Cellular and metabolic function of GIRK1 potassium channels expressed by arcuate POMC and NPY/AgRP neurons

Abstract

It is well known that the G protein-gated inwardly rectifying K⁺ (GIRK) channels are critical to maintain excitability of central neurons. GIRK channels consist of 4 subunits and GIRK1/GIRK2 heterotetramers are considered to be the neuronal prototype. We previously reported the metabolic significance of GIRK2 subunits expressed by the neuropeptide Y (NPY)/agouti-related peptide (AgRP) neurons of the arcuate nucleus of the hypothalamus (ARH). However, the role of GIRK1 subunits expressed by the neurons of ARH remains to be determined. In this study, we delineated the contribution of GIRK1 channel subunits to the excitability of the pro-opiomelanocortin (POMC) and NPY/AgRP neurons of the ARH. We further assessed the metabolic function of GIRK1 subunits expressed by these neurons. Our results provide insight into how GIRK channels regulate arcuate POMC and NPY/AgRP neurons and shape metabolic phenotypes.

INTRODUCTION

The arcuate nucleus of hypothalamus (ARH) is one of the key brain sites to control feeding behavior and metabolism (Waterson and Horvath, 2015). Two well-known neuronal populations exist in the ARH: the anorexigenic (appetite-suppressing) pro-opiomelanocortin (POMC) neurons and the orexigenic (appetite-promoting) neuropeptide Y (NPY)/agouti-related peptide (AgRP) neurons (Sohn, 2015). Several previous studies demonstrated that optogenetic/chemogenetic manipulations of POMC and/or NPY/AgRP neuronal activity dramatically alter feeding behavior (Aponte et al., 2011, Krashes et al., 2013, Zhan et al., 2013). However, far less is known about intrinsic mechanisms that control the activity of these neurons.

The G protein-gated inwardly rectifying K⁺ (GIRK or Kir3) channels are important to control intrinsic excitability as well as synaptic responses to inhibitory neurotransmitters such as gamma-aminobutyric acid (GABA) (Luján and Aguado, 2015, Lüscher and Slesinger, 2010). Mammalian cells express 4 GIRK channel subunits (GIRK1-GIRK4 or Kir3.1-Kir3.4), and GIRK1/GIRK2 heterotetramers are considered to be the neuronal prototype (Luján and Aguado, 2015, Lüscher and Slesinger, 2010). Previous studies demonstrated the importance of GIRK2 channel subunits in functional GIRK channels of several brain areas, including hippocampus, cerebellum, and midbrain (Cruz et al., 2008, Lüscher et al., 1997, Signorini et al., 1997).

Within the hypothalamus, previous studies demonstrated that neurons in the ARH and ventromedial nucleus of hypothalamus express functional GIRK channels (Acuna-Goycolea et al., 2005, Chee et al., 2010, Kelly et al., 2002, Roseberry et al., 2004). In particular, we demonstrated that GIRK2 channel subunits expressed by arcuate NPY/AgRP neurons are important to stabilize neuronal excitability and increase energy expenditure (EE) via the regulation of sympathetic activity (Oh et al., 2023). On the other hand, GIRK1 channel subunits were suggested to be more important than GIRK2 subunits in the regulation of POMC neuronal excitability (Sohn et al., 2011). Nonetheless, little information is currently available on the metabolic role of GIRK1 channel subunits in hypothalamic neurons.

In this study, we performed whole-cell patch clamp experiments to further assess the role of GIRK1 channel subunits expressed by POMC and NPY/AgRP neurons within the ARH in the regulation of neuronal excitability. We also generated conditional knockout mice to delineate the metabolic function of GIRK1 subunits expressed by POMC and NPY/AgRP neurons.

MATERIALS AND METHODS

Mice

All mice used for breeding and experiments in this study were housed in a light-dark (12 hours on/off; lights on at 7:00 am) and temperature-controlled environment with food and water available ad libitum in the Korea Advanced Institute of Science and Technology facilities with the guidelines established by the Korea Advanced Institute of Science and Technology's Institutional Animal Care and Use Committee (Protocol No. KA2021-126). Pathogen-free male POMC-humanized Renilla green fluorescent protein (hrGFP) (Jackson Laboratory, #006421) or male NPY-hrGFP (Jackson Laboratory, #006417) mice (5-13 weeks old) that express hrGFP under the promoter of Pomc or Npy genes were used for all patch-clamp experiments to localize POMC- or NPY-expressing neurons in ARH. In some patch-clamp experiments, GIRK1 (Bettahi et al., 2002) or GIRK2 (Signorini et al., 1997) global knockout mice were crossed with POMC-hrGFP or NPY-hrGFP mice to characterize the roles of specific GIRK channel subunits. NPY-hrGFP mice were fasted for 18 hours before being sacrificed for patch-clamp experiments.

POMC-Cre mice (Jackson Laboratory, #005965) or AgRP-ires-Cre mice (Jackson Laboratory, #012899) were crossed with either Girk1^flox/flox mice (Marron Fernandez de Velasco et al., 2017) or tdTomato reporter mice (Jackson Laboratory, #007905) for in situ hybridization, immunohistochemistry, and in vivo metabolic experiments. Girk1^flox/flox mice were kindly provided by Dr. Kevin Wickman (University of Minnesota). Mice were fed with standard normal chow diet (NCD, Teklad global 18% protein 2018S, ENVIGO) or 60% high fat diet (HFD, #TD.06414, ENVIGO). For some whole-cell patch-clamp experiments, mice were fed with HFD for 3 to 4 weeks before being sacrificed for experiments at 8 to 9 weeks of age.

Electrophysiology

Whole-cell path-clamp recordings from hrGFP-expressing neurons were maintained in acute hypothalamic slice preparations as previously described (Sohn et al., 2016). In brief, 5- to 13-week-old male mice were deeply anesthetized with isoflurane inhalation and transcardially perfused through ventricular catheter with a modified ice-cold artificial CSF (ACSF) (described below), in which an equiosmolar amount of sucrose was substituted for NaCl. The mice were then decapitated, and the entire brain was removed from the skull and immediately submerged in ice-cold, carbogen-saturated (95% O₂ and 5% CO₂) ACSF (126 mM NaCl, 26 mM NaHCO₃, 2.8 mM KCl, 1.25 mM NaH₂PO₄, 1.2 mM MgSO₄, 2.5 mM CaCl₂, and 5 mM glucose). Coronal sections (250 µm) were cut with a Leica VT1200S vibrating microtome and then incubated in oxygenated ACSF at 34°C for at least 1 hour before recording. Brain slices were transferred to the recording chamber and allowed to equilibrate for 10 to 20 minutes before recording. The slices were bathed in oxygenated ACSF (32-34°C) at a flow rate of around 2 ml/min. The pipette solution was modified to include an intracellular dye (Alexa Flour 594) for whole-cell recording: 120 mM K-gluconate, 10 mM KCl, 10 mM HEPES, 1 mM CaCl₂, 1 mM MgCl₂, 5 mM EGTA, 2 mM Mg-ATP, and 0.03 mM Alexa Fluor 594 hydrazide dye, pH 7.3. Recording electrodes had resistances of 2.5 to 5 MΩ when filled with the K-gluconate internal solution. Epifluorescence was briefly used to target fluorescent cells, at which time the light source was switched to infrared differential interference contrast imaging to obtain the whole-cell recording (Nikon Eclipse FN1 equipped with a fixed stage and an optiMOS scientific CMOS camera). Electrophysiological signals were recorded using an Axopatch 700B amplifier (Molecular Device), low-pass filtered at 1 to 2 kHz, and analyzed offline on a PC with pCLAMP programs (Molecular Device).

In current-clamp experiments, input resistance was assessed by measuring voltage deflection at the end of the response to hyperpolarizing rectangular current pulse steps (−50 to −10 pA or −25 to −5 pA for 500 ms each). A drug effect was required to be associated temporally with drug application, and the responses had to be stable within a few minutes. We determined membrane potential before and during drug applications by averaging membrane potential for 10 seconds in each condition. A neuron was considered to be depolarized or hyperpolarized if a change in membrane potential was larger than 2 mV in amplitude. Membrane potential values in this study were not compensated for liquid junction potentials (−8 mV).

For voltage-clamp experiments, we used the same K-gluconate pipette solutions described above and added 0.5 µM tetrodotoxin (TTX) and synaptic blockers (50 µM picrotoxin and 1 mM kynurenic acid) in oxygenated ACSF. Local applications of baclofen (a GABA_B receptor agonist) were performed using micropipettes attached to the Picospritzer III microinjection dispense system (Parker Hannifin). At a holding potential of −40 mV, a micropipette was placed 10 to 20 µm away from soma, and small volume (15-20 pl) of ACSF (containing baclofen and the blocker cocktail) was ejected with a pressure of 16 to 18 psi for 15 seconds. The currents generated by baclofen (I_Bac) were normalized by cell capacitance. Voltage ramp pulses (from −120 to −10 mV, 100 mV/s) were applied before and during the baclofen application to obtain I-V relationships of I_Bac.

Immunohistochemistry

Eight- to 12-week-old male POMC-Cre::tdTomato mice were deeply anesthetized with isoflurane and transcardially perfused with phosphate-buffered saline (PBS) and subsequently with a 4% paraformaldehyde-based fixative. Brains were removed from the skull, and submerged in cold 4% paraformaldehyde for postfixation. After a 16-hour postfixation, brains were transferred to sucrose solutions of gradients (4 hours in 10% sucrose, 12 hours in 20% sucrose, and 24 hours in 30% sucrose). Brains were embedded in Tissue-Tek O.C.T. compound (Sakura Finetek) and were frozen at −80°C. Coronal brain slices with 30 µm thickness were obtained at −23°C using a cryostat (Leica, #CM1860) and were stored in a cryo-protectant solution at −20°C.

