Thermal Treatment of a Novel Saponin-Cholesterol Nanoparticle Vaccine Adjuvant Named NanoQuil F70 Secures a Uniform Morphology and Size Distribution

Quil-A and cholesterol can, under certain conditions, form nanoparticles, but the conditions for the formation of a homogeneous population of a particular kind of particles with the same morphology and size have remained elusive. However, a well-defined uniformity, as well as a high degree of batch-to-batch consistency, are prerequisites for adjuvant formulations to be used for practical vaccines. To accomplish the above stated tasks in the present study, we demonstrated that Quil-A and cholesterol form well-defined worm-like stable particles with a mean diameter of around 40 nm by dynamic light scattering (DLS) under carefully controlled thermodynamical conditions with little or no Quil-A degradation. The nanoparticles thus prepared possessed a significantly reduced hemolytic effect in comparison to unformulated free Quil-A. However, when the same conditions were applied to QS-21/cholesterol nanoparticle formation the morphology achieved was coil spring-like particles with a mean diameter of around 70 nm. This discrepancy in particle morphology and size was attributed to the differences in hydrophobicity of the Quil-A, being a heterogenic fraction of acylated as well as deacylated saponins, and the QS-21. With the process optimizations leading to a standardized particle size distribution and nanoparticle morphology presented here, NanoQuil F70 holds the potential as a well-tolerated vaccine adjuvant for veterinary use raising a Th1/Th2 balanced immune response.


Introduction
The triterpene saponin mixture Quil-A has been used as vaccine adjuvant in animals for many years, but its lytic and surfactant activity introduce limitations for the dose to be applied in order to balance the adjuvant effect against the local adverse reactions at the injection site [1] [2] [3]. This problem was at least partially overcome by Morein et al. in 1984 by incorporating the Quil-A saponin into cage-like ISCOM particles when combined with cholesterol and phospholipids (PC) [4] and later with the ISCOMATRIX formulation [5]. Lately, the ability of triterpene saponins to interact with cholesterol alone was used by Morein, Hu et al. to design a novel Quil-A containing nanoparticle, originally referred to as G3 and later re-named NanoQuil.
This formulation of Quil-A + cholesterol, when used as a vaccine adjuvant, generated an immune response overall similar to that of the traditional ISCOM formulation with a balanced Th-1/Th-2 profile. In addition, this new adjuvant formulation when supplemented with the diterpene, Stevia glycoside, known as a commonly used and GRAS food additive (sweetener) [6], could even induce cross-protective immunity again a range of influenza virus strains and stimulated a vast array of innate immune response parameters in vitro.
However, G3-nanoparticles made in accordance with the original method gave lead to preparations with a heterogenous particle size distribution and morphology ( Figure 1) which turned out to be the main obstacle for transforming such a concept into a credible and standardized product with a uniform and reproducible particle morphology and particle size distribution. Although triterpene saponins, and QS-21 in particular, are known to be heat labile [7], we here present data that identified thermal treatment as the single most important factor for obtaining a uniform morphology and particle size distribution and the ability to maintain an acceptable low tissue lytic effect.

Formation of NanoQuil Particles at 70˚C
1.25 mg of plant-derived cholesterol was dissolved in 1 mL acetone. 0.1 mL of the cholesterol solution was transferred to a 1.5 mL Eppendorf tube. The acetone was removed while rotating the tube on a vortex mixer at 300 rpm, gradually forming a cholesterol membrane covering the V-bottom part of the Eppendorf tube. 1 mL or 0.9 mL of 70˚C PBS was added quickly to the tubes with the cholesterol deposit, followed by 10 µL Quil-A stock solution (100 mg/mL) or 100 µL QS-21 stock solution (10 mg/mL) respectively, resulting in a final saponin concentration of 1 mg/mL, and cholesterol concentration of 0.125 mg/mL (the weight ratio of Quil-A: cholesterol at this stage is 8:1). The tubes were incubated at 70˚C for 60 minutes, then cooled down to 4˚C followed by storing the tubes at 4˚C overnight. The tubes were centrifuged at 10.000 x G for 20 minutes followed by filtration through a 0.22 µM filter next day. The Quil-A + cholesterol complexes, named NanoQuil F70, were stored at 4˚C until use.

Negative Staining Electron Microscopy (TEM) for Morphology
The TEM was commercially run at BioVis Platform of Uppsala University.
A 5 µL drop of the sample was placed on a formvar and carbon coated 200-mesh copper grid. After 30 seconds, the excess solution was removed by blotting with filter paper. The sample was then directly stained for 10 seconds with 2% uranyl acetate. Excess of uranyl acetate was removed by blotting on filter paper. Dried grids were examined by TEM (FEI Tecnai G2 operated at 80 kV).

Quil-A Quantification
The

Cholesterol Quantification
Cholesterol content in the Quil-A + cholesterol products was determined using a commercially available cholesterol assay kit (ab65359 cholesterol/cholesteryl ester quantitation kit) from ABCAM according to its instruction manual.

