Pituitary Adenylate Cyclase Activating Polypeptide Inhibits A₁₀ Dopamine Neurons and Suppresses the Binge-like Consumption of Palatable Food

Abstract

Pituitary adenylate cyclase-activating polypeptide (PACAP) binds to PACAP-specific (PAC1) receptors in multiple hypothalamic areas, especially those regulating energy balance. PACAP neurons in the ventromedial nucleus (VMN) exert anorexigenic effects within the homeostatic energy balance circuitry. Since PACAP can also reduce the consumption of palatable food, we tested the hypothesis that VMN PACAP neurons project to the ventral tegmental area (VTA) to inhibit A₁₀ dopamine neurons via PAC1 receptors and K_ATP channels, and thereby suppress binge-like consumption. We performed electrophysiological recordings in mesencephalic slices from male PACAP-Cre and tyrosine hydroxylase (TH)-Cre mice. Initially, we injected PACAP (30 pmol) into the VTA, where it suppressed binge intake in wildtype male but not female mice. Subsequent tract tracing studies uncovered projections of VMN PACAP neurons to the VTA. Optogenetic stimulation of VMN PACAP neurons in voltage clamp induced an outward current and increase in conductance in VTA neurons, and a hyperpolarization and decrease in firing in current clamp. These effects were markedly attenuated by the K_ATP channel blocker tolbutamide (100 μM) and PAC1 receptor antagonist PACAP_6–38 (200 nM). In recordings from A₁₀ dopamine neurons in TH-Cre mice, we replicated the outward current by perfusing PACAP_1–38 (100 nM). This response was again completely blocked by tolbutamide and PACAP_6–38, and associated with a hyperpolarization and decrease in firing. These findings demonstrate that PACAP activates PAC1 receptors and K_ATP channels to inhibit A₁₀ dopamine neurons and sex-dependently suppress binge-like consumption. Accordingly, they advance our understanding of how PACAP regulates energy homeostasis via the hedonic energy balance circuitry.

INTRODUCTION

The hypothalamus is known for its regulation of energy balance. The regions primarily responsible for this regulation are the ventromedial nucleus (VMN), paraventricular nucleus (PVN), arcuate nucleus (ARC), dorsomedial nucleus (DMN), and lateral hypothalamic area (LHA) (Berthoud and Morrison, 2008;Wagner, 2016). Within the VMN, expression of the transcription factor steroidogenic factor 1 (SF-1) impacts the development of this region, such that the deletion of SF-1 or leptin receptors (LepR) in SF-1 neurons leads to obesity in mice (Dhillon et al., 2006;Kim et al., 2011;Majdic et al., 2002). Pituitary adenylate cyclase-activating polypeptide (PACAP) is colocalized with SF-1 in the VMN, and the transcript levels of PACAP in the VMN depend on leptin signaling. It was also reported that the anorexigenic PACAP released by the VMN can mediate leptin’s anti-obesity action (Hawke et al., 2009). VMN SF-1/PACAP neurons project to the ARC and produce excitatory inputs onto pro-opiomelanocortin (POMC) neurons (Lindberg et al., 2013). As mentioned, PACAP is highly expressed in the VMN, and it is known that PACAP stimulates c-Fos expression in POMC neurons and increases POMC and melanocortin-4 receptor (MC4R) transcripts in the ARC (Mounien et al., 2009). We have previously reported that PACAP regulates energy homeostasis via stimulation of POMC neurons (Chang et al., 2020), this stimulation may link to the activation of melanocortin system through MC4R in the PVN resulting in the reduction of food intake and increase in energy expenditure (Cheng et al., 2011;Mounien et al., 2009). Beside the VMN, PACAP also colocalizes with a subpopulation of POMC neurons within the ARC (Dürr et al., 2007), a region that projects to both intrahypothalamic (e.g. VMN, DMN, LH, and PVN) and extrahypothalamic areas (e.g. the mesolimbic reward system as well as the nucleus of the tractus solitarius hunger and satiety sites) (Berthoud and Morrison, 2008). The type I PACAP receptor is termed the PACAP-specific (PAC1) receptor, which has high binding affinity and selectivity for PACAP. The PACAP/PAC1 receptor system is highly expressed in the VMN, PVN, ARC, DMN, and LHA, which further reinforces its important role in appetite regulation.

Previous reports have shown that PACAP suppresses appetite and feeding by intracerebroventricular (ICV) injection in mice (Morley et al., 1992), goldfish (Matsuda et al., 2005), and chick (Tachibana et al., 2003), as well as by intraperitoneal (IP) injection in goldfish (Matsuda et al., 2005) and mice (Vu et al., 2015). More specifically, PACAP regulates energy balance when injected into the VMN by reducing food intake, increasing core body temperature while decreasing interscapular brown adipose tissue (iBAT) triglyceride level and retroperitoneal white adipose tissue (WAT) (Resch et al., 2011;Resch et al., 2013;Rudecki and Gray, 2016). In addition, PACAP also plays a role in regulating glycemia as ICV injected PACAP stimulate hepatic glucose production in response to hypoglycemia (Rudecki and Gray, 2016). Finally, in contrast to the role PACAP plays in glucose-induced insulin secretion (Filipsson et al., 1999), chemogenetic stimulation of VMN PACAP neurons inhibits insulin secretion and increases glucose levels (Khodai et al., 2018).

Ion channels play an important role in the central regulation of energy and glucose homeostasis (Sohn, 2013;Sun and Feng, 2013). The K⁺ channel of particular interest to this study is the ATP-sensitive K⁺ (K_ATP) channel, which has proved to be critical in the central regulation of energy expenditure, food intake, and glucose homeostasis (Sohn, 2013;Sun and Feng, 2013). K_ATP (aka. Kir6) channels belong to a subfamily of the weak inward rectifier K⁺ channels that are classified into Kir6.1 and Kir6.2 subtypes (Hibino et al., 2010). These channels are negatively gated by ATP. The acute inhibitory effects of leptin and insulin are mediated by the activation of K_ATP channels in neuropeptide Y/Agouti-related protein (NPY/AgRP) neurons in the ARC (Qiu et al., 2014;van den Top et al., 2004), and for insulin these extend to ARC POMC and VMN SF-1 neurons as well (Hill et al., 2008;Klöckener et al., 2011;Plum et al., 2006;Williams et al., 2010). In the ARC, enhanced phosphatidylinositol3,4,5-triphosphate (PIP₃) signaling in POMC neurons due to knockdown of PIP₃ phosphatase increased K_ATP channel activity and led to diet-induced obesity and hyperphagia by blocking the excitatory effects of leptin in these cells (Plum et al., 2006). This indicates that the PIP₃ produced by phosphatidylinositol 3-kinase (PI3K) signaling is a positive allosteric modulator of K_ATP channels. Similarly, stimulation of insulin receptors in VMN SF-1 neurons that play a critical role in controlling energy homeostasis suppresses cell excitability via activation of PI3K and K_ATP channels; resulting in diet-induced obesity (Klöckener et al., 2011). Moreover, it has also been reported that leptin and insulin act on distinct subpopulations of POMC neurons (Williams et al., 2010).

