Neurotensin induces hypothermia by activating both neuronal neurotensin receptor 1 and astrocytic neurotensin receptor 2 in the median preoptic nucleus

Abstract

Neurotensin (NTS) is a neuropeptide acting as a neuromodulator in the brain and is a very potent hypothermic agent. However, the cellular mechanisms of actions are not fully understood. Here we report that NTS increases the firing rate of preoptic GABAergic neurons by activating both neurotensin receptor 1 (NTSR1) and neurotensin receptor 2 (NTSR2), expressed by neurons and astrocytes, respectively. Downstream of NTSR1 the neuropeptide activated an inward current, calcium release from intracellular stores and, postsynaptically, increased frequency and amplitude of inhibitory synaptic events. NTSR2 activation in astrocytes resulted in increased excitatory input in preoptic GABAergic neurons, an effect which was dependent upon the activation of P2X4 receptors. We also found that neuromedin N acted as a selective agonist at the NTSR1. Surprisingly, activation of both NTSR1 and NTSR2 in the median preoptic nucleus was required for activating a full hypothermic response.

Graphical Abstract

1. Introduction

Neurotensin (NTS) is a 13 aminoacid peptide found in the central nervous system (CNS) as well as in the gastrointestinal tract (Vincent et al., 1999). It exerts potent effects when infused centrally including hypothermia (Mason et al., 1982; Morley et al., 1982), analgesia and antipsychotic like effects (St-Gelais et al., 2006). NTS-producing neurons and their projections are widely distributed in the brain, including the anterior hypothalamus, which explains the wide variety of effects of this peptide (Schroeder and Leinninger, 2018). The cellular mechanisms and neuronal networks modulated by NTS to induce hypothermia are not known.

NTS exerts its actions by binding to two G-protein-coupled receptors, NTSR1 and NTSR2, which are widely expressed in the brain, including regions involved in the control of thermoregulation (Schroeder and Leinninger, 2018). NTSR1 mRNA was detected in many hypothalamic regions, including the preoptic, anterior, periventricular, ventromedial and arcuate nuclei (Alexander and Leeman, 1998; Nicot et al., 1994). The highest levels of expression were found in the basal ganglia nuclei, and in the dopaminergic neurons of the ventral tegmental area and the substantia nigra (Alexander and Leeman, 1998; Nicot et al., 1994; Woodworth et al., 2018). NTSR1 is a high affinity receptor and is coupled with the phospholipase C and inositol phosphate (IP) signaling cascade (St-Gelais et al., 2006). NTSR2 is a low affinity receptor, however the signaling pathways activated in native brain tissue are less studied. In cell lines expressing NTSR2 calcium release from intracellular stores and IP formation has been reported in response to NTS (St-Gelais et al., 2006).

Numerous studies have revealed that GABAergic neurons in the median preoptic nucleus (MnPO) play an important role in the control of thermoregulatory networks that comprise also neurons of the rostral raphe pallidus and dorsomedial hypothalamus (Morrison, 2018). The firing activity of preoptic GABAergic neurons provides an inhibitory tone on downstream neurons controlling thermogenesis. A decrease in the firing rate of preoptic GABAergic neurons can account for the increased thermogenesis associated with fever (Morrison and Madden, 2014) or other hyperthermic responses (Lundius et al., 2010; Sanchez-Alavez et al., 2010). In this study we have investigated the effects of NTS on identified preoptic GABAergic neurons and the receptor subtypes involved as well as their role in the potent hypothermia induced by NTS.

2. Materials and methods

All animal work was conducted in accordance with the National Institute of Health Guide for the care and use of Laboratory animals (NIH publication No. 8023, revised 1978). The procedures were approved by the Institutional Animal Care and Use Committee of the Scintillon Institute. The standards are set forth by American Association for the Accreditation of Laboratory Animal Care (AAALAC) standards and the regulations set forth in the Animal Welfare Act. Efforts were made to minimize the number of animals used and their suffering.

2.1. Telemetry and POA injections

4–6 months old male wild-type C57BL/6 mice (Jackson Laboratory, Bar Harbor, ME, USA) were anesthetized with isoflorane (induction 3–5%, maintenance 1–1.5%) and surgically implanted with radio telemetry devices (Anipill, BodyCap, Hérouville Saint-Clair, France) into the peritoneal cavity for core body temperature (CBT) measurement. For brain injections, mice were first subject to stereotaxic placement of a guide cannula (26 Ga, 10 mm length) as previously described (Lundius et al., 2010) (Suppl Fig 5).

Coordinates for cannula (27 ga. 16 mm length) implants in the median preoptic nucleus (MnPO) were: 0.38 mm from Bregma and ventral 4.6 mm (Paxinos and Franklin, 2001). The ambient temperature was maintained at ~28 ± 0.5 °C in a 12:12-h light-dark cycle controlled room (lights on 8:00 am). NTSR1 and NTSR2 ligands were injected into the MnPO. All substances injected were dissolved in sterile aCSF. Mice were handled for at least three days before the injection for 5 minutes every day for habituation. On the day of injections, mice were held and the injector (cannulae 33 Ga, 17 mm length) was placed inside the cannula. The injector was connected to a microsyringe (0.25 μL). The injected volume was 100 nL (rate 0.1 μL/min). After this procedure the animal was returned to the home cage. Injections were always performed at 12 pm local time, during the “subjective light period”. In experiments that employed 2 injections, the first injection was performed at 10 am and the second at 2pm.

2.2. Slice Preparation

Coronal tissue slices containing the median preoptic nucleus (MnPO) were prepared from GAD65-GFP male mice (42–70 days old) as previously described (Lundius et al., 2010). This transgenic mouse line expresses enhanced green fluorescent protein (eGFP) under the control of the regulatory region of mouse glutamic acid decarboxylase (GAD) 65 gene (Bali et al., 2005). The mice were a kind gift from Dr. Gabor Szabo (Hungarian Academy of Sciences, Budapest, Hungary). The slice used in our recordings corresponded to the sections located from 0.5 mm to 0.25 mm from Bregma in the mouse brain atlas (Paxinos and Franklin, 2001).

2.3. Dissociated Preoptic Neurons from Slices

The MnPO was punched out of a brain slice and incubated in Hibernate A (Invitrogen, Temecula, CA) and papain (Worthington, Lakewood, NJ) (1 mg/ml) for 10 min at room temperature. After washing out the enzyme with Hibernate-A the cells were dissociated by gentle trituration with a fire-polished Pasteur pipette. The cell suspension was pelleted (1000 g for 2 min) and resuspended in Neurobasal medium (supplemented with 2% B27, 0.5 mM Glutamax and 10 mg/ml gentamycin) and then plated on poly-D-lysine/laminin coated coverslips (Biocoat, BD Biosciences, Bedford, MA). Cells were kept in culture for 5–10 days before being used for experiments.

