- Email: [email protected]
Spontaneously active GABAergic interneurons in the subfornical organ of rat slice preparations
Neuroscience Letters 306 (2001) 45±48
www.elsevier.com/locate/neulet
Spontaneously active GABAergic interneurons in the subfornical organ of rat slice preparations Eiko Honda a, Sheng-Hong Xu a, Kentaro Ono a, Kayoko Ito a,b, Kiyotoshi Inenaga a,* a Department of Physiology, Kyushu Dental College, Kokurakitaku, Kitakyushu 803-8580 Japan Department of Removable Prosthodontics, Kyushu Dental College, Kokurakitaku, Kitakyushu 803-8580 Japan
b
Received 2 April 2001; received in revised form 18 April 2001; accepted 19 April 2001
Abstract Inhibitory postsynaptic currents (IPSCs) were recorded from subfornical organ (SFO) neurons in slice preparations of rats, using whole-cell voltage clamp techniques. Some SFO neurons showed bimodal distributions in amplitude with the large and small IPSCs. The large IPSCs vanished in the tetrodotoxin perfusion medium, but the small did not. Both sizes of the IPSCs were completely abolished by application of bicuculline and picrotoxin. Further subpopulation of SFO neurons with the bimodal distributions showed intermittent bursts of the large IPSCs. Immunohistochemical approach revealed existence of gamma-aminobutyric acid (GABA)-immunoreactive neurons and axons in the SFO. These suggest that spontaneously-active and intermittently-burst-®ring GABA interneurons affect other SFO neurons in slice preparations of rats. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Inhibitory postsynaptic current; Synaptic input; Gamma-aminobutyric acid; Patch clamp
Several reports have suggested that gamma-aminobutyric acid (GABA) and its analogues in¯uence drinking and cardiovascular responses through actions at the subfornical organ (SFO) [1,4,8]. Anatomical [10] and electrophysiological [2,6] studies have indicated that the SFO receives extensive GABAergic synaptic inputs, and recently, we demonstrated cholinergic modulation of the GABAergic synaptic inputs in the SFO [11]. Here we suggest existence of GABAergic interneurons in the SFO and spontaneous modulation by them of SFO neurons, by using whole-cell clamp recordings in rat slice preparations and immunohistochemical techniques. The present study was carried out according to the rules of the Animal Experiment Committee, Kyushu Dental College. Techniques used in making slice preparations and whole cell patch clamp recordings were similar to those reported previously [2,12]. Male Wistar rats weighing 150±250 g were deeply anaesthetized with ketamine (250 mg/kg, s.c.) and decapitated. Slices of 300 mm in thickness were prepared in a cold bathing medium, and were trimmed to contain only the SFO and the hippocampal commissure (SFO/hippocampal commissure slices). In some recordings, * Corresponding author. Tel.: 181-93-582-1131; fax: 181-93582-8288. E-mail address: [email protected] (K. Inenaga).
they were further trimmed to contain only the SFO (SFO slices). Before electrophysiological recording, the slice was transferred to a recording chamber in which it was submerged and perfused with the bathing medium at 328C. A pipette solution used in the microelectrodes contained (in mM): K gluconate 140, MgCl2 1, CaCl2 1, EGTA 5, HEPES 10, Na2ATP 4 (pH 7.2 adjusted with KOH). Potential values were corrected for the junction potential (11 mV). Membrane potentials were clamped to 251 mV. The normal bathing and perfusion media contained (in mM): NaCl 124, KCl 5, KH2PO4 1.24, MgSO4 1.3, CaCl2 2.1, NaHCO3 20, and glucose 10. These bathing and perfusion media were continuously oxygenated with 95% O2 and 5% CO2. Amplitude histograms and time constants in postsynaptic currents were analyzed with the software of AxoGraph (Axon Instruments, Inc.). A bipolar nichrome stimulating electrode, located at fringe of the SFO, was used to evoke inhibitory postsynaptic currents (IPSCs) with pulses of 5±40 V intensity and 0.5 ms duration. When necessary, 6-cyano-7-nitroqunoxaline-2,3-dione at 3 mM or kynurenic acid at 1 mM were applied, to block spontaneous and evoked excitatory postsynaptic currents (EPSCs). The data are given as mean ^ S.E., and n represents the number of neurons tested. The Student's unpaired t-test was performed for analytical comparison (two-tailed).
