Amoxapine Inhibits the Delayed Rectifier Outward K Current

Home , Histamine, Ketanserin

0022-3565/10/3322-437–445$20.00 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 332, No. 2 Copyright © 2010 by The American Society for Pharmacology and Experimental Therapeutics 159160/3553965 JPET 332:437–445, 2010 Printed in U.S.A.

Amoxapine Inhibits the Delayed Rectifier Outward Kϩ Current in Mouse Cortical Neurons via cAMP/Protein Kinase A Pathways

Yan-Lin He, Xiao-Qin Zhan, Guang Yang, Ji Sun, and Yan-Ai Mei Institutes of Brain Science, School of Life Sciences and State Key Laboratory of Medical Neurobiology, Fudan University, Shanghai, China Received July 30, 2009; accepted November 12, 2009 Downloaded from

ABSTRACT

Ion channels are known to be modulated by antidepressant pine inhibition of IK. Amoxapine was also found to significantly drugs, but the molecular mechanisms are not known. We have increase intracellular cAMP levels. The effects of amoxapine on shown that the antidepressant drug amoxapine suppresses IK were abolished by preincubation with 5-hydroxytryptamine (5- ϩ rectifier outward K (IK) currents in mouse cortical neurons. At HT) and the antagonists of 5-HT2 receptor. Moreover, intracellular jpet.aspetjournals.org ␮ Ј ␥ a concentration of 10 to 500 M, amoxapine reversibly inhib- application of guanosine 5 -[ thio]-triphosphate increased IK am- ited IK in a dose-dependent manner and modulated both plitudes and prevented amoxapine-induced inhibition. The selec- steady-state activation and inactivation properties. The appli- tive Kv2.1 subunit blocker Jingzhaotoxin-III reduced IK amplitudes cation of forskolin or dibutyryl cAMP mimicked the inhibitory by 30% and also significantly abolished the inhibitory effect of effect of amoxapine on IK and abolished further inhibition by amoxapine. Together these results suggest that amoxapine inhib- amoxapine. N-[2-(p-Bromocinnamylamino)ethyl]-5-iso-quino- its IK in mouse cortical neurons by cAMP/PKA-dependent path- linesulphonamide (H-89), a protein kinase A (PKA) inhibitor, way associated with the 5-HT receptor, and suggest that the

␣ at ASPET Journals on January 16, 2020 augmented IK amplitudes and completely eliminated amoxa- Kv2.1 -subunit may be the target for this inhibition.

Tricyclic antidepressants (TCAs), named after their three- kir4.1 channels in astrocytes (Su et al., 2007). Imipramine ring molecular core structure, were the first successful anti- and structurally related compounds can directly inhibit tran- depressants and have been widely used for the treatment of sient KA currents in rat hippocampal neurons, and it was depression and other psychiatric disorders (Chen et al., 2004; hypothesized that there is a specific imipramine binding site Gillman, 2007). The therapeutic effects of these drugs are consisting of an aliphatic and aromatic site on the external generally thought to be mediated by inhibition of specific side of the channel pore (Kuo, 1998). Moreover, imipramine serotonin [5-hydroxytryptamine (5-HT)] and/or norepineph- and amitriptyline can inhibit G protein-activated inwardly rine transporters in the brain (Baldessarini, 2001), whereas rectifying Kϩ channels expressed in Xenopus oocytes (Taka- the adverse side effects of TCAs have been attributed to hashi et al., 2006). In addition, TCAs have been shown to interactions with muscarinic, adrenergic, and histamine re- modulate Ca2ϩ-activated Kϩ channels and ClϪ channels ceptors (Baldessarini, 2001). In addition to these pharmaco- (Terstappen et al., 2001; Maertens et al., 2002). Moreover, logic interactions, TCAs have been reported to exert addi- TCAs are known to alleviate various pain syndromes by tional effects on diverse ion channels. Nortriptyline is reported to overlapping with the local anesthetic site in Naϩ channels be a potent blocker of voltage-gated inwardly rectifying (Song et al., 2008). Although the effects on neurotransmitter transporters are held to underlie the activity of TCAs, it is possible that their affinity for ion channels may modulate the This work was supported in part by the National Basic Research Program therapeutic effects and side effects of these molecules. of China [Grant 2007CB512303]; Shanghai Leading Academic Discipline Project [Grant B111]; and the National Talent Training Fund in Basic Re- Amoxapine is a second-generation TCA previously intro- search of China [Grant J0630643]. duced as an alternative to traditional TCAs because of its Article, publication date, and citation information can be found at http://jpet.aspetjournals.org. shorter onset of action and reduced cardiotoxicity. The mol- doi:10.1124/jpet.109.159160. ecule was found to exert antipsychotic effects in animal mod-

ABBREVIATIONS: TCA, tricyclic antidepressant; 5-HT, 5-hydroxytryptamine; IK, delayed rectifier potassium current; PKA, protein kinase A; ANOVA, analysis of variance; PBS, phosphate-buffered saline; HEK, human embryonic kidney; GTP␥S, guanosine 5Ј-[␥-thio]-triphosphate; db-cAMP, dibutyryl cAMP; H-89, N-[2-(p-bromocinnamylamino)ethyl]-5-iso-quinolinesulphonamide; 4-AP, 4-aminopyridine; C6-ceramide, dihy- droceramide C6; JZTX-III, Jingzhaotoxin-III; DMSO, dimethyl sulfoxide; IA, transient potassium current. 437 438 He et al.

els (Greenblatt et al., 1978) and was tested as a single-drug actively proliferating fibroblasts for 48 h. On day 4, cultured cells treatment for psychotic depression. In systematic double- were washed and maintained in fresh medium. All the experiments blind trials, amoxapine was found to be of equal efficacy to were carried out with cortical neurons during 6 to 10 days in culture. that of an antidepressant and an antipsychotic (Anton and Patch-Clamp Recordings. Whole-cell currents of cortex neurons were recorded using a patch-clamp technique. Before I recording, Burch, 1990). A recent study found that the antipsychotic K the culture medium was replaced with a bath solution containing activity of amoxapine was comparable with that of risperi- 140 mM NaCl, 2.5 mM KCl, 10 mM HEPES, 1 mM MgCl2, and 0.001 done (Apiquian et al., 2005). Furthermore, because amoxa- mM tetrodotoxin (pH adjusted to 7.4 using NaOH). Soft glass record- pine is now off-patent and can be produced at generic prices, ing pipettes were filled with an internal solution containing 135 mM the drug could potentially rival other typical antipsychotics. potassium gluconate, 10 mM KCl, 10 mM HEPES, 1 mM CaCl2,2mM

