Tamiami virus infection in mice and cotton rats

Tamiami virus infection in mice and cotton rats W. C. WINN, JR' & F. A. MURPHY 2 Tamiami virus produces a lethal encephalitis in suckling mice, and the illness is mediated, at least in part, by cellular immunity. Infection of extraneural organs, including lymphoid organs, is limited. The same virus produces a widespread infection in its natural host, the cotton rat, but neither symptomatic illness nor cytopathology results. Since antibody is produced, as in the murine infection, suppression of cellular immunity to the virus may be responsible for the non-cytolytic infection. Lymphoid tissue is extensively infected in the cotton rat and a relationship between this lymphotropism and immunosup- pression is suggested. The relationships of arenaviruses to their animal hosts have challenged investigators since the initial isolation of lymphocytic choriomeningitis (LCM) virus. Chronic infections with continued excretion of virus are well documented laboratory phenomena and are probably important mechanisms for the maintenance of the viruses in nature. Deviation from non-cytopathic infection has been described when LCM virus is inoculated into suckling rats, rather than into the natural mouse host (1). Following studies of the encephalitis produced by Tamiami virus in the laboratory mouse, we investigated the infection produced by this virus in its natural host, the cotton rat. PATHOGENESIS OF TAMIAMI VIRUS INFECTION IN MICE Tamiami virus, one of two North American arena- viruses, was first isolated from cotton rats in the Everglades of Florida by Calisher et al. (2). As a non-human pathogen, it represents a useful model for the study of arenavirus infections. Suckling mice inoculated intraperitoneally at 2-5 days of age de- velop a lethal meningoencephalitis (3). When the mice are inoculated at less than 18 hours of age, encephalitis results, but the inflammatory response is less severe and a high proportion of the mice survive the infection. Gilden et al. (4) have demonstrated that the lethal murine infection can be prevented by I Assistant Professor of Pathology, University of Virginia Medical School, Charlottesville, VA 22903, USA. 2 Chief, Viral Pathology Branch, Center for Disease Control, Atlanta, GA 30333, USA. neonatal thymectomy. These findings reinforce the conclusion that immunological immaturity accounts for the survival of mice inoculated immediately after birth. Tamiami virus infection in mice is neurotropic. Although multiplication of virus occurred in the spleen and kidney, titres of infectious virus were significantly lower in those organs than in the brain. Viral antigen, which was widespread in the brain, was seldom seen in other organs; specifically, anti- gen was never observed in lymphoid tissue. Humoral antibody response in the murine model was delayed. After the appearance of neutralizing antibody, infectious virus in the brain slowly drop- ped to undetectable levels, but the outcome of infection had already been established. Animals inoculated at 2 days of age died before neutralizing antibody developed. Complement-fixing (CF) anti- body appeared earlier, but signs of illness and tissue destruction were found even before this. There were no significant differences in CF antibody production between the high and low mortality groups of mice. PATHOGENESIS OF TAMIAMI VIRUS INFECTION IN COTTON RATS After inoculation of cotton rats with the same pool of Tamiami virus that had been used in the murine infection, no signs of disease developed in any animal, although virus was widely distributed in the body. Most of the pathogenesis studies were done in animals 2 days old at the time of inocula- tion, using methods that have been reported previ- ously (3, 5). When the cotton rats were less than 18 3377 501 - BULL. WORLD HEALTH ORGAN., Vol. 52, 1975 W. C. WINN & F. A. MURPHY hours old at inoculation, their responses did not differ significantly from those of the 2-day old animals. After intraperitoneal inoculation, virus initially multiplied in lymphoid tissue. By day 9 after inocu- lation the titres of infectious virus in lymphoid organs had reached 104-105 plaque-forming units (PFU) per gram of tissue, viraemia had developed, and many organs yielded infectious virus. At this time, distribution of viral antigen in lymphoid organs was well defined by immunofluorescence. Antigen was concentrated in the cortex of