INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

IN MAMMALS, CIRCULATING LEVELS of the neurohypophysial hormones arginine vasopressin (AVP) and oxytocin (Oxy) are regulated by the coordinated activity of magnocellular neurons in the hypothalamic paraventricular (PVN) and supraoptic nuclei (SON) and their accessory nuclei. Diurnal variations in plasma AVP and Oxy levels have been noted in both humans (21) and rats (11, 38). This could imply entrainment of these neurosecretory neurons by the hypothalamic suprachiasmatic nucleus (SCN), the site of a major biological clock regulating circadian rhythms in the mammalian brain (19). Given the importance of AVP and Oxy in hydromineral homeostasis, there is a need to delineate the neural mechanisms involved in transferring SCN's imprint on neuroendocrine function and hormone release in the neurohypophysis.

The PVN, widely regarded as a key brain center for control of autonomic and neuroendocrine function (29), would be a suitable target for entrainment by SCN. In fact, SCN does innervate the hypothalamic area around PVN (3, 35, 36). Additionally, consistent with the immunocytochemistry that reveals GABA in the somata of most SCN neurons (24) is the localization of GABA in axon terminals of these SCN projections (4). Interestingly, although anatomic data suggest that PVN is not a major target of SCN efferents (3, 4, 35, 36), electrophysiological data indicate that at least a portion of these efferents project directly to putative AVP- and Oxy-synthesizing magnocellular neurons to reduce their excitability via monosynaptic inputs that act via postsynaptic GABA_A receptors (13, 14). Therefore, rapid neurotransmission between SCN and PVN magnocellular neurons appears to be mediated via ionotropic GABA receptors, consistent with immunocytochemical and functional impressions.

An alternate role for GABA acting at metabotropic GABA_B receptors may be anticipated given the recent evidence of expression of GABA_B receptors in hypothalamus (17) and evidence that they may contribute to modulation of synaptic inputs intrinsically within SCN itself (5) or to magnocellular neurons in the SON (18, 20, 23). To address whether GABA_B receptors might have a role in modulating a "circadian-relevant" pathway to the neuroendocrine hypothalamus, we investigated the effects of GABA_B receptor agonists and antagonists on fast inhibitory neurotransmission evoked in magnocellular PVN neurons by SCN stimulation. We conclude that this innervation, tentatively involving both AVP- and Oxy-synthesizing PVN neurons, is subject to inhibitory modulation by presynaptic GABA_B receptors.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Preparation. Experiments used brain slices prepared from Long-Evans rats (100-250 g) housed under a 12:12-h light-dark regimen for at least 2 wk. Under methoxyflurane (Janssen) anesthesia, rats were decapitated and their brains were rapidly removed and placed in ice-cold oxygenated (95% O₂, 5% CO₂) artificial cerebrospinal fluid (aCSF) composed of (in mM) 124 NaCl, 3.2 KCl, 26.2 NaHCO₃, 1.3 MgCl₂, 1 NaH₂PO₄, 10 glucose, and 2.4 CaCl₂ at pH of 7.4. Coronal slices 500 µm in thickness containing both PVN and SCN were cut with a vibratome, trimmed, and preincubated in aCSF for 1 h at room temperature and transferred to a submerged chamber for recordings at room temperature.

Electrophysiology. A concentric bipolar stimulating electrode (OD 325 µm, tip-ring separation 100 µm; FHC, Bowdoinham, ME) positioned in SCN and connected to a stimulus isolation unit was used to evoke postsynaptic events in PVN neurons. Recordings were obtained during the subjective dark cycle, using the whole cell patch clamp technique with micropipettes (borosilicate glass; OD 1.5 mm; ID 1.1 mm) filled with a solution containing (in mM) 90 K gluconate, 50 KCl, 10 HEPES, 1 EGTA, 2 Mg-ATP, and 0.2 GTP at pH 7.3 at an osmolarity of 285 mosM; biocytin 5 mM was added for cell visualization after intracellular dialysis. A higher chloride concentration was used to amplify inhibitory postsynaptic currents (IPSCs) at a holding potential near membrane resting potential. Pipettes had an open resistance of 3-7 M and seal resistances were 3 G. Values were not corrected for liquid junction potential. Data were obtained with an Axopatch 1-D amplifier (Axon Instruments) and low-pass filtered at 2 kHz. Clampex 6.0 or 7.0 (Axon Instruments) was used for online data acquisition, and stored data were analyzed using clampfit 6 (Axon Instrument) and Mini Analysis Program (version 3.0, Synaptasoft, Leonia, NJ). Results are expressed as means ± SE.

