1 Introduction

The monotypic genus Quillaja (from Quillajaceae) consists of two temperate evergreen tree species. The geographic distribution of Quillaja is quite intriguing, with the Andes Mountains being a geographical barrier isolating the two species. Quillaja saponaria Molina is found in Chile, while Quillaja lancifolia D.Don (previously known as Q. brasiliensis) is a representative tree of Araucaria forests in southern Brazil, northern Uruguay, northeastern Argentina, and eastern Paraguay [1, 2]. Quillaja saponins have a wide range of industrial applications—as adjuvants in vaccines, food additives, and cosmetic industry. According to Central Bank of Chile, quillay extracts brought US$ 52.6 million to the Chilean economy in 2019, as 515 tons were exported.

Several Q. saponaria saponins fractions were utilized as vaccine adjuvants. The most studied, Quil-A®, a crude saponin fraction extracted from Q. saponaria barks [3], is widely used in veterinary vaccines. However, Quillaja saponins has inherent toxicity. Therefore, two main approaches were applied to circumvent this matter. The first is isolating molecules with less toxicity from the saponin pool. A successful example, QS-21, was purified from Quil-A® and it has been demonstrated to have less toxicity than its parent saponin fraction. Today, QS-21 has been extensively studied in the form of a sole or combination adjuvant in several promising human vaccine formulations and it was recently approved to malaria and shingles vaccines [4]. The second approach is based on the inherent capacity of Quillaja saponins to interact with membrane lipids. Thus, in the presence of cholesterol and phosphatidylcholine, Quillaja saponins self-assembles in nanoparticles (immunostimulating complexes or ISCOM-matrices) with outstanding immunoadjuvant activity [5]. ISCOM-matrices included in human vaccines were found to be safe and well tolerated, with no serious adverse events [6]. Recently, the recombinant spike protein from SARS-CoV-2 was formulated with a Q. saponaria saponin-based nanoadjuvant (Matrix-M®). This vaccine demonstrate strong efficacy in phase 3 clinical trials demonstrating and is a possible candidate for fighting the SARS-CoV-2 pandemic [7].

The congener species Q. lancifolia (“pau-de-sabão”) has saponins that are chemically similar to those of Q. saponaria barks [8,9,10], showing a remarkable immunoadjuvant activity. The immunoadjuvant activity of purified Q. lancifolia saponin fractions (denominated QB-90 and QB-80) [11,12,13,14], as well as their poorly purified AE [15, 16], is equivalent to Quil-A®, being already demonstrated in experimental vaccines. Regarding AE, MALDI-ToF–MS (matrix-assisted laser desorption/ionization time-of-flight mass spectrometry) revealed peaks that showed similar patterns between Quil-A® in spectra regions of m/z 750–1000 and 1850–2200, which are characteristic of Quillaja saponins [15].

Purified Quillaja saponins fractions such as Quil-A® [17], QB-90, and QB-80 [13, 18,19,20] are recognized to self-assemble in cage-like nanoparticles (ISCOM) in the presence of lipids. However, the processes utilized to purify these saponin-rich fractions are expensive and time-consuming, involving silica, gel filtration, and reversed-phase chromatography. Therefore, processes that minimize and or eliminate steps in formulation of ISCOM-like nanoparticles are still desired. Here, we presented for the first time, ISCOM-matrices nanoparticles assembled using raw AE of Q. lancifolia.

2 Material and Methods

2.1 Plant Material and Aqueous Extract Preparation

Q. lancifolia D.Don leaves were collected from a 5-year-old specimen in Eldorado do Sul, RS, Brazil (29° 59′ 26″ S–51° 19′ 39″ W). A herbarium specimen is on deposit at Herbarium of the Department of Systematics and Ecology of Universidade Federal da Paraíba, João Pessoa, Brazil. AE preparation were carried out as previously described [21]. Briefly, leaves were dried in a circulating air oven at 37 °C for 5 days and submitted to a knife mill. The powered leaves (100 g) were transferred to a flat bottom balloon with 800 mL of distilled water and maintained for 8 h under slight agitation. After, the AE was filtered, and tannins were precipitated with gelatin. The AE was partitioned with ethyl acetate and the organic phase is discarded. The aqueous phase is concentrated in a rotary evaporator at 40 °C.

2.2 ISCOM-Matrices Nanoparticle Preparation

Concisely, a lipid-mix (2 mL) containing chloroform-dissolved cholesterol (Sigma-Aldrich, USA) and di-palmitoyl phosphatidylcholine (Avanti Polar Lipids, USA) (both at 100 mg/mL) was mixed with 10 mL of 20% octyl β-D-glucopyranoside (OGP, Sigma-Aldrich, USA). Five hundred microliters of the lipid-mix were added to 5, 10, 15, or 20 mg of AE dissolved in 4.5 mL of water. The mixture was dialyzed in a 12–14 kDa pre-hydrated dialysis membrane at 22 °C in 50 mM Tris–Cl (pH 8.5) containing 0.001% of thimerosal for 5 days against 20 L of buffer, changing the dialysis buffer every 24 h.

