Although epidemiologic studies have shown that substance addiction is strongly influenced by genetic factors, the number and identity of vulnerability genes remain unknown [1, 3–8]. This is the first study to examine eight candidate genes for association with substance addiction in individuals of Arab descent. These eight genes were the Dopamine receptors (DRD1 ,DRD2 ,DRD2 ,DRD3 and DRD5), Solute Carrier Family 6, Member 3 (SLC6A3), Brain-Derived Neurotrophic Factor (BDNF) and Catechol-O-Methyltransferase (COMT). Altogether 460 individuals were genotyped using 49 SNPs from these eight genes. Of the samples tested, 220 were from substance addicted male subjects of Arab descent. The control group were an ethnically homogenous Jordanian Arab population with no lifetime history of psychosis, mood disorders or substance dependence.
Both dopamine and non-dopamine neurochemical pathways through neurotransmitters (SLC6A3), neurotrophic factors (BDNF) and enzymes (COMT) are influenced by drugs and their psychoactive and addictive effects [8, 10, 12]. Dopamine is one of the main neurotransmitters involved in the stimulation of reward pathways, which is the important feature of substance addiction [13, 15, 57]. It has been suggested that dopamine receptor genes play a role in the genetics of substance addiction [58–60]. Previous studies have emphasized the importance of dopamine gene family specifically DRD2 gene as a general risk factor for substance dependence rather than a marker of risk for a particular drug [13, 15, 18]. However, various genetic association studies reported that there are inconsistencies in the frequency of alleles within DRD2 gene in different populations. For example, Barr and Kidd reported that the A1 allele frequency differs dramatically among the population studied from as low as 0.09 to as high as 0.075 .
As many studies indicated that multiple substances influence dopaminergic system activity, the investigation of substance addiction may result in a complete examination of gene risk [7, 11, 12]. In this study, none of the polymorphisms within the eight genes differed significantly for allele or genotype frequencies, with exception of six polymorphisms (rs2283265, rs10765560, rs2075654, rs1125394, rs2734836 and rs1799732) within the DRD2 gene. The strongest statistical evidence for these association signals was found within the DRD2 gene at two sites: rs1799732 (C/-C, 5'-UTR) and rs1125394 (A/G, intron 1). The strongest evidence of allelic frequency for these association signals were from rs1799732.
The rs1799732 (C/-C, 5'-UTR) is of particular interest because there is evidence that this allele has a functional effect on DRD2 gene expression . The dopaminergic system is involved in reward and reinforcing mechanisms in the brain [13, 57] specifically the positive reinforcing effects of substance addiction . Animal and human studies of addiction indicate that DRD2 plays a critical role in the mechanism of reward and reinforcement behavior [60–63]. Various animal studies reported that opiate rewarding effects were absent in mice lacking D2 receptors, while DRD2 overexpression in transgenic mice led to reduced self-administration of alcohol [60, 62]. A positron emission tomography study of human brain showed that D2 receptor density in the brain decreased significantly in alcoholic compared with control subjects [63, 64]. These findings suggest that genetically determined variation in DRD2 expression and function can alter reward responses to a variety of substances and may contribute to vulnerability to heroin dependence in humans. For example, DRD2 gene was previously studied by Xu et al. (2004) to examine the susceptibility of this gene with heroin dependence in Chinese and German population . This study found that genetic polymorphisms, specifically rs1799732 (C/-C), within DRD2 gene play a role as a susceptibility gene with heroin dependence in Chinese but not in German population .
Association with substance addiction was not seen in the studied SNPs within SLC6A3 BDNF and COMT genes. Conflicting results have been published in various studies on the influence of these genes on the increased risk of substance addiction [35–37, 42–46]. Candidate gene analysis is problematic because the prior probability of seeing true association is exceptionally low , unless a very strong case of specific phenotype for involvement of a particular gene can be made. This is not applied to substance addiction because compelling biological evidence implicating particular neurotransmitter receptors in addiction is absent, with the possible exception of the opioid receptor gene family, and prior probability is impossible to determine . Thus p-values of 0.05 are more likely to be chance occurrences, especially when using cases and controls where hidden population stratification as confounding factor is an inherent danger. However, a risk of population stratification as a confounding factor was not found in this study because the Jordanian Arab population are considered to be genetically homogenous population. This offers an advantage for genetic studies. For example, the numbers of different variations in the genes behind phenotypes are expected to be smaller than in more heterogeneous populations. This increases the probability to find genetic associations . Therefore, even a small study sample from a genetically homogenous population, like the sample of subjects used in this study, can give accurate results.
In this study, genotyping was carried out by sequenom MassARRAY® system for 49 SNPs. The NCBI, dbSNP, HapMap databases and previous published data were used to select the studied SNPs, yielding reliable candidate SNPs database for genetic association analysis. In this array we focused on genes of particular interest for drug, alcohol and neuropsychiatric researchers because they were reported to be involved in drug dependence and other neurological and psychiatric disorders [4–12]. The chosen SNPs were also selected because they showed the greatest potential to distinguish between substance addicts’ individuals and control subjects in previous studies [4–12]. Therefore, the distribution of SNPs through the selected genes was optimal.
Various studies showed a risk of false positive results due to population stratification. However, a risk of false positive results was not found in this study because genotypic frequencies of chosen SNPs in the patients and controls met HWE expectations. In addition, it is likely that there were genotyping errors. However, genotyping errors were minimized by genotyped each patient twice in order to avoid technical errors as evidenced by the low average rate of genotype discrepancy. Genotyping was conducted for patients under the same conditions and during the same period. Genotypes were also evaluated by investigators who were blind to the status of the subject and any discrepancies were resolved by test replication.
A confounding factor which could have contributed to the observed variations in the between this study and previous studies is the heterogeneity of population based on gender [66, 67]. However in our study, only male individuals with substance addiction were genotyped. Therefore, the generalisation of the results to all substance addicts’ individuals is limited. Another confounding factor is differences in phenotype in addiction such as polysubstance use, severity of addiction and the use of unstructured clinical interviews to obtain phenotypic data could affect the genetic association analysis. However, these confounding factors are not found in our study as a specific clinical structural interview was designed based on the DSM-IV criteria and the Addiction Severity Index (ASI) for collecting clinical and phenotypic data . The careful and extensive interview based phenotypic data collection has been performed by highly trained psychiatrist consultants, yielding exceptionally reliable phenotype data. In addition, the study sample is strongly enriched with regular substance addicts’ individuals giving more statistical power.