Significant outcomes
-
∙ Among 23 body mass index (BMI)-related single nucleotide polymorphisms (SNPs), we identified three SNPs showed significantly associated with psychiatric diseases.
-
∙ Six SNPs showed a significant difference on at least one of subscales of Wechsler Memory Scale – Revised (WMS-R), depending on the genotype in healthy controls.
-
∙ Carrying the A allele of rs11142387 near the Kruppel-like factor 9 (KLF9) was significantly associated with both psychiatric disease and poor memory function.
Limitations
-
∙ Caution is required in the interpretation of the results because of possible chance findings due to multiple testing.
Introductions
There were many reports showing that obesity is a risk factor for psychiatric diseases such as major depressive disorder (MDD), bipolar disorder, and schizophrenia (Reference Stunkard, Faith and Allison1–Reference Mansur, Brietzke and Mcintyre3). In a meta-analysis of 15 studies, obese people had an increased risk (55%) of developing depression. In addition, depressed people had an increased risk (58%) of becoming obese (Reference Luppino, De Wit and Bouvy4), indicating a bidirectional association between the two conditions. A similar bidirectional relationship has also been found between depression and metabolic syndrome (Reference Pan, Keum and Okereke5). An association between obesity and bipolar disorder has also been reported with 58% of bipolar patients being overweight, 21% obese (BMI≥30), and 5% extremely obese (BMI≥40) (Reference Mcelroy, Frye and Suppes6). The obesity (i.e. higher BMI) was related to the severity of bipolar disorder, as obese patients with bipolar disorder had a greater number of lifetime affective episodes, history of suicide attempts, and more severe and difficult-to-treat episodes (Reference Fagiolini, Kupfer, Houck, Novick and Frank7,Reference Calkin, Van De Velde and Ruzickova8). Furthermore, schizophrenia is also a risk factor for obesity and metabolic syndrome (Reference Deh, Schreurs, Vancampfort and Vanw9). In the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study, 35.8% of patients with schizophrenia had metabolic syndrome (Reference Meyer, Nasrallah and Mcevoy10). In a meta-analysis of 77 studies, the overall rate of metabolic syndrome was 32.5% in patients with schizophrenia, and was the strongest influence on illness duration (Reference Mitchell, Vancampfort, Sweers, Van Winkel, Yu and De Hert11). In Japan, the rates of metabolic syndrome in outpatients with schizophrenia are 48.1% (Reference Sugawara, Yasui-Furukori and Sato12). In addition, several reports have shown the association between BMI and memory performance (Reference Gunstad, Paul, Cohen, Tate and Gordon13,Reference Kumari, Brunner and Fuhrer14), which is often impaired in MDD, bipolar disorder, and schizophrenia (Reference Baune, Fuhr, Air and Hering15–Reference Gur and Gur17). Higher BMI has been reported to be associated with reduced whole brain volume, including hippocampal atrophy (Reference Bruce-Keller, Keller and Morrison18,Reference Jagust, Harvey, Mungas and Haan19).
A number of SNPs have been reported to be associated with BMI (Reference Wen, Cho and Zheng20–Reference Okada, Kubo and Ohmiya22). Psychiatric diseases and metabolic abnormalities are thought to be influenced by multiple risk genes (Reference Xia and Grant23,Reference Flint and Kendler24), and some genes may give susceptibility to both conditions (Reference Afari, Noonan and Goldberg25,Reference Samaan, Anand and Zhang26). The FTO (fat mass and obesity associated) gene, for example, has been shown to contribute to both obesity and depressive symptoms (Reference Rivera, Cohen-Woods and Kapur27). However, the causal pathways and mechanisms linking physical and psychiatric conditions have not been thoroughly examined.
