Available online at www.sciencedirect.com

ScienceDirect

Journal of Nutritional Biochemistry 26 (2015) 82–90

Ellagic acid modulates lipid accumulation in primary human adipocytes and humanhepatoma Huh7 cells via discrete mechanisms

Meshail Oklaa,b,1, Inhae Kanga,b,1, Da Mi Kima, Vishnupriya Gourinenia, Neil Shayc,Liwei Gua, Soonkyu Chunga,b,⁎

aDepartment of Food Science and Human Nutrition, University of Florida, Gainesville, FL, USAbDepartment of Nutrition and Health Sciences, University of Nebraska, Lincoln, NE, USAcDepartment of Food Science and Technology, Oregon State University, Corvallis, OR, USA

Received 12 April 2014; received in revised form 11 September 2014; accepted 11 September 2014

Abstract

Previously, we have reported that consumption of a muscadine grape phytochemical powder (MGP) decreased lipid accumulation in high-fat fed mice. Theaim of this study was to identify the responsible polyphenolic constituents and elucidate the underlying mechanisms. In mice, MGP supplementationsignificantly reduced visceral fat mass as well as adipocyte size. To determine whether MGP affects adipogenesis or hypertrophic lipid accumulation, we used ahuman adipogenic stem cell (hASCs) model. Among the MGP, ellagic acid (EA) was identified as a potent negative regulator of adipogenesis of hASCs. In addition,EA substantially decreased the conversion of [3H]-acetyl CoA into fatty acids (FAs), suggesting that EA inhibits de novo synthesis of FA in mature adipocytes.Similarly, MGP supplementation significantly decreased hepatic triglyceride (TG) levels. The TG-lowering effects of EA were confirmed in human hepatoma Huh7cells. EA reduced [3H]-oleic acid esterification into [3H]-TG as well as the de novo synthesis of FA from [3H]-acetyl CoA in Huh7 cells. Intriguingly, EA alsoincreased oxygen consumption rate and β-oxidation-related gene expression. Taken together, EA attenuated new fat cell formation and FA biosynthesis inadipose tissue, while it reduced the synthesis of TG and FA and increased FA oxidation in the liver. These results suggest that EA exerts unique lipid-loweringeffects both in adipose tissue and liver via discrete mechanisms.© 2015 Elsevier Inc. All rights reserved.

Keywords: Muscadine grape; Ellagic acid; Adipogenesis; Obesity; Liver lipid

1. Introduction

Liver andadipose tissueare twomajor organs regulatingwholebodylipid metabolism. Excessive lipid accumulation in adipose and liver is ahallmark of obesity and metabolic syndrome. Obesity is characterizedby abnormal expansion of white adipose tissue either by increasing thenumber of adipocytes from mesenchymal progenitor cells (adipocytehyperplasia) or by simply increasing cell size (hypertrophy). Thenumberof adipocytes is tightly controlled afterpuberty [1]; however, anabnormal increase in adipocyte number is evident in childhood obesityand also frequently associated with extreme obesity in adults [2]. Incontrast, enlargement of the adipocytes is the most common mecha-nism to accommodate surplus energy in the form of triglyceride (TG) inadults. It is well established that adipocyte hypertrophy occursconcurrently with inflammation [3,4]. Enlarged and inflamed adipo-

⁎ Corresponding author at: Department of Nutrition and Health SciencesUniversity of Nebraska-Lincoln, 316G Ruth Leverton Hall, P.O. Box 830806Lincoln, NE 68583, USA. Fax: +1 402 472 0806.

E-mail address: schung4@unl.edu (S. Chung).1 Equal contribution.

http://dx.doi.org/10.1016/j.jnutbio.2014.09.0100955-2863/© 2015 Elsevier Inc. All rights reserved.

,,

cytes are key contributors to the pathogenesis of obesity by impairingendocrine function of adipocytes [5]. Although it is controversial thathyperplastic expansion of subcutaneous fat could be a defensemechanism to attenuate lipotoxicity, adipocyte hyperplasia also resultsinmetabolically unfavorable ectopic fat accumulation and increases therisk for cardiovascular diseases [6], suggesting that both hyperplasticandhypertrophic expansionof adipocytes are associatedwith adipocyteremodeling during the pathogenesis of metabolic syndrome [7].

The development of hepatic steatosis typically accompaniesobesity. It has been reported that 76% of those with nonalcoholicfatty liver disease are obese [8]. Redistribution of fat in lipodystrophicconditions such as diabetes, or uncontrolled lipolysis from inflamedadipocytes increases the influx of free fatty acid (FFA) into the portalvein leading to hepatic steatosis [9]. Conversely, fatty liver is asignificant risk factor for hyperlipidemia, insulin resistance anddiabetes [10].

Controlling obesity by limiting the adipocyte's capacity for eitherhypertrophic or hyperplastic growth is a general target of weight loss-promoting dietary supplements. However, detrimental consequencesmay occur if excess FFAs are redirected into the liver. The use of somedietary compounds that were claimed to be effective in attenuatingadiposity, e.g., trans-10, cis-12 conjugated linoleic acid, had to bereevaluated due to their adverse side effects on hepatic steatosis

83M. Okla et al. / Journal of Nutritional Biochemistry 26 (2015) 82–90

[11,12]. Recently, we reported that supplementation ofmuscadine grapephytochemicals (MGP) decreased visceral fat mass in mice [13]. Inaddition, MGP supplementation was effective in reducing systemic andretinal inflammation, and glucose intolerance [13,14]. However, fattyacid (FA) partitioning and/or redistribution between adipose tissue andliver byMGP has not been fully investigated. The aims of this studywere(1) to identify the polyphenolic compounds responsible for the lipid-lowering effects of MGP and (2) to investigate the metabolic modifica-tions caused by MGP in adipocytes as well as in hepatocytes. Here wedemonstrate that EA inhibits adipogenesis and decreases lipid accumu-lation both in mature human adipocytes and hepatocytes via distinctmechanisms. Our results also suggest that EA may constitute aconsumer-friendly dietary strategy that may effective in reducing lipidaccumulation both in adipose tissue and liver.

