A Protein Preload Enhances theGlucose-Lowering Efficacy ofVildagliptin in Type 2 DiabetesDiabetes Care 2016;39:511–517 | DOI: 10.2337/dc15-2298

OBJECTIVE

Nutrient “preloads” given before meals can attenuate postprandial glycemic ex-cursions, at least partly by slowing gastric emptying and stimulating secretion ofthe incretins (i.e., glucagon-like peptide-1 [GLP-1] and glucose-dependent insulino-tropic polypeptide [GIP]). This study was designed to evaluate whether a pro-tein preload could improve the efficacy of the dipeptidyl peptidase-4 (DPP-4)inhibitor vildagliptin to increase incretin concentrations, slow gastric emptying,and lower postprandial glycemia in type 2 diabetes.

RESEARCH DESIGN AND METHODS

Twenty-two patients with type 2 diabetes treated with metformin were studiedon four occasions, receiving either 50 mg vildagliptin (VILD) or placebo (PLBO) onboth the evening before and the morning of each study day. The latter dose wasfollowed after 60min by a preload drink containing either 25 gwhey protein (WHEY)or control flavoring (CTRL), and after another 30 min by a 13C-octanoate–labeledmashed potato meal. Plasma glucose and hormones, and gastric emptying, wereevaluated.

RESULTS

Compared with PLBO/CTRL, PLBO/WHEY reduced postprandial peak glycemia,increased plasma insulin, glucagon, and incretin hormones (total and intact),and slowed gastric emptying, whereas VILD/CTRL reduced both the peak and areaunder the curve for glucose, increased plasma intact incretins, and slowed gastricemptying but suppressed plasma glucagon and total incretins (P < 0.05 each).Compared with both PLBO/WHEY and VILD/CTRL, VILD/WHEY was associatedwith higher plasma intact GLP-1 and GIP, slower gastric emptying, and lowerpostprandial glycemia (P < 0.05 each).

CONCLUSIONS

In metformin-treated type 2 diabetes, a protein preload has the capacity to en-hance the efficacy of vildagliptin to slow gastric emptying, increase plasma intactincretins, and reduce postprandial glycemia.

The “incretin” hormones glucose-dependent insulinotropic polypeptide (GIP)and glucagon-like peptide-1 (GLP-1) are major determinants of postprandial gly-cemia (1). The latter is an important target in patients with type 2 diabetes,particularly those with modestly elevated HbA1c (2). In health, both incretinsstimulate insulin secretion in a glucose-dependent manner (1). However, theeffect of GIP is substantially diminished in type 2 diabetes (3), whereas GLP-1

1Discipline ofMedicine, The University of Adelaide,Adelaide, Australia2Centre of Research Excellence in TranslatingNutritional Science to Good Health, The Universityof Adelaide, Adelaide, Australia3Department of Biomedical Science, University ofCopenhagen, Copenhagen, Denmark

Corresponding author: Tongzhi Wu, tongzhi.wu@adelaide.edu.au.

Received 21 October 2015 and accepted 24December 2015.

Clinical trial reg. no. ACTRN12612001005842,www.anzctr.org.au.

© 2016 by the American Diabetes Association.Readersmayuse this article as longas thework isproperly cited, the use is educational and not forprofit, and the work is not altered.

Tongzhi Wu,1,2 Tanya J. Little,1,2

Michelle J. Bound,1,2 Malcolm Borg,1

Xiang Zhang,1,2 Carolyn F. Deacon,3

Michael Horowitz,1,2 Karen L. Jones,1,2 and

Christopher K. Rayner1,2

Diabetes Care Volume 39, April 2016 511

CLIN

CARE/ED

UCATIO

N/N

UTR

ITION/PSYC

HOSO

CIAL

retains considerable insulinotropic activ-ity (3) and also slows gastric emptying (4)and suppresses glucagon secretion (5)and energy intake (6).Dipeptidyl peptidase-4 (DPP-4) inhib-

itors prevent rapid degradation of en-dogenous incretins, thereby elevatingplasma concentrations of the intactforms. However, the overall secretionof incretin hormones is reduced as a re-sult of negative feedback regulation (1).Given the loss of response to GIP intype 2 diabetes (3), the glycemic effectof DPP-4 inhibitors has been consideredlargely GLP-1 related. Indeed, the GLP-1receptor antagonist exendin(9–39) at-tenuates the glucose-lowering and insu-linotropic effects of the DPP-4 inhibitorsitagliptin by ;50% in patients withtype 2 diabetes and abolishes any slow-ing of gastric emptying (7). Therefore,strategies that stimulate endogenousGLP-1 secretion could potentially en-hance the efficacy of DPP-4 inhibition.Our group has developed a novel

