INTRODUCTION

Some patients do not respond to an initial antidepressant (AD) treatment or the treatment loses efficacy over time. Regardless of the class of AD administered, 30–45% of patients fail to respond adequately to drug treatment (Fava, 2000). One strategy that has been widely adopted for treatment of nonresponsive depressed patients is to switch from the initial type of AD used, such as a selective serotonin reuptake inhibitor (SSRI), to another class, such as a selective norepinephrine reuptake inhibitor (NRI) (Thase and Rush, 1997; Fredman et al, 2000; Thase et al, 2002; Rush and Ryan, 2002). The regulatory effects of switching from SSRIs to selective NRIs on serotonergic and noradrenergic parameters have not yet been explored preclinically (Thase and Rush, 1995; Fredman et al, 2000; Rush and Ryan, 2002). It would be useful to know, for instance, if the effects of the second drug were enhanced, accelerated or diminished by prior treatment with the first drug. We therefore examined the effects of switching from the SSRI, paroxetine (PRX), to the selective NRI, reboxetine (RBX), on serotonergic and noradrenergic parameters that have been shown previously to be altered by treatment with either of these drug classes alone.

Regulatory changes observed in rat brain after 2–3 weeks of SSRI treatment include a decrease in serotonin transporter (SERT)-binding sites (Benmansour et al, 1999,2002). Similar regulatory changes have been observed for the norepinephrine transporter (NET) after chronic treatment with the selective NRI, desipramine (Frazer and Benmansour, 2002), but have not yet been examined with RBX. In vitro, RBX is about 120-fold more potent at blocking norepinephrine (NE) reuptake than serotonin (5-HT) reuptake (Wong et al, 2000). After either acute or chronic treatment, RBX increases extracellular NE levels in rat brain (Sacchetti et al, 1999; Invernizzi et al, 2001), while producing no effect on extracellular 5-HT (Millan et al, 2001). Considering the specificity of RBX for the noradrenergic system, we examined its long-term regulatory effects on the NET and noradrenergic receptors. We predicted that chronic RBX treatment, like desipramine, would decrease NET-binding sites in the rat brain. Also, radioligand binding to β1 adrenergic receptors is decreased after chronic treatment of rats with NRIs (Ordway et al, 1988,1991; Nelson et al, 1991). We therefore anticipated that chronic RBX treatment would similarly result in a decrease in β1 adrenergic receptor binding. Further, a compensative regulatory effect observed in the locus coruleus after chronic treatment with AD drugs affecting the noradrenergic system such as desipramine, nortriptyline, and imipramine is a decrease in tyrosine hydroxylase (TH) gene expression levels (Nestler et al, 1990; Brady et al, 1991; Morilak et al, 1998). We predicted that chronic RBX treatment might also result in reduced mRNA expression for TH. Overall, we anticipate that selective NRIs, such as RBX, will have regulatory effects on the noradrenergic system that parallel the effects that SSRIs, like PRX or sertraline, have on the serotonergic system. Thus, we measured the effects of treating rats with PRX or RBX, either alone or sequentially, on radioligand-binding sites for the SERT, NET, and β1 adrenoceptors.

It was also of interest to determine if parameters not observed to change after monotherapy with noradrenergic reuptake inhibitors would be altered following sequential treatment. Radioligand binding to α2 adrenergic receptors in terminal fields appears to be unaltered by treatment with NE reuptake inhibitors (Kovachich et al, 1993; Jimenez-Rivera et al, 1996). Also, while α1 adrenergic responsivity can be affected by chronic AD treatment (Rogoz and Kolasiewicz, 2001), radioligand binding to α1 adrenergic receptors has not been reported to be altered by such treatments. Consequently, we measured radioligand binding to sites for α1 and α2 adrenergic receptors in terminal fields.

