Serological studies with influenza A(H1N1) viruses cultivated in eggs or in a canine kidney cell line (MDCK)

Bulletin of the World Health Organization, 65 (2): 181-187 (1987) © World Health Organization 1987 Serological studies with influenza A(H1Ni) viruses cultivated in eggs or in a canine kidney cell line (MDCK) J. S. OXFORD,' T. CORCORAN,' R. KNOTT,' J. BATES,' 0. BARTOLOMEI,2 D. MAJOR, R. W. NEWMAN,' P. YATES,' J. ROBERTSON,1 R. G. WEBSTER,3 & G. C. SCHILD' Pairs of influenza A(HINI) viruses cultivated from the same clinical specimen in canine kidney (MDCK) cells or in embryonated hens' eggs can frequently be distinguished by their reactions with monoclonal antibodies to haemagglutinin and with antibodies in ferret or human sera. Egg-adapted virus, further passaged in MDCK cultures remained "egg-like" in serological characteristics indicating that the differences in their serological reactions were not a direct result of host cell-dependent glycosylation of the haem- agglutinin. Haemagglutination-inhibiting (HI) or virus neutralizing antibodies in human sera can be detected more frequently, and to higher titre, in tests employing virus grown exclusively in MDCK cells than in tests with virus adapted to growth in embryonated eggs. Striking differences were detected in the serological reactions in HI tests when sera from ferrets infected with egg-grown virus were tested against a series of strains of influenza A(HJNI) virus isolated in 1983 and adapted to growth in eggs. In contrast, sera from ferrets infected with MDCK-derived virus failed to distinguish serologically between the same viruses that had been passaged exclusively in MDCK cells and also revealed relatively small differences between their egg-adapted counterparts. It was concluded that the cell substrate used for virus isolation and cultivation is a factor that should be considered when interpreting the results ofstrain characterization of influenza A(HINI) isolates and in sero-surveys using these viruses. Antigenic analysis of influenza A and B virus strains is most commonly carried out using virus cultivated in embryonated hens' eggs (1). Although human influenza viruses were first isolated in an experimental animal, the ferret (2), and soon after- wards adapted to mice, extensive laboratory work only became possible with the observation that viruses could be isolated in the amniotic cavity of embryonated hens' eggs (3). In the absence of tissue culture systems at that time, the embryonated hens' eggs became the established method for cultivating influenza viruses. Subsequently Burnet demonstrated that human isolates of influenza virus were mixtures of variants with differing biological properties and that the complex environment of the amnion or allantoic cavity of the egg exerted differing selective pressures (reviewed in 4). Additional selective ' Division of Viral Products, National Institute for Biological Standards and Control, Holly Hill, Hampstead, London NW3 6RB, England. Requests for reprints should be sent to Dr J. S. Oxford at this address. 2 Institute of Microbiology, University of Florence, Florence, Italy. 3 Division of Virology and Molecular Biology, St Jude Children's Hospital, Memphis, TN, USA. pressures that could result in a degree of separation of variants include naturally occurring alpha, beta and gamma inhibitors (S) and, of course, specific antibody to the virus itself (6). More recently we have presented evidence for the host-cell selection of influenza B virus (7). Viruses isolated from clinical specimens and serially cultivated in mammalian cells (a canine kidney cell line, MDCK), possessed haemagglutinins (HA) that were antigenically dis- tinguishable from those of viruses grown from the same source but in embryonated hens' eggs. Differences in the amino acid sequence of the HA of influenza B virus cultivated in MDCK cells and in eggs have been identified (8). We concluded that