New biofunctional effects of oleanane-type triterpene saponins

In the current review, we describe the novel biofunctional effects of oleanane-type triterpene saponins, including elatosides, momordins, senegasaponins, camelliasaponins, and escins, obtained from Aralia elata (bark, root cortex, young shoot), Kochia scoparia (fruit), Polygala senega var. latifolia (roots), Camellia japonica (seeds), and Aesculus hippocastanum (seeds), considering the following biofunctional activities: (1) inhibitory effects on elevated levels of blood alcohol and glucose in alcohol and glucose-loaded rats, respectively, (2) inhibitory effects on gastric emptying in rats and mice, (3) accelerative effects on gastrointestinal transit in mice, and (4) protective effects against gastric mucosal lesions in rats. In addition, we describe (5) suppressive effects of the extract and chakasaponins from Camellia sinensis (flower buds) on obesity based on inhibition of food intake in mice. The active saponins were classified into the following three types: (1) olean-12-en-28-oic acid 3-O-monodesmoside, (2) olean-12-ene 3,28-O-acylated bisdesmoside, and (3) acylated polyhydroxyolean-12-ene 3-O-monodesmoside. Furthermore, common modes of action, such as involvements of capsaicin-sensitive nerves, endogenous NO and PGs, and possibly sympathetic nerves, as well as common structural requirements, were observed. Based on our findings, a common mechanism of action might mediate the pharmacological effects of active saponins. It should be noted that the gastrointestinal tract is an important action site of saponins, and the role of the saponins in the gastrointestinal tract should be carefully considered.


Introduction
Most saponins, characterized by triterpenes or steroids with oligoglycoside linkage, are well-known to exert soap-like foaming, fish toxicity, and hemolysis, however herbs with high saponin content have been used as traditional medicines, dietary supplements, and development of cosmetics and drugs. Recently, the development of isolation techniques such as preparative HPLC, along with instruments for elucidating chemical structures such as high-resolution NMR, MS, and X-ray analyses, have facilitated the isolation and elucidation of chemical structures of saponins in a short period and from minimal quantities. Therefore, the chemical structures of numerous saponins have been elucidated. However, their biofunctional effects and the mechanism of actions remain poorly explored, except for saponins contained in several important natural medicines such as glycyrrhiza [1], ginseng [2], and bupleurum root [3].
We have previously elucidated the active constituents of natural medicines considering the aspects of chemistry and pharmacology, revealing activities against allergy and inflammation, diabetes, obesity, and proliferation and metastasis of tumor cells. During our studies, we explored a large number of triterpene saponins, comprising triterpene as a sapogenol, from various natural medicines and medicinal foods and identified interesting biofunctional effects.
In the present review, we summarize the oleanan-type triterpene saponins isolated from Aralia elata Seem. (horse chestnut seed), with suppressive or delayed effects on elevated blood ethanol and glucose levels in rats and gastric emptying (GE) in rats and mice, accelerative effects on gastrointestinal transit (GIT) in mice, protective effects on gastric mucosal lesions in rats, and anti-inflammatory activity in mice. Furthermore, we describe the anti-obesity and suppressive effects of food intake mediated by an extract and chakasaponins from Camellia sinensis (L.) O. Kuntze (flower buds) (tea-flower) in mice [4].

Effects on increased blood ethanol levels after ethanol loading in rats
Alcoholism is a major health problem globally and has been associated with considerable physiological and social challenges. Chug-a-lugging of alcoholic drinks ('Ikkinomi' in Japanese) can induce acute alcohol toxicity with acidosis, heart failure, and respiratory depression caused by autonomic nerve and cerebrum dysfunction. Long-term alcohol consumption in large quantities can induce numerous disorders, such as hepatopathy, gastrointestinal disorders, chronic pancreatitis, peripheral nerve disorder, and hypertension. Therefore, inhibitors of alcohol absorption may exert potential preventive effects against acute and chronic alcoholism.
To identify compounds that could suppress elevated blood ethanol levels, we performed ethanol loading and explored the potential of saponins from various natural medicines, including those traditionally employed for detoxification against ethanol poisoning. For screening, the test samples were orally administered to fasted rats. After 1 h, 20% aqueous ethanol (5 mL/kg) was orally (p.o.) or intraperitoneally (i.p.) administered, and blood samples were collected at 1, 2, and 3 h after ethanol loading, followed by detection of blood ethanol levels. Saponin fractions of A. elata (bark root cortex and young shoot), K. scoparia (fruit), senega, C. japonica (seeds), and horse chestnut seeds were found to suppress the elevated blood ethanol levels in ethanol-loaded rats.
Following i.p. administration of 20% aqueous ethanol to rats, 16 failed to decrease blood ethanol levels. Although the inhibitory mechanism of oleanolic acid 3-O-monodesmosides on elevated blood levels remains poorly understood, we speculated that these monodesmosides decrease the blood ethanol concentration by suppressing absorption across the cell membranes of the digestive tract or delaying absorption by inhibition of GE, as described in the section of the mode of action, but not by acceleration of ethanol metabolism. Based on the pharmacological assessments, these oleanolic acid 3-O-monodesmosides exerted more potent inhibitory activity than the olean-12-ene 3,28-O-acylated bisdesmosides (senegins and senegasaponins) and the acylated polyhydroxyolean-12-ene 3-O-monodesmosides (camelliasaponins and escins), as described in the following sections.

