A icle
Es adiol Regula es Ene gy Balance by Amelio a ing
Hypo halamic Ce amide-Induced ER S ess
G aphical Abs ac
Highligh s
dCen al E2 egula es BAT he mogenesis ia he sympa he ic
ne ous sys em
dCen al E2 educes hypo halamic ce amides and ER s ess
dHypo halamic ce amides/ER s ess media e E2 ac ions on
BAT
Au ho s
Ismael Gonza
´lez-Ga cı
´a,
C is ina Con e as,
A
´nxela Es e
´ ez-Salgue o, ...,
Ch is ophe Magnan,
Manuel Tena-Sempe e, Miguel Lo
´pez
Co espondence
[email p o ec ed]
In B ie
Gonza
´lez-Ga cı
´a e al. demons a e ha
es adiol (E2) ac s in a p ecise a ea o he
hypo halamus, named he en omedial
nucleus (VMH), o egula e b own a
he mogenesis. The ac ions o E2 a e
media ed by modula ion o hypo halamic
ce amides and ER s ess.
Gonza
´lez-Ga cı
´a e al., 2018, Cell Repo s 25, 413–423
Oc obe 9, 2018 ª2018 The Au ho (s).
h ps://doi.o g/10.1016/j.cel ep.2018.09.038
Cell Repo s
A icle
Es adiol Regula es Ene gy Balance
by Amelio a ing Hypo halamic
Ce amide-Induced ER S ess
Ismael Gonza
´lez-Ga cı
´a,
1,2
C is ina Con e as,
1,2
A
´nxela Es e
´ ez-Salgue o,
1,2
F ancisco Ruı
´z-Pino,
2,3,4
Benoi Colsh,
5
I a
´n Pensado,
1,2
Lau a Lin
˜a es-Pose,
1,2
E a Rial-Pensado,
1,2
Pablo B. Ma ı
´nez de Mo en in,
1,2
Johan Fe nø,
6
Ca los Die
´guez,
1,2
Rube
´n Noguei as,
1,2
He e
´Le S un ,
7,8
Ch is ophe Magnan,
8
Manuel Tena-Sempe e,
2,3,4,9
and Miguel Lo
´pez
1,2,10,
*
1
Depa men o Physiology, CiMUS, Uni e si y o San iago de Compos ela-Ins i u o de In es igacio
´n Sani a ia, San iago de Compos ela,
15782, Spain
2
CIBER Fisiopa ologı
´a de la Obesidad y Nu icio
´n (CIBERobn), San iago de Compos ela, 15706, Spain
3
Depa men o Cell Biology, Physiology, and Immunology, Uni e si y o Co
´ doba, Co
´ doba, 14004, Spain
4
Ins i u o Maimo
´nides de In es igacio
´n Biome
´dica (IMIBIC)/Hospi al Reina So ı
´a, Co
´ doba, 14004, Spain
5
CEA-Cen e d’E ude de Saclay, Labo a oi e d’e
´ ude du Me
´ abolisme des Me
´dicamen s, Gi -su -Y e e, F ance
6
Ho mone Labo a o y, Haukeland Uni e si y Hospi al, Be gen, 5021, No way
7
Pa is-Saclay Ins i u e o Neu oscience, CNRS UMR 9197, Uni e si e
´Pa is-Sud, Uni e si y Pa is Saclay, O say 91405 Cedex, F ance
8
Uni e
´de Biologie Fonc ionnelle e Adap a i e, CNRS UMR 8251, Uni e si e
´Pa is Dide o , So bonne Pa is Ci e
´, Pa is, 75205, F ance
9
FiDiP o P og am, Resea ch Cen e o In eg a i e Physiology and Pha macology, Uni e si y o Tu ku, Kiinamyllynka u 10, 20520 Tu ku, Finland
10
Lead Con ac
*Co espondence: [email protected]
h ps://doi.o g/10.1016/j.cel ep.2018.09.038
SUMMARY
Compelling e idence has shown ha , besides i s pu-
a i e e ec on he egula ion o he gonadal axis,
es adiol (E2) exe s a dicho omic e ec on he hypo-
halamus o egula e ood in ake and ene gy expen-
di u e. The ano ec ic e ec o E2 is mainly media ed
by i s ac ion on he a cua e nucleus (ARC), whe eas
i s e ec s on b own adipose issue (BAT) he mogen-
esis occu in he en omedial nucleus (VMH). He e,
we demons a e ha cen al E2 dec eases hypo ha-
lamic ce amide le els and endoplasmic e iculum
(ER) s ess. Pha macological o gene ic blockade o
ce amide syn hesis and amelio a ion o ER s ess
selec i ely occu ing in he VMH ecapi ula e he e -
ec o E2, leading o inc eased BAT he mogenesis,
weigh loss, and me abolic imp o emen . These ind-
ings demons a e ha E2 egula ion o ce amide-
induced hypo halamic lipo oxici y and ER s ess is
an impo an de e minan o ene gy balance, sug-
ges ing ha dys egula ion o his mechanism may
unde lie some changes in ene gy homeos asis seen
in emales.
INTRODUCTION
One o he mos in e es ing and leas unde s ood aspec s o en-
e gy balance modula ion is gende dimo phism. Al hough some
common mechanisms con ol ene gy balance in bo h males and
emales, ele a ed le els o o a ian s e oids deeply a ec me a-
bolic ne wo ks in emales (Mau ais-Ja is e al., 2013, 2017;
Palme and Clegg, 2015; Mau ais-Ja is, 2015; Mo selli e al.,
2016; Lo
´pez and Tena-Sempe e, 2017). Fo example, dimin-
ished le els o es adiol (E2) a e o a ian insu iciency o any e i-
ology, including physiological (menopause) o su gical (o a iec-
omy [OVX]), a e associa ed wi h hype phagia, educed ene gy
expendi u e, and weigh gain (Mau ais-Ja is e al., 2013,
2017; Palme and Clegg, 2015; Mau ais-Ja is, 2015; Mo selli
e al., 2016; Lo
´pez and Tena-Sempe e, 2017). E2 eplacemen
in hese condi ions p ecludes o e e s OVX-induced obesi y
by educing ene gy in ake and ele a ing ene gy expendi u e; in
keeping wi h his, E2 eplacemen also p e en s me abolic com-
plica ions, such as glucose in ole ance and diabe es. (Mau ais-
Ja is e al., 2013, 2017; Palme and Clegg, 2015; Mau ais-Ja -
is, 2015; Mo selli e al., 2016; Lo
´pez and Tena-Sempe e, 2017).
