- Split View
-
Views
-
Cite
Cite
Christopher D. Swagell, Bruce R. Lawford, Ian P. Hughes, Joanne Voisey, Gerald F.X. Feeney, Angela van Daal, Jason P. Connor, Ernest P. Noble, C. Phillip Morris, Ross McD. Young, DRD2 C957T and TaqIA Genotyping Reveals Gender Effects and Unique Low-Risk and High-Risk Genotypes in Alcohol Dependence, Alcohol and Alcoholism, Volume 47, Issue 4, July/August 2012, Pages 397–403, https://doi.org/10.1093/alcalc/ags047
- Share Icon Share
Abstract
Aims: As recent conflicting reports describe a genetic association between both the C- and the T-alleles of the dopamine D2 receptor (DRD2) C957T polymorphism (rs6277) in alcohol-dependent subjects, our aim was to examine this polymorphism and TaqIA (rs1800497) in Australian alcohol-dependent subjects. Methods: The C957T polymorphism was genotyped in 228 patients with alcohol dependence (72 females and 156 males) and 228 healthy controls. Results: The C-allele and C/C genotype of C957T was associated with alcohol dependence, whereas the TaqIA polymorphism was not. When analysed separately for C957T, males showed an even stronger association with the C-allele and females showed no association. The C957T and TaqIA haplotyping revealed a strong association with alcohol dependence and a double-genotype analysis (combining C957T and TaqIA genotypes) revealed that the relative risk of different genotypes varied by up to 27-fold with the TT/A1A2 having an 8.5-fold lower risk of alcohol dependence than other genotypes. Conclusion: Decreased DRD2 binding associated with the C-allele of the DRD2 C957T polymorphism is likely to be important in the underlying pathophysiology of at least some forms of alcohol dependence, and this effect appears to be limited to males only.
INTRODUCTION
Alcohol dependence is a common, chronic, disabling illness that is one of the most costly public health problems (Grant et al., 2004). Twin and adoption studies indicate that heritable factors account for 50–60% of the risk for developing alcoholism and potential susceptibility genes have been identified by association and linkage studies (Quickfall and el-Guebaly, 2006). Several lines of evidence implicate the DRD2 gene as a candidate gene for susceptibility. Alcohol-induced drug reward is largely mediated via DRD2 (Blum et al., 2000) and neuroimaging evidence indicates that patients with alcohol dependence have decreased striatal DRD2 density (Repo et al., 1999). Conversely, unaffected offspring of alcohol-dependent fathers have increased striatal DRD2 density compared with those without a family history of this disorder, indicating that higher levels of DRD2 availability may protect against the development of alcohol dependence (Volkow et al., 2006). Finally, animal data reveal that DRD2 over-expression in the nucleus accumbens core reduces alcohol consumption in both alcohol-preferring and non-preferring rats (Thanos et al., 2004).
In vivo positron emission tomography studies have found that the C-allele of C957T (rs6277), a polymorphism in the DRD2 gene is associated with a lower striatal DRD2-binding potential than the T-allele in healthy subjects (Hirvonen et al., 2004). Contrary findings in an in vitro study reported an increase in DRD2 mRNA stability and protein translation associated with the C-allele (Duan et al., 2003). A recent within-family association study reported an association of the T-allele and alcohol dependence, but it could not demonstrate a population-level association (Hill et al., 2008).
Another association study of alcohol-dependent patients (Berggren et al., 2006) confirmed earlier findings that a polymorphism in the DRD2 region, TaqIA, is associated with alcohol dependence (Blum et al., 1990). This association has also been confirmed by several meta-analyses (Young et al., 2004; Munafo et al., 2007; Smith et al., 2008; Le Foll et al., 2009), but it has not always been reproduced by individual studies (Anghelescu et al., 2001; Haberstick et al., 2007). TaqIA is a polymorphism in the ankyrin repeat and protein kinase domain-containing protein 1 gene (ANKK1) that is found immediately downstream from DRD2. In a pharmacogenetic study, TaqIA was associated with the treatment response to bromocriptine, a pre-synaptic DRD2 agonist (Lawford et al., 1995). The greatest improvement in craving and anxiety occurred in patients receiving bromocriptine and carrying the TaqIA A1-allele. A recent study found that both the C957T C-allele and the TaqIA A1-allele were associated with alcohol dependence, although the C957T polymorphism appeared to have a stronger effect (Ponce et al., 2009).
As C957T and TaqIA have previously been implicated in DRD2 function, dopamine response and alcohol dependence, we investigated the role of these polymorphisms in a case–control genetic association study of a treatment-seeking alcohol-dependent population.
METHODS
Subjects
A total of 228 alcohol-dependent subjects were recruited from large public hospitals in Brisbane, Australia. All subjects met DSM-IV criteria for alcohol dependence. There were 72 females and 156 males in this group with a mean age of 40.7 years (SD 10.4 years; range 20–68 years). All were hospital inpatients and represented a spectrum of disease severity. A significant proportion (n = 65) of these patients were diagnosed with two or more alcohol-related medical conditions such as pancreatitis, cirrhosis, hepatitis or peripheral neuropathy. Alcohol-dependent patients were excluded from the study if they had dementia, delirium, psychosis or any other condition that would affect their ability to provide informed consent.
