers (31). In addition, we purchased a small number of 6-month-old hamsters from the same vendor which originally supplied the aged hamsters and examined them for the presence of Helicobacter spp. in their intestines and liver and whether inflammation was associated with the presence of Helicobacter spp. in these target tissues. We document that, with the exception of fibrosis, which does not develop in mice with any frequency, all of the hepatobioliary lesions noted in unmanipulated colony-maintained hamsters described in our report are consistent with the hepatic lesions noted in male A/JCr, ABF1 mice or B6C3F1 mice infected with Helicobacter hepaticus (14, 19, 20, 39). In general, the localization of hepatic fibrosis in these aged hamsters which had Helicobacter spp. in their livers correlated with the portocentric inflammation typically observed in helicobacter-infected animal models. We also classify the novel helicobacter isolates from the ceca and liver of hamsters and additional isolates from mice and gerbils as belonging to a novel genotype in the H. bilis cluster.
Twelve hamsters aged 18 to 24 months (group A), described previously (31), and an additional five hamsters aged 6 months (group B) were euthanatized by CO2 and underwent a complete postmortem examination.
Tissue samples that had been collected from the cecum (five hamsters in group A, five hamsters in group B) and liver (five hamsters in group B) were placed in culture medium containing 20% (vol/vol) glycerol with brucella broth and were frozen at −70°C. These samples were subsequently homogenized with phosphate-buffered saline for microaerobic culture. The media used for culture included Trypticase soy agar with 5% sheep blood (BAP); medium impregnated with trimethoprim, vancomycin, and polymyxin (TVP; Remel); and medium impregnated with cefoperazone, vancomycin, and amphotericin B (CVA; Remel). A portion of each homogenized sample was directly applied to TVP and CVA, filtered through a 0.45-mm-pore-size filter, and plated on BAP. The plates were incubated at 37°C under microaerobic conditions for 2 to 4 weeks in vented jars containing N2, H2, and CO2 (80:10:10). Bacterial isolates with characteristic growth on agar and microscopic morphology were subsequently used for biochemical analysis and as sources of genomic DNA.
Bacteria isolated from the ceca or liver were cultured on blood agar plates, and the cells were harvested and washed once with 1 ml of phosphate-buffered saline. A High Pure PCR template preparation kit (Roche Molecular Biochemicals, Indianapolis, IN) was used for DNA extraction.
The sequences of the 16S rRNA genes of six hamster isolates of Helicobacter spp. were determined as described previously (7). Briefly, primers F24 (positions 9 to 27 in the forward direction) and F25 (positions 1525 to 1541 in the reverse direction) were used to amplify the 16S rRNA genes. The PCR product was concentrated and purified with QIAquick PCR purification kits (Qiagen, Valencia, CA). Purified DNA was sequenced with an ABI Prism cycle sequencing kit (BigDye Terminator cycle sequencing kit) on an ABI 3100 genetic analyzer (Applied Biosystems, Foster City, CA). The sequencing primers (7) were used in quarter-dye reactions, according to the manufacturer's instructions. The 16S rRNA gene sequences were entered into RNA, a program for analysis of 16S rRNA gene data, and were aligned as described previously (26, 33). Similarity matrices were constructed from the aligned sequences by using only those base positions for which data were available for 90% of the strains and were corrected for multiple base changes by the method of Jukes and Cantor (26). Phylogenetic trees were constructed by the neighbor-joining method (40).
Paraffin-embedded hamster liver tissues were sectioned at 5 μm. The sections were dewaxed with two xylene washes (10 min at 55°C each), two 100% ethanol washes, two 70% ethanol washes, and two distilled water washes (5 min at 55°C for each step). The samples were then rehydrated with 1× Tris-EDTA for 5 min at 55°C and Tris (1 M, pH 7.5) overnight at 55°C. After rehydration, the samples were digested with proteinase K (20 mg/ml) at 55°C for 3 h. The DNA was then extracted with a QIAamp DNA minikit (Qiagen).
DNA was extracted from the cecal tissue of all hamsters in group A and group B and the livers of hamsters in group B by using a High Pure PCR template penetration kit (Roche Prognostics, Penzberg, Germany). PCR amplifications were performed with a thermal cycler and an Expand high-fidelity PCR system (Roche Molecular Biochemicals). Each reaction mixture (50 μl) contained 1× Expand high-fidelity buffer, Expand high-fidelity enzyme mixture (2 U), a 0.5 μM concentration of primers for the Helicobacter genus (primers C97 and C98), a 200 μM concentration of each deoxyribonucleotide triphosphate, and bovine serum albumin (200 μg/ml), as described previously (13). Amplification was achieved by denaturation at 94°C for 1 min, annealing at 58°C for 1.5 min, and elongation at 72°C for 2 min with 35 amplification cycles. Five microliters of the PCR product was used as the template for the second run of the PCR with the same set of primers. A 15-μl portion of the sample was then electrophoresed through a 1% agarose gel, followed by ethidium bromide staining and UV illumination.