Brain slices were washed 3 times with PBS (10 minutes each) and incubated overnight at 4°C with primary antibodies. Anti-GIRK1 antibodies (1:100 or 1:200, APC-005, Alomone Labs) were diluted with PBS containing 0.3% Triton X-100% and 5% BSA before being applied to brain slices. Anti-GIRK2 antibodies (1:200, APC-006, Alomone Labs) were diluted in PBS containing 0.25% Triton X-100% and 10% goat serum. At the end of incubations with primary antibodies, brain tissues were washed 3 times with PBS (10 minutes each), which was followed by 2 hours of incubation at room temperature with Alexa flour 488 rabbit antimouse secondary antibodies (Thermo Fisher Scientific) that were diluted in PBS (1:500) for binding to GIRK1 antibodies or in PBS containing 0.25% Triton X-100% and 2% goat serum (1:500) for binding to GIRK2 antibodies. Brain slices were washed again with PBS (3 times, 10 minutes each), and cover-slipped using fluorescence mounting medium (DAKO). Images were obtained with a confocal microscope LSM 780 (Carl Zeiss) and were analyzed with ZEN lite (ZEN Microscopy software).

In Situ Hybridization

In situ hybridization was performed using the BaseScope Duplex Kit (ACDBio, #323800). Briefly, 12-week-old Girk1^flox/flox or POMC-Cre::Girk1^flox/flox mice were deeply anesthetized with isoflurane and perfused with RNase-free PBS and 10% neutral buffered formalin (Sigma, #HT501128). Brains were obtained and postfixed for 24 hours with 10% neutral buffered formalin. Subsequently, brains were permeated for 24 hours in 30% sucrose dissolved in RNase-free PBS and were embedded in Tissue-Tek O.C.T. compound (Sakura Finetek) before being frozen at −80°C. Coronal brain slices with 15 µm thickness were obtained at −23°C using a cryostat (Leica, #CM1860) and were stored in a cryo-protectant solution at −20°C. Brain slices were mounted on SupraFrost Plus Slides (Fisher Scientific, #12-550-15) for the assays. All reagents for these assays were purchased from ACD.

BaseScope Girk1 probe (target region 1,351-1,479, accession # NM_001355118.1) and BaseScope Pomc Probe (target region 151-269, accession # NM_008895.3) were used to visualize Pomc and Girk1 mRNA. To minimize mRNA interference, the Pomc probe was diluted 12 times in a probe diluent (ACDBio, #300041) and brain slices were incubated using RNAscope Protease Plus (ACDBio, #322331) for 15 minutes at 40°C. Procedures followed the BaseScope Duplex Kit protocol. Images of the BaseScope assays were obtained with a slide scanner (Axio Scan. Z1, Carl Zeiss) and were analyzed with the Image J software. We lined up all stained images from individual mice, based on the reference brain atlas (Allen Brain Atlas), and selected 20 brain slices that are available from all mice to be included for analyses. Cells with 4 or more dots were considered to express Pomc or Girk1 mRNAs.

Body Weight and Food Intake

Each conditional knockout mouse (GIRK1^POMC-KO or GIRK1^AgRP-KO) had its littermate control mouse (GIRK1^WT). Body weight and food intake were measured weekly (5-20 weeks of age) from single-housed male mice. A 60% HFD (#TD.06414, ENVIGO) was provided (5-20 weeks of age) for some body weight and food intake measurements. Twenty-two- to 24-week-old male mice were fasted for 18 hours for fast-refeeding experiments. Refeeding started at 10 am.

Energy Expenditure, Physical Activity, Body Composition

EE and physical activity were measured from 20- to 21-week-old male mice by an indirect calorimetric chamber (CLAMS 12; Columbus Instruments). After acclimation for 2 days, O₂ consumption (VO₂) and CO₂ production (VCO₂) were measured for 2 days to determine the EE. Simultaneously, physical activity was determined using a multidimensional infrared light beam system with beams installed on bottom and top levels of the cage. Ambulatory movement was defined as breaks of any 2 different light beams at the bottom level of the cage, while rearing was recorded once the mouse broke any light beam at the top level. Body composition was measured by time domain nuclear magnetic resonance (NMR) spectrometer (Minispec LF50, Bruker Biospin).

Organ Weight and Fat Mass

Several organs (liver, kidney, spleen, heart, and quadriceps muscle) as well as white adipose tissue (WAT; inguinal, interscapular, perigonadal, and mesenteric) and brown adipose tissue (BAT) were obtained from 26- to 28-week-old male mice. All organs and adipose tissues were stored in DPBS, and measurements were conducted after gently removing DPBS. The weights of bilateral organs and fats were averaged.

Drugs

All solutions of tertiapin-Q (STT-170, Alomone Labs), baclofen (0417, Tocris), kynurenic acid (K3375, Sigma), picrotoxin (1128, Tocris), and TTX (T-550, Alomone Labs) were made according to manufacturer’s specifications. Stock solutions of all drugs used in this study were made by dissolution in deionized water.

Data Analysis

Statistical data are expressed as mean ± s.e.m., where n represents the number of cells or mice studied. The significance of differences between groups was evaluated using 2-tailed unpaired Student’s t test or 2-way repeated measures ANOVA with Šidák’s multiple comparison tests with a confidence level of P < .05 (*), P < .01 (**), P < .001 (***), or P < .0001 (****).

RESULTS

GIRK Channels Contribute to Maintain Resting Membrane Potential of POMC Neurons

Arcuate POMC neurons reportedly express mRNA transcripts of all of the 4 GIRK subunits (Girk1 or Kcnj3, Girk2 or Kcnj6, Girk3 or Kcnj9, and Girk4 or Kcnj5) (Henry et al., 2015). Given that GIRK1/GIRK2 heterotetramers are the neuronal prototype (Luján et al., 2014, Lüscher and Slesinger, 2010), we conducted immunohistochemistry targeting GIRK1 and GIRK2 subunits using POMC-Cre::tdTomato mice ( Fig. 1a and b). We found that 43.4 ± 2.6% (n = 3) of tdTomato-positive cells (putative POMC neurons) also express GIRK1 subunits, while 55.2 ± 2.7% (n = 3) of tdTomato-positive cells express GIRK2 subunits (Fig. 1c). We did not note significant differences in the expression levels of GIRK1 and GIRK2 subunits by POMC neurons.

Fig. 1.

GIRK channels are expressed by arcuate POMC neurons and maintain RMP. (a and b) Expression of GIRK1 subunits (a) or GIRK2 subunits (b) in tdTomato-expressing ARC POMC neurons of POMC-Cre::tdTomato mice. Arrows indicate GIRK1- or GIRK2-expressing POMC neurons. 3V; third ventricle. (c) POMC neuronal populations (%) that express GIRK1 subunits (empty circle) or GIRK2 subunits (filled circle) across the bregma. Gene (df = 1, F_{1, 4} = 1.048, and P = .364), bregma (df = 7, F_{2.500, 10.00} = 3.418, and P = .067, and interaction (df = 7, F_{7, 28} = 0.605, and P = .747). (d) Whole-cell patch clamp recording of POMC-hrGFP neurons. Brightfield illumination (Brightfield), fluorescent (FITC) illumination of hrGFP, fluorescent (TRITC) illumination of Alexa Fluor 594 dye, and a successful overlap of POMC-hrGFP neurons with the complete dialysis of Alexa Fluor 594 (Merge). Arrows indicate the targeted neuron for recording. (e) Depolarization of POMC-hrGFP neurons by tertiapin-Q (300 nM). Dotted line indicates RMP. (f) Depolarization of POMC-hrGFP neurons by tertiapin-Q (300 nM) in the background of TTX and synaptic blockers (picrotoxin and kynurenic acid). Dotted line indicates RMP. (g) Percentages of depolarized POMC neurons (%) in ACSF and synaptic blockers-contained ACSF. (h) Changed amplitudes of membrane potential of POMC neurons by tertiapin-Q in ACSF (2.1 ± 1.0 mV, n = 5) and synaptic blockers-contained ACSF (2.3 ± 0.9 mV, n = 6, df = 9, t = 0.1163, and P = .910). Red circles for depolarized neurons and black circles for no response neurons. (i) Voltage deflection in response to hyperpolarizing current steps (from −50 to 0 pA by 10 pA increment) before (Control, upper trace) and during (tertiapin-Q, lower trace) the perfusion with tertiapin-Q to POMC neurons as indicated with arrows in (f). (j) Voltage-current (V-I) relationship demonstrates that tertiapin-Q increased input resistance by tertiapin-Q. Reversal potential is indicated as E_rev. Data are presented as mean ± SEM. Two-way ANOVA with Šidák correction (c) and unpaired t test (h) was used for statistical analyses. ns, not significant.

We obtained acute hypothalamic slices that contain ARH from the POMC-hrGFP mice and targeted the GFP-labeled POMC neurons for whole-cell patch clamp recordings (Fig. 1d). We tested acute effects of teritapin-Q (300 nM), a GIRK channel blocker, on membrane potential of POMC neurons (Fig. 1e and f). We found that tertiapin-Q depolarized membrane potential in 3 of 5 (60%) POMC neurons by 3.7 ± 0.3 mV (from −55.3 ± 2.3 mV to −51.7 ± 2.0 mV, n = 3, Fig. 1g and h, left bars). We also noted that tertiapin-Q depolarized 3 of 6 (50%) POMC neurons by 4.3 ± 0.3 mV (from −59.3 ± 2.2 mV to −55.0 ± 2.0 mV, n = 3, Fig. 1g and h, right bars) when 0.5 μM TTX and synaptic blockers (1 mM kynurenic acid and 50 μM picrotoxin) were added to the bath solutions to synaptically isolate neurons being recorded. We applied small hyperpolarizing current steps pulses (−50 to −10 pA for 500 ms each, Fig. 1f and i) before and during the perfusion of teritapin-Q to obtain a voltage-current (V-I) relationship (Fig. 1j). We noted that the depolarizing effects of tertiapin-Q were accompanied by increased input resistance (from 1.4 ± 0.1 GΩ to 1.6 ± 0.1 GΩ, n = 6) with a reversal potential (E_rev) of −77.1 ± 2.3 mV (n = 6). These results indicate that GIRK1 and/or GIRK2 channels are expressed by arcuate POMC neurons to be assembled into functional GIRK channels to maintain RMP.

