Article

The role of somatosensory innervation of

adipose tissues

https://doi.org/10.1038/s41586-022-05137-7 Yu Wang1,2, Verina H. Leung1, Yunxiao Zhang1,2, Victoria S. Nudell1, Meaghan Loud1,2,

M. Rocio Servin-Vences1,2, Dong Yang1, Kristina Wang1, Maria Dolores Moya-Garzon3,
Received: 3 February 2022
Veronica L. Li3, Jonathan Z. Long3, Ardem Patapoutian1,2 ✉ & Li Ye1 ✉
Accepted: 22 July 2022

Published online: 31 August 2022

Adipose tissues communicate with the central nervous system to maintain
Open access whole-body energy homeostasis. The mainstream view is that circulating hormones
Check for updates secreted by the fat convey the metabolic state to the brain, which integrates
peripheral information and regulates adipocyte function through noradrenergic
sympathetic output1. Moreover, somatosensory neurons of the dorsal root ganglia
innervate adipose tissue2. However, the lack of genetic tools to selectively target these
neurons has limited understanding of their physiological importance. Here we
developed viral, genetic and imaging strategies to manipulate sensory nerves in an
organ-specific manner in mice. This enabled us to visualize the entire axonal
projection of dorsal root ganglia from the soma to subcutaneous adipocytes,
establishing the anatomical underpinnings of adipose sensory innervation.
Functionally, selective sensory ablation in adipose tissue enhanced the lipogenic and
thermogenetic transcriptional programs, resulting in an enlarged fat pad, enrichment
of beige adipocytes and elevated body temperature under thermoneutral conditions.
The sensory-ablation-induced phenotypes required intact sympathetic function. We
postulate that beige-fat-innervating sensory neurons modulate adipocyte function
by acting as a brake on the sympathetic system. These results reveal an important role
of the innervation by dorsal root ganglia of adipose tissues, and could enable future
studies to examine the role of sensory innervation of disparate interoceptive systems.

Mammalian adipose tissue is highly innervated. This innervation has non-neuronal transient receptor potential vanilloid (TRPV1)-expressing
been primarily studied for its efferent functions through tyrosine cells11–13. Moreover, capsaicin-mediated denervation is clearly biased
hydroxylase (TH)-expressing, noradrenaline-secreting sympathetic towards the heat-sensitive and nociceptive TRPV1-expressing neurons
fibres3,4. These sympathetic fibres act on β-adrenergic receptors and (Aδ- and C-fibres)14,15, potentially dampening the loss-of-function phe-
have a well-recognized role in regulating thermogenesis and lipid notypes9. By contrast, it is now known that a prominent non-peptidergic
metabolism in both brown and beige fat5. DRG population also expresses TH16, calling into question the use of
The afferent function of adipose innervation is much less under- TH as a selective marker for sympathetic fibres in fat. These caveats
stood. Adipose tissue is one of the few internal organs that receive little motivated us to develop new imaging, molecular and circuit-specific
vagal sensory innervation2. By contrast, somatosensory fibres originat- tools to examine sensory innervation in adipose tissues.
ing from the dorsal root ganglia (DRGs)—best known for skin and muscle
sensation—have been reported to innervate adipose tissue in rats and
hamsters2,6, but the extent and importance of this innervation remain to Direct visualization and characterization of the DRG
be determined across species7. Although pioneering work using herpes projections to fat
viral tracing has elegantly mapped the central projection of fat affer- In conventional tracing studies, viruses or dyes are injected into the fat
ent8, the functional importance of this sensory innervation remains to be retrogradely transported back to the DRG somas and assessed by
less understood because traditional chemical or surgical denervation histology. This indirect measurement is susceptible to varying tracer
of sensory fibres resulted in only minor phenotypes in the hamsters2,9. efficiency across hosts, potentially contributing to the discrepancy
As we begin to appreciate the cellular and molecular heterogeneity of observed across animal species6–8. Ideally, direct visualization of the
DRG neurons10 and adipose tissue5, the specificity of earlier classic den- entire projection from DRG soma to the target organs, for example,
ervation approaches needs to be re-evaluated. For example, whereas through an axon-filling fluorophore, would provide the most reliable
surgical denervation cannot discriminate between sensory and sympa- anatomical proof. However, the peripheral branch of the mouse DRG
thetic fibres as they travel together in bundles, capsaicin denervation, travels several centimetres before reaching its targets, making it impos-
which was thought to selectively ablate sensory fibres in fat, can target sible to visualize using conventional histology. We recently developed

Department of Neuroscience, Dorris Neuroscience Center, Scripps Research, San Diego, CA, USA. 2Howard Hughes Medical Institute, Chevy Chase, MD, USA. 3Department of Pathology,
1

Stanford School of Medicine, Sarafan ChEM-H, Stanford University, Stanford, CA, USA. ✉e-mail: [email protected]; [email protected]

Nature | Vol 609 | 15 September 2022 | 569

Article
a Anterograde labelling b d
DRG-FP DRG-FP
AAV-FP
mm
DRG 14

iWAT DRG
SChG
(T13)
DRG
Fluorescence Low High DRG
HYBRiD c (L1)

42 mm
light-sheet DRG-FP
clearing
imaging
whole torso

iWAT
LN

3d view 30 um 32 mm
e Retrograde labelling CTB-647 (iWAT) f g Anterograde labelling
CTB-488 (Skin) CTB retrograde labelling Perilipin DRG-FP

iWAT DRG Skin DRGs

n = 526 n = 368
Maximum-
CTB-488 CTB-647 Overlap = 1 intensity
Skin iWAT n = 4 animals projection
DRG

h DRG-FP TH Perilipin i j DRG-FP TH

DRG-FP+ thin fibres
TH+
TH–
n = 31
n = 46 40.3%
59.7%

Total = 77 views

Fig. 1 | Adipose tissues receive robust somatosensory innervations. a, The T13). f, Quantification of CTB-positive cell numbers from labelling in the iWAT
workflow of mapping sensory innervation in adipose tissues. Tissues from and skin. n = 4 mice. g, Representative image of virally labelled DRG fibre in
mice with AAV expressing fluorescent protein (FP) injected in T13/L1 DRGs close apposition to an adipocyte. h, Representative image of TH− and TH+
(vertebral level T13 and L1) were processed for en bloc HYBRiD clearing and parenchymal DRG innervation. i, Quantification of the percentage of TH+ DRG
fluorescence microscopy imaging. b–d, Representative 3D image volume of fibres (77 views in 13 images from 3 biological samples). j, Representative
AAV-labelled DRGs (T13) by light-sheet imaging (b), DRG fibres in the iWAT by image of virally labelled DRG fibre travelling along vasculature. The white
confocal imaging (showing lymph node (LN) as a landmark) (c) and DRG axonal arrows mark DRG fibres, and the white triangle marks TH-stained sympathetic
projections in an adult mouse torso (d). The colour gradient suggests relative fibres. Scale bars, 200 μm (b and e), 500 μm (c), 5 mm (d) and 30 μm in (g, h and j).
intensity. LN, lymph node. e, Schematic of dual-colour CTB labelling from iWAT
and flank skin (left) and representative whole-mount image of DRGs (right,

HYBRiD, specifically designed for en bloc fluorescence visualization Furthermore, retrograde cholera toxin subunit B (CTB) labelling
of large tissues17, paving the way to directly characterize the intact was also used to confirm the imaging findings. As expected6,8,20,21, iWAT
sensory innervation from DRGs to fat. receives sensory innervation and sympathetic innervation mainly from
Another barrier to selectively labelling sensory innervation in the T11–L3 DRGs and paravertebral sympathetic chain ganglia (SChGs),
fat is that common pan-DRG Cre transgenic mice (such as Pirt-Cre, respectively; but not from nodose ganglion (vagal nerve) or collateral
Scn10a-Cre, Advillin-CreERT218) also target sympathetic neurons sympathetic ganglia (Extended Data Fig. 2a,b). Fat-projecting DRGs
(Extended Data Fig. 1a). We therefore adopted an intraganglionic DRG consist of multiple neuron types with an enrichment of peptidergic
surgery in mice to directly inject a recombinant adeno-associated virus and myelinated fibres (Extended Data Fig. 2h,i). Given the proximity
(AAV) expressing fluorescent protein into individual DRGs without between iWAT and skin (the primary DRG target), we examined whether
transducing sympathetic ganglia (Fig. 1a and Extended Data Fig. 1b). fat innervation is part of the cutaneous sensory circuit. Dual-colour CTB
For the rest of the study, we focused on the inguinal white adipose tis- labelling from iWAT and neighbouring flank skin showed that these
sue (iWAT), a beige fat pad with well-established roles in mouse physi- two organs are innervated by two non-overlapping DRG populations
ology19. We injected a fluorescent-protein-expressing AAV into the (Fig. 1e,f). Visceral fat (epidydimal white adipose tissue (eWAT)) also
thoracolumbar DRGs (vertebral level T13 and L1 (T13/L1)6), to target all receives sensory innervation at comparable vertebral levels but from
projecting fibres from these two ganglia. After en bloc HYBRiD clearing a separate population (Extended Data Fig. 2d,e). Together, these 3D
and light-sheet imaging of the whole torso (Fig. 1a), the entire projec- imaging and retrograde labelling data demonstrate that distinct neu-
tion from DRG soma to iWAT travelling across 1.2 cm can be resolved rons in the thoracolumbar DRGs robustly project to adipose tissues.
(Fig. 1b–d and Supplementary Video 1), unequivocally demonstrating The volumetric images enabled us to examine the morphology and
that thoracolumbar DRGs directly innervate iWAT (Extended Data topology of somatosensory terminals in adipose tissue in a high level
Fig. 1c). of detail (Fig. 1g–j). The sensory fibres in iWAT can be divided into at