Effects on elevated blood glucose levels after sugar loading in rats
The bark and root cortex of A. elata ('Taranoki' in Japanese) have been used for tonic, antiarthritic, and antidiabetic purposes in Chinese traditional medicines. It has been suggested that saponins isolated from this medicinal plant could reduce blood glucose levels. In the screening test, d-glucose (0.5 g/ kg, p.o.) or sucrose (1 g/kg, p.o.) was orally loaded after administering the test sample in fasted rats. Blood samples were collected at 0.5, 1.0, and 2.0 h after sugar administration, and serum or plasma glucose levels were determined.
Furthermore, the fresh roots and leaves of sugar beet (Beta vulgaris L.) showed potent inhibitory effects on elevated serum glucose levels in glucose-loaded rats. We isolated 10 saponins, called betavulgarosides I-X, with novel dioxolane-type or acetal-type substituents, both of which are presumed to be biosynthesized through an oxidative degradation process of a terminal monosaccharide moiety, and their inhibitory effects on elevated plasma glucose levels have been reported in glucose-loaded rats [17][18][19]. Six new oleanane-type triterpene saponins (gymnemoside-a, -b, -c, -d, -e, -f) were isolated from the leaves of Gymnema sylvestre R. Br. Gymnemic acids III and V exhibited weak inhibitory effects on elevated serum glucose levels in glucose-loaded rats. Furthermore, the inhibitory effects of gymnemosides-c, -d, -e, and -f and principal triterpene glycosides (gymnemic acids I-V) were examined on d-[U-14 C] glucose (2 mM) uptake in rat small intestinal fragments at 30℃ for 6 min. Several saponins, including gymnemic acids II and IV, as well as oleanolic acid 3-O-glucuronide (16) and escin Ia (55), inhibited glucose uptake at 0.5 mM [20,21].

Mode of action of inhibitory effects on elevated blood glucose levels
Next, the mode of action through which active saponins mediated their inhibitory effects on blood glucose elevation was examined using oleanolic acid 3-O-monodesmosides [momordin Ic (14), oleanolic acid 3-O-glucuronide (16)] [22].

Effects on blood glucose levels of normal rats, intraperitoneal glucose-loaded rats, and alloxan-induced diabetic mice
The regulation of serum glucose is controlled by several factors, such as the secretion and release of hormones (e.g., insulin and glucagon), transport of sugar in the digestive tract, and glucose absorption via membranes of the small intestine.
Tolbutamide can increase insulin secretion to reduce serum glucose levels in normal and glucose-loaded rats. Both saponins (14,16) dose-dependently inhibited elevated serum glucose levels in oral glucose-loaded rats. However, 14 and 16 at 50 mg/kg did not decrease serum glucose levels in glucose-untreated (normal) rats nor serum glucose elevation in intraperitoneal glucose-loaded rats. Insulin (1 U/kg, i.p.) strongly reduced the serum glucose levels 1 and 1 3 3 h after intraperitoneal injection in alloxan-induced diabetic mice. However, 14 and 16 (l00 mg/kg) lacked hypoglycemic effects. These results indicate that 14 and 16 have neither insulin-like activity nor insulin-releasing activity like tolbutamide, and we speculated that they impact on glucose absorption in the gastrointestinal tract [22].

