Abstract

Introduction

Individuals vary in their therapeutic and acute behavioral responses to stimulant drugs. This variability has been observed in clinical populations, such as patients treated for Binge Eating Disorder (Davis et al. 2007), Attention-Deficit Hyperactivity Disorder (Spencer etal. 1996) and narcolepsy (Mitler et al. 1993). Variability has also been reported in healthy volunteers, who receive acute doses of stimulants in laboratory studies (Brauer and de Wit 1996; Lott et al. 2005; de Wit et al. 1986; Gabbay 2003). In particular, healthy volunteers differ in subjective ratings of drug-induced mood effects, including mood effects that are linked to the drugs’ potential for abuse. One possible source of individual variability is in the dopamine transporter (SLC6A3), which is a direct target of amphetamine. Through its actions on the dopamine transporter, amphetamine blocks the reuptake of dopamine and causes reverse transport of intracellular dopamine into the synapse (Jones et al. 1998; Di Chiara and Imperato 1988, Schiffer et al. 2006). In humans, positron emission tomography studies demonstrate that cocaine-induced euphoria is correlated with the extent to which cocaine binds the dopamine transporter (Volkow et al. 2004). In animals, knockout mice lacking SLC6A3 show a reduced locomotor response to amphetamine (Spielewoy et al. 2001).

SLC6A3 is located on chromosome 5p15.33 and consists of 15 exons. The most commonly studied polymorphism in this gene is the variable number of tandem repeat polymorphism in the 3′ untranslated region of this gene (3′-UTR VNTR). This polymorphism has been associated with both Attention Deficit Hyperactivity Disorder (ADHD) and response to stimulants (Cook et al. 1995; Waldman et al. 1998; Daly et al. 1999; Bakker et al. 2005; Langley et al. 2005; Feng et al. 2005; Purper-Ouakil et al. 2005; Li et al. 2006; Lott et al. 2005). Recent studies indicate that single marker polymorphisms (SNPs) or haplotypes of SLC6A3 in the 5′ region may be also associated with ADHD and related phenotypes (Friedel et al. 2007; Brookes et al. 2006; Lasky-Su et al. 2006). These newly identified SNPs are not in linkage disequilibrium with the 3′-UTR VNTR, suggesting that there are other important polymorphisms within this gene.

In this study we evaluated four SNPs in SLC6A3. The first SNP, rs3756450, has been implicated in risk for schizophrenia (Talkowski et al. 2008). The second SNP, rs460000, is in perfect linkage disequilibrium with a number of neighboring SNPs that have been associated with ADHD and related haplotypes (Friedel et al. 2007; Lasky-Su et al. 2006). The remaining two SNPs have not been previously associated with any phenotypes. These SNPs may provide some insight into the role of variability in this gene, specifically on genetic variability related to response to amphetamine. We examined the relationship between these SNPs and measures of the subjective, cognitive and physiological response to amphetamine.

Methods

Results

Administration of d-amphetamine produced the expected responses on all 5 scales of the ARCI. It increased scores on the stimulation scale F_(2,302) = 58.7, P ≤ 0.001, ${\eta }_{\mathrm{p}}^2$ = 0.15; the intellectual efficacy and energy scale F_(2,302) = 39.1, P ≤ 0.001, ${\eta }_{\mathrm{p}}^2$ = 0.21; the dysphoria scale F_(2,302) = 10.4, P ≤ 0.001, ${\eta }_{\mathrm{p}}^2$ = 0.06; the euphoria scale F_(2,302) = 50.4, P ≤ 0.001, ${\eta }_{\mathrm{p}}^2$ = 0.25 and it decreased the sedation scale F_(2,302) = 28.8; P ≤ 0.001 ${\eta }_{\mathrm{p}}^2$ = 0.16.