570 | Nature | Vol 609 | 15 September 2022

 a c Ef1a e f
ROOT AAV9 120 0.0028 1.2 <0.0001

DRG
mCherry
DTA

CTB+ cell counts

mScarlet 1.0

CTB+ cell counts

CTB
80

Normalized
DTA
DTA
0.8

DRG
0.6
40 0.4

iWAT
YFP Cre 0.2
Cre– Cre+ 0 0
Cre– Cre+ Cre– Cre+
d Cre– Cre+ Cre– Cre+
SChG

CTB CTB

b ROOT AAV9

SChG
mScarlet

DRG
mCherry mCherry
Liver

Enhanced

Low High

Fig. 2 | Combinatorial strategy for specific iWAT-DRG manipulation. strategy for unilateral sensory ablation. Each mouse has AAV-mCherry-
a,b, Comparison of ROOT and AAV9 for retrograde labelling in the iWAT. flex-DTA injected into the bilateral T13/L1 DRGs, and ROOT-YFP or ROOT-Cre
a, Representative whole-mount images of DRGs (T13) and SChGs (T12) labelled injected into the iWAT unilaterally. Three weeks after surgery, CTB-647 was
by ROOT-mScarlet or AAV9-mScarlet and a sequential CTB-647 injection from injected into the iWAT bilaterally. d, Representative whole-mount images of
iWAT. Some of the AAV and CTB double-positive cells are highlighted by white contralateral (Cre−) and ipsilateral (Cre+) DRGs (T13) and SChGs (T12).
arrows. CTB labels 26.42 ± 6.49% of ROOT-labelled neurons, and 9.95 ± 0.92% e,f, Quantification of CTB+ cell numbers (e) and normalized cell numbers
of AAV9-labelled neurons in T13/L1 DRGs. Data are mean ± s.e.m. n = 4 mice per (f) in T13 and L1 DRGs labelled from iWAT. n = 10 mice. Statistical analysis was
group. b, Representative images of livers from animals with ROOT-mScarlet performed using two-tailed paired t-tests. P values are shown at the top. For
or AAV9-mScarlet injected into the iWAT. c–f, Combinatorial viral strategy for a,b and d, scale bars, 200 μm.
Cre-dependent ablation of iWAT DRGs. c, Schematic of the combinatorial viral

least two main types: (1) larger bundles travelling along the vasculature high retrograde potential from iWAT to DRGs (Extended Data Fig. 4a).
(Fig. 1j and Extended Data Fig. 3e) and (2) parenchymal innervation, in We adopted a published viral engineering pipeline22 to generate ran-
which the sensory nerve terminates in close apposition to adipocytes domized mutants of AAV9 (Extended Data Fig. 4b,c). Although our
(Fig. 1g,h and Extended Data Fig. 3a,d). TH has long been regarded as a initial intent was to improve retrograde efficiency from fat to DRGs,
sympathetic marker and a surrogate for adipose sympathetic innerva- the evolved new retrograde vector optimized for organ tracing (or
tion4,7. For the larger bundles, sensory fibres travel together with TH+ ROOT) is more desirable mainly for its significantly reduced off-target
sympathetic fibres along the vasculature but rarely wrap the vessel as expression such as in SChGs, contralateral DRGs and the liver (Fig. 2a,b
the latter typically do4 (Fig. 1j and Extended Data Fig. 3e). Notably, in and Extended Data Fig. 4d–g).
the parenchymal portion, nearly 40% of thin sensory terminals close ROOT provides an opportunity to specifically ablate the sensory
to adipocytes are immunopositive for TH (Fig. 1h,i and Extended Data innervation in fat—we injected Cre-dependent diphtheria toxin subunit
Fig. 3d), challenging the traditional view that TH is an exclusive sympa- A (DTA) construct (mCherry-flex-DTA) into the T13/L1 DRGs bilaterally
thetic marker in fat and, importantly, suggesting that earlier studies while injecting Cre- or YFP-expressing ROOT unilaterally in the iWATs
based on TH might be confounded, at least partially, by the sensory (Fig. 2c). On the basis of CTB quantification, we noticed a decrease of
innervation. Thus, with these findings in mind, it is imperative to estab- approximately 40% in fat-projecting neurons in Cre+ ipsilateral DRGs
lish the specific functions of sensory innervation in adipose tissue. compared with the control side (Fig. 2d–f), comparable to the efficiency
achieved by previous viral DTA ablations23. No difference was observed
in the SChGs (Fig. 2d and Extended Data Fig. 5a). Importantly, the struc-
Selectively targeting adipose sensory innervation ture and function of sensory terminals in the flank skin remained intact
Sympathetic output in adipose tissues works through β-adrenergic (Extended Data Fig. 5b–e). Together, our projection-defined, partial
receptors, enabling the use of catecholaminergic neurotoxins (such sensory ablation in the iWAT provides a specific loss-of-function model
as 6-hydroxydopamine (6-OHDA)) and adrenergic ligands to specifi- to study the adipose-innervating DRGs.
cally manipulate their activities. By contrast, sensory fibres are more
diverse and respond to different sensory modalities and therefore
cannot be manipulated based on a single signalling pathway. Thus, a Fat gene program changes after sensory ablation
neuron-specific, projection-defined genetic approach is necessary to We next investigated how the loss of sensory innervation affects the
study sensory innervation in adipose tissue. molecular programs of the iWAT. We compared the sensory-ablated
A combination of axonal target injection of retrograde Cre and iWAT (Cre+) with the non-ablated contralateral fat pad (YFP+) within
somatic expression of a Cre-dependent payload has been widely used the same animal. RNA-sequencing (RNA-seq) analysis was performed
to manipulate projection-specific circuits in the brain. However, legacy in the fat pads 3–4 weeks after viral injection (Fig. 3a–d). Unbiased Gene
peripheral viral tracers such as pseudorabies virus and herpes sim- Ontology analysis showed that both fatty acid and lipid metabolism
plex virus are highly toxic, restricting their use beyond acute anatomi- and cold-induced thermogenesis pathways (Fig. 3d) were enriched by
cal mapping. In search for newer and safer viral vectors suitable for sensory ablation. At the individual-gene level, well-established ther-
long-term functional manipulations, we found that AAV9 exhibited a mogenic brown/beige cell markers were significantly upregulated

Nature | Vol 609 | 15 September 2022 | 571

Article
a b Cre+ versus Cre– c
mCherry
-flex-DTA 25
DRG 2

Cre–
20 1

–log10[P]
iWAT 15 Cidea
Cre Fasn 0
YFP

Treatment Cre+
10 –1
3–4 Ucp1 –2
weeks 5 2

z-score
Individual
–1.5 1.5 Elovl3

Adrb3

Acacb
Acaca
Fasn

Ucp1
Elovl3
0
RNA-seq –5 0 5
Cre– Cre+ log2[fold change]
d e Thermogenesis Lipogenesis
GO pathways –log10[P]
Ucp1 Elovl3 Cidea Fasn Acacb ChREBPE
Monocarboxylic acid 12.3 0.0483 0.0029 0.0022 0.0200 0.0492 0.0071
metabolic process 210 27 25 23 23 25
Defence response 26 24 24
22

Fold change

Fold change
22

Fold change
Fold change
11.2 25

Fold change
Fold change
to protozoan 23 23
25 24 21 21
23 22 22
Triglyceride biosynthetic
7.6 22 21 20 20 21
process 20 21 20 20
Regulation of cold- 20 2–1 2–1
7.4 2–1 2–1 2–1
induced thermogenesis 2–2 2–2 2–2
2–5 2–2 2–2
Lipid storage 6.9 Cre– Cre+ Cre– Cre+ Cre– Cre+ Cre– Cre+ Cre– Cre+ Cre– Cre+
Thermogenesis
f g
mCherry Cre– Cre+ Ucp1 Elovl3 Cidea
-flex-DTA 28 0.0178 0.3684 28 0.0083 0.2574 26 0.0119 0.3538
DRG 26 26 24

Fold change

Fold change
Fold change

24 24
iWAT 22 22 22
YFP Cre 20
2–2 20 20
2–4 2–2
2–3 weeks 2–4 2–2
2–6 Cre P = 0.0088 Cre P = 0.0038 Cre P = 0.0064
2–8 OHDA P = 0.0163 2–6 OHDA P = 0.0073 2–4 OHDA P = 0.0662
Saline 6-OHDA Saline 6-OHDA Saline 6-OHDA
Lipogenesis
Cre– Cre+ Fasn Acacb ChREBPE
0.0060 0.3409 0.0147 0.6315 0.0069 0.1767
or 24 24 26
Saline 23 24
Fold change
Fold change

Fold change
22 22
21 22
20 20
6-OHDA 20
2–1
1 week Cre P = 0.0038 2–2 Cre P = 0.0134 2–2 Cre P = 0.0025
2–2 OHDA P = 0.2954
2–3 OHDA P = 0.1308
2–4
OHDA P = 0.2481
RNA expression analysis Saline 6-OHDA Saline 6-OHDA Saline 6-OHDA

Fig. 3 | Ablation of iWAT DRGs upregulates thermogenic and lipogenic Statistical analysis was performed using two-tailed paired t-tests.
transcriptional programs. a–e, Transcriptional profiling of the iWAT after f,g, Transcriptional analysis of iWAT with sensory ablation and sympathetic
Cre-dependent sensory ablation. a, Schematic of transcriptional profiling. ablation. f, Schematic of sensory and sympathetic double ablation. Mice
The iWAT from mice with Cre-dependent unilateral sensory ablation was with Cre-dependent unilateral sensory ablation were subjected to bilateral
processed for RNA-seq analysis. b, RNA-seq results. Statistical analysis was sympathetic 6-OHDA denervation. g, RT–qPCR analysis of thermogenic and
performed using Wald tests. c, Heat map of upregulated genes identified by lipogenic genes in the iWAT with or without sympathetic denervation. n = 6
RNA-seq analysis. d, Gene Ontology (GO) enrichment analysis of upregulated mice per group. Statistical analysis was performed using two-way analysis of
genes. e, qPCR with reverse transcription (RT–qPCR) analysis of thermogenic variance with Sidak’s multiple-comparisons test.
and lipogenic genes in the iWAT after unilateral sensory ablation. n = 5 mice.

