Energy state alters regulation of proopiomelanocortin neurons by glutamatergic ventromedial hypothalamus neurons: pre- and postsynaptic mechanisms

Abstract

To maintain metabolic homeostasis, motivated behaviors are driven by neuronal circuits that process information encoding the animal’s energy state. Such circuits likely include ventromedial hypothalamus (VMH) glutamatergic neurons that project throughout the brain to drive food intake and energy expenditure. Targets of VMH glutamatergic neurons include proopiomelanocortin (POMC) neurons in the arcuate nucleus that, when activated, inhibit food intake. Although an energy-state-sensitive, glutamate circuit between the VMH and POMC neurons has been previously indicated, the significance and details of this circuit have not been fully elucidated. Thus, the goal of the present work was to add to the understanding of this circuit. Using a knockout strategy, the data show that the VMH glutamate→POMC neuron circuit is important for the inhibition of food intake. Conditional deletion of the vesicular glutamate transporter (VGLUT2) in the VMH results in increased bodyweight and increased food intake following a fast in both male and female mice. Additionally, the targeted blunting of glutamate release from the VMH resulted in an ∼32% reduction in excitatory inputs to POMC cells, suggesting that this circuit may respond to changes in energy state to affect POMC activity. Indeed, we found that glutamate release is increased at VMH-to-POMC synapses during feeding and POMC AMPA receptors switch from a calcium-permeable state to a calcium-impermeable state during fasting. Collectively, these data indicate that there is an energy-balance-sensitive VMH-to-POMC circuit conveying excitatory neuromodulation onto POMC cells at both pre- and postsynaptic levels, which may contribute to maintaining appropriate food intake and body mass.

NEW & NOTEWORTHY Despite decades of research, the neurocircuitry underlying metabolic homeostasis remains incompletely understood. Specifically, the roles of amino acid transmitters, particularly glutamate, have received less attention than hormonal signals. Here, we characterize an energy-state-sensitive glutamate circuit from the ventromedial hypothalamus to anorexigenic proopiomelanocortin (POMC) neurons that responds to changes in energy state at both sides of the synapse, providing novel information about how variations in metabolic state affect excitatory drive onto POMC cells.

INTRODUCTION

The ventromedial hypothalamus (VMH) is known to modulate consummatory behaviors (1, 2), is activated during feeding (3), and lesions to the VMH can result in increased body mass and obesity (1, 4, 5), indicating a potent role for this region in the inhibition of food intake. VMH neurons express the vesicular glutamate transporter type 2 (VGLUT2), indicating that cells in this brain region release glutamate onto efferent targets, which includes anorexigenic proopiomelanocortin (POMC) neurons in the arcuate nucleus (ARC) of the hypothalamus (6, 7). It is possible that glutamate neurons in the VMH inhibit food intake, at least in part, by stimulating POMC neurons.

Terminals from POMC cells extend throughout the brain and modulate satiety, glucose metabolism, and bodyweight through the release of peptide and amino acid transmitters (8–12). Using chemo- and optogenetic approaches, the activity of POMC cells has been bidirectionally modulated to affect commensurate changes in food intake and bodyweight (13, 14). In vivo, the activity of POMC neurons changes in response to variations in circulating levels of peripheral signals that reflect the energy state of the animal such as leptin, insulin, and glucose (15–19). In addition to regulation by circulating factors, it is now recognized that amino acid transmitters rapidly modulate metabolic brain circuits including POMC cells in response to changes in energy state (20–22). For example, recent work indicates that the activity of POMC cells is altered immediately upon sensory detection of food (23, 24). Thus, it seems likely that fast synaptic transmission onto POMC cells acts to regulate the activity of these cells before peripheral postprandial signals are able to reach the brain.

Indeed, our recent work (21, 25), as well as that from others (26–29), provides compelling evidence that GABAergic transmission onto POMC cells plays an important role in modulating that activity of these cells and affecting metabolic activity. Despite recent work indicating an important role of hypothalamic glutamate signaling in the modulation of metabolic activity (30–32), there is much to learn about the origin of glutamatergic inputs onto POMC cells. Indeed, it is possible that the release of, and response to, glutamate may be changed by variations in energy state.

To reduce these gaps in our knowledge, we assessed how energy state in male and female mice affects VMH-to-POMC cell glutamate synapses. To accomplish this, we used brain region-specific deletion of the vesicular glutamate transporter (VGLUT2), or expression of channelrhodopsin (ChR2) in glutamatergic VMH neurons, followed by ex vivo electrophysiology. The present results indicate that conditional deletion of glutamate release from the VMH results in increased overall bodyweight, food intake following an overnight fast, and reduced excitatory inputs onto POMC cells. Further, results show that this input is strengthened upon food intake and that drops in leptin during caloric deficit contributes to a switch in POMC AMPA receptors from the GluR2-lacking, calcium-permeable state to a GluR2-containing, calcium-impermeable state that reduces the overall excitable drive onto POMC cells. These studies provide a better understanding of the interaction between two hypothalamic brain regions during changes in energy state and support the hypothesis that synaptic amino acid transmission plays a critical role in modulating satiety circuits during energy surfeits.

