Effect of Mutation in Immunodominant Neutralization Epitopes on the Antigenicity of Rotavirus SA-11

Transcription

1 J. gen. Virol. (1985), 66, Printed in Great Britain 2375 Key words: rotaviruses/antigenieity/antiserum selection Effect of Mutation in Immunodominant Neutralization Epitopes on the Antigenicity of Rotavirus SA-11 By DOUGLAS R. KNOWLTON AND RICHARD L. WARD* James N. Gamble Institute of Medical Research, 2141 Auburn Avenue, Cincinnati, Ohio 45219, U.S.A. (Accepted 22 July 1985) SUMMARY Exposure of rotavirus SA-11 to polyclonal neutralizing antibody from hyperimmunized guinea-pigs permitted selection of variants which were poorly neutralized by antisera against the parental virus. In one-way cross-neutralization experiments, at least 22 of 24 plaque-purified variants could be classified as belonging to a serotype different from that of the parent. Most antisera generated against the variants, however, readily neutralized the parental virus. This indicates that immunodominant neutralization epitopes in the parent differed from those in the variants. Changes in immunodominant epitopes caused the serotypic relationships between the variants and other strains of rotavirus to differ from those of the parental SA-11. The serotypic relatedness of human strain P (human serotype 3) was reduced while, in contrast to results found with the parental SA-11, several of the antisera against the variants recognized the bovine rotavirus NCDV as the same serotype. Causes for these changes are discussed. INTRODUCTION Rotaviruses are a major cause of acute diarrhoeal disease in young animals and a common cause of morbidity and mortality in man, especially in infants and young children (DuPont, 1984). Development of an effective vaccine to protect against human rotaviruses is clearly needed. There are at least four known serotypes of human rotaviruses (Hoshino et al., 1984; Wyatt et al., t983) and studies conducted with calves and pigs suggest that there is little crossprotection between animal rotavirus serotypes (Bohl et al., 1984; Gaul et al., 1982; Woode et al., 1983). Human studies have suggested that some cross-protection exists (Vesikari et al., 1984) but re-infection and illness have been documented in both children and adults (Cukor & Blacklow, 1984). Thus, effective protection against rotavirus disease by active immunization may require a vaccine composed of antigens from multiple serotypes. The number of human rotavirus serotypes that currently exist and may develop in the future is not known. One method used to determine this number has been to conduct cross-neutralization studies with numerous isolates collected throughout the world. This is both tedious and costly, and the results may be valid for only a limited time because of the potential evolution of new serotypes. An alternative approach used in this report was to study serotypic changes in rotavirus under selective laboratory conditions. The method used was to expose rotavirus to neutralizing antibodies and allow survivors to replicate in cultured cells. Based upon the frequency of mutation at any one antigenic site found with other RNA viruses (Portner et al, 1980), about one out of every infectious particles should have an alteration at one of its neutralization epitopes. Such mutants should have a survival advantage under the selective conditions employed. Therefore, after multiple cycles of treatment and replication, variants with altered antigenicities could conceivably be obtained at all dominant neutralization epitopes that elicit the production of significant amount; of neutralizing antibodies (i.e. immunodominant epitopes) provided that such changes were not lethal. This type of selection procedure has been used successfully with several viruses SGM

2 2376 D.R. KNOWLTON AND R. L. WARD employing monoclonal neutralizing antibodies (Emini et al., 1982; Gerhard et al., 1981 ; Portner et al., 1980) and with influenza and visna viruses using polyclonal neutralizing antibodies (Clements & Narayan, 1984; Fasekas de St. Groth, 1975). The rotavirus chosen for this study was the simian strain SA-11. This virus was selected because it belongs to the same serotype as certain human strains (Hoshino et al., 1984) and large numbers of infectious particles can be produced in cultured cells, thus increasing the probability of variant production and selection. METHODS Cells and viruses. Growth and infectivity assays of all strains of rotaviruses were conducted in MA-104 cells, a rhesus monkey kidney cell line. Cells were grown in monolayer cultures in Special Eagle's MEM, Richter's modification (Special MEM) from Irvine Scientific Co., Santa Ana, Ca., U.S.A., containing 10~ foetal calf serum and antibiotics (100 U penicillin, 100 gg streptomycin, 2-5 gg amphotericin B per ml). Viruses studied included SA-11 (simian) provided by M. K. Estes, Baylor College of Medicine, Houston, Tx., U.S.A., OSU (porcine) and NCDV (bovine) purchased from The American Type Culture Collection, and three human strains, Wa, P and ST3, all provided by R. G. Wyatt, NIH, Bethesda, Md., U.S.A. All viruses were grown in medium without serum