Journal of Ethnopharmacology 172 (2015) 368–376

Contents lists available at ScienceDirect

Journal of Ethnopharmacology

http://d0378-87(http://c

n CorrE-m1 Ju

journal homepage: www.elsevier.com/locate/jep

Rhizoma Anemarrhenae extract ameliorates hyperglycemia and insulinresistance via activation of AMP-activated protein kinase in diabeticrodents

Jun Han a,1, Na Yang b,1, Feng Zhang b, Chuan Zhang c, Fengying Liang b, WeiFen Xie a,Wansheng Chen b,n

a Department of Gastroenterology, Changzheng Hospital, Second Military Medical University, 415 Fengyang Road, Shanghai 200003, PR Chinab Department of Pharmacy, Changzheng Hospital, Second Military Medical University, 415 Fengyang Road, Shanghai 200003, PR Chinac New Drug Research Center, School of Pharmacy, Second Military Medical University, 325 Guohe Road, Shanghai 200433, PR China

a r t i c l e i n f o

Article history:Received 8 January 2015Received in revised form4 May 2015Accepted 7 May 2015Available online 7 July 2015

Keywords:AMP-activated protein kinaseAnimal experimentAnti-diabetic compoundsInsulin sensitivity

Chemical compounds studied in this article:Mangiferin (PubChem CID: 5281647)Neomangiferin (PubChem CID: 6918448)Pioglitazone (PubChem CID: 4829)Streptozocin (PubChem CID: 29327)Compound C (PubChem CID: 11524144)Wortmannin (PubChem CID: 312145)STO-609 (PubChem CID: 16760660)Carboxymethyl Cellulose Sodium (PubChemCID: 6328154)Pentobarbital Sodium (PubChem CID:14075609)Bacillus Calmette-Guérin (PubChem CID:6451)

x.doi.org/10.1016/j.jep.2015.05.01641/& 2015 The Authors. Published by Elseviereativecommons.org/licenses/by-nc-nd/4.0/).

esponding author. Fax: þ86 21 81886191.ail address: chenwanshengsmmu@aliyun.comn Han and Na Yang contributed equally to th

a b s t r a c t

Ethnopharmacological relevance: Rhizoma Anemarrhenae has been used in Asian countries for thousandsof years to treat diabetes. Insulin resistance (IR) is the primary cause responsible for type 2 diabetes. Theaim of this study was to to assess the hypoglycemic and insulin sensitizing properties of Rhizoma An-emarrhenae extract (TFA) in animal models of insulin resistance and/or diabetes and to delineate modesof action.Materials and methods: In-vivo studies were performed on STZ-induced diabetic mice and KK-Ay mice,the former of which were given the extract alone or in combination with insulin for 7 days, and the latterof which were given the extract for 8 consecutive weeks. Fasting blood glucose and serum insulin levelswere measured. Pancreatic tissue sections were examined using transmission electron micrographs.Further, hyperinsulinemic–euglycemic clamping study was conducted in BCG vaccine-induced insulinresistance rats, and glucose infusion rate was examined. Mechanisms of action were investigated in 3T3-L1 and Hela cells using Western blot analysis.Results: Our study showed that TFA enhanced the glucose-lowering effects of exogenous insulin ad-ministration in STZ-induced diabetic mice. Therapeutic administration of TFA significantly reducedfasting blood glucose, and serum insulin levels, and markedly increased the size and the number ofinsulin-producing beta cells in KK-Ay mice. Further, hyperinsulinemic–euglycemic clamping studyshowed that glucose infusion rate was significantly improved in TFA-treated BCG vaccine-induced insulinresistance rats. Study of mechanism of action revealed that TFA increased phosphorylation of AMPK andits downstream target, acetyl-CoA carboxylase (ACC) in 3T3-L1 cells. It activates AMPK in a LKB1-in-dependent manner, providing a unified explanation for the beneficial effects of TFA.Conclusions: This study that TFA mediates activation of AMPK and improves overall glucose and lipidmetabolism in diabetic rodents, highlights the potential utility of TFA for the management of type2 diabetes.& 2015 The Authors. Published by Elsevier Ireland Ltd. This is an open access article under the CC BY-NC-

ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Natural principles and crude extracts of plants have been a richresource for the development of novel therapeutics used to treat avariety of human diseases such as type 2 diabetes (T2D). Ane-marrhena asphodeloides Bunge. (Asparagaceae) yields Anemar-rhenae Rhizoma, which has been commonly used in Asian

r Ireland Ltd. This is an open acces

(W. Chen).is study.

countries for thousands of years to treat various ailments, likefebrile diseases, fever, cough and diabetes (Wang et al., 2014).Modern research found that the water extracts of AnemarrhenaeRhizoma displayed significant hypoglycemic bioactitvities in dia-betic animal models (Guo et al., 2011; Hoa et al., 2004; Ichiki et al.,1998; Muruganandan et al., 2002; Muruganandan et al., 2005;Nakashima et al., 1993; Wang et al., 1995,, 2014; Yoshikawa et al.,2001). The bioassay-guided fractionation of the extract resulted inisolation of active components such as mangiferin and its gluco-side (mangiferin-7-O-β-D-glucoside), pseudoprototimosaponinAIII, Anemarans A, B, C, and D, etc. Among these compounds,mangiferin exhibits not only antihyperglycemic, but also

s article under the CC BY-NC-ND license

J. Han et al. / Journal of Ethnopharmacology 172 (2015) 368–376 369

hypolipidemic and antiatherogenic properties by reducing fastingblood glucose, plasma total cholesterol, triglycerides, low densitylipoprotein-cholesterol (LDL-C) and increasing high density lipo-protein (HDL-C) in diabetic animal models (Guo et al., 2011; Ichikiet al., 1998; Muruganandan et al., 2002, 2005; Yoshikawa et al.,2001), probably through activation of liver AMPK (Niu et al., 2012).The hypoglycemic mechanism of pseudoprototimosaponin AIIImight be related to inhibition of hepatic gluconeogenesis and/orglycogenolysis (Nakashima et al., 1993). In the Chinese Pharma-copoeia 2010, Anemarrhena asphodeloides has been described tohave the effect on treating diabetes due to internal heat andconstipation (Committee, 2010). Up to now, 108 compounds havebeen isolated from Anemarrhena asphodeloides, including steroidalsaponins, flavonoids, phenylpropanoids, alkaloids, steroids, or-ganic acids, anthraquinones, and others. However, to our knowl-edge a role for Anemarrhena asphodeloides and its ingredients inobesity-induced insulin resistance has not been previouslydemonstrated.

