CA2304824A1

CA2304824A1 - Antigen vectors in the form of multilamellar vesicles and compositions containing antigens encapsulated in these vesicles

Info

Publication number: CA2304824A1
Application number: CA002304824A
Authority: CA
Inventors: Patrick Mahy; Brigitte Delord; Joelle Amedee; Olivier Freund; Didier Roux; Rene Laversanne; Dominique Kaiserlian; Sophie Gaubert
Original assignee: Individual
Current assignee: Capsulis SA
Priority date: 1997-09-29
Filing date: 1998-09-17
Publication date: 1999-04-08
Also published as: EP1024830A1; AU744308B2; FR2769022A1; ES2234152T3; DE69827837D1; EP1024830B1; FR2769022B1; JP2001518452A; AU9169798A; WO1999016468A1; DE69827837T2

Abstract

The invention concerns an antigen vector, in particular a protein antigen, consisting of a dispersion of multilamellar vesicles within which said antigen is at least partially encapsulated, said vesicles consisting of two layers comprising at least a surfactant, said two layers being concentric and providing said vesicles with an onion-like structure. The invention also concerns compositions containing at least an antigen against which it is desired to produce antibodies, said antigen being partially encapsulated in such a vector. The invention further concerns the method for preparing said compositions.

Description

Antigen vectors in the form of multilamellar vesicles and compositions containing antigens encapsulated in these vesicles The present invention relates to antigen vectors in the form of multilamellar vesicles and to compositions containing antigens encapsulated in these vesicles.
More than 100 years after the discovery of infectious agents by Louis Pasteur, the following question is still being asked: how does one prepare the immune system to be durably effective against an infection under conditions of total safety? The research objective is not only to develop new vaccines, but also to improve those already in existence by reducing the number of injections.
Currently the majority of viral vaccines are attenuated live viruses. This strategy has been less fruitful against bacteria, there being only one attenuated bacterial vaccine in use at the present time: BCG. The process of microorganism attenuation has made enormous advances through the techniques of molecular biology, which make it possible to modify the virulence function of these pathogens.
Nevertheless, it is diffcult to reconcile the attenuation of this virulence with an effective stimulation of the immune defenses. So-called molecular vaccines based on purified antigens require the participation of adjuvants, for which a whole area of research has yet to be developed. By way of example, the first vaccine of this type was developed against hepatitis B in 1987 from a protein of the viral envelope (HBs antigen) produced by animal cells. There are still only very few molecular vaccines because purified antigens are rapidly eliminated by the organism. Furthermore, purified antigens injected by themselves are rather ineffective and require the use of substances which favor their recognition by the immune system and which slow down their degradation: adjuvants. Aluminum hydroxide, Al(OH)~, discovered in 1926, is currently a permitted adjuvant for humans. New substances - synthetic derivatives of the mycobacterial wall (muramyl dipeptide and derivatives thereof) -have been discovered in recent years.
However, a multitude of tests are required to demonstrate the safety of these new agents before they can be used in human vaccines. Finally, the problem of the number of injections arises because, in the absence of boosters, the immune memory becomes inadequate.
Modern vaccination, particularly that which uses recombinant proteins, is based on the immune response associated with the administration of an antigen by different routes (intramuscular, oral). However, said immune response may be poorly stimulated, either because of the inadequate bioavailability of the antigen (too short a lifetime, proteolysis and dilution in the biological fluids), or because of the absence of capture and of an effective presentation by the antigen-presenting cells (accessibility of the dendritic cells at the immunization site, conformation of the antigen not allowing association with both class I and class II molecules of the major histocompatibility complex), or because of imperfect recognition of the antigenic determinants by the immune cells, even when the protein sequence is optimally designed (e.g.: the glycoproteins associated with HIV, or tumoral markers).
It is therefore of very great interest to vectorize these molecules in an attempt to overcome the limitations described above.
Conventionally, the aim of vaccination is to elicit a humoral immune response. This system protects the vaccinated individual from possible exposure to the pathogen, thereby preventing the dissemination of an extracellular infection.
1 S However, it is powerless against an intracellular infection. The cellular immune response is therefore the essential complement for elimination of the infected cell.
Consequently, an ideal vaccination should favor both types of immunity:
cellular and humoral. The cellular immune response makes it possible to eliminate intra-cellular pathogens such as viruses or to destroy cancerous cells. At the present time, there are still numerous difficulties associated with obtaining an appropriate immune response by administering an antigen. Thus it is still impossible to obtain a vaccination against specific tumoral antigens (e.g. breast cancer) or against weakly immunogenic antigens (e.g. glycoproteins associated with HIV).
