ORIGINAL RESEARCH
published: 06 Sep embe 2018
doi: 10.3389/ psy .2018.00411
F on ie s in Psychia y | www. on ie sin.o g 1Sep embe 2018 | Volume 9 | A icle 411
Edi ed by:
Valen ina Bassa eo,
Uni e si à degli s udi di Caglia i, I aly
Re iewed by:
Olga Val e de,
Uni e sidad Pompeu Fab a, Spain
Se gueï O. Fe isso ,
Uni e si é de Rouen, F ance
*Co espondence:
Me cè Co ea
[email p o ec ed]
†P esen Add ess:
Lau a López-C uz,
Depa men o Psychology and MRC,
Wellcome T us Beha iou al and
Clinical Neu oscience Ins i u e,
Uni e si y o Camb idge, Camb idge,
Uni ed Kingdom
Special y sec ion:
This a icle was submi ed o
Psychosoma ic Medicine,
a sec ion o he jou nal
F on ie s in Psychia y
Recei ed: 03 May 2018
Accep ed: 13 Augus 2018
Published: 06 Sep embe 2018
Ci a ion:
Co ea M, SanMiguel N,
López-C uz L, Ca a alá-Ros C,
Oli a es-Ga cía R and Salamone JD
(2018) Ca eine Modula es Food
In ake Depending on he Con ex Tha
Gi es Access o Food: Compa ison
Wi h Dopamine Deple ion.
F on . Psychia y 9:411.
doi: 10.3389/ psy .2018.00411
Ca eine Modula es Food In ake
Depending on he Con ex Tha Gi es
Access o Food: Compa ison Wi h
Dopamine Deple ion
Me cè Co ea1,2*, Noemí SanMiguel1, Lau a López-C uz1†, Ca la Ca a alá-Ros1,
Régulo Oli a es-Ga cía1and John D. Salamone2
1À ea de Psicobiologia, Campus de Riu Sec, Uni e si a Jaume I, Cas elló, Spain, 2Beha io al Neu oscience Di ision,
Uni e si y o Connec icu , S o s, CT, Uni ed S a es
Ca eine is a me hylxan hine consumed in di e en con ex s o po en ia e ale ness and
educe a igue. Howe e , ca eine can induce anxie y a high doses. Ca eine is also
a mino psychos imulan ha seems o ac as an appe i e supp essan , bu he e a e
also epo s indica ing ha i could s imula e appe i e. Dopamine also is in ol ed in ood
mo i a ion and in beha io al ac i a ion. In he p esen se ies o expe imen s, we e alua ed
he e ec s o acu e adminis a ion o ca eine on ood consump ion unde di e en access
condi ions. CD1 male adul mice had access o highly pala able ood (50% suc ose)
in a es ic ed bu habi ual con ex , unde con inuous o in e mi en access as well as
unde anxiogenic, o e o ul condi ions. Ca eine (2.5–20.0 mg/kg) inc eased in ake a
he highes dose unde amilia con inuous and in e mi en access. Howe e , his high
dose educed ood in ake in he da k-ligh pa adigm. In con as , a dopamine-deple ing
agen , e abenazine (TBZ; 1.0–8.0 mg/kg) did no a ec ood in ake in any o hose
expe imen al condi ions. In he T-maze-ba ie ask ha e alua es seeking and aking
o ood unde e o ul condi ions, ca eine (10.0 mg/kg) dec eased la ency o each he
ood, bu did no a ec selec ion o he high- ood densi y a m ha equi ed mo e e o ,
o he o al amoun o ood consumed. In con as , TBZ (4.0 mg/kg) educed selec ion
o he high ood densi y a m wi h he ba ie , hus a ec ing amoun o ood consumed.
In e es ingly, a small dose o ca eine (5.0 mg/kg) was able o e e se he ane gia-inducing
e ec s p oduced by TBZ in he T-maze. These esul s sugges ha ca eine can po en ia e
o supp ess ood consump ion depending on he con ex . Mo eo e , ca eine did no
change appe i e, and did no impai o ien a ion owa d ood unde e o ul condi ions,
bu i a he helped o achie e he goal by imp o ing speed and by e e sing pe o mance
o no mal le els when a igue was induced by dopamine deple ion.
Keywo ds: anxie y, appe i e, me hylxan hine, decision-making, suc ose, e abenazine
Co ea e al. Ca eine and Food Consump ion
INTRODUCTION
Ca eine is he psychos imulan subs ance mos widely consumed
in he wo ld, and i is ound in se e al ypes o ood and
be e ages (1). Psychos imulan s a e cha ac e ized by s imula ion
o locomo ion and eelings o in igo a ion and inc ease in
pe cei ed ene gy, al hough a high doses hey can induce
s e eo ypies and anxie y (2–5). This ca ego y o d ugs is known
o i s ano ec ic e ec s, and in ac ca eine is a common
cons i uen in o e - he-coun e weigh -loss supplemen s (6).
Howe e , ca eine in humans does no seem o ha e a consis en
pa e n o e ec s on appe i e and ene gy in ake. While some
s udies epo ha i exe s a sligh ano ec ic e ec (7), o he s do
no epo signi ican changes (8,9).