The prevalence of type 2 diabetes (T2D) and comorbiditiescontinues to increase (Association, 2014). Insulin resistance (IR) isthe primary cause responsible for type 2 diabetes. So there isconsiderable interest in the discovery of insulin sensitizing agentsto aid in the treatment of this disease. Metformin is such an insulinsensitizer against IR by increasing peripheral glucose uptake whilereducing hepatic glucose production in an AMPK-dependentmanner (Viollet et al., 2012).

AMPK is a major cellular energy sensor and a master regulatorof metabolic homeostasis (Zhang et al., 2009). In mammalian tis-sues, AMPK is a heterotrimeric protein kinase consisting of a cat-alytic (α) and two regulatory subunits (β and γ) (Viollet et al.,2006). AMPK is activated by two distinct signals: aCa2þ-dependent pathway mediated by CaMKKβ and an AMP-de-pendent pathway mediated by LKB1 (Sanders et al., 2007). Currentstudy suggests a close link between dysregulation of AMPK and IRin both rodents and humans. AMPK activity is diminished in adi-pose tissue of very obese insulin-resistant people (Xu et al., 2012).In model systems, sustained decreases in AMPK activity accom-pany IR, whereas AMPK activation increases insulin sensitivity(Steinberg and Kemp, 2009). On the other hand, AMPK can beactivated by exercises and calorie restriction, which have longbeen known to increase insulin sensitivity and decrease the pre-valence of T2D (Knowler et al., 2002). In addition, pharmacologicalagents developed for the treatment of type 2 diabetes, such asmetformin, TZDs, GLP1 agonists, and dipeptidyl peptidase IV (DPPIV) inhibitors, have all been shown to activate AMPK, and in sev-eral instances, have demonstrated utility in preventing the pro-gression of impaired glucose tolerance to type 2 diabetes (De-Fronzo et al., 2011; Knowler et al., 2002; LeBrasseur et al., 2006;Nawrocki et al., 2006; Svegliati‐Baroni et al., 2011; Viollet et al.,2012). So AMPK becomes an attractive target for treatment of T2D.

TFA, the Rhizoma Anemarrhenae extract, was composed of twomain compounds, mangiferin and neomangiferin (details seeSection 2.3). To the best of our knowledge, there are still no reportson the effect of TFA on diabetes. The objective of our study is toassess the hypoglycemic and insulin sensitizing properties of TFAin animal models of insulin resistance and/or diabetes and to de-lineate modes of action.

2. Material and methods

2.1. Chemicals and reagents

Compound C (CC), and STO-609 were purchased from Sigma-Aldrich (St. Louis, MO, USA). Neomangiferin and mangiferin stan-dards were from National Institute for the Control

of Pharmaceutical and Biological Products (NICPBP, Beijing, China).Wortmannin was from Selleck Chemicals (Houston, TX, USA). Re-ference substances for HPLC were of HPLC quality and purchasedfrom Merck (Darmstadt, Germany). Solvents used for extractionwere of analytical grade.

Rhizoma Anemarrhenae (radix of Anemarrhena asphadeloidesBge.) was purchased from Shanghai LeiYunshang PharmaceuticalCompany (Shanghai, China). The crude drug was identified byProfessor Wan-sheng Chen, and a voucher specimen (accessionnumber ZM-2009-10) has been deposited in Changzheng Hospital,Shanghai.

2.2. Cell lines

Murine 3T3-L1 preadipocytes and LKB1-deficient Hela cellswere obtained from Shanghai Institute of Cell Biology, ChineseAcademy of Sciences (Shanghai, China).

2.3. Preparation of the TFA extract

Dried Rhizoma Anemarrhenae (10 kg) was extracted with 70%aqueous ethanol (40 L) by percolation for 48 h. The collected ex-tracts were filtered and concentrated to yield liquid extract (about8 L) by a rotatory evaporator in vacuo. The ethanol extract waschromatographed over a ZTC-1 macroporous resin column elutedwith water and 20% ethanol later. The 20% ethanol extracts weredried in vacuo and grinded into powder for use as Rhizoma Ane-marrhenae extract (pale yellow powder, about 0.15 kg) (Chen, W.S., 2003).

The primary stock solution of neomangiferin (1.96 mg) wasprepared in 70% ethanol in a 2 mL volumetric flask to be0.98 mg/mL. The primary stock solution of mangiferin(0.98 mg/mL) was prepared in the same procedure. Then, themixed standard solutions of neomangiferin and mangiferin (both98 μg/mL) were obtained by diluting the primary stock solutionsof neomangiferin and mangiferin by 70% ethanol. The primarystock solution of TFA (2.20 mg) was prepared in 70% ethanol in a2 mL volumetric flask to be 1.10 mg/mL, from which obtained thetesting TFA solution at 110 μg/mL using 70% ethanol for dilution.

Fingerprint analysis and compound determination was per-formed using an established HPLC-DAD method (ZHANG et al.,2014). An Agilent 1200 liquid chromatography system was appliedfor analysis, equipped with a quaternary solvent delivery system,an autosampler and a photodiode array detector (DAD). A Dia-monsil C18 column (4.6 mm�250 mm, 5 μm, Dikma) was used foranalysis at 1.0 mL min�1 and at 25 °C. The mobile phase consistedof acetonitrile (eluent A) and water with 0.1% acetic acid (Tedia,United States of America, eluent B), using a linear gradient elution:2–20% A (0–30 min); 20–22% A (30–35 min); 22–26% A (35-40 min); 26–35% A (40-60 min); 35–55% A (60–75 min); 55–95% A(75–85 min). Representative HPLC chromatograms of TFA wereshown in Fig. 1. The contents of two main compounds in TFA,neomangiferin and mangiferin, were 15.32% and 46.82%,respectively.