“preload” concept, which primarily tar-gets the postprandial glycemic excur-sion in the management of type 2diabetes. Consumption of a macronutri-ent preload before the main meal canstimulate release of gut peptides, in-cluding GLP-1, slow gastric emptying,and reduce the glycemic response tothe meal (8). We recently reportedthat a preload of xylose (a poorly ab-sorbed pentose) augments the reduc-tion of postprandial glycemia bysitagliptin in type 2 diabetes, associatedwith sustained elevation of plasma in-tact GLP-1 concentrations (9). However,diarrhea and flatulence may limit thetolerability of xylose for long-term use.Whey protein preloads also stimulateGLP-1 and reduce postprandial glycemiain type 2 diabetes and are well tolerated(10,11). However, the doses of wheyused (;50 g) entail a substantial burdenin energy intake and are relatively ex-pensive. A smaller whey preload (25 g)was also reported to slow gastric emp-tying and reduce postprandial peakblood glucose, without causing signifi-cant weight gain with sustained use inpatients with type 2 diabetes (12), butwhether it is sufficient to increaseplasma GLP-1 concentrations and, ac-cordingly, enhance the efficacy of DPP-4inhibitors has not been established.In the current study, we hypothesized

that a protein preload would enhance

the efficacy of glucose lowering byDPP-4 inhibition and evaluated theacute effect of the DPP-4 inhibitorvildagliptin with and without a 25-gwhey preload on postprandial glycemiain patients with type 2 diabetes man-aged with metformin monotherapy, agroup in whom DPP-4 inhibitors arecommonly added (13).

RESEARCH DESIGN AND METHODS

SubjectsTwenty-two male subjects with type 2diabetes, managed bymetforminmono-therapy (500–2,000 mg/day, stable for$3 months), were studied after provid-ing written informed consent. The mean(6SE) age was 64.2 6 1.4 years, BMI27.9 6 1.7 kg/m2, HbA1c 6.6 6 0.2%(48.8 6 2.2 mmol/mol), and durationof known diabetes 5.6 6 1.2 years.None were smokers or tookmedicationsaffecting gastrointestinal function. Theprotocol was approved by the HumanResearch Ethics Committee of RoyalAdelaide Hospital and was conductedin accordance with the principles ofthe Declaration of Helsinki.

ProtocolEach subject was studied on four occa-sions, separated by 7 days, in random-ized, double-blind fashion. On the eveningbefore each study day (;1900 h), subjectsconsumed a standardized beef lasagnameal (2,472 kJ; McCain Foods Proprie-tary Ltd., Victoria, Australia) with a tabletof either 50 mg vildagliptin (Novartis,NSW, Australia) or matching placebo, to-gether with their usual evening dose ofmetformin. Compliance was reinforcedwith a phone call and evaluated by pillcount.

Subjects then fasted untilmorning andattended the laboratory at ;0730 h.Any usual morning medications, includ-ing metformin, were withheld until theend of the visit. An intravenous cannulawas inserted for repeated blood sam-pling, and subjects remained seatedthroughout the study. Vildagliptin 50 mgor a matching placebo tablet was ad-ministered orally with 30 mL water (att = 290 min), followed after 60 min(t = 230 min) by a 250-mL preload drinkcontaining either 25 g whey protein iso-late (89 kcal; Murray Goulburn, Mel-bourne, Australia) or 25 g controlflavoring (8 kcal; Cottee’s, Southbank,Australia), so that the four treatments

were vildagliptin + whey preload(VILD/WHEY), vildagliptin + control pre-load (VILD/CTRL), placebo + whey pre-load (PLBO/WHEY), and placebo +control preload (PLBO/CTRL) (Fig. 1).Thirty minutes later (between t = 0 and5 min), subjects ate a semisolid mealcomprising 65 g powdered potato (DebInstant Mashed Potato; Continental,Epping, Australia) and 20 g glucose, recon-stituted with 200 mL water and one eggyolk containing 100 mL 13C-octanoate.Breath samples were collected immedi-ately before and every 5 min after mealingestion in the 1st hour and every15 min for a further 3 h for the measure-ment of gastric emptying. Venous bloodsamples were collected at t =290,230,215, 0, 15, 30, 60, 90, 120, 180, and 240min, for measurements of plasma glu-cose, insulin, glucagon, and GIP andGLP-1 (both total and intact forms).Samples were stored on ice in tubescontaining EDTA and DPP-4 inhibi-tor (DPP4-010; Linco Research Inc.,St. Charles, MO), before centrifugationat 3,200 rpm for 15 min at 48C within15 min of collection. Plasma was sepa-rated and stored at 2808C for subse-quent analysis.

MeasurementsPlasma glucose concentrations weremeasured by the glucose oxidase tech-nique (2300 STAT Plus; YSI, YellowSprings, OH). Plasma insulin was mea-sured by ELISA immunoassay (10-1113;Mercodia, Uppsala, Sweden). Plasmaglucagon was measured by radioimmu-noassay (GL-32K; Millipore, Billerica,MA). Plasma GIP and GLP-1 analyseswere performed as previously described(14,15). Intact and total GIP wereanalyzed with the N-terminally andC-terminally directed antisera 98171 (14)and 80867 (15), respectively. Intact GLP-1was measured using a two-site ELISA(C-terminally directed GLP-1F5–catchingantibody and N-terminally directedMab26.1-detecting antibody) (14). TotalGLP-1 was assayed using antiserum89390, requiring the intact amidatedC terminus of the molecule and reactingequally with intact GLP-1 and the pri-mary (N-terminally truncated) metabo-lite (14).

