Changing perspective on immunization against influenza

Vaccine 25 (2007) 3062–3065

Changing perspective on immunization against influenza Bert E. Johansson a,∗ , Ian C. Brett b b

a Innovation Sciences, Armonk, NY 10504, USA State University of New York, Stony Brook School of Medicine, Health Sciences Center, L4, Stony Brook, NY 11794, USA

Available online 19 January 2007

Abstract Current vaccination strategies against influenza rely on decades old technology of strain selection and prolonged labor-intensive, embryonated chicken–egg based production methods. Although, containing both major surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA), the immunity engendered by these vaccines is dominated by the anti-HA response. Consequently, current vaccines are susceptible to failure resulting from significant antigenic drift or shift in the time elapsing from the selection of the vaccine candidate strain and wild-type virus exposure. Therefore, immunity may be of short duration. There must be a change in vaccine strategy to include immunization with both HA and NA to broaden the immune response against influenza. Inclusion of the more slowly evolving NA in a vaccine against influenza will reduce the vulnerability to antigenic changes in a potential emerging influenza virus. Alternative production technologies such as recombinant baculovirus and yeast should be explored to decrease vaccine production times. © 2007 Elsevier Ltd. All rights reserved. Keywords: Influenza; Vaccine; Neuraminidase

Influenza represents an important, poorly controlled public health problem. Estimates are that each year, influenza is responsible for 40,000 deaths in the United States and up to 150,000 hospitalizations, making influenza the most common cause of vaccine preventable morbidity and mortality. The impact is especially severe at the extremes of the age spectrum, in very young children and in the elderly, as well as in individuals with a variety of chronic medical conditions. In last century, the human population was affected by three influenza pandemics: 1918–1919 (“Spanish flu”), 1957–1958 (“Asian flu”) and 1968–1969 (“Hong Kong flu”). Currently, the world faces threat of a new pandemic, as avian H5N1 viruses continue to spread along the flyways of migratory waterfowl, with increasing risk of transmission to humans. The available vaccines for H5N1 influenza are poorly immunogenic in man, and the continued antigenic variation in H5 viruses precludes the ability to stockpile an optimally effective vaccine in advance of a pandemic. H5 viruses resistant to both adamantanes and neuraminidase ∗

Corresponding author. Tel.: +1 914 273 2373; fax: +1 914 273 5492. E-mail addresses: [email protected] (B.E. Johansson), [email protected] (I.C. Brett). 0264-410X/$ – see front matter © 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2007.01.030

inhibitors have been isolated. These developments suggest that our ability to control pandemic influenza may not be very different today than it was in 1918. Influenza viruses are classified into types A, B and C, based on genetic and antigenic differences among surface proteins, the hemagglutinin (HA) and neuraminidase (NA) and the major internal proteins, M and NP [1]. Influenza A viruses have been isolated from many host species including humans, pigs, horses, dogs, mink, felids, marine mammals and a wide range of domestic and wild birds. Aquatic birds are the natural reservoir for the 15 HA and 9 NA subtypes of influenza A. HA mediates the initial attachment of the virion to cells via sialic acid residues and possesses a fusion capability that enables the virus envelope to integrate with a lysoendosomal membrane allowing the internal viral components access to the host cell cytoplasm [1,2]. Antibodies to HA neutralize viral infectivity; antigenic variation in the HA is mainly responsible for frequent outbreaks of influenza and for the poor control of infection by immunization. NA is a tetrameric enzyme that cleaves terminal sialic acid residues from any oligosaccharide chain [1,2], permitting transport of virions through mucin and destroying the HA receptor on the host cell thereby allowing release of progeny virus

