Article
The Sense of Smell Impacts Metabolic Health and
Obesity
Graphical Abstract
Highlights
d Loss of adult olfactory neurons protects against diet-induced
obesity
d Loss of smell after obesity also reduces fat mass and insulin
resistance
d Loss of IGF1 receptors in olfactory sensory neurons (OSNs)
improves olfaction
d Loss of IGF1R in OSNs increases adiposity and insulin
resistance
Authors
Celine E. Riera, Eva Tsaousidou,
Jonathan Halloran, ..., Andreas Stahl,
Jens C. Br€uning, Andrew Dillin
Correspondence
bruening@sf.mpg.de (J.C.B.),
dillin@berkeley.edu (A.D.)
In Brief
Riera et al. unravel a new bidirectional
function for the olfactory system in
energy homeostasis. Ablation of olfactory
sensory neurons (OSNs) in mice protects
them from diet-induced obesity
accompanied by increased
thermogenesis, while loss of IGF1
signaling in OSNs leads to increased
adiposity and insulin resistance.
Riera et al., 2017, Cell Metabolism 26, 198–211
July 5, 2017 ª 2017 Elsevier Inc.
http://dx.doi.org/10.1016/j.cmet.2017.06.015
Cell Metabolism
Article
The Sense of Smell Impacts
Metabolic Health and Obesity
Celine E. Riera,1,2,3,10 Eva Tsaousidou,4,5,6,7,10 Jonathan Halloran,1,2 Patricia Follett,8 Oliver Hahn,6
Mafalda M.A. Pereira,4,5 Linda Engstro¨ m Ruud,4,5 Jens Alber,4,5 Kevin Tharp,9 Courtney M. Anderson,9 Hella Bro¨ nneke,4,5
Brigitte Hampel,4,5 Carlos Daniel de Magalhaes Filho,8 Andreas Stahl,9 Jens C. Br€uning,4,5,6,* and Andrew Dillin1,2,11,*
1Howard Hughes Medical Institute and Department of Molecular and Cell Biology
2The Paul F. Glenn Center for Aging Research
University of California, Berkeley, Berkeley, CA, USA
3Diabetes and Obesity Research Institute, Cedars-Sinai Medical Center, 8700 Beverly Boulevard, Los Angeles, CA, USA
4Department of Neuronal Control of Metabolism, Max Planck Institute for Metabolism Research, Gleueler Strasse 50, Cologne, Germany
5Center for Endocrinology, Diabetes and Preventive Medicine (CEDP), University Hospital Cologne, Kerpener Strasse 26, Cologne, Germany
6Max Planck Institute for Biology of Ageing, Cologne, Germany and Excellence Cluster on Cellular Stress Responses in Aging Associated
Diseases (CECAD) Cologne, Germany
7Department of Genetics and Complex Diseases and Sabri U¨ lker Center, Harvard T.H. Chan School of Public Health, Boston, MA, USA
8The Salk Institute for Biological Studies, 10010 North Torrey Pines Road, La Jolla, CA, USA
9Nutritional Sciences and Toxicology, University of California, Berkeley, Berkeley, CA, USA
10These authors contributed equally
11Lead Contact
*Correspondence: bruening@sf.mpg.de (J.C.B.), dillin@berkeley.edu (A.D.)
http://dx.doi.org/10.1016/j.cmet.2017.06.015
SUMMARY
Olfactory inputs help coordinate food appreciation
and selection, but their role in systemic physiology
and energy balance is poorly understood. Here we
demonstrate that mice upon conditional ablation of
mature olfactory sensory neurons (OSNs) are resistant
to diet-induced obesity accompanied by
increased thermogenesis in brown and inguinal fat
depots. Acute loss of smell perception after obesity
onset not only abrogated further weight gain but
also improved fat mass and insulin resistance.
Reduced olfactory input stimulates sympathetic
nerve activity, resulting in activation of b-adrenergic
receptors on white and brown adipocytes to promote
lipolysis. Conversely, conditional ablation of the IGF1
receptor in OSNs enhances olfactory performance in
mice and leads to increased adiposity and insulin
resistance. These findings unravel a new bidirectional
function for the olfactory system in controlling
energy homeostasis in response to sensory and hormonal
signals.
INTRODUCTION
The regulation of whole-body energy homeostasis relies on an
intricate balance between food intake and energy expenditure.
This balance requires the coordinated response of peripheral
and central neuronal inputs including hormones, multiple peptides,
and neurotransmitters. Notably, these peripheral factors
include leptin, insulin, ghrelin, glucagon-like peptide-1 (GLP-1),
insulin-like growth factor-1 (IGF1), and cholecystokinin (Belgardt
and Br€uning, 2010; Clemmons, 2004; Williams and Elmquist,
2012). In order to adapt quickly to variations in environmental
conditions and maintain global body homeostasis, mammalian
systems have developed neuronal circuits within the central nervous
system (CNS) that integrate internal cues into autonomic
responses via the sympathetic and parasympathetic nervous
systems (Bartness and Song, 2007; Marino et al., 2011). The
CNS is responsible for engaging these autonomic systems to
control many vital physiological functions, such as pancreatic
secretion, lipid storage, thermogenesis, peripheral glucose uptake,
and hepatic glucose flux (Marino et al., 2011; Vogt and
Br€uning, 2013). The variations in autonomic tone are orchestrated
by extensive and often reciprocal connections between
the hypothalamus and brainstem nuclei (Grill and Hayes, 2012;
Skibicka and Grill, 2009; Spencer et al., 1990). In the hypothalamus,
the melanocortin system in the arcuate nucleus (ARC) controls
feeding in response to circulating insulin and leptin levels
(Cowley et al., 2001; Fan et al., 1997; Gropp et al., 2005; Hahn
et al., 1998; Huszar et al., 1997; Luquet et al., 2005). Leptin promotes
energy expenditure via increasing sympathetic nerve activity
and subsequent catecholamine signaling in white adipose
tissue (WAT) and brown adipose tissue (BAT), to promote lipolysis
and fatty acid oxidation (Bartness and Song, 2007; Collins
et al., 1996; Haynes et al., 1997; Lafontan and Langin, 2009; Rahmouni
and Morgan, 2007). In addition, cold stimulation or pharmacological
activation of the b-adrenergic pathway enhances
the formation of brown-like or beige adipocytes predominantly
in the inguinal fat pads (iWAT) of rodents, which are characterized
by a thermogenic gene expression program similar to that
of brown adipose tissue, including high level expression of mitochondrial
uncoupling protein 1 (UCP1) (Cousin et al., 1992;
Fisher et al., 2012; Petrovic et al., 2010; Seale et al., 2011; Tsukiyama-Kohara
et al., 2001; Wu et al., 2012; Young et al., 1984).
Among the many sensory stimuli that influence behavioral
decisions about food choice, olfactory inputs are likely to
198 Cell Metabolism 26, 198–211, July 5, 2017 ª 2017 Elsevier Inc.
contribute to the regulation of energy homeostasis. Remarkably,
the sensory perception of a hidden food cue, without its ingestion,
at least transiently switches the activation state of AgRP
and POMC neurons (Betley et al., 2015; Chen et al., 2015). In
mice and other rodents, the hypothalamus receives indirect inputs
from OSNs through signals entering from the main olfactory
bulb (MOB) and transmitted to the centers of the olfactory cortex
(Bjorklund et al., 1996; Gascuel et al., 2012). Therefore, olfactory
signals may prime the activity of key homeostatic neurons in the
hypothalamus to adapt systemic metabolism under conditions
of anticipated food intake.
Nutritional status dynamically and profoundly impacts olfactory
sensitivity, suggesting the existence of a reciprocal communication
between OSNs and the CNS. For example, olfactory
acuity is stimulated in the fasted state (Albrecht et al., 2009;
Cameron et al., 2012; Soria-Go´ mez et al., 2014), whereas satiety
corresponds to reduced olfactory sensitivity (Critchley and Rolls,
1996). Interestingly, the main olfactory epithelium (MOE) and
MOB express high levels of mRNAs for receptors of anorexiasignaling
hormones, such as leptin (Ob-R), insulin (IR), and
IGF-1, as well as ghrelin receptors (Baskin et al., 1983; Elmquist
et al., 1998; Hill et al., 1986). These findings point to the possibility
that these hormonal signals not only directly regulate hypothalamic
control of peripheral metabolism but might also modulate
the sensory perception of the environment via the control of
olfaction.
We investigated the role of OSNs in the control of energy balance.
To this end, we examined the consequence of genetically
ablating the ability of animals to smell, by disrupting OSNs, on
whole-body energy homeostasis in lean and obese animals.
We find that mice with reduced olfaction, i.e., hyposmia, are
leaner upon diet-induced obesity (DIO) either before or after
the onset of obesity. These animals exhibit increased energy
expenditure and enhanced fat burning capacity as a consequence
of enhanced sympathetic nerve activity in BAT and
iWAT. Conversely, we describe that conditional ablation of the
IGF1 receptor in OSNs results in enhanced olfactory perception.
Complementing the results observed in the hyposmic animals,
these hyperosmic mice have increased adiposity and insulin
resistance. Collectively, the results reveal a critical role for olfactory
sensory perception in coordinately regulating peripheral
metabolism via control of autonomic innervation.
RESULTS
Genetic Disruption of OSNs Is Associated with
Resistance to Diet-Induced Obesity
To explore the possible relationship between olfactory perception
and energy homeostasis, we characterized the metabolic
consequences of smell disruption in mice. As defects in olfactory
perception at birth lead to neonatal lethality in mice (Belluscio
et al., 1998; Brunet et al., 1996), we conditionally ablated mature
OSNs in adult mice by expressing the inducible diphtheria toxin
receptor (DTR) in OSNs using the olfactory marker protein (OMP)
promoter (Buch et al., 2005; Buiakova et al., 1996; Chen et al.,
2005; Gropp et al., 2005; Li et al., 2004), generating OMPDTR
and control animals. OMP labels OSNs from the MOE and the
vomeronasal organ (VNO), an organ specialized in pheromone
detection to discriminate among gender- and species-specific
cues involved in the control of mating and aggressive behavior
(Dulac and Torello, 2003). As the DTR transgene of the OMPDTR
model is integrated in the genome but only expressed upon Credependent
recombination, we further verified the specificity of
the OMP promoter by crossing OMP-Cre to a Rosa26 line that
expresses tdTomato (RFP) in all Cre-expressing cells (Madisen
et al., 2010). Robust RFP detection was detected in OSNs present
in the MOE (Figure S1A), the VNO (data not shown), and
OSNs axons projecting to the MOB (Figure S1B) but absent in
non-olfactory tissues reported to express OMP transcripts
such as the thymus, the thyroid, and the bladder (Kang et al.,
2015) (Figure S1A). However, weaker OMP expression was detected
in discrete cells of the anterior hypothalamus and nerves
from the optic chiasm and the cerebellum (Figure S1B). RFP
expression overlapped with a polyclonal antibody against OMP
in the MOE (Figure S1E), which was not observed in the CNS
or non-olfactory tissues.
Ablation of OSNs in OMPDTR
mice was induced in 7-week-old
animals by diphtheria toxin (DT) injection and validated via
assessment of imunoreactive OSNs in the MOE and the VNO using
anti-OMP antibodies (Figures 1A and S2A). Such reduction in
OSN numbers was not observed upon DT treatment of control
animals (DTR+/À
littermates). OSNs regenerate from stem cell
pools which are not OMP positive and have a half-life averaging
4.2 weeks (Chen et al., 2005; Gogos et al., 2000; Margolis et al.,
1991), requiring repeated DT-injections every 3 weeks to maintain
a state of OSN ablation for chronic studies. Functional olfactory
habituation and exploration tests confirmed that OMPDTR
mice had a strong impairment in olfactory discrimination of
attractive food-based and social odors but still retain some
olfactory capacity rendering these animals hyposmic (Figures
1B, S2B, and S2C).
