In general, triterpenoid saponins and steroidal saponins could form stable complexes with cholesterol in the membrane. Studies have shown that saponins tend to bind to cholesterol on the erythrocyte membrane, leading to the formation of insoluble complexes. This alteration in osmotic pressure ultimately causes the distension and rupture of erythrocytes, resulting in hemolysis (de Groot and Muller-Goymann, 2016). Specifically, digitonin molecules have been demonstrated to bind to cholesterol in the membrane (Sudji et al., 2015). This binding process removes cholesterol from the membrane core and leads to the formation of cholesterol-digitonin complexes on the membrane surface (Frenkel et al., 2014). Additionally, α-tomatine could form similar complexes, while the aglycone tomatidine lacks this capability (Sucha et al., 2013). A study conducted by Gestetner et al. (1971) revealed that lucerne saponins have a high affinity for various sterols, particularly cholesterol and 7-dehydrocholesterol, forming water-insoluble complexes. In our research (Jiang et al., 2022), we observed that polyphyllin I can interact with cholesterol, resulting in the formation of an insoluble precipitate when dissolved in ethanol. Similarly, a separate study by Story et al. (1984) investigated the interactions between alfalfa saponins and cholesterol in-vitro. They observed that alfalfa saponins effectively bound cholesterol from both ethanol solutions and micellar suspensions. OSW-1 demonstrated cholesterol-dependent membrane activity comparable to digitonin (Malabed et al., 2017). The unique structure of OSW-1, with a partially acylated sugar chain attached at the D ring of the steroidal aglycone and a distinct flat triangular shape in its three-dimensional structure, contributes to its potent activity. This amphiphilic polarity likely differentiates the binding mode of OSW-1 to cholesterol from that of typical saponins like digitonin, which has a 3-O-sugar chain (Malabed et al., 2020).

Structurally, saponins consist of one or more hydrophilic sugar moieties and a lipophilic steroid or triterpenic component, making them amphiphilic compounds (Bottger et al., 2012). When saponins reach their critical micelle concentration (CMC), they aggregate in solution, coexisting with free monomers below the CMC. These aggregates are characterized by their “soft” or fluid-like nature, resulting from weak intermolecular forces such as hydrogen bonding, van der Waals, hydrophobic, or screened electrostatic interactions. When saponin aggregates encountered a compound capable of forming a more stable complex, they undergo dissociation and form a new complex with the compound.

Quillaja saponins can form mixed micelles with cholesterol, due to a 1000-fold enhancement in the solubility of the sterol (Orczyk and Wojciechowski, 2015). These mixed micelles are larger in size compared to pure saponin micelles, measuring 10 nm instead of 7 nm. They also exhibit a higher CMC concentration and aggregation number. It is important to note that cholesterol is a component of the lipid compartment within these micelles. In a study conducted by Demana et al. (2005) observed that the worm-like micelles assembled when different proportions of saponins and cholesterol were mixed. Additionally, Saponaria officinalis L. saponins were found to be capable of forming micelles with various shapes, including rod-like, worm-like, and spherical, in the presence of cholesterol (Lendemans et al., 2005).

Membranes provide an amphiphilic environment characterized through a hydrophobic gradient, which could provide a hydrophobic core. The artificial membrane models that have been used to study saponins has provided valuable insights into the interactions between these molecules and various membrane components within this amphiphilic environment. These studies have also offered valuable insights into the mechanisms underlying saponins’ interaction with cholesterol in these membrane models (Lorent J. H. et al., 2014).

Some saponins could insert themselves into the monolayers without cholesterol in many different conditions. However, in the case of α-tomatine (Story et al., 1984), it must be accompanied by cholesterol for it to be able to insert in the membrane layers. This insertion is effective only when the sterol’s hydroxyl function at the third position in the monolayer is in the β configuration (Bottger and Melzig, 2013). In another case, comprising cholesterol and egg yolk phosphatidylcholine, permanent insertion of digitonin was observed upon formation of an equimolar complex. In this study, crucial conditions for the combination effect of digitonin and cell membrane cholesterol was proposed. Firstly, with the increase of the ratios of the digitonin and cholesterol, both formed aggregated within the cell layer (Frenkel et al., 2014). Then rising to a certain molar ratio, an intermediate complex emerged, consisting of a mixture of aggregates and equimolecular compounds (Sudji et al., 2015). Lastly, after reaching a certain proportion, an equimolecular complex was produced in the membrane.

