Constituents of the Stem of Cucurbita moschata Exhibit Antidiabetic
Activities through Multiple Mechanisms
Chi-I Changa, Chih-Ming Hsua, Ting-Syuan Lia, Shen-Da Huanga, Chen-Chen Lina,
Chia-Hung Yena, Chang-Hung Choub,#, Hsueh-Ling Chenga,*
a Department of Biological Science and Technology, National Pingtung University of Science and
Technology, No. 1, Shuehfu Rd., Neipu, Pingtung 91201, Taiwan
b Research Center for Biodiversity and Graduate Institute of Ecology and Evolutionary Biology,
China Medical University, 91, Hsueh-Shih Road, Taichung, 40402, Taiwan.
*Corresponding author: Hsueh-Ling Cheng, Professor E-mail: [email protected]
Tel: +886-8-7703202 ext 5186 ; Fax: 886-8-774-0550
#Co-corresponding author: Chang-Hung Chou, Professor
Tel.: 886-4-2205366 ext1633; fax: 886-4-22071500. E-mail: [email protected]
Abstract
molecular mechanisms mediating its hypoglycaemic activity are not clear. This study characterized some of the hypoglycaemic constituents in the crude extract of the Cucurbita
moschata stem and investigated the mechanisms of their action. The C. moschata stem extract
showed a hypoglycaemic effect in vivo in streptozotocin-induced diabetic mice. Several
hypoglycaemic fractions of the extract were identified via cell-based screening, and two of them were confirmed to have hypoglycaemic activities in vivo. Ten compounds were afforded from the 2 fractions. Compounds 9 [(22E,24R)-24-methyl-6β-methoxy-5α-cholesta-7,22-diene-3β,5-diol] and 10 [3β-hydroxy-(22E,24R)-ergosta-5,8,22-trien-7-one] showed an insulin-like activity in normal cells that may be mediated by AMP-activated protein kinase. Compounds 4 (ferulic acid),
8 (syringaresinol) and 9 exhibited an insulin sensitizing and/or insulin substitution function in
insulin-resistant cells. Thus, the C. moschata stem contains compounds with potential for managing type 1 or type 2 diabetes.
Keywords: AMP-activated protein kinase; Cucurbita moschata; ferulic acid; insulin resistance;
Chemical compounds studied in this article
Ferulic acid (PubChem CID: 445858); 2-Hydroxybenzoic acid (PubChem CID: 338);
4-Hydroxycinnamic acid (PubChem CID: 637542); 4-Ketopinoresinol (PubChem CID: 44578390); Loliolide (PubChem CID: 100332); Pinoresinol (PubChem CID: 234817); Syringaresinol
1. Introduction
Diabetes remains to be one of the most prevalent chronic disorders worldwide. Type 1 diabetes is characterized by impaired insulin secretion of the pancreatic cells , whereas type 2 diabetes, which accounts for more than 90% of diabetic cases, is characterized by insulin
resistance and progressive -cell dysfunction. Insulin resistance is the reduced response of tissues to the circulating insulin, and initially causes a compensatory increase in insulin secretion to maintain a normal level of blood glucose, yet chronic overstimulation of cells results in damage of the cells and insufficient insulin secretion, leading to the development of hyperglycaemia and other diabetic complications . Thus, managing insulin resistance is crucial in preventing and treating type 2 diabetes. Medicines with insulin sensitizing function (i.e., agents that increase the sensitivity of cells to insulin such as thiazolidinediones [TZDs]) are often prescribed . The binding of insulin with the cell-surface insulin receptor activates the tyrosine-kinase activity of the receptor, which in turn recruits and catalyzes the tyrosine phosphorylation of insulin receptor substrates (IRSs), mainly IRS-1 and IRS-2, leading to the activation of downstream effectors PI3K (phosphatidylinositol-3-kinase), PDK1 (PI3K-dependent kinase), and subsequently Akt (also known as protein kinase B). Akt catalyzes the phosphorylation of AS160 (Akt substrate of 160 kDa), which causes the translocation of glucose transporter-4 (GLUT-4) from its intracellular pool to the cell surface. The GLUT-4 in the cell membrane facilitates the cell’s extracellular glucose uptake, resulting in the reduction of blood glucose . Obesity is a known high risk factor for the development of insulin resistance and type 2 diabetes . Proinflammatory cytokines
induced by obesity such as tumour necrosis factor-(TNF-)interfere with the insulin-signaling pathway by suppressing the tyrosine phosphorylation or expression of IRSs and insulin receptor, causing insulin resistance in cells . Thus, in previous studies, we have used TNF-to induce insulin resistance in FL83B cells and have developed a screening method using insulin-resistant cells to identify potential hypoglycaemic compounds for treating type 2 diabetes . In the current study, this cell-based method was used to identify and characterize the hypoglycaemic
constituents in Cucurbita moschata.
Pumpkin which belongs to the family Cucurbitaceae and the genus Cucurbita has received considerable attention in recent years because of its nutritional and health benefits. Species of pumpkins that are commonly cultivated include Cucurbita moschata, Cucurbita maxima,
Cucurbita pepo, Cucurbita ficifolia, and Cucurbita mixta. C. ficifolia is used in traditional
Mexican medicine to treat diabetes , and the extract of its fruit has been shown to have a hypoglycaemic action in type 2 diabetes patients . C. moschata is widely consumed in Asia. Some natural products isolated from the C. moschata fruit were reported to reduce blood glucose levels in diabetic mice . Although the fruit is the more commonly consumed and explored part of the plant, a water extract from the C. moschata stem was shown to have an antiobesity effect on high-fat-diet-induced obesity in mice , and a compound from this extract displayed an
antiadipogenic effect in cellular models . Thus, the stem of the pumpkin likely contains bioactives that are also beneficial for the health, but are mostly unexplored.
were investigated. The hypoglycaemic effect of the crude extract of the C. moschata stem was tested in streptozotocin (STZ)-induced diabetic mice. The constituents of the extract were subsequently fractionated, isolated, and subjected to screening for hypoglycaemic activity using the previously established, cell-based assay. Some of the hypoglycaemic principles were
identified, and the mechanisms of their actions were investigated.