but containing 2 lag trypsin per ml. Viral lysates were frozen and thawed, centrifuged at 1000 g to remove cell debris, and stored in aliquots at -70 C. Purification ofrotaviruses. Viral lysates were frozen and thawed, centrifuged at 1000 g for 20 rain, and layered onto an 8 ml cushion of CsC1 (1-4 g/ml). After centrifugation (SW27, r.p.m., 1 h), the banded virus was removed and the density of CsCI was adjusted to 1.37 g/ml. The virus was then purified by isopycnic gradient centrifugation (SW50.1, r.p.m., 24 h) and dialysed against phosphate-buffered saline (PBS) containing 10 mm-caci2. Purified virus was stored at 4 C until used for inoculation of guinea-pigs. Preparation ofhyperimmune sera in guinea-pigs. A series of three intraperitoneal injections of infectious purified rotaviruses was given after collection of preimmune guinea-pig (Hartley strain) serum by heart puncture. The first injection was with complete Freund's adjuvant, the second with incomplete Freund's adjuvant, and the third without adjuvant. The injections were separated by 3-week intervals, and 0-2 to 0.5 ml containing at least 5 x 106 infectious viruses was used on each injection. Animals were bled by heart puncture beginning 1 week following the final injection. Sera were heat-inactivated (56 C, 30 min) and stored at -20 C. In]ectivity (plaque) and neutralization assays. Confluent monolayers of MA-104 cells in 60 mm tissue culture plates were washed twice with Earle's balanced salt solution (EBSS) and inoculated with 0.2 ml of virus diluted in EBSS. After a 1 h adsorption at 37 C, 5 ml of overlay medium was added (Special MEM with antibiotics and 2 lag trypsin, 25 lag DEAE-dextran, and 1.8 mg agarose per ml) and incubation was continued at 37 C. After 3 to 5 days, the soft agarose overlay was poured from the plates and the cells were stained with crystal violet solution to help visualize plaques. Neutralization of virus by hyperimmune guinea-pig antiserum was performed by mixing equal volumes of diluted virus and serum followed by a 30 min incubation at 37 C. The dilution buffer was EBSS which contained 10~ tryptose phosphate broth to stabilize the viruses against heat inactivation. After neutralization, virus survival was determined by the plaque assay. Antiserum titres were expressed as the reciprocals of the dilutions required to neutralize 60~ of infectious viruses. Selection ofserotypic rariants of SA-11. An SA-I 1 isolate obtained after triple plaque purification was used to prepare hyperimmune sera in two guinea-pigs (787 and 793). The respective titres of these sera were and Antiserum 787 was used in the selection procedure. The initial selection step was to mix an equal volume of a lysate of the plaque-purified virus (5 x 108 p.f.u./ml) with a 1 : 100 dilution of antiserum and to allow the majority of infectious particles to be neutralized during 30 min at 37 C. Survivors were isolated by plaquing and 24 were picked for further selection. After growth in MA-104 cells, lysates of each were again neutralized with the same concentration of antiserum and survivors re-grown. This process was repeated for 22 passages using antiserum 787, a point at which no further selection appeared to have occurred for several passages as determined by plaque neutralization. The process was then repeated for an additional 16 passages with an antiserum against the parent of the plaque-purified isolate. The second antiserum was chosen because it had a very high titre ( ) and was made against non-plaque-purified SA-I 1 which theoretically could have a greater variety of immunodominant neutralization epitopes. Variants selected with this antiserum would potentially be less likely to contain any of the usual immunodominant SA-I1 neutralization epitopes. After 39 total passages, plaque-purified isolates were obtained from each of the 24 cultures. These were analysed to determine the effect of the selection procedure. RESULTS Analysis of serotypic variants of SA-11 The 24 plaque-purified isolates of SA-11 obtained after 39 selective passages following treatment with anti-sa-11 sera were analysed to determine their serotypic relationships to the

3 Table 1. Antigenicity of rotavirus SA-11 Neutralization titres of anti-sa-11 sera against parental virus and variants Virus Parent Variant (V) V01 V 02 V 03 V 04 V 05 V 06 V 07 V O8 V O9 V 10 VII V12 V13 Vl4 Vl5 V16 Vl7 VI8 V19 V 20 V21 V 22 V 23 V 24 Antiserum titre* A Antiserum 787 Antiserum <100 (> 128)t 800(32) 100 (128) 400(64) 100 (128) 800 (32) 200 (64) 800 (32) 200 (64) 400(64) 1600 (8) 1600 (16) < 100 (> 128) 1600 (16) 100 (128) 800(32) 200 (64) 400(64) 200 (64) 800 (32) < 100 (> 128) 400 (64) 100 (128) 800(32) 200 (64) 800(32) 100 (128) 800 (32) 200 (64) 800 (32) * Antiserum titres are expressed as the reciprocals of the serum dilution required to neutralize at least 60~ of infectious viruses as determined by plaque assay. I" Numbers in parentheses are the ratios of homologous/heterologous titres parental virus. The initial antiserum preparations used for these neutralization assays were against the plaque-purified parent. The dilution of antiserum 787 (used to