SIRT1 gene, schizophrenia and bipolar disorder in the Japanese population: an association study

Authors


T. Kishi, Department of Psychiatry, Fujita Health University School of Medicine, Toyoake, Aichi 470-1192, Japan. E-mail: tarok@fujita-hu.ac.jp

Abstract

Several lines of evidence suggest that alterations in circadian rhythms might be associated with the pathophysiology of psychiatric disorders such as schizophrenia and bipolar disorder (BP). A recent study reported that SIRT1 is a molecule that plays an important role in the circadian clock system. Therefore, to evaluate the association among the SIRT1 gene, schizophrenia and BP, we conducted a case–control study of Japanese population samples (1158 schizophrenia patients, 1008 BP patients and 2127 controls) with four tagging SNPs (rs12778366, rs2273773, rs4746720 and rs10997875) in the SIRT1 gene. Marker-trait association analysis was used to evaluate the allele and the genotype association with the χ2 test, and haplotype association analysis was evaluated with a likelihood ratio test. We showed an association between rs4746720 in the SIRT1 gene and schizophrenia in the allele and the genotype analysis. However, the significance of these associations did not survive after Bonferroni's correction for multiple testing. On the other hand, the SIRT1 gene was associated with Japanese schizophrenia in a haplotype-wise analysis (global Pall markers = 4.89 × 10−15). Also, four tagging SNPs in the SIRT1 gene were not associated with BP. In conclusion, the SIRT1 gene may play an important role in the pathophysiology of schizophrenia in the Japanese population.

Abnormalities in the circadian rhythm are hypothesized to be involved in the pathophysiology of psychiatric disorders such as schizophrenia and bipolar disorder (BP), since many psychiatric disorder patients have insomnia and sleep-wake disturbance (Barbini et al. 1998; Harmer 2008; Kato 2007; Le Francois et al. 2008; Levinson 2006; McClung 2007a,b,c). Moreover, some genetic studies have shown significant associations between schizophrenia/schizoaffective disorder and the timeless homolog gene (TIMELESS) (Mansour et al. 2006), neuronal PAS domain protein 2 gene (NPAS2) (Mansour et al. 2009) and period homolog 3 gene (PER3) (Mansour et al. 2006); between BP and the clock gene (CLOCK) (Soria et al. 2010), NPAS2 (Mansour et al. 2009), early growth response 3 (EGR3) (Mansour et al. 2009), RAR-related orphan receptor alpha gene (RORB) (Mansour et al. 2009), vasoactive intestinal peptide gene (VIP) (Soria et al. 2010), rev-erb alpha gene (NR1D1) (Severino et al. 2009), prokineticin 2 receptor gene (PROKR2) (Kishi et al. 2009), Bmal1 gene (ARNTL) (Nievergelt et al. 2006), TIMELESS (Mansour et al. 2006) and PER3 (Mansour et al. 2006; Nievergelt et al. 2006). These findings suggest a crucial relationship between the circadian rhythms and psychiatric disorders. Genes associated with the molecular clock mechanism are therefore good candidates for the etiology of psychiatric disorders. In addition, NPAS2 was shown to be associated with schizophrenia and BP (Mansour et al. 2009). These results suggest that a partial overlap in genetic predisposition between schizophrenia and BP is possible (Ivleva et al. 2009). The evidence for this is discussed in more detail in four reviews (Ivleva et al. 2009; Moskvina et al. 2009; O’donovan et al. 2009; Purcell et al. 2009). Schizophrenia and BP have approximately 80% heritability (Ivleva et al. 2009). Recent whole genome studies have showed a number of susceptibility regions that overlap in schizophrenia and BP (1q32, 10p11–15, 13q32, 18p11.2 and 22q11–13) (Ivleva et al. 2009). Schizoaffective disorder is known to be a disorder to have both characteristics of schizophrenia and BP (Ivleva et al. 2009). The following evidence supports this proof. Recent whole genome association studies (GWAS) reported that Zinc finger binding protein 804A (ZNF804A) and calcium channel, voltage-dependent, L-type, alpha 1C (CACNA1C) subunits were associated with schizophrenia and BP (Consortium 2007; Green et al. 2010; Moskvina et al. 2009; O’donovan et al. 2008; Purcell et al. 2009). We thought that schizophrenia and BP may have some shared mechanisms with respect to circadian rhythms and considered it reasonable to assess these disorders.

