Aquaporin‐4 knockout mice exhibit increased hypnotic susceptibility to ketamine

Abstract Purpose This study examines anesthetic/hypnotic effects of ketamine in AQP4 knockout (KO) and wild‐type (WT) mice with the particular focus on neurotransmission. Materials and Methods Ketamine (100 mg/kg) was intraperitoneally injected in 16 WT and 16 KO mice. The hypnotic potencies were evaluated by the loss of the righting reflex (LORR). The amino acids neurotransmitter levels in prefrontal cortex were measured by microdialysis. Results This study demonstrated that AQP4 knockout significantly shortened the latency compared with WT mice (98.0 ± 4.2 vs. 138.1 ± 15.0 s, p < .05) and prolonged duration of LORR (884.7 ± 58.6 vs. 562.0 ± 51.7 s, p < .05) compared with WT mice in LORR induced by ketamine. Microdialysis showed that lack of AQP4 markedly decreased glutamate level within 20 min (p < .05) and increased γ‐aminobutyric acid (GABA) level within 30–40 min (p < .05) after use of ketamine. Moreover, the levels of taurine were remarkably higher in KO mice than in WT mice, but no obvious differences in aspartate were observed between two genotypes. Conclusion AQP4 deficiency led to more susceptibility of mice to ketamine, which is probably due to the modulation of specific neurotransmitters, hinting an essential maintenance of synaptic activity mediated by AQP4 in the action of ketamine.

Glia express a large spectrum of neurotransmitter receptors and ionic channels to sense neuronal signals. Also, they sense and respond to neuronal activity via sophisticated calcium and sodium signaling, which are considered to be involved in the modulation of anesthetic sensitivity (Liu et al., 2016). As an attractive isoform with the features distinct from other aquaporins in central nervous system (CNS), aquaporin-4 (AQP4) is predominantly expressed in glia, especially concentrated in the astrocyte endfeet membranes adjacent to the synaptic cleft and blood vessels (Miller, Moran, & Hall, 2016). AQP4 appears to be involved in the maintenance of cerebral homeostasis and serves as a modulator of astrocytic function (Amiry-Moghaddam, Frydenlund, & Ottersen, 2004;Manley, Binder, Papadopoulos, & Verkman, 2004;Verkman, Binder, Bloch, Auguste, & Papadopoulos, 2006). Astrocytes are crucial elements implicated in neuronal activity and express receptors for most neuroactive compounds. As the importance of astrocytes in neuronal activity, AQP4 has gained sufficient attraction for its roles in the regulation of astrocytic function, such as astroglial migration (Saadoun et al., 2005) and activation (Fan et al., 2008), K + buffering (Padmawar, Yao, Bloch, Manley, & Verkman, 2005), as well as neurotransmission Fan et al., 2005;Sun et al., 2007).
Besides the contribution to support neurons, astrocytes are active participants in neuroregulation, including synaptic plasticity, neuronal communication, and neurosecretion (Chung, Allen, & Eroglu, 2015). It has been extensively demonstrated that astrocytes could "listen and talk" to synapses and mediate synaptic cross talk (Haydon & Carmignoto, 2006). Glutamate, an excitatory neurotransmitter in CNS, possesses powerful neurotoxin, and its accumulation may be implicated in the pathogenesis events such as amyotrophic lateral sclerosis and Alzheimer's disease (Karki, Smith, Johnson, Aschner, & Lee, 2015). γ-Aminobutyric acid (GABA), an inhibitory neurotransmitter known to activate high-affinity GABA receptor, elicits tonic inhibition to maintain a persistently suppressed tone (Yoon & Lee, 2014). Astrocytes participate in retrieving neurotransmitters released from active neurons, in particular glutamate and GABA (Colangelo, Cirillo, Lavitrano, Alberghina, & Papa, 2012), maintaining an appropriate environment for neuronal activity. Neurotransmitter-gated ion channels are reported to be particularly sensitive to anesthetic agents, and general anesthetic action can be induced by potentiating inhibitory neurotransmission and/or weakening excitatory neurotransmission (Belelli, Lambert, Peters, Wafford, & Whiting, 1997;Grasshoff, Rudolph, & Antkowiak, 2005). It has been reported that GABA-induced chloride transport potentiates anesthetic response mediated by anesthetic such as 1-aminoanthracene (Butts et al., 2009). In terms of the important role of AQP4 in the modulation of synaptic activity and neurotransmission, we hypothesized that AQP4 might be responsible for the modulation of general anesthesia. In this study, an AQP4 knockout (KO) mouse model was used to decipher the potential effects of AQP4 on general anesthesia by measuring loss of the righting reflex (LORR) and the levels of several amino acid-type neurotransmitters in the prefrontal cortex (PFC) in mice treated with ketamine.

