Saponins are well-known for their ability to induce systemic immune responses when co-injected with antigen via intramuscular immunization (36). Additionally, saponins are effective as mucosal adjuvants when delivered intranasally ( Table 1 ). Intranasal administration of a DNA vaccine encoding the envelope of human immunodeficiency virus type 1 (HIV-1) along with QS-21 as an adjuvant has been shown to increase both systemic and mucosal immune responses against HIV-1, including production of intestinal sIgA and cytolytic activity of mesenteric lymph node cells (50). Saponins derived from Polygala tenuifolia or Chenopodium quinoa have demonstrated enhanced antigen-specific immunoglobulin G (IgG) and local IgA responses to co-administered antigens in lungs when used as adjuvants following intranasal administration (44, 51). In another study, QS-21 loaded on liposome induced higher levels of sIgA compared to liposome without QS-21 when used as an adjuvant for a tetanus toxoid antigen after nasal administration (52). The use of oleanolic acid 28-cinnamoylglucoside as an adjuvant in a nasal-administered influenza split vaccine showed a slight but statistically significantly enhancement in anti-influenza virus sIgA in the nasal washes (46). Moreover, intranasal immunization of saponins-adjuvant vaccines showed protective effects against influenza virus and Toxoplasma gondii cysts challenge (44, 49, 53, 54). These studies highlight the potential of saponins as adjuvants for nasal vaccines.

ISCOMs was first described in the 1980s by Morein et al. as a novel type of immunostimulating complex. It has a spherical cage-like structure composed of saponin, cholesterol, phospholipids, and antigens (55, 56). On the other hand, ISCOMsATRIX (IMX), also called empty ISCOMs, has similar structure and composition to ISCOMs but without the incorporated antigens (56, 57). Both ISCOMs and IMX have been found to be highly immunostimulating due to the combination of an in-built adjuvant (Quil A) with a particulate delivery system. Moreover, they are less toxic and do not have hemolytic activity due to the embedding of saponins into cholesterol (56). The assembly of ISCOMs relies on hydrophobic interactions, making them suitable for incorporation of hydrophobic antigens derived from envelope viruses or cell membranes. In contrast, IMX has a negatively charged surface, which enables it easily to interact with a broad range of positively charged antigens to make IMX vaccines (57). However, this approach limits the binding of neutral or negatively charged hydrophilic antigens to IMX (57). Thus, further research is needed to expand the range of antigens available for IMX vaccines and simplify the production process (58). ISCOMs and IMX vaccines have been developed for a variety of diseases, including viruses, bacteria, parasites, and tumors (56). Both have shown good safety and well-tolerated in animal and human studies, with the ability to induce strong antigen-specific cellular or humoral immune responses (56, 59). There are several ISCOMs vaccines registered for veterinary use, but no ISCOMs or IMX vaccines have been approved for human use yet (60).

Studies have demonstrated that intranasal administration of ISCOMs or IMX vaccines with a variety of antigens can elicit potent mucosal cellular and humoral immune responses, including local IgG and sIgA antibodies in the respiratory tract, as well as systemic and distal mucosal responses, such as in the genital and intestinal tracts (61–70). Pulmonary delivery of ISCOMs or IMX vaccines has also been shown to induce both systemic and mucosal antibody responses against various antigens, including those of influenza virus and Mycobacterium tuberculosis (71–73). Protective efficacy of intranasal immunization with ISCOMs or IMX vaccines has been demonstrated in several studies against various pathogens such as influenza virus, Helicobacter pylori, Angiostrongylus costaricensis, and Eimeria tenella (74–80). These findings highlight the great potential of ISCOMs or IMX as adjuvants for nasal vaccine development ( Table 2 ) (22).