The control group included 228 subjects (94 females and 134 males) with a mean age of 36.8 years (SD 12.8 years; range 18–66 years). The control group was composed of volunteers from the general public, hospital staff, university staff and university students. Formal screening for alcohol dependence or other psychological disorders was not undertaken in the control population. To minimize the population stratification bias, both control and clinical subjects were recruited in the Brisbane region (a city of ∼2 million inhabitants on the East Coast of Australia). All subjects reported being of British or European descent.
Each participant gave written informed consent. Ethics approval was obtained from all institutions involved (QUT ethics number 0700000611).
Genotyping
Genotyping of the C957T polymorphism was performed by allele-specific kinetic real-time PCR (Lawford et al., 2005) using the Applied Biosystems 7000 sequence detection system (Applied Biosystems, Foster City, CA, USA). Sequence-specific primers were designed for the C-allele (5′-ATGGTCTCCACAGCACTCTC-3′), the T-allele (5′-ATGG TCTCCACAGCACTCTT-3′) and a common reverse primer (5′-CATTGGGCATGGTCTGGATC-3′). A total of 5–10 ng of genomic DNA was amplified in 1× SYBR green PCR master mix (Applied Biosystems) containing 0.4 μM of allele-specific forward primer and 0.4 μM of common reverse primer in a 25 μl volume. Amplification conditions were: step 1: 50°C for 2 min, step 2: 95°C for 10 min, step 3: 95°C for 15 s, step 4: 60°C for 1 min, steps 3 and 4 were repeated by 40 cycles. A cycle time () value was obtained by setting the threshold during the geometric phase of amplification and scored relative to the generated between the matched and mismatched primer pairs.
TaqI A genotyping was performed by restriction fragment length polymorphism analysis of PCR products. A genomic sequence of 501 bp of the coding region of ANKK1 was amplified by PCR using the forward primer 5′-GCACGTGC CACCATACCC-3′ and the reverse primer 5′-TGCAGAGC AGTCAGGCTG-3′. A total of 5–10 ng of genomic DNA was amplified in a PCR master mix containing 0.2 μM of forward primer and 0.2 μM of reverse primer, 1× PCR buffer, 1.5 mM MgCl2, 200 μM dNTPs and 1 unit of Platinum Taq DNA Polymerase (Invitrogen) in a 25 μl volume. Amplification conditions were: step 1: 94°C for 4 min, step 2: 94°C for 30 s, step 3: 68°C for 30 s, step 4: 72°C for 30 s, steps 2–4 were repeated by 40. Amplified PCR fragments were digested with TaqI restriction enzyme (New England Biolabs) and digested fragments were visualized via agarose gel electrophoresis.
Statistical analysis
For each polymorphism, a homogeneity χ2 analysis was employed to test the hypothesis of homogeneity of allele frequency distributions between alcohol dependent and control populations. An odds ratio (OR) with 95% confidence interval (CI) was also calculated. Genotype frequencies were similarly compared between alcohol dependent and control populations. An extended Mantel–Haenszel test was performed to test for a trend to alcohol dependence with increasing copy numbers (0, 1 or 2) of the allele. ORs were calculated relative to the homozygote for the non-associated allele and population attributable fractions calculated. Genotype data were entered into JLIN v1.6.0 (Carter et al., 2006) to estimate haplotype frequencies. Haplotype frequencies were compared between alcohol dependent and control groups using a homogeneity χ2 analysis. Each haplotype with a significant association was analysed in isolation by partitioning the χ2 accordingly. Generated haplotypes were analysed for linkage disequilibrium (LD) measures (D′ and r2) using the JLIN (Carter et al., 2006). The calculations for attributable risk were performed using genotypes, counting the presence of the C-allele as a risk factor. Statistical tests were performed using the COMPARE2 program from the WinPepi suite of epidemiology programs (Abramson, 2004) and SPSS version 16. Hardy–Weinberg equilibrium (HWE) was computed using Utility Programs for Analysis of Genetic Linkage (Ott, 1988). The analysis of genotypes under a recessive model involved pooling genotypes and comparing frequencies using a two-by-two contingency table. For example, for the C957T polymorphism, the T/T and C/T genotypes were pooled for both groups and compared with the C/C genotype.
