Activation of Preoptic Arginine Vasopressin Neurons Induces Hyperthermia in Male Mice

Abstract

Arginine vasopressin (AVP) is a neuropeptide acting as a neuromodulator in the brain and plays multiple roles, including a thermoregulatory one. However, the cellular mechanisms of action are not fully understood. Carried out are patch clamp recordings and calcium imaging combined with pharmacological tools and single-cell RT-PCR to dissect the signaling mechanisms activated by AVP. Optogenetics combined with patch-clamp recordings were used to determine the neurochemical nature of these neurons. Also used is telemetry combined with chemogenetics to study the effect of activation of AVP neurons in thermoregulatory mechanisms. This article reports that AVP neurons in the medial preoptic (MPO) area release GABA and display thermosensitive firing activity. Their optogenetic stimulation results in a decrease of the firing rates of MPO pituitary adenylate cyclase-activating polypeptide (PACAP) neurons. Local application of AVP potently modulates the synaptic inputs of PACAP neurons, by activating neuronal AVPr1a receptors and astrocytic AVPr1b receptors. Chemogenetic activation of MPO AVP neurons induces hyperthermia. Chemogenetic activation of all AVP neurons in the brain similarly induces hyperthermia and, in addition, decreases the endotoxin activated fever as well as the stress-induced hyperthermia.

Arginine vasopressin (AVP) is a 9 amino-acid peptide synthesized by neurons in the paraventricular nuclei (PVN) and supraoptic nuclei of the hypothalamus, which are part of the neurohypophysis. Additional, more sparsely distributed, populations of AVP-expressing neurons have been identified also in the suprachiasmatic nucleus, central nucleus of the amygdala, and medial preoptic area (MPO) (1, 2) as well as in sensory neurons (3). Although the characteristics and function of PVN and supraoptic nuclei neurons have been studied extensively (4, 5), much less is known about the other AVP-expressing groups.

AVP, released from the posterior pituitary, plays an important role in the control of plasma osmolality and blood volume/pressure (6). In addition, AVP released in the brain exerts potent effects on social and maternal behavior (7), pain sensation (8), and thermoregulation (9, 10). The peptide potently decreases the endotoxin-induced fever (11) and induces hypothermia when injected systemically (12). The cellular mechanisms and neuronal networks modulated by AVP in its thermoregulatory actions are not fully understood.

The preoptic area contains thermoregulatory neurons that control, via their projections to the dorsomedial hypothalamus and rostral raphe pallidus, heat generation or heat loss mechanisms (13). Recent studies have identified a population of pituitary adenylate cyclase-activating polypeptide (PACAP)-expressing preoptic neurons that integrate peripheral thermal information and control the core body temperature (CBT) (14) and are necessary for the fever response to endotoxin (15). This study has investigated the cellular characteristics of MPO AVP neurons and studied the influence their activity exerts on CBT, stress-induced hyperthermia, and endotoxin-induced fever.

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 (1996, 7th ed.; Washington DC: National Research Council, National Academies Press). 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 standards and the regulations set forth in the Animal Welfare Act. Efforts were made to minimize the number of animals used and their suffering.

Telemetry and MPO Injections

Four- to 6-month-old male mice were anesthetized with isoflurane (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 CBT measurement. For brain injections, mice were first subject to stereotaxic placement of a bilateral guide cannula (27 Ga) as previously described (16). Coordinates for cannula implants in the MPO were: from Bregma: 0.05 mm, 0.35 mm, and −0.35 mm lateral, and ventral 4.75 mm (Paxinos and Franklin, 2001). The ambient temperature was maintained at ~28 ± 0.5°C in a 12:12-hour light-dark cycle-controlled room (lights on 8:00 am). All substances injected were dissolved in sterile artificial cerebrospinal fluid (aCSF). For MPO injections, mice were placed in a stereotaxic frame and the injector (33 Ga) 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.”

Transgenic Animals and AAV Injections

The AVP-IRES2-Cre-D driver line (B6.Cg-Avp^{tm1.1(cre)Hze}/J; stock no: 023530) was purchased from Jackson Laboratory (Bar Harbor, ME, USA). Homozygous male mice 4 to 6 months old received bilateral stereotaxic injections (200 nL at a rate of 0.1 μL/min) of AAV5-hSyn-DIO-hM3D(Gq)-mCherry (Addgene #44361; 2.3 × 10¹³ virus molecules/mL) in the MPO to express the excitatory designer receptor hM3D(Gq) in MPO AVP neurons. These mice are referred to as MPO^AVP;hM3D(Gq). Similarly, for patch-clamp recordings in slices, other animals were given bilateral MPO injections (200 nL at a rate of 0.1 μL/min) of AAV5-EF1a-double floxed-hChR2(H134R)-EYFP-WPRE-HGHpA (Addgene #20298; 2.1 × 10¹³ virus molecules/mL) to express the excitatory opsin ChR2 in MPO AVP-neurons. These mice are referred to as MPO^AVP;hChR2. To ensure that only cre-expressing neurons were targeted, injections of the 2 viral vectors in wild-type C57/Bl6 mice (3 mice for each) were carried out and confirmed that no fluorescent signal detected in the brain of these mice. Together with the results indicating that AVP transcripts were present in the labeled neurons (see Results), it is concluded that AVP neurons were specifically targeted using these viral vectors.

The AVP-IRES2-Cre-D driver line was crossed with the B6N;129-Tg(CAG-CHRM3*,-mCitrine)1Ute/J line (Jackson Laboratory; stock no: 026220) to generate mice that express the excitatory designer receptor hM3D(Gq) in AVP neurons in all brain regions. This new line was named AVP^hM3D(Gq). Genotyping was carried out according to the provider’s instructions to identify heterozygous double transgenic mice. To activate hM3D(Gq) mice received an IP injection of 1 mg/kg clozapine N-oxide (CNO) dissolved in 0.3 mL sterile saline.

The AVP-IRES2-Cre-D driver line was also crossed with the Slc32a1^tm1Lowl (also referred to as Vgat^flox/flox) (Jackson Laboratory; stock no: 012897) to generate mice that have the GABA vesicular transporter Vgat (solute carrier family 32 member 1), knocked down in AVP neurons. Heterozygous double transgenic mice AVP^+/-;Vgat^+/- were crossed to obtain AVP^+/+;Vgat^-/-. These mice are referred to as AVP^VGAT-/-.