Therefore, an understanding of the molecular and cellular MgCl2, and 10 mM EGTA (pH adjusted to 7.3 using KOH). The mechanisms underlying the therapeutic effects and/or side pipette resistance is 5 to 7 M⍀ after filling with the internal solution. effects of amoxapine is not of purely academic interest. Al- All the recordings were performed at room temperature (23–25°C). though some reports have suggested that amoxapine may Data Acquisition and Analysis. All the currents were recorded using an Axopatch 200B amplifier (Axon Instruments, Sunnyvale, also target 5-HT or D2 receptors (Nasu et al., 2000), few studies have addressed the molecular mechanism that is CA) operated in voltage-clamp mode. A Pentium (Intel Corporation, responsible for the action of this molecule. Santa Clara, CA) computer connected to the recording equipment with a Digidata 1300 (MDS Analytical Technologies, Sunnyvale, CA) Kϩ channels are extremely diverse in structure and func- analog-to-digital interface. Current was digitally sampled at 100 ␮s tion and have been held to be among the most important (10 kHz). The current signals were filtered by a 3-kHz, three-pole Downloaded from signaling macromolecules both in neuronal and non-neuro- Bessel filter. Currents were corrected on-line for leak and residual ϩ nal cells. In the central nervous system, K channels’ activity capacitance transients by a P/4 protocol. Data acquisition and anal- determines the shape, frequency, and duration of action po- ysis were performed with pClamp 8.01 software (Axon Instruments) tentials (Melishchuk et al., 1998). In addition, Kϩ channels and/or Origin 6.1 (MicroCal, Northampton, MA). Statistical analysis play a major role in setting and maintaining the resting was performed using the Student’s t test with nonpaired comparison or paired comparisons where it was relevant. Values were given as

membrane potential, and the modification of this potential jpet.aspetjournals.org ϩ Ϯ Ͻ can alter neuronal excitability by activating Ca2 channels mean S.E.M., with n as the number of cell tested. P value 0.05 and modulating neuronal transmitter release (Roeper and was used to denote the statistical difference between groups. When multiple comparisons were made, data were analyzed by a one-way Pongs, 1996). Indeed, blockade of Kϩ channels has been analysis of variance (ANOVA) test. proposed as a potential therapy for diverse disorders, includ- cAMP Assay. cAMP levels were measured as described previ- ing myasthenia gravis, multiple sclerosis, Huntington’s cho- ously (Clark et al., 2004). In brief, 1 ϫ 105 cells were plated in each ϩ rea, and Alzheimer’s disease. The role of K channels is 35-mm dish and grown to confluence. Cells were washed with 1 ml of

underscored by the accumulation of potassium ion that takes medium and preincubated at 37°C for 10 min in the presence of the at ASPET Journals on January 16, 2020 place during repetitive neuronal firing, and anticonvulsant phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (2.5 mM) drugs modulating Kϩ channel activity have potential as a (Sigma-Aldrich, St. Louis, MO). Cells were incubated for an addi- means to control seizure activity (Liu et al., 2007). Although tional 3 min at 37°C with or without amoxapine-varying concentra- ϩ ␮ some TCAs have previously been reported to target K chan- tions of 10, 50, and 100 M. The medium was removed, and 0.5 ml of nels, few investigations have addressed the possibility that HCl (0.1 M) with 0.8% Triton X-100 (Sigma-Aldrich) was added to ϩ the plates. After 10-min incubation at room temperature, the lysate amoxapine might modify the activity of voltage-gated K was removed from the plates and centrifuged for 2 min. The super- channels. natant was collected and assayed for cAMP levels using a direct In the present study, we used patch-clamp technique, im- cAMP enzyme immunoassay kit (Sigma-Aldrich) according to the munocytochemistry, and cAMP assays to investigate the ef- manufacturer’s instructions. ϩ fects of amoxapine on the major K channels of mouse cor- Immunocytochemistry. To detect changes in cAMP immunore- tical neurons. We report that amoxapine inhibits the delayed activity of cultured cortex neurons after amoxapine treatment, cul- ␮ rectifier potassium current (IK) in cortical neurons, and the tured cortex neurons were incubated for 10 min at 37°C in 50 M inhibition is associated with the modulation of cAMP/protein amoxapine. Immunocytochemical assay was performed for the cul- kinase A (PKA)-dependent pathways and the 5-HT receptor tures using the procedure described as follows. Isolated cultured involved. Our results also suggest that amoxapine may se- cortex neurons placed on the coverslips were first washed three times with 0.01 M phosphate-buffered saline (PBS) and were then lectively target the Kv2.1 ␣-subunit. fixed with 4% paraformaldehyde for 30 min, rinsed with PBS three times, and preincubated for1hin6%normal donkey serum (v/v) in Materials and Methods PBS plus 0.1% Triton X-100 at room temperature. The cells were then incubated with rabbit anti-cAMP antibody (1:200 dilution; Mouse Cortex Neuron Culture. Primary neuronal cultures Chemicon, Temecula, CA) for2hinahumidified air chamber. After were prepared from the cerebral neocortex of 15-day-old embryonic incubation, they were rinsed with PBS three times and then further mice as originally described by di Porzio (Prochiantz et al., 1979) incubated with the secondary antibody donkey anti-rabbit IgG with several modifications. In brief, cerebral neocortices were dis- tagged with Texas Red (1:100 dilution; Jackson ImmunoResearch sected from embryonic Institute for Cancer Research (Swiss Hau- Laboratories Inc., West Grove, PA) for 30 min at room temperature. schka) mice. Dissociated cortical neurons were then plated onto Then the coverslips were rinsed twice in PBS and were mounted onto poly-L-lysine-coated (10 ␮g/ml) 35-mm-diameter Petri dishes (Corn- glass slides. ing Life Sciences, Lowell, MA) at a density of 1 ϫ 105 cells/dish. DNA Constructs and Cell Transfection. Total RNA was iso-

Cultured cells were incubated at 37°C with 5% CO2 in Dulbecco’s lated from primary cultured mouse cortical neurons according to the modified Eagle’s medium supplemented with 10% (v/v) fetal calf manufacturer’s instructions (Qiagen Mini RNeasy; QIAGEN, Valen- serum, 10% (v/v) heat-inactivated horse serum, and 1% antibiotic- cia, CA). First-strand synthesis was performed using SuperScript II antimycotic solution. After culturing for 24 h, cytosine 1-␤-D-arabino- reverse transcriptase (Invitrogen, Carlsbad, CA). Amplification was furanoside (5 ␮M) was applied to the culture medium to eliminate performed with the following primer sets: primer Kv2.1: forward, Amoxapine Inhibited IK 439

5Ј-TGGCTCGAGATGCCGGCGGGCATG-3Ј (XhoI); reverse, 5Ј-ATA- Results CAGAATTCGGATACTCTGATCCCT-3Ј (EcoRI), GenBank NM_008420. Mouse Kv2.1 cDNA was ligated into pmCherry-N1 by using XhoI and Central neurons contain two main voltage-dependent out- ϩ EcoRI restriction sites. Each gene will be expressed as fusion to the ward K currents. One is the IK, and the other is transient

