FORMULATIONS

The present invention relates to methods of stabilising live vaccines. Current vaccines must usually be stored at low temperatures to maintain their biological activity up to the point of delivery.

The developing world urgently needs new vaccines for diseases such as malaria, HIV-AIDS and tuberculosis, which together kill over 5 million people every year. Unfortunately, the countries most affected by the mortality and morbidity inflicted by these and other infectious diseases are also not well equipped with the costly infrastructure currently needed for childhood vaccination campaigns - most crucially an uninterrupted network of refrigerated storage and distribution facilities, or "cold chain", and its underlying requirement for a reliable electricity supply. With the exception of oral polio vaccine, all currently available live viral vaccines are lyophilized products, and none is sufficiently thermostable to obviate the manufacturers' requirements for storage at 4-8°C or below.

An alternative stabilisation technology for vaccine delivery (especially to the developing world) is required.

The development of heat stable vaccine products that do not require refrigeration would reduce the cost of delivery as well as increase the overall efficacy of a vaccination program, and has been identified as one of the Grand Challenges in Global Health by the Gates Foundation and its partners. Even in the developed world, the cold chain accounts for up to 14% of the cost of immunization.

The development of non-replicating recombinant viral vectors provides a promising new candidate for vaccine technologies, but presents an especially high hurdle for thermostabilisation technologies. As live organisms, these vaccines have an even more stringent requirement for retention of protein conformation (and hence viral infectivity) during storage than vaccines based on purified proteins or inactivated pathogens. Methods of vaccine stabilisation currently available involve some inherent disadvantages. Vaccines stabilised by freeze drying (lyophilisation) not only require reconstitution with sterile saline before administration but often have their biological activity impaired as a result of the protein-damaging during the procedure, which may take several days to complete. Freeze-drying processes also typically result in losses before storage of at least 10° ⁵ even when complex multi-component cryoprotectant formulations are employed. Equally vaccines stabilised by spray drying, a relatively new technique to the field, are exposed to shear stress during the atomisation step, which also has a protein-damaging effect.

Stabilisation procedures involving live vaccines necessitate that the live microorganism (eg. virus or bacterium) must retain activity once reconstituted. Since ambient temperatures in many parts of the developing world can regularly exceed 40-45°C at certain times of year, failure to achieve stability at these temperatures remains a profound problem with live vaccines.

There is therefore a need in the art for alternative and/or improved means for stabilising live vaccines.

The present invention ameliorates one or more of the above problems by providing a method of preparing stable formulations containing virus, which retain biological activity after prolonged storage (eg. at high temperatures). This technology is capable of revolutionising the deployment of viral vaccines in resource-poor settings and rural areas.

The present invention provides a method of stabilising a virus selected from modified vaccinia Ankara (MVA) virus or E1/E3-deleted human adenovirus serotype 5 (AdHu5), wherein said virus retains activity after storage, the method comprising:

(a) forming a liquid mixture of the virus with a glass-forming excipient which comprises trehalose and sucrose;

(b) applying said liquid mixture to a support; and

(c) drying said liquid mixture at a temperature that does not exceed 35"C and does not fall below 1 °C to form a glass containing the virus,

wherein said virus retains activity after storage in the glass at 45°C for at least 4 months, and

wherein said virus expresses an antigen selected from a malarial antigen, an influenza antigen, a tuberculosis antigen and/or an HIV antigen. The virus may be a naturally-occurring virus or a recombinant virus, or a mixture thereof.

The virus is the Poxviridae family virus modified vaccinia Ankara (MVA) virus or the Adenoviridae family virus E1/E3-deieted human adenovirus serotype 5 (AdHuS). In a further embodiment, a mixture of the two viruses may be employed. These two viruses are particularly suited for vaccine applications in which the virus expresses a malarial antigen, an influenza antigen, a tuberculosis antigen and/ or an HIV-AIDS antigen.

After storage and following reconstitution, the virus demonstrates a retained activity. For example, the virus is able to infect a vaccinated individual at the site of administration. Thus, after storage and reconstitution, the virus retains an ability to infect a host cell. Alternatively (or in addition), the virus retains an ability to induce a T cell and/or antibody response after storage.