RESULTS
C957T genotyping
In order to find genetic associations between alcohol dependence and polymorphisms in the DRD2 region, the DRD2 C957T polymorphism was genotyped in 228 unrelated alcohol-dependent subjects and 228 unrelated Caucasian healthy control subjects. The frequency of the C957T alleles in the alcohol-dependent and control groups (Table 1) revealed that the C-allele was found more frequently in the alcohol-dependent group (50% of alleles) compared with the control group (42% of alleles) with an OR of 1.37 (P = 0.019). The genotype frequencies of both groups are displayed in Table 2. The C/C genotype was over-represented in the alcohol-dependent group when compared with the control sample, with an OR of 1.9 (P = 0.034). A simple comparison of the genotypes between alcohol dependent and control subjects just achieved significance (χ2 = 6.33, P = 0.042) but this association became more obvious when analysed using a Mantel–Haenszel test for trend (χ2 = 5.23, P = 0.022). The OR for the C/T genotype (1.27) was not significantly different from the T/T genotype (OR = 1) indicating that the C-allele appears to behave as a recessive allele. When analysed under a C-allele recessive model, the C/C genotype is almost twice as likely to be found in those with alcohol dependence (OR = 1.75, 95% CI, 1.09–2.82; P = 0.014). The attributable fraction, which is the proportion of cases in the study population that would not have occurred had the risk factor (risk allele) not been present, is a useful measure to quantify the impact of the allele on the development of alcohol dependence. Analysis of the genotype data revealed that 13% (95% CI, 3.37–22.09%, P = 0.004) of the susceptibility to alcohol dependence could be attributed to having the C/C genotype of C957T polymorphism. Both alcohol-dependent and control groups were shown to be in HWE based on the respective genotype frequencies of each group (controls, χ2 = 0.077, P = 0.78; alcohol dependence, χ2 = 1.42, P = 0.23).
. | . | Allele frequency . | . | . | . | |
---|---|---|---|---|---|---|
C957T (rs6277) . | n . | C (%) . | T (%) . | χ2 . | P-value . | OR (95% CIa) . |
Control | 454 | 193 (42.5) | 261 (57.5) | 5.47 | 0.019 | 1.37 (1.04–1.79) |
Alcohol dependence | 456 | 229 (50.2) | 227 (49.8) | |||
TaqIA (rs1800497) | A1b (%) | A2 (%) | ||||
Control | 454 | 77 (17.0) | 377 (83.0) | 1.942 | 0.163 | 1.27 (0.90–1.80) |
Alcohol dependence | 452 | 93 (20.6) | 359 (79.4) |
. | . | Allele frequency . | . | . | . | |
---|---|---|---|---|---|---|
C957T (rs6277) . | n . | C (%) . | T (%) . | χ2 . | P-value . | OR (95% CIa) . |
Control | 454 | 193 (42.5) | 261 (57.5) | 5.47 | 0.019 | 1.37 (1.04–1.79) |
Alcohol dependence | 456 | 229 (50.2) | 227 (49.8) | |||
TaqIA (rs1800497) | A1b (%) | A2 (%) | ||||
Control | 454 | 77 (17.0) | 377 (83.0) | 1.942 | 0.163 | 1.27 (0.90–1.80) |
Alcohol dependence | 452 | 93 (20.6) | 359 (79.4) |
aFisher's exact confidence interval.
bThe A1-allele is T and the A2 allele is C.
. | . | Allele frequency . | . | . | . | |
---|---|---|---|---|---|---|
C957T (rs6277) . | n . | C (%) . | T (%) . | χ2 . | P-value . | OR (95% CIa) . |
Control | 454 | 193 (42.5) | 261 (57.5) | 5.47 | 0.019 | 1.37 (1.04–1.79) |
Alcohol dependence | 456 | 229 (50.2) | 227 (49.8) | |||
TaqIA (rs1800497) | A1b (%) | A2 (%) | ||||
Control | 454 | 77 (17.0) | 377 (83.0) | 1.942 | 0.163 | 1.27 (0.90–1.80) |
Alcohol dependence | 452 | 93 (20.6) | 359 (79.4) |
. | . | Allele frequency . | . | . | . | |
---|---|---|---|---|---|---|
C957T (rs6277) . | n . | C (%) . | T (%) . | χ2 . | P-value . | OR (95% CIa) . |
Control | 454 | 193 (42.5) | 261 (57.5) | 5.47 | 0.019 | 1.37 (1.04–1.79) |
Alcohol dependence | 456 | 229 (50.2) | 227 (49.8) | |||
TaqIA (rs1800497) | A1b (%) | A2 (%) | ||||
Control | 454 | 77 (17.0) | 377 (83.0) | 1.942 | 0.163 | 1.27 (0.90–1.80) |
Alcohol dependence | 452 | 93 (20.6) | 359 (79.4) |
aFisher's exact confidence interval.
bThe A1-allele is T and the A2 allele is C.
