Pharmacological Characterization of 5-HT1A Autoreceptor-Coupled GIRK Channels in Rat Dorsal Raphe 5-HT Neurons

G protein-activated inwardly rectifying potassium (GIRK) channels in 5-HT neurons are assumed to be principal effectors of 5-hydroxytryptamine 1A (5-HT1A) autoreceptors, but their pharmacology, subunit composition and the role in regulation of 5-HT neuron activity have not been fully elucidated. We sought for a pharmacological tool for assessing the functional role of GIRK channels in 5-HT neurons by characterizing the effects of drugs known to block GIRK channels in the submicromolar range of concentrations. Whole-cell voltage-clamp recording in brainstem slices were used to determine concentration-response relationships for the selected GIRK channel blockers on 5-HT1A autoreceptor-activated inwardly rectifying K+ conductance in rat dorsal raphe 5-HT neurons. 5-HT1A autoreceptor-activated GIRK conductance was completely blocked by the nonselective inwardly rectifying potassium channels blocker Ba2+ (EC50 = 9.4 μM, full block with 100 μM) and by SCH23390 (EC50 = 1.95 μM, full block with 30 μM). GIRK-specific blocker tertiapin-Q blocked 5-HT1A autoreceptor-activated GIRK conductance with high potency (EC50 = 33.6 nM), but incompletely, i.e. ~16% of total conductance resulted to be tertiapin-Q-resistant. U73343 and SCH28080, reported to block GIRK channels with submicromolar EC50s, were essentially ineffective in 5-HT neurons. Our data show that inwardly rectifying K+ channels coupled to 5-HT1A autoreceptors display pharmacological properties generally expected for neuronal GIRK channels, but different from GIRK1-GIRK2 heteromers, the predominant form of brain GIRK channels. Distinct pharmacological properties of GIRK channels in 5-HT neurons should be explored for the development of new therapeutic agents for mood disorders.


Introduction
It is well documented that the activity of raphe 5-HT neurons is under regulatory control by 5-HT 1A autoreceptors and K + channels. Early electrophysiological studies in vivo [1] and in midbrain slices [2] suggested that stimulation of 5-HT receptors hyperpolarize dorsal raphe 5-HT neurons by an increase in K + conductance. Using intracellular recordings, Williams et al. [3] showed that stimulation of 5-HT or GABA B receptors on dorsal raphe 5-HT neurons activates an inwardly rectifying K + conductance via a pertussis toxin sensitive G-protein. This conductance was abolished by low (100 μM) concentration of Ba 2+ . Studies using whole-cell [4] and single-channel [5] patch-clamp recording on acutely isolated dorsal raphe 5-HT neurons confirmed the sensitivity to pertussis toxin and the block by Ba 2+ . These studies further revealed that the receptor implicated belongs to the 5-HT 1A subtype and that the activation of K + channels occurs via a direct membrane-limited pathway without the involvement of soluble intracellular messengers. Similar findings were reported by Bayliss et al. [6] for 5-HT neurons in the medulla oblongata (raphe pallidus and raphe obscurus). Together, these studies clearly indicate that the inwardly rectifying K + channel activated by 5-HT 1A autoreceptors belongs to the GIRK (K ir 3) channel family. Consistently, it has been shown that GIRK1 (K ir 3.1), GIRK2 (K ir 3.2) and GIRK3 (K ir 3.3) mRNA and proteins are expressed in dorsal raphe 5-HT neurons [7][8][9][10].
In raphe 5-HT neurons, the activation of 5-HT 1A autoreceptors, besides the opening of GIRK channels, also produces an inhibition of voltage-gated Ca 2+ channels [11][12][13] and, probably, modulate other ion channel types [3,4]. The precise role of GIRK channels in the control of 5-HT neuron activity has remained unclear, in part due to the limited knowledge of their pharmacological properties. Ba 2+ , a nonselective blocker of inwardly rectifying K + channels, is the only drug so far reported to block GIRK channels in raphe 5-HT neurons, albeit with unknown EC 50 . Given the functionally important role of 5-HT 1A autoreceptor-activated inwardly rectifying K + channels in the control of 5-HT neuron activity, the availability of organic blockers would greately help further functional studies and may lead to the development of valuable new therapeutic agents. Here, we used whole-cell recording in brainstem slices to characterize the effects of available GIRK channel blockers on inwardly rectifying K + current activated by 5-HT 1A autoreceptors in 5-HT neurons. We found that inwardly rectifying K + channels coupled to 5-HT 1A autoreceptors have pharmacological properties generally expected from GIRK channels, although the profile of blocker sensitivity differs from that of GIRK1-GIRK2 heteromers, the predominant form of brain GIRK channels.