In one embodiment, the present invention provides a dual vaccine system in which two different viruses expressing the same (or similar) transgenic antigen are used sequentially in a heterologous prime-boost regimen. When said dual vaccine components are stored in accordance with the present invention, following reconstitution, the vaccine components induce a potent T cell and/or antibody response. Alternatively (or in addition), following storage, reconstituted virus demonstrates viability. The reconstituted virus formulation prepared according to the method of the invention demonstrates at least 5% or at least 10% or at least 20% or at least 30% or at least 40% or at least 50% or at least 60% or at least 70% or at least 80% activity vis-a-vis the formulation prior to storage. In a preferred embodiment, the virus is unable to replicate productively (eg. the virus is attenuated).

The glass-forming excipient consists or comprises two different sugars, namely the first sugar trehalose and the second sugar sucrose. Whilst any particular ratio of the sugars may be employed, the present inventors have identified particularly good stability results when a ratio (first sugar: second sugar) of 90:10 through to 50:50 (w/w, w/v, or v/v) is employed. Other preferred ranges include 80:20, 70:30, and 60:40 (w/w, w/v, or v/v), respectively for the first sugar:second sugar.

The glass-forming excipients vitrify upon desiccation, thereby providing a thermostable viral formulation.

The liquid mixture comprising the virus is typically applied (eg. by pipette) on to a surface such as a fibrous support surface (eg. a Pall J200 woven polypropylene (PP) or a Whatman S14 glass fiber (GF) membrane). In other embodiments, an alternative surface such as a polycarbonate (eg. fibrous) support surface may be used. The fibrous support may comprise fibers chosen from the group of materials comprising: glass, silica, metals, ceramic, silicon carbide, carbon, boron, natural fibers such as cotton, wool, hemp, linen, artificial fibers such as viscose, or cellulose fibers, synthetic fibers such as polyesters, polyamides, polyacrylics, chlorofibers, polyolefins, polyimides, synthetic rubbers, polyvinyl alcohol, aramids, fluorofibers, phenolics, etc. The support surfaces are typically inert, and/ or typically have a large surface area. The surfaces are typically at least partially hydrophilic (or comprising a hydrophilic coating) so as to help minimise crystallisation of the liquid mixture. The surface may be treated to render it hydrophilic. Such hydrophilic treatments (to the support surface) include, but are not limited to, straight plasma, corona or ozone processing, or grafting of hydrophilic polymers containing one or more carboxyl, hydroxyl or amine functional groups that serve to loosely bind water (e.g. HydroLAST). Similarly, the surface may be treated with a blocking agent to render it inert (ie. minimal or no substantial physicochemical bonding between the surface and the glass-forming excipient. Examples of such blocking agents include casein, serum albumin, surfactant such as Tween 20 or Tween 80 (RTM of ICI Americas, Inc.), or polymers such as polyvinyl pyrrolidone.

The drying step typically comprises leaving the liquid mixture for a period of time so that the liquid mixture dries into a thin coating(s) on the surface (eg. into thin glass sheaths on fibres of a membrane). This is carried out at a temperature of 1-35°C or 5-35°C or 5-30°C, and a typical ambient temperature of 20-25°C is preferred. Other suitable temperatures include 15-30°C, 15-30°C, 10-25°C or 10-30°C. The above temperature values relate to a temperature to be employed at or around ambient pressure, namely around 1 atmosphere (which equates to approximately 100 kPa; 1 bar; or 15 psi). The entire drying phase is preferably performed at a pressure that does not typically drop below atmospheric pressure - by way of example, the pressure does not typically drop below 800 mbar, and preferably not below 1 bar. The above temperature ranges relate to the temperature at atmospheric pressure and may, of course, vary according to the pressure in line with known laws of physics. The drying phase is typically a single, uninterrupted phase, and results in the formation of a glass having a reduced, low residual water content sufficient to protect the virus within the glass - the resulting glass may have a typical residual water content of less than 10% or less than 6% or less than 4% or not exceeding 2% (units in w/w or w/v or v/v). A low relative humidity during the drying phase of between about 0% and 20% (eg. 2-10%) is typical. The drying step typically takes 6- 30 hours, or 8-28 hours, or 8-26 hours, or 8-24 hours, or 8-18 hours, or up to 24 hours or up to 20 hours or up to 18 hours.