Finally, bilateral stereotaxic injections (200 nL at a rate of 0.1 μL/min) of AAV5-hSyn-DIO-hM3D(Gq)-mCherry (Addgene #44361; 2.3 × 10¹³ virus molecules/mL) were performed in the MPO of AVP^VGAT-/- to express the excitatory designer receptor hM3D(Gq) in MPO AVP^VGAT-/- neurons. These mice are referred to as MPO^{AVP;VGAT-/-;hM3D(Gq)}.

Slice Preparation

Coronal tissue slices containing the MPO were prepared from AVP-IRES2-Cre-D or MPO^AVP;hChR2 male mice 4 to 6 months old. The slice preparation was as previously described (17). The slice used in these recordings corresponded to the sections located from 0.15 to -0.05 mm from Bregma in the mouse brain atlas (18).

Dissociated Preoptic Neurons From Slices

The MPO was punched out of a brain slice from an AVP-IRES2-Cre-D mouse and incubated in Hibernate A (Invitrogen, Temecula, CA) and papain (Worthington, Lakewood, NJ) (1 mg/mL) for 10 minutes 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 (1000g for 2 minutes) 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 to 10 days before being used for experiments.

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 to 305 mOsm, equilibrated with 95% O₂ and 5% CO₂ (pH 7.4). Other salts and agents were added to this medium. 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 to 4 MΩ. All voltages were corrected for the liquid junction potential (−13 mV). 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 cell capacitance was determined and compensated using the Multiclamp Commander software.

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, whereas 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) was maintained at 36 to 37°C unless otherwise indicated. The thermal coefficient of a neuron, defined as the slope of the linear regression fitted to the firing rate versus temperature plot was determined as described previously (19). Briefly, the temperature was varied in the range 35 to 41°C, and the thermal coefficient was determined over a temperature range of at least 3°C. A neuron having a thermal coefficient higher than 0.8 impulses/s^–1/°C^–1 was defined as thermosensitive.

Synaptic activity was quantified and analyzed statistically as described previously (20). Briefly, synaptic events were detected and analyzed (amplitude, kinetics, frequency) off-line using a peak detection program (Mini Analysis program, Synaptosoft, Decatur, NJ, USA). Events were detected from randomly selected recording stretches of 2 minutes 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 2-sample 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.

Spot illumination of MPO^AVP;hChR2 neurons was carried out using a Polygon300 illumination system (Mightex, Toronto, Canada) that allows the control of the size, shape, intensity and position of the spot illuminated. In a field of view there were 1 to 3 fluorescent neurons. The illumination spot position and size were controlled using the Polygon 300 software. Typically, stimulation of at most 1 of the fluorescent neurons presents in the field of view resulted in postsynaptic effects in the recorded neuron.

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 AVP. At this excitation frequency, photobleaching and phototoxicity were minimal. Fura-2AM loading and data acquisition were carried out as described in previous studies.

Chemicals

All agonists and antagonists were purchased from Tocris (Ellisville, MO, USA). All the other chemicals were from Millipore Sigma (Carlsbad, CA, USA).

Cell Harvesting, Reverse Transcription, and PCR

MPO neurons in slices were patch-clamped and then harvested into the patch pipette by applying negative pressure as previously described (17). 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 minutes and then put on ice for 3 minutes. 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 minutes, at 50°C for 50 minutes, and then at 75°C for 15 minutes. After RT samples were immediately put on ice, 1 μL of RNAse H was added to samples and kept at 37°C for 20 minutes. PCR assays were carried out using the pairs of primers listed in Table 1. To exclude false negatives, the presence of transcripts of the housekeeping gene 18s RNA was tested. Samples that were negative for the hosekeeping gene were eliminated. Two rounds of PCR, using nested primers, were used to detect AVP, AVPr1a, and AVPr1b transcripts. For PACAP, a second round of PCR was not necessary because it did not yield additional positives when compared with the first round.

Table 1.

PCR primers for the genes studied

	Gene	External primers	Amplicon size	Internal primers	Amplicon size
1	Arginine vasopressin (AVP)	F: “CTGCCCAAGAGGCGGCAAG”	Bp:380	F: “ CAAAGGACGCTTCGGAC”	Bp:257
		R: “CCAGCTGTACCAGCCTTAGC”		R: “ GTGTGGCGTTGCTTGGCTC”
2	Arginine vasopressin receptor 1A (AVPr1a)	F: “GGCTGTGCCGGGTGGTGAAGCA”	Bp:622	F: CATACCACAGTACTTTATCTT”	Bp:402
		R: “TCGGTCCAAACGAAATTGGTAT”		R: “GGACGATGAAGAAAGGTGT”
3	Arginine vasopressin receptor 1B (AVPr1b)	F: “TCAGTGTCCAGATGTGGTCTGT”	Bp:382	F: “AGAATGCCCCTAATGAAGAT”	Bp:340
		R: “TAAGAGATGCTGGTCTCCATAG”		R: “GTCTCCATAGTGGCTTCTCC”
4	PACAP	F: “CCTACCGCAAAGTCTTGGAC”	Bp:103
		R: “TTGACAGCCATTTGTTTTCG”
5	18s RNA	F: “CGGCTACCACATCCAAGGAA”	Bp:463
		R: “GCTGGAATTACCGCGGCT”

RNAscope in situ hybridization

To detect single mRNA molecules, RNAscope was performed on fixed cultures of acutely dissociated MPO neurons and astrocytes from slices. The cells were kept in culture for 7 to 10 days. In situ hybridization was performed according to the protocol of the RNAscope Multiplex Fluorescent Reagent Kit v2 (Cat. No. 320293). Briefly, coverslips containing MPO cells were incubated in cold 4% paraformaldehyde for 30 minutes and mounted on microscope slides. The cells were then dehydrated in 50%, 70%, and 100% ethanol for 5 minutes each at room temperature. Cells were then rehydrated in 70%, 50% ethanol, and PBS. Cell were treated with RNAscope hydrogen peroxide for 10 minutes, washed 5 times with distilled water, and, for antigen accessibility, treated with Protease III for 10 minutes at room temperature 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 2 hours 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 1-day protocol has been used in all experiments to preserve the quality of the preparations.

Image acquisition and analysis

Images were collected on an Olympus BX-51 inverted microscope (Olympus, Melville, NY) using a 60× objective. Ten 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 differential interference contrast images and the DAPI staining. Neurons or astrocytes were identified using mRNAs probes for Rbfox3/NeuN or Slc32a1, respectively (Opal 570). AVPr1a (Opal 520) and AVPr1b 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 1 fluorescent punctum of minimum 0.45 µm². A vast majority of cells analyzed in the study presented 4 or more puncta for any of the probes used.