We next assessed whether OSN ablation affected body weight
under normal chow diet feeding conditions in lean mice as well
as upon induction of DIO upon high-fat-diet feeding. After one
DT injection, body weight was reduced in the OMPDTR
mice in
paired cohorts under normal chow and high-fat diet (HFD), and
repeated in an independent experiment (Figures 1C, 1F, and
S3A). Strikingly, OMPDTR
mice displayed a substantial reduction
in body weight (16%) compared to controls upon HFD, as
confirmed in an independent cohort (Figures 1F, S3A, and
S3B). Food intake was unchanged upon normal chow but was
reduced in hyposmic mice upon HFD feeding (Figures 1D, 1E,
1G, and 1H), but without affecting the diurnal feeding pattern
(Figure 1I). Importantly, the attenuation of smell did not impact
the ability of mice to feed after overnight fasting, and food intake
was even enhanced in the first hour of refeeding (Figure S3C).
Therefore, the lack of olfactory input leads to a lean phenotype
that is not due to a loss of appetite. The reduction of body weight
observed in OMPDTR
mice reflected a loss in fat mass without
alterations in lean mass upon HFD-induced obesity (Figure 1J).
In particular, the subcutaneous inguinal fat and the visceral
gonadal fat pads were smaller in OMPDTR
mice, whereas the
composition of other tissues was unaffected (Figure 1K).
Intranasal Ablation of OSNs Slows Weight Gain in
DIO Mice
Because Cre recombinase expression is not exclusively
restricted to OSNs, we investigated whether the DT injections
Cell Metabolism 26, 198–211, July 5, 2017 199
were causing cellular ablation of certain neuronal subsets in nonolfactory
regions. When administered every 3 weeks, DT had
very limited effects in brain areas other than OSNs projections
into the MOB (Figure S4A) and did not cause abnormal gliosis
(Figures S4B and S4C). However, as the DT-mediated cell death
model may ablate a few discrete neurons expressing Cre, which
could play a role in the lean phenotype observed, we investigated
an alternative route to specifically ablate OSNs in the naris.
We engineered an adenoviral vector Ad-flex-ta-Casp3-TEVp
based on published results of adeno-associated virus (Wu
et al., 2014; Yang et al., 2013) for delivery into OMP-Cre animals
(OMP-Cre+) and control littermates (OMP-CreÀ). Ad-flex-taCasp3-TEVp
was administered simultaneously with Ad-GFP.
Viral delivery was achieved by intranasal injections in order to
A
OMPGαolf
OMP
DTR
Control
MOE MOE
C
Average chow FI (g/day)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
**
Average HFD FI (g/day)
0
2
4
6
8
10
HFD cumulative FI (g)
day night day night day night
**
F
0
5
10
15
20
**
Fatmass(g)
0
5
10
15
20
25
Leanmass(g)
iW
A
TgW
A
TpW
A
T
Pancreas
LiverSpleenK
idneys
H
eartLungs
B
rain
0
1
2
3
4
***
***
Organweight(g)
Control
OMPDTR
Control
OMPDTR
H
I J
0
50
100
150
200
**
OMP+cells
Control
OMPDTR
K
Control
OMP DTR
8 10 12 14 16 18
0
2
4
6
8
Age (Weeks)
Foodintake(g)
DT DT DT DT
HFD HFD
Chow control diet
D E
G
Control
OMPDTR
Control
OMPDTR
Control
OMP DTR
8 10 12 14 16 18 20 22
0
2
4
6
8
Age (Weeks)
Foodintake(g)
DT DT DT DT DT
** *
6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
50
DT
DT
DT
Age (Weeks)
Weight(g)
DT
***
OMP DTR
Control
6 8 10 12 14 16 18 20 22
0
5
10
15
20
25
30
35
40
45
50
Control
OMP DTR
***
DT
DT
DT
DT
DT
Age (Weeks)
Weight(g)
Chow control diet
Timetofindfood(sec)
0
200
400
600
800 **
Food finding test
B
Control
OMPDTR
0
1
2
3
4
5
Figure 1. OMPDTR
Are Lean and Resistant to Diet-Induced Obesity
(A) DT ablates OMP+ neurons in the MOE of OMPDTR
, but not controls, a- OMP (red), Ga olf (green), or dapi (blue). Scale bar, 50 mm. Quantification is in right panel.
(B) Average time needed to find hidden food pellets for random fed mice.
(C–E) Weight gain and weekly and average daily food intake of DT-injected OMPDTR
and controls on chow diet (two-way ANOVA for C and D).
(F–H) Weight gain and weekly and average daily food intake of DT-injected OMPDTR
and controls when maintained on HFD (two-way ANOVA for F and G).
(I) Feeding pattern of DT-injected OMPDTR
and controls on HFD (two-way ANOVA).
(J) Fat mass and lean mass of OMPDTR
and controls on HFD.
(K) Body composition of HFD-fed OMPDTR
and controls measured by organ weight (g).
Means ± SEM, n = 5–10, ***p < 0.001, **p < 0.001, *p < 0.05.
200 Cell Metabolism 26, 198–211, July 5, 2017
robustly target OSNs and increase the expression efficacy (Figure
2A), with no observed penetrance in the MOB or other brain
tissues (Figure S4E). OMP-Cre+ mice displayed reduced thickness
in the rostral area of the olfactory epithelium and decreased
number of OSNs in MOE sections compared to Cre- controls
(Figure 2A), and showed reduction in olfactory discrimination
compared to controls (Figures 2B, S4D, and S4E). When subjected
to DIO, these animals presented reduced body weight
and adiposity compared to control mice without differences in
lean mass (Figures 2C and 2D). Interestingly, this alternative
olfactory depletion method promoted a leaner phenotype than
controls without altering food intake (Figure 2E), suggesting
that olfactory inhibition leads to reduced obesity, but the degree
of hyposmia may impact food intake, as observed in the more
complete OMPDTR
ablation model (Figure 1). Ad-flex-ta-Casp3TEVp
also ameliorated glucose tolerance in OMP-Cre+ mice
(Figure 2F) and provided beneficial effects on insulin sensitivity
(Figure 2G). Based on the similarities observed between adenoviral
and genetic ablation, these findings suggest that the
observed lean phenotype in DIO animals is indeed a specific
consequence of reduced olfactory perception.
Hyposmia Is Concomitant to Enhanced Energy
Expenditure
From our results, hyposmia results in animals with reduced fat
mass despite the ingestion of a high caloric diet. Therefore, the
A
B
E F G
C D
Figure 2. Adenoviral Disruption of OSNs Validates the Lean OMPDTR
Phenotype
(A) Strategy to deliver Ad-GFP and Ad-flex-ta-Casp3-TEVp in OMP-CreÀ and OMP-Cre+ animals. Ablation of OMP-positive neurons in the MOE of OMP-Cre+
but not controls, a- OMP (red), a- GFP (green), and dapi (blue). Arrows indicate MOE regions where OSNs are depleted.
(B) Average time needed to find hidden food pellets for random fed mice injected with intranasal Ad-GFP and Ad-flex-ta-Casp3-TEVp.
(C–E) (C) Weight gain, (D) fat mass and lean mass, and (E) weekly food intake of Ad-delivered OMP-CreÀ and OMP-Cre+ on HFD diet (two-way ANOVA for
C and E).
(F and G) Glucose tolerance and insulin tolerance testing of HFD-fed mice after 6 weeks of DIO (two-way ANOVA).
Means ± SEM, n = 5–7, ***p < 0.001, **p < 0.001, *p < 0.05.
Cell Metabolism 26, 198–211, July 5, 2017 201
excessive calories are dissipated either in the form of energy utilization
or via bodily excretion. To test both hypotheses, we
analyzed energy expenditure and energy assimilation in the
HFD-fed OMPDTR
model. Oxygen consumption and energy
expenditure relative to lean body mass were markedly increased
in HFD-fed OMPDTR
mice compared to controls as measured
2 weeks post DT administration (Figures 3A and 3B). Strikingly,
there was no significant difference in physical activity and
energy assimilation measured by fecal energy loss between
the two genotypes (Figures 3C and S3D), indicating that the
reduced weight gain was mainly due to combined effects of
increased energy expenditure and a reduction in food intake
in HFD-fed OMPDTR
mice (Figures 3B and 1D), but not due to
differences in nutrient absorption. Furthermore, OMPDTR
mice
had decreased circulating free fatty acids and fasting leptin
levels (Figures 3D and 3E) and were protected against pathological
phenotypes typical of the effects of HFD feeding after
14 weeks of HFD treatment, such as hepatic steatosis (Figure 3F)
and visceral adipocyte hypertrophy (Figure 3G).
The development of insulin resistance due to DIO is mediated
in part by chronic inflammation in adipose tissue characterized
by the recruitment of macrophages and elevation of chemokine
expression in fat (Lumeng et al., 2007; Xu et al., 2003a). Accumulation
of macrophage infiltrates was visible by hematoxylin and
eosin staining in visceral fat depots of control mice under DIO
but was far less abundant in OMPDTR
mice after 14 weeks of
BA
Light cycle Dark cycle
0
200
400
600
Totalactivity
C
0.000
0.001
0.002
0.003
0.004
FFA(mmol/L)
*
0
20
40
60
80
Leptin(mg/mL)
*
0 20 40 60 80 100 120
0
100
200
300
400
500
600
***
time (min)
Bloodglucose(mg/dl)
0 20 40 60 80 100 120
0
20
40
60
80
100
120
***
time (min)
%Initialglucose
FD
G
H
I J
ns
ns
α
tnf
ccl2
ccl3
F4/80
C
D
68
tlr2
cxcl1
0
1
2
Foldchange
* *** * *
Insulin(ng/mL)
0
2
4
6
8
*
K
Control
OMPDTR
Control OMP
DTR
Control OMP
DTR
Control
OMP DTR
Insulin tolerance testGlucose tolerance test
Control
OMP DTR
Light cycle Dark cycle
1000
2000
3000
4000
**
*
VO2(mL/hr/Kg)
Light cycle Dark cycle
0
10
20
30
40
50
**
Heat(kcal/hrFFM)
*
E
Figure 3. Increased Energy Expenditure in OMPDTR
(A–C) Oxygen consumption, energy expenditure, and total physical activity in HFD-fed DT-injected OMPDTR
and control mice measured 2 weeks post DT delivery.
(D and E) (D) Free fatty acid and (E) leptin serum levels.
(F and G) Hematoxylin and eosin sections of liver and gonadal WAT in HFD-fed mice after 14 weeks of DIO. Scale bar, 40 mm.
(H) RTqPCR of pro-inflammatory cytokines and macrophage markers in gonadal WAT from HFD-fed mice after 14 weeks of DIO.
(I and J) Glucose tolerance and insulin tolerance testing of HFD-fed mice after 6 weeks of DIO (two-way ANOVA).
(K) Fasting serum insulin levels.
Means ± SEM, n = 6–11, ***p < 0.001, **p < 0.001, *p < 0.05.
202 Cell Metabolism 26, 198–211, July 5, 2017
DIO (Figure 3G). Similarly, HFD feeding induced expression of
markers of inflammation in visceral fat tissue but was largely
attenuated in OMPDTR
mice (Figure 3H). Consistent with
decreased inflammation, OMPDTR
mice had improved insulin
sensitivity, improved glucose tolerance, and lower fasting insulin
levels after 6 weeks of DIO (Figures 3I–3K). Taken together, hyposmia
is associated with protection against many of the negative
features caused by excessive caloric intake.
Disrupted Olfaction Is Linked to Adipose Tissue
Thermogenesis and Browning of Subcutaneous Fat
Because hyposmic mice were protected from the severity of
excessive caloric intake, we began to ask how these animals
were equilibrating the intake of calories with their peripheral
metabolism. BAT dissipates energy directly as heat through uncoupling
fatty acid oxidation from ATP production by increasing
UCP1 within mitochondria (Enerb€ack et al., 1997). While control
animals on prolonged HFD feeding showed lipid accumulation in
BAT, OMPDTR
mice exhibited healthy histological features of
BAT despite prolonged HFD feeding (Figure 4A). After 14 weeks
of DIO, the brown fat of hyposmic mice had a striking reduction in
the number of fat vacuoles (Figure 4A) expected for a higher thermogenic
capacity. Additionally, OMPDTR
mice had elevated
expression for many genes involved in mitochondrial biogenesis
and genes enriched in BAT, including Ucp1, Ucp2, Cox5b, CytC,
and Pgc1a (Figure 4B). Concordantly, BAT fat pads from
OMPDTR
mice had increased mitochondrial respiration (Figure
4C). In line with increased thermogenic capacity leading to
increased fatty acid oxidation, increased expression of genes
is involved in fatty acid oxidative metabolism from adipose tissue
of OMPDTR
mice (Figure 4D).