Recently, several common genes across major mental disorders have been suggested (Reference Gatt, Burton, Williams and Schofield28,29). These common genes might share the risk of psychiatric disorders, including MDD, schizophrenia, and bipolar disorder. On the other hand, all of these psychiatric diseases indicate an association with obesity or metabolic syndrome as mentioned above, which prompted us to assume genetic risk for obesity might be the shared risk for MDD, schizophrenia, and bipolar disorder.
Aims of the study
The aim of the present study was to clarify the genetic mechanisms linking physical and psychiatric conditions. We investigated BMI-related SNPs for an association with psychiatric disease (MDD, bipolar disorder, and schizophrenia) and memory performance in a Japanese population. We hypothesised that some of the genetic variants associated with obesity may also be risk factors for psychiatric disease and poor memory performance.
Material and methods
Subjects
The subjects were 1189 healthy controls (404 men and 785 women; mean age=45.8 years, SD=16.2 years) and 1624 patients with a psychiatric disorder (801 men and 823 women; mean age=46.6, SD=15.1). There were significant differences between patients and controls in sex ratio (χ2-test, p<0.0001) and BMI (controls, mean BMI=22.3, SD=3.2; patients, mean=23.5, SD=4.9; t-test, p=0.0004), but not in age (t-test, p=0.19). The patient group consisted of 799 patients with MDD, 594 with schizophrenia, and 231 with bipolar disorder. The clinical information is shown in Supplementary file 1. All subjects were biologically unrelated Japanese people who were recruited from the outpatient clinic of the National Center of Neurology and Psychiatry (NCNP) Hospital, Tokyo, Japan, through advertisements in free local information magazines, or through an announcement on our website. Consensus diagnosis by at least two psychiatrists was made for each patient according to the Diagnostic and Statistical Manual of Mental Disorders, 4th edition criteria (30), on the basis of the Japanese version of the Mini International Neuropsychiatric Interview (M.I.N.I.) (Reference Otsubo, Tanaka and Koda31,Reference Sheehan, Lecrubier and Sheehan32), additional unstructured interviews, and information from medical records if available. The controls were healthy volunteers with no current or past history of psychiatric disorder. They were screened using the M.I.N.I. by a research psychiatrist to rule out any axis I psychiatric disorders. Participants were excluded if they had prior medical history of central nervous system disease or severe head injury, or if they met the criteria for substance abuse or dependence, or mental retardation. The study protocol was approved by the ethics committee at NCNP (A2013-132 and A2015-138). After receiving a full description of the study, written informed consent was obtained from every subject.
Neurocognitive test battery
To measure memory function in the healthy controls, we employed the WMS-R (Reference Wechsler33,Reference Sugishita34), which yields five subscale scores, verbal memory, visual memory, general memory, attention/concentration, and delayed recall.
Genotyping
Genomic DNA was prepared from venous blood according to standard procedures. We genotyped 23 BMI-related SNPs (Supplementary file 2) using the TaqMan 5′-exonuclease allelic discrimination assay. Thermal cycling conditions for polymerase chain reaction were 1 cycle at 95 °C for 10 min followed by 50 cycles of 92 °C for 15 s and 60 °C for 1 min. Allele-specific fluorescence was measured using ABI PRISM 7900 Sequence Detection Systems (Applied Biosystems, Foster City, CA, USA). Genotype data were read blind to the case control status. Ambiguous genotype data were not included in the analysis.
Statistical analysis
Deviation of genotype distributions from the Hardy–Weinberg equilibrium (HWE) was assessed using the χ2-test for goodness of fit. Genotype and allele distributions were compared between patients and controls using the χ2-test for independence. Memory performance and BMI in psychiatric patients are known to be greatly influenced by medication and illness duration. Therefore, in this study, we analysed the association between SNPs, memory, and BMI in controls only. We examined the effect of genotype on BMI and WMS-R scores, using multiple regression analysis to control for age and gender. All statistical tests were two-tailed, and p<0.05 indicated statistical significance. Analyses were performed using IBM SPSS Statistics Version 22 Japanese edition (IBM SPSS Japan, Tokyo, Japan).
