INTRODUCTION

Ethanol consumption has monoaminergic as well as nonmonoaminergic components (Gianoulakis, 2001). Dopamine and serotonin systems have been especially implicated in ethanol consumption (for review see Nutt, 1999; Grace, 2000; Li, 2000). Many features of human and rodent ethanol consumption also display substantial sex dependence, although interactions between sex and monoamine effects have been studied only rarely.

A dopaminergic role in the effects of alcohol is supported by evidence from a variety of fields. Ethanol stimulates the firing of dopaminergic neurons (Mereu et al, 1984; Mereu and Gessa, 1984; Brodie et al, 1990,1999), and increases extracellular levels of dopamine in the nucleus accumbens (Wozniak et al, 1991; Yoshimoto et al, 1992a,1992b; Kiianmaa et al, 1995). A specific role for dopamine in ethanol reward is supported by data showing that dopaminergic agents and lesions of dopamine systems modify ethanol self-administration (Levy et al, 1991; Dyr et al, 1993; Russell et al, 1996; Hodge et al, 1997; Ikemoto et al, 1997) and by results documenting that ethanol is directly self-administered into the ventral tegmental area (Gatto et al, 1994). Animal models also support a role for dopaminergic differences in the genetic determination of ethanol reward. Rats bred for ethanol preference display dopamine receptor (Stefanini et al, 1992; McBride et al, 1993) and forebrain dopamine levels different from ethanol nonpreferring rats (Zhou et al, 1995), and dopamine receptor gene knockouts (KOs) also alter ethanol self-administration (Crabbe et al, 1996; Risinger et al, 1996,1999,2000; Rubinstein et al, 1997; El-Ghundi et al, 1998; Phillips et al, 1998).

Serotonin function has also been associated with ethanol consumption and alcoholism (for reviews see Lovinger, 1999; Hoffman et al, 2001; Myrick et al, 2001; Weiss et al, 2001). Reduced serotonin function has been especially postulated to predispose to alcoholism (Myers and Melchior, 1977). Reduced cerebrospinal fluid levels of the serotonin metabolite 5-HIAA are observed in some alcoholics (Linnoila et al, 1983), and are associated with increased ethanol consumption in rhesus monkeys (Higley et al, 1996). Ethanol-preferring rats display alterations in tissue serotonin levels (McBride et al, 1990,1991; Aulakh et al, 1994; Zhou et al, 1994), responses to serotoninergic agents (Gudelsky et al, 1985; Aulakh et al, 1988a,1988b,1992,1994; Wang et al, 1988), serotonin reuptake (Arora et al, 1983; Hulihan-Giblin et al, 1993; Chen and Lawrence, 2000), and serotonin receptor densities (Hulihan-Giblin et al, 1992,1993; Wong et al, 1993; McBride et al, 1994,1997; Chen and Lawrence, 2000). A variety of serotonergic agents can reduce voluntary ethanol consumption in animal models and in humans (for a review, see Lovinger, 1999; Myrick et al, 2001), and serotonin receptor gene KOs also affect ethanol self-administration (Crabbe et al, 1996; Risinger et al, 1996,1999,2000; Rubinstein et al, 1997; El-Ghundi et al, 1998; Phillips et al, 1998).

Dopamine and 5-HT involvement in the rewarding effects of ethanol need not be independent however. Ethanol elevates both serotonin and dopamine concentrations when administered into the VTA (Yan et al, 1996), and 5-HT3 receptor antagonists attenuate the dopamine release in the nucleus accumbens that is triggered by systemic ethanol (Wozniak et al, 1990). Vesicular monoamine transporter 2 (VMAT2) (Erickson et al, 1996; Takahashi and Uhl, 1997) mediates the vesicular storage of both dopamine and serotonin in synaptic vesicles (Gasnier, 2000; Uhl et al, 2000), and is therefore well positioned to regulate both dopaminergic and serotonergic neurotransmission, and perhaps alcohol consumption as well. DAT (Donovan et al, 1995) is similarly positioned to influence dopamine function since it mediates much of the clearance of released dopamine from the synaptic cleft. While effects of altering these genes' expression on psychostimulant reward and locomotion have been examined, there are no current data describing the effects of VMAT2 or DAT KOs on ethanol consumption.

