Main

Attention Deficit/Hyperactivity Disorder (ADHD) is a common behavioral disorder found in 3–7% of school age children (American Psychiatric Association 2000; Szatmari 1992; Offord et al. 1987). ADHD is characterized by levels of increased motor activity, impulsiveness, distractibility, restlessness and inattention that is maladaptive and inconsistent with the child's developmental level. Three subtypes have been identified: inattentive, hyperactive/impulsive and combined (inattentive/hyperactive/impulsive). Children with ADHD have increased risk for low educational and vocational attainment, as well as social dysfunction, increased criminality, and drug abuse (Gittelman et al. 1985; Mannuzza et al. 1993). ADHD is frequently associated with oppositional and conduct disorders, depression, and anxiety (Downey et al. 1997; Biederman et al. 1991). ADHD often persists into adult life (Spencer et al. 1996; Gittelman et al. 1985; Barkley et al. 1990).

Dysfunction of catecholamine and particularly dopamine (DA) neuronal systems has been postulated to be involved in ADHD (Castellanos et al. 1996a; Zametkin and Rapoport 1987). Dopamine plays a key role in attentional, psychomotor, reinforcing and rewarding behaviors that are deficient in ADHD. Amphetamine and methylphenidate, which have been widely used to treat ADHD, block DA and norepinephrine (NE) transporters and thereby enhance catecholamine neurotransmission (Barkley 1977; Spencer et al. 1995; Gatley et al. 1996).

Norepinephrine has also been proposed to play a key role in the pathophysiology and pharmacotherapy of ADHD (Zametkin and Rapoport 1987; Pliszka et al. 1996; Arnsten et al. 1996; Biederman and Spencer 1999). The noradrenergic system is involved in attentional processes and has been shown to prime the prefrontal cortex (PFC) for response to sensory stimuli (Segal and Bloom 1976; Aston-Jones et al. 1991; Berridge et al. 1993). Increased basal activity of locus coeruleus noradrenergic cell bodies may decrease the response of the PFC and thus treatments that reduce locus coeruleus activity have been hypothesized to improve attentional, arousal, and cognitive processes (Pliszka et al. 1996).

Pharmacotherapy of ADHD to control the behavioral symptoms has been largely with psychostimulants such as d-amphetamine and methylphenidate (Spencer et al. 1995; Wilens et al. 1998). However, about 10–30% are non-responders or are intolerant to psychostimulant therapy (Barkley 1977; Elia et al. 1991; Greenhill 1995). Adverse reactions associated with stimulant use include insomnia, tics, anorexia, anxiety, and dysphoric mood (Gittelman 1980; Greenhill 1995). The older psychostimulants have short half lives that result in problematic multiple daily doses, particularly with concerns about insomnia after late afternoon and evening administration. In addition, the psychostimulants used in ADHD are controlled substances that have liability for drug abuse and diversion, thus limiting their usefulness (Holman 1994). Concerns about the tolerability and abuse liability of psychostimulants have led to interest in the development of drugs for pharmacotherapy of ADHD with a different mechanism of action.

The tricyclic antidepressants, desipramine and nortriptyline, have emerged as alternative therapies for the treatment of ADHD (Spencer et al. 1996; Wilens et al. 1996). Desipramine and nortriptyline have high affinity for NE transporters relative to DA and serotonin (5-HT) transporters (Wong et al. 1995) and have been shown in clinical evaluations of adults and children to be effective in ADHD (Biederman et al. 1989; Wilens et al. 1996). However, tricyclic antidepressants have significant affinity for α1-adrenergic, cholinergic and histaminergic receptors potentially resulting in sedation, dry mouth, weight gain and cognitive impairment and also have cardiovascular concerns (Wong et al. 1995; Cookson 1993; Leonard et al. 1995; Walsh et al. 1994). Therefore, the tricyclic antidepressants are also limited in usefulness.

Recently, atomoxetine (tomoxetine, LY139603) was found to be efficacious in the treatment of ADHD in a double-blind, placebo-controlled crossover study in adults (Spencer et al. 1998). In children and adolescents, atomoxetine had superior outcomes in reducing ADHD symptoms compared with placebo and had a graded dose response (Michelson et al. 2001). Atomoxetine significantly reduced core symptoms of ADHD in an open-label study in children (Spencer et al. 2001). Overall, it was well tolerated in the child and adolescent age group (Michelson et al. 2001; Spencer et al. 2001). Atomoxetine has also been shown in a 6-week open label study to produce clinically significant improvement in the symptoms in patients with major depression (Chouinard et al. 1984).