PACAP has also been shown to regulate binge-eating behavior via site-specific actions such that the injection of PACAP into the nucleus accumbens (NAc) mimicked the actions of γ-aminobutyric acid (GABA) agonists by decreasing hedonic feeding while not affecting homeostatic feeding; whereas, the injection of PACAP into the VMN mimicked the actions of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) by reducing homeostatic feeding without altering hedonic feeding (Hurley et al., 2016;Hurley et al., 2019). This indicates that PACAP can regulate both the homeostatic and the hedonic energy balance circuitries (Liu et al., 2015). Hedonic feeding is defined as pleasure-driven to consume palatable food (i.e. reward-related) (Liu et al., 2015). Hedonic eating is linked to the midbrain dopaminergic (DA) system, which is known for its regulation of reward-related behaviors (Liu et al., 2015;Volkow et al., 2017) and its role in binge eating episodes (Avena and Bocarsly, 2012;Rada et al., 2005). There is a close relationship between hedonic feeding and reward/addiction that involves the mesolimbic dopamine (A₁₀) emanating from the ventral tegmental area (VTA) and projecting to the NAc. It is generally accepted that DA is a key neurotransmitter modulating rewards through this very projection (Volkow et al., 2017). There is also a neural circuit conveying information from the homeostatic energy balance circuitry to the hedonic circuitry, where leptin reduces excitatory synaptic strength onto orexin-expressing neurons originating in the LHA project to the VTA and may affect dopamine signaling and feeding behavior (Leinninger et al., 2009;Liu et al., 2017;Liu et al., 2015). In addition, a subpopulation of ARC POMC neurons project to the VTA, where they inhibit A₁₀ dopamine neurons and thereby participate in regulating stress-induced hypophagia, and anhedonia (Qu et al., 2020). On the other hand, while orexin-expressing LHA neurons project to the VTA to regulate neuronal activity by activating orexin receptors (OX1R) in the VTA and promote appetite for a high-fat diet (HFD) (Zheng et al., 2007), MCH-expressing LHA neurons project to the NAc to elevate both food incentive salience and reward reinforcement (Chung et al., 2009;Georgescu et al., 2005;Mul et al., 2011). Incentive salience refers to the ‘wanting’ module established via neural interactions between the VTA, NAc, and amygdala, which is driven by both physiological state and previously learned associations about a reward cue (Berridge, 1996;Liu et al., 2008). Motivated behaviors then prioritize attention stimuli based off of the anticipated rewards (Liu et al., 2015).

Given the PACAP/PAC1 receptor system and its known suppressive effect on the consumption of palatable food when injected into the NAc (Hurley et al., 2019), which is a region containing the terminal fields of A₁₀ dopamine neurons emanating from the VTA (Volkow et al., 2017) - coupled with the presence of PACAP binding sites in the VTA (Masuo et al., 1991) - this is suggestive of a scenario where PACAP may be capable of inhibiting A₁₀ dopamine neurons and suppressing binge-like consumption. We therefore systematically tested the hypothesis that VMN PACAP neurons project to the VTA to inhibit A₁₀ dopamine neurons via the activation of K_ATP channels and thereby suppress binge feeding.

MATERIALS AND METHODS

Animal Models

We bought tyrosine hydroxylase (TH)-cre mice (19–28 grams; 7–10 weeks old; Stock # 008601; C57BL/6 background) and PACAP-cre mice (18–25 grams; 7–10 weeks old; Stock #030155; C57BL/6 background) from Jackson Laboratories (Bar Harbor, ME, USA), and bred them in-house. We furnished all animals with food and water ad libitum and kept them under constant temperature (25°C) and kept on a 12hour light – 12hour dark schedule (lights on 06:00–18:00). We weaned the pups when they were 21 days old and genotyped them between 21–42 days of age using established PCR procedures. After weaning, we randomly grouped animals into one of two dietary conditions: a standard chow diet (Teklad Rodent Diet, Teklad Diets, Madison, WI, USA) from which animals received 18% of their calories from fat, 24% from protein, and 58% from carbohydrates, or a HFD (Research Diets, New Brunswick, NJ, USA) – from which animals received 45% of their calories from fat, 20% from protein and 35% from carbohydrates, and kept them there for 5–8 weeks prior to experimentation. Our previous studies have shown that consuming this HFD for this duration produces evident adiposity and glucose dysregulation (Fabelo et al., 2018;Hernandez et al., 2019;Qiu et al., 2018). All of our experiments were conducted in consultation with the NIH Guide for the Care and Use of Laboratory Animals and approved by our institutional animal care and use committee.

Surgical Procedures

We ovariectomized (OVX) female TH-cre, PACAP-cre, and wildtype mice after anesthetizing them with 2% isoflurane anesthesia approximately five days prior to experimentation. We focally injected TH-Cre and PACAP-Cre mice with adeno-associated viral vector (AAV) constructs after anesthetizing them with 2% isoflurane and placing them on a stereotaxic frame (Stoelting, Wood Dale, IL, USA). We made an incision to access the skull, and drilled a hole on one or both sides of the mid-sagittal suture so we could slowly lower the injection needle into the VTA (coordinates from bregma: AP, −2.1 mm; ML, ± 0.5 mm; and DV, −4.0 mm from dura) for TH-Cre mice or into the VMN (coordinates from bregma: AP, −0.6 mm; ML, ± 0.3 mm; and DV, −5.6 mm from dura) for PACAP-Cre mice. AAV constructs containing either cation channel rhodopsin-2 (ChR2) (pAAV-Ef1a-DIO hChR2(E123A)-EYFP.YFP.WPRE.jGH; 7.2 × 10¹² genomic copies/mL; 300 nL total volume; a gift from Karl Deisseroth; Addgene plasmid #35507) or its enhanced yellow fluorescent protein (eYFP) blank control (pAAV-Ef1a-DIO EYFP; 1.0 × 10¹³; 300 nL total volume; Addgene plasmid #27056), as well as the retrograde tracer Fluorogold (Conde K et al., 2016) (Fluorochrome, LLC, Denver, Colo.,USA; 0.2 μL) were injected bilaterally over 2 minutes. We kept the injection needle in place for 10 minutes to enable adequate diffusion from the tip, after which we slowly retracted it from the brain. Animals were ready for use 2–3 weeks after administration of the vectors and one week after delivery of the tracer.

