Role of G protein coupled inwardly rectifier K+ channels (GIRK) on the neurobiology depression

Abstract

[EN] Major depression is a recurrent or chronic mood disorder with impaired psychosocial function, and an increased tendency of relapse for patients who fail to reach remission. In industrialized societies, it affects 15% of the population being the age range of affected people between 18 ¿ 44 years (Kessler et al., 2007). The World Health Organization has predicted that major depression will become a key cause of illness-induced disability by the year 2010, followed only by ischemic heart disease (Murray and Lopez, 1997).Although the disorder is thought to be the outcome of gene-environmental interactions, the causative genes and environmental factors underlying depression remain to be identified (Lee et al., 2010). On the other hand, there are many hypotheses to explain the molecular mechanism of depression. Many of them are related, but a definite ethiopathogenic mechanism is still not identified. The action of reserpine in causing depression by depleting catecholamines and the reversal of these effects by triciclic antidepressants formed the basis of the catecholamine hypothesis of depression (Schildkraut, 1965). The monoamine hypothesis proposes that the underlying biological or neuroanatomical basis of depression is a deficiency or imbalance of central monoamine neurotransmission (Schildkraut, 1965; Van Praag, 1982). This hypothesis was supported by the effects of the antidepressant drugs on these neurotransmitter systems.Summary340However this hypothesis does not explain why the clinical effects of antidepressants appear at least four weeks after the beginning of the treatment, while the antidepressants increase the monoamine levels immediately.It is now believed that an adaptation of downstream events, including lasting changes in gene expression by chronic treatment underlay the antidepressant efficacy. However, other molecular mechanisms could also be implicated since the desensitization of the ¿2-adrenoceptors seems to be necessary for observing the antidepressant effect. These receptors are coupled to G proteins, which activate the inwardly rectifier potassium channels (GIRK; also known as Kir3). Therefore, the action on the GIRK channel could be a new pharmacological target in the treatment of depression.GIRK channels are downstream effectors of G protein-coupled receptors (GPCRs), and they can be activated by direct binding of the ß¿ subunit of G protein, an interaction with important physiological consequences. In fact, GIRK channels hyperpolarize neurons in response to activation of GPCRs which leads to decrease the excitability of neurons, slow synaptic potentials and reduce volume transmission (Krapivinsky et al., 1995; Lüscher et al., 1997; Signorini et al., 1997; Wickman and Clapham, 1995; Wickman et al., 1998; Yamada et al., 1998). In mammalian four subunits (GIRK 1-4) have been cloned which assemble into homo and heterotetrameric channels (Wickman et al., 2002). GIRK1, GIRK2 and GIRK3 are broadly expressed in many areas of the centralSummary341nervous systems whereas GIRK4 expression is restricted to a small number of neuron populations (Aguado et al., 2008; Karschin et al., 1996; Wickman et al., 2000). Most neuronal GIRK channels are thought to be heteromeric containing GIRK1 and GIRK2 subunits (Luscher et al., 1997; Signorini et al., 1997, Torrecilla et al., 2002). Each GIRK subunit has a specific role in the activity of the channel. While GIRK3 subunits may regulate the availability of the GIRK channels on the plasma membrane, GIRK1 subunits are related to the time that channels remain open. However, GIRK2 subunits have a primary role controlling channel conductance. Indeed, the absence of GIRK2 subunit makes the channel become dysfunctional since mice lacking GIRK2 channels exhibit little or no GIRK current in a number of brain regions (Labouèbe et al., 2007; Slesinger et al., 1997).Up to date the availability of pharmacological agents that act on the GIRK channels has been poor. In this context, the use of GIRK2 knockout (GIRK2-/-) mice has been a key tool of insight into the role of specific GIRK channel composition on neuronal physiology, pathology and animal behaviour. Thus, GIRK channels functional relevance has been linked to susceptibility to seizures (Signorini, 1997), hyperalgesia and analgesia (Blednov et al, 2003; Marker et al., 2004, 2005; Smith et al., 2008), drug addiction and alcohol related behavioural effects (Blednov