Atomoxetine is a potent NE uptake inhibitor in vitro and in vivo with relatively low affinity for 5-HT and DA uptake processes (Wong et al.1982; Bolden-Watson and Richelson 1993). Furthermore, atomoxetine is a potent inhibitor of the presynaptic NE transporter and has minimal affinity for other neurotransmitter transporters and neuronal receptors (Gehlert et al. 1995; Tatsumi et al. 1997; Wong et al. 1982). In this study, we have further investigated the pharmacology of atomoxetine to understand its therapeutic actions in ADHD. The affinity of atomoxetine and methylphenidate for human monoamine uptake transporters and the potential interaction of atomoxetine with neuronal receptors was evaluated. To evaluate the potency and selectivity of uptake inhibitors of NE and 5-HT transporters in vivo, the ability of uptake inhibitors to block neurotransmitter depletion in rat brain induced by monoamine transporter-dependent neurotoxins was investigated. The activation of expression of the neuronal activity marker, Fos, by atomoxetine was determined in several brain regions. Finally, we compared the effects of atomoxetine administration with another NE uptake inhibitor, reboxetine, and with methylphenidate on extracellular levels of monoamines in rat brain regions potentially involved in ADHD including PFC.

MATERIALS AND METHODS

Transporter Binding Studies

Membranes from HEK 293, MDCK, and HEK293 cell lines transfected with human 5-HT, NE and DA transporters, respectively, were obtained from Receptor Biology, Inc. (Beltsville, MD). All assays were performed in triplicate in a final volume of 0.8 ml containing either a buffer consisting of 50 mM Tris Cl, pH 7.4, 150 mM NaCl, and 5 mM KCl for 5-HT and NE transporters or 50 mM Tris Cl, pH 7.4, and 100 mM NaCl for the DA transporter. The radioligands for 5-HT, NE and DA human transporters were [3H]-paroxetine (0.2 nM, 25 Ci/mmol, Dupont NEN products, Boston, MA), [3H]-nisoxetine (1.0 nM, 86 Ci/mmol, New England Nuclear) and [3H]-WIN35,428 (1.0 nM, 86 Ci/mmol, New England Nuclear), respectively. Membranes equivalent to protein in amounts of 10.3 μg, 16.9 μg or 6.2 μg, respectively, were used in the assays. After incubation at 37°C for 40 min for the 5-HT transporter and 25°C for 30 min for NE and DA transporters, the binding was terminated by rapid vacuum filtration over Whatman GF/B filters (presoaked in 0.5% polyethylenimine) and the filters were washed four times with cold 50 mM Tris Cl buffer, pH 7.4. The filters were then placed in vials containing liquid scintillation fluid and radioactivity was measured by liquid scintillation spectrometry. Non-specific binding was determined in separate samples with 1 μM duloxetine, 10 μM desipramine or 10 μM nomifensine, for 5-HT, NE and DA transporters, respectively.

Determination of Binding to Neuronal Receptors

Inhibition of binding to neuronal receptors was provided by NovaScreen (Hanover, MD) using proprietary receptor binding assays.

Blockade of p-Chloramphetamine and DSP-4 Effects

The 5-HT selective neurotoxin p-chloramphetamine hydrochloride (p-CA) was dissolved in sterile water for intraperitoneal (i.p.) administration. Male Sprague-Dawley rats (Harlan Industries, Indianapolis, IN) in groups of five weighing 140–170 grams were injected with p-CA (10 mg/kg i.p.) 2 h before cervical dislocation. Vehicle or test drugs were dissolved in sterile water and injected 1 h prior to p-CA. The brains were quickly removed, frozen on dry ice and stored at −70°C until assayed. Whole brain 5-HT concentrations were measured using high pressure liquid chromatography with electrochemical detection (HPLC-EC) as previously described (Fuller and Perry 1989).

DSP-4 (N-(2-chloroethyl)-N-ethyl-2-bromobenzylamine hydrochloride), a noradrenergic neurotoxin (Jonsson et al. 1981; Grzanna et al. 1989), was dissolved in 0.01 N HCl and injected at 30 mg/kg i.p. one hour after administration of drugs. Cortical NE concentrations were measured 6 h after administration of DSP-4 according to a modification of the method of Fuller and Perry (1977). Cortical tissue was homogenized in 2 ml 0.1 N trichloroacetic acid with 3,4-dihydrobenzylamine hydrobromide added as an internal standard. After centrifugation, 1 ml of the supernatant was added to 150 mg alumina, and the pH adjusted to 8.6 with 0.5 M Tris containing 0.1 M EDTA. Samples were mixed for 10 min, centrifuged and the supernatant decanted. The alumina was washed with 1 ml 50 mM Tris/10 mM EDTA. One ml of 0.1 N formic acid was added to the alumina, and the samples were mixed for 10 min and centrifuged. Cortical NE was measured using a HPLC-EC technique by injecting a 20 μl aliquot of the supernatant onto a BDS Hypersil C-18 analytical column (Keystone Scientific, Inc.). The elution buffer contained 75 mM sodium phosphate, 0.5 mM EDTA, 350 mg/l 1-octanesulfonate sodium, 4% acetonitrile (v/v) and 0.8% tetrahydrofuran, pH 3.0 and the flow rate was 1.2 ml/minute. Dihydrobenzylamine and NE peaks were measured at 600 mV with 10 nA sensitivity using an electrochemical method and compared with samples containing known amounts of NE.