Al hough hese bene icial e ec s o E2 we e assumed o be
mos ly exe ed a he pe iphe al le el, ecen ly gleaned da a
ha e shown ha , o a la ge ex en , hey a e exe ed a he
hypo halamus.
E2 di ec ly ac s on he CNS o modula e ene gy balance (Lo
´-
pez and Tena-Sempe e, 2015, 2017). Es ogen ecep o s (ERs)
a e widely exp essed h oughou he b ain, pa icula ly in hypo-
halamic si es, such as he a cua e (ARC), en omedial (VMH),
and pa a en icula (PVH) nuclei, as well as he p eop ic (POA)
and la e al (LHA) hypo halamic a eas (Sime ly e al., 1990; Simo-
nian and He bison, 1997; Voisin e al., 1997; Os e lund e al.,
1998; Me chen hale e al., 2004), all ha ing c i ical oles in he
egula ion o ene gy me abolism (Schneebe ge e al., 2014;
Sco e al., 2014; Magnan e al., 2015; Lo
´pez e al., 2016; Cui
e al., 2017). Se e al lines o e idence ha e shown ha E2 exe s
a nucleus-speci ic ac ion in he hypo halamus o modula e en-
e gy homeos asis. Thus, while mos o he e ec s o E2 on
eeding occu in p oopiomelanoco in (POMC) neu ons in he
ARC (Xu e al., 2011), E2 wi hin he VMH modula es b own adi-
pose issue (BAT) he mogenesis (Musa o e al., 2007; Xu
e al., 2011; Ma ı
´nez de Mo en in e al., 2014, 2015). Howe e ,
Cell Repo s 25, 413–423, Oc obe 9, 2018 ª2018 The Au ho (s). 413
This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
he exac molecula mechanism by which E2 exe s i s ac ions in
his nucleus emains unknown.
Cu en e idence has poin ed o hypo halamic endoplasmic
e iculum (ER) s ess as a cen al pa hophysiological mechanism
leading o insulin and lep in esis ance and subsequen ly o
obesi y (Zhang e al., 2008; Ma ı
´nez de Mo en in and Lo
´pez,
2010; Ozcan e al., 2009; Schneebe ge e al., 2013). I also
has been shown ha cen al ce amide-induced lipo oxici y a -
ec s ene gy balance (Ramı
´ ez e al., 2013; Pica d e al., 2013;
Tu pin e al., 2014; Con e as e al., 2014; Magnan e al., 2015).
In he hypo halamus, ele a ed ce amide concen a ion elici s
ER s ess, leading o weigh gain, insulin esis ance, hepa ic
s ea osis, dec eased sympa he ic one, and BAT he mogenesis
(Con e as e al., 2014, 2017). No ably, he cen al ac ion o ce-
amides can be e e sed by dec easing hypo halamic ER s ess,
esul ing in inc eased BAT he mogenesis and b owning o whi e
adipose issue (WAT), which ul ima ely amelio a es obesi y (Con-
e as e al., 2014, 2017). I also has been epo ed ha pe iphe al
E2 ea men educes se um ce amide concen a ion in OVX a s
(Vinaya ekhin e al., 2016). Howe e , despi e his e idence, i is
cu en ly unknown (1) whe he o a ian es ogens modula e hy-
po halamic ce amide-induced lipo oxici y, and, mo e impo -
an ly, (2) whe he impai men o his mechanism may be he
cause o obesi y in condi ions o es ogen de iciency. The e o e,
in his s udy we aim o de e mine he impo ance o hypo halam-
ic ce amides and ER s ess on he cen al e ec s o E2 on ene gy
homeos asis.
RESULTS
Cen al E2 Inhibi s Hypo halamic Ce amide-Induced
Lipo oxici y and ER S ess
OVX a s gained signi ican ly mo e weigh 15 days a e he p o-
cedu e (sham: 12.7 ±2.54 g; OVX: 48.63 ±2.38 g; p < 0.001) and
de eloped a ma ked hype phagia (sham: 16.1 ±0.42 g; OVX:
18.25 ±0.40 g a e 15 days o OVX; p < 0.001). OVX a s showed
he expec ed dec ease in u e us weigh and an inc ease in se um
lu einizing ho mone (LH) (da a no shown), con i ming he e i-
ciency o he OVX p ocedu e (Ma ı
´nez de Mo en in e al., 2014,
2015). As expec ed, cen al (in ace eb o en icula , ICV) E2
ea men elici ed a ma ked dec ease in body weigh and eeding
(Figu es 1A and 1B), as well as a educ ion in adiposi y and lean
mass (Figu e 1C). As epo ed be o e (Ca alcan i-de-Albuque -
que e al., 2014), cen al E2 adminis a ion did no a ec oxygen
consump ion (VO
2
72-h sham ehicle: 121.16 ±2.76 mL/g
lean mass; VO
2
72-h OVX ehicle: 114.66 ±3.13 mL/g lean
mass; VO
2
72-h OVX E2: 118.17 ±3.1 mL/g lean mass).
Howe e , cen al E2 signi ican ly augmen ed he ene gy
expendi u e (EE) o OVX a s (Figu e 1D). In keeping wi h his e -
idence, OVX a s ecei ing ICV E2 showed a educed espi a o y
quo ien (RQ; Figu e 1E), which is indica i e o highe lipid oxida-
ion, as well as an inc ease in body empe a u e and in BAT em-
pe a u e (Figu es S1A and S1B), uncoupling p o ein 1 (UCP1)
BAT exp ession (Figu e S1C) and a educ ion in he hepa ic lipid
con en (Figu e S1D). The selec ed dose o E2 was o me ly
demons a ed o be in he physiological ange and no o leak
o pe iphe y om he ce eb ospinal luid (Ma ı
´nez de Mo en in
e al., 2014, 2015).
