Article

An airway-to-brain sensory pathway

mediates influenza-induced sickness

https://doi.org/10.1038/s41586-023-05796-0 Na-Ryum Bin1, Sara L. Prescott1,3, Nao Horio1, Yandan Wang1, Isaac M. Chiu2 &
Stephen D. Liberles1 ✉
Received: 13 April 2022

Accepted: 3 February 2023

Pathogen infection causes a stereotyped state of sickness that involves neuronally
Published online: 8 March 2023
orchestrated behavioural and physiological changes1,2. On infection, immune cells
Open access
release a ‘storm’ of cytokines and other mediators, many of which are detected by
Check for updates neurons3,4; yet, the responding neural circuits and neuro–immune interaction
mechanisms that evoke sickness behaviour during naturalistic infections remain
unclear. Over-the-counter medications such as aspirin and ibuprofen are widely used
to alleviate sickness and act by blocking prostaglandin E2 (PGE2) synthesis5. A leading
model is that PGE2 crosses the blood–brain barrier and directly engages hypothalamic
neurons2. Here, using genetic tools that broadly cover a peripheral sensory neuron
atlas, we instead identified a small population of PGE2-detecting glossopharyngeal
sensory neurons (petrosal GABRA1 neurons) that are essential for influenza-induced
sickness behaviour in mice. Ablating petrosal GABRA1 neurons or targeted knockout of
PGE2 receptor 3 (EP3) in these neurons eliminates influenza-induced decreases in food
intake, water intake and mobility during early-stage infection and improves survival.
Genetically guided anatomical mapping revealed that petrosal GABRA1 neurons
project to mucosal regions of the nasopharynx with increased expression of
cyclooxygenase-2 after infection, and also display a specific axonal targeting pattern in
the brainstem. Together, these findings reveal a primary airway-to-brain sensory
pathway that detects locally produced prostaglandins and mediates systemic sickness
responses to respiratory virus infection.

Respiratory infections caused by influenza and other pathogens are neuro–immune crosstalk, although it is unclear when or whether each
leading causes of death and hospitalization worldwide6, and the recent of these pathways is engaged during naturalistic infections.
COVID-19 pandemic has broadly disrupted human society. Feeling sick Chemicals such as salicylic acid from willow bark, aspirin and ibu-
can be profoundly debilitating, and most people are sick several times profen block biosynthesis of key infection-induced lipid mediators
a year. The sensation of sickness represents a neural response to infec- through inhibition of cyclooxygenase enzymes, and have provided
tion that may provide a highly coordinated and adaptive strategy to a historically effective approach to manage sickness symptoms5,10.
promote recovery1,2. Animals infected with various pathogens display Prostaglandin E2 (PGE2) is a key cyclooxygenase-dependent metabo-
common behavioural and physiological responses that can include lite that evokes sickness behaviour, and knockout of other enzymes
fever, lethargy, loss of appetite, headache and pain, mood changes and downstream of cyclooxygenase in the PGE2 biosynthesis pathway
decreased socialization, suggesting a common sickness state involv- also ameliorates sickness responses11. PGE2 is detected by a small
ing shared neural circuits2,7. In addition to common symptoms, other subfamily of 4 G-protein-coupled receptors (EP1–EP4)12, and some
sickness responses are tailored to the site of infection; for example, but not all pyrogen-induced sickness responses are thought to be
some respiratory infections induce cough, congestion and bronchoc- mediated by the EP3 receptor2. The EP3 receptor is expressed in vari-
onstriction, whereas some gut infections induce nausea, diarrhoea and ous brain regions, including the hypothalamus and circumventricular
vomiting. Infection-specific behavioural responses suggest multiple organs as well as peripheral neurons, immune cells and many other
body–brain communication pathways for pathogen detection. cell types12,13. Region-specific knockout of the EP3 receptor in the
Several cytokines and immune mediators can induce sickness behav- median preoptic nucleus (MnPO) of the hypothalamus diminished
iour when administered in isolation, including interleukins, interferons, lipopolysaccharide-induced fever responses14, with PGE2 receptors
tumour necrosis factor (TNF) and eicosanoids1,2,4,8. These and other in other areas reportedly being relevant for effects on arousal and
cytokines, as well as pathogen-derived factors such as lipopolysac- feeding2. These and other findings have led to several possible mod-
charide, bacterial toxins and formyl peptides, can activate an assort- els in which (1) PGE2 can directly cross the blood–brain barrier owing
ment of central and/or peripheral sensory neurons4,9. Together, these to its hydrophobicity, (2) PGE2 can be detected or enter the brain at
observations raise the possibility that there are many pathways for circumventricular organs, and/or (3) PGE2 can be synthesized in the

Howard Hughes Medical Institute, Department of Cell Biology, Harvard Medical School, Boston, MA, USA. 2Department of Immunology, Harvard Medical School, Boston, MA, USA.
1

Present address: Department of Biology, Massachusetts Institute of Technology, Cambridge, MA, USA. ✉e-mail: [email protected]
3

660 | Nature | Vol 615 | 23 March 2023

brain itself2,15,16. Central PGE2 could then activate a distributed neural recombination at least one week before virus administration; prior
network, with different receptive brain regions such as the MnPO evok- tamoxifen treatment of control mice did not affect the subsequent
ing particular aspects of a characteristic sickness response17. behavioural responses to influenza infection (Extended Data Fig. 4b).
The roles of prostaglandin receptors in peripheral neurons have Influenza-induced decreases in feeding, drinking, movement, body
remained unclear, and furthermore, we reasoned that exogenous weight and survival were attenuated in Advillin-cre ER; Ptger3 flox mice
application of chemically defined pyrogens may produce a systemic (Fig. 2b), with effect magnitudes similar to ibuprofen treatment, but
inflammatory response, whereas natural infections in different loca- persisted in Nestin-cre; Ptger3 flox mice (Fig. 2a). Similar effects of Ptger3
tions and of varying intensities may instead trigger local neuro–immune gene deletion on sickness behaviour were observed when lower, less
response pathways that have remained unexplored. lethal doses of influenza virus were used (Extended Data Fig. 5). Mice
with attenuated sickness behaviour displayed a similar recovery time,
as normal feeding, drinking and motility were observed by about two
Influenza causes sickness through EP3 weeks after infection in all experimental groups. These observations
We developed a mouse model to characterize the neuronal mecha- indicated that influenza induces sickness responses through PGE2
nisms underlying influenza-induced sickness behaviour. Mice were action on peripheral sensory neurons.
infected by intranasal administration of influenza A virus PR/8/34 The EP3 receptor is expressed in several classes of peripheral neurons
(H1N1), and monitored for characteristic sickness responses over the marked in Advillin-cre ER mice, including vagal sensory neurons, glos-
subsequent 10–20 days. Influenza infection decreased food intake, sopharyngeal sensory neurons, spinal sensory neurons of the dorsal
water intake, locomotion, body weight and survival in wild-type mice. root ganglia, and potentially in other neuron types20,22. Glossopharyn-
Higher titres of virus inoculum increased the extent of sickness behav- geal and vagal sensory neurons were considered primary candidates
iour, with the most severe phenotypes arising six-to-seven days after for the detection of respiratory pathogens as they account for most of
infection (Fig. 1a). Influenza infection increased PGE2 levels in both the innervation in the upper and lower airways23. In mice, the soma of
plasma and bronchoalveolar lavage fluid (BALF) over a similar time vagal (nodose and jugular) and glossopharyngeal (petrosal) sensory
frame, as measured by enzyme-linked immunoassay (ELISA) (Extended neurons are fused into a large superganglion on each side of the body23.
Data Fig. 1b). PGE2 administration also acutely inhibited food intake We injected an adeno-associated virus (AAV) with a constitutive cre
(Extended Data Fig. 2a), and fibre photometry measurements showed allele (AAV-cre) bilaterally into both nodose–jugular–petrosal (NJP)
decreased activity of hypothalamic neurons expressing agouti- ganglia of Ptger3 flox mice. Effective knockout of Ptger3 in NJP ganglia was
related peptide (AGRP) (Extended Data Fig. 2c), consistent with a confirmed two weeks after AAV injection by RNA in situ hybridization
decreased motivation to eat, rather than solely a physical inability (Extended Data Fig. 6). NJP ganglion-targeted Ptger3 knockout caused a
to eat. We also observed a hypothermic response to influenza infec- marked attenuation of influenza-induced sickness behaviour (Fig. 2c).
tion (Extended Data Fig. 3a), consistent with previous observations These findings indicate that influenza induces behavioural changes
in mice18. Administration of ibuprofen (1 mg ml−1 in drinking water, through EP3 receptor expressed on vagal and/or glossopharyngeal
ad libitum) or aspirin (20 mg kg−1, daily intraperitoneal injection) sensory afferents.
to influenza-infected mice decreased plasma PGE2 levels, restored Ptger3 is expressed in a subset of vagal and glossopharyngeal sen-
feeding, water intake and body weight, and promoted survival from sory neurons, as revealed by RNA in situ hybridization (Extended
infection (Fig. 1b and Extended Data Figs. 1b and 3b). Ibuprofen- and Data Fig. 6b) and analysis of single-cell transcriptome data (Fig. 3a).
aspirin-treated mice retained low-level decreases in feeding, drinking Single-cell RNA sequencing approaches revealed dozens of molecularly
and motility without elevated PGE2, raising the possibility that other distinct NJP sensory neurons22,24,25. The highest Ptger3 expression was
neuro–immune communication pathways might also contribute to observed in 6 neuron clusters: J1, J2, J3, NP2, NP9 and NP26 (Fig. 3a),
the behavioural responses. Yet, ibuprofen and aspirin substantially with J denoting jugular and NP denoting nodose–petrosal neurons.
decreased influenza-induced sickness behaviour, consistent with a We crossed Ptger3 flox mice to Piezo2-IRES-cre mice, in which J1, J2, J3
key role for cyclooxygenase-2 metabolites in neuro–immune crosstalk. and other NJP neurons are labelled, and Phox2b-cre mice, in which NP2,
Roles for multiple prostaglandin receptors have been proposed in NP9, NP26 and other NJP neurons are labelled. Influenza-induced sick-
sickness behaviours2,19. We administered selective antagonists for each ness behaviours were attenuated in Phox2b-cre; Ptger3 flox mice across
PGE2 receptor daily after influenza infection, and measured different a range of viral titres, but not in Piezo2-IRES-cre; Ptger3 flox mice (Fig. 3b
aspects of sickness behaviour. The EP3 receptor antagonist DG-041 and Extended Data Figs. 7 and 8a), suggesting a role for either NP2, NP9
effectively blocked influenza-induced sickness in each parameter meas- or NP26 neurons. To distinguish these neuron types, we next crossed
ured and also promoted survival (Fig. 1c), with an effect magnitude Ptger3 flox mice to Pdyn-IRES-cre, Oxtr-IRES-cre and Gabra1-IRES-cre
similar to that of ibuprofen and aspirin. By contrast, antagonism of the mice, which display differential targeting of NP26, NP2 and NP9 neu-
EP1, EP2 or EP4 receptors had no effect on any measured parameter. rons. Gabra1-IRES-cre; Ptger3 flox mice displayed a marked attenuation
The EP3-selective agonist sulprostone had the opposite effect of inhib- of influenza-induced sickness behaviour similar to ibuprofen treat-
iting food intake (Extended Data Fig. 2b). Thus, mice display charac- ment, whereas no effect was observed in either Pdyn-IRES-cre; Ptger3 flox
teristic behavioural changes to influenza virus infection through the or Oxtr-IRES-cre; Ptger3 flox mice (Fig. 3b and Extended Data Fig. 8a).
action of PGE2 on EP3 receptors. Gabra1-IRES-cre; Ptger3 flox mice additionally displayed an attenuated
hypothermia response to influenza infection (Extended Data Fig. 3a).
We also tested a role for TRPV1 neurons since their deletion affects
Glossopharyngeal neurons and sickness survival following bacterial lung infections that cause lethal pneu-
The EP3 receptor is expressed in several classes of central and periph- monia26. However, influenza-induced sickness behaviour was normal
eral neurons. To determine the key site of EP3 receptor action in in Trpv1-IRES-cre; Ptger3 flox mice (Extended Data Fig. 8b,c), and fur-
influenza-induced sickness, we obtained mice with an allele (Ptger3 flox) thermore, GABRA1 neurons are predominantly Trpv1-negative. Thus,
for Cre-dependent knockout of the Ptger3 gene14, which encodes the deletion of the EP3 receptor in a small subset of Gabra1-expressing
EP3 receptor. We then crossed Ptger3 flox mice with either Nestin-cre peripheral sensory neurons (NP9 neurons) has a wide-ranging effect
or Advillin-creER mice, which target Cre recombinase to most central on behavioural responses to influenza infection.
or peripheral neurons20,21, respectively (Extended Data Fig. 4a), and Next, we examined how manipulations of sensory neurons affected
measured influenza-induced sickness behaviours. Advillin-cre ER; viral transcript levels and cytokine production. Influenza infection
Ptger3 flox mice were treated with tamoxifen to induce Cre-mediated induced PGE2 to similar levels in plasma and BALF of Gabra1-IRES-cre;