2. Materials and methods

2.1. Chemicals

Fetal bovine serum (FBS) was purchased from Cellgro. Rosiglitazone (BRL49653)waspurchased fromCaymanChemical. Ellagic acid (EA), quercetin (Quer),myricetin (My)andkaempferol (KMP) were purchased from Sigma-Aldrich.

2.2. Animals

Adipose tissue and liver samples were collected from our previous study [13].Briefly, male C57BL/6 mice were fed with low-fat (LF), high-fat (HF; 60% kcal from fat)diet alone or HF plus 0.4% MGP for 15 weeks. Diet composition is found in Gourineniet al. [13]. At the time of necropsy, freshly isolated epididymal fat and liver tissue waseither fixed in 10% buffered formaldehyde for paraffin embedding or snap-frozen withliquid nitrogen and kept at−80°C. All protocols and procedures were approved by theInstitutional Animal Care and Use Committee at the University of Florida.

2.3. Preparation of MGP and subfractionation

Preparation and composition of MGP has been described [13]. Briefly, MGPs wereprepared from Noble muscadine grapes by acidified-methanol extraction (0.5% aceticacid) and subsequent purification using Amberlite FPX66 column (Rohm Hass). Thesubfractionation of MGP into non-anthocyanin (NAcy) and anthocyanin (Acy) wascompleted as described previously [15]. The analysis of phytochemical composition ofNAcy was performed on an Agilent 1200 HPLC system (Agilent Technologies). Briefly, atotal of 5 mg of NAcy in 1 ml of HCl/methanol solution (1:1) was analyzed by HPLCusing an Agilent Zorbax Stablebond SB-C18 column (250 mm×4.6 mm, 5-μm particlesize). The binary mobile phase consisted of 0.5% aqueous formic acid (A) andacetonitrile (B); the flow rate was mL/min and a 25-min gradient was used (thegradient was 0–5 min of 10%–30% B, 5–10 min of 30%–40% B, 10–20 min of 40%–50% B,20–23 min of 50%–70% B, 23–25 min of 70%–100% B, followed by 5 min ofreequilibration of the column before the next run). The detection wavelength for theEA and flavonols was 360 nm using UV–Vis diode array detector (Fig. 3A).

2.4. Cell culture and phytochemical treatment

For isolation of human adipose-derived stem cells (hASCs), abdominal adiposetissue was obtained from females with a body mass index of ~30 kg/m2 duringliposuction or abdominoplasty surgery. Isolation of hASCs and differentiation ofadipocytes was conducted as we described previously [16,17]. All protocols andprocedures were approved by the institutional review board (#693-2011) at theUniversity of Florida and the University of Nebraska-Lincoln. To examine the effects ofdifferent MGP phytochemicals on adipocyte differentiation (Figs. 2 and 3), confluentcultures of hASCs (d0 preadipocyte) were induced to differentiate in an increasing doseof MGP subfraction (0–25 μg/ml of NAcy or Acy), or 10 μmol/l of individualphytochemicals of EA, Quer, My and KMP in human adipocyte medium (AM-1,ZenBio) in differentiation cocktail for 3 days (d1–d3 adipocyte). Upon 4 days ofdifferentiation (d4), medium was changed with AM-1 plus fresh phytochemicalwithout differentiation cocktail. Medium was then replenished with fresh addition ofeach phytochemical every other day and harvested on day 10 of differentiation. Alltreatments were paralleled with vehicle controls (0.1% DMSO). The experimentaldesign to investigate the effects of EA on lipid metabolism in mature adipocytes issummarized in Fig. 4A.

Huh7 cells were a kind gift from Dr. Jae Sung Kim (University of Florida). Huh7 cellsweremaintained inDulbecco'smodification of Eagle'smediumcontaining1% L-glutamine,10% FBS, 100 units/ml penicillin, 100g/ml streptomycin in 5% CO2 at 37°C.

2.5. Adipocyte size measurement

Hematoxylin and eosin (H&E)-stained sections of epididymal adipose tissue wereused for size determination by following thepublished protocol by Chen et al. [18]. Briefly,adipocyte sizewas examined by analyzing digital images of H&E-stained paraffin sections(n=300 cells/mouse, total=1500–2000 adipocytes from 7–9 mice/diet group) by usingImage J software.

2.6. Lipid accumulation

The colorimetric TG quantification kit (BioVision, K622‐100) was used to quantifyhepatic TG content according to the manufacturer's protocol. To measure lipidaccumulation in human adipocytes and Huh7 cells, cells were fixed with isopropanoland stained with oil-red O (ORO). For determination of relative TG content, bright fieldimages were taken by EVOS XL microscope (AMG), or ORO dye was extracted forquantification (OD 500 nm).

2.7. [3H]-oleic acid and [3H]-acetyl CoA incorporation into FFA and TG

To measure FA esterification rate into TG, we followed the previously publishedmethods [19] in cultures of mature adipocytes or human hepatoma Huh7 cells. Briefly,cells were incubated with serum-free low glucose media (1000 mg/l D-glucose)overnight before experiments. [3H]-oleic acid (OA) (Perkin Elmer; final concentrationof 0.5 μCi/ml in 0.5×106 cells) was complexed with FA-free BSA, then added to cells for3 h. Incorporation of [3H]-OA into TG was increased linearly over a 6-h period (data notshown). After 3-h incubation with [3H]-OA, medium containing unincorporatedisotope was removed by washing with phosphate-buffered saline (PBS). Cellular lipidswere extracted as published [20]. Then thin layer chromatography was performed tofractionate FFA and TG, and the [3H] was measured by liquid scintillation counting(Beckman LS 6000; Beckman Instruments, Palo Alto, CA, USA). Radioactivity wasnormalized by protein concentration quantified by bicinchoninic acid colorimetricassay (Pierce, Rockford, IL, USA). Similarly, for the measurement of de novo synthesis ofFA, [3H]-acetyl CoA (Perkin Elmer; final concentration of 0.5 μCi/ml in 0.5×106 cells)was added to cells for 3 h and unincorporated isotope was removed by washing withPBS three times.