GIRK1-Containing GIRK Channels Contribute to RMP and GABA_B-Activated K⁺ Currents of POMC Neurons

We previously reported that GIRK1-deficient POMC neurons have significantly depolarized RMP compared to wildtype POMC neurons (Sohn et al., 2011). In agreement, we confirmed that GIRK1-deficient POMC (POMC^G1KO) neurons have significantly depolarized RMP (−51.4 ± 0.9 mV, n = 52, P = .042) compared to wildtype POMC (POMC^WT) neurons (−53.9 ± 0.8 mV, n = 81) (Fig. 2a and b). The depolarized RMP of POMC^G1KO neurons was accompanied by a significantly increased input resistance (2.0 ± 0.1 GΩ, n = 52, P = .001) compared to that of POMC^WT neurons (1.7 ± 0.1 GΩ, n = 81) (Fig. 2c). These results suggest that GIRK1-containing GIRK channels are open at rest to stabilize RMP of arcuate POMC neurons.

Fig. 2.

GIRK1-containing GIRK channels contribute to RMP and GABA_B-induced inhibition of POMC neurons. (a) Spontaneous firing and RMP, which is followed by hyperpolarization of POMC^WT neurons (upper trace) and POMC^G1KO neurons (lower trace) by baclofen (100 μM). Dotted lines indicate RMP. (b and c) RMP (b) and input resistance (c) of POMC^WT neurons (black) and POMC^G1KO neurons (blue). (d) Summary of GABA_B-induced hyperpolarization of POMC^WT neurons (black) and POMC^G1KO neurons (blue). Solid lines indicate fitting of dose-response curve (Hill slope = 1.0, Y = Bottom + (Top-Bottom)/(1 + 10^(logEC50-X)). Both hyperpolarization and no responses were included in analyses. See Table 1 for amplitudes of hyperpolarization by variable doses of baclofen. (e) Outward currents recorded in POMC^WT neurons (left traces) and POMC^G1KO neurons (right traces), defined as I_Bac, by locally applying 10 μM baclofen (upper row) or 100 μM baclofen (lower row). (f and g) Normalized amplitudes of I_Bac by 10 μM baclofen (f) and by 100 μM baclofen of POMC^WT neurons and POMC^G1KO neurons (g). Data are presented as mean ± SEM. Unpaired t test was used for statistical analyses. *P < .05, **P < .01, and ****P < .0001.

Since GIRK1 subunits are known to mediate the inhibitory effects of GABA_B receptors in hippocampal neurons (Koyrakh et al., 2005), we examined the involvement of GIRK1 subunits in GABA_B-induced hyperpolarization of POMC neurons. We conducted current-clamp experiments and applied several concentrations (1, 10, 30, and 100 μM) of baclofen, a GABA_B receptor agonist, to the bath solutions. Mean amplitudes of baclofen-induced hyperpolarization were smaller in POMC^G1KO neurons compared to those in POMC^WT neurons, but the difference was statistically significant when we applied 100 μM baclofen: POMC^WT neurons were hyperpolarized by −18.9 ± 1.8 mV (n = 9) and POMC^G1KO neurons were hyperpolarized by −13.4 ± 1.3 mV (n = 9, P = .024) in response to baclofen applications (Fig. 2a and d, Table 1).

Table 1.

GABA_B-induced hyperpolarization of arcuate POMC neurons

Baclofen concentration (µM)	POMCWT	POMCG1KO	POMCG2KO
1	−8.2 ± 1.5 mV (n = 6) 5 of 6, 83% (EK = −99.5 ± 5.4 mV, n = 5)	−3.9 ± 1.4 mV (n = 8) 4 of 8, 50% (EK = −109.7 ± 9.7 mV, n = 4)	−5.0 ± 1.4 mV (n = 9) 5 of 9, 56% (EK = −84.5 ± 1.7 mV, n = 5)
10	−9.7 ± 1.7 mV (n = 21) 15 of 21, 71% (EK = −91.6 ± 2.6 mV, n = 15)	−6.0 ± 1.1 mV (n = 9) 8 of 9, 89% (EK = −97.6 ± 5.9 mV, n = 8)	−7.1 ± 1.5 mV (n = 14) 11 of 14, 79% (EK = −97.0 ± 4.7 mV, n = 11)
30	−14.5 ± 2.0 mV (n = 12) 11 of 12, 92% (EK = −89.2 ± 3.4 mV, n = 11)	−11.2 ± 1.2 mV (n = 12) 12 of 12, 100% (EK = −97.9 ± 3.5 mV, n = 12)	−12.7 ± 1.8 mV (n = 9) 9 of 9, 100% (EK = −84.9 ± 2.6 mV, n = 9)
100	−18.9 ± 1.8 mV (n = 9) 9 of 9, 100% (EK = −90.3 ± 2.2 mV, n = 9)	−13.4 ± 1.3 mV (n = 9) 9 of 9, 100% (EK = −92.5 ± 6.3 mV, n = 9)	−12.2 ± 2.2 mV (n = 7) 7 of 7, 100% (EK = −93.2 ± 6.8 mV, n = 7)

POMC, pro-opiomelanocortin.

We also performed voltage-clamp experiments to record GABA_B-activated or baclofen-induced GIRK currents. We held the membrane potential at −40 mV and applied 100 μM baclofen to observe outward currents (Fig. S1a). We also applied voltage ramp pulses from −120 to −10 mV (100 mV/s) before and during baclofen treatments as indicated by arrows a and b in Fig. S1a. We subtracted the current-voltage (I-V) relationships (b and a) to obtain the I-V relationship of baclofen-activated currents (I_Bac) (Fig. S1b), where we noted that I_Bac is an inwardly rectifying current with an E_rev of −89.9 ± 1.2 mV (n = 14) that is close to E_K. The rectification index (I_-120 mV/I_-60 mV) of I_Bac was 4.2 ± 0.7 (n = 14). These results confirm that I_Bac has key electrophysiological properties of GIRK currents.

We recorded I_Bac in POMC^WT neurons and POMC^G1KO neurons by applying 2 concentrations of baclofen (10 and 100 μM). We found that normalized amplitudes of I_Bac (10 μM baclofen) were significantly reduced in POMC^G1KO neurons (1.0 ± 0.1 pA/pF, n = 23, P = .005) compared to the results from POMC^WT neurons (1.6 ± 0.2 pA/pF, n = 17) (Fig. 2e, upper traces and Fig. 2f). We obtained similar results with 100 μM baclofen: the normalized amplitudes of I_Bac were significantly decreased in POMC^G1KO neurons (1.1 ± 0.1 pA/pF, n = 25, P < .0001) compared to those from POMC^WT neurons (2.2 ± 0.2 pA/pF, n = 27) (Fig. 2e, lower traces and Fig. 2g). Taken together, we concluded that GIRK1-containing GIRK channels contribute to RMP and GABA_B-activated K⁺ currents to inhibit arcuate POMC neurons.

GIRK2-Containing GIRK Channels Contribute to GABA_B-Activated K⁺ Currents, but not RMP, of POMC Neurons

Since GIRK2 subunits are essential for functional GIRK channels in some central neurons (Luján et al., 2014, Lüscher and Slesinger, 2010), we also assessed the contribution of GIRK2-containing GIRK channels. In current-clamp experiments, we noted slightly depolarized RMP of GIRK2-deficient POMC (POMC^G2KO) neurons (−51.6 ± 1.0 mV, n = 44) compared to that of POMC^WT neurons (−53.9 ± 0.8 mV, n = 81), but this difference was not statistically significant (P = .077) (Fig. S2a and b). The input resistance was also comparable between POMC^WT neurons (1.7 ± 0.1 GΩ, n = 81) and POMC^G2KO neurons (1.8 ± 0.1 GΩ, n = 44, P = .183) (Fig. S2c).

We subsequently examined the involvement of GIRK2 subunits in GABA_B-induced hyperpolarization of POMC neurons and noted found that amplitudes of baclofen-induced hyperpolarization were significantly reduced (P = .031) at 100 μM of baclofen: the mean amplitudes were −12.2 ± 2.2 mV (n = 7) in POMC^G2KO neurons and −18.9 ± 1.8 mV (n = 9) in POMC^WT neurons (Fig. S2a and d, Table 1). Consistent with these results, we found in voltage-clamp experiments that the amplitudes of I_Bac were significantly reduced in POMC^G2KO neurons compared to that recorded in POMC^WT neurons. At 10 μM baclofen, normalized amplitudes of I_Bac were significantly reduced in POMC^G2KO neurons to 1.1 ± 0.1 pA/pF (n = 23, P = .034) compared to that of POMC^WT neurons (1.6 ± 0.2 pA/pF, n = 17) (Fig. S2e, upper traces and Fig. S2f). The normalized amplitudes of I_Bac were also significantly reduced in POMC^G2KO neurons to 1.3 ± 0.2 pA/pF (n = 28, P = .003) compared to that of POMC^WT neurons (2.2 ± 0.2 pA/pF, n = 27) at 100 μM baclofen (Fig. S2e, lower traces and Fig. S2g). Taken together, we concluded that GIRK2-containing GIRK channels contribute to GABA_B-activated K⁺ currents to inhibit arcuate POMC neurons without significant effects on RMP.

GIRK1-Containing GIRK Channels of POMC Neurons do not Contribute to Maintain Energy Homeostasis

Given the larger contribution of GIRK1 subunits over GIRK2 subunits to maintain the excitability of arcuate POMC neurons, we chose to delineate the metabolic function of GIRK1 channel subunits expressed by arcuate POMC neurons. To this end, we crossed POMC-Cre mice (Balthasar et al., 2004) with Girk1^flox/flox mice (Marron Fernandez de Velasco et al., 2017) to obtain POMC-Cre::Girk1^flox/flox mice. We confirmed the successful deletion of GIRK1 subunits in POMC neurons by in situ hybridization experiments (Duplex BaseScope) in Girk1^flox/flox (GIRK1^WT) mice and POMC-Cre::Girk1^flox/flox (GIRK1^POMC-KO) mice (Fig. S3a and b). We noted comparable numbers of Pomc(+) neurons in the ARH (Fig. S3c), but the percentage of Girk1(+) Pomc(+) neurons was significantly decreased in GIRK1^POMC-KO mice (13.85 ± 1.22%, n = 4, P < .0001) compared to GIRK1^WT mice (69.81 ± 2.00%, n = 3). The percentage of Girk1(+) Pomc(+) neurons was decreased throughout the rostrocaudal axis of ARH (from bregma −1.22 to −2.54 mm) (Fig. S3e).