Nature | Vol 615 | 23 March 2023 | 661

Article
a Vehicle
Influenza A virus, low
b Influenza A virus
Influenza A virus + ibuprofen
c Vehicle
EP1 antagonist
Influenza A virus, medium + Influenza A
EP2 antagonist
Influenza A virus, high virus
10 10 10 EP3 antagonist
EP4 antagonist
0
0 0
–10

Food intake
(% change)
–20 –10 –10

–30 ***
–20 –20
–40 ** ***
*** –30 –30
–50 ***
–60 –40 –40
10 10 10
0
–10 0 0
–20
***
Water intake

***
(% change)

–10 –10
–30
–40 NS
–20 –20
–50 ***
–60
***
–30 –30
–70
–80 –40 –40
10 10 10
0
0 0
–10
–20
NS
–10 –10
***
***
(% change)

*** ***
Motility

–30
–20 –20
–40
–50 –30 –30
–60
–40 –40
–70
–80 –50 –50
10 10 10

0 0 0
***
Body weight
(% change)

–10 –10 *** –10

–20 –20 –20

***
–30 *** –30 –30
***
–40 –40 –40
100 100 100

80
NS
80 * 80
Survival rate

60 60 60 *
(%)

40 40 40
*
20 20 20

0 ** 0 0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Time after infection (days) Time after infection (days) Time after infection (days)

Fig. 1 | Influenza infection induces multiple sickness symptoms through P < 0.0001; survival: P = 0.0295. c, Mice were infected with influenza A virus,
the EP3 receptor. a, Mice were infected intranasally with 25 μl of influenza injected daily (intraperitoneal injection, 1 mg kg−1) with antagonists for
A virus at low (105 EID50 ml−1), medium (106 EID50 ml−1) or high (107 EID50 ml−1) EP1 (SC-51322), EP2 (PF-04418948), EP3 (DG-041) or EP4 (ONO-AE3-208), or
dose and food intake, water intake, motility and body weight were subsequently vehicle alone, and monitored as indicated. Data are mean ± s.e.m.; n = 10
monitored daily. EID50 is the 50% egg infective dose. Data are mean ± s.e.m.; mice for vehicle groups; n = 8 mice for all others. For EP3 antagonist—food
n = 6 mice per group. One-way ANOVA with Dunnett’s multiple comparison test intake: P < 0.0001; water intake: P < 0.0001; motility: P < 0.0001; body weight:
compared with vehicle control. Food intake: P = 0.0048 (low), P < 0.0001 P = 0.0002; survival: P = 0.0094. b,c, Two-tailed unpaired t-test, with
(medium), P < 0.0001 (high). Water intake: P = 0.8185 (low), P < 0.0001 (medium), comparisons between groups treated with ibuprofen or EP3 antagonist and
P < 0.0001 (high). Motility: P = 0.5231 (low), P < 0.0001 (medium), P < 0.0001 vehicle in c. Log-rank (Mantel–Cox) test for survival analyses; for behavioural
(high). Body weight: P = 0.0002 (low), P < 0.0001 (medium), P < 0.0001 (high). or physiological changes, a mean daily change in behaviour (days 1–10 after
Survival: P = 0.3173 (low), P = 0.0185 (medium), P = 0.0007 (high). b, Mice were infection or survival) was obtained for each mouse, and then used for
infected with influenza A virus (all infections are with 106 EID50 ml−1 unless comparisons across experimental groups (for more information see
otherwise indicated), given drinking water with or without 1 mg ml−1 ibuprofen, Extended Data Fig. 1a). *P < 0.05, **P < 0.005, ***P<0.0005; NS, not
and monitored as indicated. Data are mean ± s.e.m.; n = 6 mice per group. Food significant.
intake: P < 0.0001; water intake: P = 0.0001; motility: P = 0.0001; body weight:

Ptger3 flox, Phox2b-cre; Ptger3 flox, Advillin-creER; Ptger3 flox and Ptger3 flox (IFNγ), TNF and interleukin 6 (IL-6) similarly peaked in BALF of control
mice (Extended Data Fig. 9a), consistent with changes in PGE2 detection mice 5 days after infection, and were likewise decreased and delayed in
rather than PGE2 synthesis underlying the observed behavioural dif- Gabra1-IRES-cre; Ptger3 flox mice (Extended Data Fig. 9c). These findings
ferences. In control mice, viral transcript levels peaked three days after indicate that targeted EP3 receptor knockout in sensory neurons not
infection in the upper airways and five days after infection in the lungs. only affects sickness behaviour, but also the immune response and the
In Gabra1-IRES-cre; Ptger3 flox mice, viral transcript levels were partially transition from upper to lower respiratory tract infection.
reduced in the upper airways and were decreased, delayed and persis- We used a complementary approach involving diphtheria
tent in the lungs (Extended Data Fig. 9b). Levels of interferon-gamma toxin-guided cell ablation to clarify a role for Gabra1-expressing

662 | Nature | Vol 615 | 23 March 2023

 a Central b Peripheral c Vagal–glossopharyngeal
neurons neurons neurons

Ptger3flox Ptger3flox Ptger3flox; AAV-cre, vehicle

Nestin-cre; Ptger3flox Advillin-creER; Ptger3flox Ptger3flox, virus
10 10 10 Ptger3flox; AAV-cre, virus

0 0 0

–10 –10 –10 ***

***

Food intake
(% change)
***
–20
* –20 –20

–30 –30 –30

–40 –40 –40

–50 –50 –50

10 10 10

0 0 0
***
–10 –10 –10 ***
Water intake
(% change)

NS ***
–20 –20 –20

–30 –30 –30

–40 –40 –40

–50 –50 –50

10 10 10

0 0 0

–10 –10 –10 ***

(% change)

***
***
Motility

–20 –20 –20

NS
–30 –30 –30

–40 –40 –40

–50 –50 –50

10 10 10

0 0 0
Body weight
(% change)

*** ***
–10 –10 –10 ***
NS

–20 –20 –20

–30 –30 –30

100 100 100

80 80 80
Survival rate

60 60 60
(%)

* *
40 40 40
NS
20 20 20

0 0 0
0 5 10 15 20 0 5 10 15 20 0 5 10 15 20
Time after infection (days) Time after infection (days) Time after infection (days)

Fig. 2 | Peripheral EP3 receptor is required for influenza-induced sickness. exposed to influenza A virus or saline and monitored as indicated. Data are
a,b, Nestin-cre; Ptger3 flox (a), Advillin-cre ER ; Ptger3 flox mice (b) or Ptger3 flox mean ± s.e.m.; n = 8 mice per group. Ptger3 flox, virus—food intake: P < 0.0001;
(a,b) mice were infected with influenza A virus and monitored as indicated. water intake: P < 0.0001; motility: P < 0.0001; body weight: P < 0.0001;
Data are mean ± s.e.m.; n = 8 mice (Ptger3 flox), n = 6 mice (Nestin-cre and survival: P < 0.0001. Ptger3 flox; AAV-cre, virus—Food intake: P < 0.0001; water
Advillin-cre ER). Two-tailed unpaired t-test as detailed in Fig. 1 for behaviour or intake: P < 0.0001; motility: P < 0.0001; body weight: P < 0.0001; survival:
physiology analyses; log-rank (Mantel–Cox) test for survival analysis. a, Food P = 0.0376. One-way ANOVA with Dunnett’s multiple comparison test as
intake: P = 0.0126; water intake: P = 0.1006; motility: P = 0.0701; body weight: detailed in Fig. 1 for behaviour or physiology analyses; log-rank (Mantel–Cox)
P = 0.9361; survival: P = 0.7735. b, Food intake: P = 0.0004; water intake: test for survival analysis, with comparisons made between Ptger3 flox, virus and
P = 0.0004; motility: P < 0.0001; body weight: P = 0.0004; survival: P = 0.0216. Ptger3 flox; AAV-cre, vehicle (red stars) or between Ptger3 flox, virus and Ptger3 flox;
c, The NJP ganglia of Ptger3 flox mice were injected bilaterally with AAV-cre, AAV-cre, virus (blue stars).

vagal–glossopharyngeal sensory neurons in influenza-induced sick- NP9 neurons lack expression of Hoxb4(ref. 22), and GABRA1 neurons
ness, and rule out a role for other Gabra1 expression sites. Mouse cells are often clustered within NJP ganglia near the glossopharyngeal branch
are normally resistant to diphtheria toxin, but can be rendered sensi- (Extended Data Fig. 10a), suggesting they comprise part of the glos-
tive by the expression of the human diphtheria toxin receptor (DTR). sopharyngeal nerve rather than the vagus nerve. Since diphtheria toxin
NJP ganglia of Gabra1-IRES-cre; lsl-DTR mice were bilaterally injected injection similarly affected vagal and glossopharyngeal neurons, we
with diphtheria toxin (we term these Gabra1-ABLATE mice), resulting distinguished the contributions from these nerves by transecting the
in highly efficient ablation of vagal and glossopharyngeal GABRA1 glossopharyngeal nerve bilaterally while preserving the vagus nerve.
neurons (Fig. 4c). We previously demonstrated that this approach does Mice with bilateral glossopharyngeal nerve transection displayed a
not affect Cre-negative neurons in NJP ganglia, or remotely located similar attenuation of influenza-induced sickness (Fig. 4b). PGE2 was
Cre-positive neurons27. Gabra1-ABLATE mice displayed attenuated previously shown to activate and/or induce transcriptional changes in
sickness responses to influenza infection that were similar to those spinal and vagal afferents28–31, so we tested whether GABRA1 neurons
in Gabra1-IRES-cre; Ptger3 flox mice or ibuprofen-treated mice (Fig. 4a). are also directly activated. We observed that PGE2 directly evoked

