Laboratory diagnosis of lymphocytic choriomeningitis*

Laboratory diagnosis of lymphocytic choriomeningitis * JOHN HOTCHIN 1 & EDWARD SIKORA2 Current laboratory methods for the diagnosis of lymphocytic choriomeningitis are discussed, mainly from the point of view of recent improvements. The use of sentinel guinea-pigs is suggested for the detection of lymphocytic choriomeningitis infection in animal colonies. The laboratory diagnosis of lymphocytic chorio- meningitis (LCM) has improved markedly during the past few years. As late as 1968 the main methods (13) were limited to the inoculation of mice, ham- sters, guinea-pigs, cell cultures, or eggs, and to complement-fixation (CF) or virus neutralization tests for infections a few months old. Today, virus isolation (and presumptive diagnosis) in mice is possible in as little as 4 days with the endotoxin challenge method (Hotchin & Sikora, unpublished observations). By means of immunofluorescence techniques, positive serological confirmation can often be obtained within a day or so of meningitis symptoms in man, and by means of the microplaque reduction test (8) neutralizing antibody can be detected much earlier than previously believed pos- sible. The status of serological diagnosis of this virus disease was reviewed in 1971 (4, 5, 11). This review will concentrate on the newest addi- tions to the practical identification of LCM, with particular emphasis on recent experience with ham- ster-derived human infections (10). DIAGNOSTIC IMPROVEMENTS RESULTING FROM RECENTLY DEVELOPED TESTS Virus isolation The most sensitive medium for the detection of infectious LCM virus particles is still the living animal, usually the mouse, in spite of the develop- ment of other specially designed tissue culture detec- tion systems. For example, L-cell culture plus CF by the viral antigen (12) is useless for virus isolation since it is stated by the originator to be at least * From the Division of Laboratories and Research, New York State Department of Health, Albany, NY, USA. 1 Assistant Director, Virus Laboratories. 2Senior Bacteriologist. 9 times less sensitive than mouse inoculation. Methods of this type are less valuable than plaque assay, now readily available for LCM virus (9, 14). However, this method has been used to solve specific assay problems, such as the detection of virus pro- duced by isolated clones (6), and virus isolation in tissue culture, e.g., from human blood, can some- times be made by this method. If combined with the immunofluorescent technique to detect newly formed intracellular antigen, this method can be used for rapid diagnosis (in about 3 days) in many cases. When animal isolation is employed, the main difficulty has been variation in the response of infected animals. Whereas in some animals typical convulsive death may occur after as little as 5 or 6 days, some suffer nonconvulsive, gradual debilita- tion, followed by death after 10-20 days, and others develop a chronic runting disease, which may ter- minate in death or recovery; still other animals show no clinical disease at all yet high levels of viraemia may persist for many months. LCM virus strains found in persistently infected animals have been observed in this laboratory to be very liable to cause inapparent, persistent infection during isolation attempts. There is consequently a real danger of missing persistent infection, e.g., in persistently infected hamsters, in standard animal tests for LCM virus. The danger is eliminated by challenging the test mice with an Escherichia coli endotoxin prepara- tion a few days after inoculation of the test mice (Hotchin & Sikora, unpublished observations). The method depends on the observation (3) that during LCM infection mice become exquisitely (and fatally) sensitive to normally harmless doses of endotoxin; the only other virus currently known to duplicate this effect in mice is mouse hepatitis virus (2). In this laboratory the use of endotoxin for sharpening the 3383 - 555 BULL. WORLD HEALTH ORGAN., Vol. 52, 1975 J. HOTCHIN & E. SIKORA endpoint of LCM virus assays in mice is routine. The endotoxin (100 p,g in 0.2 ml of buffered saline, pH 7.2) is given by the intraperitoneal route 7 days after intracerebral virus inoculation of young adult mice. Animals are inspected 24 h later. Under these conditions, essentially all mice that have received virus die within 24 h of the endotoxin challenge. Under special circumstances the virus can be de- tected by the third day after infection by giving a larger dose of endotoxin at that time. Some titration sensitivity is lost by the earlier endotoxin challenge. During a recent occurrence of human LCM con- tracted from hamsters, the initial diagnosis of LCM was made by challenge of test mice on day 4 after infection, enabling a presumptive diagnosis to be made on day 5 (10). The agarose-suspended cell method of plaque assay can be used for virus isolation, but is not of practical value except in laboratory circumstances, e.g., the detection and assay of virus in large num- bers of test animals. Under these circumstances, the plaque assay is more accurate, quicker to perform, and much cheaper than the mouse-endotoxin method, although it is slightly less sensitive. The plating efficiency compared to intracerebral titration for LCM virus in our laboratory was found to be 0.25 and 