|Home | About | Journals | Submit | Contact Us | Français|
Preclinical data have shown that addition of the selective norepinephrine transporter (NET) inhibitor reboxetine increases the antipsychotic-like effect of the D2/3 antagonist raclopride and, in parallel, enhances cortical dopamine output. Subsequent clinical results suggested that adding reboxetine to stable treatments with various antipsychotic drugs (APDs) may improve positive, negative and depressive symptoms in schizophrenia. In this study, we investigated in rats the effects of adding reboxetine to the second-generation APD olanzapine on: (i) antipsychotic efficacy, using the conditioned avoidance response (CAR) test, (ii) extrapyramidal side effect (EPS) liability, using a catalepsy test, (iii) dopamine efflux in the medial prefrontal cortex and the nucleus accumbens, using in vivo microdialysis in freely moving animals and (iv) cortical N-methyl--aspartate (NMDA) receptor-mediated transmission, using intracellular electrophysiological recording in vitro. Reboxetine (6mg/kg) enhanced the suppression of CAR induced by a suboptimal dose (1.25mg/kg), but not an optimal (2.5mg/kg) dose of olanzapine without any concomitant catalepsy. Addition of reboxetine to the low dose of olanzapine also markedly increased cortical dopamine outflow and facilitated prefrontal NMDA receptor-mediated transmission. Our data suggest that adjunctive treatment with a NET inhibitor may enhance the therapeutic effect of low-dose olanzapine in schizophrenia without increasing EPS liability and add an antidepressant action, thus in principle allowing for a dose reduction of olanzapine with a concomitant reduction of dose-related side effects, such as EPS and weight gain.
Previous studies have shown the critical importance of optimal dopamine transmission in the prefrontal cortex for certain aspects of cognition, such as working memory, and indicate that dysfunctional dopamine D1 receptors may contribute to cognitive impairments in schizophrenia (see Abi-Dargham et al, 2002; Goldman-Rakic et al, 2004; Abi-Dargham and Moore, 2003). Moreover, glutamate, the excitatory transmitter in cortical pyramidal cells, is involved in higher mental functions, such as cognition, memory and learning, and a glutamatergic synaptic hypofunction in schizophrenia seems indicated (see Coyle et al, 2003), a notion recently supported by the demonstration of a reduced prefrontal expression of the N-methyl--aspartate (NMDA) receptor subunits NR1, NR2A and NR2C in schizophrenic patients (Beneyto and Meador-Woodruff, 2008). In addition, several other transmitter systems are in all probability involved in the control of cognitive functions (see Briand et al, 2007), as well as in the pathophysiology of schizophrenia, such as the catecholaminergic systems, because elevation of central noradrenergic and secondarily dopaminergic activity, for example, by administration of α2-adenoreceptor antagonists, has been shown to improve working memory, as well as attention, learning and memory in rodents (Lapiz and Morilak, 2006; Sara and Devauges, 1989).
Clozapine, the prototypical atypical antipsychotic drug (APD), has been found to be superior to first (FGAs) and other second-generation APDs (SGAs) in treatment-resistant schizophrenia (see Swartz et al, 2008; Kane et al, 1988; Lewis et al, 2006; McEvoy et al, 2006; Taylor and Duncan-McConnell, 2000), as well as in reducing suicidality in schizoaffective disorders and schizophrenia (see Hennen and Baldessarini, 2005; Meltzer et al, 2003). Thus, the clinical efficacy of clozapine seems superior to both FGAs and most, if not all SGAs (Davis et al, 2003; Leucht et al, 2009). Unfortunately, the use of clozapine has been associated with a risk of agranulocytosis, which limits its use. Besides its relatively low affinity for the dopamine D2 receptor, clozapine possesses high affinity for a variety of other receptors, and several serotonergic receptors, for example, 5-HT2A/C, 5-HT1A, 5-HT6, as well as α2-adrenoceptors have been suggested to contribute to its superior efficacy (see, eg, Meltzer and Huang, 2008; Svensson, 2003a; Marcus et al, 2005). The SGA olanzapine is structurally similar to clozapine and shows higher affinity for several serotonergic receptors than for the D2 receptor (Schotte et al, 1996). However, although clozapine possesses a high affinity for α2-adrenoceptors, olanzapine has a low affinity for these receptors (Shahid et al, 2009). Both clozapine and olanzapine elevate cortical dopamine and norepinephrine outflow to a greater extent than FGAs such as haloperidol (Devoto and Flore, 2006; Li et al, 1998; Moghaddam and Bunney, 1990; Nomikos et al, 1994; Westerink et al, 1998). In contrast to FGAs, both clozapine and olanzapine also markedly facilitate cortical NMDA receptor-mediated transmission (Jardemark et al, 2002; Ninan et al, 2003), an effect that may contribute to improve negative symptoms and cognitive dysfunction in schizophrenia (see eg Stone et al, 2007).