2.8. [3H] 2-deoxy-glucose uptake

To determine basal and insulin-stimulated glucose uptake, cultures of maturehuman adipocytes were incubated with or without 10 μmol/l of EA for 3 days. The daybefore the experiment, cultures were incubated with 1 ml serum-free basal mediumcontaining 1000 mg/l D-glucose and 20 pmol/l human insulin (Thermo Scientific;SH30021.01) in the presence of vehicle or treatment. After 24 h in serum-free media,culture media was removed and replaced with 1 ml of HBSS buffer containing 100nmol/l human insulin for 10 min, followed by addition of [3H]-2DOG (Perkin Elmer;final concentration was 0.5 μCi/ml) and incubation at 37°C for 90 min. Glucose uptakewas terminated by adding 1 ml of stop buffer [ice-cold Krebs–Ringers bicarbonate(KRBC) buffer supplemented with 25mmol/l D-glucose]. After washing cells with KRBCbuffer three times to reduce background radioactivity, cells were lysed in 0.1% SDS.Glucose uptake was determined by liquid scintillation counting [21].

2.9. Oxygen consumption rate

Huh7 cells were seeded into 96-well clear bottom black polystyrene sterile plate(Corning). Oxygen consumption rate (OCR) was determined by using MitoXpress(Cayman Chemical, 600800) according to the manufacturer's protocol. Briefly, anincrease of phosphorescent signal from an oxygen-sensitive probe wasmeasured every3 min for 5 h using a Synergy H1 multi-mode microplate reader (BioTek).

2.10. Quantitative polymerase chain reaction

Gene-specific primers for quantitative polymerase chain reaction (qPCR) wereobtained from Integrated DNA Technologies (Chicago, IL, USA). Total RNA of hASCs wasisolated using Trizol reagent (Invitrogen). To remove genomic DNA contamination,RNA was treated with DNase (Mediatech); 2 μg of RNA was converted into cDNA in atotal volume of 20 μl (iScript cDNA synthesis kit; Bio-Rad). Gene expression wasdetermined by real-time qPCR (CFX96; Bio-Rad), and relative gene expression wasnormalized by 36B4 (primer sequences are available upon request).

2.11. Western blot analysis

To prepare tissue lysates, 0.1 g of snap-frozen tissue was homogenized in ice-coldRIPA buffer (Thermo Scientific) with protease inhibitors (Sigma) and phosphataseinhibitors (2 mmol/l Na3VO4, 20 mmol/l β-glycerophosphate and 10 mmol/l NaF).Adipose tissue lysate was incubated on ice for 10 min to remove the solidified fat cake.To prepare total cell lysates, monolayers of differentiated cultures of human adipocyteswere harvested with RIPA buffer. Proteins were fractionated using 4%–15% gradientSDS-PAGE (Biorad), transferred to PVDF membranes with a semidry transfer unit(Hoefer TE77X) and incubated with the relevant antibodies. Chemiluminescence from

84 M. Okla et al. / Journal of Nutritional Biochemistry 26 (2015) 82–90

ECL (Western Lightning) solution was detected with FluorChem E (Cell Biosciences)imaging system. Polyclonal or monoclonal antibodies targeting to FA synthase (FAS)(3180), β-actin (4967) and H3K9Ac (AcH3, 9649) were purchased from Cell SignalingTechnology. Antibodies targeting histone deacetlylase 9 (HDAC9; ab 59718) andhistone H3 (ab1791) were purchased from Abcam. The mouse monoclonal antibodiesfor PPARγ (sc-7273) and FABP (aP2, sc-271529) were purchased from Santa CruzBiotechnology.

2.12. Statistical analysis

Results are presented as the mean±S.E.M. The data were statistically analyzedusing Student's t test or one-way analysis of variance (ANOVA) with Tukey's multiplecomparison tests. For the analysis of adipocyte size, Gaussian curve fitting and linearregression was performed. To calculate OCR, linear regression (95% confidence, Pb.05significant) was conducted. All analyses were performed with GraphPad Prism6 (Version 6.02).

3. Results

3.1. MGP supplementation attenuated adipocyte size in C57BL/6 mice

Previously, we reported that the supplementation of MGP (Vitisrotundifolia) significantly reduced HF-diet induced epididymal fat mass[13]. However, the mechanistic details bywhichMGP supplementationreduced adiposity are unknown. To determine whether MGP decreasesadipocyte hypertrophy, we first examined epididymal adipocyte size byanalyzing digital images of H&E-stained paraffin sections. Consistentwith the reduced epididymal fat mass, adipocyte size was significantlyreduced with MGP supplementation compared to HF alone (Fig. 1A;86.8±2.8, 100.6±3.3 and 86.5±1.4 μm for LF, HF and HF+MGP,respectively). Histograms of adipocyte size distribution also demon-strated a clear shift toward smaller sizes for the HF+MGF group,comparable to LF control (Fig. 1B, C).

3.2. EA in MGP is a potent negative regulator of adipogenesis

To gain insight into whether MGP supplementation also decreaseshyperplastic expansion of adipocytes (adipogenesis) as well ashypertrophic expansion of adipocytes (Fig. 1), we examined theexpression of proteins that are known to influence adipogenesis. Aswe expected, PPARγ expression levels were higher in the HF group in

Fig. 1. MGP supplementation attenuated fat mass and adipocyte size. C57BL/6 mice were feadipocytes from n=7-9 mice per group. Means not sharing a common superscript differ sigB, Adipocyte cell size histograms for all animals within each diet group. The line of best fit isadipose tissue for each diet.

comparison to the LF or HF+MGP group. Interestingly, HDAC9expression, a negative regulator of adipogenesis [22], was reduced inthe HF group compared to LF or HF+MGP. Conversely, histone lysine 9acetylation (H3K9Ac) levels, a positive epigenetic marker for adipocytedifferentiation, wasmarkedly increased in the HF group compared to LFor HF+MGP (Fig. 2A), suggesting that adipocyte differentiation mightbe reduced by MGP supplementation. To pursue this possibility, weconducted in vitro studies using hASCs. First, we fractionated MGP intoAcy and NAcy fractions. The effect of each fraction on adipocytedevelopmentwas examined. The Acy fraction had significant butminorimpact on adipogenesis. In contrast, the NAcy fraction dramaticallysuppressed lipid accumulation in a dose-dependentmanner as assessedby ORO staining (Fig. 2B, C). Consistent with reduced TG accumulation,the NAcy fraction of MGP dramatically suppressed adipogenic geneexpression including PPARγ and adipocyte protein 2 (aP2) (Fig. 2D).Next, we performed HPLC analysis and found that the NAcy fractioncomposed of four major polyphenols, EA, My, Quer and KMP (Fig. 3A).When hASCs were exposed to 10 μmol/l of these individual purepolyphenols, EA almost exclusively repressed the adipogenesiscompared to the other polyphenols. In an in vitro model of humanadipocytes, EAdecreased: (1) TGaccumulation (Fig. 3B); (2) adipogenicgene expression including PPARγ, CCAAT/enhancer binding proteinalpha (C/EBPα), aP2 and FAS (Fig. 3C); and (3) adipogenic proteinexpression including PPARγ, aP2 and FAS (Fig. 3D).

