Activation of temperature-sensitive TRPV1-like receptors in ARC POMC neurons reduces food intake

Proopiomelanocortin (POMC) neurons in the arcuate nucleus of the hypothalamus (ARC) respond to numerous hormonal and neural signals, resulting in changes in food intake. Here, we demonstrate that ARC POMC neurons express capsaicin-sensitive transient receptor potential vanilloid 1 receptor (TRPV1)-like receptors. To show expression of TRPV1-like receptors in ARC POMC neurons, we use single-cell reverse transcription-polymerase chain reaction (RT-PCR), immunohistochemistry, electrophysiology, TRPV1 knock-out (KO), and TRPV1-Cre knock-in mice. A small elevation of temperature in the physiological range is enough to depolarize ARC POMC neurons. This depolarization is blocked by the TRPV1 receptor antagonist and by Trpv1 gene knockdown. Capsaicin-induced activation reduces food intake that is abolished by a melanocortin receptor antagonist. To selectively stimulate TRPV1-like receptor-expressing ARC POMC neurons in the ARC, we generate an adeno-associated virus serotype 5 (AAV5) carrying a Cre-dependent channelrhodopsin-2 (ChR2)–enhanced yellow fluorescent protein (eYFP) expression cassette under the control of the two neuronal POMC enhancers (nPEs). Optogenetic stimulation of TRPV1-like receptor-expressing POMC neurons decreases food intake. Hypothalamic temperature is rapidly elevated and reaches to approximately 39 °C during treadmill running. This elevation is associated with a reduction in food intake. Knockdown of the Trpv1 gene exclusively in ARC POMC neurons blocks the feeding inhibition produced by increased hypothalamic temperature. Taken together, our findings identify a melanocortinergic circuit that links acute elevations in hypothalamic temperature with acute reductions in food intake.


Introduction
Body temperature changes according to the circadian rhythm, physical activities, and external factors such as food intake [1][2][3].Although it is expected that a rise in body temperature is rapidly conveyed to the hypothalamus via the blood stream, the physiological consequences of such an increase in hypothalamic temperature have not been studied.The transient receptor potential vanilloid subfamily members 1-4 (TRPV1, 2, 3, and 4) in peripheral sensory neurons play a role in accurately sensing ambient temperature and in adapting to environmental temperature changes through behavioral and physiological responses [4][5][6][7][8][9].In the early 1960s, Nakayama and colleagues showed that local hypothalamic heating from 36.7 ˚C to 38.7 ˚C increases hypothalamic neuronal activity [10], suggesting that, like peripheral sensory neurons, hypothalamic neurons are able to respond to changes in temperature, possibly via activation of temperature-sensitive transient receptor potential vanilloid (TRPV) channels.Indeed, neurophysiological, genetic, and in situ hybridization studies have described expression of TRPV receptors, in particular TRPV1 channels, specifically in hypothalamic neurons [11][12][13][14].
However, the presence of functional TRPV1 in the hypothalamus is still controversial.In fact, TRPV1 reporter mouse studies have shown very limited expression of TRPV1 receptors in the hypothalamus [12].Second, TRPV1-expressing vasopressin neurons show temperaturesensitive inward currents at temperatures above 35 ˚C [13,15,16], which is in contrast to the prior findings that recombinant TRPV1 receptors are activated only by noxious temperatures (>42 ˚C) [4].Third, transcriptome analysis of agouti-related peptide (AgRP) and proopiomelanocortin (POMC) neurons in the arcuate nucleus of the hypothalamus (ARC) shows no expression of Trpv1 transcripts [17].Interestingly, some ARC POMC neurons are situated outside the blood-brain barrier [18,19], which permits them to detect acute changes in temperature.Therefore, we sought to determine whether ARC POMC neurons express temperaturesensitive TRPV1 receptors.
Using multiple independent techniques, we found that ARC POMC neurons expressed capsaicin-and temperature-sensitive TRPV1 subunit-containing receptors (TRPV1-like).Pharmacologic and optogenetic activation of ARC POMC neurons that express TRPV1-like channels reduced food intake.Increased body temperature during exercise was transmitted to the ARC.This elevation was associated with a reduction in food intake that was blocked by a melanocortin receptor antagonist.These results support the interpretation that activation of capsaicin-sensitive TRPV1-like receptors in ARC POMC neurons contributes to the regulation of acute food intake.

Competing interests:
The authors have declared that no competing interests exist.