in the ablated side, including Ucp1 (10.4-fold, P = 1.6 × 10−4), Elovl3 sensory ablation was blunted in 6-OHDA-treated animals (Ucp1: 9.4-fold
(28.8-fold, P = 5.2 × 10−5) and Cidea (9.1-fold, P = 3.3 × 10−6). Markers (saline) to 2.6-fold (6-OHDA); Elovl3: 12.0-fold to 2.9-fold; Cidea: 4.7-fold
for de novo lipogenesis (DNL)24, such as Fasn (3.5-fold, P = 7.1 × 10−7) to 1.8-fold; Fasn: 3.5-fold to 1.6-fold; Acacb: 3.6-fold to 1.4-fold; ChREBPβ:
and Acacb (2.6-fold, P = 1.1 × 10−4), also showed elevated expression 6.7-fold to 2.5-fold) (Fig. 3g and Extended Data Fig. 6e,f), indicating
(Fig. 3b,c). The concurrent activation of opposing lipid oxidation and that sensory-regulated gene expression changes are at least partially
lipogenesis pathways is a unique but well-documented mechanism dependent on the intact sympathetic function.
in the adipose tissue to ensure fuel availability for heat production
under cold or β-adrenergic stimulation25,26. The mRNA expression of
these genes, together with that of the gene encoding ChREBPβ (Mlxipl; Sensory regulation of adipose physiology
called 'ChREBPβ' here), a transcriptional master regulator of DNL27,28, We next examined how the sensory-ablation-induced gene changes
was confirmed by quantitative PCR (qPCR), showing increases in the affect adipose functions (Fig. 4a–c). We observed higher phosphoryla-
ablated iWAT (Fig. 3e) but not in non-targeted eWAT or interscapular tion of hormone-sensitive lipase (HSL) and enrichment of multilocular
brown adipose tissues (iBAT) (Extended Data Fig. 6a,b). beige adipocytes in the unilateral sensory-ablated iWAT (Fig. 4c and
As thermogenic and lipogenesis programs are both down- Extended Data Fig. 7d–f), consistent with upregulation of the ther-
stream of sympathetic signalling24,26,29, we next tested whether the mogenic program (Fig. 3). This resembles the beiging of iWAT after
sensory-elicited gene expression changes are dependent on intact cold exposure or β-adrenergic agonism. However, under these condi-
sympathetic innervation. We bilaterally injected 6-OHDA—a catecho- tions, the wild-type fat pads normally shrink in size, presumably due
laminergic toxin for selective sympathetic denervation (Extended to stronger lipid utilization than DNL. By contrast, we observed an
Data Fig. 6c,d)—into the iWAT of mice that had previously undergone increase in fat mass of the sensory-ablated iWAT compared with the
unilateral sensory ablation (Fig. 3f). This double-ablation experiment contralateral controls (Fig. 4a,b and Extended Data Fig. 7a–c), sug-
showed that the induction of thermogenic and lipogenic genes by gesting that the sensory innervation not only counteracts sympathetic

572 | Nature | Vol 609 | 15 September 2022

 with a high-fat diet, despite showing a small difference in body weight,
the mice with adipose sensory ablation exhibited markedly improved
a b c glucose tolerance compared with the control mice (Extended Data
Cre– Cre+
Cre– Cre+
0.15
0.0008 Fig. 8j–n). Such disproportional changes in glucose tolerance and body
iWAT
weight resemble the phenotype reported in PRDM16-induced beiging

Weight (g)
0.10 in transgenic models33,34, suggesting that adipose sensory ablation
could protect mice from diet-induced glucose intolerance through
0.05
beige-fat-related mechanisms.
Cre– Cre+

d e Cre– Cre+
Cre– Cre+ 0.0142 0.0249 0.0158 Discussion
24
iWAT
23 The conventional view posits that adipose signals transmit to the brain

Fold change
Lipogenesis
22
21
through slow, diffusive circulating hormones. Here the unequivocal
20 anatomical proof of sensory innervation in fat provides a circuitry
2–1
2–2 basis for a potential fast, spatially encoded neural transmission from
Fasn Acacb ChREBPE
0.0149 0.0450 0.0095
the peripheral organs to the brain. Indeed, we go on to show evidence
24
that DRG sensory neurons act as an inhibitory break on the local sym-
Thermogenesis

23
Fold change

22
21
pathetic function, reminiscent of the role of vagal baroreceptors in
20 modulating blood pressure35,36. These findings fill an important gap
2–1
2–2 of how the central nervous system monitors and orchestrates adipose
f mCherry
Ucp1 Elovl3 Cidea
functions, highlighting the importance of an underappreciated branch
-flex-DTA
DRG
g h i of brain–body communication.
0.7736 0.5265 0.0207
100 700 38.0 It remains to be determined how sensory pathways mechanistically
(beats per min)

temperature (ºC)

iWAT 37.5
interact with sympathetic signalling. It has been suggested that cap-
at 30 ºC (%)
Time spent

Heart rate

Cre Cre 600

37.0
saicin denervation of iWAT altered sympathetic output in the distal
Rectal

22 ºC 30 ºC 80
Post-surgery Thermo- 36.5
recovery neutrality
500 iBAT through central circuits8,37; however, this observation might be
36.0
60 400 35.5
confounded by the likely increase in sympathetic tone38 due to periph-
1 week 2–3 week
Cre– Cre+ Cre– Cre+ Cre– Cre+ eral capsaicin administration. By contrast, with projection-specific
Fig. 4 | Ablation of iWAT DRGs changes the morphological and ablation, our adipose phenotypes were strictly limited in the ipsilateral
physiological properties of the iWAT. a,b, Representative images (a) and iWAT but not in the distal iBAT or eWAT (Figs. 3 and 4 and Extended Data
quantification of fat mass (b) of iWAT with Cre-dependent unilateral sensory Figs. 5 and 6), suggesting a local specificity of the sensory–sympathetic
ablation. n = 11 mice. Statistical analysis was performed using two-tailed paired interaction (Fig. 3f,g). Sensory modulation of sympathetic activity
t-tests. c, Histology of iWAT with Cre-dependent unilateral sensory ablation. has been documented in other organs39,40 at the spinal or supraspinal
d, Histology of iWAT with Cre-dependent bilateral sensory ablation. Each panel levels41–43. We did not observe a significant change in total noradrena-
is from a different mouse. e, RT–qPCR analysis of iWAT with Cre-dependent line content in sensory-ablated fat (Extended Data Fig. 8i), suggesting
bilateral sensory ablation. n = 4 mice per group. Statistical analysis was that sensory activity may act on sympathetic signalling downstream
performed using two-tailed unpaired t-tests. f–i, Physiological measurements of the adrenergic receptors, although we cannot rule out potential
after Cre-dependent bilateral sensory ablation. f, Schematic of bilateral temporospatial noradrenaline changes, which cannot be resolved by
sensory ablation, and the timeline of physiological measurement.
a single, bulk noradrenaline measurement. Moreover, remodelling
g, Two-temperature choice assay (30 °C versus 18 °C) of Cre− (n = 6) and
of sympathetic fibres is a hallmark of adipose adaptation to chronic
Cre+ (n = 6) mice. h, Heart rate of Cre− (n = 6) and Cre+ (n = 6) mice. i, Rectal
metabolic challenges, and it remains to be tested whether sensory
temperature at thermoneutrality of Cre− (n = 9) and Cre+ (n = 10) mice. For
g–i, data are mean ± s.e.m. Statistical analysis was performed using two-tailed
innervation undergoes a similar process, especially in different strains
unpaired t-tests with Welch’s correction (g–i). For c and d, scale bars, 100 μm. of mice as well as across species.
Our study raises many questions. The somatosensory nervous
system, which is characterized by clusters of first-order neurons
activity, but may also have a role in coordinating these two opposing that mainly reside in the DRG, has great molecular heterogeneity as
downstream pathways. profiled by recent single-cell transcriptomes10,44. Which DRG sub-
Unilateral ablation enabled accurate assessment of adipose phe- types innervate fat and whether distinct subtypes have different
notypes within the same animal. To determine how adipose sensory functions (that is, regulating thermogenesis versus lipogenesis)
innervation affects whole-body physiology, we further ablated sen- is currently unclear. This could potentially be determined in the
sory innervation in the iWAT bilaterally. Consistent with the unilateral future by coupling ROOT-based retrograde targeting with single-cell
ablation, we observed an enrichment of beige adipocytes (Fig. 4d) RNA-seq to establish the identities the fat-innervating neurons. This
and upregulated lipogenic and thermogenic genes in iWAT in the information could also help to identify the interoceptive signal that
ablated animals (Fig. 4e and Extended Data Fig. 8a,b). These animals fat-innervating neurons sense. Previous research showed that infu-
were housed at murine thermoneutrality (30 °C)30 (Fig. 4f) to elimi- sion of exogeneous leptin and free fatty acids could activate DRGs
nate confounding effects from other thermoregulatory mechanisms by FOS staining or ex vivo recording45,46; however, the identification
(such as background brown fat activity). Fat sensory ablation did not of endogenous signals (chemical or physical) will probably require
lead to significant changes in body weight, food intake, temperature a full understanding of the putative receptor expression using the
sensitivity or systemic sympathetic tone (Fig. 4g,h and Extended Data organ-targeted single-cell RNA-seq approach suggested above.
Fig. 8c–i), but exhibited elevated core body temperature compared Moreover, the nature of neural transmission also suggests that
with the controls (Fig. 4i), consistent with an enhanced thermogenesis these endogenous activities probably occur at second or millisec-
program in the iWAT. Interestingly, the difference in body temperature ond scales. Thus, a matching ability to readout projection-specific
was normalized at 22 °C (Extended Data Fig. 8c), suggesting that the DRG activities in vivo would be required to identify the triggering
elevated body temperature at thermoneutrality was not due to a deficit signals, for example, using emerging long-term DRG calcium imag-
in central thermoregulatory mechanisms (pyrexia)31,32. After challenge ing techniques47,48.

