Replication of avian influenza viruses in humans.

Volunteers inoculated with avian influenza viruses belonging to subtypes currently circulating in humans (H1N1 and H3N2) were largely refractory to infection. However 11 out of 40 volunteers inoculated with the avian subtypes, H4N8, H6N1, and H10N7, shed virus and had mild clinical symptoms: they did not produce a detectable antibody response. This was presumably because virus multiplication was limited and insufficient to stimulate a detectable primary immune response. Avian influenza viruses comprise hemagglutinin (HA) subtypes 1-14 and it is possible that HA genes not so far seen in humans could enter the human influenza virus gene pool through reassortment between avian and circulating human viruses.

Plaquing characteristics of influenza A virus recombinants of defined genetic composition. Brief report.

The plaque size and morphology of twenty-two influenza A virus recombinants representing seven distinct families were analyzed on MDCK cells. By examination of the genetic composition of the recombinants no relationship could be established between any gene, including those coding for the surface antigens, and the plaque size and morphology.

Human trials with wild-type H1N1 and recombinant H3N2-H1N1 influenza A viruses of 1977-1978.

A series of trials was conducted in which wild-type A/USSR/90/77 (H1N1) influenza A virus and a few of its antigenic variants were inoculated into volunteers. Infections readily occurred in people of all ages who had initial low antibody titers, but clinical effects were generally mild in comparison with those of the previously tested subtypes, H0N1, H1N1, H2N2, H3N2. There was, however, an inverse relationship between severity of symptoms and age of volunteers, although the incidence of virus excretion and of increase in anti-hemagglutinin was apparently not age related. Naturally occurring recombinant viruses with H3 hemagglutinin and one or more genes of A/USSR/098/77-like strains were likewise studied in volunteers. These clones also produced mild symptoms, providing evidence of an attenuating effect on H3N2 viruses by the substitution of some of its genes with the genes of an H1N1 virus.

A re-examination of the single radial haemolysis technique for the assay of influenza anti-neuraminidase antibodies in human sera.

A new study is described of the use of single radial haemolysis (SRH) for the measurement of antibodies to influenza virus neuraminidase (NA). The technique is known to be consistently successful in the assay of anti-haemagglutinin (HA) antibodies, subject only to the condition that the indicator virus belongs to an appropriate serotype. Its adaptation to the measurement of anti-NA is, however, more difficult. The virus used must be a recombinant which contains a specific NA and an "irrelevant" HA. However the present experiments showed that the two recombinants MRC-3 and X-38, which contain the same NA but a different HA, gave different results. Other properties of recombinants, including rates of attachment to and elution from red cells, may affect the results. The chemical NA-inhibition test (NI), although requiring the use of antigenic hybrids, did not produce these discrepancies. However it appears possible to exploit the simplicity and convenience of SRH for mass survey of anti-Na, if individual hybrid recombinants can first be shown to yield results comparable to those obtained by NI.

Further studies in man of Hsw1N1 influenza viruses.

Two subpopulations of antigenically different Hsw1N1 influenza viruses, cloned from 'swine' New Jersey virus 1976, were individually inoculated into antibody-free volunteers. One clone contained a haemagglutinin so far seen only in swine viruses prevalent in 1971 and after, the other a haemagglutinin of earlier strains dating back to a least 1957. Each of the viruses was infectious for man and intermediate in human virulence between a wild human virus and swine pathogens of 1966 and 1967, which had earlier been tested in man. Antigenically comparable clones segregated from viruses recovered in Wisconsin from a pig and a human contact, respectively, were also infectious for man; however, they were less virulent than their New Jersey counterparts. Differences between the Wisconsin clones themselves were small, but there was an indication of a relationship between human passage. and human virulence. There was no evidence of inherently greater human virulence in the newer Hsw1N1 serotype as compared with the earlier serotype. Hence, recent detection of swine influenza-like viruses in man is unlikely to be the consequence of a host-range mutation concurrent with an antigenic mutation. We believe that the Hsw1N1 viruses recovered from the human influenza outbreak at Fort Dix, New Jersey, and from recent single human infections were wholly derived from enzootic swine viruses that underwent limited human adaptation through man-to-man passage.

