key: cord-0994943-dh3vvr26
authors: FRANCE, MP; SMITH, AL; STEVENSON, R; BARTHOLD, SW
title: Granulomatous peritonitis and pleuritis in interferon‐γ gene knockout mice naturally infected with mouse hepatitis virus
date: 2008-03-10
journal: Aust Vet J
DOI: 10.1111/j.1751-0813.1999.tb13199.x
sha: 6262aa67e8ca9f194ca4e8543f78bb94e990b1d1
doc_id: 994943
cord_uid: dh3vvr26

Objective To investigate a disease outbreak in a colony of laboratory mice with targeted disruption of the gene for interferon‐γ FORMAT: A case report based on necropsy, histopathology, serology and immunohistochemistry. RESULTS: Affected mice exhibited depression and variable ascites. Necropsy revealed a granulomatous peritonitis and pleuritis with extensive adhesions although parenchymal lesions were minimal. Serum samples had high concentrations of antibody to mouse hepatitis virus and immunohistochemical examination revealed the presence of mouse hepatitis virus antigen in granuloma macrophages. Serological testing for other infectious agents and bacterial culture were negative and wild type mice kept in the same facility remained healthy. Despite the association between the disease and mouse hepatitis virus infection, the precise role played by mouse hepatitis virus was not determined. While the disease is superficially similar to feline infectious peritonitis (another coronavirus‐induced serositis), differences exist between the histopathological findings in these two conditions. CONCLUSION: This unusual disease process illustrates how new diagnostic challenges can arise in novel mouse genotypes created through molecular genetics. Furthermore, the association between the disease and mouse hepatitis virus illustrates the importance of maintaining laboratory animals under specific‐pathogen free conditions.

In recent years, the range of mice available to biomedical re s e a rchers has expanded greatly through the application of molecular genetics. In part i c u l a r, the development of mice in which a single, targeted gene has been rendered non-functional (gene knockout mice) has resulted in the creation of seve r a l h u n d red novel genotypes. 1 4 Many such mice have been deve loped for immunological studies and often the resulting phenotypes exhibit a degree of immunodeficiency. This situation raises the prospect of atypical infectious disease problems due to peculiar host responses or opportunistic infections. This report describes the investigation of an outbreak of an unusual granulomatous peritonitis and pleuritis associated with natural MHV infection in knockout mice lacking a functional gene for the i m p o rtant immunomodulatory cytokine IFN-γ. The findings draw attention to both the emerging challenges of disease diagnosis in animals with novel genotypes and the importance of maintaining SPF husbandry conditions when working with these valuable animals.

Mice rendered deficient in IFN-γ through targeted disruption of the IFN-γ gene (IFN-γ knockout mice) and backcrossed to the C57BL/6 strain we re derived from stock originally produced by Dalton et al. 15 The main breeding colony was kept under SPF conditions. A second breeding and experimental colony was then established in a separate facility where mice we re kept in open top cages under 'c o n ve n t i o n a l' conditions. While regular health monitoring was not undertaken in this f a c i l i t y, MHV, ro t a v i rus, murine encephalomyelitis viru s , Sp i ronucleus muris, Gi a rdia muris and pinworm had been detected there in the past. Despite this, all mice apart from the IFN-γ knockouts appeared clinically normal. Food (commercial rodent pellets), water, bedding (vermiculite) and caging we re unsterilised. None of the mice examined in this inve s t i g a t i o n had been used for experimental work.

Serum samples taken from affected IFN-γ knockout mice at necropsy were frozen at -70°C then shipped overnight in ice to the Murine Virus Monitoring Service, Adelaide, South Australia for ELISA testing. ELISA antigen was obtained from the Un i versity of Missouri, Columbia, USA, or was manufacture d in-house as follows. Permissive cell lines were infected with virus (strain A59 in the case of MHV) and after 72 h (approximately

To investigate a disease outbreak in a colony of laboratory mice with targeted disruption of the gene for interferon-γ

A case report based on necropsy, histopathology, serology and immunohistochemistry.

A ffected mice exhibited depression and variable ascites. Necropsy revealed a granulomatous peritonitis and pleuritis with extensive adhesions although parenchymal lesions were minimal. Serum samples had high concentrations of antibody to mouse hepatitis virus and immunohistochemical examination revealed the presence of mouse hepatitis virus antigen in granuloma macrophages. Sero-logical testing for other infectious agents and bacterial culture were negative and wild type mice kept in the same facility remained healthy. Despite the association between the disease and mouse hepatitis virus infection, the precise role played by mouse hepatitis virus was not determined. While the disease is superficially similar to feline infectious peritonitis (another coronavirus-induced serositis), differences exist between the histopathological findings in these two condition s .