Increased thermogenesis and browning of fat can be profoundly
modulated by inputs from the sympathetic nervous system
through the release of catecholamine (Bartness and Song,
2007; Cousin et al., 1992; Haynes et al., 1997; Young et al.,
1984). We found a drastic increase in circulating levels of the
catecholamine noradrenaline, a major agonist of b-adrenergic
receptors, in OMPDTR
mice consistent with increased thermogenesis
in fat (Figure 4E). Transcript levels of the b-adrenergic
receptor Adrb3, a critical mediator of fat thermogenesis and
lipolysis, was substantially elevated in BAT and iWAT from
OMPDTR
mice (Figure 4F). Prolonged b-adrenergic receptor
activity induces the emergence of brown-like adipocytes in
subcutaneous iWAT which display thermogenic properties and
contribute to energy dissipation (Petrovic et al., 2010). Strikingly,
iWAT isolated from OMPDTR
mice had large pockets of cells with
the multilocular morphology characteristic of brown-like adipocytes
(Figure 4G). Consistent with fat browning of white adipose
tissue, a thermogenic program was strongly activated in iWAT
from OMPDTR
mice. Levels of Ucp1 mRNA-expression were upregulated
by 40-fold, and levels of Cidea were also increased
(Figure 4H).
Excessive calories stored in fat can be utilized by the activation
of lipases capable of converting the large lipid stores into usable
triglycerides. One such enzyme, hormone-sensitive lipase
(HSL), hydrolyzes intracellular triacylglycerol and diacylglycerol
and is one of the key enzymes controlling lipolysis. Triggering
b-adrenergic receptors promotes lipolysis through a cAMP
cascade leading to PKA-mediated phosphorylation and activation
of HSL (Carmen and Vı´ctor, 2006). We found that the total
levels of HSL were similar among genotypes; however, levels
of activated, phosphorylated (Ser 660) HSL were increased in
WAT from OMPDTR
mice (Figures 4I and S4E), pointing toward
enhanced lipid catabolism in the hyposmic mice. Consistent
with activation of HSL, the isoproterenol-induced lipolysis rate
was increased in OMPDTR
mice compared to control adipocytes
(Figure 4J), whereas unstimulated lipolysis was similar among
genotypes (Figure S3F), confirming increased stimulation of
b-adrenergic signaling in adipose depots of the hyposmic animals.
In addition, OMPDTR
mice presented a lower respiratory
quotient compared to controls (Figure S3G) as a response to
increased fatty acid oxidative metabolism. Altogether, these results
demonstrate that the lean phenotype observed in OMPDTR
mice is due to increased sympathetic nerve activity resulting in
increased fat thermogenesis.
Hyposmia Attenuates Further Weight Gain upon
High-Fat Feeding
The results presented thus far indicate that hyposmia can mitigate
DIO in mice prior to being placed on a high caloric diet.
To explore the possibility that loss of olfaction could also mitigate
DIO of mice already obese due to increased caloric intake, we
ablated OSNs in DIO glucose-intolerant mice. We found that
one single injection of DT was sufficient to reduce adiposity in
the severely obese animals, and repeated injections maintained
reduced body weight and fat mass despite continued HFD
feeding of these mice (Figures 5A, 5B, and S5A). This strong
reduction in fat mass, but not lean mass, was achieved without
reducing food intake (Figure 5C) and was mediated by increased
energy expenditure (Figure 5F). As observed in animals in which
OSNs were ablated prior to DIO, these mice had improved
glucose clearance (Figure 5D), insulin sensitivity (Figure S5B),
increased oxygen consumption, and lower basal insulin levels
(Figures S5C and S5D).
Elevated leptin levels observed in obesity and type 2 diabetes
mellitus contribute to impaired hypothalamic leptin signaling,
thus leading to leptin’s failure to suppress appetite and promote
energy expenditure (Balthasar et al., 2004; Cowley et al., 2001).
To examine leptin action in OMPDTR
mice, we performed a leptin
sensitivity test prior and post DT injection. The anorectic effects
of leptin were detectable in both control and OMPDTR
mice
before DT injection (Figure 5E). DT administration enhanced leptin’s
ability to lower food intake in the OMPDTR
animals
compared to controls. Additionally, DT-injected OMPDTR
animals
also presented improved energy expenditure upon leptin
treatment compared to controls and untreated OMPDTR
mice
(Figure 5F), suggesting that the reduction of olfactory inputs in
mice increased leptin sensitivity upon overnutrition.
Loss of IGF1 Receptor in the Olfactory Epithelium
Promotes Increased Olfactory Sensitivity
Since our results indicated that reducing olfactory sensitivity is
associated with improved metabolic fitness, we investigated
whether the converse manipulation of olfactory acuity would
worsen metabolic health. In line with this hypothesis, the potassium-dependent
Na+
/Ca2+
exchanger Nckx4 mutant mice
present reduced body weight and impaired olfactory acuity (Stephan
et al., 2011), whereas diabetic mice lacking leptin (ob/ob) or
Cell Metabolism 26, 198–211, July 5, 2017 203
A
G
B
F
I
BAT
iW
AT
0
2
4
6
8
10 **
Foldchange
**
J
H
0
50
100
150
200
Cellnumber
*
OCR(pmol/min)
0
20
40
60
80
100
120
140
*
C
4
6
8
10
12
14
Noradrenaline(ng/mL)
**
D E
gWAT
iWAT
iWAT
BAT
A
p2
C
idea
α
Pgc1
Prdm
16
U
cp1
U
cp2
A
diponectin
D
io2
0
5
10
20
40
60
80
***
*
*
Foldchange
U
cp1
U
cp2
C
ox4il
C
ox5b
α
Pgc1
C
ox8b
C
oxIII
C
ytC
0
2
4
6
8
*
**
* * *
Foldchange
*
A
caa1b
M
cad
M
cd
LcadA
cox1A
caa2
β
A
cc-
C
pta1C
pt1b
γ
Ppar-
0
5
10
15
20
60
80
100
*
**
Foldchange
*
*
Adrb-3
Control
OMPDTR
Control
OMPDTR
Control
OMPDTR
Control
OMPDTR
Control
OMPDTR
Control
OMPDTR
Control
OMPDTR
Control
OMP DTR
Control OMP
DTR
Control OMP
DTR
BAT
0.0 0.5 1.0 1.5 2.0
0
100
200
300
400
isoproterenol (hr)
Glycerolrelease(μM)
*
**
isoproterenol
- +
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
**
Ratio
P-HSL (Ser 660)/HSL
ns
**
Figure 4. OMPDTR
Have Increased Thermogenesis in Adipose Tissue
(A) Hematoxylin and eosin staining of brown fat depots and brown adipocytes number in HFD-fed OMPDTR
and controls. Right panel, n = 3, cells counted on a
section of 300 mm2
. Scale bar, 50 mm.
(B) RTqPCR of mitochondrial genes in brown fat of HFD-fed OMPDTR
and control mice.
(C) OCR of BAT from HFD-fed OMPDTR
and control littermates analyzed by Seahorse extracellular flux analyzer after 14 weeks of DIO.
(D) RTqPCR against fatty acid oxidation genes in gonadal fat of HFD-fed OMPDTR
and control littermates after 14 weeks of DIO.
(E) Noradrenaline serum levels after 14 weeks of DIO.
(F) Adrb-3 gene expression levels (RTqPCR) in brown and inguinal fat depots after 14 weeks of DIO.
(G) Presence of multilocular adipocytes in the inguinal fat depots by hematoxylin and eosin staining after 14 weeks of DIO. Scale bar, 50 mm.
(H) RTqPCR against brown fat and thermogenic genes in inguinal fat of HFD-fed OMPDTR
and control mice after 14 weeks of DIO.
(I) Phospho (Ser 660) HSL levels normalized to total HSL in WAT from HFD-fed OMPDTR
and controls after isoproterenol or saline injection after 14 weeks of DIO.
(J) Lipolysis rates in adipocytes exposed to isoproterenol from gonadal WAT isolated from HFD-fed OMPDTR
and control mice after 14 weeks of DIO.
Means ± SEM, n = 3–12, ***p < 0.001, **p < 0.001, *p < 0.05.
204 Cell Metabolism 26, 198–211, July 5, 2017
leptin receptors (db/db) can locate food ten times faster than
wild-type, and this effect can be abrogated in ob/ob mice by
daily leptin injections (Getchell et al., 2006). As the olfactory
bulb contains a high density of insulin receptors (Hill et al.,
1986), leptin receptors (Shioda et al., 1998), and other endocrine
hormone receptors (Palouzier-Paulignan et al., 2012), we
considered that hormonal inputs might be able to adjust olfactory
perception. Additionally, insulin application in the olfactory
mucosa has been shown to regulate odorant-evoked electroolfactogram
responses, suggesting that olfactory control depends
on nutritional status (Lacroix et al., 2008). IGF1 has previously
been established as a factor influencing peripheral regeneration
of OSNs in vitro and in vivo (Pixley et al., 1998), and IGF1R ablation
in NSCs is responsible for enhanced neurogenesis in the
olfactory bulb and improved olfactory function (Chaker et al.,
2015). Based on this knowledge, we thought to create a new
model of loss of IGF1 receptors (IGF1Rs) in the MOE as a complementary
genetic tool to the OMPDTR
model to manipulate
olfactory perception and study its impact on metabolic function.
In situ hybridization confirmed the presence of IGF1R in the
MOE (Figure S6A). We then explored the consequence of
impairing IGF1R expression in MOE. Mice homozygous for the
loxP-flanked IGF1R allele (IGF1Rflox/flox
) were crossed with
OMP-Cre animals to generate IGF1RDOMP
mice. As expected,
these animals had reduced abundance of IGF-1R mRNA specifically
in OSNs as determined by RNAscope (Figure S6A) and
reduced transcript levels as measured by RT-qPCR and
RNasequencing on isolated olfactory epithelium (Figure S6B).
We also verified that mutation of IGF1R in the MOE did not
affect levels of IR as revealed by RT-qPCR analysis and
RNasequencing on isolated olfactory epithelium (Figure S6C).
Moreover, global comparative gene expression analysis on
olfactory epithelium of control and IGF1RDOMP
-mice revealed
2,123 significantly differentially regulated genes (p % 0.05) between
both genotypes (Figure S6D). Interestingly, this mRNAexpression
analysis revealed a differential regulation of genes
associated with OSN maturity, with a higher expression of genes
characteristic for immature OSNs in IGF1RDOMP
mice, whereas
expression of markers of mature OSNs were significantly
decreased in these mice compared to controls (Figure S6D).
One-sided Fisher’s exact test confirmed a significant enrichment
for immature markers in the set of upregulated genes and mature
markers in the set of downregulated genes with a p value less
than 1eÀ20
in both cases, whereas the enrichment for immature
marker genes in the downregulated set and mature marker
genes in the upregulated set was not significant (p values of
0.98 and 1, respectively, Figure S6D).
To further validate the differential regulation of markers of
immature and mature OSNs in the absence of OSN-intrinsic
IGF-1 signaling, we assessed the mRNA expression of several
immature and mature markers (Figure S6E). This analysis
revealed an enrichment in growth-associated protein 43
(GAP43), a canonical marker for immature OSNs (McIntyre
et al., 2010), and nescient helix loop helix 2 (Nhlh2), a factor
involved in neuronal development (Nickell et al., 2012). Surprisingly,
previously identified markers of mature OSNs were
reduced in IGF1RDOMP
mice (Figure S6E). In line with the
mRNA abundance profile of MOE tissue, immunostaining of
A B
before DT after DT
C
-leptin +leptin
0
1
2
3
4
ns
* *
*
-leptin +leptin
0
1
2
3
4
Foodintake(g)
ns
* *
ns
Foodintake(g)
F after DT
Control
OMP DTR
Leanmass(g)
0
5
10
15
20
25
Fatmass(g)
0
5
10
15
20
**
Control
OMPDTR
Control
OMPDTR
0.0
0.5
1.0
1.5
2.0
2.5
3.0
HFD daily FI (g/day)
Control
OMPDTR
D
0 20 40 60 80 100 120
0
100
200
300
400
500
600
***
time (min)
Bloodglucose(mg/dl)
Control
OMP DTR
E
8 10 12 14 16 18 20 22 24
0
1
2
3
4
5
DT DT DT
Age (Weeks)
Glucose tolerance test
6 8 10 12 14 16 18 20 22 24
0
4
8
12
16
20
24
28
32
36
40
44
48 DT
DT
DT
*
* * * * *
Age (Weeks)
Weight(g)
-leptin
+leptin
-leptin
+leptin
0.0
0.1
0.2
0.3
Heat(kcal/hrFFM)
ns
*
***
***
Figure 5. Olfactory Inhibition Restores Metabolic Health in Obese Animals
(A) Weight gain and adiposity of HFD-fed OMPDTR
and controls before and after DT injection after 7 weeks of DIO.
