j ourna l homepage: www.e lsev ie r .com/ locate / jpedsurg
Rotavirus particles in the extrahepatic bile duct in experimental biliary atresia
Christina Oetzmann von Sochaczewski a,⁎, Isabel Pintelon b, Inge Brouns b, Anika Dreier a,Christian Klemann a, Jean-Pierre Timmermans b, Claus Petersen a, Jochen Friedrich Kuebler a
a Department of Pediatric Surgery, University Hospital of Hannover, Hannover, Germanyb Laboratory of Cell Biology & Histology, Department of Veterinary Sciences, University of Antwerp, Antwerp, Belgium
a b s t r a c ta r t i c l e i n f o
⁎ Corresponding author. Department of General SurgWiesbaden, Germany. Tel.: +49 176 61124081.
Received 31 May 2013Received in revised form 10 August 2013Accepted 27 September 2013
Key words:Biliary atresiaMouse modelRotavirusObstructionLiverExtrahepatic bile duct
Background: Biliary atresia (BA) is the most common indication for liver transplantation in children. Theexperimental model of BA, induced by rotavirus infection in neonatal mice, has been widely used toinvestigate the inflammatory aspects of this disease. We investigated the kinetics and the localization of theviral infection in this murine model.Methods: In this study 399 animals were employed for a detailed investigation of rhesus rotavirus (RRV)-induced BA. RRV kinetics was analyzed by rtPCR and its (sub) cellular localization investigated using wholemounts which were further processed for confocal and electron microscopy.Results: The BA mouse model resulted in up to 100% induction of atresia following RRV injection. The kineticsof RRV infection differed between liver and extrahepatic bile ducts. While the virus peak up to day 10postinfection was similar in both organs, the virus remained detectable in extrahepatic bile duct cells up today 21. Interestingly, RRV particles were localized not only in cholangiocytes but also in cells of the
subepithelial layers, potentially macrophages.Conclusions: RRV remains present in the extrahepatic bile duct cells after an initial virus peak. Viral particleswere detected in subepithelial cells in contrast to the described tropism toward cholangiocytes.
Biliary atresia (BA) is the most common cause of chronicprogressive liver disease in childhood and is the leading indicationfor pediatric liver transplantation worldwide. The etiology of BAremains unknown, but one leading hypothesis proposes a virus as thetriggering event leading to BA [1,2]. In BA patients, several viral strainshave been detected in liver or blood samples [3,4]; however,simultaneous screening of BA patients for all common hepatotropicviruses, yielded positive results for only 30%–55% of patientsundergoing the Kasai procedure [5–7]. Experimental BA is inducedby postpartum intraperitoneal infection of BALB/c mice with rhesusrotavirus (RRV) [8,9]. RRV is widely prevalent in the population as themost common cause for diarrhea in infants and children and has beenone of the viruses identified in livers of patients with BA [7,10,11]. Inthis experimental model, the virus is cleared in the liver, prior to thedevelopment of the full clinical picture of BA, but triggers aninflammatory reaction that causes the fibrosing destruction of theextrahepatic bile ducts [8,9,12,13]. Most studies describing experi-mental and clinical BA have focused on liver tissues. However, wehypothesized, that there are differences in the virus kinetics betweenliver and extrahepatic biliary tissue. Therefore we assessed thedynamics of the viral load in both tissues and used electron
ery, Dr. Horst Schmidt Klinik,
n von Sochaczewski).
l rights reserved.
microscopy to localize the viral particles in the extrahepatic bileducts of affected animals.
1. Methods
1.1. Biliary atresia animal model
In total, 399 newborn BALB/c mice were divided into two groupsand injected within the first 24 h postpartum. The control group(n = 102) was injected with 50 μl saline. Two hundred ninety-sevenanimals were infected intraperitoneally with 50 μl containing2.5 × 106 pfu RRV [14]. Animals dying within 72 h because of theinjection were excluded as described in the Results section. Animalswere evaluated every other day for weight, cholestasis and fur-covered skin up to day 21 postinfection. BA was scored at the timepoint of preparation of liver and extrahepatic bile ducts.
