Effects of Flavonoids and Triterpene Analogues from Leaves of Eleutherococcus sieboldianus (Makino) Koidz. ‘Himeukogi’ in 3T3-L1 Preadipocytes

Eleutherococcus sieboldianus (Makino) Koidz. is a local product from the area in and around Yonezawa City in Yamagata Prefecture, Japan. It has been used as a medicinal plant for a long time. We isolated and identified four types of flavonoid glycosides [astragalin (1), isoquercetin (2), rhamnocitrin 3-O-glucoside (3), and nicotiflorin (4)], a triterpene [methyl hederagenin (5)], and three types of triterpene glycosides [δ-hederin (6), echinocystic acid 3-O-arabinoside (7), and cauloside B (8)] from the methanol extract of E. sieboldianus, which regulates lipid accumulation in 3T3-L1 preadipocytes. Among the compounds isolated, 2 and 8 up- and down-regulated lipid accumulation and insulin induced adipocyte differentiation in 3T3-L1 preadipocytes. Compound 2 induced up-regulation of lipid accumulation and decreased adipocyte size, while 8 down-regulated lipid accumulations without decreasing cell size. Additionally, 2 increased adipogenic proteins [peroxisome proliferator-activated receptor γ (PPARγ), CCAAT/enhancer-binding protein alpha (C/EBPα), and fatty-acid-binding protein 4 (FABP4)]. In contrast, 8 decreased the levels of all adipogenic proteins and glucose transporter type 4 (GLUT4), but increased adiponectin.


Introduction
Eleutherococcus sieboldianus (Makino) Koidz. (fiveleaf aralia) is a local product from the area in and around Yonezawa City in Yamagata Prefecture, Japan. E. sieboldianus has been used as a medicinal plant for preventing cancer and cardiovascular diseases and it is regarded as a longevity drug [1]. Additionally, E. sieboldianus has been used as a traditional Korean medicine (min gal pi) [2]. To date, constituents from E. sieboldianus have not been elucidated, except for saponins [3]. Recently, the Figure 1. Cytotoxic effects of the three extracts and eight compounds isolated from E. sieboldianus in 3T3L1 cells. Data are expressed as the mean ± SD from three independent experiments. The same letters indicate that there are no differences between those groups, and different letters indicate significant differences (p < 0.05).

Figure 2.
The effects of the three extracts and eight compounds isolated from E. sieboldianus on triglycerol levels in 3T3-L1 cells. The 3T3-L1 cells were cultured in 24-well plates and differentiated under the conditions described in the materials and methods section for each compound. Undifferentiated cells, cells with the addition of the MDI mixture (a mixture of 0.5 mM 3-isobutyl-1methyl xanthine (M), 0.1 μM dexamethasone (D), and 2 μM insulin (I)), rosiglitazone, and berberine, are indicated by CTRL, INS, ROS, and BER, respectively. On day 8 of culturing, the medium was removed and cells were lysed using Ripa buffer. Triglycerol levels were determined by the Triglycerol E-test Wako (Wako Pure Chemical, Osaka, Japan). Data are presented as the mean ± SD from three independent experiments. The same letters indicate that there are no differences between those groups, and different letters indicate significant differences (p < 0.05).

Figure 1.
Cytotoxic effects of the three extracts and eight compounds isolated from E. sieboldianus in 3T3L1 cells. Data are expressed as the mean ± SD from three independent experiments. The same letters indicate that there are no differences between those groups, and different letters indicate significant differences (p < 0.05).

Figure 2.
The effects of the three extracts and eight compounds isolated from E. sieboldianus on triglycerol levels in 3T3-L1 cells. The 3T3-L1 cells were cultured in 24-well plates and differentiated under the conditions described in the materials and methods section for each compound. Undifferentiated cells, cells with the addition of the MDI mixture (a mixture of 0.5 mM 3-isobutyl-1methyl xanthine (M), 0.1 μM dexamethasone (D), and 2 μM insulin (I)), rosiglitazone, and berberine, are indicated by CTRL, INS, ROS, and BER, respectively. On day 8 of culturing, the medium was removed and cells were lysed using Ripa buffer. Triglycerol levels were determined by the Triglycerol E-test Wako (Wako Pure Chemical, Osaka, Japan). Data are presented as the mean ± SD from three independent experiments. The same letters indicate that there are no differences between those groups, and different letters indicate significant differences (p < 0.05).