2.4. Whole-cell patch-clamp recording

The aCSF contained (in mM) the following: 130 NaCl, 3.5 KCl, 1.25 NaH₂PO₄, 24 NaHCO₃, 2 CaCl₂, 1 MgSO₄, and 10 glucose, osmolarity of 300–305 mOsm, equilibrated with 95% O₂ and 5% CO₂, pH 7.4. Other salts and agents were added to this medium. A set of experiments was carried out in Ca²⁺-free aCSF which contained (in mM) 130 NaCl, 3.5 KCl, 1.25 NaH₂PO₄, 24 NaHCO₃, 0.5 EGTA, 3 MgSO₄, and 10 glucose. In a set of experiments the extracellular solution was exchanged with a Na⁺-free buffer (“NMDG external solution”) containing 155 N-methyl-D-glucamine, 3.5 KCl, 2 CaCl₂, 1 MgSO₄, 10 glucose, and 10 HEPES (pH 7.4, adjusted with HCl). Whole-cell recordings were carried out using a K⁺ pipette solution containing (in mM) 130 K-gluconate, 5 KCl, 10 HEPES, 2 MgCl₂, 0.5 EGTA, 2 ATP and 1 GTP (pH 7.3) was used in all experiments. The electrode resistance after back-filling was 2–4 MΩ. All voltages were corrected for the liquid junction potential (−13 mV and −7 mV, respectively). Data were acquired with a MultiClamp 700B amplifier (Molecular Devices, Sunnyvale, CA, USA) digitized using a Digidata 1550 interface and the Pclamp10.6 software package. The sampling rate for the continuous recordings of spontaneous activity was 50 kHz. The recording chamber was constantly perfused with extracellular solution (2–3 mL·min⁻¹). The agonists were applied locally using a perfusion pencil system (tip diameter 100 μm, Automate Scientific) driven by gravity, while the antagonists were bath-applied. The temperature of the external solution was controlled with a TC-344B temperature controller and an inline heater (Warner Instruments, Hamden, CT, USA) and was maintained at 36–37°C.

Synaptic activity was quantified and analyzed statistically as described previously (Tabarean, 2012). Briefly, synaptic events were detected and analyzed (amplitude, kinetics, frequency) offline using a peak detection program (Mini Analysis program, Synaptosoft, Decatur, NJ, USA). Events were detected from randomly selected recording stretches of 2 min before and during incubation with pharmacological agent. Statistical significance of the cumulative distributions of the measured parameters (inter-event interval, amplitude, rise time, time constant of decay) were assessed with the Kolmogorov-Smirnov two-sample test (K-S test, P<0.05) using the Mini Analysis program. The averages for the measured parameters (frequency, amplitude, rise time, time constant of decay) for each experiment were obtained using the Mini Analysis program. Event frequency was calculated by dividing the number of events by the duration (in seconds) of the analyzed recording stretch.

2.4. Ca²⁺ imaging

Fura-2 fluorescence signals were acquired with a CCD camera (Hamamatsu ORCA-ER) connected to its frame grabber driven by Slidebook software (Intelligent Imaging Innovations, Denver, CO, USA). An ultra-high-speed wavelength switcher Lambda DG-4 (Sutter Instruments, Novato, CA, USA) provided alternating excitation for ratiometric Fura-2 measurements. The filters were 340HT15 and 380HT15. The illumination source was a standard xenon lamp. The sampling frequency of 0.2 Hz was sufficiently fast to capture responses to NTS and other agonists studied. At this excitation frequency, photobleaching and phototoxicity were minimal. Fura-2AM loading and data acquisition were carried out as described in our previous studies.

2.5. Chemicals

All agonists and antagonists were purchased from Tocris (Ellisville, MO, USA), except ML314 and levocabastine which were from Cayman Chemical (Ann Arbor, MI, USA). All the other chemicals were from Millipore Sigma (Carlsbad, CA, USA). The PubChem ID, targets and selectivity are listed below (Table2).

Table 2.

Primers used for single cell RT/PCR.

Compound	PubChem CID	Target
neurotensin	25077406	NTSR1,2 agonist
neuromedin N	9940301	NTSR1,2 agonist
TC NTR1 17	44157038	NTSR1 agonist
SR48692	119192	NTSR1 antagonist; NTSR2 agonist
NTRC824	101873359	NTSR2 antagonist
levocabastine	54385	H1 histamine antagonist; NTSR2 agonist
ML314	53245590	NTSR1 agonist
CNQX	3721046	AMPAR antagonist
AP-5	1216	NMDAR antagonist
picrotoxin	31304	GABAAR antagonist
PPADS	4881	P2X antagonist
Ro-3	11289644	P2X2/3 antagonist
A438079	11673921	P2X7 antagonist
PSB-12062	2320735	P2X4 antagonist

2.6. Cell Harvesting, Reverse Transcription and PCR

MnPO neurons in slices were patch-clamped and then harvested into the patch pipette by applying negative pressure as previously described (Lundius et al., 2010). The content of the pipette was expelled in a PCR tube. dNTPs (0.5 mM), 50 ng random primers (Invitrogen) and H₂O were added to each cell to a volume of 16 μl. The samples were incubated at 65°C for 5 min and then put on ice for 3 min. First strand buffer (Invitrogen), DTT (5 mM, Invitrogen), RNaseOUT (40 U, Invitrogen) and SuperScriptIII (200 U, Invitrogen) were added to each sample to a volume of 20 μl followed by incubation at room temperature for 5 min, at 50°C for 50 min and then at 75°C for 15 min. After reverse transcription samples were immediately put on ice. 1 μl of RNAse H was added to samples and kept at 37°C for 20 min. PCR assays were carried out using the pairs of primers listed in Table 1.

Table 1.

Primers used for single cell RT/PCR.

Neurotensin receptor 1 (NTSR1)	F: “ACACCTTCATGTCCTTCCTGTT”	Bp:400
	R: “GAGACGAGGTTGTAAAGGAT”
Neurotensin receptor 2 (NTSR2)	F: “GCCTGGTGAGACACAAGGAT”	Bp: 248
	R: “TTCACCACACAGGGAACTGA”

2.7. RNAscope in-situ hybridization

To detect single mRNA molecules, RNAscope was performed on fixed cultures of acutely dissociated MnPO neurons and astrocytes from slices. The cells were kept in culture for 7–10 days. In situ hybridization (ISH) was performed according to the protocol of the RNAscope Multiplex Fluorescent Reagent Kit v2 (Cat. No. 320293).

Briefly, coverslips containing MnPO cells were incubated in cold 4% PFA for 30 min and mounted on microscope slides. The cells were then dehydrated in 50%, 70% and 100% ethanol for 5 min each at room temperature (RT). Cells were then rehydrated in 70%, 50% ethanol and PBS. Cell were treated with RNAscope hydrogen peroxide for 10 min washed 5 times with distilled water and, for antigen accessibility, treated with Protease III for 10 min at RT and then washed twice in PBS. C2 and C3 probes were diluted in C1 probes at a 1:50 ratio and incubated on the slides for 2hrs at 40 °C. C1, C2 and C3 probes were detected with Opal520, Opal570 and Opal 650, respectively (Perkin Elmer, NEL801001KT). Before mounting the slides, DAPI (Perkin Elmer, REF 323108) was added to label the nuclei. A one-day protocol has been used in all experiments to preserve the quality of the preparations.

2.8. Image acquisition and analysis

Images were collected on an Olympus BX-51 inverted microscope (Olympus, Melville, NY) using a 60× objective. 10 different fields in each coverslip were captured, and all cells in the pictures were analyzed for their fluorescent puncta. Backgrounds were defined as the fluorescence intensity in fields with no cells and subtracted from the entire image.

