Elevated PostoperativeEndogenous GLP-1 Levels MediateEffects of Roux-en-Y GastricBypass on Neural Responsivityto Food CuesDiabetes Care 2017;40:1522–1529 | https://doi.org/10.2337/dc16-2113
OBJECTIVE
It has been suggested thatweight reduction and improvements in satiety after Roux-en-Y gastric bypass (RYGB) are partly mediated via postoperative neuroendocrinechanges. Glucagon-like peptide-1 (GLP-1) is a gut hormone secreted after food in-gestionand is associatedwith appetite andweight reduction,mediatedvia effects onthe central nervous system (CNS). Secretion of GLP-1 is greatly enhanced after RYGB.We hypothesized that postoperative elevated GLP-1 levels contribute to the im-proved satiety regulation after RYGB via effects on the CNS.
RESEARCH DESIGN AND METHODS
Effects of the GLP-1 receptor antagonist exendin 9-39 (Ex9-39) and placebo wereassessed in 10women before and after RYGB.We used functionalMRI to investigateCNS activation in response to visual food cues (pictures) and gustatory food cues(consumption of chocolate milk), comparing results with Ex9-39 versus placebobefore and after RYGB.
RESULTS
After RYGB, CNS activation was reduced in the rolandic operculum and caudatenucleus in response to viewing food pictures (P = 0.03) and in the insula in responseto consumption of palatable food (P = 0.003). GLP-1 levelswere significantly elevatedpostoperatively (P < 0.001). After RYGB, GLP-1 receptor blockade resulted in a largerincrease in activation in the caudate nucleus in response to food pictures (P = 0.02)and in the insula in response to palatable food consumption (P = 0.002).
CONCLUSIONS
We conclude that the effects of RYGB on CNS activation in response to visual andgustatory food cues may be mediated by central effects of GLP-1. Our findings pro-vide further insights into the mechanisms underlying the weight-lowering effects ofRYGB.
Bariatric surgery is currently themost effective therapeutic modality for severe obesityin terms of substantial weight loss and long-term efficacy (1). The most commonlyperformed procedure is Roux-en-Y gastric bypass surgery (RYGB), which comprises theformation of a small gastric pouch, which is connected to the midjejunum, bypassingthe duodenum and proximal jejunum. This may lead to reduced ingestive capacity and
1Diabetes Center/Department of Internal Medi-cine, VU University Medical Center, Amsterdam,the Netherlands2Department of Psychiatry, VU University Medi-cal Center, Amsterdam, the Netherlands3Department of Internal Medicine, SlotervaartHospital, Amsterdam, the Netherlands4Department of Radiology and Nuclear Medi-cine, VU University Medical Center, Amsterdam,the Netherlands5Novo Nordisk Foundation Center for Basic Met-abolic Research and Department of BiomedicalSciences, Panum Institute, University of Copen-hagen, Copenhagen, Denmark6Department of Internal Medicine/EndocrineSection, VU University Medical Center, and De-partment of Clinical Neuropsychology, VU Uni-versity, Amsterdam, the Netherlands
also some reduction in the absorption ofcalories. However, it has been suggestedthat the reduction in caloric intake afterRYGB is not only explained by these re-strictive and/or absorption-limitingmechanisms, but that RYGB has addi-tional effects on caloric intake by dimin-ishing appetite via changes in the centralnervous system (CNS) and endocrine sys-tem (2).The CNS is important in the regulation
of food intake, and it has been proposedthat altered CNS responses may contrib-ute to disturbances in this regulation. Al-tered responses to visual and gustatoryfood cues have indeed been described inobese individuals, using functional MRI(fMRI) (3–5). Interestingly, weight lossafter RYGB is paralleled by decreased re-sponsivity of the CNS tohigh-calorie visualfood cues, measured with fMRI (6,7),which may contribute to the reduced he-donic drive to consume highly palatablefood and therefore contribute to the sub-stantial weight loss after RYGB. However,the mechanism explaining this alteredCNS responsivity to food cues afterRYGB is unknown.Appetite and satiety are regulated by
the interaction of several neurologicaland hormonal signals. Gut hormones,as a part of the gut-brain axis, convey in-formation about the nutritional statusto the CNS and contribute to the centralregulation of food intake (8). RYGB is con-sistently associated with increased post-operative levels of the gut hormoneglucagon-like peptide-1 (GLP-1) (9,10),which is secreted after food ingestionfrom enteroendocrine L cells. In additionto its glucose-regulating effects, GLP-1 isassociated with reduced appetite, foodintake, and body weight (11), which is atleast partly mediated via effects in theCNS (12,13). Neuroendocrine changesafter RYGB, such as the enhanced GLP-1secretion, are regarded as possible mech-anisms to account for a part of appetiteand weight reduction and the sustainedefficacy of this procedure (14,15).We have previously shown, by means
of a GLP-1 receptor antagonist, that en-dogenous GLP-1 mediates the satiatingeffects ofmeal intake on CNS responsivityto food cues in humans (13). We there-fore hypothesized that the increasedGLP-1 response after RYGB may enhanceeffects of GLP-1 on the satiety and rewardpathways in the CNS, thereby contrib-uting to the observed postoperative
decreases in food intake andbodyweight.In the current fMRI study,we investigatedthe role of endogenous GLP-1 in the im-proved responsivity of the CNS to foodcues after RYGB by comparing the effectsof the selectiveGLP-1 receptor antagonistexendin 9-39 (Ex9-39) with placebo be-fore and after RYGB.