adaptation of influenza B virus to growth in eggs selected a virus subpopulation (4, 6, 9-11) which was antigenically and biochemically distinct from virus from the same clinical source but passaged exclusively in MDCK cells. In the present paper we describe marked differences in the serological characteristics in both haemagglutination inhibition (HI) and virus neutral- ization tests with monoclonal anti-HA antibodies of influenza A(HlNl) viruses isolated from the same 4766 -181- 182 J. S. OXFORD ET AL. clinical specimens in MDCK cells or in eggs. Moreover, in HI or virus neutralization tests with human and ferret sera, virus cultivated exclusively in MDCK cells detected antibody more frequently and at higher titre than the corresponding egg-adapted virus. Our findings suggest that interpretation of serological and antigenic analyses of influenza A(HINI) viruses, like those for influenza B viruses (7), may be complicated by selection of antigenic variants during virus cultivation in different host-cell systems. The observations may also have significance for our understanding of the antigenic structure and variation of influenza virus haemagglutinin (12) and immune responses to infection and immunization. MATERIALS AND METHODS Virus isolation The influenza A viruses studied were isolated from an influenza outbreak in a residential school, Christ Hospital School, in February 1983 (40 strains designated A/Chr/83), and from Florence (2 strains). Most viruses were isolated in Madin-Darby Canine Kidney Cell (MDCK) cultures and further passaged in these cells maintained in Eagle's minimal essential medium containing TPCK trypsin (0.2 ,g/ml). For isolation of virus in eggs, 0.1 ml of a virus suspension from a throat swab or from a primary isolation in MDCK cells was inoculated into the amniotic cavities of eleven-day old embryonated hens' eggs. After incubation at 33 °C, for 48 hours, the allantoic and amniotic fluids were harvested, tested for the presence of virus, and stored at -70 °C ("egg- virus"). Egg-adapted virus isolates were further passaged twice in the allantoic cavity and the allantoic fluids used as antigen for serological H1 tests. Monoclonal antibodies and serological tests Monoclonal antibodies to the HA of A/USSR/92/77, A/Chr/91/83, A/Brazil/ 11/78, A/England/333/80 and A/Baylor/5700/82 (HINl) viruses were prepared using standard procedures (13, 14). In brief, mice were immunized twice with purified influenza A virus prepared in eggs and spleen cells removed 4 days after the second booster dose of antigen, which was given intravenously. Antibody was prepared as mouse ascitic fluids and used for the HI tests after overnight treatment at 37 OC with 4 volumes of receptor-destroying enzymea to remove any non-specific inhibitors. The micro-HI test with 96-well microtitration plates b was used and the challenge dose of virus was carefully standardized by a From Philips, Duphar B. V., Amsterdam, Netherlands. b Linbro plates supplied by Flow Laboratories, Woodcock Hill, Herts., England. repeat titration to 8 HA units. In each experiment the homologous viruses were included as a control and reproducibility of HI titres was established by experiment. Single radial haemolysis was carried out as described previously (15). Neutralization experi- ments were carried out by mixing approximately 100 TCID5o of virus with varying dilutions of monoclonal antibody or human serum for 1 hour at 37 °C and testing the resultant antibody-virus mixtures for residual infective virus in MDCK cells. Post-infection human andferret sera Sera were obtained from non-immunized children and adults (aged 2-45 years) in the United Kingdom in November 1983. The antibody detected in the sera was assumed to result from natural influenza A(HlNI) infection during or after the time the viruses were isolated at Christ Hospital School. Ferret sera were obtained from animals infected intranasally with egg-grown