MATERIALS AND METHODS

Animals

All experiments were approved by the Colorado State University Animal Care and Use Committee and adhered to The Guide for the Care and Use of Laboratory Animals set forth by the National Institutes of Health. Mice expressing the enhanced green fluorescent protein (eGFP) driven by the Pomc promoter (POMC-eGFP, stock number 009593), mice expressing Cre-recombinase driven by the vesicular glutamate transporter type 2 (Slc17a6) gene promoter (VGLUT2-Cre, stock number 028863), and mice with Slc17a6 floxed on both alleles (VGLUT2^flox/flox, stock number: 012898) were originally obtained from Jackson Laboratories (Bar Harbor, ME). Transgenes and floxed alleles were detected with standard polymerase chain reaction (PCR) techniques ran on ear clips from the mice. Primer sequences used for PCR were those suggested by the supplier under information for the stock number for each mouse line.

Male and female mice were used for all experiments. Animals were housed on a 12-h light/dark schedule and had ad libitum access to tap water and chow except as noted for the fast and fast/refeed experiments. Individual mice were assigned to one of three experimental conditions to determine if energy state affects VMH-to-POMC glutamatergic circuitry: 1) ad libitum, mice had continuous and uninterrupted access to standard rodent chow until euthanasia; 2) overnight fast, 18–20 h before euthanasia mice were transferred to a fresh cage containing the same bedding and enrichment as ad libitum mice, but no chow was present; 3) refed, mice were fasted overnight for 18–20 h. At ∼0800 on the day of euthanasia, these mice were given access to ∼10 g of fresh rodent chow for 2 h before euthanasia. Electrophysiological experiments were conducted in mice aged 9–12 wk. No differences were seen between male and female mice for electrophysiological experiments. Thus, data from both sexes were pooled for data presentation and analysis. Mice were separated by sex for bodyweights and food consumption during a refeeding after fasting.

Stereotaxic Surgery for In Vivo Gene Delivery

Stereotaxic microinjections were performed on mice at 6–8 wk of age. A deep anesthetic plane was induced with isoflurane. Mice were then placed into a stereotaxic apparatus (David Kopf instruments) fitted with a nose cone for delivery of isoflurane throughout the procedure. To conditionally interrupt glutamate release from the VMH, an adeno-associated virus (AAV) encoding Cre-recombinase (AAV1.hSyn.Cre.WPRE.hGH; Penn Vector Core, 100 nL) was injected bilaterally into the VMH (from bregma: A/P, −1.40; M/L, ±0.42; D/V, −5.70) of VGLUT2^flox/flox mice. For bodyweight studies, control animals were VGLUT2^flox/flox mice that received AAV that did not encode Cre (100 nL of AAV9-CAG-GFP; 7E12GC/mL; Addgene Cat. No. 37825). To drive expression of ChR2 in glutamatergic neurons within the VMH, an AAV encoding Cre-dependent ChR2 [AAV9.EF1.dflox.hChR2(H134R)-mCherry.WPRE.hGH; 7.24E13 GC/mL; Addgene, Cat. No. 20297] was injected into the VMH of VGLUT2-Cre mice (100 nL each side). For all injections, virus was infused over the course of 1 min and the needle was left in place for 150 s following the end of solution delivery.

Fluorescent In Situ Hybridization

For tissue collection, mice were anesthetized with sodium pentobarbital, perfused with 10% sucrose in water followed by 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS). Brains were then removed and stored overnight in 4% PFA at 4°C. Coronal sections (50 μm) containing the VMH were prepared on a vibratome in ice-cold diethylpyrocarbonate (DEPC)-treated PBS and processed for fluorescent in situ hybridization using a vGLUT2 exon 2 probe following the protocol previously detailed (9, 33, 34) After the in situ hybridization protocol was completed, slices were mounted and cover slipped using Aqua Poly/Mount (Polysciences, Warrington, PA). Images were acquired on a Zeiss 800 laser-scanning confocal microscope.