Since the 1990s, saponins-based adjuvants have moved forward to combine saponins with other adjuvants to provide a synergistic adjuvant effect. GlaxoSmithKline (GSK) biologicals developed several Adjuvant Systems (AS) that incorporate QS-21 with other immunostimulants in different formulations (81). For instance, AS01 is a liposome-based system containing QS-21 and 3-O-desacyl–monophosphoryl lipid A (MPL), while AS02 contains QS-21 and MPL in an oil-in-water emulsion (81). AS05 is a liposome-based system with QS-21, MPL, and alum, while AS15 contains QS-21, MPL, and CpG7909 (82, 83). The U.S. Army has also developed an adjuvant called Army Liposome Formulation containing QS21 (ALFQ) which with different liposome properties to AS01 (84). MPL, a detoxifying lipid from Salmonella Minnesota LPS, is a Toll-like receptor 4 (TLR4) agonist and promotes the production of pro-inflammatory cytokines by activating APCs (85, 86). The combination of QS-21 and MPL in AS enhances innate immunity, stimulates antigen-specific T cell responses, and converts mouse antibodies to IgG2c subtype (87, 88). The use of saponins in combination with cholesterol in adjuvant complexes can also reduce hemolytic toxicity. These adjuvant systems have shown good effects in various vaccines against pathogens such as malaria, tuberculosis, SARS-CoV-2, HIV, and Campylobacter jejuni (20, 88–93). In particular, AS01 has been licensed for use in the malaria RTS,S/AS01 vaccine and herpes zoster vaccine (Shingrix®) (94). However, there are currently no reports of the use of ASs or ALFQ in nasal vaccines.

AS01, AS05, AS15, and ALFQ are liposome-based adjuvants. Liposomes can protect antigens from degradation and increase antigens absorption across the nasal epithelium, which can prolong the time antigens remain in the respiratory tract (95). They are biocompatible, biodegradable, and safe for nasal vaccine development (95). Intranasal vaccination of influenza vaccine with liposome-based adjuvants has been shown to protect mice from both homologous and heterologous influenza virus challenge (96). Therefore, as liposome-based adjuvants, ASs and ALFQ could be promising options for nasal vaccines development, which needs further investigation.

Matrix M is the third generation of ISCOMs technology that contains two matrix particles, Matrix-A and Matrix-C, each made from a different, well-characterized saponin fraction from purified Quillaja saponin fractions A and C, along with cholesterol and phospholipids (97, 98). Matrix-C is a highly adjuvant active saponin, while Matrix-A is a weaker but very well-tolerated saponin. The optimal ratio of mixture of Matrix-A and Matrix-C can be explored with co-administering the antigen to achieve the best balance of adjuvant activity and safety (98, 99). Matrix M has been shown to effectively activate and recruit immune cells such as DCs, B cells, and T cells to draining lymph nodes, resulting in strong cellular and humoral immune responses (100, 101). It has been used in clinical trials for vaccines against influenza, malaria, and SARS-CoV-2 (98, 99, 102–104). Interestingly, Matrix-M has mucosal adjuvant properties that enhance mucosal immune responses. Intranasal immunization of Matrix-M (in a ratio of 91:9 of Matrix-A and Matrix–C) adjuvanted a virosomal influenza H5N1 vaccine elicited a significant cross-reactive serum antibody response and protected against a highly pathogenic viral challenge in a mouse model (105). Another study found that intranasal immunization of mice with a Matrix-M-adjuvanted DNA vaccine encoding the P6 outer membrane protein of nontypeable Haemophilus influenza resulted in the induction of P6-specific nasal IgA and serum IgG, as well as enhanced bacterial clearance (Table 3) (106). These results highlight that Matrix-M adjuvant is a promising mucosal adjuvant for nasal vaccines formulations.

The mucosal immune system is composed of inductive and effector sites. NALT and local and regional cervical draining lymph nodes are induction sites that can trigger both systemic and local mucosal immune responses following nasal vaccination (8, 107). NALT is an organized structure that contains all the immune cells and is covered by a lymphoid epithelium with microfold (M) cells, which are specialized for antigen uptake (8, 107). M cells can transport antigens and adjuvants to APCs, such as DCs, which accumulate immediately below the epithelium and M cells. The APCs phagocytose, process and present antigens to the surrounding naive T cells (108, 109) ( Figure 2A ). Vaccines administered nasally can partially pass into the lungs, where alveolar macrophages take them into the interstitium and to draining lymph nodes to activate DCs (110, 111) ( Figure 2B ). Saponins-based adjuvants in liposome or ISCOMs can deliver antigens to epithelial cells and APCs. The negatively charged IMX nanoparticles have mucoadhesion properties, and the intranasal administration of IMX vaccine has been shown to be retained in the nasal cavity for a longer period of time, allowing antigens to access immune inductive sites in the nasal mucosa and inducing to high mucosal and systemic immune responses (57). Saponins can activate the production of cytokines by resident innate cells, including stromal cells, which then recruit neutrophils, monocyte, and DCs into the respiratory tract to take up more antigens (20).