MATERIALS AND METHODS

Animals and Chronic Drug Treatments

All animal procedures were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Texas Health Science Center at San Antonio in accordance with NIH guidelines, and efforts were made to minimize animal use and suffering during the experiment. Male Sprague–Dawley rats (Harlan, Indianapolis, IN), weighing 175–200 g, were housed three per cage with ad libitum access to rat chow and water in a temperature- and humidity-controlled environment, with a 12 h light/dark cycle (lights on at 07.00 h) for 1 week prior to surgery. At 1 day prior to surgery, osmotic minipumps delivering 2.5 μl/h (Model 2ML4, Alza Corporation, Palo Alto, CA) were filled with drug or vehicle (see Table 1), filtered through 0.8-μm nitrocellulose filters (Millipore, Bedford, MA) using a sterile technique under an air-filtered hood. Drug solution concentrations were determined based on a mean rat weight of 280 g for the first 21 days of treatment. Minipumps were then stored in sealed containers filled with sterile saline in a 30–32°C incubator to prime the pumps until the time of surgery.

Table 1 6-Week Drug Treatment Groups, Doses, and Sample Sizes
Full size table

At the time of surgery, the rats (weighing 200–250 g) were anesthetized with an intramuscular injection of a cocktail of 34 mg/kg ketamine, 1 mg/kg acepromazine, and 7 mg/kg xylazine in saline. A drug-filled minipump was implanted subcutaneously on the rat's dorsum, approximately 3 cm below the neck, and the rat was injected with penicillin. After 21 days of drug treatment, the minipumps were removed, and a new minipump with a second 3-week treatment of drug or vehicle was implanted. The approximate weight of rats at the time of surgery for the second treatment was 320 g, so drug concentrations were based on an estimated body weight of 380 g for the second 21-day treatment. Table 1 summarizes pretreatments, treatments, and sample sizes for each experimental group.

Selection of the dose of RBX was determined from a pilot study that aimed to determine steady-state serum levels after 6–12 days of drug administration via minipump. Six concentrations of RBX were tested to determine the dose needed to obtain a steady-state serum concentration within the therapeutic range observed in humans. RBX doses of 30, 20, 10, 5, and 2.5 mg/kg/day produced serum values of 1196±122, 191±38, 86±18, 40±7, and 20±2 ng/ml, respectively. The RBX concentration used in this study was 20 mg/kg/day, which produced a serum concentration close to human therapeutic values (Fleishaker et al, 1999). As the affinity of RBX for the rat and human NET is comparable (Wong et al, 2000), attaining ‘therapeutic’ serum concentrations is relevant for a study in rats. Selection of the dose of PRX was based on a similar rationale and results from a previous study (Benmansour et al, 1999). Washout times needed after drug treatments to establish serum levels of drug <5 ng/ml were also obtained from Benmansour et al (1999) for PRX (96 h) and were determined experimentally for RBX (48 h). Saline was the vehicle control for RBX, whereas 50% ethanol was the vehicle control for PRX. On all parameters measured, values obtained in the two control groups were not significantly different, so values are presented for the combined control groups.

Upon completion of the drug treatments, the minipumps were removed. After the appropriate washout times, rats were euthanized by decapitation and their brains were removed, rinsed in ice-cold saline, frozen quickly on dry ice, and stored at −80°C until use.

Brain Tissue Preparation

Coronal sections (20 μm) of frozen rat brain were cut on a –16°C cryostat (Leica Frigocut, 2800 N, Deerfield, IL) at the level of plate 18–20 for the bed nucleus of the striata terminalis, plate 28–33 for amygdala and hippocampus, plate 47–50 for dorsal raphe, and plate 56–59 for locus coeruleus, according to the stereotaxic atlas of Paxinos and Watson (1986). Sections for quantitative autoradiography were thaw-mounted onto gelatin or silane-coated, chilled microscope slides and dried under vacuum for 12 h at 4°C. Sections for in situ hybridization were thaw-mounted onto silane-coated, chilled microscope slides, fixed in 4% paraformaldehyde, washed in phosphate-buffered saline, dehydrated in ethanol, and air-dried. Sections on slides were stored at −80°C until use in assays.