Electrophysiology

For whole cell electrophysiology experiments, mice were decapitated following induction of a deep anesthetic plane with isoflurane. Brains were removed and placed into ice-cold artificial cerebral spinal fluid (aCSF) consisting of (in mM) 126 NaCl, 2.5 KCl, 1.2 MgCl₂^6H₂O, 2.4 CaCl₂^2H₂O, 1.2 NaH₂PO₄, 11.1 glucose, and 21.4 NaHCO₃, and bubbled with 95% O₂ and 5% CO₂. Sagittal slices (240 µm) containing the arcuate nucleus were cut using a Leica VT1200S vibratome (Leica Microsystems Inc.). Brain slices were allowed to rest for at least 1 h at 37°C in aCSF containing the N-methyl-d-aspartate (NMDA) receptor blocker MK-801 (15 µM). Slices were then transferred to a recording chamber and perfused with oxygenated 37°C aCSF at a flow rate of ∼2 mL/min. Recording electrodes (1.8–2.2 MΩ) were pulled with a Narishige PC-10 vertical pipette puller (Narishige International). Whole cell, voltage-clamp recordings were made using an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA). Electrophysiological data were collected at 10 kHz and filtered at 5 kHz using Axograph X software running on a Windows 7 operating system. Recordings were excluded from analysis if the series resistance exceeded 15 MΩ or changed significantly over the course of the experiment. To record miniature and light-evoked EPSCs, the internal recording solution contained (in mM) the following: Cs-methanesulfonate 135, MgCl₂ 2, HEPES 10, EGTA 1.1, Mg-ATP 2.5, Na₂-GTP 0.3, phosphocreatine 10. The pH was adjusted to 7.3 with KOH. mEPSCs were recorded in the presence of the voltage-gated sodium channel blocker, tetrodotoxin (500 nM), and the GABA_A receptor antagonist, picrotoxin (50 μM), at a holding potential of −60 mV. Paired-pulse ratio experiments were conducted at a holding potential of −60 mV in the presence of picrotoxin (50 μM). POMC cells were visually identified for recording based on the transgenic expression of eGFP. For AMPA receptor rectification experiments, 3.5 mM QX-314, 5 mM tetraethylammonium, and 0.1 mM spermine were added to the pipette solution. EPSCs were confirmed to be AMPA receptor mediated by bath perfusion of 10 µM 6,7-dinitroquinoxaline (DNQX).

Miniature excitatory postsynaptic potentials (mEPSCs) were continuously collected in 60-s epochs for 12–15 min and three to five epochs were combined for frequency and amplitude analysis. mEPSC events were detected by sliding an event template over the raw data trace, and data were then visually inspected to exclude spurious events. Paired-pulse ratio experiments were performed by light activating ChR2-expressing cells with a 2 ms 470-nm light pulse delivered through the objective centered over the patched cell in the arcuate nucleus with an interstimulus interval of 50 ms. A 470-nm LED (Thorlabs) driven by an LEDD1B driver (Thorlabs) triggered through the TTL output on an ITC-18 computer interface board (HEKA Instruments Inc.) was used. Paired-pulse ratio was assessed every 20 s for 10 min, and 10–20 consecutive sweeps were averaged for data analysis. The stimulus intensity was reduced at the beginning of the experiment so that the first EPSC was ∼50% of the maximal amplitude. To measure AMPA receptor rectification, ChR2-expressing VMH terminals were light stimulated with a 2 ms 470-nm LED light pulse while making voltage clamp recordings from POMC cells at holding potentials ranging from −80 to +40 mV in 20-mV increments. Ten consecutive stimuli were delivered at each holding potential at an interstimulus interval of 45 s. Peak amplitudes of light-evoked EPSCs were averaged over 10 consecutive stimuli for data analysis. The value of the peak amplitude of AMPA receptor-mediated EPSCs was normalized as a percent of the value at −60 mV. The rectification index represents the peak amplitude at 40 mV divided by the value at −60 mV.

Following electrophysiological recordings, slices were transferred to a 24-well plate and stored overnight at 4°C in potassium phosphate-buffed saline (KPBS) containing 4% paraformaldehyde (PFA). The next day, slices underwent three 15-min washes in KPBS and were mounted on slides for subsequent imaging. Confocal images were acquired on either a Zeiss 800 or 880 confocal microscope to confirm injection location and viral spread.

Drugs

MK-801, picrotoxin, and DNQX were purchased from Sigma-Aldrich (St Louis, MO). Drugs were prepared in DMSO as 1,000 or 10,000× concentrates and diluted in aCSF to achieve desired concentration immediately before use.

Bodyweight Measurements and Food Intake after an Overnight Fast

Bodyweight measurements from group-housed, VGLUT2^flox/flox mice commenced weekly at 6 wk of age. To conditionally delete glutamate release from the VMH, an AAV encoding Cre recombinase or a control virus was injected into the VMH at 8 wk of age. Weekly bodyweights were recorded until mice were 17 wk old. Because littermates were used for both experimental conditions and lived in the same cage, it was not possible to determine food intake for individual mice or conditions throughout the weight study. As an alternative measure of relative food intake, we assessed the amount of chow eaten in response to metabolic challenge. Mice were placed individually in fresh cages with ad libitum access to water 1 h before the dark period was set to begin. After 18 h, mice were given access to ∼10 g of fresh chow, and the amount of food consumed in a 2-h period was collected.

Statistics and Data Analysis

Data were compared using one-way and two-way ANOVAs or unpaired t tests. Tukey’s multiple comparisons test was used following one-way ANOVAs to make comparisons between individual groups. For bodyweight studies, a two-way repeated measures (RM) ANOVA was used. Individual points on figures represents a single mouse, or in the case of electrophysiology experiments, a single cell. No more than three cells were recorded from a single mouse. All data were analyzed using GraphPad Prism v. 8. Data are presented as mean ± SEM, and differences were considered significant at P < 0.05.