Nex , we in es iga ed he e ec o OVX and cen al E2 eplace-
men on ce amide con en in he mediobasal hypo halamus
(MBH). We ound ha OVX induced a ma ked ele a ion o hypo-
halamic ce amide le els and ha E2 ICV es o ed hem o he
le els obse ed in sham-ope a ed a s (Figu e 1F). O no e, in as-
socia ion o he ise in ce amide concen a ion, OVX a s showed
inc eased hypo halamic ER s ess in he MBH, as demons a ed
by he augmen ed le els o un olded p o ein esponse (UPR)
ma ke s, such as phospho yla ed inosi ol- equi ing enzyme
(pIRE), phospho yla ed PKR-like ER kinase (pPERK), phospho y-
la ed euka yo ic ini ia ion ac o 2 alpha (peIF2a), and C/EBP
homologous p o ein (CHOP) (Figu e 1G). Again, cen al E2 admin-
is a ion e e sed he inc eased ER s ess by blun ing he ele a ed
le els o he UPR ma ke s o he le els o hose de ec ed in he
sham con ols. Pa icula ly, he hypo halamic exp ession o he
chape one glucose egula ed p o ein 78 kDa (GRP78; also called
binding immunoglobulin p o ein [BiP]; see below) was inc eased
a e cen al ea men wi h E2, as a cellula mechanism o educe
ER s ess (Figu e 1G).
Pha macological Inhibi ion o Hypo halamic Ce amide
Syn hesis Recapi ula es he E ec o Cen al E2 on
Ene gy Balance h ough he SNS
Ce amide-induced lipo oxici y has been epo ed o p omo e
posi i e ene gy balance due o educed he mogenic capaci y
(Con e as e al., 2014, 2017). The e o e, we aimed o in es iga e
whe he inhibi ion o ce amide syn hesis in he hypo halamus o
OVX a s migh imp o e hei me abolic pheno ype. Ch onic ICV
adminis a ion o he ce amide inhibi o my iocin (Ramı
´ ez e al.,
2013) o OVX a s dec eased body weigh and adiposi y inde-
penden o eeding (Figu es 1H–1J). This e ec was associa ed
wi h no maliza ion o ce amide and ER s ess ma ke s le els in
he MBH (Figu es 1K and 1L), inc eased body empe a u e (Fig-
u e 1M), ele a ed BAT empe a u e (Figu e 1N) and aised UCP1
p o ein le els in he BAT (Figu e 1O), as well as educed hepa ic
lipid le els (Figu e S1E) o OVX a s. Gi en ha cen al ce amides
ha e been epo ed o induce a ma ked dec ease in he ac i i y
o he sympa he ic ne ous sys em (SNS) inne a ing BAT (Con-
e as e al., 2014), we aimed o in es iga e whe he ad ene gic
ecep o blockade a ec ed he cen al e ec o ICV my iocin
on BAT he mogenesis o OVX a s. The e ec o cen al admin-
is a ion o my iocin was e e sed (in a eeding-independen
manne ) by pha macological blockade o be a 3 ad ene gic e-
cep o (b3-AR) wi h he speci ic an agonis , SR59230A (Lo
´pez
e al., 2010; Ma ı
´nez de Mo en in e al., 2014; Con e as e al.,
2014, 2017; Ma ı
´nez-Sa
´nchez e al., 2017a; Seoane-Collazo
e al., 2018)(Figu es 2A and 2B). The inc ease in body weigh
induced by SR59230A was associa ed wi h equi alen e e sal
o ICV my iocin-induced ac i a ion o body empe a u e (Fig-
u e 2C), BAT empe a u e (Figu e 2D), and BAT UCP1 exp ession
(Figu e 2E).
Silencing o SPTLC Speci ically in he VMH Amelio a es
ER S ess and he Me abolic Pheno ype o OVX Ra s
To u he in es iga e he ole o de no o ce amide syn hesis
pa hway and o dissec in which hypo halamic nucleus hose ac-
ions ake place, we a ge ed se ine palmi oyl ans e ase long
chain base subuni 1 (SPTLC1). This enzyme ca alyzes he
414 Cell Repo s 25, 413–423, Oc obe 9, 2018
limi ing i s s ep in he de no o syn hesis o ce amides: he
condensa ion o palmi oyl-CoA and se ine, p oducing 3-ke os-
phinganine. The SPTLC is composed o wo subuni s (SPTLC1
and 2); bo h a e essen ial o enzyme unc ion because hey
cons i u e i s ca aly ic co e (Hanada, 2003; Ya d e al., 2007;
Wa son e al., 2009). The e o e, we silenced SPTLC1 exp ession
by using adeno i uses ha bo ing a small hai pin RNA (shRNA) o
con ol adeno i uses exp essing g een luo escence p o ein
(GFP) alone (Wa son e al., 2009), speci ically in he VMH, a
key hypo halamic nucleus modula ing he mogenesis (Mo ison
e al., 2014; Con e as e al., 2015). In ec ion in he VMH was as-
sessed by isualiza ion o GFP exp ession (Figu e 3A) and by
dec eased p o ein le els o SPTLC1 (Figu e 3B). Gi en ha he
s abili y o he SPTLC2 subuni is inhe en ly dependen on he
exp ession o SPTLC1 and bo h subuni s associa e wi h a 1:1
mola s oichiome y (Hanada, 2003; Wa son e al., 2009),
silencing SPTLC1 also led o an a endan educ ion in he
exp ession o SPTLC2 (Figu e 3B), as expec ed and as p e i-
ously shown (Wa son e al., 2009).
Adminis a ion o adeno i uses encoding shSPTLC1 in he
VMH induced a eeding-independen weigh loss in OVX a s,
bu no in sham a s (Figu es 3C–3F). This e ec was associa ed
Figu e 1. E ec o Cen al E2 and My iocin on Ene gy Balance in OVX Ra s
(A–G) Body weigh change (A), daily ood in ake (B), a mass change (le panel) and lean mass change ( igh panel) (C), ime cou se ene gy expendi u e (EE) (le
panel) and o al ene gy expendi u e change ( igh panel) (D), espi a o y quo ien (E), ce amide le els in he MBH (F), and ep esen a i e wes e n blo au o a-
diog aphic images (le panel) and MBH p o ein le els o UPR ( igh panel) (G) o sham a s o OVX a s ICV ea ed wi h ehicle o E2 (n = 7–9 animals pe g oup).