or MDCK cell-grown influenza A/Chr/157/83 (HINI), A/Brazil/78 or A/Chile/1/83 viruses and bled 10 days later. RESULTS Antigenic analysis of MDCK-cell-grown and egg- grown influenza A(HINI) viruses using a panel of anti-HA monoclonal antibodies Results of HI tests on nine representative influenza A(HlN1) viruses adapted to growth in eggs or cultivated exclusively in MDCK cells are presented in Table 1. Marked differences were noted in the HI reactions of the viruses against a large panel of monoclonal antibodies dependent upon the passage histories of the viruses. In all, 63 viruses were tested; the majority, 53 of them, could be distinguished antigenically when corresponding viruses adapted to growth in eggs or cultivated exclusively in MDCK cells were compared. Three patterns of serological reactions were discerned for the monoclonal antibodies. Three of the twelve antibodies reacted exclusively with egg- adapted virus (Br29, E23 and B3). In contrast, other antibodies (E61 and B7) reacted with virus cultivated exclusively in MDCK cells but not egg-adapted virus for some isolates, and for other isolates they reacted with virus grown in both substrates. A third group of antibodies reacted equally well with both egg-adapted and MDCK-cell-derived virus (e.g., Br2). Egg- or MDCK-derived A/Chr/157/83 and A/Chr/91/83 viruses, used in the subsequent serological studies with human and ferret sera (described below), were clearly distinguished with monoclonal antibodies E61, B3 and B7. SEROLOGY OF INFLUENZA A(HINI) VIRUSES Table 1. Serological analysis of MDCK-cell-grown and egg-grown influenza A(H1N1) viruses using monoclonal antibodies to HA Viruses and monoclonal antibodies A/USSR/0092/77 A/Brazil/11 /78 A/England/333/80 A/Baylor/5700/82 Virus and host cell used for cultivation' U22 U70 UW18 Br2 Br29 E23 E58 E61 B1 B3 B7 A/USSR/0092/77 E 1600 6400 3200 > 12800 > 12800 6400 < 100 1600 800 < 100 < 100 A/Brazil/ 1/78 E 1600 12800 1600 > 12800 6400 3200 800 1600 1600 < 100 400 A/lndia/6263/80 E 800 6400 100 6400 3200 400 400 1600 < 100 < 100 < 100 A/England/333/80 E 1600 6400 800 6400 3200 6400 800 3200 1600 < 100 < 100 A/England/403/80 E 1600 6400 < 100 6400 3200 3200 400 1600 800 < 100 1600 A/Hong Kong/2/82 E 1600 3200 400 >12800 6400 <100 800 1600 1600 <100 800 A/Dunedin/27/83 E < 100 < 100 < 100 3200 100 < 100 200 < 100 3200 12800 800 A/Chr/892/83 C < 100 < 100 200 3200 < 100 < 100 800 1600 1600 < 100 6400 E < 100 < 100 < 100 3200 < 100 < 100 400 < loob 3200 < 100 < 100 A/Chr/920/83 C < 100 < 100 200 3200 < 100 < 100 800 1600 1600 < 100 3200 E 400 400 400 6400 1600 b 1600 1600 3200 800 < 100 < 100 A/Chr/922/83 C < 100 < 100 200 3200 < 100 < 100 800 1600 1600 < 100 6400 E 100 < 100 < 100 6400 < 100 < 100 400 < 100 6400 12800 1600 A/Chr/965/83 C < 100 < 100 200 3200 < 100 < 100 400 1600 1600 < 100 6400 E 100 < 100 400 3200 800 3200 400 3200 1600 < 100 < 100 A/Chr/83/83 C < 100 < 100 200 3200 < 100 < 100 800 1600 1600 < 100 3200 E <100 <100 <100 3200 <100 <100 800 <100 <100 <100 <100 A/Chr/91/83 C < 100 < 100 200 3200 < 100 200 400 1600 1600 < 100 3200 E < 100 < 100 < 100 3200 < 100 < 100 200 < 100 3200 12800 < 100. A/Chr/1 57/83 C < 100 < 100 200 3200 < 100 400 800 1600 1600 < 100 3200 E < 100 < 100 < 100 3200 < 100 < 100 1600 < 100 3200 12800 < 100 A/Florence/ 1 3/83 C 800 1600 < 100 12800 NT' < 100 NT 1600 800 < 100 NT E 1600 12800 < 100 12800 NT < 100 NT < 100 1600 < 100 NT A/Florence/ 1 9/83 C 400 800 200 3200 NT < 100 NT 1600 400 < 100 NT E 400 3200 200 6400 NT 1600 NT 1600 400 < 100 NT C = MDCK cell; E= allantoic cavity of embryonated hens' eggs. b The underlined titres indicate significant serological differences between cell- and egg-grown virus. c NT = not tested. A striking finding was that very few strain-to-strain serological differences were observed between differ- ent viruses that had been cultivated exclusively in MDCK cells (Table 1). In contrast, the different