General discussion—Session V

General Discussion-Session V DR M. R. HILLEMAN. The paper by Dr Schulman (see page 647) was very impressive: the findings indicate the importance of the neuraminidase antigen in stimulating antibody and immunity to influenza. The results also raise the question of what happens to the antigenicity of the neuraminidase in the course of splitting the virus with ether or with desoxycholate. Is the neuraminidase antigen left intact? We know that inactivated whole virus induces antibody which in turn induces protective immunity. We know also that ether-split and desoxycholate- split vaccines induce haemagglutination-inhibiting antibody. But what is more important is to know about protective efficacy. What is the extent of evidence that ether-split and desoxycholate-split vaccines actually afford protection against influenza and what about the breadth of the protective effect? DR M. F. WARBURTON. Data in my paper (see page 639) showed that disruption with desoxycholate did not destroy the neuraminidase. However, no studies on the actual stimulation by the desoxy- cholate-disrupted vaccines of antineuraminidase antibodies in humans have been performed. Table 3 of my paper showed that protection could be demonstrated after the use of this type of vaccine. DR F. B. BRANDON. Ether treatment generally inac- tivates enzymes but we have no direct information on the ability of the inactivated enzyme to induce antibody. DR M. R. HILLEMAN. That is the point: the exist- ing data are few and limited. It is very important that the protective efficacy of the split vaccines should be established before they are made generally avail- able. We need to be sure about protective efficacy prior to general application or we might be creating further difficulties. DR J. L. SCHULMAN. In answer to the question that Dr Hilleman has raised, it can be said that if mice are challenged with a virus with a haemagglutinin antigen identical to that of the immunizing virus, equivalent protection is observed in animals immunized with intact virus and animals immunized only with the haemagglutinin antigen of that virus. Therefore, the antineuraminidase antibody may be of particular importance when there are appreciable antigenic differences between the haemagglutinin proteins of the challenge virus and those of the virus used for immunization. DR M. R. HILLEMAN. Dr Schulman's remarks are important. Split antigen which matches the epidemic strain induces antibody which would be expected to be protective. But, the real problem- in influenza vaccination is the constant antigenic alteration: it would appear that the neuraminidase antigen might be the only real means of providing the necessary breadth of immune response to take care of these minor changes and keep the vaccine efficacious. DR F. M. DAVENPORT. In response to Dr Hille- man's remarks, it is known that ether-split vaccine does induce antineuraminidase antibodies when given to man. DR J. T. SETO. At the 1965 meeting of the American Society of Microbiologists, we reported that specific A2 virus enzyme antiserum neutralizes the infectivity ofA2/Taiwan mouse-adapted virus in mice. The same antiserum did not significantly neutralize the infec- tivity of A2/Japan/305 virus in embryonated eggs. In studies carried out with Rott and Drzenick in Germany, we observed no neurotoxicity to mice and no cytotoxicity to HeLa cell cultures by A2/Japan/ 305 virus which was propagated in duck embryos. The appropriate control-that is, virus propagated in chicken eggs-was toxic to HeLa cells and to mice. Preliminary findings by E. Chang in our laboratory on the influence of specific X-7 enzyme antiserum on the incorporation of 14 C-uridine in kinetic experi- ments of X-7 virus replication in chick embryo tissue cultures showed an elevated uptake of 14C isotope. Would Dr Schulman venture to estimate the per- centage neutralizing capacity attributed to enzyme antibodies and to haemagglutinin antibodies present in virus antiserum? DR J. L. SCHULMAN. The best answer I can give to Dr Seto is that one can protect mice almost as well by immunizing them against the enzyme antigen as against the haemagglutinin antigen of the challenge virus. Conversely, in the presence of high levels of homotypic antibody to the haemagglutinin antigen antineuraminidase antibody does not seem to provide any appreciable additional protection. It seems, therefore, that antineuraminidase anti- body is most important when new viruses appear which have antigenically new haemagglutinins. - 6592420 GENERAL DISCUSSION-SESSION V DR N. M. TAURASO. Soon after it became evident that the Hong Kong variant was sufficiently different antigenically from previously occurring influenza virus strains, consideration was given to selecting a strain suitable for vaccine production. Seven strains underwent 9 serial