2.1. Background; Seasonal Vaccines
Influenza vaccination is the most effective prophylactic measure to prevent morbidity and mortality (reviewed in [16
]). There are two main types of influenza vaccine available; inactivated vaccine delivered deep subcutaneously or intramuscularly and live attenuated vaccine administered intranasally. Parenterally administered inactivated influenza vaccines have been used for many decades and extensive information is available on their quality and safety, being about 60-80% effective in preventing disease against homologous or closely related strains. In some cases, vaccination may not prevent influenza morbidity, but rather reduce its severity and duration. Inactivated vaccines are available in whole, split (chemically disrupted), subunit (purified surface glycoproteins) and virosomal/virus-like particles (VLPs) containing surface glycoproteins formulations.
Current inactivated vaccines are mostly produced by propagation in embryonated hens’ eggs. The allantoic fluid is harvested, and the virus is concentrated and highly purified, then inactivated by formaldehyde or beta-propiolactone, or eventually by the detergent disruption procedure itself. The availability of embryonated hens’ eggs is a limiting factor in vaccine production and the global manufacturing capacity is not expected to meet pandemic vaccine demands. It is therefore important to develop dose-sparing strategies by using effective formulations and/or adjuvants. Some manufacturers are using cell cultures as vaccine substrate (MDCK, Vero, PerC6®), either as their only platform or as a supplement, allowing more expediently to scale up of production (Table 1
The use of reverse genetics technology can save a considerable amount of time in generation of pandemic seed virus. Traditionally, influenza vaccine seed viruses have been made by classical reassortment by choosing a virus isolate which closely matches circulating strains, and introducing its surface glycoprotein gene segments (HA and NA) into the genetic background of the laboratory-adapted and high-yielding “PR8” virus, A/Puerto Rico/8/34 (H1N1). Seed viruses are produced by co-infection of the two “parent” viruses, and screening for the progeny of interest. This is a time-consuming process and not always successful. And even so, the growth characteristics may not be acceptable. The use of reverse genetics allows viruses to be constructed with a predefined set of genes, following combination of plasmid DNAs encoding separate RNA segments. Site-directed mutagenesis allows engineering of the DNA transcripts to create viruses altered in specific genes. Highly pathogenic avian strains can be attenuated to create a depathogenised virus, which can be used as a vaccine strain, e.g. NIBRG-14 derived from A/Vietnam/1194/2004 (H5N1), which has been extensively used in human clinical trials. However, some of the recent vaccine reassortants vaccine strains for the pandemic H1N1 virus, whether prepared by classical reassortment or by reverse genetics, have been reported to give low yields.
Main properties of the most widely used influenza vaccines.
Main properties of the most widely used influenza vaccines.
|Type of influenza vaccine||Substrate 1)||Route of administration||Immune response 2)||Immunological challenges||Comments|
|Inactivated||No apparent stimulation of mucosal immunity|
|Whole virion||Strong, mixed Th1/Th2|
|Split||Embryonated eggs, cell culture ||intramuscular||Mostly Th2||Should be adjuvanted||Supply of embryonated eggs is a limiting factor. Cell-grown vaccine production more scalable.|
|Subunit (HA and NA)||Mostly Th2||Least immunogenic, should be adjuvanted|
|Live ttenuated||Embryonated eggs||intranasal||Humoral, cellular, mucosal||Should perform better in the elderly||Rapid and large vaccine output. Easily scalable production. Cannot be used in the very young and the immunocompromised.|
Live attenuated seasonal vaccines have been used in Russia since the 1970s and have been licensed for use in the USA since 2003. There are two strains A/Leningrad/134/17/57 (H2N2) and A/Ann Arbor/6/60 (H2N2), which have been used as donor strains. They are attenuated, genetically stable, non-transmissible, safe, immunogenic, and provide protective immunity. Importantly, live attenuated vaccines can be rapidly manufactured and provide more vaccine doses per egg than inactivated vaccines, and they are conveniently administered by intranasal spray devices. Live attenuated vaccines induce a good mucosal response, which upon challenge with field strains may limit their initial replication in the upper respiratory tract, the viral portal of entry. These vaccines perform well in children, reduce laboratory confirmed infection in adults, but are less immunogenic in the elderly. A comprehensive review of the current status of attenuated live influenza vaccines is given by Ambrose et al.