In animal s udies, he li e a u e shows also a complex pa e n
o esul s. In non-dep i ed mice, a ecen s udy demons a ed
ha acu e adminis a ion o ca eine (6.0–24.0 mg/kg) inc eases
in ake o s anda d ee chow, and ha animals a e no mo e
ac i a ed o anxious (10). In di e en ypes o ope an condi ions
imposed on ood es ic ed a s, acu e doses o ca eine p oduce
dose dependen e ec s; doses o 40.0 mg/kg o highe educed
le e p essing o di e en ypes o ood (11,12), and doses up
o 25.0 mg/kg inc eased le e p essing independen ly i he e was
no ne inc ease on ood access (12–14).
Nucleus accumbens (Nacb) dopamine (DA) has been
implica ed in some ea u es o ood mo i a ion (15–18). Ca eine
does no clea ly induce Nacb DA elease (19,20). Howe e ,
i is a non-selec i e adenosine ecep o an agonis , and hese
ecep o s ha e a unc ional in e ac ion wi h DA ecep o s in
s ia um. Adenosine ecep o s a e co-localized wi h DA ecep o s
con e ging in o he same in acellula cascade in an opposi e
way; while A1 ecep o s a e co-localized wi h D1 ecep o s, A2A
ecep o s a e co-localized wi h D2 ecep o s in di e en g oups
o s ia al neu ons (21–23). Thus, an agonism o adenosine
ecep o s leads o he opposi e e ec o DA an agonism. Se e al
p e ious s udies ha e ocused on he unc ional in e ac ion
be ween adenosine and DA ecep o s in s udies o e o -
ela ed decision-making p ocesses ha allow access o di e en
quan i ies o ypes o ood in ood es ic ed animals (15,24–
26). Fo example, in a T-maze p ocedu e in which mice ha e
o jump a ba ie o ge access o a ela i ely la ge quan i y o
ood, co-adminis a ion o heophylline, ano he me hylxan ine
ha ac s as a non-selec i e adenosine an agonis , e e sed he
ane gia inducing e ec s o a DA D2 ecep o an agonis , es o ing
no mal le els o e o ha lead o mo e ood (26).
In he p esen s udies, we wan ed o assess he impac o
a b oad ange o doses o ca eine (2.5–20.0 mg/kg) on highly
pala able ood in ake in mice. Because anxie y has been shown
o supp ess appe i e, we compa ed he impac o ca eine on
ood in ake unde habi ual condi ions s. anxiogenic condi ions.
Since ca eine in igo a es beha io , we also s udied in ake o
pala able ood when e o was a componen o he decision-
making p ocess, such ha ood es ic ed mice could gain access
o highe quan i ies o ood by exe ion o e o . In addi ion,
we also e alua ed he abili y o ca eine o e e se he e ec s o
a DA-deple ing agen ha educes willingness o wo k o ood
(ane gia). We also assessed i his DA-deple ing agen changes,
by i sel , ood in ake unde habi ual, anxiogenic o e o ul
condi ions. These esul s could cla i y he po en ial he apeu ic
e ec s o ca eine on appe i e unde no mal o pa hological
condi ions.
MATERIALS AND METHODS
Subjec s
CD1 male mice (24–28 g) pu chased om Jan ie , F ance S.A.
we e 4 weeks old upon a i al o he labo a o y (N=47),
and a e a week o adap a ion o he colony condi ions, all
expe imen s s a ed. Mice we e housed in g oups o h ee animals
pe cage wi h ap wa e and s anda d chow ood a ailable ad
libi um in he home cage ac oss he en i e expe imen , excep
in expe imen s 5–7, in which mice we e ood- es ic ed in hei
home cage ( o a maximum o 85% ee eeding body weigh )
h oughou he s udy. These animals ecei ed be ween 7 and
8 g o s anda d chow ood pe cage du ing he week days and
be ween 13 and 14 g du ing he weekend days o allow no mal
g ow h. The colony was kep a a empe a u e o 22 ±2◦C wi h
ligh s on om 08:00 o 20:00 h. All animals we e co e ed by
a p o ocol app o ed by he Ins i u ional Animal Ca e and Use
commi ee o Uni e si a Jaume I. All expe imen al p ocedu es
complied wi h di ec i e 2010/63/EU o he Eu opean Pa liamen
and o he Council, and wi h he “Guidelines o he Ca e and Use
o Mammals in Neu oscience and Beha io al Resea ch,” Na ional
Resea ch Council 2003, USA. All e o s we e made o minimize
animal su e ing, and o educe he numbe o animals used.
Pha macological Agen s
Ca eine (1,3,7- ime hylxan hine; Sigma-Ald ich, Spain) was
dissol ed in 0.9% w/ saline and was adminis e ed 30 min be o e
es ing. Saline solu ion was used as i s ehicle con ol. The ange
o ca eine doses (2.5, 5.0, 10.0, and 20.0 mg/kg) was selec ed
based on p e ious and pilo s udies (27). Te abenazine (TBZ)
[(R,R)-3-Isobu yl-9,10-dime hoxy-1,3,4,6,7,11b-hexahyd o-
py ido[2,1-a]isoquinolin-2-one] (CIMYT Quimica SL, Spain),
adminis e ed 120 min be o e es ing, was dissol ed in a ehicle
solu ion o 0.9% saline (80%) and dime hylsul oxide (DMSO;
20%) (pH =4.5). DMSO was used as ehicle con ol. All
solu ions we e adminis e ed in ape i oneally (IP).
Appa a us and Tes ing P ocedu es
The same ype o ood was used in all he expe imen s; 45 mg
(expe imen s 1–4) o 20 mg (expe imen s 5–7) p ecision pelle s
o oden s (Tes Die TM) wi h a balanced nu ien composi ion
and a 50% suc ose con en ha ga e i pala able cha ac e is ics.