Characterization of the Compounds Isolated
All of the compounds were identified by comparison of their spectral data with the literature. Astragalin (1) was previously isolated from the Astragalus species [32]. Compounds 2, 3, and 4 also displayed flavonol glycoside characteristics in NMR and were determined as isoquercetin (2), rhamnocitrin 3-O-glucoside (3), and nicotiflorin (4), respectively [33,34]. The NMR data of compounds 5-8 possessed the characteristics of triterpene and their structures were to be methyl hederagenin (5), δ-hederin (6), echinocystic acid 3-O-arabinoside (7), and cauloside B (8) (Figure 3) [35][36][37][38]. The effects of the three extracts and eight compounds isolated from E. sieboldianus on triglycerol levels in 3T3-L1 cells. The 3T3-L1 cells were cultured in 24-well plates and differentiated under the conditions described in the materials and methods section for each compound. Undifferentiated cells, cells with the addition of the MDI mixture (a mixture of 0.5 mM 3-isobutyl-1-methyl xanthine (M), 0.1 µM dexamethasone (D), and 2 µM insulin (I)), rosiglitazone, and berberine, are indicated by CTRL, INS, ROS, and BER, respectively. On day 8 of culturing, the medium was removed and cells were lysed using Ripa buffer. Triglycerol levels were determined by the Triglycerol E-test Wako (Wako Pure Chemical, Osaka, Japan). Data are presented as the mean ± SD from three independent experiments. The same letters indicate that there are no differences between those groups, and different letters indicate significant differences (p < 0.05).

Cytotoxicity and Regulatory Effects on Adipogenesis of the Compounds Isolated from E. sieboldianus
We examined cell survival among 3T3-L1 cells cultured with these compounds for 8 days as described in 4.7. (Figure 1). Although 7 and 8 induced cytotoxicity at 50 and 100 μM, respectively, compounds 1-6 did not induce cytotoxicity at 100 μM. Therefore, the effects on adipogenesis in the 3T3-L1 cells of each compound were evaluated at 10 and 30 μM under the conditions stated in the materials and methods section.
Differentiation of the 3T3-L1 cells to adipocytes was achieved within 8 days in the presence of compounds 1-8 at concentrations of 10 and 30 μM as described in 4.6 , and the accumulation of intracellular lipids were measured ( Figure 2). When rosiglitazone (ROS) or berberine (BER) was added to the medium, with a mixture of 0.5 mM 3-isobutyl-1-methyl xanthine (M), 0.1 μM dexamethasone (D), and 2 μM insulin (I) (MDI), lipid accumulation increased by 59% and decreased by 47% respectively. Lipid accumulation was decreased by 32% when 8 (30 μM) and the MDI mixture were added to the medium compared with the addition of the MDI mixture only, but the addition of 2 (30 μM) up-regulated the levels of intracellular lipids by 43% ( Figure 2). The images shown in Figure 4 indicated that the size of the 3T3-L1 cells was diminished following the addition of ROS and also when 2 and the MDI mixture were added. In contrast, the addition of 8 and the MDI mixture led to decreased intracellular lipids, but unlike ROS, miniaturization of the adipocytes did not occur.

Cytotoxicity and Regulatory Effects on Adipogenesis of the Compounds Isolated from E. sieboldianus
We examined cell survival among 3T3-L1 cells cultured with these compounds for 8 days as described in 4.7. (Figure 1). Although 7 and 8 induced cytotoxicity at 50 and 100 µM, respectively, compounds 1-6 did not induce cytotoxicity at 100 µM. Therefore, the effects on adipogenesis in the 3T3-L1 cells of each compound were evaluated at 10 and 30 µM under the conditions stated in the materials and methods section.
Differentiation of the 3T3-L1 cells to adipocytes was achieved within 8 days in the presence of compounds 1-8 at concentrations of 10 and 30 µM as described in 4.6 , and the accumulation of intracellular lipids were measured ( Figure 2). When rosiglitazone (ROS) or berberine (BER) was added to the medium, with a mixture of 0.5 mM 3-isobutyl-1-methyl xanthine (M), 0.1 µM dexamethasone (D), and 2 µM insulin (I) (MDI), lipid accumulation increased by 59% and decreased by 47% respectively. Lipid accumulation was decreased by 32% when 8 (30 µM) and the MDI mixture were added to the medium compared with the addition of the MDI mixture only, but the addition of 2 (30 µM) up-regulated the levels of intracellular lipids by 43% ( Figure 2). The images shown in Figure 4 indicated that the size of the 3T3-L1 cells was diminished following the addition of ROS and also when 2 and the MDI mixture were added. In contrast, the addition of 8 and the MDI mixture led to decreased intracellular lipids, but unlike ROS, miniaturization of the adipocytes did not occur.