2. Materials and Methods
2.1 ChemicalsGlucose concentration was assayed using a glucose assay kit (Glucose GOD FS, Diagnostic Systems, Holzheim, Germany); foetal bovine serum (FBS) and an antibody against
phosphorylated AS160 (Thr 642) were from Invitrogen (San Diego, CA, USA); specific antibodies against total or phosphorylated IRS-1 (Tyr 895), total or phosphorylated Akt (Ser 473), total or phosphorylated acetyl-CoA carboxylase-1 (Ser 79), and total or phosphorylated AMPK subunit (Thr 172) were from Cell Signaling Technology (Beverly, MA, USA); an antibody against total AS160 was from Upstate (Salt Lake City, UT, USA); all secondary antibodies were from Santa Cruz Biotechnologies (Santa Cruz, CA, USA); mouse TNF-was acquired from R&D systems (Minneapolis, MN, USA); and streptozotocin, bovine insulin solution, troglitazone, cell culture media, dimethyl sulphoxide (DMSO), MTT
(3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide), NaF, sodium orthovanadate, and sodium pyrophosphate were from Sigma Chemical Company (St. Louis, MO, USA) as reagent grade or
cell-culture grade. The extract or fractions of C. moschata were dissolved in DMSO to a concentration of 100 or 50 mg/mL. Each pure compound (1–10) was dissolved in DMSO to 5 mg/mL or 10 mM.
2.2 General instrumental operation for the isolation and identification of compounds
Optical rotation measurements were obtained using a JASCO (Easton, MD, USA) DIP-180 digital spectropolarimeter. UV spectra weremeasured in methanol using a Shimadzu (Columbia, MD, USA ) UV-1601PC spectrophotometer. IR spectra were recorded using a Nicolet 510P FT-IR spectrometer (Thermo Scientific, Waltham, MA, USA). NMR spectra were analyzed in CDCl3
at room temperature using a Varian Mercury plus 400 NMR spectrometer (Agilent Technologies, Santa Clara, CA, USA), and the solvent resonance was used as the internal shift reference
(tetramethyl silane [TMS] as standard). Electron impact mass spectrophotometry was recorded using a Finnigan TSQ-700 mass spectrometer (Thermo Scientific). Thin-layer chromatography was performed on silica gel 60 F254 plates (Merck, Darmstadt, Germany). Silica gels (230–400 mesh ASTM, Merck) were used for column chromatography. Semipreparative HPLC was conducted on a Lichrosorb silica gel 60 (5 m) column (250 × 10 mm) using an RI detector.
2.3 Source, extraction, fractionation, and purification of C. moschata compounds
The Cucurbita moschata stems were collected in Pingtung County, Taiwan in July 2006 in a private farm in which the plant was cultivated, and purchased from the owner. No endangered or protected species was involved or sampled during the collection. The identification of the
voucher specimens was done by Prof. Sheng-Zehn Yang, Curator of the Herbarium, National Pingtung University of Science and Technology. The voucher specimens were kept in the laboratory of Dr. Chi-I Chang, Department of Biological Science and Technology, National Pingtung University of Science and Technology, Pingtung, Taiwan, for reference.
The dried C. moschata stem (71 kg) was mechanically ground to a fine powder and then sieved through a 10 mesh sieve. The obtained powder was extracted with methanol (3 × 120 L) at room temperature (7 d each) . The combined methanol extract was evaporated under reduced pressure to afford a black residue, which was suspended in H2O (6 L), and then partitioned
sequentially, using ethyl acetate (EA) and n-butanol (3 × 4 L) as the solvents. The EA fraction (780 g) was chromatographed on a silica gel column (150 × 12 cm), using solvent mixtures of n-hexane and EA with increasing polarity as eluents. Twenty-three fractions were collected. F19 (12.3 g) was further chromatographed on a silica gel column (5 × 45 cm), eluted with CH2Cl2－
EA (5:1 to 0:1) to give 8 fractions (each about 600 mL), 19A－19H. 19F was column
chromatographed over silica gel eluted with n-hexane－CH2Cl2－EA (5:3:1) and semipreparative
HPLC eluted with n-hexane－EA (7:3) to yield 2 (1.1 mg), 7 (1.1 mg), and 6 (1.1 mg). 19G was column chromatographed over silica gel eluted with n-hexane－CH2Cl2－EA (5:4:1) and
semipreparative HPLC eluted with n-hexane－EA (7:3) to yield 1 (33.1 mg), 4 (7.5 mg), and 10 (1.1 mg). 19H was column chromatographed over silica gel eluted with CH2Cl2－EA (8:1) and
semipreparative HPLC eluted with n-hexane－EA (6:4) to yield 3 (5.1 mg), 5 (3.1 mg), and 9 (2.1 mg). F22 (15.5 g) was further chromatographed on a silica gel column (5 × 45 cm), eluted
with CH2Cl2－EA (4:1 to 0:1) to resolve into 7 fractions (each about 600 mL), 22A－22G. 22D
was column chromatographed over silica gel eluted with n-hexane－CH2Cl2－EA (3:3:1) and
semipreparative HPLC eluted with n-hexane－EA (6:4) to yield 8 (4.1 mg).
2.4 Cell culture and glucose uptake assay
FL83B cells (a normal mouse hepatic cell line; Bioresource Collection and Research Center, Hsinchu, Taiwan; BCRC number 60325) were cultured in F12K medium as previously
described . Glucose uptake assays in insulin-resistant FL83B cells were performed by treating cells with 20 ng/ml TNF- for 5 h, followed by treating the cells with the investigated chemical and 100 nM insulin for 5 h. At 0, 1, 2, 3, 4, and 5 h after the addition of the chemical, glucose concentration of the culture medium was assayed, as previously described . For assays using normal FL83B cells, experiments were performed similarly except that cells were not pretreated with TNF-. Briefly, FL83B cells were seeded in 12-well plates (2 x 105 cells/well), cultured
overnight, then washed with phosphate-buffered saline (PBS, pH 7.4) and incubated in serum-free F12K medium for 3 h. After being washed twice with PBS, cells were incubated for 5 h in 450 L of Eagle’s minimum essential medium (MEM) containing 20or 40 M of the
investigated compound, an equivalent amount of DMSO (control), or 100 nM of insulin (positive control). At 0, 1, 2, 3, 4, and 5 h after the addition of the compound, 30 μL of the medium was withdrawn and centrifuged at 500g for 5 min. Ten microliters of the resulting supernatant was mixed with 250 μL of Glucose GOD FS in a 96-well plate and incubated at 37 °C for 10 min.