ARC POMC neurons express capsaicin-sensitive TRPV1-like receptors
The ARC expresses Trpv1 mRNA and [ 3 H]resiniferatoxin(RTX) binding sites [11,14].Among the class of TRPV channels, only TRPV1 channels respond to capsaicin [4,[20][21][22][23].Consequently, we examined whether ARC POMC neurons respond to the selective TRPV1 receptor agonist capsaicin [24][25][26] To further examine the expression of TRPV1 receptors in ARC POMC neurons, we used the TRPV1-Cre knock-in animal model [12].An adeno-associated virus serotype 1 (AAV1) encoding Cre-inducible GFP (AAV1-DIO-GFP) was injected into the ARC of TRPV1-Cre mice.Two weeks post viral injections, we immunostained hypothalamic sections with anti-GFP and anti-POMC antibodies.We found that the ARC contained GFP-expressing cells in the ARC, confirming the expressing of TRPV1-positive cells in this area.Importantly, GFP-positive cells also expressed POMC (Fig 1E).These results from the application of molecular, genetic, and immunohistochemical approaches strongly support the interpretation that a population of ARC POMC neurons expresses capsaicin-sensitive TRPV1 receptors and provide our rationale for subsequent investigation of the roles of TRPV1 receptors in ARC POMC neurons in the regulation of feeding in mice.

ARC POMC neurons express functional temperature-sensitive TRPV1-like receptors
Increased temperature activates capsaicin-sensitive TRPV1 receptors [4].We examined the neurophysiological impact of temperature on TRPV1-expressing POMC neuronal activity.Whole-cell voltage-clamp recordings were obtained from ARC POMC neurons in the presence of synaptic blockers (CNQX, 10 μM, and picrotoxin, 100 μM) and TTX (0.5 μM).Raising bath temperature from 34 ˚C up to 38 ˚C elicited an inward current in ARC POMC neurons (S6A and S6B Fig) .In addition, these thermal stimuli significantly depolarized ARC POMC neurons in current-clamp configurations in the presence of synaptic blockers (Fig 1F and 1G and S6C Fig).Approximately 40% of POMC neurons did not respond to increased temperature (S6D and S6E Fig) .This depolarization was blocked by the TRPV1 receptor antagonist AMG9810 and by Trpv1 gene knockdown with short hairpin RNA (shRNA) in the ARC (Fig 1F , 1G and 1H).In contrast, NPY neurons did not respond to increased temperature (S5D and S5E Fig) .It has been demonstrated that recombinant TRPV1 receptors responded only to noxious temperatures (>42 ˚C) [4].Interestingly, TRPV receptors could coassemble into heteromeric channels, and heteromeric TRPV1 receptors with other TRPV subunits had altered temperature sensitivity [23,29].For instance, TRPV1/TRPV3 channels had a lower threshold temperature, around 34 ˚C [23].Importantly, the sensitivity to capsaicin of heteromeric TRPV channels was determined by the presence of TRPV1 subunits [23].As ARC POMC neurons had Trpv3 and Trpv4 mRNA along with Trpv1 mRNA, we examined whether knockdown of the Trpv3 and Trpv4 genes in the ARC affect thermal sensitivity.POMC neurons in mice injected with Trpv3 and Trpv4 shRNA were less sensitive to increased temperature (S7A and S7B Fig) .Therefore, our results suggest that ARC POMC neurons express TRPV1 receptor subunit-containing heteromeric TRPV channels (hence, we describe them as "TRPV1-like").

Capsaicin-induced depolarization reduces food intake
Given that ARC POMC neurons play a role in the control of food intake [30][31][32][33][34], we asked whether activation of TRPV1-like receptors in ARC POMC neurons is sufficient to reduce food intake.To test this possibility, animals were fasted for 6 h (from 1 PM to 7 PM) before drug infusion and capsaicin was directly infused into the ARC.Local microinjection of capsaicin (100 nM) significantly reduced food intake (  ).Therefore, activation of capsaicin-sensitive TRPV1-like receptors in the ARC decreases food intake.
We then examined whether capsaicin-induced depolarization of ARC POMC neurons reduces food intake.We expressed inhibitory designer receptors exclusively In in vivo studies of these animals, intraperitoneal administration of CNO alone did not change food intake as described in prior studies [35].However, CNOinduced inhibition of ARC POMC neurons was sufficient to block the effect of capsaicin on food intake (Fig 3C and 3D and S1 Table), suggesting that inhibition of ARC POMC neurons counteracts capsaicin-induced excitation of ARC POMC neurons.
ARC POMC neurons synthesize and release POMC-derived peptides such as anorexigenic α-melanocyte-stimulating hormone (α-MSH) [36,37] and depolarization of ARC POMC neurons releases α-MSH, resulting in reduced food intake [36].We examined whether the anorexigenic effect of capsaicin is due to the release of α-MSH.We intracerebroventricularly injected the melanocortin 3/4 receptor antagonist SHU9119 prior to local injection of capsaicin.While injection of SHU9119 had no feeding effect when administered alone (Fig 3E ), pretreatment with SHU9119 abolished the anorexigenic effect of capsaicin (Fig 3E and S1 Table).In contrast, capsaicin was still effective in reducing feeding in the presence of the antagonist for the opioid receptor naloxone (1 mM; S8 Fig) .Our results support the interpretation that activation of TRPV1-like receptor induces the release of anorexigenic α-MSH from ARC POMC neurons and consequently reduces food intake.

