Resveratrol as Add-on Therapyin Subjects With Well-ControlledType 2 Diabetes: A RandomizedControlled TrialDiabetes Care 2016;39:2211–2217 | DOI: 10.2337/dc16-0499

OBJECTIVE

To determine whether resveratrol supplementation can improve insulin sensitiv-ity and promote overall metabolic health on top of standard diabetes care.

RESEARCH DESIGN AND METHODS

Seventeen subjects with well-controlled type 2 diabetes (T2D) were treated withplacebo and 150mg/day resveratrol (resVida) in a randomized double-blind cross-over study for 30 days. The main outcome measure was insulin sensitivity by thehyperinsulinemic-euglycemic clamp technique.

RESULTS

Hepatic and peripheral insulin sensitivity were not affected by resveratrol treat-ment. Intrahepatic lipid content also remained unaffected by resveratrol; how-ever, the change in intrahepatic lipid content correlated negatively with plasmaresveratrol levels (R =20.68, P = 0.03). Intramyocellular lipid content increased intype 2 muscle fibers (P = 0.03), and systolic blood pressure tended to decrease(P = 0.09) upon resveratrol treatment. In addition, resveratrol significantly im-proved ex vivo mitochondrial function (state 3 and state U respiration uponmalate with octanoyl-carnitine, P < 0.005). Intriguingly, a correlation was foundbetween plasma levels of a metabolite of resveratrol (dihydroresveratrol) and themetformin dose used by the patients (R = 0.66, P = 0.005), suggesting an interac-tion between metformin and resveratrol. It could be speculated that the lack of aresveratrol-induced insulin-sensitizing effect is caused by this interaction.

CONCLUSIONS

Resveratrol supplementation does not improve hepatic or peripheral insulinsensitivity. Our results question the generalized value of resveratrol as an add-ontherapy in the treatment of T2D and emphasize the need to perform studies indrug-naive patients with T2D or subjects with prediabetes.

Exercise and calorie restriction (1) are the primary treatment options for type 2diabetes (T2D). Both target the cellular energy-sensing route, with activation ofAMPK and the sirtuin (SIRT) family of transcription factors, resulting in stimulationof mitochondrial biogenesis and function (2). Nutraceutical compounds, such asresveratrol, can also target these pathways (3). Interest in resveratrol peakedwhen Howitz et al. (4) identified this polyphenolic compound as a potent SIRT1activator. Since then, resveratrol has been postulated to alleviate metabolic

1Department of Human Biology and HumanMovement Sciences, NUTRIM School of Nutritionand Translational Research in Metabolism,Maastricht University, Maastricht, the Nether-lands2DSM Nutritional Products Ltd., Kaiseraugst,Switzerland3Department of Radiology, NUTRIM School of Nu-trition and Translational Research in Metabolism,Maastricht UniversityMedical Center,Maastricht,the Netherlands

Corresponding author: Patrick Schrauwen,p.schrauwen@maastrichtuniversity.nl.

Received 7 March 2016 and accepted 18 Sep-tember 2016.

Clinical trial reg. no. NCT01638780, clinicaltrials.gov.

This article contains Supplementary Data onlineat http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc16-0499/-/DC1.

S.T. and M.d.L. contributed equally to this work.

© 2016 by the American Diabetes Association.Readers may use this article as long as the workis properly cited, the use is educational and notforprofit, and thework isnotaltered.More infor-mation is available at http://www.diabetesjournals.org/content/license.

Silvie Timmers,1 Marlies de Ligt,1

Esther Phielix,1 Tineke van de Weijer,1

Jan Hansen,1 Esther Moonen-Kornips,1

Gert Schaart,1 Iris Kunz,2

Matthijs K.C. Hesselink,1

Vera B. Schrauwen-Hinderling,1,3 and

Patrick Schrauwen1

Diabetes Care Volume 39, December 2016 2211

CLIN

CARE/ED

UCATIO

N/N

UTR

ITION/PSYC

HOSO

CIAL

consequences of consumption ofenergy-dense foods and physical inactiv-ity, including T2D (5,6). Indeed, animalstudies suggest that resveratrolmaybluntmetabolic complications induced by ahigh-fat diet (3,7–9). These encouragingfindings have stimulated the applicationof resveratrol in clinical trials to investi-gate its potency in humans with chronicmetabolic diseases. Some (10,11), but notall, studies (12–14) in obese humans havereported positive effects of resveratrol onmeasures of insulin sensitivity. In contrastto studies in obese humans, studies insubjects with T2D have been more con-sistent in reporting a beneficial effect ofresveratrol on blood glucose levels(15–17), insulin levels (16,17), markersof insulin resistance, such as the HOMAinsulin-resistance index (15,17), andHbA1c (17,18). None of these studies,however, examined the effect of resvera-trol on peripheral and hepatic insulinsensitivity by the gold standard hyper-insulinemic-euglycemic clamp technique.Hence, we performed a randomized,double-blinded, placebo-controlled cross-over study to examine if 30 days of resver-atrol (resVida) supplementation leads toan improvement in peripheral and he-patic insulin sensitivity in subjectswith well-controlled T2D.