Cells were identified using the DIC images and the DAPI staining. Neurons or astrocytes were identified using mRNAs probes for Rbfox3/NeuN or Slc32a1, respectively (Opal 570). NTSR1 (Opal 520) and NTSR2 transcripts (Opal 650) were detected with specific RNA probes (ACD Bio). Following ACD Bio’s guidelines for RNAscope data analysis, a cell was considered positive for a probe if it had at least one fluorescent punctum of minimum 0.45 μm². We should note that a vast majority of cells analyzed in the study presented 5 or more puncta for any of the probes used. Only 2 out of 252 NeuN positive neurons presented only 1 punctum, 1 cell presented 2 puncta and the rest presented 4 or more puncta. For all the other probes each positive cell presented 5 or more puncta.

2.9. Statistics

The values reported are presented as mean ± standard deviation (S.D.). Statistical significance of the results pooled from two groups was assessed with t-tests using Prism4 (GraphPad Software). One-way analysis of variance (ANOVA, Kruskal-Wallis) with Tukey’s post hoc test (P<0.05) was used for comparison of multiple groups. Comparisons of cumulative distributions were done with Kolmogorov-Smirnov test (P<0.05). All data collected as time series were compared across time points by one-way ANOVA with repeated measures (P<0.05) (Prism4, GraphPad Software), followed by unpaired t-tests (P<0.05) for comparisons at each time point. The statistical value, the degrees of freedom and the p value are reported in the figure legends or, if data is not presented in a figure, the respective values are reported in the text.

3. Results

3.1. NTS increases the firing activity of preoptic GABAergic neurons by two distinct mechanism in two subgroups of neurons

MnPO GABAergic neurons in acute slices were identified by using the transgenic mouse line GAD65-GFP, which expresses eGFP under the control of the regulatory region of mouse GAD65 gene. Neurotensin applied locally via a perfusion pencil (diameter 100 μm) potently increased the firing rate of 43% (30 out of 70) of MnPO GABAergic neurons studied at all concentrations tested (100 nM, 300 nM, 1 μM and 3 μM) and was without effect in the others. In contrast, NTS increased the firing rate of only 2 out of 37 GFP-negative neurons studied and had no effect in the rest. Among all MnPO GABAergic neurons (n=8 neurons) in which several concentrations were tested, 1 μM NTS induced the maximal effect therefore we have used this concentration to further characterize the excitatory actions of the neuropeptide (Suppl Fig 1).

Based on the pharmacological profile of the NTS responses we grouped the MnPO GABAergic neurons excited by NTS in two groups (Fig 1A, B). Type 1 neurons increased their firing rate by 282 ± 91% (n = 12). The firing rate of the neurons averaged 6.2 ± 1.2 Hz and 16.9 ± 4.1 Hz (n=12) in control and during NTS incubation, respectively. This effect was mimicked by the selective NTSR1 agonist TC NTR1 17 (300 nM) and was completely blocked by the NTSR1 antagonist SR48692 (100 nM) in all neurons studied (n=12) (Fig 1A). In contrast, the NTSR2 antagonist NTRC 824 (200 nM) did not change the excitatory effect activated by NTS. Neuromedin N (1 μM), an endogenous neuropeptide with binding affinity for NTS receptors, also increased the firing rate by a similar amount as NTS (Fig 1C).

Figure 1. Neurotensin (NTS) excites MnPO GABAergic neurons by activating differential mechanisms in two distinct groups of neurons.

A. Example of a type 1 NTS response in a MnPO GABAergic neuron. NTS (1 μM) increased the firing rate of the neuron from 1.3 Hz to 6.1 Hz (upper trace). The NTSR1 agonist TC NTR1 17 (300 nM) and neuromedin N (1 μM) also increased the firing rate of the neuron from 1.1 Hz to 5.6 Hz (2^nd trace from top) and from 1.2 Hz to 5.1 Hz (3^rd trace from top), respectively. In the presence of NTSR2 antagonist NTRC 824 (200 nM) NTS increased the firing rate of the neuron from 1.1 to 5.8 Hz (4^th trace from top). The NTSR1 antagonist SR48692 (100 nM) blocked the excitatory effect of NTS (1 μM) (bottom trace).

B. Example of a type 2 NTS response in a MnPO GABAergic neuron. NTS (1 μM) increased the firing rate of the neuron from 2.9 Hz to 5.7 Hz (upper trace). In contrast, the NTSR1 agonist TC NTR1 17 (300 nM) and neuromedin N (1 μM) decreased the firing rate of the neuron from 2.8 Hz to 1.2 Hz (2^nd trace from top) and from 2.7 Hz to 1.3 Hz (3^rd trace from top), respectively.). In the presence of the NTSR2 antagonist NTRC 824 (200 nM) the excitatory effect of NTS (1 μM) was abolished, instead NTS decreased the firing rate from 2.5 Hz to 1.1 Hz (4th trace from top). In the presence of NTSR1 antagonist SR48692 (100 nM) NTS (1 μM) increased the firing rate of the neuron from 2.8 Hz to 5.8 Hz (bottom trace).

C, D. Bar charts summarizing the effects of neurotensinergic ligands on the firing activity of the two groups of MnPO GABAergic neurons. Summary of the effects of NTS (1 μM), TC NTR1 17 (300 nM), neuromedin N (1 μM) and of the effect of NTS during incubation with the NTSR1 antagonist SR48692 (100 nM) or the NTSR2 antagonist NTRC 824 (200 nM). Bars represent means ± S.D. of the normalized firing rate relative to the control. C. Type 1 neurons. There was a statistically significant difference between groups as determined by one-way ANOVA (F(5,71)=52.59, p=1.1×10⁻¹⁶) followed by Tukey’s test relative to control; ** indicate statistical significance of P<0.01. Data pooled from n=12 neurons in each condition. D. Type 2 neurons. There was a statistically significant difference between groups as determined by one-way ANOVA (F(5,65)=37.82, p=2.2 ×10⁻¹⁰) followed by Tukey’s test relative to control; ** and * indicate statistical significance of P<0.01 and P<0.05, respectively. Data pooled from n=10 neurons in each condition.

Type 2 neuronal excitatory responses were similarly potent (the average firing rate of the neurons was 6.8 ± 1.4 Hz and 16.0 ± 3.1 Hz (n=10) in control and during NTS incubation, respectively) but had a strikingly different pharmacological profile. The increase in firing rate induced by NTS was fully blocked by the NTSR2 antagonist NTRC 824 (200 nM) and was not affected by the NTSR1 antagonist SR48692 (100 nM) in all neurons studied (n=10) (Fig 1B). Furthermore, NTSR1 agonist TC NTR1 17 (300 nM) and neuromedin N (1 μM) decreased the firing rate of the neurons (Fig 1B, D), an effect that was associated with an apparent increase in the frequency of sIPSCs. Overall type 1 and 2 neurons represented ~ 23 and 19%, respectively, of the MnPO GABAergic neurons studied (10 and 12 out of 52 neurons).

