Exp Neurobiol 2022; 31(3): 147-157
Published online June 30, 2022
© The Korean Society for Brain and Neural Sciences
Moonsun Sa1,2, Jung Moo Lee1,2, Mingu Gordon Park1,2, Jiwoon Lim2,3, Jong Min Joseph Kim4, Wuhyun Koh2, Bo-Eun Yoon4 and C. Justin Lee1,2,3*
1KU-KIST Graduate School of Converging Science and Technology, Korea University, Seoul 02841, 2Center for Cognition and Sociality, Institute for Basic Science (IBS), Daejeon 34126, 3IBS School, University of Science and Technology (UST), Daejeon 34126, 4Department of Molecular Biology, Dankook University, Cheonan 31116, Korea
Correspondence to: *To whom correspondence should be addressed.
TEL: 82-42-878-9150, FAX: 82-42-878-9151
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
The principal inhibitory transmitter, γ-aminobutyric acid (GABA), is critical for maintaining hypothalamic homeostasis and released from neurons phasically, as well as from astrocytes tonically. Although astrocytes in the arcuate nucleus (ARC) of the hypothalamus are shown to transform into reactive astrocytes, the tonic inhibition by astrocytic GABA has not been adequately investigated in diet-induced obesity (DIO). Here, we investigated the expression of monoamine oxidase-B (MAOB), a GABA-synthesizing enzyme, in reactive astrocytes in obese mice. We observed that a chronic high-fat diet (HFD) significantly increased astrocytic MAOB and cellular GABA content, along with enhanced hypertrophy of astrocytes in the ARC. Unexpectedly, we found that the level of tonic GABA was unaltered in chronic HFD mice using whole-cell patch-clamp recordings in the ARC. Furthermore, the GABA-induced current was increased with elevated GABAA receptor α5 (GABRA5) expression. Surprisingly, we found that a nonselective GABA transporter (GAT) inhibitor, nipecotic acid (NPA)-induced current was significantly increased in chronic HFD mice. We observed that GAT1 inhibitor, NO711-induced current was significantly increased, whereas GAT3 inhibitor, SNAP5114-induced current was not altered. The unexpected unaltered tonic inhibition was due to an increase of GABA clearance in the ARC by neuronal GAT1 rather than astrocytic GAT3. These results imply that increased astrocytic GABA synthesis and neuronal GABAA receptor were compensated by GABA clearance, resulting in unaltered tonic GABA inhibition in the ARC of the hypothalamus in obese mice. Taken together, GABA-related molecular pathways in the ARC dynamically regulate the tonic inhibition to maintain hypothalamic homeostasis against the HFD challenge.
The arcuate nucleus (ARC) of the hypothalamus is a part of the mediobasal hypothalamus, located near the third ventricle . It is responsible for integrating information and providing inputs to other hypothalamic nuclei as well as inputs to areas outside the hypothalamus . It is also home to essential and diverse neuronal populations that mediate different neuroendocrine and physiological functions . The diverse functions of the ARC are dependent on its diversity of neurons, but its central role is to maintain homeostasis with respect to feeding and metabolism [4-7]. Astrocytes in the ARC participate in brain energy metabolism by controlling glycogen storage, sensing glucose, and supplying fuel to neurons under physiological conditions [8-11]. When mice are challenged by chronic high-fat diet (HFD) feeding, reactive astrocytes are observed in the ARC [12-15]. However, the pathophysiological role of reactive astrocytes in maintaining homeostasis in the ARC remains to be elucidated.