RESEARCH DESIGN AND METHODS

The study protocol was approved by theMedical Review Ethics Committee ofMaastricht University and Medical Cen-tre. All study participants were informedabout the nature and risks of the exper-imental procedures before their writteninformed consent was obtained.

ParticipantsSeventeenmen (age 40–70 years, BMI 27–35 kg/m2, body fat percentage .25%)with well-controlled T2D (HbA1c , 8.0%[,64mmol/mL]) participated in the study.Sixteen participants were treated withthe oral glucose-loweringmedicationmet-formin, six of whom were treated in com-bination with sulfonylurea derivatives(SUDs) (Supplementary Table 1). Most ofthe participants received additionalmedications to lower cholesterol(n = 11) and/or blood pressure(n = 12). Exclusion criteria were unstablebody weight (weight gain or loss.3 kg inthe previous 3 months), engagement inprogrammed exercise .2 h per week,impaired renal and/or kidney function,

intake of dietary supplements (except vi-tamins and minerals), alcohol consump-tion .20 g/day, diabetes comorbidities,and insulin treatment.

OutcomesThe primary outcome measure was theeffect of resveratrol treatment on insu-lin sensitivity compared with placebo.Secondary outcome measures were in-trahepatic lipid content (IHL), intramyo-cellular lipids (IMCL), mitochondrialfunction (in vivo and ex vivo), bloodpressure, and cardiac function.

Clinical Study DesignThe study was conducted at MaastrichtUniversity Medical Center, the Nether-lands, between June 2012 and June2014. In randomized order, participantsunderwent two experimental trials: aplacebo and a resVida (150 mg/daytrans-resveratrol [99.9%]; provided byDSM Nutritional Products Ltd.) condi-tion, with a washout period of at least30 days. The resveratrol dosage wasbased on our previous study in healthyobese subjects in which we found acti-vation of the energy-sensing pathwayAMPK-SIRT1-peroxisome proliferator–activated receptor g coactivator 1-a(PGC1-a) in combination with meta-bolic improvements (10). Randomizationwas performed according to standardprocedures as described in StatisticalMethods by Snedecor and Cochran (19).

Participants were instructed to takethe first supplement on the day of thebaseline measurements (day 0) the lastsupplement in the evening of day 29,and to abstain from food and beveragescontaining substantial amounts of resvera-trol (e.g., wine, grapes, peanuts, and ber-ries) and from food supplements. Inaddition, instructions tomaintain their nor-mal living, activity, and sleeping patternswere given. At the start (day 0) and end(day 30) of both interventionperiods, bloodsamples were analyzed for general safetyparameters (creatinine, aspartate amino-transferase, alanine aminotransferase,g-glutamyl transferase, bilirubin), bloodpressure was measured after an overnightfast, as previously described (10), and a12-lead electrocardiogram was performed.Body fat percentage was determined byDXAonday0of thefirst interventionperiodfor patient characterization.

Each experimental trial lasted 30 days,and the participants came to the university

weekly (days 0, 7, 14, 21, and 29). Theweekly checkup took place in the morningin overnight fasted state and included bodyweight measurement and drawing of ablood sample for analysis of resveratrol (pa-rental and metabolites) to confirm compli-ance to the protocol. On day 29 at 4 P.M.,subjects came to the university for cardiacfunction measurements by M-mode, two-dimensional and Doppler echocardiogra-phy, using a Vivid 7 ultrasound system (GEHealthcare, Milwaukee, WI) with 3.5-MHzcardiac transducer. Results were inter-preted according to the criteria of theAmerican Societies of Echocardiography.Subsequently, by proton magnetic spec-troscopy (1H-MRS) IHL content was quanti-fied, as described earlier (20), on a 3Twhole-body scanner (Achieva Tx; PhilipsHealthcare, Best, the Netherlands) usingan echo time of 32.5 ms. Spectra were fit-tedwith a home-written script (21) inMAT-LAB R2014b (MathWorks, Natick, MA).Values are given as T2-corrected ratios(22) of the CH2 peak, relative to the unsup-pressed water resonance (as percentage).Postexercise phosphocreatine (PCr) recov-ery was assessed by 31P-MRS to estimatein vivo mitochondrial function in vastus lat-eralis muscle on a 3T whole-body scanner,described elsewhere (21). To standardizefood intake before these measure-ments, subjects consumed a standard-ized lunch at a fixed time and afterwardstayed fasted until the measurementswere completed.