To assist with confirmation of the species of the Helicobacter spp. isolated, restriction fragment length polymorphism (RFLP) analyses were performed as described previously (41). Helicobacter sp.-specific primers C97 and C05 were used to generate 16S rRNA amplicons of 1.2 kb. The Helicobacter species-specific, 1.2-kb PCR products were subjected to RFLP analysis. Briefly, 27 μl of the PCR products was digested with 10 μl each of the restriction endonucleases AluI and HhaI and 3 μl of restriction buffer (New England Biolabs, Beverly, MA) at 37°C overnight. The samples were then electrophoresed through a 6% Visigel separation matrix (Stratagene, Cedar Creek, TX), stained with ethidium bromide, and visualized under UV illumination.
Tissues from the liver and lower gastrointestinal tract were placed in 10% buffered formalin. After routine paraffin embedding of the fixed tissue, sections (5 μm) were cut and stained with hematoxylin-eosin. Sections of the livers from 12 hamsters in group A were stained with Masson's trichrome for evaluation of the degree of hepatic fibrosis on a scale of from 0 to 6 by use of the criteria established by Ishak et al. for assessment of chronic viral hepatitis in humans (24). Additionally, the degrees of associated portal/interface hepatitis and lobular hepatitis were scored on a scale of from 0 to 4.
Helicobacter species were isolated from the ceca of five hamsters in group A and from the ceca of five hamsters and the liver of one hamster in group B. Culture of the novel helicobacter from the liver of only one hamster (but from the ceca of all hamsters) early during the course of the infection is not surprising and is similar to the results of our longitudinal studies with mice infected with H. hepaticus or H. bilis (14, 16). The bacteria grew at 37°C under microaerobic conditions, formed discrete colonies, and were gram negative and motile. The bacteria were catalase and oxidase positive and urease negative. The bacteria did not reduce nitrate to nitrite, did not hydrolyze alkaline phosphatase or indoxyl acetate, and did not have gamma-glutamyltranspeptidase activity.
By 16S rRNA analysis, the Helicobacter sp. isolated from the liver tissue of one hamster in group B was identical to the cecal isolates from the hamsters in group A (Fig. (Fig.1).1). By 16S rRNA sequence analysis, the sequences of the six isolates from hamsters formed a genotypic cluster that was the most closely related to Helicobacter sp. Flexispira taxon 8 and part of the H. bilis/H. cinaedi group. The hamster isolates contained an intervening sequence (IVS) in their 16S rRNA that is essentially identical to that in H. bilis Flexispira taxa 2, 3, and 8, which are part of the H. bilis/H. cinaedi species group. The MIT 96-1001 cluster of helicobacter mouse strains first described by Shomer et al. (42) also have an IVS that is essentially identical to that in H. bilis and Flexispira taxon 8. Other mouse and hamster isolates (isolates hamster B and UNSW1.7sp) also belong to taxon 8 and possess the H. bilis-like IVS. In the cluster marked with an asterisk in Fig. Fig.1,1, the majority of helicobacters have the flexispira morphology (5). However, the flexispira morphology and the H. bilis-like IVS do not fully correlate. Helicobacter cinaedi and H. canis strains do not have either the flexispira morphology or the H. bilis-like IVS. The helicobacter cluster of Shomer et al. has a simple curved helicobacter morphology in combination with the H. bilis-like IVS (42). Isolates hamster B and UNSW1.7sp (as well as some other strains in GenBank) have the H. bilis-like IVS but have an unreported or unknown morphology.
As noted previously, the presence of Helicobacter genus-specific DNA was noted by PCR in all cecal samples from hamsters in group A surveyed (31) and were similarly identified in the cecum of all hamsters in group B. All except one (Fig. (Fig.2,2, lane 8) of the paraffin-embedded liver samples from hamsters in group A were PCR positive by analysis with Helicobacter-specific primers (Fig. (Fig.2).2). One of five liver samples from hamsters in group B was PCR positive (the same hamster whose liver was positive by culture) when Helicobacter genus-specific primers were used.
