OUP user menu


Sabah Kelaï, Franck Aïssi, Klaus Peter Lesch, Charles Cohen-Salmon, Michel Hamon, Laurence Lanfumey
DOI: http://dx.doi.org/10.1093/alcalc/agg095 386-389 First published online: 1 July 2003


Knock-out mice lacking the serotonin transporter [5-hydroxytryptamine transporter (5-HTT)] were used to assess the influence of 5-HT re-uptake on ethanol consumption. Under a free-choice paradigm, alcohol intake was lower in mutant than in wild-type mice, and pharmacological blockade of 5-HTT by fluoxetine reduced alcohol intake in wild-type mice only. These data confirm the inhibitory effect of 5-HTT inactivation on ethanol intake.


Central serotoninergic neurotransmission has been implicated in the aetiology of alcohol dependence (Heinz et al., 2001), and there is a large body of evidence in favour of the existence of an inverse relationship between brain serotonin [5-hydroxytryptamine (5-HT)] tone and alcohol consumption: low brain 5-HT levels correlate with high ethanol intake (Murphy et al., 1982; Higley et al., 1996), and drugs that increase extracellular 5-HT levels either by preventing 5-HT re-uptake, such as selective serotonin re-uptake inhibitors (SSRIs) (Borg et al., 1985; Naranjo et al., 1990; LeMarquand et al., 1994; Maurel et al., 1999), or by stimulating 5-HT release, such as dex-fenfluramine (Higgins et al., 1992), attenuate ethanol intake in animal models of alcoholism.

In addition, the gene coding for the 5-HT transporter (5-HTT) has been considered as a candidate gene in numerous studies of alcohol dependence (Gorwood et al., 2000; Heinz et al., 2000; Preuss et al., 2000, 2001; Thompson et al., 2000), and treatment with SSRIs (i.e. zimeldine, citalopram, fluoxetine and fluvoxamine) was reported to decrease the desire to drink alcohol in alcohol-dependent subjects (Lejoyeux, 1996). However, not all studies have confirmed that this effect persists throughout treatment (Gorelick and Paredes, 1992), and the usefulness of SSRIs to help alcoholic subjects to reduce their alcohol intake is still a matter of debate (Naranjo and Knoke, 2001). Because knock-out mice with targeted disruption of the 5-HTT gene can be considered as a model of whole-life treatment with SSRI (Bengel et al., 1998), they offer a relevant opportunity to assess whether or not long-term inactivation of 5-HT re-uptake yields a sustained decrease in alcohol intake. Using a free-choice procedure, we thus compared the spontaneous ethanol consumption in these mutants (5-HTT–/–) and paired wild-type (5-HTT+/+) mice, and investigated their respective response to chronic treatment with fluoxetine.



Experiments were performed using homozygous 5-HTT–/– and wild-type 5-HTT+/+ littermates born from heterozygous mutants at the seventh generation of backcrossing with C57Bl/6J mice. Genotyping was performed as described by Bengel et al.(1998). After weaning and sexing, males and females were housed separately in groups of six to eight animals per cage and maintained under standard laboratory conditions (22 ± 1°C, 60% relative humidity, 12 h–12 h light– dark cycle, food and water ad libitum).

Procedures involving animals and their care were conducted in conformity with the institutional guidelines that are in compliance with national and international laws and policies (Council directive no. 87-848, 19 October 1987, Ministère de l’Agriculture et de la Forêt, Service Vétérinaire de la Santé et de la Protection Animale; permissions No. 75-116 to M.H. and No. 6269 to L.L.).

Alcohol self-administration procedure

Male mice were used at 3 months of age when body weight in each genotype equally ranged between 28 and 30 g. Mice were then housed singly in standard laboratory cages (Macrolon type 2 cages, 22 × 16 × 14 cm). Two tubes (20 ml) were placed on each cage, one containing tap water and the other filled with water supplemented with varying concentrations of ethanol. Mice had continuous free access to the drinking tips of both tubes. The position of the tubes was changed every 2 days in order to avoid possible bias due to place preference. Tubes were filled with freshly prepared liquids every 2 days. Food was provided ad libitum.

Homozygous mutants and wild-type mice were exposed to a progressively increasing concentration of ethanol (0 to 20% ethanol in water within 24 days) under the free-choice procedure described by Crabbe et al.(1996). During the first 4 days, both tubes were filled with tap water. Then, mice were offered 3% ethanol (v/v) versus water for 4 days, and 6% ethanol versus water for the following 4 days. For the next step, the ethanol concentration was raised to 10% for 6 days, then to 15% for another 6 days, and finally to 20% for 16–24 days (stabilization period).