After the MRS measurements, partic-ipants entered the respiration chamberfor 12 h starting at 7 P.M. to allow mea-surement of sleeping metabolic rate(23). Before the respiration chambermeasurement started, a standardizedevening meal (pasta Bolognese and fruityogurt) was provided. The energy pro-vided was based on individual daily en-ergy requirements, calculated with theHarris and Benedict equation. In themorning of day 30, participants leftthe respiration chamber at 7 A.M. in theovernight fasted state and a biopsyspecimen was taken from the vastus lat-eralis muscle under local anesthesia (2%lidocaine), as previously described (24). Aportion of the muscle tissue was directlyfrozen inmelting isopentane for immuno-histochemistry and stored at280°C untilassayed. Another portion (;30 mg) wasimmediately placed in ice-cold preserva-tion medium for determination of ex vivomitochondrial respiration. From the

2212 Resveratrol and T2D Diabetes Care Volume 39, December 2016

muscle tissue in the preservation me-dium, permeabilized muscle fiberswere immediately prepared (24). Thepermeabilized muscle fibers (;2.5 gwet weight) were analyzed for mito-chondrial function using an oxygraph(OROBOROS Instruments, Innsbruck,Austria) (10). Fresh cryosections (5 mm)were stained for IMCL by Oil Red O stain-ing combined with fiber typing and im-munolabeling of the basal membranemarker laminin to allow quantificationof IMCL (25). Mitochondrial DNA copynumber and protein expression of oxida-tive phosphorylation (OXPHOS) andPGC1-a by Western blot were per-formed according to standard proce-dures as described previously (10).After the muscle biopsy was taken, a

two-step hyperinsulinemic-euglycemicclamp was performed to assess periph-eral and hepatic insulin sensitivity.Three days before the clamp, subjectswere instructed to refrain from strenu-ous activities and to continue their anti-diabetic medication, with the last doseon the morning of the test. The clampstarted by giving the subjects a primedcontinuous infusion of D-[6,6-2H2]glucose(0.04 mg/kg/min) to determine ratesof endogenous glucose production(EGP), glucose appearance (Ra), and glu-cose disposal (Rd) (26). After 120 min,a low insulin infusion was started(10 mU/m2/min) for 3 h, after whicha high insulin infusion was started(40 mU/m2/min) for 2 h. During thelast 30 min of each insulin infusion step(0, 10, and 40 mU/m2/min), blood sam-ples were collected, and substrateutilization was measured by indirect cal-orimetry. Steele’s single pool nonsteady-state equations were used to calculateglucose Ra and Rd (27). Volume of distri-bution was assumed to be 0.160 L/kg forglucose.

ComplianceTo check compliance, resveratrol me-tabolites were measured by mass spec-troscopy in plasma on days 0, 7, 14, 21,and 30 (10). In addition, unused cap-sules were counted.

Sample SizeThe sample size was determined basedon demonstrating the statistical superior-ity of resveratrol on insulin-stimulatedglucose disposal in muscle comparedwith placebo treatment. We estimated

14 subjects were required to achieve80% power, with an assumed treatmentdifference of 1.4mg/kg fat-freemass/minafter 30 days and an assumed SD of1.7 mg/kg fat-free mass/min for ahyperinsulinemic clamp. A dropout of20% was taken into account, so 17 sub-jects were recruited. The expectedeffect size and SD of a hyperinsulinemic-euglycemic clamp was based on a previ-ous study in our research group witha pre- and postintervention clamp in sub-jects with T2D (28).

Statistical AnalysisResults are presented as means 6 SEMwhen normally distributed and as me-dian (95% CI) when this was not thecase (using Shapiro-Wilk normalitytest). The Student paired t test wasused to compare placebo and resvera-trol supplementation in normally dis-tributed data; otherwise, the Wilcoxonsigned-rank test was used. Treatmentcomparisons of plasma parametersmeasured at the beginning and end in-tervention were assessed by two-wayrepeated-measures ANOVA. Linear re-gression analyses were conducted toidentify correlations between vari-ables. On normally distributed dataPearson correlation was used, other-wise Spearman correlation was used.Potential carryover effect betweentreatment and period was examinedby unpaired t test analyses accordingto Pocock (29). A P value of ,0.05was considered statistically significant.Analyses were performed using SPSS22.0 software.

RESULTS

Study Design and PlasmaBiochemistryThe study participants were 17 malesubjects with well-controlled T2D (base-line characteristics, Supplementary Table1). To check compliance with the studyprotocol and to confirm systemic conver-sion of resveratrol to dihydroresveratrol(DHR), parental resveratrol and DHR(aglycone + glucuronide conjugates) lev-els were determined weekly in theplasma. Although no resveratrol or DHRcould be detected during the placebotreatment, both compounds were pre-sent in the plasma of all resveratrol-supplemented subjects. Plasma levelswere 378.596 40.65 ng/mL for parentalresveratrol and 444.69 6 101.81 ng/mL

for DHR at day 30 of the resveratrol treat-ment. These levels are representativefor the entire study period. In addition,no carryover effect was found betweentreatment and period for the primarystudy outcome, confirming the washoutperiod was sufficient.