2.4. Protein analysis

3T3-L1 preadipocytes or Hela cells were serum starved forovernight, and then treated with TFA at 30 mg/mL for 2 h at 37 °C.CC (40 mM) or STO-609 (10 mg/mL) was added 30 min before theinitiation of treatment. Western blot analysis was performed bystandard methods with antibodies directed against total AMPK,phosphorylated AMPK, total ACC, and phosphorylated ACC thatwere obtained from Cell Signaling Technology (Danvers, MA, USA).The phospho-AMPK antibody detects endogenous AMPKα1 andα2 when phosphorylated at Thr172.

Fig. 1. Representative HPLC-DAD fingerprint chromatograms of TFA at 260 nm (A), 280 nm (B) and 230 nm (C). Peaks: 1, neomangiferin; 2, mangiferin.

J. Han et al. / Journal of Ethnopharmacology 172 (2015) 368–376370

2.5. Animal studies

Mouse experiments were conducted in 6- to 8-week-old maleand female ICR, and KK-Ay mice (Institute of Laboratory AnimalSciences, Chinese Academy of Medical Science, Beijing, China). Ratexperiments were conducted in male and female Sprague-Dawley(SD) rats (180–210 g). Experimental animals unless noted werepurchased from Slac Laboratory Animal Co. Ltd. (Shanghai, China).All animal procedures were performed in accordance with theguidelines of the institutional animal care and use committee ofthe Second Military Medical University. Animals were maintainedon a 12-h light/dark cycle and fed ad libitum unless noted. Animalswere acclimated for 7 days in the research facility before initiationof the experiments. The STZ mice model was developed as follows.Normal 6-week-old ICR mice were given a single intraperitonealinjection of 200 mg/kg STZ (Sigma-Aldrich, St. Louis, MO, USA).

One week after STZ injection, mice exhibiting 6-h fasting bloodglucose410 mM were considered to have type 2 diabetes. All thediabetic animals were randomized to treatment groups based onfasting blood glucose levels (n¼16–20). TFA were administeredonce daily by oral gavage for 7 days at 30, 90, and 270 mg/kg,respectively. 0.8% Carboxymethyl cellulose sodium in distilledwater was used as the vehicle. Pioglitazone (Pio, National MedicineGroup Shanghai Chemical Reagent Co., Ltd., Shanghai, China)30 mg/kg/d was used as a positive control. 6-h fasting blood glu-cose was monitored after 7 days. One week after drug withdrawal,mice were given insulin protamine zinc (Shanghai No. 1 Biochem-ical & Pharmaceutical Co., Ltd., Shanghai, China. 0.2 U/mouse/d, s.c.) once a day for 4 consecutive weeks. The treatment was initiated3 weeks after insulin injection. Mice with similar blood glucoseconcentrations (n¼16–20) were treated with TFA or Pio for 7 daysusing the treatment regimen mentioned above. For the studies

Table 1Effects of TFA on fasting blood glucose levels in normal SD rats.

Group n Fasting blood glucose (mmol / L)

Day 0 Day 7

Vehicle 10 2.770.11 2.570.12TFA 20 mg/kg 10 2.670.09 2.570.11TFA 60 mg/kg 10 2.770.08 2.870.18TFA 180 mg/kg 10 2.670.09 2.670.09Pio 20 mg/kg 10 2.670.13 2.570.08

TFA: total flavonoids from Anemarrhena Asphodeloides Bge. Pio: pioglitazone.Results are expressed as mean7SEM.

J. Han et al. / Journal of Ethnopharmacology 172 (2015) 368–376 371

involving KK-Ay mice, mice were treated with TFA or Pio using thesame treatment regimen previously described for 8 weeks. Six-hour fasting plasma glucose was monitored weekly. Plasma insulinlevels were measured by radioimmunoassay (LINCO, St. Charles,MO). A homeostasis model assessment of insulin resistance(HOMA-IR), an index of insulin resistance, was calculated usingfasting insulin and glucose concentrations according to the fol-lowing equation (Haffner et al., 1997; Matthews et al., 1985).

HOMA IR fasting insulin IU/mL fasting glucose nmol/L

/22.5

− = ( (μ ) × ( ))

Fig. 2. Effects of TFA on fasting blood glucose levels in mice with STZ-induceddiabetes. (A) STZ-induced diabetic mice were treated orally with TFA (30, 90,270 mg/kg) or Pio (30 mg/kg) for 7 days. (B) STZ-induced diabetic mice weretreated with insulin 0.2 IU mouse per day for 21 days, and then treated with thesame dose of insulin in combination with TFA or Pio for 7 consecutive days. Veh :Vehicle. Results are expressed as mean7SEM. **Po0.01 vs. STZ/Veh group.

At the end of the study, pancreas, livers, kidneys and hearts ofKK-Ay mice were fixed in 10% buffered formaldehyde and em-bedded in paraffin, cut in 5 mm sections, and stained by standardhematoxylin/eosin procedures. Pancreatic tissue sections werealso prepared for transmission electron microscopic (TEM) ex-amination as previously described (Carlson et al., 2003).

2.6. Hyperinsulinemic–euglycemic clamps

Experiments were conducted in SD rats. As previously de-scribed (Guo et al., 2002), rats were administered a single i.v. in-jection of 10 mg Bacillus Calmette-Guérin (BCG, Shanghai Instituteof Biological Products, Shanghai, China) per rat to establish ananimal model of immune insulin resistance, and control rats wereinjected with the same volume of saline. All the BCG vaccine-in-duced insulin resistance rats (BCG-rats) were randomized totreatment groups based on fasting blood glucose levels (n¼10).Then, TFA were administered once daily orally for 14 days at 20(BCG/TFA 20), 60 (BCG/TFA 60), and 180 mg/kg (BCG/TFA 180),respectively. 0.8% Carboxymethyl cellulose sodium in distilledwater was used as the vehicle (BCG/Vehicle). Pio 20 mg/kg/d wasused as a positive control (BCG/Pio 20). After a 14-day treatmentperiod and an overnight fast, clamps were performed on theseBCG-rats and WT SD rats (n¼10). Body weight was measured, andthen a blood sample was taken from the tail for the assessment ofthe basal glucose concentration. After being anesthetized withsodium pentobarbital (30 mg/kg i.p.), an indwelling catheter forinfusion of insulin and glucose was placed into the right jugularvein. Rats were allowed to recover for 10 min, then a hyper-insulinaemic–euglycaemic clamp study was conducted in theserats. The insulin clamp was begun at t¼0 min with a primed-continuous infusion of human insulin (16 mU/kg bolus followed by8 mU kg�1 min�1). 5 min later, a continuous infusion of glucosewas given at a rate of 1 mg kg�1 min�1 for the remainder of theexperiment. Plasma glucose concentration was monitored every10 min, and the infusion rate of exogenous glucose (GIR) was ad-justed appropriately to maintain the baseline glucose level. RecordGIR at a time when a steady blood glucose level (3.8–4.0 mM,lasted at least 30 min) was reached.