Absorbance at 500 nm was then determined using a microplate reader (Molecular Devices, Sunnyvale, CA, USA). A standard curve was established simultaneously using solutions of various glucose concentrations. Experiments were performed in triplicate and data were
statistically analyzed against the control by two-way analysis of variance (ANOVA). Meanwhile, the total amount of medium glucose consumed in each cell plate during the 5 h of treatment was calculated to determine the relative glucose uptake versus the control.
2.5 Animal experiments
Eight-week-old C57B/6J male mice were purchased from the National Laboratory Animal Center (Taipei, Taiwan), and fed with a regular laboratory rodent diet and housed under a 12-h light–dark cycle. All experimental protocols involving animals and their care were approved by the Institutional Animal Care and Use Committee of National Pingtung University of Science and Technology (Permit Number: NPUST-IACUC-101-001), and were in accordance with
international guidelines. The mice (25-30 g) were divided into groups of 6 mice, and the
induction of hyperglycaemia in the mice was performed by injecting 50 mg/kg body weight/day of STZ abdominally to the mice for five consecutive days as described previously . Blood was collected from the tail and glucose concentration was measured using a blood glucose meter (Accu-Chek Go, Roche Applied Science, Indianapolis, IN, USA). The wounds of mice were treated with a healing balm afterwards to minimize suffering. In the normal control group, mice were injected abdominally with the corresponding volume of the vehicle instead of with STZ. For
hypoglycaemic assays, the crude extract or fractions of C. moschata were dissolved in a 0.5% methyl cellulose solution. An overnight meal was provided for the mice, then the food was removed and the mice were administered with the extract or selected fraction by using a gavage tube and a force-feeding needle. In the normal control and diabetic control groups, mice were fed with the corresponding volume of the vehicle. Blood glucose levels were monitored before and 2 h after the administration. Eight hours after the administration, food was again provided
overnight. This procedure was repeated for 6 d (crude extract) or 3 d (fractions).
2.6 Western blot analysis
Cells were seeded in 35- or 60-mm plates until they reached 80–90% confluence, and were then treated with combinations of agents as described in the figure legends. Subsequently, the cells were washed twice with PBS, submerged in a lysis buffer (1x Cell Culture Lysis Reagent [Promega, Madison, WI, USA] containing 1 mM of phenylmethylsulphonyl fluoride, 1g/mL of pepstatin, 1 g/mL of leupeptin, 1 g/mL of aprotinin,10 mM of NaF, 1 mM of sodium
orthovanadate, and 10 mM of sodium pyrophosphate), and scraped off the plate on ice. The resulting suspension was centrifuged at 14,000×g for 15 min or 500×g for 5 min (for IRS and AS160) at 4 ºC. The supernatant was collected and the protein concentration analyzed using Bradford assay reagent (Bio-Rad, Hercules, CA, USA). Equal amounts of proteins were sampled and subjected to electrophoresis and Western blotting as described previously . Alternatively, immunoreactive bands on the PVDF membrane (Millipore, Bedford, MA, USA) were detected
using an imaging system (UVP BioSpectrum, UVP, LLC, Upland, CA, USA) and band intensities analyzed by the supplied software.
2.7 Cytotoxicity assay
Cytotoxicity assay was performed using MTT as described previously . Experiments were performed in triplicate. The average percentage ± standard deviation of viable cells was
calculated against the control (cells treated with the solvent).
2.8 Statistical analysis
The data of glucose uptake assays were analyzed against the control by two-way ANOVA, with cell treatment and times at which the medium glucose concentration was measured (0, 1, 2, 3, 4, and 5 h) as the 2 parameters. Significance was considered when the P value between groups of cells and the P value of interaction between the 2 parameters were both < 0.05. The results of animal tests and cytotoxicity assays were analyzed by one-way ANOVA followed by Scheffe's post hoc test. Significance was considered when P < 0.05 and F > 3.5546.
3. Results
3.1 Characterization of the hypoglycaemic activity of the crude extract and fractions of C. moschata
Methanol extract (ME) of the C. moschata stem was subjected to animal tests to evaluate its hypoglycaemic activity in vivo. STZ-induced diabetic mice were fed 50 or 100 mg/kg/d of ME
for 6 d. In the group that received 100 mg/kg of ME (Fig. 1A), most of the blood glucose levels at 2 h postadministration of ME on Days 1 to 6 were significantly lower than the blood glucose level at 0 h of the respective day. Moreover, the blood glucose levels of this group of mice at 0 h on Days 5 and 6 significantly decreased compared with that on Day 1. In the group that received 50 mg/kg of ME (Fig. 1B), the hypoglycaemic effect of ME was less obvious because the reduction of blood glucose levels at 2 h did not achieve statistical difference against the blood glucose level at 0 h of the respective day. In the diabetic control (Fig. 1C), from Days 1 to 6, none of the blood glucose levels at 2 h was apparently lower than the blood glucose level at 0 h of the respective day, and the 0-h blood glucose levels were not obviously changed between days. The blood glucose levels of the normal control were essentially maintained in a normal range (Fig. 1D). Overall, these data supported that ME possessed a hypoglycaemic effect in vivo in a dosage-dependent manner.