Fos Immunohistochemistry

Two hours after administration of atomoxetine (3 mg/kg, i.p.), the rats were deeply anesthetized with sodium pentobarbital (60 mg/kg, i.p.) and transcardially perfused with 100 ml of phosphate buffered saline (PBS) followed by 100 ml of 4% paraformaldehyde in PBS The brain was rapidly removed and postfixed for 90 min in 4% paraformaldehyde, and then was transferred to 30% sucrose at 4°C until saturated. After quick freezing, serial 30 μm sections were cut and placed in PBS until processed for immunohistochemistry. In brief, sections were incubated in PBS containing blocking serum and 0.5% Triton-X 100 for 1 h. Sections were then incubated with anti-Fos antibody (Santa Cruz Biotechnology, Inc.) at 4°C overnight. Visualization of the Fos-like immunoreactivity was performed with a Vectastain ABC Elite Kit (Vector Labs, Burlingame, CA) using the standard protocol supplied with the kit. Nickel-intensified diaminobenzidine (DAB) was used as the chromagen to yield a gray-black precipitation product. Following visualization of the Fos immunoreactivity, the sections were mounted on gelatin-coated glass slides and allowed to dry. The sections were then dehydrated and cover slipped. Fos expressing cells were quantitated using the MCID M2 imaging system (Imaging Research, St. Catherines, Ontario).

Microdialysis Studies

Male Sprague-Dawley rats from Harlan Industries (Indianapolis, IN) were used for all studies. For the microdialysis studies, rats weighing 260–300 mg were anesthetized with chloral hydrate/pentobarbital (170 mg/kg and 36 mg/kg, respectively, in 30% propylene glycol and 14% ethanol) for implantation of the dialysis probes.

The microdialysis technique used here has been described previously (Li et al. 1998; Zhang et al. 2000). In brief, a homemade loop type probe with a regenerated cellulose dialysis fiber of 3 mm tip length (6 mm total) fused into PE-10 tubing was used. Coordinates for the PFC were: A (anterior to bregma), 3.2 mm; L (lateral from the midsagittal suture), 0.8 mm; and V (ventral from the dura surface), 4 mm (Paxinos and Watson 1986). The coordinates used for the nucleus accumbens were A, 2.0 mm; L, 1.5 mm; and V, 8.0 mm. The coordinates used for the striatum (caudate putamen) were A, 0.2 mm; L, 3.0 mm; and V, 6.5 mm. At the end of experiments, the probe position was histologically verified by perfusing a dye (2,3,5 triphenyltetrazolium chloride, 5 mg/ml in water) through the dialysis probe and then sectioning the frozen brain to observe the probe location. Rats with improper probe location were not included in the statistical analysis.

Microdialysis experiments were performed two days after surgery to allow the rats to fully recover from the operation and resume normal food intake. The rat was placed in a plastic bowl and connected to a fraction collection system for freely moving animals (Raturn, BioAnalytical Systems (BAS), West Lafayette, IN). The input tube of the dialysis probe was connected to a syringe pump (BeeHive and BabyBee, BAS) which delivered an artificial cerebrospinal fluid containing 150 mM NaCl, 3 mM KCl, 1.7 mM CaCl2 and 0.9 mM MgCl2 (pH 6.0) to the probe at a rate of 1.0 μl/min. The output tubes from the rats were attached to a refrigerated fraction collector (820 microsampler, Scipro, North Tonawanda, NY). After acclimatization of the rat to the apparatus and establishment of stable monoamine baseline levels, the drugs were administered in a volume of 1 ml/kg i.p. in sterile water. Monoamines in dialysates were measured off-line by the analytical method described in Li et al. (1998). The sensitivity for DA, NE and 5-HT was 0.1 pmol/ml dialysate or 2 fmol/sample (20 μl). All microdialysis data were calculated as percent change from dialysate basal concentrations with 100% defined as the average of the final three drug preinjection values and each group had 5–6 rats.

