Original Paper
Abstract: The antihyperglycemic effect of caffeic acid, one of the phenolic compounds contained in the fruit of Xanthium strumarium, was investigated. After an intravenous injection of caffeic acid into diabetic rats of both streptozotocin-induced and insulin-resistant models, a dose-dependent decrease of plasma glucose was observed. However, a similar effect was not pro-duced in normal rats. An insulin-independent action of caffeic acid can thus be considered. Otherwise, this compound reduced the elevation of plasma glucose level in insulin-resistant rats re-ceiving a glucose challenge test. Also, glucose uptake into the isolated adipocytes was raised by caffeic acid in a concentration-dependent manner. Increase of glucose utilization by caffeic acid seems to be responsible for the lowering of plasma glucose. Key words:Xanthium strumarium, Compositae, caffeic acid, an-tihyperglycemic effect, insulin-independent action, glucose challenge test, glucose uptake.
Introduction
The fruit of Xanthium strumarium (Compositae), named as Chang©ErZi or Chang©ErChao in Chinese (1), is known to contain carboxyatractyloside, xanthanol, isoxanthanol and hydroqui-none (2). A previous report indicated the hypoglycemic effect in animals receiving the extract of this herb (3). In this paper, we report the isolation and identification of caffeic acid as one of the active principles for lowering plasma glucose.
Materials and Methods General experimental procedures
M.p.s were determined on a Yanagimoto micromelting point apparatus and are uncorrected. Optical rotations were meas-ured with a Jasco DIP-140 digital polarimeter. IR spectra were obtained with a Jasco A-100 spectrometer. The1H- and13 C-NMR spectra were taken with a Bruker AM-300WB spec-trometer operating at 300 MHz and 75 MHz, respectively, us-ing TMS as an internal standard. FAB-MS was performed in a Jeol JMS-HX 110 and X-ray crystallography was determined by
means of a CAD4 under Kappa Axis single crystal XRD. TLC was carried out on a precoated Kieselgel 60 F254 plates (0.2 mm thick, Merck) with the solvent system C6H6 -HCOOEt-HCOOH (1:7:1 and 3:6:1); TLC detection was made by UV il-lumination and by spraying with 10% H2SO4followed by heat-ing (for thiazine) and FeCl3reagent (for phenolics).
Plant material
The fruits of Xanthium strumarium (Compositae), Xanthii fruc-tus, were purchased from a folk medicinal drug store in Taipei and authenticated by Mr. M. T. Kao, Department of Botany, Na-tional Taiwan University. The voucher specimens (TMCP85-10) have been deposited in the herbarium of the School of Pharma-cy, Taipei Medical College, Taipei City, Taiwan.
Extraction and isolation
Xanthii fructus (6 kg) were chopped into small pieces and ex-tracted three times with 70 % aqueous acetone at room tem-perature. The extract, after removal of the acetone by evapo-ration in vacuo, was fractionated with column chromatogra-phy on Diaion HP-20, eluted with H2O-MeOH with increasing MeOH content to afford 3 fractions. Through the assay of hy-poglycemic activity, fr. 2 was found to be the effective one. The fr. 2 was further chromatographed on a Sephadex LH-20 column (4 80 cm) with 60% aqueous MeOH (3 L), and with acetone to afford caffeic acid (260 mg) (4). NMR analysis indi-cated the trans-form of caffeic acid. The purity of this com-pound was not less than 99.8%.
Animals
Male Wistar rats, weighing 230±250 g, were obtained from the animal center of National Cheng Kung University Medical College. They were maintained in a temperature-controlled room (25 1 8C) and kept on a 12:12 light-dark cycle (light on at 06:00 h). Food (Purina Rat Chow) and water were availa-ble ad libitum. Rats were treated in accordance with the NIH Guide for the Care and Use of Laboratory Animals.
Detection of plasma glucose
Plasma glucose was measured in duplicate with the enzymat-ic glucose method by an autoanalyzer (Quik-Lab. Ames, Miles Inc., USA) using a blood sample from the femoral vein of rats
Caffeic Acid as Active Principle from the Fruit of Xanthium
strumarium to Lower Plasma Glucose in Diabetic Rats
Feng-Lin Hsu
1, Yun-Chueh Chen
2, Juei-Tang Cheng
2,*
1Department of Medicinal Chemistry, School of Pharmacy, Taipei Medical College, Taipei City, Taiwan, R.O.C. 2Department of Pharmacology, College of Medicine, National Cheng Kung University, Tainan City, Taiwan, R.O.C.
