The 2009 H1N1 pandemic highlighted the need for rapidly scalable influenza vaccine production and presented a unique opportunity to test a new vaccine approach in a real-life pandemic setting. The rHA vaccine was produced and released in less than 12 weeks from initial virus identification. The vaccine was well tolerated and responses were influenced positively by antigen dose and adjuvant and negatively by subject age. This study included a high proportion of elderly, obese and chronic disease subjects, not well represented in other H1N1/2009 trials, which may have negatively impacted on the overall seroprotection rates achieved with the rHA vaccine. The lower vaccine response rates observed in subjects over the age of 50 years have similarly been reported by other H1N1/2009 vaccine studies [26
]. The reduced responsiveness in older subjects in our study was improved by higher antigen doses, Advax™ adjuvant and a booster immunization. With all factors optimized (maximum rHA dose, Advax™ adjuvant, and two immunizations), subjects aged over 50 years achieved a seroprotection rate of 61.3%.
On their face, the HI results obtained with the rHA vaccine are lower than results of studies performed with conventional egg-based H1N1/2009 pandemic vaccines, where in most cases a single immunization achieved over 80% seroprotection. For example, in a Chinese study a 15ug dose of inactivated H1N1 in adults 18–60 years achieved 82.1% seroprotection [28
] and in a US study a 15 or 30ug dose of inactivated H1N1 vaccine in adults 18–64 year achieved >95% seroprotection [29
]. Equally high seroprotection levels were seen in studies of either conventional egg-based or cell culture grown inactivated H1N1/2009 vaccines [26
]. This raises the important question of why the rHA vaccine required at least three fold higher antigen doses plus an adjuvant to achieve still relatively modest levels of seroprotection which, except at the highest 45ug HA dose, did not generate HI titers sufficient to meet CHMP/FDA licensing requirements, and even at the highest 45ug dose still only met 2/3 CHMP criteria? This low immunogenicity outcome is reflected in previous studies when recombinant and inactivated seasonal influenza vaccines have been compared head to head, where 45ug dose of rHA was required to produce the same HI titers as 15ug of inactivated HA antigen [17
]. Thus, an inherent property of rHA antigen appears to be that it is a third the immunogenicity of an equivalent dose of inactivated virus antigen. The reasons for this difference is not known but may reflect the fact that inactivated vaccines contain many other components including neuraminidase and nuclear proteins, plus viral RNA, that act as inbuilt adjuvants and thereby increase the immunogenicity of inactivated HA antigen [31
]. Notably, inactivated influenza vaccines lose their immunogenicity when administered to toll-like receptor (TLR)-7 or MYD88 knockout mice, in which RNA contained in inactivated vaccines is unable to activate the innate immune system and act as an adjuvant [33
Another surprising finding in our study was the limited boosting effect of the second dose of rHA vaccine. This may partly be because of the short time window between doses, as a longer dose interval of up to 12 months between boosters has previously been shown in the case of hepatitis B vaccines to improve vaccine responses [34
]. Alternatively, given the inverse relationship between cellular (Th1) and humoral (Th2) immunity [35
], it is possible that the low-responders on HI were instead making a strong T-cell response to the vaccine, which in turn inhibited the antibody response [35
]. In future vaccine studies, we plan to also measure anti-influenza T-cell responses to examine this more closely.
Another consideration when trying to compare the study responses to those obtained with inactivated H1N1/2009 vaccines, is that inter-study comparisons based on HI titers are problematic being highly dependent on factors such as the HA antigen and red blood cell species used, and HI assays remain poorly standardized between laboratories [36
]. For example, another study reported a much higher estimate of baseline seroprotection to H1N1 2009 of 32% in the South Australian population compared to our HI assay based estimate of just 9.1% baseline seroprotection, despite the two studies being performed in the same population at the same time [26
]. Our estimate of 9.1% baseline seroprotection appears more consistent with reported US, Chinese and European data obtained early in the H1N1/2009 pandemic [27
]. Seroprotection rates measured by HI assay may understate the protection achieved with adjuvanted rHA vaccine as, for example, a single dose of Advax™-formulated H5N1 vaccine in a recent ferret study completely protected against H5N1 challenge even when there was no detectable HI titer prior to challenge [23
]. Thus, HI titers alone are an imperfect measure of influenza protection. Advax™ adjuvant has been shown to enhance memory CD4 and CD8 T cell vaccine responses, and T cells make an important contribution to protection against influenza [39
]. Although our study was not designed as an infection outcome study, it is notable that there was an extremely low rate of late seroconversions between weeks 6 and 26 (3.9%), consistent with a low rate of H1N1/2009 infection in the immunized study subjects during a period when the H1N1/2009 pandemic was at its peak.