RESEARCH DESIGN AND METHODS
ParticipantsThe study (NCT01363609) was approvedby the Medical Ethics Review Committeeof the VUUniversityMedical Center. Subjectswere includedafterwritten informedconsentwas obtained. Ten female candidates forRYGBwererecruited fromtheCenter forBari-atric Surgery at the Slotervaartziekenhuis(Amsterdam, the Netherlands). Subjectswere eligible if they were 40–65 years old,had a BMI.35 kg/m2, had a stable bodyweight during the previous 1 month(i.e., ,5% reported change), and wereright handed. Subjects were not on a for-mal calorie-restricted diet prior to and/orduring the study but received generaladvice on healthy food choices. Exclusioncriteria were a history of neurologicaldisease, the use of any centrally actingagent, psychiatric disorders, or current di-abetes. Three patients used antihyper-tensive medication, one patient used acholesterol-lowering agent, and three
patients used thyroxin for the treatmentof hypothyroidism.
General Experimental ProtocolThe study consisted of four separate testvisits. The first two visits were scheduled8 weeks to 2 weeks before RYGB, and thefinal two visits were scheduled 4 weeksafter RYGB (Fig. 1A). All patients had lap-aroscopic RYGB procedures. After anovernight fast, participants arrived at8:30 A.M. at the research unit. Duringeach visit, two fMRI scans were per-formed: one while the participant wasfasted and one 30 min after intake of astandardized liquid meal. The liquid mealwas consumed over a 25-min interval.The first four participants received200 mL Nutridrink (Nutricia, Zoetermeer,the Netherlands; 300 kcal, carbohydrate37.5 g, fat 11.6 g, and protein 12.0 g) ateach visit (i.e., the two visits before andthe two visits after RYGB). However, sincethese participants reported that thisamount was very difficult to consumeduring the visits after RYGB, the protocolwas adapted during the study. The re-maining six participants received 150 mLduring all test visits. At each visit, a cath-eter was inserted into a cubital vein forinfusion (random order) of either pla-cebo (0.9% sodium chloride solution) orthe GLP-1 receptor antagonist Ex9-39
Figure 1—Study protocol. A: Study design. Ten candidates for RYGB were studied in an acute inter-vention study. All participants underwent four test visits: two before RYGB and two 4weeks after RYGB.During two visits (one before and one after RYGB), the GLP-1 receptor antagonist Ex9-39 was infused inorder to block actions of endogenous GLP-1. During the other visits, only placebo (saline) was infused.B: Test visit. The infusion started 1 h before the beginning of the scan and lasted until the end of thevisit. During each visit, two fMRI scanswere performed: onewhile fasted andone 30minafter intakeof a standardizedmeal. During both the fMRI scans, visual food cueswere presented,whereas a taskwith gustatory food cueswaspresentedonly during the postprandial fMRI scan. Blood samplesweredrawn and sensation of hunger, fullness, and appetitewere scored on a 10-point Likert scale at fixedtime points. Plac, placebo; T1, structural MRI T1 weighted sequence; V1–4, test visits 1–4.
care.diabetesjournals.org ten Kulve and Associates 1523
(Clinalfa; Bachem, Bubendorf, Switzer-land; used to block effects of endogenousGLP-1), using MRI-compatible infusionpumps (MRidium 3850 IV Pump; Ira-dimed, Winter Park, FL). Ex9-39 was di-luted in 0.9% sodium chloride solutioncontaining 0.5% human serum albuminand infused at a rate of 600 pmol/kg/min.A test visit with Ex9-39 infusion was per-formed once before and once after RYGB.In addition, a test visit with placebo in-fusion was performed once before andonce after RYGB. Each infusion started1 h before the start of the MRI and wascontinued during thewholeMRI scanningperiod. The order of infusion was deter-mined by block randomization, and theparticipants were blinded for the type ofinfusion. Blood was drawn at fixed mo-ments to measure GLP-1 and glucose lev-els. Body composition was measuredusing bioelectrical impedance analysis. Asummary of the protocol is presented inFig. 1B.
fMRI TasksAt each visit, a visual food-cue task and agustatory food-cue task were performed.The visual food-cue task was performedboth in the fasted condition and in thepostprandial condition. The gustatoryfood-cue task was performed only in thepostprandial condition (i.e., when endog-enousGLP-1 levels would be at their high-est). All the fMRI tasks were created andpresentedvia the softwareEprime1.2 (Psy-chology Software Tools, Pittsburgh, PA).
Visual Food Cues
Details of this fMRI task have been de-scribed previously (5,13,16). In brief, thefMRI task consisted of pictures selectedfrom three different categories: 1) high-calorie food, 2) low-calorie food, and 3)nonfood items. Pictures were presentedin a block design. In total, 42 pictures percategory were presented, divided in sixblocks of 21 s (Supplementary Fig. 1A).Given that each participant was scannedeight times, eight versions were createdof this paradigm with different pictures,with the images being matched betweenthe versions and between the categoriesfor type and color.
Gustatory Food Cues
Details of this fMRI task have been de-scribed previously (17). Chocolate milkwas used as a palatable food stimulus.As a neutral stimulus, a tasteless solutionwas used, designed to mimic the natural
taste of saliva (consisting of 2.5 mmol/LNaHCO3 and 25 mmol/L KCl) (4). This so-lution should provide a better neutralstimulus thanwater, which has previouslybeen shown to be able to activate thegustatory cortex (18,19). Participants re-ceived 0.4 mL of the chocolate milk ortasteless solution per “trial.” In each trial,participants were presented a picture ofan orange triangle (coupled to chocolatemilk) or a blue star (coupled to taste-less solution), which was followed bythe consumption of the coupled solution.Participants were instructed to keep thesolution within their mouth for 6 s andto refrain from swallowing until thesign “swallow” was presented afterward(Supplementary Fig. 1B).
The taste solutions were deliveredwith two programmable infusion pumps(Infusomat P; B. Braun, Melsungen,Germany) to ensure consistent volumeand timing of the solution delivery.
