Osmotic perception of GABAergic synaptic transmission in the supraoptic nucleus of rats

Extracellular osmolality plays a crucial role in controlling the activation of neurons. Hypertonic stimulation modulates glutamatergic inputs to the supraoptic nucleus (SON) magnocellular neurosecretory cells (MNCs) putative vasopressin (VP) neurons through capsaicin-insensitive transient receptor potential vanilloid (TRPV) 1 channels on the presynaptic terminals. However, it remains unclear whether osmotic stimulation modulates GABAergic inputs to VP-secreting neurons within punched-out slices containing only the SON and the perinuclear zone. To answer this question, we studied the effects of various osmotic conditions on the miniature GABAergic postsynaptic currents (mGPSCs) using the whole-cell patch-clamp technique on rat SON putative VP-secreting neurons in small slice preparations. We revealed that incubation in hypertonic solution for 2 h reduced both the frequency and amplitude of the mGPSCs to the SON putative VP neurons, whereas the mGPSCs were unaffected when the external osmolality was changed from isotonic to hypotonic. Of interest, we found that changing from a hypertonic to hypotonic environment increased the frequency of the mGPSCs. This effect was independent of TRPV4. We hypothesize that two coordinated mechanisms may play an important role in the regulation of a wide range of physiological functions of VP.: 1) the modulation of GABAA receptor properties by brain-derived neurotrophic factor (BDNF)-induced tyrosine kinase B receptor-mediated signaling under hypertonic conditions, and 2) cell swelling-induced activation of whole-cell anion currents under hypotonic conditions.


Introduction
Magnocellular neurosecretory cells (MNCs) in the supraoptic nucleus (SON) synthesize and secrete arginine vasopressin (VP) and oxytocin (OT). It has long been known that VP which is released into the systemic circulation contributes a regulatory mechanism of body fluid homeostasis. Recently, the possibility that VP may be somatodendritically released into the brain, and may play an important role in social behavior, sexual motivation and pair bonding, and maternal responses to stress (Insel, 2010). The release of VP is closely related to the electrical activity of MNCs, which is modulated by neurotransmitters and neuromodulators. Fast synaptic inputs (from glutamate and GABA A receptors, respectively) are two of the most important regulators of MNC electrical activity (Leng et al., 1999;Shibuya et al., 2000). In general, neuronal excitability in the adult brain is controlled by a balance between synaptic excitation and inhibition mediated by glutamate and GABA, respectively. However, a recent report revealed that GABA is excitatory in adult VP neurons under normal condition (Haam et al., 2012).
In this work, we investigated the effects of various osmotic conditions on mGPSCs using a whole-cell patch-clamp technique on VP neurons in rat SON slice preparations. We revealed that incubation in hypertonic solution for 2 h reduced both the frequency and amplitude of the mGPSCs to the SON VP neurons. Further, we found that the frequency of the mGPSCs increased after an abrupt drop from hyper-to hypo-extracellular osmolality and that this increase was not dependent on TRPV4. The mGPSCs were unaffected when the external osmolality was changed from isotonic to hypotonic. Although the specific biologic mechanisms of this effect remain unclear, prior evidence has suggested that hypertonic-induced brain-derived neurotrophic factor (BDNF) release and hypo-osmotic cell swelling-induced activation of whole-cell anion currents may be involved in this process.

Sampling of supraoptic nucleus slice
All tissue preparations and electrophysiological experiments were performed as previously reported (Ohbuchi et al., 2009;Yokoyama et al., 2010). Experiments were performed on young adult (3∼4-weekold) male Wistar rats weighing 80−150 g. The animals were housed in standard plastic cages at 23-25°C in a 12-h light/dark cycle. All experiments in this study were carried out in accordance with the Physiological Society of Japan under the guidance of the Ethics Committee of Animal Care and Experimentation, University of Occupational and Environmental Health, Japan.
The rats were sacrificed by decapitation, and care was taken to avoid gross contusion or hemorrhage during and after removal of the brain. Brains were quickly removed and cooled in a perfusion medium at 4°C for 1 min. A block containing the hypothalamus was cut from the brain and glued onto the stage of a vibratome-type slicer (Linearslicer Pro 7, DSK, Kyoto, Japan). After the meninges were carefully removed, coronal slices (150 μm) containing SON were cut from the block in the medium at 4°C. The slices were carefully trimmed with a circular punch (inner diameter 1.8 mm) and incubated in the isotonic (300 mOsmol/kg) or hypertonic (350 mOsmol/kg) medium at room temperature (22-24°C) for 2 h until they were used for electrophysiological measurements.