Results
Association with psychiatric diseases
Genotype and allele frequencies for the 23 SNPs in the patients and controls are shown in Supplementary file 2. Genotype distributions for three SNPs (rs2568958, rs6567160, and rs11671664) significantly deviated from HWE in the controls, which we attributed to genotyping errors. We therefore excluded these three SNPs from subsequent analyses. Of the 20 SNPs, three (rs11142387, rs12597579, and rs6548238) had significantly different genotype and/or allele frequencies between the combined psychiatric disease patient group and the controls. We then examined these three SNPs for any association with each psychiatric disorder (i.e. schizophrenia, bipolar disorder, or MDD). The frequencies of the other 17 SNPs (rs516636, rs574367, rs2815752, rs13034723, rs6545814, rs9356744, rs2206734, rs2030323, rs2074356, rs10150332, rs4776970, rs1121980, rs9939609, rs1558902, rs12149832, rs17817449, and rs2287019) were not significantly different between the psychiatric disease patients and the controls.
For rs11142387, there was the most significantly different with the frequency of the CC genotype being lower in patients with psychiatric disease (p=0.007), especially schizophrenia (p=0.002) and bipolar disorder (p=0.031), than in controls (Table 1). There was also a significant difference in the allele frequency between the patients with bipolar disorder and controls (p=0.014). There were no significant differences in the genotypes or allele frequencies of patients with MDD and controls. Additional analyses using autosomal dominant and recessive models are shown in Supplementary file 3. The A allele (dominant model) was more common in patients with psychiatric disease (particularly schizophrenia and bipolar disorder, but not MDD) than in controls.
Table 1 Genotype and allele frequencies for rs11142387 in patients with psychiatric disease and healthy controls
MDD, major depressive disorder.
a Significant difference between patients and controls using the χ2-test.
For rs12597579, the frequency of the T allele was significantly higher in the total patient group (p=0.032), particularly in those with MDD (p=0.042), than in the controls (Table 2). There were no significant differences in the genotypes or allele frequencies of patients with schizophrenia or bipolar disorder, and controls.
Table 2 Genotype and allele frequencies of rs12597579 for patients with psychiatric diseases and healthy controls
MDD, major depressive disorder.
a Significant difference between patients and controls using the χ2-test.
For rs6548238, the frequency of the CC genotype was significantly higher in the patients with any psychiatric disorder (p=0.019) than in controls (Table 3). The frequency of the C allele was significantly higher in patients with psychiatric diseases (p=0.005), especially in those with MDD (p=0.033) or schizophrenia (p=0.028), than in controls. There were no significant differences in the genotypes or allele frequencies of patients with bipolar disorder and controls.
Table 3 Genotype and allele frequencies for rs6548238 in patients with psychiatric diseases and healthy controls
MDD, major depressive disorder.
a Significant difference between patients and controls using the χ2-test.
Association with memory performance in healthy controls
Next we tested the possible association between the 20 SNPs and memory performance in healthy controls (n=716) using multiple regression analysis, controlling for age and gender. As a result, six SNPs showed genotype related significant differences in at least one subscale of the WMS-R (Supplementary file 4). Interestingly, two SNPs (rs11142387 and rs12597579) overlapped in their association with psychiatric disorder and memory performance.