MRI Acquisition and AnalysesMRI acquisition and analyses have beendescribed previously (5,13,16,17). MRIdata were acquired on a 3.0 Tesla GESigna HDxt scanner (GE Healthcare, Mil-waukee, WI). Functional images were an-alyzed with SPM8 software (WellcomeTrust Centre for Neuroimaging, London,U.K.).
Functional scans were analyzed in thecontext of the general linear model. Forthe visual food-cue task, the high-calorie,low-calorie, and nonfood block were de-fined in the model. Next, to assess CNSactivation related to food cues and, morespecifically, their hedonic quality, wecomputed two contrasts of interest:food .nonfood and high calorie .non-food, which refer to the activity duringviewing food or high-calorie food that isgreater compared with during viewingnonfood pictures. These contrast imageswere entered into three-way ANOVA(5,13,16) with factors surgery (pre-RYGBand post-RYGB), infusion (placebo andEx9-39), and state of feeding (fasted andpostprandial) to assess effects of surgeryand to compare the effect of Ex9-39 ver-sus placebo infusion before and afterRYGB in both meal states. For the gusta-tory food-cue task, the events of the con-sumption of solution were modeled andthe contrast of chocolate milk greaterthan tasteless solution consumption(chocolate .tasteless) was computed.These contrast images were entered
into a separate two-way ANOVA, compa-rable to the visual food-cue task but with-out the factor meal, since the gustatorytask was only performed in the postpran-dial state.
First we explored, using whole brainanalyses, if differences in activation in apriori regions of interest (ROIs) were pre-sent at an uncorrected P, 0.001. A prioriROIs were determined based on previousstudies (i.e., insula [including adjacentopercular cortices], striatum [i.e., puta-men and caudate nucleus], amygdala,and orbitofrontal cortex [OFC]), as theseregions are consistently shown to be in-volved in responses to food cues and arepart of the central reward circuits (3–5).CNS activations were reported as signif-icant when these survived family-wiseerror (FWE) correction for multiple com-parisons on the voxel level using smallvolume correction within the predefinedROIs, using 5-mm (for amygdala) or 10-mm(for insula, putamen, caudate nucleus, andOFC) radius spheres as described previ-ously, comparing peak voxel on group level(5,13,16,17).
Blood Sampling and AssaysThe measurement of blood glucose wasperformed using the glucose dehydroge-nasemethod (GlucoseAnalyzer;HemoCue,Angelholm, Sweden). Total GLP-1 wasanalyzed using a C-terminally directedradioimmunoassay for amidated GLP-1(antibody 89390) (20).
QuestionnairesThe participants were asked to score theirsensations of hunger, fullness, prospectivefood consumption, and nausea and theirappetite for sweet, savory, or fat fooditems on a 10-point Likert scale at fourfixed time points during visits: 1) beforestart of the first MRI session, 2) beforeintake of the meal, 3) 30 min after mealintake, and 4) 60 min after meal intake.
Statistical AnalysesClinical group data were analyzed withSPSS version 20. Data are expressed asmean 6 SEM or median [interquartilerange]. Effects of RYGB on clinical charac-teristics were analyzedwith theWilcoxonsigned rank test. To analyze the interac-tion of RYGB and the infusion of Ex9-39,and for the measurements with morethan one time point per visit, repeatedmeasurement analysis was used. Resultswere considered statistically significantwhen P , 0.05.
1524 RYGB and GLP-1 on CNS Responses to Food Cues Diabetes Care Volume 40, November 2017
Clinical CharacteristicsClinical characteristics before and afterRYGB are presented in SupplementaryTable 1. After RYGB, body weight was re-duced significantly (mean 6 SD, 28.8 61.7 kg, P = 0.005). Additionally, waist cir-cumference, body fat mass, and leanmass were significantly reduced afterRYGB (P# 0.007).Supplementary Fig. 2 shows the GLP-1
and glucose levels during the differentvisits. After RYGB, GLP-1 levels were sig-nificantly higher compared with beforesurgery (P , 0.001), but the levels didnot differ significantly while patientswere fasted (P = 0.3). GLP-1 levels alsodid not differ significantly between par-ticipants receiving 200 mL of the stan-dardized liquid meal compared with theparticipants receivingonly 150mL (beforeRYGB, P = 0.2; after RYGB, P = 0.6). DuringEx9-39 infusion, GLP-1 levels were signif-icantly higher compared with placebo,both before and after RYGB (P , 0.001),but GLP-1 levels were not significantly af-fected by Ex9-39 infusion while patientswere fasted (before RYGB, P = 0.8; afterRYGB, P = 0.1). The effect of Ex9-39 in-fusion on GLP-1 levels was larger afterRYGB compared with before surgery (in-teraction P = 0.05). Glucose levels alsodiffered significantly after RYGB com-pared with before (P , 0.001, duringplacebo infusion), but not while fasted(P = 0.3). Glucose levels were higher dur-ing Ex9-39 compared with placebo infu-sion, both before and after RYGB (P ,0.001), and this effect of Ex9-39 wasalso observed while patients were fasted(before RYGB, P, 0.001; after RYGB, P,0.004). However, no significant interac-tion of RYGB with Ex9-39 infusion wasobserved (P = 0.5).
RYGB Reduces CNS Activationin Response to Visual and GustatoryFood CuesWe first investigated if RYGB resulted in adifference in CNS activation in responseto food cues (i.e., to visual and gustatoryfood cues). We compared CNS activationduring placebo infusion before and afterRYGB. A detailed overviewof the results ispresented in Table 1.