There is substantial evidence for sex-dependent differences in alcoholism and in animal models of alcoholism (for a review see Lancaster, 1994,1995). Human alcoholism subtypes (Cloninger, 1987) may be differentially represented in males and females (Gilligan et al, 1987), and there is evidence for differential heritability of alcohol dependence in men and women in some (eg Han et al, 1999), but not all, studies (Heath et al, 1997). Furthermore, ethanol consumption is reduced by μ opiate receptor (MOR) gene KO in a sex-dependent fashion (Hall et al, 2001), so sex dependence might be expected in other alcohol-related genes as well. Despite this evidence, most human studies of dopaminergic and serotonergic gene polymorphisms have examined only males (Dobashi et al, 1997; Sander et al, 1997a; Iwata et al, 1998; Lappalainen et al, 1998,1999; Nielsen et al, 1998; Sander et al, 1999; Schuckit et al, 1999), studied subject pools that are largely male (Gelernter et al, 1991,1997; Higuchi et al, 1994; Muramatsu and Higuchi, 1995; Sander et al, 1997a,1997b; Ueno et al, 1999; Matsushita et al, 2001), omitted statements about the sex of the subjects tested (Pesonen et al, 1998; Sander et al, 1998; Gelernter and Kranzler, 1999; Blomqvist et al, 2000; Vandenbergh et al, 2000), or examined pooled data with sex assumed to be irrelevant (Goldman et al, 1997,1998; Parsian and Zhang, 1997; Franke et al, 1999; Laine et al, 2001a,2001b). When a recent study examined male and female alcoholics separately, it found sex-dependent interactions between X chromosome-linked monoamine oxidase A (MAOA) alleles and alcoholism subtypes (Schmidt et al, 2000).

Based on these considerations, we now report sex-specific assessments of ethanol consumption in VMAT2 (+/+ and +/−) and DAT (+/+, +/−, and −/−) KO strains (Takahashi et al, 1997; Sara et al, 1998). Data from heterozygous mice are especially interesting since such animals provide approximate models for the magnitude of human interindividual differences in the levels of expression of these two genes (see Uhl, 1998).

MATERIALS AND METHODS

Subjects: VMAT2 and DAT KO Mice

Mice were bred by random heterozygote crosses of mice developed in our laboratory maintained on a mixed C57/129sv background: VMAT +/+ and +/− (Takahashi et al, 1997); DAT +/+, +/−, and −/− (Sora et al, 1998). Most VMAT2 −/− mice die by the third postnatal day. The mice used in these experiments were from greater than the tenth generation of mice in these lines. All generations were produced from crosses of heterozygous mice. Mice were weaned at 21 days of age, and housed with same-sex littermates for the duration of the experiments. Standard colony conditions were used: 24°C, 50% relative humidity, and ad libitum food and water according to AALAC guidelines. Experimentation began at between 8 and 12 weeks of age, at which point mice were housed singly.

At weaning, 0.5 cm tail samples were taken for genotyping by PCR. Tail samples were incubated overnight at 55°C in tail buffer (50 mM Tris, pH 8.0; 100 mM EDTA; 100 mM NaCl; 1% SDS) containing Protease K (10 mg/ml). Supernatants were removed and lysis buffer added (0.32 M sucrose; 10 mM Tris, pH 7.5; 5 mM MgCl; 1% Triton X-100). After centrifugation the supernatant was removed and the tail DNA solution was used for PCR using PCR buffer (Lambda Biotech), 1 mM dNTP mix (Lambda Biotech), 25 mM MgCl2 (Lambda Biotech; final concentration of 4 mM), and 3.1 U/tube Tsg DNA Polymerase (Lambda Biotech, 5 U/μl).

Oligonucleotides (10 μM) for VMAT2 included a forward primer located outside the deleted region (5′GCT TAC CTC GTG GGC ATG GTG 3′), a reverse primer for the VMAT2 gene located in the region of the gene which is deleted in the KO (5′ GTC CCC AGT TTA TGT AGC ATT G 3′), and a reverse primer for the NEO gene (5′ TCG ACG TTG TCA CTG AAG CGG 3′). Amplimers from wild-type DNA were 1000 bp and amplimers from KOs were 700 bp.