N terminus of enhanced green fluorescent protein or mCherry, which potassium current (IA). We first determined whether amox- ϩ was used as the marker to detect the transfected human embryonic apine affects either IA or IK. Outward K currents were kidney (HEK) 293 cells. All the constructs were verified by sequenc- evoked by two sequential 200-ms depolarizing pulses 1 s ing. The plasmids were extracted using a Qiagen plasmid midi kit apart to ϩ40 mV from holding potentials of Ϫ100 and Ϫ50 (QIAGEN). The DNA concentration and purity were determined by mV, respectively (Fig. 1A). When the membrane potential measuring the absorbance at 260 and 280 nm. HEK293 cells were was held at Ϫ100 mV, the depolarizing pulses elicited a transfected using the calcium phosphate method. The average trans- global outward current (I plus I ) that activated rapidly fect efficiency is Ͼ80%. Two days after transfection, HEK293 cells A K (5–10 ms) and then decayed with time (Fig. 1A, left). How- with green or red fluorescence were further analyzed. Ј ␥ ␥ ever, at a holding potential of Ϫ50 mV, depolarization evoked Chemicals. Amoxapine, guanosine 5 -[ thio]-triphosphate (GTP S), ϩ cyproheptadine, dibutyryl cAMP (db-cAMP), 5-HT, forskolin, risperi- only a slight inactivating or noninactivating outward K done, TEA, tetrodotoxin, H-89, 4-aminopyridine (4-AP), dihydrocer- current described previously as a delayed rectifier IK (Fig. ␤ amide C6 (C6-ceramide), cytosine1- -D-arabinofuranoside, poly-L-ly- 1A, right). Application of amoxapine significantly accelerated sine, and Dulbecco’s modified Eagle’s medium were purchased from the decay phase of IA and reduced the IK amplitude. In the Sigma-Aldrich. Fetal calf serum, heat-inactivated horse serum, and presence of 5 mM TEA, a specific I channel blocker, I was K A Downloaded from antibiotic-antimycotic solution were obtained from Gibco Life Tech- obtained by depolarizing pulses to ϩ40 mV from a holding nologies (Carlsbad, CA). Jingzhaotoxin-III (JZTX-III; molecular potential of Ϫ100 mV, and amoxapine did not significantly mass, 3919.3 Da) is a peptide toxin containing 36 amino acid resi- inhibit the IA current (Fig. 1B, left). dues with three disulfide bridges cross-linked in the pattern of I–IV, These results indicated that amoxapine selectively modu- II–V, and III–VI (Cys4–Cys19, Cys11–Cys24, and Cys18–Cys31) lates I . Therefore, we investigated I amplitudes in the isolated from the venom of the Chinese spider Chilobrachys jingzhao, K K presence of 5 mM 4-AP, which was used to block IA currents. and was the gift from Dr. Liang (Huanan Normal University, Chang- jpet.aspetjournals.org sha, China). Amoxapine and cyproheptadine solutions were prepared IK were evoked by 200-ms depolarizing pulses over the range extemporaneously. First they were dissolved in methanol and then from Ϫ50 mV to ϩ40 mV at 10-s intervals. Application of diluted in the bath with a final methanol concentration Ͻ0.1%. amoxapine produced a clear reduction in IK amplitudes (Fig.

C6-ceramide and risperidone were first dissolved in dimethyl sulfox- 1B, right). The inhibitory effect of amoxapine on IK was ide (DMSO) and then diluted in the bath with a final DMSO concen- reversible, reaching its maximum effect within 50 to 100 s tration Ͻ0.1%. The 0.1% of methanol and DMSO did not produce the and recovered close to control levels after 2 min. The amox- significantly effect on Kϩ currents. apine-induced inhibitor effect on IK was concentration-de- The drug was ejected by gravity for 40 to 60 s from MSC-200 pendent at concentrations of 10 to 500 ␮M. The inhibition of at ASPET Journals on January 16, 2020 Manual Solution Changer (Bio-Logic SAS, Claix, France). We tested I by amoxapine at 50, 100, and 300 ␮M was 20.7 Ϯ 0.4% the exact dilution rate by perfusion of a high concentration potas- K (n ϭ 5), 31.7 Ϯ 1.6% (n ϭ 5), and 52.2 Ϯ 2.19% (n ϭ 5), sium solution. The exact dilution rate and the time for the drug to respectively (P Ͻ 0.05 by one-way ANOVA). When the con- reach maximal concentration around the cell have been evaluated ␮ from the resting membrane potential shift recorded in response to centration of amoxapine was increased to 500 M, IK was ϩ Ϯ ϭ Ͻ perfusion of 50 mM K solution. By using the Goldman-Hodgkin- inhibited by 72.5 2.2% (n 5, P 0.05 by one-way Katz equation, we figured out that the drug was diluted approxi- ANOVA). The fitting of concentration-response curve was ϭ ϩ mately 3.4 times after it got to the cell surface, and the time for the performed by using the Hill equation y 1/{1 (IC50/[amox- drug to reach maximal concentration around the cell was 3 s. apine])h}, where y is the inhibition rate in the presence of

Fig. 1. Concentration-dependent and reversible inhibition of cortical neu-

ron IK by amoxapine. A, current traces recorded in the absence and presence of 50 ␮M amoxapine (AXP). The current was evoked by two sequential 200-ms depolarizing pulses to ϩ40 mV at 1-s intervals. Holding potentials were set at Ϫ100 mV (first

pulse) for IA plus IK (Itotal) with full Ϫ activation of IA, and at 50 mV (second pulse) for activation of IK.B,IA and IK traces recorded in the absence and presence of 50 ␮M amoxapine

(AXP). IA current elicited by depolarizing pulse from Ϫ100 mV to Ϫ50 mV

with TEA in the bath solution, and IK elicited by depolarizing pulse from Ϫ50 to ϩ40 mV with 4-AP in the bath solution. C, statistical analysis of amoxapine with different concentra- Ͻ ء tions on IK. , P 0.05 compared versus different concentration by one-

way ANOVA. D, Itotal and IK recorded from cortical neurons in the presence of 1% methanol in the bath solution. 440 He et al. ϩ amoxapine, IC50 is the half-maximal inhibition concentration before a 200-ms test pulse of 40 mV (Fig. 3, A and B). After of amoxapine, [amoxapine] is the concentration of amoxa- normalizing each current peak to the maximal current am- pine, and h is the Hill coefficient. With the Hill equation, we plitude obtained from the Ϫ80-mV prepulses as a function of Ϯ got the IC50 and Hill coefficient h, which were 211.36 27.93 the conditioning prepulse potential, we obtained an inactiva- ␮ M and 0.9255, respectively. Statistical analysis of concen- tion curve of IK and calculated the Vh50 (the voltage at which tration-response data is presented in Fig. 1C. Because amox- the current amplitude was half-inactivated; Fig. 3C). This apine was dissolved by 0.1% methanol, we also examined the revealed that the inactivation curve was also significantly

