*Jun'ichi Semba, MD, Department of Psychiatry, Saitama City Hospital, 2460 Mimuro, Midori-ku, Saitama 336-8522, Japan. Email: email@example.com
Aim: Epidemiologic studies suggest the existence of a biological link between nicotine withdrawal and depression. To investigate the neuronal mechanisms of the precipitation of depression during smoking cessation, an animal model of nicotine withdrawal was used, and the expression of serotonin transporter (5HTT), abnormality of which is implicated in the pathogenesis of depression, was investigated. The effect of co-administration of bupropion, which has been clinically shown to ameliorate nicotine withdrawal symptoms, was also investigated in this model.
Methods: Male Wistar rats were implanted with a minipump s.c., which delivered nicotine at a rate of 6 mg/kg per day for 12 days (days 1–12). Rats given chronic nicotine were killed on day 13, or 2 days after the removal of minipump (withdrawal day 2). In a separate experiment, bupropion (15 or 30 mg/kg per day) was injected into the nicotine infused rats on days 2–12. The expression of mRNA for 5HTT in the dorsal raphe was determined on in situ hybridization.
Results: Chronic nicotine infusion resulted in the reduction of 5HTT mRNA expression, which lasted through withdrawal day 2. Co-administration of bupropion, however, significantly antagonized this reduction.
Conclusions: Chronic nicotine infusion reduces the synthesis of 5HTT protein, which may consequently precipitate depression during nicotine withdrawal, but co-administration of bupropion may ameliorate withdrawal symptoms by counteracting nicotine's effect on 5HTT.
EPIDEMIOLOGICAL STUDIES OF smokers have suggested a close link between nicotine consumption and depression.1–3 Smoking cessation often precipitates depressive symptoms, even in patients without a history of depression.4–6 Severe major depressive episodes after smoking cessation appear 2 days–6 weeks after the initial abstinence from smoking.1 This suggests that chronic tobacco smoking and abrupt cessation may elicit neurochemical changes in the brain producing an episode of major depression, but the biochemical mechanism of how nicotine withdrawal precipitates depression is still unclear. Because various lines of evidence have suggested alterations of serotonin (5HT) function during nicotine withdrawal,7 5HT may play a key role in the precipitation of depression in this period.
The uptake of serotonin from the synaptic cleft by 5HT transporter (5HTT) proteins is a major mechanism for terminating the activity of a transmitter. There is a close relationship between 5HTT function and depression as follows: (i) most antidepressants have an inhibitory activity on 5HTT; (ii) recent brain imaging studies show decreased binding of 5HTT using positron emission tomography8 or single photon emission computed tomography9,10 in patients with depression; and (iii) analyses of post-mortem brains from suicide victims, most of whom are assumed to be affected with depressive disorders, indicate decreased expression of 5HTT mRNA in the dorsal raphe,11,12 decreased 3H-imipramine binding in the frontal cortex and hippocampus,13,14 and decreased 3H-citalopram15 or 3H-cyanoimipramine binding in the frontal cortex.16
In the present study we hypothesized that the precipitation of depression during nicotine withdrawal resulted from the impaired function of 5HTT. To explore this hypothesis, we used an animal model of nicotine withdrawal in which continuous nicotine infusion was ceased abruptly via an s.c. implanted minipump.17–20 We investigated the function of 5HTT by measuring the expression of 5HTT mRNA in the dorsal raphe in both rats given chronic nicotine and those being forced to withdraw from nicotine administration, using in situ hybridization. Furthermore, we examined whether co-administration of bupropion, which has been clinically indicated to prevent depressive episodes during smoking cessation,21 had an antagonistic effect in this animal model.
Male Wistar rats weighing 210–230 g were used. The animals were housed under a 12-h light–dark cycle (lights on, 07.00–19.00 hours) with free access to food and water. All animal experiments were carried out under the control of the Guidelines for Animal Experiments of the University of the Air in accordance with Japanese Federal Law (No. 105) and Notification (No. 6) of the Japanese Government.