Activation of TRPV1-like receptors expressed in POMC neurons reduces feeding
Next, we investigated whether activation of TRPV1-like receptors expressed exclusively in ARC POMC neurons regulates feeding.To this end, we used the Cre-dependent clustered regularly interspaced short palindromic repeats/CRISPR-associated protein 9 (CRISPR/Cas-9) system to silence the Trpv1 gene in ARC POMC neurons.We crossed POMC-Cre with floxedstop CRISPR/Cas-9-eGFP mice to generate POMC-Cre;;CRISPR/Cas-9-eGFP mice (Fig 4A).We injected Trpv1 single guide RNA (sgRNA) into the ARC of these mice and then examined whether ARC POMC neurons still respond to capsaicin.Immunohistochemical experiments with an anti-c-fos protein antibody revealed that local injection of capsaicin induced c-fos protein expression in ARC POMC neurons in control mice, whereas mice injected with Trpv1 sgRNA to ARC POMC neurons showed no c-fos protein expression in response to capsaicin (Fig 4B , 4C and 4D).Moreover, the mean amplitude of temperature-sensitive currents in Given that the anorexigenic effect of capsaicin was abolished by the melanocortin 3 and 4 receptor (MC3/4R) antagonist, we expected that administration of capsaicin would not reduce food intake in mice silencing the Trpv1 gene in ARC POMC neurons.Indeed, capsaicin had no effect on feeding in this animal model, whereas POMC-Cre mice injected with Trpv1 sgRNA still showed a reduction in food intake following capsaicin treatment (Fig 4E and 4F).Our results appeared to be inconsistent with prior studies showing that stimulation of the whole population of ARC POMC neurons failed to decrease food intake [35,36,38].We thus investigated whether activating TRPV1-like receptor-expressing ARC POMC neurons decreases food intake.To selectively stimulate TRPV1-like receptor-expressing POMC neurons in the ARC, we generated an AAV5 carrying a Cre-dependent channelrhodopsin-2 (ChR2)-enhanced yellow fluorescent protein (eYFP) expression cassette under the control of the two neuronal POMC enhancers (nPE1 and nPE2) (AAV5-nPE-DIO-ChR2-YFP, Fig 5A).These enhancers have been shown to drive reporter gene expression to ARC POMC neurons [39].We injected this viral vector into the ARC of TRPV1-Cre mice with an expectation that only TRPV1-like receptor-expressing POMC neurons, among the entire ARC POMC neuronal population, would express ChR2.Eight weeks post viral injections, we found expression of the reporter protein in POMC neurons in the ARC (Fig 5B).Importantly, optogenetic stimulation of these neurons at 10 Hz for 1 h significantly reduced food intake (Fig 5C and 5D).Our results support the interpretation that, contrary to the prior findings [35,36,38], selective stimulation of a specific subset of ARC POMC neurons (i.e., TRPV1-like receptor-expressing) is sufficient to decrease food intake.