Pala able Food Consump ion Unde Habi ual
Condi ions: Con inuous o In e mi en Access
Du ing 4 weeks (5 days pe week) non- ood es ic ed mice
we e placed indi idually in s anda d home cages whe e hey
had ad libi um access o highly pala able pelle s (45 mg each).
Baseline sessions las ed 60 min (da a we e egis e ed e e y
30 min), s a ing 3 h p io he beginning o he da k cycle.
Thus, animals consumed ood in a amilia and epe i i e
condi ion. Du ing 2 mo e weeks, animals we e habi ua ed o
F on ie s in Psychia y | www. on ie sin.o g 2Sep embe 2018 | Volume 9 | A icle 411
Co ea e al. Ca eine and Food Consump ion
ecei e an IP ehicle injec ion once a week, and sessions
las ed 30 min (see Figu e 1). Because i has been sugges ed
ha non-con inuous access o ood inc eases ood consump ion
leading o binge ea ing, we had wo condi ions o ood in
a amilia con ex : he con inuous g oup had sessions 5 days
pe week (Monday o F iday), and he in e mi en access
g oup had sessions 3 days pe week (Monday, Wednesday
and F iday). The es phase las ed i e mo e weeks du ing
which each subjec ecei ed all doses o ca eine (Expe imen s
1 and 2) o TBZ (Expe imen s 3 and 4) in a andomly a ied
o de , once a week. Body weigh was egis e ed 3 imes pe
week.
Pala able Food Consump ion Unde Anxiogenic
Condi ions
A e comple ing he habi ual ood-consump ion expe imen s,
he same mice had access o he highly pala able ood o h ee
addi ional weeks o baseline wi h no ea men , and sho e
sessions (15 min) (see Figu e 1). Then, mice in he con inuous
and in he in e mi en condi ions we e di ided in wo ea men
g oups: saline o 20 mg/kg o ca eine (Expe imen 2), and DMSO
o TBZ (8 mg/kg; Expe imen 4). Animals ha had ecei ed
ca eine in he p e ious expe imen also ecei ed ca eine in
he p esen one, and he same was ue o animals ha had
ecei ed TBZ. Doses we e selec ed based on hei impac in he
p e ious expe imen s. In one single es day, animals we e placed
o 15 min in a da k and ligh (DL) pa adigm. In he DL box,
one chambe was enclosed and da k, and he open chambe
whe e he ood dish was placed, was in ensely illumina ed. Tes s
s a ed when each subjec was placed in he da k chambe . Since
he e we e no di e ences in ood in ake in he DL be ween
animals in he in e mi en and he con inuous g oups da a
om bo h g oups we e pooled in o de o inc ease he numbe
o animals pe g oup. The amoun o ood (mg) consumed
du ing 15 min was eco ded, as well as he ime spen in he li
compa men .
E o -Based Decision-Making o Pala able Food in a
T-Maze Wi h Ba ie
In he las g oup o expe imen s (5–7), we s udied he impac o
ca eine o TBZ in he T-maze p ocedu e ha imposes an e o
es ic ion in o de o ge access o highe quan i ies o ood. In
o de o make animals o lea n and pe o m he ask, mice we e
ood es ic ed in hei home cage. This p ocedu e is based on
p e ious published wo k (26). The T-maze appa a us consis ed
o a cen al co ido wi h wo opposed a ms (see Figu e 4). Each
a m p o ided a di e en densi y o ood: 2 pelle s (20 mg each)
in he high densi y (HD) a m, and 1 pelle (20 mg) in he low
densi y (LD) a m. Pelle s we e loca ed in dishes placed nea he
a walls o he maze a ms. The HD a m con ained a e ical
ba ie (12 cm high) ha p o ided he e o - ela ed challenge.
Hal he mice had he HD a m wi h he ba ie consis en ly
loca ed on he le side, while hal he mice had he HD a m
and ba ie on he igh side. Du ing he i s aining phase
no ba ie was p esen , and o he i s 2 days o he ini ial
aining, mice had ee access o bo h a ms o he T-maze upon
exi ing he s a a m, and we e allowed o consume all pelle s
in bo h HD and LD a ms o he maze be o e being e u ned
o he s a a m. Upon comple ion o his ini ial aining, mice
we e only allowed o choose one a m o he maze; a e he
ini ial a m choice, he o he a m was blocked. Du ing 2 weeks
mice choose be ween he wo a ms wi h no ba ie in place.
In he second aining phase, a small ba ie (6 cm high) was
in oduced in he HD a m o 1 week. Animals we e hen
ained wi h he 12 cm ba ie in he HD a m o he es o
he sessions. A aining phase o 3 weeks was allowed be o e he
d ugs we e adminis e ed. The es phase s a ed a ha momen :
in expe imen s 5 and 6 animals ecei ed a single injec ion pe
week o ca eine (10 mg/kg) o saline ehicle, and o TBZ (4.0
mg/kg) o DMSO ehicle. These lowe doses we e chosen based
on p e ious published da a om ou labo a o y (27,28) on he
impac o ca eine and TBZ on unning wheel ac i i y, an ac i i y
ha is highly ein o cing and equi es a lo o igo . Thus, TBZ
4.0 mg/kg and ca eine 10 mg/kg did no educed unning, bu
TBZ 8 mg/kg did (28). In addi ion, a pilo s udy demons a ed
ha ca eine (20.0 mg/kg) also educed unning wheel ac i i y
( - es o independen samples [ (9) =6.83; p<0.01], Saline =
1,982.3 ±128.1, Ca eine 20.0 mg/kg =1,309 ±116.5). Based
on hese esul s, o expe imen s 5–7 we used lowe doses o
bo h d ugs. In expe imen 7 a di e en g oup o animals ollowed
he same p o ocol and he pha macological in e ac ions be ween
ca eine and TBZ we e s udied.