Hemolysis Test
160 µL 5 times diluted sheep whole blood (Statens Serum Institute, Denmark) were mixed with 40 µL Quil-A standard or NanoQuil F70 sample and incubated at 37 ˚C for 45 minutes. After centrifugation at 500 x G for 5 minutes. The supernatant of 100 µL/well x 3 wells for each sample was analyzed spectrophotometrically at 540 nm. The degree of hemolysis reduction was calculated as the percentage of untreated Quil-A at the same concentration according to their respective OD values.

Quil-A + Cholesterol Particles Prepared in Accordance with the Original Method
Initially, Quil-A + cholesterol particles were formulated during overnight incubation at 37˚C, corresponding to body temperature, which was the highest temperature to think of due to concerns of saponin degradation. Under these conditions, various particle morphologies were generated, including worm-like, helices/double helices, coil spring-like etc., as shown by TEM (Figure 1(a)), and  (Figure 1(b)), demonstrating a shortage of homogeneity and reproducibility

Optimization of F70 Protocol
When Quil-A and cholesterol were incubated at 70˚C for 30 minutes, an identical population of worm-like particles was formed (Figure 2(a), Figure 2(b)). At temperatures lower than 70˚C, it was not possible to form such a homogeneous population. Further studies showed that prolonging the time of incubation to 60 minutes, an optimized result was achieved judging from TEM (Figure 2(c)) and DLS (Figure 2(d)). After this, we settled for a protocol referred to as NanoQuil

NanoQuil Particles Generated with QS-21 at 70˚C for 60 Minutes
After incubation with the optimized protocol for Quil-A i.e., at 70˚C for 60 min, QS-21 and cholesterol formed coil spring-like particles (Figure 3(a), Figure 3(b)), distinctly different from the worm-like particles seen with Quil-A + cholesterol under the same conditions.

Quil-A Recovery in the Quil-A + Cholesterol Products
The Quil-A content in the Quil-A containing NanoQuil F70 products was measured by the orcinol test, which gave, on average, a recovery rate around 97% in comparison to that of the starting materials used (Table 1) i.e., only 3% of the Quil-A was lost in the production process. The 3% Quil-A was trapped and detected (data not shown) in the pellets of cholesterol aggregates after the centrifugation step.

Cholesterol Recovery in the Quil-A + Cholesterol Products
The cholesterol recovery rate was low i.e., only about 19% (Table 2).

Reduction of Hemolysis Effect after Formulating Quil-A with Cholesterol into F70
A significant reduction of hemolysis was recorded when Quil-A was formulated into F70 nanoparticles with cholesterol as demonstrated with 5 batches of F70 with a mean hemolysis reduction value of around 31% (Table 4)

Discussion
NanoQuil particles consisting of Quil-A and cholesterol, originally referred to as G3 particles (indicating that they constituted the third generation of saponin-containing nanoparticles after the ISCOMs [4] (as the 1st generation) and ISCOMATRIX [5] [6] (as the 2nd generation), have shown interesting and promising results in several studies since 2014. Van de Sandt et al. [8], Hjertner et al. [9] and Hellmann et al. [10] [11] demonstrated that this new adjuvant formulation supplemented with DT could induce cross-protective immunity again influenza virus and in vitro stimulated a vast array of innate immune response respectively, including Th1 immunity [9]. Initially, Quil-A + cholesterol nanoparticles gave lead to mixed populations of several different kinds of particles after overnight incubation at 37˚C. Incubation temperatures higher than that had not yet been tested before the present study, largely due to the risk for saponin hydrolytic degradation. This is particularly true with QS-21 since it is highly sensitive to environmental changes such as pH, temperature etc. Table 5 summarizes all the particles that could be generated by ISCOMs (Quil-A + cholesterol + PC), Quil-A + cholesterol and QS-21 + cholesterol. Hemolysis effect has been used for in vitro evaluation of potential tissue lysis after injection of saponin products. NanoQuil alone can significantly reduce such an effect in comparison to Quil-A alone. Addition of DT to NanoQuil reduced hemolysis even further (unpublished data). This is of interest because when using Quil-A saponin in veterinary vaccines the dosing of Quil-A constitutes a balance between the adjuvant effect and the local tissue reactogenicity.
Consequently, a formulation like NanoQuil F70 with reduced tissue lytic effect will allow increasing the dose of Quil-A with the aim of achieving a higher degree of immunostimulation.
This is the first detailed study to characterize the morphology and particle size distribution of the cholesterol-containing dual component nanoparticles made from either Quil-A saponin or QS-21 with cholesterol. We have identified thermal treatment as the decisive parameter for particle morphology and size. We have devised a method for obtaining a high degree of standardization, comprising a homogenous particle morphology and size distribution for NanoQuil F70, by thermal treatment, without causing heat-derived degradation of the saponin molecules.
Taken together, NanoQuil F70 formulated with Quil-A with or without DT could be used as a more cost-friendly, simplified version of ISCOMs for use as a veterinary vaccine adjuvant. On the other hand, the same adjuvant formulated with QS-21 needs to be further investigated before being considered for human use.