By slowly drying virus suspended in a glass-forming solution on to a support (eg. a filter-like polypropylene or glass fiber support membrane) at 1-30° C (eg. preferably at ambient temperature), an ultra-thin glass is deposited on the support with a thickness of 1-85μηη, or 10-60μιη, or 20-60μιτι. Said glass provides a viscous anhydrous liquid (functionally a solid) in which virus is immobilized and thereby protected from adverse chemistry.

The activity of the virus within the glass is retained for a period of at least 6, 8, 10 or 12 months, when stored at a temperature of at least 45° C.

In the absence of the present invention, virus is typically unstable at elevated temperatures - for example, one or more of infectivity, immunogenicity, T cell response, viability and/ or antibody response is typically reduced after storage for one month or longer at 37°C, 40°C or 45°C.

According to the present invention, virus is formulated with glass-forming excipients (ie., trehalose and sucrose, for example at a final concentration of 0.5M) into liquid mixtures. These mixtures are then dried on to a support (eg. a membrane, or a fibrous support) at a temperature recited above (eg. ambient temperature). During non-abrasive application to the support, the mixture initially forms a syrup that coats the support (eg. fibres), and ultimately forms an evenly distributed ultra-thin layer of glass, with the immobilised virus lying within it. By way of example, in the case of AdHu5, the present invention permitted complete retention of immunogenicity after storage at up to 25°C for polypropylene membranes and up to 45°C for glass fibre membranes for up to 6 months. Deterioration of the vaccine stored on glass fibre membranes at 45°C was only seen after 15 months.

By way of further example, the present invention permitted thermostabilisation of MVA polypropylene and glass fibre membranes at 37°C and 45°C. When stabilised within glass on a support surface, undiminished immunogenicity was achieved at 45°C for 4 months.

AdHu5 and MVA both fully retained viability and immunogenicity after storage in glass-forming excipient at up to 45°C for 4-6 months, and showed substantial stabilisation at lower temperatures for a year or more.

By suspending virus (ie, recombinant E1/E3-deleted human adenovirus type 5 (AdHu5), or modified vaccinia Ankara (MVA) virus) in glass excipient on a support (eg. a membrane), the present inventors were able to completely recover original virus (eg. viral titer and immunogenicity) after storage at up to 45°C for 6 months, and even longer with minimal losses.

Thus, the present inventors have developed a methodology for thermostabilisation of live virus(eg. live virus vaccines) based on formation of a glass-forming excipient at 1-3CTC (eg. preferably at ambient temperature) on a support (eg. woven polypropylene (PP) or glass-fiber (GF) membranes).

Fibrous membranes are preferred as they provide a large surface area for gradual evaporation of water from the virus formulation, resulting initially in the formation of a syrup that coats the fibers, and ultimately in the deposition of an evenly distributed ultra-thin layer of sugar glass, with the immobilized virions of the formulation (eg. vaccine) lying within it (Fig. 2).

To demonstrate this technology of the virus to be stabilised, the inventors employed live AdHu5 and MVA recombinant vaccine vectors carrying a model recombinant antigen (GFP and an epitope string) to determine viral titer in vitro and CD8 ⁺ T cell responses in vaccinated mice. To gain an understanding of the intrinsic resistance of these viruses to heat and drying, the vaccine vectors were first subjected to storage at a range of temperatures, either in buffered liquid, or dried onto GF and PP membranes in the absence of glass-forming sugar stabilizers. Both viruses lost considerable titer and immunogenicity upon storage at elevated temperatures in liquid and in the dry state.