(H–O) Body weigh change (H), daily ood in ake (I), a mass change (le panel) and lean mass change ( igh panel) (J), ce amide le els in he MBH (K), ep e-
sen a i e wes e n blo au o adiog aphic images (le panel) and MBH p o ein le els o UPR ( igh panel) (L), body empe a u e (M), ep esen a i e in a ed he mal
images (le panel) and empe a u e o BAT a ea ( igh panel) (N), and ep esen a i e wes e n blo au o adiog aphic images (le panel) and p o ein le els o UCP1
in he BAT ( igh panel) (O) o sham a s o OVX a s ICV ea ed wi h ehicle o my iocin (n = 7–16 animals pe g oup o all he analyses, bu he body weigh and
ood in ake measu emen s we e n = 28–38).
All da a a e exp essed as mean ±SEM. *, **, and ***p < 0.05, 0.01, and 0.001 e sus sham ehicle; #, ##, and ### p < 0.05, 0.01, and 0.001 e sus OVX ehicle. Fo
he wes e n blo analyses, ep esen a i e images o all p o eins a e shown; in he case o he loading con ols a ep esen a i e gel is displayed o cla i y, al hough
each band o each p o ein was always co ec ed by i s own in e nal con ol band (b-ac in o a- ubulin). The bands o each pic u e always come om he same gel,
al hough hey ha e been spliced o cla i y.
See also Figu e S1.
Cell Repo s 25, 413–423, Oc obe 9, 2018 415
wi h a ma ked elie o ER s ess in he VMH o OVX a s, as
demons a ed by he educed p o ein le els o pIRE, pPERK,
peIF2a, ATF6a, and CHOP, which we e also obse ed in sham
a s, bu o a lesse ex en (Figu es 3G–3H). In keeping wi h hese
da a, injec ion o shSPTLC1 adeno i uses led o inc eased body
empe a u e (Figu es 4A and 4B), BAT empe a u e (Figu es 4C
and 4D) and UCP1 p o ein le els in he BAT (Figu es 4E and
4F) o OVX, bu no o sham, a s. These e ec s we e associa ed
wi h an imp o emen in he me abolic pheno ype o OVX a s, as
demons a ed by dec eased hepa ic s ea osis (Figu e S2A).
O e all, hese esul s, alongside he my iocin da a, indica e
ha he hypo halamic and, mo e p ecisely, VMH ce amide le els
media ed he cen al ac ions o E2 on BAT he mogenesis and
ene gy balance.
Pha macological Inhibi ion o Hypo halamic ER S ess
Recapi ula es he E ec o Cen al E2 on Ene gy
Balance
Inc eased hypo halamic ER s ess has been associa ed wi h he
de elopmen o insulin and lep in esis ance, leading o obesi y
(Zhang e al., 2008; Ozcan e al., 2009; an Dam e al., 2015;
Schneebe ge e al., 2013; Con e as e al., 2014, 2017). Bea ing
in mind ha hypo halamic ce amides elici ER s ess (Con e as
e al., 2014, 2017), we aimed o in es iga e whe he ce amide-
induced ER s ess may be a mechanism media ing he cen al
ac ions o E2 de iciency on BAT he mogenesis and ene gy ho-
meos asis. Thus, OVX a s we e ICV ea ed wi h he chemical
chape one au ou sodeoxycholic acid (TUDCA) (Zhang e al.,
2008; Ozcan e al., 2009; Schneebe ge e al., 2013; Imbe non
e al., 2016; Con e as e al., 2017; Po ei o e al., 2017). Cen al
adminis a ion o his d ug o OVX, bu no sham, a s induced
eeding-independen weigh loss (Figu es 5A–5D), dec eased
hypo halamic ER s ess (Figu es 5E and 5F), a end o inc ease
body empe a u e (Figu e 5G), ele a ed BAT empe a u e (Fig-
u e 5H) and UCP1 p o ein le els in BAT (Figu e 5I). O no e,
none o he me abolic changes we e ound in sham a s (Figu es
5G–5I), which is in line wi h he ac ha hey did no show
ele a ed ER s ess (Figu es 1G and 1L). Fu he mo e, he cen al
injec ion o TUDCA educed he hepa ic lipid con en o OVX a s
bu had no e ec s in sham a s (Figu e S3A).
Figu e 2. E ec o My iocin and b3-AR An agonism on BAT The mogenesis in OVX Ra s
(A–E) Body weigh change (A), daily ood in ake (B), body empe a u e (C), ep esen a i e in a ed he mal images (le panel) and empe a u e o BAT a ea ( igh
panel) (D), and ep esen a i e wes e n blo au o adiog aphic images (le panel) and p o ein le els o UCP1 in he BAT ( igh panel) (E) o OVX a s ICV ea ed wi h
my iocin and subcu aneously ea ed wi h ehicle o he b3-AR an agonis SR59230A (n = 7–18 animals pe g oup).
All da a a e exp essed as mean ±SEM. ** and ***p < 0.01 and 0.001 e sus OVX ehicle ehicle; #, ##, and ### p < 0.05, 0.01, and 0.001 e sus OVX my iocin
ehicle. Fo he wes e n blo analyses, ep esen a i e images o all p o eins a e shown; in he case o he loading con ols a ep esen a i e gel is displayed o
cla i y, al hough each band o each p o ein was always co ec ed by i s own in e nal con ol band (a- ubulin). The bands o each pic u e always come om he
same gel, al hough hey ha e been spliced o cla i y.
416 Cell Repo s 25, 413–423, Oc obe 9, 2018
GRP78 in he VMH Dec eases Body Weigh and Imp o es
he Me abolic Pheno ype o OVX Ra s h ough he SNS
Ou i s se o da a showed ha cen al ea men wi h E2
dec eased ER s ess and augmen ed he hypo halamic p o ein
le els o GRP78 (Figu e 1G). This was o impo ance because
his p o ein is a chape one, loca ed in he ER, ha acili a es
he p o ein olding ups eam o he UPR and he e o e dec eases
ER s ess le els (G ego and Ho amisligil, 2011; Fu e al., 2012).