0304-3940/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 1) 01 86 2- 6
46
E. Honda et al. / Neuroscience Letters 306 (2001) 45±48
Spontaneous IPSCs with both small and large amplitude were recorded from 29 SFO neurons in the voltage-clamp mode (n 25 in SFO/hippocampal commissure slices, n 4 in SFO slices). Amplitude histograms of IPSCs of these all neurons indicated bimodal distributions with large and small sizes, as represented in Fig. 1D. Mean amplitudes of the small and large IPSCs in the 29 neurons were 27.4 ^ 1.3 pA and 86.9 ^ 4.3 pA, respectively. Tetrodotoxin (TTX) at 0.3 mM was applied to nine SFO neurons to block actionpotential dependent neurotransmitter release from nerve terminals (n 5 in SFO/hippocampal commissure slices, n 4 in SFO slices). In eight of nine SFO neurons, TTX abolished the large IPSCs, but not the small IPSCs (Fig. 1A). The responses were reversible. No signi®cant differences were observed in the time constants between the small and large size of IPSCs in the control medium (P . 0:96, n 34 IPSCs for the neuron shown in Fig. 1A), or in the time constants of the small IPSCs between the control medium and the TTX containing medium (P . 0:6, n 34) (Fig. 1C). In one SFO neuron, the large IPSCs still remained in TTX containing solution, together with the small IPSCs. Focal electrical stimulation evoked IPSCs (Fig. 2A). The
sizes of the evoked IPSCs were near to those of the large IPSCs. This was found in ®ve SFO neurons. To test whether both the small and large IPSCs were due to GABAergic synaptic inputs, the GABA receptor antagonists bicuculline at 10 mM (n 3) or picrotoxin at 50 mM (n 4) were applied. Application of bicuculline and picrotoxin completely abolished both the small and large spontaneous IPSCs (data not shown) in all neurons tested, implying that both inhibitory synaptic inputs are mediated through GABAergic receptors. We occasionally found burst activities of IPSCs from 17 SFO neurons in the normal perfusion medium (Fig. 2B). The activities were short and intermittent. Most burst durations ranged from 150 ms to less than 1 min. They varied from neuron to neuron and over time. The burst activities remained in the application of kynurenic acid at 1 mM (n 2). After application of TTX-containing medium, the burst activities disappeared in three of four neurons and were still found in one neuron. Immunocytochemical results showed existence of GABA-immunoreactive (ir) perikarya and axons for antiGABA serum in the SFO (Fig. 3A). It may be worth noting that GABA-ir perikarya were densely found in the surround-
Fig. 1. TTX abolishes large IPSCs but not small IPSCs. (A) Shows that TTX at 0.3 mM completely abolished the large size of IPSCs (. 85 pA) but not the small size (, 85 pA), 45 s after the application. The holding potential was 251 mV. (B) Shows expanded current traces in the control medium (Control) and the TTX-containing medium (TTX). By using a curve-®tting procedure with the software of AxoGraph, the time constants of IPSCs were estimated (C). The curve ®tting analysis was done for 34 events of small and large IPSCs in the control and in the TTX-containing medium. There was no signi®cant difference between the small (Small) and large (Large) IPSCs in the control medium. TTX produced no signi®cant change in the time constants of the small IPSCs (Small-TTX). (D) Shows amplitude histogram in the control medium (Control) and the TTX-containing medium (TTX). Each recording segment of 300 s in duration was analyzed.
