Significant outcomes
-
∙ The frequencies of the AUTS2 haplotypes were significantly different between alcohol dependence and control subjects.
-
∙ The patients with alcohol dependence showed a significantly higher frequency of the A-A haplotype (rs6943555 and rs9886351) as compared with the control group.
-
∙ Our study suggested that the rs6943555 and rs9886351 A-A haplotype in the AUTS2 gene might affect the vulnerability to the alcohol dependence pathogenesis.
Limitations
-
∙ The sample size of alcohol-dependent patients and control subjects in our study was small and consequently have low statistical power.
-
∙ In addition, it must be considered that the structure of the subject have several bias such as gender and age.
-
∙ Larger scale comparison of subjects of the same generation with no gender bias is needed in a further study to confirm our current results.
Introduction
Alcohol dependence is a form of drug addiction that involves physical and psychological dependence on alcohol, and loss of control as to alcohol consumption (Reference Nishizawa, Han and Hasegawa1). Several previous reviews have revealed that the heritability of alcohol dependence accounts for 40–60% of the total etiologic variance (Reference Agrawal and Lynskey2–Reference Goldman, Oroszi and Ducci4). A number of genes are associated with the development of alcohol dependence through interaction with environmental factors (Reference Nishizawa, Han and Hasegawa1).
The autism susceptibility candidate 2 (AUTS2) gene has been implicated in multiple neurological disorders including autism (Reference Oksenberg and Ahituv5). Recent studies have shown that the AUTS2 gene is associated with autism (Reference Ben-David, Granot-Hershkovitz and Monderer-Rothkoff6), schizophrenia (Reference Zhang, Xu and Wei7) and suicide (Reference Chojnicka, Gajos and Strawa8). In addition, several lines of evidence suggest that AUTS2 also plays some role in the development of drug addiction. The results of a previous small-scale study indicated that AUTS2 expression in lymphoblast cell lines is significantly correlated with nicotine dependence and cannabis dependence (Reference Philibert, Ryu and Yoon9). The study also suggested that AUTS2 expression tends to be significantly associated with alcohol dependence (Reference Philibert, Ryu and Yoon9). In several very recent studies, Chen et al. discovered that the average of relative cDNA level of AUTS2 in lymphoblastoid cell lines was significantly reduced in patients with heroin dependence compared with healthy controls (Reference Chen, Liao, Lai and Chen10). Additionally, Chen et al. and Dang et al. reported that the polymorphism rs6943555 in the AUTS2 gene might increase susceptibility to the development of heroin dependence (Reference Chen, Liao, Lai and Chen10,Reference Dang, Zhang, Zhu and Lu11). Interestingly, Schumann et al. have found that AUTS2 expression in human brain tissue differs depending on genotype (Reference Schumann, Coin and Lourdusamy12). They also found significant differences in mice selected for voluntary alcohol consumption differences in expression of AUTS2 (Reference Schumann, Coin and Lourdusamy12). Furthermore, their large-scale meta-analysis including 12 population-based samples of European ancestry comprising 26 316 individuals with replication genotyping in an additional 21 185 individuals indicated that the polymorphism AUTS2 rs6943555 is significantly associated with alcohol consumption by individuals, with genome-wide significance (Reference Schumann, Coin and Lourdusamy12). They also revealed that down-regulation of an AUTS2 homolog induced reduced alcohol sensitivity in Drosophila (Reference Schumann, Coin and Lourdusamy12). Thus, these observations suggest that the AUTS2 gene is related to the regulation of alcohol drinking behaviour (Reference Chen, Liao, Lai and Chen10,Reference Schumann, Coin and Lourdusamy12), and we hypothesised that polymorphisms in the AUTS2 gene also affect the vulnerability to alcohol dependence pathogenesis.
