Coxsackievirus B4 infection of the mouse pancreas: acute and persistent infection

Transcription

1 Journal of General Virology (1992), 73, Printed in Great Britain 1387 Coxsackievirus B4 infection of the mouse pancreas: acute and persistent infection C. Vella,a*1 " C. L. Brown 2 and D. A. McCarthy 3 I Department of Immunology, The London Hospital Medical College, Turner Street, London El, 2Institute of Pathology, London Hospital, London E1 and 3School of Biological Scienees, Queen Mary and Westfield College, Mile End Road, London El, U.K. The course of infection of a pancreas-adapted isolate of coxsackievirus B4 was followed over a 10 month period in a murine model. Following intraperitoneal inoculation a typical acute infection was seen in nine of 10 inbred mouse strains. Virus rapidly infected the exocrine pancreas, titres peaking 3 to 4 days postinfection (p.i.). Lesions were almost exclusively confined to pancreatic acinar cells and varied in severity among the inbred strains. Virus shed into the bloodstream was not cell-associated. Evidence of persistent infection was found in nine mouse strains and infective virus was recovered from the pancreas of seven strains for up to 10 months p.i. Approximately 28% of pancreases examined beyond the acute phase showed focal inflammation and 22% showed focal necrosis (cell death). Virus was occasionally recovered from other organs (heart, liver and spleen), but lesions were rarely seen. Virus-specific antigen was localized to small groups of pancreatic acinar cells using an indirect immunogold silver staining technique. These observations suggested that the virus persists in pancreatic tissues because it seems unlikely that virus disseminated from distant sites would cause such localized infection. In three of these strains, the course of infection may have been influenced by superinfection with mouse hepatitis virus. Introduction Coxsackievirus B4 (CVB4) is an enterovirus of the family Picornaviridae. Since their isolation (Melnick et al., 1949), the group B viruses (B1 to B6) have been implicated in a wide spectrum of clinically distinct, acute and chronic disease syndromes (for a recent account see Bendinelli & Friedman, 1988). However, evidence linking these viruses to chronic disease has been largely seroepidemiological and persistent infection demonstrable only in immunodeficient animals (Schnurr & Schmidt, 1984; Schnurr et al., 1984). More recently, enterovirus-specific cdna probes have been used to detect RNA sequences in human heart tissue in situ (Easton & Eglin, 1988) and in RNA extracted from biopsy samples from patients with chronic heart or muscle disease (Bowles et al., 1986, 1987; Archard et al., 1987; Cunningham et al., 1990). In addition, it has been shown that these viruses can persist in the tissues of normal adult mice (Bocharov & Shalaurova, 1984; Guriunova & Savinov, 1987). t Present address: National Institute for Biological Standards and Control, Blanche Lane, South Mimms, Potters Bar, Hertfordshire EN6 3QG, U.K. This paper presents the first report of a persistent infection by an isolate of CVB4 (designated CVB4-P15) adapted to grow in the mouse pancreas. Data were collected during a 10 month study of a murine model of CVB4-induced pancreatitis using standard virological and histological techniques. These included the recovery of infective virus from the pancreas and other tissues, and the demonstration of new lesions and CVB4-specific antigen in the exocrine pancreas late after the initial infection. Methods Generalprocedures. The origins of the virus isolate and mouse strains used in this study, together with the procedures employed for Vero cell culture, infection, necropsy, virus isolation, infectivity assays and the quantification of