Da a Analyses
All expe imen s bu 2 and 4 used a wi hin-g oups design.
No mally dis ibu ed and homogenous da a o m expe imen s
1 (A and B), 3 (A and B), and 7 we e e alua ed by epea ed
measu es analysis o a iance (ANOVA). Fu he analyses we e
conduc ed by non-o hogonal planned compa isons using he
o e all e o e m o assess di e ences be ween each dose and he
con ol condi ion (29) ( he numbe o compa isons was es ic ed
o he numbe o ea men s minus one). T- es s o dependen
samples analysis we e used o expe imen s 5 and 6. Expe imen s
2 and 4 used a be ween g oups design and da a we e analyzed
by - es s o independen samples. Addi ionally, compa ison
o body weigh p og ession was analyzed independen ly wi h
epea ed measu es ANOVA o e e y g oup o animals in all hese
expe imen s, plus one-way ANOVAs we e used o compa e body
weigh be ween g oups unde di e en ood access condi ions
(expe imen 8). All da a we e exp essed as mean ±SEM, and
signi icance was se a p<0.05. STATISTICA 7 so wa e was
used.
RESULTS
Expe imen s 1 A and B. E ec o Ca eine
on Highly Pala able Food In ake Unde
Habi ual Condi ions: Con inuous o
In e mi en Access
The e ec o ca eine (0, 2.5, 5.0, 10.0, and 20.0 mg/kg) on pelle
in ake was eco ded du ing 30 min sessions. Repea ed measu es
ANOVA o con inuous access (N=8) e ealed a signi ican
e ec o ca eine [F(4, 28) =5.91; p<0.01]. Planned compa isons
F on ie s in Psychia y | www. on ie sin.o g 3Sep embe 2018 | Volume 9 | A icle 411
Co ea e al. Ca eine and Food Consump ion
FIGURE 1 | Schema ic ep esen a ion o he cage o habi ual and limi ed ood access, and he DL appa a us used in expe imen s 1–4, wi h he expe imen al
p ocedu e.
FIGURE 2 | E ec o saline o ca eine (2.5, 5.0, 10.0, and 20.0 mg/kg) on pala able ood in ake in a amilia con ex unde con inuous access (A) o unde in e mi en
access (B), and e ec o saline o ca eine (20.0 mg/kg) on ood in ake e alua ed on a DL box (C). Mean ±S.E.M. millig ams consumed. **p<0.01 signi ican ly
di e en om ehicle.
showed ha he highes dose o ca eine (20.0 mg/kg) signi ican ly
inc eased he amoun o pala able ood consumed (p<0.01;
Figu e 2A). The epea ed measu es ANOVA o he in e mi en
access g oup (N=8) e ealed also a signi ican e ec [F(4, 28)
=4.77; p<0.05], and he planned compa isons also con i med
ha he highes dose o ca eine inc eased he amoun o pelle s
consumed (p<0.01; Figu e 2B).
Expe imen 2. E ec o Ca eine on Highly
Pala able Food In ake Unde Anxiogenic
Condi ions
The same mice used in he p e ious expe imen s we e
used in o de o e alua e he e ec o ca eine on pelle
consump ion unde anxiogenic condi ions. Because con inuous
and in e mi en access exposi ion did no di e in he DL
F on ie s in Psychia y | www. on ie sin.o g 4Sep embe 2018 | Volume 9 | A icle 411
Co ea e al. Ca eine and Food Consump ion
FIGURE 3 | E ec o DMSO o TBZ (1, 2, 4, and 8 mg/kg) on pala able ood in ake in a amilia con ex unde con inuous access (A) o unde in e mi en access (B),
and e ec o DMSO o TBZ (8 mg/kg) on ood in ake e alua ed on a DL box (C). Mean ±S.E.M. millig ams consumed.
pa adigm da a om bo h g oups we e pulled o inc ease he
numbe o animals (N=16). Mice we e spli in o wo g oups;
one g oup ecei ed he dose o ca eine ha had inc eased in ake
(20.0 mg/kg) in expe imen 1 and he o he g oup ecei ed saline.
The - es o independen samples e ealed ha his dose o
ca eine signi ican ly dec eased ood in ake (mg in 15 min) in he
DL pa adigm [ (16) =2.90, p<0.01; Figu e 2C]. Howe e , his
dose o ca eine was no high enough o induce anxie y assessed
wi h classical measu es, such as ime spen in he li compa men
[ (16) =0.098; n.s]. The saline g oup spen 118.4 ±14.3, and he
ca eine g oup spen 120.0 ±8.6 s.