3.3. EA attenuated lipid accumulation in mature adipocytes

Next, we asked whether EA is the key polyphenolic componentantagonizing adipocyte hypertrophy as observed in Fig. 1, as well assuppressing adipogenesis. To answer this question, fullydifferentiatedcultures of human adipocytes were exposed to EA for 3 or 7 days,depending on the experimental design (Fig. 4A). Exposure to 10 μmol/lEA for 7 days caused a significant reduction of TG accumulationmeasured by ORO staining (Fig. 4B). To test whether the reduction ofTG accumulation was due to an alteration of lipogenic pathways, [3H]-acetyl CoA and [3H]-OA was incubated with the adipocytes, andconversion into TG was measured. The conversion of [3H]-acetyl CoAto radiolabeled FFA (Fig. 4C) and TG (Fig. 4D) was almost completely

d LF, HF, or HF +0.4% MGP for 15 weeks. A, Average size (diameter) of epididymalnificantly (Pb.05) by one-way ANOVA with Tukey's multiple comparison test (aNb).shown (Gaussian curve fitting). C, Representative H&E staining images of epididymal

Fig. 2. NAcy constituents of MGP were associated with decreased of adipocyte differentiation. (A) Protein expression of PPARγ, aP2, HDAC9, H3K9Ac and H3 by Western blot analysisfrommice fed with either LF, HF or HF+0.4%MGP for 15weeks. Cultures of hASCs were differentiated and incubated with either Acy or NAcy fractions for 10 days. (B) TG accumulationin 96 well culture plates was visualized by ORO staining (upper panel). Extracted ORO staining was quantified (OD 500 nm) (lower panel). (C), Bright-field images with ORO stainingfor cultures differentiated with different doses of NAcy and Acy. (D) Adipogenic gene expression of PPARγ and aP2 by qPCR analysis. In panels B and D, values are presented as themean±S.E.M. Means not sharing a common superscript differ significantly (Pb.05) by one-way ANOVA with Tukey's multiple comparison test (aNbNc).

85M. Okla et al. / Journal of Nutritional Biochemistry 26 (2015) 82–90

blunted by 3 days of EA treatment. However, exposure of EA exerted nosignificant impact on conversion of [3H]-OA into [3H]-TG at 3 days(Fig. 4E). In addition, basal and insulin-stimulated [3H]-2-deoxyglucoseuptake were not affected by EA treatment (Fig. 4F). Seven days of EAtreatment decreased adipocyte-specific mRNA expression includingPPARγ, C/EBPα and FAS compared to vehicle control; however, therewere no significant changes in adipogenic gene expression after EAtreatment for 3 days (data not shown). There were no significantchanges in FA oxidation-related gene expression [PPARα and carnitinepalmitoyltransferase 1 (CPT1)], but lipolysis-related gene expression(i.e., hormone-sensitive lipase and adipocyte TG lipase) was substan-tially lower in EA-treated adipocytes (Fig. 4G). Taken together, thesedata implicate that the inhibition of de novo synthesis of FA isaccompanied by transcriptional regulation of lipogenic gene expressionin EA-treated human adipocytes.

3.4. MGP decreased hepatic lipid accumulation in C57BL/6 mice

Attenuation of adipogenesis could cause hepatic steatosis if the livermishandles FA influx from adipose tissue [23]. To determine the effectsof MGP on hepatic lipid metabolism, we also measured hepatic TGcontent. H&E staining showed a reduction of hepatic lipid accumulationin HF+MGP vs. the HF group (Fig. 5A), which was confirmed bymeasurement of a ~44% reduction of hepatic TG content in theMGP-fedgroup compared to the HF group (146±6.3 vs. 83±6.8; Fig. 5B).Interestingly, FA oxidation-related genes including PPARα, fibroblastgrowth factor 21 (FGF21), acyl-CoA oxidase 1 (ACOX1) and CPT1 weresignificantly higher in the MGP group (Fig. 5C), suggesting that MGP

decreased hepatic TG accumulation, at least in part, by augmentinghepatic FA oxidation as well as attenuating adipocyte expansion.

3.5. EA attenuated lipid accumulation in Huh7 cells

To determinewhether EA is the primary polyphenolic compound inreducinghepatic TGaccumulation,we further examined effects of EAonFA esterification, de novo synthesis and FA oxidation in humanhepatocarcinomaHuh7cells. PretreatmentwithEA for 24h significantlyattenuated lipid accumulation in a dose-dependent manner in Huh7cells (Fig. 6A). In EA-treated samples, expression of mRNAs related toβ-oxidation such as PPARα and CPT1 was up-regulated; however,ACOX1 was not. In contrast, genes involved in lipogenesis, i.e., FAS anddiacylglycerol acyltransferase 2 (DGAT2), were significantly reduced byEA treatment without affecting stearoyl-CoA desaturase 1 (SCD1) geneexpression (Fig. 6B). Accordingly, the OCR, measured by an oxygen-sensitive phosphorescent probe, was higher in Huh7 cells treated withEA than vehicle control (Fig. 6C) demonstrating an up-regulation ofFA oxidation by EA in hepatocytes. Interestingly, uptake of [3H]-OAinto cells was similar between groups (Fig. 6D), while incorporation of[3H]-OA into TG was significantly lower in Huh7 cells treated with EA(Fig. 6E). Similar toadipocytes, incorporation of [3H]-acetyl CoA into FFAand TG was markedly decreased with EA treatment (Fig. 6F, G).Collectively, our data clearly showed that MGP and its activepolyphenolic constituent EA decrease hepatic lipid accumulation bytargeting multiple mechanisms including FFA synthesis, TG esterifica-tion and FFA oxidation.