neutralizing antibody are given in Vella & Festenstein (1992). Infectious centre assay (ICA). Blood was collected from the heart of infected or uninfected control anaesthetized mice 2 days post-infection (p.i.) into 2 ml syringes previously rinsed with heparin (1000 units/ml; C P Pharmaceuticals). It was immediately diluted 1:4 in RPMI 1640 medium containing 1% heparin, layered onto 2 ml Lymphosep (SeraLab) in 10 ml screw cap conical test tubes (Falcon), and centrifuged at 959 g for 25 min at room temperature. Buffy coat peripheral blood lymphocytes (PBLs) and plasma were collected SGM

2 1388 C. Vella, C. L. Brown and D. A. McCarthy separately and assayed for infectivity. Spleens were aseptically removed and single-cell suspensions prepared in medium. After two washes by centrifugation in RPM11640 (124g for 10 min), cells (PBL or spleen) were divided into two portions; one was incubated in excess monkey anti-cvb4 serum (dilution 1 : 100) for 4 h at 37 C, the other in RPM11640 alone. After three washes in RPMI t640 to remove the antiserum, the cell concentration was determined and 500 l-tl aliquots containing 10, 50 or 100 cells were transferred in duplicate to Vero cell monolayers in six-weu Linbro plates. A plasma control for infective virus was included in duplicate for each animal (final dilution 1 : 100). Duplicate samples were assayed after one cycle of freezing and thawing. Immunogold silver staining ( IG SS) of tissue sections. The protocol used was basically as described by the manufacturers (Janssen) and included pretreatment with Lugol's iodine (1% iodine in 2% aqueous potassium iodide). Briefly, following a 10 rain block with undiluted normal goat serum sections were incubated with 30 Ixl of primary antibody (monkey anti-cvb4 diluted 1 : 250, a kind gift from P.H.L.S. Colindale, U.K. or F4/80 diluted 1:40, a mature macrophage marker, a kind gift from Dr S. Gordon, Oxford, U.K.) for 1 h at room temperature and washed three times with PBS. Sections were then incubated for a further 1 h with the appropriate gold (5 nm)-conjugated second antibody (Janssen anti-human or anti-rat at 1 : 120) and washed extensively, first in PBS and then in Milli-q H~O (Millipore). Silver enhancement with Janssen Intense II was monitored by light microscopy. Indirect immunofluoreseent ( IFA ) staining of cryoseetions. The method used has been reported in detail (Vella & Festenstein, 1992), but briefly, unfixed 7 ~tm pancreas cryosections were blocked with 20% normal sheep serum, incubated for 40 rain with monkey anti-cvb4 antibody as described above, washed and incubated for a further 40 min with fluorescein isothiocyanate-conjugated sheep anti-human Ig (WeUcome) at a dilution of 1:30. After washing, sections were counterstained with propidium iodide (0-1%), washed, mounted in Uvinert (Gurr) and examined under a Zeiss fluorescence microscope. Results and Discussion Virus growth and dissemination during acute infection At 1 to 8 days p.i. virus titres in homogenates of liver and heart were low in most of the strains tested (< 2.0 log l0 p.f.u./g wet tissue); however, titres in BALB/c mouse liver homogenates 3 days p.i. were high enough (1 106 to 7 x 106 p.f.u./g wet tissue) to assume that this organ supported virus replication. Virus could not be recovered from cell-free spleen homogenates from any strain between 1 and 7 days p.i. (results not shown), although small amounts of virus were occasionally found by ICA (see below). All animals of all strains were viraemic 2 to 4 days p.i. (1 x 104 to 10 x 104 p.f.u./ml) (results not shown). Results for virus growth in the pancreas during the acute phase of infection (1 to 8 days) are presented in Vella & Festenstein (1992), but are summarized briefly here. Between 1 and 5 days p.i. CVB4-P15 