We performed guide cannula implantations into the VTA of wildtype mice similar to the above-described procedure. Briefly, we secured an anesthetized animal into a stereotaxic frame and cut a midline incision through the scalp. We then drilled a hole in the skull and lowered a 26-gauge guide cannula (Plastics One, Roanoke, VA, USA) to a point 1 mm above the VTA (AP, −2.1 mm, ML, ± 0.5 mm, DV, −4.0 mm). We affixed the guide cannulas in place with C&B Metabond cement (Parkell, Edgewood, NY, USA) that we applied to the surgical field. Finally, we inserted a stylet into the guide cannula to ensure patency of the lumen. We allowed the animals to recover 1 week before beginning the experiment. When injecting, we inserted a needle until its tip extended 1 mm below the tip of the guide cannula shaft and into the area of interest. We possess considerable experience in accessing structures deep within the brain (Borgquist et al., 2013;Conde et al., 2016;Hernandez et al., 2019). We used only those animals where the injection sites or guide cannulas were correctly located within the VMN and VTA (Figure 1).

Figure 1.

Coronal schematics illustrating the accuracy of our bilateral AAV ChR2 injections in the VMN (A) and the guide cannula implantations and AAV eYFP injections in the VTA (B).

Drugs

We acquired all drugs from Tocris Bioscience/ R&D Systems (Minneapolis, MN, USA) unless otherwise specified. For electrophysiological experiments, the following drugs were used and their stock solutions prepared as detailed below: octahydro-12-(hydroxymethyl)-2-imino-5,9:7,10a-dimethano-10aH-[1,3]dioxocino-[6,5-d]pyrimidine-4,7,10,11,12-pentol (Tetrodotoxin, TTX; 1 mM stock solution in UltraPure H₂0, diluted with aCSF to 500 nM). PACAP_1–38 (100 μM stock solution in UltraPure H₂O, diluted with aCSF to 100 nM). PACAP_6–38 (200 μM stock solution in UltraPure H₂O, diluted with aCSF to 200 nM). Tolbutamide (1-butyl-3-(4-methylphenyl) sulfonylurea; 100 mM stock solution in UltraPure H₂0, diluted with aCSF to 100 μM).

For the behavioral experiments, we made a 150 μM stock solution of PACAP1–38 by solvating it in filtered saline, and then we administered it directly into the VTA (30 pmol; 0.2 μL). Estradiol benzoate (EB; Steraloids, Newport, RI, USA) was originally formulated as a 1 mg/mL stock solution in punctilious ethanol. We then added a known quantity of this stock solution to a volume of sesame oil that produced a final concentration of 20 μg/kg upon vaporization of the ethanol. We stored all aliquots of our stock solutions at either four or −20 °C until we needed them.

Brain Slice Preparation

On the day of experimentation, the animal was sedated with 32% isoflurane and euthanized by decapitation. We extracted the brain from the skull and further dissected it to obtain hypothalamic or mesencephalic slices in the coronal plane. The resultant block was then attached to a cutting platform that we inserted into a vibratome well containing an ice-cold, oxygenated (95% O₂, 5%CO₂) sucrose-containing cutting solution (NaHCO₃ 26; dextrose 10, HEPES 10; sucrose 208; KCl 2; NaH₂PO₄ 1.25; MgSO₄ 2; CaCl₂ 1; in mM). In this way, we were able to generate two to three slices (250μm) through the rostrocaudal range of the VTA. We transferred the slices to a storage chamber filled with oxygenated aCSF at room temperature, and kept there until we started recording.

Electrophysiology

We performed whole-cell patch clamp electrophysiological recordings from VTA neurons using biocytin-containing electrodes in mesencephalic slices prepared from intact male TH-Cre and PACAP-Cre mice. We maintained the slices during recordings in a chamber perfused with warmed (35°C), oxygenated aCSF in which we increased the CaCl₂ concentration to 2 mM. All drugs (diluted with aCSF) were suffused via peristaltic pump at a rate of 1.5 mL/min. We fabricated patch electrodes from borosilicate glass (World Precision Instruments, Sarasota, FL, USA; 1.5 mm OD) pulled on a P-97 Flaming Brown puller (Sutter Instrument Co., Novato, CA, USA), and filled with an internal solution that included the following (in mM): potassium gluconate 128; NaCl 10; MgCl₂ 1; EGTA 11; HEPES 10; ATP 1; GTP 0.25; 0.5% biocytin; pH adjusted to 7.3 with KOH; osmolality: 286–320 mOsm. Electrode resistances ranged from 3 to 8 MΩ.

We made visualized recordings on an Olympus BX51 W1 fixed stage microscope equipped with infrared differential interference contrast (DIC) video imaging. Potentials were amplified and current passed current through the electrode via Multiclamp 700A or B preamplifiers (Molecular Devices). Membrane currents and voltages underwent analog-digital transformation with Digidata 1550A or B interfaces (Molecular Devices) linked to pClamp 10.6 or 11.0 software. We monitored the access resistance, resting membrane potential (RMP), and input resistance during the entire recording period. If the access resistance changed more than 10% of its original value, we ended the recording. We conducted low-pass filtering of the currents at a frequency of 2 kHz. We calculated a liquid junction potential of −10 mV, which we corrected for during data analysis using pClamp software. We performed all recordings at a holding potential of −60 mV (in voltage clamp) or at rest (in current clamp).

To ascertain the postsynaptic effects caused by bath-applied PACAP, recordings were performed in slices from TH-Cre mice injected 2–3 weeks beforehand with an eYFP blank-containing AAV into the VTA. We initially established a baseline current-voltage (I/V) relationship from a holding potential of −60 mV using a ramp protocol (75 mV/sec; from −110 to −30 mV). For voltage clamp experiments, we generated baseline I/V relationships with 500 nM TTX on board. We added PACAP_1–38 [100 nM] (along with TTX) after the baseline I/V, and monitored the membrane current continuously until a new steady-state value was reached, at which time we generated a second I/V relationship. During the PACAP washout, we again monitored the membrane current until it came back to its original baseline value, and then ran the ramp I/V protocol one final time to confirm reversibility of the PACAP-induced effect. For current clamp experiments, we monitored the membrane potential and firing rate from rest without any TTX on board until new PACAP-induced steady-state levels were achieved, and then evaluated for 10–20 minutes more to allow for the return to baseline. To examine whether these postsynaptic effects are caused by PAC1 receptors and K_ATP channels, these same recordings were conducted in slices pretreated with PACAP_6–38 [200 nM] and tolbutamide [100 μM], respectively.