Expe imen s 3 A and B. E ec o TBZ on
Highly Pala able Food In ake Unde
Habi ual Condi ions: Con inuous o
In e mi en Access
The e ec o TBZ (0, 1.0, 2.0, 4.0, and 8.0 mg/kg) on pelle
in ake was e alua ed du ing 30 min sessions. Repea ed measu es
ANOVA o con inuous access (N=7) did no show a signi ican
e ec o TBZ [F(4, 24) =0.51; n.s.] on amoun o ood consumed
(Figu e 3A). The epea ed measu es ANOVA o he in e mi en
access g oup (N=7) e ealed no signi ican e ec ei he [F(4, 24)
=1.34; n.s.; Figu e 3B].
Expe imen 4. E ec o TBZ on Highly
Pala able Food In ake Unde Anxiogenic
Condi ions
Mice used in expe imen s 3 A and B we e used o e alua e he
e ec o TBZ on ood in ake in he DL box (Figu e 3C). These
animals (N=14) om he con inuous and in e mi en g oups
we e equally dis ibu ed in o wo ea men g oups: DMSO
ehicle o TBZ (8.0 mg/kg). The - es o independen samples
e ealed no signi ican di e ences on o al ood in ake be ween
hese wo g oups [ (12) =0.55; n.s.]. Classical measu es o anxie y,
such as ime in he li compa men , showed no signi ican e ec
o TBZ ei he [ (12) =0.54; n.s.]. The DMSO g oup spen 171.3
±16.8 s in he li compa men , and TBZ g oup spen 184.4 ±
17.5 s.
Expe imen s 5 and 6. E ec o Ca eine o
TBZ on Seeking o Highly Pala able Food
Unde E o ul Condi ions
A new g oup o mice (N=9) was ained in he T-maze.
Du ing he es ing phase, some animals ecei ed ca eine (10.0
mg/kg) o i s ehicle in he 2 i s weeks and hen TBZ
(4.0 mg/kg) o i s ehicle in he las 2 weeks and ano he
g oup o animals ecei ed d ugs in he e e sed o de . In
addi ion, ehicle o d ug ea men s we e adminis e ed in a
F on ie s in Psychia y | www. on ie sin.o g 5Sep embe 2018 | Volume 9 | A icle 411

Co ea e al. Ca eine and Food Consump ion
FIGURE 4 | Schema ic ep esen a ion o he T-maze appa a us used in he
p esen s udies, and expe imen al p ocedu e. All he su aces and he
doo way we e cons uc ed ou o Plexiglas, and he ba ie (depic ed in he
high-densi y a m, o he le ) was cons uc ed o wi e mesh. The high densi y
(HD) a m con ained wo ood pelle s, and he low densi y (LD) a m con ained
one ood pelle .
andomized o de . This expe imen was also a pilo o he
ollowing s udy. The - es o dependen samples showed ha
ca eine did no modi y amoun o ood consumed [ (8) =
0.38; n.s.] o selec ion o e o ul op ions (HD a m) [ (8)
=0.43; n.s.] o ge i , al hough i educed la ency o each
he ood ac oss all he ials [ (6) =3.21; p<0.01]. In
con as , 4.0 mg/kg TBZ did signi ican ly educed quan i y o
ood consumed [ (8) =3.44; p<0.01] because i educed
he numbe o HD a m choices [ (8) =3.92; p<0.01].
Howe e , he e was no signi ican e ec on la ency o each
he ood ac oss ials [ (7) = −1.71; n.s.]. Da a can be seen in
Figu es 5A–F.
Expe imen 7. Ca eine Re e sal o TBZ
E ec s on Seeking o Highly Pala able
Food Unde E o ul Condi ions
Naï e mice (N=8) ecei ed DMSO o TBZ (4.0 mg/kg)
120 min be o e es and saline o ca eine (2.5 o 5.0 mg/kg)
30 min be o e es (Figu es 6A–C). Repea ed measu es ANOVA
yielded an o e all e ec o d ug ea men [F(3, 21) =11.91, p
<0.01] on HD a m selec ion. Planned compa isons showed
ha TBZ/VEH and TBZ/Ca eine 2.5 mg/kg condi ion we e
signi ican ly di e en om VEH/VEH con ol condi ion (p
<0.01). In addi ion, co-adminis a ion o TBZ wi h he
highes dose o ca eine (5.0 mg/kg) signi ican ly inc eased
HD a m selec ion (p<0.01) compa ed o he TBZ/VEH
condi ion, indica ing an a enua ion o he DA-deple ing agen
e ec s. LD a m selec ion was also modi ied in he e e sal
s udy. Repea ed measu es ANOVA indica ed a signi ican
e ec o d ug ea men [F(3, 21) =8.44, p<0.01]. Planned
compa isons showed ha he TBZ/VEH and TBZ/Ca eine 2.5
mg/kg condi ion we e signi ican ly di e en om VEH/VEH
con ol condi ion (p<0.01), and TBZ plus ca eine (5.0
mg/kg) signi ican ly dec eased LD a m selec ion compa ed
o TBZ/VEH condi ion (p<0.01). The same pa e n o
esul s was ound o he dependen a iable mg o pelle s
consumed du ing he T-maze pe o mance: a signi ican e ec
o d ug ea men [F(3, 21) =4.15, p<0.01], and a signi ican
di e ence be ween VEH/VEH and TBZ/VEH (p<0.01) on
he one hand, as well as wi h TBZ/Ca eine 2.5 mg/kg on
he o he (p<0.05). Mo eo e , TBZ/Ca eine (5.0 mg/kg)
signi ican ly es o ed mg consumed compa ed o TBZ/VEH
condi ion (p<0.01).