2.7. Statistics

All data are expressed as mean7SE. Statistical analysis wasconducted by using Student's t test. Statistical significance wasdefined as Po0.05.

3. Results

3.1. Effects of TFA on glycemic control in normal SD rats

To evaluate the effects of TFA on glycemic control in normalrats, TFA or Pio was administered orally once daily for 7 con-secutive days. As shown in Table 1, TFA and Pio had no significanteffects on fasting blood glucose in normal SD rats.

3.2. Effects of TFA on glycemic control in mice with STZ-induceddiabetes

To generate a nongenetic rodent model mimicking human type2 diabetes, ICR mice were given a single injection of 200 mg/kgSTZ. As shown in Fig. 2, 1–2 weeks after STZ injection, the bloodglucose of untreated STZ-induced diabetic mice (STZ/Veh) rised to14.7–16.0 mmol/L, 5–6 week after STZ treatment, the blood glu-cose of STZ/Veh remained at 21.0–23.0 mmol/L, well above that ofvehicle-treated non-diabetic mice without STZ injection (Veh).Furthermore, treatment with insulin (0.2 U/mouse/d, s.c.) incombination with Pio (30 mg/kg/d, p.o.) for 7 days significantlyreduced PP fasting blood glucose (83%) in these STZ-induced dia-betic mice, validating the suitability of the STZ mouse model forpharmaceutical testing. We therefore performed a 7-day study totest the effect of TFA or Pio on glycemic control in the STZ mousemodel. As shown in Fig. 2A, treatment with TFA (30, 90, or 270 mg/kg) or Pio (30 mg/kg) alone had little effect on fasting blood

Table 2Effects of TFA on fasting blood glucose levels (mM) in KK-Ay mice.

Treatment weeks Group

Vehicle n¼13 TFA 30 n¼12 TFA 90 n¼12 TFA 270 n¼13 Pio 30 n¼13

0 12.370.9 12.671.1 12.370.9 12.471.3 12.471.31 12.871.2 11.671.5 12.171.3 11.471.0 10.171.12 15.470.9# 10.571.0** 11.671.3* 12.071.1* 8.670.6**,#

3 15.871.4 9.870.8** 10.070.7**,# 8.670.6**,# 8.470.5**,#

4 17.471.4## 9.770.8**,# 9.370.5**,## 11.071.0** 8.070.5**,##

5 17.971.6## 9.270.7**,# 9.070.6**,## 12.571.2**,# 7.670.7**,##

6 18.971.7## 8.470.7**,## 8.970.8**,## 9.770.7**,# 7.370.5**,##

7 19.471.4## 7.970.7**,## 8.770.6**,## 9.970.6**,# 6.470.3**,##

8 20.771.6## 7.770.6**,## 8.670.6**,## 8.570.6**,## 6.570.3**,##

Results are expressed as mean7SEM.n Po0.05.nn Po0.01 vs. the vehicle-treated group.# Po0.05.## Po0.01 vs. corresponding group before treatment.

J. Han et al. / Journal of Ethnopharmacology 172 (2015) 368–376372

glucose. Likewise, treatment with insulin (0.2 U/mouse/d, s.c.) for21 days caused mild reduction in fasting blood glucose (5.6–9.8%,P40.05). However, treatment with insulin (0.2 U/mouse/d, s.c.) incombination with TFA or Pio for 7 days thoroughly attenuated theincrease of fasting blood glucose levels (n¼16–20, Po0.01 vs. STZ/Veh, Fig. 2B), suggesting that TFA and insulin acted synergisticallyto improve glucose metabolism in this nongenetic mouse model oftype 2 diabetes.

3.3. Effects of TFA on metabolic control in KK-ay mice

To fully characterize the effects of TFA on metabolic control, weperformed an 8-week study in diabetic KK-Ay mice. As illustratedin Table 2 and Figs. 3, 6-h fasting blood glucose levels of the ve-hicle-treated group increased gradually and reached a peak value

Fig. 3. Effects of TFA on fasting blood glucose levels (A), fasting plasma insulin levels (B)as mean7SEM. *Po0.05, ***Po0.001 vs. the vehicle-treated group.

of 20.7 mM by the end of treatment, while that of TFA-treatedgroups decreased gradually to 7.7–8.5 mM, significantly lowerthan that of the vehicle group and of corresponding groups beforetreatment.

The effects of TFA on plasma insulin levels were also de-termined. TFA significantly reduced plasma insulin levels (Fig. 3B).Thus, TFA effectively reduced hyperglycemia and hyperinsulinemiain this KK-Ay mouse model of type 2 diabetes.

The HOMA-IR index provides a simple method to measure in-sulin sensitivity, and its use is widespread in both clinical andanimal studies (1). In our study, the HOMA-IR index was sig-nificantly decreased in the TFA groups ( Fig. 3C). Collectively, theseresults demonstrated TFA improved glycemic control and insulinsensitivity in KK-Ay mice.

, and the insulin resistance index HOMA-IR (C) in KK-Ay mice. Results are expressed

Fig. 4. Representative hematoxylin and eosin (H&E) staining (1000� ) and trans-mission electron micrographs (7500� ) of pancreatic tissue sections from KK-Ay

mice with TFA or Pio treatment for 8 weeks.

J. Han et al. / Journal of Ethnopharmacology 172 (2015) 368–376 373

3.4. Effects of TFA on pancreas, liver, kidney, and heart morphology

We next characterized the influence of different treatment re-gimens on the morphological appearance of pancreas, liver, kid-ney, and heart tissue compartments.