select variants) necessary to cause comparable neutralization of variants and parents differed by at least 64-fold for all but one of the variants (no. 10) which differed by eightfold (Table l). Thus, 23 out of 24 variants were recognized as serotypes different from the parent serotype by this one-way test, i.e. antibody titres against the homologous and heterologous viruses differed by 20-fold or more (Hoshino et al., 1984; Wyatt et al., 1983). A similar result was obtained when a second antiserum against the plaque-purified parent (antiserum 793) was tested (Table 1). This antiserum had not been used in the selection procedure. The differences in homologous and heterologous antibody titres were generally somewhat less with this antiserum than with the antiserum used for selection. This is probably due to slight differences in antibody response in the two hyperimmunized guinea-pigs. However, 22 of the 24 variants were still recognized as serotypes different from the parent. To determine whether the variants could be truly classified as serotypes different from the parent serotype, the two-way cross-neutralization experiments were completed. That is, the neutralization titres of antisera from guinea-pigs hyperimmunized with variant viruses were determined for the homologous viruses and the parent. Six of the 24 variants were used for this study, and two guinea-pigs were hyperimmunized with each. In all but one case (antiserum 714), the antisera recognized the variants and parent as the same serotype (Table 2). Furthermore, the homologous/heterologous titres of one-half of these antisera differed by less than 2. As shown above, antisera against the parental SA-11 failed to recognize all six of these variant viruses as the same serotype. Thus, immunodominant neutralization epitopes in the parental virus differed from those of the variants. This could cause the serotypic relationships between the variants and

4 2378 D.R. KNOWLTON AND R. L. WARD Table 2. Neutralization titres of antisera made against SA-11 variants and tested against the homologous variants and the parental SA-I I Antiserum against variant (V) Antiserum titres against homologous variant or parent* 5 Variant Antiserum Variant (V) Parent (SA-11) V/SA-11 V V V V V V * Antiserum titres are the reciprocals of the dilutions required to neutralize 60% of infectious viruses as determined by graphic interpolation of plaque assays, other strains of rotavirus to differ from those observed with the parental SA-11. This possibility was examined. Serotypic relationships of SA-11 and variants with other rotavirus strains SA-11 is a simian rotavirus that belongs to the same serotype as certain canine and equine strains as well as human rotaviruses classified as serotype 3 (Hoshino et al., 1984). One of the human rotavirus representatives of serotype 3 (strain P) was tested for its serotypic relatedness to the SA-11 parent and its variants. Antisera against the parental SA-11 recognized P as the same serotype (i.e. homologous/heterologous titres < 20) and one antiserum against P (antiserum 705) recognized the parental SA-11 as the same serotype while a second antiserum (706) recognized SA-11 as a different but closely related serotype (Table 3). In contrast, 11 out of the 12 antisera against the six variants had homologous/heterologous titres against P of greater than 40. Thus, only one (antiserum 721) recognized P as the same or even a closely related serotype. Likewise, antisera against P neutralized the parental SA-11 at least five times more efficiently than they neutralized the variants. Differences in serotypic relatedness of the SA-11 parent and its variants to rotaviruses belonging to other serotypes were also measured. Three of the rotaviruses tested, all from different serotypes, were poorly neutralized by antisera against both the SA-11 parent and variants (results not shown). These included representatives of human serotypes 1 (Wa) and 4 (ST3) and a porcine rotavirus (OSU). A different observation was made when crossneutralization experiments were conducted with NCDV, a bovine rotavirus. Antisera against the parental SA-11 poorly neutralized NCDV (i.e. homologous/heterologous titres >60) as shown in Table 3. Likewise, antisera against NCDV failed to neutralize either the parental SA- 11 or the variants at the dilutions tested (i.e. homologous/heterologous titres > 29). However, all but one of the antisera made against the variants (antiserum 714) neutralized N CDV better than did antisera against the parent and five out of 12 of these antisera did not distinguish NCDV as a different serotype. DISCUSSION It has been shown that RNA viruses contain major neutralization epitopes which can mutate at a high rate to a state of neutralization resistance (Portner et al., 1980). Viruses altered in this

5 Table 3. Serotype relationships of parental SA-11 and variants with rotavirus strains P and NCDV NCDV < OOO 4OOOO 33OOO ' ~D Antiserum against strain P, parental SA-11, variant (V) or NCDV r ~ F Virus Antiserum P P * SA-I V < V <200 V < V < V < V < NCDV Antiserum titre against different strains of rotavirus SA-I1 V01 V 04 V 15 V 18 V 19 V < 800 < 800 < 800 < 800 < < 1600 < 1600 < 1600 < 1600 < < < 100 < < 200 < 500 < 500 < 500 < 500 < 500 < 500 <200 < 1000 < 1000 < 1000 < 1000 < 1000 < 1000 < 200 < 1000 < 1000 < 1000 < 1000 < 1000 < 1000 < 200 < 1000 < 1000 < 1000 < 1000 < 1000 < 1000 * Homologous reactions are in bold type.