Nature | Vol 609 | 15 September 2022 | 573

Article
Our research also has implications beyond the sensory innervation 25. Yu, X. X., Lewin, D. A., Forrest, W. & Adams, S. H. Cold elicits the simultaneous induction
of fatty acid synthesis and β‐oxidation in murine brown adipose tissue: prediction from
of fat. Interoception has been mostly studied from the perspective differential gene expression and confirmation in vivo. FASEB J. 16, 155–168 (2002).
of cranial ganglia of vagal origin49. Increasing evidence indicates 26. Mottillo, E. P. et al. Coupling of lipolysis and de novo lipogenesis in brown, beige, and
the DRG innervation of internal organs also has a crucial role in white adipose tissues during chronic β3-adrenergic receptor activation. J. Lipid Res. 55,
2276–2286 (2014).
interoception 50. The discoveries made here could serve as a 27. Herman, M. A. et al. A novel ChREBP isoform in adipose tissue regulates systemic glucose
proof-of-principle for investigating the role of DRG neurons in a metabolism. Nature 484, 333–338 (2012).
variety of other internal organs. 28. Eissing, L. et al. De novo lipogenesis in human fat and liver is linked to ChREBP-β and
metabolic health. Nat. Commun. 4, 1528 (2013).
29. Yeh, W. J., Leahy, P. & Freake, H. C. Regulation of brown adipose tissue lipogenesis by
thyroid hormone and the sympathetic nervous system. Am. J. Physiol. 265, E252–E258
Online content (1993).
30. Cannon, B. & Nedergaard, J. Nonshivering thermogenesis and its adequate measurement
Any methods, additional references, Nature Research reporting sum- in metabolic studies. J. Exp. Biol. 214, 242–253 (2010).
maries, source data, extended data, supplementary information, 31. Gordon, C. J. A review of terms for regulated vs. forced, neurochemical-induced changes
acknowledgements, peer review information; details of author con- in body temperature. Life Sci. 32, 1285–1295 (1983).
32. Dittner, C., Lindsund, E., Cannon, B. & Nedergaard, J. At thermoneutrality, acute
tributions and competing interests; and statements of data and code thyroxine-induced thermogenesis and pyrexia are independent of UCP1. Mol. Metab. 25,
availability are available at https://doi.org/10.1038/s41586-022-05137-7. 20–34 (2019).
33. Seale, P. et al. Prdm16 determines the thermogenic program of subcutaneous white
adipose tissue in mice. J. Clin. Invest. 121, 96–105 (2011).
1. Guilherme, A., Henriques, F., Bedard, A. H. & Czech, M. P. Molecular pathways linking 34. Ikeda, K. et al. UCP1-independent signaling involving SERCA2b-mediated calcium cycling
adipose innervation to insulin action in obesity and diabetes mellitus. Nat. Rev. regulates beige fat thermogenesis and systemic glucose homeostasis. Nat. Med. 23,
Endocrinol. 15, 207–225 (2019). 1454–1465 (2017).
2. Bartness, T. J., Liu, Y., Shrestha, Y. B. & Ryu, V. Neural innervation of white adipose tissue 35. Wehrwein, E. A. & Joyner, M. J. Regulation of blood pressure by the arterial baroreflex and
and the control of lipolysis. Front. Neuroendocrinol. 35, 473–493 (2014). autonomic nervous system. Handb. Clin. Neurol. 117, 89–102 (2013).
3. Bachman, E. S. et al. βAR signaling required for diet-induced thermogenesis and obesity 36. Zeng, W.-Z. et al. PIEZOs mediate neuronal sensing of blood pressure and the
resistance. Science 297, 843–845 (2002). baroreceptor reflex. Science 362, 464–467 (2018).
4. Chi, J. et al. Three-dimensional adipose tissue imaging reveals regional variation in beige 37. Nguyen, N. L. T., Xue, B. & Bartness, T. J. Sensory denervation of inguinal white fat
fat biogenesis and PRDM16-dependent sympathetic neurite density. Cell Metab. 27, modifies sympathetic outflow to white and brown fat in Siberian hamsters. Physiol. Behav.
226–236 (2018). 190, 28–33 (2018).
5. Cohen, P. & Kajimura, S. The cellular and functional complexity of thermogenic fat. Nat. 38. Ludy, M.-J., Moore, G. E. & Mattes, R. D. The effects of capsaicin and capsiate on energy
Rev. Mol. Cell Biol. 22, 393–409 (2021). balance: critical review and meta-analyses of studies in humans. Chem. Senses 37,
6. Fishman, R. B. & Dark, J. Sensory innervation of white adipose tissue. Am. J. Physiol. 253, 103–121 (2012).
R942–R944 (1987). 39. Longhurst, J. C., Tjen‐A‐Looi, S. C. & Fu, L. Cardiac sympathetic afferent activation
7. Jiang, H., Ding, X., Cao, Y., Wang, H. & Zeng, W. Dense intra-adipose sympathetic provoked by myocardial ischemia and reperfusion. Ann. N. Y. Acad. Sci. 940, 74–95
arborizations are essential for cold-induced beiging of mouse white adipose tissue. Cell (2001).
Metab. 26, 686–692 (2017). 40. Morelli, C. et al. Identification of a population of peripheral sensory neurons that
8. Song, C. K., Schwartz, G. J. & Bartness, T. J. Anterograde transneuronal viral tract tracing regulates blood pressure. Cell Rep. 35, 109191 (2021).
reveals central sensory circuits from white adipose tissue. Am. J. Physiol. 296, R501–R511 41. Jänig, W. Central organization of somatosympathetic reflexes in vasoconstrictor
(2009). neurones. Brain Res. 87, 305–312 (1975).
9. Foster, M. T. & Bartness, T. J. Sympathetic but not sensory denervation stimulates white 42. Liu, S. et al. Somatotopic organization and intensity dependence in driving distinct
adipocyte proliferation. Am. J. Physiol. 291, R1630–R1637 (2006). NPY-expressing sympathetic pathways by electroacupuncture. Neuron 108, 436–450
10. Sharma, N. et al. The emergence of transcriptional identity in somatosensory neurons. (2020).
Nature 577, 392–398 (2020). 43. Liu, S. et al. A neuroanatomical basis for electroacupuncture to drive the vagal–adrenal
11. Baboota, R. K. et al. Capsaicin induces “brite” phenotype in differentiating 3T3-L1 axis. Nature 598, 641–645 (2021).
preadipocytes. PLoS ONE 9, e103093 (2014). 44. Zeisel, A. et al. Molecular architecture of the mouse nervous system. Cell 174, 999–1014
12. Chen, J. et al. Activation of TRPV1 channel by dietary capsaicin improves visceral fat (2018).
remodeling through connexin43-mediated Ca2+ influx. Cardiovasc. Diabetol. 14, 22 45. Murphy, K. T. et al. Leptin-sensitive sensory nerves innervate white fat. Am. J. Physiol. 304,
(2015). E1338–E1347 (2013).
13. Baskaran, P., Krishnan, V., Ren, J. & Thyagarajan, B. Capsaicin induces browning of white 46. Garretson, J. T. et al. Lipolysis sensation by white fat afferent nerves triggers brown fat
adipose tissue and counters obesity by activating TRPV1 channel‐dependent thermogenesis. Mol. Metab. 5, 626–634 (2016).
mechanisms. Brit. J. Pharmacol. 173, 2369–2389 (2016). 47. Chen, C. et al. Long-term imaging of dorsal root ganglia in awake behaving mice. Nat.
14. Caterina, M. J., Rosen, T. A., Tominaga, M., Brake, A. J. & Julius, D. A capsaicin-receptor Commun. 10, 3087 (2019).
homologue with a high threshold for noxious heat. Nature 398, 436–441 (1999). 48. Ju, F. et al. Long-term two-photon imaging of spinal cord in freely behaving mice. Preprint
15. Julius, D. TRP channels and pain. Annu. Rev. Cell Dev. Biol. 29, 355–384 (2013). at bioRxiv https://doi.org/10.1101/2022.01.09.475306 (2022).
16. Li, L. et al. The functional organization of cutaneous low-threshold mechanosensory 49. Prescott, S. L. & Liberles, S. D. Internal senses of the vagus nerve. Neuron 110, 579–599
neurons. Cell 147, 1615–1627 (2011). (2022).
17. Nudell, V. et al. HYBRiD: hydrogel-reinforced DISCO for clearing mammalian bodies. 50. Marshall, K. L. et al. PIEZO2 in sensory neurons and urothelial cells coordinates urination.
Nat. Methods 19, 479–485 (2022). Nature 588, 290–295 (2020).
18. Hunter, D. V. et al. Advillin Is expressed in all adult neural crest-derived neurons. eNeuro
5, ENEURO.0077-18.2018 (2018). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
19. Wu, J. et al. Beige adipocytes are a distinct type of thermogenic fat cell in mouse and published maps and institutional affiliations.
human. Cell 150, 366–376 (2012).
20. Huesing, C. et al. Sympathetic innervation of inguinal white adipose tissue in the mouse. Open Access This article is licensed under a Creative Commons Attribution
J. Comp. Neurol. 529, 1465–1485 (2021). 4.0 International License, which permits use, sharing, adaptation, distribution
21. Giordano, A. et al. White adipose tissue lacks significant vagal innervation and and reproduction in any medium or format, as long as you give appropriate
immunohistochemical evidence of parasympathetic innervation. Am. J. Physiol. 291, credit to the original author(s) and the source, provide a link to the Creative Commons license,
R1243–R1255 (2006). and indicate if changes were made. The images or other third party material in this article are
22. Chan, K. Y. et al. Engineered AAVs for efficient noninvasive gene delivery to the central included in the article’s Creative Commons license, unless indicated otherwise in a credit line
and peripheral nervous systems. Nat. Neurosci. 20, 1172–1179 (2017). to the material. If material is not included in the article’s Creative Commons license and your
23. Wu, Z., Autry, A. E., Bergan, J. F., Watabe-Uchida, M. & Dulac, C. G. Galanin neurons in the intended use is not permitted by statutory regulation or exceeds the permitted use, you will
medial preoptic area govern parental behavior. Nature 509, 325–330 (2014). need to obtain permission directly from the copyright holder. To view a copy of this license,
24. Guilherme, A. et al. Control of adipocyte thermogenesis and lipogenesis through visit http://creativecommons.org/licenses/by/4.0/.
γ3-adrenergic and thyroid hormone signal integration. Cell Rep. 31, 107598–107598
(2020). © The Author(s) 2022

