Introduction

Adjuvants have been used by immunologists for many years to stimulate immune responses. A wide variety of adjuvants exist including mineral salts, immunoenhancing cytokines, bacterial products and other natural products such as Quil A^{1, 2}. Adjuvants have become increasingly important in the area of vaccine design with the move away from whole pathogen vaccines to subunit vaccines. This shift in vaccine design has in part been caused by the increasing demands of vaccine safety and while the peptide and protein subunit vaccines are indeed safer they are much less immunogenic^{3, 4}. The search is now on for vaccine adjuvants which can be used to stimulate protective immune responses and which can be safely used in human vaccines.

Another way of increasing the immunogenicity of subunit vaccines is to incorporate the vaccine antigen into a particulate delivery system. This increases the effectiveness of the vaccine by increasing the amount of antigen which can be delivered to the cells of the immune system and also provides protection for the antigen⁵. Specialized delivery systems can also allow for mucosal vaccine delivery. Liposomes are a commonly used delivery system for both medicines and vaccines⁶. Liposomes are enclosed bilayers made from a variety of phospholipids and cholesterol. They are a very versatile delivery system as they can deliver a wide variety of medicines or vaccine antigens. Hydrophilic molecules can be easily incorporated into the aqueous centre of the particles while lipophilic molecules can be incorporated into the lipid bilayers. The surface of the liposomes can also be modified to include targeting molecules^{7, 8, 9}.

Problems with liposomes include their inherent physical instability and lack of immunogenicity, although the lack of immunogenicity may be an asset for drug delivery. Modifications can be made to increase stability of liposomes, for example, inclusion of cholesterol and polyethylene glycol, creating sterically stabilized or stealth liposomes, increases systemic circulation and stability⁵. The lack of immunogenicity can generally be overcome with the inclusion of a range of adjuvants into liposomes. Cytokines and chemokines have been included by a number of groups, and stimulatory molecules such as CD40 have also been included^{10, 11, 12, 13}. The adjuvant Quil A is another compound which can be used to increase the immunogenicity of liposomes. Quil A is a complex mixture of chemically related triterpenoid saponins extracted from the bark of the Chilean tree Quillaja saponaria Molina¹⁴. The saponins are typically approximately 20% of the dry weight of the bark extract. Following extraction, the saponins are purified from the tannins and polyphenolics by dialysis, gel diffusion or filtration giving a largely decolourized extract enriched in adjuvant and haemolytic activity¹⁴. The extract remains a complex mixture of at least 20 chemically related but different saponins.

The adjuvant activity of Quil A is thought to be mediated through an aldehyde group on the saponins' triterpene aglycone forming a Schiff base with amino groups on costimulatory T cell receptors¹⁴. Quil A may potentially be able to replace B7.1 as a costimulatory signal and preferentially induce Th1 type immune responses¹⁵. Chemical modification of the aldehyde groups completely ablates adjuvant activity. The prototypical Quil A containing delivery systems are immune stimulating complexes (ISCOMs). These have been used widely in animals to stimulate immune responses to a variety of antigens but to our knowledge they are not yet licensed for use in humans.

Of particular interest to us is that while protein adjuvants can be easily incorporated into liposomes with little or no effect on structure, the inclusion of Quil A in liposomes alters the physicochemical properties of the particles, as well as potentially affecting their immunogenicity. Inclusion of Quil A disrupts the lipid bilayer structure of the liposomes causing a variety of particles to be formed depending on the ratios of phospholipids, cholesterol and Quil A used. In previous studies we have constructed pseudo-ternary phase diagrams to characterize the structures produced^{16, 17, 18}. A number of different colloidal structures are produced including classical ISCOM-like particles, ring-like micelles, worm-like micelles, lamellae (hexagonal arrays of ring-like micelles) and lipidic/layered structures. These particles vary in size, morphology and stability. For this study we have chosen two of the predominant structures from the phase diagram, ISCOM-like particles and ring-like micelles and have examined the antigen loading capacity of the structures and their immunogenicity in vitro.

Materials and Methods

Mice

C57Bl/6 J and OT-I mice were bred and maintained in microisolators under specific pathogen-free (SPF) conditions at the Hercus Tairei Resource Unit. All experiments were approved by the University of Otago Animal Ethics Committee.

Antibodies

Anti-CD11c, anti-CD86, anti-CD40, anti-CD25, anti-V2 and anti-V5.1,5.2 were purchased from BD PharMingen (San Diego, CA, USA) and were used according to the manufacturer's instructions.

