INTRODUCTION

In view of the inadequacies of current treatment of depression, considerable efforts are being made to find improved drugs (Millan, 2004, 2006; Morilak and Frazer, 2004). In this regard, much attention has been devoted to NK1 receptors which transduce the actions of substance P (Rupniak et al, 2001; Holmes et al, 2003). Several preclinical and clinical observations support the utility of NK1 receptor blockade in the control of depressed—and anxious—states. First, long-term treatment with various classes of antidepressant modifies central tissue levels of substance P, an effect observed in the absence of an alteration in NK1 receptor expression (Shirayama et al, 1996; Sartori et al, 2004). Second, NK1 receptors are localized in: the dorsal raphe nucleus (DRN), the origin of ascending serotonergic projections (Froger et al, 2001; Santarelli et al, 2001; Commons and Valentino, 2002; Ma and Bleasdale, 2002; Lacoste et al, 2006); the locus coeruleus, the source of adrenergic input to corticolimbic regions (Santarelli et al, 2001; Ma and Bleasdale, 2002); and the ventrotegmental area (VTA), from which dopaminergic pathways project to the frontal cortex (FCX) and limbic regions (Lessard and Pickel, 2005). Third, NK1 receptors are broadly distributed in limbic structures implicated in the control of mood, such as the FCX, hippocampus, lateral septum, nucleus accumbens, basolateral amygdala (BLA), and hypothalamus (Liu et al, 2002; Saffroy et al, 2003). Fourth, substance P exerts aversive effects in rodents, whereas mice genetically deprived of NK1 receptors reveal improved resistance to stress and display an anxiolytic phenotype (Rupniak et al, 2001; Santarelli et al, 2001). Fifth, selective NK1 receptor antagonists exert antidepressant and anxiolytic actions in several, though not all, experimental models (Rupniak et al, 2001; Varty et al, 2002; Dableh et al, 2005; Renoldi and Invernizzi, 2006), and they attenuate the behavioral response to stress (Ebner et al, 2004).

The above observations prompted clinical evaluation of several selective NK1 receptor antagonists, including MK869 (aprepitant) and L759274, which revealed beneficial effects in major depression with improved tolerance (less sexual dysfunction) as compared to SSRIs (Kramer et al, 1998, 2004; Gerald et al, 2006). Nonetheless, therapeutic trials of NK1 antagonists have not overall yielded convincing evidence of robust and reproducible antidepressant actions, tempering enthusiasm for their further development. This has triggered interest in the concept of combining NK1 antagonist properties with suppression of 5-HT reuptake, either as drug associations or in a single chemical structure (Ryckmans et al, 2002; Guiard et al, 2004; Millan, 2006). Such agents would be expected to exert complementary antidepressant properties, potentially with greater efficacy and rapidity of action than SSRIs and selective NK1 blockers, respectively. Support for this notion has been generated by both neurochemical and electrophysiological studies in rodents. Thus, the influence of SSRIs upon extracellular levels of 5-HT in FCX was reinforced in mice genetically lacking NK1 receptors, possibly due to desensitization of feedback actions at inhibitory 5-HT1A autoreceptors in the DRN (Froger et al, 2001; Santarelli et al, 2001; Guiard et al, 2005; Gobbi et al, 2007). By analogy, Guiard et al (2004) demonstrated that the effect of paroxetine on frontocortical 5-HT levels in mice were potentiated by the NK1 antagonists, GR205171 and L733060.

These elegant studies underpin the notion that NK1 receptor blockade may enhance the influence of SSRIs upon serotonergic transmission, but several important issues and questions remain to be addressed. First, neurochemical studies of the influence of NK1 antagonists on the actions of SSRIs have to date been restricted to mice, and it would be desirable to reproduce such findings in other species. Second, the influence of NK1 receptor blockade upon SSRI-elicited increases in 5-HT levels has been evaluated in the FCX, but it would be interesting to extend such work to other corticolimbic structures implicated in depressed states, like the hippocampus, BLA, nucleus accumbens, and striatum. Third, certain ‘SSRIs’ also enhance extracellular levels of noradrenaline (NA) and dopamine (DA) in FCX (Millan et al, 2000), but it has not been determined whether NK1 receptor antagonists likewise modify the influence of SSRIs upon dopaminergic and adrenergic transmission. Fourth, subactive doses of SSRIs exert antidepressant actions in the mouse forced swim (FS) test in the presence of GR205171 (Chenu et al, 2006). Remarkably, however, no other data on how NK1 antagonists may modify the behavioral actions of SSRIs are available. In this light, it would be of particular interest to examine whether the acute anxiogenic actions of SSRIs (Dekeyne et al, 2000) are modified by blockade of NK1 receptors.

The present studies addressed these questions employing combined neurochemical, electrophysiological, and behavioral approaches, together with the highly selective NK1 receptor antagonist, GR205171. In contrast to most agents, GR205171 possesses high affinity at NK1 receptors across a broad range of species, though its affinity is higher for NK1 receptors in gerbils and guinea pigs (which closely resemble those in humans) than rats and mice (Rupniak et al, 2000; Griffante et al, 2006; Engberg et al, 2007). GR205171 was also chosen in view of its well-characterized actions in vivo (Gardner et al, 1996; Millan et al, 2001b; Lejeune et al, 2002; Guiard et al, 2004; Chenu et al, 2006), and as clinical studies have shown that it relieves social phobia (Furmark et al, 2005; Michelgard et al, 2007). To underpin the stereospecificity of the actions of GR205171, its effects were compared to those of its less active distomer, GR226206 (op. cit.).

MATERIALS AND METHODS

Animals

Unless specified below, these studies employed male Wistar rats (225–250 g body weight upon arrival) and male CD mice (22–26 g upon arrival) supplied by Charles River (L’Arbresle and Saint Aubin les Elbeuf, France, respectively). Male Mongolian gerbils (50–70 g upon arrival) were acquired from CERJ (Le Genest, St-Isle, France). Male Syrian hamsters (70–80 g upon arrival) were obtained from Charles River Laboratories (Kingston, NY, USA). Rats, mice, and gerbils were housed in standard Macrolon, sawdust-lined cages and hamsters were housed in polycarbonate cages lined with sterilized pine chips and unrestricted access to food and water. Except for hamsters (see below), there was a light/dark cycle with lights on from 07 : 30 to 19 : 30 hours. All animal use procedures conformed to international European ethical standards (86/609-EEC) and the French National Committee (décret 87/848) for the care of laboratory animals. Hamster use was approved by the Institutional Animal Care and Use Committee of Valdosta State University and complied with regulations outlined in the US Animal Welfare Act.