574 | Nature | Vol 609 | 15 September 2022

Methods removed and sterilized using ethanol and iodine. After surgery, the mice
were given a subcutaneous injection of flunixin and topical antibiotic
Animals ointment for post-operative care.
Mice were group housed in standard housing under a 12–12 h light–dark
cycle with ad libitum access to chow diet and water, with the room Retrograde tracer labelling of sensory neurons. For injection into
temperature kept around 22 °C and humidity kept between 30–80% the iWAT, a lateral incision was made on the flank skin of each side.
(not controlled). Mice were single housed for food intake measurement For injection into the eWAT, a lateral incision was made on the lower
and held at 30 °C for thermoneutral exposure experiments. Mice (aged abdominal wall. For injection into the iBAT, a midline incision was made
at least 6 weeks) from the following strains were used for this study: in the interscapular region. For injection into the skin, an intradermal
wild-type (WT) C57BL/6J ( Jackson stock, 000664), B6.Cg-Gt(ROSA)2 injection was performed. A Hamilton syringe with a 31G (point type 2)
6Sortm9(CAG-tdTomato)Hze/J ( Jackson, 007909, Ai9), Pirt-Cre51, Scn10a-Cre52. needle was used for all retrograde tracing. A total of 4–5 μl 0.1% CTB-
Both genders were used for anatomical mapping studies, and male 488 or CTB-647 (Invitrogen) in PBS was injected per fat pad (2 μl for
mice were used for in vivo functional experiments. All of the experi- iBAT) or the flank skin with 8–15 injections to spread out the tracer. The
mental protocols were approved by The Scripps Research Institute abdominal wall (for eWAT injection) and skin (for iWAT, eWAT or iBAT
Institutional Animal Care and Use Committee and were in accordance injection) were sutured separately. Tissues were taken 3–5 days after
with the guidelines from the NIH. injection to allow the dye to reach the DRG soma.

Viruses Retrograde viral labelling of sensory neurons. Injections were per-

CAV2 was obtained from CNRS Vector Core. rAAVretro-Cre was formed as described above. For the viral comparison study (Extend-
obtained from Boston Children’s Hospital and Janelia. PHP.S-Cre, ed Data Fig. 3a), CAV2-Cre (4.2 × 1012 physical particles per ml, 5 μl),
PHP.S-DIO-sfGFP (gift from D. Gibbs) was obtained from Janelia. PHP.S-Cre (1.6 × 1013 VGs per ml, 3 μl), rAAVretro-Cre (9 × 1012 VGs per
PHP.S-TdTomato (Addgene, 59462), PHP.S-mScarlet (gift from the Deis- ml, 5 μl) was mixed with 0.01% FastGreen (Sigma-Aldrich) and unilater-
seroth laboratory and cloned in-house), ROOT-Cre (Addgene, 51904), ally injected into iWAT of Ai9 mice. AAV9-TdTomato (2 × 1013 VGs per ml,
ROOT-EYFP (cloned in-house), PHP.S-mCherry-flex-DTA (Addgene, 2 μl) was mixed with 0.01% FastGreen (Sigma-Aldrich) and unilaterally
58536), ROOT-mScarlet, AAV9-mScarlet and PHP.S-mScarlet were pack- injected into iWAT of WT mice.
aged in-house using published protocol53. AAVs produced in-house
were titrated by qPCR before aliquoting into 6–10 μl and flash-frozen Characterization of ROOT. Injections were performed as described
for long-term storage. above. For ROOT characterization, AAV9 or ROOT-mScarlet (4 × 1013 VGs
per ml, 2 μl) in PBS with 0.001% F-68 and 0.01% FastGreen was adminis-
In vivo selection of ROOT tered unilaterally into iWAT of WT mice. Two weeks after the first sur-
Plasmids. The plasmids used for in vivo selection were adapt- gery, 4 μl of 0.1% CTB-647 was injected into the same iWAT fat pad, and
ed from the previous publication 54. rAAV-pUBC-sfGFP-Cap and tissues were taken 3–5 days after the second injection for quantification.
AAV2/9-REP-AAP was generated from pUBC-mCherry-rAB (Addgene,
115239), pUCmini-iCAP-PHP.S (Addgene, 103006), pAAV2/8 (Ad- 6-OHDA treatment. 6-OHDA has been used previously for selective
dgene, 112864). No in-cis-Lox module or transgenic Cre lines were sympathetic denervation55,56. 6-OHDA (Tocris) (12 mg ml−1 in saline with
used given the anatomical separation of DRGs and iWAT. The ROOT 0.02% ascorbic acid (Sigma-Aldrich)) was prepared fresh before use
capsid library was generated by inserting random heptamers using and kept on ice and in the absence of light. 6-OHDA (8 μl) was injected
NNK degenerate primers (Integrated DNA technologies) between into each iWAT fat pad. Saline with 0.02% ascorbic acid was used as a
the 588 and 589 sites of AAV9 by Gibson assembly as previously de- control. Tissues were taken 7–9 days after injection.
scribed54.
Intraganglionic DRG injection. Intraganglionic DRG injection was
AAV capsid library production. The viral libraries were produced as performed according to a previous report57. A midline incision was
previously described53,54. In brief, only 10 ng of rAAV-pUBC-sfGFP-Cap made on the dorsal skin to expose the dorsal muscles. The muscles
library DNA was transfected in HEK293FT (Invitrogen R70007) cells per along the vertebra were carefully separated to expose DRGs. DRGs at
150 mm plate, and the virus was collected after 60 h for purification. the T13 and L1 vertebral level were exposed, and AAV (~1 × 1013 VGs per
ml in PBS with 0.001% F-68 and 0.01% FastGreen) was injected in the
DNA recovery and sequencing. The resulting AAV capsid library was ganglion (~200 nl per ganglion) with a pulled glass pipette using the
injected bilaterally into the iWAT of C57Bl/6J male mice at 109 viral ge- Nanoliter 2020 Injector (World Precision Instrument). Care was taken
nomes (VGs) per fat pad. rAAV genomes were recovered from T12–L3 to avoid damaging the surrounding vasculatures. Dorsal muscle and
DRGs from injected mice two weeks after injection using the DNeasy skin were sutured separately. PHP.S-TdTomato or PHP.S-mScarlet was
Blood and Tissue kit (Qiagen). The recovered viral DNA was amplified used for anterograde labelling.
for two rounds against a fragment containing the heptamer insertion
with Q5 high-fidelity polymerase (New England Biolabs). The ampli- Cre-dependent ablation of iWAT-innervating DRGs. To selectively
fied products were cleaned up and processed for complete amplicon ablate iWAT-innervating DRGs, ROOT-Cre or ROOT-YFP (4 × 1013 VGs
sequencing at Massachusetts General Hospital. per ml, 2 μl) in PBS with 0.001% F-68 and 0.01% FastGreen was inject-
ed into iWAT unilaterally or bilaterally, and PHP.S-mCherry-flex-DTA
NGS data alignment and processing. Raw FASTQ files from NGS runs (1 × 1013 VGs per ml, 200 nl per ganglion) was injected into T13 and L1
were aligned to an AAV9-template DNA fragment containing the 21 bp DRGs bilaterally as described above. Tissues were extracted, or physi-
diversified region between amino acids 588 and 589 using SAMtools ological measurements were performed 3–4 weeks after surgery.
(v.1.10). The abundance of each 21 bp sequence in all of the recovered
sequences with the heptamer insertion was quantified. 6-OHDA treatment and Cre-dependent sensory ablation. AAVs were
injected into iWAT and DRGs to achieve Cre-dependent sensory ablation
Surgery as described above. Two to three weeks after the first injection, 6-OHDA
Mice were anaesthetized with isoflurane (4% for induction and 1.5–2% or saline was injected into the iWAT again. Tissues were extracted 7–9
for maintenance), with their skin at the surgical area shaved and hair days after the second injection.
Article
blocked with 5% normal donkey serum in PBS with 0.3% Triton X-100.
En bloc HYBRiD tissue clearing and immunolabelling of iWAT Primary antibodies were prepared in the same blocking solution and
To visualize sensory nerves in mouse torso and iWAT samples, tissue incubated overnight (anti-βIII-tubulin (Abcam, ab18207, 1:1,000). The
samples were cleared using the HYBRiD method as described previ- next day, the sections were washed in PBS and then incubated for 2 h at
ously17. room temperature with secondary antibodies (anti-rabbit-647, Jackson
Immuno Research, 711-606-152, 1:500), and stained with DAPI before
Sample collection and pretreatment. Mice were terminally anaesthe- mounted with ProLong Gold antifade mountant (Invitrogen) for confo-
tized with isoflurane and intracardially perfused with ice-cold PBS and cal microscopy imaging.
ice-cold 4% PFA in PBS with 4% sucrose (Electron Microscopy Perfusion
Fixative, 1224SK). For torso samples, the skin was carefully removed Immunolabelling of DRG sections. T12 to L2 DRGs from mice with
leaving the iWAT attached to the muscle, the spinal cord was cut from CTB-647 injected in iWAT were dissected, post-fixed in PFA, dehydrated
the midline to facilitate clearing and imaging. All of the collected sam- in 30% sucrose before being embedded in OCT, then sectioned at 20 μm.
ples were post-fixed in 4% PFA at 4 °C for 1–2 days before being washed DRG sections were stained according to the procedure described above
in PBS. The torso samples were decalcified in 10% EDTA/15% imida- using the following primary antibodies: anti-CGRP (Immunostar, 24112,
zole (at 4 °C for 7 days), then decolourized in 25% N,N,N′,N′-tetraki 1:1,000), anti-neurofilament heavy polypeptide (Abcam, ab4680,
s(2-hydroxypropyl)ethylenediamine (Quadrol) (in 1× PBS) (at 37 °C 1:1,000). Secondary antibodies and dyes included anti-rabbit-594 ( Jack-
for 4 days). son Immuno Research, 711-586-152, 1:500), anti-chicken-488 ( Jackson
Immuno Research, 703-546-155, 1:500) and isolectin B4 AlexaFluor 488
Delipidation and hydrogel embedding. The torso and iWAT samples (Life Technologies, I21411, 25 µg ml−1).
were sequentially washed in 50%, 70%, 80%, 95% and 95% tetrahydro-
furan in 25% Quadrol (in 1× PBS), 100% dichloromethane (DCM), 100% Haematoxylin and eosin staining of adipose tissues. Mice were
DCM, 100% DCM, 95%, 95%, 80%, 70% and 50% tetrahydrofuran in 25% terminally anaesthetized with isoflurane. iWAT, eWAT and iBAT were
Quadrol followed by 1× PBS to wash out any remaining organic solvent. extracted and post-fixed in 4% PFA overnight at 4 °C. The tissues were
The samples were then incubated in A1P4 hydrogel (1% acrylamide, paraffin-embedded, sectioned at 5 μm and mounted onto glass slides.
0.125% Bis, 4% PFA, 0.025% VA-044 initiator (w/v), in 1× PBS) at 4 °C be- The sections were then stained with haematoxylin and eosin at San-
fore being degassed with nitrogen and polymerized (4 h at 37 °C). The ford Burnham Prebys histology core and imaged using a bright-field
samples were then removed from the hydrogel and passively cleared microscope.
with 20 mM LiOH-Boric buffer pH 8.0 containing 6% SDS at 37 °C until
the samples appeared translucent. After clearing, the samples were Imaging
washed in PBST (0.2% Triton X-100) thoroughly before proceeding to Confocal microscopy. Mounted iWAT samples, whole-mount gan-
refractive-index matching or immunolabelling. glia, stained skin sections were imaged with Olympus FV3000 con-
focal microscope using one of the following objectives: ×4/0.28 NA,
Immunolabelling. iWAT samples were incubated with primary antibod- air (XLFluor, Olympus); ×10/0.6 NA, water immersion (XLUMPlanFI,
ies diluted in PBST at room temperature. After incubation with primary Olympus). Images were acquired using Fluoview (v.2.4.1.198).
antibodies, the samples were washed in PBST and then incubated in
secondary antibodies diluted in PBST at room temperature. The sam- Light-sheet microscopy. Torso or iWAT samples were mounted us-
ples were washed extensively in PBST before refractive-index matching ing 1% agarose/EasyIndex. Mounted samples were imaged inside the
and mounting. The following antibodies were used: anti-perilipin-1 SmartSPIM chamber filled with EasyIndex and sealed with mineral oil on
(Cell Signaling, 9349, 1:400), anti-TH-647 (BioLegend, 818008, 1:300); the top. Images were acquired using a ×3.6/0.2 NA objective (LifeCan-
anti-Ucp1 (Abcam, ab10983, 1:200); anti-rabbit-488 ( Jackson Immuno vas), with a 1.79 μm, 1.79 μm, 4 μm xyz voxel size. Image acquisition
Research 711-546-152, 1;400); anti-rabbit-647 ( Jackson Immuno Re- was completed with bilateral illumination along the central plane of
search 711-606-152, 1:400). symmetry within the sample.

