Pancreatic lipase was purchased from Sigma Chemical Co. (St Louis, MO). Triglyceride E- and Total Cholesterol E-test kits were purchased from Wako Pure Chemical Co. (Osaka, Japan). Laboratory pellet chow was purchased from CLEA Japan (Osaka, Japan). Beef tallow, casein, vitamin and mineral mixtures were purchased from Oriental Yeast Co. Ltd (Tokyo, Japan). Orlistat (a lipase inhibitor) was purchased from Hong Kong Market. Other chemicals were of reagent grade.

The rhizomes of P. japonicus were obtained from Tochimoto Tennkaido Co. Ltd. (Osaka, Japan). Voucher specimens (No. PJ020817) are deposited in the Faculty of Environmental and Symbiotic Sciences, Prefectural University of Kumamoto and the Second Department of Medical Biochemistry, School of Medicine, Ehime University.

Chikusetsusaponins were isolated using by a modification of previous reported method 1516. The dried rhizomes of P. japonicus (1 kg) were extracted three times by refluxing with methanol (3 L). After removal of the solvent from the methanol solution under reduced pressure, the extract (370 g) was subjected to chromatography on reversed-phase highly porous polymer [DIAION HP-20 (4 kg), Mistubishi Chemical Ind. Ltd.: elution with H₂O (800 mL), MeOH (500 mL), and CHCl₃ (300 mL)] to provide the three fractions such as H₂O eluate (235 g), MeOH eluate (130 g), and CHCl₃ eluate (5 g). The fraction eluted with methanol was designated the "total chikusetsusaponins". The MeOH eluate (total saponin fraction, 110 g) was separated by column chromatography on silica gel [silica gel G (2 kg), Merck, CHCl₃-MeOH-H₂O] to furnish four fractions and the fractions 2 and 3 were purified by HPLC (YMC-pack ODS, MeOH-H₂O-AcOH) to give chikusetsusaponins III (13.5 g), IV (6.2 g) and V (54.5 g), 28-deglucosylchikusetsusaponins IV (15 mg) and V (110 mg). 28-Deglucosylchikusetsusaponins IV and V were also obtained by alkaline hydrolysis of chikusetsusaponins IV and V, respectively. The isolated five compounds were identified by comparison of their physical data ([α]_D, IR, ¹H- and ¹³C-NMR) with reported values 1516. The purity of each compound was over 95% based on the analysis by HPLC (Fig. 1).

Previously, we reported that varying the casein concentration (22, 31, 34 and 36 % diets) in high-fat diet containing 40 % beef tallow did not affect body weight or parametrial adipose tissue weight 17. Therefore, we substituted 1 or 3% total chikusetsusaponins isolated from P. japonicus rhizomes for an equivalent mass of casein in the high-fat diets shown in Table 1. To avoid autooxidation of the fat components, food was stored at -30°C. If the diet except fat from high-fat diet was used as control diet, carbohydrate (e.g. corn starch or sugar) instead of fat was added to the diet; and consequently this diet is high-carbohydrate diet. Black et al. reported that feeding high-carbohydrate diet containing 73% kcal corn starch for 16 weeks elevated the α-disaccharidase activity in small intestine 18. Therefore, it seems likely that the high-carbohydrate diet may affect carbohydrate metabolism. In this study, we used laboratory chow diet as a control diet.

Female ICR strain mice (3 weeks old) and male Wistar King strain rats (6 weeks old) were obtained from CLEA Japan (Osaka, Japan) and Charles River Japan (Yokohama, Japan), respectively, and housed for 1 week under a 12 h/12 h light/dark cycle in a temperature- and humidity-controlled room.

The animals were given free access to food and water. After adaptation to the lighting conditions for 1 week, the healthy animals were used in these experiments. The experimental protocol was approved by the Animal Studies Committees of Ehime University and Kumamoto Prefectural University.

Female ICR mice (3 weeks old) were divided into four groups that were matched for body weight, after 1 week of being fed laboratory pellet chow ad libitum. The control group continued to be fed laboratory pellet chow ad libitum. The mice consumed the high-fat diet or the high-fat diet containing 1 and 3% total chikusetsusaponins or 0.012 and 0.024% orlistat (positive control) for 9 weeks. The body weight of each mouse was measured once a week and the total amount food consumed was recorded 3 times per week. After the mice had been fed these diets for 9 weeks, blood was taken from each mouse by venous puncture under anesthesia with diethyl ether; the mice were then killed with an overdose of diethyl ether. Experiments were performed in a ventilated room. The plasma was prepared and frozen at -80°C until analysis. The liver and parametrial adipose tissue were dissected and weighed. Liver triacylglycerol and total cholesterol concentrations were measured using Wako Triglyceride E-Test and Total Cholesterol E-Test kits.

Female ICR mice (3 weeks old) were housed for 1 week in a room maintained at 25 ± 1°C with 60% relative humidity and given free access to standard laboratory pellet chow and water. The mice consumed the high-fat diet or the high-fat diet containing 1% or 3% total chikusetsusaponins, or 0.025% orlistat for 3 days. Wet weight and triacylglycerol content in the feces obtained during the last 24 h were measured using the Wako Triglyceride E-Test kit.