Cerebral Microdialysis and Chromatographic Procedures

The protocol used for quantification of levels of 5-HT, DA, and NA in dialysate samples was detailed elsewhere (Millan et al, 2001b). Briefly, a guide-cannula was implanted under pentobarbital anesthesia (60 mg/kg, i.p.) into the FCX, nucleus accumbens, striatum, or dorsal hippocampus of rats as previously described (Millan et al, 2001b). For the BLA, coordinates were AP, 2.8; ML, ±4.9; and DV, 6.4 from dura. Experiments were performed 5 days later after placement of the guide cannula following placement of a cuprophane CMA/11 probe: 0.24 mm in diameter, 4 mm in length for FCX and striatum, and 2 mm for the dorsal hippocampus, BLA and nucleus accumbens. Samples were taken every 20 min over 1 h, then GR205171 or vehicle was injected i.p. followed, 20 min later, by s.c. administration of citalopram, fluoxetine, 8-OH-DPAT or S15535. In further experiments, injection of citalopram was preceded by treatment with GR226206 or RP67580. Sampling was pursued for a further 3 h. Monoamine levels were quantified as described previously (above citations) by HPLC and electrochemical detection.

Electrophysiological Procedure

As detailed previously (Lejeune et al, 2002), rats were anaesthetized with chloral hydrate (400 mg/kg, i.p.) and, after cannulation of the femoral vein, placed in a stereotaxic apparatus. A tungsten microelectrode was slowly lowered into the DRN (from bregma and the sinus surface, AP, 7.2; ML, ±0.0; and DV, 5.5/6.5). After amplification (CP511; Grass Technology, USA) and A/D conversion (micro1401mkII; CED, Cambridge, UK) of electrical activity, data were recorded using Spike2 software (CED). Serotonergic neurones were identified by their waveform and spontaneous firing patterns (≈1.2 Hz). One cell was recorded per animal. The spontaneous firing rate was recorded for 5 min before i.v. administration of vehicle, GR205171, or GR226206. After 3 min, the influence of citalopram, fluoxetine, 8-OH-DPAT, S15535, (−)-pindolol, or buspirone, injected i.v. in volumes of 0.5 ml/kg, was evaluated upon administration in cumulative doses at intervals of 2–3 min. Drug effects were characterized over 60 s periods at the time of peak action.

Forced Swim Test in Mice

As previously described (Brocco et al, 2006), mice were placed in individual glass cylinders (24 cm h × 12 cm diameter) containing 6 cm of water at 25°C for 6 min. Immobility (s) was measured during the last 4 min of the test. In dose-range studies of citalopram (s.c.) and GR205171 (i.p.), drug or vehicle was administered 30 min before testing. In antagonist studies, GR205171, GR226206, RP67580, or vehicle was administered (i.p.) 30 min before treatment with citalopram (2.5 mg/kg, s.c.) or vehicle, and testing undertaken 30 min later.

Influence Upon Locomotor Activity in Mice

As described (Brocco et al, 2002), locomotor activity in mice was evaluated using white Plexiglass cages (27 × 27 × 27 cm) equipped with two rows of four photocells 2 cm above the floor and 6 cm apart connected through an interface to a microcomputer (Hesperid, Loiron, France). GR205171 or vehicle (i.p.), then citalopram or vehicle (s.c.), was administered 60 and 30 min before placing mice in individual activity chambers for 10 min. Each interruption of an infrared beam was counted as a movement.

Ultrasonic Vocalization Test in Rats

As previously (Millan et al, 2001a), rats were initially placed in a chamber equipped with a grid floor and were exposed to six randomly distributed, electric shocks (800 μA, 8 s) over a 7 min period. After 24 h, they were placed in the chamber for 2 min and received a single shock. They were returned to the chamber 30 min later and the total duration of ultrasonic vocalizations (USVs), defined as vocalizations with frequencies higher than 20 kHz, were recorded as previously (Millan et al, 1997a) over 10-min by use of a microphone linked to an ultrasound recording system (Ultravox; Noldus, Wageningen, The Netherlands). Rats emitting USVs for less than 90 s were not examined further. After 24 h, the procedure was replicated following drug administration. GR205171, GR226206, citalopram or vehicle was administered immediately after the 2 min session and, in potentiation studies, animals received GR205171, GR226206 or vehicle, followed by citalopram or vehicle, immediately after the 2 min session.

Fear-Induced Foot Tapping in Gerbils

As previously (Brocco et al, 2008), the procedure involved two sessions: (1) acquisition of fear conditioning followed, 4 h later, by (2) a test session. The apparatus was a chamber with a floor comprised of four metal plates (10 × 4 cm) connected to a shocker (Apelex, Massy, France). In the acquisition session, naive gerbils were placed for 2 min habituation in the chamber, then a shock was (1.75 mA, 0.5 s) delivered manually each time the animal crossed from one plate to another. Each animal received 10 shocks, separated by at least 10 s intervals without shock. For testing, the animal was placed again in the four-plate chamber for 3 min and the duration (s) of foot-tapping bouts recorded. No shock was delivered during the test session. GR205171, GR226206, or vehicle was administered i.p. together with either vehicle or fluoxetine (40.0 mg/kg, i.p.), 30 min before the test session.

Social Interaction Tests in Rats and Gerbils

As previously (Millan et al, 2001a; Brocco et al, 2008), male Sprague–Dawley rats of 240–260 g (Charles River, Saint-Aubin-les-Elbeuf, France) or gerbils were maintained under a 12/12 h low light (3 lux)/dark cycle for 5 days before testing, and individually housed 5 days (rats) or 3 h (gerbils) before testing. On the test day, they were placed in weight-matched pairs (±5 g for rats and ±3 g for gerbils) in opposite corners of a highly illuminated (300 lux) open-topped arena for a 10 (rats) or 5 min (gerbils) observation session. Data were the duration of active SI: ie, the time spent in grooming, following, sniffing, biting, jumping, or crawling over or under the other animal. If animals remained adjacent to each other without any movement for more than 10 s, scoring was discontinued until active SI resumed. Both animals of the same pair receive the same drug treatment. GR205171, citalopram, or vehicle was administered 30 min before testing. In interaction studies, GR205171, GR226206, or vehicle was administered 45 min before testing, and citalopram or vehicle 30 min before testing.

Light-Induced Phase-Advances in Circadian Wheel Running Rhythms in Hamsters

As previously (Gannon and Millan, 2007), Syrian hamsters were maintained in a 14 : 10 h light : dark schedule for several weeks before being transferred to conditions of constant darkness (DD) where they had access to small running wheels of 19 cm in diameter. Wheel running was recorded in 10 min bins using Actimetrics ClockLab hardware (Evanstown, IL, USA). Food and water was provided ad libitum at all times. Onset of wheel running in DD is defined as circadian time (CT 12). Data were recorded for approximately 10 days in DD and then hamsters were removed from their home cages under dim red light (<1 lux) at circadian time 18.25, weighed, and injected with either drug or vehicle and returned to their home cage. After 45 min at CT, 19 hamsters were again removed from their home cage and exposed to a 10-min pulse of white light (20 lux) and then returned to their home cage for another 10 days. Hamsters were returned to the 14 : 10 h lighting schedule at the conclusion of the experiment. Light pulses delivered at CT 19 to hamsters in DD phase advance the time of onset of wheel running. The magnitude of the phase advance was determined by fitting a line through the activity onset for several days before the light pulse and again for several days after the light pulse once the rhythm has stabilized, normally days 5–10 after light. The difference in time between the two fitted lines on the day of the experiment indicates the magnitude of the light-induced phase advance.

