WO2010103022A2

WO2010103022A2 - Oil-in-water adjuvanted influenza vaccine

Info

Publication number: WO2010103022A2
Application number: PCT/EP2010/053004
Authority: WO
Inventors: Catherine Caillet; Fabienne Piras-Douce; Inca Carola Kusters
Original assignee: Sanofi Pasteur
Priority date: 2009-03-12
Filing date: 2010-03-10
Publication date: 2010-09-16
Also published as: WO2010103022A3

Abstract

The present invention is related to the use of a composition comprising at least one antigen derived from a H₅ influenza virus strain, for the preparation of an influenza immunization composition for administration according to a regimen comprising at least a 1st and a 2nd administration steps timely separated, wherein the H₅influenza virus strain of the 1st administration step is different from the H₅ influenza virus strain of the 2nd administration step and wherein the immunization composition used in at least one of the 1st or the 2 nd administration step comprises an oil-in-water emulsion as an adjuvant, wherein said oil-in-water emulsion comprises at least squalene, an aqueous solvent, a polyoxyethylene alkyl ether hydrophilic nonionic surfactant, a hydrophobic nonionic surfactant, and wherein said oil-in-water emulsion is obtainable by a phase inversion temperature process and wherein 90 % of the population by volume of the oil drops has a size less than 200 nm, and optionally less than 150 nm.

Description

Oil-in-water adjuvanted influenza vaccine

This invention relates to the field of vaccines, more particularly to the field of influenza vaccines comprising at least one adjuvant emulsion. More precisely, this invention is related to the use of a influenza vaccine composition which is intended to be used in a regimen where there are at least 2 administrations with 2 different influenza virus strains.

Some influenza vaccine has already been proposed in the prior art, as for example in patent application US-2008/0014217 Al. Influenza, commonly known as the flu, is an infectious disease, which is due to a virus which has the property of changing by mutation or by reassortment. Mutations can cause small changes in the haemaglutinin and neuraminidase antigens on the surface of the virus. This is called antigenic drift, which creates an increasing variety of strains over time until one of the variants eventually achieves higher fitness, becomes dominant, and rapidly sweeps through the human population often causing an epidemic. In contrast, when influenza viruses reassort, they may acquire new antigens for example by reassortment between avian strains and human strains; this is called antigenic shift. If a human influenza virus is produced with entirely novel antigens, everybody will be susceptible, and the novel influenza will spread uncontrollably, causing a pandemic. The current fear is that such a new strain could come from an avian strain such as a highly pathogenic strain of H5N1 virus, that would mute or reassort into a strain capable of efficient human-to-human transmission.

In the case of seasonal influenza, the strains belong to the types A or type B. Based on predictions of the future epidemic strain, the influenza vaccine for a given season usually contains antigens from three different strains, two A strains and one B strain. This composition changes every year. Based on the fact that almost all the humans have already met at least one of this strain, when a vaccine shot is administered, the immune response is of the "booster" type, leading to an important immunological response. In the case of a strain of a type A sub-type H₅N₁, the situation is very different as most of the world population is immunologically naive vis-a-vis such a virus.

In the case of pandemic, it would be desirable to protect as many people and as fast as possible, but time and industrial yield would probably be limiting factors. It is thus desirable to get formulations and immunization regimen that would permit to immunize people very efficiently.

To achieve this aim, the present invention proposes the use of a composition comprising at least one antigen derived from a H₅ influenza virus strain, for the preparation of an influenza immunization composition for administration according to a regimen comprising at least a 1^st and a 2^nd administration steps timely separated, wherein the H₅ influenza virus strain of the 1^st administration step is different from the H₅ influenza virus strain of the 2^nd administration step and wherein the immunization composition used in at least one of the 1^st or the 2^nd administration step comprises an oil-in-water emulsion as an adjuvant, wherein said oil-in-water emulsion comprises at least squalene, an aqueous solvent, a polyoxyethylene alkyl ether hydrophilic nonionic surfactant, a hydrophobic nonionic surfactant, and wherein said oil-in-water emulsion is obtainable by a phase inversion temperature process and wherein 90% of the population by volume of the oil drops has a size less than 200 nm, and optionally less than 150 nm. Thanks to such a composition it has been possible to get a very good cross immunization, which means that the administered composition is useful not only against the administered viruses, even the ones which have been administered only once, but also against other viruses that could appear later and are still unknown. According to a specific embodiment, the oil-in-water emulsion is present in the composition used in the 1^st administration step.

Thus, for the preparation of a pandemic, it is possible to immunize the population with a pharmaceutical composition which is more complex to prepare than the composition which will be used later. According to a particular embodiment of the invention, the H₅ virus strain used in the 1^st administration step and the H₅ virus strain used in the 2^nd administration step belong to different clades or sub-clades.

According to a particular embodiment of the invention, the administration regimen comprises as the 1^st administration step, 2 shots separated by a period of time which is at least 7 days. Such a regimen provides an optimized immune response after the 2^nd administration. Another subject of the invention is a method of immunizing a human against influenza, comprising the step of administering to the human at least: - a 1^st immunization composition comprising at least one antigen derived from a H₅ influenza virus strain,

- and later, a 2^nd immunization composition comprising at least one antigen derived from a H₅ influenza virus strain, wherein the H₅ virus strain virus used in the 1^st immunization composition is different from the H₅ virus strain virus used in the 2^nd immunization composition, wherein at least one of the 1^st or the 2^nd immunization compositions comprises an oil-in- water emulsion comprising squalene, an aqueous solvent, a polyoxyethylene alkyl ether hydrophilic nonionic surfactant, a hydrophobic nonionic surfactant, wherein said emulsion has been obtained by a phase inversion temperature process and wherein 90% of the population by volume of the oil drops has a size less than 200 nm, and optionally less than 150 nm.