This unusual disease process illustrates how new diagnostic challenges can arise in novel mouse genotypes created through molecular genetics. Furthermore, the association between the disease and mouse hepatitis virus illustrates the importance of maintaining laboratory animals under specific-pathogen free conditions. M HV is a coro n a v i rus distinguishable into many strains and has a high rate of mutation. 1 , 2 It is widely re g a rded as the most important pathogen of laborat o ry mice due to its high pre va l e n c e 3 -5 and its ability to interf e re with both in-vivo and in-vitro re s e a rc h . 1 , 6 -9 M H V-related effects on re s e a rch may be insidious since natural infection is usually subclinical in immunocompetent hosts. 2 , 6 , 8 In immuno-deficient hosts, howe ve r, infection is often fatal and re p o rts continue to appear in which valuable colonies have been deva s t a t e d f o l l owing incursion by MHV. 1 0 -1 2 Indeed, a seve re disease consistent with MHV infection greatly impeded early pro p a g ation of the athymic (nude) mouse, now probably the most widely used immunodeficient mouse genotype in biomedical re s e a rc h . 1 

90% cytopathic effect) the cells we re lysed by fre eze -t h a w i n g ; cell debris was pelleted at 3000 g for 20 min and viru scontaining supernatant was concentrated by ultracentrifugation then purified by ultracentrifugation on a 20% sucrose cushion. Negative control antigen was prepared likewise from uninfected cells. Antigen concentration was determined by titration against k n own positive and negative sera. Antigens we re diluted in bicarbonate buffer (pH 9.8) and coated onto 96-well micro t itration plates at 4°C overnight and then either stored at -80°C or washed four times with phosphate-buffered saline containing 0.05% Tween-20 for immediate use. Test and control sera were diluted 1:100 in phosphate-buffered saline containing 0.5% low-fat skim milk powder and 100 µL added to coated microtitration plate wells. After incubation and washing, 100 µL of p reviously titrated affinity isolated goat anti-mouse IgG and IgM-horseradish peroxidase conjugate (Biosource International, CA, USA) was added to each well. After re-incubation and washing, 100 µL of K-Blue substrate (Graphic Scientific, Queensland) was added to each well, the reaction stopped after 10 min with 2M H 2 SO 4 and the plates read at 450 nm on an automated plate re a d e r. Corrected optical density values we re c o m p a red to values from positive and negative re f e rence sera (BioReliance, USA ) and the results expressed as a percentage of reference sera values. Equivocal (15 to 30% ) and positive (> 30%) sera we re then re-tested by the immunofluorescence antibody method using slides provided by the Un i versity of Mi s s o u r i , Columbia, USA, or pre p a red with air-dried, acetone-fixe d , v i rus-infected cells. Test and control sera we re diluted 1:10 in phosphate-buffered saline and 15 µL added to appropriate wells on the slide. After incubation, washing and air drying, 15 µL of p reviously titrated goat anti-mouse IgG and Ig M -f l u o re s c e i n conjugate (Bi o s o u rce International, CA, USA) was added to each well. After re-incubation and washing, the slides were glycerol mounted and examined by fluorescence microscopy.

Mice were euthanased with CO 2 and necropsied immediately. Ascitic fluid was collected from five mice and cytocentrifuge p reparations made then stained with Di f f -Quick (Lab-Aids, New South Wales). Swabs of peritoneal lesions and exudate from four mice were cultured aerobically on blood agar. Tissues we re fixed in 10% neutral-buffered formalin, pro c e s s e d routinely, embedded in paraffin, sectioned at 4 µm and stained with haematoxylin and eosin. Selected sections were also stained with Wa rt h i n -St a r ry silve r, Brown and Brenn, Gro c o t t's methenamine silve r, periodic acid-Schiff, Ziehl Neelsen and Giemsa methods. Immunohistochemical staining for MHV was p e rformed on 5 µm trypsinised sections of formalin fixe d , paraffin-embedded tissue mounted on silane-coated slides. Sections we re incubated with polyvalent mouse ascitic fluid containing anti-MHV antibody then labelled using biotinylated goat anti-mouse antibody followed by a stre p t a v i d i nhorseradish peroxidase conjugate and diaminobenzidine substrate. 16, 17 

Clinical signs were observed in IFN-γ knockout mice of both sexes (including previously bred females) ranging in age from 10 to 24 weeks. C57BL/6 mice housed in adjacent cages remained normal. Affected animals exhibited depression and va r i a b l e ascites, and some also developed mild diarrhoea. De t a i l e d morbidity figures were not available although a total of 10 mice, all similarly affected, we re submitted for necropsy over a period of 3 weeks, after which all remaining IFN-γ knockout mice we re euthanased. Clinical signs we re not observed in IFN-γ k no c k o u t mice kept under SPF conditions in the main breeding colony.