The extrahepatic bile duct samples were used either for whole-mount preparation (infected n = 95, control n = 80) and immuno-histochemistry imaged by means of confocal microscopy, rtPCR(infected n = 96, control n = 0) or electron microscopic analysis(infected n = 19, control n = 4). Animals used in a second studywere included in the numbers for whole-mount samples and analyzedhere only for clinical symptoms. A minimum of four animals weresacrificed daily for whole-mount analysis from day 1 to day 10 andadditionally on day 12 and day 14 postinfection. A group of eightanimals was analyzed by rtPCR at various time points. Samples were
Fig. 1. Development of RRV-induced biliary atresia in mice during the first 21 dayspostinfection. (A) RRV-infected animals show reduced weight gain. (B) Most infectedanimals show cholestasis from day 5 on followed by an oily fur around day 10. (C) Themajority of all animals develop BA between day 12 and day 14. The high incidenceobserved at days 9 and 10 may be overrepresented because of the small group size.
521C. Oetzmann von Sochaczewski et al. / Journal of Pediatric Surgery 49 (2014) 520–524
collected daily around the expected peak of infection and lessfrequently at early and late time points (Fig. 3). Extrahepatic bileducts were sampled for electronmicroscopy at days 1, 2, 3, 5, 8, 10, 12,and 14 postinfection.
All experiments of this study were approved by the governmentand performed according to the national regulations for theprotection of animal models (registration number 11-0360).
1.2. Virus production
RRV strain MMU 18006 was grown in MA-104 African greenmonkey kidney cells and assayed for concentration by infectiousplaque assay as previously described [8,9].
1.3. Whole mount of bile ducts
Bile duct whole mounts were immersion-fixed 30 min afterisolation with 4% paraformaldehyde (in 0.1 M phosphate buffer;pH 7.4). Immunohistochemical incubations were carried out at roomtemperature on free-floating bile ducts. All primary and secondaryantisera were diluted in phosphate-buffered saline (PBS; 0.01 M;pH 7.4) containing 10% normal horse serum, 0.1% bovine serumalbumin, 0.05% thimerosal, 0.01% NaN3 1% and Triton-X-100. Prior toincubation with the primary antisera, whole mounts were incubatedfor 1 h with the antibody diluent. Whole mounts were incubatedovernight with a monoclonal primary antibody raised in rat againstthe endothelial marker CD31 (1:50; Abcam ab56299, Cambridge, UK).To visualize immunoreactivity for CD31, whole mounts werefurther incubated for 4 h with Cy3-conjugated donkey antiratimmunoglobulins (DARa-Cy3; 1:200; Jackson ImmunoResearch,712-165-150, West Grove, PA). Whole mounts were then incubatedfor a consecutive night with a second primary antibody against RRV(1/2000; Sh Pc; Abcam ab35417) followed by a 4-h secondaryincubation with FITC-conjugated donkey antisheep immunoglobulins(DASh-FITC; 1:200; Jackson ImmunoResearch, 713-095-003).
High-resolution images were obtained using a microlens-en-hanced dual spinning disk confocal microscope (UltraVIEW VoX;PerkinElmer, Seer Green, UK) equipped with 488-nm and 561-nmdiode lasers for excitation of FITC and Cy3, respectively. Images wereprocessed and analyzed using Volocity software.
1.4. Transmission electron microscopy
Bile ducts were immersion-fixed after isolation in 2.5% glutaral-dehyde solution for 30 min, rinsed in 0.1 M sodium cacodylate-buffered (pH 7.4) and postfixed in 1% OsO4 solution for 2 h. Afterdehydration in an ethanol gradient (70% ethanol for 20 min, 96%ethanol during 20 min, 100% ethanol for 2 × 20 min), whole mountswere embedded in EMbed 812 (Electron Microscopy Sciences,Hatfield, PA). Ultrathin sections were stained with 2% uranyl acetateand lead citrate, and examined in a Tecnai G2 Spirit Bio TwinMicroscope (FEI, Eindhoven, the Netherlands) at 120 kV.