Drugs

Atomoxetine, reboxetine, and fluoxetine were provided by the Lilly Research Laboratories, Indianapolis, IN. Desipramine, buproprion, imipramine, nomifensine, p-CA, and DSP-4 were purchased from Sigma Chemical Company, St. Louis, MO. Methylphenidate was purchased from Mallinckrodt, St. Louis, MO. All other chemicals were reagent grade quality.

Statistics

Inhibition curves for in vitro studies were analyzed by nonlinear least-squares curve fitting to obtain IC50 values. The inhibition constant (Ki) values were calculated from IC50 and Kd values according to the method of Cheng and Prusoff (1973). The Kd values of the radioligands used for the 5-HT, NE and DA transporters were 0.18, 2.5 and 22.8 nM, respectively. All values for microdialysis studies were calculated as percentage change at each time point compared with the average of three baseline values. Significant differences for the time course of vehicle control injection on NE, DA or 5-HT were determined by a 1-way analysis of variance (ANOVA) for repeated measures with respect to time. Differences between treatment groups, including control, were determined by a 2-way ANOVA with treatment as the independent variable and time as the repeated measure. If significant, the ANOVA was followed by a post-hoc Duncan's multiple range test on the overall effect of treatment using the Statistica program (StatSoft, Tulsa, OK).

RESULTS

Transporter and Receptor Binding

Atomoxetine inhibited radioligand binding in cells transfected with human NE, 5-HT, and DA transporters with Ki values of 5, 77 and 1451 nM, respectively. (Table 1). The selective NE uptake inhibitors reboxetine (Wong et al. 2000) and desipramine also had high affinity and selectivity for NE transporters. In contrast, methylphenidate had higher affinity for human DA transporters than NE transporters and low affinity for 5-HT transporters. Nomifensine had high affinity for DA and NE transporters and buproprion had moderate affinity for DA transporters. Imipramine inhibited binding to NE and 5-HT transporters, whereas fluoxetine had appreciable affinity for 5-HT transporters only. The Hill coefficients for binding of the compounds to the transporters were not significantly different from 1. The affinity of atomoxetine for over 60 other neuronal receptors, transporters and binding sites was investigated. Atomoxetine did not have appreciable affinity for the receptors and binding sites investigated (Table 2).

Table 1 Affinity of Atomoxetine and Other Uptake Inhibitors for Human Monoamine Transporters
Table 2 Atomoxetine Has Low Affinity (>1 μM) for Neuronal Receptors, Transporters and Other Binding Sites

The ability of atomoxetine to block 5-HT and NE uptake processes in vivo was determined using the transporter specific neurotoxins p-CA and DSP-4 which deplete 5-HT and NE concentrations in rat brain regions, respectively. Atomoxetine blocked depletion of rat cortical NE concentrations by DSP-4 in a dose-dependent manner with an ED50 of 2.5 mg/kg p.o. (Table 3 and Figure 1, panel A). However, atomoxetine did not significantly block p-CA–induced depletion of rat brain 5-HT (Figure 1, panel B). Like atomoxetine, the selective NE uptake inhibitors desipramine and reboxetine also blocked depletion of cortical NE by DSP-4 and had modest effects on p-CA–induced depletion of brain 5-HT at doses up to 30 mg/kg (Table 3). Methylphenidate did not inhibit either DSP-4–induced NE depletion or p-CA–induced 5-HT depletion, but the selective 5-HT uptake inhibitor fluoxetine potently inhibited p-CA–induced 5-HT depletion (Table 3).

Table 3 Blockade of the p-Chloramphetamine (p-CA)-induced Rat Brain Serotonin Depletion and DSP-4-induced Depletion Of Cortical Norepinephrine Concentrations by Atomoxetine and Other Monoamine Reuptake Inhibitors
Figure 1
figure 1

Effect of atomoxetine on the depletion of rat cortical norepinephrine (NE) concentrations produced by DSP-4 (30 mg/kg i.p.) (panel A) and brain serotonin (5-HT) concentrations produced by p-chloroamphetamine (p-CA, 10 mg/kg i.p.) (panel B). Atomoxetine was administered to rats in groups of five 1 h prior to DSP-4 and the rats were sacrificed 6 h after DSP-4 administration. B. Atomoxetine was administered to rats in groups of five 1 h prior to p-CA and the rats were sacrificed 2 h after p-CA administration. Control concentrations of NE in cortex were 1.92 ± 0.10 nmol/g and 5-HT concentrations in whole brain were 2.43 ± 0.12 nmol/g. Means and standard errors of the percent of control are shown.*p < .05 as compared with control group; #p < .05 as compared with DSP-4 or p-CA alone groups.