Reactive astrocytes observed under various pathological conditions have been demonstrated to acquire detrimental functions by tonically releasing γ-aminobutyric acid (GABA) [16-20]. The released GABA from reactive astrocytes strongly inhibits the excitability of neighboring neurons and impairs synaptic transmission and plasticity [16, 18, 19, 21]. GABA plays a fundamental role in hypothalamic homeostasis [22-24]. Synaptic GABA is synthesized by glutamic acid in neurons and is released via vesicular exocytosis, whereas non-synaptic tonic GABA is synthesized by monoamine oxidase-B (MAOB)  and released via bestrophin-1 channel . We previously reported that reactive astrocytes in the disease state have shown a significant increase in the level of astrocytic MAOB with an increase in tonic inhibition [16, 17, 19]. Tonic inhibition, distinct from synaptic inhibition, is mediated by extrasynaptic GABAA receptors expressed in neurons . The most common configuration of extrasynaptic GABAA receptors has been shown to contain either α5 (GABRA5) or δ subunit, whereas synaptic GABAA receptors have γ subunit with α1, α2 or α3, and β2/3 subunits [27-30]. The released GABA is taken up by GABA transporters (GATs) both in neurons and astrocytes . GABA transporter 1 (GAT1) isoforms in neurons and GABA transporter 3 (GAT3) isoforms in astrocytes are thought to regulate ambient GABA levels in the brain . Interestingly, the level of extracellular GABA was found to be elevated in the mediobasal hypothalamus, which comprises both the ventromedial hypothalamus and ARC in obese mice . However, the exact extracellular ambient GABA level in the ARC has not been investigated after chronic HFD feeding. Furthermore, the source of extracellular GABA content and how chronic HFD alters extracellular GABAA receptors and GATs remain undetermined.
In this study, we hypothesized that reactive astrocytes in the ARC may play a regulatory role in maintaining homeostasis. We measured the branching of reactive astrocytes and the levels of MAOB, GABA, GABAA receptor, and GATs using immunohistochemistry. We then investigated tonic inhibition using whole-cell patch-clamp recording with drug application. We found unaltered tonic inhibition despite increased expression of MAOB and GABRA5 with elevated GAT1 in the ARC after HFD feeding. Our results propose that the increased GABA clearance compensated for the increased astrocytic GABA synthesis and neuronal GABAA receptor level, leaving the tonic inhibition unaltered.
All animal experiments were conducted according to protocols approved by the Institutional Animal Care and Use Committee of IBS (Daejeon, South Korea). All mice were maintained in a specific pathogen-free animal facility under a 12-h light-dark cycle and allowed free access to water and food. All experiments were performed on C57BL/6J background were used originated from Jackson Laboratory (USA, stock number 000664). 6-week-old male C57BL/6J mice (DBL, Chungbuk, Republic of Korea) were fed a HFD (60% kcal fat, D12492, Research Diets, USA) or chow (Teklad, 2018S, Envigo) for 20 weeks. All experiments were done with age-matched controls.
Mice were anesthetized with vaporized isoflurane and then decapitated to isolate the brain. The isolated brains were excised and submerged in ice-cold NMDG recovery solution containing (in mM): 93 of NMDG, 93 of HCl, 30 of NaHCO3, 20 of HEPES, 25 Glucose, 5 sodium ascorbate, 2.5 KCl, 1.2 NaH2PO4 (pH 7.4). All the solution was gassed with 95% O2 and 5% CO2. The brain was glued onto the stage of a vibrating microtome (Linear Slicer Pro7, D.S.K) and 250-µm-thick coronal slices were prepared. For stabilization, slices were incubated in room temperature for at least 1 h in extracellular aCSF solution containing (in mM): 130 of NaCl, 3.5 of KCl, 24 of NaHCO3, 1.25 of NaH2PO4, 1.5 of CaCl2, 1.5 of MgCl2, and 10 of d-(+)-glucose (pH 7.4) and simultaneously equilibrated with 25℃.