. | . | Genotype frequency . | . | ||
---|---|---|---|---|---|
C957T (rs6277) . | n . | C/C (%) . | C/T (%) . | T/T (%) . | χ2 (P) . |
Control | 227 | 40 (17.6) | 113 (49.8) | 74 (32.6) | 6.33a (0.042) 5.23b (0.022) |
Alcohol dependence | 228 | 62 (27.2) | 105 (46.0) | 61 (26.8) | |
OR (P) | 1.88 (0.034) | 1.27 (1.00) | 1.00 | ||
Alcohol- dependent males | 156 | 47 (30.1) | 69 (44.2) | 40 (25.7) | 8.37a (0.015) 6.72b (0.009) |
OR (P) | 2.17 (0.014) | 1.13 (1.00) | 1.00 | ||
Alcohol- dependent females | 72 | 15 (20.8) | 36 (50.0) | 21 (29.2) | 0.505a (0.777) 0.497b (0.481) |
OR (P) | 1.32 (0.955) | 1.12 (1.00) | 1.00 | ||
TaqIA (rs1800497) | A1/A1 (%) | A1/A2 (%) | A2/A2 (%) | ||
Control | 227 | 8 (3.5) | 61 (26.9) | 158 (69.6) | 1.88a (0.390) 1.86b (0.173) |
Alcohol dependence | 226 | 11 (4.9) | 71 (31.4) | 144 (63.7) | |
OR (P) | 1.88 (0.034) | 1.14 (1.00) | 1.00 |
. | . | Genotype frequency . | . | ||
---|---|---|---|---|---|
C957T (rs6277) . | n . | C/C (%) . | C/T (%) . | T/T (%) . | χ2 (P) . |
Control | 227 | 40 (17.6) | 113 (49.8) | 74 (32.6) | 6.33a (0.042) 5.23b (0.022) |
Alcohol dependence | 228 | 62 (27.2) | 105 (46.0) | 61 (26.8) | |
OR (P) | 1.88 (0.034) | 1.27 (1.00) | 1.00 | ||
Alcohol- dependent males | 156 | 47 (30.1) | 69 (44.2) | 40 (25.7) | 8.37a (0.015) 6.72b (0.009) |
OR (P) | 2.17 (0.014) | 1.13 (1.00) | 1.00 | ||
Alcohol- dependent females | 72 | 15 (20.8) | 36 (50.0) | 21 (29.2) | 0.505a (0.777) 0.497b (0.481) |
OR (P) | 1.32 (0.955) | 1.12 (1.00) | 1.00 | ||
TaqIA (rs1800497) | A1/A1 (%) | A1/A2 (%) | A2/A2 (%) | ||
Control | 227 | 8 (3.5) | 61 (26.9) | 158 (69.6) | 1.88a (0.390) 1.86b (0.173) |
Alcohol dependence | 226 | 11 (4.9) | 71 (31.4) | 144 (63.7) | |
OR (P) | 1.88 (0.034) | 1.14 (1.00) | 1.00 |
aLikelihood-ratio χ2 test.
bMantel–Haenszel χ2 test for trend.
. | . | Genotype frequency . | . | ||
---|---|---|---|---|---|
C957T (rs6277) . | n . | C/C (%) . | C/T (%) . | T/T (%) . | χ2 (P) . |
Control | 227 | 40 (17.6) | 113 (49.8) | 74 (32.6) | 6.33a (0.042) 5.23b (0.022) |
Alcohol dependence | 228 | 62 (27.2) | 105 (46.0) | 61 (26.8) | |
OR (P) | 1.88 (0.034) | 1.27 (1.00) | 1.00 | ||
Alcohol- dependent males | 156 | 47 (30.1) | 69 (44.2) | 40 (25.7) | 8.37a (0.015) 6.72b (0.009) |
OR (P) | 2.17 (0.014) | 1.13 (1.00) | 1.00 | ||
Alcohol- dependent females | 72 | 15 (20.8) | 36 (50.0) | 21 (29.2) | 0.505a (0.777) 0.497b (0.481) |
OR (P) | 1.32 (0.955) | 1.12 (1.00) | 1.00 | ||
TaqIA (rs1800497) | A1/A1 (%) | A1/A2 (%) | A2/A2 (%) | ||
Control | 227 | 8 (3.5) | 61 (26.9) | 158 (69.6) | 1.88a (0.390) 1.86b (0.173) |
Alcohol dependence | 226 | 11 (4.9) | 71 (31.4) | 144 (63.7) | |
OR (P) | 1.88 (0.034) | 1.14 (1.00) | 1.00 |
. | . | Genotype frequency . | . | ||
---|---|---|---|---|---|
C957T (rs6277) . | n . | C/C (%) . | C/T (%) . | T/T (%) . | χ2 (P) . |
Control | 227 | 40 (17.6) | 113 (49.8) | 74 (32.6) | 6.33a (0.042) 5.23b (0.022) |
Alcohol dependence | 228 | 62 (27.2) | 105 (46.0) | 61 (26.8) | |
OR (P) | 1.88 (0.034) | 1.27 (1.00) | 1.00 | ||
Alcohol- dependent males | 156 | 47 (30.1) | 69 (44.2) | 40 (25.7) | 8.37a (0.015) 6.72b (0.009) |
OR (P) | 2.17 (0.014) | 1.13 (1.00) | 1.00 | ||
Alcohol- dependent females | 72 | 15 (20.8) | 36 (50.0) | 21 (29.2) | 0.505a (0.777) 0.497b (0.481) |
OR (P) | 1.32 (0.955) | 1.12 (1.00) | 1.00 | ||
TaqIA (rs1800497) | A1/A1 (%) | A1/A2 (%) | A2/A2 (%) | ||
Control | 227 | 8 (3.5) | 61 (26.9) | 158 (69.6) | 1.88a (0.390) 1.86b (0.173) |
Alcohol dependence | 226 | 11 (4.9) | 71 (31.4) | 144 (63.7) | |
OR (P) | 1.88 (0.034) | 1.14 (1.00) | 1.00 |
aLikelihood-ratio χ2 test.
bMantel–Haenszel χ2 test for trend.
