Enterohepatic Lesions in SCID Mice Infected with Helicobacter Bilis

Enterohepatic Lesions in SCID Mice Infected with Helicobacter bilis

Craig L. Franklin, Lela K. Riley, Robert S. Livingston, Catherine S. Beckwith, Cynthia L. Besch-Williford, and Reuel R. Hook, Jr.

Abstract ͉ Helicobacter bilis is a recently identified species that colonizes the intestine and liver of mice. In immunocompetent mice, infections have been associated with mild hepatitis, and in immunocompromised mice, inflammatory bowel disease has been induced by intraperitoneal inoculation of the organism. We re- port inoculation of 6-week-old C.B-17 scid/scid mice by gastric gavage with approximately 107 H. bilis colony- forming units. Groups of mice were euthanized and necropsied 12, 24, and 36 weeks after inoculation. Mild to moderate proliferative typhlitis was evident in all mice at 12 and 36 weeks after inoculation and in most mice 24 weeks after inoculation. Mild to severe chronic active hepatitis was detected in 10 of 10 male mice and 3 of 10 female mice. These results indicate that H. bilis can cause moderate to severe enterohepatic disease in immunocompromised mice.

The genus Helicobacter is a rapidly expanding genus volved in lesion development. Culture of specimens from currently containing 17 named species. Members of this mice confirmed intestinal colonization with H. hepaticus. Fox genus are microaerophilic, have curved to spiral rod mor- et al. reported enteric lesions in immunocompetent germ- phology, and are motile by flagella that vary in number free Swiss Webster mice infected with H. hepaticus (15), and and location among various species (1). All known Cahill et al. recently reported inflammatory bowel disease Helicobacter spp. live in animal hosts where colonization in C.B-17 scid/scid mice inoculated with H. hepaticus, fur- is restricted to the intestine or liver. The type species of ther strengthening the role of this bacterium as an etiologic the genus is Helicobacter pylori. In all likelihood, this or- agent of intestinal as well as liver disease (17). ganism was present in the stomach of humans for years Much less is known about H. bilis, which has been docu- but was not recognized as a pathogen until 1985, when mented to colonize the intestine and liver of several strains Marshall et al. established H. pylori as the causal agent of of mice (8). Hepatic colonization was associated with mild chronic gastritis and peptic ulcer disease in people (2–4). hepatitis in aged mice, and hepatomas have been identi- Several species of Helicobacter have been identified in fied in aged CBA/CA mice infected with H. bilis (8). Re- rodents, including H. hepaticus, H. bilis, H. muridarum, cently, H. bilis was reported to cause inflammatory bowel H. cinaedi, H. cholecystus. H. trogontum, H. rodentium, and disease in defined flora ICR-scid/scid mice inoculated in- the related bacterium, “Flexispira rappini” (5–12). Of these, traperitoneally with the bacterium; however, liver lesions H. hepaticus and H. bilis have received the most attention were not evident (18). Here, we describe proliferative typh- because they have been associated with disease and are litis and chronic active hepatitis in conventional flora C.B- the most prevalent species in laboratory animals. 17 scid/scid mice experimentally infected with H. bilis by Helicobacter hepaticus was first recognized in 1992, when gastric gavage. hepatitis was observed in untreated A/JNCr mice during a long-term carcinogenicity study (13, 14). Examination of Materials and Methods silver-stained sections of liver revealed slender, curved to Animals: The C.B-17-scid/scid (SCID) mice used in in- spiral-shaped bacteria associated with lesions. A bacterium oculation studies were obtained from Charles River Labo- subsequently cultivated from the intestinal tract and liver ratories (Raleigh, N.C.). Vendor surveillance indicated that was found to be a novel member of the genus Helicobacter mice were free of all known bacterial, viral, and parasitic and was named Helicobacter hepaticus. In experimental pathogens, including but not limited to the enteric patho- studies, Koch’s postulates were fulfilled, confirming the role gens, mouse hepatitis virus (MHV), Citrobacter rodentium, of this bacterium in disease pathogenesis (13–15). and Clostridium piliforme. Mice were also determined to In 1996, Ward et al. reported high incidence of rectal be Helicobacter free by generic Helicobacter fecal PCR ex- prolapse and chronic proliferative typhlocolitis and procti- amination (19). Mice were housed in polycarbonate tis in immunodeficient mice (16). Helical organisms con- microisolator cages containing Paperchip Laboratory Ani- sistent with H. hepaticus were seen within the lumen of mal Bedding (Canbrands International, Ltd., Moncton, New the affected colon, suggesting that these bacteria were in- Brunswick, Canada) and were allowed ad libitum access to Autoclavable Laboratory Rodent Diet 5010 (Purina Mills, Department of Veterinary Pathobiology, University of Missouri, Columbia, Inc., Richland, Ind.) and tap water. The animal room was Missouri