et al., 2001b; Morgan et al 2003; Kobayashi et al., 1999; Cruz et al., 2008), motor activity and co-ordination and reward and anxiety-related behaviours (Pravetonti and Wickman, 2008). Abnormal GIRK function can lead to excessive or deficient neuronal excitability which is related to severalSummary342pathologies such as epilepsy or Down¿s syndrome (Lüscher and Slesinger, 2010). Therefore, GIRK channels represent a key part in many types of neuronal communication and, as outlined in the following sections, modification of GIRK channel function and number could affect central nervous systems function in health and disease.Studies of locus coeruleus (LC) noradrenergic neurons have had a key role to understand the etiology of some psychiatric illnesses related to dysfunction of this neuromodulatory system (Itoi and Sugimoto, 2010). This nucleus contains the largest population of central noradrenergic neurons and innervates almost the entire neuroaxis (Dahlstrom and Fuxe, 1964; Swanson and Hartman, 1975). In LC neurons, activation of opioid receptors and 2-adrenoceptors leads to a decreased excitability (North and Williams, 1983; Pepper and Henderson, 1980; Williams et al., 1982), mainly though activation of GIRK channels (Aghajanian and Wang, 1986; Travagli et al., 1995; 1996; Williams et al., 1988), and in particular, those formed by GIRK2 and GIRK3 subunits (Cruz et al., 2008; Torrecilla et al., 2002; 2008). Additionally, constitutive GIRK channel activity contributes to the resting membrane potential of LC neurons in vitro (Torrecilla et al., 2002; Velimirovic et al., 1995). Moreover, in mice lacking both GIRK2 and GIRK3 subunits the spontaneous firing rate of LC neurons from brain slices and cortical NA concentrations are augmented (Cruz et al., 2008). However, the role of GIRK channels controlling LC-noradrenergic function in vivo has not been investigated yet.Summary343In this line, electrophysiological studies using Xenopus oocyte expression assays have suggested that GIRK channel modulators might have therapeutic benefits in the treatment of several neurologic disorders and cardiac arrhythmias (Hashimoto et al., 2006, Kobayashi et al, 2004; Kobayashi et al., 2010). Thus, GIRK channels have been proposed as new pharmacological targets that could be effective for the treatment of several illnesses related to an altered central neurotransmission.Taken all together, the main aim of this work was to investigate the involvement of GIRK2 subunit-containing GIRK channels on the noradrenergic transmission in vivo and in vitro as well as on the mechanism of action of antidepressant drugs. To this end, we established the following concrete objectives:To determine the contribution of GIRK2 subunit-containing GIRK channels to the bioelectric properties of noradrenergic neurons in the mouse LC in vivo and to quantify the levels of monoaminergic neurotransmitters in wild type (WT) and GIRK2 mutant mice.To characterize the effects induced by the administration of the ¿2-adrenoceptor agonist clonidine, and the ¿ opioid receptor agonist morphine, on the electric activity of LC neurons in GIRK2 mutant mice in vivo.Summary344To study the role of GIRK2 subunits-containing GIRK channels on the inhibitory effects induced by antidepressant drugs on noradrenergic neurotransmission by measuring the electrical activity of the LC neurons in WT and GIRK2 mutant mice in vivo and in vitro. Additionally, the effect of antidepresants on GIRK currents were examinated in LC slices from WT mice.To study the possible implication of GIRK channels on the behavioral response to stress and to the antidepressant-induced effect, by means of behavioral tests such as tail suspension test and activity box, using WT and GIRK2 mutant mice.For that, we used C57bj6 WT, heterozygous (GIRK2+/-) and homozygous (GIRK2-/-) mice (males and females) and performed electrophysiological recordings (in vivo and in vitro), neurochemical determinations by HPLC technique and behavioral experiments.Our results indicated that the electric characteristics of mice LC neurons were similar to those previously described in rats (Cerbadaum and Aghajanian., 1976) and mice (Gobbi et al., 2007). First, recorded cells presented biphasic excitation-inhibition response to pressure applied onto the contralateral hind paw. Second, the extra-cellular signal of the action potentials recorded from all genotypes (WT, GIRK2+/- and GIRK2-/-) displayed the typical bi-phasic morphology, with a large positive part that has notch and another