0.40 for the clear and turbid LCM plaque- type variants, respectively (9). In titrations of the level of LCM viraemia in 20 mice believed to be persistently infected, 18 gave good titres on plaque assay; the sera of the other 2 gave no plaques. When each serum was used to inoculate each of 2 mice, followed by routine endotoxin challenge on day 7 after infection, the same 18 sera caused the death of both mice in all cases, whereas the 2 plaque-free sera caused no mouse deaths with endotoxin. When these 4 mice were challenged intracerebrally with 100 LD50 of LCM virus on day 7 after inoculation of endo- toxin, 1 out of 2 in each case proved to be immune. Death occurred in all 5 mice of the same age and weight that received identical virus and endotoxin. Thus the plaque assay in this study was as useful in detecting virus as inoculation of 2 mice but was less sensitive than mouse inoculation followed by LCM virus challenge. The latter combination has consis- tently proved to be the most sensitive method for detecting very low doses of LCM virus (7). Serology Improvements in the detection of neutralizing antibody for LCM virus and refinements of fluores- cent antibody (FA) serology have overtaken the CF test as the most valuable method for detecting LCM antibodies in man and probably in animals too. This has left the CF test as still the most readily available method, but the FA test is the more sensitive in the early phase of infection. The micro plaque reduction (MPR) method of detecting neutralizing antibody is the most sensitive way of detecting antibody in older cases (several months to several years after infec- tion). Immunofluorescence We have used minor but important modifications of the test originally devised by us some years ago (1). If carefully controlled and used with a standardized conjugate dilution, this test can give an accurate titre of immunofluorescent LCM antibody in a serum. Usually a 1: 8 to 1: 12 dilution is used to screen sera; a more detailed titration is performed later if required. Positive results at dilutions of less than 1: 8 are considered doubtful unless taken dur- ing the disease with a negative earlier serum. Impor- tant modifications of the method include the use of L cells as the antigen and ensuring that only about 30% are infected. The fact that only 30% of cells are positive enables this proportion to be readily identi- fied on each field. If an equal proportion of infected and uninfected cells were used, non-specific results in 50% of cells could not be identified as being present in infected as opposed to uninfected cells. Smears are prepared in bulk and stored at -70°C. This results in a standard antigen, with large cells showing a clearly defined punctate-stippled fluorescent pattern; the 70% of uninfected cells provide a " built-in" negative control on every field examined. This pro- vision is of great help in distinguishing false positives due to non-specific fluorescence. Experience soon shows that non-specific results can be obtained in at least two ways. In both cases the colour of the fluorescence is lighter than the specific fluorescence. In one form, only LCM virus- infected cells are stained, yet exhaustive tests show that no LCM specificity is in fact present. Appar- ently, LCM virus-infected cells have a greater ten- dency to pick up non-specific components in certain sera. We have seen the same type of non-specific staining of LCM virus-infected cells (but not con- trols) with certain batches of conjugate without prior serum exposure. In these cases there is no stippled pattern to the fluorescence. Only the stippled fluores- cence should be regarded as positive. In the second 556 LABORATORY DIAGNOSIS OF LCM form of non-specific staining, control uninfected cells also are stained; this type is easily distin- guished, particularly when control and infected cells can be compared in the same field. The mixed control and infected cell method also eliminates the need for two slides to be stained and examined for every test. Plaque reduction test This test has recently been described (8). While the microtitre plate method is admirable for large num- bers of tests, the standard 6-cm petri dish " macro " method is easier to perform if only a few tests are desired or if the test is only to be used intermittently. When 55 different human LCM sera (all collected relatively late after onset) were tested by MPR, FA, and CF methods (8), positive results were given by 85%, 74% and 42% of the sera, respectively. Obvi- ously, much earlier sera would be expected to give fewer MPR and more FA positives. The MPR test has the additional advantage of taking only 4 days to perform, compared with about 3 weeks for mouse neutralization methods, most of which require sub- sequent virus challenge to give clear endpoints. THE USE OF SENTINEL GUINEA-PIGS TO DETECT THE CONTAMINATION OF AN ANIMAL COLONY WITH LCM VIRUS When it became clear that the airborne route was of paramount importance in human LCM virus infection, tests were made with sentinel guinea-pigs. Young adult guinea-pigs were positioned at various points in the rooms and corridors associated with LCM work. These are classed as infected mouse rooms, non-animal LCM tissue culture rooms, and other associated rooms where infectious LCM virus was not used. The air of the animal rooms was