effect of methanol on IA and IK. As shown in Fig. 1D, neither shifted to the left by the application of amoxapine. In five ϭ Ͼ IA nor IK was affected by 1% methanol (n 10, P 0.05 by cells studied, the half-maximal inactivation voltage was paired t test). 11.4 Ϯ 2.6 mV and Ϫ3.9 Ϯ 2.0 mV in the absence and We also studied the effects of amoxapine on the activation presence of 100 ␮M amoxapine, respectively (n ϭ 5, P Ͻ 0.05

and inactivation of IK using different experimental protocols. by paired t test). In the voltage activation protocol, membrane potential was We then used a train protocol in conjunction with the Ϫ held at 80 mV, and IK was evoked by a 200-ms depolarizing protein phosphatase activator C6-ceramide (Gallego et al., pulse from a first pulse potential of Ϫ60 mV to ϩ60 mV in 2005) to investigate whether amoxapine acts directly as an 10-mV steps at 10-s intervals (Fig. 2A). When the depolariz- open-channel blocker. In this protocol amoxapine was ap- ing pulse was more positive than 10 mV, amoxapine signifi- plied to the bath solution for 1 min before channel opening

cantly reduced the current amplitude throughout the activa- (i.e., before IK was elicited by the depolarizing pulse). The Downloaded from tion voltage range (Fig. 2B). We obtained the IK activation effects of amoxapine on IK amplitudes elicited by the first and curve by plotting normalized conductance as a function of the subsequent depolarizing pulses were recorded. As shown in command potential; data were analyzed using the equation Fig. 4A, the amplitude of the peak current evoked by the first ϭ Ϫ ϩ ␮ GK IK/(Vm Vrev), where GK is the membrane K conduc- pulse was reduced in the presence of 50 M amoxapine. tance, Vm is the membrane potential, and Vrev is the reversal Thereafter, the current elicited by subsequent pulses de- ϩ Ϯ potential for K . As shown in Fig. 2C, the activation curve creased to a steady-state level approaching 79 3.3% of jpet.aspetjournals.org shifted to the left in the presence of amoxapine. The current control. This experiment indicated that current blocking can Ϯ Ϫ Ϯ was half-activated at 4.59 0.65 mV and 5.28 0.36 mV occur without previous channel opening. Furthermore, IK ␮ ␮ in the absence and presence of 100 M amoxapine, respec- amplitude was not affected by 20 MC6-ceramide alone, ϭ Ͻ tively (n 10, P 0.05 by paired t test), suggesting that whereas C6-ceramide abolished amoxapine inhibition of IK. ␮ amoxapine treatment significantly modified the voltage de- As shown in Fig. 4B, 20 MC6-ceramide alone did not modify ␮ pendence of IK channel steady-state activation. the IK amplitude. In the presence of 20 M C6-ceramide, the

To study the effects of amoxapine on steady-state inacti- amoxapine-induced inhibitory effect on IK amplitude was at ASPET Journals on January 16, 2020 ϭ Ͻ vation of IK, currents were elicited using 1-s conditioning eliminated (n 4, P 0.05 by paired t test). Statistical prepulses from Ϫ80 mV to different membrane potentials analysis of the above data is shown in Fig. 4, C and D.

Fig. 3. Amoxapine alters steady-state inactivation of IK.A,IK recorded in ␮ Fig. 2. Amoxapine modulates steady-state activation of IK.A,IK record- the absence (top) or presence (bottom) of 100 M amoxapine. A 1-s ings using an activation voltage protocol in the absence (top) or presence conditioning prepulse from Ϫ80 mV to ϩ40 mV in 10-mV increments was (bottom) of 100 ␮M amoxapine. Cells were held at Ϫ80 mV and depolar- applied before the test pulse to ϩ40 mV. The voltage protocol is shown Ϫ ϩ ized in 10-mV steps from 60 mV to 60 mV at 10-s intervals. B, IK below the current recordings. B, steady-state inactivation curves of IK in voltage-dependent activation curves in the absence or presence of amox- the absence or presence of amoxapine. The abscissa shows the condition- apine. C, plot of the normalized conductance as a function of the com- ing prepulse potential. C, peak current amplitudes normalized to the mand potential in the absence or presence of amoxapine; data points were maximal currents were plotted against the prepulse potential. Normal- fitted using the Boltzmann function. The voltage-dependent activation ized current values were fitted to a Boltzmann function. The steady-state curve was significantly shifted toward the left in the presence of amox- inactivation curve was significantly shifted by amoxapine toward the apine. Data are mean Ϯ S.E.M. from 10 cells. hyperpolarization potential. Data are mean Ϯ S.E.M. from five cells. Amoxapine Inhibited IK 441

able cAMP analog) (Sigma-Aldrich) produced a significant Ϯ reduction (21.8 1.0%) of IK and abolished amoxapine inhibition of IK. In the presence of db-cAMP, amoxapine only Ϯ inhibited IK by 4.7 1.9% (Fig. 5B). The inhibition was not statistically significant (Fig. 5D; n ϭ 5, P Ͼ 0.05 by paired t test). We then used a selective PKA antagonist, H-89, to explore

the role of PKA in the inhibition of IK by amoxapine. H-89 was added into the patch pipette solution and itself increased Ϯ IK significantly to 139.7 3.7% of control. Furthermore, 50 ␮ Ϯ M amoxapine only inhibited IK by 1.8 0.87% in the presence of H-89, indicating that H-89 also abolishes the inhibitory effects of amoxapine (Fig. 5, C and D; n ϭ 6). The effects of amoxapine on intracellular cAMP levels were then studied using an anti-cAMP antibody. As shown in Fig. 6A, cAMP immunoreactivity was detected before the addition of amoxapine (Fig. 6A, left), but the levels were significantly in-

creased when neurons were incubated with 50 ␮M amoxapine for Downloaded from 10 min (Fig. 6A, right). This indicated that amoxapine increases intracellular cAMP levels. We then used a direct cAMP assay to confirm amoxapine-induced changes in cAMP levels. As shown in Fig. 6B, cAMP levels were 16.5 Ϯ 2.5 pM (n ϭ 5) before the addition of amoxapine. When the cells were ␮ incubated with 10, 50, and 100 M amoxapine for 3 min, jpet.aspetjournals.org cAMP levels were increased significantly to 21.8 Ϯ 3.1 pM, 26.1 Ϯ 2.7 pM, and 38.9 Ϯ 4.6 pM, respectively (P Ͻ 0.05 by one-way ANOVA). It was recently reported that amoxapine may be comparable in efficacy with the atypical antipsychotic risperidone, which has been shown by positron-emission tomography to

have high serotonin (5-HT)2 occupancy (Kapur et al., 1999). at ASPET Journals on January 16, 2020 Therefore, we investigated whether the 5-HT receptor might