In vitro slice patch-clamp and data analysis
The slice was placed on a glass-bottom chamber and fixed with a grid of parallel nylon threads supported by a U-shaped stainless-steel weight. The volume of the recording chamber was 1 ml and the perfusion rate was 2 ml/minute. Continuous perfusion was performed using an eight-head peristaltic pump (MP-8, Gilson nucleus, Villiers le Bel, France). The solution volume was kept constant by a low-pressure aspiration system. Slices were moved from an incubation solution to chamber filled with the same osmotic solution. The perfusion solutions contained 1 μM TTX. To identify magnocellular neurons in the SON, we used an upright microscope (Axioskop, ZEISS, Germany) with Nomarski optics (×400). All solutions were applied to the slice preparations by using a two-way valve to switch the perfusion solution from the control buffer to a buffer with various osmolality and 4-Phorbol 12, 13-dicaprinate (4α-PDD) (HV 4-4, Hamilton, Reno, NV, USA). The electrodes used in this study were triple-pulled (P-87, Sutter Instrument Co., Novoto, CA, USA) from a glass capillary. The pipettes had a final resistance of 5-8 MΩ when filled. Whole-cell recordings were made from microscopically identified SON neurons in the upper surface layers of the slices at room temperature (22-24°C). A previous immunohistochemical study demonstrated that VP neurons are more common in the caudal and ventral parts of the SON (Rhodes et al., 1981). Moreover, according to the VP-enhanced green fluorescent protein (VP-eGFP) fusion gene/OT-monomeric red fluorescent protein 1 (OT-mRFP1) fusion gene double transgenic rat, the fluorescence of eGFP was observed in the ventral side, whereas the fluorescence of mRFP1 was observed in the dorsal side of the SON (Kato et al., 2011). Therefore, we recorded mGPSCs in the ventral part of the SON MNCs. Currents were recorded with an EPC-9 amplifier (HEKA, Lambrecht, Germany). Signals were digitized with an analog-digital converter. (MacLab/v. 3.5, Castle Hill, Australia) (Ohbuchi et al., 2009) . Only the AC components (selected using a 1-Hz high pass filter) were used for quantitative analyses of synaptic currents in AxoGraph v.3.6.1 (Axon Instruments, Foster Hill, CA, USA). Spontaneous events were automatically screened using an amplitude threshold of 15 pA and were then visually accepted or rejected based on the rise and decay time. Recordings of postsynaptic currents were begun over 5 min after membrane rupture at a holding potential of −20 mV when the currents reached a steady state. Recordings were collected during periods of stable series resistance. The average values of the frequency and amplitude of mGPSCs during the initial 10 min were calculated as baseline values. Following the initial 10 min, the stimulating solution was perfused for 10 min. And then, the wash-out solution which is the same solution as the initial solution was perfused for 10 min. The average frequency and amplitude values for the first 10 min after adding the new buffer were normalized to the baseline values. All data were expressed as mean ± standard error of the mean (SEM). Results were compared by Student's t-test or Welch's t-test. P < 0.05 was considered statistically significant. The number of neurons tested is represented as 'n'.

Hypertonic conditions reduce the frequency and amplitude of mGPSCs
In our previous experiment, we have revealed that short-term (5 min) hypertonic stimulation affected neither the frequency nor the amplitude of the mGPSCs in rat SON slice preparation which is the same-size sample we used in the present study (Yokoyama et al., 2010).