E. Honda et al. / Neuroscience Letters 306 (2001) 45±48
47
SFO neurons in the normal perfusion medium, it disappeared in the TTX containing solution. These suggest that some GABAergic interneurons in the SFO spontaneously ®re action potentials in bursts. A recent study shows that SFO neurons can burst with TTX-dependent and subthreshold sodium currents [9]. At present, although we have no direct evidence to show what type of neurotransmitters are included in these bursting SFO neurons, partly they may be GABAergic.
Fig. 2. Comparison in size of evoked IPSCs by focal electrical stimulation and spontaneously active large IPSCs, and appearance of burst activity of large IPSCs. (A) Shows current records from an SFO neuron. (A-a) Shows evoked IPSCs by focal electrical stimulation (arrows, 35 V intensity, 0.5 ms duration). (A-b) Shows a superimposition of 80 current sweeps for a 20 s recording period. Note that the size of evoked IPSCs by focal stimulation is near to that of the large IPSCs. The holding potential was 251 mV. (B) Shows occasional bursts of large IPSCs in the normal perfusion medium, in another SFO neuron. (B-a) and (B-b) in the lower traces show the respective expansions in the upper trace. The holding potential was 251 mV.
ings of vessels and within the fringe of the SFO, compared to the central part of the SFO (Fig. 3B). In this study, we observed TTX-sensitive and -resistant large IPSCs from the SFO neurons. The large IPSCs in most neurons tested in this study were TTX-sensitive. The TTXsensitive large IPSCs might be due to GABA release from the synaptic terminals by propagation of action potentials, which were related to voltage-dependent sodium channels. On the other hands, TTX-resistant large IPSCs might be due to TTX-resistant sodium [7] and calcium components. The TTX-resistant small IPSCs were so-called `miniature synaptic currents'. The present immunohistochemical evidence clearly showed the existence of GABA-ir cells in the SFO. Considering both the present electrophysiological and histological results, it reveals that spontaneously active GABAergic interneurons exist in the SFO. In addition, while burst activity of the IPSCs was observed in some
Fig. 3. Existence of GABA-immunoreactive neurons in the SFO. (B): This Is a magni®cation of (A). Scale bars in (A,B) are 200 and 80 mm, respectively. Five rats were used for GABA immunocytochemistry. Animals, which were treated with 10 ml of colchicine (10 mg/ml, Sigma, i.c.v. injection under anesthesia with sodium pentobarbital (50 mg/kg)) 24 h before killing, were again anaesthetized with the same anesthetic (80 mg/kg) and perfused transcardially with 0.1 M PBS, followed by 4% paraformaldehyde in 0.1 M PB, (pH 7.4). The brains were immediately removed, post®xed in the same ®xative, then cryoprotected in 20% sucrose in 0.1 M PB at 48C overnight. Frozen sections (40 mm thickness) were collected, and rinsed in 0.1 M TBS. The sections were incubated in the anti-GABA antibody (Chemicon, Temecula, CA) at a dilution of 1: 3000 for 24 h at 48C. After rinsing in 0.1 M TBS, they were incubated in biotinylated goat anti-rabbit IgG (Vector) diluted 1: 200 for 2 h at room temperature. The sections were rinsed again in TBS, incubated in ABC diluted 1: 200 for 2 h at room temperature. Finally, the sections were reacted in 0.05% DAB solution containing 0.06% NiCl2 and 0.01% H2O2 for 10 min and then rinsed several times in 0.05 M TB and 0.1 M PB.