Insulin Sensitivity and SubstrateKinetics Assessed byHyperinsulinemic-Euglycemic ClampPlasma glucose profiles during thehyperinsulinemic-euglycemic clampprocedure were similar for the placeboand the resveratrol conditions (Fig. 1A).Both during the low- and high-insulinstate, glucose levels were clamped;5 mmol/L (Fig. 1A). Insulin-stimulatedglucose disposal, expressed as thechange in Rd from basal to the low- orhigh-insulin state, was not significantlychanged by resveratrol supplementa-tion (Pbasal = 0.65, Plow insulin = 0.89,and Phigh insulin = 0.66) (Fig. 1B). BasalEGP, reflecting hepatic glucose output,and EGP under low insulin–stimulatedconditions were also similar betweenboth conditions, suggesting that hepaticinsulin sensitivity was unchanged(Pbasal = 0.22) (Fig. 1C). In fact, EGP wassimilarly suppressed during the low-insulin state of the hyperinsulinemic-euglycemic clamp (Plow insulin = 0.84,Phigh insulin = 0.74) (Fig. 1C), and EGP sup-pression was nearly complete uponhigh-insulin infusion in both condi-tions (Plow insulin = 0.64, Phigh insulin =0.73) (Fig. 1D). In line with the absenceof resveratrol-induced changes in Rdand EGP, nonoxidative glucose disposalalso remained unaffected by 30 days ofresveratrol supplementation (Supple-mentary Table 2). In addition, resveratroldid not exert an effect on circulating in-sulin levels (Supplementary Table 2),thereby excluding an effect of resveratrolon insulin clearance.

As expected, carbohydrate oxidationincreased during the clamp at the ex-pense of free fatty acid (FFA) oxidation(Supplementary Table 2). No statisticallysignificant changes in carbohydrate orFFA oxidation rates were observedwhen basal and insulin-stimulated oxi-dation rates were compared betweenplacebo and resveratrol (SupplementaryTable 2). At theendof thehyperinsulinemic-euglycemic clamp, suppression of FFAwas similar between both conditions(P = 0.36).

care.diabetesjournals.org Timmers and Associates 2213

In accordance with the lack of effectof resveratrol on insulin sensitivity, noimprovement in fasting glucose, insulin,or HbA1c levels were found comparedwithplacebo (Supplementary Table 3). We didobserve a time effect for HbA1c; however,this was found in both treatment condi-tions. Similarly, no resveratrol effect onother markers of metabolic health wasfound (Supplementary Table 3).

Sleeping MetabolismSleeping metabolic rate, measured in awhole-body respiratory chamber duringthe last night of the intervention, de-creased upon resveratrol supplementa-tion in 82% of subjects. However, thedifference failed to reach statistical sig-nificance (7.81 6 0.20 MJ/day for pla-cebo vs. 7.55 6 0.16 MJ/day forresveratrol, P = 0.14). During the night,substrate utilization (reflected by the re-spiratory quotient) was similar for both

treatments, at 0.81 (95% CI 0.81–0.88)for placebo vs. 0.82 (95% CI 0.80–0.85)for resveratrol (P = 0.45).

Ectopic Lipid StorageLipid area fraction, measured ex vivo inmuscle biopsy specimens by Oil Red Ostaining, increased significantly in type 2muscle fibers (P = 0.03) (Fig. 2A), and asimilar tendency was observed for totallipid content (P = 0.08) (Fig. 2A).

IHL content, determined by 1H-MRS,remained unaffected by 30 days of re-sveratrol supplementation (Fig. 2B). In-terestingly, linear regression analysisshowed that the difference in IHL con-tent between resveratrol and placebonegatively correlated with the parentalplasma resveratrol concentration(R =20.68, P = 0.03) (Fig. 2C), indicatingthat a decrease in IHLmay depend on theplasma resveratrol concentrationsachieved.

Mitochondrial FunctionMitochondrial state 3 respiration on alipid-derived substrate (malate +octanoyl-carnitine) and upon parallelelectron input to both complex I and II(malate + octanoyl-carnitine + glutamate +succinate) (Fig. 3A), and maximal FCCP-induced uncoupled respiration (P =0.001) (Fig. 3B) were significantly higherafter resveratrol supplementation. How-ever, in the absence of a lipid-derivedsubstrate, no resveratrol effect was ob-served in state 3 respiration upon complexI- and II-linked substrates (SupplementaryFig. 1A). In addition, state 4o respirationupon addition of oligomycin (reflectingmi-tochondrial proton leak) was similar be-tween the resveratrol and the placebocondition (Supplementary Fig. 1B). Mito-chondrial DNA copy number (2,251 6162 arbitrary units vs. 2,234 6 152 arbi-trary units, respectively, P = 0.91), meanprotein content of the OXPHOS com-plexes, PGC1-a protein content, andphosphorylated AMPK/AMPK ratio re-mained unaffected by resveratrol supple-mentation (Supplementary Fig. 1C–E).Mean PCr recovery half-time, a measureof in vivo mitochondrial function, wasunchanged by resveratrol comparedwith placebo (Supplementary Fig. 1F). Inline, VO2max, a measure of cardiorespi-ratory fitness, was not affected by re-sveratrol supplementation (23.88 61.10 mL z kg21 z min21 for placebo vs.23.40 6 1.34 mL z kg21 z min21 for re-sveratrol, P = 0.29).