For DAT genotyping, oligonucleotides (10 μM) included a forward primer located outside the deleted region (5′ GTG CCT AAG GTG CTC ACG GAG 3′), a reverse primer for the DAT gene located in the region which is deleted in the KO (5′ CAC AGC TCT GGC AGG TCT CAG 3′), and a reverse primer for the NEO gene (5′ GCC TCT GTC CGC AGT TCA TTC AG 3′). Amplimers from wild-type mice were 640 bp and from KOs were 900 bp.

Experiment 1. Voluntary Ethanol Consumption in VMAT2 KO Mice

Experimentally näve male and female littermates (N=15 per genotype; >10th generation) were housed singly beginning 1 week prior to the experiments. Water, food, and ethanol consumption were monitored in home cages. Initially, only food and water were available to determine baseline consumption. Subsequently, the subjects were given access to food, water, and ethanol in a standard two-bottle home-cage consumption paradigm. Fluids were made available in 50 ml polypropylene centrifuge tubes capped with rubber stoppers and standard sipper tubes (control experiments, data not shown, showed that spillage and evaporation from these tubes was less than 0.1 ml per measurement interval). Mice were weighed weekly, fluid bottles and food were weighed every 2–3 days, and consumption was calculated in g/kg/day, ml/kg/day, and g/kg body weight/day, respectively, for food, water, and ethanol. The initial ethanol concentration was 2% and concentrations were increased every 2–3 days in the following progression: 2, 4, 8, 12, 16, 24, and 32%. The positions of the bottles were switched each time the bottles were changed.

Experiment 2. Voluntary Ethanol Consumption in DAT KO Mice

Experimentally näve male and female littermates (N=9–11 per genotype; >10th generation) were studied as described previously. However, 11 of 20 DAT −/− mice died within 2 days of single housing, while none of the DAT +/+ or DAT +/− mice died. DAT −/− mice stopped drinking and eating (eg spontaneous adipsia and aphagia) under these conditions. Similar mortality, which we have previously observed in DAT −/− mice in our breeding facility (unpublished findings), may also be the result of this spontaneous adipsia/aphagia.

Data Analysis

Data were analyzed by ANOVA with the between-subjects measures of GENOTYPE and SEX. Subsequently, because ethanol consumption differed between male and female mice, and significant GENOTYPE × SEX interactions were observed in some cases, separate ANOVAs were performed on data from males and females for all data. Consumption data were analyzed as grams of ethanol per kilogram body weight per day (g/kg/day). The within-subjects factor of CONCENTRATION was used for these data. Although food and water consumption was measured throughout, only the baseline data are presented, expressed in g/kg/day and ml/kg/day, respectively. Post hoc analyses were performed using Scheffe's post hoc comparisons.

RESULTS

Experiment 1. Voluntary Ethanol Consumption in VMAT2 +/+ and +/− Mice

Baseline food and water consumption

Male mice weighed more than female mice (Table 1: F[1,56]=37.5, p<0.0001), but there was no significant effect of GENOTYPE (F[1,56]=0.7, NS). When food consumption was expressed as g/kg/day it was found that male VMAT2 +/− mice ate more than female VMAT2 +/− mice, but no such difference was observed in VMAT2 +/+ mice (Table 1: GENOTYPE × SEX F[1,56]=6.0, p<0.02). Male VMAT2 +/− consumed more food than male VMAT2 +/+ mice, and female VMAT2 +/− mice consumed less food than femaleVMAT2 +/+ mice, but neither of these effects was statistically significant according to post hoc analysis. An identical GENOTYPE × SEX effect was found for water consumption (Table 1: GENOTYPE × SEX F[1,56]=6.8, p<0.02).