Drugs and Sources

Drug doses are expressed in terms of the base. Drugs were dissolved in distilled water and administered s.c., i.p., or i.v. (electrophysiology). For i.p. administration, drugs were administered as suspensions in water with a few drops of Tween 80. GR205171 (2-methoxy-5-(5-trifluoromethyltetrazol-1-yl)-benzyl-([2S,3S]-2-phenylpiperidinyl)-amine) diHCl; GR226206 (2-methoxy-5-(5-trifluoromethyltetrazol-1-yl)-benzyl-([2R,3R]-2-phenylpiperidinyl)-amine) diHCl; citalopram HBr, RP67580 ((3aR,7aR)-imino-1(methoxy-2-phenyl)-2-diphenyl-7,7-perhydro isoindolone-4), and S15535 (1-(1,4-benzodioxan-5-yl)-4-(indan-2-yl) piperazine mesylate) were synthetized by Servier chemists. 8-OH-DPAT (8-dihydroxy-2(di-n-propylamino)tetralin) HBr, (−)-pindolol, WAY-100,635 (N-(2-(4-(2-methoxyphenyl)piperazin-1-yl)ethyl)-N-(2-pyridyl)cyclohexane carboxamide) maleate and buspirone were purchased from Sigma-RBI (Natick, USA). Fluoxetine HCl was purchased from Interchim (Montluçon, France).

RESULTS

Enhancement by GR205171 of the Influence of Citalopram and Fluoxetine Upon Dialysis Levels of 5-HT in Frontal Cortex of Freely Moving Rats

At a dose of 0.63 mg/kg, s.c., citalopram markedly increased levels of 5-HT in the FCX of freely moving rats (Figure 1; Table 1). In contrast, levels of DA and NA were not modified. In an opposite fashion, the selective NK1 antagonist, GR205171 (10.0–40.0 mg/kg, s.c.), had no effect on frontocortical levels of 5-HT, whereas it elevated levels of DA and NA. This effect of GR205171 was expressed stereospecifically in that its less active distomer, GR226206 (40.0 mg/kg, s.c.), did not significantly influence levels of DA, NA, or 5-HT (p>0.05, not shown). GR205171 (10.0–40.0 mg/kg, i.p.) dose-dependently potentiated the increase in extracellular levels of 5-HT induced by citalopram (0.63). This effect was exerted stereospecifically in that GR226206 (40.0) did not significantly modify the citalopram-induced increase in 5-HT levels. Area under the curve analyses (%, expressed relative to basal values, defined as 100%) is as follows: vehicle/citalopram (0.63)=204.8±10.7 vs GR226206 (40)/citalopram (0.63)=207.3±11.3, P>0.05. The combination of GR205171 with citalopram did not differentially influence levels of DA and NA as compared to GR205171 alone. There was no significant influence of GR226206 in combination with citalopram upon levels of DA and NA (P>0.05, not shown). GR205171 (10.0–40.0 mg/kg, i.p.) also dose-dependently potentiated the increase in extracellular levels of 5-HT induced by fluoxetine in FCX. This effect was expressed stereospecifically in that GR226206 (40.0) was inactive. Area under the curve analyses is as follows: vehicle/fluoxetine (10.0) vs GR226 206/fluoxetine (10.0), 5-HT=191.9±6.3 vs 215.5±7.7, P>0.05. In addition, at a dose of 10.0 mg/kg, s.c., fluoxetine increased levels of DA and NA. In the presence of increasing doses of GR205171 (10.0–40.0 mg/kg, i.p.), fluoxetine enhanced levels of DA and NA. GR226206 did not modify the influence of fluoxetine upon DA and NA levels. Area under the curve analysis: vehicle/fluoxetine (10.0) vs GR226206 (40.0)/fluoxetine (10.0), DA=161.0±5.2 vs 172.9±4.7, P>0.05 and NA=170.0±3.3 vs 184.1±3.6, P>0.05.

Figure 1
figure 1

Enhancement by the neurokinin (NK)1 receptor antagonist, GR205171, of the influence of citalopram upon dialysis levels of serotonin in frontal cortex of freely-moving rats. Serotonin (5-HT), dopamine (DA), and noradrenaline (NA) levels are expressed relative to basal values defined as 100%. These were 0.96±0.06, 1.12±0.07, and 1.84±0.08 pg per 20 μl for 5-HT, DA, and NA respectively. Data are means±SEMs; N=5–9 per value. Arrows denote the injection of drugs. For ANOVA with dose as between factor, see legend to Table 1.

Full size image
Table 1 Enhancement by NK1 Receptor Antagonists of the Influence of Citalopram and Fluoxetine Upon Dialysis Levels of Serotonin in Frontal Cortex
Full size table

Enhancement by GR205171 of the Influence of Citalopram Upon Dialysis Levels of 5-HT in Diverse Cerebral Structures of Freely Moving Rats

In the dorsal hippocampus, BLA, nucleus accumbens, and striatum, at a dose of 0.63 mg/kg, s.c., citalopram increased levels of 5-HT (Figure 2; Table 2). It also slightly elevated the levels of DA and NA in the BLA. By contrast, it did not affect NA (dorsal hippocampus) or DA (nucleus accumbens and striatum). GR205171 (40.0 mg/kg, s.c.), did not affect 5-HT levels in any structure. It did not modify DA levels in nucleus accumbens, and slightly decreased DA levels in striatum. However, it enhanced levels of DA in the BLA and those of NA in BLA and dorsal hippocampus. In the presence of GR205171, the influence of citalopram upon 5-HT was potentiated, but changes in DA and NA were not modified.

Figure 2
figure 2

Enhancement by GR205171 of the influence of citalopram upon dialysis levels of serotonin in dorsal hippocampus and basolateral amygdala of freely moving rats. Serotonin (5-HT), dopamine (DA) and noradrenaline (NA) levels are expressed relative to basal values defined as 100%. These were 1.42±0.10 and 0.79±0.14 pg per 20 μl for NA and 5-HT in the dorsal hippocampus, respectively, and 0.91±0.05, 0.47±0.05, and 0.62±0.10 pg per 20 μl for NA, DA, and 5-HT in the BLA, respectively. Data are means±SEM; N=5–8 per value. Arrows denote the injection of drugs. ANOVA as follows. Dorsal hippocampus, 5-HT: citalopram, F(1, 11)=33.3, P<0.01; GR205171, F(1, 13)=2.1, P>0.05 and interaction, F(1, 8)=7.5, P<0.05 and NA: citalopram, F(1, 11)=4.5, P>0.05; GR205171, F(1, 13)=7.6, P<0.05 and interaction, F(1, 8)=5.1, P>0.05. BLA, 5-HT: citalopram, F(1, 9)=7.1, P<0.05; GR205171, F(1, 9)=0.1, P>0.05 and interaction, F(1, 11)=5.6, P<0.05; DA: citalopram, F(1, 9)=10.5, P<0.05; GR205171, F(1, 8)=7.5, P<0.05 and interaction, F(1, 11)=4.1, P>0.05 and NA: citalopram, F(1, 9)=7.2, P<0.05; GR205171, F(1, 9)=16.2, P<0.01 and interaction, F(1, 11)=3.5, P>0.05. Significance (P<0.05) of drug-treated groups vs vehicle-treated groups is indicated by #, and significance (P<0.05) of GR205171/citalopram-treated groups vs vehicle/citalopram-treated groups is indicated by asterisks.