(B) Fat mass and lean mass of OMPDTR
and controls.
(C) Weekly food intake in DT-injected OMPDTR
and controls.
(D) Glucose tolerance test of HFD-fed mice (two-way ANOVA).
(E and F) Daily food intake and energy expenditure in a subgroup of OMPDTR
and control littermates before and after leptin administration 2 weeks after DT
delivery.
Means ± SEM, n = 4–10, ***p < 0.001, **p < 0.001, *p < 0.05.
Cell Metabolism 26, 198–211, July 5, 2017 205
GAP43 revealed a higher abundance of this protein in MOE section
(Figures S7A and S7B), whereas Kcnq2 was reduced (Figures
S7C and S7D). These data suggest that the MOE of
IGF1RDOMP
mice undergoes increased proliferation of immature
cells that could potentially result in reduced mature OSNs. To
address whether mature cells were under-represented, we
quantified the average number of OMP-positive neurons in tdTomato
DOMP
mice and found it to be unaffected by this MOE
remodeling (Figures S7B and S7D). Additionally, RT-qPCR revealed
no significant changes in the markers of apoptosis
caspase-3 and -4 but a significant increase in mKi67 expression,
demonstrating an increased proliferation in the olfactory epithelium
of IGF1RDOMP
mice (Figure S6E).
Interestingly, MOE remodeling upon loss of IGF1R functionally
translated into an improved olfactory performance in IGF1RDOMP
mice, as evidenced by an increased ability in locating a food
source and increased reactivity in locating a social odor (Figures
6A and 6B) without any impairment in odor discrimination and
detection as measured by an olfactory habituation and dishabituation
test (Figure S6F). Taken together, these findings indicate
that loss of IGF1 receptors in the MOE promotes an increased
proliferation of immature OSNs resulting in hyperosmia of
these mice.
Increased Adiposity in IGF1RDOMP
Mice
To define the physiological consequences of reducing IGF1R
expression in OSNs, remodeling MOE, and improving olfactory
performance, we assessed adiposity and body weight in
IGF1RDOMP
mice compared to control mice. Upon normal
chow diet feeding, these mice presented a body weight comparable
to controls until 12 weeks; however, further analysis revealed
a significant, age-dependent weight gain in these animals
C D
E
3 4 5 6 7 8 9 10 11 12
0
10
20
30
40
Age (weeks)
Bodyweight(g)
control
IGF1R
ΔOMP
Fat mass (chow diet)
Whole fat gWAT iWAT BAT
0
1
2
3
4
8
10
12
14
control
IGF1R
ΔOMP
**
Fatmass(g)
**
**
Average adipocyte surface
gWAT BAT
0
500
1000
3000
4500
6000
Surface(um2)
control
IGF1RΔOMP**
Bodyweight(g)
*
40 weeks
30
35
40
45
50
Timetofindfood(s)
control IGF1RΔOMP
0
150
300
450
600
**
control IGF1RΔOMP
0
200
400
600
800
Timetofindodor(s)
*
Food finding test Social odor finding testA B
Foodintake(g)
0
1
2
3
4
5
control IGF1RΔOM
F
Figure 6. IGF1R Ablation in the Olfactory
Epithelium Results in Obesity upon Chow
Control Diet
(A) Average time needed to find hidden food pellets
for random fed mice.
(B) Average time needed to find a hidden Q-tip
immersed in female urine for random fed mice.
(C) Growth curve of control and IGF1R DOMP
mice
on chow diet, 3–12 weeks (left panel) and 40 weeks
(right panel).
(D) Fat mass of control and IGF1R DOMP
mice at
40 weeks.
(E) Average adipocyte surface quantification of
gonadal WAT and BAT of control and IGF1RDOMP
mice at 40 weeks.
(F) Average daily food intake of control and
IGF1RDOMP
mice at 15 weeks.
Means ± SEM, n = 5–16, ***p < 0.001, **p < 0.001,
*p < 0.05.
(Figure 6C). This increased body weight
was accompanied by excessive adipose
tissue mass in the gonadal and inguinal
fat depots (Figure 6D). Particularly, the
gonadal fat pad had enlarged adipocytes
compared to control animals (Figure 6E),
accompanied by non-significant reduction
in energy expenditure, oxygen consumption, and respiratory
quotient (Figures S7E–S7G). Importantly, adipose expansion
was independent of food intake, which remained unchanged
compared to controls (Figure 6F). Globally, markers of browning
and particularly Dio2 (Figure S7H) were reduced in the subcutaneous
inguinal fat pads of these animals, and Ucp1 expression
was reduced in gWAT (Figure S7I), whereas these markers were
unaffected in brown fat (Figure S7J). Taken together, these findings
indicate that inhibition of IGF1 signaling in OSNs enhances
olfactory perception and in turn promotes adiposity.
Insulin Resistance and Reduced Suppression of Hepatic
Glucose Production Are Associated with Improved
Olfactory Capacity
To assess the regulation of peripheral glucose metabolism of
IGF1RDOMP
and control mice, we performed insulin and glucose
tolerance tests in these animals (Figures 7A and 7B). IGF1RDOMP
mice exhibited impaired insulin sensitivity compared to controls,
while glucose tolerance was unaffected. To further investigate
which aspect of peripheral glucose metabolism was differentially
modulated, we performed euglycemic-hyperinsulinemic clamp
studies comparing IGF1RDOMP
and control animals. While both
groups achieved the same degree of glycemia during the
steady-state phase of the clamp (Figure 6D) and exhibited similar
degree of glucose disappearance (Figures 7D and 7F), the
glucose infusion rate (GIR) required to maintain euglycemia
was significantly reduced in IGF1RDOMP
mice (Figures 7C
and 7E).
When we assessed hepatic glucose production (HGP),
IGF1RDOMP
and control mice exhibited no significant differences
in baseline, but IGF1RDOMP
mice exhibited a reduced ability of insulin
to suppress HGP (Figure 7G). However, insulin’s ability to
206 Cell Metabolism 26, 198–211, July 5, 2017
promote glucose uptake into peripheral tissues including brown
fat, gonadal fat, skeletal muscle, and the brain remained unaltered
(Figure S8H). Importantly, IGF1RDOMP
mice were hyperinsulinemic
and hyperleptinemic in random fed conditions
compared to the control mice (Figure S8I). Thus, inhibition of
IGF1R in OSNs promotes peripheral insulin resistance and pinpoints
the ability of OSNs to influence autonomic control of
hepatic glucose production.
DISCUSSION
Since the discovery of hypothalamic circuits influencing anabolic
and catabolic processes, very little information has been gathered
on the role of sensory inputs in the regulation of energy
homeostasis pathways. We establish that OSNs actively participate
in the modulation of peripheral metabolism in response to
external and internal cues. Prior to our work, multiple studies
have demonstrated that the initiation and cessation of feeding
behaviors are associated with changes in olfactory sensitivity.
Hunger arouses olfactory perception (Albrecht et al., 2009;
Cameron et al., 2012; Soria-Go´ mez et al., 2014), whereas a
dampening of olfactory sensitivity is observed after the ingestion
of a meal and therefore correlates with a satiety signal (Savigner
et al., 2009). Thus, given the existing precedents linking olfaction
with feeding behavior, we characterized in depth the consequence
of either impairing or improving olfaction on metabolic
homeostasis.
We discovered that reduced olfactory inputs initiate prolonged
sympathetic responses to severe energy depletion through
increased b-adrenergic signaling in adipose tissue, therefore
promoting increased lipolysis in gonadal fat and increased thermogenesis
in brown and subcutaneous fat fads. In humans,
hyposmia is commonly accompanied by the loss of flavor from
food, with subsequent loss of appetite and weight loss or
A B C
D E F
G H
Figure 7. IGF1R Ablation in OSNs Promotes Insulin Resistance and Impaired Insulin-Induced Suppression of Hepatic Glucose Production
(A and B) Insulin and glucose tolerance tests of control and IGF1RDOMP
mice at 11 and 12 weeks, two-way ANOVA test.
(C and D) Glucose infusion rate and glucose levels during a hyperinsulinemic-euglycemic CLAMP of control and IGF1RDOMP
mice at 22 weeks.
(E) Area under curve of the glucose infusion rate (CLAMP).
(F) Rate of disappearance (CLAMP).
(G) Hepatic glucose production at basal and steady state (CLAMP).
(H) Model for the regulation of energy balance by olfactory neurons.
Means ± SEM, n = 8–16, ***p < 0.001, **p < 0.001, *p < 0.05.
Cell Metabolism 26, 198–211, July 5, 2017 207
anorexia (Bonfils et al., 2005; Henkin and Smith, 1971). We
observe an adipose mass decrease in our study without overt
food aversion (Figures 5B and 5C). Thus, our data support the
notion that reduced olfactory output triggers a metabolic
response mimicking the cessation of feeding and thus leads to
catabolic energy utilization needed to protect animals from a
high caloric diet (Figure 7H).
In our second mouse model, we demonstrate that loss of
IGF1R in the olfactory epithelium confers opposing effects to
the ablation of OSNs, and increases smell perception (Figures
6D and 6E). Notably, these mice displayed increased adiposity
accompanied by hyperleptinemia, hyperinsulinemia, and insulin
resistance (Figures 7A and S7L). Strikingly, insulin’s ability to
suppress HGP is impaired in these animals, pointing to a new
role for IGF1 in OSNs in the control of peripheral insulin resistance
(Figure 7G). IGF1, which shares 48% amino acid identity
with proinsulin, has been implicated in ameliorating insulin sensitivity
in both rodent models and human subjects (Clemmons,
2004), and suppression of liver IGF1 promotes hyperinsulinemia
and insulin resistance by acting on skeletal muscle (Ferna´ ndez
et al., 2001; Yakar et al., 2001). The data presented here support
a more complex model in which IGF1 also functions in OSNs to
regulate peripheral insulin sensitivity. In addition, how OSNs
signal this information to the liver remains to be determined.
One plausible explanation implicates OSN communication with
the CNS. Indeed, central insulin action in the CNS is necessary
for the suppression of endogenous glucose production. In
particular, AgRP-expressing neurons play a critical role in the
suppression of HGP by the regulation of vagal efferent nerves
in the liver (Ko¨ nner et al., 2007; Obici et al., 2002; Pocai et al.,
2005). Our results suggest that OSNs also contribute to the regulation
of HGP and could function upstream of ARC neurons.
Despite the link that we observe between OSNs, energy
expenditure, and adiposity, we cannot fully exclude that unspecific
manipulation of non-olfactory CNS neurons expressing
OMP-Cre may also affect the observed weight phenotypes
(Figures S1C and S1D). Nevertheless, intranasal delivery of the
Ad-flex-ta-Casp3-TEVp adenovirus selectively reduced OSNs
without CNS penetrance and reproduced the metabolic benefits
of the genetic ablation model, although to a lesser extent associated
with the weaker ablation of OSNs using this method (Figure
2). Consistent with our findings, a mouse model with reduced
smell, the Nckx4 mutant mouse, presents reduced body weight
and impaired olfactory acuity (Stephan et al., 2011), whereas a
hyperosmic phenotype is observed in diabetic mice lacking leptin
(ob/ob) or leptin receptors (db/db)(Getchell et al., 2006).