Histological identification. Slices were fixed in 4% paraformaldehyde in 0.1 M PBS overnight and subsequently reimmersed in PBS containing 20% sucrose (wt/vol), then sectioned on a cryostat (40 µm), and stained with streptavidin-conjugated Texas red. Although an earlier study (13) indicated that SCN stimulation evoked monosynaptic responses in nonmagnocellular PVN neurons, the current analysis was restricted to neurons that displayed properties of magnocellular cells, in part because these cells have been reported to lack postsynaptic GABA_B receptors (but see Ref. 12) whose activation might induce alterations in resting membrane conductances.

Pharmacology. The following drugs from Tocris Cookson, Ballwin, MO, were bath applied at concentrations indicated in the text: ()-bicuculline methochloride, D()-2-amino- 5-phosphonopentanoic acid (D-AP5), 6-nitro-7-sulphamoyl- benzo(f)quinoxaline-2-3-dione disodium (NBQX), 4- amino-3-(4-chlorophenyl)butanoic acid [(RS)-baclofen], (RS)-3-amino-2-(4-chlorophenyl)-2-hydroxypropyl-sulphonic acid (2-hydroxysaclofen), and nimodipine. All slices were superfused with aCSF containing glutamate receptor antagonists NBQX (5 µM) and D-AP5 (20 µM) to pharmacologically isolate GABA-mediated responses. In experiments to test for G protein involvement, slices were incubated in oxygenated aCSF containing pertussis toxin (PTX) (List Biological Laboratories, Campbell, CA) 1 µg/ml for 16-20 h in room temperature before recording, or, alternatively, slices were preperfused with N-ethylmaleimide (Sigma, St Louis, MO) 50 µM for 5 min before application of the GABA_B receptor agonist.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Identification of magnocellular cells and SCN-evoked IPSCs. Recordings were restricted to neurons anatomically localized within the magnocellular part of the PVN. These cells displayed properties attributed to magnocellular neurons: mean membrane potential of 54.5 ± 0.8 mV, membrane resistance of 864.2 ± 51.4 M, linear current-voltage plots at resting potential levels (Fig. 1Aa), frequency-dependent action potential broadening (Fig. 1Ab), expression of a transient outward current ("A" notch) when a depolarizing current injection is initiated from a hyperpolarized resting membrane potential (Fig. 1Ac; see Refs. 14, 32). By analogy with observations in SON, where immunocytochemically characterized AVP- and Oxy-synthesizing neurons demonstrated differences in their response to hyperpolarizing current pulses triggered from "depolarized" resting levels of 50 mV (28), we tentatively classified (without immunocytochemical confirmation) 22 neurons as putative Oxy-synthesizing cells on the basis of their demonstration of a time-dependent inward rectification to an inward current pulse triggered from a holding level approximately 50 mV and a rebound depolarization of varying magnitude upon return to resting membrane potentials (Fig. 1Ba). Another 20 neurons with little or no time-dependent rectification and delayed return to baseline from hyperpolarized levels (Fig. 1Bb) were tentatively labeled as AVP-synthesizing cells. Consistent with earlier observations in this preparation (13), SCN stimulation evoked IPSCs, blockable with bicuculline (Fig. 1C); their constant latencies and ability to follow stimuli >20 Hz implied a monosynaptic connection.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   A: current clamp traces illustrate electrophysiological properties attributed to magnocellular neurosecretory neurons. Aa: response to a series of intracellular current pulses; Ab: broadened action potential; Ac: delayed onset of action potential firing when the membrane potential is stepped from a hyperpolarized level (82 mV) to a depolarized level. B: when held at a potential of 50 mV, responses to a hyperpolarizing current pulse reveal 2 patterns. Ba: bottom trace from cell illustrated in Aa displays time-dependent rectification and a small rebound depolarization (arrow), tentatively attributed to oxytocin-synthesizing neurons. Bb: trace from a cell that lacks time-dependent rectification and displays a delayed return to baseline (open arrow), tentatively identified as a vasopressin-synthesizing neuron. C: control traces (average of 5 sweeps) obtained in voltage clamp illustrate suprachiasmatic nucleus (SCN)-evoked inhibitory postsynaptic currents (shock artifact designated by *), their blockade in media containing 20 µM bicuculline, and recovery 15 min after drug washout.

