Roche). For insulin measurements, samples
were immediately placed on ice and centri-
fuged at 14,000gfor 10 min at 4°C as soon as
possible. Serum was collected and transferred
into a new clean tube. Insulin concentration
was determined using an ELISA kit (Mercodia,
Uppsala, Sweden) according to the manufac-
turer’s instructions. Homeostatic Model Assess-
ment for Insulin Resistance (HOMA-IR) was
calculated by multiplying fasting plasma insu-
lin (FPI) by fasting plasma glucose (FPG), then
dividing by the constant 22.5; HOMA-IR =
(FPI × FPG)/22.5.
iDISCO/ClearMap
Mice were single-housed around 7 p.m. and
were gavaged the next day at 10 a.m. with
MDPorMDPctr(1mgin0.1mlofLALwater).
Mice were perfused 3 hours after gavage. The
iDISCO and ClearMap protocol was followed
as detailed on the websitehttp://idisco.info
( 51 ). Whole brains were stained with anti–
c-Fos (rabbit; 1:1500; Synaptic Systems 226-
003) and Alexa Fluor 647–conjugated donkey
anti-rabbit IgG (1:1000; Thermo Fisher A-31573).
Stereotaxic surgery
Mice were deeply anesthetized after an intra-
peritoneal injection of a mixture of ketamine
(Imalgene 1000, 150 mg/kg) and xylazine
(Rompun 2%, 5 mg/kg) and injected sub-
cutaneously with buprenorphine (Buprecare;
0.1 mg/kg). After their heads were shaved, the
mice were placed in a stereotaxic apparatus.
The skin was sterilized and local anesthesia
(lidocaine, Xylovet, 2%) was injected sub-
cutaneously before the skin was incised. A
small craniotomy was performed and the viral
solution was then injected into DMH and/or
ARC through a glass micropipette attached
to a nanoinjector system (Nanoject III; 100 nl
per site; 1 nl/s) using stereotaxic coordinates
(–1.7 mm anterior from bregma, 0.3 mm lateral,
and at a depth of 5 mm for DMH and 5.9 mm
for ARC from the skull surface at bregma).
Recombinant AAV expressing GCaMP6f
(AAV9.hSyn.Flex.GCaMP6f; ~10^13 VG/ml),
GCaMP7f (AAV9.hSyn.Flex.GCaMP7f; ~10^13 VG/
ml), Cre recombinase (AAV9.hSyn.Cre; ~5 ×
1011 VG/ml), mCherry (AAV9.hSyn.Flex.mCherry;
~10^13 VG/ml), and TdTomato (AAV9.CAG.Flex.
Tdtomato; ~10^13 VG/ml) were purchased from
Addgene Vector Core.
Fiber photometry
A fiber photometry system was developed based
on ( 52 ) with modifications. Immediately after
viral injection in the right ARC nucleus, a
single optical fiber (multimode, 430mm in
diameter, NA 0.50, LC zirconia ferrule, Doric
Lenses) was implanted 100mmabovethevirus
injection site and fixed to the skull with a
liquid bonding resin (Superbond, Sun Medi-
cal) and dental acrylic (Unifast). At least 3 weeks
after injection, GCaMP was continuously ex-
cited using a 473-nm diode-pumped solid-state
(DPSS) laser (output fiber intensity, 0.1 mW;
Crystal Lasers) reflected on a dichroic filter
(452 to 490 nm/505 to 800 nm, MD498 GFP,
Thorlabs) and collimated into a 400-mm multi-
mode optical fiber (NA, 0.50; Doric Lenses)
with a convergent lens (f = 30 mm). The emit-
ted fluorescence was collected in the same fiber
and transmitted by the dichroic mirror, filtered
(525 ± 19 nm) and focused on a NewFocus
2151-fW photoreceptor (Newport; DC mode).
Reflected blue light along the light path was
also measured with another amplified photo-
detector (PDA36A, Thorlabs) for monitoring
light excitation and fiber coupling. Signals
from both photodetectors were digitized at
5000 Hz by a digital-to-analog converter (DAC;
Power 1401, CED), and recorded and analyzed
using Spike2 software. For simultaneous dual-
color photometry, a 473-nm DPSS laser and
a 561-nm DPSS laser (output fiber intensity,
0.1 mW; Crystal Lasers) were reflected via a
dichroic filter (MD498 GFP, Thorlabs) onto a
488/561-nm dual-edge dichroic beam split-
ter (Di01-R488/561, Semrock) and collimated
into a 400-mm multimode optical fiber (NA,
0.50, Doric Lenses) with a convergent lens (f =
30 mm). The emitted fluorescence was col-
lected in the same fiber and separated by a
510-nm dichroic mirror, filtered (525 ± 19 nm
for GCaMP and 624 ± 75 nm for mCherry) and
focused on two respective NewFocus 2151-fW
photoreceptors (Newport; DC mode). Animals
were habituated to the fiber optic patch cord
(Doric Lenses) and to the connection proce-
dure at least three times before the recording
sessions. Recordings with low fluorescence
signals(meanfluorescence<0.15mVand/or
spontaneous events inDF/F<1%) or display-
ing movement/laser artifacts were discarded
from the analysis. Animals were brought into
the recording room and singly housed around
7 p.m. the night before the recording session.
