Australian Public Assessment Report for purified antigen fractions of inactivated split virion A/Indonesia/05/2005 (H5N1), as03 adjuvanted

III. Nonclinical findings

Introduction

General comments

The sponsor has applied to register a prepandemic split influenza vaccine, Prepandemrix, for prophylaxis of influenza caused by the H5N1 strain with pandemic potential in adults 18 years old and above. The vaccine strain is A/Indonesia/05/2005 (H5N1/PR8-IBCDC-RG2, a clade 2 representative). A previous submission was made to register a prepandemic vaccine based on an H5N1 clade 1 strain, A/Vietnam/1194/2004 NIBRG-14, a submission was also made to register the vaccine as a mock-up pandemic vaccine (Pandemrix), which was approved on 4 June 2008. Subsequent applications were made to extend Pandemrix use to the elderly. Both Prepandemrix and Pandemrix were approved by the EU in 2008 for use in adults, the Prepandemrix strain was changed to A/Indonesia/05/2005 in 2009.

The sponsor also manufactures a split seasonal trivalent influenza vaccine, Fluarix, which is registered in Australia. Prepandemrix monovalent bulk is manufactured by an identical process to Fluarix in Dresden. In accordance with the prepandemic and pandemic guidelines (below), additional toxicity testing was not performed with the A/Indonesia/05/2005 antigen for the current application.

The sponsor submitted previously evaluated nonclinical immunogenicity, ferret homologous and heterologous lethal challenge studies, repeat dose toxicity studies for the split H5N1/AS03 vaccine; genotoxicity studies for AS03 adjuvant; and a rat embryofetal and postnatal development study with split H5N1/AS03 and whole H5N1/Al vaccines and AS03 adjuvant. Immunogenicity, repeat dose toxicity and local tolerance studies were also submitted for seasonal trivalent influenza/AS03 vaccines. Studies of the mode of action of AS03 adjuvant were also previously evaluated for Arepanrix H1N1 vaccine. Safety studies were Good Laboratory Practice (GLP) compliant.

New studies in the current submission consisted of an integrated summary of the actions of AS03, a cardiovascular and respiratory study of AS03 in conscious dogs, and a study of preimplantation loss in rats (Table 1).

Table 1: New and previously evaluated nonclinical studies.

The previous submission from the sponsor to register Prepandemrix was withdrawn by the sponsor following a negative recommendation by the Advisory Committee on Prescription Medicines (ACPM). According to the sponsor’s application letter dated 5 February 2014, primarily over: (i) inadequate information on duration of immunity and no information on booster doses, and (ii) a considered unfavourable risk/benefit ratio based upon concerns about a potential safety signal in the elderly population (a higher number of New Onset Chronic Disease cases that was reported in Study H5N1-008 in adults > 60 years old, following administration of AS03 adjuvanted H5N1 vaccine when compared to Fluarix).

The AS03 adjuvant in Prepandemrix is composed of squalene, D,L-α-tocopherol (vitamin E) and polysorbate 80. AS03 adjuvant has previously been tested in humans in Fluarix, and a candidate malaria vaccine. In the initial application for Prepandemrix, data on the mechanisms of action of the adjuvant were limited to a demonstration that it induced pro-inflammatory cytokines in human peripheral blood mononuclear cellS (PBMCs) in vitro. However, numerous nonclinical studies on the mode of action of the AS03 adjuvant were subsequently submitted in an application to register the H1N1 pandemic split influenza (AS03 adjuvanted) vaccine, Arepanrix. The current submission contained a summary of the mode of action of AS03 with new and previously evaluated data. These data were published.²

Vaccine guidelines

The European Medicines Agency (EMA) has issued guidelines for preclinical testing of vaccines,³ prepandemic vaccines,⁴ and new adjuvants in vaccines,⁵ which have been adopted by the TGA. During the course of this evaluation, a new consolidated draft guideline on influenza vaccines nonclinical and clinical module was issued by the EMA for consultation.⁶

Submission quality

The submission consisted of several new and previously evaluated nonclinical studies. Overall, the nonclinical studies met the general requirements of the relevant EMA vaccine, prepandemic influenza vaccine and adjuvant nonclinical guidelines.