334 Helicobacter bilis Infection in SCID Mice

maintained at 24ЊC, 45% humidity, and Table 1. Incidence of cecal (proliferative typhlitis) and hepatic (chronic active hepatitis) lesions in 14.5 air changes/h; a 14/10-h light/dark mice inoculated with Helicobacter bilis and in unmanipulated control mice Mice inoculated with Control mice cycle was used. Cages, water, and feed Weeks after Helicobacter bilis (Uninoculated) were autoclaved prior to use, and cage Tissue inoculation Male (n = 10)a Female (n = 9) Male (n = 10) Female (n = 2) changes were performed under a lami- Cecum 12 3/3* 3/3 0/4 0/2 nar flow hood. During the course of this 24 2/3 1/3 0/2 NTb 36 4/4* 3/3 0/4 NT study, sentinel mice housed in the ani- Total 9/10* 7/9 0/10 0/2 mal room were subjected to monthly se- Liver 12 3/3 0/3 0/4 0/2 rologic and parasitologic examinations. 24 3/3 2/3 0/2 NT 36 4/4* 1/3 0/4 NT Sentinel mice were consistently found Total 10/10* 3/9 0/10 0/2 free of MHV, Sendai virus, pneumonia aNumber of mice with lesions score Ն2/number of mice in the group virus of mice, Theiler’s murine encepha- bNT = not tested; female mice in the 24- and 36-week postinoculation groups were contaminated with lomyelitis virus, ectromelia, epizootic di- another Helicobacter species during the course of the study and were removed. Statistical analysis was not done on data from female groups at these time points. arrhea of infant mice virus, lymphocytic *Significantly different from control (P < 0.05); Fisher’s exact test choriomeningitis virus, mouse adenoviruses, polyoma virus, reovirus type 3, cilia-associ- lobe also were collected for H. bilis PCR analysis. ated respiratory bacillus, Encephalitozoon cuniculi, My- Histologic examination: Biopsy specimens of liver and coplasma pulmonis, Helicobacter hepaticus, Clostridium intestine were preserved in neutral-buffered 10% forma- piliforme, mouse parvoviruses, and endo- and ectoparasites. lin. The liver was sectioned through the medial left lobe, All animal use procedures were approved by the University the gallbladder, and the medial right lobe; the distal por- of Missouri Institutional Animal Health and Use Committee. tion of the colon was sectioned longitudinally; and the Bacteria: Helicobacter bilis, strain MU, was isolated ileocecocolic biopsy specimen was sectioned longitudinally from a colony of naturally infected BALB/c mice. Briefly, through the ileocecal junction, the cecocolic junction, most fecal pellets from mice were homogenized in phosphate- of the cecal tip, and the proximal portion of the colon. buffered saline (PBS) and incubated at room temperature Trimmed biopsy specimens were embedded in paraffin, for 30 min. Fecal slurries were filtered through a 0.80-␮m and 3-␮m sections were prepared, stained with hematoxy- filter, and a single droplet was placed on plates of trypticase lin and eosin, and examined. To objectively evaluate the soy agar that contained 5% sheep blood (blood agar). Cul- severity of proliferative typhlocolitis and chronic active tures were incubated in a microaerobic environment of 90% hepatitis, a lesion scoring system was developed (Tables