Administration of d-amphetamine also produced the typical effects on the physiological measures blood pressure and heart rate and on the DSST. d-amphetamine increased systolic blood pressure F_(2,302) = 110.92, P ≤ 0.001, ${\eta }_{\mathrm{p}}^2$ = 0.423; diastolic blood pressure F_(2,300) = 59.46, P ≤ 0.001, ${\eta }_{\mathrm{p}}^2$ = 0.029; heart rate F_(2,302) = 26.667, P ≤ 0.001, ${\eta }_{\mathrm{p}}^2$ = 0.150; and increased the number of substituted symbols F_(2,284) = 12.13, P ≤ 0.001, ${\eta }_{\mathrm{p}}^2$ = 0.079.

Table 1 shows the genotype frequencies for the 4 SNPs included in the study. The frequencies were similar to the frequencies reported in the HapMap project. Because all SNPs had a minor allele frequency of <0.1, we pooled the homozygotes of each minor allele with the heterozygotes; this approach precluded the detection of recessive effects associated with the rare allele; however, due to the sample size, we did not have sufficient power to detect such effects.

Table 2 provides a summary of demographic characteristics and drug use histories. Most of the subjects were in their twenties with at least some college education. They were moderate caffeine and alcohol drinkers and low cigarette and marijuana smokers. The genotypic groups formed by the four SNPs in SLC6A3 did not differ on most demographic characteristics. However, for rs6869645, which consisted of 18 C/T and 134 C/C individuals, there were significantly more male subjects in the C/T group (males = 14, females = 4) than in the C/C group (males = 71, females = 63). As a result gender was included as a covariate for all analyses of rs6869645.

Table 2

Demographic characteristics and drug use histories of the participants (N = 152)

Age (years; mean ± SD)	22.8 ± 3.4
Gender: Male/female	85/67
BMI (Mean ± SD)	22.6 ± 2.2
Education (n)
High school	2
Some college	64
College degree	68
Advanced degree	18
Current substance use (Mean ± SD)
Alcohol (drinks per week)	4.6 ± 03.7
Cigarettes (per week)	0.7 ± 01.8
Marijuana (uses per month)	0.9 ± 2.3
Caffeine (cups per week)	7.3 ± 7.0
Lifetime substance use (% ever used)
Marijuana	60.0
Stimulants	31.6
Opiates	27.0
Tranquilizers	7.5

Open in a separate window

In the absence of drug administration there were no robust differences on the ARCI scores between SLC6A3 genotypic groups; however, subjects in the T/T genotype group for SNP rs3756450 scored slightly lower on the baseline stimulation scale (2.02 ± 1.03) compare to subjects in the C/T+C/C genotype group (2.53 ± 1.0; P = 0.013).

For the SNP rs460000 there were significant differences in the response to amphetamine. We identified a significant genotype x dose interaction for both the stimulation scale F_(2,300) = 4.42, P = 0.015 ${\eta }_{\mathrm{p}}^2$ = 0.028 and for the euphoria scale F_(2,300) = 3.72, P = 0.025; ${\eta }_{\mathrm{p}}^2$ = 0.024. Post-hoc tests indicated that these differences were significant for both the 10 and 20 mg doses but that the two genotype groups were similar following the placebo dose. These differences are shown in Fig. 1. We did not find a difference in the means between the three genotype classes for rs460000 on stimulation scale F_(4,298) = 2.12 (P = 0.080), or euphoria scale F_(4,298) = 1.89 (P = 0.113). Visual inspection of the means did not reveal a linear trend—the A/A group did not have a lower response than A/C. There were no significant genotype-drug interactions for the ARCI scales for any of the other three SNPs. Further, none of the genotypes differed significantly in their response to amphetamine on the DSST, or on physiological (heart rate, blood pressure) measures.