Endotoxin-induced Fever and Restraint Stress-induced Hyperthermia

To induce a fever response, a dose of 0.1 mg/kg lipopolysaccharides (LPS) dissolved in 0.3 mL sterile saline was injected IP . To elicit restraint-stress hyperthermia, mice were anaesthetized with isoflurane (1.5%) and inserted in 50-mL tubes and placed horizontally in their home cages. The tubes presented perforations (4-mm diameter) every 20 mm on the entire surface to allow for air circulation. After 2 hours, the lids were removed and mice were released in their home cages.

Experimental Design and Statistical Analyses

The values reported are presented as mean ± SD. Power analyses (www.biomath.info) with values from the data (means and SDs) were undertaken to calculate that this study was powered to detect a 0.7°C change in CBT with at least > 80% reliability for all transgenic models used in this study. Statistical significance of the results pooled from 2 groups was assessed with t tests using Prism4 (GraphPad Software). 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 1-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 are not presented in a figure, the respective values are reported in the text.

Results

MPO^AVP;hChR2 Neurons Are Pacemakers and Display Thermosensitive Firing Activity

The AVP-IRES2-Cre-D mice received bilateral MPO injections of AAV5-EF1a-double floxed-hChR2(H134R)-EYFP-WPRE-HGHpA (see Methods) to express the excitatory opsin ChR2-EYFP in MPO AVP neurons. These mice are referred to as MPO^AVP;hChR2. Patch-clamp whole-cell recordings from these neurons revealed that they are pacemaker cells (i.e., they have a regular firing pattern and most action potentials arise from ramp-like “prepotentials”) (Fig. 1A) (21). The average firing rate at 36°C was 7.2 ± 1.4 Hz (n = 14). Intrinsic properties of preoptic/anterior hypothalamus neurons were investigated using injections of square current pulses. The average capacitance of the neurons studied was 21.8 ± 2.1 pF (n = 14). All neurons tested displayed a time-dependent inward rectification when hyperpolarizing current pulses were injected (Fig. 1B) that averaged 6.6 ± 0.9 mV (n = 14) during a –40 pA hyperpolarizing pulse. This depolarizing “sag” has been shown to reflect the activation of the hyperpolarization-activated cation currents (22). Injection of positive currents in all cell types generated a sustained train of action potentials with no adaptation in all neurons studied. Upon repolarization, all neurons displayed a slow afterhyperpolarization of 7.7 ± 1.6 mV (n = 21) following a 30-pA depolarizing step (Fig. 1B). Because preoptic thermosensitive neurons are pacemakers, the firing rates of MPO^AVP;hChR2 neurons during ramp changes in the bath temperature were measured (Fig. 1C). All neurons studied displayed thermosensitive firing activity, increasing the firing rate with higher temperature. The thermal coefficient of the neuron depicted in Fig. 1C was 1.29 impulses/s⁻¹/°C⁻¹ (Fig. 1D). The thermal coefficients of all MPO^AVP;hChR2 neurons studied was above 0.8 impulses/s⁻¹/°C⁻¹, hence these neurons can be classified as thermosensitive, and averaged 1.08 ± 0.26 impulses/s⁻¹/°C⁻¹ (n = 8). To verify the presence of AVP transcripts in MPO^AVP;hChR2, single-cell RT-PCR (scRT-PCR) in these neurons were carried out. AVP transcripts in 8 of 10 MPO^AVP;hChR2 studied were detected (Fig. 1E).

Figure 1.

Electrophysiological characteristics of MPO^AVP;hChR2 neurons. (A) Representative example of spontaneous firing activity of an MPO^AVP;hChR2 neuron. (B) Membrane potential responses to hyperpolarizing current steps of -40, -30, -20, and -10 pA (left) reveal the presence of a “sag” (*) and the absence of o a low threshold spike after the steps. Depolarizing current injections of 10 and 30 pA (right) elicit increased firing activity with no adaptation and the activation of a slow afterhyperpolarization (arrow). MPO^AVP;hChR2 neurons display thermosensitive firing activity. (C) The spontaneous firing rate of an AVP neuron gradually decreases from 12.8 Hz at 39.6°C to 2.9 Hz at 34.3°C. (D) The slope of the firing rate vs temperature plot yields a thermal coefficient of 1.29Hz/°C. (E) AVP transcripts are present in MPO^AVP;hChR2 neurons. Representative gel from 10 MPO^AVP;hChR2 neurons. The expected size of the PCR product is 257 bp. Negative (−) control was amplified from a harvested cell without reverse transcription, and positive control (+) was amplified using 1 ng of hypothalamic mRNA. AVP transcripts were detected in 8 of 10 neurons studied. AVP, arginine vasopressin; MPO, medial preoptic.

Optogenetic stimulation of MPO^AVP;hChR2 neurons decreases the firing rate of MPO PACAP neurons by increasing the frequency of inhibitory postsynaptic currents

Spot illumination of MPO^AVP;hChR2 neurons resulted in their instant depolarization and a robust increase in their firing rate in all neurons tested (n = 6) (Fig. 2A). In contrast, recordings from other MPO neurons revealed that spot illumination of MPO^AVP;hChR2 neurons resulted in a decrease in their firing rate of (25 of 68 MPO neurons studied) or no effect in the others. Fig. 1B illustrates the decrease in firing rate of a MPO neuron (from 7.5 Hz to 1.3 Hz) induced by optogenetic stimulation of a nearby MPO^AVP;hChR2 neuron. The effect was associated with a robust increase in the frequency of inhibitory postsynaptic potentials (IPSPs) from 0.3 Hz to 3.7 Hz (Fig. 2B). Overall, optogenetic stimulation of a nearby MPO^AVP;hChR2 neuron reduced the firing rate of the recorded MPO neurons from 7.4 ± 1.8 Hz to 2.5 ± 1.1 Hz (n = 25, paired t test t(24) = 19.6; P = 1.8 × 10^-11) and increased their IPSP frequencies from 1.5 ± 0.6 Hz to 6.8 ± 3.0 Hz (n = 25, paired t test t(24) = -10.76; P = 6.2 × 10^-10). To exclude the possibility of synaptic excitation in the activation of IPSPs downstream of MPO^AVP;hChR2, the effect of optogenetic stimulation in the presence of CNQX (20 µM) and AVP (50 µM) was studied. In the presence of the antagonists, stimulation of MPO^AVP;hChR2 neurons resulted in activation of IPSCs (Fig. 2C), indicating that GABA is released by MPO^AVP;hChR2. Prolonged optogenetic stimulation (up to 5 minutes) of MPO neurons in the presence of Gabazine (5 µM) did not activate excitatory postsynaptic currents (EPSCs) or apparent inward or outward currents in the cells studied (n = 20). Also tested was the optogenetic activation of IPSCs in the presence of the AVPr1a antagonist SR49059 (30 nM) and AVPr1b antagonist TASP0390325 (20 nM). The antagonists did not affect the frequency or amplitude of optogenetically evoked IPSCS (not shown). The frequency of the evoked IPSCs averaged 5.6 ± 2.0 Hz and 5.7 ± 1.9 Hz in control and in the presence of the antagonists, repectively (n = 7, paired-test t(6) = -0.25; P = 0.81).