1.5. Quantitative RT-PCR
Liver and bile duct were sampled independently in eachexperimental animal. RNA was isolated using the RNeasy Mini Kit(Qiagen) according to the manufacturer's instructions. RNA of 1 μgwas transcribed to cDNA with the RNA-to-cDNA Kit (AppliedBiosystems). qPCR was performed in technical triplicates on a StepOne Plus cycler (Applied Biosystems) using the Maxima Sybr Green/Rox qPCRmastermix (Fermentas) according to standard protocol. Theprimers CACCAGCGGTAGCGGCGTTAT and TTGCTTGCGTCGGCAAG-TACTGA were employed to detect RRV and the signal was normalizedagainst GAPDH in a second reaction with the primer pair CCCCAG-CAAGGACACTGAGCAAG and TGGTATTCAAGAGAGTAGGGAGGGC.
Relative RNA levels were quantified using the StepOne SoftwareVersion 2.0 (Applied Biosystems) and Excel 2007 (Microsoft).
2. Results
2.1. Biliary atresia model
Mice (n = 297) were infected with 2.5 × 106 pfu RRV within 24 hof life to induce BA.
2.1.1. LethalityPerinatal mortality—defined as death within the first three days
postinfection—resulted in the loss of 28 animals (9.4%). After day 3, 9infected animals (3.0%) died with a majority of 9 animals deceasedfrom days 15 to 17. In the placebo-injected control group (n = 102) 9animals (8.8%) died of perinatal mortality while no animals were lostafter day 3 postinfection.
2.1.2. Clinical symptomsInfected individuals showed a delayed weight development
compared to healthy controls, although results were not significantin this study (Fig. 1A). Three animals (1.0%) cleared the infection asdocumented by a weight gain and loss of clinical symptoms. Themajority of the infected mice developed cholestasis starting from day3 to day 6 postinfection complemented by an oily fur starting fromday 8 to day 11 (Fig. 1B).
2.1.3. Development and incidence of BAMicroscopic and histological evaluation of bile ducts of infected
animals showed the first complete atresia on day 8 (day 8: 1/9).Throughout the following days evidence of BA increased significantly(day 9: 5/10, day 10: 6/10, day 12: 10/14). All infected animalssacrificed after day 14 showed BA (day 14: 44/44, day 17: 8/8, day 21:8/8; Fig. 1C). A subset of the dissected bile ducts was prepared for
Fig. 2. Whole-mount preparation of extrahepatic bile duct stained for the endothelial marker CD31 (red) and RRV vp6 antigen (green). (A) Early time point (day 2 postinfection)shows isolated virus particles. (B) Middle time point during disease development (day 5 postinfection) indicates that the virus is present in high numbers throughout the tissue butthen starts to reduce (day 7 postinfection) (C).
522 C. Oetzmann von Sochaczewski et al. / Journal of Pediatric Surgery 49 (2014) 520–524
electron microcopy to evaluate morphological modifications duringdisease development. As the disease progresses in infected animals,the lumen of the extrahepatic bile duct decreases. At day 14cholangiocytes are ultimately lost when the bile duct becomescompletely obstructed (data not shown).
2.1.4. Localization and kinetics of RRVAn overview picture of RRV spread was obtained by double
staining a second subset of the dissected bile duct whole mounts forviral protein 6 (VP6, green) and the endothelial marker CD31 andsubsequently making 3D reconstructions of stacks of confocal images.Only scattered cells were infected on day 2 postinfection (Fig. 2A). Thepeak of infected cells was observed around day 5 (Fig. 2B), while thenumber of infected cells was reduced at day 7 (Fig. 2C). The onset andkinetics of virus replication were similar in liver and bile ducts up today 7, suggesting a synchronous infection of both organs (Fig. 3).Interestingly, RRV was cleared from liver cells from day 11 on(Fig. 3B), while it replicated in the bile duct at high levels up to day 10and remained detectable at 10–30 times lower levels until day 21(Fig. 3A).