RESULTS

Deletion of Glutamate Release from the VMH Results in Increased Bodyweight and Food Consumption

Increasing the activity of VMH glutamate neurons reduces food intake (35), and mice with an embryonic deletion of vGlut2 from VMH neurons are heavier than controls (36). The previous deletion studies used a constitutive knockout approach to disrupt VMH glutamate release. Here, we wanted to test the hypothesis that deletion of VMH glutamate release from adult animals would also result in increased bodyweight. To test this hypothesis, mice were weighed weekly beginning at 6 wk of age. Then, at 8 wk of age, VGLUT2^flox/flox mice received intra-VMH injections of an AAV encoding Cre recombinase to conditionally blunt glutamate release from the VMH. This conditional deletion strategy was used to avoid compensatory effects that might occur with constitutive deletion approaches. Controls were also VGLUT2^flox/flox mice, but these animals received a control virus encoding eGFP.

First, to validate that this approach results in deleted vGlut2 expression in the VMH, we performed fluorescent in situ hybridization for vGlut2 mRNA in VGLUT2^flox/flox mice that had been injected with the virus encoding Cre. Consistent with previous literature (37), Vglut2 mRNA is highly expressed in the VMH of wildtype mice (Fig. 1A). However, vGlut2 mRNA signal was notably reduced in the VMH of mice injected into the VMH with a virus encoding Cre (Fig. 1B), suggesting that this approach likely reduces or abolishes vGLTU2 production and thus, glutamate packaging and release from VMH cells.

Figure 1.

Targeted viral expression of Cre recombinase reduces vGlut2 expression in the VMH. A: representative image from a wild-type mouse showing expression of vGlut2 mRNA in the VMH and the thalamus. B: representative image from a VMH VGLUT2 KO mouse showing reduced vGlut2 mRNA in the VMH. Note the intact vGlut2 expression in the thalamus following viral expression of Cre in the VMH. D, dorsal; L, lateral; KO, knockout; VGLUT2, vesicular glutamate transporter; VMH, ventromedial hypothalamus; 3V, third ventricle.

In both male [control, n = 9 mice; VMH VGLUT2 KO, n = 12 mice; F(11,209) = 4.67; P < 0.001; two-way RM ANOVA; Fig. 2A] and female [control, n = 9 mice; VMH VGLUT2 KO, n = 10 mice; F(11,187) = 3.05; P = 0.009; two-way RM ANOVA; Fig. 2B] mice, conditional deletion of VMH vGlut2 resulted in a significant increase in bodyweight relative to control-injected mice, indicating that glutamate release from the VMH onto efferent targets is required for normal bodyweight maintenance.

Figure 2.

Conditional deletion of glutamate release in the VMH results in increased bodyweight and increased food consumption. A: summary data showing that when glutamate is deleted from the VMH at 8 wk of age (indicated by arrow), male mice show increased bodyweight compared with eGFP-injected controls [F(11,209) = 4.67; P < 0.001; two-way RM ANOVA]. B: summary data showing that when glutamate release is deleted from the VMH at 8 wk of age (indicated by arrow), female mice show increased bodyweight compared with eGFP-injected controls [F(11,187) = 3.05; P = 0.009; two-way RM ANOVA]. C: summary data showing that in male mice, deletion of VMH glutamate release results in increased food consumption during a 2-h refeed following an overnight fast (t = 2.10; P = 0.04; unpaired t test). D: summary data showing that in female mice, deletion of VMH glutamate release results in increased food consumption during a 2-h refeed following an overnight fast (t = 2.61; P = 0.01; unpaired t test). All data are presented as mean ± SEM. *P < 0.05. eGFP, enhanced green fluorescent protein; KO, knockout; O/N, overnight; RM, repeated measures; VGLUT2, vesicular glutamate transporter; VMH, ventromedial hypothalamus.

Bodyweight studies were conducted in group-housed mice with both control and VMH VGLUT2 KO littermates housed within the same cages. Thus, daily or weekly food intake measures were not possible. As an alternative means to evaluate food intake, we fasted mice overnight (18 h) and assessed food intake over a 2-h period the next morning. Consistent with the bodyweight phenotype, both male (control, n = 11; VMH VGLUT2 KO, n = 10; t = 2.10; P = 0.04; unpaired t test; Fig. 2C) and female (control, n = 11; VMH VGLUT2 KO, n = 14; t = 2.61; P = 0.01; unpaired t test; Fig. 2D) mice consumed more chow during this refeeding paradigm. Together, it is clear that conditional disruption of glutamate release from the VMH is sufficient to disturb normal energy balance regulation.

VMH cells send projections throughout the brain, and thus, phenotypes due to deletion of vGlut2 from the VMH are likely mediated by disruption of VMH glutamate release onto multiple target cells in various brain regions. However, VMH cells send direct, energy-state-sensitive glutamatergic terminals onto anorexigenic POMC cells (7). Thus, it is possible that the increased body mass observed using our conditional approach, as well as previous studies where vGlut2 was constitutively deleted from VMH cells (38), could be mediated, in part, by a loss of glutamate signaling onto POMC cells.