Nature | Vol 615 | 23 March 2023 | 663

Article
a J2

NP26
J3
0.5 3 3

UMAP 2

UMAP 2

UMAP 2
NP9 0.4
0.3 2 2
0.2 1
0.1 1
NP2

J1 Ptger3 Piezo2 Phox2b

0.5 0.5 2.0

0.4 0.4
UMAP 2

UMAP 2

UMAP 2
1.5
0.3 0.3 1.0
0.2 0.2
0.1 0.1 0.5

Pdyn Oxtr Gabra1

UMAP 1 UMAP 1 UMAP 1
b Piezo2-IRES-cre Phox2b-cre Pdyn-IRES-cre Oxtr-IRES-cre Gabra1-IRES-cre
(J1, J2, J3) (NP2, NP9, NP26) (NP26) (NP2) (NP9)
Ptger3flox Ptger3flox Ptger3flox Ptger3flox Ptger3flox
10 Piezo2-IRES-cre; Ptger3flox 10 Phox2b-cre; Ptger3flox 10 Pdyn-IRES-cre; Ptger3flox 10 Oxtr-IRES-cre; Ptger3flox 10 Gabra1-IRES-cre; Ptger3flox

0 0 0 0 0
–10 –10 –10 –10 –10
NS NS NS
*** ***
Food intake
(% change)

–20 –20 –20 –20 –20

–30 –30 –30 –30 –30
–40 –40 –40 –40 –40
–50 –50 –50 –50 –50
10 10 10 10 10
0 0 0 0 0
–10 –10 –10 –10 –10
(% change)
Motility

–20 NS –20 –20 NS –20 NS –20

*** ***
–30 –30 –30 –30 –30
–40 –40 –40 –40 –40
–50 –50 –50 –50 –50
10 10 10 10 10

0 0 0 0 0
Body weight
(% change)

–10 NS –10 –10 NS –10 NS –10

*** ***
–20 –20 –20 –20 –20

–30 –30 –30 –30 –30

100 100 100 100 100

80 80 80 80 80
Survival rate

60 60 60 60 60 *
*
(%)

40 40 40 40 40
NS NS
20 NS 20 20 20 20

0 0 0 0 0
0 5 10 15 20 0 5 10 15 20 0 5 10 15 20 0 5 10 15 20 0 5 10 15 20
Time after infection (days) Time after infection (days) Time after infection (days) Time after infection (days) Time after infection (days)

Fig. 3 | Rare transcriptome-defined sensory neurons mediate influenza group. Two-tailed unpaired t-test as detailed in Fig. 1 for behaviour or
responses. a, A uniform manifold approximation and projection (UMAP) physiology analyses; log-rank (Mantel–Cox) test for survival analysis.
plot derived from published single-cell transcriptome data of vagal and Piezo2-IRES-cre—food intake: P = 0.2631; motility: P = 0.2680; body weight:
glossopharyngeal sensory ganglia22 showing expression of indicated genes P = 0.7925; survival: P = 0.4736. Phox2b-cre—food intake: P < 0.0001; motility:
(colour shows relative expression on a natural log scale). b, Cell types, as P < 0.0001; body weight: P = 0.0002; survival: P = 0.0181. Pdyn-IRES-cre—food
highlighted in a, express the indicated gene as well as Ptger3. Mice were infected intake: P = 0.4539; motility: P = 0.4946; body weight: P = 0.2675; survival:
with influenza A virus and monitored as indicated. Data are mean ± s.e.m.; P = 0.7937; Oxtr-IRES-cre—food intake: P = 0.4786; motility: P = 0.6333;
n = 8 (Piezo2-IRES-cre), n = 6–8 (Phox2b-cre; 8 control and 6 Phox2b-cre), n = 6 body weight: P = 0.3297; survival: P = 0.7121; Gabra1-IRES-cre—food intake:
(Pdyn-IRES-cre), n = 6 (Oxtr-IRES-cre) and n = 10 (Gabra1-IRES-cre) mice per P = 0.0001; motility: P < 0.0001; body weight: P = 0.0004; survival: P = 0.0263.

calcium transients in GABRA1 neurons acutely cultured from NJP sen-

sory ganglia of Gabra1-IRES-cre; lsl-tdTomato mice (Fig. 4d, 16 out of A sensory arc from nasopharynx to brain
26 or 61.5% of tdTomato-positive, KCl-responsive neurons). Thus, rare Gabra1-IRES-cre mice provide a valuable tool for visualizing scarce
glossopharyngeal NP9 neurons detect the presence of a respiratory glossopharyngeal sensory neurons involved in neuro–immune cross-
viral infection indirectly through the EP3 receptor and, in response, talk, so we used genetic mapping approaches to mark the peripheral
evoke a multi-pronged programme of sickness behaviour. and central projections of GABRA1 neurons (Fig. 5a). NJP ganglia of

664 | Nature | Vol 615 | 23 March 2023

 a GABRA1 neuron b Glossopharyngeal c
ablation transection
Gabra1-IRES-cre; lsl-DTR –DT Sham
10 Gabra1-IRES-cre; lsl-DTR +DT 10 Nerve cut

0
0
–10
***

Food intake
(% change)
DTR
–20 –10
***
–30
–20
–40

–DT
–50 –30
10 10

0
0
Water intake

***
(% change)

–10
*** –10
–20 DTR
–20
–30

–40 –30
10 10

+DT
0
–10 0
***
(% change)

–20 ***
Motility

–10
–30
–40 –20
–50
–60 –30 d GABRA1 neuron
10 10 PGE2 response

Control
Body weight

0
(% change)

–10
***
***
–10

PGE2
–20

–30 –20
100 100

80 80 100
Survival rate

Fluorescence
intensity (AU)

60 60 *
(%)

*
40 40 50

20 20
PGE2 PGE2 KCl
0 0 0
0 5 10 15 20 0 1 2 3 4 5 6 7 8 9 10 0 100 200 300 400 500 600
Time after infection (days) Time after infection (days) Time (s)

Fig. 4 | PGE2 acts via glossopharyngeal sensory neurons. a, Gabra1-IRES-cre; Two-tailed unpaired t-test as detailed in Fig. 1 for behaviour or physiology
lsl-DTR mice were injected bilaterally in NJP ganglia with or without diphtheria analysis; log-rank (Mantel–Cox) test for survival analysis. c, Immunostaining
toxin (DT) and then infected with influenza A virus and monitored as indicated. for DTR in whole-mount preparations of NJP ganglia four weeks after injection.
Data are mean ± s.e.m.; n = 8 mice per group. Food intake: P < 0.0001; water Scale bars, 200 μm. d, Calcium transients evoked by PGE2 (1 μM) or KCl
intake: P < 0.0001; motility: P < 0.0001; body weight: P = 0.0004; survival: (150 mM) were imaged using Calbryte 520 AM in tdTomato-positive neurons
P = 0.049. b, Wild-type mice with bilateral glossopharyngeal nerve transection acutely collected from NJP ganglia of Gabra1-IRES-cre; lsl-tdTomato mice.
surgery or sham surgery were infected with influenza A virus and monitored as Scale bar, 10 μm. Images are representative of three technical replicates. AU,
indicated. Data are mean ± s.e.m.; n = 8 mice per group. Food intake: P < 0.0001; arbitrary units.
water intake: P < 0.0001; motility: P < 0.0001; body weight: P < 0.0001; P = 0.036.

Gabra1-IRES-cre mice were injected with either an AAV containing a from NTS locations targeted by axons of some other Cre-defined NJP
Cre-dependent reporter gene encoding tdTomato (AAV-flex-tdTomato) neurons, such as gut-innervating neurons marked by expression of
or alkaline phosphatase (AAV-flex-AP), as well as an AAV containing a Gpr65 or Glp1r35. These findings further support a model for topo-
Cre-independent reporter gene encoding GFP (AAV-GFP) to visual- graphic organization in the brainstem36, with neurons relaying the
ize the global NJP projection field. Fibres were then visualized in the presence of an airway infection spatially restricted from at least some
airways and the brain. other interoceptive inputs.
Centrally, sensory neurons of the nodose and petrosal ganglia cross In the periphery, labelled GABRA1 axons were observed in only a
the skull and innervate brainstem regions that include the nucleus of few internal organs known to receive vagal and glossopharyngeal
the solitary tract (NTS) and area postrema32, whereas jugular sensory innervation. We observed densest innervation in the nasopharynx
neurons innervate the paratrigeminal nucleus33. Various Cre-defined and oral cavity including taste papillae, some innervation of tracheal
vagal afferents target spatially restricted subregions of the NTS, with and pharyngeal muscles, and sparse innervation of stomach muscle
some but not all afferent types also accessing the area postrema24,34,35. (Extended Data Fig. 10b). Innervation was not observed in many other
The axons of NJP GABRA1 neurons were highly restricted in the lat- internal organs, including the heart, oesophagus, aorta and carotid
eral NTS (Fig. 5a), not observed in the area postrema, and remote sinus. The nasopharynx is a rich site for immune surveillance, as it

Nature | Vol 615 | 23 March 2023 | 665

Article
a sickness behaviours were similarly attenuated by 1) ibuprofen, aspirin
and EP3 receptor antagonism, 2) Ptger3-targeted gene knockout using
Advillin-creER, Phox2b-cre and Gabra1-IRES-cre mice, and direct AAV-cre
injection in NJP ganglia, 3) ablation of GABRA1 NJP neurons, and 4) and
All vagal–glosso neurons
glossopharyngeal nerve transection. From these combined results, we
All vagal–glosso neurons GABRA1 neurons GABRA1 neurons conclude that a small cluster of Gabra1-expressing glossopharyngeal
b c sensory neurons detects PGE2 and induces a neuronally orchestrated
state of sickness associated with a suite of behavioural responses to

Control
respiratory viral infection.
Nasopharynx Prostaglandins and many other cytokines can activate vagal and
COX2
CO COX2 other peripheral afferents in vitro8,29,31, but their physiological roles in
naturalistic infections have been difficult to parse. Vagotomy blocks
sickness responses to cytokines and lipopolysaccharide in some stud-
Influenza A virus

ies but not others42,43, and this variability may be owing to the location
or dose of pyrogen administration. Our findings indicate a clear role
COX2
OX2 COX2 for sparse glossopharyngeal sensory neurons that are anatomically
poised to detect influenza-induced PGE2. PGE2 has limited stability
Control in vivo44, so peripheral sensory neurons that directly detect PGE2 at
GABRA1 neurons