Thus, a gain o unc ion expe imen was unde aken using an
adeno i us encoding GRP78 o con ol adeno i uses exp essing
GFP alone injec ed in o he VMH o OVX a s and hei sham con-
ols. In ec ion e iciency in he VMH was assessed by exp es-
sion o GFP (simila ly o Figu e 3A; da a no shown), and by
inc eased concen a ion o GRP78 in he VMH (Figu es 6E and
6F). GRP78 adeno i uses in o he VMH induced eeding-inde-
penden weigh loss in OVX a s, bu no in sham a s (Figu es
6A–6D). Tha e ec was linked o dec eased ER s ess in he
VMH o OVX, bu no sham, a s (Figu es 6E and 6F), inc eased
body empe a u e (Figu e 6G), BAT empe a u e (Figu e 6H),
and augmen ed UCP1 p o ein le els in BAT (Figu e 6I). No ably,
cen al adminis a ion o GRP78 in he VMH educed he hepa ic
lipid con en o OVX a s (Figu e S3B).
Finally, we in es iga ed whe he egula ion o BAT ollowing
adminis a ion o GRP78 adeno i al pa icles in he VMH o
OVX a s was media ed by he SNS. Pha macological inac i a-
ion o b3-AR by SC adminis a ion o he speci ic an agonis ,
SR59230A (Lo
´pez e al., 2010; Ma ı
´nez de Mo en in e al.,
2014; Con e as e al., 2014, 2017; Ma ı
´nez-Sa
´nchez e al.,
2017a; Seoane-Collazo e al., 2018), p e en ed he e ec on
body weigh associa ed wi h cen al adminis a ion o GRP78 i-
uses (Figu e 7A) wi hou a ec ing eeding (Figu e 7B). Consis-
en ly, he ea men wi h SR59230A blun ed he GRP78-induced
Figu e 3. E ec o SPTLC1 Down egula ion in he VMH o OVX Ra s on Ene gy Balance
(A–H) Di ec luo escence o GFP (A), ep esen a i e wes e n blo au o adiog aphic images (le panel) and VMH p o ein le els o se ine palmi oyl ans e ase, long
chain subuni 1-2 ( igh panel) (B), body weigh change (C and E), daily ood in ake (D and F), and ep esen a i e wes e n blo au o adiog aphic images (le panel)
and VMH p o ein le els o UPR ( igh panel) (G and H) o sham o OVX a s s e eo axically ea ed wi h adeno i uses encoding GFP o shSPTLC1 in o he VMH
(n = 7–20 animals pe g oup).
All da a a e exp essed as mean ±SEM. *, ** and ***p < 0.05, 0.01, and 0.001 e sus sham GFP o OVX GFP. Fo he wes e n blo analyses, ep esen a i e images
o all p o eins a e shown; in he case o he loading con ols a ep esen a i e gel is displayed o cla i y, al hough each band o each p o ein was always co ec ed
by i s own in e nal con ol band (b-ac in). The bands o each pic u e always come om he same gel, al hough hey ha e been spliced o cla i y.
See also Figu e S2.
Cell Repo s 25, 413–423, Oc obe 9, 2018 417
inc ease in body empe a u e (Figu e 7C), BAT empe a u e (Fig-
u e 7D), and UCP1 p o ein le els in he BAT (Figu e 7E). O e all,
his e idence demons a es ha E2-induced dec ease o ER
s ess in he VMH is an impo an modula o o ene gy balance
by con olling BAT he mogenesis.
DISCUSSION
This s udy iden i ies a link be ween he e ec s o E2 on hypo ha-
lamic ce amide-induced lipo oxici y and ER s ess wi h BAT
he mogenesis. We demons a e ha cen al E2 eplacemen
es o ed hypo halamic ce amide le els and ER s ess in OVX
emale a s. O no e, pha macological o gene ic blockade o
ce amide syn hesis and elie o ER s ess ecapi ula e he
e ec s o E2, namely inc eased BAT he mogenesis, weigh
loss, and educed hepa ic lipids.
Condi ions o es ogen de iciency, such as o a iec omy o
menopause, a e associa ed wi h a posi i e ene gy balance, as
esul o hype phagia and dec eased ene gy expendi u e, lead-
ing o enhanced adiposi y (Mau ais-Ja is e al., 2013, 2017;
Palme and Clegg, 2015; Mau ais-Ja is, 2015; Mo selli e al.,
2016; Lo
´pez and Tena-Sempe e, 2017), hepa ic lipid accumula-
ion, and s ea osis (Paque e e al., 2007; Vo
¨lzke e al., 2007; Pa-
que e e al., 2008) in animal models and humans. Pe iphe al and
cen al es ogen eplacemen he apy e e s his pheno ype
bo h in women and in emale oden s (Mau ais-Ja is e al.,
2013, 2017; Palme and Clegg, 2015; Mau ais-Ja is, 2015;
Mo selli e al., 2016; Lo
´pez and Tena-Sempe e, 2017). Recen
e idence indica es ha AMP-ac i a ed p o ein kinase (AMPK)
in he VMH media es he cen al e ec o E2 on BAT he mogen-
esis h ough he SNS (Ma ı
´nez de Mo en in e al., 2014, 2015).
Conside ing he key ole o AMPK on a y acid me abolism
(Kahn e al., 2005; Lage e al., 2008; Lo
´pez e al., 2016), his e -
idence di ec ly links he cen al e ec s o es ogens wi h hypo-
halamic lipids. Howe e , despi e ecen da a showing sexual
dimo phism in b ain a y acid con en (Rod iguez-Na as e al.,
2016) and e en hough pe iphe al E2 ea men educes se um
ce amide le els in OVX a s (Vinaya ekhin e al., 2016), whe he
complex lipids me abolism is in ol ed in he cen al ac ions o E2
emains unclea . This is ele an , because he ho monal egula-
ion o lipid me abolism usually shows issue-speci ic e ec s
(Lo
´pez e al., 2010; Ma ı
´nez-Sa
´nchez e al., 2017a). In his
sense, we ha e ecen ly epo ed ha hy oid ho mones (THs)
modula e ce amide me abolism in an opposi e ashion wi hin
di e en b ain a eas, such as he hypo halamus and he ce eb al
co ex (Ma ı
´nez-Sa
´nchez e al., 2017a). The e o e, he ac ha
sys emic E2 dec eased se um ce amide le els (Vinaya ekhin
e al., 2016) did no imply a simila hypo halamic e ec .