6 2380 D.R. KNOWLTON AND R. L. WARD manner at a sufficient number of sites should escape detection by an immune system primed to recognize the parental strain. In the same way, viruses belonging to different serotypes should also escape detection and may be able to initiate new cycles of infection. The only limitation in number of serotypes may be the lethality of certain changes. Acceptable changes appear to be much more limited with some viruses than others as reflected by the number of their serotypes. At least seven distinct serotypes of rotavirus have been identified using hyperimmune sera in two-way cross-neutralization experiments (Hoshino et al., 1984). This implies that the immunodominant neutralization epitopes in these viruses are antigenically different. Alteration of these epitopes through mutation could change the original serotypic relationships. Virus strains that originally shared immunodominant epitopes may no longer do so and new immunodominant epitopes may be shared by viruses that previously belonged to different serotypes. These new dominant epitopes could be alterations of the old dominant epitopes or epitopes that were immunorecessive in the original virus. As shown in this report, rotavirus strain SA-11 does contain readily mutable dominant neutralization epitopes. Variants selected for these mutations were poorly neutralized by antisera made against the parental virus. However, antisera made against the variants in hyperimmunized guinea-pigs readily neutralized the parental SA-11 in 11 out of 12 cases. One explanation for these results is that the mutational events caused immunorecessive neutralization epitopes in parental SA-11 to become immunodominant in the variants. The new immunodominant epitopes could either contain part of the original dominant epitopes or be totally new epitopes. Another explanation was proposed by Fasekas de St. Groth (1975) who made similar observations with influenza viruses. He suggested that mutants selected through the use of polyclonal antiserum should contain larger amino acids in their antigenic sites than their predecessor. Antibodies made against these mutants should also neutralize the predecessors but antibodies against the predecessors should not neutralize the mutants because of steric hindrance at the antibody-binding sites. Other explanations are also possible. Changes in immunodominant neutralization epitopes caused the variants to have different serotypic relationships with other rotavirus strains from those of the parental SA-11. We and others (Hoshino et al., 1984) have shown SA-11 can be classified as the same serotype as the human strain P. This relationship was lost in the variants as determined by two-way crossneutralization experiments. The only exception noted was with one antiserum against variant 22 (antiserum 721) which readily neutralized strain P. Because this antiserum neutralized the homologous virus and the parental SA-11 equally, it apparently contained neutralizing antibodies against an epitope(s) that was similar if not the same in the parent, the variant and P. From these results, we conclude that the variants contained at least some mutations in immunodominant neutralization epitopes common to both SA-I 1 and P. Alterations in immunodominant epitopes of SA-11 also changed its serotypic relatedness to the bovine rotavirus NCDV, but in the opposite direction. Two-way cross-neutralization experiments demonstrated that the parental SA-11 and NCDV were different serotypes and antisera against NCDV also failed to neutralize the variants. However, antisera against the variants recognized NCDV as the same serotype in five out of 12 cases and as a closely related serotype in six other cases. The exception was antiserum 714 which also poorly neutralized the parental SA-11 and, therefore, must have contained neutralizing antibodies directed against antigenic determinants altered by mutation. The results found with the other 11 antisera imply that some of the immunodominant neutralization epitopes in the variants were present in SA-11 and NCDV, but were immunorecessive in both. This conclusion supports the explanation that immunodominant neutralization epitopes in the variants were primarily immunorecessive in the parental SA-11. If this conclusion can be extended, it is possible that most, if not all, serotypes of rotavirus share neutralization epitopes but many remain immunorecessive. If certain of these shared sites cannot be altered without loss of viability, as has been suggested for other viruses (Emini et al., 1984), production of antibodies against determinants in these sites may provide lasting protection against all rotaviruses that share these same determinants. The implications of this possibility are very important in production of useful vaccines against rotaviruses.