Pharmacology

Vaccine reference virus

The initial vaccine development was carried out with a vaccine containing purified antigen fractions of inactivated split virion A/Vietnam/1194/2004 NIBRG-14 (H5N1). However, based on geographical spread, epidemiology and the antigenic and genetic properties of H5N1 viruses isolated from humans during early 2007 (and WHO recommendations at that time), the sponsor decided to shift the strain used in production from A/Vietnam/1194/2004 (clade 1) to A/Indonesia/05/2005 (clade 2). Clade 2 viruses are still predominant in 2014.

Immunogenicity

Immunogenicity studies in primed mice and pigs demonstrated the capacity of the AS03 adjuvant to significantly increase HI and neutralising antibody titres, an increase in CD4 T cells was also seen in mice, and cross reactivity was seen between antibody and cell mediated responses in mice administered vaccine derived from A/Vietnam/1194/2004 and A/Indonesia/05/2005.

Ferret challenge studies

The primary pharmacology of Prepandemrix was evaluated in the previous Prepandemrix H5N1 evaluation report, and included lethal challenge studies in ferrets with heterologous and homologous H5N1 virus. The following conclusions were made by the evaluator:

In ferret lethal challenge studies with either the vaccine parent virus (A/Vietnam/1194/04 Clade 1) or a heterologous H5N1 virus (A/Indonesia/05/05, Clade 2), the split H5N1/AS03 vaccine did not prevent infection, but it reduced or prevented mortality and morbidity, and substantially reduced lung virus titres. The lowest vaccine dose of 0.6 μg HA combined with the human dose of AS03 adjuvant is ~2x and 15x human dose in mg/m², respectively, hence the studies provide limited evidence of efficacy.

Mechanisms of action of AS03 adjuvant

This was initially addressed by the sponsor in the previous Prepandemrix H5N1 submission and was followed with additional numerous studies in a submission to register the pandemic split influenza, AS03 adjuvanted Arepanrix H1N1 vaccine. The following assessment is based mainly on extracts from these previous evaluations.

The AS03 adjuvant serves as an antigen sparing measure, and to compensate for the low immunogenicity of H5N1 virus in naїve humans. AS03 adjuvant is an oil-in-water emulsion containing two biodegradable oils, squalene and d,lα-tocopherol (vitamin E). The emulsion particles are 120-180 nm in size. Squalene occurs naturally in plants, animals and humans, and shark liver oil has been a major source for vaccines. Squalene is an intermediate metabolite in the synthesis of cholesterol in humans. It is the main component (by weight) of another adjuvant, “MF59” (Novartis), in the seasonal trivalent influenza vaccine Fluad, currently marketed in 12 European countries, and Focetria, a mock up pandemic influenza vaccine approved by the EMA.

Many older adults, regardless of their vaccination history, have low titers of naturally occurring antibodies that react with squalene.⁷ Squalene was implicated in the so-called Gulf War Syndrome, although it was not a component in vaccines administered to veterans. In a review of this implication, the WHO Global Advisory Committee on Vaccine Safety (6-7 June 2006) concluded that “... fears of squalene in vaccine inducing pathological anti squalene antibodies are unfounded”, but recommended careful post market follow-up to detect any vaccine related adverse events (AEs) in other age groups.