The 13CO2 concentration in breathsamples was measured by an isotoperatio mass spectrometer (ABCA 2020;Europa Scientific, Crewe, U.K.) with an

512 Vildagliptin and Protein Preload Diabetes Care Volume 39, April 2016

online gas chromatographic purificationsystem. The half-emptying time (T50)was calculated, using the formula de-scribed by Ghoos et al. (16). This methodhas been validated against scintigraphyfor the measurement of gastric empty-ing (17).

Statistical AnalysisThe differences in fasting plasma glu-cose and hormone levels for vildagliptinversus placebo, and for whey versuscontrol preload days, were evaluatedusing two-factor repeated-measuresANOVA, with drug and preload as fac-tors. Areas under the curve (AUCs)were calculated using the trapezoidalrule for plasma glucose and hormones,which, together with postprandial peak

plasma glucose and T50, were comparedusing one-factor repeated-measuresANOVA. The variables were alsoassessed using two-factor repeated-measures ANOVA, with treatment andtime as factors. Post hoc comparisons,adjusted for multiple comparisons byBonferroni-Holm correction, were per-formed if ANOVAs revealed signifi-cant effects. Relationships betweenvariables were assessed using univari-ate linear regression analysis. Based onour previous work (9,10,12), a samplesize of 22 subjects was calculated tohave at least 80% power (at a = 0.008to enable correction for multiple posthoc testing) to detect a difference in theAUC for blood glucose of 2.5 mmol/L z hwith an SD of 3.3 mmol/L z h between

treatments. All analyses were performedwith SPSS Statistics (Version 21; IBM, NewYork, NY). Data are presented asmeans6SEM.P,0.05was considered statisticallysignificant.

RESULTS

Compliance with medication was com-plete, and all subjects tolerated thestudy well.

Plasma Glucose ConcentrationsFasting plasma glucose did not differbetweenwhey and control butwas slightlylower with vildagliptin than placebo(P = 0.003 at t = 290 min and P = 0.006at t = 230 min, respectively) (Table 1).Before the meal (t = 230 to 0 min),plasma glucose remained unchangedafter both control and whey preloads. Af-ter the meal, the plasma glucose excur-sion showed significant treatment effectson both the peak and AUC (P , 0.001for each), such that the peak was lowerfor PLBO/WHEY and VILD/CTRL versusPLBO/CTRL and was lowest after VILD/WHEY (P , 0.05 for each), whereasthe AUCwas lower for VILD/CTRL versusPLBO/CTRL and lowest after VILD/WHEY (P , 0.05 for each), withouta significant difference betweenPLBO/CTRL and PLBO/WHEY (Table 2and Fig. 2A).

Plasma Insulin ConcentrationsFasting plasma insulin (at t = 290 and230 min) did not differ across studydays (Table 1). Before the meal (t =230 to 0 min), plasma insulin remained

Figure 1—Outline of study protocol. On the preceding evening of each study visit, either 50 mgvildagliptin (VILD) or a placebo (PLBO) tablet was givenwith a standardizedmeal. On each studyday, at t = 290 min, either 50 mg vildagliptin or a placebo tablet was given. At t = 230 min, a250-mL preload drink containing either 25 g whey or control flavoring (CTRL) was consumed,followed by a 13C-octanoate–labeled mashed potato meal (at t = 0–5 min). Venous blood wasobtained at frequent intervals during the study for measurement of plasma glucose, insulin,glucagon, GLP-1, and GIP. Gastric emptying was determined by breath test.

Table 1—Fasting plasma glucose, insulin, glucagon, and GIP and GLP-1 (both total and intact forms) concentrations on thecontrol (CTRL) and whey preload days after an acute dose of 50 mg vildagliptin (VILD) or placebo (PLBO) at ∼1900 h on thepreceding evening in patients with type 2 diabetes treated with metformin (n = 22)*

PLBO (CTRL) PLBO (WHEY) VILD (CTRL) VILD (WHEY)Differencedue to drug

Differencedue to preload

At t = 290 minPlasma glucose (mmol/L) 8.8 6 0.5 8.8 6 0.5 8.4 6 0.5 8.3 6 0.4 0.003 0.501Plasma insulin (mU/L) 9.3 6 1.5 8.4 6 1.0 8.2 6 1.1 9.0 6 1.2 0.534 0.900Plasma glucagon (pg/mL) 95.5 6 4.7 93.6 6 4.0 91.8 6 4.8 90.2 6 24.5 0.092 0.249Plasma total GIP (pmol/L) 11.5 6 1.1 10.5 6 0.9 9.8 6 0.7 9.7 6 0.9 0.022 0.202Plasma intact GIP (pmol/L) 6.4 6 0.4 6.4 6 0.5 7.9 6 0.6 8.2 6 0.7 0.002 0.639Plasma total GLP-1 (pmol/L) 4.6 6 0.4 5.2 6 0.8 5.4 6 0.7 4.9 6 0.6 0.494 0.803Plasma intact GLP-1 (pmol/L) 0.2 6 0.1 0.3 6 0.1 0.7 6 0.2 1.0 6 0.4 0.003 0.340