Refractive-index matching and mounting. Cleared or immunola- Fluorescence stereomicroscopy. Freshly fixed liver samples were
belled samples were refractive-index matched in EasyIndex (RI = 1.52, imaged using the Leica M165 FC stereomicroscope and FLIR BFS-U3-
Life Canvas) and mounted in spacers (Sunjin Lab) for confocal microsco- 51S51M-C camera. Images were acquired using SpinView (v.2.5.0.80).
py imaging or mounted in agarose for light-sheet microscopy imaging.
Bright-field microscopy. Slides stained with haematoxylin and eosin
Histological analysis of whole-mount samples and cryosections were imaged using the Keyence BZ-X710 microscope with a ×40/0.6 NA
Mice were terminally anaesthetized with isoflurane and intracardially objective (CFI S Plan Fluor ELWD ADM, Nikon).
perfused with PBS and 4% PFA. For flank skin samples, skin was shaved,
and hair was removed using Nair. Transcriptional analysis
RNA preparation and RT–qPCR analysis. Adipose tissues were dis-
Whole-mount imaging of ganglia and flank skin. Ganglia of interest sected between 12:00 and 14:00 and flash-frozen in liquid nitrogen.
(DRGs, SChGs and celiac/mesenteric complex) were dissected and Total RNA was extracted from frozen tissue using TRIzol (Invitrogen)
mounted in RapiClear (Sunjin Lab) with silicone spacer (Electron Mi- and RNeasy Mini kits (Qiagen). For RT–qPCR analysis, total RNA was
croscopy) for confocal imaging. Flank skin was dissected, post-fixed in reverse-transcribed using the Maxima H Minus First Strand cDNA Syn-
PFA and washed three times in PBS before mounting in Fluoromount-G thesis Kit (Thermo Fisher Scientific). The resultant cDNA was mixed
(Invitrogen 00-4958-02) using 0.25 mm iSpacers (Sunjin Lab) for con- with primers (Integrated DNA Technology) and SyGreen Blue Mix
focal imaging. (Genesee Scientific, 17-507) for RT–qPCR using the CFX384 real-time
PCR system (BioRad). Normalized mRNA expression was calculated
Immunolabelling analysis of skin sections. Flank skin was dissected, using ΔΔCt method, using Tbp (encoding TATA-box-binding protein)
post-fixed in PFA, dehydrated in 30% sucrose before being embed- mRNA as the reference gene. Statistical analysis was performed on
ded in OCT, then sectioned at 25 μm and mounted on gelatin-coated ΔΔCt. The primer sequences (forward and reverse sequence, 5′ to 3′,
slides. For immunofluorescence analysis, skin tissue sections were respectively) were as follows: Tbp (CCTTGTACCCTTCACCAATGAC and
ACAGCCAAGATTCACGGTAGA); Ucp1 (AGGCTTCCAGTACCATTAGGT water with 0.1% formic acid; buffer B, acetonitrile with 0.1% formic acid.
and CTGAGTGAGGCAAAGCTGATTT); Elovl3 (TTCTCACGCGGGT- The LC gradient started at 5% B from 0 to 2 min. The gradient was then
TAAAAATGG and GAGCAACAGATAGACGACCAC); Cidea (ATCACAACTG- increased linearly to 5% A/95% B from 2 to 23 min. From 23 to 28 min,
GCCTGGTTACG and TACTACCCGGTGTCCATTTCT); Fasn (GGAGGTG- the gradient was maintained at 5% A/95% B; and from 28 to 29 min, the
GTGATAGCCGGTAT and TGGGTAATCCATAGAGCCCAG); Acacb (CGCT- gradient went back to the starting concentration of 5% B. The flow rate
CACCAACAGTAAGGTGG and GCTTGGCAGGGAGTTCCTC); ChREBPβ was maintained at 0.7 ml min−1 throughout the run. Multiple reaction
(TCTGCAGATCGCGTGGAG and CTTGTCCCGGCATAGCAAC). monitoring was performed for noradrenaline, looking at the transition
of m/z = 170.1 as the precursor ion to the m/z = 152 fragment. The dwell
RNA library preparation and sequencing. RNA library preparation time was 200, the fragmentor set at 60, the collision energy set at 4 and
and sequencing was performed at Scripps Genomics core. Total RNA the cell accelerator voltage set at 4.
samples were prepared into RNA-seq libraries using the NEBNext
Ultra Directional RNA Library Prep Kit for Illumina according to the Metabolic studies
manufacturer’s recommended protocol. In brief, 1 μg total RNA was Mice received bilateral sensory ablation were fed a high-fat diet (HFD)
poly(A)-selected for each sample, converted to double-stranded cDNA (D12492, Research Diets) under thermoneutrality at 30 °C. HFD feeding
followed by fragmentation and ligation of sequencing adapters. The was started at 1 month after surgery (aged around 11–12 weeks old).
library was then PCR-amplified for 8 cycles using barcoded PCR prim- Body weight was measured every week. Fasting glucose was taken and
ers, purified and size-selected using AMPure XP Beads before loading glucose tolerance tests were performed at 9 weeks on HFD, and plasma
onto an Illumina NextSeq 2000 for 100 bp single-read sequencing. insulin was measured at 13 weeks on HFD.

RNA-seq analysis. Sequenced reads were aligned to the GRCm39 Glucose tolerance test. Mice were fasted for 4 h (08:30–12:30) before
reference genome (Ensembl, v.104; http://uswest.ensembl.org/Mus_ intraperitoneal injection of glucose (1 g kg−1). Blood glucose levels
musculus/Info/Index), and gene counts were quantified using Salmon were measured at the indicated time point using the OneTouch Ultra
(v.1.5.1)59. Differential gene expression analysis and P-value calculation 2 blood glucose meter.
were performed by DESeq2 (v.1.32.0)58. Gene Ontology enrichment
analysis was performed using Metascape59 with the Gene Prioritization Plasma insulin measurement. Fasting blood samples were
by Evidence Counting setting. retro-orbitally obtained 3 h into the light cycle after fasting for 14 h.
Plasma was separated in heparin-treated tubes (Microvette CB 300LH).
Behavioural and physiological assays Plasma insulin was measured with an ELISA kit (Crystal Chem, 90080)
Mechanical threshold. Mice were acclimated for 1 h in von Frey cham- and read using the BioTek Cytation 5 Imaging Reader.
bers. The 50% mechanical threshold was measured with calibrated von
Frey filaments (0.07, 0.16, 0.4, 0.6, 1.0, 1.4, 2.0, 4.0 and 6.0 g) using the Western immunoblotting
up–down method60. Whole-tissue protein lysate was extracted in RIPA buffer (G-Biosciences)
containing Halt proteinase inhibitor (Thermo Fisher Scientific, 78430)
Two-temperature choice assay. The two-temperature choice test ap- and phosphatase inhibitor (Thermo Fisher Scientific, 78420). Protein
paratus was set up as previously described61. Lanes were evenly divided lysate was determined using the bicinchoninic acid assay (Pierce).
between two different temperature plates set at 30 °C and 18 °C, and Protein lysates were denatured in Laemmli buffer (BioRad), resolved
individual mice were placed in one of the lanes. The mice were given by 4–12% Mini-PROTEAN TGX SDS–PAGE (BioRad) and transferred to
10 min to acclimatize and were then tracked for 1 h using the EthoVi- a polyvinylidene difluoride membrane. The membrane was incubated
sion tracking system (Noldus Information Technology) in a dark room with primary antibodies diluted in EveryBlot blocking buffer (Bio-
with infrared lighting. The total time spent in each temperature zone Rad) overnight at 4 °C and then incubated with secondary antibody
was analysed. anti-rabbit HRP ( Jackson Immuno Research, 711-036-152, 1:10,000) or
anti-mouse HRP ( Jackson Immuno Research, 715-036-150, 1:10,000)
Core body temperature measurement. Core temperature was meas- diluted in EveryBlot at room temperature. The results were visual-
ured using a thermocouple rectal probe (World Precision Instruments). ized using SuperSignal West Pico PLUS Chemiluminescent Substrate
Room temperature core body temperature was taken when the mice (Invitrogen). The following antibodies were used for immunoblot-
in thermoneutral condition moved to room temperature for more ting: anti-p-HSL(Ser660) (Cell Signaling, 45804, 1:1,000), anti-HSL
than 24 h. (Cell Signaling, 4107, 1:1,000), anti-α-tubulin (Abcam, 7291, 1:10,000).
Specifically, HSL was blotted after stripping the p-HSL membrane.
Blood pressure and heart rate measurement. Blood pressure and
heart rate were measured using the tail-cuff method with the CODA Imaging analysis and quantification
High Throughput Non-Invasive Blood Pressure System (Kent Scientific) All of the images were analysed using ImageJ. 3D volume images were
as previously described62. rendered in Imaris.