A rise in body temperature elevates ARC temperature
A small elevation in temperature within the physiological range (from 37 ˚C to 38 ˚C) activated TRPV1-like receptors in ARC POMC neurons in hypothalamic slices.We asked whether ARC temperature undergoes such elevations under normal physiological conditions.It has been shown that physical exercise raises not only body but also brain temperature [40].Thus, we measured both ARC and body temperature during a 40-min bout of treadmill running in trained mice (Fig 6A and 6B and S1 Movie).This exercise rapidly raised body as well as ARC temperature (Fig 6B and 6C).ARC temperature reached a plateau after 20 min exercise and remained high for more than 1 h.
We sought to determine if this rise in ARC temperature is associated with reduced feeding through activation of TRPV1-like receptors in ARC POMC neurons.Mice were given liquid food following the completion of the exercise bout to measure acute changes in food intake (Fig 6D).We found that exercised mice consumed significantly less food than did non-exercised mice (Fig 6D).Importantly, the anorexigenic effect of exercise was abolished by administration of the MC3/4R antagonist SHU9119 (Fig 6E and S2 Table), suggesting that exercise- induced rise in ARC temperature activates ARC POMC neurons.To further determine this possibility, we measured food intake in mice silencing the Trpv1 gene in ARC POMC neurons after exercise.We found that, although mice injected with control sgRNA showed a decrease in food intake (Fig 6F ), exercise had no effect on food intake in mice silencing the Trpv1 gene in ARC POMC neurons (Fig 6G and S2 Table).Hence, our results propose that elevated hypothalamic temperature can activate temperature-sensitive TRPV1-like receptors in ARC POMC neurons, which may result in a reduction in food intake.