2.3. Immunological Challenges
In contrast to the current H1N1 pandemic, an eventual pandemic caused by the avian H5N1 would have met a virtually immunologically naïve global population and potentially caused a very high number of excess deaths. The compiled data for the zoonotic H5N1 cases suggest an overall case-fatality rate of 59% (262/442), with Indonesia standing out with 82% (115/141) [54
]. In the event of a sustained human-to-human spread, it must be assumed that the fatality rates will not reach such high levels. The antigenic and genetic diversity of the avian H5N1 is a particular challenge for vaccine development. There are substantial antigenic differences between clades and subclades having consequences for vaccine development. So far, the WHO advocates that viruses from clades 1 and 2 should be used as vaccine seeds [55
]. Whether or not an H5N1 pandemic will occur we cannot know, of course, nor is it possible to predict which clade and strain will eventually emerge as the pandemic virus. This makes pre-pandemic preparations of trial lots of H5N1 vaccines complicated. On the other hand, this uncertainty has stimulated research and development work, particularly when it comes to generating a strong cross-reactive vaccine response. Consequently, it has highlighted the importance of generating cellular immunity, predominately elicited against the internal and less variable antigens of the virus, namely the type-specific NP and M1 proteins, thus eliciting an immune response being more robust to antigenic changes.
The H1N1 pandemic situation is in many ways the complete opposite to the H5N1 threat. While the number of highly lethal zoonotic H5N1 cases has slowly continued to increase throughout the recent years, the virus still has not adapted sufficiently to allow effective human-to-human transmission to sustain community-level outbreaks. For the H5N1 virus we have been in the WHO pandemic phase 3 for six years (since 2003). It may well be that pandemic viruses not necessarily must emerge via an human adaption process from a zoonotic case, or as a result of a singular reassortment event between an animal and human strain. There could also be several reassortment incidents, all spaced in time and place, and involving different animal species, as highlighted for the recent H1N1 pandemic by Smith et al
]. We should therefore not lose sight of the H5N1 threat.
Ideally, a candidate pandemic influenza vaccine should elicit a rapid and strong humoral and cell-mediated immune response, and ideally also a strong mucosal response, which are long-lasting and exhibit broad cross-reactivity against drifted strains (within and across different clades). To date, most published results from animal studies and clinical trials of parenteral pandemic avian influenza H5N1 virus vaccines have highlighted the need for high antigen doses or an effective adjuvant to elicit a satisfactory antibody response, as highlighted by Hehme et al.
] and Treanor et al.
It is widely accepted that serum antibodies, directed against the HA, correlate with protection against seasonal influenza. However, there is convincing evidence, that both CD4+
T helper cells and CD8+
cytotoxic T lymphocytes (CTL) may play an important role in controlling viral infection and reduce the severity of disease and decrease mortality [57
]. It has been well established that purified viral proteins formulated with oil-in-water adjuvants, stimulate the Th1 arm of the post-vaccination response. The same is also seen for inactivated whole virion vaccines and VLP formulations. It is assumed that viral proteins are entering the antigen-presenting cell via the endosomal pathway into the cytosol for subsequent presentation of viral peptides on MHC Class I, thus mimicking the infectious process. Also vaccine formulations with ISCOMs have been extensively studied and demonstrated CTL responses (reviewed by Rimmelzwaan et al
]. Therefore, the use of whole virion vaccines formulations, VLPs and/or the use of Th1 stimulating adjuvants would greatly improve vaccine performance.
Current vaccine seeds, whether prepared by classical reassortment or by reverse genetics, will, - by design,- not contain the NP and M1 proteins from the epidemic/pandemic strain, but from an historic laboratory reassortment partner. One may speculate whether this could constitute a disadvantage when aiming at stimulating a more authentic CTL response. One manufacturer (Baxter) uses the unmodified field virus as seed strain for their cell-grown inactivated whole virus vaccines, and in the context of cellular response this approach may well turn out to be especially advantageous.
The three pandemic H1N1 clinical studies just published [50-52
] are mostly presenting preliminary results. The finer details of the immune response, e.g. Th1/Th2 profile and cross-reactions with emerging variant strains, will hopefully be presented in due time. Regarding the need for one or two doses, even if one dose satisfied licensing requirements, only follow-up studies of longevity of antibodies and humoral and eventual cellular memory will clarify this issue.
The poor immunogenicity of pandemic vaccines against H5 and H7 avian subtypes appears not to have repeated itself for the currently used H1N1 pandemic vaccines. This is a comforting finding.