B.E. Johansson, I.C. Brett / Vaccine 25 (2007) 3062–3065

from infected cells. Immunization of mice [4] or humans [3] with vaccines containing immunogenic NA induces a specific response to NA which, although infection-permissive across a broad range of NA antibody levels results in the reduction of pulmonary virus titers below a pathogenic threshold. Intact virions also contain a third major envelope associated protein called matrix (M1) protein. Another surface protein found in variable amounts in the viral membrane, the M2 protein functions as a pH inducible ion-channel [1]. Within the viral envelope are eight RNA segments, polymerase proteins (PA, PB1 and PB2) supplying the enzymatic machinery for viral RNA synthesis and nucleoprotein (NP) associated with RNA segments to form ribonucleoprotein. Additionally, influenza virus encodes two non-structural proteins, NS1 and PB1-F2. The major role of the NS1 protein is the inhibition of the type I interferon response and PB1-F2 has been implicated in the regulation of apoptosis of infected immune cells [2]. Increase in serum antibody response during influenza is usually demonstrable no earlier than the end of the first week of illness. Specific response to infection diagnostic of infection can be documented by comparative measurement of antibody in serum obtained during the acute phase and in convalescent phase serum obtained 10 days or more following the onset of illness. Immune responses to the more highly conserved internal proteins; M1, M2 and NP [1,5] have been studied with focus on the cross-reactive cytotoxic T-cell (CTL) and T-helper response. Antibodies to M1 and NP can be found in the sera of animals immunized with whole virus vaccines, purified protein preparations and after infection [4,5]. However, studies have failed to demonstrate a significant role for these antibodies in the amelioration of disease. Despite evidence that live and inactivated influenza vaccines induce cross-reactive T-cells in humans [5] and mice [4], reinfection with homologous or heterotypic virus occurs. Cell-mediated immunity can promote viral clearance it cannot prevent infection. In humans, the level of antiinfluenza CTLs correlates with the rate of viral clearance but not alter susceptibility to infection or subsequent infection [5]. Recently, attention has been directed toward M2 protein which contains a 23 amino-acid non-glycoslyated ectodomain, referred to as M2e, which has limited variation. Antibodies toward M2e can be induced by vaccination with a M2e vaccine which induced an immunity that did not prevent infection but did restrict viral replication, reduced illness and reduce deaths after an experimental infectious challenge [2].

1. Current influenza vaccines Influenza remains a pervasive public health problem in spite of the wide availability of two currently licensed vaccines against influenza: conventional inactivated virus vaccine (CIV) and live-attenuated vaccine (LAV). CIV is derived from inactivated high-yield reassortant viruses whose internal genes are from a A/PR/8/34 high-yield donor parent [1,6]. Similarly, the live-attenuated influenza vaccine

3063

is derived from a master donor virus (MDV) strain containing temperature sensitive (ts), cold-adapted (ca) and attenuation (att) mutations in several genes coding for internal proteins [1]. Each vaccines possesses the HA and NA of recently prevalent influenza viruses predicted to cause widespread infection. In primed individuals, vaccination with CIV induced “protective” levels of antibody against HA, which prevented infection in 89% of recipients shortly after vaccination. In unprimed recipients, only 65% developed protective levels of serum anti-HA antibodies [6]. In general, levels of anti-HA antibody were low after a single dose of vaccine but increased significantly in response to a second dose, a phenomenon not observed with inactivated H5N1 vaccine [3]. CIV and LAV are effective when the HA of the vaccine strain is closely matched antigenically to the wild-type strain HA. Immunity produced in this way is of short duration. HA and NA are subject to continuous and sequential evolution within immune or partially immune populations. Antigenic variants within a subtype emerge and are gradually selected as the predominant virus while the preceding virus is suppressed by specific antibody arising in the population (antigenic drift). Kilbourne et al. [7] studied the rate of evolution of epidemiologically important H1N1 and H3N2 antigens isolated from humans over a 30-year period and determined that HA evolved more rapidly than NA. Vaccine failure can occur if the predicted vaccine strain is not the prevalent wild-type strain, as occurred in the 2003–2004 influenza season with A/Panama/2007/99 (vaccine strain) H3N2 and A/Fujian/411/2002 H3N2 (circulating strain): 12.7% of clinical influenza isolates were antigenically similar to the vaccine strain A/Panama and 87.3% were similar to the drift variant A/Fujian [4]. Although current influenza virus seeds for vaccine production must be shown to have the appropriate NA antigen. The vaccine’s NA content is variable, as a result the frequency of antibody response to NA is poor (mean seroconversion rate of 18%) compared to the HA response (84%) [1]. Yet, mice immunized with graded doses of purified HA and NA from A/Hong Kong/1/68 (H3N2) influenza virus demonstrated equivalent responses when HA-specific and NA-specific antibodies were measured by ELISA, indicating that the HA and NA are antigenically equivalent [3]. Vaccination with CIV and infection will induce antibody formation to both HA and NA in man [2] and mice [3] but the immunologic response to NA is severely suppressed in primed subjects as a result of a lack of immunogenic NA in the vaccine and HA dominant antigenic competition [3,4]. The protective efficacy of CIV against influenza virus infection has been 70–90% in studies of vaccinated military personnel, but protection of persons in high-risk categories (e.g., the elderly) has been variable (0–80%) [1,2]. How effective is inactivated vaccine in induction of protection against challenge with a heterovariant virus? Hoskins et al. [8] showed that vaccination with the prevailing influenza virus strain provided protection against that strain, but did not induce protection against the next annual strain. Proponents of the LAV claim that the live virus