isolates that had been adapted to growth in eggs were heterogeneous in their serological reactions with individual monoclonal antibodies. For example, all nine MDCK-derived isolates reacted with mono- clonal antibody E61 but only three of the nine egg- adapted viruses isolated from the same clinical specimens reacted with this antibody. To investigate these differences further, certain viruses were examined serologically using ferret sera from animals infected with either egg-adapted virus or virus cultivated exclusively in MDCK cells. Firstly, it is apparent (Table 2) that sera such as F10/84 or F3/85, from animals infected with MDCK-cell- derived virus, generally do not distinguish sero- logically between the isolates such as A/Chr/ 157/83, A/Chr/91/83 and A/Chr/83 regardless of whether the viruses are cultivated in MDCK cells or in eggs. In contrast, in an HI analysis using sera from ferrets infected with egg-derived virus, antigenic differences are detected among this group of isolates, particularly if the viruses used in the HI test are cultivated in eggs. Thus, serum F6/84 had HI titres of 960 and 20, respectively, when reacted with egg- derived A/Chr/157/83 and A/Chr/83/83 viruses in 183 184 J. S. OXFORD ET AL. Table 2. Hi reactions of post-infection ferret sera with influenza A(H 1 N 1) viruses cultivated exclusively on either MDCK cells or eggs Hi titre with following ferret sera: Ferrets infected Ferrets infected with MDCK-derived with egg-derived A/Chr/l157/83 virus A/Chr/157/83 virus Ferrets infected with egg-derived Virus and host cell used for cultivation' F10/84 F3/85 F6/84 F6/85 A/Brazil/78 virus A/Chile/i /83 virus A/Chr/157/83 C 480 3200 80 240 80 480 A/Chr/91/83 C 640 4800 120 240 40 480 A/Chr/892/83 C 640 4800 120 240 40 480 A/Chr/922/83 C 640 3200 120 240 - - A/Chr/83/83 C 640 3200 120 240 40 480 A/Chr/157/83 E 320 800 960 > 2560 20 60 A/Chr/91/83 E 480 800 480 > 2560 40 120 A/Chr/892/83 E 480 1600 320 640 30 240 A/Chr/922/83 E 480 1200 480 > 2560 - - A/Chr/83/83 E 160 600 20 <20 <20 40 Reference viruses: A/USSR/92/77 E 75 200 < 20 <20 480 160 A/Brazil/ 11/78 E 150 1200 < 20 50 640 240 A/lndia/6263/80 E 600 1200 75 150 80 480 A/Chile/1 /83 E 200 1200 25 50 40 320 C = MDCK cell; E = hens' eggs. Table 3. Haemagglutination inhibition (HI) and virus neutralization titres of influenza A/Chr/91/83 (HiNi) virus derived exclusively from MDCK cells or passaged in eggs HI titre: Neutralization titre: Serological Egg-derived Egg-derived reactivity virus virus with egg- or MDCK- Egg-derived cultivated in MDCK- Egg-derived cultivated in cell-derived cell-derived virus MDCK cells cell-derived virus MDCK cells Antibody virus virus (M6)a (MlE5)a (M 1 E5M 1) virus (M6) (Ml E5) (Ml E5M 1) Monoclonal antibody Br2 Both 3 200 3 200 NT" 16 000 8 000 NT E61 Cell 1 600 <100 <100 6000 <500 <500 E336 Cell 12 800 < 100 < 100 22 000 < 500 < 500 B3 Egg <100 12 800 12 800 <500 64 000 >64 000 Bli Egg < 100 6 400 6 400 < 500 6 000 20 000 Post-infection human serum' 1 - 960 < 10 < 10 2 400 30 < 10 2 - 120 <10 <10 120 15 <10 3 - 240 <10 <10 240 15 <10 4 - 120 <10 <10 320 30 <10 Number of passages indicated in MDCK cells (e.g., M6) and in embryonated hens' eggs (e.g., E5M. NT = not tested. cSera collected in November 1983. SEROLOGY OF INFLUENZA A(HINI) VIRUSES the HI test, whereas the same serum failed to dis- tinguish between the same two viruses when they were cultivated in MDCK cells (HI titres of 80 and 120, respectively). The viruses with different passage histories were also clearly distinguishable in neutralization tests with the same monoclonal antibodies as used in the HI reaction (Table 3). For example, monoclonal antibody E61 neutralized only MDCK-cell-grown virus, whereas antibody B3 neutralized only egg- grown virus. Several of the monoclonal antibodies neutralized viruses grown in both substrates (e.g., Br2). In further experiments egg-grown virus passaged subsequently in MDCK