Compelling e idence indica es ha hypo halamic ce amide-
induced lipo oxici y is a pa hological mechanism leading o
obesi y by supp essing BAT he mogenesis (Con e as e al.,
2014, 2017). Impo an ly, pha macological o gene ic a ge ing
Figu e 4. E ec o SPTLC1 Down egula ion in he VMH o OVX Ra s on BAT The mogenesis
(A–F) Body empe a u e (A and B), ep esen a i e in a ed he mal images (le panel) and empe a u e o BAT a ea ( igh panel) (C and D), and ep esen a i e
wes e n blo au o adiog aphic images (le panel) and p o ein le els o UCP1 in he BAT ( igh panel) (E and F) o sham o OVX a s s e eo axically ea ed in he
VMH wi h adeno i uses encoding GFP o shSPTLC1 (n = 7–20 animals pe g oup).
All da a a e exp essed as mean ±SEM. * and ***p < 0.05 and 0.001 e sus sham GFP o OVX GFP. Fo he wes e n blo analyses, ep esen a i e images o all
p o eins a e shown; in he case o he loading con ols a ep esen a i e gel is displayed o cla i y, al hough each band o each p o ein was always co ec ed by i s
own in e nal con ol band (a- ubulin). The bands o each pic u e always come om he same gel, al hough hey ha e been spliced o cla i y.
See also Figu e S2.
418 Cell Repo s 25, 413–423, Oc obe 9, 2018
o his pa hway, causing educed ce amide concen a ion and/o
dec eased ER s ess amelio a es obesi y (Con e as e al., 2014,
2017). Howe e , i emains o ally unknown whe he luc ua ions
in ce amide le els and/o ER s ess e lec b ain changes asso-
cia ed wi h, o example, neu onal and/o glial me abolic ac i i y
(Pica d e al., 2013, 2014; Magnan e al., 2015), o whe he hey
migh espond o o in eg a e nu i ional and ho monal cues om
he pe iphe y (Lo
´pez e al., 2016). Suppo ing his la e idea,
se e al endoc ine signals, such as THs (Lo
´pez e al., 2010; Al-
a ez-C espo e al., 2016; Ma ı
´nez-Sa
´nchez e al., 2017a, b),
lep in (Wol gang e al., 2007; Tanida e al., 2013), bone-mo pho-
gene ic p o ein 8B (BMP8B) (Whi le e al., 2012; Ma ins e al.,
2016), glucagon-like pep ide 1 (GLP-1) (Bei oa e al., 2014),
and, impo an ly, E2 (Ma ı
´nez de Mo en in e al., 2014, 2015),
ha e been shown o egula e hypo halamic a y acid me a-
bolism o modula e ene gy balance.
Thus, we hypo hesized ha E2-induced ac ions on ene gy bal-
ance could be media ed by modula ion o hypo halamic ce am-
ides and ER s ess. Ou da a showed ha cen al ea men wi h
E2 induced a nega i e ene gy balance in OVX a s, associa ed
wi h a educ ion o hypo halamic ce amide-induced lipo oxici y
and ER s ess. O no e, cen al pha macological inhibi ion o
ce amide syn hesis by using he SPTLC inhibi o my iocin sup-
p essed hypo halamic ER s ess and ully ecapi ula ed he e -
ec s o E2 on ene gy balance. Nex , we aimed o u he in es-
iga e he speci ic hypo halamic nuclei whe e hose ac ions
ake place. I is known ha he VMH plays a majo ole in he
modula ion o BAT he mogenesis (Mo ison e al., 2014;
Figu e 5. E ec o Cen al TUDCA on Ene gy Balance in OVX Ra s
(A–I) Body weigh change (A and C), daily ood in ake (B and D), ep esen a i e wes e n blo au o adiog aphic images (le panel) and MBH p o ein le els o
UPR ( igh panel) (E and F), body empe a u e (G), ep esen a i e in a ed he mal images (le panel) and empe a u e o BAT a ea ( igh panel) (H), and ep e-
sen a i e wes e n blo au o adiog aphic images (le panel) and p o ein le els o UCP1 in he BAT ( igh panel) (I) o sham o OVX a s ICV ea ed wi h ehicle o
TUDCA (n = 7–8 animals pe g oup).
All da a a e exp essed as mean ±SEM.* and **p < 0.05 and 0.01 e sus sham ehicle o OVX ehicle. Fo he wes e n blo analyses, ep esen a i e images o all
p o eins a e shown; in he case o he loading con ols a ep esen a i e gel is displayed o cla i y, al hough each band o each p o ein was always co ec ed by i s
own in e nal con ol band (b-ac in o a- ubulin). The bands o each pic u e always om come om he same gel, al hough hey ha e been spliced o cla i y.
See also Figu e S3.
Cell Repo s 25, 413–423, Oc obe 9, 2018 419
Con e as e al., 2015) and ha ERais he si e o ac ion o E2 in
he VMH o d i e he modula ion o BAT he mogenic unc ion
(Musa o e al., 2007; Xu e al., 2011; Ma ı
´nez de Mo en in
e al., 2014, 2015). The e o e, ollowing a gene ic s a egy we
silenced SPTLC1 exp ession in he VMH o OVX obese a s. Spe-
ci ic down egula ion o SPTLC exclusi ely in he VMH amelio-
a ed ER s ess and obesi y and imp o ed he me abolic heal h
o OVX a s, including a educ ion in hepa ic lipid con en . O e -
all, hese da a sugges ha cen al E2 induces a nega i e ene gy
balance by diminishing ce amide-induced lipo oxici y and ER
s ess, a hypo hesis ha was p o en by ea ing OVX a s cen-
ally wi h he chemical chape one TUDCA o by gene ically
o e exp essing he chape one GRP78 wi hin he VMH. Impo -
an ly, all he abo e e ec s we e e e sed by pha macological
blockage o b3-AR, he e o e sugges ing ha VMH changes in
ce amide signaling and hypo halamic ER s ess induced by E2
o inc ease he mogenesis a e con eyed ia he sympa he ic
ou low o BAT.