Targeted detection of norepinephrine in bulk iWAT. Frozen iWAT tis- Quantification of TH + DRG nerves in the iWAT. Regions of
sues were lysed in 4× weight of 0.1 mol l−1 perchloric acid, centrifuged 40 μm × 40 μm × 40 μm (x,y,z) were randomly selected and maximally
and run through a 30 kDa filtration tube (Millipore). The filtrates were projected over z using customized ImageJ scripts in the whole stacks of
analysed on an Agilent 6470 Triple Quadrupole (QQQ) liquid chro- intraganglionically labelled iWAT with TH staining. Only regions con-
matography–mass spectrometry (LC–MS) system using electrospray taining positive TdTomato (DRG) signals were retained. The thickness of
ionization (ESI) in positive mode. The AJS ESI source parameters were TdTomato-positive fibres was measured using the ImageJ straight line
set as follows: the gas temperature was set at 250 °C with a gas flow of function across two different places in the view and averaged. Fibres
12 l min−1 and the nebulizer pressure at 25 p.s.i. The sheath gas tem- with widths of less than 2.5 μm (arbitrary cut-off) were considered to
perature was set to 300 °C with the sheath gas flow set at 12  l min−1. be thin fibres. If the TH-647 channel showed overlap with the TdTo-
The capillary voltage was set to 3,500 V. Separation of metabolites mato signal, the view was considered to be positive. We quantified 13
was conducted on an Agilent Eclipse Plus C18 LC column (3.5 μm, images from 3 biological replicates, and a total of 77 views containing
4.6 × 100 mm, 959961-902). Mobile phases were as follows: buffer A, the TdTomato positive thin fibres were quantified for TH positivity.
Article
follows: Figs. 1b–d (four), 1e,g,h,j (three),  2a–b,d (two),  4a (four), 4c,d
Quantification of nerve density in flank skin and iWAT. Regions of (two) and Extended Data Figs. 1a (three), 1b (four), 1c (two), 2a (six), 2f
80 μm × 80 μm × 20 μm (x,y,z) were randomly selected and maximally (three), 2h–i,k (two), 3a,d,e (three), 3b (two), 4a,h (two); 5b (two), 7d
projected over z using customized ImageJ scripts in the whole stacks (three), 7e (two).
of flank skin or iWAT from mice received intraganglionic labelling. Ar-
eas containing nerve fibres were automatically segmented using auto Reporting summary
thresholding in ImageJ. Nerve density was calculated as AreaNerve/AreaTotal. Further information on research design is available in the Nature
Only views containing nerve signals were retained for quantification. We Research Reporting Summary linked to this article.
quantified 466 views from 39 images from 3 biological replicates for flank
skin, and 468 views from 31 images from 2 biological replicates for iWAT.
Data availability
Quantification of intraepidermal nerve fibres. Quantification of in- Bulk RNA-seq data have been deposited at the Gene Expression Omni-
traepidermal nerve fibres with antibodies against βIII-tubulin was deter- bus under accession number GSE207664. All the numeric data in this
mined according to the number of nerve fibres crossing the basement study are included in the Supplementary Information. All other data
membrane in the cross-sections of flank skin. Scoring was performed supporting the findings of this study are too large for public deposit
in a blinded manner on all of the images and post hoc registered to the and are available from the corresponding authors. Source data are
condition. Ipsilateral and contralateral flank skin from 4 mice (10–15 provided with this paper.
sections per tissue) was quantified.

Quantification of ROOT/AAV9 labelling in ganglia. DRGs and SChGs Code availability

were taken from mice injected with AAV (ROOT or AAV9) and CTB for The code used for sequencing analysis and FIJI analysis is deposited at
whole-mount imaging as described above. T11–L3 DRGs and T12 SChGs GitHub (https://github.com/yelabscripps).
were quantified for cell numbers labelled by AAV and CTB. The percent-
age of AAV and CTB double-positive cells in all AAV labelled cells were 51. Kim, A. Y. et al. Pirt, a phosphoinositide-binding protein, functions as a regulatory subunit
of TRPV1. Cell 133, 475–485 (2008).
quantified in T13 and L1 DRGs as intraganglionic surgery was performed 52. Agarwal, N., Offermanns, S. & Kuner, R. Conditional gene deletion in primary nociceptive
on T13 and L1 DRGs for later in vivo experiments. neurons of trigeminal ganglia and dorsal root ganglia. Genesis 38, 122–129 (2004).
53. Challis, R. C. et al. Systemic AAV vectors for widespread and targeted gene delivery in
rodents. Nat. Protoc. 14, 379–414 (2019).
Quantification of CTB/ROOT/AAV9-labelled DRG soma sizes. 54. Deverman, B. E. et al. Cre-dependent selection yields AAV variants for widespread gene
Soma sizes were quantified manually in full stack images from DRG transfer to the adult brain. Nat. Biotechnol. 34, 204–209 (2016).
55. Thoenen, H. & Tranzer, J. P. Chemical sympathectomy by selective destruction of
whole-mount imaging using Fiji. We noticed that injecting CTB con-
adrenergic nerve endings with 6-hydroxydopamine. Naunyn Schmiedebergs Arch. Exp.
jugated to different fluorophores (that is, CTB-488 versus CTB-647) Pathol. Pharmakol. 261, 271–288 (1968).
into the iWAT may result in a slight difference in the distribution of the 56. Cao, Q., Jing, J., Cui, X., Shi, H. & Xue, B. Sympathetic nerve innervation is required for
beigeing in white fat. Physiol. Rep. 7, e14031 (2019).
soma diameter of labelled DRG neurons; we therefore exclusively used
57. Fischer, G. et al. Direct injection into the dorsal root ganglion: technical, behavioral, and
CTB-647 when quantifying CTB-labelled DRG soma sizes. histological observations. J. Neurosci. Methods 199, 43–55 (2011).
58. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion
for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
Quantification of liver fluorescence intensity. For liver fluorescence
59. Zhou, Y. et al. Metascape provides a biologist-oriented resource for the analysis of
determination in the characterization of ROOT, the same views were systems-level datasets. Nat. Commun. 10, 1523 (2019).
acquired with two different exposure times. The images with a longer 60. Chaplan, S. R., Bach, F. W., Pogrel, J. W., Chung, J. M. & Yaksh, T. L. Quantitative
assessment of tactile allodynia in the rat paw. J. Neurosci. Methods 53, 55–63 (1994).
exposure time were used to determine the shape of the tissue, and the
61. Dhaka, A. et al. TRPM8 is required for cold sensation in mice. Neuron 54, 371–378 (2007).
images with a shorter exposure time were used to quantify intensity 62. Daugherty, A., Rateri, D., Hong, L. & Balakrishnan, A. Measuring blood pressure in mice
to avoid over-saturation. The region of interest of the liver and the using volume pressure recording, a tail-cuff method. J. Vis. Exp. https://doi.
org/10.3791/1291 (2009).
background were drawn manually in ImageJ, and the liver fluorescence
intensity was defined as Mean Intensityliver − Mean Intensitybackground.
Acknowledgements We thank E. Saez, P. Cohen, C.-H. Lee, K. L. Marshall and I. Daou for their
Study design input; B. E. Deverman for advice with AAV vector engineering; K. Deisseroth, D. Gibbs and C.
Ramakrishnan for the mScarlet and sfGFP plasmids; J. Stirman and K. Spencer for the imaging
No statistical methods were used to calculate the sample size. The support; the staff at Scripps genomics core and Sanford Burnham Prebys histology core for
sample size was determined on the basis of previous studies and lit- sample preparation; A. Chesler and R. Z. Hill for their feedback on manuscript; and all
members of the Ye laboratory, Patapoutian laboratory and Dorris Neuroscience Center for their
erature in the field using similar experimental paradigms. No data
support and feedback. This work was supported by the Howard Hughes Medical Institute; NIH
were excluded, except for mice with deteriorating health issues after grants R35 NS105067 and R01AT012051 (to A.P.); NIH Director’s New Innovator Award
surgery or during the experiment and mice with missed viral target- DP2DK128800 (to L.Y.), NIDDK K01DK114165 (to L.Y.), Whitehall Foundation (to L.Y.) and Baxter
Foundation (to L.Y.); Y.W. was supported by the Dorris Scholars fellowship. Y.Z. is a Merck Fellow
ing assessed by histology. Data were collected in a blinded manner
of the Damon Runyon Cancer Research Foundation, DRG-2405-20. M.D.M.-G. was supported
and post hoc registered to the condition and analysed accordingly to by the Fundacion Alfonso Martin Escudero postdoctoral fellowship.
prevent any bias.
Author contributions Y.W., A.P. and L.Y. conceived and designed the study. Y.W., V.H.L., Y.Z.,
V.S.N., M.L. and M.R.S.-V. performed the experiments and analysed the data. D.Y. and K.W.
Statistics and reproducibility contributed to in vivo animal experiments. M.D.M.-G., V.L.L. and J.Z.L. performed noradrenaline
All non-RNA-seq analysis was performed using GraphPad Prism 9 measurements. Y.W., A.P. and L.Y. wrote the manuscript with input from all of the authors. All of
the authors provided input and reviewed the manuscript.
(v.9.3.1) with the indicated statistical tests. Paired two-tailed Student’s
t-tests were performed to compare one animal’s ipsilateral and con- Competing interests The authors declare no competing interests.
tralateral sides. The individual animal’s left or right sides were randomly
assigned for unilateral treatment. Mice were randomly assigned for Additional information
Supplementary information The online version contains supplementary material available at
bilateral treatment. Sample sizes for each experiment are reported https://doi.org/10.1038/s41586-022-05137-7.
in the figure legends. All in vivo experiments were performed at least Correspondence and requests for materials should be addressed to Ardem Patapoutian or
twice or grouped from two independent cohorts with the same con- Li Ye.
Peer review information Nature thanks Ru-Rong Ji and the other, anonymous, reviewer(s) for
clusion, except HFD treatment was performed in one cohort. For rep- their contribution to the peer review of this work. Peer reviewer reports are available.
resentative images, the numbers of experimental repetitions were as Reprints and permissions information is available at http://www.nature.com/reprints.
Extended Data Fig. 1 | Anterograde labelling maps somatosensory Autofluorescence (647 nm laser) is used to show the tissue outline. Scale bar:
innervation in adipose tissues. a, Representative images of DRG (T13), 200 μm. c, Representative images of iWAT and perigonadal WAT (pgWAT,
sympathetic chain ganglia SChG (T12) from Pirt-Cre;Ai9, Scn10a-Cre;Ai9 and refers to eWAT for male and periovarian WAT for female) from mice with
Scn10a-Cre with systemic viral labelling (PHP.S-DIO-sfGFP) (3 mice per line). intraganglionic injections of AAV expressing fluorescent protein (FP) in T13
Scale bar: 200 μm. b, Representative images of DRG (T13), SChG (T12) from and L1 DRGs (from 3 mice).
mice with unilateral intraganglionic viral injection in T13 DRG (from 3 mice).
Article