cells showed identical sero- logical reactivity with the monoclonal and polyclonal antibodies (see below) as did virus cultivated exclusively in eggs (Table 3). Therefore, the differ- ences detected above in serological reactions were not a manifestation of host-dependent glycosylation of the HA. Electron microscopy of negatively-stained prep- arations of influenza A/Chr/91/83 (HlNl) virus cultivated in MDCK cells or in eggs failed to detect obvious morphological differences or stages of aggregation which could otherwise provide an explanation of the differing serological reactions. Serological reactivity of human sera with influenza A(HINJ) viruses of different passage history A total of 419 sera from individuals ranging from 2 to 65 years of age were examined using the HI test and viruses which had been grown exclusively in MDCK cells, or in eggs, or which had been adapted to growth in eggs and subsequently passaged in MDCK cells. The data pertaining to a representative 94 sera are shown in Table 4. It was assumed that the antibody in the human sera was generated by natural infection: the sera were taken in the months following the 1983 epidemic. The MDCK-cell-derived influenza A(HlNl) virus A/Chr/91/83 detected HI antibody at considerably higher frequency and titre in the sera than did the corresponding virus adapted to growth in eggs. A/Chr/91/83 virus cultivated exclusively in MDCK cells reacted with 65% of sera with an HI titre of > 1/20, whereas the same virus cultivated in eggs detected antibody in 20% of the same sera (Table 4). Egg-adapted virus which was subsequently culti- vated in MDCK cultures reacted with 16% of sera and thus was similar to the egg-grown virus. In addition, exclusively MDCK-cell-grown virus detec- ted higher HI titres than egg-grown virus and, for example, 30% and 4% of sera had titres >60 when tested using MDCK-cell-grown and egg-grown virus respectively. Essentially identical data were obtained when a group of 212 sera from younger individuals (2-11 years old) was tested against exclusively MDCK-cell or egg-grown A/Chr/157/83 (HINI) virus and also when a group of 60 sera from older individuals (50-65 years old) was tested against exclusively MDCK-cell and egg-grown A/Chr/959/ 83 (HlNl) virus (data not presented). Similarly, neutralizing antibody was detected more frequently and to higher titres using exclusively MDCK-cell-derived virus in the test than using egg-adapted virus. Egg-adapted virus subsequently passaged in MDCK cells reacted similarly to virus grown exclusively in eggs (Table 4). Thus 70% of sera had neutralizing antibody titres of > 20 when analysed using MDCK-cell-grown virus in the assay, whereas only 7%1o and 10% of sera had neutralizing antibody to egg-grown virus and egg-grown virus cultivated in MDCK cells, respectively. In preliminary experiments absorption studies were performed on the adult sera to confirm the specificity Table 4. Frequency and titre of antibody in human sera to A/Chr/91 /83 (H1 N1) virus passaged exclusively in MDCK cells or adapted to growth in eggs Cumulative number of sera (out of 94 tested) b Cumulative number of sera (out of 30 tested) with Origin and with the following HI titres: the following virus neutralization titres: passage history of virus' <20 >20 >60 >120 <20 >20 >60 i120 MDCK (M6) 33(35.1)c 61(64.9) 28(29.8) 10(10.6) 9(30.0) 21(70.0) 7(23.3) 3(10.0) Egg (M1E5) 75(79.8) 19(20.2) 4(4.3) 0 28(93.3) 2(7.0) 0 0 Egg and MDCK (MlE5Ml) 79(84.0) 15(16.0) 3(3.2) 0 27(90.0) 3(10.0) 0 0 a Number of passages indicated in MDCK cells le.g., M6) and in hens' eggs )e.g., E5). bSera collected in the United Kingdom in 1983. ' Figures in parentheses are percentages. 