A series of studies investigated the actions of AS03 adjuvant. There was no detectable physicochemical interaction between antigen and adjuvant upon mixing, or intramuscular (IM) injection and transport to the draining lymph node in mice, that is, no antigen entrapment. In mice the adjuvant acted as an immunostimulant, increasing the number of antigen presenting cells (APCs), the proliferation of antigen expressing T cells, and the expression of co-stimulatory CD80, CD86 and CD40 molecules and pro-inflammatory cytokines (IL-6, IFN-γ, TNFα) by APCs (mainly macrophages and dendritic cells) in the draining lymph node. Serum levels of the pro inflammatory cytokines IL-6 and MCP-1 peaked within 10 h, and declined over the following 48h+ in mice. The induction of these pro inflammatory cytokines was not mediated by Toll like receptor 4 (TLR4) (MPL adjuvant is reportedly a TLR4 agonist). When H5N1 antigen and AS03 were injected in separate legs in mice, the pro inflammatory effects (elevated serum IL-6, MCP1) were preserved, but the adjuvant effect (HI response) was lost.

Measurement of Th1 cytokine (Fin) and Th2 cytokine (IL-5, IL-13) secretion from restimulated spleen cells in naïve mice immunised with ovalbumin antigen and AS03 indicated a mixed T helper (Th1/Th2) response.

The presence of α-tocopherol in AS03 adjuvant was shown to increase levels of pro inflammatory cytokines, and antibody responses, with no significant effects on antigen uptake or the expression of co-stimulatory molecules. A more balanced Th1/Th2 cytokine response was observed in the presence of α-tocopherol.

In conclusion, AS03 adjuvant primarily acts as an immunostimulant, with transient effects on multiple co-stimulatory molecules and pro inflammatory cytokines. Although AS03 does not act as an antigen depot, its action requires the antigen and adjuvant to be in proximity upon injection. The mode of action of AS03 is illustrated in Figure 1.

Figure 1: Mode of action of AS03 in the Prepandemrix H5N1 adjuvanted vaccine.

Safety pharmacology

A study in anaesthetised rats was previously evaluated for the Prepandemrix H5N1 submission. The following conclusions were drawn:

A safety pharmacology study in which anaesthetised rats were administered a split trivalent influenza (seasonal) vaccine adjuvanted with AS03 (3 μg HA/strain + 50 μL AS03) showed no significant cardiovascular or respiratory effects up to 2 h post dose after IM injection. The AS03 dose in rats was ~8x the human dose, adjusted for body surface area. CNS effects were not assessed in safety pharmacology studies, but pharmacodynamic and toxicity studies showed no indication of CNS effects.

The current submission contains a new safety pharmacology study investigating cardiovascular and respiratory effects of AS03 alone in conscious beagle dogs administered a full human dose (0.5 mL volume) of AS03 (~⁷x human dose on a mg/kg basis in a 10 kg dog). There was slight bodyweight loss in 2/4 dogs (~5% compared to ~23% gain in the other 2/4 AS03 treated dogs), associated with a slight decrease in food consumption. Given the relatively small total of 4 treated dogs, a relationship to treatment could not be excluded. A small but statistically significant increase (+ 0.5°C compared to pre-test) in body temperature was observed at 6 h post dose only. There were no relevant treatment related cardiovascular or respiratory findings up to 72 h post dose following the single IM injection.

Pharmacokinetics

No pharmacokinetic studies were conducted with the vaccine, in accordance with the relevant vaccine guidelines. The biodistribution of the AS03 adjuvant was previously investigated as recommended by the relevant vaccine guidelines.

Toxicology

Toxicity was assessed in the previous Prepandemrix H5N1 evaluation report. The following conclusions were drawn with regard to the repeat-dose toxicity studies:

Rabbits administered 4 consecutive fortnightly doses of the split H5N1/AS03 vaccine, containing 30 μg of H5 antigen and the human dose of AS03 adjuvant, had a transient inflammatory response to the adjuvant, but no other systemic effects. Local reactions to the split H5N1/AS03 vaccine were more marked than with a registered, seasonal trivalent influenza vaccine, due to the adjuvant. Most local effects resolved within 28 days. The toxicity studies were adequate, although the adjuvant was not fully tested in a second species, and individual components of the adjuvant were not tested, as recommended by the EMEA adjuvant guideline.