Figure 2.

Effects of optogenetic stimulation of MPO^AVP;hChR2 neurons on the activity of nearby MPO neurons. (A) Blue light activates action potential firing in an MPO^AVP;hChR2 neuron expressing hChR2(H134R)-eYFP (inset). Increasing the light intensity of the illumination spot activates increased firing activity. The neuron was held at -56 mV by hyperpolarizing current injection of -10 pA). (B) Optogenetic stimulation of an MPO^AVP;hChR2 neuron decreases the spontaneous firing rate of a nearby MPO neuron from 7.5 Hz to 1.2 Hz and increases the frequency of IPSC from 0.3 Hz to 3.7 Hz (see expanded traces). (C) Optogenetic stimulation of an MPO^AVP;hChR2 neurons activates sIPSCs in a nearby MPO neuron. The sIPSCs were abolished by gabazine (5 µM) (middle trace). After washout of gabazine prolonged optogenetic stimulation (90 seconds) of the same neuron induced a similar activation of sIPSCs (lower trace). Recordings were performed in the presence of CNQX (20 µM) and AP-5 (50 µM) to block excitatory synaptic transmission. The neuron was held at -50 mV. (D) PACAP transcripts are present in MPO neurons inhibited by optogenetic stimulation of nearby MPO^AVP;hChR2 neurons. Representative gel from 20 recorded MPO neurons. The expected size of the PCR product is 103 bp. Negative (−) control was amplified from a harvested cell without reverse transcription, and positive control (+) was amplified using 1 ng of hypothalamic mRNA. PACAP transcripts were detected in 18 of 20 neurons studied. IPSC, inhibitory postsynaptic current; MPO, medial preoptic; PACAP, pituitary adenylate cyclase-activating polypeptide.

Preoptic thermoregulatory neurons represent a subpopulation of preoptic PACAP neurons; therefore, the presence of PACAP transcripts in MPO neurons inhibited by optogenetic stimulation of MPO^AVP;hChR2 neurons was tested. PACAP transcripts were detected in 18 of 20 neurons studied (Fig. 2D), suggesting that MPO^AVP;hChR2 neurons can influence the activity of MPO thermoregulatory neurons.

Finally, MPO^AVP;hChR2 neurons were recorded and stimulated, using post illumination, a different MPO^AVP;hChR2 neuron to question the presence of reciprocal connections. Only 2 of 11 MPO^AVP;hChR2 neurons studied presented IPSCs evoked by optogenetic stimulation of other MPO^AVP;hChR2 neurons, suggesting that such connections exist but are relatively sparse.

Also examined were the brains of MPO^AVP;hChR2 mice in coronal slices to identify projection sites outside MPO; however, no fluorescent signal was detected outside the MPO, suggesting their projections are mostly local.

Local AVP Application Has Opposite Effects on the Firing Activity of Distinct Groups of MPO Neurons

AVP (1 μM) applied locally via a perfusion pencil (diameter 100 μm) increased the firing rate of 16% (20 of 126) of MPO neurons studied and decreased the firing activity of 25% (31 of 126) of them. The peptide had no effect in the remaining neurons.

The excitatory effect of AVP was associated with a depolarization of 3.6 ± 1.1 mV (n = 12) and was completely blocked by the AVPr1a antagonist SR49059 (30 nM) in all neurons studied (n = 12) (Fig. 3A). The firing rate of the neurons averaged 7.5 ± 1.8 Hz and 15.6 ± 3.7 Hz (n = 20, paired t test t(19) = -19.1; P = 1.4 × 10^-9) in control and during AVP incubation, respectively. In voltage-clamp experiments, AVP activated an inward current that averaged 21.7 ± 3.8 pA (n = 13) and in 4 of 13 neurons studied, the peptide also increased the frequency of spontaneous inhibitory postsynaptic currents (sIPSCs) (Fig. 3B). The frequency of sIPSCs was 1.4 ± 0.8 Hz and 5.4 ± 2.6 Hz (n = 4 paired t test t(3) = -4.2; P = 0.018) in control and during AVP incubation, respectively. Both effects were blocked by the AVPr1a antagonist SR49059 (30 nM) (Fig. 3B). Optogenetic stimulation of nearby MPO^AVP;hChR2 evoked IPSCs in 5 of the 12 neurons depolarized by AVP tested.

Figure 3.

AVP excites and depolarizes a group of MPO neuron by activating AVPr1a receptors. (A) Example of a response to local AVP application in an MPO neuron. AVP (1 µM) increased the spontaneous firing rate of the neuron from 1.1 Hz to 8.2 Hz (upper trace). The AVPr1a antagonist SR 49059 (30 nM) blocks the effect of AVP (lower trace). (B) AVP (1 µM) activates and inward current of 18 pA and increases the frequency of sIPSCs in an MPO neuron (upper trace). Both effects were abolished by preincubation with AVPr1a antagonist SR 49059 (30 nM) (lower trace). The neuron was held at -50 mV. (C) AVPr1a transcripts are present in MPO neurons excited by AVP. Representative gel from 20 recorded MPO neurons. The expected size of the PCR product is 402 bp. Negative (–) control was from a harvested cell without reverse transcription and positive control (+) was amplified using 1 ng of hypothalamic mRNA. AVPr1a transcripts were detected in 16 of the 20 neurons studied. AVP, arginine vasopressin; MPO, medial preoptic.

scRT-PCR experiments in 20 MPO neurons excited by AVP were carried out; AVPr1a transcripts were detected in 16 of them (Fig. 3C). In contrast, AVPr1b transcripts in any of the neurons studied were not detected. PACAP transcripts were detected in 15 of 20 neurons.