Thus, it is possible to induce in the human a cross immunization which is much more important than the one which is obtained without adjuvant.

This result is very important for the preparation of a pandemic against which it will be necessary to be able to react very quickly on a large scale.

Other advantages and embodiments of the invention will emerge upon reading the detailed description hereinafter, with reference to the figures which illustrate the results of the tests described in the examples.

Figure 1 represents HI antibody response against clade 1 RG- 14 measured at D 140 in sera of mice primed once or twice with 0.3μg of AF03-adjuvanted clade 1 RG-14 vaccine and boosted four months later with 0.3μg of either a clade 1 RG-14 or a clade 2 RG-2 monovalent vaccine, in presence or not of AF03. These results are those obtained by the test performed according to Example 4.

Figure 2 represents, as also described in Example 4, HI antibody response against clade 2 RG-2 measured at D140 in sera of mice primed once or twice with 0.3μg of AF03- adjuvanted clade 1 RG-14 vaccine and boosted four month later with 0.3μg of either a clade 1 RG-14 or a clade 2 RG-2 monovalent vaccine, in presence or not of AF03. For the purpose of the present invention, the expression "influenza virus antigen" is intended to mean all the possible antigens, whether they consist of the whole virus, "natural" fractions of virus or elements obtained by methods involving genetic recombinations. These viral antigens originate, in general, from viral cultures which are performed either in eggs or in cells. It may involve influenza which infects humans, but also influenza which infects more particularly animals, especially birds. For a description of influenza antigens that can be used, reference may be made to pages 2-9 of patent application WO2007/052058. On the basis of their nucleocapsid and M protein antigens, the influenza viruses are divided into 3 distinct immunological types (A, B, and C). Influenza A viruses occur in mammals, pigs, birds, and horses. However, only human is infected by influenza B and C.

Influenza viruses have two kinds of surface proteins: haemagglutinin and neuraminidase. There are different types of haemagglutinin (H) and neuraminidase (N) surface proteins. There are 16 types of H surface proteins and 9 types of N surface proteins (which are named Hl, H2, Nl, N2, etc.). An influenza virus always has one type of H surface protein, and one type of N surface protein.

Influenza virus subtypes are classified and named according to the types of H and N surface proteins on the virus. For example, a subtype named H5N1 would indicate that the virus has type 5 hemagglutinin surface proteins, and type 1 neuraminidase surface proteins. Subtype H5 and particularly H5N1 is one of the most notable virus that cause avian influenza.

The HA sequences of the majority of H5N1 viruses circulating in avian species are separated into a number of distinct clades. Clade 1 viruses have caused human infections in Cambodia, China Hong Kong Special Administrative Region, Thailand and Viet Nam and have been recently detected in poultry in Cambodia, Thailand and Viet Nam. Clade 2.1 viruses have continued to circulate in poultry and caused human infections in Indonesia. Clade 2.2 viruses have the most geographically diverse distribution and have caused outbreaks in birds in over 60 countries in Africa, Asia and Europe with human infections in Azerbaijan, Bangladesh, China, Djibouti, Egypt, Iraq, Nigeria, Pakistan and Turkey. Some recent clade 2.2 viruses have diverged genetically from reference strains. Clade 2.3 viruses are genetically diverse. Clade 2.3.2 and 2.3.4 viruses continue to circulate in birds in Asia ; clade 2.3.4 viruses have been responsible for human infections in China, Lao People's Democratic Republic, Myanmar and Viet Nam.

Viruses from other clades, including clade 7, have been sporadically detected in birds in

Asia.

For the present invention, each of the immunization compositions comprise at least one antigen deriving from a strain of the H5 subtype, the antigen used in the 1^st administered composition deriving from a strain different from the one used in the 2^nd administered composition.

Both viruses can derive from a different clade, or from a different sub-clade.

As examples of the different strains which are currently approved by the Regulatory

Authorities, we can cite:

In particular, in the present invention, we have used immunization candidates that were monovalent, inactivated, split-virion vaccine using the influenza ANietnam/ 1194/2004 NIBRG14 (H5N1) reassortant reference strain derived by reverse genetics from the highly pathogenic avian strain ANietnam/ 1194/2004 by the UK National Institute for Biological Standards and Control. This is one of the reference viruses considered suitable for prototype pandemic influenza vaccines. This virus belongs to clade 1. We have also used A/ Indonesia/5/2005/CDCRG-2 (H₅Ni) which belongs to clade 2.1 and A/bar-headed goose/Qinghai Lake/ 1A/05/S JRG- 163222 (H₅Ni)which belongs to clade 2.2. These viruses were propagated in embryonated hens' eggs, using the licensed manufacturing process for the interpandemic vaccines Vaxigrip®, or Fluzone®. According to the invention, at least one of the administered immunization composition comprises an oil-in-water adjuvant emulsion which comprises at least: i) squalene, ii) an aqueous solvent, iii) a polyoxyethylene alkyl ether hydrophilic nonionic surfactant, iv) a hydrophobic nonionic surfactant, said emulsion being obtained by a phase inversion temperature process and wherein 90% of the population by volume of the oil drops has a size less than 200 nm, and preferably less than 150 nm. Such an emulsion has been described in US-2007-0014805-A1 patent application. The squalene is an oil initially originating from shark liver; it is an oil whose empirical chemical formula is C30H50, comprising 6 double bonds; this oil is metabolizable and has qualities that allow it to be used in an injectable pharmaceutical product. Squalene of plant origin, extracted from olive oil, also exists. Good results have in particular been obtained using the squalene provided by the company Fluka, which is of animal origin. The amounts of squalene used for the preparation of a concentrated emulsion are advantageously between 5 and 45%; this concentrated emulsion is subsequently diluted during the preparation of the immunogenic compositions so as to prepare immunizing doses in which the amount of squalene is between 0.5 and 5%, and particularly 1 or 2.5%. This dilution can be carried out by simple mixing of the adjuvant emulsion according to the invention and the suspension comprising the antigen.