Animal Welfare and Ethical Statement
All animal care and experimental procedures complied with the European Communities Council Directive (2010/63/UE) and were approved by the Internal Committee for Animal Care and Experimental Use (IACUC) of the University of Florence and communicated to the Italian Ministry of Health, as required (D.L. 116/92). A total of 22 male Wistar rats (Harlan Italy, Milano, Italy) were used in the present study. Animals were housed under standard laboratory conditions (12 h light/dark cycle, ambient temperature 22±1°C, humidity 40-50%, standard chow and water ad libitum). Animals (25-35 days of age at the experimental day) were sacrified under isoflurane anesthesia, and all efforts were made to minimize suffering.

Whole-cell patch-clamp recordings
Experimental procedures have been previously described in detail [14]. In brief, the brain was rapidly removed and dissected in ice-cold gassed (95% O 2 and 5% CO 2 ) artificial cerebrospinal fluid (ACSF) composed of: 124 mM NaCl, 2.75 mM KCl, 1.25 mM NaH 2 PO 4 , 1.3 mM MgCl 2 , 2 mM CaCl 2 , 26 mM NaHCO 3 , 11 mM D-glucose. The brainstem was sliced coronally into 200 μm thick slices with a vibratome (DSK, T1000, Dosaka, Japan). After recovery for at least 90 min at room temperature, the slices were individually transferred to the recording chamber and superfused continuously with warmed (29-31°C; Warner Instruments in-line heater TC324-C), modified (see below) ACSF at a rate of 2 mL min -1 . Slices were allowed to equilibrate for at least 15 min before the beginning of the recording. Drugs were bath-applied through a peristaltic pump-driven perfusion system and a complete exchange of the recording chamber volume occurred in approximately 1 min. Neurons within the dorsal raphe nucleus were visualized by infrared differential interference contrast (IR-DIC) video microscopy with a Newicon camera (C2400-07; Hamamatsu, Hamamatsu City, Japan) mounted on an upright microscope (Axioskop; Zeiss, Göttingen, Germany). Recordings were made using an EPC-10 amplifier (HEKA Elektronic, Lamberecht, Germany). Patch pipettes were prepared from thick-walled borosilicate glass on a P-97 Brown-Flaming electrode puller (Sutter Instruments, Novato, CA, USA). The pipette solution consisted of: 120 mM K gluconate, 15 mM KCl, 2 mM MgCl 2 , 10 mM HEPES, 0.1 mM EGTA, 10 mM Na 2 phosphocreatine, 4 mM MgATP, 0.3 mM Na 3 GTP (pH 7.35 with 9 mM KOH). Pipettes had filled-tip resistance of 2.5-5.2 MO.