mediate the amoxapine-induced inhibition of IK. As shown in ␮ Ϯ ϭ Fig. 7A, 50 M 5-HT inhibited IK by 21.5 1.5% (n 6). In the presence of 5-HT, however, amoxapine (50 ␮M) only Ϯ ϭ inhibited IK by 2 2.7% (n 6). We further used the 5-HT2 receptor antagonist cyproheptadine to conform the hypothe- sis. The result indicated that cyproheptadine (20 ␮M) alone Ϯ ϭ significantly reduced IK amplitude by 24.4 1.58% (n 6, P Ͻ 0.05 by paired t test). With cyproheptadine in the bath solution, the 100 ␮M amoxapine-induced inhibitory effect on Ϯ Ͼ IK was decreased to 3.25 0.8% (Fig. 6B; P 0.05 by paired t test). Likewise, another 5-HT2 receptor antagonist, risperidone, has the same effect. The risperidone (0.02 ␮M) itself induced a significantly inhibition of I by 20.1 Ϯ 1.06% (n ϭ Fig. 4. Effect of the protein phosphatase activator C6-ceramide on amox- K apine inhibition of IK. A, a train protocol was used to address whether 5, P Ͻ 0.05 by paired t test). In the presence of risperidone, amoxapine inhibition was selective for open channels. In this protocol ␮ 100 M amoxapine-induced inhibitory effect on IK was only current was elicited by a depolarizing pulse (control). Amoxapine was Ϯ Ͼ then applied to the bath solution 1 min before channel opening again by 2.28 0.75% (Fig. 6C; P 0.05 by paired t test). Moreover,

depolarizing pulses. The effects of amoxapine on the IK amplitudes elic- high or low concentrations of cyproheptadine or risperidone ited by the first and subsequent pulses were measured. Pulses were had the similar effect (Fig. 7E). Because 5-HT receptors are G evoked by 200-ms steps from a Ϫ50 mV holding potential to ϩ40 mV. ␮ protein-coupled receptors that modulate multiple signal B, protein phosphatase activator C6-ceramide (20 M) abolished amoxapine inhibition of IK. C and D, statistical analysis. Data come from four transduction pathways (Wada et al., 2006), we then used P Ͻ 0.05 compared versus controls. GTP␥S to explore whether G protein-coupled receptors were ,ء ;to seven independent cells

involved in amoxapine inhibition of IK. Internal infusion of ␥ ␮ To address whether the cAMP/PKA pathway contributed GTP S (10 M) resulted in the gradual augmentation of IK to amoxapine-mediated inhibition of IK, we studied the ef- amplitudes after membrane rupture (Fig. 7D). The ampli- fects of amoxapine in the presence of forskolin, which was an tude was increased significantly to 1501.2 Ϯ 284.5 pA in adenylate cyclase agonist (Fig. 5A). Perfusion of cortical neu- controls from 1231.7 Ϯ 225.9 pA, an increase of 19.6 Ϯ 1.2% ␮ ϭ Ͻ rons with 20 M forskolin provoked a gradual decrease in IK (n 8, P 0.05 by paired t test). Moreover, after augmen- ␥ amplitude similar to that produced by amoxapine. In the tation of IK amplitudes by internal application of GTP S, 50 Ϯ ␮ Ϯ ϭ Ͼ presence of forskolin, amoxapine inhibited IK by only 5.4 M amoxapine only reduced IK by 3.9 2.1% (n 8, P 0.05 1.2%. Perfusion with 20 ␮M db-cAMP (a membrane-perme- by paired t test), significantly different from the reduction 442 He et al. Downloaded from jpet.aspetjournals.org

Fig. 5. cAMP/PKA pathways mediate amoxapine inhibition of IK in mouse cortical neurons. A, time course of changes of IK amplitudes induced by 50 ␮ ␮ M amoxapine in the presence of extracellular forskolin. Forskolin (20 M) provoked a gradual decrease in IK amplitude and abolished the effect of ␮ 50 M amoxapine on IK amplitude. The insets in the graphs show superimposed IK traces from the initial control levels (after establishment of the whole-cell configuration) and after external infusion of forskolin. The time points (1, 2, 3) noted on the curves correspond to the superimposed IK traces ␮ illustrated by the insets. B, time course of the changes in IK amplitudes induced by amoxapine in the presence of db-cAMP (20 M). db-cAMP decreased ␮ IK amplitudes and eliminated the inhibitory effect of 50 M amoxapine. C, time course of amoxapine inhibition in the presence of the PKA inhibitor, H-89. Intracellular H-89 (10 ␮M) provoked a gradual increase in I amplitudes and abolished the effect of 50 ␮M amoxapine on I . D, statistical K K at ASPET Journals on January 16, 2020 Ϯ analysis of the effects of forskolin, db-cAMP, and H-89 on IK amplitudes and amoxapine-induced IK inhibition. Data are mean S.E.M. from five to .P Ͻ 0.05 compared versus control ,ء ;nine cells

It was previously reported that IK channels are predomi- nantly composed of Kv2 subunits of the Shaker family (Mi- sonou et al., 2008). In particular, subunit Kv2.1, which is expressed at high levels in most mammalian central neurons,

is a major contributor to IK channels and plays a crucial role in regulating neuronal excitability (Murakoshi and Trimmer, 1999; Du et al., 2000). To address the potential role of Kv2.1

in amoxapine inhibition of IK channel activity, we used JZTX-III, a toxin known to specifically eliminate the Kv2.1 subunit (Liao et al., 2007). To confirm the selectivity of JZTX- III, experiments were first carried out in HEK293 cells transfected with an expression vector encoding the Kv2.1 ␣-subunit. Kv2.1 currents were evoked by depolarizing pulse to ϩ60 mV from a holding potential of Ϫ50 mV. As shown in Fig. 8A, the toxin significantly decreased the amplitude of the Kv2.1 currents, and application of 100 nM JZTX-III induced a 87.10 Ϯ 1.72% decrease in Kv2.1 currents (n ϭ 11, P Ͻ 0.05 by paired t test). Inhibition was also observed Fig. 6. Amoxapine increases intracellular cAMP levels in mouse cortical neuron. A, cAMP immunoreactivity was detected by anti-cAMP antibody in cortical neurons, where addition of 100 nM JZTX-III to ␮ Ϯ in the absence or presence of 50 M amoxapine. The scale bar represents the bath solution reduced IK amplitudes by 28.5 4.4% 20 ␮m. B, intracellular levels of cAMP measured by direct cAMP analysis (n ϭ 7; Fig. 8B). We then investigated the effects of the P Ͻ 0.05 ,ء ;in the presence of 10, 50, and 100 ␮M amoxapine for 3 min compared versus multiple concentration of amoxapine by one-way toxin on amoxapine inhibition. In the presence of 100 nM ANOVA. Data come from five independent experiments. JZTX-III, 50 ␮M amoxapine only inhibited the remaining Ϯ Ͼ IK by 8.2 3.9% (Fig. 8C; P 0.05 by paired t test). This induced by 50 ␮M amoxapine alone (Fig. 7E). The abolition of was significantly different from the inhibition obtained ␮ Ϯ amoxapine inhibition of IK by either 5-HT receptor antagonist with 50 M amoxapine alone (20.7 0.4%; Fig. 8D), indi- or GTP␥S is consistent with the interpretation that amoxapine cating that elimination of Kv2.1 substantially prevents

exerts its effects through the activation of 5-HT receptor. amoxapine inhibition of IK. Amoxapine Inhibited IK 443 Downloaded from