Action-Potential-Independent Glutamate Release onto POMC Neurons Is Not Altered by Feeding State but Is Diminished by vGlut2 Deletion from the VMH

The VMH is considered as anorexigenic nucleus (1), as stimulation of these cells reduces food intake (39) and deletion of these cells results in increased bodyweight (38). Meanwhile, reducing the activity of POMC cells increases food intake (13, 14). Considering the energy-state-sensitive glutamatergic input from the VMH to POMC cells (7), we hypothesized that VMH glutamate neurons exert their anorexigenic actions, in part, through stimulation of POMC cells. To assess this, we recorded excitatory postsynaptic currents in POMC cells in the presence of TTX (mEPSCs) in POMC-eGFP mice and mice with targeted disruption of glutamate release from the VMH (VMH VGLUT2 KO mice). VMH VGLUT2 KO mice were again generated by injections of an AAV encoding Cre recombinase into the VMH of VGLUT2^flox/flox;POMC-eGFP mice. Two to five weeks after injection, whole cell voltage clamp experiments were performed on POMC cells. There was a significant main effect of conditional VMH VGLUT2 deletion on mEPSC frequency [F(1,78)_genotype = 8.821; P = 0.004; two-way ANOVA; Fig. 3C], suggesting that the VMH delivers a portion of excitatory inputs onto POMC cells regardless of metabolic state.

Figure 3.

Conditional deletion of glutamate release in the VMH reduces the frequency of mEPSCs onto POMC cells. A: representative mEPSC recordings from POMC cells from control mice that were fed ad libitum (top), fasted overnight (middle) or refed for 2 h following an overnight fast (bottom). B: representative mEPSC recordings from POMC cells from VMH VGLUT2 KO mice. Mice were fed ad libitum (top), fasted overnight (middle), or refed for 2 h following an overnight fast (bottom). C: summary data showing the frequency of mEPSCs recorded from POMC cells during varying energy states in both control and VMH VGLUT2 KO mice. Conditional deletion of VMH glutamate release significantly reduced mEPSC frequency across all energy states [F(1,78) = 8.821; P = 0.004; two-way ANOVA]. D: summary data showing the amplitude of mEPSCs recorded from POMC cells during varying energy states in control mice. There was a significant effect of both VMH VGLUT2 KO [F(1,78) = 10.13; P = 0.002; two-way ANOVA] and feeding state [F(2,78) = 5.76; P = 0.004; two-way ANOVA]. Data are presented as mean ± SEM. **P < 0.01. KO, knockout; mEPSCs, miniature excitatory postsynaptic potentials; O/N, overnight; POMC, proopiomelanocortin; VGLUT2, vesicular glutamate transporter; VMH, ventromedial hypothalamus.

POMC neuronal activity changes based on the metabolic state of the animal (23, 24, 40) and GABAergic transmission onto POMC cells is affected by the animal’s energy state (18, 21, 25). To determine if excitatory inputs onto POMC cells are also affected by energy state, mEPSCs were examined in differing conditions. Mice were fed ad libitum, fasted overnight, or refed for 2 h following an overnight fast. Energy state did not affect mEPSC frequency [F(2,78)_{energy state} = 0.01; P = 0.99; two-way ANOVA; Fig. 3C], further indicating (18, 41) that spontaneously active glutamatergic inputs to POMC neurons coming from the VMH and other regions are insensitive to energy state. There was not a significant interaction between genotype and energy state [F(2,78)_interaction = 0.63; P = 0.53; two-way ANOVA; Fig. 3C], and post hoc analysis (Tukey’s multiple comparison) did not reveal significant differences in mEPSC frequency between individual energy states. Interestingly, both VMH VGLUT2 KO [F(1,78)_genotype = 10.13; P = 0.002; two-way ANOVA; Fig. 3D] and feeding state [F(2,78)_{energy state} = 5.76; P = 0.004; two-way ANOVA; Fig. 3D] did significantly alter mEPSC amplitude, suggesting that postsynaptic AMPA receptor complexes undergo changes in response to variations in the animal’s energy state and that synapses originating from the VMH may target these energy-state-sensitive AMPA receptors. There was not a significant interaction between genotype and energy state [F(2,78)_interaction = 1.63; P = 0.20; two-way ANOVA; Fig. 3D].

Food Intake Increases the Release Probability for Glutamate from VMH Neuron Terminals Synapsing on POMC Neurons

Although energy state did not significantly affect action potential independent glutamate release onto POMC cells, we have previously shown that the sources of spontaneous and action-potential-mediated inputs to POMC neurons can differ from each other and be distinctly regulated (25). Therefore, we examined whether evoked glutamate release from the VMH onto POMC neurons was sensitive to energy state. To evoke glutamate release only from VMH neurons, an AAV encoding Cre recombinase-dependent ChR2 was injected into the VMH of POMC-eGFP mice expressing Cre in Vglut2-expressing neurons (VGLUT2-Cre, Fig. 4A). We then made voltage clamp recordings from POMC cells while light-stimulating VMH-originating terminals in the arcuate nucleus (Fig. 4B). In recordings from slices with robust expression of ChR2, we reliably observed (112/156 cells) light-evoked EPSCs in POMC cells (Fig. 4C, top). Light-evoked EPSCs were blocked by the AMPA receptor antagonist DNQX (Fig. 4C, bottom), confirming the glutamatergic nature of this connection.

Figure 4.