ME was further partitioned successively using ethyl acetate and butanol (Fig. 2A). The ethyl acetate (EA), butanol (Bu), and water (W) portions were assayed for hypoglycaemic activity in
vitro using a cell-based method. As shown in Figs. 2B, 2C, and 2D, normal FL83B cells
exhibited increased glucose consumption when stimulated with insulin (Group 2 versus Group 1), whereas TNF--treated cells lost this response to insulin (Group 3 versus Group 2), confirming that TNF- induced insulin resistance. The addition of ME, EA, or Bu (Fig. 2B, Groups 4, 5, and 6, respectively) significantly promoted the glucose uptake of TNF--treated cells compared with the control (Group 3), whereas W (Fig. 2B, Group 7) did not, suggesting that the hypoglycaemic
principles in ME were distributed into EA and Bu after partitioning.
The EA was subjected to column chromatography, resulting in 23 fractions (Fig. 2A). These fractions were analyzed for hypoglycaemic activity in vitro. Fractions 1, 2, 3, 5–8, 10, 13, and 15–23 (Figs. 2C and 2D) all exhibited obvious hypoglycaemic activity in insulin-resistant cells compared with the control (Group 3). Thus, these fractions were considered to contain
hypoglycaemic molecules.
Fractions 19 (F19) and 22 (F22) were selected stochastically from the active fractions to test for in vivo hypoglycaemic activity. STZ-induced diabetic mice were fed 200 mg/kg/d of F19 or F22 for 3 d. As shown in Fig. 3A, the blood glucose levels of the F19 group of mice at 2 h
postadministration were all significantly reduced compared with the respective 0-h blood glucose level on each day of administration. Similar results were observed in the F22 group (Fig. 3B), and the 0-h blood glucose level of this group on Day 3 was significantly lower than that on Day 1. In the diabetic control, the 2-h blood glucose levels did not significantly reduce compared with 0-h blood glucose level of the respective day (Fig. 3C). These data support that F19 and F22 both possess hypoglycaemic activity in vivo. Thus, their constituents were further characterized.
3.2 Characterization of the hypoglycaemic activities of the constituents of Fractions 19 and 22
Nine compounds were afforded from F19 (Fig. 3E), including one apocarotenoid [loliolide (1) ], 3 phenolics [2-hydroxybenzoic acid (2) , 4-hydroxycinnamic acid (3) , and ferulic acid (4) ], 3 lignans [(+)-(1R,2S,5R,6S)-2,6-di(4'-hydroxyphenyl)-3,7-dioxabicyclo[3.3.0]octane (5) , pinoresinol (6) , and 4-ketopinoresinol (7) ], and 2 steroids
[(22E,24R)-24-methyl-6β-methoxy-5α-cholesta-7,22-diene-3β,5-diol (9) , and 3β-hydroxy-(22E,24R)-ergosta-5,8,22-trien-7-one (10) ]. One lignan, syringaresinol (8) was obtained from F22 (Fig. 4). The chemical structures of these compounds were verified by comparing their physical and spectral data (IR, MS, and NMR) with the values described in the literature (Appendix). These molecules were subjected to cell-based assays for hypoglycaemic activity.
First, the compounds were used in treating normal FL83B cells to observe whether they could increase the glucose uptake of normal cells as insulin does. This activity is defined as insulin-like activity herein. Consequently, 9 and 10, at 20 M, significantly promoted the glucose uptake of the treated cells as insulin does, whereas others did not (Fig. 4A); when the
concentration was raised to 40 M, 1-8 still did not exhibit a significant effect on promoting the glucose uptake of normal cells, and 2 and 5 reduced glucose consumption (Fig. 4B). Therefore, 9 and 10 possess insulin-like activity in normal cells.
Next, the hypoglycaemic activities of the compounds were tested in insulin-resistant cells. As a result, 4, 8, and 9 obviously elevated the glucose consumption in insulin-resistant cells (Fig. 4C) compared with the control (Group 3), suggesting that 4, 8, and 9 contain the most obvious hypoglycaemic activity in insulin-resistant cells among the 10 compounds. Thus, the
cytotoxicities of 4, 8, 9, and 10 were characterized, revealing that 2 to 50 M of these molecules did not have apparent cytotoxicity to FL83B cells (Fig. 4D).
3.3 Molecular mechanism underlying the insulin-like activities of 9 and 10 in normal cells
insulin-signaling pathway in 9- or 10-treated cells was analyzed. Insulin stimulation apparently raised the tyrosine phosphorylation of IRS-1 in FL83B cells (Fig. 5A, Lane 2 versus Lane 1; Lane 7 versus Lane 6), whereas treatment with 2, 20, or 30 M of 9 (Lanes 3, 4, and 5) or 10 (Lanes 8, 9, and 10) did not show this effect. Consistently, insulin obviously induced the phosphorylation of Akt (Fig. 5B, Lane 2 versus Lane 1; Lane 7 versus Lane 6), but 2-30 M of 9 (Lanes 3, 4, and 5) or
10 (Lanes 8, 9, and 10) did not. These results suggest that 9 and 10 cannot activate IRS and Akt
as insulin can.
Hypoglycaemic molecules isolated from bitter melons have been shown to activate AMP-activated protein kinase (AMPK) , which catalyzes the phosphorylation of AS160, resulting in the translocation of GLUT-4 to the cell surface and an increase in glucose uptake . Thus, whether
9 and 10 activate AMPK was examined. Fig. 5C shows that 9 and 10 both obviously activated
AMPK in concentrations of 2-30 M (Lanes 3, 4, and 5 versus Lane 1; Lanes 8, 9, and 10 versus Lane 6). Consistently, Fig. 5D reveals that, in cells treated with 9 (Lane 6), 10 (Lane 8), or troglitazone (Lane 4; a thiazolidinedione-type antidiabetic medicine demonstrated to activate AMPK , the phosphorylation of AS160 was obviously increased compared with the control (Lane 1) and similar to that in insulin-stimulated cells (Lane 2). Furthermore, the addition of Compound C (Fig. 5D), an AMPK inhibitor, obviously reduced the troglitazone- (Lane 5 versus Lane 4), 9- (Lane 7 versus Lane 6), or 10-promoted (Lane 9 versus Lane 8) phosphorylation of AMPK and AMPK substrate ACC-1 (acetyl-CoA carboxylase-1), confirming the suppression of AMPK activity, while Compound C also apparently decreased the troglitazone- (Lane 5 versus Lane 4),
9- (Lane 7 versus Lane 6), or 10-induced (Lane 9 versus Lane 8) elevation of AS160
phosphorylation. Moreover, in Fig. 5E, Compound C reduced the 9- (Group 6 versus Group 5) or
10-promoted (Group 12 versus Group 11) glucose uptake of cells. These results suggest that
AMPK likely contributes to the insulin-like activities of 9 and 10 in normal cells.