Fig. 4 shows typical H&E staining and transmission electronmicrographs of pancreatic tissue sections from KK-Ay mice. H&Estaining showed that treatment with TFA or Pio results in a relativenormal pancreatic structure, a more clear outline of insulin-pro-ducing beta cells which were markedly increased in numbers andsize. Using transmission electron microscope, we observed that

Fig. 5. Representative hematoxylin and eosin (H&E) staining (400� ) of liver, kidney a

beta cells in the KK-Ay pancreas showing nuclear shrinkage,chromatin condensation and margination, and granular basophiliccytoplasm, characteristic features of apoptosis. After 8-weektreatment with increasing doses of TFA or Pio, the characteristicfeatures of apoptosis were significantly alleviated.

Fatty liver disease is commonly observed in type 2 diabetes andobesity, and we therefore investigated the morphology of hepa-tocytes by light microscopy. Hepatocytes of diabetic KK-Ay micedisplayed marked morphological changes with coalescing cyto-plasmic vacuoles, which is strongly suggestive for accumulation offat droplets (Fig. 5). The increase of vacuoles was alleviated by TFAas well as Pio. Further, treatment with TFA or Pio resulted in arelative normal hepatic architecture with evenly-distributed cellsand arranged closely connective tissue, shortage of nuclearshrinkage, clear nuclear outline, characteristics of attenuation ofhepatic cellular apoptosis. Thus, these data suggest that bothagents decreased lipid accumulation associated with diabetes innon-adipose tissues such as liver.

The H&E staining of kidney tissue sections from KK-Ay mice(Fig. 5) showed significant vacuolar degeneration of renal tubularcells with displaced pycnotic nuclei, mesangial hyperplasia, andtubulointerstitial nephropathy. After 8-week treatment with TFAor Pio, the above pathological changes were apparently lessened,cellular apoptosis was alleviated, and inflammatory cell infiltrationwas also significantly attenuated.

As Fig. 5 shown, no abnormal coronary vascular and myocardialstructural changes were observed before and after treatment withTFA or Pio.

3.5. Effects of TFA on insulin sensitivity

We induced insulin resistance in SD rats by a single i.v. injec-tion of 10 mg BCG (Guo et al., 2002). Hyperinsulinemic–eu-glycemic clamp analysis also revealed these rats to be insulin re-sistant: the glucose infusion rate (GIR) was reduced by 41% in theBCG vaccine-induced insulin resistance rats (BCG-rats) comparedwith the vehicle controls (Table 3). Furthermore, GIR was sig-nificantly improved in Pio-treated BCG-rats (Table 3), demon-strating the validity of the BCG vaccine-induced insulin resistancerat model for pharmaceutical testing. As shown in Table 3, GIR rosewith increasing doses of TFA, suggesting that TFA enhance insulinaction and improve insulin-stimulated glucose disappearance inthese BCG-rats.

nd heart tissue sections from KK-Ay mice with TFA or Pio treatment for 8 weeks.

Table 3Effects of TFA on glucose infusion rates (GIR) during a hyperinsulinemic–euglycemic clamp experiment in BCG vaccine-induced insulin resistance rats.

Group Vehicle BCG/vehicle BCG/TFA 20 BCG/TFA 60 BCG/TFA 180 BCG/Pio 20

Glucose (mM) 3.870.15 4.070.14 3.870.12 3.870.13 3.870.17 3.970.14GIR (mg/kg min) 14.1770.27** 8.3970.31 13.3570.28** 14.3370.32** 14.7670.23** 14.8970.35**

BCG vaccine-induced insulin resistance rats were fasted for 6 h before the experiment. The insulin clamp was begun at t¼0 min with a primed-continuous infusion of humaninsulin 16 mU/kg bolus followed by 8 mU kg�1 min�1. Blood was acquired from the cut tail in restrained mice. Data are mean7SEM for 10 mice/group.

nn Po0.01 vs. the vehicle-treated group.

Fig. 6. TFA promote phosphorylation of AMPK and ACC in vitro. 3T3-L1 preadipocytes (A) or Hela cells (B) were treated with TFA in the presence or absence of compound Cor STO-609 for 2 h. Then cells were lysed, and phospho-specific antibodies were used to determine phosphorylation of AMPK and ACC in immunoblots. After immunoblotswere stripped, anti-AMPK and anti-ACC antibodies were used to confirm that protein loads were equal. Densitometric analyses and quantification of the ratios pAMPK/AMPKand pACC/ACC from experiments. **Po0.01 vs. basal levels. #Po0.05, ##Po0.01 vs. cells treated with the indicated compound in the absence of Compound C or STO-609.

J. Han et al. / Journal of Ethnopharmacology 172 (2015) 368–376374

3.6. Effects of TFA on AMPK phosphorylation

To investigate the mechanism of action for the antidiabeticactivity of TFA, we analyzed phosphorylation of AMPK by westernblotting. We found that TFA treatment resulted in a significantelevation of the phosphorylation at Thr-172 on the AMPKα2 sub-unit in 3T3-L1 preadipocytes. TFA also increased the phosphor-ylation of the AMPK target ACC (Fig. 6). Using an AMPK inhibitorcompound C, the ability of TFA to stimulate the phosphorylation of

AMPK and ACC was diminished (Fig. 6), suggesting that TFA pri-marily activate AMPK by increasing its phosphorylation at Thr-172.

AMPK is primarily activated by two kinases, LKB1 and CaMKKβ.We found TFA also stimulated AMPK and ACC phosphorylation inthe LKB1-deficient HeLa cells (Fig. 6). In addition, CaMKKβ in-hibitor STO-609, at a nontoxic concentration (Pang et al., 2008;Turner et al., 2008), reduced the effect of TFA on AMPK and ACCphosphorylation in 3T3-L1 preadipocytes and Hela cells (Fig. 6).These results indicated that TFA activated AMPK and ACC in a

J. Han et al. / Journal of Ethnopharmacology 172 (2015) 368–376 375

LKB1-independent fashion. Further, CaMKKβ may play a role inAMPK activation stimulated by TFA.