We measured body weight and food intake of NCD-fed GIRK1^WT and GIRK1^POMC-KO mice every week (5-20 weeks). GIRK1^POMC-KO mice weighed less than GIRK1^WT mice until the 11th week, but this difference was no longer observed starting from the 12th week (Fig. 3a). The cumulative food intake was not different between GIRK1^WT mice and GIRK1^POMC-KO mice (Fig. 3b). The NMR spectrometer analyses of body composition, which were performed when the mice were 20 to 21 weeks old, demonstrated no significant difference in lean mass, fat mass, or body fluids between genotypes (Fig. 3c-e). Consistent with these results, we did not observe any differences between genotypes in the weight of liver, kidney, spleen, heart, and quadriceps femoris (quad) muscles (Table 2a), which may constitute lean mass. In addition, there were no differences in the weight of subcutaneous WAT such as inguinal WAT (IGW) and interscapular WAT (ISCW) or visceral WAT such as perigonadal WAT (PGW) and mesenteric WAT (MWAT) between genotypes (Table 2a). We also did not see any differences in the weights of BAT between GIRK1^WT mice and GIRK1^POMC-KO mice (Table 2a).

Fig. 3.

GIRK1-containing GIRK channels expressed by POMC neurons minimally contribute to energy homeostasis of NCD-fed mice. (a) Body weights of GIRK1^WT (n = 3, black) and GIRK1^POMC-KO mouse (n = 5, blue) on NCD. Gene (df = 1, F_{1, 6} = 4.88, and P = .069), time (df = 15, F_{2.54, 15.24} = 188.1, and P < .0001), and interaction (df = 15, F_{15, 90} = 3.864, and P < .0001). (b) Cumulative food intake of GIRK1^WT (n = 3, black) and GIRK1^POMC-KO mouse (n = 5, blue) on NCD. Gene (df = 1, F_{1, 6} =0.07393, and P = .795), time (df = 14, F_{1.017, 6.101} = 829.4, and P < .0001), and interaction (df = 14, F_{14, 84} = 0.6289, and P = .834). (c) Lean mass of GIRK1^WT (21.5 ± 0.3 g) and GIRK1^POMC-KO (23.0 ± 1.9 g, df = 6, t = 1.160, P = .290). (d) Fat mass of GIRK1^WT (4.9 ± 0.7 g) and GIRK1^POMC-KO (3.7 ± 0.8 g, df = 6, t = 1.072, and P = .325). (e) Body fluid of GIRK1^WT (2.3 ± 0.1 g) and GIRK1^POMC-KO (2.2 ± 0.1 g, df = 6, t = 0.7746, and P = 0.468). (f) Oxygen consumption (VO₂) (left, dark cycle: 3.31 ± 0.09 L/h/kg, n = 3, for GIRK1^WT and 3.26 ± 0.10 L/h/kg, n = 5, for GIRK1^POMC-KO, df = 6, t = 0.3177, and P = .762; right, light cycle: 2.53 ± 0.05 L/h/kg, n = 3, for GIRK1^WT and 2.52 ± 0.12 L/h/kg, n = 5, for GIRK1^POMC-KO, df = 6, t = 0.02954, and P = .977). (g) Carbon dioxide production (VCO₂) (left, dark cycle: 3.08 ± 0.09 L/h/kg, n = 3, for GIRK1^W and 3.03 ± 0.09 L/h/kg, n = 5, for GIRK1^POMC-KO, df = 6, t = 0.3915, and P = .709; right, light cycle: 2.26 ± 0.08 L/h/kg, n = 3, for GIRK1^WT and 2.31 ± 0.11 L/h/kg, n = 5, for GIRK1^POMC-KO, df = 6, t = 0.3308, and P = .752). (h) Energy expenditure (EE) (left, dark cycle: 0.513 ± 0.008 kcal/h, n = 3, for GIRK1^WT and 0.507 ± 0.018 kcal/h, n = 5, for GIRK1^POMC-KO, df = 6, t = 0.2164, and P = .836; right, light cycle: 0.382 ± 0.010 kcal/h, n = 3, for GIRK1^WT and 0.390 ± 0.016 kcal/h, n = 5, for GIRK1^POMC-KO, df = 6, t = 0.3502, and P = .738). (i) Respiratory exchange ratio (RER) (left, dark cycle: 0.932 ± 0.005, n = 3, for GIRK1^WT and 0.929 ± 0.006, n = 5, for GIRK1^POMC-KO, df = 6, t = 0.3026, and P = .772; right, light cycle: 0.894 ± 0.015, n = 3, for GIRK1^WT and 0.916 ± 0.012, n = 5, for GIRK1^POMC-KO, df = 6, t = 1.141, and P = .297). Data are presented as mean ± SEM. Two-way ANOVA with Šidák correction (a and b) and unpaired t test (c-i) were used for statistical analysis. *P < .05, **P < .01, and ns, not significant.

Table 2.

Organ and fat tissue weight of GIRK1^WT and GIRK1^POMC-KO mice

a
	Organ weight	Liver	Kidney	Spleen	Heart	Quad muscle
NCD	GIRK1WT	1.56 $\pm$ 0.05 g (n = 3)	0.24 $\pm$ 0.02 g (n = 3)	0.07 $\pm$ 0.00 g (n = 3)	0.17 $\pm$ 0.01 g (n = 3)	0.25 $\pm$ 0.01 g (n = 3)
	GIRK1POMC-KO	1.45 $\pm$ 0.12 g (n = 5) P = .523, df = 6, t = 0.6788	0.23 $\pm$ 0.01 g (n = 5) P = .356, df = 6, t = 0.9990	0.08 $\pm$ 0.01 g (n = 5) P = .190, df = 6, t = 1.477	0.16 $\pm$ 0.01 g (n = 5) P = .761, df = 6, t = 0.3189	0.26 $\pm$ 0.01 g (n = 5) P = .341, df = 6, t = 1.035
	Fat tissue weight	IGW	ISCW	PGW	MWAT	BAT
	GIRK1WT	0.68 $\pm$ 0.03 g (n = 3)	0.27 $\pm$ 0.00 g (n = 3)	0.76 $\pm$ 0.02 g (n = 3)	0.66 $\pm$ 0.04 g (n = 3)	0.10 $\pm$ 0.01 g (n = 3)
	GIRK1POMC-KO	0.66 $\pm$ 0.06 g (n = 5) P = .822, df = 6, t = 0.2346	0.24 $\pm$ 0.02 g (n = 5) P = .235, df = 6, t = 1.319	0.67 $\pm$ 0.09 g (n = 5) P = .471, df = 6, t = 0.7687	0.58 $\pm$ 0.10 g (n = 5) P = .572, df = 6, t = 0.5972	0.10 $\pm$ 0.01 g (n = 5) P = .954, df = 6, t = 0.06005
b
	Organ weight	Liver	Kidney	Spleen	Heart	Quad muscle
HFD	GIRK1WT	2.33 $\pm$ 0.19 g (n = 15)	0.26 $\pm$ 0.00 g (n = 15)	0.12 $\pm$ 0.01 g (n = 14)	0.20 $\pm$ 0.01 g (n = 15)	0.24 $\pm$ 0.01 g (n = 15)
	GIRK1POMC-KO	2.02 $\pm$ 0.16 g (n = 14) P = .217, df = 27, t = 1.265	0.27 $\pm$ 0.01 g (n = 14) P = .687, df = 27, t = 0.4076	0.11 $\pm$ 0.01 g (n = 14) P = .441, df = 26, t = 0.7820	0.18 $\pm$ 0.01 g (n = 14) P = .110, df = 27, t = 1.651	0.23 $\pm$ 0.01 g (n = 14) P = .483, df = 27, t = 0.7112
	Fat tissue weight	IGW	ISCW	PGW	MWAT	BAT
	GIRK1WT	1.84 $\pm$ 0.13 g (n = 15)	0.49 $\pm$ 0.03 g (n = 15)	1.06 $\pm$ 0.06 g (n = 15)	1.21 $\pm$ 0.10 g (n = 15)	0.22 $\pm$ 0.02 g (n = 15)
	GIRK1POMC-KO	1.69 $\pm$ 0.16 g (n = 14) P = .439, df = 27, t = 0.7860	0.48 $\pm$ 0.03 g (n = 14) P = .835, df = 27, t = 0.2107	1.15 $\pm$ 0.07 g (n = 14) P = .305, df = 27, t = 1.045	1.17 $\pm$ 0.15 g (n = 14) P = .840, df = 27, t = 0.2038	0.21 $\pm$ 0.02 g (n = 13) P = .725, df = 26, t = 0.3553

Given no differences in the body weight, body composition, and food intake, we predicted that there would be no differences in the EE as well. We measured oxygen consumption (VO₂) and carbon dioxide production (VCO₂) with an indirect calorimetry from 20- to 21-week-old GIRK1^WT and GIRK1^POMC-KO mice, and found that there was no significant difference between GIRK1^WT mice and GIRK1^POMC-KO mice (Fig. 3f and g). Likewise, calculated EE and respiratory exchange ratio (RER) did not differ between genotypes (Fig. 3h and i). During the indirect calorimetry measurements, we also measured total horizontal movements, ambulatory movements, and rearing activities, but there was no difference between genotypes (Fig. S4).