Table 1 Baseline measures in VMAT2 KO mice
Full size table

Ethanol consumption and preference

Female mice consumed more ethanol than male mice (Figure 1 and Figure 2). An overall ANOVA for ethanol consumption, including SEX as a factor, revealed significant effects of SEX (F[1,56]=11.2, p<0.002) and CONCENTRATION (F[7,392]=101.5, p<0.0001). Ethanol preference was also affected by ethanol CONCENTRATION (Figure 1b and Figure 2b; F[1,56]=11.8, p<0.0001), but not by any other factors.

Figure 1
figure 1

Male VMAT2 KO mice: The data represent voluntary consumption of ethanol in grams per kilogram of body weight per day (g/kg/day) (a) and percent preference (b), for male +/+, and +/− VMAT2 KO mice. The data are expressed as mean±SEM. *Significant post hoc difference using Scheffe's post hoc comparisons, +/+ vs +/−, p<0.05.

Full size image
Figure 2
figure 2

Female VMAT2 KO mice: The data represent voluntary consumption of ethanol in grams per kilogram of body weight per day (g/kg/day) (a) and percent preference (b), for female +/+, and +/− VMAT2 KO mice. The data are expressed as mean±SEM.

Full size image

When the data were analyzed separately by SEX, ANOVA of the data from male mice revealed more consumption of higher-concentration ethanol solutions in VMAT2 +/− than in VMAT2 +/+ mice (Figure 1a; GENOTYPE F[1,28]=5.6, p<0.03; GENOTYPE × CONCENTRATION F[7,196]=2.5, p<0.02). There was no difference in ethanol consumption between female VMAT2 +/+ and +/− mice (Figure 2a; GENOTYPE F[1,28]=0.0, NS; GENOTYPE × CONCENTRATION F[7,196]=0.1, NS). There were no differences between genotypes in ethanol preference for males (Figure 2a; GENOTYPE F[1,28]=0.5, NS) or females (Figure 2b; GENOTYPE F[1,28]=0.5, NS).

Experiment 2. Voluntary Ethanol Consumption in DAT KO Mice

Baseline food and water consumption

Male mice weighed more than female mice (Table 2: F[1,45]=55.4, p<0.0001), independent of genotype. DAT −/− mice weighed less than either +/+ or +/− mice (Table 2: GENOTYPE F[2,45]=10.9, p<0.001), as previously reported (Sora et al, 1998). Overall, males consumed more food than females (Table 2: SEX F[1,45]=5.2, p<0.03), but this difference was not observed in DAT −/− mice (significant post hoc Scheffe's comparison between males and females in DAT +/+ and +/− mice, but not DAT −/− mice). Despite these differences in food consumption, there was no effect of either GENOTYPE or SEX on water consumption (Table 2: GENOTYPE F[2,45]=0.7, NS; SEX F[1,45]=0.6, NS).

Table 2 Baseline measures in DAT KO mice
Full size table

Ethanol consumption and preference

Overall, ANOVA for ethanol consumption, including SEX as a factor, revealed significant effects of CONCENTRATION (Figure 3 and Figure 4; F[7,315]=27.9, p<0.0001), SEX (F[1,45]=35.9, p<0.0001), and a GENOTYPE × SEX interaction (F[2,45]=3.2, p<0.05). Overall, ethanol preference was affected by CONCENTRATION (Figure 3b and Figure 4b; F[7,315]=10.5, p<0.0001) and by GENOTYPE (F[2,45]=4.0, p<0.03). As before, female mice consumed more ethanol than male mice.

Figure 3
figure 3

Male DAT KO mice: The data represent voluntary consumption of ethanol in grams per kilogram of body weight per day (g/kg/day) (a) and percent preference (b), for male +/+, +/−, and −/− DAT KO mice. The data are expressed as mean±SEM. *Significant post hoc difference using Scheffe's post hoc comparisons, +/+ vs +/−, p<0.05.

Full size image
Figure 4
figure 4

Female DAT KO mice: The data represent voluntary consumption of ethanol in grams per kilogram of body weight per day (g/kg/day) (a) and percent preference (b), for female +/+, +/−, and −/− DAT KO mice. The data are expressed as mean±SEM.

