- Email: [email protected]
Effect of priming on subsequent response to inactivated influenza vaccine
Vaccine 21 (2003) 940–945
Effect of priming on subsequent response to inactivated influenza vaccine C.W. Potter∗ , R. Jennings Section of Infection and Immunity, Division of Genomic Medicine, University of Sheffield Medical School, Sheffield S10 2RX, England, UK
Abstract Although shown to be a potent stimulator of serum antibody responses in animal models, the adjuvant immuno-stimulating complexes (ISCOMs) showed little adjuvant effect for inactivated influenza vaccines in a volunteer study. The result may be the non-comparability of the studies: animal studies were carried out chiefly in unprimed mice, while volunteers are mostly primed by previous infection and/or immunization. To test this, Balb/C mice were infected with influenza viruses or immunized with inactivated influenza vaccine, and subsequently given inactivated vaccine in saline or incorporated into ISCOMs. The serum in antibody responses was measured 1 month after immunization. The results confirm the adjuvant activity of ISCOM in unprimed mice, and show a marked reduction in adjuvant activity for primed mice. We argue that ISCOMs are important to prime the T cell response necessary for the serum antibody response to saline vaccine, but largely unnecessary where priming has been accomplished by prior exposure to influenza antigens. Further, the value of ISCOMs may lie in promoting antibody responses in unprimed subjects, and not in enhancing antibody titres. © 2002 Elsevier Science Ltd. All rights reserved. Keywords: Influenza vaccines; ISCOM; Animal models for vaccine studies
1. Introduction Despite the considerable research efforts of the past decades in developing different forms of influenza vaccine given by different routes [1–3], the only licensed vaccines available in most countries are the various forms of inactivated, saline vaccines. Although, shown to be effective in numerous clinical studies [1,4], these vaccines remain disappointing, since protective levels of haemagglutination-inhibiting (HI) antibody are only induced in 60–90% of vaccines [5]. Several factors probably contribute to this relatively low percentage: these include mismatch between vaccine formulation and the epidemic virus, since viruses continue to drift during the period of vaccine production to cause later epidemics; the relatively poor antigenicity of saline vaccines for the very young and elderly [6]; the presence of serum antibody to other influenza virus strains which can limit a response to vaccine viruses [7]; and the failure of saline vaccines to induce strong local and/or cellular immunity [3,8]. One strategy to enhance immune responses is to incorporate an adjuvant into the vaccine; and a large number of potential adjuvants have been developed and studied in laboratory animals [9]. Although, many of these preparations are too toxic for ∗
Corresponding author.
parental use in man, comparative studies of some of the remaining have indicated that immune stimulating complexes (ISCOM) are among the most potent, inducing high levels of serum antibody against both influenza and other viruses [10–14]. In addition to enhancing circulating antibody titres, ISCOM vaccines promote cellular immune responses to influenza virus antigens, local antibody production and protection against challenge virus infections [10,11,15]. The above results have encouraged volunteer trials; but in the one published report, the results were disappointing. Thus, serum haemagglutination-inhibiting (HI) antibody titres were only marginally enhanced, but this was offset by an increase in reactogenicity [16]. The reasons for the discrepancy between the results obtained in laboratory animals and in volunteers is not known, and may be due to different efficacies of adjuvants in different species; however, the differences may lie in the study design. Thus, animal studies have been conducted mostly in unprimed animals [11,12], whilst human studies include many volunteers who are primed to influenza virus antigens by prior infection or immunization. In this respect, numerous studies have demonstrated that influenza vaccines in saline are poor immunogens in experimental animals, and a good serum antibody response requires initial priming [17,18]: priming has been shown to induce Th cells essential for antibody production [15]. Thus, the animal and volunteer models are
0264-410X/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 4 - 4 1 0 X ( 0 2 ) 0 0 5 4 4 - 3
C.W. Potter, R. Jennings / Vaccine 21 (2003) 940–945
not comparable and this may explain the different results of animal and human studies. To test this, the serum HI antibody response to inactivated, trivalent influenza vaccine in saline or ISCOM was measured in both primed and unprimed Balb/c mice.