The release probability of glutamate onto POMC cells from VMH terminals is increased during a refeed after an overnight fast. A: representative confocal image demonstrating expression of ChR2 (red) in the VMH and POMC-eGFP cells in the ARC (green). B: schematic of recording parameters. ChR2 was virally expressed in VMH glutamatergic neurons, and these terminals were activated by 470-nm blue light while making voltage clamp recordings from POMC cells. C: stimulation of ChR2-expressing VMH terminals resulted in an EPSC in POMC cells (top) that was blocked by the AMPA receptor antagonist, DNQX (bottom). D: representative traces showing the paired pulse ratio of light-evoked EPSCs recorded from POMC cells during varying energy states. E: summary data (mean ± SEM) showing that refeeding significantly reduces the paired-pulse ratio of light-evoked EPSCs [F(2,56) = 4.13; P = 0.02; one-way ANOVA]. *P < 0.05, Tukey’s multiple comparisons. ARC, arcuate nucleus; C, caudal; ChR2, channelrhodopsin; D, dorsal; DNQX, dinitroquinoxaline; eGFP, enhanced green fluorescent protein; EPSC, excitatory postsynaptic potential; KO, knockout; O/N, overnight; POMC, proopiomelanocortin; R, rostral; V, ventral; VGLUT2, vesicular glutamate transporter; VMH, ventromedial hypothalamus.

POMC neuronal activity is increased during feeding (23, 40), as is the activity of VMH neurons (3, 42). Therefore, we next tested the hypothesis that VMH glutamate neurons increase glutamate release onto POMC cells during feeding. To test this hypothesis, ChR2 was virally expressed in the VMH, and paired-pulse ratio experiments were conducted in POMC cells from mice in differing energy states while light-stimulating VMH terminals (Fig. 4B) at an interstimulus interval of 50 ms to assess the paired pulse ratio (PPR). Analysis of the PPR indicated a significant effect of energy state (ad lib, n = 19 cells; fasted, n = 25 cells; refed, n = 15 cells; F(2,56) = 4.13; P = 0.02; one-way ANOVA; Fig. 4, D and E). Further analysis revealed a significant increase in release probability of glutamate at VMH-to-POMC synapses between fasted and refed states (P = 0.02; Tukey’s multiple comparison). These data collectively suggest that glutamate signaling between the VMH and POMC cells contributes to increasing the activity of POMC cells during feeding.

AMPA Receptor Complexes at VMH-to-POMC Synapses Are Remodeled during Fasting

The results presented in Fig. 4 indicate that glutamatergic terminals originating from the VMH increase release probability onto POMC neurons upon refeeding after an overnight fast. Interestingly, we also found evidence that POMC AMPA receptors show altered responses to glutamate under differing energy states (Fig. 3D), suggesting that these receptors undergo changes that alter their unitary conductance in response to refeeding. The AMPA receptor is a heteromer containing compositions of subunits GluR1–4 (43). AMPA receptor complexes containing the GluR2 subunit are calcium impermeable and have a relatively lower single-channel conductance. Conversely, GluR2-lacking AMPA receptors are calcium permeable, have higher conductance, and contribute to LTP (44–46). Recent work indicates that AMPA receptor subunit switching occurs in the nucleus accumbens following exposure to drugs of abuse or a junk food diet (47, 48), in the hippocampus following learning and memory tasks (49), and on POMC cells during variations in energy state (41). Thus, we hypothesized that this mechanism of postsynaptic plasticity occurs at VMH-to-POMC synapses in response to changes in energy state. GluR2-containing AMPA receptors show a linear current-voltage (I-V) relationship, whereas GluR2-lacking AMPA receptors display inward rectification in I-V relationships (45). To generate an I-V relationship and determine the subunit composition of POMC AMPA receptors, voltage clamp recordings were made from POMC cells while light-stimulating ChR2-expressing VMH terminals and adjusting the holding potential of POMC cells from −80 mV to +40 mV in 20-mV increments (Fig. 5, A and B). The rectification of AMPA receptor-mediated EPSCs was assessed by the rectification index (RI), a well-validated tool determining the subunit composition of AMPA receptors (50, 51). The RI was calculated by dividing the peak amplitude of light-evoked EPSCs at a holding potential of +40 mV by the peak amplitude of light-evoked EPSCs at −60 mV. Thus, the more an I-V rectifies, the lower the RI. We observed a significant effect of energy state on the RI (ad lib, n = 10 cells; fasted, n = 10 cells; refed, n = 11 cells; F(2,28) = 11.53; P = 0.002; one-way ANOVA; Fig. 5C). In recordings from ad lib mice, the I-V relationship rectified, indicating GluR2-lacking, calcium-permeable AMPA receptors at VMH-to-POMC synapses; however, a linear I-V relationship was observed in recordings from fasted mice, suggesting the presence of GluR2-containing, calcium impermeable AMPA receptors at these synapses during fasting that returns to the rectifying, GluR2-lacking state during refeeding (Fig. 5).

Figure 5.