GABAA receptor-mediated current in the ARC was measured as previously described for tonic GABA recording . Whole-cell patch-clamp recording was conducted with the holding potential of -60mV. Pipette electrode (6~8 MΩ) was filled with an internal solution (in mM): 135 of CsCl, 4 of NaCl, 0.5 of CaCl2, 10 of HEPES, 5 of EGTA, 2 of Mg-ATP, 0.5 of Na2-GTP, and 10 of QX-314, pH-adjusted to 7.2 with CsOH (278~285 mOsmol/kg). Baseline current was stabilized with D-AP5 (50 μM, 0106, Tocris, UK) and CNQX (20 μM, 0190, Tocris, UK) to isolate GABAA receptor current from NMDAR and AMPAR currents. Nipecotic acid (300 μM, 211672, Sigma-Aldrich, USA), NO711 (10 μM, N142, Sigma-Aldrich, USA), SNAP5114 (100 μM, 1561, Tocris, UK)-induced currents were recorded to isolate the contribution of neuronal GAT1 and astrocytic GAT3. Electrical signals were digitized and sampled at 10-ms intervals with Digidata 1550 B and the Multiclamp 700 B Amplifier (Molecular Devices, USA) using the pClamp10.2 software. Data were filtered at 2 kHz. Amplitude of the tonic GABA current was measured by the baseline shift in response to the bath application of bicuculline (BIC, 50 μM, 0109, Tocris, UK) using the Clampfit software (ver. 10.6.0.13.). Frequency and amplitude of spontaneous inhibitory post-synaptic currents (sIPSCs) before BIC administration was detected and measured by MiniAnalysis software (ver. 6.0.7., Synaptosoft, USA).
Mice were anaesthetized with isoflurane and perfused with 0.9% saline followed by ice-cold 4% paraformaldehyde (PFA). Excised brains were postfixed overnight at 4℃ and transferred to 30% sucrose for 48 hours and cut with a frozen microtome in coronal 30 μm sections. Brain sections were filled with blocking solution (0.3% Triton X-100, 3% donkey Serum in 0.1M PBS). Primary antibodies were added to blocking solution and slices were incubated in a shaker at 4℃ overnight. Primary antibodies for immunostaining were chicken anti-GFAP (1:500, AB5541, Millipore, USA), mouse anti-MAOB (1:100, sc-515354, Santa Cruz Biotechnology, USA), guinea pig anti-GABA (1:200, AB175, Millipore, USA), rabbit anti-GABRA5 (1:500, ab10098, Abcam, UK), guinea pig anti-NeuN (1:500, ABN90, Millipore, USA), rabbit anti-GAT1 (1:500, AB1570, Millipore, USA), rabbit anti-GAT3 (1:1,000, AB1574, millipore, USA). After washing, sections were incubated with corresponding secondary antibodies purchased from Jackson ImmunoResearch Laboratories (PA, USA) for 1 or 2 hours at room temperature. If needed, DAPI (1:3000, Thermo Fisher Scientific, USA) was added to PBS to visualize the nuclei of the cells. Sections were mounted with fluorescent mounting medium (S3023, Dako, Denmark) and dried. A series of fluorescent images were obtained by Zeiss LSM900 confocal microscope using a 20x, 40x objective. Z stack images were processed using the ZEN Digital Imaging for Light Microscopy blue system (Zeiss, ver. 3.2) and ImageJ (NIH, ver. 1.52s., USA) software.
Confocal microscopic images were analyzed in order to quantify the expression. Fluorescence intensities were calculated using the mean intensity value of each fluorescence pixels in the marker-positive area using the ImageJ (NIH, USA). Confocal images of brain sections immunostained with GFAP antibody were used for Sholl analysis. The Sholl analysis was performed on serially stacked and maximally projected confocal images and applied in Imaris software (Version 9.0.1, Oxford Instruments) constructs serially concentric circles at 10 μm intervals from the center of GFAP signal (soma) to the end of the most distal process of each astrocyte. The number of intercepts of GFAP-positive processes at each circle and the radius of the largest circle intercepting the astrocyte are analyzed.
All analysis were done blindly. The numbers and individual dots refer to the number of cells unless otherwise clarified in figure legends. The significance level is represented as asterisks (*p<0.05; **p<0.01; ***p<0.001; ****p<0.0001; n.s., not significant). Outliers were excluded by Grubb’s test or ROUT method. GraphPad Prism 9.3.1 for Windows (GraphPad Software, USA) was used for these analyses and to create the plots.