The inhibitory effect of AVP on the spontaneous firing activity of a different group of MPO neurons was associated with an increase in the frequency of spontaneous IPSCs recorded (Fig. 4A). The firing rate of the neurons averaged 7.7 ± 1.8 Hz and 2.4 ± 1.1 Hz (n = 31 paired t test t(30) = 24.3; P < 3.6 × 10^-11) in control and during AVP incubation, respectively. In this group of MPO neurons, whole-cell voltage-clamp recordings revealed that AVP (1 μM) incubation resulted in increased inhibitory and excitatory input but no inward current or outward current in all neurons studied (n = 25) (Fig. 4B). AVP increased the frequencies of both spontaneous IPSCs and spontaneous EPSCs but did not significantly affect their amplitudes (Fig. 4B). The AVPr1a antagonist SR49059 (30 nM) blocked the effect on sIPSCs almost completely but did not affect the increase in sEPSCs frequency (Fig. 4B,C). The AVPr1b antagonist TASP0390325 (20 nM) blocked the residual effect on sIPSCs frequency and abolished the effect on the sEPSCs frequency (Fig. 4B,D). Optogenetic stimulation of nearby MPO^AVP;hChR2 evoked IPSCs in 9 of the 16 neurons inhibited by AVP tested.

Figure 4.

AVP inhibits the firing activity of MPO neurons by increasing their inhibitory input. (A) Example of an inhibitory effect of local AVP application on the spontaneous firing rate of a MPO neuron. AVP (1 µM) decreased the spontaneous firing rate of the neuron from 6.1 Hz to 0.4 Hz (upper trace). Expanded fragments of the recording emphasizing the change in spontaneous IPSP frequency (asterisks, lower row). (B) AVP (1 µM) increases the frequency of both sEPSCs and sIPSCs in an MPO neuron (upper trace). In the presence of AVP, the frequency of sIPSCs increased from 0.5 Hz to 4.1 Hz and the frequency of sEPSCs from 1.1 Hz to 2.6 Hz. The AVPr1a antagonist SR 49059 (30 nM) blocked the AVP effect on sIPSCs but did not affect the effect on sEPSCs (second trace from top). In contrast, in the presence of AVPr1b antagonist TASP 0390325 (20 nM) AVP did not change the frequency of sEPSCs but increased the frequency of sIPSCs from 1.0 Hz to 2.5 Hz (third trace from top). Coapplication of the 2 antagonists abolished the effect of AVP on both sIPSCs and sEPSCs (bottom trace). The neuron was held at -50 mV. Bar charts summarizing the effects of AVP, AVPr1a, and AVPr1b antagonists on the frequency of (C) sIPSCs and (D) sEPSC in MPO neurons inhibited by AVP. Bars represent means ± SD of the normalized firing rate relative to the control. Data pooled from n = 25 neurons in each condition. (C) There was a statistically significant difference between groups as determined by 1-way ANOVA (F(3,31) = 21.8, P = 1.7 × 10^-7) followed by Tukey’s test relative to control; **statistical significance of P < 0.01. (D) There was a statistically significant difference between groups as determined by 1-way ANOVA (F(3,31) = 28.5, P = 1.2 × 10^-8) followed by Tukey’s test relative to control; **statistical significance of P < 0.01. AVP, arginine vasopressin; MPO, medial preoptic.

To test whether other mechanisms, besides the increased IPSP frequency, are involved in the inhibition of the MPO neurons’ firing rate, the effect of AVP in the presence of the GABA-A receptor antagonist gabazine was studied (5 µM). The antagonsist increased the firing rate of the neurons from 7.1 ± 1.2 Hz to 8.5 ± 1.0 Hz (n = 5 paired t test t(4) = -3.73; P = 0.024). In the presence of gabazine, local AVP (1 µM) incubation did not affect the spontaneous firing rate of the neurons (n = 5 paired t test t(4) = -0.33; P = 0.756).

The scRT-PCR experiments in 22 MPO neurons inhibited by AVP have not detected either AVPr1a or AVPr1b transcripts in any neurons studied. PACAP transcripts were detected in 14 of 22 neurons.

AVPr1a and AVPr1b Are Expressed Exclusively in MPO Neurons and in MPO Astrocytes, Respectively

The distribution of AVPr1a and AVPr1b transcripts among MPO neurons and astrocytes was studied using RNAscope technology. Neurons and astrocytes were identified as cells containing NeuN and Slc1a3 transcripts, respectively. In all preparations studied, AVPr1a and AVPr1b transcripts were present in different populations of cells (Fig. 5A,C). AVPr1a was present in ~20% of neurons and in < 2% of astrocytes. In contrast, AVPr1b was detected in ~27% of astrocytes and in < 3% of neurons (Fig. 5B,D).

Figure 5.

Subpopulations of MPO neurons and astrocytes express AVPr1a and AVpr1b transcripts, respectively. AVPr1a, AVPr1b, NeuN, and Slc1a3 transcripts visualized using RNAscope technology. (A) Differential interference contrast image (first from left) of MPO neurons and astrocytes in culture and the respective DAPI staining (gray) and the neuronal marker NeuN (blue) (second from left). AVpr1a transcripts (green) and AVPr1b transcripts (red) are present in distinct populations of cells as indicated by their superimposed image (third, right). AVPr1a transcripts are present in 1 neuron as indicated by the presence of NeuN transcripts (blue) in the same cell (right, arrow). (B) Differential interference contrast image (first from left) of MPO neurons and astrocytes culture and the respective DAPI staining (gray) and the astrocytic marker Slc1a3 (blue) (second from left). AVPr1a transcripts (green) and AVPr1b transcripts (red) are present in distinct populations of cells as indicated by their superimposed images (third from left). AVPr1b transcripts are present in one astrocyte (right, arrow). Bar chart summarizing the percentages of (C) neurons (NeuN positive cells) and (D) astrocytes (Slc1a3 positive cells) that also express AVPr1a, AVPr1b, NeuN, and Slc1a3 transcripts. Data were averaged from 40 fields of view for both staining conditions containing 2 to 7 neurons and 6 to 13 astrocytes each. Overall, 39 of 195 cells expressing NeuN also expressed AVPr1a and 72 of 267 Slc1a3-expressing cells were positive for AVPr1b. There was a statistically significant difference between groups as determined by 1-way ANOVA (C) (F(2,149) = 169.9, P = 2.4 × 10^–13 and (D) F(2,149) = 453.8, P = 1.1 × 10^–16) followed by Tukey’s test relative to the other 2 groups; **statistical significance of P < 0.01. MPO, medial preoptic.