48
E. Honda et al. / Neuroscience Letters 306 (2001) 45±48
A major excitatory synaptic input is glutamatergic in the SFO [12]. Glutamatergic neurons with intermittently-burst®rings may innervate GABAergic interneurons and generate the burst activities of inhibitory synaptic events. However, this may not be the case in the SFO because the burst activities of EPSCs have never observed so far in our laboratory, but those of IPSCs are observed even in the application of the glutamate receptor antagonist kynurenic acid in the present study. Our previous report showed that an equilibrium potential of GABA-gated ion channels was more negative than the resting membrane potential [2]. This indicates that the GABAergic synaptic inputs elicit inhibitory responses in SFO neurons. The report also showed that application of the GABAA receptor antagonists bicuculline and picrotoxin depolarized membrane. Our other reports demonstrated that the frequency of IPSCs was high. The mean frequency was 10.8 Hz. For comparison, the mean frequency of EPSCs, which were due to glutamatergic synaptic inputs, was only 0.92 Hz [11]. Taken together with these results, it is considered that GABAergic synaptic inputs must be important factors to tune up neural excitability of SFO neurons. The SFO lacks an effective blood±brain barrier [3]. Accumulating evidence suggest that bioactive substances in the plasma directly affect SFO neurons [3,5]. Considering that GABAir cell bodies were found in the neighbor of the blood vessel and in the fringe of the SFO, it is suggested that GABA interneurons tuning up SFO neurons are further modulated by humoral factors from the plasma and the cerebrospinal ¯uid. [1] Abe, M., Tokunaga, T., Yamada, K. and Furukawa, T., Gamma-aminobutyric acid and taurine antagonize the central effects of angiotensin II and renin on the intake of water and salt, and on blood pressure in rats, Neuropharmacology, 27 (1988) 309±318. [2] Inenaga, K., Nagatomo, T., Honda, E., Ueta, Y. and Yama-
[3] [4] [5]
[6] [7]
[8]
[9]
[10]
[11]
[12]
shita, H., GABAergic inhibitory inputs to subfornical organ neurons in rat slice preparations, Brain Res., 705 (1995) 85± 90. Johnson, A.K. and Gross, P.M., Sensory circumventricular organs and brain homeostatic pathways, FASEB J., 7 (1993) 678±686. Jones, D.L. and Mogenson, G.J., Central injections of spiperone and GABA: attenuation of angiotensin II stimulated thirst, Can. J. Physiol. Pharmacol., 60 (1982) 720±726. McKinley, M.J., Gerstberger, R., Mathai, M.L., Old®eld, J. and Schmid, H., The lamina terminalis and its role in ¯uid and electrolyte homeostasis, J. Clinical Neurosci., 6 (1999) 289±301. Osaka, T., Yamashita, H. and Koizumi, K., Inhibitory inputs to the subfornical organ from the AV3V: involvement of GABA, Brain Res. Bull., 29 (1992) 581±587. Sleeper, A.A., Cummins, T.R., Dib-Hajj, S.D., Hormuzdiar, W. and Tyrrell, L., Changes in expression of two tetrodotoxin-resistant sodium channels and their currents in dorsal root ganglion neurons after sciatic nerve injury but not rhizotomy, J. Neurosci., 20 (2000) 7279±7289. Unger, T., Bles, F., Ganten, D., Lang, R.E., Rettig, R. and Schwab, N.A., GABAergic stimulation inhibits central actions of angiotensin II: pressor responses, drinking and release of vasopressin, Eur. J. Pharmacol., 90 (1983) 1±9. Washburn, D.L., Anderson, J.W. and Ferguson, A.V., A subthreshold persistent sodium current mediates bursting in rat subfornical organ neurnes, J. Physiol., 529 (2000) 359± 371. Weindl, A., Bu¯er, J., Winkler, B., Arzberger, T. and Hatt, H., Neurotransmitters and receptors in the subfornical organ. Immunohistochemical and electrophysiological evidence, In A. Ermisch, R. Landgraf and H.-J. Ruhle (Eds.), Progress in Brain Research, Vol. 91, Elsevier, Amsterdam, 1992, pp. 261±269. Xu, S.-H., Honda, E., Ono, K. and Inenaga, K., Muscarinic modulation of GABAergic transmission to neurons in the rat subfornical organ, Am. J. Physiol. Regulatory Integrative Comp. Physiol., (2001) In press. Xu, S.-H., Inenaga, K., Honda, H. and Yamashita, H., Glutamatergic synaptic inputs activate neurons in the subfornical organ through non-NMDA receptors, J. Auto. Nerv. Syst., 78 (2000) 177±180.