In an acute nicotine experiment, rats received 0.5 mg/kg nicotine (as freebase) or saline i.p. 2 h before death. In a chronic nicotine experiment, rats were anesthetized with pentobarbital, and an osmotic minipump (Alzet model 2ML2; Alza, Cupertino, CA, USA) filled with nicotine was implanted s.c. behind the shoulder. (-)Nicotine di-tartrate was dissolved with saline, and the pH was adjusted to 7.0–7.5 with NaOH. The minipump was designed to deliver nicotine at a constant rate (6 mg/kg per day as a free base) for 12 days (days 1–12). Control animals were implanted with a minipump filled with neutralized tartrate solution. Plasma nicotine levels achieved with this administration model are known to resemble those seen in heavy smokers.19,20 In a withdrawal experiment, the pump in rats given chronic nicotine was removed under light ether anesthesia, and the rats were allowed to go through spontaneous withdrawal. In a bupropion experiment, bupropion (15 or 30 mg/kg per day, i.p.) was co-administered to nicotine-infused rats beginning 2 days after the pump implantation. Control rats received an equivalent volume of saline.
Non-radioactive in situ hybridization
Rats were killed by decapitation on day 13 or 2 days after the removal of minipump (withdrawal day 2). Rat brains were promptly removed, frozen in powdered dry ice, and stored at −80°C. Coronal brain sections of 12-µm thickness were cut in a cryostat, kept at −20°C, and mounted on silane-coated glass slides. The sections were dried and kept frozen until use. The RNA probe for rat 5HTT (plasmid generously donated by Dr R. D. Blakely, Vanderbilt University)22 consisting of a 6505-bp XbaI/EcoRI fragment was subcloned into pGEM-4Z. An antisense digoxigenin-labeled RNA probe was prepared using digoxigenin-11-UTP (Roche Diagnostics, Basel, Switzerland) and T7 polymerase to transcribe the 5HTT insert cDNA.
Non-radioactive in situ hybridization was performed as previously described.17 In order to avoid inter-assay variance, all sections of one brain region from each treatment group were processed together. The sections were post-fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS, pH 7.4) for 5 min. They were then rinsed in PBS and acetylated in 0.1 mol/L triethanolamine (pH 8.0)/0.25% acetic anhydride. Hybridization was performed with a digoxigenin-labeled RNA probe at 42°C for 16 h. After hybridization, the slides were washed twice in 2 × SSC, 50% formamide at 55°C for 15 min, and treated with 20 µg/mL RNase A at 37°C for 40 min. They were rinsed in 2 × SSC, 50% formamide at 55°C for 15 min, twice in 1 × SSC, 50% formamide at room temperature for 15 min, then in 2 × SSC, and finally in Tris-buffered saline. For immunological detection of digoxigenin, the slides were pre-blocked in 1% blocking reagent, and incubated with anti-digoxigenin conjugated to alkaline phosphatase at 4°C for 30 min. After incubation, the slides were washed and developed in a solution of NBT/BCIP (Roche Diagnostics) for 24 h. The color reaction was stopped using 1 mmol/L ethylenediamine tetra-acetic acid, 10 mmol/L Tris-HCl. After being washed in distilled water, the slides were dehydrated and cover-slipped.
The sections of each brain region were processed simultaneously to allow direct comparison between groups using the same digoxigenin-labeled probes. Only background labeling was observed when hybridization was conducted with the corresponding sense probes.
Analysis of hybridization signal
NIH Image 1.6 software (Wayne Rasband, NIMH, Bethesda, MA, USA) was used to collect images for semi-quantitative examination of 5HTT mRNA expression. Sampled areas were digitized through a microscope with a CCD camera monitored using a Macintosh computer. The NIH image outline tool was used to circumscribe the subregions of the dorsal raphe. The optical densities of areas of interest were measured on at least three sections from each subject, and converted to relative optical densities (ROD) by the formula of ROD = log (256/levels). The background was subtracted by measuring ROD of the periaquaductal gray.