At t = 230 minPlasma glucose (mmol/L) 8.7 6 0.4 8.6 6 0.5 8.4 6 0.5 8.1 6 0.4 0.006 0.251Plasma insulin (mU/L) 8.2 6 1.4 6.7 6 1.0 7.9 6 1.3 7.9 6 1.2 0.216 0.130Plasma glucagon (pg/mL) 88.0 6 4.5 86.5 6 3.7 83.7 6 4.2 84.2 6 3.7 0.007 0.776Plasma total GIP (pmol/L) 9.6 6 0.8 10.0 6 1.1 8.2 6 0.6 8.6 6 0.9 0.004 0.553Plasma intact GIP (pmol/L) 5.6 6 0.5 6.0 6 0.4 7.3 6 0.6 8.3 6 0.9 0.003 0.093Plasma total GLP-1 (pmol/L) 4.1 6 0.4 5.0 6 0.6 4.3 6 0.6 4.5 6 0.7 0.378 0.097Plasma intact GLP-1 (pmol/L) 0.2 6 0.1 0.2 6 0.1 0.6 6 0.2 0.7 6 0.3 0.007 0.905

Data are mean 6 SEM. *Two-factor ANOVA, with drug and preload as factors, was used to determine statistical difference.

care.diabetesjournals.org Wu and Associates 513

unchanged on the control days (PLBO/CTRL and VILD/CTRL) and increased af-ter the whey preload (PLBO/WHEY andVILD/WHEY). After the meal, plasma in-sulin concentrations showed a signifi-cant treatment effect for AUC (P ,0.001), such that plasma insulin washigher for PLBO/WHEY versus PLBO/CTRL and for VILD/WHEY versus VILD/CTRL (P, 0.05 for each), without signif-icant difference between PLBO/CTRLand VILD/CTRL or between PLBO/WHEY and VILD/WHEY (Table 2 andFig. 2C).

Plasma Glucagon ConcentrationsFasting plasma glucagon did not differbetween the whey and control daysbut was slightly lower with vildagliptincompared with placebo (P = 0.092 att =290 min and P = 0.007 at t =230 min,respectively) (Table 1). Before the meal(t = 230 to 0 min), plasma glucagonconcentrations remained unchangedon the control days (PLBO/CTRL andVILD/CTRL) but increased markedly af-ter the whey preload (PLBO/WHEY andVILD/WHEY). After the meal, plasmaglucagon showed a significant treat-ment effect for AUC (P , 0.001), suchthat plasma glucagon was higher forPLBO/WHEY versus PLBO/CTRL and forVILD/WHEY versus VILD/CTRL but lowerfor VILD/CTRL versus PLBO/CTRL and forVILD/WHEY versus PLBO/WHEY (P ,0.05 for each) (Table 2 and Fig. 2D).

Plasma Total and Intact GIPConcentrationsFasting plasma total and intact GIP didnot differ between thewhey and controldays, but plasma total GIP was slightlylower (P = 0.022 at t = 290 min and

P = 0.004 at t = 230 min, respectively)whereas intact GIP was higher (P =0.002 at t = 290 min and P = 0.003 att = 230 min, respectively) with vildaglip-tin compared with placebo (Table 1). Be-fore the meal (t = 230 to 0 min), bothtotal and intact GIP concentrations re-mained unchanged on the control days(PLBO/CTRL and VILD/CTRL) but in-creased after the whey preload (PLBO/WHEY and VILD/WHEY). After the meal,both total and intact GIP concentrationsshowed significant treatment effects forthe AUC (P , 0.001 for each), such thattotal GIP was higher for PLBO/WHEYversus PLBO/CTRL and for VILD/WHEYversus VILD/CTRL but lower for VILD/CTRL versus PLBO/CTRL and for VILD/WHEY versus PLBO/WHEY, whereas in-tact GIP was higher for PLBO/WHEY andVILD/CTRL versus PLBO/CTRL and high-est after VILD/WHEY (P, 0.05 for each)(Table 2 and Fig. 2E and F).

Plasma Total and Intact GLP-1ConcentrationsFasting plasma total GLP-1 did not differbetween study days, whereas intactGLP-1 concentrations did not differ be-tween the whey and control days butwere slightly higher with vildagliptincompared with placebo (P = 0.003at t = 290 min and P = 0.007 at t =230 min, respectively) (Table 1). Beforethemeal (t =230 to 0min), both total andintact GLP-1 concentrations remainedunchanged on the control days (PLBO/CTRL and VILD/CTRL) but increased af-ter the whey preload (PLBO/WHEY andVILD/WHEY). After the meal, therewere significant treatment effects onthe AUC for both total and intact GLP-1

(P , 0.001 for each), such that totalGLP-1 was higher for PLBO/WHEY ver-sus PLBO/CTRL and for VILD/WHEYversus VILD/CTRL but lower for VILD/CTRL versus PLBO/CTRL and for VILD/WHEY versus PLBO/WHEY, whereas in-tact GLP-1 was higher for PLBO/WHEYand VILD/CTRL versus PLBO/CTRL andhighest after VILD/WHEY (P, 0.05 foreach) (Table 2 and Fig. 2G and H).

Gastric EmptyingThere was a significant treatment effecton the half-emptying time of the meal(T50) (P, 0.001), such that gastric emp-tying was slower for PLBO/WHEY (T50:172.3 6 5.7 min) and VILD/CTRL (T50:161.3 6 5.7 min) versus PLBO/CTRL(T50: 147.76 4.4 min) and was slowestafter VILD/WHEY (T50: 193.76 7.3 min)(P , 0.05 for each) (Fig. 2B).