3064

B.E. Johansson, I.C. Brett / Vaccine 25 (2007) 3062–3065

vaccine induces a more broadly cross-reactive immunity than CIV [9]. The reported 86% efficacy of LAV is not significantly different from 60 to 80% effectiveness achieved by immunization with CIV [1,2,4]. Additionally, LAV is labeled for use only in young healthy individuals and use in vaccination against H5N1 avian influenza is obviated by the risk of reassortment with a wild-type strain.

2. An alternative strategy: immunity to neuraminidase The usual objective of immunization against influenza is complete inhibition of infection. Another strategy is infection-permissive immunization. The basic premise of this approach involves immunization with NA, which does not elicit neutralizing antibody but restricts viral replication allowing infection but not disease. NA administered as a purified protein with or without adjuvant [3,4] or in viral antigenic hybrids containing an HA novel to human experience (e.g., H7N2) [3], has been shown to engender protective immunity in mice [3] and in humans [4]. Immunization of mice with either N1, N2 or combined together induces a specific immune response to NA resulting in significant reduction of viral titers, lessened severity of viral infection, protection from homotypic infectious challenge, closely related heterotypic infectious challenge and a relatively distant heterotypic infectious challenge [4]. Comparative studies of NA-specific vaccines and CIV in populations primed to prevalent influenza virus subtypes have shown that H7N2 and H7N1 reassortant vaccines evoke greater NA antibody response than H3N2 and H1N1 vaccines [3]. Superiority of the H7N2 or H7N1 vaccines as immunogens for antibody to NA reflect different processing of NA when it is associated with a HA to which the study population had not been primed, presumably because subjects were primed to the HA and the anti-HA anamnestic response depresses the concomitant NA response by antigenic competition [3,10]. Because HA is found in greater molar amounts on the virion surface than NA, despite immunogenic equivalence as purified antigens, the immune response is skewed toward HA. The diminished antibody response to NA occurs only when both NA and HA are presented on the same viral particle [10].

3. An alternate strategy: vaccination with recombinant antigens To facilitate a rapid production scheme and examine the immune effects of purified HA and NA, the immunogenicity of baculovirus expressed influenza HA and NA have been examined in mice [4] and humans [11,12]. In human trials, N2-NA purified from influenza virions [11] and recombinant baculovirus expressed N2-NA [12] were evaluated for immunogenicity and toxicity in young adults. In each study, NI antibody and NA-specific ELISA anti-

body titers were proportional to the dose given. All doses were tolerated with respect to local and systemic reactions and were less than in the trivalent vaccine group. In comparison, studies examining the safety and immunogenicity of recombinant HA (rHA) vaccines in humans [13,14] have shown that higher doses (45 ␮g) of rHA vaccine compared to CIV (15 ␮g) were required to induce the same increases in frequency of four-fold increased anti-HA antibody titers in adults. However, the markedly poor response to CIV-like H5N1 [15] and recombinant H5 vaccine [16] suggest use of adjuvants may be required. The reduced immunogenicity of the baculovirus produced H3 and H1-HA vaccines compared to CIV can be explained by differences in production methods, eliminating the adjuvant effects of viral lipids and other proteins and the lack of cross-linking of antigens secondary to the use of inactivating agents such as formalin. The marked lack of immunogenicity of the H5-HA in rHA vaccine and CIV cannot be simply explained as lack specific priming in immunologically na¨ıve subjects, since the vaccines lack immunogenicity after multiple doses. Preliminary data from our laboratory suggest that the H5-HA reduces the expression of various chemokines (CCL2, CC3 and CCL4); these chemokines interact with receptors preferentially expressed on Th1 cells, suggesting that H5-HA may impair Th1 responses mediated by these molecules.