Effects on GE in rats
Effects on GE of rats were examined using the phenol red method. The reference drug, atropine sulfate (10 mg/kg, p.o.), can significantly inhibit GE in rats 0.5, l, and 2 h after oral administration. Momordin Ic (14) and oleanolic acid 3-O-glucuronide (16) (50 mg/kg, p.o.) strongly inhibited GE [22]. The GE suppression mediated by 14 and 16 appears critical to suppress the increased serum glucose levels after oral glucose loading.
Phlorizin is well-known as an inhibitor of the Na + / glucose co-transport system at the intestinal brush border membrane. Both saponins inhibited glucose uptake in rat small intestine fragments in vitro, similar to phlorizin. Based on the above evidence, saponins, such as 14 and 16, could inhibit the elevated serum glucose levels in oral glucoseloaded rats by suppressing glucose transfer from the stomach to the small intestine, the main site of glucose absorption, and partly by inhibiting glucose transport at the intestinal brush border membrane [22].
Regarding the mode of action of escins Ia (55) and Ila (57) and E,Z-senegins II (mixture of 35 and 36) in mediating inhibitory effects on elevated serum glucose levels in oral glucose-loaded rats, the results were similar to those of oleanolic acid 3-O-glycosides (14 and 16) [23].
There is insufficient evidence to support that active saponins potently inhibit intestinal absorption in vivo, although they do decrease the increased blood glucose levels in rats by delaying glucose absorption, primarily by inhibiting GE and partly by suppressing the intestinal glucose transport system. It has been reported that reducing postprandial hyperglycemia is an effective strategy for preventing and treating non-insulin-dependent diabetes mellitus. Therefore, these active saponins could also effectively prevent and treat diabetes.
Next, we explored the detailed action of oleanolic acid glycosides and escins Ia-IIb (55-58) on GE using mice.
Capsaicin is widely used to ablate sensory C fibers. It has been systematically used to ablate all capsaicin-sensitive C fiber. Hyperglycemia in streptozotocin-induced hypoinsulinemic rats can reduce the sensitivity of the sympathetic nervous system [25,26]. In our study, the inhibitory effect against GE in 1.5% CMC-Na test meal-loaded mice was potentiated by glucose [2 g/kg, intravenously (i.v.) or 5 g/kg, i.p.] but markedly attenuated by pretreatment with alloxan (50 mg/kg, i.v.) and streptozotocin (100 mg/kg, i.v.), in which the activity of sympathetic nervous system might be decreased, or by insulin [1 or 3 U/kg, subcutaneously (s.c.)]. The effect of insulin (1 U/kg) was markedly reduced by glucose (2 g/kg, i.v.), which can be directly utilized by the brain, but not by fructose (2 g/kg, i.v.), which cannot be used by the brain [27]. GE is also enhanced by signals from chemoreceptors in severe hypoglycemia, allowing the rapid passage of nutrients through the stomach for immediate digestion and absorption [28]. Pretreatment with capsaicin (75 mg/kg in total, s.c.) could attenuate the effect of momordin Ic (14). These results suggest that GE inhibition mediated by 14 is relative to serum glucose levels and partially mediated by capsaicin-sensitive sensory nerves and the central nervous system [27].
Escins Ia-Ilb (55-58) (12.5-200 mg/kg) inhibited GE of a 1.5% CMC-Na test meal (11.1-52.8%). Treatment with 55-58 (50 mg/kg) also inhibited GE of a 40% glucose test meal (21.1-23.5%) except for escin Ia (55), a milk test meal (18.4-33.1%), and a 30% ethanol test meal (GE in the control group was ca. 70% and inhibitions were 13.5-15.9%). Pretreatment with streptozotocin (100 mg/kg, i.v.), capsaicin (75 mg/kg in total, s.c.), or insulin (1 U /kg, s.c.) could attenuate the effects of 55-58 on GE of the CMC-Na test meal. The effect of insulin was reduced by glucose (2 g/ kg, i.v.), which can be directly used by the brain, but not by fructose (2 g/kg, i.v.), which cannot be utilized by the brain. The inhibitory effects of 55-58 (50 mg/kg) could not be observed on the GE of 60% ethanol test meal, in which the central nervous system was suppressed by ethanol. Accordingly, capsaicin-sensitive sensory nerves and the central nervous system may partially mediate the effects of 55-58 [29]. Furthermore, the GE inhibitory effects mediated by 55-58 (25 mg/kg) were markedly attenuated following pretreatment with indomethacin, an inhibitor of prostaglandins (PGs) biosynthesis, suggesting the involvement of endogenous PGs in GE inhibition [30].
Dopamine (DA) is a major neurotransmitter in the central nervous system. DA is also found in large concentrations in the stomach and is suggested to be involved in controlling GE in rats. Escin Ib (56, 25 mg/kg, p.o.) mediated GE inhibition was attenuated following pretreatment with a single bolus of dl-α-methyl-p-tyrosine methyl ester (an inhibitor of tyrosine hydroxylase), reserpine (a catecholamine depletor), 6-hydroxydopamine (a dopamine depletor). Furthermore, pretreatment with centrally-acting DA 2 receptor antagonists (e.g., spiperone, haloperidol, metoclopramide) attenuated the effect of 56. However, a peripherally-acting DA 2 antagonist, domperidone, exerted weak attenuation, whereas SCH23390 (a DA 1 receptor antagonist) did not [31]. These findings suggest that 56 could inhibit GE, at least in part, mediated via capsaicin-sensitive sensory nerves, to stimulate the synthesis and/or release of DA, to act through the central DA 2 receptor, which, in turn, causes PGs synthesis or release.