For the optogenetic experiments, we performed recordings in slices from PACAP-Cre mice that were given a ChR2-containing AAV into the VMN 2–3 weeks before the experiment. Once we encountered PACAP-expressing fibers (visualized with eYFP) impinging on VTA neurons, we ascertained functional synaptic connectivity by delivering high-frequency stimulation (10-ms pulses delivered at 20 Hz for 10 s) from a light-emitting diode (LED) blue light source (470 nm) controlled by a variable 2A driver (ThorLabs, Newton, NJ, USA) that directed the light path through the Olympus 40X water-immersion lens. For voltage clamp experiments, we generated baseline ramp I/V relationships with TTX (500 nM) on board. After doing this, we photo-stimulated to generate the slow outward current (in voltage clamp) or hyperpolarization and decrease in firing (in current clamp) Once a new steady-state current (or voltage) was achieved, we generated another ramp I/V, after which the membrane current (or voltage) was allowed to resettle to its original baseline value, at which time we executed a final ramp I/V protocol to affirm reversibility of the optogenetic stimulus.

Behavioral Studies

We conducted the behavioral studies in a four-station Comprehensive Lab Animal Monitoring System (CLAMS; Columbus Instruments, Columbus, Ohio, USA) as previously characterized and authenticated (Farhang B et al., 2010). We examined energy intake, meal pattern and energy expenditure in intact male and OVX female wildtype mice; and we injected these females with either EB (20 μg/kg; s.c.) or its sesame oil vehicle (1 mL/kg; s.c.) on alternate days during the experiment. First, we allowed the animals to acclimatize to the experimental chambers over three days. We weighed and handled the animals every afternoon before returning them to their respective chambers. After acclimation, we started a five-day monitoring phase. The binge feeding paradigm was performed as previously described (Hernandez et al., 2020;Rospond et al., 2015). Briefly, we exposed animals to the HFD from 4–5 pm for five successive days following the acclimation period, and quantified energy intake, meal pattern and energy expenditure for that 60-minute time frame. We provided standard chow ad libitum for the remaining 23 hours. Just before beginning the one-hour HFD exposure, we injected animals with PACAP_1–38 (30 pmol, 0.2 μL) or its 0.9% saline vehicle (0.2 μL) straight into the VTA.

Immunohistochemistry

We fixed slices from wildtype and TH-Cre mice overnight with 4% paraformaldehyde (4% PFM) in Sorenson’s phosphate buffer (pH 7.4). We then immersed them for three days in 20% sucrose dissolved in Sorensen’s buffer, which we replaced daily, followed by snap freezing in 2-methylbutane (EMD Millipore Corporation, Burlington, MA, USA) the next day. We cut coronal sections (20 μm) through the VTA on a cryostat and mounted them on chilled slides. We then washed these sections with 0.1 M sodium phosphate buffer (PBS; pH 7.4), and then incubated them overnight with either a monoclonal antibody directed against TH (Immunostar, Inc., Hudson, WI, USA; 1:4000 dilution) or a polyclonal antibody aimed against PACAP (Chang et al., 2020;Hu et al., 2020;Woodley et al., 2019) (Abcam, Cambridge, MA, USA; 1:200 dilution). We followed this up the next day with two 15-minute PBS washes, and then a two-hour application with either biotinylated goat anti-rabbit (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) or goat anti-mouse (Life Technologies, Carlsbad, CA, USA, 1:300 for both) secondary antibodies. Following another round of three 15-minute PBS washes, we did a last two-hour overlay with streptavidin-Alexa Flour (AF) 546 (Molecular Probes, Inc., Eugene, OR, PA, USA; 1:600), and concluded with a final series of three 30-minute PBS washes and cover slipping the slides. We evaluated the slides via fluorescence immunohistochemistry using a Zeiss Axioskop 2 Plus microscope (Carl Zeiss, Gӧttingen, Germany).

Statistical Analyses

We made comparisons amongst two groups with the Student’s t-test, Mann-Whitney U test or the signed rank test. When there were greater than two groups, we performed either the one-way or repeated measures, multifactorial analysis of variance (ANOVA; with rank transformation when necessary), and then the Least Significant Difference (LSD) test. We considered differences statistically significant when the alpha probability was less than 0.05.

RESULTS

Direct Injection of PACAP into the VTA Significantly Decreases the Consumption of Palatable Food in Wildtype Male but not Wildtype Female Mice

Previous reports have shown that PACAP significantly reduced food intake either by ICV (Matsuda et al., 1992;Mounien et al., 2009;Tachibana et al., 2003), IP (Matsuda et al., 2005;Vu et al., 2015), intra-VMN (Hurley et al., 2016;Resch et al., 2011;Resch et al., 2013), intra-PVN (Resch et al., 2013), or intra-ARC injection (Chang et al., 2020;Gastelum et al., 2021) under both ad libitum-fed and fasting conditions. We now wanted to ascertain whether PACAP could dampen the hedonic, binge-like consumption of palatable food. Here, we exposed wildtype mice to HFD for one hour every day for five days from 4–5 pm; with standard chow provided throughout the remainder of the time. Energy intake, meal pattern and energy expenditure were monitored around the clock. We found that as early as day 2 of the monitoring period, the animals started to exhibit clear signs of anticipatory locomotor activity immediately preceding the presentation of the HFD (Supplemental Figure 1). This is accompanied by an increase in metabolic heat production and O₂ consumption during the binge-feeding episode that is maintained for several hours after the HFD exposure (Figure 2A: repeated measures ANOVA/LSD, F = 3.69, df = 8, p < 0.002 and 2B: repeated measures ANOVA/LSD, F = 5.13, df = 8, p < 0.0001) (Supplemental Figures 2 and 3). This limited intermittent access to HFD causes a rapid, dramatic escalation in consumption (Figure 2C). When averaged across the five days (Figure 2D), the energy intake during these binge episodes far exceeds the consumption seen in strictly chow- or HFD-fed animals during the same time period (Hernandez et al., 2020), to such an extent that these otherwise chow-fed males end up eating over 35% of their daily caloric intake during that one-hour exposure (Figure 2E). PACAP administered directly into the VTA (30 pmol; 0.2 μL) suppresses this binge-like escalation (Figures 2C: repeated measures multifactorial ANOVA/LSD, F_day = 10.18, df = 4, p < 0.0001, F_PACAP = 16.87, df = 1, p < 0.0003, F_interaction= 0.42, df = 4, p < 0.80; 2D: Student’s t-test, t = 2.928, p < 0.006; 2E: Mann-Whitney U test, W = 121.0, p < 0.04) without affecting chow consumption during the other 23 hours (Figure 2F: Student’s t-test, t = −0.078175, p < 0.94), and this suppression is associated with reductions in meal bout size and duration (Figures 2G: Student’s t-test, t = 2.095, p < 0.05 and 2H: Student’s t-test, t = 3.137, p < 0.004). These results indicate that PACAP reduced binge-like consumption.

Figure 2.