4. Discussion

The present study has revealed that TFA is an insulin sensitizer.TFA was shown to be an effective antihyperglycaemic agent ongenetically diabetic KK-Ay mice (Fig. 3, Table 2). The blood glu-cose-lowering action of TFA was marked and sustained (Fig. 3,Table 2). It was associated with a substantial reduction of hyper-insulinaemia and an apparent increase in pancreatic insulin stores(Figs. 3 and 4). The HOMA-IR index provides a simple method tomeasure insulin sensitivity, and its use is widespread in bothclinical and animal studies. In our study, the HOMA-IR index sig-nificantly decreased in the TFA groups (Fig. 3). Furthermore, anindependent indicator of insulin sensitivity, the glucose infusionrate was improved during hyperinsulinaemic–euglycaemic clampstudies in TFA-treated BCG vaccine-induced insulin resistance rats(Table 3). So it is not surprising that TFA and insulin act sy-nergistically to reduce blood glucose of STZ-induced diabetic mice(Fig. 2B). TFA alone had little effect on the STZ-induced diabeticmice because the inability of their pancreas to synthesize insulinprecluded any potential effects mediated by altered insulin re-sistance (Szkudelski, 2001) (Fig. 2A).

The reason we did not observe a dose response between 30 and270 mg/kg of TFA may be that the dose range studied was too high.Most clinical mangiferin studies in the literature administered 10–30 mg/kg (Miura et al., 2001; Muruganandan et al., 2002, 2005),which is equal to 21–64 mg/kg of TFA (The content of mangiferinin TFA is 46.82%). Therefore, a dose-response effect of TFA mayplateau beyond 30 mg/kg. On the other hand, though there wereno significant differences among the doses, 30, 90, and 270 mg/kgof TFA all decreased hyperglycemia and hyperinsulinemia sig-nificantly compared with 0 mg/kg (vehicle). That is, doses withinthe range of 30–270 mg/kg appeared to be equally effective.

AMPK acts as a fuel sensor, sensing the body insulin sensitivityagainst IR. We found that TFA stimulated AMPKa and ACC phos-phorylation not only in 3T3-L1 cells, but also in Hela cells whichdo not express LKB1 (Fig. 6). Furthermore, AMPK activation in Helaand 3T3-L1 cells was inhibited by treatment with the CaMKKßinhibitor STO-609, indicating that TFA-induced activation of AMPKwas partly mediated by the CaMKK signaling pathway, but not byLKB1signaling (Fig. 6). Consistent report that mangiferin activatedAMPK in liver of hyperlipidemic rats (Niu et al., 2012).

Activation of AMPK may play an important role in the hy-poglycemia effects of TFA. AMPKα2 KO animals exhibited in-creased body weight and fat mass on a high-fat diet (Villena et al.,2004). In contrast, acute hepatic overexpression of a constitutivelyactive AMPKα2 subunit results in reduced blood glucose and in-creased hepatic fatty acid oxidation (Foretz et al., 2005). Firstly,AMPK activation decreases hepatic glucose output. As an AMPKactivator, metformin has been used for decades to improve gly-cemic control in diabetic patients and is thought to decrease bloodglucose levels by reducing hepatic glucose output through in-hibition of transcription of key gluconeogenic genes in the liver(Shaw et al., 2005). Secondly, AMPK activation enhances cellularglucose transport in muscle and adipose tissue via activation ofGLUT1 and GLUT4 (Barnes et al., 2002; Zheng et al., 2001). AICAR,a well-known AMPK activator, enhanced glucose uptake in skeletalmuscle, and this effect was abolished in skeletal muscle fromAMPK inactive mutant mice (Mu et al., 2001). Thirdly, activation ofAMPK increases phosphorylation of ACC, a key enzyme in theregulation of fatty acid metabolism. The phosphorylation and in-activation of ACC results in decreased levels of malonyl CoA, whichin turn activates carnitine palmitoyltransferase 1 (CPT1) to

increase fatty acid oxidation and reduces de novo fatty acidsynthesis (Ruderman et al., 2003). Hepatocytes of diabetic KK-Aymice displayed marked morphological changes with coalescingcytoplasmic vacuoles, which are strongly suggestive for accumu-lation of fat droplets (Fig. 5). The increase of vacuoles in KK-Aymice was alleviated by TFA as well as Pio (Fig. 4). Our finding thatTFA reduced hepatocellular lipids indicates that the activation ofAMPK may account for its lipid-lowering actions. Metformin-in-duced reduction in hepatic lipid content is consistent with an in-crease in fatty acid oxidation presumably acting through hepaticAMPK activation (Zang et al., 2004). On the other hand, It is in-dicated that hepatic steatosis contributes to insulin resistance(Ryysy et al., 2000). Increased hepatic fatty acid flux has beenpostulated to play a role in this process (Day and Saksena, 2002).Both TFA and Pio reduce hepatic vacuoles accumulation in KK-Aymice, thereby probably contributing to improved insulin sensitiv-ity in both treatment groups.

Taken together, our results indicate that TFA mediates activa-tion of AMPK and improves overall glucose and lipid metabolism.

5. Conclusions

In conclusion, the present study has demonstrated that8 weeks of TFA treatment in KK-Ay mice significantly alleviateshyperglycemia and hyperinsulinaemia, and markedly increasedpancreatic insulin stores. So the HOMA-IR index markedly de-creased in the TFA groups. Furthermore, an independent indicatorof insulin sensitivity, the glucose infusion rate was improvedduring hyperinsulinaemic–euglycaemic clamp studies in TFA-treated BCG vaccine-induced insulin resistance rats, illustratingthat TFA is an insulin sensitizer by enhancing insulin signaling. Astrong proof of this notion is TFA and insulin act synergistically toreduce blood glucose of STZ-induced diabetic mice (Fig. 2B). Studyof mechanism of action revealed that TFA increased phosphor-ylation of AMPK and its downstream target ACC. This study sug-gests that TFA improves insulin sensitivity by activating AMPK, in amechanism which is independent of LKB1, providing a unifiedexplanation for the beneficial effects of TFA. This highlights thepotential utility of TFA for the management of type 2 diabetes.

Acknowledgments

This study was supported by National Natural Science Foun-dation of China 81325024, Shanghai Science and TechnologyCommittee 12431900805. The authors thank Professor Jun Wangand Xinmin Liu for their technical assistance.

References

Association, A.D., 2014. Diabetes management in correctional institutions. DiabetesCare 37, S104–S111.

Barnes, K., Ingram, J.C., Porras, O.H., Barros, L.F., Hudson, E.R., Fryer, L.G., Foufelle, F.,Carling, D., Hardie, D.G., Baldwin, S.A., 2002. Activation of GLUT1 by metabolicand osmotic stress: potential involvement of AMP-activated protein kinase(AMPK). J. Cell Sci. 115, 2433–2442.