Effects on gastrointestinal transit in mice
Ileus is a common complication induced by various factors, such as laparotomy with manipulation and peritoneal irritation. Given the lack of specific therapy, ileus remains an important clinical challenge. Patients with ileus accumulate gas and secretions, leading to bloating, distention, emesis, and pain. Prokinetic drugs, such as cisapride, metoclopramide, erythromycin, and octreotide, are commonly used to combat chronic ileus. However, no medical therapy affords notable relief in advanced cases. Non-steroidal anti-inflammatory drugs, such as indomethacin, are known to block PG biosynthesis and are widely used for postoperative pain. These drugs have been shown to afford beneficial effects in treating postoperative ileus in rodents, although undesirable side effects have also been documented. Recently, a Kampo preparation, daikenchuto (大建中湯), has been used clinically to treat ileus post-abdominal surgery, and the effective constituents and detailed mechanisms of action have been revealed [32].
Screening can be performed using a 5% charcoal suspension in a 1.5% CMC-Na solution intragastrically administered (0.2 mL/mouse) to conscious mice. Thirty minutes later, mice were sacrificed by cervical dislocation. The abdominal cavity was opened, and the gastrointestinal tract was harvested. The distance traveled by the front of charcoal suspension from the pylorus was measured and expressed as a percentage of the total length of the small intestine from the pylorus to the caecum. In this condition, GIT in the control group was ca. 50%. The test samples were administered orally 60 min prior to administering the charcoal suspension.
First, the effects of oleanolic acid glycosides on the GIT of ileus were examined in normal fasted mice.  [33].
Ileus was induced by peritoneal irritation or by laparotomy with manipulation. In our experiments, GIT could be suppressed by peritoneal injection of 1% acetic acid and laparotomy with manipulation; the GITs in the control groups were ca. 14 and 23%, respectively. Momordins Ic (14, 5-25 mg/kg) and I (17, 25 mg/kg) also significantly accelerated the reduced GIT induced by the intraperitoneal acetic acid injection, with acceleration rates of 109.2-246. 8 [33].
In this experiment using normal mice, the applied interval between the saponin fraction and charcoal suspension was set from 5 to 300 min. Interestingly, the saponin fraction (25 mg/kg) demonstrated significant effects 5 min after the oral administration, which persisted until 240 min. These findings suggest that the saponin act immediately after oral administration, with actions persisting for 4 h [34].
The GIT acceleration induced by 55-58 in normal mice was completely abolished by the pretreatment with streptozotocin but not by the pretreatment with capsaicin (75 mg/ kg in total, s.c.) or atropine (10 mg/kg, s.c.). Accordingly, these results suggest that the sympathetic nervous system, but not capsaicin-sensitive sensory nerves nor the cholinergic mechanism, mediates the GIT accelerations induced by 55-58, similar to oleanolic acid 3-O-glycosides (14, 16,  and 17). The GIT acceleration induced by 55-58 may be mediated by the release of endogenous PGs and nitric oxide (NO), as determined by the results of pretreatment with indomethacin and NO synthase (NOS) inhibitor [N G -nitrol-arginine methyl ester (l-NAME)] [30]. Furthermore, the GIT acceleration mediated by escin Ib (56, 25 or 50 mg/ kg, p.o.) was attenuated following pretreatment with 5-HT 2 receptor antagonists (e.g., ritanserin, ketanserin, haloperidol, spiperone), but not by 5-HT 3 or 5-HT 4 receptor antagonists (MDL72222, metoclopramide or tropisetron). A bolus of dl-p-chlorophenylalanine methyl ester (an inhibitor of 5-HT synthesizing enzyme, tryptophan hydroxylase) and reserpine (a 5-HT depletor), but not 6-hydroxydopamine (a dopamine depletor), could attenuate the GIT acceleration effects of 56 [35]. In addition, we reported that chakasaponin II (78), classified into the acylated polyhydroxyolean-12-ene 3-O-monodesmoside like escins, stimulated the release of 5-HT from intestinal fragments in vitro, as described in the section of tea-flower.
Collectively, these saponins stimulate the synthesis or release of 5-HT to act through 5-HT 2 receptors, possibly the 5-HT 2A receptor, which, in turn, causes the release of NO and PGs, thereby accelerating GIT at the intestinal tract. The potential therapeutic effects of saponins should be explored in clinical settings for preventing the inhibition of GIT, including ileus induced by peritoneal irritation.