Relationships of Glycemia WithGastric Emptying, Insulin, GIP, andGLP-1When data from the four study visitswere pooled, postprandial glycemia att = 30, 60, and 90 min was inverselyrelated to T50 (r = 20.25, P = 0.02;r = 20.40, P , 0.001; r = 20.28, P =0.008, respectively), and at t = 240 mindirectly related to T50 (r = 0.25,P = 0.02).

Compared with placebo, the magni-tude of reduction in peak postprandialglucose with vildagliptin was relateddirectly to the slowing of gastric emp-tying (r = 0.42, P = 0.005), while thereduction in AUC for postprandialplasma glucose was related to the in-crease in AUC for plasma intact GLP-1(r = 0.31, P = 0.04), but not GIP, insulin,or glucagon.

Table 2—Effects of vildagliptin, with or without a whey preload, on postprandial glycemic peak and the AUC for plasmaglucose, insulin, glucagon, and GIP and GLP-1 (both total and intact forms) in response to a carbohydrate meal in patientswith type 2 diabetes treated with metformin (n = 22)

PLBO/CTRL PLBO/WHEY VILD/CTRL VILD/WHEY P value

Glycemic peak (mmol/L) 15.4 6 0.7 14.7 6 0.9§ 14.4 6 0.8# 13.1 6 0.8a,d,« ,0.001

Glucose AUC230 to 240min (mmol/L z h) 51.6 6 3.2 51.2 6 3.6 48.7 6 3.3# 46.3 6 3.0a,d,« ,0.001

Insulin AUC230 to 240min (mU/L z h) 130.7 6 20.4 155.0 6 21.5§ 137.9 6 23.4 166.3 6 26.5a,« ,0.001

Glucagon AUC230 to 240min (pg/mL z h) 380.5 6 16.9 496.4 6 20.8§ 358.7 6 15.8# 468.7 6 21.1a,d,« ,0.001

Total GIP AUC230 to 240min (pmol/L z h) 163.9 6 14.9 193.3 6 21.0§ 135.6 6 12.4# 167.4 6 24.5d,« ,0.001

Intact GIP AUC230 to 240min (pmol/L z h) 59.0 6 2.9 72.5 6 5.5§ 114.7 6 9.8# 144.3 6 19.6a,d,« ,0.001

Total GLP-1 AUC230 to 240min (pmol/L z h) 30.4 6 2.0 37.4 6 2.5§ 25.5 6 2.2# 31.8 6 3.1d,« ,0.001

Intact GLP-1 AUC230 to 240min (pmol/L z h) 1.7 6 0.4 3.9 6 0.8§ 6.4 6 1.0# 10.1 6 1.5a,d,« ,0.001

Data are mean 6 SEM. Post hoc comparisons were adjusted by Bonferroni-Holm correction. One-factor ANOVA was used to determine statisticaldifference. §P , 0.05, PLBO/WHEY vs. PLBO/CTRL. #P , 0.05, VILD/CTRL vs. PLBO/CTRL. aP , 0.05, VILD/WHEY vs. PLBO/CTRL. dP , 0.05,VILD/WHEY vs. PLBO/WHEY. eP , 0.05, VILD/WHEY vs. VILD/CTRL.

514 Vildagliptin and Protein Preload Diabetes Care Volume 39, April 2016

CONCLUSIONS

In this study of patients with type 2diabetes who were relatively wellcontrolled on metformin alone, weobserved that 1) a low-dose whey pre-load reduces postprandial glycemic ex-cursions in association with slowing ofgastric emptying and stimulation of

incretin hormone, insulin, and glucagonsecretion; 2) acute dosing with vildaglip-tin increases plasma intact GLP-1 andGIP, suppresses plasma glucagon andtotal GLP-1 and GIP, slows gastric emp-tying, and attenuates postprandial gly-cemia independent of enhancement ofinsulin secretion; and 3) combining

vildagliptin with a whey preload ismore effective for increasing plasmaintact GLP-1 and GIP, slowing gastricemptying, and reducing postprandial gly-cemia, compared with either treatmentalone. Remarkably, the addition of thewhey preload to treatmentwith vildaglip-tin approximately doubled the reductionin peak postprandial blood glucose.

The dose of whey, although lowerthan used previously (10,11), provedsufficient to stimulate GLP-1 and GIP se-cretion and slow gastric emptying.Whey also induced early increases inplasma insulin and glucagon, probablydue to direct b- and a-cell stimulationby absorbed amino acids, since themod-est elevation in intact GLP-1 and GIP oc-curred later. However, the effects ofinsulin and glucagon on blood glucoseappeared to be counterbalanced. Al-though the modest increase in intactGLP-1 and GIP in the absence of vildaglip-tin was not associated with significantlowering of the AUC for postprandialglycemia, the “early” phase of the post-prandial glycemic excursion was attenu-ated, in line with slowing of gastricemptying after whey (18). The latterwould also favor a reduction in themeal-induced insulin response (19). Inthe presence of vildagliptin, both theearly rise andoverall AUC for postprandialglycemia were attenuated after wheycompared with control, highlighting thecomplementary actions of a dietary strat-egy combined with DPP-4 inhibition.