Fig. 8. The Kv2.1 ␣-subunit is implicated in amoxapine-induced inhibi-

tion of IK in mouse cortical neurons. A, time course of changes in the amplitude of Kv2.1 currents induced by application of 100 nM JZTX-III; HEK293 cells were transfected with an expression vector encoding the ␣ Kv2.1 -subunit. B, time course of changes in IK amplitudes in cortical neurons treated with 100 nM JZTX-III. The toxin partially and reversibly jpet.aspetjournals.org inhibited IK amplitude. C, JZTX-III in the bath solution largely eliminated amoxapine inhibition of IK. D, statistical analysis of the effects of JZTX-III on IK amplitudes and on amoxapine-induced IK inhibition. Data P Ͻ 0.05 compared versus ,ء ;are mean Ϯ S.E.M. from five to nine cells controls.

In contrast, the results reported here indicate that exter- ␥ Fig. 7. Effects of 5-HT, 5-HT2 receptor antagonists, and GTP S on amox- nal amoxapine blocks IK with slow onset and reversible re- apine inhibition of IK. A, time course of changes in IK amplitudes induced

by amoxapine in the presence of 50 ␮M 5-HT. 5-HT provoked a gradual covery. Moreover, amoxapine inhibited IK in the absence of at ASPET Journals on January 16, 2020 ␮ decrease in IK amplitudes and abolished the inhibitory effects of 50 M depolarizing pulses, implying that amoxapine can block amoxapine. B and C, time course of changes in IK amplitudes induced by channel activity irrespective of whether the channel is al- amoxapine in the presence of 20 ␮M cyproheptadine (CPD) or 0.02 ␮M risperidone (RPD). Cyproheptadine and risperidone both provoked a ready open. This would appear to exclude direct open-chan-

gradual decrease in IK amplitudes and abolished the inhibitory effects of nel blockade and instead suggest that amoxapine is most 100 ␮M amoxapine on I . D, time course of changes in I amplitudes K K likely to inhibit IK through intracellular biochemical pro- induced by the application of 10 ␮M GTP␥S. Internal infusion of GTP␥S cesses, which in turn reduce IK channel activity. resulted in gradual augment of IK amplitude after membrane rupture ␮ Protein phosphorylation and dephosphorylation are cen- and abolished IK inhibition by 50 M amoxapine. E, statistical analysis of the effects of 5-HT, higher and lower concentration of 5-HT2 receptor tral to the regulation of ion channel function and membrane antagonist, and GTP␥SonI amplitudes and amoxapine-induced I ϩ K K excitability. Some voltage-dependent K channels, such as I P Ͻ 0.05 A ,ء ;inhibition. Data are mean Ϯ S.E.M. from five to nine cells compared versus untreated cells. #, P Ͻ 0.05 compared versus controls and human ether-a-go-go-related gene channel, are known to without amoxapine. be regulated by protein phosphorylation (Thomas et al., 1999; Discussion Anderson et al., 2000), even though it is not known whether such regulation occurs directly or via intermediary proteins. Previous studies have revealed that, in addition to specific We reported previously that activation of cAMP/PKA path- inhibition of serotonin and/or norepinephrine transporters, way in cerebellar granule cells could induce fast modulation

TCAs also modulate several ion channels. Although the of IK amplitudes (Jiao et al., 2007). In the present study we mechanism by which TCAs can affect ion channel activity have shown that activation of cAMP/PKA pathways produces

has not been elucidated, previous reports have suggested a a similar inhibitory effect on IK amplitudes in cortical neu- direct interaction between the drugs and the ion channels rons. Activation of cAMP/PKA pathways by either forskolin

(Kim et al., 2007). It was reported that the drugs can block or db-cAMP abolished amoxapine inhibition of IK. Further- ion channel conductance by binding within the pore itself or more, inhibition of the PKA pathway by intracellular H-89

by allosteric inhibition produced by binding outside the pore. produced a gradual increase in IK amplitudes after estab- For example, imipramine and diphenhydramine were re- lishment of the whole-cell configuration, and H-89 com- ϩ ported to act as an open-channel blocker on Na currents in pletely prevented amoxapine inhibition of IK. We report rat hippocampal neurons (Yang and Kuo, 2002). Further- that amoxapine treatment led to a significant increase in more, it was reported that imipramine and structurally re- intracellular levels of cAMP. We suggest that amoxapine

lated compounds can directly inhibit transient KA currents in inhibits IK by activating cellular cAMP/PKA pathways. rat hippocampal neurons. It was hypothesized that there is a Nevertheless, we have no evidence to indicate that IK specific imipramine binding site consisting of an aliphatic channels are phosphorylated after amoxapine induction of and aromatic site on the external side of the channel pore the PKA pathway. Further experiments will be required to (Kuo, 1998). test this possibility. 444 He et al.

Amoxapine, a dibenzoxazepine antidepressant, is highly onset of action and reduced cardiotoxicity (Kudo et al., 2007). lipid-soluble and membrane-permeable. It was previously Because amoxapine is a relatively weak reuptake inhibitor, it reported that externally applied TCAs inhibit ion channel was suggested that the molecule may have the properties of activity by binding to an extracellular site or by diffusing an atypical antipsychotic agent (Apiquian et al., 2005). The through the cell membrane to gain access to a binding site on upper range of therapeutic plasma concentrations of amox- the internal surface of the membrane (Kuo, 1998; Kim et al., apine has been reported from 0.2 to 0.4 ␮g/ml (0.637–1.912 ␮ 2007). However, our results argue against direct IK blocking M) in human body (Winek et al., 2001). However, acute activity. It would be interesting to explore whether amoxa- overdose has reportedly caused severe adverse reactions, in- pine activates the cAMP/PKA pathway by targeting an ex- cluding seizure, coma, metabolic acidosis, and tachycardia, tracellular receptor or whether entry of the molecule into the and some fatalities have been reported (Kudo et al., 2007). In

cell is required. the present study, amoxapine inhibited IK with an IC50 value A recent positron-emission tomography study in normal of 211.36 ␮M, a concentration significantly above the clinical volunteers showed that, at a dose of 150 mg/day, amoxapine dosage. Although it is possible that the effective cell surface

saturated 5-HT2 occupancy (98%) but displayed only modest concentration in our perfusion system might be lower than D2 occupancy (63%) (Kapur et al., 1999). Polymerase chain the concentration applied, it still appears that the concentra- reaction analysis of rat cortical neurons has revealed that the tions used in our study are above the levels obtained after 5-HT receptor is well expressed in mouse cortical neuron oral administration of the drug (Kudo et al., 2007). This