To block synaptic transmission, all recordings were done using modified ACSF supplemented with a cocktail of glutamate and GABA/glycine receptor blockers consisting of: 10 μM NBQX (2,3-dioxo-6-nitro-1,2,3,4-tetrahydrobenzo[f]quinoxaline-7-sulfonamide disodium salt), 20 μM D-AP5 (D-(-)-2-amino-5-phosphonopentanoic acid), 10 mM strychnine hydrochloride, 10 μM SR-95531 (6-imino-3-(4-methoxyphenyl)-1(6H)-pyridazinebutanoic acid hydrobromide) and 2 μM CGP-55845 (3-N[1-(S)-(3,4-dichlorophenyl)ethyl]amino-2-(S)hydroxypropyl-P-benzyl-phosphinic acid hydrochloride). The modified ACSF also contained 5.5 mM K + (the additional 2.75 mM by Na + substitution) to increase the driving force for inward K + current and to shift K + reversal potential to a more positive value, thus permitting reliable detection of inwardly rectifying K + currents at membrane potentials more negative than -85 mV, at which there is a neglibile contribution of outward rectifying K+ currents and voltage-gated ion channels. We used hyperpolarizing voltage ramps from the holding potential of -65 mV (to -125 mV, every 10 s; 100 mV s -1 ; 3 kHz cutoff frequency low-pass filter; 10 kHz sampling frequency) and measured the conductance from the slope of inward K + current in range from -110 to -90 mV (G -110/-90 mV ). To monitor access resistance throughout the recording, hyperpolarizing pulses (10 mV; 100 ms duration; 16 kHz low-pass filter; 25 kHz sampling frequency; cell capacitance cancellation circuit switched off) were interlaced with ramps. The access resistance was in 8 to 25 MO range and recordings were interrupted when it changed for more than 25%. In addition, approximately 15% of the experiments were aborted since they were compromised by spontaneous activation of a marked outward rectifying K + current, similar to that previously described by Bayliss et al. [6], which was not further examined. The membrane potential was not corrected for Donnan liquid junction potential. The effects of channel blockers were examined using a cumulative concentration-response protocol in the contunuous presence of the 5-HT 1A receptor agonist 5-carboxamidotryptamine maleate (5-CT). One experiment was performed per slice. Each concentration of blocker was bath-applied until the apparent steady-state of the online monitored G -110/-90 values was reached, i.e. typically eight to ten minutes, with the exception of SCH23390 (up to 20 min). In most of recordings three to five increasing concentrations of a blocker were succesfully applied and the concentrationresponse relationship in individual neurons could be fitted by the four parameters logistic equation y = b+ (a-b)/(1 + (EC 50 /[Blocker]) nH ), where a corresponds to G -110/-90 in the absence of blocker effect, b corresponds to G -110/-90 with the maximal blocker effect, EC 50 is the half-maximally effective concentration and n H is the Hill coefficient.

Data analysis and statistical procedures
Data were analyzed using Patchmaster 2 (HEKA Elektronic) and then with Prism 6 software (GraphPad Software, San Diego, CA, USA). All statistics are given as mean ± SD, except EC 50 values which are given as mean and 95% confidence intervals (95% C.I.).

Results
To activate 5-HT 1A autoreceptors in dorsal raphe 5-HT neurons, we used 5-CT, an agonist which in our experimental conditions selectively activates 5-HT 1A autoreceptors [15,14] and is structurally similar to the endogenous agonist, 5-HT. As shown in Fig 1A-1C, bath application of 5-CT produced a concentration-dependent increase in an inwardly rectifying K + conductance which was completely reversed upon agonist washout. To minimize 5-HT 1A receptor and GIRK channel desensitization we used 5-CT at 30 nM, a concentration which produces near-maximal effect. As shown in Fig 1D-1F, at this concentration only a limited run-down of the response was observed over a prolonged period of continuous agonist application (22.3 ± 4.1% in 1 h, mean ± SD, n = 6), permitting a fairly accurate determination of concentration-response relationships for the investigated blockers in individual neurons.
We first wanted to characterize the antagonism of 5-HT 1A receptor-activated inwardly rectifying K + conductance by Ba 2+ , an information surprisingly missing even if Ba 2+ is the only blocker so far reported to be effective in 5-HT neurons. As shown in Fig 2, Ba 2+ concentrationdependently blocked 5-CT-induced inwardly rectifying K + conductance, producing full block at 100 μM concentration. The full block by Ba 2+ was confirmed in additional recordings in which 5-CT (30 nM, 10 min) failed to induce inwardly rectifying K + conductance when coapplied in the presence of 100-150 μM Ba 2+ (10 min; n = 5, not shown). In four individual neurons in which Ba 2+ was applied in concentrations from 3 to 100 μM, fit of data with a logistic equation revealed EC 50 values ranging from 5.1 to 17.3 nM and Hill slope values ranging from -1.06 to -1.39 (Fig 2E and 2F). Concentration-response relationship of Ba 2+ on pooled data ( Fig 2G) revealed an EC 50 of 9.4 μM (95% C.I. 7.2 to 12.2 μM) and a Hill Slope of -1.21 (95% C. I. -0.83 to -1.59).