2. Materials and methods 2.1. Animals Balb/c mice were obtained from a closed, randomly-bred colony held at the University of Sheffield Field Laboratories. Mice were used at age 8–9 weeks when the weight was approximately 25 g. Viruses and virus vaccines: influenza viruses A/Texas/91 (H1N1), A/Nanchang/95 (H3N2) and B/Harbin/94 were kindly supplied by Influenza Reference Laboratory, World Health Organization, London. Virus pools were prepared and stored as described previously [10]. Virus infectivity titres for mice (MID50 ) were calculated by inoculating groups of five Balb/c mice intranasally and dropwise with logarithmic dilutions of virus in 0.1 ml volume of phosphate buffered saline (PBS). Mice were bled 21 days following infection and the sera tested for HI antibody: infectivity titres were taken as the highest dilution of virus which induced specific HI antibody in 50% of animals. Commercially obtained trivalent, subunit influenza vaccine, containing influenza A/Texas/91, A/Nanchung/95 and B/Harbin/94 antigens, were used either as a saline vaccine or incorporated into ISCOMs using previously published methods [19]: both saline and ISCOM vaccines contained 15 g haemagglutinin (HA) of each virus component in an 0.5 ml volume, as determined by single radial haemolysis tests. Both vaccines were administered to mice intramuscularly as an 0.1 ml volume containing 3 g HA of each virus component. 2.2. Haemagglutination-inhibition tests Haemagglutination-inhibition (HI) tests were carried out using a standard procedure [20]. Blood specimens, obtained from mice by retro-orbital bleeding, were left overnight at 4 ◦ C and the sera treated with cholera filtrate to remove inhibitors. Titres of HI antibody were taken as the highest serum dilution which gave 50% inhibition of haemagglutination against eight HA units of virus. 2.3. Experimental methods Four separate experiments were carried out to test the effect of priming on the subsequent reponse to saline or ISCOM vaccines. 1. Groups of Balb/c mice were primed by intranasal and dropwise inoculation with 0.1 ml of PBS (control mice)
941
or 103 MID50 of live A/Texas/91 (H1N1) virus. One month after priming, half the animals in each group, each divided into three groups of 6–8 mice, were inoculated intramuscularly with 0.1 ml of PBS, or trivalent, inactivated subunit vaccine containing 3 g HA of each of the three virus sub-type components in PBS or incorporated into ISCOM. The remaining mice were divided similarly into three groups and inoculated 3 months after priming. Blood samples were taken just before immunization and again 1 month after immunization, and tested for HI antibody against A/Texas/91 and B/Harbin/94 viruses. 2. Groups of Balb/c mice were primed with PBS or B/ Harbin/94 virus, as described above. One month later, half the animals in each group were divided into three groups, and immunized with PBS, saline or ISCOM vaccine; the remaining mice in three groups were given the same vaccines 3 months after priming. Blood samples taken before and 1 month after immunization were tested for HI antibody against B/Harbin/94 and A/Texas/91 viruses. 3. Using the same experimental design as described by experiments 1 and 2, Balb/c mice were primed with PBS (control mice) or 0.1 ml of trivalent, saline vaccine given intramuscularly. Again, three groups of primed mice were immunized with PBS, saline or ISCOM vaccine at 1 or 3 months after priming. Sera taken 1 month after immunization were tested for HI antibody against A/Texas/91 and B/Harbin/94 viruses. 4. To test if priming could be achieved by infection with influenza viruses of different sub-type, groups of Balb/c mice were primed by intranasal and dropwise infection with PBS, influenza virus A/Texas/92 (H1N1), A/Nanchung/94 (H3N2) or B/Harbin/94 at a concentration of 103 MID50 in an 0.1 ml volume of PBS. One month after infection, the animals in each group were divided into two groups and immunized with saline or ISCOM vaccine. Sera collected before and 1 month after immunization were tested for HI antibody against A/Nanchung/95 and A/Texas/91 viruses. 3. Results 3.1. Effect of priming with influenza virus A/Texas/91 (H1N1)–experiment 1 The HI antibody responses of mice primed with A/Texas/91 virus and subsequently given saline or ISCOM vaccines are shown in Fig. 1. For control mice not primed by prior infection, no detectable HI antibody was observed in serum samples following inoculation with inactivated, trivalent, saline vaccine; in contrast, mean titres for unprimed animals given ISCOM vaccine were over 1:1000. In contrast, animals primed with A/Texas/91 and given saline vaccine 1 month later exhibited a good HI antibody response, and an arithmetically better but not significantly
942
C.W. Potter, R. Jennings / Vaccine 21 (2003) 940–945
Fig. 1. H1 antibody response of mice to inactivated influenza vaccine following priming.
greater response to ISCOM vaccines; similar results were seen when animals were immunized 3 months after priming (Fig. 1). Priming had no effect on the antibody response to the B/Harbin/94 component of the vaccine; thus, control animals and animals primed with A/Texas/91 gave no response to saline vaccine but a good and comparable response to ISCOM vaccine at 1 and 3 months post-priming.
3.2. Effect of priming with B/Harbin/94–experiment 2 The results for mice primed with B/Harbin/94 prior to immunization at 1 and 3 months later with saline or ISCOM vaccine are given in Fig. 2. Again, for unprimed mice a serum HI antibody response to the B/Harbin/94 and A/Texas/ 91 components of the vaccine were only seen for ISCOM
Fig. 2. H1 antibody response of mice to inactivated influenza vaccine following priming.
C.W. Potter, R. Jennings / Vaccine 21 (2003) 940–945
943
Fig. 3. H1 antibody response of mice to inactivated influenza vaccine following priming.
vaccine; HI antibody responses to the saline vaccine were not detectable. For mice primed with B/Harbin/94, serum HI antibody titres to B/Harbin/94 were increased following inoculation with both saline and ISCOM vaccine at both 1 and 3 months after priming, and the results were similar for these two vaccine preparations. Priming with B/Harbin did not affect the response of mice to subsequent immunization with the A/Texas/91 component of the vaccine (Fig. 2).