lymph nodes with extension to the peripheral sinus (Fig. 1), in the cortex of the thymus gland (Fig. 2), and in the white pulp of the spleen, where both the marginal zone and the periarteriolar sheath contained antigen. The immunofluorescent patterns indicated an initial viral invasion of the reticular and epithelioid cells, with rapid subsequent infection of lymphoid cells. Electron microscopy revealed only scanty viral par- ticles and inclusions, but they were associated with " lymphoblastoid " cells, as described previously in Machupo virus infection of Calomys callosus (6). Additionally, in the spleen and bone marrow, mega- karyocytes containing viral antigen were observed; a productive infection was demonstrated ultrastruc- turally, with accumulation of virions in the platelet demarcation channels. By day 40 after inoculation, infectious virus was no longer present in the lym- phoid organs and antigen was limited to a few histiocytes. The major non-lymphoid organs that supported viral multiplication were the liver, salivary gland, kidney, and brain. Peak titres (up to 107 PFU/g) were reached 16 days after inoculation; there was then a fall to undetectable levels by day 40 after inoculation. The potential for salivary gland excre- tion of infectious virus was confirmed by the pre- sence of large numbers of viral particles, maturing at the apical membrane of acinar cells and shedding into the glandular space. In all these organs, however, there was a signifi- cant time-lag between the appearance of infectious virus and of viral antigen, as detected by immuno- fluorescence. The dichotomy was most apparent in the brain, in which viral antigen was very sparse until 40 days after inoculation, at which time infec- tious virus was no longer demonstrable. From 40 to 90 days after inoculation, antigen steadily accu- mulated and remained at a plateau until the end of the experiment at one year. Electron microscopy of the brain confirmed the paucity of viral particles and the identity of viral antigen accumulation with inclu- sion body formation. The inclusions were found in neurons and oligodendroglia, and consisted of amas- sed ribosome-like particles in an electron-dense matrix (Fig. 3). Complement-fixing antibody was first detected on day 13 and neutralizing antibody on day 23 after inoculation. Neutralizing antibody increased rapidly and remained at a high plateau level for the duration of the experiment. Termination of viraemia co- incided with the appearance of neutralizing antibody; infectious virus in organs was detected in high titre for an additional week. Despite the high viral titres and large amounts of viral antigen present, no cytopathology was seen. With the exception of a few perivascular collections of mononuclear cells in the brain, all organs ap- peared histologically normal during the acute stages of infection. The major anatomic finding was the development of extensive lymphoid hyperplasia, beginning two weeks after infection. Mononuclear infiltrates also appeared in several organs, especially the salivary gland and kidney (Fig. 4). These infil- trates did not appear to be neoplastic, were not associated with tissue damage, and no viral antigen could be detected by immunofluorescence. DISCUSSION Comparison of Tamiami virus infection in the cotton rat and in the mouse demonstrates major differences in pathology and pathogenesis. Ultima- tely, the same viral inoculum produces a non-cyto- pathic infection in 2-day old cotton rats and a destructive, lethal encephalitis in 2-day old mice. Several mechanisms for the differences in host re- sponse are possible: (a) viral multiplication in the cotton rats may be aborted at an early stage before infection of vital organs occurs; (b) replication of virus in the cotton rat may be defective, so that plasma membranes are not antigenically altered by viral maturation and cytoplasmic antigen is inaccess- ible to cytolytic immune mechanisms; or (c) cyto- plasmic membranes may be antigenically altered and virus produced without the immune system of the host recognizing the altered cells as " foreign ". The high titres of virus in numerous organs of the cotton rats clearly indicate that the infection is disseminated; in fact, viral multiplication in systemic organs was far more extensive than in the mouse model, in which the virus was neurotropic. A shift from production of infectious virus to accumulation 502 Fig. 1. Tamiami