There is therefore an obvious need for a method of vectorizing antigens in order to elicit a cellular response while avoiding possible anaphylactic shocks or proteolysis in the serum.
A multitude of diseases could benefit from the administration of vectorized therapeutic antigenic proteins or peptides. This would make it possible to reduce the undesirable side effects associated with the excessive amount of antigens which has to be administered in order to "provoke" the immune response (e.g.
anticancer vaccination).
The inventors of the present invention developed, for different applications in areas totally different from those of the present invention, multilamellar vesicles which have a so-called "onion" structure and consist, from their center to their periphery, of a succession of lamellar layers separated by a liquid medium.
These vesicles can be obtained by a process comprising the preparation of a lamellar liquid-crystal phase and its transformation by the application of shear. Such a process is described in particular in patent WO 93/19735 derived from French patent FR-2 689 418 or in WO 95/18601, which are introduced here by way of S reference.
According to French patent FR-2 689 418, this transformation can be effected during a homogeneous shearing step carried out on the liquid-crystal phase to give vesicles, also referred to as microcapsules, of controlled size.
However, by varying the formulation of the lamellar liquid-crystal phase, and particularly the nature of the surfactants forming part of its composition, this liquid-crystal phase can be transformed into vesicles simply by means of mechanical stress, in particular during the mixing of the constituents.
The researches conducted by the inventors led them to the discovery that the use of the technologies described above afforded the development of novel multilamellar vectors of small dimensions which have a high encapsulating efficiency and a great ease of preparation and which can be used to vectorize antigens with the result of increasing the immune response of the organism to the encapsulated antigen.
Such a vector has in many respects the advantages sought for a vaccination adjuvant.
Another advantage is that all the molecules involved in the preparation of the compositions of the invention are commercially available.
Other advantages are associated with the particularly simple process for the preparation of the multilamellar vesicles useful according to the invention, which enables a wide variety of surfactants to be used.
Another advantage, also associated with the process for the preparation of the multilamellar vesicles useful according to the invention, is that, as will be apparent from the following account, it makes it possible to coencapsulate several antigens in the same vesicle, thereby optionally providing access to polyvalent vaccines, or to coencapsulate, with one or more antigens, additives for functionalizing the vesicles, making it possible e.g. to improve the targeting of the composition or to influence the orientation (humoral or cellular) of the immune response.
Another advantage, again associated essentially with the process used to prepare the onion-structured vesicles used according to the invention, is the fact that the active ingredients and the additives are incorporated prior to the formation of the vesicles, affording an excellent encapsulation yield and hence a better efficiency and a very substantial saving of extremely expensive molecules.
Furthermore, another advantage, also associated with the process employed to prepare the vesicles used according to the invention, is that either water-soluble or liposoluble active ingredients or additives can be incorporated.
More precisely, the present invention proposes a novel means of vectorizing antigens with a view to increasing the immune response of the organism to this antigen or these antigens. The present invention relates more precisely to a method of administering an antigen whereby both the cellular and humoral immune responses are significantly improved.
Thus, according to one of its essential characteristics, the invention relates to a novel antigen vector consisting of a dispersion of multilamellar vesicles within which said antigen is at least partially encapsulated, said vesicles consisting of bilayers comprising at least one surfactant, said bilayers being concentric and giving said vesicles an onion structure.
Antigen is understood as meaning any substance which elicits a response from the immune system and/or the production of antibodies. An antigenic substance comprises at least one antigenic determinant, or epitope, composed of molecular moieties recognized by specific antibodies and/or the cells of the immune system.
The antigens used according to the invention are advantageously antigens of protein type. In particular, they may be antigens consisting of natural or recombinant proteins or of peptides.
The vesicles of the invention may also be used as vectors for antigens of glycoprotein or polysaccharide type.
The vesicles of the invention may contain one antigen or a mixture of antigens. Mixtures of antigens which may be mentioned in particular are those obtained either by synthesis, or by biotechnology, or by extraction from the microorganisms in question.
As explained previously, "onion" structure is understood as meaning a multilamellar structure in which the vesicles of substantially spherical shape consist of a succession of concentric bilayers from the center to the periphery of the vesicles, which is why the name onion structure is used by analogy to describe such structures.