Full size image
Table 2 Enhancement by GR205171 of the Influence of Citalopram Upon Dialysis Levels of Serotonin in Striatum and Nucleus Accumbens
Full size table

Attenuation by GR205171 of the Inhibitory Influence of Citalopram and Fluoxetine Upon the Firing Rate of Serotonergic Neurones

Citalopram suppressed the electrical activity of DRN-localized serotonergic neurones (Figure 3). Administered alone, GR205171 failed to modify firing rate. However, it dose-dependently (1.0–4.0 mg/kg, i.v.) and significantly displaced the dose–response curve for inhibition of serotonergic neurones by citalopram to the right. The effects of citalopram alone and of citalopram in the presence of GR205171 were both blocked by the selective 5-HT1A receptor antagonist, WAY100635 (0.1 mg/kg, i.v.) (Newman-Tancredi et al, 1998). In distinction to GR205171, its less active distomer, GR226206 (4.0 mg/kg, i.v.), failed to affect the influence of citalopram upon DRN firing: vehicle+citalopram, inhibitory dose50 (ID50) in μg/kg, i.v=66.4±8.0 vs GR226206+citalopram, ID50=71.0±9.2, P>0.05. By analogy to citalopram, fluoxetine blocked the activity of serotonergic perikarya. GR205171 dose-dependently (1.0–4.0 mg/kg, i.v.) displaced the dose–response curve of fluoxetine to the right. The effects of fluoxetine alone and of fluoxetine plus GR205171 were blocked by WAY100635. Further, GR226206 (4.0 mg/kg, i.v.) failed to affect the influence of fluoxetine: vehicle+fluoxetine, ID50 in μg/kg, i.v=280.5±45.6 vs GR226206+fluoxetine, ID50=216.7±25.6, P>0.05.

Figure 3
figure 3

Attenuation by GR205171 of the inhibitory influence of citalopram and fluoxetine upon the firing rate of serotonergic neurones in anesthetized rats. VEH=vehicle; WAY=WAY100635. The influence of drugs is expressed as the percentage change from baseline defined as 0%. Baseline firing rate, 1.2±0.2 Hz. Data are means±SEMs; N=5–9 per value. Two-way ANOVA as follows. VEH+citalopram vs GR205171 (1)+citalopram, F(8, 112)=1.8, P>0.05; VEH+citalopram vs GR205171 (2)+citalopram, F(8, 96)=10.5, P<0.01; VEH+citalopram vs GR205171 (4)+citalopram, F(8, 112)=8.7, P<0.01; VEH+fluoxetine vs GR205171 (1)+fluoxetine, F(8, 72)=0.3, P>0.05; VEH+fluoxetine vs GR205171 (2)+fluoxetine, F(10, 100)=4.5, P<0.01; VEH+fluoxetine vs GR205171 (4)+fluoxetine, F(10, 100)=10.9, P<0.01. Asterisks indicate significance (P<0.05) of GR205171-treated groups vs vehicle-treated group.

Full size image

Influence of GR205171 Upon Inhibition of Serotonergic Neurones by 5-HT1A Agonists

The 5-HT1A partial agonist, S15535 (Gobert et al, 1995; Millan et al, 1997a, 1997b; Lejeune and Millan, 2000), dose dependently reduced the firing rate of DRN neurones (Figure 4). Its dose–response curve was shifted to the right by GR205171 (4.0 mg/kg, i.v.). This effect was expressed stereospecifically in that GR226206 (4.0 mg/kg, i.v.) failed to influence the decrease in DRN firing rate induced by S15535. In contrast, GR205171 (4.0) did not modify the inhibitory influence of the weak 5-HT1A partial agonist, (−)-pindolol (Newman-Tancredi et al, 1998), nor of the 5-HT1A agonists, 8-OH-DPAT and buspirone (Gobert et al, 1995; Lejeune et al, 1997; Newman-Tancredi et al, 1998), upon DRN firing rate. Actions of all 5-HT1A agents were antagonized by WAY100635 (0.1 mg/kg, i.v.).

Figure 4
figure 4

Influence of GR205171 on the inhibition of serotonergic neurones by the 5-HT1A agonists, S15535, (−)-pindolol, 8-OH-DPAT and buspirone, in anesthetized rats. VEH, vehicle; WAY, WAY100635. Data are means±SEMs; N=5–7 per value. Two-way ANOVA as follows. Vehicle+S15535 vs GR205171 (4)+S15535, F(8, 104)=8.6, P<0.01; vehicle+S15535 vs GR226206 (4)+S15535, F(8, 72)=1.0, P>0.05; vehicle + (−)-pindolol vs GR205171 (4) + (−)-pindolol, F(6, 42)=0.6, P>0.05; vehicle+8-OH-DPAT vs GR205171 (4)+8-OH-DPAT, F(6, 72)=1.3, P>0.05; vehicle+buspirone vs GR205171 (4)+buspirone, F(7, 63)=0.4, P>0.05. Asterisk indicates significance (P<0.05) of GR205171-treated groups vs vehicle-treated groups.

Full size image

Influence of GR205171 Upon the Modulation of FCX Levels of Monoamines By 5-HT1A Agonists

S15535 decreased levels of 5-HT in the FCX of freely-moving rats, and increased levels of DA and NA (Figure 5). The inhibitory influence of S15535 upon 5-HT levels was not affected by GR205171. Interestingly, coadministration of GR205171 and S15535 was accompanied by a more pronounced elevation in levels of DA and NA than for either drug alone. 8-OH-DPAT also decreased 5-HT levels in FCX while increasing the levels of DA and NA. In the presence of GR205171, the reduction in 5-HT levels provoked by 8-OH-DPAT was unaffected, whereas the elevation in DA and NA was facilitated.

Figure 5
figure 5

Influence of GR205171 upon the modulation of frontocortical levels of monoamines by the 5-HT1A agonists, S15535 and 8-OH-DPAT, in freely moving rats. Serotonin (5-HT), dopamine (DA), and noradrenaline (NA) levels are expressed relative to basal values defined as 100% (see legend of Figure 1). Data are means±SEMs; N=5–9 per value. Arrows denote the injection of drugs. ANOVA as follows. S15535, 5-HT: S15535, F(1, 11)=50.8, P<0.01; GR205171, F(1, 15)=0.2, P>0.05 and interaction, F(1, 8)=3.0, P>0.05; DA: S15535, F(1, 12)=49.3, P<0.01; GR205171, (1, 16)=31.3, P<0.01 and interaction, F(1, 8)=12.7, P<0.01 and NA: S15535, F(1, 12)=59.8, P<0.01; GR205171, (1, 16)=29.3, P<0.01 and interaction, F(1, 8)=30.5, P<0.01. 8-OH-DPAT, 5-HT: 8-OH-DPAT, F(1, 12)=54.5, P<0.01, GR205171, F(1, 15)=0.2, P>0.05 and interaction, F(1, 10)=2.2, P>0.05; DA: 8-OH-DPAT, F(1, 13)=89.2, P<0.01; GR205171, (1, 16)=31.3, P<0.01 and interaction, F(1, 11)=2.6, P>0.05 and NA: 8-OH-DPAT, F(1, 13)=46.3, P<0.01; GR205171, (1, 16)=29.3, P<0.01 and interaction, F(1, 11)=14.8, P<0.01. Significance (P<0.05) of drug-treated groups vs vehicle-treated groups is indicated by #, and the significance (P<0.05) of GR205171/8-OH-DPAT or GR205171/S15535-treated groups vs vehicle/8-OH-DPAT- or vehicle/S15535-treated groups is indicated by asterisks.