3.4 Molecular mechanisms underlying the hypoglycaemic activities of 4, 8, and 9 in insulin-resistant cells
Compounds 4, 8, and 9 obviously increased the glucose intake of insulin-resistant cells in the presence of insulin (Fig. 4C). It could be because the compound improved the sensitivity of the cell to insulin, which is a function of an insulin sensitizer, or that the compound, instead of insulin, promoted the glucose uptake of insulin-resistant cells. The latter property is defined as insulin substitute herein. These two possibilities were examined. Fig. 6A shows that, as a positive control, the phosphorylation of Akt and that of AS160 were elevated in insulin-stimulated normal cells (Lane 2 versus Lane 1; Lane 10 versus Lane 9, respectively); however, in TNF--treated cells, the insulin-induced phosphorylation of Akt and AS160 was significantly inhibited (Lane 4 versus Lane 2, Lane 12 versus Lane 10, respectively), thereby confirming insulin resistance in these cells. When TNF--treated cells were also treated with 4 in the absence of insulin, the phosphorylation of Akt (Lane 7) and AS160 (Lane 15) was not obviously elevated. Nonetheless, in the presence of insulin, the phosphorylation of these effectors was significantly promoted by 4 (Lane 8 versus Lanes 2, 4, and 7; Lane 16 versus Lanes 10, 12, and 15). These data suggest that 4 alone did not activate the insulin-signaling pathway in TNF--treated cells; it did so in the
presence of insulin, indicating that 4 is an insulin sensitizer. The effect of 8 was similar to that of
4. Compound 8 significantly increased the phosphorylation of Akt and AS160 only in the
presence of insulin (Fig. 6B, Lane 8 versus Lane 7, Lane 16 versus Lane 15, respectively), indicating that it is an insulin sensitizer. Compound 9 displayed a different result. In Fig. 6C, 9 also activated Akt only in the presence of insulin (Lane 8 versus Lane 7), suggesting that it is an insulin sensitizer. However, 9 clearly elevated the phosphorylation of AS160 in both the absence (Lane 15) and presence of insulin (Lane 16) compared with the controls (Lanes 10, 11, and 12). The results of Lanes 7 and 15 revealed that 9 could promote the phosphorylation of AS160 in an insulin- and Akt-independent manner, indicating that it is an insulin substitute in insulin-resistant cells. Therefore, 9 may be both an insulin sensitizer and insulin substitute. The effect of
troglitazone, an insulin sensitizer, is more similar to those of 4 and 8 (Figs. 6A, 6B, and 6C, Lanes 5 and 6; Lanes 13 and 14).
4. Discussion
In this study, the C. moschata stem extract was shown to contain hypoglycaemic activity in
vivo in STZ-induced diabetic mice. The hypoglycaemic components of the stem extract were
divided into an ethyl acetate-soluble portion and a butanol-soluble portion after partitioning. This suggests that the hypoglycaemic molecules in the extract are mostly hydrophobic. Thus,
extraction using water might not effectively release these hypoglycaemic components from the plant. Subsequently, 18 out of the 23 fractions divided from the ethyl acetate portion
demonstrated hypoglycaemic activity in vitro, suggesting that the ethyl acetate portion was abundant in hypoglycaemic molecules. The hypoglycaemic activities in vivo of F19 and F22 were confirmed in animal models in this study, whereas those of other fractions still need to be
verified. Consistently, the compound afforded from F22 (8) exhibited a hypoglycaemic activity in insulin-resistant cells, and 3 out of the 9 molecules afforded from F19 (4, 9, and 10) displayed hypoglycaemic activities in either normal or insulin-resistant cells. Compounds 4 and 8 exhibited an insulin sensitizing function; 9 displayed 3 functions, i.e., insulin-like activity, insulin
sensitizing, and insulin substitution; and 10 exhibited insulin-like activity. Insulin sensitizers and insulin substitutes in insulin-resistant cells might be useful in treating type 2 diabetes, whereas insulin-like activity in normal cells might be used in type 1 diabetic patients.
The activation of AMPK has been suggested to be a potential therapeutic strategy for metabolic syndrome and diabetes . The results confirmed that 9 and 10 both activate AMPK (Fig. 5). They also revealed that 9 and 10 did not activate IRS and Akt, but they did promote the phosphorylation of AS160 and glucose uptake of normal cells, and these effects were suppressed by the AMPK inhibitor. Hence, 9 and 10 are not insulin mimetic agents because they do not activate the entire insulin-signaling pathway as insulin does. Instead, they activate AMPK, which results in an insulin-like effect that increases the glucose uptake of the cell. Resveratrol and berberine, two extensively studied natural products for various functions including anti-diabetes, were shown to activate AMPK in concentrations of 3-100M and 0.1-10 M , respectively. The abilities of 9 and 10 to activate AMPK (2-30 M, Fig. 5C) are comparable to those of these
natural products. How 9 and 10 activate AMPK is unclear and requires further investigation. In Groups 4 and 10 in Fig. 5E, Compound C suppressed the insulin-stimulated glucose uptake of the cell, but to a weaker extent compared with that in 9- or 10-treated cells (Groups 6 and 12, respectively). Insulin activates AMPK (Fig. 5C, Lanes 2 and 7), which explains this inhibition by Compound C. However, the phosphorylation of AS160 is also catalyzed by Akt in stimulated cells. Thus, Compound C only partially or weakly inhibited the insulin-promoted phosphorylation of AS160 (Fig. 5E).
Compounds 4, 8, and 9 can all function as insulin sensitizers. Their structures are quite different, representing 3 distinct types of chemicals. Whether the mechanisms underlying their insulin sensitizing functions are the same or different deserves to be addressed. Compound 9 displayed an additional activity in insulin-resistant cells. It increased the phosphorylation of AS160 without insulin (Fig. 6C). The mechanism mediating this insulin substitution function also requires further investigation.