For rs11142387, verbal memory and general memory scores showed significant differences between genotype groups, that is, individuals who carried the A allele (AA/AC) showed lower verbal memory (p=0.024) and general memory (p=0.038) scores than their CC counterparts (Fig. 1a). In addition, individuals with the AA genotype had lower verbal memory (p=0.002), general memory (p=0.010) and attention/concentration (p=0.006) scores than C carriers (CC/CA; Fig. 1b). For rs12597579, there were significant differences in verbal memory (p=0.007), general memory (p=0.026), and delayed recall (p=0.034) scores between genotype groups. Individuals with the CC genotype had lower scores than T carriers (TT/TC; Fig. 1c). For rs2815752, there was a significant difference in delayed recall scores between genotype groups (p=0.038). Individuals carrying the A allele (AA/AG) had lower scores than their GG counterparts (Fig. 1d). For rs2074356, verbal memory (p=0.038) and general memory (p=0.032) scores were significantly different between the genotype groups. A allele carriers (AA/AG) had lower scores than their GG counterparts (Fig. 1e). For rs4776970, verbal memory scores showed significant differences between genotype groups (p=0.006). Individuals with the TT genotype had lower scores than A allele carriers (AA/AT; Fig. 1f). For rs2287019, verbal memory (p=0.009), general memory (p=0.008) and delayed recall (p=0.046) scores were significantly different between genotype groups. Individuals with the CC genotype had lower scores than T carriers (TT/TC; Fig. 1g).
Fig. 1 Mean Wechsler Memory Scale – Revised (WMS-R) index scores for each SNP by genotype. (a) The rs11142387 single nucleotide polymorphism (SNP) A allele carriers (AA/AC) had lower verbal memory (p=0.024) and general memory scores (p=0.038) than their CC counterparts. (b) C carriers (CC/CA) had higher verbal memory (p=0.002), general memory (p=0.010), and attention/concentration scores (p=0.006) than their AA counterparts. (c) The allele T carriers (TT/TC) of the rs12597579 SNP had higher verbal memory (p=0.007), general memory (p=0.026), and delayed recall scores (p=0.034) than their CC counterparts. (d) The allele A carriers (AA/AG) of the rs2815752 SNP had lower delayed recall scores (p=0.038) than their GG counterparts. (e) The rs2074356 SNP A allele carriers (AA/AG) had lower verbal memory (p=0.038) and general memory (p=0.032) scores than their GG counterparts. (f) The rs4776970 SNP A allele carriers (AA/AT) had higher verbal memory scores (p=0.006) than their TT counterparts. (g) The rs2287019 SNP T allele carriers (TT/TC) had higher verbal memory (p=0.009), general memory (p=0.008), and delayed recall (p=0.046) scores than their CC counterparts. All statistical analyses examined the effect of genotype on WMS-R scores using multiple regression analyses, controlling for age and gender. Error bars indicate SD. *p<0.05, **p<0.01.
Association with BMI in healthy controls
We examined the genotype association for the 20 SNPs with BMI in the 298 controls for whom we had information on BMI, using multiple regression analysis and controlling for age and gender. We found that only two SNPs (rs2206734 and rs11142387) were significantly associated with BMI (Supplementary file 5). The C alleles of rs2206734 and rs11142387 were associated with higher BMI. The directions of associations for these SNPs were the same as that reported previously (Reference Okada, Kubo and Ohmiya22).
Discussion
Based on a possible link between metabolic function and psychiatric disease, we examined 23 BMI-related SNPs for an association with psychiatric disease. Of 23 BMI-related SNPs, three showed a significant difference in the genotype and/or allele frequency between the combined patient group and the healthy controls. Next, we tested the possibility of an association between the SNPs and memory performance in the healthy controls. As a result, six SNPs had significant differences in at least one subscale of the WMS-R depending on genotype. Notably, two SNPs (rs11142387 and rs12597579) overlapped in their association with psychiatric disorder and memory performance. Our main findings are summarised in Table 4.
Table 4 Significant associations between psychiatric disease and memory performance
SNP, single nucleotide polymorphism; BMI, body mass index; MDD, major depressive disorder; KLF9, Kruppel-like factor 9; GP2, glycoprotein 2; TMEM18, transmembrane protein 18; NEGR1, neuronal growth regulator 1; HECTD4, HECT domain containing E3 ubiquitin protein ligase 4; MAP2K5, mitogen-activated protein kinase kinase 5; QPCTL, glutaminyl-peptide cyclotransferase-like; GIPR, gastric inhibitory polypeptide receptor.