Effects of baclofen and 2-hydroxysaclofen on SCN-evoked IPSCs. In the presence of a GABA_B receptor agonist, baclofen (1-10 µM), all of 11 cells tested (holding potential 70 mV; 6 putative Oxy- and 5 putative AVP-synthesizing cells) demonstrated a dose-dependent and reversible reduction in the amplitude of SCN-evoked IPSCs (Fig. 2, A-C). Despite the amplitude reduction, time constants of decay of SCN-evoked IPSCs were not altered significantly (22.2 ± 1.3 ms in control, 22.9 ± 4.7 ms) in the presence of baclofen (n = 7, P > 0.5, paired Student's t-test). In three of three cells tested, the baclofen-induced effects could be prevented by prior perfusion with aCSF containing 2-hydroxysaclofen (300 µM). The observation that the amplitude of SCN-evoked IPSCs did not change in the presence of 2-hydroxysaclofen suggests little or no tonic activation of GABA_B receptors by endogenously released GABA at this particular synapse.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Baclofen depresses SCN-evoked (*) inhibitory postsynaptic currents (IPSCs). Aa: voltage clamp traces (average of 12 sweeps) illustrate SCN-evoked IPSCs during a control period (at time 1 in graph), at end of a 2-min exposure to 3 µM baclofen (2), and after drug washout (3). Ab: superimposition of traces 1 and 2 (latter scaled to size) reveals similarity in time course of decay of IPSCs. B: time course of baclofen (Bac)-induced change in IPSC amplitude [relative to control (Con)] for cell shown above. C: histograms of average relative changes (compared with Con) in SCN-evoked IPSC amplitudes in response to different concentrations of bath applied Bac. Nos. in bars refer to nos. of cells included.