The viruses (AdHuS and MVA) fully retained viability and immunogenicity after storage of glass-forming excipient thermostabilised membranes at < 45°C for 4 - 6 months, and showed substantial stabilization at lower temperatures for a year or more. Differential calorimetric analysis of the membranes revealed that the T _g of the sugar glasses was in excess of 50°C for MVA and 60°C for AdHu5 (Table 1 ). This is the temperature at which the glass begins to soften, and becomes liquid rather than solid. Below the transition temperature, the molecules are immobilized, and the glass-forming excipient may substitute for water in the hydrogen-bonded hydration shell of proteins, so that degradative mechanisms such as those observed during storage without stabilization (Figs 1 and 3) does not occur.

An unexpected technical effect of the present invention is that, following reconstitution, the virus is able to induce antigen-specific CD8* T cells, which may be particularly valuable for protection against pathogens such as Plasmodium spp. and HIV-1 that are able to evade the antibodies typically induced by traditional immunization approaches. Adenovirus- and MVA-based heterologous prime-boost vaccination regimes (in which different vectors expressing the same recombinant antigen are administered sequentially) also induce protective humoral immunity, especially when the recombinant antigen is a secreted or membrane-anchored protein.

The viral doses deposited on supports used according to the method of the present invention approximate those that are employed in the clinic. In more conventional units, the inventors typically dried and recovered over 10 ³ plaque-forming units of MVA (the typical human dose of the MVA-85A candidate tuberculosis vaccine) and over 10 ¹⁰ viral particles of AdHuS from single membranes. The recovery volume (500 μΙ) is one commonly used for injection of vaccines via the intramuscular or subcutaneous routes. Even substantial reduction in the effective dose of MVA did not result in impaired responses. Remarkably, after immunization of mice with MVA where only ~10 ³ CCID ₅₀ of titerable virus was recovered from the heated filter (a 4 log drop), the ELIspot response was only reduced by a factor of ~100. Nevertheless, no effect was observed on immunogenicity in the absence of an impaired recoverable titre, indicating that the thermostabilisation technology fully protect both the in vivo and in vitro viral bioactivities.

A technical strength of the support-based (eg. membrane-based) drying system of the present invention is that there is no need for the application of any potentially protein-damaging procedures such as heating or freezing, as are employed in spray-drying and lyophilisation, respectively, during desiccation of the virus- containing formulation to generate a glass with high T _g . Thus, procedures such as heating, cooling and/ or freezing are typically absent. Similarly, mechanically abrasive techniques (such as high-pressure spraying) for applying the virus- containing glass-forming excipient on to the support may be avoided. In this regard, the virus-containing glass-forming excipient is typically applied on to the support by simple dripping (such as by pipetting) or painting (eg. by a sterile applicator), or by light non-abrasive spraying. Despite its comparative simplicity, the residual moisture content that can be achieved is similar to that typically achieved by conventional drying technologies (eg. 0-10% or 0-8% or 0-5% or 0-2%). In contrast, freeze-drying processes typically result in losses before storage of at least 10° ⁵ even when complex multi-component cryoprotectant formulations are employed. Live viruses present a particularly challenging thermostabilisation target, since both protein conformation and nucleic acid integrity must be preserved in order to retain infectivity and hence immunogenicity. Thus, confirmation that the thermostabilisation technology of the present invention successfully preserves the bioactivity of live viruses indicates that the constituent proteins of these viruses are maintained in their native conformations (or at least sufficiently so for normal infectivity and immunogenicity).

The thermostabilisation performance of the glass-forming technology of the present invention exploits the large surface area present on support surfaces such as filter- like matrices (GF or PP). This provides (i) efficient evaporation during ambient temperature drying; (ii) a hydrophilic surface upon which the high-T _g glass is formed and deposited; and (iii) efficient dissolution upon reconstitution.