Compound 4 is also known as ferulic acid, which is an antioxidant and an anti-inflammatory agent , and has been suggested to be antidiabetic. Studies have reported that ferulic acid has a hypoglycaemic and protective effect in diabetic animal models , or alleviates the symptoms of diabetes in obese rats , but the underlying mechanism has not been addressed. In agreement with these previous reports, this study identified ferulic acid as one of the hypoglycaemic components in the crude extract of C. moschata. Our data further suggest that ferulic acid can be an insulin sensitizer. Whether this insulin sensitizing activity contributes to its antidiabetic effect in vivo
deserves to be addressed.
Four structurally similar lignans (5, 6, 7, and 8) were isolated in this study, but only 8 showed an obvious hypoglycaemic activity. The structure of 8 is most similar to that of 6. The difference between these 2 molecules is one additional methoxy group on each of the aromatic rings in 8, indicating that either one or both of these 2 additional methoxy groups are significant for the hypoglycaemic activity of 8. Likewise, 3 phenolic compounds (2, 3, and 4) were isolated, but only 4 exhibited an obvious hypoglycaemic effect. The structure of 4 is different from that of
3 only by one additional methoxy group on Carbon 3 of the aromatic ring, suggesting the
importance of this methoxy group in the activity of 4. It will be interesting to address the
structure-function relationship and insights of the methoxy group present in these hypoglycaemic compounds.
5. Conclusion
Our data suggest that the stem of C. moschata is a rich source of hypoglycaemic compounds, including those with an insulin-like, insulin sensitizing, or insulin substitution function that have the potential for treating or managing type 1 diabetes, type 2 diabetes, or insulin resistance. The hypoglycaemic molecules are mostly hydrophobic. Those in F19 and F22 of the ethyl acetate portion were characterized, but apparently there are other hypoglycaemic molecules in the butanol portion and in the other fractions of the ethyl acetate portion that deserve to be explored.
Conflict of interest
The authors declare no conflicts of interest.
Acknowledgments
The authors are grateful to the National Science Council of Taiwan for the grants to H. L. Cheng (NSC 99-2313-B-020-003-MY3; NSC 2317-B-020-003), to C. I. Chang (NSC 97-2317-B-020 -002), and to C. H. Chou (NSC 96-2317-B-039-002).
References
Acosta-Patino, J. L., Jimenez-Balderas, E., Juarez-Oropeza, M. A., & Diaz-Zagoya, J. C. (2001). Hypoglycemic action of Cucurbita ficifolia on Type 2 diabetic patients with moderately high blood glucose levels. Journal of Ethnopharmacology, 77, 99-101.
Andrade-Cetto, A., & Heinrich, M. (2005). Mexican plants with hypoglycaemic effect used in the treatment of diabetes. Journal of Ethnopharmacology, 99, 325-348.
Baboota, R. K., Bishnoi, M., Ambalam, P., Kondepudi, K. K., Sarma, S. M., Boparai, R. K., & Podili, K. (2013). Functional food ingredients for the management of obesity and associated co-morbidities - A review. Journal of Functional Foods, 5, 997-1012.
Chang, C. I., Chen, C. R., Liao, Y. W., Cheng, H. L., Chen, Y. C., & Chou, C. H. (2006).
Cucurbitane-type triterpenoids from Momordica charantia. Journal of Natural Products, 69, 1168-1171.
(2011). In vivo and in vitro studies to identify the hypoglycaemic constituents of Momordica
charantia wild variant WB24. Food Chemistry, 125, 521-528.
Cheng, H. L., Huang, H. K., Chang, C. I., Tsai, C. P., & Chou, C. H. (2008). A cell-based
screening identifies compounds from the stem of Momordica charantia that overcome insulin resistance and activate AMP-activated protein kinase. Journal of Agricultural and Food
Chemistry, 56, 6835-6843.
Cheng, H. L., Kuo, C. Y., Liao, Y. W., & Lin, C. C. (2012). EMCD, a hypoglycemic triterpene isolated from Momordica charantia wild variant, attenuates TNF-alpha-induced
inflammation in FL83B cells in an AMP-activated protein kinase-independent manner.
European Journal of Pharmacology, 689, 241-248.
Choi, H. J., Eo, H., Park, K., Jin, M., Park, E. J., Kim, S. H., Park, J. E., & Kim, S. (2007). A water-soluble extract from Cucurbita moschata shows anti-obesity effects by controlling lipid metabolism in a high fat diet-induced obesity mouse model. Biochemical and Biophysical
Research Communications, 359, 419-425.
Ciminiello, P., Fattorusso, E., Magno, S., Mangoni, A., & Pansini, M. (1990). Incisterols, a new class of highly degraded sterols from the marine sponge Dictyonella incise. Journal of the
American Chemical Society, 112, 3505-3509.
Ferguson, L. R., Zhu, S. T., & Harris, P. J. (2005). Antioxidant and antigenotoxic effects of plant cell wall hydroxycinnamic acids in cultured HT-29 cells. Molecular Nutrition & Food
Research, 49, 585-593.
salmantica. Phytochemistry, 34, 733-736.
Golden, S. H., & Sapir, T. (2012). Methods for insulin delivery and glucose monitoring in diabetes: summary of a comparative effectiveness review. Journal of Managed Care
Pharmacy, 18, S1-17.
Hardie, D. G. (2007). AMP-activated/SNF1 protein kinases: conserved guardians of cellular energy. Nature Reviews Molecular Cell Biology, 8, 774-785.
Harrison, L. J., Sia, G. L., Sim, K. Y., Tan, H. T. W., Connolly, J. D., Lavaud, C., & Massiot, G. (1995). A ferulic acid ester of sucrose and other constituents of Bhesa paniculata.
Phytochemistry, 38, 1497-1500.
Hirabara, S. M., Gorjao, R., Vinolo, M. A., Rodrigues, A. C., Nachbar, R. T., & Curi, R. (2012). Molecular targets related to inflammation and insulin resistance and potential interventions.