The antidepressant drug reboxetine, a selective norepinephrine reuptake inhibitor (NRI), has been claimed to improve drive and motivation, as well as social and cognitive functioning in depressed patients (Ferguson et al, 2003; Kasper, 1999; Montgomery, 1997; Schatzberg, 2000). A cognitive-enhancing effect of the selective NRI atomoxetine has also been observed in healthy volunteers (Chamberlain et al, 2006). In rodents, addition of reboxetine to the D2/3 antagonist raclopride has been shown to significantly enhance the antipsychotic-like effect of the FGA raclopride without increasing its extrapyramidal side effect (EPS) liability, as well as to markedly increase cortical dopamine output (Linner et al, 2002). Moreover, reboxetine has recently been found to enhance the effect of raclopride also on cortical NMDA receptor-mediated transmission (Jardemark et al, unpublished observation). Thus, preclinical results suggest that concomitant norepinephrine transporter (NET) inhibition may provide means to augment the efficacy of FGAs in schizophrenia.
In this study, we have examined in rats whether concomitant NET inhibition by means of reboxetine might enhance the efficacy of olanzapine, and potentially mimic some of the preclinical effects of clozapine. The antipsychotic effect was assessed by using the conditioned avoidance response (CAR) paradigm, a preclinical test with high predictive validity for clinical antipsychotic effect (Arnt, 1982; Wadenberg and Hicks, 1999). EPS liability was examined by using a catalepsy test. Dopamine outflow in the medial prefrontal cortex (mPFC) and the nucleus accumbens (NAC) was measured by in vivo microdialysis in freely moving animals. Finally, the effects on cortical NMDA receptor-mediated transmission were examined using intracellular electrophysiological recording in pyramidal cells in vitro.
Adult male Wistar rats (~250g on arrival) were used for behavioral and microdialysis experiments, whereas male Sprague–Dawley rats (~70g on arrival; B&K Universal, Sollentuna, Sweden) were used for in vitro electrophysiological experiments. The animals were housed under standard laboratory conditions with a temperature of 21°C and relative humidity of 55–65%, with food (R34, Ewos, Södertälje, Sweden) and water ad libitum. For the behavioral tests, the animals were kept on a reversed 12/12h light/dark cycle (lights off at 0600h), whereas for the other experiments, animals were maintained on a 12/12h light/dark cycle (lights on at 0600h). The animals were acclimatized for at least 5 days before experiments. Experiments were approved by, and conducted in accordance with, the local animal ethics committee, Stockholm North and the Karolinska Institutet, Sweden.