By RFLP analysis of the PCR products, the presence of the novel Helicobacter sp. was confirmed in cecal isolates and the bacterial isolate from the liver. The RFLP patterns noted in Fig. Fig.3,3, lanes 1 to 4 (by use of restriction enzyme HhaI) and lanes 5 to 8 (by use of restriction enzyme AluI), were consistent with those for the previously described novel hamster Helicobacter sp. (31).
A group of five retired hamster breeders (group B) were evaluated for histopathological changes in the liver and ileo-ceco-colic junctions. In general, there was minimal to mild hepatic inflammation involving the portal tracts and lobular parenchyma in the livers from all hamsters in group B. The inflammation was predominantly composed of lymphocytes, admixed with a few neutrophils and macrophages. The inflammatory foci were random and sporadic, and not all lobes were consistently affected. The hepatitis index scores ranged from 1.5 to 3, indicative of the low-grade nature of the inflammatory process. In at least three of the animals, discrete moderately sized portal inflammatory aggregates were consistent with a diagnosis of portal hepatitis. Few hematopoietic precursors were also scattered within the lobular parenchyma in selected livers. Three of the livers exhibited moderate centrilobular to diffuse hepatocellular swelling and cytoplasmic clearing, most likely indicative of glycogen accumulation. The diffuse nature of the glycogen distribution in the hepatic parenchyma, in most cases, is nonspecific and a consequence of altered physiological states or conditions in which there is excess glycogen synthesis or decreased/impaired enzymatic glycogen catabolism. Additionally, on histological examination, the ileo-ceco-colic junctions of the animals revealed mild mucosal inflammation comprising lymphocytes and a few granulocytes in at least three of the five hamsters from this group examined. Interestingly, these three animals also had hepatitis index scores of 3. In one of the ileo-ceco-colic junctions, mild gut-associated lymphoid tissue reactive hyperplasia was present in association with mucosal inflammation.
Histomorphological evaluation of 12 trichrome-stained livers from aged infected hamsters showed various stages of predominantly portocentric and, to a lesser extent, perivenular fibrosis, with scores ranging from 0.5 to 5 (median score, 3.0; mean score, 2.67; standard deviation, 1.24) (Fig. (Fig.4).4). Similar patterns of fibrosis are also noted in viral hepatitis of humans, and hence, the fibrosis scoring criteria developed by Ishak et al. (24) were utilized in this study. Furthermore, the portal fibrosis (zone 1) observed in these aged hamsters contrasts with the predominant centrivenular fibrosis (zone 3) that is usually manifested in experimental animal models of toxic hepatic hepatitis, especially in mice and rats (52). The intensity of collagen staining varied from pale blue to dark blue, depending upon the thickness of the fibrous connective tissue and the associated inflammation. In a few livers, mild to moderate, multifocal portal-portal and/or portal-central bridging fibrosis, fibrous septation and/or nodularity with separation into small pseudolobules, or a nodule-in-nodule appearance was also evident (Fig. 5b to d). Within these nodules, there was architectural distortion; hepatic cords two to three cells thick; sinusoidal distention; as well as cellular atypia, consistent with nodular dysplasia (Fig. 6a and b). The livers also exhibited a range of chronic active portal/interface and lobular inflammation, with the scores ranging from 0.5 to 6. Multiple foci of mild to moderate hepatocellular necrosis were present in some of the livers (Fig. (Fig.7).7). Significant portal hepatitis (scores, ≥2) was also present in 6 of 12 livers selected for trichrome staining. The inflammation was composed of a mixture of lymphocytes, neutrophils, and macrophages, indicative of its chronic active nature in these aged hamsters infected with Helicobacter spp. (Fig. (Fig.7).7). The correlation between hepatic fibrosis and associated inflammation is depicted in Fig. Fig.4.4. In some of the livers with marked fibrous thickening of the portal/periportal areas and perivenular regions (Fig. (Fig.5c),5c), the degree of inflammation was milder, whereas in other livers with significant portal inflammation, the degree of fibrosis was of lesser intensity (Fig. (Fig.5b).5b). One of the 12 infected animals showed sparse hepatic inflammation and a minimal increase in portal fibrous connective tissue (Fig. (Fig.5a).5a). These differences in the extent of fibrosis may be due different phases of inflammation expected with an ongoing chronic active inflammation, postrecovery scarring, or a combination of both. In many livers, distinct portal and lobular aggregates of yellow-pigmented macrophages and few multinucleated cells were also present (Fig. (Fig.7d).7d). Other changes included various degrees of incomplete biliary hyperplasia and/or oval cell proliferation (Fig. 7e and f).