The AUTS2 gene is located on chromosome 7q11.22 and consists of 19 exons, the first six exons being separated by very large introns and the last 13 exons being close (Reference Dang, Zhang, Zhu and Lu11). Among the several known AUTS2 gene polymorphisms, we focussed on two single nucleotide polymorphisms (SNPs), which have been studied extensively. One of these common SNPs in the AUTS2 gene is rs6943555, which comprises a single nucleotide change of T→A. Schumann et al. reported that the minor A allele of rs6943555 significantly increases AUTS2 gene expression in the prefrontal cortex of the human brain compared with the T allele (Reference Schumann, Coin and Lourdusamy12). Meanwhile, Chen et al. found that subjects with the rs6943555 A/A genotype exhibit significantly lower AUTS2 mRNA level in lymphoblastoid cell lines compared to subjects with the T/T and T/A genotypes (Reference Chen, Liao, Lai and Chen10). Although the AUTS2 gene expression might not be equal between brain tissue and lymphoblastoid cell lines (Reference Chen, Liao, Lai and Chen10), in any event, these studies suggest that the single nucleotide exchange of T to A in rs6943555 variants may influence transcriptional activity and expression of the AUTS2 gene. In addition to rs6943555, we selected the polymorphism rs9886351 which was adjacent to rs6943555 (Reference Zhang, Xu and Wei7) and comprises a single nucleotide change of A→G. The selection of this polymorphism was based on previous studies (Reference Zhang, Xu and Wei7,Reference Dang, Zhang, Zhu and Lu11). We detected SNPs of the AUTS2 gene in Haploview v4.2 using the HapMap Japanese population and a minor allele frequency (MAF) cut-off≥5%, and selected rs9886351 using the pair-wise tagging only mode and r 2≥0.8 (Reference de Bakker, Yelensky, Pe’er, Gabriel, Daly and Altshuler13) as the cut-off for the selection of tagSNPs. Zhang et al. and Dang et al. also selected the polymorphism rs9886351 as tagSNP (Reference Zhang, Xu and Wei7,Reference Dang, Zhang, Zhu and Lu11). Regarding the association of heroin dependence and the polymorphism rs9886351, an effect of this polymorphism on the development of heroin dependence was not observed, but a statistical association trend has been found (Reference Dang, Zhang, Zhu and Lu11).
Therefore, in the present study, to elucidate genetic factors for alcohol dependence, we compared the frequencies of the polymorphisms AUTS2 rs6943555 and rs9886351 between patients with alcohol dependence and healthy control subjects from a Japanese population. In addition, we examined whether or not the haplotypes consisting of these polymorphisms in the AUTS2 gene are related to alcohol dependence.
Materials and methods
The subjects of this study consisted of 64 patients (male: 50, female: 7, not available: 7; mean age±SD: 57.34±10.18 years) with a diagnosis of alcohol dependence according to DSM-IV criteria and 75 unrelated healthy people including the alcohol drinker in everyday life (male: 23, female: 52; mean age±SD: 35.36±9.06 years). Blood samples were collected from the subjects at Koutokukai Sato Hospital Group, and all subjects provided written informed consent for genetic studies. All of the patients and control subjects lived in Yamagata prefecture in Japan.
The study was approved by the ethics committees of the Tokyo Metropolitan Institute of Medical Science [Reference de Bakker, Yelensky, Pe’er, Gabriel, Daly and Altshuler13–29] and Azabu University [0648].
Two AUTS2 gene polymorphisms were genotyped by means of polymerase chain reaction (PCR)-restriction fragment length polymorphism (RFLP) according to the method of Zhang et al. (Reference Zhang, Xu and Wei7). Genomic DNA was amplified with the following primers: rs6943555 (forward: 5'-TGG GTG TTG GAA GAG TTT TGA-3', reverse: 5'-ATA CAG TAT ACA TAA ACA TTG GAA AAG AGG GAA-3') and rs9886351 (forward: 5'-GGT GGA AAA TAA GCC AGT ATG C-3', reverse: 5'-TAG GAA AAT GGA TTA AAC GTA GGA G-3'). The PCR cycling conditions were: 95°C for 10 min, 35 cycles of 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s, with final extension at 72°C for 7 min, for rs6943555; and 95°C for 10 min, 35 cycles of 95°C for 30 s, 58°C for 30 s, and 72°C for 30 s, with final extension at 72°C for 7 min, for rs9886351. The PCR products (196 bp for rs6943555 and 221 bp for rs9886351) were digested with a restrictive enzyme, Hinf I (New England Biolabs, Tokyo, Japan), and the digested products were subjected to electrophoresis on 5% polyacrylamide gels and visualised using the ethidium bromide staining method. Genotypes were determined according to fragment sizes: rs6943555: T/T=196 bp, T/A=196 bp+165 bp+34 bp, A/A=165 bp+34 bp and rs9886351: A/A=221 bp, A/G=221 bp+198 bp+26 bp, G/G=198 bp+26 bp. The major visible bands of these polymorphisms were those of 221, 198, 196, and 165 bp (Fig. 1).