Journal of Biomedicine and Biotechnology, 2012, 379024.
Hosseinian, F. S., & Mazza, G. (2009). Triticale bran and straw: Potential new sources of phenolic acids, proanthocyanidins, and lignans. Journal of Functional Foods, 1, 57-64. Ishizuka, T., Yaoita, Y., & Kikuchi, M. (1997). Sterol constituents from the fruit bodies of
Grifola frondosa (FR.) S. F. Gray. Chemical & Pharmaceutical Bulletin, 45, 1756-1760.
Jiang, Z. G., & Du, Q. Z. (2011). Glucose-lowering activity of novel tetrasaccharide glyceroglycolipids from the fruits of Cucurbita moschata. Bioorganic & Medicinal
Chemistry Letters, 21, 1001-1003.
Korbayashi, M., & Ohta, Y. (1983). Induction of stress metabolite formation in suspension cultures of Vigna angularis. Phytochemistry, 22, 1257-1261.
Lee, C. K., Lu, C. K., Kuo, Y. H., Chen, J. Z., & Sun, G. Z. (2004). New prenylated flavones from the roots of Ficus beecheyana. Journal of the Chinese Chemical Society, 51, 437-441. Lee, J., Kim, D., Choi, J., Choi, H., Ryu, J. H., Jeong, J., Park, E. J., Kim, S. H., & Kim, S.
(2012). Dehydrodiconiferyl alcohol isolated from Cucurbita moschata shows anti-adipogenic and anti-lipogenic effects in 3T3-L1 cells and primary mouse embryonic fibroblasts. Journal
of Biological Chemistry, 287, 8839-8851.
Nakasone, Y., Takara, K., Wada, K., Tanaka, J., Yogi, S., & Nakatani, N. (1996). Antioxidative compounds isolated from kokuto, non-centrifugal cane sugar. Bioscience Biotechnology and
Biochemistry, 60, 1714-1716.
Otsuka, H., Takeuchi, M., Inoshiri, S., Sato, T., & Yamasaki, K. (1989). Phenolic compounds from Coix lachryma-jobi var. Ma-yuen. Phytochemistry, 28, 883-886.
Prabhakar, P. K., Prasad, R., Ali, S., & Doble, M. (2013). Synergistic interaction of ferulic acid with commercial hypoglycemic drugs in streptozotocin induced diabetic rats. Phytomedicine,
20, 488-494.
Ramar, M., Manikandan, B., Raman, T., Priyadarsini, A., Palanisamy, S., Velayudam, M., Munusamy, A., Marimuthu Prabhu, N., & Vaseeharan, B. (2012). Protective effect of ferulic acid and resveratrol against alloxan-induced diabetes in mice. European Journal of
Pharmacology, 690, 226-235.
Saltiel, A. R., & Kahn, C. R. (2001). Insulin signalling and the regulation of glucose and lipid metabolism. Nature, 414, 799-806.
pharmacotherapy decisions for U.S. medicare enrollees with diabetes. Journal of Diabetes
and Its Complications, 26, 301-307.
Shen, N., Huan, Y., & Shen, Z. F. (2012). Berberine inhibits mouse insulin gene promoter through activation of AMP activated protein kinase and may exert beneficial effect on pancreatic beta-cell. European Journal of Pharmacology, 694, 120-126.
Song, Y., Wu, T., Yang, Q., Chen, X., Wang, M., Wang, Y., Peng, X., & Ou, S. (2014). Ferulic acid alleviates the symptoms of diabetes in obese rats. Journal of Functional Foods, 9, 141-147.
Tang, W., Li, S., Liu, Y., Huang, M. T., & Ho, C. T. (2013). Anti-diabetic activity of chemically profiled green tea and black tea extracts in a type 2 diabetes mice model via different
mechanisms. Journal of Functional Foods, 5, 1784-1793.
Tanti, J. F., & Jager, J. (2009). Cellular mechanisms of insulin resistance: role of stress-regulated serine kinases and insulin receptor substrates (IRS) serine phosphorylation. Current Opinion
in Pharmacology, 9, 753-762.
Teresa, D. Q., Moran, J. R., Hernandez, J. M., & Grande, M. (1985). Tovarol and other germacrane derivatives from Thapsia villosa. Phytochemistry, 24, 1779-1783.
Thompson, A. M., Martin, K. A., & Rzucidlo, E. M. (2014). Resveratrol induces vascular smooth muscle cell differentiation through stimulation of SirT1 and AMPK. PLoS One, 9, e85495. Whitehead, J. P., Clark, S. F., Urso, B., & James, D. E. (2000). Signalling through the insulin
receptor. Current Opinion in Cell Biology, 12, 222-228.
leaf polyphenol extracts reduced hepatic lipid accumulation involving regulation of adenosine monophosphate activated protein kinase and lipogenic enzymes Journal of Functional Foods,
5, 1620-1632.
Yang, H. J., Jang, D. J., & Hwang, J. T. (2012). Anti-diabetic effects of Korean red pepper via AMPK and PPAR-gamma activation in C2C12 myotubes. Journal of Functional Foods, 4, 552-558.
Yu, H. J., Chen, C. C., & Shieh, B. J. (1998). The constituents from the leaves of Magnolia coco.
Figure Captions
Fig. 1 - Animal tests to assess the hypoglycaemic activity of ME in STZ-induced diabetic mice.
A, B, and C, diabetic mice administered with 100 mg/kg/d of ME, 50 mg/kg/d of ME, or the vehicle, respectively. D, normal mice administered with the vehicle. On the first, third, fifth, and sixth day of administration, blood glucose was measured before (0 h) and 2 h after the
administration. Data represent mean ± standard error of the mean (SEM; N = 6). *P < 0.05 versus 0-h blood glucose of the respective day; #P < 0.05 versus 0-h blood glucose of Day 1.