According to the invention, the emulsion comprises a nonionic hydrophilic surfactant, the hydrophilic/lipophilic balance, or HLB, value of which is greater than or equal to 10, and which belongs to the chemical group of polyoxyethylene alkyl ethers (PAEs), also called polyoxyethylenated fatty alcohol ethers, or n-alcohol polyoxyethylene glycol ethers, or macrogol ethers. These nonionic surfactants are obtained by chemical condensation of a fatty alcohol and ethylene oxide. They have a general chemical formula in which n denotes the number of ethylene

oxide units (typically 10-60), and (x+1) is the number of carbon atoms in the alkyl chain, typically 12 (lauryl(dodecyl)), 14 (myristyl(tetradecyl)), 16 (cetyl(hexadecyl)), or 18 (stearyl(octadecyl)), so x is in the range of from 11 to 17. Polyoxyethylene alkyl ethers tend to be mixtures of polymers of slightly varying molecular weights. Accordingly, the emulsions of the invention will comprise a mixture of polyoxyethylene ethers. Furthermore, because polyoxyethylene alkyl ethers are mixtures, when a particular polyoxyethylene ether is recited herein for use in an emulsion, it will be understood that it is the primary but not necessarily the only polyoxyethylene alkyl ether present in the emulsion.

The emulsion according to the invention usually comprises a single hydrophilic PAE. A mixture of several PAEs is also suitable insofar as the overall HLB value is >10. The polyoxyethylenated fatty alcohol ethers that are suitable for the subject of the invention can be in a form which is liquid or solid at ambient temperature. Among solid compounds, preference is given to those which dissolve directly in the aqueous phase or which do not require substantial heating. Insofar as the number of ethylene oxide units is sufficient, polyoxyethylenated ethers of lauryl alcohol, myristyl alcohol, cetyl alcohol, oleyl alcohol and/or stearyl alcohol are particularly suitable for the subject of the invention. Some of them can be found among products known under the trade names Brij® for the products sold by the company ICI America's Inc., Eumulgin® for the products sold by the company Cognis, or Simulsol® for the products sold by the company Seppic.

An emulsion which is particularly preferred according to the invention contains, as hydrophilic nonionic surfactant, a polyoxyethylene alkyl ether chosen from the group consisting of ceteareth-12 (sold under the name Eumulgin® Bl), ceteareth-20 (Eumulgin®. B2), steareth-21 (Eumulgin® S21), ceteth-20 (Simulsol® 58 or Brij® (58), ceteth-10 (Brij® 56), steareth-10 (Brij® 76), steareth-20 (Brij®78), oleth-10 (Brij® 96 or Brij®97) and oleth-20 (Brij®98 or Brij®99). The number attributed to each chemical name corresponds to the number of ethylene oxide units in the chemical formula.

Good results have been obtained with the product BRIJ® 56. A compound that is particularly suitable and preferred because of its semi- synthetic origin is polyoxyethylene (12) cetostearyl ether, provided by the company Cognis under the name Eumulgin®Bl. This product is a mixture consisting essentially OfCH₃ (CH₂)i5-(O-CH₂- CH₂)i2 -OH and CH₃ (CH₂)I₇-(O-CH₂-CH₂)I₂ -OH, but with also some CH₃ (CH₂)i₆- (O-CH₂-CH₂)i2 -OH and some CH₃ (CH₂)₁₉-(O-CH₂-CH₂)I₂ -OH; considering also that the number 12 of ethylene oxide units is not the exact number, but it is a mean between 5 and 23.

According to the invention, the adjuvant emulsion also comprises a hydrophobic nonionic surfactant, which must be pharmaceutically acceptable; among surfactants that are suitable in this regard, mention may be made of sorbitan ester or mannide ester surfactants; the sorbitan ester surfactants are obtained by reaction of a fatty acid and of a mixture of partial esters of sorbitol and its mono- and dianhydrides; this may involve a mono-, a di- or a triester, or even a mixture; they are hydrophobic surfactants for which the overall hydrophilic- lipophilic balance (HLB) is less than 9, and preferably less than 6. Some of them can be found amoung the surfactants sold by the company ICI Americas Inc. under the name Span®, or by the company Cognis under the name Dehymuls™., or by the company ICI under the name Arlacel™; as examples of surfactants that are particularly suitable, mention may be made of the sorbitan oleate sold under the name Dehymuls SMO™. or Span®80 or Montane™80. Among the mannide ester surfactants, mention may be made of the mannide monooleate sold by the company Sigma, or by the company Seppic under the name Montanide 80™.It has to be noticed that these products are never 100% pure, and depending upon their origin, in addition to oleates (mono, di and tri-oleates), they can also contain some other esters such as palmitates or linoleates.

According to the present invention, this emulsion is prepared through a PIT (Phase Inversion Temperature) process which leads to a monodisperse emulsion, the droplet size of which is very small, which makes the emulsion very stable.

Example 1 Preparation of adjuvant emulsion

The concentrated bulk (32.5% (w/w) of squalene) of adjuvant was first manufactured, and then, diluted with PBS and sterile filtered to produce the emulsion which contains 5% (w/w) of squalene. The PIT method is used to produce the adjuvant as a very fine monodisperse emulsion with a narrow size distribution. Mean particle diameter of the oil drops is 90-100 nm, as determined by laser light scattering. The process consists in three steps:

1. Preparation of the aqueous phase;

2. Preparation of the oily phase; 3. Emulsification.

The quantities of components which are used are as follows: - squalene: 488 g - sorbitan oleate: 72.3 g macro go 1 cetostearyl ether: 92.7 g mannitol: 90 g

- PBS: 758 g For the preparation of the aqueous phase, mannitol is dissolved in a PBS solution at +40°C ± 5°C. The solution is then mixed for at least 5 minutes to ensure the total dissolution of the mannitol. The solution is kept at +40°C ± 5°C and macrogol cetostearyl ether is added. The solution is then stirred for at least 5 minutes to ensure the total dissolution of the macrogol. The aqueous phase is kept at +40°C ± 5°C until the emulsification process.