Revision accepted: September 12, 1999; Received: June 22, 1999
Planta Medica 66 (2000) 228± 230 Georg Thieme Verlag Stuttgart·New York
ISSN: 0032-0943
Received: June 22, 1999; Accepted: September 12, 1999
228
under anesthesia with pentobarbital (30 mg/kg, i.p.). Rats were made diabetic by a bolus injection of streptozotocin (60 mg/kg) dissolved in citrate buffer (pH 4.5) according to our previous method (5). In addition to polyuria, rats with the plasma glucose levels higher than 400 mg/dL were considered diabetic for the experiments. Also, insulresistance was in-duced in rats aged 8 weeks by the i.p. injections of long-acting human insulin (MonotardHM) at 0.5 IU/kg three times daily as described previously (6). Rats with the plasma glucose lev-el higher than 120 mg/dL in addition to the symptoms of hy-perphagia and/or increase of body weight were considered as diabetic. This model is considered similar to patients with non-insulin-dependent diabetes mellitus (NIDDM) (7). Assay of antihyperglycemic activity
Antihyperglycemic activity was determined in fasted rats that received an intravenous (i.v.) injection of aqueous solution containing caffeic acid at desired doses. Results were calculat-ed as percentage decrease of initial value according to the for-mula: {[(Gi ± Gt) / Gi] 100%} where Gi was the initial value and Gt indicated the value after treatment. Plasma glucose was measured every 30 min except when stated otherwise in the results. Values of plasma glucose were then compared with the control group that was injected with the same vol-ume of vehicle. After the investigation, animals were left for more than 3 days. Then, the diabetic rats with a plasma glu-cose value that had returned to the initial value within 10 mg/dL were identified as recovery from the effect and they were used for other studies. As a positive control, metformin was used in the present study.
Glucose challenge test
The glucose challenge test was performed in NIDDM rats, as described previously (8). In pentobarbital-anesthetized rats, glucose (60 mg/kg) was injected into the femoral vein using a 50% (W/V) solution. Blood samples (0.25 mL) were taken 10 min before, and every 10 min after glucose injection for 90 min. Blood loss was replaced by injection of equivalent vol-umes of isotonic saline solution. Changes of plasma glucose in rats treated with caffeic acid were compared with those from the vehicle-treated controls.
Determination of glucose uptake
The epididymal fat pads were removed from the Wistar rats under anesthesia with pentobarbital (30 mg/kg, i.p.). The adi-pose tissue was minced and digested with collagenase (Type I, Sigma) in Kreb©s Ringer buffer (KRB) containing 1% (W/V) bo-vine serum albumin as described in our previous report (9). Af-ter digestion for 30 min at 37 8C under shaking (120 cycles/ min), fatty cells were obtained from the supernatant by centri-fugation (1,200 rpm, 5 min). They were filtered and washed three times in KRB. The packed cells were adjusted to a suitable dilution (around 105 to 106 cells/ml) for experiments. Uptake of glucose was measured as described previously (15) using 2-deoxy-D-[14C]glucose (2-DG) (323 mCi/mmol, NEN) as indica-tor. Fat-cell suspensions were incubated in plastic vials (final volume 1 ml) in which drugs were added. Uptake was initiated by the addition of 2-DG (0.002 mM) and stopped by addition of 0.1 mM phloretin (0.75 ml, Sigma). Each vial was mixed with 0.5 ml silicone oil (D = 0.99) to obtain adipocytes in the
super-natant after centrifugation (6,000 rpm, 3 min). The adipocytes were placed in scintillation vials and 3 ml scintillation fluid was added and the radioactivity was counted. Nonspecific up-take, extracellular 2-DG present in the obtained fraction, was assayed with cells treated under similar conditions except that 50 mM cytochalasin B (Sigma) was incubated. Specific uptake of glucose was calculated after subtraction of the nonspecific uptake from total uptake, samples were not treated with cyto-chalasin B in parallel. The effect on spontaneous uptake of glu-cose was obtained from adipocytes that have been co-incubat-ed with the caffeic acid at the indicatco-incubat-ed concentration for 30 min followed by an incubation with 2-DG for 5 min to ob-tain the specific uptake. The specific uptake from samples incu-bated with vehicle (KRB) at the same volume was taken as 100% control. All the data of specific uptake were expressed as percentage of that from a control run in parallel.