Visual Food Cues
In the fasted condition during placeboinfusion, RYGB resulted in lower activa-tion in response to viewing food pic-tures in the left caudate nucleus and
Tab
le1—Effe
ctsofRYGBsu
rgery
andGLP
-1rece
ptorblocka
dein
response
tovisu
alandgusta
tory
foodcu
es
Used
contrast
Compariso
nRegio
nSid
eCluster
Max
tval
P-FW
EMNIcoordinates
(x,y,z)
Visualfood
cues:effectsof
RYG
B
Food.nonfood
Effectsin
fastedstate:
pre-R
YGB
.po
st-RYG
B(bo
thplacebo)
Caudate
nucleu
sL
183.15
0.03215,23,2
2Rolandic
operculum
R13
3.110.03
54,24,10
High
calorie.nonfood
Caudate
nucleu
sL
193.11
0.03213,23,2
2OFC
L13
3.130.03
233,47,2
8Rolandic
operculum
R9
2.640.09
48,21,10
Food.nonfood
Effectsinpostp
randialstate:pre-R
YGB
.post-R
YGB(both
placebo)
dd
dd
dd
High
calorie.nonfood
dd
dd
dd
Visualfood
cues:effectsof
GLP-1
Rblockad
e3
RYG
B
Food.nonfood
Fastedstate:
effectofGLP-1
post-R
YGB.pre-R
YGB
Caudate
nucleu
sL
263.34
0.0223,14,2
2
High
calorie.nonfood
Caudate
nucleu
sL
53.02
0.0826,20,1
Food.nonfood
Postp
randialstate:
effectofGLP-1
post-R
YGB.pre-R
YGB
dd
dd
dd
High
calorie.nonfood
dd
dd
dd
Gustatory
foodcues:effects
ofRYG
B
Cho
colate
.tasteless
Pre-RYG
B.post-R
YGB(bo
thplaceb
o)Insula
R52
4.270.003
51,2,28
Gustatory
foodcues:effects
ofGLP-1
Rblockade
3RYG
B
Cho
colate
.tasteless
(Ex9-39.placebo
)Effect
GLP-1
blockad
e:post-R
YGB.pre-R
YGB
Insula
R59
4.420.002
48,27,2
11
Thistab
ledescrib
esthe
areaswhere
significantdifferences
inactivations
were
observedfor
thedifferent
compariso
ns.Firstwedescribe
theeffect
ofRYG
B,i.e.,the
differenceinCNSresponses
before
andafter
RYG
B.Second,w
edescrib
ethe
effectof
RYG
Bonthe
effectofGLP-1
Rblockade
(GLP-1
Rblockade
3RYG
B[before
versusafter]).For
eachcom
parison,thetw
ocontrasts
forthe
visualfoo
dcues
(activationdu
ringfood
.nonfood
picturesand
high-calorie
food.nonfood
pictures)andthe
contrastfo
rthegustatory
foodcues
(activationduring
chocolate
.tasteless
solution)arepresented
.Theareas
with
significantdifferencesare
listed,includ
ingthe
clustersize
ofthiseffect,the
tvalu
e,andthe
FWE-corrected
Pvalu
eafter
smallvolu
mecorrection
(P-FWE).The
lastcolum
ndescribes
thecoordinates
ofthepeak
voxelofthe
observed
differenceinMNI
space.Forastepw
iseinterp
retationofthe
resultsdescrib
edinthis
table,pleasesee
RESU
LTS.GLP-1
R,G
LP-1receptor;L,left;M
NI,M
ontrealNeurologicalInstitute;R
,right;val,value.
care.diabetesjournals.org ten Kulve and Associates 1525
right rolandic operculum (cluster size =18, t value = 3.15, P = 0.03 and clustersize = 13, t value = 3.11, P = 0.03, respec-tively). In addition, the activation in re-sponse to high-calorie pictures wasdecreased after RYGB in the left caudatenucleus (cluster size = 19, t value = 3.11,
P = 0.03), right rolandic operculum (clus-ter size = 9, t value = 2.64, P = 0.09), andleft OFC (cluster size = 13, t value = 3.13,P = 0.03) (effect of RYGB, visual foodcues) (Fig. 2A). No significant effects ofRYGB were observed in the postprandialcondition.
Gustatory Food Cues
In the postprandial condition during pla-cebo infusion, RYGB resulted in decreasedCNS activation in response to the gusta-tory food cues (i.e., activation duringchocolate milk consumption) in the rightinsula (cluster size = 52, t value = 4.27,P = 0.003) (effect of RYGB, gustatory foodcues) (Fig. 2B).
Effects of GLP-1 Receptor BlockadeAfter RYGB Are Larger Versus BeforeRYGBSecond, we investigated if endogenousGLP-1 contributed to the effects of RYGBonCNSactivation in response to food cues(described above).We compared theeffect of Ex9-39 infusion versus placeboinfusion before and after RYGB on CNSactivation during the different food-cuetasks. A detailed overview of the resultsis presented in Supplementary Table 1.
Visual Food Cues
In the fasted condition, GLP-1 receptorblockade with Ex9-39 infusion resultedin a larger increase after RYGB than beforesurgery in activation in the left caudate nu-cleus in response to both food pictures andhigh-calorie food pictures (cluster size = 23,t value = 3.34, P = 0.02 and cluster size = 5,t value = 3.02, P = 0.08, respectively) (ef-fect of GLP-1 receptor blockade3 RYGB,visual food cues) (Fig. 3A). In the post-prandial condition, we did not observeany effect of Ex9-39 administration afterRYGB compared with before RYGB.
Gustatory Food Cues
In the postprandial condition, comparingGLP-1 receptor blockade before and afterRYGB, the effect of Ex9-39 was signifi-cantly larger after RYGB in the right insula(cluster size = 59, t value = 4.42, P = 0.002)(effect of GLP-1 receptor blockade 3RYGB, gustatory food cues) (Fig. 3B).