Full size image

When the data from each SEX were analyzed separately, it was found that male DAT +/− and −/− mice con-sumed more ethanol than DAT +/+ mice, particularly at high ethanol concentrations (Figure 3a; GENOTYPE F[2,23]=3.6, p<0.05). The genotype difference was only significant, by post hoc Scheffe's comparisons, at the high ethanol concentrations. Ethanol preference did not differ significantly in male mice based on GENOTYPE (Figure 3b; F[2,23]=0.6, NS) or GENOTYPE × CONCENTRATION interaction (F[7,161]=1.1, NS).

There were no differences in ethanol consumption between female DAT +/+ mice and female DAT +/− or −/− mice (Figure 4a; GENOTYPE F[2,22]=2.5, NS; GENOTYPE × CONCENTRATION F[7,154]=1.0, NS), although there was a trend for female DAT +/− to consume more, and female DAT −/− less, ethanol than female DAT +/+ mice at lower ethanol concentrations. Preference assessments did reveal (Figure 4b) significant effects of both CONCENTRATION (F[7,154]=3.8, p<0.001) and GENOTYPE (F[2,22]=4.1, p<0.04). Female DAT +/− mice displayed ethanol preferences greater than those of both +/+ and −/− mice, the latter also having trends toward reduced ethanol preference.

DISCUSSION

Deletions of either the DAT or the VMAT2 genes increased ethanol consumption in male but not female mice, although female DAT +/− mice displayed higher ethanol preferences and there was a trend for female DAT −/− mice to consume less ethanol and to have lower preferences. These observations add to the wealth of evidence that these neurotransmitter systems modulate ethanol consumption, and also provide working hypotheses about the ways in which human interindividual differences in the expression of these genes could affect human ethanol consumption.

A recent study in DAT KO mice (Savelieva et al, 2002) found decreased consumption in female DAT KO mice (where a trend toward reduced consumption was observed in the present study), while the increased consumption in male DAT KO mice observed in the present study was paralleled (in absolute terms, although the differences were not statistically significant) in the Savelieva et al (2002) study. Despite the apparent discrepancies between the two reports, the overall pattern of effects is strikingly similar even though the magnitude of differences in males and females is reversed in the two studies. The reasons for this divergence can only be speculated at, but might denote differences in genetic background as the two strains were developed independently but certainly are consistent with the view that there are multiple genetic and nongenetic determinants of ethanol consumption.

Certainly some of the basis for these effects of gene deletion might be rationalized by examining the effects of ethanol on dopamine neurons. Ethanol stimulates firing of dopaminergic neurons (Mereu et al, 1984; Mereu and Gessa, 1984; Brodie et al, 1990,1999) and increases extracellular levels of DA in the nucleus accumbens (Wozniak et al, 1991; Yoshimoto et al, 1992a,1992b; Kiianmaa et al, 1995). Ethanol supports self-administration behavior when injected directly into the ventral tegmental area, consistent with a large role of dopaminergic systems in ethanol reward (Gatto et al, 1994). However, in considering the potential consequences of DAT and VMAT2 gene KO on ethanol actions, the effects of these gene KOs on monoamine neurotransmitter dynamics should also be considered.

DAT KO mice display not only DAT loss but also compensatory reductions in dopamine synthesis (Jaber et al, 1999), autoreceptor function (Jones et al, 1999), and dopamine receptor levels (Sora et al, 2001b), making simple predictions difficult. This circumstance is even more complicated in heterozygous KOs in which compensatory potential is greater than in full KOs, depending on the degree of compensatory change and receptor reserve (Sora et al, 2001a). Such complications could account for examples of enhanced pharmacological effects in heterozygous KO mice (eg Figure 4b; Sora et al, 2001a). In any case, the simplest hypothesis is that reduced DA reuptake could potentiate the effects of ethanol-induced firing of VTA neurons by prolonging the time during which released dopamine is available to interact with extracellular receptors and by extending the extracellular distance that released dopamine can travel before it is inactivated by uptake or metabolism.