Gastromucosal protective effects in rats
As a beneficial effect against alcohol toxicity, we examined the effects of oleanolic acid glycosides on ethanol-induced gastric mucosal lesions in rats. In addition, the effects of oleanolic acid glycosides on indomethacin-induced gastric lesions and gastric secretion in pylorus-ligated rats were examined. The lesions were characterized by multiple hemorrhage red bands (patches) of different sizes along the long axis of the glandular stomach. The total length (mm) or score of lesions of each rat was measured 1 h after the administration of 99.5% ethanol (1.5 mL/rat, p.o.) or 4 h after the administration of indomethacin (30 mg/kg, s.c.). Test samples were administered orally to fasted rats 1 h prior to treatment with ethanol, indomethacin, or pyloric ligation.
PGs and NO as well as capsaicin-sensitive neurons have been shown to participate in the gastric defense mechanism. Oxygen-derived free radicals and lipid peroxidation are associated with gastrointestinal lesions, and antioxidants prevent the lesions by various ulcerogens. Gastric mucosal sulfhydryls (SHs) including glutathione (GSH) act as antioxidants, and are important for maintenance of mucosal integrity in the stomach. Ethanol-induced gastric damage is also associated with a significant decrease in the mucosal SHs level such as GSH, and pretreatment with SH-blockers prevents the gastroprotection of SH-containing compounds [38].
The gastroprotective effects of 16 and 55-58 were attenuated following pretreatment with capsaicin, L-NAME (70 mg/kg, i.p.), and indomethacin (10 mg/kg, s.c.), but not by N-ethylmaleimide (10 mg/kg, s.c.), a SHs blocker. The effects of 55-58 were also attenuated in streptozotocin-induced diabetic rats. Based on these findings, it can be suggested that the gastroprotective effects of 55-58 on ethanol-induced gastric mucosal lesions are acid-independent, potentially mediated by endogenous PGs, NO, and capsaicin-sensitive sensory nerves. Furthermore, the sympathetic nervous system partly mediates these effects, although the underlying mechanism remains unclear [36,37,39].

Antipruritic and anti-inflammatory effects
The fruit of K. scoparia has been used to treat skin diseases and cutaneous pruritus in Chinese traditional medicine. Matsuda et al. reported that the 70% aqueous ethanol extract of this natural medicine and its principal saponin constituent, momordin Ic (14), exerts antiallergic, anti-inflammatory, and antinociceptive effects. Furthermore, the 70% aqueous ethanol extract and 14 exhibited antipruritic activity, as determined by its inhibitory effect on the compound 48/80-induced pruritic mouse model [40]. As a continuing study, we examined the antipruritic activity of oleanolic acid glycosides.
The saponin mixture 'escin' obtained from the seeds of the horse chestnut seeds could afford anti-inflammatory activity [42]; however, the anti-inflammatory effects of each pure saponin of escin had not been examined, given the incomplete isolation and structural determination of saponin constituents.
We explored the effects of escins Ia-Ilb (55-58) and desacylescins I and II on acute inflammation in animals. All escins   Desacylescins I and II (200 mg/kg) did not afford inflammatory effects [43].
Our hypothesis of the potential mechanism of action saponins was established based on the experimental results using various receptor inhibitors and activators, inhibitors of NO and PG biosynthesis, and a high dose of capsaicin to mainly damage the afferent vagal nerves; however, these agents act systemically and not selectively act at each tissue and nerve. Therefore, further experiments, such as determining the concentrations of catecholamines in the brain and intestine using microdialysis-HPLC methods and selective detection of afferent vagal nerves, need to be considered. The interpretation of our experimental results should be reconsidered, given that regulation of gastrointestinal movements has been updated in detail [28,38,44,45].

Structure-activity relationship
Here, we summarize the structural requirements of active saponins to suppress elevated blood alcohol and glucose levels in rats, inhibit GE and accelerate GIT in mice, and afford gastromucosal protection in rats, as well as antipruritic and anti-inflammatory activities; however, detailed structure-activity relationships of saponins remain poorly clarified based on our reported results.

Inhibition of elevated blood alcohol and glucose levels
Regarding the inhibition of elevated blood alcohol levels, the 3-O-glycoside moiety and 28-carboxyl group in oleanolic acid glycosides are essential for mediating this activity [5,10]. The 28-O-glucopyranosyl group or the 2′-O-β-dglucopyranosyl group can reduce the suppression of elevated blood alcohol levels.
Regarding the inhibitory effects of oleanolic acid glycosides on elevated blood glucose levels, oleanolic acid 3-O-monodesmosides and 3,28-O-bisdesmosides, such as 13 and 18, comprised of a 4′-O-l-arabinofuranosyl group, which tends to inhibit elevated plasma glucose levels; this action appears distinct from inhibitory effects mediated by oleanolic acid glycosides on elevated ethanol levels mediated [5,6,10]. The 2′-O-β-d-glucopyranosyl group of an oleanolic acid 3-O-monodesmoside [glycosides B (23)] also could reduce the inhibitory activity similar to the inhibition of elevated blood ethanol levels [16].
Similar to the inhibition of elevated blood ethanol levels, acyl groups of senegasaponins, senegins, and escins are essential for inhibiting elevated blood glucose levels, and the 2′-O-β-d-xylopyranosyl group of escins is responsible for mediating potent inhibitory activity; the 2′-O-β-dglucopyranosyl group reduces this inhibitory activity [11,12,14].