To measure the morphological architecture of reactive astrocytes after chronic HFD feeding, we induced mice to attain a bodyweight reached 50 g (Fig. 1A, B). After 20 weeks, we performed immunostaining for GFAP, an astrocyte-specific cytoskeletal protein, and imaged it in the ARC using confocal microscopy in chow diet (Chow) and HFD mice (Fig. 1C). As reported previously [12-15], the intensity of GFAP was significantly increased in HFD mice (Fig. 1D). The morphological hypertrophy of astrocytes in the ARC of HFD mice could be clearly visualized by Sholl analysis (Fig. 1F). GFAP-positive astrocytes were significantly enhanced in the intersections as they moved radially away from soma (Fig. 1G), and the length of the ending branch and the total branch length were significantly increased in HFD mice (Fig. 1H, I). Additionally, there were significant increases in astrocytic MAOB levels (Fig. 1E) with the immunostained GABA signals in astrocytes (Fig. 1J, K). These findings indicate that chronic HFD challenge leads to a state of reactive astrocytes with increased MAOB and GABA in the ARC of the hypothalamus.
To investigate whether tonic inhibition is altered in HFD mice, we measured GABAA receptor-mediated currents in the ARC after chronic HFD feeding using whole-cell patch-clamp recordings (Fig. 2A-C). The tonic current was measured by a current shift during the treatment with the GABAA receptor antagonist, bicuculline (BIC, 50 μM) in Chow mice and HFD mice (Fig. 2D). Surprisingly, the BIC-sensitive tonic inhibition current was not altered in the ARC after chronic HFD (Fig. 2F). There was also no significant difference in the amplitude and frequency of spontaneous inhibitory post-synaptic currents (sIPSCs) in Chow and HFD mice (Fig. 2E, G, H), indicating that GABAergic synaptic transmission was not changed. It has been previously shown that HFD induced oxidative stress and subsequent cellular senescence in the mouse brain . In addition, dopaminergic neurons in Parkinson’s disease mice exhibited decreased cell capacitance that became more severe with age . To test whether chronic exposure to HFD changes the cell size, we measured the cellular capacitance, an indicator of membrane surface area. As a result, the membrane capacitance of the neurons in the ARC was significantly decreased in HFD mice (Fig. 2I), suggesting that chronic HFD challenges may act as toxins to neurons in the ARC and may accelerate senescence compared to age-matched Chow mice. In conclusion, both astrocytic GABA-mediated tonic inhibition and neuronal GABA-mediated phasic inhibition were not altered in chronic HFD mice.
To test the possibility of whether chronic HFD may alter extrasynaptic GABAA receptors in the ARC, we measured the response of extrasynaptic GABAA receptors to a low concentration of GABA (10 µM) (Fig. 3A, B). We induced an inward current shift with bath application of GABA in Chow and HFD mice. We unexpectedly found that there was a significant increase in GABA-induced currents by measuring BIC-induced current shifts in the presence of GABA in the ARC of chronic HFD mice (Fig. 3C, D). Notably, a previous study reported that there was little δ subunit expression of extrasynaptic GABAA receptor in the ARC . Therefore, we investigated the expression of extrasynaptic GABAA receptor α5 subunit, previously reported to be expressed in the ARC and distinct from POMC/CART and AgRP/NPY . We found that there was a significant increase in the GABRA5 expression in the ARC of chronic HFD mice using immunohistochemistry (Fig. 3E, F). However, there was no significant difference in percentages of GABRA5-positive neurons in the ARC after chronic HFD feeding (Fig. 3G). These results suggest that the increased GABA-induced full activation current is possibly due to an increased expression of GABRA5.