The wall of the extrahepatic bile duct of uninfected animalsconsisted of cholangiocytes (the innermost layer), subepithelialstructures of connective tissue, myofibroblasts, smooth muscle cellsand blood vessels and the outermost mesothelial layer of the serosa(Fig. 4A). Electronmicroscopical analysis demonstrated virus particlesin cholangiocytes on day 5 (data not shown). Interestingly, groups ofRRV particles could be observed in several cells in the subepitheliallayers as identified by location (Fig. 4A–D).
Although identification of the exact cell type is difficult based onlyon electron microscopy without immune labeling, the protrusionsobserved in these cells would indicate that these cells have the abilityto migrate (Fig. 5). Possible cell types could include macrophages,myofibroblasts or stellate cells.
3. Discussion
We closely documented the progressive development of experi-mental BA after its induction by perinatal infection of BALB/c micewith rhesus rotavirus. In the first days, no clinical or histologicalchanges were observed, apart from a few animals that died likely ofthe stress and/or traumatic impact of the infecting procedure. Thisearly lethality was not directly related to the virus infection, as it wassimilar in the placebo control group (9.4% vs. 8.8%, respectively). Thefirst sign of BA was jaundice, which mice developed in the second halfof the first week of life. Throughout this period, histological sections ofthe bile duct revealed increasing stenosis of the lumen of theextrahepatic bile ducts, which, however, remained open, suggesting
that the jaundice was caused by hepatic affection rather than bychanges of the extrahepatic biliary system. BA developed in thesecond week, with the earliest atretic animal detected on day 8 up today 14, at which time all dissected animals showed a completeobstruction of the extrahepatic bile duct (Fig. 1).
It is not clear how exactly the fibrosing stenosis in the extrahepaticbile ducts is triggered in this model. Several authors, who investigatedviral and inflammatory factors in the pathogenesis of BA, predomi-nantly worked with liver tissue samples which are easily available inthe murine experimental model as well as in human patients [11,15].However, less data are available on the changes in the extrahepaticbiliary tissue. To investigate the pathogenesis of the bile ductdestruction in experimental BA, we focused on the kinetics of theRRV infection which we correlated to the different stages of thedisease. The amount of viral RNA increased during the first days inboth liver and bile duct tissue in a similar fashion. However, the viralinfection peaked prior to the clinical development of BA at the end ofthe first week, as seen in both liver and bile duct tissue. Thereafter, thehepatic virus load declined and no viral RNA could be detectedanymore by rtPCR in the liver after day 12. These observations are inaccordance with those of a number of published studies, in which thepeak of replication in both tissues was reported at day 5 to day8 [12,13]. Although the overall kinetics were similar in liver and bileduct tissues up to day 10, reduced amounts of viral RNA persisted inthe extrahepatic biliary system throughout the observation period.This finding suggests that there are differences in the kinetics of RRVinfection between the extrahepatic bile duct and the liver.
Both whole-mount immunohistochemistry and electron micros-copy were used to localize the virus infection at a (sub)cellular level.Using double staining for CD31 and virus antigen VP6, we observed avirus cluster localized in the lumen of the extrahepatic bile duct, withsome spots of viral antigen localized in the wall structure. It is wellknown that RRV has a tropism for cholangiocytes and infection ofcholangiocytes has been postulated as an initial step in theinflammatory reaction in experimental BA [12,13,16]. These findingsare further supported by our observation of viral particle clusterswithin the cholangiocytes at days 5 and 8. Interestingly, viral particleswere also detected in several cells in the subepithelial layers, inaddition to the epithelial layer (Fig. 4). The exact nature of these cellsremains to be identified, but the protrusions observed suggest thatthese cells are migrating cells, most likely macrophages. It is knownthat there is a proliferation of CD68-positivemacrophages in the liversof BA patients [17] and the expression of macrophage-associatedantigens has been correlated to a poor outcome [18]. Moreover,immunohistochemical analysis of tissue exudates of the bile duct wallat the porta hepatis showed a high number of macrophages, but anabsence of lymphocyte infiltration [19]. Although these data are in
Fig. 3. Kinetics of RRV infection evaluated by quantitative PCR of RRV gene VP6 relative to GAPDH expression. Numbers of positive (product detectable in PCR) and negative testedanimals (copy numbers below detection limit) are given on top of each chart. Each data point reflects the average of virus presence in positive animals quantified in the bile duct (A)and liver (B). Error bars indicate standard error.