Quantitative Autoradiography

The densities of NET, α1, α2, and β1 noradrenergic receptor-binding sites, as well as those for the SERT were examined by quantitative autoradiography. Table 2 summarizes these procedures and the materials used. Preliminary experiments showed equivalent binding of either [3H]cyanoimipramine or [3H]nisoxetine in brain slices of chronic drug-treated rats if there was no preincubation or preincubations up to 60 min. For the NET and SERT incubation solutions, radioligand concentrations were 5–9 times the ligand's KD value, so as to obtain binding site densities approximating Bmax values. Incubation solutions of the iodinated ligands were at the ligand's KD value and therefore, binding site densities represent nonmaximal values. Large incubation volumes (10 ml) were used for all incubations that would produce dilution of any residual drug coming out of the brain slices. Following incubation, the sections on slides were washed in ice-cold buffer, dipped for 1 s in ice-cold distilled-deionized water to remove buffer salts, and dried on a slide warmer (Labline, Fisher Scientific, USA).

Table 2 Quantitative Autoradiography Methods Summary Table
Full size table

All quantitative autoradiography sections on slides were exposed to Hyperfilm-3H (CEA AB Sweden for Amersham) along with tritium standards (American Radiolabeled Chemicals, St Louis, MO) calibrated against brain mash for [3H] as per Geary et al (1985) or [125I] as per Artymyshyn et al (1990). Films were developed using Kodak GBX developer, stop bath, and fixer. Autoradiograms were analyzed using NIH image-based Scion Image 1.60c (public domain shareware http://scioncorp.com) on a Macintosh IICi computer. Optical densities of brain images were converted to femtomoles per milligram as per Geary et al (1985) or Artymyshyn et al (1990).

MANOVA tests were used to analyze the effects of the chronic drug treatments on all brain regions quantified using Statistica for the Macintosh (Statsoft, Tulsa, OK). Significant multivariate results were examined post hoc by ANOVA followed by Neuman–Keuls post hoc test.

In Situ Hybridization with TH Riboprobes

In situ hybridization was conducted as described previously (Sands et al, 2000a). The 275 nt TH probe was transcribed from cDNA templates provided by Dr Stanley Watson (University of Michigan), with addition of α-35S-UTP (New England Nuclear) to a specific activity of 2 × 109 cpm/μg. All prehybridization and hybridization solutions were treated with diethyl pyrocarbonate. Brains sections were hydrated, acetylated, rinsed in 2 × sodium chloride–sodium citrate buffer (SSC), dehydrated in ethanol, delipidated in chloroform, and air-dried. Sections were hybridized overnight at 60°C under glass coverslips in 50 μl hybridization buffer containing 4 × 107 cpm/ml radiolabeled riboprobe (approximately 20 ng/ml). All posthybridization solutions contained 1 mM dithiothreitol. After hybridization, sections were rinsed, and excess probe was removed by digestion with RNAse A (20 μg/ml, 30 min at 37°C), followed by a series of washes of increasing stringency: 10 min each in 1 × SSC, 0.5 × SSC, 0.2 × SSC. Sections were washed for 10 min in 0.1 × SSC and 3 × 1 h in 0.1 × SSC, 15% formamide, at 60°C. Sections were then rinsed, dehydrated, and apposed to Kodak Biomax MR X-ray film along with a set of 14C-radioactive standards for 16 h.

After developing the films, signals overlying the LC on digitized autoradiographic images were analyzed densitometrically using NIH-Image (v. 1.55, Wayne Rasband, NIH). Integrated density was calibrated in units of nCi/mg using the 14C-standards, and measured in 4–6 LC sections per brain, corresponding to plates 56–58 in the atlas of Paxinos and Watson (1986). The mean integrated density was then calculated to generate a single value for each animal. Data obtained from brains in different hybridizations were combined into groups by expressing all individual values as a percentage of the mean baseline value for the 6-week PBS-treated control rats analyzed within the same hybridization. Resulting group data were compared using one-way ANOVA, followed by post hoc comparisons with Fisher's PLSD test. Significance was determined at P<0.05.

Liquid Chromatography Analysis of Serum Drug Levels

Serum concentrations of PRX were determined by high-performance liquid chromatography (HPLC) as described in Benmansour et al (1999). Serum RBX concentrations were determined by HPLC as described in Fleishaker et al (1999).