Ne c ropsy re vealed a variable amount of cloudy peritoneal fluid and thick, white plaques on the serosae of abdominal viscera. Many of these plaques had formed ru b b e ry adhesions between adjacent structures and in advanced cases had virtually formed a cocoon around groups of organs (Fi g u re 1). T h e spleen was often the organ most affected and was always m a rkedly enlarged. Adhesions and effusion we re also found in the thoracic cavity in some animals. Culture of abdominal fluid f rom four mice resulted in no growth after 5 days. Howe ve r, cytological examination of fluid from one mouse which was not c u l t u red re vealed small coccoid and occasional ro d -s h a p e d bacteria within neutrophil cytoplasm. In this specimen, approximately 90% of the cells were neutrophils. This was in contrast to the findings in smears from other mice in which the majority of cells we re macrophages and lymphocytes and no bacteria were seen.

Histological examination revealed that the serosal plaques and adhesions consisted of sheets of macrophages forming granulomas, which in some mice had a necrotic centre. Some adhesions consisted only of a thin layer of fibrous tissue. Many granulomas we re heavily infiltrated with neutrophils, especially at the periphery, and variable numbers of lymphocytes and plasma cells we re scattered through the granulomas and surro u n d i n g c o n n e c t i ve tissue. Multinucleate giant cells we re ve ry rare. A striking feature of the inflammatory process was that it was almost entirely confined to the serosa and adjacent folds of peritoneum or pleura (Figure 2 ). Solitary parenchymal lesions were seen in the liver and kidney of one mouse although these were subcapsular and appeared to be extensions of serosal lesions. The only organ exhibiting substantial parenchymal damage was the pancreas although it was only affected in advanced cases and a p p e a red to be the result of extension of inflammation fro m adjacent mesentery into the pancreatic interstitium. Occasional granulomas we re present in visceral lymph nodes and splenic enlargement was found to be due to intense extramedullary m yelopoiesis. Dilated lymphatic vessels we re prominent adjacent to some foci of inflammation. Oc c a s i o n a l l y, near areas of intense inflammation, small muscular arteries showed a localised vasculitis or hyaline degeneration. In general, however, blood vessels appeared normal. There were no significant findings in sections of limb joints, brain or other stru c t u res of the head. No bacteria or fungi we re seen in sections stained for micro-organisms.

Sera from two affected mice gave strong positive re a c t i o n s when tested for antibodies to MHV but we re negative to 13 other agents (murine cytomegalov i rus, mouse parvov i ru s e s , Mycoplasma pulmonis, lymphocytic choriomeningitis, pneumonia virus of mice, reovirus 3, Sendai virus, Theiler's murine e n c e p h a l o m yelitis virus, mouse ro t a v i rus, ectromelia viru s , mouse adenovirus types 1 and 2, polyoma virus and Clostridium piliforme, the agent of Tyzzer's disease). Sera from a further eight affected mice we re then tested and again found to be stro n g l y p o s i t i ve to MHV and negative to a reduced panel of agents (murine cytomegalov i rus, mouse parvov i ruses, Myc o p l a s m a p u l m o n i s and lymphocytic choriomeningitis). Im m u n ohistochemical staining demonstrated MHV antigen in the c e n t re of granulomas, most of which appeared to be within macrophages (Figure 3 ).

In recent years, the use of transgenic and gene knockout mice has re volutionised the study of gene function in vivo. Applications have ranged from the study of basic cellular p rocesses right through to the creation of homologues of i m p o rtant human and animal diseases. Many of these genotypes, howe ve r, have yet to be well characterised and often the resulting phenotypes have produced surprising results which challenge prior understanding. 1 4 It is to be expected, there f o re , that some of these animals will respond to commonplace infections in novel ways.

The present case describes a striking granulomatous disease of the peritoneum and pleura in IFN-γ gene knockout mice and several findings indicate that MHV was important in its pathogenesis. These include appearance of the disease following introduction of the mice to an MHV-endemic facility in open top cages, seropositivity to MHV, and the demonstration of viral antigen in granuloma macrophages by immunohistochemistry.