Numerous studies have consistently demonstrated the significant role of cholesterol in the process of membrane permeabilization induced by saponins. In the case of most saponins, cholesterol has been found to enhance or be essential for permeabilization. The permeabilization mechanisms of saponins can be summarized as follows: The first mechanism suggested that saponins interacted with sterols, forming equimolecular complexes within the membrane (Lorent J. et al., 2014). As the density of these complexes increased, a new lipid phase was generated due to the hydrophilic interactions with the saponins’ sugar moieties. The three-dimensional structure of these compounds could form some aggregate such as spherical buds or tubules. This rearrangement of the membrane structure ultimately led to membrane disruption. The second mechanism proposed a model for the formation of annular pores in a specific lipid composition consisting of POPC/DOPE/cholesterol, using avenacin A1 as the saponin (Armah et al., 1999). In this mechanism, the hydrophilic interaction between the saponins’ sugar moieties caused the aggregation of saponins and cholesterol. This aggregation subsequently led to the formation of pores in the bilayers, adopting a toroidal shape. The third mechanism involves α-hederin, another type of saponin, and describes a concentration-dependent permeabilization process. Under the concentrations of the CMC, the α-hederin saponins could bind to the outer-layer of the membrane (Lorent J. H. et al., 2014). This binding creates an area discrepancy and curvature between the inner and outer monolayers, leading to the formation of vesicles. As the concentration of α-hederin increases, the saponin, cholesterol, and phospholipids aggregate, forming wormlike structures within the membrane. These aggregates cause transient defects in the membrane and gradual permeabilization. The dimension of the sugar chain at the C3 position of the triterpenoid ring influences the formation of domains and the speed of permeabilization. The saponin of α-hederin could form pores in the membrane directly, resulting in the loss of structure of the layer material with concentrations exceeding the CMC. This suggests that micelles or aggregates could be formed by transferring large amounts of saponins near the membrane, which can directly interact with the membrane architecture.

The combination of saponins with cholesterol-enriched domains. Immediately results in the change in membrane permeabilization and the formation of increasingly larger cores (Korchowiec et al., 2015). Specifically, α-hederin showed a higher tendency to accumulate at the pore’s rim, stabilizing it through a reduction in line tension, by virtue of its amphiphilic nature. According to this model, the external monolayer experiences a positive curvature strain. The presence of two hydrophilic sugars in the saponin molecule gives it an axe-like shape, resulting in positive curvature stress in the trans bilayer direction. This stress leads to the formation of worm-like aggregates or macroscopic pores. The model considers the concentration dependent self-aggregation of saponins, their three-dimensional shape, affinity for cholesterol, and amphiphilicity.

The foregoing exposition elaborates in detail on the manner in which saponins interact and bind with cell membrane cholesterol. Nevertheless, the preponderance of extant literature primarily centers on the validation of these effects within in vitro experiments and proffers scant delineation of the specific merits, demerits, and applications of saponins. Whether this can furnish definite guidance for the anti-tumor application of saponins remains a subject that demands further exploration.

Saponins have shown good anti-cancer potential in many tumor cells by inhibiting cell growth and inducing cell apoptosis. Cytotoxicity for most saponins, seem to be ascribed to their ability to interact with membrane lipids, especially cholesterol.

Park et al. (2010) found that Rh2 induced lipid raft internalization, which caused cell membrane Akt protein inactivation and finally induced cell apoptosis. The study demonstrated that Rh2 can modulate the cell membrane cholesterol levels and influence cell viability regardless of cholesterol addition in MDA cells. Verstraeten et al. (2018) also reported that membrane cholesterol can delay the activity of ginsenoside Rh2, suggesting that saponin cytotoxicity is linked to its interaction with membrane cholesterol. Changhong et al. (2021) indicated that Ginsenoside Rb1 could inhibit the toxicity of A-β25-35-induced pheochromocytoma (PC12) cells, and could produce protective effects, including inhibiting cell growth and inducing cell apoptosis. The mechanism proposed that ginsenoside Rb1 could serve as an agonist of peroxisom proliferator-activated receptor-γ (PPARγ) and therefore decrease the content of cholesterol. Additionally, other saponins have been shown to have similar effects. Polyphyllin D (PD) (Chen et al., 2023), a steroidal saponin extracted from the rhizomes of Paris polyphylla, has been proven to induce G2/M phase arrest and inhibit the growth of liver cells by disrupting cholesterol biosynthesis. Avicins D, a triterpenoid saponin, has been demonstrated by Xu et al. (2009) to induce cell apoptosis in cell lines lacking cell death receptors. Upon avicin D treatment, Fas translocate to lipid rafts containing cholesterol, where it associates with Fas-associated death domain (FADD) and Caspase-8 to form the death-inducing signaling complex (DISC). Lorent J. et al. (2014) further investigated the role of cholesterol and saponin structure in apoptosis and membrane permeabilization in two malignant monocytic cell lines, and showed that the activity of α-hederin is primarily determined by its interaction with membrane cholesterol and subsequent pore formation.