Cardiac FunctionSystolic blood pressure, measured inovernight fasted state on day 30,tended to decrease by resveratrolsupplementation (P = 0.09), whereas di-astolic blood pressure remained un-changed (Supplementary Table 4).Echocardiography revealed a marginalreduction in left ventricular end systolicdiameter upon resveratrol supplemen-tation (P = 0.04) (Supplementary Table4). Because the values remain wellwithin the normal range, they are notinterpreted as clinically significant. Nochange in stroke volume, cardiac out-put, or left ventricular ejection fractionwas found. Parameters of diastolic func-tion of the heart remained unaffectedby resveratrol. Furthermore, no struc-tural changes were observed in theheart by resveratrol (SupplementaryTable 4).

Figure 1—Effect of resveratrol on peripheral and hepatic insulin sensitivity. After 30 days ofresveratrol and placebo, peripheral and hepatic insulin sensitivity was assessed with a two-stephyperinsulinemic-euglycemic clamp (t = 0–120 min: D-[6,6-2H2]glucose tracer infusion; t = 120–300 min: low-insulin infusion; t = 300–420 min: high-insulin infusion). As a result of technicalproblems, data are only available for 14 subjects (in 1 subject the equipment malfunctioned, inanother subject the catheter with the D-[6,6-2H2]glucose tracer leaked, and in the last subjectaspiration of the venous catheter was no longer possible in the late phase of the clamp,which is necessary for regular blood sampling). A: Plasma glucose levels during the last30 min of the basal and low- and high-insulin state of the clamp. Glucose levels were clamped;5 mmol/L. Insulin-stimulated glucose disposal, expressed as the Rd (B), and EGP werecalculated for the last 30 min of the basal and low- and high-insulin state (C ). D: EGP sup-pression upon low- and high-insulin infusion. Data are presented as mean6 SEM. Rd, rate ofdisappearance.

2214 Resveratrol and T2D Diabetes Care Volume 39, December 2016

Plasma Resveratrol Levels andMetformin DoseA strong correlation was found betweenplasma DHR concentration andmetformindose after resveratrol treatment at day30 (R = 0.66, P = 0.005) (Fig. 4A) suggestingan interaction between metformin andthe metabolism of resveratrol. Post hocanalysis revealed that total plasma DHR

levels were higher during 30 days of re-sveratrol supplementation in subjectstaking a relatively high daily dose of met-formin (.1,000 mg/day [n = 9]; averagemetformin dose, 2,188mg/day) comparedwith patients treated with a lower dose(#1,000 mg/day [n = 8]; average metfor-min dose, 639 mg/day) (Fig. 4B), whereasthe daily dose of metformin did not seem

to influence plasma levels of parental re-sveratrol (Fig. 4C).

Adverse EventsResveratrol was well tolerated by theparticipants, and no adverse events oc-curred. Measurement of parameters forkidney and liver function (creatinine, as-partate aminotransferase, alanine ami-notransferase, g-glutamyl transferaseand bilirubin) indicated that resveratrolwas well tolerated (SupplementaryTable 3). Creatinine showed a signifi-cant, albeit very modest, decline fromday 0 to day 30 in both treatmentconditions, but no treatment effectover time was found (SupplementaryTable 3).

CONCLUSIONS

Preclinical research has suggested thatresveratrol could be used to prevent themetabolic consequences of high-fatfeeding, including insulin resistance, byimproving mitochondrial function. Wepreviously observed in obese, normo-glycemic subjects that resveratrolsupplementation exerted beneficialmetabolic effects accompanied by animprovement in mitochondrial function.The current study did not find a stimula-tory effect of resveratrol on peripheralor hepatic insulin sensitivity in subjectswith well-controlled T2D on top of stan-dard diabetes care. Other markers ofmetabolic health also largely remainedunaffected by resveratrol, althoughimprovements in ex vivo muscle oxida-tive capacity were detected. Overall, ourresults suggest that resveratrol may notbe of added value as adjunct therapy forsubjects with T2D receiving the oral glu-cose-lowering medication metformin.

The metabolic actions of resveratrolare mainly ascribed to activation of theenergy-sensing pathway AMPK-SIRT1-PGC1-a (1). Several downstream targetsof this pathway (e.g., mitochondrial bio-genesis and glucose and lipid homeosta-sis) are interesting targets for treatmentof T2D. No positive effect of 30 days ofresveratrol supplementation on hepaticEGP, (non)oxidative skeletal muscle glu-cose disposal, circulating insulin levels,and substrate utilization was detected.Also, fasting glucose, insulin, and HbA1cvalues remained unaffected by resvera-trol. It should be noted that we usedan add-on approach, in which resvera-trol was supplemented on top of the

Figure 3—Effect of resveratrol (RSV) on ex vivo mitochondrial function. After 30 days of resver-atrol and placebo, a muscle biopsy specimen was obtained from the vastus lateralis muscle. Partof the specimen was used for evaluation of ex vivo mitochondrial function (n = 17). A: ADP-stimulated respiration (state 3) upon a lipid-like substrate and upon parallel electron input intocomplex I and II. B: Maximally uncoupled respiration upon FCCP. Data are presented as means6SEM. Box plot represents minimum, first quartile, median, third quartile, and maximum. *P ,0.05. G, glutamate; M, malate; O, octanoyl-carnitine; S, succinate.