Proper induction of innate immune response in the respiratory mucosal and draining lymph nodes is crucial for the quality and magnitude of the adaptive responses and vaccine efficacy. Saponins, such as AS01 and QS-21, have been shown to activate and stimulate APCs to release inflammatory cytokines at draining lymph nodes when administrated intramuscularly (88). Saponins can activate innate immune response by binding to the lectin receptors in the innate immune cells, such as DC-SIGN on DCs, and eliciting cytokines and chemokines production in APCs (38, 88, 112) ( Figure 2A ), which leads to a rapid and substantial influx of neutrophils, monocytes, DCs and T-cell populations in the lymph node. AS01 can also enhance the expression of the costimulatory molecules such as CD86 and CD40 on the cell surface of APCs (88). The activated innate immune responses are largely resolved by day 7 (88). These immunostimulant responses could also occur in the respiratory mucosa, NALT, and cervical lymph node, when saponins are delivered intranasally ( Figure 2A ). As intranasal delivery of IMX-adjuvanted human T-cell lymphotropic virus type 1 vaccine was shown to modulate cytokines expression, with increased IFN-γ and IL-10 expression and decreased TGF-β1 levels (68).

The induction of antigen-specific resident memory cells is important for long-term protection against pathogens, which is where mucosal vaccines play a critical role. In NALT or draining lymph node, activated DCs (by saponins and antigens) are capable of activating T cells that differentiate into effector cells and later into memory cells. Activated CD4⁺ T cells, especially follicular helper T cells (Tfh) induce the development of IgA-secreting B cells in NALT or draining lymph nodes ( Figure 2A, C ). AS01 may activate Tfh responses, which are correlated with antigen specific IgG and memory B cells (67, 113–115). The resulting B lymphocytes migrate locally to the lymphatic follicles and proliferate in the germinal center, leading to mucosal and systemic immune responses characterized by the secretion of IgA and IgG, respectively (15). Antigen-specific dimeric immunoglobulin A (dIgA) is transported by epithelial cells through polymeric immunoglobulin receptors (pIgRs) and released as sIgA into the nasal lumen (116, 117) ( Figure 2A ). The antigen and adjuvant-loaded DCs the mature and migrate to the follicular B-cell areas and interfollicular T-cell zone in local draining lymph nodes of nasal tissue or lungs, where they present the antigen to neighboring naive T cells, triggering adaptive immune responses ( Figure 2C ) (106, 118). T cells and B lymphoblasts activated by APCs migrate throughout the body via the circulatory system, such as Intranasal immunization of Matrix-M adjuvanted influenza vaccine can activate antigen-specific CD4⁺ and CD8⁺ T cells responses in spleen (105). In addition, immune cells can diffuse through the common mucosal immune system that connects the induction site to the effector sites ( Figure 2D ) (119). Thus, adaptive immune responses are not confined to the site of induction, but also occur at distant mucosal region (63, 119) ( Figure 2D ). Such as, ISCOMs vaccines after nasal administration can induce local antibody secreting cells, CTL, and memory B and T cell responses in lungs (62, 63, 120, 121). The underlying mechanism of saponins to induce mucosal and systemic immune response after intranasal administration requires further study.

