OUP user menu

Role of GIRK channels on the noradrenergic transmission in vivo: an electrophysiological and neurochemical study on GIRK2 mutant mice

María Torrecilla , Irrintzi Fernández-Aedo , Aurora Arrue , Mercedes Zumarraga , Luisa Ugedo
DOI: http://dx.doi.org/10.1017/S1461145712000971 1093-1104 First published online: 1 June 2013

Abstract

Dysfunctional noradrenergic transmission is related to several neuropsychiatric conditions, such as depression. Nowadays, the role of G protein-coupled inwardly rectifying potassium (GIRK)2 subunit containing GIRK channels controlling neuronal intrinsic excitability in vitro is well known. The aim of this study was to investigate the impact of GIRK2 subunit mutation on the central noradrenergic transmission in vivo. For that purpose, single-unit extracellular activity of locus coeruleus (LC) noradrenergic neurons and brain monoamine levels using the HPLC technique were measured in wild-type and GIRK2 mutant mice. Girk2 gene mutation induced significant differences among genotypes regarding burst activity of LC neurons. In fact, the proportion of neurons displaying burst firing was increased in GIRK2 heterozygous mice as compared to that recorded from wild-type mice. Furthermore, this augmentation was even greater in the homozygous genotype. However, neither the basal firing rate nor the coefficient of variation of LC neurons was different among genotypes. Noradrenaline and serotonin basal levels were altered in the dorsal raphe nucleus from GIRK2 heterozygous and homozygous mice, respectively. Furthermore, noradrenaline levels were increased in LC projecting areas such as the hippocampus and amygdale from homozygous mice, although not in the prefrontal cortex. Finally, potency of clonidine and morphine inhibiting LC activity was reduced in GIRK2 mutant mice, although the efficacy remained unchanged. Altogether, the present study supports the role of GIRK2 subunit-containing GIRK channels on the maintenance of tonic noradrenergic activity in vivo. Electric and neurochemical consequences derived from an altered GIRK2-dependent signalling could facilitate the understanding of the neurobiological basis of pathologies related to a dysfunctional monoaminergic transmission.

Key words
  • Dorsal raphe
  • HPLC
  • locus coeruleus
  • noradrenaline
  • single-unit extracellular recording

Introduction

Noradrenergic neurotransmission alteration underlies the aetiopathology of multiple psychiatric diseases, such as attention deficit/hyperactivity disorder, anxiety, depression and drug addiction (Itoi & Sugimoto, 2010; Sara, 2009). Medications that target the noradrenergic system are considered as valuable therapeutic tools for the treatment of these pathologies, as for example the selective noradrenaline (NA) reuptake inhibitors for depression, atomoxetine for deficit/hyperactivity disorder and clonidine for opiate withdrawal syndrome and detoxification (Cipriani et al. 2009; Hanwella et al. 2011; Papakostas et al. 2008; Soyka et al. 2011). Nevertheless, the treatment efficacy is in many cases unsatisfactory, as for example in depressive patients (Blier, 2008; Lieberman et al. 2005; Turner et al. 2008).

Studies of locus coeruleus (LC) noradrenergic neurons have had a key role in understanding the aetiology of some psychiatric illnesses related to dysfunction of this neuromodulatory system (Itoi & Sugimoto, 2010). This nucleus contains the largest population of central noradrenergic neurons and innervates almost the entire neuroaxis (Dahlstrom & Fuxe, 1964; Swanson & Hartman, 1975). In LC neurons, activation of µ opioid receptors and α2 adrenoceptors leads to a decreased excitability (North & Williams, 1983; Pepper & Henderson, 1980; Williams et al. 1982), mainly through activation of G protein-coupled inwardly rectifying potassium (GIRK; also known as Kir3) channels (Aghajanian & 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 yet been investigated.

Dysfunction of GIRK channels can lead to excessive or deficient neuronal excitability, which is related to several pathologies (Lüscher & Slesinger, 2010). In this context, the use of GIRK knockout mice has been a key tool to look into the role of specific GIRK channel composition on neuronal physiology and animal behaviour. Thus, GIRK2 subunit-containing GIRK channels functional relevance has been linked to increased susceptibility to seizures (Signorini et al. 1997), hyperalgesia and analgesia (Blednov et al. 2003; Marker et al. 2004, 2005), drug addiction and alcohol-induced behaviours (Blednov et al. 2001a; Cruz et al. 2008; Kobayashi et al. 1999; Morgan et al 2003), motor activity and coordination and reward and anxiety-related behaviours (Blednov et al. 2001b; Pravetoni & Wickman, 2008). In this way, electrophysiological studies using Xenopus oocyte expression assays have suggested that GIRK channel modulators might have therapeutic benefits in the treatment of several neurological disorders and cardiac arrhythmias (Hashimoto et al. 2006, Kobayashi et al. 2004, 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.

Therefore, the electrophysiological and neurochemical study of GIRK2 mutant mice could highlight the involvement of GIRK2 subunit-containing GIRK channels on noradrenergic transmission in vivo, which might be a key point for a better understanding of the neurobiological basis of the psychiatric diseases related to improper noradrenergic signalling.