GIRK1-Containing GIRK Channels Contribute to RMP and GABA_B-Activated K⁺ Currents of NPY Neurons

We previously demonstrated that GIRK2-containing GIRK channels maintain RMP of NPY/AgRP neurons to regulate EE (Oh et al., 2023). To identify the role of GIRK1-containing GIRK channels in the regulation of NPY neuronal excitability, we recorded membrane potential of NPY neurons of wildtype and GIRK1KO NPY-hrGFP mice (NPY^WT neurons and NPY^G1KO neurons, respectively). We found in current-clamp experiments that NPY^G1KO neurons have significantly depolarized RMP (−43.4 ± 0.6 mV, n = 54, P = .0002) with statistical significance compared to NPY^WT neurons (−47.9 ± 0.9 mV, n = 64) (Fig. 4a and b). This depolarization was accompanied by increased input resistance of NPY^G1KO neurons (2.9 ± 0.1 GΩ, n = 54, P = .0014) compared to that of NPY^WT neurons (2.4 ± 0.1 GΩ, n = 64) (Fig. 4c). Therefore, GIRK1-containing GIRK channels, such as GIRK2-containing GIRK channels, also contribute to RMP of NPY neurons. In addition, we noted a significant reduction in GABA_B-induced hyperpolarization of membrane potential of NPY^G1KO neurons (−6.5 ± 1.5 mV, n = 12, P = .023) compared to that of NPY^WT neurons (−14.7 ± 3.4 mV, n = 8) when we used 1 μM baclofen (Fig. 4a and d). We did not observe significant differences when we used higher concentrations (10, 30, or 100 μM) of baclofen (Fig. 4d and Table 3). In voltage-clamp experiments, the amplitudes of normalized I_Bac were similar between NPY^WT neurons (1.4 ± 0.1 pA/pF, n = 32) and NPY^G1KO neurons (1.1 ± 0.1 pA/pF, n = 28, P = .172) when we used 10 μM baclofen (Fig. 4e, upper traces and Fig. 4f), but they were significantly decreased in NPY^G1KO neurons (1.2 ± 0.1 pA/pF, n = 31, P = .003) compared to NPY^WT neurons (1.8 ± 0.1 pA/pF, n = 53) when we used 100 μM baclofen (Fig. 4e, lower traces and Fig. 4g). These results mirror decreased GABA_B-induced hyperpolarization obtained from current-clamp experiments. Therefore, we concluded that GIRK1-containing GIRK channels contribute to RMP and GABA_B-activated K⁺ currents to inhibit arcuate NPY neurons.

Fig. 4.

GIRK-containing GIRK channels contribute to RMP and GABA_B-induced inhibition of NPY neurons. (a) Spontaneous firing and RMP, which is followed by hyperpolarization of NPY^WT neurons (upper trace) and NPY^G1KO neurons (lower trace) by baclofen (1 μM). Dotted lines indicate RMP. (b and c) RMP (b) and input resistance (c) of NPY^WT and NPY^G1KO neurons. (d) Summary of GABA_B-induced hyperpolarization of NPY^WT neurons (black) and NPY^G1KO neurons (red). Solid lines indicate fitting of dose-response curve (Hill slope = 1.0, Y = Bottom + (Top-Bottom)/(1 + 10^(logEC50-X)). Both hyperpolarization and no responses were included in analyses. Table 2 shows amplitudes of hyperpolarization by variable doses of baclofen. (e) Outward currents recorded in NPY^WT neurons (left traces) and NPY^G1KO neurons (right traces), defined as I_Bac, by locally applying 10 μM baclofen (upper row) or 100 μM baclofen (lower row). (f and g) Normalized amplitudes of I_Bac by 10 μM baclofen (f) and by 100 μM baclofen of NPY^WT neurons and NPY^G1KO neurons (g). Data are presented as mean ± SEM. Unpaired t test was used for statistical analyses. **P < .01, ***P < .001, and ns, not significant.

Table 3.

GABA_B-induced hyperpolarization of arcuate NPY neurons

Baclofen concentration (µM)	NPYWT	NPYG1KO
1	−14.7 ± 3.4 mV (n = 8) 7 of 8, 88% (EK+ = −94.9 ± 6.1 mV, n = 7)	−6.5 ± 1.5 mV (n = 12) 8 of 12, 67% (EK+ = −94.8 ± 7.6 mV, n = 8)
10	−11.9 ± 2.2 mV (n = 14) 12 of 14, 86% (EK+ = −104.3 ± 5.3 mV, n = 12)	−14.9 ± 1.2 mV (n = 12) 12 of 12, 100% (EK+ = −108.0 ± 4.4 mV, n = 12)
30	−19.1 ± 2.0 mV (n = 12) 12 of 12, 100% (EK+ = −99.6 ± 4.8 mV, n = 12)	−17.3 ± 3.0 mV (n = 11) 10 of 11, 91% (EK+ = −105.9 ± 8.7 mV, n = 10)
100	−20.0 ± 2.0 mV (n = 14) 14 of 14, 100% (EK+ = −92.7 ± 3.1 mV, n = 14)	−19.9 ± 1.4 mV (n = 12) 12 of 12, 100% (EK+ = −103.8 ± 5.4 mV, n = 12)

GIRK1-Containing GIRK Channels of AgRP Neurons do not Contribute to Maintain Energy Homeostasis

Given the role of GIRK1 subunits in the regulation of NPY neuronal excitability, we tested the metabolic role of this subunit using the Agrp-ires-Cre::Girk1^flox/flox (GIRK1^AgRP-KO) mice, which was generated and validated in our previous study (Oh et al., 2023). We measured body weight and food intake of NCD-fed GIRK1^WT and GIRK1^AgRP-KO mice every week (5-20 weeks), but there was no difference in the body weight of GIRK1^WT mice and GIRK1^POMC-KO mice during this period (Fig. 5a). There was no difference in cumulative food intake (Fig. 5b) or body composition (Fig. 5c-e) between GIRK1^WT mice and GIRK1^AgRP-KO mice. Consistent with these results, we did not observe any differences between genotypes in the weight of liver, kidney, spleen, heart, and quad muscles (Table 4a). In addition, there were no differences in the weight of subcutaneous WAT (IGW and ISCW) or visceral WAT (PGW and MWAT) between genotypes (Table 4a). We also did not see any differences in the weight of BAT between GIRK1^WT mice and GIRK1^AgRP-KO mice (Table 4a). We measured VO₂ and VCO₂ with an indirect calorimetry from 20- to 21-week-old GIRK1^WT and GIRK1^AgRP-KO mice, and found no significant differences (Fig. 5f and g). Likewise, calculated EE and RER did not differ between genotypes (Fig. 5h and i). Total horizontal movements, ambulatory movements, and rearing activities were not different between genotypes as well (Fig. S5). Taken together, we concluded that GIRK1-containing GIRK channels expressed by AgRP neurons do not regulate energy balance.

Fig. 5.

GIRK1-containing GIRK channels expressed by AgRP neurons do not contribute to energy homeostasis of NCD-fed mice. (a) Body weights of GIRK1^WT (n = 8, black) and GIRK1^AgRP-KO mice (n = 8, red). Gene (df = 1, F_{1, 14} = 0.02835, and P = .869), time (df = 5, F_{3.053, 42.74} = 316.7, and P < .0001), and interaction (df = 15, F_{15, 210} =0.6079, and P < .8669). (b) Cumulative food intake of GIRK1^WT (n = 8, black) and GIRK1^AgRP-KO mice (n = 8, red). Gene (df = 1, F_{1, 14} = 0.08833, and P = .771), time (df = 14, F_{1.119, 15.67} = 1964, and P < .0001), and interaction (df = 14, F_{14, 196} = 0.1316, and P > .9999). (c) Lean mass of GIRK1^WT (21.1 ± 0.5 g) and GIRK1^AgRP-KO (20.7 ± 0.6 g, df = 12, t = 0.5195, P = .613). (d) Fat mass of GIRK1^WT (3.1 ± 0.3 g) and GIRK1^AgRP-KO (3.2 ± 0.4 g, df = 12, t = 0.2254, and P = .825). (e) Body fluid of GIRK1^WT (1.9 ± 0.1 g) and GIRK1^AgRP-KO (1.8 ± 0.1 g, df = 12, t = 0.4378, and P = .669). (f) Oxygen consumption (VO₂) (left, dark cycle: 3.26 ± 0.11 L/h/kg, n = 7, for GIRK1^WT and 3.26 ± 0.10 L/h/kg, n = 7, for GIRK1^AgRP-KO, df = 12, t = 0.02895, P = .977; right, light cycle: 2.55 ± 0.06 L/h/kg, n = 7, for GIRK1^WT and 2.55 ± 0.07 L/h/kg, n = 7, for GIRK1^AgRP-KO, df = 12, t = 0.03107, P = .976). (g) Carbon dioxide production (VCO₂) (left, dark cycle: 3.00 ± 0.11 L/h/kg, n = 7, for GIRK1^W and 3.02 ± 0.12 L/h/kg, n = 7, for GIRK1^AgRP-KO, df = 12, t = 0.1398, P = .891; right, light cycle: 2.27 ± 0.05 L/h/kg, n = 7, for GIRK1^WT and 2.30 ± 0.08 L/h/kg, n = 7, for GIRK1^AgRP-KO, df = 12, t = 0.2446, P = .811). (h) Energy expenditure (EE) (left, dark cycle: 0.460 ± 0.015 kcal/h, n = 7, for GIRK1^WT and 0.459 ± 0.016 kcal/h, n = 7, for GIRK1^AgRP-KO, df = 12, t = 0.03941, P = .969; right, light cycle: 0.354 ± 0.009 kcal/h, n = 7, for GIRK1^WT and 0.355 ± 0.011 kcal/h, n = 7, for GIRK1^AgRP-KO, df = 12, t = 0.0921, P = .928). (i) Respiratory ratio (RER) (left, dark cycle: 0.917 ± 0.013, n = 7, for GIRK1^WT and 0.926 ± 0.017, n = 7, for GIRK1^AgRP-KO, df = 12, t = 0.3938, P = .701; right, light cycle: 0.891 ± 0.015, n = 7, for GIRK1^WT and 0.899 ± 0.013, n = 7, for GIRK1^AgRP-KO, df = 12, t = 0.3731, P = .716). Data are presented as mean ± SEM. Two-way ANOVA with Šidák correction (a and b) and unpaired t test (c-i) were used for statistical analysis. ns, not significant.