Antigen-specific T cells and IgA⁺ B cells that are induced in NALT and lymph nodes migrate to the mucosa of the respiratory tract through the thoracic duct and circulation. At these effector sites, the IgA⁺ B cells differentiate to IgA⁺ plasma cells to secret sIgA, which is very important in preventing infections by inhibiting the adhesion, invasion, and spread of pathogens to epithelial cells (15, 16, 116) ( Figure 2A ). Intranasal immunizations with ISCOMs vaccines containing different antigens induced not only local IgG and sIgA antibody production in the lungs, but also in distal mucosal system (61, 63–65). The antigen-specific T cells that homing to mucosal region can further differentiate into tissue-resident memory cells that express cell surface markers such as the CD69 and CD103. These memory cells persist in the mucosal tissue for extended periods without entering the circulation and provide a front-line defense against pathogenic invasion by rapidly reactivation in response to antigenic pathogens. Numerous studies have shown that intranasal administration of saponins can induce sIgA production in the respiratory mucosa ( Tables 1 - 3 ) ( Figure 2A ) (44, 50, 53, 61, 64, 68). Further research is needed to understand the regulation of tissue-resident memory T cell and B cell responses triggered by nasal administration of saponins.

Ensuring the safety of the effective adjuvants is the first priority in vaccine development, particularly for those based on saponins used in nasal vaccine, where their safety profile in humans is not yet fully established. Saponins are amphiphilic compounds containing both hydrophobic and hydrophilic regions and their hydrophobic regions can interact with the cell membranes of red blood cells and disrupt the membrane integrity, leading to hemolysis. To mitigate the toxicity, saponins can be formulated with other components, such as cholesterol used in AS01 and ISCOMs, which can interact with the hydrophobic region of saponins to prevent their interaction with cell membranes (122). The adjuvant effects of saponins are dependent on their immunostimulatory effect, which can stimulate transient inflammation at the injection site. The effects usually are mild and transient, returning to normal within several days after administration. However, high dose of adjuvants may cause strong inflammation or cell death. Although saponins and saponin-based adjuvants are generally considered safe for parenteral immunization (93, 102, 106 (90, 99, 102, 103, 106, 123), they have been associated with some adverse effects, including local pain, redness, swelling, and fever, especially at the injection site (124, 125). So, intranasal administration of saponin-based adjuvants may induce inflammation and cellular damage in respiratory tract and lungs at high doses, which is a safety concern that requires careful evaluation in the development of nasal vaccines (50). Another important safety concern for nasal adjuvants is their potential side-effects on the central nervous system, since they may be transported from the olfactory epithelium to central nervous system (126, 127). Thus, evaluating the local and systemic toxicities of nasal vaccines containing saponin-derived adjuvants is crucial, which should be carefully evaluated for a good balance of efficacy and side effects in pre-clinical and clinical studies.

Successful development of saponin-based adjuvants for nasal vaccines requires consideration of the unique physiological, chemical, and immunological properties of the nasal cavity. The physiobiological barrier system, including the mucus and mucociliary movement, may hinder antigens and adjuvants absorption (14). Nasal mucus containing proteases and aminopeptidases may degrade the vaccine components (14). However, saponins-based adjuvant such as ASs or ISCOMs, which use liposome as a delivery system, may overcome these barriers and transport antigens and adjuvants to epithelial cells and APCs, which require further evaluation. Additionally, administering vaccines through the nasal cavity may result in the passage of vaccines into the lungs or oral cavity, making it difficult to determine the exact amount of antigens or adjuvants that reach the immune system. To address this issue, metered-dose nasal spray devices can be used to control the amount of solution delivered per spray, allowing for the evaluation of efficacy and side effects to determine the optimal spray dose. In conclusion, overcoming these obstacles is critical when developing saponins-based adjuvant for nasal vaccines.

Different adjuvants have unique immunomodulatory effects and the protective immune responses for different pathogens can vary. To achieve enhanced and broad protective immune responses for co-administered antigens, a combination of saponins with different mucosal adjuvants or delivery systems can be used, potentially resulting in additive enhancing effects on mucosal immune responses