Expe imen 8. Body Weigh P og ession
Unde Di e en Food Access Condi ions
Body weigh p og ession om week 1 ill week 11 om
animals in expe imen s 1A-B, and 7 as well as a con ol g oup
unde s anda d house ood condi ions a e shown in Figu e 7.
Independen epea ed measu es ANOVAs o body weigh da a,
showed ha he ou g oups o animals had a s a is ically
signi ican p og ession in hei body weigh gain: s anda d
con ol [F(10, 140) =18.15, p<0.01], con inuous access [F(10, 140)
=91.55, p<0.01], in e mi en access [F(10, 140) =167.81,
p<0.01] and es ic ed access in he T-maze expe imen [F(10, 80)
=22.52, p<0.01]. Planned compa isons indica e ha animals
in he s anda d g oup, in he con inuous g oup (Expe imen s
1A +3A), and in he in e mi en g oup (Expe imen s 1B +
3B) inc eased body weigh e e y week (p<0.01). Howe e ,
because o he home ood es ic ion egime, animals in he
es ic ed access condi ion (expe imen 7), educed signi ican ly
hei ini ial body weigh (<15%) du ing he 5 i s weeks, bu
by week 7, in spi e o he ood es ic ion, hey had e u ned
o hei ini ial body weigh , and p og essi ely inc eased e e y
week (p<0.01) a e ha . Fu he compa isons using a be ween
g oups ANOVA o he a e age weigh du ing weeks 2–6, o
weeks 7–11, demons a ed ha om week 2–6 he e was a
signi ican e ec o ype o ood access on body weigh [F(3, 50)
=20.9, p<0.01], and he planned compa isons showed ha
only he ood es ic ed g oup in he T-maze expe imen was
signi ican ly di e en om he s anda d g oup (p<0.01). The
one-way ANOVA o he a e age weigh in weeks 7–11 also
showed a signi ican e ec [F(3, 50) =18.0, p<0.01], bu in
his case, all g oups we e s a is ically di e en om he s anda d
g oup (p<0.05 o he con inuous g oup, and p<0.01 o he
o he wo g oups).
DISCUSSION
The p esen expe imen s indica e ha he non-selec i e
adenosine an agonis ca eine p oduces a complex pa e n o
e ec s on swee ood consump ion depending on he condi ions
in which he ood is p esen ed o mice. In expe imen 1 and
3 we ha e used wo p ocedu es o ood access, in which a e
se e al weeks o daily o in e mi en access o pala able ood
F on ie s in Psychia y | www. on ie sin.o g 6Sep embe 2018 | Volume 9 | A icle 411
Co ea e al. Ca eine and Food Consump ion
FIGURE 5 | Impac o ca eine ( ehicle o 10 mg/kg; A–C) o TBZ ( ehicle o 4 mg/kg; D–F) on la ency o ood ac oss 30 ials (A,D), HD a m selec ion (B,E), and o al
ood consumed (C,F) in he T-maze wi h ba ie . Mean (±S.E.M.) seconds om ga e o ood, numbe o a m choices in 30 ials and mg o ood consumed. **p< 0.01
signi ican ly di e en om hei co esponding ehicle.
FIGURE 6 | E ec s o saline o ca eine (2.5 and 5.0 mg/kg) in mice co-adminis e ed wi h DMSO o TBZ (4 mg/kg) on HD (A), LD (B) a m selec ion, and o al ood
consumed (C) in he T-maze wi h ba ie . Mean (±S.E.M.) numbe o a m choices in 30 ials and mg o pelle s consumed. *p<0.05; **p<0.01 signi ican ly di e en
om VEH/VEH; ##p<0.01 signi ican ly di e en om TBZ/VEH.
ha is di e en om he s anda d chow, mice show high le els
o consump ion. The e we e no di e ences in ood consumed
o in body weigh gain be ween he wo access condi ions.
Conside ing ha hese animals a e non- ood es ic ed, an in ake
o a ound 600 mg o ood in hal an hou is ema kable. These
esul s sugges ha ou p ocedu es could ha e induced “binge
F on ie s in Psychia y | www. on ie sin.o g 7Sep embe 2018 | Volume 9 | A icle 411
Co ea e al. Ca eine and Food Consump ion
FIGURE 7 | Body weigh p og ession ac oss di e en g oups o mice exposed o di e en ood access condi ions. (A) Body weigh (g ) gain om week 1 ill week 11,
and (B) a e age body weigh o he ou g oups du ing wo pe iods o 5 weeks each. Mean (±S.E.M.) weigh in g ams. *p<0.05; **p<0.01 signi ican ly di e en
om s anda d g oup.
ea ing,” which is cha ac e ized by excessi e ood in ake du ing
a sho pe iod o ime, and ypically is induced by o e ing
a highly pala able ood o luid on a limi ed, o in e mi en
schedule (30–32). In addi ion, he highe dose o ca eine used
(20.0 mg/kg) inc eased consump ion o his pala able ood o
a ound 1,100 millig ams (25 pelle s, 45 mg each) unde bo h
access condi ions. The e ec o ca eine is in ag eemen wi h a
ecen s udy in which acu e ca eine (up o 26 mg/kg) inc eased
s anda d chow consump ion unde habi ual home condi ions
on he i s 2 h (10). In ha s udy, animals did no each such
a high le el o in ake ( he maximum consump ion o chow
was 400 mg in 2 h). Thus, ca eine unde epea ed bu limi ed
access condi ions, did no ha e an ano ec ic e ec , bu ins ead i
po en ia ed pa e ns o binge ea ing o pala able ood.