Fig. 3. EA suppressed adipogenesis in hASCs. Cultures of hASCs were induced to differentiation in the presence of either vehicle (Veh) or 10 μmol/l of NAcy components of EA, Quer, Myand KMP. (A) HPLC chromatogram showing major phenolic constituents of NAcy fraction in MGP. (B) Effects of individual polyphenols on TG accumulation measured by ORO staining.(C) Adipogenic gene expressions of PPARγ, aP2, C/EBPα and FAS by qPCR analysis. (D) Adipogenic protein expression of PPARγ, FAS and aP2 byWestern blot analysis. In panels B and C,values are presented as the mean±S.E.M. Means not sharing a common superscript differ (Pb.05) by one-way ANOVA with Tukey's multiple comparison test (aNbNc).

86 M. Okla et al. / Journal of Nutritional Biochemistry 26 (2015) 82–90

4. Discussion

Obesity and hepatic steatosis are the two manifest phenotypes ofmetabolic syndrome and are inextricably linked together. The simul-taneous reduction of lipid accumulation both in adipose tissue and liverwould be ultimate goal for the dietary intervention strategies. Thepresent studywasdesigned todetermine the TG-lowering effect ofMGPsupplementation on adipose tissue and liver, and to identify themetabolic alterations caused by EA by using a human model ofadipocytes and hepatocytes in parallel. We demonstrated that thesupplementation of MGP attenuated hypertrophic obesity (Fig. 1) andhepatic steatosis (Fig. 5) in HF-fed mice. EA has been identified as theactive polyphenolic compound that suppressed the hyperplasticexpansion of adipocytes (Figs. 2, 3); EA also exerted the distinctivelipid-lowering properties to decrease biosynthesis of FA in bothadipocytes and hepatocytes, but augmented FA oxidation only inhepatocytes (Figs. 4, 6). Our dataprovide the first evidence that EAplaysseparate roles inmanipulating excess lipid differently in adipocytes andhepatocytes, resulting in a synergistic attenuationof obesity andhepaticsteatosis. Collectively, our results suggest that EA-containing foodsmayconstitute a novel and effective dietary strategy to prevent and/or treatobesity and metabolic syndrome.

Muscadine grape (V. rotundifolia) contains an array of health-promoting bioactive phytochemicals [24–26]. Our previous study hasprovided the implication that unique lipid-loweringpropertyofMGPmaybe attributed to the high content of EA [13]. In the present study, weproved that EA produces a broad spectrum of actions by targetingmultiple mechanisms, i.e., adipocyte differentiation, de novo synthesis ofFA, FA esterification and FA oxidation, probably through differentmechanisms in liver and adipose tissue. Numerous reports suggest thatEA-containing fruits, vegetables and nuts are effective dietary sources

to attenuate obesity. However, direct evidence of the underlyingmechanism of how EA displays an antiobesity effect has not yet beenaddressed. To better understand the relationship between EA intake andadiposity, we evaluated results from the peer-reviewed literature, anddespite variation related to the differences in sources and contents ofEA, we observed that a daily intake of EA in the range of 5–88 mg/kgBW correlated with a N25% decrease of fat mass and improvement ofglucosemetabolism. (SupplementaryTable1, [13,27–33]). Theconclusiondeduced from this review is consistent with our present data that EA-enriched MGP is associated with a reduction of fat mass, adipocytehypertrophy andhepatic lipid accumulation. Interestingly, themuscadinewine extract that has almost identical phytochemical composition exceptfor a reduced EA content due to filtration [34], displayed reduced lipid-lowering effects. This provides us with additional rationale to draw theconclusion that EA is a key ingredient of MGP to reduce fat mass.

It has been well documented that EA possesses antiproliferative andanti-inflammatory characteristics in various transformed cell lines[27,35]. Moreover, EA has shown to be effective in reducing atheroscle-rotic lesions [36] and increasing cholesterol efflux inmacrophages [37]. Inthis study, we added the previously unappreciated function of EA as alipid-lowering dietary compound. The reduction of adiposity is attributedto a reduction of both hyperplastic and hypertrophic expansion ofadipocytes. The inhibitory effects ofMGP (or EA) on adipogenesis seem tobe associated, at least partly, with epigenetic modification (Figs. 2, 3).Recently, EAwas identified as anegative regulator of histone3arginine17methylation (H3R17me2) by inhibiting coactivator-associated argininemethyltransferase 1 (CARM1) [38]. Given to the fact that CARM1 activityis required for adipogenesis [39], we have recently reported thatepigenetic regulation of adipogenesis by EA in the same primary humanadipocytes [16]. In that study, we demonstrated that EA alters epigeneticmarkers of adipocytedifferentiationby alteringHDACactivity, acetylation

Fig. 4. EA decreased TG accumulation in mature human adipocytes. (A) Experimental scheme. EA (10 μmol/l) was added to the newly differentiated human adipocytes (at day 7) andincubated for 3–7 days. Experiments were conducted at the given time (arrows). (B) Lipid accumulation was visualized by ORO staining. (C) Conversion of [3H]-acetyl CoA into [3H]-FA.(D) Conversion of [3H]-acetyl CoA into [3H]-TG. (E) Conversion of [3H]-OA into [3H]-TG. (F) Basal and insulin-stimulated [3H]-2-deoxyglucose (DOG) uptake. (G) Relative geneexpression levels of PPARγ, C/EBPα, FAS, PPARα and CPT1 by qPCR. In panels C–F, data were normalized by protein concentration. All values are presented as the mean±S.E.M., *Pb.05,**Pb.01, ***Pb.001, **** Pb.0001 and ns=not significant by Student's t tests.

87M. Okla et al. / Journal of Nutritional Biochemistry 26 (2015) 82–90

and methylation levels [16]. It may align well with the study by Wanget al. [40] showing that EA inhibits differentiation of 3T3-L1 preadipocyteinto adipocytes by inhibiting mitotic clonal expansion, although primaryadipogenic progenitor cells do not entermitotic clonal expansion and theinvolvement of CARM1 needs to be verified in 3T3-L1 cells.