Open in a separate window

Fig. 1

Mean area under the curve ± SEM for stimulation scale (top panel) and euphoria scale (bottom panel) for both rs460000 genotypic groups. The C/C group (N = 83) reported greater stimulant effects at both 10 mg (***) and 20 mg (**) doses, compared to the combined A/A+A/C group (N = 69). In addition, the C/C group (N = 83) reported greater euphoric effects at both 10 mg (*) and 20 mg (*) doses, compared to the combined A/A+A/C group (N = 69). * P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001

Discussion

The main finding of this study is that subjects homozygous for the C allele of rs460000 exhibited a pattern of enhanced responsiveness to the stimulant and euphoric effects of acute amphetamine. The two genotypic groups (A/A+A/C and C/C) of rs460000 did not differ on the placebo session, but the CC group reported stimulation and euphoria about twofold higher than the combined A/A+A/C group after both doses of d-amphetamine (10 and 20 mg). This SNP did not influence physiological and cognitive measures in either the placebo or drug conditions. None of the other SNPs had significant effects on any of the measures studied.

Findings like this with an individual SNP must be interpreted with caution because of problems with multiple testing. However, one reason to believe that the results of this analysis reflect a true difference in stimulant responses between the groups is their consistency with results of recent studies describing association of either SNPs or haplotypes of the SLC6A3 with ADHD and related phenotypes (Friedel et al. 2007; Brookes et al. 2006; Lasky-Su et al. 2006). ADHD, like response to stimulants, is related to dopamine function, and stimulant drugs are an effective treatment for ADHD. Friedel et al. (2007) found a significant association between ADHD and rs463379. Our SNP (rs460000) is in perfect linkage disequilibrium with rs463379 (both D′ and r² = 1 for CEU). In addition, Lasky-Su et al. (2006) evaluated 35 SNPs in SLC6A3 in relation to substance use disorder in ADHD patients and found that six SNPs, including rs460000, were associated with that phenotype, although correction for multiple comparisons made the significance of rs460000 marginal. In addition, Brookes et al. (2006) found that the only SLC6A3 haplotype showing a trend towards significant association with ADHD included rs460000 (best nominal P = 0.075; see supplementary information). These results suggest that this locus pleiotropically influences subjective responses to stimulant drugs, ADHD and substance use in ADHD patients. These other observations lend support to the idea that our observations reflect a true biological effect.

Our results make sense in the context of the current understanding of the biological basis for the reinforcing effects of stimulants. It is widely believed that the reinforcing and addictive properties of stimulants depend on their ability to interact with the dopamine transporter, thereby increasing the extracellular concentration of the neurotransmitter dopamine within specific brain areas (Kuhar et al. 1991; Wise and Bozarth 1987). The increased extracellular dopamine is thought to mediate the drugs’ subjective (Broadbent et al. 1991) and reinforcing effects (Ritz et al. 1987; Bergman et al. 1989; Volkow et al. 1997). Although the functional consequences of rs460000 or SNPs in linkage disequilibrium with rs460000 are as yet unknown, it seems logical to postulate that the functional variant affects drug responses by altering expression of the dopamine transporter, which would be expected to alter synaptic dopamine levels.

This study had both strengths and limitations. We tested a reasonable number of subjects (N = 152) for this type of drug challenge study. Moreover, the two genotypic groups of rs460000 were of about equal size, allowing us to make the comparison between C/C and A/A+A/C groups with a maximum of power. On the other hand, our small sample size and the preliminary nature of the study did not provide enough power to support a stringent correction of multiple comparisons. If we had used the Bonferroni-corrected P value (.0014) to account for the multiple comparisons, the drug-genotype interaction on the measure of stimulation (both 10 and 20 mg condition) would remain significant, but the effects on euphoria would not. A further limitation of the study is that the selected SNPs were not tagSNPs and thus do not capture the bulk of the variability in this gene. As such, there may be other important alleles of this gene that we did not detect.

In summary, we showed that rs460000 predicts stimulant and euphoric effects of acute doses of d-amphetamine in healthy volunteers. Notably, the differences across the genotypic groups were apparent at both doses of d-amphetamine, suggesting that the effect is replicable. It is interesting that this locus appears to influence both ADHD and the response to amphetamine, a drug that is used to treat ADHD, thus providing a particularly interesting example of pleiotropy. The results of this research may aid in our understanding of individual variability in the response to drug of abuse and could help to predict individual susceptibility to stimulant drug abuse.

Acknowledgments

References