Drugs

RBX was obtained from Pharmacia (Kalamazoo, MI). PRX was obtained from SmithKline Beecham (Harlow, England). Prazosin, phentolamine, and isoproterenol were obtained from Sigma (St Louis, MO). Mazindol was obtained from Sandoz (East Hanover, NJ). Sertraline was obtained from Pfiser (Groton, CT). ICI 118,551 was obtained from Imperial Chemical Industries (Cheshire, England). [3H]nisoxetine (80–85 Ci/mmol) and [3H]idazoxan (82 Ci/mmol) were obtained from Amersham (Arlington Heights, IL). [3H]cyanoimipramine (80–85 Ci/mmol) was obtained from American Radiolabeled Chemicals (St Louis, MO). [125I]HEAT and [125I](−)iodopindolol, both 2200 Ci/mmol, were obtained from Perkin-Elmer-NEN (Boston, MA).

RESULTS

Quantitative Autoradiography of Noradrenergic Parameters

Treatment with RBX produced qualitatively similar reductions in the binding of [3H]nisoxetine in the noradrenergic terminal fields measured. For illustrative purposes, results found in the arcuate nucleus of the hypothalamus (ARC), the dentate gyrus (DG) of the hippocampus, and the lateral nucleus of the amygdala (LA) are shown in Figure 1. Also measured, but not shown, were results in the CA1 region of the hippocampus and the frontal parietal cortex. The decrease in NET binding after 3 weeks of treatment with RBX was significant in the ARC, LA, and CA1 (P<0.01), but not in the DG and cortex. Treatment with RBX for 6 weeks decreased NET-binding sites significantly in all areas measured (P<0.001). In rats treated with RBX for 3 weeks, there were significantly smaller effects on [3H]nisoxetine binding (about 40–45% reductions) than that seen after 6 weeks of treatment (50–60% reductions) in four of five brain regions measured (P<0.025), but not in the LA. Neither PRX treatment for 6 weeks, nor 3 weeks of PRX treatment followed by 3 weeks of saline had any significant effect on NET-binding sites (Figure 1).

Figure 1
figure 1

Representative autoradiograms for [3H]nisoxetine binding in terminal field areas. Typical autoradiograms for either saline or 50% ethanol vehicle controls, 3 and 6 weeks of RBX treatment, 6 weeks of PRX treatment, 3 weeks of PRX treatment followed by 3 weeks of RBX treatment, and nonspecific binding, defined by 10 μm mazindol are shown. FPCtx=frontal parietal cortex, L. Nu. Amyg=lateral nucleus of the amygdala.

Full size image

Interestingly, a 3-week treatment with PRX followed by a 3-week treatment with RBX produced decreases in [3H]nisoxetine binding significantly greater than that caused by 3 weeks of RBX alone in four of five brain regions measured (P<0.05, see Figure 1), the exception being in the LA. The effect of this sequential treatment was similar to 6 weeks of RBX treatment, with a decrease in NET binding of 60–70% in the terminal field areas.

The effects of the various treatments on β1 adrenergic receptor-binding sites in the basolateral nucleus of the amygdala (BLA) and the CA1 region of the hippocampus are shown in Figure 2. The LA and layers 1–3 of the frontal parietal cortex were also measured, but are not shown in Figure 2. Results in these areas were similar to those found in the BLA and CA1 region. RBX produced significant decreases of a similar magnitude, on the order of 40–60%, in the binding of [125I]iodopindolol to β1 adrenergic receptors in all terminal fields, both at 3 and 6 weeks of treatment (P<0.001, Figure 2). The sequential treatment of PRX followed by RBX produced decreases in [125I]iodopindolol binding not significantly different from that caused by 3 weeks of RBX treatment alone. Treatment with PRX for 6 weeks or treatment with the SSRI for 3 weeks followed by 3 weeks of saline did not significantly alter the binding of [125I]iodopindolol (Figure 2).