N2, 5% CO2, and 5% H2. Pinpoint colonies that grew under 2 and 3), and two pathologists (CSB and RSL) assigned these conditions were tested for urease activity (rapid ure- lesion scores without prior knowledge of mouse infection ase test; Remel, Lenexa, Kans.) and motility. Motile ure- status (blind study). Discrepancies in lesion scores be- ase-positive isolates were subcultured on blood agar, and a tween the two pathologists were resolved by their view- first-passage subculture was used as inoculum. The isolate ing slides together and agreeing on a single score; the lat- was confirmed to be H. bilis by sequencing of a 374-bp poly- ter discussions were also conducted without prior knowl- merase chain reaction (PCR)-amplified fragment from the edge of mouse infection status. 16S rRNA gene. Sequence analysis revealed 100% simi- Generic Helicobacter and Helicobacter bilis PCR: larity with the type species of H. bilis (GenBank acces- Feces or frozen hepatic tissue were analyzed for H. bilis by sion no. U18766). PCR analysis as described (19, 20). Briefly, DNA was iso- Experimental design: Ten female and 10 male lated from feces or tissues, using a QIAmp Tissue Kit Helicobacter-free 4-week-old SCID mice were inoculated (Quiagen, Inc., Chatsworth, Calif.). The PCR analyses were with approximately 107 H. bilis colony-forming units in 0.5 performed, using Helicobacter genus-specific primers in a ml of sterile PBS by gastric gavage. Two weeks after inocu- Perkin-Elmer 2400 thermocycler. The PCR products were lation, fecal pellets were collected from cages harboring digested in separate reactions with Mbo I, Mae I, and Hha I inoculated SCID mice and were tested for H. bilis by PCR. restriction endonucleases, and digestion products were sepa- At postinoculation weeks 12, 24, and 36, groups of two to rated by gel electrophoresis and visualized. Products cleaved four uninoculated control and three to four inoculated mice by all three restriction enzymes were diagnosed as H. bilis were euthanized by inhalation of an overdose of carbon di- (20). To confirm presence of H. bilis, select specimens were oxide and were examined by necropsy. Table 1 outlines the analyzed by use of a PCR assay containing a generic for- number of mice in each group. At necropsy, biopsy speci- ward primer (H276f) and a reverse primer specific for mens of the right and left medial liver lobes and intestine H. bilis (Hb664r). Primers used for the generic Helicobacter were prepared for histologic examination. Intestinal speci- and H. bilis-specific PCR assays are listed in Table 4. mens included a 1-cm section of the distal portion of the Statistical analyses: Prevalence of hepatic or intesti- colon and a specimen that contained the aboral 1 cm of the nal lesions was compared between experimentally inocu- ileum, the ileocecal junction, three-fourths of the cecum, lated mice and control mice in toto and at all time points and the oral 1 cm of the proximal portion of the colon. Fresh and between sexes. Fisher’s exact test was used for these fecal samples and a biopsy specimen of the left lateral liver comparisons. Mice were considered to have significant lesions if the lesion score was Ն2.