185 J. S. OXFORD ET AL. of reaction with MDCK virus noted above. Ten sera from non-immunized adults whose antibodies were presumably induced by natural infection were absorbed with purified A/Chr/ 157/83 (HIN 1) virus cultivated either in MDCK cells or in eggs and residual antibody investigated using the single radial haemolysis (SRH) technique with MDCK-cell-grown or egg-grown virus in the immunoplate. When the absorbed sera were tested on the SRH plate containing egg-grown virus, no residual antibody was detected. In contrast, when the same sera were tested on the SRH plate containing MDCK-cell-grown virus, haemolysis zones were present after the sera had been adsorbed with egg-grown virus, whereas adsorption with MDCK-cell-grown virus removed all antibody. DISCUSSION Antigenic differences were frequently observed between the haemagglutinins of influenza A(HlNl) viruses that had been isolated from the same clinical sample and cultivated exclusively in MDCK cells or adapted to egg growth. In addition, the viruses grown in different substrates could be clearly distinguished in virus neutralization tests with the monoclonal anti- bodies. Amino acid sequence changes have been detected in the haemagglutinins of influenza A(HlNl) viruses cultivated in eggs or MDCK cells in the present study. They are located on the surface of the globular head of the haemagglutinin molecule adjacent to the receptor binding site (J. Robertson et al, unpublished data). The antigenic differences that have been elucidated here between the haemagglutinins of MDCK-cell and egg-grown influenza A(HlNl) viruses with mono- clonal antibodies have been confirmed using polyclonal human and ferret sera. With human sera, both a higher frequency of antibodies and higher HI and virus neutralization titres were detected to MDCK-cell-grown virus compared to egg-grown influenza A(HlNI) virus. A qualification of the serological data is that neutralization tests were carried out using only MDCK cells and it is possible that similar experiments when carried out in embryonated hens' eggs may give different results. Such experiments are in progress. The serological reactivities of cell- and egg-grown viruses were independent of the last substrate in which the virus was cultivated. Both egg-grown virus and egg-grown virus cultivated in MDCK cells had identical serological characteristics with monoclonal and polyclonal antisera and, antigenically, were distinguishable from the exclusively MDCK-cell- grown virus: thus there could not be a direct result of the host cell of origin of carbohydrate side-chains of the haemagglutinin molecule. Finally, adsorption studies have demonstrated the presence of antibodies in human sera which react exclusively with MDCK- cell-grown virus and not with egg-grown virus. The study has implications for the interpretation of results of sero-epidemiological studies with influenza A(HlNl) viruses, since both the HI and neutral- ization tests are at present routinely performed using egg-grown virus (reviewed in I and 16). In the light of the antigenic differences between the egg- and cell- grown virus populations reported here, comparative sero-surveys which investigate the antibody to influenza A(HlNl) viruses grown in MDCK cells and eggs would seem to be of interest to determine which of these two methods of cultivation would be of greater value in an epidemiological context. Finally, experiments are in progress to investigate the implications of our findings for the design of influenza vaccines. ACKNOWLEDGEMENTS A. Douglas and J. J. Skehel, National Institute for Medical Research, Mill Hill, London, generously provided certain of the monoclonal antibodies used in the present study. J. Smith and E. Grilli (PHLS Laboratories, Guildford) kindly provided the viruses isolated at Christ Hospital School. RESUME ETUDES SEROLOGIQUES REALISEES AVEC LE VIRUS GRIPPAL A(HIN1) CULTIVE SUR CEUFS OU EN LIGNtES DE CELLULES RENALES DE CHIEN (MDCK) cellules renales de chien (MDCK) ou en ceufs de poule embryonnes par leur reaction avec des anticorps mono- On peut frpquemment distinguer des virus grippaux A(HINI) cultives a partir du meme prelevement