Nonetheless, our hypothesis is challenged by the whole-body
deletion of the voltage-gated K(+) channel (Kv) subtype Kv1.3
channel, a mouse model of enhanced olfactory sensitivity
(Tucker et al., 2012) which shows resistance to DIO (Tucker
et al., 2012; Xu et al., 2003b). The exact contribution of the olfactory
system in this particular model remains to be determined, as
Kv1.3 is highly expressed throughout the brain and non-neuronal
tissues, including liver and skeletal muscle (Xu et al., 2003b).
Therefore, whether the lean phenotype of the hyperosmic
Kv1.3 knockout animals comes from their enhanced smell acuity
or from other sites of action remains unclear.
The finding that OSNs can control peripheral metabolism is
intriguing, and multiple mechanisms could be engaged in this
circuitry. It is mainly thought that the hypothalamus receives
indirect inputs from OSNs through the MOB and transmitted
to the centers of the olfactory cortex (Purves et al., 2001;
Su et al., 2009). Interestingly, direct connections between
discrete subpopulations of OSNs and several nuclei from the
hypothalamus, such as the paraventricular nucleus (PVN), the
supraoptic nucleus (SO), and the luteinizing hormone-releasing
hormone (LHRH)-secreting neurons, have been observed
(Bader et al., 2012; Yoon et al., 2005), reinforcing the idea
that an active circuitry initiated in OSNs might influence metabolic
homeostasis. Our data strongly indicate a circuit that relays
information to autonomic neurons and may require central
neurons, including hypothalamic pro-opiomelanocortin (Pomc)
neurons in the arcuate nucleus of the hypothalamus which exert
their anorexigenic effects by activating melanocortin-4 receptors
and an autonomic response (Cowley et al., 2001; Wardlaw,
2011). In line with this hypothesis, fiber photometry recording of
AgRP and POMC neurons activity in the hypothalamus of
awake, behaving animals shows that the perception of food
rapidly switches the activation state of these neurons upon
hunger and can be immediately reversed by removing the
food cues (Chen et al., 2015). Additionally, olfactory inputs
may be integrated by a complex interplay of different hypothalamic
and brainstem nuclei expressing appetite-modulatory
neuropeptides. Regardless, the potential of modulating olfactory
signals in the context of the metabolic syndrome or diabetes
is attractive. The data presented here show that even
relatively short-term loss of smell improves metabolic health
and weight loss, despite the negative consequences of being
on a HFD.
STAR+METHODS
Detailed methods are provided in the online version of this paper
and include the following:
d KEY RESOURCES TABLE
d CONTACT FOR REAGENT AND RESOURCE SHARING
d METHOD DETAILS
B Animal Care
B Genetic mouse models
B Indirect Calorimetry, Physical Activity, and Food Intake
B Food finding test
B Social odor finding test
B Habituation-dishabituation test
B Metabolic studies
B Ex vivo lipolysis
B Protein analysis
B Leptin sensitivity
B Quantitative PCR analysis
B Adenovirus propagation and titration
B Immunofluorescence and histological studies
B Hyperinsulinemic-euglycemic CLAMPS
B Adipose tissue oxygen consumption
B Bomb fecal calorimetry
B RNA sequencing
B RNAscope analysis
B Data and Software Availability
B Statistical methods
208 Cell Metabolism 26, 198–211, July 5, 2017
SUPPLEMENTAL INFORMATION
Supplemental Information includes seven figures and one table and can be
found with this article at http://dx.doi.org/10.1016/j.cmet.2017.06.015.
AUTHOR CONTRIBUTIONS
C.E.R. executed most of the olfactory ablation experiments, and E.T. characterized
the IGF1RDOMP
model. C.E.R., E.T., A.D., and J.C.B. planned the majority
of the experiments and wrote the manuscript. J.H. and P.F. performed
animal husbandry maintenance and assistance to metabolic studies. O.H. performed
bioinformatic analyses of the RNaseq data. M.M.A.P., L.E.R., and B.H.
assisted E.T. during tissue analysis experiments. J.A. and H.B. assisted with
the setup of behavioral experiments. K.T., C.M.A., and C.D.M.F. assisted
CE.R. .during tissue analysis experiments.
ACKNOWLEDGMENTS
We thank Prof. K. Baldwin, T. Tang, Prof. H. Sook Sul, R. Fletcher, and Prof.
Daniela Kaufer for training and advice. We thank Prof. Margolis for providing
the OMP antibody. We also thank H. Dexai for assistance with tissue
sectioning and P.J. Rene for help with data analysis. This work was supported
by the Howard Hughes Medical Institute and the Glenn Center for Research on
Aging, the American Diabetes Association Pathway to Stop Diabetes Grant
1-15-INI-12 (CER). A.D. is cofounder of Proteostasis Therapeutics, Inc., and
Mitobridge, Inc., and declares no financial interest related to this work.
Received: October 7, 2016
Revised: April 9, 2017
Accepted: June 16, 2017
Published: July 5, 2017
REFERENCES
Albrecht, J., Schreder, T., Kleemann, A.M., Scho¨ pf, V., Kopietz, R., Anzinger,
A., Demmel, M., Linn, J., Kettenmann, B., and Wiesmann, M. (2009). Olfactory
detection thresholds and pleasantness of a food-related and a non-food odour
in hunger and satiety. Rhinology 47, 160–165.
Bader, A., Klein, B., Breer, H., and Strotmann, J. (2012). Connectivity from
OR37 expressing olfactory sensory neurons to distinct cell types in the hypothalamus.
Front. Neural Circuits 6, 84.
Balthasar, N., Coppari, R., McMinn, J., Liu, S.M., Lee, C.E., Tang, V., Kenny,
C.D., McGovern, R.A., Chua, S.C., Jr., Elmquist, J.K., and Lowell, B.B.
(2004). Leptin receptor signaling in POMC neurons is required for normal
body weight homeostasis. Neuron 42, 983–991.
Bartness, T.J., and Song, C.K. (2007). Thematic review series: adipocyte
biology. Sympathetic and sensory innervation of white adipose tissue.
J. Lipid Res. 48, 1655–1672.
Baskin, D.G., Porte, D., Jr., Guest, K., and Dorsa, D.M. (1983). Regional concentrations
of insulin in the rat brain. Endocrinology 112, 898–903.
Belgardt, B.F., and Br€uning, J.C. (2010). CNS leptin and insulin action in the
control of energy homeostasis. Ann. N Y Acad. Sci. 1212, 97–113.
Belluscio, L., Gold, G.H., Nemes, A., and Axel, R. (1998). Mice deficient in G(olf)
are anosmic. Neuron 20, 69–81.
Betley, J.N., Xu, S., Cao, Z.F.H., Gong, R., Magnus, C.J., Yu, Y., and Sternson,
S.M. (2015). Neurons for hunger and thirst transmit a negative-valence teaching
signal. Nature 521, 180–185.
Bjorklund, A., Hokfelt, T., and Swanson, L.W. (1996). Integrated Systems of the
CNS, Part III: Cerebellum, Basal Ganglia, Olfactory System (Elsevier).
Bonfils, P., Avan, P., Faulcon, P., and Malinvaud, D. (2005). Distorted odorant
perception: analysis of a series of 56 patients with parosmia. Arch.
Otolaryngol. Head Neck Surg. 131, 107–112.
Brunet, L.J., Gold, G.H., and Ngai, J. (1996). General anosmia caused by a targeted
disruption of the mouse olfactory cyclic nucleotide-gated cation channel.
Neuron 17, 681–693.
Buch, T., Heppner, F.L., Tertilt, C., Heinen, T.J.A.J., Kremer, M., Wunderlich,
F.T., Jung, S., and Waisman, A. (2005). A Cre-inducible diphtheria toxin receptor
mediates cell lineage ablation after toxin administration. Nat. Methods 2,
419–426.
Buiakova, O.I., Baker, H., Scott, J.W., Farbman, A., Kream, R., Grillo, M.,
Franzen, L., Richman, M., Davis, L.M., Abbondanzo, S., et al. (1996).
Olfactory marker protein (OMP) gene deletion causes altered physiological activity
of olfactory sensory neurons. Proc. Natl. Acad. Sci. USA 93, 9858–9863.
Cameron, J.D., Goldfield, G.S., and Doucet, E´ . (2012). Fasting for 24 h improves
nasal chemosensory performance and food palatability in a related
manner. Appetite 58, 978–981.
Carmen, G.-Y., and Vı´ctor, S.-M. (2006). Signalling mechanisms regulating
lipolysis. Cell. Signal. 18, 401–408.
Chaker, Z., Aı¨d, S., Berry, H., and Holzenberger, M. (2015). Suppression of
IGF-I signals in neural stem cells enhances neurogenesis and olfactory function
during aging. Aging Cell 14, 847–856.
Chen, H., Kohno, K., and Gong, Q. (2005). Conditional ablation of mature olfactory
sensory neurons mediated by diphtheria toxin receptor. J. Neurocytol.
34, 37–47.
Chen, Y., Lin, Y.-C., Kuo, T.-W., and Knight, Z.A. (2015). Sensory detection of
food rapidly modulates arcuate feeding circuits. Cell 160, 829–841.
Clemmons, D.R. (2004). The relative roles of growth hormone and IGF-1 in
controlling insulin sensitivity. J. Clin. Invest. 113, 25–27.
Collins, S., Kuhn, C.M., Petro, A.E., Swick, A.G., Chrunyk, B.A., and Surwit,
R.S. (1996). Role of leptin in fat regulation. Nature 380, 677.
Cousin, B., Cinti, S., Morroni, M., Raimbault, S., Ricquier, D., Pe´ nicaud, L., and
Casteilla, L. (1992). Occurrence of brown adipocytes in rat white adipose tissue:
molecular and morphological characterization. J. Cell Sci. 103, 931–942.
Cowley, M.A., Smart, J.L., Rubinstein, M., Cerda´ n, M.G., Diano, S., Horvath,
T.L., Cone, R.D., and Low, M.J. (2001). Leptin activates anorexigenic POMC
neurons through a neural network in the arcuate nucleus. Nature 411, 480–484.
Critchley, H.D., and Rolls, E.T. (1996). Hunger and satiety modify the responses
of olfactory and visual neurons in the primate orbitofrontal cortex.
J. Neurophysiol. 75, 1673–1686.
Dulac, C., and Torello, A.T. (2003). Molecular detection of pheromone signals
in mammals: from genes to behaviour. Nat. Rev. Neurosci. 4, 551–562.
Elmquist, J.K., Bjørbaek, C., Ahima, R.S., Flier, J.S., and Saper, C.B. (1998).
Distributions of leptin receptor mRNA isoforms in the rat brain. J. Comp.
Neurol. 395, 535–547.
Enerb€ack, S., Jacobsson, A., Simpson, E.M., Guerra, C., Yamashita, H.,
Harper, M.E., and Kozak, L.P. (1997). Mice lacking mitochondrial uncoupling
protein are cold-sensitive but not obese. Nature 387, 90–94.
Fan, W., Boston, B.A., Kesterson, R.A., Hruby, V.J., and Cone, R.D. (1997).
Role of melanocortinergic neurons in feeding and the agouti obesity syndrome.
Nature 385, 165–168.
Ferna´ ndez, A.M., Kim, J.K., Yakar, S., Dupont, J., Hernandez-Sanchez, C.,
Castle, A.L., Filmore, J., Shulman, G.I., and Le Roith, D. (2001). Functional
inactivation of the IGF-I and insulin receptors in skeletal muscle causes type
2 diabetes. Genes Dev. 15, 1926–1934.
Ferre´ , P., Leturque, A., Burnol, A.F., Penicaud, L., and Girard, J. (1985).
A method to quantify glucose utilization in vivo in skeletal muscle and white
adipose tissue of the anaesthetized rat. Biochem. J. 228, 103–110.
Fisher, S.J., and Kahn, C.R. (2003). Insulin signaling is required for insulin’s
direct and indirect action on hepatic glucose production. J. Clin. Invest. 111,
463–468.
Fisher, F.M., Kleiner, S., Douris, N., Fox, E.C., Mepani, R.J., Verdeguer, F., Wu,
J., Kharitonenkov, A., Flier, J.S., Maratos-Flier, E., and Spiegelman, B.M.
(2012). FGF21 regulates PGC-1a and browning of white adipose tissues in
adaptive thermogenesis. Genes Dev. 26, 271–281.