The CAR methodology used has previously been described in detail (Linner et al, 2002; see Wadenberg and Hicks, 1999). In brief, rats were trained and tested in conventional shuttle boxes (530 × 250 × 225mm), divided into two compartments of equal size by a partition with an opening (Salmi et al, 1994). On presentation of the 80dB white noise (Lafayette Instruments, Lafayette, IN) conditioned stimulus (CS), the rats had 10s to move from one compartment of the shuttle box into the other. If the rat remained in the same compartment for more than 10s, an intermittent electric shock (intershock interval 2.5s, shock duration 0.5s) of ~0.4mA, that is, the unconditioned stimulus (UCS), was presented to the grid floor until an escape was performed. If the rat did not respond within 50s, the trial was terminated, that is, escape failure. The position of the rat was automatically transferred to the computer and avoidance (response to CS within 10s), escape (response to CS+UCS), escape failure (failure to respond within 50s) and intertrial crosses were recorded. The animals were trained for 5 days, each session consisted of ~20 trials randomly distributed over 15min. Only animals reliably performing >85% avoidance were included in the study. Experiments were preceded by a pretest and experiment sessions, lasting 10min, were conducted at 20, 90 and 240min after last injection. Experimental days were separated by at least two non-experimental days. The animals were tested in a counterbalanced change-over design serving as their own controls (Li, 1964).
Catalepsy measurements were performed in a dimly lit room at 30, 90 and 120min after drug administration (see, eg, Linner et al, 2002). The rat was placed on an inclined grid (60°) and allowed 30s of adaptation before observations started. These were performed for a maximum of 2.5min and the time until the rat initiated a movement by one of its paws was recorded. Catalepsy was scored from 0 to 5, according to the time (square root transformation) the rat remained immobile (min): 0=0.00–0.08; 1=0.09–0.35; 2=0.36–0.80; 3=0.81–1.42; 4=1.43–2.24; 52.25 (Ahlenius and Hillegaart, 1986). The different treatments were blinded to the observer.
Our procedures for microdialysis have previously been described (Frånberg et al, 2008). In brief, rats were anesthetized with a cocktail of Hypnorm (0.315mg/ml fentanyl citrate and 10mg/ml fluanisone; Janssen-Cilag, Saunderton, UK) and Dormicum (5mg/ml midazolam; Roche AB, Stockholm, Sweden) diluted in distilled water (1:1:2; 5ml/kg, intraperitoneal (i.p.)) mounted in a stereotaxic frame, and implanted with dialysis probes in the mPFC or NAC (anteroposterior (mm): +3.2, +1.6; mediolateral: −0.6, −1.4; dorsoventral: −5.2, −8.2), respectively, relative to bregma and dural surface (Paxinos and Watson, 1998). Dialysis occurred through a semi-permeable membrane (Filtral AN69, Hospal Industrie, Meyzieu, France) with an active surface length of 4mm (mPFC) or 2mm (NAC). Dialysis experiments were conducted ~48h after surgery in awake freely moving rats. The dialysis probe was perfused with a physiological perfusion solution (in mM: 147 NaCl, 3.0 KCl, 1.3 CaCl2, 1.0 MgCl2, 1.0 NaHPO4, pH 7.4) at a rate of 2.5μl/min, set by a microinfusion pump (Harvard Apparatus, Holliston, MA). Dialysate samples were collected over 30 (mPFC) or 15min (NAC). Online quantification of dopamine was accomplished by high-performance liquid chromatography (HPLC) coupled to electrochemical detection (ESA Bioscience, Chelsford, MA), with a detection limit of ~0.2fmol/min. The injector (Valco Instruments, Houston, TX) were directed by a computerized system, Totalcrom WS version 6.3 (Perkin Elmer, Wellesley, MA). Separation of dopamine and metabolites was achieved by reversed phase liquid chromatography. The mobile phase consisted of 55mM sodium acetate buffer (pH 4.0), 12% methanol and 0.55mM octanesulfonic acid and delivered by an HPLC pump (Model 2150, Pharmacia LKB, Sweden) on a C-18 column (Nucleocil 150/75 × 4.6mm, 5μm), flow rate 0.8ml/min. After separation, the analysate was passed through a guard cell with an applied oxidizing potential of 50mV to reduce baseline. Samples were quantified by sequential oxidation and reduction in a high-sensitive analytical cell (model 5011; ESA Bioscience) that was controlled by a potentiostat (Coulochem II model 5200; ESA Bioscience) with applied potentials of 400mV and −200mV for detection of metabolites and dopamine, respectively. Injection of drug was performed after a stable outflow (<10% variation) of dopamine and metabolites. Baseline was calculated as the average of the last two (mPFC) or four (NAC) pre-injection values. The placement of the probe was later verified under microscope in slices stained with neutral red.