| ME THODS
16 AQP4 knockout (KO) CD1 mice (aged 12-14 weeks, weighed 20-30 g) were used as an experiment group, and 16 wild-type (WT) CD1 mice (12-14 weeks, 20-30 g) were treated as a control group. No significant differences were found in weights between CD1 and AQP4 null mice (24.51 ± 1.611 vs. 25.59 ± 0.739 s, p = .5464). The AQP4 KO mice were generated as described elsewhere  and had more than 99.99% homogeneity with WT mice after a series of backcrosses. All experiments were performed on body weight-matched male wild-type and AQP4 KO mice. Four mice per cage were housed under a standardized 12-h light/dark cycle, and water and food were available ad libitum. The temperature of the testing room was kept at 22°C ± 2°C, and experiments were conducted between 9:00 and 17:00 on a thermostat. The mice were used only once in all experiments. The research protocols were approved by the Institutional Animal Care and Use Committee, Nanjing Medical University, China.

| Behavioral observations
The sensitivity to ketamine was evaluated by a rating scale as previously described (Petrenko, Tsujita, Kohno, Sakimura, & Baba, 2007;Petrenko et al., 2004). Ketamine (Hengrui Pharmaceuticals Co., Ltd, Jiangsu, China) was dissolved in 0.9% saline solution, and its effective doses were chosen according to previous reports (Petrenko et al., 2004). Each mouse was intraperitoneally (ip) injected with ketamine (100 mg/kg) and placed in a 2-L glass beaker. At a 2-min interval, the beaker was tilted to an angle of approximately 45° with a horizontal plane three times to gently place animals on their backs, and the ability of each mouse to right itself was measured as the anesthetic score according to the rating scale as described by Boast et al. (Boast, Pastor, Gerhardt, Hall, & Liebman, 1988), with minor modifications as follows: 0: a normal righting reflex; +1: the mouse righted itself within 2 s in all three trials (slightly impaired righting reflex); +2: the latency to righting ranged at 2-10 s in three trails (moderately or severely impaired righting reflex); and +3:>10 s with no righting within for all three trials (i.e., LORR). Total anesthetic score (TAS) was the sum of all scores recorded after use of ketamine. The duration of LORR was calculated as the time between the LORR (as measured as a score of +3) and the time at which mice regained the ability to right themselves (as measured as a score of +2).

| The measurement of amino acids in the prefrontal cortex
Concentrations of amino acids in the PFC were measured by microdialysis according to the protocols as described elsewhere Semba, Adachi, & Arai, 2005). In brief, five weight-matched male WT and KO mice were anesthetized with chloral hydrate (350 mg/kg, i.p.) and placed in a stereotaxic apparatus. After the scalp was incised, the skull was cleaned, and two drill holes for the fixing screws were dug. The microdialysis-guided cannula was implanted into the right prefrontal cortex at the coordinates anterior-posterior (AP) +2.3 mm, medial-lateral (ML) -0.6 mm, dorsal-ventral (DV) -1.2 mm according to the stereotaxic atlas (Paxinos & Franklin, 1997). Microdialysis studies were conducted 7 days later in Plexiglas test chambers (Bioanalytical System, Inc., West Lafayette, Ind., USA). A microdialysis probe (MD-2200, Bioanalytical system, Inc., West Lafayette, Ind., USA) with an active length of 2 mm was inserted gently through the guide cannula. The animals were lightly anesthetized with ether to facilitate manual insertion of the microdialysis probe into the guide cannula. The membranes were tested on the day before use.
On the day of testing, the microdialysis probe was perfused with an artificial cerebrospinal fluid (aCSF; 140 mM NaCl, 2.7 mM KCl, 1.4 mM CaCl 2 , 1.2 mM MgCl 2 , 5.0 mM glucose, pH 7.4) at a flow rate of 2 μl/min and lasted for 2 hr to reach an equilibrium state. After that, baseline samples were collected into polypropylene microcentrifuge vials at a 20-min interval for 80 min. Following baseline sample collection, mice were injected with ketamine (100 mg/kg, i.p.), and dialysates were collected every 10 min for 200 min. In all, 20 μl of the dialysate samples was analyzed by high-performance liquid chromatography (HPLC) with fluorescent detection (Zeng et al., 2007). The peak areas were integrated and calculated by means of an external standard calibration.