Appetite-Related ScoresRYGB significantly decreased feelings ofhunger and prospective food consump-tion (P , 0.001) during placebo infusion(Supplementary Fig. 3). Appetite forsweet and savory food items was also re-duced after RYGB (P = 0.001, P = 0.006,and P = 0.003, respectively). Feelings ofnausea were increased after RYGB (P ,0.001), but no differences in sensation offullness were observed (P = 0.3). The ef-fects of GLP-1 receptor blockade on visualanalog scale (VAS) score before and afterRYGB were not significantly different(Supplementary Fig. 3).
Figure 2—Effects of RYGB on CNS activation in response to visual (A) and gustatory (B) food cues.Coronal and axial slices showing the difference between the group averages for the 10 participantsregarding activation in areas of the CNS where activation in response to viewing food pictures wasdecreased after RYGB compared with before (A) and activation in response to chocolate milkconsumption was lower after RYGB compared with before (B). The color scale reflects the t valueof the functional activity. Results are presented at the threshold of P , 0.05, FWE corrected(correction for multiple comparisons on the voxel level) on cluster extent. In the graphs, bold signalintensity is plotted (arbitrary units [a.u.]); mean and SEM are shown.
1526 RYGB and GLP-1 on CNS Responses to Food Cues Diabetes Care Volume 40, November 2017
Adverse EventsDuring the visits after RYGB, two patientscomplained about nausea shortly after in-take of the meal during placebo. Two pa-tients experienced periods of dizzinessand palpitation lasting approximately10 min after intake of the meal duringthe visit after RYGB with placebo. Whenweexcluded these patients from the anal-yses, thefindings in thepostprandial stateon CNS activation remained similar. Onepatient had diarrhea shortly after intakeof the meal on both visits after RYGB.
CONCLUSIONS
In the current study, we investigated theeffects of RYGB on CNS activation in re-sponse to food cues,measuredwith fMRI.
In addition, we evaluated the contri-bution of changes in GLP-1 levels afterRYGB to these central effects. We foundthat RYGB reduced the responsivity in ourpredefined ROIs of the CNS (involved inreward and satiety circuits) to both visualand gustatory food cues. Using the GLP-1receptor antagonist Ex9-39, we also ob-served that the effects of endogenousGLP-1 onCNS responses to both the view-ing of food pictures and the consumptionof palatable food were larger after RYGBcompared with before. These findings in-dicate that the effects of RYGBon the CNSare at least partly explained by postoper-ative changes in endogenous GLP-1.
RYGB is known for its substantial asso-ciated weight loss, which is maintained in
the long term (1). Postoperative neuroen-docrine alterations are suggested to playan important role in these effects of RYGB(21). Decreased activations in areas thatare part of the central reward circuits(among others, ventral striatum and puta-men) in response to visual food cues afterRYGB have been described (6,7). RYGB isalso associated with a deceased desire toeat highly palatable food (6,22). In accor-dancewith these studies, we observed de-creased CNS activation in response to bothvisual and gustatory food cues after RYGB,paralleled by decreased scores for hungerand appetite. We also observed increasedfeelings of nausea after RYGB. As hungerand nausea feelingsmay be related, it couldbehypothesized that thedecrease inhungeris due to an increase in nausea. However,the increase in nausea after RYGB is mainlypresent in the postprandial state (not in thefasted state), whereas the decrease in hun-ger was also present in the fasted state.Wethereforebelieve thatRYGBhasaneffect onhunger independent of increase in nausea.
In the current study, we focused on therole of enhanced postoperative GLP-1 inthe decreased CNS responses to foodcues after RYGB. GLP-1 and treatmentwith GLP-1 receptor agonists reducefood intake and body weight (11) via ef-fects in the CNS (5,12,13,16,17). In thecurrent study, we observed that the ef-fect of endogenous GLP-1 on responsivityin the caudate nucleus to viewing foodpictures was larger after RYGB in thefasted state. In the postprandial condi-tion, we found a larger effect of GLP-1on responsivity in the insula to the con-sumption of palatable food after RYGB,although responses to viewing food pic-tures after RYGB were not affected byGLP-1. The fact that we only found effectsin the postprandial condition on gusta-tory food cues suggests a larger role forGLP-1 in the central rewarding evaluationof taste perception than in the evaluationof visual food cues. Interestingly, receptorsfor GLP-1 were reported to be present inmammalian taste buds, and GLP-1 recep-tor knockout mice were shown to havereduced sweet taste sensitivity, pointingtoward an important role forGLP-1 in tasteperception in rodents (23). It is howeverunknown whether this mechanism is alsooperative in humans.
As expected, GLP-1 levels were higherafter RYGB, which may be related to rapidentry and absorption of nutrients to themore distal small intestine postoperatively
Figure 3—Effects of GLP-1 blockade (using Ex9-39), comparing before versus after RYGB, on CNSactivation in response to visual (A) and gustatory (B) food cues. Axial slices showing the difference inthe group averages in activation in areas of the CNS, depicting the difference of the effect of GLP-1blockade by infusion of Ex9-39 (versus placebo). A: GLP-1 receptor blockade resulted in a largerincrease in activation in response to viewing food pictures after RYGB compared with before. B: Acomparable effect in response to chocolatemilk consumption. The color scale reflects the t value ofthe functional activity. Results are presented at the threshold ofP, 0.05, FWE corrected (correctionfor multiple comparisons on the voxel level) on cluster extent. In the graphs, bold signal intensity isplotted (arbitrary units [a.u.]); mean and SEM are shown.