Fig. 1 Representative results for the AUTS2 gene polymorphisms with the PCR-RFLP method. (a) rs6943555: Lane 1 shows a T/T type. Lane 2 shows a T/A type. Lane 3 shows an A/A type. Lane 4 shows a 20bp DNA ladder. (b) rs9886351: Lane 1 shows an A/A type. Lane 2 shows an A/G type. Lane 3 shows a G/G type. Lane 4 shows a 20bp DNA ladder. AUTS2, autism susceptibility candidate 2; PCR-RFLP, polymerase chain reaction-restriction fragment length polymorphism.
Statistical differences in the allele and genotype frequencies of each polymorphism between the patients with alcohol dependence and healthy control subjects were assessed using the χ2 test or Yates’ correction. In addition, linkage disequilibrium (LD) coefficients (D' and r 2) and haplotype frequencies were calculated with gPLINK 2.050 (http://pngu.mgh.harvard.edu/purcell/plink/) and Haploview 4.2 (http://www.broad.mit.edu/mpg/haploview/index.php) (Reference Barrett, Fry, Maller and Daly14,Reference Purcell, Neale and Todd-Brown15). Statistical significance was defined as p<0.05.
Results
The distributions of the genotype and allele frequencies of the polymorphisms AUTS2 rs6943555 and rs9886351 in both patients with alcohol dependence and control subjects are shown in Tables 1 and 2. The genotype distributions of these polymorphisms in both the alcohol dependence and control subjects were in Hardy–Weinberg equilibrium when tested using the χ2 test (data not shown).
Table 1 Distributions of the genotype and allele frequencies of the polymorphisms AUTS2 rs6943555 in alcohol dependence and control subjects
AUTS2, autism susceptibility candidate 2.
Figures in parentheses are percentages.
Table 2 Distributions of the genotype and allele frequencies of the polymorphisms AUTS2 rs9886351 in alcohol dependence and control subjects
AUTS2, autism susceptibility candidate 2.
Figures in parentheses are percentages.
For the polymorphism AUTS2 rs6943555, there were no significant differences in the genotype or allele frequencies between alcohol dependence and control subjects (total: genotype: χ2(2)=1.87, p=0.392, allele: χ2(1)=1.17, p=0.280; male: genotype: χ2(2)=0.154, p=0.926, allele: χ2(1)=0.396, p=0.529; female: genotype: χ2(2)=3.24, p=0.197, allele: χ2(1)=0.365, p=0.546). For the polymorphism AUTS2 rs9886351, there were also no significant differences in the genotype or allele frequencies between alcohol dependence and control subjects (total: genotype: χ2(2)=1.69, p=0.430, allele: χ2(1)=1.70, p=0.193; male: genotype: χ2(2)=0.0783, p=0.962, allele: χ2(1)=0, p=0.996; female: genotype: χ2(2)=2.24, p=0.326, allele: χ2(1)=3.43, p=0.0641).
The frequencies of haplotypes consisting of the polymorphisms AUTS2 rs6943555 and rs9886351 in both the patients with alcohol dependence and control subjects are shown in Table 3. The distributions of the A-A haplotype combinations were statistically significantly different between alcohol dependence and control subjects (χ2(1)=5.53, p=0.0187). The patients with alcohol dependence showed a significantly higher frequency of the A-A haplotype (rs6943555 and rs9886351) as compared with the control group (26.73% of patients, 15.03% of controls). The frequencies of other haplotypes did not show significant differences between alcohol dependence and control subjects.