Gender differences
There has been widespread discussion of sexual dimorphism in alcohol dependence (Devaud and Prendergast, 2009); so we investigated the association of C957T separately in male and female alcohol-dependent subjects (Table 2). Despite decreased power due to reduced numbers, the strength of association in male alcohol-dependent subjects was increased (Mantel–Haenszel test for trend χ2 = 6.72, P = 0.009). Indeed, male subjects with the C/C genotype were 2.2 times as likely to have alcohol dependence (OR = 2.17, P = 0.014). In contrast, the female subjects showed no indication of association with this polymorphism (Mantel–Haenszel test for trend χ2 = 0.497, P = 0.481), although there were only 72 female subjects in the study.
TaqIA genotyping
As the TaqIA polymorphism has repeatedly been seen associated with alcohol dependence (Lawford et al., 1997), it was also genotyped in the same samples (one control and two alcohol-dependent samples failed to genotype). No association was found at the allele (χ2 = 1.94, P = 0.163; Table 1) or genotype level (χ2 = 1.88, P = 0.39; Table 2). The A1-allele of TaqIA is reported to be inherited dominantly (Lawford et al., 1997) so the genotypes were examined under a partial and complete dominance model. No trend in susceptibility to alcohol dependence was found when analysed under a partially dominant model (Mantel–Haenszel χ2 = 1.86, P = 0.173; Table 2). When analysed under an A1-dominant model, the A1-containing genotypes (A1/A1 and A1/A2) also showed no association with alcohol dependence (χ2 = 1.768, P = 0.184).
Haplotype analysis
The C957T polymorphism showed an association with alcohol dependence in our study and TaqIA has previously been associated with alcohol dependence, so haplotype analysis was also performed for both polymorphisms. Haplotype frequencies for the two polymorphisms are shown in Table 3. Overall haplotype frequency was highly associated with alcohol dependence (χ2 = 25.11, P = 0.00001). This association was due to the over representation of the 957C/TaqI A1 haplotype in alcohol-dependent patients. The 957C/TaqI A2, 957T/TaqI A2 and the 957T/TaqI A1 haplotypes were not significantly associated with alcohol dependence. The 957C/TaqI A1 haplotype was approximately twice as likely to be associated with alcohol dependence compared with the 957T/TaqI A2 haplotype. The 957T/TaqI A1 haplotype appears to be protective as it has an OR of 0.284, i.e. this haplotype is ∼3.5 times less likely to be found associated with alcohol dependence than the 957T/TaqI A2 haplotype (P = 0.002) and almost seven times less likely to be found associated with alcohol dependence than the 957C/TaqI A1 haplotype (OR = 6.782, P = 4.5 × 10−6).
Haplotypea . | Controls . | Alcohol . | ORb (P) . | χ2c . | P . |
---|---|---|---|---|---|
957T/TaqI A2 | 230 | 216 | 1 | 0.868 | 0.352 |
957T/TaqI A1 | 30 | 8 | 0.284 (0.002) | 14.124 | 0.0002 |
957C/TaqI A2 | 145 | 143 | 1.05 (1.000) | 0.020 | 0.886 |
957C/TaqI A1 | 47 | 85 | 1.926 (0.004) | 12.967 | 0.0003 |
Total | 452 | 452 | 25.11 | 0.00001 |
Haplotypea . | Controls . | Alcohol . | ORb (P) . | χ2c . | P . |
---|---|---|---|---|---|
957T/TaqI A2 | 230 | 216 | 1 | 0.868 | 0.352 |
957T/TaqI A1 | 30 | 8 | 0.284 (0.002) | 14.124 | 0.0002 |
957C/TaqI A2 | 145 | 143 | 1.05 (1.000) | 0.020 | 0.886 |
957C/TaqI A1 | 47 | 85 | 1.926 (0.004) | 12.967 | 0.0003 |
Total | 452 | 452 | 25.11 | 0.00001 |
aHaplotype frequencies were estimated using JLIN. Note: haplotypes with a missing genotype were omitted from the analysis (two control and two alcohol dependence samples).
bOR with respect to 957T/TaqI A1 haplotype which is more commonly associated with a normal phenotype.
cLikelihood ratio χ2 test comparing haplotype frequency between groups when all other haplotypes were pooled.
Haplotypea . | Controls . | Alcohol . | ORb (P) . | χ2c . | P . |
---|---|---|---|---|---|
957T/TaqI A2 | 230 | 216 | 1 | 0.868 | 0.352 |
957T/TaqI A1 | 30 | 8 | 0.284 (0.002) | 14.124 | 0.0002 |
957C/TaqI A2 | 145 | 143 | 1.05 (1.000) | 0.020 | 0.886 |
957C/TaqI A1 | 47 | 85 | 1.926 (0.004) | 12.967 | 0.0003 |
Total | 452 | 452 | 25.11 | 0.00001 |
Haplotypea . | Controls . | Alcohol . | ORb (P) . | χ2c . | P . |
---|---|---|---|---|---|
957T/TaqI A2 | 230 | 216 | 1 | 0.868 | 0.352 |
957T/TaqI A1 | 30 | 8 | 0.284 (0.002) | 14.124 | 0.0002 |
957C/TaqI A2 | 145 | 143 | 1.05 (1.000) | 0.020 | 0.886 |
957C/TaqI A1 | 47 | 85 | 1.926 (0.004) | 12.967 | 0.0003 |
Total | 452 | 452 | 25.11 | 0.00001 |
aHaplotype frequencies were estimated using JLIN. Note: haplotypes with a missing genotype were omitted from the analysis (two control and two alcohol dependence samples).
bOR with respect to 957T/TaqI A1 haplotype which is more commonly associated with a normal phenotype.
cLikelihood ratio χ2 test comparing haplotype frequency between groups when all other haplotypes were pooled.