For the preparation of the oily phase, squalene and sorbitan oleate are transferred into a brown-coloured flask and mixed for at least 5 minutes to obtain total dissolution. For the emulsification step, the oil phase is transferred into a jacketed steel reactor equipped with a stirring helix. After addition of the aqueous phase, the reactor is closed. The oil/water mixture is first stirred at 1000 rpm for approximately 5 minutes to form a coarse oil- in- water emulsion and then at 200 rpm for the rest of the process. The reactor is heated (by circulation of warm water) in order to increase the temperature of the emulsion to +63°C + 1°C. During the entire process, the temperature and the conductivity of the emulsion are monitored. Once the conductivity drops to zero (zero conductivity is characteristic of a water-in-oil emulsion), heating is stopped. The reactor is then cooled to +22°C ± 3°C, which produces the final fine oil-in-water emulsion (adjuvant AF03 concentrated bulk). This concentrated bulk is then transferred to a 2 L amber glass flask, which can be stored overnight at +5°C + 3°C before dilution.

The proportions of the different components in the concentrated bulk are as follows: - Squalene: 32.5%

- Sorbitan oleate: 4.8% Macrogol cetostearyl ether: 6.2%

- Mannitol: 6.1%

- PBS: q.s. 100% This bulk product is then diluted 6.5-fold with a PBS solution to reach a target concentration of 5% w/w of squalene. This 5% squalene emulsion will then be diluted, as needed, by the suspension containing antigens to get the final concentration which is indicated in each of the following examples.

Example 2 : Immune Response Induced in BALB/c Mice after Priming with H₅N₁ Monovalent Influenza Vaccines - Prepared with H ₅N₁Viral Strains from Different Clades Adjuvanted or not with AF03- Administered by the IM Route and Booster with a Clade 2.2 H₅N₁

The aim of this study was to evaluate in BALB/c mice the immunogenicity of three H₅N₁- monovalent vaccines (prepared according to the process used for the licensed product Fluzone® process) comprising either clade-1 ANietnam/1203/2004, clade-2.1 A/Indonesia/5/05/CDCRG-2 or clade-2.2 A/bar-headed goose/Qinghai Lake/1 A/05/SJRG- 163222 viral strains, and to evaluate their ability to prime an immune response induced by a subsequent immunization with a clade-2.2-based H5N1 monovalent vaccine.

For that purpose, 3 groups of 10 mice received a primary immunization series (1^st administration) with 2 injections, given 3 weeks apart via the intramuscular route, with 0.3 μg of each H5N1 monovalent vaccine administered in presence of AF03 (2.5% of squalene). Three additional groups of 10 mice received 0.3 μg of either vaccine administered without adjuvant. Four months later (D 147), all animals received a booster immunization (2^nd administration) with 0.3 μg of the AF03 -adjuvanted clade-2.2 monovalent immunization composition. A control group of 5 mice received PBS at primary and booster immunization. Blood samples were collected at D161 for immune response analysis by hemagglutination inhibition (HI) assay against ANietnam/1194/04/NIBRG-14, A/Indonesia/5/05/CDCRG-2 and A/bar-headed goose/Qinghai Lake/1A/O5/SJRG-163222.

To appreciate the immune response, a test called Haemagglutination Inhibition is performed.

This technique is used to titrate the functional anti-HA antibodies present in the sera of influenza immunized animals, on the basis of the ability of the virus to agglutinate red blood cells: a serum containing specific functional antibodies directed against HA inhibits the Hemagglutination activity.

In order to eliminate serum non-specific inhibitors directed against the HA, each serum was treated with a receptor-destroying enzyme (RDE) and then horse red blood cells. Serial dilutions (2 fold) of the virus A/ H₅Ni Nietnam/ 1194/04/NIBRG-14, A/Indonesia/5/05/CDCRG-2 (clarified allantoic fluid, 40 HAU/50μl and 256 HAU/50μl, respectively) and A/ bar-headed goose/Qinghai Lake/1A/O5/SJRG-163222 (inactivated whole virus vaccine, 1280 HAU/50μl) were performed in PBS in order to calibrate the viral suspension and to obtain 2 HAU (Hemagglutination Unit) in presence of horse red blood cells (1% in PBS). Calibrated virus (50 μl) was then added to the V shaped well of a 96 well plate on 50 μl of serum serial dilutions (2 fold) in PBS, and incubated one hour at room temperature. Horse red blood cells (1% in PBS) (50 μl) were then added to each well and inhibition of hemagglutination (red point) or hemagglutination (pink network) was visually read after one hour and half at 4°C. The titer in HAI antibody is the reciprocal of the last dilution giving no hemagglutination. A value of 5 corresponding to half of the initial dilution (1/10) was arbitrary given to all sera determined negative in order to perform statistical analysis. A control without virus (only serum and horse red blood cells), a control of the horse red blood cells (only horse red blood cells and PBS) and a control of the presence of 2 HAU of the viral dilution used, were performed for each series of HI test. The HI determinations were performed once. Geometric means are calculated for each group of animals.

HI immune responses against each priming vaccine strain were measured after booster at D161 in individual sera from all animals and GMTs are summarized in Table 1 hereinafter:

After the booster immunization, we can see that there is a cross reactivity of the induced antibodies, and a trend towards a higher response in the cases where the adjuvant was used in the priming administration.