care.diabetesjournals.org ten Kulve and Associates 1527
(24), which may stimulate an enhancedrelease of GLP-1. In addition, an increaseddensity of GLP-1–immunoreactive cellshas been observed after RYGB (25). Wedemonstrated that the enhanced GLP-1secretion may explain the decreasedCNS activation in response to the con-sumption of palatable food after RYGB.Noteworthy, although fasting GLP-1 lev-els were not significantly altered afterRYGB, the effects of endogenous GLP-1on responses to viewing food pictures inthe fasted condition were increased. Itcould be speculated that this might bedue to an increase in GLP-1 sensitivity,as suggested by a study in rats, showingthat administration of GLP-1 receptor ag-onist exendin-4 decreased food intakemore in RYGB than in sham-operatedrats, indicating a higher sensitivity toGLP-1 after RYGB (26). In accordance,lower BMI in humans was correlatedwith an increased incretin effect (27).This observational finding is compatiblewith the hypothesis that reductions inBMI may enhance effects of and sensitiv-ity to incretin hormones, such as GLP-1.An increased sensitivity after RYGB hasalso been described for other hormones,i.e., insulin and thyroxin, independent ofweight reduction (28,29). Although itcould be considered contradictory to theoften observed effect that increased hor-monal levels leads to desensitization ofthe corresponding hormonal receptor,as postprandial GLP-1 levels are increasedafter RYGB, we do speculate that in-creased sensitivity for GLP-1 may explainour findings in the fasted state.Although previous studies have inves-
tigated the effect of RYGB on CNS re-sponses to the viewing of food pictures,only one recent pilot study has investi-gated CNS activation in response to sweettaste after RYGB in humans, showing asignificant decrease in activation in re-sponse to sweet taste in the OFC aftersurgery (30). However, this finding wasnot conclusive, as this effect was also ob-served in control subjects. We also foundthat RYGB decreased the CNS responsesin the insula in response to chocolatemilkconsumption, which was accompanied byweight reduction. At first sight, this find-ing may be considered to be at odds withprevious studies, as several (4,17), but notall (31,32), previous studies demon-strated that leaner individuals have in-creased responsivity to the consumptionof chocolate milk in comparison with
obese individuals. However, in general,both lean and obese individuals are pre-sumed to “like” the palatable gustatoryfood cue but seem to differ in the centralresponses andprocess of the reward eval-uation of this cue. In contrast, RYGB isassociated with changes in food prefer-ences (22) and taste perception (33),with higher susceptibility for sweet tasteperception (34,35). Studies reported thatpatients after RYGB have decreased inter-est in sweet food, finding it less enjoyableor even unpleasant (34). Therefore, the“liking” of the chocolate milk consump-tion in our current study may be alteredpostoperatively, and chocolate milk mayeven be experienced as unpleasant. Ac-cording to this, we observed a significantdecrease in the appetite-related scores forsweet food items after RYGB. This may ex-plain the decreased responsivity of the in-sula to the consumption of chocolate milkobserved after RYGB in our study. In linewith this, we found that blockade of endog-enous GLP-1 effects after RYGB increasedthe CNS activations in response to cho-colate milk consumption. These increasedCNS activations may be interpreted as in-creased liking of chocolate milk, suggestingthatendogenousGLP-1decreased the likingof sweet taste, which may lead to reducedsweet palatable food consumption.
GLP-1 may affect the CNS directly viaaccess through areas with a permeableblood-brain barrier or via secretion byGLP-1–producing neurons (36). However,central effects of GLP-1 may also be in-directly mediated via activation of vagalafferents. In our current study, we usedEx9-39, which is able to cross the blood-brain barrier (37). We are therefore notable to distinguish if the observed effectsof GLP-1 in our study are mediated di-rectly or also partly indirectly.
In our current study, we focused onthe effects of the exaggerated GLP-1 re-sponse after RYGB on the CNS. It could besuggested that changes in the levels ofother hormones, such as insulin, mayalso play a role. However, we do not be-lieve that insulin levels can explain ourfindings for several reasons. First, in ourprevious study we did not find a signifi-cant difference in insulin levels betweenplacebo and Ex9-39 infusion (13). Second,we previously demonstrated, using a pan-creatic clamp, that the effects of the GLP-1receptor agonist on the CNS responses tofood stimuli were independent of changesin insulin levels (5). Third, although others
have shown that insulin levels may in-crease at 3 months after RYGB, this wasnot observed 1 week after RYGB (38). Italso could be suggested that changes inglucose levels after RYGB may affect ourobserved findings. However, the effect ofEx9-39 on glucose level did not differ sig-nificantly before, compared with after,RYGB. In addition, we have demonstratedin our previous studies that the effects ofGLP-1 on the brain are independent ofglucose and/or insulin levels (5,13).
The sample size of the current studyis relatively small. However, we used alongitudinal, within-subjects design with.90% power to detect the expected dif-ference in CNS activation (5,6,13,39). Itshould, however, be emphasized thatthis was a pilot study with only femalepatients between the ages of 40 and65 years, which limits the generalizabilitytomen and other age-groups. In addition,we investigated patients 4 weeks aftersurgery, comparable to previous studies(6,39). However, in this phase after sur-gery, patientsmay still have complaints ofthe intestinal anastomoses and may haveproblemswith a number of food products,which they can tolerate more than a yearafter surgery. Others have found reducedCNS responses several years after RYGB(7,40), but further research is needed todetermine the role for GLP-1 in theselonger-term CNS changes.
In conclusion, similar to previous stud-ies, we found that the effects of RYGB onfood intake may be mediated by de-creased activation in feeding regulatingareas in the CNS in response to food stim-uli. In addition, our findings using theGLP-1 receptor antagonist suggest thatthese effects of RYGB may be partly ex-plained by postoperative changes in thelevels of endogenousGLP-1and/orpossiblechanges in sensitivity toGLP-1. These find-ings provide further insights in theweight-lowering mechanisms of RYGB and mayultimately lead to further developmentof treatment strategies for obesity.