TRPV4 activation alone does not increase mGPSC frequency
TRPV4 is typically activated by the swelling that occurs when cells are exposed to a hypotonic environment, resulting in increased intracellular Ca 2+ (Liedtke et al., 2000). It can also be activated by a number of chemical compounds (Vriens et al., 2007). To test whether activation of TRPV4 results in changes in the mGPSCs, we treated cells with 10 μM 4α-PDD, a TRPV4 agonist, for 10 min in hypertonic conditions and measured changes in the frequency and the amplitude of mGPSCs. As of above, abruptly switching to hypotonic conditions did increase the frequency of the mGPSCs under hypertonic condition, but not under isotonic condition. Therefore, we investigated whether TRPV4 is involved the increase of frequency under hypertonic conditions. As a result, neither the frequency nor the amplitude of the mGPSCs were affected by treatment with 4α-PDD (frequency; 105 ± 4.8 % of baseline, n = 4, p > 0.05: amplitude; 98.0 ± 2.2 % of baseline, n = 4, p > 0.05: Fig. 3C and D).

Discussion
In this study, we revealed that incubation in hypertonic conditions for 2 h reduced both the frequency and amplitude of the mGPSCs. In addition, we found that the frequency of mGPSCs increased when extracellular fluids were changed from hypertonic to hypotonic, but the amplitude was not affected. These changes in mGPSCs frequency occurred in an osmolality-dependent manner, where greater reductions in osmolality were associated with greater increases in frequency. TRPV4 did not appear to contribute to this mechanism.
postsynaptic tyrosine kinase (Trk) B receptors, decreasing the surface expression of GABA A receptors and thus altering synaptic inhibition (Brunig et al., 2001;Jovanovic et al., 2004;Hewitt and Bains, 2006). Activation of the TrkB receptor and its downstream signaling pathway may trigger phosphorylation cascades that may be the mechanistic link between the attenuation of GABA responses and the internalization of GABA A receptors (Cheng and Yeh, 2003).
In agreement with these findings, our previous research demonstrated a decrease in both the frequency and amplitude of GPSCs after BDNF treatment of SON slice preparations (Ohbuchi et al., 2009). In addition, postsynaptic responses to focal GABA application were significantly attenuated after BDNF treatment of dissociated cultures of SON MNCs expressing VP-enhanced green fluorescent protein (Ueta et al., 2005;Ohbuchi et al., 2009). Thus, it appears that hyperosmotic conditions lead to the local release of BDNF in the SON, eventually inhibiting GABAergic postsynaptic inputs. The activation of the TrkB Fig. 3. Hypotonic (250 mOsmol/kg) stimulation for 10 min changed after isotonic (300 mOsmol/kg) solution led to no changes in the frequency (94.9 ± 4.3 % of baseline, n = 5, p > 0.05) or amplitude (96.5 ± 1.9 % of baseline, n = 5, p > 0.05) of mGPSCs. Representative example, cumulative probability for inter-event interval and amplitude, and average time-course of the frequency and amplitude are shown in A and B, respectively. Neither the frequency nor the amplitude of the mGPSCs were affected after treatment with 10 μM 4α-PDD, a TRPV4 agonist, under hypertonic conditions (frequency; 105 ± 4.8 % of baseline, n = 4, p > 0.05: amplitude; 98.0 ± 2.2 % of baseline, n = 4, p > 0.05). Representative example, cumulative probability for inter-event interval and amplitude, and average timecourse of the frequency and amplitude are shown in C and D, respectively.
receptor and its downstream phosphorylation cascade may lead to changes in affinity or channel conductance and the internalization of GABA A receptors, attenuating the inhibitory response. In this work, we found that GABAergic postsynaptic inputs were attenuated after 2 h of exposure to hypertonic solution alongside increases in local BDNF expression and release. Next, we revealed that hypotonic stimulation after hypertonic incubation increased mGPSCs frequency. Previous studies have shown that hypotonic conditions inhibit the firing of VP neurons and that GABA-mediated GPSCs play an essential role in this mechanism (Richard and Bourque, 1995;Wang et al., 2013). However, these studies were performed using brain explants containing the organum vasculosum of the lamina terminalis, glia, and hypothalamic SON. To our knowledge, this study is the first to report that hypotonic stimulation increases the frequency of mGPSCs in a rat SON punched-out small slice preparation.