Materials and method

Animals

Wild-type and GIRK2 adult mutant mice (aged 12–15 wk) derived from heterozygote crossing (Signorini et al. 1997) were used in electrophysiological and neurochemical experiments. Animals were housed 4–5 per cage in a colony room at 22 °C, under 12:12 h light:dark cycle (lights on 08:00 hours) with food and water provided ad libitum. Every effort was made to minimize animal suffering and to use the minimum possible number of animals. The procedures were approved by the Local Committee for Animal Experimentation at the University of the Basque Country. All the experiments were performed in compliance with the European Community Council Directive on ‘The Protection of Animals Used for Experimental and Other Scientific Purposes’ (86/609/EEC) Spanish Law (RD 1201/2005) for the care and use of laboratory animals.

Electrophysiological procedures

Mice were anaesthetized with chloral hydrate (400 mg/kg i.p.). After cannulating the trachea, the mouse was placed in the stereotaxic frame with the skull positioned horizontally. A burr hole was drilled and the recording electrode was placed 1.5 mm posterior to lambda and 0.2–1.2 mm from the midline and lowered into the LC usually encountered at a depth of between 2.7 and 4.0 mm from the brain surface (Gobbi et al. 2007). A catheter (Terumo Surflo®; Teruma Medical Products, USA) was then inserted in the peritoneo for additional administrations of anaesthetic and systemic drug. The body temperature was maintained at ∼37 °C for the entire experiment using a heating pad.

Single-unit extracellular recordings of mouse LC neurons were performed as previously described by Gobbi et al. (2007). The recording electrode was filled with 2% solution of Pontamine Sky Blue in 0.5% sodium acetate and broken back to a tip diameter of 1–2 µm. The electrode was lowered into the brain by means of a hydraulic microdrive (model 640; David Kopf® Instruments, USA). LC neurons were identified by standard criteria, which included spontaneous activity displaying a regular rhythm and firing rate between 0.5 and 5 Hz, characteristic spikes with a long-lasting (>2 ms), positive–negative waveform action potentials and the biphasic excitation–inhibition response to pressure applied on contralateral hind paw (paw pinch), as previously described in mice (Gobbi et al. 2007) and rats (Cedarbaum & Aghajanian, 1976).

The extracellular signal from the electrode was pre-amplified and amplified later with a high-input impedance amplifier and then monitored on an oscilloscope and on an audio monitor. This activity was processed using computer software (Spike2 software; Cambridge Electronic Design, UK) and the following patterns were calculated: firing rate; the coefficient of variation (percentage ratio of standard deviation to the mean interval value of an interspike time-interval histogram); percentage of spikes in burst; mean spikes/burst; percentage of cells exhibiting burst firing; response to drug administration. Basal firing rate and other electrophysiological parameters were measured for 3 min. Changes in firing rate were expressed as percentages of the basal firing rate (mean firing rate for 3 min prior to drug injection) and were measured after each dose of drug. Only one cell was studied in each animal when any drug was administered. A burst was defined according to Gartside et al. (2000) as a train of at least two spikes with the first interspike interval of ≤20 ms and a termination interval ⩾160 ms. Burst-firing of LC neurons was detected as a train of at least two spikes with the first interspike interval <80 ms and a termination interval >160 ms, as previously described in publications from our group (Miguelezet al. 2011a, b; Pineda et al. 1997). All active neurons recorded from each mouse were analysed according to the above defined criteria using the computer software Spike2 (script w_burst.s2s). Basal firing period (3 min) was used to evaluate whether the neuron displayed a burst-firing pattern or not. If electrical activity fitted burst criteria, percentage of spikes in burst and mean interspike interval of the burst train were also provided by the script.

Neurochemical procedures

The mice were anaesthetized with chloral hydrate (400 mg/kg i.p.) and decapitated and their brains were removed. The LC, dorsal raphe nucleus (DRN) prefrontal cortex (PFC), hippocampus (HPC) and amygdala (AM) were rapidly dissected on ice-cooled plate. The tissue was homogenized in 0.1 m perchloric acid, centrifuged (30 min, 12 500 g) and the supernatant was spin filtered through a 0.2 µm filter. Finally, one aliquot corresponding to 2 mg tissue was assessed by HPLC with amperometrical detection. The detector was operated at 0.6 V potential between the working electrode and the Ag/AgCl reference electrode with a sensitivity of 2 nA full-scale deflection. The mobile phase consisted of phosphoric–citric buffer (6.7 mm disodium phosphate, 13.3 mm citric acid, 0.05 µm EDTA, 3.64 mm heptanosulfonic acid, pH 3.1) and 15% methanol (v/v). The column was a 250 × 4.6 mm Symetry C18, 5 µm particle size (Waters Corporation, USA). All separations were achieved isocratically at room temperature. Flow rate was 1 ml/min. A calibration curve was performed daily within a concentration range of 0.15–2.4 ng of each standard per injection. The concentrations of NA, serotonin (5-HT) and its metabolite 5-hidroxyindolacetic acid (5-HIAA) in the tissue samples were calculated by extrapolation of the sample peak heights to the calibration curve.