Since VMAT2 KO reduces the accumulation of all monoamines into vesicles, multiple effects of VMAT2 KO could affect ethanol consumption. VMAT2 KOs alter monoamine function by reducing tissue content (Fon et al, 1997; Wang et al, 1997; Mooslehner et al, 2001), extracellular monoamine levels (Wang et al, 1997), and the amounts of monoamine that amphetamine or depolarization can release (Wang et al, 1997). VMAT2 +/− mice display enhanced locomotor effects of psychostimulants (Takahashi and Uhl, 1997) and ethanol (Wang et al, 1997), but reduced reward from amphetamine (Takahashi et al, 1997). While DAT and VMAT2 reductions can produce dissimilar effects, both may be seen as reducing the influence of ‘phasic’ levels of dopamine released by nerve impulse trains in relationship to the relatively increased influence of ‘tonic’ dopamine whose levels and distributions are less dependent on dopamine cell firing. While these considerations provide a plausible shared dopamine mechanism for the effects of DAT and VMAT2 KO, neither the effects of VMAT2 KO or DAT KO on ethanol consumption are necessarily or entirely mediated by direct alterations in dopaminergic function.

The effects of DAT and VMAT2 KO on ethanol consumption were sex dependent. Female mice have often been observed to consume more ethanol than male mice (Middaugh and Kelley, 1999; Middaugh et al, 1999), and the effects of MOR gene KO are also sex dependent (Hall et al, 2001). As modulation of MOR levels by progesterone and estrogen (Carter and Soliman, 1996,1998) provides one potential explanation for the sex interaction with MOR KO, so ovarian hormonal regulation of DAT and VMAT2 (Attali et al, 1997; Disshon et al, 1998) might also account for sex-dependent effects in DAT and VMAT2 KO mice. In addition, different consumption levels in male and female mice might affect the sensitivity with which these experiments could reveal enhanced or reduced ethanol consumption. Higher consumption may reduce the likelihood of detecting elevated ethanol consumption (ie ceiling effects), as noted for the female mice in this study. This analysis also suggests that other factors that increase or decrease ethanol consumption, such as the background strain of the mice (eg C57Bl/6 vs DBA/2), might similarly interact with the effects of gene KO.

In addition to the dependency on sex, the effects of genotype were also highly dependent on ethanol concentration. The effects of genotype were only observed at high ethanol concentrations. Such concentration dependencies have been noted previously in the effects of isolation rearing (Wolffgramm, 1990; Hall et al, 1998) and strain (Hall et al, 1998) on ethanol consumption in rats. The reasons for concentration dependencies are unknown; however it might be speculated that if ethanol produces rewarding effects through multiple mechanisms, then this might be differentially activated by different doses of ethanol. Alternatively, this might be due to differences in taste aversion, although such suggestions have not proved true for the strain differences mentioned earlier (Hall et al, 1998).

Observations in heterozygous mice, in which the expression of VMAT2 and DAT lie within the range of human variation in the expression of these genes, may have direct relevance to alcoholism, although substantial environmental impact on human individual differences in the expression of these genes is still possible. A human DAT exon 15 variable number tandem repeat marker (VNTR) has been frequently examined in association studies that have compared alcoholics or polysubstance abusers with controls. This VNTR DAT polymorphism has been associated with modestly altered DAT availability as assessed by single photon emission computed tomography (SPECT) in alcoholics and controls (Heinz et al, 2000). Although several significant associations (Muramatsu and Higuchi, 1995; Dobashi et al, 1997; Sander et al, 1997b; Ueno et al, 1999) have been reported, most studies lack such association (Parsian and Zhang, 1997; Franke et al, 1999; Heinz et al, 2000), although no study carefully examined possible sex effects. Furthermore, the VNTR marker displays little linkage disequilibrium with 5′ exons or 5′ flanking sequences that are more classical candidates to control levels of DAT expression (Vandenbergh et al, 2000). Initial human VMAT2 variants have also been studied for association with polysubstance abuse vulnerability, but these markers again are unlikely to adequately reflect all of the important variation identified at this locus (Uhl et al, 2000). The current observations of sex-dependent differences in ethanol consumption or preference in DAT and VMAT KO mice add to the weight of evidence for monoaminergic influences on alcohol consumption and motivate more careful evaluation of sex-dependent differences at these loci in humans of genetic vulnerability to alcoholism.