***
the infection site would seemingly provide a fast and robust conduit
Influenza A virus
for information transfer to the brain.
GABRA1 neurons 0 5 10 15 20 25 30 35 Sickness behaviours have been proposed to help conserve energy
COX2 immunofluorescence (AU)
to fight off infection1. However, eliminating the influenza detection
Fig. 5 | GABRA1 neurons provide an airway–brain communication route. pathway of the glossopharyngeal nerve not only attenuates sickness
a, NJP ganglia of Gabra1-IRES-cre mice were injected bilaterally with Cre- behaviour, but also promotes survival. Knockout of EP3 from a small
independent AAV-GFP and Cre-dependent AAV-flex-tdTomato, and fluorescent group of glossopharyngeal neurons is sufficient to mimic the profound
axons were visualized (green: all NJP sensory axons; red: GABRA1 NJP axons) by survival phenotype observed in full-body knockouts. One model is that
immunohistochemistry for tdTomato and GFP in fixed coronal cryosections of
pathogen-induced anorexia is beneficial in some infection models but
mouse brainstem. Scale bar, 200 μm. b, NJP ganglia of Gabra1-IRES-cre mice
harmful in others. For example, blocking PGE2 production is harmful
were injected bilaterally with AAV-flex-tdTomato, and axons were visualized by
during mycobacterium 3 infection45, but promotes survival during
immunostaining for tdTomato in fixed coronal cryosections of nasopharynx.
Scale bars, 100 μm (top), 50 μm (bottom). c, Top, immunohistochemistry of
influenza infection19,46. Moreover, glucose gavage is detrimental during
COX2 in fixed cryosections of nasopharynx in uninfected mice (control) or bacterial sepsis, where blood glucose may provide direct fuel for the
mice infected for five days with influenza A virus. Scale bars, 100 μm. Bottom, pathogen, but protective during influenza infection47. Thus, disrupting
quantification of COX2 immunofluorescence in the nasopharynx of indicated neuronal responses to infection may decrease mortality by attenuat-
mice. Data are mean ± s.e.m.; n = 29 sections from 5 mice per group over 3 ing changes in feeding behaviour. Moreover, an infection-activated,
independent experiments. Two-tailed unpaired t-test, P < 0.0001. Left panel in anorexia-promoting neural pathway may provide a net survival benefit
part a adapted with permission from ref. 35, Elsevier. Top panel in part b created during certain bacterial infections, and thus is maintained over evolu-
with BioRender.com. tion, but is harmful during viral infection, as observed here. In addition,
sickness behaviour may provide a separate population-level benefit as
sick animals seek isolation and thereby limit pathogen transmission
connects the nasal cavity to the rest of the respiratory system and to kin48.
provides an early line of mucosal immune defence37. Influenza infec- These findings indicate that targeted EP3 receptor knockout in sen-
tion increases PGE2 levels locally within the nasopharynx38, and in sory neurons not only affects sickness behaviour, but also the immune
humans leads to inflammation of nasopharyngeal tonsils39. Mice have response and viral transition from the upper to lower respiratory tract.
an orthologous system known as nasopharynx-associated lymphoid A parsimonious interpretation of these findings is that petrosal GABRA1
tissue40, so we tested whether GABRA1 axons were located near rel- neurons, on detection of PGE2, engage neural circuits that evoke coor-
evant immune mediators. GABRA1 axons were enriched dorsally in dinated responses that include sickness behaviours as well as motor
epithelial and subepithelial layers of the nasopharynx (Fig. 5b) and reflexes that affect immune function. Additionally, changes in feeding
notably, cyclooxygenase-2 expression was likewise detected in dorsal behaviour may secondarily affect immune function47,48 and activation
epithelium of the nasopharynx (Fig. 5c). Moreover, cyclooxygenase-2 of petrosal GABRA1 neurons may affect levels of cytokines other than
expression was markedly upregulated in dorsal nasopharynx after PGE2 (Extended Data Fig. 9c), which can potentially elicit further neu-
influenza infection in mice (Fig. 5c). Thus, GABRA1 neurons occupy a ronal feedback and enhance sickness behaviour. All such responses
strategic and privileged position for detection of the increased PGE2 to infection would critically depend on the EP3 receptor in petrosal
levels that occur during viral infection. GABRA1 neurons, which serves an essential role in first relaying the
presence of an upper respiratory infection to the brain, ultimately
evoking sickness behaviour.
Discussion Influenza infection-induced sickness behaviour was attenuated,
Despite seasonal vaccination and antiviral therapeutics, influenza A but not eliminated, following NSAID treatment, targeted EP3 receptor
viral infection remains one of the most severe human threats, affecting knockout, targeted neuronal ablation and glossopharyngeal nerve
millions of people worldwide each year. Cyclooxygenase-2 inhibitors transection, suggesting other routes to sickness. Of note, the kinetics of
are among the most prevalent anti-sickness and anti-pain medica- residual sickness behaviour following manipulation of petrosal GABRA1
tions, with around 30 billion doses of nonsteroidal anti-inflammatory neurons matches the time course of virus accumulation in the lungs,
drugs (NSAIDs) consumed annually in the USA alone41. Blockade of indicating that there are probably two phases of influenza-induced sick-
PGE2 production alleviates sickness, and several models have been ness behaviour. The first phase occurs when the virus is most prevalent
proposed for how the brain receives input about a peripheral infec- in the upper respiratory tract, and sickness is primarily mediated by
tion through PGE2 (ref. 2). Here, we observe that influenza-induced PGE2-detecting glossopharyngeal sensory neurons marked by GABRA1,

666 | Nature | Vol 615 | 23 March 2023

which project to the nasopharynx. The second phase of sickness occurs 20. Lau, J. et al. Temporal control of gene deletion in sensory ganglia using a tamoxifen-
inducible Advillin-Cre-ERT2 recombinase mouse. Mol. Pain 7, 100 (2011).
when the virus is most prevalent in the lower respiratory tract, and 21. Sisley, S. et al. Neuronal GLP1R mediates liraglutide’s anorectic but not glucose-lowering
the lung is primarily innervated by vagal rather than glossopharyn- effect. J. Clin. Invest. 124, 2456–2463 (2014).
geal sensory neurons. Our data raise the possibility that this second 22. Prescott, S. L., Umans, B. D., Williams, E. K., Brust, R. D. & Liberles, S. D. An airway
protection program revealed by sweeping genetic control of vagal afferents. Cell 181,
phase of sickness involves another neuronal pathway independent 574–589.e14 (2020).
of PGE2, EP3 and glossopharyngeal sensory neurons. Inhibitors of 23. Prescott, S. L. & Liberles, S. D. Internal senses of the vagus nerve. Neuron 110, 579–599
residual neuro–immune interaction pathways, once identified, could (2022).
24. Bai, L. et al. Genetic identification of vagal sensory neurons that control feeding. Cell 179,
potentially provide improved relief of influenza-induced malaise when 1129–1143.e23 (2019).
used in combination with NSAIDs. We also note that PGE2 receptors are 25. Kupari, J., Haring, M., Agirre, E., Castelo-Branco, G. & Ernfors, P. An atlas of vagal sensory
expressed in multiple NJP sensory neuron types, as well as in the brain neurons and their molecular specialization. Cell Rep. 27, 2508–2523.e4 (2019).
26. Baral, P. et al. Nociceptor sensory neurons suppress neutrophil and gammadelta T cell
and dorsal root ganglia. It is likely that the nervous system uses multiple responses in bacterial lung infections and lethal pneumonia. Nat. Med. 24, 417–426
sensory pathways to detect peripheral infections, with different sensors (2018).
perhaps specialized for particular pathogen types or infection locations 27. Min, S. et al. Arterial baroreceptors sense blood pressure through decorated aortic claws.
Cell Rep. 29, 2192–2201.e3 (2019).
in the body. In this way, gut and respiratory sensory neurons could 28. Baba, H., Kohno, T., Moore, K. A. & Woolf, C. J. Direct activation of rat spinal dorsal horn
potentially engage different neural circuits that lead, for example, to neurons by prostaglandin E2. J. Neurosci. 21, 1750–1756 (2001).
either cough or nausea31,49. Consistent with this notion, various routes 29. Coleridge, H. M. et al. Stimulation of ‘irritant’ receptors and afferent C-fibres in the lungs
by prostaglandins. Nature 264, 451–453 (1976).
to sickness are differentially sensitive to pharmacological inhibition; 30. Verzele, N. A. J. et al. The impact of influenza pulmonary infection and inflammation on
for example, gut malaise is treated clinically using antagonists of the vagal bronchopulmonary sensory neurons. FASEB J. 35, e21320 (2021).
serotonin receptor HTR3A50. Understanding the diversity of sensory 31. Mohammed, S. P., Higenbottam, T. W. & Adcock, J. J. Effects of aerosol-applied capsaicin,
histamine and prostaglandin E2 on airway sensory receptors of anaesthetized cats.
pathways to sickness and when they are engaged by different pathogen J. Physiol. 469, 51–66 (1993).
infections will provide an essential framework for deciphering this 32. Berthoud, H. R. & Neuhuber, W. L. Functional and chemical anatomy of the afferent vagal
complex and poorly understood physiological state, and may enable system. Auton. Neurosci. 85, 1–17 (2000).
33. McGovern, A. E. et al. Distinct brainstem and forebrain circuits receiving tracheal sensory
improved therapeutic interventions. neuron inputs revealed using a novel conditional anterograde transsynaptic viral tracing
system. J. Neurosci. 35, 7041–7055 (2015).
34. Chang, R. B., Strochlic, D. E., Williams, E. K., Umans, B. D. & Liberles, S. D. Vagal sensory
neuron subtypes that differentially control breathing. Cell 161, 622–633 (2015).
Online content 35. Williams, E. K. et al. Sensory neurons that detect stretch and nutrients in the digestive
Any methods, additional references, Nature Portfolio reporting summa- system. Cell 166, 209–221 (2016).
36. Ran, C., Boettcher, J. C., Kaye, J. A., Gallori, C. E. & Liberles, S. D. A brainstem map for
ries, source data, extended data, supplementary information, acknowl-
visceral sensations. Nature 609, 320–326 (2022).
edgements, peer review information; details of author contributions 37. Tacchi, L. et al. Nasal immunity is an ancient arm of the mucosal immune system of
and competing interests; and statements of data and code availability vertebrates. Nat. Commun. 5, 5205 (2014).
38. Tam, V. C. et al. Lipidomic profiling of influenza infection identifies mediators that induce
are available at https://doi.org/10.1038/s41586-023-05796-0.
and resolve inflammation. Cell 154, 213–227 (2013).
39. Eccles, R. Understanding the symptoms of the common cold and influenza. Lancet Infect.
Dis. 5, 718–725 (2005).
1. Poon, D. C., Ho, Y. S., Chiu, K., Wong, H. L. & Chang, R. C. Sickness: from the focus on
40. Kiyono, H. & Fukuyama, S. NALT- versus Peyer’s-patch-mediated mucosal immunity.
cytokines, prostaglandins, and complement factors to the perspectives of neurons.
Nat. Rev. Immunol. 4, 699–710 (2004).
Neurosci. Biobehav. Rev. 57, 30–45 (2015).
41. Green, G. A. Understanding NSAIDs: from aspirin to COX-2. Clin. Cornerstone 3, 50–60
2. Saper, C. B., Romanovsky, A. A. & Scammell, T. E. Neural circuitry engaged by
(2001).
prostaglandins during the sickness syndrome. Nat. Neurosci. 15, 1088–1095 (2012).
42. Konsman, J. P., Luheshi, G. N., Bluthe, R. M. & Dantzer, R. The vagus nerve mediates
3. Dennis, E. A. & Norris, P. C. Eicosanoid storm in infection and inflammation. Nat. Rev.
behavioural depression, but not fever, in response to peripheral immune signals; a
Immunol. 15, 511–523 (2015).
functional anatomical analysis. Eur. J. Neurosci. 12, 4434–4446 (2000).
4. Pinho-Ribeiro, F. A., Verri, W. A. Jr. & Chiu, I. M. Nociceptor sensory neuron–immune
43. Sehic, E. & Blatteis, C. M. Blockade of lipopolysaccharide-induced fever by
interactions in pain and inflammation. Trends Immunol. 38, 5–19 (2017).
subdiaphragmatic vagotomy in guinea pigs. Brain Res. 726, 160–166 (1996).
5. Funk, C. D. Prostaglandins and leukotrienes: advances in eicosanoid biology. Science
44. Forstermann, U. & Neufang, B. Elimination from the circulation of cats of
294, 1871–1875 (2001).
6-keto-prostaglandin E1 compared with prostaglandins E2 and I2. J. Pharm. Pharmacol.
6. Simonsen, L. et al. The impact of influenza epidemics on mortality: introducing a severity
35, 724–728 (1983).
index. Am. J. Public Health 87, 1944–1950 (1997).
45. Divangahi, M., Desjardins, D., Nunes-Alves, C., Remold, H. G. & Behar, S. M. Eicosanoid
7. Hart, B. L. Biological basis of the behavior of sick animals. Neurosci. Biobehav. Rev. 12,
pathways regulate adaptive immunity to Mycobacterium tuberculosis. Nat. Immunol. 11,
123–137 (1988).
751–758 (2010).
8. Zanos, T. P. et al. Identification of cytokine-specific sensory neural signals by decoding
46. Carey, M. A. et al. Contrasting effects of cyclooxygenase-1 (COX-1) and COX-2 deficiency
murine vagus nerve activity. Proc. Natl Acad. Sci. USA 115, E4843–E4852 (2018).
on the host response to influenza A viral infection. J. Immunol. 175, 6878–6884 (2005).
9. Chiu, I. M. et al. Bacteria activate sensory neurons that modulate pain and inflammation.
47. Wang, A. et al. Opposing effects of fasting metabolism on tissue tolerance in bacterial
Nature 501, 52–57 (2013).
and viral inflammation. Cell 166, 1512–1525.e12 (2016).
10. McCarthy, M. K. & Weinberg, J. B. Eicosanoids and respiratory viral infection: coordinators
48. Rao, S. et al. Pathogen-mediated inhibition of anorexia promotes host survival and
of inflammation and potential therapeutic targets. Mediators Inflamm. 2012, 236345
transmission. Cell 168, 503–516.e12 (2017).
(2012).
49. Zhang, C. et al. Area postrema cell types that mediate nausea-associated behaviors.
11. Pecchi, E., Dallaporta, M., Jean, A., Thirion, S. & Troadec, J. D. Prostaglandins and sickness
Neuron 109, 461–472.e5 (2021).
behavior: old story, new insights. Physiol. Behav. 97, 279–292 (2009).
50. Freeman, A. J., Cunningham, K. T. & Tyers, M. B. Selectivity of 5-HT3 receptor antagonists
12. Narumiya, S., Sugimoto, Y. & Ushikubi, F. Prostanoid receptors: structures, properties, and
and anti-emetic mechanisms of action. Anticancer Drugs 3, 79–85 (1992).
functions. Physiol. Rev. 79, 1193–1226 (1999).
13. Breyer, R. M., Bagdassarian, C. K., Myers, S. A. & Breyer, M. D. Prostanoid receptors: subtypes
and signaling. Annu. Rev. Pharmacol. Toxicol. 41, 661–690 (2001). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
14. Lazarus, M. et al. EP3 prostaglandin receptors in the median preoptic nucleus are critical published maps and institutional affiliations.
for fever responses. Nat. Neurosci. 10, 1131–1133 (2007).
15. Dantzer, R. et al. Molecular basis of sickness behavior. Ann. NY Acad. Sci. 856, 132–138 Open Access This article is licensed under a Creative Commons Attribution
(1998). 4.0 International License, which permits use, sharing, adaptation, distribution
16. Wang, T. A. et al. Thermoregulation via temperature-dependent PGD2 production in and reproduction in any medium or format, as long as you give appropriate
mouse preoptic area. Neuron 103, 309–322.e7 (2019). credit to the original author(s) and the source, provide a link to the Creative Commons licence,
17. Osterhout, J. A. et al. A preoptic neuronal population controls fever and appetite during and indicate if changes were made. The images or other third party material in this article are
sickness. Nature 606, 937–944 (2022). included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
18. Jhaveri, K. A., Trammell, R. A. & Toth, L. A. Effect of environmental temperature on sleep, to the material. If material is not included in the article’s Creative Commons licence and your
locomotor activity, core body temperature and immune responses of C57BL/6J mice. intended use is not permitted by statutory regulation or exceeds the permitted use, you will
Brain Behav. Immun. 21, 975–987 (2007). need to obtain permission directly from the copyright holder. To view a copy of this licence,
19. Coulombe, F. et al. Targeted prostaglandin E2 inhibition enhances antiviral immunity visit http://creativecommons.org/licenses/by/4.0/.
through induction of type I interferon and apoptosis in macrophages. Immunity 40,
554–568 (2014). © The Author(s) 2023