Figure 2
figure 2

NET density in representative brain regions. Bars indicate means±standard errors; sample size (N) is indicated within each bar. VEH.=vehicle control (50% EtOH or saline), PRX=paroxetine, SAL=saline, RBX=reboxetine. All brain regions measured had significant differences between treatment groups (P<0.001). Results in the arcuate nucleus of the hypothalamus (F5,65=22), dentate gyrus (F5,67=25), and the lateral nucleus of the amygdala (F5,67=23)are shown in this figure. Drug treatment groups indicated with an * exhibited [3H]nisoxetine binding that was significantly less than that in the vehicle control (P<0.001, Neuman–Keul's). RBX treatment groups indicated with a ♦ exhibited [3H]nisoxetine binding that was significantly less than that in the corresponding 3-week RBX treatment (P<0.05, Neuman–Keul's).

Full size image

In the terminal fields, none of the drug treatments produced any significant changes in radioligand binding to either α1 or α2 adrenergic receptors (see Table 3).

Table 3 Binding to α Adrenergic Receptors (fmol/mg Protein) in Representative Brain Regions, and Expression of mRNA for Tyrosine Hydroxylase in the Locus Coruleus (% of Normalized Control)
Full size table

Quantitative Autoradiography of the SERT

For illustrative purposes, the treatment-induced effects on [3H]cyanoimipramine binding to the SERT in two terminal field areas and one cell body area are shown in Figure 3. Chronic RBX treatment did not produce any significant decreases in [3H]cyanoimipramine binding in any region of the brain. By contrast, treatment with PRX for 6 weeks produced a 50–80% decrease (P<0.001) in binding in both serotonergic terminal areas and cell bodies, consistent with our previous data after 3 weeks of treatment (Benmansour et al, 1999). Rats treated with PRX for 3 weeks followed by either 3 weeks of saline or 3 weeks of RBX exhibited no significant decrease in [3H]cyanoimipramine binding. Also measured, but not shown, were the CA1 region of the hippocampus, arcuate nucleus of the hypothalamus, frontal parietal cortex, and the medial raphe nucleus. Results in these areas were very similar to those found in the areas shown (Figure 4).

Figure 3
figure 3

β1 adrenergic receptor binding in the basolateral nucleus of the amygdala and CA1 region of the hippocampus. See Figure 1 for abbreviations. All brain regions measured had significant differences between treatment groups (P<0.001). The CA1 region of the hippocampus (F5,67=14) and the basolateral nucleus of the amygdala (F5,65=30) are shown in this figure. Drug treatment groups indicated with an * exhibited [125I]iodopindolol binding that was significantly less than that in the vehicle control (P<0.001, Neuman–Keul's).

Full size image
Figure 4
figure 4

SERT density in representative brain regions. See Figure 1 for abbreviations. All brain regions measured had significant differences between treatment groups (P<0.001), and included the dentate gyrus (F5,66=11), lateral nucleus of the amygdala (F5,66=17), and the dorsal raphe nucleus (F5,64=12), shown in the figure. Only 6 weeks of PRX treatment produced a significant reduction in [3H]cyanoimipramine binding, which was significant in all regions measured (P<0.005, Newman–Keul's). Drug treatment groups indicated with an * exhibited [3H]cyanoimipramine binding that was significantly less than that in the vehicle control (P<0.001, Neuman–Keul's).

Full size image

In Situ Hybridization for TH

No significant main effect of treatment condition was observed on the level of mRNA expression for TH in the LC (F6,65=0.96, P=0.46; Table 3).

DISCUSSION

The primary observation in this study was that pretreatment with PRX enhanced the timing and magnitude of at least one regulatory effect of RBX treatment, namely downregulation of the NET. When 3 weeks of RBX treatment followed 3 weeks of prior treatment with the SSRI PRX, the resulting decrease in NET binding was similar to that seen after 6 weeks of RBX treatment (Figure 2). By contrast, on the other parameters measured, the effects of RBX were qualitatively similar, whether given alone or following pretreatment with PRX. There was no enhancement of RBX's inhibitory effects on β1 adrenergic receptor-binding site density, and there was no residual effect of the PRX pretreatment on the SERT-binding density in rats subsequently treated with RBX. However, it seems likely that 3 weeks of RBX treatment was sufficient to induce maximal decreases in β1 adrenergic receptor sites, thus any potential facilitation by PRX may not have been evident. This is suggested by the observation that 3 weeks of RBX decreased β1 adrenergic receptor binding to an equivalent extent as that caused by desipramine treatment in a previous study (Ordway et al, 1991) and to a extent similar to that caused by 6 weeks of RBX treatment (Figure 2).