AVP Activates Ca²⁺ Release From Intracellular Stores in Subpopulations of Neurons and Astrocytes by Activating AVPr1a and AVPr1b, Respectively

We then examined the effect of AVP on the intracellular Ca²⁺ concentrations [Ca²⁺]_i in dissociated MPO neurons and astrocytes loaded with fura-2 AM. Action potential propagation was blocked with TTX (1 μM). In 22% of neurons (33 of 150), AVP (1 μM) activated an increase in [Ca²⁺]_i, an effect that was fully blocked by preincubation with AVPr1a antagonist SR49059 (30 nM) (Fig. 6A). The AVPr1b antagonist TASP0390325 (20 nM) was without effect on the AVP responses (Fig. 6A). AVP (1 μM) activated [Ca²⁺]_i responses in 31% of astrocytes (61 of 197); however, their pharmacological profile was different. In astrocytes, the responses were blocked by preincubation with AVPr1b antagonist TASP0390325 (20 nM) but not by AVPr1a antagonist SR49059 (30 nM) (Fig. 6B).

Figure 6.

AVP activates [Ca²⁺]_i responses in neurons and astrocytes by activating AVPr1a and AVPr1b, respectively. (A) [Ca²⁺]_i responses to AVP (1 µM) from 3 acutely dissociated MPO neurons (black, blue, and green traces) and an average from 12 MPO neurons (red trace). The AVPr1a antagonist SR 49059 (20 nM) completely blocked the responses to AVP. The AVPr1b antagonist TASP 0390325 (10 nM) did not affect the AVP responses in neurons. (B) [Ca²⁺]_i responses to AVP (1 µM) from 3 MPO astrocytes (black, blue, and green traces) and an average from 10 MPO astrocytes (red trace). The AVPr1b antagonist TASP 0390325 (10 nM) completely blocked the responses. The AVPr1a antagonist SR 49059 (20 nM) did not affect the AVP response in astrocytes. (A, B) TTX (1 µM) was added to all extracellular solutions. AVP, arginine vasopressin; MPO, medial preoptic.

MPO and Intracerebroventricular AVP Injection Induce Hyperthermia

To assess the effect of AVP on CBT, the neuropeptide or aCSF (100 nL control) was injected bilaterally in the MPO (100 nL, 1 μM) or intracerebroventricularly (ICV) (300 nL, 10 μM). AVP induced a robust hyperthermia of up to 2.2 ± 0.9°C and 2.1 ± 0.8°C when applied intra-MPO or ICV, respectively, when compared with control injections (n = 6 mice in each condition) (Fig. 7A). The effect reached a maximum at ~3 hours after the injection of AVP and persisted for 6 to 7 hours postinjection. Interestingly, in both cases, AVP also significantly decreased the stress-induced hyperthermia that is observed during the first hour postinjections (Fig. 7A).

Figure 7.

AVP induces hyperthermia when injected centrally and hypothermia when injected subcutaneously. (A) Responses to bilateral intra-MPO injection (arrow) of AVP (1 µM, 100 nL, red trace), ICV injection of AVP (10 µM, 500 nL, blue trace), and bilateral intra-MPO aCSF (100 nL, black trace). AVP induced a hyperthermia of 2.1 ± 1.2 and 1.9 ± 1.4°C, respectively (1-way repeated measures ANOVA, F(2,357) = 113.5, P = 2.8 × 10^-6, followed by t tests for each time point, **P < 0.01, *P < 0.05). (B) Subcutaneous injection (arrow) of AVP (1 mM, 5 µL, blue trace) and aCSF (black trace, control). AVP induced a hypothermia of 2.8 ± 0.8°C (1-way repeated measures ANOVA F(1,238) = 47.3, P = 1.9 × 10^-8, followed by t tests for each time point relative to the aCSF injection, **P < 0.01). (A, B) The points represent averages ± SD (n = 6 mice) through the 10-hour recording period. Experiments were carried out in parallel in groups of 6 mice for each treatment. aCSF, artificial cerebrospinal fluid; AVP, arginine vasopressin; ICV, intracerebroventricular; MPO, medial preoptic.

To measure the effect of AVP when applied peripherally the mice were also injected subcutaneously (200μL, 10μM). The peptide induced a profound hypothermia of 3.1 ± 0.7°C when compared with saline injections (n = 6 mice in each condition), which lasted 4 hours (Fig. 7B).

Activation of MPO^AVP;hM3D(Gq) Neurons or AVP^hM3D(Gq) Neurons Induces Hyperthermia

Chemogenetic activation of AVP neurons in the MPO or of all AVP neurons in the brain using MPO^AVP;hM3D(Gq) mice or AVP^hM3D(Gq) mice (see Methods) resulted in hyperthermia of up to 2.1 ± 0.5°C and 1.8 ± 0.6°C, respectively (n = 6 mice in each condition) (Fig. 8A-C). Activation of AVP^hM3D(Gq) neurons in contrast with activation of MPO^AVP neurons only, resulted in a significant decrease in the initial, stress-induced phase of hyperthermia (Fig. 8C).

Figure 8.

Chemogenetic activation of AVP neurons induces hyperthermia. (A) Differential interference contrast (left) and fluorescence (right) images of an acute slice from MPO^AVP;hM3D(Gq) mouse indicating hM3D(Gq)-mCherry expression in the MPO. (B) Intraperitoneal injection (arrow) of CNO (20 mM, 3 µL, red trace) or vehicle (3 µL DMSO, black trace) in MPO^AVP;hM3D(Gq) mice. CNO induced a hyperthermia of 2.0 ± 0.5 (1-way repeated measures ANOVA, F(1,238) = 183.1, P = 7.2 × 10^-8, followed by t tests for each time point, **P < 0.01). (C) Intraperitoneal injection (arrow) of CNO (20 mM, 3 µL, red trace) or vehicle (3 µL DMSO, black trace) in mice AVP^hM3D(Gq) mice. CNO induced a hyperthermia of 1.8 ± 0.4 (1-way repeated measures ANOVA, F(1,238) = 60.5, P = 3.1 × 10^-7, followed by t tests for each time point, **P < 0.01). (D) Intraperitoneal injection (arrow) of CNO (20 mM, 3 µL, red trace) or vehicle (3 µL DMSO, black trace) in mice MPO^{AVP;VGAT-/-;hM3D(Gq)} mice. CNO induced a hyperthermia of 0.8 ± 0.3 (1-way repeated measures ANOVA, F(1,238) = 175.1, P = 5.2 × 10^-5, followed by t tests for each time point, *P < 0.05). The response to CNO was significantly smaller than that of MPO^AVP;hM3D(Gq) mice (1-way repeated measures ANOVA, F(1,238) = 220.5, P = 2.4 × 10^-8). (B-D) The points represent averages ± SD through the 10-hour recording period. Experiments were carried out in parallel in groups of 6 mice for each treatment. AVP, arginine vasopressin; CNO, clozapine N-oxide; DMSO, dimethyl sulfoxide; MPO, medial preoptic.