passages and the strains A2/Aichi/ 2/68 and A2/Philippines/303/68 produced the most haemagglutinin: they were both egg-derived. The Aichi/2 strain was selected as the production strain for the 1968-69 monovalent vaccine. When the first Hong Kong variant recombinant (X-30)-HK/l6-PR/8 (LVR 1086)-became avail- able, we compared its HA production with the Aichi/ 2 vaccine strain. The mean HA titres of these 2 strains over the period of 10 serial egg passages were approximately equal. The next recombinant we obtained was the Aichi/2-PR/8 (LVR 1119) strain X-31. The HA and CCA production of this strain was compared with the 2 previous strains passaged simultaneously. Both the mean HA and CCA values of the Aichi/2 parent and the Aichi/2-PR/8 recombinant were approximately 1.5 times higher than those of the HK/16-PR/8 recombinant. There was little difference between the Aichi/2 and the recombinant strains. The discrepancy between these results and those reported to us byDr Kilbourne might have been due to the possible bias the higher passage level had in favour of the Aichi/2 strain. Another difference was that Dr Kilbourne used human type 0 cells and we normally use cockerel cells for influenza virus haemagglutination tests. To resolve this discrepancy, experiments were performed starting with viruses which had had the same number of egg passages from isolation, with the exception of the PR/8 strain, to compare the HA production of the strains tested using cockerel, human type 0, and guinea-pig red blood cells. The mean HA values of the recombinants were only 1.5 times greater than that of the Aichi/2 strain. Using cockerel red blood cells, the mean HA titre for PR/8 was almost 6 times greater than for the Aichi/2 and approximately 4 times greater than for the recom- binant strains. The low HA titres encountered with PR/8 using human type 0 cells were surprising and inexplicable. A full account of these studies will be published later. Usually the time between the appearance of a new influenza virus variant and the time by which a suitable vaccine production strain must be selected is very short. The situation during the early weeks of the 1968 Hong Kong epidemic was no exception. We were fortunate in having been able to select a vaccine production strain much sooner than had been the case with the 1957 Asian influenza epidemic. The A2/Aichi/2/68 strain which was used for manu- facturing the 1968-69 monovalent vaccine in the USA proved to be a satisfactory production strain. Comparison of the HA and CCA production of 3 Hong Kong recombinant strains and the Aichi/2 and PR/8 parent strains showed that the recombinant strains could produce approximately 1.5 times more haemagglutinin than the Aichi/2 strain under certain conditions. However, contrary to Dr Kilbourne's observations (personal communication), the HA and CCA of the recombinant strains did not closely resemble those of the PR/8 strain. The possibility of selecting a recombinant yielding HA and CCA titres equivalent to those achievable with PR/8 could conceivably be accomplished by recombining a recombinant with PR/8 (V. Vonka, personal com- munication). A comment on the practicability of the recombina- tion procedure for selecting strains with improved growth capacity seems appropriate. If the same increased growth potential can be achieved through rapid passage in eggs, this would obviate the need for resorting to the laborious recombination pro- cedure to achieve the same end-result. Our experi- ments have not supported the suggestion that recom- binant strains would solve the vaccine production problems faced in the past. DR C. G. LOOSLI. Studies were carried out several years ago in association with Dr Paul Gerber and Dr Dorothy Hamre at the University of Chicago concerning the development of antigenic variants of influenza A virus (PR/8 strain) by serial passage in the lungs of partially immunized mice, previously vaccinated with homologous virus. Successful serial passage of PR/8 influenza A virus in immunized animals was dependent on the initial selection of mice with uniformly low HI antibody titres and the intranasal instillation of sufficient virus to favour the survival of virus particles least related to the circulating antibodies. Groups of mice, vaccinated with either the parent (PR/8-S) or the 21st (T21) passage strains and having comparable anti- body titres, showed no growth of virus following air- borne challenge with homologous strains. On the other hand, following heterologous airborne chal- lenge no deaths occurred. However, virus grew to high titre in the lungs of mice vaccinated with the parent strain and challenged with the PR/8-T21 virus 660 GENERAL DISCUSSION-SESSION V 661 which resulted in extensive pulmonary consolidation. Significantly less virus grew in the lungs of mice vac- cinated with the variant strain (T21) and challenged with the parent PR/8 virus. In these animals only microscopic evidence of lung changes due to virus growth was observed.a Four successive generations of antigenic variants of influenza PR/8-S virus, each derived from the previous one by serial passage in the lungs of mice immunized with the homologous agents, were com- pared with the original parent with respect to their serological and immunological characteristics. It was demonstrated by means of HI, complement- fixation, and in ovo neutralization tests that the variants exhibited a progressively decreasing reactiv- ity with parent (PR/8) antiserum while retaining the ability to elicit significant antibody titres to the PR/8-S influenza virus and to their respective pre- decessors. The inheritable character of the new antigenic properties of the variant strains was demonstrated by their persistence in the absence of the selective environment following 24 serial intra- nasal passages with large inocula in normal mice and following limiting dilution passage in fertile eggs.b None of the PR/8-S derived variants reacted signifi- cantly with Al antisera.C In 2 series of variants of influenza PR/8-S virus, all retained some degree of pathogenicity for mice and fertile eggs while showing progressive loss of ability to provoke antibody following vaccination or infection in mice and ferrets. The immunogenicity of the variants was therefore less than that of the original strain. Although little or no serological relationship could be demonstrated between some of the variants and the parent PR/8-S virus, a consider- able degree of cross-immunity, or protection, could be demonstrated if observations were based on death or survival of the mice. By employing the occurrence of lung lesions and the lung virus titre at 48 hours after challenge, the degree of cross-protection in mice was found to be proportional to the amount of serological cross-reaction. In general, mice vaccinated with the parent (PR/8-S) virus were significantly less resistant to infection with the variant viruses than mice vaccinated with variants and challenged with the parent virus. While these studies emphasize the importance of the immune state of the host, the a Gerber, P., Loosli, C. G. & Hamre, D. (1955) J. exp. Med., 101, 627-638. b Gerber, P., Hamre, D. & Loosli, C. G. (1956) J. exp. Med., 103, 413-424. c Hamre, D., Loosli, C. G. & Gerber, P. (1958) J. exp. Med., 107, 829-844. precise mechanism in the selection of new antigenic components of influenza virus A (PR/8 strain) in the emerging variant was not clearly defined.d DR F. M. DAVENPORT. Dr Stuart-Harris is correct in saying in his paper (see page 617) that delayed local reactions to mineral-oil adjuvant vaccines do occur. In our experience the rate is lower than the 3 or 4 per 10 000 reported in the British data. It is also correct to say that the mechanism of these reactions is unknown but their occurrence does seem to correlate with subcutaneous deposition of vaccine. Caution is always commendable but hypothetical and actual risks should be balanced against the risk of doing nothing different. By 1963 Asian influenza had caused over 150 000 excess deaths in the USA and by now the count must exceed 200 000. On this basis my own clinical and epidemiological judgement urges adoption of mineral-oil influenza vaccine now, rather than waiting until all the less important questions are answered. The bridging strains reported by Dr Fazekas de St. Groth were prevalent in the USA in 1967, yet did not prevent the 1968 outbreak of Hong Kong influenza there. Table 6 of the paper by Dr Fazekas de St. Groth (see page 657) shows that since titre values are identical after vaccination with junior or senior strains, the order of vaccination makes no practical difference. DR S. FAZEKAS DE ST. GROTH. To answer Dr Daven- port's first comment, protection would be apparent only if the same people from whom the AA/67 strain was isolated also caught influenza in 1968-69. Patently, there is no evidence for this, or Dr Daven- port would have mentioned it. In Australia the situation is completely different. These bridging strains have been prevalent there since 1962, and as Dr Warburton reported, 30%-50% of the popula- tion had anti-Hong Kong antibodies before the Hong Kong strain was even isolated. Dr Davenport's second point is well taken; but even here I have to enter a caveat. The data in Table 6 of my paper represent animals treated with very large doses of the 2 antigens, as we wanted to detect even the smallest cross-reactions. If smaller doses are used, as we learnt from bitter experience, the cross-reactions are hardly detectable and a senior primary immunity is definitely superior. These observations will be published elsewhere, in due course. d Loosli, C. G., Hamre, D. & Gerber, P.(1958) J. exp. Med., 107, 857-868.