PACAP reduces binge feeding accompanied by decreases in meal size and bout duration. A & B, Pronounced increases in energy expenditure triggered by the binge episode. C-E, The escalation of the daily consumption is significantly dampened by PACAP (30 pmol; VTA; n = 6) as compared to saline-treated controls (n = 6). F, No changes in chow consumption over the remaining 23 hours were observed. G & H, PACAP-induced changes in meal pattern. Symbols represent means and lines 1 SEM. The bars of the box-and-whisker plot represent the median, the 25^th and 75^th quartiles, whereas the lines depict the 5^th and the 95^th percentiles. #, p < 0.05 with respect to time (A & B) or day (C), repeated measures ANOVA/LSD; *, p < 0.05; with respect to saline, repeated measures multifactorial ANOVA/LSD (C), Student’s t-test (D, F-H), Mann-Whitney U-test (E).

Similar to male mice, OVX wildtype female mice displayed the anticipatory increase in locomotor activity manifested by increases in both metabolic heat production (Figure 3A) and O₂ consumption (Figure 3B) that developed during the binge hour and extended out to at least 4-hour post binge. The increase in metabolic heat production but not O₂ consumption was significantly enhanced by EB treatment (20 μg/kg; s.c.) compared to the oil-treated mice (Figures 3A: repeated measures multifactorial ANOVA/LSD, F_time = 19.22, df = 8, p < 0.0001, F_EB = 0.80, df = 1, p < 0.38, F_interaction= 0.31, df = 8, p < 0.97 and 3B: repeated measures multifactorial ANOVA/LSD, F_time = 17.79, df = 8, p < 0.0001; F_EB = 4.01, df = 1, p < 0.05; F_interaction= 0.50, df = 8, p < 0.86). Consistent with what we and others have shown previously (Richard et al., 2017; Hernandez et al., 2020), EB produced its characteristic reduction in binge consumption (Figure 3C) and the percent daily intake ingested during the binge episode (Figure 3D), with no effect on cumulative chow consumption during the remaining 23 hours (Figure 3E). In contrast to male mice, however, PACAP showed no effect on binge intake and increased the percent daily intake during the 1-hour HFD in EB-treated animals (Figures 3C: repeated measures multifactorial ANOVA/LSD, F_EB = 52.77, df = 1, p < 0.0001, F_PACAP = 0.05, df = 1, p < 0.83, F_interaction= 0.02, df = 1, p < 0.88 and 3D: rank-transformed, repeated measures multi-factorial ANOVA/LSD: F_EB = 19.65, df = 1, p < 0.0001, F_PACAP = 1.25, df = 1, p < 0.27, F_interaction= 5.24, df = 1, p < 0.03; one-way ANOVA/LSD, F = 9.56, df = 3, p < 0.0001). Interestingly, the post binge intake during other 23-hours in which OVX mice were exposed to ad libitum chow was significantly suppressed in animals treated both EB and PACAP (Figure 3E: repeated measures multi-factorial ANOVA/LSD: F_EB = 4.97, df = 1, p < 0.03, F_PACAP = 5.46, df = 1, p < 0.03, F_interaction= 17.61, df = 1, p < 0.0001; one-way ANOVA/LSD, F = 8.29, df = 3, p < 0.0001); suggesting a synergistic anorexigenic effect in OVX female mice. Overall, our results indicate a sexual dimorphism in the ability of PACAP to blunt the binge-like consumption of palatable food.

Figure 3.

PACAP shows no effect in OVX wildtype female mice. A & B, Pronounced increases in energy expenditure triggered by the binge episode; with heat production being potentiated by EB (A). C & D, EB (20 μg/kg; s.c.; n = 6) but not PACAP (30 pmol; VTA; n = 6) decreases binge intake as well as the percent of the daily intake consumed during the binge hour as compared to their sesame oil (n = 6) and saline controls (n = 6). E, Chow consumption over the remaining 23 hours is significantly reduced in EB- and PACAP-treated animals. Bars represent means and lines 1 SEM. Box-and-whisker plots illustrate the median, 25^th and 75^th quartiles, as well as the 5^th and 95^th percentiles. #, p < 0.05 with respect to time (A & B); *, p < 0.05; with respect to saline; ^, p < 0.05 with respect to sesame oil; repeated measures, multi-factorial ANOVA/LSD (A – C, E), rank-transformed, repeated measures, multi-factorial ANOVA/LSD (D).

VMN PACAP Neurons Project to the VTA

The fact that PACAP delivered directly into the VTA suppressed binge feeding behavior suggested that the peptide may also be capable of regulating the excitability of A₁₀ dopamine neurons of the hedonic energy balance circuitry. We then set to determine whether there are endogenous PACAP inputs to the VTA, and from where these inputs originated. To do this, we first performed retrograde tract tracing as described previously (Conde et al., 2016). We injected fluorogold (0.2 μL) into the VTA region of wildtype mice (Figure 4A), which is taken up by the nerve terminals in the VTA and transported back to the cell body. This revealed that the VTA receives strong innervation from the hypothalamus; including large clusters of neurons emanating from the VMN and LHA, with scattered inputs from the ARC (Figures 4B and 4C). Given the prominent population of PACAP neurons in the VMN (Chang et al., 2020;Hawke et al., 2009), we then stained for PACAP immunoreactivity and found a sizable population of immunopositive cells in this region (Figures 4D and 4F). This includes a significant number (~10%) of the fluorogold-filled somata exhibiting colocalization with immunostaining for PACAP (Figures 4E and 4F).

Figure 4.

Retrograde tract tracing revealing a subpopulation of PACAP-containing, VTA-projecting VMN neurons. A, The injection site seen within the VTA (5X; denoted by the arrow). B, Fluorogold labelling in the VMN and surrounding areas (5X). C, Fluorogold labelling in the VMN (10X). D, PACAP immunostaining in the VMN as visualized with AF546. E, Composite overlay. Bars represent means and vertical lines 1 SEM. D–E were also photographed at 10X.

In advance of being able to optogenetically stimulate PACAP nerve terminals in the VTA, we then injected a ChR2-containing AAV linked to an eYFP reporter into the VMN of PACAP-cre mice. We have shown previously that the ChR2 expression in the VMN is confined to the vast majority of PACAP-immunoreactive neurons (Chang et al., 2020). Two to three weeks after injection, we observed robust ChR2 expression in PACAP cell bodies in the VMN (Figures 5A and 5B), and equally intense ChR2 expression in PACAP nerve terminals in the VTA (Figure 5C and 5D). This effectively demonstrates connectivity between VMN PACAP neurons and postsynaptic targets in the VTA.

Figure 5.

Photostimulation of VMN PACAP neurons inhibits VTA neurons in PACAP-Cre mice. A & B, eYFP ChR2 reporter signal in the VMN (4X & 40X). C & D, eYFP ChR2 reporter signal in the VTA (4X & 40X). E, DIC image (40X) of a recorded VTA neuron. F-I, Optogenetic stimulation of VMN PACAP neurons produces an outward current associated with an increased slope conductance and reversal of polarity at ~ −90 mV in the vast majority of VTA neurons (n = 15). Arrows indicate where I/Vs were conducted.