final negative part. Third, the firing rate oscillatedSummary345between 0.5-5 Hz, and the mutation of the Girk2 gene did not alter the firing rate nor the coefficient of variation of LC neurons in anesthetized mice. However, both GIRK2+/- and GIRK2-/- animals had a higher percentage of LC neurons with burst firing pattern than WT mice. Furthermore, GIRK2-/- animals showed a greater percentage of burst firing neurons as compared with the GIRK2+/- mice. A more extended study of the burst pattern showed that in the latter genotype the mean spikes per burst as well as the mean inter-spike interval were reduced compared to those recorded from WT mice. Therefore, burst firing pattern was specifically modified in GIRK2+/- mice.Studies carried out in LC brain slices from WT mice showed that the mean basal frequency was lower than that recorded in vivo, which was in agreement with previous reports (Torrecilla et al., 2002; Miguelez et al., 2009). The application of the peptide that blocks the GIRK channels, tertiapin-Q (Jin and Lu, 1999,) increased the basal frequency of LC neurons in brain slice from WT mice indicating that GIRK channels control LC activity in a constitutive manner. In addition, as observed in the in vivo experiments, the basal frequency of LC neurons from GIRK2+/- mice in vitro was not different from that obtained from WT slices. Therefore, our in vitro experiments support the hypothesis that GIRK channels play a relevant role in the control of the basal electric activity of the LC neurons (Cruz et al., 2008; Torrecilla et al., 2002). However, simultaneous blockage of at least GIRK2 and GIRK3 subunits seems to be necessary to induce relevant changes in basal firing rate. In this sense, inSummary346LC neurons from double GIRK2 and GIRK3 knockout mice increased basal frequency is observed in vitro (Cruz et al., 2008).We next evaluated the role of GIRK2 subunits containing GIRK channels on the inhibitory effect induced by ¿2-adrenoceptor agonist, clonidine and ¿ opioid agonist, morphine in LC neurons in vivo. As expected, both agonists induced a dose-dependent inhibitory effect onto LC neurons firing rate in all genotypes. However, in mutant mice dose-response curves for clonidine and morphine were shifted to right as compared to those obtained in WT genotype. Statistical differences were also observed between GIRK2-/- and GIRK2+/- mice. In all cases, the maximal inhibitory effect was reached by agonist application. These results suggest that GIRK2 subunits participate on the inhibitory effect induced by ¿2 and ¿ agonist on LC activity although other subunits, likely GIRK3 subunits, may be necessary to reach maximal efficacy of the drug.When we tested the effect of the antidepressants reboxetine and desipramine (DMI) on LC basal activity we observed that in GIRK2+/- mice DMI was less potent though equally effective inhibiting electric activity compared to the WT group. On the other hand, the dose-response curve for reboxetine showed that this antidepressant was not only less potent but also less effective inhibiting LC neurons in GIRK2+/- animals. Furthermore, reboxetine did not have any effect onto LC neurons from GIRK2-/- mice. The antagonist of the ¿2-adrenoceptors RX821002 recovered to the same extent the inhibitory effect induced by the antidepressants in all tested genotypes.Summary347The reduced efficacy of reboxetine inhibiting LC activity from GIRK2 mutant mice was also observed in vitro. Thus, in LC slices from GIRK2+/- mice, unlike in the WT LC slices, a maximal concentration of reboxetine did not completely reduce the spontaneous firing rate. However, after reboxetine washed out noradrenaline (NE) perfusion, which was used as an internal control, produced a total inhibition of the firing rate that returned to basal values within 10 min after wash, in both WT and GIRK2+/- mice. On the other hand, perfusion of DMI onto WT LC slices produced a concentration-dependent and complete inhibition of firing rate. Tertiapin-Q not only recovered the inhibitory effect of DMI, but it also increased spontaneous firing frequency above basal values. Similarly, reboxetine completely inhibited LC firing rate in a concentration-dependent manner and tertiapin-Q recovered spontaneous firing rate toward basal values. Since reboxetine, unlike DMI, showed smaller efficacy inhibiting the firing rate of the LC neurons in GIRK2+/- animals in vivo and in vitro, it seems reasonable to argue that role of GIRK2 channels on the mechanism of action of these two antidepressants