It is noteworthy that even the mice which have received only the clade 2.2, both during the priming and the booster, are nevertheless capable of responding to a clade 1 strain, in particular when the priming composition comprises the adjuvant.

Example 3: Immunogenicity of a clade 2 HjS[^ monovalent vaccine in Balb/c mice primed or not with a clade 1 HjS[^ monovalent vaccine.

The aim of this study was to evaluate the ability of a clade 1 H5N₁ monovalent vaccine to prime the immune response induced in Balb/c mice by an H5N₁ monovalent vaccine prepared with a different H5N₁ strain/clade. For that purpose, 3 groups of 20 mice received at prime either PBS (control group) or a single dose (0.3 μg of HA) of unadjuvanted or adjuvanted clade 1 vaccine administered in presence of squalene emulsion AF03. Seven months later, animals received an H5N₁ booster of either clade 1 (10 mice/group) or clade 2.1 vaccine (10 mice/group) administered in presence of AF03. The composition of the AF03 squalene emulsion is, per dose, as follows: - squalene: 2.50 mg Eumulgin Bl : 0.48 mg

- Dehymuls SMO: 0.37 mg mannitol: 0.46 mg

Blood samples were collected every month after immunization for immune response analysis, and in particular at D273 (6 weeks after the booster) to appreciate the immune response after the boost.

The test performed is 1HA as in the previous example.

The results in GMT for each group of mice, are summarized in the following Table 2

Six weeks after the boost (D273), HI titers remained low in the PBS primed group (< 60) and were enhanced in all groups previously primed with the clade 1 vaccine with similar HI titers in all groups irrespective of the presence or not of an adjuvant at prime. One can also notice that mean HI titers against the Vietnam strain were higher in all groups that received the heterologous boost compared to those that received the homologous boost: mean HI titers were approximately 8000 and 2500 after heterologous and homologous boost, respectively.

This test shows that there is no need to have adjuvant both during the priming and the booster administration; the presence of the adjuvant is important in at least one administration. Example 4: Impact of a one vs. two doses priming schedule on the immune response induced after a booster with H ₅N₁.

The aim of this study was to evaluate in Balb/c mice the impact of the priming regimen (1 vs. 2 immunizations) with a clade 1 H₅Ni vaccine (ANietnam/I 194/2004/ NIBRG- 14) administered in presence of AF03 on the immune response induced by a booster immunization given 4 months later with either the same vaccine strain or a clade 2 H₅N ₁ monovalent vaccine (A/Indonesia/2005/CDCRG-2) administered in presence or not of AF03. For that purpose, 2 groups of 40 mice were immunized once or twice, at 3 week- intervals with 0.3 μg of AF03-adjuvanted clade 1- H5N1 monovalent vaccine (ANietnam/ 1194/2004/ NIBRG-14, clade 1). Four months later, each group was divided into four groups to receive 0.3 μg of either the same strain or a clade 2 H5N1 monovalent vaccine, each administered in presence or not of AF03. The composition of the AF03 squalene emulsion is, per dose, as follows: squalene: 2.50 mg

- Eumulgin Bl : 0.48 mg

- Dehymuls SMO: 0.37 mg mannitol: 0.46 mg Blood samples were collected at D42 and at D 140 for immune response analysis determined by inhibition (HI) assays, which are performed as described in Example 2. The HI responses elicited against clade 1 RG- 14 and clade 2 RG-2 strains were measured at D42 in pooled sera and at D 140 in individual sera collected from all animals. Geometric means for HI antibody titers as well as individual values against clade 1 -RG- 14 and clade 2 RG-2 strains are presented in Figures 1 and 2, respectively. Figure 1 represents HI antibody response against clade 1 RG- 14 while Figure 2 represents HI antibody response against clade 2 RG-2. The results are also summarized in Table 3 which follows.

From the results illustrated on Figure 1, we can see that at D 140, 3 weeks after the clade 1 booster immunization, a strong booster effect of the HI response against ANietnam/ 1194/2004/ NIBRG- 14 was measured in groups primed with 1 shot with a 30-fold increase of the HI titers (from 211 HI titer at D42 as indicated by the dotted line to 6400 at D 140) whereas this increase was minimal (from 3152 to 7000) in groups that received 2 shots at prime. The same booster effect was also measured when the clade 2 RG-2 vaccine was used as a boost. Interestingly, similar HI immune responses were measured in animals previously primed once and twice, except for the mice group boosted with AF03-adjuvanted clade 2 vaccine, where responses were higher after one priming immunization. Moreover, for each booster strain, similar HI titers were measured irrespective of the presence or not of the adjuvant at boost. We can also notice that mean HI titers against clade 1 RG- 14 were higher in all groups that received the heterologous boost compared to those that received the homologous boost: mean HI titers were approximately 15000 and 6000 after heterologous and homologous boost, respectively.

From the results illustrated on Figure 2, we can see that after the clade 1 homologous booster immunization, the HI titers against A/Indonesia/5/2005/CDCRG-2 were strongly increased to achieve similar mean titers close to 1500 in all animals previously primed once and twice, irrespective of the presence or not of the adjuvant at boost. In groups that received the clade 2 RG-2 heterologous booster, HI antibodies against the clade 2 RG-2 strain were also strongly increased, to achieve at D 140 similar HI titers in all groups, irrespective of the number of previous clade 1 immunizations and the presence of the adjuvant at boost. So, these results show globally that in animals primed with an AF03-adjuvanted immunization composition, the H5N1 specific immune response induced by a boost given 4 months later is similar whatever the priming regimen- 1 vs. 2 doses- and irrespective of the presence of the adjuvant at boost. This shows that it is not necessary to perform a boost with adjuvant, in so far as the adjuvant was already present for the priming administration.