VMN PACAP Neurons Induce a PAC1 Receptor-Mediated Outward Current Associated with Hyperpolarization in A₁₀ Dopamine Neurons via Activation of K_ATP Channels

Based off of our in vivo behavioral and tract tracing studies, we then hypothesized that PACAP inhibits A₁₀ dopamine neurons via activation of the K_ATP channels to suppress binge-like consumption of palatable food. We wanted to test this via optogenetic stimulation of PACAP nerve terminals during visualized whole-cell patch clamp recordings in mesencephalic slices through the VTA from male PACAP-cre mice (Figures 5A–5E). Optogenetic stimulation of PACAP nerve terminals in the VTA rapidly produced a powerful and reversible outward current in voltage clamp in the presence of TTX (500 nM; Figures 5F and 5H) in the majority of VTA neurons (Figure 5I). Identical to what we have explicitly shown previously (Chang et al., 2020), optogenetic stimulation in slices from PACAP-cre mice that were sham-injected or injected with an eYFP blank control AAV failed to elicit any postsynaptic response (Figure 5F and 5H). This outward current is associated with an increase in the slope conductance, and the polarity of the response reverses near the equilibrium potential for K⁺ (Figures 5G and 5H). The outward current is linked to a reversible hyperpolarization measured under current clamp, and associated with a decrease in firing (Figure 6A & 6D), which are completely reversed in the presence of the K_ATP channel blocker tolbutamide (100 μM; Figure 6B & 6D) and also significantly attenuated by the PAC1 receptor antagonist PACAP_6–38 (200 nM; Figure 6C–6D: one-way ANOVA/LSD, F = 32.22, df = 2, p < 0.0001). It should be noted that these inhibitory effects of endogenous PACAP released via optogenetic stimulation occur in A₁₀ dopamine neurons identified post-hoc via immunohistofluorescence as described previously (Wagner et al., 1998). The change in membrane potential (V_m) was significantly reversed in the presence of tolbutamide and also neutralized when PACAP_6–38 was introduced (Figure 6D).

Figure 6.

Optogenetic stimulation of VMN PACAP neurons in PACAP-Cre mice also hyperpolarizes VTA neurons and decreases their firing. A & D, Current clamp trace showing that photo-stimulation produces a reversible hyperpolarization (n = 12). B & D, The hyperpolarization is completely reversed in the presence of tolbutamide (100 μM; n = 6). C PACAP_6–38 (200 nM; n = 6) significantly attenuated the hyperpolarization. D, Bar represents means and lines 1SEM of the change in membrane potential (ΔV_m mV). *, p < 0.05 relative to PACAP alone, one-way ANOVA/LSD.

As a complimentary follow-up study, we also performed visualized patch clamp recordings in mesencephalic slices from male TH-cre mice injected with an eYFP-containing AAV injected into the VTA 2–3 weeks prior to experimentation in order to assess whether exogenously administered PACAP could replicate the inhibitory effect of the peptide on pre-identified A₁₀ dopamine neurons. The robust immunolabelling of TH perikarya was observed in the vast preponderance of eYFP-expressing neurons in the VTA (Figures 7A–7C). Voltage clamp recordings of these eYFP-containing A₁₀ dopamine neurons in the VTA (Figures 7D–7F) revealed that bath application of PACAP (100 nM) elicited a robust and reversible outward current (Figures 7G and 7N) in the considerable majority of A₁₀ dopamine neurons (Figure 7O) that was associated with an increase in slope conductance (Figures 7H & 7M). This inhibitory effect was completely negated in the presence of the tolbutamide (Figures 7I, 7J, 7M: one-way ANOVA/LSD, F = 12.04, df = 2, p < 0.0005, and 7N: one-way ANOVA/LSD, F = 18.48, df = 2, p < 0.0001) and the PAC1 receptor antagonist PACAP_6–38 (Figures 7K and 7L–7N).

Figure 7.

PACAP inhibits A₁₀ dopamine neurons in slices from male TH-cre mice via a PAC1 receptor-mediated activation of K_ATP channels. A, TH immunostaining in the VTA. B, eYFP signal seen in these A₁₀ dopamine neurons. C composite overlay. D & E, eYFP signal from A₁₀ dopamine neurons in VTA slices seen at 4X and 40X. F, DIC image of the recorded A₁₀ dopamine neuron seen in E. G-O, PACAP (100 nM; n = 12) produces a robust and reversible outward current in A₁₀ dopamine neurons that is associated with an increased K⁺ conductance and abrogated by PACAP_6–38 (n = 9) and tolbutamide (n = 6). Arrows indicate where I/Vs were conducted. Bars represent means and lines 1 SEM of the change in membrane current (M; pA) membrane slope conductance (M; nS) and current (N; pA). *p < 0.05 relative to PACAP alone, one-way ANOVA/LSD.

Similarly, the outward current is also linked to a reversible hyperpolarization measured under current clamp (Figure 8A) that significantly lowered the membrane potential (V_m; Figure 8B: Student’s t-test, t = 6.961, p < 0.001), and was associated with a decrease in firing (Figure 8C: signed rank test, test statistic = 2.10819, p < 0.04). Collectively, these data effectively demonstrate that VMN PACAP neurons inhibit A₁₀ dopamine neurons within the hedonic energy balance circuitry via PAC1 receptor-mediated activation of K_ATP channels.

Figure 8.

The outward current obtained in A₁₀ dopamine neurons from TH-cre mice is also associated with a hyperpolarization and decrease in firing. A, Current clamp trace illustrating that bath application of PACAP_1–38 produces reversible hyperpolarization that significantly lowers the membrane potential (V_m; B) and suppresses firing (C). B & C, Bars represent means and lines 1 SEM (n = 6). *, p < 0.05 relative to PACAP alone, Student’s t-test (B), signed rank test (C).

DISCUSSION

The results generated from this project demonstrate that VMN PACAP neurons can dampen the hedonic, binge-like consumption of palatable food in a sex-dependent manner through inhibition of A₁₀ dopamine neurons. These findings are based on the following observations: 1) PACAP administered directly into the VTA suppresses binge-like consumption in male but not female mice, 2) tract tracing reveals that VMN PACAP neurons innervate the VTA, 3) optogenetic stimulation of VMN PACAP neurons induces a prominent, reversible outward current in VTA A₁₀ dopamine neurons that is associated with an increase in slope conductance as well as hyperpolarization and decrease in firing, 4) the inhibitory effect of PACAP endogenously released from VTA nerve terminals on A₁₀ dopamine neurons is mimicked by exogenous administration of the peptide.