To address the possibility of synaptic mechanisms in the two groups, we also studied the effects of NTS in the absence of fast synaptic events, which were blocked by adding CNQX (20 μM), AP-5 (100 μM), and picrotoxin (50 μM) to the extracellular solution. In the presence of the antagonists, NTS increased the firing rate of the type 1 neurons by 255.8 ± 53.9% (n=5) relative to control, value that was not statistically different to the effect in the control (unpaired t-test t(15)=0.93, p=0.37) (Suppl Fig 2A). In contrast, in type 2 neurons NTS had no effect on the firing rate of the neurons when synaptic activity was blocked (Suppl Fig 2B). The firing rate of the neurons averaged 6.7 ± 1.1 Hz and 6.8 ± 0.9 Hz (n=5) in control and during NTS incubation, respectively (paired t-test t(4)=−0.83, p=045). Thus, in type 1 neurons NTS activates an intrinsic mechanism of excitation downstream of NTSR1 while in type 2 neurons the neuropeptide activates synaptic mechanisms downstream of NTSR2.

To further characterize the cellular mechanism involved in the robust excitation of MnPO GABAergic neurons by NTS we have then carried out whole-cell voltage-clamp recordings. We found that in type 1 neurons NTS (1 μM) activated an apparent inward current that averaged 22.8 ± 4.1 pA (n=26), an effect that was fully blocked by NTSR1 antagonist SR48692 (100 nM) but was not affected by incubation with the NTSR2 antagonist NTRC 824 (200 nM) in all neurons studied (n=21 and 18, respectively) (Fig 2A,B). NTSR1 agonist TC NTR1 17 (300 nM) and neuromedin N (1 μM) activated inward currents of similar amplitude (n=13 and 14, respectively) (Fig 2A, B). In contrast, levocabastine, a partial agonist of NTSR2, did not activate an inward current. Taken together these data indicate that NTS activated an inward current in type 1 MnPO GABAergic neurons downstream of NTSR1.

Figure 2. NTS induces an inward current in type 1 neurons by activating NTSR1 receptors.

A. Example of an inward current activated by NTS in a type 1 MnPO GABAergic neuron. The amplitude of the inward current activated by NTS (1 μM) was 21 pA when the neuron was held at −50 mV (upper trace). Note that NTS also induced an increase in the frequency and amplitude of sEPSCs. The NTSR1 antagonist SR48692 (100 nM) blocked the inward current activated by NTS (1 μM) but did not block the effect of NTS on sEPSCs (2^nd trace from top). In contrast, NTSR2 antagonist NTRC 824 (200 nM) did not change the inward current activated by NTS (1 μM) but abolished the effect on EPSCs (3^rd trace from top). The NTSR1 agonist TC NTR1 17 (300 nM) and neuromedin N (1 μM) activated inward currents of 22 pA and 18 pA, respectively (traces 4 and 5 from top). GRP (1 μM) did not activate any current in type 1 MnPO GABAergic neurons (bottom trace).

B. Bar chart summarizing the amplitude of the inward currents activated by NTS (1 μM), NTSR1 agonist TC NTR1 17 (300 nM), neuromedin N (1 μM) and of the effect of NTS during incubation with the NTSR1 antagonist SR48692 (100 nM) or the NTSR2 antagonist NTRC 824 (200 nM). Bars represent means ± S.D. There was a statistically significant difference between groups as determined by one-way ANOVA (F(5,105)=74.55, p=1.1×10⁻¹⁰) followed by Tukey’s test relative to NTS effect; ** indicate statistical significance of P<0.01. Cell numbers studied are indicated in brackets.

C. NTSR1 but not NTSR2 transcripts are present in type 1 MnPO GABAergic neurons. Representative gels from a batch of 16 MnPO GABAergic neurons in which NTS activated an inward current. The expected sizes of the PCR products are (in base pairs) 400 and 248, respectively. Negative (−) control was amplified from a harvested cell without reverse-transcription, and positive control (+) was amplified using 1 ng of hypothalamic mRNA. NTSR1 transcripts were detected in 15 out of 16 recorded neurons, while NTSR2 transcripts were not detected in any neuron.

In 8 out of 26 type 1 neurons studied NTS activated also an increase in the frequency and amplitudes of sEPSCs (Fig1A). This effect was not affected by NTSR1 antagonist SR48692 (100 nM) but was fully blocked by NTSR2 antagonist NTRC 824 (200 nM) (Fig 2A). Nevertheless, since the NTS-induced increase in firing rate persisted in the presence of NTSR2 antagonist (Fig 1A, C) or by blocking all synaptic activity (see above) we concluded that in type1 neurons the increased sEPSC input plays a minor role in the excitation activated by the neuropeptide.

Single cell reverse transcription-PCR (scRT/PCR) analysis was performed for 28 MnPO GABAergic neurons in which NTS activated an inward current. NTSR1 transcripts were detected in 26 out of 28 neurons while NTSR2 transcripts were not detected in any neuron. Fig 2C illustrates such an analysis of a batch of 16 neurons in which NTS activated an inward current.

To understand the nature of the inward current activated by NTS we performed recordings in NMDGCl external solution. When NTS was applied in this solution no inward current was activated (n=15 neurons) (Suppl Fig 3). This observation indicates that Na⁺ is the main ion responsible for the inward current and that Ca²⁺ has low permeability through the respective channels.

In a distinct group of GABAergic MnPO neurons which we termed type 2 neurons (see above) whole-cell voltage-clamp recordings revealed that NTS (1 μM) incubation resulted in increased inhibitory and excitatory input but no inward current in all neurons studied (n=24) (Fig 3A). NTS increased the frequencies and amplitudes of both sIPSCs and sEPSCs (Fig 3 A–C). NTSR1 antagonist SR48692 (100 nM) blocked the effect on IPSCs but did not affect the increase in sEPSCs frequency and amplitude (Fig 3A, D, E). In contrast, NTSR2 antagonist NTRC 824 (200 nM) abolished the effect on sEPSCs but left the increase in sIPSCs frequency and amplitude intact (Fig 3A, D, E). NTSR1 agonist TC NTR1 17 (300 nM) and neuromedin N (1 μM) increased the frequency and amplitude of sIPSCs but had no effect on excitatory inputs (Fig 3A, D, E). Single cell reverse transcription-PCR (scRT/PCR) analysis was performed for 33 MnPO GABAergic neurons in which NTS increased synaptic input without activating an inward current. NTSR1 or NTSR2 transcripts were not detected in any type 2 neuron (n=33 neurons).

Figure 3. NTS increases the frequencies of sEPSCs and sIPSCs in type 2 MnPO GABAergic neurons by activating NTSR2 and NTSR1, respectively.

A. NTS (1 μM) increases the frequency of both sEPSCs and sIPSCs in a type 2 MnPO GABAergic neuron (upper trace). In the presence of NTS the frequency of sEPSCs increased from 0.8 Hz to 3.3 Hz and the frequency of sIPSCs from 0.1 Hz to 1.1 Hz. The NTSR1 antagonist SR48692 (100 nM) blocked the NTS effect on sIPSCs but did not affect the effect on sEPSCs (2^nd trace from top). In contrast, in the presence of NTSR2 antagonist NTRC 824 (200 nM) NTS did not change the frequency of sEPSCs but increased the frequency of sIPSCs from 0.02 Hz to 1 Hz (3^rd trace from top). The NTSR1 agonist TC NTR1 17 (300 nM) and neuromedin N (1 μM) increased the frequency of sIPSCs from 0.01 Hz to 0.8 Hz and from 0.01 to 1.1 Hz respectively (traces 4 and 5 from top). In contrast GRP (1 μM) activated an inward current of 26 pA current (bottom trace). The neuron was held at −50 mV.