We next characterized the block of 5-HT 1A receptor-activated inwardly rectifying K + conductance by the specific GIRK blocker tertiapin-Q, a non-air-oxidizable derivative of honey bee toxin tertiapin [16]. Tertiapin-Q concentration-dependently blocked 5-CT-induced inwardly rectifying K + conductance in the submicromolar range, but the block was incomplete (e.g. Fig 3A and 3B). To better examine if tertiapin-Q can produce a complete block, in an additional set of experiments, tertiapin-Q was applied only at concentrations of 1 μM (n = 2) and/ or 3 μM (n = 5; e.g. Fig 3C and 3D). In all cases, tertiapin-Q failed to produce full block. In eight individual neurons in which tertiapin-Q was applied at concentrations from 10 or 30 nM to 0.3 or 1 μM, fit of data with a logistic equation showed that tertiapin-Q blocked 5-CTinduced inwardly rectifying K + conductance with EC 50 values ranging from 20.9 to 74.6 nM and Hill slope values ranging from -0.76 to -1.60 (Fig 3E and 3F). Concentration-response of tertiapin-Q on pooled data from all fifteen experiments (Fig 3G) revealed an EC 50 of 33.6 nM (95% C.I. 27.8 to 40.7 nM), a Hill Slope of 1.22 (95% C.I. -0.90 to -1.54) and a maximal block of 83.9% (95% C.I. 79.6 to 88.3%).
We proceeded to examine the block of 5-HT 1A receptor-activated inwardly rectifying K + conductance by organic compounds reported to block GIRK with EC 50 values in the submicromolar range. We first tested SCH23390, a classic dopamine D 1 receptor antagonist and potent 5-HT 2C receptor agonist [17] which is also direct GIRK channel blocker [18]. As shown in Fig  4A-4C, SCH23390 concentration-dependently blocked 5-CT-induced inwardly rectifying K + conductance, producing a complete block at 30 μM concentration. The ability of SCH23390 to fully block 5-CT-induced inwardly rectifying K + conductance was confirmed in additional recordings in which it was applied only at 30 μM (n = 3; Fig 4D-4F) and in recordings in which 5-CT (30 nM, 10 min) failed to induce inwardly rectifying K + conductance when coapplied in the presence of 30 μM SCH23390 (20 min; n = 3, not shown). In eight individual   rectifying K + conductance with EC 50 values ranging from 1.5 to 3.1 μM and Hill slope values ranging from -0.84 to 1.47 (Fig 4G and 4H). Concentration-response of SCH-23390 on pooled data from all experiments (Fig 4I)   We also examined the effect of two small molecules reported to block GIRK channels with EC 50 values in the submicromolar range: U73343, which blocks GIRK channels in acutely isolated rat neocortical pyramidal cells with an EC 50 of 400 nM [19], and SCH28080, reported to block mouse GIRK channels in AtT20 and HL-1 cells with EC 50 values of 200 and 300 nM, respectively [20]. As shown in Fig 5, both substances were poorly effective blockers of the 5-CT-induced inwardly rectifying K + conductance in the dorsal raphe 5-HT neurons. At 10 μM, the highest concentration tested, U73343 and SCH28080 reduced 5-CT-induced inwardly rectifying K + conductance by 31.6% (range 22.4-48.0%, n = 4) and 20.5% (range 12.3-32.3%, n = 4), respectively.

Discussion
Pharmacological properties of GIRK channels in 5-HT neurons have remained uncharacterized in spite of the evidence that GIRK channels are principal effectors of 5-HT 1A autoreceptors. We attempted to find a blocker which may serve as a tool to study the role of GIRK channels in the regulation of 5-HT neuron activity. Whole-cell recording in brainstem slices were used to directly measure GIRK conductance. We report the first quantitative assessment of the effects of a series of specific organic blockers and Ba 2+ on native inwardly rectifying K + channels activated by 5-HT 1A autoreceptor stimulation in raphe 5-HT neurons.