There is an added bonus to the use of a support (eg. a membrane), namely the ability to avoid the need to dispense with traditional, bulky, heavy and fragile vaccine vials. Instead, the formulation-coated support prepared according to the method of the invention can be aseptically packaged within an injection-moulded inline support holder as part of an all-in-one ready-to-inject vaccine delivery device such as a HydRIS system (hypodermic rehydration injection system - see, for example, WO 2007/057717). In this embodiment, the dissolution of the glass-forming excipient and reconstitution of the vaccine occur concomitantly with the flow of buffer from the syringe, through or across the in-line stabilized membrane, and into the attached needle, so that no separate reconstitution step is required in the field. Such devices occupy similar packaging space as syringes and would already contain the thermostable vaccine ready for administration. This combination of factors is particularly suited for low-tech distribution routes in rural areas, enabling better penetration of disease prevention measures into resource-poor settings.

Figures

Fig. 1 illustrates the effect of time and temperature on infectivity and immunogenicity of AdHu5 and MVA live recombinant viral vectors. Thermal instability of (A) AdHu5 and (B) MVA stored in liquid solution (buffered with Tris pH 7.4 for AdHuS and pH 9.0 for MVA; see Materials and Methods) at the indicated temperatures for 1 week (AdHuS) or 1 or 4 months (MVA; 1m, 4m) in comparison to controls stored at -80°C. Right y-axes: mean viral titer per ml (bars) measured by endpoint dilution and expressed as cell culture infectious dose (CCID ₅₀ ) with error bars showing SEM. Left y-axes: immunogenicity, one circle per mouse with horizontal line representing the mean, expressed as spot forming cells (SFC) per million splenocytes. Specific CD8 ⁺ T cell responses to Pb9 peptide (present in the recombinant antigen expressed by both viral vectors) were measured by IFN-γ ELIspot 2 weeks after intradermal immunization with 50μΙ of material (i.e. 1/20* of the indicated titer per ml, which corresponds to 6.9 x 10 ⁶ CCID _S0 of AdHu5 or 3.8 x 10 ⁷ CCID ₅₀ of MVA per mouse if undegraded). Asterisks, statistically significant difference (p < 0.05) in both immunogenicity and titer versus control by one-way ANoVA followed by Dunnet's test. LD, limit of detection of CCID ₅₀ assay reached. Fig. 2 illustrates scanning electron micrograph of trehalose-sucrose glass deposited on a GF membrane. Virus formulated in 0.5M trehalose and sucrose was pipetted onto the membrane and dried at 20 - 25°C at 2 - 10% relative humidity. During this process, a sugar glass forms and coats the membrane to form an ultra- thin layer of web-like structures between and around the individual fibers. The membrane thus provides a high surface-area to volume ratio for rapid drying and a support matrix for the resulting glass, as well access for buffer during reconstitution. Fig. 3 illustrates resistance of viruses to drying on membranes with and without glass-forming sugars. Recovery of titer and immunogenicity of (A) AdHuS or (B) MVA immediately after drying onto PP or GF membranes in the presence (stabilized) or absence (unstabilised) of trehalose-sucrose in comparison to controls stored at -80°C. Data are displayed as in Fig. 1. Asterisks, statistically significant difference (p < 0.05) in both immunogenicity and titer versus control by one-way ANoVA followed by Dunnet's test.

Fig. 4 illustrates AdHuS stability on dried membranes without sugar glass stabilization. Recovery of titer and immunogenicity of AdHuS after 1 month's storage at the indicated temperatures after drying in the absence of glass-forming sugar stabilizers onto (A) PP membranes or (B) GF membranes in comparison to controls stored at -80°C. Data are displayed as in Fig. 1. Asterisks indicate statistically significant difference (p < 0.05) in both immunogenicity and titer versus control by one-way ANoVA followed by Dunnet's test.

Fig. 5 illustrates AdHu5 sugar glass thermostabilisation on PP membranes. Recovery of titer and immunogenicity of AdHuS dried with trehalose-sucrose on PP membranes after storage for (A) 1 month, (B) 3 months, (C) 6 months, or (D) 15 months at the indicated temperatures, in comparison to control viruses stored at -80°C. Data are displayed as in Fig. 1. Asterisks, statistically significant difference (p < 0.05) in both immunogenicity and titer versus control by one-way ANoVA followed by Dunnet's test.