Drugs

Chloral hydrate, clonidine hydrochloride and morphine hydrochloride (Sigma Aldrich, USA) were prepared in 0.9% saline. Drugs were prepared on the day of the experiment, except for chloral hydrate.

Statistical analysis of data

Compiled data are expressed as the mean±s.e.m., except for the percentage of neurons with burst firing. The sample size (n) represents the number of recorded neurons in electrophysiological experiments and the number of animals in neurochemical experiments. Dose–response curves for the inhibitory effect of clonidine or morphine were constructed by systemic administration of the drug at cumulative doses until a maximal response was reached. The inhibitory effect on LC cells induced by each dose was quantified as the percentage reduction from the basal firing rate. Experimental data regarding dose–response curves from each animal were analysed for the best simple nonlinear fit to the three-parameter logistic equation (Parker & Waud, 1971) using GraphPad Prism Software (v. 5.01; GraphPad Software Inc., USA) as described previously (Miguelez et al. 2009). The following equation was used: Embedded Image where [A] is the concentration of the drug, E is the effect on the firing rate induced by A, Emax is the maximal effect, EC50 is the effective concentration for eliciting 50% of Emax and n is the slope factor of the concentration–effect curve. Statistical comparison between dose–response curves parameters was done using the extra sum-of-squares F test (GraphPad Prism 5.01). Differences in the percentage of neurons presenting burst firing were statistically evaluated by two-tailed χ2 analysis of contingency tables. Parameters derived from burst pattern were analysed by a non-parametric test, Kruskal–Wallis test followed by Dunn's post-hoc test. Spontaneous firing rate, coefficient of variation and neurochemical data were compared through genotypes by one-way analysis of variance (ANOVA) followed by Newman–Keuls post-hoc test. Unpaired t test was used for selected pair comparisons. The level of significance was considered as p < 0.05.

Results

Electrophysiological properties of LC neurons in anaesthetized GIRK2 mutant mice

Here we investigated the role of GIRK channels on the in vivo spontaneous activity of LC neurons using GIRK2 mutant mice.

All LC neurons recorded from wild-type (WT) and GIRK2 mutant mice showed a characteristic spike with a long-lasting (>2 ms) and positive–negative waveform (Fig. 1 a). These neurons also displayed a biphasic excitation–inhibition response to a pinch of the contralateral paw (Fig. 1 b) as that previously described in mice (Gobbi et al. 2007) and rats (Cedarbaum & Aghajanian, 1976). Thus, the firing frequency of LC neurons from homozygous (GIRK2−/−) mice was not significantly different from that measured in WT or heterozygous (GIRK2+/−) mice (WT: 2.60 ± 0.16 Hz, n = 109; GIRK2+/−: 2.13 ± 0.14 Hz, n = 91; GIRK2−/−: 2.08 ± 0.16 Hz, n = 47, p > 0.05; Fig. 2 a). The coefficient of variation did not differ among genotypes (WT: 50.77 ± 1.41%, n = 109; GIRK2+/−: 45.82 ± 2.09%, n = 91; GIRK2−/−: 50.79 ± 2.06%, n = 47, p>0.05; Fig. 2 b). However, the number of neurons discharging in bursts was significantly higher not only in GIRK2−/− mice [55% of total recorded cells, χ2(1) = 46.15, p < 0.0001] but also in GIRK2+/− mice [26% of total recorded cells, χ2(1) = 7.79, p < 0.01] compared to WT animals (10% of the total recorded neurons). Statistical differences were also found between mutant genotypes [χ2(1) = 17.45, p < 0.0001; Fig. 2 c].

Fig. 1

In vivo extracellular signal of action potential recorded from locus coeruleus (LC) neurons in G protein-coupled inwardly rectifying potassium (GIRK)2−/− mice. (a) Example of a single spike from a mouse LC neuron recorded in vivo. The shape and duration of the action potential signal recorded from GIRK2−/− mouse LC were similar to those previously reported in the rat and wild-type (WT) mouse (Cedarbaum & Aghajanian, 1976; Gobbi et al. 2007). (b) LC neurons from GIRK2−/− mutant mice also responded to a pinch of the contralateral paw with a brisk increase in firing rate followed by a short pause (so-called Korf's response). Arrow indicates pinch of paw.

Fig. 2

The spontaneous firing pattern of locus coeruleus (LC) neurons from wild-type (WT) and G protein-coupled inwardly rectifying potassium (GIRK)2 mutant mice. The elimination of either one or both alleles for GIRK2 subunit did not alter the firing rate (a) or the coefficient of variation (b) of LC neurons recorded in vivo. All the values represented are the mean±s.e.m. of n experiments (n = 109, n = 91, n = 47 of WT, GIRK2+/− and GIRK2−/− mice, respectively). (c) The relative amount of LC neurons with burst firing was increased in GIRK2+/− and GIRK2−/− mice. Significant differences were also found between mutant groups. Burst-firing onset and offset was defined as the concurrence of two spikes with an interspike interval between 80 and 160 ms, respectively. * p < 0.01 and ** p < 0.001 vs. WT; # p < 0.001 vs. GIRK2+/−2 test). Each cell was recorded for 3 min.