Trials of live influenza A recombinants in man during natural antigenic change in 1971-1976.

A long-term study is described of human trials with live recombinants derived from A/PR/8/34 (H0 N1) and successive virulent H3 N2 viruses. A/PR/8/34 was noninfectious for man and the H3 N2 strains all induced similar influenza-like illness. In each recombinant experiment, some of the progeny were adequately attenuated and potentially immunogenic. In addition, there appeared to be evidence of probable cross-protection against viruses with related but non-identical haemagglutinins. However, the degree of cross-protection depended upon the time interval between the appearance of the epidemic and vaccine viruses. Recombinants of A/PR/8/34 and a human Hsw1 strain of low virulence did not function satisfactorily, and it seemed that the suitability of A/PR/8/34 as a 'master' parent depended upon the presence of standard genetic properties in the wild parents. The replacement of A/PR/8/34 with the partially virulent virus, X-31 (H3 N2), produced a live vaccine which was infectious but poorly antigenic. A universal master strain for live influenza vaccine recombinants is probably not yet a practical possibility.

Formation of antibody to matrix protein in experimental human influenza A virus infections.

Antibodies to type A influenza virus matrix protein (M) were assayed by single radial diffusion in 180 paired sera of volunteers challenged intranasally with live H3N2 viruses of varying degrees of virulence. Of these volunteers 20 had had severe clinical reactions (influenza-like); there had been 19 moderate reactions (lesser degrees of constitutional illness), and the remaining 141 reactions had been graded mild, very mild, or nil. Only 2 volunteers were shown to have antibodies to M in the pre-trial serum samples, and 11 developed anti-M rises after virus inoculation. Nine of the 11 had had severe reactions, and 2 had had moderate reactions. There was, therefore, a clear correlation between severity of clinical illness and anti-M antibody formation. In general, anti-M increases coincided with increases to the hemagglutinins and nucleoprotein, and with virus shedding. However, no anti-M antibody could be demonstrated in paired sera of 18 additional volunteers of whom 12 had developed severe reactions after the inoculation of virulent H0N1 and H1N1 influenza A viruses and of whom 12 had shown laboratory evidence of infection.

Experiments in the preparation of live influenza B vaccines.

A study is described of the selection of influenza B virus recombinants. Three virulent viruses isolated in 1970, 1970 and 1973, were crossed with host-range mutants of low virulence for man, which had originally been isolated in 1940, 1959 and 1956, respectively. Nine presumptive recombinants were inoculated into volunteers with low initial HI antibody titres. Although a number proved attenuated and there was evidence of high frequency of recombination, antigenic characterization of neuraminidases proved difficult. There was always some crossing between parental surface antigens, and appropriate antigenic hybrids for preparation of reference antisera were not available. Comparative virus titres at 33 degrees C and 38 degrees C were of limited value as a genetic marker. It is suggested that biological methods are unsatisfactory for the study of influenza B virus recombinants, and should be replaced by biochemical techniques designed to trace the parental origins of individual RNA segments.

Influenza A neuraminidase antibodies in children and young adults studied by serum absorption.

A study is described of influenza A anti-neuraminidase antibodies in the sera of young people of three different groups. Each serum was individually absorbed with viruses containing the N2 neuraminidases of 1957, 1968 and 1972. Rabbit antisera prepared against the viruses were similarly absorbed. Results obtained with the animal sera suggested that these neuraminidases were antigenically distinct, but the human sera had a broader range of anti-neuraminidase activity and gave indication of asymmetric antigenic relationships. Earlier workers who surveyed anti-haemagglutinin antibodies reported that the virus of primary infection absorbed all antibodies, and the virus of secondary infection only those directed against itself. We too found that the virus of secondary infection absorbed only homologous anti-neuraminidase antibody. However, although the primary infecting virus did absorb some secondary antibody, this absorption was incomplete and it lessened with the lengthening of the time interval between the primary and secondary infecting viruses. A similar pattern was seen with anti-haemagglutinin antibodies. Absorption of anti-neuraminidase antibodies from human sera proved much more difficult than absorption of anti-haemagglutinin antibodies particularly after repeated influenza virus infections. The relative rarity of antigenic shift in the neuraminidase subunit also creates problems in the interpretation of results of serum neuraminidase antibody surveys.