Preparation of PE-fluorescein isothiocyanate-OVA

Ovalbumin (OVA; Grade V, purity approx. 98%, Sigma-Aldrich Pty Ltd, USA) was conjugated to fluorescein isothiocyanate (fitc; Sigma-Aldrich, USA) and phosphatidylethanolamine (PE; Sigma-Aldrich) using a previously described method^{16, 19}. Briefly, 100 mg of OVA was added to a solution of 20 mg of fitc in 10 mL of 20 mmol/L carbonate buffer (pH 9.5). The solution was incubated with continuous stirring at 4°C for 18 h in the dark. The fitc-OVA was purified by repetitive ultrafiltration and washing with Milli-Q water. The preparation was freeze-dried overnight and recovery quantified by gravimetry. The fitc-OVA was then conjugated to PE by linking the amino group of PE to the carboxyl groups of fitc-OVA. To achieve this, 100 mg of PE dissolved in 20 mL of 10% w/v octylglucoside (Sigma-Aldrich), 100 mg of fitc-OVA dissolved in 10 mL of Milli-Q water, 2 g of 1-ethyl-3'-(3-dimethylaminopropyl) carbodiimide hydrochloride and 100 mg of N-hydroxysuccinimide (Sigma-Aldrich Pty Ltd, USA) were added together to a total volume of 80 mL of Milli-Q water. The solution was stirred for 30 h at 4°C and the resulting PE-fitc-OVA was purified by repetitive ultrafiltration and washing with Milli-Q water. The preparation was freeze-dried and recovery quantified by gravimetry.

Preparation of formulations

All formulations were prepared by the hydration method^{18, 20}. Three formulations were made containing 0%, 20% or 70% Quil A (Figure 1). The PE and cholesterol (Sigma-Aldrich) content of the formulations was kept constant allowing the examination of the effect of including Quil A in liposomes on structure and immunogenicity. PE and cholesterol (mass ratio of 3:1) were dissolved in chloroform and a thin lipid film formed by evaporation at 45°C for 1 h under vacuum on a rotary evaporator. Varying amounts of Quil A (Superfos Biosector, Frederikssund, Denmark) (0%, 20% or 70% of the total lipid weight of 6.7 mg) and 1 mg of PE-fitc-OVA dissolved in 1 mL of PBS were added to the dried lipid films of PE and cholesterol. Formulations were stirred for 3 h at 4°C to ensure complete hydration. Formulations were then freeze-dried overnight followed by rehydration in 1 mL PBS. Particles formed were purified by sucrose density gradient ultracentrifugation (200 000 g for 18 h, 10°C)²¹. Particle morphology in each formulation was examined by transmission electron microscopy (Phillips CM100) of negatively stained samples using 2% phosphotungstic acid, pH 5.2 as a contrasting agent. Antigen incorporation was determined fluorometrically as previously described by comparison to a PE-fitc-OVA calibration curve^{19, 22}.

Figure 1.

Pseudo-ternary phase diagram of mixtures of Quil A, cholesterol (CH) and phosphatidylethanolamine (PE) in PBS, pH 7.4 characterized within 1 day of preparation (adapted from¹⁷). Areas rich in liposomes (), ISCOM-like particles () and ring-like micelles () are indicated. Three formulations were made containing either 0% (A), 20% (B) or 70% (C) Quil A. The phospholipid and cholesterol content of the formulations were kept constant (mass ratio 3:1).

Full figure and legend (35K)

We have previously shown that varying the amount of Quil A in colloidal formulations affects antigen incorporation^{16, 22}. Therefore two formulations were prepared for the 0%, 20% and 70% Quil A containing preparations; one containing PE-fitc-OVA and one without antigen. After measuring the relative antigen incorporation rates, the amounts of empty and antigen-containing formulations were adjusted to give preparations containing equal amounts of PE-fitc-OVA and total lipid.

Generation and activation of bone marrow dendritic cells

Bone marrow dendritic cells (BMDC) were generated by culturing bone marrow derived stem cells from C57Bl/6 J mice in complete Iscove's Modified Dulbecco's Medium (cIMDM; IMDM supplemented with 5% foetal bovine serum, 1% penicillin/streptomycin, 1% glutamax and 0.01% 2-mercaptoethanol (all from Invitrogen (Carlsbad, CA, USA)) with 20 ng/mL recombinant granulocyte/macrophage colony stimulating factor (clone kindly supplied by Dr G. Buchan, University of Otago) for 6 days at 37°C, 5% CO₂. The BMDC were then pulsed with titrated amounts of the formulations.