Table 3 Haplotype frequencies for the polymorphisms AUTS2 rs6943555 and rs9886351 in alcohol dependence and control subjects
AUTS2, autism susceptibility candidate 2.
The distributions of the A-A haplotype showed a statistically significant difference between alcohol dependence and control subjects.
*p<0.05.
In addition, the pairwise D' and r 2 values for the polymorphisms AUTS2 rs6943555 and rs9886351 in this study were 0.033 and 0.001, respectively. These polymorphisms showed weak LD as to each other.
Discussion
In this study, we investigated the association of AUTS2 gene polymorphisms and alcohol dependence in a Japanese population. The genotype and allele frequencies of the AUTS2 gene polymorphisms observed in our healthy subjects were consistent with the genotype (rs6943555: χ2(2)=0.0494, p=0.976; rs9886351: χ2(2)=0.352, p=0.839) and allele (rs6943555: χ2(1)=0.248, p=0.618; rs9886351: χ2(1)=0.0928, p=0.761) frequencies observed in the HapMap (http://hapmap.ncbi.nlm.nih.gov/index.html.ja) Japanese population, respectively.
Our results showed that there were no significant differences in the genotype and allele frequencies of the polymorphisms AUTS2 rs6943555 and rs9886351 between the patients with alcohol dependence and healthy control subjects. Recent studies have shown that the AUTS2 gene polymorphisms, especially rs6943555, are broadly associated with schizophrenia (Reference Zhang, Xu and Wei7), suicide (Reference Chojnicka, Gajos and Strawa8), and heroin dependence (Reference Chen, Liao, Lai and Chen10,Reference Dang, Zhang, Zhu and Lu11). All of these case-control reports stated that A/A homozygotes and/or the minor A allele of this polymorphism is a risk factor in such cases. Regarding the association of heroin dependence, Chen et al. measured the cDNA level of AUTS2 in lymphoblastoid cell lines with heroin-dependent patients and control subjects, and they discovered that it was significantly reduced in the patient group compared with the control group (Reference Chen, Liao, Lai and Chen10). In addition, they found that the rs6943555 A/A genotype lowered the AUTS2 mRNA level in lymphoblastoid cell lines compared with the T/T and T/A genotypes (Reference Chen, Liao, Lai and Chen10). For these findings, reduced AUTS2 gene expression is likely to be a risk factor for the development of heroin dependence (Reference Chen, Liao, Lai and Chen10,Reference Dang, Zhang, Zhu and Lu11). On the other hand, Schumann et al. revealed that down-regulation of an AUTS2 homolog induced reduced alcohol sensitivity in Drosophila (Reference Schumann, Coin and Lourdusamy12). The low level of response (reduced sensitivity) to alcohol contribute to an increased risk for alcohol use disorders (Reference Chen, Liao, Lai and Chen10–Reference Schumann, Coin and Lourdusamy12,Reference Schuckit, Smith and Kalmijn16). Therefore, there is a possibility that reduced AUTS2 gene expression induces an increasing risk of developing alcohol drinking disorders in humans (Reference Chen, Liao, Lai and Chen10,Reference Dang, Zhang, Zhu and Lu11). In this study, although no relationship was found between the polymorphism AUTS2 rs6943555 alone and alcohol dependence, the haplotype A-A including the minor A allele of the rs6943555 is one of the factors that reduce the AUTS2 gene expression in lymphoblastoid cell lines associated with the disease. Thus, the polymorphism rs6943555 might have affected more than a little to alcohol dependence, though not to the extent of heroin dependence.