Selection of influenza B virus recombinants and their testing in humans for attenuation and immunogenicity.

The selection of influenza B virus recombinants from plaques in bovine kidney cell monolayers is described. Two sets of recombinants were each derived from parents of high and low virulence for humans, respectively. Recombination frequency was apparently high, and reassortment of genes made it possible to obtain attenuated recombinants containing the surface antigens of the virulent parents. Attenuation and immunogenicity were demonstrated in a series of volunteer trials. However, technique proved less satisfactory than for influenza A viruses which periodically undergo antigenic shift and for which there is a wide choice of parent viruses with distinctive surface antigens. In our two influenza B recombinant series there was appreciable antigenic overlap in the neuraminidases of the parents, even though in both cases these were chronologically widely separated. Another marker used was comparative titer at 35 and 38 degrees C. In practice, technical problems might sometimes make it difficult to ensure rapid production of live influenza B vaccines by recombination.

RNAs of influenza virus recombinants derived from parents of known virulence for man.

Extensive use of recombinants made from A/PR/8/34 (H0N1) and wild, virulent H3N2 viruses as live influenza vaccines has provided a number of viruses of defined virulence for man. Clinical symptoms produced by these strains have ranged from febrile influenza to local coryzal symptoms or nil. A study was therefore made of the extent to which the PR8 genome had been incorporated into that of a number of the recombinants. By RNA--RNA hybridization it seemed that recombinants which had 55 per cent of greater homology with the PR8 parent were likely to conform an acceptable standard of attenuation. Those with lesser homology were frequently, but not always, clinically virulent. The technique seemed potentially useful, therefore, for screening PR8 live vaccine recombinants in vitro before giving them to volunteers.

Virulence for man of a human influenza-A virus antigenically similar to "classical" swine viruses.

The newly isolated human influenza-A strain containing swine antigens isolated in New Jersey, U.S.A., was inoculated into six volunteers. Clinical reactions were mild although all volunteers were infected. The longest period for which the virus was excreted was 8 days and the shortest 3 days. In its virulence for man the New Jersey strain was intermediate between a human and animal virus, and was quite clearly more virulent than known swine viruses. It seems possible that the outbreak in the U.S.A. was an isolated event and that the virus will not become established in man.

Measurement of antibody to influenza virus neuraminidase by single radial hemolysis in agarose gels.

A simple method of assaying anti-influenza neuraminidase antibodies in human sera was described. Suitable antigenic hybrid viruses were adsorbed to sheep erythrocytes, which were then incorporated into agarose gels. When sera were introduced into wells cut in the gels, zones of hemolysis were observed in the neighborhood of those containing neuraminidase antibodies. There was a direct relationship between zone size and antibody titer. No purification of adsorbed viruses was necessary. The test was rapid, required very simple reagents, gave results that agreed well with those given by conventional techniques, and appeared to be the most sensitive of four methods evaluated. Studies of cross-reactions by hyperimmune sera against homologous and heterologous neuraminidases and of absorption of neuraminidase antibodies from human sera indicated a high degree of specificity. The technique seems to be suitable for large-scale epidemiological investigations.

Trials in man with live recombinants made from A/PR/8/34 (H0 N1) and wild H3 N2 influenza viruses.

A long-term study is described of recombinant influenza viruses produced from the avirulent laboratory strain, A/PR/8/34 (H0 N1), and newly isolated H3 N2 influenza virus variants. A number of H3 H2 recombinants were found to be attenuated for man and capable of inducing antibody formation, and were therefore potentially usable as live vaccines. However, the volunteer trials as a whole suggested that, in this system, there might not be complete segregation of virulence and antigenic characteristics. No H3 N2 recombinants were detected which were non-infective for man, like the A/PR8 (H0 N1) parent, and reciprocal hybrids (H3 N1 and H0 N2) always reflected some of the virulence of the parent from which they had inherited their haemagglutinin. This property is not a feature of mouse influenza.