The corrected gray levels for each brain region were statistically analyzed using Student's t-test or one-way ANOVA followed by the Neuman–Keuls multiple comparison test.
The representative microphotographs of mRNA-positive cells for 5HTT in the dorsal raphe of rats given chronic nicotine are shown in Fig. 1. In an acute nicotine experiment, single injection of 0.5 mg/kg of nicotine did not influence the expression of mRNA for 5HTT (saline control, 0.117 ± 0.009, n = 5; acute nicotine, 0.124 ± 0.006, n = 4; t = 0.62, d.f. = 7, P > 0.05, Student's t-test). In a chronic nicotine experiment, the expression of 5HTT mRNA in rats given chronic nicotine was significantly decreased (−24.7%) compared with that in control rats (Fig. 2a; t = 3.18, d.f. = 14, P < 0.01, Student's t-test), and to a lesser extent, this decrease (−14.9%) was observed even on withdrawal day 2 (Fig. 2b; t = 2.36, d.f. = 14, P < 0.05). As Fig. 3 shows, the reduced expression of 5HTT mRNA induced by chronic nicotine administration was dose-dependently antagonized by daily co-injection of bupropion (F = 3.22, d.f. = 3, 27, P < 0.05, one-way ANOVA) with a significant recovery at a higher dose of 30 mg/kg (P < 0.05, Neuman–Keuls test) but not at 15 mg/kg (P > 0.05).
Our main finding was that chronic nicotine administration decreased the expression of 5HTT mRNA in the raphe nucleus, and this decrease was antagonized by co-administration of bupropion. This phenomenon was not due to an acute nicotine effect, because a single injection did not influence the expression of 5HTT mRNA. The mechanism of reduced expression of 5HTT after chronic nicotine administration could be explained by the functional interaction between nicotine and 5HT in the brain.
The effect of nicotine on 5HT function has been extensively studied.24 It is widely accepted that the acute systemic administration of nicotine enhances the release of 5HT in several brain regions of rats.25–27 In contrast to the acute treatment, chronic nicotine administration decreases the concentration and biosynthesis of 5HT in the hippocampus or frontal cortex.28,29
In human studies Malone et al. have reported a correlation between cigarette smoking and impaired serotonin function in depressed patients by measuring cerebrospinal fluid 5-hydroxyindoleacetic acid level and investigating prolactin response to fenfluramine.7 Reduced serotonergic function in the brain is also suggested by studying the post-mortem brains of depressed suicide victims; they have lower brainstem levels of 5HT and 5HIAA, and fewer 5HTT binding sites in the prefrontal cortex.12In vivo imaging and post-mortem brain studies conducted in depressed patients suggested that reduced 5HTT binding is widespread in the brain.12,13,16 Although the synthesis of 5HTT protein does not always reflect the expression of its mRNA, this could be caused by the reduced 5HTT gene expression in 5HT neurons in the raphe nucleus.