Cytotoxicity and Regulatory Effects on Adipogenesis of the Compounds Isolated from E. sieboldianus
We examined cell survival among 3T3-L1 cells cultured with these compounds for 8 days as described in 4.7. (Figure 1). Although 7 and 8 induced cytotoxicity at 50 and 100 μM, respectively, compounds 1-6 did not induce cytotoxicity at 100 μM. Therefore, the effects on adipogenesis in the 3T3-L1 cells of each compound were evaluated at 10 and 30 μM under the conditions stated in the materials and methods section.
Differentiation of the 3T3-L1 cells to adipocytes was achieved within 8 days in the presence of compounds 1-8 at concentrations of 10 and 30 μM as described in 4.6 , and the accumulation of intracellular lipids were measured ( Figure 2). When rosiglitazone (ROS) or berberine (BER) was added to the medium, with a mixture of 0.5 mM 3-isobutyl-1-methyl xanthine (M), 0.1 μM dexamethasone (D), and 2 μM insulin (I) (MDI), lipid accumulation increased by 59% and decreased by 47% respectively. Lipid accumulation was decreased by 32% when 8 (30 μM) and the MDI mixture were added to the medium compared with the addition of the MDI mixture only, but the addition of 2 (30 μM) up-regulated the levels of intracellular lipids by 43% ( Figure 2). The images shown in Figure 4 indicated that the size of the 3T3-L1 cells was diminished following the addition of ROS and also when 2 and the MDI mixture were added. In contrast, the addition of 8 and the MDI mixture led to decreased intracellular lipids, but unlike ROS, miniaturization of the adipocytes did not occur.

The Effects of 2 and 8 on Adipogenic Proteins
The expression levels of the adipogenic proteins examined, namely, PPARγ, C/EBPα, and FABP4 in 3T3-L1 cells that were differentiated using MDI mixture with or without ROS, BER, 2, or 8 are shown in Figure 5. ROS, a positive reference compound, and 2 showed that up-regulation of the expression levels of adipogenic proteins (PPARγ, C/EBPα, and FABP4) was found. In contrast, BER, a negative reference compound, and 8 down-regulated the expression of adipogenic proteins. the medium. Undifferentiated cells, cells with the addition of MDI mixture, rosiglitazone, and berberine are indicated by CTRL, INS, ROS, and BER respectively.

The Effects of 2 and 8 on Adipogenic Proteins
The expression levels of the adipogenic proteins examined, namely, PPARγ, C/EBPα, and FABP4 in 3T3-L1 cells that were differentiated using MDI mixture with or without ROS, BER, 2, or 8 are shown in Figure 5. ROS, a positive reference compound, and 2 showed that up-regulation of the expression levels of adipogenic proteins (PPARγ, C/EBPα, and FABP4) was found. In contrast, BER, a negative reference compound, and 8 down-regulated the expression of adipogenic proteins.  Figure 4. Protein levels were measured by electroblotting. Data are presented as the mean ± SD from three independent experiments. The same letters indicate no differences between groups, and different letters indicate significant differences (p < 0.05).

The Effects of 2 and 8 on the Expression of Lipogenic Proteins
Acetyl-CoA carboxylase (ACC) and fatty acid synthase (FAS) are enzymes that synthesize malonyl-CoA from acetyl-CoA and fatty acids and from acetyl-CoA and malonyl-CoA respectively [39]. The effects of 2 and 8 on the levels of lipogenic proteins are shown in Figure 5. ROS showed up-regulation of the expression of ACC and FAS. BER down-regulated the expression of ACC and FAS. In contrast, adding 2 or 8 did not change the expression of ACC and FAS.