Nature | Vol 615 | 23 March 2023 | 667

Article
Methods were collected over 15 min intervals with a receiver (easyTEL, Emka
Technologies) connected to a computer running iox software v.2.10.8.
Animals All data points from one mouse on a given day were averaged.
All animal procedures followed the ethical guidelines outlined in the
National Institutes of Health Guide for the Care and Use of Laboratory Pharmacological manipulations
Animals, and all protocols were approved by the institutional animal Ibuprofen (I4883) and aspirin (acetylsalicylic acid, A5376) were pur-
care and use committee (IACUC) at Harvard Medical School. Mice were chased from Sigma, and PGE2 receptor antagonists SC-51322 (2791),
maintained under constant temperature (23 ± 1 °C) and relative humid- PF-04418948 (4818), DG-041 (6240), and ONO-AE3-208 (3565) were
ity (46 ± 5%) with a 12-h light:dark cycle. Wild-type C57BL/6J (000664), purchased from Tocris. Ibuprofen (1 mg ml−1, saline) was added to drink-
Nestin-cre (003771), Phox2b-cre (016223), Advillin-creER (032027), ing water ad libitum, and aspirin (20 mg kg−1, saline) and PGE2 receptor
Trpv1-IRES-cre (017769), Pdyn-IRES-cre (027958), Piezo2-EGFP-IRES-cre antagonists (1 mg kg−1, saline) were injected (intraperitoneally) daily.
(027719), Oxtr-IRES-cre (031303), Agrp-cre (012899), lsl-DTR (007900), Administration of cyclooxygenase-2 inhibitors and PGE2 receptor
lsl-tdTomato (Ai14, 007914) were purchased from Jackson Laborato- antagonists began 3 days before virus inoculation and continued to
ries. Gabra1-IRES-cre, lsl-L10-GFP and Ptger3 flox mice were previously the end of the monitoring period. Tamoxifen was administed daily for
generated14,22,51. 5 days (intraperitoneal injection, 70 mg kg−1) at least one week before
influenza virus infection.
Influenza A virus inoculation
Influenza A/PR/8/34 (H1N1) was purchased from Charles River Avian Ganglion injection
Vaccine Services (10100374) with a EID50 of 1012 ml−1. Virus was diluted NJP ganglia were injected with AAVs or diphtheria toxin as previously
in sterile saline for the administered dosage, which was EID50 106 ml−1 described34, with minor modifications. Mice were anaesthetized
unless otherwise indicated. Viral administration was adapted from (200 mg kg−1 avertin, intraperitoneal injection), and depth of anaesthe-
previous studies52. Awake mice were manually restrained and influenza sia ensured throughout the procedure by lack of a toe pinch response.
A virus was delivered through each nostril (25 μl total volume) slowly NJP ganglia were exposed and serially injected (10 × 13.8 nl) with saline
using a P200 pipette to prevent spillage or delivery to the mouth. After solution containing either AAVs (titre > 6.7 × 1012 vg ml−1) or diphtheria
administration, the mice were held with nostrils upright for 10–15 s, toxin (5 μg ml−1, Sigma) supplemented with 0.05% Fast Green FCF Dye
and returned to their cage. (Sigma) using a sharply pulled glass pipette attached to a Nanoject
Injector (Drummond). AAV-cre (AAV-Syn-Cre-GFP, SignaGen Laborato-
Behavioural monitoring ries, SL100892, AAV9), AAV-flex-tdTomato (AAV9.CAG.Flex.tdTomato.
For analysis of influenza A virus infection-induced behavioural changes, WPRE.BGH, Addgene, 51503, AAV9), AAV-GFP (pENN.AAV.CB7.CI.eGFP.
six- to eight-week-old mice without sex bias were singly housed with WPRE.rBG, Addgene, 105542, AAV9), and AAV-flex-AP (AAV9.CAG.flex.
free access to food and water. Baseline values were obtained by meas- PLAP.WPRE.bgH, custom virus, Boston Children’s Hospital Viral Core,
uring daily food intake, daily water intake, body weight, and motility Boston, MA, AAV9) were purchased. Successful injection was verified
for three days before inoculation. After viral infection, all measure- by Fast Green FCF Dye slowly filling the entire ganglion without leak-
ments were expressed as a percentage change from baseline values. age. Surgical wounds were closed with coated VICRYL Sutures ( J392H,
Food and water intake were calculated daily by measuring the change Ethicon) and animals received Buprenorphine SR (BupSR-LAB, ZooP-
in weight of remaining food and water in the cage. Animal movement harm, 20 mg kg−1 subcutaneously at the back of the neck) as an analgesic.
was recorded (C922x Pro Stream Webcam, Logitech) in the home cage Animals recovered for at least two weeks for behavioural analysis or
for 30 min daily and motility expressed as the total distance moved, as four weeks for histological analysis.
determined by the MTrack2 plug-in available in ImageJ (1.53q, NIH). For
measurement of acute food intake, mice were fasted for 24 h, and at Nerve transection
dark onset, were administered with PGE2 or sulprostone (EP3 agonist, Mice were anaesthetized (200 mg kg−1 avertin, intraperitoneal injec-
14765, Cayman Chemical Company) via the exposure route and doses tion), and depth of anaesthesia ensured throughout the procedure by
indicated in Figure legends. Animals were then singly housed in clean lack of a toe pinch response. NJP ganglia were surgically exposed, and
cages with ad libitum access to food and water, and the amount of food the glossopharyngeal nerve was identified and severed immediately
consumed for 1 h determined by measuring the food in the cage before adjacent to the ganglia using microscissors. Sham surgeries involved
and after the test period. NJP ganglia exposure without nerve transection. Surgical wounds
were closed with coated VICRYL Sutures ( J392H, Ethicon) and animals
Body temperature measurements received 20 mg kg−1 Buprenorphine SR (BupSR-LAB, ZooPharm, subcu-
Core body temperature was monitored by a rectal probe or radio telem- taneously at the back of the neck) as an analgesic. Animals recovered
etry. Rectal probe measurements were made hourly each day from for at least two weeks before behavioural analysis.
noon to midnight using a mouse rectal temperature probe (Kent Scien-
tific, RET-3, 0.16 mm tip diameter) connected to a thermometer (Kent Histological and anatomical analysis
Scientific WD-20250-9). Probes were sterilized (70% ethanol), lubri- Immunohistochemistry and analysis of native tissue fluorescence was
cated (Aquagel lubricating gel, ParkerLabs, 57-05), and inserted performed after intracardial perfusion of fixative (10 ml PBS, then 10 ml
~5–7 mm into the rectum of physically restrained mice, and temperature 4% paraformaldehyde in PBS). For whole-mount DTR immunostaining,
was recorded for 30 s. For radio telemetry, a radio telemetry device that NJP ganglia were collected and permeabilized (11.5 g glycine, 400 ml PBS
reports body temperature (HD-X11, DSI) was surgically implanted in the with 0.2% Triton X-100, 100 ml DMSO) at 37 °C for 1 week. Ganglia were
abdominal cavity. Surgery was performed under anaesthesia with aver- then blocked (1 h, room temperature) in blocking buffer (5% normal
tin (200 mg kg−1, intraperitoneal injection) and the abdominal area was donkey serum, 017-000-121, Jackson in 0.05% Tween-20/PBS) and incu-
disinfected and shaved. A ~1 cm incision was made, and the abdominal bated (overnight, 4 °C) with 1:200 anti-DTR (HB-EGF, Goat, AF259NA,
peritoneum at the linea alba was gently exposed to insert the sterile Fisher Scientific) in blocking buffer. Samples were washed (4 × 10 min,
radio telemetry device. The surgery site was sealed with VICRYL Sutures 0.05% Tween-20 in PBS, room temperature) and incubated (2 h, room
( J392H, Ethicon). Animals received Buprenorphine SR (BupSR-LAB, temperature) with anti-goat Alexa 488 solution ( Jackson Immunore-
ZooPharm, 20 mg kg−1 subcutaneously at the back of the neck), search, 705-545-147, 1:500 in PBS with 0.05% Tween-20). Tissues were
and were allowed to recover for at least a week. Body temperature data washed again (4 × 10 min, 0.05% Tween-20 in PBS, room temperature),
mounted (Fluoromount G medium, SouthernBiotech) onto microscope for Phox2b probe) in amplification buffer (humidified chamber, room
slides (Fisher Scientific), and visualized using a Leica SP5 II confocal temperature, overnight). Tissue was then washed (5× SSCT, 2 × 30 min,
microscope and analysed using ImageJ (1.53q). For cyclooxygenase-2 1 × 5 min, room temperature), mounted in Fluoromount G medium, and
immunostaining, nasopharyngeal tissue was cryosectioned (15 μm), visualized with by confocal microscopy (Leica SP5 II).
blocked (1 h, room temperature) with blocking solution (5% normal
donkey serum, PBS with 0.01% Triton X-100), and incubated (overnight, Single-cell transcriptomics
4 °C) with anti-COX2 antibody (rabbit, ab179800, Abcam, 1:500 in PBS All UMAP plots in this manuscript were made from published single-cell
with 0.01% Triton X-100). Sections were washed (3 × 10 min, room transcriptome data (GEO Accession ID: GSE145216)22 using Seurat (4.1.0)
temperature, PBS with 0.01% Triton X-100) and incubated (2 h, room and R Studio (4.1.2).
temperature) with anti-rabbit Alexa 488 solution ( Jackson Immunore-
search, 711-545-152, 1:1,000 in PBS with 0.01% Triton X-100). Sections Quantitative PCR analysis
were washed again (3 × 10 min, room temperature, PBS with 0.01% Triton Levels of mouse Ptger3, Gapdh transcript and the influenza virus nucleo-
X-100), mounted with Fluoromount G medium, and visualized using a protein (NP) were measured in cDNA obtained from hypothalamus,
Leica SP5 II confocal microscope. To visualize brainstem axons, brain- NJP ganglia, nasal lavage (for upper respiratory tract analysis), and/or
stem cryosections (20 μm) were obtained from AAV-injected mice, BALF (for lower respiratory tract analysis). BALF was collected by surgi-
and processed as described above for cyclooxygenase-2 immunostain- cally opening the neck and cannulating a 21G catheter into the trachea.
ing in nasopharyngeal tissue, except antibodies were 1:1,000 anti-GFP In total, 3 × 1 ml of PBS was slowly injected into the lungs then solution
(chicken, GFP-1020, Aves Labs), 1:1,000 anti-RFP (rabbit, 600-401-379, was gently aspirated back to the syringe and collected. Collected BALF
Rockland), anti-chicken Alexa 647 ( Jackson Immunoresearch, 703-605- was centrifuged (7 min, 400g, 4 °C), and the supernatants were stored
155, 1:300) and anti-rabbit Cy3 ( Jackson Immunoresearch, 111-165-144, (−80 °C) until analysis. Nasal lavage was performed by inserting a needle