Male Wistar rats were also housed for 1 week in the above same conditions. After rats had been deprived of food overnight, they were orally administered 1 mL of a lipid emulsion consisting of corn oil (3 mL), cholic acid (40 mg) and cholesteryl oleate (1 g) plus physiological saline (3 mL), the lipid emulsion (1 mL) plus total chikusetsusaponins (final concentration 1000 mg/kg body) or the lipid emulsion plus orlistat (final concentration 45 mg/kg body). Blood samples were taken from the tail vein 0, 0.5, 1, 2, 3, 4 and 5 h after administration of the lipid emulsion with or without total saponins or orlistat using a capillary tube (heparinized), and centrifuged at 5500 × g for 5 min in a Model KH-120 M (Kubota Co., Osaka, Japan) centrifuge to obtain the plasma. The plasma triacylglycerol concentration was determined using a Wako Triglyceride E-Test kit.

Lipase activity in the porcine pancreas was assay as described previously 19. Enzyme activity was expressed as μmoles of oleic acid released per mL of reaction mixture per min.

All values are expressed as means ± s.e. Data were analyzed by one-way ANOVA, and then differences among means were analyzed using the Fisher's protected LSD test. Differences were considered significant at P < 0.05.

Consumption of the high fat diet for 3 days reduced the feces weight (0.14 ± 0.035 g/mouse/day) at day 3 compared with that of the control group (1.04 ± 0.017 g /mouse/day). Mice fed the high-fat diet plus 3% total chikusetsusaponins, or 0.025% orlistat for 3 days had significantly higher triacylglycerol level at day 3 compared with the high-fat diet group (Table 2).

The mean food consumption per week per mouse was significantly different between the control group and high-fat diet groups, being 409.4 ±7.0 KJ in the control group and 611.0 ± 41.6 KJ in the high-fat diet group?? but was not significantly different among the high-fat diet groups, high-fat plus 1% and 3% saponin-diet groups, and high-fat plus 0.012% orlistat-diet groups, being 611.0 ± 41.6 KJ (high-fat diet), 708.7 ± 34.6 KJ (high-fat diet plus 1% chikusetsusaponin), 686.0 ± 30.0 KJ (high-fat diet plus 3% chikusetsusaponin) and 693.1 ± 53.8 KJ (high-fat diet plus 0.012% orlistat). A high-fat plus 0.024% orlistat diet significantly reduced the body weight to 4 weeks compared to both the normal and the high-fat diet groups, but among 15 mice of high-fat plus 0.024% orlistat-diet groups, 10 mice died by the severe diarrhea until 5 weeks by the feeding high-fat diet containing 0.024% orlistat (data not shown). Orlistat is a potent inhibitor of pancreatic lipase; therefore, the absorption of dietary fat from small intestine was strongly inhibited by feeding high-fat diet containing 0.025% orlistat (Table 2). Mice treated with high-fat diet plus 0.024% orlistat for a long term died by severe diarrhea, but high-fat diet plus 0.012% orlistat-fed mice did not cause diarrhea (data not shown). This finding suggests that feeding level of 0.024% orlistat in the high-fat diet to mice for a long term may cause the physiological toxicity through the continuous inhibition of pancreatic lipase. Feeding high-fat plus 0.012% orlistat diet had no effect on the adipose tissue weight. In the present study, the discrepancy between anti-obesity action and inhibitory action of pancreatic lipase by orlistat could not sufficiently be clarified; therefore, this discrepancy is needed to clarify in future studies. Figure 2 shows the changes in body weight of the groups during the experiment. Consuming a high-fat diet for 9 weeks caused significant increases in body weight at 1 to 9 weeks compared to that of the control group (laboratory pellet chow). Consuming a high-fat diet supplemented with 1% or 3% chikusetsusaponins significantly suppressed the increase in body weight during the experimental period compared to the high-fat diet, but the high-fat diet plus 0.012% orlistat did not affect. The final parametrial adipose tissue weights of the groups are shown in Table 3. The final parametrial adipose tissue weight was significantly increased by consumption of a high-fat diet compared to the control diet, and that in animals with a high-fat diet containing 1% or 3% chikusetsusaponins was significantly reduced as compared to that in the high-fat diet group. Furthermore, mice fed a high-fat diet long term developed fatty liver, with increases of the liver weight and an accumulation of hepatic triacylyglycerol compared to the control group. Consumption of a high-fat diet containing 1% or 3% chikusetsusaponins, or 0.012% orlistat reduced the liver weight and hepatic triacylglycerol content compared to consumption of a high-fat diet group (Table 3).

The plasma triacylglycerol level (2.98 ± 0.31 mM, n = 4) at 2 h after orally administered lipid-emulsion was significantly reduced to 1.66 ± 0.28 mM (n = 4) and 0.93 ± 0.16 mM (n = 4), respectively, by the oral administration of total chikusetsusaponins (1000 mg/kg) and orlistat (45 mg/kg) (Fig. 3).

As shown in Fig. 4a, methanol extracts, total chikusetsusaponins and the positive control orlistat dose dependently inhibited pancreatic lipase activity in an assay system using triolein emulsified with lecithin, but water and CHClS₃-eluate fraction had no effect. We then examined the various chikusetsusaponins isolated from P. japonicus rhizomes. Chikusetsusaponin III and 28-deglucosylchikusetsusaponins IV and V dose-dependently inhibited the pancreatic lipase activity at concentrations of 125 to 500 μg/mL (Fig. 4b).