To assess the role played by GABA release from MPO AVP neurons in the hyperthermia induced by their chemogenetic stimulation, AVP neurons in MPO^{AVP;VGAT-/-;hM3D(Gq)} mice were activated. An increase of up to 0.8 ± 0.3°C relative to control was recorded; however, the responses was significantly smaller than that observed in MPO^AVP;hM3D(Gq) (Fig. 8B,D).

The effect of chemogenetic activation of AVP neurons in MPO^AVP;hM3D(Gq) mice or AVP^hM3D(Gq) mice was then tested on the LPS-induced fever response. Activation of MPO^AVP neurons had no effect on the LPS-induced fever (Fig. 9A), whereas activation of AVP neurons in the entire brain decreased CBT significantly in the initial and the late phases of fever (Fig. 9B). Also studied was the LPS-induced fever in AVP^VGAT-/- mice. Their fever response, relative to that of AVP^VGAT+/+ littermates, differed only during the late phase of the fever response when, surprisingly, CBT was below that of the control (Fig. 9B).

Figure 9.

Activation of MPO AVP neurons does not influence the fever response to LPS. (A) CBT responses to intraperitoneal injection (arrow) of LPS (0.03 mg/kg) + vehicle (3 µL DMSO) (black trace, control) and to LPS (0.03 mg/kg) + CNO (20 mM, 3 µL) (red trace) in MPO^AVP;hM3D(Gq). LPS induced a fever of 1.5 ± 0.3 and 1.5 ± 0.4°C, in control and in the presence of CNO, respectively (1-way repeated measures ANOVA, F(1,238) = 4.8, P = 0.24, followed by t tests for each time point, **P < 0.01, *P < 0.05). (B) CBT responses to intraperitoneal injection (arrow) of LPS (0.03 mg/kg) + vehicle (3 µL DMSO) (black trace, control) and to LPS (0.03 mg/kg) + CNO (20 mM, 3 µL, red trace) in AVP^hM3D(Gq) mice. Activation of AVP neurons decreases the initial hyperthermic response from 1.4 ± 0.2°C (control) to 0.7 ± 0.3°C as well as the late phase of fever from 1.8 ± 0.4°C (control) to 0.6 ± 0.3°C (1-way repeated measures ANOVA, F(1,238) = 35.1, P = 7.2 × 10^-6, followed by t tests for each time point, **P < 0.01). (C) LPS (intraperitoneal, 0.03 mg/kg) induces fever responses in AVP^GAT+/+ mice (black trace, control) and AVP^GAT-/- mice (red trace). The intermediate and late phases of the fever response are diminished in AVP^GAT-/- mice (1-way repeated measures ANOVA, F(1,238) = 52.1, P = 1.5 × 10^-6, followed by t tests for each time point, **P < 0.01). (A-C) The points represent averages ± SD (n = 6 mice) through the 10-hour recording period. AVP, arginine vasopressin; CBT, core body temperature; DMSO, dimethyl sulfoxide; LPS, lipopolysaccharide; MPO, medial preoptic.

Activation of AVP^hM3D(Gq) Decreases the Restraint-stress Induced Hyperthermia

Because activation of AVP^hM3D(Gq) neurons decreased the initial phase of hyperthermia (Figs. 8C and 9B), response attributed to handling stress, also tested was the effect on hyperthermia induced by restraint stress. Chemogenetic activation of AVP^hM3D(Gq) neurons decreased the hypothermia recorded during restraint stress (Fig. 10A). In contrast, AVP^VGAT-/- mice had a similar response to that of AVP^VGAT+/+ littermates. However, upon release from the restraint tube, the AVP^VGAT-/- mice displayed a hyperthermic rebound of up to 1.8 ± 0.6°C (n = 6) that was absent in control littermates or AVP^hM3D(Gq) mice (Fig. 10B).

Figure 10.

Hyperthermic responses induced by restraint stress. (A) CBT responses to 90-minute restrain stress (black bar) in AVP^hM3D(Gq) mice after intraperitoneal injection (arrow) of vehicle (3 µl DMSO, black trace, control) or CNO (20 mM, 3 µL, red trace). Activation of AVP neurons decreased the hyperthermia induced by restraint stress by 0.7 ± 0.4°C, in control and in the presence of CNO, respectively (1-way repeated measures ANOVA, F(1,238) = 39.1, P = 3.1 × 10^-7, followed by t tests for each time point, **P < 0.01). (B) CBT responses to 90-minute restrain stress (black bar) in AVP^GAT+/+ mice (black trace, control) and AVP^GAT-/- (red trace). AVP^GAT-/- mice displayed a hyperthermic rebound of 1.5 ± 0.6°C upon release from the restraint tube, which was not present in the control (1-way repeated measures ANOVA, F(1,238) = 131.9, P = 6 × 10^-8, followed by t tests for each time point, **P < 0.01). (A, B) The points represent averages ± SD (n = 6 mice) through the 10-hour recording period. AVP, arginine vasopressin; CBT, core body temperature; CNO, clozapine N-oxide; DMSO, dimethyl sulfoxide.

Discussion

In this study, a novel population of parvocellular AVP neurons in the MPO has been characterized . Interestingly, single-cell transcriptomics has also revealed a population of preoptic inhibitory AVP neurons (23). Similar populations of AVP neurons have been identified in the bed nucleus of the stria terminalis, the medial amygdaloid nucleus, and the lateral septum (24). Parvocellular AVP neurons have also been identified the MPO of lower vertebrates (25, 26). The MPO^AVP neurons had significantly different intrinsic membrane characteristics than those of magnocellular AVP neurons in the PVN and SON. Besides their much smaller membrane surface, as indicated by cell capacitance, MPO^AVP neurons had tonic firing characteristic with little adaptation and did not display a depolarizing afterpotential and bursting activity, the latter 2 being typical for magnocellular AVP neurons (27). Also found was that MPO^AVP neurons displayed thermosensitive firing activity, and that their optogenetic stimulation resulted in activation of IPSCs in postsynaptic MPO PACAP neurons as well as in other MPO^AVP neurons, indicating that these neurons send extensive local projections. Optogenetic activation of IPSCs was independent of activation of AVPr1a or AVPr1b as well as of glutamatergic neurotransmission indicating that MPO^AVP neurons release GABA. Also found was that optogenetic stimulation of MPO^AVP neurons had only inhibitory effects on the firing activity of other MPO neurons and that no inward or outward current were recorded. Thus, in spite of presence of AVP transcripts in MPO^AVP neurons, no functional evidence was found for AVP release from these neurons.