We found that rs11142387 and rs12597579 were associated with both psychiatric disease and memory performance. The rs11142387, in particular, showed the most significant results and is located close to KLF9, which encodes Kruppel-like factor 9. KLF9 is one of the transcriptional regulators and is widely expressed, including in the brain and especially in the cerebellum and hippocampus (Reference Pearson, Fleetwood, Eaton, Crossley and Bao35). KLF9 has been shown to regulate transcription for uterine endometrial cell proliferation, and also plays a key role in the development of neurons in the brain (Reference Simmen, Su, Xiao, Zeng and Simmen36–Reference Lebrun, Avci and Wehrle38). KLF9 has been suggested to regulate oligodendrocyte differentiation and myelin regeneration, promoting myelin repair after cuprizone-induced demyelinations in mice (Reference Dugas, Ibrahim and Barres39). These neuronal or myelin-related events are important in psychiatric disease and memory, as psychiatric conditions have been shown to occur when normal myelin development is interrupted (Reference Mighdoll, Tao, Kleinman and Hyde40,Reference Fields41). One antipsychotic drug, quetiapine, showed similar action to that of KLF9, enhancing oligodendrocyte regeneration and myelin repair after cuprizone-induced demyelination in mice (Reference Zhang, Zhang and Wang42). Myelin is also reported to relate to memory function as cuprizone-induced memory dysfunction in mice was mitigated, whereas myelin was repaired (Reference Mighdoll, Tao, Kleinman and Hyde40).
In our study, we found that carrying the A allele of rs11142387 is associated with a high frequency of psychiatric disease and two domains of poor memory performance (verbal memory and general memory). Contrary to our expectation, however, the rs11142387 A allele has been reported to be associated with lower, rather than higher, BMI in the East Asian population (Reference Okada, Kubo and Ohmiya22). In our sample, similar results were obtained. So it is unlikely that the possible link between rs11142387 and psychiatric disease and memory function is explained by the putative association between high BMI and psychiatric/cognitive disturbances. It is possible that carrying the A allele of rs11142387 may have detrimental effects on neurodegeneration and/or demyelination, which then confers a risk of developing psychiatric disease and having poor memory performance.
We found that carrying the T allele of rs12597579 was more frequent in the psychiatric diseases studied, and was associated with high memory performance, that is, the direction of association was inverse. It is located close to GP2, which encodes glycoprotein 2 (GP2). GP2 is mainly expressed in the pancreas (Reference Hoops, Ivanov, Cui, Colomer-Gould and Rindler43), and plays an important role in mucous membrane immunosurveillance to reject foreign bodies (Reference Gomez-Lazaro, Rinn, Aroso, Amado and Schrader44,Reference Somma, Ababneh and Ababneh45). The C allele of rs12597579 has been reported to be associated with greater BMI in the East Asian population (Reference Wen, Cho and Zheng20). Our results showed a similar trend, but did not reach statistical significance. To our knowledge, it is not known whether GP2 is expressed in the brain, and there have been no reports of GP2 being associated with psychiatric disease or memory function. Collectively, our findings relating to rs12597579 may be due to chance, and further studies are required.
The rs6548238 SNP had significantly different genotype and allele frequencies between patients with psychiatric disease and controls. It is located close to TMEM18, which encodes transmembrane protein 18 (TMEM18). Differential methylation of the TMEM18 promoter in adipose tissue has been reported to be associated with adipose tissue depot specificity (Reference Thorleifsson, Walters and Gudbjartsson46,Reference Rohde, Keller and Klos47). In the brain, it is widely expressed in neurons in all major brain regions (Reference Almen, Jacobsson and Shaik48) and modulates the migration of neural stem cells (Reference Jurvansuu, Zhao and Leung49).