Gascuel, J., Lemoine, A., Rigault, C., Datiche, F., Benani, A., Penicaud, L., and
Lopez-Mascaraque, L. (2012). Hypothalamus-olfactory system crosstalk:
orexin a immunostaining in mice. Front. Neuroanat. 6, 44.
Cell Metabolism 26, 198–211, July 5, 2017 209
Getchell, T.V., Kwong, K., Saunders, C.P., Stromberg, A.J., and Getchell, M.L.
(2006). Leptin regulates olfactory-mediated behavior in ob/ob mice. Physiol.
Behav. 87, 848–856.
Gogos, J.A., Osborne, J., Nemes, A., Mendelsohn, M., and Axel, R. (2000).
Genetic ablation and restoration of the olfactory topographic map. Cell 103,
609–620.
Grill, H.J., and Hayes, M.R. (2012). Hindbrain neurons as an essential hub in the
neuroanatomically distributed control of energy balance. Cell Metab. 16,
296–309.
Gropp, E., Shanabrough, M., Borok, E., Xu, A.W., Janoschek, R., Buch, T.,
Plum, L., Balthasar, N., Hampel, B., Waisman, A., et al. (2005). Agouti-related
peptide-expressing neurons are mandatory for feeding. Nat. Neurosci. 8,
1289–1291.
Hahn, T.M., Breininger, J.F., Baskin, D.G., and Schwartz, M.W. (1998).
Coexpression of Agrp and NPY in fasting-activated hypothalamic neurons.
Nat. Neurosci. 1, 271–272.
Harrow, J., Frankish, A., Gonzalez, J.M., Tapanari, E., Diekhans, M.,
Kokocinski, F., Aken, B.L., Barrell, D., Zadissa, A., Searle, S., et al. (2012).
GENCODE: the reference human genome annotation for The ENCODE
Project. Genome Res. 22, 1760–1774.
Haynes, W.G., Morgan, D.A., Walsh, S.A., Mark, A.L., and Sivitz, W.I. (1997).
Receptor-mediated regional sympathetic nerve activation by leptin. J. Clin.
Invest. 100, 270–278.
Henkin, R.I., and Smith, F.R. (1971). Appetite and anosmia. Lancet 1,
1352–1353.
Hill, J.M., Lesniak, M.A., Pert, C.B., and Roth, J. (1986). Autoradiographic
localization of insulin receptors in rat brain: prominence in olfactory and limbic
areas. Neuroscience 17, 1127–1138.
Huszar, D., Lynch, C.A., Fairchild-Huntress, V., Dunmore, J.H., Fang, Q.,
Berkemeier, L.R., Gu, W., Kesterson, R.A., Boston, B.A., Cone, R.D., et al.
(1997). Targeted disruption of the melanocortin-4 receptor results in obesity
in mice. Cell 88, 131–141.
Kang, N., Kim, H., Jae, Y., Lee, N., Ku, C.R., Margolis, F., Lee, E.J., Bahk, Y.Y.,
Kim, M.-S., and Koo, J. (2015). Olfactory marker protein expression is an indicator
of olfactory receptor-associated events in non-olfactory tissues. PLoS
ONE 10, e0116097.
Kim, D., Pertea, G., Trapnell, C., Pimentel, H., Kelley, R., and Salzberg, S.L.
(2013). TopHat2: accurate alignment of transcriptomes in the presence of insertions,
deletions and gene fusions. Genome Biol. 14, R36.
Ko¨ nner, A.C., Janoschek, R., Plum, L., Jordan, S.D., Rother, E., Ma, X., Xu, C.,
Enriori, P., Hampel, B., Barsh, G.S., et al. (2007). Insulin action in AgRP-expressing
neurons is required for suppression of hepatic glucose production.
Cell Metab. 5, 438–449.
Lacroix, M.-C., Badonnel, K., Meunier, N., Tan, F., Schlegel-Le Poupon, C.,
Durieux, D., Monnerie, R., Baly, C., Congar, P., Salesse, R., and Caillol, M.
(2008). Expression of insulin system in the olfactory epithelium: first approaches
to its role and regulation. J. Neuroendocrinol. 20, 1176–1190.
Lafontan, M., and Langin, D. (2009). Lipolysis and lipid mobilization in human
adipose tissue. Prog. Lipid Res. 48, 275–297.
Li, J., Ishii, T., Feinstein, P., and Mombaerts, P. (2004). Odorant receptor gene
choice is reset by nuclear transfer from mouse olfactory sensory neurons.
Nature 428, 393–399.
Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G.,
Abecasis, G., and Durbin, R.; 1000 Genome Project Data Processing
Subgroup (2009). The sequence alignment/map format and SAMtools.
Bioinformatics 25, 2078–2079.
Love, M.I., Huber, W., and Anders, S. (2014). Moderated estimation of fold
change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550.
Lumeng, C.N., Deyoung, S.M., Bodzin, J.L., and Saltiel, A.R. (2007). Increased
inflammatory properties of adipose tissue macrophages recruited during dietinduced
obesity. Diabetes 56, 16–23.
Luquet, S., Perez, F.A., Hnasko, T.S., and Palmiter, R.D. (2005). NPY/AgRP
neurons are essential for feeding in adult mice but can be ablated in neonates.
Science 310, 683–685.
Madisen, L., Zwingman, T.A., Sunkin, S.M., Oh, S.W., Zariwala, H.A., Gu, H.,
Ng, L.L., Palmiter, R.D., Hawrylycz, M.J., Jones, A.R., et al. (2010). A robust
and high-throughput Cre reporting and characterization system for the whole
mouse brain. Nat. Neurosci. 13, 133–140.
Margolis, F.L., Verhaagen, J., Biffo, S., Huang, F.L., and Grillo, M. (1991).
Regulation of gene expression in the olfactory neuroepithelium: a neurogenetic
matrix. Prog. Brain Res. 89, 97–122.
Marino, J.S., Xu, Y., and Hill, J.W. (2011). Central insulin and leptin-mediated
autonomic control of glucose homeostasis. Trends Endocrinol. Metab. 22,
275–285.
McIntyre, J.C., Titlow, W.B., and McClintock, T.S. (2010). Axon growth and
guidance genes identify nascent, immature, and mature olfactory sensory neurons.
J. Neurosci. Res. 88, 3243–3256.
McIntyre, J.C., Davis, E.E., Joiner, A., Williams, C.L., Tsai, I.-C., Jenkins, P.M.,
McEwen, D.P., Zhang, L., Escobado, J., Thomas, S., et al.; NISC Comparative
Sequencing Program (2012). Gene therapy rescues cilia defects and restores
olfactory function in a mammalian ciliopathy model. Nat. Med. 18, 1423–1428.
Nickell, M.D., Breheny, P., Stromberg, A.J., and McClintock, T.S. (2012).
Genomics of mature and immature olfactory sensory neurons. J. Comp.
Neurol. 520, 2608–2629.
Obici, S., Zhang, B.B., Karkanias, G., and Rossetti, L. (2002). Hypothalamic insulin
signaling is required for inhibition of glucose production. Nat. Med. 8,
1376–1382.
Palouzier-Paulignan, B., Lacroix, M.-C., Aime´ , P., Baly, C., Caillol, M., Congar,
P., Julliard, A.K., Tucker, K., and Fadool, D.A. (2012). Olfaction under metabolic
influences. Chem. Senses 37, 769–797.
Petrovic, N., Walden, T.B., Shabalina, I.G., Timmons, J.A., Cannon, B., and
Nedergaard, J. (2010). Chronic peroxisome proliferator-activated receptor
gamma (PPARgamma) activation of epididymally derived white adipocyte cultures
reveals a population of thermogenically competent, UCP1-containing
adipocytes molecularly distinct from classic brown adipocytes. J. Biol.
Chem. 285, 7153–7164.
Pixley, S.K., Dangoria, N.S., Odoms, K.K., and Hastings, L. (1998). Effects of
insulin-like growth factor 1 on olfactory neurogenesis in vivo and in vitro.
Ann. N Y Acad. Sci. 855, 244–247.
Pocai, A., Obici, S., Schwartz, G.J., and Rossetti, L. (2005). A brain-liver circuit
regulates glucose homeostasis. Cell Metab. 1, 53–61.
Purves, D., Augustine, G.J., Fitzpatrick, D., Katz, L.C., LaMantia, A.-S.,
McNamara, J.O., and Williams, S.M. (2001). The Organization of the
Olfactory System (Sinauer Associates).
Rahmouni, K., and Morgan, D.A. (2007). Hypothalamic arcuate nucleus mediates
the sympathetic and arterial pressure responses to leptin. Hypertension
49, 647–652.
Savigner, A., Duchamp-Viret, P., Grosmaitre, X., Chaput, M., Garcia, S., Ma,
M., and Palouzier-Paulignan, B. (2009). Modulation of spontaneous and
odorant-evoked activity of rat olfactory sensory neurons by two anorectic peptides,
insulin and leptin. J. Neurophysiol. 101, 2898–2906.
Seale, P., Conroe, H.M., Estall, J., Kajimura, S., Frontini, A., Ishibashi, J.,
Cohen, P., Cinti, S., and Spiegelman, B.M. (2011). Prdm16 determines the
thermogenic program of subcutaneous white adipose tissue in mice. J. Clin.
Invest. 121, 96–105.
Shioda, S., Funahashi, H., Nakajo, S., Yada, T., Maruta, O., and Nakai, Y.
(1998). Immunohistochemical localization of leptin receptor in the rat brain.
Neurosci. Lett. 243, 41–44.
Skibicka, K.P., and Grill, H.J. (2009). Hypothalamic and hindbrain melanocortin
receptors contribute to the feeding, thermogenic, and cardiovascular action of
melanocortins. Endocrinology 150, 5351–5361.
Soria-Go´ mez, E., Bellocchio, L., Reguero, L., Lepousez, G., Martin, C.,
Bendahmane, M., Ruehle, S., Remmers, F., Desprez, T., Matias, I., et al.
(2014). The endocannabinoid system controls food intake via olfactory processes.
Nat. Neurosci. 17, 407–415.
Spencer, S.E., Sawyer, W.B., Wada, H., Platt, K.B., and Loewy, A.D. (1990).
CNS projections to the pterygopalatine parasympathetic preganglionic
210 Cell Metabolism 26, 198–211, July 5, 2017
neurons in the rat: a retrograde transneuronal viral cell body labeling study.
Brain Res. 534, 149–169.
Stachelscheid, H., Ibrahim, H., Koch, L., Schmitz, A., Tscharntke, M.,
Wunderlich, F.T., Scott, J., Michels, C., Wickenhauser, C., Haase, I., et al.
(2008). Epidermal insulin/IGF-1 signalling control interfollicular morphogenesis
and proliferative potential through Rac activation. EMBO J. 27, 2091–2101.
Stephan, A.B., Tobochnik, S., Dibattista, M., Wall, C.M., Reisert, J., and Zhao,
H. (2011). The Na(+)/Ca(2+) exchanger NCKX4 governs termination and adaptation
of the mammalian olfactory response. Nat. Neurosci. 15, 131–137.
Su, C.-Y., Menuz, K., and Carlson, J.R. (2009). Olfactory perception: receptors,
cells, and circuits. Cell 139, 45–59.
Sultan, M., Do¨ kel, S., Amstislavskiy, V., Wuttig, D., S€ultmann, H., Lehrach, H.,
and Yaspo, M.-L. (2012). A simple strand-specific RNA-Seq library preparation
protocol combining the Illumina TruSeq RNA and the dUTP methods.
Biochem. Biophys. Res. Commun. 422, 643–646.
Tsukiyama-Kohara, K., Poulin, F., Kohara, M., DeMaria, C.T., Cheng, A., Wu,
Z., Gingras, A.C., Katsume, A., Elchebly, M., Spiegelman, B.M., et al. (2001).
Adipose tissue reduction in mice lacking the translational inhibitor 4E-BP1.
Nat. Med. 7, 1128–1132.
Tucker, K., Overton, J.M., and Fadool, D.A. (2012). Diet-induced obesity resistance
of Kv1.3-/mice
is olfactory bulb dependent. J. Neuroendocrinol. 24,
1087–1095.