Fig. 2 - Glucose uptake assays for insulin-resistant cells treated with partitions or fractions of
ME. A, a flowchart showing the partitioning and further isolation of compounds of the methanol extract of C. moschata. B, C, and D, glucose uptake assays performed using insulin-resistant cells. FL83B cells were treated with TNF- for 5 h to induce insulin resistance (Groups 3-8), followed by treatment with insulin and the indicated partition or fraction for another 5 h (Groups 4-8), during which glucose concentration in the culture medium was monitored to calculate the glucose consumption of the cells. The glucose uptake relative to Group 1 was plotted as the histograms. Data represent the mean ± standard deviation (SD) of triplicate. B, assays for cells treated with 100 g/mL of ME, EA, Bu, or W. C, assays for cells treated with 1 g/mL of fraction 1. D, assays for cells treated with 1 g/mL of fractions 2-23, respectively. *P < 0.05 against Group 3 by two-way ANOVA.
Fig. 3 - The hypoglycaemic activities in vivo and constituents of F19 and F22. A to C,
STZ-induced diabetic mice were administered with 200 mg/kg/d of F19 (A), 200 mg/kg/d of F22 (B), or the vehicle (C). D, normal mice administered with the vehicle. On the first, second and third day of administration, blood glucose was measured before (0 h) and 2 h after the administration. Data represent mean ± SEM (N = 6). *P < 0.05 versus 0-h blood glucose of the respective day; #P < 0.05 versus 0-h blood glucose of Day 1. E, the structures of compounds isolated from F19 and F22. Compound 8 was isolated from F22. The others were isolated from F19.
Fig. 4 - Glucose uptake and cytotoxicity assays for compounds 1-10. A, glucose uptake assays in
normal FL83B cells. Cells were treated with the vehicle (control), 100 nM of insulin (Ins), or 20 M of the indicated compound (1-10) for 5 h, during which glucose concentration in the culture medium was monitored to determine the glucose uptake of the cells. The glucose uptake relative to control was plotted as the histograms. Data represent the mean ± SD of triplicate experiments. *P < 0.05 versus control. B, glucose uptake assays as in A, but the concentration of the indicated compound (1-8) was 40 M. C, glucose uptake assays in insulin-resistant cells. FL83B cells were treated with TNF- for 5 h to result in insulin resistance (Groups 3-8), followed by treatment with insulin and the indicated compound for another 5 h (Groups 4-8). The glucose uptake relative to Group 1 was plotted as the histograms. Data represent the mean ± SD of triplicate experiments. *P < 0.05 versus Group 3. D, cytotoxicity assays of 4, 8, 9, and 10 in FL83B cells. Cells were treated with 0 (control), 2, 20, 30, or 50M of the compound for 24 h. The percentage of viable cells relative to control was calculated. Data represent the mean ± SD of triplicate experiments, and were statistically analyzed against the control by one-way ANOVA. None of them showed a significant difference.
Fig. 5 - Characterization of the insulin-like activities of 9 and 10 in normal cells. A, B, and C,
Western blotting of the levels of tyrosine phosphorylated and total IRS-1 (A, IRS-Y-P and IRS), phosphorylated and total Akt (B, P-Akt and Akt), as well as phosphorylated and total AMPK (C, P-AMPK and AMPK) in FL83B cells, respectively. These were treated with insulin (Lanes 2 and 7), 2, 20, or 30 M of 9 (Lanes 3, 4, and 5) or 10 (Lanes 8, 9, and 10) for 30 min. Band intensities relative to Lane 1 or Lane 6 in each blot were determined after normalization by the
corresponding total protein and plotted as the respective histogram. A and B were performed twice independently, the blots were from one of the experiments and the histograms show the mean of the two experiments. D, analysis by Western blot of the levels of phosphorylated and total AS160 (P-AS160 and AS160), AMPK, and ACC-1 (P-ACC-1 and ACC-1) in FL83B cells,
which were treated with the vehicle (Lanes 1 and 3), insulin (Lane 2), 50 M of troglitazone (TZD; Lanes 4 and 5), 20 M of 9 (Lanes 6 and 7), or 20 M of 10 (Lanes 8 and 9) for 30 min. In Lanes 3, 5, 7, and 9, 30 M of Compound C (cpd C) was added 2 h before the addition of the indicated compound. Experiments were repeated 3 times independently. The blots shown were the results of one of the experiments. Band intensities relative to Lane 1 in each blot were determined after normalization by the corresponding total protein, and the mean ± SD of the 3 experiments was plotted as the histograms. E, glucose uptake assays in normal FL83B cells as in Fig. 4A. The cells were treated with the vehicle (Groups 1, 2, 7, and 8), insulin (Groups 3, 4, 9, and 10), 9 (Groups 5 and 6), or 10 (Groups 11 and 12) for 5 h. In Groups 2, 4, 6, 8, 10, and 12, 30 M of Compound C was added 2 h before the addition of the indicated chemical. The glucose uptake relative to Group 1 (the left histogram) or Group 7 (the right histogram) is presented. Data represent the mean ± SD of triplicate experiments. *P < 0.05 versus Group 1 or Group 7; #P < 0.05 between the indicated groups.
Fig. 6 - Characterization of the hypoglycaemic activities of 4, 8, and 9 in insulin-resistant cells.
Analysis by Western blot of the levels of phosphorylated and total Akt (P-Akt and Akt, Lanes 1-8), and those of AS160 (P-AS160 and AS160, Lanes 9-16) in FL83B cells. The cells were treated with TNF-(Lanes 3-8 and 11-16) for 5 h to induce insulin resistance, followed by treatment with 50 M troglitazone (TZD; Lanes 5, 6, 13, and 14), 20 M 4 (A, Lanes 7, 8, 15, and 16), 20 M 8 (B, Lanes 7, 8, 15, and 16), or 20 M 9 (C, Lanes 7, 8, 15, and 16) for 30 min. In Lanes 2, 4, 6, 8, 10, 12, 14, and 16, 100 nM of insulin was added simultaneously with the indicated compound. Experiments were repeated twice independently. The blots shown were from one of the experiments. Band intensities relative to Lane 1 or Lane 9 in each blot were determined after normalization by total Akt or AS160, and the mean of the 2 experiments was plotted as the respective histogram.
