Our principal finding was that allelic variation in the 5′-HTTLPR was associated with differences in phenomenological characteristics of treatment-seeking alcoholics and in the biological parameters of paroxetine binding and 5-HT uptake by the 5-HTT of the platelets in these alcoholics. At the time of enrollment, SS homozygotes were younger in chronological age and had tended to drink more heavily for fewer years of lifetime drinking than their L-carrier (i.e., LL and LS) counterparts. Even more importantly, we found evidence that patterns of drinking among these alcoholics may differentially affect 5-HTT expression and function depending upon the subjects’ allelic genotype. First, we found that the 5-HTT polymorphism was functional in the platelets of alcoholics in that compared with the SS homozygotes, subjects with an L-allele (i.e., the L-carriers) actually showed reduced paroxetine binding capacity and reduced functional 5-HT uptake capacity. Second, we frequently found relationships between drinking and platelet 5-HT parameters only within the L-carriers and never within the SS homozygotes. Third, we found that for the majority of the population who were L-carriers, heavier current drinking was associated with increased 5-HT uptake but reduced paroxetine binding, but that a greater number of years of problem drinking was associated with reductions in both of these parameters.
The data showing reduced 5-HT binding and uptake among the L-carriers contradict common expectations that L-carriers should have higher expression and/or function than should SS homozygotes (Johnson, 2000
). Though we have studied the platelets and not the brains of our subject population, our findings are entirely consistent with similar observations from a neuroimaging study showing that, compared with controls, alcoholics with the LL genotype have reduced β-CIT neuronal binding to 5-HTT in raphe brain areas (Heinz et al., 2000
). Although that study showed similar levels of binding in LL vs. SS homozygotes, our L-carriers actually had less 5-HT binding and uptake than the SS subjects. Heinz et al. (2003)
further reported that in rhesus monkeys exposed to alcohol, this differential reduction in raphe binding potential for L-carriers is related to chronic alcohol drinking, suggesting an alcohol “toxicity” to the 5-HTT. The current study is the first to show that the reduced activity of the 5-HTT in human alcoholics is related to drinking patterns and that this can be demonstrated in platelet preparations. We previously reported reduced 5-HT uptake in the platelets of a smaller sample of LL alcoholics (Javors et al., 2005
). The current study substantially expands those findings, clearly showing reduced functional 5-HTT uptake in L-carriers, and adds the findings of reduced paroxetine binding as well. Heinz et al. (2004)
hypothesized that neurotoxic reductions in 5-HTT protein expression were limited to homozygous L-carriers and correlated with negative mood states as opposed to the SS genotype. Also, Sander et al. (1997)
observed a behavioral dichotomy between L-carriers and SS genotype on the basis of severity of withdrawal symptoms of alcoholics. Overall, these data clearly support the hypothesis that alcohol’s toxic effect on 5-HTT gene expression affects selectively those who are L-carriers but not those of the SS genotype.
The finding of relationships between drinking and 5-HTT function among L-carriers but not SS homozygotes needs to be characterized more fully. Somewhat paradoxically, L-carriers who are currently drinking more heavily exhibit higher 5-HTT uptake and lower affinity while exhibiting less paroxetine binding. We think that a plausible explanation for this finding is that current drinking increases 5-HT release markedly (LeMarquand et al., 1994a
), and there may be an attempt to compensate physiologically by increased uptake. Nevertheless, binding levels remain low because of the toxic damage to those of L-carrier status.
In the current study, SS subjects tended to drink more and to be of a younger current age than LL subjects, which is consistent with the finding of Herman et al. (2005)
that binge drinking may be greater in young college-aged adults with the SS genotype. However, the current study found no significant relationships between 5-HTT genotype and age of onset of alcoholism, which is consistent with family-based studies reporting no differential association of the L or S allele with early onset of alcoholism in offspring (Samochowiec et al., 2006
). Previous reports have suggested that the SS genotype is associated with increased anxiety and affective liability (Olsson et al., 2005
) and may be associated with subtypes of alcoholism complicated by a comorbid psychiatric condition (Feinn et al., 2005
). Though the current study included only patients without psychiatric comorbidity, we have reported, in another study of this population, that anxiety reduction is associated with a favorable treatment outcome (Sloan et al., 2003
). Since S-linkages to alcohol dependence severity have not been consistently supported (Köhnke et al., 2006
; Kweon et al., 2005
), these findings suggest the possibility that the SS genotype may be associated with greater tendencies to seek treatment at a younger age.
Paroxetine binding is a widely used marker for the expression rate of the 5-HTT. Previous studies have shown that the principal binding site for paroxetine in platelets is the 5-HTT but that paroxetine also binds to mitochondrial and alpha granule proteins in addition to the 5-HTT (Cesura et al., 1990
). The observed lack of correlation between Vmax
in our study probably is due in part to this lack of specificity of paroxetine binding. However, our unpublished results have indicated that only about 20% of the 5-HTTs are located at the surface of platelets, which may further account for the observed lack of correlation between Vmax
(Javors et al., 2005
). Although the density of paroxetine binding (Bmax
) is certainly an indicator of the expression rate of the 5-HTT, the mechanism by which the majority of transporters are internalized is unknown, but it is likely to include phosphorylation, glycosylation, and other cellular signaling mechanisms involved with 5-HTT uptake and turnover. It is unknown whether these internalization processes may differ among the L-carriers as opposed to the SS homozygotes, and additional research is required to understand the effects of alcohol on transporter internalization processes. Research comprising such detailed analyses of alcohol’s cellular effects on the 5-HTT is ongoing in our laboratory.
If our observation of the effects of drinking status and allelic variation at the 5-HTT in platelets parallels changes that occur in the brain, it would provide added support for our hypothesis that ondansetron treatment response might be greater among L-carriers than in those with the SS genotype (Johnson, 2000
). This is because in the brain, 5-HTTs in raphe nuclei are somatodendritic. Decreased somatodendritic function in alcoholics of L-carriers can, therefore, be expected to reduce 5-HT firing rates by increasing self-inhibition (Little et al., 1998
), resulting in reduced intrasynaptic neurotransmission and post-synaptic up-regulation of 5-HT receptors. Blockade of up-regulated post-synaptic 5-HT3 receptors might underlie ondansetron’s therapeutic efficacy. Indeed, formal study of the effect of ondansetron on drinking among alcoholics with allelic variation at the 5-HTT is ongoing.