The attenuation of the hypertrophic expansion was clearlydetectable both in epididymal fat of MGP-fed mice (Fig. 1) as wellas in the EA-treated mature adipocyte cultures (Fig. 4). A reduction ofFA biosynthesis in mature adipocytes (Fig. 4C) was an earlier eventthan transcriptional down-regulation of adipogenic gene expression(Fig. 4G). Conversely, FA esterification into TG, glucose uptake (bothbasal- and insulin-stimulated) and FA oxidation were not altered byEA at the same period of time when FA biosynthesis is markedlyreduced by EA. It is unlikely that the inhibition of FA synthesis is dueto diminished CARM1 activity, as we do not observe the differences inH3R17me2 levels in mature adipocytes before and after EA treatment(data not shown).

The impact of MGP and EA on lipid metabolism in the liver wassignificant:MGP supplementation decreased hepatic TG content by ~50%compared to HF alone, probably through enhanced FA oxidation (Fig. 5).In an agreement with our finding, Yoshimura et al. [41] reported thatsupplementation of 0.1% of EA for 68 days was effective in decreasing

hepatic steatosis by increasing mRNA expression of PPARα and CPT1αgenes inKK-Aymice, amodel of obesity and type2diabetes. Consistently,we demonstrated that EA lowered TG accumulation (Fig. 6A), increasedFA oxidation-related gene expression (Fig. 6B) and, more importantly,OCR (Fig. 6C) in human hepatoma cells. Although less evident in MGP-fedmice, EAdecreased FAuptake,de novo synthesis, and its esterificationinto TG in Huh7 cells (Fig. 6D–G). These TG-lowering effects are alsoobserved in EA-fed rats [33], and our own pilot studywithmice fed pureEA(unpublisheddata). TheexactmechanisticnaturebywhichEA targetsto multiple metabolic pathways in the liver is uncertain. Some potentialmechanisms suggesting that EA represses de novo synthesis of FA couldbe found: Sarikaya et al. [42] showed that EA is an inhibitor of carbonicanhydrase (CA) activity, and a reduction of CA activity has beenassociated with lowering hepatic de novo lipid synthesis in primary rathepatocytes [43].Weare currently investigating the twopossibilities that(1) EA may change epigenetic markers such as H3R17me2 by CARM1[38], leading to transcriptional inhibition of lipogenic gene expressionin hepatocytes, and (2) EA directly inhibits lipogenic enzyme activitiesincluding FAS and ACC activity.

Free EA can be found at maximal concentrations of ~1 μmol/l afteroral administration [44] due to limited solubility and rapid metabolismby gutmicrobes [45]. Our current study design has limitations for direct

Fig. 5. MGP supplementation decreased hepatic lipid accumulation. C57BL/6 mice were fed either HF or HF +0.4% MGP for 15 weeks. (A) Representative H&E staining for liver tissues.(B) TG content in liver (n=7–8 mice per group). (C) Relative mRNA levels of PPARα, FGF21, ACOX1 and CPT1 in the liver by qPCR. All values are presented as the mean±S.E.M. *Pb.05and ****Pb.0001 by Student's t tests.

88 M. Okla et al. / Journal of Nutritional Biochemistry 26 (2015) 82–90

translation to humans: we used 10 μmol/l EA in 0.1% DMSO andconversion of EA intometabolites by gut bacteria is not occurring in cellculture experiments. We chose 10 μmol/l of EA to mimic our previousanimal study [13] in which mice were fed diet containing 0.4% of MGP

Fig. 6. EA decreased TG accumulation in human hepatoma Huh7 cells. Huh7 cells were preincubstaining in Huh7 incubatedwith 5 or 10 μmol/l of EA. (B) Relative mRNA expression levels invol(n=9per group). Data expressed as relative phosphorescent unit per minute. (D) Uptake of [3Hinto [3H]-FA. (G) Conversion of [3H]-acetyl CoA into [3H]-TG. In panel A, means not sharing a comthe mean±S.E.M. *Pb.05, **Pb.01, and ***Pb.001 and ns=not significant by Student's t test.

(18.2mgEA/g dry extract). This dietary dose provided ~288 μg of EAperday based upon ~4 g of diet consumed per mouse. Assumingapproximately 1%–2% of EA may be absorbed by the gut and a totalblood volume of ~1.5 ml, EA concentrations will approach 3 μg/ml

ated overnight with (+) or without (−) EA. (A) Lipid accumulation quantified by OROved in FA β-oxidation and lipogenesis by qPCR. (C) OCR in the presence or absence of EA]-OA into cells. (E) Conversion of [3H]-OA into [3H]-TG. (F) Conversion of [3H]-acetyl CoAmon superscript differ, Pb.05 by one-way ANOVA. In panels B–G, data are presented as

89M. Okla et al. / Journal of Nutritional Biochemistry 26 (2015) 82–90

(equivalent to ~10 μmol/l) in blood. Secondly, we chose the lowest EAconcentrations based on the 10- to 50-μmol/l range which has beenroutinely used for cellular studies without cytotoxicity in various celltypes [35,36,46,47]. Intriguingly, treatment of physiologically achiev-able concentrations of EA (b1 μmol/l) required a longer period of time todetect measurable reductions of TG levels in hepatocytes andadipocytes (unpublished data), implicating that reduction of lipidaccumulation could be attainable via chronic supplementation of anEA-containing diet. We are currently investigating whether individualurolithins, the metabolites of EA, recapitulate the lipid-lowering effectsof EA, perhapswithgreater biological potency thanEA itself. Our currentstudy does not include informationwhether (1) supplementation of EAalone will be sufficient to exert lipid-lowering effects, or the otherphytochemicals in MGP (Acys and other floavonols such as Quer, Myand KMP) will exert synergistic roles. The latter will be more likely inphysiological conditions, especially in enhancing bioavailability byfacilitating uptake of EA in intestinal lumen or by protecting fromcatabolic degradation. Supporting this concept, it has been suggestedthat EA from natural sources exerts stronger antioxidant activity thanpure EA [48] and Quer potentiates anticancer activity of EA [49],although chronic supplementation of pure EA supplementation alonewas also effective in reducing fat pad size and liver weight in rats [33].