Mucosal immunity may also provide protection against infection per se
and not only protection against illness. Therefore, needle-free mucosal influenza vaccine is an attractive approach, which may provide immunity at the portal of virus entry. Additionally, mucosal IgA has broader specificity than serum IgG and may provide better cross-clade protection against drifted influenza strains [60
]. However, although demonstrating stimulation of secretory IgA when using intranasal route for administration of influenza antigen alone, i.e.
a non-live formulation, the immunogenicity is frequently poor [62-64
]. Use of an appropriate nasal adjuvant could improve immunogenicity and thus reduce the required dosage [65
]. One should, however, not forget the many cases of Bell’s palsy associated with a recent clinical trial involving intranasal vaccination [66
Therefore, safe and effective mucosal adjuvants are urgently needed. A recent murine study with sublingual administration of an adjuvanted H1N1 vaccine has shown good mucosal immunity reflected by high IgA levels in the respiratory tract and provided protection against viral challenge [67
]. Sublingual administration of a pandemic vaccine may significantly contribute to the pandemic preparedness through its ease of administration and better public compliance. Additionally, sublingual vaccines are less likely to induce neurological disorders and most probably a safer alternative than the intranasal vaccine for mucosal delivery, avoiding the potential exposure of the olfactory bulb [66
]. Provided not excessive use of antigenic material is required, this alternative route should have much to offer and warrants further studies.
Few reports have addressed the long-lasting immunity and memory response after pandemic influenza vaccines [38
]. Most of the published data evaluated the response in the immediate weeks after vaccination, which may underestimates the long-term protection against influenza viruses. However, it is difficult to assess the precise longevity of vaccine-induced immunity in the field, as influenza viruses undergo a continual antigenic variation (‘antigenic drft’). It is for this reason seasonal vaccination is recommended annually. Therefore, long-term humoral and cellular immunity after pandemic vaccines should be evaluated considering that more than one pandemic wave may occur [68
2.4. New Production Platforms
Despite recent decades’ many new technical-scientific advances, the influenza vaccine production platforms have essentially remained unchanged [69
]. While egg-grown influenza vaccines in the 1950’ties where initially rather reactogenic, subsequent introduction of zonal centrifugation purification steps, and also the introduction of detergent-split virions, greatly reduced its reactogenicity. Throughout the later parts of the 20th
century the global production capacity, and the actual use of vaccine against seasonal influenza, has steadily increased. Most manufacturers still grow viruses in embryonated hens’ eggs, whereas a small number of companies use cell cultures (Vero, MDCK, PER.C6®), either as their only substrate or as a supplement to eggs. The use of embryonated eggs is a critical point, as an emergency decision to scale up the production may not easily be accomplished as quality assured eggs must be ordered many months in advance. Adding to this complexity is the season-dependent quality of embryonated eggs, affecting the resultant viral yield. Production platforms based on approved cell lines are more controllable and more straightforward to scale up. One should not forget, however, that the availability of approved sites for bioreactors might potentially be a limiting factor. The fragile egg-based system is demonstrating its shortcomings during times of a pandemic urgency, such as today. Also, high pathogenic avian influenza viruses may add to the complexity by making the egg-laying flocks more vulnerable. Adding to this is the rare, but very critical consequences for vaccine output, when bacterially contaminated eggs are entering the production lines. This was clearly demonstrated in 2004 [70
]. There is therefore a need for several new, or at least supplementary, vaccine production systems being more robust, more flexible and easier to scale up.
While some plasmid DNA vaccines have shown some promising results in animal models, there is still a way to go for generating effective human DNA vaccines. There is also a public acceptance issue to consider. In contrast, recent years industrial use of recombinant DNA technology, allowing generation of viral proteins to be made in large quantities in various host cells, hold promise for future vaccine technologies. In the context of influenza both DNA vaccines and recombinant protein preparations produced in insect and E.coli
cell cultures have been used, some of them have got into clinical trials [71-75
]. Although the strategy of producing edible vaccines has not been abandoned, the transient expression of engineered vaccine proteins in plant cells has shown promise. Such low cost systems are considered safe and can easily be scaled up (see review [76
]). None of these new platforms have hitherto made any noteworthy contribution to the global annual influenza vaccine output. This will probably change in the years to come. Still, of the more than 70 pre-H1N1 registered trials with vaccines against pandemic influenza, several have used DNA vaccines, recombinant proteins made in E. coli
and by engineered baculovirus in insect cells [18
]. The most attractive property of the new production platforms is their claim to produce very large quantities of vaccine material within a short time. This is exemplified by a recent press release from VaxInnate Corp [77
]. Using engineered bacterial flagellin, a TLR agonist, carrying viral sequences and targeting antigen presenting cells, they claimed that 300 millions H1N1 doses could be produced in weeks rather than months.