335 Vol 48, No 4 Laboratory Animal Science August 1998

Table 2. Histologic criteria used to assign proliferative typhlitis scores Lesion score Criteria 0 No significant lesions; no conclusive evidence of hyperplasia; crypt portion comprises approximately 1/3 of the cecal gland; no piling up of crypt enterocytes; mucosal height may vary slightly but should retain normal architecture. 1 Few focal areas of mild mucosal hyperplasia characterized by increase in mucosa height and presence of basophilic “crypt” epithelial cells in at least the lower 2/3 of gland. 2 Mild multifocal to diffuse hyperplasia characterized by a general increase in mucosa height, general increase in basophilic staining of mucosa, presence of basophilic “crypt” epithelial cells in at least the lower 2/3 of gland, and loss of goblet cells in few crypts. 3 Mild to moderate multifocal to diffuse hyperplasia characterized by a general increase in mucosa height, general increase in basophilic staining of mucosa, presence of “crypt” features in at least the lower 1/2 to 2/3 of gland AND loss of goblet cells in several crypts OR lesion score 2 with focal areas of moderate hyperplasia characterized by at least a doubling in the mucosa height with crypt features occupying most of the gland (pseudopolyp formation). 4 Moderate multifocal to mucosal diffuse hyperplasia characterized by lesion score of 3 plus areas of marked increase in height (greater than twice normal height), pseudopolyp formation, or evidence of gland branching and tortuosity; hyperplasia may be accompanied Figure 1. (A) Photomicrograph of a section of colon from a by mild neutrophilic or mononuclear cell inflammation. C.B-17 scid/scid mouse infected with Helicobacter bilis; prolif- 5 Lesion score of 4 with moderate inflammation and/or invasion of erative typhlitic lesions are characterized by severe mucosal hy- muscularis mucosae by glands. perplasia with associated chronic inflammation. (B) Photomicro- 6 Lesion score of 5 with marked inflammation and mucosal ulcer- graph of a section of cecum from an uninfected SCID mouse for ation. comparison of mucosal height. Bars = 100 ␮m.

Table 3. Histologic criteria used to assign chronic active hepatitis scores Results Lesion Bacterial colonization: Two weeks after inoculation, score Criteria all SCID mice were positive for H. bilis colonization by fe- 0 No evidence of hepatocellular necrosis, suppurative inflammation, or oval cell hyperplasia; mild portal mononuclear cell infiltrates cal PCR analysis; all control mice were Helicobacter nega- may be present. tive. During the course of the study, one inoculated female 1 Few (1 to 5/liver section) foci of very mild (2 to 5 hepato- mouse died for reasons unrelated to H. bilis infection, and cytes affected) hepatocellular necrosis OR female mice in the 12- and 24-week postinoculation groups few small aggregates of neutrophils (5 to 10 cells). were removed from the study because they were inadvert- 2 Multiple (>5/liver section) foci of very mild (2 to 5 hepatocytes affected) hepatocellular necrosis ently contaminated by another Helicobacter species. At OR postinoculation weeks 12, 24, and 36, groups of SCID mice few (1 to 5 foci/liver section) foci of mild (5 to 10 hepatocytes af- were euthanized and necropsied. Clinical signs of disease fected) hepatocellular necrosis with associated inflammation and/ or oval cell hyperplasia. were not evident at any time during the course of the study. 3 Multiple (>5) foci of mild (5 to 10 hepatocytes affected) hepatocell- At necropsy, 18 of 19 experimentally inoculated mice were ular necrosis with associated inflammation and/or oval cell hyper- positive for H. bilis by results of fecal PCR, and all control plasia OR mice were negative for H. bilis. One of three male and one few (1 to 3) foci of moderate (10 to 50 hepatocytes affected) hepa- of three female mice of the 12-week postinoculation group, tocellular necrosis with associated inflammation and/or oval cell hyperplasia (some chronic foci may be predominantly inflamma- one of three male mice from the 24-week postinoculation tory). group, and two of four male mice from the 36-week postin- 4 Multiple (>3) foci of moderate (10 to 50 hepatocytes affected) hepa- oculation group also were positive for H. bilis by PCR of tocellular necrosis with associated suppurative inflammation and/ or oval cell hyperplasia (some chronic foci may be predominantly liver specimens. inflammatory). Intestinal histopathologic findings: Lesions in the 5 Multiple to coalescing (>3) foci of severe (>50 hepatocytes affected) cecum consisted predominantly of mucosal hyperplasia with hepatocellular necrosis with associated inflammation and/or oval cell hyperplasia (some chronic foci may be predominantly inflam- associated inflammation (Figure 1). The inflammatory matory). population varied from acute, consisting principally of neutrophils, to chronic, consisting of mononuclear cells and Table 4. Oligonucleotide primers used to amplify the 16S rRNA gene macrophages. Mild hyperplasia was sometimes evident at sequences of Helicobacter bilis the ileocecal junction in control mice. Because of this find- Primer Sequence (5 to 3) Positionsa ing, proliferative typhlitis scores were determined, using Helicobacter genus-specific PCR only mucosae distant from the ileocecal junction. Aside from H276f CTA TGA CGG GTA TCC GGC 276–293 mild hyperplasia at the ileocecal junction, no control mice H676r ATT CCA CCT ACC TCT CCC A 676–658 H. bilis-specific reverse primer had any evidence of cecal disease. Hb664r TCT CCC ATA CTC TAG AAA AGT 643–664 Mild mucosal hyperplasia with associated chronic inflam- aApproximate positions for Helicobacter-specific primers