The Effects of 2 and 8 on the Expression of a Lipolytic Protein
Perilipin coats lipid droplets in adipocytes and induces lipolysis [40,41]. The effects of 2 and 8 on the levels of perilipin are shown in Figure 5. Similar to the effects on the levels of lipogenic proteins, ROS up-regulates and BER down-regulates the expression of perilipin. Adding 2 or 8 did not change perilipin levels.

The Effects of 2 and 8 on the Expression of Adiponectin
Adiponectin is a hormone controlling energy expenditure, glucose homeostasis, and insulin resistance [9]. Insulin resistance is improved by increasing adiponectin [10]. In addition, it has been reported that insulin resistance in diabetic mice was improved by oral administration of adiponectin [12]. The effects of 2 and 8 on the levels of adiponectin are shown in Figure 5. ROS, Compound 2, and  Figure 4. Protein levels were measured by electroblotting. Data are presented as the mean ± SD from three independent experiments. The same letters indicate no differences between groups, and different letters indicate significant differences (p < 0.05).

The Effects of 2 and 8 on the Expression of Lipogenic Proteins
Acetyl-CoA carboxylase (ACC) and fatty acid synthase (FAS) are enzymes that synthesize malonyl-CoA from acetyl-CoA and fatty acids and from acetyl-CoA and malonyl-CoA respectively [39]. The effects of 2 and 8 on the levels of lipogenic proteins are shown in Figure 5. ROS showed up-regulation of the expression of ACC and FAS. BER down-regulated the expression of ACC and FAS. In contrast, adding 2 or 8 did not change the expression of ACC and FAS.

The Effects of 2 and 8 on the Expression of a Lipolytic Protein
Perilipin coats lipid droplets in adipocytes and induces lipolysis [40,41]. The effects of 2 and 8 on the levels of perilipin are shown in Figure 5. Similar to the effects on the levels of lipogenic proteins, ROS up-regulates and BER down-regulates the expression of perilipin. Adding 2 or 8 did not change perilipin levels.

The Effects of 2 and 8 on the Expression of Adiponectin
Adiponectin is a hormone controlling energy expenditure, glucose homeostasis, and insulin resistance [9]. Insulin resistance is improved by increasing adiponectin [10]. In addition, it has been reported that insulin resistance in diabetic mice was improved by oral administration of adiponectin [12]. The effects of 2 and 8 on the levels of adiponectin are shown in Figure 5. ROS, Compound 2, and 8 up-regulated the expression levels of adiponectin dose dependently. In contrast, the addition of BER down-regulated the expression levels of adiponectin.