Such structures are advantageously obtained by incorporating at least one antigen into a lamellar liquid-crystal phase comprising at least one surfactant, and then transforming this lamellar liquid-crystal phase into a dense phase of multilamellar vesicles of small dimensions.
$ Thus, according to another essential characteristic of the invention, it relates to a process for the preparation of a composition containing at least one antigen, wherein a lamellar liquid-crystal phase incorporating said antigen is prepared and said liquid-crystal phase is caused to rearrange into multilamellar vesicles by the application of shear.
This shear may be homogeneous shear, which has the advantage of producing vesicles of perfectly homogeneous size. However, simple mechanical agitation may prove sufficient to lead to the formation of the multilamellar vesicles of the invention.
According to yet another characteristic, the invention relates to the products 1$ obtainable by this process.
According to French patent FR-2 689 418, this transformation can be effected during a homogeneous shearing step carried out on the liquid-crystal phase to give vesicles or microcapsules of controlled size. However, by varying the formulation of the lamellar liquid-crystal phase, and particularly the nature of the surfactants forming part of its composition, this liquid-crystal phase can be transformed into vesicles simply by means of mechanical stress, in particular during the mixing of the constituents.
The formulation advantageously includes a mixture of surfactant molecules.
The general procedure is to use at least two different surfactants with different hydrophilic-lipophilic balances, which makes it possible to regulate the properties of the bilayers continuously and thus to control the appearance of the instability which governs the formation of the multilamellar vesicles.
It seems that this structure is responsible for the particularly advantageous results obtained and that the multilamellar vesicles of the invention enable the antigen to arrive intact at the antigen-presenting cells (APC) and make it possible to favor its capture by these cells. It does therefore seem that the function of the vesicles of the invention is to vectorize and protect the antigen and improve its capture by the immune system.
In one advantageous variant, the vesicles constituting the compositions of the present invention have diameters below 20 pm, preferably below 10 p.m, particularly preferably below 1 ~m and very particularly preferably of between 0.1 and 1 p.m.
In one advantageous variant, the membranes of the vesicles constituting the compositions of the invention contain at least one surfactant selected from the S group consisting of - hydrogenated or non-hydrogenated phospholipids, - linear or branched, saturated or mono- or polyunsaturated C6 to C,s fatty acids in the form of the acid or an alkali metal, alkaline earth metal or amine salt, - ethoxylated or non-ethoxylated esters of these same fatty acids with . sucrose, sorbitan, mannitol, glycerol or polyglycerol, or glycol, - mono-, di- or triglycerides or mixtures of glycerides of these same fatty acids, - ethoxylated or non-ethoxylated, linear or branched, saturated or mono- or poly-unsaturated C6 to C,s fatty alcohols, - ethoxylated or non-ethoxylated ethers of these same fatty alcohols with sucrose, . sorbitan, mannitol, glycerol or polyglycerol, or glycol, - hydrogenated or non-hydrogenated, polyethoxylated vegetable oils, - polyoxyethylene and polyoxypropylene block polymers (poloxamers), - polyethylene glycol hydroxystearate, and - alcohols with a sterol skeleton, such as cholesterol and sitosterol.
The chosen surfactants, particularly those mentioned above, are advantageously selected from the category of surfactants which are permitted by legislation for pharmaceutical use as a function of the route of administration.
Two surfactants with relatively different properties, particularly a different hydrophilic-lipophilic balance (HLB), will advantageously be selected from those listed above. The first surfactant will advantageously have a hydrophilic-lipophilic balance of between 1 and 6, preferably of between 1 and 4, while the second surfactant will have a hydrophilic-lipophilic balance of between 3 and 15, preferably of between 5 and 15.
As seen previously, the size of the multilayer vesicles may be better controlled by following the procedure given in patent FR 2 689 418.
The preparation obtained after transformation of the lamellar liquid-crystal phase into multilamellar vesicles can then be diluted, particularly with an aqueous solvent, e.g. a buffer solution, a saline solution or a physiological solution, to give an aqueous suspension of vesicles.
The encapsulation technique used according to the present invention easily makes it possible to achieve very high encapsulation yields, even in the order of 100%. However, depending on the intended applications, such yields are not always essential.
Thus the encapsulation yield of the antigens) in the compositions of the invention is advantageously greater than 50%, preferably greater than 80%.
Another advantage associated with the encapsulation technique of the invention is that it makes it possible to coencapsulate several antigens within the same vesicles, but also to coencapsulate one or more antigens with additives for functionalizing these vesicles, making it possible particularly to improve the targeting. Cytokines, for example, may be coencapsulated, thereby enabling the orientation (humoral or cellular) of the immune response to be influenced.
The invention provides a means of vectorizing antigens, particularly antigens of protein type, it being possible for this vectorization to be effected in vivo after intravenous or intraperitoneal injection.