Table 4.

Organ and fat tissue weight of GIRK1^WT and GIRK1^AgRP-KO mice

a
	Organ weight	Liver	Kidney	Spleen	Heart	Quad muscle
NCD	GIRK1WT	1.38 $\pm$ 0.03 g (n = 7)	0.39 $\pm$ 0.01 g (n = 7)	0.08$\pm$ 0.01 g (n = 7)	0.14$\pm$ 0.00 g (n = 6)	0.43 $\pm$ 0.01 g (n = 7)
	GIRK1AgRP-KO	1.28 $\pm$ 0.08 g (n = 7) P = .297, df = 12, t = 1.091	0.38$\pm$ 0.02 g (n = 7) P = .372, df = 12, t = 0.9275	0.08$\pm$0.01 g (n = 7) P = .735, df = 12, t = 0.3464	0.13$\pm$ 0.00 g (n = 7) P = .909, df = 11, t = 0.0071	0.44 $\pm$0.01 g (n = 7) P = .490, df = 12, t = 0.712
	Fat tissue weight	IGW	ISCW	PGW	MWAT	BAT
	GIRK1WT	0.59 $\pm$ 0.05 g (n = 7)	0.16 $\pm$ 0.02 g (n = 7)	0.56 $\pm$ 0.07 g (n = 7)	0.25 $\pm$ 0.03 g (n = 7)	0.05 $\pm$ 0.01 g (n = 7)
	GIRK1AgRP-KO	0.64 $\pm$ 0.11 g (n = 7) P = .699, df = 12, t = 0.3967	0.18 $\pm$ 0.03 g (n = 7) P = .490, df = 12, t = 0.7122	0.63 $\pm$ 0.14 g (n = 7) P = .679, df = 12, t = 0.4246	0.26 $\pm$ 0.07 g (n = 7) P = .887, df = 12, t = 0.1449	0.04 $\pm$ 0.01 g (n = 7) P = .4879, df = 12, t = 0.7157
b
	Organ weight	Liver	Kidney	Spleen	Heart	Quad muscle
HFD	GIRK1WT	2.16 $\pm$ 0.16 g (n = 14)	0.47$\pm$ 0.02 g (n = 14)	0.13 $\pm$ 0.01 g (n = 14)	0.18 $\pm$ 0.01 g (n = 14)	0.40 $\pm$ 0.02 g (n = 14)
	GIRK1AgRP-KO	2.33 $\pm$ 0.14 g (n = 14) P = .428, df = 26, t = 0.8046	0.47 $\pm$ 0.01 g (n = 14) P = .827, df = 26, t = 0.2214	0.13 $\pm$ 0.01 g (n = 14) P = .667, df = 26, t = 0.4352	0.17 $\pm$ 0.00 g (n = 14) P = .089, df = 26, t = 1.768	0.41 $\pm$ 0.01 g (n = 14) P = .551, df = 26, t = 0.6035
	Fat tissue weight	IGW	ISCW	PGW	MWAT	BAT
	GIRK1WT	3.38 $\pm$ 0.28 g (n = 14)	0.83 $\pm$ 0.06 g (n = 14)	1.92 $\pm$ 0.14 g (n = 14)	1.20 $\pm$ 0.14 g (n = 14)	0.15 $\pm$ 0.02 g (n = 14)
	GIRK1AgRP-KO	4.08 $\pm$ 0.15 g (n = 14) P = .035, df = 26, t = 2.221	1.42 $\pm$ 0.34 g (n = 14) P = .092, df = 26, t = 1.748	2.0 $\pm$ 0.11 g (n = 14) P = .666, df = 26, t = 0.4368	1.48 $\pm$ 0.06 g (n = 14) P = .079, df = 26, t = 1.827	0.18 $\pm$ 0.02 g (n = 14) P = .209, df = 26, t = 1.288

GIRK1-Containing GIRK Channels Expressed by POMC Neurons are Dispensable for the Regulation of Energy Homeostasis on HFD

We found in this study that GIRK1-containing GIRK channels influence the excitability of POMC and NPY/AgRP neurons, but the channels expressed by these neurons do not have significant roles to regulate energy homeostasis when the mice are fed NCD. These data suggested that GIRK1 channel subunits expressed by POMC and NPY/AgRP neurons are largely dispensable in basal metabolic states. To test the role of GIRK1 subunits expressed by these neurons in metabolic stress, we repeated whole-cell patch clamp and metabolic experiments using HFD-fed mice.

First, we recorded the membrane potential of POMC^WT neurons and POMC^G1KO neurons obtained from HFD-fed mice. We noted that POMC^G1KO neurons (−44.4 ± 0.8 mV, n = 34, P = .583) have comparable RMP compared to POMC^WT neurons (−43.8 ± 0.6 mV, n = 30) (Fig. 6a and b). Interestingly, the input resistance of POMC^G1KO neurons (1.7 ± 0.1 GΩ, n = 34, P = .005) was higher than that of POMC^WT neurons (1.3 ± 0.1 GΩ, n = 31) (Fig. 6c). Therefore, while the deletion of GIRK1 subunits may have resulted in increased input resistance, RMP remained unchanged probably because the conductance of other channels was also affected by HFD. We noted a significant reduction in GABA_B-induced hyperpolarization of POMC^G1KO neurons (−2.0 ± 0.8 mV, n = 24, P = .003) compared to that of POMC^WT neurons (−7.7 ± 1.5 mV, n = 27) when we used 10 μM baclofen (Fig. 6a and d). Consistent with these results, voltage-clamp experiments demonstrated that the amplitudes of normalized I_Bac were significantly decreased in POMC^G1KO neurons (0.8 ± 0.1 pA/pF, n = 18, P = .002) compared to POMC^WT neurons (1.5 ± 0.1 pA/pF, n = 22) when we used 10 μM baclofen (Fig. 6e and f). These results indicate that GIRK1-containing GIRK channels contribute to GABA_B-induced hyperpolarization of POMC neurons in HFD condition.

Fig. 6.

GIRK1-containing GIRK channels expressed by POMC neurons do not contribute to energy homeostasis of HFD-fed mice. (a) Spontaneous firing and RMP, which is followed by hyperpolarization of POMC^WT neurons (upper trace) and POMC^G1KO neurons (lower trace) by baclofen (10 μM). Dotted lines indicate RMP. (b and c) Bar graphs and dots summarize RMP (b) and input resistance (c) of POMC^WT and POMC^G1KO neurons. (d) Summary of GABA_B-induced hyperpolarization by 10 μM baclofen of POMC^WT neurons and POMC^G1KO neurons. Both hyperpolarization and no responses were included in analyses. (e) Outward currents recorded in POMC^WT neurons (upper trace) and POMC^G1KO neurons (lower trace), defined as I_Bac, by locally applying 10 μM baclofen. (f) Normalized amplitudes of I_Bac by 10 μM baclofen of POMC^WT neurons and POMC^G1KO neurons. (g) Body weights of GIRK1^WT (n = 15, black) and GIRK1^POMC-KO mice (n = 14, blue). Gene (df = 1, F_{1, 27} = 3.864, and P = .060), time (df = 15, F_{1.355, 36.60} = 346.8, and P < .0001), and interaction (df = 15, F_{15, 405} = 3.082, and P < .0001). (h) Lean mass of GIRK1^WT (25.3 ± 0.6 g) and GIRK1^POMC-KO (24.2 ± 0.5 g, df = 27, t = 1.517, P = .141). (i) Fat mass of GIRK1^WT (15.2 ± 1.0 g) and GIRK1^POMC-KO (13.2 ± 1.3 g, df = 27, t = 1.204, and P = .239). (j) Body fluid of GIRK1^WT (3.5 ± 0.1 g) and GIRK1^POMC-KO (3.1 ± 0.2 g, df = 27, t = 1.727, and P = .096). (k) Cumulative food intake of GIRK1^WT (n = 15, black) and GIRK1^POMC-KO mice (n = 14, blue). Gene (df = 1, F_{1, 27} = 0.6054, and P = .443), time (df = 14, F_{1.041, 28.12} = 75.27, and P < .0001), and interaction (df = 14, F_{14, 378} = 0.7604, and P = .712). (l) Oxygen consumption (VO₂) (left, dark cycle: 2.84 ± 0.06 L/h/kg, n = 14, for GIRK1^WT and 3.08 ± 0.13 L/h/kg, n = 14, for GIRK1^POMC-KO, df = 26, t = 1.663, and P = .108; right, light cycle: 2.47 ± 0.07 L/h/kg, n = 14, for GIRK1^WT and 2.64 ± 0.13 L/h/kg, n = 14, for GIRK1^POMC-KO, df = 26, t = 1.151, and P = .260). (m) Carbon dioxide production (VCO₂) (left, dark cycle: 2.09 ± 0.04 L/h/kg, n = 14, for GIRK1^W and 2.28 ± 0.10 L/h/kg, n = 14, for GIRK1^POMC-KO, df = 26, t = 1.706, and P = .100; right, light cycle: 1.84 ± 0.04 L/h/kg, n = 14, for GIRK1^WT and 1.99 ± 0.11 L/h/kg, n = 14, for GIRK1^POMC-KO, df = 26, t = 1.323, and P = .197). (n) Energy expenditure (EE) (left, dark cycle: 0.628 ± 0.012 kcal/h, n = 15, for GIRK1^WT and 0.593 ± 0.016 kcal/h, n = 14, for GIRK1^POMC-KO, df = 27, t = 1.743, and P = .093; right, light cycle: 0.546 ± 0.012 kcal/h, n = 15, for GIRK1^WT and 0.510 ± 0.015 kcal/h, n = 14, for GIRK1^POMC-KO, df = 27, t = 1.876, and P = .072). (o) Respiratory ratio (RER) (left, dark cycle: 0.736 ± 0.007, n = 15, for GIRK1^WT and 0.739 ± 0.005, n = 14, for GIRK1^POMC-KO, df = 27, t = 0.3656, and P = .718; right, light cycle: 0.746 ± 0.008, n = 15, for GIRK1^WT and 0.752 ± 0.006, n = 14, for GIRK1^POMC-KO, df = 27, t = 0.5568, and P = .582). Data are presented as mean ± SEM. Unpaired t test (b-d, f, h, j, and l-o) and 2-way ANOVA with Šidák correction (g and k) were used for statistical analysis. **P < .01 and ns, not significant.