Fig. 6 illustrates AdHu5 sugar glass thermostabilisation on GF membranes. Recovery of titer and immunogenicity of AdHu5 dried with trehalose-sucrose on GF membranes after storage for (A) 1 month, (B) 3 months, (C) 6 months, or (D) 15 months at the indicated temperatures, in comparison to control viruses stored at -80°C. Data are displayed as in Fig. 1. Asterisks, statistically significant difference (p < 0.05) in both immunogenicity and titer versus control by one-way ANoVA followed by Dunnet's test.

Fig. 7 illustrates MVA stability on dried membranes without sugar glass stabilization. Recovery of titer and immunogenicity after (A) 1 month or (B) 4 months storage at the indicated temperatures after drying in the absence of glass-forming sugar stabilizers onto PP membranes or GF membranes in comparison to controls stored at -80°C. Data are displayed as in Fig. 1. ND, not done. Immunogenicity for PP membranes in (B) was tested at 7 months. Asterisks, statistically significant difference (p < 0.05) in both immunogenicity and titer versus control by one-way ANoVA followed by Dunnet's test Fig. 8 illustrates MVA sugar glass thermostabilisation on GF and PP membranes. Recovery of titer and immunogenicity of MVA dried with trehalose-sucrose on (A) GF membranes or (B) PP membranes and stored for 1 month, 4 months (7 months for PP membrane immunogenicity only), or 12 months at the indicated temperatures, in comparison to control viruses stored at -80°C. Data are displayed as in Fig. 1. Asterisks, statistically significant difference {p < 0.05) in both immunogenicity and titer versus control by one-way ANoVA followed by Dunnet's test.

EXAMPLES

Materials and methods

Virus production

Recombinant MVA and E1/E3-deleted AdHu5 expressing the model antigen TIP (an epitope string) with a GFP marker (M. G. Cottingham, R. F. et al (2008) PLoS ONE, 3 (2), e1638) were derived as described in S. Sridhar, A. et al (2008) J Virol, 82 (8), 3822-33, and J. Schneider, S. C. et ai (1998) Nat Med, 4 (4), 397-402. The immunodominant CD8 ⁺ T cell epitope in this construct in BALB/c mice is the protective H2-K ^d -restricted epitope Pb9 from Plasmodium berghei circumsporozoite protein (P. Romero, J. L. et al (1990) Immunol Lett, 25 (1-3), 27-31). The sequences immediately flanking Pb9 are identical in the two constructs, but in AdHuS, the epitope string is fused to the N-terminus of GFP, while in MVA it is expressed from a separate viral promoter. MVA was grown in primary chick embryo fibroblasts and purified by centrifugation through 36% trehalose instead of the conventional sucrose and resuspended in 10mM Tris pH 9.0. AdHu5 was grown in HEK293 cells, purified by CsCI gradient ultracentrifugation and dialysed against 10mM Tris pH 7.4. Viral titration

The viruses were titered by endpoint dilution assay of infectious dose in cell culture (CCIDso) on DF-1 cells (MVA) or HEK293 (AdHu5) cells grown under standard conditions. Wells were scored positive or negative 7 days (MVA) or 10 days (AdHu5) after infection by GFP fluorescence with a BMG Fluostar 96-well plate fluorimeter with a cut-off of the mean plus three standard deviations above uninfected negative controls. Titers per ml were calculated with the Spearman- Karber method.

Thermostabilisation and recovery

Viruses were formulated for thermostabilisation by 20-fold (AdHu5) or 5-fold (MVA) dilution of the purified virus suspensions into an unbuffered 0.5M solution of mixed glass-forming disaccharides (trehalose and sucrose). This solution was used for studies of instability in liquid formulation (Fig. 3) and for subsequent input to the thermostabilisation-desiccation procedure. Formulated viruses were pipetted onto Pall J200 woven polypropylene (PP) or Whatman S14 glass fiber (GF) membranes. The quantity of virus applied per membrane was 55μΙ of the feedstock, corresponding to 3.8 x 10 ⁸ CC!D ₅₀ of MVA or 3.5 x 10 ⁷ CCID _S0 (or 1.1 x 10 ¹⁰ viral particles by A ₂₆ o) of AdHu5. These were dried overnight (up to 18 hours) at ambient temperature (20 - 25°C) with a silica gel bed to control humidity (typically 10 - 20% relative humidity). Once dried, the membranes were handled in a low humidity glove box and kept at 2 - 10% relative humidity before transfer into heat-sealed moisture barrier bags (Dri-Shield 3000, 3M) for storage and stability testing. For reconstitution, membranes (1 for MVA or 2 for AdHu5) were placed into bijou-type vials, and 500μΙ of PBS was added prior to brief agitation and incubation for 5-10 minutes.