336

Helicobacter bilis Infection in SCID Mice Lesion score Lesion

Figure 3. Photomicrograph of a section of liver from a C.B-17 Weeks after inoculation scid/scid mouse infected with H. bilis; necrotizing hepatitis lesions are characterized by coagulative necrosis (left side) with Figure 2. Proliferative typhlitis scores in mice infected with associated minimal mononuclear cell inflammation. Bar = 20 ␮m. H. bilis and in uninfected controls. Numbers above symbols indicate the number of mice with that lesion score; symbols with no number represent one animal. mation was observed in the proximal portion of the colon of 6 of 19 H. bilis-infected mice. With the exception of one female mouse from the 12-week postinoculation group, all mice with proximal colitis had concurrent typhlitis. Relevant lesions were not seen in the distal portion of the colon. Proliferative typhlitis or colitis of the proximal colonic segment (lesion score Ն2) was seen in all inoculated males from the 12- and 36-week postinoculation groups and all score Lesion inoculated females from the 12-week postinoculation group (Table 1 and Figure 2). This resulted in a significant (P < 0.05) difference in prevalence of lesions between inoculated male mice and control mice from these groups. Prevalence of cecal or proximal colonic lesions was not significantly different between inoculated and control female Weeks after inoculation mice at 12 weeks, inoculated and control male mice at 24 weeks, and total inoculated and control female mice. How- Figure 4. Chronic active hepatitis scores in mice infected with H. bilis and in uninfected controls. Numbers above symbols indi- ever, these results must be interpreted carefully because cate the number of mice with that lesion score; symbols with no the powers of these analyses were low (Յ0.605) due to the number represent one animal. low number of control mice (n = 2) at these time points. It is likely that results would be statistically significant if and control mice was compared, there were statistically sufficient numbers of control mice were available to raise significant (P < 0.05) differences in the male mouse groups the power of the test above 0.80. Statistical analyses could at postinoculation weeks 12 and 36. The prevalence of liver not be performed on female mice from the 24- and 36-week lesions was not significantly different between inoculated postinoculation groups because there were no female con- and control male mice at 24 weeks, but the power of this trols at these time points. analysis was low (0.605). Statistical analyses could not be Liver histopathologic findings: Relevant liver lesions performed on female mice from 24- and 36-week postin- consisted of some degree of chronic active hepatitis (lesion oculation groups because there were no female controls at score Ն2). Lesions ranged from few small foci of necrosis these time points. When data were pooled, the incidence of involving two to three hepatocytes per focus to large coa- lesions was significantly (P < 0.05) different between male lescing foci of coagulative necrosis with associated inflam- and female mice, suggesting that there was a sex bias in mation (Figure 3) or oval cell hyperplasia. Chronic active hepatic disease. In addition to necrosis with associated in- hepatitis was observed in all inoculated male mice, but in flammation, mild portal mononuclear cell inflammation was only two of three inoculated female mice from the 24-week observed in one of three female mice each from the 12- and postinoculation group and one of three inoculated female 24-week postinoculation groups, three of three female mice mice from the 36-week postinoculation group (Table 1 and from the 36-week postinoculation group, and three of four Figure 4). When prevalence of lesions between inoculated male mice from the 36-week postinoculation group.