Our procedures for electrophysiological experiments have previously been described (Konradsson et al, 2006). Briefly, the rats were decapitated while under halothane anesthesia (AstraZeneca AB, Södertälje, Sweden). The brains were quickly removed and cooled in ice-cold Ringer's solution (in mM: 126 NaCl, 2.5 KCl, 2.4 CaCl2, 1.3 MgCl2, 1.2 NaH2PO4, 10 -glucose, 18 NaHCO3, pH 7.4) aerated by 95% O2:5% CO2. The brains were cut coronally into 450μm slices, using a Vibroslice (Campden model MA 752, World Precision Instruments, FL), and kept submerged in aerated Ringer's solution at room temperature for >1h to allow for recovery. A slice containing mPFC (approximately anterioposterior +3.2mm from bregma) was transferred to a recording chamber (32°C), and held submerged between two nylon nets. The chamber was continuously perfused with aerated Ringer's solution, flow rate 1.5-2.5ml/min. Electrodes were pulled from borosilicate glass capillaries (i.d. 0.58mm; Clark Electromedical Instruments, Pangbourne, UK) by using a horizontal electrode puller (Model P-87, Sutter Instruments, San Rafael, CA). Recording electrodes were filled with 2M KAc (55–120MΩ), and used for recording with an Axoclamp 2A amplifier (Molecular Devices, CA). Penetrations of cells by sharp electrodes were performed blindly. The electrophysiological criteria for distinguishing presumed pyramidal from non-pyramidal neurons have been described previously (Arvanov et al, 1997; Connors and Gutnick, 1990). It is very rare to impale fast-spiking interneurons with a relatively low resistance microelectrode (McCormick et al, 1985), which might account for not recording fast-spiking non-pyramidal cells. The presumed pyramidal cells of the mPFC have relatively long spike duration (1–3ms at half-maximum spike amplitude) and show a pronounced spike-frequency adaptation in response to constant-current depolarization pulses. Single electrode voltage-clamp (holding potential −60mV) was performed in the discontinuous mode with a sampling rate of 5–6.2kHz. The voltage-clamp recordings were acquired using digital/analogue sampling and acquisition software (Clampex version 9.2, Molecular Devices). During the voltage-clamp recordings of NMDA (10–15μM)-evoked currents, tetrodotoxin (0.5μM, to block the action potentials), glycine (1μM, to enhance the NMDA-induced responses) and bicuculline (5μM, to block the GABAA responses) were included in the Ringer's solution. All drugs used were diluted in Ringer's solution and administered through bath perfusion. The amplitude of the NMDA-induced current after drug or drug combinations was divided by the amplitude of the control NMDA-induced current for calculation of drug effect on prefrontal NMDA transmission.
The data from the behavioral experiments are not normally distributed and accordingly non-parametric tests were used. Thus, for the CAR data, that is, avoidance behavior and intertrial crosses, we used the Friedman two-way analysis of variance (ANOVA) followed by Wilcoxon matched-pairs signed-ranks test, and for the catalepsy data the Kruskal–Wallis one-way ANOVA followed by Mann–Whitney U-test was used. Statistical evaluation of microdialysis data over time was performed by means of a two-way (treatment × time) ANOVA for repeated measures. To analyze difference between different treatments, we also measured the overall effects (AUC=area under curve), that is, interval 60–240min for mPFC and 45–240min for NAC. The overall effects were statistically evaluated by one-way ANOVA, followed by planned comparisons test. One-way ANOVA was also used to detect differences between baseline values. Statistical evaluation of the electrophysiology experiments were performed by paired t-test, and, for multiple group comparisons, one-way ANOVA followed by the post hoc Tukey's HSD test. In all statistical measures p<0.05 was considered significant. The statistical evaluations were performed by using Statistica version 8.0 (StatSoft).