Meanwhile, Schumann et al. reported that the rs6943555 A allele increases AUTS2 gene expression in the prefrontal cortex of the human brain compared with the T allele (Reference Schumann, Coin and Lourdusamy12). Furthermore, in genome-wide association study in Europeans, they also have revealed that the minor ancestral A allele of the rs6943555 was associated with 5.5% lower alcohol consumption (Reference Schumann, Coin and Lourdusamy12). For these observations, we expected that the A allele of this polymorphism would be associated with protective effects against alcohol dependence, although no association was found between them in a Japanese population. In molecular biological research on genetic factors influencing alcohol drinking disorders, including alcohol dependence, when consistent results can not be observed between the various countries and populations, the differences in genetic factors as well as in environmental factors such as alcohol availability, people’s thinking about alcohol, and social system constructed by law that affects alcohol drinking behavior, will need to be considered (Reference Nishizawa, Han and Hasegawa1). Incidentally, there was a tendency towards a slightly lower rs6943555 minor A allele frequency (24%) of Europian subjects in Schumann’s study (Reference Schumann, Coin and Lourdusamy12) compared with those of Japanese subjects in our study (case: 37.5%, control: 31.3%). Thus, the genotype and allele frequencies of the polymorphism AUTS2 rs6943555 might differ between races (Reference Chen, Liao, Lai and Chen10). The A/A genotype (9.3%) and A allele frequencies (31.3%) in our healthy subjects showed not much difference from that in Han Chinese subjects (A/A genotype: 8.31–13.0%; A allele: 29.3–35.8%) (Reference Zhang, Xu and Wei7,Reference Chen, Liao, Lai and Chen10,Reference Dang, Zhang, Zhu and Lu11). On the other hand, in Polish Caucasian subjects the A/A genotype accounts for only 4.7% and the A allele 21% (Reference Chojnicka, Gajos and Strawa8). The genotype and allele frequencies of the polymorphism AUTS2 rs6943555 in Polish Caucasian subjects show a significant difference from that in the Japanese determined in this study (genotype: χ2(2)=7.69, p=0.0213; allele: χ2(1)=9.34, p=0.00224) (Reference Chojnicka, Gajos and Strawa8). Therefore, if different results from those in our study on additional exploration of the polymorphism AUTS2 rs6943555 and alcohol dependence are obtained, such differences in genotype and allele frequencies among races might be one of the reasons. However, because the number of healthy subjects in our study was smaller than that in Chojnicka’s study (Reference Chojnicka, Gajos and Strawa8) of Polish Caucasians, it will be necessary to take this point into consideration.
For the polymorphism rs9886351, several studies have concerned the association with schizophrenia and heroin dependence, but they failed to reveal significant associations with these diseases (Reference Zhang, Xu and Wei7,Reference Dang, Zhang, Zhu and Lu11). Furthermore, no association was found between the polymorphism rs9886351 and alcohol dependence in this study. Although it is currently unknown whether the rs9886351 affects the function of AUTS2, it is unlikely that this polymorphism alone affects the development of alcohol dependence.
On the other hand, the frequencies of the AUTS2 haplotypes were significantly different between the patients with alcohol dependence and healthy control subjects, and the rs6943555 and rs9886351 A-A haplotype was associated with alcohol dependence. In this study, we observed that two SNPs in all subjects did not exhibit strong LD. In an association study of AUTS2 gene polymorphisms and schizophrenia, although Zhang et al. also revealed that the LD values between the polymorphisms rs6943555 and rs9886351 was low (D’=0.06 and r 2=0), they suggested that calculation of the haplotype frequencies of these SNPs is not required (Reference Zhang, Xu and Wei7). Additionally, Dang et al. investigated the correlation of heroin dependence and 21 SNPs in the AUTS2 gene, and they compared the haplotype frequencies in heroin dependence and healthy control subjects focusing on five haplotype blocks (strong LD) (Reference Dang, Zhang, Zhu and Lu11). According to that report, the haplotypes consisting of the polymorphisms rs6943555 and rs10251416 were significantly associated with heroin dependence (Reference Dang, Zhang, Zhu and Lu11). Conversely, in the presence of multiple susceptibility alleles, analyses based on haplotypes have advantages over analysis based on one SNP, particularly when the degree of LD between the SNPs is weak (Reference Morris and Kaplan17,Reference You, Yuan, Zhang, Zhang, Li and Qian18). Morris and Kaplan have suggested that the power advantage of haplotype analysis can be minimal or lost when there is strong correlation among SNPs (Reference Morris and Kaplan17). Furthermore, they also concluded that statistical methods based on haplotypes when the LD between SNPs is weak may be promising for identifying and locating disease genes in the presence of multiple susceptibility alleles with general features of complex disease genes (Reference Morris and Kaplan17). Haplotype analysis is thought to be meaningful not only when the LD between SNPs is strong but also when it is weak, and important factors that were not found on analysis to have strong correlation among SNPs might be detected with weak correlation among SNPs. Although the present study was exploratory research, the new finding that the A-A haplotype consisting of the polymorphisms rs6943555 and rs9886351 is significantly associated with alcohol dependence, may be one of the important findings in molecular biological studies for elucidation of genetic factors for alcohol dependence.