2.5. Limited Supply and Global Equity
For a pandemic, the ultimate objective is to produce enough vaccine to immunize the world's entire population (6.7 billion people) with 2 doses, within 6-9 months, anticipating 5 µg HA/dose [78
]. As mentioned previously, the total global output capacity for trivalent seasonal vaccines, requiring 15 µg HA for each strain, is estimated to reach approximately 1 billion doses by 2010 [79
]. Assuming that all production capacity is completely switched to manufacture pandemic vaccines, and that 2 doses of 5 µg HA per dose is required, this will translate to approximately 4.5 billion immunization courses, representing about 67% of the global need. Such calculations will of course depend heavily on whether 2 doses are necessary for all age groups, which dose strength is required to satisfy licensing requirements, and whether dose-sparing adjuvants will be used. (Table 3
). Considering the mixed type/subtype/strain influenza epidemiology during recent months, it is reasonable to assume that both A/H3N2 and B viruses, and possibly also the seasonal A/H1N1 virus, will continue to circulate. The calculations presented in Table 3
are therefore only representing a best case and probably also an unlikely scenario where manufacturers only have to consider one influenza vaccine strain.
Different scenarios for global availability of (monovalent) pandemic vaccine 2010. Vaccine courses in millions
Different scenarios for global availability of (monovalent) pandemic vaccine 2010. Vaccine courses in millions
|1 dose to all||2 doses to all||70% 1 dose, 30% 2 doses|
|HA per dose||No of vaccination courses||Global coverage||No of vaccination courses||Global coverage||No of vaccination courses||Global coverage|
|5 µg||9,000||>100%||4,500||67%||6,900||>100 %|
Most importantly, for these calculations it is assumed that the growth characteristics of the seed virus will be about as good as recent years’ vaccine strains. WHO published an overview of viral characteristics with recommendations for vaccine development as early as 26 May, 2009 [80
]. However, the growth characteristics of the pandemic H1N1 seed strains so far having being distributed to manufacturers have been questioned, thus the initial estimates of number of available doses may well be less than was optimistically suggested in Table 3
. Still, there is substantial dose-stretching potential if the HA dosage can be reduced. Based on recent studies [50-52
] it is also apparent that if certain age-cohorts only need one and not two vaccine doses, tentatively shown as 70% of the population in Table 2
, there is a saving to be gained. Also, the use of intradermal immunization may offer savings in the order of about 80% in antigen usage [81-83
]. When there is a global scarcity of vaccine, particularly when facing a pandemic, such an approach should be seriously considered. It should also be noted that the calculations in Table 3
show the theoretical capacity to provide a pandemic H1N1 vaccine for the global population, varying from 22% to full coverage. In reality, the situation will be quite different, since the doses will not be delivered in full at one specifically set date. According to calculations by the WHO approximately 50% of total output will be available within 6 months [84
Based on currently used technologies there is a five-six month lead-time from having identified the pandemic strain to release of the first vaccine lots [85
]. This includes generating seed virus (by classical reassortment or by reverse genetics), preparing reference sera and antigens for potency standardization purposes, optimizing manufacturing processes, bulk manufacture, quality control, vaccine filling, clinical studies and regulatory approval [86
]. For the preparation of seed virus, the use of reverse genetics in lieu
of classical reassortments, has the potential of cutting several weeks off this time-line. Furthermore, the licensing hurdle can be dramatically shortened if the manufacturer has submitted and received approval for a mock-up process with the same subtype (but a different strain) without any change of the manufacturing process.
Clearly, using classical production platforms the necessary total global influenza vaccine output could possibly be within reach within the next few years, but the long lead-time required to provide enough vaccine in time for the entire global population is not acceptable. Following the H1N1 pandemic some countries have activated their dormant contractual arrangements with the pharmaceutical industry for prioritized delivery of vaccines for their populations. Global solidarity issues will be raised, and rich countries with such exclusive contracts will be asked to share their vaccine allocations with less affluent countries. While national politicians no doubt will battle with this quandary, it is reassuring to know that several vaccine manufacturers have pledged donations of millions of doses of pandemic vaccine to the WHO for subsequent distribution to low-income countries. To what extent national authorities, particularly those hosting vaccine manufacturers, will activate emergency statutes and ban export of such a life-saving commodity before their domestic need is satisfied, remains to be seen. It will most certainly depend upon how the clinical pattern of the H1N1 pandemic develops, assuming that a less virulent strain will translate into more generosity