Fluoxetine treatment

Fluoxetine [10 mg/kg/day intraperitoneally (i.p.)] or saline (5 ml/kg/day i.p.) was injected daily at 18:00 h for 10 days under the 20% ethanol/water conditions.

Ethanol and total fluid intake

Fluid intake and body weight were determined every 2 days between 17:00 and 18:00 h. The number of grams of ethanol consumed and volumes of total liquid intake/kg body weight/24 h were determined for each mouse.

Statistical analyses

Data were analysed by a two-way analysis of variance (ANOVA) and unpaired t-test, with Welch’s corrections when necessary. Statistical significance was set at P ≤ 0.05.


Spontaneous alcohol consumption in 5-HTT–/– versus 5-HTT+/+ mice

In both mutant and wild-type mice, daily ethanol consumption increased in parallel with the concentration of ethanol available to the animals (Fig. 1A), but 5-HTT–/– mutants drank less ethanol than wild-type mice at all concentrations, except 3 and 6% (post hoc analysis); this difference reached statistical significance only for the higher concentrations of ethanol (15 and 20%). During the stabilization period (the last 16 days) under the 20% alcohol regimen, the mean spontaneous daily alcohol consumption was also markedly lower in 5-HTT–/– mutants (∼4.0 g/kg/day) than in wild-type mice (∼9.0 g/kg/day). Indeed, for the whole alcohol self-administration procedure (40 days), total alcohol consumption was 55% lower (P < 0.001) in 5-HTT–/– mutants (126.8 ± 13.0 g/kg, mean ± SEM, n = 8) than in paired 5-HT+/+ mice (281.9 ± 15.0 g/kg, mean ± SEM, n = 8) (Fig. 1B).

Fig. 1.

Comparison of ethanol consumption by 5-HTT–/– mutant and 5-HTT+/+ wild-type mice. (A) Mean ± SEM of ethanol consumption (g/kg/day) is shown as a function of increasing ethanol concentration available to the animals during the alcohol self-administration procedure. Post hoc analysis revealed that 5-HTT–/– mutant mice drank significantly less ethanol than wild-type mice at the 15 and 20% ethanol concentrations. *P < 0.05; **P < 0.01. (B) Total alcohol consumption in 5-HTT+/+ and 5-HTT–/– mice during the 40-day alcohol self-administration procedure. Each bar represents the mean ± SEM of independent determinations in eight animals per group. ***P < 0.001.

Effects of fluoxetine on ethanol consumption in 5-HTT–/– versus 5-HTT+/+ mice

In another series of experiments, other groups of 5-HTT–/– and 5-HTT+/+ mice underwent a fluoxetine treatment (10 mg/kg/day) for 10 days during the alcohol self-administration procedure. This treatment produced a sustained decrease in alcohol consumption in 5-HTT+/+ mice (–38%), but was ineffective in 5-HTT–/– mutants (Fig. 2). Thus, under fluoxetine treatment, ethanol consumption did not significantly differ in 5-HTT+/+ (7.6 ± 1.1 g/kg/day, mean ± SEM, n = 8) and 5-HTT–/– mice (5.3 ± 2.6 g/kg/day, mean ± SEM, n = 8) by as soon as the third treatment day (Fig. 2). In contrast, saline administration for 10 days affected alcohol intake in neither 5-HTT+/+ nor 5-HTT–/– mice (data not shown).

Fig. 2.

Effect of fluoxetine on alcohol consumption in 5-HTT+/+ and 5-HTT–/– mice. Fluoxetine treatment (10 mg/kg/day for 10 days, horizontal bar) decreased ethanol consumption (mean ± SEM) in 5-HTT+/+ wild-type mice, but was ineffective in 5-HTT–/– mutant mice. Ethanol consumption under fluoxetine in 5-HTT+/+ mice did not significantly differ from that in 5-HTT–/– mice by as soon as the third treatment day. *P < 0.05 as compared with ethanol intake during the 6 days just prior to starting fluoxetine treatment; §P < 0.05, §§P < 0.01, as compared with wild-type mice.

Total fluid intake

The mean daily intake of total liquid (water plus ethanol) was in the same range in 5-HTT–/– (81.3–99.2 ml/kg) and 5-HTT+/+ (92.4–111.6 ml/kg) mice, and did not vary significantly at the various stages of the progressive alcohol self-administration procedure. In addition, chronic fluoxetine treatment did not significantly affect daily total fluid intake in both mutant (+2.3 ± 5.0% in fluoxetine-versus saline-treated mice, P > 0.05) and wild-type (–5.7 ± 4.8% in fluoxetine-versus saline-treated mice, P > 0.05) animals that had free access to 20% ethanol and water.