Microdialysis Studies

Basal Levels of Monoamines in Dialysates

The basal extracellular (EX) concentrations for 5-HT, NE, DA were 0.24 ± 0.03, 0.93 ± 0.05 and 0.41 ± 0.03, pmol/ml of PFC dialysate, respectively, in freely moving rats. The basal extracellular concentrations of DA in nucleus accumbens and striatum were 3.60 ± 0.50 and 4.54 ± 0.36 pmol/ml of dialysate, respectively.

Effect of Atomoxetine on Monoamine Extracellular Concentrations in Brain Regions

Vehicle injection produced no significant changes in monoamine levels compared with baseline concentrations with respect to time as shown by a 1-way ANOVA with repeated measures (F8,40 = 0.399, 2.090, and 0.681 for NE, DA, and 5-HT respectively). These vehicle values were used as controls in all statistical comparisons with the treatment groups mentioned below.

Atomoxetine (0.3 to 3mg/kg i.p.) increased NEEX concentrations in PFC up to a peak level of 290 ± 33% of basal concentrations and the higher doses had longer duration of increase (Figure 2, panel A). The 2-way ANOVA revealed a significant effect of treatment (F3,20 = 5.840) and time (F8,160 = 22.615) and a significant interaction of treatment and time (F24,160 = 3.063). The 4-h average for NEEX at the 3-mg/kg i.p. dose was 243 ± 12% of basal concentrations (p < .0022). Atomoxetine (0.3 to 3 mg/kg i.p.) also increased DAEX up to a peak level of 323 ± 17% (F3,19 = 28.500, F8,152 = 37.538, F24,152 = 9.444 for the effects of treatment, time and interaction, respectively) and a 4-h average of 251 ± 18% of basal concentrations at 3 mg/kg i.p. (Figure 2, panel B) (p < .001). The extracellular levels of 5-HT in PFC were not significantly increased by atomoxetine at doses up to 3 mg/kg i.p. (Figure 2, panel C) (F3,18 = 1.255, F8,144 = 1.609, F24,144 = 1.263 for the effects of treatment, time and interaction, respectively). The effects of atomoxetine on DAEX in PFC, nucleus accumbens and striatum were compared (Figure 3). The levels of DAEX in PFC were increased by atomoxetine (3 mg/kg i.p.) to a peak level of 323 ± 17% of basal, but were not significantly altered in nucleus accumbens (3 mg/kg i.p.) or striatum (10 mg/kg i.p.). NorepinephrineEX levels were not quantified in the nucleus accumbens or striatum, but 5-HTEX was not altered by atomoxetine treatment at doses of 3 mg/kg or less (data not shown).

Figure 2
figure 2

Time course of the effects of control (vehicle) or atomoxetine (0.3, 1, 3 mg/kg i.p.) administration on extracellular concentrations of norepinephrine (NE) (A), dopamine (DA) (B) and serotonin (5-HT) (C) in prefrontal cortex of freely moving rat. Values are the mean ± SEM of the % of pre-drug baseline determined at −1, −0.5 and 0 h. Administration of vehicle or atomoxetine at time 0 h is indicated by the arrow. Atomoxetine significantly increased extracellular NE and DA concentrations throughout the 4-h period (* p < .025, Duncan's post hoc test).

Figure 3
figure 3

Time course of the effects of atomoxetine administration on extracellular dopamine levels in prefrontal cortex (PFC), striatum and nucleus accumbens of freely moving rat. Values are the mean ± SEM of the % of pre-drug baseline determined at −1, −0.5 and 0 h. Administration of vehicle or atomoxetine (3 mg/kg i.p. in PFC and nucleus accumbens and 10 mg/kg i.p. in striatum) at time 0 h is indicated by the arrow. Atomoxetine significantly increased extracellular norepinephrine and dopamine concentrations throughout the 4-h period only in the PFC (* p < .05, Duncan's post hoc test).

Local perfusion of 0.34 μM atomoxetine through the dialysis probe into the PFC significantly increased NEEX and DAEX to the maximal effect of 175 ± 33 and 190 ± 15% of baseline concentrations, respectively, by 1 h (Figure 4). After 2.5 h the concentration of atomoxetine was increased to 1.03 μM, but the extracellular levels were not appreciably changed from the 0.34 μM concentration. A 1-way ANOVA with repeated measures showed significant effects for NEEX and DAEX (F10,50 = 3.380 for NE and F10,50 = 6.927 for DA). The concentration of 5-HTEX was not increased by local perfusion with atomoxetine at either concentration.

Figure 4
figure 4

The effects of local perfusion of atomoxetine on extracellular monoamine levels in prefrontal cortex of freely moving rat. Values are the mean ± SEM of the % of pre-drug baseline determined at −1, −0.5 and 0 h. Atomoxetine (0.34 μM) was perfused through the dialysis probe at time 0 to 2.5 h and the concentration was increased to 1.03 μM from 2.5–5 h. Atomoxetine significantly increased extracellular norepinephrine and dopamine concentrations throughout the 5-h period (* p < .05, Duncan's post hoc test).