Narcolepsy

Epidemiological studies in several European countries including Sweden, Finland and the UK⁸ have reported that a monovalent pandemic influenza vaccine, Pandemrix H1N1 manufactured by the sponsor in Dresden, and used in an estimated 31 million Europeans during the 2009 H1N1 influenza pandemic, has been associated with narcolepsy in children aged approximately 5 to 20 years of age. Narcolepsy is a rare sleep disorder seen almost exclusively in individuals who are HLA DQB1*0602 allele carriers, which suggests that it might be an autoimmune disorder. In the under 20 years population the absolute attributable risk increase of narcolepsy was approximately 1.4 to 8 additional cases per 100,000 vaccinated individuals compared to background rates of 0.12 to 0.79 per 100,000 children/adolescents per year. The sponsor stated:

There is little doubt on a temporal association between receipt of Pandemrix H1N1 during the period of the pandemic and the occurrence of narcolepsy among children aged approximately 5 to 20 years old.

A review of this issue by the EMA concluded that a causal relationship was not established, and that further studies were required.⁹ Further EMA review concluded that the role of the vaccine antigen and its adjuvant on the association between Pandemrix antigen and narcolepsy remains unknown.¹⁰ The sponsor is conducting a study in Quebec Canada of the other GSK AS03 adjuvanted vaccine, Arepanrix H1N1, which was also used in the 2009 pandemic. No nonclinical studies were submitted in relation to narcolepsy. A clinical statement regarding the epidemiology data was proposed in the Precautions section of the Pandemrix PI, based on an approved statement for the registered Pandemrix H1N1.

Genotoxicity

The AS03 adjuvant was negative in previously evaluated, adequate in vitro and in vivo genotoxicity tests.

Carcinogenicity

No carcinogenicity studies were conducted, in accordance with the relevant vaccine guidelines.

Reproductive toxicity

A new reproductive toxicity study was performed with a pandemic AS03-adjuvanted H1N1v vaccine (Dresden, split inactivated influenza antigen [(A/California/7/2009)/AS03A]). Six daily IM injections (GD0-6) of the adjuvanted vaccine or the AS03 adjuvant alone (100 µL each, 1/5th human dose), administered to female rats during the early stage of pregnancy did not result in any statistically significant intergroup differences on pre implantation loss or on litter values in general (including implantations, resorptions [early, late and total], live embryos or pre/post implantation losses).

Mild maternal toxicity was also evident as reductions in maternal bodyweight gain/food consumption, and swelling at the injection site (which reversed at a greater rate in animals receiving AS03A alone compared with the group receiving H1N1/AS03). Both findings were more severe in the vaccine (AS03 adjuvanted) group. These findings were anticipated consequences of a daily IM injection for 6 days, a situation which would not occur clinically. Anti H1N1 antibodies were detected in all vaccine treated dams (that is, 100% seroconversion), but no anti H1N1 antibodies were detected in the AS03 only group or the control group.

Reproductive toxicity was also evaluated in the previous Prepandemrix H5N1 submission. The following is taken from the ‘Assessment’ section:

The prepandemic vaccine and adjuvant guidelines recommend animal reproductive toxicity studies. The sponsor conducted a rat embryofoetal and postnatal development study, in which females were administered 4 or 5 IM doses of the split H5N1/AS03 vaccine, containing 6 μg HA and 2/5 of the human dose of AS03 adjuvant, or whole H5N1/Al vaccine, or AS03 adjuvant only. The first dose was administered 30 days prior to mating, and the rest 6, 8, 11 and 15 days after mating (the vaccine or adjuvant were not administered during lactation). Half the rats were sacrificed on GD 20 for foetal examinations, and the remainder raised their pups to PND 25, and pup development was assessed. All vaccine treated dams, their foetuses and pups developed anti H5N1 antibodies, and increases in pup antibody levels between days 4 and 25 indicated antibody transfer in milk. There were no significant toxicological effects on the dams, or their foetuses or pups.