In summa y, in his s udy we demons a e ha E2 p omo es
an amelio a ion o ce amide-induced lipo oxici y and ER
s ess, speci ically in he VMH. This ac ion leads o inc eased
BAT he mogenesis h ough he ac i a ion o sympa he ic
b3-AR signaling, an ac ion ha is associa ed wi h eeding-inde-
penden weigh loss and educed hepa ic s ea osis, as well as
wi h a subs an ial a enua ion (o e e sal) o some o he me a-
bolic hallma ks o OVX and menopause, namely, body weigh
gain, and hepa ic s ea osis. This e idence is o ele ance
because p ope unde s anding o gende di e ences in he
Figu e 6. E ec o GRP78 O e exp ession in he VMH o OVX Ra s on Ene gy Balance and BAT The mogenesis
(A–I) Body weigh change (A and C), daily ood in ake (B and D), ep esen a i e wes e n blo au o adiog aphic images (le panel) and VMH p o ein le els o UPR
( igh panel) (E and F), body empe a u e (G), ep esen a i e in a ed he mal images (le panel) and empe a u e o BAT a ea ( igh panel) (H), and ep esen a i e
wes e n blo au o adiog aphic images (le panel) and p o ein le els o UCP1 in he BAT ( igh panel) (I) o sham o OVX a s s e eo axically ea ed in he VMH wi h
adeno i uses encoding GFP o GRP78 (n = 7–18 animals pe g oup).
All da a a e exp essed as mean ±SEM. *, **, and ***p < 0.05, 0.01, and 0.001 e sus sham GFP o OVX GFP. Fo he wes e n blo analyses, ep esen a i e images
o all p o eins a e shown; in he case o he loading con ols a ep esen a i e gel is displayed o cla i y, al hough each band o each p o ein was always co ec ed
by i s own in e nal con ol band (b-ac in o a- ubulin). The bands o each pic u e always come om he same gel, al hough hey ha e been spliced o cla i y.
See also Figu e S3.
420 Cell Repo s 25, 413–423, Oc obe 9, 2018
e e y 30 min. Animals we e placed o adap a ion o 1 week be o e s a ing he measu emen s. Fo he measu emen o body
composi ion, we used nuclea magne ic esonance (NMR) (Whole Body Composi ion Analyze ; EchoMRI; Hous on, TX) (Imbe non
e al., 2013; Ma ı
´nez de Mo en in e al., 2012, 2014; Ma ins e al., 2016; Ma ı
´nez-Sa
´nchez e al., 2017a; Seoane-Collazo e al.,
2018). Animals do no need o be anes he ized nei he o he special p epa a ion be o e measu emen . They a e placed in a holde
o cus om-de ined size du ing he measu emen (measu ing ime: 0.5-3.2 min). Two measu emen s we e done wice o animal
72 h apa .
Sample p ocessing
Ra s we e killed by ce ical disloca ion and decapi a ion. F om each animal, he MBH, VMH, li e and he in e scapula BAT we e
collec ed o wes e n blo ing and Oil Red O s aining and immedia ely homogenized on ice o p ese e phospho yla ed p o ein le els.
Those samples and he se um we e s o ed a 80C un il u he p ocessing. Dissec ion o he VMH was pe o med by mic o-punch
p ocedu e unde he mic oscope, as p e iously desc ibed (Lo
´pez e al., 2010; Whi le e al., 2012; Con e as e al., 2014, 2017; Ma -
ı
´nez de Mo en in e al., 2014; Ma ins e al., 2016; Ma ı
´nez-Sa
´nchez e al., 2017a; Seoane-Collazo e al., 2018).
Ce amide quan i ica ion
Ce amide analysis epo ed in his manusc ip we e pe o med in wo di e en labo a o ies and acco ding o he nex me hods. In he
i s one, ce amides we e ex ac ed and analyzed using liquid ch oma og aphy-mass spec ome e consis ed o a Wa e s Acqui y
UPLC Sys em connec ed o a Wa e s LCT P emie O hogonal Accele a ed Time o Fligh Mass Spec ome e (Wa e s; Mill o d,
MA, USA), ope a ed in posi i e elec osp ay ioniza ion mode. Full scan spec a om 50 o 1500 Da we e acqui ed and indi idual
spec a we e summed o p oduce da a poin s each 0.2 s. Mass accu acy and ep oducibili y we e main ained by using an indepen-
den e e ence sp ay by he LockSp ay in e e ence. The analy ical column was a 100 mmL 2.1 mm i.d., 1.7 mm C8 Acqui y UPLC
BEH (Wa e s; Mill o d, MA, USA). The wo mobile phases we e phase A: me hanol; phase B: wa e , bo h con ained 0.2% o mic acid
( / ) and 2 mM ammonium o ma e. A linea g adien was p og ammed (0.0 min: 20% B; 3 min: 10% B; 6 min: 10% B; 15min: 1% B;
18 min: 1% B; 20 min: 20% B; 22 min: 20% B). The low a e was 0.3 mL/min. The column was held a 30C. Quan i ica ion was ca ied
ou using he ex ac ed ion ch oma og am o each compound, using 50 mDa windows. The linea dynamic ange was de e mined by
injec ing s anda d mix u es. Posi i e iden i ica ion o compounds was based on he accu a e mass measu emen wi h an e o < 5
ppm and i s LC e en ion ime, compa ed o ha o a s anda d (±2%). Concen a ions we e measu ed by mul iple eac ion moni o ing
expe imen s using N-lau oyl-D-e y h o-sphingosine (C12-ce amide) and N-Hep adecanoyl-D-e y o-Sphingosine (C17-ce amide)
as in e nal s anda ds and using as pa ons N-Palmi oyl-D-e y o-sphingosine (C16-ce amide), N-S ea oyl-D-e y o-sphingosine
(C18-ce amide), N-lignoce oyl-D-e y h o-sphingosine (C24-ce amide) and N-ne onoyl-D-e y h o-sphingosine (C24:1-ce amide)
(A an i Pola Lipids; Alabas e , AL, USA). The ce amide analyses om Figu e 1K we e pe o med in he Resea ch Uni on Bioac i e
Molecules (RUBAM) o he Ins i u de Quı
´mica A anc¸ ada de Ca alunya (IQAC-CSIC) acco ding o his me hod.