GE inhibition and GIT acceleration
The 3-O-monodesmoside structure and 28-carboxyl group are essential for GE inhibition in mice, and the 28-ester glucoside moiety and 2′-O-β-d-glucopyranosyl group reduce this inhibitory activity [24], similar to the inhibitory effects on elevated blood ethanol and glucose levels in rats. All escins Ia-IIb (55-58) could inhibit GE of a 1.5% CMC-Na test meal, a 40% glucose test meal, a milk test meal, and a 30% ethanol test meal-loaded mice, except for the effect of escin Ia (55) possessing the 2′-O-β-d-glucopyanosyl group on GE of 40% glucose test meal-loaded mice. The presence of the 2′-O-β-d-glucopyranosyl group did not markedly reduce activity [29].
As observed for GE inhibitory effects, the 3-O-monodesmoside structure and 28-carboxyl group of oleanolic acid glycosides are essential for accelerating GIT in mice, except for 28-O-deglucosyl-chikusetsusaponins IV (19) and V (21) [33]. Escins could accelerate the reduced GIT mediated by intraperitoneal acetic acid administration and laparotomy with manipulation, and the 21,22-acyl groups are essential for activity [34].

Gastromucosal protection
The 3-O-glycoside moiety of oleanolic acid glycosides was found to be essential for suppressing ethanol-induced gastric lesions, and the 28-ester glucoside could reduce this inhibitory activity. Furthermore, the 2′-O-β-d-glucopyranosyl group of the glucuronic acid part decreased the activity, similar to the effects on elevated blood alcohol and glucose levels [36]. The 21,22-acyl groups of escins are crucial for affording protection against ethanol-induced gastric lesions, similar to their function in mediating the other observed effects, while the 2′-O-β-d-glucopyranosyl group did not markedly reduce activity, as described in the section of GE and GIT [37].

Antipruritic and anti-inflammatory effects
Regarding the relationship between their chemical structures and activities, the 3-O-glycoside moiety and the 28-carboxyl group of oleanolic acid glycoside were found to be essential for exerting the antipruritic effects, similar to the effects described in other sections, and the 3-O-glucuronide showed more potent activity than the corresponding 3-O-glucoside [41].
The acyl groups of escins are essential for exerting antiinflammatory effects, as described in other sections. Furthermore, escins Ib-IIb (56-58) with either the 21-O-angeloyl group or the 2′-O-d-xylopyranosyl group showed more potent activities than 55 with both the 21-O-tigloyl and the 2′-O-β-d-glucopyranosyl groups [43].
Furthermore, common modes of action, such as involvements of capsaicin-sensitive nerves, endogenous NO and PGs, and possibly sympathetic nerves, as well as common structural requirements, were observed. Based on these findings, a common mechanism of action might mediate the pharmacological effects of active saponins.