B,C. Cumulative histograms for the inter-sEPSCs interval (K-S test p < 0.01) (B) and for the sEPSC amplitude (K-S test, p < 0.01) (C) before (blue circles) and during NTS incubation (red triangles).

D. Bar chart summarizing the fold increase in sEPSCs frequency induced by NTS (1 μM), NTSR1 agonist TC NTR1 17 (300 nM), neuromedin N (1 μM). The effect of NTS (1 μM) during incubation with the NTSR1 antagonist SR48692 (100 nM) or the NTSR2 antagonist NTRC 824 (200 nM) is also presented. Bars represent means ± S.D. Cell numbers studied are indicated in brackets. There was a statistically significant difference between groups as determined by one-way ANOVA (F(4,80)=5.41, p=0.0007) followed by Tukey’s test relative to NTS effect; ** indicate statistical significance of P<0.01.

E. Bar chart summarizing the fold increase in sIPSCs frequency induced by NTS (1 μM), NTSR1 agonist TC NTR1 17 (300 nM), neuromedin N (1 μM) and of the effect of NTS during incubation with the NTSR1 antagonist SR48692 (100 nM) or the NTSR2 antagonist NTRC 824 (200 nM). Bars represent means ± S.D. There was a statistically significant difference between groups as determined by one-way ANOVA (F(4,80)=8.12, p=1.7×10⁻⁵) followed by Tukey’s test relative to NTS effect; ** indicate statistical significance of P<0.01.

To further characterize differences between type 1 and type 2 NTS-responsive neurons we tested the effect of a different hypothermic neuropeptide, gastrin related peptide (GRP). In a previous study we have established that GRP activates an inward current in a subpopulation of MnPO GABAergic neurons that express its receptor (Blais et al., 2016). Surprisingly, GRP activated inward currents only in type 2 neurons (n=6 put of 6 neurons studied), and no current in type 1 neurons (n=9 out of 9 neurons tested) (Figs 2A, 3A).

3.2. NTSR1 are expressed exclusively in MnPO neurons while NTSR2 are present only in MnPO astrocytes

The distribution of NTSR1 and NTSR2 transcripts among MnPO neurons and glia was studied using RNAscope technology. Neurons and astrocytes were identified as cells containing NeuN and Slc1a3 transcripts, respectively. In all preparations studied NTSR1 and NTSR2 transcripts were present in different populations of cells (Fig 4 A, B). NTSR1 was present in ~24% of neurons and in less than 1% of astrocytes. In contrast, NTSR2 was detected in ~33% of astrocytes and in less than 1% of neurons (Fig 4 C, D).

Figure 4. Subpopulations of MnPO neurons and astrocytes express NTSR1 and NTSR2 transcripts, respectively.

A, B. NTSR1, NTSR2, NeuN and Slc1a3 transcripts visualized using RNAscope technology.

A. DIC image (upper row ,1^st from left) of MnPO neurons and astrocytes in culture and the respective DAPI staining (gray). NTSR1 transcripts (green) and NTSR2 transcripts (red) are present in distinct populations of cells as indicated by their superimposed image (upper row, right). NTSR1 transcripts are present in one neuron as indicated by the presence of NeuN transcripts (blue) in the same cell (lower row, arrow). Transcripts of the neuronal marker NeuN and NTSR2 are expressed in different populations of cells as indicated by their superimposed images (lower row, right).

B. DIC image (upper row ,1^st from left) of MnPO neurons and astrocytes culture and the respective DAPI staining (gray). NTSR1 transcripts (green) and NTSR2 transcripts (red) are present in distinct populations of cells (upper row, right). Transcripts of the astrocytic marker Slc1a3 (blue) and NTSR1 are expressed in different populations of cells as indicated by their superimposed images (lower row, middle). NTSR2 transcripts are present in one astrocyte as indicated by their superimposed images (lower row, right, arrow).

C, D. Bar chart summarizing the percentages of neurons (NeuN positive cells) (C) and astrocytes (Slc1a3 positive cells)(D) that also express NTSR1, NTSR2, NeuN and Slc1a3 transcripts. Data were averaged from 50 fields of view for both staining conditions containing 2–8 neurons and 5–18 astrocytes each. Overall, 59 out of 252 cells expressing NeuN also expressed NTSR1 and 181 out of 550 Slc1a3-expressing cells were positive for NTSR2. There was a statistically significant difference between groups as determined by one-way ANOVA (F(2,149)=308.45, p=1.1×10⁻¹⁶ (C) and F(2,149)=478.96, p=1.0×10⁻¹⁵ (D)) followed by Tukey’s test relative to the other two groups; ** indicate statistical significance of P<0.01.

3.3. NTS activates Ca²⁺ release from intracellular stores in subpopulations of neurons and astrocytes by activating NTSR1 and NTSR2, respectively

We then examined the influence of NTS on the intracellular Ca²⁺ concentrations [Ca²⁺]_i in dissociated MnPO neurons and astrocytes loaded with fura-2 AM. Action potential propagation was blocked with TTX (1 μM). In 31% of neurons (84 out 270) NTS (1 μM) activated an increase in [Ca²⁺]_i, an effect which was fully blocked by preincubation with NTSR1 antagonist SR48692 (100 nM) (Fig 5A). Neuromedin N (1 μM) activated similar [Ca²⁺]_i responses which were also abolished by NTSR1 antagonist (Fig 5A). The NTSR2 antagonist NTRC 824 (200 nM) was without effect on either the NTS or neuromedin N responses (Fig 5A).

Figure 5. NTS activates [Ca²⁺]_i responses in neurons and astrocytes by activating NTSR1 and NTSR2, respectively.

A. [Ca²⁺]_i responses to NTS (1 μM) and neuromedin N (1 μM), from 3 acutely dissociated MnPO neurons (black, blue and green traces) and an average from 11 MnPO neurons (red trace). The NTSR1 antagonist SR48692 (100 nM) completely blocked the responses to both NTS and neuromedin N. The NTSR2 antagonist NTRC 824 (200 nM) did not affect the NTS and neuromedin N responses. B. [Ca²⁺]_i responses to NTS (1 μM), neuromedin N (1 μM), from 3 MnPO astrocytes (black, blue and green traces) and an average from 12 MnPO astrocytes (red trace). Neuromedin N was without effect in astrocytes. The NTSR2 antagonist NTRC 824 (200 nM) completely blocked the responses to NTS, while the NTSR1 antagonist SR48692 (100 nM) did not affect the NTS responses. In MnPO astrocytes high concentrations of SR48692 (10 μM) activated a response similar to the NTS response. A,B. TTX (1 μM) was added to all extracellular solutions.