According to another of its essential characteristics, the invention relates to a composition containing at least one antigen against which it is desired to produce specific antibodies or a specific cellular response, in which composition said antigens) is (are) at least partially encapsulated in the multilamellar vesicles of a vector as defined above.
The compositions containing at least one antigen encapsulated in the multilamellar vesicles described above can be either pharmaceutical compositions for use as vaccines or compositions for the production of monoclonal antibodies.
Finally, according to another of its essential characteristics, the invention relates to a process for the preparation of the compositions defined above.
This process comprises the following steps:
- preparation of a lamellar liquid-crystal phase comprising at least one 3 S surfactant and incorporating the antigen(s), g - transformation of the liquid-crystal phase into onion-structured multi-lamellar vesicles by shear, and - dispersion in an aqueous solution to give the desired concentration of antigen(s).
More precisely, the first step of this process comprises the preparation of a liquid-crystal phase consisting of a mixture of amphiphilic molecules and an aqueous solution containing the antigen(s). In particular, the aqueous phase can contain PBS or any other physiological solution. Of course, the procedure will be adapted to take account of any lack of solubility of certain antigens, in which case different additives or cosolvents, such as glycerol, polyethylene glycol, propylene glycol or urea, will advantageously be incorporated into the aqueous phase in order to favor the dissolution of the antigen. Those skilled in the art will choose the nature of the surfactants and their proportions so as to facilitate the subsequent transformation of the liquid-crystal phase into onion-structured multilamellar vesicles. A possible but non-obligatory choice of surfactants will be at least one amphiphilic molecule of a rather hydrophobic nature, preferably with an HLB
below 4, and another amphiphilic molecule of a rather hydrophilic nature, preferably with an HLB above 5. Such a choice will facilitate the transformation of the liquid-crystal phase.
In a second step, the homogeneous mixture consisting of the liquid-crystal phase formed is sheared by one of the methods described above. The shear applied will advantageously be homogeneous so as to give a uniform vesicle size.
However, it is not always found necessary to use homogeneous shear and, provided the formulation of the liquid-crystal phase is appropriately adapted, simple mixing in an industrial mixer can suffice to transform the lamellar liquid-crystal phase into multilamellar vesicles according to the invention.
In a third step, the mixture consisting of the multilamellar vesicles according to the invention is dispersed in excess aqueous solution, for example physiological serum, to bring this mixture to the desired antigen concentration.
The invention is illustrated by the Examples below, which show that it is possible to envisage using the multilamellar microvesicles of surfactants for human and animal vaccination with the following advantages:
1 ) It is possible to vectorize one or more antigens, separately or together, in multilamellar vesicles of small dimensions.
2) Given the increase in the immune response, it is possible to envisage using a smaller amount of antigen. This can be important when the use of recombinant and/or expensive antigens is envisaged.
3) The amplification of the immune response by microencapsulation makes it possible to envisage using weakly immunogenic molecules which, by themselves, do not produce a suffcient immune response.
4) The speed of development of the immune response makes it possible to envisage reducing the intervals between administrations of the successive boosters.
In fact, only 30% of treated persons attend for the final booster, so 70% of individuals are not immunized effectively even though they have had a first injection. A shorter interval between boosters ought to ensure that the population is better vaccinated.
S) The increase in the cellular immune response is a guarantee of better vaccination.
6) The fact that the amplifying effect of an encapsulated antigen on the immune response is even more pronounced with the presence of an adjuvant, such as QS21, makes it possible (without obligation) to envisage a coupling of technologies.
In conclusion, the results demonstrate a very general amplifying effect on the immune response which can potentially be applied in numerous human and animal vaccines or for the production of compositions containing antigens.
The Examples which follow are given purely by way of illustration of the invention, without implying a limitation.
Example 3 is illustrated by Figures 1 to 4 as follows:
- Figure 1 shows the results concerning the production of anti-HSA serum IgGs after immunization according to the IP/IV protocol defined in Example 3, at D = 19, D = 35 and D = 50, for four types of treatment carried out with encapsulated HSA (1), free HSA in PBS (2), free HSA mixed with empty vesicles (3), and empty vesicles (4);
- Figure 2 shows the anti-HSA isotypic response after immunization according to the IP/IV protocol defined in Example 3, expressed as the mean titers of IgG isotypes after the four types of treatment defined below (1, 2, 3 and 4);
- Figure 3 shows the results concerning the production of anti-HSA serum IgGs after immunization according to the SC protocol defined in Example 3, at D =
19, D = 35 and D = 50, for three types of treatment denoted respectively by 1 (with encapsulated HSA), 2 (with free HSA in PBS) and 3 (with empty vesicles); and - Figure 4 shows the anti-HSA isotypic response after three immunizations according to the SC protocol defined in Example 3, in the case of three treatments denoted by l, 2 and 3, with encapsulated HSA ( 1 ), free HSA in PBS (2), and empty vesicles (3).