could be detected in the pancreas of all strains of mice, other than C3H/HeJ, which had been inoculated with 10 z to 104 p.f.u./mouse. Pancreatic virus titres peaked 3 to 4 days p.i. in most strains, after which titres fell; virus could still be isolated from the pancreas of six strains 7 to 8 days p.i. (DBA/1, SWR/J, BALB/c, CBA/H-T6, C3H/Heola and C57BL/ksdbm) (Vella & Festenstein, 1992). Virus could not be detected in PBLs, spleen cells or lysates from seven NIH or five CBA/H-T6 mice 2 days p.i. by using ICA. However, small amounts of virus (which produced between one and three plaques per well) were detected occasionally (one of seven mice tested in each instance) in PBLs that had not been treated with monkey anti-cvb4 antibody and in treated spleen cell lysates from C3H/Heola mice. Similarly one of five BALB/c mouse spleen cell lysates that had not been incubated with the antiserum produced three virus plaques. Virus was present at high concentration (100% c.p.e, when plated at a dilution of 1 : 100) in comparable plasma samples from all four strains of mice. These results suggest that following intraperitoneal inoculation, CVB4-P 15 rapidly infects and replicates in the pancreas. Virus shed into the bloodstream circulates mostly free in plasma and, except for that in BALB/c mice, bloodborne virus does not subsequently undergo significant replication in non-pancreatic tissues. In agreement with these results, Gauntt et al. (1979) have reported that spleen cells do not support the replication of myocarditic or amyocarditic variants of CVB3 in vitro. Virus recovery at later stages of infection Infective virus was detected between 1 and 10 months p.i. in plasma or various organs from seven of the 10 strains examined (Table 1). In most instances the virus titre was low (three to 15 plaques at a 0 dilution, approximately 102 p.f.u./g wet tissue), but in several the titre ranged between 3.8 x 103 and p.f.u./g wet tissue. Of 45 infected organs, 31 were pancreases, but in two strains, C57BL/6Jdbm and C57BL/ksdbm, other organs were sampled only rarely. A single SJL/J mouse serum sample was virus-positive (5 x 104 p.f.u./ml) at 3 months p.i. These results show that CVB4-P 15 persists in the tissues of seven mouse strains. In most cases, the pancreas appears to be the target organ, but unlike acute phase samples, virus can be occasionally isolated from the spleen. Virus was not recovered from the tissues of C3H/HeJ, NIH or CBA/H-T6 mice beyond the acute phase (not shown). Replicate tissues from all strains were examined for histological evidence of infection. Histological changes on infection During the acute phase of infection (1 to 8 days), virusinduced lesions were found in the exocrine pancreas of all strains inoculated with 10 z to 105 p.f.u. CVB4-P15 (Table 2). However, infiltrating inflammatory cells

3 Persistent CVB4 infection of mouse pancreas 1389 Table 1. Detection of virus in various organs between 10 days and 10 months after infection* Time after infectiont Strain Organ n~ 10 to 27 days 1 to 3 months 4 to 7 months 8 to 10 months Total number of organs with a titre i>1 xl04 p.f.u./g DBA/1 6 Pancreas 2 : 2 0 : 2 1 : 2 ND Heart 1:2 0:2 0:2 ND Liver 1:2 0:2 0:2 ND Spleen ND 0 : 2 0 : 2 ND SWR/J 30 Pancreas 1:2 1:5 5:9 4 2 Heart 1:2 1: Liver 1:2 0:5 0:10 0:14 Spleen 0:2 1:5 3:10 0:14 BALB/c 17 Pancreas 0 : 3 2 : 3 3 : 11 ND Heart 0:3 0:3 0:11 ND Liver 0:3 1:3 0:11 ND Spleen 0 : 3 0 : 3 1 : 11 ND C3H/Heola 10 Pancreas 1 : 1 1 : 4 3 : 5 ND Heart 0:4 0:4 ND Liver 0:1 0:4 0:4 ND Spleen 0:1 0:4 0:4 ND SJL/JII 10 Pancreas 0:1 1:3 0:3 0:4 Heart 0:1 0:3 0:3 0:4 Liver 0:1 0:3 0:3 0:4 Spleen 0:1 0:3 0:3 1:2 C57BL/6Jdbm 11 Pancreas ND 2 : 7 1 : 2 1 : 2 Heart ~ ND ND 1 : 2 Liver ND ND ND ND Spleen No ND ND ~D C57BL/ksdbm 