A number of hamster Helicobacter spp. have previously been isolated and characterized (50). Helicobacter cinaedi and H. mesocricetorum have been isolated from clinically normal hamsters, H. aurati has been isolated from hamsters with gastritis, and H. cholecystus has been isolated from the gallbladders of young hamsters with cholangiofibrosis and centrilobular pancreatitis (17). The Helicobacter spp. isolated from the ceca and livers of hamsters described in this report cluster with Helicobacter Flexispira taxon 8, as determined by l6S rRNA analysis (5, 31). Hänninen et al. proposed the inclusion of Helicobacter spp. of Flexispira taxa 2, 3, and 8 largely on the basis of the ureB gene phylogeny in the H. bilis taxon (22). All of these taxa also have a fusiform morphology. By 23S rRNA analysis, these taxa are closely related, but a monophyletic cluster that includes these taxa also includes H. cinaedi, a species with a curved helicobacter morphology (6). Hänninen et al. did not include H. cinaedi in their phylogeny, as it is urease negative and presumably lacks the ureB gene (22). Thus, we would argue that the phylogeny and the taxonomy of this cluster of organisms are still not fully resolved and require further investigations, such as sequencing of the 23S rRNA genes of all those species in the group marked with an asterisk in Fig. Fig.1,1, sequencing of multiple informative genes such as that for HSP60, or comparison of the genome sequences among these species.
Flexispira taxa 2, 3, and 8, now classified by Hänninen et al. (22) as H. bilis, have been isolated from the feces of dogs, mice, sheep, and humans and the stomachs of pigs (5). The H. bilis taxa have thus been linked to a wide spectra of diseased hosts and have also been associated with zoonotic diseases (12). H. bilis was first isolated from the livers of aged inbred strains of mice with hepatitis (16) and subsequently from the diseased livers of outbred mice (15). H. bilis has also been used experimentally to induce colitis and hepatitis in a variety of immunocompromised mice (18, 27), as well as cholecystitis in C57L mice fed a lithogenic diet (29). The novel Helicobacter spp. from the hamsters cluster most closely by 16S rRNA analysis to Helicobacter spp. of Flexispira taxon 8 (5). Interestingly, when the novel mouse isolates first described by Shomer et al. as belonging to taxon 8 were used to experimentally infect A/J and Tac:ICR:HascidRf mice, they caused typhlocolitis and cholangohepatitis (42). Furthermore, bacteria with the H. bilis/Flexispira morphology have been identified in the intrahepatic bile ducts of rats experimentally infected with Fasciola hepatica liver flukes (11). When H. bilis isolates recovered from nude rats with inflammatory bowel disease were orally inoculated into nude rats, they induced typhlocolitis (21).
H. bilis DNA has also been isolated from gallbladder tissue of humans with cholecystitis who were at a high risk of developing gallbladder cancer (13). More recently, H. bilis was associated with biliary cancers in two high-risk populations in Japan and Thailand (28). As determined by enzyme-linked immunosorbent assay-based serological assays, Thais, some with liver fluke infections, also had serologic evidence of H. bilis infection (35). These organisms have also been cultured from the stool of diarrheic and bacteremic humans (45, 49).
Hamsters have been used extensively in the study of carcinogen-induced hepatic malignancy as well as models to study the role of hepatobiliary flukes that induce bile duct proliferation and cholangiocarcinoma. During the course of these and related studies, authors have noted the presence of necrotic foci, cholangiohepatitis, periportal lymphocytic infiltration, cirrhosis, bile duct proliferation, and cholangiocarcinoma as incidental findings in control hamsters or hamsters undergoing other experimental procedures (3). In one study conducted in England, cirrhosis, bile duct proliferation, and cholangiocarcinoma were noted in aged control hamsters over 6 months of age. The incidence of cholangiocarcinoma in hamsters analyzed from 1958 to 1963 was 33 (9 in males and 24 in females) among a total of 5,604 hamsters; the authors noted that Hymenolepsis sp. tapeworms were endemic in these hamsters, and the granulomas were ostensibly caused by the parasite seen in the liver. They did not attribute the cirrhosis, bile duct proliferation, or cholangiocarcinoma to parasites, stating that the parasites were more often noted in noncirrhotic livers. Aflatoxin contamination was also ruled out, given that the diet used was tested and not found to be toxic in the standard test (in vivo duckling toxicity assay) employed at the time. They did, however, note that outbreaks of clinical enteritis were a common occurrence and subclinical infections were suspected in almost all of the animals (3). In a study of a Hannover colony of Syrian hamsters, liver tumors were composed of hemangioendotheliomas (5.6% of all tumors), hepatocellular adenomas (16.7% of all tumors) cholangiocarcinomas (16.7% of all tumors), and cholangiomas (6.1% of all tumors) (37, 