The C allele of rs6548238 was reported to be associated with obesity (BMI≥30) in the Japanese population (Reference Hotta, Nakamura and Nakamura50). Unfortunately, our data did not reveal similar results; however, this might be attributable to the small number of subjects in our study whose BMI≥30. We found that the frequency of the CC genotype or C allele in rs6548238 was significantly higher in patients with psychiatric disease, which is in line with our hypothesis that higher BMI is a risk for psychiatric disease.
In our analysis, four SNPs (rs2815752, rs2074356, rs4776970, and rs2287019) had genotype-dependent significant differences in at least one subscale of the WMS-R, but no significant association with the psychiatric disorders investigated.
The rs2815752 SNP is located close to NEGR1, which encodes neuronal growth regulator 1 (NEGR1) or neurotractin. NEGR1 is widely expressed in the brain, especially in the hypothalamus, and loss of function leads to decreased body weight in mice, and decreases in cell adhesion and neurite outgrowth in vitro (Reference Marg, Sirim and Spaltmann51,Reference Lee, Hengstler and Schwald52). The A allele of rs2815752 has been reported to be associated with higher BMI (Reference Speliotes, Willer and Berndt53,Reference Willer, Speliotes and Loos54); however, our data did not show this.
Interestingly, the rs2815752 risk allele (A) for higher BMI has been associated with lower white matter integrity across a substantial portion of the brain in healthy young adults (Reference Dennis, Jahanshad and Braskie55). In our study, individuals with the A allele had lower delayed recall scores than their counterparts. As higher BMI has been reported to be associated with memory deficits (Reference Gunstad, Paul, Cohen, Tate and Gordon13,Reference Kumari, Brunner and Fuhrer14) and a reduction in brain volume (Reference Bruce-Keller, Keller and Morrison18,Reference Jagust, Harvey, Mungas and Haan19), the A allele of rs2815752 might play a role in the mechanism relating higher BMI to smaller brain volume and memory deficits.
We found that the A allele of rs2074356, which is reportedly the risk allele for higher BMI (Reference Kim, Go and Hu56), was associated with lower verbal memory and general memory. Rs2074356 is located close to HECTD4, which encodes HECT domain containing E3 ubiquitin protein ligase 4 (HECTD4). Although some polymorphisms of HECTD4 have been reported to be associated with metabolic diseases in Asia (Reference Kim, Go and Hu56,Reference Heo, Hwang and Uhmn57), the functions of HECTD4 remain largely unknown. Further studies are required to elucidate the role of HECTD4 in brain function.
For rs4776970, it is located close to MAP2K5, which encodes mitogen-activated protein kinase kinase 5 (MAP2K5). MAP2K5 is an enzyme that belongs to the MAP kinase kinase family, and is widely expressed in various tissues including the brain. It interacts with and activates MAPK7 (Reference Drew, Burow and Beckman58). The conditional knockout of MAPK7 mice in adult neurogenic regions of the mouse brain attenuates adult neurogenesis in the hippocampus and disrupted memory function, which is improved by MAPK7 activation (Reference Wang, Pan and Zou59). We found that the A allele of rs4776970, which is the risk allele for higher BMI (Reference Wen, Cho and Zheng20), was associated with higher, rather than lower, verbal memory.