Figure 2—Effect of resveratrol (RSV) on ectopic lipid storage. A: Muscle biopsy sections from arepresentative study participant, stained for IMCL with Oil Red O staining (in red), musclelaminin (in blue), and type 1 muscle fibers (in green). IMCL content is quantified as the percent-age area of a muscle fiber that is covered by lipids (n = 17). B: IHL content quantified by 1H-MRSafter 29 days of resveratrol and placebo supplementation (n = 10). C: ΔIHL in relation to totalplasma resveratrol concentration (n = 10). Data are presented as means 6 SEM. Box plotrepresents minimum, first quartile, median, third quartile, and maximum. *P , 0.05.

care.diabetesjournals.org Timmers and Associates 2215

patients’ normal blood glucose–lower-ing treatment, mainly metformin. It iswell known that treatment with theoral glucose-lowering medication met-formin leads to phosphorylation of thethreonine residue 172 in the a-subunitof AMPK (30). Accordingly, measure-ment of phosphorylated AMPK byWestern blot revealed no significant dif-ferences between the resveratrol andplacebo condition. Potentially, a dailydose of 150 mg of resveratrol may nothave been able to further activate AMPKin subjects with T2D receiving metfor-min, thereby explaining the lack ofeffect of resveratrol on insulin sensitiv-ity. In addition to this explanation, ourdata indicate that metformin may inter-act with the metabolism of resveratrol.This potential interaction could haveinfluenced the efficacy of the active com-ponent, thereby accounting for the lackof a resveratrol-induced improvementof insulin sensitivity.Interestingly, Movahed et al. (17) re-

cently reported that subjects with T2Dtreated with metformin only did notshow an improvement in circulating in-sulin or HbA1c levels upon resveratroladministration, whereas patients re-ceiving SUDs only or metformin in com-bination with SUDs demonstrated aresveratrol-induced improvement inboth plasma makers of insulin sensitiv-ity. Together with our study results,these findings reinforce the importanceof investigating the relationship be-tween resveratrol and metformin in fu-ture research.A potential mechanism that may partly

explain the interaction of metformin with

resveratrol can be sought in the recentlyappreciated influence of metformin ongut microbiota composition (31,32). Thegutmicrobiota is one of themajor sites ofmetabolism of trans-resveratrol to themetabolite DHR (33). Although beyondthe scope of this research, alterations ingut microbiota composition by metfor-min could have affected the metabolismof our trans-resveratrol. It would there-fore be worthwhile to examine if resver-atrol could improve insulin sensitivity indrug-naive subjects with T2D or subjectswith prediabetes.

Despite the lack of effect of resvera-trol on AMPK activity and insulin sensi-tivity, we did observe an improvementof ex vivo mitochondrial function. Theseresults are in line with Price et al. (34),who recently demonstrated that the ef-fects of resveratrol on mitochondrialmetabolism are not always paralleledby an effect on AMPK. Furthermore,we found that resveratrol increasedIMCL content, consistent with our pre-vious findings in healthy, obese subjects(10). These findings may suggest thatthe beneficial effects of resveratrol oninsulin sensitivity, as found by others,do not depend entirely on the effectson skeletal muscle mitochondrial func-tion. Alternatively, these results maysuggest that the beneficial effects ofresveratrol on muscle mitochondrialfunction may be too small to affect in-sulin sensitivity or that the duration ofthe treatment may have been too shortto affect glycemic control. In that re-spect, recent work of Thazhath et al.(35) in subjects with T2D controlledby diet only also failed to improve

glycemic control when treated with500 mg of resveratrol twice daily for2 weeks.

IHL content also remained unaffectedby resveratrol in subjects with T2D,which was in contrast to our previousstudy in which we noted a significantdecrease in healthy, obese subjects(10). However, we did observe a corre-lation between plasma parental resver-atrol levels and changes in IHL content,suggesting that higher doses of resvera-trol may be able to reduce IHL contentalso in subjects with T2D. Furtherclinical studies are needed to test thishypothesis.

Resveratrol supplementation did notaffect systolic and diastolic function ofthe heart but tended to decrease sys-tolic blood pressure. A decrease in sys-tolic blood pressure has previously beenobserved with resveratrol. For example,a recent meta-analysis of six random-ized controlled trials investigated theeffect of resveratrol on blood pressureand reported that high doses of resver-atrol ($150 mg/day) significantly re-duced systolic blood pressure but thatlower doses had no effect on bloodpressure (36).