Once phosphorylated, the TrkB receptor and its associated downstream signaling proteins remain activated for hours, even after removal of BDNF (Choi et al., 2001). Indeed, we previously elucidated that the effects of BDNF on GPSCs were still present 30 min after removal of BDNF (Ohbuchi et al., 2009). Therefore, it is not likely that hypotonic stimulation reversed the effects of BDNF on mGPSCs.
A previous study reported that exposure to a hypotonic solution simultaneously induces cell swelling and activates whole-cell anion currents in dissociated rat SON VP neurons (Sato et al., 2011). This cell swelling-induced activation of whole-cell anion currents was observed immediately after hypotonic treatment (Sato et al., 2011). Substance of the anion channels activated during hypotonic-induced cell swelling have not been identified for now, whereas it is known that aquaporin channels play important roles in H 2 O influx (Giuliani and Peri, 2014). In light of this finding, the inhibition of GABA A receptors upon switching from a hyper-to hypotonic environment may be partially driven by increased susceptibility to Cl − influx during cell swelling.
In the present study, mGPSCs to the SON MNCs were unaffected by changes both from iso-to hypotonic and from hyper-to isotonic external osmolality. These findings were consistent with the idea that there would be no increase in BDNF expression or cell swelling in the isotonic condition.
In conclusion, we hypothesize that two coordinated mechanisms regulate homeostasis in response to changing osmolality: 1) BDNF- induced TrkB receptor-mediated signaling modulates GABA A receptor properties under hypertonic conditions, and 2) cell swelling-induced activation of whole-cell anion currents may lead to increasing susceptibility of GABA A receptors under hypotonic conditions (Fig. 4).
In the present study, certain issues and limitations remain unresolved. First, the hypothesis we propose is preliminary and speculative, so that other various mechanisms cannot be ruled out. For example, it's possible the involvement of glial cells in responses observed in the current experiments. Various stimulation including the hypertonicity-induced glutamate release may retract the astrocytes process which change the microenvironment (Ohbuchi et al., 2015). The hypothesis might be also limited statistically due to small size of the samples. Second, we cannot definitely affirm that MNCs recorded from the ventral parts of the SON are authentic VP-secreting neurons. Third, hypertonic conditions could affect its response to the TRPV4 agonist 4α-PDD. It was shown that the hypertonic condition can also affect the TRPV4 functions as a transducer of tonic stimulation (Alessandri-Haber et al., 2005;Liedtke and Friedman, 2003;Mizuno et al., 2003). Finally, we cannot be completely confident that these findings reflect the physiologic properties of GABA effects in vivo. In general, the GABA action is excitatory in immature neurons. This is because the intracellular Cl − concentration ([Cl − ]i) is high, owing to high levels of the Na + -K + -2Cl − cotransporter (NKCC1), which mediates inward transport of Cl − , and to low levels of the K + -Cl − cotransporter (KCC2), which excludes Cl − from the cell. In most neurons, the GABA response switches from excitation to inhibition during early postnatal development, due to the developmental decrease of the NKCC1 and increase of the KCC2 (Watanabe et al., 2014). However, GABA is excitatory in adult VP neurons under normal condition, suggesting that the classical excitation/inhibition paradigm of synaptic glutamate and GABA control of neuronal excitability does not apply to VP neurons (Haam et al., 2012). The unique characteristics of VP neurons make an understanding more complicated.
Despite these limitations, we believe that our hypothesis is potentially plausible. Further studies of the exact mechanistic details of these interactions are required to improve our understanding of fluid homeostasis regulation in the SON.

Conflicts of interest
The authors have no conflicts of interest to disclose