In summary,we identify EA as the primary polyphenolic componentinMGPwhich lowers TG levels, and delineated themetabolic pathwaysthat affected by EA in adipocytes and hepatocytes. There remainunanswered questions regarding physiological levels of EA andgeneration of microbial metabolites, but we believe, nonetheless, thatourworkprovidesmechanistic insights into lipid-lowering effects of EA.Further study is essential to understand themultiple metabolic benefitsof EA-containing foods.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.jnutbio.2014.09.010.

Acknowledgments

We thank MeeAe Lee at the University in Florida for her technicalassistance. This work was supported by the Institute of Food andAgricultural Science at the University of Florida (USDA-Hatch).

References

[1] Spalding KL, Arner E, Westermark PO, Bernard S, Buchholz BA, Bergmann O, et al.Dynamics of fat cell turnover in humans. Nature 2008;453:783–7.

[2] Drolet R, Richard C, Sniderman AD, Mailloux J, Fortier M, Huot C, et al. Hypertrophyand hyperplasia of abdominal adipose tissues in women. Int J Obes (Lond) 2008;32:283–91.

[3] Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, Ferrante Jr AW.Obesity is associated with macrophage accumulation in adipose tissue. J ClinInvest 2003;112:1796–808.

[4] Winkler G, Kiss S, Keszthelyi L, Sapi Z, Ory I, Salamon F, et al. Expression of tumornecrosis factor (TNF)-alpha protein in the subcutaneous and visceral adipose tissuein correlation with adipocyte cell volume, serum TNF-alpha, soluble serum TNF-receptor-2 concentrations and C-peptide level. Eur J Endocrinol 2003;149:129–35.

[5] Skurk T, Alberti-Huber C, Herder C, Hauner H. Relationship between adipocyte sizeand adipokine expression and secretion. J Clin Endocrinol Metab 2007;92:1023–33.

[6] Britton KA, Fox CS. Ectopic fat depots and cardiovascular disease. Circulation2011;124:e837–41.

[7] Sun K, Kusminski CM, Scherer PE. Adipose tissue remodeling and obesity. J ClinInvest 2011;121:2094–101.

[8] Farrell GC, Larter CZ. Nonalcoholic fatty liver disease: from steatosis to cirrhosis.Hepatology 2006;43:S99–S112.

[9] YoussefW,McCullough AJ. Diabetesmellitus, obesity, and hepatic steatosis. SeminGastrointest Dis 2002;13:17–30.

[10] Savage DB, Semple RK. Recent insights into fatty liver, metabolic dyslipidaemiaand their links to insulin resistance. Curr Opin Lipidol 2010;21:329–36.

[11] Poirier H, Shapiro JS, Kim RJ, Lazar MA. Nutritional supplementation with trans-10,cis-12-conjugated linoleic acid induces inflammation of white adipose tissue.Diabetes 2006;55:1634–41.

[12] Clement L, Poirier H, Niot I, Bocher V, Guerre-MilloM, Krief S, et al. Dietary trans-10,cis-12 conjugated linoleic acid induceshyperinsulinemiaand fatty liver in themouse.J Lipid Res 2002;43:1400–9.

[13] Gourineni V, Shay NF, Chung S, Sandhu AK, Gu L. Muscadine grape (Vitisrotundifolia) and wine phytochemicals prevented obesity-associated metaboliccomplications in C57BL/6J mice. J Agric Food Chem 2012;60:7674–81.

[14] Ha Jung-Heun, Kumar Shil Pollob, Zhu Ping, Gu Liwei, Li Qiuhong, Chung Soonkyu.Ocular inflammation and endoplasmic reticulum stress are attenuated bysupplementation with grape polyphenols in human retinal pigmented epitheliumcells and in C57BL/6 mice. J Nutr 2014;144:1–14.

[15] Sandhu AK, Gu L. Antioxidant capacity, phenolic content, and profiling of phenoliccompounds in the seeds, skin, and pulp of Vitis rotundifolia (muscadine grapes) asdetermined by HPLC-DAD-ESI-MS(n). J Agric Food Chem 2010;58:4681–92.

[16] Kang I, Okla M, Chung S. Ellagic acid inhibits adipocyte differentiation throughcoactivator-associated arginine methyltransferase 1-mediated chromatin modi-fication. J Nutr Biochem 2014;25:946–53.

[17] Zhao L, Ha JH, Okla M, Chung S. Activation of autophagy and AMPK by gamma-tocotrienol suppresses the adipogenesis in human adipose derived stem cells. MolNutr Food Res 2014;58:569–79.

[18] Chen HC, Farese Jr RV. Determination of adipocyte size by computer imageanalysis. J Lipid Res 2002;43:986–9.

[19] Chung S, Gebre AK, Seo J, Shelness GS, Parks JS. A novel role for ABCA1-generatedlarge pre-beta migrating nascent HDL in the regulation of hepatic VLDLtriglyceride secretion. J Lipid Res 2010;51:729–42.

[20] Bligh EG, Dyer WJ. A rapid method of total lipid extraction and purification. Can JBiochem Physiol 1959;37:911–7.

[21] Chung S, LaPoint K, Martinez K, Kennedy A, Boysen SM, McIntosh MK. Preadipocytesmediate lipopolysaccharide-induced inflammation and insulin resistance in primarycultures of newly differentiated human adipocytes. Endocrinology 2006;147:5340–51.

[22] Chatterjee TK, Idelman G, Blanco V, Blomkalns AL, Piegore Jr MG, Weintraub DS,et al. Histone deacetylase 9 is a negative regulator of adipogenic differentiation.J Biol Chem 2011;286:27836–47.

[23] He W, Barak Y, Hevener A, Olson P, Liao D, Le J, et al. Adipose-specific peroxisomeproliferator-activated receptor gamma knockout causes insulin resistance in fatand liver but not in muscle. Proc Natl Acad Sci U S A 2003;100:15712–7.

[24] Banini AE, Boyd LC, Allen JC, Allen HG, Sauls DL. Muscadine grape products intake,diet and blood constituents of non-diabetic and type 2 diabetic subjects. Nutrition2006;22:1137–45.

[25] Pastrana-Bonilla E, Akoh CC, Sellappan S, Krewer G. Phenolic content andantioxidant capacity of muscadine grapes. J Agric Food Chem 2003;51:5497–503.

[26] Yi W, Fischer J, Akoh CC. Study of anticancer activities of muscadine grapephenolics in vitro. J Agric Food Chem 2005;53:8804–12.