Australian pregnancy category

The sponsor has proposed an Australian pregnancy category of B2 for Prepandemrix. The embryofoetal and postnatal development studies did not show evidence of foetal damage. Testing in a single species is consistent with the draft EMA influenza vaccine guideline¹¹ (not yet adopted by TGA) and the relevant FDA guideline¹² (not adopted by TGA). Since the embryofoetal and postnatal development studies were conducted with AS03 adjuvanted A/Vietnam/1194/2004, Fluarix and Flulaval (H1N1, H3N2, B) and H1N1v derived vaccines, rather than the candidate A/Indonesia/05/2005 (H5N1)/AS03 vaccine, category B2 is more appropriate than category B1.

Local tolerance

Local tolerance was assessed in the previous Prepandemrix H5N1 evaluation report. The following conclusions were drawn:

The relevant guidelines recommend local tolerance studies for prepandemic and pandemic vaccines, and for new adjuvants with and without the proposed antigen. Local tolerance was investigated in both repeat-dose toxicity studies in rabbits, and a specific local tolerance study with a seasonal trivalent influenza/AS03 vaccine in rabbits. These studies had some limitations. In the repeat-dose toxicity study with the split H5N1/AS03 vaccine, the group treated with the adjuvanted vaccine was administered 0.5 mL in each leg, whereas the groups administered only AS03 adjuvant or H5N1 antigen were administered a single dose of 0.5 mL. In the rabbit local tolerance study with the seasonal trivalent influenza/AS03 vaccine, the antigen and adjuvant were injected into opposite thighs, at two dose sites, and the recovery period was only 4 days, whereas the 2 toxicity studies had 28-day recovery periods.

The studies showed that the presence of AS03 adjuvant in either the split H5N1 or seasonal influenza vaccine resulted in some erythema and/or oedema for up to 48 h, and increased incidences and severity of injection site fasciitis and perivascular cuffing, which had not fully resolved after 28 days. The severity of local reactions was unrelated to the dose of trivalent HA antigen, indicating that it was caused by the AS03 adjuvant. Local reactions to Fluarix seasonal trivalent influenza vaccine were comparable to saline controls.

The marked antigen sparing effect of AS03 adjuvant is offset by increased local reactivity in comparison with seasonal influenza vaccines, but some increase is acceptable for a vaccine for a potentially life-threatening infection.

Paediatric use

Adults from the age of 18 years are proposed to receive two doses of 0.5 mL Prepandemrix. Clinical experience in children is limited. No specific studies in juvenile animals were submitted. A small, but statistically significant body temperature increase was observed in dogs administered a single dose of AS03 adjuvant alone, which is consistent with the pro inflammatory properties of AS03. However, the relevance to humans is unclear given the small number of treated dogs, and the ~8x (mg/kg) human dose multiple.

Thiomersal

Thiomersal 10 µg/mL (5 µg per dose) is a constituent of the multidose vial as a preservative. Current intake limits for methylmercury are 3.3 µg/kg/week in an adult and 0.67 µg/kg/week in a pregnant woman (WHO), and 0.4 µg/kg/day (2.8 µg/kg/week) in a 70 kg adult (FDA).

The reduction, elimination or substitution of thiomersal (sodium ethyl mercury thiosalicylate) has been recommended for vaccines,¹³ however, the use of thiomersal as a preservative may be considered for a mutidose presentation.¹⁴

Vaccine residuals

The vaccine residuals are formaldehyde, ovalbumin, sucrose and sodium deoxycholate. No toxicological concerns are raised by the potential residual amounts in the vaccine.

Nonclinical summary and conclusions

Summary

The sponsor has applied to register a prepandemic monovalent, split virion, inactivated vaccine, Prepandemrix, for prophylaxis of influenza caused by the H5N1 strain with pandemic potential. The vaccine strain is A/Indonesia/05/2005/PR8-IBCDC-RG2 (H5N1 clade 2), prepared by reverse genetics, and is propagated in embryonated hen’s eggs. The vaccine contains AS03 oil-in-water adjuvant, the antigen and adjuvant is mixed prior to use. After mixing the vaccine should be used within 24 h. The multidose preparation (10 doses) also contains thiomersal 10 µg/mL (5 µg/dose). The treatment regimen is 2 consecutive 0.5 mL IM doses (3.75 μg HA) given at least 3 weeks and up to 12 months apart, in adults from the age of 18 years.