4. What are the alternatives? Effective influenza vaccines must accommodate the viral potential for antigenic change, both rapid antigenic shift and the more slowly accumulated changes in antigenic drift. Currently licensed vaccines utilizing egg-based methods have long production times and lack flexibility in changing strain type after production begins. Immunization with NA and M2e are infection-permissive; consequently similar restrictions on use as LAV may apply and rHA vaccines, as single antigen preparations will be vulnerable as CIV to antigenic drift. How do we get around these problems? First, a reliable recombinant protein expression system is needed. This will allow for more rapid antigen production and facilitate quick changes in vaccine composition should a viral strain emerge that is significantly different from the predicted strain. Secondly, the likelihood of infection from a challenge virus is correlated to the antigenic relatedness of the HA from the challenge strain to the immunizing strain and the severity of infection is mediated by the antigenic relatedness of the NAs involved, suggesting that optimal protection against influenza is afforded by antibodies against both HA and NA of the circulating strain. Immunizing mice with CIV supplemented with purified NA [11,12], H3-HA and N2-NA expressed from recombinant baculovirus [13] or DNA plasmids encoding for HA and NA [17] resulted in high titers of antibodies to both HA and NA, equivalent for each antigen to titers in animals immunized with either antigen alone. NA antibody titers, measured by ELISA and neuraminidase inhibition test

B.E. Johansson, I.C. Brett / Vaccine 25 (2007) 3062–3065

(NAI) were higher in the HA–NA recombinant vaccine than in CIV, with no differences in the anti-HA antibody titers between these two vaccine preparations. Homotypic and closely related heterotypic infections were suppressed and greater reduction in viral replication was observed following a distantly related heterotypic infectious challenge than was observed with CIV. In a murine study designed to simulate the 2003–2004 influenza season, immunization with a combination of HA and NA had the broadest protection to homotypic or heterotypic infectious challenge [4]. These data suggests vaccines immunogenic for both H3-HA and N2-NA from the vaccine A/Panama strain would have offered protection from the circulating heterotypic A/Fujian strain. A Phase 2 challenge study [12] of CIV supplemented with purified N2-NA examined efficacy of vaccines by measuring reduction of viral shedding and signs/symptoms of influenza infection. No statistically significant differences in HAI titers between CIV and NA-supplemented vaccine groups were seen. NAI titers were significantly higher in the supplemented group. Twenty-eight days after vaccination, subjects were challenged with live-attenuated virus. Subjects who received NA-supplemented vaccine reported fewer symptoms, shorter duration of those symptoms, had fewer positive cultures for influenza, and a reduction in the amount of virus shed.

5. Conclusions Recombinant protein based influenza vaccines offer several advantages over currently licensed vaccines. The baculovirus expression system allows for varied combinations of dose amounts of antigens and rapid vaccine reformulation. Recombinant HA and NA are less reactogenic and allow vaccination of those most at risk for influenza related morbidity and mortality, both are immunogenic in primed adults without adjuvant. Vaccines containing NA, especially formulations containing equimolar amounts of HA and NA presented separately, are superior to CIV and LAV when challenged with a heterovariant strain of influenza, potentially ameliorating the effects of antigenic drift. Recombinant antigens are stable when stored for long periods at 4 ◦ C, allowing stockpiling of a vaccine more likely to be effective in the face of antigenic drift or shift. H5-HA is poorly immunogenic in CIV-like and rHA vaccines and likely will require adjuvant for induction of an effective immunity. Addition of NA to vaccines will ameliorate the effects of antigenic drift. NAs from currently circulating H3N2 and H1N1 strains and N1-NA from H5N1 should be included in vaccines against influenza, a shift in the influenza vaccination paradigm to recombinant antigens should be seriously considered.