www.elsevier.com/locate/neulet
Spontaneously active GABAergic interneurons in the subfornical organ of rat slice preparations Eiko Honda a, Sheng-Hong Xu a, Kentaro Ono a, Kayoko Ito a,b, Kiyotoshi Inenaga a,* a Department of Physiology, Kyushu Dental College, Kokurakitaku, Kitakyushu 803-8580 Japan Department of Removable Prosthodontics, Kyushu Dental College, Kokurakitaku, Kitakyushu 803-8580 Japan
b
Received 2 April 2001; received in revised form 18 April 2001; accepted 19 April 2001
Abstract Inhibitory postsynaptic currents (IPSCs) were recorded from subfornical organ (SFO) neurons in slice preparations of rats, using whole-cell voltage clamp techniques. Some SFO neurons showed bimodal distributions in amplitude with the large and small IPSCs. The large IPSCs vanished in the tetrodotoxin perfusion medium, but the small did not. Both sizes of the IPSCs were completely abolished by application of bicuculline and picrotoxin. Further subpopulation of SFO neurons with the bimodal distributions showed intermittent bursts of the large IPSCs. Immunohistochemical approach revealed existence of gamma-aminobutyric acid (GABA)-immunoreactive neurons and axons in the SFO. These suggest that spontaneously-active and intermittently-burst-®ring GABA interneurons affect other SFO neurons in slice preparations of rats. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Inhibitory postsynaptic current; Synaptic input; Gamma-aminobutyric acid; Patch clamp
Several reports have suggested that gamma-aminobutyric acid (GABA) and its analogues in¯uence drinking and cardiovascular responses through actions at the subfornical organ (SFO) [1,4,8]. Anatomical [10] and electrophysiological [2,6] studies have indicated that the SFO receives extensive GABAergic synaptic inputs, and recently, we demonstrated cholinergic modulation of the GABAergic synaptic inputs in the SFO [11]. Here we suggest existence of GABAergic interneurons in the SFO and spontaneous modulation by them of SFO neurons, by using whole-cell clamp recordings in rat slice preparations and immunohistochemical techniques. The present study was carried out according to the rules of the Animal Experiment Committee, Kyushu Dental College. Techniques used in making slice preparations and whole cell patch clamp recordings were similar to those reported previously [2,12]. Male Wistar rats weighing 150±250 g were deeply anaesthetized with ketamine (250 mg/kg, s.c.) and decapitated. Slices of 300 mm in thickness were prepared in a cold bathing medium, and were trimmed to contain only the SFO and the hippocampal commissure (SFO/hippocampal commissure slices). In some recordings, * Corresponding author. Tel.: 181-93-582-1131; fax: 181-93582-8288. E-mail address: [email protected] (K. Inenaga).
they were further trimmed to contain only the SFO (SFO slices). Before electrophysiological recording, the slice was transferred to a recording chamber in which it was submerged and perfused with the bathing medium at 328C. A pipette solution used in the microelectrodes contained (in mM): K gluconate 140, MgCl2 1, CaCl2 1, EGTA 5, HEPES 10, Na2ATP 4 (pH 7.2 adjusted with KOH). Potential values were corrected for the junction potential (11 mV). Membrane potentials were clamped to 251 mV. The normal bathing and perfusion media contained (in mM): NaCl 124, KCl 5, KH2PO4 1.24, MgSO4 1.3, CaCl2 2.1, NaHCO3 20, and glucose 10. These bathing and perfusion media were continuously oxygenated with 95% O2 and 5% CO2. Amplitude histograms and time constants in postsynaptic currents were analyzed with the software of AxoGraph (Axon Instruments, Inc.). A bipolar nichrome stimulating electrode, located at fringe of the SFO, was used to evoke inhibitory postsynaptic currents (IPSCs) with pulses of 5±40 V intensity and 0.5 ms duration. When necessary, 6-cyano-7-nitroqunoxaline-2,3-dione at 3 mM or kynurenic acid at 1 mM were applied, to block spontaneous and evoked excitatory postsynaptic currents (EPSCs). The data are given as mean ^ S.E., and n represents the number of neurons tested. The Student's unpaired t-test was performed for analytical comparison (two-tailed).