Linkage disequilibrium
The TaqI A1 and the 957C alleles were in partial LD in both the control population D′ = 0.321 and r2 = 0.029 and the alcohol-dependent population D′ = 0.824 and r2 = 0.173. While LD exists in the control population it is significantly larger in the alcohol group (goodness-of-fit χ2 = 13.35, P = 0.001), suggesting a functional relationship between alleles not just LD as a consequence of genetic proximity, i.e. the polymorphisms are inherited together more strongly in cases because both polymorphisms have a (functional) role in disease, not just the C957T polymorphism.
Combined C957T and TaqIA genotypes
The combined genotypes of both the C957T polymorphism and the TaqIA polymorphism were scored for each individual, e.g. if a person was homozygous C/C for C957T polymorphism and heterozygous for the TaqIA polymorphism their combined genotype was recorded as CC/A1A2. Two alcohol-dependent patients failed to genotype for TaqIA, so 226 combined alcohol genotypes were obtained. One control sample failed to genotype for C957T and a different control sample failed to genotype for TaqIA, so 226 combined control genotypes were obtained (Table 4). As there are nine possible combined genotypes, the counts for each genotype are quite low and only two reached significance, the CC/A1A2 genotype and the TT/A1A2 genotype. Because of the very low frequency of the 957T/TaqI A1 haplotype (∼7% in controls and 2% in alcohol-dependent samples), no controls or alcohol-dependent samples had the TT/A1A1-combined genotype.
Genotype . | Control . | Alcohol . | OR . | χ2a . | P . |
---|---|---|---|---|---|
CC/A1A1 | 4 | 7 | 1.750 | 0.758 | 1.000 |
CC/A1A2 | 10 | 34 | 3.400 | 10.219 | 0.01 |
CC/A2A2 | 26 | 21 | 0.808 | 0.380 | 1.000 |
CT/A1A1 | 4 | 4 | 1.000 | 0.0 | 1.000b |
CT/A1A2 | 35 | 35 | 1.000 | 0.0 | 1.000 |
CT/A2A2 | 73 | 65 | 0.890 | 0.212 | 1.000 |
TT/A1A1 | 0 | 0 | |||
TT/A1A2 | 16 | 2 | 0.125 | 10.93 | 0.007 |
TT/A2A2 | 58 | 58 | 1 | ||
Total | 226 | 226 | 28.053 | 0.0002 |
Genotype . | Control . | Alcohol . | OR . | χ2a . | P . |
---|---|---|---|---|---|
CC/A1A1 | 4 | 7 | 1.750 | 0.758 | 1.000 |
CC/A1A2 | 10 | 34 | 3.400 | 10.219 | 0.01 |
CC/A2A2 | 26 | 21 | 0.808 | 0.380 | 1.000 |
CT/A1A1 | 4 | 4 | 1.000 | 0.0 | 1.000b |
CT/A1A2 | 35 | 35 | 1.000 | 0.0 | 1.000 |
CT/A2A2 | 73 | 65 | 0.890 | 0.212 | 1.000 |
TT/A1A1 | 0 | 0 | |||
TT/A1A2 | 16 | 2 | 0.125 | 10.93 | 0.007 |
TT/A2A2 | 58 | 58 | 1 | ||
Total | 226 | 226 | 28.053 | 0.0002 |
aLikelihood-ratio χ2.
bAt least one cell has an expected frequency <5.
Genotype . | Control . | Alcohol . | OR . | χ2a . | P . |
---|---|---|---|---|---|
CC/A1A1 | 4 | 7 | 1.750 | 0.758 | 1.000 |
CC/A1A2 | 10 | 34 | 3.400 | 10.219 | 0.01 |
CC/A2A2 | 26 | 21 | 0.808 | 0.380 | 1.000 |
CT/A1A1 | 4 | 4 | 1.000 | 0.0 | 1.000b |
CT/A1A2 | 35 | 35 | 1.000 | 0.0 | 1.000 |
CT/A2A2 | 73 | 65 | 0.890 | 0.212 | 1.000 |
TT/A1A1 | 0 | 0 | |||
TT/A1A2 | 16 | 2 | 0.125 | 10.93 | 0.007 |
TT/A2A2 | 58 | 58 | 1 | ||
Total | 226 | 226 | 28.053 | 0.0002 |
Genotype . | Control . | Alcohol . | OR . | χ2a . | P . |
---|---|---|---|---|---|
CC/A1A1 | 4 | 7 | 1.750 | 0.758 | 1.000 |
CC/A1A2 | 10 | 34 | 3.400 | 10.219 | 0.01 |
CC/A2A2 | 26 | 21 | 0.808 | 0.380 | 1.000 |
CT/A1A1 | 4 | 4 | 1.000 | 0.0 | 1.000b |
CT/A1A2 | 35 | 35 | 1.000 | 0.0 | 1.000 |
CT/A2A2 | 73 | 65 | 0.890 | 0.212 | 1.000 |
TT/A1A1 | 0 | 0 | |||
TT/A1A2 | 16 | 2 | 0.125 | 10.93 | 0.007 |
TT/A2A2 | 58 | 58 | 1 | ||
Total | 226 | 226 | 28.053 | 0.0002 |
aLikelihood-ratio χ2.