NTS (1 μM) activated [Ca²⁺]_i responses in 36% of astrocytes (75 out 210), however their pharmacological profile was different. In astrocytes the responses were blocked by preincubation with NTSR2 antagonist NTRC 824 (200 nM) but not by NTSR1 antagonist SR48692 (100 nM) (Fig 5B). Furthermore, neuromedin N (1 μM) was without effect in all astrocytes tested. Since NTSR2 transcripts are present in astrocytes we also wanted to test the effect of selective NTSR2 activation. However, no selective NTSR2 agonist are available. It has been reported the NTSR1 antagonist SR48692 at high concentrations has partial agonist activity at NTSR2 (St-Gelais et al., 2006). In MnPO astrocytes SR48692 (10 μM) activated a [Ca²⁺]_i response that represented about 50% the NTS response (Fig 5B). In neurons SR48692 (10 μM) had no effect on [Ca²⁺]_i. We have also tested the effect of levocabastine, a histamine 1 receptor antagonist that has also agonist activity at NTSR2 and found that it had no action in neurons but mimicked the NTS responses in astrocytes (Suppl Fig 4A).

To confirm the presence of NTSR2 transcripts in astrocytes we have harvested astrocytes from slices with patch pipettes and carried out single cell RT/PCR experiments (Suppl Fig 4B). NTSR2 transcripts were detected in 3 out of 14 astrocytes while NTSR1 transcripts were not present in any cell.

3.4. Activation of both NTSR1 and NTSR2 is involved in NTS-induced hypothermia

To assess the action of NTS on CBT the neuropeptide (100nL, 1 μM) or aCSF (100nl control) was injected in the MnPO via a cannula. The neuropeptide induced a robust hypothermia of 5.2±1.1°C when compared to control injections (n=6 mice in each condition). The effect reached a maximum at ~1 h after the injection of NTS and recovered to control after ~3h post injections. Following the hypothermic phase, we observed also a mild hyperthermic rebound of 0.9±0.2 °C in mice injected with NTS but not in aCSF- injected mice (Fig 6A). I.c.v. injection of the same dose of NTS did not result in hypothermia (Suppl Fig 5A) indicating that this action could not be ascribed to its leaking into the ventricle and widespread brain exposure to the neuropeptide. We then carried out injections of NTSR1 agonists TC NTR1 17 (100nL, 300 nM) and ML312 (100nL, 5 μM), neuromedin N (100nL, 1 μM) and aCSF (100nM, control). Surprisingly, TC NTR1 17, ML312 and neuromedin N induced no hypothermic response, instead a hyperthermia of 1.3±0.8 °C, 0.7±0.3 °C and 2.4±0.9 °C, respectively, was observed at 1.5 h after injection (n=6 mice in each condition) (Fig 6B). Higher concentrations of TC NTR1 17 (1 or 3 μM), ML314 (15 μM) or neuromedin N (10 μM) resulted in similar hyperthermic responses (1.2±0.7 °C, 1.5±0.9 °C, 0.9±0.4 °C and 2.1±0.7 °C, respectively; n=6 mice in each condition).

Figure 6. Responses to intra-MnPO injection of NTS, NTSR1 agonist and neuromedin N.

A. Responses to intra-MnPO injection (arrow) of NTS (1 μM, 100 nl, red trace) and aCSF (100 nL, black trace). NTS induced a 5.2±1.1 °C hypothermia (one-way repeated measures ANOVA, F(1,238)=14.2, p=0.0003, followed by t-tests for each time point, ** P<0.01.

B. Intra-MnPO injection (arrow) of NTSR1 agonist TC NTR1 17 (300 nM, 100 nl, red trace), neuromedin N (1 μM, 100 nl, blue trace) and aCSF (100 nl, black trace,control). The NTSR1 agonist and neuromedin N induced a hyperthermia of induced hyperthermia of 1.3±0.8 °C and 1.9±0.9 °C, respectively (one-way repeated measures ANOVA F(2,357)=189.6, p=1×10⁻⁵, followed by t-tests for each time point relative to the aCSF injection, ** P<0.01).

C. Responses to intra-MnPO injection (2^nd arrow) of NTS (1 μM, 100 nl) 4h after an intra-MnPO injection (1^st arrow) of NTSR1 antagonist SR48692 (300 nM, 100nl, red trace) or aCSF (100 nl, black trace, control). NTS induced hypothermia of 5.2±1.2 °C and 4.7±1.1 °C, respectively (one-way repeated measures ANOVA F(1,238)=9.6, p=0.002 followed by unpaired t-tests for each time point, ** P<0.01). Note that the hypothermic phase of the NTS response was decreased by 54% and the hyperthermic phase was abolished.

D. Responses to intra-MnPO injection (2^nd arrow) of NTS (1 μM, 100 nl) 4h after an intra-MnPO injection (1^st arrow) of NTSR2 antagonist NTRC 824 (200 nM, 100nl, red trace) or aCSF (100 nl, black trace). NTSR2 antagonist potently reduced the NTS-induced hypothermia. Following a first injection of NTSR2 antagonist or aCSF, NTS decreased CBT by 4.5±0.9 °C and 0.9±0.5 °C, respectively (one-way repeated measures ANOVA F(1,238)=12.9, p=0.0005, followed by unpaired t-tests for each time point, ** P<0.01).

A–D. The points represent averages±S.D. (n = 6 mice) through the 10 h recording period. Experiments were carried out in parallel in groups of 6 mice for each treatment.

To determine the receptor subtypes involved in the NTS induced hypothermia, we studied the effect of the neuropeptide in the presence of specific antagonists. We first injected the respective antagonist intra-MnPO (or aCSF as control) followed 4h later by NTS injection. The NTSR1 antagonist SR48692 (300 nM, 100nl, red trace) decreased the hypothermia induced by NTS by 53% relative to control (Fig 6C). The antagonist also abolished the hyperthermic rebound that followed the hypothermia induced by NTS (Fig 6C). Higher concentrations of SR48692 (600 nM and 1.5 μM) resulted in similar reductions in the hypothermic response to NTS of 51 and 49%, respectively (n=6 mice in each condition).

Following intra-MnPO injections of NTSR2 antagonist NTRC 824 (200 nM, 100nl, red trace) the NTS-induced hypothermia was decreased by 81% relative to control (Fig 6D) while the hyperthermic rebound was not affected.

3.5. Activation of NTSR2 increases the frequency and amplitude of sEPSC in MnPO neurons. Role of P2X4 downstream of NTSR2 activation.

Since experiments with the selective NTSR2 antagonist NTRC 824 suggested that this receptor plays an important role in the hypothermia induced by NTS and the Ca²⁺ responses observed in astrocytes we tested the effect of NTSR2 activation on synaptic activity of MnPO neurons and on CBT. The only available compounds with agonist activity at NTSR2 are levocabastine and the NTSR1 antagonist SR48692 at high concentrations (St-Gelais et al., 2006).

In whole-cell voltage-clamp recordings high concentration of SR48692 (10 μM) and levocabastine (1 μM) mimicked the increase in sEPSCs frequency and amplitude induced by NTS (1 μM) (Fig 7A, C). In contrast, as described above, the NTSR1 agonist TC NTR1 17 (300 nM) only increased the frequency of sIPSCs (Fig7A). When the cells were pre-incubated with NTSR2 antagonist NTRC 824 (200 nM) the effects of SR48692 (10 μM) and levocabastine (1 μM) were abolished, confirming the involvement of this receptor subtype.

Figure 7. Activation of NTSR2 mimics the NTS effects on sEPSCs and induces hypothermia. The role of P2X4 receptors.