3.3. Priming with trivalent, inactivated vaccine–experiment 3 The serum HI antibody responses for mice primed with 0.1 ml of trivalent saline vaccine given intramuscularly and subsequently inoculated with saline or ISCOM vaccine at 1 and 3 months post-priming are given in Fig. 3. For unprimed mice, no detectable HI antibody response was seen to saline
Fig. 4. H1 antibody response of mice to inactivated influenza vaccine following priming.
944
C.W. Potter, R. Jennings / Vaccine 21 (2003) 940–945
vaccine but a pronounced response was seen to both the A/Texas/91 and to B/Harbin/94 when the animals were inoculated with ISCOM vaccine with mean titres greater than 100. In contrast, for mice primed with inactivated vaccine a good HI antibody response was seen to both saline and ISCOM vaccines against both the A/Texas/91 and B/Harbin/94 components of inactivated vaccine at both 1 and 3 months after priming. 3.4. Priming with different influenza virus A sub-types–experiment 4 The results given in Fig. 4 show the effects of priming with an influenza virus of a different sub-type. Control mice responded to both the A/Texas/91 and the A/Nanchang/95 components of vaccine only when it was administered in ISCOMs; no serum HI antibody to either virus type was seen in control animals given saline vaccine (Fig. 4). In contrast, animals primed with A/Texas/91produced serum HI antibody to both the A/Texas/91 and A/Nanchang/95 components of vaccines following immunization with both saline and ISCOM vaccines. The converse was also found; thus, for mice primed with A/Nanchung/95 virus and subsequently immunized with saline or ISCOM vaccines, serum HI antibodies to both the A/Nangchung/95 and A/Texas/91 viral components of both saline and ISCOM vaccines were detected at mean titres of >1000 1 month later (Fig. 4). For mice primed with B/Harbin/94, no antibody response was seen to either influenza A types following immunization with saline vaccine, but a good HI antibody response was seen in these mice given ISCOM vaccine: the results confirm the findings shown in Figs. 1 and 2.
4. Discussion Previous studies have shown that the serum HI antibody titre is the most important indicator of immunity to influenza [21], and that this antibody response is often disappointing following immunization with inactivated saline vaccines [1,2]. The HI antibody response of unprimed animals to these vaccines, including mice and ferrets, is low, and in many cases undetectable: in animals, an HI antibody response to saline inactivated influenza vaccine is dependent on previous exposure to influenza virus antigens [17,18]. This priming effect can be achieved by prior exposure to live virus or to saline vaccines [18]; but priming is not necessary if the vaccine is incorporated into adjuvants which stimulate Th cell responses [15,18]. The value of ISCOM as an adjuvant to promote the HI antibody response to influenza vaccines in saline has been demonstrated previously; and the present results confirm this marked adjuvant effect [10,12]. In comparative studies ISCOM have been shown to be among the best promoters of antibody responses to virus antigens [12,13]. However, when similar experiments were carried out in volunteers, comparing the immune re-
sponses to saline and ISCOM influenza vaccines, the results were disappointing; the moderately higher antibody titres achieved with ISCOM vaccines were offset by an increased incidence of vaccine reactions [16]. The present studies offer an explanation for the discrepancy in the results. Thus, the differential effect of high HI antibody responses to ISCOM vaccines compared to low or absent responses to saline vaccine in normal mice is largely discounted if animals have been primed by previous exposure to influenza virus; and this previous exposure can be induced by prior experience to homologous virus infection, by exposure to influenza viruses of different sub-type but not by infection or immunization with influenza viruses of different type. The results indicate that a measurable serum HI antibody response to saline vaccine required prior priming, that priming was not required for a response to ISCOM vaccine and that in primed animals, the response to saline and ISCOM vaccine was not significantly different. In addition, priming for a response to saline vaccine is not dependent upon exposure to live virus infection, as shown in Figs. 1 and 2, but can be achieved by exposure to inactivated virus antigens. Most volunteers have been primed by previous infection or immunization, and possess Th cells. In this case priming of Th cells is not required and the effect of adjuvant is not pronounced. Thus, our present results in mice are in agreement with human studies, since when animals are primed, the comparative reaction between saline and ISCOM vaccines is approximately similar. However, arithmetically higher titres are commonly found in response to ISCOM vaccines compared to saline vaccines, and this again follows the experience of volunteer studies [16]. The present results question some opinions concerning the value of adjuvants in inactivated influenza vaccines. Thus, when prior infection or immunization has occurred vaccines are required to raise HI antibody titres to protective levels of 40 for influenza A infections [21]: this can be achieved with saline vaccines which do not carry the possible risk of increased reactogenicity. This may be true for other adjuvant systems and other virus vaccines but this has not been tested to date. In addition, the