5 Figures 5 and 6 are given with reference to Example 4 and show the results of assaying the B lymphocytes producing antibodies against NP of IgG isotypes compared with the total splenic cells after application of the two immunization protocols defined in Example 4, respectively comprising one IP immunization (Figure S) and three IP immunizations (Figure 6).
EXAMPLES
Example 1 Preparation and characterization of a composition containing encapsulated antieens A composition containing multilamellar vesicles and having the following composition, in which the proportions are expressed as percentages by weight:
phosphatidylcholine (PC90 from Natterman) 45.5 potassium oleate 5 cholesterol 5 cholesterol sulfate 2.5 laureth 4 (lauropal 4 from Witco) 2 aqueous solution containing 10 mg/ml of human serum albumin (HSA) 40 is prepared by the procedure described below.
The potassium oleate, the cholesterol, the cholesterol sulfate and the laureth 4 are first mixed by magnetic stirring at 55°C for 1 h. The resulting paste is cooled to room temperature and the aqueous solution of HSA (or pure water in the case of the empty microvesicles) and then the phosphatidylcholine are added slowly, with gentle stirring. The mixture is then introduced into a shearing cell consisting of two coaxial cylinders, one of which is fixed while the other rotates. This type of device is described in patent WO 93/19735. The shear of about 10 s 1 is maintained for 1 h. The sample is then dispersed in sterile water at a rate of 250 mg of paste per ml of water.
Before injection, a second dilution is carried out at a rate of 0.5 ml of the above solution in 2 ml of sterile water. For the samples containing an adjuvant, the latter is first incorporated into the dispersing water at a rate of 5 pl of aqueous solution of adjuvant (containing 20 pg/pl) in 1.995 ml of water. A
concentration of 20 ~g of antigen per 100 pl of injectable volume is therefore obtained in the injected solution and, in the case where the adjuvant is used, its final concentration is 4 pg per 100 p.l.
The degree of encapsulation is estimated by assay after separation of the vesicles from the supernatant, followed by release of the antigen. The dispersion of vesicles is ultracentrifuged (40,000 rpm, 1 h). The vesicle residue is diluted 10-fold with 10% sodium deoxycholate solution. The suspension of vesicles destroyed by the deoxycholate salt is subjected to an intense vortex effect and then ultra-centrifuged (50,000 rpm, 60 minutes). The antigen is assayed by spectrophoto-metry of the residue by Bradford's technique (Analytical Biochemistry, 1976, 72, p.
248) using the Bio-Rad protein assay in accordance with the protocol supplied (Bio-Rad Laboratories, USA).
The values associated with the residue, which correspond to the amount of encapsulated antigen (HSA), are 80% ~ 5%.
The mean size of the multilamellar vesicles containing the HSA is measured by the dynamic light scattering technique. A size of 0.218 ~m with a polydispersity of 20% is found, showing that the sample is homogeneous in respect of the vesicle size.
Example 2 Assay of the immune response a) Vaccination~rotocol The immune response to a microencapsulated antigen administered in vivo is determined using a murine model.
An accelerated vaccination protocol was set up to test the capacity of the vectors to elicit an immune response rapidly. Three injections were given over twenty days and the total immunoglobulins (IgG, IgM) were then assayed at two and four weeks.
Four different formulations (denoted by 1 to 4) containing the same amounts of antigens, and two control formulations (denoted by 5 and 6), were tested:
1 ) the antigen by itself;
2) the antigen associated with a known adjuvant (QS21 from AQUILA, USA);
3) the antigen encapsulated in the multilamellar microvesicles of surfactants according to Example 1;
4) the antigen encapsulated in the multilamellar microvesicles of surfactants according to Example 1, associated with QS21;
5) multilamellar microvesicles of surfactants analogous to those obtained according to Example 1, but empty;