BMDC viability after the 48 h pulse was assessed by Trypan blue (Invitrogen, USA) exclusion and the percentage of live cells calculated. Uptake of the formulations by CD11c^+ve cells was measured by flow cytometry (FACScalibur, Becton Dickinson, Franklin Lakes, NJ, USA) 48 h after the addition of the various formulations. Cells were stained with antibodies and propidium iodide (PharMingen) according to the manufacturer's instructions. Data was analysed using CellQuest Pro (Becton Dickinson) and the fold increase in fitc positive CD11c^+ve cells calculated by dividing the percentage CD11c^+ve cells fitc positive after incubation with the formulations by the percentage CD11c^+ve cells fitc positive in the negative control (no formulation). The fold increase in mean fluorescence intensity (MFI) was similarly calculated by dividing the MFI of CD11c^+ve cells after incubation with the formulations by the MFI of CD11c^+ve cells in the negative control (no formulation). Activation of BMDC was measured by flow cytometric analysis of the activation markers CD86 and CD40 on propidium iodide^–ve, CD11c^+ve cells 48 h after the addition of the various formulations. The fold increase in CD86^hi or CD40^hi BMDC was calculated by dividing the percentage CD86^hi or CD40^hi, propidium iodide^–ve, CD11c^+ve cells activated by each formulation by the percentage CD86^hi or CD40^hi, propidium iodide^–ve, CD11c^+ve cells in the negative control.

T-cell proliferation and activation

To examine the ability of the formulations to stimulate antigen specific lymphocyte proliferation and activation, BMDC pulsed with the various formulations for 48 h were washed to remove excess formulation and incubated with OT-I splenocytes stained with carboxyfluorescein succinimidyl ester (CFSE; Molecular Probes, Eugene OR, USA) for 72 h. CFSE staining and CD25 expression on OT-1 transgenic T cells were measured by flow cytometry. Data was analysed using CellQuest Pro. The fold increase in proliferation was calculated by dividing the percentage proliferated OT-1 transgenic T cells incubated with formulation-pulsed-BMDC by the percentage proliferated OT-1 transgenic T cells in the negative control (T cells incubated with BMDC not pulsed with any formulation). The fold increase in CD25^hi T cells was calculated by dividing the percentage CD25^+ve OT-1 transgenic T cells incubated with formulation-pulsed-BMDC by the percentage CD25^+ve OT-1 transgenic T cells in the negative control (T cells incubated with BMDC not pulsed with any formulation).

Results

Addition of Quil A destroys the lipid bilayer structure of liposomes and produces distinct colloidal particles

Previous work has been carried out in our laboratory characterizing the pseudo-ternary phase diagram shown in Figure 1¹⁷. From this initial work, three formulations were chosen to investigate the effect of the addition of the adjuvant Quil A to liposomes on colloidal structure and immunogenicity. The formulations chosen contained 0%, 20% or 70% Quil A (percentage of the total lipid, i.e. PE, cholesterol and Quil A). In the 0% Quil A formulation (the liposome control), the PE and cholesterol formed lipidic particles with a wall of several bilayers enclosing an aqueous core. The size and lamellarity of the liposomes varied greatly but their size was usually well above 100 nm in diameter (Figure 2A). The addition of 20% Quil A disrupted the bilayer and regular spherical cage-like structures of approximately 40 nm in diameter were formed (Figure 2B). These particles represent the classical ISCOM^{6, 18}. With the addition of 70% Quil A, the structures were broken down further and small ring-like micelles of 10–12 nm in diameter were formed (Figure 2C).

Figure 2.

Transmission electron micrographs of formulations containing (A) 0%, (B) 20% and (C) 70% Quil A.

Full figure and legend (418K)

Incorporation of antigen into the formulations decreases with increasing Quil A concentration

Of key importance to any particulate delivery system is the amount of antigen that can be loaded into or onto the particles. This is important for the efficacy of the product in vivo but is also an important practical consideration as formulations with low antigen incorporation will be expensive to produce. An advantage of liposomes is the wide variety of antigens that can be incorporated into them. Hydrophilic antigens can be loaded into the aqueous core while lipophilic antigens can be incorporated into the lipid bilayer. However, as the addition of Quil A to liposomes destroys the lipid bilayer there is no longer an enclosed aqueous core in which to load hydrophilic antigens. Therefore, in order for an antigen to be incorporated into Quil A containing formulations, the antigen must have a lipophilic domain in its chemical structure. For this reason the model antigen we used was OVA conjugated to PE. A fluorescent tag (fitc) was also added to facilitate uptake studies. In all formulations, incorporation of the PE-fitc-OVA was less than 30% (Figure 3). Antigen incorporation decreased significantly with the addition of Quil A (26% incorporation with 0% Quil A down to 18% incorporation with 20% Quil A). The amount of Quil A added also had an affect on incorporation with PE-fitc-OVA incorporation decreasing with increasing Quil A concentration (18% incorporation with 20% Quil A down to 7% incorporation with 70% Quil A).

Figure 3.