Effect of Reboxetine and Methylphenidate on Monoamine Levels in Brain Regions

Reboxetine (3 mg/kg i.p.) significantly increased PFC NEEX and DAEX to peak levels of 344 ± 73 and 341 ± 28% of baseline concentrations, respectively, and had a 4 h average percentage increase of 282 ± 18 and 268 ± 21%, respectively (Figure 5, panel A). (NE 2-way ANOVA: F1,10 = 13.257, F8,80 = 6.492, F8,80 = 6.040 for treatment, time and interaction, respectively; DA 2-way ANOVA: F1,10 = 57.777, F8,80 = 15.947, F8,80 = 11.241 for treatment, time and interaction, respectively). Reboxetine did not significantly increase 5-HTEX(F1,7 = 2.113, F8,56 = 0.845, F8,56 = 1.624). Methylphenidate (3 mg/kg i.p.) significantly increased NEEX and DAEX to a peak concentration of 199 ± 16 and 253 ± 25% of PFC basal levels (p < .05) and a 4-h average of 145 ± 14 and 168 ± 17% of basal levels (p < .05), respectively (Figure 5, panel B). (1-way ANOVA for NE: F8,40 = 21.714, and for DA: F8,40 = 28.045).

Figure 5
figure 5

Time course of the effects of reboxetine (3 mg/kg i.p.) (A) and methylphenidate (3 mg/kg i.p.) administration (B) on extracellular monoamine levels in prefrontal cortex of freely moving rat. Values are the mean ± SEM of the % of pre-drug baseline determined at −1, −0.5 and 0 h. Administration of reboxetine or methylphenidate at time 0 h is indicated by the arrow. Reboxetine significantly increased extracellular norepinephrine (NE) and dopamine (DA) concentrations throughout the 4-h period (* p < .05, Duncan's post hoc test) whereas methylphenidate significantly increased NE and DA for the first 2.5 h (* p < .05, Duncan's post hoc test).

Methylphenidate significantly increased DAEX concentrations up to 253 ± 25, 209 ± 14 and 267 ± 37% of baseline in PFC, striatum and nucleus accumbens, respectively (Figure 6). The 4-h average increases in DAEX were 168 ± 17, 138 ± 14 and 169 ± 16% of baseline concentrations, respectively. A 2-way ANOVA comparing the three different areas showed that although there was a significant increase in DAEX there was no significant difference between the areas over the 4-h period. The F values were F2,15 = 1.599 for area comparison (not significant), F8,120 = 75.082 for effect over time (highly significant for all areas) and F16,120 = 1.644 for interaction between areas over time (not significant).

Figure 6
figure 6

Time course of the effects of methylphenidate (3 mg/kg i.p.) administration on extracellular dopamine levels in prefrontal cortex, striatum, and nucleus accumbens of freely moving rat. Values are the mean± SEM of the % of pre-drug baseline determined at −1, −0.5 and 0 h. Administration of methylphenidate at time 0 h is indicated by the arrow. Methylphenidate significantly increased extracellular dopamine concentrations through the 2.5-h period in all brain regions (* p < .05, Duncan's post hoc test).

Effect of Atomoxetine on Fos Expression in Brain Regions

Immunohistochemical localization of the neuronal activity marker Fos was determined 2 h after atomoxetine administration (3 mg/kg i.p.) in the same brain areas as in the microdialysis studies. Atomoxetine increased the number of cells expressing Fos-like immunoreactivity in PFC 3.7-fold, but did not significantly increase the number of Fos positive cells in striatum and nucleus accumbens (Table 4, Figure 7.

Table 4 Expression of Fos-like Immunoreactivity in the Prefrontal Cortex, Nucleus Accumbens and Striatum of Rat Following Vehicle or Atomoxetine (3 mg/kg i.p.) Administration
Figure 7
figure 7

Photomicrographs showing localization of Fos expression following vehicle or atomoxetine (3 mg/kg i.p.) administration to rats in groups of five. Boxed areas in panels A, D, and G indicate the relative area selected for quantitation for the prefrontal cortex (panels A–C), nucleus accumbens (panels D–F) and striatum (panels G–I). Higher magnification of a representative section of vehicle-treated (panels B, E, H) or atomoxetine-treated animals (panels C, F, I) for each respective area are presented. Scale bar = 1 mm.