Because ca eine can ha e anxiogenic e ec s as well (27,33),
and anxie y and s ess can a ec ood consump ion, we explo ed
he impac o he highes dose o ca eine (20.0 mg/kg) ha
was e ec i e in inc easing ood in ake, o e alua e i s e ec s
unde anxiogenic condi ions such as he b igh open a ea o he
classical DL box. Compa ing animals ha had ecei ed saline
wi h hose ha had ecei ed he 20.0 mg/kg dose o ca eine, we
obse ed an e ec ha was opposi e o ha seen in he p e ious
expe imen ; ca eine dec eased he amoun o ood consumed in
15 min. Howe e , his dose o ca eine did no modi y classical
anxie y measu es. This lack o anxie y e ec s is in acco dance
wi h p e ious expe imen s in mice using simila doses (10,27).
Thus, al hough ca eine a his dose did no show signs o being
anxiogenic in e ms o explo a ion in he DL box, i did educe
ood consump ion, showing ha his a iable is possibly mo e
sensi i e o he anxiogenic p ope ies o ca eine.
The pa e ns o esul s obse ed a e ca eine adminis a ion
we e no obse ed a e a b oad ange o TBZ doses ha
had p e iously been demons a ed o educe DA in en al
s ia um o mice (28) and, mo e speci ically, in Nacb co e o
a s (34). TBZ deple es DA in he cen al ne ous sys em by
e e sibly inhibi ing he esicle monoamine anspo e ype 2
and p e en ing monoamine up ake in o p esynap ic neu ons
(35). Thus, in he p esen esul s, DA deple ion did no a ec
consump ion o ood wi h a high con en o suc ose, nei he in a
amilia con ex no in an anxiogenic one when he e is no wo k
o e o in ol ed in he access o he ood.
As a psychos imulan , ca eine can also a ec locomo ion and
explo a ion. Inc eases in locomo ion wi h mode a e doses o
ca eine a e seen in s udies we e a low baseline ac i i y is seen,
such as mice habi ua ed o an open ield (27,33). Howe e ,
in si ua ions o high baseline le els such as olun a y unning,
high le els o locomo ion a e easie o dec ease (27). Thus, o
he ollowing T-maze e o condi ions in which explo a ion,
olun a y unning and ba ie climbing a e equi ed, lowe
doses o ca eine we e used. I has been demons a ed ha
he T-maze is a good ool o measu e e o - ela ed decision-
making in oden s (26,36,37), and ha his pa adigm is
sensi i e o beha io al manipula ions such as p e- eeding, which
de alues ood ein o cemen by educing appe i e and ood
mo i a ion inc easing omissions (26). As seen in he p esen
esul s (Figu es 5A–C) ca eine (10.0 mg/kg) dec eased la ency
o each he ood independen ly o he p esence o absence o
he ba ie . Howe e , his dose did no a ec p e e ence o he
a m ha con ained mo e pelle s and did no inc ease omissions
(da a no shown), which esul ed in no changes in pelle s ea ned
and consumed. The pa e n o esul s was qui e di e en when
ane gia was induced by adminis e ing he DA-deple ing agen
TBZ. A his dose, TBZ did no ha e a signi ican e ec on la ency
o each he ood, al hough some animals did show inc eased
la ency. TBZ- ea ed animals educed selec ion o he HD a m
bu compensa ed by inc easing he selec ion o he LD a m (see
F on ie s in Psychia y | www. on ie sin.o g 8Sep embe 2018 | Volume 9 | A icle 411
Co ea e al. Ca eine and Food Consump ion
Figu e 6B). This ea men did no change appe i e since animals
did no inc ease he numbe o omissions and ea all he pelle s
ha hey ea ned, al hough hey we e signi ican ly ewe (see
Figu es 6F,7C). In e es ingly, co-adminis a ion o a ela i e low
dose o ca eine (5.0 mg/kg), e e sed his ane gia inducing e ec
o TBZ. This e e sal o TBZ e ec s by ca eine is p obably due
o adenosine-DA ecep o in e ac ion ha ing opposi e e ec s on
he adenylyl cyclase- ela ed signal ansduc ion cascade. Thus,
ca eine ac ing on A1 ecep o s would educe he impac o
D1 ecep o s loca ed in he subs ance Pcon aining neu ons
o Nacb, and he same would be ue o ca eine ac ing on
he A2A ecep o co-localized wi h D2 ecep o s in enkephaline
con aining neu ons (21–23).
In ope an pa adigms ha equi e low e o ( a iable o
ixed in e als schedules) high doses o ca eine (≥50.0 mg/kg),
educe le e p essing o chow pelle s (11), bu doses up o
20.0 mg/kg inc eased le e p essing al hough he e was no ne
inc ease in access o highly pala able ood (12). In ope an
schedules ha equi e high e o ( ixed a io 20 o p og essi e
a io schedules), doses up o 40.0 mg/kg o ca eine educed le e
p essing o pala able ood (12), and in con as , ca eine (up o
25.0 mg/kg) imp o es pe o mance in animals esponding o
suc ose solu ions (13,14,38).
In summa y, ca eine adminis e ed acu ely o mice a
mode a e o high doses (27) can po en ia e binge ea ing when
subjec s ha e al eady s ablished a pa e n o excessi e ea ing.