0304-3940/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 1) 01 86 2- 6
46
E. Honda et al. / Neuroscience Letters 306 (2001) 45±48
Spontaneous IPSCs with both small and large amplitude were recorded from 29 SFO neurons in the voltage-clamp mode (n 25 in SFO/hippocampal commissure slices, n 4 in SFO slices). Amplitude histograms of IPSCs of these all neurons indicated bimodal distributions with large and small sizes, as represented in Fig. 1D. Mean amplitudes of the small and large IPSCs in the 29 neurons were 27.4 ^ 1.3 pA and 86.9 ^ 4.3 pA, respectively. Tetrodotoxin (TTX) at 0.3 mM was applied to nine SFO neurons to block actionpotential dependent neurotransmitter release from nerve terminals (n 5 in SFO/hippocampal commissure slices, n 4 in SFO slices). In eight of nine SFO neurons, TTX abolished the large IPSCs, but not the small IPSCs (Fig. 1A). The responses were reversible. No signi®cant differences were observed in the time constants between the small and large size of IPSCs in the control medium (P . 0:96, n 34 IPSCs for the neuron shown in Fig. 1A), or in the time constants of the small IPSCs between the control medium and the TTX containing medium (P . 0:6, n 34) (Fig. 1C). In one SFO neuron, the large IPSCs still remained in TTX containing solution, together with the small IPSCs. Focal electrical stimulation evoked IPSCs (Fig. 2A). The
sizes of the evoked IPSCs were near to those of the large IPSCs. This was found in ®ve SFO neurons. To test whether both the small and large IPSCs were due to GABAergic synaptic inputs, the GABA receptor antagonists bicuculline at 10 mM (n 3) or picrotoxin at 50 mM (n 4) were applied. Application of bicuculline and picrotoxin completely abolished both the small and large spontaneous IPSCs (data not shown) in all neurons tested, implying that both inhibitory synaptic inputs are mediated through GABAergic receptors. We occasionally found burst activities of IPSCs from 17 SFO neurons in the normal perfusion medium (Fig. 2B). The activities were short and intermittent. Most burst durations ranged from 150 ms to less than 1 min. They varied from neuron to neuron and over time. The burst activities remained in the application of kynurenic acid at 1 mM (n 2). After application of TTX-containing medium, the burst activities disappeared in three of four neurons and were still found in one neuron. Immunocytochemical results showed existence of GABA-immunoreactive (ir) perikarya and axons for antiGABA serum in the SFO (Fig. 3A). It may be worth noting that GABA-ir perikarya were densely found in the surround-
Fig. 1. TTX abolishes large IPSCs but not small IPSCs. (A) Shows that TTX at 0.3 mM completely abolished the large size of IPSCs (. 85 pA) but not the small size (, 85 pA), 45 s after the application. The holding potential was 251 mV. (B) Shows expanded current traces in the control medium (Control) and the TTX-containing medium (TTX). By using a curve-®tting procedure with the software of AxoGraph, the time constants of IPSCs were estimated (C). The curve ®tting analysis was done for 34 events of small and large IPSCs in the control and in the TTX-containing medium. There was no signi®cant difference between the small (Small) and large (Large) IPSCs in the control medium. TTX produced no signi®cant change in the time constants of the small IPSCs (Small-TTX). (D) Shows amplitude histogram in the control medium (Control) and the TTX-containing medium (TTX). Each recording segment of 300 s in duration was analyzed.