value of adjuvants such as ISCOM would be to stimulate Th cells and high levels of antibody in subjects where no previous priming infection has taken place, and this would be achieved with a single dose of adjuvant vaccine compared to two doses of saline vaccine, one for priming and one to induce the immune responses. There are many examples of vaccines where priming has not taken place, and here the value of adjuvants is not questioned. It is important to recognize that when designing animal experiments intended to mimic human experience, matching conditions as far as possible is desirable. Vaccines with and without adjuvants for human use should be tested in animals under comparable conditions where the human experience is to boost existing antibody titres an animal experiment should include priming whilst unprimed animals match the human experience for no previous experience. Without this pre-requisite, animal studies are of questionable value, and
C.W. Potter, R. Jennings / Vaccine 21 (2003) 940–945
volunteer studies can be disappointing. Finally, it should be emphasized that the present results apply to influenza virus vaccines which present some highly individual features. The results cannot be extended to other virus vaccines without specific tests; and the present results only apply to ISCOM since it has not been extended to other adjuvants/carrier systems. The implications for other vaccines in adjuvants remains speculative; and the underpinning role of Th cells for the response to saline and ISCOM vaccine is being tested by serum/lymphocyte transfer studies at the present time. References [1] Potter CW. Inactivated influenza virus vaccines. In: Beare AS, editor. Basic and applied influenza research. Boca Raton (FL, USA): CRC Press; 1982. p. 119–38. [2] Wright PF, Karzon DT. Live attenuated influenza vaccines. Prog Med Virol 1987;34:70–85. [3] Potter CW, Jennings R. Intranasal immunisation with inactivated influenza vaccine. Pharm Sci Tech Today 1999;2:402–8. [4] Keitel WA, Cate TR, Couch RB, Huggins LL, Hess KR. Efficacy of repeated annual immunisation with inactivated influenza vaccine over a 5-year-period. Vaccine 1997;15:1114–22. [5] Couch RB, Keitel WA, Cate TR. Improvement of inactivated influenza virus vaccines. J Infect Dis 1997;176(Suppl 1):S38–44. [6] Miller RA. The ageing immune response: primer and prospectus. Science 1996;273:70–4. [7] Smith DJ, Forrest S, Ackley DH, Perelson AS. Variable efficacy of repeated annual influenza immunisation. Proc Natl Acad Sci 1999;96:14001–9. [8] Askanas, BA, McMichael, AJ, Webster, RG. The response to influenza viruses and the problem of protection against infection. In: Beare AS, editor. Basic and applied influenza research. Boca Raton (FL, USA): CRC Press; 1982. p. 160–88. [9] Edelman R, Tacket CO. Adjuvants Int Rev Immunol 1990;7:51–66. [10] Ghazi HO, Potter CW, Smith TL, Jennings R. Comparative antibody responses and protection in mice to immunisation by oral or parental routes with influenza virus subunit antigen in aqueous form or incorporated into ISCOMs. J Med Microbiol 1995;42:51–60.
945
[11] Coulter A, Wong TY, Drane D, Bates J, Macfarlan R, Cox J. Studies on experimental adjuvanted influenza vaccines: comparison of immune stimulating complexes (ISCOMTM ) and oil-in-water vaccines. Vaccine 1997;16:1243–53. [12] Ben-Ahmeida ETS, Gregoriadis G, Potter CW, Jennings R. Immunopotentiation of local and systemic humoral immune responses by ISCOMs, liposomes and FCA: role in protection against influenza A in mice. Vaccine 1993;11:1302–9. [13] Stieneker F, Kersten G, Van Bloois L, Crammelin DJA, Hem SL, Lower J, Kreuter J. Comparison of 24 different adjuvants in inactivated HIV-2 split whole virus as antigen in mice. Induction of titres of binding antibodies and toxicity of the formulations. Vaccine 1995;13:45–53. [14] Sjolander S, Hansen J-ES, Lovgren-Benetsson K, Akerblom L, Morein B. Induction of homologous virus neutralising antibodies in guinea-pigs immunised with two human immunodeficiency virus type 1 glycoprotein gp120-ISCOM preparations. A comparison with other adjuvant systems. Vaccine 1996;14:344–52. [15] Deliyannis G, Jackson DC, Dyer W, Bates J, Coulter A, Harling-McNabb LE, Brown LE. Immunopotentiation of humoral and cellular responses to inactivated influenza vaccines in two different adjuvants with potential for human use. Vaccine 1998;16:2058–68. [16] Martin TJ. Enhanced immunogenicity of chiron bioline adjuvanted influenza vaccine in the elderly. In: Brown LE, Hampson AW, Webster RG, editors. Proceedings of the 3rd International Conference on Options for the Control of Influenza 3. Amsterdam: Elsevier; 1996. p. 647–52. [17] Potter CW, McLaren C, Shore SL. Immunity to influenza A in ferrets. V. Immunisation with inactivated virus in adjuvant 65. J Hyg Camb 1974;71:97–106. [18] Jennings R, Potter CW. Effect of pre-infection and pre-immunization on the serum antibody response to subsequent immunization with heterotypic influenza vaccines. J Immunol 1974;113:1834–43. [19] Cox, J, Coulter, A, Macfarlan, R, Beezum, Z, Bates, J, Wong, TW, Drane, D. Development of an influenza-ISCOM TM vaccine. In: Gregoriadis G, editor. Vaccine design: the role of cytokine networks. New York (USA): Plenum Press; 1997. p. 33–47. [20] Jennings R, Potter CW, Massey PMO, Duerden BI, Martin J, Bevan AM. Response of volunteers to inactivated influenza virus vaccines. J Hyg Camb 1981;86:1–16. [21] Potter CW, Oxford JS. Determinants of immunity to influenza infection in man. Br Med J 1979;35:69–75.