To test the possibility that GATs may alter the tonic inhibition in chronic HFD mice, we used a non-selective GAT inhibitor, nipecotic acid (NPA, 300 µM), to inhibit GATs because GAT1 and GAT3 are sometimes considered to work synergistically in the absorption of GABA . We recorded NPA-sensitive currents in the ARC of Chow and HFD mice (Fig. 4A, B). Surprisingly, NPA-induced current was significantly increased in HFD mice (Fig. 4C), indicating that GAT activity is increased after HFD. This raises the possibility that GABA uptake and clearance are enhanced in HFD mice. To investigate whether neuronal GAT1 or astrocytic GAT3 is enhanced in the ARC of HFD mice, we recorded NO711-induced currents in the ARC using a potent and selective GAT1 inhibitor [38, 39], NO711 (10 µM) (Fig. 4D, E). NO711-induced current was significantly enhanced in HFD mice (Fig. 4F). We then recorded SNAP5114-induced current using SNAP5114 (100 µM)  (Fig. 4G, H). However, there was no significant alteration in SNAP5114-induced current (Fig. 4I), suggesting a major role for GAT1, not GAT3 in the unaltered tonic inhibition in the ARC of HFD mice. We found that there was a significant increase in neuronal GAT1 expression (Fig. 4J, K), whereas no change in astrocytic GAT3 expression in the ARC of chronic HFD mice (Fig. 4L, M). Taken together, these findings indicate that the increased content of cellular GABA in the astrocytes and the increased neuronal GABAA receptor are compensated by the increased GAT1 activity.
In summary, we observed unaltered tonic inhibition in the ARC of HFD mice compared to chow diet mice. Chronic HFD increased astrocytic hypertrophy with an increase in GABA production via MAOB and increased extrasynaptic GABRA5. Nevertheless, unaltered tonic inhibition was attributed to an increase in GABA uptake in the ARC (Fig. 5). In our recent study, we have demonstrated that increased tonic inhibition from reactive astrocytes attenuated the firing rate of GABRA5-positive neurons in the lateral hypothalamus of HFD mice . However, the tonic inhibition from reactive astrocytes was preserved in the ARC of HFD mice. These results suggest that unlike in the lateral hypothalamus, the increased GABA clearance by GAT1 compensated for the increased astrocytic GABA synthesis and neuronal GABAA receptor level, leaving the tonic inhibition unaltered in the ARC (Fig. 5).
In this study, we observed that phasic inhibition was not altered in the ARC of chronic HFD mice. However, it has been previously reported that HFD increased the excitability of AgRP neurons with an increase in the frequency and amplitude of IPSCs in AgRP neurons after 8 weeks of HFD feeding . Although synaptic inhibition on AgRP neurons was significantly increased, paradoxically GABA reversal potential was depolarized by HFD . This might be due to the excitatory actions of GABA  during the development of obesity. In 8 weeks of HFD feeding, this stage might be the process of induction of obesity, so hyperexcitability of AgRP neurons can persist despite an increase in inhibitory input. We cautiously speculate that NKCC1 expression may predominate in 8 weeks of HFD, therefore activation of GABAA receptors generates an efflux of chloride and an excitation of AgRP neurons. However, we measured IPSCs in the mice after 5 months of HFD, a time when obesity is fully established and maintained. We expect that KCC2 expression may predominate in 5 months of HFD . In this stage, activation of GABAA receptors might generate an influx of chloride and an inhibition of AgRP neurons. Although we did not measure AgRP neurons specifically, the IPSCs in the ARC were not altered by HFD when obesity is established. These interesting relationships between NKCC1 predomination in the development and KCC2 predomination in the establishment of obesity await future investigation.
Our findings have not provided how tonic inhibition affects specific neuronal cell types in the ARC. Future studies are warranted to determine which specific neurons in the ARC are affected by tonic inhibition after chronic HFD feeding. We suggest that GAT1 in the ARC actively modulates inhibitory signaling by maintaining the GABA release-uptake balance in chronically established obesity. Future studies are needed to determine which neuronal cell types of GAT1 contribute to GABA clearance in the ARC. Our study is expected to provide evidence that GAT1 in the ARC could be a new molecular target for obesity.
This study was supported by the Institute for Basic Science (IBS), Center for Cognition and Sociality (IBS-R001-D2) to C.J.L.