A more extended study of the burst pattern was done by analysing several parameters, such as the mean firing rate, the percentage of spikes firing in bursts and the mean interspike interval recorded from LC burst firing neurons. Thus, the analysis of the mean firing rate showed significant differences between WT and GIRK2 mutant mice (Table 1). In fact, the burst pattern obtained from LC neurons of GIRK2+/− mice was significantly altered since the percentage of spikes that fired in burst and the mean interspike interval were reduced compared to those recorded from GIRK2−/− mice (p < 0.01 for both parameters, Kruskal–Wallis test followed by Dunn's post-hoc test). Also, statistically different results were obtained regarding the interspike interval of burst in WT and GIRK2+/− mice (p < 0.01, Kruskal–Wallis test followed by Dunn's post-hoc test) than those obtained in the GIRK2−/− mice (Table 1).

View this table:
Table 1

Electrophysiological properties of locus coeruleus neurons with burst activity: role of G protein-coupled inwardly rectifying potassium (GIRK)2 channels

WT (n = 11)GIRK2+/− (n = 25)GIRK2−/− (n = 26)
Firing rate (Hz)3.25 ± 0.621.90 ± 0.18**2.30 ± 0.16*
Spikes in burst (%)3.23 ± 0.990.87 ± 0.094.43 ± 0.87&
Mean interspike (ms)39.53 ± 4.4120.02 ± 3.20**34.71 ± 2.98&
  • * p < 0.05 and

  • ** p < 0.001 vs. wild-type (WT) mice.

  • & p < 0.01 vs. GIRK2+/− mice [Kruskal–Wallis test following Dunn's test. For firing rate analysis, one way analysis of variance (ANOVA) following Newman–Keuls test].

  • Each value represents the mean±s.e.m. of n experiments.

In neurons with no burst-firing, basal activity was not different among genotypes (WT: 2.53 ± 0.16 Hz, n = 98; GIRK2+/−: 2.34 ± 0.18 Hz, n = 66; GIRK2−/−: 1.78 ± 0.28 Hz, n = 21, F2,182 = 2.25, p = 0.108). However, the coefficient of variation of LC neurons from GIRK2+/− mice was significantly reduced with regard to that recorded from WT and GIRK2−/− mice (WT: 50.74 ± 1.50%, n = 98; GIRK2+/−: 41.55 ± 1.87%, n = 66; GIRK2−/−: 50.08 ± 2.80%, n = 21, F2,182 = 7.62, p < 0.001; data not shown).

Monoamine brain levels on GIRK2 mutant mice

We next studied whether Girk2 gene mutation affects monoamine basal levels in different brain areas. Thus, NA, 5-HT and 5-HIAA concentrations were evaluated in the LC, DRN, PFC, HPC and AM from WT, GIRK2+/− and GIRK2−/− mice.

GIRK2 subunit mutation did not alter NA levels in the LC (WT: 0.43 ± 0.03 ng/mg, n = 10; GIRK2+/−: 0.60 ± 0.09 ng/mg, n = 10; GIRK2−/−: 0.54 ± 0.04 ng/mg, n = 11; Fig. 3 a) nor in the projecting area of PFC (WT: 0.19 ± 0.01 ng/mg, n = 10; GIRK2+/−: 0.19 ± 0.01 ng/mg, n = 10; GIRK2−/−: 0.21 ± 0.02 ng/mg, n = 10; Fig. 3 c). In the HPC, NA levels were increased in the GIRK2−/− mice (WT: 0.19 ± 0.01 ng/mg, n = 18; GIRK2+/−: 0.23 ± 0.02 ng/mg, n = 7; GIRK2−/−: 0.26 ± 0.02 ng/mg, n = 10; F2,30 = 3.22, p = 0.054; Fig. 3 d) and statistical differences were revealed when comparison was restricted to the WT and GIRK2−/− genotypes (p < 0.05). Similar results were obtained in the AM (WT: 0.06 ± 0.01 ng/mg, n = 12; GIRK2+/−: 0.10 ± 0.02 ng/mg, n = 5; GIRK2−/−: 0.11 ± 0.02 ng/mg, n = 8; F2,22 = 2.13, p = 0.12; Fig. 3 e). Furthermore, comparison of only GIRK2−/− and WT genotypes revealed statistically significant differences for NA levels (p < 0.05). However, in the DRN, NA levels were significantly increased in the GIRK2+/− genotype compared to WT and GIRK2−/− mice (WT: 0.32 ± 0.03 ng/mg, n = 9; GIRK2+/−: 0.42 ± 0.02 ng/mg, n = 10; GIRK2−/−: 0.32 ± 0.03 ng/mg, n = 10; F2,26 = 4.41, p < 0.05; Fig. 3 b).