Fasting increases switches AMPA receptors on POMC cells from a calcium-permeable state to a calcium-impermeable state. A: representative traces showing light-evoked EPSCs recorded from POMC cells at holding potentials of +40 mV and −60 mV during varying energy states. B: I-V relationship of EPSCs recorded from POMC cells from mice fed ad libitum, fasted overnight, or re-feed for 2 h following an overnight fast. The value of peak amplitude EPSCs was normalized as percent of the peak at −60 mV. C: summary data showing that fasting increases the RI of POMC AMPA receptors at VMH glutamate synapses [F(2,28) = 11.53; P = 0.002; one-way ANOVA]. ***P < 0.001; **P < 0.01 by Tukey’s multiple comparisons test. EPSCs, excitatory postsynaptic potentials; I-V, current-voltage; O/N, overnight; POMC, proopiomelanocortin; RI, rectification index; V_hold, holding potential; VMH, ventromedial hypothalamus.

Leptin Administration Blocks Fasting-Induced Changes in POMC AMPA Receptor Composition

Next, we wanted to try to discern a mechanism by which fasting might cause the shift in AMPA subunit composition. Among other changes, fasting is associated with a drop in circulating leptin levels (18, 52), and leptin administration can switch AMPA receptors on hippocampal neurons from GluR2 containing to GluR2 lacking (53). Thus, it could be that the fasting-associated decrease in leptin causes a shift toward calcium-impermeable AMPA receptors on POMC neurons. This possibility is supported by a recent study where the investigators found that leptin administration led to the maintenance of calcium-permeable AMPA receptors on POMC neurons during fasting (41). In the previous study, glutamate release was evoked electrically and thus, the source of the inputs was likely diverse (29). Here, we evoked glutamate release selectively from VMH-originating terminals and asked whether leptin could alter the response to glutamate inputs from this source. Mice expressing ChR2 in VMH vGlut2 neurons were fasted overnight and administered leptin (5 mg/kg) or saline into the intraperitoneal space 2 h before euthanasia and slice preparation. Consistent with untreated fasted mice, saline-treated fasted mice had an RI of 0.65 ± 0.22, indicating that POMC AMPA receptors are calcium impermeable (Fig. 6). However, when mice were treated with leptin, the RI was 0.28 ± 0.21 (saline, n = 7 cells; leptin, n = 8 cells; t = 3.21; P = 0.006; unpaired t test; Fig. 6C), indicating a shift to calcium permeable AMPA receptors. This suggests that drops in leptin levels during fasting may contribute to changes in AMPA receptor subunit composition on POMC, switching these receptors to the lower conductance, calcium-impermeable state where glutamate inputs are less depolarizing.

Figure 6.

Leptin administration prevents the switch in POMC AMPA receptor calcium permeability observed during fasting. A: representative traces showing light-evoked EPSCs recorded from POMC cells at holding potentials of +40 mV and −60 mV in fasted mice that were treated with either saline or leptin (5 mg/kg) 2 h prior to euthanasia. B: I-V relationship of EPSCs recorded from POMC cells from fasted mice treated with either saline or leptin. The value of peak amplitude EPSCs was normalized as percent of the peak at −60 mV. C: summary data showing that leptin treatment prevents fasting-induced increases in the RI (t = 3.21; P = 0.006; unpaired t test; **P < 0.01). EPSCs, excitatory postsynaptic potentials; I-V, current-voltage; POMC, proopiomelanocortin; RI, rectification index; V_hold, holding potential; VMH, ventromedial hypothalamus.

DISCUSSION

The results presented provide additional insight into the significance of glutamate release from the VMH on bodyweight and homeostatic food intake. The weight phenotype observed in response to the conditional deletion of vGlut2 from VMH neurons in the adult mouse may be at least partially attributed to the observation that VMH-derived glutamate release is a source of glutamatergic input to POMC neurons and the loss of this input decreases excitability of the anorexigenic POMC neurons. However, considering the diversity of VMH projections sites, it is also likely that the metabolic phenotypes observed are also partially mediated by loss of VMH glutamate release into other brain regions. Regardless, the results here further indicate the ability of POMC neurons to dynamically regulate the composition of AMPA receptors on their surface and add that this is an important factor regulating POMC neuron activity in various energy states and in responding to glutamate release from VMH neurons.

Conditional Deletion of VMH Glutamate Release Results in Increased Body Mass and Food Consumption

Considering the roles of the VMH and ARC POMC cells in regulating energy balance (1, 8, 10, 54), and the glutamatergic connection between the two (6, 7), we hypothesized that conditional deletion of VMH glutamate release would result in less excitatory drive onto POMC cells and increased bodyweight. Importantly, we wanted to employ a conditional approach to avoid compensatory effects inherent to constitutive deletion. Indeed, both male and female mice gained significantly more weight following conditional deletion of VMH glutamate release. Additionally, these mice consumed more rodent chow during a 2 h refeed than eGFP-injected controls. Importantly, the VMH is a heterogeneous population of cells (55), projecting throughout the brain (56). It is likely that the loss of glutamate release from the VMH contributes to brain-wide changes that contributed to the bodyweight and food consumption phenotypes that we observed. Indeed, the VMH projects to brain regions known to be involved in food intake and manipulation of these circuits can affect metabolic activity (39). However, the increased food intake and bodyweight observed with the inducible disruption of vGlut2 in the VMH is consistent with the possibility that VMH-to-POMC glutamate synapses contribute to the inhibition of food intake. Considering the recent evidence that POMC neurons are activated immediately upon sensory detection of food (23), our data provide a rapidly acting, glutamate circuit that may contribute to this immediate activation of POMC cells that occurs before ingestion of food and mobilization of gut-derived peptides that also affect that activity of POMC cells.