The sponsor also manufactures a registered seasonal, split trivalent influenza vaccine, Fluarix, which is registered in Australia. Prepandemrix monovalent bulk H5N1 antigen is manufactured by an identical process to Fluarix in Dresden.

A previous submission to register a prepandemic vaccine drived from the influenza H5N1 (clade 1) strain A/Vietnam/1194/2004/NIBRG14 was withdrawn following a negative recommendation by the ACPM, mainly due to: (i) inadequate information on duration of immunity and no information on booster doses; and (ii) a considered unfavourable risk/benefit ratio based on concerns about a potential safety signal in the elderly population. However, an H5N1 pandemic vaccine, Pandemrix H5N1, was registered in 2008.

However, the nonclinical data in the 2007 Prepandemrix H5N1 submission was considered sufficient to support registration of the influenza vaccine as a prepandemic and “mock-up” pandemic vaccine at an HA dose of 3.75 μg. Previously submitted nonclinical studies include immunogenicity studies in mice and pigs with AS03 adjuvanted seasonal influenza vaccines, lethal homologous and heterologous challenge studies with AS03 adjuvanted H5N1 vaccine in ferrets, a safety pharmacology study of AS03 adjuvanted seasonal influenza vaccine in rats, a repeat dose toxicity study with H5N1/AS03 vaccine in rabbits, genotoxicity studies with AS03 adjuvant, and an embryofoetal and postnatal development study with H5N1/AS03 vaccine in rats. Additional toxicity testing was not performed with vaccine derived from the A/Indonesia/05/ 2005 strain.

New nonclinical studies submitted with this application consisted of an integrated summary of the mode of action of AS03 adjuvant, a safety pharmacology study of AS03 adjuvant in dogs, and a rat reproductive toxicity study with a pandemic H1N1v/AS03 vaccine (Dresden) investigating pre-implantation loss.

AS03 adjuvant primarily acts as a stimulant of the innate immune response, with transient effects on multiple co-stimulatory molecules (CD40, CD80 and CD86), pro inflammatory cytokines (for example, IL-6), fibrinogen and neutrophils. Although AS03 does not act as an antigen depot, its action requires the antigen and adjuvant to be in proximity upon injection.

The new safety pharmacology study in conscious dogs with AS03 alone showed no cardiovascular or respiratory effects up to 72 h after IM injection of the human dose. A slight, but significant increase in body temperature was evident at 6 h post dose only.

The new GLP compliant rat reproductive toxicity study demonstrated that 6 daily IM injections on GD 0-6 of AS03 alone, or AS03 adjuvanted pandemic H1N1v vaccine (1/5th human dose) during the early stage of pregnancy had no effect on pre implantation loss.

Conclusions and recommendation

A previous application was made to register the split H5N1 prepandemic vaccine (Prepandemrix) derived from the A/Vietnam/1194/2004 (clade 1) strain. The current application seeks to register the prepandemic vaccine derived from the A/Indonesia/05/2005/PR8-IBCDC-RG2 (H5N1, clade 2) strain, at an HA dose of 3.75 µg, for 2 doses, in adults.

Nonclinical immunogenicity, AS03 adjuvant mechanisms of action, ferret lethal homologous and heterologous challenge, safety pharmacology, repeat dose toxicity, reproductive toxicity and genotoxicity studies were previously evaluated for Prepandemrix and Arepanrix vaccines, and were considered adequate for registration.

New nonclinical studies in the current submission comprised additional studies on AS03 adjuvant mechanisms of action, a safety pharmacology study with AS03 adjuvant in conscious dogs, and a rat reproductive toxicity study investigating pre implantation loss. There were no new adverse toxicological findings. The nonclinical data are considered adequate to support registration.