2.3 Transmission Electron Microscopy (TEM)

For TEM analysis, 10 µL of the nanoformulation was placed on a copper grid covered with a formvar and carbon film for 2 min. After, the remaining liquid was drained, and the sample was negatively stained with uranyl acetate (2% w/v) for 2 min. The analysis was performed in a Jeol JEM 1010 transmission electron microscope (Jeol, Japan) at an accelerating voltage of 90 kV and at magnifications between 30,000 and 150,000X.

2.4 Dynamic Light Scattering (DLS) and ζ-Potential Parameters

Standard operating procedure (SOP) for the dynamic light scattering (DLS): material RI = 1.59, dispersant RI (water) = 1.33, T = 25 °C, viscosity (water) = 0.887 cP, measurement angle = 173° backscatter, measurement position = seek for optimum position, automatic attenuation. Additionally, for ζ-potential, automatic attenuation and voltage selection.

2.5 Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry (MALDI-ToF–MS)

MALDI-ToF–MS measurements were conducted on a Microflex LR MALDI-ToF–MS instrument (Bruker Daltonics, USA) with a 337-nm nitrogen laser operated in positive ion lineal mode with delayed extraction and optimized in the m/z range of 0–20 kDa. One microliter of each nanoparticle preparation was mixed with 1 μL of matrix solution (2,5-dihydroxybenzoic acid, 10 mg/mL in H2O with 1% of TFA). The sample and matrix mixes were spotted onto a 96-well stainless-steel plate and allowed to air dry for 15 min at room temperature before measurements. Calibrations were performed with a peptide calibration standard mix (Bruker Daltonics, USA). The laser was fired 100 times at each of ten locations for each sample well on a 96-well plate for a cumulative 1000 shots per sample well taken at 30% intensity.

2.6 AE Saponin Profile by Direct Injection Electrospray Ionization (ESI) and Time-of-Flight Detection (DI-ESI-ToF–MS)

For saponin identification, DI-ESI-ToF–MS were applied on a micrOTOF II high-resolution mass spectrometer (Bruker Daltonics, USA). Spectrometer analysis parameters were as follows: 4.5 kV capillary, ESI in negative ion mode, a 500 V end plate offset, a 40.6 psi nebulizer, a dry gas (N2) flow rate of 8.0 L/h, and a temperature of 200 °C. Spectra (m/z 1400–2400) were recorded every 2.0 s. The ESI Tuning mix® (Sigma Aldrich, USA) was used as the internal calibrator.

3 Results and Discussion

ISCOM-matrices nanoparticles assembly were verified by TEM and DLS and are shown in Fig. 1a. The self-assembled nanostructures formulated with AE (1–4 mg/mL) in the presence of lipids give a spherical-shaped cage-like particles of ~ 40 nm, typical of ISCOM or ISCOM-matrices nanoparticles. However, in formulations containing 3 or 4 mg/mL of AE, the number of cage-like particles is high and more homogeneous in comparison to formulations containing fewer saponins. These observations support the fact that the experimental conditions that used a high concentration of AE (it also increases the saponin’s concentration) are more adequate for the preparation of the observed micelles. Moreover, aggregates and amorphous structures were also observed, particularly in formulations containing 1 and 2 mg/mL of AE.

Fig. 1
figure 1

ISCOM-like nanoparticles assembled with Q. lancifolia AE. a TEM were obtained after negative staining with uranyl acetate. The arrowhead shows the assembly of cage-like particles. The scale bar bellows each figure represents 100 nm. b DLS measurements of average hydrodynamic diameter and ζ-potential of nanoformulations. Size distribution were stated as an average of ten measurements. All measurements were carried out at 25 ± 0.1 °C. c MALDI-ToF mass spectra of the nanoformulation containing 3 mg/mL of AE showed characteristic peaks of saponins between m/z 1900 and 2200. The products were detected as [M + Na]+ and/or [M + K]+ ions. d DI-ESI-ToF–MS data from Q. lancifolia AE. The most frequent ions were assigned, based on its homologous, well characterized, saponins from Q. lancifolia and Q. saponaria

Full size image

Concerning DLS measurements (Fig. 1b), the z-average size (size distribution by number) was consistent with TEM data. Interestingly, the size of the nanoparticles decreases when greater amounts of saponins are used in the formulation (Fig. 1b). The particle size in the formulation containing 3 or 4 mg/mL AE was similar, showing once again, that these concentrations are suitable for the formulation of ISCOM-matrices from raw Q. lancifolia AE. The nanoparticles assembled with AE has a negative zeta (ζ)-potential (fluctuating from − 10.1 to − 12.5 mV) since carrying a net negative charge due to the glucuronic acid component present in Quillaja saponins [22]. Entire ζ-potential data are presented in Table 1.