bAt least one cell has an expected frequency <5.
The TT/A2A2-combined genotype was chosen as the OR = 1 low-risk or ‘normal’ genotype. Examination of the ORs of each genotype reveals that there are five genotypes with a similar OR to TT/A2A2, i.e. close to 1.0. The OR of TT/A2A2, CC/A2A2, CT/A1A1, CT/A1A2 and CT/A2A2 vary between 0.808 and 1.000. These can be considered ‘normal’-risk genotypes. The CC/A1A1 genotype has an OR of 1.75 and the CC/A1A2 genotype has an OR of 3.4, so these genotypes can be considered high-risk genotypes. The TT/A1A2 genotype has an OR of 0.125, meaning that people with this genotype are eight times less likely to develop alcohol dependence than people with a normal-risk genotype. Compared with people with the high-risk CC/A1A2 genotype, those with the TT/A1A2 genotype are almost 30 times less likely to develop alcohol dependence (OR = 27.2, χ2 = 24.608, P = 5 × 10−6).
When the high-risk CC/A1A1 and CC/A1A2 genotypes were combined and compared with the remaining genotypes, people with these genotypes were more than three times as likely to have alcohol dependence (OR = 3.36, 95% CI = 1.72–6.87, χ2 = 15.68, P = 0.00008,) and the attributable fraction in the population was 12.7% (95% CI = 6.4–18.6%).
DISCUSSION
An analysis of individual polymorphisms identified a modest association between the presence of the C-allele of the DRD2 C957T polymorphism and alcohol dependence. The A1-allele of the TaqIA polymorphism has previously been associated with alcohol dependence and we observed that it was more abundant in alcohol-dependent patients, but this observation did not reach significance in our study. This may be because the A1-allele has a relatively low allele frequency resulting in a small effect size (Berggren et al., 2006). Consequently, the number of subjects in this study might have been too small to detect an association. However, when the data from both polymorphisms was analysed together by haplotype and combined genotype analysis, the observed association was much stronger. This observation lends support to the argument that both polymorphisms are important in alcohol dependence and confirms the DRD2 gene and adjacent region (ANKK1) as an alcohol dependence susceptibility gene.
While the functional role of DRD2 has been well defined in addiction and other reward-related disorders, it is less clear what role might be played by the TaqIA polymorphism in the ANKK1 gene. ANKK1 is a serine/threonine kinase that is likely to be involved in signal transduction. The TaqIA polymorphism causes a non-conservative (Glu to Lys) amino acid change in the 11th ankyrin repeat of ANKK1. This might affect substrate binding but it is unlikely to affect the structural integrity of the protein (Neville et al., 2004). While it is plausible that ANKK1 is involved in transduction of signals in the dopaminergic reward pathway, the failure to detect its expression in whole brain homogenates does not support this hypothesis (Neville et al., 2004). The most plausible explanation is that the TaqIA polymorphism is only associated with alcohol dependence because it is in disequilibrium with C957T, especially since the association seen for TaqIA is not as strong as the C957T polymorphism. However, the LD between C957T and TaqIA seen in alcohol-dependent cases is almost three times higher than in controls (D′ = 0.824 in cases compared with D′ = 0.321 in controls), suggesting that TaqIA has a role in disease independent of the C957T polymorphism. An alternative explanation for the previously observed association with TaqIA is that the ANKK1 gene has no direct role in addiction but the proximity of TaqIA to C957T in the DRD2 gene (they are only separated by ∼12.5 kb) may result in a change in the regulation of the DRD2 gene.
DRD2 is present as both presynaptic autoreceptors and postsynaptic receptors (Grigoriadis and Seeman, 1984). Dopaminergic activation of presynaptic autoreceptors decreases both the release and synthesis of dopamine and increase dopamine reuptake (Tissari et al., 1983). It is known that low-DRD2 binding is associated with increased drug reinforcing effects (Volkow et al., 1999a) and the intensity of this reinforcement is proportional to dopamine release (Volkow et al., 1999b). Reduced density of presynaptic DRD2 associated with the C/C genotype of C957T may result in decreased autoreceptor activity, thereby resulting in increased dopaminergic transmission and increased drug reward.