11 Pancreas ND 1:4 1:5 1:2 Heart rid ND ND ND Liver ~ ND ND ND Spleen No ND ND ND * Mice were inoculated with 102 p.f.u. CVB4-P15, except strains C57BL/6Jdbm and C57BL/ksdbm which received 104 p.f.u. t First figure indicates the number of organs from which virus was recovered, the second figure is the total number of organs sampled at that time. For each strain the number of organs with a virus titre >/l x 104 p.f.u./g wet tissue is shown in the last column, all other organs having a titre of approximately 1 x 102 p.f.u./g wet tissue (one to 15 plaques at a 0 dilution of organ homogenate). ~: n, Number of animals tested surviving at 10 days p.i. ND, Not done. II One SJL/J mouse serum sample had a titre of 5 x 104 p.f.u./ml at 3 months p.i. Virus could not be recovered from the serum of any other strain nor from other SJL/J mice beyond 4 days p.i. persisted in several strains for at least 3 weeks p.i. and, between 3 and 6 weeks, dead acinar tissue was replaced by adipose tissue. Therefore, for the purpose of describing histological changes, the time course was divided into three phases, acute (1 day to 3 weeks p.i.), convalescent (3 weeks to 6 weeks p.i.) and a late or persistently infected phase, 6 weeks to 10 months p.i. These findings are illustrated in Fig. 1 to 10, and summarized in Table 2. The extent of CVB4-P15-induced pancreatic damage varied among the strains studied from frank necrosis with interstitial oedema (compare Fig. 1 with Fig. 2) to focal pancreatitis (e.g. Fig. 3). Acute coagulative necrosis of the exocrine tissues was also seen in some strains (Table 2). In such cases the cytopathology was a direct effect of lytic virus infection. Indeed, for DBA/1 mice rapid lysis of acinar tissue with the release of pancreatic enzymes resulted in the autolysis of abdominal fat. For strain C3H/HeJ mice a virus concentration of 105 p.f.u, was required to induce focal pancreatitis. Occasional small foci of inflammatory cells were also seen in sections of liver but not of heart or spleen from several strains during the acute phase (results not

4 1390 C. Vella, C. L. Brown and D. A. McCarthy Table 2. Histological findings following CVB4-P15 infection CVB4-P15-induced pancreatic pathology at various times p.i.* Strain Acute Convalescent Late inoculumt (1 to 20 days) (3 to 6 weeks) (6 weeks to 10 months) SWR/J Frank to coagulative necrosis Inflammation (2) 102 of 60 to 80% acinar tissue (19) :~ Healed (2) New lesions (2) Myocarditis (1) n 19:1: n 7 CBA/H-T6 Frank to coagulative necrosis Inflammation, new lesions (2) 102 of 60 to 80% acinar tissue (6) n6 C3H/Heola Frank necrosis of ND t02 60 to 80% acinar tissue (7) n7 BALB/c Frank to coagulative necrosis 102 of 60 to 80% of acinar tissue (10) No apparent damage (1) nil DBA/I Frank to coagulative necrosis 102 of 100% of acinar tissue (6) Severe fat necrosis with calcification n6 C57BL/ksdbm Mild to frank necrosis of up 104 to 80% of acinar tissue (14) No apparent damage (1) n 15 SJL/J Diffuse necrosis (4) 10 z Coagulative necrosis (1) n5 NIH 102 Pancreatitis No apparent damage (16) (1) n17 C57BL/6Jdbm Pancreatitis with small 104 foci of necrosis (18) hi8 C3H/HeJ Pancreatitis with very l0 s little necrosis (4) n4 Inflammation and areas of degranulated and swollen acinar cells (2) New lesion (1) Healed (1) Inflammation and areas of degranulated and swollen acinar cells (1) No signs of earlier damage (1) No evidence of earlier damage (1) Healed (8) Inflammation (8) Recent necrosis (4) Lymphoma (1) Myocarditis (1) No evidence of prior infection (2) 4 Inflammation new lesion (1) Inflammation (1) Healed (1) Healed (1) Inflammation (1) New lesion (1) Healed (3) Inflammation (3) New lesions (3) n9 Healed (1) Massive MNC associated with calcified areas (1) Healed (2) Insulitis (1) No evidence of earlier damage (1) Healed (2) Healed (2) Inflammation (1) Inflammation (1) Healed (2) Healed (4) n5 Healed (2) Healed (3) New lesions (2) New lesion Healing endocarditis (2) (1) n4 n6 Healed (3) ND * Lesions were occasionally found in other organs as shown. i" Inoculum, p.f.u./mouse. :~ Values of n are the number of animals studied; numbers in parentheses indicate the number of mice with that condition. r,~, Not done. shown). Clearly, CVB4-P15 shows a strong tropism for the exocrine pancreas. Dead acinar cells were replaced with adipose tissue as early as 3 weeks p.i., and by 6 weeks p.i. fat was well organized. In all strains the extent to which fat replaced acinar tissue during the convalescent phase seemed to be directly proportional to the extent of acinar tissue necrosis suffered during the acute phase. Other commonly observed features were the persistence of inflammatory cells intimately associated with acinar tissue, calcification and evidence of recent acinar necrosis. Acinar cells showing degranulation with swollen and vacuolated cytoplasm (e.g. Fig. 4) were also observed in mice of strains BALB/c and C57BL/ksdbm (Table 2). Acute myocarditis was seen in an SWR/J mouse 24 days p.i. (not shown). The pattern of pathological damage seen in the pancreas of all strains between 2 and 10 months p.i. (e.g. Fig. 5 to 8) largely resembled that seen during the convalescent phase, except that the healed pancreas pre-

5 Persistent CVB4 infection of mouse pancreas ' 1391,j~ Fig. 1. An uninfected C57BL/6Jdbm mouse pancreas. One large and one small islet of Langerhans (i) are surrounded by acinar (exocrine) tissue in which a pancreatic duct (d) and a small blood vessel (b.v.) can be seen. Haemotoxylin and eosin (H & E) staining; bar marker represents 0.4 p.m. Fig. 3. C57BL/6Jdbm mouse pancreas 7 days after infection with 10a p.f.u. CVB4-P15 showing focal pancreatitis (f.p.) top right and lower right, adjacent to islet tissue. H & E staining; bar marker represents 1 p.m. F2 /t ti Fig. 2. C57BL/ksdbm mouse pancreas 5 days after infection with 104 p.f.u. CVB4-P 15 showing loss of acinar tissue and massive inflammatory infiltrate, whereas islet (endocrine), ductal and vascular tissue are largely unaffected. H & E staining; bar marker represents 0.4 p.m. sented areas of well organized adipose tissue alongside healthy acini. An insulitis was observed in one C57BL/ ksdbm mouse pancreas section and heart lesions were observed at 6 months p.i. and at 8 months p.i. in C57BL/ 6Jdbm and SWR/J mice respectively (not shown). From a total of 83 pancreases examined after the acute phase, 42 were healed (five of these showed no evidence of prior infection), 23 showed foci of inflammation and in 18, new lesions were seen (e.g. Fig. 5). In five pancreas sections evidence of recent necrosis was found in organs which showed little or no sign of previous damage (e.g. Fig. 6, 10 months p.i.). Mature (or well organized) and newer (less well organized) adipose tissue were readily distinguishable by their structure (e.g, Fig. 7). In short, evidence of continuing pathological damage was found in all mouse strains other than S J L / J and C3H/HeJ. Fig. 4. BALB/c mouse pancreas 1 month after infection with 102 p.f.u. CVB4-P 15 showing replacement of necrotized acinar cells with adipose tissue (a.d.). Surviving acinar cells are degranulated with swollen vacuolated cytoplasm. H & E staining; bar marker represents 1 p.m. Fig. 5. SWR/J mouse pancreas 6 months after infection with 102 p.f.u. CVB-P15 showing adipose tissue in loci of recent necrosis (n). The presence of inflammatory cells and acinar cells showing cytopathic changes (as in Fig. 4) is suggestive of ongoing necrosis. H & E staining, bar marker represents 1 i.tm.