Local application of AVP had mostly inhibitory effects in MPO neurons, mediated by increased synaptic inhibition. The increase in sIPSC frequency was blocked by AVPr1a antagonists almost completely, with the remaining effect blocked by AVPr1b antagonists. However, in a small proportion of MPO neurons, AVP increased the firing activity by inducing a depolarization. This group of neurons express AVP1ra.

This study has revealed through using RNAscope in situ hybridization that the expression of AVPr1a was restricted to a subpopulation of MPO neurons, whereas AVPr1b transcripts were present only in a subpopulation of MPO astrocytes. AVP induced increases of [Ca²⁺]_i in subpopulations of neurons and astrocytes by activating AVPr1a and AVPr1b, respectively. Similar responses, mediated by AVPr1b activation, have been reported also in hippocampal and cortical astrocytes (28, 29). Also observed is that activation of AVPr1b was responsible for an increase in the sEPSC frequency and, to a lesser extent, in the frequency of sIPSCs. These findings suggest that activation of AVPr1b in astrocytes may modulate synaptic activity in the MPO, possibly by release of gliotransmitters such as glutamate and ATP and subsequent activation of P2X or glutamate receptors in neurons.

AVP has antipyretic and hypothermic effects; however, the mechanisms involved are not clear. The results indicate, as previously reported in the rat (30), that MPO or ICV administration of AVP induces hyperthermia, whereas peripheral administration of the peptide induces a profound hypothermia (12). Also shown is that activation of MPO^AVP neurons, or of all AVP neurons in the brain, does not induce hypothermia. In general, models of the central control of body temperature, associate an increase in the firing activity of preoptic thermosensitive neurons with a decrease in thermogenesis and, consequently, a decrease in CBT (21). Shown here is that activation of a population of thermosensitive MPO neurons has the opposite effect, suggesting that this tenet may need to be revised.

The antipyretic action of AVP was proposed based on several observations. The plasma and central release of AVP increases during the fever response (11). Infusion of AVP ICV or in the ventral septum, or electrical stimulation of the PVN reduces the fever response (31, 32). However, the mechanism of AVP antipyresis are not clear. The data indicate that MPO^AVP neurons do not influence the fever response, but activation of all brain AVP neurons results in a decrease of the late phase of fever. The LPS-induced fever was reduced in AVP^VGAT-/- mice, relative to control littermates, suggesting that GABA release from AVP neurons contributes to the increase in CBT, and thus it appears to oppose the AVP actions.

At the cellular level, optogenetic stimulation of MPO^AVP neurons resulted in GABA-mediated inhibition of nearby PACAP neurons. Recent studies have revealed that inhibition of the activity of PACAP neurons causes hyperthermia (14, 15). Thus, I propose that inhibition of PACAP neurons is the likely mechanism by which stimulation of MPO^AVP neurons causes hyperthermia. Indeed, stimulation of MPO^{AVP;VGAT-/-;hM3D(Gq)} neurons resulted in much reduced hypothermic response, thus proving the important role of GABA release in this action. It is also interesting that GABA infusion in the MPO results in increased CBT (33).

Local application of AVP, similarly had mostly GABA-dependent inhibitory effects on the firing activity of preoptic PACAP neurons; however, it also excited a population of AVPr1a-expressing PACAP neurons. Given the larger number of PACAP neurons inhibited by the peptide, it is likely that its net effect in the preoptic area, hence the responses observed were always hyperthermic (Fig. 11). Nevertheless, AVPr1a excitation of a population of PACAP neurons may decrease or prevent the hyperthermia caused by increased GABA release. Such a mechanisms may explain why AVP infusion (but not stimulation of MPO^AVP neurons) decreased the initial, stress-induced, phase of the hyperthermic responses. An alternative explanation may be that AVPr1a-expressing PACAP neurons may recruit different downstream neuronal networks to influence thermogenesis than the other preoptic PACAP neurons.

Figure 11.

Schematic representation of a preoptic pathway controlling thermogenesis Preoptic thermoregulatory PACAP neurons are glutamatergic and project to inhibitory interneurons in other brain regions that project to thermogenic neurons. PACAP neurons’s inhibition results in increased thermogenesis (diagram adapted from (15)). Preoptic AVP neurons are proposed to project locally and modulate the activity of thermoregulatory PACAP neurons. PACAP, pituitary adenylate cyclase-activating polypeptide.

AVP release increases in stressful conditions (2). Interestingly, it appears that such conditions (late-stage pregnancy, immobilization, osmotic stimulation) are associated with reduced febrile responses (11). This study found that both MPO and ICV administration of AVP reduced the stress-induced hyperthermia that follows any injection, response that is attributed to the stress caused by handling of the mice. Stimulation of all AVP neurons in the brain resulted in a similar decrease in the initial phase of the fever response. In contrast, stimulation of MPO^AVP only, did not affect the initial phase of fever, suggesting that their activation does not result in significant AVP release.

Also revealed is that activation of AVP neurons decreased the hyperthermia observed during restraint-induced stress. In contrast, the hyperthermia observed during prolonged restraint was similar in AVP^VGAT-/- mice and their control littermates, indicating that AVP rather than GABA is responsible for this action. Upon release from immobilization AVP^VGAT-/-, mice displayed a further hyperthermic phase, suggesting that GABA release is involved in the control of CBT following stressful conditions.

In summary, this study elucidates the cellular characteristics of a population of parvocellular AVP neurons in the MPO, their role in thermoregulation and reveals a novel mechanism of excitation downstream of astrocytic AVPr1b. Also provided is evidence that AVP release from neurons antagonizes stress-induced hyperthermia and the later phase of the fever response.

Glossary

Abbreviations

aCSF

artificial cerebrospinal fluid

AVP

arginine vasopressin

CBT

core body temperature

CNO

clozapine N-oxide

EPSC

excitatory postsynaptic current

ICV

intracerebroventricular

IPSC

inhibitory postsynaptic current

IPSP

inhibitory postsynaptic potential

LPS

lipopolysaccharide

MPO

medial preoptic

PACAP

pituitary adenylate cyclase-activating polypeptide

PVN

paraventricular nuclei

scRT-PCR

single-cell RT-PCR

Additional Information

Disclosures: The author has nothing to disclose.

Data Availability

The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.