1:300). To visualize axons in nasopharynx, fixed nasopharyngeal (15 μm) (26G) into the upper trachea angled toward the nose. Sterile PBS (1 ml)
sections were stained as above for cyclooxygenase-2 immunostain- was administered through the needle via a cannulated syringe (PE20
ing, but with anti-RFP primary antibody followed by anti-rabbit Cy3 polyethylene cannula, inner diameter 0.38 mm), and collected after
secondary antibody. For GABRA1 neuron innervation to the trachealis expulsion from the nose. RNA was collected and cDNA synthesized from
and inferior pharyngeal constrictor muscles, tissues were collected BALF, nasal lavage fluid, and homogenized tissue using Trizol (15596026,
fresh, cut along the ventral axis for open-book visualization, fixed ThermoFisher) and the High-Capacity cDNA Reverse Transcription Kit
(4% paraformaldehyde/PBS either 1 h, room temperature or overnight, (4368813, ThermoFisher) according to the manufacturer’s protocols.
4 °C), and stained for tdTomato as above for tdTomato visualization qPCR analysis was performed on cDNA using gene-specific prim-
in nasopharynx, except primary antibody incubation (36 h, 4 °C) and ers, PowerTrack SYBR Green Master Mix (A46012, ThermoFisher) on
subsequent washes (3 × 12 h, 4 °C) were longer. For visualizing alka- QuantStudio 7 qPCR instrument (ThermoFisher). Gapdh primers were
line phosphatase, animals were perfused with PBS, and tissues were GGGTGTGAACCACGAGAAATATG and TGTGAGGGAGATGCTCAGTG
collected, fixed (4% paraformaldehyde, 1 h, room temperature), and TTG; Ptger3 primers were CAACCTTTTCTTCGCCTCTG and TTTCTG
washed in cold PBS. Tissues were then incubated (70 °C, 2 h) in alkaline CTTCTCCGTGTGTG; and influenza virus nucleoprotein primers were
phosphatase buffer (0.1 M Tris HCl pH 9.5, 0.1 M NaCl, 50 mM MgCl2, 0.1% GACGATGCAACGGCTGGTCTG and ACCATTGTTCCAACTCCTTT.
Tween-20, 5 mM levamisole) and washed twice in alkaline phosphatase Expression levels of Ptger3 and nucleoprotein were normalized to
buffer. Alkaline phosphatase activity was visualized with NCT/BCIP Gapdh levels. Normalization was by the 2–ΔCT method for Extended
solution (34042, ThermoFisher Scientific) according to manufacturer’s Data Fig. 4a, and the 2–ΔΔCT method53 for Extended Data Fig. 9b, with
protocols. Stained samples were post-fixed (4% paraformaldehyde, added normalization using values from uninfected controls.
overnight, 4 °C), dehydrated through a series of ethanol washes, and
cleared using a 1:2 mixture of benzyl alcohol (Sigma-Aldrich 402834- PGE2 and cytokine measurements
500ML): benzyl benzoate (Sigma-Aldrich B6630-1L). The tissue was then Mice were euthanized at different time points after infection with influ-
cut along the ventral axis for open-book visualization. Whole-mount enza A virus. Whole blood was collected by cardiac puncture using a
stainings were captured by microscopy with an AxioZoom (Zeiss) and 0.5 M EDTA-coated needle and syringe and mixed immediately with
analysed using ImageJ (1.53q). 0.5 M EDTA (50 μl). Plasma was isolated by centrifugation (3,000 rpm,
10 min, 4 °C) and the supernatant stored (−80 °C) until ELISA analy-
Analysis of mRNA expression in situ sis. BALF was collected as described above. PGE2 and cytokine levels
DNA-based hybridization chain reaction (HCR) probes against Ptger3 were measured in plasma and/or BALF by ELISA (R&D Systems, PGE2:
(lot no. PRH698) and Phox2b (lot no. PRF341) were purchased from KGE004B, TNF: DY410-5, IFNγ: DY485-05, IL-6: DY406-05) according
Molecular Instruments. Hybridization solution (30% formamide, 5× SSC, to the manufacturer’s instructions.
9 mM citric acid (pH 6), 0.1% Tween-20, 50 μg ml−1 heparin, 1× Denhardt’s
solution, 10% dextran sulfate), probe wash buffer (30% formamide, Calcium imaging
5× SSC, 9 mM citric acid pH 6.0, 0.1% Tween-20, 50 μg ml−1 heparin), NJP ganglia were acutely collected from Gabra1-IRES-cre; lsl-tdTomato
amplification buffer (5× SSC, 0.1% Tween-20, 10% dextran sulfate), and mice and incubated (37 °C, 90 min, with rotation) in dissociation solu-
fluorophore-labelled HCR amplifiers were purchased from Molecular tion (Dulbecco’s Modified Eagle’s Medium (DMEM) containing lib-
Instruments. NJP ganglia were freshly collected, mounted in OCT (Sakura erase (55 mg ml−1, Roche, 05401135001), DNase (0.004%, Worthington,
Finetek) and cryosectioned (10 μm). Tissue was post-fixed (4% PFA, LS002007)). Cells were than pelleted (3 min, 300g, 4 °C) and resus-
PBS, room temperature, 20 min), washed (3 × 10 min, PBS, room tem- pended in 0.2% ovomucoid, 0.004% DNase, DMEM, 200 μl. Using a P200
perature), treated with 1% hydrogen peroxide (PBS, room temperature, pipette, cells were triturated at least 10 times until clumps were not
20 min), and incubated with Ptger3 and Phox2b HCR probes (0.4 pmol, visible, filtered through a 40 μm mesh cell strainer, and pelleted again
hybridization buffer in a humidified chamber, 37 °C, overnight). Tissue (3 min, 300g, 4 °C). Cells were then resuspended in culture medium con-
was then washed (15 min, 37 °C) with (1) 3:1 probe wash buffer: 5× SSCT taining 5 μM Calbryte 520 AM (20651, AAT Bioquest); culture medium
(5× SSC, 0.1% Tween-20), (2) 1:1 probe wash buffer: 5× SSCT, (3) 1:3 probe contains 1× B27 (Gibco, 17504044), 1× N2 (Gibco, 17502048), 1× penicil-
wash buffer: 5× SSCT, and (4) 5× SSCT. Tissue was then incubated with lin and streptomycin (Gibco, 15140122) in Neurobasal media without
amplification buffer (30 min, room temperature), and then with fluores- phenol red (Gibco, 12348017). Resuspended cells were then plated
cent HCR amplifiers (6 pmol, Alexa594 for Ptger3 probe and Alexa 488 onto laminin-coated cover glass and incubated (37 °C, at least 30 min
Article
before imaging) in a humidified CO2 incubator. The cover glass was then
transferred to a chamber with an inlet and outlet perfusion (Warner Data availability
Instruments, RC-24N) with continuous flow of Hank’s balanced salt All data used to generate figures, including behavioural data points
solution (HBSS). The calcium responses were then imaged (488 nm) on from each individual mouse are provided as source data. All reagents
perfusion of prostaglandin E2 (1 μM in HBSS, 2 min) and KCl (150 mM, that may not be commercially available including certain trans-
at the end of the run). GABRA1-positive neurons were identified by genic mice and AAVs will be made freely available upon reasonable
tdTomato fluorescence (568 nm). Cells were identified as responsive request. Source data are provided with this paper.
if mean Calbryte 520 AM fluorescence during the stimulation period
exceeded three standard deviations above the baseline mean. 51. Krashes, M. J. et al. An excitatory paraventricular nucleus to AgRP neuron circuit that
drives hunger. Nature 507, 238–242 (2014).
52. Hanson, L. R., Fine, J. M., Svitak, A. L. & Faltesek, K. A. Intranasal administration of CNS
Fibre photometry therapeutics to awake mice. J. Vis. Exp. https://doi.org/10.3791/4440 (2013).
Fibre photometry of AGRP neurons was performed as described54 with 53. Livak, K. J. & Schmittgen, T. D. Analysis of relative gene expression data using real-time
quantitative PCR and the 2(−ΔΔCT) method. Methods 25, 402–408 (2001).
PGE2 (intraperitoneal injection, 0.5 mg kg−1) or PBS acutely injected dur- 54. Horio, N. & Liberles, S. D. Hunger enhances food-odour attraction through a neuropeptide
ing a brief manual restraint using Synapses software (Build: 94-42329P, Y spotlight. Nature 592, 262–266 (2021).
55. Ilanges, A. et al. Brainstem ADCYAP1+ neurons control multiple aspects of sickness
Tucker-Davis Technology). Data were analysed using Matlab (R2020a).
behaviour. Nature 609, 761–771 (2022).
56. Nonomura, K. et al. Piezo2 senses airway stretch and mediates lung inflation-induced
Statistical analysis apnoea. Nature 541, 176–181 (2017).
Data in graphs are represented as mean ± s.e.m., with all sample sizes
and P values provided in figure legends. Statistical analysis for behav- Acknowledgements We thank C. Saper for Ptger3flox mice and comments on the manuscript.
Cartoons were created using BioRender.com. This work was supported by grants from the
ioural experiments involved averaging daily changes in parameters
NIH (DP1 AT009497 and R01 HL132255) and the Chan Zuckerberg Initiative to S.D.L., a Banting
indicated per mouse (days 1–10 after infection or through survival). Postdoctoral Fellowship to N.-R.B. and a Harvard Medical School Goldberg Fellowship to N.H.
Statistical significance was calculated with Prism 8.4.3 software (Graph- S.L.P. was a Warren Alperts Distinguished Scholar and S.D.L. is an investigator of the Howard
Hughes Medical Institute.
Pad), and involved statistical tests described in figure legends. Probit
analysis in Extended Data Fig. 5b was performed with SPSS, 21 (IBM). Author contributions N.-R.B., S.L.P., N.H., I.M.C. and S.D.L. designed experiments. N.-R.B., N.H.,
Sample sizes were chosen based on prior expertise and publications S.L.P. and Y.W. performed experiments. N.-R.B., S.L.P., N.H., I.M.C. and S.D.L. analysed data.
N.-R.B. and S.D.L. wrote the manuscript.
in our field26,55 and are disclosed in the Figure legends. Animal groups
were randomly assigned and control animals were age-matched to Competing interests S.D.L. is a consultant for Kallyope, Inc. The other authors declare no
experimental animals. The same investigator performed genotyping competing interests.
and analysis of sickness responses, so data were not generated blind
Additional information
to genotype or experimental group. Supplementary information The online version contains supplementary material available at
https://doi.org/10.1038/s41586-023-05796-0.
Reporting summary Correspondence and requests for materials should be addressed to Stephen D. Liberles.
Peer review information Nature thanks Alice McGovern, Ruslan Medzhitov and Kevin Tracey
Further information on research design is available in the Nature Port- for their contribution to the peer review of this work. Peer reviewer reports are available.
folio Reporting Summary linked to this article. Reprints and permissions information is available at http://www.nature.com/reprints.
 a