VMH Glutamate Neurons Contribute Dynamically Regulated Glutamate Inputs to POMC Cells

Initial experiments were designed to determine the relative contribution of VMH-originating glutamatergic inputs to POMC cells to the overall glutamatergic regulation of POMC cells and test whether energy state modulated the frequency of overall- or VMH-derived mEPSCs onto POMC cells. In recordings from control mice, the average frequency for mEPSCs was 3.38 ± 0.27 Hz. In recordings from VMH VGLUT 2 KO mice, the average frequency of mEPSCs was 2.29 ± 0.23 Hz, indicating that glutamate neurons in the VMH target POMC cells and represent ∼32% of the total glutamatergic inputs to POMC cells. Considering the extent and diversity of brain regions upstream of POMC cells (29), it will be of future interest to determine which regions contribute the other 68% of excitatory inputs, and how these connections respond to changes in energy state.

Refeeding Increases Glutamate Release Probability at VMH-to-POMC Synapses

The present data suggest that changes in the strength of VMH-to-POMC synapses could contribute to the regulation of motivated behaviors related to food intake. We initially focused on VMH-to-POMC synapses due to the prior observation that VMH inputs to POMC cells are diminished during fasting (7). Complimenting this prior study, our data indicate that in the fed state, there is an increase in the strength of VMH-to-POMC glutamatergic synapses. This increase in release of glutamate onto POMC cells during refeeding may contribute to the immediate increase in POMC neuronal activity during feeding and while experiencing food-related cues (23, 40). Interestingly, the activation of POMC cells during refeeding is independent of vagal and brainstem inputs (57), and although multiple brain regions likely contribute to activation of POMC cells during refeeding, our data suggest that the VMH may play a critical role. Considering that the VMH only makes up 32% of spontaneous glutamate release onto POMC cells, it is likely that the other brain regions are also responsible to the immediate increase in POMC cellular activity during food consumption.

Leptin Administration Blocks Fasting-Induced Changes in POMC AMPA Receptor Subunit Composition

Although we did not detect fasting-induced changes in presynaptic glutamate release when examining mEPSCs from all sources, mEPSC amplitude was reduced during fasting in recordings from both POMC-eGFP mice and VMH VGLUT2 KO mice, indicating that caloric deficit reduces unitary conductance at POMC AMPA receptors. This observation, as well as previous work (41), led us to hypothesize that POMC AMPA receptors undergo postsynaptic changes in receptor subunit composition during variations in energy state. Indeed, we found that POMC AMPA receptors are in the GluR2-lacking, calcium-permeable state in mice fed ad libitum, but then switch to the GluR2-containing, calcium-impermeable state during fasting. Our data suggest that the calcium-permeable state of POMC AMPA receptors in mice fed ad libitum is one mechanism through which these neurons are stimulated to prevent overeating. Likewise, the fasting-induced switch to the lower conductance, calcium-impermeable AMPA receptors is likely one of the many mechanisms that contribute to an overall inhibition of POMC cells during caloric deficit, facilitating food intake.

Ad libitum fed mice have elevated circulating leptin levels (18). Interestingly, leptin promotes the removal of the GluR2 subunit from hippocampal neurons (53) via inhibition of the lipid phosphatase PTEN. We hypothesized that circulating leptin levels are involved in maintaining the calcium permeability of POMC AMPA receptors by promoting the GluR2-lacking configuration and thus, that drops in leptin levels during fasting may contribute to the observed switch to the GluR2-containing, calcium-impermeable state. The finding that leptin administration prevented the fasting-induced switch to calcium-permeable AMPA receptors supported this hypothesis. Together with the prior results from mixed inputs (41), it is clear that circulating hormone levels contribute to the regulation of POMC neuronal activity through modulation of AMPA receptor complexes to change the unitary conductance of these receptors.

Conclusions

In summary, glutamate release from VMH neurons contributes to normal energy balance in the adult animal such that reduced glutamate release from the VMH onto POMC neurons leads to weight gain and increased feeding. These metabolically sensitive excitatory inputs from the VMH onto POMC change in synaptic strength at both pre- and postsynaptic sites in an energy state-dependent manner.

GRANTS

This work was supported by National Institute of Health Grant DK078749 (S. T. Hentges). We declare no competing financial interests.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.R.R. and S.T.H. conceived and designed research; A.R.R. performed experiments; A.R.R. analyzed data; A.R.R. and S.T.H. interpreted results of experiments; A.R.R. prepared figures; A.R.R. and S.T.H. drafted manuscript; A.R.R. and S.T.H. edited and revised manuscript; A.R.R. and S.T.H. approved final version of manuscript.