PACAP Administered Directly into the VTA Suppresses Binge-Feeding Behavior in a Sexually Differentiated Manner

Here we uncover the role of PACAP and its predominant receptor PAC1 in the regulation of palatable food consumption. Consistent with what we and others have shown previously (Bello et al., 2003;Hernandez et al., 2020;Kreisler et al., 2017), intermittent exposure to a palatable HFD dramatically accelerates binge-like consumption. Our results further demonstrate that intra-VTA PACAP administration significantly suppresses binge-like consumption that is associated with reductions in meal size and bout duration in male but not female mice. These findings coincide with other reports that not only ICV or IP but also direct injections of PACAP into discrete hypothalamic nuclei causes a reduction in energy intake under various energy states (Chang et al., 2020;Gastelum et al., 2021;Hurley et al., 2016;Matsuda et al., 2005;Morley et al., 1992;Resch et al., 2011;Resch et al., 2013;Tachibana et al., 2003;Vu et al., 2015).

Our findings are also in agreement with the fact that microinjections of PACAP into the NAc mimic the actions of GABA agonists by reducing hedonic but not homeostatic feeding (Hurley et al., 2016), and acute microinjection of PACAP into the NAc attenuates palatable food consumption while microinjection of PACAP into the hypothalamus does not attenuate palatable feeding (Hurley et al., 2019). This further reinforces the idea that PACAP regulates binge-eating behavior in a site-specific manner. As mentioned, PACAP regulates the homeostatic energy balance circuitry in a pleiotropic, energy balance state-dependent manner – it is powerfully anorexigenic in sated animals with unlimited access to food and water (Morley et al., 1992;Mounien et al., 2009;Resch et al., 2011;Resch et al., 2013;Tachibana et al., 2003), but this can be reversed under conditions of negative energy balance (Gastelum et al., 2021). In addition, PACAP increases O₂ consumption, metabolic heat production, and core body temperature (Chang et al., 2020;Resch et al., 2011;Resch et al., 2013). Reduction in body weight in these animals is consistently observed after PACAP administration, and pair-feeding studies suggest that the weight loss is not entirely due to food intake but also involves factors like increased activity and metabolic rate (Hawke et al., 2009;Resch et al., 2013).

The lack of effect of PACAP in both oil- and E₂-treated OVX females was a great surprise to us, given that E₂ potentiates the PACAP-induced excitation of anorexigenic POMC neurons and the corresponding reduction in homeostatic feeding (Chang et al., 2020;Gastelum et al., 2021) and attenuates the N/OFQ-induced inhibition of A₁₀ dopamine neurons and the corresponding reduction in hedonic feeding (Hernandez et al., 2020). E₂ per se did decrease binge-like consumption, which is consistent with the fact that it reduces the hedonic drive to consume palatable food (Richard et al., 2017). This also coincides with previous reports that E₂ significantly reduces energy intake, body weight gain, and meal size as well as frequency (Geary and Asarian, 1999;Hernandez et al., 2019;Roepke et al., 2010), and this effect was seen at E₂ doses ranging from 2.0 and 20.0 μg, respectively (Butera et al., 2010;Geary and Asarian, 1999). The dose of EB used in the present study (20 μg/kg) administered every other day over the behavioral experiment produces circulating levels typically seen during the late follicular phase or proestrus – a stage wherein energy intake is at its nadir in humans and non-human primates (Johnson et al., 1994;Kemnitz et al., 1989)- and consistently produces anorexigenic effects in guinea pigs and other rodent species (Fabelo et al., 2018;Hernandez et al., 2019;Hernandez et al., 2020;Kellert et al., 2009;Washburn et al., 2013). Thus, the ability of E₂ to diminish hedonic feeding must occur through a neural substrate other than VMN PACAP neurons.

Binge-Feeding Behavior and the Relation to Dopamine Level and Food Addiction

Although the hedonic and homeostatic energy balance circuitries are regionally distinct, there are neural connections between the two that allow the conveyance of information from one to the other. This neural connection is established through the orexin-expressing LHA neurons in the LHA that innervate the VTA to regulate A₁₀ dopamine neurons (Liu et al., 2017;Liu et al., 2015). These MCH and orexin expressing neurons project to the NAc as well as the VTA, to promote appetite and modulate reward (Chung et al., 2009;Georgescu et al., 2005;Mul et al., 2011;Zheng et al., 2007). These two energy balance circuitries are similar in that they are both modulated by leptin action, such that leptin directly regulates a population of leptin receptor (LepR)-expressing inhibitory neurons in the LHA and decreases feeding and body weight (Leinninger et al., 2009). In addition, leptin is known for its heterogenous effects in which it inhibits ARC NPY/AgRP neurons via activation of K_ATP channels (van den Top et al., 2004), and excites both POMC and SF-1/PACAP neurons via activation of TRPC5 channels (Dhillon et al., 2006;Hawke et al., 2009;Qiu et al., 2010) to regulate food intake. On the other hand, leptin also modulates behaviors associated with dopamine reward circuit and decreases hedonic feeding via administration of leptin into the VTA (Hommel et al., 2006;Morton et al., 2009).

A₁₀ dopamine neurons play an important role on binge eating behaviors. In this study, we showed that intermittent access to HFD dramatically increases binge-like consumption, and we think that this binge eating may be associated with increased dopamine release from A₁₀ dopamine neurons. This was indeed the case in rats exposed to the intermittent dietary-induced binge eating (DIBE) schedule with a sugar solution (10% sucrose) demonstrate increased accumbens dopamine levels (Rada et al., 2005). However, with repeated exposure to the food reward, the dopamine response habituates, and this habituation of dopamine signaling can be attenuated if animals are food restricted (Bassareo and Di Chiara, 1999;Bello et al., 2003). Therefore, receiving a palatable food during periods of calorie restriction serves to not only increase accumbal dopamine release, but to prevent habituation of dopamine signaling due to repeated stimulus exposure. Recent studies also report that short-term consumption of palatable food (i.e. binge feeding) is mediated by the strengthening of excitatory synaptic transmission onto VTA dopamine neurons, and injection of insulin (an anorexigenic hormone) into the VTA suppresses these excitatory inputs and abolishes food consumption (Liu et al., 2016). This is congruent with what we show in this paper; that administration of PACAP, which is also an anorexigenic hormone, directly into the VTA significantly suppresses binge-like consumption.

As additional proof that dopaminergic stimulation augments hedonic feeding, prolonged dopamine replacement therapy in Parkinson’s disease (PD) enhances dopamine receptor activation, and in parkinsonian animal models daily treatment with levodopa (L-Dopa) significantly escalates chocolate consumption in a time-dependent manner (Mineo et al., 2019). This implies that the restoration of dopaminergic neural activity can promote binge feeding. Likewise, in PD patients, inadequate control of dopamine replacement therapy with either dopamine agonists or high doses of L-Dopa may manifest in the form of behavioral addictions. (Beaulieu-Boire and Lang, 2015).