As anticipated, vildagliptin increasedintact GLP-1 and GIP concentrations, aneffect that was enhanced by the wheypreload. The reduction in total GLP-1and GIP concentrations is consistentwith the concept of negative feedbackon incretin hormone secretion (1). Al-though there is recent evidence thatabout half of the insulinotropic andglucose-lowering effects of sitagliptin areGLP-1 independent (7), we showedthat the magnitude of reduction in post-prandial glycemia was related directlyto the increase in plasma intact GLP-1,consistent with complementary glucose-lowering actions of GLP-1 in type 2 diabetes(1). Elevated intact GIP concentrations arelikely to contribute less (3,20,21), and GIPadministration can even antagonize theglucagonostatic effect of GLP-1 (20). Inthe current study, the glucagonostaticeffect of intact GLP-1 prevailed, andmay, to some extent, have contributed

Figure 2—Effects of vildagliptin, with or without a whey preload, on plasma glucose (A), gastrichalf-emptying time (B), plasma insulin (C), plasma glucagon (D), and plasma GIP (E and F for totaland intact forms, respectively) and GLP-1 (G and H for total and intact forms, respectively) inresponse to a high-carbohydrate meal in patients with type 2 diabetes managed by metforminmonotherapy (n = 22). The four treatments were VILD/WHEY, VILD/CTRL, PLBO/WHEY, andPLBO/CTRL. Repeated-measures ANOVA was used to determine statistical difference. Resultsof ANOVA were reported as P values for differences by experiment (Tx), differences over time(Time), and differences due to interaction of experiment and time (Tx*time). Post hoc compar-isons were adjusted by Bonferroni-Holm correction. *P , 0.05, PLBO/WHEY vs. PLBO/CTRL;#P, 0.05, VILD/CTRL vs. PLBO/CTRL;aP, 0.05, VILD/WHEY vs. PLBO/CTRL; dP, 0.05, VILD/WHEYvs. PLBO/WHEY; eP, 0.05, VILD/WHEY vs. VILD/CTRL. Data are mean6 SEM.

care.diabetesjournals.org Wu and Associates 515

to glucose lowering by vildagliptin.However, relative to the increase ofglucagon after whey, this effect wasmodest. Therefore, additional mecha-nisms are likely to be important in ac-counting for the glycemic effect aftercombined treatment with whey preloadand vildagliptin. Despite augmentationof intact GLP-1 after vildagliptin, partic-ularly when combined with whey, therewas minimal increase in insulin, proba-bly because the insulinotropic activityof GLP-1 is glucose dependent (1). Theglycemic effect of vildagliptin observedin the current study was also unlikely tobe related to improvement in insulin sen-sitivity; administration of 50 mg vildaglip-tin twice daily for 9 days was previouslyreported not to alter insulin action inpatients with type 2 diabetes (22). Nev-ertheless, during pancreatic clampstudies where insulin and glucagon se-cretion are inhibited, GLP-1 remains ef-fective to suppress endogenous glucoseproduction (23) and enhance peripheralglucose uptake (24), suggesting thatGLP-1 has the capacity to lower bloodglucose independent of changes in islethormones.Vildagliptin slowed gastric emptying

modestly both in the absence and pres-ence of the whey preload, associatedwith the reduction in peak blood glu-cose. This has been reported previ-ously in patients with type 2 diabetesreceiving a single dose of vildagliptin(25) and is consistent with evidencethat endogenous GLP-1 slows gastricemptying (4,7). Several studies failedto show any effect of DPP-4 inhibitorson gastric emptying in health or type 2diabetes (9,26–28), probably related tochanges in other peptides that regulategastric emptying (e.g., reduced conver-sion of peptide YY[1-36] to [3-36]) (29),and differences in test meal, subjectcharacteristics, or duration of DPP-4 in-hibition. Sustained exposure to elevatedGLP-1 during prolonged DPP-4 inhibi-tionmay potentially cause tachyphylaxisfor its effect on gastric emptying (30).Furthermore, the magnitude of GLP-1secretion is dependent on the load andnutrient composition of themeal (1) andmay be influenced by concurrent medi-cations (e.g., metformin) (31).We studied only one dose of whey on

the basis of its established effect on gas-tric emptying (12); testing of differentdoses at various intervals before the

meal may further optimize the interac-tion with DPP-4 inhibitors. Alternativepreloads could be evaluated; Tricoet al. (32) recently reported that a pre-load of protein (egg) and fat (cheese)reduced the glycemic response to oralglucose substantially, but the effect onthe secretion of incretin hormones wasrather modest, despite higher caloricvalue of the preload compared withour whey preload. We studied onlymale patients with relatively good glyce-mic control on metformin in order tominimize heterogeneity in this “proofof concept” study. Therefore, thereshould be caution in generalizing ourfindings to the broader community ofpatients with type 2 diabetes.