[27] Adams LS, Zhang Y, Seeram NP, Heber D, Chen S. Pomegranate ellagitannin-derived compounds exhibit antiproliferative and antiaromatase activity in breastcancer cells in vitro. Cancer Prev Res (Phila) 2010;3:108–13.

[28] KohGY,McCutcheonK, Zhang F, LiuD,CartwrightCA,MartinR, et al. Improvement ofobesity phenotype by Chinese sweet leaf tea (Rubus suavissimus) components inhigh-fat diet-induced obese rats. J Agric Food Chem 2011;59:98–104.

[29] Lee JH, Johnson JV, Talcott ST. Identification of ellagic acid conjugates and otherpolyphenolics in muscadine grapes by HPLC-ESI-MS. J Agric Food Chem 2005;53:6003–10.

[30] Lei F, Zhang XN, Wang W, Xing DM, Xie WD, Su H, et al. Evidence of anti-obesityeffects of the pomegranate leaf extract in high-fat diet induced obese mice. Int JObes (Lond) 2007;31:1023–9.

[31] Lucas EA, Li W, Peterson SK, Brown A, Kuvibidila S, Perkins-Veazie P, et al. Mangomodulates body fat and plasma glucose and lipids in mice fed a high-fat diet. Br JNutr 2011;106:1495–505.

[32] Makino-Wakagi Y, Yoshimura Y, Uzawa Y, Zaima N, Moriyama T, Kawamura Y.Ellagic acid in pomegranate suppresses resistin secretion by a novel regulatorymechanism involving the degradation of intracellular resistin protein inadipocytes. Biochem Biophys Res Commun 2012;417:880–5.

[33] Panchal SK, Ward L, Brown L. Ellagic acid attenuates high-carbohydrate, high-fatdiet-induced metabolic syndrome in rats. Eur J Nutr 2013;52:559–68.

[34] Neyrinck AM, Van Hee VF, Bindels LB, De Backer F, Cani PD, Delzenne NM.Polyphenol-rich extract of pomegranate peel alleviates tissue inflammation andhypercholesterolaemia in high-fat diet-induced obese mice: potential implicationof the gut microbiota. Br J Nutr 2012;7:1–8.

[35] Losso JN, Bansode RR, Trappey A, Bawadi HA, Truax R. In vitro anti-proliferativeactivities of ellagic acid. J Nutr Biochem 2004;15:672–8.

[36] Rani UP, Kesavan R, Ganugula R, Avaneesh T, Kumar UP, Reddy GB, et al. Ellagicacid inhibits PDGF-BB-induced vascular smooth muscle cell proliferation andprevents atheroma formation in streptozotocin-induced diabetic rats. J NutrBiochem 2013;24:1830–9.

[37] Park SH, Kim JL, Lee ES, Han SY, Gong JH, Kang MK, et al. Dietary ellagic acidattenuates oxidized LDL uptake and stimulates cholesterol efflux in murinemacrophages. J Nutr 2011;141:1931–7.

[38] Selvi BR, Batta K, Kishore AH, Mantelingu K, Varier RA, Balasubramanyam K, et al.Identification of a novel inhibitor of coactivator-associated arginine methyltrans-ferase 1 (CARM1)-mediated methylation of histone H3 Arg-17. J Biol Chem 2010;285:7143–52.

[39] Yadav N, Cheng D, Richard S, Morel M, Iyer VR, Aldaz CM, et al. CARM1 promotesadipocyte differentiation by coactivating PPARgamma. EMBO Rep 2008;9:193–8.

[40] Wang L, Li L, Ran X, Long M, Zhang M, Tao Y, et al. Ellagic acid reducesadipogenesis through inhibition of differentiation-prevention of the induction ofrb phosphorylation in 3T3-L1 adipocytes. Evid Based Complement Alternat Med2013;2013:287534–45.

[41] Yoshimura Y, Nishii S, Zaima N, Moriyama T, Kawamura Y. Ellagic acid improveshepatic steatosis and serum lipid composition through reduction of serum resistin

90 M. Okla et al. / Journal of Nutritional Biochemistry 26 (2015) 82–90

levels and transcriptional activation of hepatic ppara in obese, diabetic KK-A(y)mice. Biochem Biophys Res Commun 2013;434:486–91.

[42] Sarikaya SB, Gulcin I, Supuran CT. Carbonic anhydrase inhibitors: inhibition ofhuman erythrocyte isozymes I and II with a series of phenolic acids. Chem BiolDrug Des 2010;75:515–20.

[43] Lynch CJ, Fox H, Hazen SA, Stanley BA, Dodgson S, Lanoue KF. Role of hepaticcarbonic anhydrase in de novo lipogenesis. Biochem J 1995;310(Pt 1):197–202.

[44] Seeram NP, Lee R, Heber D. Bioavailability of ellagic acid in human plasma afterconsumption of ellagitannins from pomegranate (Punica granatum L.) juice. ClinChim Acta 2004;348:63–8.

[45] Garcia-Villalba R, Beltran D, Espin JC, Selma MV, Tomas-Barberan FA. Time courseproduction of urolithins from ellagic acid by human gut microbiota. J Agric FoodChem 2013;61:8797–806.

[46] Karlsson S,NanbergE, Fjaeraa C,Wijkander J. Ellagic acid inhibits lipopolysaccharide-induced expression of enzymes involved in the synthesis of prostaglandin E2 inhuman monocytes. Br J Nutr 2010;103:1102–9.

[47] Chang WC, Yu YM, Chiang SY, Tseng CY. Ellagic acid suppresses oxidised low-density lipoprotein-induced aortic smooth muscle cell proliferation: studies onthe activation of extracellular signal-regulated kinase 1/2 and proliferating cellnuclear antigen expression. Br J Nutr 2008;99:709–14.

[48] Murugan V, Mukherjee K, Maiti K, Mukherjee PK. Enhanced oral bioavailabilityand antioxidant profile of ellagic acid by phospholipids. J Agric Food Chem 2009;57:4559–65.

[49] Mertens-Talcott SU, Talcott ST, Percival SS. Low concentrations of quercetin andellagic acid synergistically influence proliferation, cytotoxicity and apoptosis inMOLT-4 human leukemia cells. J Nutr 2003;133:2669–74.