AS01 is a successful example of combining two adjuvants, MPL and QS-21, in one formulation. Other attempt to combine different adjuvants includes intranasal administration of the ginseng stem-leaf saponins (GSLS) in combination with selenium (GSLS-Se). This combination enhanced the adjuvant effect on live vaccines for Newcastle disease virus and infectious bronchitis virus in chickens, promoting significantly higher antigen-specific antibody responses, increased lymphocyte proliferation and production of IFN-γ and IL-4 compared to GSLS alone. The increased antibody was able to neutralize corresponding viruses (128). Antigen-specific sIgA and the numbers of IgG⁺, IgA⁺, IgM⁺ plasma cells were significantly higher in GSLS-Se group than the control in the Harderian gland (129). Another example is the combination of cyclic guanosine monophosphate-adenosine monophosphate (cGAMP) and saponins, which improved protective response to influenza. Saponins can increase the permeability of cell membrane, allowing more cGAMP to enter the cells and increasing the utilization rate of saponins, leading to stronger immune effects (130, 131). Furthermore, nasal delivery systems such as microspheres or nanospheres made of chitosan, PLGA (poly[D,L-lactic-co-glycolic acid]), alginate, or cross-linked dextran have been employed to encapsulate saponins for nasal administration (132–135). These studies provide compelling evidence for the potential of developing mucosal adjuvants through combinations of saponins and other adjuvants.

ISCOMs and IMX complexes are also a promising option for the development of combined adjuvant carriers with other adjuvants to increase adjuvant activity. Combination of a mucosal adjuvant cholera toxin B (CTB) with ISCOMs has shown increased adjuvant effects. Intranasal immunization of synthetic peptide polymerized with the CTB in ISCOMs showed increased protection against polymerized synthetic peptide in ISCOMs (78). In addition, the cholera toxin subunit A1 (CTA1) fused to a dimer of the Ig-binding D-region of Staphylococcus aureus protein A (CTA1-DD) has been incorporated into ISCOMs to create CTA1-DD/ISCOMs adjuvant. This adjuvant significantly augments the immunogenicity of the antigen, as demonstrated by increased levels of specific serum antibodies, balanced Th1 and Th2 priming, and strong activation of DCs (136–138). Moreover, intranasal vaccination with CTA1-DD/ISCOMs adjuvanted antigens Ag85B-ESAT-6 from M. tuberculosis significantly reduced the M. tuberculosis burden in the lungs compared to control animals (139). However, combination of adjuvants does not always show enhanced effects. For instance, the combination of chitosan with IMX nanoparticles in a vaccine using the PR8 antigen induced a weaker immune response compared to IMX nanoparticles alone after intranasal administration (70).

Combining different routes of adjuvant delivery through prime-boost strategies has the potential to enhance both mucosal and systemic immune responses. For example, a systemic prime-intranasal boost strategy with an influenza vaccine adjuvanted with the liposomal dual TLR4/7 adjuvant has been shown to enhance both systemic and local/mucosal immunity. This regimen results in the secretion of antigen-specific sIgA and development of tissue-resident memory T cells in the respiratory tracts, as well as cross-reactive sIgA to multiple influenza virus strains (140). Similarly, this prime-boost strategy has been successful in preventing SARS-CoV-2 transmission and disease development through vaccination (141). Therefore, combining different routes of adjuvant delivery can provide comprehensive and effective immune responses at both the systemic and mucosal levels. Such strategies can also be applied to the study of saponin-based adjuvants in nasal vaccines. However, the optimal combination of administration routes may vary depending on the specific vaccine and pathogen being targeted, and thus requires thoroughly experimentation.

Achieving effective vaccine-induced immune responses requires a deeper understanding of the mode of action of saponin-based adjuvants, which can expedite the development of novel vaccine strategies. Collaborative research efforts between various disciplines, including chemistry, biochemistry, molecular biology, immunology, material science, and artificial intelligence, are crucial to achieve this goal. By using of multiomics technology such as transcriptomics, proteomics, and metabolomics at bulk and single-cell levels, researchers can uncover the function and mechanism of saponins-based adjuvants. Machine learning algorithms have also been applied to identify immune signatures associated with adjuvant formulations like AS01B, AS02A, AS03, CpG, and MF59 (142–144). Combining machine learning with in-depth profiling of vaccine-induced immune signatures including cytokine, cellular, and antibody responses can lead to identify adjuvant-specific immune response characteristics that can predict the efficacy and safety of the adjuvants in human (142–145). Moreover, immune response patterns in mice may not be predictive of responses in human (146). Therefore, organoids derived from lymphoid tissues, such as tonsils, will proved a powerful platform for studying key immune mechanisms related to human and enable rapid preclinical research on saponins-based adjuvants (146).