6) multilamellar microvesicles of surfactants analogous to those obtained according to Example 1, but empty, associated with the adjuvant QS21.
All the formulations were prepared such that an injection volume of 100 ~1 contained 20 pg of antigen (free or encapsulated) and, when present, 4 ~g of adjuvant QS21 (cf. above).
The murine model is that of 8-week-old female BALB/c mice with a mean weight of 20 g. Ten days elapse between the first and second intraperitoneal injections. The third and final administration is intravenous and takes place 10 days later. The protocol therefore extends over a total of 20 days. The same protocol is followed for the six formulations corresponding to the six groups of four mice, numbered from 1 to 6. Each group was treated with the formulation of the same number.
The animals are kept under "normal" conditions, i.e. fed ad libitum and hydrated normally.
Two blood samples (retro-orbital) are taken at 2 and 4 weeks after the last injection in order to assay the total immunoglobulins (IgG, IgM) in the serum after centrifugation of the heparin-treated whole blood.
b) Result: asst of the total antibod,~response The total antibody response is assayed by titrating the immunoglobulins by means of an ELISA (Saunders G.C. "Immunoassays in the clinical laboratory", ed.
by Nakamura et al., 1979) with the aid of rabbit immunoglobulins directed against mouse immunoglobulins (Institut Pasteur, Paris), labeled with peroxidase. The fixed immunoglobulins are disclosed by adding the enzyme substrate.
The measurement made in this test is conventionally an absorbance measurement at 405 nm, the enzymatic reaction product formed in the ELISA
absorbing light at this wavelength. The analytical method used gives identical results to measurement of the absorbance at half the saturation height, but with greater precision for the low values, and proceeds via a method which provides access to a quantity proportional to the amount of antigens present in the serum by parameterization of the curves expressing the absorbance as a function of dilution.
In fact, the dilution law can be represented by the following model equation:
IB
I(D) = Io +
( 1 + CoD) where:
I(D) is the measured absorbance at dilution D, Io is the background noise of the measurement, I$ is the measurement at saturation (plateau), Co is proportional to the desired antibody concentration (Co has the dimensions of a concentration).
Io is measured directly. IS is obtained from an experiment in which the response is suffcient to reach saturation. The value obtained is then used in the parameterization of the other measurements in the same series of experiments, even if saturation is never reached for the latter.
The parameterization of the curves makes it possible for each experiment to deduce Co, i.e. a number proportional to the amount of antibodies present in the serum studied.
This method is more discriminating than simple measurement of the dilution for obtaining the half plateau, since it enables a measurement to be obtained even in the absence of saturation. It is therefore particularly suitable for weak responses.
The results at two weeks and four weeks (after the end of the vaccination protocol) are given for all the preparations in Table 1 below.
The values obtained by this method are indicated in Table 1 below and are therefore proportional to the level of immunoglobulins present in the blood.