Table 1 ζ-potential data
Full size table

The saponin profile of ISCOM-matrices particles assessed by MALDI-ToF–MS is quite similar in all formulations (Fig. 1c), showing saponins with m/z of 1700–1950 and 2000–2200. This is in agreement with Q. lancifolia saponins MS data and shows that these saponins can form ISCOM-matrices structures [8, 9, 15]. The most frequent ion is 2012.4, which may correspond to saponin with m/z of 1988.9 (QS-21), detected as [M + Na]+ ion. In order to confirm the identity of QS-21 and tentatively identify another Quillajaceae saponins, a direct injection in a high-resolution mass spectrometer followed by electrospray ionization and time-of-flight MS detection (DI-ESI–MS) was performed.

The most frequent ions were tentatively assigned, based on its homologous and well-characterized saponins from Quillajaceae, and the full scan of DI-ESI–MS is presented in Fig. 1d and in Table 2. Detailed high-resolution MS data can be obtained in Fig. 1d. Using this approach, it was possible to identify the most frequent saponins ions in the Q. lancifolia AE. These includes the S4/S6 (QS-21; exact mass 1988.9242, detected mass, 1987.9100, error 3.5 ppm), 15b, S2, and S13 saponins [23,24,25]. Some ions could not be attributed to saponins already identified in Quillaja by spectroscopic or spectrometric analysis, indicating probable novel molecules in Q. lancifolia leaves (data not shown).

Table 2 Most frequent saponin ions tentatively identified by high-resolution mass spectrometry (DI-ESI-ToF)
Full size table

Furthermore, the ISCOM-matrices saponin profile is quite similar to QB-90 saponin fraction (Fig. 1). QB-90 is a C18 reversed-phase chromatography purified Q. lancifolia saponin-rich fraction and extensively studied by our research group. This similar profile brings to light a possible new saponin purification/enrichment method, based on dialysis. Dialysis-based methodologies are less expensive than techniques based on reversed-phase chromatography. Furthermore, dialysis-based purifications do not use organic solvents (such as methanol, used to elute QB-1290 from the C18 column), resulting in an environmentally friendly process.

The commercial Q. saponaria fraction Quil-A® has been successfully used for veterinary applications, but it is known that side effects as local reactions and granulomas can be attributed to its components. Q. lancifolia AE have some advantages. They can induce immune responses similarly to Quil-A®, including promoting dose sparing [15, 16, 26, 27]. This could be explained by the presence QS-21, easily detected in AE of Q. lancifolia leaves, that may explain the strong immunoadjuvant activity of AE, including B- and T-cell immunological responses, and their low toxicity in experimental vaccines. Furthermore, AE extraction process is simple and less expensive than the laborious and costly chemical purification steps to obtain saponin-rich fractions, involving silica and reverse-phase chromatography. Regarding forest extractivism concerns, AE was derived from leaves of Q. lancifolia that are readily renewable alternative sources of saponins compared to Q. saponaria barks.

As stated previously, one of the critical issues in the use of saponins as vaccine adjuvants is the fact that they can raise safety concerns due to their toxic effects in the injection site [28]. The results obtained in our previous studies showed that AE has a low toxicity profile both in vitro and in vivo [15]. Using a hemolysis assay, a widely accepted test to predict saponins toxicity, AE has been shown to cause membrane damage at high dose concentrations (with HD50 close to 1 mg/mL). This value is calculated 23-fold higher than Quil-A® (HD50 35.9 μg/mL) [15]. The nanoformulations presented here were submitted to the hemolysis test and no hemoglobin release was verified, indicating no membrane damage. This suggests that the membrane-toxic effects are abrogated when saponin formulations are complexed with lipids. Additionally, due to the presence of the well-characterized saponin QS-21 in Q. lancifolia AE, this can be an enriched intermediate for obtaining/purifying this potent adjuvant for use in human vaccines.

According to previous studies involving the immunoadjuvant potential of AE and the data presented here, we propose that Q. lancifolia AE and their nanoparticles has a promising saponin-based vaccine adjuvant, as it presents some advantages over commercially available saponins. Foremost, AE contains known saponins with a strong immunoadjuvant activity, as QS-21. Additionally, AE is a raw extract of saponin and, as such, is easier and cheaper to produce compared to purified saponin fractions. Furthermore, in addition to its efficacy, AE was proven to be safe and well-tolerated in mice. Finally, AE proved to be able to form cage-like structures similar in shape and size to ISCOM particles and, with that, opens up new possibilities to AE in nanoformulations. These are some advantages that are carefully analyzed when a new vaccine adjuvant is considered a candidate to be escalated from experimental to industrial uses.