For positive control samples, identical 55μΙ aliquots of feedstock were stored at - 80°C and diluted to 500μΙ_ with PBS for use. This freeze-thaw treatment and cryogenic storage did not affect viral viability, since the measured control titer per ml post-formulation and freezing (see Fig.3 to Fig.8; range across experiments from _{1 0} 8.o _ _{1 0} 8.9 _CC | _{D5o m} | _{for M} vA and from 10 ⁷⁸ - 10 ⁸⁴ CCID ₅₀ /ml for AdHu5) did not differ significantly from the expected titer per ml (10 ^{8 8} CClD507ml for MVA and 10 ^{8 1} CCID50/ml for AdHu5; also given above per filter), calculated from the formulation and reconstitution dilution factors (see above) and the titers of the original purified viral preparations, which were based on single duplicate titrations (3.5 x 10 ¹⁰ CCID50/ml_ for MVA and 1.3 x 10 ¹⁰ CCIDSO/mL for AdHu5).

Physico-chemical analyses

Differential scanning calorimetric analysis of glass transition temperature (T _g ) was carried out using a Diamond DSC Calorimeter (Perkin Elmer). An indium standard was used to calibrate the instrument and sample membranes carrying vaccine- bearing sugar glasses were placed into DSC pans (Perkin Elmer) and sealed, while an empty pan was placed in the reference cell. The sample chamber was purged with nitrogen gas in order to remove oxygen, prior to scanning from -20°C to 160 °C at a heating rate of 100°C per minute. The onset point of the glass transition was considered as the characteristic temperature of the transition. The residual moisture content of the membranes was determined with a Karl Fisher coulometer fitted with an oven (Metrohm) from which a carrier gas transfers released water to the titration cell. The instrument was calibrated using KF oven water standard (Sigma) and the water content was calculated as a percentage of total sample weight.

Immunogenicity

After reconstitution, 50μΙ per mouse of recovered material (1/10 of the reconstitution volume) was injected intradermal^ (i.d.) into groups of four Balb/c mice (25μΙ per ear) in accordance with UK Animals (Scientific Procedures) Act (1986) legislation. For AdHu5, two membranes were reconstituted in 500μΙ, so the dose was 6.9 x 10 ⁶ CCID ₅₀ (2.2 x 10 ^s viral particles) per mouse; for MVA only one membrane was used, so the dose was 3.8 x 10 ⁷ CCID ₅₀ per mouse. One week (MVA) or 2 weeks (AdHu5) after immunization, the animals were sacrificed and splenocytes were isolated. CD8 ⁺ T cell responses were measured by IFN-γ ELIspot assay by using stimulation with 1 pg/ml PbQ peptide (SYIPSAEKI) as described (M. G. Cottingham, R. F. et al (2008) PLoS ONE, 3 (2), e1638).

Example 1 - Thermal instability ofAdHuS and MVA

To determine the effect of elevated temperature upon viral viability and immunogenicity, AdHu5 or MVA recombinant vaccines were subjected to liquid storage in buffered solution (see Materials and Methods - above). Recombinant viruses expressing green fluorescent protein (GFP) and a model epitope string antigen, TIP (M. G. Cottingham, R. F. et al (2008) PLoS ONE, 3 (2), e1638), were employed to facilitate measurement of viral titer and specific CD8 ⁺ T cell responses to the P£>9 epitope (P. Romero, J. L et al (1990) Immunol Lett, 25 (1-3), 27-31) in vaccinated mice. At temperatures above ambient, adenovirus immediately begins to lose activity, such that by one week at 37°C or 45°C, no infectious virus remains and the vaccine is wholly ineffective (Fig. 1A). MVA loses infectivity and immunogenicity after storage for one month or longer at 37°C or 45°C (Fig. 1 B). Thus, these viruses exhibit significant degradation in solution over the period of a typically desirable product shelf-life.