In conclusion, 30 days of resveratrolsupplementation did not improve he-patic or peripheral insulin sensitivity insubjects with T2D treated with oral glu-cose-lowering medication. Intriguingly,levels of a metabolite of resveratrolwere affected by the dose of metforminused, suggesting that the lack of effectof resveratrol on insulin sensitivity mayhave been affected by metformin use inthese patients. These results question

Figure 4—Possible interaction of resveratrol (RSV) withmetformin. Post hoc analysis revealed a possible interaction ofmetforminwith themetabolism ofresveratrol. A: Correlation of the daily metformin dose used by the patients with T2D and plasma DHR levels. A subsequent subanalysis was performed inwhich patients with T2Dwere separated in a group using high daily doses of metformin (.1,000mg/day [n = 9]; average metformin dose, 2,188mg/day)vs. a group using low daily doses of metformin (#1,000 mg/day [n = 8]; average metformin dose, 639 mg/day). Total plasma DHR levels (B) and totalresveratrol levels (C) for the high- vs. low-dose metformin groups. Data are presented as mean6 SEM.

2216 Resveratrol and T2D Diabetes Care Volume 39, December 2016

the generalized value of resveratrol asan add-on therapy for treatment ofT2D and emphasize the need to explorethe possible interaction between re-sveratrol and metformin. Studies in sub-jects with prediabetes are needed toexamine whether resveratrol can im-prove insulin sensitivity when not com-bined with oral glucose-lowering drugtherapy.

Acknowledgments. The authors thank DSMNutritional Products Ltd. for providing the res-Vida and placebo capsules and for performingthe resveratrol and DHR analysis.Funding. This study was funded by a EuropeanFoundation for the Study of Diabetes ClinicalResearch Grant.Duality of Interest. I.K. is employed at DSMNutritional Products Ltd., Kaiseraugst, Switzerland.No other potential conflicts of interest relevant tothis article were reported.Author Contributions. S.T. and M.d.L. de-signed and performed the experiments, ana-lyzed the data, and wrote the manuscript. E.P.,T.v.d.W., J.H., E.M.-K., and G.S. assisted duringthe experiments and reviewed and edited themanuscript. I.K. delivered the resVida and pla-cebo capsules and performed the analysis ofplasma resveratrol and DHR levels. M.K.C.H.,V.B.S.-H., and P.S. contributed to the design ofthe study, analyzed and interpreted the data,and reviewed and edited the manuscript. Allauthors approved the final version of the man-uscript. P.S. is the guarantor of this work and, assuch, had full access to all the data in the studyand takes responsibility for the integrity of thedata and the accuracy of the data analysis.

References1. Canto C, Auwerx J. PGC-1alpha, SIRT1 andAMPK, an energy sensing network that controlsenergy expenditure. Curr Opin Lipidol 2009;20:98–1052. Fernandez-Marcos PJ, Auwerx J. Regulationof PGC-1a, a nodal regulator of mitochondrialbiogenesis. Am J Clin Nutr 2011;93:884S–890S3. Lagouge M, Argmann C, Gerhart-Hines Z,et al. Resveratrol improves mitochondrial func-tion and protects against metabolic disease byactivating SIRT1 and PGC-1alpha. Cell 2006;127:1109–11224. Howitz KT, Bitterman KJ, Cohen HY, et al.Small molecule activators of sirtuins extend Sac-charomyces cerevisiae lifespan. Nature 2003;425:191–1965. Baur JA, Sinclair DA. Therapeutic potential ofresveratrol: the in vivo evidence. Nat Rev DrugDiscov 2006;5:493–5066. Timmers S, Hesselink MK, Schrauwen P.Therapeutic potential of resveratrol in obesityand type 2 diabetes: new avenues for healthbenefits? Ann N Y Acad Sci 2013;1290:83–89