Extended Data Fig. 2 | Retrograde labelling confirms somatosensory and NG from mice with CTB-488 injected into iBAT. g, Quantification of CTB
innervation of adipose tissues. a, Representative images of DRG (T13), SChG labelled DRG soma size distribution from mice with CTB-647 injected into iWAT
(T12), nodose ganglion (NG), and celiac/superior mesenteric complex (Ce/M) (n = 3 mice). h–i, Representative images and quantification of cell type of CTB
from mice with CTB-647 injected into iWAT or eWAT. b–c, Quantification of labelled DRG neurons (from 4 mice). j, Quantification of total CTB labelled
CTB labelled cell numbers in DRG labelled from iWAT (b) or eWAT (c) along neurons from DRG (T11-L6) in male and female mice with CTB injected into
the vertebral levels. n = 4 per group. d, Schematics of dual-colour CTB iWAT or eWAT (n = 4 for male mice, and n = 3 for female mice). k, Representative
labelling from iWAT and eWAT (left) and representative images of T13 DRG. image of DRG (L1) from female mice with dual-colour CTB labelling from iWAT
e, Quantification of CTB positive cells from iWAT and eWAT. (n = 3 mice). and pgWAT (from 3 mice). All values are mean ± s.e.m. in b, c, j. Scale bar:
f, Representative images of DRG (T3), SChG (stellate ganglion and T1 SChG) 200 μm.
Extended Data Fig. 3 | Anterograde labelling reveals the morphological intraganglionically FP labelled DRG (T13 and L1) fibres in flank skin (466 views
features of somatosensory fibres in iWAT. a, 3D view of intraganglionically from 39 images from 3 biological replicates) and iWAT (468 views from 31
fluorescent protein (FP) labelled DRG (T13 and L1) fibres in close apposition to images from 2 biological replicates). All values are mean ± s.e.m. d–e,
adipocytes. b, Representative image of intraganglionically FP labelled DRG (T13 Representative images of intraganglionically FP labelled DRG (T13 and L1) fibres
and L1) fibres in flank skin. c, Quantification of relative nerve density of in adipose parenchyma (d) and travelling along the vessel (e). Scale bar: 30 μm.
Article

Extended Data Fig. 4 | Development and characterization of retrograde contralateral DRGs (T11-L3). f, Quantification of AAV+ cell numbers in
vector optimized for organ tracing (ROOT). a, Representative images of DRG ipsilateral SChG (T12). g, Quantification of bulk fluorescence intensity of liver.
(T13) from mice with AAV9, rAAV2-retro, PHP.S, or CAV2 injected into iWAT. All values are mean ± s.e.m. Statistics determined by two-tailed unpaired t test
Scale bar: 200 μm. b-c, Development of ROOT. b, Schematics of in vivo in f, g. h–i, Representative image of DRG (T13) (h) and quantification of cell
selection of retrograde vector and amino acid enrichment after one-round numbers (i) from mice with ROOT-mScarlet injected in iWAT and CTB injected
selection. c, Quantification of the abundance of recovered sequence with in flank skin. Scale bar: 200 μm. j–k, Quantification of AAV labelled DRG soma
peptide insertion. d–g, Comparison of ROOT and AAV9. d, Workflow of ROOT size distribution from mice with AAV9 ( j) or ROOT (k) injected into iWAT (n = 2
and AAV9 comparison. e, Quantification of AAV+ cell numbers in ipsilateral and mice for AAV9, n = 5 mice for ROOT).
Extended Data Fig. 5 | Characteriztaion of Cre-dependent unilateral flank skin from Cre- side (n = 48) and Cre+ side (n = 53). All values are
ablation of iWAT-DRGs. a, Quantification of CTB labelled cells in SChG (T12). mean ± s.e.m. Statistics determined by two-tailed unpaired t test. d, Average
n = 8. Statistics determined by two-tailed paired t test. b, Representative IENF density per animal. Statistics determined by two-tailed paired t test. e,
images of flank skin nerve fibres. Scale bar: 200 μm. c–d, Quantification of Mechanical threshold of hindpaws from mice with unilateral sensory ablation
intra epidermal nerve fibre (IENF) density of flank skin. n = 4 mice, 10-15 of iWAT innervation. n = 10 mice. Statistics determined by two-tailed paired t
non-continuous sections were quantified per sample. c, Pooled IENF density of test.
Article

Extended Data Fig. 6 | Gene expression analysis of Cre-dependent d, Quantitative RT-PCR analysis of iWAT after sympathetic chemical
unilateral ablation of iWAT-DRGs. a–b, Quantitative RT-PCR analysis of denervation. Statistics determined by two-tailed paired t test. e–f, Quantitative
eWAT (a) and iBAT (b) after Cre-dependent unilateral sensory ablation in iWAT. RT-PCR analysis of eWAT (e) and iBAT (f) after Cre-dependent unilateral sensory
n = 5 mice. Statistics determined by two-tailed paired t test. c–d, Chemical ablation and bilateral sympathetic chemical denervation. n = 6 mice per group.
denervation of sympathetic innervation of iWAT. c, Workflow of sympathetic Statistics determined by 2-way ANOVA.
chemical denervation. 6-OHDA was injected into iWAT unilaterally.
Extended Data Fig. 7 | Characterization of morphological changes of iWAT of iWAT, eWAT, iBAT from mice (n = 3) with unilateral sensory ablation of iWAT.
after sensory ablation. a–b, Fat mass of eWAT (a) and iBAT (b) after unilateral Scale bar: 100 μm. e, Digital slice views (200 µm) of UCP1 staining in iWAT from
sensory ablation of iWAT. n = 11 mice. Statistics determined by two-tailed mice (n = 3) with unilateral sensory ablation, showing lymph node (LN) as a
paired t test. c, Normalized weight of iWAT, eWAT and iBAT. n = 11 mice. landmark. Scale bar: 500 μm. f, Western blot of p-HSL, HSL and α-Tub in iWAT
Statistics determined by 2-way ANOVA. d, Representative histological images from mice (n = 4) with unilateral sensory ablation.
Article

Extended Data Fig. 8 | Characterization of physiological changes of iWAT mice with bilateral sensory ablation on high-fat diet (HFD) in thermoneutral
after sensory ablation. a–b, Quantitative RT-PCR analysis of eWAT (a) and temperature. j–k, Body weight (j) and body weight gain (k) of Cre- (n = 5) and
iBAT (b) from mice with bilateral sensory ablation of iWAT (n = 4 mice per Cre+ (n = 6) mice on HFD. l, Fasting glucose levels in Cre- (n = 5) and Cre+ (n = 6)
group). c–h, Physiological measurement of mice with bilateral sensory mice after 9 weeks of HFD. m, Fasting plasma insulin levels in Cre- (n = 5) and
ablation. c, Rectal temperature at room temperature of Cre- (n = 9) and Cre+ Cre+ (n = 6) mice after 14 weeks of HFD. n, IP-glucose tolerance test (1 g/kg) in
(n = 10). d, Body weight of Cre- (n = 9) and Cre+ (n = 10). e, 24 h food intake of Cre- (n = 5) and Cre+ (n = 6) mice after 9 weeks of HFD. c–h, j–n are shown as
Cre- (n = 11) and Cre+ (n = 13). f–h, Systolic (f), diastolic (g) and mean (h) blood mean ± s.e.m. Statistics determined by two-tailed unpaired t test in a–b;
pressure of Cre- (n = 6) and Cre+ (n = 6). i, Bulk iWAT  noradrenaline (NA) amount two-tailed unpaired t test with Welch’s correction in c–h, l–m; 2-way ANOVA in
in mice with unilateral sensory ablation (n = 9). j–n, Metabolic measurement of j, k, n.
 α