By contrast with the ‘priming’ of RBX's regulatory effects on the NET that was precipitated by PRX pretreatment, the reverse effect was not observed, that is, we did not observe any ‘conversion of effect’ of RBX treatment. That is, RBX treatment did not sustain any regulatory effects that the initial PRX treatment may have exerted on SERT radioligand-binding sites. We know from previous studies that after cessation of treatment with the SSRI sertraline, SERT densities recover to pretreatment levels within 2 weeks (Benmansour et al, 2002). Evidently, RBX treatment neither prolonged or sustained the paroxetine-induced downregulation of the SERT, nor did it interfere with the recovery of SERT density in the timeframe of our observations.

PRX is a potent inhibitor of 5-HT reuptake, but it also inhibits NE reuptake with a greater potency than that of other SSRIs (Frazer, 2001). PRX exhibits a moderate affinity for the NET (Owens et al, 1997), and has been shown to have modest effects on the NET at serum concentrations >100 ng/ml (Owens et al, 2000; Gilmor et al, 2002). Nonetheless, PRX treatment alone for 6 weeks did not affect any of the noradrenergic parameters measured in this study. PRX had no effect on NET or β1 adrenergic receptor-binding sites, whereas drugs that block NE reuptake, that is, desipramine (Frazer and Benmansour, 2002) and RBX (Figures 1 and 2), did. Thus, this dose of PRX, which produces serum concentrations of the drug in the ‘therapeutic’ range (Benmansour et al, 1999), is unlikely to have exhibited any direct long-term regulatory effects on the NE system.

Similar to the tricyclic antidepressant desipramine, RBX has been shown to selectively increase extracellular NE in the cortex, hippocampus, and other brain regions, presumably through reuptake blockade, without affecting serotonin levels (Sacchetti et al, 1999; Invernizzi et al, 2001; Page and Lucki, 2002). Based on these studies, we anticipated that RBX would not affect the SERT, and it did not. Reduction of NET binding following 3 and 6 weeks of chronic treatment with RBX was similar to the downregulation of NET observed following chronic desipramine treatment (Frazer and Benmansour, 2002). The selective downregulatory effects on the NET observed after chronic RBX treatment are therefore consistent with its function as a selective NRI.

Other interactive effects of treatment with an SSRI and a selective NRI have been observed. For example, subacute simultaneous treatment of RBX and sertraline made recovery from both 8-OH-DPAT and clonidine-induced responses more rapidly than the time course of recovery with either treatment alone, suggesting that the two drugs together accelerated changes to the responsiveness of 5-HT1A and α2-adrenoceptors (Harkin et al, 1999). Chronic coadministration of desipramine and fluoxetine caused desipramine-mediated downregulation of β adrenergic receptors to occur more rapidly (Baron et al, 1988). In microdialysis studies, a 120 min acute pretreatment with fluoxetine potentiated the effects of acute administration of bupropion (a dual NE and dopamine (DA) reuptake inhibitor) to increase extracellular NE and DA concentrations in the hypothalamus, cortex, and nucleus accumbens (Li et al, 2002).