For in vivo experiments olanzapine was dissolved in a minimal amount of acetic acid with a 5.5% glucose solution added to volume and reboxetine was dissolved in saline (0.9% NaCl) and administered by i.p. injections. Bicuculline methiodide, glycine and NMDA were purchased from Sigma-Aldrich, St Louis, MO and tetrodotoxin from Tocris, Bristol, UK. For the electrophysiological experiments, stock solutions of olanzapine and reboxetine (dissolved in dimethyl sulfoxide) were prepared. The possibility that peripheral pharmacokinetics, involving an interaction between reboxetine and low-dose olanzapine, could account for some of our findings seems unlikely, because olanzapine is metabolized by the enzyme CYP1A2, reboxetine is metabolized by an entirely different enzyme, CYP3A4.
The overall effect showed a statistical significant suppression of CAR at 20min (χ2 (5)=28. 71, p<0.001, Figure 1a) and 90min (χ2 (5)=27.24, p<0.001, Figure 1b, n=10), whereas at 240min all animals were back to baseline performance. Compared with control, that is, saline+vehicle, olanzapine 1.25mg/kg produced a small but significant suppression of CAR at 90min (p<0.01), and olanzapine 2.5mg/kg at both 20 and 90min (p<0.01). Reboxetine 6mg/kg had no effect on CAR behavior when given alone. However, the same dose of reboxetine enhanced the effect of a low dose of olanzapine 1.25mg/kg (20min; p<0.05), but not a high dose of olanzapine 2.5mg/kg.
The drug-induced suppression of CAR was accompanied by a concomitant decrease in intertrial crosses. There was a statistical significant overall effect for the intertrial crosses at 20min (χ2 (5)=11.95, p<0.05) and 90min (χ2 (5)=13.25, p<0.05, n=10, Table 1). At 20min, there was a statistically significant decrease in intertrial crosses after saline+olanzapine 1.25mg/kg, reboxetine+olanzapine 1.25mg/kg and reboxetine+olanzapine 2.5mg/kg, whereas at 90min only reboxetine+olanzapine 2.5mg/kg resulted in a statistically significant decrease. No escape failures were recorded at any time point for any of the treatments.
Reboxetine (6mg/kg), olanzapine (1.25 and 2.5mg/kg) and the combination of reboxetine and olanzapine were tested in the catalepsy test (30min: H(5)=10.15, p=0.071; 60min: H(5)=11.35, p<0.05; 120min: H(5)=3.91, p=0.56; Table 2). Compared with control, that is, saline+vehicle, reboxetine 6mg/kg (p<0.001), as well as olanzapine 2.5mg/kg (p<0.01) reached statistically significant level at 60min (Figure 2). However, the median catalepsy scores were all below 2, and in accordance with previous studies an animal was considered cataleptic only with a score 2 (Wadenberg et al, 2001).
The mean±SEM basal extracellular dopamine levels in the mPFC and NAC were 0.32±0.03fmol/min (n=26) and 2.62±0.22fmol/min (n=23), respectively. No statistically significant differences between mean baseline concentrations of dopamine output were found between different treatments group within the same brain region studied. Control injections, that is, saline and vehicle, had no effect on dopamine output in any of the regions analyzed.