Effect of priming on subsequent response to inactivated influenza vaccine C.W. Potter∗ , R. Jennings Section of Infection and Immunity, Division of Genomic Medicine, University of Sheffield Medical School, Sheffield S10 2RX, England, UK
Abstract Although shown to be a potent stimulator of serum antibody responses in animal models, the adjuvant immuno-stimulating complexes (ISCOMs) showed little adjuvant effect for inactivated influenza vaccines in a volunteer study. The result may be the non-comparability of the studies: animal studies were carried out chiefly in unprimed mice, while volunteers are mostly primed by previous infection and/or immunization. To test this, Balb/C mice were infected with influenza viruses or immunized with inactivated influenza vaccine, and subsequently given inactivated vaccine in saline or incorporated into ISCOMs. The serum in antibody responses was measured 1 month after immunization. The results confirm the adjuvant activity of ISCOM in unprimed mice, and show a marked reduction in adjuvant activity for primed mice. We argue that ISCOMs are important to prime the T cell response necessary for the serum antibody response to saline vaccine, but largely unnecessary where priming has been accomplished by prior exposure to influenza antigens. Further, the value of ISCOMs may lie in promoting antibody responses in unprimed subjects, and not in enhancing antibody titres. © 2002 Elsevier Science Ltd. All rights reserved. Keywords: Influenza vaccines; ISCOM; Animal models for vaccine studies
1. Introduction Despite the considerable research efforts of the past decades in developing different forms of influenza vaccine given by different routes [1–3], the only licensed vaccines available in most countries are the various forms of inactivated, saline vaccines. Although, shown to be effective in numerous clinical studies [1,4], these vaccines remain disappointing, since protective levels of haemagglutination-inhibiting (HI) antibody are only induced in 60–90% of vaccines [5]. Several factors probably contribute to this relatively low percentage: these include mismatch between vaccine formulation and the epidemic virus, since viruses continue to drift during the period of vaccine production to cause later epidemics; the relatively poor antigenicity of saline vaccines for the very young and elderly [6]; the presence of serum antibody to other influenza virus strains which can limit a response to vaccine viruses [7]; and the failure of saline vaccines to induce strong local and/or cellular immunity [3,8]. One strategy to enhance immune responses is to incorporate an adjuvant into the vaccine; and a large number of potential adjuvants have been developed and studied in laboratory animals [9]. Although, many of these preparations are too toxic for ∗
Corresponding author.
parental use in man, comparative studies of some of the remaining have indicated that immune stimulating complexes (ISCOM) are among the most potent, inducing high levels of serum antibody against both influenza and other viruses [10–14]. In addition to enhancing circulating antibody titres, ISCOM vaccines promote cellular immune responses to influenza virus antigens, local antibody production and protection against challenge virus infections [10,11,15]. The above results have encouraged volunteer trials; but in the one published report, the results were disappointing. Thus, serum haemagglutination-inhibiting (HI) antibody titres were only marginally enhanced, but this was offset by an increase in reactogenicity [16]. The reasons for the discrepancy between the results obtained in laboratory animals and in volunteers is not known, and may be due to different efficacies of adjuvants in different species; however, the differences may lie in the study design. Thus, animal studies have been conducted mostly in unprimed animals [11,12], whilst human studies include many volunteers who are primed to influenza virus antigens by prior infection or immunization. In this respect, numerous studies have demonstrated that influenza vaccines in saline are poor immunogens in experimental animals, and a good serum antibody response requires initial priming [17,18]: priming has been shown to induce Th cells essential for antibody production [15]. Thus, the animal and volunteer models are
0264-410X/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 4 - 4 1 0 X ( 0 2 ) 0 0 5 4 4 - 3
C.W. Potter, R. Jennings / Vaccine 21 (2003) 940–945
not comparable and this may explain the different results of animal and human studies. To test this, the serum HI antibody response to inactivated, trivalent influenza vaccine in saline or ISCOM was measured in both primed and unprimed Balb/c mice.