Carlson, E.C., Audette, J.L., Veitenheimer, N.J., Risan, J.A., Laturnus, D.I., Epstein, P.N.,2003. Ultrastructural morphometry of capillary basement membrane thicknessin normal and transgenic diabetic mice. Anat. Rec. Part A: Discov. Mol. Cell.Evol. Biol. 271, 332–341.

Committee, S.P., 2010. Pharmacopoeia of the People’s Republic of China. People’sMedical Publishing House, Beijing.

Day, C., Saksena, S., 2002. Non‐alcoholic steatohepatitis: definitions and patho-genesis. J. Gastroenterol. Hepatol. 17, S377–S384.

DeFronzo, R.A., Tripathy, D., Schwenke, D.C., Banerji, M., Bray, G.A., Buchanan, T.A.,Clement, S.C., Henry, R.R., Hodis, H.N., Kitabchi, A.E., 2011. Pioglitazone for

J. Han et al. / Journal of Ethnopharmacology 172 (2015) 368–376376

diabetes prevention in impaired glucose tolerance. N. Engl. J. Med. 364,1104–1115.

Foretz, M., Ancellin, N., Andreelli, F., Saintillan, Y., Grondin, P., Kahn, A., Thorens, B.,Vaulont, S., Viollet, B., 2005. Short-term overexpression of a constitutively ac-tive form of AMP-activated protein kinase in the liver leads to mild hypogly-cemia and fatty liver. Diabetes 54, 1331–1339.

Guo, F., Huang, C., Liao, X., Wang, Y., He, Y., Feng, R., Li, Y., Sun, C., 2011. Beneficialeffects of mangiferin on hyperlipidemia in high‐fat‐fed hamsters. Mol. Nutr.Food Res. 55, 1809–1818.

Guo, Y., Luo, M., Lin, Z., 2002. [Establishment of immune insulin resistance model inthe rats by iv injection of BCG]. Yao xue xue bao¼ Acta Pharm. Sin. 37, 321–325.

Haffner, S.M., Miettinen, H., Stern, M.P., 1997. The homeostasis model in the SanAntonio heart study. Diabetes Care 20, 1087–1092.

Hoa, N.K., Phan, D.V., Thuan, N.D., Östenson, C.G., 2004. Insulin secretion is sti-mulated by ethanol extract of Anemarrhena asphodeloides in isolated islet ofhealthy Wistar and diabetic Goto-Kakizaki rats. Exp. Clin. Endocrinol. Diabetes112, 520–525.

Ichiki, H., Miura, T., Kubo, M., Ishihara, E., Komatsu, Y., Tanigawa, K., Okada, M.,1998. New antidiabetic compounds, mangiferin and its glucoside. Biol. Pharm.Bull. 21, 1389–1390.

Knowler, W.C., Barrett-Connor, E., Fowler, S.E., Hamman, R.F., Lachin, J.M., Walker, E.A., Nathan, D.M., 2002. Reduction in the incidence of type 2 diabetes withlifestyle intervention or metformin. N. Engl. J. Med. 346, 393–403.

LeBrasseur, N.K., Kelly, M., Tsao, T.-S., Farmer, S.R., Saha, A.K., Ruderman, N.B., To-mas, E., 2006. Thiazolidinediones can rapidly activate AMP-activated proteinkinase in mammalian tissues. Am. J. Physiol. Endocrinol. Metab. 291,E175–E181.

Matthews, D., Hosker, J., Rudenski, A., Naylor, B., Treacher, D., Turner, R., 1985.Homeostasis model assessment: insulin resistance and β-cell function fromfasting plasma glucose and insulin concentrations in man. Diabetologia 28,412–419.

Miura, T., Iwamoto, N., Kato, M., Ichiki, H., Kubo, M., Komatsu, Y., Ishida, T., Okada,M., Tanigawa, K., 2001. The suppressive effect of mangiferin with exercise onblood lipids in type 2 diabetes. Biol. Pharm. Bull. 24 1091–1091.

Mu, J., Brozinick Jr., J.T., Valladares, O., Bucan, M., Birnbaum, M.J., 2001. A role forAMP-activated protein kinase in contraction-and hypoxia-regulated glucosetransport in skeletal muscle. Mol. cell 7, 1085–1094.

Muruganandan, S., Gupta, S., Kataria, M., Lal, J., Gupta, P., 2002. Mangiferin protectsthe streptozotocin-induced oxidative damage to cardiac and renal tissues inrats. Toxicology 176, 165–173.

Muruganandan, S., Srinivasan, K., Gupta, S., Gupta, P., Lal, J., 2005. Effect of man-giferin on hyperglycemia and atherogenicity in streptozotocin diabetic rats. J.Ethnopharmacol. 97, 497–501.

Nakashima, N., Kimura, I., Kimura, M., Matsuura, H., 1993. Isolation of pseudopro-totimosaponin AIII from rhizomes of Anemarrhena asphodeloides and its hy-poglycemic activity in streptozotocin-induced diabetic mice. J. Nat. Prod. 56,345–350.

Nawrocki, A.R., Rajala, M.W., Tomas, E., Pajvani, U.B., Saha, A.K., Trumbauer, M.E.,Pang, Z., Chen, A.S., Ruderman, N.B., Chen, H., Rossetti, L., Scherer, P.E., 2006.Mice lacking adiponectin show decreased hepatic insulin sensitivity and re-duced responsiveness to peroxisome proliferator-activated receptor gammaagonists. J. Biol. Chem. 281, 2654–2660.

Niu, Y., Li, S., Na, L., Feng, R., Liu, L., Li, Y., Sun, C., 2012. Mangiferin decreases plasmafree fatty acids through promoting its catabolism in liver by activation of AMPK.PloS One 7, e30782.

Pang, T., Zhang, Z.-S., Gu, M., Qiu, B.-Y., Yu, L.-F., Cao, P.-R., Shao, W., Su, M.-B., Li, J.-Y.,Nan, F.-J., 2008. Small molecule antagonizes autoinhibition and activates AMP-activated protein kinase in cells. J. Biol. Chem. 283, 16051–16060.

Ruderman, N.B., Saha, A.K., Kraegen, E.W., 2003. Minireview: malonyl CoA, AMP-activated protein kinase, and adiposity. Endocrinology 144, 5166–5171.