Recently, SIRT1 was reported to be an important molecule in the circadian rhythm mechanisms. Clock, one of the important clock molecules, makes up a complex with Bmal1 and plays an important role in the feedback loop for circadian rhythms (Asher et al. 2008; Belden & Dunlap 2008; Nakahata et al. 2008, 2009; Ramsey et al. 2009; Wijnen 2009). Some recent studies indicated that SIRT1 completes the Clock–Bmal1-dependent transcription, forming another complementary negative feedback loop with Clock–Bmal1 and nicotinamide phosphoribosyltransferase (Nampt, and influences the circadian transcription of several core CLOCK genes, including ARNTL, period homolog 2 gene (Per2) and cryptochrome 1 gene (Cry1) (Asher et al. 2008; Belden & Dunlap 2008; Nakahata et al. 2008, 2009; Ramsey et al. 2009; Wijnen 2009). From this, one may predict that SIRT1 has some role in the pathophysiology of psychiatric disorders, similar to other CLOCK genes. On the other hand, Lee et al. (2009) reported that nicotinamide, a well-known inhibitor of SIRT1 used for in vivo studies, produced similar phenotypes in mice (small size, weight loss and decreased motor activity) to those of dopamine-deficient and SIRT1 gene knockout mice, suggesting the influence of SIRT1 on the dopaminergic system.

Abnormalities in the dopamine neural transmission are known to be involved in the pathophysiology of schizophrenia (Lang et al. 2007) and BP (Berk et al. 2007). Based on the above, we considered the SIRT1 gene to be a good candidate gene for involvement in the pathogenesis of schizophrenia and BP.

The SIRT1 gene (OMIM * 604479, 11 exons in this genomic region spanning 34.521 bp) is located on 10q21.3. There are no reported gene-based association analyses among SIRT1 gene, schizophrenia and BP in the Japanese population. Recently, a GWAS using Japanese BP samples was reported (Hattori et al. 2009), but this GWAS did not include variants which are seen around the SIRT1 gene, including the tagging SNPs in our study. Recently, several GWAS (such as schizophrenia and BP) have been performed using European Caucasians samples (Consortium 2007; Green et al. 2010; Moskvina et al. 2009; Ng et al. 2009; O’donovan et al. 2008; Purcell et al. 2009). However, no association with the SIRT1 gene was found (Consortium 2007; Green et al. 2010; Moskvina et al. 2009; Ng et al. 2009; O’donovan et al. 2008; Purcell et al. 2009). We therefore conducted a moderate size case–control study with Japanese schizophrenia and BP samples.

Materials and methods

Subjects

The subjects in the association analysis were 1158 schizophrenia patients [605 males and 553 females; mean age ± standard deviation (SD) 42.6 ± 17.7 years], 1008 BP patients (495 males and 513 females; 730 patients with bipolar I disorder and 278 patients with bipolar II disorder; mean age ± SD 47.1 ± 14.2 years) and 2127 healthy controls (1026 males and 1101 females; 45.6 ± 15.2 years). All subjects were unrelated to each other, ethnically Japanese and lived in the central area of Japan. A total of 1016 schizophrenia patients and 997 BP patients were diagnosed according to DSM-IV criteria with the consensus of at least two experienced psychiatrists on the basis of unstructured interviews and a review of medical records. Of them, 142 schizophrenic patients and 11 BP patients underwent the Structured Clinical Interview for DSM-IV disorders (SCID-1). Schizophrenic patients were grouped according to the following DSM-IV subtypes of schizophrenia: paranoid type (n = 407), disorganized type (n = 232), catatonic type (n = 49), residual type (n = 385) and undifferentiated type (n = 85). A total of 1950 controls were also diagnosed according to DSM-IV criteria with the consensus of at least two experienced psychiatrists on the basis of unstructured interviews, including 46 and 131 who underwent the Mini-International Neuropsychiatric Interview (MINI) and SCID-1, respectively. None had severe medical complications such as liver cirrhosis, renal failure, heart failure or other Axis-I disorders according to DSM-IV. This sample panel was the same as used in the Collaborative Study of Mood Disorder consortium study. Seven laboratories (National Institute of Neuroscience, two laboratories of RIKEN Brain Science Institute, Kohnodai Hospital, Teikyo University, Okayama University and Fujita Health University) provided case and control samples. The study was described to the subjects and written informed consent was obtained from each. This study was approved by the ethics committees of all participating institutes.