In the mPFC, the maximal dopamine increase for reboxetine 6mg/kg was 320% (at 60min), for olanzapine 1.25mg/kg 102% (at 60min), for the combination of reboxetine 6mg/kg and olanzapine 1.25mg/kg 773% (at 60min) and for the control group (saline+vehicle injections) 16% (at 60min). Statistical evaluation for dopamine output in the mPFC (Figure 3a) revealed a significant treatment (F3,22=13.89, p<0.001), time (F8,176=30.47, p<0.001), as well as interaction (treatment × time) effect (F24,176=11.02, p<0.001). To assess the effects of the two drugs in combination, we analyzed the overall effect after injection of the second drug, AUC (60–240min). The overall effect was statistically significant (F3,22=14.21, p<0.001; Figure 3b). Compared with control, that is, saline+vehicle, reboxetine, as well as the combination of reboxetine and olanzapine were significantly higher (p<0.01–0.001). The combination of reboxetine and olanzapine was statistically higher than both reboxetine (p<0.01) and olanzapine (p<0.001), when given alone.
In the NAC, the maximal dopamine increase for reboxetine 6mg/kg was 30% (at 225min), for olanzapine 1.25mg/kg 172% (at 60min), for the combination of reboxetine 6mg/kg and olanzapine 1.25mg/kg 202% (at 60min) and for the control group (saline+vehicle injections) 18% (at 30min). Statistical evaluation for dopamine output in the NAC (Figure 3c) revealed a significant treatment (F3,19=13.24, p<0.001), time (F16,304=5.69, p<0.001), as well as interaction (treatment × time) effect (F48,304=2.33, p<0.001). To assess the effects of the two drugs in combination, we analyzed the overall effect after injection of the second drug, AUC (45–240min). The overall effect was statistically significant (F3,19=13.37, p<0.001; Figure 3d). Compared with control, that is, saline+vehicle, olanzapine, as well as the combination of reboxetine and olanzapine were significantly higher (p<0.01–0.001). The combination of reboxetine and olanzapine was statistically higher than reboxetine (p<0.001), when given alone.
Intracellular voltage-clamp was used to record NMDA-induced currents in pyramidal cells of layer V and VI in the prelimbic region of the rat mPFC. The characterization of pyramidal cells was performed according to criteria previously described (see above Arvanov et al, 1997; Konradsson et al, 2006).
The concentration-response curve for the effects of olanzapine, as well as other atypical APDs, on the NMDA-induced currents recorded has been shown to display an inverted U-shaped curve and the maximal effect of olanzapine were observed at 5nM (see Ninan et al, 2003). In this study, we examined a potential enhancement by reboxetine of a sub-maximal concentration of olanzapine, that is, 3nM. First, reboxetine in various concentrations (5, 20, 200 and 1000nM) was tested alone, and was generally found to produce very small effects on the NMDA-induced currents regardless of concentrations (data not shown). Plasma concentration in patients treated with reboxetine has been shown to be 660±346nM (mean±SD; Öhman et al, 2001). Reboxetine shows an extensive (>97%) binding to plasma proteins (Fleishaker 2000) and, as the amount of drug present in cerebrospinal fluid (CSF) tends to be equivalent at steady-state to the non-protein-bound drug concentration in plasma, 20nM was chosen as a clinically relevant concentration.
There was a small but significant facilitation of the NMDA-induced currents by 20nM reboxetine (117.7±4.2 (mean±SEM), n=7, t(6)=4.18, p<0.01, Figure 4), whereas 3nM olanzapine did not reach statistical significance (118.2±7.3, n=5, t(4)=2.49, p=0.07). The facilitating effect on NMDA-induced currents by the combination of reboxetine 20nM and olanzapine 3nM was highly significant (170.0±6.0, n=5, t(4)=11.62, p<0.001). Group comparison (F(treatment: 2,14)=26.3, p<0.001) revealed statistical significant differences between reboxetine 20nM and the combination of reboxetine 20nM and olanzapine 3nM (p<0.001) and between olanzapine 3nM and the combination of reboxetine 20nM and olanzapine 3nM (p<0.001).