Discussion
All the results from neurophysiologic, molecular, genetic, and immunohistochemical studies provided evidence that ARC POMC neurons expressed functional TRPV1 channels.ARC POMC neurons were sensitive to capsaicin but activated by physiological warm temperature around 37 ˚C, suggesting that ARC POMC neurons may express heteromeric rather than homomeric TRPV1 receptors.In fact, a subset of ARC POMC neurons had Trpv3 and Trpv4 transcripts.Pharmacological and optogenetic stimulation of TRPV1-like receptor-expressing ARC POMC neurons reduced food intake.This anorexigenic effect was blocked by an MC3/ 4R antagonist.Importantly, a transient increase in body temperature was rapidly transmitted to the ARC.This elevation of hypothalamic temperature was enough to reduce food intake, although temperature-independent effects might explain these anorexigenic effects.Mice silencing the Trpv1 gene exclusively in ARC POMC neurons showed no effect of increased hypothalamic temperature on food intake.Therefore, our results support the interpretation that ARC POMC neurons have the ability to respond not only to hormones and nutrients but also to changes in body temperature.
At first glance, our present TRPV1 receptor expression results may appear to be inconsistent with those from a prior study describing the lack of Trpv1 mRNA expression in ARC POMC neurons [17].We think this is likely due to methodological differences in the temperatures applied during transcriptome assessments.In the study of cell type-specific transcriptomics, POMC-GFP cells were sorted at room temperature [17].Importantly, such subphysiological temperatures reducedTrpv1 mRNA expression [41].Thus, transcriptomic expression of TRPV1 receptors would be expected to be greatly down-regulated under such low temperature conditions; indeed, in our hands we found that low temperature treatment (i.e., 24 ˚C) strongly decreased Trpv1 mRNA expression in ARC POMC neurons.We did find robust Trpv1 mRNA in ARC POMC neurons, but only when we incubated at a temperature in the physiological range.
The use of multiple independent techniques, including patch-clamp recording, single-cell RT-PCR, immunohistochemistry, western blotting, gene knockdown with shRNA, CRISPR/ Cas-9-mediated gene knockdown, TRPV1 KOs, and TRPV1-Cre mice, provided converging lines of evidence for the expression of TRPV1 receptor subunits in ARC POMC neurons.However, it appears that there was a discrepancy between our results and the prior findings  described in the literature [24,25].In particular, ARC POMC neurons were depolarized by warm temperature ˚C).It is unlikely that this depolarization was mediated via of homomeric TRPV1 receptors.In fact, recombinant TRPV1 receptors were activated at much higher temperatures (>42 ˚C) [4,23].These findings suggest that warm temperatureelicited depolarization may be due to the activation of other than TRPV1 channels, such as TRPV3 and TRPV4 subunits, with a low threshold temperature [5,6,25,42,43].Although we cannot rule out this possibility, the sensitivity to capsaicin suggests that ARC POMC neurons contained TRPV1 subunits that have a specific binding site for capsaicin [20][21][22].
It has been documented that TRPV channels can assemble into heteromeric complexes [23,29,44].Heteromeric TRPV channels containing TRPV1 subunits are always sensitive to capsaicin, but have altered threshold temperature [23,44].For instance, heteromeric TRPV1/ TRPV3 channels are activated at lower temperatures compared to homomeric TRPV1 channels in human embryonic kidney 293 cells (HEK293) cells (>34 ˚C versus >38 ˚C) [23], suggesting that ARC POMC neurons may express heteromeric TRPV channels.In our preparations, Trpv3 and Trpv4 mRNA were also detected along with Trpv1 transcripts in ARC POMC neurons.Interestingly, the neurons in the supraoptic and paraventricular nucleus expressing TRPV1 and TRPV4 receptor subunits were excited by warm temperature [13,16,45].The neurons in these areas in TRPV1 KO mice still responded to elevated temperature, although the amplitude of temperature-elicited currents was greatly reduced [13,16].In line with these findings, we found that knockdown of the Trpv3 and Trpv4 genes in the ARC altered the thermal sensitivity.Hence, it seems likely that TRPV receptors in the hypothalamus generally form heteromeric rather than homomeric complexes that have capsaicin-sensitive TRPV1 subunits (TRPV1-like).
Despite the well-known anorexigenic effects of α-MSH [46], stimulation of the entire ARC POMC population did not produce the canonical reduction in food intake [35,36,38].The lack of effects of selective stimulation of ARC POMC neurons would be due in part to POMC neuron heterogeneity.A large and growing body of literature supports the idea that ARC POMC neurons are neurochemically, anatomically, and functionally heterogeneous [19,31,[47][48][49][50][51].For example, activation of ARC POMC neurons by the cannabinoid receptor type 1 agonist has been shown to promote food intake through the release of opioids [31].In our preparations, optogenetic stimulation of TRPV1-expressing POMC neurons clearly reduced food intake.These findings suggest that neurochemically and/or neuroanatomically distinct subpopulations of ARC POMC neurons play distinct roles in feeding behavior.In fact, while both TRPV1 and cannabinoid type 1 receptor (CB1) agonists depolarize POMC neurons [31], the physiological consequences of ARC POMC neuron activation by TRPV1 versus CB1 agonists diverged in opposite directions.Hence, it is plausible that the net balance between anorexigenic melanocortin POMC and orexigenic opioid POMC neurons would be finely tuned by differential responses to nutrient-related molecules and signaling, as described in our prior study [19].Interestingly, long-but not short-term stimulation of ARC POMC neurons effectively decreased food intake and this anorexigenic effect was completely absent in A y mice [36], suggesting that long-term stimulation of the entire population of ARC POMC neurons could indeed shift the net balance toward the anorexigenic melanocortin outflow.Administration of capsaicin reduced food intake and prevented body weight gain in rodents and humans [52,53].TRPV1 KO mice gained more body weight than did control animals in adulthood and when fed a high-fat diet [54].Our present findings provided possible explanations as to how vanilloids contribute to the regulation of feeding behavior.In other words, capsaicin and its analog excite TRPV1-like receptor-expressing POMC neurons, resulting in lowered food intake.Importantly, TRPV1-like receptors in ARC POMC neurons were activated not only by vanilloids but also by small, physiological elevations in temperature that can also easily be induced by a brief bout of treadmill exercise.Hence, TRPV1-like receptorexpressing POMC neurons could function as a transducer of thermal stimuli.Like hormones and nutrients, temperature-induced depolarization of ARC POMC neurons drove anorexigenic, melanocortinergic neuronal outflow in our preparations.In the early 1960s, feedingrelated activity has been shown to result in an "immediate" increase in brain temperature (within less than 1 min) [55], suggesting that diet, like physical exercise, is able to increase hypothalamic temperature.As capsaicin increases satiety and fullness in humans as well [56], it is plausible that elevated hypothalamic temperature could promote satiety and thereby reduce food intake through activation of TRPV1-like receptors in ARC POMC neurons.
In summary, we identify a melanocortinergic circuit beginning in ARC POMC neurons that links acute elevations in peripheral and brain temperature with acute reductions in food intake via activation of TRPV1-like receptors.Such temperature-sensing mechanisms are ideally situated to mediate the regulation of energy intake and expenditure.

Ethics statement
All mouse care and experimental procedures were approved by the Institutional Animal Care Research Advisory Committee of the Albert Einstein College of Medicine (Protocol number: 20150203) and were performed in accordance with the guidelines described in the NIH guide for the care and use of laboratory animals.Stereotaxic surgery and viral injections were performed under isoflurane anesthesia.

