Association of alcohol dehydrogenase and aldehyde dehydrogenase Polymorphism with Spontaneous Deep Intracerebral Haemorrhage in the Taiwan population

Alcohol dehydrogenase (ADH) and aldehyde dehydrogenase (ALDH) encode essential alcohol-metabolizing enzymes. While alcohol use is associated with spontaneously deep intracerebral haemorrhage (SDICH), particularly in males, the activities and genetic variants of ADH and ALDH may affect SDICH development. This case-control study was conducted to identify the interaction of alcohol use and SDICH with five single-nucleotide polymorphisms (SNPs): ADH1B rs1229984, ADH1C rs2241894, ALDH2 rs671, ALDH2 rs886205, and ALDH2 rs4648328. We enrolled 208 patients with SDICH and 244 healthy controls in a Taiwanese population. ALDH2 rs671 was significantly associated with SDICH in the dominant (P < 0.001) and additive models (P = 0.007). ALDH2 rs4648328 was borderline significantly associated with SDICH in the recessive (P = 0.024) or additive models (P = 0.030). In alcohol-using patients, the ALDH2 rs671 GG genotype was associated with SDICH risk compared to the GA+AA genotype (P = 0.010). ADH1B rs1229984, ADH1C rs2241894, and ALDH2 rs886205 did not demonstrate association with SDICH. Thus, the ALDH2 rs671 GG genotype is a risk factor for SDICH. Because the genetic distributions of ALDH2 rs671 exhibited strong ethnic heterogeneity, further studies in different populations are needed to validate these findings.

Genotype frequency and association analysis of controls and patients with SDICH. All single-nucleotide polymorphisms (SNPs) were in Hardy-Weinberg equilibrium in the case and control groups according to the standard χ 2 test at a significance level of 0.05. The genotype frequencies of the analysed SNPs in the case and control groups are shown in Table 2. ALDH2 rs671 was significantly associated with SDICH in the dominant model (OR = 0.5, 95% CI: 0.4-0.8, P < 0.001) and additive model (OR = 0.7, 95% CI: 0.5-0.9, P = 0.007). The significance remained after adjusting for sex and age in the dominant model (OR = 0.6, 95% CI: 0.4-0.8, P = 0.003) and borderline in the additive model (OR = 1.5, 95% CI: 1.1-2.0, P = 0.015). However, these associations did not remain after further adjusting for hypertension and alcohol use. ALDH2 rs4648328 could be associated with SDICH in the recessive model (OR = 2.4, 95% CI: 1.1-5.1, P = 0.024) and additive model (OR = 1.4, 95% CI: 1.0-1.9, P = 0.030) with boardline significance. These associations were not detected after Bonferroni correction and multivariate adjustment. The genotypic frequencies of other genetic variants were similar between the SDICH and controls.
The minor allele frequencies (MAFs) of the analysed SNPs in the case and control groups are shown in Table 3. The MAF of ALDH2 rs671 (21.9%) in the SDICH group was significantly lower compared to controls (30.1%, odds ratio (OR) = 0.7, 95% confidence interval (CI): 0.5-0.9, P = 0.005). The MAFs of the other SNPs were similar between the SDICH and control groups.
We further stratified the allelic and genotypic frequencies of ADH1B rs1229984, ADH1C rs2241894, ALDH2 rs671, ALDH2 rs886205, and ALDH2 rs4648328 according to alcohol use. When stratified by alcohol use, the ALDH2 rs671 GA genotype was significantly associated with SDICH in the alcohol use group (OR = 0.2, 95% CI: 0.1-0.7, P = 0.008), indicating the interaction between the ALDH2 rs671 genotype and alcohol use. Specifically, in alcohol-free subjects, the SDICH risk was similar between genotypes. In subjects with alcohol use, SDICH was more frequently observed in individuals carrying ALDH2 rs671 GG genotype compared to rs671 GA+AA genotype (SDICH percentage: GG vs GA+AA: 70.6% vs 38.9%, OR = 0.3, 95% CI 0.1-0.8, crude P = 0.01, Fig. 1), whereas this difference was not observed after multivariable adjustment. There was no association between all tested SNPs and SDICH by stratification according to the presences of hypertension and gender (data not shown). None of the alleles and genotypes in this study showed associations with hypertension (Supplementary Table S1).