Statistical analysis
All values, expressed as mean s.e., were obtained from the number (N) of samples. Student©s t-test for paired compari-sons between prior to and after administration, or unpaired observation for between vehicle-treated control and tested data were carried out for statistical evaluation of the differen-ces; p values of 0.05 or less were considered significant. Results and Discussion
By using Diaion HP-20 and Sephadex LH-20 gel column chro-matography, caffeic acid was isolated from the aqueous ace-tone extract of Xanthii Fructus. Based on the NMR spectral da-ta of a previous report (4), the obda-tained compound was iden-tified as trans-caffeic acid.
Intravenous injection of caffeic acid (0.5±3 mg/kg) into both streptozotocinduced diabetic rats (STZ-rats) and rats with in-sulin-resistance (NIDDM-rats), produced a marked plasma glu-cose lowering effect in a dose-dependent manner (Fig.1). Anti-hyperglycemic activity by caffeic acid at 3 mg/kg in STZ-rats was 23.8 8.8% which was similar (P > 0.05) to that in NIDDM-rats (33.4 5.7%). As the positive control, oral administration of met-formin at 100 mg/kg produced 18.4 3.1% antihyperglycemic activity in STZ-rats. Similar treatment with metformin at 10 mg/ kg produced 28.2 4.5% antihyperglycemic activity in NIDDM-rats. The effects of caffeic acid reached a plateau within 30 min and were maintained for 40 min or more. The duration of effect for caffeic acid was progressively lengthened as the dosage was increased but the effects were completely reversed within 90 min (data not shown). No irreversible decrease in plasma glu-cose level was observed. However, the plasma gluglu-cose level in normal rats, Wistar rats, was not modified by caffeic acid at the same doses (Fig.1). Lowering of plasma glucose may be induced by the release of insulin, an endogenous peptide in regulation of blood sugar (10). Activity of insulin was considered negligible in this STZ-induced diabetic rat (11) and rats with insulin-resist-ance (7). An insulin-related mechanism in the obtained antihy-perglycemic effect seems unlikely. Caffeic acid is introduced as an antioxidant (12) that has been used to prevent human low-density lipoprotein oxidation (13). Actually, involvement of oxi-dative stress in diabetic disorders has been documented (14) and renal antioxidant enzyme mRNA levels were increased in the diabetic rat (15). However, antihyperglycemic effect of caffe-ic acid has not been mentioned before.
Caffeic Acid as Active Principle from the Fruit of Xanthium strumarium to Lower Plasma Glucose in Diabetic Rats Planta Med. 66 (2000) 229
Then, a glucose challenge test was performed in rats with insu-lin-resistance according to the previous method (8). An intrave-nous injection of glucose produced hyperglycemia rapidly within 10 min in insulin-resistant rats that received only vehicle and the plasma glucose level returned to the initial value at 60 min after glucose injection. Pretreatment with caffeic acid (1 mg/kg, i.v.) attenuated this elevation of plasma glucose (Fig. 2). Inhibi-tion of this glucose elevaInhibi-tion in plasma can be due to an increase of glucose utilization in animal (8). An increase of glucose uptake into isolated adipocytes by caffeic acid was also obtained in a concentration-dependent manner (Fig. 3). Caffeic acid at 10mM
made an increase of glucose uptake into adipocytes (143 21.4%) in a way similar (P > 0.05) to that induced by 0.1 IU/ml ofinsulin (159.1 34.4%). In the in vitro assay, involvement of en-dogenous insulin is negligible. Increase of glucose utilization by caffeic acid through an insulin-independent action can thus be considered.
Acknowledgements
We thank Mr. M. T. Kao (Department of Botany, National Tai-wan University) for the kind authentication of the plant and Mr. T. C. Chi (Department of Pharmacology, College of Medi-cine, National Cheng Kung University) for the kind handling of animals. The present study is supported in part by a grant from National Science Council of the Republic of China (NSC86-2314-B006-029-M13).