Slice preparation and electrophysiological recordings
Transverse brain slices were prepared as described in our previous studies [19].Slices were equilibrated with an oxygenated artificial cerebrospinal fluid (aCSF) for >1 h at 32 ˚C before transfer to the recording chamber.The slices were continuously superfused with aCSF at a rate of 1.5 mL/min containing the following (in mM): 113 NaCl, 3 KCl, 1 NaH 2 PO 4 , 26 NaHCO 3 , 2.5 CaCl 2 , 1 MgCl 2 , and 5 glucose in 95% O 2 /5% CO 2 .All neurophysiological recordings were obtained from bright fluorescent green neurons present in POMC-eGFP mice, except as noted, such as POMC-eGFP neurons in POMC-Cre;; CRISPR/Cas9 mice.The internal solution contained the following (in mM): 130 K-gluconate, 5 CaCl 2 , 10 EGTA, 10 HEPES, 2 MgATP, 0.5 Na 2 GTP, and 10 phosphocreatine.Membrane potentials were recorded in the presence of ionotropic glutamate and GABA A receptor antagonists such as CNQX and picrotoxin (10 μM and 100 μM, respectively) with a Multiclamp 700B.Membrane currents were recorded in the presence of TTX (500 nM) and synaptic blockers.Electrophysiological signals were low-pass filtered at 2-5 kHz, stored on a PC, and analyzed offline with pClamp 10 software (Molecular devices).For each recording, the membrane potentials measured from every 30 s were taken as 1 data point.A total of 10 data points before and after the application of drugs were compared using the unpaired t test.

Stereotaxic surgery, drug infusion, and measurement of food intake
Mice were maintained under isoflurane anesthesia and placed in a stereotaxic apparatus.Under aseptic conditions, sterile custom guide cannulas were stereotaxically implanted into the ARC (AP, −1.2 mm; ML, 0; DV, −5.5 mm) for local injection of drugs (1 μL of 100 nM capsaicin, Sigma-Aldrich, cat # M2028 and AMG9810, Tocris bioscience, cat # 2316) and into the lateral ventricle (AP, +0.85 mm; ML, +0.5 mm; DV, −3.0 mm) for injection of SHU9119 (Tocris bioscience, cat # 3420) or naloxone (Tocris bioscience, cat # 0599).Animals were allowed at least a week to recover from surgery.To measure food intake, animals were separated individually in single cages for 5 d and fasted for 6 h (from 1 PM to 7 PM) before drug infusion.SHU9119 (1 μL of 5 μM) or naloxone (1 μL of 1 mM) was intracerebroventricularly administrated 30 min prior to microinjection of capsaicin or saline.Capsaicin and/or AMG9810 were directly infused into the ARC at 6 PM.Food intake was measured from 7 PM to 11 PM and from 11 PM to 7 AM.The mice used for all of the experiments were randomly assigned to a treatment group.For animal studies, sample sizes were estimated according to our previous experience [19].Single housed mice with a 20% weight loss were excluded from experiments and analyses.For optogenetic experiments, a mono fiber-optic cannula of 230 μm diameter (Doric Lenses, Inc.) was implanted into the ARC and coupled to a 473-nm DPSS laser (Laserglow Technologies).

Measurement of body and ARC temperatures, and treadmill exercise
To directly measure both body and hypothalamic temperatures, while mice were running on a treadmill, a wire thermoprobe (IT-24P, Physitemp Instruments) was placed into the ARC (AP, −1.2 mm; ML, 0; DV, −5.5 mm) through a guide cannula and another wire thermoprobe (RET-4, Physitemp Instruments) was inserted into the rectum.
For treadmill exercise (Eco-6M, Columbus Instruments), mice were trained to run on the treadmill for 2 wk.The mice were always warmed up for 5 min at 7 m/min and the speed was slowly ramped up to 18 m/min.To measure food intake following exercise, animals were fasted from 1 PM to 7 PM and given liquid food (Ensure original nutrition shake, Abbott).If at any point during the experiment a mouse became exhausted, the shock grid for that lane was turned off and the mouse allowed to rest.