In he second one, o al lipid species we e ex ac ed acco ding o he me hod o Folch e al. (1957). Ce amide lipid species we e
en iched in he lowe phase (o ganic phase) and analyzed using an un a ge ed lipidomic app oach by liquid ch oma og aphy coupled
wi h high- esolu ion mass spec ome y (LC–HRMS) as desc ibed by Seye e al. (2016). The ce amide analyses om Figu e 1F we e
pe o med in he Uni e
´de Biologie Fonc ionnelle e Adap a i e (Uni e si e
´Pa is Dide o ) acco ding o his me hod.
His ological analyses
Hepa ic lipid con en was analyzed by Oil Red O s aining, as p e iously shown (Seoane-Collazo e al., 2014; Con e as e al., 2014,
2017; Imbe non e al., 2016; Po ei o e al., 2017; Ma ı
´nez-Sa
´nchez e al., 2017a). Hepa ic ozen sec ions we e cu (8 mm) and ixed in
10% bu e ed o maldehyde. Sec ions we e s ained in il e ed Oil Red O (Sigma; S . Louis, MO, USA), washed in dis illed wa e , coun-
e s ained wi h Ha is hema oxylin (Bio-Op ica; Milan, I aly) and washed in dis illed wa e again. Sec ions we e moun ed in aqueous
moun ing medium (Bio-Op ica; Milan, I aly). Images we e aken wi h a digi al came a Olympus XC50 (Olympus Co po a ion; Tokyo,
Japan) a 20X. Digi al images o li e we e quan i ied wi h ImageJ So wa e (Na ional Ins i u es o Heal h; USA). Di ec de ec ion o
GFP luo escence was pe o med a e pe usion o he animals; on 40 mm b ain sec ions. Images we e aken wi h a luo escence
mic oscope Olympus IX51, as p e iously shown (Con e as e al., 2017; Seoane-Collazo e al., 2018).
Wes e n blo ing
P o ein lysa es om hypo halamus (MBH o VMH) and BAT we e homogenized in lysis bu e (consis ing o a mix o 0.05 M T is-HCl,
0.01 M EGTA, 0.001 M EDTA, 0.016 M T i on X-100, 0.001 M sodium o ho anada e, 0.05 M sodium luo ide, 0.01 M sodium py o-
phospha e and 0.25 M suc ose, made up wi h dis illed wa e and adjus ed o 7.5 pH; all o hem om Sigma; S . Louis, MO, USA) and
eshly added p o ease inhibi o cock ail able s (Roche Diagnos ics; Indianapolis, IN, USA). The p o ein concen a ion was de e -
mined by he B ad o d Me hod (P o ein assay dye concen a e, Bio-Rad Labo a o ies; He cules, CA, USA), and he o al p o ein con-
en o he issues was calcula ed. The p o ein lysa es we e subjec ed o SDS-PAGE, elec o ans e ed o poly inylidene di luo ide
memb anes (PVDF; Millipo e; Bille ica, MA, USA) wi h a semid y blo e and p obed wi h an ibodies agains GRP78 (Cell Signaling;
Dan e s; MA, USA); ATF6a, CHOP, peIF2a(Se 52), pPERK (Th 981) (San a C uz; San a C uz, CA,USC); pIREa(Se 724), SPTLC1,
SPTLC2, UCP1 (Abcam; Camb idge, UK); a- ubulin o b-ac in (Sigma; S . Louis, MO, USA) as desc ibed (Lo
´pez e al., 2010; Va ela
e al., 2012; Ma ı
´nez de Mo en in e al., 2014; Con e as e al., 2014; Ma ı
´nez-Sa
´nchez e al., 2017a; Seoane-Collazo e al., 2018).
Cell Repo s 25, 413–423.e1–e5, Oc obe 9, 2018 e4
Each memb ane was hen incuba ed wi h he co esponding seconda y an ibody: an i-mouse o an i- abbi (all o hem om DAKO;
Glos up, Denma k). The memb anes we e exposed o an X- ay ilm (Fuji ilm; Tokyo, Japan) and de eloped using de elope (De el-
ope G150; AGFA Heal hCa e: Mo sel, Belgium) and Fixa o (Manual Fixing G354; AGFA Heal hCa e: Mo sel, Belgium).
Au o adiog aphic ilms we e scanned and he bands signal was quan i ied by densi ome y using ImageJ-1.33 so wa e (NIH;
Be hesda, MD, USA). Values we e exp essed in ela ion o b-ac in (hypo halamus) o a- ubulin (BAT). Rep esen a i e images o
all p o eins a e shown; in he case o he loading con ols a ep esen a i e gel is displayed o cla i y, al hough each band o each
p o ein was always co ec ed by i s own in e nal con ol band (b-ac in o a- ubulin). In all he Figu es showing images o gels, all
he bands o each pic u e come always om he same gel, al hough hey may be spliced o cla i y.
QUANTIFICATION AND STATISTICAL ANALYSIS
ImageJ-1.33 so wa e (NIH; Be hesda, MD, USA) was used o p ocess wes e n blo analyses and li e his ologic images (Lo
´pez e al.,
2010; Va ela e al., 2012; Ma ı
´nez de Mo en in e al., 2014; Con e as e al., 2014; Ma ı
´nez-Sa
´nchez e al., 2017a, b; Seoane-Collazo
e al., 2018). FLIR-Tools-So wa e; FLIR; Wes Malling, Ken , UK) was used o p ocess he he mal images (Whi le e al., 2012; Ma -
ı
´nez de Mo en in e al., 2012; Con e as e al., 2014, 2017; Ma ı
´nez de Mo en in e al., 2014, 2015; Ma ins e al., 2016; Ma ı
´nez-
Sa
´nchez e al., 2017a; Seoane-Collazo e al., 2018). S a is ical analysis was conduc ed using G aphPad P ism 6 So wa e (G aphPad
So wa e, Inc.; La Jolla, CA, USA). Da a a e exp essed as mean ±SEM. P o ein da a we e exp essed in ela ion (%) o con ol (Sham
o OVX, ehicle o GFP ea ed) a s. E o ba s ep esen SEM. S a is ical signi icance was de e mined by S uden ’s es (when wo
g oups we e compa ed) o ANOVA (when mo e han wo g oups we e compa ed) ollowed by pos hoc Bon e oni es . p < 0.05 was
conside ed signi ican . The numbe o animals used in each expe imen al se ing and analysis a e speci ied in each igu e legend.
e5 Cell Repo s 25, 413–423.e1–e5, Oc obe 9, 2018