E. Honda et al. / Neuroscience Letters 306 (2001) 45±48
47
SFO neurons in the normal perfusion medium, it disappeared in the TTX containing solution. These suggest that some GABAergic interneurons in the SFO spontaneously ®re action potentials in bursts. A recent study shows that SFO neurons can burst with TTX-dependent and subthreshold sodium currents [9]. At present, although we have no direct evidence to show what type of neurotransmitters are included in these bursting SFO neurons, partly they may be GABAergic.
Fig. 2. Comparison in size of evoked IPSCs by focal electrical stimulation and spontaneously active large IPSCs, and appearance of burst activity of large IPSCs. (A) Shows current records from an SFO neuron. (A-a) Shows evoked IPSCs by focal electrical stimulation (arrows, 35 V intensity, 0.5 ms duration). (A-b) Shows a superimposition of 80 current sweeps for a 20 s recording period. Note that the size of evoked IPSCs by focal stimulation is near to that of the large IPSCs. The holding potential was 251 mV. (B) Shows occasional bursts of large IPSCs in the normal perfusion medium, in another SFO neuron. (B-a) and (B-b) in the lower traces show the respective expansions in the upper trace. The holding potential was 251 mV.
ings of vessels and within the fringe of the SFO, compared to the central part of the SFO (Fig. 3B). In this study, we observed TTX-sensitive and -resistant large IPSCs from the SFO neurons. The large IPSCs in most neurons tested in this study were TTX-sensitive. The TTXsensitive large IPSCs might be due to GABA release from the synaptic terminals by propagation of action potentials, which were related to voltage-dependent sodium channels. On the other hands, TTX-resistant large IPSCs might be due to TTX-resistant sodium [7] and calcium components. The TTX-resistant small IPSCs were so-called `miniature synaptic currents'. The present immunohistochemical evidence clearly showed the existence of GABA-ir cells in the SFO. Considering both the present electrophysiological and histological results, it reveals that spontaneously active GABAergic interneurons exist in the SFO. In addition, while burst activity of the IPSCs was observed in some
Fig. 3. Existence of GABA-immunoreactive neurons in the SFO. (B): This Is a magni®cation of (A). Scale bars in (A,B) are 200 and 80 mm, respectively. Five rats were used for GABA immunocytochemistry. Animals, which were treated with 10 ml of colchicine (10 mg/ml, Sigma, i.c.v. injection under anesthesia with sodium pentobarbital (50 mg/kg)) 24 h before killing, were again anaesthetized with the same anesthetic (80 mg/kg) and perfused transcardially with 0.1 M PBS, followed by 4% paraformaldehyde in 0.1 M PB, (pH 7.4). The brains were immediately removed, post®xed in the same ®xative, then cryoprotected in 20% sucrose in 0.1 M PB at 48C overnight. Frozen sections (40 mm thickness) were collected, and rinsed in 0.1 M TBS. The sections were incubated in the anti-GABA antibody (Chemicon, Temecula, CA) at a dilution of 1: 3000 for 24 h at 48C. After rinsing in 0.1 M TBS, they were incubated in biotinylated goat anti-rabbit IgG (Vector) diluted 1: 200 for 2 h at room temperature. The sections were rinsed again in TBS, incubated in ABC diluted 1: 200 for 2 h at room temperature. Finally, the sections were reacted in 0.05% DAB solution containing 0.06% NiCl2 and 0.01% H2O2 for 10 min and then rinsed several times in 0.05 M TB and 0.1 M PB.