Acknowledgments. The authors thank SandraGassman (Department of Internal Medicine, VUUniversityMedical Center) and Ton Schweigmann(Department of Radiology and Nuclear Medi-cine, VU University Medical Center) for their assis-tance during the test visits, as well as the subjectswho participated in this study. The authors thankNutricia for providing the Nutridrink.Funding. R.G.I. is financed by the NetherlandsOrganisation for Scientific Research (NWO)
1528 RYGB and GLP-1 on CNS Responses to Food Cues Diabetes Care Volume 40, November 2017
Innovational Research Incentives Scheme Veni(no. 91613082).DualityofInterest.C.F.D.hasreceivedconsultancy/speaker fees from Bristol-Myers Squibb, BoehringerIngelheim, Eli Lilly and Company, Merck Sharp &Dohme, Novartis, and Novo Nordisk. J.J.H. hasreceived fees for consulting, lecturing, and/orbeing part of an advisory board fromAstraZeneca,Boehringer Ingelheim, Bristol-Myers Squibb, EliLilly and Company, GI Dynamics, Merck Sharp &Dohme, Novo Nordisk, Novartis, Sanofi, Takeda,and Zealand Pharma. M.D. was a consultantfor Abbott, AstraZeneca, Bristol-Myers Squibb,Boehringer Ingelheim, Eli Lilly and Company, GIDynamics, Merck Sharp & Dohme, Novo Nordisk,Poxel SA, and Sanofi andwas a speaker for Bristol-Myers Squibb/AstraZeneca, Eli Lilly and Company,Novo Nordisk, and Sanofi. Through M.D., the VUUniversity Medical Center received researchgrants from Abbott, Bristol-Myers Squibb/AstraZeneca, Boehringer Ingelheim, Eli Lilly andCompany, Medtronic, Merck Sharp & Dohme,Novo Nordisk, and Sanofi. R.G.I. is the principalinvestigator of studies sponsored by research grantsfrom Novo Nordisk and Eli Lilly and Company. M.D.and R.G.I. report receiving no personal payments inconnection to the above-mentioned activities, but allpayments were directly transferred to the DiabetesCenter (VU University Medical Center) nonprofit re-search foundation. No other potential conflicts of in-terest relevant to this article were reported.Author Contributions. J.S.t.K. designed thestudy, conducted the experiments, designed thefMRIparadigm,performeddataanalysis, andwrotethemanuscript. D.J.V. designed the fMRI paradigm,performeddataanalysis, andwrote themanuscript.V.E.A.G. contributed to the design and the perfor-mance of the study and contributed to writing themanuscript. L.v.B. designed the fMRI paradigm andcontributed to writing the manuscript. F.B. performedanalyses of all structural MRI scans and contributed towriting the manuscript. C.F.D. and J.J.H. performedlaboratory analyses and contributed to writing themanuscript. M.L.D. contributed to the design of thestudy and to writing themanuscript. M.D. designedthe study.R.G.I. designed the study, performeddataanalysis,andwrotethemanuscript.Allauthorshaveseen and approved the final version of the man-uscript. J.S.t.K. and R.G.I. are the guarantors of thiswork and, as such, had full access to all the data inthe studyand take responsibility for the integrity ofthe data and the accuracy of the data analysis.
References1. Buchwald H, Avidor Y, Braunwald E, et al. Bari-atric surgery: a systematic reviewandmeta-analysis.JAMA 2004;292:1724–17372. Rao RS. Bariatric surgery and the central ner-vous system. Obes Surg 2012;22:967–9783. Stoeckel LE,Weller RE, Cook EW3rd, TwiegDB,Knowlton RC, Cox JE. Widespread reward-systemactivation inobesewomen in response topicturesofhigh-calorie foods. Neuroimage 2008;41:636–6474. Stice E, Spoor S, Bohon C, Small DM. Relationbetween obesity and blunted striatal response tofood is moderated by TaqIA A1 allele. Science2008;322:449–4525. van Bloemendaal L, IJzerman RG, ten Kulve JS,et al. GLP-1 receptor activationmodulates appetite-and reward-related brain areas in humans. Diabe-tes 2014;63:4186–4196
6. Ochner CN, Kwok Y, Conceição E, et al. Selec-tive reduction in neural responses to high caloriefoods following gastric bypass surgery. Ann Surg2011;253:502–5077. Frank S, Wilms B, Veit R, et al. Altered brainactivity in severely obese women may recoverafter Roux-en Y gastric bypass surgery. Int JObes 2014;38:341–3488. Murphy KG, Bloom SR. Gut hormones and theregulation of energy homeostasis. Nature 2006;444:854–8599. Borg CM, le Roux CW, Ghatei MA, Bloom SR,Patel AG, Aylwin SJ. Progressive rise in gut hor-mone levels after Roux-en-Y gastric bypass sug-gests gut adaptation and explains altered satiety.Br J Surg 2006;93:210–21510. Falken Y, Hellstrom PM, Holst JJ, Naslund E.Changes inglucosehomeostasis afterRoux-en-Ygas-tric bypass surgery for obesity at day three, twomonths, and one year after surgery: role of gut pep-tides. J Clin EndocrinolMetab 2011;96:2227–223511. Vilsbøll T, ChristensenM, Junker AE, Knop FK,Gluud LL. Effects of glucagon-like peptide-1 recep-tor agonists onweight loss: systematic review andmeta-analyses of randomised controlled trials.BMJ 2012;344:d777112. Turton MD, O’Shea D, Gunn I, et al. A role forglucagon-like peptide-1 in the central regulationof feeding. Nature 1996;379:69–7213. ten Kulve JS, Veltman DJ, van Bloemendaal L,et al. Endogenous GLP-1 mediates postprandialreductions in activation in central reward and sa-tiety areas in patients with type 2 diabetes. Dia-betologia 2015;58:2688–269814. Tadross JA, le Roux CW. The mechanisms ofweight loss after bariatric surgery. Int J Obes2009;33(Suppl. 