In summary, in metformin-treatedpatients with type 2 diabetes, the effi-cacy of vildagliptin in reducing postpran-dial glycemia is substantially improvedby combination with a small whey pre-load, associated with augmentation ofplasma intact GLP-1 concentrationsand slowing of gastric emptying. Thestrategy of combining a dietary andpharmacological approach warrantsevaluation in longer-term clinical trials.

Acknowledgments. The authors thank KylieLange (National Health Medical Research Coun-cil Centre of Research Excellence in TranslatingNutritional Science to Good Health) for herexpert statistical advice and Dr. Ada Lam (RoyalAdelaide Hospital Pharmacy) for her assistancein preparation and blinding of the treatments.Funding. T.W. is supported by a Royal AdelaideHospital Research Committee Early Career Fellow-ship.K.L.J. is fundedbyaNationalHealthandMedicalResearch Council Senior Career DevelopmentAward(627011).Duality of Interest. This investigator-initiatedstudy was funded by Novartis. Novartis also ar-ranged the supply of vildagliptin and placebo.T.W. has received research funding fromAstraZeneca. T.J.L. has received research fundingfrom Novartis. C.F.D. has received consultant/lecture fees from companies with an interest indeveloping and marketing incretin-based thera-pies (BMS, Boehringer Ingelheim, Eli Lilly andCompany, Merck, Novartis, and Novo Nordisk),and her spouse is employed by Merck and holdsMerck stock. M.H. has participated in the advi-sory boards and/or symposia for Novo Nordisk,Sanofi, Novartis, Eli Lilly and Company, Merck,Boehringer Ingelheim, and AstraZeneca andhas received honoraria for this activity. K.L.J.has received research funding from Sanofiand drug supplies from Merck. C.K.R. has re-ceived research funding from AstraZeneca,Merck, Eli Lilly and Company, and Novartis.No other potential conflicts of interest relevantto this article were reported.

Author Contributions. T.W. was involved inthe conception, design, and coordination of thestudy, subject recruitment, data collection andinterpretation, statistical analysis, andwriting ofthe manuscript. T.J.L. was involved in the con-ceptionanddesignof the study.M.J.B.,M.B., andX.Z. assisted in data collection. C.F.D. performedhormone assays and was involved in datainterpretation. M.H. and K.L.J. were involvedin the conception of the study and data in-terpretation. C.K.R. was involved in the con-ception and design of the study and dataanalysis and interpretation. All authors criti-cally reviewed the manuscript and approvedthe publication of the final version of themanuscript. T.W. and C.K.R. are the guarantorsof this work and, as such, had full access to allthe data in the study and take responsibility forthe integrity of the data and the accuracy of thedata analysis.PriorPresentation.Parts of the data from thisstudy were presented at the 2015 WorldDiabetes Congress, Vancouver, Canada, 30November–4 December 2015.

References1. Wu T, Rayner CK, Horowitz M. Incretins.Handb Exp Pharmacol. 23 April 2015 [Epubahead of print]2. Monnier L, Lapinski H, Colette C. Contribu-tions of fasting and postprandial plasma glucoseincrements to the overall diurnal hyperglycemiaof type 2 diabetic patients: variations with in-creasing levels of HbA(1c). Diabetes Care 2003;26:881–8853. Nauck MA, Heimesaat MM, Orskov C, HolstJJ, Ebert R, Creutzfeldt W. Preserved incretinactivity of glucagon-like peptide 1 [7-36 amide]but not of synthetic human gastric inhibitorypolypeptide in patients with type-2 diabetesmellitus. J Clin Invest 1993;91:301–3074. Deane AM, Nguyen NQ, Stevens JE, et al. En-dogenous glucagon-like peptide-1 slows gastricemptying in healthy subjects, attenuating post-prandial glycemia. J Clin Endocrinol Metab2010;95:215–2215. Hare KJ, Vilsbøll T, Asmar M, Deacon CF,Knop FK, Holst JJ. The glucagonostatic and insu-linotropic effects of glucagon-like peptide 1contribute equally to its glucose-lowering ac-tion. Diabetes 2010;59:1765–17706. Gutzwiller JP, Goke B, Drewe J, et al. Glucagon-like peptide-1: a potent regulator of food intakein humans. Gut 1999;44:81–867. Aulinger BA, Bedorf A, Kutscherauer G, et al.Defining the role of GLP-1 in the enteroinsulinaraxis in type 2 diabetes using DPP-4 inhibitionand GLP-1 receptor blockade. Diabetes 2014;63:1079–10928. Wu T, Rayner CK, Jones K, Horowitz M.Dietary effects on incretin hormone secretion.Vitam Horm 2010;84:81–1109. Wu T, Bound MJ, Zhao BR, et al. Effects of aD-xylose preload with or without sitagliptin ongastric emptying, glucagon-like peptide-1, andpostprandial glycemia in type 2 diabetes. Diabe-tes Care 2013;36:1913–191810. Ma J, Stevens JE, Cukier K, et al. Effects of aprotein preload on gastric emptying, glycemia,and gut hormones after a carbohydrate meal indiet-controlled type 2 diabetes. Diabetes Care2009;32:1600–1602