48
E. Honda et al. / Neuroscience Letters 306 (2001) 45±48
A major excitatory synaptic input is glutamatergic in the SFO [12]. Glutamatergic neurons with intermittently-burst®rings may innervate GABAergic interneurons and generate the burst activities of inhibitory synaptic events. However, this may not be the case in the SFO because the burst activities of EPSCs have never observed so far in our laboratory, but those of IPSCs are observed even in the application of the glutamate receptor antagonist kynurenic acid in the present study. Our previous report showed that an equilibrium potential of GABA-gated ion channels was more negative than the resting membrane potential [2]. This indicates that the GABAergic synaptic inputs elicit inhibitory responses in SFO neurons. The report also showed that application of the GABAA receptor antagonists bicuculline and picrotoxin depolarized membrane. Our other reports demonstrated that the frequency of IPSCs was high. The mean frequency was 10.8 Hz. For comparison, the mean frequency of EPSCs, which were due to glutamatergic synaptic inputs, was only 0.92 Hz [11]. Taken together with these results, it is considered that GABAergic synaptic inputs must be important factors to tune up neural excitability of SFO neurons. The SFO lacks an effective blood±brain barrier [3]. Accumulating evidence suggest that bioactive substances in the plasma directly affect SFO neurons [3,5]. Considering that GABAir cell bodies were found in the neighbor of the blood vessel and in the fringe of the SFO, it is suggested that GABA interneurons tuning up SFO neurons are further modulated by humoral factors from the plasma and the cerebrospinal ¯uid. [1] Abe, M., Tokunaga, T., Yamada, K. and Furukawa, T., Gamma-aminobutyric acid and taurine antagonize the central effects of angiotensin II and renin on the intake of water and salt, and on blood pressure in rats, Neuropharmacology, 27 (1988) 309±318. [2] Inenaga, K., Nagatomo, T., Honda, E., Ueta, Y. and Yama-
[3] [4] [5]
[6] [7]
[8]
[9]
[10]
[11]
[12]
shita, H., GABAergic inhibitory inputs to subfornical organ neurons in rat slice preparations, Brain Res., 705 (1995) 85± 90. Johnson, A.K. and Gross, P.M., Sensory circumventricular organs and brain homeostatic pathways, FASEB J., 7 (1993) 678±686. Jones, D.L. and Mogenson, G.J., Central injections of spiperone and GABA: attenuation of angiotensin II stimulated thirst, Can. J. Physiol. Pharmacol., 60 (1982) 720±726. McKinley, M.J., Gerstberger, R., Mathai, M.L., Old®eld, J. and Schmid, H., The lamina terminalis and its role in ¯uid and electrolyte homeostasis, J. Clinical Neurosci., 6 (1999) 289±301. Osaka, T., Yamashita, H. and Koizumi, K., Inhibitory inputs to the subfornical organ from the AV3V: involvement of GABA, Brain Res. Bull., 29 (1992) 581±587. Sleeper, A.A., Cummins, T.R., Dib-Hajj, S.D., Hormuzdiar, W. and Tyrrell, L., Changes in expression of two tetrodotoxin-resistant sodium channels and their currents in dorsal root ganglion neurons after sciatic nerve injury but not rhizotomy, J. Neurosci., 20 (2000) 7279±7289. Unger, T., Bles, F., Ganten, D., Lang, R.E., Rettig, R. and Schwab, N.A., GABAergic stimulation inhibits central actions of angiotensin II: pressor responses, drinking and release of vasopressin, Eur. J. Pharmacol., 90 (1983) 1±9. Washburn, D.L., Anderson, J.W. and Ferguson, A.V., A subthreshold persistent sodium current mediates bursting in rat subfornical organ neurnes, J. Physiol., 529 (2000) 359± 371. Weindl, A., Bu¯er, J., Winkler, B., Arzberger, T. and Hatt, H., Neurotransmitters and receptors in the subfornical organ. Immunohistochemical and electrophysiological evidence, In A. Ermisch, R. Landgraf and H.-J. Ruhle (Eds.), Progress in Brain Research, Vol. 91, Elsevier, Amsterdam, 1992, pp. 261±269. Xu, S.-H., Honda, E., Ono, K. and Inenaga, K., Muscarinic modulation of GABAergic transmission to neurons in the rat subfornical organ, Am. J. Physiol. Regulatory Integrative Comp. Physiol., (2001) In press. Xu, S.-H., Inenaga, K., Honda, H. and Yamashita, H., Glutamatergic synaptic inputs activate neurons in the subfornical organ through non-NMDA receptors, J. Auto. Nerv. Syst., 78 (2000) 177±180.