There has been considerable debate regarding the benefits and risks of incorporating adjuvants into influenza vaccines. Aluminium adjuvants are unsuitable for influenza vaccines and paradoxically may even reduce immunogenicity[27
]. Several squalene oil adjuvanted seasonal influenza vaccines are licensed in Europe for use in the elderly[43
]. Following the commencement of our study, studies were undertaken of inactivated H1N1/2009 vaccines together with squalene adjuvants, MF59 or AS03. Squalene adjuvant provided high levels of seroprotection and dose-sparing; for example a 3.75 µg dose of inactivated HA with AS03A adjuvant achieved 94% seroprotection compared to 73% seroprotection with inactivated HA alone [44
]. Overall, vaccines containing squalene adjuvants were well tolerated apart from a propensity to increase injection site pain and muscle aches[37
]. However, the importance of developing alternative pandemic influenza vaccine adjuvants is highlighted by recent reports of an increased risk of narcolepsy in Scandinavian children immunized with the squalene-adjuvant formulated H1N1/2009 pandemic influenza vaccine (Pandemrix®) [45
]. Although the mechanism underlying the association of use of Pandemrix® with narcolepsy is not understood, and may ultimately not implicate the squalene adjuvant used, this issue reinforces the importance of developing a broader range of adjuvants for use in future influenza pandemics.
Advax™ is a novel polysaccharide adjuvant made from particles of delta inulin [20
]. In support of previous findings in animal models[21
], Advax™ significantly increased the immunogenicity of the rHA vaccine in this study. Given the difficulties of interpreting HI results from different studies we compared, in the same assay, HI titers from the H1N1/2009 vaccine study to convalescent H1N1/2009 patient sera [46
]. This confirmed that subjects given the highest dose of rHA plus Advax™ adjuvant achieved comparable HI titers to patients recovered from clinical infection with H1N1/2009.
Enhanced vaccine immunogenicity should not be at the expense of tolerability or safety[47
] and in this respect the tolerability and safety profile of Advax™ was reassuring. The lower rate of post-immunization headaches in subjects receiving Advax™ was unexpected, as studies of most adjuvants reveal an increase in headaches[48
]. Reduced headaches may relate to adjuvant-induced changes in cytokine production. Immunization headaches are likely to be mediated by inflammatory cytokines, in particular interleukin (IL)-1, as IL-1 serum levels are increased in cluster headaches and IL-1 gene polymorphisms (3953 C/T) are associated with migraine headaches[49
]. Unlike pro-inflammatory adjuvants, Advax™ adjuvant does not induce IL-1 gene expression (N. Petrovsky, unpublished data) and this might explain why it doesn’t exacerbate post-immunization headaches.
The H1N1/2009 pandemic reinforced the need for innovation in influenza vaccine design and manufacture. Our results confirm the utility and speed of a recombinant vaccine approach to pandemic vaccine production, although there remains a need to optimise the immunogenicity of the recombinant antigen. However, whilst the overall HI data obtained in the current study of recombinant hemagglutinin were modest by reference to CHMP/FDA licensing requirements, this does not necessarily mean that the recombinant vaccine was not effective. Whilst high vaccine-induced HI titers are generally predictive of protection, the converse is not necessarily true, namely low vaccine-induced HI titers do not necessarily predict lack of protection. This fact was clearly demonstrated when ferrets immunized with a H5N1 vaccine formulated with Advax™ adjuvant were completely protected against lethal H5N1 infection despite having no detectable HI titers to H5N1 prior to challenge[23
]. What this clearly demonstrates is the urgent need for better assays of immune correlates of influenza vaccine protection to replace traditional HI assays in the licensing of new influenza vaccines. Nevertheless, there is also scope for further enhancement of the rHA approach through further adjustments in antigen and adjuvant dose to help them better meet current CHMP/FDA licensing requirements, and the potential inclusion of recombinant neuraminidase protein to enhance heterotypic immunity [47
]. With such strategies to improve immunogenicity in place, a rHA pandemic vaccine may overcome many of the problems of traditional inactivated influenza vaccines and thereby provide benefit in future pandemics.