Incorporation of PE-fluorescein isothiocyanate (fitc)-OVA into colloidal formulations. Incorporation of PE-fitc-OVA into formulations containing 0%, 20% and 70% Quil A was measured by fluorometry. Data shown are the mean and standard deviations of 3 experiments. P-values were derived from unpaired, two-sided t-tests.

Full figure and legend (14K)

Uptake of formulations by BMDC decreases with increasing Quil A concentration

As shown in Figure 3, incorporation of antigen into the formulation was not the same for all formulations. However, for the studies examining uptake and activation we required that the same amounts of lipid and PE-fitc-OVA were delivered to the cells. In order to do this, two preparations of each formulation were made, one containing PE-fitc-OVA and one containing no antigen. Using the incorporation data the preparations were mixed to give a stock containing 100 g/mL of lipid and 1 g/mL PE-fitc-OVA. The uptake of antigen by BMDC could then be directly compared (Figure 4). As expected, the number of cells taking up PE-fitc-OVA increased with the addition of increasing amounts of formulation (Figure 4A). The formulation containing no Quil A was taken up by significantly more CD11c^+ve cells than the Quil A containing formulations. This was even more impressive considering over half of the 0% Quil A formulation contained no antigen. Examination of uptake of the Quil A containing formulations showed that increasing the amount of Quil A from 20% to 70% further reduced the number of cells able to take up the formulation.

Figure 4.

Uptake of PE-fitc-OVA colloidal formulations by bone marrow dendritic cells (BMDC). Formulations containing the same amount of PE-fitc-OVA, but either 0% (), 20% () or 70% () Quil A were incubated with BMDC for 48 h. Fitc fluorescence of CD11c positive cells was measured by flow cytometry. The fold increase in fitc^+ve CD11c cells (A) and the fold increase in the mean fluorescence intensity (MFI) of the fitc^+ve cells (B) is shown. Results are expressed as the mean and SD of three independent experiments. †, P < 0.01 vs cells pulsed with 20% Quil A containing formulations. ‡, P < 0.01 vs cells pulsed with 20% or 70% Quil A containing formulations. P-values were derived from unpaired, two-sided t-tests.

Full figure and legend (6K)

The amount of antigen taken up by CD11c^+ve cells was examined by measuring the MFI of fitc^+ve cells. Again it was the 0% Quil A containing formulation that had the highest uptake (Figure 4B). Interestingly, with the 20% Quil A containing formulation, although there was a 25-fold increase in the number of BMDC taking up antigen (Figure 4A), the BMDC took up significantly lower amounts of 20% Quil A containing formulation as compared with the 0% Quil A containing formulation. This data shows that the number of cells taking up antigen and the amount of antigen taken up by the cells can vary considerably in response to the composition of the formulation.

Viability of BMDC decreases with increasing Quil A concentration

Quil A has been reported to be haemolytic both in vitro and in vivo, however, haemolytic activity is reduced upon incorporation into colloidal particles^{23, 24}. In response to this, the viability of the cells in culture after 48 h incubation with the various formulations was examined (Figure 5). The viability of cells cultured for 48 h without antigen was greater than 90%. Addition of the 0% Quil A formulation at concentrations ranging from 1 to 50 g/mL resulted in a modest decrease in viability (greater than 80% viable cells at 50 g/mL) indicating that the cholesterol and PE were not themselves toxic. However, the addition of Quil A to the liposomes resulted in a significant concentration-dependent decrease in viability. At lipid concentrations of 50 g/mL the viability of cells pulsed with the 20% Quil A formulation was only 30%. Increasing the Quil A concentration from 20% to 70% did not result in a further decrease in viability (35% viability at 50 g/mL lipid).

Figure 5.

Viability of BMDC following incubation with colloidal formulations containing Quil A. Formulations containing the same amount of PE-fitc-OVA, but either 0% (), 20% () or 70% () Quil A, were incubated with BMDC for 48 h. Viability of BMDC was determined by staining cells with Trypan blue. Results are expressed as the mean and SD of three independent experiments. †, P < 0.01 vs cells pulsed with 20% Quil A containing formulations. ‡, P < 0.01 vs cells pulsed with 20% or 70% Quil A containing formulations. P-values were derived from unpaired, two-sided t-tests.

Full figure and legend (11K)

Activation of BMDC increases following incubation with Quil A containing formulations

As shown above, incubation of BMDC with Quil A containing formulations resulted in substantial cell death. Therefore, analysis of DC activation was carried out on live cells only, using propidium iodide staining to exclude dead and dying cells. Analysis of the expression of two DC activation markers, CD86 and CD40, showed that incubation of BMDC with the 0% Quil A formulation resulted in no increase in the number of CD86^hi (Figure 6A) or CD40^hi (Figure 6B) CD11c^+ve cells. Incubation of BMDC with formulations containing 20% Quil A resulted in a significant increase in CD86^hi and CD40^hi CD11c^+ve cells. Increasing adjuvant levels up to 70% did not further increase activation and instead induced lower numbers of CD86^+ve and CD40^+ve CD11c^+ve cells than did formulations containing 20% Quil A.