The Effects of 2 and 8 on the Expression of GLUT4
GLUT4 accumulates in intracellular organelles (endosomes) in an environment without insulin stimulation. Once insulin binds to its receptor in the cell membrane, phosphoinositide 3-kinase (PI3K), Akt, and other downstream components are activated in turn, followed by the translocation of GLUT4 to the cell membrane. GLUT4 in the cell membrane takes glucose up from the blood [42]. In addition, the translocation of GLUT4 to the cell membrane is closely related to adipogenesis [43]. The expression of GLUT4 in 3T3-L1 cells that were differentiated under the same conditions as mentioned in Figure 4 is shown in Figure 5. ROS and BER up-and down-regulated GLUT4 protein levels respectively. Compounds 2 and 8 showed dose-dependent up-regulation of GLUT4 protein levels.
Among the compounds isolated, 2 and 8 up-and down-regulated, respectively, lipid accumulation in 3T3-L1 preadipocytes. Compound 2 induced up-regulation of lipid accumulation and decreased adipocyte size, while 8 down-regulated lipid accumulation without decreasing cell size ( Figure 3). Additionally, 2 increased adipogenic proteins (PPARγ, C/EBPα, and FABP4) in a dose dependent manner ( Figure 5). In contrast, 8 decreased the levels of all adipogenic proteins and GLUT4, but increased adiponectin. Because it showed promotion of lipid accumulation equivalent to ROS (Figure 2), 2 may be one of the main active components from E. sieboldianus. On the other hand, the components other than 8 (with its inhibition of lipid accumulation) may be responsible for the use of E. sieboldianus, because the effect of 8 was weaker than that of the methanol extract.
Though saponins and flavonoid glycoside (nicotiflorin) are known to be constituents of E. sieboldianus [3], there have been no reports on compounds except nicotiflorin that have been identified from E. sieboldianus.
Only 2 up-regulated lipid accumulation among the four types of flavonoid glycosides, which indicates that the number and position of the hydroxyl groups binding to the B ring of the flavonol skeleton controls this effect. It has been shown that flavonoids, especially flavonol, flavan-3-ol and anthocyanins decrease glycemia, which is followed by an improvement in the secretion and sensitivity of insulin [52]. Furthermore, it has also been reported that anthocyanins improve glucose metabolism, insulin resistance, and β cell dysfunction via regulation of GLUT4 [53,54]. In addition, we showed that the presence or absence of a hydroxyl group binding to the C23 position of triterpene controls the ability to regulate lipid accumulation because the regulation of lipid accumulation significantly differed between 7 and 8. The relationship between the chemical structure of triterpenoids and anti-diabetic effects is unclear, however, it has been reported that triterpene analogues, such as ginsenosides, regulate adipogenesis [55][56][57][58]. Further studies, including the findings obtained in the present study, may be needed to clarify the mechanisms of action, even though there are many reports concerning anti-diabetic or anti-obesity effects of flavonoid and triterpene analogues.
Increases in hypertrophic adipocytes regulate the secretion of adiponectin (a hormone regulating energy expenditure), glucose homeostasis, and insulin sensitivity. In contrast, because smaller adipocytes secrete more adiponectin and improve sensitivity to insulin, decreasing the size of hypertrophic adipocytes, followed by increasing the number of smaller adipocytes, is considered to be one strategy to treat diabetes [13].
In the present study, compound 2 and ROS decreased the size of adipocytes produced by insulin stimulation, as shown in Figure 4. In contrast, unlike ROS (an agonist of PPARγ) [59], 2 decreased the size of adipocytes but was irrelevant to the expression of lipogenic and lipolytic proteins ( Figure 5); thus, 2 may show antidiabetic effects through a different mechanism of action than thiazolidinediones (TZDs) such as ROS, the drugs that are widely used to treat diabetes.
Furthermore, compound 8 showed a dose-dependent anti-obesity effect, as seen in Figure 3. The anti-obesity effect of 8 was indicated because 8 down-regulated the levels of three types of adipogenic proteins and GLUT 4 as is shown in Figure 5. Because 8 did not decrease the size of adipocytes but did up-regulate the secretion of adiponectin, 8 may induce anti-obesity effects by improving the insulin resistance of adipocytes.
PPARγ was identified as an indispensable transcription factor that regulates adiposity through analysis of gene-deficient mice [60]. Moreover, C/EBPα is a transcription factor regulating the acquisition of insulin sensitivity [61]. In contrast, FABP4, which is expressed in adipocytes and macrophages, is strongly related to inflammation and lipid homeostasis in cells [62]. In the present study, compounds 2 and 8 up-or down-regulated adipogenic proteins and they up-regulated adiponectin and down-regulated GLUT4. In contrast, 2 and 8 were irrelevant to the expression of lipogenic and lipolytic proteins ( Figure 5). This suggests that the regulation of adipogenesis by 2 and 8 were not related to the expression levels of GLUT4. Furthermore, the reason for the up-regulation of the levels of adiponectin without downsizing of adipocyte by DMI should be clarified in the future.
Anti-obesity compounds are therapeutic agents that can reduce body weight by decreasing the consumption or absorption of food, and/or increasing energy expenditure [63]. Because the causes of obesity are various, it is necessary to use various treatments, according to the characteristics of the patient. In the present study, effects that decrease lipid accumulation by stimulation using compound 8 are recognized as one type of anti-obesity treatment. The in vivo trials are also necessary to practicality access the anti-adiposity activity of compound 8.
In the present study, the effect on the glucose uptake of the cells treated with compound 2 and 8 was deduced by the evaluation of expression levels of GLUT4 in the cytoplasm. It was clarified that both compound 2 and 8 down-regulated the expression of GLUT4. On the other hand, it is necessary to evaluate the GLUT4 levels in the cell membrane and the amount of uptake of glucose to confirm the effects of compound 2 and 8 in more detail. Though TZDs like ROS are widely used to treat diabetes, they have side effects such as edema. Because E. sieboldianus contains both anti-diabetic and anti-obesity compounds such as 2 and 8, and because extracts from E. sieboldianus improve insulin resistance in type 2 diabetic mice [4,6], E. sieboldianus may be a promising natural anti-diabetic or anti-obesity agent.
Because anti-diabetic compounds stabilize and control blood glucose levels [64], the evaluation of the biological effects of compound 2 in people with diabetes or in animal models of this disease from the viewpoint of blood glucose levels is necessary to confirm the practical effects of compound 2 in detail.