Fig. 3

Brain monoamine levels in G protein-coupled inwardly rectifying potassium (GIRK)2 mutant mice. Noradrenaline (NA) levels detected on locus coeruleus (LC) (a) and prefrontal cortex (PFC) (c) from mutant mice were not statistically different from those obtained in wild-type (WT) mice. However, NA levels in hippocampus (HPC) (d) and amygdala (AM) (e) from GIRK2−/− mice were significantly increased. Surprisingly, NA levels of dorsal raphe nucleus (DRN) from GIRK2+/− mice were increased compared to those detected in WT and GIRK2−/− genotypes (b). No statistical differences were detected among genotypes regarding basal levels of serotonin (5-HT) on LC (f) and PFC (h). However, 5-HT concentration in DRN of GIRK2−/− mice was significantly reduced as compared to those measured in WT mice (g). In the LC of GIRK2−/− mice, 5-hidroxyindolacetic acid (5-HIAA) concentration was elevated compared to those detected in other genotypes (k) and comparisons between GIRK2−/− and WT genotypes revealed statistically significant differences for metabolite levels on DRN (l). * p < 0.05, **p < 0.01 vs. WT, one-way analysis of variance (ANOVA), followed Newman–Keuls test or t test for selected pair comparisons). Data are expressed as means±s.e.m of n (5–18) experiments.

No differences were observed among genotypes regarding basal 5-HT levels in LC (WT: 0.40 ± 0.04 ng/mg, n = 10; GIRK2+/−: 0.56 ± 0.09 ng/mg, n = 10; GIRK2−/−: 0.37 ± 0.05 ng/mg, n = 9; p = 0.109; Fig. 3 f), PFC (WT: 0.28 ± 0.014 ng/mg, n = 9; GIRK2+/−: 0.29 ± 0.038 ng/mg, n = 10; GIRK2−/−: 0.20 ± 0.037 ng/mg, n = 10; Fig. 3 h), HPC (WT: 0.30 ± 0.02 ng/mg, n = 17; GIRK2+/−: 0.28 ± 0.03 ng/mg, n = 7; GIRK2−/−: 0.32 ± 0.02 ng/mg, n = 8; Fig. 3 i) and AM (WT: 0.27 ± 0.04 ng/mg, n = 12; GIRK2+/−: 0.32 ± 0.06 ng/mg, n = 5; GIRK2−/−: 0.26 ± 0.03 ng/mg, n = 8; Fig. 3 j). However, in the DRN, 5-HT levels were significantly reduced in the GIRK2−/− group compared to WT mice (GIRK2−/− 0.32 ± 0.05 ng/mg, n = 9; GIRK2+/−: 0.71 ± 0.12 ng/mg, n = 10; WT: 0.54 ± 0.03 ng/mg, n = 9; F2,25 = 5.61, p < 0.01; Fig 3 g).

Finally, the levels of the 5-HT metabolite, 5-HIAA, were measured in the brain areas already mentioned. Thus, in the LC 5-HIAA levels were significantly greater in the GIRK2−/− mice compared to those from the other genotypes (WT: 0.90 ± 0.07 ng/mg, n = 10; GIRK2+/−: 1.10 ± 0.13 ng/mg, n = 10; GIRK2−/−: 1.66 ± 0.19 ng/mg, n = 11; F2,28 = 7.31, p < 0.01; Fig. 3 k). In the DRN from GIRK2−/− mice, basal levels of 5-HIAA showed a slight increase (WT: 1.12 ± 0.063 ng/mg, n = 9; GIRK2+/−: 1.23 ± 0.12 ng/mg, n = 10; GIRK2−/−: 1.54 ± 0.15 ng/mg, n = 10; F2,26 = 3.09, p = 0.06). Further comparison of only GIRK2−/− and WT genotypes revealed statistically significant differences for metabolite levels on this nucleus (p < 0.05; Fig. 3 l). Girk2 mutation did not significantly alter 5-HIAA levels from PFC (WT: 0.28 ± 0.01 ng/mg, n = 9; GIRK2+/−: 0.26 ± 0.01 ng/mg, n = 10; GIRK2−/−: 0.32 ± 0.03 ng/mg, n = 10; Fig. 3 m), HPC (WT: 0.50 ± 0.04 ng/mg, n = 18; GIRK2+/−: 0.44 ± 0.04 ng/mg, n = 7; GIRK2−/−: 0.57 ± 0.04 ng/mg, n = 8; Fig. 3 n) and AM (WT: 0.52 ± 0.05 ng/mg, n = 12; GIRK2+/−: 0.50 ± 0.05 ng/mg, n = 5; GIRK2−/−: 0.54 ± 0.06 ng/mg, n = 8; Fig. 3 o).

The inhibitory effect of morphine and clonidine on the firing rate of LC neurons from GIRK2 mutant mice

Here, we investigated the role of GIRK2 subunit on the inhibitory effect induced by the activation of α2 and µ opioid receptors on the basal firing rate of LC neurons in vivo.

Thus, increasing doses of the α2 adrenoceptor agonist, clonidine (2–130 µg/kg i.p.) caused a progressive and complete inhibition of LC firing rate of all genotypes tested (Fig. 4 a). However, in GIRK2+/− mice, the dose–response curve for clonidine was shifted to the right, so that the EC50 mean value was increased by 2-fold with regard to that obtained in the WT group (EC50: 73.26 ± 9.06 µg/kg, n = 5; EC50: 35.89 ± 3.02 µg/kg, n = 10, for GIRK2+/− and WT, respectively, p < 0.0001, nonlinear fit analysis, extra sum-of-squares F test). In GIRK2−/− mice, the shift of the clonidine dose–response curve was greater as EC50 mean value increased 3-fold (EC50: 109.60 ± 12.78 µg/kg, n = 5, p < 0.0001, nonlinear fit analysis, extra sum-of-squares F test) when compared to that obtained in WT mice. Nevertheless, the maximal effect reached by clonidine remained unaltered among tested genotypes (Fig. 4 b).