Published reports of MHV-related lesions in the peritoneum and pleura are rare although early accounts noted the accumulation of giant cells in these tissues, 1 8 and ascites and peritonitis due to MHV has been described in nude mice. 1 9 , 2 0 Fu rt h e rm o re, the experience of one of the present authors (SWB) suggests that such lesions are more common than is reported. In all these cases, howe ve r, the peritoneal and pleural changes are accompanied by substantial hepatic parenchymal lesions, unlike in the present case. More recently, a condition almost identical to the present case has been reported in two separate situations w h e re IFN-γ deficient gene knockout mice we re exposed to M H V. In the first, a granulomatous pleuro-peritonitis deve loped following natural exposure in IFN-γ/interleukin 10 double knockout mice, although interestingly IFN-γ s i n g l e knockout mice were not affected. 21 In the second report, IFN-γ knockout mice developed ascites, peritonitis and pleuritis with granulomatous inflammation following experimental inoculation of MHV by the intraperitoneal route. 22 While the findings of these two re p o rts suggest that pleuroperitonitis can be t h e major lesion induced by MHV in animals deficient in IFN-γ, other studies show that this is not always the case. In part i c u l a r, s e veral studies with gene knockout mice deficient in va r i o u s components of the IFN-γ pathway (including the IFN-γ gene) re p o rted no significant serosal lesions despite an incre a s e d susceptibility to parenchymal damage by MHV 1 1 , 2 3 -2 5 a n d similar findings were made in earlier studies using mice treated with anti-IFN-γ antibodies. 26, 27 It is possible that in the present case and in the other two with prominent pleuroperitonitis that factors such as age of host and genetic background and the dose and strain of virus contributed to the pathogenesis since all these can affect the outcome of MHV infection. 1, 2, 6 Even so, it is curious that some of the hallmark lesions of MHV-re l a t e d disease such as epithelial and endothelial syncytia were lacking. Route of inoculation can also influence disease manifestation, although this does not seem to be important in establishing peritonitis since natural exposure and intraperitoneal ro u t e s both led to a similar distribution of lesions. Overall it seems likely that factors in addition to IFN-γ deficiency and simple e x p o s u re to MHV are re q u i red to produce the sort of disease seen here.

The search for additional infectious agents in this inve s t i g ation was limited. In part i c u l a r, bacteriological culture did not include suitable techniques for the detection of myc o b a c t e r i a which are important causes of granulomatous inflammation Scientific and can be difficult to demonstrate histologically. Anaero b i c and other fastidious organisms would also have gone undetected with the culture methods used. While bacteria we re seen in smears of abdominal fluid from one mouse, that particular fluid sample was not cultured. Moreover, the very high proportion of n e u t rophils in that sample compared to samples from other mice suggests that a secondary septic peritonitis might have been present. This is further supported by the fact that despite large numbers of organisms in smears of fluid, no bacteria were seen in sections of solid granulomas. Vi rus isolation was not attempted so it is unknown whether the viral antigen seen immunohistochemically represented infectious virus particles or components of degenerate virions. Se rological testing was largely aimed at detecting antibodies to those murine viru s e s which have been found to affect macrophage function although testing was not available for lactate dehyd rogenase eleva t i n g v i rus, an agent which can replicate in peritoneal macro p h a g e s and has subtle but complex effects on the host. 6 The possibility that other, undetected infectious agents contributed to this disease therefore cannot be excluded.

The extent to which the granulomatous inflammation was restricted to the peritoneal and pleural serosae is one of the most striking features of this disease. This finding may relate in part to the IFN-γ deficient state of the animals. The peritoneal cavity contains resident and blood monocyte-derived macro p h a g e s which form an important line of defense. 28, 29 Since IFN-γ plays an important role in the activation of macrophages in the peritoneum and elsew h e re , 3 0 i m p a i red activation resulting fro m I F N -γ-deficiency could predispose to granuloma formation. The presence of MHV might then exacerbate this situation since MHV itself can affect peritoneal macrophage function 8, 31 and it has been found that peritoneal macrophage numbers i n c rease during natural MHV infection. 3 1 Loss of autocrine synthesis of IFN-γ by peritoneal macrophages resulting in an inability to induce an antiviral state 32 might further contribute to this proposed sequence of events.

In their discussion of MHV-induced pleuroperitonitis in IFN-γ knockout mice, Kyuwa et al suggest the possible involvement of B-1 cells. 22 B-1 cells are a class of B lymphocyte which is enriched in the peritoneal and pleural cavities and evidence that IFN-γ plays a role in the regulation of B-1 cells 3 3 adds to the appeal of their hypothesis. Further studies into this disease may there f o re be of value in the investigation of general immune mechanisms in the peritoneal and pleural cavities.