337 Vol 48, No 4 Laboratory Animal Science August 1998

Hepatic lesions were seen in all male mice in which tion. Susceptibility to H. hepaticus-induced disease has H. bilis was identified by PCR of liver specimens (n = 4). been documented to vary among different inbred stains (13, However, six mice had lesions in the liver, but results of 14). Although the mechanisms for strain susceptibility to PCR were negative for H. bilis colonization of this organ. H. hepaticus-induced disease is unknown, it is possible that Of the nine inoculated female mice examined, the three subtle differences between different strains or stocks of mice with lesions had negative results of PCR for H. bilis he- (e.g., differences in bacterial receptors or physiochemical patic colonization, and one mouse with no lesions had posi- state of intestinal luminal contents) could alter the patho- tive results of PCR. genesis of enteric Helicobacter-induced disease. It also is possible that the strains of H. bilis used in these Discussion two studies differed in virulence. Although these strains Little is known about the pathogenesis of H. bilis infec- were 100% identical by rRNA sequencing analysis, it is tions. The bacterium is probably transmitted by ingestion probable that they were not genetically identical. Marked of contaminated feces. In immunocompetent mice in which genetic diversity is common among isolates of H. pylori, this bacterium was originally identified, mild hepatitis and this diversity may influence virulence (22). characterized by portal, perivascular, or randomly distrib- Lastly, route of inoculation may have influenced coloni- uted aggregates of lymphocytes, macrophages, and a few zation. Intraperitoneal inoculation of H. bilis may have polymorphonuclear cells was seen in some infected mice; resulted in a higher initial dose to the colon, compared with intestinal lesions were not described (8). We describe pro- gastric route of inoculation. This, in turn, may have facili- liferative typhlitis and chronic active hepatitis in im- tated colonization of this part of the intestine. To reiterate, munodeficient mice that were experimentally inoculated the reasons for the differences in the disease induced in by gastric gavage with H. bilis. The route of inoculation is these two studies are unknown. This phenomenon, how- noteworthy because it likely mimics the natural fecal-oral ever, highlights the need for further studies about patho- route of infection. genesis of enteric Helicobacter infections of rodents. Shomer et al. recently described inflammatory bowel dis- The proliferative typhlitis in mice infected with H. bilis ease in ICR-scid/scid mice experimentally inoculated with was remarkably similar to enteric lesions in SCID H. bilis (18). It is difficult to compare the results of that mice, either naturally or experimentally infected with study with those of our study because of a number of H. hepaticus (16, 17). Both are characterized by prolifera- methodologic differences. For example, defined flora ICR- tion of cecal, and to a lesser extent, proximal colonic mu- scid/scid mice were inoculated intraperitoneally with the cosal epithelium with accompanying mild chronic inflam- type strain of H. bilis, whereas in our studies, conventional mation. This suggests that pathogenesis of these bacteria flora C.B-17 scid/scid mice were inoculated with the MU may be similar. However, elucidation of virulence mecha- strain of H. bilis by gastric gavage. Shomer et al. exam- nisms used by both bacteria awaits further studies. Al- ined only female mice at postinoculation week 7, whereas though lesions in H. bilis-infected mice were similar to those we examined male and female mice at postinoculation in H. hepaticus-infected mice, they differed in severity. We weeks 12, 24, and 36. Despite these differences in method, have used a similar experimental inoculation protocol to our results are similar to those of Shomer et al., in that infect C.B-17 scid/scid mice with H. hepaticus (unpublished inoculation of an immunodeficient murine host with H. bilis data). In H. bilis-infected C.B-17 scid/scid mice, lesion scores induced predominantly proliferative chronic inflammatory ranged from 2 to 4, whereas most SCID mice infected with bowel disease of the cecum. However, two notable differ- H. hepaticus developed lesions with scores of 4 to 5. ences were found. Shomer et al. found moderate prolifera- Mild hyperplastic lesions were observed at the ileocecal tive lesions in the distal portion of the colon of 75% and junction in some Helicobacter-free C.B-17 scid/scid mice. diarrhea in 20% of inoculated mice. In contrast, in our in- The cause of these lesions was not determined. These le- oculated mice, distal colonic lesions were not found, and sions may be associated with change in the intestinal envi- the disease was subclinical. The reasons for these discrep- ronment, either microbial flora or physiochemical state, ancies are unknown, but several possibilities exist. Flora between the ileum and cecum. Importantly, for the diag- inhabiting mice may have had an impact on the ability of nostic pathologist, the finding of mild mucosal hyperplasia H. bilis to colonize the large intestine. It is conceivable that, in our control mice highlights the fact that this area of the in defined flora mice, H. bilis more readily colonized the intestine should not be relied on when evaluating conven- distal portion of the intestine, compared with that of con- tional flora immunodeficient mice for Helicobacter-induced ventional flora mice, where the different bacterial flora may intestinal disease. have precluded H. bilis colonization. In addition to intestinal lesions, we identified chronic ac- The relative immunocompetence of mice may have in- tive hepatitis in H. bilis-infected C.B-17 scid/scid mice. Le- fluenced colonization. The C.B-17 scid/scid mice are known sions were similar to those described in immunocompromised to have more “leakiness,” compared with mice with the scid and immunocompetent mice infected with H. hepaticus (13– mutation on other inbred backgrounds (21). Hence, the mice 15). This is in contrast to the mild chronic hepatitis de- of our study may have developed some H. bilis immunity scribed in immunocompetent mice infected with H. bilis that prevented formation of distal colonic lesions. (8). In our study, hepatic lesions were more prevalent and The background of mice may have influenced coloniza- more severe in male mice, compared with inoculated fe-