Example 5: Clinical Study related to Safety and Immunogenicity of Intramuscular H₅N₁ adjuvanted, inactivated, split-virion pandemic Influenza vaccine in healthy adults subjects. In this multicenter, randomized, blind-observer phase 1 trial, the safety and immunogenicity of 4 influenza ANietnam/I 194/ 2004 IBRG-14 (H₅Ni) vaccine candidates containing 1.9, 3.8, 7.5 or 15 μg of hemagglutinin and an oil- in- water adjuvant emulsion were investigated in comparison with 7.5μg of antigen without adjuvant. The amount of adjuvant in each injected vaccine was identical. The participants were healthy 18-40 year-old volunteers .

Investigational vaccine candidates were monovalent, inactivated, split-virion vaccine produced by Sanofi Pasteur using the influenza A/Vietnam/ 1194/2004 NIBRG14 (H₅Ni, clade 1) reassortant reference strain for the 1^st administration (UK National Institute for Biological Standards and Control), and A/Indonesia/05/2005-RG2 (H₅Ni, clade 2.1) for the booster administration. These viruses were propagated in embryo nated hens' eggs, using the licensed manufacturing process for the interpandemic vaccine Vaxigrip. Before being mixed with the vaccine, the adjuvant, produced as described in Example 1 , was a 5% squalene-in- water emulsion which was prepared to be a very fine (mean particle diameter 100 nm) and monodispersed emulsion with a narrow particle size distribution.

Vaccine doses were prepared just before injection by mixing vaccine from multidose vials and adjuvant from monodose vials according to a reconstitution protocol. This protocol was devised to ensure that the amount of antigen in the final vaccine doses matched the targeted amount and that each adjuvanted vaccine dose contained an identical amount of adjuvant.

The final volume per dose was 0.3 mL (nonadjuvanted vaccine), 0.4 mL (adjuvanted 1.9-μg vaccine), or 0.6 mL (the other 3 formulations). The final antigen dose of the injected vaccines was calculated on the basis of the known antigen dose of the multidose vials.

In a preliminary step, 15 subjects were immunized twice, 21 days apart, with the adjuvanted 15-μg vaccine (i.e., the highest dosage vaccine to be injected into subsequent subjects) and were closely monitored for clinical and biological safety at study visits on days 0, 2, 8, 21, 23, and 29. Blood samples obtained at each visit were used to identify abnormalities in liver function parameters, electrolytes and proteins, lipid metabolism parameters, and hematology parameters. Safety data from this preliminary step were reviewed before continuing the trial. A total of 251 subjects were then enrolled and randomly assigned to 1 of 5 groups according to the pandemic influenza vaccine formulation used (1.9μg, 3.75μg, 7.5μg, 15μg of haemagglutinin antigen with AF03 adjuvant and 7.5μg of haemagglutinin antigen without adjuvant). Each subject received 2 injections 21 days apart, plus a booster injection 12 months after the first injection. This booster injection consisted of 3.75μg of HA with AF03 for the subjects immunized with an adjuvanted formulation during the primary series, and of 7.5μg HA without adjuvant for the subjects immunized with 7.5μg of HA without adjuvant during the primary series. Three different formulations corresponding to the four doses of influenza antigen (1.9 μg, 3.75 μg, 7.5 μg, and 15 μg of HA per vaccine dose) were tested. To prepare the administered doses, one prepared vials of adjuvant produced according to Example 1, each vial of 0.7 ml comprising: - squalene: 35 mg sorbitan oleate: 5.2 mg macro go 1 cetostearyl ether: 6.7 mg mannitol: 6.5 mg - PBS solution: qs 0.7 mL

The preparation of influenza antigens is done from 3 different formulations: 1) Formulation 1 (15 μg/mL of HA) :

The three formulations of the influenza vaccine to be administered on DO and D21 were mixed with the AF03 adjuvant as presented in the Table 4 below:

The formulation used for the booster administration had the same composition as the formulation used for the primo-immunization of Group 2 with the exception that the viral strain was A/Indonesia/05/2005CDCRG-2 (H₅Ni) which is a clade 2.1 while the clade used for the primo-immunization was a clade 1.

For the group receiving no adjuvant, the administered doses were prepared from the formulation 2 for the primo-immunization, and from a formulation resembling the Formulation 1 for the booster administration with the exception that the clade used was A/Indonesia/05/2005CDCRG-2 (H₅Ni) which is a clade 2.1.

Serum samples for antibody analyses were obtained before and 21 days after each immunization.

Immunogenicity assays. Samples were tested for haemagglutination-inhibition (HI) and neutralizing activity against the clade 1 reassortant vaccine seed virus; A/Vietnam/ 1194/2004 was used in HI assays, and A/Vietnam/ 1203/2004 was used in neutralization assays. To test for cross-reactivity on day 42, both assays were also performed with the clade 2 influenza A/Indonesia/05/2005 (H5N1) RG2 strain. Assays were performed under blinded conditions. The used HI assay has been described previously (cf. Stephenson I, Wood JM, Nicholson KG, Charlett A, Zambon MC. Detection ofanti-H5 responses in human sera by HI using horse erythrocytes following MF59-adjuvanted influenza A/Duck/Singapore/97 vaccine. Virus Res 2004; 103:91— 5;and Kendal AP, Skehel JJ, Pereira MS, eds. Concepts and procedures for laboratory- based influenza surveillance. Atlanta: Centers for Disease Control, 1982:B17-35.) Briefly, after eliminating nonspecific inhibitors and anti-species agglutinins, samples were centrifuged and supernatants were submitted to the HI method. From an initial serum dilution of 1 :8 (A/Vietnam) or 1 :10 (A/Indonesia), ten 2-fold dilutions of serum were prepared and combined with an equal volume of antigen suspension at 4 hemagglutinin units (HAU)/25 μL for 1 h at room temperature. The 4 HAU of antigen used in the assay was determined for each lot of erythrocytes by titrating the antigen (2- fold dilutions) and mixing with equine erythrocytes. After 1 h, haemagglutination was assessed for each dilution, and virus concentration containing 4 HAU was determined. After incubating serum and antibody together for 1 h, 50 μL of 1.0% equine erythrocyte suspension was added, and the reaction was left for another hour before reading. The serum titer is equal to the highest reciprocal dilution, which induced a complete inhibition of hemagglutination. The titer of each quality control serum is close to the previously assigned value (within one serial two-fold dilution limits). The RBC control (RBC suspension without antigen) and the serum control (for each sample tested) do not produce any agglutination.