References
1Cai JF. Advanced textbook on traditional Chinese medicine and
pharmacology. Vol. 2, New World Press, Beijing 1995: pp. 29±30
2Huang KC. The Pharmacology of Chinese Herbs. CRC Press, Boca
Raton 1993; 1st Ed. pp. 160±78n,
3Chang HM, But PPH. Pharmacology and applications of Chinese
Materia Medica. World Scientific Publishing Co., Singapore 1986; 1st Ed., Vol. 1: pp. 589
4Morishita H, Iwahashi H, Osaka N, Kido R. J. Chromatograph.
1984; 315: 253±60
5Hsu FL, Lai CW, Cheng JT. Planta Med. 1997; 63: 323 ±5 6Chi TC, Liu IM, Cheng JT. Chin. Pharm. J. 1998; 50: 113±21 7Taylor SI, Accili D, Imai Y. Diabetes 1994; 46: 735±40
8Spraul M, Anderson EA, Bogardus C, Ravussin E. Diabetes 1994;
43: 191±6
9Chang CJ, Kao JT, Lee TL, Lai CW, Cheng JT. J. Auton. Nerv. Syst.
1996; 61: 191±4
10Bolli GB. Diabetes Metab. 1997; 23: 29±35
11Chang SL, Lin JG, Chi TC, Liu IM, Cheng JT. Diabetologia 1999; 42:
250±5
12Kono Y, Kobayashi K, Tagawa S, Adachi K, Ueda A, Sawa Y, Shibata
H. Biochim. Biophy. Acta 1997; 1335: 335±42
13Nardini M, D©Aquino M, Tomassi G, Gentili V, Di Felice M, Scaccini
C. Free Radic. Biol. Med. 1995; 19: 541±52
14Ceriello A, Falleti E, Motz E, Taboga C, Tonutti L, Ezsol Z, Gonano F,
Bartoli E. Horm. Metab. Res. 1998; 30: 146±9
15Sechi LA, Ceriello A, Griffin CA, Catena C, Amstad P, Schambelan
M, Bartoli E. Diabetologia 1997; 40: 23±9
Prof. Juei-Tang Cheng
Department of Pharmacology, College of Medicine National Cheng Kung University
Tainan City Taiwan 70101 R.O.C.
E-mail: jtcheng@mail.ncku.edu.tw Fax: +886-6-238-6548
Fig.1 Effect of caffeic acid on the plasma glucose level in normal rats (Wistar), streptozotocin-induced diabetic rats (STZ) and rats with insulin-resistance (NIDDM). Caffeic acid was injected into rats intrave-nouslyat the indicated dose to compare with vehicle-treated control (C). Each point indicates the mean of value from 6±8 rats with error (s.e.) bar. *P < 0.05, **P < 0.01 and ***P < 0.001 vs. control (C), respectively. As a positive control, oral administration of metformin at 100 mg/kg decreased the plasma glucose by76.5 4.3 mg/dL in strep-tozotocin-induced diabetic rats (STZ) and similar treatment at 10 mg/ kg lowered 42.5 4.9 mg/dL in rats with insulin-resistance (NIDDM).
Fig. 2 Effect of caffeic acid on the plasma glucose level in insulin-re-sistant (NIDDM) rats receiving the glucose challenge test. Caffeic acid (1 mg/kg) injected into the femoral vein (open square) was compared with the control group that received a similar injection of saline at the same volume (solid square). Then the glucose challenge test was per-formed byan intravenous injection of glucose (60 mg/kg) into two groups of NIDDM rats at 30 min later and the plasma glucose in sam-ples obtained immediatelywas indicated as 0 min. Values of mean with error (s.e.) bar were obtained from 6 rats in each group. *P < 0.05 and **P < 0.01 vs. data from control group, respectively.
Fig. 3 Effect of caffeic acid on the glucose up-take into isolated adipo-cytes. Each point indi-cating the percentage of control represents the mean from 6 sam-ples with s.e. bar. *P < 0.05 and **P < 0.01 vs. control (C), respectively.
Planta Med. 66 (2000) Feng-Lin Hsu et al.
230