Figure 6.

BMDC activation following incubation with colloidal formulations containing either 0% (), 20% () or 70% () Quil . The fold increase in live CD11c positive cells expressing the activation markers CD86 (A) and CD40 (B) following incubation of the BMDC with formulations for 48 h is shown. Results are expressed as the mean and SD of three independent experiments. †, P < 0.05 vs cells pulsed with 0% Quil A containing formulations. ‡, P < 0.01 vs cells pulsed with 0% or 70% Quil A containing formulations. P-values were derived from unpaired, two-sided t-tests.

Full figure and legend (6K)

Activation and proliferation of T cells increases following incubation with Quil A containing formulations

Splenocytes from OT-I transgenic mice were incubated with DC pulsed with formulations containing 0%, 20% or 70% Quil A. Splenocytes were stained with CFSE and antigen specific proliferation was assessed by flow cytometry. Incubation of OVA transgenic T cells with DC pulsed with formulations containing no Quil A did not result in any antigen specific proliferation (Figure 7A). Similarly, there was no increase in the number of CD25^+ve transgenic T cells (Figure 7B). However incubation of transgenic T cells with DC pulsed with formulations containing either 20% or 70% Quil A did result in significant increases in T-cell proliferation and in the number of cells expressing CD25.

Figure 7.

T-cell proliferation and activation following incubation with colloidal formulations containing either 0% (), 20% () or 70% () Quil A. BMDC pulsed with the various formulations for 48 h were washed and then incubated with 5,6-carboxyfluorescein diacetate succinimidyl ester (CFSE)-stained OT-I splenocytes for 72 h. Proliferation and CD25 expression were measured by flow cytometry. (A) Fold increase in proliferation of CFSE stained OT-I T cells. (B) Fold increase in CD25^+ve OT-I T cells. Results are expressed as the mean and SD of three independent experiments. †, P < 0.05 vs cells pulsed with 20% Quil A containing formulations. ‡, P < 0.05 vs cells pulsed with 20% or 70% Quil A containing formulations. P-values were derived from unpaired, two-sided t-tests.

Full figure and legend (23K)

Discussion

The aim of this study was to investigate the effect of incorporation of the adjuvant Quil A on the physical structure and immune stimulatory capacity of liposomes. Specifically, we were interested in determining if increasing the amount of the adjuvant Quil A incorporated in the formulation from 20 to 70% would further boost immune responsiveness. These concentrations of Quil A were chosen based on the pseudo-ternary phase diagram established for phospholipids, cholesterol and Quil A¹⁸. Using the hydration method, the addition of 20% Quil A to phospholipid and cholesterol (ratio 3:1) causes the formation of predominantly ISCOM-like structures. Quil A disrupts the enclosed bilayer of the liposome forming regular cage-like structures of approximately 40 nm in diameter. When the Quil A content is increased to 70% ISCOM-like structures are replaced by small (about 10 nm) ring-like micelles. We have shown previously that increasing Quil A increases the hydrophilicity of the colloidal structures²². Therefore, in formulations containing 70% Quil A, ring-like micelles do not aggregate to form ISCOM structures or lamellae (hexagonal array of ring-like micelles) but instead disperse in the aqueous phase as individual particles^{17, 18}.

We next investigated the effect of the increase in percentage Quil A in the formulation on antigen incorporation into the resulting colloidal structures. A key feature of any delivery system is antigen incorporation. Peptide and protein antigens are commonly expensive to produce therefore it is important that antigen incorporation is efficient and does not affect the immunogenicity of the antigen. For mass manufacturing, formation of the particles and antigen incorporation should also be simple. Formation of the Quil A containing liposomes using the hydration method is indeed a simple process. However, antigen incorporation becomes more difficult with the addition of Quil A to liposomes as there is no longer an enclosed aqueous core in which hydrophilic protein and peptide antigens can be entrapped and therefore incorporation rates of hydrophilic proteins such as ovalbumin are low¹⁶. Antigens which have a lipophilic domain can, however, be incorporated into these particles. Virus coat proteins, or any antigen that is naturally expressed on the surface of a cell, will fulfil these requirements and can be easily incorporated into ISCOMs^{25, 26}. The simplest way to incorporate proteins or peptide antigens into Quil A containing formulations is to attach a lipid tail to the antigen^{21, 22, 27}. In this study we attached phosphatidylethanolamine to ovalbumin by a conjugation reaction and investigated antigen incorporation using a fluorescent tag also linked to the ovalbumin. The antigen incorporation rate for all particles was less than 30%. The addition of Quil A to the liposomes decreased incorporation, as did increasing the amount of Quil A from 20% to 70%. The decreased incorporation into ISCOM-like particles as compared to liposomes can be attributed at least partially to the loss of the enclosed aqueous core. The further decrease in incorporation with the increase in Quil A from 20% to 70% can be attributed to the increased hydrophilicity of the ring-like micelles. This will decrease the incorporation of the amphiphilic PE-fitc-OVA antigen. The incorporation of PE-fitc-OVA did not change the morphology or size of the colloidal particles in agreement with previous results¹⁶.