Mouse group and 1 2 3 4 5 6 formulation no.

Res onse at 2 - - 145 335 - -weeks - ____.

Response at 4 _ I 7 I 467 8~4 ~ -weeks ~ I

The symbol "-" corresponds to a response below 2, which is considered insignificant.
It is therefore seen that, as from 2 weeks, a significant immune response is obtained for the encapsulated formulations (with or without adjuvant). At the end of four weeks, the encapsulated formulations give a very substantial response which is more than 50 times greater than the response to the antigen associated with the adjuvant; it is more than 100 times greater when the adjuvant QS21 is added to the encapsulated antigen.
Example 3: Influence of the route of immunization Other immunization experiments carried out with HSA on BALB/c mice were conducted according to 2 immunization protocols, hereafter called IP/IV
and SC, comprising the following treatments:
- IP/IV protocol: 2 intraperitoneal (IP) injections followed by one intravenous (IV) injection, - SC protocol: 3 subcutaneous (SC) injections.
The formulations of the vesicles were identical to those of Example 1. In all cases, the dose of HSA was 20 ~g per 100 ~1 injection. All the injections were days apart, i.e. one injection at D = 0, D = 10 and D = 20. The samples (retro-orbital) were taken at D = 19, i.e. one day before the last injection, and at D = 35 andD=50.
Groups each containing 4 mice were subjected to the following treatments:
- For the IP/IV protocol:
HSA by itself, dissolved in PBS (control) HSA encapsulated in the vesicles according to the invention at a rate of 20 pg of protein in 5 mg of vesicles dispersed in PBS to a total volume of 100 ~tl HSA dissolved in PBS to which empty vesicles (20 ~g of protein and 5 mg of empty vesicles dispersed in PBS to 100 p.l) have been added r Empty vesicles (5 mg of vesicles dispersed in PBS to 100 ~1) - For the SC protocol:
~ HSA by itself, dissolved in PBS (control) r HSA encapsulated in the vesicles according to the invention at a rate of 20 ~g of protein in 5 mg of vesicles dispersed in PBS to a total volume of 100 ~l Empty vesicles (5 mg of vesicles dispersed in PBS to 100 ~1) The antibodies were assayed by ELISA on the total immunoglobulins (IgG
H+L) at D = 19, D = 35 and D = 50 and on the IgG~, IgG~, IgG26 and IgG3 isotypes at D = 50.
The results are given in Figures 1 to 4, which show the mean titers (mean of the assays of the individual sera). The following observations are made:
1) A zero antibody response when the non-encapsulated free HSA is used, and 5 with the empty vesicles.
2) A substantial production of total immunoglobulins, by both routes of administration, when the HSA is encapsulated, with a maximum at D = 35.
3) A strong amplification of the total IgG (H+L) response with the encapsulated HSA administered subcutaneously.
10 The results are shown in Table 2 below.

D=19 D=35 D=50 IP/IV rotocol 700 1700 1600 SC rotocol 10,000 35,000 40,000 15 4) No adjuvant effect of the empty vesicles because the free HSA in the presence of the empty vesicles elicits only a very weak response.
S) A dominant production of IgG,, the presence of IgG2a in both protocols with much greater titers in the case of subcutaneous administration, and the presence of IgG2b. No production of IgGz could be observed, whatever the administration protocol.
This Example completes the results of Example 2 and demonstrates that the encapsulation of an antigen in the multilamellar vesicles according to the invention amplifies the immune response for different types of injection. They further indicate that the eE~fect of the vesicles is not an adjuvant effect, since the addition of empty vesicles to the non-encapsulated antigen produces no amplification of the response.
Example 4: Experiments with the NP of measles Immunization experiments with the NP of measles (internal protein of the measles virus) were carried out with 2 immunization protocols involving intraperitoneal (IP) injection:
a single injection (the results are shown in Figure 5) three injections at 7-day intervals (the results are given in Figure 6) each with a 30 ~g dose of NP, on female BALB/c mice.
The vesicles were prepared according to the formulation and procedure of Example 1 with a 30 ~g dose of protein in 75 mg of vesicles dispersed in PBS
(300 ~1 injections containing 30 ~g of protein).
The empty vesicle injections contained a 75 mg dose of vesicles in a final volume of 300 p.l (PBS).
For each protocol and each dose, groups of 4 mice were treated with:
NP dissolved in PBS (I) NP encapsulated in the vesicles according to the invention (II) r NP dissolved in PBS to which empty vesicles have been added (III) r NP in PBS emulsified with FREUND's complete adjuvant. The mixture is prepared for immediate use, just before injection, at a rate of 1 volume of protein-containing buffer solution to 1 volume of FCA (SIGMA, St Louis USA) (IV).
The assays were performed, 7 days after the last immunization, by means of ELISPOT on IgG. This assay consists in counting the B lymphocytes producing antibodies against NP of IgG isotypes compared with the total splenic cells.
The result is therefore expressed as the ratio r of the number of cells producing the specific antibodies to the total number of cells observed.
1) Protocol for 1 immunization very slight improvement in the response with encapsulated NP (r = 500 x 10-') compared with NP by itself (r = 400 x 10-') no amplifying effect with empty vesicles associated with non-encapsulated NP
(r = 400 x 10-') 2) Protocol for 3 immunizations r encapsulated NP ve, c~ learl~perior (r = 900 x 10~') to NP by itself (r =
250 x 10-') and comparable to NP with FREUND's complete adjuvant (r = 900 x 10-') r free NP associated with empty vesicles similar to NP by itself (r = 300 x 10-') This Example completely confirms the results of Example 3 on a different model and indicates that the amplification factors of the response obtained by encapsulation in the multilamellar vesicles according to the invention are comparable to those obtained with the most potent adjuvants (FREUND's adjuvant is particularly effective but its toxicity is such that it is not authorized in formulations for vaccination or therapeutic purposes).