supplied, via perforations in the doors, from the corridor. One-way air flow was used; the air was admitted to the corridor from outside, exhausted by fans and ducts from the " sterile room " within each animal room, and vented from the roof without disinfection. No guinea-pigs ever became infected with LCM except in the LCM mouse rooms. All other rooms and corridors were apparently completely LCM risk-free. Within the animal room, however, guinea-pigs routinely became infected and usually died if the number of persis- tently LCM virus-infected mice was above a certain threshold. No accurate figure of this threshold num- ber can be given, but it was usually about 100 for a room of 24 M2. If the guinea-pig was close to the persistently infected mice, infection was more likely. A typical result is seen in Fig. 1. In the guinea-pig, Over many years and in many deliberate experi- ments we have observed that mice do not readily become infected from other mice acutely infected with LCM virus. Even with persistently infected mice as the virus source, cross-infection is sporadic only when the mice share the same cage. With the virus and mouse strains used, such infection always seems to cause subclinical disease followed by immu- nity. Transferring non-immune, non-infected mice into the cages freshly used by acutely or persistently infected mice has never resulted in cross-infection in the absence of the infected mice. Direct contact between the mice appears to be necessary. In contrast, humans working on the maintenance (but not inoculation or autopsy) of infected mice can become infected in the LCM animal room. Several such infections have occurred over a period of 20 years in this laboratory, but only when several hundred persistently infected mice were present in the room. Provision of more rapid rates of fresh-air flow through the room and the placing of dustproof paper covers over the cages have eliminated any further infections. 800 700C71 I (0 w 600[ 4 DAYS IN CAGE PREVIOUSLY I USED BY 12d' PTI MICE NEAR LCM-INFECTED l 1 ON SHELF UNDER ON SHELF UNDER. HAMSTERS 15 PTI MICE 75 PTI MICE f t : . 500O 400F 10 WEEK 15 Fig. 1. Weight change of a sentinel guinea-pig in anr animal room containing different numbers of LCM- infected animals. The location of the guinea-pig relative to the main source of infection is shown in boxes. The arrow shows the probable time of infection. PTI = persistently infected; t = death of guinea-pig from LCM. __ _ = . . _ 557 5 558 J. HOTCHIN & E. SIKORA weight loss is associated with clinical LCM. Other experiments had showed that the average time before severe symptoms occurred after being placed in a high-risk LCM room was about 2 weeks. Inocula- tion gives an incubation period of 3-4 days before weight loss caused by fluid loss begins. The experi- ment shows that the guinea-pig did not become infected owing to its proximity to several LCM- infected hamsters, nor from being in a cage previ- ously occupied by 12 persistently infected male mice. Infection did occur, however, at about the time the number of persistently infected mice on a shelf immediately above the guinea-pig's cage was in- creased from 15 to 75. The main infection risk in that animal room was evidently close proximity to exhalations or dust from large numbers of such mice. Human infection in this laboratory, as judged by antibody status change, has invariably pointed to the same source. Accidental self-inoculation with a syringe or glass pipette containing LCM virus has occurred on two occasions with no outward effect and no change in negative antibody status. In summary, the use of sentinel guinea-pigs coupled with specific FA testing is the most useful method of detecting LCM virus in animal colonies. However, the guinea-pig must be subjected to fur- ther tests, including virus isolation procedures if it should become sick or die, and it should be sub- jected either to virus challenge or to serological investigation after a suitable interval if it does not, since some strains of LCM virus are relatively non- lethal for guinea-pigs. RESUME DIAGNOSTIC DE LABORATOIRE DE LA CHORIOMENINGITE LYMPHOCYTAIRE Une des ameliorations recentes du diagnostic de labo- ratoire de la choriomeningite lymphocytaire (CML) est l'utilisation de l'endotoxine d'Escherichia coli qui permet de preciser les points finals du titrage chez la souris et d'accelerer la detection et l'identification du virus, 4 jours suffisent avec l'endotoxine contre 10 a 21 jours sans endotoxine. Le systeme le plus sensible pour identifier le virus de la CML reste celui qui utilise la souris; il est un peu plus satisfaisant que la culture de tissu ou meme la m6thode des plages pour la moyenne des souches virales, et probablement beaucoup plus pour les souches per- sistantes. Des methodes permettant d'accroitre la fiabilite diagnostique des reactions d'immunofluorescence sont decrites. Les sensibilites relatives des epreuves de reduc- tion des plages, d'immunofluorescence et de fixation du complement se sont montrees decroissantes dans cet ordre. Les cobayes constituent d'excellents animaux pour la detection de concentrations atmospheriques de virus (de la plupart des souches de CML, mais non la totalite) suffiantes pour infecter l'homme par voie aerienne. L'in- fection du cobaye peut etre suspectee lorsque l'animal presente une brusque perte de poids. D'autres epreuves serologiques ou virologiques sont necessaires pour confirmer que le virus de la CML est a l'origine de la maladie. La source principale d'infection semble etre constituee par les animaux infectes de fagon persistante. REFERENCES 1. COHEN, S. M. Er AL. J. Immunol., 96: 777-784 (1966). 