Statistics
Statistical analyses were performed using an unpaired t test or one-way ANOVA with Tukey multiple comparison test (GraphPAD 4.03 and Origin 8.5).Data were considered significantly different when the p-value was <0.05.All statistical results are given as mean ± SEM.Arrows show mRNA regions detected by primer sets (outer, blue; inner, red).Gel images show lack of Trpv1 mRNA in ARC POMC neurons from slices kept at 24 ˚C.M: PCR marker (Trpv1, 248 bp; Gad2, 205 bp; 18S rRNA, 121 bp).P: positive control single-cell samples were collected from hypothalamic slices kept at 24 ˚C and 36 ˚C for 3 h.The initial RT reaction was conducted after pressure ejection of the single-cell samples into freshly prepared RT mix A solution (20 U of RNase OUT, 300 ng of random primers, 0.5% NP-40, and RNase free water).Samples were sonicated in a total volume of 10 μL at 4 ˚C for 5 min and then incubated for 3 min at 65 ˚C before addition of 10 μL RT mix B (500 μM dNTP, 1 × RT buffer, 5 mM MgCl 2 , 10 mM DTT, and 200 U of Superscript IIl).The tubes were incubated at 25 ˚C for 5 min, at 42 ˚C for 1 h, and at 65 ˚C for 10 min.Three rounds of amplification were done for the detection of Trpv1 transcripts (two rounds with the outer primer set and one round with the inner primer set) and two rounds of amplification for the analysis of Gad2 transcripts.The primers used for qPCR were the following: Trpv1 outer primer, f5 0catgctcattgctctcatgg-3 0 and r5 0 -aaccagggcaaagttcttcc-3 0 , Trpv1 inner primer, f5 0 -catgggcgagactgtcaac-3 0 and r5 0 -ctgggtcctcgttgatgatg-3 0 ; Gad2 outer primer, f5 0 -ggcgatggaatcttttctcct-3 0 , and Gad2 inner primer f5 0 -cgcactctggaagacaatga-3 0 and r5 0 -cgaggcgttcgatttcttcaa-3 0 .ARC, arcuate nucleus of the hypothalamus; dNTP, deoxynucleotide; DTT, dithiothreitol; Gad2, glutamate decarboxylase 2; MgCl
. Membrane potentials were recorded from ARC POMC-enhanced green fluorescent protein (eGFP) neurons in hypothalamic slices from POMC-eGFP mice in the presence of synaptic blockers (6-cyano-7-nitroquinoxaline-2,3-dione [CNQX], 10 μM, and picrotoxin, 100 μM).Treatment with capsaicin (100 nM) significantly depolarized and triggered action potentials in ARC POMC neurons (S1A and S1B Fig).In addition, treatment with another potent and selective TRPV1 receptor agonist RTX [27] (100 nM) also depolarized ARC POMC, but not neuropeptide Y (NPY) neurons (S1C, S1D, S5B and S5C Figs).Capsaicin-induced depolarization was completely blocked by treatment with the TRPV1 receptor antagonist AMG9810 (10 μM) [28] (S1E and S1F Fig).In addition, the TRPV1 receptorinduced depolarization was still present in the presence of tetrodotoxin (TTX) (S1G and S1H Fig).These pharmacological results support the existence of functional capsaicin-sensitive TRPV1 channels expressed by ARC POMC neurons.We next performed single-cell reverse transcription-polymerase chain reaction (RT-PCR) and immunohistochemistry to identify the expression of capsaicin-sensitive TRPV1 receptors in ARC POMC neurons.Single-cell RT-PCR analysis of ARC POMC neurons revealed that approximately two thirds of ARC POMC neurons examined contained Trpv1 messenger RNA (mRNA) (Fig 1A, n = 21 out of 34 neurons).In addition to the expression of Trpv1 transcripts, some ARC POMC neurons expressed Trpv3 and Trpv4 mRNA (S2 Fig).The prior transcriptome study in which GFP-tagged POMC neurons were isolated at room temperature (24 ˚C) showed the lack of Trpv1 mRNA in ARC POMC neurons [17].We further investigated Trpv1 mRNA expression in ARC POMC neurons from hypothalamic slices kept at room temperature.We found that this low temperature treatment strongly down-regulated Trpv1 mRNA levels, although the number of ARC POMC neurons expressing glutamate decarboxylase 2 (Gad2) mRNA was similar to that observed at 36 ˚C (S3 Fig).Thus, methodological differences in the temperatures at which the tissue was isolated may account for prior failures to identify Trpv1 transcripts in ARC POMC neurons.Immunostaining of ARC POMC neurons with an anti-TRPV1 receptor antibody showed TRPV1 receptor-positive cells in the ARC of control animals (Fig 1B and S4A Fig).However, no labeled cells were observed in TRPV1 knock-out (KO) mice (Fig 1B).Additionally, western blot analysis showed a TRPV1-positive signal in lysates from the ARC of control but not TRPV1 KO mice (Fig 1C).Importantly, immunostaining, quantitative polymerase chain reaction (qPCR), and western blot analysis showed no expression of TRPV1 receptors around the ARC (S4B, S4C and S4D Fig).All these results provide the specificity of the anti-TRPV1 antibodies that we used in this study.Double immunolabeling experiments with an anti-POMC antibody revealed that most TRPV1 receptor-positive cells in the ARC were POMC-containing neurons (Fig 1D, 96.3 ± 0.3%; n = 3 mice), consistent with our neurophysiological results.In contrast, very few neuropeptide Y-expressing neurons in the ARC expressed TRPV1 receptors (S5A Fig; 1.6 ± 0.7% of TRPV1-postive neurons; n = 3 mice).