Vogt, M.C., and Br€uning, J.C. (2013). CNS insulin signaling in the control of energy
homeostasis and glucose metabolism - from embryo to old age. Trends
Endocrinol. Metab. 24, 76–84.
Wardlaw, S.L. (2011). Hypothalamic proopiomelanocortin processing and the
regulation of energy balance. Eur. J. Pharmacol. 660, 213–219.
Williams, K.W., and Elmquist, J.K. (2012). From neuroanatomy to behavior:
central integration of peripheral signals regulating feeding behavior. Nat.
Neurosci. 15, 1350–1355.
Wu, J., Bostro¨ m, P., Sparks, L.M., Ye, L., Choi, J.H., Giang, A.-H., Khandekar,
M., Virtanen, K.A., Nuutila, P., Schaart, G., et al. (2012). Beige adipocytes are a
distinct type of thermogenic fat cell in mouse and human. Cell 150, 366–376.
Wu, Z., Autry, A.E., Bergan, J.F., Watabe-Uchida, M., and Dulac, C.G. (2014).
Galanin neurons in the medial preoptic area govern parental behaviour. Nature
509, 325–330.
Xu, H., Barnes, G.T., Yang, Q., Tan, G., Yang, D., Chou, C.J., Sole, J., Nichols,
A., Ross, J.S., Tartaglia, L.A., and Chen, H. (2003a). Chronic inflammation in fat
plays a crucial role in the development of obesity-related insulin resistance.
J. Clin. Invest. 112, 1821–1830.
Xu, J., Koni, P.A., Wang, P., Li, G., Kaczmarek, L., Wu, Y., Li, Y., Flavell, R.A.,
and Desir, G.V. (2003b). The voltage-gated potassium channel Kv1.3 regulates
energy homeostasis and body weight. Hum. Mol. Genet. 12, 551–559.
Yakar, S., Liu, J.-L., Fernandez, A.M., Wu, Y., Schally, A.V., Frystyk, J.,
Chernausek, S.D., Mejia, W., and Le Roith, D. (2001). Liver-specific igf-1
gene deletion leads to muscle insulin insensitivity. Diabetes 50, 1110–1118.
Yang, C.F., Chiang, M.C., Gray, D.C., Prabhakaran, M., Alvarado, M., Juntti,
S.A., Unger, E.K., Wells, J.A., and Shah, N.M. (2013). Sexually dimorphic neurons
in the ventromedial hypothalamus govern mating in both sexes and
aggression in males. Cell 153, 896–909.
Yoon, H., Enquist, L.W., and Dulac, C. (2005). Olfactory inputs to hypothalamic
neurons controlling reproduction and fertility. Cell 123, 669–682.
Young, P., Arch, J.R., and Ashwell, M. (1984). Brown adipose tissue in the parametrial
fat pad of the mouse. FEBS Lett. 167, 10–14.
Cell Metabolism 26, 198–211, July 5, 2017 211
STAR+METHODS
KEY RESOURCES TABLE
REAGENT or RESOURCE SOURCE IDENTIFIER
Antibodies
anti-GAP43 Cell Signaling sc-17790
anti-Ga olf Santa Cruz Biotechnology sc-383
anti-GFP Thermo Fisher A-6455
anti-HSL Cell Signaling #18381
anti-iba1 Wako #019-19741
anti-Kcnq2 Cell Signaling sc-7793
anti-mCherry Thermo Fisher PA5-34974
anti-OMP Wako
anti-p-HSL Ser 660 Cell Signaling #4126
anti-tubulin Cell Signaling #2144
Bacterial and Virus Strains
pAAV-flex-taCasp3-TEVp Nirao Shah & Jim Wells Addgene plasmid # 45580
AdGFP Jeremy McIntyre (Univ of Florida)
McIntyre et al., 2012
N/A
Chemicals, Peptides, and Recombinant Proteins
2-deoxy-D-[1-14C] glucose (2DG) Amersham CAS: #154-17-6
BSA Sigma A8806
Complete protease inhibitor tablet Roche 11697498001
D-[3-3H] glucose Amersham Biosciences CAS: #2535-38-8
Humulin Eli Lilly
IGF1R transcript ACD; Advanced Cell
Diagnostics Inc., Hayward, CA
target region: 1231-1554,
Acc. No. NM_010513.2
Insulin Roche #11376497001
Insulin Novo Nordisk
Isoproterenol Sigma-Aldrich CAS: #51-30-9
OCT Tissue-Tek VWR CAT #25608-930
PhosSTOP Roche 4906845001
VECTOR Hematoxylin QS Vector laboratories, Burlingame, CA #H-3404
Critical Commercial Assays
UltraSensitive Insulin ELISA kit Crystal Chem #90080
Free Glycerol Determination kit Sigma-Aldrich FG0100-1KT
Quantikine ELISA kit R&D systems
Noradrenaline sensitive ELISA kit Eagle Biosciences SKU: NOU39-K01
Virapur Adenovirus mini purification Virakit Virapur, San Diego, CA
Experimental Models: Organisms/Strains
Mouse: OMP Cre, B6;129P2-Omptm4(cre)
Mom/MomJ
https://www.jax.org/strain/006668 Stock Nr: 006668
Mouse: iDTR Buch et al., 2005 N/A
Mouse: IGF1Rfl/fl, Igf1rtm2Arge
https://www.jax.org/strain/012251 012251
Mouse: TdTomato, B6;129S6-Gt(ROSA)
26Sortm9(CAG-tdTomato)Hze
/J
Madisen et al., 2010 N/A
Oligonucleotides
See Table S1
(Continued on next page)
e1 Cell Metabolism 26, 198–211.e1–e5, July 5, 2017
CONTACT FOR REAGENT AND RESOURCE SHARING
As Lead Contact, Andrew Dillin is responsible for all reagent and resource requests. Please contact Andrew Dillin at dillin@berkeley.
edu with requests and inquiries.
METHOD DETAILS
Animal Care
All procedures were approved by the Animal Care and Use Committee of UC Berkeley and MPI Cologne. Mice were housed at 22
C–
24
C using a 12 hr light/12 hr dark cycle. Animals had ad libitum access to water at all times, and food was only withdrawn if required
for an experiment. All experiments have been performed in adult mice.
Genetic mouse models
All animals described in this study have been generated elsewhere and described in the key resource table. Ablation of mature
olfactory neurons in adult mice was achieved by expressing the inducible diphteria toxin receptor (DTR) specifically in OSNs and conditional
ablation by diphtheria toxin (DT) administration as previously described (Chen et al., 2005). Heterozygous mice expressing
the Cre recombinase under the olfactory marker protein (OMP) promoter were crossed to homozygous mice expressing DTR under a
ubiquitously active promoter blocked by loxP-stop flanking cassette (Buch et al., 2005; Gropp et al., 2005) to generate OMP-DTR+/À
(OMPDTR
) and DTR+/À
(control) animals (Figure 1a). OMP-DTR+/À
and DTR+/À
littermates were generated from 5 breeders pairs of
homozygous DTRÀ/À
mutant mice crossed to OMP-Cre mice in the C57BL/6J background obtained from the Jackson laboratory
(Bar Harbor, ME). DT (40ng/g) was administered intraperitoneally every 3 weeks to maintain olfactory ablation. For DIO, all male
mice were fed a normal chow (PicoLab Rodent 20 5053*, LabDiet) until 6 weeks of age. Subsequently, mice were assigned randomly
to either HFD chow (BioServ, 60% fat calories) or control chow (BioServ, 7% fat calories). IGF1Rflox/flox
mice (Stachelscheid et al.,
2008) were crossed to OMP-Cre mice in order to obtain control (IGF1Rfl/fl) and IGF1RDOMP (IGF1Rfl/flOMPCre+/À) mice. For immunohistochemistry
mice were crossed to tdTomato reporter (B6;129S6-Gt(ROSA)26Sortm9(CAG-tdTomato)Hze
/J (Madisen et al., 2010) to
obtain tdTomato;DOMP (tdTomato+/À
;DOMP+/À
) and IGF1R;tdTomatoDOMP (IGF1Rfl/fl;tdTomato+/À
;OMP+/À
) mice.
Indirect Calorimetry, Physical Activity, and Food Intake
Indirect calorimetric studies were conducted in a Comprehensive Lab Animal Monitoring System in an eight-chamber system
(CLAMS, Columbus Instruments) at UC Berkeley and PhenoMaster System (TSE systems) at MPI. Body composition was assessed
by EchoMRI and NMR Analyzer minispeq mq7.5 (Bruker Optik). Mice were allowed to acclimatize in the chambers for at least 24 hr.
Food and water were provided ad libitum in the appropriate devices and measured by the built-in automated instruments. Locomotor
activity and parameters of indirect calorimetry were measured for at least the following 48 hr. Presented data are average values
obtained in these recordings.
Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
Software and Algorithms
Deseq2 Love et al., 2014 https://bioconductor.org/packages/
release/bioc/html/DESeq2.html
GENCODE https://www.gencodegenes.org
GraphPad Prism 5 https://www.graphpad.com/scientific-
software/prism/
org.Mm.eg.db Bioconductor http://bioconductor.org/packages/release/
data/annotation/html/org.Mm.eg.db.html
RStudio RStudio https://www.rstudio.com
samtools Li et al., 2009 http://samtools.sourceforge.net
SeqMonk Babraham Bioinformatics https://www.bioinformatics.babraham.ac.
uk/projects/seqmonk/
topGO Bioconductor https://bioconductor.org/packages/
release/bioc/html/topGO.html
Tophat2 https://ccb.jhu.edu/software/tophat/
index.shtml
Trim Galore! Babraham Bioinformatics https://www.bioinformatics.babraham.ac.
uk/projects/trim_galore/
Cell Metabolism 26, 198–211.e1–e5, July 5, 2017 e2
Food finding test
The mice were single caged a week before the experiment started. They were habituated to the small food pellets, the perforated
eppendorf tube and the round buckets that were used during the experiment. The experiment took place in three consecutive
days with two trials per day. The mice were given up to 15 min to find the hidden under 4cm of bedding perforated eppendorf
tube containing the food pellets.
Social odor finding test
Urine was collected from ten female mice. A q-tip was dipped in the urine and placed in a perforated Eppendorf tube which was
hidden under 4cm of bedding in a round bucket. The mice were given up to 15 min to find the hidden perforated Eppendorf tube
containing the food pellets hidden under 4 cm of bedding.
Habituation-dishabituation test
The mice were exposed to five smells in three consecutive 2min trials per smell. The time that the mouse spent with its nose 2cm or
closer to the odor source was measured using an automated camera system.
Metabolic studies
To perform the GTT, mice were fasted for 6h and tails were bled for the initial blood glucose concentration measured using a One
Touch Ultra glucometer (LifeScan) or Bayer Contour glucometer (Bayer). Glucose (2g/kg weight) was intraperitoneally administered
and blood glucose was measured at indicated times after injection. For the ITT, 5 hr fasted mice were injected with 0.75-1u/kg of
human insulin (Humulin; Eli Lily, Roche or Novo Nordisk) and glucose was measured as in the GTT. Serum insulin was measured using
an UltraSensitive Insulin ELISA kit (Crystal Chem). Serum levels of fasted leptin were measured using a Quantikine ELISA Kits (R&D
systems). Noradrenaline were measured using a sensitive ELISA Kits (Eagle Biosciences), and fasted levels of free fatty acids were
measured using a Free Fatty Acid Quantitation Kit (Sigma Aldrich). To measure Phospho (Ser 660) HSL levels in WAT, Isoproterenol
(10mg/kg) or saline was injected intraperitoneally 15 min prior to animal’s euthanasia.
Ex vivo lipolysis
Gonadal fat pads were isolated from mice and weighed. Each pad was cut into 3 equal pieces and incubated in Krebs-Ringer Bicarbonate
Buffer containing 1% fatty acid–free BSA (Sigma-Aldrich). The samples were treated with either PBS or isoproterenol (25 mM,
Sigma-Aldrich) at 37
C with mild shaking at 300 rpm. After 120 min, glycerol release was measured using the Free Glycerol Determination
Kit (Sigma-Aldrich).