| Statistical analysis
All values were presented as mean ± SEM. Statistical analysis for anesthetic sensitivity was performed with Student′s t test by SPSS version 20 (SPSS Inc., Chicago, IL, USA). One-way ANOVA was used to analyze the differences in amino acids in the prefrontal cortex. p < .05 was considered statistically significant.

| Effects of AQP4 knockout on hypnotic state induced by ketamine
The LORR assay was conducted to determine whether AQP4 knockout could change the anesthetic/hypnotic responses to ketamine (100 mg/kg, i.p.). As shown in Figure 1, the latency of LORR (98.0 ± 4.2 s) was shorter in KO mice than in WT mice (138.1 ± 15.0 s, p < .05). Ketamine significantly prolonged the duration of LORR in KO mice compared with WT mice (884.7 ± 58.6 vs. 562.0 ± 51.7 s, p < .01). Furthermore, KO mice exhibited a twofold higher TAS than WT mice (29 ± 0.1 vs. 15 ± 0.4 s, p < .01) (Figure 1c). These findings suggested that AQP4 might be involved in anesthesia/hypnosis initiated by ketamine.

| Effects of AQP4 deficiency on neurotransmitter release after use of ketamine
To evaluate changes in the release of neurotransmitters after use of ketamine, microdialysis was used to measure the levels of amino acids in the prefrontal cortex. Baseline levels of amino acids in WT and KO mice are summarized in Table 1. The levels of taurine were significantly increased in KO group compared with WT group (0.511 ± 0.004 vs. 0.425 ± 0.004 μM, p < .05). However, there were no significant differences in the basal levels of aspartate, glutamate, and GABA between WT and KO mice. As shown in Figure 2, the levels of amino acids in the prefrontal cortex changed rapidly in the two groups following ketamine administration. In addition to no statistical difference in the levels of aspartate between KO and WT mice (Figure 2a), the levels of glutamate were decreased significantly in KO mice within the first 20 min (Figure 2b). Moreover, the levels of GABA were increased notably in KO mice during 30-40 min after ketamine treatment (Figure 2c). And the levels of taurine were also remarkably elevated in KO mice at 60 and 90 min (Figure 2d). After F I G U R E 1 Latency of loss of righting reflex (a), duration of loss of righting reflex (b), and (c) total anesthetic scores (TASs) after intraperitoneal administration of ketamine (100 mg/kg) to WT and KO mice. The data were presented as mean ± SEM. *p < .05, **p < .01, WT vs KO mice that, the levels of aspartate, glutamate, GABA, and taurine were slowly recovered to their basal values, respectively.