2. Materials and methods 2.1. Animals Balb/c mice were obtained from a closed, randomly-bred colony held at the University of Sheffield Field Laboratories. Mice were used at age 8–9 weeks when the weight was approximately 25 g. Viruses and virus vaccines: influenza viruses A/Texas/91 (H1N1), A/Nanchang/95 (H3N2) and B/Harbin/94 were kindly supplied by Influenza Reference Laboratory, World Health Organization, London. Virus pools were prepared and stored as described previously [10]. Virus infectivity titres for mice (MID50 ) were calculated by inoculating groups of five Balb/c mice intranasally and dropwise with logarithmic dilutions of virus in 0.1 ml volume of phosphate buffered saline (PBS). Mice were bled 21 days following infection and the sera tested for HI antibody: infectivity titres were taken as the highest dilution of virus which induced specific HI antibody in 50% of animals. Commercially obtained trivalent, subunit influenza vaccine, containing influenza A/Texas/91, A/Nanchung/95 and B/Harbin/94 antigens, were used either as a saline vaccine or incorporated into ISCOMs using previously published methods [19]: both saline and ISCOM vaccines contained 15 g haemagglutinin (HA) of each virus component in an 0.5 ml volume, as determined by single radial haemolysis tests. Both vaccines were administered to mice intramuscularly as an 0.1 ml volume containing 3 g HA of each virus component. 2.2. Haemagglutination-inhibition tests Haemagglutination-inhibition (HI) tests were carried out using a standard procedure [20]. Blood specimens, obtained from mice by retro-orbital bleeding, were left overnight at 4 ◦ C and the sera treated with cholera filtrate to remove inhibitors. Titres of HI antibody were taken as the highest serum dilution which gave 50% inhibition of haemagglutination against eight HA units of virus. 2.3. Experimental methods Four separate experiments were carried out to test the effect of priming on the subsequent reponse to saline or ISCOM vaccines. 1. Groups of Balb/c mice were primed by intranasal and dropwise inoculation with 0.1 ml of PBS (control mice)
941
or 103 MID50 of live A/Texas/91 (H1N1) virus. One month after priming, half the animals in each group, each divided into three groups of 6–8 mice, were inoculated intramuscularly with 0.1 ml of PBS, or trivalent, inactivated subunit vaccine containing 3 g HA of each of the three virus sub-type components in PBS or incorporated into ISCOM. The remaining mice were divided similarly into three groups and inoculated 3 months after priming. Blood samples were taken just before immunization and again 1 month after immunization, and tested for HI antibody against A/Texas/91 and B/Harbin/94 viruses. 2. Groups of Balb/c mice were primed with PBS or B/ Harbin/94 virus, as described above. One month later, half the animals in each group were divided into three groups, and immunized with PBS, saline or ISCOM vaccine; the remaining mice in three groups were given the same vaccines 3 months after priming. Blood samples taken before and 1 month after immunization were tested for HI antibody against B/Harbin/94 and A/Texas/91 viruses. 3. Using the same experimental design as described by experiments 1 and 2, Balb/c mice were primed with PBS (control mice) or 0.1 ml of trivalent, saline vaccine given intramuscularly. Again, three groups of primed mice were immunized with PBS, saline or ISCOM vaccine at 1 or 3 months after priming. Sera taken 1 month after immunization were tested for HI antibody against A/Texas/91 and B/Harbin/94 viruses. 4. To test if priming could be achieved by infection with influenza viruses of different sub-type, groups of Balb/c mice were primed by intranasal and dropwise infection with PBS, influenza virus A/Texas/92 (H1N1), A/Nanchung/94 (H3N2) or B/Harbin/94 at a concentration of 103 MID50 in an 0.1 ml volume of PBS. One month after infection, the animals in each group were divided into two groups and immunized with saline or ISCOM vaccine. Sera collected before and 1 month after immunization were tested for HI antibody against A/Nanchung/95 and A/Texas/91 viruses. 3. Results 3.1. Effect of priming with influenza virus A/Texas/91 (H1N1)–experiment 1 The HI antibody responses of mice primed with A/Texas/91 virus and subsequently given saline or ISCOM vaccines are shown in Fig. 1. For control mice not primed by prior infection, no detectable HI antibody was observed in serum samples following inoculation with inactivated, trivalent, saline vaccine; in contrast, mean titres for unprimed animals given ISCOM vaccine were over 1:1000. In contrast, animals primed with A/Texas/91 and given saline vaccine 1 month later exhibited a good HI antibody response, and an arithmetically better but not significantly
942
C.W. Potter, R. Jennings / Vaccine 21 (2003) 940–945
Fig. 1. H1 antibody response of mice to inactivated influenza vaccine following priming.
greater response to ISCOM vaccines; similar results were seen when animals were immunized 3 months after priming (Fig. 1). Priming had no effect on the antibody response to the B/Harbin/94 component of the vaccine; thus, control animals and animals primed with A/Texas/91 gave no response to saline vaccine but a good and comparable response to ISCOM vaccine at 1 and 3 months post-priming.
3.2. Effect of priming with B/Harbin/94–experiment 2 The results for mice primed with B/Harbin/94 prior to immunization at 1 and 3 months later with saline or ISCOM vaccine are given in Fig. 2. Again, for unprimed mice a serum HI antibody response to the B/Harbin/94 and A/Texas/ 91 components of the vaccine were only seen for ISCOM
Fig. 2. H1 antibody response of mice to inactivated influenza vaccine following priming.
C.W. Potter, R. Jennings / Vaccine 21 (2003) 940–945
943
Fig. 3. H1 antibody response of mice to inactivated influenza vaccine following priming.
vaccine; HI antibody responses to the saline vaccine were not detectable. For mice primed with B/Harbin/94, serum HI antibody titres to B/Harbin/94 were increased following inoculation with both saline and ISCOM vaccine at both 1 and 3 months after priming, and the results were similar for these two vaccine preparations. Priming with B/Harbin did not affect the response of mice to subsequent immunization with the A/Texas/91 component of the vaccine (Fig. 2).