SNP selection and linkage disequilibrium evaluation

We first consulted the HapMap database (release#3, 10 May, www.hapmap.org, population: Japanese Tokyo, minor allele frequencies (MAFs) of more than 0.05) and included 21 SNPs covering the SIRT1 gene [5′-flanking regions including about 1.5 kb from the initial exon and about 3 kb bp downstream (3′) from the last exon: HapMap database contig number chr 10q21.3: 69313085 … 69349830]. Four ‘tagging SNPs' in the SIRT1 gene were then selected with the criteria of an r2 threshold greater than 0.8 in ‘pair-wise tagging only’ mode using the ‘Tagger’ program (Paul de Bakker, http://www/broad.mit.edu/mpg/tagger) in Haploview for the following association analysis (Barrett et al. 2005) (Fig. 1).

Figure 1.

LD evaluation from tagging SNPs in the SIRT1 gene. Black bars represent exons of the SIRT1 gene. Tagging SNPs selected from HapMap database are represented by black boxes. The colour scheme is based on D′ value. LDs of schizophrenia, BP and controls are almost the same. Other information can be seen at the haploview website.

SNP genotyping

Four tagging SNPs (rs12778366, rs2273773, rs4746720 and rs10997875) were genotyped using the TaqMan allelic discrimination assay and the ABI PRISM 7900 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). TaqMan SNP genotyping assays C_1340370_10, C_16179813_10, C_32338526_10 and C_1340400_10 were used for the four SNPs (rs12778366, rs2273773, rs4746720 and rs10997875), respectively (http://www.appliedbiosystems.com/absite/us/en/home.html). One allelic probe was labeled with FAM dye and the other with fluorescent VIC dye. The plates were heated for 2 min at 50 and 95°C for 10 min, followed by 45 cycles of 95°C for 15 seconds and 58°C for 1 min. Please refer to ABI for the primer sequence (http://www.appliedbiosystems.com/absite/us/en/home.html).

Statistical analysis

Genotype deviation from the Hardy–Weinberg equilibrium (HWE) was evaluated by χ2 test (SAS/Genetics, release 8.2, SAS Japan Inc, Tokyo, Japan). Marker-trait association analysis was used to evaluate allele and genotype association with the χ2 test (SAS/Genetics, release 8.2, SAS Japan Inc) and haplotype association analysis was done with a likelihood ratio test using the COCAPHASE2.403 program (Dudbridge 2003). In the haplotype analysis, we determined that the cutoff for testing haplotype frequency was 0.01. We used the permutation test option as provided in the haplotype analysis to avoid spurious results and correct for multiple testing. Permutation test correction was performed using 10 000 iterations (random permutations). In addition, Bonferroni's correction was used to control inflation of the type I error rate in the single marker association analysis. For Bonferroni's correction, we employed eight multiple tests for each sample set in the allele and genotype analysis (four tagging SNPs in SIRT1 gene and two disorders). For individual haplotype association analysis, we employed five multiple tests for rs12778366–rs2273773–rs4746720–rs10997875 (maximum number of haplotypes) for each sample set in the haplotype analysis. Power calculation was performed using a genetic power calculator (Purcell et al. 2003). We set each item in each value in the genetic power calculator as follows: prevalence 0.01 in schizophrenia and BP, user-defined 0.0125 (four SNPs examined in this study. Bonferroni's correction was used to control inflation of the type I error rate). When we calculated the statistical power using the genetic power calculator, we substituted the MAFs of case and controls and number of case and controls. The significance level for the statistical tests was 0.05.