Fig 2 .
Fig 2. Activation of TRPV1 receptors in the ARC reduces food intake.(A) Schematic drawing of our experimental configuration (left panel).Saline and drug were injected into the ARC of POMC-GFP mice (right panel).(B) Scheme depicting food intake measurement.(C) Pooled data from 17 mice of food intake following local microinjection of capsaicin into the ARC.Ã p < 0.05, ÃÃÃ p < 0.001.(D) Pooled data from 7 mice showing blockade of the effect of capsaicin with AMG9810 (3 μg/μL).ÃÃÃ p < 0.001, Ã p < 0.05.(E) Schematic drawing of our experimental

Fig 4 .
Fig 4. Knockdown of the Trpv1 gene in ARC POMC neurons blocks the anorexigenic effect of capsaicin.(A) Images of fluorescence confocal microscopy showing GFP-positive POMC neurons in the ARC of POMC-Cre::CRISPR-Cas-9-eGFP mice.Scale bar: 50 μm.(B) Images of fluorescence confocal microscopy showing that local injection of capsaicin induced expression of c-fos protein in POMC neurons in the ARC.(C) Images of fluorescence confocal microscopy showing that local injection of capsaicin did not induce c-fos protein expression in mice lacking TRPV1 receptors in ARC POMC neurons.Scale bar: 50 μm.(D) Pooled data show the percentage of POMC neurons expressing c-fos protein in control mice and mice lacking TRPV1 receptors in ARC POMC neurons following treatment with saline and capsaicin (n = 5, 5, and 3 animals, respectively).(E and F)

Fig 6 .
Fig 6.Exercise increases body and ARC temperatures.(A) Scheme depicting the mouse treadmill protocol.Mice were always warmed up for 5 min at 7 m/min and the speed was slowly ramped up to 18 m/min.(B and C) Pooled data showing changes in body and ARC temperature during treadmill running (gray line, warming up for 5 min and black line, running for 35 min [B]).Plot showing mean temperature at 0, 5, and 40 min exercise (at 40 min, mean body temperature, 38.9 ± 0.1 ˚C; mean ARC temperature, 38.6 ± 0.2 ˚C).(D) Pooled data from 9 mice showing the effect of exercise on liquid food intake for 1 h ( Ã p < 0.05, ÃÃÃ p < 0.001).(E) Pooled data from 9 mice showing the effect of exercise on food intake for 1 h with an MC3/4R antagonist.Exercise-induced anorexigenic effect was abolished by local injection of SHU9001.(F) Pooled data from 9 mice showing the effect of exercise on food intake for 1 h in mice injected with control sgRNA.Exercise-induced anorexigenic effect is absent in mice silencing the Trpv1 gene in ARC POMC neurons.Ã p < 0.05, ÃÃ p < 0.01, ÃÃÃ p < 0.001.(G) Pooled data from 9 mice showing the effect of exercise on food intake for 1 h in mice silencing the Trpv1 gene in POMC neurons.Exercise no longer reduced food intake in mice lacking the Trpv1 gene in POMC neurons.Data are shown as mean ± SEM.ARC, arcuate nucleus of the hypothalamus; Ex, exercise; MC3/4R, melanocortin 3 and 4 receptor; POMC, proopiomelanocortin; sgRNA, single guide RNA; Trpv1, transient receptor potential vanilloid 1 receptor; Veh, vehicle.https://doi.org/10.1371/journal.pbio.2004399.g006 resiniferatoxin; TRPV1, transient receptor potential vanilloid 1 receptor; TTX, tetrodotoxin.(PDF) S2 Fig.Expression of Trpv1, Trpv3, and Trpv4 transcripts in ARC POMC neurons.ARC, arcuate nucleus of the hypothalamus; lane, cell number; M, PCR marker; P, positive control; POMC, proopiomelanocortin; Trpv1, 3, 4, transient receptor potential vanilloid 1, 3, 4 receptors.(PDF) S3 Fig. Expression of Trpv1 mRNA in ARC POMC neurons under different conditions in vitro.Upper panel: simplified structure of the mRNA (numbered boxes show exons of the Trpv1 gene) encoding Trpv1.