(data not shown). Therefore, we examined whether amoxa- would suggest that therapeutic levels of amoxapine do not Downloaded from

pine induction of the inhibition of IK might be mediated by produce a major blockade of neuronal IK. the 5-HT receptor using 5-HT, 5-HT2 receptor antagonists, A recent study compared the antipsychotic effects and side ␥ and GTP S. The results showed that the 5-HT2 receptor and effect profiles of amoxapine versus haloperidol in a 6-week G proteins were associated with amoxapine inhibition of IK double-blind study of patients with schizophrenia. It was because the amoxapine-induced effect was fully eliminated reported that amoxapine may have the similar effect with ␥ by 5-HT, 5-HT2 receptor antagonists, and GTP S. Moreover, haloperidol as an antipsychotic (Chaudhry et al., 2007), as jpet.aspetjournals.org we noted that, except 5-HT, 5-HT2 receptor antagonists have predicted by its affinity for 5-HT2 receptors. We report here the inhibitory effect on IK amplitude. The effect of cyprohep- that amoxapine shifts Kv2.1 activation toward the resting tadine and risperidone on IK was similar and not concentra- membrane potential, making it more easily activated, and tion-dependent (data are not completely shown). The above thereby dampening neuronal excitability (Surmeier and Foe- results were consistent with a previous report in which ris- hring, 2004). This may represent a novel mechanism by

peridone and ketanserin (5-HT2 receptor antagonist) inhib- which amoxapine exerts atypical antipsychotic properties. ϩ

ited K currents through a direct block mechanism in human In summary, this is the first report addressing the effects at ASPET Journals on January 16, 2020

and rat myocardium (Zhang et al., 1994). Even though cypro- of amoxapine on IK in cultured mouse cortical neurons. We ϩ heptadine and risperidone could play a role as a K channel report that amoxapine inhibits IK, an effect different from the blocker, it did not hinder their usefulness as the 5-HT2 re- previously described effects of TCAs on ion channel activity. ceptor antagonists. Because amoxapine is highly lipid-solu- Amoxapine inhibition of IK was mediated by activation of ble and membrane-permeable, whether this substance inhib- cAMP/PKA pathways. Our experiments indicate that the

its IK by interacting at the intracellular faces of 5-HT Kv2.1 subunit is a major target for inhibition. Although the receptors or IK channels needs further exploration. clinical relevance of amoxapine inhibition of IK remains un- Electrophysiological studies have revealed that most mam- clear, inhibition of Kϩ channels by amoxapine may contrib- malian neurons express multiple types of voltage-gated Kϩ ute to the antidepressant and antipsychotic activity of the channels with distinct time- and voltage-dependent proper- molecule. ties (Storm, 1990). The Kv2 (Shab) genes in many species encode components of sustained potassium current channels (Coetzee et al., 1999). Previous studies have indicated that Acknowledgments neuronal I channels are composed of ␣-subunits of the Kv2 We thank Dr. Liang (The Key Laboratory of Protein Chemistry K and Developmental Biology of Ministry of Education, College of Life subfamily. In particular, Kv2.1 subunit, which is highly ex- Sciences, Hunan Normal University, Changsha, People’s Republic of pressed in most mammalian central neurons, is a major China) for the JZTX-III. contributor to IK channels and plays a crucial role in regulating neuronal excitability (Sarmiere et al., 2008). Recent References studies confirmed by polymerase chain reaction (data not Anderson AE, Adams JP, Qian Y, Cook RG, Pfaffinger PJ, and Sweatt JD (2000) shown) have indicated that Kv2.1 is a major component of Kv4.2 phosphorylation by cyclic AMP-dependent protein kinase. J Biol Chem cortical neuron I channels (Guan et al., 2007). In the 275:5337–5346. K Anton RF Jr and Burch EA Jr (1990) Amoxapine versus amitriptyline combined with present study, we report that inhibition of Kv2.1 ␣-subunits perphenazine in the treatment of psychotic depression. Am J Psychiatry 147:1203– using a specific toxin (JZTX-III) reported to selectively block 1208. ␣ Apiquian R, Fresan A, Ulloa RE, de la Fuente-Sandoval C, Herrera-Estrella M, Kv2.1 -subunits (Liao et al., 2007) eliminated the inhibitory Vazquez A, Nicolini H, and Kapur S (2005) Amoxapine as an atypical antipsy- ␣ chotic: a comparative study vs risperidone. Neuropsychopharmacology 30:2236– effect of amoxapine on IK, implying that Kv2.1 -subunits are 2244. central to amoxapine inhibition. We also report that JZTX-III Baldessarini RJ (2001) Drugs and the treatment of psychiatric disorders: depression and anxiety disorders, in The Pharmacological Basis of Therapeutics, 10th ed significantly reduced IK amplitudes in cortical neurons, ar- ␣ (Hardman JG, Limbird LE and Gilman AG eds) pp 447–483, McGraw-Hill, New guing that Kv2.1 -subunits are an integral part of cortical IK York. channels. Chaudhry IB, Husain N, Khan S, Badshah S, Deakin B, and Kapur S (2007) Amoxapine as an antipsychotic: comparative study versus haloperidol. J Clin Amoxapine, a second-generation TCA, was introduced as Psychopharmacol 27:575–581. an alternative to the traditional TCAs because of its shorter Chen YW, Huang KL, Liu SY, Tzeng JI, Chu KS, Lin MT, and Wang JJ (2004) Amoxapine Inhibited IK 445