The metastasis of cancer cells is a cascade of steps with biological cell progressions including dissemination and migration (Xiao et al., 2023), tumor angiogenesis, tumor microenvironment regulation and epithelial-mesenchymal transition, etc. Recently, saponins have been reported to be involved in the steps that constitute the cascade reaction, which appear amenable to the treatment of cancer metastasis. This is closely related to the interaction with cholesterol (Liu et al., 2021).

Barthomeuf et al. (2004) discovered that Hederacolchiside-A1 (Hcol-A1), a triterpenoid saponin from Hedera colchica Koch, has anti-melanoma activity (Kim et al., 2023) and might affect the endothelial cell network communication system required to inhibit tumor metastasis. Hcol-A1’s activity was found to be critical with plasma membrane cholesterol sequestration. Zhu et al. (2022) discovered that Pulsatilla saponin E (PSE) effectively decreased NSCLC cell survival, migration, and invasion while promoting apoptosis. This was achieved by adjusting apoptosis-related proteins, reducing the levels of free cholesterol (FC) and total cholesterol (TC), and downregulating the expression of flotillin-1, flotillin-2, Akt, and FASN in a concentration-dependent manner. However, the inhibitory impact of PSE on A549 cell survival, migration and invasion was reversed by overexpressing flotillin-2 in lipid rafts (LR), which was closely related to membrane cholesterol levels.

Chemotherapeutics and small-molecule targeted drugs are effective treatments for cancer. However, cancer cells can develop resistance to these drugs. The development of multidrug resistance is a major reason why cancer chemotherapy and targeted therapy fail. Interestingly, the reversal of drug resistance has been linked to the cholesterol content in the cell membrane (Yan et al., 2020; Kopecka et al., 2020). Controlling the cholesterol levels in tumor cells has emerged as a potential strategy for combating multidrug resistance in cancer and improving therapeutic outcomes. Saponins have been shown to have varying degrees of effectiveness in reversing drug resistance to chemotherapy drugs and small-molecule targeted drugs. This effectiveness is closely related to their ability to regulate cholesterol levels. Therefore, saponins have the potential to be used in combination therapies to reverse drug resistance.

It is well known that disruption of lipid rafts, which are cholesterols-rich microdomains, could alter membrane functions, suggesting that cholesterol plays a vital role in raft function (Ye et al., 2019). Yun et al. (2013) have proven that Ginsenoside Rp1 can reverse resistance to actinomycin D by reducing MDR-1 protein levels and Src phosphorylation while modulating lipid rafts, which has a better therapeutic efficacy compared to the positive drug MβCD. The addition of cholesterol attenuated the aggregation of lipid rafts induced by Rp1 and the redistribution of MDR-1. Similarly, Kwon et al. (2008) made a significant discovery by demonstrating that Rg3, a metabolite found in ginseng, has the unique ability to selectively reduce cholesterol dependent membrane fluidity and impede the accumulation of P-gp mediated drugs in multi-drug resistance (MDR) cells, while not affecting drug-sensitive cells. However, due to the hemolytic problem of saponins, the single-agent administration mode of saponins may lead to a relatively low bioavailability in vivo. Additionally, Ginsenosides Rg1 and CK (Qiu et al., 2022) could lead to cholesterol efflux and decrease intracellular cholesterol content by upregulating LXRαin in TMZ-resistant GBM cells, which redistributed lipid rafts and effectively regulated the resistance of GBM cells to TMZ. Polyphyllin I (Yang et al., 2018), known for its ability to coagulate cell membrane cholesterol, was reported to modulate the MALAT1/STAT3 signaling pathway, leading to apoptosis in gefitinib-resistant NSCLC cell. However, Polyphyllin I possesses relatively high toxicity and hemolytic property, which poses limitations when applied in vivo. Wang et al. (2022) discovered that Polyphyllin VII demonstrates potent and selective anti-multiple myeloma both in-vitro and in-vivo by binding to the cell membrane protein-moesin, resulting in a decrease in its protein levels and subsequently inhibiting the Wnt/β-catenin pathway. Consequently, Polyphyllin VII reduces the surplus population of cells and overcomes bortezomib resistance.