Anti-obesity effects of tea-flower and appetite inhibition in mice
Tea prepared from leaves of the plant C. sinensis (tea leaves) has been used since ancient days for medicinal purposes and is now consumed as a popular beverage. Tea has been extensively explored for its beneficial health effects, such as reducing body weight, alleviating metabolic syndrome, preventing cardiovascular diseases and cancer, and protecting against neurodegeneration. Regarding the mechanisms responsible for benefits against metabolic syndrome, tea polyphenols such as (-)-epigallocatechin 3-O-gallate can reduce intestinal lipid absorption, as well as activate AMPactivated protein kinase in the liver, skeletal muscle, and adipose tissues. The activation of AMP-activated protein kinase decreases gluconeogenesis and fatty acid synthesis and increases catabolism, resulting in body weight reduction and alleviation of metabolic syndrome [46][47][48].
Herein, we focus on the anti-obesity effects of tea-flower collected in the Fujian province of China (Fujian Chaka) and saponin constituents with anti-appetite effects.
The effects of the methanol (MeOH) extract on body weight gain in high-fat diet-fed mice and an experimental animal of metabolic syndrome, TSOD (Tsumura Suzuki Obese Diabetic) mice, were examined [61]. The MeOH extract (500 mg/kg/day, p.o.) markedly inhibited body weight gain 9-14 days after administration to high-fat dietfed mice (Fig. 6A). After two weeks, treatment with the extract (500 mg/kg/day, p.o.) significantly suppressed liver weight (p < 0.05, 1.06 g vs. control 1.23 g), liver triglycerides (p < 0.01, 30.2 mg/g wet tissue vs. control 62.1 mg/g wet tissue) and the weight of visceral fat (p < 0.05, 1.70 g vs. control 2.73 g). After one week of administration, the extract (500 mg/kg/day, p.o.) also significantly suppressed body weight gain in TSOD mice (Fig. 6C). Three weeks later, a glucose tolerance test was performed by intraperitoneal injection of glucose. The MeOH extract (250 and 500 mg/ kg/day, p.o.) significantly suppressed increased plasma glucose levels 2 h after glucose loading. After four weeks, treatment with the extract (500 mg/kg, p.o./day) significantly suppressed liver weight (p < 0.01, 1.27 g vs. control 1.48 g), weight of visceral fat (p < 0.01, 3.67 g vs. control 5.23 g), and plasma total cholesterol levels (p < 0.05, 210.4 mg/dL vs. control 254.3 mg/dL).
We speculated that the potent reduction in body weight within one week of extract treatment could be primarily attributed to reduced food intake. Therefore, the effect of the extract on food intake was examined in high-fat diet-fed and TSOD mice (Figs. 6B, D). The extract inhibited food intake in a dose-dependent manner, and this effect was also observed in normal diet-fed mice; the total intake for 5 days in the MeOH extract-treated group (500 mg/kg/day, p.o.) was 19.3 g (p < 0.01) vs. 21.0 g in the control group, although no obvious toxic effect was observed except for body weight gain [61]. The n-BuOH-soluble fraction inhibited food intake at a dose of 250 mg/kg/day, p.o., but the EtOAc-and H 2 O-soluble fractions had no such effect when administered orally according to yield.
Regarding the effect of the n-BuOH-soluble fraction on appetite signals, the effects on hypothalamic mRNA expression of neuropeptide Y (NPY) and agouti-related protein (AgRP) were examined. NPY is an important regulator of body weight that mediates its effects on food intake and energy expenditure. Several neurons expressing NPY in the hypothalamus are found within the arcuate nucleus (ARC), with most co-expressing AgRP. Ablation of NPY/AgRP neurons in young mice was shown to reduce food intake and body weight, and intracerebroventricular (i.c.v.) injection of NPY potently stimulated food intake in adult rats [70]. In our study, the n-BuOH-soluble fraction administrated at 250 mg/kg for 4 days significantly suppressed NPY mRNA expression. These findings suggest that the n-BuOH-soluble fraction inhibited food intake by suppressing appetite signals.
Furthermore, a principal saponin, chakasaponin II (78) (50 mg/kg/day, p.o.), induced a suppressive effect on food intake and the hypothalamic expression of NPY mRNA levels, similar to the n-BuOH-soluble fraction. These results D Effects on food intake of the standard laboratory chow in TSOD mice.The test sample was administered orally once daily. Each value represents the mean with the standard error of the mean (S.E.M.) (n = 6-10, *p < 0.05 **p < 0.01). This graph was taken from reference [61] with a modification suggest that the saponins are active constituents of the extract. Furthermore, the desacyl derivative of 78, desacylfloratheasaponin B, failed to exert these suppressive effects, suggesting that the 21 and 22-acyl groups are critical for the activity, as observed for the other effects described in the section of structure-activity relationship.
Recently, an anti-cancer drug, cisplatin, and selective serotonin reuptake inhibitors (SSRIs) were found to inhibit food intake, and the involvement of 5-HT 2 receptors in appetite control has been reported. Activation of the 5-HT 2B receptor in gastric smooth muscle and the 5-HT 2C receptor in the hypothalamus can suppress appetite. 5-HT produced during cisplatin or SSRI treatment binds to various receptor subtypes and is likely to stimulate the 5-HT 2B and 5-HT 2C receptors. Stimulating the 5-HT 2B receptor decreases plasma ghrelin levels, suppressing the appetite signals via afferent vagal nerves [71][72][73]. Consistent with previous reports, 5-HT (1 mg/kg, i.p.) inhibited food intake in mice. We investigated 5-HT release from isolated ilea of mice and its tissue retention in vitro. Chakasaponin II (78) at 1.0 mM significantly enhanced 5-HT release into the medium and reduced tissue retention [61]. The concentration of 78 was relatively high for in vitro experimentation, but saponin concentrations are typically considered to be elevated in the intestinal tract, given that this type of compound is poorly absorbed [42,74]. In our preliminary investigation, more than 30% of 78 was retained in the small intestinal tract 1 h after oral administration to mice. Furthermore, the effects of the n-BuOHsoluble fraction and chakasaponin II (78) on food intake were notably reduced in capsaicin-pretreated mice in which the capsaicin-sensitive sensory nerves were desensitized by pretreatment with high-dose capsaicin (Fig. 7) similar to that observed with escins (55-58) [61].
Cholecystokinin (CCK) and glucagon-like peptide 1 (GLP-1) secreted from the intestinal I-cells and L-cells stimulate each receptor, and the signals are mediated through the afferent vagal nerves and nucleus tractus solitarius (NTS) to reduce the expression of NPY and AgRP, ultimately suppressing appetite. Stimulation of the 5-HT 2B receptor in the stomach via the 5-HT released from intestinal chromaffin cells inhibits the release of ghrelin, which stimulates the appetite through the afferent vagal nerves, and stimulation of the 5HT 2C receptor in the hypothalamus stimulates proopiomelanocortin (POMC) neurons to reduce appetite [70][71][72][73][74][75]. In our preliminary examination, chakasaponin II (78) increased plasma CCK and GLP-1 levels in mice. These findings suggest that their inhibitory effects on food intake were initiated by the excretion of CCK and GLP-1, and mediated via capsaicin-sensitive sensory nerves, probably the afferent vagal nerves (Fig. 8). Chakasaponins I-III (77-79) (50-100 mg/kg) inhibited plasma glucose levels after sucrose loading in mice without inhibiting intestinal α-glucosidase and suppressed GE, similar to escins. CCK and GLP-1 were shown to inhibit GE [76,77], suggesting that CCK and GLP-1 release also participates in GE inhibition meditated by saponins such as chakasaponins and escins [29,59].
As described in the section of effects on GE in mice, the inhibitory effects of escins on GE involved DA release and DA 2 receptors via mechanisms involving capsaicinsensitive sensory nerves, probably certain vagal afferent nerves [29,31]. Tominaga et al. reported that 5-HT inhibits GE in rats [78]. Consistently, 5-HT (10 mg/kg, i.p.) significantly inhibited GE under our experimental conditions, although the effective dose of 5-HT was higher than that for food intake. Pretreatment with capsaicin partly reduced the inhibitory effects of chakasaponins I (77) and II (78) on Fig. 7 Effects of the n-BuOH-soluble fraction and chakasaponin II (78) on food intake of standard laboratory chow in normal mice and/or capsaicin-pretreated mice. Male ddY mice were fed a standard laboratory chow (MF, Oriental Yeast Co., Ltd.) for 8 days. The test sample was administered orally once daily. Each value repre-sents the mean for 5 or 6 mice. Significantly different from the control, **p < 0.01, and from the corresponding capsaicin-treated group, † p < 0.05. This graph was taken from reference [61] with a modification.