Baclofen and 2-hydroxysaclofen alter paired pulse depression by different mechanisms. Paired-pulse paradigms can be useful in the identification of presynaptic mechanisms (e.g., Refs. 7, 22). All of 12 cells tested with paired SCN stimuli revealed a reduction in the amplitude of the second response (P2) over interstimulus intervals between 100 and 700 ms, peaking around 150-200 ms, i.e., paired-pulse depression or PPD (Fig. 3). Both baclofen and 2-hydroxysaclofen attenuated PPD, although in a different manner. In the presence of 3 µM baclofen, there was a reduction in the amplitude of both evoked IPSCs, similar to the actions reported above. However, in baclofen the P2/P1 amplitude ratio also changed from 0.64 ± 0.10 (control) to 1.06 ± 0.28 (n = 7 cells; P < 0.05; Fig. 3, A and B), in effect abolishing PPD through reduction in the amplitude of the P1 response. In the presence of 300 µM 2-hydroxysaclofen, which did not alter the amplitude of P1 response, the P2/P1 ratio of the evoked IPSCs also changed from 0.56 ± 0.05 (control) to 0.89 ± 0.12 (n = 5 cells; Fig. 3, C and D). Although this also produces reduction in PPD (note the resemblance between histograms in Fig. 3, D and B), the mechanism is different because of an increase in the amplitude of the P2 response. Neither baclofen nor 2-hydroxysaclofen significantly altered resting membrane currents or membrane conductances. These features collectively support the interpretation that presynaptic GABA_B receptors modulate this inhibitory SCN input.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   A: top (control) voltage clamp traces (average of 12 sweeps) of paired (150-ms interval) SCN-evoked IPSCs (shock artifact designated by *) demonstrate paired-pulse depression (PPD) manifested as a reduction in the amplitude of the second (P2) compared with the first response (P1); on right is simultaneous test for membrane conductance. Middle traces obtained in 3 µM Bac reveal loss of PPD and overall reduction in P1 and P2 amplitudes, but no change in membrane conductance. Bottom trace indicates recovery. B: histogram data for 7 neurons tested with 3 µM Bac reflects the PPD amplitude ratio (P2/P1) in control (0.64 ± 0.10) vs. in Bac (1.06 ± 0.28; P < 0.05). C: traces obtained in control (top), during exposure to 2-hydroxysaclofen (Sac; middle), and after recovery (bottom) illustrate reduction in PPD, but no effect on membrane conductance, by Sac. D: histogram data for 5 other neurons exposed to Sac (300 µM) reflect PPD amplitude ratio in control (0.56 ± 0.05) vs. in Sac (0.89 ± 0.12; P < 0.05). Calibration in A also applies to C.

Effects of baclofen and 2-hydroxysaclofen on spontaneous IPSCs. A further indication of a presynaptic site of action was a baclofen-induced, dose-dependent, and reversible reduction in the frequency of spontaneous IPSCs (sIPSCs; n = 6 cells; Fig. 4, A and C). Tests on three cells indicated that this effect could be prevented by prior addition of 2-hydroxysaclofen (300 µM; Fig. 4A). Cumulative probability plots obtained from recordings in the presence of 1 and 10 µM baclofen (Fig. 4B) indicated no significant change in sIPSC amplitude but a marked reduction in the frequency of occurrence of sIPSC frequencies. In media containing 0.5 µM tetrodotoxin, baclofen had similar depressant actions on the frequency of occurrence, but not amplitude, of miniature IPSCs (mIPSCs; n = 4 cells; not illustrated). Interestingly, in tests on six cells with aCSF containing 300 µM 2-hydroxysaclofen, all displayed a significant increase in sIPSC frequency (from 1.13 ± 0.16 to 3.03 ± 0.61 events/s) but not amplitude (44 ± 9.0 vs. 48 ± 7.7 pA).

View larger version (25K): [in this window] [in a new window]   Fig. 4.   A: current traces from a single paraventricular nucelus (PVN) neuron. Control traces reveal abundant spontaneous IPSCs (sIPSCs). Top right, note reduced frequency of sIPSCs in 10 µM Bac. Bottom left, traces reflect blockade of Bac effect by prior and simultaneous application of 300 µM Sac. Bottom right, eventual return of Bac effect after Sac washout. B: results of cumulative probability analyses of amplitude and interevent intervals of spontaneous IPSCs in control (solid line) and during exposure to 10 µM Bac (dashed line). Note shift to right in interevent intervals but no change in the amplitude of sIPSCs. C: summary plot for spontaneous IPSCs observed in individual cells tested with different concentration of Bac (open symbols), and mean values for these data (closed symbols). Although ANOVA test did not show a significant difference (P = 0.15 for frequencies, and P = 0.91 for amplitude), results reflect trend of a dose-dependent Bac-induced depression in frequency of occurence of sIPSCs/s with little effect on their amplitudes.