Targeting toxins through ribosome inactivating proteins (RIPs) (Smith et al., 2017) has become a topic of extensive research, but their successful clinical applications as therapeutic drugs are still limited to date. One of the major factors limiting the effectiveness of these protein-based therapies is their internalization efficiency through receptor-mediated endocytosis (RME), as well as their subsequent release efficiency from lysosomal compartments into the cytoplasm, where toxin components can catalyze the action on target ribosomes. Poor internalization of the target toxin and then recycling back to the cell surface or partial transport of the toxin to the lysosome and subsequent degradation may play a role (Weng et al., 2012; Thakur et al., 2013). In a decade, the endoplasmic escape activity of triterpenoid saponins has been studied as a potential powerful tool for improving cytoplasmic penetration of protein drugs through endocytosis, thereby greatly enhancing their pharmacological effects. More importantly, cholesterol seems to play a central actor in the enhancement of toxin cytotoxicity by triterpenoid saponins (Cao et al., 2020).

Smith et al. (2017) studied the effects of plasma membrane and cellular cholesterol on the enhancement and dissolution properties of saponins on Daudi human lymphoma cell lines. The results indicated that when lipid-deprived Daudi cells supplemented their cholesterol with low-density lipoprotein (LDL), the effect of SA on enhancing BU12 SAP IT cytotoxicity depends on cholesterol (Wensley et al., 2021). Meanwhile, they suggested that SA-induced lysosome escape may depend more on the curvature of the cell membrane and the actual cholesterol content in the bilayer, and therefore how this is also influenced by membrane phospholipid content. Wang et al. (2023) found that Momordin Ic (MIC) from Kochia scoparia (L.) Schrader significantly enhanced the cytotoxicity of recombinant MAP30, a type I ribosomal inactivation protein derived from bitter melon, in breast cancer cells, including triple cancers. The presence of cholesterol and ganglioside GM1 in the cell membrane were shown to be key factors in enhancing the cytotoxicity of MAP30, while the endosomal escape activity of MIC was less important. Raddeanin A (Kang et al., 2022), an oleanane-type triterpenoid saponin, has been shown to significantly promote endosome escape because it can recruit galactose-9, an endosome escape event reporting protein. As expected, RA effectively enhanced the anti-tumor effect of MAP30. Meanwhile, the effect of RA is related to the intracellular pH value and cell membrane cholesterol content, as well as cell type indolence.

As a novel form of regulated cell death, ferroptosis triggers the production of reactive oxygen species (ROS) and lipid peroxidation resulting from iron overload. Consequently, the molecules that directly or indirectly modulate these mediators are expected to regulate the process of iron-induced cellular demise. Glutathione (GSH) plays a pivotal role in the antioxidant system, safeguarding against lipid peroxidation, while SLC7A11-mediated cystine uptake enhances GSH synthesis. Therefore, alterations in GSH levels and solute carrier family proteins (SLC7A11 and SLC40A1) serve as commonly employed indicators for assessing iron mediated death. Recent evidence suggests (Liu et al., 2021) that lymphoma cells reliant on lipoprotein-mediated cholesterol uptake also undergo iron mediated death—a mechanism of cell demise dependent on oxygen and iron—triggered by the accumulation of oxidized lipids within the cellular membrane unless glutathione peroxidase 4 (GPX4), an enzyme responsible for reducing these toxic lipids (Forcina and Dixon, 2019), intervenes.

Nowadays, with the increasing research in the mechanism of ferroptosis, more studies have found that saponins can induce ferroptosis through their unique regulation effect on cholesterol. Zhou et al. (2023) found that Tim-AIII, a steroid saponin, is able to decrease cholesterol levels in cancer cells. Further evidence has also shown that it is able to form a complex with HSP90 protein to target and degrade the expression of GPX4, which ultimately leads to induced ferroptosis in lung cancer cells. And Tim-AIII possessed relatively good safety when applied in vivo. Lu et al. (2014) found that Ruswcogenin could positively influence fatty liver changes in type 2 diabetes and in the insulin resistant state. Hence, based on its cholesterol-lowering effect, Song et al. (2020) further proved that Ruscogenin induced ferroptosis by regulating the levels of transferrin and ferroportin. Meanwhile, saponins that have coagulative properties and regulate cell membrane cholesterol have also been reported to be related to the mechanism of ferroptosis. For instance, Wu et al. (2022) showed that α-hederin could lead to ferroptosis and membrane permeabilization in non-small cell lung cancer at safe, low toxic doses by disrupting the GSS/GSH/GPX2 axis-mediated GSH redox system. Xie and Chen (2021) found that dioscin has certain effects on the expression of transferrin and ferroproteins, which could regulate iron level in cells, and thus induce ferroptosis.