7. Baur JA, Pearson KJ, Price NL, et al. Resvera-trol improves health and survival of mice on ahigh-calorie diet. Nature 2006;444:337–3428. Burgess TA, Robich MP, Chu LM, Bianchi C,Sellke FW. Improving glucose metabolism withresveratrol in a swine model of metabolic syn-drome through alteration of signaling pathwaysin the liver and skeletal muscle. Arch Surg 2011;146:556–5649. Marchal J, Blanc S, Epelbaum J, Aujard F,Pifferi F. Effects of chronic calorie restrictionor dietary resveratrol supplementation on insu-lin sensitivity markers in a primate, Microcebusmurinus. PLoS One 2012;7:e3428910. Timmers S, Konings E, Bilet L, et al. Calorierestriction-like effects of 30 days of resveratrolsupplementation on energy metabolism andmetabolic profile in obese humans. Cell Metab2011;14:612–62211. Crandall JP, Oram V, Trandafirescu G, et al.Pilot study of resveratrol in older adults withimpaired glucose tolerance. J Gerontol A BiolSci Med Sci 2012;67:1307–131212. Chachay VS, Macdonald GA, Martin JH,et al. Resveratrol does not benefit patientswith nonalcoholic fatty liver disease. Clin Gas-troenterol Hepatol 2014;12:2092–2103.e1, 613. Dash S, Xiao C, Morgantini C, Szeto L, LewisGF. High-dose resveratrol treatment for 2weeksinhibits intestinal and hepatic lipoprotein pro-duction in overweight/obese men. ArteriosclerThromb Vasc Biol 2013;33:2895–290114. Poulsen MM, Vestergaard PF, Clasen BF,et al. High-dose resveratrol supplementationin obese men: an investigator-initiated, ran-domized, placebo-controlled clinical trial ofsubstrate metabolism, insulin sensitivity, andbody composition. Diabetes 2013;62:1186–119515. Brasnyo P, Molnar GA, Mohas M, et al. Re-sveratrol improves insulin sensitivity, reducesoxidative stress and activates the Akt pathwayin type 2 diabetic patients. Br J Nutr 2011;106:383–38916. Elliott PJ, Walpole S, Morelli L, et al. Resver-atrol/SRT-501. Drugs Future 2009;34:291–29517. Movahed A, Nabipour I, Lieben Louis X,et al. Antihyperglycemic effects of short termresveratrol supplementation in type 2 diabeticpatients. Evid Based Complement Alternat Med2013;2013:85126718. Bhatt JK, Thomas S, Nanjan MJ. Resveratrolsupplementation improves glycemic control intype 2 diabetes mellitus. Nutr Res 2012;32:537–54119. Snedecor GW, Cochran WG. StatisticalMethods. Ames, Iowa State University Press,198020. Bilet L, Brouwers B, van Ewijk PA, et al.Acute exercise does not decrease liver fat inmen with overweight or NAFLD. Sci Rep 2015;5:970921. Lindeboom L, Nabuurs CI, Hesselink MK,Wildberger JE, Schrauwen P, Schrauwen-Hinderling VB. Proton magnetic resonancespectroscopy reveals increased hepatic lipidcontent after a single high-fat meal with no

additional modulation by added protein. Am JClin Nutr 2015;101:65–7122. Guiu B, Petit JM, Loffroy R, et al. Quantifi-cation of liver fat content: comparison of triple-echo chemical shift gradient-echo imaging andin vivo proton MR spectroscopy. Radiology2009;250:95–10223. Schoffelen PF, Westerterp KR, Saris WH,Ten Hoor F. A dual-respiration chamber systemwith automated calibration. J Appl Physiol(1985) 1997;83:2064–207224. Phielix E, Schrauwen-Hinderling VB,Mensink M, et al. Lower intrinsic ADP-stimulated mitochondrial respiration underliesin vivo mitochondrial dysfunction in muscle ofmale type 2 diabetic patients. Diabetes 2008;57:2943–294925. Koopman R, Schaart G, Hesselink MK. Opti-misation of oil red O staining permits combina-tion with immunofluorescence and automatedquantification of lipids. Histochem Cell Biol2001;116:63–6826. van de Weijer T, Phielix E, Bilet L, et al. Ev-idence for a direct effect of the NAD+ precursoracipimox on muscle mitochondrial function inhumans. Diabetes 2015;64:1193–120127. Steele R. Influences of glucose loading andof injected insulin on hepatic glucose output.Ann N Y Acad Sci 1959;82:420–43028. Meex RC, Schrauwen-Hinderling VB,Moonen-Kornips E, et al. Restoration of musclemitochondrial function and metabolic flexibilityin type 2 diabetes by exercise training isparalleled by increased myocellular fat storageand improved insulin sensitivity. Diabetes 2010;59:572–57929. Pocock S. Clinical Trials: A Practical Ap-proach. Chichester, John Wiley and Sons, 198730. Fryer LG, Parbu-Patel A, Carling D. The anti-diabetic drugs rosiglitazone and metforminstimulate AMP-activated protein kinasethrough distinct signaling pathways. J BiolChem 2002;277:25226–2523231. Maratos-Flier E. Metabolic disease putsup a fight: microbes, metabolism and medica-tions. Nat Med 2013;19:1218–121932. Lee H, Ko G. Effect of metformin on meta-bolic improvement and gut microbiota. Appl En-viron Microbiol 2014;80:5935–594333. Bode LM, Bunzel D, Huch M, et al. In vivoand in vitro metabolism of trans-resveratrol byhuman gut microbiota. Am J Clin Nutr 2013;97:295–30934. Price NL, Gomes AP, Ling AJ, et al. SIRT1 isrequired for AMPK activation and the beneficialeffects of resveratrol onmitochondrial function.Cell Metab 2012;15:675–69035. Thazhath SS, Wu T, BoundMJ, et al. Admin-istration of resveratrol for 5 wk has no effect onglucagon-like peptide 1 secretion, gastric emp-tying, or glycemic control in type 2 diabetes: arandomized controlled trial. Am J Clin Nutr2016;103:66–7036. Liu Y, Ma W, Zhang P, He S, Huang D. Effectof resveratrol on blood pressure: a meta-analysis of randomized controlled trials. ClinNutr 2015;34:27–34

care.diabetesjournals.org Timmers and Associates 2217