A. NTS (1 μM) increases the frequency of both sEPSCs and sIPSCs in a type 2 MnPO GABAergic neuron (upper trace). The frequency of sEPSCs increase from 2.2 Hz to 7.8 Hz. The NTSR1 agonist TC NTR1 17 (300 nM) increases the frequency of sIPSCs but does not affect the frequency of sEPSCs (2^nd trace from top). High concentration of NTSR1 antagonist SR48692 (10 μM) robustly increased the frequency of sEPSCs from 1.3 Hz to 7.1 Hz (3^rd trace from top). The partial NTSR2 agonist levocabastine (1 μM) also increased the frequency of sEPSCs from 1.9 Hz to 6.5 Hz (1^st trace from bottom).

B. The P2X4 antagonist PSB-12062 (5 μM) blocks the increase in sEPSCs frequency induced by NTS (1 μM).

C. Bar chart summarizing the fold increase in sEPSC frequency induced by NTS (1 μM), high concentration NTSR1 antagonist SR48692 (10 μM), levocabastine (1 μM), as well as co-incubation of NTSR1 and NTSR2 antagonists, and co-incubation of P2X4 antagonist and NTS. The NTSR2 antagonist NTRC 824 (200 nM) blocked the NTSR1 antagonist (10 μM) effects on sEPSCs. The P2X4 antagonist PSB-12062 (5 μM) blocked the increase in sEPSCs frequency and amplitude induced by NTS (1 μM). Bars represent means ± S.D. Cell numbers studied are indicated in brackets. There was a statistically significant difference between groups as determined by one-way ANOVA (F(4,50)=62.3, p=1.1×10⁻¹⁶) followed by Tukey’s test; ** and * indicate statistical significance of P<0.01, and P<0.05, respectively relative to the NTS effect.

D. Intra-MnPO injection (arrow) of high concentration NTSR1 antagonist SR48692 (10 μM) or of the NTSR2 ligand levocabastine (1 μM) induced a hypothermia 2.3±0.8 °C and 1.7±0.9 °C, respectively. The responses were statistically significant when compared with aCSF injection (not shown) (one-way repeated measures ANOVA, F(1,238)=5.4, p=0.037 and F(1,238)= 25.8, p=0.0002, respectively).

E. Responses to intra-MnPO injection (2^nd arrow) of NTS (1 μM, 100 nl) 4h after an intra-MnPO injection (1^st arrow) of P2X4 antagonist PSB-12062 (5 μM,100nl, red trace) or aCSF (100 nl, black trace, control). P2X4 antagonist potently decreased the NTS-induced hypothermia from 4.5±0.9 °C (control) to 0.9±0.5 °C, respectively (one-way repeated measures ANOVA, F(1,238)=16.1 p= 0.0001, followed by unpaired t-tests for each time point, ** P<0.01).

Activation of astrocytes results in release of gliotransmitters such as ATP. We therefore tested whether NTSR2 activation in astrocytes results in downstream activation of purinergic receptors. The broad P2X antagonist PPADS (10 μM) completely blocked the NTS effect on sEPSCs (not shown). We then tested antagonists selective for P2X receptor subtypes expressed in the hypothalamus. The P2X2/3 antagonist RO-3 (5 μM) as well as the P2X7 antagonist A 438079 (5 μM) did not change the NTS effect on sEPSCs. In contrast, the P2X4 selective antagonist PSB-12062 (5 μM) completely abolished the increase in sEPSCs frequency and amplitude induced by NTS (Fig 7B, C).

We then tested the effect on CBT of levocabastine and high concentration NTSR1 antagonist injected in the MnPO (Fig 7D). The partial agonists induced a hypothermia 2.3±0.8 °C and 1.7±0.9 °C, respectively, relative to aCSF injection (n=6 mice in each condition). We then tested the involvement of P2X4 receptors in the NTS-induced hypothermia (Fig 6E). Following intra-MnPO injections of PSB-12062 (5 μM, 100nl, red trace) the NTS-induced hypothermia was decreased by ~68% relative to control (Fig 7E).

4. Discussion

In this study we have revealed using RNAscope in situ hybridization that the expression of NTSR1 was restricted to a subpopulation of MnPO neurons while NTSR2 transcripts were present only in a subpopulation of astrocytes. This observation is in agreement with data obtained using NTSR1-cre and NTSR2-cre mice which reported a similar distribution of the two receptors in the ventral tegmental area (Woodworth et al., 2018). NTSR2 expression in astrocytes has also been found in the rat cortex (Nouel et al., 1999).

As reported also in other brain regions NTS had exclusively excitatory effects on MnPO neurons (Jolas and Aghajanian, 1996; Stowe and Nemeroff, 1991; Xue et al., 2007). We dissected two distinct mechanisms of excitation in separate populations of MnPO GABAergic neurons. In NTSR1-expressing neurons the neuropeptide activated an inward Na⁺- dependent current which resulted in depolarization and increased firing rate (type 1 neuron). The NTS-induced excitation of NTSR1-expressing neurons was not dependent on synaptic input. NTSR1 activation also induced Ca²⁺ release from intracellular stores in these neurons. In a different population of MnPO GABAergic neurons pharmacological blockade of synaptic inputs abolished the excitatory effect of NTS (type2 neuron). These neurons did not express either NTSR1 or NTSR2. Surprisingly, the increased frequency and amplitude of sEPSCs observed in the presence of NTS was blocked by a NTSR2 antagonist. We also observed that NTS activated Ca²⁺ release from intracellular stores in a population of astrocytes, an effect which was blocked by a NTSR2 antagonist. Taken together these results suggest that the NTS excitation of this population of neurons depends upon activation of NTSR2 in astrocytes and subsequent release of gliotransmitters that act at presynaptic locations. We further revealed that ATP release plays an important role in this mechanism since a P2X4 blocker abolished the increased synaptic excitation activated by the neuropeptide. A role of P2X4 receptors in potentiating excitatory input has been proposed also in other brain regions (Baxter et al., 2011; Boue-Grabot and Pankratov, 2017).

Among MnPO neurons NTS acted selectively at the GABAergic population of neurons. This observation was further supported by scRT/PCR results which confirmed that NTSR1 transcripts were present in GABAergic neurons (Fig 2). Postsynaptically, NTSR1 activation resulted in an increase in the frequency of sIPSCs, providing further evidence that NTSR1 are expressed in GABAergic neurons. It is interesting to note that a recent study in mouse cortex reported that NTS recruited large populations of inhibitory interneurons by a Na⁺- dependent depolarization and by increased excitatory synaptic input (Case and Broberger, 2018), suggesting that the mechanism we propose here may be present also in other brain regions.