Results

The linkage disequilibrium (LD) structure as determined from our schizophrenia, BP and control sample can be seen in Fig. 1. The LD of our control samples is similar to that of the HapMap database. Genotype frequencies of all SNPs were in the HWE (Table 1). We showed an association between rs4746720 in the SIRT1 gene and schizophrenia in the allele and genotype analysis (Table 1). However, the significance of these associations disappeared after Bonferroni's correction (Table 1). We found the SIRT1 gene to be associated with the Japanese schizophrenia in an all markers haplotype-wise analysis (global Pall markers = 4.89 × 10−15). We showed individual haplotype analysis in Table 2. Haplotype analysis to investigate both schizophrenia and controls indicated several common haplotypes as follows: rs12778366–rs2273773–rs4746720–rs10997875, T-T-T-T, T-T-T-C, T-T-C-T, T-C-T-T and C-T-T-T (Table 2). The T-T-T-T haplotype was less prevalent in controls compared with schizophrenia (PT-T-T-T = 1.63 × 10−17), while T-T-C-T was very prevalent in controls compared with schizophrenia (PT-T-C-T = 0.00421) (Table 2). Moreover, the association of these individual haplotypes with schizophrenia remained after Bonferroni's correction (corrected PT-T-T-T = 8.15 × 10−17 and corrected PT-T-C-T = 0.0211) (Table 2). We did not detect any associations between the tagging SNPs in the SIRT1 gene and BP in the allele/genotype or haplotype analyses (global Pall markers = 0.669) (Tables 1 and 3).

Table 1.  Tagging SNPs and association analysis of SIRT1 gene
SNP ID*PhenotypeMAF n Genotype distribution*HWEGenotypeAlleleCorrected P-value
M/MM/mm/m χ 2 (2 df) P-value χ 2 (2 df) P-value χ 2 (1 df) P-valueGenotypeAllele
  1. M, major allele; m, minor allele.

  2. *Major allele > minor allele.

  3. Bold numbers represent significant uncorrected P-value.

  4. Calculated by Bonfferoni's correction (eight multiple tests).

rs 12778366Controls0.11521271669428300.1840.668
T>CSchizophrenia0.1211158900236220.1970.1611.210.5450.5550.456 
5′ flanking regionBipolar disorder0.1211008776220120.6710.4131.400.4970.5290.467  
rs2273773Controls0.33721279399432450.1250.724
T>CSchizophrenia0.33411585155131300.01750.8950.07440.9630.06440.800
Exon 5Bipolar disorder0.3131008468449911.290.2564.760.09263.530.0604
rs4746720Controls0.384212780710053150.005360.942
T>CSchizophrenia0.35211584985051552.230.1358.07 0.0177 6.75 0.00947 0.1420.0758
Exon 9 UTRBipolar disorder0.39710083565031491.770.1832.270.3220.9700.325
rs10997875Controls0.16921271459618502.680.102
T>CSchizophrenia0.1691158804317370.7000.4032.840.2420.00002000.996
3′ flanking regionBipolar disorder0.1791008683290350.3760.5403.260.1960.9210.337
Table 2.  Haplotype-wise analysis between SIRT1 gene and Japanese schizophrenia
SIRT1 gene, common haplotypes rs 12778366–rs2273773–rs 4746720–rs10997875PhenotypeIndividual haplotype frequency χ 2 (1 df)OR95% CIIndividual P-value*Corrected P-value*,
  1. OR, odds ratio; CI, confidential interval.

  2. *Bold numbers represent significant uncorrected P-value.

  3. Caluculated by Bonferroni correction (number of multiple test: maximum number of haplotypes).

T-T-T-TControls0.014677.11.001.00–1.00 1.63 ×10 17 8.15 ×10 17
 Schizophrenia0.0527     
T-T-T-CControls0.1630.5250.2640.189–0.3710.469 
 Schizophrenia0.156     
T-T-C-TControls0.3818.190.2500.181–0.345 0.00421 0.0211
 Schizophrenia0.345     
T-C-T-TControls0.3350.2690.2710.196–0.3750.604 
 Schizophrenia0.328     
C-T-T-TControls0.1072.020.3070.217–0.4340.155 
 Schizophrenia0.119     
Table 3.  Haplotype-wise analysis between SIRT1 gene and Japanese bipolar disorder
SIRT1 gene, common haplotypes rs 12778366–rs2273773–rs4746720–rs 10997875PhenotypeIndividual haplotype frequency χ 2 (1 df)OR95% CIIndividual P-value
  1. OR, odds ratio; CI, confidential interval.