DISCUSSION

Atomoxetine inhibited binding of [3H]-nisoxetine to human NE transporters with a Ki value of 5 nM and with 15- and 290-fold lower affinity for human 5-HT and DA transporters, respectively. This is consistent with previously reported selectivity of atomoxetine for rat NE uptake processes (Wong et al. 1982; Bolden-Watson and Richelson 1993; Gehlert et al. 1995). Desipramine and reboxetine also were potent and selective inhibitors of binding to human NE transporters, consistent with previous reports (Tatsumi et al. 1997; Wong et al. 2000). In contrast, the psychostimulant methylphenidate had higher affinity for the human DA transporter than the NE transporter. Atomoxetine had low affinity for a number of other radioligand binding sites, suggesting a high degree of selectivity for NE transporters.

Atomoxetine potently blocked in vivo depletion of NE in rats by DSP-4 in a dose-dependent manner, thus demonstrating in vivo blockade of NE transporters. Previous studies have shown DSP-4 depletes NE by a carrier dependent mechanism (Grzanna et al. 1989; Jonsson et al. 1981). Like atomoxetine, desipramine and reboxetine blocked depletion of NE by DSP-4, but fluoxetine and methylphenidate were not effective. Atomoxetine did not appreciably block p-CA–induced depletion of 5-HT, indicating that it does not significantly block 5-HT transporters in vivo at the doses evaluated.

The extracellular concentrations of NE in PFC were increased by atomoxetine, whereas 5-HTEX was not significantly altered by atomoxetine up to 3 mg/kg i.p. in the brain regions examined. Atomoxetine increased DAEX to about the same magnitude as NEEX in the PFC. The higher doses of atomoxetine produced longer lasting increases in NEEX and greater increases in DAEX than the low dose. However, atomoxetine did not increase DAEX in the DA and DA transporter-rich areas—nucleus accumbens and striatum (Soucy et al. 1997; Coulter et al. 1995). This report is in agreement with studies using reboxetine and desipramine that have shown selective NE uptake inhibitors increase DAEX as well as NEEX in PFC, but not in the nucleus accumbens (Tanda et al. 1994; Linnè r et al. 2001) or striatum (Carboni et al. 1990; Di Chiara et al. 1992). The transporters for NE are relatively abundant compared with DA transporters in the PFC (Gehlert et al. 1993; Soucy et al. 1997; Coulter et al. 1995; Sesack et al. 1998) and it has been determined that DA is taken up non-selectively by NE transporters in the PFC (Carboni et al. 1990; Di Chiara et al. 1992; Tanda et al. 1997; Yamamoto and Novotney 1998) as well as co-released (Devoto et al. 2001). The NE transporter has similar affinities for NE and DA (Raiteri et al. 1977) and presumably extracellularly-released DA may diffuse transsynaptically to NE transporters (Yamamoto and Novotney 1998). Consistent with this hypothesis, locally administered atomoxetine increased both NEEX and DAEX in the PFC, but not to the magnitude seen with systemic administration. Thus, the increase of the catecholamines is due, at least in part, to a local effect. The onset of increase in catecholamines after local infusion was rapid, suggesting that this is not due to diffusion of the drug to other brain areas.

However, other interactions may also be involved in the increase of DA induced by atomoxetine in PFC. The selective NE transporter inhibitors nisoxetine (Wong et al. 1975) and reboxetine (Wong et al. 2000), as well as α2-adrenergic antagonists that enhance release of NE (Dennis et al. 1987; Thomas and Holman 1991), increase burst firing of ventral tegmental DA neurons which would result in increased DA release in their terminal fields (Linnè r et al. 2001; Shi et al. 2000; Grenhoff and Svensson 1993).

The immediate-early gene c-fos and its protein products have been increasingly utilized as markers for neuronal activation (Dragunow and Faull 1989; Morgan and Curran 1990; Robertson et al. 1994). In the present study, immunohistochemical localization of Fos protein allowed the quantitation of activated cells in specific forebrain nuclei following vehicle or atomoxetine administration. Atomoxetine significantly and robustly increased the number of Fos-positive cells in the PFC, but not in the nucleus accumbens or the striatum. These findings are in direct correlation with the pattern of increased concentrations of DA induced by atomoxetine in the same forebrain structures. Therefore, these studies suggest that elevations in DAEX activate postsynaptic neurons and Fos expression in the PFC, but the involvement of NE cannot be excluded. In contrast, a study by Lin et al. (1996), found that administration of methylphenidate to cats induced Fos expression in the striatum. Thus, the lack of Fos induction in the nucleus accumbens and striatum may indicate a unique method of action for atomoxetine as compared with methylphenidate.