3.3. Priming with trivalent, inactivated vaccine–experiment 3 The serum HI antibody responses for mice primed with 0.1 ml of trivalent saline vaccine given intramuscularly and subsequently inoculated with saline or ISCOM vaccine at 1 and 3 months post-priming are given in Fig. 3. For unprimed mice, no detectable HI antibody response was seen to saline
Fig. 4. H1 antibody response of mice to inactivated influenza vaccine following priming.
944
C.W. Potter, R. Jennings / Vaccine 21 (2003) 940–945
vaccine but a pronounced response was seen to both the A/Texas/91 and to B/Harbin/94 when the animals were inoculated with ISCOM vaccine with mean titres greater than 100. In contrast, for mice primed with inactivated vaccine a good HI antibody response was seen to both saline and ISCOM vaccines against both the A/Texas/91 and B/Harbin/94 components of inactivated vaccine at both 1 and 3 months after priming. 3.4. Priming with different influenza virus A sub-types–experiment 4 The results given in Fig. 4 show the effects of priming with an influenza virus of a different sub-type. Control mice responded to both the A/Texas/91 and the A/Nanchang/95 components of vaccine only when it was administered in ISCOMs; no serum HI antibody to either virus type was seen in control animals given saline vaccine (Fig. 4). In contrast, animals primed with A/Texas/91produced serum HI antibody to both the A/Texas/91 and A/Nanchang/95 components of vaccines following immunization with both saline and ISCOM vaccines. The converse was also found; thus, for mice primed with A/Nanchung/95 virus and subsequently immunized with saline or ISCOM vaccines, serum HI antibodies to both the A/Nangchung/95 and A/Texas/91 viral components of both saline and ISCOM vaccines were detected at mean titres of >1000 1 month later (Fig. 4). For mice primed with B/Harbin/94, no antibody response was seen to either influenza A types following immunization with saline vaccine, but a good HI antibody response was seen in these mice given ISCOM vaccine: the results confirm the findings shown in Figs. 1 and 2.
4. Discussion Previous studies have shown that the serum HI antibody titre is the most important indicator of immunity to influenza [21], and that this antibody response is often disappointing following immunization with inactivated saline vaccines [1,2]. The HI antibody response of unprimed animals to these vaccines, including mice and ferrets, is low, and in many cases undetectable: in animals, an HI antibody response to saline inactivated influenza vaccine is dependent on previous exposure to influenza virus antigens [17,18]. This priming effect can be achieved by prior exposure to live virus or to saline vaccines [18]; but priming is not necessary if the vaccine is incorporated into adjuvants which stimulate Th cell responses [15,18]. The value of ISCOM as an adjuvant to promote the HI antibody response to influenza vaccines in saline has been demonstrated previously; and the present results confirm this marked adjuvant effect [10,12]. In comparative studies ISCOM have been shown to be among the best promoters of antibody responses to virus antigens [12,13]. However, when similar experiments were carried out in volunteers, comparing the immune re-
sponses to saline and ISCOM influenza vaccines, the results were disappointing; the moderately higher antibody titres achieved with ISCOM vaccines were offset by an increased incidence of vaccine reactions [16]. The present studies offer an explanation for the discrepancy in the results. Thus, the differential effect of high HI antibody responses to ISCOM vaccines compared to low or absent responses to saline vaccine in normal mice is largely discounted if animals have been primed by previous exposure to influenza virus; and this previous exposure can be induced by prior experience to homologous virus infection, by exposure to influenza viruses of different sub-type but not by infection or immunization with influenza viruses of different type. The results indicate that a measurable serum HI antibody response to saline vaccine required prior priming, that priming was not required for a response to ISCOM vaccine and that in primed animals, the response to saline and ISCOM vaccine was not significantly different. In addition, priming for a response to saline vaccine is not dependent upon exposure to live virus infection, as shown in Figs. 1 and 2, but can be achieved by exposure to inactivated virus antigens. Most volunteers have been primed by previous infection or immunization, and possess Th cells. In this case priming of Th cells is not required and the effect of adjuvant is not pronounced. Thus, our present results in mice are in agreement with human studies, since when animals are primed, the comparative reaction between saline and ISCOM vaccines is approximately similar. However, arithmetically higher titres are commonly found in response to ISCOM vaccines compared to saline vaccines, and this again follows the experience of volunteer studies [16]. The present results question some opinions concerning the value of adjuvants in inactivated influenza vaccines. Thus, when prior infection or immunization has occurred vaccines are required to raise HI antibody titres to protective levels of 40 for influenza A infections [21]: this can be achieved with saline vaccines which do not carry the possible risk of increased reactogenicity. This may be true for other adjuvant systems and other virus vaccines but this has not been tested to date. In addition, the