virus antigen in lymph node cortex of the cotton rat at 16 days after inoculation. Infection involved lymphoid elements and cortical reticulum at this time. All micrographs are of tissues of cotton rats inoculated intraperitoneally at 2 days of age. For frozen section immuno- fluorescence, an FITC-conjugated anti-Tamiami virus globulin was used. The bar represents 40 ,m. Fig. 2. Tamiami virus antigen in thymus cortex of the cotton rat at 16 days. Both thymocytes and reticulum contained antigen at this peak time. The bar represents 40 1pm. Fig. 3. Tamiam? virus inclusion body (IB), but no virus particles, in an oligodendroglial cell in the brain of the cotton rat at 23 days after inoculation. Inclusions consisted of ribosome-like particles embedded in an electron-dense matrix. The bar represents 0.5 4m. 1 - Fig. 4. Tamiami virus infection of the salivary gland in the cotton rat at 30 days after inoculation. Reticuloendothelial hyperplasia displaced acinar structures, but was not otherwise damaging. The bar represents 40 jum. Ir'li1!V ''I" -1 TAMIAMI VIRUS INFECTION IN MICE AND COTTON RATS of cytoplasmic antigen may play a part in limiting cytopathology. In the cotton rat, the reciprocal rela- tionship between the production of virus early in infection and the production of cytoplasmic antigen later in infection is well defined in several organs. The mechanism is unclear, but temporal relation- ships suggest that neutralizing antibody may be involved. A similar phenomenon has been observed in cell cultures infected with measles virus and incubated in the presence of antibody (7). Cytopathology in the mouse model did correlate better with the presence of infectious virus than with viral antigen, and incomplete maturation of virus may explain the absence of cell damage late in infection. It cannot, however, explain the non-cyto- toxic course of the early, productive infection in cotton rats. Gilden et al. (4) have demonstrated that cellular immunity is responsible, at least in part, for ence- phalitis in newborn mice. In our studies (3), neutral- izing antibody was clearly not related to cytopatho- logy. CF antibody in the mouse was detected shortly after the development of encephalitis, although it did not correlate with the outcome of infection. Both neutralizing and CF antibodies were formed in the cotton rats, yet no cytotoxicity occurred; antibody production was similar in quantity and time of appearance to the murine immune response. We have not made direct measurements of cellular immunity in the cotton rat, but the implication is that suppression of T cell response to Tamiami virus is responsible for the non-cytopathic infection. This state of unresponsiveness might be permanent, but might also be temporary and restricted to the time period when cytoplasmic membranes are altered by maturing virions. The mechanism for this postulated immunosup- pression is not known, but attention must be directed to the dramatic difference in infection of lymphoid tissue between the two hosts. Extensive infection of lymphocytes was seen only in the animal model, which did not demonstrate cellular destruc- tion or symptomatic disease. A functional alteration, rather than cellular destruction, was suggested by the morphological integrity of lymphoid organs in these animals. RtSUMI2 INFECTION VIRALE A TAMIAMI CHEZ LA SOURIS ET CHEZ LE RAT DU COTON Le virus Tamiami, un des deux arenavirus de l'Ame- rique du Nord, a ete isole pour la premiere fois a partir du rat du coton dans les Everglades de Floride. Nous avons compare la pathogenese de l'infection chez son h6te naturel d'une part et chez la souris de laboratoire d'autre part. L'inoculation intraperitoneale du virus chez des souri- ceaux ages de 2 A 5 jours a provoque une encephalite mortelle tandis qu'une forte proportion de souriceaux ages de 18 heures au moment de l'inoculation ont sur- vecu a l'infection. L'antigene n'a jamais ete observe dans le tissu lymphoide. Gilden et al. ont demontre que l'infec- tion pouvait etre prevenue par la thymectomie neonatale. La production des anticorps ne coincidait pas avec I'histopathologie. Chez le rat du coton, le virus Tamiami a provoque une infection non-cytolitique. Le neuvieme jour de la maladie, les organes lymphoides presentaient des titres eleves de virus infectant; le virus a et detecte en forte concentration dans plusieurs organes malgre 1'absence de reaction cellulaire. Apres