may be different. Taken into account that reboxetine selectively inhibits NE reuptake, its reduced efficacy could be due to adaptive changes in the NE transporter derived from the Girk2 gene mutation.On the other hand, our results obtained from LC WT slices using patch-clamp recordings were in disagreement with the hypothesis that suggests that antidepressant drugs elicit a direct blocking action on GIRK channels (Kobayashi et al., 2000; 2003; 2004; 2010): First, acute perfusion of fluoxetine, reboxetine and bupropion did not modify the LC potassiumSummary348currents registered at rest (-60 mV) or at more hyperpolarized potentials (-120 mV). Second, NE produced an increase of the resting potassium current which reverted at the potassium equilibrium potential and this NE-current was greater when reboxetine was co-perfused. These results indicated that reboxetine required exogenous NE for activating GIRK. Conversely, co-administration of bupropion did not significantly increase the amplitude of the NE-current.Our results also showed that the mutation of Girk2 gene had greater relevance in the regulation of basal amine levels related to serotonergic transmission. Thus, 5-HT levels were significantly reduced in the dorsal raphe nucleus of GIRK2-/- mice and also levels of its metabolite 5-HIIA was increased in the LC of this genotype. NE levels were only elevated in dorsal raphe nucleus of GIRK2+/- animals. GIRK2 mutation had no impact on monoamine basal levels of the prefrontal cortex.After the electrophysiological characterization, we performed behavioural experiments to study the impact of GIRK2 subunit mutation on the depression-related behaviors as well was on the antidepressant-like activity of drugs. For that, we have first validated the tail suspension test (TST) by evaluating the acute effect of several antidepressants on the immobility time in WT mice. The baseline behavior in the TST was examined after injections of physiological saline, DMI, reboxetine and fluoxetine. All antidepressant tested, significantly decreased the immobility time compared to control group. Reboxetine, in fact, was the most effective drug reducing immobility. We next examined the impact of GIRK2 subunitSummary349mutation on immobility time in the TST using WT and GIRK2 mutant mice. GIRK2+/- mice as well as GIRK2-/- mice showed reduced immobility time in the TST compared to WT group. Also, the behavioral effects of reboxetine were studied in GIRK2 mutant mice. Reboxetine did not alter significantly basal immobility time in GIRK2-/- mice although it induced a slight but significant increment of immobility time in GIRK2+/- mice. Also a pharmacological approach was done in WT mice by using intracerebroventricular injections of the GIRK channel blocker, tertiapin-Q. Thus, tertiapin-Q (30 and 100 pmol) dose-dependently decreased the immobility time in the TST. It should be pointed out that the magnitude of the response obtained from these experiments were not different to that observed in the GIRK2-/- genotype. Both the genetic blocking of the GIRK2 subunit and the pharmacological blocking of the GIRK channels improved behavioral response to stress. Moreover, tertiapin-Q injection 30 minutes before reboxetine administration completely prevented the effect of the antidepressant drug in WT mice. In fact, the immobility time values in this group were not different from those obtained in the GIRK2-/- mice who had also received reboxetine at the same dose. Thus, we can conclude that blocking the inhibitory effect of GIRK channels could alter depression-related behaviors and more importantly, the efficacy of the treatment.We next tested whether the reduction of immobility time observed in the GIRK2 mutant mice as well as that induced by blocking GIRK channels were due to an alteration of animal motor activity. For that purpose we used the activity chamber to measure mice locomotion duringSummary35030 min. As previously described, the results obtained confirm that complete delection of the GIRK2 subunit results in an increase of locomotive activity since this hyperactivity is not observed in the group of GIRK2+/- animals (Blenov et al., 2001b; Blenov et al., 2002; Pravetoni and Wickman, 2008). Thus, tertiapin-Q (30 pmol and 100 pmol) significantly reduced motor activity in WT mice. It is probable that the antidepressive-like effect observed in those animals is related to greater central monoaminergic activity. For this reason the GIRK2+/- genotype might be of greater interest than the GIRK2-/- to