The ability of murine BMDC to take up the different particles was then examined. Parameters investigated were the percentage of cells taking up antigen, the amount of antigen taken up and the ability of BMDC to take up the different types of colloidal particles. To allow for the different amount of antigen in the particles and corresponding differences in fluorescence, cells were given a mixture of formulations containing antigen and empty formulations. This meant the BMDC were all pulsed with the same amount of antigen and lipid. Uptake of the particles varied considerably with the liposomes having the highest uptake and the ring-like micelles having the lowest uptake. Again, this most likely reflects changes in the lipid composition of the particles which lead to increased hydrophilicity and increasingly negative zeta potential upon increasing Quil A concentration²³. The liposomes are the least hydrophilic particles and can interact intimately with the cellular membrane whereas with increasing Quil A concentration the particles become more hydrophilic^{18, 23} and may not interact so well with the cell membranes. Liposomes also have a neutral charge, but with increasing Quil A concentration as stated in the preceding pages, the colloidal particles will develop a stronger negative charge due to presence of glucuronic acid in the Quil A molecule^{23, 28} and may therefore not be so easily taken up by the cells that also have a negative surface charge. Analysis of the fluorescence intensity of cells that had taken up antigen showed that incubation of BMDC with liposomes results in the highest amounts of antigen being taken up into cells. This is consistent with the large size of the liposomes and the higher antigen incorporation. This demonstrates one of the advantages of using a delivery system with a larger particle size instead of smaller particles such as ring-like micelles that would need to be taken up at a much higher rate to load the same amount of antigen into a cell.

As well as having adjuvant activity, saponins are also haemolytic. Therefore, it was important to determine what effect increasing amounts of Quil A in the particles had on cell viability as well as on the ability to stimulate an immune response. Not unexpectedly, we found that particles containing Quil A were more toxic to BMDC in cell culture than were particles containing only phospholipid and cholesterol. However, increasing the amount of Quil A from 20% to 70% did not result in a further decrease in cell viability. In fact the 70% Quil A containing particles were, if anything, less cytotoxic than the 20% Quil A containing particles. These differences may be at least partially explained by the decreased uptake of the 70% Quil A containing particles.

We next examined the immune stimulatory ability of the particles by examining the ability of the particles to activate BMDC and the subsequent capacity of the formulation pulsed BMDC to induce the proliferation and activation of transgenic CD8 T cells. In order to stimulate a protective acquired immune response, the vaccine (antigen plus delivery system) must be able to activate the antigen presenting cell but additionally the antigen must be presented on MHC molecules and in sufficient quantities to activate T cells. Dendritic cell (DC) activation without exogenous antigen presentation will not stimulate immune responses and indeed may induce autoimmunity and exogenous antigen presentation without DC activation will similarly not stimulate an immune response and may induce tolerance^{29, 30}. Therefore, it is important for safety as well as efficacy that the vaccine is able to activate the APC and introduce sufficient antigen into the cell in such a way that it is targeted to the appropriate antigen processing and presentation pathway. In this experimental system, we choose to look at MHC class I antigen presentation and the activation of CD8 T cells. This type of response is more difficult to stimulate than the MHC class II, CD4 T-cell-mediated immune response as exogenous antigen delivered in a non-living delivery system is generally targeted to the MHC class II antigen processing and presentation pathway. Other researchers have demonstrated that Quil A containing particles can in fact access the MHC class I presentation pathway and activate CD8 T cells³¹.