| D ISCUSS I ON
General anesthesia is a reversible, drug-interfered loss of consciousness. Accumulating evidence has demonstrated that AQP4 knockout or down-regulation leads to increased seizure duration associated with aberrant neuronal activity (Binder, Yao, Zador, et al., 2006), suggesting an implication of AQP4 in the altered brain excitability and synaptic plasticity. As little is known about the role of AQP4 in anesthetic/hypnotic action of ketamine, AQP4 knockout mice were used to illustrate the function of AQP4 in hypnotic state. Our findings demonstrated that AQP4 knockout mice exhibited reduced latency and increased duration of LORR after use of ketamine, suggesting that AQP4 plays an essential role in the susceptibility of mice to ketamine. AQP4, a predominant isoform of aquaporins in adult brain that is abundantly expressed in astrocytes' endfeet, participates in the modulation of synaptic signals and the buffering of extracellular K + (Benarroch, 2005). As demonstrated previously by our laboratory in vitro and in vivo, AQP4 was implicated in cerebral neurotransmission Zeng et al., 2007). It is well known that glutamate is the principal excitatory neurotransmitter released by presynaptic vesicular efflux and uptake by transporters (Halassa, Fellin, & Haydon, 2007;Semba et al., 2005) and that glutamate is a vital element in many physiological processes, such as the plasticity of synaptic transmission and the formation of neural networks (Matute, Domercq, & Sanchez-Gomez, 2006). The disturbance of glutamate level could cause cognition and psychiatric illnesses, such as Alzheimer's disease, Parkinson's disease, epilepsy, and schizophrenia (Karki et al., 2015;Plaitakis et al., 2010;Rucker & McGuffin, 2014). Moreover, the removal of glutamate depends on the highaffinity, sodium-dependent glutamate transporter 1 (GLT-1) on astrocytes, which promotes glutamate uptake from the extracellular space to maintain extracellular glutamate concentrations. AQP4 deficiency down-regulates the expression of GLT-1 on astrocytes and increases glutamate in the prefrontal cortex (Matute et al., 2006;Zeng et al., 2007). Thus, it is inferred that the susceptibility of mice to ketamine may be influenced by knockout of AQP4 due to the alteration in the levels of glutamate in synaptic cleft. Interestingly, using AQP4 knockout mice, we demonstrated that there was no significant difference in the basal values of glutamate between two genotypes, but that the levels of glutamate markedly decreased in KO mice compared with WT mice after use of ketamine. That seems not to be in accordance with the observations of others that AQP4 deficiency down-regulates the expression of GLT-1 and reduces glutamate uptake (Li et al., 2012;Zeng et al., 2007). In some respects, in vitro experiments are impossible to compare with in vivo experiments. Intricate mechanisms are implicated in AQP4, anesthetic state, and glutamate transport. Further explorations are needed to explain the phenomenon.
Ketamine, a phencyclidine derivative agent, has been widely used as an anesthetic that produces a wide spectrum of pharmacological effects, such as sedation, catalepsy, and somatic analgesia (Lodge & Mercier, 2015). It is found that ketamine is associated with neurocognitive performance, such as depression and schizophrenia (Dawson, Morris, & Pratt, 2013). Clinical trials have revealed that ketamine serves as an antidepressant at low doses, but that ketamine can evoke psychotomimetic actions at higher doses (Miller et al., 2016;Salvadore et al., 2012). In addition, ketamine is also an antagonist of the receptors of the N-methyl-D-aspartate (NMDA) family, an ionotropic, ligand-gated, glutamate-sensitive neurotransmitter receptor (Lodge & Mercier, 2015). As general anesthetic can benefit the patient through suppressing of excitatory neurotransmission (Jevtovic-Todorovic, 2016), our findings suggested that increased susceptibility to ketamine in KO mice might be ascribed to the lower levels of glutamate in the prefrontal cortex.
As an inhibitory neurotransmitter, GABA is released mainly from three candidate pathways: synaptic vesicle, GABA transporters, and non-vesicular channel-mediated release (Yoon & Lee, 2014). In particular, GABA transporters in glia promote the release of GABA into extracellular space, and subsequently GABA activates GABA A receptors (Del, Bustamante, Del, & Solis, 2000;Yoon & Lee, 2014).
GABA causes a long-lasting and synchronous inhibition of mitral and granule cells in the olfactory bulb as well as tonic inhibition of neurons in the cerebellum, olfactory bulb, and hippocampus (Yoon & Lee, 2014). The previous studies have demonstrated that AQP4 deletion delays the clearance of K + from the extracellular space, leading to an accumulation of extracellular K + accompanied by persistent depolarization, which induces the release of massive neurotransmitters release (Strohschein et al., 2011). Our study also indicated that the levels of GABA rose rapidly in KO mice after use of ketamine, which may elicit a longer duration of LORR. Previous studies have revealed that GABA can potentiate response to anesthetic, such as 1-aminoanthracene (Butts et al., 2009), indicating the important contribution of GABA to general anesthetic. Taken together with these coherent findings, AQP4 deficiency may enhance the hypnotic susceptibility to ketamine, which at least partially by enhanced high-K + -induced inhibitory neurotransmission.
In addition, the synthesis and release of another endogenous ag-  & Li, 2000), taurine, an essential neurotransmitter, may affect synaptic activity. The synthetic process of them depends on the cross talk among different cells from astrocytes to neurons, and the release event is sensitive to osmotic swelling and cellular volume (Strohschein et al., 2011). As previously reported by Ding et al., combining with our findings, the extracellular levels of taurine in KO mice were significant higher than those in WT mice (Fan et al., 2008). As AQP4 is involved in the regulation of cellular volume and osmotic homeostasis, we supposed that AQP4 deficiency induced the higher levels of taurine, which was responsible for increased hypnotic effect of ketamine.

| CON CLUS IONS
In summary, this study suggested that AQP4 knockout could prolong the duration and shorten the latency of LORR, and increase TAS after use of ketamine, and that lack of AQP4 altered the neurotransmission of glutamate, GABA, and taurine. As astrocyte pathology and neurotransmission disturbance are important factors that devote to the pharmacodynamic actions of general anesthesia, our findings provide direct evidence that AQP4 plays a vital role in the anesthesia/hypnosis effect of ketamine, which is probably due to the altered neurotransmission, highlighting the important contribution of astrocytes in modulating drug-induced general anesthesia.

CO N FLI C T O F I NTE R E S T S
None.