7 Antigenicity of rotavirus SA REFERENCES BOHL, E. H., THEtL, K. W. & SAIF, L. J. (1984). Isolation and serotyping of porcine rotaviruses and antigenic comparison with other rotaviruses. Journal of Clinical Microbiology 19, CLEMENTS, J. E. & NARAYAN, O. (1984). Immune selection of virus variants. In Concepts In Viral Pathogenesis, pp Edited by A. L. Notkins & M. B. A. Oldstone. Wien & New York: Springer-Verlag. CUKOR, G. & BLACKLOW, N. R. (1984). Human viral gastroenteritis. Microbiological Reviews 48, DUPONT, H. L. (1984). Rotavirus gastroenteritis - some recent developments. Journal of Infectious Diseases 149, EMINI, E. A., BRADFORD, J. A., LEWIS, A. J., LARSEN, G. R. & WIMMER, E. (1982). Poliovirus neutralization epitopes: analysis and localization with neutralizing monoclonal antibodies. Journal of Virology 43, EMINI, E. A., JAMESON, B. A. & WIMMER, E. (1984). Identification of a new neutralization antigenic site on poliovirus coat protein VP2. Journal of Virology 52, FASEKAS DE ST. GROTH, S. (1975). The phylogeny of influenza. In Negative Strand Viruses, vol. 2, pp Edited by B. W. J. Mahy & R. D. Barry. New York: Academic Press. GAUL, S. K., SIMPSON, T. F., WOODE, G. N. & FULTON, R. W. (1982). Antigenic relationships among some animal rotaviruses: virus neutralization in vitro and cross-protection in piglets. Journal of Clinical Microbiology 16, GERHARD, W., YEWDELL, J., FRANKEL, M. E. & WEBSTER, R. (1981). Antigenic structure of influenza virus hemagglutinin defined by hybridoma antibodies. Nature, London 240, HOSHINO, Y., WYATT, R. G., GREENBERG, H. B., FLORES, J. & KAP1KIAN, A. Z. (1984). Serotypic similarity and diversity of rotaviruses of mammalian and avian origin as studied by plaque-reduction neutralization. Journal of lnjectious Diseases 149, PORTNER, A., WEBSTER, R. G. & BEAN, W. J. (1980). Similar frequencies of antigenic variants in Sendai, vesicular stomatitis, and influenza A viruses. Virology 104, VESIKARI, T., ISOLAURI, D., D'HONDT, E., DELEM, A., ANDRE, F. E. & ZISSIS, G. (1984). Protection of infants against rotavirus diarrhoea, by RIT 4237 attenuated bovine rotavirus strain vaccine. Lancet i, WOODE, G. N., KELSO, N. E., SIMPSON, T. F., GAUL, S. K., EVANS, L. E. & BABIUK, L. (1983). Antigenic relationships among some bovine rotaviruses: serum neutralization and cross-protection in gnotobiotic calves. Journal of Clinical Microbiology 18, WYATT, R. G., JAMES, H. D., JR, PITTMAN, A. L., HOSHINO, Y., GREENBERG, H. B., KALICA, A. R., FLORES, J. & KAPIKIAN, A. Z. (1983). Direct isolation in cell culture of human rotaviruses and their characterization into four serotypes. Journal of Clinical Microbiology 18, (Received 27 March 1985)