516 Vildagliptin and Protein Preload Diabetes Care Volume 39, April 2016

11. Jakubowicz D, Froy O, Ahren B, et al. Incre-tin, insulinotropic and glucose-lowering effectsof whey protein pre-load in type 2 diabetes: arandomised clinical trial. Diabetologia 2014;57:1807–181112. Ma J, Jesudason DR, Stevens JE, et al. Sus-tained effects of a protein ‘preload’ on glycae-mia and gastric emptying over 4 weeks inpatients with type 2 diabetes: a randomizedclinical trial. Diabetes Res Clin Pract 2015;108:e31–e3413. American Diabetes Association. (7) Ap-proaches to glycemic treatment. Diabetes Care2015;38(Suppl. 1):S41–S4814. Vilsbøll T, Krarup T, Sonne J, et al. Incretinsecretion in relation to meal size and bodyweight in healthy subjects and people withtype 1 and type 2 diabetes mellitus. J Clin Endo-crinol Metab 2003;88:2706–271315. LindgrenO, Carr RD,Deacon CF, et al. Incretinhormone and insulin responses to oral versus in-travenous lipid administration in humans. J ClinEndocrinol Metab 2011;96:2519–252416. Ghoos YF, Maes BD, Geypens BJ, et al. Mea-surement of gastric emptying rate of solids bymeans of a carbon-labeled octanoic acid breathtest. Gastroenterology 1993;104:1640–164717. Chew CG, Bartholomeusz FD, Bellon M,Chatterton BE. Simultaneous 13C/14C dualisotope breath test measurement of gastricemptying of solid and liquid in normal subjectsand patients: comparison with scintigraphy.Nucl Med Rev Cent East Eur 2003;6:29–3318. Marathe CS, Horowitz M, Trahair LG,et al. Relationships of early and late glycemicresponses with gastric emptying during an

oral glucose tolerance test. J Clin EndocrinolMetab 2015;100:3565–357119. Gentilcore D, Chaikomin R, Jones KL, et al.Effects of fat on gastric emptying of and theglycemic, insulin, and incretin responses to acarbohydrate meal in type 2 diabetes. J Clin En-docrinol Metab 2006;91:2062–206720. Mentis N, Vardarli I, Kothe LD, et al. GIPdoes not potentiate the antidiabetic effects ofGLP-1 in hyperglycemic patients with type 2diabetes. Diabetes 2011;60:1270–127621. Wu T, Ma J, Bound MJ, et al. Effects ofsitagliptin on glycemia, incretin hormones,and antropyloroduodenal motility in responseto intraduodenal glucose infusion in healthylean and obese humans and patients withtype 2 diabetes treated with or without met-formin. Diabetes 2014;63:2776–278722. Dalla Man C, Bock G, Giesler PD, et al. Di-peptidyl peptidase-4 inhibition by vildagliptinand the effect on insulin secretion and actionin response tomeal ingestion in type 2 diabetes.Diabetes Care 2009;32:14–1823. Seghieri M, Rebelos E, Gastaldelli A, et al.Direct effect of GLP-1 infusion on endogenousglucose production in humans. Diabetologia2013;56:156–16124. Edgerton DS, Johnson KM, Neal DW, et al.Inhibition of dipeptidyl peptidase-4 by vilda-gliptin during glucagon-like peptide 1 infusionincreases liver glucose uptake in the consciousdog. Diabetes 2009;58:243–24925. Woerle HJ, Lindenberger T, Linke R, et al.A single dose of vildagliptin (VILDA) deceleratesgastric emptying (GE) in patients with type 2diabetes (T2DM). Diabetes 2007;56:A133

26. Vella A, Bock G, Giesler PD, et al. Theeffect of dipeptidyl peptidase-4 inhibitionon gastric volume, satiation and enteroendo-crine secretion in type 2 diabetes: a double-blind, placebo-controlled crossover study.Clin Endocrinol (Oxf) 2008;69:737–74427. Vella A, Bock G, Giesler PD, et al. Effects ofdipeptidyl peptidase-4 inhibition on gastroin-testinal function, meal appearance, and glucosemetabolism in type 2 diabetes. Diabetes 2007;56:1475–148028. Stevens JE, Horowitz M, Deacon CF, NauckM, Rayner CK, Jones KL. The effects of sitagliptinon gastric emptying in healthy humans - a ran-domised, controlled study. Aliment PharmacolTher 2012;36:379–39029. Witte AB, Gryback P, Holst JJ, et al. Dif-ferential effect of PYY1-36 and PYY3-36 ongastric emptying in man. Regul Pept 2009;158:57–6230. Umapathysivam MM, Lee MY, Jones KL,et al. Comparative effects of prolonged andintermittent stimulation of the glucagon-likepeptide 1 receptor on gastric emptying andglycemia. Diabetes 2014;63:785–79031. Wu T, Thazhath SS, Bound MJ, Jones KL,Horowitz M, Rayner CK. Mechanism ofincrease in plasma intact GLP-1 by metfor-min in type 2 diabetes: stimulation of GLP-1secretion or reduction in plasma DPP-4 ac-tivity? Diabetes Res Clin Pract 2014;106:e3–e632. Trico D, Baldi S, Tulipani A, et al. Mechanismsthrough which a small protein and lipid preloadimproves glucose tolerance. Diabetologia 2015;58:2503–2512

care.diabetesjournals.org Wu and Associates 517