um
um
um

i
um

ed
ow

h
i

ig
i

ed
ed

ow

-M
ow

i
h

-L
ed

-H
ow

ig
ig

Average Change in Water Intake (%)

-M
ig
-M

Average Change in Body Weight (%)

-L
-L

-H
Average Change in Food Intake (%)

-M
-H

-L

za

za

za
20

-H
10 20 10

za

za

za

le
za

za

za

en

en

en
za

za

za

Average Change in Motility (%)

c
le
le

en

en

en
en

en

en

flu

flu

flu
hi
le

en

en

en

c
c

ve
flu

flu

flu
c

hi
flu

flu

flu

In

In

In
hi

flu

flu

flu
hi

ve
ve

In

In

In
In

In

In

ve

In

In

In
0 0
0 0

-10 -20 ns

-20 ns -10
-20 -40 

-40 -20
-30 -60  

 
-40 -60  -80 -30


b
Plasma Plasma BALF
vehicle vehicle
Influenza A virus Influenza A virus vehicle
Influenza A virus + ibuprofen Influenza A virus + aspirin Influenza A virus
7 7 *** ***
*** ***
*** *** 2.5
6 6 *** ***
*** ***
2.0
PGE2 (ng/mL)

PGE2 (ng/mL)
PGE2 (ng/mL)

5 5 ***
***
4 4 1.5 ***
3 3 *** ***
*** 1.0
2 2
*** *** 0.5
1 *** *** 1
*** *** *** *** *** ***
0 0 0.0
0 1 3 5 7 9 0 1 3 5 7 9 0 1 3 5 7 9 11
Days Post Infection Days Post Infection Days Post Infection

Extended Data Fig. 1 | Influenza A infection induces behavioral changes virus-infected mice (blue) were additionally given ad libitum access to
and PGE2 production. a, Statistical analysis for behavioral experiments ibuprofen (1 mg/ml) in drinking water from 3 days prior to infection (left) or
involved averaging daily changes in parameters indicated per mouse (days 1-10 received daily aspirin administration (IP, 20 mg/kg), mean ± sem, n: 3 mice per
after infection or through survival), as represented here in bar graphs. group, ***p < 0.0005 by two-way ANOVA followed by Bonferroni’s multiple
Statistical values are identical to those in Fig. 1a. b, PGE2 levels in plasma (left, comparison test with comparisons made between red and black curves (red
middle) and BALF (right) were measured by ELISA at time points indicated after stars) or red and blue curves (blue stars). p values in b are <0.0001 for all
exposure to influenza A virus (red, blue) or vehicle control (black). Some indicated stars.
Article
a Intraperitoneal Intranasal
vehicle vehicle
0.01 PGE2 0.0625 PGE2
1.5 *** 2.0
0.05 PGE2 *** 0.125 PGE2
0.5 PGE2 0.25 PGE2
0.5 PGE2
1.5

Food Intake (g)

Food Intake (g)
**
1.0
*
*** 1.0 **
0.5 ns
0.5

0.0 0.0

b Intraperitoneal Intranasal
***
1.5 1.5 ***
vehicle vehicle
0.025 EP3 agonist 0.025 EP3 agonist
0.1 EP3 agonist

Food Intake (g)

Food Intake (g)

0.5 EP3 agonist

1.0 1.0

0.5
* 0.5

0.0 0.0

c Time After Injection (min)

0.05
*** *** **
10 20 30 40 50 60
0.00
∆F / F

-0.05

vehicle
-0.10 PGE2

-0.15
vehicle PGE2
0.25 0.25
0.2 0.2
0.15 0.15
0.1 0.1
∆F / F

∆F / F

0.05 0.05
0 0
-0.05 -0.05
-0.1 -0.1
-0.15 -0.15
-10 0 10 20 30 40 50 60 -10 0 10 20 30 40 50 60
Time (min) Time (min)

Extended Data Fig. 2 | PGE2 decreases feeding and AGRP neuron activity. c, GCaMP6s fluorescence (ΔF/F) was measured in AGRP neurons of the arcuate
a, Food intake by fasted mice administered with PGE2 (left: IP, right: intranasal) nucleus by fiber photometry before and after IP injection of PGE2 (0.5 mg/kg
at doses indicated (mg/kg), and given ad libitum access to food (1 h), mean ± in PBS) or vehicle alone (PBS). (top) Responses are depicted as the mean of
sem, n: 9 (left) mice per group, 28 (vehicle), 10 (0.0625, 0.125, 0.25), 20 (0.5) measurements made in 10-minute time intervals (for example, 10 refers to the
in (right), ***p < 0.0005, **p < 0.005, *p < 0.05, ns: not significant by two-way mean of measurements made between 0 and 10 min), mean ± sem, n: 12 mice
ANOVA Tukey’s multiple comparison test. b, Food intake by fasted mice per group, **p < 0.01, ***p < 0.001 by two-way ANOVA with Bonferroni’s
administered with the EP3 receptor agonist sulprostone (left: IP, right: multiple comparison test, (bottom) representative recording traces with red
intranasal) at doses indicated (mg/kg), and given ad libitum access to food (1 h), bar indicating time of injection. p values left to right in a: IP: <0.0001, 0.0005,
mean ± sem, n: 8 mice per group, ***p < 0.0005, *p < 0.05 by two-way ANOVA <0.0001, intranasal: <0.0001, 0.0455, 0.0008, 0.6630; b, IP: <0.0001, 0.0312,
Tukey’s multiple comparison test (left) or two-tailed unpaired t-test (right). 0.0187, intranasal: <0.0001; c: <0.0001, <0.0001, 0.0045.
 a
flox-Ptger3, vehicle
flox-Ptger3, Influenza A virus flox-Ptger3
Gabra1-ires-Cre; flox-Ptger3, Influenza A virus Gabra1-ires-Cre; flox-Ptger3

Core Body Temperature (oC)

Core Body Temperature (oC)
37.5 37.0
37.0 36.5
36.5 ***
*** 36.0
36.0
*** 35.5
35.5
35.0
35.0
34.5 34.5

34.0 34.0
33.5 33.5
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Days Post Infection Days Post Infection

b
Influenza A virus
Influenza A virus + aspirin
10 Days Post Infection 10 10 10 100
Water Intake Change (%)

Body Weight Change (%)

Food Intake Change (%)

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

Motility Change (%)

1 2 3 4 5 6 7 8 9 10
0 0 0 80
0

Survival (%)
*
-10 -10 *** -10 *** 60
***
-10
***
-20 -20 -20 40
-20
-30 -30 -30 20

-40 -40 -40 -30 0

0 1 2 3 4 5 6 7 8 9 10
Days Post Infection

Extended Data Fig. 3 | Influenza-induced hypothermia and sickness so subsequent data is represented as a dashed line and statistical analysis was
behaviors are attenuated by cell-specific Ptger3 knockout and aspirin. performed only on data from days 1-4. b, Mice were given daily injections of
a, Core body temperature in flox-Ptger3 and Gabra1-ires-Cre; flox-Ptger3 mice aspirin (IP, 20 mg/kg), red or vehicle alone (black) throughout the paradigm.
was measured daily after administration of influenza virus (red, blue) or saline After three days, mice were infected with influenza A virus and subsequently
(black) using a rectal probe (left) or radiotelemetry (right), mean ± sem, n: 6 monitored as indicated, mean ± sem, n: 6 mice per group, ***p < 0.0005 by
(left) and 3 (right) mice per group, ***p < 0.0005 by one-way ANOVA Dunnett’s two-tailed unpaired t-test as detailed in Fig. 1 for behavior/physiology analysis,
multiple comparison test (left); ***p < 0.0005 by two-tailed unpaired t-test for *p < 0.05 by a log-rank (Mantel-Cox) test for survival analysis. p values in a, left:
right as detailed in Fig. 1 for behavior/physiology analysis (right), comparisons red <0.0001, blue 0.0003, right: <0.0001; b, left to right: <0.0001, <0.0001,
between red and black curves (red stars) or between red and blue curves (blue <0.0001, <0.0001, 0.0243.
stars). For data on the right, two of three control flox-Ptger3 mice died on day 5,
Article