Intrathecal tri-cyclic antidepressants produce spinal anesthesia. Pain 112:106– Murakoshi H and Trimmer JS (1999) Identification of the Kv2.1 Kϩ channel as a 112. major component of the delayed rectifier Kϩ current in rat hippocampal neurons. Clark MC, Dever TE, Dever JJ, Xu P, Rehder V, Sosa MA, and Baro DJ (2004) J Neurosci 19:1728–1735. Arthropod 5-HT2 receptors: a neurohormonal receptor in decapod crustaceans that Nasu R, Matsuo H, Takanaga H, Ohtani H, and Sawada Y (2000) Quantitative displays agonist independent activity resulting from an evolutionary alteration to prediction of catalepsy induced by amoxapine, cinnarizine and cyclophosphamide the DRY motif. J Neurosci 24:3421–3435. in mice. Biopharm Drug Dispos 21:129–138. Coetzee WA, Amarillo Y, Chiu J, Chow A, Lau D, McCormack T, Moreno H, Nadal Prochiantz A, di Porzio U, Kato A, Berger B, and Glowinski J (1979) In vitro MS, Ozaita A, Pountney D, et al. (1999) Molecular diversity of Kϩ channels. Ann maturation of mesencephalic dopaminergic neurons from mouse embryos is en- N Y Acad Sci 868:233–285. hanced in presence of their striatal target cells. Proc Natl Acad Sci U S A 76:5387– Du J, Haak LL, Phillips-Tansey E, Russell JT, and McBain CJ (2000) Frequency- 5391. ϩ dependent regulation of rat hippocampal somato-dendritic excitability by the K Roeper J and Pongs O (1996) Presynaptic potassium channels. Curr Opin Neurobiol channel subunit Kv2.1. J Physiol 522:19–31. 6:338–341. Gallego M, Setie´n R, Puebla L, Boyano-Ada´nez Mdel C, Arilla E, and Casis O (2005) Sarmiere PD, Weigle CM, and Tamkun MM (2008) The Kv2.1 Kϩ channel targets to alpha1-Adrenoceptors stimulate a Galphas protein and reduce the transient out- ϩ the axon initial segment of hippocampal and cortical neurons in culture and in ward K current via a cAMP/PKA-mediated pathway in the rat heart. Am J situ. BMC Neurosci 9:112. Physiol Cell Physiol 288:C577–C585. Song MK, Liu H, Jiang HL, Yue JM, Hu GY, and Chen HZ (2008) Discovery of Gillman PK (2007) Tricyclic antidepressant pharmacology and therapeutic drug talatisamine as a novel specific blocker for the delayed rectifier Kϩ channels in rat interactions updated. Br J Pharmacol 151:737–748. hippocampal neurons. Neuroscience 155:469–475. Greenblatt EN, Lippa AS, and Osterberg AC (1978) The neuropharmacological Storm JF (1990) Potassium currents in hippocampal pyramidal cells. Prog Brain Res actions of amoxapine. Arch Int Pharmacodyn Ther 233:107–135. 83:161–187. Guan D, Tkatch T, Surmeier DJ, Armstrong WE, and Foehring RC (2007) Kv2 Su S, Ohno Y, Lossin C, Hibino H, Inanobe A, and Kurachi Y (2007) Inhibition of subunits underlie slowly inactivating potassium current in rat neocortical pyra- astroglial inwardly rectifying Kir4.1 channels by a tricyclic antidepressant, nor- midal neurons. J Physiol 581:941–960. triptyline. J Pharmacol Exp Ther 320:573–580. Jiao S, Liu Z, Ren WH, Ding Y, Zhang YQ, Zhang ZH, and Mei YA (2007) cAMP/ Surmeier DJ and Foehring R (2004) A mechanism for homeostatic plasticity. Nat protein kinase A signalling pathway protects against neuronal apoptosis and is Neurosci 7:691–692. associated with modulation of Kv2.1 in cerebellar granule cells. J Neurochem Takahashi T, Kobayashi T, Ozaki M, Takamatsu Y, Ogai Y, Ohta M, Yamamoto H, Downloaded from 100:979–991. and Ikeda K (2006) G protein-activated inwardly rectifying Kϩ channel inhibition Kapur S, Cho R, Jones C, McKay G, and Zipursky RB (1999) Is amoxapine an and rescue of weaver mouse motor functions by antidepressants. Neurosci Res atypical antipsychotic? Positron-emission tomography investigation of its dopa- 54:104–111. mine and serotonin occupancy. Biol Psychiatry 45:1217–1220. 2 2 Terstappen GC, Pula G, Carignani C, Chen MX, and Roncarati R (2001) Pharmaco- Kim SE, Ahn HS, Choi BH, Jang HJ, Kim MJ, Rhie DJ, Yoon SH, Jo YH, Kim MS, logical characterisation of the human small conductance calcium-activated potas- Sung KW, et al. (2007) Open channel block of A-type, kv4.3, and delayed rectifier ϩ sium channel hSK3 reveals sensitivity to tricyclic antidepressants and antipsy- K channels, Kv1.3 and Kv3.1, by sibutramine. J Pharmacol Exp Ther 321:753– chotic phenothiazines. Neuropharmacology 40:772–783. 762. Thomas D, Zhang W, Karle CA, Katho¨fer S, Scho¨lsW,Ku¨ bler W, and Kiehn J (1999)

Kudo K, Inoue H, Ishida T, Tsuji A, and Ikeda N (2007) A fatal case of amoxapine jpet.aspetjournals.org Deletion of protein kinase A phosphorylation sites in the HERG potassium channel poisoning under the influence of chronic use of psychotropic drugs. Leg Med inhibits activation shift by protein kinase A. J Biol Chem 274:27457–27462. (Tokyo) 9:63–67. ϩ Wada K, Hu L, Mores N, Navarro CE, Fuda H, Krsmanovic LZ, and Catt KJ (2006) Kuo CC (1998) Imipramine inhibition of transient K current: an external open channel blocker preventing fast inactivation. Biophys J 75:2845–2857. Serotonin (5-HT) receptor subtypes mediate specific modes of 5-HT-induced sig- Liao Z, Yuan C, Peng K, Xiao Y, and Liang S (2007) Solution structure of Jingzhao- naling and regulation of neurosecretion in gonadotropin-releasing hormone neu- toxin-III, a peptide toxin inhibiting both Nav1.5 and Kv2.1 channels. Toxicon rons. Mol Endocrinol 20:125–135. 50:135–143. Winek CL, Wahba WW, Winek CL Jr, and Balzer TW (2001) Drug and chemical Liu H, Gao ZB, Yao Z, Zheng S, Li Y, Zhu W, Tan X, Luo X, Shen J, Chen K, et al. blood-level data 2001. Forensic Sci Int 122:107–123. ϩ (2007) Discovering potassium channel blockers from synthetic compound database Yang YC and Kuo CC (2002) Inhibition of Na( ) current by imipramine and related by using structure-based virtual screening in conjunction with electrophysiological compounds: different binding kinetics as an inactivation stabilizer and as an open assay. J Med Chem 50:83–93. channel blocker. Mol Pharmacol 62:1228–1237. at ASPET Journals on January 16, 2020 Maertens C, Droogmans G, Verbesselt R, and Nilius B (2002) Block of volume- Zhang ZH, Boutjdir M, and el-Sherif N (1994) Ketanserin inhibits depolarization- regulated anion channels by selective serotonin reuptake inhibitors. Naunyn activated outward potassium current in rat ventricular myocytes. Circ Res 75:711– Schmiedebergs Arch Pharmacol 366:158–165. 721. Melishchuk A, Loboda A, and Armstrong CM (1998) Loss of shaker K channel ϩ conductance in 0 K solutions: role of the voltage sensor. Biophys J 75:1828–1835. Address correspondence to: Mei Yan-Ai, School of Life Sciences and Insti- Misonou H, Thompson SM, and Cai X (2008) Dynamic regulation of the Kv2.1 tutes of Brain Science, Fudan University, Shanghai 200433, China. E-mail: voltage-gated potassium channel during brain ischemia through neuroglial inter- [email protected] action. J Neurosci 28:8529–8538.