The rs2287019 SNP is located close to QPCTL and GIPR, which encode glutaminyl-peptide cyclotransferase-like (QPCTL) and gastric inhibitory polypeptide receptor (GIPR), respectively. QPCTL is an isoenzyme of glutaminyl-peptide cyclotransferase (QPCT) and widely expressed in various tissues. QPCTL converts the formation of glutamate to pyroglutamate at the N-terminus of various peptides, such as chemokine (C-C motif) ligand 2 (CCL2) (Reference Becker, Eichentopf and Sedlmeier60,Reference Saido, Iwatsubo, Mann, Shimada, Ihara and Kawashima61). GIPR is an important incretin hormone receptor. GIPR mediated signalling also plays an important role in the deposition of excess fat from the diet into adipose tissue (Reference Flatt62,Reference Irwin and Flatt63). GIPR knockout mice results in deficits in cognitive function and reduced hippocampal expression of mammalian target of rapamycin (mTOR) (Reference Lennox, Moffett, Porter, Irwin, Gault and Flatt64). We found that carrying the T allele of rs2287019 was associated with higher performance in verbal memory, general memory, and delayed recall. Its counter allele (C) has been associated with higher BMI (Reference Gomez-Lazaro, Rinn, Aroso, Amado and Schrader44), but our data were not in agreement with this. The T allele of rs2287019 may be associated with a loss of function of GIPR, which may result in poorer memory performance.
Of the seven SNPs in Table 4, the results of four (rs6548238, rs2815752, rs2074356, and rs2287019) agreed with our hypothesis that genetic variants associated with obesity may also confer a risk for psychiatric disease and/or poor memory performance. These are new findings suggesting a genetic commonality for obesity and brain function. Because the function of some SNPs and their nearest gene remain unclear, further studies are warranted.
However, for the other three SNPs (rs11142387, rs12597579, and rs4776970), the risk alleles for obesity did not increase the risk for psychiatric disease or poor memory performance. For example, for rs11142387 the genetic effect might have different mechanisms for increasing the risk for obesity (promoting differentiation of fat cells) and psychiatric disease or memory performance (involvement in myelination).
There are several limitations to our study. First, we made multiple comparisons, so some of the observed associations might be attributable to chance. As we examined 20 SNPs for their association with psychiatric disease, the Bonferroni corrected p-value is 0.05/20=0.0025. We also examined the association of 20 SNPs on five subscales of the WMS-R, with a corrected p-value of 0.0005. In these cases the results did not retain significance if correction was applied. However, for rs11142387, the same genotype was found to be associated with schizophrenia, bipolar disorder, and three subscales of memory performance, which is unlikely to be due to chance.
Second, of the 20 BMI-related SNPs, we found that only two were significantly associated with BMI in healthy controls. One of reasons for this may be the difference in the sample sizes. Our sample size was much smaller than in previous studies (Reference Wen, Cho and Zheng20,Reference Okada, Kubo and Ohmiya22,Reference Speliotes, Willer and Berndt53) and may not be sufficient to examine an association between BMI and the genotype frequencies of the 20 SNPs. Furthermore, owing to the size of our sample type II errors may have occurred in the psychiatric disease and memory function association analyses.
Third, there were few obese subjects in our study. Some SNPs were analysed specifically in obese subjects and showed significant differences in a previous study (Reference Hotta, Nakamura and Nakamura50). However, in our healthy controls, only five subjects were obese or extremely obese (BMI≥30).
Acknowledgements
The authors thank the patients and the healthy volunteers for their participation. Moeko Hiraishi had been greatly supportive to perform the cognitive tests for memory performance. Authors’ contributions: M.N. and H.K. designed the study. H.K. Organised recruitment of the subjects and H.K., H.H., M.O., K.H., T.T., D.S., I.I., and T.N. contributed to collection of psychiatric data. J.M. performed the cognitive tests for memory performance. M.N. performed the SNP genotyping. M.N. performed the statistical analysis, managed the literature search, and wrote the draft of the manuscript. H.K. and S.S. supervised the entire project and gave critical comments to the manuscript. All authors contributed to and have approved the final manuscript.
Financial Support
This study was supported by Intramural Research Grant for Neurological and Psychiatric Disorders of the National Center of Neurology and Psychiatry Hospital, Tokyo, Japan.
Conflicts of Interest
None.
Ethical Standards
The authors assert that all procedures contributing to this work comply with the ethics committee of the NCNP and with the Helsinki Declaration of 1975, as revised in 2008. After receiving a full description of the study, written informed consent was obtained from every participant.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/neu.2016.66