Discussion
This study, at the first time, describes the potential association between SNPs of ADH and ALDH2 with SDICH susceptibility in the Taiwanese population. Asian populations have higher incidences of SDICH with high mortality and long-term disability than Caucasians 2,3,19 . Alcohol use demonstrates the association with SDICH 7 . Alcohol is primarily metabolized by ADH and ALDH 8 . Our results support a role for ALDH2 genetic variants in SDICH. We found that the ALDH2 rs671 GG genotype could be a risk locus for SDICH, particularly in subjects who used alcohol, in Taiwanese population. Haplotype analysis further identified the association between haplotypes in rs671-rs4648328 of ALDH2 and SDICH. Further large case-control cohorts in multi-ethnicities are needed to validate this association.
The rs671 is a functional SNP (Glu504Lys) in ALDH2 20 . Minor allele (A allele) of rs671 results in reduced ALDH2 enzymatic activity. Approximately 30% of people in Asia and 47% of those in Taiwan carrying the rs671 A allele [21][22][23][24] . In the male Japanese population, the ALDH2 rs671 GG genotype is associated with cerebral lacunar infarcts 17 . ALDH2 rs671 A allele are associated with coronary artery disease in Chinese patients with hypertension 16 . Moreover, ALDH2 rs671 A allele is also associated with hypertension and cerebral amyloid angiopathy 18,25 .
Although ALDH rs671 AA genotype may be associated with alcoholism-related hypertension 26 , our results did not detect the association between ALDH2 rs671 and hypertension, supporting the primary effect of ALDH2 rs671 on SDICH. ALDH2 rs671 GG genotype tends to be a risk factor for SDICH, particularly in the group with alcohol use in Taiwanese. www.nature.com/scientificreports www.nature.com/scientificreports/ ALDH2 rs4648328, an intronic SNP, was associated with delayed alcohol metabolism in European population 27 . In our analysis, we found a potential association between rs4648328 and SDICH in the recessive and additive models. SNPs in ALDH2 demonstrated strong LD in Indian population 27 . In addition, our study showed that rs4648328 was in LD with rs671 in Taiwanese population. The haplotype "GT" of ALDH2 rs671-rs4648328 was associated with SDICH, whereas the haplotype "AC" demonstrated protective effect on SDICH. This study provides a baseline for future research about the role of the ALDH2 loci in SDICH in Taiwanese population. Further large-scale investigations are needed to confirm this result. Table 5 shows ethnicity differences in SNPs of ADH and ALDH. The genetic distributions of ALDH2 rs671 showed strong ethnic heterogeneity. The frequencies of A allele in Taiwanese (26.3%) and East Asians (17.4%) are much higher compared to Americans (0%), Europeans (0%) and global populations (3,6%). Previous studies showed that genetic variants of ALDH2 rs671 were associated with both alcohol flushing and alcohol use in Asian populations 28,29 . Additionally, the ALDH2 rs671 GG genotype is associated with cerebral lacunar infarcts in the male Japanese 17 . Our study showed that rs671 GG genotype was associated with SDICH susceptibility, particularly in the alcohol use group.
In addition to rs671, the MAFs of rs886205, rs1229984, and rs2241894 also greatly differ between Asian and Caucasians 30 . Table 5 showed the ethnic heterogeneous of the rs671, rs886205 rs1229984, and rs2241894 according to 1000 genome information.
The MAF T allele was present in 15.9% of rs1229984 in global population, while the rs1229984 C allele was present in 26.2% of Taiwanese and 30.3% of east Asian. (Table 5). While the ADH1B 1229984 CC genotype is predominant in East Asian population, it is rarely observed in Indian population 31 . The role of ADH1B rs1229984 in modulating alcohol consumption remains controversial. It has been reported that ADH1B rs1229984 C allele is associated with alcoholism31. However, a case-control study suggests that CC genotype of ADH1B rs1229984 may protect against alcohol dependence 32 . In our analysis, the ADH1B rs1229984 did not demonstrate association with alcohol consumption.   www.nature.com/scientificreports www.nature.com/scientificreports/ The MAF C allele was present in 47.2% of rs2241894 in global population, while the rs2241894 T allele was present in 28.2% of Taiwanese. A genome-wide association study also demonstrated an association between ADH1C rs2241894 and alcohol dependence in African and European Americans 14 . The MAF A allele was present in 49.1% of rs886205 in global population, while the rs886205 A allele was present in 14.1% of Taiwanese. A recent study reported that ALDH2 rs886205 is associated with alcohol-dependent patients 33 . However, we found no associations between ALDH2 rs886205, ADH1C rs2241894, alcohol use and SDICH in our analysis. This discrepancy may be contributed by the ethnic difference of genetic background, as well as the design of studies.