In this study, we first examined the ability of Quil A containing particles to induce the upregulation of the costimulatory molecules CD86 and CD40. The addition of Quil A to the delivery system resulted in an increase in the number of BMDC expressing both of these molecules. The particles containing 20% Quil A induced higher levels of activation than did the particles containing 70% Quil A. This reduced activation by the particles containing 70% Quil A is most likely due to the reduced uptake of the formulations by the DC. The particles containing 0% Quil A did not induce any BMDC activation even with high levels of uptake because liposomes are not inherently immunostimulatory. As the model antigen used in this study was OVA, we examined the ability of the particles to induce a T-cell response using transgenic OT-I CD8 T cells. The results of these experiments were striking in that the formulations containing 0% Quil A were completely unable to stimulate T-cell proliferation or the upregulation of the IL-2 receptor chain (CD25). However, both of the Quil A-containing formulations were able to induce substantial antigen-specific T-cell proliferation and the upregulation of CD25. The formulation containing 20% Quil A again performed best, probably reflecting the differences in uptake. Neither of the formulations had titrated out to background levels of proliferation indicating it may be appropriate to decrease the amount of formulation used even further, which would further reduce the cytotoxicity of the formulation and would reduce the amount of antigen used.

In summary, we have shown that while liposomes are an excellent delivery system for the delivery of large amounts of antigen to APC without harming the cell, they are not highly immunogenic and would benefit from the addition of an adjuvant. The adjuvant we chose to examine was the saponin Quil A. The addition of 20% and 70% of this hydrophilic adjuvant had profound effects on the structure of the particles, decreasing the size of the particles and disrupting the lipid bilayer. The more Quil A added, the smaller the particles became and the less antigen could be incorporated. The increased hydrophilicity of the small negatively charged particles containing high amounts of Quil A also had a negative effect on uptake. The inclusion of the adjuvant Quil A (both 20% and 70%) did greatly increase the ability of the particles to activate BMDC and T cells in vitro. Analysis of all the data leads to the conclusion that the addition of Quil A is highly desirable in terms of immunogenicity, even though it makes antigen loading more difficult. Increasing the amount of Quil A from 20% to 70% is of no advantage as there is no corresponding increase in immunogenicity. This is most likely due to the low uptake caused by the increased hydrophilicity and increased negative zeta potential of the particles.