value of adjuvants such as ISCOM would be to stimulate Th cells and high levels of antibody in subjects where no previous priming infection has taken place, and this would be achieved with a single dose of adjuvant vaccine compared to two doses of saline vaccine, one for priming and one to induce the immune responses. There are many examples of vaccines where priming has not taken place, and here the value of adjuvants is not questioned. It is important to recognize that when designing animal experiments intended to mimic human experience, matching conditions as far as possible is desirable. Vaccines with and without adjuvants for human use should be tested in animals under comparable conditions where the human experience is to boost existing antibody titres an animal experiment should include priming whilst unprimed animals match the human experience for no previous experience. Without this pre-requisite, animal studies are of questionable value, and
C.W. Potter, R. Jennings / Vaccine 21 (2003) 940–945
volunteer studies can be disappointing. Finally, it should be emphasized that the present results apply to influenza virus vaccines which present some highly individual features. The results cannot be extended to other virus vaccines without specific tests; and the present results only apply to ISCOM since it has not been extended to other adjuvants/carrier systems. The implications for other vaccines in adjuvants remains speculative; and the underpinning role of Th cells for the response to saline and ISCOM vaccine is being tested by serum/lymphocyte transfer studies at the present time. References [1] Potter CW. Inactivated influenza virus vaccines. In: Beare AS, editor. Basic and applied influenza research. Boca Raton (FL, USA): CRC Press; 1982. p. 119–38. [2] Wright PF, Karzon DT. Live attenuated influenza vaccines. Prog Med Virol 1987;34:70–85. [3] Potter CW, Jennings R. Intranasal immunisation with inactivated influenza vaccine. Pharm Sci Tech Today 1999;2:402–8. [4] Keitel WA, Cate TR, Couch RB, Huggins LL, Hess KR. Efficacy of repeated annual immunisation with inactivated influenza vaccine over a 5-year-period. Vaccine 1997;15:1114–22. [5] Couch RB, Keitel WA, Cate TR. Improvement of inactivated influenza virus vaccines. J Infect Dis 1997;176(Suppl 1):S38–44. [6] Miller RA. The ageing immune response: primer and prospectus. Science 1996;273:70–4. [7] Smith DJ, Forrest S, Ackley DH, Perelson AS. Variable efficacy of repeated annual influenza immunisation. Proc Natl Acad Sci 1999;96:14001–9. [8] Askanas, BA, McMichael, AJ, Webster, RG. The response to influenza viruses and the problem of protection against infection. In: Beare AS, editor. Basic and applied influenza research. Boca Raton (FL, USA): CRC Press; 1982. p. 160–88. [9] Edelman R, Tacket CO. Adjuvants Int Rev Immunol 1990;7:51–66. [10] Ghazi HO, Potter CW, Smith TL, Jennings R. Comparative antibody responses and protection in mice to immunisation by oral or parental routes with influenza virus subunit antigen in aqueous form or incorporated into ISCOMs. J Med Microbiol 1995;42:51–60.
945
[11] Coulter A, Wong TY, Drane D, Bates J, Macfarlan R, Cox J. Studies on experimental adjuvanted influenza vaccines: comparison of immune stimulating complexes (ISCOMTM ) and oil-in-water vaccines. Vaccine 1997;16:1243–53. [12] Ben-Ahmeida ETS, Gregoriadis G, Potter CW, Jennings R. Immunopotentiation of local and systemic humoral immune responses by ISCOMs, liposomes and FCA: role in protection against influenza A in mice. Vaccine 1993;11:1302–9. [13] Stieneker F, Kersten G, Van Bloois L, Crammelin DJA, Hem SL, Lower J, Kreuter J. Comparison of 24 different adjuvants in inactivated HIV-2 split whole virus as antigen in mice. Induction of titres of binding antibodies and toxicity of the formulations. Vaccine 1995;13:45–53. [14] Sjolander S, Hansen J-ES, Lovgren-Benetsson K, Akerblom L, Morein B. Induction of homologous virus neutralising antibodies in guinea-pigs immunised with two human immunodeficiency virus type 1 glycoprotein gp120-ISCOM preparations. A comparison with other adjuvant systems. Vaccine 1996;14:344–52. [15] Deliyannis G, Jackson DC, Dyer W, Bates J, Coulter A, Harling-McNabb LE, Brown LE. Immunopotentiation of humoral and cellular responses to inactivated influenza vaccines in two different adjuvants with potential for human use. Vaccine 1998;16:2058–68. [16] Martin TJ. Enhanced immunogenicity of chiron bioline adjuvanted influenza vaccine in the elderly. In: Brown LE, Hampson AW, Webster RG, editors. Proceedings of the 3rd International Conference on Options for the Control of Influenza 3. Amsterdam: Elsevier; 1996. p. 647–52. [17] Potter CW, McLaren C, Shore SL. Immunity to influenza A in ferrets. V. Immunisation with inactivated virus in adjuvant 65. J Hyg Camb 1974;71:97–106. [18] Jennings R, Potter CW. Effect of pre-infection and pre-immunization on the serum antibody response to subsequent immunization with heterotypic influenza vaccines. J Immunol 1974;113:1834–43. [19] Cox, J, Coulter, A, Macfarlan, R, Beezum, Z, Bates, J, Wong, TW, Drane, D. Development of an influenza-ISCOM TM vaccine. In: Gregoriadis G, editor. Vaccine design: the role of cytokine networks. New York (USA): Plenum Press; 1997. p. 33–47. [20] Jennings R, Potter CW, Massey PMO, Duerden BI, Martin J, Bevan AM. Response of volunteers to inactivated influenza virus vaccines. J Hyg Camb 1981;86:1–16. [21] Potter CW, Oxford JS. Determinants of immunity to influenza infection in man. Br Med J 1979;35:69–75.