Next, we measured body weight and food intake of HFD-fed GIRK1^WT and GIRK1^POMC-KO mice every week (5-20 weeks). GIRK1^POMC-KO mice weighed slightly less than GIRK1^WT mice during this period, but this difference was not statistically significant (Fig. 6g). Body composition and cumulative food intake did not significantly differ between genotypes (Fig. 6h-k). Consistent with these results, we did not observe any differences between genotypes in the weight of liver, kidney, spleen, heart, and quad muscles (Table 2b). In addition, there were no differences in the weight of subcutaneous WAT (IGW and ISCW) or visceral WAT (PGW and MWAT) between genotypes (Table 2b). We also did not see any differences in the weight of BAT between GIRK1^WT mice and GIRK1^POMC-KO mice (Table 2b). We measured VO₂ and VCO₂ with an indirect calorimetry from 20- to 21-week-old GIRK1^WT and GIRK1^POMC-KO mice, and found no significant differences (Fig. 6l and m). Likewise, calculated EE and RER did not differ between genotypes (Fig. 6n and o). Total horizontal movements, ambulatory movements, and rearing activities were not different between genotypes as well (Fig. S6). Taken together, we concluded that GIRK1-containing GIRK channels expressed by POMC neurons do not regulate energy balance on HFD.

GIRK1-Containing GIRK Channels Expressed by AgRP Neurons are Dispensable for the Regulation of Energy Homeostasis on HFD

We recorded the membrane potential of NPY^WT neurons and NPY^G1KO neurons obtained from HFD-fed mice and found that NPY^G1KO neurons (−42.3 ± 0.5 mV, n = 23, P = .765) have comparable RMP compared to NPY^WT neurons (−42.5 ± 0.40 mV, n = 39) (Fig. 7a and b). Consistent with these data, input resistance of NPY^G1KO neurons (3.5 ± 0.2 GΩ, n = 23, P = .941) was not different from that of NPY^WT neurons (3.6 ± 0.2 GΩ, n = 39) (Fig. 7c). Therefore, GIRK1 subunits do not contribute to either input resistance or RMP of NPY neurons on HFD. We observed no changes in GABA_B-induced hyperpolarization of NPY^G1KO neurons (−11.4 ± 2.8 mV, n = 6, P = .464) compared to that of NPY^WT neurons (−14.5 ± 2.6 mV, n = 11) when we used 10 μM baclofen (Fig. 7a and d). Consistent with these results, voltage-clamp experiments demonstrated that the amplitudes of normalized I_Bac were comparable between NPY^G1KO neurons (1.6 ± 0.2 pA/pF, n = 16, P = .127) compared to NPY^WT neurons (2.1 ± 0.2 pA/pF, n = 17) when we used 10 μM baclofen (Fig. 7e and f). These results indicate that GIRK1-containing GIRK channels do not contribute to GABA_B-induced hyperpolarization of NPY neurons in HFD condition.

Fig. 7.

GIRK1-containing GIRK channels expressed by AgRP neurons do not contribute to energy homeostasis of HFD-fed mice. (a) Spontaneous firing and RMP, which is followed by hyperpolarization of NPY^WT neurons (upper trace) and NPY^G1KO neurons (lower trace) by baclofen (10 μM). Dotted lines indicate RMP. (b and c) Bar graph and dots summarize RMP (b) and input resistance (c) of NPY^WT neurons and NPY^G1KO neurons. (d) Summary of GABA_B-induced hyperpolarization by 10 μM baclofen of NPY^WT neurons and NPY^G1KO neurons (e). Outward currents recorded in NPY^WT (upper trace) and NPY^G1KO (lower trace) neurons, defined as I_Bac, by locally applying 10 μM baclofen. (f) Normalized amplitudes of I_Bac by 10 μM baclofen of NPY^WT and NPY^G1KO neurons. (g) Body weights of GIRK1^WT (n = 14, black) and GIRK1^AgRP-KO mice (n = 15, red). Gene (df = 1, F_{1, 27} = 0.3599, and P = .554), time (df = 15, F_{1.379, 37.22} = 359.4, and P < .0001), and interaction (df = 15, F_{15, 405} = 1.492, and P = .105). (h) Lean mass of GIRK1^WT (24.6 ± 065 g) and GIRK1^AgRP-KO (24.9 ± 0.5 g, df = 27, t = 0.3849, P = .703). (i) Fat mass of GIRK1^WT (13.6 ± 1.2 g) and GIRK1^AgRP-KO (15.5 ± 1.0 g, df = 27, t = 1.273, and P = .214). (j) Body fluid of GIRK1^WT (3.2 ± 0.1 g) and GIRK1^AgRP-KO (3.4 ± 0.1 g, df = 27, t = 1.073, and P = .293). (k) Cumulative food intake of GIRK1^WT (n = 14, black) and GIRK1^AgRP-KO mice (n = 15, red). Gene (df = 1, F_{1, 27} = 0.6747, and P = .419), time (df = 14, F_{1.054, 28.44} = 68.85, and P < .0001), and interaction (df = 14, F_{14, 378} = 0.8004, and P = .669). (l) Oxygen consumption (VO₂) (left, dark cycle: 3.24 ± 0.09 L/h/kg, n = 13, for GIRK1^WT and 3.12 ± 0.08 L/h/kg, n = 15, for GIRK1^AgRP-KO, df = 26, t = 0.9658, and P = .343; right, light cycle: 2.82 ± 0.07 L/h/kg, n = 13, for GIRK1^WT and 2.70 ± 0.09 L/h/kg, n = 15, for GIRK1^AgRP-KO, df = 26, t = 1.003, and P = .325). (m) Carbon dioxide production (VCO₂) (left, dark cycle: 2.39 ± 0.07 L/h/kg, n = 13, for GIRK1^W and 2.27 ± 0.05 L/h/kg, n = 15, for GIRK1^AgRP-KO, df = 26, t = 1.462, and P = .156; right, light cycle: 2.12 ± 0.06 L/h/kg, n = 13, for GIRK1^WT and 2.01 ± 0.05 L/h/kg, n = 15, for GIRK1^AgRP-KO, df = 26, t = 1.483, and P = .150). (n) Energy expenditure (EE) (left, dark cycle: 0.709 ± 0.033 kcal/h, n = 14, for GIRK1^WT and 0.683 ± 0.017 kcal/h, n = 15, for GIRK1^AgRP-KO, df = 27, t = 0.7239, and P = .475; right, light cycle: 0.620 ± 0.029 kcal/h, n = 14, for GIRK1^WT and 0.593 ± 0.020 kcal/h, n = 15, for GIRK1^AgRP-KO, df = 27, t = 0.7991, and P = .431). (o) Respiratory ratio (RER) (left, dark cycle: 0.737 ± 0.008, n = 14, for GIRK1^WT and 0.729 ± 0.007, n = 15, for GIRK1^AgRP-KO, df = 27, t = 0.8117, and P = .424; right, light cycle: 0.750 ± 0.009, n = 14, for GIRK1^WT and 0.742 ± 0.009, n = 15, for GIRK1^AgRP-KO, df = 27, t = 0.5726, and P = .572). Data are presented as mean ± SEM. Unpaired t test (b-d, f, h-j, and l-o) and 2-way ANOVA with Šidák correction (g-k) were used for statistical analysis. ns, not significant.

Given the negligible role of GIRK1 subunits in regulating the excitability of NPY neurons from HFD-fed mice, we measured body weight and food intake of HFD-fed GIRK1^WT and GIRK1^AgRP-KO mice every week (5-20 weeks). Body weight of GIRK1^AgRP-KO mice and GIRK1^WT mice were similar throughout this period (Fig. 7g), and body composition and cumulative food intake did not significantly differ between genotypes (Fig. 7h-k). Consistent with these results, we did not observe any differences between genotypes in the weight of liver, kidney, spleen, heart, and quadriceps femoris muscles (Table 4b). In addition, there were no differences in the weight of subcutaneous WAT (IGW and ISCW) or visceral WAT (PGW and MWAT) between genotypes (Table 4b). We also did not see any differences in the weight of BAT between GIRK1^WT mice and GIRK1^AgRP-KO mice (Table 4b). We measured VO₂ and VCO₂ with an indirect calorimetry from 20- to 21-week-old GIRK1^WT and GIRK1^AgRP-KO mice, and found no significant differences (Fig. 7l and m). Likewise, calculated EE and RER did not differ between genotypes (Fig. 7n and o). Total horizontal movements, ambulatory movements and rearing activities were not different between genotypes as well (Fig. S7). Taken together, we concluded that GIRK1-containing GIRK channels expressed by AgRP neurons do not regulate energy balance on HFD.

DISCUSSION

In this study, we found that GIRK1-containing GIRK channels contribute to maintain RMP of arcuate POMC and NPY/AgRP neurons. We also noted that GIRK1-containing GIRK channels are responsible for GABA_B-activated K⁺ currents and GABA_B-induced inhibition of arcuate POMC and NPY/AgRP neurons. These results suggested significant contribution of GIRK1-containing GIRK channels to regulate the activity of these neurons. However, POMC neuron-specific deletions of GIRK1 subunits resulted in less body weight only for a short period when the mice were fed NCD. Moreover, we did not see any difference in body weight of NCD-fed mice when we deleted GIRK1 subunits specifically in AgRP neurons. We repeated similar series of experiments using HFD-fed mice and noted that GIRK1-containing GIRK channels no longer contribute to maintain RMP of arcuate POMC and NPY/AgRP neurons. In agreement with the electrophysiological findings, neither POMC nor AgRP neuron-specific deletions of GIRK1 subunits resulted in any significant difference in body weight when the mice were fed HFD. Taken together, we conclude that GIRK1 channel subunits expressed by arcuate POMC and NPY/AgRP neurons are largely dispensable for the regulation of body weight whether the mice are fed NCD or HFD.