6 1392 C. Vella, C. L. Brown and D. A. McCarthy Immunolocalization o f viral antigen i: : i Fig. 6. C57BL/6Jdbm mouse pancreas 10 months after infection showing necrosis involving an entire lobule. Such extensive necrosis is not typical of late lesions. H & E staining, bar marker represents 0.2 ~tm. Pancreas sections showing histological lesions were examined for viral antigen by IGSS or IFA (Table 3). Antigen was readily detected in acinar tissue between 2 and 4 or 5 days p.i. depending on the mouse strain (e.g. Fig. 8). Between 3 and 10 days p.i. infiltrating inflammatory cells also stained positively for virus antigen, and staining of duplicate sections with the monoclonal antibody F4/80 indicated that the majority of these virus-positive cells were mature macrophages (e.g. Fig. 9). Murine macrophages are not permissive to the replication of CVB in vitro (Kabiri & Hadaegh, 1977), but can take up and carry these viruses (Gauntt et al., 1979). In this model macrophages may play a role in virus clearance, but their abundance in the pancreas may also suggest that they contribute to tissue pathology. Antigen could not be detected in other pancreatic tissues, nor in liver, heart or spleen. However, viral R N A has been detected in pancreatic acinar cells and occasionally in heart, liver and spleen by in situ hybridization (Vella et al., 1991). Antigen was not detected in C3H/HeJ mice beyond 4 days p.i. During the later stages of infection (10 days to 10 months), small groups of acinar cells stained for virus antigen (e.g. Fig. 10, Table 3). Macrophages were also occasionally found among foci of acinar cell inflammation (not shown). Virus antigen was not detected in nonpancreatic tissues at this time, although virus RNA was found in the liver and heart of the single mouse examined at 1 month p.i. (Vella et al., 1991) Both the histological and immunocytological findings suggest that CVB4-P15 persists in pancreatic acinar cells In contrast to the diffuse lesions seen during acute infection, late lesions were mostly discrete or focal in appearance (e.g. Fig. 5). Evidence of prior damage (mature fat replacement), more recent damage (less Fig. 7. SWR/J mouse 6 months after infection showing well structured mature adipose tissue (m.a.d.) around an islet of Langerhans. Areas of newer fat replacement (centre right) are distinguished by smaller cell size and less ordered structure. H & E staining, bar marker represents 0.4 ~tm. T a b l e 3. i Detection of virus antigen in pancreatic acini* Strain Time after infection~" 2 to 3 days 4days 5 days 6 days 7 days 8 to 9 days 10 days to 6 weeks 6 weeks to 3 months 3 months to 5 months 6 months to 8 months 9 months to 10 months SWR/J BALB/c C57BL/6Jdbm 1:3:~ 0:1 1:2 3:3 4:4 5:5 0:2 0:2 5:5 2:4 1: 1 2 :4 0:4 1:2 0 :2 1: 1 C57BL/ksdbm SJL/J CBA/H-T 6 C3H/Heola DBA/1 N1H C3H/HeJ 4:4 2:3 2:3 3:4 0:3 0:1 0:1 1:4 t :1 3:4 1:3 1:2 1:3 2:4 * All results were obtained by IGSS, except for those marked which were obtained by IFA. t Only pancreas sections showing histological evidence of infection (necrosis or inflammation) were examined. First figure indicates the number found positive, the second figure is the total number of mice examined. 1:3 0:2