1):S28–S3215. Madsbad S, Dirksen C, Holst JJ. Mechanismsof changes in glucose metabolism and body-weight after bariatric surgery. Lancet Diabetes En-docrinol 2014;2:152–16416. ten Kulve JS, Veltman DJ, van Bloemendaal L,et al. Liraglutide reduces CNS activation in re-sponse to visual food cues only after short-termtreatment in patients with type 2 diabetes. Dia-betes Care 2016;39:214–22117. van Bloemendaal L, Veltman DJ, ten Kulve JS,et al. Brain reward-system activation in response toanticipation and consumption of palatable food is al-teredbyglucagon-likepeptide-1 receptoractivation inhumans. Diabetes Obes Metab 2015;17:878–88618. O’Doherty J, Rolls ET, Francis S, Bowtell R,McGlone F. Representation of pleasant and aver-sive taste in the human brain. J Neurophysiol2001;85:1315–132119. Zald DH, Pardo JV. Cortical activation inducedby intraoral stimulation with water in humans.Chem Senses 2000;25:267–27520. Orskov C, Rabenhøj L,WettergrenA, Kofod H,Holst JJ. Tissue and plasma concentrations of ami-dated and glycine-extended glucagon-like peptide Iin humans. Diabetes 1994;43:535–53921. BerthoudHR,Morrison C. The brain, appetite,and obesity. Annu Rev Psychol 2008;59:55–9222. Halmi KA, Mason E, Falk JR, Stunkard A. Ap-petitive behavior after gastric bypass for obesity.Int J Obes 1981;5:457–46423. Shin YK,Martin B, Golden E, et al.Modulationof taste sensitivity by GLP-1 signaling. J Neuro-chem 2008;106:455–46324. Jacobsen SH, Bojsen-Møller KN, Dirksen C,et al. Effects of gastric bypass surgery on glucose
absorption and metabolism during a mixed mealin glucose-tolerant individuals. Diabetologia 2013;56:2250–225425. Rhee NA, Wahlgren CD, Pedersen J, et al. Ef-fect of Roux-en-Y gastric bypass on the distribu-tion and hormone expression of small-intestinalenteroendocrine cells in obese patients withtype 2 diabetes. Diabetologia 2015;58:2254–225826. Abegg K, Schiesser M, Lutz TA, Bueter M.Acute peripheral GLP-1 receptor agonism or an-tagonism does not alter energy expenditure inrats after Roux-en-Y gastric bypass. Physiol Behav2013;121:70–7827. Muscelli E, Mari A, Casolaro A, et al. Separateimpact of obesity and glucose tolerance on theincretin effect in normal subjects and type 2 di-abetic patients. Diabetes 2008;57:1340–134828. Moulin deMoraes CM,Mancini MC, deMeloME, et al. Prevalence of subclinical hypothyroid-ism in a morbidly obese population and improve-ment after weight loss induced by Roux-en-Ygastric bypass. Obes Surg 2005;15:1287–129129. Wickremesekera K, Miller G, Naotunne TD,Knowles G, Stubbs RS. Loss of insulin resistanceafter Roux-en-Y gastric bypass surgery: a timecourse study. Obes Surg 2005;15:474–48130. Wang JL, Yang Q, Hajnal A, Rogers AM. A pilotfunctional MRI study in Roux-en-Y gastric bypasspatients to study alteration in taste functions aftersurgery. Surg Endosc 2016;30:892–89831. Stice E, Yokum S, Burger KS, Epstein LH, SmallDM. Youth at risk for obesity show greater activa-tion of striatal and somatosensory regions tofood. J Neurosci 2011;31:4360–436632. SzalayC, AradiM, SchwarczA, et al. Gustatoryperception alterations in obesity: an fMRI study.Brain Res 2012;1473:131–14033. Tichansky DS, Boughter JD Jr, Madan AK.Taste change after laparoscopic Roux-en-Y gastricbypass and laparoscopic adjustable gastric band-ing. Surg Obes Relat Dis 2006;2:440–44434. Bueter M, Miras AD, Chichger H, et al. Alter-ations of sucrose preference after Roux-en-Ygastric bypass. Physiol Behav 2011;104:709–72135. Burge JC, Schaumburg JZ, Choban PS,DiSilvestro RA, Flancbaum L. Changes in patients’taste acuity after Roux-en-Y gastric bypass for clin-ically severe obesity. J Am Diet Assoc 1995;95:666–67036. Larsen PJ, Tang-Christensen M, Holst JJ,Orskov C. Distribution of glucagon-like peptide-1and other preproglucagon-derived peptides in therat hypothalamus and brainstem. Neuroscience1997;77:257–27037. During MJ, Cao L, Zuzga DS, et al. Glucagon-like peptide-1 receptor is involved in learning andneuroprotection. Nat Med 2003;9:1173–117938. Jørgensen NB, Jacobsen SH, Dirksen C, et al.Acute and long-term effects of Roux-en-Y gas-tric bypass on glucose metabolism in subjectswith type 2 diabetes and normal glucose toler-ance. Am J Physiol Endocrinol Metab 2012;303:E122–E13139. Ochner CN, Laferrere B, Afifi L, Atalayer D,Geliebter A, Teixeira J. Neural responsivity tofood cues in fasted and fed states pre and postgastric bypass surgery. Neurosci Res 2012;74:138–14340. Scholtz S, Miras AD, Chhina N, et al. Obesepatients after gastric bypass surgery have lowerbrain-hedonic responses to food than after gastricbanding. Gut 2014;63:891–902
care.diabetesjournals.org ten Kulve and Associates 1529