We did not observe any decrease in mRNA expression for TH after chronic SSRI or selective NRI treatment (Table 3). Findings of changes in gene expression for TH with chronic antidepressant treatment are variable. Nestler et al (1990) found that after 2 weeks of chronic treatment, all major classes of antidepressant treatment decreased the expression of mRNA for TH in the locus coruleus. By contrast, Brady et al (1991) found that either 2 or 8 weeks of administration of fluoxetine, idazoxan, and phenelzine produced an increase in the expression of mRNA for the TH in the locus coruleus. Brady et al (1991) also found that 8 weeks of imipramine treatment decreased TH mRNA expression by 40%. Chronic administration of desipramine for 6 weeks was shown to decrease TH mRNA in the LC (Morilak et al, 1998), but chronic RBX did not do the same in our current study. Perhaps differences in the mode of drug administration could have contributed to this difference. In previous studies, where mRNA expression for TH was increased, it may have been due to the stress of chronic daily injections. Both chronic stress and acute stress have been shown to increase mRNA expression for TH and TH activity in the locus coruleus (Brady, 1994; Sands et al, 2000b). Our use of osmotic minipumps may have reduced the stress associated with the injections, which may explain why neither RBX nor PRX increased mRNA levels. However, we did not observe a decrease either, as was found previously with desipramine (Morilak et al, 1998). Perhaps the reduction in mRNA expression for TH that was observed after desipramine treatment may have been triggered by its effects on other receptor types (ie H1 histaminergic, α1 adrenergic, muscarinic-cholinergic, or dopaminergic). If so, then RBX's lack of effect on such receptors (Wong et al, 2000) may account for its inability to elicit changes in TH gene expression. Likewise, the lack of effect of chronic RBX treatment on α1 adrenergic receptor binding (Table 3) was consistent with the findings of others (eg Rogoz and Kolasiewicz, 2001).

RBX had no effect on α2 adrenergic receptor density in the terminal fields, where both presynaptic autoreceptors and postsynaptic receptors coincide. The effect that NE-specific reuptake blockade may have on α2 adrenergic receptors of the cell body area of the locus coruleus remains unclear and warrants further investigation. Chronic imipramine treatment produced a modest decrease in the numbers of α2 receptors in the locus coruleus (Jimenez-Rivera et al, 1996). Thus, NE-specific antidepressant-induced decreases in the density of α2 receptors are more likely to be found in the cell bodies (Kovachich et al, 1993). However, consistent with the lack of effect of RBX on terminal α2 receptor density in the present study, chronic desipramine treatment failed to produce downregulation of α2 receptors in the locus coruleus (Sacchetti et al, 2001). Nevertheless, chronic RBX treatment has been shown through microdialysis studies to desensitize presynaptic α2 adrenoceptors (Invernizzi et al, 2001). Linner et al (1999) demonstrated that acute imipramine treatment profoundly inhibited the level of firing in LC neurons, and this effect was only slightly attenuated after chronic imipramine treatment. They observed a corresponding increase in cortical NE levels with chronic imipramine treatment, which they attributed to desensitization of presynaptic α2 autoreceptors (Linner et al, 1999). Our data suggest, however, that there may be an additional mechanism responsible for the observed increase in extracellular NE with chronic NRI treatment. In addition to continuous pharmacologic blockade of the transporter and possible desensitization of α2 adrenoceptors, downregulation of the NET may also be involved in elevated extracellular NE. Evidence in support of such an interpretation has been obtained in investigations of the SERT in vivo. Decreased clearance of serotonin by the SERT was much more robust following downregulation of the SERT than that due to blockade of the SERT by SSRIs (Benmansour et al, 2002).

Overall, our findings suggest that switching from PRX to RBX may have ‘primed’ the noradrenergic system for downregulation of the NET by RBX through unknown 5-HT/NE system interactions. It is useful to know that switching from an SSRI to a selective NRI does not appear to impose any unforeseen challenges to perhaps important regulatory effects of the selective NRI. Thus, switching a nonresponding patient from an SSRI to a selective NRI such as RBX may be a viable strategy insofar as no regulatory effects of RBX on the noradrenergic system are compromised and there may even be temporal enhancement on some parameters.

While to date there have been few controlled, randomized clinical trials examining the effects of switching antidepressants in nonrespondent patients, a new initiative, Sequenced Treatment Alternatives to Relieve Depression, endeavors to do so for patients (Rush and Ryan, 2002). In turn, preclinical studies such as this one, examining the effects of sequential treatment with different classes of antidepressants on neurotransmitter regulation in the brain, can help to elucidate underlying mechanisms responsible for regulatory changes in 5-HT and/or NE neurotransmission.