Abstinent alcohol-dependent patients have decreased DRD2 availability. Low DRD2 availability is also associated with increased craving for alcohol and increased risk of relapse (Heinz et al., 2005). Furthermore, alcohol-dependent patients exhibit diminished dopamine release following amphetamine challenge, reflecting decreased dopaminergic neurotransmission (Martinez et al., 2005). Animal studies also show lower density of DRD2 in alcohol-preferring rodents than in non-alcohol preferring animals (Boehme and Ciaranello, 1981; Stefanini et al., 1992; McBride et al., 1993) and that the administration of DRD2 agonists reduces alcohol consumption in alcohol-preferring rats (Dyr et al., 1993).
The C/C genotype of DRD2 C957T has been associated with increased cigarette smoking in response to negative mood, consistent with an enhanced risk for addictive behaviour, at least in a sub-group of smokers (Perkins et al., 2008). We have also recently found an association between C957T and TaqIA genotypes and haplotypes in nicotine-dependent subjects (Voisey et al., 2012). Conversely, homozygotes for the T-allele have shown enhanced avoidance learning compared with those carrying the C-allele, indicating enhanced learning from error in this sub-group (Frank et al., 2007). Impaired capacity to learn from experience is frequently seen in those with substance misuse.
Our findings are supported by a recent study that also found a strong association between the C957T C-allele and alcohol dependence (Ponce et al., 2008). They also found a weaker association with the TaqIA A1-allele. However, two reports conflict with our observations. It has been reported that the C-allele results in higher mRNA stability in vitro (Duan et al., 2003), which could translate into higher levels of the receptor. However, it is possible that individual cells in the striatum may not respond to RNA instability elements in the same manner in vivo. Also, while we observed an association between the C-allele of C957T and alcohol dependence, a recent report found an association of the T-allele in a family study, but they could not demonstrate a population-level association using limited numbers of subjects (81 cases and 78 controls) (Hill et al., 2008).
When we analysed male and female subjects separately, we found an even stronger association between C957T and alcohol dependence in males despite decreased power due to reduced numbers, while the female subjects showed no sign of any association. It is well recognized that incidence and risk of developing alcohol dependence varies greatly between men and women (Limosin et al., 2002; Lopez-Castroman et al., 2009). This may be due to significant environmental differences or the impact of those environmental differences on different susceptibility genes. Alternatively, a more biological difference in brain function between the sexes may be a more salient contributor. In this respect, there is growing evidence for in vivo gender differences in D2 dopamine receptor characteristics. In a positron emission tomography study, women had higher D2 dopamine-binding potentials than men in the frontal cortex (Kaasinen et al., 2001) and lower D2 dopamine-binding potentials in striatal regions (Pohjalainen et al., 1998). It is suggested (Limosin et al., 2002; Lopez-Castroman et al., 2009) that these differences may be reflected in gender-associated differences found in alcohol dependence and in the present study.
Our study is one of only a few to demonstrate clear gender differences for particular susceptibility genes. Some studies have examined male or female risk alone but few studies have directly compared genetic risk for particular polymorphisms in both men and women with alcohol dependence. However, in one Spanish study of alcohol-dependent subjects, an interleukin-1 receptor polymorphism was associated in male subjects but not female subjects (Saiz et al., 2009).
By comparing the ORs of the combined genotypes of the C957T and TaqIA polymorphisms, we observed five genotypes (representing ∼87% of controls and 81% of alcohol cases) that had a neutral risk status for alcohol dependence, i.e. an OR of about one. There were two high-risk genotypes (representing ∼6% of controls and 18% of alcohol cases) that had about 3.5 times the chance of having alcohol dependence and, interestingly, we found one novel low-risk genotype (representing ∼7% of controls and <1% of alcohol cases) that had an 8.5 times lower risk of having alcohol dependence.
While our data show strong association between DRD2 genotypes and alcohol dependence, the number of subjects is still modest by current expectations. In addition, the control population was not formally screened for alcohol dependence or other psychological disorders, which would be expected to decrease the power to detect associations.
Our results and the data of others suggest that the D2 receptor is centrally involved in the development of alcohol dependence, and this has implications for pharmacological intervention strategies. Our findings support the conclusion that the DRD2 and the dopamine pathway are important in defining genetic susceptibility to alcohol dependence, but only in males.
Authors' roles
All authors made a substantial contribution to drafting the manuscript or reviewing it critically gave final approval of the version of the article to be published and can certify that no other individuals not listed as authors have made substantial contributions to the paper. In addition: R.M.Y., B.R.L., C.D.S., A.v.D. and C.P.M.—made a substantial contribution to the conception and design, the acquisition of data and the analysis and interpretation of the data; I.P.H. —made a substantial contribution to the analysis and interpretation of the data; J.P.C., F.X.F. and E.P.N.—made a substantial contribution to the acquisition of the data.
Funding
This work was supported by the Institute of Health and Biomedical Innovation (IHBI), QUT and AstraZeneca Australia. IHBI and AstraZeneca Australia had no further role in study design; in the collection, analysis and interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.
Conflict of interest statement. All authors declare that they have no conflicts of interest.
Acknowledgements
We acknowledge Aiden Morris for his assistance with data analysis and the Institute of Health and Biomedical Innovation (IHBI), QUT and AstraZeneca Australia for their financial support of this project.