338 Helicobacter bilis Infection in SCID Mice

male mice. These findings indicate that, similar to that of 5. Shen, Z., J. G. Fox, F. E. Dewhirst, et al. 1997. Helicobacter H. hepaticus, there is a sex bias in prevalence of H. bilis- rodentium sp. nov., a urease-negative Helicobacter species isolated from laboratory mice. Int. J. Syst. Bacteriol. induced liver disease. 47:627–634. Presence of liver lesions did not correlate with amplifi- 6. Franklin, C. L., C. S. Beckwith, R. S. Livingston, et al. cation of H. bilis DNA from the liver. The reasons for this 1996. Isolation of a novel Helicobacter species, Helicobacter are unknown. Because samples for PCR and histologic ex- cholecystus sp. nov., from the gallbladders of Syrian ham- amination were obtained from different liver lobes, it is sters with cholangiofibrosis and centrilobular pancreatitis. J. Clin. Microbiol. 34:2952–2958. possible that the discrepancies between these two tests were 7. Mendes, E. N., D. M. Queiroz, F. E. Dewhirst, et al. 1996. due to patchy colonization or disease. For example, a mouse Helicobacter trogontum sp. nov., isolated from the rat intes- that had liver lesions but had negative results of PCR for tine. Int. J. Syst. Bacteriol. 46:916–921. H. bilis may have had H. bilis colonization only in the liver 8. Fox, J. G., L. L. Yan, F. E. Dewhirst, et al. 1995. Helicobacter bilis sp. nov., a novel Helicobacter species iso- lobe that was collected for histologic examination. Several lated from bile, livers, and intestines of aged, inbred mice. other possibilities should be considered, including: low J. Clin. Microbiol. 33:445–454. numbers of H. bilis are sufficient to induce histologic dis- 9. Fox, J. G., F. E. Dewhirst, J. G. Tully, et al. 1994. ease, an unlikely scenario because the PCR used has been Helicobacter hepaticus sp. nov., a microaerophilic bacterium documented to detect as few as five organisms (19); H. bilis isolated from livers and intestinal mucosal scrapings from mice. J. Clin. Microbiol. 32:1238–1245. products or fragments may disseminate to the liver, and 10. Schauer, D. B., N. Ghori, and S. Falkow. 1993. Isolation these products may be adequate for lesion induction, pre- and characterization of “Flexispira rappini” from laboratory cluding need for bacterial colonization; and H. bilis may mice. J. Clin. Microbiol. 31:2709–2714. transiently colonize the liver, with initial colonization suf- 11. Lee, A., M. W. Phillips, J. L. O’Rourke, et al. 1992. Helicobacter muridarum sp. nov., a microaerophilic helical bac- ficient to induce lesion formation. Further studies are nec- terium with a novel ultrastructure isolated from the intesti- essary to determine which if any of these scenarios apply nal mucosa of rodents. Int. J. Syst. Bacteriol. 42:27–36. to H. bilis-infected mice. 12. Gebhart, C. J., C. L. Fennell, M. P. Murtaugh, et al. 1989. In conclusion, H. bilis infection induced by gastric gav- Campylobacter cinaedi is normal intestinal flora in hamsters. age can cause enterohepatic disease in immunocompromised J. Clin. Microbiol. 