Fig. 4

The inhibitory effect of clonidine on the basal firing rate of locus coeruleus (LC) neurons in wild-type (WT) and G protein-coupled inwardly rectifying potassium (GIRK)2 mutant mice. (a) Representative firing rate histograms illustrate the inhibitory effects of clonidine on LC activity from WT mice. (b) Dose–response curves of clonidine (2–130 µg/kg i.p.) show a progressive and complete inhibition of the firing rate in all genotypes tested. However, the potency of clonidine inhibiting the firing rate was reduced in the GIRK2+/− mice [effective concentration for eliciting 50% of maximal effect (EC50): 73.26 ± 9.06 µg/kg, n = 5, p < 0.0001] and even more in the GIRK2−/− animals (EC50: 109.60 ± 12.78 µg/kg, n = 5, p < 0.0001) compared to that in WT mice (EC50: 35.89 ± 3.02 µg/kg, n = 10). Each point of the curve represents the mean±s.e.m. of n experiments. Nonlinear fit analysis (extra sum-of-squares F for dose–response curves parameter statistical comparisons).

Similarly, the administration of the µ opioid receptor agonist, morphine (0.5–64 mg/kg i.p.), reduced the firing rate, progressively reaching a complete inhibition on WT mice (Fig. 5 a). In the GIRK2+/− mice, morphine showed a reduced potency inhibiting LC firing rate as EC50 value from the morphine dose–response curve was increased 1.4-fold as compared to that recorded from the WT group (EC50: 4.35 ± 0.45 mg/kg, n = 6; EC50: 6.15 ± 0.26 mg/kg, n = 4, for GIRK2+/− and WT respectively, p < 0.0001, nonlinear fit analysis, extra sum-of-squares F test). In the GIRK2−/− group, morphine showed a greater reduction in inhibitory potency, since EC50 value increased 8.3-fold (EC50: 36.30 ± 1.40 mg/kg, n = 5, p < 0.0001, nonlinear fit analysis, extra sum-of-squares F test) with regard to that obtained in WT mice. Conversely, the efficacy of morphine inhibiting LC firing rate remained unaltered among tested genotypes (Fig. 5 b).

Fig. 5

The inhibitory effect of morphine on the basal firing rate of locus coeruleus (LC) neurons in wild-type (WT) and G protein-coupled inwardly rectifying potassium (GIRK)2 mutant mice. (a) Representative firing rate histograms illustrate the inhibitory effects of morphine on LC activity from WT mice. (b) Dose–response curves of morphine (0.5–64 mg/kg i.p.) show a progressive and complete inhibition of the firing rate in all genotypes tested. However, the potency exhibited by morphine inhibiting the firing rate was reduced in the GIRK2+/− mice [effective concentration for eliciting 50% of maximal effect (EC50): 6.15 ± 0.26 mg/kg, n = 4, p < 0.0001] and difference was greater in the GIRK2−/− animals (EC50: 36.30 ± 1.40 mg/kg, n = 5, p < 0.0001) compared to that in WT mice (EC50: 4.35 ± 0.45 mg/kg, n = 6). Each point of the curve represents the mean±s.e.m. of n experiments. Nonlinear fit analysis (extra sum-of-squares F test for dose–response curves parameter statistical comparisons).

Moreover, in each case the inhibition produced by the agonist was similarly reverted in all genotypes with the corresponding antagonist: the α2 adrenoceptor antagonist RX 821002 (5–15 mg/kg i.p.; WT: 94.01 ± 16.84%, n = 4; GIRK2+/−: 78.61 ± 14.87%, n = 7; GIRK2−/−: 73.75 ± 17.91%, n = 4) and µ opioid receptor antagonist naloxone (5–12 mg/kg i.p.; WT: 69.60 ± 3.73%, n = 4; GIRK2+/−: 72.04 ± 21.05%, n = 3; GIRK2−/−: 84.06 ± 16.46%, n = 4; data not shown).

Discussion

The main purpose of the present study was to determine the role of GIRK channels containing GIRK2 subunits on the noradrenergic transmission in vivo. Our results show that Girk2 gene mutation alters electric activity of LC neurons in vivo, since a greater proportion of neurons display burst firing in mutant genotypes. In this respect, brain levels of monoamines are also modified in GIRK2 mutant mice, showing a more pronounced change in the serotonergic DRN and some projecting areas of the LC, such as HPC and AM. In fact, alteration of noradrenergic activity and the neurochemical changes observed in the serotonergic system from GIRK2 mutant mice could be a consequence of the reciprocal interactions between noradrenergic and serotonergic systems (Baraban et al. 1978; Haddjeri et al. 1997; Svensson et al. 1975).