References

1. Gupta RK, Siber GR. Adjuvants for Human Vaccines – Current Status, Problems and Future-Prospects. Vaccine 1995; 13: 1263–76. | Article | PubMed | ISI | ChemPort |
2. Cox JC, Coulter AR. Adjuvants – a classification and review of their modes of action. Vaccine 1997; 15: 248–56. | Article | PubMed | ISI | ChemPort |
3. O'Hagan DT, Valiante NM. Recent advances in the discovery and delivery of vaccine adjuvants. Nat. Rev. Drug Discov. 2003; 2: 727–35. | Article | PubMed | ChemPort |
4. Singh M, O'Hagan DT. Recent Advances in Vaccine Adjuvants. Pharm. Res. 2002; 19: 715–28. | Article | PubMed | ISI | ChemPort |
5. Gursel I, Gursel M, Ishii KJ, Klinman DM. Sterically stabilised cationic liposomes improve the uptake and immunostimulatory activity of CpG oligonucleotides. J. Immunol. 2001; 167: 3324–8. | PubMed | ISI | ChemPort |
6. Kersten GFA, Crommelin DJA. Liposomes and ISCOMs. Vaccine 2003; 21: 915–20. | Article | PubMed | ISI | ChemPort |
7. Jaafari MR, Foldvari M. Targeting of liposomes to melanoma cells with high levels of ICAM-1 expression through adhesive peptides from immunoglobulin domains. J. Pharm. Sci. 2002; 91: 396–404. | Article | PubMed | ChemPort |
8. Medina OP, Soderlund T, Laakkonen LJ, Tuominen EKJ, Koivunen E, Kinnunen PKJ. Binding of novel peptide inhibitors of type IV collagenases to phospholipid membranes and use in liposome targeting to tumor cells in vitro. Cancer Res. 2001; 61: 3978–85. | PubMed | ChemPort |
9. Arigita C, Bevaart L, Everse LA et al. Liposomal meningococcal B vaccination: Role of dendritic cell targeting in the development of a protective immune response. Infect. Immun. 2003; 71: 5210–8. | Article | PubMed | ChemPort |
10. Babai I, Barenholz Y, Zakay-Rones Z et al. A novel liposomal influenza vaccine (INFLUSOME-VAC) containing hemagglutinin-neuraminidase and IL-2 or GM-CSF induces protective anti-neuraminidase antibodies cross-reacting with a wide spectrum of influenza A viral strains. Vaccine 2001; 20: 505–15. | Article | PubMed | ChemPort |
11. BacaEstrada ME, Foldvari M, Snider M, LittelvandenHurk SV, Babiuk LA. Effect of IL-4 and IL-12 liposomal formulations on the induction of immune response to bovine herpesvirus type-1 glycoprotein D. Vaccine 1997; 15: 1753–60. | Article | PubMed | ChemPort |
12. Ninomiya A, Ogasawara K, Kajino K, Takada A, Kida H. Intranasal administration of a synthetic peptide vaccine encapsulated in liposome together with an anti-CD40 antibody induces protective immunity against influenza A virus in mice. Vaccine 2002; 20: 3123–9. | Article | PubMed | ChemPort |
13. van Slooten ML, Storm G, Zoephel A et al. Liposomes containing interferon-gamma as adjuvant in tumor cell vaccines. Pharm. Res. 2000; 17: 42–8. | Article | PubMed | ISI | ChemPort |
14. Kensil CR. Saponins as vaccine adjuvants. Crit. Rev. Ther. Drug Carrier Syst. 1996; 13: 1–55. | PubMed | ISI | ChemPort |
15. Rhodes J. Covalent chemical events in immune induction: fundamental and therapeutic effects. Immunol. Today 1996; 17: 436–41. | Article | PubMed | ChemPort |
16. Demana PH, Davies NM, Berger B, Rades T. Incorporation of ovalbumin into ISCOMs and related colloidal particles prepared by the lipid film hydration method. Int. J. Pharm. 2004; 278: 263–274. | Article | PubMed | ChemPort |
17. Demana PH, Davies NM, Berger B, Vosgerau U, Rades T. A comparison of pseudo-ternary diagrams of aqueous mixtures of Quil A, cholesterol and phospholipid prepared by lipid film hydration and dialysis. J. Pharm. Pharmacol. 2004; 56: 573–580. | Article | PubMed | ChemPort |
18. Demana PH, Davies NM, Vosgerau U, Rades T. Pseudo-ternary phase diagrams of aqueous mixtures of Quil A, cholesterol and phospholipid prepared by the lipid film hydration method. Int. J. Pharm. 2004; 270: 229–39. | Article | PubMed | ChemPort |
19. Copland MJ, Baird MA, Rades T et al. Liposomal Delivery of Antigen to Human Dendritic Cells. Vaccine 2003; 21: 883–90. | Article | PubMed | ISI | ChemPort |
20. Copland MJ, Rades T, Davies NM. Hydration of lipid films with an aqueous solution of Quil A: a simple method for the preparation of immune-stimulating complexes. Int. J. Pharm. 2000; 196: 135–9. | Article | PubMed | ChemPort |
21. Sjölander A, VantLand B, Bengtsson KL. ISCOMs containing purified Quillaja saponins upregulate both Th1-like and Th2-like immune responses. Cell Immunol. 1997; 177: 69–76. | Article | PubMed | ChemPort |
22. Könnings S, Copland MJ, Davies NM, Rades T. A method for the incorporation of ovalbumin into immune stimulating complexes prepared by the hydration method. Int. J. Pharm. 2002; 241: 385–9. | PubMed |
23. Kersten GFA, Spiekstra A, Beuvery EC, Crommelin DJA. On the Structure of Immune-Stimulating Saponin-Lipid Complexes (ISCOMs). Biochim. Biophys. Acta 1991; 1062: 165–71. | PubMed | ChemPort |
24. Barr IG, Sjölander A, Cox JC. ISCOMs and other saponin based adjuvants. Adv. Drug Deliv. Res. 1998; 32: 247–71. | ChemPort |
25. Höglund S, Dalsgaard K, Lövgren K et al. ISCOMs and immunostimulation with viral antigens. In: Harris, JR (ed.) Subcellular Biochemistry. New York: Plenum 1989, 39–68.
26. Barr IG, Mitchell GF. ISCOMs (immunostimulating complexes): The first decade. Immunol. Cell Biol. 1996; 74: 8–25. | PubMed | ChemPort |
27. Reid G. Soluble-Proteins Incorporate into Iscoms after Covalent Attachment of Fatty-Acid. Vaccine 1992; 10: 597–602. | Article | PubMed | ChemPort |
28. Higuchi R, Komori T. Structures of Compounds Derived from the Acyl Moieties of Quillajasaponin. Phytochemistry 1987; 26: 2357–60. | Article | ChemPort |
29. Foged C, Sundblad A, Hovgaard L. Targeting Vaccines to Dendritic Cells. Pharm. Res. 2002; 19: 229–38. | Article | PubMed | ChemPort |
30. Sallusto F, Lanzavecchia A. Mobilizing Dendritic Cells for Tolerance, Priming, and Chronic Inflammation. J. Exp. Med. 1999; 189: 611–4. | Article | PubMed | ISI | ChemPort |
31. Robson NC, Beacock-Sharp H, Donachie AM, Mowat AM. The role of antigen-presenting cells and interleukin-12 in the priming of antigen-specific CD4+ T cells by immune stimulating complexes. Immunology 2003; 110: 95–104. | Article | PubMed | ChemPort |

Acknowledgements

This work was supported by grants from the Genesis Oncology Trust and the New Zealand Pharmacy Education and Research Foundation.