3065

References [1] Kilbourne ED. Influenza. New York: Plenum Publishing; 1987. pp. 164–5. [2] Neuman G, Kawaoka Y. Host range restriction and pathogenicity in the context of influenza pandemic. Emerg Infect Dis 2006;12(6): 881–6. [3] Kilbourne ED, Cerini CP, Khan MW, Mitchell Jr JW, Ogra PO. Immunologic response to the influenza virus neuraminidase is influenced by prior experience with the associated viral hemagglutinin. Studies in human vaccines. J Immuno 1987;138:3010. [4] Brett IC, Johansson BE. Immunization against influenza A virus: comparison of conventional inactivated, live-attenuated and recombinant baculovirus produced purified hemagglutinin and neuraminidase vaccines in a murine model system. Virology 2005;339:273–80. [5] McMicheal A, Grotch F, Cullen P, Askonas B, Webster R. The human cytotoxic T cell response to influenza vaccination. Clin Exp Immunol 1981;43:276. [6] McLaren C, Verbonitz MW, Daniel S, Guggs G, Ennis F. Effect of priming infection on serologic response to whole and subunit influenza virus vaccines in animals. J Immuno 1977;125:2679. [7] Kilbourne E, Grajower B, Johansson B. Independent and disparate evolution in nature of influenza A virus hemagglutinin and neuraminidase. Proc Natl Acad Sci USA 1990;87:786–90. [8] Hoskins TW, Davies JR, Smith AJ, Miller CL, Allchin A. Assessment of inactivated influenza A vaccine after three outbreaks of influenza at Christ’s Hospital. Lancet 1979;1:33. [9] Belshe R, Gruber W, Mendelman P, Cho I, Reisinger K, Block S, et al. Efficacy of vaccination with live attenuated, cold-adapted, trivalent, intranasal influenza virus vaccine against a variant (A/Sydney) not contained in the vaccine. J Pediatr 2000;136(2):168–75. [10] Johansson B, Moran T, Kilbourne E. Antigen-presenting B cells and helper T cells cooperatively mediate intravirionic antigenic competition between influenza A virus surface glycoproteins. Proc Natl Acad Sci USA 1987;84:6869. [11] Hocart M, Grajower B, Donabedian A, Pokorny B, Whitaker C, Kilbourne E. Preparation and characterization of a purified influenza virus neuraminidase vaccine. Vaccine 1995;13:1793. [12] Schiff G, Kilbourne E, Smith G, Hackett C, Manoff S, Matthews J. Phase 2 clinical evaluation of an influenza A virus recombinant N2 neuraminidase. In: The First European Influenza Congress Hersonessos. 2000. [13] Treanor JJ, Schiff GM, Couch RB, Cate TR, Brady RC, Hay CM, et al. Dose-related safety and immunogenicity of a trivalent baculovirusexpressed influenza-virus. J Infect Dis 2006;193:1223. [14] Powers DC, Smith GE, Anderson EL, Kennedy DJ, Hackett CS, Wilkinson BE, et al. Evaluation of a recombinant hemagglutinin expressed in insect cells as an influenza vaccine in young and elderly adults. J Infect Dis 1996;173(June (6)):1467–70. [15] Treanor JJ, Campbell JD, Zangwill KM, Rowe T, Wolff M. Safety and immunogenicity of an inactivated subvirion influenza A (H5N1) vaccine. New England J Med 2006;354(13):1343–51. [16] Treanor JJ, Wilkinson BE, Masseoud F, Hu-Primmer J, Battaglia R, O’Brien D, et al. Safety and immunogenicity of a recombinant hemagglutinin vaccine for H5 influenza in humans. Vaccine 2001;19: 1732–7. [17] Chen Z, Matsuo K, Asanuma H, Takahashi H, Iwasaki T, Suzuki Y, et al. Enhanced protection against a lethal influenza virus challenge by immunization with both hemagglutinin- and neuraminidase-expressing DNAs. Vaccine 1999;17:653–5.