Solvent Fractionation
Raw leaves of Eleutherococcus sieboldianus (Araliaceae) was purchased commercially in Yamagata Prefecture in Japan and identified by Prof. Nishina. The voucher specimen has been deposited at the College of Science and Technology, Nihon University. Dried E. sieboldianus powder (1846 g) was immersed in hexane (500 mL) for 24 h at room temperature. The solvent containing the extracts was filtrated through a filter paper (5C; Whatman, Brentford, UK) and the filtrate was evaporated to dryness to prepare hexane extract. The residue was then stirred in chloroform (500 mL) at room temperature for 24 h, filtrated, and the filtrate was dried in vacuo to prepare chloroform extract. Then, methanol extract was obtained in the same manner.

Isolation of Active Constituents
The methanol extract (227 g) was partitioned with n-hexane, ethyl acetate, and n-butanol in that order and each fraction of n-hexane (25 g), ethyl acetate (37 g), and n-butanol (78 g) was obtained. The ethyl acetate fraction (37 g) was divided by silica gel column chromatography (CC) eluted with Purity of compound 1-9 was confirmed as more than 95% by measurement of 1 H-NMR spectra with dimethyl sulfone as the internal standard [65].

Analytical Instrument of Active Components
1 H (400 MHz) and 13 C (100 MHz) NMR spectra were recorded with a JEOL ECX 400 spectrometer with tetramethylsilane as an internal standard. MS spectra were obtained using a JEOL JMS-700/GI spectrometer and the Waters UPLC-MS system (Aquity UPLC XevoQTof). IR spectra were recorded on a JASCO FT/IR-460 Plus spectrophotometer.

Cell Culture
Murine 3T3-L1 preadipocytes were propagated in DMEM supplemented with 10% calf serum until 80% confluence (day 0) and the medium was replaced with DMEM containing 10% FBS and MDI mixture, with or without one of the test compounds. After 48 h (day 2), the medium was replaced with DMEM containing 10% FBS and 2 µM insulin. After 48 h (day 4), the medium was replaced with DMEM containing 10% FBS. Thereafter, the medium was exchanged every other day [66]. 0.1 µM or 2.7 nM of ROS or BER was used as a positive or negative reference compound respectively. Cells were maintained in a humidified atmosphere of 5% CO 2 at 37 • C

Cell Toxicity Assay
The 3T3-L1 cells were seeded in 96-well plate with DMEM supplemented with 10% calf serum until 80% confluence and the medium was replaced with DMEM containing a 10% FBS with or without one of a test compound. Thereafter, we exchanged the medium without test compounds on every other day. Cytotoxicity was measured by the use of Cell Counting Kit-8 (Dojindo, Kumamoto, Japan) according to the instructions of the manufacturer. Absorbance was measured at 450 nm by using a Sunrise Absorbance Reader (Tecan, Männedorf, Switzerland).

Measurement of Intracellular Triglycerol Level
Intracellular triglycerol levels in 3T3-L1 cells at day 8 were measured by using of E-test WAKO Triglyceride Kit (Wako Pure Chemical) according to the instructions of the manufacturer. 100 ng/mL of BODIPY 493/503 were added to culture medium followed by incubation of 10 min. Images were taken by a fluorescent cell imager (Floid Cell Imaging Solution; Life Technologies) [66].

Statistical Analysis
The results were expressed as mean ± standard deviation (SD). The significant difference between the groups compared was determined using analysis of variance (ANOVA) followed by the Tukey test.