Protein analysis
Adipose tissue was homogenized in TNET buffer (50mM Tris-HCl, 5mM EDTA, 150mM NaCl), supplemented with Complete protease
inhibitor tablet (Roche), PhosSTOP (Roche). Protein concentrations were determined using Bradford. 20-50mg of samples were
loaded on SDS/Page gels (Invitrogen) then transferred to nitrocellulose membranes. Ponceau S staining was used to confirm equal
loading. Immunoblotting was done in 3% BSA containing TBS-T with antibodies against Phospho HSL (Ser 660), HSL, a-tubulin.
Leptin sensitivity
Recombinant mouse leptin (R&D systems) was injected intraperitoneally 90 min before the onset of the dark cycle for 3 consecutive
days (3 mg/kg) in an independent cohort of obese OMPDTR
and control animals, before and 2 weeks post DT delivery. Food intake
and body weights were recorded daily.
Quantitative PCR analysis
RNA was isolated using trizol/chloroform extraction and RNEasy QIAGEN columns. RNA was converted into cDNA using quantitect
reverse-transcription kit. Gene expression was assessed by qPCR using SYBR green.
Adenovirus propagation and titration
All adenoviruses used in this study are replication-deficient (E1 deleted), serotype 5 adenovirus, and were produced following the
ViraPower protocol (Invitrogen). The amplified virus was purified with the Virapur Adenovirus mini purification Virakit (Virapur, San
Diego, CA), which was followed by dialyzation into a solution of 2.5% glycerol, 25 mM NaCl and 20 mM Tris-HCl, pH 8.0, at 4
C overnight
before storing it at À80
C. AdGFP was a gift from Jeremy McIntyre (Univ of Florida)(McIntyre et al., 2012), and Ad-taCasp3TEVp
under the EF-1a promoter was subcloned from AAV-flex-taCasp3-TEVp (Yang et al., 2013) and inserted into the adenoviral
vector pAD/PL-DEST (Invitrogen). pAAV-flex-taCasp3-TEVp was a gift from Nirao Shah & Jim Wells (Addgene plasmid # 45580).
Stock titrations were: Ad-GFP, 1.11 3 1012
particles per mL; Ad-taCasp3, 1.10 3 1012
particles per mL.
Immunofluorescence and histological studies
Whole mouse heads were fixed in 4% paraformaldehyde, then decalcified in 10% ethylenediaminetetraacetic acid dissolved in PBS
at pH 7.6 for 2-3 weeks to perform olfactory epithelium staining. Tissues were cryoprotected in 30% sucrose in PBS overnight, then
frozen in OCT (Tissue-Tek) and sectioned on a cryostat. 12-20 mm thick sections were immunoblotted with primary antibodies in PBS
e3 Cell Metabolism 26, 198–211.e1–e5, July 5, 2017
containing 2% Donkey serum and 0.4% Triton x100. Antibodies used were diluted at 1:100 or 1:200 and secondary antibodies
(Alexxa) were diluted at 1:600. Sections were imaged using a LSM 710 confocal microscope and a Keyence microscope BZX700.
For quantification of OMP+ cells, coronal sections of MOE lining the septum were chosen. Semi-quantification for OMP+ cells
were measured by point-counting morphometry using the VS-ASW software. For histological analysis, tissues were immersed in
formalin overnight, dried in 70% Ethanol and embedded in paraffin. 8mm thick sections were stained with hematoxylin and eosin
and finally imaged with a VS120 slide scanner (Olympus).
Hyperinsulinemic-euglycemic CLAMPS
Catheter implantation in the jugular vein was performed as previously described (Fisher and Kahn, 2003). Only mice that had regained
at least 90% of their preoperative body weight after 6 days of recovery were included in the experimental groups. After starvation for
4 hr the clamp was performed in awake animals. After a bolus infusion (5 mCi) of D-[3-3H] glucose (Amersham Biosciences) tracer
solution, the tracer was infused continuously (0.05 mCi/min) for the duration of the experiment. At the end of the 50 min basal period,
a blood sample (50 ml) was collected for determination of basal parameters. Insulin (human regular insulin; Novo Nordisk) solution
containing 0.1% BSA (Sigma) was infused at a fixed rate (4 mU/g/min) following a bolus infusion (20 mU/g). Blood glucose levels
were determined every 10 min (B-Glucose Analyzer, HemoCue), and physiological blood glucose levels (between 120 and
145 mg/dl) were maintained by adjusting a 20% glucose infusion (DeltaSelect). 60 min after the insulin bolus infusion, which coincides
approximately to 60 min before steady state was achieved, a bolus of 2-deoxy-D-[1-14C] glucose (2DG) (10 mCi, Amersham) was
infused. Steady state was ascertained when glucose measurements were constant for at least 30 min at a fixed glucose infusion
rate and was achieved within 100 to 130 min. During the clamp experiment, blood samples (15 ml) were collected after infusion of
the 2DG at time points 0, 5, 15, 25, 35 min, etc. until reaching the steady state. During the steady state, blood samples (50 ml)
were collected for the measurement of steady-state parameters. All blood samples were kept at room temperature until centrifugation;
serum samples were stored at À20
C. At the end of the experiment, mice were euthanized by cervical dislocation, and brain,
liver, white adipose tissue (WAT), and skeletal muscle were dissected and stored at À80
C.
Plasma [3-3H] glucose radioactivity at basal and steady state was measured as previously described (Fisher and Kahn, 2003).
Plasma deoxy-[1-14C] glucose radioactivity was directly measured in a liquid scintillation counter. WAT and skeletal muscle lysates
were processed through ion-exchange chromatography columns (Poly-Prep Prefilled Chromatography Columns, AG1-X8 formate
resin, 200–400 mesh dry; Bio-Rad) to separate 2DG from 2DG-6-phosphate (2DG6P).
Glucose turnover rate (mg/kg/min) was calculated as previously described (Fisher and Kahn, 2003). In vivo glucose uptake for
brain, WAT, and skeletal muscle (nmol/g/min) based on the accumulation of 2DG6P in the respective tissue and the disappearance
rate of 2DG from plasma was calculated as described previously (Ferre´ et al., 1985).
Adipose tissue oxygen consumption
Adipose tissue was minced into small pieces and rinsed two times with PBS through a 100 mm strainer. 10 mg of tissue was placed
into each well of an XF24-well Islet Flux plate (Seahorse Bioscience). Tissue chunks were washed with 1 mL of assay media (1x KHB,
2.5mM glucose, 0.5mM carnitine, 0.5mM HEPES, 0.24mM MgCl2; pH to 7.4 at 37
C), then 450 mL of fresh assay media was added to
every well and plate was run in Seahorse XF-24 (mix 3 min, wait 2 min, measure 3min). Results of OCR and ECAR were normalized to
protein concentration.
Bomb fecal calorimetry
Freeze-dried feces was pooled and pulverized with a ceramic mortar and pestle. Caloric content of feces was measured with an 1108
Oxygen Combustion Bomb calorimeter and normalized to replicate benzoic acid standards. n = replicate explosion of pooled feces.
RNA sequencing
RNA-seq library preparation and measurement was performed by the Cologne Center for Genomics, Cologne, Germany. Stranded
TruSeq RNA-seq library preparation was principally conducted as previously described (Sultan et al., 2012) with 2 mg of extracted
RNA and poly-A enrichment. Barcoded libraries were sequenced with 100 bp paired-end reads on an Illumina HiSeq2500.
Raw sequence reads were trimmed to remove adaptor contamination and poor-quality reads using Trim Galore! (v0.3.7,
parameters:–paired–length 25). Trimmed sequences were aligned using Tophat2 (Kim et al., 2013) (v2.0.14, parameters:–nomixed–library-type
= fr-firststrand -g 2 -r 500–mate-std-dev 525) against the mouse reference genome (build GRCm38), supplying
GENCODE annotation (Harrow et al., 2012) (release M2, main annotation) for improved mapping. Multi-mapped reads were filtered
using samtools (Li et al., 2009) (v1.2, parameters: view -F 0x100 -b –h). Data visualization and analysis was performed using
SeqMonk, custom RStudio and the following Bioconductor packages: Deseq2 (Love et al., 2014), topGO and org.Mm.eg.db. To
account for tissue specific expression, we defined all genes with an FPKM of > 2 in at least half of all samples as ‘expressed’. Unless
stated otherwise, the set of expressed genes was used as background for all conducted functional enrichment analyses with topGO.
For determining the maturity state of olfactory neurons, a previously published study (Nickell et al., 2012) was used as reference.
Mature and immature marker genes lists were used to calculate the enrichment among the entire set of significant differentially
expressed genes (p-val < 0.05 after running Deseq2) as well as enrichments among the respective sub-populations of up- and
downregulated genes, using one-sided Fisher’s exact test. Additionally, random subsets of the same size as the immature/mature
Cell Metabolism 26, 198–211.e1–e5, July 5, 2017 e4
marker sets were drawn out of all significant differentially expressed genes and their respective log2foldchange distributions were
compared to the two distributions of the immature/mature marker set.
RNAscope analysis
Chromogenic in situ hybridization for the detection of IGF1R was performed using RNAscope (ACD; Advanced Cell Diagnostics Inc.,
Hayward, CA). A custom designed probe targeting the floxed region (exon 3) of the IGF1R transcript (target region: 1231-1554, Acc.
No. NM_010513.2). Negative and positive control probes recognizing dihydrodipicolinate reductase, DapB (a bacterial transcript)
and cyclophilin (PPIB), respectively, were processed in parallel with the target probe to ensure tissue RNA integrity and optimal assay
performance. All pretreatment reagents, detection kit and wash buffer were purchased from ACD. All incubation steps at 40
C were
performed using a humidified chamber and a HybEz oven (ACD).
The day before the assay, slides were removed from the À80 freezer, allowed to thaw and were then washed for 5 min in 0.1 M
phosphate buffered saline (PBS, pH 7.4) in order to remove tissue embedding medium. Slides were then briefly dipped in diethylpyrocarbonate
(DEPC) treated Millipore water and allowed to dry at room temperature, before being incubated at 60
C overnight.
On the day of the assay, the tissue underwent a pretreatment procedure to allow for optimal probe penetration and assay performance.
First, sections were incubated with RNAscope Hydrogen Peroxide (ACD) for 20 min at room temperature. Slides were then
submerged in Target Retrieval (ACD) for 8 min, heated to 98.5-99.5
C, followed by two brief rinses in DEPC water. The slides were
quickly dehydrated in 100% ethanol and allowed to air dry for 5 min. A hydrophobic barrier was drawn around each section with an
Immedge barrier pen. Each section was then incubated with Protease Plus (ACD) for 20 min at 40
C, followed by 10 min at room
temperature. The subsequent steps, i.e., hybridization of the probe, the amplification steps and detection of the probe, were performed
according to the online protocol for RNAscope 2.5 High Definition (HD) BROWN Assay.
In brief, the procedure included the following steps; hybridization of the IGF1R probe at 40
C for 2 hr, 2x2 min washes in Wash
buffer (ACD), incubation with Amp1 at 40
C for 30 min, 2x2 min washes, Amp2 at 40
C for 15 min, 2x2 min washes, Amp3 at
40
C for 30 min, 2x2 min washes, Amp4 at 40
C for 15 min, 2x2 min washes, Amp5 at room temperature for 30 min, 2x2 min washes,
Amp6 at room temperature for 15 min, 2x2 min washes, incubation in DAB substrate for 10 min at room temperature. After this step,
sections were briefly washed in tap water two times, counterstained with VECTOR Hematoxylin QS (40 s; Vector laboratories, Burlingame,
CA), dehydrated for 2 min each in 70%, 95% and 95% ethanol. Finally, slides were defatted in xylene for 5 min before being
coverslipped with Cytoseal mounting medium.
Data and Software Availability
Transcriptome profiles of olfactory sensory neurons derived from IGF1RDOMP
mice have been deposited in the NCBI’s Gene Expression
Omnibus repository under ID code GSE85329: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE85329.
Statistical methods
GraphPad Prism 5 was used for statistical analysis. Data are expressed as mean ± SEM p Values were calculated using unpaired twotailed
unequal variance Student’s t test unless or two-way ANOVA with post hoc Bonferroni tests (when specified). P values less than
0.05 were considered significant.
e5 Cell Metabolism 26, 198–211.e1–e5, July 5, 2017