Full size image

Enhancement by GR205171 of the Antidepressant Action of Citalopram in the Forced Swim Procedure in Mice

In the mice FS test, citalopram decreased the duration of immobility, whereas GR205171 had no effect up to 40.0 mg/kg, i.p. (Figure 6). In the presence of GR205171 (40.0), however, a subactive dose of citalopram (2.5) significantly decreased immobility. This effect was not observed with GR226206 (40.0), which was inactive alone. Lower doses (2.5 and 10.0) of GR205171 did not influence the effect of citalopram (2.5) upon immobility (s): vehicle+citalopram, 154.7±11.1 (N=20) vs GR205171 (2.5)+citalopram (2.5), 145.0±19.7 (N=8), P>0.05 and vs GR205171 (10.0)+citalopram (2.5), 128.6±18.2 (N=12), P>0.05.

Figure 6
figure 6

Enhancement by GR205171 of the antidepressant action of citalopram in the forced swim procedure in mice. VEH, vehicle and CITAL, citalopram. (a) Dose-dependent reduction of immobility by citalopram as compared to GR205171; (b) stereospecific enhancement by GR205171 as compared to its active distomer, GR226206, of the action of citalopram. Data are means±SEMs (N=7–20). (a) One-way ANOVA as follows: citalopram, F(4, 30)=5.5, P<0.01; GR205171, F(3, 20)=0.4, P>0.05. Asterisks indicate significance of differences to respective vehicle values in Dunnett's test. (b) Two-way ANOVA as follows: citalopram, F(1, 75)=5.1, P<0.05; GR205171, F(1, 75)=11.6, P<0.01 and interaction, F(1, 75)=2.2, P>0.05; GR226206, F(1, 55)=0.5, P>0.05 and interaction, F(1, 55)=0.2, P>0.05. The asterisk indicates significance of differences in Dunnett's test vs vehicle/vehicle values. *P<0.05.

Full size image

Lack of Enhancement by GR205171 of the Facilitatory Influence of Citalopram Upon Locomotor Activity in Mice

Citalopram (10.0 mg/kg, s.c.) increased locomotor activity in mice (Table 3). This action was dose-dependently (2.5–40.0 mg/kg, i.p.) attenuated by GR205171. Locomotor activity was not significantly modified by GR205171 alone.

Table 3 Influence of GR205171 Upon the Induction of Locomotion in Mice by Citalopram
Full size table

Enhancement by GR205171 of the Actions of Citalopram in the Ultrasonic Vocalization Procedure in Rats

Citalopram dose-dependently decreased the duration of fear-induced USV in rats, though significance was obtained only for the highest dose tested (10.0 mg/kg, s.c.) (Figure 7). A similar profile was obtained with GR205171 (40.0 mg/kg, i.p.). GR226206 was inactive up to 40.0 mg/kg, i.p. In the presence of GR205 171 (10.0, i.p.), a subactive dose of citalopram (5.0, s.c.) significantly decreased USV, an effect not observed in the presence of GR226206 (40.0, i.p.).

Figure 7
figure 7

Enhancement by GR205171 of the actions of citalopram in the ultrasonic vocalization procedure in rats. VEH=vehicle and CITAL=citalopram. (a) Dose-dependent reduction of USV by CITAL and GR205171 as compared with its less active distomer, GR226206. (b, c) Stereospecific enhancement by GR205171 of the action of CITAL. Data are means±SEMs; N=5–15 per value. (a) ANOVA as follows: CITAL, F(4, 51)=3.0, P<0.05; GR205171, F(4, 47)=1.8, P>0.05 and GR226206, F(3, 21)=2.3, P>0.05. Asterisks indicate significance of differences to respective VEH values in Dunnett's test. (b, c) ANOVA as follows. (b) CITAL, F(1, 29)=3.8, P>0.05; GR205171, F(1, 29)=18.9, P<0.01 and interaction, F(1, 29)=1.8, P<0.05. (c) CITAL, F(1, 14)=0.3, P<0.05, GR226206, F(1, 14)=0.8, P>0.05 and interaction, F(1, 14)=0.1, P>0.05. The asterisk indicates significance of differences in Newman–Keuls test vs VEH/VEH values. *P<0.05.

Full size image

Additive Inhibition by GR205171 and Citalopram of Light-Induced Phase Advances in Circadian Wheel Running Rhythms in Hamsters

A 10-min light pulse advanced circadian rhythms by 1.6±0.1 h following vehicle (Table 4). Citalopram (5 mg/kg, i.p.) and GR205171 (5.0 mg/kg, i.p) alone did not influence light-induced phase advances. In contrast, combination of these doses significantly inhibited the effect of light (−47%). Further, citalopram (10.0 mg/kg) and GR205171 (20.0 mg/kg) significantly inhibited the effect of light by −40 and −70%, respectively. When injected simultaneously, the combination inhibited the effect of light by 60%.

Table 4 Additive Influence of Low Doses of GR205171 and Citalopram Upon the Phase Advances in Circadian Wheel Running Rhythms Elicited by Light in Hamsters
Full size table

Blockade by GR205171 of the Facilitation by Fluoxetine of Fear-Induced Foot-Tapping in Gerbils

In animals without fear conditioning (ie not submitted to shocks during the acquisition session), fluoxetine (40.0 mg/kg, i.p.) did not induce foot tapping (Figure 8). In contrast, in fear-conditioned gerbils (submitted to shocks), fluoxetine dose-dependently induced foot tapping at doses of 20.0 and 40.0 mg/kg, i.p. The dose of 40.0 mg/kg, i.p. was selected for antagonist studies. GR205171 (0.16 mg/kg, i.p.) blocked the effect of fluoxetine (40.0 mg/kg, i.p.) upon fear-conditioned foot tapping whereas GR226206 was inactive.

Figure 8
figure 8

Blockade by GR205171 of the facilitation by fluoxetine of fear-induced foot tapping in gerbils. VEH=vehicle and FLUOX=fluoxetine. (a) Induction of foot tapping by FLUOX and (b) stereospecific blockade by GR205171 of the action of FLUOX. Data are means±SEMs; N=4–10 per values. ANOVA as follows. (a) FLUOX, F(3, 15)=6.6, P<0.05. Asterisks indicate significance of differences in Dunnett's test between fear-conditioning/FLUOX and fear-conditioning /VEH values. (b) Interaction, F(2, 35)=4.4, P<0.05. The closed asterisk indicates significance of differences in Dunnett's test between VEH/FLUOX and GR205171/FLUOX values and open asterisks, between fluoxetine and vehicle values. *P<0.05.