Reboxetine increased NEEX (Sacchetti et al. 1999) and DAEX in the PFC to similar levels and to about the same magnitude as found with atomoxetine, consistent with previous reports (Linnè r et al. 2001). In the PFC, methylphenidate caused large increases in NEEX and DAEX, but the increases were of short duration in agreement with previous studies (During et al. 1992), and consistent with a short duration of action in humans. However, methylphenidate, unlike atomoxetine, also increased DAEX in nucleus accumbens which may mediate the methylphenidate-induced increases in locomotor activity and may also be involved in the rewarding aspects of the drug (Kuczenski and Segal 1997, 1999). Furthermore, methylphenidate-induced increases in DAEX in striatum may stimulate motor tracts in that area causing stereotyped behaviors and motor disturbances such as tics in humans. Similarly, imaging studies in humans have shown that methylphenidate increases extracellular levels of DA in the striatum (Volkow et al. 2001).

Atomoxetine is a development candidate for pharmacotherapy of ADHD and has been shown in early studies to be efficacious and well-tolerated for use in this disorder (Spencer et al. 1998, 2001; Michelson et al. 2001). We propose that the atomoxetine-induced increase of DAEX and NEEX in the PFC, and presumably other cortical areas, enhances catecholaminergic neurotransmission in the cortex and this may be a mechanism of action of atomoxetine in the pharmacotherapy of ADHD. Catecholamines, particularly DA, are highly involved in ADHD and enhancement of DA and NE neurotransmission in the PFC by psychostimulants and NE uptake inhibitors may play a pivotal role in the efficacy of these drugs in ADHD (Spencer et al. 1995; Biederman and Spencer 1999; Arnsten et al. 1996). The observed increase in DA transporters in brains of ADHD patients (Dougherty et al. 1999; Krause et al. 2000; Dresel et al. 2000) presumably results in decreased DA neurotransmission which may be offset by drugs that enhance dopamine neurotransmission. However, DA transporters in ADHD patients have not been shown to be increased in cortical areas due to the difficulty of imaging the low levels of DA transporters in those regions. If DA transporters are increased in cortical areas, then this may be offset by blockade of NE transporters by atomoxetine.

Enhanced catecholaminergic neurotransmission in the PFC by atomoxetine may affect attentional processes possibly by activation of DA1 receptors known to mediate working memory in the PFC (Sawaguchi and Goldman-Rakic 1994; Arnsten 1998). Activation of prefrontal cortical areas may also impact neuronal tracts projecting to subcortical areas involved in regulation of DA neurotransmission (Taber and Fibiger 1993; Taber et al. 1995). In support of the role for frontal cortex in ADHD-like behaviors, lesions of the frontal lobes cause a breakdown of goal directed activity, executive function, attention and produce hyperactivity (Benson and Stuss 1982; Petrides and Milner 1982). In imaging studies, the right frontal lobes including PFC (Hynd et al. 1990; Castellanos et al. 1996b) and the caudate nucleus of children with ADHD were smaller in volume than controls, possibly suggesting a neurodevelopmental lag in the maturation of the associated neuronal pathways and their connectivity (Castellanos et al. 1996b). Thus, increasing catecholaminergic neurotransmission in cortical areas may be involved in the efficacy of psychostimulants and atomoxetine in ADHD.

The use of selective inhibitors of NE transporters such as atomoxetine would be expected to have several advantages over the psychostimulants now commonly used for therapy of ADHD. Psychostimulants may induce dose-limiting and intolerable anxiety and dysphoric mood in some patients and do not satisfactorily alleviate symptoms of comorbid conditions such as depression and anxiety (Wilens and Spencer 2000) and may exacerbate those symptoms. Atomoxetine blocks NE uptake in brain and the antidepressant activities of NE transporter inhibitors such as desipramine have been clearly demonstrated (Delgado et al. 1993). Thus, atomoxetine may alleviate symptoms of comorbid depression and anxiety due to enhanced NE neurotransmission. In regard to adverse events compared with the psychostimulants, atomoxetine would be expected to have decreased motor effects such as induction of tics due to its lack of effect on DAEX in the striatal motor areas. Furthermore, since atomoxetine does not increase DAEX in the nucleus accumbens, a region associated with psychostimulation and rewarding behaviors, atomoxetine would not be expected to have drug abuse liability as found with the psychostimulants (Kuczenski and Segal 1997). In fact, atomoxetine did not substitute appreciably for methamphetamine in drug discrimination studies in monkeys (Tidey and Bergman 1998).

In conclusion, the selective inhibitor of NE transporters, atomoxetine, increases NEEX and DAEX in the PFC, but does not alter DAEX in nucleus accumbens and striatum. We have proposed the efficacy of atomoxetine in ADHD is related to its enhancement of cortical NE and DA neurotransmission, particularly in PFC.