The present electrophysiological study shows that neither the basal firing rate nor the coefficient of variation of LC neurons recorded in vivo was statistically different among tested genotypes. In contrast, the proportion of LC neurons with burst firing was bigger in the GIRK2+/− mice and that percentage was even greater in the GIRK2−/− group compared to that recorded from the WT group, which showed similar values to those previously reported by Gobbi et al. (2007). The resting membrane potential is depolarized in several brain areas from GIRK mutant mice, including LC neurons not only from homozygous but also from heterozygous GIRK2 mice (Chen & Johnston, 2005; Koyrah et al. 2005; Lüscher et al. 1997; Torrecilla et al. 2002). This is in agreement with the idea that a subpopulation of GIRK channels contributes to setting resting properties of LC neurons. Interestingly, GIRK1 protein levels are reduced in the brain of GIRK2+/− mice and nearly undetectable in GIRK2−/− mice (Signorini et al. 1997; Torrecilla et al. 2002). These alterations on GIRK1 and GIRK2 subunits could explain the changes observed in the basal noradrenergic transmission on GIRK2 mutant animals in vivo, because most LC GIRK channels are thought to be heteromeric, containing GIRK1, GIRK2 and GIRK3 subunits (Torrecilla et al. 2002). In LC slices taken from mice lacking both GIRK2 and GIRK3 subunits, intrinsic excitability is evidently increased only when synaptic activity is impaired (Cruz et al. 2008). Since glutamatergic input to the LC is important for burst activity (Tung et al. 1989), the increased proportion of neurons with burst patterns could be related with altered glutamatergic transmission on GIRK2−/− mice. Recently, it has been described that dopaminergic neurons from GIRK2−/− mice exhibit increased glutamatergic synaptic activity and reduced basal firing rate in vitro (Arora et al. 2010). Conversely, excitatory neurotransmission in LC brain slices from double GIRK2 and GIRK3 knockout mice is not altered with regard to that recorded in WT mice (Torrecilla et al. 2008). The lack of altered excitatory activity could, however, reflect the development of adaptive changes derived from the double mutation. The present study also shows that, while the mean firing rate of LC neurons is similar in WT and GIRK2 mutant mice, the burst activity is significantly higher in the mutant groups. These results also support the idea that burst firing analysis is an adequate approach to examine neuronal excitability and therefore the role of GIRK2 subunit-containing GIRK channels on basal properties of LC neurons in vivo.

In addition, the present study indicates that altered monoaminergic levels in the LC and DRN could be related to different burst firing patterns observed between GIRK2 mutant mice. Thus, in the heterozygous genotype, a modified burst pattern might be sufficient to specifically increase the levels of NA in the DRN. In homozygous mice, however, the greater population of burst neurons, which showed a firing pattern similar to that of WT mice, could be linked to pronounced changes on serotonergic transmission. Thus, the 5-HT decrease detected in the DRN from homozygous mice and increased levels of 5-HIAA in the LC and DRN might support the idea of an altered metabolism of 5-HT on these mice. These results indicate that partial or complete Girk2 gene mutation might differentially modulate the reciprocal interactions between LC and DRN, which regulate neuronal activity of these nuclei (Guiard et al. 2008; Labonte et al. 2011; Miguelez et al. 2011b). An explanation for the present results could be that increased burst firing from LC neurons in GIRK2+/ mice might cause a release of NA that is not sufficient to reduce 5-HT levels on the DRN by activation of local α2 adrenoceptors (Haddjeri et al. 1997; Pudovkina et al. 2003). However, in GIRK2−/− mice, the greater proportion of burst firing neurons would facilitate NA transmission that subsequently would reduce the release of 5-HT on the area. This neurotransmitter exerts an inhibitory control onto LC activity (Haddjeri et al. 1997; Szabo & Blier, 2001), pretreatment with the 5-HT synthesis inhibitor parachlorophenylalanine increases burst activity of LC neurons and short treatment with escitalopram, a serotonin selective reuptake inhibitor, has the opposite effect (Dremencov et al. 2007). Therefore, reduced inhibitory effect driven by decreased 5-HT levels on the DRN might contribute to increased burst activity of LC neurons from GIRK2−/− mice. This supports the idea that functional DRN × LC interactions should be taken into account to better explain the relationship between electrophysiological and neurochemical data obtained from the GIRK2 mutant mice. Furthermore, since 5-HT1A receptors from the DRN are coupled to GIRK, the contribution of these cellular effectors to DRN basal activity, and therefore functional consequences onto NA system, warrants further investigation.