Fig. 4. Ultrathin sections of extrahepatic bile duct at day 8 using transmission electron microscopy. Black arrows point at RRV particles. (A) Overview picture showing a section of thebile duct with an increasingly obstructed lumen (L) in the center. The lumen is surrounded by cholangiocytes. The cell marked by a black box was selected for higher amplification inpanels B, C and D. (D) Groups of viral particles possibly forming a viroplasm are visible within the cell.
523C. Oetzmann von Sochaczewski et al. / Journal of Pediatric Surgery 49 (2014) 520–524
Fig. 5. Subepithelial cell 5 days post-RRV infection. The black arrow points at RRVparticles. Cell protrusions are indicated by black triangles.
524 C. Oetzmann von Sochaczewski et al. / Journal of Pediatric Surgery 49 (2014) 520–524
contrast to the findings of strong influx of T cells in the livers of miceduring experimental BA, some studies have suggested that also inexperimental BA, macrophages may play an important role: Macro-phages were shown to be susceptible to infection by RRV in tissueculture [20] and have been described to support the inflammatoryreaction in livers of mice subjected to experimental BA [21]. Theappearance of these subepithelial virus-laden cells correlatedwith thedevelopment of stenosis and atresia of the extrahepatic bile ducts inour model, suggesting that these cells could indeed contribute to thedestruction of the extrahepatic bile duct. Based on the well-knowntropism of rotavirus to cholangiocytes, cholangiocytes have been inthe focus of the research regarding the initiation of the inflammatoryresponse in BA, with in vivo and in vitro studies [22–24]. Our resultssuggest that macrophages might also be an interesting candidate andfurther studies should explore the possibilities of alteringmacrophagefunctions in order to alleviate the course of experimental BA.
Despite a number of limitations, such as the inability ofconventional electron microscopy to detect the virus in every stateof its cycle, our study provides a detailed description of the differentstages of development of experimental BA that might be instrumentalin furthering our knowledge of this challenging disease. We clearlydemonstrated in this study that, following a comparable virus peak inliver and extrahepatic bile ducts, the virus is cleared from the liverwhile it remains detectable in the extrahepatic bile ducts, supportingour hypothesis of differences in the viral kinetics in hepatic and biliarytissues. Detection of viral particles in subepithelial cells was notpreviously reported and is in contrast to the observed tropism towardcholangiocytes. These findings could help to better understand the
primary affection of the extrahepatic biliary system and to betterunderstand the pathogenesis of BA.
References
[1] Landing BH. Considerations of the pathogenesis of neonatal hepatitis, biliary andcholedochal cyst—the concept of infantile obstructive cholangiopathy. ProgPediatr Surg 1974;6:113–39.
[2] Hertel P, Estes M. Rotavirus and biliary atresia: can causation be proven?Gastroenterology 2012;28:10–7.
[3] Morecki R, Glaser JH, Cho S, et al. Biliary atresia and retovirus type 3 infection. NEngl J Med 1982;307:481–4.
[4] Richardson SC, Bishop RF, Smith AL. Reovirus serotype 3 infection in infants withextrahepatic biliary atresia or neonatal hepatitis. Gastroenterol Hepatol 1994;9:264–8.
[5] Schukfeh N, Al-Gamrah A, Petersen C, et al. Detection of hepatotropic viruses has noimpact on the prognosis after Kasai procedure. J Pediatr Surg 2012;47:1828–32.