7 Persistent CVB4 infection of mouse pancreas 1393 organized fat replacement) and new focal lesions was found within a single pancreas section on four occasions. Lesions of the heart were rarely found and, although virus was occasionally isolated from the liver and spleen several months after infection, lesions were not seen. Moreover, virus was detected in the serum on only one occasion during the late phase of infection (Table 1). Taken together these observations suggest that persistent infection is localized to the pancreas. CVB4-specific neutralizing antibody Fig. 8. C57BL/ksdbmmouse pancreas 4 days after infection showing CVB4-P15 antigen with IGSS (black) in acinar cells adjacent to an islet. Bar marker represents 1.7 ~tm. Fig. 9. SWR/J mouse pancreas 10 days p.i. showing macrophages (black) stained with F4/80 antibody by IGSS, in association with necrosed acinar cells, and in the interstitium. Bar marker represents 1.7 ~tm. Neutralizing antibody was detected in all strains during the acute infection (Vella & Festenstein, 1992) and four strains (C57BL/ksdbm, C57BL/6Jdbm, BALB/c and SWR/J) were sampled at later stages of the infection. Between 6 weeks and 10 months p.i. reciprocal serum neutralization titres ( 5 0 ~ endpoint) were comparable with or higher than acute titres in individual animals, suggesting the continuation of a humoral immune response. Between 6 weeks and 10 months p.i. 10 of 24 C57BL/ksdbm mice had reciprocal serum neutralization titres of /> 10; individual titres ranged between 20 and 340. Similar results were seen with 10 of 28 C57BL/ 6Jdbm mice (range 15 to 120). With strain SWR/J, 25 of 34 mice showed titres in the range 10 to 680; 19 of these animals had titres >/40. Only 13 BALB/c mice were sampled between 6 weeks and 6 months p.i., but nine of these animals had titres in the range 10 to 80. Retrospectively, four of the 10 mouse strains were found to have seroconverted for mouse hepatitis virus (MHV). Clinical evidence of M H V infection was seen 10 weeks after infection with CVB4-P15 in SWR/J mice, although these animals were seronegative for M H V by external screening (The Microbiology Laboratories, Middlesex, U.K.) prior to infection with CVB4-P15. Routine internal screening (Organon Technika ELISA) revealed that strains NIH, DBA/1 and SJL/J had seroconverted for MHV by 3 months, and C57BL/6Jdbm and C57BL/ksdbm mice not earlier than 6 months after CVB4-P15 infection. Although it is possible that strains DBA/1 N I H and SJL/J became persistently infected with CVB4-P15 only when co-infected with MHV, coinfection was clearly not essential for the establishment of a persistent CVB4-P15 infection in the other strains (Tables 1, 2 and 3). Concluding remarks Fig. 10. CBA/H-T6 mousepancreas 6 weeks p.i. showingreplacement of dead acinar cells with well structured adipose tissue and CVB4-P15 antigen (as in Fig. 8) in surviving acinar cells. H & E staining; bar marker represents 1 ktm. Persistent infections with lytic viruses such as CVB have been reported in immunocompromised mice when there is extensive pathology and high mortality (e.g. Schnurr & Schmidt, 1984; Schnurr et al., 1984). However, this is the third report of persistent coxsackievirus infection of

8 1394 C. Vella, C. L. Brown and D. A. McCarthy normal adult mice producing few or no overt clinical symptoms. CVB1 has been isolated for up to 180 days (Bocharov & Shalaurinova, 1984) and CVB3 for up to 390 days in BALB/c mice (Guriunova & Savinov, 1987). With certain viruses, e.g. herpesvirus (Croen et al., 1987) and Aleutian mink disease virus (Bloom et al., 1989), persistent infections may be associated with restricted virus replication. Cunningham et al. (1990) have proposed that chronic enterovirus infection in non-inflammatory muscle disease is associated with mutant viruses defective in their control of RNA synthesis. In this model, in contrast to the