27:1692–1694. 13. Ward, J. M., J. G. Fox, M. R. Anver, et al. 1994. Chronic mice. Similar to H. hepaticus, H. bilis has the potential to active hepatitis and associated liver tumors in mice caused jeopardize research using infected mice, especially studies by a persistent bacterial infection with a novel Helicobacter involving the intestine and liver. That H. bilis-induced dis- species. J. Natl. Canc. Inst. 86:1222–1227. ease may be subclinical, and infection may insidiously 14. Ward, J. M., M. R. Anver, D. C. Haines, et al. 1994. Chronic active hepatitis in mice caused by Helicobacter hepaticus. spread throughout a mouse colony without appearance of Am. J. Pathol. 145:959–968. clinical diseases, raises additional concerns about the po- 15. Fox, J. G., L. Yan, B. Shames, et al. 1996. Persistent hepa- tential for this bacterial infection to compromise biomedi- titis and enterocolitis in germfree mice infected with cal research. Helicobacter hepaticus. Infect. Immun. 64:3673–3681. 16. Ward, J. M., M. R. Anver, D. C. Haines, et al. 1996. In- flammatory large bowel disease in immunodeficient mice naturally infected with Helicobacter hepaticus. Lab. Anim. Sci. Acknowledgements 46:15–20. We thank Sue Bingamen, Allen Maddy, and Beth Livingston 17. Cahill, R. J., C. J. Foltz, J. G. Fox, et al. 1997. Inflam- for excellent technical support and Howard Wilson for photo- matory bowel disease: an immunity-mediated condition graphic assistance. triggered by bacterial infection with Helicobacter hepaticus. Infect. Immun. 65:3126–3131. 18. Shomer, N. H., C. A. Dangler, M. D. Schrenzel, et al. 1997. Helicobacter bilis-induced inflammatory bowel disease in References scid mice with defined flora. Infect. Immun. 65:4858–4864. 1. Vandamme, P., E. Falsen, B. Pot, et al. 1990. Identifica- 19. Beckwith, C. S., C. L. Franklin, R. R. Hook, Jr., et al. tion of Campylobacter cinaedi isolated from blood and feces of 1997. Fecal PCR assay for diagnosis of Helicobacter infection children and adult females. J. Clin. Microbiol. 28:1016–1020. in laboratory rodents. J. Clin. Microbiol. 35:1620–1623. 2. Blaser, M. J., and J. Parsonnet. 1994. Parasitism by the 20. Riley, L. K., C. L. Franklin, R. R. Hook, Jr., et al. 1996. “slow” bacterium Helicobacter pylori leads to altered gastric Identification of murine Helicobacters by PCR and restric- homeostasis and neoplasia. J. Clin. Invest. 94:4–8. tion enzyme analyses. J. Clin. Microbiol. 34:942–946. 3. Marshall, B. J. 1990. Campylobacter pylori: its link to gas- 21. Nonoyama, S., F. O. Smith, I. D. Bernstein, et al. 1993. tritis and peptic ulcer disease. Rev. Infect. Dis. 12:S87–S93. Strain-dependent leakiness of mice with severe combined 4. Marshall, B. J., and J. R. Warren. 1984. Unidentified curved immune deficiency. J. Immunol. 150:3817–3824. bacilli in the stomach of patients with gastritis and peptic 22. Blaser, M. J. 1997. Heterogeneity of Helicobacter pylori. ulceration. Lancet 1:1311–1315. Eur. J. Gastroenterol. Hepatol. 9:S3–S6; discussion S6–S7.

339