Extended Data Fig. 4 | Validating approaches for cell-specific Ptger3 mice were treated with (right) or without (left) tamoxifen for 5 consecutive
knockout. a, qPCR analysis of Ptger3 expression in hypothalamus (left) or NJP days (IP, 70 mg/kg) as was done for Advillin-Cre ER; flox-Ptger3 mice. At least one
ganglia (right) of flox-Ptger3 (white), Nestin-Cre; flox-Ptger3 (red), and week later, mice were either exposed to influenza A virus (red) or saline (black)
Phox2b-Cre; flox-Ptger3 (blue) mice. Phox2b-Cre mice display Cre expression in and monitored as indicated, mean ± sem, n: 6 mice per group, ***p < 0.0005 by
nodose and petrosal but not jugular neurons56. Ptger3 transcript levels were two-tailed unpaired t-test as detailed in Fig. 1 for behavior/physiology analysis,
expressed after normalization to Gapdh expression levels, mean ± sem, n: 6 *p < 0.05 by a log-rank (Mantel-Cox) test for survival analysis. p values in a, left:
mice for flox-Ptger3 and 3 for other groups, ***p < 0.0005, ns: not significant by <0.0001, 0.7506, right: 0.8680, <0.0001; b top to bottom, left: <0.0001,
one-way ANOVA Dunnett’s multiple comparison test. b, flox-Ptger3 control <0.0001, <0.0001, <0.0001, 0.0179, right: <0.0001, 0.0003, <0.0001, 0.0179.
 a b c
LD50 ≈ EID50 105.778/ml
100 EID50 10 /ml
4 R2 = 0.926
100
EID50 104.5/ml
EID50 105/ml Viral Titre (EID50/ml) Alive Dead Survival rate (%) Estimated % lethality
80
EID50 105.5/ml
75
% Survival

EID50 106/ml 104 10 0 100 3.3

% Survival
60 EID50 106.5/ml 104.5 9 1 90 9.3
EID50 107/ml 50 8 2 80 21
105
40
105.5 5 5 50 39
25 106 4 6 40 59
20
106.5 3 7 30 77
0 0 107 1 9 10 90
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 4 5 6 7
Days Post Infection Log10 of Viral Titre (EID50/ml)

d
EID50 105/ml (LD21) flox-Ptger3
Advillin-CreER; flox-Ptger3

10 10 10 10 100
Days Post Infection Days Post Infection

% Body Weight Change

% Water Intake Change

Days Post Infection

% Food Intake Change

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Days Post Infection ns

% Motility Change
0 80
0 0 1 2 3 4 5 6 7 8 9 10

% Survival
0
-10 60
-10 -10
*** -20 40
*** *** -10 ***
-20 -20
-30 20

-30 -30 -40 -20 0

1 2 3 4 5 6 7 8 9 10
Days Post Infection
EID50 10 /ml (LD39)
5.5 flox-Ptger3
Advillin-CreER; flox-Ptger3

10 10 10 10 100
Days Post Infection

% Body Weight Change

% Food Intake Change

% Water Intake Change

Days Post Infection Days Post Infection Days Post Infection

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
% Motility Change

1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 80 *
0 0

% Survival
0
-10 60
-10 -10
-20 40
-10
-20 *** -20 ***
*** -30 20
***
-30 -30 -40 -20 0
1 2 3 4 5 6 7 8 9 10
Days Post Infection

Extended Data Fig. 5 | Ptger3 knockout in peripheral sensory neurons previously injected with tamoxifen) or flox-Ptger3 mice (black) indicated were
attenuates behavioral responses to sublethal influenza A doses. a, Wild infected with sublethal doses of influenza A virus (top: 105 EID50 or LD21, bottom:
type mice were infected with influenza A virus at titers indicated and 105.5 EID50 or LD39) and monitored daily as indicated, mean ± sem, n: 6 mice per
subsequent survival monitored daily, n: 10 per group. b, a dose-response curve group, ***p < 0.0005 by two-tailed unpaired t-test as detailed in Fig. 1 for
of viral titer (log10) vs. survival rates with non-linear fit of R 2 = 0.926. c, A table behavior/physiology analysis, *p < 0.05, ns: not significant by a log-rank
indicating survival rates to various influenza A virus inoculation doses, with (Mantel-Cox) test for survival analysis. p values left to right in d, top: <0.0001,
the estimated % lethality determined by Probit regression analysis (Statistical <0.0001, 0.0002, <0.0001, 0.4788; bottom: <0.0001, <0.0001, 0.0004,
Product and Service Solutions). d, Advillin-Cre ER ; flox-Ptger3 mice (red, 0.0004, 0.0078.
Article

Extended Data Fig. 6 | Validating AAV-driven Ptger3 knockout in NJP were subsequently examined by two-color RNA in situ hybridization to detect
ganglia. a, Cartoon depicting bilateral injection of AAV-Cre into NJP ganglia of Phox2b (green) and Ptger3 (red), scale bar: 100 μm. Images are representative
flox-Ptger3 mice. b, The NJP ganglia of flox-Ptger3 mice were injected bilaterally from three independent experiments. Part a created with BioRender.com.
with AAV-Cre (bottom) or saline (control, top), and cryosections of NJP ganglia
 EID50 105/ml (LD21) EID50 105.5/ml (LD39)
10 flox-Ptger3 10 flox-Ptger3
Phox2b-Cre; flox-Ptger3 Phox2b-Cre; flox-Ptger3
0 0
Food Intake
% Change

-10 -10
***

-20 -20 ***

-30 -30
10 10

0 0
Water Intake
% Change

-10 -10
**

-20 -20
***

-30 -30
10 10

0 0
% Change
Motility

-10 -10

-20 *** -20

***
-30 -30

-40 -40
10 10
Body Weight

0 0
% Change

-10 -10
*** ***

-20 -20
100 100

ns *
80 80
Survival Rate
% Survival

60 60

40 40

20 20

0 0
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
Days Post Infection Days Post Infection

Extended Data Fig. 7 | Cell-specific Ptger3 knockout attenuates behavioral

responses to sublethal influenza A doses. Phox2b-Cre; flox-Ptger3 mice (red)
or flox-Ptger3 mice (black) were infected with sublethal doses of influenza A
virus (top: 105 EID50 or LD21, bottom: 105.5 EID50 or LD39) and monitored daily as
indicated, mean ± sem, n: 6 mice per group, **p < 0.005, ***p < 0.0005 by
two-tailed unpaired t-test as detailed in Fig. 1 for behavior/physiology analysis,
*p < 0.05, ns: not significant by a log-rank (Mantel-Cox) test for survival
analysis. p values top to bottom left: <0.0001, <0.0001, <0.0001, 0.0002,
0.4524; right: <0.0001, <0.0001, <0.0001, 0.0002, 0.0155.
Article

Extended Data Fig. 8 | Measures of flu-induced behavioral changes in Trpv1 expression (red shading: natural log scale). c, Trpv1-ires-Cre; flox-Ptger3
cell-specific Ptger3 knockouts. a, Water intake changes after flu infection in mice (blue) or flox-Ptger3 mice (black) were infected with influenza A virus
mice of Fig. 3, mean ± sem, n: 8 (Piezo2-ires-Cre), 6-8 (Phox2b-Cre; 8 control and and monitored daily as indicated, mean ± sem, n: 6 mice per group, ns: not
6 Phox2b-Cre), 6 (Pdyn-ires-Cre), 6 (Oxtr-ires-Cre), and 10 (Gabra1-ires-Cre) mice significant by two-tailed unpaired t-test as detailed in Fig. 1 for behavior/
per group, ***p < 0.0005, ns: not significant by two-tailed unpaired t-test as physiology analysis, ns: not significant by a log-rank (Mantel-Cox) test for
detailed in Fig. 1 for behavior/physiology analysis. b, A Uniform Manifold survival analysis. p values top to bottom in a: 0.9101, 0.0004, 0.0534, 0.9544,
Approximation and Projection (UMAP) plot derived from published single-cell <0.0001 and c: 0.2485, 0.0705, 0.0888, 0.1801, 0.8412.
transcriptome data of vagal and glossopharyngeal sensory ganglia22 indicating
 a flox-Ptger3, Influenza A virus
Phox2b-Cre; flox-Ptger3, Influenza A virus
Advillin-CreER; flox-Ptger3, Influenza A virus
Gabra1-ires-Cre; flox-Ptger3, Influenza A virus
flox-Ptger3, vehicle
Plasma BALF

7 3.0

6 2.5
5
PGE2 (ng/mL)

PGE2 (ng/mL)

ns
2.0

ns
4
1.5
3
1.0
2

1 0.5

0 0.0
0 1 3 5 7 9 11 0 1 3 5 7 9 11
Days post infection Days post infection

b flox-Ptger3 (Upper Respiratory Tract)

flox-Ptger3 (Lower Respiratory Tract)
Gabra1-ires-Cre; flox-Ptger3 (Upper Respiratory Tract)
Gabra1-ires-Cre; flox-Ptger3 (Lower Respiratory Tract)

10
Normalized Viral NP
Transcript Levels

***
7.5
***

5 *** ***
***
2.5
***

0
0 1 5 10 15
Days Post Infection

c
flox-Ptger3 flox-Ptger3 flox-Ptger3
Gabra1-ires-Cre; flox-Ptger3 800 Gabra1-ires-Cre; flox-Ptger3 Gabra1-ires-Cre; flox-Ptger3
1400 1600
700
1200 1400
IFNgamma (pg/ml)

TNFalpha (pg/ml)

600
1200
1000
IL-6 (pg/ml)

500
1000
800
400 800 ***
600 300
*** *** 600
*** ***
400 200
***
400
*** *** ***
200 100 200
*** *** ***
0 0 0
0 1 3 5 7 9 11 13 15 0 1 3 5 7 9 11 13 15 0 1 3 5 7 9 11 13 15
Days Post Infection Days Post Infection Days Post Infection

Extended Data Fig. 9 | Impact of targeted Ptger3 knockout on cytokines followed by Bonferroni’s multiple comparison test with comparisons made
and viral transcript levels. a, PGE2 levels in plasma (left) and BALF (right) were between red and blue curves (blue stars) or black and green curves (green
measured by ELISA in mice indicated at various time points after exposure to stars). c, Levels of IFNγ, TNFα, and IL-6 in BALF were measured by ELISA at
influenza A virus or PBS control, mean ± sem, n: 5 mice per group, ns: not time points indicated after exposure to influenza A virus in flox-Ptger3
significant by two-way ANOVA involving analysis of influenza virus-infected (black) or Gabra1-ires-Cre; flox-Ptger3 (red), mean ± sem, n: 3 mice per group,
groups. b, qPCR analysis of viral nucleoprotein (NP) transcript levels in the ***p < 0.0005 by two-way ANOVA followed by Bonferroni’s multiple
upper and lower respiratory tract of flox-Ptger3 and Gabra1-ires-Cre; flox-Ptger3 comparison test with comparisons made between red and black curves (red
mice after influenza virus infection, normalized to Gapdh and uninfected stars). p values in a: 0.1591, 0.0662, b and c: <0.0001 for all indicated stars.
controls, mean ± sem, n: 5 mice per group, ***p < 0.0005 by two-way ANOVA
Article

Extended Data Fig. 10 | Sparse innervation of internal organs by GABRA1 phosphatase substrate (top two rows, left images in bottom two rows) or
NJP neurons. a, Native GFP fluorescent signals in wholemount preparations tdTomato immunostaining (right two images in bottom two rows). Scale bars
of NJP ganglia from Gabra1-ires-cre; lsl-L10 GFP, scale bar: 200 μm. b, NJP (left to right) top row: 500, 1000, 1000 (inset: 250) μm; 2nd row: 500, 1000,
ganglia of Gabra1-ires-Cre mice were injected bilaterally with Cre-dependent 500 μm; 3rd row: 200, 50 μm; bottom row: 200, 100 μm. Images are
AAV-flex-AP or AAV-flex-tdTomato and axons were visualized in fixed representative of three independent experiments involving GFP and tdTomato
wholemount tissue preparations using either a colorimetric alkaline and two independent experiments involving alkaline phosphatase.