Claims

1. Antigen vector, characterized in that it consists of a dispersion of multilamellar vesicles within which said antigen is at least partially encapsulated, said vesicles consisting of bilayers comprising at least one surfactant, said bilayers being concentric and giving said vesicles an onion structure.

2. Vector according to claim 1, characterized in that said vesicles have a diameter below 20 µm, preferably below 10 µm, particularly preferably below 1 µm and very particularly preferably of between 0.1 and 1 µm.

3. Vector according to claim 2, characterized in that said surfactants constituting said bilayers are selected from the group consisting of:
- hydrogenated or non-hydrogenated phospholipids, - linear or branched, saturated or mono- or polyunsaturated C6 to C18 fatty acids in the form of the acid or an alkali metal, alkaline earth metal or amine salt, - ethoxylated or non-ethoxylated esters of these same fatty acids with . sucrose, . sorbitan, . mannitol, . glycerol or polyglycerol, or . glycol, - mono-, di- or triglycerides or mixtures of glycerides of these same fatty acids, - ethoxylated or non-ethoxylated, linear or branched, saturated or mono- or polyunsaturated C6 to C18 fatty alcohols, - ethoxylated or non-ethoxylated ethers of these same fatty alcohols with . sucrose, . sorbitan, . mannitol, . glycerol or polyglycerol, or . glycol, - hydrogenated or non-hydrogenated, polyethoxylated vegetable oils, - polyoxyethylene and polyoxypropylene block polymers (poloxamers), - polyethylene glycol hydroxystearate, and - alcohols with a sterol skeleton, such as cholesterol and sitosterol.

4. Vector according to one of claims 1 to 3, characterized in that the bilayers of said vesicles comprise at least two surfactants, one of which has a hydrophilic-lipophilic balance (HLB) of between 1 and 6, preferably of between 1 and 4, while the other has a hydrophilic-lipophilic balance (HLB) of between 3 and 15, preferably of between 5 and 15.

5. Vector according to one of claims 1 to 4, characterized in that the encapsulation yield is greater than 50%, preferably greater than 80%.

6. Vector according to one of claims 1 to 5, characterized in that said antigen is a protein or peptide antigen.

7. Vector according to claim 6, characterized in that said protein antigen consists of a natural or recombinant protein.

8. Composition containing at least one antigen against which it is desired to produce specific antibodies or a specific cellular response, characterized in that said antigen(s) is (are) at least partially encapsulated in the multilamellar vesicles of a vector as defined according to one of claims 1 to 7.

9. Composition according to claim 8, characterized in that it is a pharmaceutical composition, particularly a vaccine.

10. Composition according to claim 8, characterized in that it is a composition for the production of monoclonal antibodies.

11. Process for the preparation of a composition according to one of claims 8 to 10, characterized in that it comprises the following steps:
- preparation of a lamellar liquid-crystal phase comprising at least one surfactant and incorporating the antigen(s), - transformation of the liquid-crystal phase into onion-structured multi-lamellar vesicles by shear, and - dispersion in an aqueous solution to give the desired concentration of antigen(s).

12. Process according to claim 11, characterized in that said shear is homogeneous shear.

13. Process according to claim 11 or 12, characterized in that at least 2 surfactants are used, one of which has an HLB of between 1 and 6, preferably of 1 and 4, while the other has an HLB of between 3 and 15, preferably of between 5 and 15.

14. Process according to one of claims 11 to 13, characterized in that said antigen is a protein or peptide antigen.