Our results show that the selective NET inhibitor reboxetine may significantly enhance the antipsychotic-like effect of a low, but not a high, dose of olanzapine, without inducing catalepsy. In principle, these data suggest the possibility of a dose reduction of olanzapine with maintained antipsychotic effect. The olanzapine-induced dopamine output in the mPFC, but not in the NAC, was significantly increased by reboxetine. Finally, reboxetine was also shown to potentiate the effect of a low concentration of olanzapine on prefrontal glutamatergic NMDA receptor-mediated transmission. The preferential enhancement of cortical dopamine and NMDA receptor-mediated transmission suggest that addition of reboxetine to a sub-maximal dose of olanzapine may not only enhance the antipsychotic effect, but also ameliorate negative and depressive symptoms and cognitive impairment in schizophrenia (c.f. Introduction).
Reboxetine has previously been found to enhance the antipsychotic-like effect and cortical dopamine output of a low dose of the selective D2/3 antagonist raclopride, which by itself did not produce a sufficient antipsychotic effect (Linner et al, 2002). In this study, we found that the same dose of reboxetine also enhanced the antipsychotic-like effect and cortical dopamine output induced by a low dose of olanzapine, which also when given alone was far from producing a sufficient antipsychotic-like effect in the CAR model. The low doses of both raclopride and olanzapine used in these experiments produce approximately 60–65% occupancy of the D2 receptor (Marcus et al, 2005; Wadenberg et al, 2001). The doses needed for sufficient antipsychotic effect in the CAR model of both raclopride and olanzapine, when given alone, require approximately 75% D2 receptor occupancy (Wadenberg et al, 2000, 2001). In schizophrenic patients treated with clozapine, the striatal dopamine receptor occupancy is rather low, that is, approximately 45% (Kessler et al, 2006; Nordström et al, 1995), indicating that also other receptors than D2 receptors contribute to its efficacy (see Svensson, 2003b). In schizophrenic patients treated with FGAs, the D2 receptor occupancy is approximately 70–80% (Farde et al, 1988; Nordström et al, 1995), and a level similar to that of D2 receptor occupancy is also observed in patients treated with olanzapine (Tauscher et al, 1999; Zipursky et al, 2005). At high D2 receptor occupancy levels the risk of adverse side effects such as EPS and prolactin elevation are increased (Kapur et al, 2000). In addition, a high degree of D2 blockade per se may generate detrimental effects on both cognition and mood in healthy volunteers (Saeedi et al, 2006), as well as in schizophrenic patients (Carpenter Jr, 1996). Therefore, a lower D2 receptor occupancy, which may be achieved by reducing the dose of olanzapine, may in principle even contribute to ameliorate cognitive and negative symptoms. The present data thus indicate that this goal may be attainable by adjunctive reboxetine treatment.
In similarity with reboxetine, the α2-adrenoceptor antagonist idazoxan has been shown to enhance the antipsychotic-like effects of raclopride and olanzapine, as well as haloperidol and risperidone, without inducing catalepsy. In addition, the prefrontal dopamine output induced by these APDs was markedly enhanced by idazoxan (Hertel et al, 1999; Marcus et al, 2009; Wadenberg et al, 2007). Furthermore, to a similar magnitude as that of clozapine, idazoxan enhanced the effects of both raclopride and risperidone on prefrontal NMDA receptor-mediated transmission, an effect that has been shown to be dopamine-dependent and executed through the dopamine D1 receptor (Chen and Yang, 2002; Marcus et al, 2005, 2009; Ninan and Wang, 2003). The increased prefrontal availability of catecholamines induced by adjunctive treatment with idazoxan is probably due to a co-release of dopamine and norepinephrine that is controlled by α2-adrenoceptors located on norepinephrine nerve terminals (Devoto et al, 2001).