Methodological considerations
Pharmacology of neuronal GIRK channels is largely unexplored mostly because of the paucity of GIRK blockers available. In many neuronal types, e.g. pyramidal and Purkinje neurons, studies of GIRK channels by using whole-cell recordings can be additionally complicated by the combination of dendritic localization of GIRK channels and the low cell membrane input resistance. Raphe 5-HT neurons are suitable for studying GIRK channels since (i) their dendrites are relatively short and without extensive arborization and (ii) they are electrically compact, displaying high input resistance (1 GO) in whole-cell patch-clamp recordings in brainstem slice preparation. Importantly, we found that when GIRK channels are continuously activated by a concentration of 5-HT 1A receptor agonist which produces near-complete receptor stimulation, recorded whole-cell GIRK conductances display only slow and limited rundown, presumably caused by desensitization. This allows recordings of sufficient duration for a fairly accurate determination of the blocker's concentration-response relationship. Due to the run-down, the calculated EC 50 and Hill slope values for the blockers are expectedly slightly under-and over-estimated, respectively. It is worth mentioning that when GIRK conductances in 5-HT neurons are activated by supramaximal concentrations of 5-HT 1A receptor agonists (e.g. bath application of 300 nM 5-CT or R(+)-8-OH-DPAT) or with the inclusion of nonhydrolyzable GTP analogue GTP-γ-S in the pipette solution (100 μM, instead of GTP) marked desensitization occurs ( [6], our unpublished observation) Pharmacological properties of GIRK channels in 5-HT neurons only partially match those of GIRK1-GIRK2 heteromers In the brain, GIRK1, GIRK2 and GIRK3 channel subunits are widely expressed, whereas the GIRK4 subunit is found at low levels. Although the precise subunit composition of GIRK channels in various neuronal subtypes and subcellular compartments is not entirely known (for recent reviews see [21,22]), there is evidence that GIRK1-GIRK2 heteromer is the predominant form of GIRK channels in neurons [23] and that GIRK2-containing channels mediate postsynaptic inhibition by G protein-coupled neurotransmitter receptors [24]. Neuronal GIRK channels may also be composed of different combinations of subunits, e.g. GIRK2A-GIRK2C [25], GIRK2-GIRK3 [26,27] and GIRK1-GIRK3 [28] and possibly of three different subunits [29]. All of these combinations are possible in 5-HT neurons, since they co-express GIRK1, GIRK2 and GIRK3 subunits [10].
Although the pharmacology of GIRK3-containing channels is still uncharacterized and subunit-selective GIRK blockers are presently lacking, some important conclusions regarding the molecular identity of GIRK channels in 5-HT neurons can be deduced from our findings. By comparing the effects of here characterized blockers with the published data on their activity at GIRK channels, it appears that 5-HT 1A receptor-activated inwardly rectifying K + conductance only partially matches that of GIRK1-GIRK2 heteromers. Relatively good correspondence was found for tertiapin-Q which blocked 5-HT 1A receptor-activated GIRK conductance with an EC 50 of~34 nM, a value similar to those reported for blocking GIRK current in AtT-20 cells (102 nM; [20]) which express native mouse GIRK1 and GIRK2 subunits, and in HEK239 cells heterologously expressing mouse GIRK1 and GIRK2A or GIRK2C subunits (~76% and~59% block by 100 nM, respectively; [30]). Similarly, the EC 50 of SCH23390 (~2 μM) was in the range of those reported to block native mouse GIRK channels in AtT-20 cells (EC 50 of 236 nM) and heterologously expressed human GIRK1-GIRK2 channels (EC 50 of 7.8 μM) [18]. In contrast, our finding that U73343 and SCH28080 are poorly effective blockers of GIRK channels in 5-HT neurons significantly differs from previous studies that showed that U73343 blocks native GIRK channels in rat neocortical pyramidal neurons and heterologously expressed GIRK1-GIRK2 channels with EC 50 of~400 nM (~85% block by 10 μM) [19] and that SCH28080 blocks~80% of native mouse GIRK channels in AtT-20 cells with an EC 50 of~200 nM [20]. Finally, Ba 2+ blocked 5-HT 1A receptor-activated GIRK channels with the EC 50 of 9.4 μM, a value close to that of 12 μM reported in rat CA3 pyramidal neurons, which express all GIRK subunits [31], but significantly lower than that reported for heterologously expressed mouse GIRK1-GIRK2 channels (EC 50 of~100 μM; [32]). Therefore, it seems unlikely that the 5-HT 1A receptor-coupled GIRK channels in dorsal raphe 5-HT neurons is a GIRK1-GIRK2 heteromer. In addition, GIRK2 homomer may also be excluded as a 5-HT 1A autoreceptor effector since SCH23390 blocks heterologously expressed human GIRK2 homomer with the EC 50 of 83 μM [18], a value~40 times higher than in 5-HT neurons.