PTX-sensitive G protein involvement. At the neuronal somata, GABA_B receptors are known to activate G proteins, and their actions are mediated through altered conductances involving inwardly rectifying K⁺ channels or suppression of Ca²⁺ channels (1, 9). However, there is some question as to the channels that mediate responses to activated presynaptic GABA_B receptors and whether these also involve G proteins (33). To test whether responses to baclofen might use PTX-sensitive G proteins, recordings were made from slices that had been incubated in aCSF containing PTX (1 µg/ml) for 16-20 h. Data from paired SCN-evoked IPSCs in three neurons in different slices revealed a P2/P1 ratio of 0.58 ± 0.05, indicating persisting PPD. In the presence of baclofen, the amplitude of SCN-evoked IPSCs was reduced by 45.9 ± 8.9%, altering the amplitude of the P1 response and changing the PPD ratio to 0.85 ± 0.05. In tests on two other neurons in control slices, preperfusion with N-ethylmaleimide, a sulfhydryl-alkylating agent that has been shown to block G protein-effector interactions by alkylating the -subunits of PTX-sensitive GTP-binding proteins (26, 34), did not change the ratio of PPD nor the response to baclofen (not illustrated).

Potassium vs. calcium channel involvement in baclofen's effects. Ionic mechanisms mediating responses to these presynaptic GABA_B receptors were examined briefly by perfusion of slices with different ionic channel blockers. In the presence of 2 mM Ba²⁺, a nonselective K+ channel blocker, the depressant effects of baclofen on SCN-evoked IPSCs persisted and PPD was still evident in all of three cells tested, indicating that K channels may not be involved. Among the various types of calcium channels, the observation that nimodipine (10 µM) failed to alter baclofen's action on SCN-evoked IPSCs or PPD (n = 3; data not illustrated) suggests lack of mediation via L-type channels. Tests with other blockers were not conclusive: application of both -agatoxin TK, a P/Q-type Ca²⁺ channel blocker (0.5 µM) in four cells and -conotoxin GVIA, an N-type Ca²⁺ channel blocker (2 µM) in three cells almost completely abolished SCN-evoked IPSCs. Other approaches will be required to assess whether P/Q- or N-type calcium channels mediate presynaptic GABA_B actions at this synapse.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

In mammalian brain, the SCN is recognized as the locus of the main biological clock governing a majority of circadian rhythms (19). Because the PVN is regarded as a key center for the regulation of autonomic, neuroendocrine, and cardiovascular function (29), the aim of this study was to provide further insight into the synaptic mechanisms through which SCN might entrain magnocellular neuroendocrine neurons. The electrophysiological data confirm the impression from immunocytochemical and functional studies that GABA is a principal neurotransmitter in SCN efferent pathways and that GABA acts at ionotropic postsynaptic GABA_A receptors to mediate rapid inhibitory neurotransmission from SCN to magnocellular neurons both in PVN (13) and SON (6). Consistent with the interpretation from earlier recordings in vivo (6, 14) and by analogy with the cell classification characteristics noted among immunocytochemically identified AVP- and Oxy-synthesizing cells in SON (28), we tentatively suggest that this innervation involves both AVP- and Oxy-synthesizing neurons.

In addition to its postsynaptic role in this specific pathway, we now propose that GABA also acts at presynaptic GABA_B receptors to modulate the inhibitory input from SCN to magnocellular neurons. Localization of the GABA_B receptors on the terminals of SCN afferents, rather than on the postsynaptic neurons, is supported by observations that baclofen, while markedly altering SCN-evoked IPSC amplitudes and reversing PPD, had no significant effect on either the time course of decay of SCN-evoked IPSCs or resting membrane currents and conductances (see Refs. 18, 20, 23 for SON neurons). A significant change in frequency, but not amplitude, of sIPSCs and mIPSCs in the presence of baclofen and 2-hydroxysaclofen is another reflection of a presynaptic GABA_B receptor-mediated action on GABA_A-mediated inputs to magnocellular neurons, albeit the origins of these sIPSCs cannot be defined. In each instance, receptor specificity is indicated by the blocking action of 2-hydroxysaclofen. The observation that the frequency of occurrence of sIPSCs was increased in the presence of 2-hydroxysaclofen suggests tonic activation of GABA_B receptors on magnocellular neurons. However, one might conclude that this was not the case for those synapses involved in SCN-mediated inputs because the 2-hydroxysaclofen treatment had little appreciable influence on the amplitude of the SCN-evoked compound IPSC. Alternatively, uncertainty on this issue arises because the stimulus intensities could have been supramaximal, thereby leaving little room for an amplitude increase. The fact that 2-hydroxysaclofen has little effect on the amplitude of the first of two SCN-evoked IPSCs but markedly enhances the amplitude of the second response suggests that PPD is due to endogenous action of GABA at the GABA_B receptor in the presynaptic terminals. Endogenous presynaptic activation of GABA_B receptors is also supported by the finding that 2-hydroxysaclofen increases the frequencies of spontaneous and miniature IPSCs without changing their amplitudes.

