decrease (as observed in pre-menopausal
and menopausal states), this control of food
intake would be attenuated and MDP’s ano-
rexic effects via Nod2 hypothalamic neurons
would play a more prominent role in appetite
control.
In addition to the well-studied control of
host immune response by Nod2, our work
highlights the need to consider the effects
of Nod2 activation on brain neuronal activity
in the context of using Nod2 ligands as a
potential therapeutic tool. As PG-derived me-
tabolites are already used in clinics for cancer
therapies and planned for other applications
( 43 ), a better understanding of the physio-
logical roles of Nod2 and its ligands is there-
fore extremely important. Our work reveals
a sex- and age-dependent pathway of gut-
brain cross-talk, which may open additional
avenues for the treatment of neurological and
metabolic disorders.
Materials and methods
Animals
Nod2flox/flox( 44 ),Nod2tm1Jhgt(Nod2GFP)( 45 ),
andCamK2A-iCRE( 46 )miceweregenerously
provided by P. Rosenstiel (University of Kiel),
J. P. Hugot (Université Paris Diderot), and
R. Liblau (Université Toulouse III), respectively.
Nod2tm1Flv(Nod2KO),Gt(ROSA)26Sortm1(EYFP)Cos
(ROSAYFP), andSlc32a1tm2(cre)Lowl(Vgatcre) mice
were purchased from Jackson Laboratories.
These lines were interbred and/or backcrossed
with C57Bl/6JRj at Institut Pasteur animal
facilities to obtain the final strains described.
Mice were maintained at Institut Pasteur ani-
mal facilities under specific pathogen-free con-
ditions, with lights on at 7 a.m. and off at
9p.m.Agesandsexesofmicearespecifiedin
the text or figure legends. Animal care and ex-
perimentation were approved by the com-
mittee on animal experimentation of Institut
Pasteur (project DAP200025) and by the French
Ministry of Research (MESR project 26737).
Immunofluorescence
To prepare brains for immunofluorescence
imaging, mice were deeply anesthetized and
intracardially perfused with 1× PBS for 5 min,
then with 4% paraformaldehyde (PFA) in
0.1 M phosphate buffer for 10 min. The brains
were removed and stored in PFA at 4°C until
the following day. They were then washed
three times with PBS and cryoprotected in
30% sucrose for 48 hours. Brain sections 60 μm
thick were cut on a microtome. For gut im-
munofluorescence imaging, mice were killed
by cervical dislocation and the small intestine
was removed and placed in HBSS Mg2+Ca2+
(Gibco) + 5% FCS (complete media). The in-
testine was incised longitudinally and the
luminalcontentswerewashedawayusing
complete media. The muscularis externa was
then carefully removed from the underlying
mucosa. The muscularis and mucosal tissue
were pinned down on a plate coated with
silicone (Smooth-ON; 83750B) and then fixed
for 15 to 30 min with 4% PFA. Whole-mount
samples were then permeabilized in 0.1%
Triton X-100 for 1 to 2 hours. Brain slices
and whole-mount intestinal samples followed
the same staining procedure. Briefly, the tissue
was rinsed, incubated in blocking buffer 10%
normal serum + 0.1% Triton X-100 for 2 hours.
It was then incubated with primary antibodies
diluted in 1% normal serum + 0.1% Triton
X-100 at appropriate concentrations and in-
cubated overnight at 4°C. The following day
the tissue was washed three times in 1× PBS
and then incubated in 1% normal serum +
0.1% Triton X-100 with secondary antibodies.
Samples were again washed once in 1× PBS,
once in 1× PBS + DAPI (1:5000), and then once
again in 1× PBS. Finally, tissue was mounted
on slides with Fluoromount-G, cover slip ad-
ded and sealed. Images were captured with a
confocal laser-scanning microscope (LSM 700,
Zeiss) using the 10×, 25×, or 40× Zeiss ob-
jectives [Plan-Apochromat 10×/0.3; I LCI Plan-
Neofluar 25×/0.8 Imm Korr DIC M27; LD
Plan-Neofluar 40×/1.3 Oil DIC (UV) VIS-IR
M27], or a microscope equipped with Apotome
system (Zeiss), a digital camera (Hamamatsu
ORCA-ER C4742-80), and Axiovision 4.8 soft-
ware (Zeiss), using the 10× and 20× Zeiss ob-
jectives (Fluar 10×/0.5 M27; Plan-Apochromat
20×/0.8 M27). Images were adjusted post hoc
in ImageJ. The following primary antibodies
were used: anti-GFP (chicken; 1:500; Millipore
06-896); anti-NeuN (mouse; 1:100; Millipore
MAB377); anti-Iba1 (rabbit; 1:500; Wako 019-
19741); anti-CD31 (rat; 1:50; BD Pharmaceuti-
cal 550274); anti-bIII Tubulin (rabbit; 1:1000;
Cell signaling D71G9); anti-AgRP (goat; 1:2000;
Neuromics GT15023); and anti-POMC (rabbit;
1:1,000, Phoenix Pharmaceuticals H02930).