2. GLEDHILL, A. W. J. gen. Microbiol., 18: xvii (1958). 3. HOTCHIN, J. Cold Spring Harbor Symp. Quant. Biol., 27: 479-499 (1962). 4. HOTCHIN, J. Amer. J. Pathol., 64: 747-769 (1971). 5. HOTCHIN, J. Persistent and slow virus infections. In: Melnick, J., ed. Monographs in virology, Vol. 3. Basel, Karger, 1971. 6. HOTCHIN, J. The role of transient infection in arenavirus persistence. In: Progress in medical viro- logy. Basel, Karger, 1974. 7. HOTCHIN, J. & BENSON, L. J. Immunol., 91: 460-468 (1963). 8. HOTCHIN, J. & KINCH, W. J. infect. Dis., 131: 186- 188 (1975). 9. HOTCHIN, J. ET AL. Infect. Immun., 4: 281-286 (1971). lo. HOTCHIN, J. ET AL. Science, 185: 1173-1174 (1974). 11. LEHMANN-GRUBE, F. Lymphocytic choriomeningitis virus. In: Virology monographs, Vol. 10. Basel, Karger, 1971. 12. LEHMANN-GRUBE, F. & HESSE, R. Arch. ges. Virus- forsch., 20: 256-259 (1967). 13. RHODES, A. J. & VAN ROOYEN, C. E. In: Textbook of virology, 5th ed. Baltimore, Williams & Wilkins, 1968, p. 812. 14. SEDWICK, W. D. & WIKTOR, T. J. J. Virol., 1: 1224- 1226 (1967). LABORATORY DIAGNOSIS OF LCM DISCUSSION MONATH: Do you have any data on the chronology of viraemia in human patients? HOTCHIN: As I recall, the virus starts to appear in the blood at 3 or 4 days; it slowly builds up, and then one finds antibody and the virus titre rapidly drops. In the patients I mentioned earlier that were anergic, no anti- body appeared and the titre of virus never dropped; they had fever the whole time. WELSH: I would point out that guinea-pig serum does not inactivate LCM whatever the virus passage history. Also, when talking about pathogenicity in different animals, I think one should consider whether or not there is antibody already attached to the virus. Do you know whether in the persistently infected hamster there is antibody complexed with the virus? HOTCHIN: We feel sure that there is. In man, complement seems to augment the effect of early antibody but not that of late antibody; this might provide a means of telling late from early infection. I might add that we have for many years used guinea-pigs to prepare neutralizing antibody against LCM virus. After suitable immuniza- tion, the serum of these animals can unquestionably neutralize this virus. LEHMANN-GRUBE: Methods employed for the detection of neutralizing antibody in persons infected with LCM virus have been unsatisfactory. The mouse assay often employed is expensive and time-consuming. Cell culture assays, on the other hand, suffer from the complication that heat inactivation is necessary to remove the non- specific virus-inactivating and cytotoxic substances con- tained in most human sera; this affects the titres of specific neutralizing antibody. In our laboratory a new method has been developed by Dr Ambrassat based on the detection of virus-specific sensitizing rather than neutralizing antibody. Sera to be assayed are heated for 30 minutes at 56°C. Serial dilutions are mixed with 100 PFU of LCM virus and after incubation for 2 hours at 37°C anti-human IgG antiserum raised in the rabbit is added. Incubation is continued for 30 minutes more, after which mixtures are transferred to L-cell monolayer cultures and covered with a medium containing methyl cellulose. Plaques are counted 4 days later. HOTCHIN: Contrary to Dr Lehmann-Grube's observa- tions, Mr William Kinch in my laboratory has tested several hundred different human sera and found that there were no cytotoxic or LCM virus-inactivating sub- stances present except in known LCM cases. In these, the inactivation of virus met the criteria required for antibody. In point of fact, we have observed a stabilizing effect of normal serum on LCM virus titre. K. JOHNSON: Do you know whether anyone made a worthwhile attempt to use fluorescent antibody or other techniques to detect the presence of antigen in cells that are relatively easy to obtain from the patient (e.g., from the urine, blood, or throat)? HOTCHIN: I do not think we have tried hard enough, but in the mouse model we have attempted to diagnose LCM early by immunofluorescence, looking for the antigen. I suspect that the infected white cells do not circulate; they may be removed from the circulation. We have examined a variety of organs; the best place to find antigen was the kidney, but we also detected antigen in the spleen and skin. In some of these arenavirus diseases fluorescent antibody study of skin biopsies might be interesting. MURPHY: My impression from all the experimental model studies is that there is not going to be much virus present in any ethically obtainable specimen, especially from a patient who has only a mild infection, like LCM. One is not going to biopsy the kidney in this disease. The other question I would like to ask, which may be the other side of the coin, is what diagnostic scheme would you recommend for the average state health department, not your own laboratory? HOTCHIN: The best single technique would be immuno- fluorescence. The average laboratory does not want to work with live virus. 559