Ryysy, L., Häkkinen, A.-M., Goto, T., Vehkavaara, S., Westerbacka, J., Halavaara, J.,Yki-Järvinen, H., 2000. Hepatic fat content and insulin action on free fatty acidsand glucose metabolism rather than insulin absorption are associated withinsulin requirements during insulin therapy in type 2 diabetic patients. Dia-betes 49, 749–758.

Sanders, M.J., Ali, Z.S., Hegarty, B.D., Heath, R., Snowden, M.A., Carling, D., 2007.Defining the mechanism of activation of AMP-activated protein kinase by thesmall molecule A-769662, a member of the thienopyridone family. J. Biol.Chem. 282, 32539–32548.

Shaw, R.J., Lamia, K.A., Vasquez, D., Koo, S.-H., Bardeesy, N., DePinho, R.A., Mon-tminy, M., Cantley, L.C., 2005. The kinase LKB1 mediates glucose homeostasis inliver and therapeutic effects of metformin. Science 310, 1642–1646.

Steinberg, G.R., Kemp, B.E., 2009. AMPK in health and disease. Physiol. Rev. 89,1025–1078.

Svegliati‐Baroni, G., Saccomanno, S., Rychlicki, C., Agostinelli, L., De Minicis, S.,Candelaresi, C., Faraci, G., Pacetti, D., Vivarelli, M., Nicolini, D., 2011. Glucagon‐like peptide1 receptor activation stimulates hepatic lipid oxidation and restoreshepatic signalling alteration induced by a high‐fat diet in nonalcoholic stea-tohepatitis. Liver Int. 31, 1285–1297.

Szkudelski, T., 2001. The mechanism of alloxan and streptozotocin action in B cellsof the rat pancreas. Physiol. Res. 50, 537–546.

Turner, N., Li, J.-Y., Gosby, A., To, S.W., Cheng, Z., Miyoshi, H., Taketo, M.M., Cooney,G.J., Kraegen, E.W., James, D.E., 2008. Berberine and its more biologicallyavailable derivative, dihydroberberine, inhibit mitochondrial respiratory com-plex IA mechanism for the action of berberine to activate AMP-activated pro-tein kinase and improve insulin action. Diabetes 57, 1414–1418.

Villena, J.A., Viollet, B., Andreelli, F., Kahn, A., Vaulont, S., Sul, H.S., 2004. Inducedadiposity and adipocyte hypertrophy in mice lacking the AMP-activated proteinkinase-α2 subunit. Diabetes 53, 2242–2249.

Viollet, B., Foretz, M., Guigas, B., Horman, S., Dentin, R., Bertrand, L., Hue, L., An-dreelli, F., 2006. Activation of AMP‐activated protein kinase in the liver: a newstrategy for the management of metabolic hepatic disorders. J. Physiol. 574,41–53.

Viollet, B., Guigas, B., Garcia, N.S., Leclerc, J., Foretz, M., Andreelli, F., 2012. Cellularand molecular mechanisms of metformin: an overview. Clin. Sci. 122, 253–270.

Wang, J., Ge, S., Chen, Q., Li, G., 1995. Hypoglycemic activity of polysaccharide ofroots of common Anemarrhena (Anemarrhena asphodeloides). Chin. Trad. Herb.Drugs 27, 605–606.

Wang, Y., Dan, Y., Yang, D., Hu, Y., Zhang, L., Zhang, C., Zhu, H., Cui, Z., Li, M., Liu, Y.,2014. The genus Anemarrhena Bunge: a review on ethnopharmacology, phy-tochemistry and pharmacology. J. Ethnopharmacol. 153, 42–60.

Xu, X.J., Gauthier, M.-S., Hess, D.T., Apovian, C.M., Cacicedo, J.M., Gokce, N., Farb, M.,Valentine, R.J., Ruderman, N.B., 2012. Insulin sensitive and resistant obesity inhumans: AMPK activity, oxidative stress, and depot-specific changes in geneexpression in adipose tissue. J. Lipid Res. 53, 792–801.

Yoshikawa, M., Nishida, N., Shimoda, H., Takada, M., Kawahara, Y., Matsuda, H.,2001. [Polyphenol constituents from Salacia species: quantitative analysis ofmangiferin with alpha-glucosidase and aldose reductase inhibitory activities].Yakugaku zasshi: J. Pharm. Soc. Jpn. 121, 371–378.

Zang, M., Zuccollo, A., Hou, X., Nagata, D., Walsh, K., Herscovitz, H., Brecher, P.,Ruderman, N.B., Cohen, R.A., 2004. AMP-activated protein kinase is required forthe lipid-lowering effect of metformin in insulin-resistant human HepG2 cells.J. Biol. Chem. 279, 47898–47905.

Zhang, B.B., Zhou, G., Li, C., 2009. AMPK: an emerging drug target for diabetes andthe metabolic syndrome. Cell Metab. 9, 407–416.

Zhang, F., Qi, Y., Lian-Na, S., Shou-Hong, G., Xia, T., Wan-Sheng, C., 2014. Fingerprintanalysis of Zhimu-Huangbai herb pair and simultaneous determination of itsalkaloids, xanthone glycosides and steroidal saponins by HPLC-DAD-ELSD.Chin. J. Nat. Med. 12, 525–534.

Zheng, D., MacLean, P.S., Pohnert, S.C., Knight, J.B., Olson, A.L., Winder, W.W., Dohm,G.L., 2001. Regulation of muscle GLUT-4 transcription by AMP-activated proteinkinase. J. Appl. Physiol. 91, 1073–1083.

Glossary

AMPK: AMP-activated protein kinase;;ACC: acetyl-CoA carboxylase;;BCG: Bacillus Calmette–Guérin;;CC: compound C;;CPT1: carnitine palmitoyltransferase 1;;DPP IV: dipeptidyl peptidase IV;;DAD: photodiode array detector;;GIR: glucose infusion rate;;HDL-C: high density lipoprotein;;HOMA-IR: homeostasis model assessment of insulin resistance;;IR: Insulin resistance;;LDL-C: low density lipoprotein-cholesterol;;STZ: streptozotocin;;SD: Sprague-Dawley;;S.C: Subcutaneous injection;;TFA: Rhizoma Anemarrhenae extract;;TEM: transmission electron microscopic;T2D: type 2 diabetes;;Veh: Vehicle..