The disease in these mice bears a superficial similarity to another coronaviral disease, feline infectious peritonitis. Specific a l l y, serosal inflammation and peritoneal and sometimes pleural effusions can be a major feature of FIP. In addition, different strains of both murine and feline coronaviruses tend to p roduce either enteric or systemic disease, and replication in m a c rophages is important to the pathogenesis of infection in both host species. 1, 2, 34 There are, however, some morphological differences between the two. In typical FIP, effusions are associated with vascular and perivascular lesions which we re not a major feature in the present case. Furthermore, the solid lesions which predominate in the so-called 'd ry' form of FIP (often loosely termed granulomas) are distinct from the 'classic' granulomas seen here and are not exc l u s i vely subserosal in distribution. Thus while the findings in the present case suggest that the role of IFN-γ in FIP may warrant investigation, it seems likely that some differences exist in the pathogenesis of these two diseases.

Re g a rdless of the precise pathogenesis of this condition, the clear invo l vement of MHV serves to illustrate the sort of problem which can arise when laboratory animals are not kept under appropriate standards of SPF husbandry. As in this case, e f f e c t i ve management of such problems may re q u i re undertaking the costly process of depopulation and restocking. T h e a p p ropriate use of microisolator caging combined with re g u l a r health monitoring can greatly reduce the likelihood of incurring such costs. 3

Brown AT, Gregory AR, Ellis TM and Hearnden MN. Comparative immunogenicity of two bivalent botulinum vaccines. Aust Vet J 1999;77:388-391.

Tables 1 and 2 should read as follows: 

The cellular and molecular pathogenesis of coronaviruses

Enterotropic mouse hepatitis virus

Health care for research animals is essential and affordable

Seromonitoring in small laboratory animal colonies. A five year survey: 1984-1988

Annual health report -serology of Australian rodent colonies

Mouse hepatitis virus

Infection of BALB/cByJ mice with the JHM strain of mouse hepatitis virus alters in vitro splenic T cell proliferation and cytokine production

Peritoneal macrophage alterations caused by naturally occurring mouse hepatitis virus

In vitro and long-term in vivo immune dysfunction after infection of BALB/c mice with mouse hepatitis virus strain A59

Epizootic of low-virulence hepatotropic murine hepatitis virus in a nude mice breeding colony in Taiwan

Targeted disruption of the mouse Stat1 gene results in compromised innate immunity to viral disease

Use of polymerase chain reaction to diagnose a natural outbreak of mouse hepatitis virus infection in nude mice

Nude', a new hairless gene with pleiotropic effects in the mouse

Cytokine knockouts

Multiple defects of immune cell function in mice with disrupted interferon-γ genes

Mouse hepatitis virus immunofluorescence in formalin-or Bouin's-fixed tissues using trypsin digestion

Staining method for detecting concurrent viral infections in animals with Pneumocystis carinii

Mouse hepatitis virus and its pathogenic action

Naturally occurring mouse hepatitis virus infection in the nude mouse

Ascitic disease in ICR-nude mice due to mouse hepatitis virus

Idiopathic granulomas in IFN-γ-/-, IL-10-/-double knockout mice

Murine coronavirus-induced subacute fatal peritonitis in C57BL/6 mice deficient in gamma interferon

Exacerbated viral hepatitis in IFN-γ receptor deficient mice is not suppressed by IL-12

tory macrophages do not die locally, but emigrate to the draining lymph nodes

Interferon-γ: biology and role in pathogenesis

Effect of inapparent murine hepatitis virus infections on macrophages and host resistance

IL-12 induces IFN-γ expression and secretion in mouse peritoneal macrophages

Continuous anti-interleukin 10 antibody administration depletes mice of Ly-1 B cells but not conventional B cells

Coronavirus infection in cats

Disassociation between the in vitro and in vivo effects of nitric oxide on a neurotropic murine coronavirus

The role of gamma interferon in infection of susceptible mice with murine coronavirus, MHV-JHM

In vivo depletion of interferon-gamma leads to susceptibility of A/J mice to mouse hepatitis virus 3 infection

Maintenance of peritoneal macrophages in the steady state

In vivo fate of the inflammatory macrophage during the resolution of inflammation

The excellent technical assistance of Karen Wa d well, El a i n e C h ew, Meira Bosnic and Trish Ma rtin (De p a rtment of Ve t e r in a ry Anatomy and Pathology) and Elizabeth Johnson (Ya l e School of Medicine, USA) in the investigation of this case is gratefully acknowledged.