When performing analyses to examine the association between alcohol dependence and the AUTS2 gene polymorphisms, we did not perform correction for multiple comparisons, such as Bonferroni correction. Several studies have indicated that Bonferroni adjustment would be too conservative for genetic association studies (Reference Nyholt19), and the likelihood of type II errors is increased by such adjustment, meaning that truly important differences may be deemed nonsignificant (Reference Perneger20). Furthermore, when the sample size of a study is also small, there is a low possibility of committing a type II error (Reference Nishizawa, Han and Hasegawa1). For the reasons stated above, for this exploratory study, we present results with no correction.
It must be considered that for the subjects in this study, there were several limitations and bias. First, the sample sizes of alcohol-dependent patients and control subjects in our study were small and consequently have low statistical power. Second, the alcohol-dependent subjects (n=64) were mostly males (n=50) and, contrary to this, in the healthy control subjects, the number of females (n=52) was more than twice than that of males (n=23). Third, there was a significant difference in the average ages (p<0.001) of alcohol-dependent patients (57.34±10.18) and healthy control subjects (35.36±9.06). However, all of the subjects in this study were from a particular provincial population (i.e. Yamagata prefecture) in Japan. Because alcohol drinking behavior differs among populations, genetic factors for alcohol dependence can be hidden when various populations are mixed in a study (Reference Nishizawa, Han and Hasegawa1). Furthermore, there is also a possibility that different results as to the frequency of a minor allele are obtained for different regions of the same country (Reference Nishizawa, Han and Hasegawa1). Larger scale comparison of subjects of the same generation with no gender bias is needed in a further study to confirm our current results. Furthermore, because we could not obtain detailed clinical data on the alcohol-dependent patients, it will also be worth examining the association between the clinical characteristics of alcohol dependence such as tendency to violence and the age of onset.
In conclusion, our results suggested that the frequencies of the AUTS2 haplotypes were significantly different between patients with alcohol dependence and healthy control subjects, and the A-A haplotype consisting of the polymorphisms rs6943555 and rs9886351 might be a risk factor for alcohol dependence in a Japanese population. However, the results of an exploratory study should be interpreted cautiously. If genes that are involved in the development of alcohol dependence are revealed in future studies, as a tailor-made medical procedure, early prevention and treatment based on features of the genetic background of individuals determined through genetic polymorphisms may be possible.
Acknowledgements
This research was supported by a research project grant awarded by the Azabu University. Authors’ Contributions: K. Iwahashi was responsible for the study design. K. Ikeda, D. Nishizawa, and A. Kawai contributed to the acquisition of the clinical samples. S. Narita performed the genotyping and statistical analysis. S. Narita drafted the manuscript. K. Nagahori, E. Yoshihara, D. Nishizawa, K. Ikeda, and K. Iwahashi performed critical revision of the manuscript for important intellectual content. All the authors contributed and critically reviewed the contents, and have approved the final version of the manuscript.
Financial Support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Conflicts of Interest
None.
Ethical Standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