Role of GIRK Channel Subunits Expressed by Arcuate POMC Neurons

We previously reported that GIRK1, but not GIRK2, subunits contribute to stabilize RMP of arcuate POMC neurons (Sohn et al., 2011), which we confirmed in this study (Figs. 2 and S2). The increased input resistance suggests that reduced GIRK channel activity underlies the depolarized RMP of POMC^G1KO neurons (Fig. 2), while the input resistance of POMC^G2KO neurons was comparable to that of POMC^WT neurons (Fig. S2). In the current work, however, we noted that both GIRK1 and GIRK2 subunits contribute to GABA_B-induced hyperpolarization of POMC neurons, which does not seem to agree with our previous findings stressing the role of GIRK1, but not GIRK2, subunits (Sohn et al., 2011). This discrepancy is likely to result from the larger number of cells included in the current study: the old study analyzed results from 14 wildtype POMC neurons, 16 GIRK1-deficient POMC neurons, and 7 GIRK2-deficient POMC neurons while the current study included 40 wildtype POMC neurons, 38 GIRK1-deficient POMC neurons, and 39 GIRK2-deficient POMC neurons for similar analyses. The voltage-clamp recordings also supported the idea that both GIRK1 and GIRK2 subunits underlie the GABA_B-activated K⁺ currents of POMC neurons (Figs. 2 and S2). It is interesting to note that while similar levels of GIRK1 and GIRK2 channels are expressed by POMC neurons (Fig. 1) only GIRK1 subunits significantly contribute to maintain RMP. These results suggest that GIRK1-containing GIRK channels underlie the constitutive GIRK channels of POMC neurons. Since GABA_B-activated K⁺ currents involved both GIRK1 and GIRK2 subunits, we predict that constitutive and agonist-activated GIRK currents are mediated by distinct composition of GIRK channels in the same cell. This hypothesis may need to be tested by further experiments. Given the increased excitability of the anorexigenic POMC neurons, we expected a lean phenotype of GIRK1^POMC-KO mice. However, GIRK1^POMC-KO mice weighed less than GIRK1^WT mice only for a short period of time and the body weight became similar between genotypes (Fig. 3). Therefore, we predict that increased POMC neuron activity is sufficient to result in lower body weight early in life, but this is soon compensated by other mechanisms that regulate energy homeostasis and body weight.

When the mice were fed HFD, POMC^G1KO neurons still had higher input resistance than POMC^WT neurons but RMP was no longer different between genotypes (Fig. 6). Since HFD was shown to affect gene expression of multiple ion channels (Roy et al., 2021, Stincic et al., 2021), we considered a possibility that an inward conductance may be upregulated in POMC neurons of HFD conditions. Consistent with this idea, we noted that POMC^WT neurons had more depolarized RMP in HFD conditions (−43.8 ± 0.6 mV, n = 30, P < .0001) compared to that in NCD conditions (−53.9 ± 0.8 mV, n = 81) (Fig. 8a). Indeed, POMC^WT neurons in HFD conditions had significantly lower input resistance (1.4 ± 0.1 GΩ, n = 30, P = .015) compared to that in NCD conditions (1.7 ± 0.1 GΩ, n = 81) (Fig. 8b). These data suggest that in HFD conditions POMC neurons have higher inward conductance that results in depolarized RMP and GIRK1 deficiency may not be sufficient to affect RMP and alter POMC neuron excitability. On the other hand, GABA_B-induced hyperpolarization and GABA_B-activated K⁺ currents were significantly decreased in POMC^G1KO neurons of HFD conditions (Fig. 6). While we noted a tendency toward lower body weight in HFD-fed GIRK1^POMC-KO mice compared to GIRK1^WT mice, the difference was not statistically significant and there was no difference in food intake and EE between genotypes (Fig. 6). These results suggest that GIRK1-containing GIRK channels of arcuate POMC neurons do not have a significant role in the regulation of energy balance in HFD conditions. Finally, we noted that compromised GABA_B-induced hyperpolarization and GABA_B-activated K⁺ currents observed in POMC^G1KO neurons may not be sufficient to have any metabolic effects in vivo.

Fig. 8.

High-fat diet feeding alters intrinsic electrical properties of POMC and NPY neurons. (a and b) Bar graphs and dots demonstrate RMP (a) and input resistance (b) of POMC^NCD neurons and POMC^HFD neurons. (c and d) Bar graphs and dots demonstrate RMP (c) and input resistance (d) of NPY^NCD and NPY^HFD neurons. Data are presented as mean ± SEM. Unpaired t test was used. *P < .05 and ****P < .0001.

Role of GIRK Channel Subunits Expressed by Arcuate NPY/AgRP Neurons

We previously reported that GIRK2-containing GIRK channels maintain RMP of NPY/AgRP neurons and that GIRK2 subunits expressed by these neurons regulate EE and body weight (Oh et al., 2023). In the current study, we found that NPY^G1KO neurons have significantly depolarized RMP compared to NPY^WT neurons, where increased input resistance explained this phenotype (Fig. 4). We also noted compromised GABA_B-induced hyperpolarization and GABA_B-activated K⁺ currents in NPY^G1KO neurons. However, GIRK1 subunits expressed by AgRP neurons did not have any role in the regulation of food intake, EE, and body weight (Fig. 5). When the mice were fed HFD, NPY^WT and NPY^G1KO neurons showed similar RMP and responses to GABA_B receptor stimulation, which was consistent with similar energy balance phenotypes of GIRK1^WT and GIRK1^AgRP-KO mice (Fig. 7). We noted that in HFD conditions NPY^WT neurons had significantly depolarized RMP (−42.5 ± 0.4 mV, n = 39, P < .0001) compared to those in NCD conditions (−47.9 ± 0.9 mV, n = 64) (Fig. 8c), which was similar to observations in POMC neurons. Of note, NPY^WT neurons in HFD conditions had significantly higher input resistance (3.6 ± 0.2 GΩ, n = 39, P < .0001) than those in NCD conditions (2.4 ± 0.1 GΩ, n = 64) (Fig. 8d), suggesting that some outward conductance was reduced in HFD conditions. These data suggest that in HFD conditions NPY neurons have lower outward conductance that results in depolarized RMP and the contribution of GIRK1-containing GIRK channels is not large enough to affect RMP and alter NPY neuron excitability. Taken together, these results suggest that GIRK1-containing GIRK channels of arcuate NPY/AgRP neurons do not have a significant role in the regulation of energy balance both in NCD and HFD conditions.

In our previous study, we did not see any difference in GABA_B-activated K⁺ currents between NPY^WT and NPY^G2KO neurons (Oh et al., 2023). Moreover, GABA_B-induced hyperpolarization was larger in NPY^G2KO neurons possibly due to higher input resistance of these neurons. Given the stronger electrophysiological phenotypes of NPY^G1KO neurons, we had expected a stronger body weight phenotype of GIRK1^AgRP-KO mice. It is not clear what is the reason for this apparent discrepancy, but the differences in the expression levels of GIRK1 and GIRK2 subunits by AgRP neurons may explain these observations: the expression level of GIRK2 subunits was more than 3 times as much as that of GIRK1 subunits (Oh et al., 2023). Therefore, the low expression level of GIRK1 subunits by NPY/AgRP neurons is likely to explain no body weight phenotypes of GIRK1^AgRP-KO mice.

Concluding Remarks

We found that the deletion of GIRK1 channel subunits in POMC and AgRP neurons alters electrical activity ex vivo, but is not sufficient to influence energy homeostasis in vivo. While we suggested possible explanations for this discrepancy above, it should also be considered that mammalian GIRK channels are heterotetramers (GIRK1/GIRK2, GIRK1/GIRK3, GIRK1/GIRK4, GIRK2/GIRK3) or homotetramers (GIRK2/GIRK2) (Luján et al., 2014, Lüscher and Slesinger, 2010). Therefore, deletion of GIRK1 subunits still leaves GIRK2/GIRK3 heterotetramers and GIRK2/GIRK2 homotetramers that serve as functional GIRK channels, as evidenced by the remaining baclofen-induced currents in Figures 2e and 4e. These remaining GIRK channels, as well as other mechanisms that regulate energy homeostasis, may compensate for the loss of GIRK1-containing GIRK channels in POMC and AgRP neurons in vivo. A similar phenomenon was previously reported regarding the contribution of phosphatidylinositol-3-kinase subunits expressed by arcuate POMC neurons (Hill et al., 2008). They reported that phosphatidylinositol-3-kinase p85α and p85β regulatory subunits contribute to leptin-induced activation of POMC ex vivo. However, leptin-induced anorexia was not affected in vivo, probably due to the remaining p110α or p110β catalytic subunits. Together, it appears that in vivo phenotypes do not always accompany ex vivo phenotypes, especially when the molecule being studied has multiple functional subunits.

Author Contributions

Yeeun Choi: Writing—review and editing, Writing—original draft, Investigation, Formal analysis. Eun-Seon Yoo: Methodology, Investigation. Youjin Oh: Writing—review and editing, Writing—original draft, Investigation, Formal analysis. Jong-Woo Sohn: Writing—review and editing, Writing—original draft, Supervision, Data curation, Conceptualization.

Declaration of Competing Interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

ORCID

Yeeun Choi https://orcid.org/0009-0000-2158-4592

Eun-Seon Yoo https://orcid.org/0000-0002-2401-0620

Youjin Oh https://orcid.org/0000-0003-0906-6359

Jong-Woo Sohn https://orcid.org/0000-0002-2840-4176

Appendix A. Supplemental material

Supplementary material

.

Supplementary material

.

Supplementary material

.

Supplementary material

.

Supplementary material

.

Supplementary material

.

Supplementary material

.

Supplementary material

.