disparition du virus infec- tant, on a note une accumulation d'antigenes viraux dans plusieurs organes, et plus particulierement dans le cerveau. Les anticorps de fixation du complement et les anticorps neutralisants etaient pr6sents comme dans l'infection murine. La pathogenese comparee de l'infection a virus Tamiami chez la souris d'une part et chez le rat du coton d'autre part donne a penser que c'est la suppression de l'immunit6 cellulaire a l'egard du virus qui est respon- sable du caractere non-cytolitique de l'infection. L'im- muno-suppression supposee pourrait etre liee a l'infec- tion virale du tissu lymphoIde. Comme on n'a pas observe de necrose lympholde, il est possible qu'il y ait alteration fonctionnelle plut6t que destruction des lym- phocytes. II se pourrait aussi que l'absence de cyto- pathologie au stade tardif de l'infection soit due 'a sa transformation en infection non productive. 7 505 506 W. C. WINN & F. A. MURPHY REFERENCES 1. MONJAN, A. A. ET AL. Pathogenesis of LCM disease in the rat. In: Lehmann-Grube, F., ed. Lymphocytic choriomeningitis virus and other arenaviruses. Berlin, Heidelberg, & New York, Springer, 1973, pp. 195-206. 2. CALISHER, C. H. ET AL. Tamiami virus, a new member of the Tacaribe group. Amer. J. trop. Med. Hyg., 19: 520-526 (1970). 3. WINN, W. C. ET AL. The pathogenesis of Tamiami virus meningoencephalitis in newborn mice. In: Leh- mann-Grube, F., ed. Lymphocytic choriomeningitis virus and other arenaviruses. Berlin, Heidelberg, & New York, Springer, 1973, pp. 299-311. 4. GILDEN, D. H. ET AL. Tamiami virus-induced im- munopathological disease of the central nervous system. In: Lehmann-Grube, F., ed. Lymphocytic choriomeningitis virus and other arenaviruses. Berlin, Heidelberg, & New York, Springer, 1973, pp. 287-297. 5. WEBB, P. A. ET AL. The measurement of specific antibodies in Bolivian hemorrhagic fever by neutral- ization of virus plaques. Proc. Soc. exp. Biol. Med., 130: 1013-1019 (1969). 6. MURPHY, F. A. ET AL. Ultrastructural studies of arenaviruses. In: Lehmann-Grube, F., ed. Lympho- cytic choriomeningitis virus and other arenaviruses. Berlin, Heidelberg, & New York, Springer, 1973, pp. 273-285. 7. RUSTIGIAN, R. Persistent infection of cells in culture by measles virus. II. Effect of measles antibody on persistently infected HeLa sublines and recovery of a HeLa clonal line persistently infected with incomplete virus. J. Bact., 92: 1805-1811 (1966). DISCUSSION NATHANSON: Under what circumstances do you find per- sistent viral antigen without being able to recover infec- tious virus? WINN: In the mouse model, we saw a decrease in the titre of infectious virus. An antigen remained for a con- siderable period of time, but it was never clear to me whether this was just antigen that was not being cleared from infected cells. In cotton rat pups inoculated periph- erally when they were 2 days old, infectious virus could be measured in a number of organs, especially the brain, by plaquing in Vero cells or inoculating into mice; it reached a peak early in the infection and then fell off about a month after infection. It was just about that time that we began to see an accumulation of viral anti- gen by immunofluorescence. Infectious virus was present only for about 30-40 days; antigen persisted for a year. K. JOHNSON: If your Tamiami virus is closely related to that occurring in nature and you studied it in the rodent that is its sole natural host, how, with the data you have presented, do you conceive that Tamiami virus can sur- vive in nature in Sigmodon? WINN: This comment is well taken. Our cotton rats were originally wild-trapped in Georgia and our later results differ somewhat from those of the limited studies done in wild cotton rats during the initial investigation. For instance we were unable to detect virus in the urine, despite renal infection, although that may have been for technical reasons. The host-parasite relationship that we have observed between Tamiami virus and the cotton rat is different from the relationships described for other arenaviruses, and it may be that we cannot generalize from the findings with one virus. It remains to be seen whether this relationship holds for cotton rats in the wild. Our findings indicate a noncytopathic infection with excretion of virus. It is possible that other strains of cotton rat may exhibit an even closer approach to the persistent viraemia observed with other arenaviruses.