NRIs have, in depressed patients, been suggested to improve drive and motivation, as well as cognitive impairment because of increased availability of cortical catecholamines (see Introduction). The increased availability of dopamine in the prefrontal cortex observed after administration of a selective NRI may largely be explained by the fact that dopamine has a high affinity for the NET, which thus contributes to clearance of dopamine from the extracellular space in the cortex (Carboni et al, 1990; Pozzi et al, 1994). Thus, after administration of a NET inhibitor, both norepinephrine and dopamine extracellular levels increase. The increased availability of dopamine in the prefrontal cortex may not only contribute to improve the effect on cognitive symptoms, but can probably also, in the presence of a D2 antagonist, contribute to enhance the antipsychotic effect per se. Previous clinical studies show that low doses of -dopa used as adjunct treatment to conventional APDs may enhance their therapeutic effect in schizophrenic patients (see Jaskiw and Popli, 2004), and experimental data reveal that adding low doses of -dopa to raclopride not only caused a preferentially increased prefrontal cortical dopamine output, but also potentiated the antipsychotic-like effect of the D2/3 antagonist (Eltayb et al, 2005). Therefore, addition of drugs that are able to increase the availability of cortical monoamines, for example, reboxetine and idazoxan, to APDs may both enhance the antipsychotic effect, as well as improve negative or depressive symptoms and certain aspects of cognitive impairment.
Clinical studies using adjunctive treatment with NET inhibitors to APDs have show inconsistent results. Although reboxetine when added to a stable treatment with APDs, including both FGAs and SGAs, it was found to improve both positive and negative, as well as depressive symptoms in schizophrenic patients (Raedler et al, 2004), whereas reboxetine as add-on treatment to haloperidol did not cause any significant improvement (Schutz and Berk, 2001). In addition, adjunctive treatment with the NET inhibitor atomoxetine to schizophrenic patients treated with SGAs was found to activate certain brain areas related to working memory (Friedman et al, 2008). Interestingly, a recent study found no improvement of processing speed and accuracy by co-administration of reboxetine with olanzapine in schizophrenic patients (Poyurovsky et al, 2009), but the investigators concluded that a more comprehensive test battery, including tasks such as working memory, verbal memory and social cognition, might have been more useful. Needless to say, these clinical results are obtained from small pilot studies in patients maintained on effective doses of various APDs. However, this study suggests that the potential benefits of adjunctive NET inhibition would be obtained at reduced dosage of olanzapine rather than at standard dose levels. Significantly, quetiapine, with a structure similar to clozapine and olanzapine, has an active metabolite, nor-quetiapine, which is a potent NET inhibitor that has been identified and visualized in the primate brain (Nyberg et al, 2007). Thus, the recently shown antidepressant effect of quetiapine (Bandelow et al, 2010; Bauer et al, 2009; Cutler et al, 2009), as well as its utility in bipolar depression may be related to concomitant NET inhibition in combination with a low D2 occupancy (Kessler et al, 2006). Finally, clinical data indicate that adding a NET inhibitor such as reboxetine to olanzapine in the treatment of schizophrenia attenuates the olanzapine-induced weight gain (Poyurovsky et al, 2003, 2007).
Consequently, further studies on NET inhibition as add-on treatment to APDs in the treatment of schizophrenia seem warranted. Specifically, the present data suggest that the addition of a NET inhibitor to olanzapine treatment might allow for a dose reduction, yet with maintained and even broadened therapeutic effects and a concomitant reduction of dose-related side effects of olanzapine, such as EPS and weight gain. To our knowledge no clinical study using this approach has as yet been published.
This work was supported by the Swedish Research Council (grant no. 4747), the Karolinska Institutet, Torsten och Ragnar Söderbergs Stiftelser, Magn. Bergvalls Stiftelse, Åhlén-stiftelsen and the Lundbeck Foundation. The expert technical assistance of Mrs Ann-Chatrine Samuelsson is gratefully acknowledged. Reboxetine is widely used in over 60 countries, however, the FDA has declined the approval of reboxetine in the United States.
The authors declare that there is no conflict of interest. Monica M Marcus, Anna Malmerfelt and Torgny H Svensson are employees at the Karolinska Institutet, Carl Björkholm is a PhD student at the Karolinska Institutet. Kent Jardemark is employee at the Karolinska Institutet and Pronexus Analytical AB.