Body weight

As compared with values just before treatment (28.52 ± 0.11 and 28.63 ± 0.18 g in 5-HTT+/+ and 5-HTT–/– mice, respectively, means ± SEM, n = 16 in each group), a slight decrease in body weight was noted in both wild-type (–1.4 ± 0.2%) and knock-out (–1.3 ± 0.9%) mice at the end of the 10-day treatment with fluoxetine (10 mg/kg/day). In contrast, body weight increased slightly in both wild-type (+1.8 ± 0.8%) and knock-out (+1.6 ± 0.7%) mice that had received saline for 10 days.


The present data show that inactivation of the 5-HTT by either a pharmacological or a genetic approach markedly reduced alcohol consumption in C57BL/6J mice.

Several 5-HT re-uptake inhibitors, including the long-acting fluoxetine, have been found to decrease alcohol intake in dependent alcoholic patients (Naranjo et al., 1990, 1994; Lejoyeux, 1996), as well as in animal models of ethanol intake (Haraguchi et al., 1990; Higgins et al., 1992; Maurel et al., 1999). However, although in rat models the SSRI-induced reduction in ethanol intake has been established by several studies (see Maurel et al., 1999), the long-term efficiency and usefulness of SSRIs in the treatment of alcohol dependence in human subjects continues to be a matter of debate (Gorelick and Paredes, 1992; Naranjo and Knoke, 2001). It has notably been emphasized that although all SSRIs reduce alcohol consumption, the degree of this effect varies markedly from one SSRI to another, and is not directly related to the potency of these drugs to inhibit 5-HT re-uptake. This led to the assumption that SSRI-induced reduction in alcohol intake may not be caused solely by 5-HT re-uptake inhibition, but may also involve additional effects that are highly variable among these drugs. This could be especially true for fluoxetine, because this compound, which has been shown to exert higher anti-alcohol effects than other SSRIs (Maurel et al., 1999), is the least selective of these drugs with regard to 5-HT re-uptake inhibition, and is also a rather potent 5-HT2C receptor antagonist (Sanchez and Hyttel, 1999).

In order to explore further the effect of SSRIs on ethanol intake in C57BL/6J mice, which are well characterized as alcohol-appetant (preferring) mice (Yoshimoto and Komura, 1989), we used genetically paired mutants with targeted disruption of the gene encoding the 5-HTT, responsible for 5-HT re-uptake (Bengel et al., 1998). Indeed, homozygous 5-HTT–/– mice, which can be considered as a model of whole-life treatment with SSRI, have been shown to exhibit adaptive changes in 5-HT neurotransmission similar to those observed after chronic treatment with these drugs (Fabre et al., 2000; Mannoury la Cour et al., 2001). Using a free-choice procedure of progressive alcohol self-administration, we observed that the spontaneous alcohol intake was markedly lower in 5-HTT–/– mutants than in wild-type mice. In agreement with previous observations in rats (Haraguchi et al., 1990; Maurel et al., 1999), chronic treatment with fluoxetine was found to decrease significantly alcohol consumption in wild-type mice. However, this treatment did not affect alcohol consumption in 5-HTT–/– mutants, as expected from an action of fluoxetine through the blockade of 5-HT re-uptake. Accordingly, it can be assumed that the inhibitory effect of fluoxetine on alcohol intake is really caused by 5-HTT blockade and does not involve secondary effects of the drug at other molecular targets.

Interestingly, total fluid intake (water plus ethanol solution) was not significantly different in 5-HTT–/– mutants and paired wild-type mice, and was also unaffected by fluoxetine treatment, indicating that the reduction in alcohol consumption caused by 5-HTT inactivation was not the consequence of a general decrease in drinking behaviour. However, a slight decrease in body weight was noted in fluoxetine-treated mice, as expected from a limited reduction in food intake by this drug (Heisler et al., 1999). Interestingly, this effect was observed not only in wild-type mice, but also in 5-HTT–/– mutants, suggesting that it might involve actions of fluoxetine at a target(s) other than the 5-HTT. Accordingly, 5-HT re-uptake appears to be a relevant target for treatments aimed at selectively reducing alcohol intake in dependent alcoholic subjects.

In conclusion, the present findings showed that inactivation of 5-HT re-uptake leads to a sustained decrease in alcohol intake in 5-HTT–/– mutant mice, and that the inhibitory effect of fluoxetine on alcohol intake is the consequence of its direct interaction with the transporter responsible for 5-HT re-uptake in wild-type mice.


This research was supported by grants from INSERM, ATC 2002 INSERM, Bristol-Myers Squibb Foundation (unrestricted biomedical research grant) and IREB (contract No. 2000/14). We are grateful to E.Lilly Corp (Indianapolis, IN, USA), for the generous gift of fluoxetine. S.K. was recipient of a MILDT fellowship during the course of this work.


  • * Author to whom correspondence should be addressed.


View Abstract