For the analysis under fasting conditions, food
was removed when mice were brought into
the recording room. For the analysis of steady-
state(fed)conditions,micewerefedadlibi-
tum at all times except during the recording
session. Water was available in both condi-
tions.Onthefollowingdayataround9a.m.,
food was removed for the fed state and mice
were recorded in the home cage for ~60 min
to establish baseline signal. Then mice were
either presented with a food pellet and re-
corded for ~60 min or gavaged with MDP or
MDPctr (1 mg in 0.1 ml) and recorded for
~150 min. After this period, gavaged mice
were presented with a food pellet and re-
corded for an additional 30-min period. For
signal analysis of spontaneous activity, raw
fluorescence signals were normalized (DF/F)
to the mean fluorescence (20-s sliding win-
dow) and then smoothened (0.05-s window).The standard deviation of theDF/Fbaseline
signal was calculated and all the calcium-
positive transients above 1.5 SD were auto-
matically extracted with custom scripts (Spike2,
CED, UK; minimum time interval between
events of 5 s). The frequency of the sorted
spontaneous events was then reported as the
number of events per minute and normalized
to the baseline activity before gavage. Non-
specific modulation of fluorescence signals
was examined using simultaneous dual-color
photometry of the calcium-sensitive fluorescent
reporter GCaMP7f and the calcium-insensitive
fluorescent report mCherry. mCherry traces
were qualitatively different from GCaMP, with
no noticeable fluctuation with the same laser
power configuration (peak mCherryDF/F<
1%) and no concomitant/correlated events
on both channels. To quantify the absence of
synchronous artifacts in both GCaMP and
mCherry channels, we sorted GCaMP events
and calculated a GCaMP event-triggered
mCherry trace average. We also calculated for
each sorted GCaMP event the meanDF/F(1-s
window around event) and the peak ampli-
tude in both GCaMP and mCherry channels.
With these quantifications, we did not observe
any correlation between GCaMP and mCherry
channels, ruling out any significant contribu-
tion of nonspecific events in GCaMP signals.
Animals in which post hoc histological exam-
ination showed that viral injection or im-
planted optic fiber were not in the correct
location were excluded from analysis. Inves-
tigators were not blinded to the group iden-
tity, but automatic analysis routines were
performed with the same script executed for
all experimental groups.Patch-clamp recordings
VgatcreNod2GFPmice previously injected with
a Cre-dependent Td-tomato virus in the ARC
were deeply anesthetized with intraperitoneal
injection of ketamine (100 mg/kg) and xyla-
zine (10 mg/kg) and swiftly decapitated. The
brain was rapidly dissected and placed in ice-
cold artificial cerebrospinal fluid (ACSF) con-
taining 124 mM NaCl, 3 mM KCl, 2.8 mM
MgSO 4 , 26 mM NaHCO 3 , 1.25 mM NaHPO 4 ,
20 mM glucose, and 0.5 mM CaCl [∼310 mOsm,
pH 7.3 when bubbled with a mixture of 95%
O 2 and 5% (vol/vol) CO 2 ] (all chemicals from
Sigma-Aldrich). Coronal slices (250mm thick)
of the whole brain were placed in bubbling
ACSFinawarmbathat35°Cfor30minand
then allowed to recover at room temperature
(i.e., 22° ± 1°C). For whole-cell recordings, in-
dividual slices were placed in a chamber
mounted on a Zeiss Axioskop upright micro-
scope and continuously perfused (1.5 ml/min)
with 30°C ACSF (Warner Instrument inline
heater) containing 124 mM NaCl, 3 mM KCl,
1.3 mM MgSO 4 , 26 mM NaHCO 3 , 1.25 mM
NaHPO 4 ,20mMglucose,and2mMCaCl.Gabanyiet al.,Science 376 , eabj3986 (2022) 15 April 2022 10 of 12
RESEARCH | RESEARCH ARTICLE