The conclusion that GIRK1-GIRK2 heteromer and GIRK2 homomer are unlikely to be 5-HT 1A autoreceptor-coupled channels in 5-HT neurons is further supported by expression studies that found a low level of GIRK2 mRNA and protein in rat dorsal raphe [9,10] or failed to detect GIRK2 mRNA [7] and protein [8] at all. In addition, a recent functional study in mice in vivo [33], showed that genetic deletion of GIRK2 does not abolish 5-HT 1A receptor-mediated suppression of 5-HT neuron firing, but only results in a limited rightward shift of doseresponse curves for 8-OH-DPAT, a 5-HT 1A receptor agonist, and citalopram, a selective serotonin reuptake inhibitor (SSRI).
Although further studies are clearly needed to define the molecular identity of GIRK channels in 5-HT neurons, their distinct pharmacological properties might provide the foundation for the development of 5-HT neuron-selective/preferential GIRK blockers. Such drugs would have a potentially important therapeutic role since they should increase 5-HT neuron activity without causing serious side effects by the nonselective block of GIRK channels, in particular of neuronal GIRK1-GIRK2 and cardiac GIRK1-GIRK4 heteromers. In fact, the delay in the therapeutic effect of SSRIs, the most commonly prescribed antidepressants, has in part been ascribed to the drug-induced increase in raphe 5-HT level, which decreases 5-HT neuron activity by activating 5-HT 1A autoreceptors [34]. Thus, 5-HT neuron-specific GIRK blockers in association with SSRIs, may afford a valuable therapeutic strategy to hasten the antidepressant effects.

Tertiapin-Q-resistant conductance
The finding that~16% of total 5-HT 1A autoreceptor-activated inwardly rectifying K + conductance is tertiapin-Q-resistant was unexpected. The lack of a complete block is unlikely to be caused to impurities in tertiapin-Q since identical results were observed with several different batches of tertiapin-Q obtained from two different commercial sources. We cannot exclude that part of the 5-HT 1A autoreceptor-activated tertiapin-Q-resistant conductance is mediated by ion channels other than GIRK, but this seems unlikely since 100 μM Ba 2+ and 30 μM SCH23390 produced full block. It is conceivable that a subpopulation of GIRK channels in 5-HT neurons has a distinctive GIRK subunit composition that results in tertiapin-Q-resistance. In fact, differential sensitivity of rat GIRK subunits to tertiapin-Q has been demonstrated by Ramu et al. [35], who showed that high tertiapin-Q affinity of cardiac GIRK1-GIRK4 heteromers (K d 13 nM, [16]) results from interaction of tertiapin-Q with GIRK4 subunit (K d 2 nM), while GIRK1 is relatively insensitive (K d 20 μM). Therefore, it is reasonable to assume that the tertiapin-Q-resistant conductance in 5-HT neurons is mediated by a GIRK1-containing channels. Other subunits besides GIRK1 may contribute to differential sensitivity to tertiapin-Q. As example, Walsh [20] reported that tertiapin-Q blocks native mouse GIRK1-GIRK4 and GIRK1-GIRK2 channels with a two orders of magnitude different potency (EC 50 of 1.4 and 102 nM, respectively). Finally, tertiapin-Q-resistant channels may result from promiscuous GIRK1-IRK heteromerization, which has been demonstrated feasible by Ishihara et al. [36].

Conclusions
In conclusion, tertiapin-Q, SCH23390 and Ba 2+ are useful tools for studying GIRK function in 5-HT neurons, albeit with specific limitations since tertiapin-Q does not produce full block whereas SCH23390 and Ba 2+ do not warrant selectivity of action on GIRK channels. Importantly, our data show that 5-HT 1A autoreceptor-coupled GIRK channels in 5-HT neurons have specific pharmacological properties, which may provide foundation for the development of valuable new therapeutic agents.