study pathological situations related to stress/depression and its treatment.On the basis of the aforementioned results, we draw the following conclusions:The GIRK2 subunit of GIRK channels plays a relevant role in controlling the basal electrical activity of mouse LC neurons in vivo, since in GIRK2+/- as well as in GIRK2-/- animals, a higher percentage of neurons fires in burst. Furthermore, the pharmacological blockade of the GIRK channel with tertiapin-Q increases the firing frequency of LC neurons in vitro, indicating that GIRK channels control LC activity in a constitutive manner.Summary351GIRK2 subunit-containing GIRK channels are implicated in the inhibitory effect mediated by the stimulation of ¿2-adrenoceptors and ¿ opioid receptors, since in GIRK2+/- and GIRK2-/- animals the inhibitory potency of both agonists was diminished. However, their efficacy remained unchanged suggesting that other subunits than the GIRK2 participate in the effect induced by the activation of ¿2-adrenoceptors and ¿ opioid receptors in vivo.The GIRK2 subunit of GIRK channels is implicated in the effect of the NE reuptake inhibitors, DMI and reboxetine. Although, its participation is significantly different for both antidepressants, since in GIRK2+/- mice DMI results in a total inhibition of firing rate whereas reboxetine fails to induce the maximal inhibition in this genotype and had no effect in GIRK2-/- mice. This efficacy loss was also observed in vitro. Since reboxetine selectively inhibits NE reuptake, while DMI also acts on other receptors, reboxetine reduced efficacy could be due to the presence of adaptive changes in the NE transporter derived from the GIRK2 mutation.Monoamine reuptake inhibitors do not block GIRK channels. In fact, fluoxetine, reboxetine and bupropion did not modify significantly the potassium currents registered at rest (-60 mV) or after the application of hyperpolarizing potentials (-120 mV) in LC neurons in vitro. However, reboxetine-induced small currentsSummary352were sufficient to inhibit the firing rate. Nevertheless, reboxetine increased the amplitude of the GIRK current induced by NE, probably as a consequence of increased NE levels in the synaptic cleft.GIRK2 subunit-containing GIRK channels are implicated in the regulation of amine levels in the central nervous systems. This regulation affects predominantly serotonergic neurotransmission. Thus, 5-HT levels are altered in GIRK2-/- mice in the dorsal raphe nucleus and its metabolite is altered in the LC. In contrast, NE levels are only elevated in dorsal raphe nucleus from GIRK2+/- animals. Significant changes in the prefrontal cortex were not observed.Blocking the inhibitory effect of GIRK channels could alter behavior related to depression and the efficacy of the treatment. Both the genetic blocking of the GIRK2 subunit and the pharmacological blocking of the GIRK channels results in an improvement in response to stress in mice. So, GIRK2+/- and GIRK2-/- animals, as well as WT animals treated with the channel blocking agent tertiapin-Q, showed reduced immobility in the TST, a similar effect to that observed with antidepressant drugs. Furthermore, the response to reboxetine was blocked under these experimental conditions.Summary353The GIRK2+/- genotype may be of greater interest than the GIRK2-/- genotype for studies of behavior related to stress, depression and its treatment. GIRK2+/- animals, unlike GIRK2-/- animals, do not show greater locomotor activity, and so the reduced immobility of the GIRK2+/- mice observed in the TST does not seem to be related to an alteration of locomotion, but rather seems to reflect a better response to stress.In summary, the present results demonstrate that GIRK2 subunit-containing GIRK channels participate in the tonic control of LC electric activity in vivo and in vitro and that they are implicated in the acute effects of several noradrenergic drugs onto noradrenergic transmission and antidepressant-like animal behavior. One of the classical problems associated with the pharmacological treatment of depression is the delay of the onset of clinical improvement, which is thought to be due to the desensitization of receptors coupled to GIRK channels. Therefore, given the role of the LC in depression and the fact that the reduction of the inhibitory role of GIRK channels triggers an increase in noradrenergic burst activity, the pharmacological blockage of GIRK channels constitutes a promising therapeutic candidate for faster onset of antidepressant response.