Howe e , his same dose led o clea educ ions in ood
consump ion i he con ex is p one o inc ease anxie y le els
(i.e., he DL box). Al hough lowe doses o ca eine do no
change appe i e and do no impai o ien a ion owa d ood unde
e o ul condi ions, hey help o app oach he ood by imp o ing
speed and by e e sing pe o mance o no mal le els when
ane gia/ psychomo o e a da ion was induced by DA-deple ion.
This pa e n o esul s was qui e di e en o TBZ.
Al hough TBZ shi ed choice beha io in he T-maze, he ood
consump ion s udies showed ha TBZ had no e ec on ood
in ake. Thus, TBZ-induced shi s in e o - ela ed choice do no
appea o be due o changes in p ima y ood mo i a ion, he
uncondi ioned ein o cing p ope ies o ood, o ood p e e ence.
These da a suppo p e ious esul s indica ing ha Nacb DA
deple ions and TBZ do no subs an ially impai appe i e o
ood, o p oduce a gene al dis up ion o all aspec s o p ima y
ood mo i a ion (34,39–41). Nacb DA deple ions did no educe
ood in ake o eeding a e, no did hey impai ood handling
(40). Howe e , deple ion o mesolimbic DA has been epo ed
o p oduce impai men s in beha io al ac i a ion, and, mo e
speci ically, in e o - ela ed aspec s o ood mo i a ion (16). The
shi in choice induced by DA deple ion o an agonism in animals
es ed in he T-maze ba ie ask has been demons a ed o be a
alid oden model o psychomo o slowing, a igue and ane gia
(24,26,36,37,42–45). A e DA deple ions animals shi hei
beha io owa d he lowe e o op ions, howe e , hey a e s ill
o ien ed owa d he ood. Mo eo e , p e ious wo k has also
demons a ed ha , i animals do no ha e an op ion because
bo h a ms ha e a ba ie o because he e is no ood in he no-
ba ie a m, animals pe o m a con ol le els, demons a ing
ha DA deple ion o an agonism does no block he absolu e
mo o capaci y o he animal o climb he ba ie , and does no
educe gene al ood mo i a ion (24,26,37,42).
A DA sys em ha also has been implica ed in ood
mo i a ion is he inne a ion o he la e al hypo halamus (LH).
LH DA sou ces include local hypo halamic monoamine gic
neu ons exp essing enzymes o DA syn hesis (46), as well
as ascending p ojec ions om he Ven al Tegmen al A ea
(VTA) (47). I has been sugges ed ha DA elease in he LH
egula es he consumma o y componen o eeding beha io
(48). Thus, DA inc eases du ing eeding in p opo ion o
meal size (49), and o he appea ance o sa ia ion p ocesses
(48,50). Fu he mo e, low suc ose in ake was associa ed wi h
inc eased VTA DA (51), and DA adminis a ion in o he
LH p oduced ano ec ic e ec s (52,53). Howe e , al hough
we canno comple ely discoun he idea ha TBZ could be
ha ing an impac on LH neu ons, he lack o esul s on ood
consump ion a e TBZ adminis a ion in expe imen s 3A-
C, when mice had di ec access o he ood, and he clea
impac o TBZ in expe imen s 6 and 7 when an e o ul
app oach is essen ial o ge access o ood, indica es ha ou
expe imen al condi ions a e mo e ela ed o Nacb DA han
LH DA. Mo eo e , because he hypo halamus in oden s seems
spa se in adenosine ecep o s (54,55), i does no seem likely ha
ca eine could be ac ing on LH neu ons o modula e pala able
ood in ake.
The p esen wo k has po en ial clinical ele ance, since
appe i e is impai ed in many diso de s such as ano exia
and bulimia, anxie y and dep ession. In addi ion, e o -
ela ed mo i a ional symp oms such as ane gia, a igue,
and psychomo o slowing seen in dep essed humans a e
e y esis an o classical an idep essan ea men s such
as 5-HT up ake inhibi o s (56,57), and ca eine has been
demons a ed o enhance he an idep essan -like ac i i y o
common an idep essan d ugs (58). Mo eo e , ca eine is
an an agonis a bo h A1and A2A adenosine ecep o s, and
e idence indica es ha selec i e A2A ecep o an agonis s
may be use ul as ea men s o mo i a ional de ici s seen in
dep ession and o he diso de s (16,36). Fu he esea ch on he
mo i a ional e ec s o adenosine an agonis s may con ibu e o
a g ea e unde s anding o he neu al mechanisms media ing
a ious aspec s o mo i a ion.
AUTHOR CONTRIBUTIONS
MC and JS designed, supe ised he expe imen s and w o e
he manusc ip . NS, LL-C, CC-R, and RO-G pe o med he
expe imen s, con ibu ed o he analysis o da a and he w i ing
o he me hodology and he esul s.
ACKNOWLEDGMENTS
This wo k was suppo ed by a g an o MC om Minis e io
de Economía y Compe i i idad (PSI2015-68497-R), Spain,
and o JS by NIH/NIMH (R03MH094966-01A1). The
ollowing esea che s we e suppo ed by ellowships: NS
UJI (PREDOC/2012/28), LL-C ME-FPU (AP2010-3793), CC-R
MEC (BES-2016-077177), and RO-G GV (ACIF/2017/195).
F on ie s in Psychia y | www. on ie sin.o g 9Sep embe 2018 | Volume 9 | A icle 411