Full size image

Blockade by GR205171 of the Anxiogenic Actions of Citalopram in the Social Interaction Test in Rats and Gerbils

Citalopram reduced the duration of SI between unfamiliar rats and gerbils exposed to a novel environment (Figure 9). In contrast, GR205171 alone did not modify the duration of SI in either species. However, GR205171 (10.0 mg/kg, i.p. in rat, and 0.63 in gerbils) antagonized the anxiogenic effect of citalopram. Conversely, GR226206 was inactive.

Figure 9
figure 9

Blockade by GR205171 of the anxiogenic actions of citalopram in the social interaction test in rats and gerbils. VEH=vehicle and CITAL=citalopram. (a, b) Dose-dependent reduction of SI by CITAL as compared with GR205171 in rats and gerbils, respectively, and (c–f) stereospecific blockade by GR205171 of the action of CITAL. Data are means±SEMs; N=5–18 per value. (a, b) ANOVA as follows: CITAL in rats, F(4, 38)=10.6, P<0.01, GR205171 in rats, F(3, 22)=0.4, P>0.05, CITAL in gerbils, F(4, 26)=13.9, P<0.01 and GR205171 in gerbils, F(4, 33)=1.6, P>0.05. Asterisks indicate significance of differences to respective vehicle values in Dunnett's test. (c–f) ANOVA as follows. (c) CITAL, F(1, 18)=7.1, P<0.05; GR205171, F(1, 18)=7.3, P<0.05 and interaction, F(1, 18)=1.6, P>0.05. (d) CITAL, F(1, 17)=5.3, P<0.05; GR205171, F(1, 17)=27.2, P<0.01 and interaction, F(1, 17)=8.9, P<0.01. (e) CITAL, F(1, 25)=15.4, P<0.01; GR226206, F(1, 25)=0.2, P>0.05 and interaction, F(1, 25)=0.3, P>0.05. (f) CITAL, F(1, 19)=13.6, P<0.01; GR226206, F(1, 19)=0.2, P>0.05 and interaction, F(1, 19)=0.4, P>0.05. Closed asterisks indicate significance of differences in Newman–Keuls test between VEH/CITAL and VEH/VEH values and open asterisks, between NK1 antagonist/CITAL and VEH/CITAL values. *P<0.05.

Full size image

Influence of RP67580 Upon Neurochemical and Antidepressant Actions of Citalopram

The NK1 antagonist, RP67580 (40.0 mg/kg, i.p.), potentiated citalopram-induced increases in 5-HT levels in FCX. It also mimicked GR205171 in elevating levels of DA and NA alone, but there was no difference between RP67580/citalopram vs RP67580/vehicle and vehicle/citalopram values (Table 1). In the FS test, the action of RP67580 (20.0 mg/kg, i.p.) and citalopram (2.5 mg/kg, s.c.) together was greater (P<0.05) than RP67580 or citalopram alone: vehicle/vehicle (s), 178.7±10.6; citalopram/vehicle, 168.3±9.0; vehicle/RP67580, 168.3±9.0 and citalopram/ RP67580, 122.4±17.0.

DISCUSSION

Potentiation by GR205171 of the Influence of Citalopram Upon 5-HT Levels

Several arguments underpin the specificity of the facilitation by GR205171 of increases in levels of 5-HT elicited by citalopram in FCX. First, GR205171 is a highly selective antagonist of NK1 receptors and expressed its actions across a dose range identical to that exerting other cerebral actions in rodents (Gardner et al, 1996; Millan et al, 2001b; Rupniak et al, 2001; Lejeune et al, 2002; Brocco et al, 2008). Further, its actions were stereospecific in as much as its less active isomer, GR226206, was ineffective. Second, the influence of GR205171 was reproduced with another selective NK1 antagonist, RP67580 (Garret et al, 1991; Gobbi et al, 2007), and GR205171 also stereospecifically augmented the influence of fluoxetine upon 5-HT levels in FCX. Third, GR205171 potentiates paroxetine-induced increases in extracellular levels of 5-HT in the FCX of mice (Guiard et al, 2004). Fourth, herein, GR205171 also enhanced the action of citalopram in four further structures receiving serotonergic input from the DRN and controlling affect (Millan, 2003, 2006): hippocampus, BLA, nucleus accumbens, and striatum. Fifth, the enhancement of SSRI-induced increases in FCX levels of 5-HT by GR205171 was paralleled by its suppression of their inhibition of DRN-localized serotonergic neurones. Finally, pharmacokinetic interactions are unlikely to be involved as: (1) two different NK1 antagonists facilitated the neurochemical effects of two different SSRIs; (2) compatible findings were acquired in electrophysiological studies using the i.v. route, a route which avoids first pass elimination in the liver; (3) certain responses to SSRIs were blocked by NK1 antagonists; (4) similar results were seen in mice using paroxetine (Chenu et al, 2006), and (5) genetic deletion of NK1 receptors likewise enhances the influence of SSRIs upon levels of 5-HT (Guiard et al, 2004, 2005).

Integration at the Level of the DRN: Significance of 5-HT1A Autoreceptors

NK1 antagonists are unlikely to directly affect actions of SSRIs as GR205171 and RP67580 have negligible affinities for 5-HT transporters (pKi values, <5.0), and do not modify the actions of SSRIs at synaptosomes in vitro (Lieb et al, 2005; Millan, MJ, unpublished observation). Antagonists at 5-HT1B autoreceptors on serotonergic terminals enhances increases in 5-HT levels elicited by SSRIs (Gobert et al, 1997). However, GR205171 and RP67580 have no affinity for 5-HT1B receptors, and this would not explain their abrogation of the inhibitory influence of SSRIs on serotonergic neurones in the DRN (Gardner et al, 1996; Millan et al, 2000). Rather, inhibitory 5-HT1A autoreceptors are localized on serotonergic perikarya, and 5-HT1A antagonists likewise enhance increases in FCX levels of 5-HT elicited by SSRIs (Gobert et al, 1997; Millan et al, 2000; Guilloux et al, 2006). Though GR205171 and RP67580 do not behave as 5-HT1A antagonists (pKi values, <5.0; Gardner et al, 1996; Figure 5), it is important to consider the significance of 5-HT1A autoreceptors.