[6] Rauschenfells S, Krassmann M, Al-Masri AN, et al. Incidence of hepatotropicviruses in biliary atresia. Eur J Pediatr 2009;168:469–76.
[7] Tyler KL, Sokol RJ, Oberhaus SM, et al. Detection of reovirus RNA in hepatobiliarytissues from patients with extrahepatic biliary atresia and choledochal cysts.Hepatology 1998;27:1475–82.
[8] Petersen C, Biermanns D, Kuske M, et al. New aspects in a murine model forextrahepatic biliary atresia. J Pediatr Surg 1997;32:1190–5.
[9] Riepenhoff-Talty M, Schaekel K, Clark HF, et al. Group A rotavirus produceextrahepatic biliary obstruction in orally inoculated newborn mice. Pediatr Res1993;33:393–4.
[10] Ramig RF. Pathogenesis of intestinal and systemic rotavirus infection. J Virol2004;78:10213–20.
[11] Riepenhoff-Talty M, Gouvea V, Evans MJ, et al. Detection of group C rotavirus ininfants with extrahepatic biliary atresia. J Infect Dis 1996;174:8–15.
[12] Feng N, Kim B, FenauxM, et al. Role of interferon in homologous and heterologousrotavirus infection in the intestines and extraintestinal organs of suckling mice.J Virol 2008;82:7578–90.
[13] Shivakumar P, Campbell KM, Sabla GE, et al. Obstruction of extrahepatic bile ductsby lymphocytes is regulated by IFN gamma in experimental biliary atresia. J ClinInvest 2004;114:322–9.
[14] Czech-Schmidt G, VerhagenW, Szavay P, et al. Immunological gap in the infectiousanimal model for biliary atresia. J Surg Res 2001;101:62–7.
[15] Allen SR, Jafri M, Donelly B, et al. Effect of rotavirus strain on the murine model ofbiliary atresia. J Virol 2007;81:1671–9.
[16] Feng N, Sen A, Wolf M, et al. Roles of VP4 and NSP1 in determining the distinctivereplication capacities of simian rotavirus RRV and bovine rotavirus UK in themouse biliary tract. J Virol 2011;85:2686–94.
[17] Tracy Jr TF, Dillon P, Fox ES, et al. The inflammatory response in pediatric biliarydisease: macrophage phenotype and distribution. J Pediatr Surg 1996;31:121–5.
[18] Kobayashi H, Puri P, O'Brian DS, et al. Hepatic overexpression of MHC class IIantigens and macrophage-associated antigens (CD68) in patients with biliaryatresia of poor prognosis. J Pediatr Surg 1997;32:590–3.
[19] Schweizer P, Petersen M, Jeszberger N, et al. Immunohistochemical and molecularbiological investigations regarding the pathogenesis of extrahepatic biliaryatresia. (Part 1: immunohistochemical studies). Eur J Pediatr Surg 2003;13:7–15.
[20] Mohanty S, Ivantes C, Mourya R, et al. Macrophages are targeted by rotavirus inexperimental biliary atresia and induce neutrophil chemotaxis via Mip2/Cxcl2.Pediatr Res 2010;67:345–51.
[21] Mack CL, Tucker RM, Sokol RJ, et al. Armed CD4+ Th1 effector cells and activatedmacrophages participate in bile duct injury in murine biliary atresia. Clin Immunol2005;115:200–2009.
[22] Barnes BH, Tucker RM, Wehrmann F, et al. Cholangiocytes as immune modulatorsin rotavirus-induced murine biliary atresia. Liver Int 2009;29:1253–61.
[23] Coots A, Donnelly B, Mohanty SK, et al. Rotavirus infection of human cholangiocytesparallels the murine model of biliary atresia. J Surg Res 2012;177:275–81.
[24] Mohanty SK, Donnelly B, Bondoc A, et al. Rotavirus replication in the cholangiocytemediates the temporal dependence of murine biliary atresia. PLoS One 2013;8:e69069.