Each serum sample was titrated in duplicate which should not differ by more than a 2- fold serial dilution. The final titer is equal to the geometric mean of the two results.

Neutralizing antibody activity was analyzed using a microneutralization assay based on the methods of the pandemic influenza reference laboratories of the Centers for Disease Control and Prevention and the Health Protection Agency Heat-inactivated human serum samples were preincubated with a standardized amount of influenza virus before the addition of Madin-Darby canine kidney cells. After overnight incubation, viral nucleoprotein was detected by ELISA in virally infected cells. Serum antibodies to the influenza virus haemagglutinin inhibit the viral infection of cells; therefore, the optical density results of the ELISA are inversely proportional to the serum antibody concentration. Serum was 2-fold serially diluted from a starting dilution of 1 : 10, and reciprocal dilutions of serum achieving 50% or greater neutralization of virus growth were considered positive. Each sample was tested in duplicate, and the final titer was the mean of the duplicate titers. Statistical analysis. The sample size was chosen in line with European guidelines for annual influenza vaccine trials. Results were summarized using point estimates and 2-sided 95% confidence intervals (95% CI).

The immunogenicity results of HI tests performed in all groups against the A/Indonesia/05/2005 (H5N1, clade 2.1) strain are summarized in Table 5 which follows:

Before booster immunization, the anti-HA Ab titers were not detectable in most of the subjects. Pre-immunization titers varied between 5.04 (1/dil) and 5.57 (1/dil). A total of 7 subjects (5 in Group 3 and 2 in Group 4) had detectable Ab titers at D365 . As soon as 7 days (D372) after booster immunization with A/Indonesia, a higher immune response was observed in the four adjuvanted groups with a GMTR [D372/D365] of 8.92 compared to 1.94 for the control group. This response was comparable across the adjuvanted groups with 93.8% of subjects with detectable Ab titers compared to 41.7% of subjects in the control group. However, Group 1 (1.9 μg) had a lower proportion of subjects with Ab titers >40 (1/dil) and a lower proportion of seroconverted subjects compared to the three other adjuvanted groups. In terms of distribution titers, 68.8% of subjects in the adjuvanted groups showed significant increase rate compared to 10.8% of subjects in the control group. The 7 subjects with detectable Ab titers before booster immunization had all a significant increase rate as soon as 7 days after booster immunization. 21 days after booster immunization, this immune response was amplified for the four adjuvanted groups, with a GMTR [D386/D365] of 14.6 compared to 2.24 for the control group. This response remained comparable across the adjuvanted groups with 95.8 to 100.0% of subjects with detectable Ab titers compared to 48.9% of subjects in the control group. Once again, Group 1 (1.9 μg) had a lower proportion of subject with Ab titers >40 (1/dil) and of seroconverted subjects. For the control group, the proportion of subjects with detectable Ab titers only slightly increased compared to D372 levels. Regarding EMEA criteria, all three criteria were met 21 days after booster immunization for all the adjuvanted groups, as soon as 7 days after booster immunization for Group 2 (3.75 μg) and Group 3 (7.5 μg). None of these criteria were met for the control group.

The results of the seroneutralization test against A/Indonesia/5/2005/RG-2 (H₅N₁₎ strain are presented in Table 6 which follows:

Before the booster immunization, the Ab titers were detectable for approximately 23.4% of subjects in the adjuvanted groups and 4.2% of subjects in the control group. Pre- immunization titers varied between 5.19 (1/dil) and 7.71 (1/dil). As soon as 7 days after booster immunization (D372) with A/Indonesia, an immune response was observed in the four adjuvanted groups with a GMTR [D372/D365] of 32.7 compared to 5.19 for the control group. This response was comparable across the adjuvanted groups with 99.0% of subjects with detectable Ab titers compared to 72.3% of subjects in the control group. In terms of distribution titers, the proportion of subjects with Ab titers >40 (1/dil) was comparable across the adjuvanted groups, 91.5 to 98.0% of subjects. The 2-fold and 4-fold increase was respectively 99.0% and 95.3% for all the adjuvanted groups compared to 70.2% and 53.2% for the control group. 21 days after booster immunization (D386), this immune response was amplified for the four adjuvanted groups, with a GMTR [D386/D365] of 54.3 compared to 6.24 for the control group. This response remained comparable across the adjuvanted groups with 97.9 to 100.0% of subjects with detectable Ab titers compared to 72.3% of subjects in the control group. Once again, the 2-fold and 4-fold increase was higher in all the adjuvanted groups.

As observed with HI method, the four adjuvanted groups had a higher immune response than the control group. Whatever the formulation received for primary immunization, all four adjuvanted groups had a similar immune response against A/Indonesia/05/2005 strain.