To our knowledge, this is the first study to propose that the ALDH2 rs671 GG genotype is a risk factor for SDICH, particularly in an alcohol-using population. ALDH2 rs671 and rs4648328 are particularly important in the interaction with alcohol use, one of the major environmental risk factors for SDICH. There are limitations to our study. First, this was a hospital-based study which may limit the generalization of our results to the whole population. Most  www.nature.com/scientificreports www.nature.com/scientificreports/ of patients with SDICH were recruited from the Department of Neurology; these patients may demonstrate smaller haemorrhages compared to those admitted to the Department of Neurosurgery. Additionally, the relatively small sample size and gender imbalance may limit detection of potential genetic associations with SDICH. However, our results support the potential association of genetic variants in ALDH2 rs671 GG genotype with SDICH risk in a Taiwanese population. Further studies in different populations are needed to validate our results.
conclusion. This study revealed a significant association between the genetic variants of ALDH2 and SDICH susceptibility. Carrying the ALDH2 rs671 GG genotype tends to be a risk factor for SDICH, particularly in those who use alcohol.

Materials and Methods
patients and control subjects. Patients (age > 30 years old) with SDICH at the basal ganglia, thalamus, cerebellum, or brainstem were included in the study 4 . The size and location of SDICH were confirmed by brain computed tomography (CT). Patients with traumatic cerebral haemorrhage, haemorrhagic transformation of a cerebral infarct, vascular anomaly, and secondary intracranial haemorrhage (coagulopathy or hyper-perfusion syndrome) were excluded. Controls were defined as those without medical disease such as renal failure, myocardial infarction, cancer, stroke history, and neurodegenerative disease. A history of hypertension, diabetes mellitus, smoking, alcohol use, and lipid profile were collected from all participants. Alcohol use referred to the consumption of greater than 210 g of alcohol per week 34 . Smokers were defined as former or current smokers 35 .
This retrospective case-control study was approved by the Chang Gung Memorial Hospital Institution Ethics Review Board for human studies, and patients provided written informed consent prior to study participation (IRB201600775B0). All methods were performed in accordance with the relevant guidelines and regulations.
Blood samples were collected for genotyping. The genomic DNA was extracted from peripheral leukocytes by using the Stratagene DNA extraction kit (La Jolla, CA, USA). Polymorphisms were genotyped using TaqMan SNP Assays in the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA, USA). The primer sets used for polymerase chain reaction amplification of each SNP region are as listed in Supplementary  Table S2. Each SNP was checked for Hardy-Weinberg equilibrium using the standard χ 2 test at a significance level of 0.05. Patterns of LD and haplotype analyses were evaluated using SHEsis Online Version (http://analysis.bio-x. cn/myAnalysis.php) 36 . Haplotypes with frequency <3% were excluded from association analysis.
Statistical analysis and power estimation. All data analyses were performed using SAS Software (version 9.4; SAS Institute, Cary, NC, USA). Demographic data and the distributions of genotypes of SNPs were analysed by χ 2 test, t-test, and univariate logistic regression. Multivariable logistic regression analyses were used to test the null hypothesis that the number of cases and controls did not differ with various genotypes of the five SNPs. Potential covariables included age, sex, hypertension, total cholesterol level, and alcohol use. Samples were stratified by alcohol use using multivariable logistic regression. All P values were two-tailed. While considering Bonferroni correction, the significance level was set to 0.01. Given the observed allele frequency in the present case-control study, at the 0.01 significance level, we had power greater than 0.8 to identify an association of the genetic variant with SDICH susceptibility when the per-allele genetic effect was greater than an odds ratio of 1.8 for rs886205 and 1.7 for the rest of the SNPs.  Table 5. Minor allele frequency (MAF) in different populations. SNP: Single-nucleotide polymorphism; MAF: minor allele frequency. a MAF data from 1000 genome information.