extensive pathology seen during the acute phase, persistent infection is characterized by discrete new lesions, the presence of virus antigen in small groups of acinar cells and a continuing humoral immune response with low virus titres in the pancreas. The authors wish to thank Professor K. W. Taylor and Dr T. M. Szopa for the virus used in this study, and for their help and guidance. This work was supported by the Nuffield Foundation. References ARCHARD, L. C., BOWLES, N. E., OLSEN, E. G. J. & RICHARDSON, P. J. (1987). Detection of persistent coxsackie B virus RNA in dilated cardiomyopathy and myocarditis. European Heart Journal 8 (supplement J), BENDINELLI, M. & FRIEDMAN, H. (editors) (1988). Coxsackie Viruses -- A General Update. New York: Plenum Press. BLOOM, M. E,, ALEXANDERSEN, S., MORI, S. (~ WOLFINBARGER, J. B. (1989). Analysis of parvovirus infections using strand specific hybridization probes. Virus Research 14, BOCHAROV, E. V. (~ SHALAUROVA, O. V. (1984). Persistence of coxsackie B, virus in BALB/c mice. Aeta virologica 28, 345. BOWLES, N. E., RICHARDSON, P. J., OLSON, E. G. J. & ARCHARD, L. C. (1986). Detection of coxsackie B virus specific RNA sequences in myocardial biopsy samples from patients with myocarditis and dilated cardiomyopathy. Lancet i, BOWLES, N. E., SEWRy, C. A., DUBOWITZ, V. & ARCHARD, L. C. (1987). Dermatomyositosis, polymyositis and coxsackie B virus infection. Lancet i, CROEN, K. D., OSTROVE, J. M., DRAGOVIC, L. J., SMIALEK, J. E. & STRAUS, S. E. (1987). Latent herpes simplex virus in human trigeminal ganglia. New England Journal of Medicine 317, CUtC~r~OnAM, L., BOWLES, N. E., LANE, R. J. M., DUaOWITZ, V. & ARCHARD, L. C. (1990). Persistence of enteroviral RNA in chronic fatigue syndrome is associated with the abnormal production of equal amounts of positive and negative strands of enteroviral RNA. Journal of General Virology 71, EASTON, A. J. & EGLIN, R. P. (1988). The detection of coxsackievirus RNA in cardiac tissue by in situ hybridization. Journal of General Virology 69, GAUNTr, C. J., TROUSDALE, M. D., LABADIE, D. R. L., PAQUE, R. E. & NEALON, T. (1979). Properties of coxsackie B3 variants which are amyocarditic or myocarditic for mice. Journal of Medical Virology 3, GURItmOVA, A. G. & SAVIIqOV, A. P. (1987). Persistent infection caused by the coxsackie B3 virus in adult mice. Voprosy virusologii 32, KAmRI, M. & HADAEOH, M. D. (1977). Interaction of coxsackie virus B3 with peritoneal exudate cells of adult mice treated with cyclophosphamide. Journal of Medical Virology 1, MELNICK, J. L., SHAW, E. W. & CURNEN, E. C. (1949). A virus from patients diagnosed as non-paralytic poliomyelitis or aseptic meningitis. Proceedings of the Society for Experimental Biology 71, SCHNURR, D. P. & SCHMIDT, N. J. (1984). Coxsackievirus B3 persistence and myocarditis in NFR nu/nu mice. Medical Microbiology and Immunology 173, 1-7. SCHN-ORR, D. P., CAO, Y. & SCI-IMIDT, N. J. (1984). Coxsackievirus B3 persistence and myocarditis in N :NIH(S) II nu/nu and +/nu mice. Journal of General Virology 65, V~LLA, C. & F~rmqSTEIr~, H. (1992). Coxsackievirus B4 infection of the mouse pancreas: the role of natural killer cells in the control of virus replication and resistance to infection. Journal of General Virology 73, VELLA, C., EASTON, A. J., EGLXN, R. P., BROWN, C. L. & PERRY, L. (1991). Coxsackievirus B4 infection of the mouse pancreas. 1. Detection of virus specific RNA in the pancreas by in situ hybridization. Journal of Medical Virology 35, (Received 11 November 1991 ; Accepted 5 February 1992)