clinique en 186 SEROLOGY OF INFLUENZA A(HINI) VIRUSES 187 clonaux antihemagglutinine et avec des anticorps de serum de furet ou de serum humain. Les virus avianises, apres passage ulterieur en culture de MDCK, conservent leurs caracteristiques serologiques initiales (de type "ceuf"), ce qui montre que la difference de reactivite serologique ne resulte pas directement de la glycosylation de l'hemag- glutinine effectuee par la cellule hote. Les anticorps neutrali- sants ou les anticorps inhibant l'hemagglutination (HI) presents dans le serum humain peuvent etre deeles plus souvent, et a un titre superieur, par des epreuves utilisant des virus cultives exclusivement en cellules MDCK que lors d'epreuves executees avec des virus adaptes a la replication en ceufs embryonnes. On a observe des differences frappantes en HI lorsqu'on a eprouve, par comparaison avec une serie de souches de virus grippal A(HIN1) isolees en 1983 et avianisees, du serum de furets infectes par des virus cultives sur ceuf. En revanche, le serum de furets infectes par un virus cultive sur MDCK ne permettait pas d'etablir une distinction serologique entre les memes virus cultives exclusivement sur cellules MDCK et ne revelait que des differences relativement faibles entre leurs homologues avianises. On peut conclure de cette etude que le substrat cellulaire utilise pour l'isolement et la culture du virus est un facteur dont il faut tenir compte lorsqu'on interprete les resultats de la caracterisation des souches de virus grippal A(H 1NI) et lors des enquetes serologiques utilisant de tels virus. REFERENCES 1. SCHILD, G. C. & DOWDLE, W. T. Influenza virus characterization and diagnostic serology P315-372. In: Kilbourne, E. D., ed. The influenza viruses and influenza. New York, Academic Press, 1975. 2. SMITH, W. ET AL. A virus obtained from influenza patients. Lancet, 2: 66 (1933). 3. BURNET, F. M. Influenza virus on the developing egg. British journal of experimental pathology, 17: 282 (1936). 4. BURNET, F. M. & CLARKE, E. Walter and Eliza Hall Institute Monograph No 4. Melbourne, Macmillan, 1942. 5. COHEN, A. & BIDDLE, F. The effect of passage in different hosts on the inhibitor sensitivity of an Asian influenza virus strain. Virology, 11: 458-473 (1960). 6. YEWDELL, J. W. ET AL. Antigenic variation in three distinct determinants of an influenza A haemagglutinin molecule. Nature, 279: 246 (1979). 7. SCHILD, G. C. ET AL. Evidence for host cell selection of influenza virus antigenic variants. Nature, 303: 706-709 (1983). 8. ROBERTSON, J. S. ET AL. Alterations in the haemag- glutinin associated with adaptation of influenza B virus to growth in eggs. Virology, 143: 166-174 (1985). 9. KILBOURNE, E. D. Genetic dimorphism in influenza viruses: characterization of stably associated haemag- glutinin mutants differing in antigenicity and biological properties. Proceedings of the National Academy of Sciences (USA), 75: 6258-6262 (1978). 10. HOLLAND, J. ET AL. Rapid evolution of RNA genomes. Science, 215: 1577-1585 (1982). 11. BRAND, C. & PALESE, P. Sequential passage of influenza virus in embryonated eggs or tissue culture: emergence of mutants. Virology, 107: 424-433 (1980). 12. WILEY, D. C. ET AL. Structural identification of the antibody-binding sites of Hong Kong influenza haemagglutinin and their involvement in antigenic variation. Nature, 289: 373-378 (1981). 13. KOHLER, G. & MILSTEIN, C. Continuous culture of fused cells secreting antibody of pre-defined specificity. Nature, 256: 495-497 (1975). 14. Lu BAO-LAN ET AL. Heterogeneity of influenza B viruses. Bulletin of the World Health Organization, 61: 681-687 (1983). 15. OXFORD, J. S. ET AL. Strain specificity of serum antibody to the haemagglutinin of influenza A (H3N2) viruses in children following immunization or natural infection. Journal of hygiene, 86: 17 (1981). 16. STUART-HARRIS, C. H. ET AL. Influenza: the viruses and the disease. London, Edward Arnold, 1985.