The immunogenicity results of HI tests performed in all groups against the ANietnam/ 1194/2004 (H5N1, clade 1) strain are summarized in Table 7 which follows:

Before booster immunization, the proportion of subjects with detectable Ab titers ranged from 31.9 to 63.3% in the adjuvanted groups compared to 25.0% in the control group. As soon as 7 days after booster immunization (D372) with A/Indonesia, a higher immune response was observed in the four adjuvanted with GMTR [D372/D365] of 6.69 compared to 2.14 for the control group. This response was comparable across the adjuvanted groups with 97.9% of subjects with detectable Antibody (Ab) titers compared to 66.7% of subjects in the control group. In terms of distribution titers, 72.4% of subjects in the adjuvanted groups showed significant increase rate compared to 12.8% of subjects in the control group. 21 days after booster immunization with A/Indonesia/05/2005, GMTs increased to 79.9 for all adjuvanted groups and only slightly increased to 14.6 for the control group. The proportion of subjects with detectable Ab titers increased to 99.0% for all adjuvanted, with 81.3% of subjects showing significant increase rate. For the control group, the proportion of subjects with detectable Ab titers remained comparable to D372. Regarding EMEA criteria, all three criteria were met as soon as 7 days after booster immunization for all the adjuvanted groups, whereas none of these criteria were met for the control group, not even 21 days after booster immunization After booster immunization with A/Indonesia, the immune response against the ANietnam strain was similar to the response observed against the A/Indonesia strain. The four adjuvanted groups had a much higher immune response than the control group.

The results of the seroneutralization test against ANietnam/I 203/2004 strain are presented in Table 8 which follows:

Before the booster, the proportion of subjects with detectable Ab titers ranged from 61.7 to 85.7% in the adjuvanted groups compared to 31.3% in the control group. As soon as 7 days after booster immunization (D372) with A/Indonesia, an immune response was observed in the four adjuvanted groups with a GMTR [D372/D365] of 11.30 compared to 4.15 for the control group. This response was comparable across the adjuvanted groups with 99.0% of subjects with detectable Ab titers compared to 81.3% of subjects in the control group. In terms of distribution titers, the proportion of subjects with Ab titers >40 (1/dil) was comparable across the adjuvanted groups, 93.8 to 95.9% of subjects. The 2-fold and 4-fold increase was respectively 95.3% and 86.5% for all the adjuvanted groups compared to 72.3% and 44.7% for the control group.

21 days after booster immunization (D386), this immune response was amplified for the four adjuvanted groups, with a GMTR [D386/D365] of 17.0 compared to 5.18 for the control group. This response remained comparable across the adjuvanted groups with 97.9 to 100.0% of subjects with detectable Ab titers compared to 87.2% of subjects in the control group. Once again, the 2-fold and 4-fold increase was higher in all the adjuvanted groups.

As observed with HI method, the immune response against the ANietnam strain after booster immunization with A/Indonesia was similar to the response observed against the A/Indonesia strain. The four adjuvanted groups had a higher immune response than the control group.

Altogether, immunogenicity analyses performed using HI horse test and SN method provided convergent results regarding the high immune response observed after booster immunization with adjuvanted H₅N₁. The following conclusions could be drawn regarding D386 results:

Booster immunization with the 3.75 μg+AF03 induced a higher immune response in the four adjuvanted groups compared to the control group (HI and SN method). The immune response against the ANietnam strain was similar to the response observed against the A/Indonesia strain (HI and SN method). All three EMEA criteria were met 21 days after booster immunization for all the adjuvanted groups, as soon as 7 days after booster immunization for Group 2 (3.75 μg) and Group 3 (7.5 μg). None of these criteria were met for the control group (HI method). Thus, thanks to the adjuvant emulsion of the invention, it was possible to get a strong cross- reactivity between the strains which have been used, even with a relatively low dose of antigen. Moreover, regarding EMEA criteria, the vaccine was considered as safe whatever the dose.

Claims

1. Use of a composition comprising at least one antigen derived from a H₅ influenza virus strain, for the preparation of an influenza immunization composition for administration according to a regimen comprising at least a 1^st and a 2^nd administration steps timely separated, wherein the H₅ influenza virus strain of the 1^st administration step is different from the H₅ influenza virus strain of the 2^nd administration step and wherein the immunization composition used in at least one of the 1^st or the 2^nd administration step comprises an oil-in-water emulsion as an adjuvant, wherein said oil-in-water emulsion comprises at least squalene, an aqueous solvent, a polyoxy ethylene alkyl ether hydrophilic nonionic surfactant, a hydrophobic nonionic surfactant, and wherein said oil-in-water emulsion is obtainable by a phase inversion temperature process and wherein 90% of the population by volume of the oil drops has a size less than 200 nm, and optionally less than 150 nm.

2. Use according to claim 1, wherein the oil-in-water emulsion is present in the composition used in the 1^st administration.

3. Use according to one of the preceding claims wherein the H₅ influenza virus strain of the 1^st administration step and the H₅ influenza virus strain of the 2^nd administration step belong to different clades or sub-clades.

4. Use according to one of the preceding claim wherein the administration regimen comprises as the 1^st administration steps, 2 shots separated by a period of time which is at least 7 days.

5. A method of immunizing a human against influenza, comprising the step of administering to the human at least: - a 1^st immunization composition comprising at least one antigen derived from a H₅ influenza virus strain,

- and later, a 2^nd immunization composition comprising at least one antigen derived from a H₅ influenza virus strain, wherein the H₅ virus strain virus used in the 1^st immunization composition is different from the H₅ virus strain virus used in the 2^nd immunization composition, wherein at least one of the 1^st or the 2^nd immunization compositions comprises an oil- in- water emulsion comprising squalene, an aqueous solvent, a polyoxyethylene alkyl ether hydrophilic nonionic surfactant, a hydrophobic nonionic surfactant, wherein said emulsion has been obtained by a phase inversion temperature process and wherein 90% of the population by volume of the oil drops has a size less than 200 nm, and optionally less than 150 nm.