In agreement with most studies of NK1 antagonists (Haddjeri and Blier, 2000; Conley et al, 2002; Lejeune et al, 2002; Guiard et al, 2005; Gobbi et al, 2007), GR205171 did not affect the spontaneous activity of DRN-localized serotonergic neurones. However, its reduction of the inhibitory influence of citalopram and fluoxetine upon DRN firing mimics the effects of chronic administration of NK1 antagonists (Haddjeri and Blier, 2001; Guiard et al, 2005), and of genetic elimination of NK1 receptors in mice (Froger et al, 2001; Santarelli et al, 2001; Gobbi et al, 2007). A subpopulation (some 30%) of 5-HT neurones in the DRN bear NK1 receptors (Froger et al, 2001; Ma and Bleasdale, 2002; Lacoste et al, 2006), raising the possibility that NK1 antagonists (via signalling crosstalk) disrupt the inhibitory influence of SSRI-engaged 5-HT1A autoreceptors upon DRN firing. Indeed, the inhibitory influence of 5-HT1A agonists on the DRN is blunted by chronic treatment with NK1 antagonists (Haddjeri and Blier, 2001; Guiard et al, 2005), and by genetic deletion of NK1 receptors (Froger et al, 2001; Santarelli et al, 2001). However, acute blockade of NK1 receptors does not generally desensitize 5-HT1A autoreceptors (Blier et al, 2004; Guiard et al, 2007; Gobbi et al, 2007), and GR205171 did not modify the influence of 5-HT1A agonists on DRN firing rate in vitro (Guiard et al, 2005). Correspondingly, the lack of influence of GR205171 upon the inhibition of DRN firing (Figure 4) by two 5-HT1A agonists, 8-OH-DPAT and buspirone (Lejeune et al, 1997), suggests that it does not markedly affect the sensitivity of 5-HT1A autoreceptors. Conversely, GR205171 attenuated the inhibitory influence of a lower efficacy (ca 30–40% vs 5-HT) 5-HT1A ligand, S15535 (Millan et al, 1997a, 1997b; Newman-Tancredi et al, 1998) upon DRN firing. Thus, this partial agonist may reveal a subtle alteration by NK1 receptor blockade of coupling at 5-HT1A autoreceptors. However, changes in firing rate do not invariably translate into alterations in release (Gobert et al, 1995; Millan and Gobert, 1999; Millan et al, 2000; Artigas et al, 2001; Gobbi et al, 2007), and the reduction by S15535 (and 8-OH-DPAT) of 5-HT release in FCX was not modified by GR205171. Moreover, GR205171 did not modify the influence on serotonergic neurones of (−)-pindolol, a low efficacy 5-HT1A ligand (ca 20%) which only submaximally decreases firing rate (Newman-Tancredi et al, 1998; Millan and Gobert, 1999; Guilloux et al, 2006). Recruitment of postsynaptic 5-HT1A receptors in FCX triggers long-loop, inhibitory feedback to serotonergic cell bodies in raphe nuclei (Haddjeri et al, 2000; Celada et al, 2001; Sharp et al, 2007). However, they are unlikely to be involved in the activations of GR205171 as the sensitivity of postsynaptic 5-HT1A receptors is unaffected by NK1 antagonists (Froger et al, 2001; Haddjeri and Blier, 2001; Santarelli et al, 2001). Further, though microinjection of substance P into the FCX suppressed 5-HT release, this action did not involve frontocortical 5-HT1A receptors (Guiard et al, 2007).

Possible Involvement of Gabaergic, Glutatamergic, and Adrenergic Mechanisms

NK1 receptors have been identified on GABAergic neurones surrounding 5-HT cell bodies in the DRN, and studies in the septum and cortex suggest that NK1 receptor antagonists may indirectly excite serotonergic neurones via a reduction of inhibitory GABAergic tone (Sloviter et al, 2001; Ma and Bleasdale, 2002; Stacey et al, 2002; Szeidemann et al, 1995; Ebner et al, 2008). This would enhance their responsiveness to SSRIs, by analogy to GABAB antagonists (Millan, 2006; Cremers et al, 2007). Glutamatergic terminals targeting serotonergic DRN neurones also bear NK1 receptors (Liu et al, 2002; Valentino et al, 2003), and their stimulation by intraraphe perfusion of substance P enhances local release of 5-HT via recruitment of AMPA receptors (Guiard et al, 2007). This results in a (delayed) decrease in cortical 5-HT release due to activation of 5-HT1A autoreceptors. That is, in line with the present observations, NK1 receptors are ‘upstream’ of 5-HT1A autoreceptors. Interestingly, by analogy to lateral septum (Ebner et al, 2008), frontocortical perfusion of substance P suppressed local release of 5-HT in mice, an effect blocked by GR205171 and absent in mice lacking NK1 receptors (Guiard et al, 2007). Thus, NK1 antagonists conceivably also enhance SSRI-induced increase in 5-HT levels by actions in the FCX and other structures innervated by the DRN. Finally, Gobbi et al (2007) found that functionally intact adrenergic terminals are essential for the facilitatory influence of NK1 receptor antagonists upon DRN firing. This is coherent with the excitatory impact of NK1 receptor blockade upon electrical activity of LC-derived adrenergic neurones (Millan et al, 2001b; Gobbi et al, 2007).

Influence of NK1 Receptor Blockade Upon NA and DA Levels

Elevations in levels of NA in FCX are seen in mice genetically deprived of NK1 receptors and are elicited by NK1 antagonists in parallel with an excitation of the LC (Millan et al, 2001b; Maubach et al, 2002; Blier et al, 2004; Ebner and Singewald, 2007; Gobbi et al, 2007). Correspondingly, GR205171 and RP67580 increased extracellular levels of NA in FCX, as well as in the BLA and hippocampus—structures likewise innervated by the LC (Millan et al, 2001b). The VTA possess a high density of NK1 sites (Lessard and Pickel, 2005), and electrical activity of VTA-localized dopaminergic neurons is accelerated by blockade of NK1 receptors, an effect accompanied by increases in DA levels in FCX (Lejeune et al, 2002). This finding was confirmed herein for GR205171, reproduced with RP67580 and extended to the BLA. Interestingly, neither agent affected DA levels in striatum or nucleus accumbens which are innervated by different subpopulations of dopaminergic cell bodies: that is, substantia nigra pars compacta (nigrostriatal) and paranigral-VTA (mesolimbic) vs parabrachial-VTA (mesocortical), respectively (Lejeune et al, 1997; Lejeune and Millan, 2000). These observations accord with the differential control of these clusters of dopaminergic neurons (Millan et al, 2000; Adell and Artigas, 2004; Alex and Pehek, 2007), including 5-HT1A receptor agonism which disinhibits VTA-derived mesocortical dopaminergic and LC-derived adrenergic pathways via GABAergic interneurones (Millan et al, 2000; Invernizzi et al, 2007). Interestingly, 8-OH-DPAT and S15535-induced elevations in FCX levels of DA and NA (Millan et al, 1997b; Gobert et al, 1995, 1999), were additive to those of GR205171. This indicates that 5-HT1A agonism and NK1 receptor blockade offer complementary strategies for improvement of affect (Lucki et al, 1994; Millan, 2003, 2006; Czeh et al, 2006). GR205171 and fluoxetine likewise additively augmented dialysate levels of NA and DA in FCX, suggesting complementary mood-elevating properties independent of 5-HT (Millan et al, 2000). Accordingly, association of NK1 receptor blockade with high doses of SSRIs, with the NA reuptake inhibitor, reboxetine, or with 5-HT/NA reuptake inhibitors like venlafaxine (Morilak and Frazer, 2004; Millan, 2006), may lead to more pronounced elevations in NA levels in FCX and improved antidepressant efficacy.