There is considerable information on the nature of the coupling of postsynaptic GABA_B receptors with G proteins and their effector ion channels. Much less is known about their presynaptic receptors and whether these may be a different subtype of GABA_B receptors as suggested by recent observations (e.g., Ref. 39). Given the evidence for the presynaptic location of the GABA_B receptors under study in this report, their resistance to pretreatment with agents that alter PTX-sensitive G proteins indicates the likelihood that these GABA_B receptors couple to ion channels via a PTX-insensitive G protein. Although our study was not definitive, the lack of effect of barium suggests mediation through Ca²⁺ rather than K⁺ channels. Similar conclusions have been reached in other central nervous system sites, including the calyces of Held in the medial nucleus of the trapezoid body (30) and hippocampal CA1 neurons (10). In the latter study, it was concluded that the presynaptic action of baclofen was mediated through both -conotoxin GVIA-sensitive and -agatoxin IVA-sensitive, but not dihydropyridine-sensitive, calcium channels. Although our preliminary data on this issue are inconclusive, they appear to eliminate involvement of L-type Ca²⁺ channels. It is interesting to note that the situation may vary with the area under study; recent studies (27) reveal that GABA_B receptors facilitate L-type and inhibit N-type calcium channels in salamander retinal neurons.

The presence of presynaptic GABA_B metabotropic receptors suggests that information transfer may not be linearly related to firing frequency of SCN neurons. Increased firing of SCN neurons, as noted during subjective light periods (15, 37) should lead to greater release of GABA. This would have the potential for competing consequences at the level of PVN magnocellular neurons: membrane hyperpolarization resulting from activation of postsynaptic GABA_A receptors, yet reduction in the effectiveness of postsynaptic inhibition due to the activation of presynaptic GABA_B receptors. This will be a matter for further investigations.

Although we and others (18, 20, 23) have seen no influence of baclofen on resting membrane currents and/or conductances in magnocellular cells, results from a recent study (12) indicate that magnocellular neurons in SON do have postsynaptic GABA_B receptors, specifically to modulate selective voltage-dependent Ca²⁺ channels. It remains to be demonstrated whether magnocellular neurons in PVN possess similar postsynaptic GABA_B receptors. If so, these would likely have escaped detection, because our protocols held neurons at levels below threshold for their activation. The effects of such postsynaptic GABA_B receptors in these neurons could become evident when two conditions are met: magnocellular neurons are generating action potentials, as seen during hormone release, and in the presence of an ambient concentration of GABA. With ~50% of synapses on magnocellular neurons containing GABA (8) and with intracellular data revealing an abundance of GABA_A-mediated inhibition (25), the latter condition is likely. Postsynaptic GABA_B receptors could be viewed as having a general influence to reduce action potential discharge among actively firing magnocellular neurons, whereas presynaptic GABA_B receptors may be distributed more selectively on specific inputs, such as those arising from SCN neurons. Verification of the latter will require that comparisons be made of the effects of GABA_B receptors on different inputs, both excitatory and/or inhibitory, to PVN magnocellular neurons. Another issue that further investigation will somehow need to address is the nature of the signaling mechanism(s) at presynaptic GABA_B receptors.

Perspectives