In contrast to the present study, Awtry and Werling reported that chronic nicotine exposure significantly increased [3H]5HT uptake and [3H]paroxetine binding in the synaptosomes prepared from the prefrontal cortex and hippocampus of rats.30 They suggested that chronic nicotine increases 5HT uptake due to an increase in 5HTT density. The discrepancy between their findings and the present ones could be partially ascribed to the use of different tissue preparations. They used the synaptosomal preparations from brain regions, which do not always reflect in vivo function. For example, Rattray et al. reported biphasic changes in the level of 5HTT mRNA with no modifications in the concentration of the 5HTT protein labeled by [3H]-citalopram in the dorsal raphe of rats during the first 2 weeks after an acute treatment with the 5HT synthesis inhibitor p-chlorophenylalaine.31 Recently, Slotkin et al. found no significant changes of 5HTT binding in the frontal cortex and brainstem 3 days after nicotine withdrawal in adolescent rats, although the effects of chronic nicotine in the adolescent may be different from those in the adult.19 In addition, Xu et al. have already demonstrated no significant changes in [3H] paroxetine binding or 5HT turnover (5HIAA/5HT ratio) in the cerebral cortex comprising major 5HT terminal fields, as well as the brainstem, which contains the majority of 5HTT cell bodies, following chronic nicotine administration.32 But because we measured only 5HTT mRNA in the cell bodies, the function of terminal 5HTT protein may not directly parallel the expression of 5HTT mRNA. For example, cortical 5HT is likely to be indirectly regulated by nicotinic receptors located on cortically projecting cell bodies in the dorsal raphe nucleus.33 In contrast, there is a direct effect of nicotine on 5HT release in the striatum.34 In the hippocampus, nicotinic–serotonergic interaction is much more complicated, as demonstrated by Seth et al.24
We hypothesized that serotonergic function played a role in the precipitation of depression during nicotine withdrawal through the following mechanism. Continuous nicotine exposure might reduce 5HT release and synthesis in the brain, as shown in animal and human studies.7,28,29 Reduced availability of 5HT might lead to reduction in the expression of 5HTT mRNA, which would cause fewer transporter sites. During nicotine consumption, however, this decrease of 5HT transmitter in the synaptic cleft might be compensated by an intermittent release of 5HT induced by frequent nicotine ingestion. Once nicotine ingestion was stopped abruptly (i.e. smoking cessation), this compensation would be disrupted and 5HT availability would be decreased in a synaptic cleft. Lowered availability might persist at least for a few days, as we found in the present study (5HTT mRNA expression was low even during the withdrawal period). Eventually, reduced 5HT function might precipitate depression during nicotine withdrawal. This hypothesis could also explain an epidemiological finding that people with a past history of depression have a heightened risk of relapsing into a new episode of major depression during nicotine withdrawal.5,6
Bupropion, which was developed originally as an antidepressant, is now approved as non-nicotine pharmacotherapy for smoking cessation.21,35 A unique feature of bupropion for smoking cessation is that it can be started before the target quit date.36–38 Smoking cessation using the combination of nicotine patches and bupropion has a higher success rate than that using placebo.37,39,40 In our animal study, co-administration of bupropion with nicotine prevented the decrease of 5HTT mRNA expression. Using a similar nicotine-infusion model, Malin et al. demonstrated that co-infusion of bupropion attenuates nicotine abstinence syndrome in rats.41 But the effect of bupropion on 5HT function is not clear,42 because it does not have a significant affinity against 5HTT sites43 and does not affect extracellular 5HT in rat brain.44 Previously, bupropion's mechanism of action in the treatment of nicotine dependence has been considered to involve its modest blockade of dopamine and norepinephrine re-uptake.45 Recent studies, however, have suggested that it also inhibits the function of nicotinic receptors in human or mice as a non-competitive nicotine receptor antagonist.46,47 Bupropion's antagonistic action on the reduction of 5HTT mRNA induced by chronic nicotine administration may be indirectly mediated by its antagonist-like action on nicotine receptors. Otherwise, the expression of 5HTT mRNA could be regulated indirectly by dopamine and norepinephrine through their own pathways. Slotkin and Seidler found a close relationship between 5HT and serotonin in the striatum after chronic nicotine administration, as shown by measuring 5HT and dopamine levels and their turnover.20 A previous study demonstrated that the dopaminergic agonist apomorphine increases 5-HT content of dorsal raphe cell bodies, which receive a prominent dopamine input.48 In addition, a recent electrophysiological study by Haj-Dahmane showed that 5HT neurons in the dorsal raphe were activated by dopamine D2 receptor stimulation.49
In conclusion, the present study suggests that chronic nicotine administration decreased the expression of 5HTT mRNA, and co-administration of bupropion counteracted this decrease. Although the mechanism of reduced expression of 5HTT or bupropion's prophylactic action was not clear, there might be a close interaction among smoking cessation, development of depression and 5HT function in the brain.
This study was partly supported by a grant from the Smoking Research Foundation, Tokyo. We wish to thank Ms K. Yoshida and Mr N. Sakuta for their excellent technical assistance.