3
GE, suggesting that certain afferent vagal nerves, at least in part, participate in the inhibition of GE and food intake [61]. Bugajski et al. have observed that long-term vagal electrical stimulation could reduce food intake and body weight in rats [79]. Therefore, in addition to the release of 5-HT, CCK, and GLP-1, other mechanisms of action, including the direct stimulation of the vagal afferent nerves by saponins, should be explored.
Based on our experimental evidence, various health/ functional foods and beverages prepared from tea-flower have been recently developed in Japan. Recent clinical studies have demonstrated that food comprising the tea-flower extract could effectively reduce postprandial blood triglyceride levels and body fat [80,81].
Regarding anti-obesity effects of the extract highly containing oleanolic acid glycosides, Han et al. reported that the ethanol extract of K. scoparia fruit prevented obesity induced by a high-fat diet for 9 weeks in mice. Briefly, the ethanol extract of K. scoparia fruit prevented the increases in body weight and parametrial adipose tissue weight induced by the high-fat diet. Furthermore, consumption of a highfat diet containing 1% or 3% extract significantly increased the fecal content and the fecal triglyceride levels at day 3. The ethanol extract (250 mg/kg/day) and total saponins (100 mg/kg/day) of K. scoparia inhibited increased plasma triglyceride levels 2 or 3 h after the oral administration of the lipid emulsion. Total saponins, momordin Ic (14), 2′-O-β-dglucopyranosyl momordin Ic and 2′-O-β-d-glucopyranosyl momordin IIc inhibited the pancreatic lipase activity (in vitro). They concluded that the anti-obesity actions of K. scoparia extract in mice fed a high-fat diet might be partly mediated through delaying the intestinal absorption of dietary fat by inhibiting pancreatic lipase activity [82].  8 Appetite signals in the gastrointestinal-brain system. NPY neuropeptide Y, AgRP agouti-related protein, MSH melanocyte-stimulating hormone, POMC proopiomelanocortin, NTS nucleus tractus solitarius, ARC arcuate nucleus, PVN paraventricular nucleus, LHA lateral hypothalamic area, CCK cholecystokinin, GLP-1 glucagon-like peptide 1. This figure is taken from reference [85] with a modification It should be noted that the gastrointestinal tract is an important action site of saponins, with rapid action observed before absorption after hydrolysis by intestinal bacteria; however, further pharmacological effects of various sapogenols of saponins should be investigated. As described in the reports that several flavonoid glycosides, but not aglycone, and certain carbohydrate chains in polysaccharides activated the immunity in the intestinal tract [83,84], the role of the glycosides including saponins in the gastrointestinal tract should be carefully considered.