Neuromedin N is a neuropeptide derived from the same precursor polypeptide as neurotensin, and with similar but distinct expression and actions (Vincent, 1995). Here we report, for the first time, that neuromedin N acts as a selective NTSR1 agonist in MnPO neurons. Neuromedin N mimicked all the NTS effects in all neurons studied (activation of inward current and Ca²⁺ release from intracellular stores) and its effects were blocked by NTSR1 antagonist and were not affected by NTSR2 antagonist. In contrast, neuromedin N did not activate Ca²⁺ release in astrocytes and did not increase synaptic excitation in MnPO neurons, effects which we ascribed to NTSR2 activation. There are no selective NTSR2 agonists available, however there is evidence that the histamine 1 receptor antagonist levocabastine and the NTSR1 antagonist SR48692 at high concentration act as a partial agonist at mouse NTSR2 (St-Gelais et al., 2006). Both compounds increased the frequency and amplitude of sEPSCs mimicking NTS effects in the same neurons, actions which were blocked by NTSR2 antagonist. We also found that high concentrations of SR48692 induced Ca²⁺ release in MnPO astrocytes, an effect which was blocked by NTSR2 antagonist. These results are similar with those previously reported in other brain regions and point to a role of NTSR2 in astrocytes in the observed actions.

Neurotensin is one of the most potent hypothermic agents when infused intracerebroventricularly (St-Gelais et al., 2006), an effect which was attributed to a reduction in heat production and no change in heat loss mechanisms (Gordon et al., 2003). Studies in NTSR1 ko mice have reported a lack of the hypothermia to NTS therefore attributing this receptor subtype a key role in this response(Mechanic et al., 2009; Pettibone et al., 2002; Remaury et al., 2002). However, receptor knockdown using antisense oligodeoxynucleotides in adult mice indicated that NTSR1 knockdown did not affect the hypothermic response to NTS, instead NTSR2 knockdown resulted in a decreased response to the neuropeptide (Dubuc et al., 1999). Given the wide expression of NTSR1 during development (Woodworth et al., 2018) it is possible that the constitutive NTSR1 ko mouse models may present abnormalities in thermoregulatory networks. Furthermore, intra MnPO injection of low doses of NTS, likely acting at the high affinity NTSR1, resulted in hyperthermia and only large doses of the neuropeptide, likely acting at both NTSR1 and NTSR2 induced hypothermia (Benmoussa et al., 1996). Nevertheless, it is possible that NTS can induce hypothermia by acting also at other brain sites where NTSR1 activation may be sufficient to achieve this effect. However, this situation is unlikely since intracerebroventricular infusion of NTSR1 selective agonists TC NTR1 17 (200 nl, 10 μM) or ML314 (200 nl, 100 μM), in contrast to NTS infusion, failed to induce hypothermia (n=6 mice in each condition, not shown).

In MnPO we found that the NTSR1 and NTSR2 antagonists decreased the NTS induced hypothermia by 53% and 80%, respectively. Further support for the role of NTSR2 in is the observation that levocabastine as well as high concentrations of SR48692 induced, substances with partial agonist action at mouse NTSR2, induced small but significant hypothermic responses. Taken together our findings represent a first report of a receptor expressed selectively in astrocytes exerting a potent modulatory action on thermoregulation.

It has been established that MnPO GABAergic neurons control thermogenesis (Morrison and Madden, 2014) while a population of glutamatergic neurons expressing NK3 receptors control heat loss mechanisms (Mittelman-Smith et al., 2015). Our data revealed that in MnPO NTS acts primarily on GABAergic neurons, similar with observations in cortex (Case and Broberger, 2018), and in agreement with the concept that NTS reduces heat production (Gordon et al., 2003). While type 1 neurons were characterized by NTSR1 expression, type 2 neurons can be distinguished by their responsiveness to GRP, further suggesting that they represent GABAergic neurons that control thermogenesis (Blais et al., 2016). Based on our results we propose a cellular mechanism in which type 2 MnPO GABergic neurons send potent inhibitory tone to downstream neurons that drive thermogenesis (see visual abstract). In our proposed circuit type 2 neurons also receive direct inhibitory projections from type 1 neurons. In this model NTSR1 activation results in overall excitation of neurons driving thermogenesis, while activation of NTSR2 or of both NTSR1 and NTSR2 results in inhibition of these neurons, decreased thermogenesis and CBT.

In summary, this study elucidates the cellular mechanisms by which NTS depolarizes MnPO GABAergic neurons and reveals a novel mechanism of excitation downstream of astrocytic NTSR2.

Supplementary Material

1. Suppl Fig 1. Dose response relationship of the NTS effect on spontaneous firing rate of MnPO neurons.

Percentage change in firing rate relative to control in response to 4 concentrations of NTS. Data pooled from 8 MnPO GABAergic neurons.

2. Suppl Fig 2. Differential sensitivity of type 1 and type 2 NTS responses to blockade of synaptic inputs.

A. NTS (1 μM) depolarizes and increases the firing rate of a type 1 MnPO GABAergic neuron (upper trace). In the presence of CNQX (20 μM), AP-5 (100 μM), and picrotoxin (50 μM), NTS induced a depolarization and increased the firing rate of the neuron (lower trace).

B. NTS (1 μM) increases the firing rate of a type 2 MnPO GABAergic neuron (upper trace). The excitatory effect of NTS was abolished in the presence of CNQX (20 μM), AP-5 (100 μM), and picrotoxin (50 μM) (lower trace).

3. Suppl Fig 3. The NTS-induced inward current depends upon extracellular Na⁺.

A. NTS (1 μM) activates an inward current of 24 pA in a MnPO GABAergic neuron. The neuron was held at −50 mV (upper trace). Note that NTS also induced an increase in the frequency and amplitude of sEPSCs. When the extracellular solution was replaced with NMDGCl external solution NTS (1 μM), added to this solution, did not activate any inward current (lower trace).

5. Suppl Fig 4. Levocabastine activates [Ca²⁺]_i responses in astrocytes by activating NTSR2.

A. [Ca²⁺]_i responses to NTS (1 μM), neuromedin N (1 μM) and levocabastine (1 μM), from 3 MnPO astrocytes (black, blue and green traces) and an average from 10 MnPO astrocytes (red trace). Neuromedin N was without effect in astrocytes. The NTSR2 antagonist NTRC 824 (200 nM) completely blocked the responses to both NTS and levocabastine. TTX (1 μM) was added to all extracellular solutions.

B. NTSR2 but not NTSR1 transcripts are present in MnPO astrocytes harvested from slices. Gels from a batch of 14 MnPO astrocytes. The expected sizes of the PCR products are (in base pairs) 248 and 400, respectively. Negative (−) control was amplified from a harvested cell without reverse-transcription, and positive control (+) was amplified using 1 ng of hypothalamic mRNA. NTSR1 transcripts were detected in 3 out of 14 astrocytes, while NTSR2 transcripts were not detected in any astrocyte.

6. Suppl Fig 5. Intracerebral injection of low dose NTS does not induce hypothermia.

A. Responses to i.c.v. injection (arrow) of NTS (1 μM, 100 nl, red trace) and aCSF (100 nL, black trace). NTS had no significant effect at this dose.

B. Location of cannula aiming the MnPO and a typical dye spot (arrow). 100 nl Alcian blue (2%) was injected via the cannula at the end of the experiment. The slice corresponds to the section 0.5 mm anterior from Bregma.

Highlights.

• A population of GABAergic preoptic neurons express NTSR1.

• NTSR2 are expressed exclusively in preoptic astrocytes

• NTSR1 and NTSR2 activate Ca²⁺ release from intracellular stores in neurons and astrocytes, respectively

• Activation of both NTSR1 and NTSR2 in the median preoptic nucleus is necessary to induce hypothermia