Example 2 - Desiccation in the presence or absence of glass-forming stabilizer

Before embarking on long-term storage studies, it was first determined whether MVA and AdHu5 could be dried onto membranes at ambient temperature and reconstituted without loss of infectivity and immunogenicity. Two substrates were investigated: woven polypropylene (PP) membranes or glass fiber (GF) membranes (Fig. 3). The membranes differ in their surface chemistry: detergent treatment of the hydrophobic PP membranes with Crillet-1 (Croda Inc.) is preferred to allow the surface to receive the vaccine solution. For drying, the viruses were formulated with or without glass-forming sugar stabilizers (trehalose and sucrose at 0.5M; see Materials and Methods). The formulated solution was pipetted onto the membranes, dried overnight at ambient temperature (20 - 25°C) with low relative humidity (2 - 10%), and then reconstituted into 500μΙ PBS. As a control, equal volumes of formulated solution were stored at -80°C and diluted to 500μΙ with PBS for use. MVA could be dried onto membranes and reconstituted without loss of infectivity or immunogenicity, even in the absence of trehalose-sucrose, but AdHuS lost activity upon drying onto both PP and GF membranes unless the glass-forming sugars were present during desiccation (Fig. 3). When formulated with stabilizers, the present inventors were able to recover all of the activity applied to the membrane after reconstitution, indicating that the losses during processing are zero or very close to zero (Fig. 3).

Example 3 - Physico-chemical properties of vaccine-bearing glasses

After drying of vaccine formulated in trehalose and sucrose onto membranes, an ultra-thin sugar glass is formed on and between the fibers of the support matrix (Fig. 2). To further characterize the MVA- and AdHu5-containing glasses formed on GF and PP membranes, the glass transition temperatures (T _g ) were determined by differential scanning calorimetry and the sample residual moisture content by the Karl Fischer method. As shown in Table 1, there was little difference in either of these parameters between the two types of membrane, but VA-containing glasses had a slightly lower T _g and correspondingly higher residual moisture content than the AdHu5-bearing glasses.

TABLE 1

Table I illustrates the physico-chemical properties of virus-containing sugar glasses formed on membranes. Glass transition temperatures, measured by differential scanning calorimetry, and percentage residual moisture content, measured by the Karl Fischer method, of AdHuS and MVA vaccines formulated in 0.5M trehalose- sucrose immediately following drying onto polypropylene (PP) or glass fiber (GF) membranes. Data shown are the mean plus or minus the standard error of the mean of two replicates.

Example 4 - Long-term storage stability: AdHu5

The recovery of infectious, immunogenic AdHu5 from trehalose-sucrose stabilized, dried PP membranes or GF membranes after storage for extended periods of time at 4°C, 25°C, 37°C or 45°C under conditions of low relative humidity (2 - 5%) was then examined. On unstabilised membranes of either type, AdHu5 showed worsening losses of titer and immunogenicity with increasing temperature after one month's storage (Fig. 4), with complete loss of bioactivity at 37°C or higher and substantial losses even when refrigerated. In contrast, glass stabilization during drying allowed complete retention of titer and immunogenicity after storage at up to 25°C for PP membranes (Fig. 5 A-C) or up to 45°C for GF membranes (Fig. 6 A-C) for up to six months. Even after extended storage (15 months) only small, nonsignificant reductions in either parameter at 4°C - 37°C were observed.

Example 5 - Long term storage stability: MVA

A similar study with MVA thermostabilised on PP and GF membranes at 37°C and 45°C was then performed. With glass thermostabilisation, undiminished viral titer and immunogenicity was achieved at up to 45°C for 4 months.