Enhancement of the Effects of Citalopram in a Forced Swim Procedure

Addition of GR205171 to a subactive dose of citalopram led to a significant antidepressant effect, an action exerted stereoselectively vs GR226206. These observations support the finding of a recent study of GR205171 and paroxetine by Chenu et al (2006)—though stereospecificity was not demonstrated. Apart from the hippocampus, the septum is implicated in antidepressant actions of SSRIs (Sheehan et al, 2004; Millan, 2006), and NK1 antagonists may enhance 5-HT release in this region via inhibition of GABAergic interneurons (Szeidemann et al, 1995; Millan, 2006; Ebner et al, 2008). As GR205171 blunted the increase in locomotor activity elicited by citalopram herein (Table 3), potentiation of the actions of citalopram in the FS procedure is unlikely to reflect stimulation of motor behavior (Redrobe and Bourin, 1998; Brocco et al, 2006). Likewise, augmentation of paroxetine-induced decreases in immobility in the FS test occurred at a dose of GR205171 which did not increase its motor actions (Chenu et al, 2006).

Enhancement of the Action of Citalopram in the Stress-Related Ultrasonic Vocalization Test

Exposure of rats to stress elicits USVs at a characteristic frequency of 22 kHz (De Vry et al, 1993; Sánchez, 2003). This model is sensitive to SSRIs and, accordingly, citalopram reduced USVs (Molewijk et al, 1996; Sanchez and Meier, 1997; Millan et al, 2001a; Sánchez, 2003). Curiously, actions of NK1 antagonists have not, as yet been documented. In fact, periaqueductal grey administration of substance P reduced USVs elicited by isolation in rats (Bassi et al, 2007), but NK1 antagonists suppress isolation-induced (nonultrasonic) vocalizations in juveniles via actions in the amygdala (Molewijk et al, 1996; Rupniak et al, 2000; Brocco et al, 2008). This observation is coherent with the present finding that GR205171 attenuates stress-induced USVs. Moreover, though the difference in potency between GR205171 and GR226206 was, for unclear reasons, less marked in this procedure than for other paradigms, only GR205171 potentiated the inhibitory influence of citalopram on USVs. These data suggest that NK1 receptor blockade may enhance the abrogation of stress-elicited behaviors by SSRIs.

Additive Influence on Circadian Rhythms

A normalization of perturbed diurnal scheduling of behavior is implicated in the therapeutic actions of certain antidepressants (Duncan, 1996; Millan, 2006). Light-induced phase shifts of activity rhythms in hamsters are blunted by several classes of antidepressant, including SSRIs (Millan, 2006; Gannon and Millan 2007). They likely act in the suprachiasmatic nucleus, which contains a high density of 5-HT transporters on serotonergic terminals originating in the median raphe nucleus (Legutko and Gannon, 2001). Intriguingly, a similar suppression of light-induced circadian phase-shifts is seen with NK1 antagonists, which act upstream in the raphe to intensify serotonergic input to the suprachiasmatic nucleus (Valentino et al, 2003; Gannon and Millan, 2005). Consistent with contrasting mechanisms of action, subthreshold doses of citalopram and GR205171 additively prevented light-triggered phase-shifts. Accordingly, combined NK1 receptor blockade and suppression of 5-HT reuptake may mutually restore circadian rhythmicity in certain depressed patients.

Blockade of the Anxiogenic Effects of Citalopram and Fluoxetine

SSRIs elicits anxiogenic effects in the SI procedure in rats, an observation extended herein to gerbils (Dekeyne et al, 2000; Millan, 2003). By contrast, NK1 receptor antagonists either fail to affect, or enhance, SI, probably dependent upon the degree of substance P release (File, 1997; Dekeyne et al, 2000; Varty et al, 2002; Millan, 2003; Brocco et al, 2008; Ebner et al, 2008). GR205171 stereospecifically abolished the reduction in SI elicited by citalopram in rats, and similar results were acquired in gerbils, a species in which NK1 receptors are potently blocked by GR205171 (Griffante et al, 2006; Engberg et al, 2007). Serotonin2C receptors in the hippocampus, amygdala and/or locus coeruleus mediate anxiogenic actions of SSRIs (Dekeyne et al, 2000; Millan 2003; Campbell and Merchant, 2003; Salchner and Singewald, 2006). However, GR205171 does not affect the reduction in SI provoked by 5-HT2C agonists (Dekeyne A, unpub. obs.), suggesting an action ‘upstream’ of 5-HT2C sites. These observations are supported by a further paradigm based on enhancement of fear-conditioned anxiety by SSRIs. Association of a mild aversive stimulus with fluoxetine triggers a characteristic foot-thumping response in gerbils (Millan, 2003; Rupniak et al, 2003; Burghardt et al, 2007; Brocco et al, 2008). By analogy to other NK1 antagonists (op. cit.), this behavior was stereospecifically abrogated by GR205171, supporting the contention that acute anxiogenic effects of SSRIs are tempered by NK1receptor blockade. SSRI-induced nervousness reduces compliance, triggers early abandonment of therapy and is associated with a long delay to efficacy (Millan, 2003, 2006; Morilak and Frazer, 2004). The present observations suggest that NK1 receptor antagonists, which show anxiolytic properties in patients with major depression or social phobia (Kramer et al, 1998, 2004; Michelgard et al, 2007), may alleviate anxiogenic effects of SSRIs at the onset of therapy.

General Discussion

First, the present data show that the antidepressant efficacy of SSRIs is enhanced by blockade of NK1 receptors, yet that their acute anxiogenic actions are abrogated. Thus, the present observation support the association of SSRIs and selective NK1 antagonists and the development of mixed NK1 antagonists/5-HT reuptake inhibitors for treatment of depression and anxiety (Ryckmans et al, 2002; Chenu et al, 2006; Millan, 2006; Brocco et al, 2008). Second, association of GR205171 with fluoxetine led to additive increases in levels of NA and DA suggesting that interactions between NK1 antagonists and antidepressants may also involve dopaminergic and adrenergic mechanisms. GR205171 did not potentiate antidepressant actions of the NA reuptake inhibitor, desipramine (Chenu et al, 2006), and desipramine did not more markedly increase NA levels in NK1 knock-out mice (Herpfer et al, 2005). Nonetheless, it would be of interest to examine the influence of GR205171 upon actions of mixed 5-HT/NA reuptake inhibitors like venlafaxine, and the atypical agent, mirtazapine (Millan et al, 2000; Morilak and Frazer, 2004; Millan, 2006). This argument is underpinned by the marked elevation in FCX levels of DA and NA elicited by GR205171 together with 8-OH-DPAT or S15535, 5-HT1A receptor agonists possessing antidepressant and anxiolytic properties (Lucki et al, 1994; Millan et al, 1997a, 1997b; Millan, 2006). Third, NK2 and NK3 receptors influence mood and modulate monoaminergic pathways, possibly in interaction with NK1 receptors (Steinberg et al, 2001; Bert et al, 2002; Léger et al, 2002; Spooren et al, 2005). Accordingly, it would be interesting to examine how blockade of NK2 and NK3 receptors affects the functional actions of antidepressants.

CONCLUSIONS

The present observations strongly suggest that combining NK1 receptor blockade with suppression of 5-HT reuptake may yield benefits in the treatment of anxious and depressed states relative to selective NK1 antagonists and SSRIs.