The following secondary antibodies from
Molecular Probes were used at 1:1000: Alexa
Fluor 488–conjugated goat anti-chicken IgG
(A11039); Alexa Fluor 568-conjugated goat anti-
mouse IgG (A21134); Alexa Fluor 568 or 647–
conjugated goat anti-rabbit IgG (A11036 or
A21244); and Alexa Fluor 647–conjugated goat
anti-rat IgG (A21449).Cell counting
Brain images were captured using a micro-
scope equipped with Apotome system (Zeiss), a
digital camera (Hamamatsu ORCA-ER C4742-
80) and Axiovision 4.8 software (Zeiss), using
the 10× (Fluar 10×/0.5 M27) or 20× (Plan-
Apochromat 20×/0.8 M27) Zeiss objectives.
Cell counting was performed manually using
Fiji or the Icy open-source platform to draw
regions and mark counted cells (http://icy.
bioimageanalysis.org). To count GFP+cells in
the cortex, striatum, thalamus, and hypothala-
mus,threefemalemicebrainswerestained
with GFP, NeuN, and Iba1. For each animal,
three to six random spots were selected and
a minimum of 100 NeuN+and 100 Iba1+cells
were then counted. Of those, the number of
GFP+cells was counted. Neuron size was
obtained by measuring a line drawn between
the farthest points in the neuronal cell body
using Fiji (9 to 12 neurons per animal). To
count the GFP+neurons in the ARC and DMH,
12 brain sections from eachNod2GFPmouse
were selected. The slices were stained with
GFP, NeuN, and AgRP (the latter being used
to help determine the area of interest). The
sex and age of the mice were revealed only
after quantification. Data are shown as means
per animal.RNAscope in situ hybridization
RNAscope in situ hybridization was performed
using mouse probes (Bio-Techne) against
Nod2(Mm-Nod2, #433391),Vgat(Mm-Slc32a1-
C2, #319191-C2), andNpy(Mm-Npy-C2, #313321-
C2) on 15-mm sections of brain tissue according
to the manufacturer’s instructions.Radiolabeling and purification ofLactobacillus
rhamnosusstrain Lr32 peptidoglycan
L. rhamnosusstrain Lr32 was inoculated into
500 ml of 25% strength BD Difco MRS broth
supplemented with 100mM GlcNAc (Sigma-
Aldrich). For radiolabeling, 10mCi per liter of
[^14 C]N-acetylglucosamine (^14 C-GlcNAc) was ad-
ded. A nonlabeled control culture of Lr32 was
prepared in parallel. Cultures were incubated
overnight at 37°C without aeration and har-
vested at an OD 600 of 2.5. Peptidoglycan (PG)
was purified according to standard protocols
for Gram-positive bacteria ( 47 ). Purified PG
was stored short-term at–20°C. Approximate-
ly 40 mg of Lr32 PG was solubilized by re-
suspension with 12.5 mM NaH 2 PO 4 pH 5.6
and incubated overnight at 37°C with 50 U of
Streptomyces globisporusATCC 21553 muta-
nolysin (Merck) per milligram of PG. MiceGabanyiet al.,Science 376 , eabj3986 (2022) 15 April 2022 8 of 12
24 hours (n=4or5).(L) Nest-building test. Amount of cotton used to build
the nest (unrolled cotton) is shown (n=9or10).(MandN) Virus-injected
mice treated with ABX. (M) Body weight weekly measures during and after
ABX treatment (females, 2 to 4 months at week 0;n= 13 per group). (N)
Food eaten in 40 hours at week 13 (during ABX treatment) and week 24 (after
ABX treatment) (n= 7 per group). Data are averages ± SEM. *P≤0.05
[unpairedttest in (A), (J), and (K); two-way ANOVA in (I), (M), and (N);
FischerÕs exact test in (L)]. Abbreviations: ARC, arcuate nucleus of the
hypothalamus; DMH, dorsomedial nucleus of the hypothalamus; PHY,
perihypoglossal nuclei; ABX, antibiotics.RESEARCH | RESEARCH ARTICLE