T-T-T-TControls0.01462.171.001.00–1.000.782
 Bipolar disorder0.0155    
T-T-T-CControls0.1631.980.6230.416–0.9290.159
 Bipolar disorder0.178    
T-T-C-TControls0.3810.3260.5820.394–0.8590.568
 Bipolar disorder0.399    
T-C-T-TControls0.3352.730.53370.790–2.730.0984
 Bipolar disorder0.313    
C-T-T-TControls0.1071.920.5080.335–0.7710.166
 Bipolar disorder0.0951    

In the power analysis, we obtained more than 80% power for the detection of association when we set the genotype relative risk at 1.23 and 1.33 in schizophrenia and 1.24 and 1.34 in BP, respectively, for SIRT1 under a multiplicative model of inheritance.

Discussion

We first performed a gene-based association analysis among the SIRT1 gene, schizophrenia and BP in the Japanese population. We found a significant association between variants in the SIRT1 gene and schizophrenia in the Japanese population. Therefore, we suggest that SIRT1 gene plays a role in schizophrenia in the Japanese population. On the other hand, we did not detect any associations between the SIRT1 gene and Japanese BP patients in the allele, genotype and haplotype analyses. However, because our samples are rather small to detect the susceptibility genes for common complex disease such as BP and schizophrenia (Committee 2009), the possibility of a statistical error in our results exists. To overcome this limitation, it will be necessary to conduct a further study using larger samples in the future.

We detected some association between some haplotypes (the T-T-T-T haplotype and the T-T-C-T haplotype on rs12778366–rs2273773–rs4746720–rs10997875) in the SIRT1 gene and schizophrenia. We also found a marginal association between rs4746720 in the SIRT1 gene and schizophrenia in the allele analysis (corrected P = 0.0758). Based on these results, we thought that some of the association between these haplotypes in the SIRT1 gene and schizophrenia in this study might be reflected in the genotype on rs4746720.

Rs4746720 is located in the exon 9 UTR in the SIRT1 gene. Since we did not perform mutation screening, however, a further study will be necessary. Several investigations have suggested that SIRT1 is related to dopaminergic neural transmission (Green et al. 2008; Liu et al. 2009). Abnormalities in the dopamine neural transmission are known to be involved in the pathophysiology of schizophrenia (Lang et al. 2007). If the variations in the SIRT1 gene associated with schizophrenia in this study have a biologically functional effect, the mechanism may be a disruption of circadian rhythms and abnormalities in the dopamine neural transmission, leading to the development of schizophrenia.

A few points of caution should be mentioned with respect to our results. First, because our samples were small for a case–control study, statistical errors are possible in the results of these association analyses. Most of our subjects (1016 schizophrenia patients, 997 BP patients and 1950 controls) did not undergo structured interviews. However, in this study patients were carefully diagnosed according to DSM-IV criteria with consensus of at least two experienced psychiatrists on the basis of a review of medical records. When we found a patient who had been misdiagnosed, we promptly excluded the misdiagnosed case to maintain the precision of our sample. Second, we did not perform a mutation scan of SIRT1 gene. Because we consider it to be difficult to evaluate the association of extremely rare variants from the viewpoint of statistical power, a replication study using a larger sample will be required for conclusive results.

In conclusion, our results suggest that the SIRT1 gene may play a role in the pathophysiology of schizophrenia in the Japanese population. Because our samples were relatively small, statistical errors are possible in the results of these association analyses. To overcome these limitations, a replication study using a larger sample may be required for conclusive results.

Acknowledgments

We thank Ms M. Miyata and Ms S. Ishihara for their technical support. This work was supported in part by research grants from the Ministry of Education, Culture, Sports, Science and Technology (Japan), the Ministry of Health, Labor and Welfare and the Health Sciences Foundation (Research on Health Sciences focusing on Drug Innovation).

Ancillary