38). That colony also had a high incidence of intestinal adenocarcinoma with lymphogenic metastasis to the lymphatic and mesenteric lymph nodes. Interestingly, chronic colitis with hyperplasia was noted in 50% of the hamsters. This finding is also reminiscent of the findings of the current study, in which all of the aged hamsters (18 to 24 months old) with hepatobiliary disease also had moderate to severe typhocolitis (31). These lesions may have been a manifestation, in part, of chronic helicobacter infection, along with wet tail syndrome or transmissible intestinal hyperplasia (caused by Lawsonia intracellularis) commonly diagnosed in hamsters during that era (4, 25). Also, 16% of the hamsters with intestinal tumors had tumors of bile duct origin. In a subsequent paper, the authors evaluated three strains of hamsters (strains CW, WH, and AH) from the Eppley Institute maintained under the same conditions. A direct relationship between age and the time of appearance of multiple tumors was evident (36). Polyps of the ceca were detected in all three groups of hamsters and were invariably associated with chronic proliferative colitis. One percent, 8%, and 8% of all tumors noted in the three groups (the strain CW, WH, and AH hamsters, respectively) were digestive tumors. Bile duct proliferation, related to cholangiomas in all three groups, was diagnosed as well. Hemangioendotheliomas of the liver were noted in female WH and AH hamsters. Hepatocellular adenomas were noted in a small percentage of WH and AH hamsters. Diseases such as colitis and cholangitis had similar incidences among the groups analyzed (36).
A review of the literature ascribing the carcinogenic potential of liver flukes and the development of cholangiocarcinoma in humans was based in part on data obtained from fluke-infected hamsters (23). In several of the hamster studies, although fluke infection per se did not induce liver cancer, the combination of liver fluke infection and exposure to various carcinogenic nitrosamines induced cholangiofibrosis, cholangiocarcinomas, and hepatocellular nodules (9, 10, 46-48). The high prevalence of Helicobacter spp. in hamsters and the associated prevalence of hepatobilary disease also raise the question of whether Helicobacter spp. could have been acting as tumor promoters in the livers of hamsters infected with the liver flukes Opisthorchis viverrini and Clonorchis sinensis. Given the persistent infection of the hamsters with the novel Helicobacter sp. (being present in the ceca and liver of 6-month-old and 18- to 24-month-old hamsters from the same vendor) described in this report allows us to argue that the hamsters previously used in the liver fluke studies had a background of helicobacter-associated hepatobiliary disease. The chronic hepatic lesions ascribed to liver fluke infection in both hamsters and humans, i.e., bile duct hyperplasia and metaplasia, thickening of bile duct walls, and suppurative cholangitis, could indeed be due, in part, to helicobacter-associated hepatobiliary disease. These findings are similar to the enhanced liver tumorgenicity of H. hepaticus infection in B6C3F1 mice that were treated with the carcinogenic nitrosamine N-nitrosodimethylamine. In that tumor promotion model, H. hepaticus not only stimulated the growth of liver tumors from initiated hepatocytes but also enhanced the progression to malignancy (8). Interestingly, exposure to N-nitroso compounds in human populations at high risk of liver cancer in Thailand was also suggested to play a role in the development of cholangiocarcinoma (30). The findings of that epidemiologic study are supported by the observation that patients infected with liver flukes have increased urinary secretion of nitrate and N-nitrosoproline, and the authors argued that the combination of fluke infection plus nitrosamine exposure increased the risk of development of cholangiocarcinoma in the Thai population (43, 44). These data are also consistent with the observation that hamsters infected with O. viverrini had increased levels of nitrosamine and nitrate biosynthesis mediated by nitric oxide synthase (32). Repeated infection with O. viverrini also induced inflammation-associated DNA damage in the bile ducts of hamsters via inducible nitric oxide synthase (24, 34).
The isolation of a novel Helicobacter genotype, its identification by PCR from the diseased livers of all aged hamsters, and its taxonomic classification in the Helicobacter bilis cluster further strengthen the argument that H. bilis and closely related Helicobacter spp. play an etiological role in hepatobiliary disease in both animals and humans. The results of the current study should stimulate further experiments with hamsters infected with organisms in the Helicobacter bilis cluster and liver flukes as well as investigations regarding coinfections of hepatic helicobacters and liver flukes in human populations at high risk of developing cholangioscarcinoma.
This work was supported by NIH grants R01CA67529 (to J.G.F.), P30ES102109 (to J.G.F.), RR07036 (to J.G.F.), and DE016937 (to F.E.D.).
Published ahead of print on 16 September 2009.