Moreover, functional correlation of electrical activity and monoamine content in projecting areas is usually complex. Thus, genetic and pharmacological disruption of neurokinin 1 receptor function augments DRN 5-HT neuronal firing by increasing the burst firing activity of LC neurons, but does not alter 5-HT basal levels (Gobbi et al. 2007; Santarelli et al. 2001). Recently, it has been shown that 5-HT1A autoreceptor levels determine vulnerability to stress and response to antidepressants (Richardson-Jones et al. 2010). In fact, a lower level of 5-HT1A autoreceptors greatly affects DRN firing rates, but has no effect on basal forebrain 5-HT levels. In our study NA levels in some LC projecting areas, such as HPC and AM, were increased in GIRK2 mutant mice although mean basal firing rate of LC neurons was not altered. In this respect, simultaneous deletion of GIRK2 and GIRK3 subunits significantly increases both NA concentration in the cortex and LC firing rate in vitro (Cruz et al. 2008). Taken together, these data suggest that ablation of GIRK2 subunits is not sufficient to simultaneously modify LC mean firing rate and NA levels in projecting areas although burst-firing pattern was facilitated in a greater proportion of neurons.

An altered burst firing pattern is linked to depression-like behaviours and a beneficial response to antidepressants (Covington et al. 2010; Dremencov et al. 2009). Sustained administration of serotonin selective reuptake inhibitors decreases burst activity from noradrenergic and dopaminergic neurons, which may account for an antidepressant-resistant response in some patients (Dremencov et al. 2007, 2009). In contrast, chronic administration of imipramine increases burst firing in LC slices (Maubach et al. 2002). Atypical antipsychotics, which increase burst activity of LC neurons, are effective adjuncts in antidepressant-resistant patients (Rapaport et al. 2006). In mice lacking neurokinin 1 receptors, in which a greater percentage of LC neurons display burst firing, the rewarding effects of opioids are absent, physical response to withdrawal is reduced and anxiety-related behaviours are decreased (Murtra et al. 2000; Santarelli et al. 2001). Therefore, medications that selectively increase burst activity of monoaminergic neurons, such as drugs that target GIRK channels, might be a more efficacious treatment of depressive behaviour and other psychiatric diseases. In this sense, GIRK2 knockout mice show blunted behavioural responses to ethanol (Blednov et al. 2001a), diminished cocaine self-administration (Morgan et al. 2003) and reduced anxiety-related behaviour (Blednov et al. 2001b; Pravetoni & Wickman, 2008). In line with this, a recent study found that the acute administration of drugs that block GIRK channels have antidepressant activity in the forced swimming test without modifying locomotive activity in rats (Kawaura et al. 2010).

The present study also shows that µ and α2 receptors of LC from GIRK2 mutants are desensitized. In this respect, the dose–response curves for the inhibitory effect on LC firing rate induced by both morphine and clonidine were shifted to the right in GIRK2 mutant mice. Importantly, the reduced function of α2 and 5-HT1A autoreceptors has been reported as a key consequence of long-term antidepressant treatment, which is related to the slow onset of the clinical response to the treatment (Artigas et al. 1996; Grant & Weiss, 2001; Invernizzi et al. 2001; Lacroix et al. 1991; McMillen et al. 1980; Szabo et al. 2000). Thus, all the pharmacological or genetic strategies that faster desensitize the above-mentioned autoreceptors have been of great therapeutic interest (Pérez et al. 1997; Portella et al. 2011; Richardson-Jones et al. 2010; Sanacora et al. 2004; Santarelli et al. 2001). Further investigations are needed to determine whether disruption of the inhibitory function of GIRK channels, which are coupled to both α2 and 5-HT1A autoreceptors, could be a new strategy in the treatment of depression.

Dose–response curves also indicate that, although agonist potency is reduced, the inhibitory efficacy remained maximal. A similar range of doses has been used to study the antinociceptive properties of clonidine and morphine in GIRK mutant mice (Blednov et al. 2003; Cruz et al. 2008; Mitrovic et al. 2003). Thus, while analgesic potency of morphine is diminished, its efficacy is preserved in GIRK2 and GIRK2/GIRK3 knockout mice (Cruz et al. 2008; Mitrovic et al. 2003). The activation of µ opioid receptors in LC brain slices from GIRK2 knockout mice induces a decreased GIRK current, which hyperpolarizes neuron membrane potential (Torrecilla et al. 2002). Conversely, in mice lacking GIRK2 and GIRK3 subunits, opioid current is greatly diminished and therefore membrane hyperpolarization is barely detectable (Cruz et al. 2008; Torrecilla et al. 2002). This suggests that GIRK3 subunit may be important for the inhibitory effect of opioids and α2 agonist on LC activity in vivo. In fact, morphine-induced post-synaptic inhibition of LC neurons, which is required for induction of dependence, is absent in GIRK2/GIRK3−/− mice (Cruz et al. 2008).

Therefore, a genetic or pharmacological blockade of GIRK channels may constitute a new approach to treatment of psychiatric pathologies such as depression and drug addiction.

The present study further supports the role of GIRK channels as important mediators of neuronal excitability in vivo and brings up the need for further investigations to study the potential links with pathologies related to improper monoaminergic transmission.

Statement of Interest

None.

Acknowledgements

The present work has been supported by specific grants from the Basque Government (S-PE09UN27 and IT436-10), the University of the Basque Country UPV/EHU (UFI 11/32) and the Spanish Government (SAF-2009-08664). We thank Dr Kevin Wickman (Department of Pharmacology, University of Minnesota) for generously providing GIRK2 mutant mice and Dr Cristina Miguélez (University of the Basque Country) for helpful comments on this manuscript.

References

View Abstract