Our findings reveal that PACAP is without effect on binge intake in OVX female mice but significantly suppresses binge intake in male mice. This provides yet another example of sex differences in the regulation of energy homeostasis. Although the prevalence of obesity is similar between men and women, women seem to have a greater risk of developing eating disorders and extreme obesity (Hoek, 2006;Yang and Colditz, 2015). The investigation of food addiction in young women using the Yale Food Addiction Scale (YFAS) revealed that participants who possessed a greater number of food addiction features exhibited greater activation in the amygdala, dorsolateral prefrontal cortex, and medial orbitofrontal cortex (OFC) during exposure to a palatable food cue such as a picture of a milkshake (Gearhardt et al., 2011). In addition, binge-prone female rats were found to tolerate significantly high levels of foot shock for access to Oreo cookies than their binge-resistant counterparts, confirming the “continued use despite negative consequences” criteria for substance dependence (Oswald et al., 2011). This is consistent with other lines of evidence illustrating that women have a reduced ability to control food desire, higher cortical and limbic activation when presented with visual, gustatory, olfactory cues as well as increased susceptibility to episodes of food-craving compared to men (Uher et al., 2006;Wang et al., 2004;Wang et al., 2009;Weingarten and Elston, 1991).

VMN PACAP Neurons Inhibits A₁₀ Dopamine Neurons Via the PAC1 Receptor-Mediated Activation of K_ATP Channels

PACAP is highly expressed in the VMN, and colocalizes extensively with SF-1 in VMN neurons (Chang et al., 2020;Hawke et al., 2009). These SF-1/PACAP neurons project to the ARC and synapse with anorexigenic POMC neurons (Chang et al., 2020;Lindberg et al., 2013). Presently, we have demonstrated through both anterograde and retrograde tract tracing that a subpopulation of VMN PACAP neurons projects to the VTA and forms functional synapses with A₁₀ dopamine neurons. We show that the optogenetic stimulation of VMN PACAP neurons results in a robust but reversible outward current that is abrogated by PAC1 receptor antagonism and K_ATP channel blockade. While PACAP_6–38 has long been touted as a PAC1 receptor antagonist, it is also capable of blocking VPAC2 receptors (Robberecht et al., 1992). PACAP binds the PAC1 receptors with high affinity and selectivity, whereas both PACAP and vasoactive intestinal peptide (VIP) bind the VPAC2 receptor with comparable affinity (Vaudry et al., 2009). While both PACAP and VIP inhibit food intake in goldfish and chick, only the appetite suppressant effect of PACAP is antagonized by PACAP_6–38 (Matsuda et al., 2005;Tachibana et al., 2003) This indicates that the endogenous PACAP released by photo-stimulation binds and activates PAC1 receptors in A₁₀ dopamine neurons within the hedonic energy balance circuit, which leads to the opening of K_ATP channels and K⁺ efflux. The fact that this was done in the presence of TTX confirms that it was a direct effect of PACAP itself, and not through some intermediary neurotransmitter such as GABA that could conceivably activate K_ATP channels and inhibit the A₁₀ dopamine neurons. This is consistent with the inhibitory effects on these cells previously reported for insulin and leptin that also occur through the activation of K_ATP channels (Hommel et al., 2006;Morton et al., 2009). It is also congruent with other examples of PACAP-induced activation of K_ATP channels in vascular smooth muscle, as well as PACAP-induced increases in K_ATP channel subunit expression in the striatum, that accounts in part for the respective vasodilatory and neuroprotective effects of the peptide (Koide et al., 2014;Wang et al., 2008). The outward current and corresponding hyperpolarization elicited by optogenetic stimulation of VMN PACAP neurons were fully replicated when PACAP is perfused exogenously. In both instances the hyperpolarization caused by K⁺ efflux through K_ATP channels greatly suppresses neuronal firing in A₁₀ dopamine neurons. Therefore, the PACAP/PAC1 receptor system can impact the neuronal activity of A₁₀ dopamine neurons and the function of hedonic energy balance circuitry.

The activation of K_ATP channels play an important role in the regulation of energy homeostasis. It has been demonstrated in vitro that PIP₃ activates K_ATP channels directly (MacGregor et al., 2002). PIP₃ is the product generated from PI3K activity, which also is closely associated with downstream signaling molecules like protein kinase B, or Akt. (Denley et al., 2009). The PI3K/Akt pathway can be elicited by the activation of the PAC1 receptor via G_q signaling (Chang et al., 2020;Vaudry et al., 2009), which implies that PACAP may couple to K_ATP channels directly through this pathway. Furthermore, we have previously shown that the PAC1 receptor-mediated activation of TRPC5 channels in POMC neurons occurs through the G_q-coupled stimulation of PI3K/PLC/PKC signaling (Chang et al., 2020). Within the hedonic energy balance circuitry, activation of another important energy sensing molecule, AMP-dependent protein kinase (AMPK), potentiates K_ATP channel-mediated currents (Wu et al., 2017) and the PACAP/PAC1R-mediated activation of K_ATP channels inhibits A₁₀ dopamine neurons, which is associated with hyperpolarization and decrease in firing. This inhibition leads to decreased dopamine neuronal activity and suppresses binge-like consumption. Within the homeostatic energy balance circuitry, however, elevated PIP₃ levels in POMC neurons due to the selective knockout of PIP₃ phosphatase causes activation of K_ATP channels to induce hyperphagia and sexually dimorphic diet-sensitive obesity (Plum et al., 2006). Similarly, augmented AMPK activity in POMC neurons promotes a switch in PAC1 receptor coupling from TRPC5 channels under ad libitum-fed conditions to K_ATP channels under fasting conditions; resulting in a hyperphagic response seen during refeeding (Gastelum et al., 2021). Taken together, this clearly demonstrates that precise control of PI3K and AMPK signaling as well as K_ATP channel activity is crucial for the maintenance of energy homeostasis.

While we certainly acknowledge that bath application and intra-VTA administration of exogenous PACAP do not entirely replicate the physiological effects of endogenous PACAP released from nerve terminals of VMN PACAP neurons in the VTA, the convergence of the evidence set forth here leads us to conclude that PACAP regulates hedonic feeding via inhibition of A₁₀ dopamine neurons. It does this by postsynaptically activating PAC1 receptors and K_ATP channels in these cells in a sex-dependent manner.

Highlights.

• VMN PACAP neurons project to the VTA

• These neurons inhibit the excitability of A₁₀ dopamine neurons

• This occurs via activation of PAC1 receptors and K_ATP channels

• PACAP suppresses binge-feeding behavior in a sexually dimorphic way

• Thus, PACAP inhibits the hedonic feeding circuitry and the intake of palatable food